paperID stringlengths 36 36 | pwc_id stringlengths 8 47 | arxiv_id stringlengths 6 16 ⌀ | nips_id float64 | url_abs stringlengths 18 329 | url_pdf stringlengths 18 742 | title stringlengths 8 325 | abstract stringlengths 1 7.27k ⌀ | authors stringlengths 2 7.06k | published stringlengths 10 10 ⌀ | conference stringlengths 12 47 ⌀ | conference_url_abs stringlengths 16 198 ⌀ | conference_url_pdf stringlengths 27 199 ⌀ | proceeding stringlengths 6 47 ⌀ | taskID stringlengths 7 1.44k | areaID stringclasses 688 values | embedding stringlengths 9.26k 12.5k | umap_embedding stringlengths 29 44 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
aea0940e-f59d-4b1b-b597-9b1bfd0ed6b7 | fisherposes-for-human-action-recognition | null | null | https://ieeexplore.ieee.org/abstract/document/8219411 | http://sharif.edu/~hoda/papers/action_sensors.pdf | Fisherposes for Human Action Recognition Using Kinect Sensor Data | This paper proposes a new method for view-invariant action recognition that utilizes the temporal position of skeletal joints obtained by Kinect sensor. In this method, the actions are represented as sequences of several pre-defined poses. After pre-processing, which includes skeleton alignment and scaling, the appropriate feature vectors are obtained for recognizing and discriminating the pose of every frame by the proposed Fisherposes method. The proposed regularized Mahalanobis distance metric is used in order to recognize both the involuntary and highly made-up actions at the same time. Hidden Markov model (HMM) is then used to classify the action related to an input sequence of poses. For taking into account the motion in the actions which are not separable by solely their temporal poses, histograms of trajectories are also proposed. The proposed action recognition method is capable of recognizing both the voluntary and involuntary actions, as well as pose-based and trajectory-based ones with a high accuracy rate. The effectiveness of the proposed method is experimented on three publicly available data sets, TST fall detection, UTKinect, and UCFKinect data sets. | ['Mozhgan Mokari', 'Hoda Mohammadzade', 'Benyamin Ghojogh'] | 2018-02-15 | null | null | null | null | ['3d-human-action-recognition'] | ['computer-vision'] | [ 3.4337252e-01 -5.7138079e-01 -3.0029452e-01 -3.2504696e-01
-5.7406276e-01 -2.4055798e-01 4.9709624e-01 -2.0526898e-01
-7.7833056e-01 5.8633381e-01 2.8719768e-01 3.6750516e-01
-4.3451533e-01 -5.2915144e-01 -1.6636252e-01 -9.2055380e-01
-1.6286870e-02 5.7332337e-01 3.6730751e-01 -6.2252991e-02
3.9151093e-01 7.8847736e-01 -1.8718330e+00 1.2540743e-01
3.8623258e-01 8.4913242e-01 6.3861258e-02 6.4498001e-01
2.7486151e-01 4.7762683e-01 -3.5927758e-01 1.5824687e-01
3.2427007e-01 -6.7185432e-01 -5.0641763e-01 5.6880021e-01
8.9096427e-02 -4.5694369e-01 -2.1932450e-01 7.2163206e-01
3.6975530e-01 5.2898341e-01 8.7397933e-01 -9.8290545e-01
1.3080399e-01 -1.1985836e-01 -4.3191126e-01 3.4697095e-01
6.9856369e-01 1.3698857e-01 6.7569882e-01 -7.4210119e-01
6.2371069e-01 1.0694298e+00 2.5591502e-01 4.3492767e-01
-8.7104553e-01 -4.5065048e-01 -2.3781556e-01 8.0512035e-01
-1.5097327e+00 -2.5267318e-01 8.4174949e-01 -6.2837744e-01
7.9596007e-01 4.2332938e-01 8.8324994e-01 9.5600897e-01
4.6527958e-01 6.4856029e-01 1.1156971e+00 -3.7775737e-01
3.0791965e-01 -5.6135315e-01 2.6247475e-01 5.3169924e-01
1.5929408e-01 2.4117082e-02 -5.2798289e-01 -7.1190372e-02
6.2542385e-01 5.5893189e-01 -2.1632987e-01 -4.4216156e-01
-1.1198988e+00 5.2070552e-01 -5.1373459e-02 5.6127596e-01
-6.9580936e-01 -2.1378876e-01 4.1884747e-01 -1.2157263e-01
1.2525508e-01 -4.6709847e-01 -1.4456558e-01 -5.9579331e-01
-8.9169019e-01 8.9512192e-02 4.3486598e-01 6.3387281e-01
5.0990486e-01 -1.0462831e-01 -9.5776722e-02 3.5086262e-01
3.2169718e-01 5.9556293e-01 8.7258977e-01 -5.7224566e-01
4.9679381e-01 8.8271755e-01 1.4959833e-01 -1.0524883e+00
-5.1259494e-01 1.9267014e-01 -5.0253183e-01 2.6459906e-01
5.6221265e-01 1.9265521e-01 -9.7743130e-01 1.2508961e+00
7.1573281e-01 3.1494360e-02 1.3787876e-02 1.0194031e+00
3.4690136e-01 2.5112811e-01 -5.5779491e-02 -5.2400136e-01
1.4010183e+00 -4.8467371e-01 -8.9204895e-01 5.9338383e-02
4.3465248e-01 -5.9820062e-01 7.0984483e-01 4.6054685e-01
-5.7497925e-01 -6.2016839e-01 -9.6990204e-01 2.1888779e-01
-1.3429980e-01 4.5949385e-01 1.1240598e-01 5.6222987e-01
-2.7954629e-01 6.9942969e-01 -1.6068691e+00 -6.3464642e-01
-7.8178443e-02 3.1446809e-01 -7.4276263e-01 1.2833859e-01
-7.0537162e-01 8.9847195e-01 4.4845939e-01 4.1192871e-01
-8.0326116e-01 4.6187487e-01 -8.5032207e-01 -3.1672972e-01
3.0049038e-01 -9.7436704e-02 6.8787539e-01 -8.5051507e-01
-1.4640011e+00 6.3119388e-01 -2.7724198e-01 -2.3923162e-01
7.2721970e-01 -2.8218198e-01 -4.7388566e-01 4.8864672e-01
2.2197884e-02 -3.5914254e-01 9.6701926e-01 -5.5299777e-01
-4.7965199e-01 -1.0171341e+00 -2.9635972e-01 6.4341652e-01
-2.8902185e-01 3.7565872e-02 -3.8688645e-01 -2.7493697e-01
6.8930769e-01 -1.0459963e+00 8.2787313e-02 -1.6874868e-01
-1.9510691e-01 -1.9347137e-01 7.5540096e-01 -1.0187038e+00
1.1142077e+00 -2.2837951e+00 4.7558448e-01 1.8536884e-01
-1.3851321e-01 1.6656913e-01 3.5770932e-01 4.9205223e-01
6.0286336e-02 -6.5180498e-01 -2.2745426e-01 -9.2240393e-02
-2.8555968e-01 6.3713479e-01 4.2776668e-01 8.7063783e-01
-2.6670584e-01 2.1077332e-01 -6.5769124e-01 -7.3874420e-01
6.3709927e-01 4.1392449e-01 -1.1257617e-03 2.9779333e-01
2.7887437e-01 7.1336973e-01 -6.0764527e-01 6.2491614e-01
3.2155919e-01 5.4630101e-01 3.5366827e-01 -2.4913052e-01
-7.5175367e-02 -1.0688976e-01 -1.5053430e+00 1.4968693e+00
5.7117671e-02 1.8899131e-01 -1.2707403e-01 -9.9337739e-01
9.3945771e-01 5.0908852e-01 8.4861368e-01 -4.5705584e-01
3.4227964e-01 2.8572562e-01 -1.4997702e-02 -8.2046038e-01
2.2367254e-01 -2.0696928e-01 1.8612310e-01 3.3319759e-01
5.9161149e-02 5.4789364e-01 3.3440048e-01 -4.3307272e-01
1.0929708e+00 5.3205341e-01 6.8967456e-01 1.0853107e-01
7.1152794e-01 -1.6135487e-01 6.0317236e-01 7.6762989e-02
-4.5583037e-01 5.6008691e-01 -5.8734724e-03 -4.2061955e-01
-6.5290087e-01 -1.0063576e+00 8.0338657e-02 6.4795142e-01
1.7572486e-01 -1.8761267e-01 -7.1238905e-01 -6.9960296e-01
1.7410152e-02 4.3348214e-01 -6.4551806e-01 -3.5212159e-01
-6.8786931e-01 -4.0650156e-01 3.7267640e-01 6.9833803e-01
5.5699372e-01 -1.1639872e+00 -1.2403585e+00 1.4262888e-01
-3.1337327e-01 -8.5246575e-01 -3.4572998e-01 4.9074892e-02
-1.1813872e+00 -1.4846159e+00 -7.2570831e-01 -4.0808517e-01
5.8863479e-01 1.8294883e-01 1.7163065e-01 -1.8285596e-01
-4.4116181e-01 6.0822558e-01 -5.5422902e-01 1.9349121e-01
-1.8841656e-01 -3.7875536e-01 3.9349058e-01 6.3842934e-01
5.9180862e-01 -5.7172555e-01 -6.3309509e-01 6.0229611e-01
-7.8736734e-01 -2.8320831e-01 6.5211147e-01 5.5295038e-01
7.8675097e-01 1.7021552e-02 -3.3862248e-02 -1.4736906e-01
2.8517076e-01 -1.5565865e-01 -1.0258904e-01 3.0633005e-01
-4.3395439e-01 -6.2855775e-03 2.7240461e-01 -4.8009956e-01
-1.1401187e+00 5.4599220e-01 -1.7589673e-01 -4.2749679e-01
-6.7578083e-01 3.6042130e-01 -3.8815838e-01 1.3278890e-01
5.8006036e-01 6.2582260e-01 3.3064875e-01 -6.7774630e-01
8.4418103e-02 7.7544451e-01 6.0973197e-01 -3.9246085e-01
5.6326908e-01 7.2613442e-01 1.4126156e-01 -9.8138469e-01
-1.9221902e-01 -9.0819126e-01 -1.3763695e+00 -6.9344664e-01
1.2870044e+00 -4.1553274e-01 -6.6901869e-01 7.2233778e-01
-9.0097678e-01 3.4779969e-01 -1.2355808e-01 1.1104265e+00
-9.2408872e-01 8.3792877e-01 -4.9964720e-01 -1.2501558e+00
-2.2684582e-01 -9.5027786e-01 9.5714563e-01 4.6535045e-02
-3.0540481e-01 -6.5216869e-01 2.6705974e-01 5.5567575e-01
-2.4901795e-01 7.4023545e-01 4.7602490e-01 -7.1965134e-01
-1.4652425e-01 -7.1617180e-01 4.4606924e-01 4.0622568e-01
5.2603924e-01 -1.1027267e-01 -3.6854979e-01 -2.3792478e-01
3.9064655e-01 -7.3806994e-02 5.9347987e-01 1.6409311e-01
4.2129806e-01 -4.7225557e-02 -1.5888229e-01 2.7079561e-01
1.2430904e+00 6.5517056e-01 7.3395807e-01 2.9084575e-01
7.4480569e-01 3.8300163e-01 1.1570535e+00 5.7270545e-01
-2.6225006e-02 8.5088086e-01 4.7531655e-01 3.1623030e-01
2.2935714e-01 -5.2284678e-03 5.9653687e-01 7.3076493e-01
-1.0153639e+00 2.6933274e-01 -7.1965009e-01 2.7183107e-01
-2.0012612e+00 -1.2672716e+00 -2.7230358e-01 2.4808118e+00
4.8014185e-01 5.1946290e-02 3.8392133e-01 6.3472301e-01
8.0067497e-01 6.6385858e-02 -5.9316945e-01 -1.3023409e-01
2.3609535e-01 2.5225341e-01 4.8964414e-01 2.8027055e-01
-1.1338340e+00 4.2449114e-01 5.1816912e+00 7.5309449e-01
-9.4595391e-01 5.9207123e-02 -1.5638490e-01 -9.8485477e-02
4.2713854e-01 -1.8313549e-01 -5.2109516e-01 3.9790547e-01
7.5850940e-01 9.6145742e-02 9.1480918e-02 8.7266874e-01
4.8358384e-01 -5.3336138e-01 -8.9882475e-01 1.0214206e+00
2.7279836e-01 -4.8100489e-01 -1.2619819e-01 5.7816770e-02
5.9599556e-02 -3.8346463e-01 -2.8816560e-01 -8.1644431e-02
-5.2775145e-01 -5.9460384e-01 7.4443853e-01 9.1429001e-01
5.4806578e-01 -7.3255718e-01 6.9985217e-01 6.1524206e-01
-1.3885143e+00 -8.6879998e-02 -2.8658066e-02 -2.6379073e-01
2.1645187e-01 3.2800075e-02 -5.4249877e-01 9.4406223e-01
3.7339932e-01 8.7706655e-01 -4.2586705e-01 9.0074682e-01
-2.7382967e-01 5.3153807e-01 -3.4394854e-01 1.6923971e-02
-6.4369462e-02 -4.7700584e-01 7.2384191e-01 8.1931138e-01
4.3637717e-01 2.3090646e-01 2.2396311e-01 2.3958090e-01
8.8970369e-01 2.6579896e-01 -5.5944633e-01 -9.4939969e-02
-1.8197017e-02 9.9783725e-01 -9.1963673e-01 -8.6313903e-02
-2.0851338e-01 1.2854072e+00 -2.0381823e-01 1.5871041e-01
-7.5987691e-01 -1.5475024e-01 4.8152512e-01 1.5791263e-01
1.8436758e-01 -7.1481633e-01 2.2591329e-01 -1.1239200e+00
2.8588530e-01 -6.3403291e-01 6.0348105e-01 -3.8104191e-01
-7.0681494e-01 3.0635151e-01 3.4095588e-01 -1.7964864e+00
-4.7226438e-01 -5.1007640e-01 -3.3319160e-01 6.8730420e-01
-6.4046723e-01 -9.8538125e-01 -3.1657195e-01 9.8053777e-01
7.1686977e-01 4.3218970e-02 8.4901309e-01 2.5457528e-01
-6.2336534e-01 2.4819170e-01 1.5912649e-01 2.0621847e-01
4.9713829e-01 -8.8764918e-01 -2.5381589e-01 8.9435995e-01
2.4336301e-01 5.5672169e-01 7.3456848e-01 -1.0067468e+00
-1.3466667e+00 -6.6164094e-01 7.7629882e-01 -3.3134374e-01
3.4558329e-01 1.1952034e-01 -5.1036537e-01 7.2294635e-01
-5.4320872e-01 -6.4958110e-02 5.4786122e-01 -3.7796107e-01
4.9386416e-02 -1.6206825e-02 -1.1382966e+00 2.2912136e-01
9.5936835e-01 -4.7802538e-01 -9.5694029e-01 2.3444553e-01
-9.0929173e-02 -5.0922269e-01 -7.7549672e-01 3.5119656e-01
1.0114970e+00 -1.2469525e+00 7.7054143e-01 -5.3782868e-01
8.3243109e-02 -3.5230774e-01 -2.5989634e-01 -6.9627398e-01
-1.3305934e-01 -4.8630472e-02 -5.1535986e-02 9.3281043e-01
-3.4517601e-01 -4.7369406e-01 9.2034197e-01 3.2975823e-01
7.6077461e-02 -5.3866929e-01 -1.4497149e+00 -1.1608360e+00
-7.1241552e-01 -3.7558901e-01 8.1138864e-02 5.1720285e-01
3.9114948e-02 -5.5326376e-02 -5.4336792e-01 3.9970454e-02
7.0386672e-01 5.6699093e-02 6.7581666e-01 -1.0723237e+00
-3.1964079e-01 4.0516950e-02 -1.2995975e+00 -6.0411119e-01
-1.3897163e-01 -5.0008553e-01 1.1047991e-01 -1.4244244e+00
1.5576200e-01 4.9694899e-01 -4.0126824e-01 3.6854941e-01
7.0904098e-02 1.3083616e-01 1.3120605e-01 4.4856754e-01
-3.3142745e-01 5.9807134e-01 1.0249579e+00 1.3621870e-01
-2.1131961e-01 4.9244210e-01 3.8872916e-01 9.0343404e-01
5.9293699e-01 -3.5479116e-01 -4.0180206e-01 1.3326967e-01
-5.5853534e-01 3.9981303e-01 4.3444106e-01 -1.4112653e+00
-1.6053081e-01 -3.7241486e-01 3.4265676e-01 -7.8930211e-01
5.8510572e-01 -1.0490298e+00 5.0978577e-01 8.9034700e-01
-1.2821150e-01 -6.5373749e-02 -1.8251577e-01 8.5420805e-01
-1.6139182e-01 -2.5078821e-01 5.8511692e-01 -1.2183434e-01
-8.6364776e-01 1.5697713e-01 -5.0524682e-01 -4.2835626e-01
1.3694279e+00 -7.3863316e-01 2.4004683e-01 -3.1634456e-01
-1.3199809e+00 -3.3467391e-01 2.7787036e-01 1.4439689e-01
7.9878002e-01 -1.4580046e+00 -2.7022675e-01 1.8221033e-01
3.2767940e-01 -4.5451397e-01 4.6425313e-01 1.3037072e+00
-4.4551942e-01 1.6569857e-01 -8.0514002e-01 -5.6265730e-01
-1.9322015e+00 3.6855799e-01 1.9451573e-01 -9.9927887e-02
-7.3008102e-01 1.9280672e-01 -5.1133072e-01 2.1147897e-02
1.5330057e-01 -2.0476231e-01 -6.3951474e-01 1.1195463e-01
4.1050091e-01 7.6831341e-01 3.8392041e-02 -1.2596122e+00
-5.5553401e-01 7.6135057e-01 3.6575419e-01 -3.0529481e-01
1.0041867e+00 -1.0146021e-01 4.6419268e-03 7.3467964e-01
1.1314820e+00 -9.3012534e-02 -1.2035666e+00 4.5122402e-03
2.0535943e-01 -6.3217741e-01 -4.5120478e-01 -4.1466185e-01
-7.9711306e-01 8.9775175e-01 1.0216942e+00 -7.2771020e-02
1.1280112e+00 -2.5904074e-01 7.5027937e-01 9.9102288e-02
8.3071184e-01 -1.2403765e+00 -4.4630710e-02 1.6629720e-01
8.7603951e-01 -7.8513366e-01 9.1383666e-02 -2.3868011e-01
-6.5636152e-01 1.3501619e+00 4.0211967e-01 -2.1976565e-01
4.5300233e-01 -2.5331974e-01 -9.3688086e-02 -1.5623079e-01
-1.5767505e-02 -5.5836797e-01 4.6601340e-01 6.1957991e-01
1.6227455e-01 3.3269617e-01 -1.1212953e+00 4.2040384e-01
1.2871709e-01 6.8572134e-02 2.7141485e-01 1.2768203e+00
-6.4169908e-01 -9.6551025e-01 -6.7438865e-01 2.7300727e-01
-1.9576441e-01 6.5014619e-01 -3.3906066e-01 1.0365343e+00
2.4653941e-01 9.3858880e-01 -9.0114124e-02 -8.1868303e-01
5.3271341e-01 4.5254880e-01 6.8785673e-01 -5.0779116e-01
-2.6820087e-01 2.7473804e-01 9.5508629e-03 -9.2714238e-01
-6.1908102e-01 -1.1057895e+00 -1.5950595e+00 2.0156564e-01
-1.5206917e-01 2.0669756e-02 5.9871876e-01 1.3092167e+00
-1.5757233e-01 6.6876739e-02 5.2337909e-01 -1.0399436e+00
-8.4100181e-01 -9.8474258e-01 -9.9546236e-01 9.8396844e-01
7.1631193e-02 -1.1236240e+00 -4.8077211e-01 1.8829577e-01] | [7.818530559539795, 0.3194855749607086] |
0cdbe157-82fa-4287-a037-20cb925c0abc | online-heavy-tailed-change-point-detection | 2306.09548 | null | https://arxiv.org/abs/2306.09548v2 | https://arxiv.org/pdf/2306.09548v2.pdf | Online Heavy-tailed Change-point detection | We study algorithms for online change-point detection (OCPD), where samples that are potentially heavy-tailed, are presented one at a time and a change in the underlying mean must be detected as early as possible. We present an algorithm based on clipped Stochastic Gradient Descent (SGD), that works even if we only assume that the second moment of the data generating process is bounded. We derive guarantees on worst-case, finite-sample false-positive rate (FPR) over the family of all distributions with bounded second moment. Thus, our method is the first OCPD algorithm that guarantees finite-sample FPR, even if the data is high dimensional and the underlying distributions are heavy-tailed. The technical contribution of our paper is to show that clipped-SGD can estimate the mean of a random vector and simultaneously provide confidence bounds at all confidence values. We combine this robust estimate with a union bound argument and construct a sequential change-point algorithm with finite-sample FPR guarantees. We show empirically that our algorithm works well in a variety of situations, whether the underlying data are heavy-tailed, light-tailed, high dimensional or discrete. No other algorithm achieves bounded FPR theoretically or empirically, over all settings we study simultaneously. | ['Narayanaswamy', 'Balakrishnan', 'Abishek Sankararaman'] | 2023-06-15 | null | null | null | null | ['change-point-detection'] | ['time-series'] | [-4.16176617e-02 -3.87512535e-01 -1.46241054e-01 1.15344115e-01
-1.11912858e+00 -8.83200288e-01 5.07593095e-01 3.88868004e-01
-4.25056338e-01 9.62520838e-01 -3.99270386e-01 -6.19855046e-01
-8.86566490e-02 -5.83254695e-01 -1.00183833e+00 -7.69318461e-01
-7.62220025e-01 4.80791003e-01 4.84831333e-01 2.29603946e-01
2.15679452e-01 7.60928035e-01 -1.49090838e+00 -4.53760177e-01
5.40418029e-01 1.01561666e+00 -2.12668315e-01 1.41655636e+00
6.47713989e-02 2.80652702e-01 -5.30765593e-01 -3.07638466e-01
5.20844817e-01 -3.47856998e-01 -1.51844814e-01 6.23807162e-02
2.60444880e-01 -3.41142923e-01 2.01244265e-01 1.27291250e+00
5.95591128e-01 -2.32308153e-02 8.72676075e-01 -1.67288661e+00
-1.25894934e-01 1.98738530e-01 -9.49141502e-01 4.10349220e-01
3.46978962e-01 -4.75361012e-03 6.66763544e-01 -1.05756867e+00
4.35051739e-01 1.02757406e+00 8.70986104e-01 3.35982323e-01
-1.38218677e+00 -5.22180915e-01 3.17878187e-01 -2.64984190e-01
-1.28984046e+00 -2.50554711e-01 4.27261055e-01 -5.37597656e-01
5.86713195e-01 4.67657298e-01 4.66244131e-01 8.07756245e-01
2.55140483e-01 9.89174664e-01 8.72502029e-01 -5.47030032e-01
7.86041617e-01 7.15152994e-02 1.26237616e-01 6.10530078e-01
7.93914557e-01 9.09067467e-02 -2.65130192e-01 -7.40786433e-01
5.93050361e-01 1.68697745e-01 -1.61666200e-01 -5.09093642e-01
-9.54011142e-01 7.89151192e-01 -5.91570854e-01 1.57578751e-01
-3.59642535e-01 2.57367402e-01 3.08266521e-01 5.44132411e-01
7.47672498e-01 -2.14797780e-01 -6.74124002e-01 -4.06864077e-01
-1.01991069e+00 4.45821434e-01 1.22851944e+00 1.27664340e+00
3.42104435e-01 -3.11287288e-02 -4.23225820e-01 4.62402970e-01
9.95229185e-02 1.06040490e+00 1.63596779e-01 -7.20733047e-01
6.06612206e-01 -1.05167978e-01 9.01452661e-01 -8.50559592e-01
-1.28938809e-01 -5.54944336e-01 -7.48780012e-01 3.18530023e-01
5.53819478e-01 -4.90859866e-01 -4.82961655e-01 1.93951774e+00
6.24672234e-01 1.71626955e-01 -2.78316528e-01 4.49353427e-01
-3.59265715e-01 6.75962687e-01 -5.97220540e-01 -8.42708230e-01
7.51455605e-01 -3.37848753e-01 -6.32777452e-01 2.11999536e-01
7.88447142e-01 -3.97633553e-01 1.03393245e+00 6.65374577e-01
-1.07385302e+00 3.84610631e-02 -9.40171719e-01 5.26296079e-01
-4.12604883e-02 -2.84950197e-01 8.17803070e-02 8.01867127e-01
-8.97310317e-01 4.12533998e-01 -9.64653671e-01 -1.94795176e-01
2.39256725e-01 -7.49215856e-02 4.86599058e-02 1.15133673e-01
-7.06423461e-01 1.95478424e-01 1.48918657e-02 -1.53869092e-01
-9.72967565e-01 -7.38509536e-01 -3.13633919e-01 -8.80422816e-02
4.57065403e-01 -5.21116316e-01 1.32192993e+00 -7.14162409e-01
-1.04887974e+00 6.52428806e-01 -3.99267226e-01 -7.37335920e-01
1.29888487e+00 -1.35080680e-01 -1.90837950e-01 -3.38966399e-02
1.74475536e-01 -3.51920635e-01 1.05944633e+00 -1.08768225e+00
-1.05315495e+00 -4.57776725e-01 -5.29320002e-01 -1.58625096e-01
-1.78773955e-01 8.37359857e-03 -3.74585778e-01 -4.90856379e-01
-1.98339596e-01 -9.15482104e-01 -2.92203486e-01 2.73662657e-01
-3.74612331e-01 -1.83435768e-01 7.28919923e-01 -2.99951196e-01
1.39316356e+00 -2.22026634e+00 -6.41257346e-01 2.97188997e-01
-8.66445079e-02 2.34632120e-01 1.57795280e-01 7.89656520e-01
2.08985314e-01 2.84166723e-01 -3.47694814e-01 -6.04091585e-01
2.69894868e-01 -5.45245502e-03 -6.42574012e-01 1.13274121e+00
-1.80249229e-01 2.21581936e-01 -1.05385566e+00 -1.48983866e-01
-1.14079855e-01 -2.50606388e-02 -4.46333408e-01 2.27889456e-02
-1.66291237e-01 -6.07472584e-02 -2.61906624e-01 3.05610985e-01
1.10623932e+00 -1.79324672e-01 -2.81936705e-01 5.20769119e-01
-2.42282674e-01 -3.12439591e-01 -1.81844616e+00 8.70517731e-01
-4.45470691e-01 6.57317758e-01 3.66103202e-01 -8.12580168e-01
7.65702367e-01 2.86132880e-02 5.14863014e-01 -1.28147844e-02
-7.52841905e-02 4.49377596e-01 -5.11932552e-01 -2.13252023e-01
1.63216606e-01 -2.35498443e-01 1.02120936e-01 3.95778179e-01
-3.76232117e-01 3.94563042e-02 3.25072229e-01 1.72389403e-01
1.29534960e+00 -3.48615170e-01 2.84665525e-01 -2.99216449e-01
2.79683113e-01 -4.00480807e-01 4.44463819e-01 1.39664197e+00
-3.19272608e-01 3.86967629e-01 8.57513785e-01 1.44639751e-02
-1.02693498e+00 -1.21221912e+00 -2.70975262e-01 9.80929673e-01
1.01203375e-01 -3.66749801e-02 -4.89171863e-01 -5.63562453e-01
4.16306734e-01 8.62987161e-01 -6.17503762e-01 1.26705840e-01
-3.68483774e-02 -6.58884943e-01 1.88484520e-01 2.34279126e-01
1.64164811e-01 -1.95808381e-01 -5.48348844e-01 4.07600880e-01
4.55364138e-01 -8.87783587e-01 -7.34534681e-01 2.09192783e-01
-9.36420083e-01 -9.99179602e-01 -7.67587006e-01 -4.72777039e-01
6.55507445e-01 2.68651783e-01 8.95111382e-01 -4.07596350e-01
-1.74925789e-01 7.67111361e-01 -1.46927714e-01 -6.95675731e-01
-3.12079906e-01 -5.60048282e-01 2.28257537e-01 2.61513174e-01
1.15580052e-01 -3.77198815e-01 -4.12060469e-01 2.30450004e-01
-9.72821653e-01 -5.49930751e-01 6.60206303e-02 6.23890817e-01
8.57088566e-01 4.26360965e-01 5.86304903e-01 -7.02887595e-01
8.14623773e-01 -7.47769773e-01 -1.23656547e+00 1.49092853e-01
-7.57995367e-01 -5.65616488e-02 7.65829444e-01 -6.33700132e-01
-6.01880014e-01 -1.90538660e-01 7.34873265e-02 -5.73352933e-01
3.47434636e-03 3.46288353e-01 3.47873792e-02 2.27389455e-01
5.82921743e-01 5.48976839e-01 -1.17151119e-01 -5.93471050e-01
2.49033138e-01 6.43600464e-01 5.64511001e-01 -6.00451827e-01
7.63597727e-01 8.17877829e-01 4.23540205e-01 -1.04467773e+00
-5.72121680e-01 -7.94331133e-01 -3.45226586e-01 -9.13540125e-02
-5.43545038e-02 -8.34082484e-01 -6.65564477e-01 7.22901881e-01
-1.01133692e+00 -4.22360957e-01 -4.67863321e-01 4.59945798e-01
-8.08793187e-01 4.95633572e-01 -2.30219021e-01 -1.69808769e+00
-3.90708834e-01 -3.54521036e-01 9.26217020e-01 -2.84927249e-01
8.72578546e-02 -1.05988383e+00 3.12797725e-01 -5.96664250e-01
2.28967920e-01 5.59744656e-01 4.36169565e-01 -1.02806103e+00
-1.77679226e-01 -8.70412946e-01 -4.47184481e-02 3.64863813e-01
1.89136624e-01 3.88872713e-01 -3.64161313e-01 -7.71355808e-01
2.29159519e-01 1.76398456e-01 6.00361943e-01 7.95704067e-01
1.23260844e+00 -9.18870330e-01 -5.17462790e-01 3.74076903e-01
1.55086923e+00 2.70052929e-04 1.26741603e-01 1.71277478e-01
7.77043328e-02 2.87625104e-01 8.92620564e-01 1.23612249e+00
3.00511643e-02 3.24428409e-01 4.03052866e-01 1.96090847e-01
3.81363243e-01 -2.60933876e-01 6.11357689e-01 3.49091440e-01
6.14943981e-01 -5.80076098e-01 -7.77881324e-01 9.10811543e-01
-1.88214254e+00 -1.07303262e+00 -5.84221184e-01 3.03390098e+00
1.02697814e+00 4.47848618e-01 6.75781667e-01 2.30929002e-01
1.08996427e+00 -3.42521787e-01 -1.00223625e+00 -3.98154706e-01
-1.31130204e-01 -1.20488852e-01 9.83947456e-01 6.51496232e-01
-1.02851975e+00 1.54986858e-01 6.57524490e+00 9.21455503e-01
-8.96671832e-01 1.94573537e-01 3.73828202e-01 -3.64297569e-01
-1.28004506e-01 -2.43091434e-01 -1.12743366e+00 8.88259888e-01
1.13542783e+00 -6.63284361e-01 6.54772446e-02 1.09712994e+00
3.90995890e-01 -2.12571234e-01 -1.15798140e+00 1.07416534e+00
-2.53616691e-01 -9.88242924e-01 -4.62985784e-01 1.27875611e-01
9.70918536e-01 3.91082279e-02 1.72980040e-01 7.31569082e-02
6.94777012e-01 -2.13314131e-01 6.54542446e-01 5.65067828e-01
7.55814433e-01 -1.02332306e+00 5.80694020e-01 6.73165143e-01
-9.89662170e-01 -1.32212639e-01 -1.83596849e-01 3.84151824e-02
1.41778588e-01 1.49873507e+00 -7.63901114e-01 1.28392830e-01
4.28568095e-01 2.64153063e-01 -7.97933564e-02 1.62155461e+00
1.18876688e-01 1.01074433e+00 -1.08497131e+00 -3.87187004e-01
2.15418125e-03 2.20661648e-02 1.02109408e+00 1.54361761e+00
6.17282391e-01 -2.70748407e-01 3.13400269e-01 4.08366263e-01
-1.69800539e-02 2.10723102e-01 -6.27916992e-01 5.89672513e-02
8.18050861e-01 7.34433949e-01 -7.34087408e-01 -5.01792908e-01
-3.65498364e-01 9.38411474e-01 -1.52890518e-01 4.50006694e-01
-7.07679033e-01 -7.94185400e-01 5.51133335e-01 1.93580359e-01
8.56021166e-01 -4.76154000e-01 -2.41215453e-01 -1.14685881e+00
4.48508471e-01 -5.97121835e-01 6.50139928e-01 -1.76920891e-01
-1.63312173e+00 -2.50881672e-01 -1.00357704e-01 -1.24197888e+00
-5.37583709e-01 -3.79849166e-01 -5.50134242e-01 8.02377403e-01
-1.07835758e+00 -4.22657043e-01 2.77904719e-01 6.17240131e-01
3.57301652e-01 2.59792179e-01 2.61398017e-01 4.83580353e-03
-4.44482565e-01 7.28983521e-01 1.02629018e+00 -6.23416901e-02
3.87871444e-01 -1.49732482e+00 2.20762804e-01 1.23738289e+00
-8.31463039e-02 4.79754955e-01 1.26000166e+00 -8.14905822e-01
-1.62801230e+00 -1.15085030e+00 5.14385462e-01 -2.39303157e-01
1.05867589e+00 -5.18393219e-01 -8.20146501e-01 4.67888564e-01
-5.05846918e-01 3.76572967e-01 5.01247585e-01 1.23597451e-01
-8.73782858e-02 -2.30525672e-01 -1.45180583e+00 5.24386644e-01
7.74200320e-01 -3.00586015e-01 -4.40581322e-01 6.49176002e-01
4.61811692e-01 -1.70105726e-01 -6.61547840e-01 2.31197834e-01
3.17919999e-01 -7.02876568e-01 5.64754605e-01 -7.92390764e-01
-3.05413544e-01 -4.56558466e-01 -2.67968595e-01 -1.04626894e+00
-9.29588452e-02 -1.17786145e+00 -7.19573498e-01 1.26369071e+00
3.20230484e-01 -7.63270676e-01 5.03672123e-01 4.43831444e-01
3.27139437e-01 -6.60469830e-01 -1.19705772e+00 -1.65333259e+00
3.34417485e-02 -8.45117629e-01 2.63654083e-01 5.01412749e-01
-2.10779026e-01 -1.11531503e-01 -4.07847852e-01 4.38682556e-01
9.25497830e-01 1.33691907e-01 9.86258209e-01 -1.08869016e+00
-4.91042852e-01 -3.72282267e-01 -3.81656289e-01 -1.13136601e+00
-1.39434069e-01 -3.54293674e-01 3.12212259e-01 -1.11955953e+00
1.13111295e-01 -3.46761853e-01 -3.06797534e-01 1.03159584e-01
-1.12721018e-01 -3.03718626e-01 -3.02094609e-01 2.25260764e-01
-7.33015001e-01 4.14356440e-01 6.51758611e-01 2.87635386e-01
-4.54233229e-01 5.95309436e-01 -2.71496564e-01 6.17565751e-01
5.65885365e-01 -5.68022370e-01 -3.13957244e-01 1.75843403e-01
3.74736756e-01 1.22748204e-01 2.45319992e-01 -9.30238008e-01
2.75005370e-01 -3.21668684e-01 1.79849416e-01 -1.13552284e+00
-2.29751036e-01 -6.89265549e-01 -1.12341493e-02 8.12891781e-01
-5.18202126e-01 -1.40437784e-04 1.34621069e-01 1.29481614e+00
3.37354779e-01 -3.83086205e-01 8.95184278e-01 2.78703928e-01
-1.15761027e-01 6.05965912e-01 -5.64077079e-01 4.42075044e-01
1.21685195e+00 1.01191297e-01 -5.81132099e-02 -7.10638940e-01
-4.64003503e-01 2.92837322e-01 3.88014287e-01 -3.17454860e-02
3.81307900e-01 -1.25773084e+00 -8.64375532e-01 4.73825783e-02
-3.52255292e-02 -2.35773876e-01 1.22003362e-03 9.21086907e-01
-7.15886354e-02 1.41413227e-01 5.45806825e-01 -7.81698763e-01
-1.18410075e+00 1.12945807e+00 8.56768414e-02 -6.35444596e-02
-4.73477900e-01 8.22506309e-01 -1.19045600e-01 3.28019150e-02
4.36963856e-01 -3.56890708e-01 6.48843229e-01 1.38492703e-01
8.92892063e-01 8.28496695e-01 1.80405751e-01 2.99763083e-02
-2.75064677e-01 1.83965430e-01 -6.85469136e-02 -5.75232923e-01
1.17186868e+00 -5.08801103e-01 6.91498145e-02 9.85643566e-01
1.45129871e+00 4.29302782e-01 -1.46183324e+00 -3.37497979e-01
-1.02848047e-03 -7.17787504e-01 -2.22236097e-01 -5.43915927e-01
-5.29095352e-01 7.16422915e-01 8.25741172e-01 8.10541868e-01
1.00961185e+00 -1.09946229e-01 5.19228458e-01 4.21110630e-01
5.98448336e-01 -1.21271455e+00 -1.78862229e-01 4.64928567e-01
7.82911479e-01 -8.94777298e-01 -1.32325292e-01 5.20514771e-02
-1.88594207e-01 1.03916025e+00 5.71742393e-02 -2.84022778e-01
8.37136567e-01 4.76010174e-01 -5.51941872e-01 3.39275569e-01
-9.47498083e-01 -2.46374354e-01 -4.00579460e-02 5.14200509e-01
-7.59626776e-02 1.24359518e-01 -6.09571576e-01 3.79072607e-01
3.10677523e-03 1.92969128e-01 6.39030397e-01 1.12226105e+00
-8.36109042e-01 -5.45851469e-01 -5.05781174e-01 6.50220215e-01
-7.97686517e-01 1.32312641e-01 -7.53724128e-02 6.11371458e-01
-4.39210057e-01 1.03229654e+00 5.25207147e-02 7.13899285e-02
2.44713485e-01 -3.20948139e-02 2.93408483e-01 -1.75827280e-01
2.17163414e-01 1.00654714e-01 -2.07168646e-02 -3.88001084e-01
1.57581881e-01 -9.86731291e-01 -1.03333676e+00 -6.14130974e-01
-4.09470320e-01 2.99507678e-01 7.70549178e-01 9.07791495e-01
4.80394185e-01 -2.01420292e-01 1.23498118e+00 -3.07501256e-01
-1.10187781e+00 -6.42066419e-01 -9.98096347e-01 1.35095790e-01
9.52725768e-01 -4.39976960e-01 -1.05778491e+00 -1.03054404e-01] | [7.110558032989502, 3.997058629989624] |
73a22e98-4940-44a9-ae79-88410ce9b087 | on-planetary-systems-as-ordered-sequences | 2105.09966 | null | https://arxiv.org/abs/2105.09966v1 | https://arxiv.org/pdf/2105.09966v1.pdf | On planetary systems as ordered sequences | A planetary system consists of a host star and one or more planets, arranged into a particular configuration. Here, we consider what information belongs to the configuration, or ordering, of 4286 Kepler planets in their 3277 planetary systems. First, we train a neural network model to predict the radius and period of a planet based on the properties of its host star and the radii and period of its neighbors. The mean absolute error of the predictions of the trained model is a factor of 2.1 better than the MAE of the predictions of a naive model which draws randomly from dynamically allowable periods and radii. Second, we adapt a model used for unsupervised part-of-speech tagging in computational linguistics to investigate whether planets or planetary systems fall into natural categories with physically interpretable "grammatical rules." The model identifies two robust groups of planetary systems: (1) compact multi-planet systems and (2) systems around giant stars ($\log{g} \lesssim 4.0$), although the latter group is strongly sculpted by the selection bias of the transit method. These results reinforce the idea that planetary systems are not random sequences -- instead, as a population, they contain predictable patterns that can provide insight into the formation and evolution of planetary systems. | ['Michael Collins', 'David Kipping', 'Emily Sandford'] | 2021-05-20 | null | null | null | null | ['unsupervised-part-of-speech-tagging'] | ['natural-language-processing'] | [-3.62496555e-01 4.83452201e-01 -6.41364083e-02 -1.75809547e-01
2.67264336e-01 -8.25761974e-01 1.10603106e+00 -2.53788441e-01
-8.62168893e-02 6.59282029e-01 1.17170662e-02 -8.26378405e-01
-1.62125349e-01 -9.38727796e-01 -9.61304784e-01 -9.04608548e-01
-3.71275038e-01 1.00479639e+00 4.01155561e-01 -4.12362039e-01
1.43102571e-01 5.12021065e-01 -1.69112301e+00 -2.95912743e-01
8.25703323e-01 6.11862302e-01 4.41833436e-01 5.37776291e-01
2.33059585e-01 5.08031189e-01 -4.78179514e-01 1.24709964e-01
4.34197664e-01 -4.42896068e-01 -7.67585099e-01 -1.56377345e-01
2.61540085e-01 2.40674034e-01 -7.36272573e-01 8.43043268e-01
-6.87148515e-03 3.04514796e-01 1.13663924e+00 -6.64511681e-01
-7.61630654e-01 9.34413493e-01 -2.21275464e-01 1.22542948e-01
-1.70123726e-01 1.75946280e-01 1.28632820e+00 -3.94527078e-01
4.49781597e-01 1.12659431e+00 8.11666787e-01 9.88855436e-02
-9.60694373e-01 -2.17975795e-01 -1.12378068e-01 6.12332150e-02
-1.35658598e+00 -5.68170905e-01 4.39845890e-01 -7.78807521e-01
1.35858333e+00 3.87649447e-01 8.52796137e-01 4.05349970e-01
4.21937346e-01 -9.17366892e-02 6.75427318e-01 -7.11173594e-01
1.69783682e-01 3.18941712e-01 6.64026812e-02 7.02080667e-01
7.57523000e-01 4.71133530e-01 -1.42840534e-01 -3.94474506e-01
5.07653117e-01 -6.87942922e-01 -1.51766047e-01 3.08322143e-02
-1.36812830e+00 5.34711719e-01 1.94878012e-01 5.75009584e-01
-2.92984903e-01 2.47043073e-01 -8.63007978e-02 1.49689168e-01
1.11774616e-01 9.56584394e-01 -7.27583349e-01 1.31115168e-01
-5.95743954e-01 2.19168320e-01 1.20919681e+00 7.09104717e-01
9.63590980e-01 1.97853357e-01 5.05978346e-01 7.01388001e-01
7.26920784e-01 7.76707768e-01 7.49651253e-01 -9.02675211e-01
-9.04846713e-02 2.64254987e-01 3.07435542e-01 -9.19793069e-01
-5.03405571e-01 -3.94580215e-02 -2.43220150e-01 -6.07687570e-02
5.22557259e-01 -1.25499576e-01 -6.37310505e-01 1.76932943e+00
4.01222736e-01 -1.79312155e-01 1.47719741e-01 5.58549881e-01
6.67620122e-01 7.73826838e-01 -9.29063093e-03 -9.08532813e-02
1.41231811e+00 -7.41086602e-01 1.27240181e-01 -4.48604077e-01
3.76216888e-01 -5.74810505e-01 7.93871641e-01 1.71436742e-01
-6.11459494e-01 -4.96524364e-01 -1.15687549e+00 3.47109079e-01
-3.68079007e-01 -6.39900044e-02 8.92761707e-01 6.76214516e-01
-8.90219986e-01 8.12996566e-01 -8.50837052e-01 -5.21024108e-01
-6.91711307e-01 2.43988335e-01 1.09600022e-01 9.74784136e-01
-1.38885307e+00 1.07051992e+00 6.35500729e-01 -4.76847529e-01
-6.99203968e-01 -3.12561214e-01 -5.99302530e-01 -7.73012489e-02
-1.41514450e-01 -5.06314278e-01 1.34359550e+00 -6.54763997e-01
-1.32857335e+00 9.21374083e-01 -5.90482913e-02 -5.74309230e-01
-7.93241486e-02 2.18684673e-01 -6.01580679e-01 1.25984460e-01
8.34527053e-03 5.59407294e-01 8.49670172e-01 -1.33471501e+00
-5.12796521e-01 1.10319681e-01 -1.88267723e-01 1.36357591e-01
-1.27066299e-01 1.17225638e-02 -1.27815530e-01 -2.77019054e-01
2.32323542e-01 -1.39771521e+00 -2.95183752e-02 -4.45064992e-01
-5.71049213e-01 -7.62861133e-01 3.87317151e-01 -6.26861930e-01
1.02436674e+00 -1.99033606e+00 -2.08193272e-01 3.23305696e-01
-2.54625408e-03 -1.15429401e-01 2.89444417e-01 3.67864698e-01
-2.81194478e-01 1.79791972e-01 -1.88533574e-01 1.31304607e-01
3.17344517e-01 4.23244864e-01 -3.73698473e-01 6.24082327e-01
-3.66738409e-01 3.61234993e-01 -7.11113274e-01 -2.91131169e-01
2.02735692e-01 8.58094767e-02 5.39723672e-02 -1.04562625e-01
-6.12731397e-01 3.26000661e-01 -2.72942483e-01 4.23874348e-01
3.68005544e-01 -5.06509602e-01 1.42704129e-01 1.90057233e-01
-6.50195420e-01 1.03151917e+00 -6.29585981e-01 9.05641317e-01
-6.99039102e-02 8.48777592e-01 -2.98210979e-01 -7.78605819e-01
1.14926028e+00 2.04432175e-01 5.84340394e-01 -2.82696098e-01
1.88339055e-01 3.08477879e-01 6.91040993e-01 -5.24374068e-01
9.91684258e-01 -2.55142182e-01 -1.00230023e-01 6.79313421e-01
-2.30608195e-01 -3.81368011e-01 3.32054794e-01 6.65698797e-02
7.83124387e-01 9.65693146e-02 1.57907650e-01 -6.88450813e-01
-6.10540807e-02 1.08576648e-01 7.40246952e-01 7.98958600e-01
2.10653111e-01 5.51212311e-01 4.06332970e-01 -5.44568479e-01
-1.48236585e+00 -7.87849963e-01 -6.60291314e-01 9.20571923e-01
2.33673096e-01 -3.53774220e-01 -3.67683768e-01 5.79889445e-03
2.84250885e-01 8.43834460e-01 -6.23621166e-01 -3.36450785e-01
-5.51875293e-01 -1.20045578e+00 6.12020671e-01 7.29821948e-03
-8.71333107e-02 -1.44558048e+00 -7.69732237e-01 -4.50606011e-02
-1.57105073e-01 -7.70445943e-01 -9.09963548e-02 3.16445112e-01
-4.27756935e-01 -9.51262355e-01 -1.22416578e-01 -6.97026074e-01
5.87392509e-01 5.88137768e-02 1.34281254e+00 2.74755269e-01
1.29965201e-01 7.28255464e-03 -4.72911626e-01 -5.14378667e-01
-8.00677061e-01 2.01397985e-02 9.95162964e-01 -4.17113900e-01
4.02916312e-01 -4.97634530e-01 -4.41831648e-01 5.94558120e-01
-2.58744627e-01 -1.77831292e-01 4.39489126e-01 4.11143035e-01
3.31684887e-01 4.47092205e-01 1.53771952e-01 -2.68100858e-01
1.37649953e-01 -9.51393485e-01 -5.52505016e-01 4.23381597e-01
-8.19491148e-01 1.98449105e-01 6.22794271e-01 -6.34535372e-01
-8.72536719e-01 -5.65979034e-02 7.32118636e-02 -9.76089388e-02
-2.21354753e-01 3.87672216e-01 -3.01818340e-03 -3.03109791e-02
9.32945192e-01 2.53584236e-01 -9.48772132e-02 -4.62390214e-01
3.71830314e-01 8.25829744e-01 8.23715925e-01 -1.07211614e+00
9.01310861e-01 3.13094795e-01 -2.94264108e-01 -1.37098050e+00
-3.48295271e-01 -2.22364232e-01 -3.66571724e-01 -1.92623317e-01
5.66763878e-01 -8.73099446e-01 -5.75552166e-01 2.77525783e-01
-9.48492825e-01 -7.28045762e-01 -4.77551132e-01 7.07231224e-01
-5.83095908e-01 6.90275848e-01 -4.22935516e-01 -7.79017091e-01
-3.75286847e-01 -6.65492475e-01 5.29023468e-01 5.78694701e-01
-6.19690061e-01 -8.03798437e-01 4.52915281e-01 -3.01188618e-01
6.51142560e-03 -1.07584754e-02 9.70000565e-01 -7.26860344e-01
-2.51235783e-01 1.77168563e-01 9.45353955e-02 -1.52864143e-01
2.24012315e-01 6.13667130e-01 -5.54532290e-01 -1.10189833e-01
1.48239762e-01 1.91620439e-01 7.99589515e-01 4.72482562e-01
6.32803619e-01 -4.95940804e-01 -6.07648194e-01 6.08018160e-01
1.08091366e+00 6.55183315e-01 4.28045094e-01 7.41391778e-01
4.33257401e-01 8.67628753e-01 3.47413778e-01 4.29875821e-01
3.39048415e-01 3.34118813e-01 2.48572767e-01 6.70954525e-01
-2.67512128e-02 -3.53586555e-01 4.88583237e-01 9.58438039e-01
-8.46835598e-02 -2.34032705e-01 -1.22891140e+00 9.66596901e-01
-1.62504160e+00 -8.63936722e-01 -4.81937289e-01 2.34677482e+00
6.87839448e-01 2.67703265e-01 3.32036406e-01 -3.82857084e-01
6.76019490e-01 7.50708133e-02 -8.02175462e-01 -4.29705828e-01
-2.66688675e-01 -5.18078059e-02 7.41938889e-01 3.67712200e-01
-8.56047332e-01 9.47148323e-01 7.83979368e+00 3.64530653e-01
-1.12484145e+00 -3.47733468e-01 3.60252738e-01 3.20470929e-01
-6.23294234e-01 3.62567842e-01 -8.76037419e-01 6.94465458e-01
1.39984870e+00 -2.68295944e-01 3.78127664e-01 7.25086510e-01
1.84903055e-01 -1.15885407e-01 -8.63876939e-01 3.41879308e-01
-1.90752193e-01 -9.97086763e-01 -2.70821989e-01 2.94381022e-01
5.13021469e-01 7.00441241e-01 1.86070204e-02 4.14161235e-02
6.42210484e-01 -8.97528887e-01 1.09884262e+00 6.59352481e-01
7.43000269e-01 -3.02904516e-01 9.77608860e-02 4.08998817e-01
-1.16142738e+00 7.57965595e-02 -9.78309929e-01 -2.53581673e-01
1.20120898e-01 5.88724136e-01 -1.11599982e+00 3.33820432e-01
1.15979600e+00 4.06529605e-01 -3.70018601e-01 9.32568073e-01
-3.05905312e-01 9.15509701e-01 -9.50255215e-01 -3.11021596e-01
-8.77476297e-03 -6.58804834e-01 8.45462978e-01 9.91460264e-01
7.08271682e-01 2.74504662e-01 -3.95062417e-01 1.00393748e+00
2.91389585e-01 -1.01031102e-01 -6.54949188e-01 -4.29848701e-01
1.02167976e+00 1.28995955e+00 -8.94464254e-01 -4.11209524e-01
-3.62543553e-01 -4.32797819e-02 -2.52945237e-02 -1.17842697e-01
-6.37660742e-01 -4.63492244e-01 6.96243525e-01 6.59133866e-02
7.07130790e-01 -5.06429017e-01 -5.27569354e-01 -1.02724659e+00
-3.05143744e-01 -6.82861447e-01 -2.47365218e-02 -9.30625081e-01
-1.19866669e+00 7.54478395e-01 -7.58586526e-02 -1.27942276e+00
-6.52868748e-01 -7.49741316e-01 -9.89318371e-01 6.08711779e-01
-5.83092630e-01 -6.57719374e-01 4.34227437e-02 -3.26246955e-02
-3.48864272e-02 -5.59745550e-01 6.61877275e-01 -3.38660121e-01
-4.22243685e-01 3.22303534e-01 7.20158637e-01 -6.46966100e-02
5.88751733e-01 -1.43150413e+00 9.57384408e-01 6.67873204e-01
4.17450517e-01 5.73573649e-01 1.39451385e+00 -8.09002876e-01
-1.07000422e+00 -7.73969531e-01 1.02161264e+00 -6.61980748e-01
1.06538379e+00 2.30398998e-02 -9.73120034e-01 7.55174279e-01
3.50944221e-01 -5.33008575e-01 5.12192965e-01 3.62089276e-01
-3.00115705e-01 2.40464866e-01 -9.33157563e-01 3.59441370e-01
7.54357278e-01 -3.88848931e-01 -1.15381312e+00 6.74626291e-01
1.02598369e+00 -9.20304507e-02 -1.11116838e+00 2.69725412e-01
6.55438006e-01 -8.78165007e-01 9.69223022e-01 -3.79735708e-01
1.63653731e-01 -5.18943846e-01 -5.48341051e-02 -1.23822236e+00
-6.86046124e-01 -6.54805958e-01 -1.74573705e-01 8.33139896e-01
7.16152549e-01 -8.29306304e-01 6.70644104e-01 8.14460278e-01
-5.56304812e-01 -3.53779048e-01 -7.63385296e-01 -9.09668207e-01
6.42072678e-01 -3.15608978e-02 9.88080680e-01 1.07481039e+00
2.45502219e-01 2.54808694e-01 -1.82649821e-01 2.82411277e-01
6.17495954e-01 4.92378652e-01 4.44378704e-01 -1.40728092e+00
-8.20526481e-01 -4.92802769e-01 1.24856137e-01 -1.09302485e+00
2.21505523e-01 -6.37524366e-01 4.76317108e-01 -1.11946309e+00
-3.22666258e-01 -1.19114602e+00 1.05908304e-03 4.98172849e-01
3.27106059e-01 -1.61728323e-01 -6.06780350e-02 6.63551271e-01
-1.42458096e-01 4.03570026e-01 5.41830242e-01 7.40471184e-02
-2.07284436e-01 -3.15503702e-02 -7.75421381e-01 9.80189502e-01
9.86114740e-01 -5.08648396e-01 3.39201614e-02 -2.68455058e-01
7.09589541e-01 -5.72842695e-02 -1.60470344e-02 -9.61325884e-01
3.07747032e-02 -2.41017714e-01 3.62293646e-02 -6.02706015e-01
1.52809486e-01 -3.61697376e-01 5.65650165e-01 5.34669280e-01
2.68346220e-01 -1.26478985e-01 -3.35084181e-03 3.64218265e-01
2.31127933e-01 -8.08089137e-01 6.99339628e-01 -2.70852387e-01
-6.69741750e-01 1.07312083e-01 -8.68553460e-01 -3.13735664e-01
7.48563051e-01 -3.64383571e-02 -5.05990028e-01 -2.25261480e-01
-5.67709744e-01 -1.01655209e-03 8.60137641e-01 4.52255368e-01
-1.54760048e-01 -9.59354579e-01 -5.42337418e-01 -2.39943247e-03
1.91618148e-02 7.45081231e-02 -3.40818018e-01 5.26567757e-01
-8.13459218e-01 6.05387390e-01 6.15870804e-02 -4.39533442e-01
-7.55904317e-01 3.88530463e-01 6.99636519e-01 2.74305314e-01
-5.72648168e-01 7.95815051e-01 1.18511632e-01 -7.66251922e-01
-3.97253871e-01 -1.77076980e-01 -1.68691441e-01 -1.34708464e-01
1.24110229e-01 1.49337262e-01 -2.49525741e-01 -1.14153957e+00
-1.55658364e-01 3.86145025e-01 1.33708060e-01 2.01898208e-03
1.30430627e+00 -1.65368110e-01 -6.20285809e-01 6.78525686e-01
5.60185015e-01 7.57257864e-02 -9.72259223e-01 -2.76044518e-01
4.65438589e-02 -1.33960843e-02 -7.33050048e-01 -4.93438125e-01
-4.92797971e-01 1.91111788e-01 8.38738978e-02 8.47269416e-01
3.86758149e-01 6.69071376e-01 6.63516998e-01 7.57608652e-01
5.19587815e-01 -8.99584532e-01 -5.34713626e-01 8.39958966e-01
6.93473399e-01 -8.91290069e-01 1.51528016e-01 8.52087662e-02
-2.00227693e-01 9.64033544e-01 5.64033568e-01 -3.26495349e-01
7.78175890e-01 -8.96958262e-02 -2.39409581e-01 -1.53738096e-01
-8.74088347e-01 -6.11594506e-02 3.46045196e-01 4.34772879e-01
3.67144674e-01 5.75735033e-01 -5.18505335e-01 4.44904745e-01
-1.22348022e+00 -8.73623610e-01 5.79181850e-01 4.15759742e-01
-1.00218570e+00 -9.06424046e-01 -7.51677513e-01 5.82684219e-01
-9.21930149e-02 -1.65677201e-02 -6.31098449e-01 7.90352881e-01
3.53825182e-01 9.63322341e-01 5.43981016e-01 -6.91938996e-01
-1.80387601e-01 5.25786504e-02 3.16577047e-01 -4.08149540e-01
-1.42637745e-01 -1.56881645e-01 5.41346788e-01 2.78244913e-01
-3.01279545e-01 -1.27661729e+00 -1.43617439e+00 -4.49227273e-01
-4.33466226e-01 7.21610069e-01 6.41216755e-01 1.09984815e+00
1.24383911e-01 5.67996651e-02 5.77128887e-01 -9.36366856e-01
-5.24452388e-01 -1.30064476e+00 -8.58404934e-01 2.01554984e-01
5.82805984e-02 -6.75192058e-01 -7.80957460e-01 8.44287500e-02] | [6.84966516494751, 3.299739360809326] |
c53c2791-ded0-4660-8028-bdab21c4c1b3 | revisit-out-of-vocabulary-problem-for-slot | 2302.13584 | null | https://arxiv.org/abs/2302.13584v1 | https://arxiv.org/pdf/2302.13584v1.pdf | Revisit Out-Of-Vocabulary Problem for Slot Filling: A Unified Contrastive Frameword with Multi-level Data Augmentations | In real dialogue scenarios, the existing slot filling model, which tends to memorize entity patterns, has a significantly reduced generalization facing Out-of-Vocabulary (OOV) problems. To address this issue, we propose an OOV robust slot filling model based on multi-level data augmentations to solve the OOV problem from both word and slot perspectives. We present a unified contrastive learning framework, which pull representations of the origin sample and augmentation samples together, to make the model resistant to OOV problems. We evaluate the performance of the model from some specific slots and carefully design test data with OOV word perturbation to further demonstrate the effectiveness of OOV words. Experiments on two datasets show that our approach outperforms the previous sota methods in terms of both OOV slots and words. | ['Weiran Xu', 'Xinyue Cui', 'Keqing He', 'Zechen Wang', 'Xuefeng Li', 'LiWen Wang', 'Tingfeng Hui', 'Chen Zeng', 'Yuxiang Wu', 'Dayuan Fu', 'Guanting Dong', 'Daichi Guo'] | 2023-02-27 | null | null | null | null | ['slot-filling'] | ['natural-language-processing'] | [ 4.41952422e-02 8.82298708e-01 -5.71119130e-01 -1.95688039e-01
-4.92403418e-01 -4.45830822e-01 7.05210626e-01 3.00174266e-01
-4.56011415e-01 9.01500583e-01 4.59187061e-01 -5.65502882e-01
1.15956850e-01 -8.97689819e-01 -3.77951026e-01 -9.26292688e-02
2.37884477e-01 8.25792134e-01 5.16455233e-01 -8.64547610e-01
1.65875033e-01 -2.81076699e-01 -1.33868480e+00 1.30737543e-01
1.33308315e+00 6.93879008e-01 2.33593494e-01 1.86677709e-01
-1.12097883e+00 4.36917096e-01 -8.89555335e-01 -4.14965689e-01
3.69505882e-01 -4.31237817e-02 -1.03230309e+00 2.54077822e-01
2.66637921e-01 -1.92401502e-02 -5.56872070e-01 8.14396620e-01
5.98668337e-01 4.58465487e-01 4.73399818e-01 -1.12222457e+00
-7.55581021e-01 1.05811262e+00 -2.33171284e-01 1.49118766e-01
1.68272391e-01 -2.27916360e-01 1.33178174e+00 -1.07408261e+00
6.86752021e-01 1.47889149e+00 6.50521696e-01 9.06611621e-01
-1.49648046e+00 -6.05423570e-01 5.09689808e-01 -1.40648127e-01
-1.13277054e+00 -3.14633638e-01 6.84208512e-01 -6.12872913e-02
1.14316392e+00 2.99753726e-01 4.84223455e-01 1.14889908e+00
-2.24962771e-01 1.34416819e+00 8.96586478e-01 -6.65000796e-01
3.48604441e-01 4.60259408e-01 7.76553869e-01 4.52005386e-01
3.24054152e-01 -1.52130410e-01 -6.40774548e-01 -4.38203245e-01
4.70456690e-01 -4.03818905e-01 -8.94721672e-02 -7.22278714e-01
-5.23323298e-01 1.30486715e+00 -1.98868498e-01 1.47594716e-02
5.05607575e-02 -3.47908080e-01 5.93677759e-01 4.25970018e-01
7.50669599e-01 8.06109846e-01 -9.88641143e-01 -1.17506064e-01
-4.36966270e-01 6.63673937e-01 9.45678592e-01 1.10675371e+00
8.13452065e-01 1.02394514e-01 -7.89978862e-01 1.53541040e+00
4.54979479e-01 3.74727219e-01 8.62775683e-01 -4.40203100e-01
6.40568018e-01 7.53880978e-01 8.63733143e-02 -6.78662300e-01
-6.45429730e-01 -3.39100569e-01 -5.42148411e-01 -4.58723158e-01
2.47434035e-01 9.54260398e-03 -1.19895935e+00 2.17003441e+00
6.24425173e-01 5.59213273e-02 3.40258867e-01 5.54317474e-01
1.31588936e+00 7.11702943e-01 4.01634157e-01 -2.43323207e-01
1.64766765e+00 -1.05877841e+00 -1.40041101e+00 -8.37506354e-01
1.04949856e+00 -3.91887307e-01 1.58193135e+00 -7.63066709e-02
-7.81334043e-01 -4.06561524e-01 -1.08353579e+00 -1.85077459e-01
-5.96915603e-01 -1.97419226e-01 8.48463714e-01 8.78787816e-01
-6.82372272e-01 1.38966203e-01 -3.07582676e-01 -2.55447209e-01
1.71170697e-01 2.45862857e-01 -1.42159805e-01 3.79915163e-02
-2.01287293e+00 7.14536011e-01 7.50006020e-01 -4.13371921e-01
-4.72964585e-01 -7.60835946e-01 -1.50109982e+00 6.55729771e-02
9.05690312e-01 -2.96330810e-01 1.24990726e+00 -7.60331452e-02
-1.35890591e+00 6.72820568e-01 -4.22446817e-01 -6.44264102e-01
3.58456552e-01 -2.39790425e-01 -1.68900564e-01 -6.25688970e-01
3.63639235e-01 6.42169774e-01 7.28846431e-01 -1.16512811e+00
-5.46366274e-01 -1.98823646e-01 4.60250415e-02 3.70202512e-01
-5.61088681e-01 -6.98490679e-01 -6.69766188e-01 -9.21324551e-01
3.82383615e-01 -8.55374336e-01 -3.61487180e-01 -5.23144484e-01
-6.22075379e-01 -9.07357275e-01 5.74852228e-01 -4.53075469e-01
1.76472175e+00 -1.99497342e+00 -4.55916785e-02 1.83282092e-01
1.63202152e-01 5.20516038e-01 -3.78238499e-01 3.92510206e-01
2.34673485e-01 1.61966622e-01 -7.26118237e-02 -5.75513899e-01
2.42252946e-01 5.94160914e-01 -5.74217141e-01 -1.80877000e-01
9.46523845e-02 1.03455484e+00 -7.45873630e-01 -4.27210063e-01
-8.95568449e-03 -1.75801545e-01 -8.64728928e-01 3.06348294e-01
-4.60446149e-01 -1.33428499e-01 -3.72026086e-01 4.91121978e-01
6.55566573e-01 -1.64752811e-01 5.84447920e-01 2.38693625e-01
4.46348757e-01 7.85248756e-01 -1.36378312e+00 1.64127278e+00
-5.17591178e-01 2.79799968e-01 -1.57560751e-01 -1.06875324e+00
1.21801877e+00 3.36361587e-01 5.63663505e-02 -9.73295569e-01
1.16552338e-01 8.49735737e-02 -2.17728674e-01 -1.46044880e-01
1.08200455e+00 5.75094745e-02 -5.20987749e-01 2.44780824e-01
4.90849853e-01 4.18152958e-02 3.40502918e-01 4.40864503e-01
7.83512831e-01 -3.13595235e-01 4.76775825e-01 -4.49941009e-01
4.04468656e-01 3.94001082e-02 8.56619656e-01 1.26567113e+00
-3.62761170e-01 1.77220583e-01 5.73978186e-01 -2.84304291e-01
-1.02379513e+00 -9.70934331e-01 -5.02663136e-01 1.43384409e+00
5.91730654e-01 -8.86862099e-01 -7.57958591e-01 -8.79606247e-01
4.00974393e-01 9.41606462e-01 -7.17565835e-01 -3.92582625e-01
-3.69076639e-01 -8.01588416e-01 2.90317446e-01 3.88258457e-01
3.67290765e-01 -1.19090807e+00 1.56421289e-01 4.96895462e-01
-4.16493744e-01 -1.24558091e+00 -3.17271590e-01 3.72415274e-01
-7.30423748e-01 -7.70962715e-01 -4.26444858e-01 -7.69339859e-01
2.42562801e-01 4.29578543e-01 1.27548778e+00 3.19453739e-02
-1.29720345e-01 1.47365212e-01 -6.19937062e-01 -6.00280404e-01
-5.32231569e-01 6.06553733e-01 3.29932064e-01 -4.10458684e-01
9.58317995e-01 -2.36938164e-01 6.32139519e-02 3.56672794e-01
-1.09288824e+00 -1.27219670e-02 1.80047691e-01 1.51659095e+00
4.30558354e-01 -1.48676872e-01 8.47249329e-01 -1.37916410e+00
1.15625751e+00 -6.00025833e-01 -3.36685747e-01 2.60015309e-01
-9.79947925e-01 4.01643962e-01 1.71622589e-01 -7.21329033e-01
-8.96267176e-01 -3.04347426e-01 -1.51911855e-01 6.59406856e-02
7.06532672e-02 4.63489085e-01 -4.29189235e-01 3.28467131e-01
7.05664158e-01 3.84713620e-01 3.81337991e-03 -6.62069261e-01
6.83185339e-01 8.24737906e-01 1.01996191e-01 -7.06987321e-01
8.33473444e-01 6.58406690e-02 -7.38803506e-01 -1.08408356e+00
-9.85669672e-01 -5.55014491e-01 -4.10792500e-01 2.68506050e-01
5.86464942e-01 -9.37910080e-01 -1.85681522e-01 1.66570961e-01
-9.90495801e-01 -4.46061462e-01 -4.91838366e-01 7.08644167e-02
-3.26974928e-01 5.44109643e-01 -5.45768142e-01 -1.05333745e+00
-2.70652384e-01 -9.85047817e-01 9.33307469e-01 3.35399032e-01
-4.24197704e-01 -1.05390215e+00 1.19418643e-01 4.23734307e-01
3.39883119e-01 -6.80199087e-01 1.14137244e+00 -1.34014666e+00
-1.70743853e-01 -9.58492383e-02 -6.34507462e-02 -1.24655619e-01
8.08291137e-02 -1.00533175e+00 -8.81960392e-01 -2.47181997e-01
-2.88298756e-01 -6.46557450e-01 1.05444753e+00 2.67077893e-01
8.46180260e-01 -2.21016243e-01 -3.80557954e-01 2.87774771e-01
1.00942552e+00 1.90506920e-01 5.85775316e-01 7.58315921e-01
4.26972300e-01 6.83708072e-01 1.11469924e+00 6.38518810e-01
4.82110351e-01 8.26956391e-01 2.13465422e-01 -1.24710530e-01
3.87777239e-02 -6.52867556e-01 -1.45427451e-01 6.71191156e-01
7.36028731e-01 -4.68444645e-01 -7.45371521e-01 7.71386802e-01
-1.95385110e+00 -3.38197172e-01 2.31219232e-01 1.84898901e+00
1.14606857e+00 4.48632300e-01 -5.75563461e-02 4.85825911e-02
5.60825348e-01 5.86673677e-01 -5.51780641e-01 -3.54709566e-01
-2.00231776e-01 2.57417321e-01 4.14250970e-01 6.38229072e-01
-1.29203570e+00 1.75986218e+00 6.76457453e+00 1.24407685e+00
-4.52693522e-01 1.81672439e-01 3.81830961e-01 3.03073674e-01
-6.91764951e-01 -1.27010554e-01 -1.32611549e+00 3.35902363e-01
4.47428584e-01 -7.42686316e-02 1.22118607e-01 1.16283679e+00
-1.82973459e-01 1.93652779e-01 -6.01132512e-01 8.85356426e-01
-3.60964350e-02 -1.32230103e+00 3.33394647e-01 6.63681701e-02
5.49330056e-01 -3.09946597e-01 -9.80356336e-02 1.13413179e+00
5.53273439e-01 -9.35815692e-01 3.77595484e-01 -2.14279369e-01
5.84165096e-01 -4.76247966e-01 8.00327063e-01 4.56465334e-01
-1.01711500e+00 -1.55207478e-02 -8.61004710e-01 -8.23395327e-02
1.73277393e-01 2.96610713e-01 -1.06242085e+00 2.14046910e-01
4.93816584e-01 4.40614074e-02 -5.53420424e-01 4.72138017e-01
-1.34380028e-01 5.34894288e-01 -2.13956505e-01 -2.24040672e-01
3.68361413e-01 -8.91972631e-02 5.47276556e-01 9.42767143e-01
-1.98910221e-01 2.31282607e-01 4.94014084e-01 8.02278161e-01
6.22267928e-03 6.05161071e-01 -5.88612199e-01 -8.96077752e-02
1.17301571e+00 9.88973677e-01 -3.98424119e-01 -3.95401388e-01
-4.92287785e-01 7.55442142e-01 6.35469377e-01 3.13268423e-01
-3.97992164e-01 -3.22325021e-01 1.14309895e+00 -2.47590870e-01
3.71781081e-01 -2.92903576e-02 -4.49507862e-01 -1.41277802e+00
1.39417136e-02 -9.58416343e-01 4.21566308e-01 -2.75717862e-02
-1.11427140e+00 6.16792500e-01 -5.64051345e-02 -9.21442568e-01
-2.82419413e-01 -5.40714979e-01 -2.95197785e-01 7.38988936e-01
-1.64017975e+00 -8.30010355e-01 2.87221000e-02 2.06355974e-01
1.10881734e+00 -6.74575627e-01 9.63777721e-01 -2.93429848e-03
-6.41227126e-01 1.13687456e+00 1.32272184e-01 2.83768654e-01
7.20205903e-01 -1.30276716e+00 8.81813645e-01 5.08416712e-01
2.54665434e-01 6.35148704e-01 8.81579220e-01 -8.34226012e-01
-9.98235941e-01 -8.72429430e-01 1.02418768e+00 -5.87137401e-01
4.11939174e-01 -7.99423397e-01 -1.24240267e+00 7.07920849e-01
-1.28941849e-01 -1.77900136e-01 8.46203983e-01 6.86400414e-01
-3.67984593e-01 4.95405465e-01 -1.00184357e+00 7.64177978e-01
1.10014808e+00 -5.09944618e-01 -1.23724961e+00 2.02675506e-01
1.25714386e+00 -5.80048382e-01 -5.34842312e-01 5.73194325e-01
1.04145266e-01 -5.52028418e-01 1.04886520e+00 -1.14701426e+00
3.15841809e-02 1.36038572e-01 -2.15230659e-01 -1.48057795e+00
-2.80196041e-01 -5.13309181e-01 -4.90286201e-01 1.45945966e+00
7.09778130e-01 -5.44350863e-01 1.12557817e+00 6.98051393e-01
1.86818056e-02 -7.68868327e-01 -1.00228977e+00 -7.66703010e-01
2.06393212e-01 -6.37771130e-01 6.55881882e-01 1.09247577e+00
3.99140328e-01 8.48763824e-01 -7.80483246e-01 -1.93639367e-04
2.39874303e-01 -1.23189509e-01 1.01877201e+00 -1.34938121e+00
-1.45734087e-01 -3.74050364e-02 -1.22161508e-01 -1.51742268e+00
4.95435506e-01 -7.94636309e-01 6.18335269e-02 -1.40973318e+00
-8.05992335e-02 -7.45853066e-01 -1.35041863e-01 5.02955675e-01
-7.82889128e-01 -1.27786901e-02 2.58306414e-01 4.33868989e-02
-6.45434380e-01 9.06506836e-01 9.01655555e-01 -2.99627393e-01
-6.52271509e-01 -1.70820162e-01 -8.87892127e-01 4.06390339e-01
5.02821445e-01 -3.30225557e-01 -7.20656276e-01 -1.79890633e-01
-1.45116627e-01 -4.37678210e-02 -5.56124210e-01 -3.78210723e-01
-1.36053398e-01 -4.14819755e-02 1.03433325e-03 -5.27218163e-01
4.35738146e-01 -4.46789235e-01 -7.18074977e-01 3.36230308e-01
-7.50552952e-01 -4.07582402e-01 4.06399518e-01 7.06251204e-01
-1.56565011e-01 -3.83499444e-01 5.76893866e-01 2.13672612e-02
-1.12931240e+00 2.63451308e-01 -4.99781430e-01 7.20912278e-01
7.09540963e-01 3.11958976e-02 -3.06927949e-01 -5.37023544e-01
-7.56976068e-01 8.12379062e-01 1.17983960e-01 7.73376167e-01
3.89349610e-01 -1.51082659e+00 -5.02941132e-01 6.36505783e-01
5.77722967e-01 5.06874137e-02 4.07947570e-01 2.34353557e-01
1.03346603e-02 7.21143067e-01 1.89479470e-01 -3.45994264e-01
-1.10356569e+00 7.22407699e-01 2.22346544e-01 -7.11437166e-01
-5.85406899e-01 8.46098244e-01 5.12047112e-01 -1.32350254e+00
6.03904366e-01 -1.23353347e-01 -6.72643781e-01 2.10126221e-01
6.30195081e-01 -3.35808396e-02 -4.64843884e-02 -3.16991031e-01
-3.11602708e-02 6.53714389e-02 -5.25829494e-01 -8.45782086e-02
8.62672925e-01 -4.65492010e-01 3.81218016e-01 3.00482273e-01
7.31650591e-01 3.97634879e-02 -8.60391557e-01 -1.02357316e+00
4.49900836e-01 -5.45689642e-01 -1.53034553e-01 -6.49379730e-01
-4.97417629e-01 5.47327101e-01 4.92802560e-01 5.01247406e-01
3.39965135e-01 1.60244688e-01 1.09531212e+00 6.11993492e-01
3.27706635e-01 -1.48252785e+00 -4.30552475e-03 1.02708113e+00
4.02629912e-01 -1.46351254e+00 -3.58419925e-01 -7.67173052e-01
-8.76174152e-01 7.11215794e-01 1.09433210e+00 1.62882179e-01
6.68341458e-01 -2.98452675e-02 3.49763751e-01 -1.07990623e-01
-8.76809180e-01 -3.48433971e-01 2.42928579e-01 7.43063390e-01
2.22987100e-01 -7.61812776e-02 -8.90797377e-01 1.08104956e+00
-2.93516815e-01 -4.92793232e-01 4.71758127e-01 8.45195532e-01
-6.50590181e-01 -1.45121109e+00 -2.09965259e-01 4.98336792e-01
-2.61062950e-01 -3.14768702e-01 -3.98048311e-01 1.00066102e+00
-3.39261115e-01 7.72607505e-01 1.51189551e-01 -5.28361797e-01
5.43936372e-01 6.59255683e-01 -1.98736966e-01 -9.25868273e-01
-1.32227674e-01 1.13092503e-02 3.83014917e-01 -4.73907620e-01
3.71962674e-02 -3.13154072e-01 -1.12455261e+00 -4.34784358e-03
-6.90756142e-01 6.70054436e-01 3.30998003e-01 1.18557262e+00
3.46342564e-01 5.32683671e-01 4.65631723e-01 -4.37027365e-01
-9.49566722e-01 -1.27673149e+00 -1.03716671e+00 8.12516868e-01
9.91666876e-03 -1.14274371e+00 -2.90814966e-01 -9.02406514e-01] | [12.55858325958252, 7.384539604187012] |
c7d2bed6-2f5a-49da-af77-51a470eb14c1 | extracting-drug-drug-interactions-from | null | null | https://www.sciencedirect.com/science/article/pii/S1532046415000441 | https://www.sciencedirect.com/science/article/pii/S1532046415000441 | Extracting drug–drug interactions from literature using a rich feature-based linear kernel approach | Identifying unknown drug interactions is of great benefit in the early detection of adverse drug reactions. Despite existence of several resources for drug–drug interaction (DDI) information, the wealth of such information is buried in a body of unstructured medical text which is growing exponentially. This calls for developing text mining techniques for identifying DDIs. The state-of-the-art DDI extraction methods use Support Vector Machines (SVMs) with non-linear composite kernels to explore diverse contexts in literature. While computationally less expensive, linear kernel-based systems have not achieved a comparable performance in DDI extraction tasks. In this work, we propose an efficient and scalable system using a linear kernel to identify DDI information. The proposed approach consists of two steps: identifying DDIs and assigning one of four different DDI types to the predicted drug pairs. We demonstrate that when equipped with a rich set of lexical and syntactic features, a linear SVM classifier is able to achieve a competitive performance in detecting DDIs. In addition, the one-against-one strategy proves vital for addressing an imbalance issue in DDI type classification. Applied to the DDIExtraction 2013 corpus, our system achieves an F1 score of 0.670, as compared to 0.651 and 0.609 reported by the top two participating teams in the DDIExtraction 2013 challenge, both based on non-linear kernel methods. | ['W. John Wilbur', 'Lana Yeganova', 'Haibin Liu', 'Sun Kim'] | 2015-03-19 | null | null | null | journal-of-biomedical-informatics-2015-3 | ['drug-drug-interaction-extraction'] | ['natural-language-processing'] | [ 1.55502006e-01 -2.61021823e-01 -6.47007465e-01 -2.49246091e-01
-8.47873390e-01 -6.74692154e-01 6.82544649e-01 1.08060956e+00
-2.56488919e-01 1.08057618e+00 -2.62273669e-01 -7.26067364e-01
-3.40265989e-01 -5.18397987e-01 -4.97323155e-01 -7.89857626e-01
-2.15142071e-01 6.97067499e-01 9.32785869e-02 1.99903026e-01
3.14072669e-01 7.00158238e-01 -1.38206542e+00 5.55192113e-01
1.25393331e+00 8.42867732e-01 1.70468181e-01 4.60480064e-01
-2.89737105e-01 6.06958508e-01 -5.05323172e-01 -2.90724397e-01
-1.89317644e-01 -4.87432629e-01 -8.24697554e-01 -6.65868163e-01
2.06797749e-01 1.60390362e-01 -1.76829379e-02 8.13930452e-01
6.55097723e-01 -5.28064489e-01 1.00694895e+00 -9.51763332e-01
-1.74473956e-01 4.43196118e-01 -4.77414250e-01 2.79047996e-01
6.86393499e-01 -1.32952288e-01 9.56283987e-01 -1.00514627e+00
7.39570737e-01 7.66056359e-01 5.99461079e-01 3.59660447e-01
-1.11467361e+00 -7.95866132e-01 -3.79250616e-01 4.38195467e-01
-1.27025986e+00 -2.48296201e-01 4.74597186e-01 -6.85938954e-01
1.48672676e+00 4.04423535e-01 3.86519521e-01 9.45076287e-01
4.60526526e-01 6.75434947e-01 1.10794771e+00 -5.04563034e-01
3.47849995e-01 5.27291656e-01 3.66493821e-01 4.68939185e-01
4.45858210e-01 -2.06753798e-02 -5.11013329e-01 -9.04259682e-01
1.10276826e-01 1.08055346e-01 -1.42970279e-01 -7.44268075e-02
-1.05122101e+00 9.00535882e-01 -7.08881766e-02 4.99826461e-01
-5.12028873e-01 -6.78117335e-01 7.09671259e-01 1.44052818e-01
7.01571584e-01 5.48779964e-01 -9.90305305e-01 -1.61792383e-01
-8.38111699e-01 1.82501748e-01 1.12943339e+00 5.31838655e-01
1.69456005e-01 -6.03986919e-01 -1.52313057e-02 9.19904888e-01
1.90451145e-01 4.58080769e-01 3.55617195e-01 2.70937413e-01
5.22510350e-01 8.79069805e-01 -4.18389263e-03 -7.46179760e-01
-6.08736694e-01 -2.42051944e-01 -9.09052789e-01 -9.67886951e-03
4.06220973e-01 -2.21848022e-02 -7.42814124e-01 1.27822626e+00
6.55666947e-01 1.76193211e-02 2.76155293e-01 4.15973604e-01
8.39184701e-01 5.35811007e-01 4.42173213e-01 -3.04185897e-01
1.46778846e+00 -3.75783771e-01 -7.67807543e-01 1.65997148e-01
1.03676999e+00 -1.13176501e+00 4.62830007e-01 6.30612075e-01
-7.02194214e-01 4.18464020e-02 -1.20960462e+00 4.18806881e-01
-5.87687552e-01 3.26841444e-01 6.88184619e-01 7.22333848e-01
-3.00539404e-01 6.99547708e-01 -6.04761362e-01 -2.08239630e-01
7.19658732e-01 6.18174255e-01 -6.17608428e-01 -1.29636928e-01
-1.39818704e+00 1.33046687e+00 3.27772170e-01 -2.93707579e-01
-3.07866633e-01 -1.13528991e+00 -7.43382215e-01 -1.85615450e-01
1.92768067e-01 -4.50736135e-01 7.72742093e-01 -3.86958688e-01
-1.24114084e+00 7.65675783e-01 -1.19930722e-01 -4.88751948e-01
4.08759207e-01 -9.91773233e-02 -4.41899478e-01 1.99526697e-01
1.00296021e-01 -1.48004219e-01 4.76701170e-01 -6.13375366e-01
-7.63771355e-01 -7.86512434e-01 -5.02264023e-01 -7.48896822e-02
-2.59062409e-01 3.23720515e-01 9.85062048e-02 -4.80788231e-01
-1.89504892e-01 -8.72499704e-01 -1.73449948e-01 -3.83937359e-01
-5.37716150e-01 -6.59283161e-01 6.46442771e-01 -6.96071923e-01
1.53840399e+00 -1.74629009e+00 -1.92969281e-03 3.35516244e-01
4.36731189e-01 8.07583034e-01 3.08815658e-01 6.11297846e-01
-5.36255121e-01 -1.52899576e-02 -2.08250076e-01 1.19446471e-01
-2.64557391e-01 -2.10138947e-01 -1.16337374e-01 7.94516146e-01
4.71790791e-01 7.14076817e-01 -9.08794999e-01 -4.43294197e-01
3.04880977e-01 5.76314270e-01 -1.87002927e-01 3.01795274e-01
-2.21111163e-01 2.79257983e-01 -5.83685994e-01 7.70330310e-01
7.67692327e-01 -3.49116325e-01 4.68105435e-01 -2.06906289e-01
-1.19734198e-01 4.00182694e-01 -9.72819030e-01 1.24950337e+00
-1.83212280e-01 2.40257904e-01 -5.30013204e-01 -1.32306075e+00
7.34632134e-01 6.25512242e-01 7.82003582e-01 -4.25518215e-01
-4.90859002e-02 6.37096882e-01 1.82247087e-01 -5.56172132e-01
-4.20140922e-01 -1.82139173e-01 6.95222393e-02 -9.58516449e-03
1.04005218e-01 4.36771482e-01 2.14498639e-01 2.56040189e-02
1.34509945e+00 -2.44813785e-01 7.58170366e-01 -3.83654684e-01
7.97916770e-01 9.93812755e-02 3.65349680e-01 3.83851290e-01
6.29951581e-02 1.85784891e-01 6.35989308e-01 -6.44718468e-01
-9.01450813e-01 -6.10773683e-01 -8.48646283e-01 3.63089025e-01
-3.44166785e-01 -4.65377748e-01 -6.75490320e-01 -7.97234595e-01
2.54876733e-01 5.26122153e-01 -5.67877948e-01 5.02492413e-02
-2.99111605e-01 -1.29808307e+00 6.07071519e-01 9.26124007e-02
5.08678481e-02 -7.95481861e-01 -2.96280116e-01 3.86236578e-01
1.96091443e-01 -9.00058568e-01 -1.67034119e-01 5.29251635e-01
-6.58337474e-01 -1.50124681e+00 -8.16182315e-01 -6.82986498e-01
3.88852030e-01 -1.71951756e-01 9.53327537e-01 -2.69674152e-01
-7.92341232e-01 -3.53909135e-01 -3.56881201e-01 -7.22332358e-01
-5.43211877e-01 2.06900790e-01 1.27995059e-01 -2.36773938e-01
9.25052166e-01 -3.73756468e-01 -5.99759519e-01 1.10094450e-01
-7.45715261e-01 -2.15653926e-01 7.26236165e-01 8.74035776e-01
3.15136880e-01 -2.35417977e-01 8.06710839e-01 -1.22023737e+00
6.33091331e-01 -9.47453022e-01 -4.52604055e-01 3.44679594e-01
-1.10432577e+00 3.32961649e-01 7.86677003e-01 -5.80014408e-01
-7.33617544e-01 2.69640684e-01 -5.02189815e-01 1.35362566e-01
-3.98778796e-01 7.27834642e-01 -1.79029450e-01 -8.67069215e-02
7.64697075e-01 1.62073180e-01 7.16006267e-04 -5.57635069e-01
-1.00055374e-01 1.10963416e+00 -2.52564430e-01 -2.16280937e-01
2.44848415e-01 -6.92732930e-02 1.38547823e-01 -8.28467011e-01
-7.31726170e-01 -7.76777327e-01 -3.65656316e-01 4.29489881e-01
7.21453488e-01 -6.49883449e-01 -1.12888372e+00 5.73104024e-01
-1.16232157e+00 1.96176171e-01 3.02901298e-01 7.08661020e-01
-2.05817983e-01 5.34105480e-01 -5.59035361e-01 -6.73824131e-01
-6.51533246e-01 -1.28834021e+00 6.80972457e-01 3.38340811e-02
-5.65333486e-01 -9.15374756e-01 6.21379495e-01 2.59228915e-01
1.99482501e-01 3.12420905e-01 1.30672538e+00 -1.49373686e+00
9.58154947e-02 -4.74999160e-01 -2.25427702e-01 8.12719390e-03
3.74205619e-01 -1.64377004e-01 -7.73558915e-01 -3.56441401e-02
-1.08782016e-01 -8.89284611e-02 6.88410461e-01 4.05175596e-01
7.27551639e-01 -3.27366292e-01 -7.37952948e-01 2.68889576e-01
1.37247598e+00 7.30730712e-01 4.45753992e-01 3.84098053e-01
6.37486875e-01 4.47541088e-01 5.01574636e-01 6.58330560e-01
2.41680115e-01 8.87583077e-01 7.23260492e-02 -3.61686110e-01
2.94332474e-01 8.74002501e-02 -9.90910679e-02 3.86102609e-02
9.59391519e-02 -5.30241504e-02 -1.26707542e+00 5.12354791e-01
-1.84898055e+00 -6.88061059e-01 -4.71006572e-01 2.34231162e+00
1.56428504e+00 -1.17609017e-02 2.20175967e-01 1.97493628e-01
4.20555919e-01 -4.71259326e-01 -6.17940903e-01 -4.79961127e-01
-1.17732078e-01 7.30533600e-01 5.61490774e-01 2.33085260e-01
-1.32016373e+00 6.19009435e-01 5.40445089e+00 1.09045827e+00
-1.17306840e+00 -1.89778551e-01 6.24262929e-01 2.84726858e-01
1.41107664e-01 -3.06152225e-01 -1.08360684e+00 7.48937130e-01
1.29300463e+00 -1.63920090e-01 -1.64555207e-01 7.90138900e-01
7.39705265e-02 -2.09154904e-01 -1.14065087e+00 1.13848269e+00
-4.11577001e-02 -1.70172763e+00 -4.47376892e-02 1.75840914e-01
4.71688181e-01 -8.05582106e-02 -2.17271526e-03 -2.44060457e-01
-1.03182524e-01 -1.19171596e+00 -2.72906087e-02 2.00245172e-01
7.22556233e-01 -1.00925243e+00 1.02770090e+00 2.82365620e-01
-8.27653348e-01 1.73094630e-01 -1.53803483e-01 2.35968769e-01
-2.22213417e-01 1.17085397e+00 -1.20363832e+00 6.94022059e-01
5.13246477e-01 9.06157494e-01 -2.12904245e-01 1.14452958e+00
8.87507871e-02 5.35000861e-01 -2.94513285e-01 -2.44846150e-01
8.82384926e-02 -4.15608250e-02 2.62601167e-01 1.47616768e+00
6.73485994e-02 2.06339151e-01 3.86744482e-03 2.72842556e-01
1.47961944e-01 6.80543065e-01 -6.35273576e-01 -3.48964751e-01
2.30178103e-01 9.92786765e-01 -4.88804132e-01 -5.07306933e-01
-5.66612124e-01 1.06915939e+00 3.09303720e-02 -8.44623968e-02
-7.15883374e-01 -7.44847119e-01 9.31108057e-01 3.56187969e-02
2.96282411e-01 2.29483262e-01 -3.70413870e-01 -1.04315782e+00
3.63600929e-03 -1.14815688e+00 5.94864607e-01 8.01219419e-02
-1.55937839e+00 7.38274992e-01 1.07346820e-02 -1.19836879e+00
-2.28509933e-01 -8.88126910e-01 -2.00721279e-01 1.18284059e+00
-1.49135983e+00 -9.27735686e-01 3.14134508e-01 4.09034252e-01
4.21728790e-01 -1.79013327e-01 1.27476478e+00 6.04495943e-01
-6.41728580e-01 6.34398043e-01 4.00944173e-01 -1.03928261e-01
1.01800883e+00 -1.28510189e+00 1.52152717e-01 4.33861353e-02
-5.82365394e-02 7.84376502e-01 7.92521477e-01 -7.37699032e-01
-1.40878403e+00 -7.76621759e-01 1.51810193e+00 -4.82060879e-01
6.30230546e-01 -1.43734440e-01 -1.10786831e+00 5.57991154e-02
3.20120854e-03 -2.61718303e-01 1.42220914e+00 8.06224346e-02
-4.65631336e-01 2.32010528e-01 -1.33738256e+00 1.25533655e-01
4.54041481e-01 -3.00793678e-01 -6.22053504e-01 8.18808973e-01
1.37979567e-01 -3.83537740e-01 -1.20214629e+00 4.75950003e-01
7.13553727e-01 -6.26599133e-01 1.27178192e+00 -1.04632747e+00
5.31905055e-01 -1.72010884e-01 2.01655999e-01 -1.04245973e+00
7.84310028e-02 -4.00582820e-01 -1.22370131e-01 7.46094644e-01
9.38616276e-01 -7.50549853e-01 7.51894414e-01 7.43883669e-01
3.10779721e-01 -1.21690536e+00 -8.66752803e-01 -6.17673934e-01
1.02436423e-01 -1.09990560e-01 2.38471016e-01 1.14034307e+00
6.55880570e-01 5.60874283e-01 -2.86670834e-01 -1.63332820e-02
6.41131639e-01 1.04961969e-01 2.73551196e-01 -1.51366115e+00
-2.38302499e-01 -3.76623183e-01 -5.87967157e-01 -3.57661307e-01
1.80700779e-01 -9.90075529e-01 -4.50654358e-01 -1.33252645e+00
5.75622261e-01 -4.52104598e-01 -3.48194242e-01 7.09864140e-01
-2.67985612e-01 2.68519949e-02 -3.62012893e-01 1.49518654e-01
-1.10508136e-01 1.74928661e-02 4.49368715e-01 -2.37649038e-01
-5.31004071e-01 1.21997416e-01 -4.74905521e-01 6.14522636e-01
8.17698181e-01 -6.40970230e-01 -2.10330293e-01 4.83073503e-01
2.12920904e-01 -5.06886654e-02 -1.45281345e-01 -4.77367610e-01
2.24391550e-01 -2.45678380e-01 6.35204375e-01 -6.54135466e-01
-1.00733861e-01 -5.84875762e-01 3.48149627e-01 7.84490347e-01
-1.90380782e-01 -1.23709507e-01 5.40263891e-01 4.38534439e-01
-1.76561564e-01 -1.83566302e-01 7.17110515e-01 6.67608455e-02
-2.40751609e-01 2.27168933e-01 -3.92528951e-01 -2.33487144e-01
1.15880525e+00 -1.04781620e-01 -1.17457293e-01 4.33720052e-01
-6.19494498e-01 -1.08271278e-01 9.59896296e-02 2.28264451e-01
4.57503408e-01 -9.46601331e-01 -7.92620122e-01 6.32246286e-02
4.60768729e-01 -5.35990000e-01 1.82808563e-01 1.00583041e+00
-5.32580197e-01 8.82565558e-01 2.97514815e-02 -5.04251301e-01
-1.80627060e+00 5.86526155e-01 6.69953153e-02 -7.19937623e-01
-5.66824257e-01 6.99415326e-01 6.38098493e-02 -1.82442650e-01
2.15445027e-01 -1.42768323e-01 -3.31956953e-01 4.69976068e-01
8.93322587e-01 1.44102097e-01 7.15517998e-01 -3.20609152e-01
-1.03685820e+00 4.57330287e-01 -4.37353283e-01 3.64846468e-01
1.57410491e+00 5.79389453e-01 -3.10711324e-01 1.44041806e-01
1.54035676e+00 -1.07000217e-01 -3.17316473e-01 -1.09240085e-01
4.98227239e-01 -3.09130847e-01 -3.72747108e-02 -1.11032903e+00
-4.42633003e-01 6.47746027e-01 7.30304360e-01 3.66235599e-02
8.48027527e-01 2.11572200e-01 7.46043622e-01 2.92228341e-01
6.47075027e-02 -6.61774695e-01 -4.84727293e-01 1.81716844e-01
5.56027830e-01 -1.38941145e+00 6.09909259e-02 -5.52682757e-01
-5.31451166e-01 1.20037115e+00 -4.97159027e-02 2.89608210e-01
9.27283168e-01 4.82737482e-01 -6.92592487e-02 -3.29709232e-01
-6.39711320e-01 5.10100333e-04 4.21255708e-01 4.45560724e-01
8.09605002e-01 6.71751127e-02 -9.44212794e-01 4.93450552e-01
4.13882077e-01 9.53556523e-02 6.90775365e-02 8.11370730e-01
-1.94742814e-01 -1.72526681e+00 -1.44067377e-01 7.09224939e-01
-9.13577139e-01 -4.44657534e-01 -4.62758243e-01 5.75628281e-01
-3.62334624e-02 8.35083902e-01 -5.63882887e-01 -1.52948573e-01
5.34996569e-01 2.43136719e-01 4.44156915e-01 -5.92567265e-01
-7.53831208e-01 -3.73822637e-02 8.12233314e-02 -4.84012872e-01
-3.93186390e-01 -5.82288742e-01 -1.23962080e+00 -5.89178503e-02
-4.17702764e-01 2.87328303e-01 9.21753764e-01 1.01419258e+00
5.85286677e-01 1.54170454e-01 6.41327441e-01 -2.89581239e-01
-5.87108433e-01 -7.19075918e-01 -5.10960281e-01 3.12886119e-01
3.77565295e-01 -6.84443772e-01 -3.51604939e-01 -1.97430253e-02] | [8.342612266540527, 8.666462898254395] |
9e4ed0a4-b054-4c5a-a3bb-93fc2268256f | seeing-the-forest-and-the-trees-detection-and | 2005.02966 | null | https://arxiv.org/abs/2005.02966v1 | https://arxiv.org/pdf/2005.02966v1.pdf | Seeing the Forest and the Trees: Detection and Cross-Document Coreference Resolution of Militarized Interstate Disputes | Previous efforts to automate the detection of social and political events in text have primarily focused on identifying events described within single sentences or documents. Within a corpus of documents, these automated systems are unable to link event references -- recognize singular events across multiple sentences or documents. A separate literature in computational linguistics on event coreference resolution attempts to link known events to one another within (and across) documents. I provide a data set for evaluating methods to identify certain political events in text and to link related texts to one another based on shared events. The data set, Headlines of War, is built on the Militarized Interstate Disputes data set and offers headlines classified by dispute status and headline pairs labeled with coreference indicators. Additionally, I introduce a model capable of accomplishing both tasks. The multi-task convolutional neural network is shown to be capable of recognizing events and event coreferences given the headlines' texts and publication dates. | ['Benjamin J. Radford'] | 2020-05-06 | null | null | null | null | ['cross-document-coreference-resolution'] | ['natural-language-processing'] | [ 2.69009292e-01 5.15460260e-02 -3.99097502e-01 -6.24767244e-01
-1.63318372e+00 -1.03578222e+00 1.48949468e+00 8.74783874e-01
-8.81270170e-01 9.26115692e-01 1.02449059e+00 -3.12206954e-01
-5.69234014e-01 -6.43449485e-01 -3.82737041e-01 -1.94145203e-01
-5.13135493e-02 1.04452014e+00 4.59224805e-02 -4.00776684e-01
4.72201258e-01 5.80061495e-01 -9.74930704e-01 7.94930100e-01
1.21410817e-01 3.25118154e-01 -4.16753553e-02 6.15189970e-01
-4.84103978e-01 9.00261462e-01 -9.10391927e-01 -5.64745247e-01
-1.73083603e-01 -4.13809806e-01 -1.51371932e+00 -5.61323643e-01
6.27377748e-01 1.42978668e-01 -6.35708094e-01 6.67538285e-01
5.35515130e-01 2.59254754e-01 7.35771060e-01 -1.19739950e+00
-1.19758844e-01 1.47389209e+00 -4.42242742e-01 1.33822083e+00
6.87465906e-01 -8.09231222e-01 1.45034218e+00 -6.35093212e-01
1.40833449e+00 1.42406631e+00 7.91586995e-01 7.19564185e-02
-1.02513504e+00 -8.66357446e-01 -2.01229006e-02 3.26109558e-01
-1.02953827e+00 -5.67395866e-01 6.89353526e-01 -4.86449182e-01
1.36566103e+00 3.22390288e-01 1.53180376e-01 1.44162047e+00
2.84011483e-01 4.87345010e-01 4.37621981e-01 -5.47297716e-01
-3.69150907e-01 -4.21400964e-01 6.64532363e-01 2.77504772e-01
1.13261180e-04 -5.95218092e-02 -1.06842077e+00 -5.29259145e-01
1.70589924e-01 -3.11193049e-01 -2.39389285e-01 3.85968864e-01
-1.32634377e+00 9.85895157e-01 6.50890991e-02 1.03552198e+00
-2.63213009e-01 -7.35919923e-02 1.07043302e+00 3.48992467e-01
6.73026681e-01 6.38095558e-01 -5.21933615e-01 -1.89905003e-01
-1.06121767e+00 6.44026756e-01 1.12643802e+00 6.24033451e-01
3.28989506e-01 -6.80118620e-01 -3.77456099e-01 7.05326200e-01
-9.61842909e-02 1.72067434e-01 2.84123927e-01 -8.24188530e-01
1.05235934e+00 4.90789473e-01 3.00671995e-01 -1.41143453e+00
-1.18674397e+00 -1.92887306e-01 -3.02011311e-01 -5.33368528e-01
6.00396514e-01 -4.34733093e-01 -1.71337292e-01 1.83532512e+00
1.98659584e-01 -3.34725603e-02 2.75137037e-01 5.68697572e-01
1.47985339e+00 7.44713962e-01 2.00189933e-01 -7.85670996e-01
1.61343515e+00 -2.43774608e-01 -1.10102689e+00 -1.95272639e-01
7.28788376e-01 -1.21800876e+00 2.81843930e-01 -1.25124544e-01
-1.21325672e+00 -1.58560112e-01 -7.06286490e-01 -3.51107866e-01
-5.64662576e-01 -1.59742087e-01 5.10989010e-01 -1.75406069e-01
-4.75695133e-01 6.21497989e-01 -6.23335421e-01 -8.37212265e-01
1.47810578e-01 6.71851933e-02 -4.55527157e-01 5.67615390e-01
-1.74177480e+00 1.27631783e+00 6.47357881e-01 -2.66516328e-01
-5.82204878e-01 -7.46309996e-01 -8.85449052e-01 1.62008867e-01
2.30418235e-01 -1.34589896e-01 1.59443021e+00 -5.68860769e-01
-4.90987748e-01 1.54150617e+00 -2.49346644e-01 -7.11263061e-01
3.16356391e-01 -2.68390357e-01 -9.42188144e-01 3.48650008e-01
8.97823930e-01 1.35891154e-01 1.38235494e-01 -8.40354979e-01
-1.17118490e+00 -3.05683345e-01 2.04726704e-03 2.84313828e-01
-2.85421133e-01 1.11841202e+00 -2.04777390e-01 -6.56720400e-01
6.07693791e-02 -6.39719784e-01 2.65188277e-01 -9.15837288e-01
-5.55662632e-01 -9.82545555e-01 8.00852656e-01 -7.00203717e-01
1.22655332e+00 -1.75435913e+00 -9.41387005e-03 -7.28648901e-02
8.35917369e-02 -3.31479639e-01 1.04163781e-01 9.28098202e-01
-5.04299641e-01 1.76836655e-01 3.44295532e-01 -2.71226138e-01
-8.18473694e-04 5.46738617e-02 -8.55870724e-01 6.35484159e-01
-5.16021401e-02 5.59263229e-01 -1.14414322e+00 -8.88470292e-01
-2.07238346e-01 2.32721984e-01 6.69226944e-02 -2.87785470e-01
1.81898579e-01 1.37632415e-01 -5.06128728e-01 1.13666989e-01
1.22812204e-01 -7.64206871e-02 4.23713297e-01 -5.20969808e-01
-4.90727603e-01 1.02539051e+00 -9.13830996e-01 1.55113518e+00
-2.22148970e-01 1.38955283e+00 6.52901083e-02 -1.13076842e+00
8.18910122e-01 8.76913905e-01 6.77095890e-01 -4.88910049e-01
4.14980978e-01 1.66739374e-01 -2.14940131e-01 -3.87142986e-01
8.22135806e-01 -9.07530487e-02 -8.70150208e-01 8.97635162e-01
2.92371541e-01 9.59819257e-02 6.58857822e-01 7.00479209e-01
1.12791455e+00 -3.83596897e-01 3.92747462e-01 -2.34342560e-01
3.03071469e-01 4.97110397e-01 7.67531633e-01 1.03331923e+00
-2.01854751e-01 2.72572875e-01 4.84666854e-01 -7.47299254e-01
-1.00929821e+00 -8.93211544e-01 -4.95294273e-01 1.36924958e+00
-1.10469088e-01 -6.89154446e-01 -3.22517842e-01 -5.98028481e-01
-3.19531918e-01 9.19997454e-01 -7.21841455e-01 4.72565323e-01
-1.09720397e+00 -6.32118702e-01 1.03384399e+00 3.62141728e-01
2.71256834e-01 -1.14294708e+00 -6.82905912e-01 5.23091912e-01
-1.04962921e+00 -1.14725733e+00 -5.21121144e-01 3.96126509e-01
-2.73788031e-02 -1.68490207e+00 -5.24381176e-02 -1.13022768e+00
-2.65091308e-03 -3.25147342e-03 1.23710334e+00 -3.86599481e-01
-2.18063861e-01 3.02965224e-01 -3.37411314e-01 -5.77490270e-01
-5.50595701e-01 3.95808995e-01 1.63001120e-01 -3.39671791e-01
8.70922208e-01 -3.72821659e-01 1.25362292e-01 -5.53236455e-02
-7.15759933e-01 -2.87126362e-01 -2.99295366e-01 8.60046566e-01
-6.88971626e-03 -1.06171072e-01 4.93751019e-01 -9.80901837e-01
9.65682685e-01 -7.39844859e-01 -1.67951688e-01 3.33418578e-01
-1.16727464e-01 -3.89047503e-01 1.08848549e-01 -2.02905327e-01
-1.45828629e+00 -3.46162945e-01 2.05714524e-01 1.60491630e-01
-2.52918929e-01 9.65305805e-01 4.46257949e-01 7.72494912e-01
9.06097889e-01 -3.50063145e-01 -6.01806819e-01 -3.84757400e-01
2.42583886e-01 7.15905070e-01 1.18593299e+00 -7.22694099e-01
6.19012237e-01 5.27915061e-01 -3.32459688e-01 -6.44837618e-01
-1.45303059e+00 -8.14486623e-01 -7.71080673e-01 -3.79175216e-01
8.35175753e-01 -8.52665067e-01 -7.95712829e-01 -9.57781076e-02
-1.78889501e+00 2.67421037e-01 -7.97930881e-02 7.47938573e-01
-4.72291410e-01 2.15530410e-01 -9.62877750e-01 -5.21202147e-01
-4.44881469e-01 -5.22002101e-01 9.34930146e-01 3.06000084e-01
-1.06101680e+00 -1.04444802e+00 4.51251686e-01 4.09012556e-01
-3.09257895e-01 5.71157217e-01 8.65226150e-01 -1.50881183e+00
4.14519697e-01 -3.15071732e-01 -9.99113098e-02 -8.52697968e-01
1.51125342e-01 2.06132025e-01 -5.27491450e-01 -2.11566523e-01
-3.14987153e-01 -4.15142328e-01 8.13081563e-01 5.09205163e-01
2.13562623e-01 -3.21907043e-01 -9.57581460e-01 1.76189005e-01
7.92754471e-01 3.79058212e-01 3.86661999e-02 1.10340095e+00
3.68281193e-02 8.59061599e-01 5.10527909e-01 4.72296804e-01
2.91609764e-01 6.92921996e-01 -2.16932505e-01 6.97175637e-02
2.71291938e-02 -1.10687613e-01 1.27536520e-01 4.74942982e-01
1.22519441e-01 -4.51519340e-01 -1.23065531e+00 1.08789086e+00
-1.97564590e+00 -1.69118738e+00 -5.10026157e-01 1.32419264e+00
1.15817928e+00 5.03024817e-01 -1.68120921e-01 2.50999462e-02
1.24326539e+00 3.05499524e-01 -4.04654182e-02 -3.65703642e-01
-6.35722399e-01 2.13421956e-01 2.64456093e-01 7.01740742e-01
-1.60615623e+00 1.02438807e+00 6.81094933e+00 4.26595926e-01
-5.28324127e-01 1.72744066e-01 1.08542614e-01 -3.96952957e-01
4.05691192e-03 9.67959762e-02 -1.21226537e+00 2.55851030e-01
1.07848465e+00 -8.57328415e-01 -9.26382020e-02 1.48128107e-01
3.12944919e-01 9.11267325e-02 -1.20946097e+00 5.15822470e-01
1.83947474e-01 -2.11797357e+00 -1.26716867e-01 -2.88277000e-01
4.56848353e-01 2.43677244e-01 -5.35968781e-01 2.97583818e-01
9.46462035e-01 -6.35631204e-01 9.35966849e-01 2.50088900e-01
5.07707536e-01 -7.62116730e-01 7.96886146e-01 2.42333055e-01
-1.01601088e+00 8.03211480e-02 1.35493621e-01 1.65190604e-02
7.32595026e-01 6.97543472e-02 -1.01253176e+00 6.28643215e-01
9.89405334e-01 7.12369144e-01 -9.41666514e-02 7.18900919e-01
-1.79776266e-01 6.67927861e-01 -4.18873489e-01 5.29046431e-02
4.71431941e-01 3.95220637e-01 1.01654565e+00 1.73295724e+00
1.16519891e-01 4.87827599e-01 4.61003900e-01 5.01908243e-01
-2.52325594e-01 8.29252228e-02 -6.43043220e-01 -1.86709732e-01
9.42187905e-01 1.18000042e+00 -9.37857509e-01 -3.28130722e-01
-2.00989082e-01 4.14964825e-01 4.57174271e-01 2.53703892e-01
-5.90097845e-01 -7.12346792e-01 4.44176108e-01 -2.28087068e-01
-5.97768165e-02 -2.67111659e-01 1.05394376e-02 -8.97448599e-01
-5.48588336e-01 -6.09449685e-01 1.39247727e+00 -8.05302441e-01
-1.46295416e+00 6.96672678e-01 3.11484396e-01 -7.01874197e-01
-5.04433334e-01 -9.57411602e-02 -8.65102172e-01 7.32832193e-01
-9.36724424e-01 -1.11783326e+00 1.45186096e-01 7.32516110e-01
7.78510451e-01 -3.20750564e-01 8.63592803e-01 3.63584399e-01
-5.30176640e-01 2.80315071e-01 2.94003543e-02 9.00095463e-01
1.26037169e+00 -8.98013353e-01 3.52247894e-01 6.80287600e-01
3.71835738e-01 7.91760147e-01 1.20415819e+00 -9.90815341e-01
-6.90992832e-01 -7.68616736e-01 1.94355071e+00 -6.37855113e-01
1.26190877e+00 -2.43357476e-02 -6.78993165e-01 1.17834580e+00
6.88027322e-01 -7.45130599e-01 9.58666205e-01 6.69201493e-01
-4.76774275e-01 2.86892444e-01 -8.44934642e-01 5.14721572e-01
9.57345426e-01 -8.70012343e-01 -1.71826673e+00 1.00280821e+00
5.61921000e-01 -6.93908393e-01 -9.58539486e-01 8.96065235e-02
1.92371726e-01 -1.49909288e-01 9.23502088e-01 -1.38100266e+00
5.85747957e-01 1.12600597e-02 -6.17897846e-02 -1.07717443e+00
-4.04845744e-01 -8.27823520e-01 3.21047664e-01 1.86009824e+00
5.59274971e-01 -2.65619159e-01 3.24205428e-01 4.00086433e-01
-2.91047096e-01 3.74984026e-01 -1.41820312e+00 -4.16772246e-01
1.64846018e-01 -5.00151992e-01 2.27490500e-01 1.74586618e+00
7.01179385e-01 9.86915171e-01 -6.02672109e-04 2.19928801e-01
3.83749098e-01 5.87176859e-01 2.67428160e-01 -1.61826837e+00
2.94960946e-01 -6.09968841e-01 2.11363643e-01 -2.07616225e-01
7.35608041e-01 -1.16039610e+00 7.12205917e-02 -1.58710039e+00
3.20959121e-01 -1.79592937e-01 -6.19165748e-02 5.46668947e-01
2.13585585e-01 6.40013590e-02 -5.88441156e-02 6.15633011e-01
-5.59977949e-01 -1.06022671e-01 3.01647753e-01 -4.12449390e-01
-2.56042808e-01 -3.83730084e-01 -4.96747792e-01 9.68684316e-01
5.59189498e-01 -8.21465075e-01 3.35612744e-01 -3.79588693e-01
3.88165891e-01 5.04157364e-01 9.45957080e-02 -5.95507145e-01
7.13354170e-01 -3.06735903e-01 2.60214776e-01 -1.11337364e+00
-2.42735762e-02 -1.75882205e-01 1.23461686e-01 1.93828478e-01
-1.07559752e+00 2.24821627e-01 3.04426819e-01 4.99400884e-01
-4.00292754e-01 -4.25933689e-01 3.98773074e-01 -1.61617905e-01
-6.09875143e-01 -1.46364868e-01 -9.21224952e-01 5.32167733e-01
8.13634694e-01 2.09244877e-01 -7.95591772e-01 -3.37503821e-01
-9.10557091e-01 3.84398907e-01 -2.35888347e-01 7.02230513e-01
1.07256763e-01 -1.33186460e+00 -1.23110819e+00 -7.10581899e-01
3.17381248e-02 -4.34731990e-01 -7.81120826e-03 5.48790216e-01
-2.89904207e-01 6.84566140e-01 -1.03309318e-01 -1.73718750e-01
-1.72095132e+00 2.78273523e-01 1.24758229e-01 -4.73997235e-01
-7.35004187e-01 8.31800520e-01 -2.90669858e-01 -3.72763455e-01
3.67283642e-01 4.85719182e-02 -7.41069078e-01 1.02911925e+00
6.49966776e-01 1.74062535e-01 4.63834219e-02 -1.00398231e+00
-5.76892495e-01 5.93435653e-02 -4.50866610e-01 -3.87578547e-01
1.69055855e+00 -2.28105143e-01 -9.97757688e-02 5.86791217e-01
1.10949993e+00 -4.14697528e-02 -4.30827022e-01 -6.70102239e-01
6.17176771e-01 1.00190774e-01 1.76261827e-01 -7.30347216e-01
-3.89359266e-01 3.76469553e-01 -3.55892107e-02 7.77076185e-01
3.35286736e-01 4.60440785e-01 8.06320906e-01 6.75925434e-01
1.05399294e-02 -1.39153814e+00 -9.91976038e-02 1.10610509e+00
1.16175807e+00 -1.04802203e+00 8.16218853e-02 -2.01520935e-01
-3.36114794e-01 1.56139207e+00 2.82729805e-01 1.94194898e-01
5.68646848e-01 4.62852448e-01 3.55860293e-02 -8.75404716e-01
-7.38661706e-01 -1.39370650e-01 9.45011228e-02 2.08608121e-01
7.32972443e-01 -2.89950430e-01 -8.82632077e-01 7.02386081e-01
-3.26927543e-01 -4.14210767e-01 6.62612617e-01 8.49395335e-01
-3.50877464e-01 -7.98115790e-01 -5.43376684e-01 3.14440370e-01
-8.70820582e-01 -1.35120004e-01 -7.73334861e-01 9.58414972e-01
4.84385900e-02 1.26105857e+00 7.89923489e-01 6.61346763e-02
3.46345574e-01 2.91067123e-01 8.55140463e-02 -5.20751834e-01
-1.21177077e+00 2.04735342e-02 1.07401133e+00 -1.99219882e-01
-9.42241311e-01 -1.31804204e+00 -1.39696157e+00 -4.79347944e-01
-2.26380855e-01 7.44839072e-01 3.71609271e-01 1.33950841e+00
5.41136041e-02 4.85429466e-01 3.12372386e-01 -6.81308866e-01
-9.41682905e-02 -1.02300441e+00 -4.03122097e-01 7.86099613e-01
4.85626981e-02 -6.88387990e-01 -2.54604518e-01 1.99464113e-01] | [9.129959106445312, 9.538833618164062] |
b143a603-e31d-4376-b46c-4481fc5c5ebb | variational-quantum-classifiers-for-natural | 2303.02469 | null | https://arxiv.org/abs/2303.02469v1 | https://arxiv.org/pdf/2303.02469v1.pdf | Variational Quantum Classifiers for Natural-Language Text | As part of the recent research effort on quantum natural language processing (QNLP), variational quantum sentence classifiers (VQSCs) have been implemented and supported in lambeq / DisCoPy, based on the DisCoCat model of sentence meaning. We discuss in some detail VQSCs, including category theory, DisCoCat for modeling sentence as string diagram, and DisCoPy for encoding string diagram as parameterized quantum circuit. Many NLP tasks, however, require the handling of text consisting of multiple sentences, which is not supported in lambeq / DisCoPy. A good example is sentiment classification of customer feedback or product review. We discuss three potential approaches to variational quantum text classifiers (VQTCs), in line with VQSCs. The first is a weighted bag-of-sentences approach which treats text as a group of independent sentences with task-specific sentence weighting. The second is a coreference resolution approach which treats text as a consolidation of its member sentences with coreferences among them resolved. Both approaches are based on the DisCoCat model and should be implementable in lambeq / DisCoCat. The third approach, on the other hand, is based on the DisCoCirc model which considers both ordering of sentences and interaction of words in composing text meaning from word and sentence meanings. DisCoCirc makes fundamental modification of DisCoCat since a sentence in DisCoCirc updates meanings of words, whereas all meanings are static in DisCoCat. It is not clear if DisCoCirc can be implemented in lambeq / DisCoCat without breaking DisCoCat. | ['Daniel T. Chang'] | 2023-03-04 | null | null | null | null | ['coreference-resolution'] | ['natural-language-processing'] | [ 4.47408736e-01 3.50540340e-01 2.67473534e-02 -5.16544044e-01
-5.38092732e-01 -9.55582857e-01 7.60386288e-01 5.68415999e-01
-3.24463040e-01 6.56253219e-01 3.15646410e-01 -5.61438799e-01
-4.85533237e-01 -9.44122612e-01 -5.34150004e-01 -6.41571999e-01
5.05138457e-01 3.84331584e-01 4.32108432e-01 -8.29724073e-01
6.51937425e-01 1.00889035e-01 -1.73315811e+00 8.68300676e-01
7.61180937e-01 6.50428951e-01 1.30632013e-01 8.72738242e-01
-6.65204644e-01 8.29061389e-01 -5.54043472e-01 -7.62745619e-01
-9.83273759e-02 -5.86190462e-01 -1.24952471e+00 -3.00278723e-01
2.46568903e-01 5.02278984e-01 -8.27463344e-02 1.39105296e+00
3.58206928e-01 -1.35496199e-01 7.86000133e-01 -1.13382721e+00
-6.72007263e-01 1.02705920e+00 -1.07131429e-01 3.62383842e-01
7.42639065e-01 -5.23060262e-01 1.41919136e+00 -4.88995314e-01
7.33666182e-01 1.58890438e+00 4.57583427e-01 4.96758878e-01
-1.41234183e+00 -4.30263400e-01 -3.12372237e-01 2.61905909e-01
-1.13432920e+00 -2.07035959e-01 5.23397207e-01 -4.13087100e-01
1.42790926e+00 6.68345630e-01 6.34431422e-01 8.59622777e-01
8.28977942e-01 4.26110059e-01 1.05390704e+00 -8.47611785e-01
4.50174272e-01 4.29663837e-01 7.56881833e-01 2.03133419e-01
2.75749743e-01 -1.06222361e-01 -6.71235919e-01 -4.17771578e-01
1.14551760e-01 -5.24473190e-01 -1.90425456e-01 -2.14405000e-01
-1.02568972e+00 1.01375473e+00 -8.52568373e-02 8.03675473e-01
-5.78525215e-02 6.83895722e-02 7.24542618e-01 8.00446153e-01
3.91908705e-01 3.33102614e-01 -3.80462348e-01 -1.79224372e-01
-6.93886757e-01 3.78320396e-01 9.35412169e-01 9.62912619e-01
5.96545696e-01 -6.51170850e-01 -1.63473368e-01 5.71713924e-01
3.51538479e-01 5.53665876e-01 3.90385032e-01 -1.08462632e+00
2.93539196e-01 4.37454641e-01 -1.70196503e-01 -6.83497012e-01
-5.03552258e-01 1.61354486e-02 -5.95682919e-01 -2.06844941e-01
1.46376386e-01 9.68287960e-02 -2.59974748e-01 1.71058846e+00
1.90493330e-01 -2.52977192e-01 6.09464228e-01 5.72041452e-01
8.92266631e-01 6.00496531e-01 -1.44684196e-01 -5.87853968e-01
1.60537839e+00 -1.93290129e-01 -9.39326644e-01 4.19905812e-01
1.30604410e+00 -1.03646541e+00 5.52141726e-01 6.50060475e-01
-8.86991203e-01 -2.42895856e-01 -1.04305887e+00 -2.68181056e-01
-5.71915746e-01 -6.04405224e-01 6.69961333e-01 9.89945471e-01
-1.09476578e+00 7.02661514e-01 -4.20553088e-01 -5.48913479e-01
-1.79371908e-01 4.87809271e-01 -1.52667016e-01 4.31567043e-01
-1.63932705e+00 9.67912912e-01 5.25486946e-01 -3.71254422e-02
-1.90264955e-02 -3.38681728e-01 -9.18262661e-01 3.01643666e-02
1.73955604e-01 -4.86693323e-01 1.29525924e+00 -8.77927363e-01
-1.49399292e+00 1.06725228e+00 -2.51805574e-01 -5.68623424e-01
2.96878457e-01 3.34843546e-01 -5.73324859e-01 1.79749727e-01
2.22635686e-01 4.89583939e-01 5.77740967e-01 -8.26962352e-01
-5.30615628e-01 -2.92412370e-01 3.71788204e-01 1.65294349e-01
6.67144218e-03 2.07457915e-01 4.09772575e-01 -9.30497646e-02
3.20198715e-01 -8.08468878e-01 4.51656610e-01 -5.12390494e-01
-2.47036085e-01 -8.88872087e-01 8.36592078e-01 1.30414233e-01
1.08089149e+00 -2.04494381e+00 3.00030738e-01 -2.65412647e-02
3.24439518e-02 -4.45060767e-02 5.71931414e-02 9.81281698e-01
-3.47308993e-01 3.86931032e-01 -3.28703284e-01 -2.19984472e-01
3.56552452e-01 6.96391165e-01 -7.20872521e-01 4.83388811e-01
5.28022572e-02 6.77964032e-01 -9.72477138e-01 -7.91210592e-01
1.42310306e-01 1.29408717e-01 -7.62356102e-01 -5.43410659e-01
-1.89049676e-01 1.07143752e-01 -5.31316340e-01 6.52065724e-02
9.04381990e-01 1.27229527e-01 4.54672456e-01 -1.88939393e-01
-5.91727734e-01 6.08847618e-01 -1.27507818e+00 1.65371823e+00
-2.12781757e-01 6.24027491e-01 1.56735197e-01 -1.39303374e+00
6.90473497e-01 5.81356943e-01 3.15790832e-01 -6.21465385e-01
3.03106129e-01 2.90918112e-01 7.12245032e-02 -7.04211593e-01
1.00789273e+00 -6.81124330e-01 -4.39863205e-01 6.17682755e-01
3.08873385e-01 -8.78313601e-01 5.09158909e-01 5.44031203e-01
8.54911983e-01 1.63499743e-01 2.41173849e-01 -7.59408832e-01
8.29538107e-01 -7.55311688e-03 4.35818374e-01 9.06495571e-01
-3.62720788e-01 4.43843722e-01 1.03908646e+00 -1.04829773e-01
-8.15975964e-01 -1.04499674e+00 -8.95988226e-01 8.39509189e-01
3.16415548e-01 -7.11405993e-01 -7.66637683e-01 -4.94033039e-01
-3.02614301e-01 1.14499366e+00 -3.99483383e-01 -1.22728020e-01
-2.99169898e-01 -8.81084383e-01 4.25529838e-01 -8.13082978e-02
2.09833443e-01 -9.72716808e-01 -5.93446910e-01 1.02822870e-01
-3.14261168e-01 -1.10769773e+00 -1.20286755e-01 3.17580879e-01
-6.10486627e-01 -1.03853750e+00 -7.88613409e-02 -5.71988583e-01
1.96495056e-01 2.18422681e-01 8.77212763e-01 -3.77975293e-02
3.51994112e-02 5.11968613e-01 -8.45903814e-01 -5.78917861e-01
-9.75628138e-01 -1.46631405e-01 2.80356288e-01 -1.35918949e-02
6.59987926e-01 -4.76833999e-01 -1.97023526e-01 -1.81436926e-01
-1.13996494e+00 -9.09649208e-02 -5.95771298e-02 6.94970846e-01
1.94100261e-01 3.34470421e-02 3.48159224e-01 -9.98069823e-01
6.69186652e-01 -2.84037352e-01 -3.26625258e-01 4.00639892e-01
-4.62363362e-01 1.79275423e-01 5.10879993e-01 -7.91742355e-02
-1.03412080e+00 -5.75186133e-01 -3.32974494e-01 1.12693273e-01
2.05772072e-02 5.02762794e-01 1.31698847e-01 1.95183694e-01
7.24285424e-01 1.37465343e-01 1.45410525e-03 -4.46994044e-02
3.25371444e-01 1.25079846e+00 -4.28350968e-03 -6.36786342e-01
2.14569911e-01 6.10016704e-01 6.80906102e-02 -1.21045041e+00
-9.33213592e-01 -4.70859259e-01 -7.51023114e-01 -4.55765389e-02
1.07761502e+00 -4.50038522e-01 -9.83693540e-01 1.23395272e-01
-1.52885044e+00 2.41270617e-01 -4.76680458e-01 7.64674008e-01
-7.15826750e-01 7.88346350e-01 -4.61595029e-01 -1.05458927e+00
-2.61059344e-01 -1.09410322e+00 1.21362615e+00 5.99020645e-02
-5.64924359e-01 -1.03977299e+00 4.95461375e-02 5.26011586e-01
5.01451716e-02 -2.08338201e-01 1.09948444e+00 -9.19518948e-01
-3.92121151e-02 -1.25314236e-01 3.89351547e-02 3.93416047e-01
-1.52800679e-01 5.11995256e-02 -9.02955890e-01 -2.15153396e-01
4.91672635e-01 -3.12134385e-01 6.69412196e-01 2.37574846e-01
3.91103566e-01 2.43856400e-01 -8.41814503e-02 -1.01891950e-01
1.24377978e+00 2.46242702e-01 5.74316442e-01 1.01430394e-01
-3.16015892e-02 9.43951845e-01 4.37414199e-01 9.10525173e-02
4.31969732e-01 5.75539887e-01 2.27998286e-01 5.32752752e-01
1.09011017e-01 1.80407956e-01 5.76459467e-01 1.25392747e+00
8.38273242e-02 -1.29470512e-01 -6.66492403e-01 1.63650528e-01
-1.87111485e+00 -1.29609549e+00 -9.20242548e-01 1.91313946e+00
7.66346633e-01 4.25150722e-01 -3.25640231e-01 2.63340145e-01
8.25304747e-01 1.86460868e-01 3.41270566e-02 -1.33778691e+00
-2.63427317e-01 4.60451514e-01 1.34288862e-01 7.67429650e-01
-7.44730234e-01 9.25350249e-01 5.80535126e+00 9.30657268e-01
-8.14971805e-01 5.35089493e-01 -1.80045411e-01 1.07369214e-01
-5.47120273e-01 5.62363684e-01 -6.74909174e-01 2.57196277e-01
1.06947577e+00 -3.39259356e-01 2.06151411e-01 2.50036508e-01
2.36630961e-01 -4.42376971e-01 -1.11895359e+00 8.44841063e-01
1.03493482e-01 -1.39497614e+00 1.58595979e-01 6.40941262e-02
4.31735724e-01 -8.40728655e-02 -3.54372352e-01 4.10524756e-01
-1.60077438e-01 -4.30743277e-01 1.01465154e+00 5.06396353e-01
4.21718061e-01 -4.81894672e-01 1.09227657e+00 4.25112635e-01
-1.01984978e+00 -1.48159927e-02 -6.61293983e-01 -3.64615709e-01
2.24819407e-01 3.34650844e-01 -7.36768777e-03 1.14536655e+00
5.76864481e-01 4.96012241e-01 -1.84479535e-01 2.77470976e-01
-4.67792302e-02 4.35154349e-01 -3.56433541e-02 -5.10299146e-01
1.73075020e-01 -5.95027626e-01 8.37588251e-01 1.21229863e+00
1.44540310e-01 2.68322468e-01 -2.12642416e-01 9.56090450e-01
4.27216440e-01 2.38000035e-01 -4.86578375e-01 -8.92311633e-02
2.88864285e-01 1.07235730e+00 -8.16955566e-01 -4.76763785e-01
-3.45672786e-01 8.04744780e-01 -1.24627687e-01 -7.46946782e-02
-7.25136995e-01 -6.18508816e-01 3.40959847e-01 -2.35310733e-01
1.99842468e-01 4.41460386e-02 -1.35001749e-01 -1.40920556e+00
-1.26792774e-01 -6.51251674e-01 2.39510283e-01 -7.80614197e-01
-1.23287570e+00 4.44489568e-01 4.29856539e-01 -9.80267167e-01
5.83941713e-02 -7.88451314e-01 -6.85901403e-01 6.93876028e-01
-9.94671404e-01 -9.08297002e-01 4.20847625e-01 4.44714010e-01
3.24947953e-01 1.94810927e-01 1.02902865e+00 -1.33438945e-01
-3.27327609e-01 2.16836289e-01 2.21072346e-01 -1.91962078e-01
5.14190316e-01 -1.38789487e+00 -9.02439468e-03 4.65732634e-01
2.32158169e-01 7.66770184e-01 1.27139556e+00 -4.38169420e-01
-1.26193178e+00 -5.72972417e-01 1.65035272e+00 -5.90393484e-01
9.06912029e-01 -8.32620561e-01 -8.43543649e-01 3.30473661e-01
5.25561929e-01 -5.09992719e-01 7.84958363e-01 8.81139487e-02
-1.94117934e-01 6.94109276e-02 -9.91620779e-01 3.60910147e-01
8.98987174e-01 -9.42982137e-01 -1.42059946e+00 8.65935981e-01
1.01782572e+00 -2.68971473e-01 -9.94756699e-01 2.30276555e-01
4.00696874e-01 -1.18681014e+00 3.07899982e-01 -4.26368207e-01
1.78185061e-01 -1.29231334e-01 -4.78780091e-01 -9.08315361e-01
2.36977056e-01 -7.98216224e-01 5.94933689e-01 1.25459218e+00
3.87539476e-01 -1.08004463e+00 2.45391458e-01 5.44329405e-01
-2.50052780e-01 -2.67433792e-01 -1.45410883e+00 -7.52938688e-01
6.64279580e-01 -9.76899803e-01 2.88050741e-01 9.67523634e-01
8.80098820e-01 7.52193272e-01 2.25292057e-01 -3.37561786e-01
4.86601114e-01 2.10953027e-01 1.98851258e-01 -1.34242952e+00
-1.92437336e-01 -3.56048942e-01 -4.57150340e-01 -5.40022016e-01
4.15693194e-01 -1.40109730e+00 -2.04674408e-01 -1.41966784e+00
2.16903836e-01 5.17213829e-02 4.32708720e-03 -1.42886725e-04
5.98682702e-01 -1.37631670e-01 3.57493877e-01 2.57290721e-01
-5.27838349e-01 5.08565128e-01 9.75506544e-01 -1.52821764e-01
-1.51703030e-01 -1.56424895e-01 -5.75827599e-01 6.39603376e-01
5.72855473e-01 -8.25714886e-01 -3.96463096e-01 4.47374582e-02
8.42351377e-01 1.78233847e-01 2.68654525e-01 -5.36612809e-01
3.12603235e-01 5.36389537e-02 -4.25744236e-01 -5.84018409e-01
3.37812662e-01 -2.50576466e-01 -8.68765339e-02 6.13859951e-01
-3.07457149e-01 -1.10891357e-01 2.21451130e-02 5.02450466e-01
-1.53963909e-01 -8.22706640e-01 5.66031635e-01 -2.17558086e-01
-3.14823627e-01 -6.04460120e-01 -8.10311258e-01 -4.17368598e-02
9.32500482e-01 -2.52213955e-01 -5.73680043e-01 -1.17588773e-01
-8.33252132e-01 2.77077258e-01 2.91166663e-01 4.75871414e-01
3.36330444e-01 -8.11800659e-01 -5.22194147e-01 -3.78558226e-02
-2.32785027e-02 -3.06805104e-01 3.83921266e-01 1.19741023e+00
-3.78990322e-01 1.09539545e+00 2.31122971e-01 -7.02068627e-01
-1.34886849e+00 6.69539034e-01 3.84167284e-01 -1.05230119e-02
-3.96543533e-01 6.17330074e-01 3.47779803e-02 -6.69687450e-01
-3.21087033e-01 -5.01823902e-01 -4.56067115e-01 4.11969125e-01
3.31245065e-01 8.95933434e-02 1.89903736e-01 -1.14819908e+00
-4.71113712e-01 7.66797841e-01 -5.18616326e-02 -3.84155810e-01
9.25932050e-01 -2.64206350e-01 -9.31055248e-01 9.00169253e-01
1.24627125e+00 -7.83803239e-02 -1.57835022e-01 8.50250199e-03
1.69184834e-01 2.50069708e-01 9.59522948e-02 -4.21840847e-01
5.99363744e-02 8.15580189e-01 4.05024767e-01 9.17554915e-01
7.17560172e-01 2.24839643e-01 4.72672611e-01 5.01628518e-01
6.45182073e-01 -1.28480971e+00 -3.67227197e-01 7.15942323e-01
9.23649609e-01 -7.91989982e-01 -1.44087926e-01 -6.51592195e-01
-4.64836091e-01 1.45123243e+00 -3.21202949e-02 -3.84568311e-02
7.99615681e-01 -1.29101560e-01 -1.53564066e-01 -5.13025820e-01
-9.43082929e-01 -2.00866178e-01 -9.91079211e-03 1.07849047e-01
7.63586044e-01 1.07736379e-01 -1.30552018e+00 5.16223848e-01
-5.60189843e-01 -1.44561365e-01 9.57648277e-01 1.20837939e+00
-5.36751211e-01 -1.42033923e+00 -6.04308724e-01 2.68522173e-01
-5.31695485e-01 -2.60435015e-01 -5.65741003e-01 7.06808507e-01
5.31770051e-01 1.66979277e+00 3.25394869e-01 -3.89757842e-01
3.09973478e-01 3.78785461e-01 6.94375277e-01 -9.01248395e-01
-8.04608643e-01 8.91277716e-02 2.52530366e-01 -2.12260872e-01
-8.48680437e-01 -9.03200030e-01 -1.60935128e+00 -3.46987009e-01
-7.01931477e-01 8.33997965e-01 7.08252609e-01 1.42823553e+00
1.70946300e-01 3.38647366e-01 2.74895370e-01 -6.45887792e-01
-8.33778381e-01 -1.09467220e+00 -8.91316295e-01 2.35446557e-01
1.55401155e-01 -5.38523793e-01 -6.07015550e-01 -2.37671092e-01] | [5.751733303070068, 5.047903060913086] |
47c52457-29b0-4bdd-a9ea-73913267adff | a-cad-system-for-colorectal-cancer-from-wsi-a | 2301.02608 | null | https://arxiv.org/abs/2301.02608v1 | https://arxiv.org/pdf/2301.02608v1.pdf | A CAD System for Colorectal Cancer from WSI: A Clinically Validated Interpretable ML-based Prototype | The integration of Artificial Intelligence (AI) and Digital Pathology has been increasing over the past years. Nowadays, applications of deep learning (DL) methods to diagnose cancer from whole-slide images (WSI) are, more than ever, a reality within different research groups. Nonetheless, the development of these systems was limited by a myriad of constraints regarding the lack of training samples, the scaling difficulties, the opaqueness of DL methods, and, more importantly, the lack of clinical validation. As such, we propose a system designed specifically for the diagnosis of colorectal samples. The construction of such a system consisted of four stages: (1) a careful data collection and annotation process, which resulted in one of the largest WSI colorectal samples datasets; (2) the design of an interpretable mixed-supervision scheme to leverage the domain knowledge introduced by pathologists through spatial annotations; (3) the development of an effective sampling approach based on the expected severeness of each tile, which decreased the computation cost by a factor of almost 6x; (4) the creation of a prototype that integrates the full set of features of the model to be evaluated in clinical practice. During these stages, the proposed method was evaluated in four separate test sets, two of them are external and completely independent. On the largest of those sets, the proposed approach achieved an accuracy of 93.44%. DL for colorectal samples is a few steps closer to stop being research exclusive and to become fully integrated in clinical practice. | ['Jaime S. Cardoso', 'Isabel M. Pinto', 'Inti Zlobec', 'Stefan Reinhard', 'Sofia Gonçalves', 'Liliana Ribeiro', 'João Monteiro', 'Ana Monteiro', 'João Fraga', 'Domingos Oliveira', 'Sara P. Oliveira', 'Diana Montezuma', 'Pedro C. Neto'] | 2023-01-06 | null | null | null | null | ['whole-slide-images'] | ['computer-vision'] | [ 2.23373383e-01 5.06947577e-01 -3.13040167e-02 -1.16058059e-01
-7.93413520e-01 -2.68089414e-01 5.59091032e-01 4.45285916e-01
-4.55437958e-01 6.36751175e-01 7.64672756e-02 -4.76663083e-01
-4.39299166e-01 -6.28349483e-01 -4.12172854e-01 -9.62586462e-01
-5.62566370e-02 6.05800211e-01 2.34003738e-01 1.12678967e-01
2.21541021e-02 7.74100482e-01 -1.45570421e+00 2.99416900e-01
8.01208675e-01 1.09718573e+00 2.92845935e-01 5.20835042e-01
-1.13992438e-01 7.17451513e-01 -4.69648540e-01 -1.45195484e-01
1.20164633e-01 -3.22846889e-01 -7.51434684e-01 2.86689490e-01
1.18800901e-01 -1.38928801e-01 4.82719466e-02 7.77642429e-01
5.16883314e-01 -3.19540918e-01 5.05227387e-01 -7.61360347e-01
-2.97991514e-01 4.72454548e-01 -5.38213968e-01 -8.08065534e-02
7.99048990e-02 3.34542245e-01 8.41645896e-01 -7.09454954e-01
9.62907195e-01 5.96708775e-01 6.80767655e-01 2.56875932e-01
-8.42045784e-01 -3.20671827e-01 -1.78564191e-01 7.76458830e-02
-1.51806152e+00 -2.64348328e-01 6.39340043e-01 -5.09002507e-01
6.39870882e-01 4.04730916e-01 1.00637126e+00 7.93065429e-01
7.70263895e-02 7.30777860e-01 1.18813729e+00 -7.98021257e-01
5.24840117e-01 7.17584670e-01 1.73981190e-02 6.57784641e-01
4.79852885e-01 -1.30829528e-01 -9.38226208e-02 -4.85744560e-03
4.39714402e-01 5.54803461e-02 -1.97216824e-01 -3.41470420e-01
-1.14271319e+00 7.05312490e-01 4.59527224e-01 7.40728796e-01
-3.86037081e-02 -3.45744401e-01 5.21833181e-01 1.24809127e-02
4.65896010e-01 4.42318946e-01 -1.79835975e-01 2.00841278e-01
-1.02872837e+00 -1.57926247e-01 7.09960639e-01 5.26061296e-01
4.58562642e-01 -4.24598098e-01 1.65455163e-01 4.87615079e-01
2.15902954e-01 2.63109624e-01 7.45778620e-01 -2.80707389e-01
7.48455822e-02 1.12598085e+00 -7.77412206e-02 -9.88427043e-01
-8.03370416e-01 -7.00679600e-01 -1.05336797e+00 1.07794277e-01
6.70143664e-01 -3.72082405e-02 -8.40948522e-01 1.37550914e+00
4.45422232e-01 -7.43529806e-03 -1.33241534e-01 8.38448703e-01
6.09077573e-01 1.33721143e-01 -9.07863900e-02 5.07438816e-02
1.48294032e+00 -6.78517044e-01 -5.86262465e-01 7.48662651e-02
1.11229014e+00 -5.66234946e-01 9.28486824e-01 5.06604135e-01
-7.40177095e-01 -4.89897728e-01 -1.22134101e+00 6.61934316e-02
-7.05092311e-01 5.55178523e-01 7.88872242e-01 7.33224094e-01
-1.19042528e+00 1.79473445e-01 -9.28159416e-01 -7.74213195e-01
5.90725124e-01 4.03850675e-01 -6.20657742e-01 -1.31127924e-01
-7.67953813e-01 8.60379636e-01 2.48004884e-01 3.56147230e-01
-6.32481635e-01 -5.76530099e-01 -6.26042366e-01 -6.97463676e-02
2.86069572e-01 -6.25895679e-01 4.37609076e-01 -1.13825703e+00
-1.32121170e+00 1.23233068e+00 8.71325433e-02 -2.44831264e-01
7.36236393e-01 1.29581556e-01 -4.22923118e-01 2.05452487e-01
1.96220800e-02 4.53566700e-01 2.15152338e-01 -1.08249366e+00
-6.52949870e-01 -6.54206634e-01 -1.44684151e-01 1.25205121e-03
-5.75476825e-01 -4.53733832e-01 -5.93614459e-01 -3.86022508e-01
-2.42186442e-01 -1.10187423e+00 -3.45267743e-01 4.20464307e-01
-5.44420183e-01 1.08045414e-01 5.47444940e-01 -5.80709994e-01
1.07871473e+00 -2.61036110e+00 9.83791873e-02 3.84576917e-01
3.06248784e-01 4.64776844e-01 1.43931448e-01 2.76592910e-01
-6.07170127e-02 9.18718055e-02 -5.38996495e-02 -3.11974883e-01
-2.61386096e-01 2.74204854e-02 2.30520487e-01 7.67405212e-01
2.34469891e-01 7.19162405e-01 -8.13243210e-01 -8.11449587e-01
3.02472979e-01 2.69995391e-01 -4.18463528e-01 1.26806721e-01
-1.99359618e-02 3.34541529e-01 -4.07224685e-01 8.02703142e-01
5.45030534e-01 -3.80847663e-01 4.49586034e-01 -1.62032649e-01
-1.78217903e-01 -2.19180629e-01 -1.28970575e+00 1.64553535e+00
-2.81510383e-01 4.81013060e-01 2.10609242e-01 -8.87674212e-01
7.60162234e-01 5.14875114e-01 6.69357657e-01 -4.55633819e-01
2.17926830e-01 4.14833754e-01 1.46804675e-01 -6.79115713e-01
1.66082576e-01 5.40265962e-02 1.30558297e-01 3.21153224e-01
-6.08451106e-02 4.52207066e-02 2.28706136e-01 -1.49055049e-01
1.21127486e+00 -1.40011594e-01 6.15109682e-01 -3.81745189e-01
6.28241360e-01 3.14190656e-01 4.46961373e-01 5.12909949e-01
-4.96970475e-01 4.96668875e-01 3.43314201e-01 -7.14361846e-01
-9.25430894e-01 -7.65394807e-01 -3.92696798e-01 3.78405273e-01
-8.61962512e-03 1.05040342e-01 -5.32143772e-01 -6.89826488e-01
-4.66917120e-02 2.45864496e-01 -9.86088932e-01 8.33662823e-02
-3.10098559e-01 -9.47272301e-01 6.04603052e-01 3.52096140e-01
2.89796144e-01 -9.17978346e-01 -7.17022955e-01 3.43281589e-02
1.31340101e-01 -1.00144064e+00 3.56723852e-02 3.84640545e-01
-7.68426299e-01 -1.20646155e+00 -6.79430127e-01 -6.72256887e-01
9.69192803e-01 7.02038333e-02 6.76898122e-01 3.92452478e-01
-8.86686623e-01 -8.29194486e-02 -3.20320070e-01 -6.18236363e-01
-5.13504505e-01 2.79689699e-01 -2.15075493e-01 1.31761283e-01
4.63431239e-01 -2.16519788e-01 -7.63918996e-01 1.67407870e-01
-1.17706716e+00 1.16465218e-01 1.17452908e+00 8.80903304e-01
5.70757270e-01 -1.69155770e-03 5.11416018e-01 -1.08749270e+00
6.54710159e-02 -3.87881458e-01 -4.27234411e-01 3.34985286e-01
-3.50042403e-01 -2.56845772e-01 8.02980125e-01 -2.41637021e-01
-1.10137045e+00 1.37584612e-01 -3.18649232e-01 1.08141467e-01
-4.64084059e-01 6.75748706e-01 -3.78220566e-02 -5.01027703e-02
7.48700321e-01 -4.40350100e-02 4.00079906e-01 -2.69080162e-01
2.24944074e-02 7.70937681e-01 1.54212743e-01 -6.95297942e-02
5.97675920e-01 7.56744981e-01 1.92435086e-01 -7.99604833e-01
-6.78327680e-01 -5.00509918e-01 -7.34706044e-01 -1.84014902e-01
7.44222939e-01 -8.63782048e-01 -3.50973964e-01 4.62262183e-01
-4.91327643e-01 -3.63887012e-01 -4.05148923e-01 5.98136663e-01
5.10187168e-03 2.94824779e-01 -7.17077792e-01 -4.79680091e-01
-3.20955426e-01 -1.25783837e+00 1.05043662e+00 1.88845143e-01
-2.81423658e-01 -1.16155589e+00 6.88758343e-02 4.70782340e-01
5.50920963e-01 5.70631206e-01 9.76688385e-01 -7.14008331e-01
-5.30204654e-01 -6.24337912e-01 -3.31527084e-01 8.15015361e-02
3.83661956e-01 3.71116549e-01 -1.15036333e+00 -3.64625126e-01
-1.56264994e-02 -3.01723897e-01 5.48232377e-01 3.64088982e-01
1.00304806e+00 6.66858479e-02 -7.50278533e-01 6.12760246e-01
1.64341605e+00 1.08939894e-01 5.59519291e-01 5.62506795e-01
2.93942720e-01 6.01324141e-01 6.83356822e-01 3.63533616e-01
1.71481982e-01 6.75257981e-01 5.36360443e-01 -7.67788410e-01
-3.36290330e-01 8.75335112e-02 -1.03694007e-01 7.32518554e-01
-5.31207491e-03 -6.15664348e-02 -9.89875734e-01 7.87935972e-01
-1.39473057e+00 -5.04995763e-01 -7.31914630e-03 2.05133700e+00
6.51144564e-01 2.34430879e-01 -1.29784510e-01 3.34170908e-01
2.96482354e-01 -2.24733472e-01 -4.63875473e-01 -2.51359433e-01
2.02229545e-02 -6.77249953e-02 2.82474607e-01 1.99961200e-01
-9.48348403e-01 4.25448596e-01 5.84259748e+00 8.70675623e-01
-1.51325548e+00 -1.59645885e-01 9.86645222e-01 -2.31820960e-02
-7.94016123e-02 -3.21393847e-01 -7.98116684e-01 4.03737754e-01
6.47206485e-01 1.09606862e-01 -1.89570561e-02 7.84528434e-01
1.41982749e-01 -3.06594998e-01 -9.90273893e-01 4.13830310e-01
9.44492295e-02 -1.38259220e+00 -1.32063538e-01 4.22692776e-01
5.41252911e-01 6.04894618e-03 -2.16296073e-02 -7.74948951e-03
-6.72665685e-02 -1.01261103e+00 4.12672848e-01 4.56086606e-01
9.12505090e-01 -5.14166892e-01 1.34633064e+00 3.46601456e-01
-7.33049691e-01 -2.07239673e-01 -6.11263812e-02 7.20742643e-02
-1.97268710e-01 1.02241611e+00 -1.22544026e+00 8.10565770e-01
6.05463564e-01 4.46452260e-01 -8.22301209e-01 9.39315379e-01
1.34449065e-01 5.51060677e-01 -4.13370132e-01 -2.31011644e-01
1.86571509e-01 8.41215029e-02 1.32031009e-01 1.31389880e+00
3.81618619e-01 -2.75697380e-01 -7.66589940e-02 5.30976832e-01
2.89597213e-01 1.64674118e-01 -4.68687564e-01 4.81811687e-02
3.39403927e-01 1.64310586e+00 -1.04529452e+00 -6.61095157e-02
-5.06747305e-01 4.43762720e-01 3.55765998e-01 6.57004938e-02
-7.06571162e-01 -2.34603003e-01 2.61210769e-01 3.51912886e-01
1.54253855e-01 2.31812507e-01 -5.87396860e-01 -7.54514575e-01
1.04201488e-01 -8.86121929e-01 4.06303316e-01 -3.03326428e-01
-1.06920695e+00 8.35995197e-01 -3.17810833e-01 -1.33746505e+00
-1.87733620e-01 -7.46980250e-01 -3.87193441e-01 6.86922967e-01
-1.64099491e+00 -1.23576200e+00 -6.27672732e-01 2.60601252e-01
-8.78483336e-03 -1.83754563e-01 1.17562199e+00 5.11692464e-01
-7.55652249e-01 5.94347894e-01 2.93337703e-01 1.69671297e-01
7.68269062e-01 -1.18725252e+00 -2.43979245e-01 5.18875718e-01
-2.52431720e-01 5.23306429e-01 3.94190043e-01 -2.31670305e-01
-1.29961312e+00 -1.00975502e+00 8.60158563e-01 -3.23240429e-01
6.91462517e-01 -3.93936157e-01 -6.94875121e-01 6.49474621e-01
-6.39978871e-02 1.28995687e-01 1.08431828e+00 -5.12841195e-02
1.96091026e-01 -4.17181522e-01 -1.36225390e+00 6.17534637e-01
4.38060194e-01 -2.44217128e-01 -1.09819971e-01 2.51549810e-01
1.99804440e-01 -3.42145652e-01 -9.87545609e-01 3.86554956e-01
5.62349558e-01 -9.93075132e-01 7.15863585e-01 7.16351420e-02
4.20406491e-01 -4.54936445e-01 2.14179203e-01 -1.08368552e+00
-3.25014234e-01 4.60076667e-02 3.90806794e-01 1.14930892e+00
4.93844151e-01 -5.83951294e-01 8.37889850e-01 4.12891030e-01
-1.89174697e-01 -1.16006863e+00 -9.58863914e-01 -2.36525431e-01
-1.13521703e-01 -6.81559592e-02 3.87572378e-01 8.99664283e-01
1.74076378e-01 -1.34211496e-01 1.10154562e-01 4.87064980e-02
3.41747314e-01 8.70928094e-02 7.23110199e-01 -1.19187105e+00
-4.11570936e-01 -4.46299255e-01 -6.25065207e-01 -5.24016857e-01
-4.10429984e-01 -8.27171504e-01 -1.24139860e-01 -1.35308146e+00
3.64778697e-01 -6.35392129e-01 -1.31764621e-01 3.72905493e-01
-1.67832822e-01 3.45048755e-01 -2.06463471e-01 3.07065457e-01
-4.47314471e-01 1.69471384e-03 1.16209996e+00 -3.47999409e-02
-6.34970516e-02 -1.32884040e-01 -8.01903903e-01 7.50372171e-01
4.93506432e-01 -1.90278485e-01 -2.39092544e-01 -1.69753954e-01
2.27879301e-01 -7.85200112e-03 1.49345279e-01 -1.18997622e+00
3.45664024e-01 3.53338048e-02 7.08300114e-01 -4.99379635e-01
9.30315480e-02 -1.06655419e+00 5.04050076e-01 6.40331566e-01
-1.35033861e-01 -4.40113544e-01 2.42608353e-01 4.54560429e-01
-4.01269466e-01 -2.09472373e-01 8.37471366e-01 -1.64870352e-01
-4.21034753e-01 -7.59003982e-02 -2.70537764e-01 -3.91955048e-01
1.42687631e+00 -5.14647245e-01 -3.32216084e-01 2.06925407e-01
-6.05367422e-01 -1.05727883e-02 6.13809526e-01 -5.78818917e-02
1.34765223e-01 -8.02192032e-01 -6.84935689e-01 2.10229665e-01
2.57980138e-01 3.41321915e-01 4.15331393e-01 1.14340472e+00
-9.57122922e-01 6.55592918e-01 -1.43663749e-01 -6.77392006e-01
-1.24150014e+00 6.71833694e-01 3.03179890e-01 -1.04563534e+00
-7.34477937e-01 6.84643686e-01 1.28227264e-01 -3.29105258e-01
3.41367245e-01 -3.98304433e-01 -2.91055828e-01 4.94446941e-02
5.88999689e-01 1.10317953e-01 3.83969992e-01 -3.37302744e-01
-4.00066316e-01 3.13317746e-01 -2.79149950e-01 2.27906615e-01
1.54546225e+00 3.99716049e-02 -3.60216126e-02 3.55531693e-01
1.02081037e+00 1.47895679e-01 -1.01342189e+00 -9.90538821e-02
1.96844488e-01 -2.27513343e-01 -4.03383235e-03 -1.08220661e+00
-1.08445275e+00 6.55230165e-01 7.73866832e-01 2.49841094e-01
1.26174521e+00 -2.04994515e-01 4.59287643e-01 -3.16077955e-02
3.65945131e-01 -9.06544089e-01 -1.38280541e-01 -1.13490082e-01
5.03137946e-01 -1.24468434e+00 2.86382049e-01 -4.39681262e-01
-3.47490311e-01 1.17787087e+00 2.70094335e-01 7.83647224e-02
6.96000636e-01 7.15137124e-01 2.78799623e-01 -3.37653756e-01
-7.69518018e-01 2.32482031e-02 9.05144140e-02 3.69267434e-01
5.31905115e-01 8.75474438e-02 -3.30748767e-01 6.66236758e-01
1.55264795e-01 6.20805144e-01 3.30981284e-01 9.14918184e-01
-1.82540253e-01 -7.17598498e-01 -3.32012475e-01 4.68365997e-01
-4.90732163e-01 2.23230988e-01 -3.55610400e-01 1.15562916e+00
3.49095941e-01 6.70655191e-01 -5.22712506e-02 -9.47627649e-02
2.21145377e-01 -1.67654246e-01 1.30709335e-01 -3.95072252e-01
-6.08742833e-01 -1.71688795e-02 -6.90989941e-02 -2.90083885e-01
-3.67442846e-01 -6.66677833e-01 -1.06202698e+00 -1.13645971e-01
-3.96382451e-01 -2.18225662e-02 7.15138018e-01 1.02913713e+00
3.90054911e-01 7.05205083e-01 5.05742431e-01 -5.52231312e-01
-4.84961897e-01 -1.09065795e+00 -8.47192705e-01 4.66479272e-01
1.06507815e-01 -4.10217911e-01 -5.51919281e-01 7.21694948e-03] | [15.05111026763916, -2.988070487976074] |
80359d0d-eda4-43e9-8fcf-8b80821e4478 | image-reconstruction-using-superpixel | 2305.09564 | null | https://arxiv.org/abs/2305.09564v1 | https://arxiv.org/pdf/2305.09564v1.pdf | Image Reconstruction using Superpixel Clustering and Tensor Completion | This paper presents a pixel selection method for compact image representation based on superpixel segmentation and tensor completion. Our method divides the image into several regions that capture important textures or semantics and selects a representative pixel from each region to store. We experiment with different criteria for choosing the representative pixel and find that the centroid pixel performs the best. We also propose two smooth tensor completion algorithms that can effectively reconstruct different types of images from the selected pixels. Our experiments show that our superpixel-based method achieves better results than uniform sampling for various missing ratios. | ['Andrzej Cichocki', 'Zaher Al Aghbari', 'Salman Ahmadi-Asl', 'Anh Huy Phan', 'Maame G. Asante-Mensah'] | 2023-05-16 | null | null | null | null | ['image-reconstruction'] | ['computer-vision'] | [ 2.77798057e-01 -2.59865612e-01 -4.68494385e-01 -2.18068168e-01
-1.02718961e+00 -8.64799768e-02 1.82637185e-01 -1.05144903e-01
-3.23715776e-01 2.64183640e-01 3.64725292e-01 1.55795276e-01
3.22938412e-02 -7.90822029e-01 -5.67900002e-01 -8.26773345e-01
-1.93235371e-02 4.43281680e-01 6.23056531e-01 1.36520863e-01
6.48404300e-01 5.63848853e-01 -1.38603258e+00 7.08075404e-01
7.18143165e-01 1.23963308e+00 5.59932292e-01 6.21234238e-01
-4.22347635e-01 8.11274052e-01 -3.66490155e-01 -4.20183614e-02
4.87724096e-01 -2.13413566e-01 -1.30938387e+00 8.25419426e-01
6.47751749e-01 -4.65837628e-01 -6.21759929e-02 1.21816647e+00
-8.69412869e-02 2.93060154e-01 5.24139285e-01 -7.28078067e-01
-6.42461956e-01 7.78034568e-01 -9.71177101e-01 2.00285032e-01
1.67149067e-01 -3.75551999e-01 9.34514999e-01 -1.13992047e+00
9.11567688e-01 1.32451999e+00 5.32535255e-01 4.72108945e-02
-1.44001281e+00 -1.46665484e-01 7.85851404e-02 1.42407209e-01
-1.67162359e+00 -2.42054090e-01 1.05353796e+00 -1.70074299e-01
3.42074335e-01 6.10206366e-01 4.99288797e-01 4.95119810e-01
1.32051066e-01 1.14137602e+00 1.41880703e+00 -4.24253702e-01
2.95861363e-01 -2.04608023e-01 4.47767079e-01 7.94717789e-01
1.82263851e-01 -4.91303533e-01 -3.46887916e-01 -3.52416337e-01
1.18750751e+00 4.52036649e-01 -2.44393766e-01 -2.12674230e-01
-1.70727515e+00 6.65556610e-01 4.13498104e-01 4.61321205e-01
-7.89969563e-01 4.90510017e-01 2.06781760e-01 -7.81545939e-04
5.06769598e-01 1.56790674e-01 -8.02808404e-02 1.12849191e-01
-1.34033346e+00 1.55858230e-02 4.70758796e-01 9.11331058e-01
1.04861593e+00 -3.10838632e-02 -2.79123604e-01 1.12598252e+00
3.17783356e-02 3.38069320e-01 2.17138723e-01 -1.63061118e+00
2.11092487e-01 3.46075356e-01 2.07517013e-01 -1.36283636e+00
1.51087046e-01 -6.67219311e-02 -6.90869272e-01 -9.82768834e-02
3.68770003e-01 4.44709718e-01 -1.02884257e+00 8.55989277e-01
3.30831051e-01 3.22419181e-02 -2.35206082e-01 1.08928573e+00
5.76618016e-01 8.05290699e-01 2.92744860e-02 -1.01906218e-01
1.41812682e+00 -1.09993041e+00 -5.85465312e-01 2.40054220e-01
1.62883133e-01 -9.90717590e-01 1.05731034e+00 7.10220456e-01
-1.06825757e+00 -5.13089120e-01 -8.23338807e-01 -1.89823583e-01
3.66361514e-02 4.11367238e-01 1.00068319e+00 3.37694436e-01
-1.02496648e+00 7.49290943e-01 -9.67703283e-01 -2.37748519e-01
3.70839953e-01 9.56060216e-02 -2.89715528e-01 -2.05686152e-01
-4.51822728e-01 2.01924577e-01 2.78573483e-01 -8.07804465e-02
-8.94159734e-01 -2.57302791e-01 -3.76763016e-01 -2.33814701e-01
5.55043705e-02 -4.41473246e-01 7.48967290e-01 -1.19374943e+00
-1.05507267e+00 8.67364585e-01 -4.58090335e-01 -4.66504872e-01
2.10015565e-01 1.34752765e-01 -1.92529745e-02 8.11969042e-01
2.39667684e-01 8.53740752e-01 1.25402999e+00 -1.47694528e+00
-6.48942292e-01 -3.01747113e-01 9.44365934e-03 7.73153603e-02
-2.91250587e-01 8.42359811e-02 -9.48703110e-01 -1.09075546e+00
9.51896310e-01 -7.60942459e-01 -6.30634248e-01 8.11740085e-02
-6.77909017e-01 2.23218668e-02 7.20922649e-01 -7.98834562e-01
1.04201853e+00 -2.16275811e+00 8.80426690e-02 3.85170758e-01
5.34690797e-01 -5.50341189e-01 -2.35921126e-02 9.64658931e-02
8.20808783e-02 1.16280215e-02 -8.73619914e-02 -3.73730898e-01
-4.35834467e-01 3.07947636e-01 -3.00061285e-01 3.86207581e-01
-3.37237537e-01 4.24206436e-01 -7.11899102e-01 -1.02584541e+00
2.21721634e-01 5.30014932e-01 -3.99508268e-01 -2.57800043e-01
-1.54011279e-01 3.09791893e-01 -8.62393856e-01 8.58102262e-01
1.04044592e+00 -3.10448557e-01 4.97277081e-03 -7.10059285e-01
-9.76506397e-02 -1.49090188e-02 -1.46503139e+00 1.55698824e+00
2.58511095e-03 6.39118314e-01 2.55352020e-01 -9.43148255e-01
1.11263585e+00 -9.53175053e-02 8.73471320e-01 -2.91664451e-01
3.52409817e-02 2.30902582e-01 -6.64476216e-01 -4.47203517e-01
1.08419549e+00 1.72343496e-02 2.73481131e-01 6.12258375e-01
-4.61818665e-01 -1.21450081e-01 2.28949517e-01 2.91279048e-01
9.02289569e-01 1.05316110e-01 -2.86906868e-01 -6.08651161e-01
3.37123543e-01 4.31864232e-01 6.96908653e-01 6.00191057e-01
-2.35532165e-01 1.03320813e+00 3.17551792e-01 -6.00979745e-01
-1.22696960e+00 -8.59183192e-01 -2.72973269e-01 1.01233339e+00
4.77965921e-01 -2.32342154e-01 -8.78389120e-01 -3.33091289e-01
-1.24991551e-01 2.45119840e-01 -6.60208583e-01 5.71983635e-01
-4.48847234e-01 -7.63921618e-01 1.34526640e-01 5.43705940e-01
9.22916532e-01 -5.38540483e-01 -4.05591249e-01 -8.42727795e-02
-7.00182736e-01 -1.00732517e+00 -8.15388680e-01 -1.43458605e-01
-1.34303784e+00 -8.94026995e-01 -1.01508927e+00 -1.10868573e+00
1.24433947e+00 1.14817262e+00 7.87476063e-01 9.12621841e-02
-6.25485554e-02 4.43375796e-01 -5.85922301e-01 5.43874502e-01
-7.74201453e-02 -2.73771614e-01 -3.14763963e-01 5.46352267e-01
1.04345217e-01 -2.28674024e-01 -7.84245372e-01 4.06361401e-01
-1.16894174e+00 2.44484931e-01 4.05019462e-01 6.35351062e-01
1.33374977e+00 4.49353755e-01 -4.91015851e-01 -1.12889934e+00
5.66177130e-01 -1.71437442e-01 -3.97163719e-01 3.05111289e-01
-2.50826955e-01 3.62916619e-01 3.02220464e-01 -4.17319775e-01
-9.91568983e-01 2.80089319e-01 4.98442382e-01 -7.02189684e-01
1.09464541e-01 4.63075697e-01 1.84520200e-01 -2.86538869e-01
4.90802169e-01 3.69239986e-01 -7.98328295e-02 -8.44878912e-01
4.57105577e-01 5.14649332e-01 4.12041694e-01 -7.83638418e-01
3.92483979e-01 1.08131599e+00 -1.64511368e-01 -1.12477195e+00
-5.18375874e-01 -5.88964224e-01 -8.00819457e-01 -1.60081729e-01
8.85470092e-01 -7.77243495e-01 -3.06953967e-01 3.62879038e-01
-1.12304449e+00 -1.85371995e-01 -2.67247558e-01 5.52111268e-01
-3.42004478e-01 8.15388560e-01 -1.05738521e+00 -5.59160888e-01
-1.99560151e-01 -1.42454684e+00 1.25900602e+00 -1.62441477e-01
2.69866865e-02 -7.96745241e-01 -1.76529005e-01 6.58358037e-01
1.71863943e-01 2.19038486e-01 5.79920828e-01 1.52812317e-01
-1.22403574e+00 -1.40660122e-01 -5.56382716e-01 2.65868813e-01
-1.54902544e-02 2.77808398e-01 -7.69524276e-01 -1.30414799e-01
-2.65795272e-03 6.52119070e-02 1.31402934e+00 7.12722301e-01
1.47689795e+00 -4.31959778e-01 -4.43663418e-01 7.23230064e-01
1.69711459e+00 -6.21135943e-02 9.39414740e-01 3.67299914e-01
9.01725709e-01 4.51138824e-01 7.22309649e-01 2.20664710e-01
3.44892263e-01 4.72017050e-01 1.88935801e-01 -1.85136810e-01
-3.27546000e-01 -1.46726295e-02 2.95942128e-01 1.00881505e+00
-3.91610742e-01 3.40943009e-01 -5.31228423e-01 7.02182591e-01
-1.74275291e+00 -9.66420174e-01 -3.84149045e-01 2.23334765e+00
7.99033284e-01 -2.84066349e-02 2.40129530e-01 1.89044282e-01
8.60189021e-01 3.87426645e-01 -9.66581926e-02 -2.09739655e-01
-3.49147648e-01 9.08613801e-02 1.06804729e+00 5.88810921e-01
-1.20924115e+00 1.04582024e+00 7.65750933e+00 1.20236349e+00
-9.76628006e-01 1.68700129e-01 9.83165026e-01 3.89068455e-01
-6.72375321e-01 3.41549367e-01 -6.50412321e-01 1.13949664e-01
3.17257792e-01 1.70069069e-01 4.20033187e-01 8.10063004e-01
3.49121094e-01 -5.99595487e-01 -4.25902754e-01 1.03793776e+00
6.77455515e-02 -1.43109226e+00 4.24619257e-01 1.00998525e-02
1.14278579e+00 -1.36996597e-01 1.07860252e-01 -7.23235965e-01
2.99732476e-01 -6.79578543e-01 9.23389673e-01 8.84084523e-01
5.07440150e-01 -4.78667796e-01 2.25654140e-01 -1.77456304e-01
-1.31735480e+00 3.42871279e-01 -8.17074478e-01 2.20873207e-01
-3.12407047e-01 9.04607177e-01 -6.57591045e-01 2.52380699e-01
7.46920109e-01 7.88657069e-01 -6.49435461e-01 1.03012252e+00
1.19210899e-01 6.27136946e-01 -4.46242899e-01 1.01942224e-02
3.31859469e-01 -7.14047432e-01 5.59973657e-01 1.10329211e+00
1.94155008e-01 -5.12053296e-02 5.73208451e-01 9.64988828e-01
2.30941206e-01 3.19763958e-01 -4.41827834e-01 2.27420442e-02
4.89642769e-01 1.25027013e+00 -1.56492662e+00 -5.71231246e-01
-3.06795955e-01 1.29650462e+00 -4.15714718e-02 6.57664716e-01
-4.36193198e-01 -3.25058609e-01 3.02194178e-01 2.54272103e-01
5.87385416e-01 -6.95480645e-01 -6.37613535e-01 -1.15556777e+00
4.54440787e-02 -5.70678890e-01 4.38559353e-01 -9.03392971e-01
-8.22156131e-01 5.99396527e-01 7.12256134e-02 -1.40129304e+00
3.65477651e-01 -5.19091368e-01 -2.96530336e-01 5.59543610e-01
-1.10312355e+00 -1.17350233e+00 -3.50249708e-01 7.17948020e-01
6.72662497e-01 1.98738277e-01 5.22556365e-01 2.79004931e-01
-5.42763948e-01 1.26067981e-01 4.35985148e-01 2.34705836e-01
1.57555044e-01 -1.11969662e+00 4.26608622e-02 9.70803082e-01
1.33053213e-01 6.99554563e-01 6.36802733e-01 -6.48671687e-01
-1.52666259e+00 -9.42327261e-01 5.80040455e-01 -6.04648851e-02
4.17110622e-01 1.17590860e-01 -6.31487012e-01 6.19295597e-01
4.47161756e-02 -9.56702009e-02 2.52332509e-01 -2.15646654e-01
-3.73574495e-01 -3.23978543e-01 -1.13928616e+00 7.66138256e-01
5.83166957e-01 -5.00358522e-01 -2.40427807e-01 5.11563480e-01
5.66842020e-01 -9.34867039e-02 -1.06657434e+00 -4.67291884e-02
3.67598891e-01 -8.72302949e-01 1.21893489e+00 -7.71807656e-02
4.84060585e-01 -5.88079333e-01 -5.08826315e-01 -8.25462222e-01
-4.35833156e-01 -5.06843984e-01 4.85446811e-01 9.40876067e-01
2.82887340e-01 -3.70068878e-01 1.03113735e+00 6.98704898e-01
2.27741227e-01 -5.07969081e-01 -6.05736613e-01 -6.27280712e-01
-4.56194252e-01 -6.11306429e-01 5.54386675e-01 7.55701303e-01
-2.21791327e-01 -3.00856363e-02 -3.03183019e-01 -1.71585247e-01
1.11185598e+00 8.28896701e-01 6.14443660e-01 -8.60672116e-01
-8.25343207e-02 -4.47388083e-01 -2.86241770e-01 -1.42590785e+00
-3.58617157e-01 -8.48868549e-01 -4.75641303e-02 -1.59643555e+00
5.97504914e-01 -7.58345425e-01 -2.05455050e-01 4.00930464e-01
3.08127031e-02 6.90710604e-01 -5.75653911e-02 8.33248794e-01
-8.06664348e-01 1.94177091e-01 1.72692728e+00 -4.75314498e-01
-1.36357561e-01 -2.02534497e-01 -6.00386500e-01 5.76875269e-01
4.92413431e-01 -1.42056525e-01 -2.31804505e-01 -6.20558321e-01
-5.72363399e-02 1.06144801e-01 2.78842747e-01 -7.55625308e-01
-4.63638408e-03 -4.97334301e-01 4.00888860e-01 -8.31154287e-01
4.19201851e-01 -7.30822444e-01 3.22287560e-01 2.75184691e-01
-4.15961176e-01 -1.09335244e-01 -2.56872237e-01 5.42127728e-01
-1.96597040e-01 -2.92115837e-01 8.32089067e-01 -4.19154704e-01
-1.06991518e+00 2.95290232e-01 -3.36102962e-01 -3.29288721e-01
7.09822357e-01 -5.91201007e-01 3.15929055e-02 -3.61981839e-01
-9.06161606e-01 -1.01890013e-01 7.62982428e-01 2.10871696e-02
1.06453240e+00 -1.52280879e+00 -6.92995012e-01 1.93804428e-01
-5.81504703e-02 -3.18569928e-01 -4.99718115e-02 9.67825651e-01
-1.24124146e+00 2.47464001e-01 -1.81642264e-01 -7.49312580e-01
-1.24069786e+00 5.17165244e-01 7.51511306e-02 7.70019293e-02
-9.43305135e-01 9.04977083e-01 1.59536451e-01 1.76164404e-01
1.94369704e-01 -7.74016500e-01 -1.21698439e-01 -1.05232485e-01
5.07129788e-01 7.86472857e-01 -1.66307926e-01 -9.79876876e-01
-8.04109052e-02 7.75856256e-01 -8.32422730e-03 -2.71919787e-01
1.27259815e+00 -3.14551979e-01 -6.21550798e-01 5.13045967e-01
1.38355625e+00 -5.15986457e-02 -1.17377508e+00 -3.53723288e-01
-1.38591275e-01 -1.14364088e+00 6.03243470e-01 2.34354120e-02
-1.63096106e+00 4.99724805e-01 4.36103344e-01 3.16011757e-01
1.25016916e+00 -2.20034588e-02 1.14298034e+00 5.76181076e-02
7.62751043e-01 -1.34327614e+00 4.77822050e-02 1.71757311e-01
8.19730997e-01 -9.73154068e-01 4.00969058e-01 -8.48499894e-01
-7.38405228e-01 1.21370101e+00 4.23989706e-02 -6.19970262e-01
7.93000221e-01 -1.35939598e-01 -8.68980587e-02 -4.71049279e-01
-3.33532810e-01 -2.97971666e-01 3.19045871e-01 4.34595466e-01
3.56958419e-01 4.72594261e-01 -6.33288503e-01 -6.65385947e-02
4.15797383e-02 -2.72120327e-01 4.49165910e-01 8.56635809e-01
-7.17178702e-01 -9.60506678e-01 -9.85966802e-01 5.31422198e-01
-4.24436629e-01 2.13892832e-02 -1.75485179e-01 2.63641208e-01
-1.82472855e-01 6.86513484e-01 2.11075813e-01 -4.42631632e-01
-1.44734845e-01 -2.36821204e-01 6.72352791e-01 -2.37038225e-01
-9.38183516e-02 7.58982956e-01 -6.47908822e-02 -7.55585849e-01
-8.15821767e-01 -8.69189858e-01 -1.27094817e+00 -4.63504851e-01
-3.21757495e-01 1.57549202e-01 4.51127946e-01 4.65050310e-01
1.51489586e-01 5.74822389e-02 8.94992352e-01 -9.66001809e-01
-9.11319256e-02 -7.59481788e-01 -1.11596048e+00 5.38638532e-01
1.27289191e-01 -4.86587882e-01 -6.71995431e-02 5.12991011e-01] | [10.607281684875488, -1.5488932132720947] |
e95ca604-5678-4ad0-91bb-cc1e8e7d37b5 | an-unsupervised-multiple-task-and-multiple-1 | null | null | https://aclanthology.org/2022.acl-long.14 | https://aclanthology.org/2022.acl-long.14.pdf | An Unsupervised Multiple-Task and Multiple-Teacher Model for Cross-lingual Named Entity Recognition | Cross-lingual named entity recognition task is one of the critical problems for evaluating the potential transfer learning techniques on low resource languages. Knowledge distillation using pre-trained multilingual language models between source and target languages have shown their superiority in transfer. However, existing cross-lingual distillation models merely consider the potential transferability between two identical single tasks across both domains. Other possible auxiliary tasks to improve the learning performance have not been fully investigated. In this study, based on the knowledge distillation framework and multi-task learning, we introduce the similarity metric model as an auxiliary task to improve the cross-lingual NER performance on the target domain. Specifically, an entity recognizer and a similarity evaluator are first trained in parallel as two teachers from the source domain. Then, two tasks in the student model are supervised by these teachers simultaneously. Empirical studies on the three datasets across 7 different languages confirm the effectiveness of the proposed model. | ['Richong Zhang', 'Wenyi Qin', 'Junfan Chen', 'Xiaohui Guo', 'Chunming Hu', 'Zhuoran Li'] | null | null | null | null | acl-2022-5 | ['cross-lingual-ner'] | ['natural-language-processing'] | [-1.13171004e-01 -7.83105493e-02 -3.08935434e-01 -4.61390346e-01
-1.00074041e+00 -7.50527680e-01 6.41531408e-01 1.07910387e-01
-9.63764489e-01 8.70959461e-01 5.78079410e-02 -4.46725935e-01
1.22643746e-01 -6.20831668e-01 -6.22967780e-01 -4.85716611e-01
4.41217422e-01 6.40640378e-01 2.90909618e-01 -2.53880829e-01
-1.54206961e-01 2.31509253e-01 -1.19515371e+00 2.72433132e-01
1.55160236e+00 4.27213907e-01 3.93795878e-01 1.25898346e-01
-4.77325916e-01 4.87558693e-01 -5.35024643e-01 -8.59886348e-01
8.05350170e-02 -3.86806130e-01 -8.24957192e-01 -5.83569646e-01
5.72946250e-01 -7.41521344e-02 1.32829651e-01 1.11548793e+00
8.99816751e-01 2.25787327e-01 6.49158061e-01 -1.02540398e+00
-1.25792611e+00 1.04561055e+00 -5.04019558e-01 9.97759774e-02
1.57447934e-01 -2.96015948e-01 8.92536163e-01 -1.08256972e+00
4.92094129e-01 1.20948827e+00 6.21439040e-01 6.09911263e-01
-1.01591766e+00 -1.09386480e+00 -8.08856264e-02 3.36697012e-01
-1.51610148e+00 -3.17038804e-01 6.47275031e-01 -2.83968985e-01
9.53589380e-01 -3.90025854e-01 -3.94451581e-02 9.45063055e-01
-2.43232414e-01 8.90538275e-01 1.49270976e+00 -7.09416568e-01
-3.38380247e-01 8.65715504e-01 2.78497692e-02 4.82694954e-01
3.22251320e-01 1.27156571e-01 -4.85506892e-01 2.53820419e-01
2.74611950e-01 -3.55196506e-01 -2.31468961e-01 -3.13410789e-01
-1.30637825e+00 6.93436682e-01 1.92623839e-01 7.33918965e-01
-8.34271386e-02 -5.10755002e-01 5.68093479e-01 5.84355295e-01
6.05650306e-01 5.11667967e-01 -9.74595845e-01 -8.28199312e-02
-7.16340840e-01 -2.62843847e-01 7.02087104e-01 1.15392208e+00
8.59353006e-01 1.04148060e-01 -2.48626843e-01 1.14914048e+00
2.48854697e-01 7.93495893e-01 8.69960308e-01 -1.40337601e-01
7.25346386e-01 7.21636236e-01 -3.00034076e-01 -5.18648565e-01
3.99808325e-02 -4.38535929e-01 -5.16301572e-01 -1.31069869e-03
2.61622518e-01 -5.54400206e-01 -5.45813143e-01 1.90293086e+00
5.63062668e-01 4.82542366e-01 6.01366460e-01 4.27493811e-01
1.07278836e+00 5.15689075e-01 6.70572579e-01 -1.19301237e-01
1.49359751e+00 -1.15832269e+00 -7.70079851e-01 -8.17777514e-02
1.27712357e+00 -1.11870098e+00 1.03236592e+00 8.38285014e-02
-7.99933195e-01 -9.00731385e-01 -8.60729933e-01 -3.36844414e-01
-9.87804711e-01 5.55306256e-01 3.04491788e-01 6.72827780e-01
-7.22258091e-01 2.05587640e-01 -5.31014740e-01 -3.78674507e-01
-7.72254542e-02 1.13893606e-01 -5.79366863e-01 -2.77089715e-01
-1.58484602e+00 1.40777433e+00 9.50906575e-01 -3.01695853e-01
-7.82881677e-01 -8.88748825e-01 -8.81276608e-01 1.17768444e-01
1.02060534e-01 -3.39867443e-01 1.28345191e+00 -8.55137050e-01
-1.58074510e+00 1.19369841e+00 2.76288658e-01 -7.74739608e-02
3.33817244e-01 -3.84831220e-01 -6.84470415e-01 -3.60852391e-01
4.45788890e-01 4.25834268e-01 1.00897968e-01 -1.00645137e+00
-1.02515030e+00 -2.38819778e-01 -1.13555290e-01 7.87934303e-01
-6.30014002e-01 1.49337634e-01 -2.21047059e-01 -7.47007847e-01
-4.76003140e-01 -7.51807749e-01 1.69376254e-01 -6.82528555e-01
-1.24767115e-02 -7.76966274e-01 6.28812432e-01 -7.85519898e-01
1.28987801e+00 -2.03548431e+00 -1.10174520e-02 -2.07244501e-01
-3.84301752e-01 6.98782980e-01 -3.62319142e-01 2.10250452e-01
-1.50749758e-01 1.48445904e-01 4.79858778e-02 -2.02293441e-01
-1.79821029e-01 1.04127415e-01 6.35412261e-02 1.04150027e-01
1.34078652e-01 8.05670440e-01 -1.08858621e+00 -6.87945545e-01
8.76759831e-03 4.17605132e-01 -2.78732747e-01 6.50309205e-01
2.81133741e-01 6.32184386e-01 -3.99448872e-01 2.56305963e-01
5.43435991e-01 1.69832423e-01 2.70439297e-01 -2.59229749e-01
-2.43688270e-01 5.62847495e-01 -1.16223633e+00 2.08297706e+00
-9.92472172e-01 2.11453736e-01 -1.32220104e-01 -1.08232403e+00
1.12753189e+00 7.95127571e-01 2.72437006e-01 -8.07408154e-01
-4.42508375e-03 6.46607816e-01 2.77944684e-01 -5.16442001e-01
4.53228444e-01 -5.61987638e-01 -2.10602656e-01 6.69993341e-01
6.29782677e-01 -1.14616394e-01 1.04667284e-01 -1.48321465e-01
6.26622379e-01 4.03382063e-01 5.44311762e-01 -2.63553500e-01
7.92735159e-01 8.42219219e-02 7.45928228e-01 3.68361205e-01
-4.10982013e-01 -5.14688008e-02 -3.59632969e-01 1.09174140e-02
-7.36838818e-01 -1.13314998e+00 -1.93004623e-01 1.74990594e+00
6.88142627e-02 -1.75679520e-01 -5.63570023e-01 -1.11206472e+00
3.68357189e-02 1.02354491e+00 -3.19462448e-01 -1.46161929e-01
-6.34643912e-01 -4.38296527e-01 1.03140986e+00 4.66087580e-01
7.12942123e-01 -1.00340354e+00 -6.77962676e-02 2.62143582e-01
-1.99986503e-01 -1.34843254e+00 -4.87278461e-01 1.89982533e-01
-6.70315444e-01 -7.52529681e-01 -7.72206008e-01 -1.32251692e+00
5.31392217e-01 2.99466878e-01 1.15714550e+00 -3.12311918e-01
2.22826123e-01 2.14420959e-01 -3.97081524e-01 -4.52231973e-01
-5.79254329e-01 4.08146113e-01 3.51962775e-01 -2.02796817e-01
1.02874494e+00 -4.59612519e-01 1.14996493e-01 3.17107409e-01
-6.19144917e-01 7.17514083e-02 9.59474266e-01 7.87476838e-01
5.52921116e-01 -1.43591046e-01 8.57208252e-01 -1.09754264e+00
7.68261909e-01 -6.17609739e-01 -3.72894287e-01 9.06399071e-01
-6.07795894e-01 4.10354346e-01 6.33504510e-01 -7.59946048e-01
-1.75655520e+00 -9.05159488e-02 2.74077859e-02 -1.95054337e-01
-4.08026695e-01 7.53119171e-01 -4.19474453e-01 -7.58624375e-02
5.84132373e-01 2.45892569e-01 -5.24835587e-01 -7.78112471e-01
6.28135204e-01 8.33498001e-01 4.46506023e-01 -1.10902214e+00
8.44947755e-01 -2.84431726e-01 -5.42870700e-01 -5.39375603e-01
-1.04111910e+00 -4.83106613e-01 -1.13781500e+00 1.78004906e-01
9.66753721e-01 -1.33176136e+00 -4.24309403e-01 3.71328741e-01
-1.37505674e+00 -2.42438003e-01 -1.38735890e-01 1.13430440e+00
-4.97766845e-02 -3.58960070e-02 -3.81865472e-01 -3.21839243e-01
-3.39306116e-01 -1.10180247e+00 6.73392117e-01 5.04212379e-01
2.23241255e-01 -1.45262480e+00 6.16841793e-01 4.22685713e-01
5.68761408e-01 -3.60094637e-01 1.05991590e+00 -1.58221352e+00
5.30675501e-02 1.16089411e-01 -1.82699561e-01 6.21834040e-01
5.06658494e-01 -2.44419679e-01 -1.00029552e+00 -2.48105243e-01
-2.78007854e-02 -6.75664723e-01 2.49700069e-01 -3.01079690e-01
5.63934088e-01 -9.88566279e-02 -3.27448338e-01 4.74126190e-01
1.36201036e+00 1.72016338e-01 -3.58578865e-03 4.28115278e-01
8.80013645e-01 7.54931748e-01 6.52157068e-01 -2.49883577e-01
8.10280144e-01 5.14377534e-01 -4.04718608e-01 -1.08558394e-01
-5.15853107e-01 -4.14819151e-01 6.82795525e-01 1.84025943e+00
-6.70079514e-02 6.08412027e-02 -1.23139107e+00 8.09132040e-01
-1.34952879e+00 -6.63096547e-01 8.03076252e-02 2.17210221e+00
1.50582504e+00 -3.12254697e-01 -4.20070797e-01 -7.01993465e-01
1.04626143e+00 -1.63530424e-01 -4.06282663e-01 -4.19205278e-01
-1.33007124e-01 4.50523525e-01 3.75402063e-01 4.56007510e-01
-9.88032818e-01 1.45273268e+00 5.23418140e+00 1.01184309e+00
-1.26053238e+00 5.68516612e-01 1.57920510e-01 3.43620986e-01
-2.20835984e-01 1.00667797e-01 -1.42260826e+00 1.75097570e-01
1.13616943e+00 -6.43719614e-01 -6.38446435e-02 1.02618933e+00
-2.16487586e-01 2.60808498e-01 -1.14313209e+00 7.46439934e-01
2.73905128e-01 -5.21892369e-01 3.21483880e-01 -1.77075908e-01
1.01823771e+00 1.77225366e-01 -1.35272428e-01 1.00717533e+00
7.33443499e-01 -7.37143815e-01 1.47853136e-01 1.59159899e-01
8.78758907e-01 -7.96435058e-01 8.01460743e-01 4.29003567e-01
-1.24386752e+00 5.02471030e-01 -2.82854766e-01 2.08331823e-01
1.52786359e-01 2.28065133e-01 -1.22497451e+00 8.64016771e-01
5.45494616e-01 5.57063520e-01 -7.86341488e-01 8.29659045e-01
-8.12844753e-01 6.90961778e-01 -1.82464242e-01 1.29413292e-01
2.98705343e-02 -2.69573361e-01 1.44248039e-01 1.42395961e+00
4.36126322e-01 -1.04470693e-01 3.55800897e-01 4.50633734e-01
-3.77480149e-01 7.64290571e-01 -1.00709140e+00 -5.51799722e-02
7.28834689e-01 1.32240331e+00 -1.25335291e-01 -5.12077987e-01
-1.01673901e+00 9.89971101e-01 8.08400154e-01 3.12688887e-01
-6.68814003e-01 -8.17700684e-01 6.42058313e-01 -3.06214422e-01
7.92922080e-02 -2.33752683e-01 -1.32685319e-01 -1.36061537e+00
-1.84557363e-01 -9.50501144e-01 5.64053595e-01 -4.51592207e-01
-1.65746057e+00 5.73248625e-01 8.57632235e-02 -1.11720943e+00
-1.89567089e-01 -6.18590772e-01 -7.52624929e-01 1.19410038e+00
-1.99733102e+00 -1.54008925e+00 -6.53188825e-02 9.73171949e-01
3.50424916e-01 -5.83781421e-01 1.07971442e+00 8.06680858e-01
-5.95690727e-01 1.04009891e+00 3.11174601e-01 5.49283922e-01
1.54470026e+00 -1.25410140e+00 9.69733372e-02 1.00730991e+00
1.81176141e-01 7.46348560e-01 2.41029724e-01 -7.26659775e-01
-9.53161061e-01 -1.12004006e+00 1.45898497e+00 -4.20995623e-01
8.70875835e-01 -1.97016850e-01 -1.29339302e+00 9.90938425e-01
6.93751991e-01 -1.17604345e-01 1.27219915e+00 3.71140718e-01
-4.72227365e-01 -4.49315496e-02 -9.47874367e-01 4.46199387e-01
7.03331411e-01 -7.83021212e-01 -1.28627121e+00 1.80582181e-01
8.61740887e-01 -1.82823300e-01 -1.31236041e+00 4.34578776e-01
1.31825849e-01 -3.04394752e-01 7.56692469e-01 -9.08269227e-01
1.06870711e-01 -2.11631000e-01 1.82166863e-02 -1.74786580e+00
-9.73149315e-02 -1.28808334e-01 3.29034001e-01 1.95008373e+00
6.14880502e-01 -6.60839140e-01 1.58795998e-01 2.84802735e-01
-2.96377987e-01 -2.63847321e-01 -8.13727140e-01 -1.07148111e+00
7.42457390e-01 -8.97186175e-02 4.27794099e-01 1.74206889e+00
-3.65614258e-02 9.29859579e-01 -5.05688787e-02 2.83589959e-01
4.50014740e-01 6.94077313e-02 7.49602497e-01 -1.34508455e+00
-7.27898255e-02 -2.55784422e-01 7.42913783e-02 -8.24625015e-01
6.73150122e-01 -1.47176969e+00 1.60587549e-01 -1.21080649e+00
1.32997334e-01 -9.36935723e-01 -5.74846268e-01 7.97096074e-01
-6.03355348e-01 -2.44675636e-01 -4.77725230e-02 5.46575934e-02
-5.18315494e-01 7.93790758e-01 1.25756359e+00 5.44141345e-02
-1.55354813e-01 -1.69259742e-01 -6.30378067e-01 7.19301760e-01
6.89507306e-01 -8.31123054e-01 -4.69631672e-01 -9.56903756e-01
-1.44211650e-01 -2.73907870e-01 -3.84679645e-01 -9.04311121e-01
5.00489175e-01 -3.14747900e-01 1.05593326e-02 -5.90047985e-02
-4.22007740e-02 -8.36866856e-01 -3.69331956e-01 2.43193746e-01
-3.52355719e-01 3.36395502e-01 4.05322939e-01 -1.34390248e-02
-5.56758106e-01 -4.23508078e-01 8.34833145e-01 -6.87034205e-02
-9.22542751e-01 2.41624072e-01 9.74944532e-02 5.07170141e-01
1.18966293e+00 1.88129112e-01 -3.35978478e-01 2.56757736e-01
-5.49104095e-01 3.70293856e-01 2.15856377e-02 7.71048069e-01
1.71356425e-01 -1.68422520e+00 -1.11506450e+00 5.11296391e-02
3.79848510e-01 -1.02386087e-01 1.10540725e-01 7.43381083e-01
-1.82913497e-01 5.05882084e-01 -5.70845783e-01 -2.80110508e-01
-1.23615336e+00 6.09945357e-01 2.34200910e-01 -7.63062656e-01
8.89273509e-02 7.63075531e-01 3.37013513e-01 -1.41040742e+00
1.32190108e-01 3.49476747e-02 -3.83371919e-01 2.47297570e-01
3.29363585e-01 3.09628487e-01 5.65107763e-02 -7.93795407e-01
-2.47243643e-01 6.98263466e-01 -3.16281229e-01 -2.14118019e-01
1.18639338e+00 -2.71199465e-01 8.03704858e-02 5.81671476e-01
1.24513936e+00 2.72830367e-01 -4.97766405e-01 -8.45517814e-01
3.89867216e-01 3.77826206e-02 -1.62726298e-01 -1.04684532e+00
-8.12348843e-01 1.22676504e+00 5.37800372e-01 -3.66458088e-01
9.30103779e-01 -1.48856625e-01 6.35999620e-01 4.89211887e-01
4.11698192e-01 -1.18099248e+00 -1.10834360e-01 9.55756664e-01
4.24124509e-01 -1.58976495e+00 -2.90106297e-01 -2.65785068e-01
-7.33106673e-01 9.39952672e-01 1.26275742e+00 2.59891152e-01
6.86164558e-01 6.23126216e-02 4.59312290e-01 1.23477250e-01
-5.12314439e-01 -3.88662219e-01 5.34137130e-01 5.81761897e-01
1.20532584e+00 1.32556602e-01 -5.89799762e-01 7.42053032e-01
-2.02546611e-01 -2.41586670e-01 1.34674475e-01 8.40897381e-01
-3.16516876e-01 -1.55310380e+00 -1.28671125e-01 -2.67990232e-01
-6.13195896e-01 -4.39639896e-01 -1.99757695e-01 8.33058655e-01
4.49843973e-01 7.47879982e-01 -2.12158263e-01 -1.97251514e-01
5.81651390e-01 5.44665456e-01 3.12406808e-01 -1.10514820e+00
-1.08101082e+00 -4.60016847e-01 -2.01648604e-02 -1.42742649e-01
-5.18318713e-01 -2.84961432e-01 -1.08028972e+00 5.48767149e-02
-5.03151715e-01 6.46835625e-01 8.22643936e-01 1.14917862e+00
3.40474606e-01 4.41757917e-01 4.21337843e-01 -6.47238418e-02
-6.31097019e-01 -1.31826961e+00 -4.22934532e-01 5.28331757e-01
-1.88062951e-01 -7.33386934e-01 -1.74494505e-01 1.28989697e-01] | [9.980419158935547, 9.676006317138672] |
939d29e0-2450-4e5e-b2a6-a348dfed97c2 | aggregation-of-local-parametric-candidates | 1407.5759 | null | http://arxiv.org/abs/1407.5759v1 | http://arxiv.org/pdf/1407.5759v1.pdf | Aggregation of local parametric candidates with exemplar-based occlusion handling for optical flow | Handling all together large displacements, motion details and occlusions
remains an open issue for reliable computation of optical flow in a video
sequence. We propose a two-step aggregation paradigm to address this problem.
The idea is to supply local motion candidates at every pixel in a first step,
and then to combine them to determine the global optical flow field in a second
step. We exploit local parametric estimations combined with patch
correspondences and we experimentally demonstrate that they are sufficient to
produce highly accurate motion candidates. The aggregation step is designed as
the discrete optimization of a global regularized energy. The occlusion map is
estimated jointly with the flow field throughout the two steps. We propose a
generic exemplar-based approach for occlusion filling with motion vectors. We
achieve state-of-the-art results in computer vision benchmarks, with
particularly significant improvements in the case of large displacements and
occlusions. | ['Patrick Bouthemy', 'Denis Fortun', 'Charles Kervrann'] | 2014-07-22 | null | null | null | null | ['occlusion-handling'] | ['computer-vision'] | [ 1.87673029e-02 -4.71473157e-01 -1.17512189e-01 -1.13417907e-03
-7.35883355e-01 -5.20182550e-01 4.73679155e-01 1.46830574e-01
-5.41376472e-01 7.51027226e-01 2.33775869e-01 7.20319748e-02
4.63600382e-02 -4.75291580e-01 -5.78031361e-01 -7.55118787e-01
4.96835820e-03 2.55510956e-01 6.97932959e-01 6.60755439e-03
3.90868306e-01 7.07284749e-01 -1.45621538e+00 1.16008267e-01
1.01433575e+00 9.01650250e-01 1.88145563e-01 7.69406796e-01
-1.73423849e-02 8.87642145e-01 -2.46271685e-01 -9.62290168e-02
5.25433064e-01 -4.25976545e-01 -9.57751930e-01 6.25849128e-01
9.87006545e-01 -5.73136866e-01 -2.92537034e-01 7.95964539e-01
4.15847033e-01 4.82658774e-01 3.49401414e-01 -9.52796578e-01
5.86993136e-02 -1.34005100e-01 -7.07848668e-01 2.77292371e-01
3.51136982e-01 3.59040052e-01 1.06527495e+00 -1.06372237e+00
1.15191841e+00 1.18475723e+00 4.71527845e-01 2.79876918e-01
-1.48531663e+00 6.71089292e-02 1.63815290e-01 2.64998376e-01
-1.22840166e+00 -5.25262833e-01 8.31944406e-01 -7.06515193e-01
7.55671024e-01 7.54825324e-02 7.93865323e-01 2.92532951e-01
9.44558308e-02 7.21138000e-01 6.52326822e-01 -3.71697307e-01
2.78879106e-01 -2.20584810e-01 1.86469063e-01 8.48933756e-01
1.76294461e-01 2.53089182e-02 -4.74894166e-01 -9.01905149e-02
1.02757478e+00 -1.93050563e-01 -4.87584352e-01 -6.29287302e-01
-1.30942583e+00 7.95474946e-01 4.64745104e-01 1.85213774e-01
-5.75965881e-01 2.61298567e-01 2.01553807e-01 1.86441779e-01
5.12611449e-01 2.95917153e-01 -2.09174231e-01 6.13891659e-03
-1.11727321e+00 5.53776503e-01 8.04423273e-01 6.83625460e-01
1.27374637e+00 -7.07662031e-02 -3.05385709e-01 6.41506433e-01
1.92357048e-01 1.64886907e-01 5.34395222e-03 -1.58092487e+00
3.98797750e-01 2.95613378e-01 6.02969289e-01 -1.17268360e+00
-2.26249382e-01 -3.01380545e-01 -5.24610937e-01 3.97157520e-01
7.62786806e-01 7.24307224e-02 -7.48482347e-01 1.53562284e+00
8.19382429e-01 5.50073445e-01 -2.72153080e-01 9.73897219e-01
3.81447971e-01 5.94209492e-01 -2.63355728e-02 -5.11119187e-01
1.16927242e+00 -1.42587543e+00 -6.71247303e-01 -2.92971194e-01
5.58791876e-01 -1.00038338e+00 5.37797868e-01 2.35112295e-01
-1.45362008e+00 -7.05582559e-01 -9.11381364e-01 -3.97987574e-01
2.35352218e-01 2.13120580e-01 4.14271653e-01 3.23509157e-01
-1.10161805e+00 8.29122245e-01 -9.84256983e-01 -6.24101348e-02
4.04920906e-01 3.50236773e-01 -3.46069962e-01 -2.62773037e-01
-6.25543714e-01 7.47643471e-01 1.18401768e-02 3.97059172e-01
-6.48578942e-01 -6.58797443e-01 -9.39562678e-01 -8.81083608e-02
2.25147337e-01 -9.93132412e-01 8.05113077e-01 -6.49603546e-01
-1.29862833e+00 7.28521824e-01 -7.43062675e-01 -3.39666218e-01
8.41419280e-01 -3.23782176e-01 3.00829411e-01 3.91584635e-01
1.64489269e-01 9.04801488e-01 9.17693913e-01 -1.23436010e+00
-7.49126196e-01 -1.25636086e-01 -3.09068467e-02 1.55811951e-01
-4.90704365e-02 -5.84323891e-02 -8.32390070e-01 -6.64970636e-01
2.30671540e-01 -9.40003991e-01 -5.02874732e-01 2.98165411e-01
-1.52829587e-01 -5.45055233e-02 8.22697699e-01 -7.70780206e-01
1.27735484e+00 -1.76499903e+00 4.72415090e-01 4.88239489e-02
2.53774941e-01 3.33641887e-01 -1.40299067e-01 -5.62457778e-02
2.24341080e-01 -2.95423448e-01 -3.39774817e-01 -9.29793537e-01
-3.18248093e-01 2.08890855e-01 -1.77482083e-01 7.76407957e-01
4.36266392e-01 9.38643456e-01 -9.97959435e-01 -6.28690243e-01
6.18706346e-01 6.82052612e-01 -8.31347227e-01 2.12189332e-01
-1.61371872e-01 8.18100929e-01 -4.58487600e-01 3.74331027e-01
8.16278219e-01 -1.07503429e-01 -7.43883476e-02 -4.06434864e-01
-4.12740201e-01 1.01383649e-01 -1.61894977e+00 1.93178439e+00
-1.74972221e-01 7.95643091e-01 2.28778094e-01 -8.62732530e-01
6.46526217e-01 2.13681757e-01 8.29023004e-01 -1.72875389e-01
2.19045371e-01 3.75940114e-01 -2.26776540e-01 -5.22668958e-01
5.36448181e-01 1.17339477e-01 5.81763625e-01 2.09913194e-01
-6.19036481e-02 -1.76241845e-01 5.74077189e-01 -4.41110358e-02
1.03136766e+00 6.18974686e-01 1.81988329e-01 -3.50366682e-01
9.91392612e-01 4.86146398e-02 8.25043797e-01 3.79555851e-01
-3.84168595e-01 1.00933123e+00 4.68610644e-01 -8.69394839e-01
-1.05379665e+00 -7.47138619e-01 -6.24889433e-02 4.52520579e-01
4.26987410e-01 -3.84943724e-01 -7.23441720e-01 -5.65628409e-01
-1.42279342e-01 2.60503050e-02 -5.02950132e-01 3.66909504e-01
-1.14669609e+00 -6.19881868e-01 -3.54968905e-01 4.55178618e-01
4.59618568e-01 -9.30037916e-01 -8.35563242e-01 5.59481740e-01
-5.07209361e-01 -1.50105584e+00 -7.74996698e-01 -2.68207669e-01
-9.62779164e-01 -9.20566678e-01 -7.69835532e-01 -7.94893622e-01
8.15948308e-01 4.21502948e-01 1.08110201e+00 4.55170304e-01
-5.43735802e-01 -2.00011730e-02 -4.75186631e-02 4.35909063e-01
-2.01705575e-01 -1.85927689e-01 -3.55173767e-01 4.88713354e-01
-1.35941997e-01 -5.77052772e-01 -1.08477104e+00 4.90490168e-01
-8.09543312e-01 -2.04559505e-01 2.48216912e-01 7.49459624e-01
7.65702546e-01 -3.84323150e-01 3.25066745e-02 -5.40035725e-01
-5.64271212e-03 -3.20996647e-03 -8.41087341e-01 1.66882098e-01
1.79180503e-02 8.10614824e-02 4.00288641e-01 -4.10175204e-01
-8.94084156e-01 7.02814758e-01 -6.82586730e-02 -6.47883296e-01
-8.22803825e-02 2.91763395e-02 -5.29264621e-02 -4.31419104e-01
4.82305735e-01 -8.25641751e-02 -5.30344658e-02 -3.79801661e-01
3.53704035e-01 3.37638296e-02 7.45436907e-01 -5.82677901e-01
8.27068925e-01 9.40292716e-01 3.42975914e-01 -9.54775572e-01
-6.12597942e-01 -9.38690364e-01 -1.01206398e+00 -2.65512228e-01
1.05375850e+00 -7.66382694e-01 -6.50625706e-01 5.26739299e-01
-1.49234784e+00 -3.54621708e-01 -3.87240887e-01 6.44401550e-01
-7.21110821e-01 6.67785645e-01 -5.93610823e-01 -6.36780202e-01
-1.78683490e-01 -1.60097289e+00 1.27136445e+00 9.34846178e-02
-4.29014079e-02 -1.12082350e+00 2.95468599e-01 2.37497181e-01
2.22506493e-01 4.63864475e-01 3.35737169e-01 8.08853135e-02
-1.27442670e+00 -6.94930181e-02 -1.58362016e-01 2.91810483e-01
7.81281218e-02 1.17803313e-01 -8.65318120e-01 -3.04835379e-01
7.83545226e-02 -2.84997728e-02 1.02395391e+00 7.43432045e-01
6.94689274e-01 4.39187288e-02 -2.90769607e-01 8.50124478e-01
1.57510829e+00 -2.05943123e-01 8.87793541e-01 2.02694505e-01
9.27403271e-01 7.10742772e-01 7.46868014e-01 3.76172811e-01
2.95831919e-01 1.11336124e+00 2.07070664e-01 -7.30288168e-03
-4.88737434e-01 1.02634035e-01 1.70962587e-01 5.73033333e-01
-2.08737493e-01 -1.60822690e-01 -6.09749734e-01 8.37776124e-01
-2.14501381e+00 -8.72594893e-01 -4.14632499e-01 2.38239050e+00
6.06449902e-01 -1.43057317e-01 8.40730742e-02 1.05518557e-01
6.62378967e-01 4.15170610e-01 -1.68609485e-01 7.14183673e-02
-1.47794709e-01 8.16276893e-02 2.55583107e-01 1.11284518e+00
-1.25493503e+00 1.01200175e+00 6.15641403e+00 5.43060124e-01
-9.86834824e-01 1.87313870e-01 6.52800024e-01 6.48465008e-02
-2.64819920e-01 4.15169984e-01 -8.87067676e-01 3.76295716e-01
3.04908007e-01 3.09270203e-01 1.97489664e-01 3.17472011e-01
3.73177201e-01 -4.07582253e-01 -8.96445453e-01 8.60072494e-01
-3.54944989e-02 -1.56778073e+00 -1.30951375e-01 9.02042687e-02
1.19285846e+00 -8.84913430e-02 -3.77001613e-01 -4.70263392e-01
-1.04314551e-01 -6.12592757e-01 6.11687124e-01 3.89868498e-01
3.01614583e-01 -5.15232980e-01 4.09574211e-01 1.44898649e-02
-1.47327042e+00 1.62046596e-01 -3.55922997e-01 8.81770104e-02
6.98992491e-01 7.31209219e-01 -1.26295194e-01 6.07413173e-01
3.33988488e-01 9.49678540e-01 -2.10918248e-01 1.48507583e+00
-1.87040567e-01 4.55455929e-02 -5.26945531e-01 5.00678122e-01
3.06284070e-01 -4.92420822e-01 7.49823928e-01 9.90671158e-01
1.36150867e-01 -2.62905210e-02 3.97613049e-01 7.35415637e-01
1.52019858e-01 1.56057209e-01 -2.03592315e-01 6.15917087e-01
1.35055870e-01 1.35322452e+00 -1.00489581e+00 -4.24222499e-01
-5.17349303e-01 1.15161026e+00 4.99983698e-01 4.89930302e-01
-6.39658391e-01 6.47203624e-03 7.87783921e-01 1.71701446e-01
5.51352680e-01 -5.75502574e-01 -4.04574834e-02 -1.41734624e+00
3.46237004e-01 -4.01481897e-01 2.25649104e-01 -4.14786845e-01
-7.16839731e-01 5.84342182e-01 -1.75713718e-01 -1.37862217e+00
-3.21985483e-01 -4.06109691e-01 -7.04768062e-01 9.14554834e-01
-1.86722612e+00 -9.26978290e-01 -6.00776494e-01 6.29238188e-01
6.87404692e-01 4.59973484e-01 4.02086467e-01 5.24799943e-01
-4.60343719e-01 6.91479519e-02 -9.33211073e-02 8.51297379e-02
6.51250601e-01 -1.09742618e+00 4.28470314e-01 1.29197907e+00
1.54727876e-01 3.40964317e-01 7.34928429e-01 -5.55827200e-01
-1.17587459e+00 -9.72377658e-01 1.12021029e+00 -3.65306884e-01
4.25282627e-01 -6.46610931e-02 -9.99143243e-01 5.17981052e-01
-5.82997426e-02 7.04837024e-01 7.77689321e-03 -5.59420288e-01
-4.48380634e-02 -8.74036476e-02 -9.37058210e-01 3.33589256e-01
1.01070988e+00 -2.20571712e-01 -2.19641283e-01 2.49582931e-01
4.63732570e-01 -5.90286613e-01 -8.60020578e-01 3.29740375e-01
5.15093803e-01 -1.05629814e+00 1.25318110e+00 -3.07433128e-01
3.95108730e-01 -6.65161669e-01 1.02644920e-01 -9.28997099e-01
-1.07505403e-01 -1.31599724e+00 -2.25705028e-01 9.28087413e-01
2.15389337e-02 -4.30447459e-01 9.15842175e-01 5.37154377e-01
-6.25915453e-02 -6.05944514e-01 -9.52675819e-01 -6.98861599e-01
-1.75868303e-01 -1.72881678e-01 6.56221434e-02 6.99043810e-01
-3.57727438e-01 2.26610294e-03 -5.23088753e-01 3.16807851e-02
9.41184938e-01 3.27176362e-01 1.00706196e+00 -1.01420605e+00
-3.38076055e-01 -3.56950045e-01 -6.14747226e-01 -1.58153296e+00
1.90402627e-01 -3.38683516e-01 2.30402008e-01 -1.51336455e+00
-1.59045160e-01 -4.31892842e-01 3.30014192e-02 3.91849391e-02
-6.43193126e-01 5.34991145e-01 2.74219722e-01 5.12268841e-01
-4.64845061e-01 3.82053554e-01 1.47054434e+00 9.94722024e-02
-5.48953533e-01 -9.28836968e-03 -1.03026656e-02 6.14121974e-01
4.29991096e-01 -3.04957032e-01 -3.70022476e-01 -6.52453721e-01
-2.85100847e-01 1.44569188e-01 5.32504559e-01 -1.15496516e+00
3.80927771e-01 -1.64654672e-01 6.74000531e-02 -6.18460357e-01
5.53357959e-01 -6.22513235e-01 2.44172546e-03 5.76409400e-01
-1.59642752e-02 5.97532578e-02 1.43069960e-02 5.14187455e-01
-3.64034116e-01 -2.07977772e-01 9.54693675e-01 9.27590355e-02
-7.90284753e-01 5.91533780e-01 3.20408703e-03 9.42510217e-02
9.51128542e-01 -1.45453677e-01 -1.67360350e-01 -2.98975706e-01
-8.75357151e-01 1.81564242e-01 5.58236301e-01 2.27598608e-01
5.14637530e-01 -1.20086062e+00 -9.19967651e-01 3.21495950e-01
-2.57744879e-01 2.04994962e-01 1.87211886e-01 1.20057094e+00
-1.08669829e+00 2.06932008e-01 -1.89194456e-01 -8.42694521e-01
-1.21582031e+00 4.60854530e-01 4.58580881e-01 -3.43447089e-01
-6.95965588e-01 8.35524201e-01 2.83147752e-01 3.74524087e-01
2.29046613e-01 -3.83865088e-01 -1.07760563e-01 5.91209717e-02
7.11645246e-01 4.85207379e-01 7.05545172e-02 -1.02354646e+00
-3.08582038e-01 1.22400379e+00 1.77365094e-01 -2.08942086e-01
1.14323843e+00 -3.28830004e-01 -2.39826858e-01 -3.30914259e-02
1.54458654e+00 8.41926783e-02 -1.67301047e+00 -2.44555488e-01
-1.29011467e-01 -1.05760968e+00 1.69128254e-02 -5.10806553e-02
-1.32136607e+00 8.02390277e-01 2.69447714e-01 5.85278962e-03
1.12321544e+00 -2.09401995e-01 9.23333764e-01 -1.13972150e-01
2.07936972e-01 -8.54661465e-01 -1.72735564e-03 3.50908101e-01
6.78378224e-01 -1.19388604e+00 2.61026412e-01 -8.72046888e-01
-2.72785693e-01 1.18398750e+00 3.36640805e-01 -3.90345246e-01
5.79618812e-01 9.75463241e-02 2.24736799e-02 1.00095935e-01
-6.79783344e-01 -4.02132422e-01 5.63915014e-01 3.61388028e-01
3.26324850e-01 -4.56595898e-01 -5.97754300e-01 -4.29651380e-01
3.58116329e-01 1.57488510e-02 2.93653369e-01 8.90817225e-01
-4.97456789e-01 -1.37437320e+00 -3.72848302e-01 -9.00739506e-02
-4.19805288e-01 1.40661508e-01 3.20086479e-02 5.66861928e-01
1.11492530e-01 8.90054107e-01 1.92058459e-01 3.24308611e-02
2.47305676e-01 -1.92013517e-01 5.87967098e-01 -2.95389205e-01
-5.32736957e-01 7.08232880e-01 8.97938162e-02 -1.07072198e+00
-8.43588829e-01 -8.10572326e-01 -1.08118522e+00 -2.29297161e-01
-2.38120377e-01 -9.79197957e-03 4.72028762e-01 8.09055030e-01
3.40128928e-01 2.31189549e-01 6.62027538e-01 -1.29482138e+00
-1.14759192e-01 -4.73848283e-01 -1.95808694e-01 7.00559914e-01
6.58131063e-01 -6.92252755e-01 -4.83243793e-01 3.94148827e-01] | [8.793898582458496, -1.7970376014709473] |
8306b73a-18a8-4d41-966e-8790054a07ba | programmatically-interpretable-reinforcement | 1804.02477 | null | http://arxiv.org/abs/1804.02477v3 | http://arxiv.org/pdf/1804.02477v3.pdf | Programmatically Interpretable Reinforcement Learning | We present a reinforcement learning framework, called Programmatically
Interpretable Reinforcement Learning (PIRL), that is designed to generate
interpretable and verifiable agent policies. Unlike the popular Deep
Reinforcement Learning (DRL) paradigm, which represents policies by neural
networks, PIRL represents policies using a high-level, domain-specific
programming language. Such programmatic policies have the benefits of being
more easily interpreted than neural networks, and being amenable to
verification by symbolic methods. We propose a new method, called Neurally
Directed Program Search (NDPS), for solving the challenging nonsmooth
optimization problem of finding a programmatic policy with maximal reward. NDPS
works by first learning a neural policy network using DRL, and then performing
a local search over programmatic policies that seeks to minimize a distance
from this neural "oracle". We evaluate NDPS on the task of learning to drive a
simulated car in the TORCS car-racing environment. We demonstrate that NDPS is
able to discover human-readable policies that pass some significant performance
bars. We also show that PIRL policies can have smoother trajectories, and can
be more easily transferred to environments not encountered during training,
than corresponding policies discovered by DRL. | ['Swarat Chaudhuri', 'Vijayaraghavan Murali', 'Rishabh Singh', 'Pushmeet Kohli', 'Abhinav Verma'] | 2018-04-06 | programmatically-interpretable-reinforcement-1 | https://icml.cc/Conferences/2018/Schedule?showEvent=2203 | http://proceedings.mlr.press/v80/verma18a/verma18a.pdf | icml-2018-7 | ['carracing-v0'] | ['playing-games'] | [ 9.64636430e-02 7.13175237e-01 -5.90243280e-01 -4.02182281e-01
-6.97594583e-01 -7.97721922e-01 8.28162670e-01 2.99489610e-02
-3.88145387e-01 1.11621737e+00 -3.39974687e-02 -9.08916056e-01
-2.48332828e-01 -7.97063351e-01 -1.45649338e+00 -5.61210215e-01
-3.79210472e-01 8.30835581e-01 8.12743753e-02 -1.70982420e-01
1.35740027e-01 5.69398224e-01 -1.57023871e+00 6.97542056e-02
8.47377598e-01 5.60127020e-01 1.20431669e-01 8.06154966e-01
2.85101801e-01 1.37184036e+00 -4.12340373e-01 1.54040053e-01
3.17255855e-01 -2.92199880e-01 -9.35926735e-01 -3.24119478e-01
1.82377234e-01 -6.14955127e-01 -2.31977418e-01 1.20055091e+00
-6.90010265e-02 2.23017320e-01 3.07220161e-01 -1.75387621e+00
-7.64272690e-01 8.36321294e-01 1.41227126e-01 -3.54904413e-01
2.97675073e-01 8.60357463e-01 1.02067888e+00 -2.00371221e-01
5.26814044e-01 1.73950219e+00 2.07794309e-01 9.52794611e-01
-1.44047856e+00 -4.08193886e-01 2.13543832e-01 2.25570407e-02
-6.86819196e-01 -1.47888705e-01 3.69491011e-01 -5.55525720e-01
1.12119794e+00 2.20310301e-01 5.50969541e-01 1.28585768e+00
2.33363599e-01 8.97860765e-01 1.32412922e+00 -3.55738699e-01
7.06293523e-01 1.54068723e-01 -9.22413170e-02 1.18407857e+00
1.74891904e-01 1.12476480e+00 -2.38935366e-01 -4.63042527e-01
7.90842831e-01 -3.48283648e-01 -1.02292567e-01 -6.09354198e-01
-1.28401077e+00 1.02758086e+00 1.91442326e-01 -3.95260811e-01
-4.94147807e-01 9.13522243e-01 5.69569230e-01 4.86398041e-01
-2.34020784e-01 1.12918186e+00 -6.24746323e-01 -4.46989566e-01
-1.39821127e-01 8.84390533e-01 1.10171759e+00 7.97290027e-01
7.46645391e-01 6.15165055e-01 -3.97152275e-01 1.27217427e-01
3.13325703e-01 8.06823552e-01 3.56834292e-01 -1.62001574e+00
1.73603341e-01 2.99372047e-01 4.77794677e-01 -6.96365416e-01
-5.11066727e-02 2.80449837e-01 -2.28930980e-01 1.04565823e+00
2.14615181e-01 -5.93315065e-01 -7.15790391e-01 1.89114726e+00
1.62199780e-01 5.65325730e-02 4.22940701e-01 8.64707112e-01
1.30918771e-01 9.86870468e-01 1.63652733e-01 2.28089932e-02
8.70770037e-01 -1.11119664e+00 -1.19923659e-01 -3.68902802e-01
5.65299392e-01 3.52787375e-01 1.39737606e+00 4.85227853e-01
-1.05237460e+00 -3.21451336e-01 -9.86150086e-01 4.03116852e-01
-1.55640662e-01 -1.92160293e-01 7.53343284e-01 4.04315323e-01
-1.31609809e+00 7.93200314e-01 -1.11621487e+00 8.80522951e-02
4.47782874e-01 5.62624812e-01 -7.75646120e-02 4.18684870e-01
-1.01254869e+00 1.02464533e+00 8.77507925e-01 -2.72842497e-01
-2.00451636e+00 -5.24026632e-01 -1.09236336e+00 2.34487742e-01
7.79505312e-01 -4.23827767e-01 1.97892034e+00 -1.32546222e+00
-2.02768826e+00 5.16821980e-01 1.23808067e-02 -1.05068505e+00
4.93057430e-01 -9.77934431e-03 -1.30264238e-01 5.04839830e-02
-3.41463797e-02 8.89771104e-01 9.04252350e-01 -1.39095712e+00
-6.85505211e-01 1.45930275e-01 6.39015734e-01 5.87182567e-02
6.22943863e-02 -7.40311444e-02 2.21565560e-01 -2.70190358e-01
-7.55263925e-01 -1.24897134e+00 -5.92451155e-01 5.40927798e-02
-4.21365887e-01 -5.53169072e-01 8.55667293e-01 -3.92434776e-01
5.93312204e-01 -1.85910988e+00 1.70094788e-01 4.39026326e-01
8.12166184e-02 4.29329664e-01 -4.33879584e-01 1.43169090e-01
7.80405253e-02 2.48780102e-01 -1.68519035e-01 2.29914129e-01
5.99994063e-01 7.91659474e-01 -7.87857592e-01 2.28703424e-01
2.97139287e-01 1.13099766e+00 -1.30518830e+00 -8.52849558e-02
3.15516219e-02 -2.39109129e-01 -8.40054750e-01 3.93083602e-01
-1.19015741e+00 5.20698130e-01 -7.48775780e-01 3.92477334e-01
-1.00185268e-01 -2.70086657e-02 3.06890517e-01 6.31422460e-01
-5.17525151e-02 2.96055943e-01 -8.82282317e-01 1.09032238e+00
-3.91811937e-01 7.72894382e-01 -4.77064215e-03 -1.20714974e+00
1.12782168e+00 2.79486626e-01 1.89936504e-01 -7.61747718e-01
3.57724540e-02 1.09661750e-01 -1.58728808e-01 -6.64782524e-01
3.39856505e-01 5.22920787e-02 -3.25834513e-01 6.10284626e-01
-1.96522593e-01 -3.95056039e-01 1.88626230e-01 -1.69467703e-01
1.20166659e+00 4.77625787e-01 2.39102587e-01 -4.34261590e-01
4.50984746e-01 4.37068850e-01 6.52922571e-01 1.40922952e+00
-2.99299985e-01 -3.08203042e-01 1.04482901e+00 -6.92701757e-01
-1.21602046e+00 -1.07712567e+00 4.10994858e-01 1.12344682e+00
-1.90419145e-02 -3.14996690e-02 -8.03321123e-01 -8.91257167e-01
3.86333704e-01 1.11836159e+00 -4.97269034e-01 -3.03524435e-01
-8.22413862e-01 6.20828494e-02 8.07693541e-01 4.71031696e-01
5.19493401e-01 -1.73323250e+00 -9.73194063e-01 1.42231673e-01
3.77915502e-01 -7.50268102e-01 -2.70412594e-01 3.49436134e-01
-6.83600962e-01 -1.16395688e+00 -8.95270407e-02 -9.58370686e-01
8.84518266e-01 -4.70857978e-01 1.01536465e+00 4.02275892e-03
2.81332940e-01 5.58802545e-01 6.90574050e-02 -4.63087618e-01
-1.18836486e+00 -1.74509376e-01 3.60940337e-01 -4.93578702e-01
2.40009259e-02 -2.82213300e-01 -6.33621663e-02 9.86535922e-02
-5.93252420e-01 1.91711023e-01 3.39003950e-01 1.10461318e+00
5.97934902e-01 2.13595152e-01 5.97990394e-01 -8.87949288e-01
1.15091836e+00 -2.64575630e-01 -1.43346298e+00 3.53672236e-01
-8.08914781e-01 7.66284704e-01 1.24178195e+00 -5.35336614e-01
-8.86055291e-01 1.79077923e-01 2.00620666e-01 -4.64144975e-01
-4.08798665e-01 3.01080108e-01 -3.33351865e-02 -1.20020457e-01
1.05709136e+00 2.83206910e-01 3.40880811e-01 -1.14919208e-02
5.29941976e-01 2.63767362e-01 7.80751169e-01 -1.40219641e+00
9.56406534e-01 -3.04668676e-02 -6.45353943e-02 -1.76762372e-01
-3.25599343e-01 5.36914945e-01 1.28434613e-01 -1.07839443e-01
7.14151382e-01 -6.14161789e-01 -1.46168709e+00 8.25784057e-02
-9.98130262e-01 -1.23051250e+00 -6.86051548e-01 1.79160759e-01
-1.24073374e+00 -9.11179483e-02 -4.59764779e-01 -7.68249035e-01
-2.01748729e-01 -1.51942217e+00 5.89710653e-01 2.33145863e-01
-4.28727239e-01 -8.83294344e-01 3.21403831e-01 -1.37521237e-01
3.72229695e-01 4.35100406e-01 1.41092825e+00 -7.91945517e-01
-8.38424325e-01 1.51989743e-01 1.52152985e-01 5.64190388e-01
-2.73203045e-01 1.09617688e-01 -6.68901980e-01 -4.58560228e-01
-3.15968215e-01 -7.66981423e-01 7.48497620e-02 3.20870191e-01
1.61651957e+00 -1.17511725e+00 -1.43502742e-01 4.92497236e-01
1.16554832e+00 7.93537199e-01 3.40804130e-01 7.39744008e-01
3.45855713e-01 3.86404693e-01 6.15592897e-01 2.79499829e-01
2.91361034e-01 3.60609651e-01 7.97081888e-01 1.10619411e-01
5.04429996e-01 -7.63604522e-01 9.12722409e-01 -1.73747227e-01
2.83788368e-02 1.10005185e-01 -9.81579781e-01 3.44540358e-01
-2.26581264e+00 -1.03118658e+00 5.15424252e-01 1.89893579e+00
9.93842900e-01 3.47194642e-01 3.05881679e-01 -4.03275520e-01
4.45391268e-01 -1.86677918e-01 -1.16179693e+00 -1.17755544e+00
1.99763626e-01 3.40058804e-01 6.79330587e-01 6.93193614e-01
-8.60756934e-01 1.09651124e+00 6.78028011e+00 5.60325205e-01
-1.05729568e+00 -1.62549883e-01 6.79313600e-01 2.16915533e-01
-4.59297925e-01 1.64031491e-01 -4.75942850e-01 3.45883757e-01
1.06675303e+00 -5.32723546e-01 1.27936172e+00 1.58765268e+00
6.39414966e-01 4.73386757e-02 -1.49636519e+00 4.20083016e-01
-5.33006251e-01 -1.58536315e+00 -8.50338340e-02 -1.48358360e-01
8.08949053e-01 3.01098805e-02 1.63493961e-01 9.61875856e-01
1.46408927e+00 -1.41665316e+00 1.07471228e+00 2.33024314e-01
5.49888670e-01 -9.58647966e-01 1.22663714e-01 5.53951383e-01
-5.28719902e-01 -5.87481976e-01 -2.47442245e-01 -9.98460129e-02
-2.95716703e-01 -2.58854032e-01 -1.18733001e+00 -1.63202867e-01
5.04193544e-01 5.95895827e-01 -1.75993726e-01 7.29978144e-01
-8.50205004e-01 7.34331608e-01 -8.67406875e-02 -5.58065891e-01
6.53802156e-01 -9.53101814e-02 6.27931416e-01 8.67489874e-01
8.57848823e-02 -1.92672133e-01 4.37755942e-01 1.55852628e+00
1.33299604e-01 -5.63062727e-01 -9.00415301e-01 -3.84292722e-01
4.36807960e-01 8.52715075e-01 -1.43609241e-01 -4.94308889e-01
1.48510695e-01 4.67733234e-01 5.74065745e-01 6.08343601e-01
-1.04871118e+00 -2.27851644e-01 9.52417254e-01 -3.71542394e-01
2.25192219e-01 -3.40138048e-01 2.33197268e-02 -7.58134484e-01
-2.74614543e-01 -1.55068231e+00 2.30104360e-03 -9.45355058e-01
-7.75901496e-01 4.90367800e-01 1.21401213e-01 -7.62512326e-01
-7.31400549e-01 -7.71170378e-01 -6.97971046e-01 5.99089086e-01
-1.32582676e+00 -6.50813758e-01 1.72408059e-01 5.98789930e-01
3.92766118e-01 -4.75059092e-01 9.01809335e-01 -3.76603007e-01
-5.76739073e-01 4.35967863e-01 7.96835646e-02 -8.02332815e-03
7.64203295e-02 -1.55178726e+00 4.35895026e-01 6.70103073e-01
-2.08099425e-01 3.74043107e-01 8.79230917e-01 -3.85327101e-01
-1.77736509e+00 -1.33707094e+00 2.20195949e-01 -3.43782961e-01
8.28467429e-01 -1.55983999e-01 -6.68440521e-01 1.11208141e+00
1.31233543e-01 -2.96958685e-01 -2.20557317e-01 -1.37522817e-01
-2.49034837e-01 -6.42793253e-02 -1.25522089e+00 1.00128806e+00
6.56944931e-01 -5.79610169e-01 -7.02456117e-01 4.56346482e-01
1.15633714e+00 -6.23152971e-01 -5.66006780e-01 1.95256680e-01
1.90690473e-01 -5.15506804e-01 8.65999937e-01 -1.22434437e+00
5.89155614e-01 -3.43343794e-01 1.11973010e-01 -1.61429870e+00
-1.68014988e-01 -1.10352409e+00 -2.73597479e-01 6.13173962e-01
4.12181288e-01 -9.23527360e-01 8.95220995e-01 9.19022083e-01
-2.30229229e-01 -7.46900797e-01 -7.25289822e-01 -1.08899415e+00
2.35214978e-01 -2.78202862e-01 8.97084296e-01 6.83955371e-01
3.21797788e-01 -7.88613260e-02 -3.19221854e-01 3.09014082e-01
4.68163788e-01 9.01357681e-02 8.46032798e-01 -5.40056527e-01
-8.28292370e-01 -5.87891102e-01 3.19742300e-02 -1.09585047e+00
8.85065913e-01 -1.06444919e+00 6.80754483e-01 -1.10336769e+00
-2.28381678e-01 -6.97882354e-01 -9.21371356e-02 1.14588070e+00
2.82569379e-01 -7.84495831e-01 2.14235947e-01 -9.95231718e-02
-6.35132372e-01 6.41019583e-01 1.37959826e+00 -4.24301058e-01
-2.63101071e-01 -2.83990928e-04 -6.28330886e-01 8.30766261e-01
1.14686728e+00 -4.96469736e-01 -6.23129606e-01 -4.20491040e-01
1.01087697e-01 4.72315520e-01 7.52499461e-01 -7.38828182e-01
1.10686511e-01 -1.05757523e+00 -2.45313436e-01 -1.60496514e-02
-1.25829518e-01 -6.64137363e-01 2.16556489e-02 9.44278002e-01
-8.92211318e-01 3.47256452e-01 3.87851298e-01 5.00991881e-01
5.15581295e-02 -3.80847871e-01 6.51584506e-01 -1.54698193e-01
-9.90187526e-01 2.56418347e-01 -9.42333102e-01 1.14892371e-01
1.30089009e+00 2.98809707e-01 -5.22104979e-01 -4.52507883e-01
-5.67337036e-01 7.40986109e-01 3.77174437e-01 3.53343934e-01
6.95622385e-01 -1.05549562e+00 -4.32768792e-01 2.66902685e-01
-8.53484422e-02 2.32400503e-02 -5.14895737e-01 1.85939267e-01
-7.72714734e-01 5.05997181e-01 -3.13065529e-01 -3.96145701e-01
-7.70211816e-01 6.58849835e-01 8.20611894e-01 -2.25203842e-01
-7.02296615e-01 4.90281433e-01 4.42207232e-02 -1.02893448e+00
4.73058105e-01 -6.44441724e-01 1.45498917e-01 -1.05742347e+00
3.31030548e-01 4.92531545e-02 -5.00702322e-01 6.74559250e-02
1.87436603e-02 -1.93387225e-01 1.17411472e-01 -1.93581924e-01
1.26191425e+00 6.87649131e-01 -1.45415381e-01 -2.10723534e-01
7.20611334e-01 -6.07285976e-01 -1.79360271e+00 -2.69848201e-02
2.30213404e-01 -2.09225118e-01 -2.03064695e-01 -9.22806501e-01
-8.20108831e-01 4.27267939e-01 3.43864620e-01 2.79637694e-01
6.33731961e-01 -2.51062065e-01 4.31539148e-01 1.22795618e+00
5.26183486e-01 -1.08880794e+00 2.58961111e-01 8.44970345e-01
9.32752132e-01 -1.15874910e+00 -4.44850892e-01 4.10929233e-01
-8.96168649e-01 1.17271090e+00 9.85910475e-01 -4.55840319e-01
8.54718313e-02 4.28761274e-01 -3.01264554e-01 -8.27985927e-02
-9.75279093e-01 2.35384449e-01 -2.22562954e-01 8.07377756e-01
-6.41720831e-01 3.17056745e-01 4.07010466e-01 2.17419863e-01
-3.23167562e-01 2.39644736e-01 5.85204601e-01 1.07913947e+00
-7.33386099e-01 -1.09384012e+00 -3.09245229e-01 2.35406563e-01
1.14404492e-01 2.52689689e-01 -6.78387806e-02 7.54009485e-01
-1.38842076e-01 6.66514695e-01 -3.81982513e-02 -2.91994780e-01
2.40912616e-01 -1.55099794e-01 3.91940594e-01 -4.57436204e-01
-4.09652561e-01 -7.22941816e-01 2.78123319e-01 -9.02294278e-01
-5.63542061e-02 -4.21246678e-01 -1.59319949e+00 -5.10502100e-01
5.21888137e-01 2.39302441e-01 4.90938097e-01 8.51785064e-01
3.35012347e-01 3.90967071e-01 6.86563373e-01 -7.37224758e-01
-1.17347968e+00 -2.63258698e-04 -1.96876854e-01 2.42014930e-01
6.34453773e-01 -4.36551034e-01 -1.80429757e-01 -2.66659912e-02] | [4.276098728179932, 1.7747491598129272] |
24d3735c-05f6-4111-aa25-211bd22396eb | end-to-end-interpretable-learning-of-non | 2007.01769 | null | https://arxiv.org/abs/2007.01769v2 | https://arxiv.org/pdf/2007.01769v2.pdf | End-to-end Interpretable Learning of Non-blind Image Deblurring | Non-blind image deblurring is typically formulated as a linear least-squares problem regularized by natural priors on the corresponding sharp picture's gradients, which can be solved, for example, using a half-quadratic splitting method with Richardson fixed-point iterations for its least-squares updates and a proximal operator for the auxiliary variable updates. We propose to precondition the Richardson solver using approximate inverse filters of the (known) blur and natural image prior kernels. Using convolutions instead of a generic linear preconditioner allows extremely efficient parameter sharing across the image, and leads to significant gains in accuracy and/or speed compared to classical FFT and conjugate-gradient methods. More importantly, the proposed architecture is easily adapted to learning both the preconditioner and the proximal operator using CNN embeddings. This yields a simple and efficient algorithm for non-blind image deblurring which is fully interpretable, can be learned end to end, and whose accuracy matches or exceeds the state of the art, quite significantly, in the non-uniform case. | ['Jean Ponce', 'Jian Sun', 'Thomas Eboli'] | 2020-07-03 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2749_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123620307.pdf | eccv-2020-8 | ['blind-image-deblurring'] | ['computer-vision'] | [ 1.75272197e-01 -1.88890636e-01 1.92072079e-01 -6.08847886e-02
-7.07003534e-01 -4.10766959e-01 5.74064493e-01 -3.14946830e-01
-5.19358993e-01 6.19627714e-01 4.10280228e-01 -1.69497490e-01
-3.07200879e-01 -1.46513626e-01 -7.69243896e-01 -9.67603683e-01
7.75252411e-04 2.08070129e-01 -1.65547758e-01 1.14570417e-01
3.54516327e-01 4.37695652e-01 -1.37770391e+00 -1.32955983e-01
9.99342322e-01 1.04233229e+00 2.77720511e-01 9.09262002e-01
2.33009413e-01 1.11231399e+00 6.72275200e-02 -2.10497111e-01
2.52273500e-01 -3.62178475e-01 -8.63983154e-01 2.76070803e-01
8.64553869e-01 -6.58789396e-01 -5.55404484e-01 1.09697509e+00
4.92958575e-01 1.60263881e-01 6.45689428e-01 -7.47372270e-01
-9.62015569e-01 1.79233789e-01 -7.50937998e-01 7.87045658e-02
2.77659506e-01 1.16815090e-01 7.16916978e-01 -1.13972020e+00
3.79660577e-01 1.09677505e+00 1.08983159e+00 3.11577260e-01
-1.47693455e+00 -3.03378016e-01 -2.36982852e-01 5.05416751e-01
-1.45035005e+00 -6.60561204e-01 5.24757624e-01 -5.82284033e-01
6.58592463e-01 3.22020173e-01 5.17709136e-01 7.94427335e-01
-4.91331257e-02 6.53267801e-01 1.32214379e+00 -4.94434297e-01
2.98913181e-01 1.63682058e-01 9.83141214e-02 6.65869772e-01
-6.03731629e-03 1.95203528e-01 -4.19930667e-01 -3.59856963e-01
9.26905274e-01 2.67303213e-02 -9.52568889e-01 -4.91291583e-01
-1.36010730e+00 6.84156835e-01 6.38523698e-01 2.08692685e-01
-6.16656244e-01 3.20757568e-01 7.84071013e-02 2.55441487e-01
6.47927403e-01 3.24125588e-01 -4.33053464e-01 2.97454782e-02
-1.31309593e+00 1.82158157e-01 9.51815248e-01 5.27620852e-01
1.17071438e+00 1.15850801e-02 -8.70953873e-02 6.68490469e-01
2.99836695e-01 4.82289284e-01 4.57440317e-01 -1.12592268e+00
-2.61661708e-02 -1.68632194e-02 6.14890516e-01 -8.96370828e-01
-3.44250560e-01 -5.75986207e-01 -1.25714171e+00 4.04210448e-01
6.52680814e-01 -1.75894514e-01 -6.97309434e-01 1.61038268e+00
3.65605563e-01 7.58080125e-01 -7.88106024e-02 1.35883892e+00
2.74996787e-01 6.24353349e-01 -3.88290197e-01 -3.46561432e-01
1.48328722e+00 -1.12860310e+00 -7.64699042e-01 -4.94677395e-01
4.35356945e-01 -8.82913470e-01 1.01137352e+00 3.79488677e-01
-1.01578474e+00 -3.95557880e-01 -9.31070626e-01 -4.76046354e-01
2.59236409e-03 4.40259129e-01 3.51564616e-01 4.78736073e-01
-1.46390069e+00 7.90255845e-01 -7.64501333e-01 -2.29811698e-01
3.05008054e-01 2.22468033e-01 -3.98393333e-01 -3.59317511e-01
-9.76019204e-01 1.03719378e+00 -7.78312683e-02 4.16990340e-01
-8.55632663e-01 -1.04935098e+00 -9.50411439e-01 1.41406953e-01
1.05827257e-01 -1.06709206e+00 9.32079256e-01 -1.29359305e+00
-1.80936730e+00 7.67979860e-01 -2.57767856e-01 -3.11172336e-01
7.02448606e-01 -6.59527242e-01 1.40196756e-01 2.19421789e-01
-1.91994116e-01 2.64328897e-01 1.70517206e+00 -1.12658906e+00
3.83607559e-02 -3.08611274e-01 -7.12964172e-03 2.69783437e-01
-3.59774739e-01 5.16183116e-02 -3.50798756e-01 -8.62921119e-01
2.41669212e-02 -9.93846834e-01 -3.86730194e-01 4.16855454e-01
-1.19026333e-01 4.62223589e-01 5.20515025e-01 -1.36726236e+00
9.99956787e-01 -2.31398463e+00 6.08635783e-01 -1.42261952e-01
3.09549540e-01 9.65085775e-02 -2.20346853e-01 1.86677381e-01
-4.28629756e-01 -5.45070291e-01 -6.26946509e-01 -4.78472292e-01
-1.41662151e-01 -3.90994027e-02 -3.31894934e-01 1.20123363e+00
-5.85046411e-03 7.58445323e-01 -8.64276588e-01 1.37430117e-01
3.91384065e-01 9.29051757e-01 -5.25279999e-01 4.54076409e-01
1.38574854e-01 6.48958087e-01 -1.37854248e-01 -1.77656099e-01
1.04118681e+00 -4.79546189e-01 -1.44832060e-01 -5.63917398e-01
-3.27266634e-01 -5.47388196e-02 -1.43444324e+00 1.85691392e+00
-9.11479235e-01 8.28432798e-01 7.11076617e-01 -1.03148210e+00
3.90039742e-01 3.77468646e-01 1.71895847e-01 -2.59703696e-01
-9.79305059e-02 3.20083827e-01 -5.75097680e-01 -5.67322195e-01
4.43174362e-01 -1.67502582e-01 6.20739400e-01 5.57494521e-01
7.25952312e-02 -2.11477861e-01 -1.27255410e-01 1.16534308e-01
8.90747488e-01 1.48968309e-01 2.48924747e-01 -6.88459992e-01
8.82699132e-01 -2.96459407e-01 1.01640575e-01 5.97739756e-01
2.55275667e-01 9.17000115e-01 1.20936722e-01 -4.02187526e-01
-1.00072742e+00 -7.03004837e-01 -2.41949946e-01 8.74475896e-01
2.46913657e-01 -1.46196261e-01 -1.11929607e+00 -2.24319622e-01
-1.42801285e-01 4.82879192e-01 -5.14583528e-01 3.53056416e-02
-5.09704232e-01 -8.64781022e-01 6.38558716e-02 1.56433105e-01
6.73454881e-01 -4.80899751e-01 -3.38579923e-01 9.02590752e-02
-1.80312559e-01 -1.13543296e+00 -8.84909570e-01 -6.06389120e-02
-8.06369126e-01 -1.00248134e+00 -1.51026511e+00 -7.57918358e-01
7.60274589e-01 5.83456337e-01 8.40445876e-01 3.32909003e-02
-2.70314336e-01 5.60689449e-01 -6.68618008e-02 5.43593168e-01
-1.93838537e-01 -3.70255917e-01 6.48767427e-02 6.48817122e-01
-3.72119427e-01 -5.85670412e-01 -9.15294111e-01 3.46934676e-01
-9.50023770e-01 1.38781577e-01 5.99354982e-01 1.05631757e+00
2.68156230e-01 -1.79448664e-01 7.42217004e-02 -5.24838150e-01
4.81636703e-01 -2.32842132e-01 -7.69198179e-01 1.69657052e-01
-6.06317401e-01 2.77303278e-01 8.08153510e-01 -4.43878919e-01
-1.21245384e+00 9.23057869e-02 1.19209655e-01 -6.00346625e-01
-6.47264197e-02 3.58018547e-01 2.15629324e-01 -6.37150347e-01
9.18851435e-01 3.32101762e-01 1.18011169e-01 -8.49414945e-01
8.93766999e-01 6.22977555e-01 7.12358057e-01 -3.12600732e-01
1.04247606e+00 7.15523958e-01 -8.83713067e-02 -1.05659747e+00
-6.90307617e-01 -6.25603259e-01 -5.02546072e-01 -8.22132006e-02
9.27993119e-01 -1.25165200e+00 -6.01026535e-01 1.05153477e+00
-1.23427629e+00 -6.05947077e-01 -2.43749797e-01 6.74108922e-01
-7.26214290e-01 8.98309827e-01 -9.27686751e-01 -5.86728215e-01
-2.62876093e-01 -1.10164917e+00 8.89318228e-01 1.30413383e-01
7.02607865e-03 -1.15436757e+00 3.60180177e-02 2.75489599e-01
8.82486284e-01 -2.77100950e-01 5.59335172e-01 2.53472447e-01
-6.57589197e-01 -6.75083324e-02 -6.63062036e-01 6.07323170e-01
2.45808177e-02 -5.43487191e-01 -1.15968597e+00 -7.24199533e-01
5.95750213e-01 -1.64157718e-01 1.02936316e+00 7.38707125e-01
8.46349955e-01 -7.51768470e-01 -1.93674900e-02 9.97903943e-01
1.52652502e+00 -9.02959526e-01 4.93782699e-01 1.13860883e-01
7.96500564e-01 4.00268734e-01 5.54499626e-02 5.25982320e-01
3.15532029e-01 8.32872093e-01 2.83802092e-01 -3.50626588e-01
-4.14158314e-01 2.40639552e-01 5.71842074e-01 6.64824128e-01
-1.15629546e-01 1.87422425e-01 -6.88480794e-01 5.56591451e-01
-1.85099435e+00 -8.67079496e-01 -4.15801823e-01 2.42847061e+00
9.50100303e-01 -6.36705399e-01 -2.14554802e-01 4.96069565e-02
6.36788309e-01 3.36373657e-01 -4.50691253e-01 -6.30352693e-03
-1.21096663e-01 9.60434303e-02 7.77304590e-01 1.00353980e+00
-1.16813385e+00 7.46014237e-01 6.37708092e+00 8.67840350e-01
-1.23086023e+00 4.74612087e-01 5.39014459e-01 1.53120607e-01
-1.29040927e-01 1.92548573e-01 -9.58836228e-02 2.88645774e-01
7.87084997e-01 1.01163544e-01 1.34349716e+00 5.93467414e-01
3.66861939e-01 -1.64135754e-01 -9.87890124e-01 1.32054317e+00
7.80020952e-02 -1.30698967e+00 -3.98953736e-01 -6.77305534e-02
9.24875140e-01 1.78129897e-01 2.24544674e-01 -3.09248954e-01
8.30678195e-02 -7.71847665e-01 8.20342481e-01 6.50131464e-01
8.76898229e-01 -2.62261897e-01 4.28585649e-01 3.31827193e-01
-7.62062311e-01 -1.19085200e-01 -2.51929939e-01 -2.02991456e-01
1.45719618e-01 8.86771858e-01 -2.85798013e-02 3.90305638e-01
7.39037275e-01 8.82639527e-01 -2.50073463e-01 1.14803374e+00
-3.25348169e-01 5.72982669e-01 -2.50506371e-01 5.75568616e-01
2.14269251e-01 -6.13565385e-01 7.23372102e-01 1.24177802e+00
4.94281977e-01 2.09549628e-02 -1.67084068e-01 9.30175662e-01
1.01197533e-01 -3.57659794e-02 -6.56306297e-02 3.63388658e-01
-1.67089671e-01 1.29141414e+00 -2.53869593e-01 -3.58774126e-01
-6.16634488e-01 1.58230448e+00 2.25848779e-01 9.71996188e-01
-6.19953752e-01 -2.63759166e-01 8.37177277e-01 4.87351567e-02
5.22232473e-01 -3.28338116e-01 -1.81712478e-01 -1.68754756e+00
4.65064012e-02 -1.01672924e+00 4.71665040e-02 -9.98956680e-01
-1.35867667e+00 4.76757169e-01 -3.60641927e-01 -1.00295091e+00
-1.18859001e-01 -7.57594943e-01 -5.90817034e-01 1.36825860e+00
-1.92368674e+00 -1.03965425e+00 -5.35882115e-01 8.07179093e-01
2.07064703e-01 3.25829804e-01 6.93232834e-01 3.63971561e-01
-5.10869443e-01 3.46749127e-01 5.33133626e-01 -2.15859711e-01
8.82655501e-01 -1.29573834e+00 1.44445643e-01 1.09998369e+00
-2.11604357e-01 5.91731668e-01 1.05260801e+00 -1.03896581e-01
-1.73625946e+00 -8.12052369e-01 6.90073133e-01 1.77939422e-02
8.82297516e-01 -1.50509328e-01 -1.21468282e+00 5.82438827e-01
4.41325307e-01 3.81045043e-01 1.80631265e-01 -1.49987876e-01
-4.92141724e-01 -1.00679435e-01 -1.00443089e+00 3.16480041e-01
5.67740977e-01 -9.05640364e-01 -2.08241165e-01 6.15318060e-01
3.01909894e-01 -4.92013246e-01 -6.62920237e-01 -6.98742121e-02
4.11967933e-01 -1.03617787e+00 1.22218871e+00 -2.52117753e-01
3.12058568e-01 -4.75661963e-01 5.03238812e-02 -1.42451167e+00
-6.87587976e-01 -1.21557522e+00 -3.44525784e-01 9.00810778e-01
4.81091328e-02 -6.82344854e-01 4.77430016e-01 5.88889837e-01
-1.29931550e-02 -3.68057609e-01 -9.65033174e-01 -5.44406950e-01
-9.32591483e-02 -3.13540727e-01 1.79011449e-01 1.06058192e+00
-2.09405437e-01 1.96292829e-02 -8.84670675e-01 6.05425775e-01
1.00627708e+00 9.34267938e-02 6.07721210e-01 -7.45453835e-01
-6.45374656e-01 -4.24969018e-01 -2.28120267e-01 -1.78280616e+00
2.15399921e-01 -7.89735854e-01 2.50858963e-01 -1.25874960e+00
1.41470134e-01 -2.31342643e-01 2.74971984e-02 2.24075496e-01
-3.90365392e-01 5.05404115e-01 -7.17177093e-02 3.59802306e-01
-8.15822035e-02 6.37809038e-01 1.17676401e+00 -1.85110271e-01
-5.59235811e-02 -8.77878666e-02 -4.40462261e-01 6.74513578e-01
2.91087598e-01 -2.58549631e-01 -5.87818101e-02 -8.05710196e-01
3.58180255e-01 2.44476497e-01 8.50465238e-01 -8.74699056e-01
4.63828146e-01 1.57741204e-01 9.16750059e-02 2.25394979e-01
3.39031488e-01 -7.26282001e-01 2.26003572e-01 2.97787249e-01
-2.46241897e-01 -3.80428851e-01 7.11395741e-02 6.43839180e-01
-1.91305488e-01 -4.14522767e-01 1.10636449e+00 7.87565187e-02
-4.77226079e-01 2.75567591e-01 -3.74251038e-01 -1.29130766e-01
4.45291877e-01 -8.34091231e-02 -8.80125836e-02 -6.36291087e-01
-7.30105758e-01 -2.36578941e-01 8.11024368e-01 -1.03373416e-02
4.36441392e-01 -1.16959083e+00 -9.18669164e-01 2.90773809e-01
-3.04808646e-01 -2.53019691e-01 6.28344595e-01 1.46607065e+00
-7.17576385e-01 -3.89609076e-02 5.68180159e-02 -5.00597894e-01
-9.22946632e-01 6.39029086e-01 5.93527794e-01 -5.79596721e-02
-6.44716144e-01 9.24922049e-01 4.47269499e-01 -1.30688608e-01
6.52618334e-02 -1.89434156e-01 8.36831778e-02 -4.27340455e-02
9.12559509e-01 6.27951264e-01 1.22919176e-02 -8.27841759e-01
-2.30872780e-01 9.73014534e-01 1.73006609e-01 -1.20036930e-01
1.29651248e+00 -4.64905947e-01 -6.67367816e-01 2.37765443e-03
1.49333262e+00 1.50495350e-01 -1.67298651e+00 -4.33295846e-01
-2.42921457e-01 -6.19721055e-01 7.65567362e-01 -5.20822108e-01
-1.21693408e+00 8.48359942e-01 7.95200646e-01 1.39598651e-02
1.23050773e+00 -1.72732174e-01 7.07704544e-01 1.20962381e-01
1.20547719e-01 -4.02250946e-01 -1.83497563e-01 4.24277663e-01
1.19892585e+00 -1.12246180e+00 1.93856135e-01 -2.63957471e-01
-1.59378231e-01 1.25188851e+00 -1.87313229e-01 -4.78302062e-01
8.06402743e-01 1.21124558e-01 2.42222220e-01 2.09749147e-01
-2.31710255e-01 5.32697625e-02 5.54530501e-01 4.26044643e-01
3.27605546e-01 -2.01321423e-01 -2.32945040e-01 -2.51481347e-02
2.65398651e-01 1.87889174e-01 4.81046170e-01 2.32306167e-01
-2.27117002e-01 -7.79414713e-01 -6.33363128e-01 1.64808020e-01
-2.24602461e-01 -4.21045184e-01 1.49868265e-01 1.80017069e-01
-1.46499619e-01 9.07308578e-01 -1.34195343e-01 1.32091179e-01
-6.99447319e-02 -1.70829386e-01 5.82666457e-01 -1.08336374e-01
-4.48896706e-01 4.69345972e-02 -3.04788500e-01 -8.01877379e-01
-5.12421489e-01 -7.49538422e-01 -7.11079240e-01 -3.13337475e-01
-4.59568024e-01 5.16940691e-02 5.15538752e-01 9.56321657e-01
4.65620548e-01 8.13448057e-03 6.08763337e-01 -1.65507913e+00
-7.95549393e-01 -9.83051836e-01 -5.86323857e-01 4.51614052e-01
8.66286635e-01 -3.57290477e-01 -8.30806851e-01 2.86768854e-01] | [11.639678001403809, -2.6214439868927] |
e0ae5ee1-5083-4649-a1bd-b03744138042 | time-delay-multi-feature-correlation-analysis | 2305.09478 | null | https://arxiv.org/abs/2305.09478v2 | https://arxiv.org/pdf/2305.09478v2.pdf | Time delay multi-feature correlation analysis to extract subtle dependencies from EEG signals | Electroencephalography (EEG) signals are resultants of extremely complex brain activity. Some details of this hidden dynamics might be accessible through e.g. joint distributions $\rho_{\Delta t}$ of signals of pairs of electrodes shifted by various time delays (lag $\Delta t$). A standard approach is monitoring a single evaluation of such joint distributions, like Pearson correlation (or mutual information), which turns out relatively uninteresting - as expected, there is usually a small peak for zero delay and nearly symmetric drop with delay. In contrast, such a complex signal might be composed of multiple types of statistical dependencies - this article proposes approach to automatically decompose and extract them. Specifically, we model such joint distributions as polynomials, estimated separately for all considered lag dependencies, then with PCA dimensionality reduction we find the dominant joint density distortion directions $f_v$. This way we get a few lag dependent features $a_i(\Delta t)$ describing separate dominating statistical dependencies of known contributions: $\rho_{\Delta t}(y,z)\approx \sum_{i=1}^r a_i(\Delta t)\, f_{v_i}(y,z)$. Such features complement Pearson correlation, extracting hidden more complex behavior, e.g. with asymmetry which might be related with direction of information transfer, extrema suggesting characteristic delays, or oscillatory behavior suggesting some periodicity. There is also discussed extension of Granger causality to such multi-feature joint density analysis, suggesting e.g. two separate causality waves. While this early article is initial fundamental research, in future it might help e.g. with understanding of cortex hidden dynamics, diagnosis of pathologies like epilepsy, determination of precise electrode position, or building brain-computer interface. | ['Jarek Duda'] | 2023-04-24 | null | null | null | null | ['dimensionality-reduction'] | ['methodology'] | [-9.45586413e-02 -9.99827832e-02 4.32626128e-01 -1.93615615e-01
-4.17985857e-01 -6.42325699e-01 5.99225402e-01 1.95786327e-01
-2.17328742e-01 1.20114017e+00 -1.13596104e-01 -2.62079448e-01
-9.49937761e-01 -5.80144346e-01 -5.40561318e-01 -1.26108587e+00
-1.04747033e+00 4.45449322e-01 3.40535827e-02 -1.94572359e-01
2.31866106e-01 5.59405625e-01 -1.55209887e+00 -5.41747808e-02
6.22268736e-01 9.04151797e-01 4.02140692e-02 4.94671792e-01
-2.72036977e-02 2.60498673e-01 -9.05286014e-01 -7.36425221e-02
-1.98534727e-01 -5.31040609e-01 -4.58048999e-01 -3.52244467e-01
-6.83812678e-01 3.22787702e-01 1.02196068e-01 1.08891058e+00
2.90886462e-01 -1.16547599e-01 1.21271372e+00 -1.32623255e+00
-1.21572219e-01 6.08632803e-01 -6.81620121e-01 5.79421461e-01
5.35850644e-01 -1.69606730e-02 6.04804158e-01 -5.95674038e-01
3.73147845e-01 6.96300089e-01 5.28597295e-01 -3.18847746e-02
-1.60792434e+00 -7.42701530e-01 -1.72006592e-01 3.06653023e-01
-1.60680723e+00 9.56309065e-02 9.09210801e-01 -6.86929882e-01
1.08008015e+00 3.11861455e-01 8.52522612e-01 1.06553626e+00
7.95470715e-01 3.01334232e-01 1.50134373e+00 -1.93046302e-01
1.02975592e-01 2.64632702e-01 4.94699866e-01 2.33338848e-01
2.95378357e-01 1.71370462e-01 -5.11634171e-01 -1.92081600e-01
6.04301393e-01 -1.49932519e-01 -6.94646299e-01 2.86079813e-02
-1.21020937e+00 3.21305960e-01 -1.06114455e-01 9.19005692e-01
-4.99915779e-01 -1.94754958e-01 9.45565552e-02 5.77100039e-01
1.97888747e-01 1.78436100e-01 -8.99386287e-01 -4.01560128e-01
-8.48985076e-01 2.10483089e-01 9.23101842e-01 7.41016984e-01
9.50481117e-01 -1.66425258e-01 8.20080116e-02 4.56906170e-01
1.28646538e-01 4.23370808e-01 3.70430470e-01 -3.74523312e-01
1.81905195e-01 3.46984953e-01 -1.02751695e-01 -7.14963078e-01
-8.67104053e-01 -4.23141241e-01 -1.12301230e+00 1.64661974e-01
7.67411709e-01 -2.75189042e-01 -2.53702849e-01 1.69706082e+00
-3.12363952e-01 -2.31819805e-02 -1.82623342e-01 5.24694681e-01
3.10956091e-01 6.26055658e-01 -2.19624236e-01 -8.52619946e-01
1.55228949e+00 2.46987581e-01 -8.51169109e-01 4.02246982e-01
2.19118729e-01 -7.14731038e-01 4.86073017e-01 7.30558097e-01
-1.07424092e+00 -1.76361501e-01 -6.32035255e-01 7.39082158e-01
-4.70761418e-01 -1.66307509e-01 5.11698484e-01 5.23781538e-01
-1.17979133e+00 9.36962068e-01 -9.40847516e-01 -4.13589180e-02
-6.46080822e-02 6.19408190e-01 -5.19948483e-01 3.64516586e-01
-1.18569922e+00 8.64242256e-01 3.27665582e-02 2.80597687e-01
-4.83797491e-01 -6.83598697e-01 -3.67567986e-01 -1.19292200e-01
-2.08725229e-01 -3.28319341e-01 5.37002683e-01 -6.25835955e-01
-1.11418283e+00 4.77980465e-01 -2.54282206e-01 -2.56065369e-01
2.98628986e-01 1.88958928e-01 -7.58755624e-01 -9.82705727e-02
1.24043450e-01 1.95008174e-01 8.64884198e-01 -1.20047498e+00
-7.38318190e-02 -7.49314666e-01 -5.51970541e-01 -2.90961832e-01
-2.48820242e-02 5.91778085e-02 3.24516475e-01 -4.26298261e-01
2.96811402e-01 -6.66405797e-01 7.86356106e-02 -7.31519401e-01
-3.95424783e-01 -1.76156476e-01 5.55686116e-01 -7.41043210e-01
1.34197021e+00 -1.85692692e+00 1.18172631e-01 6.66005015e-01
3.70796263e-01 -3.22382122e-01 3.23241323e-01 6.73806667e-01
-6.82887316e-01 -8.83522257e-02 -5.89431047e-01 2.02187479e-01
-1.54389158e-01 6.74877018e-02 -2.55116238e-03 8.90349984e-01
3.12971085e-01 4.81113672e-01 -5.56770265e-01 1.86642688e-02
-1.01864459e-02 9.20515001e-01 -1.72389030e-01 3.63310762e-02
2.00860247e-01 9.39226031e-01 -1.70496091e-01 3.51124406e-01
6.29939377e-01 1.31052034e-02 -3.81739847e-02 -2.04951003e-01
-5.61941445e-01 1.88210130e-01 -1.09931326e+00 1.18353558e+00
-3.78687903e-02 9.73994374e-01 4.58984375e-02 -1.57386446e+00
1.21663404e+00 5.45078456e-01 9.30897713e-01 -6.34457588e-01
3.53501618e-01 1.67240307e-01 4.83799309e-01 -5.51228821e-01
-1.37523711e-01 -2.52479941e-01 2.44276017e-01 3.81842226e-01
3.15239847e-01 1.15517572e-01 4.94708978e-02 -2.42163479e-01
1.48776209e+00 -1.22451618e-01 2.74052054e-01 -6.28587902e-01
5.61795592e-01 -4.19647634e-01 8.58483389e-02 2.68727899e-01
-1.20890819e-01 3.28717619e-01 1.30336607e+00 2.93555553e-03
-7.90930867e-01 -1.08844948e+00 -8.47963750e-01 2.70936996e-01
-1.02759534e-02 -3.04323763e-01 -6.42754674e-01 -3.63177322e-02
-1.57825559e-01 6.17524207e-01 -6.69349074e-01 -1.29479349e-01
-4.60950971e-01 -1.28749061e+00 3.61114740e-01 8.94441679e-02
2.12636679e-01 -1.21310544e+00 -7.67783880e-01 2.83540428e-01
-4.94189672e-02 -5.74258626e-01 2.89407343e-01 9.28654134e-01
-1.17832541e+00 -9.30042148e-01 -7.17270255e-01 -2.02124327e-01
4.08793420e-01 -3.29937518e-01 1.11072230e+00 -3.53314579e-01
-5.86138904e-01 4.91291583e-01 -2.43293807e-01 -2.94991910e-01
9.37730744e-02 -6.03864670e-01 2.56310195e-01 -1.47978440e-01
4.93857116e-01 -1.50929880e+00 -7.43956864e-01 3.17662805e-01
-6.75942004e-01 -7.92659938e-01 5.74282706e-01 6.83231235e-01
5.22804916e-01 4.95910078e-01 4.86252159e-01 -3.29539716e-01
8.92158806e-01 -7.97240794e-01 -4.92118925e-01 -1.04299709e-01
-5.77906847e-01 2.46257827e-01 6.25471711e-01 -4.83648211e-01
-6.22405529e-01 -3.61451775e-01 -6.86650872e-02 -3.42566848e-01
-6.48814321e-01 4.97567445e-01 -1.00612395e-01 2.91000873e-01
6.12935960e-01 5.78575671e-01 -9.91292857e-03 -4.26623374e-01
-3.37770060e-02 3.47410530e-01 5.28131723e-01 -5.37902474e-01
4.15615350e-01 3.07290167e-01 2.63911247e-01 -1.16582155e+00
2.82487214e-01 -3.53073984e-01 -5.88146031e-01 -1.44911930e-01
8.52476358e-01 -5.12166858e-01 -1.04169071e+00 2.49525279e-01
-1.43414235e+00 -7.25322869e-03 -1.29732862e-01 8.93647492e-01
-4.92175341e-01 5.20101935e-02 -3.88385236e-01 -1.21886814e+00
-6.67224154e-02 -9.89020467e-01 7.31232107e-01 7.64660165e-02
-6.12231016e-01 -1.16902900e+00 2.94851810e-01 -2.54609883e-01
1.52761787e-01 2.51610935e-01 1.14105785e+00 -6.90815568e-01
-3.17875355e-01 -4.26053762e-01 3.71126533e-02 1.55568704e-01
1.27459511e-01 1.27565220e-01 -8.66231799e-01 -1.05564974e-01
3.41379374e-01 5.15668988e-01 5.08758485e-01 8.48290682e-01
8.62638831e-01 -2.34911293e-01 -4.71722782e-01 3.74660164e-01
1.41077542e+00 7.99366713e-01 7.50462651e-01 -4.40702364e-02
6.94564506e-02 7.38944709e-01 -3.85276452e-02 7.89758444e-01
4.50026952e-02 4.94193941e-01 3.12026531e-01 2.59797454e-01
4.06182557e-01 4.52031493e-01 5.76657534e-01 7.49668121e-01
-7.52881408e-01 1.65473580e-01 -8.59615922e-01 4.33777481e-01
-1.32942855e+00 -1.06324220e+00 -7.25542665e-01 2.47843385e+00
4.99081492e-01 2.04535335e-01 3.00526440e-01 5.63327551e-01
6.99999154e-01 -2.09529757e-01 -1.79778516e-01 -2.53867239e-01
-2.74063617e-01 8.27285469e-01 2.88217217e-01 3.72201324e-01
-4.12556797e-01 3.77215184e-02 5.62153912e+00 5.20530164e-01
-1.15087450e+00 9.87079665e-02 3.05218995e-01 -2.30487790e-02
-4.65276122e-01 2.01770321e-01 -5.46399355e-01 8.67888510e-01
1.01486945e+00 -2.44638443e-01 3.36267769e-01 2.38600194e-01
3.46894145e-01 -6.03467584e-01 -1.19561517e+00 1.27136362e+00
-2.24706993e-01 -7.62323797e-01 -3.48798871e-01 4.54255313e-01
3.66068244e-01 -2.92546153e-01 -3.06793563e-02 -6.45994162e-03
-1.58593327e-01 -1.18849671e+00 5.06678760e-01 8.96192610e-01
7.29659855e-01 -8.38325322e-01 7.03834534e-01 5.19234359e-01
-1.26624715e+00 6.84703737e-02 -9.27930772e-02 -3.51978660e-01
2.39155337e-01 1.21282029e+00 -2.94869095e-01 7.77718425e-01
8.91774595e-01 6.64967000e-01 -1.59550548e-01 9.72937942e-01
-6.97719902e-02 5.34411907e-01 -4.09227073e-01 -2.56922513e-01
-2.12010160e-01 -7.09409952e-01 9.88985360e-01 1.16118836e+00
7.45588422e-01 4.63243425e-01 -6.90296173e-01 1.15095329e+00
7.39265144e-01 -1.55816853e-01 -9.74061906e-01 4.06642556e-02
2.85568148e-01 1.22480106e+00 -1.08579326e+00 1.65107578e-01
-3.00273687e-01 8.24701250e-01 -1.22672185e-01 5.03916085e-01
-8.26472998e-01 -4.83515710e-01 7.91061580e-01 3.38397115e-01
6.49880841e-02 -3.77816468e-01 -4.19535816e-01 -1.04363167e+00
3.11428040e-01 -2.78400272e-01 3.41268908e-03 -5.96925259e-01
-1.31671429e+00 9.31356370e-01 3.35475475e-01 -1.37616825e+00
-4.62263077e-01 -6.16987288e-01 -8.09247375e-01 1.16100144e+00
-1.04815221e+00 -2.94305086e-01 1.53615177e-01 1.16595316e+00
-2.64124349e-02 -2.34758869e-01 9.11853433e-01 1.98756725e-01
-3.81488383e-01 1.77560836e-01 2.68024821e-02 -2.90129781e-01
3.54466230e-01 -1.29442453e+00 -4.30351198e-01 4.35953975e-01
7.22515061e-02 7.11950958e-01 1.12335515e+00 -5.04515350e-01
-1.13951969e+00 -2.85828084e-01 8.85071456e-01 -4.03129816e-01
9.75550771e-01 -2.22604856e-01 -9.79434848e-01 3.46667171e-01
4.79022145e-01 -1.41142309e-01 7.51591325e-01 9.64830890e-02
1.25975877e-01 -1.09661549e-01 -9.37520206e-01 2.39376664e-01
7.07069278e-01 -3.90224874e-01 -5.39817691e-01 2.11529121e-01
4.08475101e-03 2.19331160e-01 -1.13963068e+00 3.12799931e-01
5.70954800e-01 -1.62535691e+00 6.12911642e-01 -4.16420996e-02
1.86813667e-01 -1.73108816e-01 -1.01731764e-02 -1.41034734e+00
-2.32391506e-01 -8.14418733e-01 2.61377573e-01 1.00616801e+00
5.44239104e-01 -1.04721653e+00 3.41395646e-01 3.68231893e-01
2.10674275e-02 -7.06122816e-01 -1.32495141e+00 -8.20423186e-01
1.10921800e-01 -7.32494473e-01 3.06721538e-01 9.96213675e-01
5.89916766e-01 2.04119027e-01 7.42036551e-02 2.15092734e-01
6.59620345e-01 -2.27783084e-01 1.94416001e-01 -1.42247868e+00
-3.85867566e-01 -7.19411433e-01 -7.40510106e-01 -5.07585406e-01
2.43264791e-02 -7.37300873e-01 -2.13684827e-01 -1.12987459e+00
-5.61884865e-02 -4.29418474e-01 -4.01830792e-01 1.39086202e-01
3.91034871e-01 -2.29044184e-01 -3.68395954e-01 1.43451393e-01
2.00953454e-01 3.63330990e-01 9.40829217e-01 2.05750212e-01
-4.29310292e-01 2.46497557e-01 -1.41138092e-01 7.68980920e-01
5.63262761e-01 -5.11447132e-01 -3.88156921e-01 1.36288524e-01
3.81367683e-01 8.18185687e-01 4.53068584e-01 -1.14109743e+00
3.22196454e-01 2.87082970e-01 5.08971453e-01 -5.83005667e-01
3.09983075e-01 -8.59432995e-01 6.39721155e-01 2.03889698e-01
3.02336574e-01 4.32038099e-01 1.30317837e-01 4.65821564e-01
-4.63397920e-01 -2.69560050e-02 4.27239776e-01 -9.90720093e-02
-1.77571401e-02 6.89842999e-02 -7.19740450e-01 -4.33474004e-01
1.14322937e+00 -3.99143368e-01 -5.92630589e-04 -4.17877197e-01
-1.13271999e+00 -1.93059444e-01 -1.55756444e-01 -1.16326347e-01
4.77762312e-01 -1.07828057e+00 -4.78644073e-01 3.80894929e-01
-2.74277121e-01 -1.11393802e-01 3.89709055e-01 1.69991183e+00
2.23264727e-03 4.22232747e-01 -4.35676277e-01 -8.86718512e-01
-9.87801075e-01 2.41182089e-01 2.61684209e-01 -1.61469564e-01
-4.55170244e-01 8.92128527e-01 1.28140794e-02 2.17290968e-01
-5.09389304e-02 -5.75765550e-01 -5.84068716e-01 5.87319374e-01
3.41564715e-01 4.09856200e-01 2.39462242e-01 -5.91223598e-01
-6.46803677e-01 7.89593041e-01 2.33522117e-01 -3.94432575e-01
1.50800574e+00 -3.84828746e-02 -7.48728216e-01 8.42016876e-01
1.40209019e+00 1.22295553e-03 -9.96334255e-01 5.08180380e-01
1.03502899e-01 1.42851518e-02 -4.96104419e-01 -7.09467530e-01
-9.66972709e-01 9.47418809e-01 6.16763711e-01 7.59242773e-01
1.34742868e+00 2.38301575e-01 4.23920341e-02 -2.02080864e-03
6.60845518e-01 -5.77464283e-01 -1.82639897e-01 3.96381915e-01
9.43580031e-01 -5.66681981e-01 -2.91793376e-01 -1.63840264e-01
-3.21464270e-01 1.44835901e+00 -4.31199409e-02 -4.36238945e-01
1.31421244e+00 6.97888076e-01 -4.09992903e-01 -3.99999976e-01
-8.14580083e-01 -8.47816020e-02 1.63799375e-01 6.13147020e-01
8.38619232e-01 3.13604534e-01 -8.27386022e-01 9.25843775e-01
-4.30271477e-01 -2.00069606e-01 3.83906096e-01 5.85467935e-01
-1.84100911e-01 -1.26515114e+00 -6.48575187e-01 6.01041853e-01
-3.74744564e-01 -8.26268420e-02 1.78167745e-02 9.01899576e-01
3.99505913e-01 8.29002500e-01 3.38547766e-01 -5.30228615e-01
3.62642497e-01 3.78904313e-01 7.78489709e-01 -9.96151101e-03
-3.53563607e-01 4.89916593e-01 -3.00702840e-01 -5.28099716e-01
-3.95349681e-01 -1.00912952e+00 -1.36943638e+00 -6.54347315e-02
-1.38136223e-01 3.20137262e-01 8.43247473e-01 1.10030496e+00
-1.31807059e-01 5.55579603e-01 3.95948380e-01 -9.70428050e-01
-3.00546344e-02 -9.78666246e-01 -1.33346462e+00 1.44745141e-01
5.35942614e-01 -8.65253329e-01 -7.18221962e-01 1.02848232e-01] | [12.905491828918457, 3.454606056213379] |
b00c638e-bd22-4a1c-86a0-7938a5727707 | alignment-uniformity-aware-representation | 2203.15381 | null | https://arxiv.org/abs/2203.15381v1 | https://arxiv.org/pdf/2203.15381v1.pdf | Alignment-Uniformity aware Representation Learning for Zero-shot Video Classification | Most methods tackle zero-shot video classification by aligning visual-semantic representations within seen classes, which limits generalization to unseen classes. To enhance model generalizability, this paper presents an end-to-end framework that preserves alignment and uniformity properties for representations on both seen and unseen classes. Specifically, we formulate a supervised contrastive loss to simultaneously align visual-semantic features (i.e., alignment) and encourage the learned features to distribute uniformly (i.e., uniformity). Unlike existing methods that only consider the alignment, we propose uniformity to preserve maximal-info of existing features, which improves the probability that unobserved features fall around observed data. Further, we synthesize features of unseen classes by proposing a class generator that interpolates and extrapolates the features of seen classes. Besides, we introduce two metrics, closeness and dispersion, to quantify the two properties and serve as new measurements of model generalizability. Experiments show that our method significantly outperforms SoTA by relative improvements of 28.1% on UCF101 and 27.0% on HMDB51. Code is available. | ['Mao Zheng', 'Kaili Zhao', 'Shi Pu'] | 2022-03-29 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Pu_Alignment-Uniformity_Aware_Representation_Learning_for_Zero-Shot_Video_Classification_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Pu_Alignment-Uniformity_Aware_Representation_Learning_for_Zero-Shot_Video_Classification_CVPR_2022_paper.pdf | cvpr-2022-1 | ['zero-shot-action-recognition'] | ['computer-vision'] | [ 2.36678854e-01 -5.11787012e-02 -4.88778114e-01 -6.19153142e-01
-7.11039782e-01 -5.10429382e-01 6.49525106e-01 4.38755229e-02
-2.29117692e-01 5.23241460e-01 3.33550185e-01 1.65599078e-01
4.71156724e-02 -5.66677392e-01 -8.87065530e-01 -6.89178467e-01
9.99905169e-02 8.38893279e-02 1.64120361e-01 2.13468954e-01
-1.94609314e-02 2.66169906e-01 -1.75397718e+00 5.74382126e-01
7.72366822e-01 1.08663881e+00 2.48458564e-01 3.31467479e-01
1.78286850e-01 9.06471789e-01 -4.73011404e-01 -2.95124948e-01
3.08771551e-01 -6.20880306e-01 -6.93854094e-01 5.97034812e-01
7.23416150e-01 -4.20439273e-01 -6.61041141e-01 9.86722350e-01
2.32450023e-01 4.20518279e-01 1.00087273e+00 -1.63988602e+00
-1.05046070e+00 2.59229392e-01 -5.32637298e-01 2.37385273e-01
1.64568618e-01 1.01541281e-01 1.21341836e+00 -8.53831649e-01
5.70929408e-01 1.16211271e+00 3.69279891e-01 7.50702441e-01
-1.24256361e+00 -5.77989757e-01 3.60608637e-01 3.48122120e-01
-1.39265203e+00 -5.35095930e-01 6.75627410e-01 -5.27965665e-01
5.76690674e-01 2.99824268e-01 3.46012980e-01 1.33407176e+00
-1.54879257e-01 1.07925558e+00 7.39627898e-01 -3.68677646e-01
3.57011050e-01 3.10893625e-01 2.18405828e-01 5.29705226e-01
2.73096651e-01 4.40031961e-02 -4.47754055e-01 -1.40781969e-01
7.06259072e-01 4.35555130e-01 -4.26084250e-01 -9.35373306e-01
-1.05748284e+00 9.39847112e-01 5.96353352e-01 1.79096237e-02
-2.28360444e-01 2.09227968e-02 3.50454181e-01 1.20890722e-01
3.51738483e-01 1.34237945e-01 -1.53070524e-01 6.85343966e-02
-7.71477759e-01 2.06283301e-01 1.98060736e-01 1.29364276e+00
8.96424770e-01 -4.82732095e-02 -4.74170595e-01 1.10205472e+00
1.29725590e-01 5.19627154e-01 7.03129470e-01 -8.44602287e-01
4.17905986e-01 4.64369476e-01 -3.98455188e-02 -9.34678912e-01
-8.50576982e-02 -4.47808892e-01 -6.84358299e-01 -1.46804377e-01
2.24459544e-01 1.29340425e-01 -1.11997545e+00 2.13653994e+00
8.67051855e-02 4.45430070e-01 2.26340145e-01 1.01245236e+00
8.61128807e-01 6.30157590e-01 1.97100177e-01 -4.71968809e-03
1.05375469e+00 -1.15954125e+00 -5.85449934e-01 -2.98259884e-01
5.62786162e-01 -4.18523729e-01 1.20828795e+00 -5.99478744e-02
-8.97800148e-01 -6.33062959e-01 -9.30838346e-01 -5.67802563e-02
-1.32085025e-01 2.88720634e-02 3.61672252e-01 4.07899559e-01
-7.10998833e-01 5.12000680e-01 -8.77133369e-01 -3.71649057e-01
7.26115108e-01 -9.74785239e-02 -3.72066796e-01 -2.74023324e-01
-9.00145710e-01 4.03377771e-01 3.43096107e-01 -3.87919962e-01
-1.17565775e+00 -7.62908816e-01 -1.03211355e+00 2.38197565e-01
4.31016445e-01 -7.44900703e-01 1.07206249e+00 -1.11621594e+00
-9.66563761e-01 8.24027658e-01 -3.36194277e-01 -4.61514860e-01
4.59255040e-01 -2.19220310e-01 -3.60053092e-01 2.36657009e-01
2.19300106e-01 8.67797315e-01 8.27819407e-01 -1.54172003e+00
-7.41390049e-01 -2.10659847e-01 1.21546343e-01 2.70960331e-01
-6.94345534e-01 -2.81751603e-01 -5.87852120e-01 -8.30799162e-01
2.25984544e-01 -8.37230921e-01 -5.24691232e-02 3.02470177e-01
-3.85803521e-01 -6.37582466e-02 7.51382351e-01 -3.77226472e-01
1.15405440e+00 -2.49838686e+00 6.93568960e-02 1.20391555e-01
3.29773635e-01 -3.79174016e-02 -4.06584322e-01 1.94848120e-01
-6.18285835e-02 -1.94953158e-02 -3.03775072e-01 -3.85117024e-01
6.78552315e-02 3.73212218e-01 -4.42916185e-01 4.09470737e-01
2.65383035e-01 7.67812967e-01 -9.32615101e-01 -2.23559901e-01
2.06134483e-01 4.15188909e-01 -8.38395953e-01 3.16663802e-01
-3.72979380e-02 1.44542396e-01 -3.55223924e-01 4.71871108e-01
6.45253003e-01 -5.29121161e-01 7.46629909e-02 -9.61839855e-02
3.97867560e-01 5.86141087e-02 -9.83310759e-01 1.50009716e+00
-4.88232583e-01 5.64691126e-01 -5.46820402e-01 -1.02771318e+00
9.86655593e-01 1.36980325e-01 2.30541572e-01 -5.61132550e-01
1.34769127e-01 -8.67069587e-02 -2.32161149e-01 -4.06236380e-01
3.79655302e-01 -1.24353513e-01 2.47760285e-02 1.33743390e-01
2.50445217e-01 3.79978091e-01 -8.12996365e-03 2.63228655e-01
9.08323824e-01 -3.39579657e-02 3.24203521e-01 -2.30342314e-01
1.90747574e-01 -3.38844568e-01 6.37616277e-01 8.15303028e-01
-2.82003969e-01 1.04867852e+00 5.55899680e-01 -4.36399579e-01
-1.07859731e+00 -1.44204426e+00 -1.43909693e-01 1.12245393e+00
4.43773091e-01 -4.07192647e-01 -6.12975359e-01 -1.02699625e+00
1.56652644e-01 8.98486555e-01 -9.19697464e-01 -5.05809009e-01
-6.86622411e-02 -4.71013457e-01 1.34136125e-01 7.99347222e-01
2.88749635e-01 -8.46645534e-01 -4.73968714e-01 -9.08990279e-02
-3.27399731e-01 -1.09446073e+00 -4.99815196e-01 -1.32829440e-03
-6.37169242e-01 -9.67046022e-01 -7.98927367e-01 -8.62492323e-01
7.86924779e-01 7.24950850e-01 9.60885525e-01 -3.39640602e-02
-2.20110163e-01 3.05168152e-01 -6.09907150e-01 -7.45498538e-02
-9.13991556e-02 -4.77239303e-02 5.47122862e-03 3.52717221e-01
4.55818176e-01 -3.34834635e-01 -5.03420830e-01 5.18796861e-01
-8.71033549e-01 6.03488423e-02 3.22423160e-01 1.01219881e+00
7.60006726e-01 -2.11962521e-01 4.94363189e-01 -7.73378670e-01
1.33310050e-01 -8.03484678e-01 -1.60974577e-01 4.01189446e-01
-3.24485511e-01 5.54803461e-02 6.06839001e-01 -6.00345671e-01
-8.57459545e-01 -4.15231846e-02 2.95334816e-01 -9.95294631e-01
-3.18959087e-01 1.78865448e-01 -4.14323211e-01 3.60013664e-01
6.08714640e-01 2.82457471e-01 -3.27029862e-02 -3.17562759e-01
4.93765265e-01 6.96492493e-01 5.41037977e-01 -5.43816566e-01
6.74075305e-01 6.43883765e-01 -4.19701368e-01 -8.66480649e-01
-1.11599576e+00 -6.22203231e-01 -4.13382649e-01 -2.33776961e-02
7.42067397e-01 -1.09964502e+00 -1.72820479e-01 2.50795960e-01
-7.35882461e-01 -2.45274454e-01 -4.33719218e-01 5.58402479e-01
-8.51199448e-01 3.77901822e-01 -4.48307335e-01 -6.10918999e-01
1.52740017e-01 -1.01292574e+00 1.15379059e+00 2.16615662e-01
-2.56924480e-01 -8.94463539e-01 -1.01539232e-01 2.05000415e-01
1.53284565e-01 2.82115400e-01 8.23874414e-01 -7.27073193e-01
-4.45835978e-01 -2.09154144e-01 -4.22010213e-01 5.04304171e-01
2.44464740e-01 -6.49937838e-02 -1.19073331e+00 -5.50274372e-01
-1.79693371e-01 -4.90306050e-01 1.15748596e+00 4.18583721e-01
1.52387464e+00 -4.83790845e-01 -4.12807375e-01 6.14829421e-01
1.18966293e+00 1.62583217e-01 6.93229973e-01 2.70640641e-01
7.12356925e-01 5.89405894e-01 5.91462553e-01 6.92794502e-01
3.01691890e-01 7.94069350e-01 3.47763687e-01 -1.15679484e-03
-1.73629731e-01 -5.18837512e-01 2.20168740e-01 4.40431535e-01
2.03142494e-01 -5.58993280e-01 -6.07460916e-01 7.30616510e-01
-1.83371532e+00 -1.08472228e+00 2.56509304e-01 2.26631093e+00
5.46163380e-01 1.35571674e-01 2.68386543e-01 6.90897740e-03
9.48051631e-01 1.98065713e-01 -5.10054588e-01 2.29415949e-02
-1.04199097e-01 -7.73855671e-02 3.07631016e-01 3.74191642e-01
-1.22964156e+00 8.19597960e-01 5.42563915e+00 9.51811016e-01
-8.25249314e-01 1.65618304e-02 6.25259042e-01 -2.83356786e-01
-4.25153136e-01 -6.90018237e-02 -6.70763850e-01 6.76295877e-01
5.96262336e-01 -2.81503946e-01 2.37521306e-01 1.09548140e+00
-1.20984456e-02 3.84170800e-01 -1.33514977e+00 9.63563800e-01
2.99368560e-01 -1.17451560e+00 5.80838859e-01 5.93363419e-02
9.07038629e-01 -2.57235587e-01 2.57130861e-01 3.47520709e-01
1.15330271e-01 -7.24661767e-01 9.03689265e-01 6.06030941e-01
6.97802424e-01 -8.72195363e-01 6.06547475e-01 1.67103663e-01
-1.07067990e+00 -2.45347947e-01 -6.35518551e-01 9.53948200e-02
3.85282608e-03 3.11966032e-01 -6.04480207e-01 4.32482988e-01
5.89303434e-01 1.03408778e+00 -6.14366829e-01 1.15276253e+00
-1.53217345e-01 5.30847490e-01 1.25375800e-02 1.96390197e-01
2.84075052e-01 1.59342475e-02 3.35000038e-01 1.03271925e+00
2.92786181e-01 -1.02339223e-01 4.24168974e-01 7.84525275e-01
-1.40091300e-01 1.81836844e-03 -5.41408837e-01 6.12607747e-02
6.48555338e-01 9.41451490e-01 -4.31786180e-01 -6.03069186e-01
-5.73003948e-01 1.14436197e+00 4.28588569e-01 5.31313658e-01
-9.89441037e-01 -3.92589539e-01 8.12878191e-01 2.22954899e-01
5.29243171e-01 1.98929489e-01 -7.86795095e-02 -1.32968760e+00
2.08543897e-01 -7.00763464e-01 5.95862269e-01 -5.84878683e-01
-1.65015471e+00 6.92214668e-01 6.89764619e-02 -1.68110585e+00
-2.01345652e-01 -5.33606946e-01 -5.01583934e-01 4.45881844e-01
-1.29829121e+00 -1.07683575e+00 -5.74496567e-01 4.87169683e-01
6.07247829e-01 -1.94521159e-01 5.74898958e-01 2.57954627e-01
-5.34392715e-01 1.01051712e+00 3.38227481e-01 1.79360896e-01
6.54238343e-01 -9.60206747e-01 3.24290842e-01 6.47869527e-01
3.68136406e-01 5.69268584e-01 5.26743233e-01 -3.77285928e-01
-7.73084581e-01 -1.40204453e+00 7.54320562e-01 -3.37910831e-01
5.54582715e-01 -4.60781276e-01 -1.16615176e+00 8.60638201e-01
-2.08236992e-01 4.03597504e-01 7.11409450e-01 -6.49879547e-03
-8.40211630e-01 -4.07995246e-02 -1.07874560e+00 6.26269758e-01
1.30528951e+00 -5.76295853e-01 -5.57382941e-01 3.45509201e-01
6.14456594e-01 -1.29354671e-01 -6.03376687e-01 4.45030212e-01
5.59738278e-01 -9.89216685e-01 9.00996983e-01 -9.84020412e-01
5.03422141e-01 -2.90469706e-01 -6.35994017e-01 -1.31227708e+00
-5.84336042e-01 -6.77055866e-02 -1.79107770e-01 1.19664145e+00
4.00225967e-01 -5.31790197e-01 7.47910559e-01 4.39829469e-01
-1.90018177e-01 -6.39190614e-01 -7.46101201e-01 -1.27342260e+00
7.43430704e-02 -1.70123681e-01 5.06784856e-01 1.10000968e+00
-4.47770171e-02 2.62922078e-01 -5.04265189e-01 1.90923482e-01
6.54100120e-01 1.55450940e-01 7.10976243e-01 -9.11263347e-01
-4.22988474e-01 -3.62470537e-01 -7.67912090e-01 -1.24314868e+00
2.69850403e-01 -9.94989097e-01 1.45680495e-02 -1.31889498e+00
8.48934889e-01 -3.53573650e-01 -6.32708251e-01 4.66439426e-01
-3.85292441e-01 2.54205525e-01 2.76819021e-01 3.02902579e-01
-8.86455178e-01 9.57715034e-01 1.11557913e+00 -4.56304997e-02
-3.55879627e-02 -1.45368561e-01 -7.05964804e-01 6.37818456e-01
7.05036998e-01 -4.24240857e-01 -6.11850381e-01 -4.64324325e-01
-5.38027048e-01 -1.69012740e-01 6.06245995e-01 -1.23396695e+00
-2.06698447e-01 -1.83817104e-01 5.43483913e-01 -3.36167276e-01
4.08318102e-01 -7.10427046e-01 -3.03728189e-02 2.63253659e-01
-7.86387622e-01 -2.76199371e-01 -4.63240780e-02 8.94742370e-01
-2.51507223e-01 -1.20503247e-01 9.68501687e-01 1.73359841e-01
-6.99221790e-01 5.37280440e-01 -4.34737429e-02 2.79686213e-01
1.31582487e+00 -2.15949327e-01 -3.85335207e-01 -5.92585504e-01
-7.28288710e-01 2.18710706e-01 6.62417412e-01 7.34568298e-01
6.94321215e-01 -1.70212185e+00 -7.42623985e-01 3.56400132e-01
6.49659872e-01 -3.13371837e-01 4.75921869e-01 4.15008724e-01
-1.35932192e-01 2.01279461e-01 -2.29418680e-01 -8.10566962e-01
-1.17777693e+00 7.97332823e-01 1.17540546e-01 2.13811532e-01
-5.78159928e-01 7.99964845e-01 8.07498276e-01 -2.78421700e-01
3.79661411e-01 -2.21170112e-01 -4.40225117e-02 -6.30080402e-02
7.06080198e-01 2.57485241e-01 -1.71006173e-01 -6.79598451e-01
-3.71324152e-01 3.76316696e-01 -3.91324162e-01 1.62258044e-01
1.13141227e+00 -1.27314493e-01 4.36116874e-01 4.61335063e-01
1.59213996e+00 -2.44976118e-01 -1.73470819e+00 -4.41734105e-01
-1.29857764e-01 -8.26101482e-01 -1.99853569e-01 -5.29042006e-01
-1.06583095e+00 9.36262071e-01 5.93672097e-01 1.08007729e-01
1.02122104e+00 3.47591639e-01 5.87689281e-01 2.64739156e-01
1.67353481e-01 -7.46595442e-01 3.03112686e-01 3.15642357e-01
7.74672091e-01 -1.16568387e+00 -2.46044055e-01 -4.92855251e-01
-1.04933190e+00 8.64200771e-01 7.50880361e-01 -1.48847461e-01
3.73306662e-01 -1.33507743e-01 -4.71988320e-02 8.21161494e-02
-7.37758577e-01 -2.70023644e-01 4.72897291e-01 6.17929399e-01
2.01322511e-01 1.51197359e-01 5.42492010e-02 6.80379212e-01
3.26047875e-02 -1.48216993e-01 3.57410908e-01 8.77762794e-01
-5.83263993e-01 -5.80818653e-01 -6.44476786e-02 5.00741541e-01
-1.12709865e-01 -3.51160066e-03 -1.69949993e-01 8.19201648e-01
-9.54318675e-05 9.01932597e-01 5.08992434e-01 -5.65871835e-01
3.07596713e-01 -1.22165389e-01 3.48009050e-01 -7.31563270e-01
1.80384144e-01 2.48258058e-02 -9.38791409e-02 -3.39441568e-01
-1.81137174e-01 -6.29835367e-01 -1.07758415e+00 -1.24420017e-01
-3.01893413e-01 4.20371331e-02 1.86291292e-01 7.67282486e-01
4.16268975e-01 4.41428423e-01 9.83622909e-01 -6.47367537e-01
-8.15134943e-01 -7.76086211e-01 -6.07793868e-01 9.20397937e-01
3.58839691e-01 -1.02419543e+00 -6.58706665e-01 1.94506824e-01] | [9.773507118225098, 2.3816916942596436] |
062862b6-1c22-4746-bae6-043cf186dcd9 | aggression-identification-and-multi-lingual | null | null | https://aclanthology.org/W18-4409 | https://aclanthology.org/W18-4409.pdf | Aggression Identification and Multi Lingual Word Embeddings | The system presented here took part in the 2018 Trolling, Aggression and Cyberbullying shared task (Forest and Trees team) and uses a Gated Recurrent Neural Network architecture (Cho et al., 2014) in an attempt to assess whether combining pre-trained English and Hindi fastText (Mikolov et al., 2018) word embeddings as a representation of the sequence input would improve classification performance. The motivation for this comes from the fact that the shared task data for English contained many Hindi tokens and therefore some users might be doing code-switching: the alternation between two or more languages in communication. To test this hypothesis, we also aligned Hindi and English vectors using pre-computed SVD matrices that pulls representations from different languages into a common space (Smith et al., 2017). Two conditions were tested: (i) one with standard pre-trained fastText word embeddings where each Hindi word is treated as an OOV token, and (ii) another where word embeddings for Hindi and English are loaded in a common vector space, so Hindi tokens can be assigned a meaningful representation. We submitted the second (i.e., multilingual) system and obtained the scores of 0.531 weighted F1 for the EN-FB dataset and 0.438 weighted F1 for the EN-TW dataset. | ['Efstathios Charitos', 'Thiago Galery', 'Ye Tian'] | 2018-08-01 | null | null | null | coling-2018-8 | ['aggression-identification'] | ['natural-language-processing'] | [ 1.97539292e-02 1.57009855e-01 -1.60983011e-01 -2.74424106e-01
-5.37899613e-01 -4.52396840e-01 7.35661268e-01 3.35105687e-01
-9.40978765e-01 4.98994410e-01 7.38047302e-01 -6.94902539e-01
-1.83011591e-02 -5.36852181e-01 -4.80894059e-01 -5.09053826e-01
-4.01442358e-03 3.79420161e-01 -1.04089968e-01 -3.65110308e-01
1.04232416e-01 2.38701284e-01 -1.35125315e+00 4.15465236e-01
7.70313382e-01 2.04619721e-01 4.07300480e-02 8.73513520e-01
-1.57288257e-02 7.66931772e-01 -4.89266068e-01 -7.69643605e-01
9.11970958e-02 -4.64821607e-01 -1.10669124e+00 -4.95362341e-01
5.71565688e-01 -1.81941658e-01 -5.34293950e-01 8.08807790e-01
5.77744663e-01 3.39239895e-01 7.29938149e-01 -9.84525502e-01
-8.55225146e-01 1.06313813e+00 -2.77778834e-01 2.49164268e-01
4.70190257e-01 2.90661186e-01 1.22688532e+00 -1.01532161e+00
1.08687377e+00 1.18411112e+00 6.40081704e-01 5.12578189e-01
-1.54505420e+00 -8.67557347e-01 -1.14774182e-01 4.40927207e-01
-1.33815265e+00 -4.49176759e-01 5.02918899e-01 -6.53288126e-01
1.26124358e+00 1.58685461e-01 8.66970420e-01 1.66604638e+00
2.77741134e-01 5.54123342e-01 1.09795260e+00 -4.95919019e-01
-1.72122687e-01 3.41450483e-01 3.80756497e-01 2.55439401e-01
1.86937764e-01 1.30790994e-01 -6.61496580e-01 -2.51321584e-01
1.41152963e-01 -3.30859095e-01 1.24044150e-01 7.59722432e-03
-9.39620256e-01 1.34882283e+00 2.05443472e-01 7.99602687e-01
-2.31449023e-01 -1.14350714e-01 8.60318303e-01 4.80212241e-01
2.69305378e-01 5.15851855e-01 -2.04280213e-01 -5.56832373e-01
-7.41334677e-01 2.10690752e-01 5.07254660e-01 2.42330790e-01
5.52847087e-01 9.74157155e-02 -1.14513911e-01 1.30911160e+00
3.08358788e-01 4.83329371e-02 1.11957407e+00 -4.74457383e-01
6.11058414e-01 1.73171863e-01 -4.82162595e-01 -8.86308491e-01
-4.08123553e-01 2.33713444e-02 -3.18954855e-01 6.98592812e-02
5.91053367e-01 -2.84855902e-01 -7.88546026e-01 1.95160711e+00
2.11045921e-01 -2.59375796e-02 1.66157082e-01 6.32684231e-01
7.65296936e-01 6.35271788e-01 3.06595832e-01 -5.26776537e-02
1.21071148e+00 -7.11494684e-01 -6.43176317e-01 -3.14455569e-01
1.19283044e+00 -9.62636292e-01 1.17277575e+00 2.61196524e-01
-7.40416706e-01 -5.98071754e-01 -1.03700304e+00 -2.87283659e-01
-6.16172433e-01 -2.49830157e-01 4.55077797e-01 8.74109328e-01
-8.61574471e-01 5.16456842e-01 -5.74379325e-01 -8.08262408e-01
6.74224598e-03 4.79499958e-02 -6.69797719e-01 1.94437020e-02
-1.50239444e+00 1.28244674e+00 4.50093567e-01 -1.57460764e-01
-5.30156493e-01 -4.92026687e-01 -1.01911283e+00 -2.42698580e-01
-7.26035014e-02 3.27108800e-02 7.94852436e-01 -8.19036782e-01
-1.25979006e+00 1.22153139e+00 2.78989170e-02 -2.45589212e-01
2.66694546e-01 -2.55205154e-01 -6.88688517e-01 -1.58170909e-01
3.69880050e-01 3.20213437e-01 6.55058801e-01 -7.56002367e-01
-4.38694656e-01 -6.21631205e-01 -1.16681352e-01 1.00452274e-01
-5.30753672e-01 3.04910153e-01 1.14515536e-01 -7.41907120e-01
-1.97724298e-01 -1.14636230e+00 2.18009785e-01 -4.01312739e-01
-3.88542593e-01 -5.34262598e-01 4.48920220e-01 -1.01147008e+00
1.45807934e+00 -2.30336261e+00 1.87958106e-01 4.60960239e-01
2.69221574e-01 4.59641606e-01 -4.44234312e-01 8.51868212e-01
-6.57276392e-01 4.07482594e-01 3.08651794e-02 -1.61001012e-01
-2.99497787e-02 4.02913988e-01 -4.33171988e-02 6.54523551e-01
1.15674928e-01 5.71334720e-01 -7.10259855e-01 -2.33074814e-01
2.70328689e-02 4.69291896e-01 -8.41155350e-01 8.76975730e-02
6.86598063e-01 2.94688106e-01 2.96519399e-01 1.03751086e-02
3.57824773e-01 6.98510826e-01 4.26054299e-01 1.25251353e-01
-4.36428159e-01 7.12963104e-01 -1.19577467e+00 1.50264788e+00
-7.69968927e-01 9.39136386e-01 -9.35029313e-02 -7.86929905e-01
7.59685993e-01 3.49881470e-01 2.98838973e-01 -7.13512719e-01
2.88119286e-01 2.98910677e-01 6.75700307e-01 -6.48377717e-01
7.39251375e-01 -2.89009839e-01 -2.70639807e-01 6.09465539e-01
3.93533558e-01 -3.01708281e-03 2.91600734e-01 2.60432810e-01
1.09548271e+00 -3.16042863e-02 1.90960243e-01 -7.79537186e-02
2.12033331e-01 -3.84708315e-01 5.87438762e-01 6.93947554e-01
-2.85771698e-01 5.31876028e-01 7.04425752e-01 -3.08075637e-01
-9.67363477e-01 -1.13280761e+00 -4.58701283e-01 1.50568938e+00
-6.28665566e-01 -9.39141095e-01 -5.23852468e-01 -4.42104816e-01
-9.49697644e-02 1.21088827e+00 -8.75319958e-01 -5.14835656e-01
-5.44851959e-01 -4.62650865e-01 9.84398961e-01 2.68948853e-01
-1.98468626e-01 -1.08053529e+00 -4.69986230e-01 2.49153048e-01
-1.51866868e-01 -7.29063094e-01 -3.60465527e-01 4.31652874e-01
-2.79819101e-01 -9.50381517e-01 -4.29265350e-01 -6.93425536e-01
1.29567474e-01 -1.27807483e-01 5.88221252e-01 8.04946646e-02
-3.38341355e-01 1.78435370e-01 -6.75907314e-01 -1.58550620e-01
-5.64887464e-01 2.61180848e-01 3.56512487e-01 5.32309944e-03
5.87817013e-01 -4.17895019e-01 4.61029261e-02 -1.94123819e-01
-9.76410449e-01 -2.59861797e-01 2.70391077e-01 1.14100182e+00
-4.96148244e-02 -4.75045621e-01 1.83298871e-01 -9.47073221e-01
9.07313347e-01 -7.33157694e-01 1.24190398e-01 -9.28262770e-02
-4.13760096e-01 -6.50622025e-02 5.05470216e-01 -6.44318581e-01
-4.56960827e-01 -3.14017713e-01 -7.33757019e-01 -1.18260704e-01
-1.85321435e-01 7.21276641e-01 1.72910970e-02 3.33246469e-01
7.72110283e-01 -2.04381999e-02 -1.48984129e-02 -4.82012659e-01
4.65912521e-01 9.40677583e-01 9.34626833e-02 -6.38781190e-01
8.81620407e-01 -3.79776895e-01 -6.60543621e-01 -1.20522332e+00
-3.55809093e-01 -2.55885720e-01 -8.45294118e-01 -1.18490949e-01
1.35161269e+00 -5.77833652e-01 -4.92250741e-01 2.80885726e-01
-1.25992048e+00 -3.75521302e-01 -2.49282539e-01 8.02114964e-01
-2.93069214e-01 2.50356376e-01 -6.64320230e-01 -7.50574648e-01
5.61992414e-02 -1.24673891e+00 2.83605099e-01 -1.75613731e-01
-9.55501378e-01 -8.63930225e-01 5.35638452e-01 2.54110426e-01
3.19006681e-01 5.60584068e-02 1.29245043e+00 -1.28568733e+00
6.84635460e-01 -1.48777917e-01 -1.55417910e-02 4.14065778e-01
-6.68979436e-02 2.14982510e-01 -8.97432804e-01 -3.20686251e-01
-4.57750261e-02 -6.03751123e-01 6.69969261e-01 -2.73078501e-01
7.16319144e-01 -5.10800064e-01 1.56574428e-01 3.14400345e-01
1.07447016e+00 1.86170727e-01 6.21948361e-01 4.64017272e-01
8.03187728e-01 7.77460515e-01 5.76897413e-02 2.48215497e-01
3.02418798e-01 8.76584053e-01 -9.20170844e-02 1.79479495e-01
-4.71666977e-02 -6.64012790e-01 7.12711632e-01 1.17005575e+00
1.26052991e-01 1.55358687e-01 -1.12670255e+00 8.07568908e-01
-1.45932221e+00 -1.04723740e+00 -4.36774015e-01 2.32729650e+00
7.67289400e-01 1.40542582e-01 3.28230470e-01 2.31668994e-01
5.27737260e-01 2.48668268e-01 -1.75227910e-01 -1.40833151e+00
1.45106301e-01 7.37541497e-01 2.68048137e-01 6.17673218e-01
-6.43891513e-01 1.06530452e+00 5.78847313e+00 8.11299384e-01
-1.22672188e+00 4.56811517e-01 1.90948039e-01 -1.60034150e-01
-2.68375367e-01 8.40429440e-02 -5.84480643e-01 6.27125084e-01
1.51539361e+00 -1.58652812e-01 4.33008015e-01 4.44213063e-01
-1.63943067e-01 7.04496801e-02 -8.65253687e-01 8.56086433e-01
1.38132468e-01 -7.27340281e-01 2.99910251e-02 2.98136145e-01
3.07005733e-01 3.31312031e-01 1.30746603e-01 7.42705524e-01
4.87063110e-01 -1.29321599e+00 7.54013717e-01 1.23632230e-01
7.63349473e-01 -7.39916682e-01 7.52951145e-01 2.11754829e-01
-8.26421499e-01 8.40167515e-03 -3.08692366e-01 -2.76579916e-01
-2.45505180e-02 5.61886802e-02 -8.51032615e-01 3.25500101e-01
7.14320481e-01 6.36316776e-01 -7.46791422e-01 5.24540842e-01
-3.89266700e-01 1.00032318e+00 -3.26622367e-01 -1.17828816e-01
5.22103965e-01 -2.32504889e-01 6.17729366e-01 1.39889693e+00
8.91198739e-02 -3.94651502e-01 -2.55249552e-02 4.65185791e-01
1.68941393e-01 5.66246569e-01 -1.03047478e+00 -2.96894431e-01
3.72587055e-01 1.12866628e+00 -4.22686160e-01 -4.69410904e-02
-5.84973454e-01 9.82313693e-01 7.30767667e-01 2.58198917e-01
-7.69663393e-01 -7.32848108e-01 1.02070642e+00 7.53380135e-02
2.02476561e-01 -3.66028100e-01 -1.41225636e-01 -9.97579873e-01
-4.52968210e-01 -8.89445186e-01 4.77553159e-01 -4.51604158e-01
-1.27497721e+00 6.73497081e-01 -4.72238101e-02 -8.44040632e-01
-4.46931273e-01 -7.29880571e-01 -7.33581245e-01 9.56149459e-01
-6.80072367e-01 -1.18715549e+00 3.76352251e-01 5.51449835e-01
3.62949729e-01 -3.42091441e-01 1.03399360e+00 4.74665701e-01
-8.95018041e-01 9.47154105e-01 2.07050502e-01 4.71762627e-01
9.70096529e-01 -1.04555559e+00 2.33179122e-01 6.48151398e-01
4.40252393e-01 1.06646526e+00 6.75654888e-01 -5.45430124e-01
-9.88165200e-01 -8.22043598e-01 1.40345836e+00 -5.44559717e-01
1.10386324e+00 -6.59811795e-01 -9.94814992e-01 8.74266565e-01
4.68742996e-01 -5.04828513e-01 1.30377555e+00 5.91424406e-01
-5.80508590e-01 3.02595049e-01 -9.19687390e-01 8.34786236e-01
9.66574132e-01 -8.61958027e-01 -9.35762167e-01 2.29563013e-01
4.65912580e-01 -2.93726064e-02 -1.01641750e+00 -2.09692493e-01
7.59273648e-01 -8.60188961e-01 5.70749462e-01 -1.22722721e+00
7.82022953e-01 3.67622346e-01 -3.18093508e-01 -1.67227769e+00
-3.57479125e-01 -2.66415030e-01 5.59988678e-01 1.39861238e+00
5.05599082e-01 -7.94747889e-01 1.32575855e-01 6.19617522e-01
-1.10245347e-01 -6.88793480e-01 -1.40012312e+00 -7.62984037e-01
6.88664734e-01 -7.72607028e-01 2.21934915e-01 1.28942847e+00
3.29879999e-01 6.94346488e-01 -3.20992589e-01 -4.05726731e-01
-3.03427177e-03 -6.95251107e-01 8.12140584e-01 -1.09624290e+00
-5.44685721e-02 -6.11347079e-01 -7.06668139e-01 -2.68341511e-01
6.63458943e-01 -1.62608755e+00 -2.22403422e-01 -1.35842192e+00
-8.04435089e-02 -2.78961509e-01 -2.51290411e-01 7.34550893e-01
-4.28712033e-02 3.31082880e-01 5.15020728e-01 -1.17761537e-01
-6.24502227e-02 5.37621081e-01 5.03373802e-01 1.67317130e-02
-1.94497257e-01 -5.39965272e-01 -7.86232591e-01 5.59494972e-01
7.73216665e-01 -6.29353762e-01 -3.04028876e-02 -4.32413280e-01
1.40578672e-01 -3.50569069e-01 8.08701441e-02 -8.79254282e-01
-7.65894912e-03 9.08949133e-03 2.92093992e-01 -5.02915122e-03
3.75985652e-01 -5.92616498e-01 -5.96906766e-02 5.57186902e-01
-4.98618931e-01 2.17220455e-01 3.96485031e-01 -7.18233967e-03
-8.94467682e-02 -6.34016633e-01 6.77491426e-01 1.55069917e-01
-4.83449161e-01 -1.68280497e-01 -9.17353570e-01 2.03232750e-01
8.65737200e-01 -3.75919372e-01 4.27512489e-02 -3.33921283e-01
-7.37366915e-01 -5.13501950e-02 2.30759427e-01 8.98695648e-01
4.97158974e-01 -1.46314669e+00 -9.57619309e-01 5.92206478e-01
2.55654395e-01 -1.02907181e+00 1.47661015e-01 8.64463866e-01
-3.85814339e-01 1.46087006e-01 -4.44915384e-01 -1.22914612e-01
-1.25726223e+00 2.28317767e-01 -1.09344736e-01 -2.75554776e-01
-5.36280930e-01 7.56507695e-01 -1.72118783e-01 -9.23541009e-01
-1.57715425e-01 1.94643945e-01 -2.91128665e-01 6.21898532e-01
4.19336289e-01 4.37887102e-01 1.30682699e-02 -1.32451594e+00
-3.09011191e-01 3.65721822e-01 -5.12396157e-01 -4.07486439e-01
1.60764277e+00 1.29519686e-01 -1.82647601e-01 1.01778829e+00
1.66308880e+00 4.43561286e-01 -1.49596721e-01 -5.21166027e-02
-6.45530270e-03 -4.75623101e-01 -1.44212529e-01 -5.29178917e-01
-8.13463569e-01 9.20460820e-01 7.06246138e-01 2.34130561e-01
2.80983567e-01 -6.17717244e-02 7.30252147e-01 1.23718619e-01
1.68376928e-03 -1.36533713e+00 -3.06478858e-01 1.01374042e+00
9.78532910e-01 -7.47333109e-01 -2.06862211e-01 2.05021292e-01
-9.81584489e-01 1.22541571e+00 6.81837142e-01 -1.48927808e-01
5.69119871e-01 -1.85905188e-01 9.70113128e-02 -1.58927843e-01
-6.06126070e-01 -1.69570893e-01 1.94929227e-01 6.17706954e-01
9.09923255e-01 1.13710426e-01 -1.01268494e+00 6.73423052e-01
-8.35467279e-01 -4.32042062e-01 6.06099963e-01 7.63929784e-01
2.54016109e-02 -1.23509812e+00 -4.53248233e-01 7.80894995e-01
-4.70215499e-01 -2.32913420e-02 -5.63799918e-01 8.85445893e-01
2.61622310e-01 1.01484954e+00 1.88781679e-01 -7.64364421e-01
3.79324853e-01 6.07213795e-01 4.29574661e-02 -8.87764692e-01
-1.05663490e+00 -3.41693491e-01 2.25526795e-01 -4.44569468e-01
2.85404563e-01 -7.24305093e-01 -1.04879510e+00 -3.57550114e-01
-5.09238429e-02 1.89553037e-01 6.55006647e-01 8.99804473e-01
7.00337254e-03 4.24821585e-01 3.15427601e-01 -7.50154078e-01
-3.27015728e-01 -1.42768312e+00 -6.32640958e-01 7.08814383e-01
8.01995024e-02 -6.93390012e-01 -4.18957800e-01 -4.38686132e-01] | [9.096390724182129, 10.656740188598633] |
e102ed8d-7732-4343-8583-e8ff3f4ad5ab | spatio-temporal-crop-classification-on | 2103.10050 | null | https://arxiv.org/abs/2103.10050v1 | https://arxiv.org/pdf/2103.10050v1.pdf | Spatio-temporal Crop Classification On Volumetric Data | Large-area crop classification using multi-spectral imagery is a widely studied problem for several decades and is generally addressed using classical Random Forest classifier. Recently, deep convolutional neural networks (DCNN) have been proposed. However, these methods only achieved results comparable with Random Forest. In this work, we present a novel CNN based architecture for large-area crop classification. Our methodology combines both spatio-temporal analysis via 3D CNN as well as temporal analysis via 1D CNN. We evaluated the efficacy of our approach on Yolo and Imperial county benchmark datasets. Our combined strategy outperforms both classical as well as recent DCNN based methods in terms of classification accuracy by 2% while maintaining a minimum number of parameters and the lowest inference time. | ['Abubakr Muhammad', 'Murtaza Taj', 'Salar Saeed', 'Muhammad Usman Qadeer'] | 2021-03-18 | null | null | null | null | ['crop-classification'] | ['miscellaneous'] | [ 2.12325469e-01 -6.14433885e-01 -1.45061314e-01 -1.86772153e-01
-4.14044857e-01 -6.45949483e-01 6.94374084e-01 3.93516034e-01
-5.81116617e-01 9.94798958e-01 -4.19646323e-01 -4.84831452e-01
-3.35028470e-01 -1.50293219e+00 -7.03429878e-01 -7.72846282e-01
-4.57845777e-01 8.30170363e-02 2.77454287e-01 -2.80945927e-01
-1.90259218e-01 6.87457979e-01 -1.78495502e+00 2.39417106e-02
1.03059518e+00 1.21153617e+00 3.60893965e-01 7.01219320e-01
1.15521401e-01 4.72701043e-01 -4.04872388e-01 1.10816203e-01
2.84746557e-01 7.83534050e-02 -7.31221437e-01 -8.07360336e-02
5.28077960e-01 -3.82498711e-01 1.85148358e-01 9.10265028e-01
3.91628861e-01 6.89235702e-02 5.46844363e-01 -9.77839887e-01
-5.41664898e-01 4.86925930e-01 -8.21686327e-01 2.28129402e-01
-2.19447076e-01 8.36260766e-02 6.67683005e-01 -4.52552110e-01
5.47432482e-01 9.76825714e-01 1.06698072e+00 -3.26648623e-01
-1.47842371e+00 -8.13398063e-01 2.23612979e-01 2.96330392e-01
-1.34890783e+00 1.91693619e-01 4.15616930e-01 -4.60920691e-01
1.17364538e+00 -1.08253904e-01 1.01762795e+00 8.54023218e-01
5.82447410e-01 6.67777002e-01 1.39308453e+00 -4.45528775e-01
9.02789831e-02 -4.48581517e-01 2.62546152e-01 5.99732161e-01
2.64502108e-01 5.31666875e-01 -2.37203613e-01 -3.56339328e-02
7.90621400e-01 9.12845656e-02 2.56480694e-01 -3.02107841e-01
-1.06836128e+00 1.16060138e+00 9.70866740e-01 3.56871754e-01
-8.32268238e-01 2.65191644e-01 4.86211360e-01 1.25226185e-01
8.44873548e-01 2.74907708e-01 -8.40938091e-01 2.71060973e-01
-1.27999210e+00 5.41884124e-01 2.95875609e-01 7.09689200e-01
7.43560672e-01 1.28459275e-01 1.21080205e-01 8.37702692e-01
-2.74298210e-02 7.00535595e-01 1.23905055e-01 -5.60852885e-01
1.10996157e-01 6.36548221e-01 1.04388716e-02 -1.17702413e+00
-7.34791994e-01 -8.07193339e-01 -1.34993982e+00 4.79249656e-01
4.69039410e-01 -1.99351981e-01 -1.16926980e+00 1.43830276e+00
5.35638705e-02 9.97044817e-02 3.81900854e-02 6.79352641e-01
5.27677476e-01 5.75900555e-01 4.13940609e-01 -1.00860469e-01
1.39863217e+00 -8.19097757e-01 -6.73744678e-01 -8.21288526e-02
4.29676354e-01 -7.56490886e-01 6.38125837e-01 4.65737373e-01
-5.44246554e-01 -6.02971017e-01 -1.16026652e+00 2.54299492e-01
-9.33580697e-01 5.06520450e-01 1.00768077e+00 5.40905654e-01
-9.70848858e-01 7.97192395e-01 -8.89568746e-01 -6.56333745e-01
8.86749804e-01 2.15422824e-01 -3.86259466e-01 -8.12726319e-02
-1.21389687e+00 7.99579442e-01 6.38146043e-01 6.42028451e-01
-1.00149703e+00 -4.88593280e-01 -7.24679947e-01 -1.23164274e-01
3.43710661e-01 -4.41092551e-01 9.93057847e-01 -8.80342364e-01
-1.22178614e+00 5.69929600e-01 1.45400856e-02 -9.88944352e-01
3.45145673e-01 -5.09333491e-01 -2.98565537e-01 -9.31052417e-02
2.97244996e-01 6.96502030e-01 4.15840387e-01 -9.84183192e-01
-9.01385367e-01 -6.30762637e-01 1.75123334e-01 -3.83621864e-02
-4.78177406e-02 -7.08141327e-02 4.57471460e-01 -8.15113604e-01
1.87660620e-01 -1.00248337e+00 -4.95281100e-01 1.71737194e-01
-1.90482989e-01 -3.81927192e-02 9.37975824e-01 -7.02910662e-01
9.52506125e-01 -1.81841505e+00 -1.06258117e-01 -8.56915712e-02
-6.24369606e-02 4.75802988e-01 -1.19398363e-01 2.06593916e-01
8.37722793e-03 1.76400393e-01 -5.57899952e-01 3.04595232e-01
-4.25520092e-01 5.64956293e-02 -3.56198102e-02 5.25768340e-01
5.92760563e-01 1.03127038e+00 -7.96638787e-01 -2.93735832e-01
5.02276003e-01 6.00890696e-01 -8.53189975e-02 -2.28812452e-02
-4.18273985e-01 -1.33874854e-02 -2.26771921e-01 9.54140067e-01
1.19852102e+00 1.49273174e-02 2.91845411e-01 -1.66575775e-01
-6.18340254e-01 -2.28308484e-01 -9.71002996e-01 1.56966317e+00
-7.08318293e-01 8.16350102e-01 -3.13464627e-02 -1.27212751e+00
1.01455796e+00 2.29009569e-01 2.42315337e-01 -5.35722673e-01
-6.89688027e-02 3.28430355e-01 4.62538972e-02 -1.85762540e-01
4.67842996e-01 6.89220130e-02 1.87768742e-01 6.10797144e-02
7.06526190e-02 -3.13478447e-02 6.56841174e-02 -4.57161367e-01
7.25517392e-01 6.59898579e-01 6.98747694e-01 -6.44993365e-01
2.03926578e-01 6.69256985e-01 3.26046497e-01 7.60722458e-01
-3.59866977e-01 3.60321462e-01 3.91360193e-01 -8.20883393e-01
-8.02985072e-01 -7.08331585e-01 -3.89632672e-01 9.00707960e-01
3.93889705e-03 9.43983868e-02 -3.37567389e-01 -6.27935290e-01
2.57473856e-01 4.91923749e-01 -9.47764575e-01 4.51167971e-01
-5.80554068e-01 -1.37841094e+00 8.21232796e-01 7.92579234e-01
1.18765116e+00 -1.16140866e+00 -9.47778821e-01 4.14079130e-01
-8.61564837e-03 -1.09489453e+00 4.50479418e-01 8.16654265e-01
-1.12115228e+00 -1.02224469e+00 -8.29321504e-01 -7.54429698e-01
7.70001411e-02 5.75746298e-01 1.12294579e+00 -3.20552170e-01
-5.03971934e-01 -5.22501707e-01 -5.05605221e-01 -6.83232963e-01
5.19590490e-02 6.81406736e-01 -4.64427412e-01 -3.18571448e-01
5.69237292e-01 -7.88185835e-01 -5.26625335e-01 1.01332031e-01
-7.96528637e-01 -2.26759594e-02 8.49742889e-01 9.74973083e-01
5.10647535e-01 3.43507200e-01 6.71489179e-01 -6.64847314e-01
2.51489967e-01 -5.60965061e-01 -9.29992497e-01 1.45562589e-01
-6.49161041e-01 -1.12335384e-01 6.16438389e-01 -3.21084589e-01
-9.50141251e-01 4.80575919e-01 -2.71008611e-02 -2.10418850e-01
-7.17400908e-01 8.75020444e-01 1.53249398e-01 4.61562872e-02
6.52587354e-01 2.54533052e-01 -9.01213959e-02 -4.94271010e-01
2.78355241e-01 5.81764877e-01 2.88046449e-01 -1.07120581e-01
6.94115579e-01 6.98005855e-01 2.81946450e-01 -9.71221626e-01
-8.05532098e-01 -3.95068109e-01 -1.15422606e+00 -1.51447847e-01
1.08638144e+00 -1.14236307e+00 -5.93511283e-01 1.01443422e+00
-1.14115059e+00 -5.18453360e-01 8.66412967e-02 4.41748589e-01
-2.55158275e-01 7.86788613e-02 -3.07864040e-01 -8.03288043e-01
-5.06534159e-01 -8.78944337e-01 1.11461556e+00 1.16543025e-01
1.88000485e-01 -9.24526691e-01 -8.16611350e-02 -1.24690026e-01
8.53308141e-01 7.76748836e-01 8.25500548e-01 -6.81755394e-02
-4.77510780e-01 -1.98528647e-01 -7.92179644e-01 1.56580076e-01
2.91440457e-01 1.82236969e-01 -1.16722679e+00 -2.29004428e-01
-5.69027245e-01 -2.32765332e-01 1.30347490e+00 1.08115625e+00
1.03181553e+00 3.20484966e-01 -3.32382113e-01 6.33165956e-01
2.01155734e+00 2.85342872e-01 4.44701850e-01 8.45195353e-01
4.35186148e-01 5.20065129e-01 9.08966243e-01 1.94065616e-01
5.29003218e-02 5.66731155e-01 8.76772940e-01 -3.38076025e-01
1.49028540e-01 2.23718345e-01 -3.46838981e-02 6.49525449e-02
-5.19916475e-01 -2.72661269e-01 -1.11093247e+00 9.14146245e-01
-2.01327205e+00 -1.14667797e+00 -4.04898465e-01 1.84957242e+00
3.80852640e-01 1.30671427e-01 1.75037935e-01 2.50966132e-01
5.31220138e-01 4.12256807e-01 -4.84180212e-01 -2.34751895e-01
-6.55097246e-01 6.64463997e-01 1.20848215e+00 1.11499541e-01
-1.90745759e+00 1.07788265e+00 6.45974731e+00 7.53947794e-01
-1.33829963e+00 9.01310220e-02 3.29970539e-01 1.28163099e-01
4.47559744e-01 2.20154971e-02 -7.36176252e-01 -3.28868590e-02
8.06287110e-01 1.42786995e-01 1.05019197e-01 8.41239274e-01
1.67227849e-01 -5.49394906e-01 -2.12108403e-01 3.14155668e-01
-3.83212417e-01 -1.44493842e+00 -2.49300689e-01 2.33102441e-01
7.49508977e-01 3.27067167e-01 -2.14307934e-01 2.15016559e-01
4.97137070e-01 -1.22233295e+00 4.39375371e-01 2.76683927e-01
6.63310647e-01 -8.48337412e-01 1.24042797e+00 2.06926778e-01
-1.70056438e+00 -2.66664296e-01 -2.39781082e-01 -3.23174089e-01
-1.15592130e-01 7.76321232e-01 -3.32660556e-01 1.05484486e+00
1.10047472e+00 1.16375172e+00 -6.71747088e-01 9.23785388e-01
7.87874497e-03 8.00765157e-01 -4.73987609e-01 -8.09931606e-02
7.82946885e-01 -1.70563862e-01 2.62868136e-01 1.17781031e+00
2.89317012e-01 -2.93500572e-01 3.42433691e-01 5.53691089e-01
3.31251383e-01 6.81534559e-02 -8.79603386e-01 9.77074653e-02
1.19615324e-01 1.22154999e+00 -9.26145196e-01 -1.47180229e-01
-2.48027831e-01 6.15247428e-01 1.53517887e-01 1.22620594e-02
-6.42961442e-01 -5.64807415e-01 6.77934706e-01 -9.47074220e-02
5.95796287e-01 -5.97565532e-01 -2.54871815e-01 -8.71708691e-01
-5.72442114e-02 -6.15097165e-01 3.09858978e-01 -5.83685398e-01
-9.81203675e-01 8.08937728e-01 1.08831070e-01 -1.35705280e+00
-4.29100767e-02 -8.74672592e-01 -3.84625643e-01 9.56686914e-01
-2.14149451e+00 -1.75549650e+00 -9.24122095e-01 2.18190193e-01
6.81921959e-01 -2.44502470e-01 1.27222288e+00 1.97383568e-01
-5.70320070e-01 2.71016993e-02 3.61697763e-01 1.32457778e-01
4.20799881e-01 -1.30841875e+00 4.60590392e-01 9.01290476e-01
-2.91729689e-01 1.44301072e-01 3.80603552e-01 -5.62403798e-01
-8.30724120e-01 -1.53941143e+00 7.28190541e-01 2.17113048e-01
7.07166791e-01 7.00303819e-03 -7.23492086e-01 5.47085106e-01
4.85371083e-01 1.15457609e-01 3.02924186e-01 3.35305631e-01
-3.13478082e-01 -2.86637932e-01 -1.14501202e+00 1.26291052e-01
9.16878641e-01 -2.23902717e-01 1.46947801e-01 2.36394793e-01
5.96761942e-01 -5.27781667e-03 -9.76459086e-01 7.82732785e-01
9.70088601e-01 -8.45374167e-01 7.61624932e-01 -3.56976688e-01
7.33516932e-01 -2.96856433e-01 -4.57597494e-01 -1.51624346e+00
-5.44364572e-01 1.04379132e-01 2.54247248e-01 8.52901101e-01
2.61341006e-01 -4.88782614e-01 6.72766984e-01 -6.93932474e-01
2.07133681e-01 -7.07322538e-01 -6.89975798e-01 -1.10596800e+00
4.21440631e-01 -4.68515098e-01 6.69097543e-01 8.51382196e-01
-8.31572950e-01 8.05721134e-02 -3.00082088e-01 3.68160367e-01
6.20279074e-01 4.99032110e-01 5.37424386e-01 -1.54175699e+00
3.08232009e-01 -5.65347970e-01 -5.50257385e-01 -4.50905323e-01
1.52235910e-01 -6.72873139e-01 -5.68375885e-02 -1.51565146e+00
8.61381963e-02 -5.48690975e-01 -1.67201683e-01 6.56259477e-01
-1.99302305e-02 3.77600282e-01 9.15307272e-03 -3.56489606e-02
1.08912334e-01 3.42885911e-01 8.62616897e-01 -2.80884236e-01
-1.37700677e-01 2.22768977e-01 -9.78613719e-02 4.66166854e-01
1.19576621e+00 -4.57924992e-01 4.97894473e-02 -5.34721613e-01
1.08486938e-03 -2.25735828e-01 7.77236581e-01 -1.15507925e+00
-2.47962356e-01 -2.69627869e-01 5.92046678e-01 -1.16778839e+00
3.06555144e-02 -7.26414204e-01 2.17117161e-01 5.92225790e-01
-1.10481888e-01 1.06995367e-01 6.15463972e-01 5.60870230e-01
-3.50440890e-01 1.36437193e-01 9.95356262e-01 -1.70761198e-01
-9.87541795e-01 2.01554656e-01 -6.88673913e-01 -6.99984252e-01
1.12991166e+00 -5.43866791e-02 -3.47812653e-01 5.65207889e-03
-8.23023558e-01 2.93304086e-01 1.67669654e-01 4.17258292e-01
2.42278695e-01 -1.09278762e+00 -9.59799826e-01 2.06547990e-01
2.40075558e-01 6.39619231e-02 1.86317205e-01 7.81854749e-01
-1.07837677e+00 7.14678586e-01 -7.04003751e-01 -8.12616467e-01
-1.37561572e+00 3.68053943e-01 6.04229152e-01 -6.92997754e-01
-4.77514654e-01 5.13939500e-01 -3.10906380e-01 -7.53800154e-01
-3.05415913e-02 -4.24571902e-01 -5.27776718e-01 5.21582544e-01
1.60488822e-02 3.08510274e-01 2.13587016e-01 -4.49689478e-01
-4.76911545e-01 5.38104117e-01 1.70941681e-01 9.92050841e-02
1.81976569e+00 2.56374180e-01 -4.33257632e-02 5.50164998e-01
1.02366817e+00 -7.49542832e-01 -8.35205674e-01 -1.48432836e-01
1.19189337e-01 -4.33568656e-01 5.88871717e-01 -9.67551410e-01
-1.18752551e+00 1.01690984e+00 1.28490365e+00 4.83268887e-01
1.41114891e+00 -5.77883780e-01 5.61517835e-01 3.60749811e-01
3.90433848e-01 -8.55406761e-01 -5.31076252e-01 5.80165029e-01
7.63615429e-01 -1.69456661e+00 3.25058669e-01 -4.36692595e-01
-2.94038236e-01 1.37431765e+00 5.98955095e-01 -4.92067814e-01
1.19514263e+00 3.14640909e-01 1.38619170e-01 -1.90180510e-01
-6.74832404e-01 -8.56901765e-01 -1.56938151e-01 7.57273436e-01
5.92522264e-01 4.76721019e-01 -1.27027288e-01 1.61889911e-01
9.79009345e-02 4.47431386e-01 2.03482911e-01 1.18995786e+00
-2.46367261e-01 -7.79137850e-01 -2.25463524e-01 5.74509323e-01
-4.01182473e-01 -2.01572061e-01 -6.55881539e-02 1.15941370e+00
3.74025732e-01 8.10767114e-01 1.82596162e-01 -2.66174883e-01
3.46993536e-01 -9.62771326e-02 2.69733310e-01 -3.07493210e-01
-8.84323835e-01 8.15104768e-02 3.97623405e-02 -3.87259662e-01
-1.12112570e+00 -7.91065753e-01 -4.13035333e-01 -4.11018044e-01
-8.70234907e-01 -3.98275614e-01 8.83249700e-01 7.92869866e-01
2.63748825e-01 6.47310615e-01 5.73163569e-01 -9.64591801e-01
-3.41863990e-01 -1.33161795e+00 -8.63506377e-01 -2.86856890e-01
3.82730454e-01 -7.77375281e-01 4.30102553e-03 -1.18201353e-01] | [9.386289596557617, -1.5526238679885864] |
4bd90a8c-866c-4fd7-9f85-3e0df4dad19b | optimal-visual-search-based-on-a-model-of | null | null | http://proceedings.neurips.cc/paper/2020/hash/691dcb1d65f31967a874d18383b9da75-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/691dcb1d65f31967a874d18383b9da75-Paper.pdf | Optimal visual search based on a model of target detectability in natural images | To analyse visual systems, the concept of an ideal observer promises an optimal response for a given task. Bayesian ideal observers can provide optimal responses under uncertainty, if they are given the true distributions as input. In visual search tasks, prior studies have used signal to noise ratio (SNR) or psychophysics experiments to set the distributional parameters for simple targets on backgrounds with known patterns, however these methods do not easily translate to complex targets on natural scenes. Here, we develop a model of target detectability in natural images to estimate the parameters of target-present and target-absent distributions for a visual search task. We present a novel approach for approximating the foveated detectability of a known target in natural backgrounds based on biological aspects of human visual system. Our model considers both the uncertainty about target position and the visual system's variability due to its reduced performance in the periphery compared to the fovea. Our automated prediction algorithm uses trained logistic regression as a post processing phase of a pre-trained deep neural network. Eye tracking data from 12 observers detecting targets on natural image backgrounds are used as ground truth to tune foveation parameters and evaluate the model, using cross-validation. Finally, the model of target detectability is used in a Bayesian ideal observer model of visual search, and compared to human search performance. | ['Lars Kulik', 'Andrew Turpin', 'Krista Ehinger', 'Shima Rashidi'] | 2020-12-01 | null | null | null | neurips-2020-12 | ['foveation'] | ['computer-vision'] | [ 3.64450127e-01 -1.28110364e-01 2.37628203e-02 -3.31528127e-01
-3.24106365e-01 -5.18172860e-01 5.40811241e-01 5.17012775e-02
-1.01874626e+00 6.41049623e-01 -1.81593195e-01 -4.25828695e-01
-1.67067200e-01 -1.59479648e-01 -6.39981151e-01 -6.01806998e-01
2.84170121e-01 1.09599560e-01 7.64717996e-01 9.98362303e-02
3.86070609e-01 6.05445683e-01 -1.84515703e+00 1.96764871e-01
5.39158940e-01 1.45381510e+00 6.60506308e-01 1.02161825e+00
4.96899903e-01 4.71812695e-01 -7.09647000e-01 -1.76718727e-01
5.25335312e-01 -4.33151186e-01 -1.69956416e-01 -1.86387122e-01
6.11768305e-01 -2.42493197e-01 -2.63753772e-01 1.24784255e+00
8.47778738e-01 2.14030266e-01 9.53112662e-01 -9.92028594e-01
-8.46668363e-01 9.78303179e-02 -4.68216538e-01 7.18061328e-01
3.83949965e-01 6.85222805e-01 8.23853374e-01 -4.57502097e-01
-9.72242840e-03 1.22091627e+00 1.14268333e-01 4.79847580e-01
-1.70338404e+00 -2.57941753e-01 2.49234959e-03 5.51822126e-01
-1.52882040e+00 -6.56063020e-01 3.23100179e-01 -7.32581437e-01
8.39556396e-01 1.81481257e-01 7.02133656e-01 9.61593330e-01
5.90371370e-01 3.36468369e-01 1.58773804e+00 -7.78792918e-01
4.03903335e-01 3.93757015e-01 -7.42192715e-02 1.76414445e-01
2.38140285e-01 8.93357933e-01 -4.36912537e-01 1.03552332e-02
7.25867569e-01 -6.90968990e-01 -5.90067983e-01 -1.50801301e-01
-1.00449550e+00 6.15078151e-01 8.23468328e-01 9.65308249e-02
-5.96085310e-01 -2.94372328e-02 -1.51945114e-01 1.56791210e-01
1.03741407e-01 5.38926899e-01 -2.32156679e-01 1.55925378e-01
-8.22709799e-01 1.40701964e-01 6.31799221e-01 6.74815476e-01
2.85265088e-01 1.46589021e-03 -7.72029877e-01 8.67134094e-01
4.85648155e-01 8.96342039e-01 6.10346794e-01 -7.88942695e-01
-3.24517399e-01 1.39662102e-01 6.17666721e-01 -8.64011109e-01
-5.05588651e-01 -5.76105297e-01 -1.57341123e-01 9.31448400e-01
9.00142729e-01 1.83531612e-01 -1.10024703e+00 1.78567064e+00
9.95128900e-02 -4.44353223e-01 -1.22989714e-01 1.44008934e+00
3.87718827e-01 3.40218008e-01 7.92423636e-02 -6.54961288e-01
1.40993452e+00 -3.33769649e-01 -5.69466531e-01 -5.94842196e-01
-1.64326653e-01 -9.72361207e-01 1.14993489e+00 5.08963168e-01
-9.03561652e-01 -8.40472937e-01 -8.66227448e-01 3.57612610e-01
-1.18974462e-01 2.42700040e-01 1.10597216e-01 7.49354661e-01
-1.04180610e+00 2.25911900e-01 -6.27657235e-01 -2.59744018e-01
2.69527286e-01 2.79423177e-01 -4.42333445e-02 2.05507591e-01
-1.03874147e+00 1.59796417e+00 5.01408935e-01 2.16556937e-01
-1.05419350e+00 -3.62500608e-01 -6.94573402e-01 6.03076406e-02
3.61353397e-01 -6.97969496e-01 1.49180448e+00 -1.26120973e+00
-1.31335402e+00 9.59245265e-01 -3.99767637e-01 -6.38128936e-01
2.52262980e-01 8.20356905e-02 -3.72832298e-01 -8.07752013e-02
-1.08921058e-01 7.02706635e-01 1.13436770e+00 -1.25100803e+00
-5.90729117e-01 -4.37233537e-01 -1.66395441e-01 2.70330638e-01
5.51769972e-01 2.95619398e-01 -9.72636044e-03 -2.18443170e-01
7.70422537e-03 -6.99641466e-01 -1.56286597e-01 2.47035369e-01
-1.46330878e-01 -2.14537770e-01 1.02865696e-01 -5.09436667e-01
1.10938692e+00 -2.18185997e+00 -3.45519274e-01 1.91869020e-01
2.91297972e-01 3.25512767e-01 -1.77586794e-01 -1.97554052e-01
-1.38561185e-02 -4.15831983e-01 1.43170655e-01 3.94415975e-01
1.71310768e-01 -3.75104547e-01 -2.56516039e-01 6.63997769e-01
1.97753385e-01 5.76746106e-01 -7.62958169e-01 -2.20542923e-01
1.69064477e-01 3.62924099e-01 -2.27152511e-01 4.42895740e-01
-3.10848862e-01 2.54335940e-01 1.41165838e-01 4.87071157e-01
6.62292004e-01 8.88749883e-02 -3.90220106e-01 -2.84204274e-01
-3.14093083e-01 -7.25667179e-02 -9.86028671e-01 8.22742820e-01
-3.33945632e-01 1.15745747e+00 -1.08088411e-01 -5.01145244e-01
1.07288110e+00 -1.39441222e-01 -3.57037246e-01 -1.00198555e+00
4.53735322e-01 3.64778519e-01 9.51026499e-01 -4.32419121e-01
2.12934032e-01 -1.73907235e-01 4.71234977e-01 2.14675725e-01
1.47331050e-02 -3.22138369e-01 -4.98922635e-03 -3.20902675e-01
7.41987765e-01 1.03673279e-01 5.68326294e-01 -2.04273403e-01
2.91479260e-01 -9.44314301e-02 1.53727636e-01 1.13833439e+00
-5.97497821e-01 6.00741208e-01 4.06321585e-01 -4.32947695e-01
-9.69722033e-01 -1.29400957e+00 -5.88795722e-01 9.70033467e-01
6.61883056e-02 3.36160839e-01 -7.79771447e-01 1.43160462e-01
-2.19176024e-01 1.28325033e+00 -6.26919329e-01 -5.04567742e-01
1.10571072e-01 -6.78805888e-01 1.30774140e-01 1.38225034e-01
1.34214833e-01 -1.14462543e+00 -1.48798954e+00 -1.13439336e-01
7.35373572e-02 -1.01296484e+00 -4.12939668e-01 2.75815189e-01
-4.35163490e-02 -1.17283225e+00 -8.16284239e-01 -3.14321399e-01
5.08329749e-01 1.26150697e-01 1.00900054e+00 -3.75315428e-01
-6.62167132e-01 5.31624734e-01 -7.75110945e-02 -1.01726460e+00
-3.92871320e-01 -7.78780520e-01 4.70844150e-01 6.71121776e-02
7.68573284e-01 -1.03023343e-01 -8.90360594e-01 7.02494800e-01
-5.19158244e-01 -3.63195330e-01 7.56095350e-01 9.76623654e-01
5.40908694e-01 -1.55725004e-02 -3.07479296e-02 1.21369019e-01
8.40461314e-01 -1.68853089e-01 -1.22676027e+00 3.99342537e-01
-5.41067541e-01 -1.38444593e-02 1.79766372e-01 -1.10866237e+00
-1.12604690e+00 -1.93498062e-03 3.98569614e-01 -4.82420415e-01
-3.39109302e-01 2.69108534e-01 1.86480910e-01 -2.40939140e-01
1.67415214e+00 2.08003744e-01 9.71587971e-02 2.03277186e-01
1.00973457e-01 8.39253604e-01 4.50739384e-01 -2.72086829e-01
3.25990289e-01 1.69289649e-01 1.70782715e-01 -8.53996456e-01
-1.01181006e+00 -1.79390147e-01 -6.24420285e-01 -4.59402770e-01
5.89690506e-01 -6.53339386e-01 -1.17811286e+00 3.55132669e-01
-1.21786761e+00 -5.54364681e-01 -2.41310611e-01 1.14645791e+00
-9.78286147e-01 -3.15738320e-02 3.14146668e-01 -1.50233471e+00
1.96166672e-02 -1.13961649e+00 8.30796897e-01 7.07719088e-01
-1.61787257e-01 -7.25329936e-01 -4.81376192e-03 -5.94233237e-02
5.85546017e-01 -3.24843191e-02 6.45727277e-01 -7.36787379e-01
-5.51458955e-01 -3.44745606e-01 -4.63574439e-01 3.75036508e-01
-1.90049838e-02 -7.07005803e-03 -1.32018113e+00 -1.48071751e-01
3.55266869e-01 -3.25691491e-01 7.83792496e-01 1.28271246e+00
6.61549211e-01 -7.17298612e-02 -2.66614735e-01 4.49299812e-01
1.29788518e+00 4.82178599e-01 4.29388642e-01 1.58648700e-01
-1.77218974e-01 8.68495464e-01 6.78666949e-01 4.35383469e-01
-2.36351922e-01 8.30178976e-01 5.73050678e-01 1.27685100e-01
-1.45579979e-01 -5.77702234e-03 9.50203687e-02 -3.78138304e-01
-1.67352911e-02 -3.66654426e-01 -9.12605226e-01 3.27398330e-01
-1.40266895e+00 -9.87627923e-01 2.82251388e-02 2.89925051e+00
6.75750852e-01 4.48696166e-01 2.41401821e-01 -1.80449277e-01
7.45466053e-01 -4.44071800e-01 -8.84914935e-01 -3.70814115e-01
-3.61651063e-01 -1.53863244e-03 5.97269654e-01 6.73747659e-01
-7.09343553e-01 6.58687055e-01 7.30269718e+00 8.49890471e-01
-1.14998460e+00 -1.95874453e-01 5.39683640e-01 -1.87752888e-01
2.55551994e-01 2.42734402e-02 -1.03558969e+00 3.24928641e-01
1.11521482e+00 -4.27130073e-01 7.49544978e-01 5.74678361e-01
4.77983326e-01 -9.27200198e-01 -1.19326472e+00 1.21785116e+00
1.24106281e-01 -6.51303232e-01 -3.40939581e-01 -4.24595661e-02
6.33676648e-02 -4.88311090e-02 4.51510996e-01 6.32826984e-02
2.25989863e-01 -1.31902122e+00 9.05660093e-01 1.02743733e+00
6.18665755e-01 -1.98429376e-01 6.68524802e-01 6.27432525e-01
-4.12862301e-01 -3.06650460e-01 -9.19143796e-01 -2.05769524e-01
-1.43136054e-01 2.41011247e-01 -1.26697135e+00 -2.77651280e-01
7.05556870e-01 -1.48986727e-01 -8.73919666e-01 1.73308098e+00
-1.06323943e-01 3.09583753e-01 -6.47944868e-01 -3.75418484e-01
-1.13145351e-01 1.85822964e-01 7.80419171e-01 9.65770006e-01
3.04541558e-01 3.44386064e-02 -3.67179692e-01 1.40705526e+00
7.85101414e-01 3.98285203e-02 -5.26077688e-01 1.20150648e-01
5.44978023e-01 1.03094149e+00 -5.79986989e-01 2.00431645e-01
-1.95309326e-01 5.58213890e-01 2.20655054e-01 6.12274289e-01
-6.42923355e-01 -3.54357123e-01 5.61755896e-02 2.12332338e-01
4.22166735e-02 7.63352737e-02 -5.96788749e-02 -4.86217916e-01
1.64521430e-02 -9.14350212e-01 2.29124755e-01 -1.41872728e+00
-1.26366818e+00 9.21841085e-01 4.37555313e-01 -1.03444266e+00
-3.08907032e-01 -1.19406736e+00 -3.61662418e-01 1.64966607e+00
-1.11476016e+00 -8.43047976e-01 -2.47872621e-01 5.63570797e-01
3.10966551e-01 -4.80753779e-01 5.59706390e-01 -4.57316697e-01
-8.19711983e-02 5.82538366e-01 -1.74243540e-01 -2.69373775e-01
7.75157988e-01 -1.12164307e+00 -9.74887982e-03 8.19911242e-01
1.06758349e-01 5.03942013e-01 1.15633786e+00 -4.46703672e-01
-7.39629388e-01 -2.10643619e-01 2.87620485e-01 -6.89053357e-01
6.90189183e-01 -1.51769757e-01 -7.87630260e-01 2.82021791e-01
8.89603570e-02 1.85731322e-01 5.95870376e-01 1.71169899e-02
-1.18934855e-01 -1.02968983e-01 -9.76839125e-01 7.21401393e-01
5.55825591e-01 -3.59252453e-01 -7.41980255e-01 2.26645783e-01
3.53965551e-01 -4.25200343e-01 -3.54621679e-01 2.89640814e-01
8.48056138e-01 -1.24707544e+00 7.38560319e-01 -6.59779787e-01
-4.07402962e-01 -4.26527441e-01 -3.33946496e-01 -1.69094503e+00
-4.92659926e-01 -4.76925284e-01 1.78590119e-01 5.23621857e-01
4.67868060e-01 -6.34438813e-01 1.97717264e-01 5.68251848e-01
2.82894284e-01 -3.98343593e-01 -1.19720173e+00 -7.57516563e-01
-2.87486583e-01 -4.83635396e-01 -1.47724003e-01 -7.75594041e-02
-3.32519978e-01 3.08562428e-01 -4.75920662e-02 3.73257518e-01
8.95136118e-01 -2.69261181e-01 6.32717431e-01 -1.25490975e+00
-4.56556201e-01 -7.36736536e-01 -8.22158098e-01 -9.96499121e-01
3.77796143e-02 -3.44775617e-01 4.04268831e-01 -1.10378504e+00
1.07230604e-01 1.05997529e-02 -3.83416742e-01 2.08441541e-03
-1.54718071e-01 -6.56927302e-02 5.47154322e-02 2.04021722e-01
-8.89006481e-02 3.39759380e-01 1.39183593e+00 -5.29373698e-02
-3.26109380e-01 6.06776357e-01 -6.49214327e-01 6.25838995e-01
4.95959789e-01 -4.01629001e-01 -4.36027646e-01 -1.30596086e-01
2.73980677e-01 -1.24851540e-02 8.42308640e-01 -1.06525874e+00
2.95703828e-01 -3.71405274e-01 9.20926213e-01 -2.97881365e-01
4.85515267e-01 -9.32870865e-01 -1.33857116e-01 4.44229335e-01
-4.71618384e-01 -3.23024362e-01 5.23900747e-01 7.74589121e-01
6.24943934e-02 -6.64241195e-01 1.22480440e+00 8.68109334e-03
-4.95409548e-01 -2.10590333e-01 -4.84415472e-01 -1.33289201e-02
1.02087963e+00 -3.69726926e-01 -5.38654745e-01 -5.37259042e-01
-1.07542741e+00 -9.73146409e-02 2.44126767e-01 1.36551425e-01
7.42838502e-01 -8.75678599e-01 -7.15961039e-01 2.87427604e-01
1.78094879e-01 -5.77988207e-01 -6.27311133e-03 1.02371705e+00
-3.83462340e-01 5.36351562e-01 -5.25303781e-01 -1.06298590e+00
-1.21660924e+00 9.53687906e-01 9.56435382e-01 5.36992311e-01
2.99680144e-01 9.91545677e-01 5.44444978e-01 4.56781924e-01
3.37887198e-01 -2.78910786e-01 -2.83863932e-01 -2.86872923e-01
7.88952172e-01 6.66239560e-02 -3.32046777e-01 -6.26724601e-01
-1.25329435e-01 3.76784325e-01 8.83835927e-02 -3.34751129e-01
5.82455337e-01 -2.22635284e-01 2.13954806e-01 8.88036966e-01
3.55281770e-01 -1.58660904e-01 -1.56313252e+00 -4.74098593e-01
-2.23581448e-01 -8.31760526e-01 4.28531617e-01 -1.14515746e+00
-2.17820421e-01 9.96948004e-01 1.27255797e+00 3.37312549e-01
1.22339380e+00 1.54494002e-01 -4.84370917e-01 4.82234240e-01
8.27785805e-02 -9.53379214e-01 1.74922734e-01 1.13848127e-01
1.23857903e+00 -1.43211651e+00 -6.52933195e-02 6.17062487e-03
-8.13151836e-01 1.10684299e+00 8.16593885e-01 1.66287988e-01
6.82248414e-01 1.11619830e-02 1.70890689e-01 8.23637918e-02
-8.23634088e-01 -7.47347891e-01 8.57942343e-01 1.17461133e+00
2.17712954e-01 5.65636251e-03 3.81603539e-02 4.87309456e-01
-3.10823351e-01 -1.28065735e-01 2.96786129e-01 4.18116570e-01
-8.18184078e-01 -2.27039769e-01 -7.74997354e-01 4.18882996e-01
-3.33229601e-01 -3.33972245e-01 -3.00001562e-01 6.73143089e-01
3.80247161e-02 1.14797580e+00 4.47160274e-01 -1.24611728e-01
6.61732674e-01 -7.34599959e-03 9.11087871e-01 -5.86850047e-01
4.22955714e-02 4.40335542e-01 -3.41341794e-01 -1.90278932e-01
-2.97865391e-01 -5.67076683e-01 -5.13234735e-01 3.06556493e-01
-5.76610446e-01 -1.16504200e-01 7.29351819e-01 7.09416509e-01
-2.41901670e-02 2.91453481e-01 2.90981025e-01 -8.92808378e-01
-9.36234534e-01 -1.24310136e+00 -6.18077338e-01 -6.13348782e-02
4.47053224e-01 -1.06496286e+00 -8.12586427e-01 4.59606498e-02] | [9.981282234191895, 2.0468661785125732] |
a2a724dd-a4e5-4c14-864d-5c826cdd8bf7 | from-pixels-to-sentiment-fine-tuning-cnns-for | 1604.03489 | null | http://arxiv.org/abs/1604.03489v2 | http://arxiv.org/pdf/1604.03489v2.pdf | From Pixels to Sentiment: Fine-tuning CNNs for Visual Sentiment Prediction | Visual multimedia have become an inseparable part of our digital social
lives, and they often capture moments tied with deep affections. Automated
visual sentiment analysis tools can provide a means of extracting the rich
feelings and latent dispositions embedded in these media. In this work, we
explore how Convolutional Neural Networks (CNNs), a now de facto computational
machine learning tool particularly in the area of Computer Vision, can be
specifically applied to the task of visual sentiment prediction. We accomplish
this through fine-tuning experiments using a state-of-the-art CNN and via
rigorous architecture analysis, we present several modifications that lead to
accuracy improvements over prior art on a dataset of images from a popular
social media platform. We additionally present visualizations of local patterns
that the network learned to associate with image sentiment for insight into how
visual positivity (or negativity) is perceived by the model. | ['Xavier Giro-i-Nieto', 'Brendan Jou', 'Victor Campos'] | 2016-04-12 | null | null | null | null | ['visual-sentiment-prediction'] | ['computer-vision'] | [ 1.57526597e-01 9.23878625e-02 -1.54778391e-01 -4.09288138e-01
1.90177828e-01 -5.67116380e-01 8.58599365e-01 3.39859217e-01
-1.37696907e-01 3.56282234e-01 7.20495343e-01 -2.22129479e-01
4.27580714e-01 -6.38817549e-01 -5.45480251e-01 -3.86398584e-01
-2.13261038e-01 -1.63948044e-01 -3.76261562e-01 -6.81369901e-01
2.01049760e-01 4.35783654e-01 -1.66153693e+00 9.57118750e-01
1.05716623e-02 1.15880167e+00 -3.51100117e-01 6.68812275e-01
-5.21501601e-02 1.18146741e+00 -4.11626965e-01 -7.24224925e-01
-1.91548213e-01 -1.01594023e-01 -5.49318016e-01 4.04626206e-02
6.95384622e-01 -9.66702551e-02 -1.82948038e-01 8.72597873e-01
2.94149339e-01 -1.82629019e-01 6.72964990e-01 -1.19532037e+00
-1.03429866e+00 2.86234081e-01 -7.42925227e-01 5.56929529e-01
3.81570011e-01 2.28168577e-01 9.84924495e-01 -8.73839200e-01
9.45089638e-01 1.37695229e+00 7.35272288e-01 2.38321930e-01
-1.28724861e+00 -5.15529871e-01 2.23390982e-02 1.52646452e-01
-6.43090189e-01 -3.70672792e-01 1.14316225e+00 -9.06933308e-01
9.18088615e-01 2.90878803e-01 1.27350962e+00 1.64161968e+00
1.89149380e-01 8.60478282e-01 1.43305492e+00 -2.53600985e-01
2.23610015e-03 5.26298583e-01 8.89857113e-02 8.84825647e-01
-1.75337002e-01 -6.12791628e-02 -9.08463538e-01 1.48428693e-01
4.48600501e-01 6.54879734e-02 -3.28754000e-02 -4.24551725e-01
-1.14843762e+00 9.42246020e-01 9.13405001e-01 2.39114523e-01
-4.76482779e-01 3.98088813e-01 6.30445182e-01 2.16811746e-01
1.02469349e+00 3.47371668e-01 -2.57791698e-01 -1.15792684e-01
-7.46351421e-01 5.39904162e-02 4.44444805e-01 2.74474591e-01
9.09764230e-01 4.32313196e-02 -2.06761271e-01 7.61131108e-01
3.33767831e-01 3.32205385e-01 2.84434468e-01 -4.39332008e-01
-2.54160631e-02 9.32585239e-01 -3.16976272e-02 -1.69866467e+00
-7.32367039e-01 -3.43315303e-01 -6.79694831e-01 5.39792240e-01
2.89220840e-01 -2.69297719e-01 -7.57998705e-01 1.58741534e+00
3.11027437e-01 -3.19783360e-01 -4.17449564e-01 8.39478374e-01
9.35604990e-01 6.36344790e-01 4.39699322e-01 -1.36159882e-02
1.45386326e+00 -6.05685532e-01 -7.44890869e-01 -4.59914893e-01
4.52078581e-01 -8.13961923e-01 1.40637040e+00 4.32255596e-01
-1.19039679e+00 -5.68289340e-01 -1.27125537e+00 -2.98366517e-01
-9.22343314e-01 1.91051550e-02 8.01985383e-01 5.98483324e-01
-1.18687356e+00 5.21716237e-01 -6.44112408e-01 -5.95833957e-01
9.77003396e-01 -5.86883584e-03 -3.82549077e-01 4.62721676e-01
-8.09203267e-01 9.31024373e-01 -1.30514398e-01 1.10768862e-01
-5.41001916e-01 -7.51934707e-01 -8.25110376e-01 9.39080864e-03
-2.26151988e-01 -4.97826695e-01 1.02366042e+00 -1.98787367e+00
-1.03781950e+00 1.37677908e+00 -4.65689190e-02 -1.65906847e-01
2.40323484e-01 -2.05000281e-01 -5.88683724e-01 3.77426416e-01
-1.05735257e-01 9.19171512e-01 9.74407315e-01 -1.34016073e+00
-1.46387279e-01 -3.73302490e-01 8.17575380e-02 -1.81506768e-01
-1.19112957e+00 2.30337679e-01 -1.65810868e-01 -6.32477999e-01
-3.03030670e-01 -7.06431210e-01 1.08540311e-01 3.15920442e-01
-3.52602541e-01 1.56083465e-01 1.05812061e+00 -6.24825358e-01
1.00464296e+00 -2.19936013e+00 9.97912511e-02 4.06584501e-01
5.36442518e-01 6.22987635e-02 6.74295872e-02 5.03138781e-01
-3.29758734e-01 2.69594043e-01 1.43968120e-01 -5.44315398e-01
1.70800582e-01 -4.12000000e-01 -3.04332823e-01 6.35743916e-01
2.24033147e-01 1.20246434e+00 -9.05679107e-01 -3.97286177e-01
3.07513535e-01 9.90868449e-01 -4.21528280e-01 -1.41317053e-02
-3.62354577e-01 3.90893817e-01 -4.96441312e-02 6.64002240e-01
4.24803227e-01 -5.83185375e-01 2.51108050e-01 -5.93864799e-01
-1.98159203e-01 -9.68153700e-02 -2.49791652e-01 1.33750737e+00
-4.33605969e-01 1.49973738e+00 1.94860902e-02 -6.93914235e-01
8.36048245e-01 5.32967597e-02 3.33496749e-01 -1.11858749e+00
5.36483407e-01 -5.33333540e-01 -3.48416567e-01 -8.91613245e-01
6.09565556e-01 -3.37610662e-01 1.09725613e-02 5.05040348e-01
5.41506195e-03 4.24051702e-01 -9.14633349e-02 2.95868993e-01
6.48862362e-01 -2.79801618e-03 1.23942539e-01 -2.99289674e-01
4.79456931e-02 3.02244839e-03 -1.28966048e-01 1.15591094e-01
-3.41780543e-01 5.46839356e-01 9.81184721e-01 -1.12165737e+00
-1.19627905e+00 -7.44682372e-01 -1.37552945e-02 1.44097161e+00
-9.15847495e-02 -4.76261705e-01 -4.30925041e-01 -2.91477293e-01
-8.52179825e-02 2.96081990e-01 -1.36461544e+00 -2.70759650e-02
-2.50443846e-01 -7.57658184e-01 1.78074628e-01 5.62146544e-01
2.10169822e-01 -1.63164699e+00 -8.21655810e-01 -1.40352145e-01
6.50158152e-02 -1.10791743e+00 2.33256757e-01 -6.82001188e-02
-3.34768891e-01 -8.69928837e-01 -5.27854860e-01 -7.24860609e-01
5.46439111e-01 1.34033635e-01 1.57396197e+00 2.40146294e-01
-4.40174818e-01 4.79585767e-01 -2.68375993e-01 -5.90924561e-01
-3.79796475e-01 -8.73035416e-02 -3.85795623e-01 3.76685053e-01
5.54845572e-01 -7.07437992e-01 -9.92392361e-01 -2.26085082e-01
-1.01122379e+00 4.21669900e-01 2.40463048e-01 3.78467858e-01
1.48860157e-01 -6.88827991e-01 2.64271200e-01 -1.01248014e+00
7.07284927e-01 -8.00821722e-01 -1.64917588e-01 -4.82862145e-02
-2.02838913e-01 -4.39586341e-01 3.28887045e-01 -2.89345562e-01
-9.88576114e-01 -5.54261655e-02 1.34506568e-01 -4.74317491e-01
-2.56999940e-01 6.56110466e-01 3.06122899e-01 -1.27509581e-02
8.29099476e-01 -4.23892081e-01 1.46401346e-01 -4.42630239e-02
6.04181170e-01 4.33894038e-01 3.79114300e-01 -6.11821413e-02
4.12957788e-01 1.10019088e+00 4.82001565e-02 -8.54310811e-01
-1.05776072e+00 -3.09976161e-01 -5.12060940e-01 -9.23234642e-01
1.23762310e+00 -9.32951927e-01 -1.10017347e+00 2.56581008e-01
-8.38700175e-01 -5.95550716e-01 2.06477642e-02 -2.00391635e-01
-4.16631877e-01 -5.29744383e-03 -6.44630194e-01 -7.02532411e-01
-3.17931861e-01 -5.86052358e-01 8.27494562e-01 2.15096250e-01
-8.21634710e-01 -1.31100321e+00 2.89097577e-01 3.16623956e-01
5.56200802e-01 8.72674465e-01 9.55258071e-01 -2.35055044e-01
-7.33759180e-02 -9.58710909e-02 -5.78367889e-01 9.50043723e-02
-2.72706538e-01 4.06592607e-01 -1.47146404e+00 -1.87270001e-01
-3.76978159e-01 -7.56825507e-01 9.35883701e-01 3.41685414e-01
1.22192240e+00 -4.84115630e-01 -1.96612373e-01 6.30288839e-01
1.43766510e+00 -3.35072875e-01 7.59445727e-01 5.81561804e-01
9.92267728e-01 9.36632454e-01 2.50975072e-01 5.35751939e-01
3.50950509e-01 4.40007627e-01 7.59096980e-01 -7.49056876e-01
-6.27448261e-02 -1.66375816e-01 3.36694419e-01 4.05952543e-01
-5.93521893e-02 -9.18727592e-02 -1.06970501e+00 6.48326635e-01
-1.82977343e+00 -1.07741916e+00 -2.79446721e-01 1.53800464e+00
5.01024842e-01 1.04808681e-01 4.13426310e-01 -1.89023197e-01
4.60292310e-01 5.50353825e-01 -1.59251511e-01 -8.79123628e-01
-3.14494133e-01 2.56708860e-01 2.04480156e-01 1.41073555e-01
-1.19709635e+00 8.41451764e-01 6.65590143e+00 1.49369299e-01
-1.71830702e+00 -2.75453348e-02 1.13829243e+00 -4.43763822e-01
-3.99740577e-01 -4.81944561e-01 -7.09029287e-02 1.86290488e-01
8.57434392e-01 5.97409189e-01 3.04345727e-01 8.08554173e-01
2.89134800e-01 8.08347464e-02 -1.03001797e+00 1.02622020e+00
3.48021120e-01 -1.74963057e+00 -6.30396083e-02 -3.47856022e-02
9.28584576e-01 -2.55493429e-02 8.12107980e-01 -1.01437792e-01
2.32397369e-03 -1.25043690e+00 9.77383018e-01 6.35764956e-01
7.67421782e-01 -7.90591836e-01 3.42406273e-01 -4.38874245e-01
-6.88838482e-01 -2.12575704e-01 -1.03684776e-01 -4.88301218e-01
1.01156577e-01 5.86111665e-01 -7.76283264e-01 -3.98561895e-01
1.14168823e+00 1.25711477e+00 -8.96564424e-01 4.95231450e-01
-1.26596838e-01 4.12886798e-01 3.23092312e-01 -3.65215272e-01
2.30773196e-01 1.08247593e-01 2.46855661e-01 1.75407803e+00
1.92780904e-02 -3.09197128e-01 -3.47633868e-01 9.91024256e-01
-1.67918384e-01 2.00647593e-01 -8.33510518e-01 -2.92095542e-01
-1.65447295e-01 1.89793897e+00 -1.07081425e+00 -3.91683429e-01
-3.55919391e-01 8.14290226e-01 6.54325306e-01 5.20681858e-01
-5.95840216e-01 1.66710377e-01 7.76681304e-01 2.65995115e-01
3.15860927e-01 -1.63952205e-02 -5.33805072e-01 -9.59703088e-01
-1.22003362e-01 -9.02876794e-01 2.83021182e-02 -1.46775115e+00
-1.48267961e+00 4.51353908e-01 -4.02002066e-01 -1.00190258e+00
-1.44778378e-02 -8.66230130e-01 -7.05859303e-01 4.61609006e-01
-1.38925409e+00 -1.60624802e+00 -5.91685176e-01 6.42673731e-01
3.60235795e-02 1.07238814e-01 7.70177066e-01 6.90339729e-02
-4.95300680e-01 1.93432242e-01 -3.51861924e-01 1.68175206e-01
7.20105827e-01 -1.20561206e+00 2.51910597e-01 5.03331721e-01
1.54777706e-01 3.77721637e-01 8.73725176e-01 -4.73376065e-01
-1.21017039e+00 -8.31843019e-01 5.30300856e-01 -5.95652223e-01
1.03863633e+00 -7.14786589e-01 -5.43024659e-01 9.73114550e-01
7.97665894e-01 -1.55311793e-01 1.07661271e+00 4.89499539e-01
-5.86921632e-01 1.01520047e-01 -8.48740160e-01 8.72681260e-01
7.04151094e-01 -8.16537321e-01 -2.85490066e-01 2.38054216e-01
3.38071942e-01 -9.71795842e-02 -5.54006815e-01 1.25870243e-01
9.27534819e-01 -1.37173653e+00 1.15400839e+00 -6.67227030e-01
1.34247613e+00 7.23293722e-02 1.19864471e-01 -1.34086347e+00
-2.46093705e-01 -4.96351957e-01 -8.64036977e-02 1.13730729e+00
4.34302866e-01 -1.69777304e-01 6.94343925e-01 4.05537993e-01
3.68799567e-02 -6.80865705e-01 -2.79576927e-01 3.49974185e-01
-8.10013637e-02 -5.39350688e-01 1.40422806e-01 1.28614712e+00
3.13785672e-01 4.69370395e-01 -4.61828768e-01 -2.06695080e-01
1.79891646e-01 -2.40931399e-02 7.62798607e-01 -1.05424762e+00
2.59402879e-02 -7.41087854e-01 -5.24817288e-01 -2.42051259e-01
-3.63109680e-03 -6.40951276e-01 -4.08495396e-01 -1.39503622e+00
3.86827886e-01 5.34760989e-02 -5.10095000e-01 4.94502753e-01
1.80302024e-01 1.09915733e+00 3.01617175e-01 -2.59250075e-01
-7.09131837e-01 2.25004211e-01 1.31833029e+00 -2.60590494e-01
-8.16065297e-02 -7.23822355e-01 -9.76766527e-01 1.05649769e+00
8.07865381e-01 -4.80902046e-02 -1.89984694e-01 -1.71705976e-01
1.43327928e+00 -3.91469985e-01 7.50990629e-01 -7.54790843e-01
-1.31008595e-01 1.41501114e-01 1.09518719e+00 -5.58704674e-01
6.79212868e-01 -6.40265107e-01 -1.19844675e-01 1.34937569e-01
-5.89434087e-01 2.67301589e-01 5.09592474e-01 3.64334196e-01
-1.40462443e-01 5.61181724e-01 6.46758080e-01 -1.36088923e-01
-7.38827109e-01 -5.45161664e-02 -5.04657507e-01 -2.01459348e-01
8.45598221e-01 -1.28180906e-01 -7.08674133e-01 -7.00548112e-01
-9.91492450e-01 -1.45551383e-01 7.09325552e-01 5.45905352e-01
6.07776821e-01 -1.30202138e+00 -3.48444611e-01 1.51638851e-01
2.41031364e-01 -7.05805123e-01 4.49206829e-01 8.97305608e-01
-4.98012424e-01 -5.22597991e-02 -8.11616778e-01 -6.48442507e-01
-1.44274437e+00 7.31271982e-01 8.75735134e-02 1.58217221e-01
-4.39003170e-01 8.67955923e-01 4.84956622e-01 -4.52106670e-02
-3.12050316e-03 -2.31906205e-01 -7.44860709e-01 7.06887126e-01
6.62305057e-01 1.72788218e-01 -8.15054774e-02 -9.02715087e-01
-2.01493546e-01 4.58950609e-01 1.79516539e-01 -1.10877216e-01
1.69198990e+00 -2.41633698e-01 -3.11389893e-01 8.08130920e-01
1.56991923e+00 2.64112446e-02 -1.33448768e+00 3.04927945e-01
-3.21550131e-01 -3.26290637e-01 1.31526932e-01 -9.87799823e-01
-1.25399232e+00 1.27894235e+00 7.66707540e-01 7.30849028e-01
1.06453335e+00 1.19685933e-01 5.02409518e-01 1.04966305e-01
-4.16745812e-01 -1.10991716e+00 7.40131140e-01 2.01337084e-01
9.28220689e-01 -1.48037910e+00 1.29797846e-01 -1.03728220e-01
-8.61861408e-01 1.35945189e+00 4.78109330e-01 -3.05675536e-01
7.64995337e-01 7.61283003e-03 4.23733830e-01 -9.21908200e-01
-7.38871038e-01 -1.23896815e-01 5.83230197e-01 4.76681441e-01
6.36316538e-01 2.15259805e-01 1.52294412e-01 2.92016715e-01
-4.82743949e-01 -1.13110468e-01 6.22312546e-01 7.19140530e-01
-2.33215854e-01 -3.91308904e-01 -2.95473903e-01 2.39844754e-01
-7.83285141e-01 -7.60832578e-02 -8.23735952e-01 7.82786965e-01
2.18785360e-01 6.65718913e-01 5.40633976e-01 -4.75053251e-01
-8.27312544e-02 3.98578937e-04 3.76262367e-01 -2.59922236e-01
-8.80131006e-01 7.76880234e-02 3.00397187e-01 -8.70718598e-01
-9.53048289e-01 -5.27372360e-01 -7.94965744e-01 -5.60424089e-01
4.35145944e-01 -7.40982294e-01 9.21086669e-01 7.60708094e-01
2.70200104e-01 6.71993375e-01 3.27256054e-01 -1.14012718e+00
6.21339798e-01 -8.72278512e-01 -2.51052439e-01 9.55055058e-01
6.96219921e-01 -4.78055805e-01 -2.56323278e-01 3.72915417e-01] | [11.04300308227539, 2.7011077404022217] |
5fddf519-5de8-4ddf-815a-6d9e7ab77a5f | ad-vo-scale-resilient-visual-odometry-using | 2001.02090 | null | https://arxiv.org/abs/2001.02090v1 | https://arxiv.org/pdf/2001.02090v1.pdf | AD-VO: Scale-Resilient Visual Odometry Using Attentive Disparity Map | Visual odometry is an essential key for a localization module in SLAM systems. However, previous methods require tuning the system to adapt environment changes. In this paper, we propose a learning-based approach for frame-to-frame monocular visual odometry estimation. The proposed network is only learned by disparity maps for not only covering the environment changes but also solving the scale problem. Furthermore, attention block and skip-ordering scheme are introduced to achieve robust performance in various driving environment. Our network is compared with the conventional methods which use common domain such as color or optical flow. Experimental results confirm that the proposed network shows better performance than other approaches with higher and more stable results. | ['Tae-young Chung', 'Sangwon Hwang', 'Kyungjae Lee', 'Woo Jin Kim', 'Sangyoun Lee', 'Joosung Lee', 'Junhyeop Lee'] | 2020-01-07 | null | null | null | null | ['monocular-visual-odometry'] | ['robots'] | [-4.06288773e-01 -4.54036027e-01 -1.68286845e-01 -5.83669305e-01
1.20712809e-01 -2.97241330e-01 3.25453758e-01 -3.85194093e-01
-7.25508988e-01 1.04392302e+00 -6.07858561e-02 -8.15160871e-02
1.14383243e-01 -5.55426598e-01 -5.38948894e-01 -4.29478168e-01
1.13017477e-01 2.74043083e-01 7.07985640e-01 -4.06411886e-01
5.22234023e-01 6.58750653e-01 -1.62214971e+00 -6.57333493e-01
9.37939525e-01 6.08879268e-01 5.80216408e-01 6.41601443e-01
-5.69246989e-03 8.57470751e-01 -3.77539814e-01 3.43269289e-01
6.43345356e-01 -3.20632041e-01 -3.41159403e-01 -2.06454601e-02
5.67698658e-01 -3.22619826e-01 -7.78285325e-01 1.34329116e+00
7.98288226e-01 3.53350103e-01 2.49063835e-01 -1.36954081e+00
-3.27588350e-01 -1.28175244e-01 -5.17038167e-01 3.33760172e-01
4.64541137e-01 3.20121884e-01 3.83305520e-01 -8.32625210e-01
7.82669544e-01 1.31980443e+00 5.89663029e-01 2.79789597e-01
-6.52858675e-01 -6.81910753e-01 1.21075779e-01 7.73850560e-01
-1.51898623e+00 -4.16736484e-01 7.26980269e-01 -3.50334793e-01
9.69217122e-01 -3.03379685e-01 7.80713618e-01 2.25631818e-01
5.52642167e-01 3.75456750e-01 1.17056012e+00 -3.13719004e-01
-2.87126824e-02 1.18539762e-02 -6.19777851e-02 9.22923863e-01
7.35934198e-01 2.60313272e-01 -6.81051672e-01 4.39860910e-01
8.40788484e-01 2.43079849e-02 -4.80907530e-01 -8.99649382e-01
-1.33531058e+00 6.15308106e-01 7.61586726e-01 -1.27825946e-01
-3.08239341e-01 3.65529865e-01 2.89987028e-01 3.39764446e-01
-2.22104485e-03 1.48688510e-01 -1.97670847e-01 -3.24619472e-01
-7.14570701e-01 1.50006920e-01 5.98492026e-01 1.27962875e+00
1.42064309e+00 5.55379272e-01 3.23435754e-01 4.87834305e-01
6.02919757e-01 7.13811636e-01 7.51394510e-01 -9.25077200e-01
4.05775368e-01 5.58460236e-01 3.46145123e-01 -1.22290421e+00
-7.12421536e-01 -3.71428281e-01 -6.65838003e-01 5.45478702e-01
9.28386226e-02 -2.87138462e-01 -1.15696454e+00 1.43028057e+00
5.03516018e-01 2.97351003e-01 1.11405589e-01 1.25551903e+00
7.64912426e-01 5.06529391e-01 -5.12542188e-01 -6.34742901e-02
9.63714182e-01 -1.12683666e+00 -1.12407136e+00 -6.63252354e-01
3.96962166e-01 -9.17917311e-01 5.96530199e-01 2.01276541e-01
-6.73188627e-01 -8.20567012e-01 -1.54729891e+00 -2.11038083e-01
-3.60215724e-01 1.00587621e-01 6.15195155e-01 5.58868647e-01
-1.19044948e+00 2.42621288e-01 -7.13300049e-01 -7.23138154e-01
-1.46407798e-01 5.88422477e-01 -4.47528422e-01 -1.56430677e-01
-1.34947503e+00 1.39395010e+00 6.91483676e-01 2.94990540e-01
-7.21735060e-01 4.18879874e-02 -1.29829931e+00 -4.19555098e-01
1.39780477e-01 -7.25381076e-01 1.02542520e+00 -5.88993788e-01
-1.97142529e+00 4.82733071e-01 -7.07368970e-01 -6.13751113e-01
5.36768258e-01 -1.46341860e-01 -2.57393241e-01 -7.89947063e-02
-4.35838476e-03 8.10884476e-01 6.75427496e-01 -1.15354681e+00
-1.05069482e+00 -2.01778322e-01 2.63878882e-01 8.54366302e-01
3.72543931e-01 -6.14522934e-01 -5.66989660e-01 1.31487578e-01
9.37090158e-01 -1.13219702e+00 -4.14504409e-01 9.55002755e-02
4.08187741e-03 2.17260897e-01 8.72292936e-01 -2.18119100e-01
9.40202951e-01 -1.77292573e+00 -1.04188561e-01 -3.02124936e-02
4.84968536e-02 1.96524009e-01 3.52274418e-01 2.86580563e-01
3.82025331e-01 -6.41986966e-01 3.21502954e-01 -3.45095515e-01
-4.39341888e-02 4.75026935e-01 -1.11804216e-03 1.02501857e+00
-2.79253811e-01 6.29003108e-01 -8.19458544e-01 -4.89379108e-01
7.02511907e-01 4.42003280e-01 -4.51325983e-01 1.14805132e-01
2.98488349e-01 8.65710974e-01 -1.89613700e-01 4.89163369e-01
1.12446201e+00 1.74567118e-01 -1.89456925e-01 -2.84025490e-01
-4.91131127e-01 2.92708874e-01 -1.62112010e+00 2.00828552e+00
-5.28254151e-01 9.89258051e-01 -1.07829779e-01 -7.28036582e-01
1.18863606e+00 -2.34740853e-01 1.32141620e-01 -8.25108409e-01
2.87191838e-01 3.75995815e-01 1.75291166e-01 -5.54644823e-01
9.52300012e-01 7.90204555e-02 1.80038437e-01 -1.54428944e-01
1.82169229e-02 -3.53876948e-01 2.50279993e-01 3.43879312e-02
6.57505155e-01 5.71454585e-01 6.26859248e-01 -2.93901920e-01
1.04056561e+00 2.12898239e-01 1.08929300e+00 6.05260491e-01
-8.37699115e-01 3.03200454e-01 -1.07172005e-01 -6.20590627e-01
-9.16604340e-01 -7.92899609e-01 -6.46608025e-02 3.10136378e-01
1.27279317e+00 2.02356968e-02 -2.20500365e-01 -3.06858122e-01
1.98574275e-01 -2.80498527e-02 -1.96615294e-01 -1.51395202e-01
-7.51177728e-01 -4.55000222e-01 1.96758211e-01 2.54921138e-01
1.12945962e+00 -6.24887705e-01 -8.22974503e-01 2.33901754e-01
-2.75817662e-01 -1.31081748e+00 -4.33572799e-01 4.38069291e-02
-8.57224107e-01 -1.04146290e+00 -3.68198603e-01 -8.45540166e-01
6.42451346e-01 8.93347085e-01 6.12280309e-01 -1.01373196e-01
-3.52303460e-02 4.63221073e-02 -1.05805278e-01 -3.80884111e-01
5.48599325e-02 1.47237433e-02 4.23440963e-01 -9.18633938e-02
4.53773707e-01 -4.78335589e-01 -8.58427048e-01 4.48230058e-01
-3.86933476e-01 4.08692434e-02 6.51846230e-01 8.13050926e-01
3.80880117e-01 -3.25051427e-01 2.41195843e-01 -2.69448221e-01
3.24679613e-01 -1.56541705e-01 -1.05065608e+00 -2.05556124e-01
-9.11481619e-01 2.03591496e-01 3.13554555e-01 -1.76136151e-01
-1.07830524e+00 4.33201373e-01 1.22395396e-01 -2.70209432e-01
-1.19435750e-01 2.83222526e-01 6.36051968e-02 -8.71797681e-01
6.11938417e-01 4.04704422e-01 1.35094315e-01 1.30294310e-02
1.85151741e-01 7.04747319e-01 6.46040618e-01 -4.08277847e-02
1.06778419e+00 8.59777927e-01 4.00112301e-01 -6.73450410e-01
-2.40346536e-01 -8.42827559e-01 -9.42686856e-01 -3.69550824e-01
6.80592597e-01 -1.38274968e+00 -1.03616738e+00 4.64376628e-01
-1.22393382e+00 1.50058910e-01 2.39087343e-01 1.10608470e+00
-5.90545118e-01 4.73934531e-01 -4.65240210e-01 -7.15023816e-01
1.59298796e-02 -1.25370800e+00 6.72765374e-01 6.18456900e-01
3.53263199e-01 -1.00734448e+00 4.51329857e-01 -1.55241728e-01
4.82495040e-01 1.76964775e-01 -1.78176865e-01 9.05263424e-02
-1.12463224e+00 -8.53693783e-02 -3.64658594e-01 -1.32711112e-01
2.86602080e-01 -3.22670966e-01 -8.37640107e-01 -4.33367580e-01
-1.53663740e-01 -3.35274599e-02 8.30526412e-01 3.34145159e-01
2.43785888e-01 1.82097688e-01 -3.70857418e-01 1.15383029e+00
1.81407821e+00 5.10493815e-01 7.27568567e-01 8.96647692e-01
8.94246995e-01 4.29230809e-01 9.61117923e-01 2.93361783e-01
8.48161578e-01 6.59913361e-01 6.64454818e-01 -3.47721064e-03
-1.84133984e-02 -2.50820398e-01 3.22705150e-01 8.13797772e-01
1.46831617e-01 1.10852458e-01 -7.70549119e-01 6.69432342e-01
-2.22232413e+00 -7.98286796e-01 -2.42418498e-01 1.92642236e+00
3.62190038e-01 2.80297935e-01 -3.93909484e-01 -1.38006926e-01
7.54683554e-01 4.16693419e-01 -6.79271877e-01 -3.77866864e-01
-1.36112109e-01 -2.23838165e-01 1.16231751e+00 9.96694982e-01
-9.30712402e-01 1.23359549e+00 6.46103573e+00 2.09975168e-01
-1.43336821e+00 1.01608053e-01 -4.30480391e-01 6.39728382e-02
1.63663507e-01 3.63376498e-01 -1.13316047e+00 3.07164639e-01
4.49779242e-01 -3.54013085e-01 3.43654990e-01 8.22319210e-01
4.16082203e-01 -6.35203242e-01 -5.00700355e-01 1.38882446e+00
1.76273257e-01 -1.13566673e+00 -3.24282676e-01 -1.07607611e-01
8.91176701e-01 4.49973494e-01 -3.66927981e-01 1.04862094e-01
4.27615643e-01 -5.03943563e-01 6.27030551e-01 4.30325359e-01
5.26153326e-01 -6.91738844e-01 9.50421870e-01 3.76781195e-01
-1.61300993e+00 6.22714534e-02 -7.40705788e-01 -6.72371447e-01
3.75552654e-01 4.05849665e-01 -9.28180754e-01 7.91834950e-01
5.55620492e-01 1.18922508e+00 -5.09095490e-01 1.51199698e+00
-3.74084800e-01 -6.65390790e-02 -2.33383819e-01 -2.11142614e-01
3.78997028e-01 -2.19964519e-01 5.08521557e-01 1.04257774e+00
4.55425501e-01 -2.44992390e-01 3.47028166e-01 3.37181985e-01
4.22718883e-01 1.50620397e-02 -1.15900993e+00 7.44748950e-01
4.63742822e-01 1.01633203e+00 -2.62570113e-01 -4.36182916e-01
-3.74935180e-01 8.68479013e-01 2.07936212e-01 5.59784710e-01
-6.65397763e-01 -7.66185760e-01 9.05287087e-01 -1.95965290e-01
2.04565376e-01 -8.10143769e-01 7.21925348e-02 -1.37906480e+00
-7.11144879e-02 -2.95407295e-01 -9.88921151e-02 -9.67802048e-01
-5.51952362e-01 5.58286786e-01 -1.74777359e-01 -1.76289761e+00
-3.07651758e-01 -5.80713272e-01 -2.27035746e-01 1.10849309e+00
-2.26537395e+00 -1.00358415e+00 -9.83427823e-01 7.61695147e-01
4.61842120e-01 -1.42247334e-01 1.70333907e-01 4.27068502e-01
-2.21037313e-01 2.44117565e-02 1.33609220e-01 -1.74265563e-01
1.13718331e+00 -1.13724065e+00 2.83408672e-01 1.29026651e+00
-1.98428154e-01 7.21397400e-01 9.22674477e-01 -6.58912420e-01
-1.40548265e+00 -8.07292342e-01 8.39973450e-01 -6.02433197e-02
3.48567367e-01 -1.32043600e-01 -6.11267030e-01 8.53454173e-01
4.28208381e-01 1.66798279e-01 -1.81158558e-01 -3.69309515e-01
1.33201510e-01 -4.65407908e-01 -1.09061444e+00 4.31728274e-01
1.11816204e+00 -4.10804480e-01 -4.27392006e-01 -7.91800991e-02
6.50171518e-01 -1.18196166e+00 -2.93160379e-01 4.39988494e-01
6.54878080e-01 -1.28566194e+00 7.47857094e-01 -2.79340427e-04
-3.57344478e-01 -1.18386698e+00 -7.54024833e-02 -1.31942499e+00
-5.24391890e-01 -6.16045296e-01 9.94090885e-02 6.66488647e-01
-1.26980886e-01 -1.18436480e+00 9.34685051e-01 -3.34680416e-02
-2.43946522e-01 -1.39882714e-01 -1.05862987e+00 -8.06917489e-01
-6.32322431e-01 -1.23930365e-01 5.83468974e-01 9.01173174e-01
-2.62169093e-01 1.65650561e-01 -7.44449914e-01 5.67869186e-01
7.61575758e-01 -1.36414424e-01 1.31854701e+00 -1.06134713e+00
1.32982686e-01 2.20585745e-02 -1.22539544e+00 -1.52924407e+00
1.78049922e-01 -4.76558715e-01 4.90901738e-01 -1.69752753e+00
-2.88137108e-01 -2.50729144e-01 -2.69981116e-01 -1.68263376e-01
-6.69064596e-02 3.25864464e-01 3.95279983e-03 3.91866207e-01
-5.97979903e-01 5.76881707e-01 1.40461373e+00 6.32639378e-02
-3.54110658e-01 -1.43069342e-01 4.96510752e-02 7.77033508e-01
8.28595817e-01 -3.56432617e-01 -6.15723789e-01 -7.01204062e-01
1.33587971e-01 -9.57524478e-02 2.47198060e-01 -1.47243094e+00
8.44437659e-01 -3.82194757e-01 3.61142486e-01 -1.10225165e+00
4.01171505e-01 -8.33026886e-01 9.25295278e-02 9.71717775e-01
3.67202878e-01 5.05948901e-01 1.71875358e-01 5.90634406e-01
-5.88198245e-01 -2.47232810e-01 7.96909869e-01 -1.62445679e-01
-1.67054367e+00 4.02866989e-01 -1.83573201e-01 -2.25485682e-01
9.87590730e-01 -6.14444494e-01 -6.18930399e-01 -6.17721498e-01
-2.61696219e-01 4.80356812e-01 6.80281758e-01 4.32551473e-01
7.22367167e-01 -1.51222777e+00 -1.44928992e-01 4.43952322e-01
3.80815029e-01 1.66280463e-01 1.79459363e-01 9.14182782e-01
-1.31154370e+00 7.50573695e-01 -6.44392014e-01 -9.60872352e-01
-9.66795743e-01 1.70851678e-01 6.87112391e-01 1.50286868e-01
-4.23206002e-01 4.64696407e-01 1.07662380e-01 -6.33126557e-01
1.13135874e-01 -3.17369491e-01 -1.98102936e-01 -3.66522759e-01
2.69641817e-01 3.74438256e-01 -2.53867775e-01 -9.43851352e-01
-5.74457705e-01 1.07789087e+00 3.43760788e-01 -3.39235008e-01
7.76254833e-01 -9.02181566e-01 -8.49539712e-02 5.20375788e-01
1.02302527e+00 -1.44647509e-02 -1.57197213e+00 -3.67430747e-01
-1.09423935e-01 -8.80097449e-01 -2.19612438e-02 -1.91917136e-01
-6.96322560e-01 8.26806426e-01 9.63236570e-01 -4.01794106e-01
8.13623488e-01 -7.40837514e-01 5.77996492e-01 6.45242333e-01
8.06906700e-01 -1.13066888e+00 -5.26222885e-02 1.00213039e+00
4.95164305e-01 -1.69681048e+00 3.19327712e-01 -1.20868683e-01
-4.52125341e-01 1.13824642e+00 1.11832130e+00 -4.09099102e-01
5.70664227e-01 1.12545053e-02 5.75049102e-01 1.17596425e-01
-5.34877658e-01 -6.04838908e-01 1.05413064e-01 7.55624473e-01
6.22496307e-02 -4.45880622e-01 -4.35126930e-01 -5.92768610e-01
-9.05656219e-02 1.05095707e-01 8.09260368e-01 1.16326272e+00
-1.00103009e+00 -9.74350333e-01 -4.18605059e-01 -1.76842570e-01
2.04181746e-02 1.70453951e-01 4.87044901e-02 1.08776677e+00
3.12162519e-01 7.33063877e-01 -3.78154591e-02 -3.86531681e-01
4.56964880e-01 -1.82689637e-01 5.69489777e-01 -3.67410153e-01
-4.63142283e-02 -1.12287246e-01 -4.91985003e-04 -8.29966664e-01
-8.35055530e-01 -3.85312498e-01 -1.41219735e+00 -3.80657107e-01
-4.14806157e-01 -1.51395202e-02 9.97630417e-01 8.38322282e-01
3.22942019e-01 4.64298546e-01 5.81484199e-01 -9.04111624e-01
-1.21875748e-01 -9.07254577e-01 -4.63017434e-01 1.46075651e-01
8.66478562e-01 -8.97753060e-01 -3.29018742e-01 -7.77223557e-02] | [7.699239253997803, -2.1064963340759277] |
eb659e9f-67bd-4491-af30-01cb737d179b | efficient-multivariate-sequence | 1409.8211 | null | http://arxiv.org/abs/1409.8211v2 | http://arxiv.org/pdf/1409.8211v2.pdf | Efficient multivariate sequence classification | Kernel-based approaches for sequence classification have been successfully
applied to a variety of domains, including the text categorization, image
classification, speech analysis, biological sequence analysis, time series and
music classification, where they show some of the most accurate results.
Typical kernel functions for sequences in these domains (e.g., bag-of-words,
mismatch, or subsequence kernels) are restricted to {\em discrete univariate}
(i.e. one-dimensional) string data, such as sequences of words in the text
analysis, codeword sequences in the image analysis, or nucleotide or amino acid
sequences in the DNA and protein sequence analysis. However, original sequence
data are often of real-valued multivariate nature, i.e. are not univariate and
discrete as required by typical $k$-mer based sequence kernel functions.
In this work, we consider the problem of the {\em multivariate} sequence
classification such as classification of multivariate music sequences, or
multidimensional protein sequence representations. To this end, we extend {\em
univariate} kernel functions typically used in sequence analysis and propose
efficient {\em multivariate} similarity kernel method (MVDFQ-SK) based on (1) a
direct feature quantization (DFQ) of each sequence dimension in the original
{\em real-valued} multivariate sequences and (2) applying novel multivariate
discrete kernel measures on these multivariate discrete DFQ sequence
representations to more accurately capture similarity relationships among
sequences and improve classification performance.
Experiments using the proposed MVDFQ-SK kernel method show excellent
classification performance on three challenging music classification tasks as
well as protein sequence classification with significant 25-40% improvements
over univariate kernel methods and existing state-of-the-art sequence
classification methods. | ['Pavel P. Kuksa'] | 2014-09-29 | null | null | null | null | ['music-classification'] | ['music'] | [ 8.00601900e-01 -7.99846411e-01 -1.22108437e-01 -2.55325884e-01
-6.47991121e-01 -7.81931937e-01 2.29766384e-01 5.47494829e-01
-6.45179451e-01 6.13723814e-01 -3.98561835e-01 -4.18129504e-01
-3.88240188e-01 -2.71168590e-01 -4.51790929e-01 -1.08311689e+00
-3.39948863e-01 1.57647327e-01 2.03910828e-01 -5.59175722e-02
5.38494647e-01 5.56483448e-01 -1.54427803e+00 2.19257250e-02
7.39214301e-01 1.00939190e+00 5.48370361e-01 1.14727426e+00
-2.78277695e-01 5.22111118e-01 -7.43074834e-01 -3.44837606e-02
-6.94732964e-02 -4.97948408e-01 -4.59066868e-01 -1.31639853e-01
8.37706849e-02 3.33527505e-01 -1.69570729e-01 1.26193285e+00
4.27971512e-01 4.38031763e-01 1.22243619e+00 -1.19833159e+00
-7.27375031e-01 2.66639590e-01 -5.68852425e-01 3.10232580e-01
3.59635592e-01 -2.70285636e-01 8.32131505e-01 -8.09402049e-01
3.70959073e-01 1.00743628e+00 7.16399908e-01 1.40735149e-01
-1.55830455e+00 -7.58743644e-01 -3.48835409e-01 6.26884639e-01
-1.59129214e+00 4.18926263e-03 7.00939894e-01 -8.12592626e-01
1.04219520e+00 2.86637396e-01 2.54226029e-01 9.01918650e-01
3.65308583e-01 6.47443593e-01 9.49580371e-01 -5.56631505e-01
3.06545258e-01 -1.40471727e-01 3.53136420e-01 5.74261665e-01
-2.39699692e-01 -3.69777441e-01 -3.31853896e-01 -8.84510756e-01
6.60692930e-01 1.35949761e-01 -5.51616371e-01 -2.53261566e-01
-1.36565459e+00 9.49757516e-01 -3.77968639e-01 4.86110806e-01
-3.36142451e-01 4.44013700e-02 8.91648412e-01 5.68408668e-01
2.84080535e-01 1.81855887e-01 -4.62705612e-01 -5.11936605e-01
-8.97613883e-01 3.24566633e-01 5.80864727e-01 8.59041810e-01
7.49531507e-01 6.18935972e-02 1.18781447e-01 1.27817702e+00
-1.39150202e-01 5.99901497e-01 9.17949617e-01 -5.20737886e-01
1.38265371e-01 2.62384146e-01 8.67108554e-02 -1.01718056e+00
-5.00444353e-01 -3.64994705e-02 -1.04269588e+00 -2.59975363e-02
4.01066631e-01 5.79166114e-01 -6.13567472e-01 1.67571414e+00
8.35195556e-02 5.92234313e-01 2.62182593e-01 6.53620720e-01
3.87006134e-01 9.63461936e-01 -1.35720283e-01 -4.68047827e-01
1.40738869e+00 -5.15521228e-01 -5.53222597e-01 4.79139805e-01
8.11225355e-01 -1.09294939e+00 1.10143566e+00 6.57222033e-01
-5.40978730e-01 -5.96040845e-01 -1.16289854e+00 3.79685163e-01
-3.98121506e-01 2.54690796e-01 4.88996387e-01 7.33802497e-01
-6.83785021e-01 7.67289698e-01 -5.42501032e-01 -3.78759682e-01
-5.86627796e-02 3.85928124e-01 -6.59825504e-01 -2.31356546e-03
-1.21469963e+00 4.54357326e-01 7.20034659e-01 -3.52490127e-01
-4.41493362e-01 -4.98766065e-01 -8.67241144e-01 -2.07780786e-02
-7.24945916e-03 -7.67949373e-02 9.94339049e-01 -9.80557442e-01
-1.45792270e+00 7.68632948e-01 -1.36741221e-01 -3.16202909e-01
-1.44938663e-01 2.71424562e-01 -6.24137461e-01 3.49919677e-01
-1.09217003e-01 7.37718865e-02 1.17138362e+00 -3.94231766e-01
-4.00865763e-01 -2.49644279e-01 -5.89808464e-01 -1.58900842e-01
-4.10961241e-01 1.79310694e-01 4.10775654e-02 -1.17615032e+00
-1.72251716e-01 -9.88011539e-01 -3.26290056e-02 -3.09333980e-01
9.42296311e-02 -4.76069063e-01 8.92098367e-01 -7.60970831e-01
1.35371697e+00 -2.49247766e+00 4.50366616e-01 3.64816993e-01
-1.57880917e-01 2.83563465e-01 -2.51758963e-01 5.93341649e-01
-5.18871367e-01 -3.52575302e-01 -4.25045371e-01 1.64492905e-01
-3.81943807e-02 3.16582114e-01 -2.63829619e-01 8.17309737e-01
7.24532679e-02 5.49928069e-01 -7.60554373e-01 -2.36186936e-01
1.49102435e-01 4.08113569e-01 5.67291901e-02 1.01850621e-01
-6.04754426e-02 2.32625365e-01 -4.00532395e-01 5.20080209e-01
5.93168736e-01 -3.50293726e-01 2.36525689e-03 -2.88107216e-01
8.70200172e-02 -3.74930710e-01 -1.25255466e+00 1.86541569e+00
-1.55574093e-02 6.82455480e-01 -2.13814408e-01 -1.51038170e+00
1.12236869e+00 4.06899661e-01 7.19649196e-01 -2.80438930e-01
-1.73509926e-01 3.33061486e-01 1.87949732e-01 -4.14705634e-01
3.67533058e-01 -1.85024351e-01 -7.77893737e-02 4.95386422e-01
1.70088306e-01 1.86143920e-01 1.41719013e-01 5.53107709e-02
1.24086642e+00 -2.63375670e-01 7.57420540e-01 -4.42824870e-01
9.92112279e-01 -2.46765360e-01 4.72852677e-01 4.90430802e-01
-2.42813841e-01 4.07395035e-01 4.31534976e-01 -3.78725864e-02
-1.27172863e+00 -1.05193591e+00 -2.66987920e-01 1.18345761e+00
-1.35548964e-01 -4.34289962e-01 -5.91919243e-01 -1.97435290e-01
2.52394646e-01 3.64845842e-01 -5.22842050e-01 -2.75345296e-01
-6.07230783e-01 -8.35424006e-01 1.13544452e+00 6.28024638e-01
-1.67243853e-01 -9.50185001e-01 -4.21646029e-01 5.54817796e-01
6.48860168e-03 -9.06564116e-01 -9.50038493e-01 6.28807843e-01
-7.90329337e-01 -1.01306248e+00 -1.10794783e+00 -9.13156927e-01
1.57029331e-01 1.45094424e-01 4.79890436e-01 -4.87314016e-01
-8.25310707e-01 3.84185702e-01 -5.28514326e-01 -8.95828456e-02
-3.76737803e-01 -2.56411701e-01 5.09352982e-01 9.84315872e-02
5.02974689e-01 -7.20405161e-01 -5.69769859e-01 4.39003080e-01
-1.46635151e+00 -5.36564589e-01 5.71461856e-01 1.23048389e+00
6.44999146e-01 1.46659747e-01 8.79077196e-01 -4.22271252e-01
8.80021036e-01 -3.68189484e-01 -3.82228553e-01 3.36726576e-01
-4.54159498e-01 1.93783298e-01 1.13392067e+00 -1.12068641e+00
-3.18815231e-01 -7.04392567e-02 2.53153518e-02 -7.66921222e-01
-2.30954126e-01 5.83197236e-01 8.92394334e-02 -2.58728653e-01
8.62087429e-01 1.09846854e+00 3.02790523e-01 -4.21757460e-01
1.80377096e-01 1.12099755e+00 6.14035308e-01 -7.04930067e-01
3.12299281e-01 2.20823139e-01 1.50669664e-01 -1.10976911e+00
6.94734901e-02 -1.17198861e+00 -5.95836699e-01 1.25324786e-01
8.71220946e-01 -4.49375600e-01 -1.05993497e+00 5.73953509e-01
-8.52603495e-01 -4.40421365e-02 1.92873374e-01 8.15818965e-01
-1.10456932e+00 1.19243765e+00 -8.04753840e-01 -9.91342962e-01
-4.31906998e-01 -1.20465553e+00 1.16256225e+00 1.04721282e-02
-2.53719717e-01 -8.70367110e-01 1.89372256e-01 -6.12918958e-02
1.60090506e-01 1.86993092e-01 1.68400872e+00 -8.99983943e-01
9.49997231e-02 -3.17241490e-01 2.20892034e-04 6.64883018e-01
3.48451704e-01 -1.75380766e-01 -5.40951729e-01 -5.58506846e-01
-1.26471758e-01 -4.63071346e-01 7.24268138e-01 2.29588747e-01
1.16008914e+00 1.37924924e-01 -3.10112089e-01 4.99618500e-01
1.23762238e+00 7.65637219e-01 3.12237710e-01 8.18730444e-02
3.19511056e-01 3.75952959e-01 7.73453653e-01 7.62221873e-01
-2.23636359e-01 7.43285775e-01 -1.40027359e-01 4.77488816e-01
5.44650018e-01 3.12768191e-01 5.63299954e-01 1.04288399e+00
8.01283494e-02 -2.26049349e-02 -6.21854186e-01 2.08873555e-01
-2.17719674e+00 -1.01430631e+00 -3.43638547e-02 2.37898159e+00
8.45527470e-01 -1.63627043e-01 3.50698560e-01 4.56201077e-01
7.57909834e-01 -6.77394196e-02 -7.48324811e-01 -5.46497345e-01
-3.38132471e-01 5.23996413e-01 4.13425535e-01 -7.49467462e-02
-1.21180022e+00 5.89717865e-01 5.56968498e+00 1.56932640e+00
-1.14654160e+00 -1.34449065e-01 6.96731592e-03 3.76590729e-01
1.41857043e-01 -3.57618898e-01 -4.68824923e-01 5.55698991e-01
9.76424217e-01 -2.20929295e-01 5.01748800e-01 8.49766731e-01
-9.37801749e-02 -6.72958195e-02 -1.07690072e+00 1.67034483e+00
1.24560885e-01 -1.14115179e+00 4.92975488e-02 -2.02381108e-02
3.19987565e-01 -3.84293348e-01 1.79486290e-01 2.38006458e-01
-3.90395015e-01 -1.15384626e+00 3.49303007e-01 5.82636297e-01
1.01381159e+00 -1.01868403e+00 5.22948980e-01 5.19824207e-01
-1.36715329e+00 6.06464967e-02 -7.57556200e-01 2.67382741e-01
-9.17885229e-02 6.99770212e-01 -7.74621487e-01 6.81840479e-01
4.85864908e-01 7.26729095e-01 -1.89974368e-01 8.32778990e-01
7.37565160e-01 7.92605579e-01 -2.03329086e-01 -3.11185062e-01
3.89213741e-01 -5.43837726e-01 5.08002281e-01 1.54595935e+00
4.07604754e-01 2.55364805e-01 2.53562033e-01 4.09249872e-01
3.76471728e-01 6.68729782e-01 -4.75733340e-01 -5.86866319e-01
1.67959496e-01 1.04528201e+00 -7.87449837e-01 -4.52113122e-01
-5.38280249e-01 1.34502828e+00 1.72935143e-01 4.89898503e-01
-7.36510038e-01 -1.04506397e+00 1.22421706e+00 -5.10449588e-01
6.00611329e-01 -5.76157928e-01 3.18344563e-01 -1.15481615e+00
-2.42268369e-02 -8.90432954e-01 4.90516871e-01 -4.29992855e-01
-1.51902151e+00 2.74702907e-01 -1.46762297e-01 -1.37188172e+00
-4.01567161e-01 -1.08117056e+00 -1.55170754e-01 1.11937177e+00
-1.08674037e+00 -6.00362241e-01 2.16879249e-01 9.72439051e-01
4.14044052e-01 -4.80526239e-01 1.26387036e+00 3.51590991e-01
-8.18652436e-02 7.99011290e-01 1.03463829e+00 6.26823381e-02
9.26640332e-01 -1.21707416e+00 2.92290390e-01 1.02070607e-01
1.69757113e-01 7.46312141e-01 5.28729081e-01 -5.68361044e-01
-1.66810715e+00 -7.96604216e-01 2.54611224e-01 -2.92317700e-02
7.65959740e-01 -5.41265607e-01 -1.25973547e+00 1.19708329e-01
-4.54408616e-01 -5.08196354e-02 1.28241050e+00 -4.35938090e-01
-7.05254555e-01 1.16094723e-01 -1.10043287e+00 4.23382431e-01
7.33766139e-01 -8.67791653e-01 -3.77941877e-01 2.94332892e-01
4.84699875e-01 1.16956227e-01 -1.27231133e+00 5.12574375e-01
9.23486054e-01 -6.89713120e-01 1.11439109e+00 -8.02584112e-01
-2.24673331e-01 -6.75605237e-01 -5.91133356e-01 -1.19939804e+00
-5.60753524e-01 -4.08772439e-01 -3.39556336e-02 5.88528991e-01
-1.11255143e-02 -6.63568199e-01 5.86992919e-01 -4.55296636e-01
7.44417682e-02 -6.75831378e-01 -1.17770338e+00 -1.18428731e+00
-7.53073171e-02 -5.46604693e-01 2.13927731e-01 1.11751890e+00
2.73000807e-01 5.77744842e-02 -3.61561626e-01 5.70436679e-02
7.99097300e-01 2.88350761e-01 5.98429322e-01 -1.08145261e+00
-7.64498711e-01 -6.66073561e-01 -1.36302364e+00 -1.08842516e+00
4.23965663e-01 -1.00542092e+00 -5.10999598e-02 -7.11095929e-01
1.68545902e-01 -1.40635133e-01 -5.36884665e-01 7.90147036e-02
-2.82816529e-01 7.25607947e-03 -1.68791637e-01 2.85808355e-01
-2.55617976e-01 4.24261183e-01 7.46933758e-01 -2.90406853e-01
3.60003556e-03 1.33407060e-02 1.37861833e-01 3.03260773e-01
5.07190108e-01 -1.71314627e-01 -4.48713332e-01 4.09778416e-01
-5.48368953e-02 2.65559882e-01 1.78991303e-01 -8.32055569e-01
9.16253328e-02 -2.10954517e-01 4.31332707e-01 -4.29590583e-01
4.14903551e-01 -5.65809906e-01 2.01452017e-01 4.07026559e-01
-2.03815281e-01 2.31465399e-01 9.00758952e-02 1.01739311e+00
-3.58823478e-01 -3.48479986e-01 8.32978189e-01 1.41326502e-01
-7.56900609e-01 2.31829938e-02 -7.16861904e-01 -3.55602741e-01
1.16431916e+00 -5.27007282e-01 6.43698648e-02 -2.74433792e-01
-7.13960230e-01 -2.72285283e-01 3.65798384e-01 3.26996237e-01
7.61196554e-01 -1.29459763e+00 -6.00661397e-01 2.37347797e-01
6.07267261e-01 -6.71055973e-01 2.88228691e-01 1.03687990e+00
-5.42504609e-01 5.85170805e-01 -2.81192154e-01 -8.97171199e-01
-1.69612122e+00 7.82207072e-01 -8.66259113e-02 -4.18627299e-02
-5.60342312e-01 7.10971534e-01 2.64516026e-01 -3.61068606e-01
2.95221776e-01 -4.99922991e-01 -7.73167685e-02 1.53704435e-02
7.60510266e-01 3.37192714e-01 1.07856221e-01 -6.62075460e-01
-4.80203897e-01 8.65226567e-01 -1.79537982e-01 -4.84169573e-02
9.40219641e-01 2.58053005e-01 -2.37050116e-01 9.25988317e-01
1.76575243e+00 -3.16338867e-01 -6.73022628e-01 -3.77192110e-01
3.04456592e-01 -2.67415375e-01 -6.77068055e-01 -3.18286628e-01
-1.63597777e-01 7.98013449e-01 7.21884966e-01 1.67573914e-01
1.19322348e+00 -2.05994900e-02 9.05745685e-01 5.70102930e-01
5.20947278e-01 -9.30725873e-01 -4.63005714e-03 5.75170279e-01
4.90405381e-01 -7.44711041e-01 -3.83392960e-01 -1.37171164e-01
-4.48607564e-01 1.68243372e+00 -4.90338840e-02 4.16798368e-02
7.05334961e-01 3.36434662e-01 -7.51857683e-02 1.96312621e-01
-5.17589331e-01 3.61181423e-03 3.58432382e-01 5.32458901e-01
6.64337695e-01 2.98973829e-01 -5.67736685e-01 5.59411228e-01
1.31993589e-03 -1.43818229e-01 1.52733624e-01 1.07543242e+00
-6.20851874e-01 -1.16941082e+00 -6.55548871e-01 6.31758153e-01
-3.98482919e-01 -1.19758941e-01 -1.16901286e-01 2.22052559e-01
-1.89432770e-01 6.56747162e-01 -2.45660767e-01 -5.86422026e-01
1.23118825e-01 6.01853430e-01 6.15776122e-01 -2.97167540e-01
-3.10257077e-01 2.22681955e-01 -5.05228043e-01 -1.65753469e-01
-3.69571984e-01 -6.44665658e-01 -1.56090701e+00 6.59405217e-02
-3.91724795e-01 4.02488172e-01 7.66261458e-01 7.09890246e-01
1.12356544e-01 1.66498467e-01 5.88422596e-01 -6.95397675e-01
-8.70053232e-01 -1.06404495e+00 -1.16117740e+00 6.62185788e-01
2.81064689e-01 -7.61011899e-01 -2.15323761e-01 2.40483433e-01] | [7.458225250244141, 3.9595859050750732] |
fa18f0d1-ee8d-4aac-9efe-3db8bbbbe2a9 | jointist-joint-learning-for-multi-instrument | 2206.10805 | null | https://arxiv.org/abs/2206.10805v2 | https://arxiv.org/pdf/2206.10805v2.pdf | Jointist: Joint Learning for Multi-instrument Transcription and Its Applications | In this paper, we introduce Jointist, an instrument-aware multi-instrument framework that is capable of transcribing, recognizing, and separating multiple musical instruments from an audio clip. Jointist consists of the instrument recognition module that conditions the other modules: the transcription module that outputs instrument-specific piano rolls, and the source separation module that utilizes instrument information and transcription results. The instrument conditioning is designed for an explicit multi-instrument functionality while the connection between the transcription and source separation modules is for better transcription performance. Our challenging problem formulation makes the model highly useful in the real world given that modern popular music typically consists of multiple instruments. However, its novelty necessitates a new perspective on how to evaluate such a model. During the experiment, we assess the model from various aspects, providing a new evaluation perspective for multi-instrument transcription. We also argue that transcription models can be utilized as a preprocessing module for other music analysis tasks. In the experiment on several downstream tasks, the symbolic representation provided by our transcription model turned out to be helpful to spectrograms in solving downbeat detection, chord recognition, and key estimation. | ['Dorien Herremans', 'Ju-Chiang Wang', 'Amy Hung', 'Minz Won', 'Bochen Li', 'Qiuqiang Kong', 'Keunwoo Choi', 'Kin Wai Cheuk'] | 2022-06-22 | null | null | null | null | ['chord-recognition', 'instrument-recognition'] | ['audio', 'audio'] | [ 3.62075478e-01 -1.95944414e-01 -6.93766475e-02 1.72504053e-01
-1.09742546e+00 -9.96344209e-01 3.08353513e-01 1.94644388e-02
8.40950198e-03 2.93992519e-01 1.98646769e-01 -1.75795350e-02
-4.53283012e-01 -3.28382373e-01 -2.11220786e-01 -6.25628054e-01
1.00454085e-01 2.95404077e-01 -9.76715237e-02 -5.08914649e-01
3.55420738e-01 4.66099381e-01 -1.67499125e+00 4.90757406e-01
5.12470365e-01 1.00412560e+00 1.92782044e-01 1.05184674e+00
-4.80188726e-04 8.00543606e-01 -8.45107377e-01 -2.06859186e-01
2.31854632e-01 -7.84726977e-01 -6.81133986e-01 -7.93603212e-02
2.50083745e-01 -3.89957689e-02 5.24330616e-01 8.57535422e-01
6.40675366e-01 1.35345280e-01 5.46799183e-01 -9.48708177e-01
5.51551841e-02 1.14874256e+00 -2.92162478e-01 -3.51256318e-02
6.41343236e-01 -1.46277383e-01 1.48238456e+00 -7.97037959e-01
3.12770993e-01 9.16237354e-01 7.17791378e-01 1.10095114e-01
-1.16243458e+00 -5.42791784e-01 -1.94206208e-01 -5.18617779e-03
-1.25141490e+00 -6.23899043e-01 1.01733434e+00 -4.53289717e-01
5.36025047e-01 7.62255788e-01 7.58028209e-01 9.15754497e-01
-1.40787050e-01 9.38531935e-01 7.35644162e-01 -6.75132930e-01
3.85423042e-02 -7.23761022e-02 2.42337763e-01 2.49910533e-01
-2.00736374e-01 -7.23900199e-02 -7.58551896e-01 -2.88459092e-01
7.98597932e-01 -4.00260508e-01 -5.22637904e-01 -7.62165785e-02
-1.24406552e+00 5.88841796e-01 -1.85571626e-01 6.17190599e-01
-4.21678931e-01 4.23165821e-02 7.26774752e-01 5.42609274e-01
3.10912337e-02 9.32149231e-01 -3.61146480e-01 -4.49102551e-01
-1.57183135e+00 4.62544233e-01 1.00410366e+00 5.34306884e-01
4.31579888e-01 1.68311268e-01 -2.53811181e-01 9.00690496e-01
-1.98336747e-02 1.62107036e-01 7.53198266e-01 -1.03294313e+00
3.23306590e-01 4.08432603e-01 7.67924115e-02 -7.74847925e-01
-4.57311660e-01 -8.06816161e-01 -6.68378711e-01 1.15054980e-01
5.92593431e-01 -4.91625704e-02 -2.52734929e-01 1.72965980e+00
-2.41305660e-02 1.70611218e-01 -2.09704675e-02 1.04513240e+00
5.65303683e-01 4.51692969e-01 -4.14485425e-01 -5.49899757e-01
1.58501959e+00 -1.02119482e+00 -9.33495462e-01 9.36145857e-02
4.41848844e-01 -1.21537793e+00 1.27483869e+00 1.00765824e+00
-1.34468186e+00 -9.09463048e-01 -1.15573335e+00 -9.27671343e-02
2.02626809e-01 7.25056827e-01 4.83071923e-01 5.59010923e-01
-5.76489091e-01 8.07432413e-01 -4.51805890e-01 -1.45679951e-01
-3.86375695e-01 1.73949480e-01 -7.39207640e-02 6.65167332e-01
-1.03923130e+00 6.05480373e-01 3.59519124e-01 6.38734102e-02
-5.94661772e-01 -3.50808710e-01 -5.75927019e-01 2.99877405e-01
4.03971314e-01 -6.66133463e-01 1.50288105e+00 -1.28009880e+00
-1.92448115e+00 7.68733740e-01 -1.11020297e-01 -3.62668127e-01
4.59328175e-01 -4.56586242e-01 -4.66952920e-01 1.98642954e-01
-1.40046462e-01 -1.00884356e-01 1.29752719e+00 -9.55136716e-01
-5.45173466e-01 -2.91019708e-01 -8.16832706e-02 2.12001830e-01
-1.11644231e-01 1.78876385e-01 -5.87099135e-01 -1.24349296e+00
1.88558325e-01 -1.04054856e+00 1.80746466e-01 -6.90035105e-01
-6.72008038e-01 1.16973281e-01 4.08172756e-01 -7.10779607e-01
1.61438417e+00 -2.39248300e+00 5.86980760e-01 4.59151268e-01
-2.62627929e-01 3.41254622e-02 -1.66188702e-01 7.08836496e-01
-3.92596364e-01 -2.60352045e-01 -5.79004698e-02 -5.08451104e-01
-3.19595411e-02 -1.78168416e-01 -7.75208056e-01 2.18233660e-01
2.31389590e-02 6.23219788e-01 -5.47633290e-01 -2.79762506e-01
-6.34407848e-02 2.82007456e-01 -5.14004588e-01 1.08382948e-01
-2.83640157e-02 6.05131507e-01 -2.17773438e-01 7.24254906e-01
2.08970696e-01 1.42121240e-01 4.11614746e-01 -3.20555687e-01
-2.83553660e-01 4.06449020e-01 -1.81522584e+00 2.07716227e+00
-3.40935618e-01 4.37890530e-01 3.88492346e-01 -7.25994825e-01
9.64176297e-01 6.55966341e-01 4.39804256e-01 -1.36544585e-01
2.21202850e-01 4.09305841e-01 3.30461025e-01 -5.60623586e-01
1.01019633e+00 -1.71044081e-01 -3.04935753e-01 5.48265398e-01
1.71131730e-01 -5.00814974e-01 5.13092816e-01 -1.11928582e-01
7.65977859e-01 3.26166809e-01 6.00944161e-01 -1.10432319e-01
7.51590669e-01 -1.31186113e-01 3.70023608e-01 5.60655892e-01
6.90208599e-02 7.74566352e-01 4.41844851e-01 2.33142711e-02
-5.97959101e-01 -8.87963653e-01 2.64101207e-01 1.45415795e+00
-1.70949280e-01 -9.53802824e-01 -6.52267635e-01 -7.48849064e-02
-2.05452457e-01 4.84277636e-01 -2.95254409e-01 1.94826145e-02
-7.24033594e-01 -3.76923680e-01 9.97314870e-01 5.44527888e-01
-1.50400877e-01 -1.00391269e+00 -6.58477902e-01 3.22628111e-01
-6.01218462e-01 -6.63550854e-01 -6.15127027e-01 4.89093751e-01
-8.44487011e-01 -1.14223647e+00 -6.43223882e-01 -5.71880937e-01
-2.75366157e-01 9.53782648e-02 9.13144648e-01 -1.53810546e-01
7.32906833e-02 2.60671228e-01 -5.03731549e-01 -5.51208913e-01
-4.19674397e-01 3.38367015e-01 1.75066039e-01 1.75495505e-01
-1.75643519e-01 -1.00318778e+00 -2.13848189e-01 2.47662559e-01
-8.54849637e-01 -1.77770350e-02 5.11407793e-01 5.87431729e-01
4.83892739e-01 -7.48796538e-02 6.52017295e-01 -4.91081655e-01
1.04087138e+00 -6.55030832e-02 -3.51750880e-01 1.76135808e-01
-6.98373690e-02 -1.87632293e-02 9.12818551e-01 -6.75987422e-01
-6.60063744e-01 1.96484938e-01 -3.18347096e-01 -3.16944301e-01
7.29231238e-02 6.31719291e-01 -3.16661775e-01 3.11530888e-01
6.80264831e-01 2.70444810e-01 -1.09186165e-01 -8.98615539e-01
4.08470750e-01 7.08719730e-01 9.71470952e-01 -7.34589040e-01
6.42517388e-01 1.81604207e-01 -1.59382746e-01 -8.47485483e-01
-7.23937869e-01 -7.51443446e-01 -6.93250418e-01 -1.90645695e-01
4.19787586e-01 -8.36247325e-01 -9.44147706e-01 2.65754789e-01
-1.19973886e+00 -4.17252593e-02 -5.43983340e-01 6.05503857e-01
-9.61108506e-01 2.97990322e-01 -7.02228367e-01 -9.32940662e-01
-5.18336236e-01 -1.21046567e+00 1.12986124e+00 -1.23025201e-01
-9.03518975e-01 -5.96963108e-01 2.99741864e-01 2.78123200e-01
8.10153596e-03 2.76017804e-02 8.86811793e-01 -6.86438084e-01
-2.27143511e-01 -1.58883795e-01 3.56955886e-01 4.86551076e-01
9.59254727e-02 2.40213290e-01 -1.38788390e+00 -1.90960810e-01
2.42003918e-01 -3.25045586e-01 8.92461300e-01 2.07455114e-01
1.01576674e+00 -1.16169434e-02 2.27784649e-01 6.19358957e-01
1.19356668e+00 7.28619993e-02 4.71353084e-01 9.28494558e-02
4.45344627e-01 6.13455176e-01 6.55054271e-01 5.48271239e-01
-7.52935186e-02 1.05861521e+00 1.59013793e-01 1.34177446e-01
-1.37475222e-01 -1.81069389e-01 6.83881640e-01 1.28810453e+00
-3.93484950e-01 -5.43855913e-02 -6.62775457e-01 4.55638975e-01
-1.96170902e+00 -1.18429160e+00 -2.82802671e-01 2.13595581e+00
8.90047610e-01 7.99057772e-04 6.12206101e-01 9.25095558e-01
5.00996530e-01 3.42774354e-02 -2.82209545e-01 -5.70901453e-01
-1.95921600e-01 5.08042872e-01 -1.33557737e-01 2.87865013e-01
-1.01989162e+00 5.95732331e-01 6.56563616e+00 1.01721454e+00
-1.23432887e+00 -1.45849958e-01 -1.78975254e-01 -2.17072442e-01
-2.26963777e-02 4.59417440e-02 -4.87836510e-01 1.92385346e-01
7.54134178e-01 2.77633499e-02 5.78020334e-01 6.94643557e-01
3.17634851e-01 -9.40570235e-02 -1.32399988e+00 1.18601954e+00
3.18479463e-02 -1.08836186e+00 1.50901452e-01 -1.54796138e-01
2.27292717e-01 -3.70644152e-01 -9.70555022e-02 2.16431141e-01
-3.23633432e-01 -8.36425960e-01 1.32937610e+00 6.45984173e-01
6.95092201e-01 -8.33999097e-01 3.78389657e-01 5.97866833e-01
-1.37482011e+00 -2.11158231e-01 2.48885870e-01 -4.45151865e-01
1.89524576e-01 5.61442077e-01 -7.32819676e-01 1.00418663e+00
2.11049631e-01 4.97968704e-01 -3.69685024e-01 1.01365173e+00
-3.51851135e-01 8.64381790e-01 -2.14084491e-01 5.50094187e-01
-1.19958542e-01 -3.17146868e-01 8.07649016e-01 1.34059334e+00
5.88337302e-01 -2.45923907e-01 2.56853729e-01 6.62712574e-01
2.54883349e-01 6.43528402e-01 -2.85763144e-01 -1.22222252e-01
3.45649660e-01 1.42444885e+00 -9.11884129e-01 -2.23976359e-01
1.65195670e-02 9.81566787e-01 6.33562133e-02 2.76259452e-01
-8.23927879e-01 -5.01450896e-01 6.50081396e-01 -1.88739195e-01
2.82147825e-01 -2.38474384e-01 -4.75308508e-01 -1.22386253e+00
-1.92095600e-02 -1.30154300e+00 2.97059059e-01 -7.64959037e-01
-8.24749291e-01 4.92505789e-01 -3.37514192e-01 -1.56821060e+00
-7.37754464e-01 -3.88123125e-01 -7.07300425e-01 9.41875815e-01
-1.09992659e+00 -1.03306782e+00 -2.22782027e-02 5.41900277e-01
4.95930821e-01 -8.47793669e-02 1.18833661e+00 3.91634852e-01
-5.23130596e-01 5.00226021e-01 -6.02345467e-02 1.17295325e-01
1.00017571e+00 -1.22018325e+00 4.14234735e-02 7.83840418e-01
7.36894310e-01 8.37463677e-01 8.18425417e-01 -3.40340823e-01
-1.31532753e+00 -6.18619263e-01 7.42658675e-01 -2.47130364e-01
5.19902706e-01 -2.45159522e-01 -7.59648442e-01 5.34426153e-01
-3.43780257e-02 -6.25623882e-01 1.04080403e+00 3.98420364e-01
-3.26547474e-01 -1.98521875e-02 -5.49039900e-01 4.14592296e-01
7.70199955e-01 -8.39403152e-01 -9.31342602e-01 -2.21531332e-01
2.88536280e-01 -3.21265966e-01 -8.25765014e-01 2.08683893e-01
9.81105685e-01 -1.16898584e+00 9.23713207e-01 -2.84494787e-01
4.24991250e-01 -5.45855641e-01 -2.58175880e-01 -1.35752773e+00
-2.60221034e-01 -9.31136966e-01 -2.84271926e-01 1.46891820e+00
2.68615782e-01 -8.30739439e-02 2.39611134e-01 -1.66462332e-01
-1.33131504e-01 -2.16104522e-01 -6.27063811e-01 -8.81755769e-01
-2.87061036e-01 -9.78432715e-01 5.70650637e-01 9.09896433e-01
2.25152433e-01 6.87716484e-01 -6.35861516e-01 -1.50666200e-03
1.79995984e-01 7.24670708e-01 1.00688028e+00 -1.41983223e+00
-1.00057864e+00 -5.88589966e-01 -7.66275600e-02 -9.78409231e-01
-6.55265003e-02 -9.17379856e-01 -1.66441247e-01 -9.48488057e-01
-3.41913141e-02 -5.18141538e-02 -4.28527027e-01 3.68794531e-01
-5.36302710e-03 3.74402136e-01 7.64340818e-01 5.00268638e-01
-2.64087766e-01 2.45849982e-01 1.08807349e+00 -1.60020024e-01
-6.71638012e-01 4.68845457e-01 -7.49621391e-01 9.88569438e-01
6.45522594e-01 -3.29751283e-01 -3.73261929e-01 -7.86514133e-02
4.05708045e-01 3.56004864e-01 1.54010609e-01 -1.19763517e+00
2.25461304e-01 1.68675825e-01 4.92963232e-02 -6.17187560e-01
6.28901124e-01 -5.49824953e-01 3.21377099e-01 2.89160579e-01
-3.20695788e-01 -8.32612365e-02 2.04893842e-01 -1.92158520e-02
-6.03049397e-01 -3.91773582e-01 4.97229517e-01 -7.06141144e-02
-3.80939007e-01 -3.87881726e-01 -4.83231425e-01 -1.64496765e-01
4.29191858e-01 -4.00511682e-01 2.68345535e-01 -4.33054775e-01
-1.06288540e+00 -3.10191453e-01 2.87352681e-01 3.35080445e-01
3.16902101e-01 -1.24475324e+00 -6.39074683e-01 3.82002503e-01
1.71608508e-01 -3.92260164e-01 1.15970187e-01 1.04688799e+00
-1.33583367e-01 2.48453021e-01 -1.61612988e-01 -4.26256984e-01
-1.57266641e+00 3.05679739e-01 1.70733571e-01 -4.10678327e-01
-3.35006624e-01 6.47607803e-01 3.70737985e-02 -2.46445090e-01
3.80741060e-01 -7.97652066e-01 -3.16351354e-01 7.65874267e-01
6.64005697e-01 4.93460953e-01 1.87961161e-01 -6.04737282e-01
-2.05524549e-01 7.33147681e-01 3.44292969e-01 -6.31874561e-01
1.15786374e+00 -5.04571199e-02 -2.36668140e-01 1.18392360e+00
7.23358572e-01 7.26808846e-01 -5.87841213e-01 9.22825634e-02
2.48467371e-01 1.79771089e-03 -1.69421628e-01 -9.45122242e-01
-4.76768196e-01 7.74827778e-01 1.38412207e-01 4.46821153e-01
1.44046414e+00 -3.73394817e-01 6.53365612e-01 3.66903514e-01
2.99137294e-01 -1.08708167e+00 -7.87193701e-02 7.53413677e-01
1.18452537e+00 -6.22672319e-01 -1.23224936e-01 -3.54437411e-01
-5.65283895e-01 1.49304879e+00 -3.71232033e-02 7.86458477e-02
4.12241578e-01 4.60871369e-01 1.33615762e-01 -4.31573652e-02
-5.48118949e-01 -3.71909499e-01 5.62550545e-01 1.53274804e-01
7.67688394e-01 8.47327635e-02 -3.28764707e-01 1.25887918e+00
-7.58874357e-01 1.10985368e-01 3.71836364e-01 5.85397482e-01
-3.74385327e-01 -1.40844345e+00 -8.90827954e-01 -7.23657012e-02
-6.34160280e-01 -1.70296401e-01 -7.13194370e-01 5.25706887e-01
3.78567725e-01 1.02676952e+00 -2.54972607e-01 -4.75354612e-01
4.80045110e-01 4.40859944e-01 6.42958403e-01 -6.14222109e-01
-1.23982334e+00 6.23038828e-01 1.59619182e-01 -4.87482786e-01
-5.28838336e-01 -6.77428484e-01 -1.16451228e+00 1.83773980e-01
-3.93311143e-01 3.98810178e-01 7.13120461e-01 8.91158760e-01
8.44338089e-02 8.44836473e-01 5.38648665e-01 -1.07556808e+00
-6.88156724e-01 -1.16319752e+00 -9.92612839e-01 3.56920302e-01
4.55625564e-01 -2.46902063e-01 -8.87706056e-02 2.78645784e-01] | [15.785526275634766, 5.4031291007995605] |
a43b1ba0-31a8-445c-8800-361987e39b41 | domain-expert-platform-for-goal-oriented | null | null | https://aclanthology.org/2021.eacl-demos.35 | https://aclanthology.org/2021.eacl-demos.35.pdf | Domain Expert Platform for Goal-Oriented Dialog Collection | Today, most dialogue systems are fully or partly built using neural network architectures. A crucial prerequisite for the creation of a goal-oriented neural network dialogue system is a dataset that represents typical dialogue scenarios and includes various semantic annotations, e.g. intents, slots and dialogue actions, that are necessary for training a particular neural network architecture. In this demonstration paper, we present an easy to use interface and its back-end which is oriented to domain experts for the collection of goal-oriented dialogue samples. The platform not only allows to collect or write sample dialogues in a structured way, but also provides a means for simple annotation and interpretation of the dialogues. The platform itself is language-independent; it depends only on the availability of particular language processing components for a specific language. It is currently being used to collect dialogue samples in Latvian (a highly inflected language) which represent typical communication between students and the student service. | ['Gunta Ne{\\v{s}}pore-B{\\=e}rzkalne', 'Normunds Gruzitis', 'Inguna Skadina', 'Arturs Znotins', 'Didzis Go{\\v{s}}ko'] | 2021-04-01 | null | null | null | eacl-2021-2 | ['goal-oriented-dialog'] | ['natural-language-processing'] | [-1.85966194e-01 7.22893238e-01 2.40609720e-01 -7.72027373e-01
-1.57713369e-01 -6.31722808e-01 9.23356056e-01 4.65191245e-01
-5.23377419e-01 9.31535602e-01 2.91942000e-01 -5.57407379e-01
5.51484637e-02 -9.07529235e-01 2.73352992e-02 -1.89778164e-01
1.16926983e-01 1.34547234e+00 2.39766166e-01 -1.00000906e+00
1.42470062e-01 3.71306270e-01 -1.28996861e+00 4.61115479e-01
7.58354485e-01 6.39874995e-01 4.82253462e-01 8.89995635e-01
-7.71303236e-01 9.94156837e-01 -9.45181310e-01 -2.65624702e-01
-1.74564242e-01 -7.42243826e-01 -1.44577610e+00 1.52488083e-01
-1.63176477e-01 -4.31104302e-01 9.58926827e-02 7.56022811e-01
3.34717244e-01 4.29988056e-01 7.11237907e-01 -8.98575246e-01
3.93376090e-02 1.06356823e+00 7.55879402e-01 -1.88712910e-01
5.13559699e-01 1.71935707e-01 8.46987307e-01 -4.49674278e-01
7.06747353e-01 1.28465998e+00 2.92790771e-01 1.02180040e+00
-1.02688742e+00 1.01851691e-02 -2.78811634e-01 -8.39545485e-03
-5.61321557e-01 -6.84948206e-01 8.44799459e-01 -5.20346105e-01
1.04889917e+00 2.57362217e-01 7.05933750e-01 1.11983025e+00
-4.02408749e-01 8.40724766e-01 9.13554966e-01 -1.01120615e+00
4.90512848e-01 7.66364872e-01 6.27202928e-01 5.74960530e-01
-3.41999799e-01 -4.79799300e-01 -2.72673994e-01 -3.88331078e-02
7.09132016e-01 -7.59048402e-01 -2.78344780e-01 -2.45627701e-01
-8.16127837e-01 1.14440954e+00 -2.55432539e-02 8.88029635e-01
-4.36645478e-01 -6.34193480e-01 1.04544818e+00 6.12119615e-01
3.51150095e-01 7.15530396e-01 -6.92512035e-01 -5.99177003e-01
-5.06080806e-01 2.61143714e-01 1.83379090e+00 7.43551672e-01
5.16574860e-01 1.54393703e-01 -1.27626076e-01 1.19450390e+00
3.28768700e-01 -1.91700503e-01 7.65977025e-01 -8.61664772e-01
3.41107547e-01 9.59921300e-01 2.94159532e-01 -4.84088987e-01
-7.62610734e-01 1.67620718e-01 -6.49787068e-01 9.74097475e-02
1.07873011e+00 -6.36543870e-01 -4.20912474e-01 1.55612803e+00
4.08051997e-01 -6.68751717e-01 5.97841322e-01 7.07984924e-01
1.43593335e+00 5.23601055e-01 3.99952456e-02 -2.19541430e-01
1.48731768e+00 -8.50008130e-01 -9.03596580e-01 -9.84472185e-02
9.79351997e-01 -6.08267665e-01 1.39726317e+00 4.14516598e-01
-1.22552323e+00 -4.03977364e-01 -7.12828219e-01 -1.87544271e-01
-8.04573238e-01 8.69177431e-02 4.84728634e-01 8.32780242e-01
-1.12664044e+00 4.75058973e-01 -2.63641000e-01 -6.63474858e-01
-4.74195182e-01 3.35749686e-01 -4.26015884e-01 4.74133193e-01
-1.57576573e+00 1.24220264e+00 8.87002826e-01 8.17120075e-04
-5.10046482e-01 -1.05697162e-01 -1.12062228e+00 1.35461822e-01
1.80543914e-01 -2.87100434e-01 1.66873431e+00 -1.05998850e+00
-2.28035259e+00 1.00838232e+00 2.58367747e-01 -6.49518311e-01
6.42519116e-01 1.44909516e-01 -1.47164956e-01 6.76706210e-02
-2.07317114e-01 5.62737525e-01 2.52406478e-01 -1.02629101e+00
-5.72213054e-01 -2.61020064e-01 4.91926938e-01 4.45674032e-01
-4.45527017e-01 7.92286918e-02 -3.78741622e-01 -1.27518654e-01
-3.21731836e-01 -5.36951005e-01 -1.44162193e-01 -6.46826744e-01
-4.85659331e-01 -6.16591573e-01 4.89453375e-01 -8.62274826e-01
9.75209653e-01 -1.65417445e+00 1.40062869e-01 6.90174624e-02
-9.57677811e-02 6.04682446e-01 8.06788206e-02 8.02481413e-01
1.20468698e-01 -2.98441857e-01 -9.92063135e-02 -2.70373136e-01
3.76029283e-01 4.78412837e-01 6.41276985e-02 -1.43417433e-01
1.66865867e-02 4.70479816e-01 -7.45276570e-01 -3.97233278e-01
7.19536483e-01 1.77963212e-01 -2.37576380e-01 6.77109063e-01
-8.31571996e-01 6.34276628e-01 -4.01001722e-01 -3.55447084e-02
1.41835511e-01 1.42452985e-01 5.01954794e-01 1.17744975e-01
-1.78632364e-01 8.58815312e-01 -9.31687772e-01 1.59028232e+00
-9.28060174e-01 6.38060331e-01 3.36245745e-01 -1.15063417e+00
1.36769915e+00 7.15855718e-01 2.67383130e-03 -4.42466706e-01
5.20230651e-01 2.62091488e-01 1.70544297e-01 -6.63519204e-01
8.84735346e-01 5.39665259e-02 -1.62537053e-01 8.13264906e-01
5.17956197e-01 -2.34214127e-01 8.11158180e-01 -9.04017687e-02
6.03976369e-01 4.47729044e-02 5.08665144e-01 -3.17713261e-01
1.18236518e+00 2.74486572e-01 9.66332629e-02 4.96797770e-01
3.79889421e-02 8.68513659e-02 8.02103758e-01 -6.30527794e-01
-1.11529171e+00 -2.83928871e-01 -2.45818049e-02 1.34772623e+00
-6.07657909e-01 -2.58640498e-01 -1.17322540e+00 -7.54925489e-01
-6.67469203e-01 1.13344276e+00 -1.84669897e-01 2.00877130e-01
-6.10818505e-01 -2.18525499e-01 6.06025279e-01 -1.16510252e-02
6.34143770e-01 -1.59056854e+00 -5.84594548e-01 5.80280781e-01
-2.39537001e-01 -9.63907361e-01 3.06452602e-01 3.51900548e-01
-6.81070149e-01 -9.35618460e-01 -5.96053958e-01 -7.86958814e-01
4.28687036e-01 -5.02443433e-01 1.40051985e+00 8.13784823e-02
2.35624313e-01 2.79329211e-01 -3.76710445e-01 -6.20214403e-01
-1.32546723e+00 5.06797969e-01 -3.14010918e-01 -3.18050802e-01
4.97416794e-01 -2.98793256e-01 4.37440351e-02 1.24385998e-01
-7.16497600e-01 2.06467077e-01 8.19740072e-02 9.31043327e-01
-3.67061973e-01 -1.01246975e-01 6.29083395e-01 -1.11492193e+00
1.45541191e+00 -1.72431454e-01 -5.28385997e-01 3.42760295e-01
-1.66152582e-01 5.74239269e-02 9.99843299e-01 -1.89272799e-02
-1.26790404e+00 -7.84691125e-02 -8.20937634e-01 4.23829138e-01
-1.07190299e+00 5.39001048e-01 -4.45759058e-01 2.00433791e-01
7.30493605e-01 1.99621528e-01 3.48724782e-01 -6.76869094e-01
3.88652056e-01 1.12143719e+00 2.19188288e-01 -8.45117211e-01
1.13085210e-01 -3.57008994e-01 -3.88937533e-01 -1.35189700e+00
-2.66607016e-01 -4.01872396e-01 -9.62446153e-01 -5.12975276e-01
7.10431874e-01 -4.81079966e-01 -1.05283666e+00 3.89943093e-01
-1.33095515e+00 -8.73665750e-01 -3.34326178e-01 2.28492081e-01
-7.14262843e-01 1.31614566e-01 -7.66517699e-01 -1.01453090e+00
-6.19069695e-01 -1.00534463e+00 4.46570605e-01 4.09395188e-01
-6.63829088e-01 -1.57280028e+00 6.39561145e-03 4.36828941e-01
3.70766014e-01 -2.02252090e-01 1.15085518e+00 -1.58224499e+00
1.89750999e-01 -2.36120731e-01 8.58273432e-02 5.97720265e-01
-3.71921882e-02 -5.76451905e-02 -9.94706690e-01 2.38479286e-01
1.43011272e-01 -8.24211299e-01 1.01643868e-01 2.97711957e-02
4.90859002e-01 -4.82774049e-01 2.80064493e-01 -3.09328645e-01
8.85950387e-01 4.69652772e-01 3.38751435e-01 4.43655580e-01
2.69569814e-01 1.35806608e+00 5.89588821e-01 2.85802245e-01
6.10098541e-01 7.66244352e-01 1.17211905e-03 -4.21723016e-02
2.49902278e-01 1.74814150e-01 2.63877273e-01 8.56845260e-01
1.17433697e-01 -1.04948752e-01 -1.16005361e+00 5.60471296e-01
-1.91058254e+00 -6.62216485e-01 -1.79842278e-01 1.96271515e+00
1.19039869e+00 1.91649899e-01 5.79260528e-01 1.61654800e-01
5.71574867e-01 9.16701853e-02 -5.78382164e-02 -1.11974478e+00
4.27621067e-01 1.64572090e-01 -2.51666099e-01 9.89083529e-01
-9.41012204e-01 1.12413061e+00 5.93546438e+00 6.15351081e-01
-9.52030897e-01 -1.30181015e-01 4.65116709e-01 5.11246860e-01
-5.04585728e-02 -3.26632977e-01 -7.65198588e-01 1.64149299e-01
1.44060528e+00 -1.47428304e-01 2.58806825e-01 1.02736199e+00
3.85931551e-01 -2.49530107e-01 -1.12007713e+00 2.70784467e-01
-1.23395212e-01 -1.32342839e+00 -2.43953243e-02 -2.42029876e-01
-1.03580803e-01 -2.27275729e-01 -6.81468964e-01 5.34371316e-01
5.93091905e-01 -8.71778905e-01 3.12743485e-01 2.28616029e-01
2.59837121e-01 -7.33713984e-01 1.03964567e+00 8.25107813e-01
-3.90347332e-01 2.12250561e-01 -2.49329850e-01 -2.98120648e-01
1.00152887e-01 2.16399416e-01 -1.51093352e+00 4.33029592e-01
1.76338568e-01 1.15439378e-01 -2.74845988e-01 6.03424191e-01
-3.25841546e-01 5.55614829e-01 -4.25484478e-01 -5.78639507e-01
4.55352217e-01 -5.16416609e-01 4.91478056e-01 1.46732831e+00
-1.60430223e-01 -1.64684802e-01 3.24715853e-01 5.21106482e-01
1.82890788e-01 6.85392320e-01 -7.10670471e-01 -8.97449851e-02
2.81906813e-01 1.39996147e+00 -7.44098186e-01 -3.44966203e-01
-2.29472905e-01 6.60835445e-01 3.16215605e-01 7.71091655e-02
-1.70247510e-01 -5.92112303e-01 4.28304136e-01 -1.69183403e-01
-1.89089596e-01 -1.57818496e-01 -1.74099237e-01 -9.19921935e-01
-2.42156774e-01 -1.17197824e+00 2.23202437e-01 -5.69329262e-01
-8.86501551e-01 9.01166379e-01 -1.52814481e-02 -5.88867307e-01
-1.16012311e+00 -8.85334253e-01 -8.35750759e-01 1.19515932e+00
-1.00782466e+00 -1.08695483e+00 -1.93502277e-01 6.09883964e-01
9.11579549e-01 -6.56651616e-01 1.59092677e+00 -2.40952112e-02
-6.67844176e-01 3.90659533e-02 -2.29303285e-01 5.89142680e-01
4.72557664e-01 -1.53960693e+00 1.93625063e-01 3.22019249e-01
9.63446870e-02 4.89934057e-01 9.82939363e-01 -2.59776592e-01
-8.43913794e-01 -4.98001218e-01 1.27050412e+00 -6.60951659e-02
7.17184603e-01 -3.55897903e-01 -1.00397635e+00 5.05067945e-01
7.12634444e-01 -8.42044234e-01 9.57760453e-01 3.16205740e-01
2.66624749e-01 1.78894773e-01 -1.18951833e+00 6.35900974e-01
1.97383478e-01 -5.23701191e-01 -1.12396955e+00 5.56896687e-01
3.06005061e-01 -5.91097772e-01 -9.17498946e-01 -7.27513731e-02
1.25549972e-01 -1.24621439e+00 6.05731726e-01 -7.94562817e-01
3.82575333e-01 1.49105579e-01 3.23005378e-01 -1.37851596e+00
4.39325362e-01 -7.07668066e-01 1.56230479e-01 1.42762852e+00
5.68004608e-01 -5.01559496e-01 7.71289349e-01 9.99779642e-01
-2.28947904e-02 -2.32188329e-01 -7.08751678e-01 -3.04114044e-01
2.00336128e-01 -4.74094510e-01 4.40354288e-01 9.82688665e-01
5.07770896e-01 1.01026928e+00 -1.43615723e-01 -3.44311893e-01
8.77541155e-02 -1.23054862e-01 1.07192469e+00 -1.35454440e+00
-1.56844139e-01 -5.70322990e-01 -7.89204165e-02 -9.14772511e-01
3.92708153e-01 -5.54300249e-01 5.53133860e-02 -1.63196146e+00
-7.81145453e-01 -2.79024214e-01 4.98852670e-01 4.98604238e-01
2.74758071e-01 -2.70824105e-01 -4.74181138e-02 2.33648587e-02
-2.53104061e-01 4.52114135e-01 8.19464445e-01 1.55995548e-01
-6.52843058e-01 3.59602839e-01 -2.39483714e-01 1.03717744e+00
1.08650708e+00 3.86532322e-02 -5.71842194e-01 -7.26118730e-03
-1.10673703e-01 3.00314844e-01 -5.49532892e-03 -8.66005123e-01
2.26042449e-01 2.89397165e-02 6.17891178e-02 -4.12916392e-01
3.70450169e-01 -8.17371368e-01 -3.39044929e-01 3.68268281e-01
-7.75165379e-01 -2.93954253e-01 1.67711139e-01 -2.24328786e-01
-5.40290773e-01 -1.03572917e+00 7.39020586e-01 -6.41104579e-01
-7.85755575e-01 -3.68969440e-01 -7.88164079e-01 2.40260107e-03
9.45505559e-01 -7.58824572e-02 -6.63787350e-02 -7.38477826e-01
-1.01630819e+00 3.23790401e-01 3.14206094e-01 2.30382189e-01
1.97313070e-01 -7.29789257e-01 -5.38761973e-01 1.91857740e-01
-5.27046137e-02 5.96957728e-02 1.00640781e-01 2.73895115e-01
-9.57890153e-01 7.10968852e-01 -5.73761940e-01 -2.94052869e-01
-1.52119505e+00 -1.22269671e-02 5.19156814e-01 -5.28862417e-01
-6.22981668e-01 5.21356165e-01 -3.10340345e-01 -9.72409666e-01
3.97047192e-01 -6.66388795e-02 -1.04631007e+00 4.68136042e-01
6.55878186e-01 4.69196402e-02 1.43662378e-01 -6.31080925e-01
1.05874732e-01 -2.62841314e-01 -1.01504996e-01 -5.87021351e-01
1.31941867e+00 -7.19865710e-02 -2.06658915e-01 8.06966841e-01
6.66810930e-01 -3.43201816e-01 -7.80832827e-01 -1.71099663e-01
3.75598252e-01 1.20862231e-01 -1.77560017e-01 -9.14161503e-01
-5.47187507e-01 1.02529991e+00 6.35341853e-02 1.06867635e+00
7.20339417e-01 -2.31681541e-01 4.05206025e-01 9.23021197e-01
3.37929815e-01 -1.37413871e+00 -3.16689074e-01 1.18448532e+00
1.03977597e+00 -1.16036284e+00 -4.01339442e-01 -2.16906935e-01
-8.41023684e-01 1.61761105e+00 7.12631524e-01 2.76734501e-01
3.88988942e-01 1.36665314e-01 5.04831433e-01 -2.57406890e-01
-8.49417686e-01 -3.77755761e-01 5.70473559e-02 7.98912346e-01
9.15455937e-01 3.55345123e-02 -7.16028273e-01 4.85811532e-01
-6.19852602e-01 -4.62329537e-02 7.31699109e-01 7.35732973e-01
-5.65020502e-01 -1.60440421e+00 -2.74909824e-01 1.06047019e-01
-2.80706048e-01 5.03632193e-03 -9.18285072e-01 1.04341900e+00
-2.71360278e-01 1.06010771e+00 1.02303706e-01 -1.33285761e-01
5.37562430e-01 7.41953731e-01 1.80225983e-01 -9.09465194e-01
-1.24402642e+00 -2.07578897e-01 1.12973213e+00 -1.40121490e-01
-3.32771778e-01 -3.11985016e-01 -1.20147002e+00 -2.74633169e-01
-8.86651725e-02 7.49925911e-01 9.62756395e-01 1.08469355e+00
-1.50052011e-01 3.25187415e-01 3.13318163e-01 -7.62436092e-01
-4.48059201e-01 -1.62524736e+00 -4.73485470e-01 3.25711280e-01
-2.41657849e-02 -1.83589280e-01 4.27091643e-02 8.39131325e-03] | [12.906583786010742, 7.9586005210876465] |
972d8edc-1d23-4e1f-8386-962a9b964ce2 | persistent-homology-in-sparse-regression-and | 1409.0177 | null | http://arxiv.org/abs/1409.0177v2 | http://arxiv.org/pdf/1409.0177v2.pdf | Persistent Homology in Sparse Regression and Its Application to Brain Morphometry | Sparse systems are usually parameterized by a tuning parameter that
determines the sparsity of the system. How to choose the right tuning parameter
is a fundamental and difficult problem in learning the sparse system. In this
paper, by treating the the tuning parameter as an additional dimension,
persistent homological structures over the parameter space is introduced and
explored. The structures are then further exploited in speeding up the
computation using the proposed soft-thresholding technique. The topological
structures are further used as multivariate features in the tensor-based
morphometry (TBM) in characterizing white matter alterations in children who
have experienced severe early life stress and maltreatment. These analyses
reveal that stress-exposed children exhibit more diffuse anatomical
organization across the whole white matter region. | ['Moo. K. Chung', 'Jamie L. Hanson', 'Richard J. Davidson', 'Seth D. Pollak', 'Jieping Ye'] | 2014-08-31 | null | null | null | null | ['brain-morphometry'] | ['medical'] | [-5.62492125e-02 -2.51460969e-02 -3.04707140e-01 -4.08806980e-01
-5.81624135e-02 -3.54572117e-01 1.87270671e-01 4.04794455e-01
-1.64150581e-01 3.83225352e-01 3.33345324e-01 2.41945803e-01
-5.76027155e-01 -5.90828180e-01 -3.77155215e-01 -8.63826036e-01
-7.03106165e-01 3.67121339e-01 1.01995744e-01 -1.79006219e-01
5.25982559e-01 5.27541518e-01 -1.23066425e+00 6.10561445e-02
1.08151853e+00 6.65599287e-01 4.43070233e-02 2.24172056e-01
2.70792358e-02 6.30813763e-02 -1.32817894e-01 -3.27980071e-01
2.09734470e-01 -2.60488361e-01 -7.39898443e-01 1.65137187e-01
7.12550342e-01 1.76225323e-02 -3.23817059e-02 1.41546226e+00
4.09127980e-01 1.98361501e-01 8.40818048e-01 -7.91163683e-01
-3.66992682e-01 9.12279963e-01 -6.41131401e-01 7.06414223e-01
2.89015949e-01 -3.03334296e-01 7.87745953e-01 -1.04532349e+00
7.20502138e-01 1.21095645e+00 5.46066165e-01 3.21640000e-02
-1.53590572e+00 -4.99212176e-01 -1.22387586e-02 5.04446566e-01
-1.47826397e+00 -4.11477268e-01 1.20097089e+00 -1.20362508e+00
7.40976453e-01 3.58890146e-02 1.06165922e+00 9.06780481e-01
3.35363418e-01 -9.71995220e-02 1.25647986e+00 -2.53268331e-01
2.97906876e-01 -3.51931185e-01 2.25201890e-01 1.03497374e+00
7.17811525e-01 -2.34543830e-02 -5.77568650e-01 -2.33258352e-01
7.40141273e-01 -7.47592896e-02 -8.56063291e-02 -5.18650234e-01
-1.11139333e+00 9.19575751e-01 3.15029383e-01 6.06447399e-01
-3.78125906e-01 3.75750698e-02 4.03296262e-01 2.88568467e-01
4.77003783e-01 5.88073194e-01 -1.65392771e-01 -2.52118446e-02
-1.22676814e+00 6.22074977e-02 4.11349028e-01 4.07584012e-01
7.61579812e-01 2.67993063e-01 2.57339925e-01 9.61199760e-01
2.94366837e-01 2.18110919e-01 5.45961797e-01 -9.64931428e-01
5.33299387e-01 8.07816684e-01 -6.79868698e-01 -1.54457164e+00
-8.29478800e-01 -4.92987573e-01 -1.18913567e+00 1.15698636e-01
2.15154111e-01 -1.50877029e-01 -5.19977927e-01 1.78867733e+00
8.05510402e-01 1.65380668e-02 -5.25773764e-01 9.49635684e-01
4.79633838e-01 1.58164188e-01 6.54507279e-02 -1.72831923e-01
1.42389572e+00 -4.13924962e-01 -4.09522265e-01 5.79363517e-02
5.64918280e-01 -6.23942971e-01 5.31571567e-01 4.28696692e-01
-1.15786242e+00 6.37020394e-02 -8.61580372e-01 2.17571989e-01
-2.60289311e-01 1.81359708e-01 9.08947945e-01 4.59658027e-01
-9.57433641e-01 8.70476782e-01 -9.57544982e-01 -4.50664639e-01
4.58166122e-01 5.24459720e-01 -7.24420071e-01 2.44184107e-01
-1.01225770e+00 8.87814879e-01 3.66291136e-01 1.38791919e-01
-5.92867851e-01 -8.17299426e-01 -8.13569665e-01 -7.86979944e-02
-7.07501546e-02 -5.61723709e-01 3.96397650e-01 -4.48138863e-01
-1.32847619e+00 9.41841185e-01 2.53324956e-01 -8.38729367e-02
7.20944908e-03 1.37206271e-01 -2.16210589e-01 7.74382293e-01
2.10296720e-01 2.13575751e-01 1.31718981e+00 -6.61507964e-01
3.27834874e-01 -6.52381659e-01 -1.31958321e-01 4.30190824e-02
-2.48464748e-01 1.98625073e-01 2.78255224e-01 -9.62294817e-01
9.19261038e-01 -8.08849275e-01 -4.22839999e-01 -1.18495837e-01
-5.72842062e-01 7.62346834e-02 3.64780754e-01 -7.82143772e-01
1.17192256e+00 -1.94769621e+00 6.64649725e-01 6.88818514e-01
4.56820101e-01 -1.98945940e-01 -1.34272918e-01 5.18343270e-01
-5.16387641e-01 -6.27135783e-02 -3.16099882e-01 1.64228603e-01
-2.95167863e-01 -1.83397800e-01 9.01317820e-02 1.00381970e+00
7.60902017e-02 5.07758617e-01 -6.84726954e-01 -7.29407907e-01
-9.99134704e-02 3.44251215e-01 -7.79711306e-01 2.11250689e-03
3.74841899e-01 6.94949865e-01 -6.04727089e-01 6.60920441e-01
6.44361854e-01 -5.92134660e-03 1.61265776e-01 -4.09216762e-01
-2.35282272e-01 2.46022120e-02 -1.30541945e+00 1.50246143e+00
5.02343252e-02 1.90627396e-01 6.09892607e-01 -1.47580612e+00
8.35133195e-01 2.90560901e-01 5.83191991e-01 -2.60220259e-01
3.39307070e-01 1.51857391e-01 5.20025849e-01 -7.76541710e-01
-2.03180611e-02 -1.55715123e-01 2.77147532e-01 4.17640179e-01
2.36771807e-01 -1.07523330e-01 4.39365715e-01 8.52964148e-02
9.53423202e-01 -3.63029599e-01 3.03848475e-01 -9.39468682e-01
6.26477301e-01 -1.81053132e-01 6.89736545e-01 -7.36675272e-03
-2.65456866e-02 2.54218698e-01 6.96003377e-01 -3.79093677e-01
-1.18761146e+00 -8.05727899e-01 -6.12004161e-01 1.00257230e+00
-4.18243349e-01 -1.13224633e-01 -9.11866069e-01 -1.76571205e-01
6.59207702e-02 -6.81682527e-02 -7.90241241e-01 -1.58237219e-01
-8.27694595e-01 -9.96122360e-01 3.96600574e-01 2.10286364e-01
2.22509444e-01 -7.34581947e-01 -5.11541307e-01 -2.28120852e-02
2.31321827e-01 -1.03741074e+00 -6.07074618e-01 4.75567626e-03
-1.34629118e+00 -8.95245135e-01 -6.64609492e-01 -5.72414398e-01
1.21414268e+00 -2.03478932e-01 4.87074763e-01 2.21136108e-01
-5.22035658e-01 1.74381986e-01 -2.30014637e-01 3.93570542e-01
-9.08251628e-02 1.57217607e-01 4.63408321e-01 7.98397884e-02
-4.69058394e-01 -1.19843113e+00 -6.78138435e-01 -3.11206449e-02
-8.86162698e-01 7.10260347e-02 4.50177163e-01 6.73937440e-01
4.09839302e-01 -9.27846180e-04 4.40367639e-01 -6.06410861e-01
5.81335664e-01 -7.14020312e-01 -6.02212191e-01 -3.21711898e-02
-4.46746290e-01 2.11722746e-01 5.30041635e-01 -7.50563085e-01
-6.85038626e-01 -6.42083678e-03 8.26996416e-02 -2.03879148e-01
-2.65008751e-02 5.90183854e-01 -2.68344972e-02 -6.28307819e-01
3.55719835e-01 -6.09982852e-03 -1.71837747e-01 -8.77441704e-01
2.37652197e-01 -2.11981684e-02 2.54155874e-01 -8.24963808e-01
7.56792307e-01 3.46121997e-01 7.01044083e-01 -1.07139647e+00
-6.27941072e-01 -1.70969307e-01 -9.87805188e-01 2.95478757e-02
8.89944851e-01 -4.13960010e-01 -5.28508902e-01 3.48141372e-01
-8.93886924e-01 -1.19504534e-01 6.95700943e-02 7.10046828e-01
-3.73861074e-01 4.99140769e-01 -6.38436019e-01 -3.22481364e-01
-5.64575195e-01 -1.03983700e+00 6.43684924e-01 -1.47355333e-01
-2.35639632e-01 -1.21018815e+00 5.69964588e-01 4.66591209e-01
4.71756727e-01 5.59159100e-01 1.55223095e+00 -5.62114835e-01
-2.39378333e-01 -8.99368152e-02 5.56877255e-02 9.09571648e-02
-2.56688863e-01 1.69956028e-01 -2.77515531e-01 -3.62657249e-01
1.80614263e-01 1.33660391e-01 4.89238352e-01 2.49829605e-01
1.00178134e+00 -5.66240668e-01 3.07742152e-02 1.13244998e+00
1.38680923e+00 -1.18833356e-01 -5.92180640e-02 7.35413358e-02
9.03283477e-01 1.03917742e+00 7.68265650e-02 4.24730211e-01
2.94597417e-01 5.73325992e-01 3.24787259e-01 3.58175337e-01
-8.19788203e-02 -1.47639010e-02 1.69759825e-01 1.47019446e+00
-2.96446025e-01 7.54515409e-01 -1.08396399e+00 6.32870495e-01
-1.39655757e+00 -9.19017613e-01 -4.24172580e-02 1.91874492e+00
1.08076000e+00 -3.59970838e-01 1.55313626e-01 9.97444093e-02
8.48973989e-01 1.02714077e-01 -4.62277949e-01 -5.00780284e-01
-1.28414765e-01 4.99905616e-01 2.36279711e-01 4.87376988e-01
-7.89667606e-01 8.50236118e-01 7.01778793e+00 6.58471704e-01
-1.30593002e+00 1.57932147e-01 2.35531017e-01 -1.01086840e-01
-2.16057479e-01 7.13207871e-02 -3.99824023e-01 5.52812338e-01
7.58985460e-01 6.60179108e-02 6.81213915e-01 6.26507282e-01
2.73287714e-01 -2.88729489e-01 -7.84581602e-01 8.07750702e-01
-1.74280107e-01 -1.09192050e+00 -1.52663365e-01 9.00427178e-02
8.62958789e-01 -1.25375807e-01 1.39887020e-01 -4.04215962e-01
-2.84134865e-01 -1.07111013e+00 6.32244349e-01 7.01180637e-01
7.30862617e-01 -5.27724922e-01 2.19539344e-01 1.58132792e-01
-1.04735518e+00 -1.26659334e-01 -3.61487716e-01 -3.57899144e-02
1.10749509e-02 1.08100331e+00 -6.39198184e-01 -8.72801766e-02
6.74601376e-01 4.49597418e-01 -6.64755940e-01 1.05655956e+00
-1.67715296e-01 7.70591676e-01 -3.63616526e-01 3.38170946e-01
-1.10251285e-01 -7.21460462e-01 9.88955379e-01 1.18639088e+00
3.88624996e-01 2.57238150e-01 -1.71466455e-01 1.09133863e+00
1.88514441e-01 4.45125908e-01 -5.79069793e-01 -2.78159082e-01
3.03996950e-01 1.64611113e+00 -9.13031340e-01 -2.42840461e-02
1.73712410e-02 4.09433246e-01 5.43935120e-01 3.28566492e-01
2.40672976e-02 -1.16345398e-01 6.89280689e-01 3.12935889e-01
3.68541867e-01 -5.97218931e-01 -4.81750548e-01 -1.39883375e+00
-1.67818621e-01 -7.14959323e-01 9.70450714e-02 -2.71240801e-01
-1.21437120e+00 3.82979423e-01 3.25697869e-01 -7.29894698e-01
-2.43676916e-01 -3.01732510e-01 -9.71945524e-01 5.60318708e-01
-8.02319348e-01 -7.18093097e-01 2.97156535e-03 6.65727556e-01
-2.14094281e-01 -2.28841394e-01 6.64273381e-01 2.73044229e-01
-1.10620451e+00 3.06531787e-01 -1.93625852e-01 -5.57461716e-02
1.40130952e-01 -1.20917904e+00 -3.16332549e-01 9.01358128e-01
-2.11743429e-01 1.10309768e+00 8.35854053e-01 -8.32727551e-01
-1.42762959e+00 -7.92367995e-01 6.77497804e-01 1.72078624e-01
1.26140404e+00 -5.89022577e-01 -1.08705461e+00 3.32848310e-01
-5.37946485e-02 1.64136410e-01 8.13628495e-01 3.09508257e-02
-4.40130293e-01 -2.88573988e-02 -1.46142256e+00 3.55632007e-01
9.78695393e-01 -5.10285974e-01 -6.56424046e-01 3.87397528e-01
4.98471856e-01 -5.11532426e-02 -1.53463066e+00 3.29675585e-01
3.13158542e-01 -8.30615342e-01 9.72198248e-01 -6.56755090e-01
2.43005261e-01 -5.86434938e-02 -1.10803895e-01 -1.41798830e+00
-4.79098290e-01 -7.83344150e-01 -3.22963387e-01 1.15608859e+00
-1.05769083e-01 -6.21493638e-01 6.29644036e-01 3.19746137e-01
7.56595284e-02 -1.19416523e+00 -1.11890590e+00 -4.85840529e-01
3.93357933e-01 -2.72086859e-02 6.01466954e-01 1.24548566e+00
2.04296142e-01 4.07145321e-02 1.35206074e-01 1.73858464e-01
9.29763258e-01 1.06902719e-01 -8.55715126e-02 -1.28508544e+00
-6.11813515e-02 -5.46851873e-01 -8.43298137e-01 -5.13537638e-02
1.26127452e-01 -1.14578784e+00 -5.45737982e-01 -8.57143939e-01
3.01246852e-01 -3.75400573e-01 -1.56701863e-01 3.53492528e-01
2.74732918e-01 1.35340512e-01 -2.51704119e-02 1.32933900e-01
-2.74607033e-01 4.63451415e-01 1.34581566e+00 2.74865150e-01
-9.34064463e-02 -2.10809544e-01 -5.50915956e-01 1.13854432e+00
8.51128459e-01 -5.74235439e-01 -1.35600910e-01 -2.23265976e-01
5.94940066e-01 2.71106869e-01 3.11194867e-01 -8.92022312e-01
7.40029663e-02 -4.22478288e-01 3.34451437e-01 -3.13151568e-01
3.21786225e-01 -7.48644054e-01 1.76563952e-02 4.26559746e-01
-2.60891676e-01 3.54979575e-01 -2.57779926e-01 -1.42450503e-03
-6.86996952e-02 -4.42656428e-01 1.15044129e+00 2.65158981e-01
-9.82983261e-02 5.63067496e-01 -3.78763050e-01 1.55990019e-01
8.62550855e-01 -1.26961945e-02 -7.26963282e-02 1.71597928e-01
-1.10250747e+00 -1.79637432e-01 2.90515482e-01 -4.35387418e-02
6.87601864e-01 -1.33570302e+00 -5.79032063e-01 3.61633718e-01
-3.10359240e-01 -4.22675461e-01 4.16187882e-01 1.54504180e+00
-5.48092365e-01 1.60526812e-01 -6.36487603e-01 -4.75383461e-01
-1.22576547e+00 3.50423992e-01 1.80300295e-01 -1.16256349e-01
-6.91058457e-01 7.85998523e-01 -1.85526349e-02 -1.44075677e-01
3.07890829e-02 -3.58342886e-01 -5.39834619e-01 4.35644031e-01
1.40648261e-01 8.82646143e-01 -6.17108382e-02 -1.12768781e+00
-3.34544212e-01 1.12468541e+00 2.51964062e-01 1.73756946e-02
1.76170874e+00 -2.36802012e-01 -1.03296888e+00 2.96127617e-01
1.33983362e+00 4.62071076e-02 -7.28914857e-01 -1.94789156e-01
1.42895073e-01 -2.43689284e-01 8.83129090e-02 -2.17215136e-01
-1.49343431e+00 8.63945484e-01 6.67282701e-01 2.03235582e-01
9.78269398e-01 1.60206139e-01 5.93597591e-01 2.97303498e-01
3.39187950e-01 -9.79060709e-01 -7.48951361e-02 5.27057469e-01
1.18685389e+00 -6.90831304e-01 3.73441279e-01 -5.24744153e-01
-2.38625199e-01 1.16299343e+00 3.54811460e-01 -6.70257330e-01
8.86131346e-01 2.62355000e-01 -4.59485918e-01 -6.53249323e-01
-5.56382775e-01 1.16034567e-01 8.52016926e-01 4.33851838e-01
3.54727894e-01 2.10386574e-01 -6.92349255e-01 5.89325607e-01
-6.68317497e-01 -5.77915013e-01 3.87046963e-01 5.01419365e-01
-5.93826056e-01 -1.04694271e+00 -6.28915191e-01 5.78789234e-01
-7.29009092e-01 -1.69280395e-01 -2.24675626e-01 3.10265064e-01
4.38132972e-01 4.36015964e-01 -3.09197575e-01 -1.91752791e-01
-5.01784571e-02 1.43022928e-02 7.88513899e-01 -7.59579003e-01
-6.70825899e-01 -2.03027487e-01 -2.08853707e-01 -7.83238590e-01
-2.45685577e-01 -9.69610989e-01 -1.18516576e+00 -1.68803245e-01
-1.46152610e-02 -2.09854022e-02 8.79071116e-01 8.21909487e-01
2.27919459e-01 1.41834155e-01 5.14010489e-01 -9.00753021e-01
-4.57935840e-01 -9.74604428e-01 -7.44472623e-01 3.11653823e-01
4.77075040e-01 -1.15612733e+00 -4.08301741e-01 -1.79638371e-01] | [12.411349296569824, 3.3731727600097656] |
34441838-1dc7-43f6-89b1-2423a4cdd0e0 | unsupervised-rgb-to-thermal-domain-adaptation | 2210.04367 | null | https://arxiv.org/abs/2210.04367v1 | https://arxiv.org/pdf/2210.04367v1.pdf | Unsupervised RGB-to-Thermal Domain Adaptation via Multi-Domain Attention Network | This work presents a new method for unsupervised thermal image classification and semantic segmentation by transferring knowledge from the RGB domain using a multi-domain attention network. Our method does not require any thermal annotations or co-registered RGB-thermal pairs, enabling robots to perform visual tasks at night and in adverse weather conditions without incurring additional costs of data labeling and registration. Current unsupervised domain adaptation methods look to align global images or features across domains. However, when the domain shift is significantly larger for cross-modal data, not all features can be transferred. We solve this problem by using a shared backbone network that promotes generalization, and domain-specific attention that reduces negative transfer by attending to domain-invariant and easily-transferable features. Our approach outperforms the state-of-the-art RGB-to-thermal adaptation method in classification benchmarks, and is successfully applied to thermal river scene segmentation using only synthetic RGB images. Our code is made publicly available at https://github.com/ganlumomo/thermal-uda-attention. | ['Soon-Jo Chung', 'Connor Lee', 'Lu Gan'] | 2022-10-09 | null | null | null | null | ['scene-segmentation'] | ['computer-vision'] | [ 4.80137974e-01 4.87614125e-02 -3.27338539e-02 -7.14271426e-01
-9.05780971e-01 -8.93348575e-01 3.77568454e-01 -2.98987001e-01
-7.10587919e-01 5.29719710e-01 -2.59229988e-01 -5.05445227e-02
2.18167737e-01 -7.85706818e-01 -7.77857721e-01 -9.82701659e-01
3.13890576e-01 7.13606000e-01 1.26792222e-01 -2.70861030e-01
-9.55401137e-02 4.01868135e-01 -1.44557488e+00 1.32498458e-01
8.55741262e-01 1.11610281e+00 4.15326506e-01 6.24128759e-01
-5.44716269e-02 3.79214376e-01 -2.85590738e-01 1.18462786e-01
7.14666486e-01 -4.66755986e-01 -1.40722299e+00 1.09501503e-01
5.50983906e-01 -3.26858521e-01 -6.28651008e-02 7.87442148e-01
4.67854649e-01 4.74157900e-01 6.35278404e-01 -1.41991305e+00
-7.14845359e-01 9.00770277e-02 -6.62403584e-01 -1.31845370e-01
-5.54597266e-02 2.99716473e-01 8.16220105e-01 -3.87666523e-01
5.49518168e-01 9.40520287e-01 8.17169130e-01 7.50266135e-01
-1.47531986e+00 -6.49166644e-01 1.46831945e-01 1.57590732e-01
-1.19404233e+00 -2.70144135e-01 8.70747626e-01 -2.76896924e-01
1.08992100e+00 9.03226510e-02 7.79980242e-01 1.15353286e+00
-2.65053302e-01 7.30800927e-01 1.47186863e+00 -2.46745989e-01
3.46530646e-01 1.55326370e-02 -1.31917000e-01 4.36742276e-01
-1.60424471e-01 9.88018140e-03 -4.86771584e-01 2.01171100e-01
7.56232977e-01 -9.09189582e-02 -1.12956733e-01 -5.54357290e-01
-1.07685137e+00 7.27615952e-01 9.50026095e-01 1.89264402e-01
-2.53751457e-01 3.28387827e-01 4.29594219e-01 2.64678359e-01
7.21667588e-01 4.71448570e-01 -9.01608884e-01 -7.32827559e-02
-7.02371120e-01 -1.73567399e-01 5.12970984e-01 1.05698299e+00
1.52886438e+00 -1.22119829e-01 4.98506218e-01 8.16136360e-01
2.21133560e-01 9.41909730e-01 4.81394410e-01 -1.33150291e+00
2.79669017e-01 5.39539039e-01 -1.71447024e-01 -4.24012035e-01
-5.79702199e-01 1.18723921e-01 -8.06702554e-01 3.25638503e-01
4.30002987e-01 -1.69021413e-01 -1.52440739e+00 1.63174593e+00
3.15121472e-01 3.47999819e-02 2.72559464e-01 1.16725850e+00
7.93285906e-01 5.68923295e-01 1.39622211e-01 4.70341623e-01
1.15984452e+00 -1.00522316e+00 -2.43538305e-01 -9.56720710e-01
5.73684931e-01 -4.27574873e-01 1.24798489e+00 7.30268098e-03
-4.16662276e-01 -5.53326428e-01 -9.39382613e-01 -4.93861198e-01
-7.78818667e-01 -1.11935899e-01 7.98517585e-01 7.19395220e-01
-1.20208561e+00 5.04435658e-01 -1.15482020e+00 -1.08518565e+00
6.62892997e-01 5.68664610e-01 -5.59449494e-01 -4.85051498e-02
-9.97627199e-01 9.22732413e-01 4.72193211e-01 2.52653867e-01
-9.09689367e-01 -6.71826601e-01 -1.03371811e+00 -5.87264121e-01
2.91304588e-02 -7.14400530e-01 1.28192222e+00 -1.63411486e+00
-1.82021821e+00 1.19765019e+00 1.12679042e-02 -1.45041555e-01
3.46061021e-01 -4.10217285e-01 8.46248120e-02 3.09496552e-01
3.11785072e-01 1.20623577e+00 6.61126137e-01 -1.51537478e+00
-2.95012385e-01 -5.73541701e-01 1.94628146e-02 5.86801529e-01
-3.97262454e-01 -3.44500780e-01 -3.29156667e-01 -2.35180065e-01
1.62198663e-01 -1.15321815e+00 -4.27399367e-01 2.47762963e-01
-1.69500366e-01 1.49285913e-01 9.54020441e-01 -6.27456903e-01
7.10122436e-02 -2.05037761e+00 3.56082350e-01 1.81720406e-01
-2.75701582e-01 3.03723514e-02 -2.50783741e-01 1.83229849e-01
-2.58270139e-03 9.98211354e-02 -9.53867555e-01 -4.70111966e-01
-1.02545887e-01 6.79710507e-01 1.54002542e-02 7.62111723e-01
1.75728276e-01 9.09679532e-01 -9.70681727e-01 -4.87838089e-01
4.84732717e-01 3.82250071e-01 -3.27908337e-01 3.19008857e-01
-2.90543735e-01 1.07173550e+00 -3.74961168e-01 5.39061189e-01
8.88907075e-01 -1.38408124e-01 1.41329363e-01 -2.85436600e-01
-1.01229362e-03 1.92190379e-01 -6.88829422e-01 2.49187231e+00
-7.41818964e-01 8.18868458e-01 1.65133506e-01 -1.27940071e+00
9.60217714e-01 -5.02086133e-02 6.92618728e-01 -1.06102586e+00
7.96259120e-02 1.92249537e-01 -5.71293592e-01 -4.34085935e-01
5.58395386e-01 -2.75746822e-01 -4.39681500e-01 5.90226293e-01
2.58596063e-01 -7.34828949e-01 -2.64947385e-01 -1.15035236e-01
7.22285271e-01 7.69484997e-01 -8.48760679e-02 -1.48470297e-01
7.38392994e-02 4.99213040e-01 5.77452242e-01 4.30349052e-01
-2.36012325e-01 8.73544633e-01 -8.02193955e-03 -4.34764117e-01
-1.13369346e+00 -1.11072612e+00 6.26158863e-02 1.46098220e+00
4.92993563e-01 2.30459273e-01 -6.79717004e-01 -6.55433238e-01
3.86256911e-02 3.86350453e-01 -8.59679520e-01 -1.88764274e-01
-4.31890368e-01 -6.95887268e-01 7.45134413e-01 9.13566113e-01
9.53046083e-01 -1.07443380e+00 -9.48044896e-01 -1.38192222e-01
-4.18169856e-01 -1.11282229e+00 -1.99641943e-01 7.90851176e-01
-1.00587773e+00 -9.33825612e-01 -7.38367379e-01 -8.43136191e-01
7.40602195e-01 1.25297457e-01 1.12754464e+00 -3.91949117e-01
-3.31103325e-01 7.35601366e-01 -4.93211418e-01 -1.79162800e-01
-1.36958376e-01 4.03036624e-01 -2.48595521e-01 -2.16928288e-01
4.49499398e-01 -5.06729126e-01 -6.78352654e-01 3.35758120e-01
-8.74775529e-01 5.05141206e-02 3.75003517e-01 8.12265694e-01
5.85697889e-01 -4.31257993e-01 1.38604745e-01 -6.53667927e-01
-6.51276410e-02 -2.12485254e-01 -5.26362240e-01 1.15805723e-01
-1.74793258e-01 1.37919769e-01 4.61979270e-01 -1.33681491e-01
-1.38811290e+00 7.78807580e-01 1.10359035e-01 -3.16664696e-01
-7.24911809e-01 1.44841120e-01 -1.25333905e-01 -3.27751517e-01
8.52325022e-01 2.52610475e-01 1.05336465e-01 -3.85116816e-01
8.73338759e-01 6.29469097e-01 6.34306073e-01 -7.15629697e-01
8.41257453e-01 8.65832627e-01 -1.25051901e-01 -9.62380648e-01
-5.71282685e-01 -7.58740306e-01 -1.26239133e+00 -2.49528989e-01
1.30043256e+00 -1.05753458e+00 -3.71807367e-01 8.73333156e-01
-8.97871256e-01 -1.11896789e+00 -2.70658016e-01 1.82665363e-01
-8.97204578e-01 3.49757105e-01 -4.35055047e-01 -4.52475250e-01
-2.55715877e-01 -9.26596344e-01 1.29135323e+00 3.21433842e-01
2.12227255e-02 -1.08223021e+00 2.19324633e-01 4.51495051e-01
4.45584238e-01 3.58856827e-01 4.15703118e-01 -2.23864183e-01
-4.99838382e-01 3.38929087e-01 -2.66797781e-01 5.10592043e-01
3.74240726e-01 -3.67599487e-01 -1.36947763e+00 -2.17227921e-01
-4.23197269e-01 -8.93933654e-01 1.10557735e+00 2.82352418e-01
1.08113241e+00 7.75836930e-02 -3.00429046e-01 1.09277999e+00
1.54496574e+00 -5.48713617e-02 5.43826580e-01 7.04506397e-01
9.90036070e-01 7.26899564e-01 6.48135781e-01 1.93941668e-01
7.22500920e-01 5.55035591e-01 6.83356762e-01 -7.68230796e-01
1.21044427e-01 1.99720845e-01 1.85049251e-01 2.60714978e-01
-3.21657598e-01 -2.39160727e-03 -1.30476868e+00 8.63361657e-01
-1.82251525e+00 -4.91678327e-01 8.67735371e-02 2.01448822e+00
8.97532463e-01 -5.34141779e-01 1.35691583e-01 -3.16261530e-01
4.86312717e-01 7.92945921e-02 -8.39682639e-01 -4.13297147e-01
-1.24844305e-01 4.72450197e-01 1.08993304e+00 2.71633357e-01
-1.53313172e+00 1.30150092e+00 5.56293726e+00 3.45232189e-01
-1.19271564e+00 3.25576991e-01 7.06593752e-01 8.46665353e-02
-2.39000451e-02 -5.13824187e-02 -7.54160061e-02 -2.61607096e-02
8.41962039e-01 4.17785078e-01 5.67920148e-01 7.29649544e-01
-1.36856750e-01 -3.25101703e-01 -1.01209784e+00 8.18664968e-01
-6.03620261e-02 -7.11604238e-01 -5.02043962e-01 -6.37407899e-02
8.08142781e-01 5.90815902e-01 1.37244418e-01 -8.03091601e-02
6.00868940e-01 -9.26166773e-01 6.77398980e-01 3.08649361e-01
9.12991405e-01 -5.47868609e-01 6.68456376e-01 -1.85586825e-01
-1.37278557e+00 6.13838769e-02 -3.17825526e-01 1.50877953e-01
-1.40200943e-01 1.08479284e-01 -5.99129617e-01 7.72398949e-01
1.32297230e+00 1.02533031e+00 -4.90222961e-01 6.06720209e-01
-2.85009712e-01 3.31312418e-01 -6.42319620e-01 2.58999258e-01
3.17665666e-01 -3.15078884e-01 4.36482467e-02 1.25468981e+00
2.05262467e-01 4.10627387e-02 1.63677275e-01 8.96038413e-01
4.27566692e-02 -2.00345770e-01 -7.60814369e-01 1.21016964e-01
3.20593804e-01 1.29425406e+00 -1.06475031e+00 -2.93388337e-01
-2.97334999e-01 1.64842963e+00 2.74673164e-01 6.36446357e-01
-8.22148919e-01 -3.02203476e-01 9.12920356e-01 -4.24659669e-01
3.39408189e-01 -4.52053964e-01 -6.58490658e-01 -9.46265876e-01
-1.32246420e-01 -4.39353883e-01 3.38036776e-01 -9.44185495e-01
-1.38255286e+00 5.60266793e-01 4.05670851e-02 -1.36577094e+00
-1.51838688e-02 -7.58769691e-01 -1.42516479e-01 8.94776344e-01
-1.66208947e+00 -1.66399765e+00 -7.27984369e-01 8.00069034e-01
2.73033589e-01 4.13946062e-01 1.03864026e+00 6.94334805e-02
-3.01295251e-01 3.48440528e-01 2.34952018e-01 2.09667429e-01
1.08509851e+00 -1.49778533e+00 4.19742018e-01 6.56329393e-01
-2.93811917e-01 2.22804785e-01 4.25984591e-01 -2.18397200e-01
-1.27033377e+00 -1.35500443e+00 9.70981270e-02 -5.17430425e-01
4.44275767e-01 -5.20904839e-01 -9.02786016e-01 8.76906037e-01
4.75275844e-01 2.27868482e-01 6.61503971e-01 2.81321183e-02
-5.74214935e-01 -4.47611868e-01 -1.24782991e+00 1.24431342e-01
1.04580832e+00 -7.35479355e-01 -3.96492898e-01 3.87500018e-01
5.11429906e-01 -5.74737847e-01 -9.98153925e-01 2.22741336e-01
4.93947119e-01 -8.87234807e-01 9.72079813e-01 -1.26595274e-01
3.99690896e-01 -4.83856648e-01 -1.04620457e-01 -1.57973695e+00
-3.83285396e-02 -1.72345161e-01 7.17963755e-01 1.10907042e+00
5.27302444e-01 -8.09666812e-01 8.35041583e-01 6.68426275e-01
-4.15508687e-01 1.94229409e-01 -1.14481854e+00 -8.51011574e-01
4.07848746e-01 -4.27950501e-01 4.50480014e-01 1.14377403e+00
-1.68539673e-01 3.40205878e-02 -2.25315019e-01 3.55208188e-01
6.69827402e-01 4.35833842e-01 7.85886765e-01 -9.94231105e-01
-8.17440152e-02 -2.80897588e-01 -3.81399482e-01 -9.68481719e-01
2.82904744e-01 -8.75578523e-01 6.71932638e-01 -1.70968640e+00
9.35323834e-02 -5.95671237e-01 -2.04933539e-01 1.20519340e+00
2.06455603e-01 7.76330233e-01 3.94231128e-03 1.44422039e-01
-6.52493298e-01 7.96800494e-01 1.06130147e+00 -3.91570598e-01
-3.16435754e-01 -5.13108253e-01 -4.15800452e-01 4.51199055e-01
1.24262822e+00 -3.91309768e-01 -2.22898766e-01 -7.83904314e-01
-8.95920992e-02 -4.89120960e-01 4.89805222e-01 -1.02536762e+00
1.98684365e-01 -2.35991150e-01 3.93601686e-01 -2.45679811e-01
5.44821560e-01 -1.11174262e+00 9.58637968e-02 2.12121814e-01
-4.03462984e-02 -2.63705462e-01 5.81346214e-01 4.92894679e-01
-2.78271493e-02 1.59454361e-01 9.22758996e-01 -3.68791640e-01
-1.31650436e+00 7.13434443e-02 -4.78027850e-01 -2.65548211e-02
9.80707943e-01 -4.15969312e-01 -4.30260599e-01 -1.81015447e-01
-6.39514744e-01 3.10521066e-01 9.72533405e-01 4.22844321e-01
4.49272394e-01 -1.04054475e+00 -3.30700278e-01 1.72600836e-01
5.10823131e-01 5.38058996e-01 3.51966023e-01 6.72207594e-01
-8.04321706e-01 1.09060347e-01 -6.43984437e-01 -1.11015701e+00
-1.11108851e+00 8.78745988e-02 7.12354064e-01 1.91588372e-01
-4.74387139e-01 1.09460008e+00 2.25021526e-01 -1.11416554e+00
-7.77020901e-02 -5.07996261e-01 2.36000791e-01 2.42043175e-02
-1.58770233e-01 1.28385082e-01 7.51588270e-02 -8.99728775e-01
-6.82477593e-01 1.09813774e+00 3.64824384e-01 -1.59479037e-01
1.51911700e+00 -3.73322338e-01 -3.43468279e-01 4.45390105e-01
1.29128408e+00 -8.86663795e-01 -1.69231188e+00 -2.87503541e-01
-2.84263611e-01 -2.40911946e-01 1.02013215e-01 -9.90831852e-01
-1.47263443e+00 1.02842307e+00 1.06234932e+00 -1.11011848e-01
1.63520515e+00 2.48053029e-01 6.09146833e-01 5.10768771e-01
4.39796567e-01 -1.30902982e+00 2.28531152e-01 6.48753643e-01
6.50693357e-01 -1.61624825e+00 -5.83583415e-02 -1.81180567e-01
-9.28489923e-01 9.71937954e-01 9.43624794e-01 2.60956939e-02
4.26776350e-01 1.40730813e-01 5.96701145e-01 -1.11066155e-01
-1.83746696e-01 -6.38336539e-01 -1.51087604e-02 1.17801917e+00
2.32521325e-01 1.34174779e-01 4.99178559e-01 -5.81909195e-02
-8.08288828e-02 -2.19110250e-01 3.03516269e-01 1.16312194e+00
-2.28290528e-01 -1.07423365e+00 -4.32068616e-01 3.24832052e-02
-1.27460072e-02 -1.51617110e-01 -6.78489983e-01 8.05984497e-01
1.41162336e-01 9.47820127e-01 4.01055336e-01 -4.81357962e-01
1.36551023e-01 1.30726933e-01 6.07342780e-01 -5.04262745e-01
-4.62507486e-01 -1.93078257e-02 5.59697801e-04 -8.40424001e-01
-1.15851736e+00 -7.92792320e-01 -1.39826465e+00 -7.84238279e-02
-1.36022940e-01 -1.66577935e-01 7.62310565e-01 8.33383381e-01
3.26696545e-01 4.52493489e-01 5.74629664e-01 -1.45010448e+00
2.43101195e-01 -1.05933046e+00 -4.21701282e-01 4.73819435e-01
5.04423022e-01 -5.82190573e-01 -2.31261119e-01 4.88618016e-01] | [8.84432601928711, -1.8350147008895874] |
2725e026-9fa7-44e1-966e-e66222c9d1a7 | training-transformers-with-4-bit-integers | 2306.11987 | null | https://arxiv.org/abs/2306.11987v2 | https://arxiv.org/pdf/2306.11987v2.pdf | Training Transformers with 4-bit Integers | Quantizing the activation, weight, and gradient to 4-bit is promising to accelerate neural network training. However, existing 4-bit training methods require custom numerical formats which are not supported by contemporary hardware. In this work, we propose a training method for transformers with all matrix multiplications implemented with the INT4 arithmetic. Training with an ultra-low INT4 precision is challenging. To achieve this, we carefully analyze the specific structures of activation and gradients in transformers to propose dedicated quantizers for them. For forward propagation, we identify the challenge of outliers and propose a Hadamard quantizer to suppress the outliers. For backpropagation, we leverage the structural sparsity of gradients by proposing bit splitting and leverage score sampling techniques to quantize gradients accurately. Our algorithm achieves competitive accuracy on a wide range of tasks including natural language understanding, machine translation, and image classification. Unlike previous 4-bit training methods, our algorithm can be implemented on the current generation of GPUs. Our prototypical linear operator implementation is up to 2.2 times faster than the FP16 counterparts and speeds up the training by up to 35.1%. | ['Jun Zhu', 'Jianfei Chen', 'Changhao Li', 'Haocheng Xi'] | 2023-06-21 | null | null | null | null | ['machine-translation'] | ['natural-language-processing'] | [ 1.65653795e-01 -3.60657781e-01 -3.91683310e-01 -4.05651271e-01
-5.45757174e-01 -1.84781268e-01 1.62712522e-02 2.79458493e-01
-8.16513181e-01 3.40233237e-01 -1.92171618e-01 -8.90386760e-01
4.20389384e-01 -7.46150374e-01 -1.01569438e+00 -4.01619762e-01
-1.38409913e-01 -1.46737350e-02 1.99587032e-01 -1.17388532e-01
2.36049935e-01 3.98808122e-01 -1.49116290e+00 7.36076713e-01
7.63727009e-01 1.55366206e+00 -1.30608097e-01 8.46769750e-01
-1.46882907e-01 8.35548282e-01 -6.81542635e-01 -5.55790126e-01
5.31679928e-01 4.55736704e-02 -4.37313139e-01 -5.00677466e-01
9.75239575e-01 -7.70926178e-01 -3.63078713e-01 1.20341086e+00
5.04821241e-01 -1.60392597e-01 2.07560733e-01 -9.65593696e-01
-3.50453556e-01 8.80589128e-01 -5.06785810e-01 4.14170891e-01
-2.93902516e-01 1.63674878e-03 9.24123824e-01 -1.05746591e+00
1.16220057e-01 1.11667514e+00 1.11393106e+00 3.88225019e-01
-9.52005267e-01 -9.50300992e-01 9.14222300e-02 4.09694433e-01
-1.48673069e+00 -5.06022334e-01 3.80580246e-01 -1.31059065e-01
1.37942874e+00 2.51585752e-01 8.21861267e-01 7.23468065e-01
1.15921296e-01 7.88447261e-01 8.12928379e-01 -3.90359491e-01
3.36551934e-01 -1.19418567e-02 2.49922752e-01 1.02005148e+00
3.96045387e-01 -1.10276252e-01 -9.98166740e-01 -1.09720141e-01
6.16310835e-01 5.55047393e-02 -2.02540442e-01 2.54824251e-01
-1.06065369e+00 8.98177981e-01 6.10030234e-01 -1.59735847e-02
-1.45418584e-01 7.74684668e-01 7.53248632e-01 4.21866566e-01
2.91665018e-01 1.78349271e-01 -5.33123374e-01 -2.95272768e-01
-1.31884122e+00 -2.50361376e-02 7.24093437e-01 8.29193532e-01
8.10653031e-01 4.46557492e-01 -1.70740306e-01 6.50821924e-01
6.71561658e-02 4.65828836e-01 7.89422333e-01 -8.78032088e-01
7.86756635e-01 4.13646460e-01 -3.66493613e-01 -8.09310198e-01
-3.66533071e-01 -5.50468087e-01 -1.23645997e+00 2.23544389e-01
4.83372390e-01 -1.83394417e-01 -1.00706339e+00 1.20960784e+00
2.41965652e-01 4.14558083e-01 -2.41625473e-01 7.68708527e-01
5.01402140e-01 7.62661099e-01 -2.35161677e-01 1.62246391e-01
1.21814764e+00 -1.09826922e+00 -4.96241897e-01 -3.31342995e-01
1.08980215e+00 -6.77264690e-01 1.14505029e+00 7.25557566e-01
-1.16346157e+00 -4.68210936e-01 -1.57221341e+00 -6.17078662e-01
-2.14871690e-01 5.01492262e-01 1.00964153e+00 9.97666597e-01
-1.16257143e+00 1.07857776e+00 -1.33098555e+00 4.56621349e-01
6.16504788e-01 7.08237886e-01 2.31594569e-03 9.43936780e-02
-9.86788929e-01 5.96666455e-01 5.28404474e-01 7.27084875e-02
-2.02256694e-01 -9.78167892e-01 -7.16231823e-01 1.95244208e-01
-2.27862805e-01 -4.20782179e-01 1.23553574e+00 -8.48786473e-01
-1.74334931e+00 5.02178609e-01 -2.09606007e-01 -1.02674019e+00
3.32735449e-01 -2.82164723e-01 -1.45212144e-01 2.34762114e-02
-4.37630445e-01 7.37093091e-01 1.01087153e+00 -1.27492979e-01
-5.74683368e-01 -2.88672745e-01 -3.70709985e-01 -1.42981419e-02
-1.06150794e+00 -1.84899911e-01 -4.10351992e-01 -7.75747180e-01
3.37830931e-01 -6.94536448e-01 -2.09235013e-01 3.58641148e-01
-1.55871332e-01 7.99840689e-02 5.00554979e-01 -4.51055974e-01
1.30883026e+00 -2.31833267e+00 -4.48977560e-01 5.30891895e-01
2.13713452e-01 5.18711329e-01 2.54319280e-01 -1.76817074e-01
1.67269260e-01 -2.18858019e-01 -4.97176871e-02 -6.84790075e-01
1.73103601e-01 2.73561299e-01 -6.79994345e-01 6.81633055e-01
1.93174183e-01 5.89800000e-01 -7.26341367e-01 -2.30826601e-01
-9.36615244e-02 6.47724628e-01 -1.22587800e+00 -2.22372830e-01
1.20693212e-02 -2.05444857e-01 -5.91792017e-02 6.44746363e-01
8.21326017e-01 -3.37450832e-01 1.04821570e-01 -5.68633795e-01
-2.22471088e-01 9.31804836e-01 -1.16997492e+00 1.75393784e+00
-4.35683131e-01 8.23888838e-01 -2.02935040e-02 -7.97538280e-01
8.93805504e-01 1.10071423e-02 1.25025779e-01 -6.75131619e-01
2.17278197e-01 6.14798725e-01 -9.13215950e-02 1.86482117e-01
7.33786941e-01 9.24007520e-02 2.26493970e-01 3.63555700e-01
5.57923727e-02 -1.34233668e-01 1.59884825e-01 -4.00960855e-02
1.12302065e+00 -4.07953262e-01 -8.33284557e-02 -1.03560217e-01
1.70587942e-01 -1.15995653e-01 5.35161197e-01 7.66479373e-01
-2.61294931e-01 4.40492690e-01 4.26120490e-01 -6.77323103e-01
-9.88981903e-01 -8.56994092e-01 -3.32260013e-01 1.48094940e+00
-3.26415777e-01 -8.59746456e-01 -6.88583314e-01 -8.20065364e-02
1.58909917e-01 2.63139486e-01 -9.40594599e-02 -1.91073298e-01
-6.99711323e-01 -1.00093663e+00 1.11209834e+00 6.90215647e-01
7.37345040e-01 -3.60524565e-01 -8.25320065e-01 1.84779838e-01
3.82834673e-01 -1.11716616e+00 -5.79754889e-01 7.18308449e-01
-1.31326914e+00 -4.44897830e-01 -4.98021305e-01 -9.11018491e-01
5.37427425e-01 -5.67834675e-02 8.56859505e-01 1.73011512e-01
-9.53392535e-02 -5.02264261e-01 9.04290602e-02 -3.22303891e-01
-1.08604707e-01 2.20627800e-01 2.33778656e-01 -2.53917783e-01
3.78411680e-01 -6.45775914e-01 -7.59140134e-01 -2.29034886e-01
-6.80437982e-01 1.50427356e-01 6.92539155e-01 1.09378409e+00
6.74343109e-01 -3.34834005e-03 -1.05505101e-01 -8.70381951e-01
4.17270452e-01 -8.46944824e-02 -6.96018636e-01 -8.06500539e-02
-6.70043468e-01 4.29541171e-01 1.06590962e+00 -7.30015934e-01
-5.12271285e-01 2.33856469e-01 -6.01025760e-01 -5.10603905e-01
4.09700483e-01 3.26075584e-01 5.69993913e-01 -5.65750897e-01
9.86946523e-01 -3.12862247e-02 -1.93721309e-01 -3.77794027e-01
3.52634430e-01 6.47440016e-01 6.65169775e-01 -5.54676175e-01
4.58096415e-01 4.99650091e-01 -2.63021514e-02 -6.59605622e-01
-7.51202345e-01 -3.28892797e-01 -1.49269566e-01 3.43601435e-01
2.02544659e-01 -1.18639171e+00 -8.98758411e-01 4.71135587e-01
-1.30012774e+00 -5.61756313e-01 -1.25494123e-01 5.91121733e-01
-2.07522251e-02 2.84790039e-01 -1.23632598e+00 -5.27422667e-01
-8.76025259e-01 -1.25064194e+00 9.65034366e-01 -1.52304202e-01
-8.92216712e-02 -5.99963903e-01 -2.54727781e-01 -3.31205229e-04
6.99563682e-01 -2.83235192e-01 5.96941650e-01 -1.01635903e-01
-6.02721810e-01 -6.96005747e-02 -3.89189571e-01 5.32116354e-01
-4.82632697e-01 -1.53554697e-02 -1.00627995e+00 -4.36452776e-01
2.57257789e-01 -5.05276203e-01 1.25414264e+00 1.36192694e-01
1.61975145e+00 -5.54955781e-01 1.70592945e-02 1.42624307e+00
1.21050394e+00 -2.89462954e-01 4.67585862e-01 2.35642582e-01
7.54621744e-01 -6.63305596e-02 1.78836703e-01 8.60988498e-01
2.38298655e-01 3.76851946e-01 2.13231191e-01 3.25540043e-02
-9.63812545e-02 -2.26084739e-01 6.13609672e-01 1.26438236e+00
2.02618644e-01 3.16517264e-01 -8.33917022e-01 1.26558870e-01
-1.31554639e+00 -5.97401917e-01 -1.81289002e-01 2.25200629e+00
1.34950101e+00 5.99737763e-01 -2.33020678e-01 5.71696997e-01
3.09291393e-01 1.31536856e-01 -5.48110008e-01 -8.47131550e-01
9.32031423e-02 9.90340769e-01 1.32193327e+00 4.39489961e-01
-1.05478692e+00 9.49588358e-01 6.56492043e+00 1.05986130e+00
-1.73319221e+00 1.95541918e-01 8.72320354e-01 -4.59174126e-01
-7.55487159e-02 -2.74910331e-01 -1.26773655e+00 4.30827171e-01
1.32040823e+00 3.16362739e-01 6.04227006e-01 1.07611918e+00
-1.53573558e-01 2.53121436e-01 -1.01944613e+00 1.36363471e+00
-1.37189925e-01 -1.44239795e+00 5.67792393e-02 -3.67063433e-01
8.54445219e-01 3.52508724e-01 5.23626387e-01 3.07228178e-01
-7.79714584e-02 -1.16707170e+00 8.46095622e-01 -6.76497966e-02
1.19827080e+00 -8.05145264e-01 3.97654951e-01 1.54782966e-01
-1.01863444e+00 -1.44606590e-01 -7.13194847e-01 -4.38479334e-01
-2.34168962e-01 1.02540791e+00 -8.52445781e-01 -2.88153738e-01
8.42774987e-01 4.99081820e-01 -5.20528793e-01 9.27343011e-01
-9.20178667e-02 8.56109679e-01 -8.52167487e-01 -2.85545886e-01
3.19758624e-01 5.66528551e-03 -1.69872254e-01 1.47230554e+00
6.16097927e-01 -1.79339826e-01 -2.49263123e-01 5.61498702e-01
-3.98567080e-01 1.72281235e-01 -7.46055841e-02 -2.43920125e-02
5.67117512e-01 9.63277757e-01 -5.44501960e-01 -5.80079913e-01
-3.66848826e-01 1.20823908e+00 7.33523190e-01 5.21043316e-02
-9.17859674e-01 -8.35050106e-01 7.63159215e-01 -1.60221040e-01
4.39282864e-01 -4.90864247e-01 -9.74123895e-01 -1.25090420e+00
2.62689263e-01 -1.04266667e+00 1.88243195e-01 -2.35661626e-01
-7.51557350e-01 4.46592689e-01 -5.67326725e-01 -8.66991401e-01
-1.01386502e-01 -1.18316019e+00 -4.60032821e-01 7.18378007e-01
-1.71079719e+00 -5.54545879e-01 -2.26523042e-01 5.96995950e-01
4.09756713e-02 -1.76975317e-02 8.19458604e-01 8.32027972e-01
-5.96759677e-01 1.31963062e+00 3.62221122e-01 9.15118381e-02
6.56387329e-01 -9.62173104e-01 9.39190388e-01 9.09850478e-01
3.33367437e-01 8.85444164e-01 3.91049594e-01 -3.65257382e-01
-1.86821735e+00 -1.17192721e+00 8.66344690e-01 3.46173704e-01
7.21310258e-01 -4.99012738e-01 -9.24307227e-01 5.67383945e-01
-3.09047967e-01 3.81985098e-01 6.05888247e-01 8.32218453e-02
-7.27145433e-01 -4.98738378e-01 -9.05226052e-01 7.67773032e-01
9.24989581e-01 -6.08525753e-01 3.44191082e-02 6.76128030e-01
6.08115971e-01 -1.15408480e+00 -6.86893523e-01 1.62824526e-01
5.35349727e-01 -9.73419607e-01 1.15536237e+00 -2.86588281e-01
3.37655514e-01 -1.94221273e-01 -1.71321645e-01 -9.76854146e-01
-9.20915753e-02 -8.31576765e-01 -6.32636547e-01 4.03566450e-01
2.75283515e-01 -6.81229770e-01 1.32411063e+00 3.72086495e-01
-3.58543515e-01 -8.75896096e-01 -1.04836833e+00 -5.37125230e-01
8.31452534e-02 -9.45527971e-01 5.71476221e-01 5.96980989e-01
2.26597428e-01 -2.29222067e-02 -3.30708563e-01 1.81846499e-01
5.64093471e-01 -9.22750756e-02 4.14164037e-01 -5.93061328e-01
-7.13910580e-01 -6.93346202e-01 -5.69110692e-01 -1.79800081e+00
-5.00220507e-02 -1.02873957e+00 -4.23006527e-02 -6.62755609e-01
-4.25359160e-01 -8.52813840e-01 -2.92208284e-01 5.98835349e-01
-7.29746148e-02 8.30102623e-01 -1.28839493e-01 -2.85358075e-02
-3.83603215e-01 4.29609388e-01 7.44570136e-01 -2.01963037e-01
-1.40051275e-01 -1.91351265e-01 -3.49310398e-01 7.70211935e-01
9.09963012e-01 -3.35817009e-01 -1.40877545e-01 -9.56532657e-01
5.64505577e-01 -4.25442070e-01 1.58945620e-01 -1.47844303e+00
7.07928419e-01 2.83895791e-01 5.31297624e-01 -6.32489920e-01
5.55588424e-01 -4.68195230e-01 -2.59268999e-01 6.97857022e-01
-2.74411082e-01 3.46958160e-01 4.66328055e-01 1.39506578e-01
-1.36542559e-01 -3.24380815e-01 8.07066262e-01 2.56187707e-01
-5.49067199e-01 3.38356465e-01 -2.75327027e-01 1.46348506e-01
3.52390021e-01 4.70073288e-03 -2.69298553e-01 2.43268465e-03
-1.27006724e-01 -1.21276334e-01 1.97726294e-01 -3.33530664e-01
8.19631279e-01 -1.33057070e+00 -3.60895663e-01 6.52873099e-01
-3.87155741e-01 1.92720711e-01 -3.25064845e-02 7.86968231e-01
-1.11862755e+00 4.91771996e-01 -1.48016244e-01 -5.74158251e-01
-1.15555716e+00 1.50969654e-01 2.52788156e-01 -1.95859969e-01
-5.55896223e-01 1.33375728e+00 -5.09851933e-01 3.45774516e-02
7.62581110e-01 -1.11017239e+00 3.53981078e-01 -2.67665863e-01
8.91039908e-01 3.76299918e-01 6.84735596e-01 -1.32587686e-01
-2.21134976e-01 3.01244527e-01 -2.18815014e-01 -8.03262275e-03
1.12895977e+00 5.10149181e-01 -1.52600557e-01 2.72882700e-01
1.54876709e+00 -3.16579752e-02 -1.24766695e+00 -2.15428367e-01
-9.90270302e-02 -3.79584908e-01 4.39600855e-01 -1.46794319e-01
-1.23395789e+00 1.23573542e+00 6.19181156e-01 -2.49009594e-01
1.22598493e+00 -7.97014534e-01 1.40553415e+00 8.60169590e-01
2.98560292e-01 -1.10632861e+00 -1.61107089e-02 1.01871729e+00
2.27048054e-01 -1.11114383e+00 1.73603594e-01 -3.84522080e-01
-5.93333580e-02 1.41647410e+00 3.20684344e-01 -2.53089875e-01
5.97334743e-01 8.05372119e-01 -6.21561743e-02 2.60382712e-01
-8.05365384e-01 1.74463570e-01 3.75359505e-01 1.05432257e-01
5.75326324e-01 -3.85737419e-02 -9.98988748e-02 3.28314304e-01
-6.88796639e-01 1.97022930e-02 2.90761322e-01 9.12490308e-01
-4.71223742e-01 -1.02824199e+00 -3.34855139e-01 6.80863738e-01
-7.28902459e-01 -7.88306177e-01 2.68507689e-01 7.96824694e-02
1.68103665e-01 5.76348960e-01 3.57489407e-01 -6.64475679e-01
-1.83716491e-02 8.29013661e-02 5.47343314e-01 -4.26219493e-01
-9.29677606e-01 -4.17725623e-01 -1.47756353e-01 -6.74891949e-01
1.76505953e-01 -2.76421219e-01 -1.46933019e+00 -8.78979862e-01
-1.94944516e-01 -2.35723510e-01 1.01223493e+00 5.30172288e-01
4.66219038e-01 3.16820651e-01 2.61717707e-01 -8.90075684e-01
-1.12733233e+00 -5.37087917e-01 -3.42666894e-01 1.59844175e-01
4.41854715e-01 -2.85492480e-01 -3.95016015e-01 -1.94361985e-01] | [8.53760814666748, 3.127624988555908] |
98ebee25-ad76-4cd2-b0ac-a66f9ec1506c | continuous-variable-neural-network-quantum | 2107.07105 | null | https://arxiv.org/abs/2107.07105v1 | https://arxiv.org/pdf/2107.07105v1.pdf | Continuous-variable neural-network quantum states and the quantum rotor model | We initiate the study of neural-network quantum state algorithms for analyzing continuous-variable lattice quantum systems in first quantization. A simple family of continuous-variable trial wavefunctons is introduced which naturally generalizes the restricted Boltzmann machine (RBM) wavefunction introduced for analyzing quantum spin systems. By virtue of its simplicity, the same variational Monte Carlo training algorithms that have been developed for ground state determination and time evolution of spin systems have natural analogues in the continuum. We offer a proof of principle demonstration in the context of ground state determination of a stoquastic quantum rotor Hamiltonian. Results are compared against those obtained from partial differential equation (PDE) based scalable eigensolvers. This study serves as a benchmark against which future investigation of continuous-variable neural quantum states can be compared, and points to the need to consider deep network architectures and more sophisticated training algorithms. | ['Giuseppe Carleo', 'Shravan Veerapaneni', 'Saibal De', 'James Stokes'] | 2021-07-15 | null | null | null | null | ['variational-monte-carlo'] | ['miscellaneous'] | [ 3.41242105e-01 -1.67556837e-01 -1.05092771e-01 -2.36424655e-01
-5.94091535e-01 -4.28243816e-01 8.55168939e-01 -6.18506610e-01
-6.97594285e-01 1.26164913e+00 -2.71144986e-01 -6.62811220e-01
-2.80968994e-01 -1.03542399e+00 -3.01080644e-01 -1.32331312e+00
-1.94695488e-01 8.58721614e-01 -3.55806172e-01 -6.55027688e-01
3.87351334e-01 5.45468986e-01 -1.22434819e+00 -2.21987337e-01
5.26832640e-01 5.71771443e-01 -2.42810443e-01 8.23034465e-01
2.13741481e-01 6.92466974e-01 -1.03998788e-01 4.45634350e-02
1.47478759e-01 -9.80636954e-01 -1.18660247e+00 -4.97160643e-01
4.42760974e-01 -9.45657492e-02 -7.47103393e-01 1.35617983e+00
5.65823615e-01 7.49448895e-01 9.31717873e-01 -5.88451028e-01
-9.09543157e-01 3.21948975e-01 2.56133348e-01 5.38175225e-01
-6.35065213e-02 5.79200804e-01 1.27914572e+00 -4.87661868e-01
8.72040391e-01 8.46602798e-01 6.56031668e-01 9.79509473e-01
-1.80576706e+00 -4.02232736e-01 -1.11570454e+00 4.37885255e-01
-1.48401606e+00 -6.25856102e-01 6.50928319e-01 -2.01970607e-01
1.67963159e+00 9.11730379e-02 7.65927613e-01 1.08099103e+00
8.72206688e-01 1.58326745e-01 1.39912164e+00 -7.06751943e-01
5.54450691e-01 -3.55763584e-01 4.69777137e-01 1.08833385e+00
2.40830123e-01 8.37590575e-01 -5.49282193e-01 -3.60985994e-01
7.11801887e-01 -6.39164805e-01 -1.02554694e-01 -6.26632214e-01
-1.34055185e+00 1.04700398e+00 3.98344636e-01 3.55929345e-01
-4.71231550e-01 6.57400846e-01 4.78126228e-01 2.42534131e-01
-1.80333346e-01 8.32043767e-01 -1.51576608e-01 -3.30922037e-01
-1.18989670e+00 4.99234498e-01 9.54596043e-01 2.40099430e-01
9.36306000e-01 6.86761022e-01 -7.74234831e-02 2.70335916e-02
1.80184603e-01 1.02403677e+00 2.66985029e-01 -1.30364561e+00
-2.56390125e-01 -3.30340922e-01 2.95933098e-01 -3.91906470e-01
-5.34067929e-01 -5.22424206e-02 -1.23149979e+00 2.23013118e-01
2.27168173e-01 -1.53077487e-02 -7.87756324e-01 1.66192102e+00
2.18672678e-02 1.31637067e-01 4.06971306e-01 9.60096061e-01
4.82032120e-01 6.59831882e-01 -2.34874740e-01 -5.80755711e-01
8.62378240e-01 -4.11562711e-01 -8.43817234e-01 5.97724855e-01
8.96488428e-01 -4.97489572e-02 5.56782663e-01 4.53349531e-01
-1.12779772e+00 -3.67637515e-01 -1.15106690e+00 -3.82952727e-02
-3.07209432e-01 -5.65940082e-01 1.23313749e+00 8.95036817e-01
-1.34853983e+00 1.52847660e+00 -1.33874369e+00 -9.69957039e-02
4.96994741e-02 6.49674237e-01 -2.77677208e-01 4.83259976e-01
-1.65488994e+00 1.19747376e+00 5.33739805e-01 1.63761735e-01
-1.15372574e+00 1.01938155e-02 -5.89416444e-01 -3.49741399e-01
-1.97719947e-01 -9.05982196e-01 1.35495961e+00 -4.97487634e-01
-1.83202577e+00 8.98216307e-01 -3.93178254e-01 -5.23456633e-01
-3.59270960e-01 5.67418337e-01 -4.33272004e-01 7.90156201e-02
-6.47547618e-02 2.04244226e-01 5.10073602e-01 -5.07981181e-01
2.87978828e-01 -2.48587951e-01 2.29820255e-02 -2.02762693e-01
2.12822095e-01 -3.63700420e-01 5.38786769e-01 3.20792168e-01
3.67986470e-01 -1.29382896e+00 -3.81716818e-01 -8.32823753e-01
-3.30403775e-01 -3.03948849e-01 9.02031511e-02 -1.98435262e-01
1.16645610e+00 -1.49352849e+00 7.22382367e-01 3.64002675e-01
1.21950351e-01 2.34784827e-01 8.77880082e-02 5.78859210e-01
-2.16537341e-01 -3.74925584e-02 -3.35467130e-01 1.45459613e-02
2.70783305e-01 4.28360879e-01 -4.07334298e-01 7.49121189e-01
1.56641573e-01 1.19868290e+00 -9.53564405e-01 -3.73432398e-01
-2.64619328e-02 5.04270375e-01 -8.56108725e-01 -2.90395319e-01
-2.22251743e-01 6.79591358e-01 -5.59508562e-01 2.21605539e-01
3.55086744e-01 -4.92881060e-01 2.17990473e-01 -2.62932181e-01
-4.37520206e-01 6.55118406e-01 -9.49014843e-01 1.76490569e+00
-1.19063295e-01 5.67103684e-01 6.80391714e-02 -1.03213978e+00
5.71844816e-01 5.44874251e-01 3.92129093e-01 -9.60416913e-01
4.63310212e-01 4.01383340e-01 5.53719640e-01 -4.72212911e-01
9.13807452e-01 -8.66981447e-01 -1.93662301e-01 7.80358434e-01
4.41304475e-01 -6.04726553e-01 3.01172972e-01 2.39270180e-01
9.77887213e-01 2.85837382e-01 3.11493903e-01 -7.31796861e-01
5.81486404e-01 1.60181090e-01 2.83368286e-02 1.18839669e+00
-6.65353060e-01 1.38925195e-01 2.17013761e-01 -7.80597866e-01
-1.37447083e+00 -1.24718201e+00 -7.02082753e-01 5.62111795e-01
1.73308134e-01 -5.27105331e-01 -7.33178556e-01 3.65917593e-01
-3.46489221e-01 7.53995240e-01 -6.20223582e-01 -2.06153184e-01
-5.90552032e-01 -1.19357085e+00 8.07959259e-01 -3.38380749e-04
3.44538689e-01 -1.48223448e+00 -6.09299958e-01 3.19452405e-01
7.48264864e-02 -4.52325732e-01 3.80761474e-01 9.21136200e-01
-8.89402092e-01 -8.94919872e-01 -4.40867320e-02 -4.19706047e-01
-9.11757499e-02 -4.20284659e-01 1.20151639e+00 -8.34273547e-02
-4.78548616e-01 2.17709973e-01 2.25534916e-01 3.19059372e-01
-8.08765113e-01 3.35518777e-01 8.11860085e-01 -6.43451154e-01
7.85647333e-01 -6.55659318e-01 -3.85141313e-01 -4.57991570e-01
-6.81182206e-01 -3.71299274e-02 6.34540766e-02 1.23395586e+00
7.12176681e-01 1.77965276e-02 -3.03114555e-03 -5.82738698e-01
6.55758202e-01 -2.04706594e-01 -6.95218682e-01 3.61920819e-02
-7.45018065e-01 8.77832770e-01 7.44763613e-01 1.36575297e-01
-7.95422316e-01 -3.32286954e-01 -1.66150525e-01 3.38312872e-02
-1.40582487e-01 5.38963199e-01 6.44620597e-01 -4.09322619e-01
1.06054628e+00 6.06026471e-01 8.02130699e-02 2.57591400e-02
3.72407496e-01 3.78492713e-01 5.89334488e-01 -8.52404833e-01
9.52992022e-01 5.43894947e-01 9.01912630e-01 -9.45882976e-01
-6.95653737e-01 -4.54787500e-02 -9.78428841e-01 1.43697500e-01
1.25714839e+00 -3.18989187e-01 -1.20378613e+00 5.17377555e-01
-1.05996788e+00 -4.35951620e-01 -3.65028083e-01 5.06990254e-01
-1.08359635e+00 5.09935260e-01 -1.01282370e+00 -1.05671263e+00
-4.41262037e-01 -1.22228420e+00 7.66593337e-01 1.20855302e-01
-1.53949410e-01 -1.28405917e+00 8.32334995e-01 9.97032076e-02
6.82107806e-01 1.22079819e-01 9.20000434e-01 -3.64369676e-02
-8.59503627e-01 -1.09886535e-01 3.93449724e-01 1.89790487e-01
-4.10976350e-01 -4.34603309e-04 -8.66083205e-01 -5.41730821e-01
1.30378619e-01 -7.62138069e-01 1.15450478e+00 4.49235141e-01
6.36945844e-01 1.40679151e-01 -2.77946126e-02 8.08110476e-01
1.51347518e+00 -1.82360858e-02 6.47752464e-01 2.81415761e-01
5.85240304e-01 -1.31471679e-01 -2.82795638e-01 2.39956319e-01
4.11651991e-02 4.66869980e-01 1.40567914e-01 3.93420696e-01
2.70681202e-01 3.37160885e-01 4.78812367e-01 1.28606987e+00
-6.79342151e-01 1.49115399e-01 -9.31271970e-01 7.88862929e-02
-1.37424874e+00 -1.72694671e+00 -4.44752991e-01 1.98235321e+00
8.71019125e-01 -2.98925377e-02 -1.35058999e-01 1.25433400e-01
3.29569042e-01 4.52979028e-01 -6.07828200e-01 -8.16016197e-01
-1.73909888e-01 1.05243206e+00 5.10856390e-01 9.05881047e-01
-1.10436141e+00 1.09500217e+00 7.45835066e+00 6.93148375e-01
-1.11024761e+00 3.19773465e-01 4.59772535e-02 -2.64604390e-01
-2.58056372e-01 2.97651321e-01 -6.91371679e-01 1.25489876e-01
1.68420267e+00 -6.86408058e-02 1.43260324e+00 3.27705681e-01
1.32631496e-01 -1.10737979e-01 -1.02514839e+00 9.74085748e-01
-2.36879542e-01 -1.73236954e+00 -2.16913238e-01 2.66705662e-01
1.05652165e+00 6.69698596e-01 1.27920851e-01 5.20384490e-01
1.96693063e-01 -1.27158952e+00 5.09303510e-01 6.90871656e-01
1.02559114e+00 -6.81582153e-01 3.92981380e-01 1.43620580e-01
-7.19675422e-01 2.89893299e-01 -5.16823709e-01 -5.81268609e-01
4.50042814e-01 3.38666141e-02 -2.91943997e-01 2.46987224e-01
9.37599689e-02 4.20756072e-01 -1.13220237e-01 3.02836746e-01
-2.47181058e-02 6.84417784e-01 -2.08687097e-01 -4.53315765e-01
6.14969015e-01 -7.51565754e-01 5.02428234e-01 8.13606679e-01
-6.63049296e-02 2.90883094e-01 -2.08561689e-01 1.49092090e+00
2.22125649e-01 -3.32660377e-01 -7.10609019e-01 -4.15236205e-01
1.74261600e-01 1.12966502e+00 -6.14738286e-01 -3.20812076e-01
-4.27555554e-02 9.57779348e-01 3.95735711e-01 5.37620783e-01
-6.57789171e-01 -4.98969764e-01 5.56385934e-01 -2.36088946e-01
3.41001630e-01 -7.72854388e-01 -9.83432457e-02 -1.65144539e+00
-6.23758912e-01 -6.03137672e-01 -1.41311496e-01 -6.53531313e-01
-1.06093633e+00 5.81319988e-01 -5.31146266e-02 -4.00787324e-01
-7.59780169e-01 -1.01215935e+00 -5.59119403e-01 1.25111449e+00
-1.25039566e+00 -4.62733418e-01 3.54664475e-01 5.38162172e-01
-6.23893797e-01 -2.63863742e-01 1.66234255e+00 -2.89929658e-01
-6.82632506e-01 1.46179363e-01 8.65389228e-01 1.01335734e-01
-2.09806189e-01 -1.31601679e+00 5.56649685e-01 7.09130764e-01
4.09302294e-01 1.17634869e+00 9.65643883e-01 -3.80668759e-01
-1.97919369e+00 -3.81780595e-01 6.56320810e-01 -8.07843924e-01
8.58381689e-01 -1.17831036e-01 -7.20716596e-01 3.95171225e-01
2.28593707e-01 -2.02320851e-02 6.13074481e-01 4.55884278e-01
-1.07515633e-01 3.96247655e-01 -8.60325217e-01 3.45267355e-01
8.23727846e-01 -1.44756937e+00 -5.57784915e-01 3.68743062e-01
1.68211069e-02 -2.46539012e-01 -8.05216193e-01 2.69598842e-01
6.25529289e-01 -1.12701714e+00 8.20961773e-01 -1.06683397e+00
2.81320244e-01 -1.62349612e-01 -3.95700634e-01 -1.06534016e+00
-7.85382688e-01 -9.18621242e-01 -2.84144819e-01 9.48579237e-02
5.53723276e-02 -4.66990829e-01 7.76877880e-01 5.46577394e-01
1.06386223e-03 -5.53624570e-01 -1.55335605e+00 -7.52101719e-01
7.17232585e-01 -5.51501632e-01 4.25879285e-02 1.01151407e+00
2.32277140e-01 5.15641868e-01 -2.60494441e-01 -6.90716282e-02
8.34823668e-01 2.68046468e-01 2.82230794e-01 -1.02510774e+00
-5.03529608e-01 -5.15976906e-01 -4.77306753e-01 -7.27351904e-01
7.21141458e-01 -1.60029137e+00 9.82068703e-02 -1.15849543e+00
4.92683887e-01 -2.91843060e-02 -7.17208803e-01 8.13040733e-02
3.02564293e-01 6.31305516e-01 -3.35457474e-01 3.16056877e-01
-8.39943409e-01 7.69018710e-01 1.20363748e+00 9.25945342e-02
-6.47074059e-02 -5.22623003e-01 9.71308574e-02 2.17218652e-01
6.63826048e-01 -5.27416945e-01 3.73054296e-02 -2.84960531e-02
6.56323671e-01 2.61322498e-01 6.30869687e-01 -1.14606690e+00
1.34100080e-01 -2.82925934e-01 9.64871868e-02 -3.82578194e-01
5.20171165e-01 9.56893414e-02 1.34041920e-01 8.43381345e-01
-3.43359590e-01 2.93218996e-02 -2.50856280e-01 2.23188147e-01
1.64987415e-01 -4.02449340e-01 1.09361386e+00 -5.58475196e-01
-7.32600451e-01 3.71948034e-01 -7.59261668e-01 1.03712358e-01
2.70293444e-01 -1.64256096e-01 -2.34097525e-01 -1.19020253e-01
-8.13673079e-01 -4.56931710e-01 6.89537942e-01 -4.51879591e-01
2.11142719e-01 -1.37223279e+00 -3.63581181e-01 4.68724459e-01
-2.02830777e-01 -5.93756974e-01 3.26908529e-01 1.01622760e+00
-9.21806753e-01 9.14797187e-01 -3.91656607e-01 -5.14443398e-01
-4.39810127e-01 4.63633001e-01 8.58713508e-01 -1.58204421e-01
-3.62155616e-01 7.15974331e-01 -3.66838187e-01 -4.94741976e-01
-5.06247044e-01 -1.37694821e-01 4.56903547e-01 -3.66068065e-01
3.06894422e-01 2.90147364e-01 -1.27796322e-01 -1.05824077e+00
-3.33916575e-01 7.42974132e-02 3.85032862e-01 -4.40105379e-01
1.38836861e+00 2.02978894e-01 -5.97925365e-01 7.36261189e-01
1.16547251e+00 -4.77054477e-01 -8.05827260e-01 -1.84538677e-01
-1.79367483e-01 2.87529558e-01 5.24460971e-01 -3.70220661e-01
-4.96627003e-01 1.13201618e+00 6.90910876e-01 3.97823840e-01
3.63592476e-01 -1.90955460e-01 7.43471205e-01 1.25647533e+00
7.84119368e-01 -1.13422477e+00 -2.88248986e-01 1.19977498e+00
1.75928012e-01 -1.10493219e+00 -2.49218233e-02 7.93923438e-01
-9.94861498e-02 1.48586988e+00 2.62014605e-02 -5.58274806e-01
7.05085635e-01 6.14959858e-02 -2.67976075e-01 -7.33847976e-01
-7.25511789e-01 -2.27541160e-02 1.37838215e-01 3.95991683e-01
7.27419794e-01 3.42956573e-01 -2.30267659e-01 6.96408078e-02
-5.22707105e-01 1.59291457e-02 7.90371776e-01 6.94234431e-01
-4.70323205e-01 -1.15811145e+00 -2.15379357e-01 4.29178625e-01
-3.72941643e-01 -4.79184270e-01 1.35585085e-01 5.97412646e-01
5.39766140e-02 4.42460597e-01 -1.14828184e-01 -2.33579785e-01
-3.19038868e-01 7.47759879e-01 1.25228071e+00 -4.99252409e-01
-1.99286029e-01 -6.24559462e-01 -1.18228178e-02 -5.48786700e-01
-6.34887695e-01 -7.33196914e-01 -1.47077858e+00 -7.77492523e-01
-6.32601678e-01 4.47402686e-01 6.20285809e-01 1.13261938e+00
-6.14319630e-02 3.09981853e-01 1.80602342e-01 -1.31458151e+00
-1.10066748e+00 -8.70488226e-01 -1.07974517e+00 3.18876177e-01
6.37970507e-01 -6.63989663e-01 -3.70064229e-01 -2.47284621e-01] | [5.505845069885254, 4.998385429382324] |
cf540f0a-724e-4303-a276-01f3e75ac591 | text-data-augmentation-towards-better | 2007.02033 | null | https://arxiv.org/abs/2007.02033v2 | https://arxiv.org/pdf/2007.02033v2.pdf | Text Data Augmentation: Towards better detection of spear-phishing emails | Text data augmentation, i.e., the creation of new textual data from an existing text, is challenging. Indeed, augmentation transformations should take into account language complexity while being relevant to the target Natural Language Processing (NLP) task (e.g., Machine Translation, Text Classification). Initially motivated by an application of Business Email Compromise (BEC) detection, we propose a corpus and task agnostic augmentation framework used as a service to augment English texts within our company. Our proposal combines different methods, utilizing BERT language model, multi-step back-translation and heuristics. We show that our augmentation framework improves performances on several text classification tasks using publicly available models and corpora as well as on a BEC detection task. We also provide a comprehensive argumentation about the limitations of our augmentation framework. | ['Sébastien Goutal', 'Mehdi Regina', 'Maxime Meyer'] | 2020-07-04 | null | null | null | null | ['text-augmentation'] | ['natural-language-processing'] | [ 7.24293590e-01 4.13080841e-01 -4.17004600e-02 -2.56461114e-01
-9.74682450e-01 -6.32786751e-01 1.08991146e+00 8.13793063e-01
-7.46322989e-01 7.24458039e-01 3.11992079e-01 -6.93883777e-01
1.36654109e-01 -7.34286308e-01 -6.12795651e-01 -1.05270460e-01
7.16336608e-01 9.51254249e-01 -1.03026733e-01 -5.12123227e-01
4.30913627e-01 5.68588734e-01 -1.16615903e+00 7.70565391e-01
9.25385952e-01 7.36767054e-01 -9.13916975e-02 6.43716216e-01
-7.54657924e-01 7.30837464e-01 -5.20940900e-01 -8.75103593e-01
2.80814976e-01 -2.13813391e-02 -1.38880062e+00 3.87586802e-01
2.55055845e-01 1.26168221e-01 1.06125735e-01 7.65453160e-01
8.58654603e-02 -1.61544178e-02 6.04783356e-01 -1.30049336e+00
-3.81483495e-01 9.84186947e-01 -3.40557128e-01 7.51613304e-02
2.14220718e-01 -3.68748382e-02 8.73654544e-01 -1.10143411e+00
8.41031909e-01 9.44286525e-01 6.49812639e-01 5.34567237e-01
-1.31364417e+00 -1.65149122e-01 2.39239931e-01 -9.51480493e-02
-8.40977311e-01 -3.54321301e-01 7.11614847e-01 -4.31932271e-01
1.02513099e+00 4.38166231e-01 1.91661000e-01 1.40196729e+00
-8.21363702e-02 9.13372636e-01 1.25100291e+00 -1.15531158e+00
-1.38280513e-02 8.44529450e-01 5.18778026e-01 2.57155716e-01
2.47167736e-01 -5.16747594e-01 -3.25888813e-01 -3.10960352e-01
1.05030514e-01 -3.70557368e-01 1.18126377e-01 -5.18108979e-02
-1.09446275e+00 8.54773700e-01 -1.46437243e-01 7.89264679e-01
-4.62850869e-01 -2.74087012e-01 8.57870281e-01 3.12932134e-01
7.35897303e-01 9.28851426e-01 -8.52927089e-01 -1.75566956e-01
-7.22568631e-01 1.41314507e-01 1.03701234e+00 1.18790734e+00
7.08822727e-01 -3.32002074e-01 -3.24245065e-01 9.23033893e-01
5.22704720e-02 2.80174501e-02 7.08419681e-01 -2.99781024e-01
9.60396886e-01 1.06873000e+00 8.09290707e-02 -5.89812875e-01
-5.32444656e-01 -3.92528057e-01 -6.78925157e-01 -3.13621342e-01
3.97850960e-01 -2.06346244e-01 -6.92555904e-01 1.30256069e+00
3.02946419e-01 -3.68331015e-01 1.57081783e-01 2.32035443e-01
5.93709409e-01 5.27921498e-01 2.41409630e-01 -4.85010684e-01
1.55958390e+00 -1.15790486e+00 -8.53693128e-01 -4.74911273e-01
1.30409276e+00 -1.11339474e+00 1.30250478e+00 4.49086815e-01
-8.30566227e-01 -3.14885795e-01 -7.26543128e-01 -2.20684171e-01
-9.19063747e-01 3.94522846e-01 4.17723268e-01 9.68425751e-01
-6.15652561e-01 2.05450013e-01 -2.99214184e-01 -8.98904979e-01
9.56034139e-02 2.31620193e-01 -3.86340648e-01 1.84429631e-01
-1.15550423e+00 9.23570991e-01 6.31224632e-01 -1.27581477e-01
-1.83236808e-01 -5.04352927e-01 -6.16053462e-01 -4.53262702e-02
7.42702842e-01 -7.59385765e-01 1.37684798e+00 -9.67746913e-01
-1.32056308e+00 1.07389164e+00 -6.02793805e-02 -9.20715451e-01
7.64580548e-01 -4.46416199e-01 -4.73080814e-01 -7.49081448e-02
-1.23170301e-01 2.46099859e-01 7.18888640e-01 -1.27476895e+00
-6.88889682e-01 -3.24789912e-01 -2.28917133e-02 -5.08983247e-03
-1.06730127e+00 5.30710995e-01 -1.69342488e-01 -8.62432480e-01
-2.83853114e-01 -9.48454380e-01 -5.45618176e-01 -5.55774629e-01
-7.06512153e-01 -1.90777406e-01 8.68267417e-01 -9.61266696e-01
1.43044412e+00 -1.75063550e+00 -1.83519989e-01 4.22948062e-01
4.49726246e-02 3.64952981e-01 -2.36533836e-01 6.64576352e-01
-2.65197664e-01 7.41555512e-01 -3.61360133e-01 -6.65341735e-01
1.46114647e-01 7.32523203e-02 -3.74177039e-01 -2.76039541e-03
4.62930381e-01 9.26009834e-01 -4.87009108e-01 -6.64280057e-01
8.29037353e-02 3.51084657e-02 -3.63957703e-01 -8.70762169e-02
-4.61042464e-01 2.44012028e-01 -4.06140149e-01 4.75359261e-01
4.49784040e-01 4.68668528e-02 7.61638135e-02 7.39372969e-02
-2.90774882e-01 4.97386783e-01 -1.02202153e+00 1.48915648e+00
-7.69475937e-01 5.15338778e-01 -1.34706452e-01 -8.79601657e-01
9.43541348e-01 4.30954158e-01 5.21355212e-01 -2.40334123e-01
4.23176676e-01 2.56419480e-01 -1.07801937e-01 -4.88390058e-01
9.78433609e-01 -4.70904596e-02 -1.14014104e-01 7.47716784e-01
2.29224674e-02 -1.51921540e-01 5.86371541e-01 2.62804866e-01
9.70024109e-01 -1.22769680e-02 5.62450349e-01 -1.52146488e-01
9.94649708e-01 4.47788149e-01 3.40936892e-02 6.37879729e-01
8.26158747e-02 3.09064686e-01 6.07872009e-01 -2.22281069e-01
-1.47724187e+00 -6.29667491e-02 -1.22095086e-01 1.20058882e+00
-5.63899636e-01 -6.43877506e-01 -9.21191573e-01 -1.25814307e+00
-1.77270561e-01 1.03297818e+00 -5.94429970e-01 -4.48103659e-02
-7.59446263e-01 -9.39092755e-01 6.94940984e-01 3.41444910e-01
1.24292210e-01 -1.00882113e+00 -5.87421432e-02 4.98702854e-01
-5.49940228e-01 -1.73614824e+00 -2.87711084e-01 3.25518489e-01
-9.19467151e-01 -8.37244391e-01 -2.77631283e-01 -7.19075382e-01
5.86666226e-01 -2.69383285e-02 1.05370259e+00 -5.72767993e-03
-2.15688318e-01 6.43796742e-01 -7.57562816e-01 -7.51274049e-01
-1.17322695e+00 6.46274388e-01 -8.52231830e-02 1.12677790e-01
4.68949497e-01 -1.37271166e-01 1.61548734e-01 2.36889884e-01
-1.12211752e+00 1.55575931e-01 7.55371511e-01 8.80313277e-01
2.89949566e-01 -9.42531377e-02 6.34933650e-01 -1.26279938e+00
1.01778054e+00 -2.22327188e-01 -3.49786580e-01 4.79689300e-01
-9.05251324e-01 1.98368412e-02 7.41410434e-01 -4.19907987e-01
-1.26800609e+00 3.01553071e-01 -4.83865321e-01 7.47069195e-02
-5.42782426e-01 5.89440346e-01 -2.24077538e-01 -1.64111201e-02
9.51479733e-01 2.61931360e-01 -2.61402071e-01 -6.21602893e-01
5.96134543e-01 9.65709865e-01 1.62453130e-01 -5.24971604e-01
1.02795780e+00 2.47783661e-01 -2.06884578e-01 -7.45288789e-01
-6.23450875e-01 -8.13640475e-01 -1.19958949e+00 -8.33824128e-02
5.84177494e-01 -3.61332804e-01 -3.40819418e-01 9.01507437e-02
-1.49561441e+00 6.63898606e-03 -5.20846426e-01 3.27419579e-01
-5.21366954e-01 4.74309862e-01 -4.66770351e-01 -8.72357011e-01
-6.86839461e-01 -7.00464547e-01 9.54795361e-01 -2.48148963e-01
-4.90197241e-01 -8.18451047e-01 -1.38562649e-01 9.17855442e-01
1.64964467e-01 3.24216634e-02 1.22858036e+00 -1.42807472e+00
-3.15093249e-03 -6.85387254e-01 -8.04081783e-02 5.55418015e-01
6.80454001e-02 1.81224987e-01 -9.79868889e-01 -8.70852321e-02
2.65833944e-01 -3.42365295e-01 5.50322354e-01 -2.98270553e-01
1.06746387e+00 -6.91175640e-01 -2.69141018e-01 -1.18472114e-01
1.12398779e+00 5.29528968e-02 5.48626542e-01 8.93395066e-01
5.05787969e-01 1.03569603e+00 8.83661866e-01 4.32279527e-01
7.77759552e-02 5.86287200e-01 1.93491012e-01 -1.29572824e-01
1.41167000e-01 -1.35459155e-01 2.60331392e-01 6.85280800e-01
1.20956972e-01 -4.49781418e-01 -1.32010865e+00 3.57447058e-01
-1.76158404e+00 -5.40435135e-01 -6.44747913e-01 1.89672625e+00
8.19699407e-01 4.61381346e-01 2.65081614e-01 5.27314067e-01
6.93097591e-01 -4.10865784e-01 -6.44819513e-02 -7.80035794e-01
-1.29797325e-01 1.81102067e-01 4.73756254e-01 2.79496729e-01
-1.19123721e+00 1.01882970e+00 5.48109055e+00 7.37339199e-01
-8.24993730e-01 4.63741124e-01 6.47641242e-01 2.95250207e-01
-1.13444820e-01 -3.78074534e-02 -9.18401241e-01 1.55782536e-01
9.01563287e-01 -4.51720059e-01 5.47201522e-02 8.06475282e-01
1.62674785e-01 2.25801080e-01 -1.16847193e+00 5.16803086e-01
1.77002370e-01 -1.33205342e+00 3.59505326e-01 1.01469241e-01
5.86110473e-01 -1.93529561e-01 -7.86780119e-02 4.81240094e-01
1.45871893e-01 -5.55851340e-01 6.03393078e-01 5.50035350e-02
3.46047223e-01 -6.71829104e-01 1.10310209e+00 6.42921269e-01
-7.77535915e-01 -2.12347791e-01 2.13683695e-01 1.71301395e-01
9.57549661e-02 4.77901459e-01 -1.35376430e+00 7.80248642e-01
4.00885582e-01 1.05450809e-01 -9.43127155e-01 7.19496608e-01
-8.88801292e-02 5.55609584e-01 -2.38172054e-01 -1.32153153e-01
4.39882189e-01 -7.73750097e-02 6.89223111e-01 1.52638936e+00
1.01627432e-01 -1.27466649e-01 1.06330894e-01 6.42464399e-01
-3.91451448e-01 1.02062058e+00 -5.36791980e-01 -4.41413254e-01
1.33611292e-01 1.62486172e+00 -7.30752110e-01 -5.14534652e-01
-3.91866386e-01 7.78546929e-01 7.23486543e-02 8.75711143e-02
-5.46991229e-01 -3.12549204e-01 -2.18500257e-01 2.97519356e-01
-3.17930192e-01 -4.13626619e-02 -9.01240647e-01 -1.10767412e+00
3.83848041e-01 -1.14351034e+00 4.60389048e-01 -5.66511989e-01
-1.22608054e+00 8.92135799e-01 -2.68450826e-01 -1.28809714e+00
-2.82337815e-01 -6.44463301e-01 -2.95328945e-01 7.26196945e-01
-1.33383870e+00 -1.55504453e+00 -2.45148331e-01 4.34904873e-01
7.97340572e-01 -2.80685991e-01 8.68540764e-01 5.50983965e-01
-6.09362602e-01 6.29107356e-01 -1.26074225e-01 1.61885113e-01
8.67329299e-01 -1.27743888e+00 5.81611514e-01 1.00242519e+00
1.98385626e-01 5.27416885e-01 6.66944683e-01 -6.35724187e-01
-1.11367631e+00 -1.32707763e+00 1.40914905e+00 -6.20183289e-01
1.17983270e+00 -7.29693830e-01 -9.87213314e-01 7.63526797e-01
3.63155454e-01 -5.86313128e-01 7.36080527e-01 2.23373741e-01
-1.14619337e-01 1.78721786e-01 -1.18027091e+00 7.73894489e-01
6.91656053e-01 -4.81849492e-01 -7.71901429e-01 8.70792747e-01
7.56106973e-01 -1.76833868e-01 -8.59960556e-01 2.42867991e-01
1.44128814e-01 -2.68874586e-01 6.38692141e-01 -1.07738745e+00
6.14656568e-01 1.57614440e-01 8.89939629e-03 -1.10479379e+00
7.65661970e-02 -8.19401622e-01 6.87465817e-02 1.77518272e+00
9.04312968e-01 -7.29427278e-01 7.30654836e-01 9.32565570e-01
-1.48819521e-01 -4.88222659e-01 -7.56292522e-01 -8.37354660e-01
2.79398769e-01 -8.07894051e-01 5.51289439e-01 1.16165078e+00
2.82770604e-01 6.74310207e-01 -2.29245856e-01 -5.39782532e-02
1.74355790e-01 -2.76813418e-01 1.06298888e+00 -1.31079221e+00
1.33587003e-01 -4.73052591e-01 5.33808842e-02 -5.98319232e-01
2.57806391e-01 -1.02040851e+00 -1.04162768e-01 -1.44638288e+00
2.13501245e-01 -3.96492451e-01 1.02471970e-01 7.08028615e-01
-2.82671303e-02 2.97738370e-02 3.38824809e-01 2.66912758e-01
-1.50261432e-01 3.41432899e-01 9.32876408e-01 -3.27755630e-01
-2.90689915e-01 1.19532093e-01 -8.20583522e-01 9.32554603e-01
1.01852775e+00 -7.43740141e-01 1.61273852e-01 -1.82743549e-01
3.74769509e-01 -2.67149031e-01 9.87699479e-02 -5.54789424e-01
2.92095006e-01 -6.55588210e-02 -9.83479694e-02 -5.93686879e-01
1.14935853e-01 -1.12316549e+00 -4.31657314e-01 5.25721073e-01
-7.28374839e-01 1.57417059e-01 4.05265331e-01 3.29685152e-01
-2.09088951e-01 -9.52915192e-01 6.54310465e-01 -1.43727157e-02
-2.99498320e-01 -5.22359088e-02 -7.81943142e-01 -2.04418302e-01
1.00047982e+00 -1.24417983e-01 -4.43194002e-01 -5.50461113e-02
-8.00394714e-01 1.26062468e-01 2.76477009e-01 5.93557000e-01
2.23406911e-01 -1.05555749e+00 -9.12445247e-01 -1.26019761e-01
5.00361145e-01 -3.37920815e-01 -1.76718622e-01 9.12566543e-01
-3.39976102e-01 6.49635434e-01 -4.85010631e-02 -1.99182779e-01
-1.63184226e+00 7.22564697e-01 4.82239053e-02 -9.63475585e-01
-3.91791999e-01 4.50589567e-01 -2.89137304e-01 -7.98271298e-01
9.02635381e-02 -5.37011623e-01 -5.34659863e-01 2.63280571e-01
2.55790770e-01 3.73486102e-01 6.75895572e-01 -5.77616870e-01
-9.21093002e-02 -1.81794260e-02 -4.63243663e-01 -2.21712023e-01
1.13900208e+00 -2.03189403e-01 -3.52524102e-01 3.69480312e-01
8.68017554e-01 1.28656819e-01 -9.15942267e-02 -5.89475036e-01
8.38613391e-01 -1.66051462e-01 -1.60059482e-01 -1.02339160e+00
-4.13343042e-01 7.15339065e-01 2.74061143e-01 5.82776904e-01
9.88619447e-01 -1.77009389e-01 7.92062581e-01 8.66653562e-01
7.59128481e-02 -1.61470091e+00 1.10528685e-01 7.25274444e-01
1.15100980e+00 -1.29957998e+00 -6.32695407e-02 -8.38666916e-01
-7.25518405e-01 1.39839375e+00 7.17902303e-01 6.39303982e-01
4.27033812e-01 2.93647289e-01 -4.53534611e-02 1.04660347e-01
-7.86814570e-01 -3.00143480e-01 2.70260721e-01 2.91806370e-01
5.52979767e-01 -2.74630010e-01 -7.23649621e-01 7.51657367e-01
-2.99830675e-01 8.11304059e-03 8.29541981e-01 1.07883930e+00
-1.16740167e-01 -1.47375584e+00 -6.20190263e-01 7.25420833e-01
-7.16242433e-01 -3.05824995e-01 -1.02993894e+00 9.55266833e-01
9.87997428e-02 1.15255272e+00 -2.45381638e-01 -2.60398239e-01
5.97335458e-01 5.13400257e-01 1.13302827e-01 -1.02294695e+00
-1.29656219e+00 6.08400330e-02 5.78956962e-01 9.68379378e-02
-3.21985781e-01 -8.44823956e-01 -1.00742054e+00 -8.29675198e-02
-6.42953992e-01 2.46968701e-01 1.01435339e+00 1.18439329e+00
4.41427603e-02 5.13853788e-01 7.30741680e-01 -2.77536243e-01
-4.77748215e-01 -1.31576884e+00 -1.55667335e-01 5.20899236e-01
-2.48172611e-01 -2.15076543e-02 -2.36146256e-01 4.61022079e-01] | [10.678742408752441, 8.597139358520508] |
bf354900-fc49-4926-aa32-24658a87b49a | s2f2-self-supervised-high-fidelity-face | 2203.07732 | null | https://arxiv.org/abs/2203.07732v2 | https://arxiv.org/pdf/2203.07732v2.pdf | S2F2: Self-Supervised High Fidelity Face Reconstruction from Monocular Image | We present a novel face reconstruction method capable of reconstructing detailed face geometry, spatially varying face reflectance from a single monocular image. We build our work upon the recent advances of DNN-based auto-encoders with differentiable ray tracing image formation, trained in self-supervised manner. While providing the advantage of learning-based approaches and real-time reconstruction, the latter methods lacked fidelity. In this work, we achieve, for the first time, high fidelity face reconstruction using self-supervised learning only. Our novel coarse-to-fine deep architecture allows us to solve the challenging problem of decoupling face reflectance from geometry using a single image, at high computational speed. Compared to state-of-the-art methods, our method achieves more visually appealing reconstruction. | ['Louis Chevallier', 'Philippe-Henri Gosselin', 'Cedric Thebault', 'Junghyun Ahn', 'Abdallah Dib'] | 2022-03-15 | null | null | null | null | ['3d-face-reconstruction', 'face-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 2.83186167e-01 1.81229785e-01 4.10898864e-01 -4.88765627e-01
-7.17767715e-01 -1.83479235e-01 6.29667878e-01 -6.98726237e-01
3.19234990e-02 7.14266658e-01 1.80514514e-01 -6.29966904e-04
2.84267068e-02 -1.08753109e+00 -1.07744479e+00 -4.29744422e-01
3.60512793e-01 6.22762978e-01 -2.85731584e-01 -2.17654482e-01
-8.65331292e-02 1.04058933e+00 -1.90612376e+00 2.92098492e-01
2.89522439e-01 9.59922194e-01 7.20129758e-02 6.56155944e-01
2.92130932e-02 9.19063449e-01 1.08832270e-01 -4.09761071e-01
3.85900557e-01 -4.96811360e-01 -6.76204026e-01 2.68847436e-01
1.07633257e+00 -1.00495863e+00 -6.39197469e-01 6.36301041e-01
5.22016287e-01 -2.17879981e-01 6.67302668e-01 -6.65832937e-01
-1.00645840e+00 -8.90219435e-02 -5.12123525e-01 -2.66473293e-01
7.60827780e-01 6.98342994e-02 7.45424151e-01 -1.19933975e+00
7.57326603e-01 1.40199423e+00 8.39438975e-01 1.03898537e+00
-1.60176897e+00 -5.80984414e-01 -2.36781448e-01 -1.80043697e-01
-1.33267832e+00 -1.04129171e+00 9.45900977e-01 -4.01036561e-01
8.66679430e-01 -6.49033114e-02 6.05224371e-01 1.04225063e+00
-3.03123862e-01 4.25530583e-01 1.30625367e+00 -4.03690994e-01
-1.98986959e-02 -2.50259697e-01 -6.87373579e-01 1.37835538e+00
-1.31826028e-01 5.57040155e-01 -6.71657860e-01 7.91821256e-03
1.48705876e+00 2.12216392e-01 -4.85830098e-01 -3.02850604e-01
-7.26348460e-01 5.50225377e-01 4.25616264e-01 1.69520780e-01
-3.81020606e-01 5.16898274e-01 -2.38044992e-01 2.91842222e-01
7.26205111e-01 -5.12984246e-02 -4.54347789e-01 5.57826273e-02
-1.06293511e+00 -2.30068192e-02 6.70623660e-01 6.92070425e-01
1.02238345e+00 5.32066643e-01 2.63869286e-01 5.59039056e-01
5.19483268e-01 6.66487098e-01 3.29990126e-02 -1.22088623e+00
-1.33163556e-01 2.99120814e-01 5.51089495e-02 -6.33557975e-01
-3.72213423e-01 -3.11042607e-01 -7.05952108e-01 7.99403489e-01
3.40835869e-01 -6.50457442e-02 -8.73886883e-01 1.64364159e+00
5.60938954e-01 3.91682833e-01 -4.83650807e-03 8.24706614e-01
9.05214190e-01 4.71906990e-01 -4.59506005e-01 -2.10704848e-01
1.07209384e+00 -7.62872756e-01 -5.83765328e-01 2.41921484e-01
-1.59289181e-01 -8.59881938e-01 1.02823317e+00 5.53885221e-01
-1.53453600e+00 -4.96666014e-01 -7.50443399e-01 -3.82668585e-01
1.16870642e-01 1.71683222e-01 7.37570167e-01 7.77883232e-01
-1.63841021e+00 9.10456955e-01 -8.17684650e-01 -2.22009331e-01
8.20008814e-01 3.80038381e-01 -6.98034227e-01 -2.25752845e-01
-4.77102906e-01 5.43458045e-01 -4.80570644e-01 -1.12399794e-02
-1.36967421e+00 -1.20115387e+00 -9.63024259e-01 -1.30089968e-01
1.26994416e-01 -1.03786540e+00 1.30999124e+00 -1.06421375e+00
-2.25442195e+00 1.03759217e+00 -4.12791669e-01 -1.44569069e-01
5.43201745e-01 -3.14191163e-01 -1.54024631e-01 5.14062226e-01
-3.34913969e-01 6.14553988e-01 1.14786267e+00 -1.74277997e+00
-1.67063233e-02 -5.69660425e-01 1.43047348e-01 1.81570761e-02
-3.68951224e-02 9.81778558e-03 -2.40822971e-01 -3.54404211e-01
6.17148466e-02 -4.40054864e-01 2.23601580e-01 8.16733897e-01
-1.28616737e-02 1.56524584e-01 8.56533766e-01 -7.56417334e-01
3.19184721e-01 -1.86623096e+00 1.47121504e-01 -2.22855359e-01
3.24236423e-01 2.03266338e-01 -1.99103072e-01 4.75981832e-01
-1.75243348e-01 -1.98997825e-01 -4.20209467e-01 -1.07641459e+00
-1.56151548e-01 9.00410265e-02 -1.92092627e-01 7.99880862e-01
2.96514571e-01 9.95091617e-01 -7.49937057e-01 -2.75295436e-01
4.66543227e-01 1.34963143e+00 -7.66069472e-01 5.43488562e-01
-3.60207587e-01 8.93227220e-01 1.16144530e-01 7.65695274e-01
8.35203409e-01 -1.95716828e-01 1.71125934e-01 -5.29626012e-01
-1.32395014e-01 7.99628198e-02 -8.70429397e-01 2.09325433e+00
-9.91720200e-01 5.53191960e-01 5.08359432e-01 -7.26215005e-01
8.00069690e-01 5.23842335e-01 7.10121453e-01 -8.26032341e-01
-1.05693258e-01 -2.68881321e-02 -6.56877518e-01 -2.83331633e-01
7.93227255e-02 -7.14410484e-01 8.18194687e-01 6.55374527e-01
2.84953117e-01 -3.20732653e-01 -3.98880333e-01 -2.16604248e-02
9.07808959e-01 8.19083571e-01 -6.58664107e-02 -1.72853112e-01
4.41078991e-01 -5.63379407e-01 1.59927815e-01 2.27853227e-02
3.50898147e-01 1.17806911e+00 3.64474021e-02 -6.54594839e-01
-1.18120503e+00 -1.31723917e+00 -2.99927890e-01 6.30245268e-01
-3.23905885e-01 -3.33479732e-01 -8.84589076e-01 -3.37495446e-01
-2.20186841e-02 2.23213896e-01 -8.20416152e-01 2.98680633e-01
-7.23865271e-01 -3.77974778e-01 5.29678762e-01 3.21779042e-01
5.70578754e-01 -9.03207660e-01 -3.56159091e-01 5.76451235e-02
2.44343445e-01 -1.34319448e+00 -2.90848374e-01 -4.71857399e-01
-8.42664838e-01 -1.15502763e+00 -8.16623807e-01 -7.17272460e-01
9.03147817e-01 2.72011191e-01 1.38921261e+00 2.78684527e-01
-8.42986882e-01 8.38764846e-01 2.88282871e-01 1.06756233e-01
-3.18629295e-01 -6.25874341e-01 5.51630929e-02 3.22595030e-01
-2.72216886e-01 -1.09709370e+00 -9.08860326e-01 4.99485172e-02
-8.10198426e-01 2.13023797e-01 2.71674663e-01 7.95429885e-01
7.27283597e-01 -1.18100256e-01 3.46029878e-01 -9.92741942e-01
3.74789424e-02 -1.56719059e-01 -8.00214410e-01 -2.57419311e-02
-4.63955879e-01 1.00453973e-01 9.35642779e-01 5.55110984e-02
-1.50012708e+00 3.07487279e-01 -5.46807468e-01 -5.62273502e-01
-3.10152352e-01 -1.11749224e-01 -6.88788667e-02 -5.81752717e-01
6.65683389e-01 3.72084469e-01 3.54617745e-01 -7.63471842e-01
4.96733040e-01 2.10933343e-01 6.49636805e-01 -5.91208696e-01
9.26677346e-01 1.20297813e+00 3.76030624e-01 -9.87327099e-01
-7.37432063e-01 5.43452427e-02 -7.57121503e-01 -2.69222170e-01
7.97609687e-01 -1.14930665e+00 -1.09695232e+00 5.94737291e-01
-1.26289749e+00 -6.70960784e-01 -3.97655189e-01 -1.98669173e-02
-8.77045691e-01 4.79105771e-01 -8.03101897e-01 -8.67709994e-01
-2.96245068e-01 -7.79872179e-01 1.55762005e+00 8.55454504e-02
3.95601988e-01 -9.83399451e-01 1.99796900e-01 4.39259708e-01
4.73403096e-01 4.48969752e-01 3.44138324e-01 5.99714518e-01
-1.01867831e+00 3.11907619e-01 -4.92315918e-01 2.48418882e-01
3.93313974e-01 3.15600000e-02 -1.41760218e+00 -4.35931563e-01
6.78629279e-02 -4.87186790e-01 7.31466591e-01 1.94664463e-01
1.27518451e+00 -3.03646594e-01 3.10316794e-02 1.23796999e+00
1.77488768e+00 -2.79162496e-01 7.50630796e-01 -3.56774956e-01
1.01018047e+00 5.45173705e-01 -2.07767963e-01 4.65611786e-01
5.73571742e-01 7.65705884e-01 5.53071439e-01 -3.63454789e-01
-8.05353761e-01 -4.79587495e-01 2.48049974e-01 4.38764155e-01
-3.97402167e-01 3.07163060e-01 -5.81526995e-01 3.47090662e-01
-1.50171721e+00 -1.02431893e+00 1.12371162e-01 2.02159309e+00
8.54009926e-01 -4.98694986e-01 -1.19775748e-02 5.26059680e-02
1.53226256e-01 2.69077849e-02 -3.80879968e-01 -1.28790960e-01
-1.00713514e-01 6.82255745e-01 1.85683653e-01 9.47986722e-01
-7.23284960e-01 9.79858577e-01 6.95987701e+00 2.99236119e-01
-1.18427956e+00 2.97240376e-01 4.53499138e-01 -2.72532135e-01
-7.66240954e-01 -1.92488506e-01 -6.27451062e-01 8.78063515e-02
8.01987290e-01 4.28111732e-01 1.05701792e+00 4.85041469e-01
4.67256829e-02 1.75437748e-01 -1.28374684e+00 1.40345991e+00
5.00993609e-01 -1.62264299e+00 9.58031341e-02 9.57219526e-02
9.13774967e-01 1.69396715e-03 1.65548652e-01 -3.52604419e-01
3.38976562e-01 -1.53096712e+00 5.89793444e-01 8.52271914e-01
1.38245559e+00 -7.40383267e-01 -8.06826800e-02 4.46553491e-02
-1.09663546e+00 2.25142196e-01 -4.06266928e-01 -9.30288658e-02
2.76706666e-01 8.22132170e-01 -4.30279732e-01 6.70624495e-01
6.56349838e-01 9.15710270e-01 -7.86052942e-02 4.40313518e-01
-1.61162645e-01 3.37154925e-01 -3.66584986e-01 5.92760682e-01
-2.13867158e-01 -3.36232513e-01 2.16541663e-01 8.72825742e-01
2.73773998e-01 4.53301072e-01 -1.64069504e-01 1.26574814e+00
-4.12054628e-01 -9.30696055e-02 -1.00683212e+00 2.74154007e-01
1.35595798e-02 1.16512561e+00 -2.83548474e-01 -4.22555134e-02
-5.69594204e-01 1.27216482e+00 7.68111765e-01 3.16896588e-01
-4.64225262e-01 2.20409244e-01 8.03883135e-01 4.40999329e-01
4.46700007e-01 -5.63324690e-01 -2.24930748e-01 -1.18665981e+00
1.13714166e-01 -5.56725740e-01 -2.30802611e-01 -8.59684706e-01
-1.27293432e+00 7.88347006e-01 -4.23983425e-01 -7.30926514e-01
-2.52174586e-01 -7.89398313e-01 -3.48210573e-01 8.48989129e-01
-2.08828688e+00 -1.54997778e+00 -4.42202836e-01 9.53876257e-01
3.92998695e-01 -3.64701748e-01 1.13832891e+00 5.06108761e-01
-2.00935885e-01 6.43304825e-01 3.15437242e-02 -2.02221982e-02
4.28604633e-01 -1.13650334e+00 5.53134143e-01 5.91801226e-01
3.75127852e-01 5.53388715e-01 3.63910913e-01 -4.63238209e-01
-1.97985125e+00 -8.77734482e-01 5.16693592e-01 -4.40423459e-01
1.44722015e-01 -3.80420089e-01 -6.43384337e-01 7.26563215e-01
1.80559352e-01 4.40999180e-01 5.51756144e-01 -7.02393502e-02
-6.73058450e-01 -3.15517366e-01 -1.55185235e+00 5.57318032e-01
1.54979968e+00 -9.69569623e-01 -1.79376930e-01 4.08783764e-01
5.09309232e-01 -4.80343401e-01 -8.65250051e-01 1.87534779e-01
6.42471611e-01 -1.34294248e+00 1.35253751e+00 -2.30715990e-01
6.97343230e-01 -1.94606498e-01 -1.76316917e-01 -1.12706780e+00
-2.09697977e-01 -9.92571771e-01 -3.52302313e-01 1.01882899e+00
-2.92547066e-02 -5.85842073e-01 1.04740965e+00 3.49710405e-01
-1.36084810e-01 -7.64689684e-01 -9.81872261e-01 -4.77727920e-01
-3.50930244e-02 -2.24513263e-01 8.35661829e-01 6.51962578e-01
-5.76938450e-01 1.27309322e-01 -5.67126095e-01 1.51368037e-01
1.14160657e+00 2.31413260e-01 6.81327939e-01 -1.28711748e+00
-4.55491394e-01 -2.01013610e-01 -2.41869643e-01 -1.13444757e+00
5.23387849e-01 -7.07254410e-01 -1.58248529e-01 -1.54533470e+00
9.19896886e-02 -2.22612083e-01 3.48302215e-01 4.56650317e-01
3.33062947e-01 9.69901741e-01 3.91333513e-02 -2.87474617e-02
-1.22911029e-01 7.86437273e-01 1.35245264e+00 1.89882413e-01
2.71524698e-01 -3.42274696e-01 -6.51224077e-01 7.47307599e-01
5.20232677e-01 -7.92220756e-02 -4.13456142e-01 -8.45026791e-01
1.76776201e-01 1.58523336e-01 6.34514809e-01 -9.17021692e-01
1.33695945e-01 4.07559089e-02 6.56100690e-01 -3.24532866e-01
8.57674718e-01 -9.57004428e-01 3.52825910e-01 1.86975583e-01
1.47333920e-01 -1.69787407e-01 2.57068694e-01 4.46058482e-01
1.14749819e-01 1.95415065e-01 1.07623732e+00 -4.07831371e-01
-2.77888000e-01 6.74212635e-01 1.41584098e-01 -2.10120484e-01
6.37822866e-01 -1.28864631e-01 -8.25086534e-02 -4.55179691e-01
-6.25682592e-01 -5.16307771e-01 8.53159249e-01 5.98185770e-02
1.09184420e+00 -1.43353772e+00 -1.04601085e+00 6.29928231e-01
-2.79105157e-01 1.66161969e-01 3.57684314e-01 2.22621620e-01
-9.37326729e-01 2.77322561e-01 -3.94383222e-01 -6.28098726e-01
-1.12786567e+00 3.42505813e-01 6.85485065e-01 8.87555555e-02
-1.08461988e+00 8.63810182e-01 3.90026242e-01 -7.44189918e-01
-6.69846386e-02 3.57417017e-01 1.79842338e-01 -4.62610513e-01
6.58094108e-01 2.10055441e-01 2.42565423e-01 -7.12930083e-01
-1.27720699e-01 1.17003584e+00 3.10542703e-01 -2.78658628e-01
1.63409185e+00 -1.79101020e-01 -2.65254617e-01 1.17149182e-01
1.42280769e+00 1.00886531e-01 -1.93800890e+00 -2.57171929e-01
-7.68882453e-01 -8.22093487e-01 3.67678046e-01 -7.54052103e-01
-1.57619452e+00 9.73080516e-01 4.13869590e-01 -4.37777638e-01
1.12698579e+00 -1.35633005e-02 8.21503639e-01 9.69840363e-02
6.11757874e-01 -5.11658788e-01 3.81200373e-01 3.34947497e-01
1.03909767e+00 -1.35666227e+00 1.26416013e-01 -5.55778146e-01
-2.19521374e-02 1.27332115e+00 5.55358768e-01 -3.37234050e-01
8.84687364e-01 5.36358118e-01 -6.74966723e-02 -4.55881894e-01
-6.23804867e-01 -2.29781017e-01 2.92783618e-01 8.81330013e-01
6.59137487e-01 -8.02621692e-02 2.86631376e-01 -5.49802110e-02
-2.56319165e-01 1.91813469e-01 3.33205968e-01 6.83544993e-01
1.13199055e-02 -1.10647500e+00 -3.36872131e-01 1.02627590e-01
-4.25606757e-01 -9.41623449e-02 -1.37513563e-01 4.76069897e-01
-1.28872022e-01 6.29617572e-01 2.38766566e-01 3.60650942e-02
2.41009697e-01 -1.13278329e-01 1.31081748e+00 -5.98318279e-01
-2.40472600e-01 -1.25940681e-01 -7.02553689e-02 -8.78129363e-01
-3.96181464e-01 -6.60791516e-01 -1.16293132e+00 -5.01264513e-01
2.05672562e-01 -4.41660285e-01 7.32601702e-01 8.58373642e-01
4.57530886e-01 3.90274495e-01 8.97005141e-01 -1.30285823e+00
-2.16738269e-01 -5.55606425e-01 -5.49391925e-01 4.68002111e-01
7.60252774e-01 -6.19448781e-01 -1.08405821e-01 1.48465678e-01] | [12.872145652770996, -0.2965932786464691] |
c670d4fd-7397-46a1-b7b4-69750abd6803 | darker-than-black-box-face-reconstruction | 2106.14290 | null | https://arxiv.org/abs/2106.14290v2 | https://arxiv.org/pdf/2106.14290v2.pdf | Darker than Black-Box: Face Reconstruction from Similarity Queries | Several methods for inversion of face recognition models were recently presented, attempting to reconstruct a face from deep templates. Although some of these approaches work in a black-box setup using only face embeddings, usually, on the end-user side, only similarity scores are provided. Therefore, these algorithms are inapplicable in such scenarios. We propose a novel approach that allows reconstructing the face querying only similarity scores of the black-box model. While our algorithm operates in a more general setup, experiments show that it is query efficient and outperforms the existing methods. | ['Aleksandr Petiushko', 'Igor Udovichenko', 'Klim Kireev', 'Anton Razzhigaev'] | 2021-06-27 | null | null | null | null | ['face-reconstruction'] | ['computer-vision'] | [ 1.28825858e-01 2.00570002e-01 9.29627195e-02 -5.72455287e-01
-4.96401787e-01 -4.63299602e-01 5.69332242e-01 -5.48556328e-01
-2.07315490e-01 3.65450829e-01 -2.28562340e-01 -2.53812134e-01
-6.01040423e-02 -7.38325298e-01 -7.90673018e-01 -6.52004838e-01
3.74492288e-01 6.30556524e-01 -4.41944189e-02 2.43140236e-02
1.22397885e-01 7.61573493e-01 -1.95785522e+00 6.78415671e-02
2.37750903e-01 1.01975286e+00 1.54338367e-02 5.58863044e-01
-1.63235649e-01 3.24049681e-01 -1.97509632e-01 -1.05540252e+00
6.12346828e-01 -3.51696223e-01 -6.78287208e-01 1.33398652e-01
1.03253889e+00 -5.70836306e-01 -5.24956465e-01 1.15348828e+00
4.38926965e-01 -2.21437022e-01 4.79717433e-01 -1.27856648e+00
-4.82960045e-01 1.34441331e-01 -2.71330506e-01 -2.76625663e-01
5.90918660e-01 -1.97023883e-01 8.18194389e-01 -1.45203376e+00
4.72368479e-01 1.37873423e+00 7.14105606e-01 8.41568112e-01
-1.45693946e+00 -6.95566893e-01 -5.56478091e-02 2.47708812e-01
-1.43308914e+00 -9.23299193e-01 8.95724416e-01 -1.34958252e-01
7.52381682e-01 2.84079641e-01 4.94154453e-01 1.09153223e+00
-3.34573030e-01 7.03203619e-01 1.08375001e+00 -5.08629560e-01
1.60475969e-01 2.21296236e-01 -3.94962132e-02 9.42135692e-01
3.36410612e-01 2.17426032e-01 -7.50991940e-01 -3.88056338e-01
7.82089770e-01 1.58394843e-01 -1.36413023e-01 -8.80194783e-01
-8.84406507e-01 7.29893923e-01 1.00940332e-01 1.31384581e-01
-3.08457166e-01 1.67567208e-01 4.75931476e-04 4.40378755e-01
3.39944959e-01 -1.71528846e-01 -1.88581884e-01 7.80601874e-02
-1.29132569e+00 2.78482199e-01 1.11224186e+00 6.95431948e-01
9.77132201e-01 -1.51840925e-01 2.07183868e-01 5.55742681e-01
4.72181112e-01 3.92649621e-01 1.87897310e-01 -9.57922995e-01
1.82902649e-01 2.25347012e-01 9.30758193e-02 -1.18456542e+00
-1.51012957e-01 4.23916504e-02 -6.08786106e-01 3.58355105e-01
5.49627304e-01 3.71095017e-02 -8.07467461e-01 1.74956012e+00
4.83885854e-01 6.98888600e-01 -1.46993175e-01 5.89270890e-01
7.32903361e-01 3.88993174e-01 -3.10977399e-01 -9.56027880e-02
1.24224079e+00 -9.10730541e-01 -7.88617134e-01 -1.10639557e-02
1.90993264e-01 -8.97874415e-01 7.94828594e-01 4.51410502e-01
-1.29893398e+00 -7.02562571e-01 -1.06714642e+00 -2.57747900e-02
-2.75184453e-01 1.67646915e-01 4.66567844e-01 1.11600530e+00
-1.49043906e+00 7.32221782e-01 -7.50670016e-01 -4.79336053e-01
4.36806679e-01 8.41686428e-01 -7.88439989e-01 -9.38076228e-02
-6.51386678e-01 1.10908091e+00 2.55806800e-02 4.41044062e-01
-9.39820290e-01 -3.68633032e-01 -8.42796326e-01 1.63410366e-01
2.61276901e-01 -5.64127922e-01 1.09823418e+00 -9.17684257e-01
-1.87784660e+00 9.53620970e-01 -6.07520998e-01 -2.54137665e-01
5.50398707e-01 -1.31486669e-01 -1.61794737e-01 1.81807205e-01
-5.63325226e-01 7.12544560e-01 1.13966298e+00 -1.23613417e+00
-1.10080771e-01 -6.97881699e-01 3.19460630e-01 -2.28900746e-01
-5.69010496e-01 1.83015183e-01 -4.95385170e-01 -4.37520266e-01
3.09524685e-01 -8.88771236e-01 1.42412437e-02 6.35432363e-01
3.85395661e-02 -1.99082226e-01 1.01135910e+00 -6.35480464e-01
8.60419750e-01 -2.06152177e+00 2.09265381e-01 3.74054819e-01
7.69264847e-02 5.46125233e-01 -3.53979439e-01 3.59815806e-01
-2.60273814e-01 -1.17064692e-01 -3.53831321e-01 -7.39086568e-01
2.15556249e-01 3.73600245e-01 -3.71294677e-01 6.50193930e-01
1.84963912e-01 8.25894594e-01 -5.13291001e-01 -4.13418263e-01
2.22623527e-01 9.49720800e-01 -1.04381323e+00 4.26273167e-01
1.77207142e-01 3.76001477e-01 -1.39021939e-02 6.23475850e-01
1.32597280e+00 1.96927562e-01 5.00476003e-01 -1.53004974e-01
1.87690720e-01 1.87109545e-01 -1.51255167e+00 1.71527600e+00
-6.07720971e-01 4.23649430e-01 4.05482531e-01 -1.15885293e+00
9.05277729e-01 5.58033645e-01 2.48437673e-01 -2.17681080e-01
-9.63262096e-02 1.33989155e-01 -1.31446838e-01 -4.85706776e-01
1.19761445e-01 -3.51739466e-01 5.25684357e-01 6.43481851e-01
3.76111567e-01 -8.28029960e-03 -6.35025129e-02 -2.40639970e-01
1.05729032e+00 5.18026352e-01 1.77049816e-01 1.27520310e-02
7.79143870e-01 -8.00148964e-01 4.11613137e-01 5.52092910e-01
-2.10890844e-01 1.00028670e+00 4.77153301e-01 -5.80829561e-01
-7.84135222e-01 -1.16171157e+00 -1.81139544e-01 9.52627718e-01
-1.34262443e-01 -7.67977297e-01 -1.15381002e+00 -9.51795340e-01
9.95766595e-02 2.20589519e-01 -7.42388904e-01 1.36061653e-01
-5.98469496e-01 -3.88878286e-01 5.53766847e-01 3.61355007e-01
2.88679332e-01 -9.20956075e-01 -4.34823841e-01 -7.47530758e-02
-1.91537719e-02 -1.11578715e+00 -2.85123795e-01 -2.75633723e-01
-1.01574397e+00 -1.07295322e+00 -6.91554546e-01 -9.47030485e-01
1.12687266e+00 2.23966047e-01 9.81927991e-01 4.47039455e-01
-2.50183403e-01 3.30495387e-01 5.93083389e-02 -5.45415021e-02
-1.29341021e-01 -3.60930592e-01 2.12161854e-01 5.39451659e-01
5.16264260e-01 -6.95596695e-01 -4.39338535e-01 3.59987408e-01
-8.84235084e-01 -3.06324005e-01 4.56598043e-01 1.00618052e+00
2.41825268e-01 -3.69409144e-01 4.52730656e-01 -9.39087808e-01
2.40445569e-01 -1.34974450e-01 -8.00482810e-01 4.29666936e-01
-4.37405348e-01 1.77261993e-01 6.94607377e-01 -3.85928929e-01
-1.11367595e+00 4.97740895e-01 -5.62390864e-01 -6.54376864e-01
-2.41168544e-01 2.66416557e-02 -6.56630933e-01 -5.60884595e-01
3.84211123e-01 2.05712140e-01 3.15153420e-01 -8.04741800e-01
1.09868087e-01 6.83757067e-01 4.31797117e-01 -5.48722208e-01
1.21751463e+00 6.00486577e-01 1.01927884e-01 -6.83914602e-01
-3.73260021e-01 -2.06404135e-01 -9.31103945e-01 -9.14901495e-02
3.09748620e-01 -5.45861244e-01 -8.76579881e-01 3.92350107e-01
-1.40152442e+00 2.63128459e-01 -4.38014157e-02 2.62627572e-01
-6.26852810e-01 6.89941406e-01 -1.04070768e-01 -8.94798279e-01
7.53546925e-03 -1.23962700e+00 1.14533794e+00 1.64742135e-02
-5.84943406e-02 -9.67890203e-01 1.12722374e-01 2.53896683e-01
3.86773229e-01 -1.73865095e-01 7.34269917e-01 -5.84288061e-01
-4.24912781e-01 -3.93903255e-01 -1.64139986e-01 4.50507909e-01
6.51328787e-02 -1.72255877e-02 -1.58313692e+00 -4.44350421e-01
1.84735388e-01 -2.58852154e-01 5.89545310e-01 -1.25893787e-01
1.51468527e+00 -3.55247021e-01 -2.22518712e-01 7.67886817e-01
1.19698393e+00 -2.12719455e-01 9.67362940e-01 -1.79627150e-01
3.21646154e-01 7.91870713e-01 2.16833353e-01 2.83014625e-01
2.65299439e-01 8.85399163e-01 5.72720289e-01 4.39964607e-02
-4.81939018e-02 -4.39113468e-01 6.82805240e-01 5.94544411e-01
-1.19614094e-01 1.44214988e-01 -5.98626733e-01 3.04509938e-01
-1.52289629e+00 -1.10284972e+00 3.76047373e-01 2.43997860e+00
6.10239804e-01 -2.09138274e-01 2.66122520e-02 2.31698945e-01
6.27026439e-01 2.09879071e-01 -3.52532178e-01 -3.46568018e-01
3.97372693e-02 7.45366395e-01 -8.04086626e-02 6.28811061e-01
-8.79480660e-01 7.59331167e-01 7.84725094e+00 3.87126386e-01
-9.26376224e-01 2.70093530e-01 2.64323533e-01 -1.00950800e-01
-3.56886119e-01 3.17663223e-01 -8.71183693e-01 2.36276776e-01
9.06725347e-01 4.08060029e-02 5.98702073e-01 6.45589411e-01
-2.87343174e-01 6.84879795e-02 -1.55745447e+00 1.42894828e+00
5.31315148e-01 -9.93592739e-01 1.69344470e-01 8.11061040e-02
2.77856410e-01 -5.44203579e-01 3.14635992e-01 2.30738983e-01
-3.84764761e-01 -1.15444243e+00 5.07664204e-01 5.61242342e-01
7.65078902e-01 -6.35732055e-01 4.98427153e-01 3.81437451e-01
-8.96278501e-01 1.34024486e-01 -5.12630224e-01 3.66476327e-02
-1.33858591e-01 3.43954802e-01 -7.81061828e-01 4.11176890e-01
4.70615387e-01 3.34715188e-01 -6.48662508e-01 8.95268261e-01
-1.43262178e-01 3.83872837e-01 -3.57435703e-01 3.96853417e-01
-2.24633247e-01 -2.43065789e-01 2.89609551e-01 1.05127215e+00
5.57025790e-01 -3.73170637e-02 -1.83410376e-01 7.97636807e-01
-2.40454555e-01 -4.28864658e-02 -8.13959599e-01 1.77797023e-02
1.92059278e-01 1.44339120e+00 -4.28569496e-01 -1.81711614e-01
-6.62155032e-01 1.13964009e+00 4.22938704e-01 1.75737262e-01
-8.99959445e-01 -3.47872078e-01 8.27227294e-01 2.88803011e-01
5.91711223e-01 -3.21791112e-01 -1.09401256e-01 -1.23617876e+00
2.27289617e-01 -8.97174597e-01 1.64785817e-01 -5.55855155e-01
-1.03535104e+00 6.16525292e-01 -1.91633347e-02 -1.08504295e+00
-3.99567336e-01 -1.05278695e+00 -4.44477588e-01 7.89405406e-01
-1.55919075e+00 -1.03715265e+00 -2.37481534e-01 6.74916208e-01
3.68227243e-01 -2.29966149e-01 1.24353611e+00 3.56246412e-01
-4.44478542e-01 9.60854709e-01 -4.10331879e-03 -8.94398615e-02
7.14697480e-01 -9.05826151e-01 4.17767137e-01 8.53812039e-01
6.01610303e-01 1.02308393e+00 6.07651949e-01 -2.63486564e-01
-1.79721606e+00 -7.07051039e-01 1.07902861e+00 -3.16548318e-01
2.40213871e-01 -6.82710469e-01 -9.22422945e-01 6.90590322e-01
3.63354176e-01 4.58165735e-01 8.20220530e-01 3.08838058e-02
-5.83038449e-01 -3.10455382e-01 -1.54380131e+00 5.26484609e-01
1.20396113e+00 -8.62385035e-01 -5.11527359e-01 2.50857025e-01
-1.08770199e-01 -2.89679378e-01 -5.76674342e-01 4.68253076e-01
8.92647743e-01 -1.09446812e+00 1.04986310e+00 -5.29595435e-01
5.66907646e-03 -4.23146069e-01 -2.16064245e-01 -9.60371256e-01
-3.32242413e-03 -7.75424898e-01 -5.39868116e-01 9.84142423e-01
1.11796930e-01 -8.10073793e-01 9.15988326e-01 6.05071127e-01
4.35766369e-01 -5.44742882e-01 -1.28425503e+00 -7.12072551e-01
-1.93777919e-01 -2.54732877e-01 8.18526268e-01 4.89680737e-01
-3.16853642e-01 -2.89290398e-02 -4.54275340e-01 1.97976097e-01
6.83167160e-01 1.08950846e-01 8.88932645e-01 -1.41272652e+00
-2.90419102e-01 -2.31251106e-01 -5.68951488e-01 -1.00332570e+00
6.43165767e-01 -1.07768762e+00 -4.25283201e-02 -1.08013892e+00
2.11845830e-01 -2.10364684e-01 -2.99538344e-01 5.68366587e-01
1.63863286e-01 6.93004489e-01 1.70328394e-01 -1.29252613e-01
-2.27691293e-01 4.48439002e-01 7.64345467e-01 -1.14096224e-01
4.48990524e-01 1.01502508e-01 -5.20902693e-01 9.05956507e-01
5.96613526e-01 -5.24139941e-01 -3.10806066e-01 -4.95337278e-01
4.22031283e-02 -1.67437911e-01 5.90341747e-01 -9.89437163e-01
4.19171989e-01 3.49100083e-02 4.53689039e-01 -5.11544466e-01
6.98574781e-01 -1.06737030e+00 5.22656441e-01 2.94789076e-01
1.37549564e-02 1.97113171e-01 8.01457465e-02 3.54211271e-01
-2.18872294e-01 -6.29664063e-01 8.85416627e-01 1.10841429e-04
-2.09504291e-01 3.98006588e-01 5.22247218e-02 -5.00188768e-01
9.30030823e-01 -4.73345339e-01 2.42801294e-01 -4.61328179e-01
-7.66517520e-01 -4.89307582e-01 5.60922623e-01 2.58671999e-01
9.34040010e-01 -1.38616884e+00 -4.85577673e-01 8.83209884e-01
-1.13518931e-01 -3.49965096e-01 2.27529034e-02 8.02652299e-01
-5.21525860e-01 3.94351035e-01 -3.00940990e-01 -3.57661664e-01
-1.45371413e+00 6.45776808e-01 3.48863989e-01 -2.79781036e-03
-3.96245152e-01 7.90839970e-01 3.92436743e-01 -8.02050292e-01
4.46601838e-01 2.10445702e-01 -6.91364482e-02 1.19727075e-01
7.95208037e-01 2.15600923e-01 3.99796277e-01 -8.67529988e-01
-4.58569318e-01 8.07607114e-01 -7.90740363e-03 -3.00408393e-01
1.36533129e+00 9.48405191e-02 -4.71984565e-01 5.54764085e-02
1.30434644e+00 1.24785028e-01 -1.12619400e+00 -2.53679037e-01
-1.86406188e-02 -9.59861040e-01 -1.04932807e-01 -3.14581454e-01
-1.30406940e+00 1.18959630e+00 7.27066696e-01 -9.30458605e-02
1.27661395e+00 -3.58894348e-01 5.67009568e-01 6.81845486e-01
4.79740739e-01 -8.33174467e-01 9.61910281e-03 4.32625890e-01
9.75832462e-01 -1.05326366e+00 -7.05643594e-02 -3.92196327e-01
-7.66036958e-02 1.43273652e+00 5.76071799e-01 -1.32263362e-01
9.28160787e-01 4.08670664e-01 -9.79785398e-02 1.91371694e-01
-6.94200635e-01 -1.55893072e-01 6.65185153e-02 7.13987887e-01
5.10094523e-01 -3.11107725e-01 -1.39498398e-01 2.23246038e-01
-1.81061596e-01 9.62700024e-02 2.58368373e-01 9.54172432e-01
-5.53884543e-02 -1.68521512e+00 -6.43221438e-01 7.72663280e-02
-4.33184803e-01 6.63249418e-02 -3.77065420e-01 5.18440664e-01
1.14034332e-01 8.95292461e-01 8.24378878e-02 -3.42694700e-01
3.30823451e-01 4.93140519e-01 9.60102499e-01 -4.95946974e-01
-3.74154985e-01 -2.53876239e-01 -2.95582324e-01 -6.57068729e-01
-5.03377914e-01 -7.19637513e-01 -7.56280541e-01 -4.28161830e-01
-3.15078288e-01 3.48560773e-02 8.24293971e-01 7.93308735e-01
3.78292501e-01 -2.21707389e-01 8.84629011e-01 -1.10376322e+00
-7.96052635e-01 -8.82639408e-01 -4.88211244e-01 2.44472861e-01
3.81652236e-01 -8.23380053e-01 -2.64055967e-01 5.53929955e-02] | [13.125481605529785, 0.30603325366973877] |
84f1d3b1-29a7-404c-8381-c1fe943f9798 | fastventricle-cardiac-segmentation-with-enet | 1704.04296 | null | http://arxiv.org/abs/1704.04296v1 | http://arxiv.org/pdf/1704.04296v1.pdf | FastVentricle: Cardiac Segmentation with ENet | Cardiac Magnetic Resonance (CMR) imaging is commonly used to assess cardiac
structure and function. One disadvantage of CMR is that post-processing of
exams is tedious. Without automation, precise assessment of cardiac function
via CMR typically requires an annotator to spend tens of minutes per case
manually contouring ventricular structures. Automatic contouring can lower the
required time per patient by generating contour suggestions that can be lightly
modified by the annotator. Fully convolutional networks (FCNs), a variant of
convolutional neural networks, have been used to rapidly advance the
state-of-the-art in automated segmentation, which makes FCNs a natural choice
for ventricular segmentation. However, FCNs are limited by their computational
cost, which increases the monetary cost and degrades the user experience of
production systems. To combat this shortcoming, we have developed the
FastVentricle architecture, an FCN architecture for ventricular segmentation
based on the recently developed ENet architecture. FastVentricle is 4x faster
and runs with 6x less memory than the previous state-of-the-art ventricular
segmentation architecture while still maintaining excellent clinical accuracy. | ['Jesse Lieman-Sifry', 'Sean Sall', 'Matthieu Le', 'Felix Lau', 'Daniel Golden'] | 2017-04-13 | null | null | null | null | ['cardiac-segmentation'] | ['medical'] | [-4.46944647e-02 1.24236003e-01 1.54198959e-01 -4.66884464e-01
-4.81522381e-01 -7.27214932e-01 -2.35906944e-01 3.44981968e-01
-5.85684776e-01 5.81548154e-01 -2.91374147e-01 -7.55600572e-01
3.31902087e-01 -6.15657926e-01 -2.66756505e-01 -1.83443740e-01
-8.92661735e-02 8.01557183e-01 4.15773362e-01 2.07477465e-01
-9.01135989e-03 6.41745627e-01 -7.15521872e-01 2.16375619e-01
8.26836586e-01 8.61502945e-01 1.64683834e-01 8.84500384e-01
-2.27748916e-01 7.41508782e-01 -4.86861736e-01 -2.01334372e-01
2.17751533e-01 -4.10508096e-01 -9.01377201e-01 1.35137424e-01
1.86253175e-01 -4.92292374e-01 -4.93870415e-02 7.09903538e-01
7.75698304e-01 -3.49666804e-01 1.36254519e-01 -7.48413384e-01
-3.11876863e-01 6.32290959e-01 -2.13510886e-01 6.55909121e-01
-3.91408026e-01 1.27254874e-01 3.91773909e-01 -7.84192562e-01
6.27683103e-01 5.02551258e-01 1.15246892e+00 6.01912498e-01
-1.10491002e+00 -5.43951154e-01 -2.83145249e-01 -3.50097567e-01
-1.30806816e+00 -1.73373967e-01 5.66136718e-01 -7.41010070e-01
9.21975791e-01 1.76044703e-01 1.11861646e+00 7.20858425e-02
3.49660844e-01 5.74537814e-01 7.61591375e-01 -1.63684383e-01
1.69843674e-01 -2.70431429e-01 1.25129268e-01 8.96233976e-01
3.13998371e-01 -1.41072258e-01 7.52242953e-02 -1.78132772e-01
1.36588585e+00 5.54818381e-03 -4.60507005e-01 -5.03423572e-01
-1.31807530e+00 8.71235371e-01 4.20570225e-01 4.10095215e-01
-3.81307542e-01 4.10657257e-01 8.13934982e-01 -1.05782063e-03
3.76417696e-01 5.52679956e-01 -3.84798050e-01 -2.92117268e-01
-1.42961299e+00 9.48685929e-02 4.45793837e-01 6.49460196e-01
1.52596638e-01 2.07323939e-01 -7.59032965e-02 7.20280826e-01
2.00716004e-01 1.13376848e-01 4.84714329e-01 -1.13844788e+00
2.20265046e-01 5.97665787e-01 -1.44317836e-01 -6.06320679e-01
-8.69146287e-01 -7.26472676e-01 -1.00658667e+00 4.83465612e-01
4.85854149e-01 -4.90319520e-01 -1.30817521e+00 1.05668318e+00
1.82694778e-01 -1.28598064e-01 -3.87681693e-01 1.19009900e+00
9.80223179e-01 1.36841401e-01 2.37869129e-01 -1.71507597e-01
1.19803202e+00 -9.21822011e-01 -6.05996430e-01 -2.08355531e-01
8.68758857e-01 -6.71719134e-01 7.31598735e-01 2.54343271e-01
-1.19510770e+00 -3.61274928e-01 -1.10830975e+00 1.87648371e-01
5.05991913e-02 1.95347577e-01 7.66709447e-01 8.00283730e-01
-1.23737311e+00 9.90893185e-01 -1.29588997e+00 8.97492766e-02
9.09156382e-01 5.15710175e-01 -3.02287340e-01 2.60538250e-01
-9.35590327e-01 9.30927575e-01 4.92704183e-01 3.95300329e-01
-3.42876136e-01 -1.11873531e+00 -7.20983624e-01 -3.29163782e-02
3.56773466e-01 -8.08973312e-01 1.43127608e+00 -8.26237321e-01
-1.31573355e+00 1.00088692e+00 3.20933610e-01 -5.14114082e-01
9.28844392e-01 -6.37652501e-02 -2.26355627e-01 5.41816831e-01
8.09469447e-02 8.37627530e-01 6.23510480e-01 -7.45939255e-01
-9.90882739e-02 -1.43488258e-01 -3.47182274e-01 -1.91078093e-02
2.33240053e-01 1.56798124e-01 -5.50489902e-01 -7.56590188e-01
4.23281461e-01 -1.01808226e+00 -6.32472515e-01 4.12989467e-01
-1.09957643e-01 7.67583475e-02 9.12346542e-01 -9.05757248e-01
1.08917463e+00 -2.00671124e+00 -4.34492558e-01 1.38871342e-01
8.55605006e-01 7.10912645e-01 3.57343942e-01 -3.00978124e-01
-3.31585437e-01 4.36160952e-01 -4.98230785e-01 1.76349096e-02
-5.80285192e-01 5.92549555e-02 3.22220057e-01 4.92481530e-01
1.11412168e-01 1.19709277e+00 -7.40148246e-01 -8.43137443e-01
4.28537309e-01 5.79654396e-01 -4.85628545e-01 1.14232913e-01
1.44471645e-01 7.24132895e-01 -2.29186922e-01 5.82565963e-01
5.19280732e-01 -6.85889900e-01 3.79304439e-01 -3.38220298e-02
7.81116716e-04 -3.12768342e-03 -1.02578986e+00 1.76187158e+00
-1.47181004e-01 5.70903361e-01 1.82951212e-01 -9.73164976e-01
8.14436376e-01 7.47713208e-01 7.92158782e-01 -2.79486269e-01
2.77496725e-01 4.14639145e-01 2.45487764e-01 -3.68493855e-01
8.66757557e-02 -2.97174752e-01 2.45541424e-01 6.21268809e-01
-2.07752496e-01 -3.27324599e-01 1.57490835e-01 -1.11959260e-02
1.05906320e+00 1.93846330e-01 3.66074473e-01 -3.83502334e-01
1.31611004e-01 2.26157293e-01 8.26039612e-01 5.10808349e-01
-6.56111300e-01 1.13835371e+00 3.41830522e-01 -1.25491905e+00
-1.22874475e+00 -9.41009939e-01 -2.22093031e-01 4.58333910e-01
-3.69818479e-01 -2.15190321e-01 -9.44295764e-01 -7.53057599e-01
-2.17796400e-01 5.34097292e-02 -4.02606398e-01 4.22320992e-01
-9.92047191e-01 -6.21798813e-01 6.75126553e-01 1.23726070e+00
5.73708057e-01 -1.32768273e+00 -1.52693808e+00 7.15892136e-01
-2.34438255e-01 -1.04130864e+00 -5.54663122e-01 7.89249092e-02
-1.33323824e+00 -1.05460358e+00 -1.11325812e+00 -7.11612105e-01
7.27309644e-01 -2.80775696e-01 1.59403026e+00 6.13972247e-01
-7.38237798e-01 -9.90915298e-02 -1.56310245e-01 -3.89149666e-01
-2.98977077e-01 8.93710405e-02 -2.79457301e-01 -5.61550319e-01
-5.25910296e-02 -4.84424442e-01 -9.37049210e-01 1.41194025e-02
-8.23543429e-01 2.87000865e-01 3.84315282e-01 8.24127197e-01
7.39004552e-01 -3.51238728e-01 6.46690071e-01 -1.27445531e+00
3.66877019e-01 9.09910426e-02 -6.32916570e-01 3.42152938e-02
-6.82556331e-01 -1.53298751e-01 4.79413241e-01 -1.62184745e-01
-6.55256152e-01 3.53777200e-01 -8.38048384e-02 -6.82575524e-01
-8.32993165e-02 6.54013574e-01 6.15209758e-01 -1.77033618e-01
4.11380619e-01 -1.01406403e-01 2.04484358e-01 -2.51466751e-01
1.02849528e-01 4.77735370e-01 7.81463206e-01 -1.48624912e-01
3.35056782e-01 2.36215219e-01 4.93293181e-02 -4.21022683e-01
-4.81233746e-01 -3.12311977e-01 -1.01829147e+00 -4.70562875e-01
1.16857648e+00 -5.98655999e-01 -5.70124567e-01 2.09808305e-01
-1.13393855e+00 -3.96340370e-01 -2.22410887e-01 3.43930393e-01
-4.40802306e-01 5.47606349e-01 -8.28137398e-01 -3.88371557e-01
-1.02182174e+00 -1.23094368e+00 6.08636022e-01 2.46975228e-01
-5.99838734e-01 -9.67191696e-01 -1.77550778e-01 3.73723239e-01
8.21729243e-01 5.94383597e-01 9.61461067e-01 -4.23065990e-01
-4.31091607e-01 -2.91292727e-01 -4.04191643e-01 2.78123051e-01
-6.88915104e-02 -1.40182540e-01 -6.11869752e-01 -2.36713767e-01
-1.63708568e-01 -1.68645933e-01 5.93682587e-01 9.35294986e-01
1.41041458e+00 3.74527633e-01 -3.11990052e-01 6.02530181e-01
1.22099113e+00 4.25745547e-01 5.10422647e-01 2.61278778e-01
7.28004754e-01 1.34950325e-01 3.73904139e-01 2.83043951e-01
7.42379203e-02 2.45864019e-01 2.33402386e-01 -9.40570414e-01
-7.99199939e-02 3.66042823e-01 -5.06386638e-01 8.45174849e-01
-4.39858586e-01 2.35820696e-01 -1.44170916e+00 5.95976293e-01
-1.73744071e+00 -5.52525520e-01 -2.55103439e-01 1.82899880e+00
8.55297983e-01 1.92737356e-01 1.04318462e-01 1.62072167e-01
7.29860485e-01 -1.71470433e-01 -4.69866544e-01 -3.82927835e-01
4.37478215e-01 5.62233686e-01 4.47041214e-01 2.27743179e-01
-1.34275162e+00 6.75854564e-01 6.84505844e+00 2.10757330e-01
-1.52341080e+00 1.40798062e-01 1.05622685e+00 -5.36084548e-02
2.70890504e-01 -2.15091214e-01 -2.26798221e-01 2.46748134e-01
7.03595161e-01 1.37817577e-01 5.40743247e-02 9.47905838e-01
2.90239483e-01 5.95390573e-02 -9.02224362e-01 1.00338697e+00
-1.70681685e-01 -1.80869222e+00 -4.69666630e-01 -1.98316351e-01
3.22570592e-01 2.75480032e-01 -3.78422529e-01 1.34818241e-01
-1.19016804e-01 -1.23171449e+00 4.19544727e-01 2.99562782e-01
1.24269271e+00 -8.06885779e-01 1.12144303e+00 5.43798395e-02
-1.19974172e+00 3.16338032e-01 -1.00187249e-01 1.91577002e-01
5.15422344e-01 7.00747192e-01 -1.12324154e+00 1.92947775e-01
7.56088376e-01 1.59134626e-01 -4.14094657e-01 1.17593551e+00
1.52556553e-01 6.60377860e-01 -2.33667538e-01 4.58429486e-01
2.33842224e-01 1.45588024e-02 3.96721214e-01 1.24440253e+00
1.73054099e-01 3.75883371e-01 7.07725108e-01 9.71878111e-01
-9.07823816e-02 1.60841390e-01 -2.45909348e-01 -2.07230344e-01
1.89040139e-01 1.50641167e+00 -1.53698480e+00 -6.40257180e-01
-2.53038317e-01 7.78266788e-01 9.75557044e-02 -6.62206337e-02
-7.72921979e-01 -3.87342840e-01 -4.25681658e-02 5.17085850e-01
3.07478577e-01 -2.38968179e-01 -7.52221227e-01 -8.07790279e-01
-1.53813381e-02 -6.77626789e-01 3.22539240e-01 -6.96691930e-01
-8.39469373e-01 8.84035647e-01 -4.38412189e-01 -1.14853203e+00
-2.15288773e-01 -4.28171784e-01 -5.75191915e-01 9.78004456e-01
-1.20041442e+00 -8.96065116e-01 -4.33648825e-01 1.96232691e-01
5.64091384e-01 4.15529013e-02 1.02123499e+00 4.67272907e-01
-3.06363255e-01 3.58710825e-01 -5.85276783e-01 6.72095835e-01
4.40907419e-01 -1.39447582e+00 4.99117106e-01 6.94083631e-01
-2.15024680e-01 4.72755969e-01 2.41802111e-01 -8.13058317e-01
-7.17867017e-01 -1.30628002e+00 8.56280148e-01 -2.40021691e-01
2.17464000e-01 1.26313582e-01 -9.58211958e-01 5.77718437e-01
1.07979169e-02 7.65672982e-01 8.17776501e-01 -1.91408589e-01
1.68080002e-01 2.76341498e-01 -1.17839706e+00 4.35616463e-01
5.92662156e-01 -2.40281507e-01 -5.45134962e-01 1.58274621e-01
5.29163718e-01 -9.08757865e-01 -1.16802227e+00 6.28240049e-01
5.58782876e-01 -8.76307070e-01 6.88545108e-01 -2.74103343e-01
4.81354862e-01 -4.42969859e-01 5.83506703e-01 -9.62516785e-01
-3.18532258e-01 -5.65526783e-01 1.05379120e-01 6.03330135e-01
5.83957374e-01 -4.43151146e-01 1.13844454e+00 9.29014862e-01
-5.55404484e-01 -9.77117538e-01 -9.74783778e-01 -4.73391354e-01
1.44020468e-01 -4.09240693e-01 4.60904330e-01 1.07702279e+00
-4.73080017e-02 1.58757567e-01 4.50244136e-02 -1.42158896e-01
4.77493495e-01 1.64847851e-01 2.38271996e-01 -1.39854026e+00
-3.55377406e-01 -4.29102361e-01 -4.87162441e-01 -4.36753064e-01
-3.31298560e-01 -9.44430172e-01 9.18048099e-02 -1.51400638e+00
1.84414625e-01 -5.96899152e-01 -2.27404013e-01 7.76648521e-01
-2.50490874e-01 8.26678872e-01 3.34696352e-01 2.03700200e-01
-3.43454838e-01 -2.01752976e-01 1.49600697e+00 1.39746651e-01
-4.48931366e-01 -1.01212248e-01 -3.60464364e-01 9.77542102e-01
9.32985723e-01 -4.56965983e-01 -2.63509899e-01 -4.01477426e-01
7.93996826e-02 5.41711628e-01 2.57772297e-01 -1.17487645e+00
3.52254994e-02 3.13323438e-01 7.79953122e-01 -9.15296555e-01
-4.52750288e-02 -6.20816648e-01 1.95439816e-01 8.07800770e-01
-1.51668206e-01 3.85818064e-01 3.45070362e-01 -4.85730432e-02
-2.17197254e-01 -1.82358012e-01 1.08969593e+00 -5.77374995e-01
-2.42509723e-01 4.81737822e-01 -7.12514699e-01 1.50776729e-01
9.44405913e-01 -4.74269629e-01 1.49075344e-01 -1.37913913e-01
-1.18545818e+00 1.34466514e-01 2.05538586e-01 -1.10822782e-01
6.07082725e-01 -9.59759116e-01 -6.46302223e-01 3.71169969e-02
-3.02752018e-01 4.16881710e-01 4.23033774e-01 1.12110901e+00
-1.43299365e+00 5.16704082e-01 -2.70708889e-01 -9.68831241e-01
-1.31942427e+00 2.96090424e-01 7.02981830e-01 -5.62507987e-01
-1.40416873e+00 7.72621274e-01 -1.42697960e-01 -4.94601786e-01
-1.60225764e-01 -5.57706416e-01 -2.15997323e-01 -2.72407472e-01
3.98216784e-01 2.36376420e-01 4.08074528e-01 -9.89864096e-02
-4.49587077e-01 3.06456596e-01 -1.28279552e-01 1.28751798e-02
1.23778856e+00 1.81317553e-01 -1.17066249e-01 1.04228720e-01
5.83937705e-01 -3.76413584e-01 -1.11501086e+00 1.24870725e-01
7.64152482e-02 -1.28384128e-01 5.21316826e-01 -1.04566896e+00
-1.44496918e+00 1.00281894e+00 8.38730395e-01 5.43123148e-02
1.05449092e+00 -3.19838792e-01 1.06680238e+00 -4.60056104e-02
1.58985779e-01 -1.09321010e+00 -2.99373299e-01 2.98412234e-01
6.42721474e-01 -1.09929657e+00 -1.31373499e-02 -5.95067561e-01
-7.38768518e-01 1.53463638e+00 5.68494081e-01 -1.87725708e-01
7.15569794e-01 5.96310198e-01 7.47723877e-01 -5.10446370e-01
-2.94517100e-01 2.64330119e-01 9.94728282e-02 4.52889025e-01
9.14315641e-01 3.47795546e-01 -4.51831132e-01 4.95114952e-01
-7.32318833e-02 4.40284163e-01 5.91099441e-01 1.18974006e+00
-2.74903119e-01 -9.99706805e-01 -2.78672278e-01 6.69086099e-01
-8.99187684e-01 -2.59232551e-01 -1.17635570e-01 6.52356267e-01
1.61826879e-01 5.72622895e-01 1.73837140e-01 2.16059893e-01
1.51157394e-01 2.37162918e-01 3.71685028e-01 -8.02482605e-01
-9.91645694e-01 3.99996966e-01 1.38687357e-01 -5.16713083e-01
-2.03894958e-01 -4.27614987e-01 -1.71186936e+00 -7.87000507e-02
-2.79262722e-01 -5.44347474e-03 5.93353570e-01 1.00435865e+00
4.27579194e-01 9.19940948e-01 5.13128042e-02 -7.14575708e-01
-2.97488183e-01 -1.00764763e+00 -4.73148018e-01 1.29870132e-01
7.12319836e-02 -4.43220854e-01 3.31874520e-01 4.15835232e-01] | [14.258268356323242, -2.5152790546417236] |
0f851c87-73a5-4611-a046-268e72e4b365 | talking-head-generation-with-probabilistic | 2212.04248 | null | https://arxiv.org/abs/2212.04248v1 | https://arxiv.org/pdf/2212.04248v1.pdf | Talking Head Generation with Probabilistic Audio-to-Visual Diffusion Priors | In this paper, we introduce a simple and novel framework for one-shot audio-driven talking head generation. Unlike prior works that require additional driving sources for controlled synthesis in a deterministic manner, we instead probabilistically sample all the holistic lip-irrelevant facial motions (i.e. pose, expression, blink, gaze, etc.) to semantically match the input audio while still maintaining both the photo-realism of audio-lip synchronization and the overall naturalness. This is achieved by our newly proposed audio-to-visual diffusion prior trained on top of the mapping between audio and disentangled non-lip facial representations. Thanks to the probabilistic nature of the diffusion prior, one big advantage of our framework is it can synthesize diverse facial motion sequences given the same audio clip, which is quite user-friendly for many real applications. Through comprehensive evaluations on public benchmarks, we conclude that (1) our diffusion prior outperforms auto-regressive prior significantly on almost all the concerned metrics; (2) our overall system is competitive with prior works in terms of audio-lip synchronization but can effectively sample rich and natural-looking lip-irrelevant facial motions while still semantically harmonized with the audio input. | ['Baoyuan Wang', 'Finn Wong', 'Duomin Wang', 'Deyu Zhou', 'Zixin Yin', 'Zhentao Yu'] | 2022-12-07 | null | null | null | null | ['talking-head-generation'] | ['computer-vision'] | [ 3.32728736e-02 1.11291394e-01 -1.59461275e-01 -3.04086745e-01
-1.25472963e+00 -4.29746002e-01 6.28435731e-01 -6.02598965e-01
1.64979056e-01 5.75270712e-01 8.54361296e-01 5.12807846e-01
1.67006910e-01 -2.77991235e-01 -6.94809496e-01 -9.59570587e-01
3.21515352e-01 -1.09019363e-02 2.25457307e-02 -2.79726028e-01
-1.68541148e-01 3.74128908e-01 -2.14105320e+00 2.40185246e-01
4.28456187e-01 1.09537673e+00 -8.00537691e-02 8.02238524e-01
3.12592626e-01 8.33020568e-01 -4.51304227e-01 -4.81529534e-01
5.86384498e-02 -6.11181378e-01 -4.74737436e-01 1.74057439e-01
7.88917184e-01 -4.20878291e-01 -4.65346605e-01 9.17485952e-01
1.09773684e+00 1.46976858e-01 5.36046326e-01 -1.57049155e+00
-4.65743661e-01 4.26756799e-01 -5.44377923e-01 -4.41141218e-01
8.35901797e-01 5.69897830e-01 1.06509340e+00 -1.16139960e+00
6.83246374e-01 1.53633082e+00 5.80727875e-01 8.64243150e-01
-1.48041761e+00 -9.64520752e-01 -4.07473259e-02 2.89921254e-01
-1.63237035e+00 -1.43620861e+00 1.11803246e+00 -3.29148978e-01
4.00859773e-01 3.85405183e-01 5.45363188e-01 1.54452360e+00
-9.10188705e-02 6.60679698e-01 7.18995690e-01 -3.23375374e-01
6.44296855e-02 6.25426322e-02 -6.03295267e-01 3.27818125e-01
-4.44783032e-01 5.50873019e-02 -1.01689637e+00 -1.53651729e-01
4.75939602e-01 -5.34092307e-01 -5.02465904e-01 -3.52800399e-01
-1.27000964e+00 5.67111611e-01 -1.60920210e-02 1.39962628e-01
-3.42316777e-01 3.16716522e-01 3.49324554e-01 -4.80089039e-02
3.50760579e-01 -7.02238306e-02 2.12754887e-02 -1.69084072e-01
-1.18567073e+00 4.35064405e-01 5.28971195e-01 9.39473033e-01
5.15823305e-01 4.39702332e-01 -5.71569324e-01 9.06229615e-01
2.27351025e-01 8.98828804e-01 4.14020360e-01 -1.27677000e+00
1.49224773e-01 -2.06574991e-01 1.24201439e-01 -9.70061481e-01
-2.69317150e-01 -8.10192432e-03 -8.85022461e-01 2.24274442e-01
1.99012056e-01 -3.43021333e-01 -3.89550149e-01 2.39911151e+00
4.43435222e-01 3.07986587e-01 -1.04282565e-01 9.72930551e-01
1.03828418e+00 6.91647828e-01 1.14355735e-01 -5.35309613e-01
1.39020836e+00 -7.49769032e-01 -1.33362937e+00 8.98173749e-02
-1.28680602e-01 -1.13536894e+00 1.12161326e+00 2.38986403e-01
-1.49410963e+00 -8.22191417e-01 -7.54089892e-01 -2.49122873e-01
4.55237329e-01 2.54412293e-01 3.25663447e-01 5.35169601e-01
-1.36144769e+00 1.95700258e-01 -4.35331732e-01 -1.54692009e-01
7.13481084e-02 2.04407886e-01 -5.18337011e-01 1.02725610e-01
-1.10418499e+00 3.92994374e-01 -1.82062089e-01 -1.94191843e-01
-9.91211295e-01 -8.16887498e-01 -1.02954257e+00 4.31018360e-02
2.97175735e-01 -8.19209158e-01 1.41754293e+00 -1.15462136e+00
-2.34775972e+00 6.66556060e-01 -7.23859549e-01 -4.25539427e-02
5.52046359e-01 -2.27589577e-01 -3.85063946e-01 3.73642981e-01
-7.52018914e-02 1.33786631e+00 1.30064023e+00 -1.35541642e+00
-2.64523804e-01 7.12707043e-02 -3.53919715e-01 3.57888907e-01
-4.08898801e-01 1.71795368e-01 -6.46243155e-01 -8.79959881e-01
-2.08938405e-01 -1.08956981e+00 2.90755957e-01 3.17969680e-01
-4.64875370e-01 -1.50886029e-01 9.06721056e-01 -5.15021503e-01
1.03870881e+00 -2.31267142e+00 3.04248810e-01 -2.79125780e-01
2.15649395e-03 1.95978638e-02 -2.41084814e-01 4.19652820e-01
-3.87622535e-01 -1.57273903e-01 5.95000982e-02 -9.20328021e-01
7.26299137e-02 -3.26811612e-01 -7.25718200e-01 6.60835028e-01
2.96340644e-01 8.80221903e-01 -7.83045530e-01 -7.23295450e-01
2.08758190e-01 1.00366735e+00 -9.21685636e-01 4.77937579e-01
-1.12717122e-01 6.56050861e-01 1.09982796e-01 5.75532079e-01
6.67586625e-01 2.86074698e-01 -2.32381616e-02 -4.17527705e-01
1.46273047e-01 1.51264653e-01 -1.22422969e+00 1.90513790e+00
-5.74622035e-01 9.39374387e-01 3.59745234e-01 -2.93621480e-01
1.00015485e+00 9.51110303e-01 6.85873687e-01 -4.07623202e-01
-7.69726560e-02 -9.09021050e-02 -3.60275388e-01 -5.40013611e-01
3.25457752e-01 -3.49109977e-01 1.79631431e-02 4.41218048e-01
3.78858507e-01 -5.84288538e-01 -3.17020237e-01 -7.24882334e-02
4.88386601e-01 2.94423550e-01 6.27899244e-02 -1.11376069e-01
6.97490156e-01 -7.71291733e-01 5.04399657e-01 1.06170490e-01
-2.39196748e-01 1.01508880e+00 5.30352473e-01 1.69262484e-01
-8.93949509e-01 -1.21337020e+00 1.96781263e-01 1.30404925e+00
-7.53467530e-02 -5.15743196e-01 -1.04283941e+00 -1.69061625e-03
-4.41259503e-01 6.71400189e-01 -4.22190934e-01 -1.61186904e-01
-4.69408959e-01 -7.54111335e-02 8.74395967e-01 2.01305002e-01
2.49472082e-01 -1.07469249e+00 -1.86395794e-01 -1.30770179e-02
-7.86292791e-01 -1.23283362e+00 -1.00939047e+00 -6.12950623e-01
-2.82497436e-01 -4.64498222e-01 -1.16712189e+00 -7.10250974e-01
2.39081711e-01 2.39329785e-01 8.07552278e-01 -5.46786904e-01
-2.00478777e-01 4.11286145e-01 1.97441816e-01 -3.67343426e-01
-4.27629977e-01 -3.67077649e-01 3.26047838e-01 6.42838240e-01
-1.58389255e-01 -8.26538980e-01 -7.57593930e-01 3.46693367e-01
-6.74644709e-01 1.06797270e-01 7.73731992e-02 6.95527732e-01
4.61110622e-01 -5.89033365e-02 7.33965158e-01 9.66496859e-03
7.29834378e-01 -3.42546970e-01 -2.68067032e-01 -1.49667755e-01
-3.90320607e-02 -2.08453313e-01 5.12587607e-01 -9.33341622e-01
-1.26611876e+00 2.48940274e-01 -3.58715266e-01 -8.43436301e-01
-1.87238708e-01 -3.01342189e-01 -5.83796680e-01 2.03365967e-01
5.94153702e-01 1.63825169e-01 2.60696620e-01 -2.47227713e-01
6.66891336e-01 6.69101894e-01 9.60711181e-01 -5.62602878e-01
5.76569736e-01 7.81126022e-01 -9.66354832e-03 -1.01581597e+00
-4.36048239e-01 -1.78692594e-01 -3.52140546e-01 -5.12076557e-01
8.38366210e-01 -1.20273435e+00 -1.24263847e+00 5.98803639e-01
-1.32969034e+00 -2.73315370e-01 -3.52667451e-01 5.03910959e-01
-1.03285468e+00 3.32148284e-01 -5.10484993e-01 -1.10344017e+00
-3.38025033e-01 -1.40495074e+00 1.50733054e+00 -2.22087428e-02
-6.55553877e-01 -5.76249778e-01 1.61270678e-01 2.74156094e-01
4.80522633e-01 3.17444235e-01 5.35087824e-01 6.00846410e-02
-3.06781381e-01 -8.19405392e-02 1.57386914e-01 2.90658534e-01
2.99508363e-01 3.94819677e-01 -1.55655932e+00 -1.61928982e-01
8.31468627e-02 -5.47208071e-01 4.75271165e-01 5.86736798e-01
8.73737216e-01 -6.07195616e-01 1.23949666e-02 4.67668325e-01
9.04699504e-01 -1.46898732e-01 6.25985742e-01 -5.91382861e-01
4.69764888e-01 1.16274774e+00 4.44321573e-01 5.99402606e-01
3.46670836e-01 1.23847079e+00 3.33080471e-01 -2.28444725e-01
-6.13705456e-01 -6.37696445e-01 8.71078312e-01 6.95048392e-01
1.64543405e-01 -2.33275697e-01 -5.16815722e-01 5.59352756e-01
-1.60479164e+00 -1.27300131e+00 2.46628344e-01 2.09423280e+00
1.10767257e+00 -1.84565246e-01 4.83442962e-01 3.51802170e-01
8.09928358e-01 2.86631972e-01 -4.02562410e-01 -2.52205700e-01
-2.84377158e-01 3.11102480e-01 -1.66719466e-01 7.62750626e-01
-8.25743318e-01 9.42781806e-01 6.00507641e+00 1.09600341e+00
-1.47432554e+00 2.00921565e-01 5.59330225e-01 -5.51019609e-01
-4.59322810e-01 -3.07717323e-01 -6.68146193e-01 3.01741779e-01
8.91202033e-01 -2.24418074e-01 4.55110878e-01 6.49242580e-01
7.12842524e-01 1.55196860e-01 -1.22662377e+00 1.24179661e+00
2.44261459e-01 -1.21618283e+00 1.23706438e-01 -9.82227176e-02
7.05485046e-01 -4.90010113e-01 4.85751987e-01 -2.80685816e-02
5.34782894e-02 -1.23262060e+00 1.26040781e+00 7.21505582e-01
1.44342041e+00 -9.26031113e-01 2.26574883e-01 1.80425197e-01
-1.19480050e+00 2.03806236e-01 1.84481725e-01 2.84989685e-01
3.09781581e-01 1.90022752e-01 -6.75880671e-01 2.73540437e-01
6.64310753e-01 3.60384256e-01 -5.84600791e-02 5.99956751e-01
-3.81842077e-01 5.16804516e-01 -2.46434376e-01 2.92590290e-01
-2.30540693e-01 2.90633321e-01 8.50325167e-01 1.17083538e+00
3.18798810e-01 -5.80758601e-02 -2.32844457e-01 8.43580902e-01
-6.00324310e-02 2.09688291e-01 -6.90038443e-01 2.18184650e-01
6.08098805e-01 1.14679539e+00 -1.16695955e-01 -6.86859787e-02
-5.17282486e-02 8.53046060e-01 -3.11045200e-01 5.21181762e-01
-1.04335225e+00 -3.19001228e-01 1.10521293e+00 1.05041727e-01
2.21713409e-01 -4.58694808e-03 -6.24038428e-02 -8.53855848e-01
4.67824712e-02 -1.00015128e+00 -1.27865508e-01 -1.16527450e+00
-9.39359248e-01 7.46654391e-01 -8.34573582e-02 -1.34376979e+00
-7.35462487e-01 1.31606394e-02 -5.68554699e-01 9.56533611e-01
-1.36537910e+00 -1.26804662e+00 -3.02896112e-01 1.10937369e+00
6.48202717e-01 5.60002290e-02 9.62094724e-01 3.04144323e-01
-3.75964433e-01 9.43909407e-01 -3.98123920e-01 -1.25725001e-01
1.23288190e+00 -5.85926056e-01 1.34014770e-01 5.33891618e-01
1.88982740e-01 4.47385520e-01 1.07037508e+00 -2.14114726e-01
-1.32830703e+00 -8.92435431e-01 8.70692372e-01 -3.18302572e-01
4.13246781e-01 -6.40544116e-01 -6.71565652e-01 2.64460593e-01
4.48491603e-01 7.72368833e-02 7.32540905e-01 -3.77652675e-01
-4.28074807e-01 -4.46103871e-01 -1.02189445e+00 9.26169157e-01
8.00135434e-01 -7.25469410e-01 -3.26481402e-01 2.54963368e-01
8.73767495e-01 -2.46612400e-01 -6.07698679e-01 4.27429199e-01
8.15921009e-01 -1.20561481e+00 1.07543874e+00 -2.62886614e-01
3.59325409e-01 -1.98181286e-01 -2.62216717e-01 -9.43635702e-01
-1.39557540e-01 -1.53369296e+00 -2.09283903e-02 1.74535322e+00
6.53732568e-02 -2.52230048e-01 4.49439198e-01 4.03723776e-01
2.18607619e-01 -4.95794773e-01 -1.00412059e+00 -5.81478357e-01
-2.45009400e-02 -6.39312863e-01 7.06951976e-01 6.24055982e-01
2.57393960e-02 5.24997115e-01 -8.41663659e-01 1.62108198e-01
6.06665015e-01 -1.45075858e-01 1.20638442e+00 -8.66606116e-01
-2.41720259e-01 -5.78019500e-01 -1.77969605e-01 -1.03325415e+00
4.07642305e-01 -4.35440689e-01 2.09004983e-01 -9.23056424e-01
3.39866877e-02 5.78988008e-02 9.70498398e-02 3.01458806e-01
3.07203736e-02 3.79084975e-01 3.24130565e-01 3.52813274e-01
-2.08997473e-01 9.69416678e-01 1.32708025e+00 1.28337517e-01
-3.06151003e-01 6.82883412e-02 -6.32935345e-01 8.25816810e-01
3.21083575e-01 -3.36022645e-01 -5.46240449e-01 -1.17764071e-01
-1.68772191e-01 6.26890063e-01 3.56105387e-01 -8.92103434e-01
3.38414878e-01 -1.61369488e-01 6.52524317e-03 -3.35665822e-01
1.08315587e+00 -6.19753242e-01 2.45902851e-01 1.50131965e-02
-5.12434244e-01 -3.53401154e-01 3.62178594e-01 5.01026630e-01
-3.37162465e-01 4.17122006e-01 1.07809043e+00 4.01280850e-01
-7.27486685e-02 3.55139434e-01 -2.50867695e-01 1.18128255e-01
9.41032290e-01 -3.94148417e-02 1.28449291e-01 -1.29478860e+00
-5.97452819e-01 -2.51298279e-01 4.30083841e-01 5.87239087e-01
5.06188095e-01 -1.62917531e+00 -9.52888250e-01 2.66972214e-01
2.21574735e-02 -2.29828522e-01 4.86016124e-01 8.68215442e-01
-2.58691479e-02 4.06375378e-01 -2.08519831e-01 -9.41835105e-01
-1.51304471e+00 5.69128454e-01 1.79121047e-01 3.29226673e-01
-5.48133969e-01 8.71554315e-01 5.75946152e-01 1.43628538e-01
6.46105111e-01 4.54703299e-03 -4.12956253e-02 1.68643326e-01
5.79868495e-01 1.72078326e-01 -1.70796171e-01 -1.04685795e+00
-3.38896900e-01 6.99984968e-01 5.81412375e-01 -7.11765051e-01
1.05790198e+00 -3.38386536e-01 2.17659637e-01 5.41016519e-01
1.30299556e+00 7.69059122e-01 -1.62908208e+00 -1.81500927e-01
-5.00672460e-01 -2.85368860e-01 -2.95155458e-02 -3.82418483e-01
-1.11512756e+00 1.21666229e+00 3.96788269e-01 -2.15087891e-01
1.34830856e+00 -8.77303723e-03 6.21099710e-01 -9.13162977e-02
2.83876900e-02 -8.09890926e-01 5.28959036e-01 -3.55697377e-03
1.33449757e+00 -8.57752502e-01 -3.38699132e-01 -4.22546566e-01
-1.00311017e+00 1.11498284e+00 2.15287507e-01 1.27119243e-01
5.21978021e-01 3.85618895e-01 1.63473383e-01 2.63335794e-01
-1.02116668e+00 -9.16750878e-02 5.23517668e-01 7.38687515e-01
4.95792389e-01 -2.77892053e-01 3.55113506e-01 5.74480176e-01
-5.61581790e-01 -3.13504972e-02 1.91048786e-01 2.40408152e-01
-6.14322796e-02 -6.74723804e-01 -5.32123804e-01 -6.41970694e-01
-4.97204244e-01 -8.13159794e-02 -4.31796968e-01 6.52903557e-01
-3.55240852e-02 1.30200934e+00 -3.66139896e-02 -2.55487144e-01
2.58997440e-01 2.48421356e-01 4.62199718e-01 -2.72859246e-01
-1.44345731e-01 6.64066911e-01 -4.30358499e-02 -7.83826053e-01
-2.66040713e-01 -7.80342281e-01 -8.90956879e-01 -3.78603876e-01
-1.99223250e-01 -1.96728572e-01 6.20463192e-01 5.60593188e-01
5.42950630e-01 3.36551279e-01 8.94281149e-01 -1.25838447e+00
-4.61088747e-01 -9.68469620e-01 -2.79678613e-01 4.96712804e-01
6.97762132e-01 -7.07850993e-01 -5.16792357e-01 3.92862380e-01] | [13.224620819091797, -0.41813868284225464] |
f80d13bd-c872-42d8-bb96-a702b41f4a50 | face-image-quality-assessment-a-literature | 2009.01103 | null | https://arxiv.org/abs/2009.01103v3 | https://arxiv.org/pdf/2009.01103v3.pdf | Face Image Quality Assessment: A Literature Survey | The performance of face analysis and recognition systems depends on the quality of the acquired face data, which is influenced by numerous factors. Automatically assessing the quality of face data in terms of biometric utility can thus be useful to detect low-quality data and make decisions accordingly. This survey provides an overview of the face image quality assessment literature, which predominantly focuses on visible wavelength face image input. A trend towards deep learning based methods is observed, including notable conceptual differences among the recent approaches, such as the integration of quality assessment into face recognition models. Besides image selection, face image quality assessment can also be used in a variety of other application scenarios, which are discussed herein. Open issues and challenges are pointed out, i.a. highlighting the importance of comparability for algorithm evaluations, and the challenge for future work to create deep learning approaches that are interpretable in addition to providing accurate utility predictions. | ['Torsten Schlett', 'Julian Fierrez', 'Olaf Henniger', 'Javier Galbally', 'Christoph Busch', 'Christian Rathgeb'] | 2020-09-02 | null | null | null | null | ['face-image-quality', 'face-image-quality-assessment'] | ['computer-vision', 'computer-vision'] | [ 4.12968069e-01 -2.05793485e-01 2.10767929e-02 -7.26381302e-01
-6.39233351e-01 -4.19852257e-01 4.36545044e-01 -1.30217686e-01
-2.07909152e-01 4.27539021e-01 9.11687464e-02 -3.59911025e-02
-6.55321717e-01 -8.77084970e-01 -1.73141479e-01 -1.05905187e+00
8.74240696e-02 3.74149024e-01 -7.36462891e-01 -3.80639806e-02
3.28075022e-01 1.07449543e+00 -2.08269620e+00 4.65774238e-01
4.93730873e-01 1.41511643e+00 -4.27457392e-01 5.06627560e-01
9.49599221e-02 2.73072481e-01 -7.46963859e-01 -8.51498902e-01
2.27898076e-01 -4.46111619e-01 -6.76860690e-01 3.17765921e-01
7.56179988e-01 -4.40238684e-01 4.27171635e-03 9.33216453e-01
9.25086021e-01 -1.94855914e-01 8.29988778e-01 -1.12261856e+00
-7.55333900e-01 -3.61053720e-02 -3.24164242e-01 4.01847929e-01
4.21691716e-01 5.72419822e-01 8.91188562e-01 -9.64334905e-01
3.24634939e-01 1.44868159e+00 5.27310014e-01 7.58006573e-01
-1.20233572e+00 -7.77022600e-01 -3.94744873e-01 5.94621718e-01
-1.08553517e+00 -1.13080704e+00 7.95052767e-01 -5.23728132e-01
8.71924758e-01 2.92243451e-01 5.45124888e-01 9.78117704e-01
2.41685882e-01 3.38263810e-01 1.35792470e+00 -5.41955173e-01
1.34988457e-01 -4.00491245e-03 -6.62366152e-02 8.63034010e-01
3.49747777e-01 6.47356510e-01 -8.81312311e-01 -1.06098890e-01
3.52148145e-01 -3.98440272e-01 7.99569115e-02 -1.31732851e-01
-5.62398434e-01 6.61191642e-01 2.51863331e-01 3.60656530e-01
-3.33248019e-01 -7.67553970e-02 2.21463233e-01 6.56120360e-01
4.55286741e-01 3.37737173e-01 -2.24066943e-01 1.57692939e-01
-1.26825047e+00 9.05441120e-02 4.09992725e-01 1.87094823e-01
7.88612306e-01 3.26757252e-01 -3.18978399e-01 9.58712280e-01
7.73518145e-01 7.39092886e-01 -2.60056779e-02 -1.05979049e+00
-6.98877573e-02 4.87501770e-01 -1.62470117e-01 -9.25992489e-01
-5.37485719e-01 -3.33967626e-01 -5.16106188e-01 9.44196224e-01
3.07522357e-01 -3.50049958e-02 -8.58323872e-01 1.26736736e+00
1.71968624e-01 -2.03050941e-01 -1.66914299e-01 7.62614965e-01
1.32116067e+00 5.44387624e-02 1.01296157e-01 -3.87098879e-01
1.38895130e+00 -1.33796602e-01 -7.15791583e-01 1.99596599e-01
-6.84986115e-02 -1.00164890e+00 7.93458223e-01 5.18587053e-01
-1.14388776e+00 -6.45211637e-01 -1.05587876e+00 1.01100020e-01
-3.10324907e-01 4.54050899e-01 5.03666580e-01 1.57757759e+00
-1.37388051e+00 6.10696793e-01 -5.82375109e-01 -3.73262465e-01
8.69900107e-01 8.47727418e-01 -5.06739914e-01 -1.03588521e-01
-8.87771249e-01 1.03781176e+00 3.33494656e-02 3.72376591e-01
-9.37836647e-01 -7.31899381e-01 -6.93746090e-01 -4.56549935e-02
-3.98436002e-03 -5.29830217e-01 1.06335115e+00 -1.07714260e+00
-1.59871495e+00 1.20810080e+00 -3.37596238e-01 -3.28254364e-02
4.55316871e-01 3.36574823e-01 -7.82018244e-01 2.90985495e-01
-3.31359208e-01 3.93048435e-01 1.17204320e+00 -1.13880980e+00
-4.85958487e-01 -9.08121407e-01 -6.91969097e-02 -1.10715717e-01
-3.08474451e-01 5.36905766e-01 -5.75467907e-02 -2.31014580e-01
-1.00292131e-01 -6.02582932e-01 2.84630775e-01 3.09176296e-01
3.30922753e-02 -4.27596748e-01 7.71358192e-01 -5.45592248e-01
1.02070892e+00 -1.84889758e+00 -1.96971938e-01 3.06860685e-01
1.83818087e-01 3.35131258e-01 -3.38629991e-01 3.77976716e-01
-1.74272671e-01 1.29747674e-01 -1.12837374e-01 -2.29518011e-01
-7.77593255e-02 -1.89703524e-01 1.05884559e-01 7.78459549e-01
7.90462375e-01 9.46195602e-01 -4.29332167e-01 -1.90426499e-01
4.84809965e-01 8.54250014e-01 -2.66315281e-01 1.36507705e-01
6.72962889e-02 5.58128178e-01 -1.65086403e-01 1.05349517e+00
9.07477200e-01 4.20989357e-02 8.00432637e-02 -5.77977836e-01
3.93840857e-03 2.47329608e-01 -1.01304984e+00 9.37090397e-01
-4.57376212e-01 7.07208693e-01 1.79710284e-01 -9.19114172e-01
9.21987474e-01 4.51844513e-01 7.31188476e-01 -1.02073765e+00
2.69644886e-01 2.52352208e-01 3.65852535e-01 -6.86942816e-01
5.97376004e-02 -5.28555930e-01 6.35525703e-01 5.25432467e-01
3.74841511e-01 1.88693136e-01 -2.68736798e-02 -5.34344196e-01
7.50552893e-01 -8.67251772e-03 1.92735687e-01 -2.42792830e-01
8.48090947e-01 -4.92317706e-01 1.82953864e-01 3.26136440e-01
-5.25508523e-01 4.58220661e-01 3.75480294e-01 -5.27741909e-01
-7.69037664e-01 -8.66137743e-01 -7.79624879e-01 8.94780338e-01
-2.65312910e-01 -9.17167738e-02 -8.54812205e-01 -4.18500185e-01
9.89584103e-02 6.85276836e-02 -8.32365990e-01 -2.84543157e-01
-4.15019423e-01 -1.26772559e+00 4.92578685e-01 2.74729103e-01
3.59543324e-01 -1.27543616e+00 -6.55393958e-01 7.09094107e-03
2.37691253e-01 -8.94352615e-01 4.01357234e-01 -3.01050991e-01
-8.88053477e-01 -1.35749781e+00 -4.17203546e-01 -2.37687096e-01
3.94063443e-01 1.18458740e-01 1.24022079e+00 5.00922799e-01
-5.38739979e-01 7.05702603e-01 -1.03051379e-01 -4.12625700e-01
-5.15166581e-01 -3.50618750e-01 1.49214834e-01 2.35528216e-01
7.92292476e-01 -1.41769886e-01 -8.22122574e-01 2.41460055e-01
-7.52052844e-01 -6.67631447e-01 5.83949685e-01 9.48893487e-01
3.62119794e-01 2.32178137e-01 7.41954267e-01 -7.16692746e-01
6.37321889e-01 -2.26482734e-01 -5.34774423e-01 4.37463492e-01
-1.18807638e+00 -9.38649029e-02 2.06290424e-01 2.85360366e-01
-1.28686786e+00 -2.49601021e-01 -5.52325845e-01 7.14712143e-02
-4.57953513e-01 2.79370397e-01 -4.49251354e-01 -6.85658514e-01
8.54673743e-01 -9.64578316e-02 4.17617708e-01 -4.54407007e-01
8.62739682e-02 6.77439809e-01 2.16478646e-01 -4.81550395e-01
6.03761911e-01 6.04616702e-01 2.73635089e-01 -1.22668934e+00
-3.03593934e-01 -2.21735582e-01 -7.62123704e-01 -5.98172843e-01
5.79713643e-01 -5.36429286e-01 -1.03728282e+00 5.38665056e-01
-8.13775063e-01 9.78772417e-02 7.75873661e-02 8.87414739e-02
-5.10272145e-01 3.48593891e-01 -4.48554993e-01 -1.11409009e+00
-5.01820743e-01 -1.50048780e+00 1.19719350e+00 2.44738609e-01
5.53712212e-02 -9.28283155e-01 -2.35217646e-01 8.90800238e-01
4.52748984e-01 3.60186920e-02 1.07934690e+00 -2.51445085e-01
-4.75059241e-01 -1.59373954e-01 -2.84863919e-01 3.54583085e-01
4.48617131e-01 4.11855757e-01 -1.79079413e+00 -5.25002480e-01
-1.56583358e-02 -3.56785625e-01 9.13138330e-01 5.50560236e-01
1.09823966e+00 -1.09200045e-01 -9.33552720e-03 4.56018269e-01
1.44816875e+00 2.03979149e-01 9.00819600e-01 5.55145293e-02
1.63087666e-01 1.01613665e+00 3.97981882e-01 2.21022576e-01
-1.14899494e-01 7.53227234e-01 5.20287395e-01 -8.97345990e-02
-5.25132000e-01 4.06391650e-01 3.20453405e-01 5.08544110e-02
-4.30480450e-01 -5.71793206e-02 -8.30774724e-01 -1.88975763e-02
-1.08542681e+00 -1.27617121e+00 4.58461158e-02 2.19874930e+00
4.86481667e-01 -3.12585413e-01 4.18520033e-01 5.40633202e-01
4.77965415e-01 -2.43511423e-02 -6.41360343e-01 -3.05769861e-01
-1.29274413e-01 5.97094297e-01 -5.61269466e-03 4.11652714e-01
-8.98248553e-01 6.27999723e-01 7.79654169e+00 5.52899957e-01
-1.42589915e+00 4.32391167e-02 1.04787457e+00 -8.65448713e-02
-2.16367021e-01 -5.55287719e-01 -6.86744928e-01 3.71934623e-01
1.17520022e+00 9.79387537e-02 4.75640893e-01 2.53110528e-01
4.27823514e-01 -6.07256368e-02 -1.10105801e+00 1.29788351e+00
2.66411692e-01 -1.14375699e+00 -2.19375342e-02 2.19318002e-01
4.72617447e-01 -1.50396049e-01 5.99680364e-01 -1.49374023e-01
-4.22813773e-01 -1.51349163e+00 4.10844862e-01 5.55232108e-01
9.71702039e-01 -8.82997155e-01 7.00382888e-01 -3.92979175e-01
-8.88633788e-01 -2.90236831e-01 -5.51031888e-01 9.95767117e-02
-4.34384197e-01 6.82113707e-01 -5.27829230e-01 5.86717665e-01
9.18442488e-01 4.61521029e-01 -7.24361360e-01 1.01876390e+00
1.15069740e-01 4.71474171e-01 8.48489106e-02 1.24534957e-01
-3.11653942e-01 -3.29151511e-01 2.73225546e-01 1.02687109e+00
1.99197918e-01 -9.17320549e-02 -3.51228893e-01 9.61534500e-01
-5.31632379e-02 1.53991371e-01 -5.23702085e-01 -1.33298621e-01
1.79077521e-01 1.48715723e+00 -6.38929248e-01 1.11032948e-02
-6.77808762e-01 5.67844987e-01 7.54016265e-02 1.94484413e-01
-2.24769831e-01 1.95217609e-01 9.80855107e-01 2.14857638e-01
-2.07884818e-01 1.53042087e-02 -4.06657517e-01 -6.48253441e-01
1.61324769e-01 -1.10099494e+00 5.26387334e-01 -4.05988872e-01
-1.42963111e+00 7.46805251e-01 -1.77668318e-01 -9.16185379e-01
-1.12298064e-01 -1.22120869e+00 -3.73715609e-01 1.02930117e+00
-1.85720074e+00 -1.21687675e+00 -3.72024179e-01 2.30508983e-01
3.90039086e-01 -5.64695954e-01 8.54681313e-01 3.65047157e-01
-4.57954437e-01 8.09720516e-01 9.72448066e-02 -1.44224524e-01
6.25578523e-01 -8.73413682e-01 3.65455486e-02 4.98827219e-01
2.06610844e-01 4.89967167e-01 5.89954615e-01 -2.35709310e-01
-1.49465561e+00 -7.48624623e-01 5.10740280e-01 -6.83094144e-01
7.38673508e-02 3.19479443e-02 -7.17573464e-01 4.60281596e-02
2.54450917e-01 -1.14088871e-01 1.26823902e+00 2.32366174e-01
-3.62353593e-01 -5.35176337e-01 -1.63940871e+00 2.28478551e-01
9.70626473e-01 -7.45662391e-01 -4.61059995e-02 2.30811432e-01
-3.82101506e-01 1.22481301e-01 -9.35851216e-01 6.73156202e-01
1.00044560e+00 -1.54614460e+00 1.03894460e+00 -6.45745933e-01
2.78172165e-01 -2.75822598e-02 -1.79500561e-02 -1.06539774e+00
-4.95033175e-01 -1.13385789e-01 -5.00696041e-02 1.08917344e+00
3.32374573e-01 -4.89970446e-01 9.38019335e-01 8.00189435e-01
3.94994169e-01 -7.42052913e-01 -1.11989093e+00 -6.19289219e-01
1.40645653e-01 -4.49969441e-01 9.18709397e-01 4.93560374e-01
-4.04792309e-01 9.21023488e-02 -1.54259667e-01 1.34205759e-01
8.08767915e-01 6.91980943e-02 3.05862904e-01 -1.59357440e+00
1.71090305e-01 -8.87771964e-01 -7.54897118e-01 -7.32677281e-02
2.69592702e-01 -7.61926889e-01 -2.42126822e-01 -1.39269805e+00
1.67174965e-01 -2.50997454e-01 -4.80503738e-01 1.86324105e-01
-5.12946546e-02 8.23599160e-01 -5.84002621e-02 6.99860826e-02
1.36257634e-02 4.48888719e-01 1.18772900e+00 -3.35714489e-01
1.74174517e-01 1.55735701e-01 -8.47106218e-01 4.17335331e-01
7.76920617e-01 1.08667612e-02 -3.76379192e-01 -4.12536800e-01
1.25849843e-01 -1.35608345e-01 3.69845271e-01 -1.20537543e+00
-1.70565605e-01 -1.74094364e-01 9.60066915e-01 -1.27960026e-01
4.42257047e-01 -1.00697303e+00 1.34902373e-01 4.82105732e-01
-2.10852370e-01 5.47090396e-02 1.51705950e-01 2.47447148e-01
-6.96066171e-02 -2.63959020e-01 1.33591843e+00 -1.21511072e-01
-6.67902052e-01 5.33715665e-01 -3.21926683e-01 -5.77975571e-01
8.51023912e-01 -5.25400758e-01 -2.16596454e-01 -3.34317803e-01
-6.93618536e-01 -2.19499946e-01 3.16129714e-01 4.34844285e-01
7.86389470e-01 -1.28550851e+00 -8.15421999e-01 5.78639925e-01
3.86921823e-01 -8.84580791e-01 4.40047920e-01 6.12737477e-01
-3.16692322e-01 4.98697847e-01 -5.64866126e-01 -5.92205346e-01
-1.66951692e+00 4.19846535e-01 7.36865938e-01 3.33863199e-01
3.71167511e-02 6.67736769e-01 -1.90384015e-01 -2.57346064e-01
8.00946578e-02 1.36346251e-01 -6.10682786e-01 2.82351404e-01
9.25587833e-01 5.84055543e-01 8.10795903e-01 -1.01581550e+00
-5.52232444e-01 7.76327312e-01 2.45450228e-01 2.44233795e-02
1.37287676e+00 -5.24912365e-02 -4.74483371e-01 2.41851006e-02
1.04271400e+00 -2.07227260e-01 -9.49369729e-01 -2.80007459e-02
1.20731376e-01 -6.76883340e-01 2.42147684e-01 -1.21998310e+00
-1.41330802e+00 1.22723913e+00 1.49146366e+00 -2.05000620e-02
1.50093579e+00 -5.46858497e-02 4.12996039e-02 2.45085686e-01
3.37250561e-01 -1.04593217e+00 3.03888589e-01 -5.82580827e-02
1.07778490e+00 -1.56745410e+00 1.69081748e-01 -4.47396517e-01
-1.47912279e-01 1.51900458e+00 5.70972979e-01 2.92835355e-01
9.05381680e-01 7.64395818e-02 3.45411062e-01 -6.18891358e-01
-5.35128713e-01 -3.40550721e-01 6.77627623e-01 1.04136336e+00
9.26778734e-01 -5.19842841e-02 -2.72265255e-01 1.30500674e-01
-1.42701507e-01 -4.46449928e-02 1.30023018e-01 5.08693695e-01
-3.88817519e-01 -1.53839827e+00 -5.38747013e-01 8.53745937e-01
-6.28586829e-01 2.43803307e-01 -5.52792013e-01 1.53398335e-01
2.64581323e-01 1.32998013e+00 -3.51592787e-02 -3.84360760e-01
2.31153980e-01 3.66844893e-01 8.53342354e-01 -5.63556731e-01
-5.61134160e-01 -1.07591525e-01 1.14981793e-02 -4.12395775e-01
-8.24162185e-01 -7.63509274e-01 -7.14345515e-01 -5.48230171e-01
-3.03727984e-01 -1.83047280e-01 7.78713584e-01 1.07961583e+00
3.55171829e-01 3.37346762e-01 6.01652861e-01 -6.71479821e-01
-4.41844285e-01 -8.67237031e-01 -5.92611969e-01 2.61641175e-01
5.75054646e-01 -9.42525864e-01 -2.11362243e-02 6.74725929e-03] | [13.107956886291504, 0.8547645211219788] |
6d67f1b1-dbd6-4759-99e4-b2262764a053 | multi-scale-spatial-temporal-interaction | 2306.10239 | null | https://arxiv.org/abs/2306.10239v2 | https://arxiv.org/pdf/2306.10239v2.pdf | Multi-scale Spatial-temporal Interaction Network for Video Anomaly Detection | Video Anomaly Detection (VAD) is an essential yet challenging task in signal processing. Since certain anomalies cannot be detected by isolated analysis of either temporal or spatial information, the interaction between these two types of data is considered crucial for VAD. However, current dual-stream architectures either confine this integral interaction to the bottleneck of the autoencoder or introduce anomaly-irrelevant background pixels into the interactive process, hindering the accuracy of VAD. To address these deficiencies, we propose a Multi-scale Spatial-Temporal Interaction Network (MSTI-Net) for VAD. First, to prioritize the detection of moving objects in the scene and harmonize the substantial semantic discrepancies between the two types of data, we propose an Attention-based Spatial-Temporal Fusion Module (ASTFM) as a substitute for the conventional direct fusion. Furthermore, we inject multi-ASTFM-based connections that bridge the appearance and motion streams of the dual-stream network, thus fostering multi-scale spatial-temporal interaction. Finally, to bolster the delineation between normal and abnormal activities, our system records the regular information in a memory module. Experimental results on three benchmark datasets validate the effectiveness of our approach, which achieves AUCs of 96.8%, 87.6%, and 73.9% on the UCSD Ped2, CUHK Avenue, and ShanghaiTech datasets, respectively. | ['Zile Wang', 'Zhengliang Guo', 'Liang Song', 'Zhangxun Li', 'Zhiyuan Ning'] | 2023-06-17 | null | null | null | null | ['video-anomaly-detection', 'anomaly-detection'] | ['computer-vision', 'methodology'] | [ 2.29033642e-02 -5.21823049e-01 1.66672945e-01 -2.53381670e-01
-3.02672416e-01 -1.73003495e-01 5.31221211e-01 2.05974802e-01
-3.55179369e-01 3.14125389e-01 5.80203719e-02 -2.45635271e-01
-9.93578807e-02 -6.42440677e-01 -4.52631474e-01 -7.08559632e-01
-3.32288057e-01 -2.70347744e-01 5.60782611e-01 -1.63233206e-01
2.39322260e-02 4.74673688e-01 -1.63148808e+00 5.45644224e-01
1.10901058e+00 1.45957816e+00 -1.36820778e-01 4.69320625e-01
-2.69575506e-01 8.76243830e-01 -7.38580525e-01 -1.61968336e-01
2.41324484e-01 -4.37303483e-01 -2.70212680e-01 1.16494671e-01
4.23297554e-01 -4.29716587e-01 -5.90077341e-01 1.14067054e+00
4.59538758e-01 1.47195458e-01 4.16194171e-01 -1.59767890e+00
-2.48515919e-01 2.88679510e-01 -8.47354293e-01 8.93572032e-01
2.56424695e-01 2.77301133e-01 7.22804189e-01 -9.66538250e-01
6.09368198e-02 1.09519613e+00 5.05654871e-01 2.52033025e-01
-9.25884128e-01 -6.79886281e-01 6.75560534e-01 6.19550049e-01
-1.37528563e+00 -3.85453045e-01 1.02232218e+00 -4.98999566e-01
8.16960692e-01 2.30058566e-01 8.41594994e-01 1.28816628e+00
2.40287721e-01 1.04375112e+00 6.29617393e-01 -4.20314185e-02
6.99504167e-02 -3.07133108e-01 1.52337238e-01 4.57588524e-01
2.76392967e-01 -7.03020468e-02 -7.69210935e-01 2.66710762e-02
8.17198277e-01 3.42123449e-01 -4.29840237e-01 -2.16394499e-01
-1.17327273e+00 3.49016458e-01 3.71627301e-01 4.06461626e-01
-6.05894804e-01 -1.92589596e-01 6.95335269e-01 3.97997230e-01
5.84322572e-01 4.93103042e-02 -1.81912675e-01 -7.64695257e-02
-9.66540337e-01 -2.54546888e-02 9.97794867e-02 6.55414462e-01
2.80457795e-01 4.28147376e-01 -4.87427533e-01 6.12071693e-01
3.91242713e-01 3.44556659e-01 4.73312497e-01 -5.20110846e-01
5.37551582e-01 8.59212041e-01 -7.53365606e-02 -1.32902682e+00
-3.78152221e-01 -6.48163497e-01 -1.11385691e+00 3.28130811e-01
5.70334733e-01 7.45600909e-02 -1.04765534e+00 1.66033590e+00
4.29085642e-01 8.29211414e-01 -2.21609641e-02 1.12921858e+00
7.37103462e-01 6.33998871e-01 2.25046054e-01 -3.10024112e-01
1.26605070e+00 -8.89483392e-01 -8.87218595e-01 -1.49989083e-01
4.85384047e-01 -4.25889701e-01 8.42994273e-01 3.82737964e-01
-1.08788121e+00 -7.26202309e-01 -1.14660108e+00 1.93745732e-01
-2.81048864e-01 -2.04837080e-02 3.17111641e-01 1.88420922e-01
-7.52766311e-01 4.02561426e-01 -9.48550880e-01 -5.13853431e-02
5.57450771e-01 1.31538808e-01 -3.02729309e-01 6.58430681e-02
-1.23408246e+00 4.58880991e-01 2.77367026e-01 5.03785729e-01
-7.55955696e-01 -6.17345512e-01 -7.52069712e-01 5.87802902e-02
4.06981319e-01 -4.71978664e-01 8.99029732e-01 -1.18891406e+00
-9.38918531e-01 3.87814015e-01 -1.19192421e-01 -5.65403402e-01
6.62981749e-01 -3.22650850e-01 -9.53376532e-01 3.98625523e-01
1.64615810e-02 3.63168120e-01 1.12782681e+00 -1.05233741e+00
-1.06051135e+00 -3.91444147e-01 -3.64973024e-03 3.19158167e-01
-5.00154495e-01 -1.75859839e-01 -6.47684574e-01 -9.84188616e-01
3.20573092e-01 -3.68572444e-01 6.78006746e-03 2.82632280e-02
-1.78776532e-01 -2.04243556e-01 1.25833344e+00 -8.58005822e-01
1.58774090e+00 -2.77550077e+00 8.72853994e-02 2.73429155e-01
4.46714193e-01 5.77934682e-01 -1.22993745e-01 -1.99594632e-01
-3.68946046e-01 -2.89920628e-01 -3.12626809e-01 -3.04808438e-01
-2.99647689e-01 8.33851323e-02 -2.58428365e-01 4.19952005e-01
3.51424724e-01 6.75857008e-01 -8.93151760e-01 -4.97896999e-01
3.08342159e-01 3.76291782e-01 -4.62096721e-01 2.18353346e-01
9.17536393e-02 5.86549401e-01 -4.01167125e-01 7.63517439e-01
6.21254563e-01 2.11263113e-02 -2.71492332e-01 -3.24923605e-01
-7.20461011e-02 6.10549524e-02 -1.30688393e+00 1.68627584e+00
4.86591980e-02 4.99522030e-01 2.06425358e-02 -1.14638865e+00
8.13828290e-01 4.62900877e-01 8.45634758e-01 -1.12098634e+00
1.17291234e-01 1.86699152e-01 9.99459922e-02 -5.14402568e-01
2.96592683e-01 2.88367897e-01 1.79289281e-01 1.19343050e-01
6.85188323e-02 7.04380989e-01 1.20273493e-01 2.78513879e-01
1.20061970e+00 1.05511968e-03 5.40057346e-02 -7.68057927e-02
7.44384408e-01 -2.96381235e-01 7.96077132e-01 6.53632224e-01
-6.44009173e-01 6.07620656e-01 5.57779193e-01 -6.52549803e-01
-7.47902393e-01 -1.08169150e+00 8.68540723e-03 7.24710941e-01
4.06043202e-01 -3.78671169e-01 -6.12345517e-01 -9.29345727e-01
-1.40481189e-01 4.84282851e-01 -5.48674643e-01 -5.13342083e-01
-6.72778308e-01 -8.83152306e-01 6.17116570e-01 7.19251037e-01
7.81106651e-01 -8.76633227e-01 -7.66850412e-01 3.37726802e-01
-5.50801754e-01 -1.19725740e+00 -4.51293200e-01 4.28574011e-02
-7.70646751e-01 -1.10846424e+00 -5.01581371e-01 -4.59602505e-01
5.39811790e-01 5.14686525e-01 8.45858395e-01 2.28639930e-01
-1.17959827e-01 3.18390667e-01 -4.64368552e-01 -3.20409685e-01
-4.39575650e-02 -3.09997112e-01 1.38320863e-01 6.16269290e-01
4.23281252e-01 -8.36160481e-01 -7.61638105e-01 4.01792079e-01
-9.50074553e-01 -2.73640230e-02 5.64646840e-01 6.94915652e-01
4.26577717e-01 2.44460180e-01 6.70549214e-01 -3.05530399e-01
3.89818460e-01 -6.57689810e-01 -4.31311905e-01 -3.44405770e-02
-3.72159898e-01 -3.08503628e-01 5.68238974e-01 -5.89477956e-01
-1.03577530e+00 -2.35275850e-02 -1.89341217e-01 -9.44180071e-01
-2.72377282e-01 3.50144535e-01 -4.02618527e-01 3.16277355e-01
4.62721318e-01 3.70379418e-01 9.36644990e-03 -2.69843131e-01
-8.03981796e-02 4.45598125e-01 7.60411263e-01 -2.08304644e-01
7.99721658e-01 6.16566479e-01 -2.37469867e-01 -7.69298434e-01
-5.69736004e-01 -5.18170178e-01 -6.12451136e-01 -4.20381516e-01
1.06268132e+00 -1.03122532e+00 -4.46666569e-01 7.69385397e-01
-1.03625715e+00 -6.43550307e-02 -4.63687122e-01 4.83938068e-01
-1.31760746e-01 5.05821466e-01 -4.75421131e-01 -8.14041853e-01
-2.30631903e-01 -1.01304281e+00 8.85236442e-01 2.70225197e-01
-1.34682208e-01 -6.21488452e-01 -3.13732773e-01 8.10539797e-02
3.58204842e-01 3.88028830e-01 7.21644342e-01 -7.86339104e-01
-6.09491765e-01 -2.97996402e-01 -3.50712955e-01 3.57445836e-01
1.56005561e-01 -5.11083268e-02 -1.10316467e+00 -1.46489039e-01
6.97634295e-02 2.02184916e-01 9.07791555e-01 4.85384971e-01
1.39783001e+00 -3.94939855e-02 -2.10944548e-01 5.68101466e-01
8.94556284e-01 6.05629444e-01 6.52925432e-01 3.68221432e-01
8.07376623e-01 5.03786027e-01 6.17670059e-01 5.08944929e-01
2.63722807e-01 6.40864372e-01 6.47109926e-01 -2.95330107e-01
-1.68603241e-01 -1.67289041e-02 5.78412533e-01 6.83242857e-01
-1.14711322e-01 -2.29593828e-01 -7.63369441e-01 5.93041003e-01
-2.04157829e+00 -1.07546043e+00 -4.53735650e-01 2.18641043e+00
3.95861268e-01 4.90857929e-01 1.70182064e-01 5.66035926e-01
6.69912934e-01 3.05242807e-01 -5.69553018e-01 -1.21914893e-02
-3.60122681e-01 -2.16676697e-01 4.68634702e-02 7.93448687e-02
-1.39313483e+00 5.05565882e-01 5.40464544e+00 8.47074091e-01
-1.21672666e+00 8.57611597e-02 6.20396018e-01 -3.74248505e-01
8.05376470e-02 -4.47864622e-01 -4.21933502e-01 8.18840802e-01
7.96679735e-01 1.60452873e-01 1.49979904e-01 4.99247909e-01
4.00352806e-01 -1.21873036e-01 -1.04093027e+00 1.00841832e+00
-2.96066832e-02 -9.17757332e-01 5.06571047e-02 -1.36255011e-01
2.71420389e-01 -1.41281322e-01 1.02962315e-01 3.17160249e-01
-1.93021744e-01 -7.70606101e-01 8.25186908e-01 4.96026278e-01
4.31701809e-01 -7.38923788e-01 8.13368738e-01 2.36065730e-01
-1.49551582e+00 -3.03041041e-01 7.92256743e-02 1.74942553e-01
2.22068340e-01 7.71997750e-01 -2.27249369e-01 8.03212821e-01
1.13482845e+00 9.10487235e-01 -6.60766959e-01 1.10005653e+00
-2.50298698e-02 6.47994637e-01 -2.62176991e-01 6.24497771e-01
2.81797945e-01 -9.66767222e-02 9.33091938e-01 1.16315770e+00
2.62231708e-01 7.15473592e-02 1.29376158e-01 6.51072860e-01
2.70827025e-01 -1.40738189e-01 -5.63302279e-01 2.09166765e-01
2.67912298e-01 1.10286260e+00 -6.01369083e-01 -3.87745380e-01
-6.23149753e-01 1.04008293e+00 -5.58653884e-02 5.18466115e-01
-1.22134459e+00 -2.48359248e-01 9.08036232e-01 -1.16529400e-02
1.73269659e-01 -1.95117399e-01 -2.13616177e-01 -1.29717088e+00
3.99210304e-01 -9.52694893e-01 6.92232788e-01 -4.52245921e-01
-1.24141943e+00 6.57108128e-01 -7.66851157e-02 -1.62315559e+00
1.08048193e-01 -4.17899847e-01 -7.62846053e-01 7.10588694e-01
-1.50079834e+00 -9.00444269e-01 -6.33901834e-01 1.00384271e+00
7.07308948e-01 -2.20185801e-01 3.66214305e-01 9.24981236e-01
-9.46642935e-01 6.78213298e-01 -3.29115778e-01 3.54317367e-01
6.95325077e-01 -1.10699356e+00 3.32392007e-01 1.49142301e+00
6.87004104e-02 1.81410849e-01 4.53464061e-01 -6.59983635e-01
-1.02137756e+00 -1.29046607e+00 4.06305850e-01 -3.37398797e-01
5.22515357e-01 -2.38803372e-01 -1.41005552e+00 4.69820470e-01
6.36732280e-02 2.74400204e-01 4.76616502e-01 -1.45703480e-01
-3.61098379e-01 -2.06087038e-01 -8.03184271e-01 6.47788286e-01
1.20521700e+00 -3.96396011e-01 -4.95282739e-01 -1.97458297e-01
6.38350368e-01 -4.16880012e-01 -5.71347356e-01 6.85815334e-01
3.82965386e-01 -1.05800009e+00 1.10384464e+00 -5.61652184e-01
4.27868515e-01 -7.55982339e-01 -1.97428063e-01 -1.16077363e+00
-3.87688130e-01 -2.38881573e-01 -5.88376880e-01 1.26127183e+00
1.25015616e-01 -4.62045521e-01 4.06436145e-01 1.76829726e-01
-5.05423129e-01 -8.03198218e-01 -1.06097388e+00 -7.73116767e-01
-4.88238961e-01 -8.45486820e-01 6.65186107e-01 1.05878496e+00
-2.75007129e-01 1.36365399e-01 -5.19092798e-01 6.45640671e-01
4.96590525e-01 -1.89673260e-01 6.04117274e-01 -1.28629541e+00
-1.05525456e-01 -6.37463033e-01 -6.68674052e-01 -9.35842037e-01
-1.84900180e-01 -6.42224610e-01 -2.03264505e-01 -1.16715574e+00
-1.49513528e-01 -1.88353777e-01 -7.83367634e-01 4.74390298e-01
-2.64817417e-01 2.53472835e-01 -6.15466535e-02 2.37068638e-01
-6.84844971e-01 7.61416078e-01 1.00627208e+00 -1.24355674e-01
-3.28449398e-01 -1.37950107e-01 -5.23251474e-01 9.54704046e-01
6.27821088e-01 -2.06461832e-01 -5.18129647e-01 -5.15062034e-01
-2.09452942e-01 -1.25815719e-02 5.60918629e-01 -1.26720238e+00
3.78460109e-01 -4.35884297e-02 7.02259660e-01 -7.87671566e-01
2.93605179e-01 -9.98917401e-01 -2.14414194e-01 2.97333479e-01
4.16993015e-02 1.34248868e-01 3.46992284e-01 8.77151072e-01
-5.83445668e-01 3.79994363e-01 7.51985848e-01 2.14905798e-01
-9.85608757e-01 5.41281641e-01 -4.85941052e-01 -8.91305506e-02
1.24910688e+00 -5.35504341e-01 -2.59742349e-01 -1.89123601e-01
-7.39002943e-01 5.01481116e-01 2.02204497e-03 7.60530055e-01
8.32985938e-01 -1.47844946e+00 -5.91724098e-01 6.46786690e-01
2.42478594e-01 1.84743375e-01 6.86629295e-01 1.34684014e+00
-2.18989149e-01 4.27936018e-02 -3.97085339e-01 -8.83184373e-01
-1.11890173e+00 5.33320427e-01 5.27495027e-01 -1.37165144e-01
-9.71840799e-01 6.28170073e-01 3.70415062e-01 2.40980774e-01
4.39978302e-01 -2.06459075e-01 -4.89930987e-01 2.62532890e-01
7.49719381e-01 3.61886203e-01 1.97658107e-01 -6.88136101e-01
-4.52589184e-01 1.81991339e-01 -2.77992040e-02 1.04905166e-01
1.09663939e+00 -3.06413978e-01 5.40247597e-02 4.59983170e-01
7.93681860e-01 -1.84583470e-01 -1.38979316e+00 -4.56490219e-01
-6.17206432e-02 -5.17801464e-01 1.00376755e-01 -4.93845105e-01
-1.36451638e+00 9.51730549e-01 8.56035352e-01 3.46932381e-01
1.65893734e+00 -3.62666517e-01 9.06760156e-01 -1.36003941e-02
-1.21820338e-01 -9.96460378e-01 3.53792518e-01 3.13473672e-01
6.82777941e-01 -1.17875385e+00 -3.23177129e-01 -2.41215840e-01
-6.46519601e-01 9.51324701e-01 1.01741600e+00 -2.13708486e-02
6.14899099e-01 2.39833832e-01 3.00094541e-02 -1.87919736e-01
-6.99561834e-01 -4.64355238e-02 5.86057603e-01 3.79892439e-01
1.50403619e-01 -3.13951254e-01 -6.77919537e-02 7.23626614e-01
3.89370531e-01 -3.18091661e-01 2.62946427e-01 9.57690001e-01
-3.50472212e-01 -6.45333290e-01 -4.29736763e-01 5.10872900e-01
-5.12010813e-01 5.50716668e-02 -1.14642553e-01 5.75227439e-01
3.84316921e-01 8.94239426e-01 4.50403690e-01 -5.84477961e-01
4.99388874e-01 2.29296654e-01 -7.64664114e-02 -3.85982729e-02
-4.12294209e-01 4.17382985e-01 -2.32682034e-01 -8.65851641e-01
-2.97122419e-01 -6.61212027e-01 -1.20086682e+00 -1.70148790e-01
-1.10961765e-01 7.98706859e-02 2.07267269e-01 1.05851042e+00
5.12987733e-01 1.08482337e+00 5.82378924e-01 -5.87936342e-01
-1.76899686e-01 -7.64431000e-01 -5.43281317e-01 4.87383753e-01
7.05404103e-01 -8.36385012e-01 -3.25231433e-01 2.15625335e-02] | [7.898193359375, 1.5592492818832397] |
5d604986-51d1-43ac-a35c-4fbed1d9a38e | integrating-semantic-scenario-and-word | null | null | https://aclanthology.org/2021.emnlp-main.196 | https://aclanthology.org/2021.emnlp-main.196.pdf | Integrating Semantic Scenario and Word Relations for Abstractive Sentence Summarization | Recently graph-based methods have been adopted for Abstractive Text Summarization. However, existing graph-based methods only consider either word relations or structure information, which neglect the correlation between them. To simultaneously capture the word relations and structure information from sentences, we propose a novel Dual Graph network for Abstractive Sentence Summarization. Specifically, we first construct semantic scenario graph and semantic word relation graph based on FrameNet, and subsequently learn their representations and design graph fusion method to enhance their correlation and obtain better semantic representation for summary generation. Experimental results show our model outperforms existing state-of-the-art methods on two popular benchmark datasets, i.e., Gigaword and DUC 2004. | ['Hu Zhang', 'XiaoLi Li', 'Ru Li', 'Shaoru Guo', 'Yong Guan'] | null | null | null | null | emnlp-2021-11 | ['abstractive-sentence-summarization'] | ['natural-language-processing'] | [ 3.31946820e-01 2.93039858e-01 -3.93456310e-01 -1.89045757e-01
-4.93985653e-01 -3.03957433e-01 5.35721779e-01 6.69014692e-01
-1.81367263e-01 1.02480352e+00 1.27294374e+00 8.59081522e-02
-1.93575144e-01 -9.49909329e-01 -4.46714848e-01 -2.96469003e-01
3.57727140e-01 2.70033479e-01 4.12183046e-01 -5.49592972e-01
4.99247193e-01 -9.37183276e-02 -1.06709731e+00 2.76382089e-01
1.19086647e+00 4.49925840e-01 2.35801205e-01 6.26502097e-01
-6.22960925e-01 1.12722826e+00 -9.90031242e-01 -4.66557533e-01
-4.28781033e-01 -9.03738320e-01 -1.07594621e+00 3.23828608e-01
1.90171912e-01 -9.54450071e-02 -9.56303418e-01 1.32182503e+00
6.97507322e-01 5.80907941e-01 6.55109406e-01 -7.83027291e-01
-1.04433179e+00 1.07549489e+00 -7.34753370e-01 5.12721241e-01
9.22265649e-01 -2.38836184e-01 1.46757555e+00 -3.26818943e-01
7.44205117e-01 1.62772310e+00 4.13890034e-01 3.67716819e-01
-8.11193943e-01 -2.56481022e-01 5.58555841e-01 4.81248796e-01
-9.20452654e-01 -4.50631291e-01 1.12002754e+00 1.94670796e-01
1.30921018e+00 4.87802655e-01 8.18909407e-01 8.73781264e-01
2.28877097e-01 1.09530282e+00 5.02685726e-01 -2.59104937e-01
-1.60689160e-01 -6.38056338e-01 5.25679231e-01 8.90269339e-01
7.18152583e-01 -8.18425715e-01 -5.52436531e-01 -7.56926462e-02
5.00701666e-01 -4.95126843e-02 -6.26460731e-01 7.25095198e-02
-1.18974626e+00 8.15958619e-01 4.53336149e-01 4.08354133e-01
-4.81047273e-01 6.03209622e-02 7.76518822e-01 1.48862302e-01
7.71389961e-01 4.34577823e-01 -9.72657204e-02 1.20667867e-01
-6.89170539e-01 4.01753843e-01 7.62243092e-01 1.07547128e+00
7.04475284e-01 1.10327736e-01 -7.94908464e-01 9.26093102e-01
3.78726840e-01 2.50027895e-01 7.49260068e-01 -5.40280342e-01
9.45496678e-01 1.14107192e+00 -2.85436153e-01 -1.57748103e+00
-4.40795481e-01 -5.15901744e-01 -1.17529213e+00 -1.03862154e+00
-5.22165835e-01 -1.81148037e-01 -8.75756383e-01 1.28729069e+00
1.08928988e-02 3.37626517e-01 5.16281903e-01 7.09055245e-01
1.89692187e+00 9.19450402e-01 -1.09938532e-01 -5.38432956e-01
1.31938982e+00 -1.29351807e+00 -1.17804122e+00 -3.36907864e-01
6.26926422e-01 -5.53788483e-01 7.69660056e-01 -8.17918479e-02
-1.26446199e+00 -4.21890974e-01 -1.27541029e+00 -2.72739351e-01
-8.15216973e-02 -3.28784324e-02 6.38960183e-01 3.43872942e-02
-9.77086544e-01 7.62472510e-01 -5.94791830e-01 -5.54083824e-01
5.57740510e-01 2.04980299e-01 -3.75020385e-01 -1.85378999e-01
-1.42342961e+00 8.53330076e-01 1.02019227e+00 1.72592983e-01
-3.10754508e-01 -2.21252218e-01 -1.19157338e+00 3.00750136e-01
6.04961812e-01 -1.32231927e+00 1.10856187e+00 -5.10280490e-01
-1.35956788e+00 4.57308620e-01 -2.25083455e-01 -5.02135694e-01
-7.18130246e-02 -1.58197477e-01 -4.05137479e-01 4.46086437e-01
1.23539582e-01 2.27768436e-01 3.44020158e-01 -1.17358780e+00
-4.49513286e-01 -3.79523784e-01 3.19693327e-01 8.03574920e-01
-5.55074334e-01 -1.08335607e-01 -4.15212482e-01 -7.51802087e-01
1.63291037e-01 -1.97306082e-01 -3.15463364e-01 -1.00768411e+00
-8.20960999e-01 -6.05222881e-01 8.75959158e-01 -9.15989339e-01
1.75656545e+00 -1.50156617e+00 6.78048491e-01 -2.89260447e-01
5.70584178e-01 4.97760892e-01 -2.94794798e-01 8.10204327e-01
1.47199512e-01 1.06067345e-01 -3.11541706e-01 -3.38115484e-01
-2.00621903e-01 4.28825974e-01 -1.91139191e-01 1.53554499e-01
-1.32945135e-01 1.17458403e+00 -1.32243526e+00 -9.29576755e-01
1.59175619e-01 6.32244200e-02 -3.43542039e-01 3.43442321e-01
-3.62093419e-01 1.53825969e-01 -9.50188220e-01 1.66046903e-01
5.48997521e-01 -1.94885612e-01 2.90619433e-01 -5.51468670e-01
3.97827417e-01 4.01999056e-01 -8.69490385e-01 2.16918135e+00
-2.68276423e-01 2.05950752e-01 -3.59011799e-01 -1.30983162e+00
1.15934038e+00 1.47923723e-01 2.34314293e-01 -3.73195499e-01
3.20339531e-01 -1.30768150e-01 -2.45498002e-01 -6.55049264e-01
9.55143511e-01 8.04156885e-02 -5.37257940e-02 2.45655492e-01
3.45842898e-01 -5.10991573e-01 5.13088226e-01 7.87624121e-01
1.22322810e+00 8.29924792e-02 6.49709582e-01 -3.33680511e-01
1.00492275e+00 -1.15462057e-01 6.25445127e-01 4.63891923e-01
2.12891906e-01 6.27283275e-01 8.71923089e-01 -1.19753480e-01
-5.40575802e-01 -7.79135227e-01 7.85450637e-01 6.29251003e-01
5.85668445e-01 -1.01594174e+00 -7.57875025e-01 -9.12732542e-01
-2.78122246e-01 8.94379079e-01 -3.86053920e-01 -6.62088394e-01
-7.21979439e-01 -7.67084718e-01 4.39868927e-01 4.60837334e-01
9.89077449e-01 -1.00732327e+00 1.79523155e-02 3.37549150e-01
-6.30735576e-01 -1.12148345e+00 -7.82658994e-01 -6.43106580e-01
-9.21827018e-01 -1.02461338e+00 -4.36363459e-01 -9.18565571e-01
5.84699333e-01 5.99272966e-01 1.15872896e+00 4.07723337e-01
1.04707792e-01 4.07335639e-01 -7.07264662e-01 -2.24302337e-01
-2.59486228e-01 3.62807184e-01 -3.14175636e-01 1.20542087e-02
-8.16418789e-03 -6.75120473e-01 -5.87496102e-01 -3.26224566e-01
-9.64278281e-01 2.93403089e-01 5.16536832e-01 8.01136732e-01
4.35769916e-01 -4.49392805e-03 8.92163694e-01 -1.09671974e+00
1.47627962e+00 -4.07676846e-01 5.56451790e-02 5.77633440e-01
-3.84431511e-01 2.10701168e-01 7.98658669e-01 1.68600276e-01
-1.42732298e+00 -6.26039565e-01 -8.15375149e-02 -1.75797671e-01
3.49349767e-01 1.05664325e+00 -3.12975228e-01 2.86472708e-01
3.78391653e-01 4.84861314e-01 -1.72257751e-01 -2.74045616e-01
6.15360081e-01 6.31982565e-01 6.10218048e-01 -5.15139103e-01
3.77646953e-01 1.99142948e-01 1.17746927e-01 -8.71422470e-01
-1.23370421e+00 -4.55708355e-01 -3.88328105e-01 -5.79093099e-02
1.02638507e+00 -9.24608529e-01 -3.35147589e-01 2.91651130e-01
-1.66325617e+00 4.35243130e-01 -2.69789606e-01 3.63905251e-01
-4.72584575e-01 1.02872789e+00 -5.54097831e-01 -3.00578892e-01
-8.94154549e-01 -8.15881133e-01 1.14305925e+00 6.43755198e-01
-1.24687634e-01 -1.35627627e+00 1.77295253e-01 4.31537837e-01
7.75630027e-02 4.39092666e-01 9.20689166e-01 -7.72657216e-01
-2.47255877e-01 -1.98608205e-01 -5.41249096e-01 3.38910729e-01
4.61449325e-01 -2.84708411e-01 -3.14680040e-01 -1.43421277e-01
-1.46781176e-01 -1.27673507e-01 1.45373774e+00 2.67677426e-01
1.15814018e+00 -7.18760908e-01 -4.26397115e-01 2.55297780e-01
1.14881575e+00 -2.01621249e-01 7.37791240e-01 -1.93923548e-01
1.26317406e+00 4.63359237e-01 4.44533139e-01 3.89416516e-01
7.17636645e-01 2.68738329e-01 2.29530737e-01 1.22690268e-01
-4.66971695e-01 -5.88775218e-01 5.76205738e-02 1.61696827e+00
-3.67450535e-01 -8.12242031e-01 -4.59208012e-01 3.88042450e-01
-2.48121285e+00 -8.95252585e-01 -3.51155311e-01 1.47147226e+00
6.50008082e-01 1.13480069e-01 -1.28927946e-01 -1.76517442e-01
1.02689946e+00 9.05999720e-01 -2.19795153e-01 -3.34708720e-01
-4.40873235e-01 1.91933036e-01 3.19782138e-01 5.08792579e-01
-8.01511586e-01 1.31296170e+00 5.67540932e+00 1.14959133e+00
-3.65525663e-01 -4.81581837e-02 3.86307269e-01 2.45276526e-01
-9.90788043e-01 1.11159623e-01 -3.13220412e-01 4.64432724e-02
4.05962139e-01 -8.94491196e-01 1.67722002e-01 3.17511469e-01
9.16757882e-02 3.65880914e-02 -6.06308520e-01 1.10245752e+00
6.81599200e-01 -1.71521807e+00 7.44271100e-01 -4.76300925e-01
9.29693520e-01 -3.59346837e-01 -6.47452831e-01 3.68644565e-01
4.44898278e-01 -8.53363335e-01 2.10719347e-01 7.96567261e-01
2.18388498e-01 -7.72714794e-01 1.00921953e+00 1.83573827e-01
-1.42468834e+00 2.81915128e-01 -5.01128912e-01 -2.15144679e-01
3.00853193e-01 5.63990295e-01 -3.59610468e-01 1.75883925e+00
-4.64165695e-02 1.44380140e+00 -7.05031455e-01 6.54218614e-01
-5.81039727e-01 5.59560061e-01 2.15495929e-01 -4.92755890e-01
3.91541064e-01 -3.73952836e-01 8.71328950e-01 1.21138787e+00
2.45567158e-01 4.43279266e-01 5.49424946e-01 5.65323174e-01
-3.93075734e-01 3.77730399e-01 -6.09940648e-01 -3.32915008e-01
2.91850567e-01 1.29056156e+00 -7.53976226e-01 -6.19673789e-01
-4.40729439e-01 1.12420118e+00 4.91567463e-01 3.82250518e-01
-6.43428147e-01 -7.06988752e-01 1.67398527e-01 -2.72679359e-01
4.99841161e-02 -3.01259130e-01 -5.08157611e-02 -1.70890570e+00
-3.65920993e-03 -6.48386598e-01 6.79607511e-01 -8.47082973e-01
-1.09892905e+00 5.24389625e-01 3.81003767e-01 -7.04553187e-01
-6.60450980e-02 -2.85063256e-02 -1.10655200e+00 6.01405978e-01
-1.38496673e+00 -1.33687222e+00 -3.72125149e-01 4.58680570e-01
8.08696091e-01 -2.15572193e-01 5.98092198e-01 -1.36321425e-01
-7.12344348e-01 1.44203380e-01 -2.36554638e-01 2.63947397e-02
2.63150841e-01 -1.13907623e+00 6.08261764e-01 9.16948676e-01
6.29468933e-02 5.03970027e-01 7.71699607e-01 -9.99610126e-01
-1.48960268e+00 -1.27732944e+00 8.16268504e-01 7.99414441e-02
5.40259719e-01 6.60917535e-02 -8.72463882e-01 7.41044283e-01
8.16606760e-01 -3.18462819e-01 3.49956214e-01 -9.27749947e-02
-6.63799867e-02 1.57072246e-02 -8.11566174e-01 8.36616516e-01
1.63297129e+00 -8.24934915e-02 -1.24132752e+00 6.80735350e-01
1.42022228e+00 -4.07773435e-01 -6.14512920e-01 6.20171249e-01
1.81715637e-02 -7.54526079e-01 8.30086291e-01 -8.57558310e-01
6.63770735e-01 -2.37497851e-01 -6.69338331e-02 -1.87461805e+00
-5.05828738e-01 -6.94337189e-01 -2.58815229e-01 1.55465829e+00
2.24170446e-01 -7.45596826e-01 6.03974819e-01 -2.55745947e-01
-5.21956146e-01 -8.81290257e-01 -5.76255381e-01 -6.03977382e-01
-2.70343304e-01 2.91757822e-01 8.33417952e-01 7.21144140e-01
2.77477115e-01 1.20390940e+00 -2.69085705e-01 -4.13739443e-01
4.11765009e-01 2.99990654e-01 6.39665246e-01 -1.12007844e+00
-4.46162708e-02 -5.77871561e-01 -5.99268913e-01 -1.09395838e+00
7.68809199e-01 -1.14514983e+00 -2.20057026e-01 -2.60414100e+00
4.66894299e-01 4.74298000e-01 -1.50824979e-01 1.11748725e-01
-7.02897012e-01 -3.85322511e-01 4.46242690e-02 -8.37328807e-02
-1.07123542e+00 1.20798755e+00 1.59307039e+00 -4.11537975e-01
-1.25205413e-01 -3.74191880e-01 -1.11024737e+00 7.36930668e-01
7.25142121e-01 -6.06560968e-02 -8.27307761e-01 -6.39468491e-01
1.34800777e-01 3.99613261e-01 1.33124650e-01 -7.50398874e-01
3.79970431e-01 -2.45771572e-01 1.89265292e-02 -6.27235115e-01
2.47372594e-02 -1.19852968e-01 -2.74741977e-01 3.29223365e-01
-4.77452457e-01 1.18556842e-01 -1.37043744e-02 1.01442993e+00
-5.14203072e-01 -3.60105515e-01 2.25870922e-01 -5.17127752e-01
-6.11547112e-01 4.41328585e-01 3.77682634e-02 4.55099314e-01
7.08743751e-01 1.26048341e-01 -7.55502701e-01 -5.68199694e-01
-4.63118643e-01 5.65745711e-01 -4.81811650e-02 5.27554631e-01
9.67703462e-01 -1.41735637e+00 -1.17826295e+00 -3.96716952e-01
-2.48439293e-02 2.71633059e-01 4.78525311e-01 5.57808518e-01
-6.11641407e-01 2.99535215e-01 -9.81497318e-02 -1.02133758e-01
-1.36839056e+00 4.21282947e-01 1.22775184e-02 -6.75773501e-01
-7.45136499e-01 7.25425363e-01 1.58908486e-01 -1.46665365e-01
-2.96762735e-01 -2.95338154e-01 -8.12327683e-01 1.32214785e-01
3.80485058e-01 5.29569626e-01 -2.66797245e-02 -7.52167702e-01
-2.80104250e-01 5.44374287e-01 -2.14670122e-01 7.02470988e-02
1.24101019e+00 -2.65775084e-01 -6.93622828e-01 1.71953335e-01
1.21400201e+00 -9.12231505e-02 -3.25604469e-01 -3.93323541e-01
-6.35968447e-02 -2.90459126e-01 2.37619057e-02 -1.13602966e-01
-1.05537748e+00 5.75643837e-01 -5.62457740e-01 3.26527178e-01
1.11989045e+00 9.80761275e-02 1.29127622e+00 5.82463861e-01
-5.24016060e-02 -9.95680034e-01 3.00949216e-01 4.97920781e-01
1.23487985e+00 -7.37060130e-01 6.69873893e-01 -7.85334766e-01
-6.64345264e-01 1.19119108e+00 7.04712033e-01 -5.06915271e-01
2.42974311e-01 -5.21695256e-01 -6.21992528e-01 -3.82887870e-01
-7.65429795e-01 -3.19335878e-01 5.11337340e-01 4.47009802e-01
4.19067323e-01 1.20751217e-01 -9.32690203e-01 8.71866167e-01
-1.45148724e-01 -2.00453192e-01 8.32961738e-01 7.25719333e-01
-7.88801491e-01 -9.05470908e-01 1.55922174e-02 7.67244399e-01
-2.79830217e-01 -2.17605188e-01 -7.23121464e-01 3.39671612e-01
-2.98243046e-01 1.28811646e+00 -1.94238335e-01 -5.30662000e-01
6.94484293e-01 -3.46218109e-01 7.70366371e-01 -9.60185707e-01
-5.24223626e-01 -1.18591473e-01 6.58307552e-01 -2.28241727e-01
-7.49038160e-01 -4.91933614e-01 -1.42632222e+00 -2.65861452e-01
-4.29059505e-01 4.81324345e-01 8.88491645e-02 1.00481474e+00
3.97166461e-01 1.18709886e+00 3.90434116e-01 -5.83081067e-01
-3.56620640e-01 -1.26979387e+00 -5.54471016e-01 5.46134651e-01
-2.39277780e-02 -4.79409367e-01 -2.00901106e-01 -2.25912169e-01] | [12.609848022460938, 9.556604385375977] |
1b803022-58ab-48c9-8e5a-ceb63fbe6d8c | on-feature-diversity-in-energy-based-models-1 | 2306.01489 | null | https://arxiv.org/abs/2306.01489v1 | https://arxiv.org/pdf/2306.01489v1.pdf | On Feature Diversity in Energy-based Models | Energy-based learning is a powerful learning paradigm that encapsulates various discriminative and generative approaches. An energy-based model (EBM) is typically formed of inner-model(s) that learn a combination of the different features to generate an energy mapping for each input configuration. In this paper, we focus on the diversity of the produced feature set. We extend the probably approximately correct (PAC) theory of EBMs and analyze the effect of redundancy reduction on the performance of EBMs. We derive generalization bounds for various learning contexts, i.e., regression, classification, and implicit regression, with different energy functions and we show that indeed reducing redundancy of the feature set can consistently decrease the gap between the true and empirical expectation of the energy and boosts the performance of the model. | ['Moncef Gabbouj', 'Alexandros Iosifidis', 'Jenni Raitoharju', 'Firas Laakom'] | 2023-06-02 | on-feature-diversity-in-energy-based-models | https://openreview.net/forum?id=ks3Q08yy66r | https://openreview.net/pdf?id=ks3Q08yy66r | iclr-workshop-ebm-2021-5 | ['generalization-bounds'] | ['methodology'] | [ 1.77506760e-01 -9.97998640e-02 -9.33450162e-02 -2.84396857e-01
-8.49103749e-01 -5.08168519e-01 5.70511222e-01 2.91753411e-02
-1.70941412e-01 7.02927232e-01 -1.00590838e-02 -3.35352011e-02
-3.68412256e-01 -8.14464569e-01 -1.08234310e+00 -1.11980486e+00
1.10072151e-01 2.44890228e-01 1.17691174e-01 -6.03496134e-02
3.28126729e-01 6.01343393e-01 -1.67620349e+00 -2.23252848e-02
8.73501956e-01 9.09507871e-01 1.55974135e-01 6.83791220e-01
1.34487420e-01 2.09023729e-01 -6.38844073e-01 -2.15399042e-01
2.41665527e-01 -7.09300697e-01 -4.49033618e-01 -3.41178268e-01
8.68983939e-02 3.00563395e-01 -3.85691315e-01 8.62400293e-01
3.86627525e-01 5.10672808e-01 1.15974653e+00 -1.30511236e+00
-5.30861795e-01 4.21497554e-01 -3.50424916e-01 1.48163512e-01
-9.84586682e-03 -6.27988949e-02 9.99864519e-01 -9.02690172e-01
3.53104323e-01 8.79478216e-01 6.19937420e-01 7.32124984e-01
-1.44760096e+00 -5.46036959e-01 -1.87448800e-01 1.30665407e-01
-1.43821168e+00 -3.82682204e-01 9.73913729e-01 -3.78041595e-01
9.39093053e-01 4.72682655e-01 7.08634615e-01 9.94125247e-01
5.14005840e-01 6.42588913e-01 1.02266645e+00 -9.00264502e-01
4.59596157e-01 3.04696292e-01 1.15442000e-01 8.33373368e-01
3.63480091e-01 4.97124344e-01 -8.39097321e-01 -4.21313822e-01
5.70524693e-01 -1.11791864e-01 -2.72808045e-01 -7.83134639e-01
-4.12355095e-01 9.65852201e-01 4.29508179e-01 2.27380186e-01
-1.35540992e-01 4.54368770e-01 3.60744931e-02 4.56936471e-02
4.17744666e-01 5.53463936e-01 -3.74686241e-01 5.61835319e-02
-8.75606239e-01 2.16792271e-01 7.60770380e-01 5.68768382e-01
9.64359105e-01 7.46630132e-03 -4.63432260e-02 8.47473145e-01
2.39265829e-01 3.32516342e-01 5.17004609e-01 -5.79818845e-01
9.97614190e-02 5.81856012e-01 -1.37503207e-01 -6.52325630e-01
-3.44064146e-01 -7.10556805e-01 -7.17063785e-01 2.26920366e-01
8.54641721e-02 -2.66567171e-01 -9.20408249e-01 2.00122142e+00
1.50253639e-01 1.35219574e-01 -1.52195305e-01 5.58020651e-01
3.95662934e-01 6.76463306e-01 8.34104195e-02 -3.90107900e-01
3.90082061e-01 -8.06241810e-01 -2.20400468e-01 -2.12273508e-01
4.76861447e-01 -2.83180177e-01 8.39692414e-01 1.72942534e-01
-1.11731887e+00 -5.51497638e-01 -1.18593705e+00 2.11893052e-01
-4.23793435e-01 -8.26309919e-02 3.98178041e-01 6.45045459e-01
-1.00211477e+00 1.06391966e+00 -8.88632596e-01 1.22710392e-01
2.20877811e-01 4.33550090e-01 -1.43824324e-01 3.83473113e-02
-7.79901266e-01 9.58817184e-01 3.62041652e-01 -1.54141545e-01
-7.21054852e-01 -7.30606437e-01 -7.42006242e-01 1.37444690e-01
1.10018894e-01 -7.79594541e-01 8.21430683e-01 -1.15021586e+00
-1.10016751e+00 4.28467035e-01 -3.11030775e-01 -3.02035987e-01
2.91726440e-01 -2.09135279e-01 -7.31657967e-02 -4.45097923e-01
-3.29000562e-01 7.64541864e-01 1.11469865e+00 -1.25807905e+00
-3.33954781e-01 -2.68253207e-01 -4.33184922e-01 1.61910892e-01
-4.30823229e-02 -5.04076958e-01 -3.44761051e-02 -5.21637917e-01
5.06060123e-02 -1.10056353e+00 -2.98388749e-01 -3.45541298e-01
-3.28011513e-01 -2.58903325e-01 5.40013313e-01 -4.30577874e-01
1.21855688e+00 -2.27130079e+00 6.26894116e-01 5.47630787e-01
-1.29353866e-01 -1.35428354e-01 -3.24216671e-02 4.50144351e-01
-9.00976881e-02 1.41027600e-01 -4.55102891e-01 -2.09181175e-01
6.82459548e-02 4.48863924e-01 -2.38697648e-01 4.15454268e-01
3.24430287e-01 1.03826928e+00 -6.63529575e-01 -2.23473415e-01
1.38376251e-01 7.21166790e-01 -5.75445950e-01 2.79767632e-01
-8.97200853e-02 3.47414255e-01 -2.77366668e-01 2.09079191e-01
3.64785433e-01 -1.87430099e-01 1.44085184e-01 -1.82392180e-01
1.47197083e-01 1.72701806e-01 -9.22154844e-01 1.42559373e+00
-7.65544891e-01 6.28679574e-01 -4.13703144e-01 -1.02218282e+00
7.98518121e-01 -8.24827626e-02 3.59998584e-01 -4.79867369e-01
-5.08534461e-02 2.52376914e-01 1.13470860e-01 -3.78458574e-02
1.82484731e-01 -1.52659342e-01 -1.75304875e-01 3.96782637e-01
3.39809358e-01 6.03375360e-02 -2.37708036e-02 -2.60659188e-01
8.32098603e-01 3.66887420e-01 5.39161742e-01 -3.38082910e-01
3.11393350e-01 -3.09081554e-01 4.14625019e-01 8.34871292e-01
2.99962133e-01 4.21490461e-01 3.04463923e-01 -2.88004458e-01
-1.14562690e+00 -1.10273755e+00 -3.43754828e-01 1.19135034e+00
2.13049829e-01 -4.07926828e-01 -7.12699234e-01 -8.20964336e-01
9.82193947e-02 1.05270290e+00 -7.35682070e-01 -8.64916265e-01
-5.30918658e-01 -1.11226666e+00 2.08320886e-01 6.53059721e-01
2.19777226e-01 -7.83350885e-01 -9.91087496e-01 1.18604556e-01
6.31560683e-02 -5.25677741e-01 -5.49252093e-01 7.30066121e-01
-9.64976847e-01 -1.01420081e+00 -1.17405131e-01 -5.86866438e-01
6.62072241e-01 -2.91467726e-01 1.18016315e+00 8.62275437e-02
-4.03503358e-01 3.91563177e-01 -2.92169005e-01 -1.72794491e-01
-6.21481001e-01 3.75629514e-02 -8.92278329e-02 -1.29239380e-01
1.96927749e-02 -6.46187246e-01 -4.67154354e-01 1.06964588e-01
-7.12094307e-01 1.92984015e-01 7.34915733e-01 1.03894734e+00
8.68848801e-01 1.25722781e-01 5.46547413e-01 -5.54385781e-01
5.19380629e-01 -5.73769152e-01 -4.23655182e-01 5.17896175e-01
-9.88923728e-01 8.02791119e-01 5.52258790e-01 -5.45261323e-01
-7.81771958e-01 1.86873525e-01 -1.34188429e-01 -3.26697320e-01
1.15953289e-01 2.00007677e-01 -6.53193220e-02 -2.99107999e-01
6.67274535e-01 5.70997775e-01 -4.40098614e-01 -5.14076114e-01
3.94013971e-01 2.54774928e-01 3.82970333e-01 -7.59551048e-01
7.52623856e-01 -4.67728712e-02 5.57005167e-01 -5.38615465e-01
-6.69258952e-01 -2.98129749e-02 -4.76214230e-01 -3.20427865e-01
4.97783154e-01 -2.56603301e-01 -2.86400944e-01 1.32305130e-01
-7.76567101e-01 -3.57568800e-01 -5.63808024e-01 2.61367798e-01
-6.72592878e-01 -1.14296958e-01 -7.49266893e-02 -9.85325634e-01
-4.38345999e-01 -1.04448509e+00 9.76452887e-01 4.15560395e-01
-5.09167425e-02 -1.04033172e+00 4.02963847e-01 -3.91802728e-01
3.32188725e-01 4.95491713e-01 1.28855383e+00 -6.71875119e-01
-3.64492238e-01 -1.44835725e-01 2.74024785e-01 3.90088916e-01
-3.02804913e-02 3.08001563e-02 -9.42285478e-01 -2.31822222e-01
-6.28197044e-02 -2.26808369e-01 1.12336397e+00 5.61476052e-01
1.18658173e+00 -4.99930531e-01 -5.73702157e-01 6.58236265e-01
1.67751157e+00 2.94887424e-01 5.42391717e-01 3.31507698e-02
4.79933679e-01 3.50644052e-01 1.91907734e-01 3.34862888e-01
-1.99978113e-01 7.95815945e-01 1.65939793e-01 1.09760739e-01
1.76656414e-02 -4.57811743e-01 3.90631437e-01 5.66632748e-01
-5.41708954e-02 -2.48949543e-01 -5.77273309e-01 2.11454093e-01
-1.82366908e+00 -8.57985973e-01 3.52859497e-01 2.43998122e+00
8.95691097e-01 3.23580913e-02 7.91180730e-02 6.28895462e-02
4.81309474e-01 -1.30252272e-01 -8.32403243e-01 -5.82628250e-01
-2.75269691e-02 7.37329483e-01 4.71336037e-01 5.17750502e-01
-8.25927138e-01 3.43698382e-01 7.41731310e+00 9.02853131e-01
-9.08612132e-01 1.62089586e-01 6.05123043e-01 -1.45797014e-01
-4.40721959e-01 -1.34407386e-01 -6.67538345e-01 6.05926394e-01
1.10787547e+00 -3.57766300e-01 6.59298122e-01 1.01691747e+00
-3.61941487e-01 -9.09715295e-02 -1.44963777e+00 9.09611344e-01
3.25307548e-02 -1.33134699e+00 -6.68988898e-02 2.01411679e-01
7.70361841e-01 5.05512841e-02 1.32126287e-01 3.80877525e-01
8.70394334e-02 -1.19615293e+00 6.75110638e-01 7.59799123e-01
7.43167639e-01 -1.05586755e+00 6.14573419e-01 5.57092905e-01
-1.11101997e+00 -1.87414795e-01 -2.81616092e-01 3.54125798e-01
-2.41826490e-01 4.89015400e-01 -4.21923071e-01 4.24003661e-01
5.39830685e-01 3.33171755e-01 -7.12349772e-01 9.77890968e-01
-8.81370530e-03 5.55262327e-01 -4.82280284e-01 -1.59053028e-01
-7.83930570e-02 -3.82777601e-01 5.23678064e-01 1.18027413e+00
3.38007838e-01 -1.12928964e-01 -5.29619418e-02 9.15730596e-01
-1.54504895e-01 -6.92601204e-02 -4.68866169e-01 1.77194148e-01
6.23399436e-01 1.05728650e+00 -4.32751745e-01 2.78672576e-02
-1.62725791e-01 8.92099440e-01 5.27730107e-01 2.14946717e-01
-9.07914162e-01 -8.50518197e-02 4.69846666e-01 -5.16743995e-02
3.78890932e-01 -4.34738547e-02 -1.80969387e-01 -7.54976392e-01
-5.97371254e-03 -4.74543959e-01 3.46014321e-01 -5.50478876e-01
-1.18707395e+00 3.54410559e-01 2.96537519e-01 -8.24704170e-01
-5.94241977e-01 -4.59452689e-01 -6.32540464e-01 7.71206856e-01
-1.06408179e+00 -6.90620542e-01 -1.57657355e-01 2.25086853e-01
4.83890951e-01 -4.07671407e-02 8.04213583e-01 -6.56818449e-02
-6.21853888e-01 8.35053086e-01 4.05542195e-01 -2.05608279e-01
2.02465579e-01 -1.33998847e+00 1.70003608e-01 5.51812232e-01
4.94601935e-01 4.86419111e-01 8.00324082e-01 -3.37052047e-01
-1.49668324e+00 -1.05996072e+00 4.63465840e-01 -6.01551294e-01
1.91625774e-01 -1.95179448e-01 -8.61962199e-01 4.77887094e-01
-9.54872146e-02 1.81350131e-02 7.43347526e-01 -8.06897730e-02
-3.24169576e-01 -4.73445542e-02 -1.23839033e+00 3.82488251e-01
1.14716733e+00 -3.49833071e-01 -3.62543732e-01 5.83187677e-02
4.28107798e-01 -3.63282800e-01 -6.90929353e-01 5.87036133e-01
6.73309147e-01 -9.98617887e-01 8.78717840e-01 -7.53109694e-01
2.59490699e-01 -1.06384438e-02 -4.06479686e-01 -1.50505722e+00
-3.89561385e-01 -3.52665514e-01 -7.03721106e-01 8.87625515e-01
5.08539975e-01 -6.22659624e-01 5.28019845e-01 4.10612762e-01
-1.57293767e-01 -1.35685933e+00 -1.17960632e+00 -9.62998509e-01
2.11606249e-01 -2.80312002e-01 4.74501699e-01 3.55219066e-01
-1.65681511e-01 4.37292188e-01 -1.44339651e-01 -7.39040673e-02
5.38734674e-01 2.32000262e-01 2.96880841e-01 -1.22282290e+00
-6.77237868e-01 -6.38259172e-01 -4.45172012e-01 -6.86222494e-01
2.85894096e-01 -9.79592979e-01 1.52114416e-02 -1.22713411e+00
4.91684020e-01 -5.16631484e-01 -3.95408869e-01 3.15281510e-01
-2.10364625e-01 3.48307155e-02 -7.38602504e-02 2.26118609e-01
-1.91303492e-01 5.51940978e-01 6.78090274e-01 1.46053240e-01
-3.70614141e-01 -9.69136308e-04 -6.22850597e-01 5.31528115e-01
8.36800814e-01 -8.13259840e-01 -4.91086274e-01 5.98681532e-02
4.77320313e-01 -2.79522657e-01 4.54191834e-01 -9.28576767e-01
-3.49948332e-02 -2.11219192e-01 8.15341949e-01 -2.29160041e-01
1.99072376e-01 -6.58621252e-01 4.65238750e-01 5.39588273e-01
-6.01294637e-01 2.25027371e-02 8.09113383e-02 7.30540633e-01
1.73822552e-01 -5.51795185e-01 9.35781300e-01 1.95116997e-01
-2.71967381e-01 8.95257443e-02 1.41226143e-01 1.01516552e-01
8.95264506e-01 -4.14910838e-02 -8.31085965e-02 -1.20134890e-01
-4.99316812e-01 -1.59872323e-01 4.52323914e-01 2.41932273e-01
5.64780593e-01 -1.36675429e+00 -5.83789766e-01 3.28863621e-01
-1.01542585e-01 -2.49034181e-01 -1.27372071e-01 5.98981738e-01
4.47478332e-02 1.87492713e-01 6.06323220e-02 -4.41162050e-01
-9.20916200e-01 3.69934320e-01 8.08209479e-01 -2.82341093e-01
-4.83952612e-01 6.99978828e-01 1.01843864e-01 -4.18300554e-02
5.79444952e-02 2.60659140e-02 1.65811792e-01 -1.48306295e-01
1.06076293e-01 4.79024053e-01 8.97317231e-02 -3.93572837e-01
-5.64771533e-01 6.89813852e-01 2.35899672e-01 -2.40281448e-01
1.25288153e+00 3.76390964e-01 2.18180969e-01 4.86996680e-01
1.30374181e+00 3.68433297e-02 -1.37820601e+00 5.85411489e-02
1.06110526e-02 -2.80635774e-01 1.18186541e-01 -7.26606011e-01
-9.87043858e-01 6.06790006e-01 8.35193574e-01 2.07793042e-01
1.24211729e+00 4.39233452e-01 4.20919269e-01 2.64931440e-01
3.36215794e-01 -1.13952756e+00 6.33163974e-02 8.15618634e-02
8.97008359e-01 -9.83318985e-01 7.56186992e-02 -1.07068047e-01
-3.78718674e-01 1.12838876e+00 5.50945163e-01 -3.39446545e-01
7.41434753e-01 2.54473209e-01 -6.88035190e-01 -2.41485611e-02
-8.80450189e-01 -1.15746960e-01 7.98495710e-01 5.76202035e-01
3.23466361e-01 1.58805385e-01 -3.04825157e-01 7.73332894e-01
-2.59832859e-01 -4.88818467e-01 2.31210534e-02 8.70100498e-01
-5.00394940e-01 -1.04999518e+00 -8.91778432e-03 4.83785570e-01
-1.71798784e-02 3.58753912e-02 -4.73530561e-01 8.14981222e-01
2.61809021e-01 6.94729507e-01 1.92726403e-01 -5.83587766e-01
1.56245470e-01 5.58416009e-01 1.03425014e+00 -4.87636358e-01
-3.62544656e-01 -1.48056239e-01 -1.51207253e-01 -3.19444329e-01
-1.15211874e-01 -6.22088253e-01 -1.16607761e+00 -5.13362661e-02
-6.28353298e-01 1.93356499e-01 7.67441094e-01 1.13580704e+00
3.32046747e-01 4.28467661e-01 9.43287611e-01 -7.00007141e-01
-6.64155781e-01 -7.20697761e-01 -3.01094860e-01 1.01876155e-01
4.30146664e-01 -8.46734405e-01 -6.22851253e-01 -2.33448505e-01] | [7.29017448425293, 3.8291916847229004] |
4ddb6fb0-799a-4cf5-ad1f-89358010f6bb | gazeformer-scalable-effective-and-fast | 2303.15274 | null | https://arxiv.org/abs/2303.15274v3 | https://arxiv.org/pdf/2303.15274v3.pdf | Gazeformer: Scalable, Effective and Fast Prediction of Goal-Directed Human Attention | Predicting human gaze is important in Human-Computer Interaction (HCI). However, to practically serve HCI applications, gaze prediction models must be scalable, fast, and accurate in their spatial and temporal gaze predictions. Recent scanpath prediction models focus on goal-directed attention (search). Such models are limited in their application due to a common approach relying on trained target detectors for all possible objects, and the availability of human gaze data for their training (both not scalable). In response, we pose a new task called ZeroGaze, a new variant of zero-shot learning where gaze is predicted for never-before-searched objects, and we develop a novel model, Gazeformer, to solve the ZeroGaze problem. In contrast to existing methods using object detector modules, Gazeformer encodes the target using a natural language model, thus leveraging semantic similarities in scanpath prediction. We use a transformer-based encoder-decoder architecture because transformers are particularly useful for generating contextual representations. Gazeformer surpasses other models by a large margin on the ZeroGaze setting. It also outperforms existing target-detection models on standard gaze prediction for both target-present and target-absent search tasks. In addition to its improved performance, Gazeformer is more than five times faster than the state-of-the-art target-present visual search model. | ['Minh Hoai', 'Gregory Zelinsky', 'Dimitris Samaras', 'Seoyoung Ahn', 'Zhibo Yang', 'Sounak Mondal'] | 2023-03-27 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Mondal_Gazeformer_Scalable_Effective_and_Fast_Prediction_of_Goal-Directed_Human_Attention_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Mondal_Gazeformer_Scalable_Effective_and_Fast_Prediction_of_Goal-Directed_Human_Attention_CVPR_2023_paper.pdf | cvpr-2023-1 | ['scanpath-prediction', 'eye-tracking'] | ['computer-vision', 'computer-vision'] | [ 5.26250482e-01 1.62197500e-01 -5.31680703e-01 -3.30390334e-01
-5.37713170e-01 -4.48959544e-02 4.51977164e-01 -2.14120790e-01
-1.50043562e-01 4.21276122e-01 3.02520156e-01 -2.69487083e-01
-9.60596055e-02 -2.20288724e-01 -5.53622782e-01 -3.00174713e-01
2.32117891e-01 2.83116758e-01 5.22182703e-01 -4.52070445e-01
8.52831721e-01 -6.15096986e-02 -2.34791255e+00 3.34985375e-01
6.76748753e-01 1.06014729e+00 6.65177822e-01 8.51626873e-01
-1.46642243e-02 7.20006645e-01 -2.39142612e-01 -6.76701590e-02
-1.84271246e-01 -7.46180117e-01 -9.53745961e-01 -5.77792823e-01
8.55461776e-01 -1.92832232e-01 8.87468904e-02 7.67611444e-01
5.86897433e-01 3.28603595e-01 6.54385269e-01 -1.67926550e+00
-1.21861041e+00 1.07153639e-01 -8.53094935e-01 5.60470164e-01
9.27060664e-01 3.76374990e-01 1.42296803e+00 -1.20379376e+00
5.34715354e-01 1.30455053e+00 4.76630956e-01 1.19898796e+00
-1.24527144e+00 -7.99215257e-01 4.03461605e-01 7.17188060e-01
-1.30470645e+00 -8.76503468e-01 5.49966276e-01 -3.50119710e-01
1.45126116e+00 5.67214429e-01 4.60960120e-01 1.50199687e+00
1.72999874e-01 1.02911377e+00 7.05481827e-01 -8.90448868e-01
4.71436186e-03 -4.39267568e-02 1.51428267e-01 5.33617675e-01
-1.86202228e-01 3.46992910e-01 -1.30072606e+00 2.15355054e-01
4.50802743e-01 1.89035580e-01 -6.83773279e-01 -3.89306396e-01
-1.27756429e+00 5.01121521e-01 8.46993625e-01 1.37439340e-01
-3.72116655e-01 3.19072723e-01 -1.42179830e-02 4.01628986e-02
4.45972681e-01 4.81201231e-01 -2.19745219e-01 -3.34536999e-01
-9.87299740e-01 3.17953646e-01 5.80099821e-01 1.35315657e+00
7.37455249e-01 -2.20987752e-01 -7.80246556e-01 5.63164234e-01
7.51830518e-01 7.68322885e-01 4.91642565e-01 -6.01542354e-01
3.79125059e-01 4.96299088e-01 3.20917994e-01 -8.03259492e-01
-4.14350092e-01 7.24590123e-02 -2.29032457e-01 3.43023390e-01
1.88178509e-01 3.40860575e-01 -1.14770806e+00 1.81541657e+00
1.00713283e-01 3.41800094e-01 -3.90399396e-01 1.23148096e+00
8.53695393e-01 5.94669521e-01 5.44360757e-01 -2.42776901e-01
1.45829105e+00 -9.86425638e-01 -1.04456079e+00 -5.92511296e-01
6.37070537e-01 -5.30051708e-01 1.76331556e+00 3.04550439e-01
-1.01455355e+00 -5.27915835e-01 -8.45078170e-01 -5.71321666e-01
-3.22349995e-01 -2.43825123e-01 3.87839615e-01 5.96571743e-01
-1.45081425e+00 2.22852096e-01 -5.01987517e-01 -7.87773013e-01
5.19338191e-01 4.65845138e-01 -1.54396771e-02 1.74675100e-02
-9.17097926e-01 1.13782394e+00 -9.53914002e-02 -2.14405581e-02
-5.26534498e-01 -8.65032434e-01 -9.87113714e-01 4.82941806e-01
4.49316710e-01 -6.77959263e-01 1.65992510e+00 -1.04729414e+00
-1.33436918e+00 7.98939705e-01 -1.13693237e+00 -5.91228865e-02
-3.61580878e-01 -3.88806790e-01 -5.71411014e-01 -2.25009307e-01
1.15614362e-01 1.11442947e+00 1.15998638e+00 -9.75555778e-01
-9.43911195e-01 -3.55343074e-01 -1.28342494e-01 3.48648340e-01
-3.86920720e-01 2.95680583e-01 -5.02989054e-01 -1.29176393e-01
-3.24496567e-01 -7.60591030e-01 2.74356484e-01 3.35369825e-01
-3.48109454e-01 -8.20875466e-01 1.09114981e+00 -1.92151278e-01
1.75955105e+00 -2.16965532e+00 8.09699371e-02 -1.02268010e-01
5.60775697e-01 1.73863366e-01 -2.88281322e-01 3.90662283e-01
-1.82544574e-01 7.96353966e-02 3.05325806e-01 -5.97950161e-01
6.70722350e-02 -4.21017408e-01 -4.77273196e-01 5.46820043e-03
3.33041996e-01 1.36905301e+00 -1.11613631e+00 -5.01898229e-01
9.87227336e-02 3.96506310e-01 -5.99895537e-01 2.67604917e-01
-3.96111429e-01 4.83198613e-02 -1.44131541e-01 6.74419880e-01
2.28999630e-01 -7.13891029e-01 -2.87116230e-01 -1.43427951e-02
-3.18151951e-01 4.45550770e-01 -4.61668074e-01 1.83106983e+00
-4.30176526e-01 1.04154396e+00 -4.93422061e-01 -1.12750433e-01
7.06418574e-01 1.78754210e-01 6.80518299e-02 -1.35717726e+00
1.16508327e-01 -3.55710760e-02 7.59984031e-02 -6.89241409e-01
6.81828141e-01 2.14824244e-01 1.63273156e-01 8.13501358e-01
1.68405712e-01 5.22551954e-01 4.30109613e-02 2.05870688e-01
9.75607812e-01 4.28849459e-01 3.06877673e-01 -1.51026905e-01
1.49744973e-01 -6.65354878e-02 1.10277720e-01 5.75761437e-01
-4.58446532e-01 8.26824367e-01 3.86932075e-01 -2.51004249e-01
-6.41183197e-01 -7.16462910e-01 3.24937165e-01 2.12356925e+00
6.05447114e-01 -6.99058235e-01 -7.08371222e-01 -5.59575379e-01
-3.28757644e-01 1.20881879e+00 -1.02354586e+00 -3.56790483e-01
-3.59133452e-01 -9.80404094e-02 3.78993005e-02 5.98366201e-01
-6.12314157e-02 -1.63046122e+00 -1.24578786e+00 -2.51272917e-01
-2.16606960e-01 -4.03877318e-01 -9.64373827e-01 -3.54030393e-02
-3.61952782e-01 -1.32429338e+00 -8.30576479e-01 -7.53015220e-01
4.35592383e-01 8.10433447e-01 1.25074434e+00 1.50476918e-01
-2.36805171e-01 5.34736097e-01 -3.65286082e-01 -9.81561065e-01
3.25835109e-01 2.62004524e-01 3.73474732e-02 -1.13692671e-01
1.25895298e+00 4.02274600e-04 -9.20433998e-01 2.74610400e-01
-2.60183126e-01 1.38569534e-01 4.81710732e-01 8.21122825e-01
1.99001983e-01 -9.66561079e-01 3.63449872e-01 -5.76202214e-01
8.88520718e-01 -6.71050012e-01 -3.04820269e-01 5.51493049e-01
-1.03363597e+00 7.33627155e-02 -2.06292346e-01 -5.42182922e-01
-1.18499446e+00 -2.42394641e-01 1.53741047e-01 -7.74767399e-01
-1.05204053e-01 2.22007647e-01 1.95634961e-01 -3.71146314e-02
1.15580606e+00 2.89145708e-02 8.53111297e-02 -3.09376240e-01
3.26711327e-01 6.38732195e-01 1.80526704e-01 2.12709233e-01
2.66650766e-01 -1.57983676e-02 -7.07204938e-02 -5.47487736e-01
-8.02072406e-01 -6.82622790e-01 -4.65040982e-01 -3.78818959e-01
8.73118103e-01 -6.19838595e-01 -1.37161946e+00 7.30914548e-02
-1.32173240e+00 -4.74693835e-01 -1.46727234e-01 1.80048004e-01
-5.39014757e-01 -2.56738067e-01 -1.27650630e-02 -1.08720195e+00
-5.66626072e-01 -8.83644581e-01 1.50577319e+00 4.34251398e-01
-7.62168348e-01 -7.81961501e-01 3.92546803e-02 -3.43986213e-01
6.58172369e-01 -4.09433395e-01 7.00171649e-01 -4.52761889e-01
-6.71937943e-01 2.20890418e-01 -6.07912719e-01 -5.58165669e-01
8.10928866e-02 -2.05174401e-01 -1.16831768e+00 -1.31716341e-01
-2.86636084e-01 -4.20345217e-01 7.99377263e-01 5.38350642e-01
9.00091648e-01 -1.68029845e-01 -1.04068482e+00 4.62093800e-01
1.10681105e+00 1.81217432e-01 6.68420732e-01 2.45983705e-01
8.98072183e-01 7.59791017e-01 8.83471310e-01 2.28937581e-01
7.30268240e-01 9.73670423e-01 7.60623634e-01 9.37137529e-02
-3.89068276e-01 -3.26377094e-01 1.74460903e-01 -1.75479427e-01
-1.68608263e-01 -6.66426182e-01 -1.18202734e+00 6.22257471e-01
-2.00276113e+00 -1.26738012e+00 -4.59163129e-01 2.14485526e+00
6.48446679e-01 -1.01232678e-01 2.31995583e-01 -1.16381444e-01
4.74162370e-01 7.40122497e-02 -7.71666467e-01 -4.86591846e-01
3.94448429e-01 3.34631264e-01 6.44682273e-02 3.43013406e-01
-8.43890786e-01 1.10357249e+00 6.92606974e+00 4.74958360e-01
-1.25829196e+00 3.33557457e-01 5.41014448e-02 -5.99146008e-01
-2.36736223e-01 -1.88908815e-01 -1.14644265e+00 6.36905670e-01
1.11185646e+00 1.64769553e-02 3.59196931e-01 9.18526411e-01
3.86002026e-02 -3.13939005e-01 -1.49345601e+00 1.45671415e+00
6.46066904e-01 -1.20750558e+00 -2.26399913e-01 3.42744552e-02
1.23711534e-01 5.10694198e-02 4.93184209e-01 4.44415003e-01
-6.49518296e-02 -1.27346933e+00 8.21224630e-01 8.33834708e-01
1.35984409e+00 -3.18786114e-01 2.34361216e-01 3.77753198e-01
-9.40162599e-01 -4.15291339e-01 -1.00135870e-01 -2.56424636e-01
3.04618776e-01 -4.35267895e-01 -9.00417089e-01 -3.02240193e-01
1.03942513e+00 7.73708582e-01 -7.81044245e-01 1.25516140e+00
-2.21814930e-01 4.66070414e-01 -4.79834452e-02 -4.59131122e-01
-2.73107719e-02 5.98591149e-01 3.87878239e-01 8.76956761e-01
5.74633479e-01 1.73373803e-01 -4.12289470e-01 1.05418158e+00
1.43112689e-01 -1.62356943e-01 -5.87305188e-01 3.59249890e-01
7.54035413e-01 7.60904193e-01 -1.75743356e-01 -1.38756046e-02
-5.18918693e-01 1.07068956e+00 4.23647672e-01 5.98254323e-01
-6.68387234e-01 -6.23279214e-01 7.26554036e-01 4.46393669e-01
4.13195670e-01 3.02659363e-01 -3.58355641e-01 -6.86373651e-01
-6.30530715e-02 -5.75360000e-01 2.25882396e-01 -1.43437135e+00
-1.16803336e+00 8.29414368e-01 5.38821556e-02 -1.40396118e+00
-7.38781214e-01 -6.03040993e-01 -5.99991560e-01 1.43530869e+00
-1.56826603e+00 -1.55814362e+00 -6.60210073e-01 8.84557545e-01
8.24648142e-01 -6.42910376e-02 1.15521061e+00 -2.03062475e-01
-4.64758396e-01 7.51459181e-01 -5.73931754e-01 -6.46472752e-01
1.08529556e+00 -1.22376752e+00 6.92582905e-01 7.48398542e-01
6.94510117e-02 8.75849783e-01 7.17868209e-01 -6.28144205e-01
-1.31173944e+00 -6.02597773e-01 1.47506881e+00 -1.15670741e+00
3.57479632e-01 -5.91978312e-01 -1.03814065e+00 9.19504106e-01
6.58853114e-01 -1.01979896e-02 7.80108273e-01 7.98038542e-01
-4.76920813e-01 1.59397870e-01 -6.22956395e-01 8.24517787e-01
1.39847648e+00 -5.72066486e-01 -8.16568971e-01 3.48215215e-02
6.98074579e-01 -4.30356026e-01 6.05631322e-02 -2.34367605e-02
7.55164444e-01 -1.02092171e+00 6.81495488e-01 -5.75815320e-01
2.21773118e-01 -1.17545784e-01 3.67666095e-01 -1.18243778e+00
-8.80266249e-01 -7.26536751e-01 -7.26272643e-01 9.22147036e-01
3.57648045e-01 -2.94807613e-01 7.30426431e-01 9.23240960e-01
-1.00125611e-01 -5.42442381e-01 -6.35961771e-01 -4.43578452e-01
-6.40153766e-01 -4.47349191e-01 6.75450802e-01 5.53747833e-01
5.22646606e-01 9.47517514e-01 -5.87276340e-01 -1.84929520e-01
4.61018980e-01 -1.04190417e-01 7.49660492e-01 -1.48697114e+00
2.37054691e-01 -6.70211613e-01 -3.13964874e-01 -1.50294375e+00
8.34689662e-02 -4.00459945e-01 5.10474980e-01 -1.43463361e+00
3.37813720e-02 -3.81087065e-01 -5.37927389e-01 8.06146085e-01
-4.91636932e-01 4.30317551e-01 3.47537339e-01 5.78238606e-01
-9.15483713e-01 4.79857832e-01 1.05898631e+00 5.52203581e-02
-5.30380428e-01 5.29995523e-02 -9.84673560e-01 2.26461425e-01
3.02114785e-01 -2.59721816e-01 -7.41997480e-01 -6.08531415e-01
5.35738766e-01 -3.27710599e-01 5.22006929e-01 -9.07568157e-01
7.28893518e-01 -2.74540246e-01 2.51946867e-01 -6.63609982e-01
6.35439396e-01 -6.63471222e-01 -4.49175507e-01 -6.11210428e-02
-5.51508307e-01 2.00405642e-02 2.29894236e-01 7.01467097e-01
5.52421175e-02 -1.06859975e-01 1.70307145e-01 2.40920633e-01
-1.32858038e+00 2.75584817e-01 -1.50239646e-01 -2.30290353e-01
1.06848538e+00 -7.32151508e-01 -5.53451240e-01 -3.97327125e-01
-3.96461934e-01 2.72871673e-01 3.85404736e-01 9.78992879e-01
8.74556899e-01 -1.19727099e+00 -1.75897896e-01 4.21806157e-01
8.90002012e-01 -3.04593801e-01 1.76103383e-01 1.05702829e+00
1.57864645e-01 7.83040822e-01 -3.15163910e-01 -7.88549662e-01
-1.39543307e+00 1.04364085e+00 -6.95666522e-02 4.30101126e-01
-2.95486242e-01 1.26447725e+00 5.86135149e-01 3.13422948e-01
4.17463452e-01 -1.58815876e-01 -5.53340554e-01 -5.12815341e-02
1.09543717e+00 1.64365336e-01 -1.94122225e-01 -8.25348556e-01
-5.38959503e-01 4.87405658e-01 -1.01431929e-01 3.70487161e-02
9.41618204e-01 -5.45241654e-01 2.08769038e-01 6.50266290e-01
7.06276178e-01 -2.62895495e-01 -1.25988257e+00 -2.61657596e-01
1.40954092e-01 -5.90766370e-01 1.36983186e-01 -1.03003490e+00
-3.38395625e-01 1.27824044e+00 7.75517106e-01 2.13913500e-01
1.25381649e+00 2.37139598e-01 5.38783133e-01 2.97233939e-01
3.02028179e-01 -7.05349565e-01 2.27469340e-01 5.61947286e-01
1.05247796e+00 -1.67955124e+00 -3.96744817e-01 -3.19554567e-01
-9.13669407e-01 7.87690222e-01 1.17038047e+00 3.73515576e-01
5.73802412e-01 -2.86028147e-01 5.21101430e-02 -5.60216248e-01
-1.21283925e+00 -5.38821101e-01 8.93707097e-01 1.04716885e+00
6.56747878e-01 -4.41399425e-01 1.97961271e-01 4.48375404e-01
4.54308055e-02 3.21703821e-01 -6.79999664e-02 8.65239441e-01
-5.15426576e-01 -8.07253361e-01 -2.65509963e-01 5.31814694e-01
-3.08791716e-02 -4.87900168e-01 -3.55497003e-01 6.46153271e-01
-1.05966270e-01 1.05290842e+00 4.38643634e-01 -5.17965376e-01
3.47296208e-01 2.64901876e-01 3.78928214e-01 -9.48886454e-01
-3.85160953e-01 -1.47022992e-01 -2.59399861e-01 -1.08174789e+00
-4.53430861e-01 -5.40060580e-01 -8.09354186e-01 -9.23145637e-02
-6.57032967e-01 -2.03978702e-01 4.70879376e-01 8.59323084e-01
9.47366059e-01 4.93331522e-01 7.41565153e-02 -1.25936019e+00
-1.34002462e-01 -1.28315973e+00 -3.68481636e-01 2.91920632e-01
6.19850993e-01 -1.18959522e+00 5.28834574e-02 2.20051050e-01] | [14.002553939819336, 0.05459184944629669] |
8c9e44e6-102a-4f9e-a5fb-e1d9120d6113 | rfnet-4d-joint-object-reconstruction-and-flow | 2203.16482 | null | https://arxiv.org/abs/2203.16482v2 | https://arxiv.org/pdf/2203.16482v2.pdf | RFNet-4D: Joint Object Reconstruction and Flow Estimation from 4D Point Clouds | Object reconstruction from 3D point clouds has achieved impressive progress in the computer vision and computer graphics research field. However, reconstruction from time-varying point clouds (a.k.a. 4D point clouds) is generally overlooked. In this paper, we propose a new network architecture, namely RFNet-4D, that jointly reconstruct objects and their motion flows from 4D point clouds. The key insight is that simultaneously performing both tasks via learning spatial and temporal features from a sequence of point clouds can leverage individual tasks, leading to improved overall performance. To prove this ability, we design a temporal vector field learning module using unsupervised learning approach for flow estimation, leveraged by supervised learning of spatial structures for object reconstruction. Extensive experiments and analyses on benchmark dataset validated the effectiveness and efficiency of our method. As shown in experimental results, our method achieves state-of-the-art performance on both flow estimation and object reconstruction while performing much faster than existing methods in both training and inference. Our code and data are available at https://github.com/hkust-vgd/RFNet-4D | ['Duc Thanh Nguyen', 'Sai-Kit Yeung', 'Quang-Hieu Pham', 'Binh-Son Hua', 'Tuan-Anh Vu'] | 2022-03-30 | null | null | null | null | ['3d-human-reconstruction', 'object-reconstruction', 'human-dynamics'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-4.08290446e-01 -6.53239310e-01 -3.33703786e-01 -2.87164748e-01
-2.12572172e-01 -5.51590919e-01 5.21816015e-01 -2.14702874e-01
-8.28641653e-02 4.45727468e-01 1.33418947e-01 -2.48571947e-01
-6.31656870e-02 -8.50412130e-01 -8.90089035e-01 -4.45339173e-01
-3.17567945e-01 3.54613811e-01 4.06428874e-01 2.77333111e-01
3.74969423e-01 1.00142169e+00 -1.34706938e+00 -7.13373348e-02
7.51288235e-01 1.03585553e+00 2.08658814e-01 8.15795064e-01
-3.26594234e-01 9.75843787e-01 1.75999850e-02 -1.61236912e-01
5.19178987e-01 1.71463728e-01 -8.41172159e-01 8.51503536e-02
7.51894057e-01 -7.93618143e-01 -8.58761549e-01 6.57910228e-01
2.02976733e-01 3.24444264e-01 5.90506017e-01 -1.30716980e+00
-4.87266898e-01 6.72734156e-02 -6.45748019e-01 3.52004945e-01
1.37341425e-01 4.44480628e-01 8.24299812e-01 -1.03768337e+00
6.10699832e-01 1.27361584e+00 7.40877867e-01 3.85686755e-01
-1.26705086e+00 -8.72966528e-01 1.67328909e-01 1.85404286e-01
-1.27655542e+00 -4.01233524e-01 9.78021920e-01 -7.33362436e-01
1.01818383e+00 -1.10190473e-01 8.84622395e-01 8.62719297e-01
-2.88339201e-02 1.03625965e+00 5.31787872e-01 4.40273806e-02
1.18606433e-01 -2.32755914e-01 -4.85361069e-02 9.24108744e-01
1.94782928e-01 3.06824267e-01 -6.03407323e-01 -4.26824205e-02
1.42796183e+00 4.06592309e-01 -3.79471898e-01 -6.63812101e-01
-1.40840733e+00 7.07129836e-01 8.11395407e-01 -1.94455534e-02
-4.58987921e-01 6.63839579e-01 4.66238201e-01 -5.24606183e-02
5.53729475e-01 -1.46177799e-01 -4.38072771e-01 -1.86984882e-01
-8.08695853e-01 4.82413709e-01 6.25675559e-01 1.15604937e+00
7.67932832e-01 1.61089629e-01 -3.62940468e-02 5.31398296e-01
5.81251979e-01 6.25780642e-01 -7.81408325e-02 -1.37249720e+00
5.13637722e-01 6.12797856e-01 3.02525491e-01 -1.22759581e+00
-2.07597241e-01 -2.56736904e-01 -1.02750969e+00 3.77963096e-01
4.88763332e-01 1.20875083e-01 -8.10324609e-01 1.38493931e+00
6.71827018e-01 1.31809974e+00 -1.80427790e-01 1.19814336e+00
8.68365407e-01 8.94316494e-01 7.35655520e-03 -2.77854614e-02
8.42339575e-01 -8.78553927e-01 -3.36420447e-01 2.21548036e-01
4.65444386e-01 -5.23086071e-01 6.87001944e-01 8.25101584e-02
-1.13696778e+00 -7.44100630e-01 -7.41944671e-01 -1.68356448e-01
1.51148468e-01 -9.21782777e-02 9.79456782e-01 5.67655787e-02
-8.77688527e-01 9.22274172e-01 -1.34103668e+00 -1.13540374e-01
1.07438946e+00 3.33867818e-01 -4.38568413e-01 -1.84719637e-01
-5.27887046e-01 4.24998462e-01 9.83899385e-02 1.95259571e-01
-8.37996066e-01 -1.28126109e+00 -7.90142655e-01 -2.20183823e-02
1.61922753e-01 -1.11012709e+00 1.27392948e+00 -3.26178938e-01
-1.44172812e+00 6.15392923e-01 -3.67359728e-01 -5.22588015e-01
6.10606790e-01 -3.68596375e-01 9.15289819e-02 2.58095741e-01
2.05385119e-01 7.33120382e-01 7.29556739e-01 -1.26288462e+00
-7.71735787e-01 -3.88145864e-01 -1.39779866e-01 -6.35859445e-02
7.79474080e-02 -3.04014355e-01 -6.86004341e-01 -5.04648447e-01
1.72015756e-01 -8.46407354e-01 -3.00652981e-01 7.12534785e-01
-3.08593482e-01 -4.44023997e-01 9.83237207e-01 -4.71862942e-01
7.54688740e-01 -1.96406543e+00 2.07042862e-02 8.30717236e-02
4.84828234e-01 4.34547335e-01 1.80472955e-01 7.69420639e-02
1.78277478e-01 -3.25972736e-02 -2.69977123e-01 -6.07373893e-01
-1.06623352e-01 2.91086346e-01 -5.91005921e-01 8.20379257e-01
3.14544141e-01 1.16809964e+00 -1.05876684e+00 -4.49754119e-01
7.52787411e-01 6.76416814e-01 -7.08087862e-01 1.67139053e-01
-2.42696717e-01 8.50332439e-01 -7.47384131e-01 7.67563581e-01
1.00671399e+00 -6.74386382e-01 -2.87373453e-01 -3.07086170e-01
-2.66643673e-01 1.16281323e-01 -1.08251190e+00 2.16845036e+00
-2.47219697e-01 7.17326224e-01 -2.60399759e-01 -1.03997993e+00
7.72208035e-01 3.29480797e-01 9.84642923e-01 -5.16100049e-01
3.85598168e-02 7.80223608e-02 -4.18017030e-01 -5.75392187e-01
2.23073125e-01 9.38448776e-03 3.57091278e-01 2.67581642e-01
1.37496054e-01 -9.55816731e-02 -1.08287156e-01 2.65459806e-01
1.13883042e+00 5.38002729e-01 -1.31887542e-02 6.11077435e-02
5.53523660e-01 1.34281874e-01 6.93845332e-01 6.53954446e-01
-4.55329150e-01 5.17380893e-01 1.90800861e-01 -7.12742150e-01
-1.12572420e+00 -1.21000934e+00 -2.21526921e-01 4.70252663e-01
5.44089198e-01 -4.44159240e-01 -1.59793850e-02 -7.31926441e-01
6.18510783e-01 4.05355662e-01 -4.83707458e-01 1.62525982e-01
-1.00570881e+00 -1.43030420e-01 2.91047961e-01 7.67773151e-01
4.94136512e-01 -1.01790762e+00 -6.88111484e-01 2.29701355e-01
-3.79644777e-03 -1.33905375e+00 -3.41490954e-01 -4.53721315e-01
-1.40227842e+00 -1.09468353e+00 -6.37008011e-01 -6.39966667e-01
4.92260277e-01 7.00412452e-01 1.14797986e+00 1.47124574e-01
-3.38363379e-01 5.55671930e-01 -2.36554772e-01 -3.01428169e-01
4.74636033e-02 -2.39519805e-01 -2.89797299e-02 1.85241196e-02
2.40254074e-01 -1.08995295e+00 -9.27940011e-01 3.21762532e-01
-6.68010652e-01 9.11815464e-02 2.62788475e-01 4.71052468e-01
7.57265329e-01 -2.06007734e-01 2.12973401e-01 -4.65132654e-01
-1.17922314e-02 -5.79791963e-01 -9.16866124e-01 -1.66873962e-01
-1.74908355e-01 -4.38259691e-02 4.89144862e-01 -2.77325422e-01
-9.12721336e-01 2.99376070e-01 -2.85998695e-02 -1.38595796e+00
-2.34310791e-01 9.19213071e-02 2.37591341e-01 -2.72388041e-01
3.49450946e-01 3.32792699e-01 1.07025445e-01 -5.78265727e-01
2.95806080e-01 1.96923167e-01 6.27737522e-01 -5.16008914e-01
1.08724654e+00 1.05039024e+00 3.19724023e-01 -5.63384593e-01
-7.37152278e-01 -8.12338412e-01 -8.05244863e-01 -5.37675321e-01
7.35984921e-01 -1.03494728e+00 -1.04998040e+00 5.05466819e-01
-1.33258903e+00 -3.89498830e-01 -1.58925533e-01 7.38279521e-01
-7.55056143e-01 2.50252008e-01 -4.79938984e-01 -6.36439383e-01
-2.99275905e-01 -1.06565917e+00 1.20406973e+00 1.76795810e-01
1.24526843e-01 -1.24203789e+00 2.22331211e-01 1.65442273e-01
1.96087524e-01 4.77415770e-01 4.00623083e-01 2.72184387e-02
-1.41927242e+00 2.29095481e-02 -4.74521458e-01 2.55049586e-01
1.62388653e-01 1.16593063e-01 -8.82047057e-01 -1.76310912e-01
-1.26776353e-01 -1.59211591e-01 9.97615993e-01 7.33895659e-01
1.43189085e+00 -8.32363218e-02 -5.31573296e-01 1.16193879e+00
1.67783785e+00 -1.38910323e-01 5.10256171e-01 6.05943017e-02
1.12453890e+00 2.84487993e-01 5.74335575e-01 6.33407176e-01
4.05567676e-01 6.16546929e-01 7.55203545e-01 2.48195127e-01
-3.33595306e-01 -4.59951013e-01 -6.43640608e-02 6.82469845e-01
-3.28470767e-01 -5.12286313e-02 -1.05529273e+00 6.61190629e-01
-2.11138415e+00 -1.07655799e+00 -3.82038057e-01 1.98150432e+00
3.92520249e-01 2.50865463e-02 1.27368689e-01 -6.19983412e-02
5.22335708e-01 2.66011000e-01 -7.76042521e-01 2.60752141e-01
2.11398572e-01 2.06492439e-01 5.21201849e-01 4.49109644e-01
-1.17626417e+00 1.10900283e+00 5.15988874e+00 4.74220932e-01
-1.30500042e+00 7.21705705e-02 2.67580956e-01 -2.00547263e-01
-1.25790745e-01 5.16557507e-02 -7.19415069e-01 4.48218971e-01
4.56307054e-01 -1.42094836e-01 1.37773916e-01 7.12627888e-01
4.97091860e-01 1.68798059e-01 -1.04330468e+00 1.38290429e+00
-2.96149135e-01 -2.03670669e+00 9.45430025e-02 9.41001773e-02
6.09928489e-01 4.08301562e-01 -1.37104169e-01 1.67783991e-01
2.82639027e-01 -7.90105999e-01 7.05169082e-01 6.03074491e-01
7.12299824e-01 -5.95805645e-01 3.22225660e-01 3.81378412e-01
-1.37681115e+00 1.78615972e-01 -4.10089344e-01 -9.24463719e-02
5.55496693e-01 4.84532088e-01 -7.04122066e-01 5.97158134e-01
9.39863145e-01 1.63709903e+00 -1.21184856e-01 1.44977105e+00
4.61314246e-02 7.43794143e-01 -4.89951938e-01 2.52947837e-01
2.98203200e-01 -2.69243360e-01 9.18634176e-01 1.06624377e+00
2.31330529e-01 2.00989202e-01 2.12930173e-01 1.13034022e+00
-1.78317949e-01 -1.81991458e-01 -7.68035054e-01 2.89171845e-01
4.78140593e-01 1.11530077e+00 -6.63149059e-01 -2.58939803e-01
-4.51216370e-01 6.52438819e-01 5.19197941e-01 3.96032959e-01
-9.67958570e-01 -6.83706328e-02 1.04769099e+00 2.04021826e-01
6.54298484e-01 -6.39930546e-01 -2.99875855e-01 -1.32816708e+00
2.25261971e-01 9.72675085e-02 2.18147635e-02 -7.00854123e-01
-1.44465840e+00 1.09670855e-01 7.51925334e-02 -1.68523896e+00
8.07519406e-02 -5.90981007e-01 -6.71947777e-01 6.97229326e-01
-1.79290640e+00 -1.02992094e+00 -6.48125648e-01 8.64300489e-01
5.64652622e-01 2.76107690e-04 2.44058937e-01 4.08496141e-01
-3.26164156e-01 1.74398676e-01 -1.12273306e-01 3.93909007e-01
3.81817311e-01 -1.00777018e+00 7.28626728e-01 7.56705225e-01
2.14393914e-01 3.22574943e-01 3.10842931e-01 -6.19048536e-01
-1.87355542e+00 -1.50895298e+00 4.10193384e-01 -5.39742231e-01
6.12426758e-01 -2.54354835e-01 -1.11116219e+00 8.24188352e-01
-2.40393564e-01 6.68726981e-01 2.98566967e-01 -3.26223135e-01
-3.15592051e-01 -1.21239997e-01 -1.01328290e+00 4.06984389e-01
1.43876469e+00 -3.07883382e-01 -1.99572399e-01 3.63918632e-01
9.00333345e-01 -6.43447816e-01 -9.41067696e-01 5.42994320e-01
4.14444178e-01 -9.37784612e-01 1.34951866e+00 -6.49610817e-01
5.62442064e-01 -5.28235137e-01 -1.38654873e-01 -9.18086410e-01
-4.75601047e-01 -5.33937871e-01 -7.94927180e-01 8.35873723e-01
-1.17583215e-01 -4.88347560e-01 1.06475353e+00 3.27514291e-01
-2.19018951e-01 -7.74936914e-01 -8.41006398e-01 -7.26105571e-01
2.97050104e-02 -7.71380961e-01 5.36874831e-01 9.79976714e-01
-5.83448708e-01 1.49716169e-01 -3.10952008e-01 4.65136945e-01
1.05658650e+00 3.76836210e-01 1.08532596e+00 -1.20776880e+00
-7.63992313e-03 -4.63550478e-01 -8.33934009e-01 -1.61698139e+00
5.38749516e-01 -1.15563107e+00 -1.07385039e-01 -1.66732955e+00
-1.58212006e-01 -8.01954687e-01 -6.21916354e-02 2.72845745e-01
-9.83850956e-02 1.78792596e-01 3.00055265e-01 5.62970161e-01
-6.12904966e-01 6.90675437e-01 1.60259509e+00 -3.81792970e-02
-3.73461962e-01 1.76266968e-01 -1.96407706e-01 5.52508414e-01
8.30823779e-01 -3.98835182e-01 -4.09090638e-01 -8.40932071e-01
-2.71900415e-01 3.80251378e-01 9.02239919e-01 -1.01165533e+00
4.22553748e-01 -2.94893593e-01 5.21015763e-01 -9.68205571e-01
4.55210388e-01 -9.28775191e-01 1.11785330e-01 2.92678893e-01
-4.64423820e-02 -1.22151487e-01 3.36829543e-01 9.38372970e-01
-8.60171691e-02 2.15815425e-01 6.58127964e-01 -1.91472292e-01
-9.46271777e-01 1.15830767e+00 2.11821973e-01 -1.18203811e-01
9.82995629e-01 -3.18264186e-01 -4.68228273e-02 -1.03964254e-01
-4.91060853e-01 3.11037242e-01 3.96866828e-01 5.03541589e-01
1.01158524e+00 -1.52252984e+00 -8.32918823e-01 3.45224231e-01
-7.95436930e-03 7.15746701e-01 3.22995603e-01 8.48832488e-01
-8.80297780e-01 4.54717249e-01 -1.71215847e-01 -1.37579358e+00
-8.91777396e-01 3.71714443e-01 3.32970321e-01 9.92229059e-02
-1.23345006e+00 8.14070821e-01 1.89226404e-01 -3.94626677e-01
2.54860997e-01 -4.38895762e-01 3.75031121e-02 -4.98386055e-01
4.84304041e-01 5.10378063e-01 -2.19398588e-01 -6.21876657e-01
-3.04109186e-01 7.78640628e-01 -1.56645440e-02 1.67044610e-01
1.56542945e+00 4.03931551e-02 4.45958562e-02 4.16147530e-01
1.45901513e+00 -3.50774109e-01 -1.74515581e+00 -4.16308254e-01
-2.64876753e-01 -1.07196128e+00 2.65677512e-01 -1.64086238e-01
-1.47491181e+00 8.89438033e-01 3.24916095e-01 -1.56553194e-01
8.55676591e-01 -1.67129003e-03 1.00653768e+00 2.02018663e-01
5.23128271e-01 -2.83164769e-01 3.29262838e-02 5.30641973e-01
6.52060032e-01 -1.35900342e+00 1.22345230e-02 -5.66494823e-01
-2.65350968e-01 1.09535921e+00 5.66531658e-01 -7.35302567e-01
9.96130824e-01 3.11340988e-02 -2.78459996e-01 -2.21896291e-01
-7.98037291e-01 4.04030234e-02 2.63457268e-01 4.55748647e-01
7.71228224e-02 -5.85513376e-02 4.95206118e-01 -7.16238096e-02
5.48184291e-02 4.87219751e-01 1.77890673e-01 1.18382323e+00
-2.37905279e-01 -7.47458339e-01 -1.35034099e-01 4.25072521e-01
-1.99393645e-01 2.79369086e-01 1.99609399e-01 7.06038594e-01
-1.61360398e-01 5.59488833e-01 4.70281065e-01 -2.05531046e-01
3.78044277e-01 -3.83612245e-01 6.32828116e-01 -2.20606282e-01
-1.87238097e-01 -1.83766842e-01 -4.41724598e-01 -1.01996469e+00
-7.29892373e-01 -7.22115040e-01 -1.30774713e+00 -7.09644020e-01
-5.55456206e-02 -1.17830671e-01 6.20721340e-01 6.52862668e-01
6.49159729e-01 3.12657923e-01 7.16075599e-01 -1.36403883e+00
-3.00563365e-01 -5.24527907e-01 -4.73815650e-01 4.96774465e-01
7.83771098e-01 -8.69828224e-01 -3.69998723e-01 2.95166701e-01] | [8.533296585083008, -2.0291359424591064] |
e69a9e39-4694-478e-9a22-65141a075ae4 | the-cl-scisumm-shared-task-2018-results-and | 1909.00764 | null | https://arxiv.org/abs/1909.00764v1 | https://arxiv.org/pdf/1909.00764v1.pdf | The CL-SciSumm Shared Task 2018: Results and Key Insights | This overview describes the official results of the CL-SciSumm Shared Task 2018 -- the first medium-scale shared task on scientific document summarization in the computational linguistics (CL) domain. This year, the dataset comprised 60 annotated sets of citing and reference papers from the open access research papers in the CL domain. The Shared Task was organized as a part of the 41st Annual Conference of the Special Interest Group in Information Retrieval (SIGIR), held in Ann Arbor, USA in July 2018. We compare the participating systems in terms of two evaluation metrics. The annotated dataset and evaluation scripts can be accessed and used by the community from: \url{https://github.com/WING-NUS/scisumm-corpus}. | ['Min-Yen Kan', 'Muthu Kumar Chandrasekaran', 'Kokil Jaidka', 'Dragomir Radev', 'Michihiro Yasunaga'] | 2019-09-02 | null | null | null | null | ['scientific-article-summarization'] | ['natural-language-processing'] | [ 3.99168901e-04 2.02699080e-01 -4.45437193e-01 5.77699021e-02
-1.56306219e+00 -7.74955630e-01 7.64459968e-01 8.60758424e-01
-5.44787288e-01 1.05245721e+00 9.78327751e-01 -2.16116041e-01
-1.84569925e-01 -1.66347995e-01 -4.89150465e-01 -2.38131866e-01
1.49210170e-02 4.66001272e-01 -3.45519260e-02 -7.60371611e-02
1.03423607e+00 6.31853715e-02 -1.39975190e+00 5.22717834e-01
1.39270258e+00 2.39303008e-01 4.37733442e-01 9.22161222e-01
-1.86674550e-01 3.43419760e-01 -8.61359835e-01 -4.96673912e-01
-2.29124859e-01 -4.57184076e-01 -1.18950951e+00 -7.32758999e-01
7.74227023e-01 3.41504961e-01 -3.07822555e-01 1.13842487e+00
8.55132282e-01 2.54575461e-01 5.23075342e-01 -9.50086653e-01
-5.57670832e-01 1.16944945e+00 -5.06569624e-01 7.90669203e-01
5.74210763e-01 -4.02739495e-01 1.40908194e+00 -8.85511279e-01
1.12096214e+00 1.16263843e+00 3.23932201e-01 3.28992665e-01
-6.46533191e-01 -8.21508229e-01 1.42317060e-02 4.42700714e-01
-1.13065720e+00 -6.21800601e-01 5.85768461e-01 -3.31555158e-01
9.59987283e-01 6.89418256e-01 2.30877683e-01 1.15944541e+00
-9.53002181e-03 1.13267243e+00 8.36607218e-01 -5.34972727e-01
1.94382444e-01 -1.66844293e-01 7.32805371e-01 2.61872023e-01
9.96671259e-01 -8.94183576e-01 -8.61305773e-01 -5.06292284e-01
-2.18004793e-01 -6.35893106e-01 -6.75788462e-01 3.43503177e-01
-1.45340466e+00 7.02498972e-01 2.84197256e-02 7.21671343e-01
-4.92230177e-01 -7.43671730e-02 7.46419311e-01 2.93871462e-01
1.21023750e+00 9.99086797e-01 -4.76137102e-01 -4.46207225e-01
-1.06844699e+00 7.03948796e-01 1.17897761e+00 7.82644451e-01
4.39461134e-02 -5.41709006e-01 -4.70187157e-01 9.22272325e-01
1.01966806e-01 3.32515925e-01 5.66776514e-01 -1.13570881e+00
7.94598877e-01 4.76737738e-01 8.69439691e-02 -7.57034421e-01
-4.11546260e-01 -4.89227206e-01 -6.23721182e-01 -5.72080493e-01
3.30345221e-02 -5.88749886e-01 -2.94720978e-01 1.39986813e+00
8.01216885e-02 6.93748891e-02 1.85956165e-01 6.90632641e-01
1.89682925e+00 8.55170786e-01 -3.05733108e-03 -7.41224229e-01
1.32859814e+00 -8.99328947e-01 -1.31162214e+00 1.23247363e-01
9.81659710e-01 -1.06779182e+00 5.62002182e-01 2.49970019e-01
-1.45883632e+00 3.94952670e-02 -1.26581550e+00 -3.94884408e-01
-7.26144969e-01 2.73414850e-01 2.70938069e-01 -1.89018697e-01
-1.26660621e+00 5.79062402e-01 -5.39388478e-01 -7.25408852e-01
3.62277448e-01 -2.94784099e-01 -1.69793442e-01 -5.31682968e-02
-1.16084766e+00 1.12530112e+00 2.27070481e-01 -2.78706014e-01
-3.61473143e-01 -1.00707626e+00 -4.41450089e-01 -8.17268863e-02
4.30594414e-01 -5.22808552e-01 1.40361309e+00 -2.96872169e-01
-1.03374910e+00 1.20080471e+00 -4.45735276e-01 -4.35650200e-01
2.57920146e-01 -6.04842365e-01 -3.46011072e-01 6.87640607e-02
4.67370480e-01 1.87323213e-01 -8.57875347e-02 -1.07249892e+00
-5.02273977e-01 -4.84709322e-01 -2.26698756e-01 3.03633809e-01
-1.95023581e-01 5.77433228e-01 -5.59567392e-01 -7.86511004e-01
-1.96189761e-01 -7.73967385e-01 2.04663854e-02 -8.50515008e-01
-6.30474865e-01 -9.37843561e-01 5.00041246e-01 -9.46315408e-01
1.59227526e+00 -1.62114763e+00 4.78351861e-01 -5.31203568e-01
1.96965218e-01 2.74161756e-01 -7.08307279e-03 9.94327724e-01
-9.89391208e-02 5.21832228e-01 -7.81003609e-02 -7.23225296e-01
-1.19487420e-01 -4.56251889e-01 -4.46397126e-01 3.71836990e-01
-4.71064389e-01 8.00790370e-01 -1.18964314e+00 -5.58470011e-01
-4.68899488e-01 3.83242182e-02 2.73428649e-01 -1.68386903e-02
-1.42374068e-01 2.69347280e-01 -7.92384148e-01 2.18202159e-01
5.45999885e-01 -1.80448890e-01 1.00884706e-01 -2.89722495e-02
-6.40576303e-01 8.35853934e-01 -6.05645955e-01 2.02649474e+00
-7.82805309e-02 9.53631163e-01 4.93623354e-02 -7.58942485e-01
5.55759668e-01 5.25931299e-01 4.87516284e-01 -2.99728960e-01
2.08194897e-01 5.50262451e-01 -8.48857462e-02 -3.43491971e-01
9.42002773e-01 9.52779591e-01 -1.90142050e-01 7.33036995e-01
2.67595321e-01 -4.09483492e-01 1.08754456e+00 1.02389181e+00
1.15193152e+00 1.19260838e-02 2.83134401e-01 -7.34952390e-01
5.32661498e-01 2.18263909e-01 5.13409615e-01 7.74331152e-01
1.82949509e-02 5.59009135e-01 5.01931131e-01 5.47432760e-03
-9.04565871e-01 -4.36163634e-01 -3.44697386e-01 1.07104254e+00
-2.21227482e-01 -1.04190421e+00 -7.87609637e-01 -2.76079267e-01
-1.69846922e-01 1.22601032e+00 -6.91847622e-01 4.61707450e-02
-4.35754657e-01 -6.19417787e-01 6.95708036e-01 -1.25011787e-01
4.50625002e-01 -1.22504270e+00 -2.88881928e-01 -1.52968839e-01
-6.71824455e-01 -1.02170396e+00 -7.13024020e-01 6.77970499e-02
-6.82311356e-01 -1.04255974e+00 -1.10447979e+00 -7.14285553e-01
3.11179996e-01 1.92729920e-01 9.78173316e-01 2.30470169e-02
-3.31294894e-01 3.95873964e-01 -5.63031793e-01 -1.06035817e+00
-2.90652275e-01 5.14631927e-01 -2.58992136e-01 -5.33753633e-01
4.15067434e-01 -2.02257752e-01 -4.33022857e-01 -6.25678480e-01
-5.03063917e-01 1.20993957e-01 1.74075052e-01 4.25462216e-01
4.12667096e-01 -8.09434712e-01 1.02556908e+00 -9.47091341e-01
1.30893040e+00 -6.26848638e-01 -4.13052946e-01 6.00072920e-01
-6.50327444e-01 -2.64652550e-01 -4.92660236e-03 1.43797979e-01
-1.07225084e+00 -7.37833798e-01 1.97607011e-01 1.99250892e-01
2.17777133e-01 1.19401097e+00 1.02050006e-01 4.48583513e-01
5.51060855e-01 -7.79638216e-02 -3.37258756e-01 -7.61160910e-01
3.77898037e-01 1.17115128e+00 6.37300313e-01 -6.22400820e-01
2.07785100e-01 -4.82846536e-02 -5.95835410e-02 -7.95874655e-01
-1.37347233e+00 -8.07768345e-01 -3.32483828e-01 -2.63716936e-01
8.08434248e-01 -9.76452589e-01 -2.47955754e-01 4.28506583e-01
-1.41888189e+00 6.16569780e-02 -2.03570232e-01 3.89738232e-01
1.53425425e-01 3.56562972e-01 -7.78434038e-01 -5.94536066e-01
-1.10722017e+00 -7.58456469e-01 9.20696080e-01 4.93554175e-01
-5.76738179e-01 -1.06840110e+00 3.79955262e-01 6.97206438e-01
2.91850448e-01 2.12634414e-01 4.68402267e-01 -1.31399632e+00
2.76372015e-01 -3.43805283e-01 -1.95356291e-02 -8.21211562e-03
-1.29571468e-01 3.42831701e-01 -7.76488483e-01 -1.88373074e-01
-4.27963406e-01 -5.36969721e-01 1.30438662e+00 5.71304679e-01
1.14949727e+00 -1.76631689e-01 -4.60803330e-01 -1.98229656e-01
1.00014985e+00 -8.42694789e-02 3.85771155e-01 4.55844074e-01
3.55593354e-01 5.64437747e-01 5.23027182e-01 4.86152709e-01
5.86785972e-01 4.46919143e-01 -1.85461372e-01 4.44161683e-01
-3.75216365e-01 1.12254089e-02 1.06257917e-02 1.58391345e+00
-2.15574622e-01 -8.12569380e-01 -1.02787340e+00 7.40273476e-01
-2.11241889e+00 -1.00776780e+00 -6.29936039e-01 2.27944040e+00
1.26688278e+00 -1.44980311e-01 -1.41480327e-01 -3.62685293e-01
6.82626009e-01 4.03878063e-01 -1.80411890e-01 -6.66745663e-01
-6.33874953e-01 2.12650523e-01 4.30639088e-01 6.07102513e-01
-9.13458169e-01 7.40216553e-01 5.16518307e+00 1.13430071e+00
-7.19181597e-01 4.97881889e-01 3.83857638e-01 -3.20738345e-01
-5.09199858e-01 1.35164440e-01 -9.20249224e-01 4.89386052e-01
1.38446856e+00 -1.30956411e+00 -1.55791610e-01 9.98462886e-02
1.56505659e-01 -4.82609242e-01 -6.14995122e-01 4.27417308e-01
4.74273771e-01 -1.67248774e+00 -7.86791146e-02 -2.99919154e-02
9.69266355e-01 6.46631181e-01 -1.57365769e-01 2.76340425e-01
4.52483535e-01 -7.99330890e-01 4.36767727e-01 5.28768837e-01
7.22631693e-01 -6.13824725e-01 8.69041383e-01 3.29113364e-01
-4.70863253e-01 3.70877296e-01 -9.44261029e-02 1.97851866e-01
9.36073288e-02 7.32602358e-01 -1.57295659e-01 9.99626219e-01
7.74777234e-01 8.92759085e-01 -5.59112132e-01 1.32764781e+00
-2.36578703e-01 1.02136993e+00 -5.22561483e-02 -4.52708840e-01
2.02384084e-01 -1.30368963e-01 1.12574422e+00 1.70012879e+00
1.87353685e-01 2.82171249e-01 2.49302853e-02 5.27069867e-01
-6.83317542e-01 3.24416131e-01 -3.50160837e-01 -4.06136036e-01
7.52475977e-01 1.59642410e+00 -4.75569814e-01 -5.21097660e-01
4.26075459e-02 6.90503180e-01 3.29402953e-01 3.70812118e-01
-4.18158323e-01 -7.66431093e-01 -1.94816664e-02 -4.96183276e-01
-9.99189913e-03 9.97990072e-02 -5.21492660e-01 -1.11584246e+00
-3.55035178e-02 -6.29055917e-01 5.69880664e-01 -9.14140046e-01
-1.26852977e+00 5.89455426e-01 3.12501878e-01 -8.61039937e-01
-1.10854886e-01 1.04238719e-01 -8.37467432e-01 9.99831378e-01
-1.09960449e+00 -1.05436790e+00 -1.01792485e-01 -9.92162451e-02
8.86376560e-01 -2.90303707e-01 8.77859533e-01 1.07182682e-01
-7.25616574e-01 4.87069786e-01 7.00237811e-01 -2.74088711e-01
1.12643123e+00 -1.32969952e+00 3.34521383e-01 6.14306033e-01
-1.63776591e-01 6.94398820e-01 1.04245865e+00 -9.15441155e-01
-1.26169884e+00 -8.93165171e-01 1.85240066e+00 -4.30808395e-01
9.13120031e-01 -1.48245960e-01 -9.41307127e-01 5.53256989e-01
1.30947137e+00 -5.58281958e-01 7.48477697e-01 1.84551954e-01
-6.46287724e-02 3.43923122e-01 -7.52427816e-01 6.08448982e-01
8.24731529e-01 -1.64321244e-01 -8.77587974e-01 9.66241121e-01
8.86353016e-01 -6.78101718e-01 -1.00514328e+00 3.01529050e-01
2.81213492e-01 5.40263169e-02 4.72906411e-01 -5.19226730e-01
9.76410329e-01 6.32434562e-02 1.78083330e-01 -1.53457487e+00
-2.76767939e-01 -6.61038220e-01 -2.06622317e-01 1.48859489e+00
7.93210983e-01 -5.37635267e-01 2.56511942e-02 1.89993203e-01
-6.35710776e-01 -7.53815174e-01 -9.76537228e-01 -3.61334294e-01
6.38927400e-01 -1.49397552e-01 3.39141153e-02 1.02047598e+00
5.63605130e-01 5.99512458e-01 2.43243411e-01 -2.97002256e-01
7.27895856e-01 5.52691855e-02 4.19762582e-01 -1.27339220e+00
4.09792572e-01 -8.47305954e-01 1.13060519e-01 -3.36901605e-01
5.19692540e-01 -1.26535177e+00 -1.63577348e-01 -2.17946315e+00
6.16188526e-01 1.72591254e-01 -3.49114478e-01 4.92426455e-01
-4.91060525e-01 4.85542491e-02 1.52202636e-01 5.91097951e-01
-1.23894930e+00 3.33987832e-01 1.05509591e+00 -1.48043960e-01
1.79752573e-01 -2.45112687e-01 -1.15095282e+00 2.03175575e-01
7.51026392e-01 -3.81386429e-01 -9.52761844e-02 -4.67604190e-01
1.45029277e-01 -8.47687200e-02 -5.66449575e-02 -6.92451358e-01
6.00311995e-01 8.85656178e-02 -1.95853949e-01 -9.33391392e-01
4.98108789e-02 1.93625197e-01 -4.20096330e-02 2.34123588e-01
-9.92207229e-01 1.69293329e-01 5.65658987e-01 1.17725581e-01
-2.29754642e-01 -5.56029081e-01 5.74932955e-02 -2.16824993e-01
-1.78637922e-01 -9.50680822e-02 -4.37420607e-01 5.75172901e-01
7.83700228e-01 5.08599997e-01 -1.08629334e+00 -4.26870614e-01
-2.17947796e-01 7.11674213e-01 6.88569108e-03 7.01279581e-01
2.65075028e-01 -8.09184253e-01 -1.60313332e+00 -8.82141531e-01
1.67195126e-01 -1.83075219e-01 3.63185346e-01 1.16340911e+00
-3.83178920e-01 9.41362917e-01 1.95107937e-01 5.79869449e-02
-1.31678331e+00 9.78680775e-02 -2.51186967e-01 -5.79503655e-01
-5.89136481e-01 7.08782792e-01 -2.69615650e-01 -3.85747612e-01
2.57032394e-01 2.71284699e-01 -8.00024748e-01 3.85255933e-01
9.29298520e-01 7.50162125e-01 4.08328116e-01 -4.72914279e-01
-1.64295733e-01 9.08727497e-02 -3.48726898e-01 -4.09270525e-01
1.63851774e+00 -2.69285917e-01 -6.58717871e-01 9.08408225e-01
1.35136759e+00 2.95363180e-02 -6.43626228e-02 -3.61533999e-01
5.01231372e-01 1.90332502e-01 6.30526483e-01 -1.12417340e+00
-6.69780493e-01 4.27958786e-01 7.54893646e-02 1.90707728e-01
5.32414556e-01 8.60094056e-02 6.53527796e-01 4.22330230e-01
-6.43434823e-02 -1.27699733e+00 -4.65350986e-01 7.41962671e-01
1.55891490e+00 -9.50961649e-01 6.27636194e-01 -1.87040389e-01
-8.00868869e-01 1.13968658e+00 3.64788651e-01 -5.13712950e-02
5.67609310e-01 7.41913095e-02 -5.47969043e-02 -5.44792652e-01
-1.19460058e+00 1.40270308e-01 7.62507737e-01 -1.60794258e-02
1.19296920e+00 -1.16066828e-01 -1.53820860e+00 7.32718110e-01
-2.11558878e-01 4.12510298e-02 8.48572969e-01 1.09505844e+00
-3.02699983e-01 -8.35019648e-01 4.77096476e-02 7.70134568e-01
-8.69844794e-01 -2.33169138e-01 -8.72046113e-01 2.38622874e-01
-4.19828236e-01 1.12489820e+00 9.03964117e-02 1.14646787e-02
1.30144224e-01 -3.57254893e-02 2.88631111e-01 -5.87088346e-01
-8.43328595e-01 -7.16397315e-02 7.07621217e-01 -1.89298198e-01
-5.57537198e-01 -1.11626017e+00 -1.17338741e+00 -5.19701064e-01
-2.08023876e-01 9.99284089e-01 9.45507824e-01 6.71667099e-01
6.97198808e-01 6.85629368e-01 1.52680561e-01 -9.24651146e-01
-3.55867058e-01 -1.58564425e+00 -5.28820276e-01 2.88816937e-03
-2.28250232e-02 -3.99321556e-01 -6.50024056e-01 -9.71289203e-02] | [12.386625289916992, 9.572855949401855] |
1c7da8e7-c98b-45b8-a2ba-676a3d0059a6 | zero-shot-visual-slot-filling-as-question | 2011.12340 | null | https://arxiv.org/abs/2011.12340v2 | https://arxiv.org/pdf/2011.12340v2.pdf | Zero-Shot Visual Slot Filling as Question Answering | This paper presents a new approach to slot filling by reformulating the slot filling task as Question Answering, and replacing slot tags with rich natural language questions that capture the semantics of visual information and lexical text often displayed on device screens. These questions are paired with the user's utterance, and slots are extracted from the utterance using a state-of-the-art Transformer-based deep learning Question Answering system. An approach to further refine the model with multi-task training is presented. The multi-task approach facilitates the incorporation of a large number of successive refinements and transfer learning across tasks. New visual slot datasets and a visual extension of the popular ATIS dataset are introduced to support research and experimentation on visual slot filling. Results show the new approach not only maintains robust accuracy for sparse training conditions but achieves state-of-the-art F1 of 0.97 on ATIS with approximately 1/10th the training data. | ['Simon Heck', 'Larry Heck'] | 2020-11-24 | null | null | null | null | ['zero-shot-slot-filling'] | ['natural-language-processing'] | [ 4.24803138e-01 6.94930077e-01 -3.49155664e-01 -4.84359145e-01
-1.30692232e+00 -4.29050803e-01 4.72842276e-01 3.02568704e-01
-3.52525949e-01 4.88529295e-01 4.62378770e-01 -7.37563014e-01
2.63907284e-01 -3.81233364e-01 -5.05622447e-01 8.80965143e-02
4.82194930e-01 1.04287446e+00 7.76225448e-01 -4.27475065e-01
-3.23825739e-02 -7.65695199e-02 -1.72298348e+00 1.11739147e+00
4.25073415e-01 1.19820654e+00 5.86083472e-01 6.15597367e-01
-9.92317677e-01 1.10445535e+00 -7.69174278e-01 -2.04200819e-01
-8.18351582e-02 -1.27680987e-01 -1.27682257e+00 3.87799114e-01
7.35744894e-01 -3.48286718e-01 -1.45638585e-01 2.80399472e-01
5.04680693e-01 2.81601399e-01 1.25406593e-01 -1.20989680e+00
-5.07200301e-01 9.67573524e-02 -3.89023274e-01 1.24894254e-01
6.09641910e-01 -2.98625976e-01 1.25497615e+00 -1.18343902e+00
7.84191728e-01 1.41816258e+00 7.14846015e-01 7.14893281e-01
-1.20015311e+00 -4.24163252e-01 1.41250834e-01 1.52865171e-01
-9.72281456e-01 -4.76270765e-01 6.60087764e-01 -2.76097506e-01
1.48124909e+00 3.40596288e-01 6.14514351e-01 8.87646616e-01
-2.69925684e-01 1.40555656e+00 1.01178980e+00 -8.49336267e-01
1.19457282e-01 5.02044022e-01 2.01901108e-01 9.82074797e-01
-3.65044296e-01 -2.95033753e-01 -9.71970499e-01 -1.82815105e-01
8.10951531e-01 -2.47931197e-01 3.01168799e-01 -8.22727144e-01
-8.10158432e-01 9.24807131e-01 2.30399638e-01 1.56835377e-01
-1.60887823e-01 -1.56309620e-01 7.50006020e-01 2.93769777e-01
5.76274693e-01 2.19565615e-01 -5.60238957e-01 -1.85736999e-01
-9.78646278e-01 4.30698484e-01 5.65163016e-01 9.83834922e-01
7.88519204e-01 2.06184052e-02 -5.91343164e-01 1.18471360e+00
4.85200435e-01 3.18146944e-01 4.50663775e-01 -6.65703654e-01
6.36845589e-01 6.88924372e-01 1.96663395e-01 -3.70809525e-01
-6.11419141e-01 -8.79963785e-02 -3.48032750e-02 1.92810789e-01
5.83249032e-01 1.89111188e-01 -1.58214128e+00 1.46637547e+00
6.09451056e-01 -2.68299103e-01 -3.31616402e-02 5.41212440e-01
1.30430365e+00 5.10356486e-01 8.11929584e-01 1.37485072e-01
2.09320164e+00 -1.09173775e+00 -9.62698162e-01 -8.25363159e-01
6.29822552e-01 -7.77483344e-01 1.68690193e+00 -3.17864567e-02
-1.03600061e+00 -9.04303551e-01 -9.10867989e-01 -7.85503328e-01
-5.15587986e-01 1.89513952e-01 7.07350552e-01 7.75109529e-01
-1.17790306e+00 -1.79007456e-01 -5.99160969e-01 -4.70555931e-01
5.90666771e-01 3.13423038e-01 -3.02357197e-01 -7.06575662e-02
-1.04721069e+00 9.45768297e-01 1.29902378e-01 -5.99580944e-01
-6.66003942e-01 -8.03445220e-01 -1.18069911e+00 -3.71655375e-02
5.72226286e-01 -7.38979757e-01 2.08469868e+00 -7.57467389e-01
-1.17199337e+00 1.42467260e+00 -6.50817275e-01 -3.56933117e-01
4.97775339e-02 -4.32420611e-01 -3.46165478e-01 3.61053765e-01
6.15537465e-01 1.13456929e+00 1.05052221e+00 -1.21745169e+00
-1.05169535e+00 -4.46403086e-01 3.47352058e-01 6.31427288e-01
-3.13530937e-02 -1.41459316e-01 -7.33456850e-01 -3.94848675e-01
1.96877703e-01 -7.00226665e-01 1.09618388e-01 2.03611702e-01
-1.37531117e-01 -4.82558638e-01 1.20413601e+00 -8.91599476e-01
1.11218703e+00 -2.36609364e+00 -2.99891055e-01 2.13383306e-02
3.27983111e-01 2.11735070e-01 -1.99801505e-01 3.98074418e-01
-1.95580766e-01 -2.76015073e-01 3.19898576e-01 -8.43353570e-01
2.41505459e-01 2.78467923e-01 -5.62927067e-01 -1.51916042e-01
6.99470611e-03 1.25863779e+00 -5.48058748e-01 -7.23102331e-01
4.57683653e-01 3.39277416e-01 -4.22789782e-01 3.53523403e-01
-6.93371236e-01 1.81295559e-01 -3.67267132e-01 7.68569052e-01
1.13308720e-01 -6.36099815e-01 2.83193290e-01 -4.09601450e-01
2.50005275e-01 7.63439834e-01 -8.40738654e-01 2.05931544e+00
-4.40735728e-01 7.14695573e-01 -9.21902526e-03 -1.01616478e+00
7.75439560e-01 5.85416973e-01 4.98057336e-01 -1.35102963e+00
-5.48671223e-02 -1.22630566e-01 -7.78097272e-01 -5.17021596e-01
9.46872294e-01 -4.45064783e-01 -3.42032969e-01 3.89420986e-01
2.37720639e-01 -2.42569298e-01 9.05851200e-02 4.08499181e-01
8.50385129e-01 2.88487345e-01 3.51069391e-01 9.61146429e-02
-9.16329492e-03 5.45597851e-01 -6.12151213e-02 6.60653651e-01
-2.03847215e-01 4.29452926e-01 2.93875396e-01 -6.23134315e-01
-1.23825014e+00 -1.33168340e+00 -3.17423865e-02 1.83955741e+00
3.47995237e-02 -5.56693554e-01 -5.52518427e-01 -7.17459440e-01
5.18412935e-03 6.28400028e-01 -7.47233391e-01 1.76742762e-01
-1.62828699e-01 -6.35627061e-02 3.80606383e-01 8.13837767e-01
4.54566211e-01 -1.31276965e+00 -1.08606482e+00 1.51819900e-01
-6.64111674e-01 -1.35638237e+00 -8.98521394e-02 4.50418442e-01
-8.34590018e-01 -1.04825091e+00 -7.81673849e-01 -1.30270505e+00
3.13685119e-01 3.41572911e-01 1.62200987e+00 -7.50094801e-02
-3.98528486e-01 8.24145079e-01 -4.17063117e-01 -3.78721565e-01
-1.44607335e-01 -3.46777476e-02 -5.94893217e-01 -5.31245530e-01
7.13103056e-01 1.76676027e-02 -3.64658952e-01 2.00044066e-01
-8.90961528e-01 6.38113976e-01 3.32772523e-01 9.53872859e-01
8.74887407e-01 -3.59365255e-01 2.55041152e-01 -1.01973057e+00
6.84281230e-01 -1.01132028e-01 -2.43533298e-01 4.36491460e-01
-3.72470438e-01 2.07197964e-01 -2.84844041e-01 -3.15714687e-01
-1.38787746e+00 3.32354933e-01 -2.41757780e-01 -1.99464723e-01
-3.07210952e-01 3.48208487e-01 2.24935547e-01 2.95080423e-01
8.43351662e-01 -1.34341687e-01 -9.32992473e-02 -4.59523827e-01
5.69752574e-01 6.21216774e-01 7.17263401e-01 -4.15301114e-01
3.60401630e-01 4.17839736e-01 -4.43632156e-01 -8.79126668e-01
-1.13286245e+00 -8.39995205e-01 -3.80181491e-01 -1.75032780e-01
1.04247999e+00 -1.08552670e+00 -5.61868191e-01 1.72886238e-01
-1.00513089e+00 -6.67873681e-01 -8.51991594e-01 -1.02711737e-01
-7.50393569e-01 1.23918481e-01 -5.56869090e-01 -9.33822155e-01
-3.18322390e-01 -7.66354024e-01 1.72179914e+00 1.80678427e-01
-5.94579339e-01 -9.00380254e-01 9.22953635e-02 9.48462307e-01
2.16713101e-01 -2.42902458e-01 1.20208251e+00 -6.75048292e-01
-3.79263848e-01 4.33053561e-02 -6.20184004e-01 -4.26079392e-01
5.68917505e-02 -9.45191801e-01 -1.23460400e+00 -1.65047228e-01
-3.40896368e-01 -9.69994426e-01 6.56832516e-01 5.31768978e-01
9.19266582e-01 3.19580227e-01 -4.52314734e-01 6.61616623e-02
1.14571607e+00 4.52053726e-01 8.34505379e-01 5.61371505e-01
6.03755951e-01 6.83129907e-01 9.37245548e-01 3.34463537e-01
7.36164987e-01 8.63796413e-01 1.85447812e-01 -5.81644177e-01
-3.56277764e-01 -7.04951167e-01 -2.79771477e-01 -9.32980478e-02
5.84233284e-01 8.43683109e-02 -9.31785882e-01 6.65868163e-01
-1.76650250e+00 -7.28858471e-01 2.04349846e-01 1.93336606e+00
8.41913283e-01 2.61456430e-01 4.61176842e-01 3.28810841e-01
5.09082079e-01 2.71806866e-01 -5.76314449e-01 -3.01907420e-01
2.00287998e-01 7.88663149e-01 1.37631834e-01 6.79235280e-01
-1.32361066e+00 1.38414001e+00 7.08856153e+00 7.68794954e-01
-8.18500459e-01 2.61451632e-01 7.03531623e-01 7.95448869e-02
-5.20582557e-01 -3.07864934e-01 -8.48429263e-01 -3.02406907e-01
8.03533614e-01 2.55713165e-01 7.72953853e-02 9.23022091e-01
-2.06615239e-01 -4.32558477e-01 -7.50083625e-01 1.22214544e+00
1.28897518e-01 -1.44048452e+00 -1.32110864e-01 -2.50766903e-01
2.66113251e-01 -1.79460675e-01 3.54437202e-01 7.28626192e-01
1.84889615e-01 -9.71042335e-01 8.66897404e-01 -1.68747455e-02
1.29660475e+00 -3.33559304e-01 1.97342649e-01 -1.91542849e-01
-1.57477152e+00 -1.78356886e-01 -1.51428401e-01 2.12247632e-02
3.52339804e-01 -8.23963340e-03 -1.25721729e+00 7.93584213e-02
8.87246788e-01 2.02542469e-01 -6.97860241e-01 7.98946798e-01
9.16216895e-03 2.52839893e-01 -2.70401806e-01 7.25659430e-02
2.61728317e-01 6.16505325e-01 -2.28152368e-02 1.19021392e+00
-2.84058060e-02 6.30117655e-02 3.66326690e-01 3.13610226e-01
3.17337930e-01 2.93753803e-01 -4.52583969e-01 -1.24924272e-01
3.60027522e-01 9.99852955e-01 -9.08520162e-01 -6.91012144e-01
-6.84408545e-01 1.12960899e+00 3.64384741e-01 4.22908545e-01
-4.88935918e-01 -1.50292233e-01 4.11608309e-01 2.73614913e-01
5.46280563e-01 5.55201154e-03 -5.33490479e-01 -6.97970927e-01
3.05434056e-02 -1.01498711e+00 7.10234702e-01 -1.29116523e+00
-6.78924263e-01 6.97342634e-01 1.33353949e-01 -1.00194108e+00
-6.09176338e-01 -8.52410972e-01 -2.48996597e-02 8.00997794e-01
-1.30102026e+00 -1.55401778e+00 -3.59071851e-01 8.75877857e-01
1.08806777e+00 -3.31456304e-01 1.31560290e+00 2.35920921e-01
2.05474913e-01 4.80593145e-01 -4.87210065e-01 -6.04139082e-02
5.34240305e-01 -1.38586080e+00 8.25589955e-01 2.03361005e-01
4.93391722e-01 2.12576650e-02 7.05491722e-01 -7.26882935e-01
-7.68015862e-01 -5.11488974e-01 1.17240512e+00 -5.52788317e-01
5.46994090e-01 -5.23604512e-01 -8.09536040e-01 7.86926568e-01
2.02388972e-01 -1.97265834e-01 9.39803898e-01 5.77635586e-01
-6.45746648e-01 2.94038117e-01 -1.14643407e+00 5.01087189e-01
9.96871114e-01 -1.34662938e+00 -1.17379379e+00 4.26377386e-01
8.54568958e-01 -9.90084708e-01 -2.21262634e-01 2.34386832e-01
7.21186876e-01 -7.41804004e-01 1.19893503e+00 -8.11115682e-01
1.84680164e-01 7.28095230e-03 -2.60593325e-01 -9.25784647e-01
-1.83838174e-01 -2.22628444e-01 -1.29183158e-01 7.38427401e-01
5.75682521e-01 -2.14946419e-02 1.37434924e+00 7.03100622e-01
-2.01325074e-01 -4.64516997e-01 -1.13070583e+00 -1.30233347e-01
-4.30991471e-01 -7.04350650e-01 1.87444136e-01 6.08040094e-01
4.11604971e-01 8.82663727e-01 -3.36276799e-01 -1.94381788e-01
2.14888230e-01 -1.47139439e-02 7.09866941e-01 -1.25607324e+00
-3.12188983e-01 9.63254198e-02 -1.58945888e-01 -1.48703170e+00
8.93319547e-02 -5.22695124e-01 8.09850767e-02 -2.02628517e+00
-5.69291562e-02 -4.02821064e-01 -8.04479867e-02 9.83893692e-01
-2.94147711e-02 3.79174978e-01 2.08499074e-01 -4.57374081e-02
-1.02104735e+00 3.85345042e-01 1.10973561e+00 -3.68658811e-01
-3.63111198e-01 -1.56983525e-01 -7.28148580e-01 3.99581671e-01
5.75054646e-01 -4.43034694e-02 -9.51917768e-01 -5.03655255e-01
-4.54512313e-02 2.91027844e-01 3.51440519e-01 -9.46777940e-01
-2.06483696e-02 2.16862336e-01 5.67610264e-01 -1.00640786e+00
1.04447985e+00 -8.62417996e-01 -2.72655964e-01 8.94882753e-02
-4.69357193e-01 2.65666246e-01 9.45212364e-01 3.08365852e-01
-2.30124816e-01 8.29750225e-02 5.12331843e-01 -2.35015064e-01
-1.32573330e+00 -1.50978729e-01 -4.95045185e-01 3.12325209e-01
6.58466816e-01 -3.83657336e-01 -4.85454232e-01 -6.67637944e-01
-1.22296500e+00 8.43414366e-02 -1.07414320e-01 7.34876454e-01
6.79511726e-01 -1.37522042e+00 -6.29754439e-02 4.75313395e-01
5.23441672e-01 -2.50715286e-01 4.10302520e-01 5.02866283e-02
-2.53444016e-01 8.13334227e-01 -3.38203311e-01 -6.90439522e-01
-1.59182680e+00 3.79150301e-01 1.85071871e-01 -5.00034153e-01
-7.15459764e-01 1.00104451e+00 4.50367987e-01 -3.74599576e-01
5.40729761e-01 -2.46068954e-01 -3.48453760e-01 2.25863636e-01
7.14391172e-01 -2.13797718e-01 1.53947949e-01 -6.32208288e-01
-2.21775010e-01 2.98166305e-01 -1.25963494e-01 -6.97276592e-01
9.36106086e-01 -3.46841216e-01 6.25785530e-01 5.74315071e-01
9.10235286e-01 -3.83164614e-01 -1.16902685e+00 -5.73422372e-01
1.04012214e-01 -4.12907064e-01 -6.09444007e-02 -1.14659977e+00
-4.75144327e-01 8.85301471e-01 1.06946194e+00 3.08249742e-01
1.01000762e+00 5.38039088e-01 8.10427725e-01 4.06670392e-01
2.12435499e-01 -1.43957293e+00 6.15823805e-01 4.77728665e-01
7.40051568e-01 -1.28715169e+00 -2.95577228e-01 -4.55317080e-01
-9.14220035e-01 5.72093010e-01 7.91935205e-01 3.65996003e-01
3.74686629e-01 2.08760947e-01 6.23416543e-01 -6.37744546e-01
-5.78655064e-01 -6.02335393e-01 5.05467176e-01 1.01732600e+00
6.43811822e-01 -7.64779141e-03 8.74241143e-02 3.87932360e-01
-2.92918444e-01 8.47536102e-02 -3.43035042e-01 1.22081828e+00
-7.40961432e-01 -1.20226634e+00 -3.17300439e-01 4.68113095e-01
-3.76337856e-01 -2.82324642e-01 -1.36453420e-01 8.29222679e-01
-1.43572047e-01 9.57191646e-01 4.58049029e-01 -3.29745352e-01
6.03610039e-01 7.71453679e-01 4.46397454e-01 -9.52635229e-01
-2.92606205e-01 1.69446796e-01 6.10715449e-01 -7.44618654e-01
-3.08758885e-01 -5.96043706e-01 -1.35526419e+00 2.67164439e-01
3.70661393e-02 1.55325904e-01 6.80051208e-01 1.18692362e+00
2.72059411e-01 7.36022949e-01 -1.28362417e-01 -6.27116561e-01
-1.62724946e-02 -7.50885427e-01 -4.72005516e-01 6.19054854e-01
3.32778394e-01 -9.76917684e-01 2.97990710e-01 1.56225920e-01] | [12.539170265197754, 7.314788818359375] |
1fa181bb-06c2-4f9b-918e-700f44b64799 | multimodal-generation-of-novel-action | 2208.01910 | null | https://arxiv.org/abs/2208.01910v1 | https://arxiv.org/pdf/2208.01910v1.pdf | Multimodal Generation of Novel Action Appearances for Synthetic-to-Real Recognition of Activities of Daily Living | Domain shifts, such as appearance changes, are a key challenge in real-world applications of activity recognition models, which range from assistive robotics and smart homes to driver observation in intelligent vehicles. For example, while simulations are an excellent way of economical data collection, a Synthetic-to-Real domain shift leads to a > 60% drop in accuracy when recognizing activities of Daily Living (ADLs). We tackle this challenge and introduce an activity domain generation framework which creates novel ADL appearances (novel domains) from different existing activity modalities (source domains) inferred from video training data. Our framework computes human poses, heatmaps of body joints, and optical flow maps and uses them alongside the original RGB videos to learn the essence of source domains in order to generate completely new ADL domains. The model is optimized by maximizing the distance between the existing source appearances and the generated novel appearances while ensuring that the semantics of an activity is preserved through an additional classification loss. While source data multimodality is an important concept in this design, our setup does not rely on multi-sensor setups, (i.e., all source modalities are inferred from a single video only.) The newly created activity domains are then integrated in the training of the ADL classification networks, resulting in models far less susceptible to changes in data distributions. Extensive experiments on the Synthetic-to-Real benchmark Sims4Action demonstrate the potential of the domain generation paradigm for cross-domain ADL recognition, setting new state-of-the-art results. Our code is publicly available at https://github.com/Zrrr1997/syn2real_DG | ['Rainer Stiefelhagen', 'Alina Roitberg', 'David Schneider', 'Zdravko Marinov'] | 2022-08-03 | null | null | null | null | ['multimodal-generation'] | ['natural-language-processing'] | [ 3.10405225e-01 2.14209333e-01 -2.50933319e-01 -4.09806281e-01
-5.83195090e-01 -5.77975869e-01 8.51965487e-01 -3.54716897e-01
-3.95887077e-01 8.95048797e-01 2.20629022e-01 2.57876694e-01
8.46837163e-02 -5.58449030e-01 -9.73958433e-01 -8.36037397e-01
-8.05396140e-02 5.67588389e-01 1.89906999e-01 -2.22918093e-01
-2.68468380e-01 6.59633219e-01 -1.87434673e+00 2.02093929e-01
6.00893080e-01 1.09710228e+00 -7.45099932e-02 6.80001080e-01
1.64940819e-01 7.50122607e-01 -5.30630410e-01 -1.17047705e-01
4.84901994e-01 -4.98469323e-01 -5.52138686e-01 3.34575653e-01
5.99732220e-01 -2.23513871e-01 -5.62068045e-01 7.04721451e-01
4.24380600e-01 2.78478622e-01 5.81755757e-01 -1.73924530e+00
-2.97953874e-01 -2.01743916e-01 -2.78743297e-01 -2.13187858e-01
6.93404853e-01 4.43744421e-01 6.30968451e-01 -6.64550424e-01
9.59398389e-01 1.17876744e+00 4.93770331e-01 8.80226016e-01
-1.29113066e+00 -6.53576791e-01 4.23209034e-02 4.30375904e-01
-1.22280455e+00 -7.00930178e-01 9.03735578e-01 -4.21466947e-01
6.43694878e-01 1.29264101e-01 1.01623499e+00 1.65032268e+00
1.12935178e-01 8.85093272e-01 1.08979380e+00 -1.10530704e-01
3.61628860e-01 1.98464915e-01 -2.29348406e-01 6.73874021e-01
2.51833171e-01 8.05768296e-02 -1.02360940e+00 1.29845396e-01
6.48745775e-01 -6.03446327e-02 -3.22063208e-01 -1.01013207e+00
-1.48398340e+00 5.03120959e-01 2.57902056e-01 2.83253659e-02
-3.20308417e-01 1.50436237e-01 2.10058689e-01 3.52468163e-01
2.55926162e-01 2.44411066e-01 -4.22003418e-01 -3.48047346e-01
-6.20289087e-01 3.71445119e-01 7.16839731e-01 8.16806614e-01
9.81926918e-01 -9.21902508e-02 -3.81327197e-02 7.77863562e-01
1.55108914e-01 7.01269746e-01 4.91851717e-01 -1.30725312e+00
4.39198941e-01 7.40549147e-01 1.23219974e-01 -7.96391726e-01
-4.84665781e-01 -2.48638123e-01 -6.28649592e-01 5.22610068e-01
8.23989630e-01 -4.51700836e-02 -9.26372588e-01 1.95572901e+00
5.93331695e-01 2.38727108e-01 1.05026849e-01 8.21903825e-01
5.53631902e-01 4.91401374e-01 -7.60383382e-02 1.29184708e-01
1.15253246e+00 -8.54821801e-01 -5.11382163e-01 -5.02413332e-01
5.27206302e-01 -2.59281725e-01 9.72254932e-01 3.79932314e-01
-9.18322444e-01 -5.31793177e-01 -1.12217689e+00 7.26362169e-02
-4.22547013e-01 9.96352080e-03 3.99102420e-01 5.69520891e-01
-9.05348063e-01 4.64924395e-01 -1.00484550e+00 -6.61733270e-01
5.00305057e-01 3.32807600e-01 -9.41162109e-01 -1.79898664e-01
-1.07212269e+00 1.03086281e+00 3.95803213e-01 -9.58662704e-02
-8.81789923e-01 -7.72492528e-01 -1.13192761e+00 -5.40028572e-01
4.80654836e-01 -6.73137426e-01 9.98444438e-01 -1.32207942e+00
-1.48564756e+00 9.10231054e-01 -1.35231152e-01 -4.59519565e-01
8.45990181e-01 -4.47714478e-02 -5.14941096e-01 1.68270558e-01
6.35577226e-03 1.10289097e+00 9.82418060e-01 -1.16672814e+00
-6.26225173e-01 -2.24265173e-01 4.41954751e-03 3.03905010e-01
-1.01289712e-01 -5.34832060e-01 -4.18367326e-01 -4.37229425e-01
-1.49727076e-01 -1.25062823e+00 2.07442015e-01 5.96454203e-01
3.42037566e-02 1.69741325e-02 9.17217672e-01 -7.15741098e-01
5.21476924e-01 -2.07875109e+00 3.54659408e-01 1.92507625e-01
1.24059595e-01 6.86628073e-02 -2.72977382e-01 9.30108055e-02
-2.93975056e-04 -5.48455417e-01 -4.11414266e-01 -4.73181516e-01
1.06785171e-01 5.73809206e-01 1.87015846e-01 7.56820142e-01
4.56864446e-01 9.15243387e-01 -9.83205736e-01 -3.43232244e-01
5.64166605e-01 7.04896450e-01 -4.10067111e-01 2.25175649e-01
-2.81971961e-01 9.27389443e-01 1.45271085e-02 6.79752588e-01
4.71413314e-01 4.20149043e-02 8.04056898e-02 -3.59691888e-01
2.59876341e-01 1.22308418e-01 -1.20656455e+00 2.07210612e+00
-4.56141174e-01 6.43333972e-01 -1.88743234e-01 -1.01531410e+00
6.50357723e-01 1.38456553e-01 9.34855580e-01 -9.90501523e-01
-3.94559056e-02 1.31062791e-01 -8.48169103e-02 -3.46812218e-01
2.59729177e-01 7.08277449e-02 -2.09895641e-01 2.33452946e-01
2.92909950e-01 -1.62318721e-02 3.74972433e-01 1.21293012e-02
1.08567882e+00 6.79363370e-01 2.09734052e-01 2.58371998e-02
3.73021185e-01 6.77052364e-02 6.89926028e-01 3.85987312e-01
-4.31131750e-01 7.12386489e-01 4.63729054e-01 -4.14905488e-01
-1.15082550e+00 -1.39296710e+00 1.35773674e-01 7.33511984e-01
2.18614757e-01 -9.88405868e-02 -7.69611776e-01 -8.08132112e-01
7.54462853e-02 4.88503307e-01 -7.98037291e-01 -4.73545521e-01
-6.62485301e-01 -4.61862803e-01 7.50113010e-01 5.28326750e-01
6.64769113e-01 -1.02554142e+00 -9.93787706e-01 4.38549630e-02
-4.34300274e-01 -1.37511921e+00 -3.15049380e-01 -4.14878875e-02
-5.81481099e-01 -1.32013321e+00 -8.64436030e-01 -4.18421805e-01
6.67964339e-01 -8.75216424e-02 1.05982530e+00 -3.33264500e-01
-4.05997157e-01 7.65481174e-01 -2.86269635e-01 -8.34170505e-02
-5.87913811e-01 -1.59036070e-01 2.92551905e-01 4.50208724e-01
2.03692526e-01 -6.46263599e-01 -7.48423576e-01 4.26422566e-01
-8.89175534e-01 2.54582763e-01 5.70268273e-01 7.42475271e-01
4.72442180e-01 -3.02685022e-01 3.65576535e-01 -3.68780315e-01
1.13038465e-01 -3.99896830e-01 -4.72683221e-01 9.92818847e-02
-3.83603871e-01 2.44439915e-01 4.45633352e-01 -6.82901263e-01
-1.15651608e+00 4.03524011e-01 2.07118830e-03 -5.44511616e-01
-5.24300516e-01 3.63253132e-02 -3.51490140e-01 -8.55942219e-02
7.03483105e-01 2.39945680e-01 4.44451123e-01 -4.06792164e-01
5.02452016e-01 2.78959423e-01 7.38111377e-01 -6.52594984e-01
8.59656453e-01 8.19848597e-01 2.18611538e-01 -1.01671863e+00
-3.88692200e-01 -4.14349645e-01 -6.50341749e-01 -5.72997332e-01
8.08045089e-01 -9.33344543e-01 -4.91593271e-01 8.08221936e-01
-8.81603658e-01 -6.32990122e-01 -5.28838992e-01 4.01585072e-01
-7.91913033e-01 3.21326107e-01 -7.50732794e-02 -6.10437393e-01
1.69550881e-01 -9.99383926e-01 1.23507047e+00 2.18181297e-01
-4.27112341e-01 -9.18865263e-01 2.39588350e-01 5.06241083e-01
1.26096398e-01 7.08765209e-01 6.23653769e-01 -3.14396590e-01
-4.97744560e-01 -1.02643929e-01 1.01173662e-01 5.43815553e-01
3.51760894e-01 -2.13734999e-01 -1.09387207e+00 -2.81989485e-01
-4.53864068e-01 -3.64631772e-01 5.52587569e-01 1.23215698e-01
8.15214813e-01 -2.84742147e-01 -2.69075155e-01 5.62306523e-01
9.99676287e-01 1.67931944e-01 8.76380265e-01 2.73288935e-01
7.82243967e-01 8.17744613e-01 6.69493556e-01 4.38306183e-01
4.77158546e-01 9.77560043e-01 4.31661367e-01 -1.02530599e-01
-5.31422079e-01 -2.74083942e-01 8.06846082e-01 2.72075474e-01
-4.30513099e-02 1.37736481e-02 -7.94501245e-01 6.81559205e-01
-1.85523474e+00 -9.68370736e-01 1.85439125e-01 2.47778654e+00
7.05155611e-01 -2.38482058e-02 4.67287958e-01 3.68990675e-02
3.88023406e-01 9.89671499e-02 -8.21236074e-01 3.02730557e-02
-2.10056558e-01 1.91494524e-01 5.46810865e-01 3.42109114e-01
-1.04634881e+00 5.34459710e-01 5.22847462e+00 5.23071766e-01
-1.11700881e+00 1.58367112e-01 3.18829894e-01 -4.73506749e-01
-5.97906709e-02 -1.90444633e-01 -5.30310273e-01 6.04636610e-01
8.46885443e-01 1.81422547e-01 5.34615874e-01 7.53077269e-01
2.35445425e-01 -2.67231435e-01 -1.24519038e+00 1.20905626e+00
1.48910642e-01 -1.08466196e+00 -5.88722117e-02 1.34385496e-01
6.04423642e-01 1.34560362e-01 -1.51210114e-01 1.42042249e-01
9.72071663e-02 -7.74658740e-01 9.84685242e-01 6.79587007e-01
9.07552302e-01 -5.84584773e-01 3.70752901e-01 2.22747967e-01
-1.19693732e+00 3.80046628e-02 5.23658991e-02 9.78527889e-02
1.72472030e-01 4.39249784e-01 -7.34862447e-01 4.30160016e-01
7.71041751e-01 9.08405066e-01 -6.93869472e-01 7.76880264e-01
-9.24797654e-02 2.77885258e-01 -5.20375073e-01 1.80011347e-01
-1.80656359e-01 -1.79511145e-01 7.50433326e-01 1.04690254e+00
2.81960994e-01 -5.37981510e-01 -8.28963220e-02 8.32536578e-01
-9.26318578e-03 -4.26008999e-01 -7.83627272e-01 7.81370550e-02
2.32866123e-01 1.08405900e+00 -3.61514509e-01 -2.13932440e-01
-3.78545374e-01 1.25880277e+00 9.50227827e-02 4.05099213e-01
-1.09133899e+00 -3.23953554e-02 1.31217742e+00 2.59133071e-01
1.14024132e-01 -2.43922114e-01 3.93363498e-02 -1.33566010e+00
3.48958075e-01 -1.02495456e+00 2.70544946e-01 -7.70931959e-01
-1.05143476e+00 1.53596461e-01 4.05923784e-01 -1.48790526e+00
-6.16009176e-01 -7.12153018e-01 -1.89437211e-01 4.40930694e-01
-1.41549850e+00 -1.29167891e+00 -7.90690601e-01 8.09492409e-01
6.05973542e-01 -7.40334168e-02 5.56353569e-01 4.62712377e-01
-2.97374725e-01 5.46253741e-01 -1.31921489e-02 1.28501192e-01
8.42910707e-01 -1.08356261e+00 4.40716535e-01 7.97556520e-01
1.85484216e-01 2.29842693e-01 7.60768652e-01 -4.47009981e-01
-1.45744669e+00 -1.17508066e+00 3.96163344e-01 -6.22196496e-01
4.12726700e-01 -5.85465968e-01 -6.64468288e-01 5.99768579e-01
-7.19055459e-02 2.21089706e-01 3.30641717e-01 -4.25670892e-01
-3.54653507e-01 -3.85541320e-01 -1.20203912e+00 6.75437033e-01
1.35392058e+00 -3.19644243e-01 -3.10177118e-01 7.39997923e-02
3.15465629e-01 -4.64126140e-01 -8.77671063e-01 4.33275610e-01
9.97170925e-01 -1.02174413e+00 1.08677602e+00 -3.80841136e-01
1.22087680e-01 -6.14061236e-01 -1.69749960e-01 -1.40050328e+00
1.41704410e-01 -4.77958828e-01 -4.86120403e-01 9.31285918e-01
2.14819431e-01 -8.88647974e-01 9.74910736e-01 5.79745054e-01
7.47662261e-02 -4.57874209e-01 -1.19405091e+00 -9.86880183e-01
-2.09668934e-01 -4.97126877e-01 5.95932245e-01 7.83610463e-01
-2.99126565e-01 -1.29473885e-03 -4.69148904e-01 -5.69089800e-02
8.09032023e-01 -3.67846251e-01 1.18769217e+00 -1.13615584e+00
-3.78895462e-01 -2.30509073e-01 -7.46077478e-01 -9.57725286e-01
2.41442740e-01 -5.97327709e-01 9.47255343e-02 -1.41958117e+00
-5.32446317e-02 -4.17051941e-01 -2.12926105e-01 6.54196262e-01
3.35526705e-01 4.83016908e-01 2.38639176e-01 2.69187149e-02
-4.34511453e-01 6.39145553e-01 1.05729640e+00 -3.59227538e-01
-1.99795172e-01 -1.30901411e-01 -2.51388967e-01 6.02161646e-01
7.45643914e-01 -3.78934234e-01 -6.33370221e-01 -1.33608714e-01
4.56742346e-02 -1.46618009e-01 9.03708458e-01 -1.48808622e+00
-2.20828801e-02 -1.72337577e-01 4.14043516e-01 -2.14293033e-01
8.03986609e-01 -9.95930314e-01 5.67594290e-01 4.11304235e-01
9.71584860e-03 -2.06420079e-01 1.45776361e-01 6.03833675e-01
9.44260322e-03 2.38326490e-01 8.34141552e-01 -8.83166119e-02
-1.10096633e+00 2.55925477e-01 -2.52493739e-01 7.87547678e-02
1.12304258e+00 -7.31871068e-01 -2.43323311e-01 -6.05953276e-01
-7.11803555e-01 7.04685645e-03 7.84246325e-01 8.60775292e-01
3.98651123e-01 -1.48244357e+00 -5.39510608e-01 4.78357047e-01
5.29677987e-01 1.54736847e-01 1.70100808e-01 9.43527281e-01
-4.66727883e-01 2.21346036e-01 -5.64505219e-01 -8.12412858e-01
-1.22751474e+00 4.09750640e-02 6.49444401e-01 -5.82865775e-02
-6.60093307e-01 3.85932177e-01 3.06193799e-01 -5.87872326e-01
2.37107098e-01 -3.63917917e-01 1.55084580e-01 4.43927310e-02
2.47078583e-01 5.25279045e-01 1.44140095e-01 -9.29811001e-01
-5.56128144e-01 5.66443086e-01 3.12993407e-01 -2.36418381e-01
1.07352245e+00 -5.62013127e-02 3.84580612e-01 4.18413609e-01
1.27594304e+00 -2.64014602e-01 -1.88582838e+00 -1.56565271e-02
-2.07010180e-01 -4.01399791e-01 -3.68671864e-01 -9.53283131e-01
-1.07022715e+00 5.52330852e-01 8.61048579e-01 -4.12641138e-01
1.17199886e+00 1.09551363e-01 8.20918977e-01 3.45316410e-01
6.13013268e-01 -1.27961814e+00 3.21677297e-01 1.98035851e-01
8.12402487e-01 -1.17204976e+00 -1.08787060e-01 -9.93195251e-02
-8.16169381e-01 7.98239529e-01 6.51965022e-01 3.61635648e-02
2.62281597e-01 6.98381960e-02 1.05165161e-01 6.52573928e-02
-4.79109257e-01 -2.69167244e-01 3.15077096e-01 1.05315220e+00
-7.80779645e-02 -1.04371281e-02 1.60029426e-01 3.68436351e-02
-1.44364955e-02 1.14778787e-01 3.76078427e-01 1.00052416e+00
-2.17848793e-02 -1.25455451e+00 -4.79833543e-01 2.79479653e-01
5.22429757e-02 5.16932249e-01 -4.15089875e-01 9.67998385e-01
4.41458762e-01 7.19588757e-01 6.94001839e-02 -3.71381015e-01
4.57605302e-01 2.52705216e-01 7.93636084e-01 -2.69533694e-01
1.89534444e-02 -4.25882936e-01 3.02129358e-01 -9.72934902e-01
-7.14285731e-01 -8.75206351e-01 -1.19797528e+00 -1.51813194e-01
2.48180717e-01 -2.83012509e-01 6.90925241e-01 9.71607864e-01
4.79897708e-01 4.33823615e-01 3.17324758e-01 -1.12170613e+00
-1.12858757e-01 -6.49145544e-01 -4.13651019e-01 9.09012794e-01
5.78152955e-01 -1.18273640e+00 -3.14249635e-01 3.39391649e-01] | [8.078276634216309, 0.5704615116119385] |
daaa86a0-6a23-4217-afdf-d2f1247980df | teaching-language-models-to-support-answers | 2203.11147 | null | https://arxiv.org/abs/2203.11147v1 | https://arxiv.org/pdf/2203.11147v1.pdf | Teaching language models to support answers with verified quotes | Recent large language models often answer factual questions correctly. But users can't trust any given claim a model makes without fact-checking, because language models can hallucinate convincing nonsense. In this work we use reinforcement learning from human preferences (RLHP) to train "open-book" QA models that generate answers whilst also citing specific evidence for their claims, which aids in the appraisal of correctness. Supporting evidence is drawn from multiple documents found via a search engine, or from a single user-provided document. Our 280 billion parameter model, GopherCite, is able to produce answers with high quality supporting evidence and abstain from answering when unsure. We measure the performance of GopherCite by conducting human evaluation of answers to questions in a subset of the NaturalQuestions and ELI5 datasets. The model's response is found to be high-quality 80\% of the time on this Natural Questions subset, and 67\% of the time on the ELI5 subset. Abstaining from the third of questions for which it is most unsure improves performance to 90\% and 80\% respectively, approaching human baselines. However, analysis on the adversarial TruthfulQA dataset shows why citation is only one part of an overall strategy for safety and trustworthiness: not all claims supported by evidence are true. | ['Nat McAleese', 'Geoffrey Irving', 'Lucy Campbell-Gillingham', 'Susannah Young', 'Mia Glaese', 'Martin Chadwick', 'Francis Song', 'John Aslanides', 'Vladimir Mikulik', 'Maja Trebacz', 'Jacob Menick'] | 2022-03-21 | null | null | null | null | ['natural-questions'] | ['miscellaneous'] | [-3.55076820e-01 8.66671860e-01 -2.50722706e-01 -1.89159989e-01
-1.56700528e+00 -1.05243158e+00 7.13065147e-01 3.24140728e-01
-5.12517452e-01 1.25685108e+00 5.13780951e-01 -9.30873990e-01
-8.15543905e-02 -9.48642373e-01 -1.02513957e+00 -1.19625621e-01
6.32872641e-01 7.58723497e-01 1.85705721e-01 -5.46137035e-01
5.38544178e-01 -4.09558564e-02 -6.45240545e-01 6.65187061e-01
1.11733055e+00 7.25123346e-01 -5.52997589e-01 6.33574486e-01
-1.04058404e-02 1.59874046e+00 -9.10634339e-01 -1.46715403e+00
9.63382125e-02 -3.63973707e-01 -1.30729282e+00 -6.35448217e-01
8.35632324e-01 -5.25848508e-01 -1.55115217e-01 1.08620179e+00
4.03728157e-01 -3.00120413e-01 8.95340562e-01 -1.36368811e+00
-1.19600737e+00 8.63656878e-01 -1.44657210e-01 4.00766939e-01
7.17957199e-01 7.41753459e-01 1.38039494e+00 -6.30350888e-01
8.05661380e-01 1.36564827e+00 4.99893010e-01 6.24235451e-01
-1.14298129e+00 -9.30088460e-01 -3.95849109e-01 3.81392121e-01
-6.70114160e-01 -4.63280529e-01 4.32705373e-01 -4.39808428e-01
8.41372848e-01 2.46381119e-01 3.52819376e-02 1.60572290e+00
6.10833943e-01 5.23852646e-01 1.59507024e+00 -1.74275100e-01
3.69473308e-01 5.25032818e-01 3.61971885e-01 5.48412085e-01
5.03664315e-01 3.66058707e-01 -5.65128624e-01 -8.29480112e-01
-4.69001867e-02 -6.60566151e-01 -2.58681417e-01 1.84838355e-01
-1.00994349e+00 1.41552758e+00 6.00282669e-01 1.44284397e-01
-6.29934728e-01 1.32925406e-01 1.70754611e-01 6.14318907e-01
1.48219675e-01 9.17229116e-01 -3.44684154e-01 -4.11113799e-02
-8.40414762e-01 8.60758603e-01 1.26465213e+00 4.27767366e-01
2.93312103e-01 -1.39470845e-01 -4.40896660e-01 4.28940654e-01
2.50495523e-01 1.05913627e+00 3.68834764e-01 -1.58224356e+00
6.30329311e-01 3.62951934e-01 7.07913816e-01 -1.27383900e+00
-6.55873790e-02 -3.82735670e-01 -4.52403247e-01 2.75708139e-01
6.46215618e-01 -3.37138593e-01 -4.02627498e-01 1.65834463e+00
-8.57816041e-02 -3.71940970e-01 4.60472316e-01 9.01800811e-01
9.69984353e-01 7.53789186e-01 4.23417538e-01 2.13035326e-02
1.50419033e+00 -3.16583902e-01 -6.49113595e-01 -1.49861783e-01
5.09582162e-01 -5.40537596e-01 1.29945934e+00 6.48427308e-01
-1.22709298e+00 -1.08475290e-01 -1.16539001e+00 -2.63556749e-01
-1.47850126e-01 -4.76838768e-01 1.95354804e-01 4.98137265e-01
-7.02913940e-01 3.37234288e-01 1.76386133e-01 3.85595858e-01
6.88184023e-01 -1.69021681e-01 -2.55890936e-01 -1.49607480e-01
-1.99082196e+00 1.56876910e+00 8.67050961e-02 -4.64274794e-01
-1.15258002e+00 -6.34532511e-01 -5.48236310e-01 2.83234417e-01
6.68080330e-01 -7.04457164e-01 1.57438111e+00 -5.71561813e-01
-1.03545749e+00 8.43785226e-01 -4.72749993e-02 -1.05862582e+00
8.24361265e-01 -2.87054956e-01 -7.43876398e-01 4.47305322e-01
6.47829294e-01 5.42982578e-01 7.67867863e-01 -1.39508426e+00
-2.89870411e-01 -1.48802102e-01 4.59121376e-01 -2.40196317e-01
1.51537836e-01 1.16653405e-01 6.35169685e-01 -4.81487125e-01
-5.57364881e-01 -7.40446687e-01 1.01072483e-01 -1.32950097e-01
-5.73660314e-01 -6.65684044e-01 5.03075778e-01 -8.93204212e-01
9.18097556e-01 -1.67642069e+00 -6.68337762e-01 2.73578644e-01
3.80096376e-01 2.62369722e-01 -1.57933697e-01 4.43508506e-01
3.66337486e-02 6.51866794e-01 -7.62328655e-02 4.01669979e-01
3.83033544e-01 1.73287600e-01 -1.18451738e+00 1.45786166e-01
3.15126091e-01 1.20387495e+00 -1.04031456e+00 -5.79975009e-01
-4.04781580e-01 -6.63003325e-02 -4.80504125e-01 9.80373248e-02
-6.38880730e-01 -2.58463800e-01 -2.71741807e-01 3.04472744e-01
3.12214822e-01 -5.44603527e-01 -3.93256605e-01 1.33336252e-02
4.04200912e-01 8.13850462e-01 -6.91852808e-01 9.24300313e-01
-3.30638111e-01 4.59780574e-01 -1.48973346e-01 -3.04422200e-01
7.04728305e-01 6.94247842e-01 -5.02374411e-01 -9.76778269e-01
1.21428341e-01 4.57808435e-01 1.57901421e-01 -7.08334386e-01
3.52332473e-01 -7.12416410e-01 -2.10621223e-01 9.89712358e-01
-6.62888661e-02 -6.36265099e-01 -1.85979772e-02 1.02092588e+00
1.11008584e+00 -3.57735157e-01 7.15623647e-02 -2.56880492e-01
4.24156070e-01 5.45428514e-01 2.27738529e-01 1.13750660e+00
-2.23316044e-01 2.25130901e-01 7.29589939e-01 -4.23098356e-01
-1.05577183e+00 -1.16041589e+00 3.88015099e-02 7.16121376e-01
-1.55525401e-01 -1.17017969e-01 -6.52496815e-01 -1.10282660e+00
8.43292773e-02 1.95614278e+00 -7.26363242e-01 -1.80262029e-01
-6.23588003e-02 -9.55499485e-02 9.40550864e-01 1.36324793e-01
4.91831243e-01 -1.33645034e+00 -8.93071651e-01 1.34142056e-01
-7.87920594e-01 -7.80140162e-01 -2.82777399e-01 -1.29053086e-01
-2.97794968e-01 -1.41383445e+00 -4.30856615e-01 4.71851695e-03
-1.14709930e-03 -1.69432729e-01 1.43490410e+00 3.41716021e-01
5.05141377e-01 2.40731493e-01 -1.07382059e-01 -6.77270472e-01
-1.11146557e+00 -4.15308416e-01 -8.79379213e-02 -4.69953030e-01
6.42934859e-01 -1.52504221e-01 -4.08887714e-01 1.59761146e-01
-9.59397972e-01 -5.01104414e-01 4.26579177e-01 8.37169588e-01
5.07750250e-02 -2.79534608e-01 1.11951852e+00 -1.29097378e+00
1.47145319e+00 -9.52185750e-01 -3.38168234e-01 4.76782173e-01
-8.98458421e-01 2.55640984e-01 7.27475524e-01 -2.99435079e-01
-1.02236259e+00 -9.33898389e-01 -1.67099401e-01 -1.23693056e-01
-3.11033502e-02 4.66096371e-01 1.48902699e-01 3.08473706e-01
1.57265770e+00 -2.47940004e-01 -5.46090864e-02 1.93543360e-01
6.48019910e-01 6.51207328e-01 7.08671093e-01 -8.59610617e-01
9.52451289e-01 1.52120799e-01 -4.27870929e-01 -1.87239945e-01
-1.33779895e+00 9.85959768e-02 2.84865171e-01 1.47016659e-01
5.38398623e-01 -6.11168385e-01 -1.19537342e+00 -3.15336287e-01
-1.47030067e+00 -2.77373195e-01 -2.57399082e-01 1.26585856e-01
-5.16560197e-01 3.23954016e-01 -5.81592619e-01 -8.81920993e-01
-6.00337505e-01 -8.09035897e-01 4.49162304e-01 -5.53493053e-02
-8.37666333e-01 -6.72778904e-01 2.69579947e-01 9.16948617e-01
4.06972796e-01 6.68676570e-02 1.43044329e+00 -1.21764302e+00
-1.85729921e-01 -3.66511077e-01 -2.09225804e-01 4.29291666e-01
-4.37678397e-01 -1.29278168e-01 -1.00123775e+00 1.75204486e-01
3.06746036e-01 -1.16201282e+00 5.90329647e-01 -6.96858987e-02
6.37489259e-01 -1.18199825e+00 2.47135982e-01 -5.81263721e-01
1.22206402e+00 -4.04113978e-02 7.12763071e-01 3.67436439e-01
-1.89018339e-01 9.18173850e-01 3.59456271e-01 -2.64906026e-02
5.93329370e-01 9.54057500e-02 4.52528834e-01 3.63443851e-01
1.39487594e-01 -6.95364654e-01 3.59384209e-01 -1.28386617e-01
2.07870543e-01 -3.81160229e-01 -8.31832111e-01 7.01165855e-01
-1.38746059e+00 -1.56460667e+00 -2.74837941e-01 1.86749041e+00
1.26447046e+00 6.90553010e-01 -2.16889428e-03 3.60381678e-02
2.46569574e-01 -7.25533888e-02 -5.93972683e-01 -7.86428630e-01
-4.55869079e-01 2.95530647e-01 1.81564271e-01 9.00922418e-01
-3.98653597e-01 7.00004578e-01 6.30302525e+00 8.89905930e-01
-5.37142992e-01 2.42096201e-01 9.85477686e-01 6.55198768e-02
-1.20256841e+00 2.06793115e-01 -2.28332832e-01 6.05071843e-01
1.32194149e+00 -3.72449130e-01 1.65830165e-01 7.75467634e-01
8.82139876e-02 -3.37646276e-01 -9.48072731e-01 4.93564308e-01
1.37607783e-01 -1.69611287e+00 3.86851907e-01 -1.16962261e-01
7.40697682e-01 -9.73495543e-02 1.63593695e-01 5.88557780e-01
1.08913553e+00 -1.47517991e+00 1.05582607e+00 5.97217262e-01
5.12156844e-01 -7.91937351e-01 1.03186309e+00 8.87653947e-01
3.19761038e-01 -1.36615619e-01 -2.83928275e-01 1.09386910e-02
2.35503346e-01 4.89973903e-01 -1.02753735e+00 1.66871369e-01
6.52854621e-01 -8.65989178e-02 -6.54771984e-01 3.45491737e-01
-8.09672892e-01 1.16257668e+00 -1.81995630e-01 -4.47864026e-01
5.33245504e-01 3.10142279e-01 4.02639031e-01 9.48694050e-01
-1.64483428e-01 6.68310642e-01 -3.78475249e-01 1.31620419e+00
-6.43224299e-01 -2.00965494e-01 -7.41994977e-01 -2.62924582e-02
4.87126827e-01 9.31452990e-01 2.68318415e-01 -6.09274268e-01
-2.09546998e-01 5.13068020e-01 2.61306196e-01 2.85612375e-01
-6.33343697e-01 -2.53424853e-01 -2.00922236e-01 2.10037217e-01
-2.56507427e-01 3.54594827e-01 -3.42147678e-01 -9.47544992e-01
-3.32175456e-02 -1.63362610e+00 6.27068460e-01 -1.58306324e+00
-1.70333600e+00 6.91271544e-01 -8.42097774e-02 -5.84928811e-01
-5.87614715e-01 -5.23785293e-01 -6.83102190e-01 1.05868006e+00
-1.46866763e+00 -8.34937572e-01 2.65442729e-01 5.81966460e-01
1.33702815e-01 -5.66010503e-03 6.81483865e-01 -3.81489575e-01
3.22535455e-01 5.27969897e-01 -4.89112437e-01 2.82087892e-01
9.13014650e-01 -1.19809330e+00 2.06302077e-01 5.86959600e-01
4.25689131e-01 1.03384185e+00 1.11260831e+00 -8.99959922e-01
-8.55374157e-01 -4.00262296e-01 1.38984025e+00 -1.27412868e+00
9.74098980e-01 3.17088217e-01 -1.30913210e+00 5.04494131e-01
8.32184613e-01 -5.13323903e-01 7.83307672e-01 -3.36553231e-02
-9.28459167e-01 2.97639549e-01 -1.62833703e+00 7.68776894e-01
2.34767348e-01 -8.79051208e-01 -1.62261271e+00 4.90682125e-01
8.01372170e-01 3.89006846e-02 -4.97272879e-01 1.31719867e-02
3.59625667e-01 -8.39503467e-01 6.59056187e-01 -1.23474145e+00
1.07395577e+00 -2.15457708e-01 -1.63870007e-01 -1.37359393e+00
-1.43991664e-01 -6.32698953e-01 -1.30966976e-01 8.26117039e-01
9.53478813e-01 -6.09839618e-01 4.54437315e-01 1.11604035e+00
3.63199055e-01 -4.93243873e-01 -1.07925260e+00 -3.21660936e-01
7.37367868e-01 -4.66820687e-01 3.26861918e-01 9.89538372e-01
2.40045175e-01 7.33956039e-01 -2.57939667e-01 8.53605643e-02
8.53767395e-01 2.01505333e-01 5.34668744e-01 -9.17607486e-01
-2.17995048e-01 -1.68704271e-01 4.52136308e-01 -5.79744756e-01
2.82807738e-01 -7.96677649e-01 -1.74169630e-01 -1.50319004e+00
3.74018282e-01 -6.00442924e-02 1.95603013e-01 5.43784142e-01
-2.81218797e-01 7.56646916e-02 1.95550546e-01 1.21504955e-01
-4.01251495e-01 3.36871862e-01 1.19309020e+00 -4.23660755e-01
5.30857325e-01 -9.77393538e-02 -1.32331872e+00 7.65485466e-01
8.15182388e-01 -6.45841777e-01 -2.31824204e-01 -2.51921773e-01
9.94981349e-01 3.66854012e-01 1.02772117e+00 -5.39656222e-01
3.68255556e-01 -3.31871510e-01 2.91607261e-01 -4.25412238e-01
-2.08266638e-02 -6.90577328e-01 -2.67026395e-01 6.89295709e-01
-1.03496599e+00 1.89243361e-01 7.91683942e-02 6.35510087e-01
2.83644591e-02 -6.60351992e-01 6.14760816e-01 -4.70068187e-01
-3.96322794e-02 -2.08850890e-01 -3.85883808e-01 8.28928530e-01
3.78966272e-01 2.04057917e-01 -9.22001779e-01 -1.06238842e+00
-4.64667290e-01 3.94253910e-01 1.74022689e-01 1.29843652e-01
6.85849905e-01 -1.05568206e+00 -1.32015538e+00 -5.57590127e-01
1.32243693e-01 -5.00455618e-01 8.37062001e-02 3.49720657e-01
-2.70653009e-01 5.76031744e-01 2.57583689e-02 8.66451189e-02
-5.67228913e-01 7.69988537e-01 2.37429217e-01 -3.23924154e-01
-1.33536458e-01 6.73172057e-01 -7.36038461e-02 -2.88618326e-01
-1.33771807e-01 -1.53029650e-01 -2.67958999e-01 -2.41183579e-01
6.37557924e-01 4.17409927e-01 -1.00708127e-01 -4.23374712e-01
-4.32380289e-02 -7.20029473e-02 -1.43953115e-01 -6.36273980e-01
9.75990593e-01 2.03278005e-01 -1.04094937e-01 1.60918981e-01
6.47064030e-01 5.95176041e-01 -7.20266998e-01 -1.59929499e-01
-9.98694543e-03 -3.82087678e-01 -2.20329210e-01 -1.84592164e+00
-3.46030295e-01 1.02937090e+00 -3.48285809e-02 4.79941219e-01
2.24215254e-01 5.78916632e-02 9.31774259e-01 8.00722241e-01
4.84724790e-01 -9.59788322e-01 3.48719865e-01 2.79563040e-01
1.60928082e+00 -1.38471496e+00 -7.78158307e-02 2.36085460e-01
-1.24866509e+00 7.77609885e-01 6.45661831e-01 -2.51334250e-01
1.17708147e-01 -1.58360630e-01 3.16592246e-01 -5.82741737e-01
-1.09821129e+00 6.46123886e-01 2.99812496e-01 2.90096730e-01
9.96943042e-02 9.23126116e-02 -4.31328267e-01 1.15097642e+00
-5.80084145e-01 -1.30676836e-01 8.97043705e-01 2.53718048e-01
-5.68151534e-01 -5.30307770e-01 -5.23087263e-01 5.50073922e-01
-8.51587653e-01 -4.00284648e-01 -7.74091423e-01 5.40982246e-01
-1.57101214e-01 1.51093161e+00 -6.96502209e-01 -1.17505975e-01
2.48157039e-01 2.73454130e-01 7.49099925e-02 -3.75179470e-01
-9.51716959e-01 -6.25772357e-01 5.81935287e-01 -5.00168324e-01
-1.09618440e-01 -1.91523269e-01 -1.19894111e+00 -7.72423208e-01
-2.78537780e-01 8.04086566e-01 2.93918759e-01 1.05780280e+00
2.49675989e-01 -2.30610102e-01 3.04589242e-01 3.08924258e-01
-1.46296155e+00 -9.27434325e-01 1.77657884e-02 6.22745275e-01
4.69215691e-01 -3.65413845e-01 -6.51013970e-01 -1.43192083e-01] | [10.699180603027344, 7.865109443664551] |
b3fb821b-a183-477f-9cac-38e73bd250ca | detection-and-resolution-of-rumours-in-social | 1704.00656 | null | http://arxiv.org/abs/1704.00656v3 | http://arxiv.org/pdf/1704.00656v3.pdf | Detection and Resolution of Rumours in Social Media: A Survey | Despite the increasing use of social media platforms for information and news
gathering, its unmoderated nature often leads to the emergence and spread of
rumours, i.e. pieces of information that are unverified at the time of posting.
At the same time, the openness of social media platforms provides opportunities
to study how users share and discuss rumours, and to explore how natural
language processing and data mining techniques may be used to find ways of
determining their veracity. In this survey we introduce and discuss two types
of rumours that circulate on social media; long-standing rumours that circulate
for long periods of time, and newly-emerging rumours spawned during fast-paced
events such as breaking news, where reports are released piecemeal and often
with an unverified status in their early stages. We provide an overview of
research into social media rumours with the ultimate goal of developing a
rumour classification system that consists of four components: rumour
detection, rumour tracking, rumour stance classification and rumour veracity
classification. We delve into the approaches presented in the scientific
literature for the development of each of these four components. We summarise
the efforts and achievements so far towards the development of rumour
classification systems and conclude with suggestions for avenues for future
research in social media mining for detection and resolution of rumours. | ['Rob Procter', 'Kalina Bontcheva', 'Ahmet Aker', 'Maria Liakata', 'Arkaitz Zubiaga'] | 2017-04-03 | null | null | null | null | ['rumour-detection'] | ['natural-language-processing'] | [-3.81750047e-01 2.73704976e-01 -4.21281725e-01 -9.16401222e-02
-1.77525338e-02 -3.84951591e-01 9.66716468e-01 8.50464880e-01
-7.70411789e-02 7.45480359e-01 6.96309268e-01 -3.98147613e-01
7.40347728e-02 -8.11828136e-01 -1.43974498e-01 -1.29994899e-01
-4.74915773e-01 2.89699644e-01 2.17366502e-01 -7.58340240e-01
8.28989744e-01 2.70257264e-01 -1.52132511e+00 5.71870208e-01
4.63888854e-01 6.85377359e-01 -1.94193408e-01 6.15703881e-01
-5.13023496e-01 1.75498760e+00 -1.14496982e+00 -3.42812240e-01
-4.48809952e-01 -7.52103150e-01 -1.04753947e+00 4.15689290e-01
-1.20431654e-01 -2.24344641e-01 -2.21734136e-01 8.24271142e-01
1.08651333e-01 1.39975131e-01 3.40289176e-01 -1.13476992e+00
-8.67441356e-01 7.78374851e-01 -5.46706498e-01 9.34317112e-01
9.50601220e-01 -4.48699385e-01 7.79651940e-01 -7.76537120e-01
9.61247742e-01 1.00389385e+00 8.95778477e-01 6.98136687e-02
-8.26749682e-01 -5.65945387e-01 -2.03149304e-01 3.25539768e-01
-8.91641259e-01 -4.72970158e-01 7.06736743e-01 -9.23055112e-01
7.28821695e-01 4.79768902e-01 7.63738096e-01 1.12735677e+00
6.45691574e-01 5.79686701e-01 1.43416941e+00 -1.77439332e-01
1.64702371e-01 6.42564654e-01 4.09993231e-01 3.85696650e-01
9.15186182e-02 -2.09648624e-01 -1.00609231e+00 -8.63045454e-01
2.96361744e-01 1.28471673e-01 -3.17447543e-01 3.92790854e-01
-1.05729723e+00 1.37866163e+00 2.40642771e-01 4.20449078e-01
-8.90477061e-01 -9.22524691e-01 7.76615202e-01 1.20104504e+00
1.23721421e+00 6.28598392e-01 6.86994940e-02 -2.07715079e-01
-1.24935651e+00 4.44184512e-01 1.47960484e+00 4.58988309e-01
4.15873647e-01 -8.05089101e-02 1.59257561e-01 1.27237070e+00
1.66252032e-01 -4.40525860e-02 7.14792490e-01 -4.01807815e-01
5.26938885e-02 4.79149431e-01 4.60914493e-01 -1.65518785e+00
-6.38339818e-01 -2.55180150e-01 -5.10120988e-01 7.63882175e-02
1.18360326e-01 -2.40429372e-01 -1.74639627e-01 9.35651362e-01
3.53443593e-01 -2.72031963e-01 -6.86360821e-02 8.11290920e-01
9.37641203e-01 7.38771498e-01 -3.76850367e-01 -8.64937603e-01
1.46696389e+00 -8.21156383e-01 -1.09553719e+00 -1.96886770e-02
2.35780075e-01 -1.20504725e+00 3.96276832e-01 4.46880996e-01
-1.17055237e+00 1.55829445e-01 -9.87937510e-01 2.72459984e-01
-2.59053141e-01 -1.11159742e+00 4.20805305e-01 5.28225064e-01
-6.04258955e-01 7.47518480e-01 -2.86349595e-01 -5.05558848e-01
2.76705652e-01 -4.55687523e-01 6.57210648e-02 2.16224909e-01
-1.47482157e+00 1.03797913e+00 -5.29161096e-02 -2.76962996e-01
-7.44117856e-01 -3.53754312e-01 -4.16906565e-01 -5.45197904e-01
4.46801990e-01 -2.27489248e-01 1.34926987e+00 -1.20813179e+00
-1.28588045e+00 9.89665151e-01 -9.69944671e-02 -5.96354961e-01
7.73858368e-01 -2.01426726e-03 -9.58360314e-01 7.15236142e-02
3.42391014e-01 -4.81248826e-01 9.60245967e-01 -8.51718605e-01
-5.63221872e-01 -3.32742870e-01 -2.35387415e-01 -1.15428127e-01
-2.10600972e-01 9.84621346e-01 5.78275144e-01 -6.51370466e-01
2.45336115e-01 -6.66552901e-01 2.06012115e-01 -6.88006639e-01
-3.91750991e-01 -2.32427105e-01 6.52549446e-01 -7.74044216e-01
1.41143429e+00 -1.66570187e+00 -3.27787578e-01 -7.81957656e-02
7.24264145e-01 4.24046293e-02 5.64216137e-01 1.25430858e+00
1.96962118e-01 3.88013661e-01 2.82125741e-01 -3.25368717e-02
-3.87037009e-01 -2.46191137e-02 -6.87806189e-01 7.55333960e-01
-3.23987722e-01 4.78426516e-01 -1.25501239e+00 -9.25130844e-02
-3.86327922e-01 8.24862272e-02 -3.27728279e-02 2.85621554e-01
1.05467461e-01 3.21211457e-01 -2.70023733e-01 5.84811449e-01
4.39248413e-01 -3.84310395e-01 2.10595913e-02 6.35591686e-01
-7.91864395e-01 1.02574646e+00 -5.72769701e-01 4.56627488e-01
-1.26127884e-01 9.20944095e-01 3.89748603e-01 -4.69784170e-01
1.18903899e+00 5.97543001e-01 3.37367922e-01 -2.96032846e-01
3.64422560e-01 2.89955139e-01 -3.63349855e-01 -5.82193017e-01
1.12282026e+00 -6.89014316e-01 -4.28472050e-02 1.35697711e+00
-3.50647479e-01 1.44077107e-01 -6.79211393e-02 4.02958691e-01
1.07981074e+00 -9.22526836e-01 9.74581599e-01 -1.56506941e-01
2.71279424e-01 3.38949382e-01 2.68479556e-01 8.19875956e-01
-3.47185314e-01 3.12193722e-01 8.30819547e-01 -5.57505429e-01
-1.02501643e+00 -3.71435344e-01 -3.06464761e-01 1.53256607e+00
-1.23235494e-01 -3.91523927e-01 -1.09947763e-01 -3.28234524e-01
1.75269112e-01 6.01111293e-01 -8.28857303e-01 3.62485588e-01
-6.07307665e-02 -1.05962658e+00 4.38258559e-01 -1.91630989e-01
2.40807295e-01 -1.43931115e+00 -3.31565708e-01 5.34978449e-01
-4.09550756e-01 -6.25053763e-01 3.30802333e-03 -1.51422217e-01
-7.90772498e-01 -1.05468547e+00 -6.56113029e-01 -3.87512803e-01
2.44675547e-01 7.16245234e-01 1.10500467e+00 4.96699721e-01
3.40537667e-01 2.22425777e-02 -7.63725340e-01 -6.62927389e-01
-1.09878564e+00 -2.85746962e-01 2.33688414e-01 8.15614872e-03
3.55259210e-01 -5.85362673e-01 -6.65424317e-02 5.69572449e-01
-9.17335153e-01 -1.80389687e-01 -9.62404758e-02 6.58909976e-01
-2.64732867e-01 6.13199398e-02 1.25139940e+00 -1.29276216e+00
1.47065139e+00 -1.77202392e+00 7.64378831e-02 -4.89702851e-01
-8.66674006e-01 -7.39836216e-01 1.60239130e-01 -3.28417391e-01
-8.33769619e-01 -1.33108902e+00 3.30981947e-02 1.87162295e-01
8.30487683e-02 1.06634498e+00 9.85069156e-01 -1.30444178e-02
1.13183928e+00 1.78147703e-02 5.95006764e-01 -6.66119814e-01
-1.56198274e-02 1.41583431e+00 2.56455410e-02 2.03028455e-01
3.10232252e-01 6.31891251e-01 -8.86268973e-01 -1.21002710e+00
-1.14132309e+00 -9.09652054e-01 -5.94720617e-03 -5.85981488e-01
6.30140007e-02 -6.72794998e-01 -4.65435684e-01 7.38598943e-01
-1.06425190e+00 9.36817378e-02 6.28072917e-02 1.54334351e-01
-2.75629073e-01 3.89873594e-01 -1.28305364e+00 -9.74554777e-01
-4.80736762e-01 -2.12499201e-01 -3.97383690e-01 -3.17505859e-02
-1.01019728e+00 -1.07771242e+00 4.44726855e-01 7.94152200e-01
6.26573145e-01 4.07621026e-01 4.07383084e-01 -1.05018389e+00
1.11908212e-01 -2.87017316e-01 -6.46485165e-02 1.21979259e-01
1.53098255e-01 -2.39320397e-01 -6.43604457e-01 -2.04398543e-01
6.60693884e-01 -5.37187338e-01 6.12745047e-01 -1.14936933e-01
4.09737825e-02 -8.89628768e-01 -1.68402959e-02 -2.23392144e-01
6.69502020e-01 -9.73932147e-02 3.93910348e-01 9.15194154e-01
1.76964793e-02 9.34015810e-01 5.40908813e-01 9.75035071e-01
4.94927108e-01 2.54633695e-01 4.65892971e-01 4.74562436e-01
2.10810512e-01 -1.11350663e-01 5.92150152e-01 1.20169580e+00
-3.17812890e-01 -1.86010107e-01 -8.11388850e-01 7.34861255e-01
-1.64665294e+00 -1.36112618e+00 -6.56415164e-01 2.03193784e+00
8.90957832e-01 3.06746393e-01 5.57766140e-01 2.59740353e-01
8.90088022e-01 6.96971536e-01 -8.98350850e-02 -1.06721759e+00
-6.00753315e-02 -6.29253805e-01 1.87066048e-01 6.04089141e-01
-4.94270980e-01 3.45819443e-01 7.31237316e+00 -2.10392609e-01
-1.07888639e+00 4.61118132e-01 6.62280470e-02 -1.59449100e-01
-2.06593916e-01 -1.96614534e-01 -4.28301543e-01 5.22397161e-01
1.20957208e+00 -4.74055648e-01 4.85722959e-01 6.18649840e-01
5.67483127e-01 -3.68905216e-01 -3.80431443e-01 2.92472541e-01
4.75290596e-01 -1.47197342e+00 -2.95664310e-01 4.62912805e-02
9.19981420e-01 6.29878044e-01 -2.44782791e-01 2.22184971e-01
2.31497005e-01 -5.40882111e-01 8.76426935e-01 3.09028745e-01
-1.15162395e-01 -5.82885563e-01 8.13439429e-01 8.16604137e-01
-1.28741577e-01 -1.21088497e-01 -5.19440591e-01 -8.92848253e-01
5.76071501e-01 1.12129104e+00 -1.20780504e+00 1.88174620e-01
5.86352587e-01 9.41557527e-01 -5.40650040e-02 9.70785737e-01
-2.32809752e-01 8.66907895e-01 1.85406953e-01 -2.80289471e-01
2.41440699e-01 8.41554850e-02 1.26332128e+00 1.28301871e+00
-5.95132522e-02 1.99502166e-02 3.93965915e-02 7.29860783e-01
-7.18380064e-02 2.81419784e-01 -3.96795034e-01 -3.58415663e-01
5.54222465e-01 9.59957063e-01 -6.50093853e-01 -3.18954349e-01
-3.99532586e-01 7.45030582e-01 3.15987676e-01 -3.72848473e-02
-4.97144550e-01 -4.21685129e-02 5.25827467e-01 8.39361370e-01
-1.92391321e-01 -7.73250312e-02 -2.32752889e-01 -1.14496124e+00
-2.74605602e-01 -1.02853370e+00 4.74983931e-01 -4.39820498e-01
-1.74493372e+00 6.71492636e-01 -2.75311947e-01 -7.59678006e-01
-3.11819822e-01 2.58153856e-01 -9.03989971e-01 6.60750985e-01
-1.52162373e+00 -3.84490162e-01 -2.66690433e-01 3.52736056e-01
5.53502262e-01 -1.68529734e-01 7.13719130e-01 7.84954876e-02
-2.33796552e-01 -1.15498453e-01 1.19030677e-01 -9.62281227e-02
8.41534138e-01 -7.83154130e-01 3.74302417e-01 1.33943960e-01
-2.70754009e-01 6.37198508e-01 1.26264107e+00 -1.00343549e+00
-8.37549210e-01 -7.30277658e-01 1.43488896e+00 -4.61637110e-01
1.42744040e+00 -3.59916538e-02 -1.39919162e+00 6.91111386e-01
2.09325641e-01 -5.04129112e-01 1.03608000e+00 7.03140497e-01
-4.17288303e-01 6.53820753e-01 -1.15726769e+00 1.91754848e-01
4.95181084e-01 -6.02620304e-01 -1.20148563e+00 6.13549292e-01
1.67730913e-01 -2.35081203e-02 -9.57050502e-01 -3.65220666e-01
5.64005554e-01 -1.33439529e+00 3.98680300e-01 -8.20232451e-01
7.86403358e-01 2.20872566e-01 3.98941606e-01 -1.45430827e+00
-5.45963049e-01 -1.17031395e+00 -1.90478578e-01 1.19730914e+00
3.02523613e-01 -1.02206731e+00 3.81199270e-01 4.42640074e-02
-4.95736822e-02 -5.00441372e-01 -4.71020341e-01 -5.31679153e-01
-5.38743101e-02 -2.68105358e-01 1.79625750e-01 1.64265597e+00
9.03968692e-01 4.86424714e-01 -8.84435117e-01 -2.63491064e-01
4.37285036e-01 2.20695391e-01 6.63166583e-01 -1.43775332e+00
9.05239210e-02 -4.41240072e-01 -4.77079749e-01 -7.44726837e-01
-2.90709883e-01 -7.86660790e-01 -1.44480199e-01 -1.21613193e+00
3.05493146e-01 -2.83825547e-01 -2.18311608e-01 1.76942553e-02
2.48707175e-01 3.17706704e-01 -2.09137537e-02 9.84836280e-01
-4.32727218e-01 1.23249590e-01 1.13557613e+00 2.14542344e-01
-5.86713493e-01 3.06153923e-01 -1.00019348e+00 9.45420980e-01
8.30488026e-01 -8.66097927e-01 2.16092225e-02 4.31590527e-01
1.03027308e+00 5.20192325e-01 3.07824850e-01 -4.17730063e-01
2.37601206e-01 -4.06774819e-01 -3.34412903e-01 -5.78603983e-01
-1.13792479e-01 4.79559153e-02 1.34107485e-01 1.53810441e-01
-4.02215660e-01 3.24322432e-02 -1.81605488e-01 6.95991278e-01
-4.22874957e-01 -3.67618412e-01 8.79938126e-01 -1.89868331e-01
-2.99897611e-01 -2.73897648e-01 -1.35626602e+00 1.74888700e-01
9.61317003e-01 -3.76641788e-02 -5.80441892e-01 -9.57916677e-01
-8.92277122e-01 -9.25641805e-02 5.84360421e-01 8.26215982e-01
4.26394075e-01 -8.10032070e-01 -1.33436191e+00 -5.33531308e-02
1.08728416e-01 -8.31185639e-01 1.46176741e-01 1.32589781e+00
-3.67490590e-01 1.11757778e-01 -4.62453961e-01 4.48246635e-02
-1.25512874e+00 5.15970767e-01 3.65508534e-02 1.76885482e-02
-9.33712065e-01 4.79113936e-01 -4.48504955e-01 -4.81779426e-02
-1.75203592e-01 3.25217187e-01 -6.71939015e-01 8.41638088e-01
1.38299942e+00 8.78447592e-01 1.85258970e-01 -9.33390319e-01
-1.14515401e-01 -6.34060264e-01 -5.33039272e-01 1.32643580e-01
1.38866103e+00 -6.96687579e-01 -8.53078365e-01 1.21577728e+00
9.10248995e-01 3.87294561e-01 -3.55437577e-01 -3.62899899e-01
2.68808722e-01 -4.20279622e-01 3.45305115e-01 -7.34326899e-01
-5.06677866e-01 7.05078244e-02 -5.56916416e-01 1.39114761e+00
2.77178228e-01 3.29774052e-01 9.72657263e-01 -1.58539433e-02
2.27854148e-01 -9.86947119e-01 -6.66554598e-03 9.16784227e-01
1.26011658e+00 -1.09047472e+00 4.59470063e-01 -3.72420162e-01
-7.59477258e-01 1.07164073e+00 -2.99179345e-01 -2.00834915e-01
8.93017411e-01 8.15450251e-02 1.99467897e-01 -7.61294901e-01
-9.43246841e-01 4.07735258e-01 5.96127026e-02 1.06326863e-01
6.41319394e-01 3.22230965e-01 -8.19644570e-01 5.79051137e-01
-4.75753278e-01 7.14968294e-02 1.40722537e+00 1.12361729e+00
-1.14517593e+00 -4.84141707e-01 -8.70112777e-01 6.93032205e-01
-1.01089430e+00 8.75443742e-02 -9.10938740e-01 4.09450352e-01
-3.98390234e-01 1.35508406e+00 -1.38981774e-01 -4.46125031e-01
-9.61534381e-02 1.57058552e-01 -1.86104447e-01 -8.12877178e-01
-9.61314082e-01 -3.97475243e-01 7.99472988e-01 -2.87737846e-01
-3.93288225e-01 -9.91669476e-01 -8.69592190e-01 -1.13222361e+00
-6.18955493e-01 5.39219558e-01 6.74173534e-01 9.89355326e-01
1.30667642e-01 1.53170124e-01 8.37667286e-01 -5.17351151e-01
-5.83069146e-01 -1.29391587e+00 -1.01717806e+00 3.04509133e-01
7.10519493e-01 -6.17818415e-01 -7.24641919e-01 -1.97487548e-01] | [8.241066932678223, 10.117903709411621] |
5e499a49-68ed-4d96-acc2-70d734e0c74f | visual-context-aware-convolution-filters-for | 1906.09986 | null | https://arxiv.org/abs/1906.09986v1 | https://arxiv.org/pdf/1906.09986v1.pdf | Visual Context-aware Convolution Filters for Transformation-invariant Neural Network | We propose a novel visual context-aware filter generation module which incorporates contextual information present in images into Convolutional Neural Networks (CNNs). In contrast to traditional CNNs, we do not employ the same set of learned convolution filters for all input image instances. Our proposed input-conditioned convolution filters when combined with techniques inspired by Multi-instance learning and max-pooling, results in a transformation-invariant neural network. We investigated the performance of our proposed framework on three MNIST variations, which covers both rotation and scaling variance, and achieved 1.13% error on MNIST-rot-12k, 1.12% error on Half-rotated MNIST and 0.68% error on Scaling MNIST, which is significantly better than the state-of-the-art results. We make use of visualization to further prove the effectiveness of our visual context-aware convolution filters. Our proposed visual context-aware convolution filter generation framework can also serve as a plugin for any CNN based architecture and enhance its modeling capacity. | ['Suraj Tripathi', 'Abhay Kumar', 'Chirag Singh'] | 2019-06-15 | null | null | null | null | ['rotated-mnist'] | ['computer-vision'] | [ 2.18769684e-01 -2.14227378e-01 1.74052149e-01 -5.01778662e-01
4.85879593e-02 -6.31737590e-01 6.50358379e-01 -1.48708925e-01
-8.60927641e-01 5.90702772e-01 6.71997219e-02 -3.28728169e-01
-1.71989053e-01 -8.64204526e-01 -9.80653107e-01 -4.70465511e-01
1.86922833e-01 -3.29535544e-01 3.88939500e-01 -2.04865828e-01
3.29048008e-01 9.10694897e-01 -1.77228963e+00 8.06499660e-01
5.41118443e-01 1.03494120e+00 4.67197895e-01 8.15428495e-01
-1.06681839e-01 7.55763888e-01 -6.00391626e-01 -2.72758693e-01
3.19947183e-01 2.36337781e-02 -7.73127735e-01 -4.85947728e-02
9.60726976e-01 -1.00569203e-01 -6.03085041e-01 9.15172100e-01
4.41180468e-01 3.38217974e-01 3.15620422e-01 -9.67030048e-01
-1.01652610e+00 5.79181850e-01 -2.60740221e-01 4.23029661e-01
-1.91463992e-01 3.17249358e-01 6.74483657e-01 -9.83010352e-01
7.67603040e-01 1.20407474e+00 4.81944174e-01 4.88375038e-01
-1.10240710e+00 -7.19139338e-01 5.33465803e-01 4.13711190e-01
-1.15833938e+00 -1.43195897e-01 6.25300288e-01 -3.79913986e-01
1.26368105e+00 3.15516800e-01 9.08617020e-01 1.08049738e+00
2.80848652e-01 5.17187834e-01 1.13826239e+00 -3.73243392e-01
-1.79854613e-02 7.06054643e-02 4.66864705e-02 7.67850697e-01
3.32759231e-01 1.92736655e-01 -6.25887990e-01 3.43739837e-01
1.19955099e+00 3.55857253e-01 -1.96238503e-01 -1.28978593e-02
-1.51489377e+00 5.55448234e-01 9.60032880e-01 2.21859947e-01
-1.53388798e-01 6.56300247e-01 2.80453652e-01 3.53050530e-01
1.53315365e-01 4.48341161e-01 -6.75628066e-01 2.67387837e-01
-9.75671113e-01 2.24114791e-01 4.37125891e-01 1.12668097e+00
9.91201818e-01 2.56498128e-01 -5.40722787e-01 6.15724206e-01
4.13649343e-02 3.74705553e-01 4.68865842e-01 -7.04155326e-01
2.96800494e-01 7.54486322e-01 -1.39816791e-01 -9.01690602e-01
-6.95567548e-01 -9.23119843e-01 -8.66316259e-01 5.30963540e-01
4.01301444e-01 -2.39380747e-01 -1.32874596e+00 1.58120477e+00
-3.85164581e-02 4.86647904e-01 4.94549870e-02 1.02759993e+00
1.18823528e+00 3.36280823e-01 1.22620270e-01 2.27316067e-01
1.49048114e+00 -1.26677811e+00 -5.47946155e-01 -3.22525740e-01
9.89934877e-02 -1.17859483e+00 1.27285647e+00 2.85427988e-01
-8.04021955e-01 -9.63562608e-01 -1.30195808e+00 -1.96157143e-01
-6.66031182e-01 7.86418259e-01 7.76636243e-01 4.29597467e-01
-1.06632161e+00 7.27616549e-01 -7.69515157e-01 -3.60769182e-01
5.43727636e-01 3.25702220e-01 -5.00026643e-01 1.12558022e-01
-6.94950283e-01 8.11095297e-01 4.48319525e-01 2.69908547e-01
-9.21895325e-01 -8.13366175e-01 -4.98469830e-01 2.52496362e-01
3.98238778e-01 -7.40181863e-01 1.04005289e+00 -9.64999139e-01
-1.33361447e+00 4.48570579e-01 -2.13705614e-01 -3.83848697e-01
5.79874575e-01 -4.72332209e-01 -6.01730585e-01 1.23642601e-01
-4.49220687e-01 8.02457273e-01 9.34605122e-01 -9.99140918e-01
-7.18855679e-01 7.02930838e-02 3.19016784e-01 4.58109286e-03
-4.79511619e-01 -6.08678050e-02 -5.35033226e-01 -9.54385996e-01
1.48296580e-01 -7.74428606e-01 -2.71273315e-01 4.56501037e-01
-3.97809952e-01 2.63129681e-01 1.11258101e+00 -3.39576840e-01
9.87454891e-01 -2.01117158e+00 -7.09839687e-02 1.78176522e-01
1.60581723e-01 6.89652383e-01 -3.28652442e-01 2.11784810e-01
-2.95376986e-01 8.41073692e-02 4.65660542e-02 -3.21692497e-01
-2.95632631e-01 1.95020199e-01 -1.79563656e-01 2.05704585e-01
7.28920519e-01 9.46388006e-01 -7.16659188e-01 -5.76087944e-02
4.85229790e-01 8.33792806e-01 -5.82319081e-01 -4.43234704e-02
-1.91864014e-01 2.94439673e-01 8.56416374e-02 3.97742331e-01
8.98860812e-01 -2.73781180e-01 2.36035213e-01 -7.06390083e-01
-3.37870181e-01 -9.45541039e-02 -1.50025189e+00 1.89108896e+00
-5.31288683e-01 9.70288754e-01 -3.11938196e-01 -7.52136886e-01
1.06489944e+00 1.09383501e-01 -9.12627131e-02 -6.39640570e-01
8.73867124e-02 -1.37463436e-01 4.67128307e-02 -4.71897602e-01
6.24968767e-01 3.65352958e-01 4.80683297e-01 1.83964998e-01
6.46571457e-01 3.72075200e-01 1.89065143e-01 5.02129309e-02
8.91910613e-01 2.74624228e-01 8.30394775e-02 -5.02141893e-01
5.11774242e-01 -2.44530499e-01 3.59217376e-01 8.50499272e-01
9.55885556e-03 7.85613298e-01 4.79177266e-01 -9.17712629e-01
-1.08971751e+00 -8.67211044e-01 -1.41378328e-01 1.12943792e+00
6.80411085e-02 -5.01038313e-01 -4.86436486e-01 -5.77323973e-01
-2.83926213e-03 3.39627594e-01 -9.38629091e-01 9.06569257e-05
-7.22408950e-01 -7.43882179e-01 4.57133770e-01 8.24405849e-01
8.87049675e-01 -1.07967544e+00 -7.78931081e-01 1.46494150e-01
3.87162000e-01 -1.28055048e+00 -2.28007004e-01 4.48376358e-01
-8.64023149e-01 -1.14061451e+00 -5.52215338e-01 -8.18228126e-01
6.68441474e-01 2.86221951e-01 8.75991940e-01 1.33639708e-01
-4.40768033e-01 1.27316162e-01 -2.30335861e-01 -3.76858532e-01
8.90169442e-02 2.26539358e-01 -3.00801486e-01 5.08956574e-02
1.82702448e-02 -7.61426449e-01 -1.01891708e+00 3.26508015e-01
-9.64030743e-01 2.79352307e-01 7.44352460e-01 8.17698181e-01
6.20056391e-01 -2.03950450e-01 3.18249971e-01 -9.29658234e-01
4.38510001e-01 -4.50821929e-02 -6.67772770e-01 4.42757308e-01
-3.54481041e-01 3.70563179e-01 7.49664009e-01 -8.45013201e-01
-1.05131173e+00 2.73802608e-01 7.87300169e-02 -4.98893410e-01
-3.01260203e-01 2.47195870e-01 1.65537968e-01 -4.24053311e-01
9.58496392e-01 -1.92299094e-02 -3.07933897e-01 -5.53996086e-01
6.80600941e-01 2.61320829e-01 7.74293423e-01 -4.45182443e-01
7.35475957e-01 5.93348384e-01 2.14534238e-01 -7.15867579e-01
-5.04837513e-01 -2.50095367e-01 -6.86786294e-01 -1.41663685e-01
8.87068808e-01 -1.05154932e+00 -8.76691878e-01 4.96131331e-01
-1.23506153e+00 -3.40586960e-01 -1.90575570e-01 4.76742357e-01
-2.00325608e-01 -1.02976479e-01 -3.83882314e-01 -5.06197631e-01
-4.57154572e-01 -1.43107009e+00 6.18470192e-01 4.94313955e-01
2.20540419e-01 -7.27842510e-01 -3.85705233e-01 -2.11953759e-01
9.34934199e-01 3.03546399e-01 6.77345514e-01 -4.00335670e-01
-7.40651429e-01 9.84763354e-02 -7.89038181e-01 5.38111985e-01
6.65412024e-02 3.68369728e-01 -1.27921593e+00 -3.44161659e-01
-5.69608986e-01 -1.67090476e-01 1.20148396e+00 1.87970653e-01
1.39500010e+00 -1.38475567e-01 -1.28430009e-01 9.76494789e-01
1.80870605e+00 3.76371443e-02 6.22962892e-01 3.91865551e-01
7.61391222e-01 8.14160407e-02 1.23891570e-01 4.48494464e-01
7.04867989e-02 5.74235022e-01 6.84164345e-01 -4.32149023e-01
-5.87453425e-01 4.37072180e-02 -2.19882816e-01 3.02711785e-01
-4.87324923e-01 -3.97980809e-02 -7.80141592e-01 4.82244760e-01
-1.93668175e+00 -8.66321325e-01 -1.92139074e-02 1.81620979e+00
4.57249552e-01 1.97833166e-01 -1.94894046e-01 -3.06370179e-03
5.97005665e-01 8.11696276e-02 -3.79213750e-01 -4.50878620e-01
-2.23988578e-01 8.19686770e-01 7.12223589e-01 3.35534245e-01
-1.29021466e+00 1.04474902e+00 6.15409994e+00 6.07050896e-01
-1.55572486e+00 -1.02131618e-02 2.81879723e-01 -3.16365153e-01
4.35256725e-03 -5.76280467e-02 -7.61463940e-01 7.06786215e-02
5.76849163e-01 1.14436492e-01 5.40930688e-01 9.24830317e-01
9.54953507e-02 1.39638722e-01 -9.56129491e-01 9.72758770e-01
-1.01703048e-01 -1.78700328e+00 2.06823185e-01 -2.27708459e-01
9.55716789e-01 1.77165344e-01 2.63877660e-01 1.37691855e-01
2.20826060e-01 -1.18861878e+00 9.88785625e-01 8.16220045e-01
8.31053436e-01 -1.00944769e+00 6.66482151e-01 -2.41383970e-01
-1.41914248e+00 -2.84692168e-01 -7.40997732e-01 -2.39459887e-01
-3.10983568e-01 5.10958791e-01 -8.19249272e-01 5.95500410e-01
9.67527390e-01 6.22631848e-01 -8.88270319e-01 1.28898549e+00
-3.31711918e-01 2.98336059e-01 -1.86560228e-01 -1.51643306e-01
3.92362684e-01 3.02848756e-01 1.63474634e-01 1.60824382e+00
2.42251083e-01 -2.34612048e-01 -2.45028781e-03 8.13367069e-01
-2.52991974e-01 9.72361714e-02 -2.40566358e-01 1.69385046e-01
3.96502167e-01 1.59994364e+00 -9.67069924e-01 -4.80410606e-01
-5.83251178e-01 1.03727245e+00 5.87222993e-01 6.99339628e-01
-7.70938933e-01 -7.49571323e-01 1.02896166e+00 -1.21561706e-01
8.37854028e-01 -4.20127094e-01 -3.40563506e-01 -1.19282603e+00
1.34232208e-01 -7.12033033e-01 3.00590582e-02 -7.96924233e-01
-8.19744825e-01 9.30945814e-01 -1.20650910e-01 -1.05983436e+00
2.07720965e-01 -1.15144885e+00 -7.23132432e-01 9.71872747e-01
-1.74986732e+00 -1.46262276e+00 -8.54191899e-01 7.25601614e-01
3.01185042e-01 -2.69951224e-01 8.59695554e-01 3.33182186e-01
-4.31063563e-01 7.68863440e-01 -4.46818136e-02 2.35074833e-01
7.72929430e-01 -1.12224126e+00 6.62133217e-01 1.07120156e+00
3.20562780e-01 9.56740141e-01 4.27149653e-01 -2.62935162e-01
-1.22924399e+00 -1.48696220e+00 3.98498535e-01 -8.34718272e-02
4.46186692e-01 -4.74322915e-01 -7.77764320e-01 7.48579502e-01
4.28433627e-01 6.81949675e-01 3.94334584e-01 1.75685231e-02
-6.84790850e-01 -3.27259213e-01 -1.06991374e+00 7.56582022e-01
1.33792734e+00 -4.67321515e-01 -3.13267469e-01 1.86070964e-01
9.48235512e-01 -6.19262278e-01 -6.49898887e-01 4.92580444e-01
6.82496011e-01 -9.67970669e-01 1.16473365e+00 -7.76074171e-01
4.75343168e-01 -5.71010351e-01 -2.01634273e-01 -1.22904050e+00
-6.59192383e-01 -2.53368407e-01 6.91591576e-02 8.18706393e-01
4.57685113e-01 -5.68465352e-01 5.88896930e-01 2.46705208e-02
-3.01662862e-01 -7.59183884e-01 -7.07071543e-01 -6.36767805e-01
-8.91807303e-02 -7.37021983e-01 7.78124332e-01 7.96788752e-01
-5.85361540e-01 -8.99177492e-02 -3.10749382e-01 3.84332985e-01
3.35768074e-01 5.86757390e-03 6.98130071e-01 -8.68017852e-01
-3.85511816e-01 -5.17255366e-01 -6.66105986e-01 -7.81018436e-01
-2.64588356e-01 -8.51960242e-01 -5.27946055e-01 -1.46518922e+00
2.24223733e-02 -9.86850336e-02 -6.35163844e-01 9.58090007e-01
-9.60583538e-02 6.68395400e-01 4.43871558e-01 -2.28258818e-01
-2.90196419e-01 1.80882365e-01 1.49967062e+00 -1.14411958e-01
-1.27081648e-01 -3.17530483e-01 -7.22738147e-01 5.63899457e-01
7.81418562e-01 -1.89797118e-01 -5.76289356e-01 -6.69426024e-01
2.29196414e-01 -5.83183646e-01 7.38714933e-01 -1.45478415e+00
3.79839480e-01 -1.79391176e-01 8.96587193e-01 -5.27646244e-01
9.58311409e-02 -7.98544109e-01 2.45848089e-01 4.28881854e-01
-2.89406896e-01 1.79969624e-01 7.50810564e-01 3.56707454e-01
-9.72172916e-02 -4.67992350e-02 7.31191695e-01 -1.87589541e-01
-9.67177212e-01 1.90512866e-01 -2.91425973e-01 -2.87378669e-01
7.83252478e-01 -6.14967346e-02 -7.51061320e-01 7.96819329e-02
-9.83252943e-01 -2.20124334e-01 1.21017046e-01 4.02317673e-01
7.89971352e-01 -1.33451831e+00 -5.92487097e-01 4.73859251e-01
2.20615283e-01 -4.78206575e-02 3.78811389e-01 3.76743346e-01
-7.61478901e-01 2.81339437e-01 -6.14875376e-01 -5.39112031e-01
-1.27176380e+00 4.90420252e-01 2.95400441e-01 -3.94144580e-02
-6.62535012e-01 9.57238436e-01 -5.62044159e-02 -3.43100905e-01
2.47549489e-01 -8.98664594e-01 -9.04970020e-02 -9.69408080e-02
7.68594623e-01 2.62442946e-01 3.94922793e-01 -1.44539818e-01
-3.61668408e-01 5.32286942e-01 -1.77680001e-01 3.32130678e-02
1.43892801e+00 3.96637022e-01 7.53443837e-02 2.61135269e-02
1.09489858e+00 -2.84160495e-01 -1.43685186e+00 -3.95469338e-01
-3.76079142e-01 -4.91793573e-01 9.88415107e-02 -1.05066156e+00
-1.46826470e+00 9.78100061e-01 9.46838260e-01 -1.73858091e-01
1.23982632e+00 -3.97529960e-01 1.40758410e-01 7.12908268e-01
1.29349884e-02 -8.99805605e-01 2.95972943e-01 6.13943398e-01
1.18908298e+00 -1.10867345e+00 8.78601894e-02 -3.31646085e-01
-3.30003917e-01 1.63147473e+00 8.68211329e-01 -4.84101415e-01
7.15403020e-01 3.92875522e-01 2.10491434e-01 -8.80435333e-02
-8.22707593e-01 -3.37162733e-01 5.78291297e-01 6.01031482e-01
5.58482170e-01 -2.86482181e-02 -9.62128118e-02 3.93224359e-01
-2.55650043e-01 1.20955698e-01 4.19406235e-01 1.05002844e+00
-3.89613688e-01 -9.79042172e-01 -1.91633865e-01 2.53949881e-01
-2.78205812e-01 -3.19087774e-01 1.79484114e-02 9.01092231e-01
5.56145668e-01 6.20412230e-01 1.91902936e-01 -6.18684769e-01
5.46870708e-01 7.69325718e-02 6.02220654e-01 -4.52616066e-01
-8.94775450e-01 -1.51621372e-01 -1.77945510e-01 -7.61272669e-01
-5.90962887e-01 -1.82149351e-01 -1.05513561e+00 -3.28156084e-01
-2.64236063e-01 -6.00411236e-01 8.84106517e-01 7.77343154e-01
4.37199295e-01 1.13969457e+00 2.19646677e-01 -9.06248093e-01
-7.81069323e-02 -1.02194226e+00 -1.57237828e-01 3.67145061e-01
4.00375158e-01 -6.52431369e-01 -1.51439413e-01 1.48085266e-01] | [9.047469139099121, 2.168177366256714] |
f727ae2c-2de2-4234-8291-cb332501f250 | visual-reference-resolution-using-attention | 1709.07992 | null | http://arxiv.org/abs/1709.07992v3 | http://arxiv.org/pdf/1709.07992v3.pdf | Visual Reference Resolution using Attention Memory for Visual Dialog | Visual dialog is a task of answering a series of inter-dependent questions
given an input image, and often requires to resolve visual references among the
questions. This problem is different from visual question answering (VQA),
which relies on spatial attention (a.k.a. visual grounding) estimated from an
image and question pair. We propose a novel attention mechanism that exploits
visual attentions in the past to resolve the current reference in the visual
dialog scenario. The proposed model is equipped with an associative attention
memory storing a sequence of previous (attention, key) pairs. From this memory,
the model retrieves the previous attention, taking into account recency, which
is most relevant for the current question, in order to resolve potentially
ambiguous references. The model then merges the retrieved attention with a
tentative one to obtain the final attention for the current question;
specifically, we use dynamic parameter prediction to combine the two attentions
conditioned on the question. Through extensive experiments on a new synthetic
visual dialog dataset, we show that our model significantly outperforms the
state-of-the-art (by ~16 % points) in situations, where visual reference
resolution plays an important role. Moreover, the proposed model achieves
superior performance (~ 2 % points improvement) in the Visual Dialog dataset,
despite having significantly fewer parameters than the baselines. | ['Andreas Lehrmann', 'Paul Hongsuck Seo', 'Bohyung Han', 'Leonid Sigal'] | 2017-09-23 | visual-reference-resolution-using-attention-1 | http://papers.nips.cc/paper/6962-visual-reference-resolution-using-attention-memory-for-visual-dialog | http://papers.nips.cc/paper/6962-visual-reference-resolution-using-attention-memory-for-visual-dialog.pdf | neurips-2017-12 | ['parameter-prediction'] | ['miscellaneous'] | [ 7.51608796e-03 1.46266803e-01 2.19849870e-01 -2.51436472e-01
-1.03418648e+00 -6.44629717e-01 5.98401010e-01 2.42999136e-01
-5.70533097e-01 5.15820265e-01 3.48873287e-01 -4.00531352e-01
1.29115075e-01 -4.44716424e-01 -8.61256421e-01 -5.41789532e-01
5.72639763e-01 5.84510446e-01 6.52669549e-01 -3.35105866e-01
5.65642416e-01 3.18356425e-01 -1.40438056e+00 4.99178737e-01
7.28579164e-01 1.04260707e+00 9.50892746e-01 7.29991317e-01
-6.46907270e-01 1.15590012e+00 -8.67968619e-01 -4.16943640e-01
-2.07630560e-01 -5.49677610e-01 -1.25374544e+00 1.60970151e-01
5.62687278e-01 -4.21415597e-01 -2.37227902e-01 9.65534747e-01
1.17608644e-01 4.82394069e-01 4.15788114e-01 -1.14230251e+00
-1.15418458e+00 5.02991527e-02 -6.85051203e-01 5.96797585e-01
5.28544247e-01 4.99670893e-01 1.09528172e+00 -1.35378361e+00
6.07206523e-01 1.64967096e+00 -7.39830211e-02 4.63504106e-01
-1.09994650e+00 -4.82480645e-01 6.35391772e-01 6.14859343e-01
-1.21299446e+00 -3.74268591e-01 7.91705966e-01 -3.73926491e-01
9.32412744e-01 2.37006932e-01 3.94705594e-01 8.70993555e-01
1.75146714e-01 7.28941977e-01 9.96233284e-01 -3.62604767e-01
1.26961783e-01 2.25722238e-01 3.38848859e-01 6.08476520e-01
-4.60112453e-01 -2.46993765e-01 -5.04590273e-01 -7.14145880e-03
4.91652250e-01 -4.60020155e-02 -4.16159481e-01 -2.42402434e-01
-1.15441895e+00 7.97411203e-01 9.39932227e-01 2.19607383e-01
-5.15842855e-01 -7.74890855e-02 -1.47380129e-01 1.54221803e-01
-1.15146607e-01 3.37469012e-01 -2.95198336e-02 3.10270160e-01
-5.40312350e-01 2.18271315e-01 4.78722543e-01 8.61521661e-01
9.90085065e-01 -2.55087346e-01 -7.49755740e-01 7.60486484e-01
4.91358072e-01 4.21855420e-01 2.61715293e-01 -1.08167934e+00
7.72900760e-01 7.77723432e-01 3.81693035e-01 -9.73551154e-01
-1.87103778e-01 1.30719421e-02 -5.64820886e-01 2.27422357e-01
6.46638870e-01 2.30614573e-01 -1.04567420e+00 1.77979612e+00
4.63828444e-01 -1.68967038e-01 2.58203179e-01 1.31751466e+00
1.24637961e+00 9.44345832e-01 4.23935711e-01 -2.57426649e-01
1.74329579e+00 -1.19703507e+00 -9.77900386e-01 -6.39437735e-01
-3.47042456e-02 -8.92570734e-01 1.41392708e+00 -3.94701958e-02
-1.21904743e+00 -8.38811219e-01 -9.08277631e-01 -6.53439403e-01
-2.59100884e-01 5.11587150e-02 -3.28035466e-02 -4.40312512e-02
-1.29470909e+00 -1.95858449e-01 -8.16569403e-02 -5.16271412e-01
6.37365580e-02 1.31684661e-01 -2.49984428e-01 -2.02917517e-03
-1.23090422e+00 9.85964954e-01 1.62949890e-01 2.16662660e-01
-1.03869402e+00 -4.93874639e-01 -8.25790346e-01 2.94060647e-01
8.08268130e-01 -8.21742535e-01 1.30188668e+00 -7.92686701e-01
-1.30954289e+00 8.10837507e-01 -7.17217922e-01 -4.69550729e-01
2.46838540e-01 -2.40765765e-01 -1.19843125e-01 4.90080327e-01
1.10461228e-01 1.13018370e+00 1.03149474e+00 -1.66709220e+00
-6.10960841e-01 -3.94605815e-01 5.89307010e-01 4.55313295e-01
1.55186102e-01 -1.29219741e-01 -9.59674358e-01 -3.05020481e-01
2.42954493e-02 -8.72154593e-01 -2.23681927e-02 7.18716085e-02
-2.26874247e-01 -3.99985075e-01 9.12640750e-01 -9.22885954e-01
1.07424915e+00 -2.12881660e+00 4.57948208e-01 -2.29615018e-01
3.00676137e-01 2.18380362e-01 -1.85747281e-01 3.88232648e-01
1.81333944e-01 -1.13620430e-01 -7.27689937e-02 -3.97863418e-01
-2.61349261e-01 1.59506321e-01 -6.42195404e-01 -1.01879891e-03
2.88570315e-01 1.24551213e+00 -6.75373971e-01 -7.74886966e-01
3.00013483e-01 3.37106824e-01 -2.67492056e-01 7.42601752e-01
-5.82787752e-01 6.21897578e-01 -2.32960537e-01 3.61509144e-01
6.47541821e-01 -6.36411726e-01 1.67826936e-01 -4.00018841e-01
-2.10109092e-02 -3.31273004e-02 -6.94184244e-01 1.59487450e+00
-4.46430445e-01 6.32474303e-01 1.76394299e-01 -5.41004062e-01
1.12368643e+00 7.77598694e-02 -2.10324004e-01 -1.19272196e+00
1.10302880e-01 -3.27304900e-01 -4.15144227e-02 -6.68465078e-01
6.45747662e-01 2.32425258e-01 -6.83755726e-02 3.69629294e-01
1.08224146e-01 1.22750886e-01 3.77817340e-02 6.47737384e-01
5.49961686e-01 -9.25148800e-02 2.76821047e-01 2.42414754e-02
1.06818247e+00 -2.53601540e-02 1.04473941e-01 7.59925723e-01
-4.30449307e-01 5.68569362e-01 6.37286246e-01 -3.52230936e-01
-8.00307870e-01 -1.17332625e+00 5.62904358e-01 1.40267503e+00
7.58099854e-01 -1.28247350e-01 -7.10139513e-01 -7.57793128e-01
-3.82845215e-02 8.95257294e-01 -8.64016652e-01 -3.94841880e-02
-6.26892865e-01 -5.14055826e-02 -1.03225648e-01 5.66192091e-01
6.14953578e-01 -1.51025367e+00 -8.30132127e-01 -1.00115053e-01
-4.97534990e-01 -1.13577211e+00 -7.17219174e-01 -2.31829181e-01
-3.97520512e-01 -1.07597899e+00 -8.66824389e-01 -9.76094306e-01
5.16124427e-01 6.21100068e-01 1.14778376e+00 2.87475646e-01
1.53660566e-01 6.45647049e-01 -1.50170833e-01 -1.16234265e-01
-3.60610157e-01 -1.34536266e-01 -6.27159357e-01 3.11069399e-01
2.31423587e-01 2.38038250e-03 -8.57224286e-01 3.64404738e-01
-8.80023599e-01 1.83936283e-01 6.52778625e-01 8.50177884e-01
7.32674539e-01 -8.15158665e-01 6.59657240e-01 -4.86470908e-01
8.17637265e-01 -4.47559088e-01 -6.30427599e-01 7.63686836e-01
-1.42889932e-01 4.54783469e-01 3.38346004e-01 -4.24849182e-01
-1.42864668e+00 -3.40838172e-02 3.98036372e-03 -7.08311558e-01
-6.28951266e-02 1.93287700e-01 -2.11247861e-01 2.12915704e-01
3.19056749e-01 2.41746813e-01 5.36583886e-02 -4.19032782e-01
6.73226058e-01 3.78521562e-01 8.38694930e-01 -3.35726768e-01
4.46161628e-01 4.39181924e-01 -2.94097751e-01 -6.58832967e-01
-8.49825501e-01 -4.32204962e-01 -6.62295282e-01 -4.88383979e-01
1.31953287e+00 -6.54710352e-01 -1.22289360e+00 4.59164716e-02
-1.61188233e+00 -1.50490060e-01 2.84817573e-02 4.96951304e-02
-3.57473999e-01 3.10186803e-01 -3.68698508e-01 -9.85719681e-01
-3.58478844e-01 -1.37516069e+00 1.07637811e+00 6.22753501e-01
-1.77944884e-01 -6.59128428e-01 -1.42621040e-01 6.84504032e-01
2.10649878e-01 -1.29904479e-01 1.09228063e+00 -5.40134013e-01
-1.02180457e+00 4.84896690e-01 -6.88736796e-01 -1.90580919e-01
-7.66554624e-02 -3.77055436e-01 -9.10239995e-01 -1.71956137e-01
9.49117839e-02 -3.61755252e-01 9.90337610e-01 1.11639023e-01
9.03507173e-01 -2.84813464e-01 -2.81480193e-01 1.21526487e-01
1.13424695e+00 6.46418095e-01 6.18014276e-01 1.71288565e-01
7.58059561e-01 7.93007970e-01 8.98702621e-01 1.53081477e-01
8.07807922e-01 8.01343620e-01 8.43315959e-01 -1.29887328e-01
-3.97185534e-01 -2.45940506e-01 -2.73213685e-02 3.30484629e-01
4.08691108e-01 -3.21583360e-01 -9.08650100e-01 7.76360750e-01
-1.96269274e+00 -8.33478570e-01 -2.17453614e-02 2.13906097e+00
6.82407439e-01 -7.89726824e-02 3.77238169e-02 -3.29753727e-01
8.54490697e-01 4.04516757e-01 -7.27297425e-01 -3.94817740e-01
-3.78684290e-02 -2.61693690e-02 -1.60572320e-01 8.77279401e-01
-7.55496085e-01 1.06840396e+00 5.70525169e+00 3.77362788e-01
-9.92759049e-01 1.19787790e-01 6.76485658e-01 1.10146791e-01
-3.32274288e-01 1.29069284e-01 -7.50679612e-01 3.17622215e-01
5.16042948e-01 -1.39386073e-01 4.86374527e-01 4.41526234e-01
-2.70661227e-02 -5.27217507e-01 -9.17489707e-01 9.56443369e-01
4.19509202e-01 -1.26332855e+00 4.27898079e-01 -2.88326383e-01
3.15014988e-01 -5.20235896e-01 3.80929440e-01 2.96977550e-01
1.23477608e-01 -9.89026487e-01 8.23218942e-01 7.97448635e-01
5.38636088e-01 -7.64653265e-01 5.58476806e-01 3.80077332e-01
-1.27790701e+00 -2.21902296e-01 -2.52577126e-01 1.46760926e-01
3.44283938e-01 -2.03996509e-01 -1.02667212e+00 4.61308479e-01
9.13817763e-01 1.89621001e-01 -8.70969594e-01 6.73323393e-01
-4.71733719e-01 3.87547165e-02 3.41880202e-01 -1.98353827e-01
3.60322863e-01 3.29523347e-02 5.03535509e-01 7.87298858e-01
8.61054808e-02 4.30123448e-01 1.83738898e-02 1.01463437e+00
-1.46714449e-01 1.48113236e-01 -2.04378769e-01 4.60397422e-01
4.86241341e-01 1.21356070e+00 -5.10131955e-01 -4.62078393e-01
-5.14401317e-01 1.18500364e+00 8.09963763e-01 7.17220247e-01
-8.82033765e-01 -2.09661067e-01 2.72790551e-01 -7.59418458e-02
6.12818837e-01 -2.81899162e-02 7.59969354e-02 -6.97782934e-01
2.63547916e-02 -8.37296605e-01 7.57296205e-01 -1.48015845e+00
-9.78134215e-01 1.00889289e+00 -8.01211447e-02 -8.76784742e-01
-4.57023889e-01 -3.58672470e-01 -4.69949007e-01 1.31366241e+00
-1.56411982e+00 -1.12541735e+00 -6.88877642e-01 7.39662588e-01
9.22373056e-01 1.78307697e-01 5.14167309e-01 -1.90387875e-01
-3.05657893e-01 3.81381750e-01 -5.31142592e-01 -9.97405648e-02
9.49773610e-01 -1.22278190e+00 3.22714478e-01 6.91667020e-01
2.46733770e-01 5.82571626e-01 6.21432424e-01 -6.24438882e-01
-1.19563305e+00 -8.21370065e-01 9.97441530e-01 -5.94533384e-01
5.36541343e-01 -3.29841554e-01 -1.48744488e+00 6.95898533e-01
9.36944842e-01 -1.68065578e-01 2.76359409e-01 -2.82842100e-01
-5.40166557e-01 -8.19167942e-02 -8.90313268e-01 7.93938041e-01
6.31955564e-01 -6.71844661e-01 -1.14517808e+00 -2.21348479e-02
1.02048314e+00 -5.93834579e-01 -3.75947773e-01 2.18603849e-01
4.19326216e-01 -9.80544209e-01 1.13710594e+00 -4.84894872e-01
2.81085759e-01 -4.87599492e-01 -5.56417704e-02 -1.00325692e+00
-3.58165920e-01 -4.64672565e-01 -1.34060875e-01 1.31657290e+00
3.05598050e-01 -2.28038773e-01 3.56776536e-01 5.62024951e-01
5.64077087e-02 -5.69277763e-01 -9.21129882e-01 -1.91240758e-01
-1.99130476e-01 1.85821250e-01 5.12566984e-01 5.31472802e-01
-2.16265008e-01 9.45799530e-01 -4.88322437e-01 3.72460812e-01
2.80613273e-01 5.54336727e-01 9.08879399e-01 -9.87867415e-01
-1.44102365e-01 -2.76123554e-01 1.01844527e-01 -1.51739097e+00
2.69660681e-01 -4.75682259e-01 2.14424312e-01 -1.79252517e+00
3.86049867e-01 1.56761408e-01 -2.42755473e-01 2.29602471e-01
-7.96203852e-01 5.61163947e-02 6.52608454e-01 3.22698534e-01
-9.53330457e-01 5.20090938e-01 1.42066467e+00 -4.46131766e-01
-4.80610907e-01 -3.80285054e-01 -7.44178236e-01 5.21585882e-01
4.52782094e-01 -3.82832848e-02 -5.40599644e-01 -5.45991063e-01
-1.36046857e-01 6.48799598e-01 7.00659633e-01 -7.13602602e-01
6.55740440e-01 -2.02389464e-01 5.32933235e-01 -1.06472516e+00
6.99767888e-01 -7.35005558e-01 1.64131615e-02 2.99789429e-01
-5.91777265e-01 5.28380752e-01 3.80651206e-01 7.23005414e-01
-3.37067425e-01 -1.47040069e-01 6.66343629e-01 -1.81386650e-01
-1.15955973e+00 7.74503350e-02 -6.58141598e-02 2.35397369e-01
1.11983049e+00 1.38748199e-01 -7.16426551e-01 -7.07725763e-01
-9.04620230e-01 6.07976675e-01 3.29718918e-01 6.34293139e-01
9.74672198e-01 -1.22275043e+00 -5.20727217e-01 -1.85028717e-01
3.77277642e-01 -2.14726001e-01 7.50935018e-01 6.04981124e-01
-2.09972322e-01 5.53650558e-01 -1.84877366e-01 -8.00157189e-01
-1.47168565e+00 1.30011880e+00 2.19021142e-01 -1.81543484e-01
-3.70656013e-01 7.99685121e-01 9.56544876e-01 2.01886207e-01
3.44344527e-01 -2.07415923e-01 -7.74762273e-01 1.96648881e-01
7.40077913e-01 -2.48239795e-03 -1.86181754e-01 -1.00746500e+00
-5.08605599e-01 6.94430292e-01 -2.27196470e-01 -4.45426404e-01
6.84540570e-01 -5.58460474e-01 1.41881704e-01 5.64109266e-01
8.78292501e-01 -2.57595498e-02 -1.52998412e+00 -4.30526406e-01
-2.17679024e-01 -5.36387086e-01 -4.13661659e-01 -9.44373250e-01
-8.89961779e-01 1.16319263e+00 4.50032443e-01 1.57799184e-01
1.23653138e+00 5.92096269e-01 5.15485883e-01 2.72169381e-01
2.43749581e-02 -6.60628915e-01 6.70130074e-01 4.60175931e-01
1.55572760e+00 -1.28095615e+00 -2.71036476e-01 -4.15333748e-01
-1.08604574e+00 7.48123109e-01 1.11295640e+00 1.74315244e-01
1.96805239e-01 -2.77990758e-01 3.58412564e-01 -3.25947940e-01
-1.05371571e+00 -4.42775846e-01 5.14357686e-01 5.02526879e-01
4.51714732e-03 -4.13602293e-01 -1.08189866e-01 4.80682492e-01
2.09719881e-01 -5.31314075e-01 1.95011750e-01 5.60783863e-01
-6.10588431e-01 -6.42357945e-01 -7.02353656e-01 -6.06501922e-02
-8.59135911e-02 -1.41738772e-01 -7.55993783e-01 8.00109327e-01
-2.24722460e-01 1.27629328e+00 3.49027455e-01 -1.57627150e-01
5.86462736e-01 2.05062956e-01 2.40921929e-01 -2.91136056e-01
-6.82834089e-01 -3.21980529e-02 -2.89228052e-01 -6.13912344e-01
-4.17914063e-01 -3.60308349e-01 -1.28320289e+00 -5.54887438e-03
-6.49121031e-02 2.38671854e-01 2.17696175e-01 9.44183826e-01
4.72295970e-01 6.65456593e-01 3.59890342e-01 -5.33763111e-01
-2.42653772e-01 -8.68363798e-01 -2.70419661e-02 6.56947553e-01
5.45360804e-01 -8.09682190e-01 -3.25292438e-01 5.36130667e-02] | [10.804866790771484, 1.6226246356964111] |
7a292b3c-85c1-4e57-9ed5-7306d8dc2dc1 | stock-price-forecast-with-deep-learning | 2103.14081 | null | https://arxiv.org/abs/2103.14081v1 | https://arxiv.org/pdf/2103.14081v1.pdf | Stock price forecast with deep learning | In this paper, we compare various approaches to stock price prediction using neural networks. We analyze the performance fully connected, convolutional, and recurrent architectures in predicting the next day value of S&P 500 index based on its previous values. We further expand our analysis by including three different optimization techniques: Stochastic Gradient Descent, Root Mean Square Propagation, and Adaptive Moment Estimation. The numerical experiments reveal that a single layer recurrent neural network with RMSprop optimizer produces optimal results with validation and test Mean Absolute Error of 0.0150 and 0.0148 respectively. | ['Ikhlaas Gurrib', 'Linda Smail', 'Firuz Kamalov'] | 2021-03-21 | null | null | null | null | ['stock-price-prediction'] | ['time-series'] | [-1.08055842e+00 -1.19594619e-01 -1.50680467e-01 -2.99023658e-01
-3.83591592e-01 -5.29024720e-01 5.30969381e-01 -1.98199958e-01
-4.07783478e-01 1.04692471e+00 1.35172531e-01 -6.40742958e-01
-3.08990479e-01 -8.31615746e-01 -3.17975372e-01 -4.96791363e-01
-6.34747982e-01 3.57532561e-01 1.64567865e-02 -5.04018903e-01
8.78329396e-01 7.37074494e-01 -8.87803674e-01 -1.24533057e-01
2.94363469e-01 2.03945041e+00 -5.27048588e-01 6.78457260e-01
-5.41807227e-02 1.40510011e+00 -8.38948429e-01 -7.95586050e-01
7.54996240e-01 -1.08416982e-01 -3.28814775e-01 -6.19875610e-01
-2.75290936e-01 -5.13002932e-01 -3.72394890e-01 9.86891806e-01
5.95073044e-01 3.37876737e-01 6.77394748e-01 -9.38756585e-01
-1.00842583e+00 1.12184775e+00 -5.99038482e-01 1.06549048e+00
-3.35479885e-01 3.33352759e-02 1.05515599e+00 -8.10244679e-01
2.47897774e-01 7.86701798e-01 9.99858439e-01 -1.33345157e-01
-9.13338363e-01 -7.90902019e-01 2.93813705e-01 1.06059603e-01
-1.28730905e+00 -1.15354076e-01 9.65662658e-01 -2.68896163e-01
1.58087134e+00 -5.86621696e-03 6.28248096e-01 4.43833619e-01
9.76217270e-01 3.92223626e-01 9.11708832e-01 7.95421451e-02
1.13732949e-01 2.28626996e-01 1.92116514e-01 4.55478847e-01
3.07535112e-01 4.14021969e-01 -3.15443844e-01 -1.05766833e-01
7.30875313e-01 3.43687162e-02 1.65912122e-01 6.24789715e-01
-9.64738548e-01 1.01466835e+00 4.49430883e-01 3.89679164e-01
-8.97989273e-01 3.66581768e-01 3.69928300e-01 6.97432697e-01
8.53078187e-01 6.31365657e-01 -9.66146052e-01 -2.10241318e-01
-1.47675622e+00 4.09294546e-01 1.15545928e+00 3.98934752e-01
1.34271845e-01 9.81311202e-01 -2.87224222e-02 3.35952133e-01
3.91623199e-01 5.32051563e-01 9.95744407e-01 -6.66342199e-01
3.42420667e-01 1.74633950e-01 3.03246588e-01 -1.06360471e+00
-7.00019062e-01 -1.11740696e+00 -8.20214748e-01 4.02487695e-01
1.11714773e-01 -1.01134133e+00 -7.42631912e-01 9.51915681e-01
-5.27009308e-01 4.65160310e-01 4.63355958e-01 3.29414845e-01
5.86727560e-01 8.30881417e-01 -3.31362098e-01 -3.32823545e-01
6.08669579e-01 -1.06759512e+00 -6.64015353e-01 1.56528249e-01
2.29142576e-01 -8.69132876e-01 4.99590002e-02 3.71840984e-01
-1.45659888e+00 -3.56470853e-01 -1.04674459e+00 5.32818258e-01
-8.17690849e-01 1.09071188e-01 4.36867446e-01 4.03515309e-01
-1.04265392e+00 1.35126805e+00 -5.88130653e-01 8.20743024e-01
3.41153085e-01 4.54425722e-01 5.86956322e-01 1.47837818e+00
-1.44013488e+00 1.16519856e+00 4.57488477e-01 5.72652161e-01
-3.30510557e-01 -8.37786734e-01 -1.78569213e-01 1.85436860e-01
-3.56242120e-01 -3.14133972e-01 1.49471176e+00 -7.16226280e-01
-2.01632190e+00 2.18147248e-01 2.09585741e-01 -1.55168617e+00
5.75942039e-01 -2.98513204e-01 -8.03935170e-01 -1.55067042e-01
-3.63690138e-01 3.03727567e-01 7.14886129e-01 -3.70692462e-01
-7.27255523e-01 -1.29079476e-01 -5.11553228e-01 -1.97012499e-01
2.17768773e-01 1.07426882e-01 4.00777012e-01 -1.17034209e+00
2.21788418e-02 -7.12159812e-01 -4.51600641e-01 -7.48484015e-01
-3.40893179e-01 -2.60178655e-01 4.41680372e-01 -9.07250524e-01
1.30293059e+00 -1.50903320e+00 -4.76406574e-01 8.08935463e-01
-2.15900540e-01 5.03951050e-02 3.32506359e-01 3.02256852e-01
-2.25789815e-01 1.78900450e-01 1.24901041e-01 -4.38615456e-02
1.48037940e-01 -2.87028313e-01 -9.73405421e-01 1.58862695e-01
1.66576177e-01 1.47764909e+00 -1.13684759e-01 1.35408491e-02
-1.00084960e-01 5.29962182e-01 -4.86000143e-02 -2.09303066e-01
-3.34588587e-01 -1.32230446e-01 -4.83576000e-01 6.72692835e-01
5.26197851e-01 -5.01103222e-01 -3.26616555e-01 5.06741107e-02
-4.74736959e-01 4.06817764e-01 -1.13066900e+00 6.76384509e-01
-3.52821022e-01 9.00478244e-01 -6.65834725e-01 -8.54065835e-01
1.45172036e+00 2.17358515e-01 3.37321967e-01 -6.12773240e-01
5.59354186e-01 6.87743604e-01 6.60013929e-02 1.26316547e-01
6.06962502e-01 -2.17030555e-01 2.92049140e-01 6.71308219e-01
-2.98563689e-01 1.84173658e-01 2.14989454e-01 -3.40229005e-01
4.14163440e-01 -1.31055698e-01 -1.54278325e-02 -2.89146632e-01
4.77125317e-01 -2.40691081e-01 3.16614628e-01 7.74520516e-01
-2.21294448e-01 2.04212680e-01 7.72147596e-01 -9.41873848e-01
-9.54499781e-01 -9.43193555e-01 -3.70265484e-01 7.46975064e-01
-4.24793810e-01 2.99886078e-01 -2.74142891e-01 -3.15942556e-01
4.10117626e-01 1.11253381e+00 -8.26291203e-01 1.54332206e-01
-5.73177040e-01 -1.29250610e+00 4.92105067e-01 7.44729638e-01
8.51685226e-01 -1.30292809e+00 -6.96765602e-01 5.03367543e-01
6.26584709e-01 -6.55264854e-01 -1.79069176e-01 4.51334625e-01
-1.25254917e+00 -5.03980517e-01 -1.10987103e+00 -5.50996304e-01
-1.33600747e-02 -8.37840557e-01 1.23230636e+00 -2.14425534e-01
3.69537830e-01 -1.80649400e-01 -7.12853521e-02 -7.28674591e-01
1.61993340e-01 2.66690850e-01 -9.22004580e-02 -1.35874018e-01
5.41151047e-01 -6.97532296e-01 -7.94744492e-01 -4.84427921e-02
-3.57914120e-01 -6.68323636e-01 5.39935589e-01 5.46978116e-01
5.40156245e-01 1.83665767e-01 8.46459389e-01 -5.16988635e-01
1.18796480e+00 -5.84037483e-01 -1.30095375e+00 9.08914953e-02
-1.48326027e+00 3.45447570e-01 5.96423686e-01 -2.31533363e-01
-8.52011800e-01 -3.74077797e-01 -1.84405297e-01 -3.57349128e-01
6.25169873e-01 9.24641490e-01 1.13351047e+00 -7.42295980e-02
2.55341589e-01 5.18394470e-01 5.78349736e-03 -6.67251289e-01
7.16107264e-02 3.39314789e-01 5.22528529e-01 2.08830252e-01
6.92441523e-01 2.50586778e-01 6.53575212e-02 -3.26100558e-01
-9.51121449e-01 2.27908716e-01 -6.09181702e-01 -8.78671855e-02
4.77038115e-01 -6.94808960e-01 -9.70661044e-01 7.19209552e-01
-9.13319588e-01 -1.56171933e-01 -3.55531961e-01 7.04669952e-01
-4.22222584e-01 -2.40334973e-01 -1.11827159e+00 -9.45028901e-01
-1.15010798e+00 -8.58809888e-01 3.41591537e-01 5.12248397e-01
-1.41549230e-01 -1.45705736e+00 2.84591645e-01 -3.84683192e-01
1.02542293e+00 4.84169215e-01 4.43458825e-01 -1.18299484e+00
-2.83171058e-01 -8.01861644e-01 -1.18525855e-01 4.90822732e-01
-3.60727876e-01 3.05060446e-01 -6.44112229e-01 8.69788602e-02
2.11132631e-01 3.89604792e-02 9.70650136e-01 9.74404871e-01
8.46529841e-01 -4.71902549e-01 2.09236965e-01 8.74033868e-01
1.52998757e+00 6.39596105e-01 5.92368782e-01 1.03297567e+00
1.20417669e-03 9.68546495e-02 9.00837258e-02 8.07449996e-01
5.76664209e-02 -1.02755539e-01 1.90679923e-01 4.06983227e-01
6.75418317e-01 -1.45490989e-01 3.12789977e-01 8.00134242e-01
-4.37129021e-01 4.55751419e-02 -8.38365436e-01 3.82004641e-02
-1.64612424e+00 -1.24910355e+00 2.22167835e-01 1.63742197e+00
5.79779804e-01 7.80641973e-01 1.95199355e-01 1.36544574e-02
3.24309409e-01 4.53278542e-01 -8.80412400e-01 -6.67815030e-01
-3.54732603e-01 4.38174516e-01 1.17900312e+00 3.62449169e-01
-1.20563102e+00 8.08668733e-01 7.60446835e+00 4.57341343e-01
-1.34409821e+00 -3.82583588e-01 1.28390729e+00 -3.40180099e-01
-1.67797487e-02 -4.15766001e-01 -1.20485365e+00 8.24582040e-01
1.62006342e+00 -6.77347839e-01 2.97102720e-01 1.00163269e+00
2.24881247e-01 1.99877605e-01 -4.46804017e-01 9.08422887e-01
-3.09815347e-01 -1.88849092e+00 3.80369760e-02 5.64218760e-02
1.14093792e+00 7.06089675e-01 6.46778524e-01 4.30067986e-01
4.78095472e-01 -1.09517384e+00 7.52736092e-01 1.37480175e+00
-1.12953544e-01 -1.31847024e+00 1.40847778e+00 2.65095439e-02
-1.08582449e+00 -2.56994218e-01 -5.16492069e-01 -4.07438755e-01
3.48717391e-01 6.71205401e-01 -2.75025636e-01 1.63341790e-01
6.90338910e-01 7.37880170e-01 -4.77013677e-01 9.31944549e-01
-1.42416134e-01 4.46094513e-01 -5.36031306e-01 -6.23454094e-01
8.29144478e-01 -4.78058487e-01 3.16303194e-01 9.50346649e-01
4.12123114e-01 -1.08425748e-02 -2.99048245e-01 8.99527252e-01
-6.04579635e-02 1.56102628e-01 -1.70997173e-01 3.30914967e-02
2.38425598e-01 7.96655297e-01 -8.09649587e-01 -5.69344044e-01
-3.19004744e-01 5.32433093e-01 -8.87101814e-02 5.03291488e-01
-8.38354528e-01 -9.08816636e-01 5.43452859e-01 -2.93172210e-01
8.25603604e-01 -1.51541471e-01 -7.04955935e-01 -9.66927946e-01
-5.34512736e-02 -3.10530663e-01 3.07186216e-01 -6.28619611e-01
-1.36033642e+00 9.10643041e-01 -2.12807253e-01 -9.80860174e-01
-8.67618859e-01 -1.02884793e+00 -1.11651397e+00 1.17169392e+00
-1.86122680e+00 -1.65710896e-01 6.10416532e-01 2.46238440e-01
2.28218824e-01 -1.08335495e+00 4.11453664e-01 -1.19883366e-01
-8.86724532e-01 5.16513467e-01 7.90649474e-01 5.88580728e-01
-2.84128845e-01 -1.15715373e+00 9.83842552e-01 6.27036631e-01
3.47123109e-02 4.34719294e-01 7.04977155e-01 -7.34628201e-01
-7.85151660e-01 -9.57415521e-01 9.27042842e-01 1.91809740e-02
1.42421472e+00 6.20664716e-01 -8.27686429e-01 9.04496193e-01
5.45093358e-01 -1.15294449e-01 2.74637997e-01 -1.75228938e-01
-1.11604027e-01 -1.69604346e-01 -1.08853686e+00 3.90106469e-01
5.93479834e-02 -1.98350638e-01 -6.92772865e-01 1.92937165e-01
4.70080107e-01 -2.71239817e-01 -1.29624343e+00 2.81106174e-01
6.85734928e-01 -1.16126740e+00 8.81846011e-01 -5.95939755e-01
3.10374320e-01 1.52742624e-01 1.27037361e-01 -1.24946761e+00
-1.69904009e-01 -8.66748929e-01 -5.80827177e-01 5.88515282e-01
1.20942461e+00 -1.35826409e+00 1.03538477e+00 8.33717406e-01
2.58896142e-01 -1.42876947e+00 -8.02479267e-01 -9.84118402e-01
6.91640735e-01 -4.47240025e-01 9.61884141e-01 5.77163815e-01
-6.00200631e-02 -6.89633051e-03 -2.30520606e-01 -4.68659736e-02
3.55723172e-01 5.51872373e-01 3.33019830e-02 -1.08581555e+00
-2.57633448e-01 -1.17573988e+00 -3.41313779e-01 -9.30875838e-01
2.52552062e-01 -5.54045856e-01 -6.47435665e-01 -1.08807957e+00
-4.41894472e-01 7.74425268e-02 -1.02958107e+00 4.23162170e-02
4.03516293e-01 1.00099020e-01 2.76611093e-02 4.99136388e-01
-8.40307102e-02 4.79113579e-01 7.46715426e-01 6.69637322e-02
-3.93180639e-01 6.14149690e-01 -2.47181967e-01 9.16423142e-01
1.70012581e+00 -1.55509517e-01 -5.42481728e-02 -1.29195094e-01
6.65549695e-01 1.09334225e-02 -3.84627469e-02 -7.68585503e-01
1.12782843e-01 1.53184101e-01 9.35120881e-01 -1.22020805e+00
7.80397505e-02 -3.02203864e-01 3.40013835e-03 8.43749702e-01
-5.57193160e-01 1.03915274e+00 2.55022556e-01 2.33908564e-01
-4.77014512e-01 -6.07498705e-01 7.57512629e-01 -2.87421495e-01
-4.12852556e-01 2.94818819e-01 -3.76416117e-01 7.44500160e-02
9.08536911e-01 -4.23294008e-02 -1.56060725e-01 -6.53001249e-01
-8.69888604e-01 3.26355547e-01 -4.39376980e-01 2.78155953e-01
7.34280646e-01 -1.25712109e+00 -7.87700117e-01 2.95936763e-02
-8.35661888e-01 -8.36753726e-01 -4.25020635e-01 7.78557658e-01
-1.02689457e+00 1.01631570e+00 -4.82446812e-02 3.04366857e-01
-4.99554098e-01 3.48668963e-01 1.08685601e+00 -5.92716694e-01
-4.72155035e-01 1.04736066e+00 -7.94036329e-01 2.12544892e-02
3.59515518e-01 -6.43567622e-01 -6.36574328e-01 4.46493089e-01
5.92340469e-01 8.70039940e-01 9.73071232e-02 -5.40608704e-01
-9.45333764e-02 7.00231910e-01 -3.20105970e-01 -3.45451921e-01
1.70399261e+00 1.43839628e-01 -8.58685449e-02 8.49800408e-01
1.31878340e+00 -4.25120622e-01 -1.08448505e+00 -1.20363645e-01
5.61075628e-01 1.36206493e-01 4.68215048e-01 -8.82037044e-01
-1.56340599e+00 4.96355742e-01 7.10303366e-01 7.54643559e-01
8.90423894e-01 -4.88566816e-01 1.07445431e+00 8.47122192e-01
-8.27065930e-02 -1.50153792e+00 -5.07707357e-01 8.66850436e-01
7.81694114e-01 -1.02779043e+00 1.59182563e-01 6.22287095e-01
-6.81628704e-01 1.52713990e+00 -1.20630518e-01 -9.18964446e-01
1.67698610e+00 3.91445667e-01 4.09468263e-01 -2.99550503e-01
-1.18867564e+00 7.30279312e-02 4.01160389e-01 -1.38137534e-01
6.02229416e-01 -3.48331667e-02 -1.27344340e-01 5.31621635e-01
-1.03904724e+00 -9.37331915e-02 4.84994322e-01 6.85374796e-01
-5.05737305e-01 -3.28856617e-01 -2.04564318e-01 9.37696874e-01
-1.23458993e+00 -4.76767242e-01 -2.28091747e-01 8.10338676e-01
-7.13518679e-01 4.35258448e-01 5.96503735e-01 -2.13559389e-01
3.47064972e-01 4.14679319e-01 -1.25227004e-01 5.56133054e-02
-1.03247762e+00 1.18923798e-01 -1.28379002e-01 -2.31097728e-01
-1.88713539e-02 -8.93596113e-01 -1.34580815e+00 -5.73733866e-01
-2.67391354e-01 2.89741099e-01 7.91057944e-01 7.43863225e-01
3.00365597e-01 4.98745501e-01 9.70084548e-01 -6.65490031e-01
-1.20794952e+00 -1.13319564e+00 -9.73188877e-01 -3.09203386e-01
5.09013772e-01 -2.06198260e-01 -7.02443898e-01 -2.01133415e-01] | [4.45042610168457, 4.232295989990234] |
6e179221-cc97-42ba-9cee-6816318dc6e1 | multi-resolution-autoregressive-graph-to | 1907.11223 | null | https://arxiv.org/abs/1907.11223v2 | https://arxiv.org/pdf/1907.11223v2.pdf | Hierarchical Graph-to-Graph Translation for Molecules | The problem of accelerating drug discovery relies heavily on automatic tools to optimize precursor molecules to afford them with better biochemical properties. Our work in this paper substantially extends prior state-of-the-art on graph-to-graph translation methods for molecular optimization. In particular, we realize coherent multi-resolution representations by interweaving the encoding of substructure components with the atom-level encoding of the original molecular graph. Moreover, our graph decoder is fully autoregressive, and interleaves each step of adding a new substructure with the process of resolving its attachment to the emerging molecule. We evaluate our model on multiple molecular optimization tasks and show that our model significantly outperforms previous state-of-the-art baselines. | ['Wengong Jin', 'Regina Barzilay', 'Tommi Jaakkola'] | 2019-06-11 | null | https://openreview.net/forum?id=rJeeKTNKDB | https://openreview.net/pdf?id=rJeeKTNKDB | null | ['graph-to-graph-translation'] | ['graphs'] | [ 6.75404966e-01 3.20915610e-01 -6.31526291e-01 -2.22334251e-01
-9.00022507e-01 -6.77220047e-01 3.62160623e-01 8.75645459e-01
-3.66226315e-01 1.17365718e+00 3.68315995e-01 -7.51993656e-01
2.62366295e-01 -5.92544854e-01 -1.32973361e+00 -5.33506155e-01
-1.22849397e-01 6.42952681e-01 -1.54670373e-01 -2.73002207e-01
2.91881889e-01 6.52045250e-01 -4.10502613e-01 3.86152893e-01
8.91837955e-01 2.60694504e-01 -3.96297015e-02 7.73151338e-01
-1.10481210e-01 5.82483709e-01 -3.74702662e-01 -7.74493933e-01
-4.36238572e-02 -6.82452440e-01 -8.09552670e-01 -3.18468899e-01
5.30898809e-01 1.20934965e-02 -5.31872869e-01 8.21447074e-01
5.27441204e-01 3.63127552e-02 7.44143844e-01 -4.98991668e-01
-1.23172128e+00 4.93628085e-01 -5.85240901e-01 2.57394910e-01
4.20852244e-01 2.88424283e-01 1.33311665e+00 -8.34789038e-01
1.15802705e+00 1.05960250e+00 3.68213058e-01 5.66352129e-01
-1.80119371e+00 -4.56149906e-01 3.76833916e-01 2.80957013e-01
-1.22969115e+00 -7.13721991e-01 5.05334735e-01 -3.80106270e-01
1.79460001e+00 -1.16480909e-01 4.62303370e-01 9.37337399e-01
8.20106983e-01 -1.70873646e-02 4.34341669e-01 -2.29313001e-01
4.50200401e-04 -6.16290748e-01 7.17489421e-02 1.18109548e+00
3.56847018e-01 -1.04833841e-01 -6.92851126e-01 -3.50460112e-01
5.16751170e-01 -1.29483804e-01 -1.17367558e-01 -2.48473331e-01
-9.79778945e-01 8.19613576e-01 5.61706841e-01 4.31410410e-02
-6.24390066e-01 4.92009789e-01 -1.23865129e-02 8.18869025e-02
3.69938254e-01 7.48264670e-01 -4.80982214e-01 1.70007020e-01
-6.92266405e-01 1.47566959e-01 8.44840288e-01 8.74125183e-01
5.82843840e-01 4.25616046e-03 -2.62086719e-01 3.67803276e-01
3.89234126e-01 -1.76223412e-01 -1.23501346e-01 -5.78411937e-01
4.46914256e-01 3.71533453e-01 -1.05347961e-01 -5.09200811e-01
-4.22471881e-01 -6.64572179e-01 -5.13824165e-01 -1.83293417e-01
9.26823616e-02 -6.70792013e-02 -1.03732765e+00 1.83746123e+00
4.45060998e-01 3.96710843e-01 8.92385989e-02 4.29512858e-01
8.68526816e-01 9.39681590e-01 6.26526475e-01 -4.81509805e-01
1.28626561e+00 -1.16878068e+00 -6.23725891e-01 -1.56707317e-01
6.08690321e-01 -7.63757646e-01 4.82407987e-01 1.70022786e-01
-1.30097365e+00 -2.85924464e-01 -1.17447984e+00 -4.34473932e-01
-3.51126045e-01 -2.50614494e-01 9.37545061e-01 2.31667146e-01
-9.53956485e-01 1.01814985e+00 -7.60144889e-01 1.29213437e-01
5.14756858e-01 6.36568129e-01 -6.26233876e-01 -1.50351031e-02
-1.03183341e+00 1.08129132e+00 3.64910454e-01 -6.17658682e-02
-9.50456083e-01 -1.18789315e+00 -7.82272041e-01 2.25731283e-02
3.99086654e-01 -1.30615127e+00 1.09769619e+00 -5.57262361e-01
-1.69776142e+00 6.32662296e-01 -6.53669298e-01 -5.27847767e-01
-5.56675382e-02 1.31701693e-01 -2.87505358e-01 -5.21265604e-02
-5.13538532e-02 6.15757585e-01 6.15925670e-01 -7.15206504e-01
-7.33870268e-02 -3.25011879e-01 6.67836575e-04 1.62598118e-01
2.46712640e-01 5.86207807e-02 -6.80338502e-01 -7.04765201e-01
-2.21158430e-01 -9.23613846e-01 -8.42403650e-01 -1.89293250e-01
-6.19999468e-01 1.49907589e-01 -1.56104982e-01 -7.11027741e-01
1.34833050e+00 -1.55793118e+00 1.10597491e+00 2.30503455e-01
5.99705935e-01 1.29206076e-01 -4.00595039e-01 7.53569722e-01
-6.12839580e-01 5.43756485e-01 -2.31985062e-01 -3.16727549e-01
-2.50118256e-01 2.06891764e-02 -7.55942389e-02 5.05688369e-01
4.65804845e-01 1.26464593e+00 -1.00788140e+00 -2.64747947e-01
-2.99848497e-01 7.23266304e-01 -1.02976596e+00 8.63900036e-02
-5.92675030e-01 6.00817144e-01 -6.83264375e-01 7.15960979e-01
1.27854526e-01 -6.32462263e-01 6.75936639e-01 -4.83617485e-01
-2.97235535e-03 7.37857103e-01 -4.35520142e-01 1.92878950e+00
-4.25153784e-02 2.26957109e-02 -2.63733238e-01 -6.75563037e-01
4.15888131e-01 1.59585550e-01 5.34159422e-01 -4.31720585e-01
-1.60382211e-01 4.19169255e-02 3.69054079e-01 9.53848809e-02
4.60414529e-01 -1.77554965e-01 2.21551582e-01 1.91885054e-01
1.88369513e-01 -9.08865929e-02 3.19122463e-01 4.05168623e-01
1.17118096e+00 5.04453182e-01 6.28886998e-01 7.23356903e-02
4.47004765e-01 -1.61994562e-01 4.28197384e-01 4.39415306e-01
2.59987086e-01 1.83209062e-01 6.44826055e-01 -4.45517033e-01
-1.24515212e+00 -1.00481582e+00 1.10999368e-01 1.12253034e+00
-3.81441087e-01 -9.15634155e-01 -7.40235031e-01 -7.09010303e-01
-1.31566850e-02 5.78313529e-01 -5.61267614e-01 -2.72895485e-01
-7.52314329e-01 -1.10332191e+00 5.40950537e-01 1.22371048e-01
-3.35226089e-01 -6.15381658e-01 3.24158937e-01 9.00441289e-01
2.07887813e-01 -9.78826642e-01 -1.10279262e+00 4.22235936e-01
-9.07101572e-01 -9.83375847e-01 -3.60918224e-01 -6.55062199e-01
8.36578906e-01 -3.01924758e-02 1.26044714e+00 2.14719623e-01
-3.73509496e-01 -4.08745259e-01 4.28300463e-02 -3.82873684e-01
-6.43892169e-01 2.27377906e-01 -1.67284042e-01 -2.28954256e-01
-1.06766470e-01 -6.84872687e-01 -5.44059098e-01 -2.73531944e-01
-7.97380745e-01 3.07359606e-01 5.82031369e-01 6.04338884e-01
1.32127845e+00 -5.92999995e-01 5.20080268e-01 -1.24375641e+00
6.46228969e-01 -3.22310209e-01 -6.78341687e-01 4.09983367e-01
-6.80024743e-01 8.57459784e-01 8.15206885e-01 -3.09408784e-01
-5.07977307e-01 2.77874321e-01 -3.33606362e-01 -9.14857015e-02
3.87937963e-01 7.80790150e-01 -3.67690772e-02 -4.56357032e-01
5.04768670e-01 9.84891355e-02 -1.93646327e-01 -4.27003205e-01
8.39535654e-01 -8.61552171e-03 4.24792677e-01 -6.24611914e-01
5.59960425e-01 6.22037351e-02 5.97352624e-01 -6.98990703e-01
-5.48356056e-01 -1.11825958e-01 -6.48177564e-01 3.89633238e-01
1.02369273e+00 -9.03211176e-01 -8.62884998e-01 2.93625835e-02
-1.50643790e+00 -2.11013108e-01 -6.10364676e-02 2.64850855e-01
-4.14375246e-01 5.02596319e-01 -9.08707082e-01 -1.59757376e-01
-5.74334323e-01 -1.58856666e+00 1.14060581e+00 -5.73207512e-02
-1.38169006e-01 -9.01277304e-01 5.53679645e-01 4.00565386e-01
9.54757445e-03 3.48921776e-01 1.44660580e+00 -5.17336369e-01
-9.38710153e-01 1.26515115e-02 9.64444205e-02 -3.19917232e-01
1.44660249e-01 1.84320286e-01 -3.54210109e-01 -1.43612817e-01
-7.81701624e-01 -8.81034732e-02 1.13110018e+00 4.13756996e-01
9.85297441e-01 -5.45623541e-01 -5.16892672e-01 1.00378942e+00
1.15730286e+00 2.88228750e-01 6.58638299e-01 9.39782262e-02
1.13354027e+00 1.20896585e-01 1.25707120e-01 1.08801290e-01
3.67025495e-01 1.00872254e+00 3.51897210e-01 -3.45231384e-01
-1.79253101e-01 -5.48372567e-01 4.41064179e-01 7.20947921e-01
-3.69520724e-01 -4.81551766e-01 -5.00617445e-01 1.19796908e-02
-1.69639373e+00 -7.96961129e-01 1.16480432e-01 1.96036041e+00
1.33720720e+00 1.22958273e-01 1.37416488e-02 -7.20451474e-01
4.26607847e-01 3.56744051e-01 -9.18029785e-01 -6.50532842e-01
-4.85839657e-02 8.63703609e-01 9.54081714e-01 1.11854565e+00
-8.47145796e-01 1.53048146e+00 7.12348509e+00 6.44590735e-01
-8.33781302e-01 -1.11360602e-01 8.02854836e-01 -1.34194434e-01
-4.51326460e-01 2.26005614e-01 -9.49739218e-01 1.63498074e-01
1.29667163e+00 -1.89813212e-01 8.09333265e-01 4.08607990e-01
3.12363863e-01 6.07342482e-01 -1.46084571e+00 6.73417509e-01
-4.39632088e-02 -2.14527488e+00 5.79103649e-01 3.33652139e-01
6.39430523e-01 6.94488212e-02 -9.91437212e-03 -1.64655522e-01
5.41773021e-01 -1.43985999e+00 4.35013920e-01 6.57692671e-01
8.17428052e-01 -7.60246396e-01 1.85978204e-01 -2.83549249e-01
-1.18343568e+00 6.08267069e-01 -2.37674922e-01 1.91506788e-01
2.86547005e-01 2.70057023e-01 -1.06814587e+00 8.23397100e-01
-2.62748957e-01 9.47492182e-01 -4.70333874e-01 6.11923218e-01
-4.82356280e-01 3.07354450e-01 7.69007802e-02 5.81844263e-02
2.61090279e-01 -7.16509283e-01 5.15714228e-01 1.21446538e+00
1.26085822e-02 4.24778104e-01 3.26335013e-01 9.25093770e-01
-8.04102838e-01 3.25178027e-01 -5.77234447e-01 -6.95117295e-01
3.32545877e-01 9.00947869e-01 -3.06533337e-01 -4.42094892e-01
-4.98225600e-01 1.18479085e+00 8.06885898e-01 4.52115059e-01
-9.33106720e-01 -2.53348470e-01 9.99201715e-01 -4.55740504e-02
3.17461133e-01 -4.07672673e-01 -1.00521088e-01 -1.12331092e+00
-4.00590032e-01 -1.26657569e+00 3.79184932e-01 -6.38692677e-01
-9.21352327e-01 4.17615592e-01 -5.82439423e-01 -3.17458361e-01
-1.32846078e-02 -4.73095715e-01 -3.97996157e-01 1.04544544e+00
-1.46704316e+00 -1.06511843e+00 6.58200264e-01 8.29394385e-02
4.73880023e-01 3.97311039e-02 1.01387835e+00 3.20720643e-01
-9.08287764e-01 7.58767724e-01 2.90780840e-03 -3.54680508e-01
6.93155050e-01 -1.02437603e+00 1.09195077e+00 8.06631505e-01
2.94582158e-01 1.04162121e+00 7.85353482e-01 -1.03343570e+00
-1.84603834e+00 -1.18419385e+00 9.23096538e-01 -6.77711368e-01
8.82362902e-01 -2.50602782e-01 -9.26397681e-01 7.99687564e-01
4.76956144e-02 -3.32297623e-01 8.79763663e-01 1.62077889e-01
-4.98401076e-01 2.50978708e-01 -6.95801079e-01 8.64865899e-01
1.17587662e+00 -6.43679380e-01 -3.00862312e-01 8.05954576e-01
1.22266090e+00 -5.42001903e-01 -1.12897396e+00 2.28369042e-01
4.67960566e-01 -2.19398528e-01 1.52253330e+00 -1.57458627e+00
5.76018035e-01 -3.00007015e-01 -6.19699545e-02 -1.23622203e+00
-7.76744902e-01 -1.20200706e+00 -3.77454937e-01 4.52333838e-01
1.15959466e+00 -4.88892436e-01 8.08758438e-01 2.94742435e-01
-4.54029560e-01 -1.02247345e+00 -8.63757551e-01 -2.92288780e-01
5.11785485e-02 9.20947418e-02 4.59998250e-01 7.37429917e-01
9.25675854e-02 1.13404930e+00 -5.32550216e-01 1.19250730e-01
5.05728781e-01 -3.63927037e-02 6.80969000e-01 -7.31945872e-01
-9.68477666e-01 -5.92899501e-01 -1.15414217e-01 -1.04365325e+00
2.24152833e-01 -1.27178144e+00 -3.60735059e-01 -1.56745696e+00
4.59208250e-01 2.09967077e-01 -4.17576522e-01 4.85550284e-01
-6.00264847e-01 6.95385262e-02 9.03726965e-02 5.63616939e-02
-5.45603633e-01 5.37301779e-01 1.33103335e+00 -4.10562575e-01
-5.09308159e-01 -4.27068114e-01 -1.02445102e+00 2.90507108e-01
5.57011187e-01 -7.35121787e-01 -1.17731027e-01 -2.94778407e-01
4.02871221e-01 3.08768898e-01 -1.84924111e-01 -2.06476733e-01
1.24814086e-01 -3.71872157e-01 3.24969828e-01 -6.98254406e-02
4.60377306e-01 -2.14096084e-01 6.37100756e-01 7.60475695e-01
-5.34738302e-01 3.35077167e-01 2.49479055e-01 9.07675564e-01
3.12597007e-01 6.20294921e-03 8.12942803e-01 -1.57855287e-01
-3.27580911e-03 7.93583810e-01 -2.22210661e-01 -1.25749871e-01
6.69267237e-01 8.16652104e-02 -3.62562060e-01 -5.13430350e-02
-8.11175287e-01 -6.90866262e-02 7.05224931e-01 1.27866000e-01
4.70282137e-01 -8.42818379e-01 -6.67340636e-01 -1.18571632e-01
-8.27913452e-03 -4.37455595e-01 -2.77629524e-01 5.77060938e-01
-8.15733790e-01 7.40007758e-01 5.32052219e-02 4.74957153e-02
-1.40918028e+00 7.94711530e-01 3.77305448e-01 -7.01007843e-01
-1.69516563e-01 7.79545486e-01 3.68904412e-01 -1.86151974e-02
-1.76750012e-02 -3.36898625e-01 -1.84913091e-02 -2.92797923e-01
3.58056098e-01 -7.93310534e-03 1.56609163e-01 -5.88492572e-01
-5.72710514e-01 6.97384119e-01 -6.36352181e-01 2.53949106e-01
1.54703617e+00 4.61266309e-01 -3.48238707e-01 -7.10546747e-02
1.20744109e+00 1.44781291e-01 -9.91781473e-01 -4.50959653e-02
-1.34788752e-01 1.30959630e-01 8.16702172e-02 -8.84108782e-01
-5.17449975e-01 4.29900855e-01 1.15054734e-01 -7.71146894e-01
5.85896671e-01 -1.72898471e-02 6.94165766e-01 8.18901479e-01
1.30125701e-01 -7.36628830e-01 3.57363261e-02 4.61113304e-01
7.21534610e-01 -9.76884902e-01 6.03536010e-01 -5.94969630e-01
-3.80970895e-01 9.13698196e-01 2.80089676e-02 -7.27462117e-03
2.28467807e-01 -1.48563385e-02 -6.36478066e-01 -4.53252077e-01
-1.02266395e+00 -1.45990163e-01 6.30295753e-01 3.22111845e-01
9.74546671e-01 1.18649311e-01 -5.32830298e-01 3.99377704e-01
1.85901016e-01 -1.23767108e-01 2.99210638e-01 8.70637536e-01
-4.76383269e-01 -1.87199128e+00 3.22066061e-02 2.39417478e-01
-9.38682556e-01 -9.61089671e-01 -9.58830953e-01 3.67915362e-01
-1.65488809e-01 7.56252646e-01 -5.96777499e-01 -1.70864418e-01
4.97240365e-01 9.31787416e-02 9.78067398e-01 -9.39581871e-01
-6.73134267e-01 3.68228257e-01 4.32602227e-01 -6.95483208e-01
-7.82144740e-02 -3.58495623e-01 -1.45276022e+00 -3.04682910e-01
-2.25810409e-01 3.57656538e-01 6.62805617e-01 6.52123332e-01
9.96987879e-01 9.74848807e-01 2.81196743e-01 -7.86623061e-01
-1.61641747e-01 -4.44151253e-01 2.07716655e-02 2.31091995e-02
3.64959836e-01 -2.90257663e-01 3.25946510e-01 3.06710064e-01] | [4.770712375640869, 5.928679466247559] |
bcc58847-db67-41f9-8d2e-a73fbd169569 | the-news-delivery-channel-recommendation | 2306.10022 | null | https://arxiv.org/abs/2306.10022v1 | https://arxiv.org/pdf/2306.10022v1.pdf | The News Delivery Channel Recommendation Based on Granular Neural Network | With the continuous maturation and expansion of neural network technology, deep neural networks have been widely utilized as the fundamental building blocks of deep learning in a variety of applications, including speech recognition, machine translation, image processing, and the creation of recommendation systems. Therefore, many real-world complex problems can be solved by the deep learning techniques. As is known, traditional news recommendation systems mostly employ techniques based on collaborative filtering and deep learning, but the performance of these algorithms is constrained by the sparsity of the data and the scalability of the approaches. In this paper, we propose a recommendation model using granular neural network model to recommend news to appropriate channels by analyzing the properties of news. Specifically, a specified neural network serves as the foundation for the granular neural network that the model is considered to be build. Different information granularities are attributed to various types of news material, and different information granularities are released between networks in various ways. When processing data, granular output is created, which is compared to the interval values pre-set on various platforms and used to quantify the analysis's effectiveness. The analysis results could help the media to match the proper news in depth, maximize the public attention of the news and the utilization of media resources. | ['Wong-Hing Lam', 'Jiaxuan Liu', 'Rui Li', 'Lin Wu'] | 2023-05-30 | null | null | null | null | ['collaborative-filtering', 'machine-translation'] | ['miscellaneous', 'natural-language-processing'] | [-4.14751738e-01 -5.16982675e-01 -5.31111538e-01 -4.13458735e-01
-3.46375965e-02 -1.36994943e-01 6.27544701e-01 3.47426564e-01
-2.14860648e-01 4.29672331e-01 6.53422356e-01 -2.42098927e-01
-3.67916644e-01 -1.25593781e+00 -5.11293769e-01 -5.02100468e-01
2.28080571e-01 3.95485371e-01 2.13657349e-01 -4.16195214e-01
4.70827430e-01 2.29128301e-01 -1.86691332e+00 9.74866807e-01
6.33465409e-01 1.73848546e+00 2.48804256e-01 -1.28257684e-02
-8.71535897e-01 7.47171938e-01 -6.76922202e-01 -2.23913953e-01
1.31997511e-01 -2.92308122e-01 -2.58298606e-01 -1.01839051e-01
-1.16851889e-01 -4.98447925e-01 -3.60078663e-01 1.27856207e+00
4.11140233e-01 2.82482803e-01 4.13664371e-01 -5.34768999e-01
-1.00036418e+00 1.19576502e+00 -2.42213473e-01 4.92367536e-01
-2.67431848e-02 -4.55721915e-01 8.38492811e-01 -7.65308619e-01
3.00307363e-01 1.09325325e+00 4.39708024e-01 6.03246242e-02
-2.94552118e-01 -7.47799695e-01 2.24324822e-01 3.84445161e-01
-8.86285067e-01 -2.74573475e-01 7.13140965e-01 -5.45680106e-01
4.66908514e-01 -2.31690481e-02 8.26187015e-01 8.74232709e-01
3.01381856e-01 6.43296361e-01 9.74000096e-01 -9.89901721e-02
4.10364747e-01 2.55769461e-01 1.99845478e-01 3.95539761e-01
1.46547973e-01 -1.55911762e-02 -4.40891862e-01 8.78091082e-02
7.98313916e-01 6.68672502e-01 -8.23473409e-02 3.08352888e-01
-9.01280940e-01 1.03790998e+00 4.74410057e-01 3.28731656e-01
-5.83690464e-01 -4.92917448e-01 6.93643987e-01 4.96909171e-01
7.71055520e-01 2.00173289e-01 -3.21828634e-01 2.27873892e-01
-8.85452867e-01 1.23629808e-01 6.12936020e-01 5.77600539e-01
3.94446969e-01 5.75700067e-02 -1.74271658e-01 8.40591908e-01
2.54908144e-01 1.28546983e-01 1.04296601e+00 -3.33754569e-01
4.43682730e-01 8.33409607e-01 -8.20155255e-03 -1.39551389e+00
-2.71153033e-01 -9.68816876e-01 -9.77388442e-01 -1.31080225e-01
1.41059816e-01 -3.16532224e-01 -7.08789051e-01 1.08183098e+00
3.59444290e-01 1.62107557e-01 -4.40255553e-02 1.06410277e+00
1.10564542e+00 1.12267017e+00 -6.06879853e-02 -1.39413670e-01
1.24987841e+00 -6.53158426e-01 -6.99600279e-01 4.46675941e-02
3.06215972e-01 -8.97458196e-01 8.27320993e-01 5.94980240e-01
-7.34668791e-01 -7.32324362e-01 -9.87906098e-01 2.73503423e-01
-4.60056037e-01 1.99448809e-01 8.23922038e-01 2.68503636e-01
-6.38724983e-01 6.95482552e-01 -2.34146550e-01 -1.26088634e-01
5.87054908e-01 1.10005043e-01 2.52991408e-01 -9.21919271e-02
-1.65250206e+00 5.83211839e-01 5.89382648e-01 2.57279903e-01
-7.20488548e-01 -2.35546172e-01 -2.45546550e-01 4.03547823e-01
2.17226803e-01 -3.77755195e-01 9.01075363e-01 -1.17986917e+00
-1.42884135e+00 1.48151278e-01 6.57774925e-01 -4.62825030e-01
2.81371653e-01 -2.21009832e-02 -8.05399239e-01 -2.46908590e-01
-1.17281623e-01 -6.51849657e-02 5.24424851e-01 -8.00934672e-01
-1.22866750e+00 -1.88938543e-01 2.50360906e-01 3.94462705e-01
-7.39048660e-01 4.05911267e-01 -4.67485249e-01 -6.61188781e-01
6.96973726e-02 -4.04105127e-01 -2.73429789e-02 -4.40726668e-01
9.43455752e-03 -2.42785618e-01 6.32616341e-01 -5.58950543e-01
1.15186763e+00 -2.28421688e+00 -1.14979908e-01 5.51843643e-01
1.13318704e-01 1.95336878e-01 9.85926539e-02 1.96908027e-01
4.39149678e-01 -7.66853094e-02 5.15434146e-01 1.37474999e-01
-1.30493402e-01 1.68981507e-01 -6.73570931e-01 1.72852024e-01
-4.51657802e-01 3.15204948e-01 -6.78057492e-01 -2.46218458e-01
5.84323071e-02 3.11546147e-01 -3.30844581e-01 3.55897367e-01
-4.84201133e-01 2.13696688e-01 -8.55755806e-01 5.34947455e-01
2.02745169e-01 -3.88320118e-01 -7.35982060e-02 -3.40449452e-01
-3.10738534e-01 4.63521332e-01 -1.08483708e+00 1.18459308e+00
-3.15598100e-01 3.85724127e-01 -1.24704041e-01 -1.19413030e+00
1.20963717e+00 2.76248395e-01 3.79680902e-01 -1.00412512e+00
6.21182501e-01 1.09268576e-01 6.67535216e-02 -5.99166989e-01
6.06034935e-01 1.03120007e-01 3.01386677e-02 6.45124316e-01
-2.68591553e-01 4.93593544e-01 6.70655593e-02 -5.00934571e-02
4.81087983e-01 -4.30061966e-01 -9.53307077e-02 -3.83732356e-02
1.34041592e-01 -1.52753061e-02 4.18788761e-01 5.51199913e-01
3.04324836e-01 2.03936324e-01 1.52838886e-01 -7.71123827e-01
-1.07039511e+00 -3.24038684e-01 -1.62418708e-01 1.43341756e+00
3.42005908e-01 1.71682518e-02 -5.11265159e-01 -2.46911392e-01
-1.74638778e-01 3.90553147e-01 -6.10930145e-01 -2.39117652e-01
9.03926138e-03 -7.67315328e-01 1.89622521e-01 3.66711378e-01
6.42418325e-01 -1.28018987e+00 -1.47140905e-01 4.92044955e-01
-6.93883793e-03 -6.20520413e-01 -3.59070301e-01 4.68589133e-03
-8.15804482e-01 -6.11710966e-01 -5.80564439e-01 -6.17685437e-01
3.91791344e-01 4.35139567e-01 6.95826948e-01 3.05568367e-01
4.93620545e-01 -3.66354257e-01 -8.66630852e-01 -5.01510322e-01
-2.26708636e-01 5.65767139e-02 1.81388333e-01 3.15798521e-01
3.16633940e-01 -3.08860153e-01 -6.71516836e-01 3.12574506e-01
-9.93717551e-01 1.49082407e-01 5.98635316e-01 7.24158883e-01
6.14755571e-01 6.57444179e-01 6.89661622e-01 -9.91971850e-01
1.28340983e+00 -1.06247771e+00 -5.00675857e-01 2.04470679e-01
-7.66519606e-01 -1.76162928e-01 9.98542964e-01 -7.46630192e-01
-9.73490536e-01 -7.26769924e-01 -8.93378183e-02 -1.96640357e-01
1.05128668e-01 1.35774660e+00 -7.95086026e-02 1.52261034e-01
5.42982280e-01 3.06368798e-01 -2.32341677e-01 -4.47989225e-01
2.19983339e-01 1.04432809e+00 4.21265475e-02 -4.79333073e-01
-1.33014983e-03 1.61035687e-01 -6.67356431e-01 -1.99334115e-01
-1.16192305e+00 -2.35472262e-01 1.44596711e-01 -4.99998659e-01
5.66939056e-01 -6.49257600e-01 -2.17218786e-01 4.15927351e-01
-9.95478272e-01 3.33734439e-03 -9.58309323e-02 8.71159017e-01
1.95259959e-01 -7.94508979e-02 -7.82695234e-01 -5.18179178e-01
-4.52271134e-01 -1.10553551e+00 1.93443179e-01 5.57857990e-01
2.03052193e-01 -8.78395677e-01 -2.96514630e-01 3.47631752e-01
8.01890314e-01 -2.85685986e-01 9.35488582e-01 -1.04743111e+00
-3.69444788e-01 -4.70162690e-01 -4.95150626e-01 3.70873183e-01
3.95344198e-02 -1.07695475e-01 -6.59939051e-01 3.47766466e-02
1.51935279e-01 -1.67915374e-01 6.93702936e-01 6.14076793e-01
1.47024667e+00 -6.79465711e-01 -5.15282620e-04 5.99386454e-01
1.08291197e+00 7.20208645e-01 4.91139799e-01 6.23062015e-01
5.20601988e-01 5.77482343e-01 4.08216387e-01 5.97489953e-01
2.59665012e-01 1.99942648e-01 4.25388038e-01 -1.36532381e-01
1.09214418e-01 -1.29951596e-01 2.05058753e-01 1.32271695e+00
-4.16770816e-01 -2.43261442e-01 -5.97092271e-01 1.88645065e-01
-1.96989048e+00 -1.26356781e+00 4.46694046e-01 1.98047960e+00
8.21378529e-01 4.75352645e-01 -5.40925264e-02 -1.35196391e-02
9.66973543e-01 3.45709808e-02 -5.19753635e-01 -3.00558686e-01
-2.46764552e-02 -2.23405749e-01 4.22367573e-01 -1.28111690e-01
-8.95922184e-01 5.32117248e-01 5.84625101e+00 1.04164779e+00
-1.64656711e+00 1.29859805e-01 9.45269585e-01 -1.76966876e-01
-4.63410884e-01 -3.28133345e-01 -8.26473892e-01 1.01333129e+00
8.45334530e-01 -2.33641937e-01 5.45353293e-01 1.03137624e+00
3.08694124e-01 2.14877054e-01 -6.33193493e-01 8.11467707e-01
-2.89157927e-02 -2.06067109e+00 3.81321251e-01 7.10845292e-02
9.59623516e-01 2.11629584e-01 3.00206393e-01 3.91918063e-01
3.88581097e-01 -7.87950456e-01 8.54322672e-01 7.51913667e-01
4.13762033e-01 -8.37110758e-01 1.26334822e+00 4.84112412e-01
-1.01996624e+00 -3.54329675e-01 -7.98045754e-01 -3.58413458e-01
-6.63855076e-02 9.28702474e-01 -4.66237664e-01 4.56137359e-01
8.17718565e-01 5.22577703e-01 4.47786301e-02 1.18305886e+00
1.58906624e-01 8.46444011e-01 -2.01958880e-01 -4.78852242e-01
2.86843896e-01 -3.96736592e-01 -3.28469649e-02 9.40718532e-01
7.07161784e-01 9.61268097e-02 4.99905914e-01 3.41484398e-01
-5.09266496e-01 6.67673707e-01 -1.72800452e-01 -1.90439358e-01
6.64370656e-01 1.02018905e+00 -8.93683553e-01 -6.09819651e-01
-5.59665620e-01 1.18655957e-01 2.68666476e-01 2.41651475e-01
-6.47953451e-01 -3.64808172e-01 1.10255495e-01 2.97702760e-01
2.25706741e-01 1.20239019e-01 -4.22721356e-01 -9.42923248e-01
-3.36381555e-01 -8.93690526e-01 4.07984614e-01 -5.76852202e-01
-1.49563420e+00 8.47936988e-01 -1.92399740e-01 -1.31258404e+00
1.77440301e-01 -3.43869835e-01 -7.43146181e-01 8.36532950e-01
-1.31132567e+00 -6.56450868e-01 -2.75926560e-01 6.64344549e-01
5.16729355e-01 -8.50736320e-01 5.53537250e-01 6.66010320e-01
-4.32244658e-01 3.40322673e-01 6.85433865e-01 3.23002875e-01
4.85333771e-01 -4.89816606e-01 -2.11327150e-01 6.95277095e-01
1.14078507e-01 6.04754031e-01 5.62046289e-01 -5.95083117e-01
-1.06800759e+00 -1.14026392e+00 4.97902215e-01 4.31989819e-01
9.01440263e-01 1.74020261e-01 -8.23998928e-01 2.35755786e-01
1.42422365e-02 -3.90409604e-02 8.86096895e-01 4.68200654e-01
-2.33909249e-01 -5.07808149e-01 -9.30836082e-01 4.02824461e-01
3.46724600e-01 -2.33262852e-01 -5.94422638e-01 3.52754116e-01
8.14071536e-01 -5.25175571e-01 -7.80960917e-01 -4.97317314e-02
6.73614323e-01 -8.53267014e-01 5.84122360e-01 -6.48373127e-01
6.94104970e-01 -1.17847390e-01 -1.32726744e-01 -1.43351829e+00
-5.53183258e-01 2.25024521e-01 -9.40379053e-02 1.04940450e+00
3.54745567e-01 -5.23367047e-01 5.35888433e-01 3.38496417e-01
-3.48602474e-01 -9.96821821e-01 -5.58929205e-01 -1.03744723e-01
-4.88186896e-01 -3.34832907e-01 1.13418329e+00 1.18206322e+00
2.48764679e-02 1.10858142e-01 -5.21287858e-01 2.31170300e-02
1.38826385e-01 4.10071373e-01 3.59380394e-01 -1.32845044e+00
-3.19900841e-01 -7.14273095e-01 -2.08497211e-01 -1.12935603e+00
-2.96164125e-01 -7.93647349e-01 -1.43477365e-01 -1.86793876e+00
1.37153696e-02 -8.32533240e-01 -8.14350545e-01 3.34562182e-01
2.29943752e-01 8.26375484e-02 -2.72520721e-01 7.54932702e-01
-5.31164289e-01 2.38598585e-01 1.44086266e+00 -3.86339724e-01
-2.15758026e-01 4.12227631e-01 -1.07933021e+00 7.57291853e-01
7.92361617e-01 -3.74857575e-01 -5.43338180e-01 -7.47158229e-01
7.74137735e-01 1.49291784e-01 -2.54632533e-01 -8.71559501e-01
6.38473690e-01 -5.95901966e-01 6.56347275e-01 -4.12336260e-01
-1.17035411e-01 -7.28335977e-01 2.08252877e-01 1.81883611e-02
-6.89168811e-01 -3.63761298e-02 -1.20286882e-01 4.39829618e-01
-4.21128988e-01 -2.58342147e-01 4.52310950e-01 -2.44204149e-01
-8.00620735e-01 7.61787474e-01 -2.66749382e-01 -3.01423609e-01
6.34866714e-01 -1.62683964e-01 -3.46739858e-01 -5.64601123e-01
-5.86372435e-01 7.78489485e-02 -8.37372057e-03 3.21673781e-01
7.37675011e-01 -1.36153841e+00 -6.71469092e-01 1.17986903e-01
-6.44549131e-02 7.33102188e-02 3.98005724e-01 4.70685780e-01
-4.70698684e-01 -6.33009002e-02 -3.77068281e-01 -1.27450198e-01
-7.95651555e-01 4.35368687e-01 1.29087999e-01 -1.27941906e-01
-6.61629379e-01 6.74479961e-01 -6.94988221e-02 4.14563231e-02
4.42312300e-01 -2.66470373e-01 -1.00930667e+00 3.38486582e-01
9.11760628e-01 4.31094952e-02 2.34845504e-01 -3.17251086e-01
1.60736308e-01 1.49840936e-01 -4.79562461e-01 2.34241948e-01
1.55854285e+00 -1.17920503e-01 -1.84818491e-01 4.53992635e-01
6.60682797e-01 6.21783435e-02 -1.03976238e+00 -6.13282859e-01
-3.22430462e-01 -2.32580140e-01 5.81592202e-01 -7.91274011e-01
-1.44908893e+00 5.37472248e-01 4.80687141e-01 6.94967866e-01
1.00176871e+00 -5.52620590e-02 9.93312061e-01 3.30955327e-01
1.78673863e-01 -1.52231252e+00 -2.19927743e-01 6.27297103e-01
5.27975917e-01 -1.06944299e+00 9.52545032e-02 3.24856639e-01
-4.29441303e-01 1.31874108e+00 2.97278851e-01 -2.92116374e-01
9.72651780e-01 4.58404690e-01 8.68886486e-02 -7.91528970e-02
-5.51651835e-01 1.15645260e-01 3.94650102e-01 1.33887842e-01
4.94218647e-01 2.37496831e-02 -3.40065360e-01 1.16764867e+00
-2.65872270e-01 2.03902364e-01 1.92626372e-01 4.44103271e-01
-1.05213916e+00 -7.70379305e-01 -4.08782661e-01 9.96126831e-01
-5.54592490e-01 -2.46151671e-01 2.06743538e-01 6.61587790e-02
5.22228301e-01 9.01002169e-01 3.34239900e-01 -8.12704802e-01
3.21777537e-02 -3.68196815e-01 -8.69406909e-02 -7.45142102e-01
-5.17874599e-01 1.73703700e-01 -2.17048172e-03 7.35763134e-03
-4.13413942e-01 -2.44451851e-01 -1.22339153e+00 -5.51279247e-01
-4.38231170e-01 5.52733183e-01 9.17687774e-01 1.35520315e+00
4.35875118e-01 6.53399229e-01 8.26359212e-01 -6.29795730e-01
-4.73286837e-01 -9.87954140e-01 -5.83179533e-01 3.72833252e-01
-8.80824476e-02 -7.35843837e-01 -2.52143234e-01 -1.90713286e-01] | [10.139934539794922, 5.623075485229492] |
5694d657-aed4-479f-94ba-87adb7ba437a | clexis2-a-new-corpus-for-complex-word | null | null | https://aclanthology.org/2021.ranlp-main.121 | https://aclanthology.org/2021.ranlp-main.121.pdf | CLexIS2: A New Corpus for Complex Word Identification Research in Computing Studies | Reading is a complex process not only because of the words or sections that are difficult for the reader to understand. Complex word identification (CWI) is the task of detecting in the content of documents the words that are difficult or complex to understand by the people of a certain group. Annotated corpora for English learners are widely available, while they are less common for the Spanish language. In this article, we present CLexIS^2, a new corpus in Spanish to contribute to the advancement of research in the area of Lexical Simplification, specifically in the identification and prediction of complex words in computing studies. Several metrics used to evaluate the complexity of texts in Spanish were applied, such as LC, LDI, ILFW, SSR, SCI, ASL, CS. Furthermore, as a baseline of the primer, two experiments have been performed to predict the complexity of words: one using a supervised learning approach and the other using an unsupervised solution based on the frequency of words on a general corpus. | ['Arturo Montejo-Ráez', 'Jenny A. Ortiz Zambrano'] | null | null | https://aclanthology.org/2021.ranlp-1.121 | https://aclanthology.org/2021.ranlp-1.121.pdf | ranlp-2021-9 | ['lexical-simplification', 'complex-word-identification'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.60575265e-01 2.09015206e-01 1.58453077e-01 -2.10728616e-01
-4.34201866e-01 -5.35541952e-01 6.16764247e-01 1.00049305e+00
-6.75591350e-01 6.57632589e-01 4.94712383e-01 -4.69501585e-01
-3.40377808e-01 -5.98511815e-01 -3.36935997e-01 -3.87222260e-01
5.28990746e-01 5.93022108e-01 3.34726982e-02 -4.35549736e-01
7.66668200e-01 5.21543562e-01 -1.90784669e+00 1.60550356e-01
1.06807756e+00 3.24656069e-01 5.58193684e-01 4.23424125e-01
-6.00384414e-01 6.83215857e-01 -9.12843704e-01 -3.40788275e-01
-3.96888465e-01 -4.25806195e-01 -9.80836630e-01 -4.70666997e-02
3.93231720e-01 4.09678966e-01 2.34904110e-01 9.66839373e-01
4.17851239e-01 1.22195296e-01 9.42162752e-01 -4.70154643e-01
-9.89346392e-03 9.03376818e-01 -3.09326332e-02 3.40708673e-01
8.20887268e-01 -4.52232897e-01 7.51762033e-01 -7.67502367e-01
8.45307946e-01 1.53608739e+00 4.21487898e-01 1.96256354e-01
-7.56891310e-01 -4.39029008e-01 4.87044454e-02 5.82616210e-01
-1.29378140e+00 -3.00882548e-01 3.20785344e-01 -6.44171417e-01
1.01614392e+00 3.24445844e-01 6.80802822e-01 7.73071766e-01
2.83473343e-01 4.37822014e-01 1.10319912e+00 -1.34229600e+00
1.65705740e-01 2.89690673e-01 8.24903786e-01 3.84937257e-01
5.62647283e-01 -3.24160308e-01 -5.06240726e-01 3.90790343e-01
-1.55903161e-01 -3.74574512e-01 -2.35168040e-01 5.37669938e-03
-8.21191549e-01 1.09458899e+00 -5.48675179e-01 1.01988399e+00
-1.25323579e-01 -5.67720890e-01 6.15696192e-01 3.84729654e-01
3.32766980e-01 7.28586912e-01 -4.63332444e-01 -3.97261202e-01
-8.63657713e-01 3.25040430e-01 1.13164353e+00 9.07292664e-01
3.52057308e-01 -3.19598973e-01 1.53192440e-02 6.34616017e-01
2.42568627e-01 5.23086488e-01 9.28995252e-01 -5.52182555e-01
4.52246904e-01 8.88786435e-01 -3.66144836e-01 -8.02043378e-01
-4.89557683e-01 -1.60892621e-01 -5.00980854e-01 6.64392412e-02
5.54974496e-01 2.28246838e-01 -5.78587890e-01 1.29612660e+00
5.61077520e-02 -7.38767207e-01 3.26960906e-02 -1.95283219e-02
1.02656770e+00 7.42409706e-01 1.62276804e-01 -6.49614215e-01
1.34594762e+00 -8.22293699e-01 -1.08877981e+00 5.29113524e-02
8.33958626e-01 -1.18738127e+00 1.25453281e+00 7.39147246e-01
-1.11955488e+00 -5.33978343e-01 -1.09076035e+00 -2.00460181e-01
-1.08030558e+00 7.11302757e-02 1.99100628e-01 7.92363465e-01
-5.58123469e-01 5.02666235e-01 -4.37300324e-01 -4.30695534e-01
-3.67018655e-02 -5.18355407e-02 -2.18209043e-01 -8.40033591e-02
-9.19184268e-01 1.44110692e+00 6.12701237e-01 -4.72107470e-01
-2.48338982e-01 -5.47964275e-01 -8.96029472e-01 2.61845797e-01
3.45013648e-01 6.31694775e-03 1.06563401e+00 -1.04722202e+00
-1.32566953e+00 1.09326315e+00 -2.66682893e-01 -1.20947987e-01
4.48338419e-01 -2.92201877e-01 -3.54070067e-01 7.79309571e-02
4.74597871e-01 2.56041531e-02 7.12609887e-01 -8.75859797e-01
-7.01910913e-01 -5.41538179e-01 -1.81928560e-01 1.24953508e-01
-1.88889161e-01 3.41415644e-01 -2.92492479e-01 -7.69015074e-01
-6.18795827e-02 -7.19796658e-01 2.85499096e-01 -5.55714071e-01
-1.66103169e-01 -8.88548434e-01 3.03285658e-01 -1.11421883e+00
1.77513528e+00 -2.05725884e+00 3.82364541e-01 4.27975982e-01
1.22834332e-01 5.66302180e-01 6.98700622e-02 5.13751924e-01
-2.32191205e-01 3.85442764e-01 -1.07954800e-01 -3.14952247e-02
4.63401414e-02 1.76454559e-01 1.58727199e-01 6.85823783e-02
-2.17015103e-01 2.73004055e-01 -7.58571208e-01 -6.44056380e-01
1.19828135e-01 1.25775546e-01 -3.42492849e-01 1.01326648e-02
-2.41617318e-02 -1.81775521e-02 -9.27424058e-02 3.02583575e-01
2.49671012e-01 3.09325755e-01 3.36780250e-01 1.50155649e-01
-7.07851768e-01 4.20250058e-01 -1.20705247e+00 1.12353337e+00
-6.85080647e-01 8.50541592e-01 -2.78537393e-01 -1.06109238e+00
7.10287094e-01 2.78363168e-01 -1.64778531e-01 -8.92612040e-01
4.31814075e-01 3.92585576e-01 1.79628417e-01 -7.18893468e-01
5.61680317e-01 2.71196753e-01 -1.65033549e-01 3.23044449e-01
2.54484057e-01 -4.35531765e-01 1.24519050e+00 1.58882603e-01
6.79368675e-01 -1.28649592e-01 8.13765943e-01 -6.80550516e-01
9.90336180e-01 -7.85056278e-02 1.57155737e-01 5.28739333e-01
2.97505468e-01 1.95640713e-01 5.24758637e-01 -1.65531233e-01
-9.83989775e-01 -6.83491826e-01 -3.11884195e-01 1.20618904e+00
-2.06407756e-01 -7.81275153e-01 -1.05746233e+00 -3.13811719e-01
-3.13435704e-01 1.18929231e+00 -3.19667071e-01 1.34878997e-02
-7.88970411e-01 -2.89375007e-01 1.50890201e-01 -5.17373495e-02
1.53386686e-02 -1.25951970e+00 -6.67758942e-01 1.98138058e-01
-3.48950952e-01 -9.22492266e-01 -1.39532506e-01 2.95252800e-01
-4.77767587e-01 -1.24049366e+00 -3.94867986e-01 -1.05465412e+00
5.20028055e-01 4.06922735e-02 1.31545544e+00 4.38812345e-01
-4.39141244e-01 3.06329012e-01 -8.14957440e-01 -1.09360099e+00
-9.61472988e-01 4.41615582e-01 -1.70008257e-01 -6.48060858e-01
5.71747303e-01 4.11478318e-02 3.19947004e-01 -1.29791573e-01
-1.05246937e+00 -2.36949801e-01 3.61532003e-01 5.90492129e-01
3.32163662e-01 2.32286423e-01 1.72997102e-01 -1.01444030e+00
9.59687531e-01 -2.34138966e-01 -3.52976203e-01 5.66693842e-01
-7.55724549e-01 7.85249695e-02 6.06259525e-01 -3.60552460e-01
-9.81264830e-01 -3.08609337e-01 -4.23812032e-01 6.32368803e-01
-3.73670489e-01 6.48214221e-01 -3.96626472e-01 -5.03802076e-02
5.59233308e-01 -3.53216343e-02 4.33087088e-02 -6.30517244e-01
-8.47868435e-03 6.99474454e-01 1.98980898e-01 -3.74858618e-01
5.44995010e-01 -3.26316953e-01 -1.06700487e-01 -1.46503294e+00
-9.04704511e-01 -6.74133122e-01 -8.96131873e-01 -2.95524925e-01
7.48892844e-01 -6.20496154e-01 -4.13870603e-01 4.56716269e-01
-1.15518129e+00 -6.27074614e-02 -3.57795358e-01 4.88552064e-01
-1.98357821e-01 7.59920359e-01 -2.20324457e-01 -3.35544229e-01
-4.06764805e-01 -1.06291628e+00 5.91789186e-01 3.81913751e-01
-8.12984705e-01 -1.02209437e+00 1.53598404e-02 2.38283843e-01
2.20086873e-01 -3.01987268e-02 1.67214119e+00 -8.74627352e-01
1.84294537e-01 5.31433798e-05 1.84722349e-01 5.57283342e-01
-2.53797978e-01 3.90686393e-01 -2.01279908e-01 1.84946992e-02
2.11375847e-01 -1.93857074e-01 5.08027136e-01 5.96644394e-02
8.13669384e-01 -1.13321714e-01 -1.04278028e-01 1.22319974e-01
1.31474936e+00 3.27997446e-01 5.57383537e-01 7.28424966e-01
3.22182894e-01 6.89590514e-01 7.51261115e-01 2.74705917e-01
2.71848738e-01 4.52622175e-01 -8.50346610e-02 2.24180862e-01
-2.34887853e-01 3.48396339e-02 3.67788643e-01 1.59699285e+00
-6.81890547e-02 -2.27471665e-01 -1.14211607e+00 4.45445538e-01
-1.11942589e+00 -7.33380973e-01 -5.14500678e-01 1.91784227e+00
7.98328340e-01 3.35915983e-01 -7.27741420e-02 8.73254776e-01
5.14680862e-01 -1.59328401e-01 4.23949808e-01 -1.01559818e+00
-2.50375867e-01 8.86343181e-01 1.73554599e-01 7.29680657e-01
-6.93913162e-01 9.91327167e-01 5.83574009e+00 1.08242440e+00
-9.45120215e-01 -9.15588513e-02 1.30891144e-01 3.04813325e-01
-2.57666767e-01 -1.89072728e-01 -1.04242742e+00 4.36433643e-01
1.07155943e+00 -2.74105161e-01 3.35706294e-01 5.44135094e-01
2.47222751e-01 -6.78600073e-01 -8.63857329e-01 8.09771955e-01
6.13467038e-01 -8.17853451e-01 3.40422213e-01 -3.63973022e-01
4.86373991e-01 -3.65611792e-01 -2.97788739e-01 3.57506365e-01
-2.50191450e-01 -9.54197407e-01 1.07149100e+00 7.16507018e-01
4.60409760e-01 -8.45721960e-01 9.27449286e-01 5.22057235e-01
-6.75649226e-01 1.03001676e-01 -3.57644469e-01 -2.21006736e-01
-1.97073743e-01 3.04205596e-01 -4.24490899e-01 3.12234759e-01
6.43906176e-01 3.52502286e-01 -1.15046763e+00 9.18303847e-01
-4.57230121e-01 6.37022913e-01 -1.68532044e-01 -7.55931020e-01
1.71803162e-02 -3.31594050e-01 5.77019572e-01 1.48151672e+00
4.40260231e-01 3.12613659e-02 -1.66069910e-01 3.30098778e-01
3.67828995e-01 9.18696105e-01 -2.88978994e-01 -1.89028159e-01
5.31850040e-01 9.86704230e-01 -7.25565374e-01 -3.72476250e-01
-3.08151811e-01 6.41082823e-01 1.81541637e-01 -6.11636154e-02
-3.43421549e-01 -8.51306796e-01 1.56668663e-01 5.61640449e-02
1.37472183e-01 -4.00762230e-01 -4.36198175e-01 -9.30309355e-01
-9.02450532e-02 -1.28704739e+00 2.11334497e-01 -4.46027905e-01
-9.52351928e-01 5.33820808e-01 3.01543266e-01 -6.87058568e-01
-3.24086607e-01 -1.03042781e+00 -5.77902079e-01 7.51418591e-01
-1.28895020e+00 -5.53530693e-01 -3.08146119e-01 5.04487514e-01
7.39742577e-01 -4.92858261e-01 1.09802222e+00 2.97467500e-01
-5.19689023e-01 4.78200436e-01 3.78128082e-01 -1.93029284e-01
4.68607157e-01 -1.25261652e+00 -8.44606105e-03 6.32012844e-01
4.44889218e-02 2.61014789e-01 7.92376876e-01 -4.33688402e-01
-7.63259113e-01 -4.42594826e-01 1.87160325e+00 -1.02509245e-01
5.55444598e-01 -1.88507602e-01 -7.29599714e-01 2.30821535e-01
2.67623037e-01 -1.07859921e+00 7.99874246e-01 -9.67661962e-02
3.45393084e-02 1.29797310e-01 -8.16590130e-01 6.17288649e-01
6.55117452e-01 -3.44038993e-01 -1.15189898e+00 4.26441163e-01
4.32520270e-01 -7.52663985e-02 -9.03847396e-01 -1.09788943e-02
3.43046248e-01 -7.85779178e-01 6.36925101e-01 -3.73702556e-01
4.36849624e-01 1.32759228e-01 1.55132890e-01 -1.35202765e+00
-2.16819808e-01 -3.74644488e-01 4.35175896e-01 1.26550817e+00
5.29413939e-01 -2.88654476e-01 3.75193834e-01 7.13616014e-02
-2.09904343e-01 -2.51426727e-01 -8.78009260e-01 -5.56799054e-01
3.75338286e-01 -3.15989554e-01 3.82814139e-01 8.08660448e-01
1.04984552e-01 6.11305356e-01 4.53315824e-01 -4.21460181e-01
2.77642667e-01 -2.55274624e-01 5.23316681e-01 -1.70817125e+00
1.74163476e-01 -7.55459249e-01 -3.07461470e-01 -3.62787932e-01
3.86464417e-01 -8.40387523e-01 -6.97963014e-02 -1.32202375e+00
-1.65725559e-01 6.05986938e-02 4.98107672e-01 7.69002587e-02
-2.11278066e-01 -2.11035028e-01 5.41787744e-01 7.89618585e-03
-3.12081069e-01 2.10523218e-01 9.76962268e-01 -2.17273328e-02
-4.87457365e-02 -1.61877973e-03 -3.30928653e-01 1.10073364e+00
7.55198240e-01 -5.27583778e-01 -4.06931043e-02 -2.82676518e-01
4.03022736e-01 -3.89648527e-01 -4.39620733e-01 -1.07551062e+00
6.88606780e-03 8.46254528e-02 1.35549203e-01 -6.74920321e-01
-2.22985297e-01 -7.04437852e-01 -9.54919010e-02 5.59772372e-01
-1.67913228e-01 6.19299829e-01 7.11568594e-02 -1.84367388e-01
-4.04168695e-01 -1.29954398e+00 7.46485114e-01 -3.04206610e-01
-6.61820292e-01 -4.45637077e-01 -1.01957095e+00 4.42115217e-01
9.75093186e-01 -6.70853332e-02 -7.36273378e-02 -2.71723539e-01
-6.71232164e-01 -2.85851628e-01 2.14550540e-01 3.35480928e-01
3.85863364e-01 -6.41416430e-01 -7.08219469e-01 1.02932967e-01
8.54690969e-02 -1.88213617e-01 -2.20913455e-01 5.70504248e-01
-1.27946854e+00 5.62606752e-01 -3.65010262e-01 2.82731708e-02
-1.61717045e+00 4.80628073e-01 -3.33831571e-02 -6.16706371e-01
-2.02947706e-01 3.76309425e-01 -3.68305057e-01 -3.43022376e-01
6.23188257e-01 -3.69526327e-01 -1.08161807e+00 7.38837361e-01
5.93747139e-01 7.82203197e-01 3.53198022e-01 -9.16209340e-01
7.52058625e-02 6.37474239e-01 2.41112057e-02 1.14257969e-02
1.21469462e+00 -2.79823393e-01 -6.50077164e-01 6.33798242e-01
1.20493424e+00 3.48324120e-01 1.06362693e-01 -8.77180994e-02
6.77428722e-01 1.23335011e-02 -6.08646497e-02 -7.40077674e-01
-2.55296707e-01 9.05851603e-01 5.83422542e-01 4.04865295e-01
9.12191570e-01 -3.22585166e-01 3.78077596e-01 5.06745815e-01
-4.59647588e-02 -1.65182245e+00 -1.04451708e-01 1.16649091e+00
9.07297850e-01 -9.09176767e-01 1.18806347e-01 -7.03584313e-01
-3.14224362e-01 1.62154412e+00 1.37952372e-01 2.88418978e-01
8.28094304e-01 -6.02690540e-02 -2.76545179e-03 -1.95411801e-01
-2.13101253e-01 -3.76714289e-01 3.02856445e-01 3.68280083e-01
7.63128996e-01 1.81484818e-01 -1.57439470e+00 6.65680408e-01
-6.86977208e-01 -2.93234795e-01 7.77166486e-01 8.39824677e-01
-6.26058578e-01 -1.25562322e+00 -5.27263045e-01 3.80908459e-01
-7.38510072e-01 -3.39898348e-01 -5.53021371e-01 1.18704045e+00
4.20213640e-01 1.11929607e+00 4.27599289e-02 1.58093899e-01
4.64940071e-01 3.92055333e-01 5.68712294e-01 -7.95785546e-01
-7.81846404e-01 -2.18846262e-01 2.55521446e-01 -2.92977281e-02
-3.52243811e-01 -9.19958293e-01 -1.04034209e+00 -4.87974018e-01
-2.92176008e-01 3.38736981e-01 1.11326361e+00 1.28411698e+00
-4.26421136e-01 4.86436993e-01 1.74334720e-01 -6.01209044e-01
-4.55088705e-01 -1.21707356e+00 -5.32657862e-01 5.33986449e-01
-2.59678245e-01 -3.56575370e-01 -4.61905539e-01 8.38110298e-02] | [10.738299369812012, 10.345609664916992] |
0adf03f2-159c-4484-a5df-c21a48e6b12a | learning-adaptive-receptive-fields-for-deep | null | null | http://openaccess.thecvf.com/content_cvpr_2017/html/Wei_Learning_Adaptive_Receptive_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Wei_Learning_Adaptive_Receptive_CVPR_2017_paper.pdf | Learning Adaptive Receptive Fields for Deep Image Parsing Network | In this paper, we introduce a novel approach to regulate receptive field in deep image parsing network automatically. Unlike previous works which have stressed much importance on obtaining better receptive fields using manually selected dilated convolutional kernels, our approach uses two affine transformation layers in the network's backbone and operates on feature maps. Feature maps will be inflated/shrinked by the new layer and therefore receptive fields in following layers are changed accordingly. By end-to-end training, the whole framework is data-driven without laborious manual intervention. The proposed method is generic across dataset and different tasks. We conduct extensive experiments on both general parsing task and face parsing task as concrete examples to demonstrate the method's superior regulation ability over manual designs.
| ['Zhen Wei', 'Jinqiao Wang', 'Si Liu', 'Yao Sun', 'Hanjiang Lai'] | 2017-07-01 | null | null | null | cvpr-2017-7 | ['face-parsing'] | ['computer-vision'] | [ 4.63952482e-01 5.61291397e-01 5.69486292e-03 -8.56853306e-01
-1.39893666e-01 -7.00480342e-01 4.14854556e-01 -4.10812050e-01
-7.26582229e-01 5.05558610e-01 6.01637848e-02 -1.21666059e-01
-3.07927523e-02 -8.66584599e-01 -8.20879579e-01 -6.71538234e-01
1.23482376e-01 -1.69075966e-01 2.81378120e-01 -3.46949786e-01
3.70694190e-01 6.90063000e-01 -1.04479992e+00 1.48924425e-01
5.18888414e-01 7.97045171e-01 9.37905386e-02 7.34047234e-01
-2.03496307e-01 7.28640378e-01 -5.22189260e-01 -4.63363022e-01
5.90006649e-01 -2.17496529e-01 -9.47671115e-01 2.46198058e-01
4.83038396e-01 -2.16891751e-01 -1.31468028e-01 1.11610198e+00
5.57266593e-01 2.89195534e-02 5.20372510e-01 -5.63316107e-01
-1.11593747e+00 5.87819874e-01 -7.19908595e-01 2.80226260e-01
-1.17888987e-01 -2.52881870e-02 6.25503659e-01 -6.04344964e-01
4.95914191e-01 1.36809361e+00 5.97236872e-01 1.14294839e+00
-1.42271638e+00 -7.43532360e-01 4.16366607e-01 -4.37767684e-01
-9.53626871e-01 -4.23796356e-01 1.05630529e+00 -3.37153047e-01
8.44685674e-01 -1.97108667e-02 1.53833881e-01 9.39082503e-01
2.03520194e-01 3.26587141e-01 1.09164321e+00 -4.91939962e-01
9.21768919e-02 1.90368816e-01 1.67501286e-01 7.31642783e-01
-3.02829780e-02 -2.57127825e-02 -2.22059116e-01 9.38710123e-02
1.28816068e+00 -2.39066660e-01 -8.40744749e-02 -4.60889012e-01
-8.33132625e-01 9.83431041e-01 6.22473598e-01 3.31162095e-01
-2.59484887e-01 2.38214076e-01 5.35913289e-01 2.98044294e-01
3.54605764e-01 2.50320256e-01 -7.79720366e-01 3.16976368e-01
-5.70957005e-01 2.27787778e-01 4.33241189e-01 9.32222664e-01
6.97894990e-01 -7.99529552e-02 -3.03378761e-01 1.22914946e+00
1.18599854e-01 -1.85938727e-04 6.26429737e-01 -1.00071633e+00
4.20350134e-01 5.80758393e-01 -1.97269082e-01 -8.74490082e-01
-5.74044347e-01 -3.07102561e-01 -1.06502175e+00 3.28075469e-01
4.90759134e-01 -3.47677439e-01 -1.36384130e+00 1.96749437e+00
3.20360929e-01 -1.41730890e-01 1.16416968e-01 7.44155169e-01
9.29623127e-01 4.02112156e-01 6.13339126e-01 -1.43719288e-02
1.45948517e+00 -9.21116233e-01 -6.06743395e-01 -1.79049999e-01
4.86611396e-01 -6.41275048e-01 1.21015716e+00 2.21612900e-01
-1.10844207e+00 -9.30808783e-01 -9.00754452e-01 -2.70507276e-01
-5.07434189e-01 4.46885049e-01 1.04825115e+00 6.06594682e-01
-1.35021973e+00 6.29047334e-01 -5.84425867e-01 -3.77452701e-01
9.36811566e-01 9.42858219e-01 -6.35903060e-01 4.07891184e-01
-1.01122868e+00 5.22306025e-01 3.91231894e-01 2.46107593e-01
-3.56398433e-01 -6.45769417e-01 -7.92317331e-01 1.64861511e-02
3.24744210e-02 -4.98124123e-01 1.26934838e+00 -1.19663024e+00
-1.97481775e+00 9.44161177e-01 2.56437473e-02 -2.37525031e-01
2.75126338e-01 -8.38329867e-02 6.72417730e-02 5.42368889e-02
-1.18242554e-01 1.34301412e+00 1.10018146e+00 -9.94374335e-01
-4.94566053e-01 -3.43660653e-01 3.81288111e-01 -7.06114024e-02
-3.93862516e-01 6.69180974e-02 -2.94781297e-01 -7.24590063e-01
3.45521688e-01 -7.48573065e-01 -3.41750950e-01 -1.06745109e-01
-1.50653318e-01 -3.99724215e-01 7.03499258e-01 -4.79012221e-01
1.01260161e+00 -2.05997300e+00 -1.69379413e-02 4.28459421e-02
2.30839044e-01 4.25190181e-01 -2.69024253e-01 -1.81036308e-01
-4.22738194e-01 2.86420196e-01 -2.84377635e-01 1.10427558e-01
-2.69804120e-01 2.57290214e-01 -2.25379243e-01 3.72751027e-01
7.16378927e-01 8.24620783e-01 -6.50426567e-01 -4.65056956e-01
1.95958599e-01 7.73789704e-01 -8.29101980e-01 1.27223730e-02
1.61155269e-01 6.56987250e-01 -6.74171925e-01 6.47366881e-01
1.00664365e+00 1.94654837e-01 -9.16243121e-02 -3.04426968e-01
-4.57250923e-02 -1.09384000e-01 -7.94080138e-01 1.74372613e+00
-5.66918433e-01 4.36127007e-01 6.86847419e-02 -1.08475125e+00
1.18348086e+00 1.13798954e-01 2.17942223e-01 -6.04642034e-01
3.40556592e-01 -2.96811312e-01 1.74065918e-01 -5.56019723e-01
1.66961506e-01 -1.57627210e-01 -4.38078977e-02 4.87215221e-02
3.16949934e-01 1.30867720e-01 -5.77232689e-02 -1.54261991e-01
1.10129106e+00 4.91306424e-01 2.29339913e-01 -6.03397608e-01
9.99283135e-01 -1.51769862e-01 6.26758754e-01 5.43348968e-01
-5.92574000e-01 4.73937422e-01 8.12084258e-01 -7.68554986e-01
-1.12572300e+00 -8.39933932e-01 -5.31105220e-01 1.61901474e+00
-1.30322620e-01 -1.36578735e-02 -1.37217724e+00 -1.05459070e+00
-1.87472016e-01 1.05530858e-01 -1.14710927e+00 8.19573849e-02
-8.88724446e-01 -9.16113317e-01 6.11383498e-01 8.15143943e-01
1.03254998e+00 -1.44445968e+00 -6.34650350e-01 1.26313046e-01
3.48075747e-01 -1.05505753e+00 -3.48440051e-01 4.06091511e-01
-1.00069249e+00 -8.79902065e-01 -6.17405593e-01 -1.29366946e+00
1.23763847e+00 -6.80005702e-04 8.95670652e-01 -5.39167188e-02
-3.66304338e-01 8.37008841e-03 -1.61231413e-01 -3.41466904e-01
-1.61701784e-01 3.88919532e-01 -4.23942029e-01 -1.16753645e-01
5.58056235e-01 -5.29769957e-01 -6.56312525e-01 2.30073407e-01
-8.76198351e-01 -1.12628803e-01 6.97949708e-01 8.51503372e-01
4.88946378e-01 -2.21659407e-01 7.16765046e-01 -1.33909309e+00
8.76180887e-01 2.12315340e-02 -7.81651855e-01 7.73554668e-02
-1.72290221e-01 3.59945208e-01 8.50824118e-01 -4.92426306e-01
-1.42586529e+00 5.40268183e-01 -2.80567199e-01 -2.55218931e-02
-4.24071133e-01 -2.87867337e-01 -1.00520030e-01 -4.28537041e-01
7.82197833e-01 -1.36447027e-01 -2.71933466e-01 -3.98423642e-01
3.18576753e-01 5.92586040e-01 6.27199829e-01 -5.95364749e-01
8.12229753e-01 3.62973779e-01 1.25316992e-01 -5.63022435e-01
-7.52273500e-01 1.71180628e-02 -1.21828580e+00 -4.73604836e-02
1.12020242e+00 -6.27548635e-01 -8.90516281e-01 3.31653446e-01
-1.15051639e+00 -1.96670160e-01 2.51946114e-02 2.28993192e-01
-4.00155365e-01 -5.56047782e-02 -5.37405849e-01 -4.45506722e-01
-5.21148801e-01 -1.16861379e+00 1.21454191e+00 6.32375479e-01
2.15933789e-02 -1.02164078e+00 7.00139701e-02 1.73746809e-01
5.29841721e-01 2.10266590e-01 9.73991632e-01 -4.17562991e-01
-1.11374371e-01 -2.65226029e-02 -3.96564901e-01 5.72172582e-01
2.87292182e-01 2.65498227e-03 -1.40037644e+00 -8.85717496e-02
-1.29268263e-02 -2.91603714e-01 1.08927333e+00 5.22292435e-01
1.55866206e+00 -3.49419527e-02 -1.39194101e-01 9.90184188e-01
1.38285100e+00 2.66816974e-01 7.32556164e-01 3.68544698e-01
6.69891119e-01 9.13539529e-01 3.53238583e-01 -3.16044837e-02
1.92293469e-02 3.03956896e-01 2.89144009e-01 -3.93601149e-01
4.55892049e-02 -1.37002366e-02 1.64815620e-01 2.68991262e-01
-3.73424113e-01 1.29072234e-01 -4.47665840e-01 2.95709282e-01
-1.56361449e+00 -7.88456619e-01 1.81488857e-01 1.85095823e+00
9.31497872e-01 2.18786255e-01 -4.42341901e-02 -1.64246947e-01
8.58547628e-01 1.08680725e-01 -6.02420986e-01 -1.07435024e+00
1.99273378e-01 7.34410167e-01 7.93659270e-01 5.37242234e-01
-1.43527699e+00 1.52853942e+00 6.83032322e+00 4.36779648e-01
-1.25844538e+00 -1.44013343e-02 9.23813939e-01 1.65379927e-01
1.56423852e-01 -1.27636611e-01 -8.96085858e-01 6.59228787e-02
5.37383080e-01 4.66535747e-01 4.64988708e-01 9.95749354e-01
1.13581412e-01 8.43272954e-02 -8.43149900e-01 8.38843524e-01
-3.22297037e-01 -1.04912710e+00 -1.31624252e-01 -1.22450985e-01
4.08806801e-01 -3.55583310e-01 1.51142493e-01 4.72748637e-01
2.81145573e-01 -1.14752054e+00 1.91260457e-01 2.80798912e-01
6.42045081e-01 -9.62366164e-01 5.38795531e-01 -5.49792536e-02
-1.01274717e+00 -7.90241286e-02 -9.22131479e-01 -7.17000291e-02
-4.74394798e-01 9.88662243e-02 -7.10578322e-01 -1.47014543e-01
7.59996414e-01 2.44232669e-01 -7.86760330e-01 2.54584134e-01
-2.24756181e-01 6.04406357e-01 -3.70962982e-04 8.26528892e-02
4.94498581e-01 -2.11933523e-01 8.25902389e-04 1.36887085e+00
-2.24605381e-01 1.19836129e-01 -3.34273756e-01 8.07658792e-01
-2.77847826e-01 3.68339390e-01 -7.14189112e-01 1.85227856e-01
1.26728266e-01 1.75503266e+00 -1.04150236e+00 -1.54587016e-01
-5.25569260e-01 9.58567500e-01 6.23877943e-01 3.67722094e-01
-1.06589925e+00 -5.85664928e-01 6.21366203e-01 -3.77137288e-02
5.69782913e-01 1.16801873e-01 -3.70594651e-01 -7.79645503e-01
-1.37264520e-01 -4.28975314e-01 8.19023848e-02 -4.39106047e-01
-9.78773952e-01 8.10458958e-01 1.65414177e-02 -6.55572236e-01
-8.97240862e-02 -1.03684807e+00 -7.44137168e-01 9.05193925e-01
-1.58204055e+00 -1.12124336e+00 -3.43397796e-01 6.00777388e-01
6.40467644e-01 -2.32089654e-01 7.15590000e-01 2.63857007e-01
-6.92541540e-01 8.37970078e-01 -4.06404942e-01 4.78437811e-01
6.02937698e-01 -1.24153519e+00 5.21249533e-01 6.71943128e-01
-2.85846353e-01 8.68581712e-01 3.63904297e-01 -3.99003953e-01
-8.81068289e-01 -1.22250175e+00 5.69474757e-01 -2.17422694e-01
4.51739281e-01 -7.08575308e-01 -9.76330876e-01 6.85978651e-01
4.32401896e-01 3.80268633e-01 5.24154842e-01 3.66856642e-02
-2.98619509e-01 -2.44020328e-01 -1.43892133e+00 4.61946368e-01
1.22402179e+00 -1.01923548e-01 -3.85901988e-01 2.14726537e-01
5.31566978e-01 -4.32904035e-01 -8.15471709e-01 6.50072098e-01
5.40063739e-01 -8.24891090e-01 8.99420559e-01 -8.24331224e-01
4.42706645e-01 -2.28536949e-01 1.03129126e-01 -1.04066682e+00
-7.71092534e-01 -5.83593726e-01 4.98403251e-01 1.33970201e+00
3.09990406e-01 -6.73549294e-01 7.92145133e-01 4.96402055e-01
6.98162764e-02 -8.43941987e-01 -6.35696590e-01 -2.15179637e-01
4.16112512e-01 -4.82127890e-02 6.33248508e-01 8.02871704e-01
-3.21591109e-01 3.88381809e-01 -1.63213283e-01 2.77451247e-01
4.21812654e-01 -3.17115456e-01 6.73638940e-01 -1.08653963e+00
-1.28932640e-01 -4.24894720e-01 -5.03253698e-01 -7.78333664e-01
2.99354017e-01 -6.76906586e-01 4.03597839e-02 -1.01268005e+00
1.47952318e-01 -3.28015029e-01 -3.10662627e-01 8.42303574e-01
-1.69621930e-01 2.89574444e-01 1.58493165e-02 -1.46141127e-01
-1.21488154e-01 1.56729341e-01 1.60129845e+00 1.12909619e-02
-3.77777517e-01 -8.83815661e-02 -1.09324837e+00 1.01604521e+00
1.19534099e+00 -3.44493002e-01 -7.34823167e-01 -5.53002059e-01
-1.05124392e-01 -4.06946033e-01 3.07119220e-01 -6.98644042e-01
-1.48686051e-01 -9.73896012e-02 8.29335213e-01 -1.37391582e-01
-1.89606860e-01 -7.13981688e-01 -4.54488575e-01 2.61663705e-01
-7.20246732e-01 1.66265480e-02 3.82920206e-01 1.97397709e-01
1.45512344e-02 -3.96171719e-01 1.15927541e+00 -2.61478961e-01
-7.24950075e-01 1.94558725e-01 -1.67908400e-01 -1.72735348e-01
1.09680533e+00 -2.63467431e-01 -2.04253480e-01 1.48055375e-01
-9.37747598e-01 -8.23768824e-02 1.27313018e-01 4.72575516e-01
4.25345272e-01 -1.12753856e+00 -3.75895262e-01 5.51496685e-01
-1.63973406e-01 3.18940371e-01 1.20322198e-01 4.15102959e-01
-6.61327958e-01 4.50516492e-01 -5.13018131e-01 -4.95673180e-01
-1.21534336e+00 6.05334461e-01 5.75990558e-01 -1.37201026e-01
-5.31066775e-01 1.10501432e+00 7.64306486e-01 -6.58653140e-01
2.66990781e-01 -7.59552836e-01 -7.07691073e-01 -2.98754185e-01
4.72517967e-01 -1.01459004e-01 -1.44884765e-01 -5.17633259e-01
-2.98741817e-01 9.53680992e-01 -4.71848577e-01 2.20427364e-01
1.48361337e+00 4.16634977e-03 -4.60579619e-02 -2.23427415e-01
1.15357208e+00 -1.14242777e-01 -1.64969599e+00 6.97268918e-02
-2.54978955e-01 -3.70651096e-01 1.67072043e-01 -6.50941432e-01
-1.40966165e+00 8.17805827e-01 9.11576271e-01 4.81775813e-02
1.52530491e+00 1.46156270e-02 4.34572607e-01 4.78611827e-01
-3.96972448e-02 -1.15865421e+00 -2.15699121e-01 3.74953121e-01
8.64663661e-01 -1.24116874e+00 -8.97784531e-02 -6.36371255e-01
-7.06095219e-01 1.38314772e+00 1.07757080e+00 -5.60018063e-01
5.42529583e-01 4.77819562e-01 2.70028979e-01 -1.29453555e-01
-5.33090472e-01 -1.51324838e-01 5.38322628e-02 8.30183208e-01
8.14134836e-01 -2.56983936e-01 -4.91001070e-01 5.05634069e-01
-2.71240026e-01 4.01301384e-02 2.46309832e-01 1.03794599e+00
-5.11694372e-01 -1.29343247e+00 -2.32130423e-01 2.53094107e-01
-6.97236598e-01 1.19288981e-01 -4.28179234e-01 1.04643631e+00
5.47609210e-01 5.92540145e-01 1.45663619e-01 -2.84974754e-01
4.47935343e-01 6.14914559e-02 8.30217123e-01 -5.29777825e-01
-7.81924844e-01 2.04126030e-01 -3.54782671e-01 -5.38601518e-01
-6.96720541e-01 -4.53400224e-01 -1.57847512e+00 -2.41826847e-02
-1.26105607e-01 -2.26553127e-01 6.56324387e-01 7.79101908e-01
2.42318615e-01 8.65973055e-01 6.77687407e-01 -6.10573828e-01
-4.72829223e-01 -1.18906891e+00 -2.76196837e-01 4.01149541e-01
1.54008284e-01 -6.94408178e-01 -1.12265972e-02 4.24767911e-01] | [13.429525375366211, 0.7268819212913513] |
a0897295-2f8f-4c4e-9699-60ec457960f0 | visually-grounded-few-shot-word-learning-in | 2306.11371 | null | https://arxiv.org/abs/2306.11371v2 | https://arxiv.org/pdf/2306.11371v2.pdf | Visually grounded few-shot word learning in low-resource settings | We propose a visually grounded speech model that learns new words and their visual depictions from just a few word-image example pairs. Given a set of test images and a spoken query, we ask the model which image depicts the query word. Previous work has simplified this few-shot learning problem by either using an artificial setting with digit word-image pairs or by using a large number of examples per class. Moreover, all previous studies were performed using English speech-image data. We propose an approach that can work on natural word-image pairs but with less examples, i.e. fewer shots, and then illustrate how this approach can be applied for multimodal few-shot learning in a real low-resource language, Yoruba. Our approach involves using the given word-image example pairs to mine new unsupervised word-image training pairs from large collections of unlabelledspeech and images. Additionally, we use a word-to-image attention mechanism to determine word-image similarity. With this new model, we achieve better performance with fewer shots than previous approaches on an existing English benchmark. Many of the model's mistakes are due to confusion between visual concepts co-occurring in similar contexts. The experiments on Yoruba show the benefit of transferring knowledge from a multimodal model trained on a larger set of English speech-image data. | ['Herman Kamper', 'Dan Oneata', 'Leanne Nortje'] | 2023-06-20 | null | null | null | null | ['few-shot-learning'] | ['methodology'] | [ 3.65558267e-01 1.19908072e-01 4.04621884e-02 -4.53820229e-01
-1.19461274e+00 -2.23970011e-01 8.25277448e-01 -6.12071082e-02
-7.79538810e-01 6.50833607e-01 2.29082897e-01 -6.20289892e-03
4.29668903e-01 -6.68092310e-01 -7.62523532e-01 -5.93106627e-01
1.69827431e-01 5.39014220e-01 5.53553402e-01 -4.54771638e-01
2.91968733e-01 -1.99323464e-02 -1.80785728e+00 5.71809709e-01
3.73712718e-01 6.15180135e-01 7.32593894e-01 8.32372665e-01
-5.17253995e-01 8.76102388e-01 -7.83063591e-01 -4.77630734e-01
1.22849360e-01 -8.02809656e-01 -8.98373187e-01 7.29790986e-01
4.85808134e-01 -3.52634341e-01 -2.33984932e-01 1.05455279e+00
6.65202856e-01 7.04473972e-01 5.49240649e-01 -1.23738372e+00
-9.01339233e-01 4.10946280e-01 -3.90551090e-01 3.16568166e-01
5.63181281e-01 3.68946433e-01 7.79170573e-01 -1.33586872e+00
8.21836174e-01 1.39259946e+00 2.09094748e-01 7.65013695e-01
-1.06521153e+00 -4.97400701e-01 -6.47408664e-02 6.35337532e-01
-1.55727744e+00 -7.04693139e-01 6.21296048e-01 -1.20732851e-01
1.22098947e+00 6.34237751e-03 6.71655476e-01 1.25969267e+00
-3.72910947e-01 8.95199358e-01 8.56759548e-01 -8.43390942e-01
3.69044989e-01 5.33751547e-01 -1.30603381e-03 8.21102917e-01
-3.20743531e-01 3.26767080e-02 -5.57243884e-01 1.00415491e-01
5.19270599e-01 2.01244354e-01 -4.14986938e-01 -5.74880958e-01
-1.15742683e+00 9.60534990e-01 4.05340225e-01 5.55334270e-01
-1.30288079e-01 1.01860866e-01 3.09799880e-01 2.99584985e-01
4.47053760e-01 3.35982203e-01 -9.64404643e-03 -1.24962024e-01
-9.54894662e-01 -1.65117815e-01 7.15535581e-01 9.24661577e-01
1.11617875e+00 1.69603899e-01 -1.85476542e-01 1.22948599e+00
1.02745987e-01 6.13081217e-01 9.72956479e-01 -6.81811750e-01
4.48357522e-01 1.91625193e-01 -1.00633882e-01 -7.92287052e-01
7.18754828e-02 3.17282796e-01 -3.69729340e-01 7.72481710e-02
1.70351997e-01 2.69509889e-02 -1.55998552e+00 1.54813099e+00
1.58864465e-02 3.88494581e-01 5.91858149e-01 9.99291658e-01
1.19459581e+00 1.00663209e+00 5.68008749e-03 -2.09138259e-01
1.61585557e+00 -1.23415005e+00 -7.77555585e-01 -6.26691103e-01
5.67327321e-01 -6.75688088e-01 1.60043895e+00 8.11430365e-02
-9.91324782e-01 -5.68352222e-01 -1.00869679e+00 -1.72935352e-02
-9.24374759e-01 -2.21641362e-01 6.48035556e-02 6.35703444e-01
-9.69584823e-01 2.22469643e-01 -3.52695435e-01 -9.79869723e-01
1.79336086e-01 -5.19307442e-02 -4.69592750e-01 -4.57501054e-01
-1.29785764e+00 9.51681614e-01 5.84932804e-01 -4.19122815e-01
-1.19948184e+00 -2.32389301e-01 -1.43127549e+00 1.87321186e-01
7.57684410e-01 -3.38050425e-01 1.43787146e+00 -1.25402439e+00
-1.38520837e+00 9.79052722e-01 -1.93520948e-01 -4.08339202e-01
-1.71986893e-02 2.12950081e-01 -5.75701475e-01 6.83444798e-01
1.46169335e-01 1.13609838e+00 1.07561564e+00 -1.62096238e+00
-4.04544473e-01 -1.20522425e-01 -1.22161461e-02 3.49429995e-01
-4.59236771e-01 -1.27247917e-02 -8.03444445e-01 -5.81508160e-01
-2.49871776e-01 -6.72731757e-01 -1.34294972e-01 -1.71473980e-01
-3.29526514e-02 -1.60129333e-03 7.52982080e-01 -3.84295523e-01
7.99840987e-01 -2.21588469e+00 -6.32018745e-02 -7.31545910e-02
-5.78007586e-02 3.55841219e-01 -6.06471777e-01 6.86041057e-01
-9.44144949e-02 1.34751555e-02 -4.23296034e-01 -4.87434119e-01
-2.84640014e-01 5.04973590e-01 -1.82544500e-01 6.59503788e-02
2.86132902e-01 1.17593622e+00 -1.10318577e+00 -8.38911593e-01
3.84307593e-01 3.02133054e-01 -1.91527203e-01 2.92673588e-01
-1.55338943e-02 -6.79916292e-02 2.48806193e-01 6.26750290e-01
1.82341799e-01 -2.54930973e-01 -6.53028712e-02 -1.40284851e-01
2.99629331e-01 -4.31928247e-01 -9.57798719e-01 1.92370331e+00
-5.46451390e-01 8.75790834e-01 -4.07744169e-01 -1.20976973e+00
9.00565803e-01 5.14533758e-01 4.77188304e-02 -9.41935241e-01
1.47193328e-01 3.52410153e-02 -1.75950617e-01 -1.08424926e+00
4.96757954e-01 -6.37305140e-01 -1.84877783e-01 5.51247895e-01
5.69891036e-01 -5.43850720e-01 3.93801212e-01 4.60713089e-01
8.18336010e-01 -2.93969154e-01 4.74319786e-01 2.55250484e-01
3.66884410e-01 1.20717242e-01 4.45920452e-02 8.25543702e-01
-3.38682443e-01 1.02022564e+00 5.67313954e-02 -3.53920847e-01
-1.23403323e+00 -1.04936326e+00 1.76987275e-01 1.31368053e+00
2.72390634e-01 -3.57514441e-01 -6.00120246e-01 -5.95011830e-01
-4.89001185e-01 9.94922638e-01 -7.22809732e-01 -2.42417991e-01
-1.39009401e-01 -3.54134649e-01 3.84349704e-01 4.44831580e-01
4.28631037e-01 -1.47013509e+00 -8.32180560e-01 2.85571925e-02
-1.80225879e-01 -1.33291781e+00 -5.92622936e-01 2.36329764e-01
-3.46317947e-01 -8.45883667e-01 -1.33374047e+00 -1.32100892e+00
6.54037535e-01 6.49018943e-01 1.20194244e+00 -8.14690590e-02
-7.13651180e-01 9.48803246e-01 -7.56146550e-01 -5.04186928e-01
-4.52744156e-01 -5.23326516e-01 -7.20610619e-02 1.64453939e-01
6.29545748e-01 -5.62237725e-02 -3.64529431e-01 1.51420325e-01
-1.10436141e+00 6.94925562e-02 6.69465303e-01 1.22812700e+00
2.87505031e-01 -4.39051419e-01 5.18031180e-01 -5.92010021e-01
6.70878649e-01 -5.03897488e-01 -2.42227297e-02 5.87092221e-01
-1.79008424e-01 2.66579054e-02 1.35412276e-01 -8.18115592e-01
-1.04914188e+00 2.48510465e-01 1.94357142e-01 -8.51752520e-01
-4.37566131e-01 3.44650000e-01 1.06349014e-01 1.58089101e-01
7.10762084e-01 5.10863602e-01 3.16585690e-01 -4.52680476e-02
7.70517886e-01 9.31533575e-01 6.50346100e-01 -1.42873555e-01
4.86702919e-01 4.04097438e-01 -5.24602234e-01 -1.39307523e+00
-5.73444307e-01 -9.23008025e-01 -6.02097213e-01 -2.45039061e-01
1.05769801e+00 -9.29550052e-01 -1.41397014e-01 7.36015737e-02
-1.17753160e+00 -2.06271738e-01 -4.86745864e-01 6.05246961e-01
-7.03457236e-01 5.60658216e-01 -3.63836139e-01 -1.02346921e+00
-2.11221561e-01 -1.10441327e+00 1.36363792e+00 1.49816081e-01
-7.57547989e-02 -8.91392231e-01 1.24007359e-01 1.99538201e-01
9.17244256e-02 -3.07277560e-01 6.40755892e-01 -9.16712642e-01
-1.93144023e-01 -1.18930772e-01 -2.34027460e-01 3.44969183e-01
2.00146854e-01 -3.48091453e-01 -1.01993096e+00 -2.15332866e-01
9.51545089e-02 -9.15376067e-01 9.82228577e-01 6.60826787e-02
7.23305464e-01 -2.05788851e-01 -1.66467294e-01 6.08719513e-02
1.48422670e+00 4.69772458e-01 6.43858612e-01 -8.54474492e-05
5.53888142e-01 8.33992302e-01 7.84942508e-01 2.41001531e-01
2.30080694e-01 6.46481931e-01 1.98268756e-01 -4.06948715e-01
-1.96298078e-01 -1.44527376e-01 4.02560234e-01 9.37801719e-01
2.84684032e-01 -3.37955147e-01 -1.19266653e+00 1.01760590e+00
-1.93209839e+00 -1.19727767e+00 5.65123379e-01 1.98935640e+00
7.44270921e-01 -6.77612871e-02 1.37463823e-01 -1.73218340e-01
8.70411754e-01 1.29755139e-01 -3.43662560e-01 -3.23684067e-01
-1.11385271e-01 2.03265414e-01 1.22761562e-01 4.40355003e-01
-8.96988809e-01 1.31690538e+00 6.10394382e+00 7.85271406e-01
-1.03328609e+00 3.35280567e-01 5.84387898e-01 -3.11789989e-01
-1.43226877e-01 -1.87935039e-01 -3.60790521e-01 2.17933148e-01
8.84872735e-01 -2.17846021e-01 3.32301617e-01 8.32289755e-01
-9.04359445e-02 -2.85775423e-01 -9.96540666e-01 1.57261550e+00
1.10848892e+00 -1.29484475e+00 3.60347092e-01 -3.00634444e-01
5.40093958e-01 -3.73884737e-02 -1.19564943e-01 5.22588909e-01
2.10266992e-01 -1.09832013e+00 4.47376430e-01 5.42058110e-01
7.39077806e-01 -7.29006708e-01 7.08369017e-01 3.46444547e-01
-1.08560109e+00 -8.62731189e-02 -4.94079918e-01 1.14309706e-01
3.06224108e-01 -2.89759338e-01 -1.23317647e+00 2.34559581e-01
8.40087533e-01 5.06704926e-01 -7.01871395e-01 8.71156871e-01
-8.32320899e-02 1.38529435e-01 4.87068035e-02 -5.05105138e-01
5.79423666e-01 3.98730859e-02 2.91835636e-01 1.26724231e+00
3.51407379e-01 3.00586075e-01 2.47056693e-01 6.88588321e-01
3.10007669e-02 4.13407803e-01 -1.13420117e+00 -2.13820785e-01
3.62480730e-01 1.25195026e+00 -9.22757745e-01 -9.92123067e-01
-8.72983456e-01 1.35664296e+00 2.92282671e-01 4.53505903e-01
-5.22742450e-01 -7.01331198e-01 2.17007220e-01 -1.17695615e-01
6.32022560e-01 -7.93103948e-02 4.90602732e-01 -1.16631150e+00
-3.24875504e-01 -9.14372861e-01 2.43284360e-01 -1.38262618e+00
-1.33109379e+00 7.89171815e-01 2.75108248e-01 -1.35184479e+00
-5.49143672e-01 -5.41711748e-01 -7.32733071e-01 4.45153654e-01
-1.27593637e+00 -1.01715541e+00 -3.69911313e-01 8.76284778e-01
1.38543046e+00 -5.02849698e-01 9.64972019e-01 6.53235316e-02
-1.27568394e-01 3.89885187e-01 -1.18813001e-01 2.21391469e-01
8.49618793e-01 -1.11127198e+00 4.54397082e-01 7.17380524e-01
9.22447026e-01 3.52328211e-01 6.56099081e-01 -4.92469519e-01
-1.10293937e+00 -6.56239688e-01 6.97565258e-01 -4.29679930e-01
6.83017671e-01 -5.40920377e-01 -1.04976261e+00 4.34926450e-01
8.76677692e-01 6.59987256e-02 8.51311922e-01 -3.52777958e-01
-1.50112554e-01 2.53228515e-01 -9.50586021e-01 7.31357038e-01
7.62960255e-01 -8.31964910e-01 -1.23035145e+00 4.25543845e-01
8.99271607e-01 1.48325354e-01 -4.44242477e-01 1.17462635e-01
3.12074095e-01 -6.39106929e-01 9.86779869e-01 -8.01347196e-01
3.01844120e-01 -1.88050956e-01 -4.52122658e-01 -1.54719329e+00
1.22206368e-01 -2.62892753e-01 1.66670024e-01 1.10310853e+00
3.69371980e-01 -1.00593127e-01 5.38170516e-01 2.48023033e-01
7.44992569e-02 -4.16587055e-01 -8.71512473e-01 -9.23649907e-01
-2.05468625e-01 -4.85433608e-01 2.94874579e-01 8.82846713e-01
3.04104507e-01 9.19139564e-01 -6.31976068e-01 -2.03784943e-01
3.71409565e-01 -7.91499317e-02 9.09397125e-01 -7.18570292e-01
-2.46963441e-01 -9.08377096e-02 -6.89687073e-01 -6.65826917e-01
1.58907458e-01 -6.82837188e-01 6.34196281e-01 -1.56894469e+00
4.76711065e-01 4.68555838e-01 -1.43384859e-01 5.94880760e-01
-3.04454267e-01 5.72642446e-01 5.98109961e-01 1.67521209e-01
-8.92697036e-01 7.53980398e-01 1.14627671e+00 -5.34784913e-01
-7.97847882e-02 -5.49393654e-01 -2.07067341e-01 5.99857032e-01
4.94883776e-01 -3.66058022e-01 -6.42784059e-01 -2.02353731e-01
-2.77165055e-01 -4.55142651e-03 4.92368519e-01 -9.02805328e-01
3.79834086e-01 4.04667892e-02 3.76000911e-01 -5.66998780e-01
8.73670220e-01 -8.39028597e-01 -3.70285332e-01 4.20307666e-01
-2.39918888e-01 9.57758576e-02 1.45113826e-01 8.14532161e-01
-5.36765337e-01 -6.51955783e-01 7.85928130e-01 -7.44791746e-01
-1.58184159e+00 -3.86298448e-02 -5.05048811e-01 5.81173226e-02
1.43800807e+00 -4.59319949e-01 -3.12842518e-01 -7.25832999e-01
-1.02121580e+00 1.37386173e-01 4.39422160e-01 7.65997052e-01
1.16705310e+00 -1.45850551e+00 -5.55450916e-01 3.17289561e-01
8.11739326e-01 -4.68569368e-01 2.41788149e-01 5.44470131e-01
-3.26867640e-01 2.58690804e-01 -3.20669562e-01 -7.88661420e-01
-1.58389831e+00 9.94364202e-01 1.33071989e-01 4.24313933e-01
-6.51248932e-01 8.00756454e-01 2.83528477e-01 -3.44998628e-01
4.31769639e-01 -2.82542594e-02 -2.48819366e-01 3.85462046e-01
7.32353866e-01 -9.26582217e-02 -2.32595086e-01 -9.36972737e-01
-2.82994837e-01 6.78792894e-01 -5.93429580e-02 -6.76019311e-01
1.16942668e+00 -4.17688042e-01 4.26608235e-01 1.03488410e+00
1.45879078e+00 -3.64491999e-01 -1.03155196e+00 -6.07195735e-01
-2.06866823e-02 -5.19975781e-01 -1.87172890e-01 -5.10365248e-01
-8.05757821e-01 1.22741365e+00 9.83692527e-01 1.24637097e-01
9.08375204e-01 5.89378655e-01 5.68193138e-01 7.86685169e-01
1.68754563e-01 -1.24065816e+00 9.74259675e-01 4.60778743e-01
7.60631561e-01 -1.87442660e+00 -2.88924754e-01 3.73059958e-02
-1.36241448e+00 1.08937335e+00 5.47907770e-01 1.94525719e-01
3.53122473e-01 -8.36400613e-02 4.01234239e-01 -3.26816499e-01
-7.31537819e-01 -9.11633193e-01 3.04562360e-01 7.76743352e-01
8.62391964e-02 -2.26782054e-01 1.41534284e-01 3.88169110e-01
1.04183368e-01 -1.96342692e-01 5.40010035e-01 1.10405684e+00
-8.87324572e-01 -8.19907308e-01 -3.88942987e-01 2.09903553e-01
-6.35772804e-03 -5.51740944e-01 -5.53061783e-01 8.89894843e-01
-1.62286460e-01 1.03357935e+00 4.41282660e-01 -2.61658013e-01
3.13488036e-01 5.28574765e-01 4.09668058e-01 -1.21199048e+00
-1.65791392e-01 1.83918148e-01 -6.20110966e-02 -4.04523104e-01
-6.55091345e-01 -3.95389825e-01 -1.25765538e+00 2.90110439e-01
-2.32213542e-01 2.25330487e-01 6.37005508e-01 9.84383941e-01
1.03053555e-01 4.11714524e-01 5.04062235e-01 -1.00163412e+00
-1.73403025e-01 -1.09721839e+00 -6.52339935e-01 7.21942723e-01
3.01362157e-01 -6.17520034e-01 -3.50575089e-01 3.28044951e-01] | [10.404559135437012, 1.9685567617416382] |
c060b0be-89d3-43bd-b34c-1e073503a430 | dynamic-message-propagation-network-for-rgb-d | 2206.09552 | null | https://arxiv.org/abs/2206.09552v1 | https://arxiv.org/pdf/2206.09552v1.pdf | Dynamic Message Propagation Network for RGB-D Salient Object Detection | This paper presents a novel deep neural network framework for RGB-D salient object detection by controlling the message passing between the RGB images and depth maps on the feature level and exploring the long-range semantic contexts and geometric information on both RGB and depth features to infer salient objects. To achieve this, we formulate a dynamic message propagation (DMP) module with the graph neural networks and deformable convolutions to dynamically learn the context information and to automatically predict filter weights and affinity matrices for message propagation control. We further embed this module into a Siamese-based network to process the RGB image and depth map respectively and design a multi-level feature fusion (MFF) module to explore the cross-level information between the refined RGB and depth features. Compared with 17 state-of-the-art methods on six benchmark datasets for RGB-D salient object detection, experimental results show that our method outperforms all the others, both quantitatively and visually. | ['Jing Qin', 'Mingqiang Wei', 'Haoran Xie', 'Jun Xu', 'Xiaowei Hu', 'Zhilei Chen', 'Baian Chen'] | 2022-06-20 | null | null | null | null | ['rgb-d-salient-object-detection'] | ['computer-vision'] | [ 1.94541857e-01 7.76503906e-02 1.80854678e-01 -4.84746128e-01
-4.11538661e-01 -1.70663908e-01 4.90905285e-01 4.46403384e-01
-4.28689152e-01 1.39295131e-01 2.52763212e-01 1.06052207e-02
-1.87365651e-01 -9.07394826e-01 -7.57216871e-01 -6.90246582e-01
-4.50905621e-01 1.26942590e-01 9.35457587e-01 -3.05992335e-01
4.97373879e-01 7.50102460e-01 -1.91098332e+00 2.30528429e-01
5.58260560e-01 1.48872888e+00 4.03756917e-01 9.04409409e-01
-1.76098958e-01 9.71912622e-01 9.11719278e-02 4.38493155e-02
5.09519517e-01 -6.60262108e-02 -9.48660254e-01 3.03237021e-01
4.06054229e-01 -3.11218023e-01 -4.76140052e-01 1.28824377e+00
4.62457925e-01 3.18467170e-01 2.17393339e-01 -1.21967292e+00
-8.66782546e-01 2.56547242e-01 -8.44398677e-01 3.42333853e-01
4.06888932e-01 1.99145138e-01 7.44499266e-01 -1.17472458e+00
8.04989934e-01 1.50162411e+00 3.65749151e-01 2.74084330e-01
-8.62510741e-01 -3.04965258e-01 5.89927614e-01 4.43699837e-01
-1.18872154e+00 1.15797585e-02 1.30995166e+00 -2.17772633e-01
8.06859910e-01 3.41265917e-01 1.23109376e+00 4.34004128e-01
1.76990002e-01 1.09032011e+00 8.44392717e-01 -1.13193549e-01
2.10756660e-01 -1.54348120e-01 8.06045309e-02 1.42710018e+00
-5.18299378e-02 -2.85353488e-03 -9.83709157e-01 -6.21111430e-02
9.87576783e-01 2.02214241e-01 -2.16336280e-01 -7.21681952e-01
-1.35310125e+00 8.15185368e-01 1.14700580e+00 1.08396560e-01
-4.53227252e-01 1.73485294e-01 -5.10568209e-02 4.48366702e-02
4.15443659e-01 2.14617252e-01 -2.78178751e-01 4.50435102e-01
-6.91192806e-01 2.03659385e-01 5.69516063e-01 8.53943050e-01
1.23189187e+00 -3.23720425e-01 -3.75154376e-01 3.07435900e-01
8.59569252e-01 4.09964174e-01 1.92355692e-01 -9.01982546e-01
3.58115315e-01 1.22734702e+00 1.13557808e-01 -1.66170251e+00
-7.46428728e-01 -3.99393886e-01 -7.14767516e-01 4.48119968e-01
1.13776498e-01 3.94367516e-01 -1.17788935e+00 1.33628488e+00
9.33134675e-01 3.05487484e-01 -1.23892710e-01 1.70212615e+00
1.18180811e+00 6.93367124e-01 -7.95502122e-03 1.63782731e-01
1.17731905e+00 -9.13822412e-01 -4.10289228e-01 -2.75793821e-01
1.50875106e-01 -6.82706416e-01 9.69786048e-01 -1.93326846e-01
-1.34878492e+00 -7.17704594e-01 -1.30448198e+00 -7.10352242e-01
-7.05256641e-01 -8.54403004e-02 9.75230575e-01 -1.51729241e-01
-1.19585943e+00 5.09538949e-01 -1.11437035e+00 -2.93985069e-01
4.54425395e-01 5.85470974e-01 -2.55513132e-01 1.40824631e-01
-1.28335774e+00 7.73367584e-01 1.91341013e-01 4.90909725e-01
-9.25349891e-01 -6.25149250e-01 -1.09690666e+00 -1.08005561e-01
-2.12843204e-03 -9.55264807e-01 6.55296981e-01 -7.41581440e-01
-1.51065111e+00 1.07557642e+00 9.72179919e-02 -2.19441891e-01
4.06667143e-01 8.62947181e-02 1.09704599e-01 3.12334150e-01
-8.41918588e-02 1.00085914e+00 9.39043462e-01 -1.34866691e+00
-1.16055632e+00 -6.29160404e-01 2.94932365e-01 5.04174531e-01
-9.65866596e-02 -2.49435097e-01 -6.57162189e-01 -3.73284876e-01
8.11907113e-01 -5.56920767e-01 -3.04478556e-01 4.62599725e-01
-5.86586833e-01 -2.08281577e-01 1.03526413e+00 -5.00522196e-01
7.06580341e-01 -1.97438753e+00 7.42591023e-01 3.98143381e-01
5.87119341e-01 -1.88113093e-01 -5.62109053e-02 -1.72198266e-01
3.74457031e-01 -4.13217902e-01 -4.30690944e-01 -5.91420949e-01
5.96085601e-02 1.34673327e-01 5.04201241e-02 8.70282292e-01
5.25341272e-01 1.22474527e+00 -1.14953125e+00 -6.33874178e-01
6.04352176e-01 8.34114015e-01 -4.81250376e-01 3.88331950e-01
-1.35093763e-01 3.61556679e-01 -7.97935307e-01 8.68430674e-01
8.76474857e-01 -4.69645619e-01 -5.58428109e-01 -4.11556333e-01
-2.46675864e-01 9.27693769e-02 -1.27153850e+00 2.16432333e+00
-4.23169397e-02 5.33026338e-01 1.82152137e-01 -7.10507810e-01
9.61701870e-01 -5.20689547e-01 4.28665578e-01 -6.93446755e-01
3.47952217e-01 -1.85608834e-01 -4.99213487e-01 -5.30864418e-01
7.16623545e-01 3.40934336e-01 -3.08200940e-02 1.54917091e-01
-6.54403642e-02 -2.97481120e-01 -2.73296624e-01 3.07556719e-01
8.93004715e-01 3.43179144e-02 -3.27568293e-01 -3.09287816e-01
8.81101072e-01 -1.20748915e-01 3.93300325e-01 5.76241076e-01
-2.74992317e-01 5.84426761e-01 1.45532429e-01 -7.54875243e-01
-6.47805154e-01 -1.25253904e+00 1.92366183e-01 1.21610713e+00
1.08891642e+00 -1.98800606e-03 -5.39257050e-01 -5.66237986e-01
4.00772244e-01 1.21049732e-01 -1.07853234e+00 -4.05211508e-01
-4.16538715e-01 -5.69907427e-01 -8.15508291e-02 3.24396014e-01
9.44332778e-01 -8.84894788e-01 -9.91788983e-01 1.09984212e-01
-1.79613708e-04 -9.76345539e-01 -2.81382620e-01 2.67456770e-01
-6.45154297e-01 -9.61726904e-01 -4.25806195e-01 -1.08234954e+00
8.52782071e-01 3.31738472e-01 8.74387622e-01 2.91164577e-01
-5.18833518e-01 3.60282362e-01 -1.71734273e-01 -1.99857801e-01
-2.23337095e-02 1.69388920e-01 -4.63469565e-01 3.17324609e-01
1.26687005e-01 -4.95029509e-01 -1.05758774e+00 6.60990030e-02
-9.77469683e-01 4.88424838e-01 7.55704939e-01 5.13454556e-01
8.87501955e-01 -2.77837604e-01 -3.65891382e-02 -2.19834790e-01
3.32799882e-01 -2.98300475e-01 -8.91721427e-01 2.66272813e-01
-2.72640616e-01 1.78814292e-01 -2.00150773e-01 -1.75248802e-01
-8.44111621e-01 4.74750370e-01 2.96595339e-02 -3.86914402e-01
1.63514838e-01 1.73234731e-01 6.06277287e-02 -6.55426800e-01
2.44477332e-01 3.07820469e-01 -1.32671803e-01 -2.10613638e-01
5.26261270e-01 3.19407970e-01 6.26405418e-01 -9.79505703e-02
1.02399647e+00 9.28748369e-01 2.20236406e-01 -4.99929935e-01
-1.10629010e+00 -5.28299332e-01 -8.31550896e-01 -4.76595193e-01
1.23852932e+00 -8.58057916e-01 -9.19244409e-01 6.28064036e-01
-1.15482485e+00 -3.66099268e-01 -2.02582583e-01 2.36274600e-01
-7.01442361e-01 3.98294292e-02 -8.49368274e-01 -5.34187019e-01
-3.77362132e-01 -1.23787308e+00 1.70078576e+00 5.92056811e-01
2.87814617e-01 -8.65425706e-01 -1.67152330e-01 -2.20577553e-04
3.74348879e-01 6.69964254e-01 7.68679202e-01 2.53242970e-01
-1.31013370e+00 -5.84579110e-02 -7.44040906e-01 -1.35122523e-01
-3.80620249e-02 -8.79178494e-02 -8.12571526e-01 -1.33822992e-01
-6.54453412e-02 -1.16793558e-01 1.05544436e+00 4.96284842e-01
1.20093751e+00 -5.23402356e-02 -3.21420103e-01 1.12507844e+00
1.43455756e+00 -4.92436171e-01 3.47641438e-01 4.88333851e-01
1.06000316e+00 4.85164613e-01 5.82034588e-01 5.14199913e-01
8.07663262e-01 3.26327652e-01 1.04063988e+00 -5.26758730e-01
-1.85733750e-01 -3.55792522e-01 -6.82637021e-02 5.83404481e-01
1.85166821e-02 2.41294160e-01 -6.70728028e-01 4.94360775e-01
-2.15941191e+00 -5.11618674e-01 1.59274507e-02 1.71732247e+00
7.87629008e-01 2.91488826e-01 -8.39959234e-02 3.87809649e-02
6.51107728e-01 2.68486679e-01 -7.44293928e-01 2.45263595e-02
-3.40241373e-01 -1.84730008e-01 7.13172376e-01 8.13275039e-01
-1.35876322e+00 1.07918990e+00 5.37915516e+00 3.69208634e-01
-1.24857461e+00 -8.92983079e-02 5.08847415e-01 -2.59175539e-01
-4.59612697e-01 -1.88459694e-01 -6.32460356e-01 1.28418848e-01
1.30896404e-01 2.20150843e-01 4.03445780e-01 6.62606597e-01
7.03876046e-03 -3.24810565e-01 -8.65023673e-01 1.07758236e+00
1.10850140e-01 -1.63859344e+00 2.32577696e-01 -1.86792120e-01
8.63056183e-01 3.48828524e-01 1.77922606e-01 -2.00646773e-01
2.85243243e-01 -5.58552384e-01 1.08389747e+00 8.97362292e-01
7.76441991e-02 -8.04247677e-01 5.49279690e-01 -9.42735374e-03
-1.47512925e+00 -3.02294135e-01 -5.37252247e-01 1.77979562e-02
-4.81017679e-03 4.62936401e-01 -3.34960043e-01 3.35575521e-01
1.03831208e+00 1.15212119e+00 -8.73531580e-01 9.70926583e-01
-2.70915270e-01 -2.13056892e-01 -3.63325775e-01 -3.78377676e-01
6.27610326e-01 1.45008981e-01 6.64350092e-01 1.15792000e+00
-3.14083993e-02 1.17094271e-01 2.10308835e-01 1.16743588e+00
6.34508505e-02 -6.46149591e-02 -2.31758177e-01 5.59250593e-01
2.13924676e-01 1.59752190e+00 -1.17880380e+00 -2.59967208e-01
-2.55963087e-01 1.16203225e+00 5.89865565e-01 4.54556972e-01
-7.31387854e-01 -3.98047179e-01 8.53770733e-01 1.40956361e-02
3.77315462e-01 -3.64464343e-01 -3.65540832e-01 -9.86408234e-01
-6.61067888e-02 -6.04047962e-02 3.21537763e-01 -1.07087564e+00
-1.03135967e+00 4.12842959e-01 -1.99283317e-01 -1.01887250e+00
2.95480579e-01 -5.64354956e-01 -2.84009904e-01 8.28979194e-01
-2.11094069e+00 -1.40060365e+00 -7.04038262e-01 1.04211843e+00
1.88225284e-01 2.58489549e-01 2.62342900e-01 -3.30558270e-02
-1.41536713e-01 7.58315623e-02 -3.90505672e-01 1.94184199e-01
6.02677576e-02 -1.18937242e+00 3.46651524e-01 8.92801702e-01
-2.44254991e-02 2.86845088e-01 5.19313693e-01 -4.27302450e-01
-1.87024021e+00 -1.19847119e+00 4.67846274e-01 -3.11739862e-01
5.75948477e-01 -6.32546425e-01 -7.26855755e-01 3.05101663e-01
1.40114930e-02 3.51303995e-01 8.38785619e-02 -5.20986140e-01
-7.38305822e-02 -1.61362588e-01 -1.04272234e+00 4.65431809e-01
1.39852798e+00 -6.54339433e-01 -6.00259662e-01 2.48081088e-01
1.16157937e+00 -8.83028924e-01 -8.99079680e-01 4.91689771e-01
3.81411463e-01 -1.03510976e+00 1.28879023e+00 -3.89118701e-01
4.20086294e-01 -7.69213319e-01 -2.68331766e-01 -9.02715266e-01
-4.99528825e-01 -4.71680105e-01 -2.42938608e-01 8.44243944e-01
2.75995433e-01 -3.13603908e-01 8.34907889e-01 6.41957581e-01
-2.42536113e-01 -9.77933347e-01 -1.07968879e+00 1.57991216e-01
-6.05061829e-01 -3.87220502e-01 7.43733466e-01 7.75264442e-01
-1.48197472e-01 -2.48546265e-02 5.10591157e-02 5.86963534e-01
1.02387214e+00 7.43363321e-01 4.86994028e-01 -1.12041199e+00
1.74202651e-01 -6.92974091e-01 -1.05079365e+00 -1.11272049e+00
9.15356874e-02 -1.06471145e+00 1.82204172e-01 -1.71021295e+00
2.02140287e-01 -3.75157624e-01 -6.36368215e-01 3.87883872e-01
-4.22127306e-01 4.08828616e-01 1.96964055e-01 -5.51863462e-02
-9.91705835e-01 8.49329174e-01 1.63023901e+00 -4.59942222e-01
-4.40703511e-01 -4.11091030e-01 -6.00120723e-01 6.48706734e-01
4.93437022e-01 -2.35342413e-01 -1.86128750e-01 -7.09389687e-01
4.09285843e-01 -1.64273039e-01 9.12732661e-01 -1.05468667e+00
7.22087145e-01 -6.94742948e-02 7.97944784e-01 -8.79996836e-01
2.79350042e-01 -7.76633441e-01 -4.87159848e-01 4.11729276e-01
-6.06238306e-01 -7.99940825e-02 1.90722272e-01 7.62217522e-01
-9.86870155e-02 6.87971473e-01 8.37586164e-01 -1.82576120e-01
-1.15152466e+00 7.42279232e-01 -9.23539847e-02 -3.37678171e-03
1.04544091e+00 -3.32762837e-01 -2.16875941e-01 -1.75264448e-01
-7.59461343e-01 4.61836159e-01 5.27828991e-01 7.74793506e-01
1.22812283e+00 -1.43492424e+00 -4.38288361e-01 6.52080536e-01
1.87765017e-01 4.64057475e-01 2.03206703e-01 7.74113595e-01
-5.86073279e-01 -1.28538504e-01 -4.25938815e-01 -1.05713272e+00
-9.38058674e-01 4.25507158e-01 4.90921170e-01 3.14958617e-02
-5.63788295e-01 1.49408185e+00 4.26104926e-02 -3.69456142e-01
4.54695523e-01 -8.12431455e-01 -2.11426750e-01 -1.41724303e-01
5.26655972e-01 1.70600772e-01 -6.28667623e-02 -7.92156100e-01
-5.78931153e-01 9.09382761e-01 2.28459954e-01 1.73234358e-01
1.44817579e+00 -6.42450571e-01 -5.21395862e-01 4.18419451e-01
1.53343856e+00 -5.82260609e-01 -1.76346612e+00 -3.66398811e-01
-2.58678347e-02 -5.97763121e-01 5.60341656e-01 -3.10266137e-01
-1.31305206e+00 7.31873691e-01 9.53651726e-01 1.67232290e-01
1.12832022e+00 3.70426804e-01 7.66540349e-01 1.93491921e-01
2.22319826e-01 -1.22027397e+00 2.59422213e-01 3.97530288e-01
1.00509727e+00 -1.39721906e+00 1.51947409e-01 -4.13461030e-01
-2.71090299e-01 9.38726783e-01 6.05299294e-01 -4.56721395e-01
1.11015499e+00 1.00967780e-01 1.33627608e-01 -4.82526898e-01
-5.08818090e-01 -5.36291599e-01 5.03635228e-01 5.34658253e-01
-1.74346462e-01 -1.73964903e-01 3.24106216e-01 1.01738967e-01
-9.84121934e-02 -1.57204375e-01 3.78054716e-02 9.03997540e-01
-7.57900178e-01 -3.71750116e-01 -2.25040570e-01 2.59347260e-01
1.21973559e-01 7.33679067e-03 -4.74970251e-01 6.88928485e-01
2.38209069e-01 7.61889398e-01 3.08292896e-01 -6.22530878e-01
2.87160009e-01 -3.81207407e-01 4.14900601e-01 -3.14248741e-01
-5.90560198e-01 -1.94357008e-01 -4.62062538e-01 -1.08398378e+00
-9.46159661e-01 -4.48665231e-01 -1.59759784e+00 -1.94499955e-01
-2.25959420e-01 -2.52759814e-01 7.73276329e-01 9.26712453e-01
4.23301786e-01 6.43619299e-01 6.70647144e-01 -1.49695051e+00
6.76086918e-02 -5.22729933e-01 -5.22984982e-01 4.32455808e-01
7.98053563e-01 -8.46441746e-01 -3.61859351e-01 -1.83250710e-01] | [9.733587265014648, -0.7327616214752197] |
470ef5bc-52f8-44e7-a75e-d33940d9474c | information-theoretic-limits-of-exact | 2101.12369 | null | https://arxiv.org/abs/2101.12369v1 | https://arxiv.org/pdf/2101.12369v1.pdf | Information Theoretic Limits of Exact Recovery in Sub-hypergraph Models for Community Detection | In this paper, we study the information theoretic bounds for exact recovery in sub-hypergraph models for community detection. We define a general model called the $m-$uniform sub-hypergraph stochastic block model ($m-$ShSBM). Under the $m-$ShSBM, we use Fano's inequality to identify the region of model parameters where any algorithm fails to exactly recover the planted communities with a large probability. We also identify the region where a Maximum Likelihood Estimation (MLE) algorithm succeeds to exactly recover the communities with high probability. Our bounds are tight and pertain to the community detection problems in various models such as the planted hypergraph stochastic block model, the planted densest sub-hypergraph model, and the planted multipartite hypergraph model. | ['Jean Honorio', 'Chuyang Ke', 'Jiajun Liang'] | 2021-01-29 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 4.63885516e-01 4.52449709e-01 -5.40559828e-01 3.05317074e-01
-7.55234063e-01 -5.69475234e-01 -5.70504367e-02 1.87152833e-01
9.34700891e-02 6.97417676e-01 -4.26049858e-01 -5.77777863e-01
-5.13186038e-01 -1.11378610e+00 -6.72025442e-01 -9.12750244e-01
-6.88765943e-01 8.75459790e-01 3.63898069e-01 1.27603874e-01
1.82140902e-01 1.27854005e-01 -6.19622886e-01 -1.27507761e-01
7.86940813e-01 5.88732101e-02 2.89966375e-01 1.08063185e+00
1.57083854e-01 3.02731007e-01 -2.45723814e-01 -3.01487625e-01
5.18438876e-01 -5.95563948e-01 -8.47577333e-01 4.49483275e-01
5.29043861e-02 -7.89211690e-02 -8.09007764e-01 1.51390445e+00
1.84717342e-01 -6.62412286e-01 8.40152025e-01 -1.59815884e+00
-2.52275437e-01 1.44165301e+00 -1.64137292e+00 2.34526426e-01
4.43183064e-01 -2.13131741e-01 1.13745975e+00 -4.43407208e-01
7.72430360e-01 1.38723016e+00 8.09510231e-01 -4.15132120e-02
-1.64505792e+00 -9.67672229e-01 -8.49396661e-02 -8.42658505e-02
-2.04497313e+00 -5.08603416e-02 2.46753201e-01 -5.52973926e-01
5.29404998e-01 2.31068581e-01 4.42509174e-01 4.02106494e-01
4.23816182e-02 6.80836976e-01 1.14224994e+00 -8.39274883e-01
1.89731300e-01 -3.63178849e-02 4.63159323e-01 9.23291266e-01
1.39827335e+00 1.33650780e-01 -1.73580036e-01 -6.95842743e-01
8.73570800e-01 -2.14796573e-01 -5.21797180e-01 -3.26091290e-01
-9.18213725e-01 9.35550570e-01 2.45639980e-01 2.56735355e-01
-2.08511770e-01 4.78963912e-01 1.43459653e-02 3.95864069e-01
2.07026899e-01 -1.04915686e-01 5.02925254e-02 5.56114912e-01
-1.08480394e+00 -9.16895792e-02 9.81067359e-01 1.57590020e+00
8.99355531e-01 -2.12210938e-01 1.03987031e-01 3.81730944e-01
4.07265276e-01 1.07908487e+00 -7.19883263e-01 -8.65231216e-01
7.36714542e-01 3.86900753e-01 2.00133950e-01 -1.04979813e+00
-3.15622166e-02 -5.14980018e-01 -1.31682289e+00 -2.62579441e-01
3.29440415e-01 2.17596740e-02 -8.92392576e-01 1.78515553e+00
1.85075864e-01 3.58832717e-01 -3.99608277e-02 3.60781282e-01
2.20456660e-01 5.14030635e-01 -3.73884618e-01 -6.69926822e-01
1.12825465e+00 -4.84289229e-01 -4.09992576e-01 -2.82237202e-01
4.92308170e-01 -4.22302663e-01 2.38928840e-01 8.19683373e-02
-1.16463161e+00 1.81743294e-01 -9.58636403e-01 8.39270532e-01
2.35575423e-01 -4.97978956e-01 2.13411361e-01 9.54086840e-01
-1.26803124e+00 1.35368422e-01 -6.72148943e-01 -6.66818559e-01
1.65996984e-01 3.28137338e-01 -3.20528030e-01 -8.50295842e-01
-7.80925572e-01 2.64848620e-01 7.04316139e-01 -9.16374922e-02
-1.18897009e+00 -1.26414999e-01 -5.64688206e-01 2.24426195e-01
5.95615566e-01 -6.82436466e-01 5.43121219e-01 -4.70920205e-01
-1.25038341e-01 1.02484095e+00 -3.73163134e-01 -5.52895546e-01
8.49567950e-02 6.55186772e-01 9.76128131e-02 6.28546655e-01
4.21032518e-01 -3.85315754e-02 4.60212290e-01 -1.49717033e+00
-3.54102522e-01 -3.63155931e-01 1.16140880e-01 -2.07428306e-01
1.37047067e-01 1.18350359e-02 -2.84414649e-01 -1.53006434e-01
7.96741664e-01 -1.20700669e+00 -7.14617074e-01 -6.96535408e-01
-1.06272495e+00 2.97038376e-01 4.37718183e-01 -2.57089347e-01
1.41550136e+00 -1.99212050e+00 1.48539990e-01 9.69955802e-01
8.10807645e-01 -1.01905778e-01 -3.33081603e-01 1.09101903e+00
-9.00357291e-02 5.93543589e-01 -3.78341258e-01 5.81295490e-02
-1.01357765e-01 1.92150902e-02 5.78434654e-02 8.50699246e-01
-5.15064716e-01 4.64166135e-01 -8.46458733e-01 -4.77040082e-01
-3.32979470e-01 -2.34598398e-01 -5.08334756e-01 -1.68118805e-01
1.83860004e-01 -4.15529497e-02 -5.70909560e-01 5.64348042e-01
1.25894666e+00 -9.03090477e-01 7.98454583e-01 5.48738182e-01
2.48017237e-01 -4.08393592e-01 -1.59593010e+00 8.06932867e-01
2.14893177e-01 4.30078655e-01 7.61760890e-01 -8.18227470e-01
4.33538288e-01 5.71772993e-01 4.14446354e-01 2.73382932e-01
-1.56483486e-01 2.99841404e-01 1.11878648e-01 -1.84901536e-01
2.24511176e-01 -1.58924714e-01 -1.69009596e-01 9.54034925e-01
-2.52937645e-01 3.30325484e-01 2.42999285e-01 1.10550332e+00
1.76908302e+00 -9.42281961e-01 5.24202883e-01 -7.13063240e-01
2.12257113e-02 1.64086558e-02 5.14680862e-01 1.60230577e+00
-2.24066108e-01 4.68658715e-01 8.24942410e-01 3.87963653e-01
-1.19429564e+00 -1.18095708e+00 1.88669432e-02 4.33689594e-01
5.56958437e-01 -5.86641610e-01 -7.66018927e-01 -1.33681357e-01
2.38015912e-02 1.33991256e-01 -4.14431781e-01 -7.99891800e-02
7.46009499e-02 -1.32616818e+00 4.46971893e-01 3.56077962e-02
4.94619519e-01 -5.16524911e-01 1.13890603e-01 2.49429420e-01
-6.16012096e-01 -1.14407301e+00 -6.04097486e-01 2.29703188e-01
-8.37320864e-01 -1.55087340e+00 -5.36694527e-01 -8.75503004e-01
9.04696047e-01 9.38954353e-01 9.75610852e-01 3.86250108e-01
-2.89928436e-01 4.00585502e-01 -3.79268348e-01 3.19797933e-01
-5.74732244e-01 1.57125786e-01 -9.01490226e-02 -3.35561126e-01
1.83843434e-01 -6.07038736e-01 -4.97619212e-01 2.83067435e-01
-6.73370063e-01 -1.92244090e-02 3.88313770e-01 6.06752455e-01
4.24319059e-01 7.00735569e-01 2.63810635e-01 -1.11537445e+00
3.32085401e-01 -9.13750052e-01 -7.24914908e-01 4.02978748e-01
-7.24251986e-01 -1.42922491e-01 -7.12125674e-02 -7.75616169e-02
-3.91616225e-01 -1.14514343e-02 3.58887374e-01 -2.67910153e-01
2.23468617e-01 8.40630770e-01 -2.68177807e-01 -2.21700057e-01
4.43397790e-01 1.18894041e-01 -2.58669049e-01 -2.05485672e-01
4.08687666e-02 6.74875557e-01 1.97636515e-01 -5.00275373e-01
1.19948506e+00 5.86496413e-01 4.45964247e-01 -9.07606900e-01
-4.64453340e-01 -9.01726782e-01 -5.60329914e-01 -6.14111200e-02
4.36984420e-01 -1.07704723e+00 -6.42544508e-01 4.15691853e-01
-1.02091563e+00 -4.04719561e-01 3.08530450e-01 2.82785177e-01
-5.01623452e-01 7.74482489e-01 -9.46237028e-01 -1.49504173e+00
-3.70709486e-02 -9.34656858e-01 8.84406388e-01 -1.88108847e-01
1.68269321e-01 -7.47689128e-01 2.15486869e-01 1.06566362e-01
-1.33379743e-01 5.30455589e-01 1.10140395e+00 -3.95553350e-01
-1.20522118e+00 -2.92124867e-01 -6.71198428e-01 -4.46195334e-01
-2.46262386e-01 -3.97960134e-02 -3.10465991e-01 -7.72437513e-01
-5.03630280e-01 2.85154879e-02 1.05714190e+00 8.49624634e-01
6.67951882e-01 -4.39915538e-01 -1.19273651e+00 4.74554658e-01
1.91722775e+00 6.56253099e-02 8.81303668e-01 3.35907526e-02
4.04074043e-01 2.74972320e-01 2.01856285e-01 3.65741521e-01
2.64635921e-01 1.31787375e-01 3.97149235e-01 9.66050252e-02
2.80168504e-01 -2.98825234e-01 5.99673092e-02 8.11046898e-01
1.71417832e-01 -9.46781278e-01 -9.51726317e-01 7.89058805e-01
-1.93954360e+00 -1.26782334e+00 -1.01081491e+00 2.47019219e+00
6.24882877e-01 1.36125803e-01 3.15277755e-01 1.53776988e-01
1.52839327e+00 -1.52813746e-02 -9.84808654e-02 9.76065919e-02
-2.70535469e-01 1.65806428e-01 1.00974214e+00 8.64151120e-01
-8.07632267e-01 7.29246378e-01 7.59580946e+00 9.66540277e-01
1.26587197e-01 2.37872243e-01 5.59985340e-01 2.37455875e-01
-4.17788386e-01 4.51461226e-01 -8.35084677e-01 2.58741200e-01
9.93011773e-01 -4.63331074e-01 4.37548339e-01 5.80055356e-01
-3.95313837e-02 -4.26249892e-01 -7.05400884e-01 8.58037531e-01
-2.62574792e-01 -1.04549587e+00 -3.64301205e-01 8.57840538e-01
1.19465613e+00 -1.51513526e-02 -4.09752071e-01 -8.67004618e-02
1.11344194e+00 -8.13378036e-01 4.17435169e-02 1.17765822e-01
1.02408206e+00 -8.09489906e-01 5.13071954e-01 7.16616631e-01
-1.64578736e+00 3.71273831e-02 -5.17533898e-01 2.01982826e-01
1.59098372e-01 9.34783459e-01 -8.29882562e-01 5.25118649e-01
4.65961725e-01 2.03383744e-01 -5.18633910e-02 1.41784298e+00
2.26135805e-01 8.56738567e-01 -6.53461814e-01 2.13070780e-01
1.02513857e-01 -4.21839267e-01 9.00761127e-01 1.11453676e+00
4.57473010e-01 4.17570055e-01 5.13618112e-01 7.06888914e-01
-1.40953928e-01 -2.26013228e-01 -8.53967071e-01 -2.61881411e-01
8.23977411e-01 7.74764478e-01 -1.14086080e+00 -2.53027916e-01
1.93869844e-02 8.25725436e-01 1.48740664e-01 4.69266474e-01
-4.71206486e-01 -5.02882898e-01 3.00993770e-01 4.93454069e-01
5.77635407e-01 -2.53041595e-01 -7.39712045e-02 -1.08332455e+00
-4.66697603e-01 -7.84472764e-01 5.11465192e-01 -4.56742316e-01
-1.32190907e+00 2.77976185e-01 2.66653150e-01 -6.06043458e-01
-6.96161911e-02 -6.36258349e-02 -6.87519670e-01 8.00645769e-01
-7.54561603e-01 -7.23226845e-01 5.40746488e-02 3.39785129e-01
-3.79850417e-01 1.32807657e-01 5.31473100e-01 6.13290854e-02
-4.97995436e-01 3.01884532e-01 5.48263371e-01 1.32634372e-01
7.66116828e-02 -1.21732271e+00 2.13313818e-01 1.27192092e+00
-9.84182507e-02 5.55554390e-01 9.15677726e-01 -9.72160339e-01
-1.32413852e+00 -1.08174336e+00 8.65227222e-01 1.44608676e-01
6.10953808e-01 -5.28121114e-01 -6.42369509e-01 8.42892408e-01
1.64023802e-01 -2.40875870e-01 5.70924819e-01 2.95122415e-01
-5.70424199e-01 1.14568919e-01 -1.41921806e+00 3.69106770e-01
1.36961257e+00 -4.37197536e-01 2.09061787e-01 7.48496592e-01
6.02514565e-01 1.79302171e-01 -6.83533192e-01 3.76540124e-01
3.86992753e-01 -7.52675414e-01 8.63969028e-01 -2.16544271e-01
1.68343216e-01 -2.15695411e-01 -5.08072197e-01 -9.58258748e-01
-7.85859108e-01 -7.33454049e-01 4.46988344e-02 9.48765099e-01
4.36838210e-01 -6.55373216e-01 9.53660011e-01 -4.04230542e-02
6.44570649e-01 -2.11511061e-01 -1.04365265e+00 -8.47601056e-01
-5.05053662e-02 2.19397973e-02 3.77505243e-01 8.58053446e-01
2.59025097e-01 2.75339246e-01 -4.23038960e-01 6.93437994e-01
1.44709837e+00 3.95978987e-01 7.40660906e-01 -1.32260966e+00
-7.00688124e-01 -3.78341556e-01 -3.94413561e-01 -1.06119180e+00
1.38856500e-01 -9.85917449e-01 1.86038151e-01 -1.63150883e+00
1.45125353e+00 -5.97163856e-01 -5.03880270e-02 3.39319222e-02
-2.54779547e-01 3.33839893e-01 1.89865362e-02 3.22260141e-01
-8.18636477e-01 -1.97693199e-01 8.34958136e-01 -1.69464096e-01
1.60708964e-01 2.59172201e-01 -8.56268167e-01 3.57801437e-01
6.54299438e-01 -8.99402916e-01 -3.19188625e-01 -4.30060886e-02
3.54847372e-01 8.88051093e-01 3.97991776e-01 -7.16053188e-01
2.09359348e-01 -2.22330973e-01 -1.75322846e-01 -9.17387962e-01
1.16388090e-01 -6.26323819e-01 6.80394351e-01 1.02203727e+00
-1.26404494e-01 -1.30633131e-01 -2.67006516e-01 1.27993286e+00
2.31599897e-01 -6.09481633e-01 9.41032290e-01 -2.25995854e-01
-2.61332467e-02 6.57992125e-01 -6.11085176e-01 1.84219986e-01
1.21246731e+00 -4.51825500e-01 -4.80755836e-01 -8.92031968e-01
-8.23637605e-01 4.84811604e-01 5.91560483e-01 -5.62529564e-01
3.78031671e-01 -9.47180092e-01 -1.02613187e+00 -2.63075978e-02
1.57446638e-01 -3.26163918e-01 2.67463207e-01 8.39075387e-01
-4.66257900e-01 3.04300457e-01 4.98176485e-01 -8.86927545e-01
-1.54299843e+00 6.26630366e-01 3.18429679e-01 -5.80740035e-01
-3.39755535e-01 7.31179059e-01 4.21025068e-01 -8.93625990e-03
-3.77004617e-03 3.53150010e-01 6.98516071e-01 -6.46526217e-01
3.49849880e-01 5.90353727e-01 -5.56176722e-01 -6.26396179e-01
-3.04067880e-01 3.24085325e-01 -3.85458022e-02 -3.84239465e-01
1.02806258e+00 -5.36435127e-01 -6.07014894e-01 7.20035285e-02
1.01368988e+00 1.69066280e-01 -7.00657189e-01 -4.34339166e-01
5.89507557e-02 -6.99790120e-01 -1.10744730e-01 -3.17121238e-01
-1.00726485e+00 5.03622472e-01 1.03147171e-01 7.60378897e-01
8.81308794e-01 4.41512883e-01 2.19925389e-01 1.59419924e-01
1.13171160e+00 -4.74480778e-01 -2.39659175e-01 4.10098612e-01
3.96851480e-01 -8.51514578e-01 1.05946414e-01 -9.93273497e-01
-8.01789761e-02 7.95548201e-01 4.13294196e-01 -2.18224645e-01
1.06149995e+00 5.84332883e-01 -9.36091721e-01 -4.19104040e-01
-7.38196075e-01 -4.64735270e-01 -4.20607775e-01 7.42142141e-01
-5.92061430e-02 5.89083135e-01 -1.78802729e-01 1.62658036e-01
1.98160961e-01 -4.09630567e-01 8.64290178e-01 5.19741058e-01
-1.01623833e+00 -1.13379014e+00 -7.06184208e-01 6.88905418e-01
-2.31664121e-01 -3.96727681e-01 -4.81179953e-01 7.26025045e-01
-2.35737309e-01 1.32627428e+00 -9.76422802e-02 -3.38715345e-01
-2.92492479e-01 -2.19981909e-01 7.19759762e-01 -8.43129814e-01
-7.04701617e-02 5.16834855e-01 -5.42171188e-02 4.87238765e-02
-3.56211096e-01 -6.16330862e-01 -6.86615288e-01 -1.04605746e+00
-9.04671252e-01 4.21318412e-01 -9.16276276e-02 6.67636871e-01
5.26541136e-02 -1.48498237e-01 8.66141140e-01 -2.58113816e-02
-4.37346339e-01 -9.87523496e-01 -1.42590952e+00 -3.85161757e-01
1.51987851e-01 -2.28867486e-01 -6.48733556e-01 -2.38215730e-01] | [6.8741912841796875, 5.109037399291992] |
7731c757-803d-45af-8bc3-7bbf0c88aa19 | mnemonic-descent-method-a-recurrent-process-1 | null | null | https://www.ibug.doc.ic.ac.uk/media/uploads/documents/trigeorgis2016mnemonic.pdf | https://www.ibug.doc.ic.ac.uk/media/uploads/documents/trigeorgis2016mnemonic.pdf | Mnemonic Descent Method: A recurrent process applied for end-to-end face alignment | Cascaded regression has recently become the method of choice for solving non-linear least squares problems such as deformable image alignment. Given a sizeable training set, cascaded regression learns a set of generic rules that are sequentially applied to minimise the least squares problem. Despite the success of cascaded regression for problems such as face alignment and head pose estimation, there are several shortcomings arising in the strategies proposed thus far. Specifically,(a) the regressors are learnt independently,(b) the descent directions may cancel one another out and (c) handcrafted features (eg, HoGs, SIFT etc.) are mainly used to drive the cascade, which may be sub-optimal for the task at hand. In this paper, we propose a combined and jointly trained convolutional recurrent neural network architecture that allows the training of an end-to-end to system that attempts to alleviate the aforementioned drawbacks. The recurrent module facilitates the joint optimisation of the regressors by assuming the cascades form a nonlinear dynamical system, in effect fully utilising the information between all cascade levels by introducing a memory unit that shares information across all levels. The convolutional module allows the network to extract features that are specialised for the task at hand and are experimentally shown to outperform hand-crafted features. We show that the application of the proposed architecture for the problem of face alignment results in a strong improvement over the current state-of-the-art. | ['P. Snape', 'E. Antonakos', 'S. Zafeiriou', 'M. A. Nicolaou', 'G. Trigeorgis'] | 2016-06-01 | null | null | null | cvpr-2016-6 | ['head-pose-estimation'] | ['computer-vision'] | [ 8.65965188e-02 2.81522661e-01 6.35323487e-03 -3.64072293e-01
-4.31787997e-01 -1.78455308e-01 6.81373894e-01 -2.21614823e-01
-6.08363330e-01 3.85959864e-01 6.10285960e-02 -6.11679964e-02
-7.63444304e-02 -2.71566361e-01 -8.56121123e-01 -7.84507751e-01
-1.14886329e-01 3.89506668e-01 -5.24707437e-02 -5.66568255e-01
8.75990689e-02 7.51986682e-01 -1.74363863e+00 9.18929800e-02
3.90488774e-01 1.01685226e+00 1.05311252e-01 7.33637750e-01
2.65657514e-01 9.13819313e-01 -2.06656411e-01 -3.94364953e-01
4.41383928e-01 -5.01520991e-01 -6.68367803e-01 1.64930850e-01
7.05865800e-01 -2.18563378e-01 -1.58910349e-01 5.81021547e-01
6.58119082e-01 3.61679345e-02 6.07331932e-01 -1.03061926e+00
-1.39863521e-01 1.47202447e-01 -4.91026461e-01 1.77464232e-01
1.67444482e-01 -8.26020725e-03 8.67837548e-01 -1.08372045e+00
4.89807338e-01 1.22392702e+00 7.38259733e-01 8.34180236e-01
-1.20301712e+00 -5.20197570e-01 1.16745643e-01 6.21633679e-02
-1.34654331e+00 -8.82299483e-01 6.89941406e-01 -3.34833264e-01
1.19376159e+00 1.83671936e-01 6.22606933e-01 8.96099925e-01
2.42692620e-01 5.03932416e-01 8.89635146e-01 -4.80027258e-01
5.85491676e-03 2.76261181e-01 -3.35772455e-01 8.76970470e-01
-4.59083803e-02 -3.71132256e-03 -6.05974257e-01 3.91517505e-02
8.53790283e-01 -1.49370313e-01 5.25159463e-02 -5.60939968e-01
-7.26473808e-01 8.93781960e-01 5.82047403e-01 4.39629018e-01
-5.25870323e-01 6.10725246e-02 2.81820208e-01 4.12992805e-01
6.87593758e-01 2.73320854e-01 -4.98515368e-01 3.82287323e-01
-1.23396111e+00 3.35004956e-01 8.11226070e-01 5.18536270e-01
7.90106356e-01 1.65661037e-01 7.94657171e-02 7.85327196e-01
5.44763029e-01 1.80897847e-01 5.25303364e-01 -5.75949192e-01
4.87289071e-01 6.67598367e-01 -2.52796393e-02 -1.08242607e+00
-6.87916756e-01 -4.39654142e-01 -8.86454880e-01 4.32214081e-01
4.73929584e-01 -3.33773017e-01 -8.66732001e-01 1.84756076e+00
5.26933849e-01 2.56879985e-01 2.29967311e-02 1.01359391e+00
6.40949368e-01 3.53006959e-01 1.27436956e-02 -2.78514564e-01
1.09742928e+00 -1.01619947e+00 -4.70709145e-01 -4.50541258e-01
4.81046170e-01 -7.70846844e-01 4.48991507e-01 1.16767764e-01
-1.29739881e+00 -6.30605042e-01 -1.01499844e+00 -1.06547937e-01
-2.71892160e-01 1.96033448e-01 3.42939645e-01 3.72278780e-01
-1.47273099e+00 5.78226507e-01 -8.29948127e-01 -3.12998205e-01
2.48479292e-01 9.83849943e-01 -4.64674532e-01 3.98628116e-01
-9.79720056e-01 1.26566112e+00 8.83960426e-02 7.32841313e-01
-7.16481268e-01 -4.44964886e-01 -9.30910528e-01 -8.16915259e-02
2.11545348e-01 -7.60619879e-01 1.04234910e+00 -1.32515156e+00
-1.92515397e+00 1.00296450e+00 -2.17425525e-01 -5.25789440e-01
8.35695803e-01 -3.88787240e-01 6.69935066e-03 -3.70336138e-02
-2.70427138e-01 7.19106972e-01 1.25404632e+00 -7.93387949e-01
-4.88475859e-01 -3.95271152e-01 -9.55990553e-02 4.00808334e-01
-5.07042468e-01 3.28969747e-01 -5.18127739e-01 -5.93352854e-01
-1.20634452e-01 -1.22956812e+00 -4.93196696e-01 -1.24375895e-01
-2.24293903e-01 -1.68935612e-01 8.52006614e-01 -6.33484006e-01
1.06748056e+00 -1.98444259e+00 7.95171201e-01 2.15639576e-01
-4.10694480e-02 3.90267015e-01 -1.05937876e-01 2.83774376e-01
-4.24302995e-01 -3.29436243e-01 -1.15056098e-01 -6.77449286e-01
-3.44705433e-01 1.09663107e-01 -3.42295691e-02 7.31231928e-01
5.51464677e-01 8.11227560e-01 -4.09144759e-01 -3.57221514e-01
2.87822366e-01 9.23567295e-01 -5.49012125e-01 4.21138465e-01
-9.55521837e-02 8.09676230e-01 -2.50726670e-01 3.88731956e-02
4.30094063e-01 2.22478267e-02 1.06068596e-01 -1.40649319e-01
-3.05782646e-01 1.11357339e-01 -1.40978634e+00 1.59084225e+00
-6.11417234e-01 4.97541755e-01 1.09853171e-01 -1.21967518e+00
8.85473967e-01 4.96688098e-01 5.54366469e-01 -4.44198072e-01
2.08837986e-01 1.17091671e-01 4.33938242e-02 -4.68150735e-01
1.15350522e-01 -1.48451552e-01 1.27506435e-01 3.29848230e-01
4.45638686e-01 -6.36165440e-02 -1.31319880e-01 -3.16877991e-01
7.66364574e-01 3.38955522e-01 3.62432659e-01 -2.62934029e-01
1.08223486e+00 -2.65565902e-01 4.30440843e-01 2.53627956e-01
2.61266261e-01 5.92964172e-01 3.74777704e-01 -6.33114278e-01
-1.21284854e+00 -2.80002564e-01 6.27187416e-02 1.28911948e+00
-2.92454720e-01 -6.23634681e-02 -8.86593699e-01 -4.77405071e-01
-1.17936976e-01 3.41583565e-02 -7.50160098e-01 1.82115883e-02
-1.01088941e+00 -7.15613663e-01 4.40152258e-01 4.36321467e-01
4.06811237e-01 -1.03342736e+00 -8.71902704e-01 3.18880737e-01
2.38330513e-01 -1.20999312e+00 -3.25006276e-01 2.89116561e-01
-9.01512921e-01 -8.71683061e-01 -7.96965897e-01 -7.84900486e-01
8.16420496e-01 5.16040763e-03 8.03155899e-01 1.30294740e-01
-3.16244870e-01 2.79824167e-01 3.89240086e-02 -3.87067497e-01
-3.25110048e-01 3.56682271e-01 1.24936536e-01 3.07044685e-01
1.21351262e-03 -4.11134660e-01 -6.72060668e-01 2.93300062e-01
-6.74000263e-01 -9.68978703e-02 6.75160766e-01 9.00092840e-01
3.51822704e-01 -2.31392726e-01 3.32789063e-01 -8.27774644e-01
3.87205392e-01 -2.09152773e-01 -7.10215867e-01 1.53783336e-01
-3.76124352e-01 3.59215826e-01 6.56873882e-01 -3.96026462e-01
-9.88994658e-01 5.64208150e-01 -2.17221558e-01 -4.02050495e-01
-8.59483406e-02 2.77028501e-01 -9.11875516e-02 -5.72596848e-01
3.95689428e-01 -3.85933556e-02 2.84987986e-01 -3.90456349e-01
3.31112176e-01 4.19078887e-01 3.81201267e-01 -8.60579237e-02
8.70323181e-01 3.77559245e-01 3.12370747e-01 -9.33306754e-01
-6.24912977e-01 -4.49993670e-01 -1.09356415e+00 -3.70204240e-01
8.33399773e-01 -1.05925643e+00 -7.61445463e-01 7.02515781e-01
-9.94122028e-01 -3.09140712e-01 -8.57103392e-02 3.15169662e-01
-7.21720934e-01 9.58738327e-02 -5.28044581e-01 -8.47550988e-01
-4.86655593e-01 -9.69998598e-01 1.00931740e+00 4.66785997e-01
-3.33696157e-01 -1.10242116e+00 6.16963990e-02 1.39735982e-01
4.45230782e-01 3.01189274e-01 6.22803271e-01 -6.68668985e-01
-2.54995018e-01 -2.82237232e-01 2.01670587e-01 4.56507921e-01
-4.53271270e-02 6.32688478e-02 -1.12467873e+00 -5.26972592e-01
2.35202402e-01 -2.98543602e-01 6.82939053e-01 3.14647526e-01
6.10200644e-01 -5.33002198e-01 -1.20098814e-01 5.06376266e-01
1.22640026e+00 -2.96249539e-01 4.96368259e-01 3.24856192e-01
7.36144364e-01 9.17461455e-01 2.41738394e-01 2.02757537e-01
4.11580890e-01 1.05621743e+00 4.63368624e-01 -4.94541377e-01
-2.50600614e-02 1.55434050e-02 5.46608448e-01 7.11702168e-01
-4.82447535e-01 3.59149009e-01 -6.96763158e-01 4.43898916e-01
-2.01103616e+00 -8.02810729e-01 3.88005264e-02 2.31839967e+00
6.59588754e-01 3.22843827e-02 4.56886053e-01 2.00814381e-01
4.90692079e-01 -3.37489089e-03 -4.57532585e-01 -7.01996744e-01
-1.51544642e-02 4.87150103e-01 4.64435875e-01 5.31099856e-01
-1.24518561e+00 1.05448556e+00 6.01681280e+00 2.75921255e-01
-1.55142748e+00 -5.31570464e-02 3.51077288e-01 -2.17982326e-02
3.73998076e-01 -2.14488298e-01 -1.01118624e+00 1.42487511e-01
9.21249866e-01 3.53719413e-01 3.86426568e-01 6.40528917e-01
2.62839019e-01 1.43186525e-01 -1.17202115e+00 6.62793040e-01
2.29724661e-01 -7.73778975e-01 -1.80254474e-01 9.21951905e-02
6.78830266e-01 3.17733027e-02 4.08924460e-01 4.74186167e-02
4.38463353e-02 -1.15640771e+00 7.14077115e-01 5.52251458e-01
4.10573244e-01 -8.40494037e-01 6.59113765e-01 4.12137866e-01
-9.90971565e-01 -2.10765526e-01 -2.74848312e-01 -1.19848073e-01
-1.41597018e-01 2.76162386e-01 -1.01781487e+00 4.86044168e-01
5.59981287e-01 5.70037067e-01 -5.26183665e-01 8.83733213e-01
-3.06588262e-01 2.48949096e-01 -5.03454804e-01 1.18798673e-01
4.64149445e-01 -2.00535934e-02 5.26107013e-01 1.33987546e+00
2.50534825e-02 -2.38344923e-01 -2.16455802e-01 4.55651313e-01
1.54601887e-01 2.69309193e-01 -5.32922328e-01 4.04186934e-01
-1.90339833e-01 1.46267033e+00 -5.64129055e-01 2.60551274e-02
-5.48619270e-01 1.01295948e+00 5.95799506e-01 3.56243998e-01
-6.75349474e-01 -8.83024558e-03 7.04799473e-01 -1.03807950e-03
7.71657288e-01 -3.06110471e-01 -7.43608028e-02 -1.12740469e+00
-4.08819765e-02 -9.22213554e-01 3.31775337e-01 -3.20878625e-01
-9.94823217e-01 7.36793637e-01 -1.69757292e-01 -9.85177755e-01
-8.43308389e-01 -6.60116255e-01 -5.36939979e-01 9.81170952e-01
-1.51586354e+00 -1.38254642e+00 -5.82256652e-02 7.00265229e-01
5.25443256e-01 -2.21381962e-01 9.27459061e-01 2.09978372e-01
-7.83047974e-01 5.81001461e-01 -2.13021427e-01 8.81778598e-02
5.77268958e-01 -1.03421164e+00 2.75179684e-01 8.57009530e-01
2.48930380e-01 5.50990105e-01 9.56441581e-01 -3.25564384e-01
-1.30159521e+00 -9.10820484e-01 9.44644272e-01 -2.29330018e-01
4.66944456e-01 -4.66712505e-01 -8.58540833e-01 5.68929076e-01
1.07268445e-01 1.59456119e-01 4.58202481e-01 -6.13740422e-02
-2.14493111e-01 -2.55758584e-01 -9.57275152e-01 4.77364928e-01
7.43826210e-01 -4.94621992e-01 -3.84600848e-01 1.26316950e-01
-7.19586983e-02 -6.72898293e-01 -7.54259169e-01 4.38009113e-01
7.29034305e-01 -9.15480256e-01 9.99395549e-01 -8.22455168e-01
4.10942525e-01 -2.40638524e-01 2.71551430e-01 -1.20243335e+00
-1.32371768e-01 -7.35438287e-01 -1.89274192e-01 1.18836153e+00
5.40924430e-01 -5.18780947e-01 8.43503654e-01 7.22519696e-01
3.56003314e-01 -8.77170742e-01 -1.10898423e+00 -3.27321261e-01
-6.51633069e-02 9.18699279e-02 2.74212748e-01 6.77465498e-01
-2.00202093e-01 6.31612659e-01 -7.09452152e-01 1.79431841e-01
4.06278342e-01 -1.50611714e-01 8.68090987e-01 -1.12750685e+00
-2.38925129e-01 -3.22433680e-01 -6.48536742e-01 -9.66311157e-01
3.77892107e-01 -6.97844744e-01 5.13539463e-02 -9.88786221e-01
-6.31604716e-02 -3.35895747e-01 -1.81201890e-01 6.49539888e-01
-2.08160535e-01 2.58858383e-01 2.23130122e-01 2.57070363e-01
-2.89195836e-01 2.89747834e-01 1.03497255e+00 1.31194279e-01
-4.05214041e-01 3.39006007e-01 -5.97275317e-01 7.91112423e-01
6.72572076e-01 -4.06781375e-01 -1.51430532e-01 -4.87535745e-01
3.32826108e-01 -1.20479754e-05 4.40869391e-01 -7.94082463e-01
5.43416083e-01 2.88761824e-01 5.69076061e-01 -4.69777808e-02
4.39860761e-01 -1.01893795e+00 5.61911389e-02 5.58848441e-01
-2.96381027e-01 1.74584806e-01 2.42192358e-01 2.56009877e-01
-2.14686185e-01 -2.02415287e-01 8.78374815e-01 8.72905646e-03
-3.86856019e-01 1.14480384e-01 -2.55050391e-01 -3.77408564e-01
9.99520898e-01 -1.45932570e-01 4.22484517e-01 -4.17578608e-01
-7.31369913e-01 8.10646564e-02 2.26293325e-01 4.95917410e-01
3.27086926e-01 -1.03345692e+00 -9.02842879e-01 3.37809205e-01
-3.30527455e-01 1.03404745e-01 1.12908857e-03 9.76323843e-01
-2.87516981e-01 4.76374865e-01 -4.64011014e-01 -6.19193554e-01
-1.48587155e+00 3.78541410e-01 6.41053736e-01 -3.78137141e-01
-4.89722431e-01 9.38959897e-01 -3.02604195e-02 -3.63766462e-01
2.10992068e-01 -9.16691720e-02 -6.71590507e-01 3.53580624e-01
3.31838369e-01 2.01637343e-01 3.46032798e-01 -1.16632128e+00
-3.61170143e-01 9.84777033e-01 -1.92086697e-01 -1.14605827e-02
1.86907935e+00 -1.29528224e-01 -2.11873740e-01 1.70396432e-01
1.19020379e+00 -2.05038443e-01 -1.33123088e+00 -2.25485891e-01
1.45356372e-01 6.52567446e-02 7.64532806e-03 -4.31281060e-01
-1.19546068e+00 8.55414808e-01 5.90211093e-01 -4.52568457e-02
1.23009574e+00 -2.09278852e-01 4.00501758e-01 2.88207263e-01
6.60146400e-02 -8.98367405e-01 -5.52875549e-02 4.93873775e-01
9.83696640e-01 -1.01903856e+00 1.12964012e-01 -2.24211827e-01
-5.64988732e-01 1.45648932e+00 4.44674790e-01 -4.77156371e-01
5.79663932e-01 4.46584523e-01 7.88519084e-02 -9.58625153e-02
-7.32973397e-01 -2.73991734e-01 5.88242471e-01 2.09025458e-01
5.59045255e-01 -4.19015020e-01 -3.11554104e-01 1.66930288e-01
-1.94417506e-01 3.54362987e-02 6.67104349e-02 7.12003946e-01
-2.04740092e-01 -1.20018268e+00 -3.92424345e-01 7.83801526e-02
-6.28950894e-01 2.64448494e-01 -5.14921069e-01 7.99521446e-01
2.05733716e-01 7.63434172e-01 -1.07369088e-01 -1.29800469e-01
5.31514525e-01 9.79724452e-02 7.24232495e-01 -4.72366899e-01
-9.61279452e-01 1.51020795e-01 -7.48683214e-02 -5.52443981e-01
-6.73581064e-01 -7.89469182e-01 -8.17641437e-01 5.85972220e-02
-5.18076360e-01 -1.44230619e-01 8.59001398e-01 1.19795907e+00
2.34370545e-01 3.87528300e-01 7.93133497e-01 -1.14815176e+00
-5.23520768e-01 -9.26995635e-01 -6.33921027e-02 2.37166777e-01
6.46780670e-01 -6.24817133e-01 -1.90275967e-01 1.47911355e-01] | [13.532952308654785, 0.30707037448883057] |
c0c426d7-6f9b-4c14-b429-be14114a1b3b | common-phone-a-multilingual-dataset-for | 2201.05912 | null | https://arxiv.org/abs/2201.05912v2 | https://arxiv.org/pdf/2201.05912v2.pdf | Common Phone: A Multilingual Dataset for Robust Acoustic Modelling | Current state of the art acoustic models can easily comprise more than 100 million parameters. This growing complexity demands larger training datasets to maintain a decent generalization of the final decision function. An ideal dataset is not necessarily large in size, but large with respect to the amount of unique speakers, utilized hardware and varying recording conditions. This enables a machine learning model to explore as much of the domain-specific input space as possible during parameter estimation. This work introduces Common Phone, a gender-balanced, multilingual corpus recorded from more than 11.000 contributors via Mozilla's Common Voice project. It comprises around 116 hours of speech enriched with automatically generated phonetic segmentation. A Wav2Vec 2.0 acoustic model was trained with the Common Phone to perform phonetic symbol recognition and validate the quality of the generated phonetic annotation. The architecture achieved a PER of 18.1 % on the entire test set, computed with all 101 unique phonetic symbols, showing slight differences between the individual languages. We conclude that Common Phone provides sufficient variability and reliable phonetic annotation to help bridging the gap between research and application of acoustic models. | ['Juan Rafael Orozco-Arroyave', 'Elmar Nöth', 'Paula Andrea Pérez-Toro', 'Tomás Arias-Vergara', 'Philipp Klumpp'] | 2022-01-15 | null | https://aclanthology.org/2022.lrec-1.81 | https://aclanthology.org/2022.lrec-1.81.pdf | lrec-2022-6 | ['acoustic-modelling'] | ['speech'] | [-1.83217779e-01 1.16652347e-01 -6.44180253e-02 -6.33096099e-01
-1.21787071e+00 -8.14190805e-01 3.21086049e-01 -2.46093258e-01
-4.25299972e-01 4.25148875e-01 9.45137665e-02 -4.85847682e-01
3.35382879e-01 -2.28818446e-01 -6.72932982e-01 -6.17284298e-01
4.60748896e-02 9.40278530e-01 -4.19020392e-02 -1.41096309e-01
-4.48558815e-02 2.06051469e-01 -1.69475770e+00 2.41613071e-02
5.72280109e-01 9.29402173e-01 6.49328947e-01 8.79815757e-01
-3.05627853e-01 -1.90922767e-02 -9.96304154e-01 -6.14609361e-01
1.31230429e-01 -1.00083746e-01 -5.49131751e-01 -2.16662392e-01
7.94259369e-01 5.50370067e-02 1.74954921e-01 8.94487679e-01
8.58549595e-01 2.04616010e-01 4.80916649e-01 -9.11857665e-01
-4.31363881e-01 1.37035263e+00 7.59608448e-02 2.42794663e-01
-5.52635454e-03 1.29370481e-01 1.21120107e+00 -1.08262110e+00
3.70661974e-01 1.27412498e+00 6.95421040e-01 6.93144441e-01
-1.08311391e+00 -8.97167265e-01 -2.93095242e-02 1.68671653e-01
-1.66609144e+00 -8.98571551e-01 6.36139095e-01 -5.98154247e-01
1.15785241e+00 4.39348370e-01 6.42871857e-01 1.39871705e+00
-2.37108782e-01 3.25110972e-01 8.08253944e-01 -6.77101731e-01
1.74319640e-01 5.56016207e-01 1.08423844e-01 1.75115034e-01
-5.94216138e-02 -1.82233304e-01 -5.80110848e-01 -4.30747829e-02
4.12881076e-01 -9.85749602e-01 7.44374329e-03 2.05128957e-02
-9.65894938e-01 7.45281398e-01 -3.69798422e-01 4.35526341e-01
2.40231119e-02 5.37240915e-02 7.32834756e-01 1.99088126e-01
2.00148314e-01 6.49567187e-01 -8.54018211e-01 -7.49184489e-01
-9.09875154e-01 9.15985778e-02 7.76038527e-01 8.57877493e-01
4.85864997e-01 7.58669674e-01 4.33647931e-01 1.71686816e+00
3.17802221e-01 9.39319015e-01 1.02783036e+00 -9.99922931e-01
5.18833995e-01 -3.79097015e-02 -3.55725884e-01 -6.55349731e-01
-1.25332415e-01 -6.06224835e-01 -4.59439546e-01 -1.73034191e-01
5.59011042e-01 -1.84213713e-01 -8.06621611e-01 1.80779588e+00
8.38964656e-02 6.99871927e-02 -4.26921323e-02 4.52586859e-01
7.14841723e-01 7.87238777e-01 1.71184376e-01 1.10061178e-02
1.29488337e+00 -8.22613180e-01 -6.40894175e-01 -4.38330710e-01
6.24504209e-01 -1.02941525e+00 1.61823249e+00 6.71624959e-01
-1.17719638e+00 -9.71896350e-01 -9.87906277e-01 6.12632856e-02
-4.96867865e-01 4.79442656e-01 3.95015687e-01 1.40274203e+00
-8.67058337e-01 1.05599165e-01 -5.84386468e-01 -1.49163693e-01
1.08404152e-01 4.38947231e-01 -2.92036206e-01 2.01676235e-01
-1.04544747e+00 8.03805113e-01 4.76352632e-01 -1.17832623e-01
-6.15041673e-01 -8.99259567e-01 -8.43404770e-01 1.43113375e-01
-9.74947512e-02 -8.03684369e-02 1.47498190e+00 -6.48035645e-01
-1.67014301e+00 8.21199417e-01 -9.55282971e-02 -5.06615162e-01
2.88032591e-01 -3.33197951e-01 -8.46217155e-01 -4.10302341e-01
-3.98274884e-02 5.43644190e-01 5.26794851e-01 -1.06614435e+00
-6.95859790e-01 -2.14053094e-01 -9.57097054e-01 1.07346423e-01
-4.36586469e-01 1.60521552e-01 -7.15158224e-01 -5.91697812e-01
-1.43528908e-01 -1.07717359e+00 -3.53278965e-02 -9.02432144e-01
-3.44792783e-01 -1.55366153e-01 4.44414407e-01 -1.05584598e+00
1.38255191e+00 -2.38492441e+00 -2.15130523e-01 3.50733429e-01
-5.10417402e-01 4.30788934e-01 -5.68468422e-02 1.24121219e-01
1.20303407e-01 2.66617298e-01 -2.32711315e-01 -6.53454304e-01
2.05447018e-01 4.21686530e-01 -4.52339113e-01 2.31151491e-01
-1.13911323e-01 5.41134953e-01 -4.05365407e-01 -3.25929999e-01
3.85270685e-01 5.47258377e-01 -5.04597247e-01 1.04022980e-01
1.44138597e-02 3.29311758e-01 1.99438989e-01 5.42506874e-01
6.87576234e-01 6.82329535e-01 1.46373764e-01 8.80229846e-02
-1.18177831e-01 5.94039321e-01 -1.30182016e+00 1.52023900e+00
-9.30921495e-01 8.15767050e-01 2.11973116e-01 -8.95513058e-01
1.31882417e+00 3.01809609e-01 2.58688003e-01 -6.43446624e-01
1.69632852e-01 7.27315068e-01 4.72017199e-01 -3.43097329e-01
6.06761694e-01 4.10250984e-02 -3.61031890e-01 1.46477465e-02
4.77573723e-01 -6.08249009e-01 1.34971201e-01 -3.98627847e-01
4.42994565e-01 -3.45907420e-01 2.11384818e-02 -3.85717958e-01
4.06772286e-01 -2.75375098e-01 5.22995830e-01 7.69070864e-01
-1.44748524e-01 6.97571635e-01 -4.81724106e-02 -1.43085405e-01
-1.29019713e+00 -1.21791422e+00 -6.11904204e-01 1.37874532e+00
-5.81061184e-01 -4.55758274e-01 -9.62154925e-01 -2.32777014e-01
-9.21658948e-02 9.75641012e-01 -1.84629828e-01 1.32595077e-01
-8.55066836e-01 -6.70167506e-01 1.14776528e+00 3.11937571e-01
3.30577157e-02 -1.09568250e+00 -9.54164490e-02 3.01721603e-01
-1.11447580e-01 -1.34459972e+00 -3.38791788e-01 2.76122957e-01
-3.64327013e-01 -6.71962798e-01 -5.23950338e-01 -1.03770244e+00
-7.75869340e-02 -4.86074328e-01 1.34116459e+00 -4.66482192e-01
-1.96691781e-01 4.95273098e-02 -1.43266797e-01 -7.23599434e-01
-9.62983549e-01 5.26938319e-01 4.23364341e-01 -1.57729104e-01
4.52204078e-01 -5.14342248e-01 1.74234822e-01 2.96785504e-01
-2.16225669e-01 -4.23464775e-01 4.21234399e-01 7.65681863e-01
7.67281473e-01 4.03568987e-03 9.52834070e-01 -6.26990438e-01
4.77476180e-01 -4.21692669e-01 -4.88298565e-01 -1.00244686e-01
-4.58151311e-01 -3.30787450e-01 6.86451674e-01 -5.91636598e-01
-1.02053571e+00 4.16032337e-02 -8.92156959e-01 -2.34936461e-01
-4.08223867e-01 2.90166020e-01 -5.36877751e-01 4.00976002e-01
6.45193338e-01 3.08457524e-01 -2.00623065e-01 -9.49989438e-01
4.78693247e-01 1.30157256e+00 9.51915383e-01 -5.83377600e-01
3.46193731e-01 -2.11553186e-01 -6.16491437e-01 -1.41303062e+00
-2.37609178e-01 -4.14478838e-01 -6.81823730e-01 -5.38891777e-02
5.54569185e-01 -8.81661713e-01 -5.58527470e-01 6.63254678e-01
-8.81086051e-01 -5.78863323e-01 -4.07962471e-01 7.85808802e-01
-4.35982943e-01 1.07656969e-02 -3.87631029e-01 -9.80111301e-01
-2.70855486e-01 -1.28935730e+00 8.61610055e-01 -1.24775209e-01
-5.72347045e-01 -9.32150245e-01 1.44154832e-01 4.94280726e-01
4.88352835e-01 -3.96743774e-01 8.58594120e-01 -1.01737237e+00
1.16772041e-01 -2.08618179e-01 4.65852380e-01 8.83022547e-01
6.62378147e-02 3.14031214e-01 -1.40713024e+00 -1.14089567e-02
-1.97142422e-01 -1.84104487e-01 5.54724336e-01 5.73900223e-01
1.31625986e+00 -9.57846716e-02 9.76784602e-02 5.40412128e-01
8.38865757e-01 4.65029657e-01 5.23277223e-01 1.49281770e-01
8.33510160e-01 5.70614934e-01 4.12063181e-01 1.64456233e-01
2.73409903e-01 9.39407229e-01 4.99465577e-02 2.62337625e-01
-2.66702563e-01 -2.89884895e-01 4.56816405e-01 1.67765009e+00
3.09566379e-01 -6.94032088e-02 -1.25980175e+00 8.54653239e-01
-9.87451732e-01 -6.64391279e-01 -9.21319351e-02 2.46378446e+00
1.11037815e+00 2.24810719e-01 2.24537045e-01 4.38848287e-01
6.99514031e-01 -1.18356317e-01 -1.50891721e-01 -1.17085218e+00
-3.35452020e-01 4.05103296e-01 4.13429677e-01 6.23697639e-01
-9.28725541e-01 1.24350512e+00 6.63359308e+00 1.29971218e+00
-1.30197644e+00 1.41678810e-01 6.51857138e-01 -2.30820253e-01
-1.73400536e-01 -5.24127662e-01 -1.21588171e+00 7.06949115e-01
1.71477449e+00 1.01966582e-01 3.88693631e-01 1.13872087e+00
6.72009960e-02 2.67156065e-01 -7.99949884e-01 1.08722794e+00
4.64713834e-02 -1.12431633e+00 -1.24156468e-01 2.00165018e-01
6.32476270e-01 4.22990113e-01 4.00729120e-01 6.56972170e-01
2.19455063e-01 -1.29929042e+00 1.06778300e+00 6.98218644e-02
1.06978273e+00 -1.03878427e+00 6.75318062e-01 3.31776947e-01
-1.05017531e+00 -1.57662705e-02 -5.34811318e-01 1.28379971e-01
2.90014237e-01 5.42617738e-01 -1.33575308e+00 1.08393461e-01
8.28273356e-01 1.38028070e-01 -6.60199761e-01 8.76773715e-01
2.08108738e-01 1.42317533e+00 -5.88712156e-01 6.71863928e-02
4.12211195e-02 -5.28211296e-02 4.79160964e-01 1.61706519e+00
6.06440365e-01 -7.22958207e-01 -4.76930961e-02 2.81503290e-01
6.60722777e-02 7.48624265e-01 -4.49095041e-01 -1.34023383e-01
1.09583902e+00 9.44908500e-01 -2.56697893e-01 -2.88859338e-01
-2.00300604e-01 5.00453591e-01 1.98601678e-01 7.81601444e-02
-6.91867411e-01 -2.80685037e-01 9.63519454e-01 -1.43655851e-01
3.02173704e-01 -4.02992427e-01 -5.52305162e-01 -5.44128239e-01
-4.44630086e-02 -1.10526621e+00 7.75693506e-02 -3.95686209e-01
-1.18037093e+00 8.24605405e-01 -2.62285560e-01 -7.94568539e-01
-6.87883973e-01 -7.06494868e-01 -3.84271294e-01 1.25516474e+00
-9.64620829e-01 -1.22860837e+00 1.37608275e-01 1.53375328e-01
8.82269084e-01 -7.61547744e-01 1.26896465e+00 6.71887934e-01
-6.44801676e-01 1.19731963e+00 2.05245391e-01 7.34302625e-02
8.12632978e-01 -1.36640918e+00 7.44821191e-01 5.04556239e-01
5.50277174e-01 5.53993404e-01 6.25303268e-01 -3.18636328e-01
-1.01096606e+00 -9.80883002e-01 1.09234917e+00 -6.57025278e-01
5.32064140e-01 -7.52647638e-01 -1.01204026e+00 6.68882132e-01
2.81042662e-02 -1.92360491e-01 1.06725919e+00 4.09343362e-01
-3.61947894e-01 -1.74241617e-01 -7.45637298e-01 6.54701352e-01
6.39777601e-01 -5.90586722e-01 -3.41796041e-01 -1.97996512e-01
7.31484473e-01 -3.87071401e-01 -9.33837831e-01 1.52006209e-01
7.66454160e-01 -6.31057322e-01 1.01703513e+00 -3.67839724e-01
-3.83815244e-02 1.38164774e-01 -6.37296319e-01 -1.45432997e+00
6.87918738e-02 -5.32083809e-01 1.19182825e-01 1.86753631e+00
8.91344666e-01 -4.39589024e-01 4.99723345e-01 2.14744225e-01
-7.60554314e-01 -5.91931343e-01 -1.23641026e+00 -1.01143980e+00
5.72920561e-01 -1.28577530e+00 8.96671295e-01 9.17755425e-01
-2.60503739e-01 1.90721691e-01 -3.44326824e-01 6.73320889e-02
1.61891669e-01 -5.75077236e-01 9.54036713e-01 -1.25219142e+00
-5.51260889e-01 -6.84865654e-01 -4.18635011e-01 -7.26622581e-01
4.13817495e-01 -9.62031126e-01 1.55329883e-01 -8.90636027e-01
-4.52280581e-01 -9.53385830e-01 -1.70863837e-01 3.31062406e-01
1.21617496e-01 4.74970698e-01 8.90994631e-03 -2.66728154e-03
1.36372119e-01 2.78247982e-01 8.37852836e-01 9.73424688e-03
-3.72959077e-01 2.63198704e-01 -5.21129608e-01 7.60081291e-01
1.17697310e+00 -2.33749121e-01 -3.75873953e-01 -4.15089011e-01
-2.12878644e-01 -4.05520201e-01 -1.63321748e-01 -1.10077190e+00
-1.40651986e-01 -1.12615794e-01 1.64009944e-01 -3.82329077e-01
7.35084951e-01 -5.17737985e-01 2.18759850e-01 7.91021958e-02
-3.74486268e-01 -1.61339417e-01 4.90688413e-01 -6.14467710e-02
-2.75272757e-01 -3.15820038e-01 7.62592077e-01 9.14700255e-02
-7.88452208e-01 -1.59233391e-01 -4.94480342e-01 2.89440125e-01
5.43057919e-01 -2.32896239e-01 -9.24656750e-04 -9.33483988e-02
-7.57237434e-01 -2.77768373e-01 1.89304724e-01 7.15850830e-01
-2.38673482e-02 -1.24373472e+00 -7.62245774e-01 5.96666396e-01
1.05121784e-01 -2.61089772e-01 4.25081253e-01 4.27275151e-01
-4.37128246e-01 4.78312463e-01 -6.13596700e-02 -6.93236887e-01
-1.36408412e+00 2.44960580e-02 3.48266184e-01 1.04164235e-01
-4.04568255e-01 1.25912392e+00 -7.63320401e-02 -8.52810621e-01
3.71157795e-01 -2.10347816e-01 -2.94947386e-01 2.39833593e-01
4.18772817e-01 4.11539406e-01 3.64679873e-01 -1.20446062e+00
-3.89778614e-01 4.35144067e-01 2.60474026e-01 -6.36351645e-01
9.55238998e-01 -2.82864571e-01 2.70333886e-01 9.28720236e-01
1.37499607e+00 6.79995716e-01 -9.06345129e-01 1.04359880e-01
-1.01434089e-01 -3.55808347e-01 -2.63845338e-03 -9.06159699e-01
-9.06879485e-01 1.23474407e+00 6.71940029e-01 1.56129867e-01
6.04377806e-01 4.42902092e-03 7.97670841e-01 2.15261281e-01
1.53122380e-01 -1.47809958e+00 -4.32297856e-01 7.12167382e-01
7.85073102e-01 -1.09920359e+00 -6.83818996e-01 -4.12922055e-01
-9.07708526e-01 8.23958576e-01 5.32345355e-01 2.06974372e-01
6.25812888e-01 4.28127497e-01 5.86853564e-01 2.33450249e-01
-4.21226203e-01 7.82541111e-02 3.62125278e-01 9.27995086e-01
5.69642067e-01 5.53264260e-01 -5.60740083e-02 1.12239313e+00
-1.47169614e+00 -7.68980026e-01 3.35950047e-01 2.04872921e-01
-4.80084002e-01 -1.25960624e+00 -3.63426447e-01 3.19195777e-01
-7.05216050e-01 -2.15264887e-01 -1.45731568e-01 7.22997010e-01
3.74038428e-01 1.08082843e+00 4.70934868e-01 -3.93970698e-01
2.83613384e-01 6.28203928e-01 1.35770947e-01 -7.57093906e-01
-6.32204771e-01 3.13920438e-01 4.14968729e-01 -1.45452678e-01
3.44872206e-01 -8.88779700e-01 -1.21806657e+00 -4.25266996e-02
-2.24647969e-01 3.38815033e-01 1.38648975e+00 6.64104521e-01
1.15664154e-01 5.25428057e-01 4.87429082e-01 -7.90424168e-01
-5.71114779e-01 -1.29949319e+00 -7.36082494e-01 1.33959413e-01
9.06410366e-02 -4.67360705e-01 -4.14381295e-01 2.16952845e-01] | [14.50281810760498, 6.745894908905029] |
acc3eb4f-20fb-4d6c-a9ea-4a8baa9b0a07 | nu-wave-a-diffusion-probabilistic-model-for | 2104.02321 | null | https://arxiv.org/abs/2104.02321v2 | https://arxiv.org/pdf/2104.02321v2.pdf | NU-Wave: A Diffusion Probabilistic Model for Neural Audio Upsampling | In this work, we introduce NU-Wave, the first neural audio upsampling model to produce waveforms of sampling rate 48kHz from coarse 16kHz or 24kHz inputs, while prior works could generate only up to 16kHz. NU-Wave is the first diffusion probabilistic model for audio super-resolution which is engineered based on neural vocoders. NU-Wave generates high-quality audio that achieves high performance in terms of signal-to-noise ratio (SNR), log-spectral distance (LSD), and accuracy of the ABX test. In all cases, NU-Wave outperforms the baseline models despite the substantially smaller model capacity (3.0M parameters) than baselines (5.4-21%). The audio samples of our model are available at https://mindslab-ai.github.io/nuwave, and the code will be made available soon. | ['Seungu Han', 'Junhyeok Lee'] | 2021-04-06 | null | null | null | null | ['audio-super-resolution', 'audio-super-resolution'] | ['audio', 'music'] | [ 1.21468179e-01 7.67061785e-02 -2.21565083e-01 -8.87954682e-02
-1.50065947e+00 -4.70373809e-01 3.74583691e-01 -2.95284718e-01
-6.79257661e-02 7.48242855e-01 7.11209476e-01 -2.57214487e-01
5.18840142e-02 -6.03561282e-01 -6.14162803e-01 -3.79784703e-01
-2.43117705e-01 -4.06844094e-02 3.08910757e-01 6.13559633e-02
-1.99732214e-01 -2.45555453e-02 -1.59914470e+00 6.33579314e-01
5.79406500e-01 1.12672269e+00 2.72919238e-01 1.23562777e+00
5.12160063e-01 4.96581703e-01 -7.96405435e-01 -1.87942877e-01
3.60833816e-02 -4.38966721e-01 -5.84300041e-01 -5.97988248e-01
4.12736654e-01 -7.70099640e-01 -6.42884612e-01 1.00753796e+00
9.70662177e-01 1.05925292e-01 5.70041597e-01 -1.02692103e+00
-7.19825089e-01 1.38440371e+00 -3.18118334e-01 5.68675160e-01
3.75528336e-01 2.33360082e-01 1.01120555e+00 -1.00722134e+00
2.46493563e-01 1.38564050e+00 8.17252636e-01 6.78297877e-01
-1.23546124e+00 -1.20703161e+00 -3.36352408e-01 2.51121461e-01
-1.72906363e+00 -9.43205297e-01 5.74611366e-01 -2.34052077e-01
9.51591730e-01 2.92678833e-01 3.95755649e-01 1.70883441e+00
1.86878499e-02 5.65198839e-01 8.99868667e-01 -1.72286630e-01
3.06505620e-01 -2.44560331e-01 -4.22894396e-02 1.14010125e-01
-1.74298376e-01 2.81546474e-01 -1.22587967e+00 -3.19705963e-01
1.04021168e+00 -7.74079502e-01 -5.10320604e-01 6.64342225e-01
-8.82668018e-01 4.67474490e-01 1.64808944e-01 1.60547554e-01
-4.33153361e-01 6.48827910e-01 1.73829138e-01 2.35882178e-01
4.56864387e-01 2.16091797e-01 -2.32401639e-01 -7.85329580e-01
-1.11256826e+00 2.17847079e-01 4.68933105e-01 1.01471210e+00
1.87193289e-01 8.02160025e-01 -2.46293440e-01 8.17950845e-01
3.44208419e-01 5.90239406e-01 5.99250257e-01 -1.49556732e+00
4.43150550e-01 -5.04133582e-01 9.14019421e-02 -5.58682144e-01
-1.12768419e-01 -6.13211036e-01 -9.15573478e-01 -9.21562314e-02
3.63828003e-01 -5.37107825e-01 -7.52328515e-01 1.77058256e+00
-8.46068189e-03 6.08395100e-01 8.56251805e-04 8.62768173e-01
8.83606315e-01 9.76400733e-01 -8.42772052e-02 -5.95441088e-02
1.21451640e+00 -8.30234468e-01 -9.20090556e-01 -1.88051574e-02
-8.81779641e-02 -8.75416875e-01 1.21764088e+00 7.27375031e-01
-1.44804966e+00 -7.71386027e-01 -1.13821185e+00 -5.39222732e-03
1.92938104e-01 -5.78295812e-02 1.81239322e-01 8.98017526e-01
-1.36809289e+00 7.18100250e-01 -6.43945098e-01 1.60652548e-01
3.77098322e-01 8.77066180e-02 9.52726454e-02 2.89629430e-01
-1.68353438e+00 1.08917668e-01 1.43642530e-01 -2.33301491e-01
-1.21127105e+00 -1.18901205e+00 -5.69714010e-01 1.37135014e-01
-3.27412002e-02 -4.30224031e-01 1.78705728e+00 -4.81532872e-01
-1.93786275e+00 6.14936985e-02 -1.55718163e-01 -8.57174397e-01
3.37987721e-01 -6.03035033e-01 -7.85078824e-01 2.52813160e-01
-3.59055638e-01 7.96823442e-01 1.08745074e+00 -9.04749990e-01
-4.48631465e-01 1.69074491e-01 -2.78500050e-01 -1.12904862e-01
-3.69432867e-01 1.03414981e-02 -3.53001565e-01 -1.03359997e+00
-2.92429924e-01 -6.38015747e-01 3.87604199e-02 -3.76674533e-01
-3.51571798e-01 2.59286501e-02 6.99589789e-01 -8.25629771e-01
1.46443331e+00 -2.46235585e+00 -3.03259313e-01 -6.29918277e-02
5.18176295e-02 3.96528900e-01 -3.70364577e-01 3.95669848e-01
-1.33964662e-02 3.26257527e-01 -1.58143491e-02 -4.28795636e-01
1.90701798e-01 -3.95195216e-01 -7.61077225e-01 1.27645880e-01
1.70725256e-01 6.05486572e-01 -8.43862593e-01 -2.97647920e-02
-6.27164170e-02 1.15831459e+00 -6.62589192e-01 4.83570015e-03
-2.40817554e-02 2.13981718e-01 1.06656693e-01 4.98717010e-01
5.36273301e-01 -3.96241732e-02 -1.12592697e-01 -2.60461390e-01
2.53157336e-02 7.94506907e-01 -1.36359715e+00 1.75246882e+00
-4.66987610e-01 9.36241210e-01 1.74601227e-01 2.09171191e-01
8.78079295e-01 6.36923194e-01 -3.10209468e-02 -5.30005395e-01
6.49375394e-02 2.53675133e-01 -1.54605761e-01 -3.02162748e-02
6.10362351e-01 1.81815580e-01 1.28558621e-01 3.60468984e-01
2.27801070e-01 -3.03055584e-01 -6.16778433e-02 1.14483453e-01
1.02105081e+00 -1.75626174e-01 -1.30999625e-01 -5.52438349e-02
-8.04739967e-02 -6.16837680e-01 3.60775083e-01 8.93435359e-01
-2.62536138e-01 7.97096729e-01 1.30462185e-01 2.68732727e-01
-1.10109019e+00 -1.57281458e+00 -3.81445020e-01 9.13792491e-01
-2.16352209e-01 -7.24959314e-01 -1.00883436e+00 1.11228578e-01
-1.58090293e-01 8.62723172e-01 -2.25347988e-02 -1.94493867e-02
-2.36142144e-01 -3.77256721e-01 1.37778604e+00 5.12778044e-01
4.28312391e-01 -9.30749059e-01 -2.38247871e-01 2.69468635e-01
-2.98830599e-01 -1.03685033e+00 -6.94182813e-01 2.76834704e-02
-7.41199136e-01 -3.20611507e-01 -7.87837505e-01 -3.98215860e-01
-2.26960540e-01 -1.48317710e-01 8.59341085e-01 -4.90183383e-01
-2.15867460e-02 4.59323898e-02 -2.66956210e-01 -3.84171277e-01
-4.60213989e-01 7.39028379e-02 4.35363233e-01 -8.15010071e-02
1.35133073e-01 -1.06144345e+00 -6.75025523e-01 -2.84584649e-02
-7.98756063e-01 -1.07461222e-01 5.03974855e-01 4.98501241e-01
6.24582052e-01 2.54042208e-01 9.10433412e-01 -4.01693672e-01
1.01535976e+00 -4.09728467e-01 -4.48574573e-01 -5.40335894e-01
-4.49793726e-01 -3.00061524e-01 6.52436495e-01 -8.60954106e-01
-9.61946905e-01 -2.61403948e-01 -6.14528418e-01 -5.36651671e-01
-3.24367940e-01 3.50669354e-01 -5.12586199e-02 4.13623244e-01
9.03553665e-01 6.41173199e-02 -5.69737136e-01 -6.86975241e-01
4.81255323e-01 1.23403764e+00 8.49404514e-01 -3.64189982e-01
6.16579652e-01 2.44606018e-01 -6.00068569e-01 -9.78586674e-01
-4.28918928e-01 -5.64105362e-02 -1.37164339e-01 -6.89319968e-02
5.00798225e-01 -1.28241718e+00 -4.00369644e-01 5.63174188e-01
-1.01358187e+00 -7.03369141e-01 -3.21714818e-01 8.05295587e-01
-4.83811140e-01 1.35187685e-01 -1.26814663e+00 -1.15973628e+00
-6.20006084e-01 -8.72132003e-01 9.53613460e-01 2.31257930e-01
-7.51393199e-01 -3.64751905e-01 -1.44823566e-01 1.23673365e-01
7.50440538e-01 -1.55223146e-01 5.75567007e-01 -2.82339275e-01
-3.65157574e-01 1.80000022e-01 -1.56471375e-02 5.11861444e-01
-1.29956633e-01 6.53680488e-02 -1.58586657e+00 -1.50149494e-01
-1.62928328e-02 -2.85039872e-01 7.96528697e-01 6.87006354e-01
1.17809963e+00 -4.50488389e-01 1.43719926e-01 5.66176713e-01
1.03205669e+00 3.16267073e-01 7.10210383e-01 -1.38195828e-01
2.39368051e-01 7.92114511e-02 1.29318520e-01 7.33810902e-01
2.14428633e-01 6.28385365e-01 1.99341312e-01 2.26117298e-01
-7.82453597e-01 -5.10643601e-01 7.75167704e-01 1.43696094e+00
1.39856428e-01 -2.46975616e-01 -7.83906281e-01 6.94996595e-01
-1.17786145e+00 -1.06756246e+00 -3.11478913e-01 2.16405416e+00
1.39038897e+00 3.71670693e-01 3.60100359e-01 4.43066925e-01
6.56898022e-01 2.05196589e-01 -4.30856287e-01 -4.80994940e-01
-9.57034156e-02 4.97330487e-01 3.25896084e-01 8.53833139e-01
-8.63878191e-01 7.13367462e-01 6.51760340e+00 1.27615464e+00
-1.06354940e+00 3.36583316e-01 6.62762523e-01 -5.61138928e-01
-3.11174005e-01 -4.39458966e-01 -9.27809179e-01 6.12643540e-01
1.90645766e+00 -3.60261261e-01 8.22699368e-01 5.50984740e-01
3.98281634e-01 1.62248418e-01 -8.83644164e-01 1.07994282e+00
-3.07570904e-01 -1.24116683e+00 -6.04174705e-03 -2.62677409e-02
6.11322880e-01 2.29513168e-01 6.38610601e-01 3.66420865e-01
3.26067477e-01 -1.12234986e+00 9.27643955e-01 5.30808926e-01
1.45831633e+00 -9.24190044e-01 3.42949867e-01 2.07316652e-01
-1.21204829e+00 9.32173431e-02 -2.78511465e-01 -5.33340909e-02
3.66072208e-01 8.41244876e-01 -8.49665046e-01 1.88994572e-01
9.95606482e-01 1.49038419e-01 -8.76137912e-02 1.04656935e+00
-2.39820018e-01 1.33083689e+00 -4.42653149e-01 5.58237545e-02
-7.30281100e-02 4.23637629e-01 7.09070444e-01 1.52030694e+00
6.79497182e-01 6.51533380e-02 -2.95655012e-01 9.21144843e-01
-3.65695953e-01 -3.26223046e-01 -2.40849823e-01 -9.83182341e-02
1.09015393e+00 8.33983719e-01 -4.80305441e-02 -2.14530692e-01
3.51045281e-02 7.27812707e-01 -2.45044380e-01 6.20975256e-01
-1.11602819e+00 -9.50563908e-01 8.43880355e-01 1.21787943e-01
3.59359711e-01 -2.60115445e-01 -2.07580447e-01 -6.09836519e-01
-1.45483240e-01 -1.07941639e+00 -6.25186116e-02 -9.84282851e-01
-1.23248410e+00 1.01460600e+00 -5.44031076e-02 -8.99057746e-01
-6.43856823e-01 -1.53192163e-01 -1.71885207e-01 1.15163112e+00
-1.23730230e+00 -6.86095297e-01 -4.57077287e-02 3.93462718e-01
7.14964509e-01 1.09008268e-01 1.05948257e+00 5.58433592e-01
-2.86303490e-01 9.83136058e-01 3.94654386e-02 4.14665863e-02
7.38842666e-01 -1.12742496e+00 9.35739875e-01 7.57226348e-01
2.19707370e-01 4.01122630e-01 9.01703060e-01 -5.30385792e-01
-1.14195848e+00 -1.25316608e+00 8.55639338e-01 -4.24829513e-01
9.97470856e-01 -5.74060678e-01 -9.99137878e-01 2.45447144e-01
4.51905578e-01 -2.41399273e-01 7.59468436e-01 -6.47821948e-02
-6.29868388e-01 -1.02826573e-01 -8.96150172e-01 8.12802255e-01
9.54714656e-01 -1.00149405e+00 -3.06288004e-01 -1.30389079e-01
1.28752017e+00 -6.33427083e-01 -1.19753551e+00 2.54447699e-01
6.67119026e-01 -8.57980907e-01 9.17899191e-01 -9.63919461e-02
1.61764562e-01 -4.11775470e-01 -4.02377993e-01 -1.33199883e+00
-4.13209260e-01 -1.20892131e+00 -5.92635512e-01 1.51116574e+00
7.00734854e-01 -4.62705433e-01 4.88431692e-01 3.28586251e-02
-1.30934373e-01 -6.02849782e-01 -1.14563870e+00 -9.74507213e-01
7.98900202e-02 -1.02230024e+00 7.46457636e-01 6.17488027e-01
-6.22429140e-02 2.93874919e-01 -4.96047884e-01 4.66225833e-01
5.98075688e-01 -4.55358744e-01 3.47859681e-01 -8.31881642e-01
-7.04073966e-01 -5.50406933e-01 2.30364650e-02 -1.25565326e+00
-1.92157626e-01 -5.58886230e-01 -1.63882710e-02 -1.24328494e+00
-2.95306891e-01 -4.82344925e-02 -3.62175524e-01 3.38120580e-01
1.45435661e-01 6.05322421e-01 2.47274876e-01 -6.40942305e-02
-1.75639674e-01 5.76838493e-01 8.64016473e-01 -1.21484436e-01
-5.04612148e-01 -4.55435924e-02 -7.30632603e-01 6.82369173e-01
1.34577775e+00 -5.42943358e-01 -5.33027411e-01 -4.62217093e-01
-1.56052515e-01 7.47485980e-02 2.07222357e-01 -1.36007059e+00
1.63049206e-01 2.06341639e-01 3.19139332e-01 -4.17805582e-01
8.10422659e-01 -2.12475866e-01 4.96349722e-01 1.38449892e-01
-5.38957119e-01 -1.83235984e-02 5.36961317e-01 5.07233679e-01
-1.60986885e-01 -1.79754887e-02 6.26259446e-01 3.42076838e-01
-3.74707758e-01 1.13255665e-01 -7.41119564e-01 2.27016076e-01
3.28018188e-01 1.37920067e-01 -6.58596218e-01 -8.71854365e-01
-5.56678474e-01 -4.37696636e-01 6.09369799e-02 4.70645159e-01
7.32046366e-01 -1.45277226e+00 -9.51269150e-01 1.55964583e-01
-3.04410815e-01 -1.11590721e-01 5.02860427e-01 5.78787327e-01
-1.86326817e-01 3.57411683e-01 8.07045028e-02 -3.68709564e-01
-1.08367860e+00 5.01854531e-02 2.08781466e-01 8.76732841e-02
-5.80554068e-01 1.17927635e+00 -1.91847920e-01 2.59264290e-01
7.23583281e-01 -4.35784131e-01 1.22761995e-01 -1.98298290e-01
1.10146773e+00 5.84814310e-01 -4.18634824e-02 -4.75473583e-01
-2.63165027e-01 9.90087464e-02 2.56846458e-01 -8.05846572e-01
1.05763459e+00 -7.94070885e-02 3.85224968e-01 6.72256529e-01
9.85044122e-01 3.99307281e-01 -1.43674004e+00 -7.66566843e-02
-3.05287838e-01 -3.02157670e-01 3.19926769e-01 -7.35262692e-01
-6.93197429e-01 1.03436768e+00 8.04713190e-01 2.18392059e-01
1.29840922e+00 -1.21564016e-01 1.16672170e+00 -1.40147567e-01
4.06057388e-01 -9.84784186e-01 1.15391314e-01 6.77204370e-01
1.04941297e+00 -3.07819992e-01 -3.54718953e-01 -2.13615615e-02
-5.71556568e-01 8.94847512e-01 4.72580701e-01 -5.32044321e-02
6.35133505e-01 8.97499144e-01 2.50861049e-01 4.06482369e-01
-1.07503986e+00 2.51698848e-02 3.03984940e-01 8.89214039e-01
6.57514632e-01 1.10743336e-01 3.05099934e-01 1.09195209e+00
-8.16073060e-01 -2.16580313e-02 5.80059171e-01 2.90540367e-01
-4.36420619e-01 -8.62655282e-01 -4.64766294e-01 2.64606625e-01
-7.10083604e-01 -5.00602245e-01 -2.16071188e-01 1.91488594e-01
-2.64874667e-01 1.31126750e+00 9.87883583e-02 -7.98587143e-01
2.57934362e-01 1.26780629e-01 1.78380653e-01 -2.41073772e-01
-5.03063560e-01 6.46478534e-01 2.96083242e-01 -5.18073916e-01
-4.07567844e-02 -4.44716632e-01 -1.22345805e+00 -5.14226139e-01
-2.81382442e-01 1.12487629e-01 4.85653907e-01 2.46995240e-01
6.62777424e-01 9.49156463e-01 3.23254317e-01 -9.59438801e-01
-7.12820470e-01 -1.23949516e+00 -7.12743402e-01 -1.71568975e-01
5.91904283e-01 -8.22671577e-02 -5.16827106e-01 2.62939215e-01] | [15.423030853271484, 5.836530685424805] |
266f4577-baf1-429b-892c-35c037341397 | 3d-pictorial-structures-for-multiple-human | null | null | http://openaccess.thecvf.com/content_cvpr_2014/html/Belagiannis_3D_Pictorial_Structures_2014_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2014/papers/Belagiannis_3D_Pictorial_Structures_2014_CVPR_paper.pdf | 3D Pictorial Structures for Multiple Human Pose Estimation | In this work, we address the problem of 3D pose estimation of multiple humans from multiple views. This is a more challenging problem than single human 3D pose estimation due to the much larger state space, partial occlusions as well as across view ambiguities when not knowing the identity of the humans in advance. To address these problems, we first create a reduced state space by triangulation of corresponding body joints obtained from part detectors in pairs of camera views. In order to resolve the ambiguities of wrong and mixed body parts of multiple humans after triangulation and also those coming from false positive body part detections, we introduce a novel 3D pictorial structures (3DPS) model. Our model infers 3D human body configurations from our reduced state space. The 3DPS model is generic and applicable to both single and multiple human pose estimation. In order to compare to the state-of-the art, we first evaluate our method on single human 3D pose estimation on HumanEva-I [22] and KTH Multiview Football Dataset II [8] datasets. Then, we introduce and evaluate our method on two datasets for multiple human 3D pose estimation. | ['Nassir Navab', 'Mykhaylo Andriluka', 'Vasileios Belagiannis', 'Slobodan Ilic', 'Sikandar Amin', 'Bernt Schiele'] | 2014-06-01 | null | null | null | cvpr-2014-6 | ['3d-multi-person-pose-estimation'] | ['computer-vision'] | [-8.65033865e-02 1.04497001e-01 1.30057037e-01 -5.65280765e-02
-3.30719411e-01 -5.48665643e-01 4.42482144e-01 -2.42453635e-01
-5.78605294e-01 6.91673756e-01 7.42775798e-02 3.77858520e-01
1.61028564e-01 -2.21802771e-01 -8.07163477e-01 -2.51740456e-01
-1.05549254e-01 1.19451189e+00 6.41862631e-01 -2.33775839e-01
-1.13429546e-01 4.58257794e-01 -1.46621978e+00 -6.32383302e-02
2.41945446e-01 5.62925458e-01 -1.44316956e-01 7.62625456e-01
5.77897608e-01 2.11694479e-01 -6.70517147e-01 -4.14827198e-01
6.62465513e-01 -3.18671077e-01 -6.85544789e-01 7.10048318e-01
9.14921105e-01 -5.00003576e-01 -8.37839171e-02 8.13258588e-01
5.43244004e-01 1.93231195e-01 6.35473788e-01 -1.60207367e+00
3.61031562e-01 -1.02595977e-01 -8.92464101e-01 -3.23066190e-02
9.70535815e-01 9.42089260e-02 5.97136915e-01 -8.19500625e-01
1.00276721e+00 1.69585383e+00 8.12574625e-01 6.22362137e-01
-1.29186749e+00 -3.57760161e-01 3.31317931e-01 5.70701100e-02
-1.37335670e+00 -4.02322829e-01 5.08932650e-01 -7.73998559e-01
7.00063467e-01 9.68001932e-02 1.13955021e+00 1.30559611e+00
2.90726483e-01 8.31147015e-01 9.62897539e-01 -3.81810993e-01
-5.10875434e-02 -2.48229116e-01 1.42889947e-01 8.32779944e-01
7.19076991e-01 1.96170118e-02 -5.89628875e-01 -2.42873982e-01
1.05902684e+00 4.13476042e-02 -1.01057112e-01 -1.23953819e+00
-1.57082987e+00 4.62851614e-01 1.39984295e-01 -3.21897089e-01
-5.02104223e-01 9.47837159e-02 2.39723697e-01 -1.06670082e-01
3.90374921e-02 -3.19570838e-03 -3.83300692e-01 1.28408670e-01
-6.98036611e-01 8.80190313e-01 8.19381654e-01 1.08318388e+00
4.04740661e-01 -1.83068991e-01 1.30632609e-01 5.13031006e-01
5.20037174e-01 6.80180013e-01 1.27275825e-01 -9.67825353e-01
6.89116776e-01 5.83754122e-01 5.15920579e-01 -8.72230113e-01
-6.92829251e-01 -2.51826882e-01 -7.88637817e-01 3.08356613e-01
7.71640122e-01 -1.10290237e-01 -1.02284586e+00 1.68394506e+00
9.35044587e-01 -3.61193866e-01 -2.10468486e-01 1.27842331e+00
4.84175354e-01 2.48230815e-01 -2.22589120e-01 -2.38453727e-02
1.77879941e+00 -8.46291959e-01 -6.22111142e-01 -7.29507565e-01
7.03336671e-02 -5.66868544e-01 1.51070699e-01 5.11332929e-01
-1.16482532e+00 -7.78226078e-01 -9.54553127e-01 5.90404794e-02
-9.87621397e-02 1.90839440e-01 1.91461489e-01 4.55453992e-01
-6.85171068e-01 3.84631515e-01 -1.17488706e+00 -7.31992483e-01
-2.06296131e-01 5.37617385e-01 -8.41338575e-01 7.33197182e-02
-1.12415564e+00 1.43728983e+00 4.40177232e-01 3.72886956e-01
-5.92516005e-01 -2.73657292e-02 -1.12208581e+00 -5.16241729e-01
9.49050725e-01 -1.39719522e+00 1.23728955e+00 -3.25807601e-01
-1.30623686e+00 1.00968277e+00 -4.07446325e-02 -3.59717965e-01
9.48043764e-01 -7.85648286e-01 7.52913719e-03 1.84681147e-01
1.53081983e-01 5.14582634e-01 8.84395182e-01 -1.29110169e+00
-4.91504759e-01 -1.01844013e+00 -1.23200126e-01 5.98114908e-01
6.59497023e-01 -3.64572316e-01 -7.93372929e-01 -4.68125314e-01
6.85603261e-01 -1.52157581e+00 -4.36048418e-01 2.63523519e-01
-5.77610433e-01 2.16651261e-01 4.43495214e-01 -7.59126067e-01
7.42368340e-01 -1.66302907e+00 9.09757972e-01 1.27886787e-01
2.76485085e-01 2.17024326e-01 2.14661196e-01 3.33080441e-01
1.03267238e-01 -4.00792181e-01 6.54631034e-02 -5.23480475e-01
-9.30948590e-04 5.00130594e-01 2.67829597e-01 8.20494771e-01
2.13362486e-03 5.36778092e-01 -6.40498281e-01 -7.08392024e-01
3.78049701e-01 3.51750940e-01 -4.12062824e-01 2.44692683e-01
1.22203268e-01 6.58969522e-01 -5.05365014e-01 7.20711350e-01
5.95487058e-01 -9.67849344e-02 3.71931881e-01 -3.89402717e-01
-2.74635255e-02 -1.72740862e-01 -1.95123887e+00 1.77139187e+00
8.71622413e-02 -2.51340330e-01 3.07222962e-01 -6.79412544e-01
4.44785088e-01 5.93523502e-01 5.53343415e-01 -1.54307649e-01
1.88560575e-01 -9.71208606e-03 -4.72874753e-02 -2.86973119e-01
4.43541825e-01 -3.23716372e-01 -3.54898125e-01 1.49517432e-01
3.83772314e-01 -1.17446020e-01 2.35236183e-01 2.60153245e-02
7.84127176e-01 5.98612249e-01 9.24907029e-01 -4.02461477e-02
5.33319771e-01 -9.52892080e-02 7.48702526e-01 5.54920852e-01
-4.93069023e-01 7.53058195e-01 3.06182384e-01 -6.09430909e-01
-9.30914402e-01 -1.43018913e+00 2.07702354e-01 5.26770532e-01
3.53815466e-01 -5.74486315e-01 -6.45504713e-01 -6.21260107e-01
2.62527943e-01 -9.98998433e-03 -5.38072586e-01 3.28703262e-02
-7.51780570e-01 -5.05949438e-01 2.45073676e-01 5.20858645e-01
3.30354452e-01 -4.35316831e-01 -1.31059957e+00 1.91468224e-01
-6.93051696e-01 -1.45184505e+00 -3.78056526e-01 -7.49529898e-02
-9.53642011e-01 -1.43084407e+00 -9.03916955e-01 -4.39351469e-01
4.86147523e-01 1.47082299e-01 1.20117819e+00 -2.33362839e-01
-4.42675889e-01 6.97078824e-01 -2.23807007e-01 -3.86437386e-01
-3.46396655e-01 -2.75404096e-01 8.27699840e-01 -1.12787358e-01
-1.75810214e-02 -2.67965019e-01 -4.97229218e-01 7.75337100e-01
-3.10092986e-01 2.01762915e-01 5.11483550e-01 7.04389095e-01
6.81150675e-01 -2.60433018e-01 -2.24542603e-01 -5.26776195e-01
1.57329196e-03 1.04969554e-01 -7.34698594e-01 3.34861726e-01
7.15510920e-03 1.30489646e-02 -7.96323083e-03 -4.91002440e-01
-8.47522497e-01 7.96149135e-01 8.94854739e-02 -4.89725322e-01
-4.69345212e-01 -7.02951401e-02 -1.96787834e-01 9.59580839e-02
4.89356488e-01 -1.72829106e-01 3.28796625e-01 -6.65894330e-01
2.76635796e-01 1.24715842e-01 7.46779501e-01 -5.07453620e-01
9.36635137e-01 5.16119421e-01 5.73937416e-01 -7.97324419e-01
-7.52475798e-01 -8.29307854e-01 -1.45833218e+00 -4.79465306e-01
1.05550635e+00 -1.20471060e+00 -8.21012259e-01 5.76236784e-01
-1.41602492e+00 1.40814736e-01 -1.59077533e-02 6.62671626e-01
-8.29897642e-01 7.64351249e-01 -5.57153821e-01 -1.05170202e+00
-1.24181248e-01 -1.18118083e+00 1.51663816e+00 -1.68652415e-01
-6.69356346e-01 -5.95534682e-01 3.95964742e-01 4.11941200e-01
-4.91040915e-01 7.97422886e-01 2.40866318e-01 -3.08023840e-01
-2.65569746e-01 -4.51189876e-01 5.12743413e-01 1.08839134e-02
-1.47477657e-01 -3.07535350e-01 -4.75759655e-01 -5.00089705e-01
-6.73558041e-02 -2.15741605e-01 4.61204648e-01 3.76814157e-01
-4.31288220e-02 -7.56694004e-02 -5.77039301e-01 1.83044806e-01
1.04179156e+00 -2.30164006e-01 1.82862967e-01 2.96190083e-01
8.01548898e-01 9.04138029e-01 8.69008780e-01 5.93055606e-01
5.04554272e-01 1.30627370e+00 5.85912347e-01 1.10001557e-01
-2.69073416e-02 -3.46540630e-01 3.63866419e-01 4.93618101e-01
-5.41449130e-01 -1.16755463e-01 -1.04241848e+00 2.27829173e-01
-2.07458830e+00 -8.42605412e-01 -3.58734667e-01 2.48887873e+00
1.80233046e-01 1.91431001e-01 7.42391586e-01 1.71884492e-01
9.63033676e-01 -1.36498109e-01 -4.58322853e-01 3.60947311e-01
3.08723032e-01 -1.93429723e-01 3.89477670e-01 3.77055764e-01
-1.22401321e+00 7.19012558e-01 5.90881109e+00 7.58006517e-03
-3.49680007e-01 -5.14821224e-02 -1.73547834e-01 -2.06083208e-01
6.33979976e-01 5.69093190e-02 -1.13165414e+00 -5.32077178e-02
3.01378995e-01 3.72427642e-01 -6.88945055e-02 5.40830374e-01
-2.32320085e-01 -4.09479648e-01 -1.34791827e+00 1.22177136e+00
3.49527538e-01 -4.32409108e-01 -3.74821126e-02 3.18307161e-01
4.02723670e-01 -3.14393282e-01 -3.64784300e-01 1.86247423e-01
9.96963531e-02 -4.10465091e-01 1.23409069e+00 5.28999507e-01
2.57775217e-01 -4.75282550e-01 6.77310348e-01 8.58980417e-01
-1.39103425e+00 1.66601524e-01 -1.02521390e-01 -2.05103576e-01
7.10742831e-01 3.65331650e-01 -5.70358813e-01 7.68425584e-01
6.62706196e-01 5.70140839e-01 -5.93222857e-01 1.04901099e+00
-2.92337000e-01 -1.73863947e-01 -6.35497034e-01 4.61338699e-01
-3.02039355e-01 -6.80702552e-02 9.24116075e-01 7.26562142e-01
1.76376760e-01 2.52500534e-01 6.31354094e-01 5.07582724e-01
4.72081631e-01 -3.32026422e-01 -5.77921033e-01 4.35797364e-01
-5.57644777e-02 1.17311931e+00 -8.17230582e-01 -4.03505534e-01
-1.77233309e-01 1.33988547e+00 1.89013228e-01 1.66054636e-01
-8.17462921e-01 1.85918972e-01 5.66491723e-01 3.60233039e-01
3.09611619e-01 -6.05738342e-01 4.12948430e-01 -1.73999012e+00
2.66776055e-01 -9.38972652e-01 5.75363755e-01 -7.07160532e-01
-1.09242034e+00 3.80764097e-01 5.80971777e-01 -1.52917290e+00
-7.58482814e-01 -6.77567482e-01 8.13053399e-02 6.54778957e-01
-6.77244782e-01 -1.19008291e+00 -1.00964241e-01 5.30170321e-01
5.19398093e-01 4.06025261e-01 6.68519616e-01 -4.01475579e-02
-3.75863761e-01 5.18067479e-02 -6.84208393e-01 4.66965921e-02
8.88870358e-01 -1.23907506e+00 8.37742567e-01 9.22309339e-01
9.51854587e-02 6.20319366e-01 9.25183475e-01 -9.89534736e-01
-1.63487327e+00 -5.61673105e-01 6.85243130e-01 -1.11907005e+00
1.19738795e-01 -6.66702747e-01 -4.53379750e-01 1.05374134e+00
-4.19987351e-01 9.51076448e-02 3.36757421e-01 1.20115608e-01
-1.25422359e-01 2.92851567e-01 -1.05724955e+00 5.14216900e-01
1.27114236e+00 9.67538059e-02 -1.00047100e+00 1.28950641e-01
1.48391977e-01 -9.25325394e-01 -8.99099469e-01 6.38512731e-01
1.01680660e+00 -1.02332366e+00 1.42895293e+00 -4.38063532e-01
-2.26452887e-01 -5.23640692e-01 -1.90487429e-01 -1.08816230e+00
-2.57824302e-01 -3.43318582e-01 -3.19441676e-01 6.64788663e-01
-1.11482844e-01 -2.37999022e-01 8.40177000e-01 4.77348179e-01
3.71622413e-01 -2.19772205e-01 -1.12182486e+00 -9.14936841e-01
-3.73190850e-01 -6.22922070e-02 4.42268364e-02 4.21468675e-01
-1.06238261e-01 5.32615483e-01 -8.72249424e-01 5.18440306e-01
1.11716509e+00 2.12954953e-01 1.48727620e+00 -1.58960223e+00
-5.73253632e-01 9.11228731e-02 -8.55813205e-01 -1.08932507e+00
-1.62435815e-01 -1.47345424e-01 3.55531037e-01 -1.32451928e+00
3.69069040e-01 4.61717695e-01 3.40093613e-01 3.18882048e-01
-2.06891730e-01 2.84289539e-01 7.06585705e-01 1.51175380e-01
-8.73637259e-01 3.35270852e-01 1.23129392e+00 2.69703656e-01
7.32805282e-02 3.39103669e-01 7.97394812e-02 1.09685171e+00
2.41309643e-01 -4.54908758e-01 -5.96039630e-02 -2.25561962e-01
1.62021413e-01 6.95331991e-01 9.97182608e-01 -1.30255115e+00
1.25532508e-01 1.13439761e-01 8.16744030e-01 -1.03721380e+00
7.76338100e-01 -9.68646169e-01 6.34542584e-01 8.88823748e-01
1.99948847e-01 2.42214411e-01 1.22224338e-01 8.21686685e-01
5.54562919e-03 1.25901466e-02 6.53006494e-01 -7.27435112e-01
-6.65676951e-01 2.64759213e-01 -3.95135939e-01 4.74693850e-02
9.12634015e-01 -4.09160703e-01 3.14779103e-01 -3.41695547e-01
-1.23265254e+00 3.92696083e-01 6.42837465e-01 4.91427928e-01
4.67121959e-01 -1.26214182e+00 -4.62839782e-01 3.19674373e-01
2.02591881e-01 1.67703897e-01 3.21387261e-01 9.46076035e-01
-3.63544345e-01 4.02890354e-01 -6.08255684e-01 -9.77803409e-01
-1.72576928e+00 5.31725883e-01 3.28137308e-01 -3.20306242e-01
-5.93901157e-01 3.05080771e-01 2.28088200e-01 -7.12534606e-01
1.39998272e-01 -3.78530562e-01 -1.57059237e-01 -5.39819002e-02
2.77763098e-01 7.46076584e-01 -1.47163600e-01 -1.18862689e+00
-6.65347576e-01 1.11598480e+00 1.74060062e-01 -4.75487560e-01
8.91981661e-01 -4.08808708e-01 2.38660172e-01 5.25827587e-01
7.37725139e-01 -1.78947687e-01 -1.28034687e+00 -1.36613712e-01
-2.53564328e-01 -4.05982822e-01 -7.15761244e-01 -6.10720754e-01
-5.63207269e-01 8.95822763e-01 6.61958218e-01 -3.09701234e-01
5.56616366e-01 3.71377356e-02 5.82242310e-01 4.38553572e-01
8.60697031e-01 -1.03212798e+00 3.90275605e-02 4.39729333e-01
9.86088336e-01 -1.22055840e+00 4.64247048e-01 -6.37832344e-01
-6.49373591e-01 1.01364112e+00 8.24076355e-01 -3.71273123e-02
3.78301024e-01 -1.34486808e-02 -3.85421179e-02 -2.70334721e-01
-4.47699636e-01 -3.30500126e-01 5.86868882e-01 5.43594599e-01
1.35395527e-01 -4.16177238e-04 -1.43015966e-01 4.59976047e-01
-7.82610476e-02 -5.67825027e-02 4.04827893e-01 1.22259820e+00
-3.14564109e-01 -1.18867671e+00 -9.60164011e-01 -1.50748312e-01
-3.26324016e-01 6.70672596e-01 -6.59538388e-01 1.27984500e+00
2.46989027e-01 7.35268891e-01 -1.27857134e-01 -3.40244174e-01
9.91058528e-01 2.63770759e-01 1.02274334e+00 -6.97909474e-01
-2.96071500e-01 4.68056887e-01 1.16994284e-01 -8.56206477e-01
-6.72885299e-01 -1.08533204e+00 -1.01302886e+00 1.48847282e-01
-4.18853849e-01 -1.90335080e-01 4.99250144e-01 9.93955612e-01
-5.48396148e-02 1.80002496e-01 -1.77597404e-01 -1.51119900e+00
-8.40327263e-01 -8.14008772e-01 -7.52864301e-01 6.11908913e-01
3.74883175e-01 -1.22705257e+00 -1.34700298e-01 7.06346780e-02] | [7.017344951629639, -0.9885551333427429] |
06b87819-0c6f-47d6-be92-cb196678aa5f | personalized-food-recommendation-as | 2101.01775 | null | https://arxiv.org/abs/2101.01775v1 | https://arxiv.org/pdf/2101.01775v1.pdf | Personalized Food Recommendation as Constrained Question Answering over a Large-scale Food Knowledge Graph | Food recommendation has become an important means to help guide users to adopt healthy dietary habits. Previous works on food recommendation either i) fail to consider users' explicit requirements, ii) ignore crucial health factors (e.g., allergies and nutrition needs), or iii) do not utilize the rich food knowledge for recommending healthy recipes. To address these limitations, we propose a novel problem formulation for food recommendation, modeling this task as constrained question answering over a large-scale food knowledge base/graph (KBQA). Besides the requirements from the user query, personalized requirements from the user's dietary preferences and health guidelines are handled in a unified way as additional constraints to the QA system. To validate this idea, we create a QA style dataset for personalized food recommendation based on a large-scale food knowledge graph and health guidelines. Furthermore, we propose a KBQA-based personalized food recommendation framework which is equipped with novel techniques for handling negations and numerical comparisons in the queries. Experimental results on the benchmark show that our approach significantly outperforms non-personalized counterparts (average 59.7% absolute improvement across various evaluation metrics), and is able to recommend more relevant and healthier recipes. | ['Mohammed J. Zaki', 'Ching-Hua Chen', 'Ananya Subburathinam', 'Yu Chen'] | 2021-01-05 | null | null | null | null | ['food-recommendation'] | ['miscellaneous'] | [-7.22405454e-03 2.89819539e-01 -6.85314357e-01 -7.43198037e-01
-6.28294468e-01 -7.27869213e-01 -3.48402232e-01 9.01895344e-01
-2.71929979e-01 4.19637203e-01 8.67276132e-01 -1.83648199e-01
-3.40396583e-01 -1.32376766e+00 -6.71271861e-01 -1.69645786e-01
1.52348697e-01 3.41022730e-01 3.79565537e-01 -7.65223145e-01
5.59693202e-02 -3.39204192e-01 -1.59751105e+00 4.10328567e-01
1.33517325e+00 7.69888580e-01 1.60189226e-01 4.03434277e-01
-1.79487035e-01 4.95101243e-01 1.26387263e-02 -5.17945766e-01
5.61742000e-02 -4.97976363e-01 -7.31931925e-01 -1.03929959e-01
5.40884376e-01 -7.05928206e-01 3.50718386e-02 1.27928412e+00
4.88880217e-01 4.66334432e-01 3.72140795e-01 -7.69555330e-01
-1.48422456e+00 1.34590971e+00 1.69058740e-01 -9.28965807e-02
9.30004537e-01 -1.17165044e-01 1.22283435e+00 -4.59017634e-01
2.22682789e-01 1.08532190e+00 5.48685372e-01 6.66173756e-01
-8.58528793e-01 -1.29902333e-01 6.10201597e-01 4.09872264e-01
-1.13891327e+00 -9.08592567e-02 4.89669472e-01 -2.78049886e-01
1.08040714e+00 7.17558026e-01 8.13556790e-01 6.26642942e-01
-1.58940807e-01 3.73205334e-01 6.63412809e-01 -2.95428514e-01
4.91769075e-01 1.56159535e-01 9.19721186e-01 7.24785447e-01
6.57480955e-01 2.61519533e-02 -8.87714699e-02 -2.61341691e-01
1.84976846e-01 2.42143020e-01 -4.49794114e-01 -3.32302690e-01
-1.13268900e+00 1.19871032e+00 5.68161845e-01 1.45209625e-01
-5.53619325e-01 -2.48594299e-01 2.51613259e-01 3.43543619e-01
2.51768380e-01 4.10153151e-01 -8.12020361e-01 7.23098636e-01
-2.93108970e-01 4.88520294e-01 1.26345253e+00 9.69403446e-01
6.93586767e-01 -4.40688014e-01 -4.03700262e-01 5.51281631e-01
7.77108610e-01 9.60451186e-01 3.27822208e-01 -7.36175299e-01
1.68107376e-01 7.80894637e-01 6.59007549e-01 -1.47207606e+00
-6.67714655e-01 -2.73816019e-01 -6.13431931e-01 -6.68534994e-01
3.76175821e-01 -4.24044430e-02 -5.97635269e-01 1.85970116e+00
7.85197496e-01 -3.42606455e-01 2.26129308e-01 1.31114864e+00
1.43258917e+00 5.37138045e-01 6.12100542e-01 -2.85814196e-01
1.95114696e+00 -9.26091850e-01 -9.63751435e-01 7.33100921e-02
8.69771481e-01 -5.26530445e-01 1.30625045e+00 3.93983632e-01
-8.03286493e-01 -4.94503587e-01 -8.58433366e-01 -3.89978141e-01
-7.05410182e-01 1.34181321e-01 7.02557266e-01 9.08738494e-01
-8.74215901e-01 2.97532380e-01 -4.06919003e-01 -6.94177687e-01
-3.44019234e-01 3.84358197e-01 1.26517460e-01 -5.78305185e-01
-1.74656832e+00 6.70802832e-01 6.85160875e-01 -1.41581863e-01
-3.07247251e-01 -9.56509531e-01 -1.17001367e+00 2.22532198e-01
8.84161890e-01 -1.25887156e+00 1.22275090e+00 -4.78345245e-01
-1.52129972e+00 2.74162412e-01 2.36160219e-01 -1.72027081e-01
-2.50280708e-01 -3.74197513e-01 -9.18222368e-01 1.48067791e-02
9.59347188e-02 5.87388456e-01 2.52593160e-01 -7.51160026e-01
-6.87432051e-01 -4.71131206e-01 7.65711546e-01 2.25628421e-01
-4.71396357e-01 -3.05967927e-02 -2.35905990e-01 -5.50816298e-01
-1.43339247e-01 -6.22928441e-01 -5.09959221e-01 -5.82528748e-02
-3.90781194e-01 -5.57725251e-01 6.48187473e-02 -9.43767071e-01
1.59213734e+00 -1.84313834e+00 5.82954101e-03 2.51603335e-01
-3.29371095e-02 1.50222495e-01 -2.04358608e-01 5.42062104e-01
2.57131815e-01 7.63769224e-02 1.24419980e-01 5.41900218e-01
6.44024193e-01 3.80414456e-01 -1.10209480e-01 9.19165649e-03
-3.05168539e-01 9.03928399e-01 -1.37247562e+00 -3.40048194e-01
1.90113366e-01 3.87688696e-01 -1.23363209e+00 -3.96113582e-02
-1.08627164e+00 7.83042908e-02 -8.83380771e-01 5.60543239e-01
4.82837200e-01 -4.51140612e-01 4.74954337e-01 -1.04695511e+00
4.43062857e-02 4.03360337e-01 -1.27900326e+00 1.83929014e+00
-5.96448109e-02 -1.03825188e+00 -2.28989780e-01 -6.98282540e-01
6.61423087e-01 2.92442948e-01 3.87181491e-01 -8.38029981e-01
4.93405834e-02 5.76289482e-02 -2.10142642e-01 -1.05733728e+00
5.97594917e-01 1.39363065e-01 -1.63726151e-01 3.98351103e-01
-2.27571458e-01 2.52343506e-01 7.09614754e-01 1.65620983e-01
7.59686708e-01 2.41914108e-01 6.48609102e-01 -5.89622319e-01
9.31021631e-01 3.78162920e-01 5.00503719e-01 5.90229988e-01
3.16697024e-02 -1.36843935e-01 -2.22194508e-01 -5.36000490e-01
-5.36022842e-01 -7.51526177e-01 2.95661271e-01 1.64489281e+00
1.04089119e-01 -7.88656294e-01 -4.08669978e-01 -1.02893496e+00
1.13807581e-01 9.55227315e-01 -6.11286461e-01 -1.85382545e-01
-5.22943079e-01 -7.34701693e-01 -1.87193155e-01 6.14489794e-01
3.35240453e-01 -7.28580177e-01 -3.85208488e-01 5.59773088e-01
-7.15265572e-01 -6.17583752e-01 -1.03909886e+00 -3.06625605e-01
-7.25489438e-01 -1.40356743e+00 -4.01565135e-01 -8.21572423e-01
5.35085618e-01 4.05247927e-01 1.52591109e+00 1.93328649e-01
1.59811765e-01 8.18555653e-01 -8.99730206e-01 -5.68055250e-02
-4.51850802e-01 -1.57825381e-01 -8.78686979e-02 -3.57238442e-01
8.75726998e-01 -1.93411723e-01 -1.29211307e+00 5.65953612e-01
-1.14979291e+00 -2.70607233e-01 1.72565103e-01 4.44535136e-01
9.57005262e-01 7.44628906e-02 1.02595472e+00 -1.08238685e+00
5.72278857e-01 -8.75663161e-01 -5.09642482e-01 6.61907554e-01
-8.33330691e-01 1.59093421e-02 7.18566835e-01 -4.46905732e-01
-8.94028842e-01 1.71779260e-01 -6.18468106e-01 6.91824794e-01
-2.87514180e-01 8.26502204e-01 -4.76675302e-01 2.13163897e-01
8.12350810e-01 -1.24110043e-01 -5.48024714e-01 -6.72296405e-01
1.11768115e+00 3.76580179e-01 4.20255303e-01 -5.91274679e-01
3.60827863e-01 -6.56687934e-03 -3.37594777e-01 -3.02351713e-01
-1.26012075e+00 -7.60761619e-01 -6.36635572e-02 1.53797746e-01
1.09413266e+00 -7.53261983e-01 -9.73620832e-01 -2.40153477e-01
-5.71968436e-01 -1.33177772e-01 -2.72125125e-01 7.33249366e-01
-2.90513963e-01 7.28296876e-01 -6.05898619e-01 -2.22608268e-01
-7.85638392e-01 -7.17636287e-01 7.75846064e-01 1.56491324e-01
-2.75085956e-01 -1.01374149e+00 1.61864132e-01 5.72659314e-01
5.95486343e-01 8.82532969e-02 1.47223437e+00 -7.10376263e-01
-2.39369065e-01 1.00143149e-01 -1.67395607e-01 1.47671625e-01
5.28895080e-01 -5.26602983e-01 -2.70931751e-01 -1.64271414e-01
-9.96670350e-02 -2.21869797e-01 5.99462926e-01 3.72626811e-01
7.60533392e-01 -6.89783275e-01 -1.56068265e-01 -1.04369773e-02
1.59376323e+00 4.36023213e-02 3.58351469e-01 -5.00849485e-02
6.03665113e-01 7.64589608e-01 8.33024442e-01 4.66554403e-01
1.20958364e+00 7.07614362e-01 4.64991480e-01 2.42502734e-01
-2.09199041e-01 -3.32049936e-01 4.57490027e-01 1.12392056e+00
1.77092105e-01 -4.18021381e-01 -1.57429904e-01 4.04916793e-01
-1.88799822e+00 -6.05392456e-01 -3.71189773e-01 2.03280282e+00
1.06626773e+00 -6.06339335e-01 3.03305000e-01 -2.23317184e-02
2.35812351e-01 -4.34139460e-01 -8.59866440e-01 -1.33929372e-01
6.94025978e-02 6.01301938e-02 3.28484148e-01 4.94212955e-01
-1.16438150e+00 7.76122153e-01 5.92646742e+00 3.19604099e-01
-6.15754247e-01 1.81648478e-01 4.69698086e-02 3.42274606e-01
-7.93315709e-01 -5.69352329e-01 -9.00342047e-01 2.44716823e-01
8.94490063e-01 -3.06405704e-02 5.94125092e-01 6.22702003e-01
5.90933263e-02 1.76215336e-01 -1.25939107e+00 4.87577856e-01
1.58533439e-01 -9.65513647e-01 1.91576868e-01 -3.91654611e-01
4.95490074e-01 -2.17213929e-01 -7.92086571e-02 5.66618919e-01
8.50219965e-01 -4.50015485e-01 5.63931286e-01 6.60344779e-01
3.24249655e-01 -4.37620312e-01 6.63285136e-01 3.07146013e-01
-1.36088157e+00 -1.10910982e-01 -7.89830744e-01 9.24356803e-02
4.31367978e-02 6.81711853e-01 -4.08990651e-01 1.08289599e+00
7.59938657e-01 7.76325703e-01 -4.13032323e-01 8.84703577e-01
-2.46130258e-01 5.42923331e-01 -4.69041884e-01 -1.02614596e-01
-5.84585331e-02 -2.80357808e-01 1.28061976e-02 9.48132277e-01
5.95842540e-01 7.01109886e-01 5.72093785e-01 6.59576654e-01
1.37058228e-01 1.08435988e+00 -5.01647703e-02 -7.11084902e-02
-1.06467463e-01 1.14369619e+00 -5.29511154e-01 -3.89852405e-01
-7.32372344e-01 7.91988909e-01 6.54517934e-02 2.88396090e-01
-6.87783539e-01 1.39674723e-01 4.94807243e-01 -2.18979090e-01
5.63685000e-01 3.06952327e-01 5.76095700e-01 -1.31934679e+00
-2.34756351e-01 -1.10709524e+00 1.06325400e+00 -7.16581762e-01
-1.64134145e+00 2.59388179e-01 -5.35807125e-02 -7.58499026e-01
-1.63048759e-01 -3.79953623e-01 5.92724681e-02 5.18048227e-01
-1.82742548e+00 -1.16458786e+00 -2.22136855e-01 9.55911160e-01
1.66258380e-01 5.90356767e-01 1.35194433e+00 8.78257573e-01
-3.19429010e-01 5.47966301e-01 -2.85251200e-01 -3.19967449e-01
8.01061332e-01 -1.34306252e+00 -1.29702613e-01 4.53321815e-01
-2.78670549e-01 8.29814672e-01 7.89655685e-01 -9.56393421e-01
-1.88932943e+00 -1.45776081e+00 1.16924238e+00 -3.52198541e-01
3.28588992e-01 1.12781852e-01 -8.41429412e-01 5.78298390e-01
7.27881938e-02 -4.31675166e-01 1.49060047e+00 6.42898530e-02
-6.05437994e-01 -1.31817788e-01 -1.57486188e+00 4.54061359e-01
1.08750904e+00 -5.99189254e-04 -7.03741312e-01 8.08948338e-01
1.29951704e+00 -1.27435967e-01 -1.45821834e+00 4.03849632e-01
4.32733715e-01 -5.22249579e-01 1.27762473e+00 -8.37939799e-01
1.22031420e-02 -6.78019106e-01 -6.78467572e-01 -1.25425100e+00
-8.07589054e-01 -9.59490761e-02 -2.33011812e-01 9.90517735e-01
2.94078350e-01 -5.00946999e-01 4.61270005e-01 8.05207312e-01
-3.39256048e-01 -4.53397572e-01 -9.33707207e-02 -3.16695958e-01
-2.76344299e-01 -8.73752609e-02 1.24296176e+00 1.12134957e+00
3.45833391e-01 2.71502823e-01 -5.42319953e-01 6.35386288e-01
4.98037785e-01 4.34392035e-01 1.77920356e-01 -1.09559488e+00
-6.34542048e-01 -1.37587517e-01 2.72709608e-01 -1.28664684e+00
-4.15703267e-01 -1.11685658e+00 6.43029809e-02 -2.03783822e+00
1.48585856e-01 -2.56206661e-01 -6.12266898e-01 7.77419806e-01
-3.33486766e-01 2.66560465e-01 -1.48530573e-01 -3.57513428e-01
-9.09668744e-01 1.64014667e-01 1.35243440e+00 -3.83525878e-01
-2.38339081e-01 -8.34396705e-02 -1.39982545e+00 6.16704047e-01
7.90826380e-01 -3.70189518e-01 -8.52735162e-01 -4.38054562e-01
1.14773023e+00 1.40450567e-01 1.00437842e-01 -5.47568142e-01
2.10258797e-01 -3.63468826e-01 -3.44533697e-02 -5.13138473e-01
-2.71637827e-01 -1.07079434e+00 3.08625489e-01 6.46183848e-01
-4.96346712e-01 -1.88418835e-01 -1.07723586e-01 6.81182027e-01
1.91760957e-01 -2.09517315e-01 3.25197470e-03 -2.12068677e-01
-7.26519108e-01 3.27301472e-01 -1.81442887e-01 -1.75313979e-01
5.92333913e-01 3.15959007e-01 -5.08002400e-01 -4.64409851e-02
-1.17343438e+00 6.13884807e-01 1.43849358e-01 5.92074156e-01
4.61661041e-01 -1.38711154e+00 -6.68806672e-01 3.75037566e-02
6.16243184e-01 -6.15275085e-01 6.54286146e-01 6.04201913e-01
-2.46844873e-01 6.51176274e-01 4.26594131e-02 -2.58672256e-02
-9.02116597e-01 1.36001849e+00 2.36204952e-01 -3.50911379e-01
-6.13642752e-01 4.28838074e-01 1.17204070e-01 -7.91522443e-01
4.57240760e-01 -1.27168036e+00 -8.37268949e-01 -6.31536767e-02
8.51420999e-01 2.65647680e-01 1.82536449e-02 -4.30551231e-01
-7.93852508e-02 7.13096440e-01 6.41185120e-02 7.83172548e-01
1.18840635e+00 -5.36721468e-01 -1.32342666e-01 -1.15926616e-01
5.11026263e-01 1.76675901e-01 -4.53418761e-01 -5.58789611e-01
-1.06105044e-01 -1.66883200e-01 -1.76122002e-02 -1.57417917e+00
-1.19103992e+00 2.08756521e-01 7.80880272e-01 5.13777673e-01
1.29493368e+00 -5.83326630e-02 1.28090608e+00 6.90660000e-01
3.57837170e-01 -7.18839228e-01 -2.70684332e-01 8.71560276e-02
8.92787158e-01 -1.14497650e+00 -1.91281199e-01 -9.46793854e-01
-2.73999661e-01 1.05290008e+00 3.58442217e-01 1.49678558e-01
1.12150359e+00 -3.37624490e-01 1.44479394e-01 -2.77105451e-01
-4.51375037e-01 -2.97338873e-01 6.74577177e-01 6.75377369e-01
4.81076777e-01 3.42402279e-01 -7.84750044e-01 1.31466234e+00
-8.98471475e-02 5.60795367e-01 3.68580669e-02 5.40289342e-01
-7.38712430e-01 -1.36784136e+00 -1.43640310e-01 5.19919872e-01
-3.02110344e-01 -3.33868533e-01 6.03252836e-02 2.34017327e-01
5.26831925e-01 1.50851846e+00 -6.21895552e-01 -3.50225031e-01
7.71499574e-01 1.88886467e-02 5.51073194e-01 -7.87923157e-01
-1.04919136e+00 1.07337706e-01 2.73281574e-01 -7.72616327e-01
-7.77440727e-01 -2.10872352e-01 -1.39853358e+00 -2.44283661e-01
-3.78578335e-01 3.89563471e-01 3.32128286e-01 7.94082165e-01
3.88018847e-01 6.17994428e-01 1.15142196e-01 1.78350657e-02
-9.09569681e-01 -7.43440211e-01 -5.89285970e-01 7.37555504e-01
-8.27479586e-02 -4.30922955e-01 8.14559683e-03 2.06507698e-01] | [11.526190757751465, 4.513442516326904] |
5f7ab7e9-b356-428a-aa14-8f9490d6d570 | defending-against-patch-based-backdoor | 2304.01482 | null | https://arxiv.org/abs/2304.01482v1 | https://arxiv.org/pdf/2304.01482v1.pdf | Defending Against Patch-based Backdoor Attacks on Self-Supervised Learning | Recently, self-supervised learning (SSL) was shown to be vulnerable to patch-based data poisoning backdoor attacks. It was shown that an adversary can poison a small part of the unlabeled data so that when a victim trains an SSL model on it, the final model will have a backdoor that the adversary can exploit. This work aims to defend self-supervised learning against such attacks. We use a three-step defense pipeline, where we first train a model on the poisoned data. In the second step, our proposed defense algorithm (PatchSearch) uses the trained model to search the training data for poisoned samples and removes them from the training set. In the third step, a final model is trained on the cleaned-up training set. Our results show that PatchSearch is an effective defense. As an example, it improves a model's accuracy on images containing the trigger from 38.2% to 63.7% which is very close to the clean model's accuracy, 64.6%. Moreover, we show that PatchSearch outperforms baselines and state-of-the-art defense approaches including those using additional clean, trusted data. Our code is available at https://github.com/UCDvision/PatchSearch | ['Liang Tan', 'Hamed Pirsiavash', 'Hamed Firooz', 'Sinong Wang', 'Qifan Wang', 'Maziar Sanjabi', 'Ajinkya Tejankar'] | 2023-04-04 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Tejankar_Defending_Against_Patch-Based_Backdoor_Attacks_on_Self-Supervised_Learning_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Tejankar_Defending_Against_Patch-Based_Backdoor_Attacks_on_Self-Supervised_Learning_CVPR_2023_paper.pdf | cvpr-2023-1 | ['data-poisoning'] | ['adversarial'] | [-1.04481816e-01 5.05499765e-02 -3.63521159e-01 -3.68659832e-02
-1.20900393e+00 -1.17457962e+00 5.78958452e-01 2.91258097e-01
-5.12493789e-01 6.04172111e-01 -2.98211873e-01 -5.08863747e-01
4.12693590e-01 -7.96963632e-01 -1.12450945e+00 -9.41841125e-01
6.22511618e-02 3.39829624e-01 5.48961639e-01 -8.16821381e-02
2.87221253e-01 7.05196500e-01 -9.03257430e-01 3.97560537e-01
5.70872843e-01 5.58740437e-01 -6.77770257e-01 4.65219319e-01
2.57315725e-01 6.72142327e-01 -9.66146648e-01 -5.10029674e-01
6.40825450e-01 -2.81475157e-01 -8.13957751e-01 -4.96460795e-01
6.03326082e-01 -4.45055932e-01 -5.21347284e-01 1.37849736e+00
2.96604931e-01 -3.45973253e-01 2.89003462e-01 -1.55396509e+00
-4.34110314e-01 7.86108434e-01 -5.91043591e-01 7.44088273e-03
1.76828369e-01 5.08029163e-01 5.66796660e-01 -5.74186742e-01
4.85798538e-01 1.12874246e+00 3.78405273e-01 1.03042245e+00
-1.22197628e+00 -1.18115866e+00 -1.02253579e-01 5.22492304e-02
-1.35540605e+00 -4.04765218e-01 5.60259521e-01 -1.79256853e-02
4.48682278e-01 4.23833728e-01 1.46382237e-02 1.55581248e+00
3.04413259e-01 9.29985225e-01 1.39122820e+00 -4.28750932e-01
5.91478407e-01 5.08262515e-01 6.53827310e-01 6.47219419e-01
5.86523473e-01 5.03464162e-01 -4.79819447e-01 -8.78597558e-01
-2.82923821e-02 1.75585435e-03 -2.27091178e-01 -3.16787541e-01
-4.10901815e-01 1.00440848e+00 5.15762150e-01 1.10822104e-01
-3.13998051e-02 -8.47921446e-02 5.23719788e-01 3.52044076e-01
3.19523990e-01 4.52296227e-01 -3.87236178e-01 4.71724957e-01
-6.70124173e-01 3.88083905e-02 1.13859248e+00 4.00502980e-01
6.59001648e-01 -8.32797885e-02 1.79825217e-01 3.04207295e-01
2.69807935e-01 7.97546566e-01 2.32672289e-01 -5.98844767e-01
3.27260405e-01 3.15229625e-01 1.14494199e-02 -7.03683496e-01
1.08330362e-01 -4.67927277e-01 -5.76973915e-01 5.23499012e-01
4.06229615e-01 -1.90494463e-01 -1.01052284e+00 1.80297256e+00
7.83452451e-01 3.28557432e-01 3.28363419e-01 6.56143188e-01
4.29735154e-01 5.31223834e-01 2.34401956e-01 4.74495105e-02
1.19672787e+00 -8.06909919e-01 -2.98431993e-01 -1.24446735e-01
8.84413898e-01 -3.45311999e-01 8.88945580e-01 7.12115884e-01
-7.61055231e-01 -1.83486789e-01 -1.32027090e+00 4.21975106e-01
-6.60673916e-01 -5.42483270e-01 3.16249132e-01 1.22121489e+00
-6.68249547e-01 6.38549805e-01 -1.00021017e+00 -3.15218717e-01
9.80957747e-01 3.89273524e-01 -6.50382221e-01 -2.59346366e-01
-1.39976346e+00 7.15974689e-01 2.18452081e-01 -4.32040691e-01
-1.82713330e+00 -5.69740653e-01 -7.28644192e-01 -1.72546804e-02
4.51100349e-01 -2.51953661e-01 1.18400967e+00 -4.06862885e-01
-9.58978176e-01 7.40690410e-01 1.56244338e-01 -8.35453451e-01
4.86573726e-01 -4.55076963e-01 -3.02373827e-01 3.95124704e-01
1.29999727e-01 2.14027613e-01 1.17229819e+00 -1.61069238e+00
-5.45884430e-01 -5.35189211e-01 1.61196083e-01 -2.93600529e-01
-5.91723561e-01 2.03493565e-01 -3.65665793e-01 -3.62822413e-01
-2.56437093e-01 -1.04985487e+00 -1.94496453e-01 -3.36842030e-01
-8.75136077e-01 5.24899401e-02 1.24152017e+00 -3.86472344e-01
1.06460702e+00 -2.37376904e+00 -4.09707636e-01 4.93921518e-01
1.94635570e-01 9.58306789e-01 -2.57519513e-01 6.46774769e-01
-1.69911042e-01 5.44354677e-01 -2.50460833e-01 -6.21086895e-01
-1.21518783e-01 8.12115446e-02 -9.54071105e-01 9.44321215e-01
-1.33086130e-01 6.05028033e-01 -8.56917679e-01 -1.84342369e-01
4.65156101e-02 2.73170143e-01 -4.45789933e-01 2.15829894e-01
-1.70077845e-01 3.67309153e-01 -4.13252950e-01 7.31011093e-01
9.76773560e-01 7.53922760e-02 3.59350070e-02 6.75004870e-02
2.85810798e-01 2.11146533e-01 -9.27958846e-01 1.21258831e+00
-7.61937052e-02 1.96782857e-01 1.55249134e-01 -4.31357235e-01
6.11181915e-01 2.39431620e-01 5.84463254e-02 -4.12190318e-01
2.64093727e-01 9.63866711e-02 -3.08808774e-01 -1.73317924e-01
-1.40981674e-01 1.05621658e-01 -3.33263546e-01 1.04992127e+00
-8.35111141e-02 1.32166594e-01 -2.36020014e-01 7.74962187e-01
1.59059989e+00 -1.62871048e-01 -8.14197883e-02 -9.04370379e-03
6.45463169e-01 1.19259052e-01 5.18884599e-01 1.15141296e+00
-2.71643758e-01 1.16690479e-01 5.86838126e-01 -3.47571880e-01
-7.85841644e-01 -1.16840696e+00 7.41804391e-02 7.64271975e-01
1.74388602e-01 -7.28640258e-01 -1.33236873e+00 -1.64623737e+00
2.14352012e-01 8.84549618e-01 -8.01685691e-01 -6.96417034e-01
-3.66400599e-01 -4.24615949e-01 1.14509439e+00 4.97523323e-02
7.31724918e-01 -9.66444612e-01 -3.21667880e-01 -3.48602712e-01
1.43821314e-01 -7.28839517e-01 -2.96815574e-01 3.74360144e-01
-6.01562381e-01 -1.49353993e+00 1.18366562e-01 -2.81607419e-01
8.98508191e-01 4.46311146e-01 4.76114392e-01 5.32695413e-01
-1.74925581e-01 2.13830411e-01 -4.20869023e-01 -5.55195212e-01
-9.81269181e-01 1.21699639e-01 3.77616882e-01 7.83687457e-02
5.61384737e-01 -2.71091193e-01 -2.57160544e-01 2.94674456e-01
-1.30042565e+00 -6.29270673e-01 3.08592290e-01 7.20997930e-01
4.82615650e-01 2.91875154e-01 3.24398786e-01 -1.42311013e+00
3.63244355e-01 -6.30099475e-01 -7.23229587e-01 3.47713411e-01
-4.82479066e-01 5.48811592e-02 8.58897686e-01 -6.32272601e-01
-6.48956239e-01 2.09155440e-01 -1.97648391e-01 -8.04325521e-01
-4.28869009e-01 7.22020566e-02 -4.51868892e-01 -3.47770184e-01
1.16391098e+00 1.78537875e-01 5.52837402e-02 -6.44996166e-01
3.40541869e-01 7.39110351e-01 5.42837381e-01 -4.40310866e-01
1.55265474e+00 6.91545784e-01 -1.88246630e-02 -4.46691573e-01
-1.17452562e+00 -2.54092693e-01 -3.60312313e-01 1.23299383e-01
4.87129182e-01 -6.04340494e-01 -5.55321276e-01 7.83982098e-01
-8.82698238e-01 -5.60077310e-01 -1.98017687e-01 5.74160889e-02
-4.84087169e-02 7.09242463e-01 -6.98628545e-01 -8.32578003e-01
-6.55933917e-01 -9.57295537e-01 7.02409744e-01 1.19711891e-01
-3.50501314e-02 -7.95162201e-01 2.67055362e-01 4.87532049e-01
1.97845340e-01 3.62497449e-01 5.72072625e-01 -1.48870993e+00
-3.79435271e-01 -7.00425863e-01 2.81581670e-01 6.78114057e-01
1.17463745e-01 -8.96464735e-02 -1.32854855e+00 -8.28126669e-01
6.39663398e-01 -7.34719098e-01 9.37613726e-01 -2.76252657e-01
1.08454967e+00 -5.32987893e-01 -5.33212423e-01 7.21108198e-01
1.32529151e+00 7.61537701e-02 7.33390391e-01 5.41679978e-01
7.20403969e-01 3.25879574e-01 7.36870706e-01 1.62452832e-01
-1.42019048e-01 1.75658479e-01 7.77496815e-01 -1.07136801e-01
1.53156757e-01 -5.53940237e-01 5.94218612e-01 -9.40064434e-03
8.40174258e-01 -1.68927222e-01 -8.72943640e-01 1.01183817e-01
-1.67523026e+00 -9.54878569e-01 -1.38407379e-01 2.54827762e+00
8.80385339e-01 4.98885214e-01 8.97512957e-02 2.61312366e-01
4.77131158e-01 1.56562068e-02 -7.42186725e-01 -4.14579809e-01
-5.86767718e-02 4.85436618e-01 8.77171755e-01 7.80145824e-01
-1.55655015e+00 1.46890128e+00 5.83537769e+00 1.04234326e+00
-1.07235825e+00 3.35969955e-01 5.39065182e-01 -2.44276643e-01
1.68639398e-03 2.35719129e-01 -1.02074742e+00 4.94051039e-01
1.27297306e+00 -9.45300087e-02 4.82800037e-01 1.02101350e+00
-2.19651058e-01 -1.42339855e-01 -9.94987369e-01 5.54693997e-01
3.29858929e-01 -1.13101614e+00 -1.08640216e-01 3.41597825e-01
4.70884502e-01 3.16184133e-01 1.53399244e-01 3.68897498e-01
8.74264061e-01 -9.25705373e-01 5.33743739e-01 6.91630915e-02
6.02593124e-01 -1.08856881e+00 5.03146589e-01 8.19158256e-01
-5.39934039e-01 -1.92091480e-01 -2.93471754e-01 4.60191190e-01
-2.23083138e-01 3.76101434e-01 -8.23172510e-01 4.12452132e-01
7.93595314e-01 2.09593922e-01 -9.32405770e-01 8.80029321e-01
-8.52533519e-01 1.18187857e+00 -4.97619390e-01 3.32226366e-01
1.76434278e-01 1.90983489e-01 4.95386064e-01 1.09409904e+00
-3.65365326e-01 8.30246434e-02 2.88464993e-01 6.56134307e-01
-2.86370248e-01 -2.85897642e-01 -8.78281236e-01 2.69694589e-02
7.34453082e-01 1.22568655e+00 -4.16839898e-01 -4.12086457e-01
6.25061616e-03 1.09993410e+00 4.50928539e-01 1.75877690e-01
-9.82963622e-01 -3.30398858e-01 6.67783499e-01 -1.64646432e-01
7.96819031e-02 3.69095355e-02 -1.63042426e-01 -1.12995374e+00
-3.13363761e-01 -1.46218371e+00 7.83835649e-01 -3.28993499e-01
-1.38956738e+00 7.08713770e-01 -1.53983340e-01 -1.01697135e+00
6.40701801e-02 -2.89272249e-01 -6.95509791e-01 7.36215770e-01
-1.10082245e+00 -1.18917596e+00 1.13110580e-01 8.33466172e-01
9.66605172e-02 -4.12847579e-01 1.07785749e+00 -1.44051909e-01
-9.55843508e-01 9.35451090e-01 -2.33570822e-02 4.03492063e-01
8.79461586e-01 -1.00005269e+00 7.93593526e-01 1.27613211e+00
4.47148740e-01 1.01217997e+00 5.60593188e-01 -1.09739590e+00
-1.49376285e+00 -1.13120699e+00 5.38583815e-01 -9.17386472e-01
6.82673156e-01 -6.85823560e-01 -1.29969037e+00 7.42113471e-01
1.32763103e-01 1.54970199e-01 9.96958137e-01 -2.62079686e-01
-1.17773998e+00 2.17155702e-02 -1.71451926e+00 5.79918206e-01
5.46564996e-01 -6.61879003e-01 -4.65525478e-01 5.15843213e-01
7.25019693e-01 -3.18892449e-01 -4.12375778e-01 7.45578334e-02
2.60486037e-01 -9.02642310e-01 8.53662908e-01 -8.72514546e-01
-1.54205739e-01 -3.41897517e-01 -5.33882417e-02 -1.04427016e+00
-1.47074446e-01 -8.41012418e-01 -5.11901796e-01 1.22674966e+00
2.56359279e-01 -9.36105490e-01 9.85161602e-01 5.98842502e-01
3.62033278e-01 -5.17820477e-01 -8.42517614e-01 -1.00652885e+00
3.93719018e-01 -3.53700906e-01 4.96209919e-01 1.09543383e+00
-7.80651197e-02 6.28110692e-02 -3.86521667e-01 8.08400214e-01
1.36530018e+00 -2.42938280e-01 1.18841374e+00 -9.38697815e-01
-2.98087209e-01 2.15135142e-01 9.18213800e-02 -4.97142494e-01
3.47163707e-01 -9.20953453e-01 -3.49277794e-01 -9.35332537e-01
1.96598038e-01 -4.24920827e-01 -4.61844355e-01 1.12698328e+00
-2.49700770e-01 4.98990804e-01 3.74771923e-01 4.14350659e-01
-5.40215731e-01 5.51101416e-02 4.84748572e-01 -1.88761085e-01
-9.37114134e-02 1.01055831e-01 -9.32388723e-01 5.82178891e-01
1.39831007e+00 -1.19444275e+00 -2.48442605e-01 -1.10452587e-03
-3.71419251e-01 -5.34595907e-01 4.57211614e-01 -1.03309023e+00
4.01247233e-01 -8.91657993e-02 2.30682880e-01 -3.82023513e-01
1.73861563e-01 -8.82403076e-01 -4.43066144e-03 1.08066046e+00
-2.92673826e-01 -4.26688820e-01 3.18768948e-01 5.04678667e-01
2.66029269e-01 -4.20988113e-01 1.10573673e+00 -4.52874824e-02
-3.19722272e-03 4.31407750e-01 -1.40599534e-01 -1.90478861e-02
1.31989336e+00 1.43447787e-01 -7.98820496e-01 -1.45995900e-01
-5.02828836e-01 2.65827775e-01 9.10258830e-01 1.76881999e-01
5.22364080e-01 -9.40645576e-01 -6.24118030e-01 4.82542038e-01
-3.22720818e-02 -3.11285019e-01 3.69278751e-02 3.35910976e-01
-3.02628815e-01 4.27039303e-02 7.84056708e-02 -2.99239427e-01
-1.68292594e+00 1.15633857e+00 2.48608440e-01 -2.88840145e-01
-4.79185134e-01 8.17818940e-01 -1.18827820e-03 -2.48487577e-01
5.22065997e-01 3.48166853e-01 1.75737619e-01 -3.39405894e-01
8.54937851e-01 1.67501956e-01 -1.01130202e-01 -5.10780573e-01
-6.01203978e-01 1.70036525e-01 -5.34868240e-01 -1.93485796e-01
1.15501869e+00 4.92266446e-01 -1.84553608e-01 -1.38101012e-01
1.25478148e+00 5.54701447e-01 -1.11066079e+00 -3.17623258e-01
-4.54898365e-02 -6.24728084e-01 -8.40574875e-02 -9.42311227e-01
-1.07426023e+00 6.15815401e-01 4.27042574e-01 3.57514441e-01
9.23129141e-01 1.07750468e-01 8.23487043e-01 4.87109154e-01
6.49328649e-01 -6.54606640e-01 6.51643053e-02 2.92626023e-01
5.45488894e-01 -1.11113954e+00 1.03499569e-01 -4.49502945e-01
-6.37417614e-01 5.61781704e-01 6.71555221e-01 -3.91359627e-01
5.82867563e-01 4.67218935e-01 3.55314791e-01 -1.22116603e-01
-8.32532406e-01 3.19730639e-01 -1.47350177e-01 7.60496497e-01
-5.02825856e-01 -2.09936723e-01 1.17726654e-01 4.23057824e-01
-6.55047074e-02 -3.30475688e-01 4.73525524e-01 1.23155367e+00
-3.97672087e-01 -1.55311477e+00 -7.97357082e-01 6.32211566e-02
-8.16710949e-01 7.16710538e-02 -8.11456323e-01 5.41317880e-01
-3.53380144e-02 1.22402596e+00 -5.85715592e-01 -7.56124914e-01
3.39497805e-01 1.31028250e-01 9.86824930e-02 -8.98557842e-01
-1.14137459e+00 -9.07189324e-02 -3.32581960e-02 -8.59315395e-01
1.58390477e-01 -5.42862236e-01 -1.27087963e+00 -4.98743355e-01
-3.80326271e-01 5.42628467e-01 7.80261159e-01 5.89576364e-01
3.16271067e-01 -9.66666192e-02 1.15219319e+00 -4.53168154e-01
-1.05591011e+00 -5.30156910e-01 -4.41944152e-01 5.09175181e-01
4.75817353e-01 -2.49062583e-01 -9.32094753e-01 -2.84889728e-01] | [5.819960594177246, 7.566900730133057] |
49d6ac03-acb8-4bfc-a0d0-da99b45f8664 | one-configuration-to-rule-them-all-towards | 2202.07631 | null | https://arxiv.org/abs/2202.07631v1 | https://arxiv.org/pdf/2202.07631v1.pdf | One Configuration to Rule Them All? Towards Hyperparameter Transfer in Topic Models using Multi-Objective Bayesian Optimization | Topic models are statistical methods that extract underlying topics from document collections. When performing topic modeling, a user usually desires topics that are coherent, diverse between each other, and that constitute good document representations for downstream tasks (e.g. document classification). In this paper, we conduct a multi-objective hyperparameter optimization of three well-known topic models. The obtained results reveal the conflicting nature of different objectives and that the training corpus characteristics are crucial for the hyperparameter selection, suggesting that it is possible to transfer the optimal hyperparameter configurations between datasets. | ['Elisabetta Fersini', 'Raphael Troncy', 'Pasquale Lisena', 'Ismail Harrando', 'Silvia Terragni'] | 2022-02-15 | null | null | null | null | ['topic-models'] | ['natural-language-processing'] | [-1.35196149e-02 5.15131280e-02 -5.32063067e-01 -4.48073655e-01
-8.08396995e-01 -4.86547619e-01 9.13581252e-01 5.10616243e-01
-8.26545954e-02 8.14990520e-01 3.80697638e-01 9.73732844e-02
-4.78265077e-01 -6.48753822e-01 -2.92359591e-01 -1.07007754e+00
-5.24120331e-02 7.66242027e-01 1.18154205e-01 -3.59107591e-02
6.33325756e-01 3.07176024e-01 -1.69778991e+00 4.34243754e-02
9.97203410e-01 7.72059977e-01 4.30047512e-01 4.69724923e-01
-6.02487922e-01 -9.76330042e-02 -1.01811945e+00 -1.79098040e-01
-2.53243834e-01 -2.31939048e-01 -4.80957657e-01 2.05337122e-01
-1.13635913e-01 1.33908182e-01 1.31311417e-01 9.17466998e-01
1.43207446e-01 4.74053413e-01 1.23066115e+00 -1.19822991e+00
-1.52983680e-01 7.56423116e-01 -4.46360648e-01 1.57473698e-01
-5.78410663e-02 -2.65343368e-01 1.26239288e+00 -6.85433269e-01
5.38003445e-01 1.22039855e+00 2.65228581e-02 2.06806734e-01
-1.31672549e+00 -4.87502187e-01 4.18521136e-01 -1.73191220e-01
-1.40496933e+00 -1.96034059e-01 1.00081885e+00 -7.71769643e-01
4.45829719e-01 3.07280302e-01 5.23470521e-01 1.33047795e+00
5.48229814e-01 7.58651376e-01 6.88645065e-01 -4.30557877e-01
6.15544796e-01 7.54235923e-01 4.56443101e-01 -7.75824413e-02
5.27851403e-01 -4.97718722e-01 -7.60354936e-01 -7.42885172e-01
3.99184018e-01 -1.25183821e-01 -4.60906565e-01 -4.70210820e-01
-1.12246954e+00 1.36400545e+00 -3.85462083e-02 6.75902367e-01
-6.12989962e-01 -1.46509320e-01 3.59028190e-01 -8.49121213e-02
5.77907443e-01 8.68611753e-01 -2.66695589e-01 -6.17342852e-02
-9.85966384e-01 3.86244714e-01 7.85129249e-01 1.05001998e+00
6.46734178e-01 -1.16932042e-01 -4.61427540e-01 8.66759300e-01
4.82468754e-01 2.72659332e-01 9.08675313e-01 -2.97377110e-01
5.17804146e-01 5.56450069e-01 3.17797720e-01 -1.04852605e+00
-3.33487242e-01 -5.11929870e-01 -6.35416090e-01 -4.67680782e-01
1.70039281e-01 -1.64046973e-01 -7.88504303e-01 1.55930459e+00
3.33300591e-01 -1.62975311e-01 2.02554569e-01 8.98901641e-01
7.60064125e-01 1.14869797e+00 6.98226631e-01 -4.33738768e-01
1.64592516e+00 -4.34323072e-01 -9.79617953e-01 -2.32178822e-01
5.32758117e-01 -7.37830937e-01 1.03433335e+00 5.55034339e-01
-9.03347313e-01 -2.47893080e-01 -9.13840413e-01 2.69603759e-01
-5.08473516e-01 4.27792013e-01 9.19395506e-01 6.04325950e-01
-5.68874359e-01 1.40925914e-01 -4.90851253e-01 -4.03446823e-01
-7.26067647e-02 4.04159606e-01 6.19877642e-03 3.29037040e-01
-1.27306545e+00 4.92009461e-01 9.96228635e-01 -7.13134632e-02
-8.19952846e-01 -5.32906890e-01 -4.85599458e-01 5.75447917e-01
2.96239108e-01 -5.35612404e-01 1.02954960e+00 -4.60425645e-01
-1.46054840e+00 4.67961848e-01 -3.02782476e-01 -3.95513684e-01
7.88019896e-02 -1.88417464e-01 -1.80333421e-01 6.21210523e-02
-1.43714026e-01 4.62166518e-01 9.55369651e-01 -1.37052250e+00
-7.12709725e-01 -3.57507706e-01 -3.34463656e-01 3.52045596e-01
-8.16744983e-01 -3.32953520e-02 -5.08574724e-01 -7.35376298e-01
2.41327718e-01 -9.73219991e-01 -2.82202572e-01 -6.35044396e-01
-7.97146142e-01 -6.45495772e-01 6.38277471e-01 -3.13434631e-01
1.40404379e+00 -2.02283239e+00 4.07983124e-01 5.27953029e-01
1.42156839e-01 8.33796617e-03 2.91189373e-01 5.22744298e-01
2.03575686e-01 2.02173918e-01 5.55654690e-02 -1.90346345e-01
-2.02574302e-02 -2.43009806e-01 -6.40270770e-01 3.06193680e-01
-5.51699325e-02 3.70089680e-01 -4.28119689e-01 -6.33309484e-01
-6.65093511e-02 4.88613576e-01 -2.05504492e-01 5.11012733e-01
-7.30740130e-01 3.10043722e-01 -1.16654098e+00 2.97206134e-01
4.03973192e-01 -3.57504994e-01 2.62967288e-01 -5.62346838e-02
-8.03172681e-03 4.06809479e-01 -1.07686853e+00 1.37587273e+00
-3.25488240e-01 8.29254746e-01 -1.83775097e-01 -9.99284327e-01
1.36168993e+00 5.69955289e-01 6.33719027e-01 -8.75650644e-02
2.89163858e-01 2.07111202e-02 -2.01379612e-01 -2.54012108e-01
1.05616009e+00 -5.11741228e-02 -1.24649778e-01 6.19161189e-01
1.78370371e-01 -3.05413574e-01 4.50480729e-01 6.12316690e-02
2.92587817e-01 -5.77855408e-01 4.28026766e-01 -6.58396661e-01
3.46132219e-01 5.89438081e-02 2.97157079e-01 6.91890717e-01
4.85158384e-01 4.14757252e-01 7.87122726e-01 -1.03125162e-01
-8.88380408e-01 -7.18058407e-01 -4.86295611e-01 1.15651715e+00
2.72409823e-02 -4.85097408e-01 -5.88426173e-01 -2.24095866e-01
-1.13396369e-01 9.98065472e-01 -5.89521885e-01 -1.97723895e-01
-3.41099590e-01 -9.33847904e-01 1.49680406e-01 2.96484768e-01
-1.71608329e-02 -6.88188374e-01 -6.28595471e-01 4.98340756e-01
-2.51271278e-01 -7.18138278e-01 -3.07789057e-01 3.96438450e-01
-1.19940460e+00 -7.71305382e-01 -1.05384684e+00 -4.28651959e-01
4.99273330e-01 3.06977421e-01 1.02946568e+00 -2.86239624e-01
-6.62396178e-02 1.37336060e-01 -3.30466002e-01 -7.64105976e-01
-3.44846755e-01 8.36287439e-01 -1.01917401e-01 4.77969227e-03
4.70069587e-01 -3.71642888e-01 -4.07108635e-01 3.62546176e-01
-1.08007073e+00 -2.31976852e-01 5.30945599e-01 7.16953397e-01
4.75352228e-01 4.07943457e-01 5.82021832e-01 -9.48164284e-01
1.20484412e+00 -7.94485509e-01 -7.62551606e-01 5.05005598e-01
-7.75474012e-01 1.76237494e-01 3.32191914e-01 -7.15058982e-01
-1.33626878e+00 -4.61009085e-01 3.78462881e-01 -1.89088136e-01
-3.60615879e-01 5.96749723e-01 -2.88309872e-01 5.73492169e-01
5.88971138e-01 4.51516569e-01 -3.45402151e-01 -4.55404222e-01
2.83153117e-01 6.60566330e-01 -2.06880495e-01 -7.79287159e-01
3.52841526e-01 3.19990009e-01 -1.78043678e-01 -1.18498325e+00
-5.93763888e-01 -6.79715276e-01 -2.85492241e-01 -4.12913375e-02
7.55899012e-01 -6.80094838e-01 -3.97539854e-01 1.03179468e-02
-1.18972671e+00 1.62365362e-01 -6.62099645e-02 8.23932230e-01
-5.75994194e-01 -2.60589302e-01 -8.75366926e-02 -9.61924136e-01
-4.88223702e-01 -1.26190317e+00 1.20080590e+00 4.93077517e-01
-5.93311965e-01 -1.25568843e+00 2.09221765e-01 -7.67113194e-02
3.26150626e-01 -1.21351734e-01 1.36973774e+00 -1.13431680e+00
-4.52826291e-01 -1.63115218e-01 8.12874883e-02 -3.07787955e-01
2.04330802e-01 1.60004497e-01 -9.21637952e-01 -3.10849011e-01
1.15942635e-01 6.37441576e-02 7.95777977e-01 9.05436516e-01
1.09851861e+00 -3.97942781e-01 -7.92645335e-01 3.61955673e-01
1.18987930e+00 4.85108286e-01 4.70051199e-01 5.03470361e-01
2.68676318e-02 9.08719480e-01 7.47339785e-01 5.95460057e-01
1.31144738e-02 8.32722962e-01 -6.80629164e-02 1.31238565e-01
6.21934533e-01 -2.99513023e-02 -6.42520338e-02 4.68222290e-01
2.64168441e-01 -7.48310626e-01 -8.67310524e-01 6.71005428e-01
-1.70978701e+00 -6.32055223e-01 -3.14234830e-02 2.16788554e+00
8.57747257e-01 1.86571509e-01 1.66342154e-01 -7.23130032e-02
9.08901215e-01 2.22509861e-01 -4.59168285e-01 -3.37175161e-01
-6.70004040e-02 -3.58217925e-01 1.56840086e-01 1.35758743e-01
-1.08494735e+00 8.80837321e-01 6.58602476e+00 7.98039436e-01
-1.11169660e+00 -2.81983674e-01 7.16345966e-01 -2.32996807e-01
-6.33427739e-01 -3.38950977e-02 -1.27576244e+00 4.76195812e-01
9.43898976e-01 -7.97486007e-01 -2.86916673e-01 9.84669387e-01
3.50205034e-01 -2.78131425e-01 -8.85627449e-01 6.00919604e-01
-1.34682462e-01 -1.18102145e+00 3.21184367e-01 2.77655780e-01
7.20767915e-01 -4.88352239e-01 2.01351508e-01 6.57629874e-03
9.73666385e-02 -8.55221212e-01 4.01637614e-01 3.52992594e-01
1.06751785e-01 -7.48941422e-01 7.28224456e-01 3.22302073e-01
-7.91884243e-01 -2.93513685e-02 -5.86733162e-01 7.84688771e-01
3.79733831e-01 1.06705546e+00 -1.07820046e+00 4.01183516e-01
5.12448847e-01 2.39801511e-01 -2.92924792e-01 1.34779668e+00
-7.42598176e-02 8.35352480e-01 -3.51709992e-01 -3.79217774e-01
3.01352650e-01 -4.40008909e-01 9.25830364e-01 1.21868324e+00
6.24042928e-01 -9.32228100e-03 -5.57812527e-02 1.09762084e+00
2.05114946e-01 5.70206642e-01 -5.90284467e-01 -3.65016311e-01
6.28016770e-01 9.64551032e-01 -9.15349543e-01 -2.63136476e-01
8.35449025e-02 1.75952703e-01 -8.06864500e-02 4.80147421e-01
-4.71377045e-01 -3.27572495e-01 7.83548057e-01 -1.21487446e-01
3.62181664e-01 -2.76794225e-01 -5.36525667e-01 -1.00687766e+00
-1.61569968e-01 -6.95029378e-01 5.68714440e-01 -3.52179527e-01
-1.01813579e+00 7.04896033e-01 6.16442561e-01 -1.14113736e+00
-4.82163578e-01 -2.58413583e-01 -8.04687679e-01 8.40223432e-01
-1.26779866e+00 -7.78784752e-01 -1.15193613e-01 2.49048203e-01
6.85435772e-01 -2.97393829e-01 7.48261809e-01 -4.83670175e-01
-6.06544137e-01 2.99777925e-01 4.38319743e-01 -5.17310083e-01
5.45482337e-01 -1.14808595e+00 -5.16337715e-02 3.21398079e-01
1.13510214e-01 1.14827657e+00 1.13411331e+00 -6.32413030e-01
-1.21040833e+00 -7.66255677e-01 7.10768044e-01 1.16354384e-01
3.89085323e-01 -4.86046046e-01 -1.29316640e+00 3.08470905e-01
1.89526826e-01 -1.09057605e+00 1.08292460e+00 6.06791198e-01
-1.98090207e-02 2.47166138e-02 -8.53360236e-01 6.23999655e-01
1.00172743e-01 3.91046517e-02 -4.99357313e-01 4.47671741e-01
6.80429816e-01 -1.04986755e-02 -9.21213686e-01 -1.10415228e-01
2.52289236e-01 -5.75675428e-01 8.39249969e-01 -7.81496644e-01
1.90738648e-01 2.27153808e-01 -6.50954023e-02 -1.60546601e+00
-4.45272550e-02 -5.68250299e-01 -2.09552303e-01 1.67977822e+00
5.61854661e-01 -6.47767603e-01 8.78366351e-01 9.70741272e-01
3.18944871e-01 -5.70147157e-01 -6.39896333e-01 -6.05230331e-01
2.39557043e-01 1.37721642e-03 8.57637286e-01 6.58555210e-01
1.51083946e-01 4.58275586e-01 -8.58674422e-02 9.42916870e-02
5.92092395e-01 3.68813097e-01 7.35892713e-01 -1.78871155e+00
9.23718363e-02 -6.92188084e-01 -9.67382416e-02 -1.02865922e+00
2.57033646e-01 -4.25680161e-01 1.42017096e-01 -1.28461361e+00
2.58157581e-01 -7.34609306e-01 -9.45514441e-02 -4.59843203e-02
-2.84733295e-01 -6.95883095e-01 -1.94379359e-01 5.25034308e-01
-2.42884681e-01 9.66275871e-01 7.76966989e-01 -1.29864559e-01
-8.03139269e-01 5.10958195e-01 -9.00504947e-01 4.36576903e-01
1.06592441e+00 -7.00596929e-01 -7.43229449e-01 -2.47267798e-01
4.67976406e-02 7.50052929e-03 -2.93247312e-01 -4.82592374e-01
2.23988697e-01 -5.21027386e-01 1.28666326e-01 -7.02489078e-01
5.87876439e-01 -6.99991822e-01 -4.61985804e-02 2.21938151e-03
-7.15864360e-01 -2.54609078e-01 1.70965821e-01 8.42576444e-01
-4.14664924e-01 -6.70779288e-01 5.93673289e-01 -6.57601980e-03
-2.57434517e-01 1.00989282e-01 -6.64008558e-01 -3.92771885e-02
9.78410423e-01 1.47474883e-02 -1.32955372e-01 -5.88827908e-01
-5.09142578e-01 1.55971542e-01 1.57433614e-01 5.75836718e-01
3.50776136e-01 -1.00253081e+00 -6.57420218e-01 -3.10444795e-02
2.41161317e-01 2.40176786e-02 5.93276210e-02 4.68234122e-01
1.31889895e-01 1.11236560e+00 1.46873713e-01 -7.06209302e-01
-1.13500476e+00 3.31233472e-01 -1.92356817e-02 -3.14859599e-01
-5.51920712e-01 5.33743441e-01 5.68272471e-01 1.21589549e-01
3.15877050e-01 -1.73456743e-01 -6.85867965e-01 4.60253090e-01
5.94605803e-01 3.51802081e-01 5.95746227e-02 -2.73159653e-01
-2.91909464e-02 4.21118110e-01 -3.79648149e-01 -2.12828279e-01
1.49929810e+00 -7.73054883e-02 3.76118757e-02 6.18712544e-01
1.05354953e+00 -4.58997428e-01 -9.32720065e-01 -3.76059115e-01
3.63750666e-01 -4.59543467e-01 3.63086313e-01 -1.76257446e-01
-7.29366660e-01 7.44684219e-01 2.61738420e-01 8.08631718e-01
8.88704419e-01 3.16079944e-01 2.80033618e-01 4.57028449e-01
1.89296857e-01 -1.21903098e+00 1.55477244e-02 2.21159652e-01
9.78772044e-01 -1.05267930e+00 2.66581327e-01 -6.36342704e-01
-7.46189117e-01 1.15830934e+00 5.17645001e-01 2.96717703e-01
7.23672211e-01 -6.79821298e-02 -1.28862515e-01 -4.86293554e-01
-9.80037510e-01 9.85505134e-02 6.77374959e-01 4.71476108e-01
6.41240299e-01 4.55654860e-02 -7.23722756e-01 7.07126498e-01
-4.60454255e-01 -5.08378506e-01 2.91308135e-01 6.77380681e-01
-6.27151310e-01 -1.00868261e+00 -6.30042374e-01 5.13377547e-01
-3.06688726e-01 1.88574925e-01 -3.88093740e-01 7.33680785e-01
-6.59138024e-01 1.07192016e+00 6.63023349e-03 2.80988991e-01
8.88285860e-02 3.96865040e-01 -1.07385911e-01 -7.09470570e-01
-2.99948841e-01 3.77206594e-01 -7.25755170e-02 -4.00262587e-02
-6.45437613e-02 -7.16260374e-01 -8.93050671e-01 1.04391046e-01
-9.84687924e-01 9.09835279e-01 1.20773149e+00 8.72484684e-01
3.13030064e-01 3.27723712e-01 6.79546177e-01 -6.11948311e-01
-5.87068260e-01 -1.26490152e+00 -7.65544713e-01 8.09825435e-02
-8.36622119e-02 -9.53223884e-01 -4.63776857e-01 3.58709060e-02] | [10.396595001220703, 6.979869365692139] |
beb38c1b-8f40-4e7c-9a2d-bee9584ef9fb | fooling-ocr-systems-with-adversarial-text | 1802.05385 | null | http://arxiv.org/abs/1802.05385v1 | http://arxiv.org/pdf/1802.05385v1.pdf | Fooling OCR Systems with Adversarial Text Images | We demonstrate that state-of-the-art optical character recognition (OCR)
based on deep learning is vulnerable to adversarial images. Minor modifications
to images of printed text, which do not change the meaning of the text to a
human reader, cause the OCR system to "recognize" a different text where
certain words chosen by the adversary are replaced by their semantic opposites.
This completely changes the meaning of the output produced by the OCR system
and by the NLP applications that use OCR for preprocessing their inputs. | ['Vitaly Shmatikov', 'Congzheng Song'] | 2018-02-15 | null | null | null | null | ['adversarial-text'] | ['adversarial'] | [ 8.36084008e-01 -1.26755506e-01 2.23253950e-01 -2.26557553e-01
-3.03431898e-01 -1.44470274e+00 6.23915970e-01 -2.27414295e-01
-7.61044443e-01 6.53785408e-01 -2.38069281e-01 -5.64609230e-01
6.05030775e-01 -7.76540220e-01 -1.14165473e+00 -6.07493520e-01
6.26517951e-01 2.44405106e-01 9.57162082e-02 -2.16766000e-01
6.92444861e-01 9.44880188e-01 -8.56983244e-01 8.27454448e-01
3.68779510e-01 6.84407711e-01 -1.35648519e-01 1.46301782e+00
-2.23946750e-01 6.36071742e-01 -1.36539531e+00 -8.84687364e-01
4.78581309e-01 -2.44130477e-01 -8.69576514e-01 1.05178088e-01
9.47629869e-01 -5.85599184e-01 -9.88808811e-01 1.63262796e+00
6.83699250e-02 -1.19068012e-01 7.02215612e-01 -8.72656345e-01
-1.75057828e+00 4.67250973e-01 -8.48947018e-02 -1.32450730e-01
7.48678446e-01 3.89262527e-01 4.48194236e-01 -5.62206030e-01
6.86277628e-01 1.34160399e+00 3.64165932e-01 1.00529432e+00
-9.38530743e-01 -4.65164930e-01 -7.26306736e-02 -7.45399818e-02
-1.23373556e+00 -2.90243536e-01 5.17303169e-01 -1.47987038e-01
9.84041512e-01 5.91547906e-01 3.61499667e-01 1.32814646e+00
8.10126364e-01 7.21994698e-01 1.08221257e+00 -6.46795273e-01
1.57371268e-01 3.13611552e-02 1.17531098e-01 6.08522713e-01
3.62482786e-01 -7.04981461e-02 -1.85502306e-01 -6.14054948e-02
5.29330671e-01 -1.88975826e-01 -3.98278832e-01 2.77449451e-02
-1.05863011e+00 5.44375360e-01 3.23208794e-02 -1.66529547e-02
1.62728414e-01 4.09102023e-01 2.63873875e-01 6.81191921e-01
-5.17348468e-04 1.00567758e+00 -2.66459852e-01 1.76402405e-01
-7.24377811e-01 -1.09418266e-01 8.88874054e-01 1.07190919e+00
5.13576865e-01 2.03166321e-01 1.06547691e-01 3.55848372e-01
2.91136801e-01 1.19582736e+00 7.28161097e-01 -5.89089870e-01
3.11932296e-01 4.22422528e-01 1.15106598e-01 -1.03958452e+00
2.28375211e-01 2.64795929e-01 -4.14780945e-01 6.46023631e-01
5.35796642e-01 -1.94837749e-01 -1.60428023e+00 8.07401240e-01
-4.78687465e-01 -4.01803434e-01 5.34199059e-01 8.47358704e-01
6.48431778e-01 8.77308309e-01 -2.16275349e-01 5.45655906e-01
1.10315049e+00 -7.93146431e-01 -8.71752441e-01 -6.30362988e-01
4.66617435e-01 -1.08010876e+00 1.09382522e+00 4.44652766e-01
-8.19382966e-01 -4.45736051e-01 -1.68730819e+00 -3.85983586e-01
-1.19664454e+00 -9.62683465e-03 2.63596892e-01 8.96970034e-01
-9.82549727e-01 5.66763341e-01 -3.13654363e-01 -1.62955001e-01
5.31096041e-01 4.71056551e-01 -6.21869802e-01 -1.95823684e-01
-1.28382027e+00 1.08794260e+00 3.93972456e-01 2.50132412e-01
-8.89513135e-01 -3.09419364e-01 -8.86819482e-01 -7.56332427e-02
1.69376731e-01 2.60302536e-02 1.29394281e+00 -1.65634298e+00
-1.66330945e+00 1.22109246e+00 -1.17311366e-02 -4.40450668e-01
8.35439444e-01 -4.24925834e-01 -7.58151174e-01 4.93898541e-01
-4.31280196e-01 6.43030703e-01 1.50002217e+00 -1.28401673e+00
-2.59325981e-01 -1.36245638e-01 9.99856857e-04 -1.35479853e-01
-4.36590686e-02 2.15878397e-01 -6.17628336e-01 -8.85331392e-01
-2.94943124e-01 -9.99499857e-01 7.24909231e-02 2.81855613e-01
-1.02613854e+00 4.95178223e-01 1.28918529e+00 -7.56880462e-01
7.28856683e-01 -2.27644563e+00 -2.43532568e-01 3.06779861e-01
-7.46430755e-02 7.94919074e-01 -4.96236771e-01 3.07475269e-01
-1.89363301e-01 6.31028473e-01 -1.91730216e-01 -1.15300909e-01
-1.34484932e-01 2.71406025e-01 -8.09436142e-01 7.86739707e-01
2.99541831e-01 1.29095006e+00 -8.05060804e-01 -1.39917150e-01
1.52995646e-01 3.19445461e-01 -1.78666916e-02 3.61947328e-01
-3.23429734e-01 -1.39236346e-01 -2.56728888e-01 6.49141550e-01
7.97511697e-01 1.04160115e-01 1.48432041e-02 1.60962567e-01
1.89515695e-01 1.18565306e-01 -6.58572614e-01 1.03562903e+00
1.73603781e-02 1.35101998e+00 -4.11559701e-01 -5.02070844e-01
9.01968479e-01 8.16521719e-02 -5.71131051e-01 -3.61544102e-01
-9.46539396e-04 1.18952326e-01 7.05282390e-02 -4.20426816e-01
8.89859855e-01 2.09640101e-01 -2.02128798e-01 6.53561115e-01
-3.81196290e-01 -4.67553467e-01 9.26267542e-03 1.83219656e-01
1.06547761e+00 1.81393489e-01 -9.66001004e-02 6.78992122e-02
6.20851159e-01 2.41793171e-01 3.93188708e-02 1.36799026e+00
-1.78027645e-01 8.44999552e-01 4.02534068e-01 -8.00430715e-01
-1.38860488e+00 -1.12249768e+00 1.73550695e-01 7.46700406e-01
1.25113040e-01 2.73546010e-01 -8.95993888e-01 -8.71122301e-01
2.15457276e-01 9.53409493e-01 -9.00925398e-01 -4.45133477e-01
-6.06936753e-01 -2.23060414e-01 1.29456270e+00 4.61147159e-01
5.07943332e-01 -1.37597096e+00 -4.14057165e-01 -1.35291010e-01
3.28823924e-01 -1.29920626e+00 -8.72979224e-01 2.83856511e-01
-5.23823380e-01 -9.78794932e-01 -8.08383107e-01 -9.99712348e-01
1.14859307e+00 1.10792564e-02 6.57098532e-01 3.08020413e-01
-1.39664546e-01 4.83616203e-01 -5.66226840e-01 -4.67540354e-01
-1.34699190e+00 -8.80663320e-02 -2.24913314e-01 6.98335320e-02
5.08533001e-01 5.22723272e-02 -1.53883398e-01 7.86204487e-02
-1.60088015e+00 -2.45455712e-01 5.01942754e-01 2.76584417e-01
2.96599507e-01 1.21586472e-01 -1.94812134e-01 -1.09027946e+00
6.73008680e-01 3.94222617e-01 -5.70708394e-01 5.70287585e-01
-2.42244899e-01 3.12391728e-01 1.08620048e+00 -1.02475119e+00
-7.13842273e-01 7.85382092e-02 2.41969615e-01 -4.26769406e-01
-4.18391675e-01 9.26697685e-05 -4.01216924e-01 -4.83812183e-01
5.60891032e-01 6.99592054e-01 -2.47414991e-01 -2.59968966e-01
3.08566988e-01 9.44473267e-01 1.17477965e+00 -2.56528616e-01
1.27518415e+00 4.69283134e-01 -2.33687967e-01 -8.60830963e-01
-2.79753119e-01 2.00264871e-01 -7.33589232e-01 1.28657788e-01
1.00573969e+00 -4.17325407e-01 -5.89744508e-01 1.48473501e+00
-1.63564193e+00 -3.49471331e-01 -1.88368306e-01 -7.19865337e-02
-2.52732515e-01 6.73315346e-01 -6.38592601e-01 -5.20493329e-01
-1.68274984e-01 -1.28427184e+00 8.72117519e-01 3.50102603e-01
-3.66398022e-02 -8.78312349e-01 -2.50647604e-01 2.96200067e-01
2.53306925e-01 -2.59625092e-02 1.32320881e+00 -1.12124276e+00
-4.10239935e-01 -7.75875092e-01 -1.86962053e-01 8.04750025e-01
3.99073422e-01 5.89409888e-01 -1.01267874e+00 -1.80231869e-01
-3.52732480e-01 -2.17668280e-01 7.28972912e-01 -1.62638798e-01
1.05253923e+00 -5.96811414e-01 -1.74176887e-01 7.65336156e-01
1.33803403e+00 6.88154280e-01 1.46830380e+00 6.37292624e-01
9.32905316e-01 2.05532357e-01 -1.20261654e-01 -2.98638970e-01
-5.13070047e-01 1.63586438e-01 3.43322515e-01 -1.91547602e-01
4.92548794e-02 -4.07769173e-01 7.11833835e-01 4.43401188e-03
6.20404541e-01 -9.57020283e-01 -8.29913080e-01 1.44056916e-01
-1.31315744e+00 -7.70921588e-01 9.00753513e-02 1.96916187e+00
1.04697430e+00 4.59275991e-01 -7.60603011e-01 -5.92882819e-02
1.18087089e+00 4.03371274e-01 -7.08831251e-01 -1.29913604e+00
-5.25776446e-01 3.93800110e-01 1.09080100e+00 6.42132998e-01
-1.31341183e+00 1.48914850e+00 7.81372118e+00 5.38478136e-01
-1.38421154e+00 -7.97444284e-01 5.66049814e-01 2.75409102e-01
-2.25056961e-01 -9.39226747e-02 -6.06932580e-01 3.80232960e-01
8.96336615e-01 -5.55415452e-02 8.46606493e-01 6.13466024e-01
-1.29190283e-02 -4.41004848e-03 -1.38597596e+00 4.64266211e-01
3.80232990e-01 -1.34781539e+00 8.34416449e-01 -1.60692248e-03
6.30702436e-01 -3.97940040e-01 4.59559768e-01 -1.65476590e-01
4.95192975e-01 -1.46530473e+00 7.67996609e-01 7.70707011e-01
1.01141346e+00 -5.59667170e-01 7.82003641e-01 -2.38305509e-01
-2.89546162e-01 1.71567753e-01 -5.31072080e-01 8.03235397e-02
-4.20803696e-01 8.32981914e-02 -7.07297206e-01 -1.59224004e-01
2.85568446e-01 4.38751653e-02 -1.00079274e+00 4.62858230e-01
-6.12589538e-01 3.33821476e-01 -3.41739953e-02 -3.54140043e-01
4.18341815e-01 1.18407525e-01 6.58103824e-01 1.49130583e+00
-2.60641158e-01 -1.62188277e-01 -2.99092561e-01 9.26504850e-01
-7.63093650e-01 -7.05853030e-02 -8.48141134e-01 -8.94422174e-01
1.62574232e-01 7.36750185e-01 -6.34750366e-01 -4.70860660e-01
-2.08541304e-01 1.78577340e+00 -2.03828141e-01 6.60423279e-01
-6.01957083e-01 -1.08408654e+00 7.38958895e-01 -1.63519472e-01
4.45214123e-01 -3.64494264e-01 -2.18608886e-01 -8.86215866e-01
-2.46875249e-02 -1.40461731e+00 -1.07669510e-01 -1.49507248e+00
-1.27405465e+00 4.76073325e-01 -6.34554148e-01 -7.68192172e-01
7.42741674e-02 -1.37438536e+00 -6.75350189e-01 9.82873559e-01
-1.18298614e+00 -9.16461945e-01 1.99969977e-01 4.36676025e-01
4.68158513e-01 -5.06563246e-01 8.51859212e-01 -4.31926697e-01
-2.97830999e-01 1.06113982e+00 3.79325509e-01 1.12606299e+00
7.00691760e-01 -1.26638544e+00 1.11512423e+00 1.43937600e+00
2.43920609e-01 5.85557401e-01 8.41508150e-01 -9.41646159e-01
-1.68734491e+00 -9.43694413e-01 8.03441644e-01 -6.72978878e-01
7.23355770e-01 -5.03522217e-01 -1.02445281e+00 7.72569418e-01
4.19065982e-01 -3.76028828e-02 5.05291879e-01 -8.20592761e-01
-8.77870560e-01 3.22713703e-01 -1.25411677e+00 9.89215076e-01
3.51253569e-01 -1.25403678e+00 -9.37472105e-01 2.84944475e-01
9.71134543e-01 -6.06763482e-01 -1.98513404e-01 -1.43327251e-01
8.93245816e-01 -3.77067536e-01 6.31555915e-01 -1.42857218e+00
7.73586214e-01 -4.13772225e-01 -3.48482758e-01 -1.09955645e+00
1.05949067e-01 -9.17378187e-01 7.70062581e-02 6.64400995e-01
6.91499174e-01 -7.91096926e-01 6.64948046e-01 9.28780138e-01
1.59265682e-01 7.28447363e-02 -7.31326044e-01 -7.03261971e-01
3.52116376e-01 -3.63802314e-01 5.64705431e-01 8.23083103e-01
-2.18408659e-01 -2.58584768e-01 -4.27974194e-01 6.83459699e-01
1.46399081e-01 6.23260951e-03 6.20581031e-01 -5.24513006e-01
-1.78091586e-01 -2.62155831e-01 -7.31064618e-01 -1.13218260e+00
2.42698386e-01 -5.65715849e-01 3.22825462e-01 -9.14712846e-01
-1.55825898e-01 3.24589491e-01 -2.77322173e-01 5.71005344e-01
-1.07140698e-01 5.46000242e-01 4.64486152e-01 2.50534117e-01
-3.89813215e-01 9.09908786e-02 1.21170652e+00 -7.70633399e-01
3.10146492e-02 -2.60317177e-01 -5.62837243e-01 8.67212236e-01
7.85644472e-01 -8.16593051e-01 1.37995481e-01 -7.77557790e-01
4.47965056e-01 -5.01314938e-01 3.19309235e-01 -8.76206279e-01
1.57486603e-01 -2.23315656e-01 1.02434456e+00 -1.97356731e-01
-3.79573517e-02 -8.99798810e-01 -3.03741992e-01 5.17935634e-01
-9.18086588e-01 3.33813488e-01 3.99824351e-01 3.33838344e-01
1.79704070e-01 -8.56686771e-01 8.76338542e-01 -1.71511486e-01
-5.88438511e-01 -1.81411400e-01 -1.11591947e+00 -7.98817575e-02
5.44583499e-01 -5.72536349e-01 -7.83565938e-01 -5.94786704e-01
-4.57351744e-01 -3.21192414e-01 9.76056516e-01 5.19874811e-01
7.87733614e-01 -7.99100518e-01 -4.70186442e-01 3.34923893e-01
-2.22731829e-02 -3.90710145e-01 -2.78344840e-01 -3.12631220e-01
-1.24784160e+00 4.94543344e-01 -3.46841872e-01 1.12412535e-02
-1.15450871e+00 7.20361471e-01 6.92538202e-01 7.05377460e-02
-5.23804069e-01 6.18413985e-01 -6.54348210e-02 -1.39301032e-01
1.23767763e-01 -2.02760026e-01 2.02682033e-01 -4.82188255e-01
8.28242958e-01 -4.52613048e-02 1.20450538e-02 -4.41152006e-01
-2.61037886e-01 3.94330561e-01 -5.39327741e-01 -2.32461676e-01
9.59459424e-01 1.67516574e-01 4.65781603e-04 1.67411938e-01
1.18613362e+00 5.11492431e-01 -1.17536461e+00 8.25141091e-03
-2.69462436e-01 -5.17539561e-01 -7.82113597e-02 -1.29853451e+00
-9.83654141e-01 9.74103451e-01 6.49129212e-01 9.76154357e-02
8.69617105e-01 -4.77738976e-01 9.54415798e-01 1.28867400e+00
-7.89298341e-02 -1.22468853e+00 5.50948046e-02 7.75834203e-01
8.14439774e-01 -1.05781007e+00 7.16591477e-02 -3.94313876e-03
-6.02844119e-01 1.72039711e+00 4.26489502e-01 -2.83604503e-01
1.34893045e-01 5.28825223e-01 6.85016572e-01 1.18048437e-01
-2.54102498e-01 2.63756096e-01 2.80062973e-01 9.96948123e-01
-5.03090210e-02 -1.85529783e-01 1.77171603e-01 1.29979020e-02
-1.19430766e-01 -2.59810388e-01 1.30440831e+00 1.05304897e+00
-2.21230522e-01 -1.12062979e+00 -7.04951346e-01 7.52437562e-02
-7.66042650e-01 -5.65539479e-01 -1.18332696e+00 8.01926196e-01
-1.16752304e-01 8.44627380e-01 2.04146370e-01 -4.23774928e-01
1.21889345e-01 4.95386183e-01 3.64519149e-01 -5.48136234e-01
-6.53558314e-01 -5.43621361e-01 -8.76772851e-02 -2.38771930e-01
1.59586117e-01 -2.30812207e-01 -1.38509309e+00 -4.04374301e-01
1.22111179e-01 -2.72298127e-01 8.26210260e-01 1.14930284e+00
1.54467404e-01 1.17095403e-01 6.65348589e-01 -3.91937494e-01
-5.73746026e-01 -6.39405370e-01 -4.05046910e-01 4.11612839e-01
5.96698999e-01 2.70753503e-01 -6.07759118e-01 6.35134399e-01] | [5.827850341796875, 8.031594276428223] |
8dde5546-5c8b-4549-93d2-6321d44d592f | recent-advances-in-direct-speech-to-text | 2306.11646 | null | https://arxiv.org/abs/2306.11646v1 | https://arxiv.org/pdf/2306.11646v1.pdf | Recent Advances in Direct Speech-to-text Translation | Recently, speech-to-text translation has attracted more and more attention and many studies have emerged rapidly. In this paper, we present a comprehensive survey on direct speech translation aiming to summarize the current state-of-the-art techniques. First, we categorize the existing research work into three directions based on the main challenges -- modeling burden, data scarcity, and application issues. To tackle the problem of modeling burden, two main structures have been proposed, encoder-decoder framework (Transformer and the variants) and multitask frameworks. For the challenge of data scarcity, recent work resorts to many sophisticated techniques, such as data augmentation, pre-training, knowledge distillation, and multilingual modeling. We analyze and summarize the application issues, which include real-time, segmentation, named entity, gender bias, and code-switching. Finally, we discuss some promising directions for future work. | ['Jingbo Zhu', 'Tong Xiao', 'Mingxuan Wang', 'Tom Ko', 'Chengqi Zhao', 'Qianqian Dong', 'Rong Ye', 'Chen Xu'] | 2023-06-20 | null | null | null | null | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 2.45275810e-01 5.79032786e-02 -5.21955311e-01 -5.61787486e-01
-1.36136484e+00 -3.52041721e-01 6.10521197e-01 -2.88823575e-01
-3.55712891e-01 1.00253224e+00 2.49647081e-01 -6.76854253e-01
4.69038576e-01 -1.18920483e-01 -4.86181706e-01 -4.61487234e-01
5.05255997e-01 8.02809834e-01 -5.55428304e-02 -2.79793918e-01
1.30991668e-01 3.38882655e-02 -1.07309282e+00 3.88804942e-01
1.40286922e+00 5.17848611e-01 2.86828876e-01 3.67354184e-01
-5.49695671e-01 5.17488122e-01 -8.11066866e-01 -9.22800362e-01
-4.64682952e-02 -5.77287257e-01 -7.86887705e-01 6.66019320e-02
1.97310373e-01 -6.42149821e-02 -9.19775665e-02 9.89211202e-01
1.06584430e+00 2.77986079e-02 4.76329714e-01 -1.26812220e+00
-7.55490482e-01 8.90303075e-01 -4.69509840e-01 2.56430835e-01
2.93071508e-01 -4.22398984e-01 5.61737299e-01 -1.19676745e+00
5.70174277e-01 1.27051806e+00 4.88769650e-01 7.92667627e-01
-6.35198712e-01 -7.75139272e-01 3.49995553e-01 3.27438742e-01
-1.40375328e+00 -1.01874173e+00 5.54153800e-01 -3.73631835e-01
1.10469186e+00 2.42102906e-01 1.37932107e-01 1.40029299e+00
-1.50974050e-01 1.10305738e+00 1.42181039e+00 -8.54551733e-01
-1.94327056e-01 4.44767028e-01 -2.26199895e-01 4.80242223e-01
-2.47758850e-01 -1.20246708e-01 -7.27845848e-01 -1.09238632e-01
5.30521810e-01 -6.19422197e-01 -8.46146047e-02 3.48697603e-02
-1.26618552e+00 5.34154177e-01 -4.08622742e-01 4.30969238e-01
1.13848992e-01 -4.55259681e-01 5.84399939e-01 2.61037350e-01
7.63563395e-01 1.36667624e-01 -8.31768811e-01 -5.81480861e-01
-1.13785374e+00 6.69855345e-03 8.06448460e-01 1.49498940e+00
4.18805212e-01 3.96895617e-01 -1.93123937e-01 1.29292417e+00
3.21863323e-01 7.28799105e-01 6.91455483e-01 -3.27886462e-01
1.24133432e+00 1.40239611e-01 -2.52089053e-01 -4.54401493e-01
-8.26350227e-02 -3.58401835e-01 -7.68170595e-01 -8.69346857e-01
4.16853502e-02 -6.05660081e-01 -8.59050691e-01 1.60528731e+00
2.85311610e-01 8.36300403e-02 1.63324520e-01 5.49297273e-01
1.02189934e+00 6.91534162e-01 1.77275389e-01 -4.80178237e-01
1.37512815e+00 -1.31860459e+00 -1.34030688e+00 -4.40297931e-01
6.41047955e-01 -1.30203533e+00 8.99979293e-01 -1.91838175e-01
-1.21917570e+00 -5.84717870e-01 -5.64877689e-01 -2.71864176e-01
-5.72934330e-01 6.59033716e-01 4.97954398e-01 9.12001312e-01
-9.10463274e-01 9.31102410e-02 -9.67674315e-01 -5.20714045e-01
8.88652820e-03 4.68819171e-01 -1.22970060e-01 1.39483377e-01
-1.49003470e+00 1.11701977e+00 2.50520587e-01 2.67330166e-02
-5.70455372e-01 -3.80426526e-01 -9.04155016e-01 -2.59556770e-01
2.92432189e-01 -5.98656654e-01 1.53619623e+00 -8.03676844e-01
-1.80964673e+00 9.34895635e-01 -6.85836375e-01 -7.64873810e-03
3.90374243e-01 -1.54335424e-01 -8.75121713e-01 -5.07244110e-01
1.16318800e-01 2.85584241e-01 4.48257864e-01 -1.02624500e+00
-8.87680113e-01 -5.74877024e-01 -3.14607650e-01 5.58915675e-01
-4.66728538e-01 9.17177200e-01 -8.59480441e-01 -9.86764789e-01
-1.95282429e-01 -9.84799922e-01 -1.61995925e-02 -7.59192228e-01
-5.21202385e-01 -1.95986807e-01 5.35739541e-01 -1.17809629e+00
1.78567207e+00 -2.16641235e+00 3.08205187e-01 -4.27869439e-01
-3.41579229e-01 4.96091872e-01 5.46177626e-02 6.26183331e-01
8.49999264e-02 1.75536618e-01 -2.19167247e-01 -8.25180531e-01
-9.62987915e-02 2.31360525e-01 -2.31746525e-01 1.28768414e-01
4.11508530e-02 9.82841671e-01 -5.50742745e-01 -7.34176815e-01
-3.37990820e-02 3.46580833e-01 5.69377514e-03 3.46134871e-01
8.87460187e-02 7.20755935e-01 -4.07503188e-01 1.07730341e+00
6.83653593e-01 2.40881190e-01 1.65657505e-01 1.33980453e-01
-4.70541984e-01 8.06342840e-01 -9.06234801e-01 1.75064492e+00
-6.79790735e-01 4.49629188e-01 3.83879691e-01 -9.07451749e-01
9.59819734e-01 7.31114745e-01 1.77346528e-01 -6.91219985e-01
1.61574170e-01 8.04958642e-01 -6.25359267e-02 -6.83072209e-01
7.32069314e-01 -1.02627762e-01 -8.13213289e-02 1.36067122e-01
1.64068952e-01 1.13558970e-01 6.25723675e-02 -2.53226221e-01
3.51040781e-01 2.70396173e-01 2.74507701e-01 -9.12588164e-02
5.31348467e-01 6.53406158e-02 8.04608762e-01 2.44095013e-01
-2.25808278e-01 4.84783620e-01 5.54559082e-02 8.88195261e-02
-9.49580431e-01 -6.94057405e-01 -2.92455107e-02 1.46183705e+00
-1.19778536e-01 -2.76804119e-01 -1.02750075e+00 -8.03174436e-01
-3.62031966e-01 5.73505461e-01 -3.14245760e-01 1.64785206e-01
-8.50408852e-01 -1.08898771e+00 9.08814669e-01 5.29645801e-01
6.49103045e-01 -7.57877290e-01 3.63580674e-01 2.54419386e-01
-8.50752890e-01 -1.47359157e+00 -5.30669153e-01 4.91224863e-02
-1.08631086e+00 -3.83575678e-01 -1.04004776e+00 -1.32477295e+00
5.61375082e-01 1.36890411e-01 1.05836165e+00 -3.90135378e-01
3.43003809e-01 -1.30575538e-01 -5.29246509e-01 -4.94513720e-01
-4.89930511e-01 7.31725991e-01 2.99568884e-02 -1.47216350e-01
4.95606720e-01 -3.50206196e-01 -1.05618602e-02 4.12471682e-01
-3.45093459e-01 2.37047091e-01 1.01929903e+00 8.16549122e-01
5.36239862e-01 -6.16003156e-01 6.71265483e-01 -1.14586425e+00
8.09436142e-01 -4.69337225e-01 -1.91277429e-01 7.04149783e-01
-6.10363245e-01 -9.37341303e-02 4.85504359e-01 -2.99207330e-01
-1.50333965e+00 -3.62314768e-02 -5.04850388e-01 -9.30853486e-02
-2.39829093e-01 6.74815238e-01 -5.62803805e-01 -3.56252864e-02
2.04200178e-01 5.70915997e-01 -1.97630912e-01 -9.79006410e-01
3.99138063e-01 1.53113961e+00 2.93972015e-01 -8.45546246e-01
5.38999021e-01 -1.25531793e-01 -7.41501272e-01 -7.53825784e-01
-7.04616249e-01 -5.58048904e-01 -7.75737405e-01 8.29274207e-02
7.02220559e-01 -1.17367339e+00 1.90566890e-02 6.06763482e-01
-1.45357001e+00 -2.21730009e-01 1.21373802e-01 6.71140671e-01
-3.63506913e-01 2.52871573e-01 -7.98987448e-01 -7.77276456e-01
-4.67160314e-01 -1.43104339e+00 1.13200057e+00 1.09465607e-01
2.54178774e-02 -1.00636244e+00 3.98011617e-02 9.33706760e-01
5.01002729e-01 -1.18812166e-01 9.81831133e-01 -9.51891899e-01
-1.96124166e-01 1.44950151e-01 -2.76990980e-02 2.04697117e-01
1.18425429e-01 -2.60916017e-02 -8.29776287e-01 -1.74511984e-01
-5.21144345e-02 -2.00430751e-01 2.61351258e-01 2.32700825e-01
7.48766959e-01 -2.76548207e-01 -4.86405522e-01 5.72888911e-01
8.69577110e-01 4.90154386e-01 3.33234280e-01 1.97089791e-01
8.44232142e-01 7.66223133e-01 6.70253813e-01 1.72229856e-01
1.01354909e+00 9.28801060e-01 -3.07409406e-01 -3.26947212e-01
-2.19119251e-01 -3.36419523e-01 4.61428881e-01 1.93216276e+00
-2.31961071e-01 -2.56607294e-01 -9.52669561e-01 6.34408236e-01
-1.88044500e+00 -3.53420556e-01 -2.49571413e-01 2.02056336e+00
1.00499904e+00 -2.63616830e-01 1.84432730e-01 -2.53550857e-01
1.06857681e+00 9.38782096e-02 -2.46357977e-01 -6.38935983e-01
-2.39204317e-01 -2.04197671e-02 4.70019668e-01 5.49218297e-01
-9.56540942e-01 1.49319160e+00 6.96921015e+00 1.15189731e+00
-1.12142956e+00 6.00589335e-01 7.11979270e-01 3.02060634e-01
-1.76099971e-01 7.88522884e-02 -1.27305639e+00 6.41159296e-01
1.04227102e+00 -1.69123963e-01 4.21539724e-01 7.88512528e-01
1.13797635e-01 2.09314033e-01 -8.30901742e-01 1.02060008e+00
5.07976830e-01 -8.70869577e-01 8.73007327e-02 2.31592003e-02
7.35634089e-01 3.00856948e-01 1.22183912e-01 6.41738296e-01
4.29214574e-02 -6.53150797e-01 5.31036735e-01 -3.93481292e-02
1.05756652e+00 -7.16124296e-01 7.48052597e-01 5.35512388e-01
-1.15920460e+00 2.10701823e-01 -1.27872318e-01 -1.23904245e-02
5.29333532e-01 4.73390609e-01 -5.31745553e-01 1.08888876e+00
3.66660833e-01 5.42360306e-01 -4.61177707e-01 7.97081411e-01
-2.57611066e-01 6.04723632e-01 -6.25765473e-02 -1.28929093e-02
5.16426228e-02 -2.72550702e-01 3.00919026e-01 1.52769005e+00
4.67851758e-01 -1.73789769e-01 3.70983839e-01 2.17352137e-01
-1.14198118e-01 5.42769969e-01 -2.76275337e-01 -1.67297348e-01
7.54449844e-01 9.13204134e-01 -3.33517820e-01 -5.59870660e-01
-8.78648758e-01 1.05110204e+00 2.74754018e-01 4.17343825e-01
-7.36707568e-01 -3.70073497e-01 5.22933424e-01 -9.61815864e-02
-1.18535534e-02 -5.23460805e-01 -4.67297435e-01 -1.53131771e+00
1.99169949e-01 -1.37968111e+00 2.25601539e-01 -3.80219609e-01
-9.21204805e-01 8.81810963e-01 2.13896278e-02 -9.74834919e-01
-3.53766441e-01 -3.61974597e-01 -2.15595558e-01 1.01793540e+00
-1.78723419e+00 -1.32884848e+00 2.09537700e-01 5.15871644e-01
1.01331890e+00 -5.22126853e-01 8.90727520e-01 1.01065922e+00
-8.92065525e-01 1.03649950e+00 3.45045835e-01 3.50394249e-01
1.09146202e+00 -9.01875615e-01 7.37075210e-01 8.62718701e-01
6.91831186e-02 6.57201171e-01 3.87335598e-01 -6.65695429e-01
-1.27124536e+00 -9.57019746e-01 1.81745791e+00 -4.34698164e-01
5.07021189e-01 -7.35848248e-01 -6.44806504e-01 8.06078255e-01
4.86781657e-01 -3.74815971e-01 9.12492931e-01 2.83201784e-01
2.17784308e-02 3.72718349e-02 -8.31900358e-01 6.40958488e-01
9.56293821e-01 -4.55440402e-01 -4.65004146e-01 3.65686268e-01
6.35301352e-01 -8.08470309e-01 -8.19488168e-01 4.34284091e-01
3.30833405e-01 -4.23714787e-01 5.27297795e-01 -6.22832835e-01
1.37229159e-01 8.68267715e-02 -1.67213738e-01 -1.40413415e+00
-2.36780569e-02 -1.01092768e+00 -1.30332589e-01 1.75963867e+00
7.75220096e-01 -4.65936512e-01 4.95293081e-01 5.66960156e-01
-5.36014557e-01 -8.88580084e-01 -1.05376828e+00 -8.26403916e-01
2.45096982e-01 -2.14742810e-01 6.74757481e-01 1.25264812e+00
1.04164146e-01 7.91026115e-01 -6.72415078e-01 -7.63615593e-02
1.39559984e-01 4.90730740e-02 6.05758071e-01 -8.03661466e-01
-1.79395616e-01 -2.10994601e-01 1.63140059e-01 -1.60429251e+00
2.36223280e-01 -7.72088468e-01 2.72217263e-02 -1.54590225e+00
1.47112474e-01 -4.08956587e-01 9.40674394e-02 5.21625876e-01
-3.16462457e-01 -9.98866707e-02 1.89784229e-01 2.47248933e-01
-4.87371504e-01 6.66723788e-01 1.23108709e+00 4.43210304e-02
-1.97534710e-01 2.89292097e-01 -6.85430467e-01 4.17567492e-01
9.29146051e-01 -5.66042244e-01 -1.52758822e-01 -1.09783614e+00
-1.60792582e-02 3.87428969e-01 -5.51270127e-01 -5.69072664e-01
2.44883150e-01 -2.30906516e-01 -5.01579046e-02 -5.39637387e-01
3.59120727e-01 -5.70128679e-01 -1.24724820e-01 9.87550169e-02
-1.83291718e-01 5.62168360e-01 1.36525065e-01 5.10017462e-02
-6.66138172e-01 -1.07054345e-01 4.97443229e-01 -1.53310314e-01
-4.77460414e-01 3.21360648e-01 -6.08023703e-01 3.62861812e-01
7.55897701e-01 6.02280768e-03 -1.18924588e-01 -2.03174472e-01
-6.31661654e-01 3.24237287e-01 -4.00954857e-02 8.99038792e-01
2.06805006e-01 -1.27424896e+00 -9.13935423e-01 1.76901475e-01
1.86939359e-01 -3.76307309e-01 4.32452336e-02 9.97931004e-01
-1.18022434e-01 6.96967065e-01 4.93502058e-02 -2.02315196e-01
-1.36295104e+00 2.49666929e-01 1.52963579e-01 -3.73049051e-01
1.33131787e-01 8.37119460e-01 -1.52455151e-01 -1.05723238e+00
3.20033491e-01 -5.80148920e-02 -2.24113062e-01 8.69852304e-02
1.29609808e-01 5.49198270e-01 3.98944318e-01 -9.65189278e-01
-5.95953643e-01 4.48645860e-01 -2.08621308e-01 -4.00996566e-01
8.04260612e-01 -5.64012527e-01 -1.68996885e-01 5.13446569e-01
9.65315938e-01 4.75242622e-02 -5.20622730e-01 -3.65183651e-01
2.12853655e-01 -3.66296440e-01 -2.13105127e-01 -1.12593639e+00
-9.47634697e-01 1.24600589e+00 3.21720719e-01 -1.77623853e-01
8.75499964e-01 -1.66590407e-01 1.09004831e+00 2.04492569e-01
3.84257853e-01 -1.63638639e+00 -4.63566542e-01 1.06625736e+00
5.07866681e-01 -1.31347179e+00 -3.73327315e-01 -7.08034635e-01
-8.36011469e-01 7.40129769e-01 7.10905373e-01 6.80368006e-01
6.19499266e-01 4.34161216e-01 5.33870816e-01 3.49545568e-01
-4.70110118e-01 -1.48344874e-01 1.67498618e-01 7.25936890e-01
1.04859293e+00 1.37413993e-01 -6.76916540e-01 8.08610499e-01
-4.02941316e-01 -1.37380376e-01 7.95074403e-02 7.58238912e-01
-1.20616786e-01 -1.71529877e+00 -3.31659079e-01 1.38945699e-01
-7.97126174e-01 -5.28416514e-01 -5.75093210e-01 6.41248465e-01
1.60311610e-01 1.10964131e+00 -1.40802413e-01 -4.08227444e-01
4.83664215e-01 4.50461119e-01 2.80828595e-01 -8.53414834e-01
-6.50178075e-01 3.54298145e-01 4.63473439e-01 -4.29164246e-02
-5.26106060e-01 -7.41758049e-01 -7.98262179e-01 -1.40951425e-01
-7.09484339e-01 4.74923939e-01 1.11552238e+00 1.10086441e+00
6.87745988e-01 4.62355316e-01 4.21418041e-01 -4.24394280e-01
-4.51912671e-01 -1.29840171e+00 -2.12702736e-01 -1.68519601e-01
1.64621007e-02 -4.58782911e-01 -7.74336979e-02 2.12713555e-01] | [14.480711936950684, 7.267546653747559] |
30918fcc-8ab3-49dd-b1fb-4ab1eb09ed3f | an-empirical-investigation-of-cash-conversion | 2005.09482 | null | https://arxiv.org/abs/2005.09482v1 | https://arxiv.org/pdf/2005.09482v1.pdf | An Empirical Investigation of Cash Conversion Cycle of Manufacturing Firms and its Association with Firm Size and Profitability | The purpose of this empirical study is to investigate Cash Conversion Cycle of thirty manufacturing firms listed in Dhaka Stock Exchanges under six different categories, which are, Food and allied, Pharmaceuticals and chemical, Cement, Textile, Engineering and Miscellaneous. This paper sets industry average Cash Conversion Cycle for these six industries and examines the relationship of Cash Conversion Cycle with firm size and profitability. This study did not find statistically significant differences among the Cash Conversion Cycle of varying manufacturing industries. The result of this study indicates a statistically significant negative relationship between the Cash Conversion Cycle and profitability, especially in terms of Return on Equity. The result also shows that the Cash Conversion Cycle of manufacturing firm also has significant negative relationship with firm size, when measured in terms of net sales. The present study contributes to the literature on working capital management written in the context of Bangladesh. | ['Nusrat Jahan'] | 2020-05-15 | null | null | null | null | ['miscellaneous'] | ['miscellaneous'] | [-6.00563049e-01 1.16804995e-01 -4.81115192e-01 2.50910282e-01
2.31832802e-01 -6.45695686e-01 4.76401836e-01 9.13901925e-02
-1.20303892e-01 8.92350793e-01 3.49252194e-01 -6.82012379e-01
-3.85533690e-01 -1.10356140e+00 -1.37469769e-01 -6.56554163e-01
2.28761226e-01 2.81272158e-02 -3.15648198e-01 -2.34228566e-01
1.29561496e+00 6.90093338e-01 -4.29344535e-01 -3.93964171e-01
5.23348272e-01 8.06181073e-01 3.88577551e-01 1.72908381e-01
3.47623490e-02 1.06766462e+00 -3.51749867e-01 -3.74360651e-01
4.46229994e-01 -5.94906926e-01 -6.08045399e-01 2.95812935e-01
-2.71566868e-01 -5.92953026e-01 2.47195587e-01 5.15456319e-01
-8.80838037e-02 -5.25035784e-02 7.57307708e-01 -1.18375754e+00
-1.14486742e+00 8.06262136e-01 -8.95569384e-01 4.30611700e-01
-1.54961953e-02 1.02344425e-02 8.40656400e-01 -9.60612297e-01
5.29581845e-01 1.30948806e+00 3.74949455e-01 -2.58110315e-01
-4.46911097e-01 -1.06675851e+00 -4.03389513e-01 -3.48309487e-01
-7.67395556e-01 -1.79706439e-01 1.69121727e-01 -4.75512505e-01
1.26132917e+00 -1.87670171e-01 6.47289932e-01 -3.64779681e-01
1.01480126e+00 -2.91447937e-01 1.33118308e+00 -6.74645305e-01
-1.14932969e-01 3.69098425e-01 1.15954928e-01 -1.81733593e-02
1.30138016e+00 8.03098362e-03 4.61457521e-02 4.81283128e-01
1.39350784e+00 1.41475067e-01 2.91678965e-01 4.20123011e-01
-1.00614870e+00 9.36576724e-01 2.01481625e-01 5.99520266e-01
-7.26182699e-01 3.02580059e-01 2.63508528e-01 8.29787195e-01
-7.53961131e-02 1.39436081e-01 -6.89357102e-01 -2.89501458e-01
-2.44573981e-01 -1.88147008e-01 1.03200674e+00 9.64956403e-01
4.24524635e-01 3.78886312e-01 2.21978158e-01 2.30454430e-01
5.47954917e-01 4.32657033e-01 1.98118567e-01 -1.21681273e+00
6.61318004e-01 4.95900691e-01 2.68499583e-01 -9.27815616e-01
-1.62975281e-01 -4.86934811e-01 -5.01289070e-01 3.18776369e-02
2.63635755e-01 -5.21670997e-01 -6.24209642e-01 8.81181598e-01
-4.07619923e-01 -4.84246463e-01 4.90799546e-01 3.81718189e-01
4.00996745e-01 5.89557171e-01 2.13204131e-01 -6.13434255e-01
1.14945745e+00 -1.22267246e+00 -9.11042154e-01 4.94888574e-02
1.37923524e-01 -7.50610054e-01 5.11747181e-01 3.40364158e-01
-1.46950257e+00 -5.20917833e-01 -5.09950578e-01 4.78388608e-01
-1.72559470e-01 2.08149895e-01 1.13611376e+00 7.39895821e-01
-9.32362258e-01 4.50846046e-01 -7.54700899e-01 -4.89268273e-01
1.26382224e-02 4.16180074e-01 -2.41182536e-01 6.53024316e-02
-5.47364295e-01 1.11323643e+00 5.69862664e-01 -1.95354801e-02
-2.57642865e-01 -2.68874913e-01 -5.24397254e-01 2.66146392e-01
1.55688599e-01 -4.52752233e-01 9.99624968e-01 -9.71876621e-01
-9.48749661e-01 1.09101668e-01 3.36505383e-01 -2.88616866e-01
1.49941415e-01 -1.34084776e-01 -3.63466442e-01 1.31612420e-01
5.02109826e-01 2.45131299e-01 -2.11062536e-01 -8.43507469e-01
-1.11354733e+00 -3.85377556e-01 8.98237377e-02 -7.01734126e-02
-4.43284074e-03 3.08073610e-01 3.05209368e-01 -8.07119668e-01
4.67745334e-01 -9.05403018e-01 -2.92903453e-01 -1.18136692e+00
1.99730992e-01 -1.62558571e-01 4.55832630e-01 -8.33230317e-01
1.15663612e+00 -1.59598565e+00 -6.65470243e-01 3.04436833e-01
-5.04292846e-01 -6.51147425e-01 5.80776632e-01 1.02074158e+00
-2.09088802e-01 4.22423422e-01 2.36260891e-01 6.66786313e-01
-1.17347963e-01 3.36808413e-01 4.78982627e-02 1.88551873e-01
6.08780265e-01 6.31041288e-01 -7.46903479e-01 -1.95464745e-01
3.60659659e-02 -1.34495899e-01 -2.96180733e-02 -4.06289190e-01
6.76800609e-01 8.02536607e-02 -3.64927232e-01 1.29719758e+00
1.09591186e+00 1.54173478e-01 4.50473428e-01 2.81484306e-01
-1.00945687e+00 4.99429584e-01 -1.02613521e+00 4.32553589e-01
-3.01423788e-01 5.19313574e-01 1.90786365e-02 -9.07590806e-01
9.91333485e-01 4.22345817e-01 1.45620003e-01 -6.16744876e-01
2.06628546e-01 5.28165460e-01 5.46845734e-01 -1.71217754e-01
9.41705167e-01 -6.60788774e-01 1.42210722e-01 3.09095830e-01
-1.31588221e-01 -1.05090454e-01 5.09909332e-01 2.61438519e-01
6.39297128e-01 -1.23653613e-01 5.73557198e-01 -5.58583140e-01
1.77433625e-01 -2.77109027e-01 7.30038881e-01 1.22664370e-01
-3.79526056e-02 -2.14781076e-01 8.75869215e-01 4.61691022e-01
-7.36736059e-01 -8.25280607e-01 -1.72838986e-01 6.12598062e-01
6.05532974e-02 7.33224690e-01 -2.48535648e-01 -9.23777297e-02
3.76985520e-01 4.06006753e-01 -2.29347587e-01 4.84543979e-01
-2.98899114e-01 -4.61862534e-01 -1.38156742e-01 9.15883601e-01
9.03961182e-01 -1.22455561e+00 -9.33375359e-01 2.29926303e-01
6.56305730e-01 -5.58339357e-01 -6.63454980e-02 3.30825001e-01
-1.50102651e+00 -9.38970029e-01 -5.08731782e-01 -9.20420587e-01
7.03121483e-01 6.59181297e-01 9.13892508e-01 2.75955975e-01
2.50276715e-01 4.68400605e-02 -3.50634485e-01 -1.14725792e+00
-1.86658517e-01 -7.24736080e-02 -2.18299031e-01 -7.55163372e-01
4.19908464e-01 -2.62523979e-01 -4.99360740e-01 -7.49669001e-02
-4.24953401e-01 -5.62740386e-01 1.03009057e+00 5.17938793e-01
5.75759560e-02 1.03520262e+00 1.66163218e+00 -6.89232349e-01
7.06679821e-01 -9.11551178e-01 -4.34026390e-01 1.75693315e-02
-1.22630560e+00 -4.55475211e-01 -8.37104768e-03 6.75668865e-02
-1.48697591e+00 -6.00446641e-01 4.23660278e-01 4.86362696e-01
1.77784353e-01 1.14554060e+00 7.67093748e-02 7.57644475e-02
-2.78852224e-01 -3.67680430e-01 7.49260187e-02 -2.26739764e-01
-4.74836022e-01 6.54585898e-01 5.98866522e-01 -1.42265093e-02
6.99574590e-01 1.62993625e-01 3.08037519e-01 -4.10735637e-01
9.21894424e-03 -4.74369824e-01 -5.47144592e-01 -2.70096689e-01
4.71332401e-01 -1.18066084e+00 -8.58858526e-01 3.98465335e-01
-4.59668964e-01 -1.66030422e-01 1.04207195e-01 1.32494187e+00
-2.32347146e-01 -8.81292596e-02 -1.04578710e+00 -1.20473683e+00
-2.45868564e-01 -5.22176147e-01 -5.51433340e-02 6.59767389e-01
-5.10877697e-04 -1.09360814e+00 -3.41239989e-01 4.83570606e-01
4.98806059e-01 8.43653262e-01 8.93319726e-01 2.18907073e-01
-9.35587823e-01 4.12734179e-03 -4.58409786e-01 4.50161576e-01
9.35220242e-01 4.43420649e-01 -1.94647368e-02 -2.72057176e-01
1.99391484e-01 3.54039147e-02 7.36971319e-01 7.31025815e-01
1.18897157e-02 -2.98496425e-01 4.23828587e-02 -1.45405576e-01
2.09720373e+00 9.75860775e-01 9.97726738e-01 9.90126491e-01
1.58245444e-01 1.03964090e+00 1.54094362e+00 6.64648414e-01
2.44351104e-01 -4.36873138e-01 2.96658278e-01 1.83754444e-01
3.35900128e-01 2.28776261e-01 1.78372979e-01 8.43922794e-01
-9.18396175e-01 2.03636512e-01 -8.33416462e-01 1.16762674e+00
-1.23450887e+00 -1.30472541e+00 -6.80680513e-01 1.77139890e+00
4.59016144e-01 3.45779270e-01 4.31908607e-01 4.77521658e-01
3.97137821e-01 -4.33546633e-01 -1.24288850e-01 -9.66301620e-01
8.60844478e-02 3.86779159e-01 1.22834349e+00 2.38068163e-01
-3.11543941e-01 5.58888495e-01 6.28836107e+00 -1.74483016e-01
-1.11098576e+00 -3.56554300e-01 7.72143543e-01 3.22859824e-01
-3.39184314e-01 6.42838001e-01 -6.83702171e-01 3.64568293e-01
9.61685359e-01 -4.93556917e-01 -7.62442872e-02 6.18067682e-01
5.73105514e-01 -8.78171861e-01 -3.50745320e-01 1.90881521e-01
-5.29113710e-01 -1.22193265e+00 -5.20767391e-01 6.61264062e-01
1.19721985e+00 -4.46841955e-01 1.35966226e-01 -9.94586479e-03
2.12686062e-01 -9.35585856e-01 8.58378351e-01 5.43613017e-01
3.15081209e-01 -1.23761988e+00 1.09210527e+00 1.05179764e-01
-1.10279870e+00 -8.41899753e-01 -2.70736516e-01 -8.72133493e-01
2.47117788e-01 1.96820423e-01 -8.86868119e-01 5.78394294e-01
6.93914533e-01 2.51826376e-01 5.31847402e-02 4.30136859e-01
2.47216702e-01 8.70017529e-01 4.05326575e-01 3.78080487e-01
4.29690361e-01 -1.24384367e+00 -5.46474934e-01 1.02770233e+00
5.73233724e-01 4.03502107e-01 -5.71481526e-01 4.48087454e-01
6.56639040e-02 5.78339137e-02 -6.41122878e-01 -6.39384627e-01
8.90985131e-01 6.80350542e-01 -1.37873912e+00 -2.15382740e-01
-7.20512509e-01 2.91695416e-01 -5.51434696e-01 -6.10876270e-02
-4.01997834e-01 -9.34908867e-01 1.32087484e-01 3.84262532e-01
5.46025872e-01 -5.78190148e-01 -8.90432537e-01 -2.99197286e-01
-2.18296766e-01 -4.52040672e-01 -4.36886288e-02 -4.38615590e-01
-6.22998476e-01 -5.12037754e-01 3.26328218e-01 -5.90381444e-01
-7.14693666e-02 -4.31733787e-01 -6.96173906e-01 1.19042075e+00
-1.45070958e+00 -8.75602067e-01 -2.75338352e-01 -1.44142613e-01
3.64721715e-01 -3.50313067e-01 1.54754817e-01 -6.91053048e-02
-3.85058939e-01 -8.53497814e-03 3.42082411e-01 5.70402481e-02
1.09507933e-01 -1.26000154e+00 1.54200539e-01 6.90589666e-01
-9.80859399e-01 1.07971895e+00 1.52666584e-01 -1.41022658e+00
-9.86529708e-01 -6.10104620e-01 1.33030534e+00 -8.08100849e-02
6.84749544e-01 6.33871734e-01 -5.44393003e-01 9.41907167e-01
7.98371434e-01 -8.96971583e-01 5.08628666e-01 -1.54051989e-01
5.51394343e-01 -2.16067240e-01 -1.57880735e+00 2.41096951e-02
6.17563248e-01 -2.70964652e-01 -5.08470893e-01 -1.49528965e-01
3.32401872e-01 1.04274660e-01 -1.57444227e+00 -1.31642679e-03
7.41827607e-01 -9.61362839e-01 8.47591281e-01 -1.11439556e-01
7.30159283e-01 3.21323842e-01 7.24257529e-02 -5.51494837e-01
-9.21201468e-01 -2.63709158e-01 3.62723976e-01 1.79749203e+00
2.91424483e-01 -1.08269060e+00 6.96287274e-01 6.39222026e-01
-3.00264210e-01 -6.15027308e-01 -4.43863332e-01 -8.35020006e-01
2.33525619e-01 5.81691802e-01 7.11353004e-01 1.47452712e+00
-4.54257504e-04 1.77318320e-01 3.00395280e-01 -2.82658398e-01
4.27557319e-01 2.35578924e-01 6.72198057e-01 -1.16369295e+00
-2.55484972e-02 -9.92680192e-02 9.26307496e-03 -1.97219715e-01
-2.82623857e-01 -3.10340434e-01 -7.39556670e-01 -2.17765284e+00
2.87656814e-01 -5.64143300e-01 -3.65454108e-01 2.87618816e-01
3.11436206e-01 -2.04384834e-01 6.39755070e-01 2.57190496e-01
6.82225287e-01 -2.64199615e-01 1.38109016e+00 4.30139869e-01
-5.79420805e-01 1.32014509e-02 -1.41006708e+00 3.63513917e-01
1.23688197e+00 -1.14290632e-01 -5.64379871e-01 -1.83911204e-01
2.36149237e-01 6.22656882e-01 -1.55837256e-02 -5.68849444e-01
-4.29330438e-01 -1.24787021e+00 2.99644887e-01 -1.00878727e+00
-3.71812791e-01 -1.15427423e+00 8.53095353e-01 1.02895522e+00
1.68370485e-01 8.07958782e-01 -1.54270738e-01 4.85433005e-02
-2.33577818e-01 -5.72812915e-01 4.04506356e-01 -3.43053758e-01
-2.66657114e-01 -4.14139599e-01 -5.25298059e-01 -5.97141206e-01
1.18372643e+00 -1.15973604e+00 -4.22749430e-01 -3.54198995e-03
-2.26270676e-01 1.62500724e-01 6.40894294e-01 -4.04265001e-02
2.13848084e-01 -1.24786651e+00 -4.81536865e-01 -4.27825928e-01
-4.67005283e-01 -1.53332770e-01 -1.40145421e-01 1.09428692e+00
-1.13696563e+00 1.03447056e+00 -1.01916754e+00 4.35983807e-01
-8.36669028e-01 1.69833109e-01 2.32458431e-02 -8.87705386e-02
6.37407154e-02 5.92689514e-01 8.28860030e-02 5.36040306e-01
-5.33179402e-01 -9.23252523e-01 -4.31214541e-01 4.72684354e-01
4.24082242e-02 1.34142137e+00 -1.02673948e-01 -6.31700814e-01
-3.55416834e-01 4.51773345e-01 2.10401058e-01 -4.10415977e-01
1.55622005e+00 -3.95229578e-01 -7.60986984e-01 6.60640061e-01
7.80930817e-01 3.94490004e-01 -1.11050129e+00 4.14392620e-01
5.44688642e-01 -1.07559729e+00 -2.53490269e-01 -5.93249440e-01
-1.22788298e+00 1.77416831e-01 2.87743032e-01 4.46620673e-01
1.40862608e+00 -3.00570756e-01 3.82676333e-01 1.09576732e-01
3.29260021e-01 -1.80662775e+00 1.91573486e-01 4.65440184e-01
9.14589107e-01 -9.08737183e-01 2.71876846e-02 -4.00071949e-01
-7.36431718e-01 1.12026405e+00 4.26558405e-01 -4.09870595e-01
9.17937517e-01 6.93724379e-02 -2.80639052e-01 -1.93891689e-01
-5.23394048e-01 2.83084571e-01 -5.59994876e-01 6.09382510e-01
1.25602198e+00 3.55575055e-01 -1.32469714e+00 4.34047788e-01
-4.59377199e-01 2.63446957e-01 1.47062516e+00 1.41817451e+00
-7.72616029e-01 -9.60902333e-01 -6.51134074e-01 9.49091315e-01
-1.41626823e+00 -4.90380004e-02 -2.65233159e-01 1.43644416e+00
2.84765393e-01 1.07083654e+00 5.59149861e-01 2.77541339e-01
2.29630411e-01 -4.49108660e-01 5.71084082e-01 -5.26184618e-01
-1.01777744e+00 2.83292472e-01 3.82160038e-01 5.60414493e-01
-7.64625251e-01 -1.07645726e+00 -1.68835902e+00 -9.09090102e-01
-8.44876111e-01 2.45576769e-01 9.83670473e-01 6.50935888e-01
-3.00207973e-01 4.04170662e-01 9.76091027e-01 -3.04176390e-01
-3.68413717e-01 -1.28838694e+00 -1.31281304e+00 -3.77701372e-02
1.64411459e-02 -5.80977619e-01 -8.71264189e-02 2.66680449e-01] | [9.033390998840332, 6.1504621505737305] |
57b56834-9d79-46e1-9000-1f7b3d5c3e24 | evaluating-text-coherence-at-sentence-and-1 | 2006.03221 | null | https://arxiv.org/abs/2006.03221v1 | https://arxiv.org/pdf/2006.03221v1.pdf | Evaluating Text Coherence at Sentence and Paragraph Levels | In this paper, to evaluate text coherence, we propose the paragraph ordering task as well as conducting sentence ordering. We collected four distinct corpora from different domains on which we investigate the adaptation of existing sentence ordering methods to a paragraph ordering task. We also compare the learnability and robustness of existing models by artificially creating mini datasets and noisy datasets respectively and verifying the efficiency of established models under these circumstances. Furthermore, we carry out human evaluation on the rearranged passages from two competitive models and confirm that WLCS-l is a better metric performing significantly higher correlations with human rating than tau, the most prevalent metric used before. Results from these evaluations show that except for certain extreme conditions, the recurrent graph neural network-based model is an optimal choice for coherence modeling. | ['Sujian Li', 'Shuang Zeng', 'Sennan Liu'] | 2020-06-05 | evaluating-text-coherence-at-sentence-and | https://aclanthology.org/2020.lrec-1.210 | https://aclanthology.org/2020.lrec-1.210.pdf | lrec-2020-5 | ['sentence-ordering'] | ['natural-language-processing'] | [ 2.64872909e-02 4.46925089e-02 -9.62112173e-02 -4.42634672e-01
-8.19903612e-01 -5.15742302e-01 7.88841188e-01 5.45794666e-01
-5.52808225e-01 9.43186522e-01 7.69653499e-01 -2.08981812e-01
-4.19171005e-01 -4.48392421e-01 -2.57600665e-01 -3.15794587e-01
-1.15749784e-01 5.16485631e-01 2.89373815e-01 -3.23819399e-01
8.31176937e-01 1.11639276e-01 -1.37834644e+00 4.24927711e-01
1.07507479e+00 6.21996403e-01 4.66869250e-02 8.75373423e-01
2.01287791e-01 9.09146965e-01 -9.22440171e-01 -5.57424486e-01
-1.62048757e-01 -5.46145976e-01 -1.27786028e+00 4.96152136e-03
4.55373913e-01 -1.20080486e-01 -1.67292565e-01 7.33900189e-01
4.84581888e-01 3.12469035e-01 7.33109832e-01 -6.00161731e-01
-7.17533171e-01 1.10707247e+00 -1.64488673e-01 5.35870075e-01
1.06666195e+00 -1.18152030e-01 1.46300352e+00 -5.74842453e-01
7.88612187e-01 9.44157660e-01 6.47107601e-01 6.21964037e-02
-1.16681707e+00 -1.31937325e-01 2.37335525e-02 3.83878261e-01
-1.09555089e+00 -3.65048498e-01 9.52005267e-01 -2.21246764e-01
1.28169739e+00 5.46300352e-01 5.36588669e-01 1.45149004e+00
3.11952353e-01 5.19728124e-01 1.26024091e+00 -7.06530273e-01
4.20095958e-02 2.37668633e-01 7.80809999e-01 3.20776820e-01
3.04292113e-01 -2.15892673e-01 -6.29502773e-01 -2.71231681e-02
7.31499940e-02 -6.95369899e-01 -3.88025135e-01 -1.26157282e-02
-1.33042061e+00 7.11278439e-01 4.09406684e-02 7.35972404e-01
-1.77546620e-01 -4.10935879e-01 6.24239683e-01 6.37759686e-01
5.89996397e-01 8.60928714e-01 -1.35470212e-01 -2.63040781e-01
-1.05642545e+00 1.47421733e-01 1.02865911e+00 8.35014045e-01
2.52921671e-01 -3.25837046e-01 -7.41250455e-01 8.96364093e-01
5.27397133e-02 6.42332295e-03 7.61675477e-01 -8.56332302e-01
7.60257244e-01 5.49685478e-01 8.12156126e-02 -1.37656033e+00
-8.02007020e-01 -8.44497383e-01 -8.17517221e-01 -5.94797969e-01
2.00386047e-01 1.10484578e-01 -2.14036092e-01 1.60901201e+00
-2.67361939e-01 -2.13703409e-01 3.72477807e-03 5.08857906e-01
9.66971934e-01 4.70380187e-01 -8.95882864e-03 -7.52876580e-01
1.08563197e+00 -1.03182518e+00 -8.32075536e-01 2.20503360e-01
8.48384261e-01 -8.91400576e-01 1.41213405e+00 6.48022950e-01
-1.13363755e+00 -7.53363788e-01 -1.32609355e+00 -1.37871489e-01
-2.49514073e-01 2.34170258e-01 4.88129616e-01 5.44154227e-01
-1.25979149e+00 8.59384418e-01 -3.70961517e-01 -5.59605777e-01
-2.97167391e-01 3.26101571e-01 -3.20054919e-01 3.71198416e-01
-1.49412990e+00 1.12209463e+00 3.41558844e-01 1.07246071e-01
-3.35559607e-01 -1.51249155e-01 -4.62960184e-01 1.58528671e-01
-1.28562570e-01 -8.29138517e-01 1.09436429e+00 -3.88377577e-01
-1.56240189e+00 9.33479130e-01 3.98452766e-02 -7.30168164e-01
4.81659383e-01 -3.17305356e-01 -5.94435155e-01 5.48361242e-02
1.21732049e-01 2.05091372e-01 3.55933487e-01 -1.04244959e+00
-3.16638350e-01 1.07118919e-01 2.01734185e-01 3.63831788e-01
-7.32490063e-01 -1.33316115e-01 -1.35747612e-01 -5.38262367e-01
-6.73216283e-02 -7.55737185e-01 1.12022489e-01 -1.08173108e+00
-7.73051560e-01 -4.86989319e-01 2.21747100e-01 -7.68168390e-01
2.12095070e+00 -1.71623993e+00 1.91745341e-01 3.29960763e-01
1.01243265e-01 1.08452067e-02 -2.74483472e-01 9.05296028e-01
2.40522083e-02 3.21241945e-01 3.35132368e-02 -4.65077192e-01
1.89496186e-02 -1.82681918e-01 -1.40394762e-01 2.41744325e-01
-1.59746140e-01 6.66551530e-01 -6.46846592e-01 -7.83267200e-01
-1.61285356e-01 1.31230603e-03 -4.39294010e-01 2.99868345e-01
1.85324270e-02 3.90791833e-01 -1.77246243e-01 1.42877162e-01
2.30551064e-01 -4.40517515e-01 3.96983832e-01 -2.30624855e-01
5.65723376e-03 7.48416901e-01 -7.74302661e-01 1.62315488e+00
-3.02386433e-01 8.36465657e-01 -8.50524127e-01 -7.73022473e-01
9.53608751e-01 3.10752869e-01 2.10911393e-01 -9.58624363e-01
1.68577567e-01 -2.98612770e-02 2.53013819e-01 -8.11405897e-01
1.10007179e+00 2.76173383e-01 -2.81359822e-01 6.28109932e-01
1.97978429e-02 -5.10848649e-02 8.13241899e-01 5.85047185e-01
1.20260978e+00 -2.35276669e-01 4.20750707e-01 -4.61281985e-01
6.78860903e-01 -1.14876971e-01 1.26254678e-01 9.67419803e-01
-1.46645397e-01 7.53866315e-01 8.18139851e-01 -3.89539637e-02
-1.13298416e+00 -8.57716799e-01 -1.89093411e-01 9.85326052e-01
1.15077086e-02 -6.51606739e-01 -8.46347809e-01 -6.91012740e-01
-3.58996332e-01 1.00025022e+00 -5.85204780e-01 -3.88393924e-02
-6.04291141e-01 -8.32320035e-01 5.13930142e-01 2.87157625e-01
3.41978163e-01 -9.43177462e-01 -3.50288540e-01 1.09868348e-01
-5.69685459e-01 -1.30309999e+00 -3.58750284e-01 8.13787654e-02
-7.69496918e-01 -9.68236208e-01 -3.12247157e-01 -7.79863477e-01
2.70289570e-01 1.10317521e-01 1.58408773e+00 2.50444561e-01
2.81387240e-01 1.82388827e-01 -7.16767788e-01 1.71155632e-01
-5.70287466e-01 7.35555112e-01 2.90226433e-02 -4.03231919e-01
2.09845692e-01 -5.59193432e-01 -5.10012150e-01 7.63432682e-02
-8.18259060e-01 5.25181442e-02 4.22356665e-01 7.60487378e-01
-9.45372060e-02 -1.04465829e-02 8.41447532e-01 -1.13573539e+00
1.56737256e+00 -3.97069842e-01 -1.75343320e-01 4.84414458e-01
-8.62225592e-01 1.96835957e-02 6.75163567e-01 -2.04912916e-01
-9.02230918e-01 -5.72456956e-01 -1.26362322e-02 3.77943277e-01
2.69840413e-04 8.08579683e-01 1.95301756e-01 4.47142750e-01
8.07991147e-01 -4.18463647e-02 -4.02050763e-01 -3.35442990e-01
2.34109029e-01 7.46266067e-01 4.88288224e-01 -3.83356065e-01
5.84491134e-01 8.57673213e-02 -2.72009760e-01 -7.48259008e-01
-1.05872762e+00 -4.81450438e-01 -8.48217189e-01 -2.46218637e-01
6.92010164e-01 -6.30996406e-01 -5.20593762e-01 1.53174102e-01
-1.28488839e+00 8.92813411e-03 6.48986027e-02 4.64703113e-01
-4.13063467e-01 7.97525525e-01 -8.02887559e-01 -6.15957737e-01
-5.74819982e-01 -8.75000477e-01 7.36933231e-01 -3.46887670e-03
-9.45443988e-01 -1.33823431e+00 5.93072414e-01 4.18532908e-01
2.97871798e-01 2.70633370e-01 1.11981630e+00 -9.23558772e-01
-3.41539867e-02 -2.57793546e-01 1.18539169e-01 2.93255806e-01
6.57401532e-02 3.44174504e-01 -8.77273083e-01 -1.89632013e-01
-3.61131132e-02 -3.70308787e-01 9.27005768e-01 2.53635556e-01
7.86383629e-01 -2.86131352e-01 -1.00992545e-02 3.48446369e-02
1.32462418e+00 8.36601108e-03 6.49812043e-01 6.41303539e-01
2.83397377e-01 6.75008595e-01 5.03498733e-01 4.62251276e-01
4.53954488e-01 7.05011129e-01 -2.35600416e-02 1.19843781e-01
-1.52327558e-02 -2.26098612e-01 3.43393892e-01 1.71778607e+00
1.11791119e-01 -7.57468045e-01 -8.40148985e-01 4.30721492e-01
-1.91486681e+00 -1.06939542e+00 -2.98017621e-01 2.01660442e+00
1.08327484e+00 6.60631895e-01 2.98330814e-01 4.65056479e-01
4.91468132e-01 2.47538820e-01 2.03286409e-01 -7.61163175e-01
-4.99541938e-01 4.99551706e-02 -1.17105067e-01 7.17670619e-01
-1.01475966e+00 7.87645042e-01 7.50422668e+00 7.08747268e-01
-6.76966012e-01 -1.39407784e-01 6.08657122e-01 -1.84203625e-01
-5.06542027e-01 -1.22420512e-01 -5.78484714e-01 5.36505878e-01
1.16037834e+00 -1.71308324e-01 9.41500440e-02 3.54055107e-01
5.05796731e-01 -3.03610891e-01 -1.21609521e+00 5.48423290e-01
3.25897604e-01 -1.21341479e+00 1.61362961e-01 -3.79793227e-01
9.32539403e-01 -1.80931091e-01 -6.24730214e-02 2.76370913e-01
1.32635608e-01 -9.92299318e-01 4.84406382e-01 7.26972997e-01
2.54127979e-01 -5.53434968e-01 1.03845167e+00 3.21621716e-01
-7.82596469e-01 7.80735388e-02 -3.03074121e-01 -2.71731794e-01
8.57987031e-02 8.46391976e-01 -7.58539617e-01 8.66513610e-01
4.87019390e-01 7.92290986e-01 -1.28478479e+00 1.00735188e+00
-2.58746892e-01 6.97790742e-01 -7.54804835e-02 -3.66789222e-01
1.75759435e-01 -2.02162787e-01 4.68596756e-01 1.76341534e+00
9.13291350e-02 -2.88231850e-01 -1.09749340e-01 4.62516069e-01
-3.42823751e-02 4.43931192e-01 -6.71254694e-01 1.09650038e-01
5.09436667e-01 1.22490895e+00 -7.65400410e-01 -2.96113402e-01
-2.25210235e-01 8.67329419e-01 7.13554025e-01 3.33726197e-01
-9.21149492e-01 -3.71664405e-01 -2.47942269e-01 -2.37957373e-01
-8.81853420e-03 -3.06886673e-01 -6.21831894e-01 -1.21219981e+00
1.88806966e-01 -8.52547169e-01 3.24721545e-01 -6.54499531e-01
-1.21970725e+00 1.04029262e+00 2.39230826e-01 -1.39814639e+00
-4.45779651e-01 -2.46410415e-01 -7.57333279e-01 5.03216505e-01
-1.36882257e+00 -8.60980392e-01 -2.28366002e-01 1.79758906e-01
4.27479506e-01 -1.38632670e-01 6.75116718e-01 2.69912213e-01
-7.36093700e-01 6.44246936e-01 6.69820607e-02 -1.95252135e-01
7.97534466e-01 -1.30568087e+00 2.19746307e-01 6.90432906e-01
4.31884825e-01 8.12542439e-01 1.08865011e+00 -5.22580385e-01
-6.63097858e-01 -5.74364007e-01 1.49809825e+00 -5.68893313e-01
7.24439144e-01 -3.32730800e-01 -9.52360451e-01 3.72025728e-01
9.44122910e-01 -1.05061579e+00 7.85654187e-01 4.92608786e-01
-9.82522741e-02 5.11335544e-02 -5.66600919e-01 6.14540875e-01
1.01211524e+00 -5.05717158e-01 -8.82092834e-01 5.16878009e-01
7.42956281e-01 -1.72505885e-01 -1.02424848e+00 5.31015873e-01
4.68532860e-01 -1.43779314e+00 5.88461637e-01 -6.08972371e-01
8.23425114e-01 -4.13572453e-02 -4.95248325e-02 -1.24398851e+00
-5.29502690e-01 -4.35511142e-01 4.26307097e-02 1.53675091e+00
8.16886127e-01 -3.23945493e-01 5.63508809e-01 2.18674228e-01
-5.53548075e-02 -9.27019835e-01 -7.33205497e-01 -7.57317364e-01
1.50944665e-01 -1.88225120e-01 2.56895632e-01 8.90608311e-01
5.37912428e-01 1.00486267e+00 -4.13153738e-01 -3.09072495e-01
1.65064454e-01 1.05496354e-01 6.27191007e-01 -1.11733162e+00
-3.43707889e-01 -6.99173093e-01 -3.12113389e-02 -9.14600194e-01
4.06863064e-01 -8.59070599e-01 -8.49501118e-02 -1.64544225e+00
4.89350677e-01 2.75586341e-02 -3.82360309e-01 -1.67287320e-01
-3.22404623e-01 1.21664822e-01 2.10104994e-02 3.73065144e-01
-9.13178265e-01 5.70768893e-01 1.08855212e+00 -1.53097674e-01
-2.51607150e-01 7.33307153e-02 -7.03516364e-01 4.91864771e-01
8.93274128e-01 -2.33164594e-01 -6.56314850e-01 -4.67230946e-01
6.93584263e-01 3.18659432e-02 -7.29516968e-02 -1.05090940e+00
4.05545890e-01 2.21008658e-01 1.22277923e-01 -6.68374717e-01
1.70126017e-02 -3.62592161e-01 -5.45117110e-02 2.18340471e-01
-1.02079630e+00 4.72050011e-01 -1.09667525e-01 3.56874168e-01
-3.49270314e-01 -4.59829718e-01 4.52862829e-01 7.89798275e-02
-2.44879678e-01 -3.72607619e-01 -3.79537970e-01 1.51899517e-01
5.24930775e-01 -2.65145391e-01 -5.18674016e-01 -6.72160208e-01
-6.07438922e-01 1.10334180e-01 2.39481017e-01 5.67290843e-01
4.54651386e-01 -1.01738679e+00 -8.52383971e-01 -2.05922499e-01
-1.27435550e-02 -7.09357202e-01 -1.00415528e-01 1.15012670e+00
-4.77002203e-01 7.77292669e-01 -4.17108834e-02 -4.87364620e-01
-1.31375217e+00 4.41771954e-01 -1.06542788e-01 -9.18569446e-01
-3.59803587e-01 4.72479761e-01 -3.48841757e-01 -1.91717505e-01
2.71740615e-01 -5.12550771e-01 -8.01972747e-01 2.25675985e-01
9.54759400e-03 5.09187102e-01 3.96008700e-01 -3.50313425e-01
-1.67390391e-01 4.75247085e-01 -2.25259081e-01 -1.47978708e-01
1.06983745e+00 -4.67851043e-01 -1.78403974e-01 7.70851672e-01
1.23459399e+00 7.90175721e-02 -3.66831511e-01 -4.36245091e-02
6.07763708e-01 -3.40300128e-02 -2.60924339e-01 -6.91187024e-01
-3.62871736e-01 5.03973365e-01 1.84986740e-01 8.75200570e-01
9.90505815e-01 -2.56062001e-01 4.79831606e-01 5.31091988e-01
8.99604037e-02 -1.33612978e+00 3.23073834e-01 6.88030124e-01
9.90039825e-01 -1.11314535e+00 4.69799936e-01 -2.71364242e-01
-7.42630720e-01 1.15953243e+00 4.41006422e-01 -2.30745748e-01
5.23752809e-01 -6.23950213e-02 -9.05605257e-02 -1.60705760e-01
-1.12780285e+00 2.82141473e-02 5.52548349e-01 9.22397375e-02
9.13290739e-01 -1.35670781e-01 -1.04367876e+00 4.33859378e-01
-7.75624871e-01 -3.02394539e-01 7.78077006e-01 5.34298003e-01
-5.05701959e-01 -1.08617866e+00 5.19253574e-02 4.94505495e-01
-3.84347826e-01 -2.12822720e-01 -7.59229958e-01 9.46402311e-01
-3.96605760e-01 1.32704484e+00 -1.36068776e-01 -7.32290983e-01
3.43150079e-01 -5.88898398e-02 4.37275022e-01 -5.73314548e-01
-1.04906690e+00 -1.00674860e-01 7.21421480e-01 -3.35197687e-01
-7.76415646e-01 -5.87461531e-01 -8.35651577e-01 -3.42746377e-01
-6.52705312e-01 4.14236933e-01 2.55085051e-01 1.04791594e+00
9.31040272e-02 6.06562614e-01 8.87597740e-01 -3.46250594e-01
-6.52777970e-01 -1.51203215e+00 -4.90604997e-01 6.68826759e-01
7.68434107e-02 -3.30249697e-01 -5.64290226e-01 2.13034656e-02] | [12.03528881072998, 9.368117332458496] |
a7907ce4-3f99-4b8c-9cb1-368c78510ade | can-generative-large-language-models-perform | 2307.04172 | null | https://arxiv.org/abs/2307.04172v1 | https://arxiv.org/pdf/2307.04172v1.pdf | Can Generative Large Language Models Perform ASR Error Correction? | ASR error correction continues to serve as an important part of post-processing for speech recognition systems. Traditionally, these models are trained with supervised training using the decoding results of the underlying ASR system and the reference text. This approach is computationally intensive and the model needs to be re-trained when switching the underlying ASR model. Recent years have seen the development of large language models and their ability to perform natural language processing tasks in a zero-shot manner. In this paper, we take ChatGPT as an example to examine its ability to perform ASR error correction in the zero-shot or 1-shot settings. We use the ASR N-best list as model input and propose unconstrained error correction and N-best constrained error correction methods. Results on a Conformer-Transducer model and the pre-trained Whisper model show that we can largely improve the ASR system performance with error correction using the powerful ChatGPT model. | ['Kate Knill', 'Mark Gales', 'Potsawee Manakul', 'Mengjie Qian', 'Rao Ma'] | 2023-07-09 | null | null | null | null | ['speech-recognition'] | ['speech'] | [ 5.01927137e-01 1.71045080e-01 1.68650493e-01 -5.42862535e-01
-1.32021177e+00 -3.31527442e-01 5.48490584e-01 1.60186991e-01
-7.11128294e-01 2.25845397e-01 3.99218231e-01 -7.26878166e-01
4.52882439e-01 -1.68072030e-01 -4.43852276e-01 -2.27833942e-01
2.30266169e-01 8.54698598e-01 3.87369633e-01 -6.97880447e-01
2.37468436e-01 3.78868058e-02 -1.21019185e+00 3.62603039e-01
7.20108688e-01 4.12754297e-01 3.97481531e-01 1.44686639e+00
-2.57080913e-01 9.02994215e-01 -9.23921108e-01 -3.85989517e-01
1.14844941e-01 -8.78218830e-01 -9.05397415e-01 -2.50177741e-01
8.29677358e-02 -2.13778377e-01 -5.92218697e-01 8.84948432e-01
7.41708100e-01 5.18523037e-01 2.67051578e-01 -5.31599462e-01
-4.76551533e-01 1.08878338e+00 -3.52419913e-02 6.23648107e-01
4.31026697e-01 -3.27615775e-02 9.46473122e-01 -1.07003021e+00
4.37125534e-01 1.19134164e+00 5.80138981e-01 1.06028521e+00
-1.25157177e+00 -4.10910517e-01 -1.63550094e-01 1.75070152e-01
-1.40411031e+00 -1.33193493e+00 3.55828196e-01 6.19455874e-02
1.88603973e+00 5.72609603e-01 2.70033866e-01 7.80263424e-01
2.97510773e-02 7.92208374e-01 5.04327297e-01 -1.00581217e+00
2.67312884e-01 3.96359749e-02 3.37292910e-01 4.66435879e-01
-2.85141617e-01 9.09018889e-02 -8.52267563e-01 8.63180086e-02
1.24034442e-01 -2.87127048e-01 -2.28872776e-01 7.07363307e-01
-7.51856983e-01 7.64440715e-01 -1.68838337e-01 6.04009449e-01
-1.36303887e-01 2.15299547e-01 3.75999689e-01 7.78374672e-01
8.29329014e-01 4.43726540e-01 -4.09181476e-01 -8.13066661e-01
-1.48533440e+00 -1.56324685e-01 9.54460382e-01 1.03305256e+00
3.50100845e-01 4.16931659e-01 -4.19401556e-01 1.43647158e+00
2.84625232e-01 3.90438229e-01 7.92878389e-01 -6.57970965e-01
5.06466925e-01 5.32130934e-02 -5.30516952e-02 -1.72758222e-01
2.32260972e-02 -3.01542372e-01 -2.54324704e-01 -5.96509390e-02
2.75687397e-01 3.48797664e-02 -1.46750057e+00 1.38269699e+00
-9.47504863e-02 1.68766737e-01 3.35250944e-01 4.96853441e-01
5.49346030e-01 1.10113096e+00 -7.37740323e-02 -6.02738619e-01
9.04292643e-01 -1.10381520e+00 -1.06865442e+00 -4.53216344e-01
1.34149563e+00 -1.10579336e+00 9.65263307e-01 2.65928894e-01
-1.38722479e+00 -2.13989541e-01 -8.46064091e-01 -1.68089375e-01
-1.16555251e-01 -5.13403118e-02 -1.09459497e-01 8.23786080e-01
-1.34659040e+00 6.12096131e-01 -9.66725826e-01 -5.59551895e-01
-8.03153664e-02 1.25825584e-01 -2.41559073e-02 -2.54302174e-01
-1.34242499e+00 1.26255798e+00 1.54170036e-01 -1.42058730e-01
-8.17044735e-01 -5.09709001e-01 -6.62423074e-01 1.29931524e-01
3.71411949e-01 8.92747268e-02 2.15080452e+00 -8.95235240e-01
-2.02588081e+00 8.09134960e-01 -6.57344818e-01 -5.69571435e-01
3.08428913e-01 -2.09988251e-01 -5.67949593e-01 -1.02699079e-01
-4.11119074e-01 5.57488156e-03 6.29068315e-01 -7.23921359e-01
-6.25757933e-01 -1.79472849e-01 -4.82621878e-01 2.29027912e-01
6.04349971e-02 9.48523223e-01 -3.27602178e-01 -6.92128897e-01
1.97802946e-01 -8.16812336e-01 -1.28231063e-01 -5.75712562e-01
-1.43753231e-01 -3.37790698e-01 6.06223106e-01 -1.09763491e+00
1.77535772e+00 -2.05707335e+00 -5.80344908e-02 1.87690273e-01
-2.31039479e-01 8.93041909e-01 -1.68796346e-01 7.14959204e-01
3.94864418e-02 3.46559137e-01 -2.39361569e-01 -8.63391161e-01
-1.03889942e-01 3.53208691e-01 -2.90412039e-01 1.50068447e-01
1.96744397e-01 7.88513660e-01 -7.74010181e-01 -2.32473552e-01
1.61081240e-01 5.06827295e-01 -5.96897781e-01 6.24390304e-01
9.47678909e-02 1.10396408e-01 1.79304674e-01 2.15384796e-01
4.59497906e-02 2.87418336e-01 5.12422435e-02 3.54474455e-01
-3.41720194e-01 1.16547120e+00 -8.45798194e-01 1.64706743e+00
-4.75237280e-01 8.00404251e-01 1.77840039e-01 -8.24151814e-01
1.01159227e+00 8.65158975e-01 -5.06392866e-02 -7.55245388e-01
2.88033843e-01 5.42570770e-01 4.86399978e-01 -2.59004891e-01
7.21336067e-01 -5.01437247e-01 7.84739926e-02 5.64502239e-01
3.33833307e-01 -4.12235290e-01 2.30644718e-02 3.97684425e-01
1.35806239e+00 -4.21396464e-01 4.43774372e-01 6.01037927e-02
3.54881793e-01 1.18854918e-01 2.11391911e-01 1.15432847e+00
-1.88743472e-01 8.40692997e-01 -1.85404941e-01 1.39205173e-01
-1.23345566e+00 -6.11767173e-01 2.44023222e-02 1.41507280e+00
-4.66889739e-01 -5.95662296e-01 -9.89630520e-01 -1.94817588e-01
-6.04670465e-01 1.55003297e+00 -6.86313435e-02 -4.51316208e-01
-9.00187194e-01 -3.53284568e-01 1.16759455e+00 3.62523407e-01
-1.55263126e-01 -8.61699343e-01 -1.36171177e-01 6.07116163e-01
-9.02865529e-02 -8.19893539e-01 -8.02834511e-01 7.47319221e-01
-6.33697212e-01 -3.27893227e-01 -6.51246130e-01 -8.03370774e-01
2.43949220e-01 2.75047809e-01 8.40961933e-01 4.72672462e-01
8.04745127e-03 2.52414078e-01 -8.75238419e-01 -3.40100884e-01
-1.18111908e+00 9.79658961e-02 2.62285352e-01 -1.90980852e-01
6.36237681e-01 -4.30177778e-01 1.57660022e-01 1.54975951e-01
-5.29918373e-01 -1.14480488e-01 2.13798210e-01 9.14249122e-01
1.60821944e-01 -4.78517056e-01 5.22449553e-01 -8.40250313e-01
7.53694057e-01 -3.11415344e-01 -2.26462767e-01 4.25844818e-01
-7.48619556e-01 7.05314279e-02 5.84542096e-01 -4.73971337e-01
-1.12897098e+00 -1.63284943e-01 -7.94907093e-01 -4.51783925e-01
8.92164484e-02 6.04601085e-01 -1.20564759e-01 -5.79719730e-02
7.61675596e-01 4.91532415e-01 -1.80271845e-02 -7.96649516e-01
2.76714414e-01 1.33027709e+00 3.22378874e-01 -2.34171107e-01
4.80434090e-01 -3.88345331e-01 -8.40473473e-01 -1.16374600e+00
-5.81881881e-01 -6.39022171e-01 -4.55635995e-01 -2.33478412e-01
4.22609776e-01 -6.19666398e-01 -1.96876347e-01 4.56956178e-01
-1.61769676e+00 -5.42319059e-01 -4.52709675e-01 6.45219803e-01
-3.64864975e-01 3.05335522e-01 -8.72176349e-01 -1.37316835e+00
-4.83077496e-01 -9.23229992e-01 8.76811206e-01 6.55700937e-02
-3.69107395e-01 -8.15021932e-01 3.91113520e-01 9.77433324e-02
6.79834187e-01 -1.16748762e+00 6.37388885e-01 -1.44328606e+00
-2.02477574e-01 -5.01876116e-01 2.31528163e-01 5.42709112e-01
-1.78605154e-01 -3.63085382e-02 -1.11201859e+00 -2.69025773e-01
1.17292166e-01 -8.53274986e-02 8.40264797e-01 1.24195792e-01
5.84165275e-01 -4.49389011e-01 -1.59521937e-01 3.99061054e-01
9.96410847e-01 6.50078595e-01 7.31110990e-01 -2.56893128e-01
4.44776386e-01 5.12782812e-01 3.69681597e-01 1.43558756e-01
-1.91072132e-02 7.43292749e-01 -1.75072506e-01 4.57631975e-01
-2.89179444e-01 -3.32557291e-01 6.69414043e-01 1.88400495e+00
8.41116384e-02 -8.11703384e-01 -1.20333588e+00 5.82828343e-01
-1.67432368e+00 -1.17940855e+00 -1.62514821e-01 2.28806901e+00
1.01221144e+00 1.32309064e-01 -2.37079069e-01 2.13272378e-01
7.84364700e-01 4.19950783e-02 -3.74180526e-02 -9.73157942e-01
1.84529424e-02 4.40087408e-01 3.57029825e-01 1.19710779e+00
-3.88529629e-01 1.31906796e+00 7.21159124e+00 8.01949084e-01
-1.09154308e+00 5.95061898e-01 2.02767178e-01 -3.09268355e-01
-2.43997738e-01 1.34607613e-01 -9.43927288e-01 4.78655100e-01
1.89735711e+00 -3.59440714e-01 7.65864670e-01 5.28064430e-01
5.15361190e-01 -7.04075396e-02 -1.18785250e+00 1.07902706e+00
3.61928940e-01 -9.70963299e-01 -1.76399231e-01 -3.27161491e-01
2.14165226e-01 3.02948385e-01 -4.32270139e-01 6.51138484e-01
5.30622065e-01 -1.07909572e+00 7.66523421e-01 3.04627895e-01
9.24804211e-01 -4.54611063e-01 5.50353587e-01 6.60040677e-01
-8.76101911e-01 9.20892581e-02 -3.45026404e-01 -2.58350104e-01
6.11715436e-01 2.21665695e-01 -1.21196377e+00 1.66078553e-01
3.48836213e-01 3.85059506e-01 -1.24758862e-01 1.06579602e+00
-1.41974417e-02 1.40159452e+00 -5.23633659e-01 -2.30902597e-01
3.24123241e-02 1.31013200e-01 9.19809997e-01 1.67998517e+00
5.92434704e-01 5.57494819e-01 5.76143153e-02 2.59610564e-01
-1.46195725e-01 2.33524963e-01 -3.58045250e-01 -3.35074335e-01
6.67043507e-01 6.52156889e-01 -3.75430971e-01 -6.98205531e-01
-3.67305756e-01 1.16144550e+00 3.46321732e-01 2.21925318e-01
-4.41916943e-01 -6.23951972e-01 8.26216280e-01 2.52853427e-02
1.67550996e-01 -4.01522368e-01 -1.08570598e-01 -1.22316182e+00
-3.75073433e-01 -9.42752659e-01 1.35516590e-02 -8.91119301e-01
-9.18588877e-01 8.00055683e-01 -3.51598412e-01 -7.32548714e-01
-5.14253438e-01 -3.61643314e-01 -7.46304512e-01 1.14283526e+00
-1.24949825e+00 -5.74310482e-01 2.99963742e-01 3.08121949e-01
1.11074436e+00 -2.78278708e-01 9.58845317e-01 5.14626265e-01
-5.27109921e-01 9.05880272e-01 3.68337959e-01 2.22626492e-01
6.75366342e-01 -1.00171721e+00 6.94127858e-01 1.28620470e+00
3.64510626e-01 6.35471761e-01 1.10393858e+00 -8.23264599e-01
-1.13136125e+00 -9.67595398e-01 1.59798706e+00 -5.84244907e-01
6.19806051e-01 -4.35000986e-01 -1.35566628e+00 8.43750775e-01
2.91046113e-01 -1.41619384e-01 6.74270093e-01 1.02714337e-01
-1.93555877e-01 -1.09625673e-02 -7.62930810e-01 5.13287783e-01
1.13457394e+00 -1.02841747e+00 -1.17215407e+00 1.33491680e-01
1.05672824e+00 -4.49622184e-01 -4.70183283e-01 -1.62304446e-01
3.01203370e-01 -4.25460726e-01 2.60381967e-01 -7.52146244e-01
-7.53944442e-02 -6.65650368e-02 -4.16651249e-01 -1.64387858e+00
-2.58467108e-01 -1.19740331e+00 -1.57832712e-01 1.39974916e+00
6.16548240e-01 -4.42660570e-01 2.87366509e-01 7.51689732e-01
-6.73111200e-01 -1.54533252e-01 -1.30274689e+00 -8.22466433e-01
-1.80692792e-01 -8.16489935e-01 2.22380862e-01 5.94494045e-01
3.99626017e-01 6.24921978e-01 -5.04290879e-01 -1.09463744e-01
1.53893307e-01 -8.46595466e-01 4.25936610e-01 -9.70751941e-01
-3.86673838e-01 -2.05396846e-01 -2.27027938e-01 -1.37412488e+00
-1.97452288e-02 -1.02938545e+00 9.29149151e-01 -1.22561705e+00
-2.18293265e-01 -3.82572085e-01 -2.31847569e-01 5.29212952e-01
-3.51660587e-02 3.90205607e-02 2.90879041e-01 3.12757343e-01
-4.78089482e-01 4.43329722e-01 4.95286852e-01 2.29171719e-02
-4.61491972e-01 1.09129511e-01 -3.35977316e-01 3.59200269e-01
6.82499230e-01 -9.60837662e-01 -1.78151414e-01 -6.21395409e-01
-2.11138707e-02 2.92925417e-01 -1.38500437e-01 -8.75089109e-01
6.93342745e-01 1.22480027e-01 -3.52977008e-01 -3.23053926e-01
4.97240305e-01 -5.50422490e-01 2.39391695e-03 3.18662524e-01
-7.34208047e-01 1.51228189e-01 4.17072959e-02 3.18506271e-01
-2.08962619e-01 -6.87367737e-01 1.00051403e+00 -1.11284636e-01
-4.58180517e-01 -1.81472585e-01 -1.03708458e+00 1.47264168e-01
4.59361643e-01 -3.51288795e-01 -2.44858816e-01 -5.13753712e-01
-7.87476242e-01 -6.32177740e-02 3.01584721e-01 4.19196159e-01
6.07704818e-01 -7.77952850e-01 -8.66691887e-01 4.02879715e-01
1.06595337e-01 -3.49405915e-01 -8.97525474e-02 8.68337452e-01
-1.67175218e-01 4.30540681e-01 3.65566134e-01 -4.51480776e-01
-1.62959826e+00 2.29726553e-01 5.54720342e-01 -6.54017106e-02
-5.42830884e-01 1.09637249e+00 -4.27211851e-01 -3.49561363e-01
3.66749257e-01 -2.16257215e-01 1.16501831e-01 -2.14535907e-01
9.07414913e-01 2.98500061e-01 5.37767649e-01 -7.73517430e-01
-3.89334083e-01 -1.59017241e-03 -3.49169761e-01 -6.45327568e-01
1.17281282e+00 -5.01559317e-01 2.06591651e-01 9.24096763e-01
9.81801689e-01 1.21694088e-01 -6.11744881e-01 -4.56237048e-01
1.97001919e-01 -2.70818502e-01 2.99440175e-01 -8.53989899e-01
-3.91106278e-01 1.17971289e+00 3.22843850e-01 2.39193380e-01
6.61941707e-01 8.84149596e-02 8.43955696e-01 5.37626445e-01
2.88285017e-01 -1.46148956e+00 -2.30254069e-01 1.26299703e+00
6.59424722e-01 -1.03584754e+00 -6.27970874e-01 -7.40243718e-02
-4.71365541e-01 1.08107054e+00 3.80106688e-01 -2.14307266e-03
6.34520113e-01 4.51679796e-01 3.25946480e-01 2.07696721e-01
-1.12110817e+00 -3.58403683e-01 -6.56546280e-03 4.62107569e-01
7.79212296e-01 -3.33052129e-02 -3.64731520e-01 5.15968144e-01
-2.30802074e-01 -8.10371265e-02 6.76974297e-01 9.23171341e-01
-9.85744953e-01 -1.10903513e+00 -2.18538120e-01 5.62768161e-01
-5.45478761e-01 -6.23367250e-01 -4.35846090e-01 2.24947527e-01
-4.94015902e-01 1.37506735e+00 9.67389271e-02 -5.56116283e-01
3.65083128e-01 7.69638717e-01 2.83502370e-01 -1.21775496e+00
-9.73239243e-01 1.83026835e-01 4.42789227e-01 -4.16045040e-01
6.33983687e-02 -5.61502635e-01 -1.30050361e+00 -2.78381824e-01
-8.17762554e-01 4.64070469e-01 7.80227005e-01 1.23156071e+00
2.84430861e-01 4.86849517e-01 4.70822662e-01 -7.36198902e-01
-9.09285486e-01 -1.32673562e+00 -4.74409461e-01 2.07855757e-02
4.14824605e-01 -8.81091952e-02 -4.25752312e-01 5.01287393e-02] | [14.37130069732666, 6.826721668243408] |
8694ed63-c3a0-4106-a4a7-3072cf3ae4d6 | distinguishing-noise-from-chaos-objective | 1401.2139 | null | http://arxiv.org/abs/1401.2139v1 | http://arxiv.org/pdf/1401.2139v1.pdf | Distinguishing noise from chaos: objective versus subjective criteria using Horizontal Visibility Graph | A recently proposed methodology called the Horizontal Visibility Graph (HVG)
[Luque {\it et al.}, Phys. Rev. E., 80, 046103 (2009)] that constitutes a
geometrical simplification of the well known Visibility Graph algorithm [Lacasa
{\it et al.\/}, Proc. Natl. Sci. U.S.A. 105, 4972 (2008)], has been used to
study the distinction between deterministic and stochastic components in time
series [L. Lacasa and R. Toral, Phys. Rev. E., 82, 036120 (2010)].
Specifically, the authors propose that the node degree distribution of these
processes follows an exponential functional of the form $P(\kappa)\sim
\exp(-\lambda~\kappa)$, in which $\kappa$ is the node degree and $\lambda$ is a
positive parameter able to distinguish between deterministic (chaotic) and
stochastic (uncorrelated and correlated) dynamics. In this work, we investigate
the characteristics of the node degree distributions constructed by using HVG,
for time series corresponding to $28$ chaotic maps and $3$ different stochastic
processes. We thoroughly study the methodology proposed by Lacasa and Toral
finding several cases for which their hypothesis is not valid. We propose a
methodology that uses the HVG together with Information Theory quantifiers. An
extensive and careful analysis of the node degree distributions obtained by
applying HVG allow us to conclude that the Fisher-Shannon information plane is
a remarkable tool able to graphically represent the different nature,
deterministic or stochastic, of the systems under study. | ['Osvaldo A. Rosso', 'Bruna Amin Gonçalves', 'Martín Gómez Ravetti', 'Alejandro C. Frery', 'Laura C. Carpi'] | 2014-01-09 | null | null | null | null | ['information-plane'] | ['methodology'] | [-1.30256817e-01 2.38521546e-02 3.27625990e-01 9.73670557e-02
-3.48673649e-02 -7.16034472e-01 7.86047339e-01 2.98937052e-01
-3.05240422e-01 8.35828543e-01 -6.28167570e-01 -6.98521495e-01
-5.41028917e-01 -1.23254049e+00 -2.51430899e-01 -1.03324878e+00
-5.52725613e-01 3.49025935e-01 1.54297262e-01 -3.71189743e-01
2.36798629e-01 6.84863031e-01 -1.58850574e+00 -5.33149183e-01
7.59274721e-01 8.52437437e-01 1.72036737e-01 8.42228174e-01
-4.67325002e-02 3.07911843e-01 -8.14257503e-01 -2.81616628e-01
2.09511384e-01 -6.41962409e-01 -4.66623634e-01 -4.32649285e-01
-5.39193034e-01 3.93598467e-01 -4.05089587e-01 1.27135515e+00
2.19708949e-01 9.48344264e-03 1.05684268e+00 -1.21906066e+00
-3.46176922e-01 4.68326211e-01 -4.88641292e-01 6.47171378e-01
4.04811561e-01 9.54896063e-02 6.05417013e-01 -3.65466326e-01
8.20209622e-01 7.95955598e-01 4.13781583e-01 -5.07534482e-02
-1.28696191e+00 -6.51598036e-01 -4.80744928e-01 2.04699144e-01
-1.69325912e+00 3.34110498e-01 7.88142204e-01 -6.83612823e-01
5.91011226e-01 4.33118969e-01 9.68413293e-01 6.53794885e-01
6.84040248e-01 -3.44865993e-02 1.68635929e+00 -5.67996025e-01
2.28874281e-01 1.91608965e-01 4.32410151e-01 7.41194785e-01
5.52555859e-01 3.77889872e-01 6.87317848e-02 -2.11867228e-01
4.37215924e-01 -2.55185336e-01 -2.13881806e-01 -3.04477066e-01
-7.33516395e-01 6.76603019e-01 2.46422321e-01 8.78610551e-01
-3.63857657e-01 7.52275586e-02 3.81741710e-02 5.57150006e-01
2.88514316e-01 1.67346612e-01 2.07438812e-01 -1.66514471e-01
-6.60157561e-01 2.37705722e-01 9.33810771e-01 5.92745900e-01
9.64731753e-01 -7.04710484e-02 1.33137062e-01 4.19310443e-02
1.83215261e-01 9.39996660e-01 6.49618655e-02 -7.15211630e-01
4.73263748e-02 5.32309473e-01 4.97456118e-02 -1.15981913e+00
-6.31198645e-01 -6.49067283e-01 -1.05456161e+00 3.93760711e-01
6.75541878e-01 -1.22695990e-01 -3.59647572e-01 1.66187871e+00
-6.29272684e-02 -1.97853014e-01 1.59639552e-01 4.05646473e-01
3.93765032e-01 7.40863144e-01 -1.82113603e-01 -5.85605681e-01
1.13928223e+00 -3.00263781e-02 -7.35779405e-01 5.88321924e-01
3.53098512e-01 -6.94483936e-01 7.26697683e-01 3.02260846e-01
-1.05152571e+00 -2.13807762e-01 -1.00851488e+00 7.32261658e-01
-6.01382017e-01 -1.52325287e-01 2.66830623e-01 1.01561105e+00
-1.46751440e+00 6.40467644e-01 -7.67385721e-01 -5.20414114e-01
-1.91583782e-01 1.38651550e-01 -4.45238948e-02 1.69781521e-01
-1.04244530e+00 6.72810435e-01 1.76524054e-02 4.84253988e-02
-5.67577362e-01 -1.03294542e-02 -4.95064676e-01 7.62751177e-02
-2.48618778e-02 -2.65865922e-01 5.42781651e-01 -4.33630437e-01
-1.20519710e+00 9.32513297e-01 -6.44385815e-02 -4.31120455e-01
4.65176314e-01 6.75719559e-01 -5.79000592e-01 3.23410809e-01
-8.16300884e-02 -1.39514491e-01 5.11629462e-01 -1.26181114e+00
-9.40180048e-02 -4.45331037e-01 -5.19652516e-02 -3.23296338e-01
1.11714318e-01 6.14885055e-02 -2.73799035e-03 -3.62598419e-01
1.63435906e-01 -1.05756199e+00 -5.48490062e-02 -6.34114325e-01
-4.03108627e-01 -6.03722744e-02 6.60175443e-01 -3.66696328e-01
1.29562700e+00 -1.98991573e+00 4.54431698e-02 9.68207896e-01
4.61277246e-01 1.10099353e-01 2.60480285e-01 9.50331211e-01
-1.74004268e-02 3.67142320e-01 -5.92229724e-01 8.29479545e-02
-1.70508698e-01 -2.91590369e-03 6.33718148e-02 8.81124139e-01
-3.46748292e-01 4.06476080e-01 -6.60985351e-01 -8.90576914e-02
3.81034687e-02 5.53287745e-01 7.79301301e-02 -2.37829179e-01
1.11307196e-01 5.20368814e-01 -2.61644363e-01 3.43378633e-01
7.28918493e-01 -2.01959476e-01 1.62531510e-01 2.28707403e-01
-6.69578850e-01 -3.35166216e-01 -1.06181335e+00 7.42349327e-01
-7.30528384e-02 1.01346481e+00 2.00599849e-01 -1.05281568e+00
1.38339281e+00 2.50718564e-01 4.37237829e-01 -6.47561252e-01
6.63700521e-01 3.43066543e-01 2.74513751e-01 -2.34697193e-01
2.87281096e-01 -8.31171647e-02 -1.54862538e-01 7.13037848e-01
-1.73378214e-01 -3.64510655e-01 5.32498181e-01 1.89847082e-01
1.32533598e+00 -3.27443510e-01 5.19901395e-01 -7.96837866e-01
6.48172796e-01 -8.09098594e-03 -7.20115229e-02 7.20537186e-01
-4.01329398e-01 2.07371488e-02 1.11826825e+00 -4.93628383e-02
-1.01839244e+00 -1.14922118e+00 -1.67978436e-01 2.32862309e-01
5.63259602e-01 -2.72147298e-01 -7.39029825e-01 -3.13808918e-02
-6.99827224e-02 7.66879737e-01 -7.09032953e-01 -1.75781697e-01
-4.34447676e-01 -8.68544817e-01 6.01126850e-01 -4.14637923e-01
4.06122774e-01 -1.08850992e+00 -6.65137470e-01 -2.44822782e-02
1.87348455e-01 -8.37868392e-01 3.18089098e-01 -5.24489395e-02
-7.32474685e-01 -1.04217172e+00 -4.80180085e-01 -1.67545930e-01
4.36652511e-01 1.95760399e-01 9.86648977e-01 1.40590608e-01
-5.03983080e-01 4.90960568e-01 -4.65976804e-01 -1.09066732e-01
-8.21396470e-01 -1.97807282e-01 2.59020459e-02 -1.52677685e-01
7.95861930e-02 -8.90941858e-01 -5.03915071e-01 3.87717664e-01
-7.70387232e-01 -5.79715669e-01 4.80449170e-01 3.28609973e-01
4.52288121e-01 3.67896497e-01 2.10635707e-01 -4.52506453e-01
8.09294283e-01 -5.99147022e-01 -1.01807845e+00 4.76642922e-02
-8.83038640e-01 1.24096230e-01 7.90321350e-01 1.72969028e-01
-5.08581758e-01 -6.67613924e-01 1.10871032e-01 -1.01749040e-01
-2.50393040e-02 4.48142439e-01 1.17181785e-01 -3.07808369e-01
4.86860752e-01 5.20688057e-01 7.09568039e-02 -2.58944005e-01
7.23854154e-02 2.72547275e-01 2.72142261e-01 -3.26606721e-01
1.03313017e+00 6.71648681e-01 6.05528951e-01 -1.07407796e+00
8.52927268e-02 -3.31221879e-01 -3.50945383e-01 -5.66322803e-01
7.82412052e-01 -2.28093311e-01 -1.02704597e+00 5.81862271e-01
-1.07185161e+00 -1.74037501e-01 -3.04796606e-01 5.66390872e-01
-4.45217162e-01 2.77116209e-01 -4.40312684e-01 -1.33548045e+00
-1.51531333e-02 -9.28333104e-01 3.18504602e-01 3.82621884e-01
1.42935649e-01 -9.55885112e-01 4.11594540e-01 -2.43749246e-01
4.80102718e-01 5.06561399e-01 1.05466056e+00 -4.16109622e-01
-6.11349881e-01 -3.77663374e-01 -3.46856236e-01 -2.91536339e-02
-9.65895206e-02 4.61987913e-01 -6.24806106e-01 -3.15045327e-01
1.48553640e-01 6.24455214e-01 5.94599128e-01 4.43640798e-01
5.48829973e-01 -6.31065527e-03 -3.45250309e-01 4.70653325e-01
1.71351910e+00 5.60278594e-01 6.48973405e-01 1.98232040e-01
1.55175164e-01 6.51373267e-01 4.48992670e-01 4.97332245e-01
1.53991610e-01 5.66354871e-01 3.92346054e-01 3.32434356e-01
1.53671801e-01 1.28549770e-01 3.49094212e-01 8.23982239e-01
-6.27184212e-01 -6.90245271e-01 -1.09568572e+00 2.38612100e-01
-1.19514918e+00 -1.11273265e+00 -6.05565667e-01 2.22814751e+00
2.72825565e-02 3.25338095e-01 2.61396050e-01 2.34938607e-01
9.15350258e-01 3.60074133e-01 7.33005404e-02 -6.24072909e-01
-5.23808956e-01 4.49652582e-01 7.01193273e-01 7.40713596e-01
-5.34145117e-01 3.36823791e-01 4.93285084e+00 6.93663001e-01
-1.08563197e+00 1.94302052e-01 1.35143101e-01 2.51026839e-01
-5.35509884e-01 2.67761171e-01 -3.95888001e-01 7.37457275e-01
1.11526024e+00 -7.31046438e-01 3.61588210e-01 2.35274017e-01
2.03935876e-01 -6.12064898e-01 -1.75609559e-01 9.11585152e-01
-8.88131484e-02 -9.86025453e-01 -3.31272244e-01 4.84348118e-01
5.24357259e-01 -6.53716549e-02 1.65904149e-01 -8.06855112e-02
3.62033069e-01 -9.01668966e-01 5.15366077e-01 8.25518370e-01
6.86290264e-01 -6.19782627e-01 8.77282262e-01 3.39415282e-01
-1.39617848e+00 7.73463547e-02 1.45055786e-01 -1.12684824e-01
4.36039954e-01 7.24191904e-01 -4.29340571e-01 8.60977232e-01
6.04388118e-01 1.73668772e-01 -4.40416127e-01 9.00122046e-01
-1.01935588e-01 5.65044165e-01 -5.58202267e-01 -5.59443235e-01
1.87417984e-01 -8.49615633e-01 1.15788507e+00 8.73392820e-01
7.65669882e-01 9.59278196e-02 -4.69354630e-01 8.97308648e-01
4.89516765e-01 2.71866500e-01 -9.52965617e-01 -1.27258003e-01
4.41533774e-01 9.35361624e-01 -1.27465963e+00 -1.25241950e-01
5.28447144e-02 5.20463169e-01 -9.59699899e-02 3.63285780e-01
-6.53684914e-01 -6.33691847e-01 3.57146293e-01 2.89286435e-01
2.55325258e-01 -7.75041938e-01 -1.44310713e-01 -7.92424798e-01
-4.98887934e-02 -8.80147368e-02 1.66487083e-01 -4.82416064e-01
-7.70617008e-01 8.32835972e-01 3.34706843e-01 -1.18279696e+00
-2.42895722e-01 -5.03195703e-01 -5.78080654e-01 1.21272099e+00
-9.18999970e-01 -4.17039841e-01 -2.89494455e-01 4.96424645e-01
-3.11540872e-01 -7.02644885e-02 5.20946741e-01 -9.93768964e-03
-4.79453236e-01 1.70073882e-01 5.87704360e-01 -3.57435137e-01
1.46293240e-02 -1.14968586e+00 1.21228777e-01 8.83054972e-01
-8.90410393e-02 4.92766738e-01 1.26844609e+00 -6.71314418e-01
-1.26055467e+00 -4.25419986e-01 8.24076593e-01 -2.32191354e-01
8.81044686e-01 -4.50229704e-01 -6.98667288e-01 3.22705328e-01
3.68379593e-01 -1.30047724e-01 4.14378136e-01 -5.42704523e-01
-7.13514909e-02 4.95038666e-02 -1.33577740e+00 3.70044678e-01
8.10107529e-01 -5.16841829e-01 -4.54609282e-02 1.83395416e-01
3.10882390e-01 2.43630350e-01 -9.36562777e-01 1.67169526e-01
4.69820887e-01 -1.62866414e+00 5.76894879e-01 1.21663742e-01
-4.48859185e-02 -2.21630991e-01 -1.35775909e-01 -1.05409455e+00
-3.89211066e-02 -7.56297469e-01 2.85873026e-01 9.04792547e-01
3.93002152e-01 -1.25226974e+00 3.92610013e-01 -1.80049390e-01
1.73273683e-01 -5.44561267e-01 -1.20129204e+00 -9.72905934e-01
9.72087458e-02 -3.76094341e-01 4.12512839e-01 6.98919177e-01
-1.14182867e-02 -1.42252475e-01 1.59407645e-01 2.27441102e-01
7.89501548e-01 7.24299625e-02 4.91152704e-01 -1.43168652e+00
-3.41791868e-01 -7.08436787e-01 -7.15138614e-01 -3.57289582e-01
-1.29443675e-01 -7.32669175e-01 -3.84263933e-01 -1.36259687e+00
-2.02544585e-01 -6.82837427e-01 -1.89011931e-01 -3.60589981e-01
3.22338283e-01 1.07117675e-01 9.95502844e-02 4.98839825e-01
-1.41117364e-01 3.55493009e-01 1.05484533e+00 2.55103230e-01
-1.46895424e-01 1.38393044e-01 -9.39317793e-02 5.11239290e-01
1.01291752e+00 -5.21761954e-01 -2.68282801e-01 3.06935430e-01
5.20746946e-01 4.54377502e-01 5.46086788e-01 -1.08809447e+00
1.43317446e-01 8.95891786e-02 -3.57998312e-01 -5.89194775e-01
3.37565303e-01 -6.55285656e-01 5.42756498e-01 8.50755513e-01
1.58803821e-01 5.26655853e-01 -7.23783448e-02 7.90803909e-01
-2.13416114e-01 -3.70637953e-01 7.09277272e-01 -2.45907512e-02
-3.14991236e-01 4.51348461e-02 -8.28762650e-01 -7.79243037e-02
1.24983203e+00 -2.64300168e-01 -5.67385852e-01 -6.45683646e-01
-7.49961734e-01 -4.07982096e-02 6.19192600e-01 -2.59336978e-01
1.23296201e-01 -6.97077572e-01 -4.66985703e-01 5.41695207e-02
-6.99015260e-02 -6.35564923e-01 4.55179393e-01 1.29496670e+00
-8.90417397e-01 4.87086028e-01 -1.48841530e-01 -5.07047892e-01
-1.02479184e+00 5.61811447e-01 2.68785954e-01 -3.71941686e-01
-2.77968496e-01 4.74393725e-01 -2.64291734e-01 -3.78997326e-02
-4.97456938e-01 -1.28015548e-01 -3.20405722e-01 1.89719573e-01
1.04975350e-01 6.53479636e-01 -1.01700462e-01 -8.79086792e-01
-3.06247503e-01 8.07094157e-01 4.54626441e-01 -5.82013011e-01
9.43803191e-01 -4.28642988e-01 -5.59515476e-01 6.66252375e-01
1.16592860e+00 4.53971505e-01 -6.33669078e-01 1.40310645e-01
-1.44941464e-01 -1.25313640e-01 -3.29551160e-01 -3.50171655e-01
-8.47809851e-01 8.02440822e-01 7.14099348e-01 1.38717270e+00
1.00086403e+00 1.82075039e-01 3.31702977e-01 6.18629679e-02
7.49380887e-01 -6.74071491e-01 -5.08006871e-01 5.09855270e-01
6.86118424e-01 -5.07091582e-01 -2.55737692e-01 -6.60521448e-01
-6.67825267e-02 1.08294904e+00 -6.32813424e-02 -2.71826029e-01
9.39035118e-01 2.85799742e-01 -2.06063822e-01 -4.96194601e-01
-2.55219340e-01 -4.17142928e-01 -9.12321508e-02 3.63296181e-01
4.09161188e-02 4.07653451e-01 -1.03306997e+00 1.77308053e-01
-5.73635519e-01 -1.91992864e-01 8.22382629e-01 8.74224603e-01
-2.95561701e-01 -1.07069230e+00 -5.41727424e-01 2.70312726e-01
-6.90806210e-02 7.93604702e-02 -4.20526236e-01 1.22515571e+00
1.60277590e-01 9.48020458e-01 1.46702141e-01 -4.52471048e-01
1.86263055e-01 5.42904027e-02 5.00802279e-01 -7.48668090e-02
-4.74062353e-01 -3.21473122e-01 -1.89579844e-01 -2.39387050e-01
-4.22271878e-01 -7.91860878e-01 -1.08075583e+00 -5.78393757e-01
-7.10864738e-02 6.77214265e-01 9.06427026e-01 7.52213061e-01
9.96834561e-02 4.06356573e-01 7.84635067e-01 -5.72463453e-01
-1.52693242e-02 -7.63944924e-01 -1.11680174e+00 3.57812643e-02
2.42775038e-01 -9.27781463e-01 -9.72351789e-01 -5.40218234e-01] | [6.938035488128662, 4.496188163757324] |
21a3c8b1-4dbc-4187-a5a0-00156130e7e5 | robust-tracking-of-respiratory-rate-in-high | 1705.06628 | null | http://arxiv.org/abs/1705.06628v2 | http://arxiv.org/pdf/1705.06628v2.pdf | Robust tracking of respiratory rate in high-dynamic range scenes using mobile thermal imaging | The ability to monitor respiratory rate is extremely important for medical
treatment, healthcare and fitness sectors. In many situations, mobile methods,
which allow users to undertake every day activities, are required. However,
current monitoring systems can be obtrusive, requiring users to wear
respiration belts or nasal probes. Recent advances in thermographic systems
have shrunk their size, weight and cost, to the point where it is possible to
create smart-phone based respiration rate monitoring devices that are not
affected by lighting conditions. However, mobile thermal imaging is challenged
in scenes with high thermal dynamic ranges. This challenge is further amplified
by general problems such as motion artifacts and low spatial resolution,
leading to unreliable breathing signals. In this paper, we propose a novel and
robust approach for respiration tracking which compensates for the negative
effects of variations in the ambient temperature and motion artifacts and can
accurately extract breathing rates in highly dynamic thermal scenes. It has
three main contributions. The first is a novel Optimal Quantization technique
which adaptively constructs a color mapping of absolute temperature to improve
segmentation, classification and tracking. The second is the Thermal Gradient
Flow method that computes thermal gradient magnitude maps to enhance accuracy
of the nostril region tracking. Finally, we introduce the Thermal Voxel method
to increase the reliability of the captured respiration signals compared to the
traditional averaging method. We demonstrate the extreme robustness of our
system to track the nostril-region and measure the respiratory rate in high
dynamic range scenes. | ['Nadia Bianchi-Berthouze', 'Simon J. Julier', 'Youngjun Cho', 'Nicolai Marquardt'] | 2017-05-08 | null | null | null | null | ['physiological-computing', 'thermal-infrared-object-tracking'] | ['computer-vision', 'computer-vision'] | [ 5.86601794e-01 -5.45010626e-01 -2.46284697e-02 -1.29375577e-01
-5.46622276e-01 -5.24194658e-01 -1.01593859e-01 -1.74101800e-01
-3.65555555e-01 5.80573797e-01 -1.38398275e-01 2.88648438e-02
1.16373993e-01 -5.64463973e-01 4.81072254e-02 -9.93857622e-01
2.72736579e-01 -2.17731565e-01 4.32395488e-01 2.17412710e-01
1.15128748e-01 4.58160609e-01 -1.42292237e+00 -4.85703945e-02
1.06097364e+00 1.06916440e+00 2.12914407e-01 9.07386422e-01
1.33340538e-01 3.16981435e-01 -5.87087691e-01 8.22485164e-02
3.62055749e-02 -7.53265858e-01 -2.11534768e-01 -6.12894539e-03
3.44266385e-01 -5.25828123e-01 3.57170170e-03 7.25607514e-01
5.85492969e-01 2.69783944e-01 3.41978133e-01 -9.22913253e-01
-1.34676605e-01 -1.41212642e-01 -7.24375188e-01 3.28793794e-01
3.44547272e-01 1.69660330e-01 2.05210268e-01 -2.47425571e-01
4.69636954e-02 6.57301962e-01 7.63851583e-01 7.99639165e-01
-9.71401751e-01 -6.13231421e-01 -3.83219928e-01 9.99258906e-02
-1.25961947e+00 -4.60362256e-01 7.63857543e-01 -2.70679712e-01
6.49476051e-01 9.48791385e-01 8.76558304e-01 9.27506268e-01
3.93370479e-01 3.46602947e-02 1.51485729e+00 -1.53840601e-01
3.29612166e-01 3.06961954e-01 -3.47001821e-01 7.63301730e-01
2.86558747e-01 -9.39021483e-02 -3.68075967e-01 -9.46722701e-02
7.44156718e-01 2.22418606e-01 -7.30578661e-01 -2.44912639e-01
-1.05786121e+00 2.31763899e-01 2.30544567e-01 7.15573251e-01
-2.34536782e-01 3.46478313e-01 1.66417480e-01 -3.48875403e-01
3.29293579e-01 3.63552153e-01 -1.11727968e-01 -6.52294278e-01
-1.13811660e+00 -5.16772091e-01 4.82614338e-01 3.60900670e-01
3.17942679e-01 -5.99485748e-02 -1.73909828e-01 8.34524393e-01
4.94063437e-01 7.81113446e-01 7.33429193e-01 -1.13321435e+00
1.31064162e-01 3.79629970e-01 2.57764548e-01 -9.22544599e-01
-3.67312998e-01 -1.02099106e-01 -7.66484320e-01 4.31242660e-02
3.39718461e-01 5.27116470e-02 -8.38171721e-01 1.40070999e+00
7.69018233e-01 2.71267384e-01 -3.47144723e-01 1.08732319e+00
4.34364915e-01 3.31645340e-01 1.39087001e-02 -6.65225983e-01
1.40761709e+00 -5.86628437e-01 -1.13805997e+00 -1.20275766e-01
3.71864498e-01 -6.45795941e-01 1.20918500e+00 3.88125718e-01
-9.52317595e-01 -4.77498412e-01 -1.15144992e+00 3.80581357e-02
-1.49053067e-01 -1.10657616e-02 3.12304437e-01 1.40377855e+00
-9.23467636e-01 7.01951265e-01 -1.28652871e+00 -3.65885735e-01
1.71492428e-01 2.40552381e-01 1.04852088e-01 6.57112524e-02
-9.24089789e-01 6.48686469e-01 -6.06168732e-02 4.71070707e-01
-2.10794851e-01 -6.27653003e-01 -7.27051497e-01 -1.73213750e-01
3.22310835e-01 -5.56337476e-01 9.79181111e-01 -6.93548381e-01
-1.98092115e+00 4.75021571e-01 -3.72543216e-01 1.28664821e-01
6.75414443e-01 -2.45224893e-01 -4.65872854e-01 5.34264803e-01
-2.07764387e-01 1.31970391e-01 1.06155169e+00 -9.89031851e-01
-4.19574715e-02 -5.07352293e-01 -6.30109012e-01 3.71336520e-01
-5.31263888e-01 -1.54991701e-01 -5.60822546e-01 -2.09996626e-01
3.79582554e-01 -1.01432288e+00 -1.13527663e-01 2.53281564e-01
-2.92163738e-03 3.55727434e-01 1.03865540e+00 -6.28725231e-01
1.37580466e+00 -2.15474367e+00 -3.88444483e-01 1.36817068e-01
9.57363322e-02 2.59639382e-01 5.16120970e-01 -7.67244771e-02
2.67449945e-01 1.92833334e-01 -6.00312531e-01 -1.60190806e-01
-3.72393131e-01 1.90115064e-01 1.59452707e-01 7.05072939e-01
-3.78095359e-01 7.28209376e-01 -9.09713924e-01 -7.17854202e-01
8.05598140e-01 8.17554295e-01 5.37595130e-04 1.33121848e-01
2.99223065e-01 8.07627797e-01 -3.02492917e-01 6.20192230e-01
5.83846807e-01 1.71404313e-02 1.88331977e-01 -4.05976653e-01
-2.70769864e-01 2.25631401e-01 -1.08133662e+00 1.56819701e+00
-6.31241620e-01 3.50345314e-01 3.25389236e-01 -3.78616661e-01
7.59215176e-01 5.74887753e-01 9.78612483e-01 -9.51140225e-01
1.99456885e-01 2.81665057e-01 -4.25117940e-01 -9.73659337e-01
5.92604041e-01 -4.84984547e-01 1.60946190e-01 3.48687053e-01
-7.03287959e-01 -4.12966609e-01 -2.88299918e-01 -1.50444552e-01
8.12623024e-01 3.09925586e-01 1.31899267e-01 -2.88332738e-02
4.99312401e-01 -3.68171126e-01 5.52484989e-01 3.93273115e-01
-7.01944828e-01 7.29811728e-01 -1.40876159e-01 1.20526850e-02
-6.39955282e-01 -9.69038904e-01 -2.79127866e-01 5.57188094e-01
3.32654744e-01 -2.38815829e-01 -8.50440800e-01 -5.46534002e-01
-4.56244349e-01 4.43257898e-01 -4.45555270e-01 -1.62322626e-01
-7.10847318e-01 -8.70842397e-01 3.89230013e-01 5.25522470e-01
6.37907684e-01 -6.93127275e-01 -1.30849898e+00 1.80182993e-01
-6.95698500e-01 -1.17633581e+00 -6.41542137e-01 -6.80648610e-02
-1.14562690e+00 -9.58143175e-01 -7.06297696e-01 -6.66555837e-02
4.91618633e-01 4.28452432e-01 7.60383666e-01 -1.44404741e-02
-8.46546233e-01 7.40547240e-01 -2.57204115e-01 -2.44529665e-01
-1.54831514e-01 -2.42041439e-01 -9.29837599e-02 4.46107909e-02
1.63211953e-02 -4.62711155e-01 -1.22784698e+00 5.69312096e-01
-8.31737161e-01 -1.95753798e-01 8.66339654e-02 1.81673124e-01
7.24402606e-01 -5.12642367e-03 2.38721967e-01 -5.84327757e-01
3.10537100e-01 -1.66179746e-01 -4.15464818e-01 5.77898584e-02
-6.47082150e-01 -2.15586200e-01 5.09661317e-01 -2.95068145e-01
-1.06388271e+00 8.82528499e-02 -8.70918557e-02 -2.93840557e-01
-6.38359487e-02 -2.32798457e-01 7.79805183e-02 -2.19582558e-01
6.88074052e-01 1.53869197e-01 1.07647642e-03 -1.44024864e-01
9.32918265e-02 7.52957880e-01 7.26115942e-01 -2.23744512e-01
6.90354526e-01 8.43278587e-01 2.04468384e-01 -1.23111951e+00
-5.46187103e-01 -7.26610005e-01 -5.54997981e-01 -7.81725109e-01
1.15533054e+00 -5.38449049e-01 -9.18400645e-01 4.79961872e-01
-5.19407272e-01 -4.41158295e-01 -2.09578589e-01 7.65949309e-01
-3.50272536e-01 8.58096719e-01 -4.20508891e-01 -1.17622268e+00
-5.67296207e-01 -8.46315503e-01 9.29603219e-01 5.93041956e-01
-2.81301439e-01 -1.07537925e+00 1.13556303e-01 8.03989470e-01
7.94673324e-01 7.15951145e-01 3.24527889e-01 6.36204123e-01
-4.15464103e-01 8.32250044e-02 1.15796916e-01 2.29158029e-01
6.15498602e-01 1.54317189e-02 -1.34818757e+00 -2.96819866e-01
6.48064971e-01 -1.22251831e-01 5.30472577e-01 5.92591465e-01
1.56484950e+00 6.31288737e-02 -3.55947673e-01 7.37672806e-01
1.51684284e+00 3.51437360e-01 7.36268163e-01 -3.18390913e-02
8.28294694e-01 6.31014764e-01 7.33618796e-01 2.75568336e-01
5.19881696e-02 7.54729569e-01 4.62389678e-01 -2.39730492e-01
-1.03480831e-01 1.69540301e-01 3.43873918e-01 9.45233703e-01
-2.36087993e-01 -1.71874732e-01 -4.80894417e-01 3.36875707e-01
-1.30327237e+00 -7.50825047e-01 -5.21446645e-01 2.74842215e+00
7.83485413e-01 -3.68120372e-01 -3.93681899e-02 2.59510398e-01
5.75876176e-01 5.69826476e-02 -5.88290215e-01 -5.94073296e-01
5.52952826e-01 3.33730161e-01 6.75015807e-01 4.12724495e-01
-8.29507470e-01 2.21802413e-01 6.42632818e+00 4.07774419e-01
-1.44038820e+00 3.12810451e-01 4.64374900e-01 -2.81284630e-01
-2.72067279e-01 -4.69603479e-01 -3.53099316e-01 6.96563959e-01
1.11014485e+00 3.86667639e-01 4.04377282e-01 4.01971817e-01
7.03656495e-01 -7.59985149e-01 -7.27398515e-01 1.13876939e+00
1.47754148e-01 -3.37581426e-01 -7.15560555e-01 1.22392148e-01
3.04582030e-01 -1.93405077e-01 1.87832981e-01 -3.87040228e-01
-5.60405791e-01 -9.89682317e-01 1.40528485e-01 3.86601597e-01
1.18790185e+00 -5.00666082e-01 2.90206611e-01 1.16706714e-01
-1.28046644e+00 2.27136046e-01 -1.05111986e-01 4.29311693e-01
6.07015491e-02 8.65897655e-01 -8.73454809e-01 5.30238867e-01
8.13529730e-01 2.57006198e-01 -4.87446547e-01 1.13109720e+00
-1.25494361e-01 5.54519534e-01 -5.98322034e-01 -7.82447383e-02
-9.58981141e-02 -4.91531849e-01 3.17607105e-01 1.00345266e+00
5.52201033e-01 2.25010648e-01 -2.86171407e-01 8.83457541e-01
2.71485150e-01 5.28628677e-02 -4.27869350e-01 2.84150034e-01
2.28861302e-01 1.62036276e+00 -1.13131368e+00 -1.53367922e-01
-2.80887008e-01 1.28659153e+00 -4.29714948e-01 2.71854699e-01
-1.16702342e+00 -3.37405026e-01 2.83886194e-01 3.80406469e-01
-1.22790625e-02 -3.95090908e-01 -4.35726583e-01 -8.96959364e-01
3.57508153e-01 -5.46751499e-01 2.15596884e-01 -7.99529195e-01
-5.34290910e-01 1.77003115e-01 -1.89802632e-01 -1.18062627e+00
7.52770081e-02 -2.12291375e-01 -4.85591322e-01 7.51265943e-01
-1.51053035e+00 -4.97508913e-01 -8.65798891e-01 7.11848676e-01
3.93936515e-01 1.00273156e+00 8.20995212e-01 5.68211138e-01
-5.65446973e-01 4.43955988e-01 1.91536546e-01 -3.23512435e-01
8.06378901e-01 -1.30382788e+00 -1.70589298e-01 7.70385444e-01
-2.63861030e-01 5.33785105e-01 4.73880440e-01 -5.13804376e-01
-1.58970177e+00 -1.09984970e+00 3.75954121e-01 -4.83200133e-01
1.59689873e-01 -3.73470426e-01 -1.01087737e+00 -7.32135202e-04
-2.19227642e-01 2.41665378e-01 7.98394859e-01 -5.03492177e-01
9.84961092e-02 -4.44879651e-01 -1.58074343e+00 3.37463647e-01
4.49640125e-01 -5.57160079e-01 -5.78492917e-02 1.62589625e-01
2.60502845e-01 -5.93592048e-01 -1.04411495e+00 2.61286676e-01
9.62570906e-01 -1.04421949e+00 7.44890451e-01 6.83271766e-01
-1.37922168e-01 -6.36837184e-01 3.17029715e-01 -7.76881278e-01
5.29763438e-02 -7.82933235e-01 -1.30043119e-01 1.12340391e+00
6.68577328e-02 -7.22612798e-01 8.51158679e-01 8.70305240e-01
-5.59651898e-03 -4.33388174e-01 -1.08549452e+00 -6.55071616e-01
-6.01171494e-01 -4.66315448e-01 1.21016994e-01 8.51432443e-01
-1.51546039e-02 -5.43532260e-02 -3.30413818e-01 9.54409838e-02
7.45059252e-01 -9.93983671e-02 1.90460712e-01 -7.57867754e-01
-3.65083516e-01 5.10382354e-02 -1.96721405e-02 -7.67803907e-01
-6.66488171e-01 -2.81274915e-01 4.34959531e-01 -1.51132429e+00
6.27438873e-02 -4.61930633e-01 -3.83074939e-01 7.29952455e-02
-3.75170261e-01 8.04341435e-01 -1.54976010e-01 1.25345007e-01
-3.59598666e-01 2.02082440e-01 1.49435115e+00 2.11231589e-01
-5.79972684e-01 5.26440628e-02 -2.22772837e-01 5.00803471e-01
9.09453809e-01 -4.91130799e-01 -6.33460462e-01 -7.36361146e-02
2.31616274e-01 1.37294918e-01 2.53449291e-01 -1.29381955e+00
-3.01196259e-02 -8.03852677e-02 5.64831138e-01 -3.96369725e-01
4.94576305e-01 -1.06939781e+00 3.63631159e-01 7.56821454e-01
4.19947840e-02 -1.08397104e-01 6.48986623e-02 5.61158061e-01
1.67816669e-01 -1.15483940e-01 1.24421191e+00 -2.70049334e-01
-9.88259614e-02 1.13542549e-01 -6.30602598e-01 -1.26425609e-01
1.05175722e+00 -6.71197355e-01 -1.19896479e-01 -3.40310186e-01
-4.11201000e-01 -4.31047613e-03 6.77580953e-01 1.36600673e-01
5.39894104e-01 -1.05175948e+00 -7.37279430e-02 1.18246190e-01
-1.04371272e-02 -1.28211416e-02 5.45127809e-01 1.46069777e+00
-7.24545598e-01 1.25623122e-01 1.52681351e-01 -8.37952852e-01
-1.52820265e+00 5.13319373e-01 7.29214072e-01 -1.32903367e-01
-6.14200711e-01 5.44279993e-01 -1.86057448e-01 2.13296354e-01
-1.40049057e-02 -6.35636270e-01 -1.24560900e-01 -2.12708950e-01
5.73171616e-01 7.07310796e-01 2.07304955e-01 -3.44230980e-01
-4.30191368e-01 1.07250834e+00 4.51132178e-01 -1.74588501e-01
6.02483809e-01 -6.61247373e-01 -8.59555304e-02 6.98631763e-01
1.00650346e+00 2.21309513e-01 -1.20746756e+00 4.31875646e-01
-3.46772969e-01 -6.38701141e-01 8.21522474e-02 -9.13629532e-01
-1.08499467e+00 1.02460480e+00 1.22136664e+00 3.54494095e-01
1.54654491e+00 -4.77671266e-01 1.09974730e+00 -2.97295690e-01
8.75616893e-02 -1.34907508e+00 1.92823231e-01 -2.08512351e-01
1.66083887e-01 -9.16836143e-01 1.41038284e-01 -6.37597740e-01
-3.93079877e-01 1.13709641e+00 4.78802651e-01 4.54384655e-01
1.97642505e-01 3.27806681e-01 3.39585125e-01 -4.70229723e-02
-2.24343210e-01 -2.51133405e-02 1.66471168e-01 6.78140819e-01
6.46604955e-01 2.14272905e-02 -5.37027657e-01 -3.09133023e-01
6.72514737e-02 1.27307490e-01 5.46890140e-01 9.72479880e-01
-4.68224853e-01 -8.49143684e-01 -7.80564785e-01 3.71606469e-01
-1.01992726e+00 2.31605813e-01 -1.34740144e-01 4.25109744e-01
1.74375936e-01 1.29133260e+00 -1.30359262e-01 -1.72254264e-01
1.44494087e-01 1.91939592e-01 7.00910449e-01 -2.94138074e-01
-6.66602969e-01 4.17557091e-01 -1.86501473e-01 -8.48381400e-01
-7.29901612e-01 -4.48708385e-01 -1.25357282e+00 6.80037066e-02
-3.92676085e-01 3.57697904e-02 1.20782900e+00 6.55564010e-01
8.03419650e-02 7.24408031e-01 7.83516943e-01 -4.88667727e-01
-8.84771347e-02 -7.07483351e-01 -7.84514904e-01 2.81632841e-01
4.06643391e-01 -3.25622290e-01 -4.97972429e-01 -5.06330393e-02] | [13.900979995727539, 2.8118507862091064] |
e8a72f4b-c664-474f-98ec-4adccd1cb251 | peek-inside-the-closed-world-evaluating | 1912.05590 | null | https://arxiv.org/abs/1912.05590v3 | https://arxiv.org/pdf/1912.05590v3.pdf | Peek Inside the Closed World: Evaluating Autoencoder-Based Detection of DDoS to Cloud | Machine-learning-based anomaly detection (ML-based AD) has been successful at detecting DDoS events in the lab. However published evaluations of ML-based AD have used only limited data and provided minimal insight into why it works. To address limited evaluation against real-world data, we apply autoencoder, an existing ML-AD model, to 57 DDoS attack events captured at 5 cloud IPs from a major cloud provider. We show that our models detect nearly all malicious flows for 2 of the 4 cloud IPs under attack (at least 99.99%) and detect most malicious flows (94.75% and 91.37%) for the remaining 2 IPs. Our models also maintain near-zero false positives on benign flows to all 5 IPs. Our primary contribution is to improve our understanding for why ML-based AD works on some malicious flows but not others. We interpret our detection results with feature attribution and counterfactual explanation. We show that our models are better at detecting malicious flows with anomalies on allow-listed features (those with only a few benign values) than flows with anomalies on deny-listed features (those with mostly benign values) because our models are more likely to learn correct normality for allow-listed features. We then show that our models are better at detecting malicious flows with anomalies on unordered features (that have no ordering among their values) than flows with anomalies on ordered features because even with incomplete normality, our models could still detect anomalies on unordered feature with high recall. Lastly, we summarize the implications of what we learn on applying autoencoder-based AD in production: training with noisy real-world data is possible, autoencoder can reliably detect real-world anomalies on well-represented unordered features and combinations of autoencoder-based AD and heuristic-based filters can help both. | ['Anh Cao', 'John Heidemann', 'Hang Guo', 'Geoff Outhred', 'Xun Fan'] | 2019-12-11 | null | null | null | null | ['counterfactual-explanation'] | ['miscellaneous'] | [-5.61765611e-01 -2.57729977e-01 -5.22222295e-02 -2.01985806e-01
-1.99345082e-01 -7.46721148e-01 5.34449518e-01 7.43945986e-02
2.72473525e-02 4.92596447e-01 -1.10279039e-01 -9.68778074e-01
-7.98293799e-02 -1.00394917e+00 -5.40483832e-01 -3.73307377e-01
-7.33169317e-01 6.48782551e-01 1.85412034e-01 1.65319408e-03
1.30630732e-01 1.08273780e+00 -1.16601992e+00 6.34900928e-01
3.19641143e-01 9.78422046e-01 -9.91052270e-01 8.45251679e-01
-2.56942570e-01 1.16538656e+00 -1.34562504e+00 -4.97997075e-01
7.49922693e-01 -1.72543555e-01 -5.24089873e-01 -4.35902588e-02
5.11437833e-01 -9.07589495e-01 -7.00885415e-01 7.93572724e-01
9.64705124e-02 -2.88408995e-01 6.84315443e-01 -1.98500180e+00
-5.57201207e-01 5.01907289e-01 -2.21244708e-01 1.10445440e+00
4.66391146e-02 7.26797462e-01 1.18449700e+00 -4.29711819e-01
3.05296987e-01 1.29712439e+00 6.50666654e-01 3.93062055e-01
-1.28009653e+00 -8.53756547e-01 1.55474752e-01 2.41559371e-01
-9.11733329e-01 -1.60995260e-01 5.26912332e-01 -4.03344274e-01
1.33174825e+00 3.01808357e-01 5.39428234e-01 1.23171866e+00
3.97691131e-01 2.74445593e-01 7.53630221e-01 -1.37007888e-02
2.97925353e-01 1.10816218e-01 9.65877026e-02 3.21781933e-01
7.94624567e-01 5.28152347e-01 4.78144735e-02 -9.33534682e-01
6.70918763e-01 5.01950443e-01 5.71417138e-02 3.29314888e-01
-9.82701957e-01 1.21638608e+00 4.07673508e-01 4.08655077e-01
-7.83664823e-01 2.67643243e-01 5.02158225e-01 7.90293276e-01
2.67597705e-01 9.48776186e-01 -9.24820781e-01 -8.24610293e-02
-7.65373588e-01 1.75910175e-01 1.26733005e+00 5.49047351e-01
7.86167860e-01 1.09421182e+00 1.18532993e-01 2.57497698e-01
2.12355241e-01 6.80690467e-01 3.90811086e-01 -1.03837419e+00
3.90720442e-02 3.27336490e-01 -6.63055480e-02 -1.01599813e+00
-1.77283660e-01 -4.96627897e-01 -2.77643710e-01 3.03300112e-01
4.94880021e-01 -6.37847245e-01 -8.85278046e-01 1.35380769e+00
-3.28476578e-01 5.79121470e-01 1.25418454e-01 4.26209062e-01
6.08578399e-02 6.92213655e-01 1.60294846e-01 -7.88771510e-02
9.86102521e-01 -2.59482324e-01 -5.26288986e-01 -1.39991507e-01
6.84159219e-01 -5.46770155e-01 5.48468173e-01 2.49052882e-01
-3.28239650e-01 1.60413105e-02 -8.86260092e-01 1.07054675e+00
-6.38178527e-01 -6.65118694e-01 8.10622931e-01 1.06010246e+00
-7.93244898e-01 6.75154805e-01 -8.08565378e-01 -4.01463330e-01
6.87733769e-01 2.77709991e-01 -9.75201279e-02 2.58195788e-01
-1.19686556e+00 5.92148483e-01 6.97951689e-02 -3.08043480e-01
-1.23969936e+00 -8.80459845e-01 -5.31944513e-01 3.31731856e-01
1.48831740e-01 -1.63196385e-01 1.02591729e+00 -9.50818896e-01
-7.11406350e-01 2.88110405e-01 7.34177083e-02 -1.03359365e+00
2.45492950e-01 9.44992825e-02 -1.38958204e+00 4.20548320e-01
1.19896762e-01 1.89818040e-01 1.01409054e+00 -1.30124700e+00
-6.80071950e-01 -1.04002029e-01 -1.36540756e-01 -8.21112990e-01
-4.64134663e-01 8.23331997e-02 7.56549776e-01 -5.89308262e-01
-1.72432527e-01 -6.85442626e-01 -6.58073351e-02 -5.46997905e-01
-4.48220134e-01 6.48372397e-02 1.47305906e+00 -3.84163082e-01
1.24101222e+00 -1.78244269e+00 -1.31541729e+00 7.86960065e-01
4.03730303e-01 5.07284105e-01 -1.78230032e-01 5.90848923e-01
-4.39790785e-01 6.12328708e-01 -5.40773757e-02 4.04442042e-01
1.34165520e-02 3.49337786e-01 -1.24198306e+00 3.17861289e-01
6.85840666e-01 4.76967692e-01 -6.21165335e-01 -2.48927306e-02
4.07952666e-01 8.47618803e-02 -8.00351799e-01 4.07274395e-01
-1.33057147e-01 1.32607566e-02 -4.36765850e-01 1.06741309e+00
5.93563378e-01 -2.83885777e-01 -1.23833954e-01 1.17643684e-01
1.96129978e-01 1.67819098e-01 -9.39855933e-01 -2.44892389e-02
-3.48381549e-01 9.03668642e-01 -3.78725082e-01 -9.01339293e-01
9.78313327e-01 3.17501932e-01 7.51756430e-01 -4.51035053e-01
-9.53381881e-02 3.71264964e-01 5.71178019e-01 -4.29748654e-01
-2.06063077e-01 -1.77764773e-01 2.21073359e-01 1.06303322e+00
7.12942630e-02 1.99133307e-01 9.61199328e-02 4.18056399e-01
1.90501320e+00 -9.03175592e-01 1.65848240e-01 1.59681849e-02
3.54194880e-01 1.89376622e-01 7.85331726e-01 1.22413754e+00
-6.61150932e-01 3.02770615e-01 1.09381652e+00 -1.09490764e+00
-1.15337145e+00 -1.42175913e+00 -1.65770546e-01 7.16290057e-01
-3.99714738e-01 -2.72573709e-01 -2.77730465e-01 -1.37295806e+00
5.04126012e-01 1.07934487e+00 -4.55213308e-01 -1.62052065e-01
-5.93413174e-01 -1.23665023e+00 1.21573818e+00 3.97418857e-01
3.27409446e-01 -1.18043089e+00 -3.22584897e-01 3.94281745e-01
5.03530025e-01 -1.25514936e+00 1.03503183e-01 3.14320683e-01
-8.01802933e-01 -1.42758310e+00 2.35874712e-01 -1.28517315e-01
4.70943034e-01 -7.07247294e-03 1.29559207e+00 3.36989075e-01
-4.23397422e-01 4.70018238e-01 -1.30342767e-01 -7.57881403e-01
-6.72568798e-01 -3.23096156e-01 3.47915471e-01 2.14774176e-01
1.13226593e+00 -9.60277259e-01 -2.22174466e-01 3.04654390e-01
-8.43626440e-01 -1.49750280e+00 5.81353903e-01 7.91798592e-01
-6.85945302e-02 1.00301668e-01 1.07533932e+00 -9.99037564e-01
7.90257096e-01 -1.12298155e+00 -3.71146619e-01 -1.78187340e-01
-9.86407042e-01 -1.45843085e-02 1.05219781e+00 -4.57022220e-01
-6.15038455e-01 -5.18817425e-01 -4.66112867e-02 -1.08974838e+00
-6.27184331e-01 -1.72568202e-01 1.59206450e-01 5.57166524e-03
1.08584940e+00 5.62158599e-02 3.71510759e-02 -5.73425815e-02
-2.33022094e-01 7.80720890e-01 4.97336835e-02 -6.13944948e-01
1.02222943e+00 7.22549498e-01 -2.76490338e-02 -8.12996626e-01
-3.95247996e-01 -1.09948903e-01 -7.27918297e-02 7.18876570e-02
2.51495600e-01 -7.82615900e-01 -8.71215761e-01 1.97271615e-01
-9.10423517e-01 -4.75713052e-02 -3.70060384e-01 5.17584980e-01
-1.31943032e-01 2.01972678e-01 -1.01839697e+00 -8.00075412e-01
-1.80560648e-01 -1.02229440e+00 4.82080102e-01 -5.84875606e-02
-4.28999752e-01 -1.23566341e+00 1.28966287e-01 -1.81272000e-01
8.45359206e-01 9.55358520e-02 1.10296750e+00 -2.17462325e+00
-5.31872749e-01 -5.34562767e-01 -3.18349153e-01 6.60127699e-01
2.07251951e-01 5.42708158e-01 -1.15446484e+00 -2.22532123e-01
9.48863756e-03 1.04553394e-01 5.59649587e-01 1.75942302e-01
1.21530533e+00 -6.93264067e-01 -2.23886549e-01 4.05539960e-01
1.32621515e+00 6.42811120e-01 4.54988211e-01 4.70953912e-01
4.62616801e-01 3.30344111e-01 5.90685941e-02 7.07519710e-01
-2.42987454e-01 1.19245835e-02 8.36859107e-01 2.10746601e-01
3.99691671e-01 -1.36309400e-01 5.73549807e-01 8.64112526e-02
3.69036943e-01 -4.81980026e-01 -1.13470244e+00 6.62922025e-01
-1.13722622e+00 -1.37208080e+00 -3.59891802e-01 1.82441461e+00
2.54414361e-02 5.17775893e-01 4.81757611e-01 1.40144765e-01
8.59920561e-01 2.29533330e-01 -6.06858134e-01 -1.00448275e+00
-1.07887246e-01 3.86378378e-01 6.51723087e-01 4.65490967e-01
-1.20551717e+00 6.46359503e-01 7.16351986e+00 3.30019176e-01
-1.16555583e+00 -1.66612014e-01 7.32066989e-01 -9.72568691e-02
-2.06413269e-01 2.31876969e-01 -4.80601817e-01 8.07385623e-01
1.59540355e+00 -2.47313634e-01 3.87504011e-01 1.21662068e+00
9.90053713e-02 4.64439571e-01 -1.17680752e+00 6.70806706e-01
-1.09042622e-01 -1.38718534e+00 5.46782315e-01 3.42979789e-01
7.60156929e-01 1.06520541e-01 9.50615183e-02 5.63970745e-01
7.51665115e-01 -1.12616062e+00 1.53654054e-01 5.36949709e-02
1.47189587e-01 -8.76558602e-01 1.20191598e+00 -9.79729593e-02
-5.30791879e-01 -6.14606202e-01 -1.87312648e-01 -2.06303716e-01
-1.21265061e-01 1.06537569e+00 -1.36551476e+00 8.40740185e-03
6.02803707e-01 6.48578346e-01 -3.98839474e-01 7.47740567e-01
-1.18998006e-01 1.30060875e+00 -5.02400458e-01 1.61758319e-01
3.40892017e-01 2.50540853e-01 1.00213516e+00 1.33830166e+00
2.14342952e-01 -2.88856864e-01 2.11210594e-01 9.05614972e-01
3.59035544e-02 -3.58594000e-01 -1.02936125e+00 -3.37197602e-01
8.79016101e-01 1.01267934e+00 -4.88489389e-01 -4.10236806e-01
-4.22989458e-01 5.45256555e-01 -3.39494616e-01 5.70086956e-01
-6.15001798e-01 -6.60004258e-01 1.40903926e+00 3.32795203e-01
5.85992709e-02 2.75264055e-01 -3.86556804e-01 -1.30591023e+00
-2.28144169e-01 -1.06075692e+00 9.22453463e-01 -3.23015004e-01
-1.95811772e+00 7.37368047e-01 -3.02317619e-01 -1.38311338e+00
-8.21314991e-01 -8.69609356e-01 -1.25883889e+00 5.28636336e-01
-1.07541537e+00 -6.48943722e-01 1.03062958e-01 5.72458506e-01
2.78414816e-01 -7.75385678e-01 7.43023455e-01 2.81127572e-01
-7.42811978e-01 5.00042200e-01 -1.41394585e-01 8.50503087e-01
8.43901753e-01 -1.30737269e+00 5.24241745e-01 1.04227340e+00
3.74868244e-01 7.02888429e-01 7.06496835e-01 -7.97042847e-01
-8.82796884e-01 -1.18899703e+00 5.13852954e-01 -7.39753366e-01
1.10186100e+00 2.19948024e-01 -1.03146136e+00 1.43439054e+00
-3.78913693e-02 5.20708084e-01 7.16674626e-01 4.79907058e-02
-5.60368598e-01 -1.61488041e-01 -1.62702763e+00 4.27515417e-01
5.86002231e-01 -5.02939701e-01 -7.11832166e-01 3.63821536e-01
3.93617064e-01 3.01278710e-01 -8.61812115e-01 2.80305326e-01
1.37529194e-01 -1.21774006e+00 9.52611148e-01 -1.31635618e+00
1.77189812e-01 -3.35117966e-01 -2.50103503e-01 -1.37869859e+00
-1.65786564e-01 -3.97726864e-01 -4.89141852e-01 1.10710919e+00
4.18111235e-01 -1.35252178e+00 7.29198515e-01 3.75908643e-01
1.38652056e-01 -5.42760074e-01 -7.81385660e-01 -1.06400347e+00
2.77562663e-02 -7.83622026e-01 9.81751978e-01 1.42772222e+00
-2.52714336e-01 -3.26266021e-01 -2.76927017e-02 6.76891685e-01
7.35131800e-01 -4.11158986e-02 5.76335132e-01 -1.35921025e+00
-1.80304572e-01 -3.97228539e-01 -9.51136410e-01 -1.46086961e-01
3.98783863e-01 -5.98526061e-01 -7.44595587e-01 -5.95170736e-01
-2.72034556e-01 -4.03631151e-01 -3.94987136e-01 6.23730004e-01
1.57805473e-01 1.44470379e-01 1.76037103e-02 3.25657576e-01
-2.50693768e-01 1.67840272e-02 2.07236499e-01 -6.02654889e-02
-1.65166438e-01 -6.67521283e-02 -6.87908351e-01 8.01358640e-01
1.01511490e+00 -7.87210882e-01 -3.20168585e-03 -5.35177626e-02
-2.21929207e-01 -1.53752610e-01 6.89023674e-01 -1.09010172e+00
-9.94488075e-02 -2.41980791e-01 8.26285005e-01 -2.76154518e-01
-2.85013974e-01 -9.32038069e-01 -1.01243734e-01 8.00968766e-01
-1.58397719e-01 4.75300252e-01 2.27445871e-01 6.50410771e-01
-2.13726297e-01 6.13010824e-02 6.67371929e-01 -1.85228094e-01
-7.47696638e-01 4.61282969e-01 -8.85211885e-01 2.76802123e-01
1.20692980e+00 -2.40198672e-02 -6.16763890e-01 -7.89006531e-01
-8.57160985e-01 1.43625215e-01 2.44220078e-01 3.37448418e-01
4.77180243e-01 -1.22367966e+00 -7.13452220e-01 7.25092232e-01
-1.03554673e-01 -7.25377262e-01 -1.31140143e-01 3.95934522e-01
-8.57877195e-01 5.40769398e-01 -5.05501866e-01 -5.47480166e-01
-6.65167749e-01 6.10613108e-01 6.96118116e-01 -2.58055538e-01
-2.91908205e-01 2.59109914e-01 -2.47662529e-01 -2.93835878e-01
-1.23626791e-01 2.06425488e-01 2.92496830e-01 -2.06471533e-02
6.40009522e-01 7.54977643e-01 1.97742835e-01 -2.55350262e-01
-7.10183442e-01 -8.86203274e-02 -4.27416146e-01 2.25735437e-02
1.29996908e+00 4.86774743e-01 -9.30539742e-02 2.00445935e-01
1.10365570e+00 2.61795074e-01 -9.23214912e-01 2.08657414e-01
7.32114241e-02 -7.10640788e-01 -1.03737675e-01 -7.34754324e-01
-1.43592358e+00 9.99235570e-01 5.41013360e-01 7.87588716e-01
8.17285180e-01 -1.85994461e-01 6.67512059e-01 7.69173384e-01
1.09652087e-01 -5.99294126e-01 2.01127008e-01 4.40836519e-01
2.18821406e-01 -1.11886322e+00 -1.57146081e-01 -8.65956992e-02
-6.72260642e-01 1.40085542e+00 9.79467273e-01 -5.87873220e-01
7.51818776e-01 3.33717585e-01 3.44724149e-01 -3.02545965e-01
-1.16836953e+00 -4.50302958e-02 -5.30246019e-01 7.26000905e-01
-1.95687022e-02 1.19274691e-01 2.79515713e-01 3.72025043e-01
-1.93876043e-01 -5.84254920e-01 1.11513257e+00 7.83467770e-01
-3.87587011e-01 -8.08530509e-01 -6.20396316e-01 1.22623277e+00
-1.02915108e+00 -7.36756474e-02 -5.40028453e-01 1.04317582e+00
-1.16572514e-01 1.31057513e+00 7.81727612e-01 -6.82766140e-01
1.92637891e-01 3.99274737e-01 -4.03459400e-01 -4.64214355e-01
-7.51047909e-01 -3.61706465e-01 8.28882679e-02 -7.94300795e-01
3.74231845e-01 -7.76608467e-01 -1.10776782e+00 -8.41234028e-01
6.43360317e-02 2.59050786e-01 2.84187138e-01 7.89633632e-01
7.09190130e-01 4.81351703e-01 1.17746735e+00 -1.69577017e-01
-7.56580472e-01 -8.58469486e-01 -7.32128441e-01 5.78683436e-01
7.25717366e-01 -4.20013666e-01 -1.30200708e+00 -3.27783018e-01] | [5.3961873054504395, 7.386047840118408] |
6d98fd08-cd64-4cbf-b27d-6c140eb4a886 | std-stable-triangle-descriptor-for-3d-place | 2209.12435 | null | https://arxiv.org/abs/2209.12435v2 | https://arxiv.org/pdf/2209.12435v2.pdf | STD: Stable Triangle Descriptor for 3D place recognition | In this work, we present a novel global descriptor termed stable triangle descriptor (STD) for 3D place recognition. For a triangle, its shape is uniquely determined by the length of the sides or included angles. Moreover, the shape of triangles is completely invariant to rigid transformations. Based on this property, we first design an algorithm to efficiently extract local key points from the 3D point cloud and encode these key points into triangular descriptors. Then, place recognition is achieved by matching the side lengths (and some other information) of the descriptors between point clouds. The point correspondence obtained from the descriptor matching pair can be further used in geometric verification, which greatly improves the accuracy of place recognition. In our experiments, we extensively compare our proposed system against other state-of-the-art systems (i.e., M2DP, Scan Context) on public datasets (i.e., KITTI, NCLT, and Complex-Urban) and our self-collected dataset (with a non-repetitive scanning solid-state LiDAR). All the quantitative results show that STD has stronger adaptability and a great improvement in precision over its counterparts. To share our findings and make contributions to the community, we open source our code on our GitHub: https://github.com/hku-mars/STD. | ['Fu Zhang', 'Xiaoping Hong', 'Zuhao Zou', 'Jiarong Lin', 'Chongjian Yuan'] | 2022-09-26 | null | null | null | null | ['3d-place-recognition'] | ['computer-vision'] | [-2.55153716e-01 -5.45961261e-01 -2.11748868e-01 -4.31471199e-01
-7.09745944e-01 -6.63600624e-01 5.94834208e-01 2.12071046e-01
-2.18312070e-01 3.32396001e-01 -1.78351954e-01 -2.08138838e-01
-7.32883289e-02 -1.02236295e+00 -6.10380709e-01 -5.21282852e-01
-1.12029977e-01 5.45132637e-01 5.28966486e-01 -3.05183321e-01
6.22058034e-01 1.00706959e+00 -1.75032747e+00 -3.67590964e-01
8.50760281e-01 1.14933932e+00 1.21405020e-01 -5.34026287e-02
-3.70004959e-02 -8.98952559e-02 -7.62571469e-02 -2.42583989e-03
5.00747621e-01 2.58711278e-01 -3.74825299e-01 -2.47621790e-01
4.23282325e-01 -1.73283771e-01 -2.81022042e-01 9.91146743e-01
5.17690659e-01 6.83621615e-02 6.05717003e-01 -1.41551101e+00
-5.89088142e-01 -9.81627405e-02 -6.02558613e-01 -1.67062327e-01
6.59313679e-01 -5.69721349e-02 8.32378209e-01 -1.39730382e+00
6.76412046e-01 1.09372866e+00 8.27131331e-01 -1.08499065e-01
-9.60083663e-01 -1.08236802e+00 -1.64201781e-01 4.45047140e-01
-2.11818385e+00 -5.44924557e-01 8.66328061e-01 -3.15405190e-01
6.29277825e-01 4.01631981e-01 5.76790929e-01 6.87609613e-01
1.18766971e-01 4.29120094e-01 9.11329389e-01 -3.15678656e-01
1.07574768e-01 -2.87573516e-01 -8.01033974e-02 7.24376500e-01
3.25668454e-01 3.01519334e-01 -3.61954868e-01 -2.59880841e-01
8.37104440e-01 2.75381446e-01 -6.20349757e-02 -8.05641651e-01
-1.51247990e+00 5.62887967e-01 6.93844020e-01 1.99967846e-01
-2.06043884e-01 -6.76700175e-02 9.24691707e-02 9.09026861e-02
1.07452199e-01 -1.39470190e-01 -5.33735119e-02 -7.04681873e-02
-7.02957690e-01 2.53735691e-01 4.63259250e-01 1.54814112e+00
1.21387661e+00 -3.10299903e-01 2.83209592e-01 7.58401275e-01
5.73569834e-01 1.14711678e+00 2.14062974e-01 -6.01385593e-01
4.94188964e-01 8.07949007e-01 1.11203417e-01 -1.45992684e+00
-4.38260049e-01 1.34893686e-01 -8.19425225e-01 -7.25691542e-02
1.26959365e-02 3.62887353e-01 -7.70335257e-01 1.20795131e+00
6.28019512e-01 3.80158782e-01 -1.67643845e-01 7.88065195e-01
9.47074115e-01 4.40540224e-01 -3.96100909e-01 2.86130786e-01
1.29228234e+00 -5.25011718e-01 -2.87089527e-01 -1.33709043e-01
5.44450998e-01 -1.01220608e+00 6.02167070e-01 -1.17583491e-01
-5.01772165e-01 -6.03268504e-01 -1.12240958e+00 -2.41289362e-01
-3.38515431e-01 2.70340264e-01 2.47375399e-01 2.63718337e-01
-8.03787410e-01 4.30856913e-01 -9.49608803e-01 -6.59237921e-01
9.82539207e-02 4.46130991e-01 -6.71090543e-01 -1.23019338e-01
-9.09357011e-01 8.66282821e-01 2.67723888e-01 1.14195399e-01
-2.67906278e-01 -5.28472006e-01 -1.02438307e+00 -3.65999401e-01
1.16285913e-01 -4.12977874e-01 9.32619214e-01 6.96948543e-03
-1.20868158e+00 9.87611473e-01 -5.60567617e-01 -5.78209478e-03
3.66074353e-01 5.18753007e-03 -4.94002789e-01 -5.56812063e-03
5.55292666e-01 5.26341796e-01 4.51273710e-01 -1.27460802e+00
-6.96999729e-01 -6.91490173e-01 -3.49428833e-01 4.47397493e-02
1.31412014e-01 -1.59990504e-01 -5.71473181e-01 -2.08200961e-01
8.87753606e-01 -1.29686701e+00 -7.12771341e-02 3.32125306e-01
-4.41568822e-01 -2.13532194e-01 1.01759672e+00 -2.76422024e-01
9.70883310e-01 -2.50848293e+00 -3.16518009e-01 6.87600493e-01
-1.10226601e-01 5.82891405e-02 9.14875269e-02 6.36734784e-01
4.55822088e-02 -1.18214311e-02 -2.28646174e-01 -2.62088388e-01
1.20002322e-01 2.88474619e-01 -2.15014502e-01 8.32969308e-01
-8.79945010e-02 8.29940259e-01 -8.96688104e-01 -6.30345583e-01
5.24547279e-01 4.23062772e-01 -1.24426976e-01 -1.70165300e-01
3.89608145e-01 4.38766927e-01 -7.82935977e-01 1.08353686e+00
1.16955376e+00 1.22478753e-01 -1.54098064e-01 -2.41866782e-01
-5.66698372e-01 2.18272239e-01 -1.56508410e+00 1.62935257e+00
-2.38463044e-01 2.72153080e-01 -2.46627405e-01 -4.71584439e-01
1.59611607e+00 3.67596769e-03 5.95656812e-01 -8.27240169e-01
-9.67634469e-03 6.49752915e-01 -3.15408647e-01 4.02064063e-02
6.28113925e-01 3.65523815e-01 -8.76554474e-02 1.95493564e-01
-4.13821578e-01 -3.97034824e-01 -1.21907599e-01 -3.19763571e-01
7.49357402e-01 -5.40146828e-02 5.33944786e-01 -2.62180239e-01
7.27722943e-01 1.35266766e-01 4.52923268e-01 3.29867154e-01
-2.43740499e-01 7.09731221e-01 -6.95479289e-02 -3.54004353e-01
-1.05563343e+00 -1.19488788e+00 -6.05642676e-01 3.55813026e-01
6.34931326e-01 -4.80136245e-01 -1.73267215e-01 -2.34741390e-01
5.73772013e-01 3.80462021e-01 -3.95343065e-01 4.61625755e-02
-6.36572540e-01 4.17799689e-02 6.09498143e-01 5.49253464e-01
7.71569490e-01 -6.49969757e-01 -4.62388992e-01 -7.24153891e-02
-1.09515086e-01 -1.10488272e+00 -5.16996682e-01 -2.03596175e-01
-9.48116720e-01 -1.06174135e+00 -3.68282437e-01 -9.00001407e-01
7.31413186e-01 8.80961478e-01 7.05047011e-01 1.86785191e-01
1.10061519e-01 1.82572201e-01 -3.36710244e-01 -3.60705584e-01
1.39308408e-01 1.93925083e-01 2.78112352e-01 -1.65251240e-01
5.26804984e-01 -8.21684897e-01 -7.03740776e-01 9.47449207e-01
-5.72904348e-01 -1.48868114e-01 4.77576494e-01 4.53324676e-01
1.19200206e+00 -4.29795265e-01 -6.62174150e-02 -2.35460669e-01
2.19789952e-01 -4.89435941e-01 -8.52162182e-01 4.81179729e-02
-4.98956114e-01 -4.46955375e-02 1.79844409e-01 3.56507301e-02
-3.31979066e-01 3.06825280e-01 -1.74661264e-01 -5.82311153e-01
-2.77506888e-01 4.82179314e-01 -1.63222224e-01 -5.39178133e-01
4.37632859e-01 4.30958003e-01 4.12470987e-03 -5.05369067e-01
3.58452141e-01 8.98576260e-01 6.03931606e-01 -7.09304690e-01
1.27682984e+00 7.16480017e-01 3.59342545e-01 -9.12743449e-01
-3.03467989e-01 -7.95199573e-01 -1.19233608e+00 -8.76869354e-03
5.30658424e-01 -8.82880747e-01 -7.92535722e-01 4.41714019e-01
-1.13077998e+00 3.15682501e-01 1.50563670e-02 4.12386000e-01
-5.82724392e-01 5.36313236e-01 3.57658081e-02 -6.53985262e-01
-3.23599786e-01 -1.08694565e+00 1.33637452e+00 1.68149874e-01
-4.15726341e-02 -7.43200839e-01 3.53521258e-01 -1.18760159e-02
2.22767621e-01 5.31695068e-01 5.44740617e-01 -4.86665577e-01
-8.23875844e-01 -4.36425596e-01 -2.80187786e-01 -1.49509519e-01
2.29196936e-01 3.66738528e-01 -6.77435219e-01 -1.63509324e-01
-3.21328938e-01 2.55355835e-01 4.35661167e-01 -1.15431920e-02
8.89663398e-01 1.47698328e-01 -6.32438660e-01 8.44700813e-01
1.34757769e+00 2.00128093e-01 7.28106678e-01 7.79066563e-01
7.36479819e-01 1.57495856e-01 1.07230926e+00 4.98996675e-01
8.34208012e-01 1.01112020e+00 4.13462132e-01 9.32431296e-02
2.52957612e-01 -5.62713444e-01 1.69060826e-01 8.85690689e-01
-3.88412178e-01 3.10485303e-01 -1.38283074e+00 4.65384215e-01
-1.89695621e+00 -8.94261599e-01 -5.20830929e-01 2.39629412e+00
4.34428900e-01 -6.82162195e-02 -1.45689383e-01 1.25967428e-01
9.39275026e-01 2.60735333e-01 -3.54914099e-01 -1.19231410e-01
-8.31313804e-02 2.60663718e-01 8.47309828e-01 3.23358923e-01
-1.09061337e+00 9.62297320e-01 5.66551447e+00 7.68688083e-01
-1.30726373e+00 -8.19430426e-02 -9.94319543e-02 3.47118288e-01
-4.53579910e-02 1.35627523e-01 -8.23727548e-01 4.80035990e-01
7.09753215e-01 -5.04990816e-01 1.62668467e-01 1.06218314e+00
1.61889523e-01 -5.17162494e-02 -1.09208369e+00 1.14780128e+00
-9.51525792e-02 -1.14859629e+00 -1.14669517e-01 2.70584941e-01
6.18949354e-01 3.87394726e-01 9.03251097e-02 4.88043297e-03
-3.91464196e-02 -8.04278016e-01 8.27792346e-01 4.27844375e-01
1.08069706e+00 -8.38253200e-01 7.16357410e-01 2.98330426e-01
-1.80923665e+00 2.12307706e-01 -6.10843003e-01 9.53380018e-02
1.23369649e-01 4.58433181e-01 -8.02170336e-01 1.03167796e+00
5.34600854e-01 1.00532639e+00 -6.22123659e-01 1.27184057e+00
-2.50005573e-01 1.79224219e-02 -7.09138632e-01 2.20039338e-01
1.45373940e-01 -4.65218693e-01 5.43371141e-01 8.61008704e-01
8.26085567e-01 1.33455887e-01 2.71704316e-01 7.48943388e-01
5.24690785e-02 2.82124996e-01 -9.22154844e-01 4.14171576e-01
1.08453417e+00 1.19940031e+00 -4.31916296e-01 -6.04419634e-02
-3.94759506e-01 6.23391807e-01 7.42885917e-02 -8.36056396e-02
-7.78935373e-01 -4.98405874e-01 9.08682227e-01 2.94399261e-01
3.97860318e-01 -8.26153696e-01 -4.10009325e-01 -9.60450351e-01
2.18288213e-01 -4.84671891e-01 8.13016146e-02 -7.66393244e-01
-1.04735124e+00 4.56836700e-01 5.27341664e-02 -1.87321818e+00
-1.15737431e-01 -3.97256970e-01 -6.06054127e-01 8.65118086e-01
-1.57059979e+00 -1.21653414e+00 -6.59607530e-01 8.98143768e-01
-9.64744017e-02 1.52751192e-01 7.53220081e-01 5.35515189e-01
-2.18990341e-01 5.20954251e-01 3.83711815e-01 3.60193908e-01
7.88812101e-01 -7.75945008e-01 8.61730874e-01 6.18378282e-01
2.00629160e-01 9.05923128e-01 2.75120944e-01 -7.82112122e-01
-1.66014934e+00 -1.04030645e+00 9.19351161e-01 -5.37786067e-01
6.86839283e-01 -3.62993479e-01 -7.10968614e-01 7.57157981e-01
-5.21430731e-01 2.25319847e-01 4.09840018e-01 -8.38236436e-02
-3.62928689e-01 -2.41360888e-01 -1.19773281e+00 3.68860304e-01
1.34075725e+00 -5.62166870e-01 -5.54348409e-01 1.68000802e-01
5.20248592e-01 -8.24472249e-01 -1.05260336e+00 5.98723114e-01
6.91704690e-01 -9.45312023e-01 1.14517140e+00 1.11021154e-01
3.96220274e-02 -7.76520550e-01 -5.51114321e-01 -9.68778610e-01
-4.30787951e-01 -1.94681048e-01 2.40092874e-01 1.12459481e+00
8.05146098e-02 -9.79315877e-01 5.79650283e-01 3.42260212e-01
-2.80164540e-01 -6.31638229e-01 -1.29979253e+00 -9.98552620e-01
-1.99430540e-01 -4.58677381e-01 1.17505968e+00 8.54491591e-01
-2.94440061e-01 -1.64775982e-01 -4.19232398e-02 6.21008933e-01
6.66612685e-01 4.51524168e-01 1.17112875e+00 -1.45755780e+00
6.45871520e-01 -1.46411881e-01 -1.12104356e+00 -1.11088204e+00
7.10856244e-02 -1.02513039e+00 -2.66566761e-02 -1.35146856e+00
-1.81662425e-01 -9.63371217e-01 -1.31918430e-01 5.02244473e-01
2.44262889e-01 3.75464410e-01 2.33151868e-01 7.49528170e-01
-4.86749470e-01 7.10067451e-01 1.16527176e+00 7.54439235e-02
-1.72424629e-01 3.47085968e-02 -3.20669174e-01 6.07309937e-01
7.25659132e-01 -5.17283559e-01 1.00976869e-01 -4.25075531e-01
-1.59526601e-01 -1.68285295e-01 4.52506483e-01 -1.23324645e+00
4.18833941e-01 -3.72100264e-01 2.74821222e-01 -1.23240650e+00
3.47023934e-01 -1.09061909e+00 5.16121447e-01 4.20527220e-01
3.50141376e-01 4.52966541e-01 8.36612210e-02 3.35809052e-01
-2.76119739e-01 -3.34617421e-02 5.78273833e-01 7.18128830e-02
-8.39330733e-01 6.44627333e-01 3.24672490e-01 -3.55654716e-01
1.24751234e+00 -5.84341645e-01 -3.38588953e-01 -1.13877721e-01
-2.53152251e-01 3.72899383e-01 1.06618333e+00 4.47966397e-01
8.59508038e-01 -1.90658855e+00 -6.14985883e-01 4.74150062e-01
6.39248908e-01 2.80657172e-01 -5.66904768e-02 9.92460012e-01
-7.33715117e-01 6.15565658e-01 -3.79626185e-01 -1.12297809e+00
-1.23168373e+00 3.03313732e-01 9.78068039e-02 3.54838729e-01
-6.76366210e-01 2.46490449e-01 -2.90716201e-01 -8.00921917e-01
-1.23056717e-01 -5.54006100e-01 1.17785491e-01 -1.75795425e-02
3.00825328e-01 4.23634917e-01 2.78992504e-01 -1.26594627e+00
-9.21603739e-01 1.28998399e+00 2.76992381e-01 3.48578170e-02
1.35056698e+00 -1.01546846e-01 -1.88992575e-01 4.03481722e-01
1.23832595e+00 3.45668137e-01 -6.87809467e-01 -3.70675594e-01
-9.79495980e-03 -9.48691666e-01 -2.70769387e-01 -7.72733167e-02
-9.15846467e-01 6.41718507e-01 7.64814138e-01 -9.39650908e-02
6.71261489e-01 7.75536224e-02 8.38109076e-01 5.24141252e-01
9.91672099e-01 -6.46719575e-01 -3.95332694e-01 8.01196158e-01
9.86986339e-01 -1.23140287e+00 1.99598238e-01 -5.80398262e-01
-3.10608149e-01 1.17968202e+00 2.57101297e-01 -3.25052947e-01
7.81476796e-01 1.06572062e-02 9.32149664e-02 -2.27595866e-01
-1.30647406e-01 -3.16439092e-01 3.79926980e-01 7.76980937e-01
1.23004176e-01 1.88983157e-01 -1.82121202e-01 1.06985584e-01
-6.25491858e-01 -9.00686756e-02 1.06367514e-01 1.04545617e+00
-6.00874543e-01 -1.20965505e+00 -7.05118179e-01 8.86632353e-02
3.64831179e-01 1.11586049e-01 -3.20354253e-01 9.28288996e-01
1.18664108e-01 7.16332734e-01 1.52780265e-01 -6.45978749e-01
7.12562561e-01 -1.39923975e-01 2.60933071e-01 -4.51465011e-01
-7.01458454e-02 -3.41414958e-01 -1.82996839e-01 -8.41243207e-01
-2.89876103e-01 -9.51985419e-01 -1.49133527e+00 -7.86934555e-01
-2.96441436e-01 5.22723421e-02 9.41464126e-01 5.92530608e-01
6.64049089e-01 -2.62397021e-01 8.35990548e-01 -1.02589500e+00
-4.91828650e-01 -5.93720257e-01 -5.83318770e-01 2.95523703e-01
9.12909657e-02 -1.15675998e+00 -2.09921002e-01 -3.79975319e-01] | [7.640600204467773, -2.5790302753448486] |
741e7053-ce4d-438d-8998-f755582c299f | enforcing-mutual-consistency-of-hard-regions | 2109.09960 | null | https://arxiv.org/abs/2109.09960v4 | https://arxiv.org/pdf/2109.09960v4.pdf | Mutual Consistency Learning for Semi-supervised Medical Image Segmentation | In this paper, we propose a novel mutual consistency network (MC-Net+) to effectively exploit the unlabeled data for semi-supervised medical image segmentation. The MC-Net+ model is motivated by the observation that deep models trained with limited annotations are prone to output highly uncertain and easily mis-classified predictions in the ambiguous regions (e.g., adhesive edges or thin branches) for medical image segmentation. Leveraging these challenging samples can make the semi-supervised segmentation model training more effective. Therefore, our proposed MC-Net+ model consists of two new designs. First, the model contains one shared encoder and multiple slightly different decoders (i.e., using different up-sampling strategies). The statistical discrepancy of multiple decoders' outputs is computed to denote the model's uncertainty, which indicates the unlabeled hard regions. Second, we apply a novel mutual consistency constraint between one decoder's probability output and other decoders' soft pseudo labels. In this way, we minimize the discrepancy of multiple outputs (i.e., the model uncertainty) during training and force the model to generate invariant results in such challenging regions, aiming at regularizing the model training. We compared the segmentation results of our MC-Net+ model with five state-of-the-art semi-supervised approaches on three public medical datasets. Extension experiments with two standard semi-supervised settings demonstrate the superior performance of our model over other methods, which sets a new state of the art for semi-supervised medical image segmentation. Our code is released publicly at https://github.com/ycwu1997/MC-Net. | ['Jianfei Cai', 'Yong Xia', 'Lei Zhang', 'Minfeng Xu', 'Donghao Zhang', 'ZongYuan Ge', 'Yicheng Wu'] | 2021-09-21 | null | null | null | null | ['semi-supervised-medical-image-segmentation'] | ['computer-vision'] | [ 4.32607889e-01 8.37137759e-01 -4.97515082e-01 -6.20025694e-01
-1.13377917e+00 -2.67506570e-01 1.64333418e-01 -2.00001583e-01
-3.10833186e-01 6.98637605e-01 -6.32284582e-02 -2.90025175e-01
3.15764725e-01 -5.39143622e-01 -8.46786380e-01 -8.69291306e-01
3.63571525e-01 6.67877734e-01 3.88017356e-01 3.23846549e-01
-1.74279660e-01 -7.47872069e-02 -9.69721019e-01 3.79063189e-01
1.30059266e+00 9.96926129e-01 4.69256043e-01 4.41287071e-01
5.25651574e-02 5.88419497e-01 -3.78743798e-01 -3.37977469e-01
1.29426301e-01 -5.90689301e-01 -9.05849099e-01 3.31578463e-01
1.81031227e-02 -3.07508111e-01 -1.75756529e-01 1.38973129e+00
4.16328639e-01 -2.72612840e-01 6.06977046e-01 -9.75894809e-01
-5.65148056e-01 8.86367559e-01 -6.37149990e-01 -9.56814587e-02
-2.03386977e-01 2.40486905e-01 8.08447421e-01 -6.48535311e-01
6.02684736e-01 8.70248675e-01 6.25656188e-01 8.38663340e-01
-1.17797685e+00 -6.65567577e-01 1.26630232e-01 -2.00404927e-01
-1.50598896e+00 -2.16496766e-01 6.86047256e-01 -4.36404556e-01
2.15495884e-01 2.14186937e-01 3.87501329e-01 1.07894170e+00
2.54016846e-01 1.37235022e+00 1.10141373e+00 -2.14929789e-01
2.74870962e-01 3.29359204e-01 2.08212584e-01 9.59580183e-01
1.39930949e-01 3.02399173e-02 -1.05524160e-01 -9.39275473e-02
6.69189990e-01 4.82400181e-03 -3.77662390e-01 -1.73426583e-01
-1.13715804e+00 6.64806843e-01 4.03850734e-01 2.84039587e-01
-1.89180020e-02 -4.63239588e-02 2.70291924e-01 -1.86099216e-01
5.97078085e-01 1.08062424e-01 -5.88979125e-01 2.64637709e-01
-1.21570921e+00 -1.74581230e-01 5.25748670e-01 1.08682442e+00
6.98567390e-01 -3.34639668e-01 -4.24759775e-01 8.96683693e-01
6.86911941e-01 3.87359798e-01 5.78757048e-01 -7.91382194e-01
4.41401273e-01 5.47214270e-01 -3.06289166e-01 -5.94745934e-01
-4.96135920e-01 -6.31117940e-01 -1.04845309e+00 -1.48006037e-01
3.89791667e-01 -3.44326168e-01 -1.43590879e+00 1.81252539e+00
3.57708097e-01 3.31281632e-01 2.56458577e-02 9.13238823e-01
8.06460738e-01 4.65071559e-01 -9.37314704e-02 -1.55605540e-01
1.08294713e+00 -1.25562382e+00 -7.03345954e-01 -2.72359401e-01
7.88283765e-01 -5.12165427e-01 8.11717212e-01 2.22004265e-01
-1.04312527e+00 -4.04168725e-01 -1.00267720e+00 1.70986995e-01
-4.32775803e-02 4.29992020e-01 3.95843953e-01 5.12965977e-01
-8.05063784e-01 6.24920547e-01 -1.36559546e+00 1.92919135e-01
7.79774487e-01 3.66762310e-01 -4.78730090e-02 -5.24828359e-02
-1.10241830e+00 5.33260286e-01 5.31988204e-01 2.65601546e-01
-9.07051861e-01 -5.46181321e-01 -1.02523744e+00 -3.40889245e-01
5.86439490e-01 -5.66852093e-01 1.24000084e+00 -1.14254534e+00
-1.48940492e+00 1.10526800e+00 -2.25289792e-01 -4.10122097e-01
8.69547427e-01 -7.78268045e-03 -1.45301476e-01 1.48429528e-01
2.94748485e-01 9.97193098e-01 7.50197232e-01 -1.51405251e+00
-4.47350681e-01 -2.91881591e-01 -3.71836275e-01 4.40023430e-02
1.70248724e-03 -3.55874240e-01 -9.37355518e-01 -7.78930366e-01
4.24475551e-01 -1.16401875e+00 -6.11796856e-01 1.43277958e-01
-1.12538576e+00 2.28793416e-02 4.48251605e-01 -5.87175786e-01
1.18133724e+00 -2.13857889e+00 9.00913924e-02 4.73115206e-01
3.36898983e-01 2.67035872e-01 6.73984289e-02 -3.71888727e-01
1.89838052e-01 4.21309412e-01 -1.02384210e+00 -6.46263063e-01
-1.38860106e-01 5.24868667e-01 1.05660960e-01 4.43433046e-01
2.47354209e-01 1.03363013e+00 -9.44756627e-01 -1.05188382e+00
1.31450504e-01 3.32021475e-01 -4.46834773e-01 2.23114520e-01
-3.14634085e-01 8.63519192e-01 -6.68317854e-01 8.76739800e-01
7.91891515e-01 -7.56486535e-01 2.54716605e-01 7.96532724e-04
3.85067165e-01 1.00921825e-01 -1.07978594e+00 1.74478781e+00
-1.91860646e-01 1.79111391e-01 3.54091972e-02 -7.99505413e-01
5.61987817e-01 3.98657262e-01 4.81664270e-01 -3.68317753e-01
2.44515255e-01 4.17904943e-01 -6.24159388e-02 -3.26212019e-01
1.51624121e-02 -6.87138885e-02 7.32317939e-02 4.06890422e-01
1.40963927e-01 -1.45379379e-01 -3.77251208e-02 1.72687724e-01
7.19528854e-01 2.03805700e-01 1.41900048e-01 -2.80010134e-01
4.62724477e-01 -1.57263070e-01 1.12706077e+00 7.93954074e-01
-4.96256620e-01 1.02989233e+00 5.15555859e-01 -7.46743381e-02
-7.25999355e-01 -1.06908476e+00 -5.23828983e-01 4.60652411e-01
4.02904689e-01 -1.96984470e-01 -1.00265181e+00 -1.22727776e+00
-2.03827590e-01 5.22766650e-01 -8.15719366e-01 -1.81897953e-01
-4.23040330e-01 -9.43606377e-01 6.62136555e-01 5.61520696e-01
5.24915874e-01 -8.77425373e-01 -3.06026757e-01 1.43639445e-01
-3.68030339e-01 -1.12233901e+00 -7.01241314e-01 3.76075178e-01
-9.89551544e-01 -1.08023775e+00 -1.02386355e+00 -7.84705043e-01
1.16289043e+00 -1.82686016e-01 1.06866634e+00 2.69273639e-01
-1.49389535e-01 -1.06707752e-01 -2.94939786e-01 -3.30244631e-01
-7.15379238e-01 1.53695330e-01 -2.73884624e-01 -8.26199260e-03
9.99130607e-02 -1.05716109e-01 -8.21042895e-01 6.97673440e-01
-1.10171282e+00 5.18509388e-01 6.58275485e-01 1.19667649e+00
1.21776533e+00 -9.51801166e-02 4.76513177e-01 -1.30655921e+00
1.29988909e-01 -6.05117083e-01 -4.52796131e-01 4.23312277e-01
-7.35640526e-01 2.19979987e-01 4.79516149e-01 -3.78182024e-01
-1.07396710e+00 3.73599052e-01 -4.28046376e-01 -3.30971658e-01
-1.16888352e-01 4.12336737e-01 -2.03927860e-01 1.75048262e-01
3.69028270e-01 1.75748229e-01 5.13068214e-02 -2.57062972e-01
1.76965281e-01 7.57270217e-01 4.55663204e-01 -4.71220732e-01
5.03104508e-01 5.85632443e-01 -3.19446385e-01 -2.41684183e-01
-1.19669759e+00 -4.50297177e-01 -5.39930046e-01 -4.21532765e-02
1.05210102e+00 -8.38508248e-01 -1.55008927e-01 6.23556852e-01
-9.47307825e-01 -7.19000816e-01 -2.26989523e-01 4.89517778e-01
-4.69348639e-01 4.70856458e-01 -8.45549047e-01 -4.85570937e-01
-3.94742489e-01 -1.62101996e+00 1.35246313e+00 3.85539830e-01
-1.05117172e-01 -1.06590652e+00 -4.44159955e-02 6.18778765e-01
3.28762121e-02 2.16843247e-01 5.72523713e-01 -7.97739625e-01
-4.07862365e-01 -3.79034802e-02 -1.68447167e-01 6.55060172e-01
1.75139293e-01 -1.26099365e-03 -1.08613992e+00 -9.26235244e-02
1.12660583e-02 -4.67123985e-01 1.15069532e+00 6.75131261e-01
1.60765457e+00 -1.27585426e-01 -7.10610807e-01 8.37097645e-01
1.14104152e+00 2.31524780e-02 6.64102495e-01 6.78494200e-02
7.12643564e-01 3.14396501e-01 7.18756795e-01 2.95743763e-01
3.93840730e-01 4.62280095e-01 3.81911367e-01 -4.75675106e-01
-2.60259886e-03 -2.22283959e-01 8.55076984e-02 7.59021759e-01
3.65245014e-01 -2.85266966e-01 -8.91445756e-01 4.59146172e-01
-1.91390169e+00 -3.52827638e-01 -5.51184043e-02 2.03614616e+00
1.37776601e+00 2.69879073e-01 -2.50616521e-01 -2.11017594e-01
8.43902886e-01 7.15158433e-02 -7.85502374e-01 1.16724685e-01
1.24445744e-02 1.11333370e-01 5.54439187e-01 5.83599985e-01
-1.30263591e+00 9.02108431e-01 5.52928495e+00 1.03564215e+00
-1.00599575e+00 2.05768645e-01 1.33468592e+00 1.44097209e-01
-4.18065012e-01 -2.12177008e-01 -7.89212048e-01 8.75521302e-01
6.62982821e-01 3.38657588e-01 -1.10506862e-01 8.39786112e-01
2.01092027e-02 -2.06888944e-01 -1.02841854e+00 6.26382887e-01
9.17096902e-03 -1.25221443e+00 -1.35946661e-01 1.42181925e-02
1.06255150e+00 1.33883730e-01 1.83015317e-01 7.18841404e-02
4.25766230e-01 -9.76467609e-01 6.61355197e-01 3.55199516e-01
9.24403548e-01 -3.69696230e-01 1.04602623e+00 4.79306370e-01
-8.02563071e-01 2.78702438e-01 -1.57911003e-01 6.66512728e-01
4.25106227e-01 1.11568558e+00 -7.78768778e-01 5.35512388e-01
5.96581697e-01 6.74958408e-01 -3.79390895e-01 8.20072770e-01
-4.11723524e-01 8.65943849e-01 -3.68227988e-01 3.85317087e-01
2.10346490e-01 -1.53707460e-01 3.69039774e-01 1.19852698e+00
6.82987869e-02 1.93213504e-02 2.40505293e-01 1.26529884e+00
-2.50109702e-01 -2.87579358e-01 -1.63450494e-01 2.69820005e-01
2.71455258e-01 1.19887102e+00 -9.91468370e-01 -4.87948090e-01
-1.69708639e-01 1.10273910e+00 1.97215796e-01 2.09464595e-01
-1.14711332e+00 2.88010556e-02 1.80150971e-01 -1.39503643e-01
4.30037007e-02 1.66204348e-01 -6.97527766e-01 -1.15468228e+00
9.21360478e-02 -7.03048110e-01 4.33261752e-01 -4.50013131e-01
-1.25992572e+00 6.98977768e-01 -1.20170064e-01 -1.22112024e+00
-9.12854820e-02 -4.64135498e-01 -4.12486047e-01 6.27312064e-01
-1.63033640e+00 -1.18048739e+00 -1.49144471e-01 3.55462193e-01
4.52722549e-01 1.62078530e-01 5.98361373e-01 2.50710309e-01
-7.86853135e-01 8.58495593e-01 1.09765232e-01 3.39017898e-01
7.54574776e-01 -1.45047092e+00 2.53775597e-01 7.79750228e-01
-4.76362603e-03 4.27189261e-01 2.84162968e-01 -7.95013070e-01
-7.41999447e-01 -1.31051755e+00 5.95879793e-01 -2.93358266e-01
3.20071220e-01 -3.08112174e-01 -9.34168160e-01 7.54800498e-01
-1.92040682e-01 3.96278739e-01 8.03049803e-01 -3.06439757e-01
-6.57224357e-02 1.23641856e-01 -1.34437299e+00 5.14589131e-01
8.15363467e-01 -3.36016387e-01 -3.61939430e-01 5.33827305e-01
8.77492189e-01 -8.27790499e-01 -7.53190041e-01 7.07773328e-01
4.07733828e-01 -8.68635774e-01 7.07337797e-01 -1.07920147e-01
5.95585108e-01 -2.73564190e-01 3.05705871e-02 -1.18291557e+00
1.16016380e-02 -5.06604970e-01 -1.29404038e-01 1.06588769e+00
9.09691453e-01 -6.44283712e-01 9.35592055e-01 6.80367112e-01
-4.76398975e-01 -1.24900031e+00 -8.81219506e-01 -5.48493028e-01
2.07159311e-01 -4.87831503e-01 3.30727220e-01 8.55259478e-01
-1.09776892e-01 -2.35856488e-01 -3.16878915e-01 2.81027973e-01
6.12568974e-01 -5.00469431e-02 2.63076991e-01 -8.92504573e-01
-6.30212069e-01 -2.11110950e-01 -5.83974384e-02 -1.28950274e+00
2.47839570e-01 -1.03843939e+00 4.52613771e-01 -1.36630630e+00
4.93241221e-01 -7.13617384e-01 -3.67405951e-01 6.51637435e-01
-5.52551568e-01 2.97028124e-01 -8.03698376e-02 3.49130183e-01
-7.45207846e-01 3.87567580e-01 1.67451942e+00 -3.06112826e-01
-3.09746295e-01 3.15582454e-01 -7.19614327e-01 9.62357521e-01
9.09173131e-01 -6.54303312e-01 -3.20970297e-01 -4.28508103e-01
1.28941061e-02 3.74975763e-02 2.86945939e-01 -7.62961924e-01
2.11730272e-01 2.69629015e-03 3.53291720e-01 -5.55832744e-01
-1.06187202e-01 -7.05335617e-01 -1.28801644e-01 5.82719505e-01
-5.48050463e-01 -6.15545750e-01 5.13831154e-02 6.40546858e-01
-3.10411543e-01 -3.00530016e-01 1.04621482e+00 -2.56838024e-01
-3.03314567e-01 4.12082732e-01 -1.58099905e-01 1.81693852e-01
1.09092176e+00 -7.17604235e-02 -1.87359884e-01 -8.03361610e-02
-1.02764988e+00 6.53670013e-01 3.79384845e-01 1.20708108e-01
6.36114240e-01 -9.88170445e-01 -5.70497453e-01 1.63592935e-01
1.15209967e-01 7.26950943e-01 2.99569070e-01 1.14137840e+00
-4.11538571e-01 1.61527857e-01 3.43459934e-01 -9.99294877e-01
-1.04840076e+00 2.70855993e-01 5.85314691e-01 -6.24953508e-01
-6.86464310e-01 9.79666710e-01 4.22435462e-01 -7.43026435e-01
2.30459556e-01 -5.90141773e-01 8.50432962e-02 -4.54911351e-01
2.92956382e-01 1.01571985e-01 -4.31590974e-02 -4.17166680e-01
-3.71953756e-01 3.48815441e-01 -3.51211458e-01 -2.94234566e-02
9.22255337e-01 -1.48350120e-01 -5.14456742e-02 4.10939783e-01
1.17717254e+00 -2.80952483e-01 -1.58307528e+00 -4.25159693e-01
-7.24740252e-02 -1.37910321e-01 9.72250625e-02 -9.69137669e-01
-1.32701600e+00 7.01654375e-01 6.62141621e-01 -3.01810414e-01
9.99177694e-01 2.63191015e-01 9.54528272e-01 1.85784977e-03
2.75507390e-01 -1.31202245e+00 -8.37770924e-02 2.80291796e-01
3.41923624e-01 -1.73005128e+00 -1.55457690e-01 -6.88605487e-01
-9.70021486e-01 7.64243126e-01 6.46064341e-01 2.44868994e-01
7.57966995e-01 5.07232785e-01 2.66335666e-01 -1.25187069e-01
-3.70047569e-01 -1.04981162e-01 3.94991875e-01 3.09492677e-01
4.21403319e-01 3.08282673e-01 -1.71567813e-01 9.88767862e-01
8.40545297e-02 1.27361238e-01 2.89152265e-01 8.79451334e-01
-2.03231484e-01 -9.29276466e-01 -1.71718910e-01 5.92535555e-01
-6.57373726e-01 -1.70265332e-01 -2.63534188e-01 5.52555442e-01
3.37349385e-01 8.67484570e-01 -1.30816132e-01 -3.59015673e-01
-1.90318763e-01 -1.11205295e-01 1.48997411e-01 -7.53175676e-01
-3.39371413e-01 2.91667730e-01 -3.16131771e-01 -5.07709861e-01
-3.51169884e-01 -5.13214171e-01 -1.72974718e+00 3.35808456e-01
-5.71157455e-01 3.50224189e-02 4.50688124e-01 1.04288912e+00
2.54684508e-01 6.65465772e-01 6.63344800e-01 -6.01453960e-01
-6.29057109e-01 -8.53005946e-01 -5.48029006e-01 3.18415582e-01
3.17376196e-01 -4.00626063e-01 -3.58340025e-01 1.73296437e-01] | [14.6248779296875, -2.0580923557281494] |
54f3e4dc-c840-4f9b-a477-6e6c98d2a4c3 | open-domain-question-answering-goes | 2010.04898 | null | https://arxiv.org/abs/2010.04898v3 | https://arxiv.org/pdf/2010.04898v3.pdf | Open-Domain Question Answering Goes Conversational via Question Rewriting | We introduce a new dataset for Question Rewriting in Conversational Context (QReCC), which contains 14K conversations with 80K question-answer pairs. The task in QReCC is to find answers to conversational questions within a collection of 10M web pages (split into 54M passages). Answers to questions in the same conversation may be distributed across several web pages. QReCC provides annotations that allow us to train and evaluate individual subtasks of question rewriting, passage retrieval and reading comprehension required for the end-to-end conversational question answering (QA) task. We report the effectiveness of a strong baseline approach that combines the state-of-the-art model for question rewriting, and competitive models for open-domain QA. Our results set the first baseline for the QReCC dataset with F1 of 19.10, compared to the human upper bound of 75.45, indicating the difficulty of the setup and a large room for improvement. | ['Srinivas Chappidi', 'Stephen Pulman', 'Shayne Longpre', 'Zhucheng Tu', 'Svitlana Vakulenko', 'Raviteja Anantha'] | 2020-10-10 | null | https://aclanthology.org/2021.naacl-main.44 | https://aclanthology.org/2021.naacl-main.44.pdf | naacl-2021-4 | ['question-rewriting'] | ['natural-language-processing'] | [ 6.64544329e-02 5.39636672e-01 5.90451777e-01 -2.87425786e-01
-1.83560395e+00 -1.13124502e+00 7.52958655e-01 4.38242331e-02
-4.40275788e-01 6.50658429e-01 8.51879835e-01 -6.23606145e-01
2.43175253e-02 -3.21182102e-01 -6.22086048e-01 -3.83180939e-02
1.79328844e-01 8.13769877e-01 4.92619872e-01 -8.35390985e-01
1.36760116e-01 -3.19534838e-01 -1.16084588e+00 1.12537801e+00
1.04062915e+00 7.17288673e-01 3.44834149e-01 1.61134708e+00
-4.25990820e-01 1.17407894e+00 -1.05623627e+00 -1.03816950e+00
-2.49649897e-01 -5.30974627e-01 -1.84946895e+00 -2.97532797e-01
8.50641727e-01 -3.54067266e-01 -2.34817117e-01 3.96752328e-01
4.49992150e-01 3.40165973e-01 5.82877576e-01 -9.41424727e-01
-8.48392308e-01 4.20499891e-01 2.98061401e-01 2.55798131e-01
1.21633136e+00 6.66657090e-02 1.34954023e+00 -7.64349401e-01
5.82674861e-01 1.38204741e+00 3.77288669e-01 9.25218403e-01
-7.02590764e-01 -3.32155712e-02 -9.58226621e-03 3.25915396e-01
-6.73556566e-01 -6.23681068e-01 -4.25607339e-03 -1.19251922e-01
1.36089182e+00 7.13572741e-01 1.16160430e-01 1.22534740e+00
4.13653515e-02 6.67122364e-01 7.12220728e-01 -4.87722814e-01
-4.64383559e-03 -2.10746448e-03 4.59846348e-01 5.48723519e-01
-4.09210354e-01 -5.83131671e-01 -5.21931112e-01 -4.66406763e-01
-5.73980287e-02 -6.09266579e-01 -4.72808212e-01 2.82540858e-01
-9.66418386e-01 7.20609546e-01 1.42270848e-01 -1.14930458e-02
-8.29103217e-02 -8.59441794e-03 6.19399726e-01 9.64329243e-01
5.70063591e-02 1.00129819e+00 -7.01790392e-01 -6.83843017e-01
-6.18464202e-02 8.10648024e-01 1.69670188e+00 1.14821243e+00
3.08735013e-01 -9.21363354e-01 -6.13259614e-01 1.28306544e+00
-2.31695292e-03 7.47640908e-01 1.96345180e-01 -1.69058740e+00
1.21568620e+00 4.90348428e-01 4.89575833e-01 -6.11119211e-01
-2.04311416e-01 1.38980046e-01 -2.01185733e-01 -6.18859589e-01
8.19893479e-01 -4.29142267e-01 -3.02399457e-01 1.51029205e+00
2.81285107e-01 -5.30723035e-01 4.87380624e-01 6.35134816e-01
1.38020778e+00 9.14348662e-01 -1.26980515e-02 1.61763951e-01
1.80773878e+00 -1.52429152e+00 -6.50510609e-01 -2.54206687e-01
7.26834536e-01 -1.08430719e+00 1.41595447e+00 5.20412773e-02
-1.21075320e+00 -3.96314323e-01 -5.64291060e-01 -6.74998820e-01
-1.71427861e-01 -3.18516493e-01 -3.80064212e-02 2.68251836e-01
-1.28998208e+00 -2.10559350e-02 -1.28729329e-01 -6.01945758e-01
-3.64698917e-01 -1.28392547e-01 -8.27330351e-02 -3.94939035e-01
-1.41771126e+00 1.14144421e+00 -3.66566598e-01 -8.53874460e-02
-7.75771797e-01 -5.48497021e-01 -6.81675196e-01 1.27706125e-01
4.42532480e-01 -7.31664777e-01 2.20243168e+00 -5.85710645e-01
-1.69126368e+00 9.54100788e-01 -4.15371478e-01 -3.86050045e-01
2.38466650e-01 -5.72225451e-01 -3.25431734e-01 6.14614189e-01
2.46852383e-01 5.89886785e-01 3.47802192e-01 -1.02634406e+00
-8.02456439e-01 -1.98371243e-03 8.98074687e-01 5.56180537e-01
2.77822018e-01 2.63264805e-01 -6.70965850e-01 -2.64069866e-02
-2.64525235e-01 -9.10129786e-01 2.77985241e-02 -5.59031844e-01
-9.23959017e-02 -7.64260113e-01 4.08932924e-01 -1.32494497e+00
1.15223491e+00 -1.64855230e+00 1.05541116e-02 -3.63180071e-01
-1.06461765e-02 2.52153695e-01 -7.47755885e-01 1.13221097e+00
3.11370879e-01 1.22073270e-01 -1.25472963e-01 -5.07603765e-01
1.68242976e-01 2.14125499e-01 -5.40583074e-01 -3.22483838e-01
1.78282171e-01 1.32529187e+00 -9.97525811e-01 -3.38887423e-01
-4.08398598e-01 -1.45147577e-01 -3.96036863e-01 7.72179723e-01
-7.42178559e-01 2.96335638e-01 -4.17883039e-01 3.00123245e-01
2.81758636e-01 -3.82304043e-01 -1.72272441e-03 4.66987997e-01
2.26409212e-01 1.13333654e+00 -4.27584827e-01 1.83904254e+00
-7.71976233e-01 7.15365648e-01 4.43772018e-01 -2.79121131e-01
5.40206313e-01 7.21820891e-01 -1.45102248e-01 -9.85632360e-01
-2.53213376e-01 1.15411595e-01 -9.67357755e-02 -9.15424407e-01
8.13005567e-01 3.35561246e-01 -5.17243743e-01 8.12835515e-01
2.23236084e-02 -4.37375516e-01 1.91903755e-01 6.04743898e-01
1.55451751e+00 -2.40718976e-01 -1.49273425e-01 -1.31487533e-01
8.57383728e-01 2.08384067e-01 -1.57865450e-01 1.16433942e+00
-1.60700530e-01 5.90690315e-01 7.18356669e-01 -1.20783433e-01
-7.93689191e-01 -1.01695549e+00 3.09832662e-01 1.50728106e+00
-1.08332917e-01 -5.99326015e-01 -1.13621259e+00 -8.49067092e-01
-2.16838598e-01 7.35550880e-01 -2.63933688e-01 1.18935764e-01
-8.90043318e-01 2.75984425e-02 7.17545152e-01 1.09294482e-01
5.02979279e-01 -1.16440749e+00 -2.92904794e-01 2.98395306e-01
-1.20337498e+00 -1.58556473e+00 -9.46294188e-01 -4.62121695e-01
-4.47415799e-01 -1.35053015e+00 -7.05900490e-01 -1.04817426e+00
5.72391972e-02 4.86519396e-01 1.95214450e+00 4.25442994e-01
1.57258794e-01 1.13985240e+00 -8.51488709e-01 -1.52619943e-01
-9.31348622e-01 5.68478882e-01 -6.67266369e-01 -4.51613784e-01
2.86206096e-01 -3.20995063e-01 -8.20121467e-01 4.81925815e-01
-4.22969699e-01 -1.77458435e-01 1.56350985e-01 7.19488502e-01
-1.94965854e-01 -1.02477813e+00 1.07088506e+00 -6.91840351e-01
1.29053760e+00 -4.35496897e-01 -3.11310381e-01 7.91009784e-01
-6.94612116e-02 -1.03315473e-01 6.31216943e-01 -9.41671357e-02
-1.27448988e+00 -5.18123806e-01 -5.66280186e-01 5.55794954e-01
-4.97651212e-02 2.65303969e-01 2.72045564e-03 3.62199694e-01
8.02540064e-01 4.83924188e-02 5.82361482e-02 -4.54106241e-01
8.25473845e-01 1.03658605e+00 8.34292054e-01 -8.52707446e-01
3.45446289e-01 -1.05345339e-01 -7.17858672e-01 -7.81744659e-01
-1.22882879e+00 -8.78775775e-01 -1.82983667e-01 -2.87309885e-01
1.04363632e+00 -9.43728566e-01 -1.22837126e+00 2.70134062e-01
-1.74676418e+00 -6.96964800e-01 -9.54092294e-02 -1.58754185e-01
-5.57030678e-01 6.37461603e-01 -9.28078949e-01 -7.89797962e-01
-8.84566009e-01 -7.20770538e-01 1.15714371e+00 1.87765330e-01
-6.19086444e-01 -9.41781282e-01 2.64193237e-01 1.33858192e+00
4.18346018e-01 -1.97555229e-01 1.01769042e+00 -8.77013206e-01
-5.60743034e-01 -1.15245841e-01 -1.28442258e-01 4.20285910e-01
-6.48135543e-02 -3.71434510e-01 -8.79628837e-01 -1.01047069e-01
-3.59698897e-03 -8.66625726e-01 5.00269413e-01 -3.59409422e-01
7.08114803e-01 -5.84784627e-01 1.81091055e-01 -2.24594444e-01
8.63497674e-01 6.34025037e-02 8.19316328e-01 2.29454815e-01
2.99760789e-01 1.01419485e+00 4.55944300e-01 -9.42568704e-02
1.12549126e+00 6.19734347e-01 1.65237159e-01 2.85785824e-01
-1.95024103e-01 -2.27105513e-01 4.45525885e-01 1.32747078e+00
4.22589064e-01 -6.02675855e-01 -9.63634193e-01 8.77244890e-01
-1.84177852e+00 -8.83176088e-01 -5.72127283e-01 1.91137981e+00
1.12633479e+00 -2.01731786e-01 -2.09670570e-02 -5.85925996e-01
4.46177393e-01 1.08080611e-01 -2.57901341e-01 -1.01841450e+00
-1.55385258e-02 4.44697589e-01 -1.70153871e-01 1.17517102e+00
-6.52345598e-01 6.60767078e-01 6.86280060e+00 4.63219792e-01
-2.17302144e-01 2.35445172e-01 3.88971686e-01 1.97487250e-01
-4.48648423e-01 3.93162714e-03 -7.62948811e-01 2.90304840e-01
1.50888348e+00 -7.24539310e-02 7.65993059e-01 3.99661869e-01
1.27594080e-02 -4.46493030e-01 -1.17577624e+00 6.71113133e-01
4.51620132e-01 -1.04226589e+00 2.59875301e-02 -4.08826858e-01
7.85748720e-01 1.41279235e-01 -3.65308166e-01 9.12038088e-01
6.31598353e-01 -1.06470203e+00 1.17332466e-01 4.91188556e-01
3.84358823e-01 -4.95948017e-01 8.06462348e-01 5.54614842e-01
-7.07832277e-01 -8.12061727e-02 -2.48593405e-01 -2.27433920e-01
3.36967230e-01 -6.07583672e-02 -1.04593766e+00 4.96544600e-01
7.32986271e-01 -3.15747373e-02 -6.59160852e-01 5.85384667e-01
-4.01163489e-01 6.59014463e-01 -1.64916143e-01 -4.97939825e-01
3.82898301e-01 -1.48539424e-01 4.85471457e-01 1.27120650e+00
-8.91248584e-02 3.67746711e-01 -1.10787235e-01 3.60709786e-01
-5.19340932e-01 2.10601404e-01 -1.96368143e-01 2.75566816e-01
6.53768241e-01 1.17835832e+00 2.36131489e-01 -4.62211341e-01
-5.45056403e-01 1.39043140e+00 7.28649855e-01 4.86673474e-01
-6.22752607e-01 -6.11135423e-01 5.94295979e-01 -2.29044855e-01
3.53410989e-01 -2.00280458e-01 6.07300103e-02 -1.08359063e+00
6.30198956e-01 -1.68466902e+00 5.71491182e-01 -8.57142389e-01
-1.48568761e+00 6.63105071e-01 -2.05463141e-01 -5.49294114e-01
-6.05288744e-01 -5.75643778e-01 -8.20554316e-01 1.08147216e+00
-1.55755246e+00 -6.97476447e-01 -4.34284359e-01 4.69448239e-01
8.67356181e-01 2.06261501e-01 1.14025760e+00 1.68692067e-01
-4.33187187e-02 6.76237226e-01 2.90122390e-01 2.33912274e-01
1.02220702e+00 -1.49487305e+00 8.94707739e-01 4.27914470e-01
-9.30399895e-02 6.67541504e-01 5.68101168e-01 -2.36445978e-01
-1.40987515e+00 -8.44627976e-01 1.65141439e+00 -1.41722929e+00
6.54497743e-01 -6.25221014e-01 -1.12018359e+00 5.41010439e-01
8.97894859e-01 -8.12646985e-01 6.97797060e-01 3.94213885e-01
-4.88655239e-01 -1.02274895e-01 -7.52206445e-01 8.38601887e-01
8.36717308e-01 -1.09740770e+00 -1.18680406e+00 6.96854115e-01
1.35364521e+00 -6.60469770e-01 -9.97962534e-01 -8.33064169e-02
4.60939646e-01 -6.71416640e-01 7.57557690e-01 -8.33663762e-01
6.69407725e-01 1.47366524e-01 -1.94568232e-01 -1.12432301e+00
1.48832783e-01 -9.48310852e-01 -1.18191622e-01 1.30220437e+00
8.90915155e-01 -4.95675832e-01 3.91172290e-01 1.09152174e+00
-3.11163425e-01 -6.65275991e-01 -1.08302450e+00 -4.64701235e-01
4.95488763e-01 -6.80893809e-02 5.37424743e-01 3.84001583e-01
3.68900895e-01 1.12766182e+00 -5.86796552e-02 4.00655642e-02
9.14068297e-02 9.46532860e-02 9.99713659e-01 -7.39711046e-01
-3.89794528e-01 -1.48207173e-01 5.24476707e-01 -1.83237445e+00
1.45426363e-01 -6.66371107e-01 3.82875025e-01 -2.06008339e+00
9.76068825e-02 6.42469004e-02 5.14735103e-01 -2.31482275e-02
-6.27585411e-01 -3.05483282e-01 2.09605351e-01 5.25840074e-02
-1.22987223e+00 5.61253309e-01 1.44479072e+00 -8.32995996e-02
-2.69214865e-02 -6.15824619e-03 -9.42450941e-01 1.38363510e-01
6.76817596e-01 -1.46200687e-01 -5.62867999e-01 -8.78188312e-01
2.48240590e-01 6.14963055e-01 2.48056963e-01 -5.74172378e-01
4.13323998e-01 2.38277197e-01 -3.25465024e-01 -4.97377515e-01
4.43173826e-01 -2.86709547e-01 -5.61176360e-01 1.23164229e-01
-8.44468713e-01 5.22432089e-01 7.72290900e-02 4.80466872e-01
-3.57610583e-01 -5.36403954e-01 2.66217262e-01 -3.76419067e-01
-2.24812999e-01 -3.16660136e-01 -7.81553864e-01 9.99261141e-01
3.22299033e-01 4.01187360e-01 -9.42093492e-01 -1.29787481e+00
-4.17448789e-01 9.84039366e-01 -9.53750778e-03 7.86977828e-01
4.32234704e-01 -9.40255642e-01 -1.11541808e+00 -7.14546800e-01
3.11117768e-01 -1.47586139e-02 2.75613695e-01 2.31290102e-01
-5.45364261e-01 9.12666142e-01 4.98880595e-01 -3.90286297e-01
-1.41629672e+00 1.53557569e-01 5.57800829e-01 -7.66986251e-01
-3.84201765e-01 8.29529762e-01 -6.41151816e-02 -1.05152416e+00
3.65385264e-01 -3.44652534e-01 -3.24797750e-01 -1.15302697e-01
8.83615911e-01 3.71839613e-01 3.55122924e-01 -1.04853101e-01
-1.56855389e-01 3.40243787e-01 -2.72227585e-01 -5.13491094e-01
6.37075663e-01 -5.70636809e-01 -3.57379258e-01 2.17125088e-01
1.47163069e+00 -6.46296144e-02 -6.77425385e-01 -3.34648430e-01
2.78258651e-01 -2.24436074e-02 -8.84398341e-01 -1.40316737e+00
8.18614587e-02 6.41761601e-01 1.40888616e-01 3.90708059e-01
6.99314594e-01 3.55597973e-01 1.31935680e+00 1.16869581e+00
1.26049668e-01 -1.09609699e+00 4.93437350e-01 1.37171984e+00
1.53869331e+00 -1.20326138e+00 -4.96446341e-01 -2.44667172e-01
-7.91922688e-01 8.94662976e-01 7.75052428e-01 1.59973919e-01
3.85988653e-02 -3.17868233e-01 6.73622966e-01 -2.74389982e-01
-1.35880983e+00 -6.98907748e-02 3.39900315e-01 4.70147461e-01
5.29547095e-01 -1.40200481e-01 -3.21124375e-01 5.33009827e-01
-6.11203909e-01 -5.27230799e-01 4.91532266e-01 7.56299913e-01
-4.70141441e-01 -1.10544944e+00 -2.37008214e-01 1.67449042e-01
-5.32040358e-01 -3.55611920e-01 -8.15037251e-01 5.20229340e-01
-7.04825103e-01 2.01830268e+00 6.33466542e-02 -1.52123913e-01
8.59721601e-01 3.70963693e-01 2.89635241e-01 -5.61938047e-01
-1.26365805e+00 -6.63591743e-01 1.11959171e+00 -5.03477931e-01
-6.53779060e-02 -4.88401055e-01 -9.84817207e-01 -2.32713550e-01
-1.41589850e-01 8.18350196e-01 3.39528769e-01 9.08900201e-01
6.62359893e-01 2.13864699e-01 6.77537441e-01 2.48792082e-01
-8.50012302e-01 -1.27528703e+00 2.66091824e-01 6.46334291e-01
5.54282367e-01 3.48793387e-01 -3.58316004e-01 4.09835540e-02] | [11.925176620483398, 8.017422676086426] |
e305aec1-b5ae-4dad-9e0f-574d09b18d26 | molecular-dynamics | 2307.02176 | null | https://arxiv.org/abs/2307.02176v2 | https://arxiv.org/pdf/2307.02176v2.pdf | Molecular Dynamics | While many good textbooks are available on Protein Structure, Molecular Simulations, Thermodynamics and Bioinformatics methods in general, there is no good introductory level book for the field of Structural Bioinformatics. This book aims to give an introduction into Structural Bioinformatics, which is where the previous topics meet to explore three dimensional protein structures through computational analysis. We provide an overview of existing computational techniques, to validate, simulate, predict and analyse protein structures. More importantly, it will aim to provide practical knowledge about how and when to use such techniques. We will consider proteins from three major vantage points: Protein structure quantification, Protein structure prediction, and Protein simulation & dynamics. We know that many proteins have functional motions, and in Chapter "Structure Determination" we already introduced the famous example of the allosteric cooperative binding of oxygen to the haem group in hemoglobin. However, experimentally, such motions are hard to observe. Here, we will introduce MD simulations to investigate the dynamic behaviour of proteins. In a simulation the forces and interactions between particles are used to numerically derive the resulting three-dimensional movement of these particles over a certain time-scale. We will also highlight some applications, and will see how simulation results may be interpreted. | ['K. Anton Feenstra', 'Sanne Abeln', 'Jocelyne Vreede', 'Arriën Symon Rauh', 'Ali May', 'Qingzhen Hou', 'Jose Gavaldá-Garciá', 'Juami H. M. van Gils', 'Halima Mouhib'] | 2023-07-05 | null | null | null | null | ['protein-structure-prediction'] | ['miscellaneous'] | [ 2.84753233e-01 5.33944555e-02 8.34074616e-02 -1.75151438e-01
-8.81903693e-02 -3.40693176e-01 1.66503236e-01 4.30620074e-01
-3.37856889e-01 1.30621469e+00 -2.01401561e-01 -4.89661664e-01
-1.20785862e-01 -3.77715290e-01 -5.72887123e-01 -1.43847978e+00
-3.16596806e-01 6.44216716e-01 3.49335164e-01 -5.75509548e-01
2.51726538e-01 8.37522626e-01 -1.53424489e+00 -3.95822600e-02
9.04446959e-01 1.92467347e-01 5.92441320e-01 9.13364589e-01
-7.45137921e-03 6.62060022e-01 -5.70681512e-01 -3.63702416e-01
-3.23468506e-01 -8.82436335e-01 -9.66580391e-01 1.50384948e-01
-2.84375399e-01 2.85274863e-01 2.81529009e-01 6.14323735e-01
7.37165034e-01 4.57829274e-02 7.13219702e-01 -4.65229690e-01
-3.88127357e-01 1.77224707e-02 -2.56148160e-01 2.03885511e-01
6.95051968e-01 2.58277744e-01 5.52308679e-01 -5.09508550e-01
1.02842247e+00 1.07000828e+00 6.15879476e-01 7.97072709e-01
-1.37913322e+00 2.27319792e-01 -3.43636274e-01 2.83069491e-01
-1.01668942e+00 -2.88284481e-01 5.12693346e-01 -5.18374383e-01
1.44195759e+00 4.35728043e-01 9.70697224e-01 6.49565279e-01
8.24134886e-01 3.59444827e-01 1.11600041e+00 -6.35630012e-01
3.52335364e-01 -1.41749084e-01 4.29829031e-01 6.62789941e-01
2.01421142e-01 2.16178223e-01 -4.16772157e-01 -5.38202882e-01
2.42457464e-01 -8.58692154e-02 -2.53955036e-01 -8.98608088e-01
-8.83278012e-01 6.81487799e-01 -1.68529391e-01 4.41641450e-01
-5.79951465e-01 -2.44835347e-01 4.31421012e-01 9.13685129e-04
7.72415474e-02 2.96609819e-01 -6.91584349e-01 -3.49979758e-01
-4.78727013e-01 7.53513932e-01 1.03540790e+00 2.48293728e-01
5.05835891e-01 -4.15799201e-01 4.52662289e-01 6.75693691e-01
3.57907772e-01 3.11180383e-01 3.68537426e-01 -8.07823956e-01
-3.16390872e-01 2.48889834e-01 3.56293619e-01 -4.62099224e-01
-6.40141308e-01 3.78670692e-01 -6.73491538e-01 4.82672691e-01
6.30726099e-01 -2.00216230e-02 -5.54336011e-01 1.35826814e+00
5.21187603e-01 -3.61424923e-01 2.41656482e-01 8.10146749e-01
7.41219938e-01 5.05355060e-01 3.16888571e-01 -1.09753191e+00
1.44370127e+00 -5.60578525e-01 -8.19226444e-01 3.93046647e-01
8.23037982e-01 -8.82154286e-01 4.12425637e-01 3.96002263e-01
-1.46789610e+00 -2.76401162e-01 -9.31484401e-01 -1.45735547e-01
-4.21490401e-01 -2.74014950e-01 6.21819139e-01 5.57950497e-01
-7.21805215e-01 1.02713358e+00 -1.26051593e+00 -9.04208839e-01
-2.02706769e-01 4.28277075e-01 -2.77893245e-01 4.25244570e-01
-1.24837744e+00 1.49301302e+00 2.99532801e-01 -1.31540477e-01
-3.26697469e-01 -4.87493217e-01 -4.09058779e-01 -3.76258016e-01
-4.95307371e-02 -8.81514132e-01 1.10877359e+00 -3.79602492e-01
-1.50055158e+00 1.29533315e+00 -8.33719194e-01 -5.37357152e-01
1.37678608e-01 2.70813257e-01 -1.18997410e-01 1.97427824e-01
-2.92906344e-01 2.30705455e-01 -1.61690891e-01 -1.09122324e+00
-2.07206219e-01 -5.08185446e-01 -1.15989462e-01 3.25932086e-01
6.48964107e-01 3.29424411e-01 2.11715296e-01 -2.45629460e-01
1.40300110e-01 -7.47478187e-01 -6.03709757e-01 -2.87459046e-01
-2.01699078e-01 -4.09690857e-01 4.70785707e-01 -3.76385927e-01
9.07146811e-01 -1.74465561e+00 5.96010089e-01 7.56076500e-02
3.85589987e-01 3.65808010e-01 6.17204428e-01 1.07870007e+00
-4.76484358e-01 -9.75370184e-02 -4.77482170e-01 4.02077645e-01
-2.09624544e-01 3.35495144e-01 4.34056371e-02 6.65913761e-01
-2.22915813e-01 8.65783751e-01 -7.28557467e-01 -2.85431921e-01
4.89501894e-01 8.27659667e-01 -2.16351599e-01 7.99160153e-02
-1.50532395e-01 5.27563274e-01 -3.16029817e-01 4.69246268e-01
7.70650387e-01 -3.56667995e-01 7.64932096e-01 -1.85176909e-01
-4.44131881e-01 4.71845597e-01 -6.98031068e-01 1.03926766e+00
4.53427523e-01 -4.32307385e-02 3.08233872e-02 -1.14281940e+00
9.16008770e-01 4.46178079e-01 9.49677348e-01 -3.41818482e-01
-6.51929453e-02 1.38818815e-01 5.14888525e-01 -6.34416997e-01
-1.80629827e-03 -5.56833088e-01 4.65261370e-01 4.16121930e-01
-4.30853695e-01 5.58729507e-02 4.81482476e-01 1.85797140e-01
9.87272680e-01 4.45707500e-01 6.99020624e-01 -5.46697259e-01
9.72269237e-01 3.87681276e-01 3.42525184e-01 2.34323829e-01
-5.37430048e-01 1.69344842e-01 7.43416011e-01 -8.60374093e-01
-1.29423022e+00 -7.68715978e-01 -4.83041853e-01 1.10989368e+00
2.48433188e-01 -7.04429507e-01 -1.26662982e+00 2.17994139e-01
-5.89865036e-02 7.14465156e-02 -4.77977067e-01 3.62544209e-02
-7.42139518e-01 -1.62521493e+00 -2.78085507e-02 -7.73283020e-02
-5.51622845e-02 -1.43704224e+00 -1.04793298e+00 5.66719174e-01
1.25320256e-01 -5.27198851e-01 2.02187821e-01 6.36988163e-01
-9.56079304e-01 -1.52378011e+00 -5.47164679e-01 -7.15094030e-01
4.20319885e-01 1.48550496e-01 1.09977090e+00 5.24338722e-01
-6.00486279e-01 5.67473471e-03 -1.97740152e-01 -4.03399497e-01
-6.90732837e-01 -1.73189178e-01 2.83703834e-01 -6.47323608e-01
9.94184732e-01 -8.24305832e-01 -7.03585684e-01 3.50122094e-01
-7.03121960e-01 1.56271785e-01 2.69032270e-01 8.37171495e-01
8.13586056e-01 -1.72938079e-01 7.01711327e-02 -9.40957129e-01
6.99982345e-01 5.07322997e-02 -3.08560610e-01 4.04830158e-01
-5.84910572e-01 2.00065494e-01 5.01724601e-01 -1.03131384e-02
-9.38311160e-01 3.76998574e-01 -6.49605274e-01 6.07098281e-01
-3.77322912e-01 2.91387856e-01 -3.08026612e-01 -2.79756427e-01
7.72493362e-01 6.03629351e-01 6.09589934e-01 -5.99203527e-01
3.63336480e-03 2.74152905e-01 1.98941410e-01 -6.63781822e-01
3.47141147e-01 5.27940631e-01 2.96319515e-01 -1.13573527e+00
-3.08891326e-01 -4.73078251e-01 -1.02561677e+00 -6.28011003e-02
8.18312228e-01 -6.00638911e-02 -1.58148539e+00 3.99048775e-01
-9.97813940e-01 -2.91294694e-01 -1.17790501e-03 4.14755225e-01
-1.14059114e+00 1.06746233e+00 -7.08924890e-01 -8.35128427e-01
-3.24840426e-01 -1.54018390e+00 9.69888508e-01 1.13120586e-01
-2.20586434e-01 -1.25225186e+00 5.02729952e-01 3.14082086e-01
2.00619251e-01 5.36117792e-01 1.07503939e+00 -2.02072501e-01
-1.75619885e-01 3.77698272e-01 4.61323798e-01 -2.05762789e-01
7.02892616e-02 3.71873468e-01 -4.98506755e-01 -3.26785445e-01
2.38201976e-01 -1.36531502e-01 6.47370040e-01 6.80213809e-01
8.39804709e-01 -1.41439028e-02 -8.15830469e-01 4.10334408e-01
1.33910263e+00 5.70644498e-01 9.58377421e-01 4.58215475e-01
2.86846697e-01 1.01303935e+00 8.58217299e-01 2.93391079e-01
-1.93425536e-01 7.71259844e-01 2.37799704e-01 -1.06958307e-01
2.10320443e-01 3.47086757e-01 2.73601949e-01 8.18735063e-01
-9.67784524e-01 -1.63351186e-03 -8.46816242e-01 -1.38711914e-01
-1.72139788e+00 -1.23952425e+00 -6.04628623e-01 2.07624125e+00
1.23958325e+00 -2.44378932e-02 6.02212071e-01 1.12011388e-01
5.71846187e-01 -3.20049286e-01 -7.67594814e-01 -5.30111194e-01
-3.15855265e-01 3.46330702e-01 5.67574143e-01 8.20243657e-01
-7.95811236e-01 8.14079762e-01 8.20071411e+00 4.71604437e-01
-9.74214733e-01 -1.30467817e-01 2.91300535e-01 5.84630482e-02
9.42614898e-02 1.54491976e-01 -8.27080965e-01 4.60805416e-01
1.33144748e+00 -2.45210499e-01 2.61457153e-02 5.24607301e-01
6.93698287e-01 -4.60365832e-01 -7.07377791e-01 6.18673801e-01
-4.01167065e-01 -1.39658821e+00 -2.17192546e-01 4.29390788e-01
1.27116188e-01 -3.27988476e-01 -4.51006263e-01 -2.45908260e-01
1.34107009e-01 -1.11554182e+00 -8.41795132e-02 6.92701936e-01
1.63259640e-01 -7.65605867e-01 7.36148894e-01 6.11457348e-01
-9.92704809e-01 6.01856411e-01 -9.71432686e-01 -3.17710608e-01
5.67986012e-01 5.84288776e-01 -4.99635488e-01 3.84785265e-01
4.88751143e-01 5.65376341e-01 -2.02864736e-01 9.25692022e-01
1.19260706e-01 2.17255518e-01 -2.15962112e-01 -5.24896979e-01
-8.35742280e-02 -8.10801923e-01 2.59683847e-01 9.60018694e-01
-3.27397138e-01 6.60853207e-01 -1.68303892e-01 5.38462698e-01
6.34709656e-01 2.93842405e-01 -2.75976360e-01 8.21043029e-02
-1.38083547e-01 1.05923140e+00 -7.06371427e-01 -3.15283835e-01
-3.88508946e-01 7.99707353e-01 2.39835367e-01 2.62399495e-01
-7.64357746e-01 -3.55866820e-01 1.08075333e+00 4.37669694e-01
2.14340940e-01 -2.44163036e-01 3.47048968e-01 -8.33503127e-01
-2.97834933e-01 -7.65208662e-01 -3.42917554e-02 -6.78080261e-01
-1.07627654e+00 7.54808858e-02 -1.09348714e-01 -2.93863326e-01
-6.78946450e-02 -1.02068055e+00 -1.69326201e-01 1.02987695e+00
-1.02164555e+00 -5.27065814e-01 2.37418354e-01 1.57631338e-01
1.85341656e-01 1.69562578e-01 1.00862920e+00 -8.25775638e-02
-5.35294592e-01 2.10052118e-01 8.36358070e-01 -3.72182190e-01
6.18540645e-01 -1.43204427e+00 3.69751364e-01 2.04883292e-01
-7.08611548e-01 1.05091715e+00 1.41635668e+00 -8.38767409e-01
-1.53025651e+00 -4.58480328e-01 9.69214499e-01 -5.77231050e-01
4.22024369e-01 -2.78620362e-01 -1.21300530e+00 3.71414304e-01
2.70162612e-01 -4.18151915e-01 1.03890944e+00 -1.86350942e-01
3.88600558e-01 4.70046699e-01 -1.21828592e+00 3.48693341e-01
1.07387280e+00 -1.21890143e-01 -7.25298941e-01 8.77141356e-01
1.86731458e-01 -5.31800330e-01 -1.38271260e+00 3.07131857e-01
8.60099614e-01 -1.47326219e+00 1.14873767e+00 -9.92148638e-01
-3.22513171e-02 -3.45874786e-01 2.67028004e-01 -6.57818317e-01
-5.22841334e-01 -8.27984452e-01 -2.61083066e-01 4.17447954e-01
2.05442831e-01 -7.10986197e-01 1.04808331e+00 6.00913167e-01
4.02035117e-02 -1.19285619e+00 -7.37129152e-01 -6.21770978e-01
4.10519630e-01 4.39653218e-01 2.71666348e-01 7.11937845e-01
6.27016604e-01 3.99087369e-01 -2.43065581e-01 -3.89211118e-01
7.04018414e-01 1.27672881e-01 6.46491826e-01 -1.23612559e+00
-2.55113959e-01 -1.99528992e-01 -6.03534997e-01 -1.07938647e+00
-6.07998595e-02 -5.12301207e-01 -2.55182415e-01 -1.52870297e+00
5.63217103e-01 1.51193634e-01 1.88171402e-01 -4.39359946e-03
-1.05008483e-01 -6.34848401e-02 -3.52590770e-01 3.94057900e-01
-4.16405499e-01 2.18919516e-01 1.33900654e+00 3.85385066e-01
-1.49053082e-01 -1.00600861e-01 -5.24938583e-01 5.17436385e-01
9.17994380e-01 -3.66211087e-01 -1.46247417e-01 6.24849617e-01
2.52570599e-01 -3.02017890e-02 -3.68957557e-02 -5.63956201e-01
-1.02640510e-01 -4.16822791e-01 2.44263679e-01 -9.34680521e-01
3.66387457e-01 -6.13577068e-01 5.69781125e-01 1.10192335e+00
-8.21205508e-03 4.27676924e-02 -1.67743236e-01 3.97255808e-01
7.62390271e-02 -4.71152723e-01 1.18406677e+00 -4.77641135e-01
-2.86683142e-01 4.67156991e-02 -1.06134272e+00 -3.50130528e-01
1.37784314e+00 -6.09382987e-01 -3.27927828e-01 6.03960408e-03
-1.36635482e+00 4.35731038e-02 1.06850159e+00 -4.92584765e-01
3.22888315e-01 -8.04073930e-01 -2.46316791e-01 -4.25873809e-02
-8.89821276e-02 -3.56467932e-01 3.56787443e-01 1.13332736e+00
-1.38526058e+00 8.98817003e-01 -3.73897552e-01 -6.71488166e-01
-1.82655454e+00 1.05064881e+00 8.82258236e-01 -1.76658124e-01
-4.28447157e-01 4.42254275e-01 2.63404161e-01 -3.33449632e-01
-1.92775577e-01 9.60985757e-03 -3.61242801e-01 -4.60145772e-01
5.98932981e-01 3.66283238e-01 1.15930013e-01 -9.15919542e-01
-5.11698723e-01 8.99198174e-01 -1.55681744e-01 4.33786631e-01
1.36075187e+00 -3.59418988e-01 -6.31531656e-01 2.99946338e-01
8.23055744e-01 -2.31284171e-01 -9.93195474e-01 9.30448249e-02
1.41560078e-01 9.45815910e-03 -6.93220198e-01 -5.49771845e-01
-3.74481708e-01 8.00146401e-01 6.27049208e-01 4.93830264e-01
8.87203991e-01 2.36960456e-01 7.71047592e-01 5.52838266e-01
2.47326732e-01 -9.35057819e-01 -4.40910518e-01 4.72278297e-01
5.37106752e-01 -8.12105596e-01 3.27014178e-01 -7.28851020e-01
-2.86617696e-01 1.26018763e+00 2.96358854e-01 3.63261439e-02
6.27972901e-01 4.10691857e-01 6.59652799e-03 -5.30233681e-01
-1.04497361e+00 -2.11982086e-01 -2.83437461e-01 8.64000976e-01
1.17438650e+00 -4.11016382e-02 -1.14526653e+00 -2.63373628e-02
-3.17552656e-01 -2.76881635e-01 5.96192718e-01 1.29419410e+00
-9.57370937e-01 -1.78965247e+00 -5.46678185e-01 6.84897527e-02
-7.14574337e-01 2.25793831e-02 -9.43198204e-01 7.41908133e-01
-8.68127793e-02 6.01228833e-01 -4.06186134e-01 2.00438365e-01
2.73422629e-01 5.11819780e-01 9.73274410e-01 -2.53243655e-01
-4.47162509e-01 1.32171795e-01 4.26208675e-02 -3.87682974e-01
-1.20981860e+00 -7.42342889e-01 -1.61540973e+00 -9.75116134e-01
-2.85809547e-01 9.32861567e-01 5.82248867e-01 8.63915205e-01
2.48850018e-01 3.25552106e-01 3.31047699e-02 -7.35499263e-01
-1.71348989e-01 -8.05816591e-01 -9.59325612e-01 4.91227090e-01
2.51044750e-01 -8.02536011e-01 -2.21473247e-01 3.61678988e-01] | [4.735240459442139, 5.274872779846191] |
478be648-6043-4d4c-9c1f-11bb13c99cda | a-new-framework-for-experimental-design-using | 2105.05539 | null | https://arxiv.org/abs/2105.05539v1 | https://arxiv.org/pdf/2105.05539v1.pdf | A new framework for experimental design using Bayesian Evidential Learning: the case of wellhead protection area | In this contribution, we predict the wellhead protection area (WHPA, target), the shape and extent of which is influenced by the distribution of hydraulic conductivity (K), from a small number of tracing experiments (predictor). Our first objective is to make stochastic predictions of the WHPA within the Bayesian Evidential Learning (BEL) framework, which aims to find a direct relationship between predictor and target using machine learning. This relationship is learned from a small set of training models (400) sampled from the prior distribution of K. The associated 400 pairs of simulated predictors and targets are obtained through forward modelling. Newly collected field data can then be directly used to predict the approximate posterior distribution of the corresponding WHPA. The uncertainty range of the posterior WHPA distribution is affected by the number and position of data sources (injection wells). Our second objective is to extend BEL to identify the optimal design of data source locations that minimizes the posterior uncertainty of the WHPA. This can be done explicitly, without averaging or approximating because once trained, the BEL model allows the computation of the posterior uncertainty corresponding to any new input data. We use the Modified Hausdorff Distance and the Structural Similarity index metrics to estimate the posterior uncertainty range of the WHPA. Increasing the number of injection wells effectively reduces the derived posterior WHPA uncertainty. Our approach can also estimate which injection wells are more informative than others, as validated through a k-fold cross-validation procedure. Overall, the application of BEL to experimental design makes it possible to identify the data sources maximizing the information content of any measurement data. | ['Thomas Hermans', 'Eric Laloy', 'Robin Thibaut'] | 2021-05-12 | a-new-framework-for-experimental-design-using-1 | https://www.sciencedirect.com/science/article/pii/S0022169421009537 | https://www.sciencedirect.com/science/article/pii/S0022169421009537 | null | ['multi-target-regression'] | ['miscellaneous'] | [ 4.80686612e-02 3.71638387e-02 6.47987872e-02 5.16890921e-02
-8.67974997e-01 -5.60776591e-01 7.02737808e-01 4.64043945e-01
-3.68793368e-01 9.16698277e-01 7.55151808e-02 -4.39088494e-01
-8.89919698e-01 -1.02380526e+00 -8.04267883e-01 -9.62636590e-01
-2.49494433e-01 4.53365088e-01 4.45046067e-01 2.42112726e-01
5.61101317e-01 1.01604855e+00 -1.22219861e+00 -2.47346312e-02
8.26582968e-01 1.05850828e+00 5.12896597e-01 5.20612478e-01
3.29485647e-02 4.59250838e-01 -5.90545297e-01 5.57156578e-02
9.82698947e-02 -2.71357656e-01 -7.32712507e-01 -2.55690455e-01
-2.51631528e-01 -3.49382371e-01 3.22384596e-01 7.00813711e-01
2.69404501e-01 1.61871761e-01 1.28038180e+00 -7.47120321e-01
5.03453128e-02 6.22950971e-01 -4.34548736e-01 -7.19257956e-03
-8.91747251e-02 8.47277045e-02 6.97757721e-01 -7.74554014e-01
2.35038370e-01 1.02736211e+00 8.65186930e-01 -2.46129557e-01
-1.57641721e+00 -4.63740677e-01 -2.79140681e-01 1.18672088e-01
-1.63250411e+00 -1.73440259e-02 6.76744819e-01 -8.26314807e-01
4.47554469e-01 7.69803748e-02 7.49584317e-01 4.66365337e-01
3.99782568e-01 1.62472308e-01 1.21852565e+00 -6.60133839e-01
8.10870945e-01 2.42780209e-01 -2.18551099e-01 1.43584907e-01
4.04350787e-01 3.95096898e-01 -2.62803316e-01 -3.32449317e-01
6.08833253e-01 -2.05456704e-01 -3.54572177e-01 -2.34471202e-01
-5.79416573e-01 8.27887416e-01 3.19644421e-01 1.77681431e-01
-3.78233790e-01 8.35582018e-02 -6.16288278e-03 -1.04842514e-01
2.54296303e-01 7.05324292e-01 -6.38299346e-01 8.85648429e-02
-1.08379126e+00 3.75529617e-01 9.20911729e-01 3.69943857e-01
1.05282223e+00 -8.61975849e-02 1.44272955e-04 7.75981724e-01
6.65692031e-01 9.77675021e-01 -1.03853866e-01 -1.17633033e+00
3.60137641e-01 3.52620572e-01 4.98738974e-01 -7.19755828e-01
-2.49899209e-01 -2.60107100e-01 -3.63488793e-01 3.14479351e-01
8.05173635e-01 -5.12527049e-01 -5.01880825e-01 1.25792825e+00
2.72405088e-01 4.13576663e-02 2.07906738e-02 3.91974211e-01
-1.72937706e-01 9.90882576e-01 1.52533159e-01 -3.19921225e-01
6.93075776e-01 1.31149171e-02 -2.27798805e-01 -2.08833084e-01
5.46395183e-01 -3.82112414e-01 7.62142897e-01 2.11313188e-01
-5.57755113e-01 -2.27061033e-01 -9.11160529e-01 7.65583992e-01
-3.96860540e-01 2.16517359e-01 2.45489940e-01 3.99747789e-01
-4.30511177e-01 1.16372406e+00 -8.75058532e-01 -1.98363680e-02
3.36326867e-01 4.94926758e-02 -2.47088410e-02 2.26736322e-01
-1.05809271e+00 1.27768123e+00 5.15302479e-01 5.78254461e-01
-9.21932876e-01 -1.07646239e+00 -7.06928074e-01 8.97636935e-02
1.78446129e-01 7.74608850e-02 8.13089192e-01 -2.23148406e-01
-1.47889400e+00 6.89358115e-02 5.29737920e-02 -3.89936656e-01
5.56919694e-01 -1.15608573e-01 -3.46133835e-03 1.26329631e-01
9.97416750e-02 1.95068762e-01 5.29362559e-01 -1.61443412e+00
-4.44720358e-01 -2.68115312e-01 -6.05635822e-01 -1.53824359e-01
-6.52564690e-02 -3.53584468e-01 1.66694313e-01 -1.16079360e-01
1.54540509e-01 -7.68954933e-01 -3.57213080e-01 -6.16928376e-02
-3.14268678e-01 5.11991233e-02 4.76902664e-01 -8.39223206e-01
1.11951268e+00 -2.06975079e+00 -7.61670480e-03 7.67641962e-01
-2.06912592e-01 5.20067662e-02 3.04388732e-01 8.09386611e-01
3.08001667e-01 8.90096277e-02 -8.05739224e-01 1.52227402e-01
-9.05734673e-02 1.96921185e-01 -4.13745135e-01 6.16868496e-01
2.76790738e-01 3.87741268e-01 -6.62690163e-01 -2.98616439e-01
5.53982377e-01 3.07609677e-01 -1.79536462e-01 3.12957227e-01
-8.13778192e-02 3.78023148e-01 -4.58251089e-01 1.05210893e-01
7.78121471e-01 1.34774700e-01 3.43092941e-02 -1.45275787e-01
-4.23497736e-01 -1.33731008e-01 -1.27836084e+00 9.15275335e-01
-6.94781303e-01 5.31475365e-01 -1.82664558e-01 -8.95588875e-01
1.48747742e+00 2.74763387e-02 4.42593932e-01 -2.10981086e-01
1.34314923e-02 3.74236912e-01 -3.02659925e-02 -4.57998425e-01
1.41785800e-01 -4.34816897e-01 5.04436865e-02 1.92473248e-01
-5.30404598e-02 -4.76445824e-01 -1.31139174e-01 -2.23765284e-01
7.26207674e-01 1.35439426e-01 2.57405460e-01 -7.16821492e-01
3.04823756e-01 -1.09395720e-01 3.67351323e-01 7.64188290e-01
3.07680935e-01 4.16502178e-01 8.84002030e-01 -1.59406781e-01
-1.17751777e+00 -1.37363315e+00 -7.43387878e-01 2.16396824e-01
2.50537336e-01 2.22445354e-01 -4.99105513e-01 -2.41121262e-01
4.13185090e-01 1.27467811e+00 -6.52896523e-01 -2.74409771e-01
-3.32759291e-01 -9.48104978e-01 4.18959379e-01 6.33340240e-01
3.58406484e-01 -7.21058667e-01 -7.76668966e-01 3.75357062e-01
1.80831373e-01 -6.47741735e-01 3.67976636e-01 6.03150070e-01
-9.21349347e-01 -1.18387461e+00 -5.50124049e-01 3.25090624e-02
3.96270543e-01 -5.83061457e-01 5.07466674e-01 -3.66281718e-01
7.34454766e-02 1.57315850e-01 -4.10047844e-02 -6.05622888e-01
-6.76851511e-01 -1.79992333e-01 -3.46852869e-01 -1.54632062e-01
1.36624202e-01 -4.16118503e-01 -4.02736664e-01 5.71655452e-01
-7.04132557e-01 -3.51790428e-01 3.97370577e-01 5.29145658e-01
6.72129750e-01 4.12798077e-01 7.91454315e-01 -4.81989086e-01
4.46437806e-01 -5.69661438e-01 -1.00389028e+00 5.31064689e-01
-7.44094431e-01 4.03558195e-01 6.66892469e-01 -3.94810230e-01
-1.24043322e+00 -1.13517409e-02 -1.28306434e-01 -1.27741382e-01
-3.57360572e-01 8.88827085e-01 -1.56912863e-01 1.00758374e-01
7.50228524e-01 -1.53288230e-01 -7.41831660e-02 -5.89487910e-01
1.00817820e-02 7.80377626e-01 3.25243413e-01 -9.14195716e-01
4.86071289e-01 2.92814434e-01 5.70540249e-01 -1.09248400e+00
-6.44737959e-01 -2.10781068e-01 -9.52966511e-01 -5.21845579e-01
5.11947036e-01 -4.08738375e-01 -6.89838409e-01 4.35272396e-01
-8.51704419e-01 -6.84251130e-01 -5.27338147e-01 7.57939875e-01
-5.12690663e-01 1.57487214e-01 -6.31593987e-02 -1.31160390e+00
-1.71369195e-01 -9.83239710e-01 8.46856236e-01 5.43646216e-02
-2.72613227e-01 -1.27660382e+00 1.99483156e-01 -1.54010624e-01
2.18779773e-01 4.89191473e-01 9.59363461e-01 -4.69264477e-01
-2.12846726e-01 -4.98134375e-01 5.51912077e-02 5.34984231e-01
-3.77164059e-03 3.70253325e-01 -9.24333155e-01 -8.80341530e-02
9.58336517e-02 1.67228222e-01 8.10282528e-01 9.13159013e-01
9.77670193e-01 -2.09223792e-01 -4.52093750e-01 3.34843159e-01
1.75481927e+00 2.87448794e-01 7.53222823e-01 4.38286126e-01
2.13806659e-01 8.43209147e-01 5.27507961e-01 6.52112484e-01
-1.09731942e-01 5.38819611e-01 2.57413596e-01 3.76466781e-01
2.59419501e-01 -4.72814620e-01 2.08738789e-01 -3.72495688e-02
-2.46534050e-02 -7.01045319e-02 -1.36459959e+00 6.25810504e-01
-1.56799197e+00 -7.78929889e-01 -1.70613647e-01 2.69216919e+00
5.72516501e-01 3.59839052e-02 -3.43829319e-02 3.05063784e-01
6.61493123e-01 -2.02231109e-01 -3.53712201e-01 -4.09505665e-01
1.68992445e-01 7.35925436e-02 9.50583637e-01 7.54660547e-01
-5.80141544e-01 2.00060114e-01 6.23001385e+00 8.83325934e-01
-9.92713332e-01 -4.54556018e-01 6.48458362e-01 2.49631435e-01
-3.24485958e-01 3.09988230e-01 -1.10506499e+00 7.50245154e-01
1.33331299e+00 -2.59171516e-01 3.30784798e-01 5.77189684e-01
5.39549828e-01 -8.51907730e-01 -9.54657137e-01 1.96732238e-01
-4.77551728e-01 -1.30317092e+00 -2.11507082e-01 2.44206518e-01
5.41865468e-01 -1.82181299e-01 -2.71334589e-01 -1.25871480e-01
3.60831112e-01 -8.36006999e-01 8.74334574e-01 1.13228202e+00
6.00509286e-01 -7.60915637e-01 1.04028702e+00 5.67375064e-01
-1.07374573e+00 -4.37611282e-01 -3.98185432e-01 -4.00530808e-02
3.13066274e-01 8.37155402e-01 -1.03143120e+00 2.86778241e-01
6.40309453e-01 4.21274364e-01 -4.54488665e-01 1.24576831e+00
-3.21200132e-01 9.93085921e-01 -9.62327063e-01 -2.70997256e-01
4.73249704e-02 -5.09690583e-01 3.24567139e-01 8.27072382e-01
6.55023038e-01 -4.69745249e-02 -2.93073654e-01 1.40923882e+00
5.23030877e-01 4.02315557e-02 -5.02898812e-01 2.24406347e-01
1.09880126e+00 6.78980172e-01 -7.00023115e-01 2.56257296e-01
1.02686875e-01 1.37519687e-01 1.18202619e-01 4.75784123e-01
-4.58415478e-01 -3.20183337e-01 7.78898001e-02 3.58361095e-01
3.25659007e-01 1.50729939e-02 -5.21048605e-01 -2.76809841e-01
-7.87196606e-02 -1.28833547e-01 2.68541783e-01 -7.83120751e-01
-1.23852921e+00 -6.47510886e-02 7.89250016e-01 -1.02256322e+00
-4.01875824e-01 -6.49890244e-01 -8.20657730e-01 1.30180943e+00
-1.14297509e+00 -6.79212332e-01 -3.43051255e-02 -1.40114114e-01
-2.23442614e-02 1.78701118e-01 6.43072605e-01 -2.67770559e-01
-4.41327989e-01 1.06916107e-01 8.59148562e-01 -8.42846557e-02
1.30234867e-01 -1.18906522e+00 -3.07103306e-01 5.81504762e-01
-3.83053064e-01 1.47633672e-01 9.36858654e-01 -8.54137123e-01
-6.60755336e-01 -1.06877136e+00 4.84817326e-01 -3.12550396e-01
9.04138803e-01 1.24198638e-01 -1.06392860e+00 2.95347869e-01
-5.50753176e-01 -6.38959855e-02 3.92745972e-01 -1.69427902e-01
3.14323843e-01 -5.32424390e-01 -1.32924306e+00 3.19562614e-01
1.00184001e-01 -4.13309664e-01 -3.80075157e-01 -3.47189307e-02
8.31045061e-02 6.42927885e-02 -1.40408874e+00 6.40339732e-01
6.64269507e-01 -7.79861867e-01 6.76527739e-01 -1.78061470e-01
4.98230845e-01 -2.82884419e-01 -2.26287842e-01 -1.42198360e+00
-6.80545345e-02 -1.26806451e-02 2.33630359e-01 1.41298211e+00
7.54077494e-01 -5.68039060e-01 6.00674510e-01 8.88154328e-01
7.72158653e-02 -9.95792449e-01 -1.13117087e+00 -8.20139110e-01
6.59438014e-01 -5.37466645e-01 6.02173686e-01 2.42108971e-01
-1.10478744e-01 -1.93514407e-01 -1.10878162e-01 6.42819405e-01
9.12214398e-01 -2.25961432e-02 4.94177341e-01 -1.38741052e+00
-3.14765632e-01 -1.38816342e-01 -2.39128709e-01 -5.22432923e-01
-3.27080651e-03 -4.44373488e-01 3.15064728e-01 -1.49308217e+00
-1.56856909e-01 -7.11039722e-01 1.74569443e-01 2.68567890e-01
3.11721805e-02 -4.63747054e-01 -5.93022779e-02 2.32785195e-01
3.94415379e-01 6.75671160e-01 9.32331920e-01 8.11090469e-02
-5.18070579e-01 2.53076404e-01 1.79315591e-03 7.88421988e-01
8.77333999e-01 -6.01058960e-01 -2.05509320e-01 -9.82170403e-02
1.59801468e-01 4.56634372e-01 4.72769469e-01 -1.06931412e+00
-2.99027041e-02 -4.67510730e-01 6.24996901e-01 -6.87830448e-01
2.44049798e-03 -1.07735825e+00 6.33106709e-01 4.16487157e-01
-4.91463155e-01 -5.53676128e-01 3.96100283e-01 6.19828701e-01
9.51856896e-02 -9.41317677e-01 8.87246132e-01 1.16488077e-01
-3.00322324e-01 -1.91718385e-01 -6.12436295e-01 -2.77200133e-01
1.00797868e+00 -1.94061190e-01 -5.39324060e-02 -9.89596844e-02
-6.53713286e-01 1.33176357e-01 2.39937618e-01 -3.04312110e-01
4.04062301e-01 -8.12167287e-01 -6.35738611e-01 5.49483448e-02
-1.59673557e-01 1.49062395e-01 8.77887607e-02 6.38071835e-01
-6.82192564e-01 4.22415137e-02 3.45299430e-02 -5.89337587e-01
-5.38501918e-01 2.44029775e-01 9.50449288e-01 -1.96957588e-01
-4.10103291e-01 4.35869575e-01 -4.38762873e-01 -3.17130744e-01
-1.50789738e-01 -2.42748782e-01 -2.09733218e-01 6.63758889e-02
3.52395564e-01 9.06076908e-01 -6.50779158e-02 -4.07568187e-01
-1.20518200e-01 5.57485819e-01 5.17086089e-01 -1.92979082e-01
1.63253713e+00 -1.14709549e-01 -1.18605532e-01 7.29029357e-01
1.08480608e+00 -3.46797034e-02 -1.79311395e+00 1.88117668e-01
2.64629424e-01 -3.94698232e-01 2.56581545e-01 -8.51363242e-01
-7.25796759e-01 8.38552177e-01 6.73256397e-01 1.08639196e-01
7.19466329e-01 1.46616399e-01 -1.33661656e-02 2.75132090e-01
2.64946282e-01 -1.11512291e+00 -4.47739184e-01 7.56824389e-02
9.42222834e-01 -8.44650865e-01 1.42523378e-01 -2.46052593e-01
-4.14251179e-01 1.19889522e+00 2.60548353e-01 5.45008704e-02
1.05384922e+00 5.81066847e-01 -1.94173813e-01 -1.60362631e-01
-2.40661472e-01 2.58416653e-01 1.49408162e-01 5.81798851e-01
-1.71861514e-01 1.59324810e-01 -2.33010665e-01 5.45891643e-01
-1.96770266e-01 -3.40894461e-02 5.33900261e-01 5.36956787e-01
-7.09145367e-01 -7.10196614e-01 -6.14014566e-01 7.12261975e-01
-7.84332082e-02 1.65840000e-01 -3.17762569e-02 5.91312468e-01
1.08026832e-01 8.69255841e-01 1.71009153e-01 -1.32353261e-01
3.31206232e-01 -8.18678923e-03 8.82746652e-02 -2.21468419e-01
1.18614994e-01 -1.03849426e-01 -1.36135416e-02 -2.11979318e-02
-1.30792364e-01 -8.30875814e-01 -1.19797182e+00 -1.69755399e-01
-7.12087512e-01 6.90219581e-01 7.68683255e-01 1.17869997e+00
-1.58252537e-01 4.52871695e-02 6.96253061e-01 -8.63441169e-01
-8.76577795e-01 -9.85565305e-01 -1.19617307e+00 -2.85482127e-02
1.00887597e-01 -7.91485250e-01 -9.20418799e-01 -2.36306265e-01] | [6.464364051818848, 3.3578972816467285] |
8cd48ccb-a025-4780-8b7a-0e760a4eb341 | self-supervised-video-representation-learning-5 | 2008.13426 | null | https://arxiv.org/abs/2008.13426v2 | https://arxiv.org/pdf/2008.13426v2.pdf | Self-supervised Video Representation Learning by Uncovering Spatio-temporal Statistics | This paper proposes a novel pretext task to address the self-supervised video representation learning problem. Specifically, given an unlabeled video clip, we compute a series of spatio-temporal statistical summaries, such as the spatial location and dominant direction of the largest motion, the spatial location and dominant color of the largest color diversity along the temporal axis, etc. Then a neural network is built and trained to yield the statistical summaries given the video frames as inputs. In order to alleviate the learning difficulty, we employ several spatial partitioning patterns to encode rough spatial locations instead of exact spatial Cartesian coordinates. Our approach is inspired by the observation that human visual system is sensitive to rapidly changing contents in the visual field, and only needs impressions about rough spatial locations to understand the visual contents. To validate the effectiveness of the proposed approach, we conduct extensive experiments with four 3D backbone networks, i.e., C3D, 3D-ResNet, R(2+1)D and S3D-G. The results show that our approach outperforms the existing approaches across these backbone networks on four downstream video analysis tasks including action recognition, video retrieval, dynamic scene recognition, and action similarity labeling. The source code is publicly available at: https://github.com/laura-wang/video_repres_sts. | ['Yun-hui Liu', 'Jianbo Jiao', 'Wei Liu', 'Linchao Bao', 'Jiangliu Wang', 'Shengfeng He'] | 2020-08-31 | null | null | null | null | ['scene-recognition'] | ['computer-vision'] | [-1.07695095e-01 -6.18366182e-01 -4.97776330e-01 -2.07603425e-01
-3.05719048e-01 -4.93137598e-01 2.35142872e-01 -1.08230844e-01
-2.70216942e-01 4.63195860e-01 4.91696775e-01 -1.23676755e-01
-1.08407997e-01 -4.85700876e-01 -6.02742255e-01 -6.65832400e-01
-3.60045224e-01 -2.30142310e-01 3.97961408e-01 2.01177910e-01
4.41328913e-01 5.40009499e-01 -1.53029048e+00 3.47126812e-01
4.69380826e-01 1.29398882e+00 3.31608206e-01 6.07943833e-01
2.30871886e-02 1.23557663e+00 -4.49294746e-01 4.24853295e-01
3.25889409e-01 -4.88479942e-01 -6.61044359e-01 4.32430267e-01
3.84669065e-01 -8.71186912e-01 -1.06319511e+00 1.00249839e+00
3.94764155e-01 5.93973994e-01 6.58794820e-01 -1.32006454e+00
-7.74747908e-01 1.60419792e-01 -9.37067986e-01 8.37132573e-01
5.24804294e-01 1.61514685e-01 6.93648636e-01 -7.79700518e-01
5.42143643e-01 1.13513362e+00 3.33327055e-01 3.87579024e-01
-6.11500204e-01 -5.75363100e-01 5.17923176e-01 5.77959001e-01
-1.63002825e+00 -3.44969749e-01 9.10269678e-01 -3.80939484e-01
7.01891661e-01 1.61692556e-02 5.57559013e-01 9.42974627e-01
5.49856685e-02 1.04394984e+00 8.87151897e-01 -6.29980266e-02
2.34372124e-01 -2.86377072e-01 1.31267644e-02 7.40236878e-01
-2.09221058e-02 -1.70810968e-01 -5.38871467e-01 7.58994371e-02
1.13711703e+00 5.25677860e-01 -3.95403564e-01 -3.32661062e-01
-1.26851737e+00 3.71647388e-01 4.57599342e-01 2.78552264e-01
-3.66667509e-01 1.87118426e-01 5.68012059e-01 1.19159527e-01
4.59388733e-01 -1.65352747e-01 -2.47777000e-01 -1.87693059e-01
-8.72261643e-01 1.23810694e-01 2.60239661e-01 1.06895816e+00
5.82792461e-01 1.58821642e-01 -1.84461802e-01 8.27446401e-01
2.71782696e-01 3.92310083e-01 6.63364947e-01 -1.29075444e+00
6.29644990e-01 4.80874538e-01 5.48140891e-02 -1.44804430e+00
-2.31004402e-01 -4.22972776e-02 -1.07339847e+00 -7.18144104e-02
2.68714041e-01 -1.38793215e-01 -9.48660314e-01 1.62571335e+00
2.26379916e-01 7.06370473e-01 2.86406409e-02 1.21964920e+00
1.03686905e+00 1.03076816e+00 9.06754583e-02 -3.17618310e-01
1.02604437e+00 -8.41613114e-01 -5.05249858e-01 9.34253111e-02
5.55127621e-01 -4.58133578e-01 7.89799213e-01 2.37793386e-01
-1.06278670e+00 -8.80741060e-01 -8.72238696e-01 8.07022825e-02
-2.52251685e-01 4.01128203e-01 4.50354815e-01 5.12994565e-02
-1.15356147e+00 5.19504905e-01 -7.63276577e-01 -5.09249628e-01
5.49477279e-01 -4.07817699e-02 -4.24609691e-01 -3.19047838e-01
-1.07559383e+00 3.48250031e-01 6.29875302e-01 1.33712366e-01
-9.47148800e-01 -2.97195047e-01 -7.76699483e-01 -9.09758434e-02
4.58340853e-01 -3.39408636e-01 9.15716052e-01 -9.94154394e-01
-1.19118369e+00 6.83466315e-01 -1.65555865e-01 -2.04413384e-01
2.21404210e-01 -1.60658225e-01 -3.48551482e-01 4.90071356e-01
1.83371589e-01 7.14029491e-01 7.75833905e-01 -9.91366267e-01
-7.48366535e-01 -3.57280493e-01 2.12999225e-01 5.14984667e-01
-3.39319944e-01 9.39089134e-02 -8.67964029e-01 -8.74830842e-01
2.93383002e-01 -7.86042094e-01 -1.00579575e-01 6.20267652e-02
-4.22798663e-01 -2.65972137e-01 9.95179594e-01 -8.01576674e-01
1.25473154e+00 -2.47803235e+00 1.76209703e-01 2.45024666e-01
1.39450327e-01 2.67418057e-01 -1.40858561e-01 2.33372211e-01
-1.95564985e-01 2.39637569e-02 1.27790570e-01 1.27425432e-01
-3.43983650e-01 2.84355991e-02 -3.17661554e-01 5.20890653e-01
-1.86944142e-01 6.27125442e-01 -9.79200125e-01 -6.39509916e-01
3.94728363e-01 2.49044329e-01 -3.83669943e-01 2.96698600e-01
-8.15762803e-02 3.11045140e-01 -7.15071380e-01 6.76019669e-01
6.14689708e-01 -4.92526889e-01 1.16751432e-01 -4.20516908e-01
-1.62340477e-01 -1.69155262e-02 -1.36691904e+00 1.81875336e+00
-7.88100809e-02 7.22757161e-01 -4.64627862e-01 -1.21789074e+00
8.30530286e-01 2.39312977e-01 8.22556555e-01 -8.30066562e-01
1.63603500e-01 -1.49157822e-01 -4.20884192e-01 -8.02980721e-01
4.42553848e-01 4.99765128e-01 8.40884224e-02 5.21730781e-01
-1.46229610e-01 5.41931152e-01 3.26742262e-01 3.61491680e-01
1.05258977e+00 4.31216329e-01 3.49832743e-01 5.82418554e-02
5.39845586e-01 -8.99921581e-02 8.04087102e-01 5.41544259e-01
-3.85167420e-01 8.21308613e-01 5.80602467e-01 -6.24380827e-01
-8.99460852e-01 -1.04907155e+00 4.10394430e-01 1.17714095e+00
6.22643530e-01 -5.02518713e-01 -5.96069038e-01 -6.07125342e-01
-1.68118268e-01 2.86332875e-01 -5.97599685e-01 -1.22281887e-01
-5.91477156e-01 -3.41263622e-01 3.38644326e-01 7.39613712e-01
9.52227652e-01 -1.10400617e+00 -5.84116817e-01 -7.25828111e-02
-3.53494048e-01 -1.07069266e+00 -7.61031866e-01 -3.52439880e-01
-9.16454136e-01 -1.17103863e+00 -8.64990830e-01 -9.81843293e-01
7.64499068e-01 9.24055994e-01 6.12907290e-01 1.66821733e-01
-1.49915472e-01 3.74486417e-01 -5.61407149e-01 2.74798721e-01
1.87103182e-01 -2.94574916e-01 2.64969379e-01 4.64798883e-02
3.30269188e-01 -6.68610275e-01 -9.04901505e-01 4.54153091e-01
-9.10906315e-01 2.10486710e-01 4.29019213e-01 4.41310942e-01
7.82859743e-01 2.91267216e-01 1.72067627e-01 -4.24265862e-01
4.11691517e-01 -7.53161848e-01 -2.14204565e-01 1.93930238e-01
1.10929273e-01 -2.71208346e-01 6.04362667e-01 -4.71872121e-01
-8.04925144e-01 7.78654739e-02 2.82834142e-01 -1.07544768e+00
-5.48330426e-01 4.49637383e-01 -1.69366434e-01 3.19401592e-01
3.50995362e-01 6.73607528e-01 -7.60481805e-02 -3.41944605e-01
3.31869215e-01 6.15704775e-01 6.06751621e-01 -4.29720938e-01
5.89201748e-01 4.97713864e-01 -2.89721310e-01 -8.86856318e-01
-7.35930741e-01 -5.53898454e-01 -7.48931944e-01 -5.01965165e-01
1.00171530e+00 -1.02450109e+00 -5.41062772e-01 6.64475918e-01
-1.15714729e+00 -3.27834338e-01 -2.80823410e-02 6.38552964e-01
-6.23539269e-01 6.72992766e-01 -5.31938434e-01 -4.92398113e-01
-4.81093973e-02 -8.60522330e-01 7.81126201e-01 3.99379104e-01
-6.10742718e-02 -7.66832948e-01 -1.20355546e-01 1.55153364e-01
9.03466344e-02 3.06467086e-01 8.17775846e-01 -3.82114798e-01
-8.25598180e-01 1.49766177e-01 -4.98500437e-01 5.06708860e-01
4.34842497e-01 1.05226278e-01 -5.10753214e-01 -1.75969571e-01
-2.55528539e-01 -2.62022614e-01 8.42016757e-01 7.51703620e-01
1.75787330e+00 -4.06933010e-01 -2.29500651e-01 7.58957922e-01
1.13219953e+00 5.04526377e-01 6.83520377e-01 2.42830247e-01
7.87872732e-01 3.87229145e-01 7.92074084e-01 6.99659884e-01
4.28694338e-01 5.20195365e-01 2.41892338e-01 5.17622940e-02
-1.96881682e-01 -3.50098342e-01 3.87018085e-01 8.29576433e-01
-3.93791676e-01 -3.56310070e-01 -6.63560331e-01 4.10322905e-01
-2.01662827e+00 -1.32292318e+00 3.08370113e-01 2.06695819e+00
4.11866605e-01 8.48519243e-03 2.08541557e-01 -3.23567167e-02
8.65571976e-01 6.73955977e-01 -7.41506755e-01 1.59945592e-01
-1.25834554e-01 -4.73397195e-01 3.81817758e-01 5.69135435e-02
-1.29173911e+00 8.69488060e-01 5.62884808e+00 8.75135601e-01
-1.13668633e+00 -1.17920190e-01 8.93910885e-01 -2.45480046e-01
7.92930499e-02 -2.85692483e-01 -4.03748721e-01 7.45665848e-01
3.93661320e-01 -1.12246104e-01 4.27863210e-01 7.50481546e-01
5.66837311e-01 -2.17509627e-01 -1.01318657e+00 1.39130163e+00
3.77354652e-01 -1.48409629e+00 4.12412971e-01 -2.52278745e-01
6.49857342e-01 3.62831838e-02 6.56995624e-02 6.94352612e-02
9.89571959e-02 -7.74824858e-01 4.81153816e-01 7.21857011e-01
8.42126727e-01 -6.88445568e-01 3.81631583e-01 2.27120101e-01
-1.48892093e+00 -1.67802840e-01 -3.54117423e-01 -3.88317667e-02
7.09923431e-02 1.59455076e-01 -2.40872562e-01 3.50980878e-01
1.03381503e+00 1.41286790e+00 -4.93964761e-01 1.08674872e+00
-3.83671187e-03 4.04564589e-01 -8.83790776e-02 1.09536983e-01
3.90599489e-01 -2.93785155e-01 3.85582268e-01 1.05930305e+00
5.47766685e-01 5.83832800e-01 3.60601455e-01 4.21105772e-01
-5.00370227e-02 7.60643929e-02 -7.85168052e-01 6.37564510e-02
4.66499776e-01 9.55357790e-01 -9.41490531e-01 -4.74525899e-01
-4.53042120e-01 9.55155194e-01 1.40087590e-01 7.45292604e-01
-8.99987042e-01 -2.85290390e-01 5.81743062e-01 3.43927853e-02
4.31206942e-01 -4.66866136e-01 2.16859564e-01 -1.06101453e+00
2.15494856e-02 -7.63530731e-01 6.21462822e-01 -1.16769838e+00
-1.00956464e+00 4.91972744e-01 1.86094135e-01 -1.68138719e+00
-1.98195919e-01 -4.47181314e-01 -7.30068445e-01 4.24052268e-01
-1.15043318e+00 -6.38293624e-01 -5.53372502e-01 9.82705891e-01
7.49051869e-01 -4.67852861e-01 2.69792140e-01 3.63624007e-01
-6.80738807e-01 3.05415183e-01 1.86763793e-01 4.16801244e-01
5.47826827e-01 -7.82099307e-01 2.15852082e-01 8.50272119e-01
5.64081892e-02 3.93129677e-01 3.44751149e-01 -6.04960322e-01
-1.34890497e+00 -1.31359220e+00 3.58516395e-01 -1.49034634e-01
5.39931178e-01 7.30300322e-02 -8.67529690e-01 5.53144038e-01
5.92357889e-02 1.97743773e-01 5.35507679e-01 -4.58021373e-01
-2.38236010e-01 -2.61680156e-01 -8.19370806e-01 6.67695761e-01
1.38015878e+00 -5.15341818e-01 -4.71217811e-01 4.90908504e-01
7.02806056e-01 -3.71069193e-01 -5.98462343e-01 2.36017153e-01
4.34816778e-01 -1.04767168e+00 9.85129952e-01 -5.71139336e-01
5.38502634e-01 -5.63683212e-01 -3.63439888e-01 -1.02415931e+00
-4.43376601e-01 -3.24773133e-01 -1.70211002e-01 9.33505237e-01
-7.97735453e-02 -1.95829540e-01 9.11243737e-01 2.53587574e-01
-8.29646885e-02 -8.76195312e-01 -7.82954752e-01 -5.97125351e-01
-5.12993574e-01 -3.77637804e-01 3.56289208e-01 8.68621469e-01
-1.54455498e-01 -7.55468905e-02 -6.18678272e-01 2.60952950e-01
4.39805716e-01 2.59795070e-01 7.03892946e-01 -7.55822957e-01
-5.19842170e-02 -3.73999029e-01 -6.78583324e-01 -1.64474833e+00
1.43799961e-01 -6.91890001e-01 -1.38682261e-01 -1.53183603e+00
4.00606990e-01 -2.50295281e-01 -7.42573798e-01 3.64441067e-01
6.27941042e-02 3.06983203e-01 2.10781187e-01 3.97437751e-01
-1.24679160e+00 4.90225166e-01 1.30101359e+00 -9.66966227e-02
-3.13935429e-01 -4.42687199e-02 -4.78316933e-01 8.70810807e-01
9.07380164e-01 -1.42478943e-01 -8.18552136e-01 -5.89727163e-01
-1.22473828e-01 2.81144440e-01 4.73271996e-01 -1.08168483e+00
4.64715302e-01 -4.98073131e-01 7.13149369e-01 -9.21364188e-01
2.99746901e-01 -7.32449114e-01 -4.83995974e-02 1.96371332e-01
-4.64451313e-01 3.49669039e-01 7.84703642e-02 6.14663005e-01
-2.96998322e-01 8.95381272e-02 4.41273421e-01 -2.22254276e-01
-1.39187384e+00 7.76129067e-01 -4.43596840e-01 9.94957611e-02
1.14225030e+00 -5.16855299e-01 -3.40478092e-01 -5.67822993e-01
-5.68878055e-01 2.72043437e-01 3.77479672e-01 5.72631717e-01
1.11775112e+00 -1.61875403e+00 -4.56144303e-01 2.46156663e-01
-1.34487068e-02 -1.15260765e-01 7.64657557e-01 7.02106535e-01
-6.72476411e-01 3.67544025e-01 -5.08584440e-01 -6.36851609e-01
-1.29247200e+00 5.50427616e-01 1.94911733e-01 2.86597371e-01
-7.43519247e-01 6.70310736e-01 2.93404639e-01 1.55922666e-01
5.45879781e-01 -3.57836604e-01 -4.32893544e-01 -1.34406939e-01
7.09670663e-01 5.53422928e-01 -4.98825878e-01 -8.29726160e-01
-3.81553769e-01 7.39853263e-01 -1.60703182e-01 2.79576898e-01
1.20950127e+00 -4.00543541e-01 -5.39443083e-02 4.24310029e-01
1.41304195e+00 -4.23498482e-01 -1.47600424e+00 -5.36302567e-01
-3.81506443e-01 -7.81582594e-01 -6.63136840e-02 -2.75619388e-01
-1.44340456e+00 7.24089444e-01 6.49362266e-01 4.77927215e-02
1.32904732e+00 3.75133306e-02 7.44168162e-01 4.21774566e-01
2.54634738e-01 -1.04067707e+00 3.98850441e-01 4.90717143e-01
7.40017474e-01 -1.11365938e+00 1.87002301e-01 -7.89052770e-02
-8.56758773e-01 9.76614177e-01 8.68478358e-01 -2.95622230e-01
7.37730920e-01 -1.59968019e-01 8.35467037e-03 -1.77494079e-01
-6.89696908e-01 -8.51196423e-02 2.63818473e-01 5.68531275e-01
3.33209515e-01 -1.84371725e-01 1.42899841e-01 1.05692044e-01
2.16006011e-01 5.01637682e-02 4.73024458e-01 9.12826300e-01
-5.00419796e-01 -4.48409736e-01 -1.80714995e-01 4.98268813e-01
-2.77268380e-01 1.86588421e-01 -1.09198444e-01 5.99837601e-01
1.15333602e-01 7.65010953e-01 3.22043687e-01 -7.34525383e-01
2.67462462e-01 -2.81961441e-01 2.44095981e-01 -4.32932884e-01
2.71914303e-01 5.19559458e-02 -3.01249206e-01 -8.95148754e-01
-7.27252901e-01 -7.19449520e-01 -1.40375412e+00 -3.01150531e-01
2.62038022e-01 6.52494580e-02 1.83615461e-01 7.93041885e-01
3.60188961e-01 4.54393923e-01 8.82890582e-01 -1.02232778e+00
-1.65168852e-01 -7.40218282e-01 -7.50593245e-01 6.03367329e-01
3.26103151e-01 -7.46482611e-01 -2.18535528e-01 3.37797254e-01] | [8.761964797973633, 0.5932644605636597] |
00a8f193-9838-49e8-91c2-b4b5398a75c2 | self-positioning-point-based-transformer-for | 2303.16450 | null | https://arxiv.org/abs/2303.16450v1 | https://arxiv.org/pdf/2303.16450v1.pdf | Self-positioning Point-based Transformer for Point Cloud Understanding | Transformers have shown superior performance on various computer vision tasks with their capabilities to capture long-range dependencies. Despite the success, it is challenging to directly apply Transformers on point clouds due to their quadratic cost in the number of points. In this paper, we present a Self-Positioning point-based Transformer (SPoTr), which is designed to capture both local and global shape contexts with reduced complexity. Specifically, this architecture consists of local self-attention and self-positioning point-based global cross-attention. The self-positioning points, adaptively located based on the input shape, consider both spatial and semantic information with disentangled attention to improve expressive power. With the self-positioning points, we propose a novel global cross-attention mechanism for point clouds, which improves the scalability of global self-attention by allowing the attention module to compute attention weights with only a small set of self-positioning points. Experiments show the effectiveness of SPoTr on three point cloud tasks such as shape classification, part segmentation, and scene segmentation. In particular, our proposed model achieves an accuracy gain of 2.6% over the previous best models on shape classification with ScanObjectNN. We also provide qualitative analyses to demonstrate the interpretability of self-positioning points. The code of SPoTr is available at https://github.com/mlvlab/SPoTr. | ['Hyunwoo J. Kim', 'Yunyang Xiong', 'Sihyeon Kim', 'Sanghyeok Lee', 'Jinyoung Park'] | 2023-03-29 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Park_Self-Positioning_Point-Based_Transformer_for_Point_Cloud_Understanding_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Park_Self-Positioning_Point-Based_Transformer_for_Point_Cloud_Understanding_CVPR_2023_paper.pdf | cvpr-2023-1 | ['scene-segmentation', '3d-point-cloud-classification', '3d-part-segmentation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-1.11847781e-01 2.49307859e-03 6.90577552e-02 -5.73696375e-01
-8.75674188e-01 -4.84304368e-01 3.84219259e-01 2.20619757e-02
-6.28875047e-02 -7.07305372e-02 -1.03382409e-01 -2.51426011e-01
-4.20416445e-01 -8.03383768e-01 -1.12521541e+00 -5.53690314e-01
1.49874568e-01 8.00866902e-01 4.35546100e-01 -2.09994122e-01
2.11165294e-01 9.20341372e-01 -1.70144749e+00 1.95876405e-01
1.11714017e+00 1.14245558e+00 7.24547923e-01 5.07127047e-01
-1.91482887e-01 1.11921653e-01 -2.21880272e-01 -3.53634596e-01
3.17117631e-01 3.87350172e-01 -6.07332349e-01 1.04923919e-01
6.47969186e-01 -1.95444390e-01 -2.17932701e-01 9.52444494e-01
4.31147397e-01 3.31727639e-02 3.56812418e-01 -1.37258744e+00
-1.08901036e+00 4.45138931e-01 -7.50180423e-01 2.71611780e-01
-1.88074231e-01 3.54692847e-01 1.35262203e+00 -1.15740526e+00
1.13806263e-01 1.40977824e+00 7.98824847e-01 2.12596744e-01
-1.06589377e+00 -9.27003622e-01 4.14851904e-01 3.02014679e-01
-1.42799473e+00 -1.38240635e-01 1.06861532e+00 -3.95460635e-01
1.15272677e+00 3.38823408e-01 6.79078996e-01 6.64142728e-01
-1.45695537e-01 8.62814367e-01 6.63449466e-01 3.98847349e-02
-8.88017297e-04 -2.56452739e-01 3.02725047e-01 5.39928675e-01
-2.53739990e-02 -9.49157402e-02 -3.02918196e-01 -2.67370651e-03
1.12084270e+00 3.97264510e-01 -8.36820155e-02 -4.42947626e-01
-1.09384239e+00 7.61827707e-01 1.14061964e+00 2.22885758e-01
-4.19755697e-01 3.74662757e-01 3.14238258e-02 -1.99510574e-01
6.52108610e-01 4.63106841e-01 -7.02564895e-01 1.15002342e-01
-6.03419185e-01 2.53866185e-02 6.64420351e-02 1.37604964e+00
8.64715993e-01 -1.42331600e-01 -2.43656009e-01 1.02577829e+00
4.18032914e-01 8.92348647e-01 2.88578421e-01 -8.17629755e-01
5.44025362e-01 1.00951040e+00 -1.47453666e-01 -1.15287900e+00
-3.35702807e-01 -7.37814009e-01 -7.36323893e-01 1.08608194e-01
-3.83321382e-02 2.95096427e-01 -1.14273679e+00 1.65336871e+00
3.67863089e-01 3.83948445e-01 -2.60432899e-01 1.04502547e+00
9.21915889e-01 5.95316768e-01 9.90763456e-02 4.79130864e-01
1.42408359e+00 -1.12153602e+00 -2.74697542e-01 -3.20225000e-01
1.78499252e-01 -4.90844786e-01 1.32317984e+00 -1.55404240e-01
-1.11828649e+00 -6.29542589e-01 -8.13887358e-01 -5.60539007e-01
-3.06376934e-01 2.13791236e-01 5.40799558e-01 1.31160840e-01
-1.00602794e+00 5.70851207e-01 -1.08863151e+00 -2.27741137e-01
8.89886558e-01 6.09243453e-01 -1.79047525e-01 7.52307102e-02
-6.20299280e-01 3.46682400e-01 -4.06733043e-02 2.43020192e-01
-3.92042249e-01 -1.13689160e+00 -8.00779223e-01 4.74507600e-01
2.07567617e-01 -7.47329533e-01 1.26662624e+00 -7.99097836e-01
-1.29242456e+00 7.74273515e-01 -3.20293486e-01 -2.33217523e-01
2.87018359e-01 -3.20943773e-01 9.88432914e-02 1.52607486e-01
3.61861914e-01 8.76969099e-01 7.36201525e-01 -1.31812346e+00
-5.10866225e-01 -7.66896486e-01 1.05539955e-01 4.05258179e-01
-4.33304310e-02 -1.68878570e-01 -8.95216823e-01 -5.81483245e-01
5.64999223e-01 -8.57881844e-01 -9.63806137e-02 1.13089763e-01
-3.41296524e-01 -5.44229090e-01 1.02976799e+00 -2.85604984e-01
6.10785007e-01 -2.34605885e+00 3.75864916e-02 7.80758401e-03
4.00473535e-01 1.34087518e-01 -3.25177282e-01 3.58039588e-02
-2.53939003e-01 3.66506785e-01 -4.18598711e-01 -5.82982540e-01
3.80779356e-02 2.47560591e-01 -4.83687520e-01 3.20411593e-01
5.87606490e-01 1.52138722e+00 -7.39138067e-01 -3.06357950e-01
4.61318314e-01 6.63616061e-01 -6.92899466e-01 3.86569723e-02
-2.54914135e-01 3.37083787e-01 -6.02617383e-01 8.53507102e-01
8.78214240e-01 -6.30647421e-01 -4.14405555e-01 -4.58833605e-01
-1.09575480e-01 1.96126953e-01 -6.73054397e-01 1.65193963e+00
-5.06629825e-01 4.33203042e-01 5.81560172e-02 -5.25597572e-01
8.23781729e-01 1.70222986e-02 6.18272603e-01 -7.54204750e-01
1.36154965e-02 2.58484501e-02 -4.79167067e-02 -2.68392175e-01
4.22533721e-01 4.30442058e-02 7.54708573e-02 1.51809439e-01
-1.42954206e-02 -1.82052091e-01 -4.19432372e-01 4.73878980e-02
9.24558282e-01 9.34696943e-02 -3.51939537e-03 -2.01448590e-01
2.81233013e-01 -6.27721623e-02 5.64276099e-01 5.19781470e-01
-2.42754910e-02 9.90828395e-01 4.16286811e-02 -3.38669032e-01
-1.03688443e+00 -1.06031585e+00 -2.66146779e-01 1.14288294e+00
4.37692761e-01 -2.89682716e-01 -4.23468143e-01 -5.23660004e-01
3.23230028e-01 6.26938939e-01 -6.49641216e-01 1.06631845e-01
-5.63715816e-01 -4.37468469e-01 3.17106336e-01 1.05386341e+00
4.83289242e-01 -1.16749966e+00 -6.77530766e-01 -8.71676803e-02
-1.42219409e-01 -1.06562519e+00 -7.04112232e-01 -2.87860185e-02
-1.04987621e+00 -9.77916598e-01 -4.98471618e-01 -4.92106706e-01
7.85055876e-01 5.77426553e-01 1.07582867e+00 1.45160168e-01
-7.42262080e-02 6.52588308e-01 -3.29147726e-01 -5.25233507e-01
4.08879459e-01 3.25645328e-01 -3.11491162e-01 7.63449371e-02
3.72111827e-01 -7.81191111e-01 -7.42583871e-01 5.12578905e-01
-7.18436182e-01 1.31075323e-01 8.33767593e-01 7.53612280e-01
1.02769935e+00 -3.17079902e-01 3.18545610e-01 -7.72887766e-01
1.53038770e-01 -5.05414188e-01 -6.22547984e-01 7.73580447e-02
-2.42679492e-01 -1.01392316e-02 2.97537237e-01 -2.82139391e-01
-8.01121533e-01 5.32314442e-02 -2.70834178e-01 -1.07637846e+00
-2.52319366e-01 2.44047597e-01 -4.12798971e-01 -2.35754669e-01
1.85368374e-01 2.36765310e-01 -1.27196878e-01 -6.87073588e-01
4.22931552e-01 5.55586874e-01 4.22743022e-01 -7.02864766e-01
8.06380808e-01 6.23524904e-01 -1.67214826e-01 -8.68976474e-01
-7.77865589e-01 -7.64623702e-01 -7.06752956e-01 2.76327766e-02
8.96904945e-01 -9.24640775e-01 -9.52160895e-01 4.17293191e-01
-1.20083213e+00 -3.08447748e-01 -4.15505558e-01 -3.86160724e-02
-5.63016772e-01 2.18724787e-01 -3.10755193e-01 -6.99063480e-01
-5.35710871e-01 -1.23733354e+00 1.76130903e+00 1.50108472e-01
2.25625902e-01 -8.82458389e-01 -3.09163541e-01 4.59106505e-01
3.60142082e-01 4.45304774e-02 7.25423992e-01 -6.82518542e-01
-1.09486961e+00 -2.01032553e-02 -5.67879498e-01 -3.00841429e-03
3.70984785e-02 -2.04936445e-01 -1.11869812e+00 -2.31830582e-01
-3.12822424e-02 6.45464798e-03 7.07175076e-01 4.34228480e-01
1.75347173e+00 -1.61276907e-01 -4.53403533e-01 1.06603396e+00
1.42322397e+00 7.92901665e-02 4.85138386e-01 1.07989892e-01
1.24083054e+00 3.73268068e-01 5.11694491e-01 1.47586927e-01
7.09641695e-01 7.29840100e-01 7.97571540e-01 -2.35602811e-01
-1.34114683e-01 -3.94013703e-01 -2.55037069e-01 8.42036605e-01
-2.20861077e-01 -1.21294446e-01 -1.17270470e+00 6.85839176e-01
-1.94342577e+00 -7.14208126e-01 -2.19425321e-01 2.01969099e+00
2.46968165e-01 -8.42525810e-02 -2.32747719e-01 -1.69868395e-01
6.36146963e-01 1.34590149e-01 -8.71166885e-01 7.27717280e-02
1.84431288e-03 2.00688988e-01 6.82542861e-01 4.95084792e-01
-1.14236617e+00 1.08591092e+00 5.30898523e+00 8.81803989e-01
-1.27670145e+00 2.83518672e-01 5.22204638e-01 -1.25778079e-01
-4.11096424e-01 -2.28447691e-01 -6.07486606e-01 5.53807020e-01
4.63328362e-01 1.78244617e-02 2.92901486e-01 9.76966441e-01
7.07141450e-03 4.38558251e-01 -9.63189363e-01 1.10412359e+00
-5.35080917e-02 -1.19587135e+00 1.26458213e-01 8.31984431e-02
4.45171535e-01 6.24800920e-01 2.90666074e-01 2.51591802e-01
1.77293673e-01 -9.65754867e-01 8.64122987e-01 4.39638138e-01
7.85824060e-01 -7.29772627e-01 5.74291110e-01 3.84127378e-01
-1.50284994e+00 -1.98494241e-01 -4.68764871e-01 7.60843754e-02
4.16703969e-02 3.50710630e-01 -5.80755234e-01 7.68007576e-01
1.01381016e+00 8.06723535e-01 -6.61464155e-01 1.02876961e+00
-8.86780396e-02 4.00836945e-01 -6.93181515e-01 2.05834329e-01
3.00679535e-01 -2.12370604e-01 7.70528257e-01 9.57549810e-01
4.14426267e-01 4.16384071e-01 9.20189470e-02 1.21858752e+00
-3.53819281e-02 -4.02539708e-02 -4.89260912e-01 3.17404240e-01
6.02393806e-01 1.34139180e+00 -7.34550118e-01 -1.79925114e-01
-3.26536208e-01 7.19744086e-01 6.02806091e-01 3.25224787e-01
-9.10225928e-01 -2.17344522e-01 8.93615127e-01 4.30272162e-01
6.78912699e-01 -3.16685200e-01 -7.34503567e-01 -9.88083184e-01
1.53084517e-01 -4.54263985e-01 1.42197147e-01 -1.18191409e+00
-1.48503578e+00 7.22729087e-01 3.21042277e-02 -1.10037398e+00
3.87256205e-01 -6.26501262e-01 -7.06497908e-01 8.55169237e-01
-1.36610436e+00 -1.64305639e+00 -5.83536029e-01 4.76208121e-01
7.48436987e-01 2.10931882e-01 3.49571973e-01 2.42202953e-01
-4.92164642e-01 6.64720953e-01 -2.35010147e-01 8.06530789e-02
2.44467378e-01 -1.37410188e+00 7.82934546e-01 6.36486113e-01
2.07570583e-01 7.51840115e-01 2.60842115e-01 -6.74129963e-01
-1.36634183e+00 -1.30712092e+00 5.30805707e-01 -7.72831559e-01
4.36763644e-01 -4.69767958e-01 -1.25094080e+00 9.56919372e-01
-5.96940182e-02 2.71440238e-01 2.96299607e-01 2.52673864e-01
-4.64394182e-01 -1.81707457e-01 -1.04877126e+00 5.36456406e-01
1.41084337e+00 -4.88480389e-01 -5.46915531e-01 2.86327511e-01
1.10524786e+00 -7.06430554e-01 -7.37551987e-01 6.71423435e-01
1.63270965e-01 -8.01117063e-01 1.22533119e+00 -2.05597699e-01
3.34375173e-01 -3.88119012e-01 -1.88246578e-01 -1.03583658e+00
-7.99839318e-01 -1.53137907e-01 1.02860548e-01 1.12038267e+00
2.66384989e-01 -9.29861724e-01 7.37543702e-01 7.51297235e-01
-6.00300252e-01 -1.06999958e+00 -1.03925741e+00 -6.53158903e-01
7.92037249e-02 -7.10646868e-01 1.07422328e+00 6.97585762e-01
-5.26739120e-01 4.29814398e-01 1.99490130e-01 7.52187610e-01
6.34072006e-01 5.68113208e-01 6.11996651e-01 -1.36881387e+00
-1.52022809e-01 -5.48639894e-01 -5.92155993e-01 -1.40886962e+00
1.15806662e-01 -1.07784545e+00 5.69933094e-02 -1.61610651e+00
1.05969809e-01 -7.98569798e-01 -2.40233377e-01 7.53481150e-01
-1.96222484e-01 2.89627910e-01 3.98560286e-01 4.38135922e-01
-6.28958225e-01 8.38262200e-01 1.32741594e+00 -2.40644976e-01
-1.50119826e-01 1.36531696e-01 -7.39720821e-01 7.02344835e-01
7.76288629e-01 -3.77273232e-01 -4.31459725e-01 -8.89514863e-01
-3.37552652e-02 -2.44780660e-01 6.73021674e-01 -9.69062626e-01
3.98226500e-01 6.20958507e-02 3.31391782e-01 -1.13118625e+00
6.19114995e-01 -1.04463041e+00 1.52288318e-01 5.59742674e-02
-3.90881598e-02 2.85159588e-01 4.02420133e-01 7.23600745e-01
-4.62652668e-02 7.53563643e-02 6.38382912e-01 -4.49875221e-02
-5.95091164e-01 8.58303010e-01 4.04782504e-01 -1.94877490e-01
1.03622198e+00 -2.54866868e-01 -2.13583097e-01 -2.42656861e-02
-5.66731095e-01 4.85457212e-01 6.08422756e-01 5.90120614e-01
7.68341064e-01 -1.32187128e+00 -4.18027371e-01 4.07168806e-01
2.07386166e-01 7.48604596e-01 6.50495470e-01 7.18721032e-01
-4.37644660e-01 4.96433169e-01 -8.84230584e-02 -1.14951336e+00
-1.19134831e+00 6.69362962e-01 3.62702191e-01 7.33773112e-02
-9.97567832e-01 9.23352957e-01 7.05687881e-01 -6.31100595e-01
-4.70619611e-02 -7.65731514e-01 -1.06064923e-01 -3.14884931e-01
1.45654902e-01 2.37235859e-01 1.33147940e-01 -8.07148635e-01
-5.42529047e-01 1.15283489e+00 3.66044044e-02 8.65322575e-02
1.47780776e+00 -3.95151153e-02 -1.56296253e-01 4.26521957e-01
1.06431246e+00 -4.43346128e-02 -1.66777658e+00 -2.81960785e-01
-2.26622313e-01 -6.78135455e-01 1.77319184e-01 -6.94701612e-01
-1.38881350e+00 9.91658390e-01 4.91323560e-01 9.94564295e-02
1.13906336e+00 3.84239405e-01 6.60189331e-01 1.14010885e-01
4.36542779e-01 -4.47378069e-01 -6.82775602e-02 5.93482673e-01
1.31167042e+00 -1.26741576e+00 -2.91233689e-01 -6.29584491e-01
-5.95395803e-01 6.88882530e-01 8.44661176e-01 -3.35992485e-01
7.38729894e-01 2.13289067e-01 -7.72671998e-02 -5.85986197e-01
-7.51403809e-01 -3.30405980e-01 6.01045251e-01 6.01755083e-01
5.64220063e-02 2.16062620e-01 2.95583040e-01 7.76451707e-01
-2.44144753e-01 -3.89526814e-01 -8.39697570e-02 5.40928483e-01
-2.18536466e-01 -6.77591324e-01 -2.78550208e-01 4.14837480e-01
-1.14319347e-01 -7.60114044e-02 -3.29785198e-01 7.28479803e-01
1.11859761e-01 5.96983969e-01 5.71633756e-01 -3.83009225e-01
5.87150335e-01 -1.55005887e-01 2.96811849e-01 -6.38466895e-01
-6.16690397e-01 2.45547984e-02 -5.38481891e-01 -7.63042748e-01
-2.05877706e-01 -6.01143003e-01 -1.32657421e+00 -1.94399148e-01
-3.06233376e-01 -1.28246263e-01 6.44703031e-01 6.77349627e-01
1.01684082e+00 8.00940752e-01 3.79326552e-01 -1.13408279e+00
-4.18958038e-01 -9.83698606e-01 -2.12260321e-01 4.16953534e-01
3.24558377e-01 -8.41400206e-01 -1.80394128e-01 -2.94711739e-01] | [7.890538692474365, -3.4414405822753906] |
720e3be6-d523-43f5-8de4-b09d827461c9 | deep-head-pose-estimation-using-synthetic | null | null | http://openaccess.thecvf.com/content_ICCV_2019/html/Kuhnke_Deep_Head_Pose_Estimation_Using_Synthetic_Images_and_Partial_Adversarial_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Kuhnke_Deep_Head_Pose_Estimation_Using_Synthetic_Images_and_Partial_Adversarial_ICCV_2019_paper.pdf | Deep Head Pose Estimation Using Synthetic Images and Partial Adversarial Domain Adaption for Continuous Label Spaces | Head pose estimation aims at predicting an accurate pose from an image. Current approaches rely on supervised deep learning, which typically requires large amounts of labeled data. Manual or sensor-based annotations of head poses are prone to errors. A solution is to generate synthetic training data by rendering 3D face models. However, the differences (domain gap) between rendered (source-domain) and real-world (target-domain) images can cause low performance. Advances in visual domain adaptation allow reducing the influence of domain differences using adversarial neural networks, which match the feature spaces between domains by enforcing domain-invariant features. While previous work on visual domain adaptation generally assumes discrete and shared label spaces, these assumptions are both invalid for pose estimation tasks. We are the first to present domain adaptation for head pose estimation with a focus on partially shared and continuous label spaces. More precisely, we adapt the predominant weighting approaches to continuous label spaces by applying a weighted resampling of the source domain during training. To evaluate our approach, we revise and extend existing datasets resulting in a new benchmark for visual domain adaption. Our experiments show that our method improves the accuracy of head pose estimation for real-world images despite using only labels from synthetic images.
| [' Jorn Ostermann', 'Felix Kuhnke'] | 2019-10-01 | null | null | null | iccv-2019-10 | ['head-pose-estimation'] | ['computer-vision'] | [ 3.18599105e-01 3.19700927e-01 -9.17691961e-02 -8.24137807e-01
-8.70945871e-01 -6.04449451e-01 6.26958191e-01 -1.96800381e-01
-4.95704383e-01 9.38982606e-01 3.73727590e-01 2.77094722e-01
3.23383600e-01 -5.33662736e-01 -8.21867645e-01 -3.86314601e-01
1.10699952e-01 8.06560934e-01 2.37962097e-01 -2.16414630e-01
-1.91250704e-02 7.19301224e-01 -1.73831367e+00 8.98773447e-02
4.58555728e-01 7.20994711e-01 -1.59689248e-01 4.43604618e-01
-2.28974149e-02 1.88807964e-01 -8.99576306e-01 -4.03748661e-01
3.67074400e-01 -3.44165891e-01 -8.16744745e-01 3.51276726e-01
9.17837441e-01 -4.58664000e-01 -1.49795935e-01 9.84146476e-01
8.40470970e-01 9.59241465e-02 9.02707815e-01 -1.69170201e+00
-5.30510664e-01 -1.62087277e-01 -3.32583666e-01 -2.63312459e-01
8.51304770e-01 1.99837778e-02 3.46051425e-01 -9.26727891e-01
1.03035462e+00 1.32342780e+00 9.19337571e-01 1.14911091e+00
-1.37630129e+00 -8.25755775e-01 2.28321806e-01 2.33319789e-01
-1.54352021e+00 -7.27070332e-01 9.78198588e-01 -4.77452368e-01
6.23166084e-01 1.93515405e-01 6.36520743e-01 1.54865253e+00
-1.71960592e-01 6.07338667e-01 1.38936448e+00 -5.43068886e-01
3.66884708e-01 1.42189920e-01 -3.52497309e-01 4.11064476e-01
8.35635141e-02 1.62689269e-01 -6.84824407e-01 -2.60316044e-01
6.13872290e-01 -3.03923130e-01 -3.79335940e-01 -1.02208829e+00
-9.04120505e-01 8.61614525e-01 4.14323509e-01 -2.22552437e-02
-7.28741214e-02 -2.01568067e-01 4.93488938e-01 2.04033121e-01
6.25819802e-01 2.70299673e-01 -3.26916307e-01 2.34668061e-01
-1.00897503e+00 5.27239084e-01 7.59066820e-01 1.08783400e+00
6.71291709e-01 4.37264144e-02 -1.90718457e-01 9.06653643e-01
4.18669701e-01 6.45033360e-01 6.10795379e-01 -1.03033400e+00
1.88661024e-01 3.14163268e-01 1.72242433e-01 -7.40231395e-01
-6.49498582e-01 -1.69444904e-01 -3.77665102e-01 6.50868535e-01
8.39157104e-01 -7.90335760e-02 -1.29587483e+00 2.06844234e+00
6.27460301e-01 6.42189234e-02 -1.17896825e-01 9.50503170e-01
9.54686701e-01 2.46277917e-02 3.33217412e-01 -9.15007964e-02
1.11350262e+00 -5.77808857e-01 -6.53867841e-01 -5.66261709e-01
3.70685786e-01 -6.46963179e-01 1.04962599e+00 1.32759675e-01
-9.50892746e-01 -3.30514550e-01 -1.02602124e+00 -1.32298455e-01
-4.48679984e-01 -1.08598419e-01 1.55100808e-01 8.94991934e-01
-1.19455194e+00 2.97075391e-01 -5.76143563e-01 -5.53900480e-01
3.98036927e-01 5.97835779e-01 -7.98314035e-01 4.13102098e-03
-1.13967788e+00 1.17639601e+00 1.01718575e-01 -3.11962396e-01
-8.22755933e-01 -6.24568403e-01 -1.24465799e+00 -5.31665981e-01
5.63416108e-02 -6.64475203e-01 1.39545679e+00 -1.11456978e+00
-1.56252122e+00 1.43843699e+00 -1.60353854e-01 -3.25472444e-01
8.23447406e-01 -1.30545059e-02 -3.73074889e-01 7.58809000e-02
1.07860379e-01 1.06990850e+00 1.24999774e+00 -1.57943583e+00
-2.58401811e-01 -6.68306887e-01 -9.13062021e-02 2.93032527e-01
-1.93468764e-01 -3.52912545e-02 -2.47746378e-01 -6.59739614e-01
2.16003120e-01 -1.09477341e+00 1.39598146e-01 5.15558720e-01
-1.34251028e-01 -3.59481052e-02 8.56530607e-01 -7.47690916e-01
5.85296571e-01 -1.92017686e+00 1.36197418e-01 9.81975868e-02
2.03515857e-01 2.45215148e-01 -5.42779602e-02 -2.99883895e-02
-4.04976547e-01 -3.64534199e-01 -2.21046671e-01 -8.09324324e-01
1.70657799e-01 4.11678761e-01 -1.70918822e-01 9.24343586e-01
-9.45434421e-02 6.96080506e-01 -8.49866450e-01 -6.90926015e-01
3.39834720e-01 6.80389106e-01 -6.71346128e-01 2.97449350e-01
-1.76329985e-01 9.08211648e-01 5.80719998e-03 4.39000100e-01
9.28602457e-01 2.76133448e-01 7.30000213e-02 -2.19959959e-01
3.35916817e-01 2.27579862e-01 -1.16256106e+00 1.82817197e+00
-3.75096262e-01 4.89406437e-01 1.07271679e-01 -7.94404805e-01
9.33723092e-01 3.11629266e-01 4.46195096e-01 -6.66783988e-01
1.05922572e-01 1.54994577e-01 -3.76190305e-01 -3.77331048e-01
2.17697963e-01 -5.53913176e-01 -1.76104218e-01 2.28614837e-01
1.75997406e-01 -4.69043553e-01 -2.94449061e-01 -2.25966230e-01
5.98149240e-01 5.08737266e-01 2.89610445e-01 1.79482028e-02
3.72462809e-01 -1.70351535e-01 5.43916762e-01 1.42319679e-01
-5.77814162e-01 1.20269275e+00 3.25705260e-01 -4.85542506e-01
-1.17329788e+00 -1.31296277e+00 -2.50608861e-01 1.29939842e+00
-8.95281285e-02 -2.14715023e-02 -1.14082587e+00 -9.27646995e-01
1.13564521e-01 7.42343426e-01 -9.70613003e-01 -2.15340093e-01
-7.60727465e-01 -4.28278327e-01 6.83718145e-01 7.34719217e-01
2.75573939e-01 -1.00182080e+00 -5.28623581e-01 -7.03636482e-02
-1.89159796e-01 -1.11010218e+00 -5.05782545e-01 6.74732961e-03
-5.62362015e-01 -1.00778794e+00 -1.15977204e+00 -7.74046302e-01
8.99200559e-01 -2.98764139e-01 1.29916942e+00 -2.71316350e-01
-4.95590419e-02 6.72939062e-01 -1.91893309e-01 -6.17209792e-01
-4.17457968e-01 -4.78250645e-02 3.64908844e-01 -7.10578412e-02
4.50345069e-01 -5.55414557e-01 -6.30373120e-01 3.02793473e-01
-5.65816998e-01 -1.52672648e-01 1.08157031e-01 8.17844391e-01
4.23370808e-01 -8.91172051e-01 4.96030480e-01 -7.66260326e-01
5.65369427e-01 -2.18578458e-01 -3.25751990e-01 7.09553286e-02
-3.47064674e-01 2.36365467e-01 3.48513067e-01 -6.65121138e-01
-1.08718646e+00 4.63445723e-01 -2.78773576e-01 -4.32360291e-01
-6.32036567e-01 -2.34582961e-01 -4.65673327e-01 -1.94509938e-01
1.05568445e+00 8.30381587e-02 2.75816709e-01 -3.99891615e-01
2.90028721e-01 4.79508311e-01 7.74038732e-01 -5.83633006e-01
8.78385782e-01 6.98597252e-01 4.05524857e-02 -6.11448228e-01
-6.27993166e-01 -8.34047124e-02 -1.01024103e+00 -2.18366921e-01
7.38609910e-01 -7.82991350e-01 -3.64080191e-01 4.56079006e-01
-1.00820398e+00 -3.33158374e-01 -3.46806288e-01 3.79499555e-01
-8.51318657e-01 3.93408090e-01 -1.16795629e-01 -4.71884549e-01
-5.92359267e-02 -1.02180266e+00 1.43318820e+00 -9.68828723e-02
-6.79793239e-01 -9.84571397e-01 2.14363247e-01 2.11830541e-01
2.42872015e-01 4.95963901e-01 5.12054861e-01 -6.72454834e-01
1.97059996e-02 -3.15314293e-01 -3.13112289e-02 1.54220715e-01
2.11156327e-02 -3.91103327e-01 -1.35168171e+00 -4.97887760e-01
-1.62539244e-01 -4.85924751e-01 2.89863259e-01 3.54127198e-01
9.30113733e-01 -5.55292144e-02 -3.41624767e-01 6.28830135e-01
9.54250097e-01 -2.10937783e-02 4.96668279e-01 4.45375353e-01
5.50001562e-01 8.88874531e-01 5.77002108e-01 4.13869530e-01
6.17187321e-01 1.15954268e+00 5.19569695e-01 -2.14382529e-01
-4.59484875e-01 -4.96634126e-01 2.19377711e-01 1.54993743e-01
7.31644258e-02 1.83395460e-01 -9.93662477e-01 5.55087686e-01
-1.32082915e+00 -6.10086918e-01 4.12641943e-01 2.45198131e+00
9.71383214e-01 -3.26347575e-02 4.95252907e-01 2.37933323e-01
7.01637328e-01 -3.41307521e-02 -6.50814414e-01 -2.68876433e-01
2.62039691e-01 2.98839301e-01 5.08532345e-01 5.87713778e-01
-1.20406127e+00 7.89310336e-01 6.72120142e+00 3.88104856e-01
-1.29105151e+00 3.92491817e-01 1.57036856e-01 -2.34842673e-01
-9.75036547e-02 -5.26375175e-01 -7.89107859e-01 3.49545747e-01
7.18611598e-01 -4.39022779e-02 1.66694894e-01 1.03223431e+00
-1.06142223e-01 1.46995887e-01 -1.39142239e+00 1.18311071e+00
4.89736319e-01 -6.41037166e-01 -2.29756996e-01 -1.74402129e-02
5.56141376e-01 -2.38184422e-01 2.80419856e-01 3.99643809e-01
1.63139910e-01 -1.22205424e+00 8.42955828e-01 2.98076063e-01
1.27965569e+00 -7.35587537e-01 4.49679077e-01 2.62016028e-01
-1.04035830e+00 2.46859223e-01 -2.36417919e-01 2.24295687e-02
4.62784953e-02 -8.67729038e-02 -1.13104749e+00 6.71758549e-03
6.81306362e-01 2.86445946e-01 -6.50942028e-01 9.14390683e-01
-1.81146249e-01 9.02496502e-02 -4.63467240e-01 2.96027869e-01
-2.17633858e-01 3.13290060e-01 4.84028101e-01 1.06544375e+00
5.30460067e-02 -2.93374807e-01 8.05103034e-02 4.99825418e-01
2.24824771e-02 -5.79810031e-02 -9.66154337e-01 8.65604639e-01
4.46521282e-01 7.31545508e-01 -4.65325028e-01 -1.87702253e-01
-3.97493869e-01 1.12929082e+00 3.63124758e-01 3.39075923e-01
-8.94094288e-01 -1.64986905e-02 9.04966950e-01 4.46843326e-01
-8.89772177e-02 -1.84205920e-01 -4.84170914e-01 -9.39860225e-01
3.06209307e-02 -8.73810410e-01 2.14722425e-01 -7.37623215e-01
-1.05707157e+00 6.15717590e-01 4.89530861e-01 -1.46224117e+00
-7.88781524e-01 -6.37385726e-01 -5.47705069e-02 8.36353004e-01
-1.32224131e+00 -1.50885236e+00 -5.23622096e-01 7.99218297e-01
5.49072027e-01 -1.99415293e-02 1.13204312e+00 3.70243162e-01
1.29655944e-02 1.09173810e+00 -1.46775663e-01 1.86987504e-01
1.26355731e+00 -1.26919353e+00 5.81008971e-01 3.51999819e-01
-4.37531509e-02 3.19680393e-01 1.14644825e+00 -4.87117559e-01
-8.81359875e-01 -1.12222791e+00 8.16072166e-01 -8.15502584e-01
-3.17541435e-02 -6.14239514e-01 -9.73918319e-01 9.13384378e-01
-1.27625495e-01 4.37780708e-01 5.39369047e-01 -1.20629393e-01
-5.67097306e-01 8.44223723e-02 -1.71505034e+00 5.81854045e-01
1.31422269e+00 -5.77569067e-01 -7.91825116e-01 3.92799884e-01
2.39925086e-01 -8.34113002e-01 -7.66355574e-01 5.55232286e-01
7.44745791e-01 -9.66795623e-01 1.25058687e+00 -5.82893252e-01
-4.27039154e-03 -2.81515658e-01 -1.63750812e-01 -1.56461275e+00
2.07748101e-03 -5.20166457e-02 1.45610375e-02 1.02573860e+00
4.94631305e-02 -6.05458319e-01 1.17470443e+00 8.66786659e-01
5.63377887e-02 -2.92138010e-01 -1.12706745e+00 -8.26327801e-01
4.46218759e-01 -3.94029796e-01 7.68914759e-01 9.34371531e-01
-1.68395966e-01 4.07874435e-02 -2.51990706e-01 1.28765360e-01
7.99877286e-01 -2.32843444e-01 9.30089772e-01 -1.33773172e+00
5.59746288e-02 -1.42693117e-01 -7.18017936e-01 -7.36574054e-01
7.31447577e-01 -7.05235660e-01 4.24946994e-02 -1.31241477e+00
-4.47147563e-02 -1.70661524e-01 2.60711521e-01 6.07330561e-01
1.56166077e-01 8.14707041e-01 2.08271563e-01 7.14792758e-02
-2.74844259e-01 5.67781091e-01 1.23265636e+00 -6.46509379e-02
5.54130189e-02 1.21297722e-03 -2.06735983e-01 1.11966765e+00
7.02755094e-01 -4.26617146e-01 -4.40357894e-01 -3.92775059e-01
-1.61649853e-01 -1.26894712e-01 4.49156940e-01 -1.13972914e+00
6.52326941e-02 1.69438273e-02 8.02964449e-01 -2.87449211e-01
7.79546916e-01 -8.94053102e-01 1.12537928e-01 3.68989021e-01
-2.14248344e-01 -1.40864551e-01 3.25273097e-01 3.40435892e-01
-1.05092764e-01 -2.17372701e-01 1.19947851e+00 -7.96761289e-02
-9.15484428e-01 2.01671153e-01 -9.85140949e-02 2.14333877e-01
1.09196520e+00 -5.53182483e-01 6.35663867e-02 -7.44840384e-01
-1.27883804e+00 -2.55082726e-01 9.31469977e-01 5.12181699e-01
6.44196272e-01 -1.47648263e+00 -7.49187291e-01 5.92490971e-01
3.96582514e-01 -5.05345613e-02 -3.02501139e-03 3.55904400e-01
-4.67398942e-01 2.51835376e-01 -5.47720373e-01 -6.50284469e-01
-1.60623622e+00 4.82515514e-01 6.09098494e-01 1.49236947e-01
-4.34596419e-01 7.94310153e-01 2.53956228e-01 -9.07857418e-01
4.19953346e-01 -9.03797224e-02 -2.48427883e-01 1.17844842e-01
4.29157078e-01 2.45791227e-01 2.68956304e-01 -1.18041790e+00
-5.46983540e-01 7.97661364e-01 1.41387060e-01 -3.39228511e-01
1.02955782e+00 -1.32926747e-01 3.95349026e-01 1.05853364e-01
1.30470574e+00 -4.42535467e-02 -1.51759756e+00 -2.28877261e-01
-1.80454582e-01 -5.31600356e-01 -2.69265473e-01 -7.37110078e-01
-8.59369218e-01 9.00015712e-01 1.09646547e+00 -4.00477856e-01
1.09639168e+00 5.69841787e-02 4.63811457e-01 9.05849785e-03
6.75452232e-01 -1.13120484e+00 1.59356549e-01 3.21369559e-01
1.17125285e+00 -1.33636820e+00 3.18335854e-02 -4.05678928e-01
-6.99364841e-01 7.41263092e-01 8.03723693e-01 8.49624630e-03
5.29890776e-01 2.98324466e-01 3.75669569e-01 8.27641711e-02
-9.81390253e-02 -1.73944131e-01 3.98016363e-01 1.36108935e+00
4.66350287e-01 1.12863958e-01 -2.63638586e-01 1.92384481e-01
-4.85055476e-01 8.63640085e-02 1.38809547e-01 9.36720490e-01
-1.56183124e-01 -1.43362141e+00 -9.10694838e-01 1.15694501e-01
-3.32643807e-01 3.36553663e-01 -4.01342303e-01 9.77649629e-01
3.04958761e-01 6.34593010e-01 9.06351507e-02 -1.10072978e-01
5.46117544e-01 4.38075006e-01 8.70215118e-01 -8.17465246e-01
-1.19724207e-01 -1.69840723e-01 -6.70841709e-02 -3.70104879e-01
-3.36275011e-01 -7.82612503e-01 -1.09282517e+00 -1.60580575e-01
1.33817300e-01 -2.27422372e-01 7.26897955e-01 7.58831918e-01
1.52689174e-01 2.43676707e-01 2.92793870e-01 -1.42105818e+00
-6.94806933e-01 -1.10864520e+00 -4.93887246e-01 8.93225431e-01
5.01500249e-01 -1.28725564e+00 -2.80394703e-01 2.74462700e-01] | [13.401111602783203, 0.1380205601453781] |
14e76535-5c42-4471-b228-fefd294ce7eb | why-is-this-misleading-detecting-news | 2302.05852 | null | https://arxiv.org/abs/2302.05852v1 | https://arxiv.org/pdf/2302.05852v1.pdf | "Why is this misleading?": Detecting News Headline Hallucinations with Explanations | Automatic headline generation enables users to comprehend ongoing news events promptly and has recently become an important task in web mining and natural language processing. With the growing need for news headline generation, we argue that the hallucination issue, namely the generated headlines being not supported by the original news stories, is a critical challenge for the deployment of this feature in web-scale systems Meanwhile, due to the infrequency of hallucination cases and the requirement of careful reading for raters to reach the correct consensus, it is difficult to acquire a large dataset for training a model to detect such hallucinations through human curation. In this work, we present a new framework named ExHalder to address this challenge for headline hallucination detection. ExHalder adapts the knowledge from public natural language inference datasets into the news domain and learns to generate natural language sentences to explain the hallucination detection results. To evaluate the model performance, we carefully collect a dataset with more than six thousand labeled <article, headline> pairs. Extensive experiments on this dataset and another six public ones demonstrate that ExHalder can identify hallucinated headlines accurately and justifies its predictions with human-readable natural language explanations. | ['Marc Najork', 'Michael Bendersky', 'Negar Rahmati', 'Dan Finnie', 'Jialu Liu', 'Jiaming Shen'] | 2023-02-12 | null | null | null | null | ['headline-generation'] | ['natural-language-processing'] | [ 1.83058120e-02 4.71394569e-01 -1.22294098e-01 -3.46826881e-01
-9.01618600e-01 -3.93708438e-01 8.03174615e-01 3.29142541e-01
-5.17894253e-02 9.45348680e-01 1.03649747e+00 -6.10715300e-02
1.64696276e-01 -6.00008726e-01 -6.36117935e-01 5.98445162e-03
3.04629356e-01 5.44820011e-01 1.26989871e-01 -5.08127034e-01
4.24602807e-01 -8.64413381e-02 -1.48624671e+00 9.73421395e-01
1.03471339e+00 8.16218615e-01 2.82200485e-01 5.71945965e-01
-1.75250307e-01 1.32520056e+00 -9.43847537e-01 -7.36886799e-01
-3.05506259e-01 -6.23379171e-01 -9.38693643e-01 3.68940830e-01
1.25517070e-01 -3.44853371e-01 -1.03657633e-01 1.00503671e+00
6.54905856e-01 -1.70926213e-01 5.51422119e-01 -1.30646229e+00
-9.15125072e-01 9.95883167e-01 -3.36529732e-01 3.15518886e-01
9.57662880e-01 3.74012478e-02 1.15729845e+00 -1.09552693e+00
9.43903685e-01 1.23902857e+00 5.28294921e-01 5.58643937e-01
-8.25491846e-01 -5.20711780e-01 -9.31351110e-02 5.31069219e-01
-1.06254971e+00 -4.09060091e-01 6.56419158e-01 -3.83398652e-01
1.00051606e+00 4.68567699e-01 4.65432793e-01 2.01032257e+00
6.74105287e-02 1.00286114e+00 9.39926565e-01 -2.17796534e-01
1.15435176e-01 6.43451333e-01 1.34982407e-01 4.61276531e-01
-3.11433541e-04 -3.25605631e-01 -9.05670881e-01 -2.68868268e-01
2.13663697e-01 -1.49164900e-01 -6.39528751e-01 4.45219010e-01
-1.37406862e+00 1.11175525e+00 3.14048231e-01 -1.39594555e-01
-5.34118772e-01 -6.51064277e-01 6.29227698e-01 2.21019000e-01
6.69009447e-01 1.02664244e+00 -1.58770129e-01 -1.45524755e-01
-6.81455970e-01 5.93826354e-01 1.21553731e+00 1.17129838e+00
7.21496865e-02 -2.89656788e-01 -4.15804595e-01 1.07612026e+00
-7.22547919e-02 4.74432051e-01 9.87710416e-01 -5.21095932e-01
5.98702490e-01 6.49817109e-01 1.69173881e-01 -1.36052501e+00
-2.42244855e-01 -7.53618419e-01 -7.57387400e-01 -5.23259640e-01
3.74031067e-02 -2.36177519e-01 -1.67940214e-01 1.41658223e+00
1.20815691e-02 -1.60288960e-01 2.22617164e-01 8.86712074e-01
1.03345239e+00 9.24651623e-01 -2.75915563e-01 -5.22319615e-01
1.55190396e+00 -1.00709617e+00 -9.04017091e-01 -5.29900908e-01
2.93207258e-01 -9.68832254e-01 1.44416690e+00 5.88174224e-01
-7.43394434e-01 -1.94880947e-01 -1.05496323e+00 -4.35813479e-02
-2.51942396e-01 -1.38410240e-01 3.00306946e-01 -1.13469452e-01
-3.76146942e-01 1.31975457e-01 -6.68958202e-02 -7.68387735e-01
1.19714633e-01 -5.50034523e-01 -2.99630821e-01 -1.09678926e-03
-1.56532800e+00 9.97252285e-01 5.08102417e-01 -3.48549336e-01
-8.45830739e-01 -4.97874290e-01 -6.52009428e-01 -1.08347433e-02
4.38087255e-01 -5.50252020e-01 1.59632206e+00 -5.60963511e-01
-8.99379253e-01 6.70813739e-01 -1.99536547e-01 -7.61516929e-01
9.19547319e-01 -4.18944061e-01 -8.38126242e-01 6.89820200e-02
5.93764722e-01 3.88351768e-01 8.13919008e-01 -1.13632214e+00
-4.42285269e-01 -2.03987330e-01 -2.10890308e-01 1.74781084e-01
-6.28098011e-01 1.98698476e-01 5.61629683e-02 -6.59112990e-01
-3.85310687e-02 -5.45599043e-01 1.01674780e-01 -4.69440162e-01
-1.21437299e+00 -4.90274310e-01 7.42252707e-01 -7.48091459e-01
1.33550966e+00 -2.02209306e+00 -3.31407964e-01 -2.57283837e-01
5.09803116e-01 -1.02398813e-01 1.07954435e-01 7.41018891e-01
1.21853143e-01 5.53655513e-02 6.40203711e-03 -1.84198305e-01
-2.17824113e-02 1.07507855e-01 -1.11284292e+00 -1.92792211e-02
1.14460848e-01 6.40281916e-01 -1.16524565e+00 -5.15371263e-01
-3.29578042e-01 1.25493482e-01 -5.45078874e-01 5.82093537e-01
-4.43841130e-01 1.36621416e-01 -4.52638716e-01 4.56965625e-01
2.71499634e-01 -6.41868949e-01 -3.32005203e-01 8.91557559e-02
-1.57703221e-01 6.73706114e-01 -6.46296382e-01 1.02127469e+00
-2.98734218e-01 7.85674036e-01 -6.17177665e-01 -3.40232879e-01
1.10974586e+00 6.64513946e-01 -1.39071375e-01 -7.35125482e-01
-5.95408864e-02 1.77728295e-01 -3.96042377e-01 -1.14771986e+00
5.96837699e-01 -3.45260561e-01 -3.20077628e-01 6.17449701e-01
-1.35957882e-01 -6.99311644e-02 1.55459866e-01 6.25834346e-01
1.14723933e+00 -7.27947891e-01 5.98597765e-01 1.99499726e-01
1.48580641e-01 3.37490559e-01 4.42554653e-01 1.04174352e+00
6.62203059e-02 8.85239899e-01 6.01470292e-01 -4.40596223e-01
-1.10983920e+00 -8.80046010e-01 -4.27967124e-02 1.11477757e+00
2.58199368e-02 -7.28633642e-01 -6.19970322e-01 -7.08933830e-01
-2.21293226e-01 1.53263140e+00 -5.24542928e-01 -1.63242415e-01
-3.86537574e-02 -6.20147824e-01 7.07256317e-01 1.96692139e-01
6.48439825e-01 -1.46972978e+00 -5.16614139e-01 4.22215372e-01
-9.53648269e-01 -1.30177999e+00 -3.58055085e-01 -1.08535782e-01
-2.08299607e-01 -9.64404464e-01 -4.90643084e-01 -6.16085410e-01
4.39815730e-01 3.40877146e-01 1.12415338e+00 -4.42409515e-03
-2.11805161e-02 -1.95425406e-01 -7.76845038e-01 -7.20164120e-01
-8.17953289e-01 6.62229881e-02 1.35026261e-01 -9.92375836e-02
3.88394535e-01 -4.32401121e-01 -3.32796037e-01 2.42869437e-01
-9.93392229e-01 5.47401309e-01 4.03536886e-01 8.14485788e-01
-4.99961618e-03 2.47069169e-02 9.64041770e-01 -1.18493700e+00
1.55325854e+00 -1.06609857e+00 2.41187453e-01 1.96886927e-01
-6.00879788e-01 -3.42281498e-02 9.36460137e-01 -4.71696466e-01
-1.05806935e+00 -2.91366518e-01 -2.85273492e-01 4.40656096e-02
-1.76468015e-01 8.44484448e-01 1.09892666e-01 9.00297821e-01
1.15033340e+00 5.66705167e-01 -1.57299027e-01 -5.23591876e-01
2.73133039e-01 1.11148131e+00 8.39141011e-01 8.61986191e-05
6.50670826e-01 2.80120313e-01 -1.02781260e+00 -8.81711960e-01
-1.66406274e+00 -4.73126441e-01 -7.61435833e-03 -3.92942190e-01
6.99368596e-01 -8.33708286e-01 -3.15005243e-01 3.32929264e-03
-1.77677906e+00 4.61445123e-01 1.14354029e-01 1.93393171e-01
-5.41054666e-01 2.11601973e-01 -4.71137971e-01 -6.79799318e-01
-5.55184007e-01 -7.57297456e-01 7.21113801e-01 -1.00370180e-02
-1.04947174e+00 -6.85651839e-01 1.62931934e-01 7.04949796e-01
3.33816707e-01 1.80673480e-01 1.05801809e+00 -1.36790490e+00
5.19869179e-02 -5.08774340e-01 -2.19352558e-01 6.19425215e-02
-2.22026691e-01 -2.96511233e-01 -9.72137809e-01 1.66526690e-01
3.30860645e-01 -9.06344354e-01 4.96312141e-01 -3.10762018e-01
1.09364069e+00 -1.23468339e+00 2.43328288e-02 8.30012485e-02
8.03754210e-01 -2.76241988e-01 6.01255298e-01 5.05766213e-01
3.09887946e-01 6.95531905e-01 4.85836208e-01 1.03544831e+00
6.13278866e-01 4.54604089e-01 3.70091289e-01 1.11291192e-01
2.64482033e-02 -1.11156666e+00 5.36002338e-01 9.52513039e-01
4.35442895e-01 -6.46112919e-01 -6.74369633e-01 4.31818277e-01
-1.84854341e+00 -1.43542957e+00 -2.43813261e-01 1.74259305e+00
1.04567170e+00 3.27884048e-01 -8.20494294e-02 1.23595499e-01
8.46103251e-01 2.16536522e-01 -5.32737970e-01 -3.76680166e-01
-3.69018137e-01 -6.32764399e-01 -5.59890196e-02 4.21682149e-01
-6.25481665e-01 7.85237730e-01 6.03143501e+00 5.97532928e-01
-1.07053173e+00 1.38156593e-01 4.66244966e-01 1.04722500e-01
-5.32217622e-01 -2.06760392e-01 -8.55834663e-01 6.37597263e-01
7.32248664e-01 -7.19601631e-01 2.20929429e-01 1.18621171e+00
5.54618001e-01 3.20392489e-01 -1.06748939e+00 9.38403964e-01
7.10819244e-01 -1.40480793e+00 5.12502909e-01 -3.69801074e-01
5.48381329e-01 3.07953339e-02 -8.49380344e-02 5.40104389e-01
9.84575003e-02 -1.03056633e+00 9.42327797e-01 2.74452537e-01
4.35529709e-01 -4.14600164e-01 8.17950785e-01 1.10308886e+00
-3.04609179e-01 -1.23819143e-01 -4.09209967e-01 -3.52378488e-01
5.31488717e-01 7.80201375e-01 -1.70087290e+00 -1.78787798e-01
4.33186650e-01 6.79352045e-01 -7.74721384e-01 1.01791418e+00
-6.39220417e-01 5.66380084e-01 9.38132927e-02 -5.39914191e-01
6.61019385e-02 7.02870429e-01 9.12789345e-01 1.32996762e+00
4.61576849e-01 1.71257034e-02 1.50379211e-01 1.07514691e+00
-2.65206546e-01 1.72965601e-01 -7.12922573e-01 -2.08213538e-01
6.30406618e-01 1.18957710e+00 -2.91661888e-01 -4.22935694e-01
-2.47852132e-01 1.14344585e+00 4.22758967e-01 1.46655336e-01
-9.14146066e-01 -3.89159709e-01 1.08626867e-02 5.23346245e-01
-4.37210947e-01 3.28980178e-01 -2.71673292e-01 -1.47676146e+00
-1.14327446e-02 -1.23076010e+00 4.41325575e-01 -1.37224090e+00
-1.72514331e+00 1.19866621e+00 -2.18870655e-01 -1.28549492e+00
-6.78215981e-01 -1.11685745e-01 -8.42835307e-01 4.26038712e-01
-1.27932584e+00 -7.31276631e-01 -5.64690232e-01 6.23006344e-01
1.25353575e+00 -1.81299970e-01 8.55920494e-01 -1.01961240e-01
-4.88646090e-01 3.67503345e-01 -4.34374154e-01 5.89306764e-02
8.83178532e-01 -1.08951592e+00 3.48075897e-01 7.20158637e-01
7.99061060e-02 6.95390344e-01 1.56533182e+00 -9.07871246e-01
-9.30660546e-01 -1.03785312e+00 1.63096869e+00 -4.98622388e-01
1.01806462e+00 -3.98989260e-01 -1.08525169e+00 6.24055386e-01
4.97453421e-01 -5.09834647e-01 8.67819667e-01 -1.68070581e-03
-5.37606597e-01 4.66017306e-01 -8.59596312e-01 7.11916447e-01
8.59032273e-01 -5.47840595e-01 -1.23984575e+00 1.04218400e+00
8.82983625e-01 -1.42435208e-01 -2.33085811e-01 1.55481964e-01
2.63054192e-01 -1.00909746e+00 4.69158202e-01 -9.12147462e-01
1.31596291e+00 -9.68775302e-02 -2.87941541e-03 -1.56954753e+00
-9.39958170e-02 -6.52899444e-01 -3.32723595e-02 1.27547622e+00
9.42447603e-01 -3.92309159e-01 3.25386882e-01 6.15508318e-01
-1.85641974e-01 -5.30570328e-01 -4.54817712e-01 -6.27623558e-01
-4.59312946e-01 -4.36649531e-01 4.29861486e-01 1.11577761e+00
8.21248770e-01 9.69062030e-01 -8.00196350e-01 1.46346733e-01
4.06863302e-01 6.61685765e-02 7.76685655e-01 -1.03248382e+00
-4.04824704e-01 -1.21469647e-01 -4.30489592e-02 -1.15371108e+00
9.57530141e-02 -9.10240471e-01 2.83941686e-01 -1.58618903e+00
7.78928459e-01 2.64142901e-01 1.96115911e-01 3.42689276e-01
-1.73195541e-01 4.16544974e-02 -3.68441492e-02 3.87600929e-01
-9.50949967e-01 7.48701930e-01 1.23339033e+00 2.31355671e-02
2.51415670e-02 7.79504627e-02 -1.20568812e+00 9.88294005e-01
6.09460294e-01 -5.37952423e-01 -5.19139588e-01 -3.36557031e-01
7.32221425e-01 3.52899879e-01 4.55621898e-01 -8.15174878e-01
4.03564513e-01 -9.19254646e-02 3.07776988e-01 -5.92291534e-01
8.41585472e-02 -2.54127443e-01 -1.90947801e-01 2.04183757e-01
-1.12585473e+00 4.62959498e-01 -3.94386619e-01 6.01258695e-01
-4.31670338e-01 -2.64921755e-01 4.98210281e-01 -3.48709166e-01
-5.88353395e-01 4.41644434e-03 -5.74774683e-01 5.24260223e-01
8.33632529e-01 8.98613408e-02 -7.77893126e-01 -1.15621841e+00
-5.79430163e-01 1.64872423e-01 1.80009410e-01 8.77276897e-01
1.07031631e+00 -1.09529066e+00 -1.20095587e+00 7.64136463e-02
8.28313529e-01 -3.54155988e-01 3.00108735e-02 6.38647497e-01
-2.67227739e-01 2.18665719e-01 5.62499836e-02 -1.52686104e-01
-9.73915815e-01 5.08706868e-01 -1.49541199e-01 -5.76146059e-02
-9.64092433e-01 7.54022598e-01 3.89186218e-02 -1.70799792e-01
2.75267571e-01 -6.59971610e-02 -4.22426522e-01 2.40600742e-02
1.26876891e+00 1.26382887e-01 -8.18438455e-02 -4.43807989e-01
5.08932695e-02 -5.70064902e-01 -5.02459407e-01 -1.94420621e-01
1.28461576e+00 -3.35214376e-01 -4.71323654e-02 5.89633346e-01
1.13234365e+00 1.85761139e-01 -7.82732904e-01 -1.57338247e-01
2.50347286e-01 -3.50839347e-01 -2.91987658e-01 -1.13283169e+00
-3.51189256e-01 5.17083883e-01 -2.83207506e-01 8.30774367e-01
5.35261095e-01 4.30516988e-01 1.33369219e+00 5.07233739e-01
2.00577930e-01 -9.61507320e-01 4.83651429e-01 8.83098006e-01
1.65472889e+00 -1.30214262e+00 -2.75901198e-01 -3.84755820e-01
-1.28949749e+00 1.02605128e+00 7.14153588e-01 1.52818650e-01
1.07371710e-01 5.83976321e-02 2.18772098e-01 -3.26825559e-01
-1.21332252e+00 2.75995910e-01 2.02992037e-01 1.84585735e-01
5.01651287e-01 9.33306217e-02 -3.27438682e-01 1.04858601e+00
-9.72042561e-01 -2.90766716e-01 9.16799426e-01 3.77425969e-01
-7.50283003e-01 -4.23787683e-01 -2.98220187e-01 5.09854555e-01
-3.06490123e-01 -1.25881165e-01 -9.08309221e-01 4.29093271e-01
-2.14110315e-01 1.19568920e+00 -2.24906281e-01 -5.00431001e-01
3.84144336e-01 6.12259768e-02 -3.07379186e-01 -8.43350649e-01
-3.74303430e-01 -3.20436895e-01 4.39483762e-01 -3.49634051e-01
2.12957934e-01 -3.53741556e-01 -1.21396065e+00 -3.08301687e-01
-9.89393219e-02 3.67657423e-01 3.86772573e-01 1.18673313e+00
3.63563955e-01 1.19328506e-01 7.05486894e-01 -1.93857729e-01
-9.51523602e-01 -1.23386550e+00 -4.72622633e-01 9.88891602e-01
2.22756207e-01 -3.16670746e-01 -6.18688285e-01 5.79590201e-02] | [12.237536430358887, 9.315686225891113] |
0a5518f3-d3f3-4f4e-8c96-824509b4e50d | scaling-speech-technology-to-1000-languages | null | null | https://research.facebook.com/publications/scaling-speech-technology-to-1000-languages/ | https://research.facebook.com/micro_site/url/?click_from_context_menu=true&country=US&destination=https%3A%2F%2Fresearch.facebook.com%2Ffile%2F6486163864750413%2FScaling-Speech-Technology-to-1%2C000%2B-Languages.pdf&event_type=click&last_nav_impression_id=09L70osMgoDOIV1W3&max_percent_page_viewed=71&max_viewport_height_px=969&max_viewport_width_px=1367&orig_http_referrer=https%3A%2F%2Ft.co%2F&orig_request_uri=https%3A%2F%2Fresearch.facebook.com%2Fpublications%2Fscaling-speech-technology-to-1000-languages%2F®ion=noam&scrolled=true&session_id=0fZiQ44z0KJcVr8QE&site=mc_research | Scaling Speech Technology to 1,000+ Languages | Expanding the language coverage of speech technology has the potential to improve access to information for many more people. However, current speech technology is restricted to about one hundred languages which is a small fraction of the over 7,000 languages spoken around the world. The Massively Multilingual Speech (MMS) project increases the number of supported languages by 10-40x, depending on the task. The main ingredients are a new dataset based on readings of publicly available religious texts and effectively leveraging self-supervised learning. We built pre-trained wav2vec 2.0 models covering 1,406 languages, a single multilingual automatic speech recognition model for 1,107 languages, speech synthesis models for the same number of languages, as well as a language identification model for 4,017 languages. Experiments show that our multilingual speech recognition model more than halves the word error rate of Whisper on 54 languages of the FLEURS benchmark while being trained on a small fraction of the labeled data. The MMS models are available at https://github.com/pytorch/fairseq/tree/master/examples/mms. | ['Michael Auli', 'Alexis Conneau', 'Wei-Ning Hsu', 'Xiaohui Zhang', 'Yossi Adi', 'Alexei Baevski', 'Maryam Fazel-Zarandi', 'Apoorv Vyas', 'Zhaoheng Ni', 'Ali Elkahky', 'Sayani Kundu', 'Arun Babu', 'Paden Tomasello', 'Bowen Shi', 'Andros Tjandra', 'Vineel Pratap'] | 2023-05-23 | null | null | null | arxiv-2023-5 | ['speech-synthesis'] | ['speech'] | [-4.98242825e-01 1.18107021e-01 -3.93799633e-01 -3.89031291e-01
-1.38148916e+00 -7.15821624e-01 7.52986848e-01 -5.27949594e-02
-6.22684300e-01 6.40827298e-01 9.22380745e-01 -7.31779099e-01
7.87507296e-01 -4.52437699e-01 -5.65143645e-01 -2.60493785e-01
1.37350783e-01 7.56999552e-01 2.09490433e-02 -5.98681629e-01
-3.26996237e-01 1.13306552e-01 -1.08650303e+00 3.76494020e-01
7.90067375e-01 5.56455314e-01 2.73973346e-01 8.61352146e-01
-2.47025862e-01 8.72860134e-01 -5.17443419e-01 -4.96420860e-01
1.65242888e-02 -1.77272514e-01 -9.96634781e-01 -3.57883610e-02
4.36307222e-01 -2.67913163e-01 -6.06810510e-01 9.00566757e-01
7.38456845e-01 -2.15859879e-02 9.74621028e-02 -7.78683782e-01
-7.44439781e-01 1.32493424e+00 -2.05363140e-01 4.34082866e-01
3.17684084e-01 -9.48727597e-03 1.20057368e+00 -1.17043161e+00
5.52479744e-01 1.52753997e+00 4.41025794e-01 5.12751818e-01
-8.75019193e-01 -7.37345338e-01 -1.78236291e-01 1.19466498e-01
-1.56271529e+00 -1.19423485e+00 3.85233641e-01 -2.95201689e-01
1.44362795e+00 2.20856309e-01 4.54270601e-01 1.52177453e+00
-2.23176122e-01 8.93238187e-01 9.78313923e-01 -5.91340303e-01
-7.95566961e-02 4.37553912e-01 2.86725372e-01 7.49524176e-01
-1.30185485e-01 -7.97076970e-02 -7.03633666e-01 -2.89899945e-01
2.64042169e-01 -6.24390185e-01 -2.42146760e-01 2.54863709e-01
-1.49698663e+00 1.01887441e+00 -1.32117391e-01 6.27270639e-01
1.10842377e-01 -1.98964596e-01 7.16470778e-01 5.05414009e-01
7.42041171e-01 1.76380903e-01 -7.15525508e-01 -3.25963736e-01
-8.07074726e-01 -5.74555062e-02 9.47010696e-01 8.02105546e-01
6.02787852e-01 6.44610465e-01 1.25129193e-01 1.47298801e+00
2.76895076e-01 1.04163456e+00 8.66533518e-01 -6.30012572e-01
7.80359447e-01 3.16026896e-01 -3.37415278e-01 -4.77718174e-01
-3.50424916e-01 -3.86003524e-01 -5.64544439e-01 -4.64244157e-01
3.96183491e-01 -3.06156456e-01 -7.83758938e-01 1.72896910e+00
2.17616275e-01 -3.22919101e-01 2.59548277e-01 5.25418639e-01
9.44356978e-01 1.17168987e+00 -1.89612806e-01 5.52249104e-02
1.31125593e+00 -1.30041659e+00 -5.33351839e-01 -4.27829266e-01
7.91121125e-01 -1.11753106e+00 1.26557040e+00 2.31567219e-01
-1.18633616e+00 -4.20120209e-01 -8.13058197e-01 -3.14326108e-01
-5.14041960e-01 1.37435213e-01 5.21792233e-01 8.46163630e-01
-1.36006010e+00 -2.60093570e-01 -7.18297839e-01 -5.47711551e-01
2.88897753e-02 8.67142454e-02 -4.94968504e-01 -1.49974063e-01
-1.32706439e+00 9.71206844e-01 3.60962212e-01 -3.90273988e-01
-9.23877537e-01 -7.01191366e-01 -1.08309436e+00 -5.57841845e-02
1.07498802e-01 -1.75401360e-01 1.44581306e+00 -8.31947565e-01
-1.59167898e+00 1.04522145e+00 -2.38126531e-01 -4.95111227e-01
2.00330436e-01 1.50627485e-02 -1.01240551e+00 -2.65865266e-01
1.80086285e-01 6.14592314e-01 3.39408517e-01 -7.53219724e-01
-6.18740261e-01 -2.16931954e-01 -4.07155633e-01 2.38659948e-01
-5.74258983e-01 4.71379459e-01 -6.11623168e-01 -6.91416860e-01
-2.73679018e-01 -9.47940350e-01 1.66187286e-02 -6.95767522e-01
-4.47541416e-01 -2.72652477e-01 5.32037914e-01 -1.27280283e+00
1.16003406e+00 -2.16290569e+00 7.90560171e-02 -2.54275557e-02
-2.21612543e-01 2.77473420e-01 -4.20071989e-01 7.17581630e-01
9.48429853e-02 1.07697800e-01 -1.82537556e-01 -5.53945720e-01
-1.64756011e-02 2.99195319e-01 -5.14457405e-01 5.01785457e-01
-8.57128724e-02 7.09049225e-01 -8.08222651e-01 -6.86118146e-03
1.22086607e-01 6.55701220e-01 -4.39214319e-01 3.93122621e-02
3.14265490e-02 3.86076123e-01 8.23192373e-02 5.98312974e-01
4.07200366e-01 1.30629733e-01 2.38158524e-01 3.03129733e-01
-3.51972997e-01 1.10945523e+00 -8.57662141e-01 1.92163134e+00
-7.59808064e-01 6.85093462e-01 2.59989500e-01 -6.82201087e-01
8.74210417e-01 6.16632283e-01 2.03164220e-01 -6.71753109e-01
6.49017990e-02 6.58234656e-01 1.17499478e-01 -1.39692441e-01
4.69854683e-01 -6.69893026e-02 -3.29067081e-01 3.90771806e-01
5.00608444e-01 -2.02432245e-01 1.08576395e-01 2.06650391e-01
9.57755983e-01 -5.23556471e-01 4.60758805e-01 -4.47998375e-01
5.39835155e-01 1.81491543e-02 5.36155224e-01 4.02660519e-01
-8.68997127e-02 3.82445961e-01 -6.65000547e-03 -4.88497704e-01
-1.19219208e+00 -1.07404566e+00 -2.63920486e-01 1.44199789e+00
-5.95065653e-01 -4.72068548e-01 -7.27921486e-01 -3.14429373e-01
-3.12363543e-02 9.00799155e-01 -5.56937791e-02 1.02262773e-01
-5.24360836e-01 -6.03972971e-01 1.11329377e+00 1.52735472e-01
4.26263332e-01 -1.10245311e+00 3.71262908e-01 1.08617455e-01
-4.42083150e-01 -1.35076404e+00 -8.44629526e-01 -5.91327399e-02
-1.97404101e-01 -5.78069627e-01 -8.38075697e-01 -8.65222096e-01
-1.95806790e-02 2.27190450e-01 1.23039508e+00 -1.60713792e-01
-3.42629328e-02 2.39099041e-01 -2.92049438e-01 -2.59290516e-01
-1.05520833e+00 4.54467624e-01 5.05556583e-01 -2.72218101e-02
3.80065441e-01 -2.42045105e-01 9.63936821e-02 -1.77895069e-01
-2.89289057e-01 4.28929692e-03 2.06138968e-01 8.13647985e-01
1.59681112e-01 -4.46783781e-01 6.87457561e-01 -7.12001264e-01
3.26328486e-01 -7.15512812e-01 -5.75662792e-01 2.14969859e-01
-2.90266335e-01 2.72560250e-02 5.62403917e-01 -3.46665204e-01
-1.08484614e+00 -1.20670743e-01 -6.80857122e-01 -2.32194420e-02
-5.76207489e-02 6.32712185e-01 -6.32093176e-02 2.48632595e-01
5.19189000e-01 4.00537640e-01 -1.41762823e-01 -6.23608649e-01
6.69932544e-01 1.30521905e+00 5.28883398e-01 -4.37054515e-01
5.21545231e-01 5.20684719e-02 -9.55036581e-01 -1.55697477e+00
-7.14920104e-01 -6.83548987e-01 -4.14265901e-01 1.33392751e-01
7.35090673e-01 -1.42536449e+00 -3.52224380e-01 5.09912550e-01
-1.21279609e+00 -5.63154161e-01 -2.06441075e-01 7.22516656e-01
-1.45962849e-01 2.93066591e-01 -9.06741381e-01 -7.89322197e-01
-6.01546764e-01 -1.19333279e+00 1.00461495e+00 -2.81889260e-01
-3.17219347e-01 -1.03911400e+00 4.55712229e-01 7.95718968e-01
5.46427250e-01 -5.67583323e-01 8.26186717e-01 -1.05833209e+00
-1.54961497e-01 1.06639892e-01 2.29185626e-01 6.68518603e-01
1.28215179e-01 -2.37455711e-01 -1.09516239e+00 -5.26467860e-01
-2.71282881e-01 -6.21882141e-01 8.44641805e-01 1.62673786e-01
5.60535848e-01 -4.12716687e-01 9.04718488e-02 3.81088644e-01
8.85387778e-01 2.30582967e-01 3.25949490e-01 1.06314093e-01
8.26356173e-01 5.38394272e-01 -1.90060854e-01 3.06372851e-01
7.33433902e-01 6.01828516e-01 -1.12445265e-01 7.25751966e-02
-3.85970980e-01 -4.40067738e-01 8.69192362e-01 2.03830957e+00
4.73899275e-01 -2.85400063e-01 -1.47989750e+00 9.43731844e-01
-1.25506103e+00 -8.40790987e-01 4.98606302e-02 2.32875776e+00
1.19864452e+00 -1.09274641e-01 2.42345840e-01 -1.42408162e-01
6.18739843e-01 3.08531702e-01 -2.61152834e-01 -5.18696070e-01
-4.94613916e-01 2.26650640e-01 6.40617251e-01 1.17163634e+00
-9.48359430e-01 1.51736772e+00 6.29348230e+00 1.01993084e+00
-1.32923853e+00 3.85937303e-01 5.95488727e-01 -1.12049773e-01
-6.41835690e-01 -1.55369461e-01 -1.13183701e+00 4.92920995e-01
1.54123878e+00 -4.12348717e-01 8.53055835e-01 7.63258755e-01
2.19411001e-01 4.14980352e-01 -5.52792966e-01 9.76275265e-01
3.69722873e-01 -1.36374021e+00 2.21274063e-01 7.74008408e-02
8.00218344e-01 1.12619007e+00 -2.39252988e-02 6.45487189e-01
7.79115438e-01 -1.00400746e+00 9.47558343e-01 4.50979359e-02
9.73414183e-01 -8.50911558e-01 3.73915642e-01 5.29629767e-01
-1.00888526e+00 6.80324361e-02 -2.85552412e-01 1.74007952e-01
1.60285205e-01 4.40332025e-01 -1.17129457e+00 3.15430880e-01
7.88168967e-01 4.97201592e-01 -4.56274003e-01 4.53672051e-01
-1.19495608e-01 1.33839071e+00 -4.75975722e-01 -1.03161195e-02
3.64865899e-01 1.49912965e-02 6.83182478e-01 1.49302793e+00
4.10564214e-01 -2.40258977e-01 2.83662468e-01 2.70100027e-01
-4.04329240e-01 5.83542585e-01 -7.52541840e-01 -3.72188151e-01
8.59245420e-01 1.05681181e+00 -1.54775739e-01 -5.46559036e-01
-8.66221070e-01 8.16007674e-01 5.02696216e-01 2.19555512e-01
-4.20869350e-01 -9.16970447e-02 7.65455246e-01 -2.37892680e-02
-5.00170030e-02 -4.74208862e-01 5.20769954e-02 -1.41709197e+00
-1.35334581e-01 -1.50514102e+00 3.01912814e-01 -5.97789407e-01
-1.27520466e+00 8.18203449e-01 -5.26999295e-01 -6.27048731e-01
-5.09910643e-01 -6.00497603e-01 -3.24633271e-01 1.13444018e+00
-1.28220463e+00 -1.29679978e+00 2.52952009e-01 5.44139743e-01
1.09987843e+00 -9.25051093e-01 1.10167027e+00 5.56621790e-01
-5.87503910e-01 7.06853628e-01 3.04302394e-01 4.65316176e-01
8.09720457e-01 -1.08141327e+00 1.00317943e+00 6.38468027e-01
6.12929821e-01 5.62914431e-01 3.38563323e-01 -4.07431781e-01
-1.41331470e+00 -9.81762290e-01 1.59289956e+00 -4.08457369e-01
1.18431473e+00 -7.67046094e-01 -9.58597839e-01 9.82465684e-01
7.67038405e-01 -1.85191318e-01 7.71061778e-01 3.46522748e-01
-6.24653280e-01 -9.69946105e-03 -6.67139590e-01 7.22180188e-01
7.35062540e-01 -9.97294664e-01 -4.26536798e-01 5.14943719e-01
1.09899426e+00 -3.31291646e-01 -8.37053776e-01 -6.83403835e-02
4.43014652e-01 -4.22654569e-01 7.76757419e-01 -6.40546262e-01
3.97820882e-02 2.64340863e-02 -7.05889702e-01 -1.59766793e+00
-6.98617324e-02 -7.44473219e-01 1.92301825e-01 1.44441199e+00
8.22892070e-01 -7.75047660e-01 2.11431786e-01 -7.40065426e-02
-4.51477766e-01 -4.31005210e-01 -1.28887224e+00 -9.31225836e-01
4.98461276e-01 -7.33944774e-01 5.07933080e-01 1.29945171e+00
7.98620135e-02 7.51071632e-01 -4.06230390e-01 1.47873431e-01
4.38710392e-01 -5.01734018e-01 6.38817728e-01 -7.46546745e-01
-5.19951344e-01 -4.17292237e-01 -2.13737324e-01 -9.44682658e-01
5.35808563e-01 -1.52091801e+00 -2.68139333e-01 -1.30978894e+00
7.61521161e-02 -3.44432712e-01 2.92960480e-02 7.63955653e-01
2.32044950e-01 1.31271213e-01 5.26032038e-02 -1.71889160e-02
-1.53647974e-01 6.19248152e-01 7.12042451e-01 -4.26056564e-01
-1.20975278e-01 -2.12980568e-01 -6.29850626e-01 5.96777320e-01
9.15188491e-01 -1.39050111e-01 -2.20616251e-01 -8.78058910e-01
-3.98569591e-02 7.78831393e-02 -1.14214495e-02 -8.10506225e-01
1.78916797e-01 1.67108685e-01 -1.23943612e-01 -4.71874148e-01
3.96505535e-01 -2.53098845e-01 -1.43657282e-01 3.89496684e-01
-4.05997664e-01 6.74450919e-02 4.33678031e-01 -3.29161026e-02
-3.01236480e-01 1.58884805e-02 7.45019913e-01 -1.60191238e-01
-7.33483136e-01 2.32180104e-01 -6.99829638e-01 4.24738646e-01
4.33139682e-01 4.46779400e-01 -4.99220043e-01 -5.29225767e-01
-4.63216245e-01 1.11327395e-01 5.99804744e-02 9.81984675e-01
7.85903111e-02 -1.34082818e+00 -1.20536065e+00 4.15176958e-01
2.69721866e-01 -4.42131519e-01 7.14682415e-02 4.60427731e-01
-3.82402331e-01 7.44122684e-01 2.95102090e-01 -3.77758533e-01
-1.13837934e+00 2.05220491e-01 3.05896252e-01 -4.39491607e-02
-4.84388918e-01 9.56998765e-01 -1.12672299e-01 -1.09601378e+00
1.85560808e-01 -2.07003623e-01 4.27253172e-02 -6.26355410e-02
6.60992265e-01 4.38892394e-01 2.96326101e-01 -1.21839392e+00
-4.98560011e-01 3.18329111e-02 -2.06516191e-01 -7.49043584e-01
1.21813238e+00 -1.69072837e-01 -1.23839878e-01 8.49130571e-01
1.36440480e+00 6.81327641e-01 -5.66085219e-01 -4.84921932e-01
7.88517892e-02 1.30718887e-01 1.90032706e-01 -8.24107587e-01
-8.82417202e-01 8.63385797e-01 2.72557735e-01 3.12949866e-02
5.58251321e-01 2.63460338e-01 1.03884113e+00 4.41229820e-01
5.50774455e-01 -1.17616200e+00 -2.97464907e-01 1.06099272e+00
9.77107525e-01 -1.31015348e+00 -5.35225868e-01 -1.16500951e-01
-8.13357711e-01 7.22332001e-01 8.60853270e-02 2.00174063e-01
6.92952752e-01 5.70755899e-01 5.19546747e-01 9.94633958e-02
-8.83333981e-01 -1.29611939e-01 2.57641077e-01 3.67497146e-01
9.82082486e-01 4.81906146e-01 -1.30826190e-01 5.33362567e-01
-8.60123515e-01 -6.18666410e-01 4.00649428e-01 3.83733928e-01
-6.31632090e-01 -1.26460135e+00 -3.18274200e-01 1.21601112e-01
-5.41869462e-01 -6.59850240e-01 -2.68732637e-01 5.56016624e-01
-2.70796776e-01 1.23914826e+00 6.13663048e-02 -3.09662133e-01
2.42845133e-01 3.93416137e-01 6.96671158e-02 -8.11880410e-01
-4.03471351e-01 2.06021056e-01 4.98922884e-01 -3.45970392e-01
1.03136688e-01 -8.65012288e-01 -1.06976199e+00 -5.45577526e-01
8.70340690e-02 1.94868892e-01 8.30630422e-01 7.92041600e-01
1.96494326e-01 1.10796228e-01 7.16416717e-01 -5.52436888e-01
-6.28831565e-01 -1.19428456e+00 -5.17159522e-01 -1.08661361e-01
2.92390943e-01 -3.67582291e-02 -4.42810297e-01 1.18191494e-02] | [14.278338432312012, 6.939574241638184] |
8f5355c8-54d1-49e0-925a-fea5e0a4c46d | heater-an-efficient-and-unified-network-for | 2205.15448 | null | https://arxiv.org/abs/2205.15448v3 | https://arxiv.org/pdf/2205.15448v3.pdf | FeatER: An Efficient Network for Human Reconstruction via Feature Map-Based TransformER | Recently, vision transformers have shown great success in a set of human reconstruction tasks such as 2D human pose estimation (2D HPE), 3D human pose estimation (3D HPE), and human mesh reconstruction (HMR) tasks. In these tasks, feature map representations of the human structural information are often extracted first from the image by a CNN (such as HRNet), and then further processed by transformer to predict the heatmaps (encodes each joint's location into a feature map with a Gaussian distribution) for HPE or HMR. However, existing transformer architectures are not able to process these feature map inputs directly, forcing an unnatural flattening of the location-sensitive human structural information. Furthermore, much of the performance benefit in recent HPE and HMR methods has come at the cost of ever-increasing computation and memory needs. Therefore, to simultaneously address these problems, we propose FeatER, a novel transformer design that preserves the inherent structure of feature map representations when modeling attention while reducing memory and computational costs. Taking advantage of FeatER, we build an efficient network for a set of human reconstruction tasks including 2D HPE, 3D HPE, and HMR. A feature map reconstruction module is applied to improve the performance of the estimated human pose and mesh. Extensive experiments demonstrate the effectiveness of FeatER on various human pose and mesh datasets. For instance, FeatER outperforms the SOTA method MeshGraphormer by requiring 5% of Params and 16% of MACs on Human3.6M and 3DPW datasets. The project webpage is https://zczcwh.github.io/feater_page/. | ['Chen Chen', 'Guo-Jun Qi', 'Taojiannan Yang', 'Matias Mendieta', 'Ce Zheng'] | 2022-05-30 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zheng_FeatER_An_Efficient_Network_for_Human_Reconstruction_via_Feature_Map-Based_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zheng_FeatER_An_Efficient_Network_for_Human_Reconstruction_via_Feature_Map-Based_CVPR_2023_paper.pdf | cvpr-2023-1 | ['2d-human-pose-estimation'] | ['computer-vision'] | [-1.80354431e-01 2.12075874e-01 1.52093619e-01 -2.88420141e-01
-5.91907561e-01 -2.03603152e-02 2.83991724e-01 -2.27290913e-01
-3.97735864e-01 3.36263150e-01 2.92475641e-01 1.45703107e-01
9.28289816e-02 -9.63311613e-01 -9.65352833e-01 -1.74482584e-01
7.96505809e-02 8.53327751e-01 3.51289511e-01 -3.45362753e-01
-1.15126409e-01 5.48059404e-01 -1.68360543e+00 4.50390987e-02
4.64406013e-01 1.26595473e+00 2.86709487e-01 4.83357787e-01
3.24806780e-01 4.75017220e-01 -3.21214467e-01 -4.40988719e-01
3.81058007e-01 -8.08052868e-02 -7.99668849e-01 -1.96080022e-02
4.08859074e-01 -5.09250283e-01 -7.23516941e-01 7.24292338e-01
7.97028840e-01 1.60967067e-01 4.56985593e-01 -1.19850659e+00
-4.34757560e-01 2.14516297e-01 -6.75573826e-01 -2.57733107e-01
5.13318300e-01 2.63911724e-01 7.41377831e-01 -1.37723851e+00
7.13850558e-01 1.42760742e+00 9.84885871e-01 4.71269071e-01
-9.72086370e-01 -6.71701133e-01 -2.55102992e-01 1.54019549e-01
-1.60878897e+00 -2.68004894e-01 7.28182912e-01 -3.93547177e-01
1.02146351e+00 1.97752193e-01 1.09063637e+00 1.15539658e+00
3.28073412e-01 8.92815828e-01 8.29952002e-01 -6.04810752e-02
-3.87943648e-02 -2.18722016e-01 -1.46167442e-01 1.14943314e+00
-9.83725563e-02 1.87037095e-01 -6.98052466e-01 5.29764500e-03
1.23794985e+00 1.46921650e-01 -3.36379141e-01 -5.40392816e-01
-1.25062656e+00 5.62942564e-01 8.28180373e-01 -2.61470407e-01
-5.28872609e-01 3.30006599e-01 3.37497771e-01 -1.87859649e-03
3.85651857e-01 3.10921729e-01 -3.95873129e-01 -1.34201959e-01
-6.11278236e-01 5.96001923e-01 5.53197682e-01 1.06700349e+00
7.52985537e-01 -2.28762165e-01 -1.04392178e-01 8.87107730e-01
2.03907520e-01 7.89542675e-01 2.60988086e-01 -1.02905416e+00
4.26995516e-01 6.91133559e-01 8.46433714e-02 -1.39359927e+00
-6.92601800e-01 -2.32035205e-01 -1.00593901e+00 6.02481216e-02
2.96253890e-01 2.52093256e-01 -1.20440722e+00 1.66958320e+00
4.79725569e-01 -1.37893647e-01 -5.33812642e-01 1.19251978e+00
8.75652850e-01 6.69722438e-01 -1.14157340e-02 3.97861332e-01
1.51585579e+00 -1.01549375e+00 -4.32347208e-01 -2.14609310e-01
1.27700493e-01 -5.70988476e-01 1.21055853e+00 2.70563453e-01
-1.23820353e+00 -7.64718831e-01 -8.71231496e-01 -5.61021268e-01
-1.60054326e-01 1.97970286e-01 6.10493362e-01 4.96637970e-02
-8.73486400e-01 7.21128047e-01 -9.99398530e-01 -3.23629528e-01
4.60567236e-01 4.37936157e-01 -6.73636794e-01 -1.42940238e-01
-1.27550054e+00 9.57815707e-01 8.54084566e-02 4.53920215e-01
-7.27409780e-01 -7.29911804e-01 -1.10503161e+00 2.63946448e-02
3.65687609e-01 -1.00605249e+00 1.06180227e+00 -4.97572005e-01
-1.45790541e+00 9.37754035e-01 -2.98151113e-02 -3.05700302e-01
8.15788507e-01 -6.44010961e-01 -1.82958506e-02 2.03621522e-01
1.41724244e-01 1.01705742e+00 7.94743359e-01 -9.31759536e-01
-3.71730208e-01 -5.73849022e-01 -2.07915381e-02 3.24826717e-01
-7.49707967e-02 -2.48905823e-01 -7.86290348e-01 -6.66480422e-01
1.72255933e-01 -1.05392003e+00 -1.89482182e-01 4.86497790e-01
-5.42198539e-01 -3.14562917e-01 6.75605178e-01 -9.15458560e-01
9.87558544e-01 -1.98465192e+00 5.83359003e-01 2.13327348e-01
4.49969292e-01 7.75819570e-02 -2.84291916e-02 1.90096200e-01
1.50719106e-01 -2.04220325e-01 -1.75280467e-01 -6.20640278e-01
9.98238549e-02 2.28379592e-01 -9.81715843e-02 5.73129833e-01
1.26353845e-01 1.24303401e+00 -5.97496748e-01 -6.13692820e-01
4.26343262e-01 8.56535017e-01 -6.81005776e-01 3.46464872e-01
-1.92980058e-02 3.97690564e-01 -2.79117256e-01 7.47254670e-01
4.95357454e-01 -4.46782053e-01 -5.75960949e-02 -6.11186326e-01
5.04315235e-02 8.71055722e-02 -1.01574409e+00 1.92282975e+00
-4.35693920e-01 2.69095898e-01 -4.46867757e-02 -5.75398147e-01
7.19568610e-01 2.37631783e-01 5.98844171e-01 -7.89507031e-01
3.22574675e-01 8.57810527e-02 -4.03385311e-01 -2.33128741e-01
5.06833851e-01 3.38925309e-02 -1.49551079e-01 2.20155865e-01
1.42235318e-02 -3.16792205e-02 -2.56761611e-01 -6.69274554e-02
1.04064012e+00 3.96193147e-01 2.90008634e-01 -7.49240965e-02
1.34778634e-01 -1.22809060e-01 5.85755527e-01 2.62203246e-01
-1.69203743e-01 9.52504754e-01 1.57420665e-01 -7.71646321e-01
-1.22200692e+00 -1.35527635e+00 -5.09612337e-02 8.25447559e-01
3.24566573e-01 -7.17050970e-01 -7.80278385e-01 -4.89020407e-01
1.76646903e-01 1.18770443e-01 -7.45933592e-01 -2.38789722e-01
-9.97344434e-01 -2.64841020e-01 4.57614690e-01 8.26862395e-01
7.30757356e-01 -1.24894226e+00 -1.10761309e+00 8.56568888e-02
-4.37035352e-01 -9.52454507e-01 -6.65149868e-01 -7.45838061e-02
-5.17434120e-01 -1.06723344e+00 -9.43967581e-01 -6.77237093e-01
6.43019736e-01 -1.19993892e-02 1.11204028e+00 9.45286378e-02
-6.00212514e-01 3.08891147e-01 -1.64677963e-01 -1.76831543e-01
1.87220931e-01 1.53064653e-01 2.15206966e-01 -2.37365067e-01
7.67071769e-02 -6.30093455e-01 -8.48536134e-01 5.84939718e-01
-4.69960898e-01 3.80137980e-01 6.05278373e-01 9.25665617e-01
8.14962029e-01 -1.24429516e-01 1.59597054e-01 -6.13507152e-01
2.47703999e-01 -2.71211267e-01 -3.26928526e-01 1.05952710e-01
-3.13571632e-01 -8.42418671e-02 4.37360168e-01 -4.62585926e-01
-7.69647658e-01 2.58347631e-01 -3.63955319e-01 -8.87937188e-01
2.15174872e-02 3.19107741e-01 -2.13629842e-01 9.33156237e-02
5.34530401e-01 1.10482827e-01 1.94403082e-01 -6.29008830e-01
1.77535906e-01 3.64326000e-01 7.65912950e-01 -5.38462639e-01
8.74336720e-01 3.67692590e-01 -7.75340060e-03 -6.86918318e-01
-8.18104804e-01 -1.31139398e-01 -5.73653400e-01 -2.81209797e-01
1.01802278e+00 -9.62877214e-01 -8.70483577e-01 6.27217472e-01
-1.07485795e+00 -2.61362106e-01 -3.61734986e-01 2.89754212e-01
-8.28723907e-01 2.38526672e-01 -7.88471222e-01 -4.38645631e-01
-5.81613898e-01 -1.29123485e+00 1.43629992e+00 -1.31501968e-03
-4.63023722e-01 -4.47159469e-01 -2.09736288e-01 4.37837541e-01
2.61212349e-01 4.23203170e-01 7.69310653e-01 3.15936864e-03
-5.65243781e-01 -3.48430127e-01 -2.89545149e-01 5.20633534e-03
-1.38184309e-01 -3.74140561e-01 -7.38484263e-01 -3.45218867e-01
-2.78619051e-01 -4.59225863e-01 5.28499961e-01 5.16012311e-01
1.30591083e+00 -1.78263456e-01 -3.64366293e-01 8.49813104e-01
9.91388202e-01 -2.71014273e-01 7.61328876e-01 1.71360329e-01
1.05873179e+00 5.16328931e-01 7.03344822e-01 5.41094124e-01
6.70347631e-01 1.06433570e+00 4.07067657e-01 -3.08203965e-01
-3.62458408e-01 -7.43935406e-01 1.31991968e-01 6.96232021e-01
-3.24496269e-01 4.84918542e-02 -9.66217637e-01 3.02947968e-01
-1.82557118e+00 -6.43310666e-01 1.85232982e-01 2.09346962e+00
6.81384683e-01 9.25779715e-02 2.45905831e-01 -2.11699829e-02
6.00552678e-01 4.22153808e-02 -5.93446136e-01 1.60201326e-01
1.62287846e-01 3.26756865e-01 3.00763905e-01 2.04551086e-01
-1.05937505e+00 9.01196361e-01 5.48729324e+00 6.64821267e-01
-9.61027205e-01 -2.76539922e-02 3.85254353e-01 -8.87576416e-02
7.41807371e-02 -3.45054418e-01 -6.47493660e-01 3.61338228e-01
5.17339110e-01 1.08103462e-01 4.38413858e-01 9.68674779e-01
-6.64740354e-02 2.50230059e-02 -1.17327690e+00 1.43534088e+00
-5.06177917e-02 -1.08647645e+00 1.06793098e-01 1.67397901e-01
1.59586176e-01 -9.34844166e-02 6.16298318e-02 3.79037410e-01
1.37114068e-02 -1.24398077e+00 9.56718385e-01 6.50955319e-01
1.01628578e+00 -9.11221623e-01 5.94689488e-01 3.02762479e-01
-1.51456451e+00 4.80410419e-02 -4.36322242e-01 1.01264408e-02
4.27643001e-01 4.07421410e-01 -4.80912060e-01 4.28564072e-01
1.04822206e+00 7.22823560e-01 -5.64675629e-01 8.08716536e-01
-1.97438389e-01 5.22599518e-02 -4.99161243e-01 3.14352214e-01
-1.71091825e-01 1.61384031e-01 5.23312628e-01 8.25606167e-01
2.99888134e-01 2.14963347e-01 3.72583300e-01 9.21771646e-01
-1.22980140e-01 -8.42526853e-02 -4.64782625e-01 2.88757563e-01
5.10966122e-01 1.15741491e+00 -5.99846125e-01 -2.00653031e-01
-1.83590241e-02 1.16155922e+00 5.51448166e-01 3.82438302e-02
-1.14278877e+00 -3.14024568e-01 6.61954582e-01 6.28434360e-01
2.46988103e-01 -1.85901597e-01 -1.31137028e-01 -1.03749871e+00
1.89436942e-01 -8.35073650e-01 1.90152720e-01 -8.62566829e-01
-1.16137779e+00 5.93736351e-01 7.28092045e-02 -1.19588041e+00
-1.74237579e-01 -5.00653505e-01 -1.80661619e-01 6.82651222e-01
-8.95690739e-01 -1.35412776e+00 -5.03412485e-01 6.96819484e-01
5.87833583e-01 1.54730871e-01 7.13459551e-01 4.56915259e-01
-5.07376611e-01 6.80123985e-01 -7.66801775e-01 3.10137630e-01
5.17879188e-01 -1.06023228e+00 7.38143861e-01 4.31587845e-01
-1.09169073e-01 6.15836501e-01 5.60128689e-01 -6.77786767e-01
-1.61010337e+00 -1.07888436e+00 7.24035382e-01 -6.52027726e-01
1.76308796e-01 -6.01396143e-01 -6.73182189e-01 7.13189185e-01
-3.40637922e-01 2.87722439e-01 1.46872848e-01 2.07922775e-02
-4.00229961e-01 5.61640076e-02 -1.10427558e+00 6.67496502e-01
1.47689092e+00 -5.04353940e-01 -5.13500333e-01 1.78708926e-01
7.06934631e-01 -7.95488715e-01 -1.23626578e+00 6.77985847e-01
8.82893980e-01 -8.91940057e-01 1.27493739e+00 -2.61269540e-01
6.39262140e-01 -2.73282081e-01 -2.55608141e-01 -1.05362546e+00
-4.57077652e-01 -2.34344900e-01 -3.78885448e-01 6.43544257e-01
7.90940300e-02 -3.88216496e-01 8.00932407e-01 6.20889127e-01
-2.11714968e-01 -1.19396269e+00 -9.80470240e-01 -4.87845808e-01
-1.88300684e-01 -1.97410822e-01 5.56141496e-01 5.81318438e-01
-1.39077812e-01 2.95811564e-01 -6.90025389e-01 5.61374538e-02
7.56259918e-01 9.15101022e-02 1.05699408e+00 -1.23388898e+00
-2.73275912e-01 -1.36671782e-01 -5.92375457e-01 -1.20132518e+00
7.35641718e-02 -7.57424951e-01 1.75399080e-01 -1.48138535e+00
1.82486147e-01 -2.74521768e-01 8.53218213e-02 6.30894363e-01
2.57619079e-02 4.28933948e-01 3.52410883e-01 2.25860193e-01
-5.43180943e-01 7.49718308e-01 1.45181024e+00 -1.25505803e-02
2.60314438e-02 -1.22768998e-01 -1.83320642e-01 8.54633987e-01
4.99800533e-01 -2.84180284e-01 -3.60730678e-01 -5.00516534e-01
9.03873518e-02 2.29487464e-01 9.15415406e-01 -1.27902424e+00
1.93030566e-01 1.76450476e-01 8.49890888e-01 -8.63969386e-01
7.29249477e-01 -7.65101254e-01 5.76637506e-01 5.73905170e-01
-3.67471054e-02 4.05639380e-01 -8.17176476e-02 4.31792855e-01
-4.93968576e-02 4.95878100e-01 9.36725676e-01 -3.05373967e-01
-7.42556155e-01 6.78991914e-01 6.12103529e-02 1.30547816e-03
8.64366293e-01 -3.44995767e-01 -5.70106460e-03 -2.39007756e-01
-7.70422816e-01 1.72633171e-01 6.35556340e-01 5.93832970e-01
1.08524489e+00 -1.46846509e+00 -4.06757206e-01 4.29202348e-01
2.58598253e-02 4.27641422e-01 4.97162342e-01 9.00181770e-01
-6.14352286e-01 1.64982259e-01 -4.62852508e-01 -6.67591751e-01
-1.09627116e+00 5.04442334e-01 5.14339626e-01 -2.49013379e-01
-1.29722857e+00 7.98316002e-01 3.43207687e-01 -6.34606004e-01
2.96637416e-01 -1.86999738e-01 1.21318020e-01 -1.98711991e-01
3.81335706e-01 5.69450676e-01 -5.22367395e-02 -8.24629068e-01
-3.64999771e-01 9.11806464e-01 3.35033685e-02 9.42462608e-02
1.34669220e+00 -1.02914423e-02 2.36199033e-02 1.59709826e-01
1.23889899e+00 -2.76867896e-01 -1.39041495e+00 -1.41969249e-01
-3.66043478e-01 -4.64502096e-01 -1.82704464e-01 -6.00742161e-01
-1.25967145e+00 7.79693604e-01 6.40424669e-01 -4.36334878e-01
1.03464782e+00 1.03325993e-01 1.27675915e+00 1.39758214e-01
9.39297915e-01 -1.01082623e+00 2.11680859e-01 4.46744919e-01
1.23284626e+00 -1.00083792e+00 6.10317327e-02 -4.77653414e-01
-5.31205654e-01 8.13221157e-01 9.17485774e-01 -3.20698291e-01
7.98177302e-01 3.16474140e-01 -1.59972057e-01 -5.95607102e-01
-4.48525101e-01 -4.05647121e-02 3.95129919e-01 5.19113958e-01
2.01478377e-01 1.53953716e-01 -3.15747410e-02 6.19904220e-01
-5.70448339e-01 3.81872145e-04 -8.28179568e-02 8.83019686e-01
-3.54038417e-01 -6.51369214e-01 -2.19591483e-01 3.58003736e-01
-1.32228568e-01 1.80559650e-01 -2.60185987e-01 9.37472939e-01
1.22278444e-01 2.90735960e-01 5.96648976e-02 -7.38858759e-01
7.61812687e-01 -1.63282171e-01 5.52278280e-01 -4.64726895e-01
-5.66314936e-01 -2.35844925e-02 -3.04221008e-02 -1.13505197e+00
-5.67990467e-02 -2.75641799e-01 -1.42120981e+00 -5.23091674e-01
4.56461869e-03 -2.22588405e-01 3.96959960e-01 6.63465261e-01
4.31254655e-01 4.38279599e-01 2.16928780e-01 -1.33983684e+00
-4.49630171e-01 -8.49497199e-01 -4.89982605e-01 6.30656719e-01
-1.36897257e-02 -1.21896708e+00 4.78030853e-02 -2.18049809e-01] | [7.081048488616943, -1.0579566955566406] |
2519f1b7-2d38-4e86-968f-611a8b0e46c0 | when-to-read-documents-or-qa-history-on | 2306.04176 | null | https://arxiv.org/abs/2306.04176v1 | https://arxiv.org/pdf/2306.04176v1.pdf | When to Read Documents or QA History: On Unified and Selective Open-domain QA | This paper studies the problem of open-domain question answering, with the aim of answering a diverse range of questions leveraging knowledge resources. Two types of sources, QA-pair and document corpora, have been actively leveraged with the following complementary strength. The former is highly precise when the paraphrase of given question $q$ was seen and answered during training, often posed as a retrieval problem, while the latter generalizes better for unseen questions. A natural follow-up is thus leveraging both models, while a naive pipelining or integration approaches have failed to bring additional gains over either model alone. Our distinction is interpreting the problem as calibration, which estimates the confidence of predicted answers as an indicator to decide when to use a document or QA-pair corpus. The effectiveness of our method was validated on widely adopted benchmarks such as Natural Questions and TriviaQA. | ['Moontae Lee', 'Seung-won Hwang', 'Sang-eun Han', 'Kyungjae Lee'] | 2023-06-07 | null | null | null | null | ['natural-questions', 'triviaqa', 'open-domain-question-answering'] | ['miscellaneous', 'miscellaneous', 'natural-language-processing'] | [ 6.41786158e-02 4.60062027e-01 8.84953421e-03 -3.86552334e-01
-1.41430616e+00 -9.49172378e-01 8.62594843e-01 3.81839365e-01
-4.79644179e-01 8.73405516e-01 3.11243355e-01 -4.97949302e-01
-4.87102211e-01 -8.26203287e-01 -5.08391142e-01 -3.01251948e-01
3.32657248e-01 8.93150091e-01 6.79904461e-01 -7.85649538e-01
5.89846313e-01 7.57071674e-02 -1.32401681e+00 3.85944158e-01
1.06045210e+00 1.14155757e+00 8.26013535e-02 6.68698609e-01
-6.82642639e-01 1.06238186e+00 -6.09661996e-01 -9.35828030e-01
1.39488339e-01 -2.65116334e-01 -1.49590266e+00 -3.93523127e-01
5.53081930e-01 -7.89070427e-02 -7.35948905e-02 8.39345992e-01
3.32205802e-01 3.22132036e-02 4.89464581e-01 -1.05051732e+00
-5.81300974e-01 2.22657874e-01 1.21296206e-02 5.18700719e-01
9.46694314e-01 2.64501601e-01 1.46063089e+00 -6.14520133e-01
6.46808863e-01 1.13006961e+00 4.37638402e-01 4.16646570e-01
-9.10419822e-01 -7.34308735e-02 -2.00723395e-01 3.30434978e-01
-9.18051302e-01 -4.14950669e-01 5.19915164e-01 -2.75006682e-01
9.30232465e-01 3.77981961e-01 -1.42807635e-02 8.53761077e-01
-4.88819145e-02 7.13933766e-01 1.18758690e+00 -7.16979921e-01
2.64835656e-01 4.47682023e-01 5.80326021e-01 5.02218246e-01
5.88510484e-02 -7.50553310e-02 -3.94646764e-01 -3.85678738e-01
1.97975099e-01 -3.29145104e-01 -5.81331313e-01 -7.72291198e-02
-7.53843546e-01 9.61537302e-01 4.09572244e-01 5.46842217e-01
-3.37885827e-01 -2.90473402e-01 3.29989552e-01 8.33631694e-01
3.67667079e-01 1.03021538e+00 -7.13709295e-01 -4.06677514e-01
-8.52859497e-01 6.16024315e-01 1.40088224e+00 6.02740467e-01
9.37865436e-01 -7.44643271e-01 -3.89763087e-01 8.66048396e-01
1.39117420e-01 2.75213510e-01 6.51453614e-01 -1.01852357e+00
7.89811134e-01 9.20296848e-01 4.21469420e-01 -7.51367807e-01
-1.28273264e-01 -1.62730366e-01 -6.35210890e-03 -2.89924711e-01
8.92381847e-01 -2.11677089e-01 -6.46708250e-01 1.54211152e+00
4.11407948e-01 -2.93463528e-01 2.53555059e-01 7.79783249e-01
8.26374829e-01 6.47955656e-01 -9.14147347e-02 6.79031014e-04
1.59764850e+00 -1.17123234e+00 -5.24663806e-01 -4.28218573e-01
5.64648449e-01 -8.27560842e-01 1.13764238e+00 3.27034265e-01
-1.00661290e+00 -4.27157134e-01 -8.96321893e-01 -3.16274464e-01
-5.63350558e-01 -3.45800936e-01 1.55639753e-01 6.44300997e-01
-1.03621423e+00 3.88478369e-01 -1.65759578e-01 -4.61540401e-01
2.15818845e-02 -8.60244036e-03 -1.41836464e-01 -5.27936280e-01
-1.38186443e+00 1.11034369e+00 6.69810772e-02 -2.43771732e-01
-5.08286178e-01 -5.69160044e-01 -4.66378778e-01 3.04482490e-01
6.41055346e-01 -7.91189194e-01 1.58246922e+00 -8.74125421e-01
-1.49832726e+00 9.68590796e-01 -1.40452608e-01 -7.23602891e-01
5.48803091e-01 -4.73362952e-01 -3.80675405e-01 3.80019873e-01
2.68124133e-01 5.08249879e-01 7.60745466e-01 -1.04109263e+00
-5.44251084e-01 -5.48326135e-01 6.93171501e-01 1.55736268e-01
-2.71695584e-01 -2.86287665e-02 -2.61114359e-01 -1.85482129e-01
-1.11708499e-01 -6.29484415e-01 -6.27105907e-02 -2.80125827e-01
-1.13735326e-01 -5.72242379e-01 6.38952672e-01 -7.80515254e-01
1.29249310e+00 -1.69631994e+00 -1.06091134e-01 3.79021862e-03
1.18205689e-01 3.78239274e-01 -3.16924274e-01 9.28492188e-01
1.94259897e-01 1.66181937e-01 -2.74968594e-01 2.55828321e-01
1.17280610e-01 1.87786043e-01 -6.78706586e-01 -2.37612247e-01
3.83372873e-01 1.05463302e+00 -9.10386145e-01 -4.89398211e-01
-4.55570042e-01 -1.92624241e-01 -4.16010827e-01 6.71648979e-01
-8.78166199e-01 1.43296108e-01 -6.97678030e-01 3.55260819e-01
3.36653799e-01 -5.01238406e-01 -4.22657356e-02 1.13040596e-01
2.46617794e-01 7.19348371e-01 -9.02964234e-01 1.65770388e+00
-6.41181529e-01 3.81389618e-01 -1.58782918e-02 -8.32922339e-01
9.57148492e-01 4.77144510e-01 -3.03127486e-02 -9.05557215e-01
-6.72925711e-02 4.30214971e-01 -7.65865818e-02 -1.09945989e+00
6.35753632e-01 -1.26958117e-01 1.38171557e-02 5.80815673e-01
2.92783439e-01 -3.60457212e-01 3.47325683e-01 3.68100733e-01
1.47472870e+00 2.18554452e-01 2.76105285e-01 -2.97381431e-01
8.21243882e-01 4.26511317e-01 7.09287962e-03 9.83071089e-01
-1.96072713e-01 5.41059911e-01 6.30025446e-01 -1.25550508e-01
-4.73370522e-01 -1.02668750e+00 2.77740043e-03 1.16763926e+00
8.27914029e-02 -3.57472986e-01 -6.77605450e-01 -1.15898597e+00
-6.73442781e-02 7.39367485e-01 -5.78992903e-01 1.09179448e-02
-5.33658803e-01 -1.04923010e-01 5.40590703e-01 2.70797282e-01
4.04725045e-01 -8.87596846e-01 -7.66959429e-01 1.64713204e-01
-6.54199779e-01 -1.02314007e+00 -1.42082959e-01 1.77717745e-01
-9.15856004e-01 -1.44592106e+00 -7.59840071e-01 -5.61245024e-01
2.86871076e-01 1.69942126e-01 1.72671139e+00 3.12204361e-01
1.68857560e-01 8.89094770e-01 -6.99012399e-01 -2.89843857e-01
-4.84158784e-01 2.30026633e-01 -5.06144583e-01 -6.23926744e-02
4.80193347e-01 -3.77805889e-01 -6.99191630e-01 3.08879882e-01
-1.02845204e+00 -5.52627504e-01 5.70501089e-01 9.22912240e-01
4.16303724e-02 -5.36751211e-01 9.83002603e-01 -1.23326671e+00
1.18488681e+00 -8.97600174e-01 -3.12033534e-01 6.93154156e-01
-6.34386778e-01 4.59887058e-01 6.03513002e-01 -1.42387390e-01
-1.27504456e+00 -6.59650564e-01 -5.19210637e-01 -8.45217034e-02
-3.22701633e-01 7.48716533e-01 -4.45805080e-02 3.80055159e-02
1.03482938e+00 -5.24718985e-02 -8.71303603e-02 -5.13374269e-01
5.38955390e-01 7.14069843e-01 5.60113013e-01 -8.84129465e-01
7.30093539e-01 1.00292638e-01 -5.24102092e-01 -4.44418877e-01
-1.13573277e+00 -9.22037482e-01 -2.30138287e-01 4.75897230e-02
7.49329031e-01 -4.02002215e-01 -4.47010547e-01 -1.68950975e-01
-1.19826388e+00 -3.71779948e-02 -4.44455624e-01 -2.78091747e-02
-4.57772374e-01 4.90646183e-01 -4.84400481e-01 -7.63665557e-01
-3.59685957e-01 -8.79348218e-01 9.44121897e-01 3.51717860e-01
-3.37002873e-01 -1.09153843e+00 5.51351368e-01 9.65920866e-01
6.41671598e-01 -4.25010696e-02 1.27020419e+00 -1.46179438e+00
-4.58804637e-01 -4.37835664e-01 -1.20546058e-01 2.58823484e-01
1.25616752e-02 -5.64559877e-01 -1.15960705e+00 -1.39959857e-01
4.02426809e-01 -9.19647992e-01 5.96590102e-01 -4.61398125e-01
6.80876017e-01 -3.46669972e-01 7.79611841e-02 -2.57602692e-01
1.42759383e+00 -9.20059439e-03 6.31427586e-01 3.73043507e-01
-1.74572155e-01 9.01506007e-01 6.65781558e-01 -4.51591462e-02
4.68467206e-01 5.54404497e-01 4.04290706e-01 5.61083257e-01
-3.70674441e-03 -3.58754367e-01 7.31897503e-02 7.16596305e-01
2.30895862e-01 -4.46882159e-01 -1.09494114e+00 6.87225342e-01
-1.73853195e+00 -7.96063423e-01 -7.79171959e-02 2.19714069e+00
8.58365297e-01 2.08530426e-01 7.65845552e-02 7.57206753e-02
2.02708244e-01 1.38843790e-01 -4.06935334e-01 -3.22330832e-01
1.39009699e-01 6.08667433e-01 -1.06824331e-01 4.69831020e-01
-4.51618463e-01 5.44456065e-01 6.15395641e+00 9.06912804e-01
-7.09998310e-01 1.82755336e-01 5.48016548e-01 4.63367552e-01
-5.05304277e-01 4.03785318e-01 -5.98821461e-01 2.79398650e-01
1.23367894e+00 -1.38046984e-02 1.55231759e-01 5.47455370e-01
-4.15870756e-01 -5.26409507e-01 -1.30229330e+00 4.67949182e-01
1.24838255e-01 -1.27644110e+00 1.79963678e-01 -3.12181532e-01
3.64113808e-01 -1.28247347e-02 -1.20246954e-01 7.34280765e-01
4.11843807e-01 -8.14747036e-01 4.44858760e-01 6.48473918e-01
3.79268050e-01 -2.39312559e-01 9.07883584e-01 9.03258145e-01
-6.54859185e-01 -1.35401905e-01 -6.57447502e-02 -1.77326486e-01
1.63591743e-01 2.12205887e-01 -1.02594745e+00 9.31870282e-01
5.98116815e-01 -7.69910291e-02 -9.85520363e-01 9.56430256e-01
-2.50455499e-01 9.09578860e-01 -2.82979727e-01 -1.91132262e-01
4.50352877e-01 -1.73603579e-01 4.54059422e-01 9.88917351e-01
1.70581508e-02 3.40748668e-01 -1.42139748e-01 8.18273187e-01
-6.88616484e-02 2.48341084e-01 -3.75794858e-01 -8.90088379e-02
3.63001168e-01 1.14439154e+00 -2.31816068e-01 -3.79487246e-01
-6.69664800e-01 7.99803317e-01 6.24928534e-01 3.46443176e-01
-4.49644536e-01 -4.32454973e-01 4.61010374e-02 2.41404116e-01
3.02961349e-01 1.44776404e-01 1.79342274e-02 -1.25211298e+00
2.68241614e-01 -1.22765684e+00 8.15630555e-01 -9.82862473e-01
-1.55671990e+00 7.30874300e-01 -8.26290548e-02 -9.61221337e-01
-6.30950868e-01 -6.07302666e-01 -6.44395411e-01 9.45631266e-01
-1.96312404e+00 -8.03540647e-01 -2.68938035e-01 5.27401507e-01
6.53217316e-01 1.12254351e-01 8.62879336e-01 1.86690852e-01
-1.53400853e-01 3.47959906e-01 -2.28394428e-03 -1.17247170e-02
7.96037614e-01 -1.43501401e+00 1.21931225e-01 6.97304726e-01
5.08878171e-01 5.87701559e-01 7.47168779e-01 -3.16818237e-01
-1.36679876e+00 -6.42847300e-01 1.15393281e+00 -9.61347222e-01
7.25819528e-01 -2.40131319e-02 -1.28443050e+00 4.50136453e-01
5.37340045e-01 -8.56508911e-02 6.49049997e-01 2.37523139e-01
-6.28560662e-01 -1.30435094e-01 -1.22042131e+00 2.15987250e-01
5.71641028e-01 -7.83828557e-01 -1.36439180e+00 3.68517756e-01
8.73058379e-01 -2.90349483e-01 -9.30999219e-01 2.80047357e-01
2.14160621e-01 -1.22409606e+00 8.10066938e-01 -9.22985315e-01
5.67550898e-01 1.89301055e-02 -3.37712973e-01 -1.09788692e+00
2.10189566e-01 -4.40000355e-01 -1.08699888e-01 1.39354062e+00
5.96671700e-01 -7.57260740e-01 7.02264488e-01 7.77535856e-01
9.93852168e-02 -8.17844927e-01 -9.68339205e-01 -5.62715888e-01
3.39952916e-01 -2.00092271e-01 5.36334634e-01 8.37059259e-01
1.15043767e-01 7.60832608e-01 1.67816728e-01 -8.73880684e-02
-7.09548518e-02 3.20071965e-01 9.44373906e-01 -1.06789327e+00
-4.96109366e-01 -2.47189298e-01 2.47190576e-02 -1.58848751e+00
2.34926809e-02 -7.70479560e-01 2.14251634e-02 -1.43945658e+00
-7.68785551e-02 -3.79466057e-01 5.03031490e-03 1.03624746e-01
-4.10934329e-01 -1.25846729e-01 2.15456262e-01 1.89136758e-01
-7.92995989e-01 3.51243466e-01 1.01701891e+00 -1.03175551e-01
6.93456680e-02 2.10469395e-01 -8.66441190e-01 5.99009573e-01
4.71377671e-01 -3.26211423e-01 -5.03592074e-01 -5.63200057e-01
5.47271550e-01 7.88495123e-01 3.69838893e-01 -8.77219796e-01
4.65877801e-01 5.70080727e-02 -3.72328669e-01 -1.58260718e-01
2.16407731e-01 -9.44628716e-01 -3.69103163e-01 2.59397805e-01
-5.95985711e-01 3.13077897e-01 -6.76986296e-03 8.46708119e-01
-7.02115536e-01 -7.59944916e-01 4.26048636e-01 -3.19689155e-01
-6.33521080e-01 -2.27888394e-02 -2.93952762e-03 9.41401720e-01
4.56987560e-01 -8.82214606e-02 -5.51243365e-01 -6.81841671e-01
-4.41924185e-01 5.49719334e-01 6.80328682e-02 5.09724379e-01
4.29018974e-01 -7.67617643e-01 -6.60505891e-01 -1.65952310e-01
4.19380218e-01 -1.32290721e-01 4.11163121e-02 6.94440126e-01
-3.83265525e-01 7.19857335e-01 1.85411036e-01 -5.73597372e-01
-6.58645213e-01 5.64531267e-01 3.81659061e-01 -9.76616442e-01
-2.47467980e-02 7.06859648e-01 -1.07231379e-01 -5.92697382e-01
1.68516889e-01 -3.01236123e-01 -4.11952883e-01 1.25259623e-01
3.80484134e-01 3.08640808e-01 4.86054689e-01 -1.59754530e-01
-8.35555866e-02 2.43951067e-01 -1.63223282e-01 -2.49954164e-01
1.03810704e+00 -2.50793695e-01 -8.37449133e-02 4.47751641e-01
9.97028708e-01 5.48084341e-02 -7.55898535e-01 -6.30857706e-01
6.65157735e-01 -3.32258075e-01 -4.72810715e-01 -1.26686060e+00
-4.59994584e-01 9.71074104e-01 3.46274406e-01 8.25951934e-01
8.86085689e-01 1.43916890e-01 1.05345106e+00 7.71758139e-01
4.56983268e-01 -7.32444048e-01 3.91499966e-01 6.28070891e-01
9.48973119e-01 -1.53569877e+00 -3.13803226e-01 -1.59223929e-01
-5.86045742e-01 1.13905323e+00 6.80953324e-01 -8.06794390e-02
3.91161472e-01 -2.33082592e-01 2.86861420e-01 -5.21544695e-01
-9.34334219e-01 -2.77460456e-01 2.44407222e-01 2.67830282e-01
3.36403072e-01 -4.65326577e-01 -4.85640138e-01 6.56290293e-01
-1.83395818e-01 2.46302388e-03 2.85620421e-01 1.11347544e+00
-5.70139825e-01 -1.05043185e+00 -3.72094452e-01 5.22511840e-01
-6.15745306e-01 -2.03886837e-01 -4.77227986e-01 6.78707361e-01
-2.24787131e-01 1.37656438e+00 -2.33506963e-01 2.76800692e-02
4.18525815e-01 5.81719398e-01 3.50472778e-01 -6.29026055e-01
-1.07785189e+00 -7.34544992e-01 4.22345459e-01 -3.67403895e-01
-4.79970425e-01 -2.54626542e-01 -8.32157075e-01 3.06664586e-01
-6.16063118e-01 9.14633691e-01 3.25462729e-01 1.29658747e+00
4.16821212e-01 -2.44183540e-02 4.67097878e-01 -4.27630655e-02
-1.28718424e+00 -9.65209424e-01 -4.55495603e-02 5.60077190e-01
4.71302509e-01 -4.21694398e-01 -4.88473117e-01 -1.07111290e-01] | [11.330581665039062, 8.018545150756836] |
3b9d4a1f-3772-4366-8f2f-2c88a45fb170 | sequential-latent-variable-models-for-few | null | null | https://openreview.net/forum?id=7C9aRX2nBf2 | https://openreview.net/pdf?id=7C9aRX2nBf2 | Sequential Latent Variable Models for Few-Shot High-Dimensional Time-Series Forecasting | Modern applications increasingly require learning and forecasting latent dynamics from high-dimensional time-series. Compared to univariate time-series forecasting, this adds a new challenge of reasoning about the latent dynamics of an unobserved abstract state. Sequential latent variable models (LVMs) present an attractive solution, although existing works either struggle with long-term forecasting or have difficulty learning across diverse dynamics. In this paper, we first present a conceptual framework of sequential LVMs to unify existing works, contrast their fundamental limitations, and identify an intuitive solution to long-term forecasting for diverse dynamics via meta-learning. We then present the first framework of few-shot forecasting for high-dimensional time-series: instead of learning a single dynamic function, we leverage data of diverse dynamics and learn to adapt latent dynamic functions to few-shot support series. This is realized via Bayesian meta-learning underpinned by: 1) a latent dynamic function conditioned on knowledge derived from few-shot support series, and 2) a meta-model that learns to extract such dynamic-specific knowledge via feed-forward embedding of support set. We compared the presented framework with a comprehensive set of baseline models trained 1) globally on the large meta-training set with diverse dynamics, and 2) individually on single dynamics, both with and without fine-tuning to k-shot support series used by the meta-models. We demonstrated that the presented framework is agnostic to the latent dynamic function of choice and, at meta-test time, is able to forecast for new dynamics given variable-shot of support series. | ['Linwei Wang', 'Zhiyuan Li', 'Ryan Missel', 'Xiajun Jiang'] | 2023-05-05 | null | null | null | iclr-2023-5 | ['univariate-time-series-forecasting'] | ['time-series'] | [-2.55242772e-02 -1.39360830e-01 -4.63871777e-01 -2.14863464e-01
-6.88681960e-01 -5.07740378e-01 1.25440347e+00 -1.18809320e-01
7.38108456e-02 5.65191031e-01 3.18047941e-01 -1.74786046e-01
-5.11759281e-01 -6.39809549e-01 -6.62682474e-01 -9.04691279e-01
-6.39906406e-01 8.18541229e-01 1.66974902e-01 -2.80987769e-01
1.36564597e-01 4.13464814e-01 -1.78500605e+00 2.16184035e-01
4.39890504e-01 9.09004688e-01 1.24984402e-02 9.42166269e-01
-2.24101722e-01 6.90471947e-01 -3.72009873e-01 1.91159502e-01
1.65366933e-01 -2.92846054e-01 -2.86620766e-01 5.05257323e-02
1.74363375e-01 -4.22151476e-01 -3.16319078e-01 3.39058518e-01
1.79413348e-01 3.39063346e-01 9.50666845e-01 -1.42992103e+00
-4.50304806e-01 5.61357200e-01 5.02927229e-02 4.80930746e-01
-3.34836193e-03 4.98350978e-01 7.38176942e-01 -8.54685843e-01
8.32850456e-01 1.22035825e+00 9.10597920e-01 5.86759865e-01
-1.63528121e+00 -4.17210400e-01 4.02775615e-01 4.45833467e-02
-7.71480858e-01 -4.81075674e-01 9.51655447e-01 -8.55247021e-01
1.25819457e+00 -1.36360005e-01 6.29639089e-01 1.70327497e+00
5.49574256e-01 5.71315467e-01 1.08875930e+00 -3.13630015e-01
5.88996649e-01 6.87662065e-02 3.10592353e-01 3.30811471e-01
-2.27183968e-01 6.07279122e-01 -5.21902859e-01 -5.11955678e-01
5.43194115e-01 5.44197857e-01 2.22817048e-01 -5.94978750e-01
-1.30700314e+00 9.20724213e-01 -2.01205805e-01 4.28854555e-01
-5.59880972e-01 3.56993407e-01 6.31643593e-01 6.90403104e-01
8.51506174e-01 1.34112030e-01 -8.98518205e-01 -3.21757704e-01
-1.25166333e+00 3.78722996e-01 1.03824687e+00 5.95137537e-01
7.49799311e-01 6.56576872e-01 -1.62239701e-01 2.66328037e-01
2.14064091e-01 6.24408185e-01 9.21829462e-01 -8.13356817e-01
2.24887833e-01 6.81570470e-02 4.17627782e-01 -7.56522775e-01
-3.46807629e-01 -5.19903362e-01 -8.41146827e-01 2.06500709e-01
2.64052153e-01 -2.52910227e-01 -9.96118307e-01 1.88626599e+00
2.74392128e-01 9.30854559e-01 2.03081608e-01 3.62918705e-01
2.44772673e-01 1.07637930e+00 -1.87706903e-01 -7.70801604e-01
6.77688837e-01 -9.71653461e-01 -7.13682950e-01 1.63253352e-01
5.13391614e-01 -2.39828378e-01 8.79675150e-01 2.67072707e-01
-9.84104216e-01 -8.69755387e-01 -8.89951050e-01 3.81315470e-01
-4.83817190e-01 -3.15170079e-01 3.66745949e-01 3.42452943e-01
-1.19126642e+00 9.28845584e-01 -1.35928881e+00 -2.96116799e-01
-2.90717006e-01 2.35277459e-01 -2.59520095e-02 4.24244463e-01
-1.37171972e+00 9.49468255e-01 2.63754845e-01 -1.97263718e-01
-1.44020128e+00 -9.78125989e-01 -6.28935099e-01 1.39253736e-01
3.35181862e-01 -8.64907146e-01 1.37438405e+00 -7.92352915e-01
-2.00885820e+00 1.72391236e-01 -3.57437819e-01 -7.43214130e-01
6.22269094e-01 -1.70387983e-01 -7.49801755e-01 -3.54291610e-02
-9.80680957e-02 8.34838226e-02 1.55669904e+00 -1.01979601e+00
-3.33490968e-01 -1.25284478e-01 -3.12162548e-01 -3.94365698e-01
-3.54645699e-01 -4.75597054e-01 2.83194214e-01 -6.00649714e-01
9.95441899e-03 -8.06294143e-01 -2.32957646e-01 -3.56671453e-01
1.45575702e-01 -2.82403857e-01 1.09384286e+00 -3.79618675e-01
1.38837504e+00 -1.87924635e+00 4.33262944e-01 4.79440242e-02
1.92921922e-01 1.15964726e-01 -1.30424619e-01 1.05443549e+00
-2.38100633e-01 -2.11528212e-01 -1.75554320e-01 -6.76373780e-01
2.17055008e-01 3.92993927e-01 -1.26294005e+00 3.23554784e-01
1.22262239e-01 8.98912251e-01 -1.06106377e+00 7.48487189e-02
5.96118271e-01 5.33142388e-01 -2.51146019e-01 1.97602406e-01
-6.23016596e-01 5.83235502e-01 -2.99761891e-01 3.31684202e-01
2.71157652e-01 -4.50876534e-01 1.60915896e-01 1.04692906e-01
-2.97091842e-01 -4.66186665e-02 -1.18683624e+00 1.44051611e+00
-6.72039747e-01 3.49200130e-01 -5.26240826e-01 -1.44619405e+00
1.18890357e+00 5.86507797e-01 9.20222342e-01 -3.02664310e-01
-1.85874000e-01 4.07085776e-01 -3.95816922e-01 -6.19767904e-01
1.95786700e-01 -5.48649311e-01 -1.08829118e-01 6.57253742e-01
5.57044446e-01 -2.47283671e-02 -5.88781349e-02 -6.65008947e-02
1.06350327e+00 4.42915857e-01 3.01719069e-01 -6.35676319e-03
3.32090825e-01 -1.54469430e-01 3.81953537e-01 8.11548412e-01
1.76786873e-02 5.85475601e-02 4.08484668e-01 -1.01378834e+00
-1.20620561e+00 -1.32484877e+00 -3.89463678e-02 1.09102845e+00
-4.34525758e-01 -5.23866534e-01 1.31986871e-01 -2.83186644e-01
2.02964604e-01 8.76250327e-01 -8.65631521e-01 -1.40788734e-01
-5.38318038e-01 -6.51865780e-01 1.09161817e-01 4.53447044e-01
-2.62386382e-01 -9.85878170e-01 -8.12503219e-01 7.05125213e-01
3.62535685e-01 -7.26229608e-01 -2.13867892e-02 4.12981898e-01
-1.35558069e+00 -4.83855724e-01 -8.16094995e-01 -2.02627778e-01
1.15395181e-01 -6.85639158e-02 1.09264159e+00 -5.05390584e-01
-1.24359280e-01 9.25025165e-01 -1.73886850e-01 -2.28249863e-01
-6.27279222e-01 1.58585072e-03 6.07464790e-01 1.45310670e-01
1.80895492e-01 -1.13130820e+00 -2.93061763e-01 -1.55515531e-02
-8.57305467e-01 -1.57460630e-01 2.39729419e-01 1.07406664e+00
4.07072842e-01 -5.79379685e-02 7.71317482e-01 -6.18016720e-01
3.96039039e-01 -1.04920542e+00 -8.11245978e-01 3.41936260e-01
-8.61809194e-01 2.97107160e-01 8.46883476e-01 -9.27400708e-01
-8.55642498e-01 -9.07622278e-02 2.50853151e-01 -1.03724027e+00
3.70596088e-02 6.36462212e-01 6.19251430e-01 5.98013103e-01
5.96531272e-01 5.51170886e-01 1.90111026e-01 -5.90531528e-01
6.48600638e-01 2.13865504e-01 3.75654459e-01 -6.65861189e-01
7.98122346e-01 7.40678072e-01 2.41931215e-01 -6.77597463e-01
-6.80958152e-01 -4.81272489e-01 -9.57675159e-01 -2.25862011e-01
4.83435392e-01 -9.41876471e-01 -5.85232973e-01 4.95958418e-01
-8.05856943e-01 -5.42043447e-01 -6.65147066e-01 6.55935526e-01
-1.06591880e+00 -2.54235268e-02 -6.55703425e-01 -1.15522671e+00
-5.36875688e-02 -7.49672055e-01 1.06421316e+00 -2.27186888e-01
-3.22699785e-01 -1.68775034e+00 8.27947080e-01 -4.13917571e-01
6.84996247e-01 5.66000819e-01 8.81838262e-01 -6.86592698e-01
-4.28128630e-01 -3.42469543e-01 4.85091537e-01 1.89881220e-01
1.61838949e-01 2.10960537e-01 -1.04699838e+00 -5.33471107e-01
2.27799833e-01 -2.82568872e-01 9.98882174e-01 7.14589179e-01
6.55937970e-01 -3.38562638e-01 -4.37719136e-01 5.86925268e-01
1.31716132e+00 -6.69052526e-02 6.47596717e-02 2.15216741e-01
3.38750839e-01 5.88261306e-01 2.35904664e-01 7.41034865e-01
2.65175134e-01 6.69375956e-01 1.75860852e-01 4.45709825e-01
2.34929442e-01 -3.86586070e-01 9.22724426e-01 1.20848954e+00
-7.54427537e-02 -2.35755608e-04 -9.72778797e-01 6.12448156e-01
-2.16718864e+00 -1.42336476e+00 1.31004497e-01 2.11390615e+00
5.57022929e-01 2.93944299e-01 3.49212945e-01 -7.35616386e-02
3.55059594e-01 3.95296216e-01 -9.98690784e-01 -2.43829846e-01
-3.68981324e-02 1.37748614e-01 1.59360338e-02 7.12697327e-01
-1.00409055e+00 7.50353396e-01 7.08339453e+00 5.33678710e-01
-1.51414216e+00 4.04337972e-01 2.85209775e-01 -2.89595246e-01
-3.98842424e-01 2.74384826e-01 -1.16047430e+00 6.24013126e-01
1.72832692e+00 -4.30948377e-01 2.77657956e-01 8.92235816e-01
2.40367979e-01 5.27856410e-01 -1.20113850e+00 5.94054759e-01
-3.92216630e-02 -1.65140760e+00 1.02061555e-01 -7.35064447e-02
1.05791032e+00 2.30747208e-01 3.35116088e-01 8.18445385e-01
3.76571268e-01 -7.53176272e-01 4.90279615e-01 1.19511306e+00
6.04808092e-01 -2.76188731e-01 3.76575112e-01 7.51527905e-01
-1.14388120e+00 -5.43685973e-01 -2.36494258e-01 -3.81622642e-01
4.75384891e-01 6.67071939e-01 -5.27109861e-01 4.22352076e-01
4.52254921e-01 1.21968210e+00 -7.91998953e-02 5.05850792e-01
2.40908787e-01 8.93731534e-01 -4.95589226e-01 2.84844458e-01
2.62364417e-01 -6.23601228e-02 8.45549643e-01 9.33464289e-01
7.46267617e-01 -3.94393474e-01 4.45434362e-01 8.30088973e-01
6.51446223e-01 -3.86024535e-01 -8.12685609e-01 -3.37763965e-01
3.04977417e-01 8.83695483e-01 -4.85135496e-01 -7.36084342e-01
-3.30749303e-01 7.41933167e-01 1.60215437e-01 6.65980518e-01
-6.96919620e-01 1.77755490e-01 4.68631744e-01 1.07314184e-01
6.52986586e-01 -5.80273926e-01 9.65667367e-02 -1.52439737e+00
-3.15266430e-01 -4.99595761e-01 4.63689208e-01 -6.28129005e-01
-1.62392199e+00 5.92548966e-01 5.76935470e-01 -1.50532389e+00
-1.28687036e+00 -4.44262862e-01 -8.32240701e-01 9.62646365e-01
-1.39399064e+00 -1.38570368e+00 1.73589125e-01 4.78447467e-01
7.98217893e-01 -5.31258762e-01 1.04477847e+00 -1.78198174e-01
-1.81256518e-01 1.16297089e-01 7.29574680e-01 -6.71486437e-01
5.47861397e-01 -1.10756242e+00 5.78971803e-01 5.05740523e-01
1.02780230e-01 6.97592854e-01 1.09590530e+00 -8.25350821e-01
-1.41240656e+00 -8.98105145e-01 8.99078488e-01 -8.55891883e-01
1.23760808e+00 -4.20715719e-01 -1.10476637e+00 8.92288864e-01
-2.37946138e-01 1.54194206e-01 5.78710079e-01 2.71122605e-01
-3.15259218e-01 -2.60302145e-02 -7.16993749e-01 3.13941598e-01
7.49614477e-01 -7.05392003e-01 -9.10562456e-01 2.62467712e-01
8.87552738e-01 -9.68500972e-02 -9.20026302e-01 4.49636102e-01
7.33211160e-01 -7.64964819e-01 1.12640834e+00 -1.11605108e+00
3.60679507e-01 -1.18843161e-01 -3.45222801e-01 -1.42002094e+00
-5.46628475e-01 -8.82396758e-01 -1.20797217e+00 9.31975424e-01
5.28512411e-02 -7.74149418e-01 4.19909596e-01 2.60652781e-01
-1.19323984e-01 -8.89883101e-01 -9.84322369e-01 -1.13823092e+00
2.03321934e-01 -4.60039377e-01 5.58629632e-01 9.36915398e-01
-1.74246758e-01 1.91487819e-02 -8.84063601e-01 1.64058194e-01
5.62222064e-01 4.93975908e-01 8.99970472e-01 -1.42295277e+00
-7.39057243e-01 -3.36389452e-01 -2.57688731e-01 -8.75032961e-01
3.55288118e-01 -6.81320548e-01 -2.43968442e-01 -1.06049252e+00
-1.12828068e-01 -1.08020864e-01 -6.41391039e-01 1.66313753e-01
1.34850368e-01 -1.95629343e-01 8.22843909e-02 7.08653092e-01
-3.44146043e-01 9.29620326e-01 7.27088094e-01 2.37669833e-02
-4.99602348e-01 3.48674387e-01 1.70846328e-01 5.61341524e-01
5.33601522e-01 -4.16799873e-01 -7.55697548e-01 1.41034141e-01
1.99527130e-01 5.75153828e-01 6.26274109e-01 -8.89026999e-01
2.14507788e-01 -3.39599073e-01 3.21807116e-01 -8.51909995e-01
6.55469775e-01 -6.93998098e-01 2.93653488e-01 6.62791133e-01
-3.24993312e-01 3.13707381e-01 2.84063339e-01 1.05536175e+00
-8.02609548e-02 9.99468844e-03 2.92341471e-01 -1.32148132e-01
-9.07964766e-01 4.32125896e-01 -5.57039618e-01 -2.02534661e-01
9.52212393e-01 -2.49119580e-01 -6.39286861e-02 -6.18869543e-01
-1.38958144e+00 7.58625418e-02 2.33247161e-01 5.93525767e-01
3.87129754e-01 -1.31131053e+00 -5.11508644e-01 4.63899255e-01
6.22051097e-02 -6.66467130e-01 5.01667261e-01 8.25587332e-01
2.87582427e-01 4.02254134e-01 -3.08351457e-01 -9.71204400e-01
-6.07967973e-01 8.81909668e-01 2.19978452e-01 -4.81095761e-01
-9.48583364e-01 4.25596833e-01 -1.07592940e-01 -4.52680349e-01
6.21221140e-02 -4.89139050e-01 5.06836176e-03 4.53845620e-01
5.24101794e-01 3.30428988e-01 -3.28811020e-01 -5.06999314e-01
-4.14448865e-02 7.00353384e-01 1.34163439e-01 -4.23065126e-01
1.73832881e+00 -2.68076986e-01 2.69267499e-01 1.81884956e+00
1.14250576e+00 -4.71686661e-01 -1.79921639e+00 -4.74534899e-01
2.74425060e-01 -2.04908177e-01 -8.77864063e-02 -5.12175858e-01
-3.64023298e-01 1.09720683e+00 6.57564461e-01 3.32369685e-01
6.60780728e-01 -1.69426501e-01 7.08602667e-01 4.42504019e-01
3.94357413e-01 -8.82232368e-01 5.13756812e-01 7.74504364e-01
7.05153942e-01 -1.17701757e+00 -1.40762165e-01 5.02270758e-01
-4.21571195e-01 1.51357329e+00 1.24870934e-01 -4.62396383e-01
1.34208834e+00 2.15736508e-01 -8.33127275e-03 -1.72223315e-01
-1.62928069e+00 6.44843280e-02 4.60948318e-01 4.68856573e-01
1.90663710e-01 -1.82996858e-02 3.90854359e-01 3.86428058e-01
-2.11666501e-03 3.92784864e-01 1.86844513e-01 8.87561560e-01
-4.27428126e-01 -9.91376102e-01 -2.22500488e-01 4.94093627e-01
1.28228575e-01 2.37428248e-01 3.50594044e-01 6.69275641e-01
-4.22166958e-02 5.32838881e-01 3.07638824e-01 -3.32695007e-01
-1.19687296e-01 7.90827870e-01 2.64139771e-01 -6.73355460e-01
-2.79953271e-01 1.30344331e-01 -4.73933905e-01 -4.27452296e-01
-4.36087668e-01 -8.99686277e-01 -7.36073673e-01 -2.16922730e-01
-5.55113032e-02 -9.23817009e-02 6.41726375e-01 1.15849721e+00
4.60164189e-01 4.97873515e-01 7.87469983e-01 -1.34747076e+00
-1.21637881e+00 -1.04011536e+00 -5.74276507e-01 3.23359966e-01
8.85600030e-01 -9.36636806e-01 -7.52186000e-01 3.51800978e-01] | [6.952516078948975, 3.300468683242798] |
2201cd36-b6bf-4898-a9be-0b0940b8c30f | words-hk-a-comprehensive-cantonese-dictionary | null | null | https://aclanthology.org/2022.dclrl-1.7 | https://aclanthology.org/2022.dclrl-1.7.pdf | Words.hk: A Comprehensive Cantonese Dictionary Dataset with Definitions, Translations and Transliterated Examples | This paper discusses the compilation of the words.hk Cantonese dictionary dataset, which was compiled through manual annotation over a period of 7 years. Cantonese is a low-resource language with limited tagged or manually checked resources, especially at the sentential level, and this dataset is an attempt to fill the gap. The dataset contains over 53,000 entries of Cantonese words, which comes with basic lexical information (Jyutping phonemic transcription, part-of-speech tags, usage tags), manually crafted definitions in Written Cantonese, English translations, and Cantonese examples with English translation and Jyutping transliterations. Special attention has been paid to handle character variants, so that unintended “character errors” (equivalent to typos in phonemic writing systems) are filtered out, and intra-speaker variants are handled. Fine details on word segmentation, character variant handling, definition crafting will be discussed. The dataset can be used in a wide range of natural language processing tasks, such as word segmentation, construction of semantic web and training of models for Cantonese transliteration. | ['Lilian Suet-ying Chan', 'Raymond Ka-wai Tse', 'Grace Wing-yan Chan', 'Chaak-ming Lau'] | null | null | null | null | dclrl-lrec-2022-6 | ['transliteration'] | ['natural-language-processing'] | [ 2.32781976e-01 -1.51794270e-01 -4.56231326e-01 -4.17797565e-01
-5.95360875e-01 -8.54047835e-01 4.51463193e-01 9.02010575e-02
-8.39102209e-01 1.03953803e+00 2.90386856e-01 -7.60322988e-01
8.52920339e-02 -4.10601735e-01 -1.83631733e-01 -3.62189561e-01
5.69326282e-01 1.02657461e+00 1.69149816e-01 -5.41800737e-01
3.32314968e-01 1.31391019e-01 -1.07007539e+00 2.50235915e-01
8.98925364e-01 5.97780049e-01 3.68698627e-01 4.84324604e-01
-4.65311855e-01 4.48961586e-01 -1.10640013e+00 -9.53035414e-01
1.14396416e-01 -7.79455006e-01 -1.00798535e+00 3.65902126e-01
4.33571041e-01 3.65046144e-01 2.63653576e-01 1.24764311e+00
3.54945391e-01 -1.93028405e-01 5.68105280e-01 -7.83856571e-01
-7.17118680e-01 1.19965291e+00 1.16080195e-01 3.30778569e-01
3.49251419e-01 1.63022417e-03 1.15776420e+00 -1.21986854e+00
1.01604939e+00 1.30541992e+00 5.91252267e-01 5.69177926e-01
-1.02346122e+00 -6.94132507e-01 -2.12270960e-01 9.43868309e-02
-1.59709477e+00 -4.56367791e-01 3.02059233e-01 -6.34971857e-01
1.24098420e+00 4.79529917e-01 9.34101760e-01 1.03444910e+00
1.55202746e-01 6.40602529e-01 1.30331433e+00 -1.00667548e+00
-5.85493259e-03 5.75797200e-01 1.09650888e-01 3.17499340e-01
1.54003099e-01 -1.57678843e-01 -5.56664169e-01 1.10688552e-01
5.18094897e-01 -4.81323421e-01 -1.07928395e-01 4.81198341e-01
-1.64643717e+00 8.77521217e-01 -5.30080676e-01 6.05182588e-01
-8.29849243e-02 -2.24536747e-01 1.08755898e+00 5.91051757e-01
2.86220610e-01 8.18346441e-01 -1.02487528e+00 -4.78241086e-01
-9.37938750e-01 -9.52288285e-02 6.94040358e-01 1.70389402e+00
5.95741928e-01 3.10699463e-01 9.06342864e-02 1.30255103e+00
-7.35288188e-02 7.51178324e-01 1.01564848e+00 -4.23555493e-01
6.43045425e-01 4.08686966e-01 -4.01064217e-01 -3.37169766e-01
-3.45474184e-02 1.01778537e-01 -4.65377480e-01 -6.48411214e-01
2.50086546e-01 -4.52323556e-01 -1.10059273e+00 1.13462901e+00
2.30132133e-01 -9.61396575e-01 2.79014800e-02 6.31909370e-01
7.29740798e-01 6.52260900e-01 -4.73214723e-02 -3.72807235e-01
1.70473039e+00 -8.82553816e-01 -1.20971155e+00 1.64154209e-02
6.05061471e-01 -1.53713727e+00 1.24356818e+00 5.91809154e-01
-9.02778924e-01 -5.19363403e-01 -7.25022197e-01 -6.00543432e-02
-9.36612308e-01 2.60920197e-01 4.79521394e-01 1.16107798e+00
-5.78681111e-01 1.06957883e-01 -3.66372734e-01 -4.97267157e-01
-1.28704041e-01 2.54182309e-01 -1.85053155e-01 1.67517915e-01
-1.47793663e+00 9.48289752e-01 8.84308577e-01 -7.89980590e-02
-5.95261812e-01 -3.98242623e-01 -1.06551051e+00 -5.69438457e-01
6.26463175e-01 2.23139912e-01 1.21324837e+00 -1.24691808e+00
-1.65701687e+00 1.10431755e+00 2.55967900e-02 -2.36982107e-01
3.71758699e-01 -1.62988141e-01 -1.12533844e+00 -1.34896040e-01
3.22973341e-01 6.13228798e-01 4.71855819e-01 -6.58455729e-01
-1.01180983e+00 -1.43136427e-01 -4.74773198e-01 5.34804054e-02
-3.36174443e-02 8.08217883e-01 -5.47422647e-01 -1.39717937e+00
-1.05374195e-01 -1.11288619e+00 1.07298918e-01 -7.87691593e-01
-4.82641190e-01 -3.70894164e-01 3.97832870e-01 -1.18605983e+00
1.55468273e+00 -2.05221605e+00 -6.42621145e-02 2.59234577e-01
-3.63597721e-01 4.02129084e-01 1.83474153e-01 6.51337802e-01
9.02690142e-02 3.16767901e-01 -4.14165676e-01 -1.42174885e-01
1.92972153e-01 5.51157176e-01 -1.76276267e-01 3.96830499e-01
1.10603005e-01 9.17112827e-01 -8.36151659e-01 -5.47289670e-01
2.51082867e-01 1.82294905e-01 -2.51138117e-02 -1.64476767e-01
-2.53641754e-01 3.66176367e-01 1.97339296e-01 8.99386644e-01
2.27287352e-01 7.00410008e-01 3.19058418e-01 3.86330277e-01
-4.49214697e-01 9.02111769e-01 -1.04666579e+00 1.23927343e+00
-8.46399546e-01 7.33771384e-01 -5.36909811e-02 -5.19128144e-01
1.02898765e+00 5.12279034e-01 -2.34559700e-02 -6.39881253e-01
3.57406944e-01 1.00146258e+00 3.06109995e-01 -3.90739352e-01
1.03659546e+00 -1.75629318e-01 -4.92010802e-01 1.54199436e-01
3.10244530e-01 -8.77332628e-01 5.51373184e-01 -2.04836890e-01
4.28820908e-01 -7.58411661e-02 7.55796254e-01 -6.09691024e-01
6.86178565e-01 6.75812423e-01 7.93416262e-01 -1.52647197e-02
-7.00552464e-02 6.57919228e-01 2.55882651e-01 -3.70425314e-01
-1.20394897e+00 -4.59757209e-01 -3.98187846e-01 1.33297253e+00
-2.10939154e-01 -6.72189355e-01 -9.16595817e-01 -6.26281321e-01
-4.82302815e-01 9.30846930e-01 -3.07051808e-01 2.65513539e-01
-9.83341515e-01 -4.80072826e-01 1.05635464e+00 2.28689849e-01
6.01719260e-01 -1.41012645e+00 -2.72298008e-01 5.26751816e-01
2.28202902e-02 -1.19463301e+00 -9.93184924e-01 3.37899685e-01
-3.08866292e-01 -1.03199589e+00 -7.76534379e-01 -1.34824109e+00
2.30310932e-01 -3.42417598e-01 1.15010250e+00 -6.48541227e-02
-2.47497663e-01 -8.17962643e-03 -7.45413780e-01 -5.96536875e-01
-8.87096465e-01 2.39805862e-01 1.07639343e-01 -4.07069206e-01
1.02278030e+00 3.18918467e-01 2.59515315e-01 4.46061581e-01
-8.65917802e-01 -4.40194041e-01 3.04369539e-01 1.00322831e+00
6.60821557e-01 -2.38186941e-01 3.19827378e-01 -1.47021854e+00
5.82868695e-01 -1.28489658e-01 -7.69678414e-01 2.84579903e-01
-5.59061408e-01 -2.58290678e-01 9.12381709e-01 -6.25185907e-01
-1.00635898e+00 9.05421972e-02 -5.14600456e-01 2.72179276e-01
-4.30378050e-01 3.59862894e-01 -4.86313254e-01 2.70491391e-01
3.49227339e-01 4.73556727e-01 -4.88849461e-01 -7.80693710e-01
3.90190691e-01 1.37693679e+00 6.76913679e-01 -5.47174931e-01
3.55128706e-01 -2.56820291e-01 -7.68714249e-01 -1.30347073e+00
-5.83481014e-01 -7.59862840e-01 -1.17356646e+00 1.79916546e-01
1.04217064e+00 -6.89677894e-01 -2.74950713e-02 5.87432146e-01
-1.27811003e+00 -2.96891481e-01 -5.08146584e-01 6.25398695e-01
-2.40455896e-01 2.50579536e-01 -8.56898963e-01 -5.53428233e-01
-3.05493742e-01 -1.21293008e+00 1.02242708e+00 -2.41243303e-01
-8.14049244e-01 -1.19105804e+00 1.17588244e-01 4.00229216e-01
2.18848954e-03 -2.97117263e-01 1.17338419e+00 -9.18225944e-01
2.21658885e-01 1.42525509e-01 1.46897644e-01 7.73910582e-01
1.90434605e-01 7.28114843e-02 -4.17335421e-01 -7.13970512e-02
-1.62471026e-01 -3.60520542e-01 5.71945608e-01 -7.89587423e-02
6.13886237e-01 -4.73780543e-01 3.10990691e-01 4.65080559e-01
1.07303369e+00 5.48056126e-01 3.79264921e-01 2.54752249e-01
6.94989741e-01 7.61917472e-01 8.05068374e-01 2.33825549e-01
2.69620478e-01 6.74960852e-01 -3.64141464e-01 -3.34277228e-02
-5.63766658e-01 -1.46999806e-01 5.63410103e-01 1.66168451e+00
1.40818939e-01 -1.79451242e-01 -1.15452862e+00 9.35498416e-01
-1.26947558e+00 -3.71996999e-01 -5.55115283e-01 2.15947080e+00
1.40083754e+00 9.56656039e-02 2.25299686e-01 3.72595489e-01
1.08623207e+00 -1.65888622e-01 1.86739955e-02 -1.01102257e+00
-4.70663160e-01 6.44777477e-01 9.39314425e-01 7.36852884e-01
-9.34637487e-01 1.81746006e+00 6.94161034e+00 1.48892546e+00
-1.04801118e+00 3.59499186e-01 2.64287055e-01 5.53822041e-01
-2.91276067e-01 -6.29249495e-03 -1.23341203e+00 6.29533648e-01
1.04059291e+00 2.53637493e-01 3.95231605e-01 6.66682184e-01
1.99293066e-03 -1.03744045e-01 -9.96612132e-01 1.17867196e+00
3.75801176e-01 -1.19520998e+00 7.05676824e-02 1.23681933e-01
9.37857628e-01 9.46264938e-02 -2.78640985e-01 3.19850892e-01
3.69863510e-01 -9.85981882e-01 1.34414816e+00 -3.09652358e-01
1.52802718e+00 -8.50173950e-01 1.14064169e+00 -4.51478995e-02
-1.07632899e+00 4.76086706e-01 -4.23680484e-01 -5.08738682e-02
1.79415584e-01 5.22531927e-01 -9.91863430e-01 2.92006314e-01
7.73236826e-02 5.10473669e-01 -6.24125957e-01 7.90453672e-01
-4.53542411e-01 1.12764275e+00 -1.81628510e-01 -4.76825774e-01
5.11657596e-01 -5.53413451e-01 4.91035581e-01 1.92595541e+00
1.17988557e-01 -2.97278672e-01 3.93377990e-01 2.51492172e-01
1.55782208e-01 7.99738884e-01 -1.18430629e-01 -3.63923967e-01
7.02560246e-01 7.51216590e-01 -1.24589670e+00 -6.62880599e-01
-3.23771417e-01 1.25826502e+00 -8.68356228e-02 1.92177862e-01
-6.22785687e-01 -1.03199756e+00 5.16562104e-01 1.92984175e-02
2.79276937e-01 -5.24289131e-01 -3.84311765e-01 -9.46988702e-01
-1.61226630e-01 -1.37363267e+00 2.36628368e-01 -9.87512991e-02
-1.18817854e+00 8.93396258e-01 -6.80539161e-02 -8.56613815e-01
-1.66107386e-01 -9.97436345e-01 -3.25026289e-02 6.72257602e-01
-1.05123496e+00 -1.05124462e+00 4.60287839e-01 4.42995101e-01
1.28027081e+00 -6.59614742e-01 8.21927547e-01 5.83042443e-01
-6.84726954e-01 5.90779781e-01 5.51549383e-02 5.92897594e-01
7.22316623e-01 -1.42546093e+00 6.47893846e-01 8.02671134e-01
4.82800514e-01 5.30416608e-01 5.18435121e-01 -9.36377466e-01
-8.97545218e-01 -1.27749181e+00 1.70870268e+00 -4.11711812e-01
1.00561774e+00 -8.50484490e-01 -5.11308670e-01 6.05233490e-01
3.93201828e-01 -4.34156686e-01 9.93763626e-01 1.46483758e-03
-1.64916173e-01 2.79114902e-01 -8.46768320e-01 5.10887682e-01
9.96913850e-01 -6.68323815e-01 -9.20630276e-01 6.55987382e-01
8.62890065e-01 -2.90274531e-01 -7.69703805e-01 -3.34434599e-01
3.17025036e-01 -3.98361117e-01 1.94623157e-01 -4.78447914e-01
-1.62556306e-01 -5.86073622e-02 -2.49851137e-01 -1.40407825e+00
2.15632811e-01 -1.17064500e+00 6.91897154e-01 1.59189904e+00
1.00915480e+00 -4.84762609e-01 1.57666221e-01 7.02346712e-02
-5.70092559e-01 -2.11002693e-01 -1.09520531e+00 -1.22289968e+00
2.59175777e-01 -7.31217146e-01 8.31787229e-01 1.12717092e+00
2.36663580e-01 7.22003758e-01 -2.84018606e-01 -5.29687643e-01
-5.03051318e-02 -2.60121077e-01 4.87365156e-01 -8.64528894e-01
-2.24316657e-01 -4.32853609e-01 -4.72099245e-01 -1.15376830e+00
2.80705541e-01 -9.86608207e-01 3.01509112e-01 -8.98839653e-01
-3.17499280e-01 -6.09081984e-01 5.54281414e-01 4.73617554e-01
6.76296949e-02 6.33570015e-01 1.86724499e-01 5.33639193e-02
-3.56697112e-01 5.23964047e-01 1.03062189e+00 2.38451641e-02
-3.26592922e-01 -3.46030705e-02 -7.87367746e-02 7.59013653e-01
7.09775388e-01 -6.31120980e-01 1.17157146e-01 -4.65241969e-01
4.02499318e-01 -4.02115583e-01 -4.66789275e-01 -6.33198082e-01
-7.41407201e-02 -4.02745634e-01 -1.08383205e-02 -6.37723982e-01
1.52073547e-01 -7.93181837e-01 6.15612417e-02 4.48930234e-01
-2.05967680e-01 5.62317371e-01 -1.44160852e-01 -1.89339966e-01
-6.39562011e-01 -5.89943171e-01 8.29530120e-01 -4.74463940e-01
-8.26022089e-01 -4.95872907e-02 -9.56827164e-01 5.24977624e-01
9.46746945e-01 -4.74721402e-01 2.27661699e-01 4.56889123e-02
-4.96545762e-01 -4.15383130e-02 4.60167557e-01 4.87926602e-01
2.52906650e-01 -1.16521561e+00 -6.86801910e-01 5.58486402e-01
3.40070069e-01 -3.23118091e-01 -7.15250015e-01 6.94312930e-01
-1.06991756e+00 7.28293836e-01 -4.40660939e-02 -2.00703934e-01
-1.16033530e+00 3.92153829e-01 -1.07760597e-02 1.68279007e-01
-4.81098861e-01 8.80570292e-01 -5.10975838e-01 -8.29812467e-01
2.52409577e-01 -6.01562440e-01 -1.64118960e-01 3.48193705e-01
4.07844126e-01 6.05747640e-01 4.10095423e-01 -1.44102502e+00
-4.99644458e-01 5.83306015e-01 4.03146371e-02 -4.35662746e-01
8.13242137e-01 -5.36186039e-01 -3.98914784e-01 8.10482085e-01
1.06224704e+00 6.86418474e-01 -2.99795896e-01 -7.69026801e-02
5.46411395e-01 -3.44619691e-01 -1.53452307e-01 -8.74329031e-01
-6.67520344e-01 7.41516650e-01 1.25003820e-02 9.21019632e-03
6.39062285e-01 -1.17674482e-03 1.38904488e+00 3.56463581e-01
3.85169774e-01 -1.87982500e+00 -5.46648502e-01 1.32266808e+00
6.70812011e-01 -1.23002708e+00 -4.28658932e-01 -6.97355330e-01
-9.55651462e-01 1.19271386e+00 1.98165730e-01 1.16405770e-01
6.52980506e-01 1.70220047e-01 4.95318502e-01 -1.77548215e-01
-1.64270073e-01 -3.74072284e-01 2.07885832e-01 7.16154039e-01
7.47245371e-01 6.36955738e-01 -9.43135202e-01 4.60142702e-01
-1.13219619e+00 -6.21864498e-01 6.28746986e-01 5.97620368e-01
-3.32374275e-01 -1.34798658e+00 -5.19858241e-01 2.07899809e-01
-8.64537477e-01 -8.31575155e-01 -9.02652740e-01 1.01234365e+00
7.44990885e-01 1.02296519e+00 1.68205202e-01 -1.32027701e-01
2.34550536e-01 1.79573268e-01 2.50696123e-01 -1.27249610e+00
-9.30739641e-01 3.99912000e-01 6.57054543e-01 -5.58771677e-02
-1.42817587e-01 -6.57866001e-01 -1.20717037e+00 -1.99829727e-01
-4.82460022e-01 5.46922982e-01 7.47422934e-01 1.04652429e+00
-3.92366737e-01 1.00423597e-01 2.41837621e-01 -2.88468450e-01
-3.33097726e-01 -1.24426270e+00 -1.01600480e+00 2.47604251e-01
6.02804534e-02 -1.18822500e-01 -3.39426517e-01 4.70181614e-01] | [10.93042278289795, 10.279072761535645] |
7ee610ad-937e-4ceb-8b4b-5a4e672dc6d2 | unseen-object-amodal-instance-segmentation | 2109.11103 | null | https://arxiv.org/abs/2109.11103v2 | https://arxiv.org/pdf/2109.11103v2.pdf | Unseen Object Amodal Instance Segmentation via Hierarchical Occlusion Modeling | Instance-aware segmentation of unseen objects is essential for a robotic system in an unstructured environment. Although previous works achieved encouraging results, they were limited to segmenting the only visible regions of unseen objects. For robotic manipulation in a cluttered scene, amodal perception is required to handle the occluded objects behind others. This paper addresses Unseen Object Amodal Instance Segmentation (UOAIS) to detect 1) visible masks, 2) amodal masks, and 3) occlusions on unseen object instances. For this, we propose a Hierarchical Occlusion Modeling (HOM) scheme designed to reason about the occlusion by assigning a hierarchy to a feature fusion and prediction order. We evaluated our method on three benchmarks (tabletop, indoors, and bin environments) and achieved state-of-the-art (SOTA) performance. Robot demos for picking up occluded objects, codes, and datasets are available at https://sites.google.com/view/uoais | ['Kyoobin Lee', 'Seongho Bak', 'Raeyoung Kang', 'Sangjun Noh', 'Taewon Kim', 'Joosoon Lee', 'Seunghyeok Back'] | 2021-09-23 | null | null | null | null | ['amodal-instance-segmentation', 'unseen-object-instance-segmentation'] | ['computer-vision', 'computer-vision'] | [ 2.22345963e-01 3.56772751e-01 -5.24287820e-02 -4.02401417e-01
-4.03629929e-01 -7.45658934e-01 4.21656698e-01 2.34638616e-01
8.22673962e-02 4.64567482e-01 -2.75930345e-01 5.40362634e-02
-2.00828776e-01 -4.01221037e-01 -8.85179758e-01 -6.23988867e-01
-1.19598009e-01 1.00657964e+00 6.20185196e-01 -8.77595246e-02
2.27939323e-01 6.39754474e-01 -2.01785851e+00 5.47431469e-01
1.10998869e+00 1.05364501e+00 8.12742591e-01 5.55947781e-01
-7.70443976e-02 5.22801101e-01 -6.27441704e-01 -2.34580729e-02
5.57293653e-01 2.08652452e-01 -8.09352160e-01 6.17230117e-01
6.51214123e-01 -3.54468346e-01 -1.12843096e-01 1.01701367e+00
8.15038849e-03 1.90346509e-01 7.30956733e-01 -1.66391563e+00
-3.69569182e-01 5.45192242e-01 -5.42235732e-01 1.35524020e-01
3.67625713e-01 1.97954252e-01 5.56221962e-01 -9.96853232e-01
8.89209330e-01 1.53500569e+00 9.15154293e-02 2.88746685e-01
-1.17636049e+00 -2.96382040e-01 6.54584169e-01 2.03350425e-01
-1.23332250e+00 -3.42422873e-01 4.66234982e-01 -7.46209204e-01
7.57865846e-01 3.80319059e-01 4.71724093e-01 9.24932718e-01
3.63512903e-01 1.15540767e+00 9.19527769e-01 -8.44025314e-02
4.22973633e-01 1.90820679e-01 4.68287736e-01 4.47149694e-01
5.01883268e-01 -1.84698462e-01 -3.15213025e-01 4.99927923e-02
5.86344481e-01 2.92349100e-01 -3.45040523e-02 -9.47375059e-01
-1.29053044e+00 4.35473561e-01 7.44091392e-01 1.45392085e-03
-5.49152076e-01 6.03770502e-02 2.77314093e-02 -5.39091304e-02
2.53829539e-01 3.82567018e-01 -4.83671844e-01 2.56731302e-01
-5.05570889e-01 2.79425174e-01 8.85673344e-01 1.60103250e+00
8.49971473e-01 -3.99667531e-01 -2.30177358e-01 5.21175146e-01
1.47752568e-01 3.28163594e-01 2.35639624e-02 -1.15113533e+00
5.75980723e-01 8.51062000e-01 6.33785129e-01 -6.75950766e-01
-5.81780076e-01 -2.71708727e-01 -3.39500308e-01 5.35320997e-01
3.00889641e-01 -3.68207903e-03 -1.60850406e+00 1.07055998e+00
8.12218130e-01 -2.68580347e-01 9.07592028e-02 1.20161641e+00
9.55936730e-01 4.51364726e-01 6.97177872e-02 2.10312024e-01
1.35697806e+00 -1.25309074e+00 -8.17891955e-01 -7.33246744e-01
2.97736794e-01 -7.28643537e-01 7.55821705e-01 6.65368855e-01
-9.64659750e-01 -7.34353662e-01 -9.60076749e-01 -2.94666141e-01
-6.31120682e-01 1.41081408e-01 6.49771631e-01 1.60626471e-01
-7.23422408e-01 5.30748606e-01 -1.06588054e+00 -4.53132212e-01
6.12028122e-01 5.25845230e-01 -4.27649498e-01 -3.38869840e-01
-4.44265842e-01 7.20846355e-01 7.05981374e-01 2.52640963e-01
-1.24425185e+00 -2.30207622e-01 -8.79610777e-01 -9.06208009e-02
8.46969128e-01 -4.05025631e-01 1.03024447e+00 -5.67671657e-01
-8.70220721e-01 6.28183722e-01 -1.36194378e-01 -4.12564456e-01
6.34382248e-01 -7.07536936e-01 7.56929591e-02 5.26158996e-02
1.32809296e-01 1.20029199e+00 8.91104996e-01 -2.05781269e+00
-9.60189104e-01 -7.23225355e-01 2.72921413e-01 5.30033827e-01
2.40638003e-01 -4.09223318e-01 -5.61067879e-01 -3.04908305e-01
9.25854266e-01 -9.48320985e-01 -3.83816987e-01 -5.22957519e-02
-8.47111821e-01 -3.56699377e-01 1.17750406e+00 -5.61991274e-01
6.81993783e-01 -2.19470501e+00 3.77905369e-01 -5.27538024e-02
2.09318727e-01 -1.18441097e-01 1.57373734e-02 3.57799739e-01
3.08348089e-01 -1.80602804e-01 -1.67325541e-01 -5.80831349e-01
3.14051270e-01 4.62788582e-01 -3.75729889e-01 5.30213118e-01
1.09113149e-01 7.17011809e-01 -9.73719597e-01 -6.09550297e-01
6.93996012e-01 2.83023715e-01 -4.16458964e-01 2.11433008e-01
-5.24175227e-01 5.99851668e-01 -4.75531220e-01 1.23617506e+00
9.50453401e-01 7.83791021e-02 3.42531577e-02 -1.18134044e-01
-5.36967963e-02 -3.13794017e-01 -1.50517023e+00 1.72073507e+00
4.73111831e-02 4.38568383e-01 3.22248220e-01 -4.61088479e-01
8.65401864e-01 -4.18792889e-02 4.98184592e-01 -2.14133129e-01
2.49423102e-01 2.89296240e-01 -2.30658531e-01 -8.50053191e-01
8.92241895e-01 6.89007461e-01 -1.78694725e-01 -3.00503194e-01
-1.25536591e-01 -3.65473926e-01 4.08635855e-01 1.63890183e-01
1.05335939e+00 4.15504634e-01 1.64434314e-01 -3.84755760e-01
4.51233471e-03 4.75409746e-01 3.72421086e-01 1.06127834e+00
-3.33668858e-01 7.23042846e-01 2.09216982e-01 -5.90052426e-01
-6.40471578e-01 -1.25884378e+00 -3.85436833e-01 1.09753966e+00
1.24421632e+00 -2.68788580e-02 -7.12518454e-01 -6.06665134e-01
2.21702352e-01 6.94967628e-01 -5.46733022e-01 3.43116790e-01
-3.88323635e-01 -2.97452897e-01 -3.99332941e-01 4.58712906e-01
2.82354355e-01 -1.37057543e+00 -1.02258241e+00 9.89431143e-02
-3.23800296e-01 -1.32784390e+00 -6.07421994e-02 6.22359991e-01
-7.58894861e-01 -1.20639288e+00 -4.26342607e-01 -9.00627553e-01
9.76265430e-01 5.17489076e-01 9.49630380e-01 -2.68451184e-01
-6.26695812e-01 5.09526670e-01 -5.11665761e-01 -7.01084435e-01
-9.29493755e-02 5.96506372e-02 1.11234523e-01 -1.56708419e-01
1.92932263e-01 -2.00166285e-01 -7.67380297e-01 6.80637777e-01
-8.89904082e-01 1.18318193e-01 6.25012338e-01 2.41953731e-01
7.85871625e-01 -5.98144764e-03 4.97348830e-02 -6.31442487e-01
8.53747055e-02 -5.53474903e-01 -7.23352075e-01 2.57586420e-01
-1.58069208e-01 -1.47779524e-01 1.61640774e-02 -5.99365652e-01
-9.41583514e-01 6.51644886e-01 5.44081211e-01 -5.57069004e-01
-7.81359792e-01 -1.23061903e-01 -2.92054325e-01 1.26773983e-01
7.20992267e-01 -4.15938646e-01 -5.50717831e-01 -5.94097733e-01
3.79367411e-01 8.54624212e-01 6.54255867e-01 -6.01183772e-01
5.16326547e-01 7.65704691e-01 -2.13654786e-01 -6.44925833e-01
-7.32895136e-01 -7.74130225e-01 -9.37419236e-01 -3.52096707e-01
9.88575101e-01 -8.79325509e-01 -5.39933920e-01 1.07743792e-01
-1.21408570e+00 -4.30139810e-01 -3.48926753e-01 3.36802185e-01
-6.14810407e-01 9.55119580e-02 -2.41768420e-01 -1.12952662e+00
7.31201172e-02 -1.41701734e+00 1.51587248e+00 2.83356011e-01
-1.43653154e-01 -2.09913135e-01 -4.94183034e-01 4.93798733e-01
-1.64714530e-01 6.11105323e-01 4.80816245e-01 -5.63141584e-01
-1.20688784e+00 -3.09944540e-01 -1.89090475e-01 -1.86922461e-01
2.04774411e-03 -2.79312193e-01 -1.05870175e+00 -3.31778646e-01
-2.63854712e-01 -1.70387730e-01 8.78512204e-01 3.99302065e-01
1.22470796e+00 -1.38217598e-01 -9.47884560e-01 2.88674623e-01
1.23705208e+00 3.67008626e-01 5.44268966e-01 1.18500270e-01
7.82036722e-01 7.83878505e-01 1.23316967e+00 5.27363598e-01
2.71733671e-01 7.91821718e-01 1.12496996e+00 -4.38043028e-02
2.03482956e-02 1.34887010e-01 4.76156734e-02 -9.25101265e-02
7.35705346e-02 -7.77398050e-01 -8.43988955e-01 7.74272501e-01
-2.14572215e+00 -2.68143505e-01 -4.96684551e-01 2.02709007e+00
1.65103629e-01 2.61096656e-01 -3.75781059e-02 -7.27459341e-02
1.00094509e+00 -6.07561693e-02 -8.95059884e-01 -7.64700547e-02
1.68692261e-01 -4.24352050e-01 4.45735514e-01 4.66632277e-01
-1.45794392e+00 1.08571732e+00 5.04525900e+00 5.90701878e-01
-4.81787980e-01 2.35964116e-02 3.35671186e-01 -8.13864395e-02
1.43083051e-01 -5.42864576e-02 -1.00253654e+00 2.63147742e-01
-2.62663178e-02 4.86098021e-01 4.33727950e-01 1.31016791e+00
-1.98410884e-01 -5.51362753e-01 -1.28798485e+00 9.91592169e-01
-8.17083269e-02 -6.78351521e-01 -1.58410326e-01 5.24438620e-02
8.06426108e-01 1.93791255e-01 1.10211037e-02 2.57434517e-01
3.85570496e-01 -7.49891818e-01 1.09311402e+00 3.71760547e-01
1.98166624e-01 -2.34066904e-01 6.30488873e-01 4.27131146e-01
-1.14356697e+00 -3.85859042e-01 -4.14493650e-01 7.53166005e-02
1.15497090e-01 4.33906257e-01 -8.84315193e-01 5.13215423e-01
1.01696634e+00 4.10469830e-01 -4.64244992e-01 1.30516446e+00
-2.58516550e-01 -2.30168939e-01 -4.17212129e-01 1.72481090e-01
1.75871149e-01 -1.96130872e-02 8.05109978e-01 8.99661124e-01
1.24041498e-01 2.08781138e-01 7.31147349e-01 7.94264674e-01
2.18940735e-01 -2.54728585e-01 -3.61128390e-01 2.90579617e-01
5.07905483e-01 1.40643585e+00 -1.36139870e+00 -3.55897605e-01
1.33758530e-01 1.24062872e+00 1.86531112e-01 5.45218050e-01
-8.77295196e-01 -2.90402234e-01 5.49780965e-01 1.96890131e-01
5.22238255e-01 -3.96705776e-01 -3.55442524e-01 -9.03503239e-01
3.72741789e-01 -4.57553089e-01 3.99757951e-01 -1.00157261e+00
-9.24716055e-01 6.51187062e-01 2.80835658e-01 -1.33663893e+00
1.73721835e-01 -7.75935471e-01 -2.34788386e-05 3.83197606e-01
-1.17548108e+00 -1.01707983e+00 -1.00284386e+00 2.14695066e-01
1.13030386e+00 4.24297810e-01 4.75702703e-01 4.52868752e-02
-3.58515739e-01 -2.37239093e-01 -1.25697836e-01 -2.79734731e-01
4.81529504e-01 -1.33020294e+00 1.31955683e-01 6.53995395e-01
1.74140707e-01 2.10830241e-01 1.01623094e+00 -1.01779008e+00
-1.29479730e+00 -1.04134643e+00 3.34034801e-01 -8.70076835e-01
1.82465136e-01 -6.99114025e-01 -8.01229894e-01 8.38339210e-01
1.07036121e-01 2.09371448e-01 -1.06043741e-01 -1.02196045e-01
1.20237954e-01 2.21944451e-01 -1.37582469e+00 5.67307651e-01
1.40500903e+00 3.01255763e-01 -5.41941941e-01 7.08318412e-01
7.82538533e-01 -1.07708061e+00 -6.86551094e-01 7.24376738e-01
5.69142520e-01 -9.54451263e-01 1.01251483e+00 -3.41300547e-01
3.05062741e-01 -6.43119156e-01 -3.50908667e-01 -8.70417595e-01
-1.74090609e-01 -4.49029058e-01 -5.39218545e-01 6.91772938e-01
2.46228874e-01 -4.22760308e-01 8.23752224e-01 9.47822750e-01
-6.23416483e-01 -6.94102824e-01 -6.23764753e-01 -9.26647723e-01
-5.50873697e-01 -1.41802773e-01 6.10489726e-01 5.19370556e-01
-2.12943897e-01 -5.13980985e-02 2.29783747e-02 8.96587610e-01
6.90286100e-01 3.88003677e-01 9.01278675e-01 -1.44505000e+00
1.23412646e-01 -1.06774114e-01 -6.07071757e-01 -1.07107508e+00
-1.50872231e-01 -6.07710302e-01 5.65587759e-01 -2.13188267e+00
1.22823074e-01 -5.08640885e-01 -1.03570029e-01 5.41636765e-01
-7.24935681e-02 2.39696145e-01 3.44444782e-01 3.70382875e-01
-1.16791666e+00 3.03296030e-01 1.21453917e+00 -3.17665964e-01
-5.97423732e-01 -7.17620477e-02 -2.30711430e-01 7.81575441e-01
9.05599594e-01 -5.31355798e-01 -2.56111294e-01 -6.45050049e-01
-2.39224643e-01 -1.19925611e-01 6.12817347e-01 -1.45433295e+00
2.22666100e-01 -1.61546290e-01 5.89022458e-01 -1.25146210e+00
8.10423553e-01 -1.13781321e+00 2.49805063e-01 5.23591220e-01
-8.17414671e-02 -3.50418091e-01 3.00279707e-01 6.80224419e-01
1.51247576e-01 -3.87825310e-01 4.44970429e-01 -2.20834747e-01
-1.17712748e+00 1.89351991e-01 -2.01624408e-01 -2.08497643e-01
1.34779084e+00 -4.17827457e-01 -5.02977550e-01 3.59904096e-02
-1.18607223e+00 6.07071519e-01 5.70503712e-01 7.94357777e-01
6.78979158e-01 -7.73093045e-01 -3.96851033e-01 1.20369293e-01
3.23968858e-01 6.18262291e-01 3.51374060e-01 6.15622461e-01
-5.87882102e-01 2.68399149e-01 -1.86586305e-02 -1.06558716e+00
-1.22539890e+00 8.58830512e-01 3.04034818e-02 1.86475322e-01
-6.32701159e-01 8.03697586e-01 6.45604491e-01 -3.83668184e-01
7.26238787e-01 -8.08072209e-01 -1.42128035e-01 2.33722124e-02
1.35145336e-01 5.82484484e-01 -6.08703494e-03 -4.47593302e-01
-3.19228411e-01 4.19597059e-01 -2.26781312e-02 2.30002746e-01
1.05073369e+00 -3.19950074e-01 9.23129842e-02 4.09151167e-01
5.42453647e-01 -4.28534031e-01 -1.65573120e+00 -1.14086226e-01
1.52857099e-02 -5.91296375e-01 -2.61407584e-01 -7.30586231e-01
-6.19851887e-01 5.21974087e-01 7.05601037e-01 3.72709274e-01
7.52620995e-01 4.80462164e-01 3.84119838e-01 7.31928408e-01
7.61363208e-01 -1.11814976e+00 1.56709716e-01 3.18331540e-01
1.10858703e+00 -1.52634537e+00 1.39474049e-01 -1.02724195e+00
-6.27598286e-01 9.61863756e-01 1.01454568e+00 1.52724348e-02
4.27060604e-01 1.85459122e-01 -7.62451291e-02 -4.80685711e-01
-2.21137792e-01 -5.56179285e-01 2.54155159e-01 6.09139621e-01
-2.86331028e-01 2.22261861e-01 1.16844274e-01 3.53428215e-01
-6.88148513e-02 -2.40868762e-01 3.41089994e-01 1.36346769e+00
-7.91347325e-01 -4.12253559e-01 -6.59324765e-01 4.13573265e-01
-5.68020940e-02 3.75451058e-01 -4.50919986e-01 9.41263199e-01
4.64508742e-01 1.32503033e+00 1.59070894e-01 -1.94539547e-01
3.92256230e-01 -1.55502498e-01 4.90433574e-01 -8.53907883e-01
-2.36870602e-01 9.00453851e-02 -2.87267603e-02 -8.14285100e-01
-2.97867060e-01 -5.99013448e-01 -1.38800049e+00 4.29904729e-01
-5.29989660e-01 -1.96633846e-01 9.86911416e-01 6.99122965e-01
4.22071785e-01 7.91716218e-01 2.73716539e-01 -1.63480580e+00
-8.30662344e-03 -8.69023740e-01 -4.81142730e-01 4.51039046e-01
3.32514942e-01 -1.03293812e+00 -3.94406885e-01 1.28159389e-01] | [6.1657795906066895, -1.0615936517715454] |
d7257a37-fcd3-41cd-a18f-1262bf2e2348 | augmenting-hessians-with-inter-layer | 2306.04879 | null | https://arxiv.org/abs/2306.04879v1 | https://arxiv.org/pdf/2306.04879v1.pdf | Augmenting Hessians with Inter-Layer Dependencies for Mixed-Precision Post-Training Quantization | Efficiently serving neural network models with low latency is becoming more challenging due to increasing model complexity and parameter count. Model quantization offers a solution which simultaneously reduces memory footprint and compute requirements. However, aggressive quantization may lead to an unacceptable loss in model accuracy owing to differences in sensitivity to numerical imperfection across different layers in the model. To address this challenge, we propose a mixed-precision post training quantization (PTQ) approach that assigns different numerical precisions to tensors in a network based on their specific needs, for a reduced memory footprint and improved latency while preserving model accuracy. Previous works rely on layer-wise Hessian information to determine numerical precision, but as we demonstrate, Hessian estimation is typically insufficient in determining an effective ordering of layer sensitivities. We address this by augmenting the estimated Hessian with additional information to capture inter-layer dependencies. We demonstrate that this consistently improves PTQ performance along the accuracy-latency Pareto frontier across multiple models. Our method combines second-order information and inter-layer dependencies to guide a bisection search, finding quantization configurations within a user-configurable model accuracy degradation range. We evaluate the effectiveness of our method on the ResNet50, MobileNetV2, and BERT models. Our experiments demonstrate latency reductions compared to a 16-bit baseline of $25.48\%$, $21.69\%$, and $33.28\%$ respectively, while maintaining model accuracy to within $99.99\%$ of the baseline model. | ['Yu Emma Wang', 'Siddharth Joshi', 'Caitlin Stanton', 'Elfie Guo', 'Jian Li', 'Tom Jablin', 'Chiachen Chou', 'Xiaofan Zhang', 'Navid Lambert-Shirzad', 'Clemens JS Schaefer'] | 2023-06-08 | null | null | null | null | ['quantization'] | ['methodology'] | [-1.74604841e-02 -4.91260231e-01 -2.41271004e-01 -5.33376336e-01
-1.04929423e+00 -6.54796183e-01 4.67558717e-03 2.24528134e-01
-7.15243101e-01 5.91722548e-01 -1.78760156e-01 -6.60294950e-01
-3.14927310e-01 -5.35325587e-01 -7.16595471e-01 -3.17984313e-01
-3.09932411e-01 2.58425415e-01 2.76598990e-01 1.35777071e-01
3.55080843e-01 6.33392930e-01 -1.34633613e+00 3.48108143e-01
8.05539846e-01 1.56429672e+00 -9.17998478e-02 7.95655906e-01
-8.68723169e-02 7.73916841e-01 -7.39199996e-01 -4.27412599e-01
5.50311029e-01 1.39902711e-01 -6.50104463e-01 -4.29354250e-01
8.96078885e-01 -7.72122681e-01 -1.75962210e-01 1.14673483e+00
4.52227086e-01 -1.76071927e-01 5.17899990e-01 -1.34731567e+00
1.66517049e-01 5.84620357e-01 -5.74076355e-01 3.92105132e-01
-3.89634728e-01 2.56789953e-01 1.01967156e+00 -6.77740276e-01
2.28301749e-01 1.11725223e+00 9.46638525e-01 3.19374263e-01
-1.54926288e+00 -1.26109028e+00 1.08397260e-01 1.81136906e-01
-1.93725181e+00 -7.65450776e-01 2.89475113e-01 -3.74361753e-01
1.40869355e+00 1.49227202e-01 5.26727557e-01 3.49366307e-01
3.44079584e-01 3.13489258e-01 7.61654317e-01 -7.53546655e-02
5.61161280e-01 -1.07141845e-01 3.74968231e-01 7.88650274e-01
3.17210674e-01 -8.22583139e-02 -6.65026188e-01 -2.85653919e-01
8.73000741e-01 -4.07963902e-01 -1.55068114e-01 1.75161753e-02
-7.06456423e-01 5.57998002e-01 6.39200687e-01 -1.78019807e-01
-3.91588241e-01 8.37589800e-01 4.89462256e-01 1.35049418e-01
3.15145940e-01 5.38960457e-01 -7.29802132e-01 -6.08374834e-01
-1.42593563e+00 3.25174451e-01 8.31939578e-01 1.02292097e+00
8.49522591e-01 3.76988381e-01 -9.32630450e-02 7.97639430e-01
4.14214015e-01 3.44019920e-01 1.53613776e-01 -1.43476272e+00
7.62153268e-01 3.91988903e-01 -4.71105464e-02 -9.72618163e-01
-5.22859037e-01 -7.46707022e-01 -8.53830099e-01 2.05711484e-01
4.35791314e-01 -1.89264968e-01 -1.03224015e+00 1.80932295e+00
1.81696862e-01 1.42088041e-01 -9.66300443e-02 7.41027653e-01
1.69515952e-01 6.78597450e-01 2.72485226e-01 -1.01704421e-02
1.25122762e+00 -7.07852244e-01 -3.23549658e-01 -2.08776504e-01
8.96814466e-01 -6.99073732e-01 1.00870728e+00 4.52082336e-01
-1.27763724e+00 -2.29068816e-01 -1.40956140e+00 -4.95598130e-02
3.91038992e-02 8.02513584e-02 3.51055175e-01 8.18082690e-01
-1.42520249e+00 7.61883676e-01 -1.28597987e+00 9.00995880e-02
5.47704875e-01 8.87313783e-01 3.87535542e-01 1.22195587e-01
-9.02708232e-01 6.00899398e-01 3.08615416e-01 1.30452625e-02
-6.36628687e-01 -1.30486655e+00 -4.65139955e-01 4.55259711e-01
1.28289253e-01 -4.31557506e-01 1.34789515e+00 -6.60246789e-01
-1.29467797e+00 -4.60995063e-02 -1.74027964e-01 -9.20665741e-01
3.57553780e-01 2.98836967e-03 -2.50586838e-01 1.42303795e-01
-3.61471534e-01 1.02599919e+00 6.44948483e-01 -9.89915729e-01
-7.16956496e-01 -3.46568912e-01 1.78999886e-01 1.81446344e-01
-7.74594188e-01 -1.67916358e-01 -7.02439368e-01 -5.30631900e-01
2.04975948e-01 -9.75440979e-01 -2.94945925e-01 2.46737495e-01
-1.81002289e-01 3.09166044e-01 4.77381885e-01 -5.99287033e-01
1.63132250e+00 -1.93938231e+00 -5.01134872e-01 6.06799483e-01
5.32818615e-01 3.11649114e-01 -3.15943249e-02 3.10500413e-02
5.75294077e-01 5.20501971e-01 4.28183191e-03 -3.20751339e-01
4.75776270e-02 1.86526492e-01 -1.16552398e-01 2.46259704e-01
1.80239841e-01 5.31098664e-01 -4.62052047e-01 -5.74003696e-01
-1.49722278e-01 7.71846831e-01 -1.07477760e+00 -2.71674335e-01
-1.20970726e-01 -3.55112910e-01 -2.28130400e-01 5.53376317e-01
8.05628181e-01 -6.40626013e-01 3.43958765e-01 -6.89316928e-01
-1.17686011e-01 5.89248717e-01 -1.34119117e+00 1.20095861e+00
-6.69290304e-01 6.47360981e-01 3.78492981e-01 -4.14757371e-01
7.02319503e-01 6.71229511e-02 3.70641410e-01 -7.66533911e-01
1.26857415e-01 4.50182140e-01 2.33456001e-01 1.81131214e-01
7.18989134e-01 1.49359256e-01 2.79319942e-01 2.42861927e-01
-2.24646062e-01 8.77538547e-02 2.66022474e-01 1.18679620e-01
1.27679205e+00 -4.45394307e-01 -1.92453668e-01 -5.77660918e-01
4.24911268e-02 2.61304639e-02 5.45914471e-01 6.79895818e-01
-4.20147926e-01 3.02639037e-01 5.52083194e-01 -1.24756292e-01
-1.15797222e+00 -1.02173221e+00 -3.28121722e-01 9.15690660e-01
-8.92701522e-02 -6.09490514e-01 -9.24774170e-01 -1.15373641e-01
1.41065875e-02 4.67073381e-01 -1.49782002e-01 -8.44620541e-02
-6.10683143e-01 -6.10912204e-01 9.77707863e-01 7.64415741e-01
9.11310315e-01 -2.37984881e-01 -8.34799230e-01 2.66835332e-01
1.53179586e-01 -1.20944142e+00 -5.73129714e-01 3.82975727e-01
-1.08491468e+00 -6.29489362e-01 -3.31297487e-01 -5.23277640e-01
4.70639229e-01 -3.34683917e-02 1.18599248e+00 1.70293450e-01
8.47513080e-02 -1.34588614e-01 2.53758520e-01 -1.23421967e-01
-2.73788512e-01 5.84447265e-01 1.64554581e-01 -4.12187576e-01
1.64387584e-01 -5.99094033e-01 -7.14380324e-01 1.22748949e-01
-7.34584570e-01 1.28782317e-01 5.77167094e-01 6.49431407e-01
6.78720534e-01 1.83883741e-01 3.10386181e-01 -5.90967417e-01
6.30740941e-01 -2.53389001e-01 -1.07113922e+00 3.64892371e-02
-1.22468841e+00 4.08162504e-01 8.27212214e-01 -4.70868945e-01
-5.23486316e-01 -1.63703356e-02 -4.80649360e-02 -7.56884813e-01
4.74916905e-01 5.60160756e-01 9.66036022e-02 -4.07153308e-01
6.97092533e-01 -8.33146721e-02 -6.44492060e-02 -3.15053672e-01
-2.69268099e-02 4.97039586e-01 4.30885851e-01 -5.63093424e-01
4.74809825e-01 1.16021119e-01 1.63856789e-01 -5.77679634e-01
-3.26521665e-01 -1.88408852e-01 -1.89690828e-01 -5.98942675e-02
3.78460318e-01 -1.05002570e+00 -9.04270113e-01 2.69166976e-01
-1.10223567e+00 -5.78244090e-01 7.58553073e-02 4.73767847e-01
-1.65027678e-01 1.05721295e-01 -8.80747616e-01 -7.34118819e-01
-7.09715664e-01 -1.50135434e+00 8.10617924e-01 -2.14735679e-02
-8.87166336e-02 -7.31686592e-01 -5.91998339e-01 6.32863864e-02
7.70113528e-01 -1.39872566e-01 1.01282692e+00 -3.53826225e-01
-9.12629306e-01 -2.45309286e-02 -8.23514163e-01 4.22397256e-01
-2.05681860e-01 6.29837736e-02 -8.53926420e-01 -4.00588721e-01
-2.31880799e-01 -1.62264243e-01 6.26725435e-01 4.07792479e-01
1.26060033e+00 -7.04606593e-01 -5.00768647e-02 8.57039511e-01
1.70141828e+00 2.44792268e-01 3.35379124e-01 1.89195067e-01
7.18047976e-01 9.79946107e-02 1.18789569e-01 6.00308239e-01
4.47507322e-01 7.51339436e-01 3.27805996e-01 2.38952488e-01
-1.08093172e-01 -6.45706104e-03 1.65301830e-01 8.69160950e-01
4.52565432e-01 -2.31665373e-01 -1.40226150e+00 3.86553466e-01
-1.32557821e+00 -4.24618572e-01 1.75618947e-01 2.25960684e+00
1.10179770e+00 7.19907343e-01 3.36093381e-02 3.22940677e-01
3.78837198e-01 9.79767554e-03 -7.85146534e-01 -6.09177589e-01
4.06094700e-01 2.00109735e-01 1.34685361e+00 6.05864584e-01
-6.98947191e-01 8.10821176e-01 6.05189466e+00 1.02539933e+00
-1.44327116e+00 -1.77037192e-03 7.50137269e-01 -5.52488267e-01
-4.28712815e-02 -2.13312477e-01 -1.10666740e+00 5.36997497e-01
1.62033844e+00 -2.20372304e-01 7.79522836e-01 7.71163404e-01
3.52414250e-01 4.85176556e-02 -1.08668017e+00 1.09833777e+00
-3.27583939e-01 -1.56380093e+00 8.52533281e-02 2.46232450e-01
6.84214413e-01 3.31571609e-01 2.62966156e-01 1.35925934e-01
2.04609111e-01 -1.00727499e+00 8.34392190e-01 1.85553744e-01
9.98202026e-01 -1.13147104e+00 5.45590937e-01 1.52544126e-01
-1.51651728e+00 -1.79667771e-01 -3.11566770e-01 -1.06175937e-01
-1.01429626e-01 5.37913442e-01 -1.15132880e+00 -1.36915684e-01
8.02927792e-01 1.35292098e-01 -6.06380880e-01 9.86876011e-01
4.20001864e-01 8.21078122e-01 -8.11070025e-01 -3.05246830e-01
3.71159762e-01 1.94123253e-01 3.52028795e-02 1.11173940e+00
1.19431399e-01 8.42064023e-02 3.68942460e-03 7.83357263e-01
-3.82089138e-01 -7.96394330e-03 1.38695210e-01 1.28080308e-01
1.15811503e+00 8.68638873e-01 -7.58190572e-01 -4.96930182e-01
-5.81389852e-02 5.44651985e-01 3.98789912e-01 3.80831361e-01
-1.01229537e+00 -4.58323538e-01 1.03445709e+00 1.70483083e-01
2.30368987e-01 -5.39844036e-01 -9.59362626e-01 -7.38024414e-01
1.52319461e-01 -8.16452622e-01 2.63396557e-02 -3.68973285e-01
-6.77442372e-01 6.89656913e-01 8.19124952e-02 -1.02495658e+00
-3.55053127e-01 -4.00130779e-01 1.13199651e-01 9.76038575e-01
-1.40331483e+00 -5.64518869e-01 -1.92748711e-01 1.50136322e-01
2.90685654e-01 2.00910270e-01 4.84881967e-01 8.06622565e-01
-5.96204340e-01 1.41405129e+00 2.60269791e-01 -1.22430503e-01
2.84696221e-01 -9.15568352e-01 6.15434825e-01 7.05416739e-01
-2.49444053e-01 8.31602037e-01 6.31965160e-01 -4.69191849e-01
-1.46619689e+00 -1.38688588e+00 7.57609367e-01 -5.75003214e-02
5.94271302e-01 -3.45597804e-01 -9.90442991e-01 4.44668978e-01
-4.10497397e-01 7.56114628e-03 3.67304593e-01 -9.11480412e-02
-4.80166078e-01 -5.18218994e-01 -1.22076178e+00 8.94358993e-01
9.38533902e-01 -5.26736259e-01 4.30512488e-01 1.51051711e-02
8.23306978e-01 -6.19645059e-01 -1.32700825e+00 3.52972478e-01
8.04961205e-01 -7.21196830e-01 9.00906861e-01 -2.38150910e-01
1.54317841e-01 -1.51824459e-01 -5.37725270e-01 -8.17475915e-01
-9.67013389e-02 -4.95678037e-01 -2.27636382e-01 1.06573606e+00
6.72194481e-01 -6.32589936e-01 1.14998174e+00 1.20929217e+00
-1.58500150e-01 -1.11791992e+00 -1.07527530e+00 -8.54416788e-01
1.75937042e-01 -8.31915677e-01 8.61133397e-01 4.81150955e-01
-3.61335605e-01 1.72242507e-01 7.41044655e-02 3.91749054e-01
7.36362398e-01 -5.01463234e-01 5.34315050e-01 -9.08412158e-01
-1.89571068e-01 -8.58985901e-01 -5.02444088e-01 -1.39004242e+00
-1.03817955e-01 -5.72574914e-01 -4.36187442e-03 -1.32488692e+00
-1.78068116e-01 -9.33146894e-01 -2.86910385e-01 4.65231240e-01
2.53302187e-01 5.33761501e-01 4.02468145e-01 3.29744071e-01
-5.34884512e-01 2.26341680e-01 7.04571962e-01 2.98901945e-02
-2.66442776e-01 -4.48184520e-01 -3.95513535e-01 5.61246216e-01
9.56310749e-01 -3.57687861e-01 -5.95064223e-01 -1.00989807e+00
1.92153066e-01 5.42887934e-02 2.67195463e-01 -1.53002143e+00
5.80479980e-01 -7.12519279e-03 3.55807722e-01 -5.84521830e-01
5.56613863e-01 -7.91943967e-01 1.17891803e-01 5.56648254e-01
-5.20519733e-01 4.53877419e-01 6.09645367e-01 2.18001500e-01
1.19467691e-01 1.10824890e-01 9.32656050e-01 4.04209822e-01
-6.44449174e-01 3.17144424e-01 -2.19434649e-01 7.61253685e-02
3.63190860e-01 -2.28629962e-01 -2.99740195e-01 -2.15825945e-01
-2.54822940e-01 4.03829843e-01 4.73734379e-01 -6.41574562e-02
4.60196763e-01 -1.14517963e+00 -3.40795755e-01 2.32292041e-01
-2.47409269e-01 -6.66079819e-02 2.38005310e-01 6.62146807e-01
-9.72117424e-01 6.14112079e-01 -2.61917729e-02 -7.19546020e-01
-1.15844047e+00 3.84267420e-02 6.76516235e-01 -2.89449692e-01
-3.21467989e-04 1.00165558e+00 -2.67710060e-01 -9.46199894e-03
6.02204621e-01 -8.78935337e-01 3.54851514e-01 -2.66850442e-01
5.08859515e-01 7.44447768e-01 4.98790205e-01 -4.06325310e-01
-4.55018461e-01 4.13110912e-01 -1.56901121e-01 -3.36780936e-01
8.27096343e-01 -1.25195682e-01 1.44269228e-01 1.70948192e-01
1.67150033e+00 -4.71085042e-01 -1.48832226e+00 -1.53545216e-01
-7.93864429e-02 -2.63348699e-01 5.89534819e-01 -8.17175746e-01
-1.28857017e+00 8.42166483e-01 9.74518120e-01 5.89177236e-02
1.28870094e+00 -4.67639208e-01 9.61507738e-01 3.23038071e-01
4.78630960e-01 -1.13361347e+00 -2.32543692e-01 5.27836084e-01
3.61252993e-01 -8.51831317e-01 1.14801563e-01 -3.05500269e-01
-1.09154142e-01 9.36469138e-01 8.49933505e-01 2.12537143e-02
8.23525071e-01 7.07976460e-01 7.44430199e-02 3.70086953e-02
-1.14453101e+00 3.55846494e-01 1.92254156e-01 2.03278095e-01
3.17578912e-01 2.90779639e-02 -2.10135251e-01 3.19528133e-01
-5.69137037e-01 2.97554620e-02 3.59972358e-01 9.32806432e-01
-3.79170299e-01 -7.56303966e-01 -2.07235694e-01 9.60129321e-01
-5.32699704e-01 -4.29875493e-01 1.83318168e-01 6.34111106e-01
-1.44066468e-01 9.65826631e-01 5.34635305e-01 -7.74810255e-01
2.23929346e-01 -1.05467752e-01 2.99434364e-01 -1.98405161e-01
-6.92270994e-01 1.75294708e-02 2.29018763e-01 -7.30156004e-01
1.85444728e-01 -3.73660773e-01 -1.54756105e+00 -9.08510983e-01
-2.02027589e-01 2.22556014e-02 1.08336163e+00 6.09931111e-01
7.65353978e-01 5.28137326e-01 3.12551141e-01 -6.50070429e-01
-1.05755007e+00 -6.34589136e-01 -3.09500337e-01 -2.42882103e-01
4.77143764e-01 -4.65722531e-01 -2.96295971e-01 -2.36250192e-01] | [8.659713745117188, 2.983760118484497] |
94c3d672-f5f8-48a2-a422-f4373868d5d1 | dsfnet-dual-space-fusion-network-for-1 | 2305.11522 | null | https://arxiv.org/abs/2305.11522v1 | https://arxiv.org/pdf/2305.11522v1.pdf | DSFNet: Dual Space Fusion Network for Occlusion-Robust 3D Dense Face Alignment | Sensitivity to severe occlusion and large view angles limits the usage scenarios of the existing monocular 3D dense face alignment methods. The state-of-the-art 3DMM-based method, directly regresses the model's coefficients, underutilizing the low-level 2D spatial and semantic information, which can actually offer cues for face shape and orientation. In this work, we demonstrate how modeling 3D facial geometry in image and model space jointly can solve the occlusion and view angle problems. Instead of predicting the whole face directly, we regress image space features in the visible facial region by dense prediction first. Subsequently, we predict our model's coefficients based on the regressed feature of the visible regions, leveraging the prior knowledge of whole face geometry from the morphable models to complete the invisible regions. We further propose a fusion network that combines the advantages of both the image and model space predictions to achieve high robustness and accuracy in unconstrained scenarios. Thanks to the proposed fusion module, our method is robust not only to occlusion and large pitch and roll view angles, which is the benefit of our image space approach, but also to noise and large yaw angles, which is the benefit of our model space method. Comprehensive evaluations demonstrate the superior performance of our method compared with the state-of-the-art methods. On the 3D dense face alignment task, we achieve 3.80% NME on the AFLW2000-3D dataset, which outperforms the state-of-the-art method by 5.5%. Code is available at https://github.com/lhyfst/DSFNet. | ['Robby T. Tan', 'Mohan Kankanhalli', 'Yu Cheng', 'Bo wang', 'Heyuan Li'] | 2023-05-19 | dsfnet-dual-space-fusion-network-for | http://openaccess.thecvf.com//content/CVPR2023/html/Li_DSFNet_Dual_Space_Fusion_Network_for_Occlusion-Robust_3D_Dense_Face_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_DSFNet_Dual_Space_Fusion_Network_for_Occlusion-Robust_3D_Dense_Face_CVPR_2023_paper.pdf | cvpr-2023-1 | ['face-alignment'] | ['computer-vision'] | [-2.12459445e-01 7.38039985e-02 -1.02877930e-01 -4.46357667e-01
-3.25764447e-01 -2.60653675e-01 4.33288753e-01 -7.83081949e-01
8.76636878e-02 2.25321189e-01 2.33849645e-01 4.20869663e-02
2.48478413e-01 -5.78682840e-01 -5.75218499e-01 -7.12946594e-01
4.53586876e-01 4.14506555e-01 -2.27500096e-01 -2.13541448e-01
-3.55550274e-02 8.55430484e-01 -1.61869967e+00 -4.56662737e-02
6.30768538e-01 1.23827279e+00 5.30640222e-02 1.51293263e-01
-3.86497527e-02 2.14516401e-01 1.36753786e-02 -4.62797284e-01
5.00124931e-01 -1.22527309e-01 -2.04101607e-01 3.93451750e-01
1.05419707e+00 -5.38410962e-01 -5.68804383e-01 9.53894079e-01
5.93002677e-01 -3.24326545e-01 3.69769335e-01 -1.10216975e+00
-4.37872529e-01 -3.64977270e-01 -9.18121636e-01 -3.39194328e-01
4.23712164e-01 1.41029924e-01 6.73321307e-01 -1.48564231e+00
8.21642578e-01 1.42957461e+00 5.26469588e-01 7.33734608e-01
-1.07065749e+00 -1.07899725e+00 4.81873393e-01 4.29982208e-02
-1.58627629e+00 -9.96994317e-01 1.09389889e+00 -4.64240342e-01
7.79313326e-01 6.27715215e-02 7.84768462e-01 8.36497784e-01
9.76145640e-02 5.35675526e-01 1.06607914e+00 -2.45516479e-01
-1.22009411e-01 -1.25785798e-01 -4.34254527e-01 1.15896130e+00
3.90387587e-02 3.34579319e-01 -6.97779417e-01 3.55809964e-02
1.00461972e+00 3.11868578e-01 -3.14272791e-01 -7.51357138e-01
-6.96659088e-01 4.92360830e-01 5.38955748e-01 -5.16456105e-02
-2.36453623e-01 -1.57637954e-01 -2.58149743e-01 -1.31837660e-02
8.87917995e-01 -7.84471855e-02 -6.19271994e-01 1.90579116e-01
-9.22785163e-01 2.25553051e-01 5.17095327e-01 1.09287226e+00
9.47695792e-01 2.59720683e-01 2.14551166e-01 8.77163291e-01
7.95191944e-01 9.12779093e-01 -1.48470246e-03 -1.09189296e+00
6.25893950e-01 7.30129838e-01 -9.77353528e-02 -1.15312064e+00
-5.18504262e-01 -4.06939924e-01 -7.92884767e-01 2.73872197e-01
2.35165715e-01 -1.00336485e-02 -1.12891448e+00 1.79916835e+00
7.95565844e-01 1.97358325e-01 -2.01742873e-01 1.04093564e+00
9.25152779e-01 3.83771420e-01 -5.47697067e-01 -3.94571543e-01
1.22574735e+00 -9.09379065e-01 -6.84900224e-01 -4.18796331e-01
2.97863960e-01 -1.13246703e+00 6.71860397e-01 1.39172748e-01
-1.08273959e+00 -5.06637931e-01 -8.86324584e-01 -1.81531265e-01
8.74241143e-02 3.46608847e-01 5.17607391e-01 5.57853937e-01
-1.08277643e+00 1.43225208e-01 -9.71856117e-01 -2.81810611e-01
6.34936869e-01 5.69012403e-01 -6.86775506e-01 -4.75461394e-01
-6.00268900e-01 7.89948881e-01 -3.96252066e-01 4.98360127e-01
-6.71866655e-01 -8.86081159e-01 -1.10008109e+00 -8.80460218e-02
3.62032920e-01 -7.89615631e-01 9.27206099e-01 -6.00746930e-01
-1.69893479e+00 9.46629345e-01 -7.15441704e-01 2.86834896e-01
5.35693765e-01 -3.37424904e-01 -2.72670329e-01 1.32559389e-01
-1.59969896e-01 7.67791867e-01 9.18588400e-01 -1.38715756e+00
-1.81875616e-01 -9.25245464e-01 -8.88926312e-02 3.91952515e-01
-2.87994355e-01 -2.36146510e-01 -9.64748740e-01 -4.46285397e-01
5.67592502e-01 -1.09730828e+00 -2.48382706e-02 5.25713861e-01
-2.44175717e-01 1.72279522e-01 9.99711275e-01 -8.66489887e-01
8.02166045e-01 -2.07820296e+00 2.91444272e-01 1.47671506e-01
2.72442997e-01 2.23990455e-01 -2.37236753e-01 1.03910021e-01
-2.83480808e-03 -1.96483091e-01 3.55953304e-03 -8.34996045e-01
-1.32293060e-01 5.56788296e-02 -3.99973206e-02 7.34002829e-01
1.36915401e-01 7.79347122e-01 -3.14654320e-01 -3.51240337e-01
5.11609077e-01 9.23641920e-01 -7.27416694e-01 3.29458594e-01
-6.77672923e-02 6.77496076e-01 -4.29701477e-01 9.44966376e-01
1.26008964e+00 -1.17966406e-01 3.18802655e-01 -6.03536963e-01
-7.57345706e-02 2.72001829e-02 -1.00210023e+00 2.03070664e+00
-6.55630112e-01 2.96287030e-01 4.22179073e-01 -5.73565841e-01
1.05934811e+00 3.39159101e-01 5.77259600e-01 -8.61995935e-01
2.10453141e-02 1.65934697e-01 -2.83885837e-01 -2.56096423e-01
-5.05087227e-02 -1.25166282e-01 3.57961059e-01 5.02589382e-02
1.91934094e-01 -2.00214297e-01 -4.22503710e-01 -7.84699991e-02
5.06607533e-01 5.81568837e-01 9.31059793e-02 -3.38687114e-02
6.07505500e-01 -6.66544020e-01 8.59275579e-01 -1.40756309e-01
8.00415426e-02 7.50978112e-01 2.98222750e-01 -6.09028637e-01
-1.00249362e+00 -9.80244517e-01 -2.28944972e-01 5.09878218e-01
2.16012105e-01 -5.01296282e-01 -7.03038394e-01 -6.13765419e-01
1.99299857e-01 1.11854210e-01 -4.84337360e-01 -3.39964591e-02
-7.32533991e-01 -5.94847381e-01 -9.54502895e-02 3.38436991e-01
5.83619058e-01 -3.89151096e-01 -4.74101380e-02 -1.83963746e-01
-9.74555984e-02 -1.50741875e+00 -5.03580570e-01 -6.27903759e-01
-8.22498500e-01 -9.37866390e-01 -5.29911220e-01 -5.07321179e-01
1.03354919e+00 4.61271822e-01 8.47080648e-01 2.78015405e-01
-2.55549043e-01 3.09572905e-01 7.13607222e-02 -1.37247622e-01
-3.44520807e-02 -8.11917111e-02 3.55950266e-01 2.51605928e-01
1.35162354e-01 -8.57202590e-01 -8.69877994e-01 6.69685245e-01
-3.65117133e-01 4.76174265e-01 5.29390872e-01 7.07240760e-01
6.66029334e-01 -4.11052495e-01 1.27019122e-01 -6.62999392e-01
-4.36359972e-01 -1.96276218e-01 -8.48283410e-01 1.30654555e-02
-6.42269731e-01 -1.73200190e-01 3.08840573e-01 -8.77902880e-02
-1.26074469e+00 3.58327210e-01 -3.60682786e-01 -7.50762045e-01
-8.86014029e-02 3.21472180e-03 -8.90900016e-01 -4.61089909e-01
1.16928630e-01 2.02448759e-02 3.99908870e-01 -7.21323013e-01
2.96900004e-01 4.90964800e-01 1.48024797e-01 -4.08519715e-01
1.09098566e+00 8.13143432e-01 3.85788232e-01 -9.16525483e-01
-9.57211614e-01 -2.67037779e-01 -7.74242640e-01 -3.40043128e-01
7.13607252e-01 -1.38582337e+00 -6.52032793e-01 6.12509012e-01
-1.20528245e+00 -3.60483713e-02 2.41326585e-01 4.63554621e-01
-4.51702446e-01 3.64859790e-01 -4.93044645e-01 -6.28841162e-01
-2.65601516e-01 -1.35467017e+00 1.37030637e+00 2.69307524e-01
1.62027285e-01 -7.50110686e-01 -3.05613488e-01 7.86647320e-01
2.24970698e-01 2.43495181e-01 6.72278941e-01 8.75227749e-02
-8.65887463e-01 -1.05199166e-01 -2.04337120e-01 2.70257264e-01
2.65088558e-01 -1.34004774e-02 -1.22839749e+00 -3.94515246e-01
1.16815552e-01 -5.97184859e-02 6.36530519e-01 4.38637406e-01
1.00076020e+00 -2.93728709e-01 -2.81627089e-01 1.13125789e+00
1.34715831e+00 -8.94176811e-02 5.15890956e-01 -1.24402314e-01
1.19144464e+00 6.96505606e-01 6.83729708e-01 4.66156363e-01
6.11568332e-01 1.26128769e+00 6.95230901e-01 -4.45084363e-01
-3.48942101e-01 -4.39893216e-01 2.93885350e-01 9.25140738e-01
-2.75203675e-01 9.11471099e-02 -6.30392730e-01 1.03094332e-01
-1.74103737e+00 -6.11708164e-01 1.59832120e-01 2.16339517e+00
5.47347903e-01 -2.80663729e-01 -1.63140193e-01 -2.31708825e-01
4.82556999e-01 3.68631959e-01 -5.79839587e-01 -7.36162364e-02
-1.81594975e-02 1.77915901e-01 2.36058772e-01 6.95252597e-01
-9.05488431e-01 1.18200338e+00 5.11669207e+00 7.50335395e-01
-1.19191885e+00 -1.62198748e-02 5.96002817e-01 -4.22906399e-01
-3.67305130e-01 4.63751471e-03 -1.20494235e+00 3.34983289e-01
3.90042692e-01 3.55107576e-01 5.12186885e-01 5.83772421e-01
2.75006056e-01 1.92915916e-01 -1.01229465e+00 1.11193919e+00
4.09677744e-01 -1.17675996e+00 6.93772808e-02 5.82605243e-01
8.50581169e-01 -9.55756530e-02 2.68105388e-01 -1.34239838e-01
-2.34114453e-01 -1.10996854e+00 6.55016005e-01 5.35306215e-01
1.03167629e+00 -6.84780121e-01 5.55351377e-01 2.63082564e-01
-1.20984745e+00 9.82102975e-02 -3.75180632e-01 -3.08773071e-02
2.00185105e-01 7.79759228e-01 -5.27554452e-01 5.94743788e-01
6.23674333e-01 9.46568668e-01 -3.07879508e-01 5.39642155e-01
-3.72378558e-01 2.61736929e-01 -4.38881785e-01 6.74329758e-01
-1.96570233e-01 -4.34271634e-01 4.83296186e-01 6.41879976e-01
5.50231814e-01 1.61552966e-01 6.60595894e-02 7.59775221e-01
-1.61682442e-01 1.12080149e-01 -7.03236401e-01 3.89234155e-01
4.53869343e-01 1.41550267e+00 -2.40858912e-01 8.74050334e-02
-6.77899778e-01 8.25118184e-01 4.20400560e-01 5.02522469e-01
-6.67486310e-01 3.18143159e-01 1.11703861e+00 3.53362501e-01
6.02373779e-01 -4.78065550e-01 -1.36709884e-01 -1.34815538e+00
4.27372605e-01 -8.37995529e-01 -1.07257292e-01 -6.72629952e-01
-9.43380237e-01 5.75685143e-01 -1.97723210e-01 -1.20301926e+00
-1.00923724e-01 -7.69662261e-01 -3.02420348e-01 9.80263889e-01
-1.76900530e+00 -1.65238762e+00 -5.15795112e-01 5.77436328e-01
3.66809934e-01 -2.70166993e-01 7.91878700e-01 3.45767856e-01
-5.96132636e-01 6.72692001e-01 -2.09412560e-01 2.43146271e-02
7.92896569e-01 -6.57744944e-01 4.55438852e-01 7.31359422e-01
1.51142508e-01 5.39767981e-01 3.31763804e-01 -5.08459449e-01
-1.75300753e+00 -9.96730387e-01 8.25874925e-01 -3.53587300e-01
2.12079391e-01 -6.58393621e-01 -6.72962487e-01 5.61687410e-01
-1.68519676e-01 4.46787387e-01 5.77910602e-01 1.42269060e-01
-6.35171115e-01 -5.35039067e-01 -1.11757076e+00 5.95881939e-01
1.45751905e+00 -5.37190437e-01 -4.89051379e-02 1.85388044e-01
5.35443902e-01 -8.09451103e-01 -9.64154482e-01 7.65740573e-01
1.00457084e+00 -1.08840907e+00 1.17212522e+00 -3.07132065e-01
3.66164982e-01 -4.08225566e-01 -3.84904385e-01 -1.19853461e+00
-1.48232162e-01 -5.27798474e-01 -3.24347436e-01 1.30594110e+00
2.32772321e-01 -6.48781896e-01 9.46406186e-01 5.21125853e-01
-7.22418576e-02 -9.50245023e-01 -1.25697064e+00 -4.47249413e-01
-1.37445897e-01 -2.87838370e-01 8.42579722e-01 7.82894671e-01
-4.58604127e-01 2.66061455e-01 -5.63071311e-01 4.64404911e-01
6.43987954e-01 4.42280293e-01 1.02476561e+00 -1.13625860e+00
-6.12059683e-02 -1.12832315e-01 -6.47498786e-01 -1.34872091e+00
5.45309901e-01 -7.36559451e-01 -2.86807686e-01 -1.17738116e+00
1.55355960e-01 -4.03014213e-01 1.47986650e-01 4.78967875e-01
1.79374032e-02 5.57429433e-01 4.65975314e-01 1.41317338e-01
-8.75940919e-02 8.45917225e-01 1.53396940e+00 1.47244707e-01
-1.49686392e-02 -9.30679142e-02 -6.09905660e-01 8.91641200e-01
5.41187406e-01 -1.66867062e-01 -3.45014542e-01 -7.72680044e-01
9.87334270e-03 1.39494851e-01 4.52652603e-01 -7.47296929e-01
2.19446458e-02 -2.38540784e-01 7.73616612e-01 -5.31023741e-01
9.89026666e-01 -1.02333403e+00 2.35666648e-01 1.96523204e-01
3.17344517e-01 -7.89702088e-02 2.47139141e-01 3.77544075e-01
-1.75233170e-01 3.87064546e-01 8.35821629e-01 4.23764363e-02
-3.42311472e-01 9.42709029e-01 3.22046787e-01 -2.26156637e-01
9.08001125e-01 -2.48198867e-01 -2.42130712e-01 -4.31740224e-01
-5.90900004e-01 1.25592649e-01 9.79509294e-01 5.05866110e-01
8.51388633e-01 -1.50215662e+00 -6.62418544e-01 8.22267473e-01
4.66539860e-02 2.33126104e-01 4.82207984e-01 9.96247053e-01
-4.19498980e-01 4.02996868e-01 -2.07665384e-01 -8.30556154e-01
-1.51858199e+00 3.47507089e-01 4.37729239e-01 7.09939525e-02
-6.02584720e-01 7.47326791e-01 6.65224195e-01 -7.01357186e-01
-7.56266415e-02 2.57158596e-02 -4.45824899e-02 -2.05333024e-01
3.70344639e-01 9.00654718e-02 1.49913356e-01 -1.07212615e+00
-5.27423263e-01 1.45520890e+00 -1.11583479e-01 7.16889277e-02
1.35547209e+00 -3.05376112e-01 -8.89226049e-02 -1.11191675e-01
1.37296689e+00 2.46218383e-01 -1.46573615e+00 -2.78178334e-01
-6.64716065e-01 -9.02251184e-01 1.82074040e-01 -5.46140730e-01
-1.59868968e+00 1.17354572e+00 6.59007430e-01 -6.87464297e-01
1.20958281e+00 1.47168590e-02 6.88384295e-01 -1.77578703e-01
4.58496273e-01 -6.58578455e-01 -2.59017833e-02 3.03227484e-01
1.03719294e+00 -1.25656533e+00 3.82061213e-01 -1.02869105e+00
-4.24186528e-01 1.03808320e+00 8.75897825e-01 1.23960823e-01
8.72321606e-01 2.15820014e-01 1.86597839e-01 -2.91208297e-01
-6.17391169e-01 3.73956822e-02 5.83598256e-01 4.83821124e-01
3.88547122e-01 -1.53938392e-02 1.91005215e-01 1.87794849e-01
-2.76704580e-01 -1.81439564e-01 -9.60555580e-03 5.38418710e-01
-7.76717365e-02 -1.13039565e+00 -3.46333325e-01 8.26603249e-02
-2.18375057e-01 2.29571536e-02 -2.77978957e-01 8.35901499e-01
2.60119796e-01 7.20750391e-01 1.75674275e-01 -4.42538142e-01
3.91661227e-01 -9.45348516e-02 7.88658082e-01 -5.48214078e-01
8.75495151e-02 3.95945191e-01 -6.93846121e-03 -9.82403517e-01
-4.64737386e-01 -6.87767446e-01 -9.55162287e-01 -4.20612365e-01
-3.61702800e-01 -3.17906588e-01 7.47658670e-01 8.14388633e-01
6.81511819e-01 -7.96731636e-02 9.36612427e-01 -1.13365722e+00
-2.88409948e-01 -8.30028117e-01 -4.38078642e-01 9.70796272e-02
4.42952216e-01 -1.05948758e+00 -3.51347446e-01 -1.88555643e-01] | [13.257974624633789, 0.17709293961524963] |
ac6509b8-7938-4a4a-8d21-8cb2eb58b698 | iitk-at-semeval-2021-task-10-source-free | null | null | https://aclanthology.org/2021.semeval-1.53 | https://aclanthology.org/2021.semeval-1.53.pdf | IITK at SemEval-2021 Task 10: Source-Free Unsupervised Domain Adaptation using Class Prototypes | Recent progress in deep learning has primarily been fueled by the availability of large amounts of annotated data that is obtained from highly expensive manual annotating pro-cesses. To tackle this issue of availability of annotated data, a lot of research has been done on unsupervised domain adaptation that tries to generate systems for an unlabelled target domain data, given labeled source domain data. However, the availability of annotated or labelled source domain dataset can{'}t always be guaranteed because of data-privacy issues. This is especially the case with medical data, as it may contain sensitive information of the patients. Source-free domain adaptation (SFDA) aims to resolve this issue by us-ing models trained on the source data instead of using the original annotated source data. In this work, we try to build SFDA systems for semantic processing by specifically focusing on the negation detection subtask of the SemEval2021 Task 10. We propose two approaches -ProtoAUGandAdapt-ProtoAUGthat use the idea of self-entropy to choose reliable and high confidence samples, which are then used for data augmentation and subsequent training of the models. Our methods report an improvement of up to 7{\%} in F1 score over the baseline for the Negation Detection subtask. | ['Ashutosh Modi', 'Vaibhav Jindal', 'Priyanshu Gupta', 'Nidhi Hegde', 'Jinang Shah', 'Harshit Kumar'] | 2021-08-01 | null | null | null | semeval-2021 | ['source-free-domain-adaptation', 'negation-detection'] | ['computer-vision', 'natural-language-processing'] | [ 4.83942091e-01 7.66966939e-01 -2.41238654e-01 -7.74321914e-01
-8.66679072e-01 -2.26099372e-01 5.52036941e-01 6.61656737e-01
-8.78864646e-01 1.17098737e+00 1.33268237e-01 2.79689468e-02
1.91518337e-01 -7.45498776e-01 -5.33701420e-01 -6.54091358e-01
3.36236209e-01 8.65573943e-01 2.16599941e-01 -2.73943752e-01
-1.50373548e-01 8.43251720e-02 -1.39081943e+00 6.38540089e-01
8.74146879e-01 1.08291495e+00 -3.37346867e-02 2.88010538e-01
-3.25160384e-01 6.75844491e-01 -7.72713602e-01 -5.81934810e-01
1.48811311e-01 -6.37020469e-01 -9.68097687e-01 -3.31465863e-02
-2.28924546e-02 -8.20878670e-02 2.42424697e-01 1.22702026e+00
4.43890005e-01 1.29826916e-02 6.64480150e-01 -1.29335070e+00
-3.36958349e-01 7.25973606e-01 -8.36802572e-02 1.35666937e-01
7.50262737e-02 -2.34953403e-01 7.37732470e-01 -5.98648548e-01
9.75124538e-01 9.22124505e-01 4.53659534e-01 1.00999451e+00
-1.24241197e+00 -6.21948302e-01 -3.05183023e-01 1.49962664e-01
-1.21252203e+00 -5.65291762e-01 8.47753227e-01 -3.36913466e-01
7.65439928e-01 5.52624352e-02 1.24112830e-01 1.49989426e+00
-8.97516534e-02 8.06478620e-01 1.07295752e+00 -7.11834073e-01
5.92634439e-01 7.60705471e-01 8.85013938e-02 1.99655414e-01
3.19653749e-01 1.19438708e-01 -3.36443067e-01 -2.61581361e-01
3.54231000e-01 -4.17518407e-01 -2.13787854e-01 -5.45596600e-01
-1.05017877e+00 9.68599796e-01 3.29942435e-01 4.48395908e-01
-2.94790268e-01 -4.74403560e-01 8.72562468e-01 4.88707751e-01
7.63141870e-01 6.18508220e-01 -8.26080680e-01 -4.25377153e-02
-8.72453034e-01 3.46999228e-01 9.53050852e-01 9.13022757e-01
3.01239848e-01 -1.92197889e-01 -9.80483741e-02 9.02504325e-01
1.30194262e-01 2.27081209e-01 7.94439554e-01 -5.07326007e-01
6.29412293e-01 7.35084772e-01 2.83798635e-01 -6.84639156e-01
-4.41897094e-01 -1.39177412e-01 -1.00564289e+00 3.42604518e-02
6.36677802e-01 -3.81587446e-01 -1.07038820e+00 1.82282543e+00
4.68465060e-01 -2.06162781e-01 6.31180882e-01 7.72227883e-01
8.59365225e-01 3.22169542e-01 3.88490468e-01 -1.94655418e-01
1.33695602e+00 -5.38822889e-01 -9.93939281e-01 -2.09110886e-01
7.84489572e-01 -3.84608626e-01 8.65407944e-01 3.98858756e-01
-6.67389154e-01 -4.04923081e-01 -1.01599741e+00 -2.38955058e-02
-6.84917212e-01 4.38920110e-02 1.86691254e-01 6.85471535e-01
-7.91059911e-01 5.92627764e-01 -7.32716382e-01 -4.06792909e-01
7.72795379e-01 3.95219952e-01 -6.90153420e-01 -1.41657248e-01
-1.60039103e+00 9.76228952e-01 1.13790560e+00 -2.37358347e-01
-4.74677026e-01 -5.35333216e-01 -9.20612812e-01 -4.31830734e-02
4.71746862e-01 -3.97927344e-01 1.24826384e+00 -1.46474576e+00
-1.25134492e+00 1.23305655e+00 2.36208752e-01 -8.87291670e-01
7.66668797e-01 -9.66667384e-02 -6.95527911e-01 3.63286063e-02
4.23941314e-02 6.52724028e-01 7.50048220e-01 -9.85801935e-01
-5.51090479e-01 -4.49825585e-01 -2.59157300e-01 5.26814535e-02
-5.25329471e-01 1.01621345e-01 7.79987276e-02 -7.08254218e-01
-3.11468720e-01 -7.31617570e-01 -3.15405667e-01 -1.35234118e-01
-6.09086037e-01 -2.51212358e-01 6.97092533e-01 -6.62432194e-01
8.89826775e-01 -2.21066093e+00 -4.70262729e-02 2.08622962e-01
1.20261451e-02 8.11109543e-01 -4.60678153e-03 7.45843053e-02
-5.13945043e-01 -6.37825578e-04 -4.77004230e-01 -2.69742727e-01
-1.73978060e-01 3.25603962e-01 -2.09727913e-01 1.85975015e-01
4.64549780e-01 5.25395393e-01 -9.13943708e-01 -6.34771109e-01
-3.53868008e-02 3.40858579e-01 -5.98257244e-01 3.45050007e-01
-5.35358906e-01 5.64966202e-01 -4.08921540e-01 4.12385136e-01
5.87797284e-01 -2.23392006e-02 2.18932077e-01 3.65285650e-02
3.25585306e-01 3.08004498e-01 -1.10026145e+00 1.66657472e+00
-1.31300986e-01 1.95969239e-01 -4.66100872e-04 -1.31687343e+00
1.13451898e+00 6.43243194e-01 6.23921394e-01 -5.70902884e-01
4.42128628e-01 4.99428630e-01 -8.08917880e-02 -4.59210277e-01
3.24128032e-01 -5.26759148e-01 -2.52853692e-01 5.63722290e-02
2.83211946e-01 -6.84692189e-02 8.44455659e-02 -1.70199603e-01
1.09910822e+00 6.48431778e-02 5.64221978e-01 -3.33331048e-01
7.75884688e-01 3.12829763e-01 6.84796154e-01 3.52782845e-01
-3.41863334e-01 5.04797518e-01 5.79291344e-01 -4.38697010e-01
-1.19828689e+00 -6.32363558e-01 -3.63633513e-01 7.61426210e-01
-3.30616564e-01 -2.65034795e-01 -1.02725649e+00 -1.25359023e+00
-1.51739582e-01 1.04172385e+00 -6.71182156e-01 -5.24708390e-01
-3.61476660e-01 -8.32206607e-01 6.88677192e-01 5.08085489e-01
5.68726540e-01 -1.32530236e+00 -7.67327428e-01 3.89965296e-01
-3.64817321e-01 -1.17813683e+00 -3.32258479e-03 6.50462747e-01
-8.09316397e-01 -1.01024151e+00 -5.41464984e-01 -6.63949430e-01
6.75393522e-01 -7.35613883e-01 1.14928555e+00 -3.93551290e-01
-6.37525022e-02 -1.54134184e-01 -5.62931776e-01 -9.76653934e-01
-8.73839140e-01 2.88359404e-01 1.48389060e-02 -4.59298119e-03
1.06051815e+00 -2.00145483e-01 -1.30539626e-01 1.39713079e-01
-1.14315295e+00 -1.58011913e-01 4.89111632e-01 1.08418894e+00
7.15060651e-01 5.45958653e-02 9.46353197e-01 -1.39374161e+00
5.79118013e-01 -6.31361306e-01 -4.81986582e-01 -9.66003910e-03
-6.68130338e-01 2.45301634e-01 8.14933836e-01 -3.35329503e-01
-1.11388421e+00 3.61139804e-01 -4.35061932e-01 -3.14394176e-01
-7.17509747e-01 3.44812870e-01 -5.10197103e-01 4.16976362e-01
1.05338335e+00 -1.09045163e-01 5.91464303e-02 -5.95221400e-01
1.65788651e-01 1.12131357e+00 5.25269866e-01 -3.03197920e-01
4.52219129e-01 1.92495346e-01 -2.90922582e-01 -6.22155249e-01
-1.05188942e+00 -3.26722354e-01 -7.41431594e-01 3.48546147e-01
9.68459845e-01 -6.89077020e-01 -5.39115965e-02 3.94737512e-01
-1.05457497e+00 -1.38244957e-01 -7.79317439e-01 2.85989910e-01
-5.44048309e-01 8.96013826e-02 -2.19887078e-01 -6.58434331e-01
-4.74409759e-01 -8.35850596e-01 9.60788488e-01 -1.00236267e-01
-3.91912341e-01 -9.13195074e-01 1.31761029e-01 3.28933358e-01
3.38679850e-01 4.45446789e-01 1.01215148e+00 -1.69233537e+00
9.95504335e-02 -4.97809291e-01 -5.50555252e-02 8.75176191e-01
2.26535797e-01 -7.27760553e-01 -1.10348320e+00 -1.41556054e-01
3.17253679e-01 -6.41793370e-01 6.31814718e-01 6.47173375e-02
1.16660833e+00 -4.09952134e-01 -3.24564755e-01 1.40927538e-01
1.29235101e+00 1.79959655e-01 5.46960890e-01 3.15083385e-01
3.85008812e-01 8.33360553e-01 9.80616570e-01 4.80631918e-01
2.13636696e-01 6.60006106e-01 3.83271605e-01 -1.97907276e-02
1.06932849e-01 -9.50554162e-02 9.45318937e-02 1.92139596e-01
1.78384721e-01 -3.99781525e-01 -1.05300391e+00 8.08940887e-01
-1.73661971e+00 -7.34473169e-01 -1.01033881e-01 2.25295830e+00
1.35451841e+00 3.04985821e-01 1.60026520e-01 4.24790859e-01
6.76441848e-01 -3.56001437e-01 -6.12521589e-01 -4.02566552e-01
4.26989086e-02 2.92592168e-01 4.87029523e-01 1.16873108e-01
-1.45601547e+00 7.41462231e-01 5.43879366e+00 6.53678417e-01
-9.73222911e-01 1.68403268e-01 7.78041065e-01 4.20872942e-02
-3.15882117e-02 -3.62393349e-01 -8.50800633e-01 6.16466582e-01
1.43675041e+00 -7.31380433e-02 -1.93367302e-01 1.18034196e+00
-2.90799811e-02 -9.37851816e-02 -1.25829601e+00 8.39815617e-01
1.55934021e-01 -9.80115652e-01 -1.50538877e-01 9.53094810e-02
4.65160817e-01 2.66929250e-02 -2.38787264e-01 4.05496836e-01
3.54269773e-01 -8.86142135e-01 3.95436525e-01 2.34661222e-01
7.46384621e-01 -9.69808519e-01 1.34482741e+00 5.88206291e-01
-4.81448889e-01 -1.06417000e-01 -3.98019314e-01 3.77178073e-01
-4.17018235e-02 8.88090789e-01 -1.47306967e+00 5.43545604e-01
7.22707331e-01 4.69409823e-01 -3.71477246e-01 8.77382040e-01
-1.96835950e-01 5.36804795e-01 -3.52154702e-01 1.12179413e-01
2.94516832e-02 1.88691914e-01 4.36714530e-01 1.19830596e+00
1.06707983e-01 -7.39228204e-02 1.97082251e-01 8.11307311e-01
-2.29915172e-01 4.21574503e-01 -7.25251853e-01 -2.49603525e-01
1.64446115e-01 9.90228236e-01 -5.69614172e-01 -4.19948220e-01
-2.15867966e-01 1.06947291e+00 2.17088401e-01 -1.19922362e-01
-5.97437143e-01 -4.56465393e-01 4.77991253e-01 2.90555418e-01
2.44811788e-01 2.68076271e-01 -2.87947059e-01 -1.09569263e+00
3.53668481e-02 -9.79033709e-01 8.09820712e-01 -3.50061476e-01
-1.45202756e+00 8.32331419e-01 1.25596911e-01 -1.04169929e+00
-6.51535690e-01 -5.76133847e-01 2.70123947e-02 8.29620779e-01
-1.61172795e+00 -9.50761318e-01 -6.31926507e-02 7.45284736e-01
5.12118220e-01 -3.16468418e-01 1.24844801e+00 3.97325814e-01
-3.20765615e-01 7.62181222e-01 1.69530824e-01 4.22349453e-01
1.04008698e+00 -1.40004969e+00 1.40985936e-01 6.49584234e-01
-9.53171626e-02 1.62656099e-01 8.57893348e-01 -7.91306973e-01
-6.76295161e-01 -1.36650503e+00 1.29474950e+00 -6.58887565e-01
3.09571564e-01 -4.50839579e-01 -1.24258077e+00 5.93044043e-01
-5.97812794e-02 2.04733104e-01 8.86590004e-01 -8.84250402e-02
-2.74808317e-01 -6.27228022e-02 -1.88971519e+00 1.29098609e-01
6.49456501e-01 -2.19849408e-01 -9.52010155e-01 3.79898667e-01
6.25180542e-01 -4.24089700e-01 -9.21608806e-01 4.84768987e-01
-1.53734848e-01 -7.09100008e-01 7.63139069e-01 -8.17196488e-01
4.92821872e-01 -2.87728384e-03 -5.92980534e-02 -1.37237883e+00
3.09664547e-01 -4.75803390e-02 5.24513423e-02 1.41489840e+00
4.03484911e-01 -3.54204834e-01 9.79337513e-01 9.40496743e-01
4.43858467e-03 -2.74109483e-01 -1.07760096e+00 -7.03954816e-01
2.27181062e-01 -3.92354369e-01 6.09233618e-01 1.30378771e+00
1.87083408e-01 3.37876797e-01 -2.37701446e-01 -1.01112396e-01
4.37998801e-01 -3.21781516e-01 4.69650179e-01 -1.52040517e+00
-1.43670797e-01 6.54252023e-02 -5.25917411e-01 -8.21649730e-02
2.47821942e-01 -1.04389572e+00 2.52528340e-01 -1.21990776e+00
-1.72543019e-01 -4.23627913e-01 -3.80179882e-01 9.99843836e-01
8.93909112e-03 1.17771022e-01 -1.16208024e-01 -1.22901320e-01
-4.60958958e-01 5.27876139e-01 7.31833458e-01 9.27598327e-02
-3.31643194e-01 1.75374411e-02 -6.75949574e-01 6.54509425e-01
8.84312034e-01 -7.68710196e-01 -6.27139449e-01 -1.32950991e-01
-4.48168293e-02 -7.81813562e-02 3.00164163e-01 -9.98006821e-01
-2.38116369e-01 -9.15900543e-02 3.05494070e-01 -3.80085826e-01
2.86328077e-01 -1.21419656e+00 -7.29000568e-02 3.72831494e-01
-5.80359221e-01 -3.67085755e-01 2.85451740e-01 5.31386256e-01
-4.10284758e-01 -4.77524638e-01 1.06425107e+00 -3.17855418e-01
-6.36523485e-01 -6.40028566e-02 -1.70895562e-01 3.05911452e-01
1.17405105e+00 2.56197415e-02 -5.43463305e-02 -8.33416916e-03
-9.69502509e-01 7.03359246e-02 3.66322577e-01 5.41931391e-01
3.60761285e-01 -1.22107613e+00 -8.21385443e-01 3.64973962e-01
4.82838213e-01 3.14577997e-01 1.29241385e-02 4.45418358e-01
-1.65915966e-01 4.07879174e-01 -3.28772485e-01 -3.75706434e-01
-1.12751269e+00 8.39908719e-01 2.38312453e-01 -4.76925433e-01
-5.20976901e-01 8.24828207e-01 -2.10207969e-01 -5.25794387e-01
1.91429257e-01 -3.78905892e-01 -4.44773108e-01 3.02974999e-01
6.95828140e-01 3.88285033e-02 4.78987187e-01 -5.57417154e-01
-4.17403102e-01 -3.13962579e-01 -2.63769597e-01 5.11140414e-02
1.46880424e+00 2.33255327e-01 8.55758879e-03 1.99192703e-01
1.12350917e+00 -4.17756438e-01 -8.71760845e-01 -4.01317596e-01
5.96840739e-01 -2.58396596e-01 -1.16193317e-01 -1.18651175e+00
-9.75707054e-01 7.89965570e-01 8.22542071e-01 1.16986737e-01
1.13936067e+00 -6.18045479e-02 6.85891032e-01 4.26086277e-01
3.16864997e-01 -1.39304686e+00 -2.12536812e-01 3.85861427e-01
7.60481834e-01 -1.59592295e+00 -4.03514594e-01 -2.94530302e-01
-8.95930886e-01 9.29320574e-01 6.30705833e-01 6.03947826e-02
6.71785355e-01 1.90070748e-01 1.02777310e-01 -7.08358362e-02
-4.64917243e-01 -4.26208340e-02 2.32419983e-01 9.02624071e-01
3.75310928e-01 -2.13832576e-02 -3.96978855e-01 1.04105639e+00
-8.87038410e-02 3.53508502e-01 4.24078107e-01 9.61355507e-01
-1.99851632e-01 -1.59538770e+00 -4.09887135e-01 5.85709095e-01
-7.33422518e-01 -6.37811720e-02 -6.17404580e-01 6.67496979e-01
3.54919940e-01 8.28923225e-01 -1.50533661e-01 -1.15285158e-01
4.88851368e-01 5.77639401e-01 -3.25801000e-02 -9.44962680e-01
-4.88903403e-01 -1.42363369e-01 5.08827329e-01 -3.54656160e-01
-5.79542875e-01 -8.01647663e-01 -1.28321886e+00 2.81367362e-01
-1.94085777e-01 4.14795727e-01 6.98890269e-01 1.08001328e+00
3.51276577e-01 3.68377835e-01 1.95685267e-01 -3.44767869e-01
-7.66462207e-01 -9.14242387e-01 -3.44972134e-01 7.41538644e-01
2.71742791e-01 -5.14603734e-01 -3.35114039e-02 1.63953274e-01] | [10.309295654296875, 3.3009631633758545] |
4b8d5c11-dbc3-4578-aa7f-44a4f63be66a | end-to-end-attention-based-distant-speech | 1610.05361 | null | http://arxiv.org/abs/1610.05361v1 | http://arxiv.org/pdf/1610.05361v1.pdf | End-to-end attention-based distant speech recognition with Highway LSTM | End-to-end attention-based models have been shown to be competitive
alternatives to conventional DNN-HMM models in the Speech Recognition Systems.
In this paper, we extend existing end-to-end attention-based models that can be
applied for Distant Speech Recognition (DSR) task. Specifically, we propose an
end-to-end attention-based speech recognizer with multichannel input that
performs sequence prediction directly at the character level. To gain a better
performance, we also incorporate Highway long short-term memory (HLSTM) which
outperforms previous models on AMI distant speech recognition task. | ['Hassan Taherian'] | 2016-10-17 | null | null | null | null | ['distant-speech-recognition'] | ['speech'] | [ 1.64399505e-01 1.74916700e-01 4.31169346e-02 -7.07794309e-01
-1.08268893e+00 -4.57034893e-02 5.05232155e-01 -3.91892284e-01
-5.34286737e-01 4.89114970e-01 5.30859649e-01 -8.87756467e-01
4.92904097e-01 -1.16452426e-01 -4.95122910e-01 -5.06635308e-01
2.05663234e-01 3.86422962e-01 3.52796763e-01 -1.71493649e-01
2.02440187e-01 5.58802366e-01 -1.03860331e+00 5.60047746e-01
5.21105886e-01 7.43192852e-01 7.92819440e-01 1.47901320e+00
6.28672168e-02 1.10603333e+00 -6.37499332e-01 -2.78629363e-01
-2.79822081e-01 -4.14621562e-01 -9.12557960e-01 -9.55158919e-02
1.41053066e-01 -4.60189044e-01 -8.82166028e-01 7.10082233e-01
9.17847931e-01 5.99088132e-01 4.82527941e-01 -5.54267824e-01
-1.31559718e+00 3.94276798e-01 -1.25385880e-01 7.47674108e-01
2.89188296e-01 1.03167862e-01 1.03487313e+00 -1.44355166e+00
1.53889015e-01 1.43935347e+00 1.71442300e-01 8.72960687e-01
-6.40611887e-01 -2.17675164e-01 4.29780126e-01 7.73538709e-01
-1.22249556e+00 -9.45078373e-01 6.17770016e-01 2.29468197e-02
2.01140070e+00 2.95483142e-01 1.30953163e-01 1.37593782e+00
2.66935289e-01 1.24865222e+00 5.15988350e-01 -7.12497175e-01
1.49257079e-01 -4.59013641e-01 2.20408827e-01 2.66409755e-01
-6.62915051e-01 2.85854697e-01 -5.58946311e-01 2.14081526e-01
6.71930909e-01 -1.70178339e-02 -1.76960707e-01 5.78497469e-01
-9.70750332e-01 7.08401561e-01 3.45461965e-01 4.00825322e-01
-7.73495197e-01 3.75018194e-02 5.15958965e-01 3.97388458e-01
6.24024928e-01 -1.97959930e-01 -5.79394758e-01 -5.11344612e-01
-9.57924128e-01 -4.93539035e-01 4.95534807e-01 8.61998320e-01
1.84515402e-01 7.29659915e-01 -5.56953669e-01 1.21839607e+00
5.08582234e-01 3.61110389e-01 9.80381846e-01 -3.54637653e-01
6.91392303e-01 -3.82069647e-01 -2.30028063e-01 -1.51707366e-01
1.20828725e-01 -6.09525084e-01 -8.56325865e-01 7.82990232e-02
-3.53683561e-01 -1.56034604e-01 -1.43704724e+00 1.28965104e+00
-1.98426560e-01 5.24775624e-01 6.14109218e-01 1.05799592e+00
6.62155867e-01 1.38731205e+00 2.78586656e-01 -1.43342733e-01
1.09244442e+00 -1.67681146e+00 -1.14115608e+00 -4.76802170e-01
5.77917159e-01 -8.69580925e-01 9.54395115e-01 3.18160772e-01
-1.34702110e+00 -9.36800957e-01 -7.95907617e-01 -4.95013565e-01
-4.10034090e-01 1.04385428e-01 -3.30818057e-01 4.79211152e-01
-1.33127046e+00 1.40896514e-01 -8.07982922e-01 -2.77804911e-01
-1.33884754e-02 1.90505594e-01 -1.45743102e-01 5.03891185e-02
-1.26706707e+00 1.17650342e+00 4.22294557e-01 3.96964461e-01
-1.20749402e+00 -1.87014118e-02 -6.63134098e-01 4.32282686e-01
-1.31556047e-02 -4.13285077e-01 1.80622661e+00 -9.11913633e-01
-2.11511970e+00 3.89636457e-01 -6.86309993e-01 -7.31900156e-01
-1.10039853e-01 -2.74180025e-01 -6.73370957e-01 -2.93131694e-02
-7.50657380e-01 6.31359577e-01 6.96278811e-01 -7.94844389e-01
-5.93751848e-01 -2.69649267e-01 -4.11326796e-01 4.42659378e-01
-2.60615647e-01 5.50836325e-01 -2.99877465e-01 -1.03742254e+00
-2.14388430e-01 -4.96300995e-01 -2.84721971e-01 -5.77841580e-01
-2.82678187e-01 -4.89568233e-01 1.02716780e+00 -1.45290887e+00
1.38559139e+00 -2.28856516e+00 2.52740711e-01 -4.40220892e-01
-3.87024611e-01 1.09909511e+00 -4.34567541e-01 5.24066269e-01
-1.98377185e-02 -4.36553769e-02 2.35726368e-02 -9.45037425e-01
-2.08958257e-02 4.39395696e-01 -3.13462615e-01 5.77923618e-02
3.12538952e-01 1.12504661e+00 -5.24746239e-01 -1.47132993e-01
5.21601915e-01 6.83930337e-01 -5.16108759e-02 7.65387535e-01
-7.23419413e-02 2.64798939e-01 6.17129207e-02 5.01285374e-01
2.67094493e-01 2.06034973e-01 -2.21365511e-01 5.76733053e-01
-3.92265469e-02 7.45765805e-01 -3.48687768e-01 1.52474678e+00
-7.48935759e-01 8.90092790e-01 1.50442317e-01 -9.55144048e-01
1.30748010e+00 9.72833097e-01 -6.19014144e-01 -8.40443492e-01
4.51364219e-02 2.82447875e-01 2.94173270e-01 -3.07109892e-01
5.74495435e-01 -3.47949624e-01 3.01244110e-01 -1.35613352e-01
1.76164031e-01 4.58319485e-01 -6.25715554e-01 -2.02363446e-01
9.81035948e-01 -4.71702605e-01 3.65620345e-01 1.98043928e-01
7.84949362e-01 -6.05331838e-01 3.17730397e-01 7.19065905e-01
-4.17252034e-01 8.94847810e-01 -1.64381608e-01 -2.34379753e-01
-1.39168739e+00 -9.79542494e-01 3.27783674e-01 1.38938701e+00
-5.47699869e-01 8.58154818e-02 -5.89826524e-01 -4.27126050e-01
-5.35249889e-01 1.00716043e+00 -1.33558393e-01 -5.52838929e-02
-5.89073658e-01 9.77466404e-02 8.85093808e-01 1.08239627e+00
3.07925373e-01 -1.38885868e+00 1.52604133e-01 5.74674904e-01
-1.64503843e-01 -1.15760708e+00 -7.43192315e-01 3.63011807e-01
-7.73121417e-01 1.60251595e-02 -1.52337873e+00 -1.40287781e+00
1.29290134e-01 1.77315980e-01 6.54678822e-01 -2.47550413e-01
3.13953012e-01 -1.23235539e-01 -7.46376872e-01 -2.31045768e-01
-5.49244642e-01 3.84883314e-01 -1.05636321e-01 1.47803441e-01
4.97145236e-01 -3.77871186e-01 -2.86708951e-01 1.97753057e-01
-6.10197127e-01 -1.97220683e-01 9.22699273e-01 9.06356275e-01
4.41713244e-01 -5.64655721e-01 1.03497493e+00 -2.71092206e-01
6.80616260e-01 -4.39940959e-01 -2.11324289e-01 3.89799893e-01
-9.33909938e-02 -1.10166028e-01 1.00517499e+00 -4.48542118e-01
-1.20088947e+00 -1.36462316e-01 -1.14529598e+00 -7.74169505e-01
-6.66407347e-01 7.23780096e-01 -3.08608592e-01 2.47478306e-01
1.18374020e-01 8.16580474e-01 -2.26358354e-01 -9.35104251e-01
3.08824450e-01 1.58074021e+00 6.05190694e-01 1.92881450e-01
1.05831372e-02 -3.48995805e-01 -7.74051785e-01 -1.28015971e+00
-3.11842680e-01 -7.58154809e-01 -5.31419694e-01 1.17565945e-01
1.12516880e+00 -1.04165936e+00 -4.68922675e-01 6.34906352e-01
-1.68028092e+00 -6.55091286e-01 2.11858466e-01 7.85184979e-01
-5.66889882e-01 4.51224983e-01 -1.21122217e+00 -1.31328654e+00
-6.28818452e-01 -1.06407261e+00 9.04240489e-01 -4.81181070e-02
5.93371456e-03 -1.07555652e+00 4.73185815e-02 3.01443249e-01
6.98801696e-01 -7.55318582e-01 6.23773932e-01 -1.16689014e+00
-4.27227169e-01 -1.08449012e-01 -2.10571252e-02 9.18049753e-01
-2.38430396e-01 -1.25421315e-01 -1.09295499e+00 -3.28221381e-01
-1.00383453e-01 -1.71324372e-01 9.15669858e-01 4.86089140e-01
1.13843226e+00 -4.19002146e-01 -9.33419392e-02 2.63202667e-01
1.07452929e+00 9.10091460e-01 1.15317786e+00 -1.26183813e-03
6.70746148e-01 4.09578919e-01 3.92475933e-01 6.50117770e-02
1.38251275e-01 7.13834286e-01 -3.00915558e-02 -2.97746211e-01
-5.15859783e-01 -3.29755574e-01 7.41445005e-01 1.63661265e+00
3.69269907e-01 -9.96577144e-01 -1.01829934e+00 8.69865775e-01
-1.86245525e+00 -9.56830561e-01 6.63397759e-02 1.87215626e+00
3.49575579e-01 2.32370973e-01 1.15964852e-01 -1.26942247e-01
1.00356245e+00 3.39926183e-01 -2.65442312e-01 -1.10796428e+00
4.94115660e-03 2.63315052e-01 2.46245340e-01 1.06318867e+00
-6.06796801e-01 1.25489819e+00 7.01945543e+00 1.25152600e+00
-1.10884798e+00 4.07290995e-01 6.61106825e-01 1.96893960e-01
-6.95404634e-02 -1.79870859e-01 -1.04345381e+00 4.31398749e-01
1.81763351e+00 2.63427645e-01 2.39631906e-01 9.66066182e-01
4.60736573e-01 4.95202184e-01 -9.53057945e-01 9.90554690e-01
1.49451017e-01 -9.70864832e-01 3.22869778e-01 5.59389554e-02
4.04566854e-01 5.44392943e-01 2.02346444e-01 5.03479362e-01
1.47580415e-01 -1.09448946e+00 6.05668604e-01 4.75445569e-01
6.44611418e-01 -1.01697683e+00 9.38361287e-01 6.40200257e-01
-1.12337708e+00 -9.44215506e-02 -4.44130033e-01 -2.28173673e-01
7.34859645e-01 1.41291261e-01 -1.21133912e+00 3.74128759e-01
1.70446068e-01 4.13978845e-01 -1.21170178e-01 1.08451891e+00
-3.52858901e-01 1.02904046e+00 1.09681189e-01 -4.13874835e-01
7.52113104e-01 2.94693500e-01 5.00685394e-01 1.70504236e+00
4.92238283e-01 2.80106813e-01 -7.44945034e-02 3.95945758e-01
-1.48218542e-01 -1.82067119e-02 -6.09912753e-01 -2.68694788e-01
4.49665308e-01 5.10789871e-01 -1.29850060e-01 -5.86984754e-01
-7.24338174e-01 1.47885084e+00 4.94448245e-01 6.40288413e-01
-6.37509227e-01 -7.97455430e-01 6.88201964e-01 -3.24646294e-01
1.06126285e+00 -6.52258933e-01 -5.85467815e-02 -9.44070697e-01
-3.49507518e-02 -8.27362001e-01 8.41270667e-03 -1.14917147e+00
-1.22273457e+00 1.04970479e+00 -6.82784259e-01 -7.48382449e-01
-4.19651806e-01 -6.71774209e-01 -7.44381726e-01 1.34535813e+00
-1.80219221e+00 -1.10245979e+00 4.08370674e-01 4.85853702e-01
1.52465200e+00 -3.50940049e-01 7.67719567e-01 4.21386600e-01
-5.49445868e-01 7.89429843e-01 4.96756017e-01 4.90842044e-01
2.54173428e-01 -9.70850945e-01 1.42426896e+00 1.20689499e+00
2.91635901e-01 4.12466854e-01 2.91746169e-01 -6.57074749e-01
-9.11544740e-01 -1.07230961e+00 1.52296996e+00 -1.67932391e-01
3.86896551e-01 -4.05662596e-01 -1.20565760e+00 1.03617454e+00
6.11551940e-01 -1.97500005e-01 6.16257846e-01 1.41081344e-02
-8.65113139e-02 1.51865885e-01 -6.04928851e-01 5.39593339e-01
8.15628827e-01 -1.10212004e+00 -9.66930091e-01 -2.00985551e-01
1.11143196e+00 -8.40186030e-02 -6.63301706e-01 2.01238856e-01
3.38734895e-01 -5.11963665e-01 8.28614354e-01 -6.85102403e-01
1.24699593e-01 -1.09507762e-01 -5.48147619e-01 -1.51447988e+00
-6.55645788e-01 -7.19395697e-01 -3.59692723e-01 1.17615044e+00
7.02545643e-01 -5.49462378e-01 4.81921762e-01 1.24789760e-01
-9.39643919e-01 -6.26831532e-01 -1.30685532e+00 -1.21815646e+00
3.12636673e-01 -4.13251966e-01 2.98543394e-01 4.55284715e-01
8.48018751e-02 6.09133124e-01 -7.58024693e-01 4.40511465e-01
-2.85010459e-03 -5.66315889e-01 1.79656476e-01 -5.10768294e-01
-5.49326360e-01 -4.20987934e-01 -5.70053458e-01 -2.11140442e+00
2.57225841e-01 -5.98037601e-01 6.11273885e-01 -1.64872491e+00
-1.08172096e-01 3.09224546e-01 -5.09887159e-01 3.19427788e-01
-1.77735180e-01 -1.59166574e-01 3.05937588e-01 -7.75688291e-02
-6.73092246e-01 1.04552662e+00 8.95785868e-01 1.07754223e-01
6.42537996e-02 7.09185079e-02 -5.14650829e-02 2.19771758e-01
7.58396685e-01 -3.06765616e-01 -1.43092245e-01 -8.77754092e-01
-7.27398872e-01 5.97075045e-01 1.04841858e-01 -9.89395976e-01
5.92768431e-01 2.25370377e-01 1.42724246e-01 -1.02151775e+00
7.15152323e-01 -4.63232309e-01 -3.93023908e-01 4.41787720e-01
-6.92507267e-01 1.34527698e-01 3.76632363e-02 6.07748330e-01
-7.85434127e-01 -3.00959468e-01 7.89324641e-01 -1.21079519e-01
-8.04584205e-01 1.75281495e-01 -1.17692804e+00 -4.31488216e-01
7.04644382e-01 -2.79584646e-01 -1.39625818e-01 -7.18858242e-01
-1.08389080e+00 -4.73649651e-02 -6.36496097e-02 7.02890217e-01
1.09484375e+00 -1.30596948e+00 -9.85376298e-01 3.64155471e-01
-3.89369912e-02 -4.13679689e-01 4.24025893e-01 4.74026173e-01
-3.02115023e-01 9.90733027e-01 3.37245092e-02 -2.80496895e-01
-1.53540969e+00 7.56784558e-01 1.99192390e-01 2.80565973e-02
-2.99016863e-01 1.18225145e+00 8.46972838e-02 -4.04651016e-01
7.04544187e-01 -4.40260351e-01 -7.35609010e-02 -4.74430084e-01
8.40794563e-01 3.59774679e-01 2.58340120e-01 -8.36803496e-01
-4.33594435e-01 8.36918280e-02 -4.01482642e-01 -5.10806084e-01
1.11491001e+00 -6.01935863e-01 5.16809583e-01 6.90118730e-01
1.40959179e+00 -5.40698767e-01 -1.15472889e+00 -2.39237502e-01
3.88377942e-02 -2.89630532e-01 3.75504255e-01 -8.32238555e-01
-6.91446245e-01 1.62032151e+00 4.67071712e-01 2.32102498e-01
9.37380493e-01 -9.54938121e-03 1.42259824e+00 3.67565602e-01
-4.44343984e-02 -1.09432912e+00 7.85347372e-02 1.10875869e+00
9.70168829e-01 -1.10943961e+00 -9.93671000e-01 -1.10270781e-02
-7.34242558e-01 1.18759584e+00 4.97592509e-01 8.77928548e-03
5.17251074e-01 2.12409317e-01 1.54741198e-01 3.72753620e-01
-1.17438519e+00 -5.30473530e-01 2.62168109e-01 7.53532887e-01
6.26099408e-01 1.12805724e-01 3.86734270e-02 3.56511325e-01
1.89921692e-01 -1.96167037e-01 3.22756171e-01 7.86069036e-01
-8.67603779e-01 -1.06738377e+00 -2.71058500e-01 -1.53221712e-01
-5.95672250e-01 -6.38802648e-01 -4.16821033e-01 -4.23187483e-03
-4.81456876e-01 1.28917706e+00 1.05707034e-01 -4.00384307e-01
3.16139072e-01 5.36437631e-01 1.92213729e-02 -7.21624255e-01
-5.54359734e-01 3.89090300e-01 2.81835645e-01 1.95160341e-02
1.38068303e-01 -2.97037959e-01 -9.72779334e-01 -1.69808492e-01
-4.94463950e-01 5.98544478e-02 6.51821673e-01 1.04557967e+00
6.09042704e-01 6.67558074e-01 6.23379707e-01 -8.34910929e-01
-5.61966121e-01 -1.40647221e+00 -3.79415542e-01 -1.10934615e-01
9.35602784e-01 7.40735268e-04 -2.91122019e-01 4.35811654e-02] | [14.426094055175781, 6.925102233886719] |
b3a6ab75-4b19-40cd-b810-9fea1b2e6b9a | sherf-generalizable-human-nerf-from-a-single | 2303.12791 | null | https://arxiv.org/abs/2303.12791v1 | https://arxiv.org/pdf/2303.12791v1.pdf | SHERF: Generalizable Human NeRF from a Single Image | Existing Human NeRF methods for reconstructing 3D humans typically rely on multiple 2D images from multi-view cameras or monocular videos captured from fixed camera views. However, in real-world scenarios, human images are often captured from random camera angles, presenting challenges for high-quality 3D human reconstruction. In this paper, we propose SHERF, the first generalizable Human NeRF model for recovering animatable 3D humans from a single input image. SHERF extracts and encodes 3D human representations in canonical space, enabling rendering and animation from free views and poses. To achieve high-fidelity novel view and pose synthesis, the encoded 3D human representations should capture both global appearance and local fine-grained textures. To this end, we propose a bank of 3D-aware hierarchical features, including global, point-level, and pixel-aligned features, to facilitate informative encoding. Global features enhance the information extracted from the single input image and complement the information missing from the partial 2D observation. Point-level features provide strong clues of 3D human structure, while pixel-aligned features preserve more fine-grained details. To effectively integrate the 3D-aware hierarchical feature bank, we design a feature fusion transformer. Extensive experiments on THuman, RenderPeople, ZJU_MoCap, and HuMMan datasets demonstrate that SHERF achieves state-of-the-art performance, with better generalizability for novel view and pose synthesis. | ['Ziwei Liu', 'Lei Yang', 'Haiyi Mei', 'Liang Pan', 'Fangzhou Hong', 'Shoukang Hu'] | 2023-03-22 | null | null | null | null | ['3d-human-reconstruction'] | ['computer-vision'] | [-1.26125380e-01 -3.19033682e-01 -2.23636981e-02 -3.40556383e-01
-3.81079167e-01 -4.01587158e-01 5.43284476e-01 -3.07572126e-01
-2.29631573e-01 2.88763881e-01 5.78967273e-01 7.05972970e-01
2.18038127e-01 -6.23212039e-01 -5.85933745e-01 -4.46325183e-01
1.57906920e-01 4.50564444e-01 6.93932697e-02 -3.24512124e-01
-2.98864216e-01 7.61014163e-01 -1.71740115e+00 3.87839079e-01
3.36068392e-01 8.89042974e-01 3.44973989e-02 7.49815941e-01
3.52139801e-01 6.05985701e-01 -3.64862859e-01 -4.11276400e-01
6.55909956e-01 -3.61857235e-01 -3.45555991e-01 6.43698931e-01
8.43318284e-01 -8.24523330e-01 -8.83617878e-01 6.47304833e-01
6.01751506e-01 2.95352638e-01 4.62750703e-01 -1.02224588e+00
-6.21027112e-01 -2.95369834e-01 -5.89435220e-01 -4.62161265e-02
1.19613111e+00 5.56031644e-01 7.91170895e-01 -1.06870627e+00
9.30865049e-01 1.73247051e+00 6.54256523e-01 6.90055251e-01
-1.06965244e+00 -4.25194830e-01 7.74625018e-02 1.06493108e-01
-1.38661456e+00 -3.78920197e-01 9.61302757e-01 -4.85203683e-01
1.02310741e+00 1.84690490e-01 1.31613386e+00 1.45220470e+00
2.77267963e-01 8.55820715e-01 1.15126252e+00 -1.39350131e-01
-2.67598540e-01 -1.63430393e-01 -2.75453657e-01 1.11176705e+00
3.28424871e-01 3.63613635e-01 -1.00940108e+00 -6.36469424e-02
1.45586312e+00 5.02246141e-01 -5.29218376e-01 -7.58802831e-01
-1.55210769e+00 5.59802234e-01 4.90454853e-01 -1.27831355e-01
-6.01658762e-01 3.31668928e-02 -6.33110804e-03 -1.27856638e-02
2.04123810e-01 2.78267831e-01 -2.95150846e-01 5.19892946e-03
-5.44524014e-01 5.52821279e-01 3.01227272e-01 1.09643686e+00
6.15273416e-01 5.14017344e-02 -1.88437372e-01 6.06397688e-01
2.70889342e-01 8.43218148e-01 2.31994346e-01 -1.30465674e+00
4.01924342e-01 8.10713351e-01 1.24457985e-01 -1.38067210e+00
-5.49107671e-01 -2.19176590e-01 -1.11012900e+00 1.31566927e-01
2.93260574e-01 3.03985745e-01 -8.60076308e-01 1.57331121e+00
6.21792555e-01 -1.52599648e-01 -2.22971216e-01 1.49143314e+00
9.24691260e-01 5.59583783e-01 -2.61823684e-01 1.84398934e-01
1.58464372e+00 -8.86268795e-01 -4.65430647e-01 -3.12620491e-01
-4.23092209e-02 -4.01999980e-01 8.75122488e-01 4.36304003e-01
-1.18083036e+00 -1.01892459e+00 -9.84319389e-01 -5.11634469e-01
-4.53306884e-02 -3.80911864e-02 4.10162568e-01 3.18750143e-01
-6.43960238e-01 3.39134127e-01 -7.92867959e-01 -3.83793682e-01
8.20765123e-02 9.67672467e-02 -1.14079726e+00 -4.34858054e-01
-1.07154763e+00 7.63645351e-01 3.02311908e-02 1.79209888e-01
-7.57908642e-01 -4.31058675e-01 -1.13859427e+00 -1.48318619e-01
3.85874897e-01 -1.40470624e+00 7.76410758e-01 -3.37219745e-01
-1.30100095e+00 1.08585489e+00 -1.58950448e-01 -1.16793558e-01
6.79507792e-01 -4.37016696e-01 -1.02044225e-01 6.95947230e-01
3.91040258e-02 8.39031577e-01 1.02676785e+00 -1.39666712e+00
-4.13145930e-01 -8.35779607e-01 9.16169211e-02 5.49600244e-01
-8.09149891e-02 -3.21994692e-01 -7.01974928e-01 -9.81634736e-01
2.56017864e-01 -9.13149655e-01 -1.17219292e-01 4.94252533e-01
-2.66128749e-01 1.53035060e-01 6.53933287e-01 -9.00728941e-01
7.36288369e-01 -2.06783414e+00 7.56608367e-01 4.10278328e-02
6.37367070e-01 -4.19752486e-02 -1.26221120e-01 1.66310653e-01
1.96716204e-01 -3.16483051e-01 7.27872699e-02 -4.68561083e-01
8.86339322e-02 3.09905708e-01 1.01685986e-01 5.11052370e-01
1.68659419e-01 1.08844030e+00 -8.69625449e-01 -5.50514162e-01
6.31292164e-01 9.02187407e-01 -7.89812684e-01 5.36850989e-01
1.46324128e-01 7.30508268e-01 -5.35559177e-01 8.12935293e-01
4.44040209e-01 -3.25514108e-01 -1.10587776e-01 -4.73563939e-01
3.69527131e-01 -2.30507106e-01 -1.13149893e+00 2.00415826e+00
-1.39605820e-01 1.99769258e-01 1.02220751e-04 -4.12492365e-01
8.52878392e-01 2.17600405e-01 5.33759058e-01 -6.14167809e-01
1.01224326e-01 -2.22670361e-01 -5.89130342e-01 -3.82719815e-01
4.98172432e-01 -3.98618579e-02 -3.06407422e-01 1.15733035e-01
2.09561348e-01 -1.73220575e-01 -2.50658751e-01 1.45623296e-01
9.54665601e-01 3.41289252e-01 6.36217475e-01 2.62815356e-01
3.68478328e-01 -3.78260851e-01 6.25770748e-01 3.85067284e-01
-3.30616742e-01 1.07686782e+00 9.65422913e-02 -8.64715993e-01
-1.18919659e+00 -1.33628058e+00 3.15803826e-01 7.19772816e-01
2.74433583e-01 -6.52747631e-01 -6.24687731e-01 -6.66298330e-01
1.40941113e-01 9.39635485e-02 -7.14799643e-01 -1.17432415e-01
-8.49723220e-01 -1.35086760e-01 3.59525770e-01 5.01903594e-01
7.59706676e-01 -8.73570144e-01 -1.19936574e+00 2.92065907e-02
-7.07243085e-01 -1.38070047e+00 -8.52444470e-01 -3.44273061e-01
-6.56451762e-01 -1.08151197e+00 -9.47225153e-01 -4.75692332e-01
6.99824095e-01 6.79436505e-01 1.10364532e+00 5.28806299e-02
-5.74674010e-01 6.04935110e-01 -4.02886361e-01 2.75240690e-01
1.01619981e-01 -4.62752998e-01 4.77093190e-01 4.35255393e-02
2.19572783e-01 -5.84244311e-01 -6.78839505e-01 4.18814480e-01
-5.39457858e-01 4.14393485e-01 5.67565858e-01 1.01442695e+00
7.87986994e-01 -1.71940997e-01 -6.08990975e-02 -4.23951417e-01
1.36775309e-02 9.49052200e-02 -1.71075374e-01 1.94267526e-01
4.70009930e-02 -1.03577487e-01 5.82734942e-01 -4.27583694e-01
-9.85456169e-01 3.40278625e-01 5.14115654e-02 -1.04308999e+00
-3.87963414e-01 -1.47178874e-03 -4.82281327e-01 2.69216876e-02
6.73705161e-01 3.01046431e-01 8.80239997e-03 -5.37936211e-01
3.68218392e-01 2.27385953e-01 8.88227403e-01 -6.49741113e-01
9.21323180e-01 6.17030382e-01 1.40225384e-02 -9.18939710e-01
-8.92738402e-01 -5.06554842e-01 -1.11973584e+00 -5.07948458e-01
1.13984156e+00 -1.40271533e+00 -6.44435227e-01 6.84754372e-01
-1.13109815e+00 -1.39001757e-03 -2.87674576e-01 5.02255261e-01
-8.00076902e-01 5.83119512e-01 -7.68417180e-01 -5.08548856e-01
-2.29235306e-01 -1.11447871e+00 1.64340734e+00 7.97326639e-02
-4.48445797e-01 -6.26858711e-01 -4.15593162e-02 8.38461041e-01
-2.44012922e-01 7.09156036e-01 5.71910441e-01 9.49392095e-02
-7.11333692e-01 -2.82916695e-01 -6.90640230e-03 2.19456986e-01
5.49069420e-02 -3.96055669e-01 -7.19759703e-01 -3.99250299e-01
-1.56195310e-03 -4.41986233e-01 6.34080648e-01 2.47899845e-01
8.76430988e-01 -2.96047837e-01 -1.09987840e-01 9.83369768e-01
8.83292496e-01 -3.75673562e-01 3.55118364e-01 -6.24609627e-02
1.10991621e+00 5.92666090e-01 4.81375337e-01 8.14162195e-01
6.84553146e-01 9.35167253e-01 1.55440614e-01 -7.78249837e-03
-4.23098505e-01 -7.60042250e-01 5.55591583e-01 8.51730525e-01
-5.01127720e-01 2.87002549e-02 -6.28086448e-01 2.09087089e-01
-1.71380532e+00 -1.23026097e+00 2.42575929e-01 2.02532387e+00
5.39198279e-01 -3.21070366e-02 3.34757298e-01 -3.20059285e-02
5.84607065e-01 3.07408750e-01 -5.58728039e-01 3.76036167e-01
-3.10739130e-01 -6.72597736e-02 5.41711822e-02 2.90158898e-01
-9.66838479e-01 7.67130733e-01 5.47765446e+00 4.13818747e-01
-6.54794633e-01 -1.12125710e-01 1.21691942e-01 -3.63188982e-01
-2.57709652e-01 -3.57553422e-01 -7.40352154e-01 1.32287160e-01
2.07371488e-01 1.53126448e-01 3.63295197e-01 7.53529787e-01
1.53102353e-02 2.06769630e-01 -1.21830928e+00 1.60512948e+00
5.22394478e-01 -1.00238490e+00 3.82885754e-01 3.57674062e-01
6.41658127e-01 -4.90928710e-01 7.26453289e-02 6.62104413e-02
5.79498447e-02 -8.91769707e-01 9.44819033e-01 8.74015450e-01
8.45744550e-01 -8.23655605e-01 4.03115630e-01 4.30656612e-01
-1.37915015e+00 5.82760796e-02 -3.61182302e-01 2.03207899e-02
4.08441454e-01 3.61445189e-01 -3.09386820e-01 6.57274008e-01
1.09457839e+00 1.05187964e+00 -6.54243529e-01 5.28885543e-01
-1.89021155e-01 -2.35346347e-01 -3.64603102e-01 3.48232001e-01
-5.76157197e-02 -4.43912186e-02 6.71579063e-01 1.03999567e+00
3.50273639e-01 8.03212404e-01 4.26708817e-01 6.73042476e-01
3.79136875e-02 -2.55870283e-01 -6.67681873e-01 2.17758983e-01
2.04511315e-01 1.23003340e+00 -4.40001607e-01 -3.00196707e-01
-3.73249322e-01 1.45039403e+00 3.31232965e-01 2.83205241e-01
-8.50454211e-01 -7.42254630e-02 8.50134373e-01 4.28486526e-01
3.49512219e-01 -5.82208455e-01 1.10628091e-01 -1.75382984e+00
1.67697802e-01 -1.13825941e+00 3.07341218e-01 -1.01924455e+00
-1.41129553e+00 6.88370764e-01 2.06221730e-01 -1.40400636e+00
-4.11450744e-01 -7.07198083e-01 3.65235098e-02 5.61093271e-01
-9.76803720e-01 -1.50453901e+00 -8.76071513e-01 1.02786517e+00
6.31828845e-01 -1.92316636e-01 7.32190728e-01 5.62641136e-02
-2.27244392e-01 5.91693878e-01 -5.89315116e-01 2.49510363e-01
8.94832194e-01 -1.01960695e+00 6.42640531e-01 6.99645281e-01
2.43429378e-01 6.89634621e-01 4.36299860e-01 -7.19036341e-01
-1.80608821e+00 -8.88436019e-01 6.54179990e-01 -9.22379494e-01
-9.12154764e-02 -5.08932292e-01 -7.24878252e-01 5.49921155e-01
-2.19208434e-01 3.57136279e-01 4.58854198e-01 -1.72562748e-01
-6.74720109e-01 1.42074928e-01 -1.26329732e+00 5.92242539e-01
1.59851909e+00 -6.20118141e-01 -8.16900253e-01 6.88443109e-02
6.00103319e-01 -5.81587493e-01 -1.16393447e+00 4.73244816e-01
1.00437355e+00 -1.31765735e+00 1.47193670e+00 -5.04492223e-01
4.14353907e-01 -4.39464897e-01 -5.26465595e-01 -1.17916894e+00
-7.12645531e-01 -4.72884715e-01 -4.34662342e-01 7.07612872e-01
-3.95582497e-01 -2.05290630e-01 7.60612845e-01 4.15803969e-01
1.06641069e-01 -6.71898246e-01 -7.63967931e-01 -6.05563223e-01
-3.28124315e-01 -9.54501331e-02 5.23169398e-01 7.41960943e-01
-1.83695480e-01 3.56879354e-01 -9.13430572e-01 5.22496328e-02
1.03177416e+00 1.28507465e-01 1.23659158e+00 -1.30201292e+00
-4.36002523e-01 -2.95714363e-02 -7.46410728e-01 -1.44236374e+00
9.27439034e-02 -5.33105731e-01 -9.01235640e-02 -1.31624103e+00
4.75867242e-01 4.28947434e-02 1.90206483e-01 3.74424696e-01
-2.03118622e-01 4.44546759e-01 5.74903965e-01 2.54822671e-01
-5.54189503e-01 8.84925187e-01 1.65956151e+00 1.98997799e-02
6.14906959e-02 -3.69323254e-01 -3.27647388e-01 9.32775259e-01
3.05724561e-01 -4.34582457e-02 -2.15027973e-01 -5.25639832e-01
-9.72745717e-02 3.88215512e-01 1.00014317e+00 -1.13499928e+00
1.48054451e-01 -2.42034137e-01 1.26984930e+00 -8.25090706e-01
9.36460316e-01 -7.98017025e-01 4.43123281e-01 3.18808615e-01
-8.42346549e-02 4.00258601e-01 -3.22346449e-01 8.62392902e-01
-1.40566811e-01 6.02709115e-01 8.71126115e-01 -5.76855481e-01
-8.69565904e-01 7.29766428e-01 6.71386868e-02 1.12821408e-01
8.11864138e-01 -4.26493347e-01 -3.67647409e-02 -5.09918690e-01
-7.52076805e-01 5.48311099e-02 1.02077353e+00 6.94656372e-01
1.26781642e+00 -1.70356929e+00 -7.57698238e-01 7.34969020e-01
2.46266529e-01 1.38463378e-01 6.92919493e-01 4.06964749e-01
-6.23718083e-01 3.19509238e-01 -7.12088108e-01 -7.38042951e-01
-1.27165234e+00 6.55237317e-01 2.67471194e-01 -1.05241701e-01
-1.07955790e+00 7.15046525e-01 6.79572999e-01 -5.51313221e-01
1.08480349e-01 -2.48481825e-01 1.63524374e-01 -1.59679562e-01
6.23547375e-01 3.70001644e-01 -3.23803753e-01 -1.28098857e+00
-3.39887440e-01 1.11161673e+00 1.57869473e-01 -1.62835270e-01
1.18152642e+00 -3.12348306e-01 2.00572252e-01 3.69780064e-01
1.23060083e+00 5.87001406e-02 -1.61196363e+00 -4.65437293e-01
-8.53709459e-01 -9.24987614e-01 -2.88455069e-01 -4.80092049e-01
-1.03446984e+00 1.03348422e+00 2.75453329e-01 -4.93037075e-01
1.14094865e+00 1.66512392e-02 1.16615987e+00 3.79097462e-01
8.63057554e-01 -8.45619380e-01 5.47196269e-01 2.64789343e-01
1.12807167e+00 -1.06411839e+00 2.88232565e-01 -4.50142056e-01
-8.43058527e-01 1.00040686e+00 9.49490309e-01 -1.08735442e-01
3.65068853e-01 -1.93077028e-01 -2.53578424e-01 -2.65096545e-01
-6.64629579e-01 -1.00267097e-01 6.87453866e-01 7.90495396e-01
1.25427293e-02 1.81601062e-01 4.88441974e-01 7.90724635e-01
-4.13616031e-01 -2.84977138e-01 2.65130758e-01 6.48049295e-01
-4.14613336e-01 -6.36492252e-01 -6.12743139e-01 1.16555400e-01
-6.82365969e-02 2.99298555e-01 -5.39186597e-01 9.02970493e-01
2.32772872e-01 6.16227567e-01 -4.21355478e-02 -6.78781152e-01
7.89237618e-01 -1.10531062e-01 1.02360499e+00 -4.40794557e-01
-3.37647110e-01 2.73830950e-01 -1.24901801e-01 -9.54763591e-01
-4.78272170e-01 -7.63939738e-01 -1.07619405e+00 -5.10226846e-01
1.28265694e-01 -3.35871220e-01 -4.01041172e-02 6.89723611e-01
3.21355760e-01 2.10118487e-01 4.24274772e-01 -1.45189738e+00
-3.36141318e-01 -5.50477922e-01 -6.51927114e-01 7.70495892e-01
5.67056477e-01 -1.03771019e+00 -2.54574418e-01 3.21119338e-01] | [7.277997016906738, -0.7095832824707031] |
f11e5361-30d5-4f49-adca-5faffcd63692 | towards-interpretable-math-word-problem | null | null | https://openreview.net/forum?id=euGE2v_UZgT | https://openreview.net/pdf?id=euGE2v_UZgT | Towards Interpretable Math Word Problem Solving with Grounded Linguistic Logic Reasoning | Automatically math word problem (MWP) solving is a challenging artificial intelligence task since a machine should be able to not only understand problem comprehensively on linguistics but also the grounded math logic entailed in problem. Recently, lots of deep learning models have made great progress in MWP solving on answer accuracy, they rely on shallow heuristics to achieve high performance, lacking of grounded math logic reasoning, which makes them uninterpretable. To address this issue and push the research boundary of MWPs to interpretable MWP solving, we construct a large-scale and high-quality MWP dataset named InterMWP which consists of 11507 MWP data and annotates interpretable algebraic knowledge formulas as the grounded linguistic logic of each solving equation and asks for a solver to output the formulas when it decides current predicted node is a inner-node (operator) during expression reasoning. We further propose a strong baseline called InterSolver to show the effectiveness of our constructed dataset and show how to harvest these logic knowledge by fusing logic knowledge with semantic representation to improve problem solving and make a step towards providing interpretability. Experimental results shows that our InterSolver has strong logical formula-based interpretability while achieving high answer accuracy simultaneously. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['math-word-problem-solving', 'math-word-problem-solving', 'math-word-problem-solving'] | ['knowledge-base', 'reasoning', 'time-series'] | [-2.81406678e-02 8.26469183e-01 -3.54019970e-01 -6.67965412e-01
-6.96984887e-01 -6.86304688e-01 -9.16474238e-02 1.52364448e-01
2.32030258e-01 7.94404566e-01 2.37687945e-01 -7.65704513e-01
-3.21761549e-01 -1.62777352e+00 -1.09168279e+00 8.02901313e-02
2.45622873e-01 1.14782667e+00 9.05409381e-02 -5.79105377e-01
-6.87678829e-02 -1.99860781e-02 -1.16608500e+00 1.00926971e+00
1.35427618e+00 1.03726423e+00 -2.14748338e-01 3.01482975e-01
-7.62399971e-01 1.41623032e+00 -2.86253572e-01 -9.56101298e-01
1.73949301e-01 -2.60887176e-01 -1.64582050e+00 -5.49774408e-01
4.63754445e-01 -1.81660786e-01 -3.98996137e-02 1.15830922e+00
-2.24811897e-01 -2.90951073e-01 3.75614434e-01 -1.54321241e+00
-1.29664552e+00 1.37745285e+00 -6.03942946e-02 -1.86180040e-01
8.24304223e-01 3.35510075e-01 1.71422529e+00 -5.60915828e-01
7.01711893e-01 1.79545152e+00 7.73874879e-01 4.56646234e-01
-9.21882212e-01 -5.54323912e-01 1.31451443e-01 9.40214574e-01
-1.19586301e+00 1.95265952e-02 7.79402852e-01 -2.16186736e-02
1.30871236e+00 3.74420732e-01 6.72980964e-01 6.02771163e-01
9.05825123e-02 8.28520000e-01 8.61321270e-01 -5.32690167e-01
6.83245882e-02 1.63162604e-01 8.13002646e-01 1.42344546e+00
1.29772261e-01 -5.43724298e-01 -4.88764286e-01 1.26203641e-01
3.61050814e-01 -4.36058640e-01 -2.90539503e-01 2.69932866e-01
-9.33059275e-01 9.64976013e-01 9.01014566e-01 9.30826440e-02
-3.89431529e-02 4.20347869e-01 3.12256783e-01 5.90529978e-01
3.65133397e-02 1.00294828e+00 -8.23152959e-01 8.24729577e-02
-3.53506416e-01 6.18246496e-01 1.22519910e+00 9.58189189e-01
8.25292587e-01 -3.10313433e-01 -1.63462877e-01 2.84628749e-01
3.82493049e-01 6.64331734e-01 1.52598485e-01 -1.28959692e+00
9.29005623e-01 1.58869433e+00 -2.29370862e-01 -1.39107788e+00
-4.42686439e-01 1.22156590e-01 -4.01741952e-01 -3.20481300e-01
2.64769435e-01 -2.80003548e-02 -4.23387438e-01 1.63479066e+00
3.08081120e-01 2.77637720e-01 2.91830331e-01 1.04026020e+00
1.34622526e+00 1.05554926e+00 -9.03571770e-02 1.69009000e-01
1.54485440e+00 -1.13407087e+00 -7.55116582e-01 -2.56711900e-01
1.09820926e+00 3.88497673e-02 1.54202878e+00 4.54779118e-01
-1.08700240e+00 -1.35414407e-01 -1.08299994e+00 -7.18530059e-01
-5.19031227e-01 -1.92485973e-01 1.16222394e+00 1.50468901e-01
-8.43729258e-01 2.74336696e-01 -4.41170454e-01 4.69033644e-02
7.34935880e-01 3.36303324e-01 -2.96520799e-01 -4.47774053e-01
-1.74443388e+00 1.13912559e+00 9.11952138e-01 9.60527211e-02
-4.39641148e-01 -1.02700150e+00 -1.26703560e+00 3.29645336e-01
8.65169525e-01 -8.48360837e-01 1.11664188e+00 -8.53068352e-01
-1.26142943e+00 8.86465013e-01 -4.85582739e-01 -8.75971198e-01
4.74701859e-02 -1.28985345e-01 -4.11323100e-01 2.03537345e-01
6.74876049e-02 6.40402794e-01 -4.92637977e-02 -9.12727237e-01
-2.72422463e-01 -3.72532934e-01 1.02672899e+00 -3.34954076e-02
-1.51720539e-01 -2.11127296e-01 -6.15633503e-02 2.43747253e-02
3.61619681e-01 -5.02498209e-01 1.55284494e-01 7.59859830e-02
-4.56433266e-01 -8.14622641e-01 5.35643339e-01 -6.61083817e-01
1.26650774e+00 -1.69135046e+00 3.73750478e-01 1.71861440e-01
4.62382346e-01 1.69720501e-01 -2.16451168e-01 3.43858987e-01
-8.75173137e-03 4.63839680e-01 -2.23800942e-01 2.31804863e-01
5.88765562e-01 8.55075598e-01 -9.29074883e-01 -2.23657250e-01
6.89924598e-01 1.53602397e+00 -1.05236542e+00 -6.30030692e-01
3.61742638e-02 -3.70789528e-01 -1.09482145e+00 9.25371423e-02
-1.10110962e+00 -5.51264100e-02 -5.78603864e-01 8.10843408e-01
7.58933544e-01 -4.96397555e-01 2.17690870e-01 -3.79118323e-01
5.24814785e-01 6.03633225e-01 -1.10035157e+00 1.70294511e+00
-8.11857641e-01 5.68029404e-01 -2.53653526e-01 -1.17815483e+00
7.69242287e-01 1.42074808e-01 -1.62836716e-01 -7.16942608e-01
-3.51874530e-02 2.58648783e-01 9.73562747e-02 -8.84085357e-01
2.14870423e-01 -4.53184277e-01 -9.79654565e-02 2.84299672e-01
1.28696769e-01 -5.95246255e-01 2.81877011e-01 4.54091460e-01
1.11479068e+00 3.79819199e-02 7.41093531e-02 -3.92579913e-01
9.68500853e-01 5.29929161e-01 6.55040860e-01 5.17479181e-01
1.71354756e-01 1.17073972e-02 1.01060152e+00 -1.02169323e+00
-3.91485184e-01 -9.42180336e-01 -1.66358706e-02 7.65388429e-01
4.20659393e-01 -8.36726069e-01 -7.32363224e-01 -7.19483256e-01
-1.73611345e-03 1.07938910e+00 -3.93270940e-01 -3.07539284e-01
-8.49165261e-01 -2.19546482e-01 8.92020583e-01 5.67791581e-01
9.77110803e-01 -1.10917163e+00 -1.72498986e-01 -3.29857990e-02
-9.02603030e-01 -1.38453853e+00 4.08271730e-01 -8.25922564e-02
-5.18251956e-01 -1.45677328e+00 6.89747751e-01 -8.79596174e-01
6.91693485e-01 -3.65329564e-01 1.61522615e+00 6.24576509e-01
-1.10773435e-02 3.73214073e-02 -3.20822120e-01 -6.72503531e-01
-2.43267119e-01 8.26824158e-02 -1.25775427e-01 -3.44065696e-01
7.59210885e-01 -3.16701829e-01 3.33360694e-02 -1.32455379e-01
-6.58505976e-01 5.21926522e-01 -5.23210037e-04 5.98316312e-01
4.21181440e-01 4.34905857e-01 3.60287577e-01 -1.10151029e+00
6.19484365e-01 -5.55437863e-01 -6.20378613e-01 7.92883098e-01
-4.05658066e-01 3.20088863e-01 1.13439417e+00 1.54702991e-01
-9.81478214e-01 -4.57871586e-01 -1.76479727e-01 -4.68093567e-02
1.33315012e-01 9.33987141e-01 -3.26813608e-01 6.04855642e-02
7.44708896e-01 -4.47925515e-02 -3.53771091e-01 9.57207680e-02
6.81731761e-01 2.48554185e-01 5.49791634e-01 -1.55005980e+00
9.49140429e-01 1.63977146e-01 2.50177234e-01 -4.52111326e-02
-1.70640624e+00 2.82933325e-01 -3.72986019e-01 3.45863253e-01
8.15031230e-01 -7.55038381e-01 -1.26885366e+00 -2.11849317e-01
-1.56261504e+00 -4.52676892e-01 -2.50582516e-01 -1.43769220e-01
-4.15883660e-01 2.69370954e-02 -7.60186851e-01 -7.16049254e-01
-4.00187582e-01 -1.07457149e+00 9.69133914e-01 4.65946123e-02
-6.22077763e-01 -1.16566133e+00 -2.05234423e-01 8.91163051e-01
1.19701855e-01 3.13440502e-01 1.89149511e+00 -6.06503248e-01
-8.15583050e-01 3.99835855e-02 -6.31847560e-01 1.74083620e-01
-2.61365771e-01 -1.12653516e-01 -7.49421954e-01 5.53905725e-01
-9.23879966e-02 -7.76870668e-01 4.87355769e-01 1.34984940e-01
1.66881275e+00 -7.91351676e-01 8.26614723e-02 6.45604789e-01
1.49514937e+00 -3.41056228e-01 7.57623613e-01 2.05273777e-01
6.20396078e-01 4.48332548e-01 4.24232572e-01 -6.30637184e-02
1.02988565e+00 3.62829030e-01 5.10236681e-01 1.91503003e-01
1.22196294e-01 -5.60246944e-01 1.05232716e-01 5.59320450e-01
-4.20826636e-02 -3.76450084e-02 -1.36794984e+00 3.77825916e-01
-2.02035356e+00 -8.84586394e-01 -5.37559390e-01 1.22550988e+00
1.42390108e+00 1.97449122e-02 -5.83760142e-01 2.36092716e-01
-1.28456466e-02 -1.40469328e-01 -3.19488704e-01 -7.21602082e-01
-1.50324389e-01 6.93130910e-01 -1.38062850e-01 1.03962564e+00
-7.19213307e-01 1.43969059e+00 5.32088375e+00 8.12790155e-01
-7.63163269e-01 -9.43721682e-02 4.32643712e-01 2.17790812e-01
-6.84475541e-01 6.06728606e-02 -6.96437538e-01 5.98818809e-02
6.71731472e-01 -1.83214858e-01 8.77345979e-01 8.56592417e-01
-1.26376987e-01 1.69283617e-02 -1.52292752e+00 8.95784855e-01
-1.08250350e-01 -1.75074708e+00 4.96376932e-01 -5.74176013e-01
5.82181454e-01 -5.17857969e-01 -2.69420236e-01 8.91691327e-01
4.24838126e-01 -1.67208409e+00 5.70204914e-01 4.08256531e-01
3.17892909e-01 -6.47937357e-01 9.77096021e-01 5.19948781e-01
-1.18820250e+00 -3.40590149e-01 -5.02836883e-01 -9.78320539e-01
-1.96384825e-02 5.02376616e-01 -7.46339679e-01 8.13746810e-01
3.73489767e-01 8.18932235e-01 -4.80613083e-01 2.25762889e-01
-1.00233591e+00 5.22335768e-01 -3.21154326e-01 -1.00807332e-01
5.11357903e-01 -2.08636791e-01 1.00113511e-01 8.07472289e-01
6.71099797e-02 7.45110571e-01 1.80429354e-01 1.74417555e+00
-3.07111949e-01 -2.32183784e-01 -5.33344924e-01 -2.50257671e-01
5.55609643e-01 1.27847373e+00 -2.79578954e-01 -6.39020324e-01
-1.56846151e-01 6.41516626e-01 7.43063390e-01 2.96429694e-01
-1.26438916e+00 -2.06185862e-01 5.52360177e-01 -4.52426523e-02
-3.30108404e-01 3.90532203e-02 -8.53581309e-01 -1.42990935e+00
3.71866643e-01 -9.40394819e-01 6.31654024e-01 -1.21914101e+00
-1.18619037e+00 2.75467545e-01 6.39023185e-02 -3.45972121e-01
6.99231997e-02 -1.23383319e+00 -6.74097002e-01 8.91522527e-01
-1.66510856e+00 -1.46444607e+00 -4.69799370e-01 4.78224814e-01
4.64216143e-01 8.11201185e-02 9.89896715e-01 2.76874788e-02
-4.55752313e-01 4.64218199e-01 -1.04342711e+00 3.56037289e-01
1.52323008e-01 -1.34131312e+00 -3.27971652e-02 6.33750081e-01
2.83323705e-01 9.18125212e-01 4.95988011e-01 -2.93005794e-01
-2.06719804e+00 -1.04845667e+00 1.19852912e+00 -9.53664243e-01
8.62807751e-01 -1.05877452e-01 -9.60473776e-01 1.13226891e+00
1.44365460e-01 1.25600956e-03 5.28945506e-01 4.38701451e-01
-5.72734118e-01 -1.81670681e-01 -1.24864352e+00 6.98512077e-01
1.23365533e+00 -5.42990029e-01 -1.34500170e+00 7.75424778e-01
1.28467095e+00 -8.00519943e-01 -6.95280731e-01 5.69360733e-01
4.03717272e-02 -6.98275924e-01 9.48580265e-01 -1.17380345e+00
1.29854441e+00 -5.35825729e-01 -2.99525499e-01 -9.89388347e-01
-1.82919025e-01 -1.68485850e-01 -4.19988215e-01 1.14415205e+00
6.72430992e-01 -6.24054253e-01 6.91155016e-01 1.30709505e+00
-2.54719734e-01 -1.23819435e+00 -7.04600334e-01 -3.54015678e-01
1.91918135e-01 -8.95162582e-01 9.81797814e-01 1.27287602e+00
6.19934916e-01 5.38965523e-01 2.78695315e-01 4.67062950e-01
2.34934598e-01 7.15133727e-01 5.51649570e-01 -1.06207347e+00
-3.36767793e-01 -3.64831448e-01 -1.38704598e-01 -8.31501722e-01
9.65443373e-01 -1.63107598e+00 -2.99139768e-01 -1.81271482e+00
3.07431314e-02 -4.02611345e-01 4.80882302e-02 1.11050379e+00
1.53934176e-03 -9.73797515e-02 6.16437905e-02 -4.94895399e-01
-8.01764667e-01 3.41396600e-01 1.40785861e+00 -5.65775573e-01
2.41986290e-01 -7.39367604e-01 -1.01203239e+00 1.14074707e+00
3.78651977e-01 -6.00886978e-02 -6.76530004e-01 -1.11940396e+00
1.06466234e+00 4.69045565e-02 5.77733278e-01 -8.04906726e-01
3.81977260e-01 -6.19793594e-01 -9.63854194e-02 -1.19909637e-01
5.79243898e-02 -9.71650183e-01 -1.54082745e-01 3.69094342e-01
-4.13295895e-01 -4.98241419e-03 2.78726369e-01 3.55291478e-02
-4.37231511e-01 -4.57327396e-01 3.21407944e-01 -4.07467335e-01
-1.07028770e+00 5.16176522e-02 2.36614987e-01 6.01016998e-01
7.80478716e-01 2.46530741e-01 -5.54530323e-01 -1.18481383e-01
-3.31996769e-01 8.54397893e-01 7.68998414e-02 3.57118882e-02
6.87456369e-01 -1.37618172e+00 -4.54906672e-01 8.11180100e-03
1.19006097e-01 6.45297408e-01 -1.78359654e-02 7.29948699e-01
-1.18757951e+00 6.04289770e-01 6.69275150e-02 -1.77074879e-01
-9.80867326e-01 4.19899762e-01 5.92331350e-01 -6.36848152e-01
-5.42711139e-01 1.18380225e+00 -8.85430910e-03 -9.49354947e-01
1.36276275e-01 -1.12159133e+00 -2.78470181e-02 -4.11553085e-01
6.87722385e-01 -4.72813435e-02 -3.90263423e-02 1.17024677e-02
-5.31148672e-01 5.31979501e-01 3.07687610e-01 5.14273167e-01
1.29851985e+00 3.40760827e-01 -1.03095186e+00 1.72198832e-01
7.51230657e-01 -2.10300609e-01 -1.75601259e-01 -3.47506911e-01
4.95864861e-02 -3.09154272e-01 -2.36174688e-01 -1.06592870e+00
-9.20433283e-01 9.39673781e-01 -4.17767882e-01 2.00072043e-02
9.33836520e-01 1.82994291e-01 1.14153624e+00 1.13147283e+00
7.13105977e-01 -8.62767339e-01 -2.50731129e-03 1.11875284e+00
9.91882265e-01 -1.34723926e+00 -7.76169598e-02 -9.30895865e-01
-5.46848714e-01 1.35918725e+00 1.03914022e+00 -2.35429540e-01
1.66178674e-01 2.76201636e-01 -1.74678743e-01 -5.27278543e-01
-8.54163170e-01 1.22594148e-01 3.37522447e-01 2.46973246e-01
3.49754125e-01 8.98161829e-02 -8.27316567e-02 1.29269326e+00
-8.48265588e-01 2.25230798e-01 8.88438001e-02 3.35730910e-01
-3.74273688e-01 -9.43338275e-01 -1.23512909e-01 2.40164414e-01
-2.50166096e-02 -4.06963021e-01 -6.27316475e-01 6.59691274e-01
4.73446667e-01 8.03367794e-01 -2.37965465e-01 -2.03717381e-01
1.48877501e-01 2.90604442e-01 7.97444344e-01 -6.46531343e-01
-4.85151678e-01 -1.26269925e+00 5.63361347e-01 -7.91518509e-01
-1.18519500e-01 2.75881767e-01 -2.16529560e+00 -7.75537729e-01
-2.17547044e-02 2.86690414e-01 -3.81526649e-02 1.65451133e+00
1.93641603e-01 5.83974481e-01 -4.36265059e-02 -1.14300866e-02
-4.31561053e-01 -4.57251489e-01 -1.50496677e-01 7.13132262e-01
-7.97495618e-02 -3.49760026e-01 -7.55918398e-02 8.10789540e-02] | [9.505708694458008, 7.495360374450684] |
94d5be59-1da5-4598-97a0-70e150d2ece7 | adopting-the-multi-answer-questioning-task | 2303.01064 | null | https://arxiv.org/abs/2303.01064v1 | https://arxiv.org/pdf/2303.01064v1.pdf | Adopting the Multi-answer Questioning Task with an Auxiliary Metric for Extreme Multi-label Text Classification Utilizing the Label Hierarchy | Extreme multi-label text classification utilizes the label hierarchy to partition extreme labels into multiple label groups, turning the task into simple multi-group multi-label classification tasks. Current research encodes labels as a vector with fixed length which needs establish multiple classifiers for different label groups. The problem is how to build only one classifier without sacrificing the label relationship in the hierarchy. This paper adopts the multi-answer questioning task for extreme multi-label classification. This paper also proposes an auxiliary classification evaluation metric. This study adopts the proposed method and the evaluation metric to the legal domain. The utilization of legal Berts and the study on task distribution are discussed. The experiment results show that the proposed hierarchy and multi-answer questioning task can do extreme multi-label classification for EURLEX dataset. And in minor/fine-tuning the multi-label classification task, the domain adapted BERT models could not show apparent advantages in this experiment. The method is also theoretically applicable to zero-shot learning. | ['Mohammed Ali Al-Garadi', 'Ying Wah Teh', 'Li Wang'] | 2023-03-02 | null | null | null | null | ['multi-label-text-classification', 'extreme-multi-label-classification', 'multi-label-text-classification'] | ['methodology', 'methodology', 'natural-language-processing'] | [ 2.05428332e-01 2.53754914e-01 -3.54521513e-01 -8.14255238e-01
-9.03185606e-01 -5.49927354e-01 8.71141627e-02 3.80592376e-01
-6.61557615e-01 6.30933940e-01 -1.21755578e-01 -1.98167995e-01
-5.24925053e-01 -5.34131110e-01 3.01319122e-01 -5.93017638e-01
6.07340455e-01 7.08157957e-01 3.50557566e-01 -3.14192623e-01
5.36306024e-01 -9.40946024e-03 -2.04815459e+00 8.07358384e-01
7.58708298e-01 1.11308801e+00 1.99709192e-01 7.17802465e-01
-6.02692008e-01 1.03042269e+00 -7.39033759e-01 -3.86260301e-01
1.37608245e-01 -4.63379323e-01 -1.08779657e+00 1.41747827e-02
6.67030692e-01 1.09399468e-01 4.71304417e-01 1.04092610e+00
7.32103050e-01 3.12567353e-01 1.20221949e+00 -1.73572707e+00
-6.56780720e-01 4.59437758e-01 -7.74197757e-01 3.58678736e-02
3.95500362e-01 -7.17419863e-01 1.30689526e+00 -5.12816668e-01
6.05833411e-01 1.19459367e+00 9.83557403e-01 6.08564019e-01
-9.62685645e-01 -6.99154615e-01 -5.79158030e-02 6.75762355e-01
-1.38181841e+00 4.89097089e-02 6.30063593e-01 -7.06187308e-01
7.27267206e-01 3.54177028e-01 -1.61941439e-01 6.80589259e-01
3.02066833e-01 3.25478822e-01 1.66296971e+00 -8.45373154e-01
3.72940600e-01 5.37171483e-01 1.30525351e+00 7.95972109e-01
-3.96527722e-02 -4.87253666e-01 -9.28112194e-02 -3.64509046e-01
-2.74971396e-01 -5.56981191e-02 6.90548047e-02 1.70737281e-02
-6.33082986e-01 1.25621223e+00 -1.90799370e-01 6.70006394e-01
1.80731550e-01 -2.38218680e-01 8.89513195e-01 7.44225681e-01
7.34592617e-01 4.47077215e-01 -7.17295647e-01 1.44462064e-01
-9.82366264e-01 -4.95840935e-03 9.59352911e-01 1.16309750e+00
9.10653949e-01 -5.28599620e-01 -3.96831036e-01 1.30153120e+00
1.91687956e-01 -1.46094441e-01 7.84686327e-01 -1.20996404e+00
2.32358202e-01 5.96739233e-01 -2.88328081e-01 -7.48976171e-01
-1.03276968e+00 -2.47377649e-01 -3.98239017e-01 3.03313345e-01
5.03452778e-01 -2.13469297e-01 -5.52292347e-01 1.53627574e+00
6.04412794e-01 -2.24606544e-01 -1.24580218e-02 2.94384420e-01
1.06251621e+00 5.43682516e-01 4.72378135e-01 -6.63523495e-01
1.73573267e+00 -1.30940843e+00 -1.15639424e+00 2.45168418e-01
1.42337739e+00 -6.88222170e-01 1.15680540e+00 3.30162227e-01
-4.97804880e-01 -8.84241879e-01 -9.89606142e-01 -3.27592701e-01
-8.21937919e-01 -1.60447098e-02 5.58454216e-01 1.03655469e+00
-7.25758255e-01 3.79552782e-01 3.60984027e-01 -5.60863316e-01
-1.62526488e-01 2.13259220e-01 -2.35065326e-01 -3.14558186e-02
-1.57468975e+00 9.98657405e-01 7.33857870e-01 -6.43123686e-01
-1.50921240e-01 -4.13178027e-01 -7.90926874e-01 2.56087054e-02
2.41515711e-01 -2.38726869e-01 1.23614907e+00 -7.72642195e-01
-1.27972054e+00 1.22189844e+00 1.70585196e-02 1.73944727e-01
2.39730164e-01 5.08337021e-01 -6.22757912e-01 1.75750613e-01
7.25543797e-01 5.35924494e-01 5.78557551e-01 -1.12827361e+00
-1.29809153e+00 -4.12638038e-01 6.89487681e-02 2.31979117e-01
-6.32988453e-01 2.34489173e-01 3.98858726e-01 -3.83523345e-01
1.91009015e-01 -9.33806300e-01 1.20602176e-01 -4.66024756e-01
1.68122843e-01 -1.12221563e+00 8.06356192e-01 -2.94449538e-01
1.38432467e+00 -1.96104467e+00 -1.41095743e-01 -2.75405318e-01
2.67527461e-01 -1.19674429e-01 -8.83202776e-02 3.76912713e-01
-2.94217199e-01 5.91060929e-02 1.62362576e-01 -2.77178526e-01
3.22256505e-01 1.72587350e-01 8.43904447e-03 5.91852188e-01
-6.66547000e-01 6.03393912e-01 -7.20415413e-01 -1.16855216e+00
-7.88354427e-02 -2.84844935e-01 -3.59618127e-01 1.19458027e-01
-2.80449707e-02 -8.14865977e-02 -2.33395547e-01 6.93835855e-01
7.46030271e-01 -1.22627757e-01 1.00101493e-01 -2.73835123e-01
-2.25240104e-02 -6.40055776e-01 -1.32989299e+00 1.60046446e+00
-4.99556869e-01 3.35211486e-01 -1.89569488e-01 -1.17898500e+00
9.17317450e-01 4.37145799e-01 6.43791914e-01 -5.74410021e-01
4.79119301e-01 1.80768490e-01 -1.16330750e-01 -8.94468307e-01
4.76933330e-01 -6.41672432e-01 -5.91436207e-01 6.86898589e-01
4.37684447e-01 7.30320290e-02 4.57406044e-01 -6.15990944e-02
8.17599595e-01 1.07573606e-01 5.28890669e-01 -6.09353542e-01
6.22432828e-01 8.51434246e-02 7.51399159e-01 7.60972977e-01
-6.40267134e-01 3.11741531e-01 5.21722436e-01 -1.98142350e-01
-6.27725780e-01 -5.40843070e-01 -7.56928086e-01 2.25647950e+00
1.33216441e-01 -4.23301011e-01 -6.24810874e-01 -1.26114821e+00
5.65730892e-02 1.09822297e+00 -7.46296406e-01 -1.18792281e-01
-5.13098463e-02 -7.37356007e-01 6.66654408e-01 -4.20793369e-02
4.83959436e-01 -7.30209827e-01 -4.85374093e-01 2.52027541e-01
-4.13408309e-01 -7.51568854e-01 -4.30017114e-01 7.49165297e-01
-3.73349011e-01 -1.11405540e+00 -7.63883829e-01 -1.23652816e+00
1.70576006e-01 2.84681886e-01 8.62537265e-01 -2.52199978e-01
-7.12775439e-02 4.95045364e-01 -8.01637292e-01 -4.47189838e-01
-4.52540278e-01 3.90300512e-01 4.10776921e-02 6.34473711e-02
9.96086478e-01 -2.68689960e-01 -8.31523836e-02 6.75544381e-01
-7.43638456e-01 -2.01295123e-01 4.81167226e-04 8.09445620e-01
2.40932554e-01 1.84298754e-01 1.33530962e+00 -1.29144454e+00
9.85853255e-01 -7.27242649e-01 1.06395528e-01 9.20468509e-01
-1.15194952e+00 -2.97044460e-02 5.46216547e-01 -5.73095560e-01
-1.15080380e+00 -2.39697352e-01 -3.12055945e-01 2.02725202e-01
-3.54364157e-01 1.91134751e-01 -5.72672747e-02 -1.81105375e-01
1.04647696e+00 -3.96540314e-01 -2.25253329e-01 -5.26517034e-01
5.01806319e-01 1.50393152e+00 -3.25788140e-01 -4.54486817e-01
-3.35690863e-02 1.26626194e-01 2.13177830e-01 -6.99795961e-01
-1.55271482e+00 -1.18474638e+00 -8.84970009e-01 -5.47800720e-01
1.18773293e+00 -5.58943093e-01 -7.89332092e-01 2.48000860e-01
-9.75505352e-01 4.14347500e-02 -1.53842941e-01 1.86164334e-01
-6.53295875e-01 6.56139195e-01 -9.31468785e-01 -6.95193112e-01
-2.85660088e-01 -9.69476461e-01 8.98459554e-01 2.20104754e-01
-3.87480229e-01 -1.10391140e+00 1.71996191e-01 7.87670851e-01
6.97634518e-02 2.57791989e-02 1.38144410e+00 -1.10077655e+00
5.84807277e-01 -1.64905697e-01 -2.10495234e-01 3.41087162e-01
1.61710456e-02 -5.42797506e-01 -1.12777126e+00 -3.85427058e-01
5.19347429e-01 -9.81101513e-01 8.45844150e-01 1.08678564e-01
9.76431251e-01 3.82276982e-01 -2.55121171e-01 3.29740971e-01
1.45080268e+00 3.58074158e-01 3.96571755e-02 4.84120905e-01
5.97391784e-01 1.13224626e+00 1.13542509e+00 2.16576487e-01
4.86996442e-01 5.60464084e-01 -2.48066857e-02 2.98991650e-01
-1.62519082e-01 3.08815032e-01 -1.70481473e-01 1.14904082e+00
3.42077434e-01 -1.62074238e-01 -9.52381015e-01 8.24010819e-02
-1.99744296e+00 -1.01978493e+00 -5.49182177e-01 1.69147062e+00
6.93637908e-01 -9.85670760e-02 -6.05022348e-02 2.90708452e-01
1.14243281e+00 2.86136344e-02 -2.80672044e-01 -8.18125665e-01
6.65569007e-02 -1.08217590e-01 3.64670068e-01 5.84310532e-01
-1.13976300e+00 5.87906063e-01 6.14870405e+00 1.36237562e+00
-5.24590254e-01 8.47218037e-01 5.54748416e-01 1.50586933e-01
2.34316476e-02 -7.93250278e-02 -1.37982070e+00 4.07966077e-01
1.22188795e+00 -1.67597458e-01 -1.73023418e-01 9.59534466e-01
-3.12223524e-01 -3.72350186e-01 -1.03928947e+00 1.11268365e+00
5.73679268e-01 -6.10285878e-01 -1.89845711e-01 -6.74965186e-03
7.18224108e-01 -5.52720726e-01 -1.85660005e-01 1.09581399e+00
2.77450830e-02 -5.73818445e-01 5.21584034e-01 2.79386044e-01
1.04879618e+00 -8.01067650e-01 8.16361845e-01 7.35690951e-01
-1.12988687e+00 -6.51221931e-01 -8.05268407e-01 -2.18256831e-01
4.99575175e-02 2.76537538e-01 -4.48176384e-01 5.68417549e-01
2.43715197e-01 3.23815435e-01 -9.46428835e-01 6.75649762e-01
2.77531922e-01 3.76315594e-01 1.79742664e-01 -3.23308647e-01
4.12279874e-01 -2.31366307e-01 -1.45357147e-01 1.33498108e+00
2.79837459e-01 -8.67151469e-02 5.90967059e-01 -9.22194403e-03
2.60093182e-01 7.61729479e-01 -5.29283226e-01 6.09975576e-01
1.91292346e-01 1.55550516e+00 -1.04010475e+00 -4.41813469e-01
-7.19208241e-01 1.07598746e+00 6.05343819e-01 1.57084107e-01
-9.88982975e-01 -8.54223549e-01 3.26141925e-03 -1.12809502e-01
-3.57779503e-01 3.27121466e-01 -4.46849078e-01 -9.54359293e-01
-5.15492320e-01 -5.88599622e-01 1.08777225e+00 -5.67177415e-01
-1.45234871e+00 4.38039452e-01 2.15800002e-01 -1.12549448e+00
-7.47370794e-02 -7.20794141e-01 -3.30224894e-02 7.21983194e-01
-1.19419706e+00 -1.21701682e+00 -3.91290247e-01 4.11671489e-01
8.60465527e-01 -3.21906388e-01 1.08084834e+00 5.68332255e-01
-4.59310353e-01 7.83037663e-01 4.36989367e-01 -3.08564037e-01
1.32632351e+00 -1.48978043e+00 -7.40397990e-01 -3.04319058e-02
-6.66647702e-02 4.41409498e-02 5.86052656e-01 -5.54828405e-01
-3.43421012e-01 -8.91900420e-01 1.29233694e+00 -3.94394070e-01
4.91664588e-01 -3.13375562e-01 -7.52604961e-01 3.79442126e-01
4.99345452e-01 -3.00083548e-01 1.63234174e+00 5.07816374e-01
-4.75548714e-01 -8.72778445e-02 -1.56828785e+00 1.64496958e-01
8.87323558e-01 -5.64530730e-01 -1.03019071e+00 7.94537008e-01
6.93193078e-01 3.07414919e-01 -1.03074324e+00 2.47928038e-01
4.42330807e-01 -8.91342640e-01 4.27276909e-01 -7.33542144e-01
3.73341948e-01 2.29307376e-02 -4.15895760e-01 -1.22896004e+00
-8.61491501e-01 1.97151601e-01 2.21099705e-01 1.57427025e+00
4.17505443e-01 -6.48622155e-01 5.83791196e-01 5.95826447e-01
-3.77843827e-02 -5.33888042e-01 -1.02248442e+00 -8.16146076e-01
5.66179037e-01 -1.22248791e-01 9.16940868e-02 1.55201209e+00
3.85888845e-01 1.00289261e+00 -2.64313310e-01 -3.49074215e-01
8.33556354e-01 2.99312025e-01 7.44480044e-02 -1.66918039e+00
-3.17201257e-01 -1.95645884e-01 -2.40063369e-01 -4.85636264e-01
8.25982630e-01 -1.53345656e+00 -1.37052208e-01 -1.61188936e+00
3.21111500e-01 -5.49512088e-01 -6.79411709e-01 2.81039774e-01
-2.77679950e-01 1.67461157e-01 2.13705346e-01 1.87550828e-01
-1.04708612e+00 1.47541031e-01 9.80263591e-01 -2.67299235e-01
1.08048081e-01 -4.16439511e-02 -7.65526652e-01 6.68623745e-01
9.41753507e-01 -9.61250544e-01 -7.53768086e-01 1.06372990e-01
2.25515708e-01 1.18417844e-01 -6.00523353e-01 -1.20205569e+00
3.81565630e-01 -1.51809871e-01 2.70460751e-02 -3.42483521e-01
1.58386022e-01 -9.94323432e-01 -1.60450533e-01 3.62892121e-01
-7.71088719e-01 -5.13570942e-02 -3.19708347e-01 6.41686916e-01
-5.45958504e-02 -1.38668144e+00 1.10025346e+00 -4.44621861e-01
-9.15461838e-01 -5.87579682e-02 -4.85241354e-01 4.50333118e-01
1.73636019e+00 -2.00852290e-01 -2.65360981e-01 3.03744618e-03
-1.06556451e+00 3.60789448e-01 1.45955533e-01 4.16411281e-01
3.16089280e-02 -1.43917012e+00 -6.53955638e-01 -2.69158036e-01
5.39351463e-01 -8.49131525e-01 4.52332228e-01 4.31603223e-01
-3.13464820e-01 3.60177964e-01 -2.41823569e-01 -3.67354423e-01
-1.66971028e+00 9.93238807e-01 2.44209051e-01 -4.15389478e-01
-2.13311508e-01 7.02229202e-01 4.88177314e-02 -7.92200565e-01
3.44626576e-01 2.71868646e-01 -9.05603349e-01 1.22671974e+00
5.05120456e-01 7.24481046e-01 -1.31122200e-02 -7.48549044e-01
-6.19261824e-02 8.70029628e-01 -2.52756745e-01 -1.00231707e-01
7.25563407e-01 -4.05639023e-01 -2.89486289e-01 1.16191983e+00
1.70821667e+00 -4.46845859e-01 -4.18690354e-01 -6.36153445e-02
5.56997657e-01 -1.04685754e-01 -1.51164075e-02 -6.70279741e-01
-2.92555928e-01 8.27152610e-01 8.85208726e-01 6.57390952e-01
7.83205152e-01 -5.97295947e-02 3.74794781e-01 4.72215325e-01
6.56844020e-01 -1.84825289e+00 1.51167288e-01 6.74070597e-01
4.93428022e-01 -1.33011508e+00 -1.97667435e-01 -5.16344845e-01
-6.56999171e-01 1.19195020e+00 8.11662614e-01 3.91197383e-01
1.07915533e+00 5.46338931e-02 3.37231249e-01 -3.10800701e-01
-6.15716040e-01 -3.83483112e-01 -4.81731966e-02 4.26547647e-01
5.98680079e-01 -1.12583172e-02 -1.25320423e+00 6.41775906e-01
2.65867651e-01 -1.29732370e-01 5.67210317e-01 9.96133685e-01
-1.09923089e+00 -1.21916974e+00 -3.31901461e-01 9.46411908e-01
-5.69594681e-01 6.48888871e-02 -5.65990470e-02 4.06474531e-01
7.80137777e-01 1.50543964e+00 -1.52094558e-01 -4.76536542e-01
3.46009731e-01 9.59765911e-01 2.46698305e-01 -9.63217974e-01
-8.09925199e-01 -3.39013338e-01 2.56512165e-01 -6.54726289e-04
-3.90446663e-01 -5.72294950e-01 -1.04025674e+00 -8.89191851e-02
-6.72691822e-01 6.80788517e-01 5.57602644e-01 8.48586977e-01
5.26297316e-02 6.43162966e-01 7.01201439e-01 -2.67084301e-01
-1.00688219e+00 -1.34126437e+00 -1.26934850e+00 8.09259892e-01
-2.42765814e-01 -8.71212304e-01 -6.91546679e-01 -2.36695215e-01] | [9.592610359191895, 4.387002944946289] |
32af5def-c3c8-40c4-a969-022555801420 | warped-convolution-networks-for-homography | 2206.11657 | null | https://arxiv.org/abs/2206.11657v2 | https://arxiv.org/pdf/2206.11657v2.pdf | Warped Convolutional Networks: Bridge Homography to sl(3) algebra by Group Convolution | Homography has an essential relationship with the special linear group and the embedding Lie algebra structure. Although the Lie algebra representation is elegant, few researchers have established the connection between homography and algebra expression in neural networks. In this paper, we propose Warped Convolution Networks (WCN) to effectively learn and represent the homography by SL(3) group and sl(3) algebra with group convolution. To this end, six commutative subgroups within the SL(3) group are composed to form a homography. For each subgroup, a warping function is proposed to bridge the Lie algebra structure to its corresponding parameters in homography. By taking advantage of the warped convolution, homography learning is formulated into several simple pseudo-translation regressions. By walking along the Lie topology, our proposed WCN is able to learn the features that are invariant to homography. Moreover, it can be easily plugged into other popular CNN-based methods. Extensive experiments on the POT benchmark, S-COCO-Proj, and MNIST-Proj dataset show that our proposed method is effective for planar object tracking, homography estimation, and classification. | ['Jianke Zhu', 'Wenyu Liu', 'Yang Li', 'Xinrui Zhan'] | 2022-06-23 | null | null | null | null | ['homography-estimation'] | ['computer-vision'] | [-1.77601874e-01 -8.34293514e-02 -2.07135618e-01 -3.03797662e-01
1.57709464e-01 -4.36012506e-01 8.54582548e-01 -6.30877376e-01
-9.89684388e-02 3.74408633e-01 1.26650974e-01 -2.44394362e-01
-5.36358580e-02 -1.05199170e+00 -1.08919823e+00 -7.55567014e-01
-6.28815964e-02 1.84585035e-01 1.30166665e-01 -2.49446124e-01
4.10639197e-02 5.81890941e-01 -9.65584993e-01 -1.64293110e-01
9.90851104e-01 1.00301945e+00 -2.49150209e-02 9.03804004e-02
1.97637126e-01 6.53979301e-01 2.00469866e-02 -4.40210253e-01
5.98539352e-01 -4.92573261e-01 -3.79544914e-01 2.44359642e-01
7.07548380e-01 -2.88969785e-01 -9.95424390e-01 1.31008637e+00
1.42665476e-01 1.12406619e-01 4.89580572e-01 -1.46189678e+00
-1.05049217e+00 7.17969060e-01 -4.58715349e-01 -1.73356295e-01
-4.45546806e-02 2.74902210e-02 1.04808867e+00 -8.63856673e-01
7.35484481e-01 1.54758966e+00 6.79124832e-01 1.70920864e-01
-1.22889173e+00 -9.84290838e-01 -3.29134136e-01 3.32717299e-01
-1.29046893e+00 1.09974481e-01 9.26443517e-01 -4.26786929e-01
6.98055625e-01 2.35808548e-02 9.82399762e-01 8.79401088e-01
3.81997824e-01 5.85268497e-01 8.24023545e-01 -1.08499736e-01
-1.50621653e-01 -2.49314234e-01 1.39028773e-01 1.10128570e+00
3.74488622e-01 2.15771466e-01 -1.78310871e-01 1.59637287e-01
1.24804902e+00 3.12239468e-01 -5.96568823e-01 -9.19385314e-01
-1.53588378e+00 9.21521842e-01 1.07950640e+00 2.29697436e-01
-1.29818311e-02 1.99642301e-01 2.77911812e-01 1.80066600e-01
-1.69933937e-03 2.69600719e-01 3.32695693e-01 4.72474754e-01
-5.17370284e-01 3.25147919e-02 9.47900712e-01 1.18251884e+00
9.82627690e-01 1.80193692e-01 1.06778681e-01 5.02079427e-01
3.08150619e-01 5.28706253e-01 4.88086224e-01 -8.82222533e-01
5.24775207e-01 7.83388615e-01 -4.61317509e-01 -1.47949421e+00
-2.39322081e-01 -6.15353525e-01 -1.42425883e+00 3.21445316e-02
3.44882667e-01 1.30034715e-01 -5.77822626e-01 1.54691768e+00
-5.14827250e-03 5.78064740e-01 -8.16177949e-02 9.95534122e-01
5.06289899e-01 5.18254399e-01 -2.28168324e-01 2.36398518e-01
1.46835876e+00 -9.44852233e-01 -6.07088268e-01 -7.25857094e-02
5.24263620e-01 -5.63820541e-01 6.37587249e-01 -1.79498330e-01
-8.43023181e-01 -5.29822528e-01 -1.71104252e+00 -3.92555296e-01
-3.00412178e-01 2.49747276e-01 6.90522730e-01 3.71796966e-01
-7.12308049e-01 1.04229259e+00 -8.48742425e-01 -3.14126760e-01
3.47607672e-01 4.58875120e-01 -7.03234851e-01 1.03193261e-01
-1.27246737e+00 8.13235104e-01 7.04917550e-01 1.89180478e-01
-4.28137571e-01 -7.35467970e-01 -1.03379238e+00 1.93681940e-01
1.70665056e-01 -9.65953708e-01 7.57367134e-01 -5.49056470e-01
-1.37030888e+00 7.32698262e-01 3.76821518e-01 -8.09154928e-01
6.84987307e-01 1.71628490e-01 -2.81165779e-01 1.06028676e-01
-2.42292285e-01 6.11992955e-01 8.08724105e-01 -7.10826635e-01
-4.74312276e-01 -5.07855952e-01 1.79132428e-02 2.72511452e-01
-3.76095772e-01 -2.76892304e-01 -4.26929444e-01 -5.66131532e-01
6.43466175e-01 -1.10927010e+00 9.00812969e-02 1.44726709e-01
-4.67743188e-01 -2.07336366e-01 9.13957536e-01 -7.01858044e-01
1.00586355e+00 -2.17473960e+00 3.38784635e-01 3.95214468e-01
5.28176129e-01 2.70960003e-01 1.15875825e-01 1.87781245e-01
-3.28588754e-01 -2.01271191e-01 -1.67776942e-02 9.09923539e-02
1.12943418e-01 2.05562294e-01 -5.20175576e-01 8.32405210e-01
1.49526477e-01 1.20555019e+00 -7.23207712e-01 -3.98157686e-01
3.30181807e-01 7.09145784e-01 -5.48243523e-01 -8.09801891e-02
9.14613008e-02 3.02201509e-01 -4.57601488e-01 2.16072336e-01
9.04887676e-01 -1.20024361e-01 9.10437256e-02 -7.52528191e-01
-2.02405378e-01 1.23409973e-02 -1.00585783e+00 1.51093745e+00
-3.11582804e-01 7.46697366e-01 -2.64676005e-01 -1.11154306e+00
1.23747790e+00 -3.70390043e-02 4.38645422e-01 -4.24263358e-01
6.24103487e-01 2.23077163e-01 2.18966812e-01 -2.23572105e-01
8.81958678e-02 2.47068718e-01 2.95397639e-01 6.81336001e-02
3.47642690e-01 5.39255142e-02 1.40714958e-01 -3.58263031e-02
6.85145736e-01 7.11596757e-02 4.45959091e-01 -4.18212026e-01
8.76540124e-01 -3.15037251e-01 5.50366521e-01 3.66975605e-01
1.07273616e-01 4.50330257e-01 5.74722230e-01 -7.86823988e-01
-1.34161747e+00 -9.27973449e-01 -2.48147786e-01 2.83607572e-01
1.54129922e-01 -2.59343088e-01 -8.78524721e-01 -2.30656400e-01
7.79172257e-02 2.40765378e-01 -6.05312169e-01 -5.14990032e-01
-8.99007976e-01 -3.62749338e-01 5.57020485e-01 5.49450576e-01
1.51160598e+00 -9.08798158e-01 -2.49425232e-01 1.92679942e-01
-4.56679724e-02 -1.26125681e+00 -9.61179137e-01 -1.73348173e-01
-9.85027075e-01 -1.25659931e+00 -7.02706158e-01 -1.01585698e+00
7.08499491e-01 5.00467718e-01 5.30975819e-01 -2.96678338e-02
-7.48122111e-02 1.75753683e-01 1.18276872e-01 7.10821594e-04
-2.29372501e-01 5.58157079e-02 -5.63987494e-02 4.41361398e-01
4.44204628e-01 -8.85272622e-01 -5.41393459e-01 6.75118983e-01
-7.50018179e-01 2.20802158e-01 5.41532874e-01 9.98519301e-01
4.77294713e-01 -1.42080143e-01 -1.27565622e-01 -5.73428810e-01
3.07815045e-01 -2.42732957e-01 -1.09287274e+00 3.40325654e-01
-4.01793897e-01 3.15552890e-01 6.70592308e-01 -6.06335044e-01
-8.21891725e-01 5.22654414e-01 2.70920098e-01 -9.63480175e-01
3.67794663e-01 4.17254984e-01 -4.76889670e-01 -7.55060196e-01
5.25057912e-01 3.25478554e-01 3.62620920e-01 -4.36389685e-01
4.37792391e-01 2.15065017e-01 9.90658939e-01 -3.03835362e-01
1.37977755e+00 6.41809464e-01 5.54143190e-01 -7.88832486e-01
-6.60617888e-01 -5.91153093e-02 -9.26111877e-01 -1.49431061e-02
1.06473017e+00 -9.95453417e-01 -1.07884872e+00 5.40334284e-01
-1.28324294e+00 1.59299865e-01 1.49472147e-01 7.50587344e-01
-6.33064926e-01 8.05797040e-01 -4.66983736e-01 -2.96586335e-01
-2.54390419e-01 -1.14141309e+00 6.84245825e-01 2.46341050e-01
1.81780949e-01 -1.00666928e+00 -2.76236713e-01 2.24635556e-01
9.82805192e-02 2.94487089e-01 9.26238239e-01 -5.30336976e-01
-1.17781448e+00 -3.27041894e-01 -4.14862186e-01 4.78500634e-01
-2.70317882e-01 -1.75925359e-01 -4.70554382e-01 -2.72607237e-01
2.88315415e-01 1.10185258e-01 9.75685775e-01 1.71110272e-01
1.07318723e+00 -3.66311759e-01 -1.76918432e-01 1.36118078e+00
1.30334342e+00 3.24522048e-01 8.44134748e-01 4.13145661e-01
1.07916415e+00 3.10817003e-01 7.36249462e-02 1.31268753e-02
2.25753888e-01 6.49085104e-01 5.06548941e-01 1.12469152e-01
-1.20397359e-01 -7.31338501e-01 3.14680070e-01 1.02756882e+00
-3.41477364e-01 4.11235631e-01 -6.96328044e-01 -2.51848064e-02
-1.96664584e+00 -8.97583425e-01 -2.28147775e-01 2.20192838e+00
4.56939310e-01 -4.91278283e-02 -1.58793867e-01 -9.31510404e-02
9.07106757e-01 1.85008183e-01 -4.59020793e-01 3.12201500e-01
-4.71979439e-01 -6.53908923e-02 7.84056306e-01 4.07740176e-01
-1.28606164e+00 9.63740230e-01 5.04685783e+00 8.25479984e-01
-1.07755256e+00 -1.37094617e-01 5.92241846e-02 6.15703583e-01
-4.18841355e-02 2.40582198e-01 -8.47065210e-01 2.69397229e-01
1.65224671e-01 -6.74033344e-01 8.02250385e-01 9.19469833e-01
-1.75081432e-01 5.74163139e-01 -1.27525198e+00 1.21345401e+00
3.30201328e-01 -1.50459921e+00 2.78949678e-01 2.52973467e-01
7.50521064e-01 -1.28950909e-01 1.58288613e-01 1.66583702e-01
-2.86156428e-03 -7.74279058e-01 5.13786495e-01 3.59631866e-01
5.88794529e-01 -5.94235480e-01 6.05819881e-01 2.37425402e-01
-1.45409203e+00 -1.02207914e-01 -8.17373693e-01 1.25862181e-01
-6.98364154e-02 1.00355059e-01 -6.52763546e-01 8.06602538e-01
4.66330647e-01 1.18927348e+00 -5.01231313e-01 1.21583879e+00
-2.37792179e-01 8.68155658e-02 -2.04492420e-01 -4.23080437e-02
3.52890402e-01 -9.14325237e-01 5.74428022e-01 7.54317284e-01
6.57662988e-01 2.44732559e-01 7.34340260e-03 1.23716450e+00
-3.16006958e-01 1.77236527e-01 -9.97260332e-01 -4.67093438e-02
2.66849637e-01 1.28603041e+00 -6.51615202e-01 -1.75282568e-01
-4.20817226e-01 5.58515310e-01 1.11549109e-01 4.36965942e-01
-7.53211081e-01 -5.26768982e-01 6.36890709e-01 -1.42555147e-01
3.21608573e-01 -4.84520614e-01 -9.11587179e-02 -1.50534761e+00
-5.10665253e-02 -8.07478964e-01 1.07880987e-01 -7.92031229e-01
-1.13918924e+00 5.27996063e-01 4.10181209e-02 -1.51344657e+00
-4.56050523e-02 -1.00624847e+00 -8.25687468e-01 8.08926523e-01
-1.08421457e+00 -1.25367916e+00 -6.41528010e-01 8.37550759e-01
1.08689472e-01 -5.04123449e-01 4.26220447e-01 4.30999219e-01
-4.09139574e-01 4.24292564e-01 4.09755707e-02 5.91315866e-01
5.07851601e-01 -1.10553861e+00 4.43548858e-01 6.18048608e-01
1.98528871e-01 8.25338125e-01 3.86193067e-01 -5.78007579e-01
-1.60912764e+00 -1.14891458e+00 7.78546572e-01 -1.33227348e-01
1.09443235e+00 -4.67402935e-01 -1.06378365e+00 1.06792355e+00
2.14665547e-01 1.47628365e-03 7.53393620e-02 -3.72366548e-01
-7.66962290e-01 -3.43121648e-01 -6.72588766e-01 8.94649446e-01
1.30735934e+00 -5.86160064e-01 -5.24700046e-01 4.42119420e-01
7.60614216e-01 -6.33992851e-01 -8.97247195e-01 4.47152019e-01
6.53084695e-01 -9.56947505e-01 1.15720201e+00 -6.05556846e-01
4.71444219e-01 -3.43966663e-01 3.73465731e-03 -1.24834037e+00
-4.69783366e-01 -7.11411476e-01 1.34385601e-01 8.80137622e-01
-3.77435386e-02 -8.16126645e-01 6.13252044e-01 9.73365456e-02
-1.74395934e-01 -4.61495936e-01 -8.13047588e-01 -9.23334002e-01
1.47831753e-01 7.74086565e-02 7.47865617e-01 1.09224641e+00
-1.38747603e-01 4.13396448e-01 -6.17230892e-01 2.25542367e-01
8.53753924e-01 7.72128478e-02 8.81957412e-01 -1.32703876e+00
-2.10715950e-01 -5.70101380e-01 -9.22239125e-01 -1.19476008e+00
4.76616144e-01 -1.43075192e+00 -3.72783154e-01 -8.88276458e-01
1.18464887e-01 4.92761992e-02 -1.60594471e-02 3.84884536e-01
4.42780614e-01 1.62342519e-01 4.53487158e-01 2.91920036e-01
1.91599913e-02 6.83075905e-01 1.73517489e+00 -2.70846516e-01
-3.57237607e-02 -1.03404969e-01 -1.76878303e-01 6.84431791e-01
7.49733984e-01 6.04703054e-02 -2.65265286e-01 -3.37758958e-01
8.29153135e-02 -2.71660369e-02 6.89990819e-01 -1.23878729e+00
6.58970475e-01 2.02815562e-01 3.10744226e-01 -6.49204016e-01
1.80787399e-01 -9.81407464e-01 4.09754664e-01 5.88577628e-01
-1.25574172e-01 -3.63678299e-02 -1.70470327e-01 4.45301861e-01
-3.79303038e-01 -1.93016633e-01 8.61363947e-01 -1.86874662e-02
-5.07481039e-01 7.23509789e-01 3.54986519e-01 -1.62231416e-01
9.14450824e-01 -1.64942220e-01 -3.87221217e-01 -3.63104403e-01
-6.80374622e-01 3.32578689e-01 2.39604861e-01 4.56229657e-01
5.24909914e-01 -1.88902259e+00 -2.12839514e-01 5.95682263e-01
-1.45463571e-01 -1.62644535e-01 1.01479918e-01 9.50254202e-01
-9.12689090e-01 7.25995183e-01 -8.28763962e-01 -7.31231034e-01
-1.17219269e+00 5.49587488e-01 7.04005480e-01 6.17237762e-02
-9.54133689e-01 1.49256930e-01 5.49374402e-01 -3.48496228e-01
5.53595304e-01 -3.35775197e-01 -2.21664295e-01 -1.31430984e-01
1.85948595e-01 4.06934142e-01 -3.83886814e-01 -8.68414760e-01
8.83028209e-02 9.50625241e-01 2.53431290e-01 9.69756991e-02
1.20140088e+00 2.74673462e-01 -5.25681674e-01 6.19946308e-02
1.65380859e+00 -4.94599760e-01 -1.20652699e+00 -4.87812191e-01
-1.67695284e-02 -3.33154798e-01 -1.33297876e-01 1.40414506e-01
-1.33733654e+00 1.13740802e+00 3.16099077e-01 8.37362111e-02
7.92302668e-01 -1.96157530e-01 6.71336949e-01 7.76054442e-01
1.87839389e-01 -6.27303421e-01 7.96265826e-02 6.80712581e-01
1.27070582e+00 -8.58393371e-01 -1.73241928e-01 -8.47994447e-01
-3.83376956e-01 1.49079037e+00 6.60625815e-01 -8.30456316e-01
8.60804677e-01 -2.70879328e-01 -2.70329297e-01 -9.35866684e-02
-1.03505999e-01 6.24978766e-02 6.14662588e-01 4.61737692e-01
-3.91367637e-03 -7.14160055e-02 -3.48117352e-01 3.05624008e-01
-6.81201935e-01 -1.80183142e-01 3.68749171e-01 2.81961352e-01
-4.37707096e-01 -1.02960622e+00 -3.69678736e-01 4.30293269e-02
4.99116071e-02 1.52608752e-01 -3.72164369e-01 1.18991458e+00
9.98794511e-02 2.01474383e-01 2.67222077e-01 -4.88679111e-01
3.43486995e-01 1.48620782e-02 6.05112612e-01 -1.68470323e-01
-1.83833897e-01 2.25703105e-01 -3.64391208e-01 -6.71336770e-01
-2.58425981e-01 -4.23192263e-01 -1.01663339e+00 -6.12379074e-01
-3.53093892e-02 -1.56361610e-01 5.77553570e-01 7.69486785e-01
-4.43921983e-02 4.05983478e-01 6.18346691e-01 -8.08880746e-01
-9.92371678e-01 -7.59066224e-01 -7.30617821e-01 4.93370354e-01
1.09690212e-01 -8.23175609e-01 -4.99583483e-01 -3.82192358e-02] | [8.621180534362793, -2.2042086124420166] |
f6d69605-1504-483e-80ff-e9c1d7031e9e | semi-supervised-node-splitting-for-random | null | null | http://openaccess.thecvf.com/content_cvpr_2013/html/Liu_Semi-supervised_Node_Splitting_2013_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2013/papers/Liu_Semi-supervised_Node_Splitting_2013_CVPR_paper.pdf | Semi-supervised Node Splitting for Random Forest Construction | Node splitting is an important issue in Random Forest but robust splitting requires a large number of training samples. Existing solutions fail to properly partition the feature space if there are insufficient training data. In this paper, we present semi-supervised splitting to overcome this limitation by splitting nodes with the guidance of both labeled and unlabeled data. In particular, we derive a nonparametric algorithm to obtain an accurate quality measure of splitting by incorporating abundant unlabeled data. To avoid the curse of dimensionality, we project the data points from the original high-dimensional feature space onto a low-dimensional subspace before estimation. A unified optimization framework is proposed to select a coupled pair of subspace and separating hyperplane such that the smoothness of the subspace and the quality of the splitting are guaranteed simultaneously. The proposed algorithm is compared with state-of-the-art supervised and semi-supervised algorithms for typical computer vision applications such as object categorization and image segmentation. Experimental results on publicly available datasets demonstrate the superiority of our method. | ['Luming Zhang', 'Zicheng Liu', 'DaCheng Tao', 'Xiao Liu', 'Mingli Song', 'Chun Chen', 'Jiajun Bu'] | 2013-06-01 | null | null | null | cvpr-2013-6 | ['object-categorization'] | ['computer-vision'] | [ 1.45290479e-01 -1.75376255e-02 -7.59834945e-01 -7.17896521e-01
-6.87292457e-01 -4.13644791e-01 1.45895511e-01 -2.64644474e-01
-3.48362774e-01 8.11215043e-01 -1.00323096e-01 -2.06894740e-01
-2.49447986e-01 -6.21260166e-01 -3.92974496e-01 -9.70518947e-01
4.26086754e-01 6.86036110e-01 1.78632215e-01 4.37889934e-01
2.75651634e-01 5.06240189e-01 -1.40483057e+00 -2.11452350e-01
1.55970335e+00 7.95462608e-01 8.90803188e-02 2.13163793e-01
-2.68303573e-01 -8.04054961e-02 -1.11374550e-01 1.78592190e-01
5.53449094e-01 -2.66951412e-01 -9.57923949e-01 9.58334923e-01
4.49406952e-01 -2.38010511e-01 1.23876408e-01 1.25484133e+00
1.15816206e-01 3.23820934e-02 1.11549902e+00 -1.48275375e+00
-3.06714803e-01 3.08106065e-01 -1.00620270e+00 -1.89390764e-01
-3.04740876e-01 -2.60198236e-01 8.49224448e-01 -9.82581496e-01
7.45179176e-01 8.94282401e-01 4.46479797e-01 3.58809918e-01
-1.55579162e+00 -6.70407236e-01 2.44854152e-01 7.02176467e-02
-1.56960607e+00 -3.37383091e-01 9.71296728e-01 -6.04571640e-01
1.84706151e-01 1.60474911e-01 4.20120865e-01 3.89245182e-01
-2.16179162e-01 9.53608453e-01 1.13844645e+00 -4.81080204e-01
4.37261105e-01 4.48306739e-01 5.86585462e-01 8.11564744e-01
2.90806383e-01 -1.05923258e-01 -2.59273440e-01 -2.88489342e-01
7.61393547e-01 2.02138290e-01 -2.30722949e-01 -1.30366230e+00
-1.09425545e+00 9.49934900e-01 3.31283063e-01 1.96115851e-01
-2.25191444e-01 -6.23270035e-01 5.66844270e-02 6.69694096e-02
5.58680952e-01 -1.08074032e-01 -4.02690291e-01 4.20700967e-01
-1.45880985e+00 -1.99084312e-01 7.15501070e-01 9.83180046e-01
1.02233624e+00 -1.03415839e-01 1.55197382e-01 1.01938224e+00
5.22016287e-01 3.42509031e-01 3.21657538e-01 -1.11440408e+00
5.06702423e-01 8.78122807e-01 1.07622363e-01 -6.19459152e-01
-2.57417828e-01 -4.37112898e-01 -1.07835531e+00 2.95771152e-01
6.01694643e-01 -4.28414941e-02 -1.09575844e+00 1.38848186e+00
9.43858445e-01 1.41173676e-01 -6.40449822e-02 9.50924158e-01
4.41388518e-01 5.32119811e-01 -1.60115361e-01 -6.80826724e-01
8.45977783e-01 -1.20933259e+00 -5.81573606e-01 -6.81207851e-02
5.65849602e-01 -5.33551991e-01 9.23673570e-01 2.92298585e-01
-6.70567334e-01 -5.96209109e-01 -1.09157908e+00 1.42600372e-01
-5.00870161e-02 5.99168777e-01 5.02142370e-01 6.91500545e-01
-5.53475022e-01 4.94834930e-01 -1.05305660e+00 -3.48222941e-01
5.06903052e-01 5.92779517e-01 -4.96261567e-01 -3.86315919e-02
-5.53406477e-01 4.01605606e-01 3.73754263e-01 2.06792310e-01
-5.39108992e-01 -3.73563647e-01 -7.21575618e-01 -1.51185662e-01
4.92275447e-01 -4.74834532e-01 9.56828713e-01 -8.92937005e-01
-1.45007324e+00 8.91956210e-01 -3.62469673e-01 -1.35880724e-01
5.80062568e-01 -1.46095395e-01 5.25799878e-02 2.17914343e-01
4.00912791e-01 6.43998504e-01 1.00694084e+00 -1.33757102e+00
-7.89770067e-01 -1.03743172e+00 -5.99821329e-01 4.46278036e-01
-6.74507797e-01 -3.12291145e-01 -4.78255063e-01 -3.67728770e-01
8.56407583e-01 -9.43107426e-01 -5.63784897e-01 1.98487580e-01
-6.23895645e-01 -2.60303825e-01 1.14776576e+00 -5.42534471e-01
1.07416487e+00 -2.02391171e+00 3.04130763e-01 4.04196352e-01
2.58186936e-01 1.23359971e-01 1.98881105e-01 6.39067311e-03
2.08703354e-01 7.82886073e-02 -6.49940312e-01 -1.20552845e-01
-3.42606068e-01 1.77656874e-01 -3.47940773e-02 9.39843774e-01
-9.64530855e-02 3.89417291e-01 -6.64772391e-01 -9.59332764e-01
1.54979244e-01 1.24108382e-01 -3.22722137e-01 5.23901284e-02
6.02402911e-02 5.00289261e-01 -8.26241434e-01 7.95001686e-01
1.04347265e+00 -2.66513675e-01 1.76584661e-01 -5.08594178e-02
-2.39152531e-03 -5.27962558e-02 -1.61326337e+00 1.61990571e+00
-1.49471527e-02 1.57950461e-01 5.08229792e-01 -1.45509875e+00
1.00161469e+00 7.77071863e-02 7.57775903e-01 2.29365692e-01
-4.06304337e-02 2.81569660e-01 -2.30737194e-01 -2.88838089e-01
1.07648917e-01 -2.54607379e-01 1.61184132e-01 3.54094684e-01
1.43957660e-01 -1.92611217e-01 2.68337369e-01 6.87072575e-02
4.71530318e-01 1.67169407e-01 5.04769504e-01 -5.46875477e-01
6.07320130e-01 1.57969192e-01 1.03024948e+00 4.24407601e-01
-4.15761262e-01 5.46764672e-01 2.05525443e-01 -1.29879788e-01
-8.52178335e-01 -1.00303423e+00 -5.18542111e-01 8.22402954e-01
4.07271326e-01 -4.66764010e-02 -9.82870042e-01 -1.19053423e+00
1.42515227e-01 4.75184411e-01 -4.21448857e-01 1.86481342e-01
-3.66184801e-01 -7.86795259e-01 8.08731765e-02 6.12407863e-01
5.34967065e-01 -6.37516439e-01 -2.67121583e-01 5.85648641e-02
-2.03341022e-01 -8.30035031e-01 -5.14925301e-01 4.54667389e-01
-1.45644498e+00 -1.00312591e+00 -6.54034257e-01 -1.12801170e+00
1.22227144e+00 6.88723624e-01 6.16149008e-01 -7.25333020e-02
-8.26336071e-02 -1.15976967e-01 -8.55603144e-02 1.92404538e-02
-2.43747771e-01 4.87548083e-01 1.46290928e-01 8.13875273e-02
4.13316399e-01 -5.64402103e-01 -3.21523458e-01 7.77498007e-01
-7.77410746e-01 1.34271592e-01 5.76440573e-01 1.10112476e+00
8.24003339e-01 2.85731643e-01 6.80581927e-01 -9.92117524e-01
2.14671612e-01 -3.01713735e-01 -7.77171969e-01 3.67437273e-01
-9.10680354e-01 1.62103042e-01 6.26776814e-01 -3.45316678e-01
-1.18411446e+00 6.63555622e-01 3.39608014e-01 -2.71122038e-01
-5.24975300e-01 2.72220910e-01 -5.56394160e-01 -2.42008138e-02
4.98475790e-01 2.72004277e-01 1.37987062e-01 -6.68501794e-01
5.45329690e-01 1.19545138e+00 3.64007562e-01 -5.07525325e-01
1.06910157e+00 8.12670350e-01 -5.36926230e-03 -8.57330739e-01
-8.52647841e-01 -1.17314386e+00 -1.36997378e+00 -1.71817224e-02
5.40151179e-01 -5.91013432e-01 -1.15059234e-01 3.93094718e-01
-5.68453789e-01 3.95362154e-02 -3.99155885e-01 6.72609091e-01
-5.50436497e-01 8.01184654e-01 -4.10841942e-01 -8.33043098e-01
-3.47966820e-01 -1.02423227e+00 9.30549502e-01 4.32051748e-01
2.46912360e-01 -7.77229011e-01 1.59106761e-01 7.45656729e-01
-2.86552697e-01 -1.77766442e-01 6.90257013e-01 -6.65329516e-01
-3.54103565e-01 -3.13563317e-01 -1.70141235e-01 4.52638805e-01
4.56899047e-01 1.28854886e-01 -5.86605072e-01 -3.47411335e-01
2.31967717e-01 -5.10477841e-01 1.01470041e+00 6.53620064e-01
1.04635310e+00 -1.15691178e-01 -7.32587039e-01 7.66659915e-01
1.27074111e+00 7.08131418e-02 8.37971494e-02 1.55302837e-01
7.54477561e-01 8.83397758e-01 9.19716477e-01 1.83796853e-01
1.86713964e-01 4.28902566e-01 1.09809794e-01 -2.14477733e-01
3.01327527e-01 -1.92417249e-01 -5.03890887e-02 7.87608385e-01
1.07026845e-01 9.93620083e-02 -8.39798689e-01 5.68809927e-01
-1.95170772e+00 -6.36368990e-01 -1.97567418e-01 2.45358610e+00
8.89116943e-01 1.01417974e-01 2.29369640e-01 4.20896798e-01
1.06070459e+00 -5.95488138e-02 -8.11247826e-01 7.24281147e-02
1.60756618e-01 -3.56175810e-01 6.23914778e-01 5.55790842e-01
-1.38869572e+00 9.68729675e-01 5.93025255e+00 9.57662880e-01
-1.04821110e+00 -2.28700340e-01 7.21380949e-01 2.41184756e-01
-1.07699215e-01 2.09244072e-01 -1.06227696e+00 1.51355132e-01
3.13354254e-01 8.94654095e-02 2.11438581e-01 1.05182624e+00
1.23389855e-01 -1.74337775e-01 -8.88140798e-01 6.07752323e-01
-7.23772421e-02 -8.83717775e-01 -1.56446218e-01 2.16093555e-01
8.56149435e-01 -2.74831921e-01 -2.17577264e-01 4.71674614e-02
2.09897354e-01 -8.05094540e-01 3.03182155e-01 1.01992778e-01
6.15342617e-01 -5.44179142e-01 4.34896290e-01 9.01176572e-01
-1.12191343e+00 1.02800734e-01 -4.79730278e-01 1.39005020e-01
5.59727401e-02 8.30765009e-01 -9.25930679e-01 6.05298340e-01
4.35666442e-01 7.54071176e-01 -4.98970896e-01 1.08756888e+00
-1.09708227e-01 7.85604596e-01 -5.93186736e-01 2.27454901e-01
1.18008465e-01 -8.80685031e-01 5.32704175e-01 7.38255560e-01
3.16147693e-03 8.23188946e-02 6.89738095e-01 6.55582190e-01
1.00329064e-01 4.58284557e-01 -4.75316554e-01 9.37548205e-02
4.94985878e-01 1.43254054e+00 -1.17647314e+00 -1.96035877e-01
-4.28602785e-01 9.10526514e-01 5.50486207e-01 4.75757539e-01
-4.35133785e-01 -1.56520024e-01 6.63416088e-02 9.29128379e-02
4.60012823e-01 -2.49766439e-01 -6.26783609e-01 -1.32663381e+00
7.94245005e-02 -6.02206945e-01 4.72028434e-01 -1.33611277e-01
-1.21221864e+00 2.76299089e-01 -3.51357250e-03 -1.44925010e+00
-1.78551331e-01 -3.44723403e-01 -4.05238628e-01 6.13441646e-01
-1.29867399e+00 -1.24000716e+00 -2.55610079e-01 4.31364000e-01
6.04707778e-01 -1.05801746e-01 6.70732677e-01 -1.21216610e-01
-8.07572246e-01 3.01263481e-01 6.34412527e-01 1.06064819e-01
7.21317708e-01 -1.35876155e+00 -1.83990121e-01 8.31017673e-01
3.42239290e-01 5.25769591e-01 4.52197224e-01 -6.88882887e-01
-9.21150744e-01 -9.11947787e-01 5.59929609e-01 2.07770422e-01
2.94005305e-01 -2.80654281e-01 -9.39552188e-01 5.50577343e-01
-2.44802684e-01 2.83103973e-01 9.33984518e-01 2.10023612e-01
-1.48676023e-01 -2.56382346e-01 -1.33602190e+00 3.47044647e-01
8.39235425e-01 -9.67502519e-02 -6.08284593e-01 3.20607334e-01
2.42314309e-01 -2.47092322e-02 -6.93243384e-01 6.78157568e-01
4.74900633e-01 -8.50176215e-01 7.41267741e-01 -6.64476395e-01
4.96630147e-02 -5.15693545e-01 -9.06815678e-02 -1.08696401e+00
-3.58519644e-01 -2.51410186e-01 1.13045201e-02 1.38642240e+00
3.23617697e-01 -4.50540155e-01 1.39875853e+00 5.05558789e-01
1.99138433e-01 -8.95410299e-01 -9.44217443e-01 -7.57065356e-01
1.56602561e-01 7.60936067e-02 8.26803520e-02 8.54950607e-01
2.78281625e-02 6.24253273e-01 -2.15430945e-01 2.00821996e-01
1.21351564e+00 7.42217720e-01 8.51969719e-01 -1.62331045e+00
-1.26382904e-02 -1.22755572e-01 -2.97676235e-01 -1.26853454e+00
4.15503412e-01 -8.32618296e-01 2.09034145e-01 -1.65191460e+00
6.08879626e-01 -6.22940123e-01 -1.55323282e-01 2.19275981e-01
-2.98005372e-01 -3.41129005e-02 -1.03771307e-01 6.89663291e-01
-5.05446196e-01 8.04171205e-01 1.29380906e+00 -1.49363697e-01
-5.45828342e-01 4.61551964e-01 -5.22039652e-01 1.00057185e+00
7.40274668e-01 -6.12732232e-01 -5.31534672e-01 -8.99381191e-02
-7.61326194e-01 2.21001044e-01 4.81061414e-02 -7.29251981e-01
1.58063158e-01 -5.51999688e-01 4.40370381e-01 -9.21029806e-01
8.77139717e-02 -1.06370485e+00 -1.44167170e-01 3.86290163e-01
-3.53636593e-01 -7.80278623e-01 -5.02127767e-01 8.45102012e-01
-2.08599821e-01 -5.35211802e-01 1.07523715e+00 1.56914502e-01
-3.04152071e-01 3.97147000e-01 -1.03623383e-01 -4.59916443e-02
1.23705637e+00 -4.53529000e-01 9.27302837e-02 -1.73722669e-01
-7.27792740e-01 5.13695657e-01 7.15000749e-01 5.10627441e-02
4.89242405e-01 -1.20210254e+00 -5.88161886e-01 3.38528216e-01
-6.77048936e-02 4.03227031e-01 -2.95369849e-02 7.56744862e-01
-2.95329273e-01 3.98496062e-01 -1.88148677e-01 -8.80873322e-01
-1.30970836e+00 6.14614069e-01 1.58946902e-01 -3.09180737e-01
-5.07701874e-01 4.84981686e-01 1.65527180e-01 -8.57325494e-01
3.33759636e-01 -2.77988791e-01 -2.55956024e-01 7.95603171e-02
6.23503551e-02 5.85029900e-01 -2.12013438e-01 -6.58154130e-01
-3.35828662e-01 6.89002156e-01 -1.88927680e-01 -1.98779926e-01
1.29556847e+00 -3.00404876e-01 -1.47216944e-02 4.85748440e-01
1.19348085e+00 -3.54398340e-02 -1.50716150e+00 -3.96243691e-01
8.98493752e-02 -5.45656979e-01 2.04119280e-01 -1.75440177e-01
-1.15938270e+00 8.07354450e-01 5.64008355e-01 1.85414001e-01
1.00533962e+00 -4.41022143e-02 6.13949418e-01 3.19722801e-01
3.06925386e-01 -1.34786093e+00 -2.31012702e-01 -8.55614990e-02
4.22113627e-01 -1.43175352e+00 3.34725469e-01 -1.04275727e+00
-5.58691800e-01 1.08265102e+00 8.11136186e-01 -1.57773584e-01
8.27665925e-01 3.18576656e-02 1.02560863e-01 9.86053571e-02
-2.58708924e-01 -7.43540823e-02 3.84708732e-01 5.63448966e-01
1.79981098e-01 1.64408639e-01 -6.10492945e-01 6.35580003e-01
7.37086982e-02 8.61534625e-02 2.08276302e-01 1.04434204e+00
-6.70224369e-01 -1.11530149e+00 -5.40643573e-01 6.14634871e-01
-2.08919257e-01 4.02782500e-01 -3.44352096e-01 5.28834641e-01
-7.43806735e-02 9.10753012e-01 -2.94517130e-01 2.41500754e-02
-5.90296835e-02 1.30642995e-01 3.37346226e-01 -7.15773523e-01
6.37093410e-02 5.03013015e-01 -2.67259270e-01 -1.94793269e-01
-6.47408366e-01 -7.81738579e-01 -1.31332004e+00 2.53843695e-01
-1.02344084e+00 3.34914416e-01 5.36081374e-01 8.91423404e-01
8.56147185e-02 -1.14608012e-01 9.79070604e-01 -7.16451705e-01
-1.17380726e+00 -9.77779627e-01 -1.00787163e+00 2.90954113e-01
7.02624768e-02 -7.28902876e-01 -5.66094160e-01 3.31840366e-01] | [14.7891263961792, -1.9806690216064453] |
f696785d-ae39-4087-9d7c-603e4edbbe41 | a-dual-source-attention-transformer-for-multi | 2306.05807 | null | https://arxiv.org/abs/2306.05807v1 | https://arxiv.org/pdf/2306.05807v1.pdf | A Dual-Source Attention Transformer for Multi-Person Pose Tracking | Multi-person pose tracking is an important element for many applications and requires to estimate the human poses of all persons in a video and to track them over time. The association of poses across frames remains an open research problem, in particular for online tracking methods, due to motion blur, crowded scenes and occlusions. To tackle the association challenge, we propose a Dual-Source Attention Transformer that incorporates three core aspects: i) In order to re-identify persons that have been occluded, we propose a pose-conditioned re-identification network that provides an initial embedding and allows to match persons even if the number of visible joints differs between the frames. ii) We incorporate edge embeddings based on temporal pose similarity and the impact of appearance and pose similarity is automatically adapted. iii) We propose an attention based matching layer for pose-to-track association and duplicate removal. We evaluate our approach on Market1501, PoseTrack 2018 and PoseTrack21. | ['Juergen Gall', 'Andreas Doering'] | 2023-06-09 | null | null | null | null | ['pose-tracking'] | ['computer-vision'] | [-1.75578028e-01 -3.96821499e-01 1.89886149e-02 -1.78872794e-01
-3.62460226e-01 -5.58598876e-01 6.62651896e-01 -1.71458900e-01
-6.84773684e-01 5.01984119e-01 5.43269694e-01 6.34279668e-01
-5.52538745e-02 -2.69482344e-01 -8.36256087e-01 -1.48231789e-01
-2.35063788e-02 6.49450004e-01 3.06027830e-01 -6.12587072e-02
-2.83496618e-01 5.27120769e-01 -1.66796720e+00 -2.01493591e-01
5.13724744e-01 8.19570899e-01 -9.73320380e-02 5.80217242e-01
3.42020363e-01 3.49839896e-01 -5.41021228e-01 -7.57773221e-01
6.03870332e-01 -1.17373198e-01 -4.18655723e-01 3.08653384e-01
1.32112646e+00 -4.66356128e-01 -6.68135226e-01 1.17261291e+00
6.93388760e-01 3.72859389e-01 5.13333619e-01 -1.46661627e+00
-2.77278095e-01 7.19048008e-02 -8.24433506e-01 2.71494120e-01
6.83023334e-01 1.38929278e-01 6.51268423e-01 -9.10957396e-01
8.88857782e-01 1.56973541e+00 1.01085222e+00 6.59026325e-01
-1.06787443e+00 -7.04203963e-01 4.75232661e-01 4.75920916e-01
-1.45594096e+00 -3.46528590e-01 6.24599457e-01 -6.23141527e-01
4.08954710e-01 1.36482686e-01 9.74052131e-01 1.41064203e+00
-6.77486975e-03 7.23073661e-01 3.72003049e-01 -1.64717063e-01
-2.11280584e-01 -2.38887444e-01 1.85456499e-02 5.07458746e-01
5.69853902e-01 9.99691337e-02 -6.81900561e-01 -1.52126504e-02
8.70505512e-01 3.97661209e-01 -2.75402158e-01 -5.74419975e-01
-1.30919862e+00 4.76318687e-01 5.52414417e-01 -1.14897145e-02
-4.44301873e-01 4.15297061e-01 5.06470978e-01 7.70783275e-02
2.90283471e-01 3.07958961e-01 -2.48927906e-01 4.05213842e-03
-7.30651081e-01 5.89828491e-01 5.16318738e-01 1.17444277e+00
5.50249100e-01 -2.55198538e-01 -6.67686343e-01 6.06323123e-01
3.65227014e-01 6.19959652e-01 2.17862889e-01 -7.39428520e-01
5.55462122e-01 6.71188593e-01 5.18264115e-01 -1.26194561e+00
-3.80864561e-01 -5.33068538e-01 -6.76273823e-01 7.33994832e-03
4.66425478e-01 -7.15318918e-02 -7.98309743e-01 1.99812841e+00
6.96710706e-01 5.16966224e-01 -6.27145112e-01 1.15652478e+00
6.77516460e-01 6.63043931e-02 1.32851005e-01 1.22528419e-01
1.73839593e+00 -1.18231308e+00 -9.27934349e-01 -5.16430616e-01
9.47808847e-02 -7.98549175e-01 4.18994099e-01 -3.00697554e-02
-8.85665715e-01 -1.04763472e+00 -8.77700031e-01 -1.77589014e-01
-1.98496029e-01 4.34583575e-01 3.65820676e-01 6.34639978e-01
-7.55003691e-01 5.36750615e-01 -8.19715023e-01 -5.52060962e-01
1.81726202e-01 5.01236856e-01 -6.00293100e-01 -1.58559635e-01
-1.14304793e+00 8.70712996e-01 1.62221774e-01 4.93877500e-01
-6.15879476e-01 -5.21000266e-01 -1.03643870e+00 -9.93961319e-02
5.85212886e-01 -1.13027632e+00 8.77544999e-01 -6.14329040e-01
-1.08035874e+00 7.71629274e-01 -2.62533665e-01 -2.44058460e-01
1.08011138e+00 -1.09291339e+00 -5.20034075e-01 -7.94962421e-02
3.18904936e-01 5.72018027e-01 9.95600581e-01 -9.13046539e-01
-7.35798001e-01 -7.31660068e-01 1.96691677e-02 4.52355832e-01
-5.13568878e-01 1.34395435e-01 -1.33350039e+00 -8.59903812e-01
1.27903625e-01 -1.28673947e+00 -2.13537946e-01 5.18539906e-01
-3.79980832e-01 -2.33454153e-01 6.04029655e-01 -1.01162267e+00
9.86923218e-01 -2.00540376e+00 5.81743896e-01 9.27265063e-02
3.14953417e-01 1.91621155e-01 -4.82586958e-02 8.97837132e-02
2.72995308e-02 -4.78068084e-01 5.08553207e-01 -8.69368494e-01
2.39383012e-01 -1.40073210e-01 1.42194614e-01 7.73424268e-01
2.16881514e-01 9.18162942e-01 -8.02482665e-01 -5.19554257e-01
4.55414861e-01 8.13666701e-01 -6.18890345e-01 2.83478588e-01
1.36780934e-02 7.50661314e-01 -2.03964233e-01 5.53840220e-01
7.88980544e-01 -1.47787079e-01 -2.36709863e-02 -8.13141525e-01
-6.57967031e-02 -2.17847228e-01 -1.78696752e+00 2.01286936e+00
1.30892605e-01 4.56019044e-01 -2.16024183e-02 -3.38033617e-01
5.36485851e-01 3.30972403e-01 6.93661213e-01 -3.74312967e-01
3.92719120e-01 -3.46740559e-02 -2.42299989e-01 -3.39635253e-01
6.53527379e-01 5.34714878e-01 -9.37465876e-02 2.38414869e-01
7.73647577e-02 7.79663086e-01 5.00724256e-01 3.47366333e-02
8.71638596e-01 4.78995204e-01 5.98857254e-02 3.93814109e-02
6.52555764e-01 -4.35778230e-01 8.26980531e-01 5.92810929e-01
-4.40974772e-01 7.93360531e-01 -1.83124289e-01 -8.75037014e-01
-1.01677704e+00 -1.00828004e+00 2.20878765e-01 1.11764717e+00
5.46014428e-01 -3.93123537e-01 -5.13290942e-01 -5.32701135e-01
2.22585231e-01 -3.05360317e-01 -8.31252158e-01 -1.83009669e-01
-8.47497404e-01 -3.29695970e-01 2.21808106e-01 7.82669961e-01
4.96612787e-01 -5.98693907e-01 -4.55205351e-01 2.18675584e-01
-4.34396207e-01 -1.43357909e+00 -1.25539362e+00 -5.53999722e-01
-3.83818060e-01 -1.25099504e+00 -1.09618866e+00 -6.26082838e-01
7.03524828e-01 2.45849326e-01 1.00704503e+00 1.91929229e-02
-4.74002391e-01 6.81448758e-01 -1.97969392e-01 -2.16173947e-01
2.06067234e-01 4.26990679e-03 5.46836734e-01 4.24035132e-01
4.77699488e-01 -2.77822971e-01 -8.73115957e-01 5.68024457e-01
-3.59176397e-01 -1.86251462e-01 4.38808888e-01 4.82109398e-01
4.77304280e-01 -2.79483557e-01 -3.95806041e-03 -3.96972209e-01
1.53641343e-01 1.31488238e-02 -5.15447855e-01 4.93877381e-01
1.32714084e-03 -7.03120083e-02 1.38378844e-01 -7.28129148e-01
-7.69314408e-01 5.17495334e-01 -5.08078840e-03 -8.95676672e-01
-2.07567848e-02 -1.95606172e-01 -3.65224004e-01 -1.74296960e-01
3.49672765e-01 -4.40183096e-02 -9.10908580e-02 -7.10749924e-01
3.64733666e-01 1.41431615e-01 1.05915320e+00 -4.92785215e-01
1.16085625e+00 6.75662994e-01 -2.01042831e-01 -5.48155248e-01
-8.42781723e-01 -8.41497064e-01 -1.01174939e+00 -5.59302986e-01
1.14142895e+00 -1.37817347e+00 -1.04142559e+00 4.30592716e-01
-1.47765982e+00 3.63200009e-01 -9.93372351e-02 7.12397099e-01
-2.82803804e-01 5.98157883e-01 -4.87255394e-01 -8.45960498e-01
-4.37011391e-01 -1.05546033e+00 1.30436873e+00 3.18238765e-01
-4.53548521e-01 -6.57440841e-01 3.00730556e-01 2.63807327e-01
1.56075548e-04 5.06079257e-01 -1.92190617e-01 -3.63854319e-01
-6.77412212e-01 -5.02365649e-01 -1.77594289e-01 -2.26875231e-01
1.46404207e-01 -3.54904085e-01 -7.48903990e-01 -8.23290288e-01
-3.19506079e-01 3.09074167e-02 7.41938710e-01 3.58727664e-01
6.29347146e-01 -1.97576687e-01 -6.61088288e-01 8.48063350e-01
9.95071471e-01 -3.36757451e-01 4.18955773e-01 4.72676337e-01
1.15159881e+00 5.63667476e-01 5.63571155e-01 5.61476409e-01
4.36109126e-01 1.25376034e+00 3.08036536e-01 -9.09667090e-03
-4.29033458e-01 -2.83133060e-01 3.44051570e-01 5.14702380e-01
-3.74131620e-01 8.84630904e-03 -4.85360503e-01 5.26229382e-01
-2.07563162e+00 -1.09110713e+00 -1.10817187e-01 2.40967846e+00
4.19977784e-01 1.84022710e-02 5.07138491e-01 -1.13310233e-01
1.22075260e+00 -2.78426819e-02 -6.38881743e-01 5.91905951e-01
-5.26772626e-03 -8.18906426e-02 4.65328991e-01 2.13126972e-01
-1.43260133e+00 6.86213970e-01 5.43440533e+00 3.84591550e-01
-6.33556962e-01 1.47635415e-01 1.28393034e-02 -3.03407788e-01
3.03880453e-01 -3.99191886e-01 -1.17666185e+00 6.38696313e-01
1.33072972e-01 1.05112053e-01 1.38913423e-01 6.59898758e-01
-1.77244574e-01 3.14723849e-01 -1.39441013e+00 1.30512190e+00
3.21871608e-01 -9.95637894e-01 -1.76462770e-01 1.92192681e-02
7.33629167e-01 -3.30547273e-01 2.14625336e-02 2.33494386e-01
6.43727481e-02 -6.44909263e-01 9.65707660e-01 6.05493367e-01
5.75153470e-01 -6.78447127e-01 6.94734097e-01 -1.03943817e-01
-1.88890934e+00 -1.43767595e-01 -2.77936757e-01 1.25483349e-01
3.91776294e-01 2.41293117e-01 -3.52226496e-01 8.56535912e-01
9.94934261e-01 8.13621461e-01 -7.64816284e-01 1.46285319e+00
-9.57896635e-02 -2.87077844e-01 -4.26501304e-01 4.31184381e-01
-3.61598551e-01 5.69063202e-02 6.33154750e-01 9.30835843e-01
3.04887891e-01 -4.36120868e-01 6.03236556e-01 6.43655241e-01
3.21764476e-03 -1.77675813e-01 -1.00558184e-01 4.12347734e-01
5.82905769e-01 1.08530474e+00 -5.32135546e-01 -1.70216665e-01
-4.53716666e-01 1.45801830e+00 3.69160920e-01 2.19942868e-01
-1.15332365e+00 1.08588852e-01 1.02022660e+00 2.12839663e-01
4.13820863e-01 -2.83351839e-01 2.98376143e-01 -1.31763458e+00
4.00705308e-01 -8.07366014e-01 6.36332154e-01 -3.98671508e-01
-1.47043753e+00 5.76097190e-01 -2.06196651e-01 -1.51494527e+00
-1.72929138e-01 -4.14483130e-01 -3.95648092e-01 6.52815998e-01
-1.13659370e+00 -1.49778962e+00 -6.52131677e-01 6.67994499e-01
4.50525254e-01 -7.38575235e-02 4.54922825e-01 1.02517521e+00
-8.79826069e-01 1.07075715e+00 -1.81796402e-01 6.28773808e-01
1.05982804e+00 -9.34911132e-01 7.34397113e-01 1.04988456e+00
2.04830796e-01 8.47448409e-01 9.01727617e-01 -9.86428797e-01
-1.54793310e+00 -1.32015622e+00 8.31660509e-01 -8.22077870e-01
2.90387392e-01 -5.92116475e-01 -6.39288962e-01 9.13808227e-01
-1.39938951e-01 1.62149265e-01 4.60335940e-01 1.99315697e-01
-4.34570998e-01 -1.39947757e-01 -6.77017212e-01 6.60914063e-01
1.53362954e+00 -3.41303080e-01 -4.78801638e-01 3.38016659e-01
5.30768275e-01 -7.90273130e-01 -7.59517968e-01 3.19560945e-01
8.97756636e-01 -7.09264815e-01 1.55777633e+00 -4.86274332e-01
-1.63947254e-01 -6.07021570e-01 1.49439961e-01 -8.54318321e-01
-6.98745549e-01 -7.43030667e-01 -4.38760638e-01 1.18070459e+00
-2.82269448e-01 -2.52189964e-01 9.53314006e-01 8.69168580e-01
2.32497036e-01 -1.71860859e-01 -1.05688989e+00 -1.05736029e+00
-6.94189131e-01 -1.14972241e-01 5.71689487e-01 6.55446172e-01
-5.97403467e-01 2.79825449e-01 -1.10153866e+00 4.86206919e-01
9.86770570e-01 1.64152738e-02 1.30427897e+00 -1.42527473e+00
-3.46713930e-01 -2.05452248e-01 -8.03377807e-01 -1.24377847e+00
-4.88856249e-02 -3.92424017e-01 1.08688585e-02 -1.26861238e+00
4.15596485e-01 -9.79797617e-02 -1.30810261e-01 1.72270253e-01
-5.02057672e-01 4.53520864e-01 5.41669667e-01 3.64636958e-01
-1.02315748e+00 7.35539556e-01 9.36346650e-01 -1.89661235e-01
8.58518202e-03 4.53644805e-02 -3.04764867e-01 5.85311413e-01
6.78835139e-02 -4.14392680e-01 2.96433475e-02 -6.71817541e-01
2.29309082e-01 -3.09768051e-01 8.40905368e-01 -1.57048118e+00
6.78310871e-01 2.68599004e-01 8.27603936e-01 -9.17752564e-01
7.82605231e-01 -1.07041740e+00 5.98978937e-01 6.03496373e-01
-7.10967183e-03 4.60103869e-01 1.78165212e-01 9.43891108e-01
1.36367958e-02 7.90264904e-02 4.80991542e-01 -1.45874796e-02
-8.90305281e-01 8.86214316e-01 2.84369856e-01 6.81888238e-02
1.01608241e+00 -4.44849670e-01 1.57861747e-02 -3.50873262e-01
-7.48329818e-01 5.30251086e-01 5.73859811e-01 9.00970459e-01
3.93512279e-01 -1.86068261e+00 -7.83164918e-01 9.86739621e-02
3.22414011e-01 -1.71239182e-01 5.85260749e-01 8.72748315e-01
-2.15553239e-01 9.54320803e-02 -3.39466065e-01 -7.62644410e-01
-1.45842516e+00 5.37754357e-01 2.93811560e-01 -1.85567901e-01
-7.22945154e-01 9.09360707e-01 2.17940718e-01 -1.78282812e-01
5.93057215e-01 1.51640102e-01 -3.20035338e-01 2.38851637e-01
8.25630844e-01 5.34686685e-01 -1.74821228e-01 -1.09921086e+00
-6.59175754e-01 1.06023371e+00 -1.24612257e-01 7.74954557e-02
1.15158284e+00 -3.07992846e-01 2.44047627e-01 7.70654380e-02
1.09557128e+00 -1.36712924e-01 -1.69699275e+00 -4.42991376e-01
-1.33738980e-01 -8.90067399e-01 -5.37655950e-01 -2.03216732e-01
-1.13172805e+00 5.12163818e-01 1.00156713e+00 -4.32719439e-01
5.63197851e-01 -1.34664163e-01 8.48366439e-01 1.91549763e-01
2.58256674e-01 -1.27197552e+00 3.25399220e-01 4.91393715e-01
9.07107830e-01 -1.27266157e+00 1.88370109e-01 -4.55367953e-01
-2.37905398e-01 8.67780924e-01 8.18077862e-01 -8.81505609e-02
3.00610334e-01 -5.17609939e-02 -1.83797404e-01 -1.70149192e-01
-1.01968706e-01 -4.52834725e-01 7.49105513e-01 7.03200877e-01
2.66767025e-01 -3.58411372e-01 -7.14987293e-02 5.24939597e-01
-8.24748725e-02 -9.32286456e-02 -1.01201475e-01 7.51477122e-01
-4.31875400e-02 -1.01154339e+00 -8.21995914e-01 4.68926467e-02
-1.31649613e-01 4.13391560e-01 -3.76346052e-01 6.57610893e-01
4.17744815e-01 6.02265775e-01 1.54883191e-01 -4.34605449e-01
6.17068529e-01 -1.12520576e-01 6.72920883e-01 -5.09178400e-01
-6.41573370e-01 1.74969658e-01 -4.95890677e-02 -6.38133585e-01
-6.04977906e-01 -8.07607770e-01 -6.42481089e-01 -3.66085738e-01
-4.03217047e-01 -2.86087066e-01 3.32255751e-01 8.11849058e-01
4.70846117e-01 6.57739699e-01 1.98826030e-01 -1.34768665e+00
-4.54526186e-01 -8.51727545e-01 -2.80282944e-01 1.14181697e+00
2.85400063e-01 -1.17934573e+00 8.82737339e-02 1.35852829e-01] | [6.896444320678711, -1.0988194942474365] |
e0f34ad8-12c3-42f5-a6d5-8749817199d8 | entailment-relation-aware-paraphrase | 2203.10483 | null | https://arxiv.org/abs/2203.10483v1 | https://arxiv.org/pdf/2203.10483v1.pdf | Entailment Relation Aware Paraphrase Generation | We introduce a new task of entailment relation aware paraphrase generation which aims at generating a paraphrase conforming to a given entailment relation (e.g. equivalent, forward entailing, or reverse entailing) with respect to a given input. We propose a reinforcement learning-based weakly-supervised paraphrasing system, ERAP, that can be trained using existing paraphrase and natural language inference (NLI) corpora without an explicit task-specific corpus. A combination of automated and human evaluations show that ERAP generates paraphrases conforming to the specified entailment relation and are of good quality as compared to the baselines and uncontrolled paraphrasing systems. Using ERAP for augmenting training data for downstream textual entailment task improves performance over an uncontrolled paraphrasing system, and introduces fewer training artifacts, indicating the benefit of explicit control during paraphrasing. | ['Rachel Rudinger', 'Balaji Vasan Srinivasan', 'Abhilasha Sancheti'] | 2022-03-20 | null | null | null | null | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 7.48875380e-01 4.80749100e-01 -2.37081915e-01 -6.23481095e-01
-1.10685802e+00 -8.17078233e-01 8.47762406e-01 2.13745877e-01
-2.89597154e-01 9.54140484e-01 8.36529791e-01 -7.17903912e-01
1.19014770e-01 -8.23079169e-01 -1.15496302e+00 1.59834340e-01
7.45688260e-01 7.48792470e-01 -1.85202077e-01 -5.88574052e-01
5.40094674e-01 1.21416546e-01 -1.24850130e+00 1.00226188e+00
9.88893867e-01 4.24997658e-01 1.04600772e-01 6.29343092e-01
-6.04661182e-02 1.26971102e+00 -6.20231509e-01 -8.60799730e-01
1.11973062e-01 -9.50887799e-01 -1.60004675e+00 -1.73216283e-01
4.62771088e-01 -2.81900644e-01 -2.51848012e-01 6.32260680e-01
6.54620156e-02 9.59041640e-02 6.64352834e-01 -1.10402703e+00
-1.08909941e+00 1.02372169e+00 8.01336840e-02 2.52631843e-01
1.12615001e+00 2.73322523e-01 1.67314732e+00 -1.08649850e+00
7.93246388e-01 9.82277453e-01 6.46000206e-01 6.66041017e-01
-1.69229722e+00 -3.03950042e-01 -2.34562576e-01 5.33272088e-01
-1.00608850e+00 -5.62579930e-01 5.80395281e-01 -4.85070096e-03
1.98156059e+00 4.23037201e-01 6.98652506e-01 1.37633419e+00
2.78626680e-01 7.76532590e-01 9.61478055e-01 -6.80298328e-01
9.53521281e-02 9.67867523e-02 2.65284538e-01 5.34216583e-01
2.06801742e-01 1.50665000e-01 -5.99243522e-01 -1.61478668e-01
3.38593125e-01 -3.68144661e-01 -1.57179758e-01 -2.43182331e-01
-1.29846394e+00 9.04618263e-01 3.82887185e-01 2.30862081e-01
-1.63719177e-01 1.42648828e-03 7.83014059e-01 1.06470227e+00
2.83645868e-01 1.46880174e+00 -4.48715895e-01 -1.20303854e-01
-1.14491749e+00 6.39186323e-01 1.08575881e+00 1.22269738e+00
6.64735079e-01 -2.01355532e-01 -5.36465406e-01 9.52116430e-01
-1.60133392e-01 3.67355466e-01 9.20914292e-01 -1.30196714e+00
9.20479774e-01 7.82212734e-01 3.25500607e-01 -3.95378739e-01
3.74777317e-02 -1.67362511e-01 -4.86700654e-01 -2.14636445e-01
2.14117110e-01 -5.59270121e-02 -3.61230254e-01 1.85201573e+00
-3.00490916e-01 -1.39982849e-01 6.52379215e-01 5.12566566e-01
8.43408585e-01 8.03623319e-01 -1.40557110e-01 -2.54552811e-01
1.28167593e+00 -1.23389399e+00 -4.63688850e-01 -7.61689425e-01
8.99171948e-01 -6.42560363e-01 1.72724640e+00 7.30307847e-02
-1.57317865e+00 -5.77443182e-01 -1.13019896e+00 -4.56792980e-01
7.94621091e-03 -1.20425165e-01 3.97726715e-01 3.00337486e-02
-9.65072274e-01 8.73875976e-01 -2.29045719e-01 -2.76752025e-01
4.71058413e-02 -5.87978885e-02 -4.19438392e-01 -3.22884351e-01
-1.58633339e+00 1.45571244e+00 4.55113620e-01 -2.82732457e-01
-6.26904309e-01 -1.01698852e+00 -1.29875171e+00 2.16477796e-01
2.57336646e-01 -1.27003884e+00 1.85083818e+00 -1.14503241e+00
-1.48713064e+00 1.18732464e+00 -4.79425430e-01 -8.31595898e-01
4.65518773e-01 -3.97811800e-01 2.04259325e-02 -2.70867236e-02
3.55922312e-01 5.87538838e-01 6.15198433e-01 -5.91211557e-01
-2.40690783e-01 1.26289934e-01 4.79978561e-01 6.44095480e-01
5.99518828e-02 1.26319110e-01 2.52654999e-01 -6.67193770e-01
-1.45017713e-01 -9.20383155e-01 -9.85706151e-02 -2.94764787e-01
-6.04310155e-01 -3.41900975e-01 1.00094453e-01 -7.79711962e-01
1.13726890e+00 -1.60654128e+00 6.23095453e-01 -2.18065962e-01
-4.15394157e-01 1.69439659e-01 -4.87693787e-01 9.88679826e-01
-4.84625548e-01 -9.22375359e-03 -4.42626566e-01 -3.17323178e-01
7.00434670e-02 4.99938875e-01 -6.95127130e-01 -2.30814561e-01
5.00260413e-01 1.50422597e+00 -1.18127823e+00 -4.91829723e-01
7.89617896e-02 -5.58039904e-01 -8.47536862e-01 6.45006061e-01
-3.54573697e-01 2.51496881e-02 1.83176309e-01 1.77461356e-01
1.41427755e-01 -2.64741004e-01 1.34092554e-01 -2.15678871e-01
4.35535908e-01 9.40991163e-01 -6.08655214e-01 2.04603457e+00
-1.31894863e+00 4.94922876e-01 -5.75983822e-01 -9.94519293e-01
7.34065294e-01 4.28724140e-01 -3.40661854e-01 -7.26470172e-01
-2.62486607e-01 2.84659088e-01 -1.19670592e-01 -5.83241999e-01
7.13439584e-01 -7.78788388e-01 -3.00009936e-01 1.09296477e+00
2.66774923e-01 -8.48145306e-01 5.13529420e-01 7.20745265e-01
1.19126248e+00 2.15925843e-01 7.91321158e-01 -1.44207552e-01
7.25799978e-01 3.44689429e-01 1.26148164e-01 1.08806920e+00
1.66787684e-01 6.36429191e-01 3.80960613e-01 -3.11743945e-01
-1.36699474e+00 -1.20018101e+00 2.00854406e-01 9.21539664e-01
-2.55212728e-02 -4.92118359e-01 -6.22741938e-01 -5.85875690e-01
-1.03335224e-01 1.39083993e+00 -3.89643222e-01 -6.80212259e-01
-8.14023852e-01 -4.86084372e-02 8.97384465e-01 7.32020617e-01
4.16709512e-01 -1.62068439e+00 -4.31939363e-01 2.18697444e-01
-9.16940153e-01 -1.21315503e+00 -6.49081230e-01 9.50502232e-02
-9.26129103e-01 -1.01787782e+00 -4.06886905e-01 -1.04196119e+00
5.46904445e-01 6.32635057e-02 1.75612175e+00 6.37742579e-02
1.74455062e-01 -1.92998722e-01 -4.65174347e-01 1.64911672e-02
-1.28634143e+00 3.91071886e-01 -2.51326174e-01 -6.26993239e-01
3.32766652e-01 -6.63011312e-01 -3.04557949e-01 7.09059462e-02
-6.96474671e-01 6.36121631e-01 6.96967304e-01 1.13943195e+00
2.45527714e-01 -7.40269542e-01 8.73749793e-01 -1.01720452e+00
1.36320221e+00 -3.06771070e-01 -1.25214411e-02 5.29442728e-01
-7.84014285e-01 5.51513553e-01 1.20109689e+00 -2.33184651e-01
-1.35736823e+00 -1.26262605e-01 -3.50845605e-01 1.24286013e-02
-2.14654505e-01 5.99805057e-01 1.28905639e-01 5.37801147e-01
1.29362333e+00 5.88747442e-01 7.86937848e-02 1.52885716e-03
8.72502506e-01 6.53548777e-01 7.63861895e-01 -8.00727189e-01
8.01901460e-01 -2.72369441e-02 -2.87233233e-01 -1.89644113e-01
-1.33806002e+00 -3.19494456e-01 -4.15445030e-01 4.95187283e-01
2.75285840e-01 -9.25974131e-01 -1.64805502e-01 -2.68519670e-02
-1.44453442e+00 -5.59923649e-01 -8.77595067e-01 5.94973601e-02
-1.04336226e+00 6.24989390e-01 -1.10885525e+00 -3.46312784e-02
-9.76592362e-01 -9.13190961e-01 9.43638265e-01 -2.46528745e-01
-1.30626464e+00 -7.36294985e-01 4.06217963e-01 6.80766463e-01
4.63361442e-01 -2.88230985e-01 1.31013906e+00 -8.96169603e-01
-1.17064968e-01 -1.91736862e-01 -2.73647070e-01 6.40588999e-01
3.00894856e-01 -2.41315141e-01 -4.69686598e-01 -4.10159007e-02
1.94053492e-03 -1.06176519e+00 5.23533523e-01 -1.25337183e-01
6.65067494e-01 -9.65957165e-01 1.80527255e-01 2.48411551e-01
9.46482360e-01 -3.32500845e-01 6.43713236e-01 2.58475095e-01
1.44669741e-01 6.36970460e-01 5.87215662e-01 1.36719376e-01
1.01796143e-01 6.34861469e-01 -6.81976229e-02 4.84020486e-02
-9.30247009e-02 -8.22671115e-01 3.47576976e-01 5.36542296e-01
4.96752858e-01 1.14652388e-01 -3.99271369e-01 7.73186684e-01
-1.84917605e+00 -1.39913929e+00 5.04453033e-02 1.75357187e+00
1.70379019e+00 2.99135417e-01 -8.53054821e-02 2.73419052e-01
3.35168123e-01 7.49128684e-02 -6.44102693e-01 -1.20921648e+00
9.43970233e-02 6.57982886e-01 -1.68484390e-01 9.80795443e-01
-3.82700801e-01 1.19912839e+00 6.24029636e+00 6.92752540e-01
-4.97561634e-01 -2.10834101e-01 6.36341721e-02 -1.82016566e-01
-6.41820729e-01 3.22211117e-01 -1.94582105e-01 1.35273784e-01
9.67143595e-01 -7.73242772e-01 8.41549754e-01 7.91454494e-01
2.05881551e-01 1.39704198e-01 -1.82026601e+00 5.17505527e-01
2.03965828e-01 -1.70782995e+00 6.78561628e-01 -7.21368968e-01
5.71553707e-01 -1.91220790e-01 -4.42711174e-01 6.64189994e-01
5.10929227e-01 -1.09971642e+00 7.89139569e-01 6.58472115e-03
6.59654915e-01 -6.46750331e-01 7.71220684e-01 6.16483450e-01
-8.17418396e-01 7.26191103e-02 -3.26307803e-01 -5.95604062e-01
4.71338779e-01 2.69064546e-01 -1.13191438e+00 6.33874118e-01
7.59749906e-03 8.78547788e-01 -7.13584244e-01 4.08510566e-01
-1.16147316e+00 6.23386025e-01 9.84580144e-02 -1.59389287e-01
1.28905103e-01 -9.04025510e-02 4.22100037e-01 1.52035165e+00
-7.46096596e-02 -4.46149334e-03 -3.43419969e-01 1.21747375e+00
-3.49159509e-01 8.38908330e-02 -9.02039409e-01 1.10800773e-01
5.82451582e-01 1.01089609e+00 1.66988745e-01 -7.59741604e-01
-3.22532982e-01 1.62028158e+00 6.16949260e-01 2.80948132e-01
-7.61438131e-01 -7.25998282e-01 3.87573749e-01 -8.24936479e-02
-7.37720281e-02 3.00741136e-01 -5.32129407e-01 -1.46936047e+00
2.08830014e-01 -1.32210195e+00 4.23241705e-01 -1.37326884e+00
-1.46704495e+00 5.95641196e-01 2.47035384e-01 -9.93498266e-01
-8.36183429e-01 -2.62603968e-01 -1.06458759e+00 1.02350390e+00
-1.27753913e+00 -1.11451840e+00 -4.08540405e-02 3.63014698e-01
1.11681342e+00 8.24704319e-02 9.84713376e-01 -2.44863898e-01
-1.23841621e-01 6.87161326e-01 -5.51751137e-01 -2.49608941e-02
7.27488041e-01 -1.35810220e+00 1.11018503e+00 8.61389101e-01
5.45695961e-01 1.08199906e+00 1.08072233e+00 -3.65717232e-01
-9.83639359e-01 -1.14429951e+00 1.71041083e+00 -6.64444923e-01
8.34309459e-01 -1.92917570e-01 -1.01364958e+00 9.84555900e-01
5.92759669e-01 -6.61056697e-01 6.16459668e-01 1.49356395e-01
-7.64452159e-01 1.25748649e-01 -1.08402693e+00 1.02355921e+00
1.53777635e+00 -1.07898653e+00 -1.81211460e+00 6.60796463e-01
1.06420302e+00 -4.44218755e-01 -5.80743492e-01 2.70006746e-01
4.25479144e-01 -6.36531711e-01 1.08336008e+00 -1.33277345e+00
1.45463955e+00 4.25634347e-02 -1.76195288e-04 -1.71977365e+00
-4.21132475e-01 -4.83937472e-01 -3.75745371e-02 8.46568465e-01
9.23673332e-01 -2.57970482e-01 6.96758389e-01 7.48444259e-01
-3.81049126e-01 -6.39974296e-01 -9.10631776e-01 -9.95182276e-01
4.31534499e-01 -1.57429621e-01 4.33678001e-01 7.97094107e-01
1.13653862e+00 1.31641722e+00 -2.96572655e-01 -3.42990994e-01
2.84830809e-01 7.15275824e-01 7.67412066e-01 -4.82819796e-01
-8.76414180e-01 -2.66390383e-01 2.62482941e-01 -1.11722195e+00
7.33354747e-01 -1.58305430e+00 1.08203933e-01 -1.99719644e+00
3.70455831e-01 2.79836744e-01 3.07372153e-01 6.52776897e-01
-2.76255071e-01 1.96423396e-01 -7.39550740e-02 2.38112226e-01
-3.82950246e-01 4.80666816e-01 1.07386672e+00 -2.83381760e-01
-1.60356864e-01 5.08610494e-02 -1.00500381e+00 4.84441549e-01
8.55412543e-01 -3.65847766e-01 -7.31416881e-01 -5.50574005e-01
5.81771374e-01 4.03511584e-01 3.09783608e-01 -4.81411844e-01
1.78772181e-01 -3.30967337e-01 -2.12116048e-01 -1.03199251e-01
3.24916914e-02 -2.58996278e-01 7.98083693e-02 6.33061171e-01
-1.30194032e+00 4.39261675e-01 2.09966183e-01 2.01884866e-01
-3.35319996e-01 -7.86113441e-01 7.13407040e-01 -4.46357340e-01
-4.40818757e-01 -5.43690085e-01 -6.95545197e-01 7.42998838e-01
6.55496061e-01 -2.22894609e-01 -3.78140599e-01 -6.19913280e-01
-4.36925977e-01 1.09173104e-01 5.29474854e-01 2.99869686e-01
7.52755702e-01 -1.20351410e+00 -9.40577388e-01 6.32730573e-02
6.31670952e-01 -2.16408700e-01 -2.50587940e-01 9.83615369e-02
-3.71750563e-01 4.57984209e-01 -2.44870052e-01 -1.15598422e-02
-1.36001456e+00 5.75004876e-01 3.82661343e-01 -8.38576674e-01
-6.09397471e-01 7.14863956e-01 -6.38985634e-02 -7.54850447e-01
-1.42066941e-01 -7.24459469e-01 1.27797812e-01 -5.15471995e-01
4.07233566e-01 8.55601393e-03 3.51084560e-01 -1.00933224e-01
-2.56407082e-01 2.08263963e-01 -3.77891511e-01 -4.33090299e-01
8.69258165e-01 7.05792308e-02 6.40421510e-02 6.20718524e-02
1.06973064e+00 -2.81299710e-01 -5.14357567e-01 -4.60879594e-01
1.70196161e-01 -2.13513538e-01 -7.42279708e-01 -1.06586254e+00
-2.31258735e-01 6.82434857e-01 -5.63379765e-01 -8.10648054e-02
6.35146499e-01 1.88001648e-01 1.01035082e+00 1.23475349e+00
3.37349623e-01 -8.46221745e-01 5.06805241e-01 7.68339396e-01
1.62303984e+00 -9.12980199e-01 -3.26144043e-03 -2.36892089e-01
-1.11745369e+00 7.83704042e-01 7.09068000e-01 -4.63916898e-01
-2.27774099e-01 6.92478642e-02 -1.99640304e-01 6.23585731e-02
-1.12170184e+00 2.42855340e-01 1.40391767e-01 4.23637062e-01
5.76107740e-01 -3.02976280e-01 -5.93867779e-01 4.38343614e-01
-7.35504329e-01 1.43691152e-01 7.50479698e-01 7.97358572e-01
-1.73476189e-01 -1.23099196e+00 2.04560280e-01 5.32218814e-01
-6.99027181e-02 -7.30146825e-01 -9.88804936e-01 3.45457494e-01
-5.89236200e-01 1.10894167e+00 -1.65039703e-01 -5.27359657e-02
6.96312964e-01 3.80804926e-01 8.88760388e-01 -1.12750518e+00
-9.76901650e-01 -1.03924859e+00 8.52867901e-01 -5.18504977e-01
3.53372097e-02 -2.43516490e-01 -1.26172841e+00 -2.56964147e-01
-1.99894652e-01 4.76951718e-01 2.29163980e-03 1.30193102e+00
4.20348495e-01 1.61381006e-01 5.04316926e-01 -5.01312196e-01
-1.20087969e+00 -1.07957184e+00 8.43311995e-02 1.21214342e+00
6.60430938e-02 1.21425062e-01 -4.31391507e-01 3.18074346e-01] | [11.59582805633545, 9.209705352783203] |
047eded3-4192-4ec8-a416-0ee428e626e0 | a-generative-approach-to-question-answering | 1711.06238 | null | http://arxiv.org/abs/1711.06238v2 | http://arxiv.org/pdf/1711.06238v2.pdf | A Generative Approach to Question Answering | Question Answering has come a long way from answer sentence selection,
relational QA to reading and comprehension. We shift our attention to
generative question answering (gQA) by which we facilitate machine to read
passages and answer questions by learning to generate the answers. We frame the
problem as a generative task where the encoder being a network that models the
relationship between question and passage and encoding them to a vector thus
facilitating the decoder to directly form an abstraction of the answer. Not
being able to retain facts and making repetitions are common mistakes that
affect the overall legibility of answers. To counter these issues, we employ
copying mechanism and maintenance of coverage vector in our model respectively.
Our results on MS-MARCO demonstrate it's superiority over baselines and we also
show qualitative examples where we improved in terms of correctness and
readability | ['Rajarshee Mitra'] | 2017-11-16 | null | null | null | null | ['generative-question-answering'] | ['natural-language-processing'] | [ 4.20407921e-01 7.51792192e-01 5.68987846e-01 -5.07242501e-01
-1.29062045e+00 -8.36403668e-01 7.23219335e-01 1.84501290e-01
-1.26213446e-01 9.31981742e-01 8.75886440e-01 -6.26453698e-01
9.62592736e-02 -1.07838726e+00 -9.74394262e-01 -1.54176988e-02
3.04786026e-01 5.20115674e-01 2.12614775e-01 -6.00606978e-01
4.45220023e-01 -3.03409517e-01 -1.09031487e+00 7.48366833e-01
1.10255885e+00 5.53589046e-01 3.85500133e-01 1.35629916e+00
-4.55431968e-01 1.65732169e+00 -1.00845885e+00 -8.16500604e-01
3.43011855e-03 -1.20266163e+00 -1.79904437e+00 7.74572566e-02
5.86881459e-01 -3.19339871e-01 -3.10678542e-01 8.39700520e-01
2.82476395e-01 2.95763671e-01 7.77597368e-01 -7.63481438e-01
-1.35694432e+00 7.17647433e-01 7.27598369e-02 4.26241577e-01
1.05023611e+00 4.36925232e-01 1.21674609e+00 -6.39079273e-01
7.00486362e-01 1.30095780e+00 4.83924717e-01 6.18733227e-01
-1.00287414e+00 1.61503375e-01 -6.97660670e-02 3.82206738e-01
-8.33014727e-01 -5.01522064e-01 3.65522712e-01 -2.35515639e-01
1.46183169e+00 5.70796013e-01 2.40668789e-01 9.93929803e-01
4.59870547e-01 6.82698309e-01 7.63847768e-01 -5.29881001e-01
6.78558722e-02 -4.20594066e-02 4.30522054e-01 6.68009043e-01
-2.42207959e-01 -3.58340532e-01 -2.27775812e-01 1.64731041e-01
4.00851727e-01 -4.24308956e-01 -3.71674418e-01 3.21292907e-01
-1.00193763e+00 9.98690546e-01 5.49181581e-01 -1.43875673e-01
-4.90112394e-01 1.53947592e-01 2.26039678e-01 9.24682081e-01
-1.09491199e-01 9.02721405e-01 -2.08953947e-01 -2.02394843e-01
-6.11726463e-01 7.11515129e-01 1.09758818e+00 1.15242374e+00
5.51706493e-01 -1.73732802e-01 -9.50701177e-01 5.75931132e-01
7.56102353e-02 3.24689478e-01 5.52660048e-01 -1.14491248e+00
9.75180030e-01 7.77131557e-01 2.30943531e-01 -7.20785260e-01
-1.31058529e-01 -5.46176493e-01 -6.07919276e-01 -3.74465466e-01
4.33231235e-01 -3.29208940e-01 -8.28943551e-01 1.75649738e+00
-1.95051879e-02 -2.51114935e-01 3.81973535e-01 7.18897820e-01
1.07364976e+00 9.99724388e-01 6.96551502e-02 5.86603303e-03
1.38198900e+00 -1.38513780e+00 -8.09939325e-01 -4.22611088e-01
7.38225818e-01 -6.81935668e-01 1.52379119e+00 9.66579691e-02
-1.57560790e+00 -8.34953904e-01 -9.49078321e-01 -7.64785647e-01
-1.17112771e-01 1.29546663e-02 1.21733263e-01 6.45028800e-02
-1.11091936e+00 4.40238804e-01 -3.56739104e-01 -3.18757117e-01
3.35722744e-01 2.12126538e-01 -1.75456315e-01 -7.48150349e-02
-1.25408936e+00 1.23487496e+00 2.10504681e-01 -1.28982455e-01
-8.50256979e-01 -3.01307976e-01 -8.57483029e-01 1.81757152e-01
3.28013748e-01 -1.40703130e+00 1.81490731e+00 -7.03138590e-01
-1.44415081e+00 6.24098539e-01 -4.49173361e-01 -7.18214035e-01
2.27262110e-01 -4.96929049e-01 -2.86429852e-01 1.42286032e-01
1.94885656e-01 7.86394298e-01 4.94263619e-01 -9.40171123e-01
-5.60977101e-01 1.22523252e-02 6.21524632e-01 4.24350262e-01
3.77434969e-01 -2.62655228e-01 -4.27821800e-02 -3.39004844e-01
7.52134621e-02 -5.56523740e-01 -1.03987597e-01 -7.14319706e-01
-7.28987694e-01 -4.77780193e-01 3.42570126e-01 -1.27308774e+00
1.49658728e+00 -1.52286136e+00 4.47679758e-01 -2.44985893e-01
1.84703127e-01 -1.66197354e-03 -4.58544433e-01 9.56809044e-01
7.94495121e-02 1.75145268e-01 -2.95455664e-01 -2.96817929e-01
1.05071433e-01 3.28661501e-01 -6.74032688e-01 -7.05026537e-02
6.02615237e-01 1.57379782e+00 -8.15250874e-01 -3.24788719e-01
-3.22001398e-01 -4.50488813e-02 -7.80032218e-01 8.46025646e-01
-9.12149727e-01 3.82547319e-01 -3.42118651e-01 1.27422199e-01
2.45918408e-01 -4.41095561e-01 -3.24133992e-01 2.42182493e-01
2.94189632e-01 1.21929109e+00 -7.20807254e-01 1.73380923e+00
-5.81729233e-01 5.35587251e-01 -3.82528096e-01 -7.29721129e-01
7.79865503e-01 4.39003885e-01 -7.75705218e-01 -9.13473606e-01
8.80673155e-02 -1.35395989e-01 2.04856575e-01 -1.25531793e+00
8.40289831e-01 -2.41554126e-01 2.79101566e-03 6.90592170e-01
1.78084716e-01 -3.94170552e-01 2.98979998e-01 6.36210322e-01
1.11164141e+00 3.23889434e-01 2.90247738e-01 -1.75964400e-01
7.50866771e-01 2.14711100e-01 -1.33289501e-01 1.10028994e+00
2.69007176e-01 7.92687535e-01 6.84822619e-01 -6.86733201e-02
-1.14678752e+00 -1.10420609e+00 5.76732218e-01 8.29787552e-01
-2.31351614e-01 -4.45949167e-01 -9.81995165e-01 -7.67995596e-01
-3.52698565e-01 1.41825628e+00 -7.12412715e-01 -2.12768421e-01
-9.18748558e-01 -1.30247772e-01 5.46822190e-01 5.77101707e-01
5.23173511e-01 -1.25485361e+00 -5.64598620e-01 4.09959584e-01
-6.68523133e-01 -7.49239385e-01 -4.42093045e-01 7.33432099e-02
-7.82896996e-01 -9.18112934e-01 -3.99726927e-01 -9.33221936e-01
5.10268629e-01 -9.93166938e-02 1.72799814e+00 3.71764868e-01
1.84995100e-01 1.89792499e-01 -5.07914603e-01 -3.44889581e-01
-9.09245789e-01 4.84887958e-01 -7.77496874e-01 -3.68574142e-01
2.50308722e-01 -4.99799669e-01 -7.65796483e-01 -2.54194617e-01
-1.09237075e+00 2.69638896e-01 3.19545805e-01 7.40360856e-01
2.28023365e-01 -5.44009387e-01 9.99543548e-01 -1.21631646e+00
1.33301938e+00 -7.91555583e-01 -2.05319207e-02 5.50085366e-01
-6.68248683e-02 4.32649583e-01 8.75284195e-01 2.32904609e-02
-1.01104736e+00 -4.90095675e-01 -6.53918207e-01 5.52026272e-01
-2.13345319e-01 4.73455131e-01 -2.47754641e-02 6.50555432e-01
8.99124265e-01 4.34947342e-01 -2.45208628e-02 -4.19681937e-01
8.47976387e-01 6.80616677e-01 8.14009428e-01 -4.84291971e-01
7.70467281e-01 -1.30293891e-01 -2.91425318e-01 -4.83951777e-01
-1.29668772e+00 -1.63490787e-01 -2.51036465e-01 7.39161521e-02
1.12758410e+00 -6.30067885e-01 -5.29710352e-01 -1.22241504e-01
-1.52153099e+00 -2.16887355e-01 -4.61800426e-01 -1.60687432e-01
-6.80739522e-01 2.85522312e-01 -7.34177291e-01 -7.71550059e-01
-6.15687072e-01 -9.62874174e-01 7.59641290e-01 3.19455624e-01
-7.72127569e-01 -1.00481451e+00 2.22690046e-01 6.45973265e-01
7.21924484e-01 1.84386835e-01 1.49621737e+00 -6.72966301e-01
-7.66343772e-01 3.20101529e-02 -1.52440658e-02 4.91980046e-01
7.91382492e-02 -4.51259077e-01 -8.39472115e-01 -4.44739722e-02
4.00219262e-01 -6.29185736e-01 7.05395103e-01 -8.72520134e-02
8.56275499e-01 -8.24765801e-01 2.31033415e-01 9.95765403e-02
1.49204838e+00 1.08965211e-01 1.02874434e+00 2.28466392e-01
4.87666488e-01 7.76223481e-01 2.08152961e-02 -1.63223684e-01
7.98850596e-01 3.26993495e-01 3.30524355e-01 2.11461380e-01
-5.84279537e-01 -7.33796120e-01 2.31431082e-01 1.25366378e+00
2.72950202e-01 -6.34513736e-01 -6.12896144e-01 7.23424971e-01
-1.60154438e+00 -1.05056262e+00 -5.74023128e-01 1.90924144e+00
1.05581403e+00 1.19564667e-01 1.91684831e-02 -2.09269263e-02
1.50300354e-01 -3.63360867e-02 -2.20897511e-01 -7.76977003e-01
-1.28993616e-01 4.93279099e-01 -1.20472871e-01 1.12102449e+00
-4.64737207e-01 8.86972249e-01 6.86012840e+00 3.10474664e-01
-5.56097567e-01 1.53927147e-01 6.04965985e-01 2.86006574e-02
-7.07908511e-01 2.11231768e-01 -4.75236297e-01 2.60026604e-01
1.08376157e+00 -5.60184494e-02 5.08000731e-01 2.54848868e-01
8.08835551e-02 -2.56122857e-01 -1.43607771e+00 4.72481161e-01
2.93018043e-01 -1.41185451e+00 5.05057752e-01 -5.57108879e-01
7.56473780e-01 -2.86043495e-01 -1.69954866e-01 6.15131199e-01
4.62723702e-01 -1.54785514e+00 7.86760688e-01 8.46418798e-01
2.66691506e-01 -5.84788263e-01 7.75988460e-01 5.95855236e-01
-5.80858946e-01 -5.07600456e-02 -5.31985402e-01 -7.21758306e-01
4.66110706e-01 -4.71591726e-02 -8.81240368e-01 6.07433200e-01
2.31598973e-01 6.16036355e-02 -1.07264435e+00 8.48969281e-01
-6.96162164e-01 7.83962965e-01 8.22183564e-02 -3.50451380e-01
4.84287292e-01 -6.31639883e-02 3.53483528e-01 1.04188299e+00
3.30540091e-01 2.79947400e-01 -2.78055578e-01 1.11427927e+00
-2.76832581e-01 5.38191311e-02 -4.89666700e-01 1.19679898e-01
3.68642926e-01 6.55874193e-01 -1.69886857e-01 -6.37998044e-01
-4.05698329e-01 1.32564294e+00 6.99495912e-01 5.03654361e-01
-7.10570812e-01 -4.92834926e-01 5.03017865e-02 1.11253425e-01
2.51478553e-01 -3.05908006e-02 -4.57885832e-01 -1.19993305e+00
2.89392442e-01 -1.35070419e+00 2.65688002e-01 -1.12328160e+00
-1.00934947e+00 9.04697001e-01 -2.60984004e-01 -6.28326118e-01
-5.01798272e-01 -1.40051857e-01 -9.15681779e-01 1.30320549e+00
-1.41994619e+00 -8.68171096e-01 -5.36329187e-02 2.87422568e-01
9.35835779e-01 1.22017056e-01 8.12522590e-01 -8.39370042e-02
-2.23582521e-01 4.83157486e-01 -2.79119104e-01 4.12691720e-02
2.80434310e-01 -1.48590744e+00 8.59908402e-01 1.02093387e+00
3.82830888e-01 9.28822339e-01 9.74914849e-01 -4.11043406e-01
-1.36743712e+00 -9.52395976e-01 1.52759862e+00 -1.06540573e+00
4.49298263e-01 -3.08513582e-01 -1.11486328e+00 8.21343362e-01
1.00854158e+00 -9.90533471e-01 4.95361328e-01 -1.87093347e-01
-1.77667007e-01 2.27936566e-01 -8.53440285e-01 6.88203692e-01
7.85822392e-01 -7.58221149e-01 -1.41821086e+00 5.04319131e-01
1.14280391e+00 -4.43114191e-01 -4.14096028e-01 -9.09012556e-02
-1.18391141e-01 -1.01823413e+00 5.57359874e-01 -9.95326519e-01
1.17695832e+00 -4.77628648e-01 -1.36511937e-01 -1.27182317e+00
-3.46386731e-01 -6.65840626e-01 -2.51387477e-01 1.23594654e+00
7.90994942e-01 -2.65096247e-01 5.02262771e-01 5.50487041e-01
-2.77446181e-01 -8.65782678e-01 -7.55364656e-01 -3.93296540e-01
5.10295153e-01 -1.64857239e-01 9.45663810e-01 3.42943877e-01
2.96402648e-02 1.24105251e+00 -3.00441295e-01 -8.53745081e-03
1.08126797e-01 1.97568223e-01 7.23172307e-01 -5.76101005e-01
-7.72486091e-01 -2.56652117e-01 2.23600626e-01 -1.72632325e+00
-2.90635794e-01 -1.05244017e+00 1.65603712e-01 -2.17784524e+00
1.62630036e-01 3.02152157e-01 3.81793022e-01 -5.15151769e-03
-6.94663823e-01 -7.96263739e-02 2.02407107e-01 2.80867964e-02
-8.13678086e-01 4.65158761e-01 1.58360922e+00 -6.31013536e-04
1.07767090e-01 -8.04388002e-02 -1.25964642e+00 2.60142803e-01
6.48433566e-01 -2.98667759e-01 -7.59897113e-01 -1.19501817e+00
8.25046778e-01 5.77876210e-01 2.27884606e-01 -9.10202682e-01
2.83739954e-01 1.58554330e-01 1.20539062e-01 -3.63538086e-01
1.53681980e-02 -2.90031940e-01 -6.40604720e-02 3.84516001e-01
-9.61579680e-01 6.49986565e-01 -6.14062920e-02 3.47739160e-01
-3.99017602e-01 -6.14228845e-01 4.07083333e-01 -4.40474331e-01
-2.23974869e-01 -1.77299738e-01 -4.09161061e-01 6.18427217e-01
4.96787697e-01 -1.15557373e-01 -5.34073234e-01 -9.80132699e-01
-5.25957167e-01 5.65815568e-01 1.15565971e-01 4.14791822e-01
6.02631152e-01 -1.18202627e+00 -1.10467470e+00 -7.11507052e-02
-3.81534249e-02 1.39996886e-01 2.29158863e-01 3.26579392e-01
-9.13817346e-01 7.24695325e-01 2.22393617e-01 -2.02963632e-02
-8.79007697e-01 4.35060501e-01 5.14060736e-01 -3.81412268e-01
-5.74799478e-01 1.19272006e+00 4.52538952e-04 -4.26559925e-01
2.35213384e-01 -7.79947221e-01 -6.09126270e-01 -3.27469260e-01
6.66637480e-01 2.42527306e-01 6.04615025e-02 -3.11129451e-01
1.35134920e-01 2.52370358e-01 -2.15528339e-01 -1.20804474e-01
1.03990495e+00 -5.18987060e-01 -3.08040977e-01 2.26224333e-01
1.23136401e+00 -8.33691657e-02 -1.01834214e+00 -6.75806627e-02
1.20474420e-01 -3.85873299e-03 -5.91874540e-01 -1.22364390e+00
-2.23192781e-01 9.76382971e-01 1.75266489e-02 4.96066511e-01
8.96842837e-01 2.31038585e-01 1.26928556e+00 6.58181667e-01
-1.51235178e-01 -7.22020745e-01 2.98457503e-01 1.04724324e+00
1.20313585e+00 -1.12807155e+00 -5.27818739e-01 -2.36196920e-01
-6.92792296e-01 9.86471713e-01 6.38856828e-01 -3.51135015e-01
-1.25860959e-01 -1.66705951e-01 2.45400786e-01 -3.58441919e-01
-1.23097813e+00 -4.14930545e-02 3.19396257e-01 5.23992658e-01
6.75047755e-01 -7.55477250e-02 -3.92525971e-01 6.72647476e-01
-1.02337325e+00 -1.21776499e-02 7.21350431e-01 6.87235534e-01
-6.59379184e-01 -1.05104482e+00 -9.86033976e-02 6.14806056e-01
-6.44809484e-01 -3.29679519e-01 -4.03144509e-01 4.24024343e-01
-7.42973089e-02 1.33929992e+00 -6.65252581e-02 -2.51977801e-01
5.23410380e-01 5.26192844e-01 4.83194232e-01 -9.92341757e-01
-9.71332788e-01 -6.29775465e-01 5.30796051e-01 -2.11177692e-01
6.51886035e-03 -2.41687700e-01 -1.18608415e+00 -2.39644945e-01
-1.74514025e-01 5.59181273e-01 1.10097706e-01 1.34161150e+00
2.98523098e-01 7.95722425e-01 2.84490556e-01 3.23150247e-01
-1.03374016e+00 -1.39628232e+00 5.77173941e-02 6.47345185e-01
5.55395603e-01 2.99024940e-01 -8.77992734e-02 3.25803459e-01] | [11.422576904296875, 8.160077095031738] |
953f3620-2817-4a75-8dd2-72f7806d8101 | feanet-feature-enhanced-attention-network-for | 2110.08988 | null | https://arxiv.org/abs/2110.08988v1 | https://arxiv.org/pdf/2110.08988v1.pdf | FEANet: Feature-Enhanced Attention Network for RGB-Thermal Real-time Semantic Segmentation | The RGB-Thermal (RGB-T) information for semantic segmentation has been extensively explored in recent years. However, most existing RGB-T semantic segmentation usually compromises spatial resolution to achieve real-time inference speed, which leads to poor performance. To better extract detail spatial information, we propose a two-stage Feature-Enhanced Attention Network (FEANet) for the RGB-T semantic segmentation task. Specifically, we introduce a Feature-Enhanced Attention Module (FEAM) to excavate and enhance multi-level features from both the channel and spatial views. Benefited from the proposed FEAM module, our FEANet can preserve the spatial information and shift more attention to high-resolution features from the fused RGB-T images. Extensive experiments on the urban scene dataset demonstrate that our FEANet outperforms other state-of-the-art (SOTA) RGB-T methods in terms of objective metrics and subjective visual comparison (+2.6% in global mAcc and +0.8% in global mIoU). For the 480 x 640 RGB-T test images, our FEANet can run with a real-time speed on an NVIDIA GeForce RTX 2080 Ti card. | ['Tin Lun Lam', 'Xiyue Guo', 'Junjie Hu', 'Junfeng Chen', 'Yuan Gao', 'Yong Yang', 'Hongmin Wang', 'Mingjian Liang', 'Hua Feng', 'Fuqin Deng'] | 2021-10-18 | null | null | null | null | ['thermal-image-segmentation'] | ['computer-vision'] | [ 3.27485442e-01 -3.09398770e-01 2.39754274e-01 -3.64911288e-01
-7.31412351e-01 -1.16248190e-01 1.13913551e-01 -4.24734265e-01
-6.83154821e-01 4.16993260e-01 -2.30293944e-01 -2.99026310e-01
-1.11422613e-01 -1.04378974e+00 -5.54478943e-01 -7.70914078e-01
2.98404515e-01 -7.86381960e-02 6.00505650e-01 -2.62789756e-01
1.98104501e-01 3.95584166e-01 -1.77253425e+00 1.85312837e-01
1.16253126e+00 1.42645383e+00 5.56572914e-01 6.67414486e-01
-8.82455036e-02 3.60870063e-01 -4.40309316e-01 1.89971533e-02
3.50110590e-01 -1.16702959e-01 -8.59909296e-01 1.00903779e-01
2.16677219e-01 -4.15815830e-01 -3.50798428e-01 1.00207698e+00
4.58399951e-01 3.84079367e-01 1.92779109e-01 -1.33058202e+00
-5.29986918e-01 1.14296235e-01 -9.90362346e-01 3.35643381e-01
4.82470095e-02 3.54073495e-01 5.94967246e-01 -9.71219957e-01
1.35523409e-01 1.34339786e+00 5.00157237e-01 -2.76289023e-02
-8.20701897e-01 -7.36234128e-01 2.02113777e-01 4.47252244e-01
-1.51336718e+00 4.83702756e-02 8.37000787e-01 3.68309319e-02
9.58928227e-01 3.89870137e-01 9.40019369e-01 4.49826926e-01
2.11919755e-01 1.02503526e+00 1.38563478e+00 -7.73277506e-02
1.21175036e-01 -4.35848862e-01 -2.40590144e-02 7.97602296e-01
-3.77244316e-02 1.21776380e-01 -5.32689035e-01 5.14020979e-01
1.14488101e+00 1.29638061e-01 -2.40423381e-01 9.81972590e-02
-1.10049200e+00 6.03163362e-01 1.15146434e+00 3.63548756e-01
-3.85765374e-01 3.93416137e-01 1.92987919e-01 -1.75605491e-01
5.37352622e-01 1.69694245e-01 -5.43728948e-01 -6.97208941e-02
-7.90798783e-01 -1.84617937e-01 -6.35376573e-02 6.58058882e-01
9.84521151e-01 1.81811988e-01 -1.98560078e-02 5.90422511e-01
4.28424716e-01 8.17756057e-01 4.98052567e-01 -1.22086871e+00
4.08459097e-01 7.12152660e-01 -9.19565111e-02 -1.00878847e+00
-5.28362334e-01 -3.26694727e-01 -8.62407088e-01 3.51841986e-01
2.56335020e-01 1.35007486e-01 -1.29447770e+00 1.22338617e+00
3.97040635e-01 1.12852179e-01 -6.28842562e-02 1.35838008e+00
1.05653560e+00 8.60702813e-01 1.38694376e-01 1.55953199e-01
1.24960506e+00 -1.14648509e+00 -5.01536667e-01 -5.21165013e-01
3.48219842e-01 -5.17297864e-01 1.34225082e+00 3.36445302e-01
-7.14189351e-01 -8.05137932e-01 -1.04568207e+00 -4.99822378e-01
-5.44257998e-01 1.84655592e-01 9.06551778e-01 7.65020490e-01
-9.93812144e-01 3.90796483e-01 -1.14690173e+00 -2.24985942e-01
6.56191587e-01 4.15308356e-01 -2.22952902e-01 -3.02072793e-01
-1.15616453e+00 5.05760550e-01 5.11761308e-01 5.53312361e-01
-5.39226949e-01 -5.36662221e-01 -8.02419007e-01 -2.98242131e-03
5.93361318e-01 -4.97584492e-01 9.26476955e-01 -9.07926738e-01
-1.59582567e+00 5.87667644e-01 -1.28719166e-01 6.19450100e-02
2.58136690e-01 -3.53282690e-01 -1.46642879e-01 4.21966106e-01
3.28110218e-01 9.56345260e-01 5.95084965e-01 -1.31109536e+00
-8.38657558e-01 -7.05985308e-01 -1.04907617e-01 4.72374290e-01
-1.72375485e-01 -3.25854540e-01 -8.00996423e-01 -5.46855569e-01
4.76871878e-01 -5.89781702e-01 -3.05520982e-01 1.63270622e-01
-5.40192664e-01 -2.21195500e-02 1.22193396e+00 -7.28848040e-01
8.79088402e-01 -2.02920842e+00 -4.80369888e-02 1.47009432e-01
3.59593928e-02 5.22199869e-01 5.56053966e-02 -3.56593579e-01
6.62800372e-02 1.38884336e-01 -4.70715672e-01 -3.30678463e-01
-2.04366535e-01 4.29739714e-01 1.26091957e-01 4.67194468e-01
-1.69725844e-03 1.22513878e+00 -8.32846165e-01 -6.41547620e-01
8.34156513e-01 6.32260323e-01 -3.41489613e-01 1.67329691e-03
1.08358577e-01 6.10438049e-01 -7.98408449e-01 9.38366592e-01
8.95292759e-01 -2.92604446e-01 -4.72184837e-01 -2.84019977e-01
-3.86941165e-01 -2.43946850e-01 -9.84757781e-01 2.02911997e+00
-4.78244066e-01 7.06784129e-01 1.00578725e-01 -7.90644050e-01
7.32244849e-01 -2.03261957e-01 4.98437524e-01 -1.37101173e+00
4.28068310e-01 2.32906759e-01 -3.60352963e-01 -3.77028972e-01
7.87213683e-01 1.46964878e-01 -1.38839990e-01 3.55861276e-01
-2.51938373e-01 -3.35013598e-01 -1.73892751e-01 -1.54736927e-02
6.15006208e-01 2.58278757e-01 7.77090415e-02 -1.01035036e-01
4.66774434e-01 1.93433482e-02 6.47219956e-01 5.28782248e-01
-3.41813505e-01 7.63924420e-01 -1.12383869e-02 -5.38722932e-01
-8.76715422e-01 -1.06695485e+00 -9.61558670e-02 1.04464865e+00
7.39349484e-01 -3.26685719e-02 -7.53548443e-01 -5.51111460e-01
-1.60631642e-01 5.18380582e-01 -6.51829541e-01 -2.20993217e-02
-6.13144875e-01 -8.31550717e-01 2.82357395e-01 8.69517207e-01
1.50565481e+00 -1.14599705e+00 -1.34625220e+00 1.48877814e-01
-2.86325753e-01 -1.08192074e+00 -4.23133761e-01 1.40167221e-01
-9.10278082e-01 -8.47230315e-01 -6.59911454e-01 -4.05882061e-01
4.31977063e-01 6.62702024e-01 7.00612903e-01 1.12713441e-01
-4.74539816e-01 9.44982916e-02 -5.26819766e-01 -4.75580916e-02
4.76574868e-01 1.03442021e-01 -5.31133831e-01 -1.16486296e-01
1.66231662e-01 -3.26766133e-01 -9.05873835e-01 4.07106727e-01
-1.00475156e+00 5.57552338e-01 6.78187668e-01 7.24882305e-01
7.97241330e-01 3.13045681e-01 1.39578879e-01 -3.01808089e-01
3.70956026e-02 -8.55296254e-02 -6.30392134e-01 1.56505972e-01
-5.36836565e-01 -1.62819460e-01 4.04806852e-01 6.55569136e-02
-1.18224919e+00 8.55469108e-02 -3.78847361e-01 -4.26064938e-01
-1.31516054e-01 3.41918379e-01 -4.79804665e-01 -2.03424439e-01
7.95411766e-02 3.82279187e-01 -1.60878077e-01 -3.55448961e-01
3.42450976e-01 7.73158371e-01 8.10721219e-01 -5.10317087e-01
6.17278397e-01 6.88877285e-01 -9.91024077e-02 -7.90967703e-01
-9.16602552e-01 -3.37108225e-01 -6.07006252e-01 -3.59932661e-01
1.22080517e+00 -9.15609896e-01 -7.61761904e-01 9.80500519e-01
-8.63521695e-01 -5.72902560e-01 -9.35737640e-02 2.24127010e-01
-5.26981413e-01 2.94887573e-01 -4.88357902e-01 -6.77058160e-01
-5.67622423e-01 -1.51943684e+00 1.56006086e+00 7.87219107e-01
3.78887475e-01 -6.26285076e-01 -4.71024752e-01 6.55149877e-01
5.13932705e-01 3.19856793e-01 6.39162362e-01 3.08152348e-01
-7.79317498e-01 1.51402220e-01 -8.99572492e-01 2.19661117e-01
-1.63729079e-02 -2.58259565e-01 -1.16296041e+00 3.48183303e-03
-1.09068222e-01 -1.99718848e-01 1.00942779e+00 6.94095850e-01
1.75308621e+00 1.62088096e-01 -3.40478837e-01 1.10880613e+00
1.60742879e+00 2.59403169e-01 7.76491165e-01 6.06368780e-01
1.17114007e+00 2.48734787e-01 9.48882401e-01 2.89407760e-01
6.29141390e-01 6.64040744e-01 7.04551399e-01 -7.06253469e-01
-1.37364894e-01 -4.24385443e-03 -7.00116009e-02 4.57029790e-01
-1.06803738e-01 -1.84054106e-01 -9.07550991e-01 5.70182502e-01
-1.82133877e+00 -4.63920712e-01 -2.40431353e-01 1.85448754e+00
6.05030239e-01 -5.45423329e-02 -7.32497051e-02 3.90997022e-01
7.03959703e-01 1.42608985e-01 -8.43546689e-01 -1.58539832e-01
-1.68706387e-01 3.83637816e-01 7.25759268e-01 3.44511688e-01
-1.17435837e+00 1.10262084e+00 4.98706055e+00 1.07558084e+00
-1.37276316e+00 2.30224326e-01 1.02512646e+00 3.22380774e-02
-1.15943439e-01 -2.23624006e-01 -2.43399978e-01 5.36871731e-01
5.11659026e-01 2.21659899e-01 2.87088484e-01 8.48454177e-01
2.35229582e-01 -6.07782066e-01 -3.99708211e-01 1.20208108e+00
-1.46960825e-01 -9.51647162e-01 -2.61199594e-01 1.44896105e-01
6.20489597e-01 1.38507023e-01 3.11008334e-01 2.86106735e-01
1.36703387e-01 -1.21190417e+00 8.50269139e-01 3.52872699e-01
9.93953228e-01 -1.07816756e+00 8.55731964e-01 6.30466491e-02
-1.57764983e+00 -1.09092973e-01 -2.17282370e-01 1.43238381e-01
3.09940010e-01 5.03878832e-01 -1.17781527e-01 8.37771893e-01
1.21708548e+00 8.39591026e-01 -7.31642008e-01 8.11528921e-01
-2.86505252e-01 2.73217261e-01 -5.96537650e-01 1.23970516e-01
6.85238242e-01 -2.26201057e-01 1.47686735e-01 8.37746680e-01
4.48321998e-01 4.86082077e-01 2.12395713e-01 8.61292362e-01
2.96242505e-01 -2.13724911e-01 1.39256176e-02 2.97489643e-01
2.99925327e-01 1.28352332e+00 -1.24902272e+00 -2.97720432e-01
-2.17122421e-01 1.26210058e+00 8.70519876e-02 5.03274083e-01
-9.68713045e-01 -6.01742864e-01 5.75184166e-01 -7.63624907e-02
6.17771745e-01 -1.64140716e-01 -6.42542183e-01 -9.54077959e-01
-1.79455578e-01 -4.77094144e-01 1.52615219e-01 -1.35580599e+00
-8.06219280e-01 6.78548694e-01 -1.27656907e-01 -8.98920834e-01
3.96303087e-01 -7.25449204e-01 -4.09257799e-01 8.93994808e-01
-1.68715942e+00 -1.34576941e+00 -8.90615106e-01 7.93011725e-01
4.52298760e-01 4.01739687e-01 4.21695054e-01 2.40769640e-01
-8.27484131e-01 3.78100425e-01 -7.90766180e-02 1.57740518e-01
1.69339329e-01 -1.21270180e+00 4.21524554e-01 1.01733029e+00
-2.78753549e-01 2.34227672e-01 2.72841573e-01 -5.02502859e-01
-1.33380961e+00 -1.38752759e+00 2.58179456e-01 -5.79516515e-02
8.12564716e-02 4.93831094e-03 -7.91752756e-01 4.10653740e-01
2.15269141e-02 4.36045527e-01 1.91992193e-01 -4.45512831e-01
-1.26027763e-01 -2.80930370e-01 -1.22148597e+00 4.05621350e-01
9.65398490e-01 -3.84644866e-01 -2.36977443e-01 -9.32601243e-02
8.24076891e-01 -6.20149910e-01 -8.35355818e-01 4.45800781e-01
5.02581060e-01 -1.24217772e+00 1.13887811e+00 2.23302200e-01
3.54882747e-01 -6.18857563e-01 -3.05323422e-01 -1.14063334e+00
-9.18734670e-02 -2.24651873e-01 4.22160089e-01 7.81302392e-01
-1.16000064e-01 -6.13931000e-01 8.27351451e-01 4.44539309e-01
-4.75698799e-01 -7.75085807e-01 -1.33971667e+00 -4.53667283e-01
-7.77160674e-02 -8.34486842e-01 7.30171561e-01 5.36795437e-01
-5.00832081e-01 1.14787154e-01 -1.15875043e-01 2.02123135e-01
6.40060186e-01 4.43221241e-01 5.42240143e-01 -9.61065233e-01
1.37886390e-01 -5.44052720e-01 -4.41591501e-01 -1.25949562e+00
-9.89459977e-02 -5.11174798e-01 2.08441600e-01 -1.71783245e+00
1.47941783e-01 -5.25201738e-01 -2.40076348e-01 6.41169667e-01
-4.41961318e-01 6.47512078e-01 1.65354714e-01 -1.67995691e-01
-7.31926918e-01 9.04911041e-01 1.57640481e+00 -8.30787271e-02
-1.04174457e-01 -3.80932748e-01 -6.27699733e-01 6.46067739e-01
6.88977897e-01 -2.49292538e-01 -2.84477204e-01 -6.88503981e-01
-1.17620103e-01 -3.57834734e-02 6.07009351e-01 -1.23113525e+00
1.40044466e-01 -1.70530468e-01 7.44531691e-01 -1.11423850e+00
5.61912179e-01 -7.33031869e-01 -7.50571489e-02 5.92685878e-01
2.57262468e-01 -4.99576144e-02 2.54030377e-01 4.65065807e-01
-2.18682081e-01 2.74989784e-01 8.11864376e-01 -2.36168399e-01
-1.24867606e+00 4.38556522e-01 -2.64587998e-01 -1.66577920e-01
1.07781065e+00 -6.78055584e-01 -3.81326914e-01 -1.50929689e-01
-3.28009307e-01 4.06102866e-01 5.62441945e-01 2.22072050e-01
9.30257320e-01 -1.27763093e+00 -2.98636794e-01 4.46921170e-01
4.18165028e-02 5.49147606e-01 6.83620512e-01 9.30090129e-01
-8.20515513e-01 4.11030412e-01 -4.57784742e-01 -8.59634578e-01
-1.00264919e+00 2.33198434e-01 3.38995904e-01 5.84289432e-03
-7.84768581e-01 9.53882098e-01 2.31063783e-01 -2.92025864e-01
-4.46817353e-02 -6.17261171e-01 -5.49419150e-02 -2.07166880e-01
3.77946764e-01 4.99846250e-01 5.76818213e-02 -8.79905820e-01
-4.84398663e-01 1.00118518e+00 1.56534150e-01 -2.45326027e-01
1.35571086e+00 -5.16378939e-01 4.47278237e-03 1.37343317e-01
1.22912884e+00 -5.95880330e-01 -1.59522796e+00 -1.35251239e-01
-4.91819978e-01 -8.98028851e-01 7.42258668e-01 -6.78101122e-01
-1.70136273e+00 1.08289969e+00 9.76356268e-01 -7.87599385e-02
1.59378111e+00 -1.94995940e-01 1.20267355e+00 2.81414799e-02
5.31370103e-01 -1.09385943e+00 2.53989875e-01 5.00572860e-01
5.65987229e-01 -1.41040730e+00 1.19738452e-01 -5.61601877e-01
-6.37327731e-01 9.35318291e-01 8.38107467e-01 9.66132339e-03
4.81184125e-01 1.15716524e-01 1.07311666e-01 -2.90070385e-01
-9.87092182e-02 -5.23847818e-01 2.01991871e-01 6.39200211e-01
1.21157557e-01 1.41966209e-01 1.79895490e-01 4.17877555e-01
-2.09044963e-01 -2.26917952e-01 2.42774799e-01 9.39086735e-01
-5.52000046e-01 -5.66891253e-01 -5.58395147e-01 3.02774042e-01
-1.67212993e-01 1.90190896e-02 2.12769303e-02 9.06814396e-01
2.94731021e-01 1.16593230e+00 3.26334298e-01 -5.71764290e-01
1.73587635e-01 -3.87152255e-01 4.03969854e-01 -3.09461147e-01
-5.51249504e-01 3.26224625e-01 -3.57673556e-01 -1.06955910e+00
-5.83691537e-01 -4.75155652e-01 -1.49420583e+00 -3.99685830e-01
-3.26924324e-01 -2.09791392e-01 8.02160025e-01 1.06467152e+00
1.88446328e-01 8.55989873e-01 4.84074026e-01 -1.20490956e+00
2.83477306e-01 -9.01021540e-01 -6.68689311e-01 9.91045386e-02
1.82608843e-01 -8.18965554e-01 -1.18388265e-01 -2.81273097e-01] | [9.461923599243164, -1.0434285402297974] |
f4651463-e13b-443c-a06b-fa7381a66c42 | mpox-aism-ai-mediated-super-monitoring-for | 2303.09780 | null | https://arxiv.org/abs/2303.09780v2 | https://arxiv.org/pdf/2303.09780v2.pdf | Mpox-AISM: AI-Mediated Super Monitoring for Forestalling Monkeypox Spread | The challenge on forestalling monkeypox (Mpox) spread is the timely, convenient and accurate diagnosis for earlystage infected individuals. Here, we propose a remote and realtime online visualization strategy, called "Super Monitoring" to construct a low cost, convenient, timely and unspecialized diagnosis of early-stage Mpox. Such AI-mediated "Super Monitoring" (Mpox-AISM) invokes a framework assembled by deep learning, data augmentation and self-supervised learning, as well as professionally classifies four subtypes according to dataset characteristics and evolution trend of Mpox and seven other types of dermatopathya with high similarity, hence these features together with reasonable program interface and threshold setting ensure that its Recall (Sensitivity) was beyond 95.9% and the specificity was almost 100%. As a result, with the help of cloud service on Internet and communication terminal, this strategy can be potentially utilized for the real-time detection of earlystage Mpox in various scenarios including entry-exit inspection in airport, family doctor, rural area in underdeveloped region and wild to effectively shorten the window period of Mpox spread. | ['Yang Li', 'Jinbao Liu', 'Jialong Xu', 'Xinyue Zhang', 'Zhenzhang Li', 'Yubiao Yue'] | 2023-03-17 | null | null | null | null | ['specificity'] | ['natural-language-processing'] | [-8.83841664e-02 -5.20538509e-01 -2.11614698e-01 -1.32625341e-01
-2.12005809e-01 -4.34976757e-01 4.55740631e-01 2.40088925e-01
-2.40790263e-01 7.36168742e-01 -4.19805557e-01 -6.75698459e-01
-6.11285031e-01 -8.53177667e-01 4.60428894e-02 -6.59579158e-01
-8.31838548e-01 6.05491817e-01 1.00384735e-01 -5.92951998e-02
2.50772208e-01 7.14177787e-01 -1.40407574e+00 3.30031216e-01
1.23376226e+00 6.37059033e-01 1.69781163e-01 7.41121531e-01
-6.75946102e-02 4.56101269e-01 -6.49181902e-01 -2.09543575e-02
2.83922672e-01 4.17030565e-02 -5.23333430e-01 -3.38665843e-01
-7.91695863e-02 -7.09041238e-01 -1.26470163e-01 6.22116685e-01
2.46185407e-01 -2.16309756e-01 7.02884555e-01 -1.57519341e+00
-3.01605880e-01 -1.02428176e-01 -1.09959626e+00 6.16961896e-01
2.61539131e-01 8.13464224e-01 3.77317965e-01 -4.35690641e-01
6.02836788e-01 8.94723117e-01 1.07505488e+00 2.10931540e-01
-9.75843608e-01 -6.10114872e-01 -1.19850866e-01 4.73410845e-01
-1.36267519e+00 -1.50563130e-02 4.26468819e-01 -5.10807693e-01
8.61738443e-01 7.40134060e-01 7.84277499e-01 8.19048285e-01
2.95412481e-01 6.59153536e-02 1.37608540e+00 -1.60287336e-01
-1.20634623e-01 3.15014750e-01 2.25294098e-01 8.14653695e-01
3.00377280e-01 1.57011613e-01 -3.67931545e-01 -5.66438913e-01
6.46013319e-01 5.49870372e-01 -1.07059859e-01 3.52911323e-01
-1.18841445e+00 7.49106526e-01 5.39133549e-01 2.65053779e-01
-8.23541820e-01 -2.82692850e-01 4.50324833e-01 1.31336823e-01
3.95701945e-01 4.56834026e-03 -2.83144355e-01 -2.61075646e-01
-5.75540543e-01 1.87070578e-01 3.52576405e-01 3.75947565e-01
5.45308411e-01 -2.07028762e-01 -1.13554977e-01 5.58971643e-01
3.90162885e-01 7.28084266e-01 1.10983327e-01 -5.53642392e-01
5.40978193e-01 8.85132432e-01 1.94391847e-01 -1.17588663e+00
-7.47540951e-01 -3.98998231e-01 -9.33528364e-01 4.41221625e-01
2.67599732e-01 -3.23658675e-01 -9.70743358e-01 1.39995301e+00
8.19479108e-01 2.92437941e-01 -3.55798811e-01 9.04044747e-01
5.44936717e-01 8.71082425e-01 5.56382835e-01 -3.06405783e-01
1.89115894e+00 -3.97985846e-01 -7.13088453e-01 1.07545391e-01
3.29610288e-01 -4.85016435e-01 1.10255873e+00 3.07797462e-01
-3.57890785e-01 -2.10602775e-01 -9.92523670e-01 3.71066809e-01
-5.65225661e-01 2.98924446e-01 1.15320730e+00 5.90408206e-01
-8.28381479e-01 5.23704112e-01 -8.82199287e-01 -9.61272836e-01
7.24856019e-01 4.58230466e-01 -3.04682523e-01 2.24754930e-01
-1.02932239e+00 8.76408517e-01 1.70108736e-01 4.37049329e-01
-7.33891904e-01 -9.93351877e-01 -4.80419286e-02 -2.14872390e-01
1.24686740e-01 -7.65604019e-01 5.39965570e-01 -2.89817095e-01
-9.12571788e-01 7.42806673e-01 1.47524789e-01 -3.59353632e-01
5.32500446e-01 1.68311372e-02 -5.39133787e-01 3.82040948e-01
-5.34391031e-02 3.87555987e-01 4.90397483e-01 -8.11062992e-01
-1.09524047e+00 -7.64409840e-01 -8.93860385e-02 1.36101395e-01
-2.70258542e-02 5.29535949e-01 5.03240637e-02 -2.58371651e-01
-3.81966799e-01 -3.78090173e-01 -1.89560607e-01 5.13777733e-01
-2.31202140e-01 -5.50080001e-01 1.13306439e+00 -1.46540260e+00
1.17932451e+00 -1.70033705e+00 -5.45140088e-01 5.18991232e-01
6.20300174e-01 6.16985857e-01 1.57199875e-01 6.84036911e-01
8.96951258e-02 -6.55316412e-02 -1.27769187e-01 3.03423017e-01
-2.56065875e-01 -5.72753400e-02 3.22629139e-02 6.06624067e-01
5.38486421e-01 5.74866474e-01 -8.71258020e-01 -5.99060595e-01
2.28992045e-01 7.18662560e-01 -9.47836563e-02 4.75391090e-01
-1.10237733e-01 4.30091858e-01 -5.48012972e-01 1.02058697e+00
8.37105870e-01 -3.79274428e-01 2.59371579e-01 2.72255726e-02
-3.50416362e-01 -1.71066806e-01 -1.00232446e+00 1.07394326e+00
-1.84490487e-01 6.37001395e-01 1.83026224e-01 -5.99571705e-01
8.49190295e-01 2.15847909e-01 4.53147918e-01 -4.46827531e-01
2.55979806e-01 3.72764766e-02 -2.41100058e-01 -1.18018997e+00
2.48372778e-02 3.45313162e-01 6.34399831e-01 8.28740120e-01
-4.66155738e-01 1.12458062e+00 3.53023820e-02 1.30987063e-01
1.11410534e+00 4.96161804e-02 3.78070593e-01 -2.13452633e-02
3.43946666e-01 4.45717990e-01 2.47305661e-01 4.60670531e-01
-5.47049701e-01 -3.71787623e-02 4.39803958e-01 -4.71763551e-01
-8.82309735e-01 -1.12391031e+00 -2.53604233e-01 1.01407719e+00
2.69821025e-02 -1.26399184e-02 -5.75971603e-01 -8.35131466e-01
5.97255118e-02 1.84000596e-01 -6.53884530e-01 1.01978570e-01
-4.45893735e-01 -1.01434875e+00 2.52131552e-01 1.95468441e-01
7.50431120e-01 -9.69992578e-01 -6.93515420e-01 -1.51712690e-02
3.78058776e-02 -9.14990678e-02 -1.85554335e-03 -2.80105531e-01
-6.54101551e-01 -1.42233014e+00 -6.92078233e-01 -7.20713317e-01
6.26801610e-01 2.63917089e-01 5.07022142e-01 3.21664244e-01
-8.82915437e-01 -6.71324804e-02 -1.03286065e-01 -4.70282063e-02
-5.30199409e-02 -1.67445078e-01 -3.56789269e-02 1.26722515e-01
6.04839981e-01 -7.13631868e-01 -1.04467762e+00 -8.75373706e-02
-3.02288026e-01 5.19071743e-02 5.81685364e-01 3.91270906e-01
2.43142962e-01 -1.43403381e-01 7.61827171e-01 -8.89244020e-01
7.43592739e-01 -1.20054317e+00 -6.13185585e-01 6.24693096e-01
-5.29195309e-01 -5.69577694e-01 4.58374500e-01 -5.61271846e-01
-1.03531301e+00 -2.88083404e-01 -3.67598087e-02 -1.33513182e-01
-2.29837134e-01 3.56573075e-01 1.74464032e-01 -2.78947372e-02
6.91020250e-01 1.52996853e-01 3.79311144e-01 -7.79721320e-01
9.48715359e-02 1.23201191e+00 6.09878004e-01 1.30052239e-01
6.81897104e-01 4.16154802e-01 -1.10441543e-01 -6.87632620e-01
1.93545163e-01 -3.31799060e-01 -1.27294719e-01 -5.34251928e-01
8.64289761e-01 -5.02429843e-01 -1.36698794e+00 5.82184255e-01
-1.21224344e+00 5.38770668e-02 5.29144526e-01 4.27993119e-01
1.28157079e-01 2.76767373e-01 -4.91106570e-01 -1.21851695e+00
-9.81759250e-01 -3.61968637e-01 5.86785734e-01 7.02516973e-01
-2.43789375e-01 -8.09648991e-01 9.24783126e-02 3.34103316e-01
7.40009189e-01 5.99193156e-01 1.08317733e+00 -6.90304935e-01
-7.38206685e-01 -2.93389887e-01 -7.48679101e-01 -3.39226484e-01
2.83348680e-01 3.72436911e-01 -6.94050372e-01 -1.17518000e-01
-1.03269756e-01 3.75175253e-02 1.81418195e-01 6.54585958e-02
7.81082690e-01 -5.91897190e-01 -7.38547862e-01 5.12283027e-01
1.45817888e+00 6.70150518e-01 2.43923515e-01 1.22758552e-01
4.38552976e-01 9.45030391e-01 8.93114030e-01 6.14003181e-01
4.35196847e-01 4.65201497e-01 2.03747720e-01 -2.01024190e-01
4.02351022e-02 -2.92212307e-01 -1.98566288e-01 3.45001161e-01
-3.16818804e-01 -1.15332589e-01 -1.27739048e+00 4.44520503e-01
-1.58021975e+00 -1.00167012e+00 -3.68894488e-01 2.15629315e+00
4.83745873e-01 -2.86214203e-01 6.03928685e-01 -3.88796069e-02
1.06481838e+00 -2.83034801e-01 -6.03974581e-01 -1.87448978e-01
4.05286133e-01 6.12571537e-02 1.52299941e-01 3.42979550e-01
-9.02586758e-01 1.95468277e-01 5.75888443e+00 8.99525464e-01
-1.41624117e+00 4.59518075e-01 7.89075375e-01 4.68797646e-02
4.29500416e-02 -1.93953902e-01 -3.89530540e-01 8.32111299e-01
7.77218103e-01 -3.94722223e-02 6.25375926e-01 7.15755939e-01
5.05953610e-01 -1.64691612e-01 -5.16935945e-01 7.62765646e-01
-4.12001282e-01 -1.48785353e+00 -3.74464005e-01 1.11090235e-01
-3.46315745e-03 -1.71505511e-02 -1.91975996e-01 1.89876691e-01
-1.92516372e-01 -8.00874531e-01 -1.52555883e-01 8.26760292e-01
1.08285522e+00 -7.71424711e-01 6.35142207e-01 3.82373571e-01
-1.21924639e+00 -1.31692529e-01 8.00584778e-02 2.13478118e-01
1.27859056e-01 9.84294340e-03 -1.03349280e+00 3.30370814e-01
8.80296052e-01 2.55660117e-01 -5.85149586e-01 1.16438794e+00
3.07415634e-01 4.56205159e-01 -6.90881193e-01 -4.33552712e-01
2.89049465e-02 -2.59183258e-01 5.21448731e-01 9.93643761e-01
8.67343396e-02 2.25428686e-01 -1.34378538e-01 7.23233581e-01
5.67889392e-01 2.51449406e-01 -4.22352970e-01 -1.13199063e-01
8.06719720e-01 1.27423942e+00 -7.10307002e-01 -3.74075085e-01
-1.00897379e-01 5.12166739e-01 2.66548783e-01 3.35313946e-01
-1.12100589e+00 -9.99417663e-01 3.86067033e-01 4.81456935e-01
-9.29591954e-02 2.23893881e-01 -3.34746718e-01 -5.88024259e-01
3.97470705e-02 -6.85280383e-01 7.55367637e-01 -7.11308599e-01
-1.16329491e+00 6.45543635e-01 -6.27668872e-02 -1.31750131e+00
-9.37614739e-02 -5.57273865e-01 -1.23327744e+00 9.09539282e-01
-1.11588514e+00 -1.54327536e+00 -4.32757705e-01 5.50366223e-01
1.48366004e-01 -3.47638845e-01 8.16291988e-01 4.25302416e-01
-8.70283961e-01 4.31152940e-01 3.53795551e-02 -2.53815234e-01
2.22383350e-01 -9.05928016e-01 -2.76857048e-01 7.08259106e-01
-7.05274343e-01 9.65569139e-01 4.73141938e-01 -1.07468438e+00
-1.42010868e+00 -9.98242259e-01 1.07793331e+00 -1.11861825e-01
6.37109756e-01 -2.25228518e-01 -8.09823334e-01 3.52072746e-01
2.24042207e-01 -2.94699967e-01 6.69463038e-01 2.28847697e-01
-2.88203079e-02 -6.35261893e-01 -1.67105913e+00 6.18641913e-01
7.02685654e-01 -3.06912899e-01 -3.97781491e-01 9.19789374e-01
3.59750658e-01 6.11438528e-02 -7.60187566e-01 1.83416650e-01
5.88796318e-01 -7.99345970e-01 7.97731578e-01 -8.41589272e-01
1.41421869e-01 -3.03353697e-01 4.25001264e-01 -4.29220468e-01
-1.67550832e-01 -7.85029233e-01 -1.68168902e-01 1.33962691e+00
1.53062910e-01 -9.00606751e-01 8.38177383e-01 5.28025746e-01
2.34524384e-01 -1.14289486e+00 -1.01577652e+00 -1.94119468e-01
-8.03214312e-01 -1.18314758e-01 7.30302751e-01 8.44647229e-01
-1.18867263e-01 -1.68140344e-02 -5.75201333e-01 5.49631894e-01
6.96628094e-01 1.92355052e-01 6.23172820e-01 -9.69284177e-01
-1.79429799e-01 -1.28575906e-01 -4.27757651e-01 -2.89832085e-01
-7.96934187e-01 -5.80807745e-01 -3.81398320e-01 -1.40941668e+00
2.98090696e-01 -7.84057498e-01 -1.40497684e-01 5.58629811e-01
-2.33728752e-01 3.87014717e-01 -1.05146788e-01 3.10867965e-01
-3.26409847e-01 5.90038626e-03 8.48076940e-01 2.03554332e-01
-3.26405525e-01 3.29338551e-01 -4.56968695e-01 4.40242529e-01
8.07074249e-01 -4.55567479e-01 -3.15133125e-01 -2.35589057e-01
-1.47625998e-01 2.29331136e-01 6.13607347e-01 -6.56096816e-01
2.53775597e-01 -5.56854784e-01 4.18084472e-01 -9.16948676e-01
2.39464000e-01 -8.69691670e-01 4.69487786e-01 1.02738440e+00
3.62231910e-01 7.60068744e-02 4.48750611e-03 3.83403420e-01
4.05926347e-01 -1.09208440e-02 5.43534815e-01 1.68765083e-01
-6.44899189e-01 3.91666234e-01 -5.68558693e-01 -4.43504453e-01
1.59512889e+00 -3.22403729e-01 -9.82756436e-01 3.99244875e-02
-5.10984182e-01 5.42616069e-01 -6.01207651e-02 1.15906358e-01
4.47591156e-01 -1.11629176e+00 -7.62282550e-01 5.61539352e-01
1.10376133e-02 -5.64627051e-01 8.70944917e-01 1.06907058e+00
-1.31862628e+00 3.19148809e-01 -7.04957187e-01 -6.79539680e-01
-1.46503842e+00 4.17378455e-01 1.97271526e-01 -2.89254963e-01
-5.11215270e-01 4.89982665e-01 -5.07345438e-01 -1.84447750e-01
3.77245396e-01 1.37806609e-01 -4.93757665e-01 1.10421784e-01
8.49088907e-01 8.08547974e-01 -2.80801535e-01 -2.21016496e-01
-6.84425533e-01 3.58320057e-01 -1.71352074e-01 3.16805422e-01
1.47010016e+00 -6.80577978e-02 -2.71411836e-01 8.35699141e-02
8.30728054e-01 -1.42515033e-01 -1.07882941e+00 2.56272525e-01
-2.86591854e-02 -7.03461528e-01 -3.99511307e-01 -1.18897164e+00
-8.30007255e-01 8.13442171e-01 1.28194451e+00 4.07124609e-01
1.21640325e+00 -3.21452925e-03 7.04010069e-01 1.31808937e-01
2.52216667e-01 -5.84936142e-01 -5.54263234e-01 -1.63723096e-01
7.57363737e-01 -1.15977859e+00 2.45248713e-02 -2.00783566e-01
-6.05920017e-01 8.17894042e-01 8.12913001e-01 -8.53268728e-02
4.89984274e-01 5.14375158e-02 4.90284204e-01 -3.24685186e-01
-1.05413449e+00 1.98386475e-01 -1.35759031e-02 1.24725366e+00
-1.97492957e-01 1.64742649e-01 -5.93723416e-01 6.01226449e-01
1.67088643e-01 1.66259170e-01 -3.62841725e-01 1.01460230e+00
-5.52484512e-01 -6.70816720e-01 -4.69742924e-01 9.26582277e-01
-1.31078213e-01 2.13156655e-01 -5.09241410e-02 1.09520245e+00
6.49469674e-01 6.88184381e-01 4.46659178e-01 -3.76901656e-01
8.07203427e-02 -2.57826090e-01 1.01994939e-01 -1.04912408e-01
-8.78364086e-01 -2.10014805e-01 2.75986642e-01 -3.01656276e-01
3.56491283e-02 -4.45447505e-01 -1.01790953e+00 -9.26332533e-01
-1.31204784e-01 1.63367793e-01 7.63549447e-01 7.83373058e-01
4.86050278e-01 7.50080273e-02 8.54538202e-01 -2.91312784e-01
-4.75661039e-01 -1.13920903e+00 -4.98376608e-01 1.09558739e-01
2.05069005e-01 -3.98122162e-01 -1.95714667e-01 -2.85476476e-01] | [15.576133728027344, -1.6934728622436523] |
2f62645d-88be-4f88-b8db-72432e7bd38b | consistent-multimodal-generation-via-a | 2307.01425 | null | https://arxiv.org/abs/2307.01425v1 | https://arxiv.org/pdf/2307.01425v1.pdf | Consistent Multimodal Generation via A Unified GAN Framework | We investigate how to generate multimodal image outputs, such as RGB, depth, and surface normals, with a single generative model. The challenge is to produce outputs that are realistic, and also consistent with each other. Our solution builds on the StyleGAN3 architecture, with a shared backbone and modality-specific branches in the last layers of the synthesis network, and we propose per-modality fidelity discriminators and a cross-modality consistency discriminator. In experiments on the Stanford2D3D dataset, we demonstrate realistic and consistent generation of RGB, depth, and normal images. We also show a training recipe to easily extend our pretrained model on a new domain, even with a few pairwise data. We further evaluate the use of synthetically generated RGB and depth pairs for training or fine-tuning depth estimators. Code will be available at https://github.com/jessemelpolio/MultimodalGAN. | ['Derek Hoiem', 'Soeren Pirk', 'Zhixin Shu', 'Krishna Kumar Singh', 'Weijie Lyu', 'Yijun Li', 'Zhen Zhu'] | 2023-07-04 | null | null | null | null | ['multimodal-generation'] | ['natural-language-processing'] | [ 3.47012550e-01 2.74889916e-01 1.26532644e-01 -5.63999951e-01
-1.24687505e+00 -8.34698737e-01 8.65344584e-01 -5.48799694e-01
-1.44523545e-03 8.05006087e-01 3.75087947e-01 -8.44263658e-02
4.14233655e-01 -8.13676298e-01 -8.68022442e-01 -6.82072937e-01
3.24650615e-01 4.66725171e-01 -3.52257304e-02 -6.13410287e-02
-1.28141448e-01 3.30508471e-01 -1.45793617e+00 3.86295825e-01
4.69423115e-01 1.04798305e+00 -9.89001840e-02 1.15042722e+00
1.32630661e-01 5.65539122e-01 -4.98661011e-01 -4.41979170e-01
5.49186587e-01 -8.37144375e-01 -7.18770444e-01 2.05936775e-01
6.22348428e-01 -7.31102288e-01 -4.28631932e-01 7.48813868e-01
7.48611391e-01 1.77214090e-02 7.22060442e-01 -1.49917197e+00
-7.23270655e-01 3.41423959e-01 -4.05177861e-01 -4.98720855e-01
5.30274868e-01 7.48978376e-01 7.64173627e-01 -7.89653659e-01
8.92232239e-01 1.37737489e+00 5.40894985e-01 1.20360637e+00
-1.38352120e+00 -5.22782326e-01 -1.17484681e-01 -3.94792795e-01
-1.29430830e+00 -6.00051343e-01 5.67753315e-01 -3.68475199e-01
6.50681198e-01 1.42232880e-01 6.08500421e-01 1.80698967e+00
-2.99808681e-01 6.54941142e-01 1.15910327e+00 -3.89274240e-01
1.01818815e-01 5.43673113e-02 -6.15800560e-01 7.29398608e-01
-1.31833479e-01 2.60379165e-01 -5.94275594e-01 -3.43903922e-03
1.12216675e+00 -3.00322235e-01 -3.77352595e-01 -4.56583321e-01
-1.37772655e+00 7.01851070e-01 3.97143543e-01 -5.34959733e-02
-2.91692205e-02 7.04457760e-01 -1.80627573e-02 2.37894997e-01
2.33612671e-01 2.53614664e-01 -3.27955633e-01 -1.78563267e-01
-7.45270610e-01 4.77090269e-01 6.11305356e-01 1.30118871e+00
9.61107016e-01 -4.67701107e-02 -1.43459767e-01 5.22463620e-01
3.62030238e-01 9.18687761e-01 1.43782124e-01 -1.73148084e+00
3.37353736e-01 1.52611226e-01 2.22004373e-02 -3.83485913e-01
-1.37307197e-01 3.19583304e-02 -6.83473527e-01 4.01036918e-01
5.32960057e-01 -4.84405607e-01 -1.33489382e+00 2.06858516e+00
2.86956519e-01 6.94745108e-02 1.37819648e-01 1.01218951e+00
9.01860535e-01 5.83511174e-01 -9.47544649e-02 4.85295862e-01
1.04480207e+00 -9.60806370e-01 -2.30251044e-01 -1.31035686e-01
4.44098592e-01 -7.54354417e-01 1.23872387e+00 1.77583486e-01
-1.46934390e+00 -5.62481999e-01 -8.28396082e-01 -5.67218781e-01
-1.77642643e-01 7.19527379e-02 4.80263472e-01 3.81157577e-01
-1.38748205e+00 5.38755655e-01 -9.52030420e-01 -5.50097942e-01
2.18763247e-01 6.48589730e-02 -4.29749787e-01 -2.90019631e-01
-1.03523445e+00 7.79960573e-01 3.46369296e-01 -1.21424235e-01
-1.36786520e+00 -7.35741496e-01 -1.07372832e+00 -4.89419401e-01
-1.72330186e-01 -1.33111989e+00 1.45852649e+00 -1.06630838e+00
-1.75524962e+00 1.12101805e+00 -8.37915465e-02 1.09520018e-01
8.16334724e-01 2.35245049e-01 -1.44507796e-01 2.19096854e-01
-4.93220277e-02 1.58216465e+00 7.16933548e-01 -1.60004997e+00
-3.71963650e-01 -1.21415742e-01 1.89226344e-01 2.63294518e-01
1.46574214e-01 -4.71945226e-01 -7.44336784e-01 -4.47646439e-01
-3.78993456e-04 -1.13184536e+00 -4.47171107e-02 3.78163487e-01
-7.07837224e-01 2.29647964e-01 3.86901945e-01 -5.66601038e-01
4.30240691e-01 -2.15663004e+00 3.89202088e-01 3.12246531e-01
5.67015223e-02 -4.37197536e-01 -4.70090985e-01 3.67017716e-01
1.77476525e-01 1.72771960e-01 -5.93360603e-01 -8.83982122e-01
4.02959526e-01 2.29868934e-01 -1.74119368e-01 2.67967492e-01
4.61005241e-01 1.06530905e+00 -8.83647919e-01 -3.49296838e-01
2.15866774e-01 8.64749551e-01 -5.93350589e-01 3.62557173e-01
-5.14603734e-01 1.10134459e+00 -1.25536606e-01 8.06321740e-01
7.42499292e-01 -4.41426218e-01 2.00557366e-01 -3.11951458e-01
1.80269718e-01 3.88617337e-01 -9.89276052e-01 2.39806414e+00
-6.41486406e-01 5.92635036e-01 3.95394340e-02 -2.35249758e-01
7.27529883e-01 3.07490617e-01 1.43212542e-01 -6.70040727e-01
2.61031955e-01 3.94030064e-01 -3.74387622e-01 -2.29735360e-01
4.47488457e-01 -2.37525806e-01 -1.59808278e-01 6.27207756e-01
4.66400385e-01 -8.06820571e-01 1.20722957e-01 2.59758115e-01
8.96928906e-01 7.22026348e-01 -1.93431273e-01 1.92869797e-01
5.36650456e-02 -1.12797759e-01 8.38898569e-02 5.57340145e-01
1.89771563e-01 1.44712412e+00 6.05443656e-01 -7.08225593e-02
-1.53267193e+00 -1.43578219e+00 -4.00497094e-02 6.47235215e-01
1.62590370e-01 -2.41350934e-01 -6.99845910e-01 -6.31849110e-01
9.29871202e-02 6.82794094e-01 -7.90232718e-01 6.65825456e-02
-2.41117492e-01 -3.13928097e-01 8.43005478e-01 5.01459002e-01
4.34873313e-01 -8.46759319e-01 -3.69521111e-01 -3.91571492e-01
-1.88518181e-01 -1.16038215e+00 -3.54283154e-01 9.69606638e-02
-7.06893981e-01 -8.93879533e-01 -9.72781897e-01 -5.98745823e-01
7.77389944e-01 -5.83402887e-02 1.25157678e+00 1.11639887e-01
9.05174613e-02 6.08507335e-01 -1.45707279e-01 -7.52854049e-02
-6.50197864e-01 1.91563889e-01 -3.30198050e-01 -2.50963330e-01
-1.39516816e-01 -7.57256329e-01 -8.27364564e-01 2.30177104e-01
-1.18049634e+00 5.07860065e-01 2.94246733e-01 6.18808985e-01
7.03836977e-01 -6.73211277e-01 1.68304332e-02 -6.93488717e-01
3.26584786e-01 -4.25098240e-01 -5.16107380e-01 4.27896418e-02
-1.77628115e-01 1.31136432e-01 3.31346095e-01 -1.28627017e-01
-1.12990844e+00 5.34120798e-02 -2.93102294e-01 -6.49749756e-01
-3.35658520e-01 -2.21466254e-02 -3.60087305e-01 5.71197234e-02
7.49139607e-01 6.64165318e-02 3.16161774e-02 -3.50010633e-01
8.27352047e-01 4.81697768e-01 6.78635836e-01 -9.74872649e-01
8.99374485e-01 6.46312237e-01 5.26341125e-02 -3.70204717e-01
-5.34262359e-01 1.58501744e-01 -4.81259286e-01 -5.47846034e-02
8.82766485e-01 -1.28251386e+00 -6.25185788e-01 6.71106339e-01
-1.15922606e+00 -1.09164703e+00 -5.07527292e-01 3.66112441e-01
-8.41436625e-01 1.74624007e-02 -6.44728720e-01 -3.42558891e-01
-2.59452518e-02 -1.21061933e+00 1.58092821e+00 3.43559504e-01
-1.98688731e-01 -1.14382243e+00 2.56264508e-01 2.87294656e-01
4.14554209e-01 5.58803022e-01 5.60697675e-01 2.23373491e-02
-8.14739466e-01 1.90342739e-01 -3.59408677e-01 5.05650878e-01
3.84955741e-02 2.36457542e-01 -1.39338875e+00 -1.92625925e-01
-4.56025451e-01 -8.46550941e-01 8.48839343e-01 2.30770126e-01
1.06763363e+00 -1.08258091e-02 -5.51428199e-02 1.14260375e+00
1.37576473e+00 -2.67008513e-01 7.88910210e-01 6.70287907e-02
8.73108745e-01 4.71713483e-01 8.15840662e-02 3.67343545e-01
7.41699696e-01 4.79849190e-01 4.98422652e-01 -8.41531232e-02
-5.46959221e-01 -5.80246031e-01 3.95893037e-01 5.26385486e-01
-1.18294328e-01 -6.29678667e-01 -6.96478486e-01 5.17633319e-01
-1.57225561e+00 -8.51951003e-01 1.33930773e-01 1.86938643e+00
1.09487557e+00 -1.44122809e-01 1.36068374e-01 -2.88279772e-01
4.27861542e-01 7.30456933e-02 -6.32775068e-01 -3.27029288e-01
-4.29349452e-01 4.69623595e-01 4.43514705e-01 7.53312051e-01
-7.83441424e-01 8.79685640e-01 6.71567965e+00 4.20917571e-01
-1.23148954e+00 -3.88836977e-03 7.66587138e-01 -3.31392586e-01
-1.14362288e+00 8.68346617e-02 -6.02976501e-01 3.88085097e-01
8.12023997e-01 3.91321987e-01 5.86838007e-01 4.83118445e-01
-7.47304410e-02 -1.34119213e-01 -1.28069651e+00 9.50923622e-01
2.78705209e-02 -1.37043452e+00 1.21140748e-01 -3.71120125e-02
1.14949703e+00 1.14871189e-01 1.85082376e-01 -4.43356782e-02
6.35506392e-01 -1.11386120e+00 8.51837158e-01 6.24896407e-01
1.35042918e+00 -4.32858586e-01 2.86343008e-01 -1.05459832e-01
-6.79978728e-01 4.71502483e-01 -6.52327687e-02 2.62281299e-01
3.18026423e-01 3.32161009e-01 -6.82079971e-01 4.28961247e-01
5.85539579e-01 7.62235105e-01 -4.57352817e-01 6.12050056e-01
-7.71373272e-01 1.33856207e-01 -5.03740191e-01 3.54163468e-01
7.41026253e-02 -5.63842990e-02 2.57714987e-01 1.00699091e+00
7.56288707e-01 -1.13320544e-01 -3.09144527e-01 1.25617445e+00
-3.82148713e-01 -4.80625987e-01 -6.26382053e-01 6.61819130e-02
5.74773669e-01 1.32193387e+00 -3.56463104e-01 -2.74709404e-01
-2.59923220e-01 1.28215814e+00 7.79628456e-02 6.36711717e-01
-1.06511641e+00 -1.98598057e-01 8.20070565e-01 -2.08934899e-02
2.91796505e-01 -3.28465223e-01 -4.32566613e-01 -1.36156285e+00
-4.92588542e-02 -9.47307944e-01 1.26403794e-01 -1.31722808e+00
-1.28362894e+00 5.38965821e-01 -3.29707824e-02 -1.13416433e+00
-6.17307007e-01 -6.05225563e-01 -5.06455541e-01 1.12652850e+00
-1.51995242e+00 -1.53857160e+00 -7.92973101e-01 8.35698903e-01
-1.41893318e-02 2.02363193e-01 9.22719657e-01 2.44990677e-01
-2.49316499e-01 7.15550601e-01 -1.90988675e-01 3.73014845e-02
9.39724565e-01 -1.34835982e+00 7.10028350e-01 6.65534675e-01
-9.87816378e-02 4.13215190e-01 6.15240574e-01 -3.33360016e-01
-1.41284668e+00 -9.78013754e-01 4.99909580e-01 -6.98719263e-01
3.87615174e-01 -4.66398925e-01 -3.83384973e-01 1.06612182e+00
4.85387087e-01 1.23974137e-01 4.08365995e-01 -3.43074292e-01
-5.09199083e-01 9.25692990e-02 -1.35532677e+00 7.07289040e-01
1.33793020e+00 -6.82912350e-01 3.61454412e-02 1.88296750e-01
8.54663372e-01 -9.34591293e-01 -1.06377029e+00 2.78626800e-01
7.00666130e-01 -1.17324638e+00 9.32696819e-01 -3.37266564e-01
8.48414481e-01 -4.79439497e-01 -5.11944175e-01 -1.44064093e+00
1.30429342e-01 -5.86475551e-01 -2.11583916e-02 1.21324325e+00
6.72882795e-01 -3.81140560e-01 8.64078641e-01 7.07468808e-01
-1.73692167e-01 -4.57208902e-01 -6.56864047e-01 -5.00825286e-01
2.76057363e-01 -5.13696134e-01 8.46988559e-01 7.64067233e-01
-4.55815226e-01 2.03895897e-01 -4.86908197e-01 1.17386855e-01
6.41974688e-01 8.37598667e-02 1.25680780e+00 -4.72201824e-01
-6.70714676e-01 -3.58072072e-01 -1.80045128e-01 -1.18044448e+00
-7.95984715e-02 -8.48292112e-01 6.99779093e-02 -1.53180444e+00
5.29866703e-02 -4.99221802e-01 1.35462344e-01 6.47388041e-01
1.24690078e-01 7.80111551e-01 3.00585896e-01 -6.22760551e-03
-3.32596421e-01 5.92711389e-01 1.68313181e+00 -5.46989180e-02
-1.64699435e-01 -3.32460016e-01 -7.36501873e-01 3.80867004e-01
9.21770155e-01 -2.61026889e-01 -5.68653464e-01 -9.27549958e-01
2.74155825e-01 4.50742617e-02 7.67055452e-01 -9.92969632e-01
-1.04980640e-01 -1.19135588e-01 7.09255159e-01 -2.26652458e-01
7.32585311e-01 -5.24864495e-01 5.23395598e-01 -1.18131209e-02
-4.25743520e-01 -5.24773970e-02 2.01475918e-01 8.92875344e-02
-2.56185979e-02 1.27497137e-01 8.74080122e-01 -1.84198245e-01
-5.71810365e-01 4.25952047e-01 1.32419303e-01 2.78084159e-01
7.45934844e-01 -6.21179007e-02 -5.78324139e-01 -7.49030113e-01
-7.91860759e-01 1.52454361e-01 1.24759078e+00 4.08109039e-01
6.99189842e-01 -1.64108384e+00 -7.03059137e-01 3.13354373e-01
2.77926236e-01 3.89963299e-01 2.39509508e-01 6.08058810e-01
-7.94959486e-01 -8.42053816e-02 -3.15885961e-01 -7.27388501e-01
-6.49912119e-01 1.06300108e-01 6.69822574e-01 1.25956431e-01
-2.94997483e-01 1.00576997e+00 1.88157931e-01 -8.39452565e-01
8.05429667e-02 -3.98131311e-01 7.68804789e-01 -2.88722783e-01
2.83197343e-01 7.54963234e-02 -2.04529911e-01 -3.38249594e-01
-3.63104671e-01 7.07622409e-01 6.16981387e-01 -6.96746886e-01
1.22070396e+00 -2.66567647e-01 3.99453789e-02 4.17839259e-01
1.41220951e+00 -2.79644076e-02 -1.80521619e+00 -2.32159253e-02
-7.24123299e-01 -3.81606907e-01 -1.63205341e-01 -1.05304289e+00
-1.28199685e+00 9.17799950e-01 5.03161490e-01 -1.00699127e-01
1.13033640e+00 2.19552964e-01 8.75267088e-01 3.28641795e-02
1.55935556e-01 -6.88238740e-01 1.93188384e-01 4.21830028e-01
8.55614960e-01 -1.31011057e+00 -2.56806076e-01 -1.02457870e-02
-7.15567112e-01 1.07376432e+00 6.33711755e-01 -1.28186643e-01
4.12303448e-01 4.01696652e-01 3.43122780e-01 -5.96754393e-03
-7.40022719e-01 -2.42380053e-01 1.83713257e-01 8.72011304e-01
6.17202044e-01 -3.68048903e-03 3.30464482e-01 -1.36222830e-02
-2.96290994e-01 1.13765951e-02 4.57255930e-01 8.42656910e-01
5.39463125e-02 -1.39538527e+00 -2.61209846e-01 -4.50983346e-02
-3.24528962e-02 -1.47292539e-01 -4.49224591e-01 9.12217319e-01
2.68564194e-01 6.28808856e-01 2.64354050e-01 -4.99652117e-01
1.36073798e-01 2.77532190e-02 1.03166509e+00 -4.67690349e-01
-2.42997706e-01 -1.21081565e-02 2.01414585e-01 -7.62525439e-01
-5.62269509e-01 -5.78441620e-01 -1.30612016e+00 -7.37675190e-01
2.51476049e-01 -2.58790046e-01 9.05032218e-01 4.94731128e-01
5.55185974e-01 3.47802699e-01 5.99977493e-01 -1.22869885e+00
-1.66512296e-01 -8.84031892e-01 -4.19794470e-01 6.82269454e-01
4.86338735e-01 -3.98082227e-01 -4.57977653e-01 2.22525910e-01] | [11.619135856628418, -0.48022836446762085] |
a4b7e0c8-b672-4ed9-a981-9d1b45b23e43 | how-are-policy-gradient-methods-affected-by | 2206.06863 | null | https://arxiv.org/abs/2206.06863v1 | https://arxiv.org/pdf/2206.06863v1.pdf | How are policy gradient methods affected by the limits of control? | We study stochastic policy gradient methods from the perspective of control-theoretic limitations. Our main result is that ill-conditioned linear systems in the sense of Doyle inevitably lead to noisy gradient estimates. We also give an example of a class of stable systems in which policy gradient methods suffer from the curse of dimensionality. Our results apply to both state feedback and partially observed systems. | ['Nikolai Matni', 'Henrik Sandberg', 'Anastasios Tsiamis', 'Ingvar Ziemann'] | 2022-06-14 | null | null | null | null | ['policy-gradient-methods'] | ['methodology'] | [-1.00619607e-01 1.53368473e-01 -5.60557902e-01 5.79439327e-02
-5.23357451e-01 -4.92961586e-01 6.59136713e-01 -2.60807604e-01
-6.11884177e-01 1.39976954e+00 9.41247270e-02 -6.52582645e-01
-9.54517573e-02 -2.05530554e-01 -6.31797075e-01 -7.30278134e-01
-2.86643624e-01 -1.21381171e-01 7.62627795e-02 -5.82616508e-01
3.07721078e-01 3.67054760e-01 -1.01024103e+00 -4.79188681e-01
7.29355037e-01 8.22557092e-01 7.90303871e-02 9.22701776e-01
3.08437705e-01 7.83042014e-01 -4.95119810e-01 5.82587004e-01
6.74604237e-01 -6.35508597e-01 -2.96601653e-01 2.05057949e-01
1.70424089e-01 -4.64468092e-01 -6.27821505e-01 1.39494336e+00
8.51218998e-01 4.01615590e-01 5.56870937e-01 -1.12484050e+00
-4.02357936e-01 4.21253592e-01 -1.18845917e-01 2.85024554e-01
2.03399971e-01 5.46761692e-01 6.10430896e-01 -8.61234784e-01
6.51084483e-01 1.42381990e+00 5.72787344e-01 8.84884298e-01
-1.23810887e+00 -1.39303714e-01 5.13521373e-01 -3.26419175e-02
-9.51057613e-01 -7.17116833e-01 5.49469471e-01 -5.19486964e-01
5.12472451e-01 1.66296855e-01 5.51576316e-01 8.96414936e-01
6.10233009e-01 9.32366848e-01 1.55025482e+00 -2.53739536e-01
8.00014496e-01 2.32395142e-01 -1.56381980e-01 8.24071050e-01
2.99843818e-01 9.02362049e-01 -1.78600341e-01 -4.65276420e-01
7.22341537e-01 -2.39074200e-01 -4.90082830e-01 -6.59586549e-01
-1.05509031e+00 9.89767551e-01 2.39876300e-01 6.27546981e-02
-5.07937312e-01 4.02038425e-01 2.10872203e-01 9.06387508e-01
4.67824340e-01 7.15381801e-01 -3.44982743e-01 -1.02863930e-01
-4.45396334e-01 5.20369589e-01 1.06566107e+00 8.83716166e-01
3.64519179e-01 8.22223544e-01 -2.09052518e-01 1.85764119e-01
4.29005265e-01 1.06734943e+00 3.46330136e-01 -1.24208212e+00
2.57982910e-01 -3.90675180e-02 9.49454665e-01 -5.70289791e-01
-4.39376175e-01 -5.59419990e-01 -6.18599713e-01 5.86748838e-01
4.92953360e-01 -9.26937878e-01 -7.08890319e-01 1.82843137e+00
3.21225911e-01 -1.62435174e-01 1.21322557e-01 1.09926188e+00
-2.29772314e-01 6.27716541e-01 -4.94512767e-01 -1.04355359e+00
1.92788735e-01 -4.54771936e-01 -1.17619443e+00 -3.96898031e-01
4.68959898e-01 -4.55266505e-01 1.04135358e+00 1.86639264e-01
-1.07506061e+00 1.80641025e-01 -1.06490612e+00 4.96310145e-01
-1.22454345e-01 -2.96295937e-02 1.44405320e-01 2.08565608e-01
-1.21466386e+00 8.35514486e-01 -8.22863817e-01 -3.01597476e-01
-2.18769148e-01 2.39180461e-01 3.30233932e-01 5.28552651e-01
-9.85786855e-01 1.33830345e+00 6.90694973e-02 2.87057996e-01
-1.03025031e+00 -4.24265414e-01 -5.41595340e-01 -3.00299734e-01
7.49718845e-01 -4.14973766e-01 1.61883831e+00 -6.90360010e-01
-2.12372446e+00 -1.59243215e-02 -2.56605685e-01 -4.21657562e-01
9.69218612e-01 -2.81406820e-01 1.18566692e-01 -1.10618979e-01
-3.73546383e-03 -1.90977022e-01 1.09328496e+00 -8.98822010e-01
-5.43788433e-01 -2.31519148e-01 -2.83923984e-01 4.33505118e-01
7.78962101e-04 -1.89960450e-01 7.23229825e-01 -1.83301881e-01
-4.19555828e-02 -1.21152723e+00 -9.18988466e-01 9.93406773e-02
-4.15569395e-01 -1.06280990e-01 8.33955050e-01 -6.98941574e-02
1.16643107e+00 -1.63415039e+00 4.11922336e-01 1.08845420e-01
-6.13504322e-03 2.18416363e-01 2.33426377e-01 6.17777586e-01
1.50674462e-01 1.11405790e-01 -1.63138568e-01 9.54418406e-02
3.27901781e-01 2.04656318e-01 -8.02335620e-01 1.16147709e+00
-5.23112044e-02 5.98577976e-01 -1.21097243e+00 -2.81413436e-01
3.11102569e-01 -1.89518467e-01 -3.78857702e-01 3.46774026e-03
-2.46566728e-01 4.35060412e-01 -9.20482337e-01 1.40117943e-01
1.74805205e-02 -3.26840393e-02 6.18524244e-03 5.47716737e-01
-5.71111977e-01 7.32450783e-02 -1.20613360e+00 9.93869781e-01
-3.03783268e-01 6.97617769e-01 6.70422852e-01 -1.12799156e+00
3.24614584e-01 5.04189968e-01 5.66891909e-01 -1.63103893e-01
4.04359072e-01 2.80532718e-01 2.62720995e-02 -4.98881102e-01
2.71369368e-01 -3.60411197e-01 -6.55520558e-02 4.49505866e-01
-2.07483813e-01 -7.40051150e-01 2.09790751e-01 1.71705127e-01
8.49525571e-01 -1.25949785e-01 6.36666656e-01 -9.53898072e-01
2.72987962e-01 -6.51485622e-02 5.65334320e-01 1.13888454e+00
-7.34814465e-01 -1.47796571e-01 5.98591566e-01 -2.36391723e-01
-1.27905667e+00 -7.12573290e-01 5.74014746e-02 8.47065449e-01
-2.73938030e-02 -1.76934656e-02 -4.04529810e-01 -2.21953943e-01
3.66536587e-01 5.05985558e-01 -6.90675020e-01 -2.20936328e-01
-5.01192153e-01 -6.81881905e-01 -6.42733974e-03 6.38898537e-02
1.44841045e-01 -6.48451269e-01 -6.88255608e-01 4.63347077e-01
3.13667148e-01 -7.27938712e-01 -6.06812835e-01 -3.21189426e-02
-1.01524389e+00 -9.24760103e-01 -9.30030644e-01 -1.62271813e-01
6.35677040e-01 -1.32523239e-01 5.63298404e-01 -3.46836179e-01
4.31508273e-02 7.43624926e-01 3.88074785e-01 -5.53930700e-01
-7.15966940e-01 -1.53973326e-01 9.76846278e-01 -1.09002806e-01
-4.16312665e-01 -2.01578304e-01 -4.58584338e-01 4.53145921e-01
-3.60956252e-01 -3.92727613e-01 1.32150963e-01 1.18539214e+00
4.94681805e-01 -5.02363622e-01 6.34735525e-01 -5.19866347e-01
1.46429682e+00 -4.43697482e-01 -1.37189782e+00 9.66384262e-02
-9.83815551e-01 3.67159694e-01 6.41433954e-01 -7.61631250e-01
-8.41663897e-01 3.04045022e-01 5.53506374e-01 -2.89509773e-01
4.43824857e-01 3.55655760e-01 5.92130065e-01 -2.50327855e-01
1.17902803e+00 3.97515148e-01 4.07534689e-01 -1.05306312e-01
5.26389837e-01 4.96111125e-01 -2.74891816e-02 -5.55612803e-01
5.37077725e-01 4.44528580e-01 3.74958515e-01 -1.09574497e+00
-6.60720110e-01 -2.11497337e-01 -1.23333089e-01 -4.14790064e-01
2.54966021e-01 -5.50605655e-01 -8.81353855e-01 3.07265997e-01
-7.47062147e-01 -9.09119964e-01 -8.45453262e-01 5.62029779e-01
-1.19903529e+00 1.53242454e-01 -8.85357320e-01 -1.55900919e+00
-1.57234501e-02 -1.01393914e+00 5.46085536e-01 3.32118988e-01
2.07720444e-01 -1.02239811e+00 4.25272942e-01 -1.00638616e+00
6.21416569e-01 1.02455199e-01 2.20360026e-01 1.29448110e-02
-1.46010756e-01 -2.62657195e-01 3.41343135e-01 1.82427227e-01
8.10243487e-02 -4.68624793e-02 -4.97942835e-01 -6.22052848e-01
5.27149856e-01 -2.24113867e-01 6.30561352e-01 8.54056537e-01
2.56519884e-01 -6.57094777e-01 -4.06107455e-01 1.48594335e-01
1.33932805e+00 2.59608984e-01 -2.45227888e-01 7.59290159e-02
4.01567072e-01 4.62120652e-01 5.84517360e-01 6.03604794e-01
-1.96589842e-01 2.96609789e-01 1.49803057e-01 1.87100261e-01
5.03579438e-01 -2.63964385e-01 6.31935835e-01 7.05762327e-01
3.71560007e-01 -8.04769620e-02 -5.60668766e-01 6.92871392e-01
-2.22791791e+00 -9.93625045e-01 1.15312517e-01 2.19400263e+00
9.31575477e-01 -1.43242657e-01 3.15096706e-01 -1.25979051e-01
8.59533906e-01 7.78104737e-02 -1.17345202e+00 -5.00572324e-01
-1.63037017e-01 -4.70614433e-01 1.12417459e+00 1.06819057e+00
-9.11806881e-01 8.77680302e-01 8.32611561e+00 4.37081039e-01
-1.09422421e+00 -1.08047538e-02 3.21873933e-01 -3.89931262e-01
2.66341418e-01 -3.22193280e-02 -6.62276208e-01 5.11477113e-01
9.98081684e-01 -9.21795011e-01 7.44768679e-01 9.54637647e-01
8.85579348e-01 -3.42810184e-01 -6.61136329e-01 6.59461379e-01
-5.23502350e-01 -8.13864887e-01 -7.75692046e-01 3.31797719e-01
1.13459241e+00 2.45744810e-01 3.85363221e-01 3.17219913e-01
9.40174282e-01 -5.38074136e-01 5.99884510e-01 4.71791327e-01
3.68093342e-01 -5.29522538e-01 3.35574299e-01 6.08013034e-01
-6.81041300e-01 -6.17117405e-01 -5.93779922e-01 -5.60408294e-01
5.20435929e-01 6.24875128e-01 -7.62421072e-01 -1.72185361e-01
2.43778923e-03 6.26713514e-01 1.27354980e-01 1.03234613e+00
-2.88764209e-01 6.93560839e-01 -9.08277988e-01 -8.46649885e-01
5.44093430e-01 -2.80212671e-01 1.13231421e+00 7.86839664e-01
2.13075355e-01 2.30912313e-01 4.94790733e-01 6.81397617e-01
5.39598167e-01 8.33755508e-02 -1.28923488e+00 5.32799512e-02
1.58252671e-01 6.54515505e-01 -4.24910933e-01 -6.58667207e-01
1.35997623e-01 5.67775726e-01 1.07271813e-01 7.66648889e-01
-2.43819788e-01 -2.85912007e-01 9.40882444e-01 -2.39434093e-01
1.12655312e-01 -5.38630307e-01 -3.94040048e-02 -1.39679265e+00
-1.62730198e-02 -5.14922142e-01 -5.88194840e-02 -1.38412431e-01
-1.03407502e+00 1.28558517e-01 6.30593896e-02 -1.01766491e+00
-1.05148542e+00 -4.21470255e-01 -3.87476504e-01 5.16309679e-01
-9.63176548e-01 1.04682721e-01 6.73457742e-01 4.34099644e-01
4.87407774e-01 3.31188142e-02 3.73071015e-01 -2.17497051e-01
-6.11073256e-01 -6.76054582e-02 1.05899286e+00 -4.37908262e-01
4.24815416e-01 -1.65918374e+00 1.14585049e-01 9.59041715e-01
-3.85016352e-01 4.18677479e-01 1.44440365e+00 -5.39784133e-01
-1.70723081e+00 -8.03507984e-01 3.52634728e-01 -1.89300731e-01
1.07927549e+00 -1.96502954e-01 -3.94974142e-01 6.25691772e-01
-7.92119727e-02 1.79167941e-01 -3.25978041e-01 -2.16592342e-01
4.86300319e-01 2.12593928e-01 -1.03836560e+00 8.78290296e-01
9.12352741e-01 -5.77991188e-01 -3.27302843e-01 6.75734282e-01
5.03930807e-01 -4.72552210e-01 -4.69133824e-01 1.22619629e-01
3.94039840e-01 -3.99452358e-01 4.90219414e-01 -1.08920074e+00
-2.43451416e-01 -1.57908320e-01 3.82787026e-02 -1.97743428e+00
-3.27077396e-02 -1.50691271e+00 -4.20986056e-01 1.70478225e-01
3.12182933e-01 -8.91569257e-01 5.25667012e-01 5.17413318e-01
1.72021434e-01 -6.17324710e-01 -1.11535347e+00 -1.24588048e+00
4.18770015e-01 -7.73848295e-02 -3.32016379e-01 7.53286183e-01
8.05132270e-01 3.81610721e-01 -6.54856145e-01 -9.28216204e-02
7.45048463e-01 -1.52892128e-01 4.21967655e-01 -5.82927108e-01
-1.03925556e-01 -6.07989550e-01 9.85626783e-03 -1.17941427e+00
2.90352225e-01 -1.53598636e-01 7.44109809e-01 -1.10718369e+00
-4.75336164e-01 -1.75255805e-01 -3.96923125e-01 -2.00432703e-01
-2.85557359e-01 -2.52040058e-01 3.29458117e-01 1.45748436e-01
-6.50117934e-01 8.70943546e-01 1.16223526e+00 9.39724036e-03
-5.77466965e-01 5.15553772e-01 -1.72874823e-01 6.78465307e-01
9.97957945e-01 -5.92977345e-01 -6.72312975e-01 -2.22066447e-01
1.72218800e-01 5.76378047e-01 2.32674759e-02 -5.86364985e-01
2.09335119e-01 -7.19612598e-01 -2.53213525e-01 -1.85549632e-01
1.59586743e-01 -3.57514471e-01 -4.92954642e-01 1.14500272e+00
-6.62124455e-01 3.95064563e-01 -2.47444659e-02 8.88467073e-01
1.51436508e-01 -1.82082102e-01 8.64733636e-01 -3.17578733e-01
-2.88938910e-01 3.80291641e-01 -1.12882614e+00 3.46998364e-01
6.74986124e-01 3.88667345e-01 -1.15439124e-01 -1.07012320e+00
-9.04192984e-01 4.43598241e-01 2.78122813e-01 -9.24197808e-02
3.73368055e-01 -9.84106183e-01 -5.98629177e-01 -1.94999069e-01
-5.87910473e-01 -7.13709056e-01 -3.16253304e-01 8.13407302e-01
5.14336824e-02 6.23276889e-01 1.37129933e-01 -5.10009706e-01
-6.38091624e-01 6.69374526e-01 7.61081874e-01 -3.58900093e-02
-4.31599259e-01 5.12287319e-01 -8.15364644e-02 -3.39635223e-01
1.41969204e-01 -6.38221920e-01 3.62736732e-01 -1.93891481e-01
5.90334475e-01 5.16821861e-01 -4.05645132e-01 -8.84828418e-02
-3.90899815e-02 2.43607014e-01 3.22167158e-01 -1.12107766e+00
9.74885821e-01 -4.75953907e-01 2.77328253e-01 8.48973036e-01
8.74550879e-01 -4.62916017e-01 -1.76499522e+00 -3.93986970e-01
3.02725971e-01 -2.20686540e-01 2.28089690e-01 -3.04927677e-01
-5.94892502e-01 7.00681686e-01 8.85738552e-01 3.95282030e-01
5.42215407e-01 -4.33123589e-01 1.06963396e-01 1.07425773e+00
4.87437040e-01 -1.47888005e+00 -3.72378796e-01 8.16111684e-01
8.69326353e-01 -1.29665029e+00 2.38502279e-01 2.90815592e-01
-4.57619220e-01 8.53899121e-01 3.35506350e-01 -8.12226355e-01
7.77967036e-01 2.68266171e-01 -6.34339005e-02 2.19610170e-01
-1.17363620e+00 -3.99964660e-01 -1.27557144e-01 5.41805446e-01
-5.49848145e-03 1.57949120e-01 -8.83150876e-01 -4.41259086e-01
1.32502675e-01 1.75910383e-01 1.01255715e+00 1.40958464e+00
-8.29870164e-01 -5.57423651e-01 -2.69852072e-01 2.80066788e-01
-4.46947604e-01 2.12198153e-01 -3.72993767e-01 7.60212064e-01
-9.57654238e-01 9.63529408e-01 -3.10508132e-01 1.56473815e-01
3.09571654e-01 -5.82553074e-02 4.53949690e-01 -3.27683806e-01
-1.91506967e-01 2.46189266e-01 -1.10591665e-01 -8.30808640e-01
-1.53791025e-01 -6.44070029e-01 -7.01161683e-01 -2.80806273e-01
-5.99935353e-01 3.83199155e-01 8.14352810e-01 8.92080486e-01
1.48149997e-01 1.80276826e-01 9.02309835e-01 -8.07132959e-01
-1.69848168e+00 -1.18141365e+00 -6.53797865e-01 -1.45830333e-01
1.08140087e+00 -7.65277147e-01 -7.80357420e-01 -3.92313063e-01] | [4.429690361022949, 2.567796468734741] |
a87c3c2c-78b7-48c4-aa15-75df06ea3189 | out-of-distribution-detection-for-generalized | 1904.08703 | null | https://arxiv.org/abs/1904.08703v2 | https://arxiv.org/pdf/1904.08703v2.pdf | Out-of-Distribution Detection for Generalized Zero-Shot Action Recognition | Generalized zero-shot action recognition is a challenging problem, where the task is to recognize new action categories that are unavailable during the training stage, in addition to the seen action categories. Existing approaches suffer from the inherent bias of the learned classifier towards the seen action categories. As a consequence, unseen category samples are incorrectly classified as belonging to one of the seen action categories. In this paper, we set out to tackle this issue by arguing for a separate treatment of seen and unseen action categories in generalized zero-shot action recognition. We introduce an out-of-distribution detector that determines whether the video features belong to a seen or unseen action category. To train our out-of-distribution detector, video features for unseen action categories are synthesized using generative adversarial networks trained on seen action category features. To the best of our knowledge, we are the first to propose an out-of-distribution detector based GZSL framework for action recognition in videos. Experiments are performed on three action recognition datasets: Olympic Sports, HMDB51 and UCF101. For generalized zero-shot action recognition, our proposed approach outperforms the baseline (f-CLSWGAN) with absolute gains (in classification accuracy) of 7.0%, 3.4%, and 4.9%, respectively, on these datasets. | ['Saikumar Dwivedi', 'Fahad Shahbaz Khan', 'Devraj Mandal', 'Vikram Gupta', 'Shuaib Ahmed', 'Sanath Narayan', 'Ling Shao'] | 2019-04-18 | out-of-distribution-detection-for-generalized-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Mandal_Out-Of-Distribution_Detection_for_Generalized_Zero-Shot_Action_Recognition_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Mandal_Out-Of-Distribution_Detection_for_Generalized_Zero-Shot_Action_Recognition_CVPR_2019_paper.pdf | cvpr-2019-6 | ['zero-shot-action-recognition'] | ['computer-vision'] | [ 7.50361741e-01 2.21174419e-01 -2.74823993e-01 -2.82063663e-01
-8.39377880e-01 -4.24456984e-01 7.25460768e-01 -4.38786596e-01
-2.02192381e-01 6.57118499e-01 2.08619863e-01 1.30557850e-01
2.04913244e-01 -6.64725780e-01 -1.00979841e+00 -9.28218424e-01
2.58896738e-01 2.68529862e-01 4.59512889e-01 1.38584236e-02
1.75251737e-01 3.88191909e-01 -1.81221485e+00 6.28633738e-01
4.16115016e-01 9.51654434e-01 -3.76652271e-01 8.12441587e-01
1.78929910e-01 1.36281645e+00 -8.90154600e-01 -2.62364477e-01
4.72694397e-01 -1.00457335e+00 -5.43565452e-01 6.35235012e-01
6.85284615e-01 -6.33048117e-01 -6.21703207e-01 9.75364566e-01
2.88426876e-01 3.62902850e-01 8.37003887e-01 -1.69261742e+00
-7.06891298e-01 2.59924769e-01 -3.91464174e-01 2.92679012e-01
4.98011678e-01 4.08950508e-01 6.72567129e-01 -8.93119216e-01
7.38206863e-01 1.19936526e+00 2.85806209e-01 1.16219294e+00
-9.24496770e-01 -6.51453793e-01 1.77818850e-01 4.62351292e-01
-1.18296695e+00 -6.53139949e-01 6.32237136e-01 -5.76464593e-01
9.60507095e-01 1.56994864e-01 5.06669641e-01 1.53743756e+00
2.34818816e-01 1.16362596e+00 8.51782441e-01 -3.85615796e-01
5.57545662e-01 -2.68115759e-01 -4.83450107e-02 2.68429697e-01
-7.13868299e-03 2.43647158e-01 -5.25151253e-01 1.06821932e-01
6.47043884e-01 2.11318403e-01 -1.20679617e-01 -6.65243328e-01
-9.87071455e-01 7.44010806e-01 1.53417245e-01 2.01610088e-01
-3.72270912e-01 3.58362198e-01 4.93180573e-01 3.64165068e-01
4.95403409e-01 1.03034534e-01 -2.42179960e-01 -5.53632736e-01
-6.23911202e-01 1.04718581e-01 4.26309645e-01 8.88456643e-01
2.45385870e-01 4.44307059e-01 -5.52148402e-01 6.58052266e-01
-2.11874228e-02 6.74254000e-01 7.85527885e-01 -8.20038080e-01
3.38381410e-01 5.70669711e-01 3.53897884e-02 -7.95306146e-01
9.21795815e-02 -9.79688466e-02 -6.17206454e-01 1.87458754e-01
6.32751465e-01 -2.66154204e-02 -1.46047795e+00 1.59532142e+00
2.30240658e-01 6.68442369e-01 5.63918591e-01 8.58331680e-01
8.34670007e-01 4.94907886e-01 1.41587526e-01 1.65317193e-01
1.10341859e+00 -1.05613911e+00 -5.96131444e-01 -3.39318633e-01
6.55068338e-01 -4.32977259e-01 7.88471341e-01 3.64470273e-01
-6.26926720e-01 -6.84241593e-01 -9.39662099e-01 2.76163787e-01
-2.11270437e-01 1.48122072e-01 2.59365201e-01 5.79869449e-01
-5.69557190e-01 5.37483513e-01 -9.88579571e-01 -4.81528908e-01
7.47097075e-01 1.67908251e-01 -4.80100155e-01 -3.72567445e-01
-1.00239837e+00 5.56213498e-01 3.46017867e-01 -1.70569316e-01
-1.38254797e+00 -2.91059256e-01 -9.45068359e-01 -6.86367452e-02
7.34037817e-01 -1.81855202e-01 1.23065567e+00 -1.56330049e+00
-1.35695076e+00 8.02802265e-01 1.20893478e-01 -6.73286378e-01
4.48491722e-01 -1.01373792e-01 -6.40905619e-01 2.98374116e-01
1.06603153e-01 4.24575686e-01 1.14291775e+00 -8.99970531e-01
-8.05686712e-01 -4.49031532e-01 2.02063516e-01 1.27499849e-01
-9.72826257e-02 -1.30043283e-01 -7.66171217e-02 -8.07676017e-01
4.51012813e-02 -1.02951014e+00 1.34391561e-01 -1.80904977e-02
-3.93985510e-01 -2.16133028e-01 1.07073212e+00 -5.65546632e-01
8.12998176e-01 -2.40512967e+00 1.02398343e-01 -3.29562813e-01
-9.68099982e-02 5.23783803e-01 -2.42765963e-01 3.26441288e-01
-1.56639293e-01 -2.50321925e-01 -3.97920869e-02 -3.96476984e-02
1.52940713e-02 5.29640973e-01 -4.24204409e-01 5.51940024e-01
3.82958591e-01 8.60061407e-01 -1.05278945e+00 -1.13150991e-01
3.82696837e-01 3.49950165e-01 -4.31373864e-01 1.85117602e-01
2.45653335e-02 6.65891647e-01 -4.37162608e-01 8.41498792e-01
3.39221537e-01 2.36801937e-01 4.76845093e-02 5.88837080e-02
3.87426466e-01 -8.04740638e-02 -1.12794483e+00 1.47410178e+00
-8.58430937e-02 3.96714956e-01 -6.75003767e-01 -1.22538757e+00
8.37285519e-01 4.47675020e-01 5.19340158e-01 -5.92435896e-01
1.42486200e-01 1.55568197e-01 2.30277538e-01 -4.93188441e-01
3.30071479e-01 -4.67766047e-01 -3.72179687e-01 2.93848932e-01
5.10913193e-01 3.54885250e-01 2.68573821e-01 -1.46552557e-02
1.35867810e+00 3.65866393e-01 4.77455407e-01 1.77075133e-01
4.88603503e-01 -1.22915126e-01 7.29488909e-01 9.43540514e-01
-6.54904842e-01 7.75923729e-01 5.86781263e-01 -4.51102138e-01
-9.26108778e-01 -1.07032895e+00 1.11483797e-01 9.75894630e-01
3.74380015e-02 1.14456322e-02 -6.97207808e-01 -1.10243881e+00
2.05857269e-02 9.03580546e-01 -7.89444983e-01 -7.27584243e-01
-4.62069392e-01 -2.58464903e-01 6.81338608e-01 8.64161432e-01
5.47598362e-01 -1.31676841e+00 -7.38865614e-01 1.93859175e-01
-1.94498245e-02 -1.22618163e+00 -3.88043523e-01 -1.20814450e-01
-6.54993713e-01 -1.49055886e+00 -8.66671622e-01 -5.52515984e-01
6.66976929e-01 3.17196965e-01 8.06434333e-01 -1.22650430e-01
-3.29673767e-01 7.80657768e-01 -8.91206741e-01 -3.43898863e-01
-6.37419760e-01 -4.16824579e-01 1.10633008e-01 7.19129860e-01
8.68314743e-01 -1.82159141e-01 -5.65563798e-01 5.52246332e-01
-1.01615930e+00 -2.36793026e-01 4.93671119e-01 9.92528677e-01
5.73790610e-01 -9.32960138e-02 4.84226406e-01 -7.00115263e-01
-7.51095563e-02 -5.12320578e-01 -1.10322841e-01 1.67813778e-01
3.42979431e-02 -1.87068820e-01 7.12086320e-01 -6.18073642e-01
-9.68008697e-01 3.27772319e-01 -5.15942946e-02 -9.25842643e-01
-5.83571494e-01 -7.93240219e-02 -4.37077194e-01 1.16242193e-01
6.06377840e-01 5.67012727e-01 8.94720480e-02 -3.05223733e-01
2.75197327e-01 8.35585892e-01 5.46450377e-01 -1.33017018e-01
6.63294435e-01 6.32385254e-01 -1.38733909e-01 -8.90957773e-01
-9.44635630e-01 -6.49408996e-01 -7.58018315e-01 -4.26649451e-01
1.10046196e+00 -1.06220889e+00 3.04316767e-02 9.26731646e-01
-5.79843342e-01 -3.24891955e-01 -8.55222940e-01 4.90015924e-01
-9.59014416e-01 4.79749322e-01 -3.84794325e-01 -7.74286926e-01
-1.52560351e-02 -1.10319018e+00 1.33378863e+00 1.11788630e-01
-1.63157627e-01 -6.70602322e-01 -1.02650180e-01 4.66354042e-01
1.25737349e-02 5.84029794e-01 2.35812232e-01 -1.01752019e+00
-3.92772853e-01 -6.07192397e-01 2.58910328e-01 8.62796068e-01
3.39776754e-01 -1.47035137e-01 -9.12936211e-01 -3.93855423e-01
4.31697331e-02 -6.39393508e-01 9.14461255e-01 2.56290138e-01
1.10991395e+00 -2.49662489e-01 -1.90816343e-01 3.66132051e-01
1.17249715e+00 6.55756652e-01 1.10274994e+00 8.56845081e-02
6.43983364e-01 1.12283662e-01 1.04434669e+00 5.46159804e-01
-1.98199823e-01 7.96964824e-01 3.53040516e-01 2.31723845e-01
-3.19855928e-01 -4.56036627e-01 7.56574690e-01 1.66687176e-01
-1.61519065e-01 -6.65291846e-01 -5.84039390e-01 5.10772467e-01
-1.84892797e+00 -1.43013430e+00 -1.00355297e-02 2.26346564e+00
3.34477842e-01 1.49474859e-01 2.00668707e-01 3.43308806e-01
7.47168422e-01 2.38043949e-01 -7.69137383e-01 -2.12771058e-01
5.85983396e-02 2.48916343e-01 4.00976539e-01 2.80507025e-03
-1.58814657e+00 9.24934030e-01 4.86281586e+00 8.99734080e-01
-8.27078700e-01 9.04541165e-02 4.74192351e-01 -2.06956387e-01
4.29725230e-01 -2.56646108e-02 -7.81239092e-01 6.19175851e-01
8.24045002e-01 -7.46770129e-02 1.02658764e-01 1.13656533e+00
-8.90465081e-02 -1.73694491e-02 -1.14956546e+00 9.04437363e-01
6.46449029e-01 -7.01541007e-01 1.02397874e-01 1.06336460e-01
7.02088952e-01 -2.88404226e-01 -5.14687859e-02 7.95051038e-01
1.63559750e-01 -7.87673414e-01 6.59965992e-01 5.49446046e-01
8.40939045e-01 -6.75980508e-01 6.94308162e-01 3.72776777e-01
-9.34586823e-01 -3.16470534e-01 -5.05932629e-01 -2.05327243e-01
9.24072936e-02 1.98216699e-02 -5.22226810e-01 3.78521889e-01
4.77119267e-01 1.23957646e+00 -4.97138798e-01 8.10983598e-01
-4.48391199e-01 7.30685711e-01 1.47147760e-01 2.19591632e-01
2.75645971e-01 2.12032884e-01 6.18218362e-01 6.62553012e-01
4.18529630e-01 3.33053261e-01 3.19033265e-01 4.68290925e-01
2.58014142e-03 -2.40610808e-01 -7.71855712e-01 -2.63784885e-01
-1.09442212e-01 7.38993526e-01 -6.23376548e-01 -6.67418838e-01
-5.45824587e-01 1.33419299e+00 -1.10482350e-01 1.91546991e-01
-1.02757490e+00 -2.39334017e-01 7.61488140e-01 4.49872389e-03
7.30407119e-01 2.43489429e-01 4.22710717e-01 -1.54492688e+00
-2.41817031e-02 -1.10448873e+00 4.86511350e-01 -5.92150033e-01
-1.28140914e+00 4.01638716e-01 5.26500382e-02 -1.88225365e+00
-4.55740958e-01 -6.84325039e-01 -4.60185289e-01 2.35457823e-01
-8.76058221e-01 -1.10973871e+00 -3.39352727e-01 5.13485253e-01
1.02757192e+00 -6.10127151e-01 7.92661428e-01 2.29899421e-01
-4.04488802e-01 5.94091952e-01 4.16954041e-01 4.78010088e-01
5.20509958e-01 -9.98329520e-01 3.60101104e-01 1.03109372e+00
2.79761583e-01 -8.40401649e-02 6.24273360e-01 -6.76667392e-01
-1.27799070e+00 -1.34581363e+00 6.13066673e-01 -5.59900522e-01
5.97018301e-01 -3.61723632e-01 -7.79395998e-01 8.80096674e-01
-2.11256191e-01 3.44974220e-01 7.92957127e-01 -4.64062065e-01
-2.32889265e-01 8.64944607e-02 -1.17756426e+00 3.80684525e-01
1.26441431e+00 -2.90456623e-01 -7.78892517e-01 4.13720995e-01
2.20563009e-01 -3.61160725e-01 -7.76973903e-01 4.42366600e-01
6.46659732e-01 -1.02344584e+00 1.00763726e+00 -1.06714642e+00
6.95291996e-01 -2.02969342e-01 -3.99881035e-01 -1.30680323e+00
-1.09776586e-01 -7.60152377e-03 -3.28089714e-01 1.06777799e+00
-1.86923385e-01 -5.33496201e-01 8.47734153e-01 3.65598470e-01
-2.34574005e-01 -5.01877666e-01 -1.21558857e+00 -1.24228692e+00
-3.31174172e-02 -2.56313771e-01 4.21404421e-01 7.84725666e-01
-2.65880406e-01 1.65872961e-01 -7.58223832e-01 -2.41691306e-01
6.35461748e-01 9.72303003e-02 1.10104144e+00 -6.92070723e-01
-6.08963013e-01 -5.22575192e-02 -1.19546938e+00 -9.64353621e-01
1.63828045e-01 -5.96578896e-01 2.42683008e-01 -1.30044138e+00
2.93795258e-01 1.79272205e-01 -5.70044875e-01 7.33927011e-01
-5.46403304e-02 6.12816453e-01 4.86841351e-01 1.00903392e-01
-6.78884089e-01 7.75424778e-01 1.26719260e+00 -4.25782382e-01
1.31391257e-01 1.93260446e-01 -3.29531610e-01 7.91936040e-01
9.35796201e-01 -6.03043735e-01 -5.19774854e-01 -1.39210209e-01
-7.26755738e-01 8.53146687e-02 7.12384462e-01 -1.37444615e+00
-2.96011329e-01 -2.48338789e-01 3.64914149e-01 -4.19179708e-01
5.47123551e-01 -7.04913259e-01 1.14573367e-01 6.59791410e-01
-2.87294418e-01 -6.90842569e-01 3.07652950e-02 8.63766372e-01
-2.75100321e-01 -2.34064117e-01 7.86371231e-01 -2.02814400e-01
-1.33200800e+00 2.02957317e-01 -5.67608714e-01 1.73324987e-01
1.60844994e+00 -5.86851358e-01 -3.11591506e-01 -3.87541026e-01
-1.14713335e+00 -1.44662991e-01 4.60807234e-01 7.49721348e-01
8.34701061e-01 -1.43942976e+00 -6.84819818e-01 3.67781013e-01
4.58215922e-01 -4.26537633e-01 5.74533880e-01 7.84316182e-01
-2.69984871e-01 2.71848649e-01 -5.04347920e-01 -5.05140483e-01
-1.41540146e+00 7.59153247e-01 4.01344419e-01 -1.41177341e-01
-7.22704172e-01 7.72399485e-01 3.29352587e-01 -2.46590123e-01
1.71802774e-01 -2.07056835e-01 -1.48037821e-01 -7.92397484e-02
6.62408948e-01 4.33754534e-01 -2.48843297e-01 -1.14540446e+00
-4.19741184e-01 3.31998855e-01 8.60790536e-02 3.47699165e-01
1.14730203e+00 2.50410140e-01 5.14370680e-01 6.56777263e-01
1.18128097e+00 -5.03869474e-01 -1.52036297e+00 -4.09423746e-02
-2.66116410e-01 -8.80416393e-01 -3.68012220e-01 -7.37611353e-01
-1.14176762e+00 8.23781312e-01 8.71194601e-01 -3.45220864e-02
1.12186599e+00 5.13995737e-02 8.56047213e-01 1.18557364e-01
6.28142893e-01 -1.02075541e+00 3.93367857e-01 2.78843611e-01
8.76096845e-01 -1.21367049e+00 -1.63044885e-01 -3.09863031e-01
-9.82813597e-01 9.70135570e-01 8.35409820e-01 -5.24769664e-01
3.50426674e-01 -2.10683629e-01 -1.55347807e-03 -1.63148716e-01
-5.99232078e-01 -3.93579125e-01 3.65371764e-01 7.43935108e-01
-2.39308979e-02 8.58723372e-02 -2.56383181e-01 4.64362472e-01
2.63235807e-01 2.79681474e-01 7.31598794e-01 1.28056383e+00
-4.16089386e-01 -1.03194809e+00 -3.32539767e-01 6.69984043e-01
-4.67864901e-01 2.90176332e-01 -4.80634421e-01 8.87542844e-01
2.06565380e-01 9.72157836e-01 2.03331694e-01 -6.38796508e-01
5.65280497e-01 3.32766145e-01 5.32139719e-01 -7.65284956e-01
-9.17484388e-02 -1.18420944e-01 1.01833597e-01 -6.13909125e-01
-6.38990641e-01 -8.91445339e-01 -1.05479133e+00 2.40376204e-01
-1.64979145e-01 -2.34506860e-01 3.61183099e-02 8.79811287e-01
3.20489526e-01 5.54683685e-01 5.56180835e-01 -6.84457302e-01
-9.46527600e-01 -1.01160228e+00 -9.19649184e-01 9.75145698e-01
1.00440554e-01 -1.07026672e+00 -6.17033780e-01 2.77732402e-01] | [8.473299026489258, 0.8916239142417908] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.