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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
ced1b479-59a2-47f8-a88e-7317da9d802e | multi-geometry-spatial-acoustic-modeling-for | 1903.06539 | null | http://arxiv.org/abs/1903.06539v2 | http://arxiv.org/pdf/1903.06539v2.pdf | Multi-Geometry Spatial Acoustic Modeling for Distant Speech Recognition | The use of spatial information with multiple microphones can improve
far-field automatic speech recognition (ASR) accuracy. However, conventional
microphone array techniques degrade speech enhancement performance when there
is an array geometry mismatch between design and test conditions. Moreover,
such speech enhancement techniques do not always yield ASR accuracy improvement
due to the difference between speech enhancement and ASR optimization
objectives. In this work, we propose to unify an acoustic model framework by
optimizing spatial filtering and long short-term memory (LSTM) layers from
multi-channel (MC) input. Our acoustic model subsumes beamformers with multiple
types of array geometry. In contrast to deep clustering methods that treat a
neural network as a black box tool, the network encoding the spatial filters
can process streaming audio data in real time without the accumulation of
target signal statistics. We demonstrate the effectiveness of such MC neural
networks through ASR experiments on the real-world far-field data. We show that
our two-channel acoustic model can on average reduce word error rates (WERs)
by~13.4 and~12.7% compared to a single channel ASR system with the log-mel
filter bank energy (LFBE) feature under the matched and mismatched microphone
placement conditions, respectively. Our result also shows that our two-channel
network achieves a relative WER reduction of over~7.0% compared to conventional
beamforming with seven microphones overall. | [] | 2019-04-28 | null | null | null | null | ['distant-speech-recognition'] | ['speech'] | [ 3.37890208e-01 -3.11039448e-01 6.42952263e-01 -2.55299121e-01
-1.47830486e+00 -3.83550048e-01 7.75287300e-02 -6.30619749e-02
-5.88042378e-01 2.42443800e-01 6.22710645e-01 -4.57191914e-01
-1.66966811e-01 -3.18385363e-01 -7.94237077e-01 -7.65757263e-01
-8.17887336e-02 -3.14806044e-01 1.44003714e-02 -1.04172446e-01
-2.04006489e-03 3.40785712e-01 -1.60137916e+00 3.59720170e-01
5.28977096e-01 1.23287141e+00 5.70896924e-01 1.17904091e+00
8.66150185e-02 3.86828929e-01 -1.27729416e+00 3.30941677e-02
2.62069821e-01 6.72015473e-02 4.40927818e-02 -1.45935029e-01
4.57672924e-01 -4.03921932e-01 -7.07437277e-01 1.10044849e+00
1.40539920e+00 4.34178650e-01 3.15798044e-01 -5.34486175e-01
-4.08368051e-01 8.36963534e-01 -4.36518133e-01 2.96959519e-01
2.79335827e-01 3.95579748e-02 6.92293286e-01 -1.21382654e+00
-3.73583704e-01 1.27438295e+00 7.39287019e-01 4.46136892e-01
-7.88501263e-01 -8.93416345e-01 7.44726509e-02 7.95898438e-02
-1.54368746e+00 -1.17033124e+00 6.88438952e-01 -9.76843983e-02
1.47834623e+00 3.62460703e-01 2.55193114e-01 8.43236685e-01
-4.43891957e-02 6.18685305e-01 7.13217616e-01 -4.69529867e-01
3.47215116e-01 -2.37843350e-01 6.24880893e-03 7.28835687e-02
-2.00517133e-01 4.15279657e-01 -9.89538014e-01 3.18302140e-02
4.48329300e-01 -2.83551306e-01 -4.46544379e-01 6.56409860e-01
-1.10182369e+00 3.80287111e-01 2.30563566e-01 5.35187185e-01
-1.94622070e-01 4.95005071e-01 -6.15876243e-02 3.77025664e-01
5.79158723e-01 5.93077958e-01 -5.05725443e-01 -3.89831960e-01
-1.22654688e+00 -2.37745687e-01 4.89534408e-01 7.28874147e-01
2.52589881e-01 7.19500065e-01 -1.07768729e-01 1.44955993e+00
6.44676030e-01 1.07711720e+00 5.06634533e-01 -8.87794197e-01
5.63620448e-01 -4.64575291e-01 -6.67318478e-02 -8.32938612e-01
-2.62440711e-01 -1.12506998e+00 -8.83013189e-01 -4.31836247e-02
1.03083812e-01 -5.31111121e-01 -8.84558916e-01 1.76687014e+00
-2.30048597e-02 3.89540821e-01 1.64116547e-01 7.60468304e-01
7.16720343e-01 9.59992945e-01 -4.33621705e-01 -2.70283163e-01
8.96950364e-01 -8.92607331e-01 -1.24551392e+00 -4.07412559e-01
3.54512572e-01 -1.08613670e+00 8.30842376e-01 5.66352785e-01
-1.35958135e+00 -6.90931678e-01 -1.32554185e+00 3.87808293e-01
-3.10219109e-01 9.36541036e-02 2.94175558e-02 1.27768552e+00
-1.61878109e+00 2.34548926e-01 -7.87905753e-01 2.26333782e-01
-7.48405904e-02 5.26106060e-01 -6.87223226e-02 4.86663692e-02
-1.17535150e+00 4.00075346e-01 -1.75668210e-01 4.38723058e-01
-7.78413355e-01 -9.80768919e-01 -7.82629907e-01 3.10069561e-01
7.40960687e-02 -1.98165566e-01 1.54041600e+00 -5.82620561e-01
-1.82225430e+00 -8.87807086e-03 -6.13441348e-01 -2.77916133e-01
-2.36156136e-01 -4.09026504e-01 -9.49498773e-01 -3.05243637e-02
-4.27628249e-01 5.17801702e-01 8.49650145e-01 -1.11335862e+00
-6.40866697e-01 -2.39432946e-01 -3.29350710e-01 2.88508594e-01
-7.52066851e-01 3.09915751e-01 -3.34731162e-01 -1.08520830e+00
2.49396026e-01 -3.86990935e-01 -3.18366617e-01 -4.60206360e-01
-2.43335068e-01 3.27711552e-01 6.81810081e-01 -9.75174427e-01
1.73905575e+00 -2.46402311e+00 -3.30051512e-01 2.82886833e-01
6.34941906e-02 4.93816555e-01 -3.49364042e-01 6.01776801e-02
-1.69532642e-01 1.77246556e-01 -9.92791802e-02 -4.80997771e-01
-2.52673719e-02 -3.75315338e-01 -1.18675031e-01 3.97936791e-01
-5.15732095e-02 4.64051783e-01 -6.29696369e-01 5.34306727e-02
4.64537710e-01 8.33384395e-01 -7.66865730e-01 2.59670466e-01
3.93092424e-01 2.74458379e-01 2.13367283e-01 5.28541446e-01
7.72630632e-01 1.64409950e-01 -1.86227441e-01 -1.33783072e-01
-6.43830448e-02 4.59953755e-01 -1.46315420e+00 1.69176328e+00
-9.29073393e-01 9.58757699e-01 8.07260394e-01 -7.47493625e-01
9.60787833e-01 7.01852560e-01 2.53835529e-01 -9.47816968e-01
-6.32601529e-02 4.97621655e-01 3.02672982e-01 -1.68788165e-01
4.37597364e-01 -1.23400934e-01 1.96328983e-01 1.63806379e-01
1.20210215e-01 9.66393575e-02 -3.97448093e-01 -1.41698360e-01
1.44665658e+00 -8.68082464e-01 -1.73120186e-01 -8.61587003e-02
5.03885448e-01 -8.49752843e-01 3.67139757e-01 9.32717741e-01
-9.05688256e-02 6.44873202e-01 -5.70238292e-01 2.89338082e-01
-8.28112125e-01 -1.20126331e+00 -5.84846959e-02 1.19059241e+00
-1.09641336e-01 -4.23297614e-01 -7.67839372e-01 1.37525782e-01
-3.93790424e-01 7.71172225e-01 1.62989408e-01 -1.20424200e-02
-7.52360702e-01 -7.67993033e-01 9.10028636e-01 5.96072018e-01
4.99582112e-01 -4.26200777e-01 -1.95661739e-01 3.62630695e-01
-1.15658492e-01 -1.17870474e+00 -6.81392729e-01 4.10222530e-01
-3.98604602e-01 -1.31621107e-01 -8.63560438e-01 -6.89186931e-01
3.21953326e-01 4.93329942e-01 7.13341773e-01 -1.68593183e-01
6.24940023e-02 5.37443519e-01 -4.00990307e-01 -3.77800554e-01
-2.81294703e-01 -2.16491252e-01 3.38824242e-01 6.58871531e-02
1.85792923e-01 -8.48109961e-01 -7.34836698e-01 4.62843925e-01
-7.39265323e-01 -2.86674738e-01 6.14858031e-01 7.71635592e-01
2.48856828e-01 5.16954899e-01 8.56295943e-01 1.27662987e-01
7.23215044e-01 -2.59111762e-01 -3.95409733e-01 2.68763937e-02
-2.89885551e-01 -1.73247054e-01 4.70618278e-01 -5.00255585e-01
-1.31823647e+00 -1.69911399e-01 -8.73634338e-01 -3.57374489e-01
-3.34159732e-01 4.66880500e-01 -6.37545228e-01 -2.71967761e-02
5.52389324e-01 1.83434978e-01 -3.30520421e-01 -6.60992861e-01
1.83508769e-01 1.47916758e+00 4.93342876e-01 -2.68129081e-01
6.26489997e-01 1.67641252e-01 -2.66757429e-01 -1.36003256e+00
-1.18713342e-01 -7.41908073e-01 -9.74438861e-02 -2.97404349e-01
6.47234976e-01 -1.11940658e+00 -5.03643274e-01 6.74144626e-01
-1.17705715e+00 -3.52716327e-01 1.77600667e-01 9.02427912e-01
-1.53808966e-01 1.80750310e-01 -7.09385574e-01 -1.31636226e+00
-3.93072307e-01 -1.35932148e+00 1.23089075e+00 -1.24181382e-01
4.69145775e-02 -8.33358824e-01 -3.46714228e-01 2.54635245e-01
9.74960923e-01 -5.39724946e-01 4.46295083e-01 -5.13488412e-01
-2.13154033e-01 -1.04627229e-01 2.48934075e-01 5.55436373e-01
6.32236302e-02 -4.84341919e-01 -1.58439326e+00 -4.44943160e-01
3.49559456e-01 2.56631255e-01 6.39043331e-01 9.78162825e-01
1.29296398e+00 -2.94021755e-01 -1.20443374e-01 7.07041860e-01
1.12524247e+00 7.88318515e-01 6.41090870e-01 -2.05921575e-01
5.26821136e-01 3.75146896e-01 2.86302865e-01 3.50531518e-01
-2.83287436e-01 7.84312010e-01 1.31956786e-01 -2.31433004e-01
-5.78086317e-01 -1.41175658e-01 7.36490846e-01 1.41523969e+00
6.79399133e-01 -7.93374181e-01 -9.09916162e-01 5.56293070e-01
-1.23411179e+00 -7.77321279e-01 -1.30491983e-02 2.42410278e+00
5.74893057e-01 -1.28883883e-01 -3.03156167e-01 5.50159514e-01
1.00411546e+00 1.44795015e-01 -3.15689474e-01 -3.82532299e-01
-3.81460518e-01 5.64272702e-01 7.78091371e-01 1.06303656e+00
-8.96162748e-01 4.86695945e-01 6.11657333e+00 1.35458016e+00
-1.29497588e+00 4.71154988e-01 6.67022586e-01 -5.53139806e-01
-2.97457337e-01 -6.85391843e-01 -7.06695318e-01 3.42087924e-01
1.51812220e+00 3.13545346e-01 7.33333528e-01 3.55969936e-01
6.58040166e-01 1.52478755e-01 -1.04589963e+00 1.49035585e+00
1.11876339e-01 -1.00435662e+00 -3.44541609e-01 1.03802830e-01
3.94111872e-01 1.69747040e-01 5.23094296e-01 1.06899343e-01
-3.03692813e-03 -1.45461452e+00 8.59407127e-01 1.71696678e-01
1.04168856e+00 -6.64787769e-01 5.67422450e-01 3.84112716e-01
-1.42839575e+00 -4.49624836e-01 -8.67320895e-02 -2.06446379e-01
4.16695654e-01 9.62639868e-01 -8.16266298e-01 2.19067350e-01
9.46114123e-01 1.29686907e-01 -2.06550002e-01 1.19746053e+00
9.53554288e-02 1.12501931e+00 -6.55383170e-01 -9.70973149e-02
2.71222759e-02 3.15204024e-01 9.40674245e-01 1.46023893e+00
8.01748216e-01 1.12622604e-01 -3.33380073e-01 4.09868121e-01
-4.72912798e-03 -8.20209011e-02 -4.33773160e-01 1.15567781e-01
8.76397848e-01 8.23276460e-01 -2.67474473e-01 -3.82935740e-02
-3.22949111e-01 6.34136260e-01 -3.98065627e-01 7.46356785e-01
-7.71341324e-01 -5.74065328e-01 7.93610513e-01 1.35287464e-01
2.89494604e-01 -4.29208547e-01 -5.09799063e-01 -7.16734350e-01
1.39121443e-01 -9.63529110e-01 -2.82783687e-01 -7.85577297e-01
-1.01483607e+00 5.96062779e-01 -3.46798122e-01 -9.37309802e-01
-8.37330222e-02 -7.57446527e-01 -5.32099783e-01 1.10105371e+00
-1.31046438e+00 -7.32850194e-01 1.78100824e-01 4.87585515e-01
8.21699917e-01 -2.49435902e-01 6.67346716e-01 1.06407583e+00
-3.86833519e-01 1.10761058e+00 4.57714736e-01 9.53944325e-02
5.64856231e-01 -1.16589785e+00 4.70852822e-01 1.24103773e+00
6.00063130e-02 7.28138924e-01 7.28901744e-01 -4.12928492e-01
-1.46217978e+00 -1.14887476e+00 6.82887793e-01 -1.14925489e-01
4.28068757e-01 -7.28670359e-01 -7.49139845e-01 1.39843836e-01
5.07545471e-01 -1.58264443e-01 9.00473416e-01 5.04257940e-02
-7.50116482e-02 -3.72040689e-01 -8.82574975e-01 4.93927211e-01
1.06398821e+00 -5.70218265e-01 -3.58142436e-01 1.45501718e-01
1.15494263e+00 -2.94302851e-01 -7.35921323e-01 6.23526514e-01
4.16627795e-01 -6.74367905e-01 1.18524814e+00 1.24159910e-01
-3.60156447e-02 -5.58043540e-01 -9.41059530e-01 -1.57471478e+00
-1.64851248e-01 -9.91126120e-01 1.60442814e-02 1.40179443e+00
8.44016790e-01 -6.01862133e-01 3.61939639e-01 2.77013421e-01
-7.17606664e-01 -4.85529035e-01 -1.12604737e+00 -7.64043212e-01
5.58684915e-02 -1.20593178e+00 5.56107819e-01 1.74036384e-01
-1.49934337e-01 3.50211531e-01 -3.36220413e-01 7.24615932e-01
4.55930263e-01 -7.80800283e-01 3.36846292e-01 -6.12826228e-01
-7.43071973e-01 -5.91167510e-01 -1.30134881e-01 -1.65931404e+00
-4.83498201e-02 -5.67972660e-01 5.95538557e-01 -1.33211422e+00
-4.84629363e-01 -3.76423240e-01 -4.93772537e-01 -6.88035265e-02
-1.90232337e-01 3.97235975e-02 6.65483922e-02 -4.53919321e-01
-3.85407984e-01 5.61222136e-01 6.52498245e-01 -3.19469035e-01
-2.15516642e-01 2.08756700e-01 -6.51021957e-01 6.23575807e-01
7.15688407e-01 -1.81116149e-01 -4.49940085e-01 -1.00934315e+00
1.22828461e-01 2.67988175e-01 1.07261613e-01 -1.37690687e+00
7.24312663e-01 3.68239939e-01 3.49502444e-01 -5.60249448e-01
7.40320206e-01 -8.51688743e-01 1.88051283e-01 1.74355149e-01
-4.97768939e-01 -1.26264216e-02 4.87649053e-01 6.42619014e-01
-4.54558730e-01 1.43573001e-01 6.90131307e-01 2.37513199e-01
-4.46266919e-01 -1.68254420e-01 -7.96321452e-01 -3.26582283e-01
2.31159359e-01 -2.35409498e-01 -1.88032076e-01 -7.99712479e-01
-5.70081890e-01 -6.64129704e-02 -2.19091550e-01 2.10170627e-01
9.08482373e-01 -1.16531539e+00 -6.09753549e-01 3.09258342e-01
-4.06712234e-01 -1.44808099e-01 6.21401310e-01 6.16862357e-01
-8.76082573e-03 6.49965405e-01 5.43839812e-01 -7.22261250e-01
-1.35367739e+00 2.14012206e-01 6.35744631e-01 1.33401617e-01
-1.66425541e-01 1.34965670e+00 3.18389863e-01 -4.28429484e-01
7.18747199e-01 -3.13953578e-01 3.30810696e-02 -4.40040171e-01
7.87736297e-01 5.72987258e-01 6.06098056e-01 -5.37909627e-01
-2.90044695e-01 5.51853001e-01 5.02386212e-01 -8.66526902e-01
1.10116005e+00 -4.43276018e-01 2.60906667e-01 3.08063090e-01
1.40298665e+00 5.95499218e-01 -8.57133567e-01 -1.55539379e-01
-4.04630482e-01 -4.87746269e-01 6.94565892e-01 -1.04947829e+00
-1.00627136e+00 1.29905200e+00 1.07044101e+00 6.41078968e-03
1.58899164e+00 -1.78352699e-01 9.43321049e-01 4.09950435e-01
1.84829310e-01 -1.17803097e+00 1.27937660e-01 5.13349295e-01
8.55441213e-01 -7.82348335e-01 -8.44461858e-01 -1.25778362e-01
-8.87166858e-02 8.47436547e-01 3.89861345e-01 3.76678050e-01
9.99529481e-01 1.06044590e+00 3.00883800e-01 1.06398880e-01
-6.34211361e-01 -7.97824860e-02 2.17825055e-01 6.89753830e-01
4.58924264e-01 1.38478041e-01 2.69534975e-01 7.88493514e-01
-5.67654610e-01 -6.51553094e-01 1.33678317e-01 6.65078461e-01
-8.24744463e-01 -8.06078255e-01 -9.86376524e-01 4.03772116e-01
-6.79659724e-01 -5.24177551e-01 -9.58416313e-02 -3.30164358e-02
-6.61049709e-02 1.98222017e+00 2.20694184e-01 -8.94573927e-01
3.69564563e-01 -2.49677096e-02 5.08888848e-02 -4.80084985e-01
-6.72746241e-01 8.38016629e-01 5.87677211e-02 -3.57943416e-01
-1.31862178e-01 -4.23969805e-01 -1.01985395e+00 -2.35610992e-01
-6.68511987e-01 1.75370410e-01 1.07036984e+00 8.44964921e-01
4.98190045e-01 1.21359563e+00 7.34168828e-01 -7.62006879e-01
-4.91816700e-01 -1.11810827e+00 -6.11907721e-01 -2.65122473e-01
5.65286040e-01 -3.57859343e-01 -5.63622773e-01 -1.29522219e-01] | [14.946783065795898, 5.949449062347412] |
20854e07-c327-48a7-b65c-6f0757c02f4d | dru-net-an-efficient-deep-convolutional | 2004.13453 | null | https://arxiv.org/abs/2004.13453v1 | https://arxiv.org/pdf/2004.13453v1.pdf | DRU-net: An Efficient Deep Convolutional Neural Network for Medical Image Segmentation | Residual network (ResNet) and densely connected network (DenseNet) have significantly improved the training efficiency and performance of deep convolutional neural networks (DCNNs) mainly for object classification tasks. In this paper, we propose an efficient network architecture by considering advantages of both networks. The proposed method is integrated into an encoder-decoder DCNN model for medical image segmentation. Our method adds additional skip connections compared to ResNet but uses significantly fewer model parameters than DenseNet. We evaluate the proposed method on a public dataset (ISIC 2018 grand-challenge) for skin lesion segmentation and a local brain MRI dataset. In comparison with ResNet-based, DenseNet-based and attention network (AttnNet) based methods within the same encoder-decoder network structure, our method achieves significantly higher segmentation accuracy with fewer number of model parameters than DenseNet and AttnNet. The code is available on GitHub (GitHub link: https://github.com/MinaJf/DRU-net). | ['Jonathan Garibaldi', 'Susan Francis', 'Mina Jafari', 'Dorothee Auer', 'Xin Chen'] | 2020-04-28 | null | null | null | null | ['skin-lesion-segmentation'] | ['medical'] | [ 2.26693064e-01 5.49697101e-01 -1.49590239e-01 -4.18358862e-01
-4.86378163e-01 -9.93189588e-02 2.26869613e-01 -1.02030382e-01
-6.66299939e-01 6.34494305e-01 2.87480384e-01 -4.04745817e-01
1.27385706e-01 -6.66831911e-01 -5.66696823e-01 -3.61737818e-01
8.29659700e-02 2.11919650e-01 3.74938935e-01 4.52302694e-02
-5.74464314e-02 3.28927279e-01 -7.39296377e-01 2.34237239e-01
1.03458035e+00 1.00820136e+00 4.97653574e-01 5.33861935e-01
4.75459592e-03 1.21554458e+00 -2.13994995e-01 -3.31603795e-01
1.34542540e-01 -3.80978018e-01 -1.31094956e+00 -3.92851472e-01
3.75287980e-01 -2.91619986e-01 -6.52943909e-01 1.21548545e+00
7.56731808e-01 -2.24277400e-03 3.86683673e-01 -7.63046801e-01
-7.36222267e-01 8.89754176e-01 -2.35832438e-01 5.46249688e-01
-4.17040020e-01 4.38876823e-02 5.74936926e-01 -5.77898800e-01
5.35273552e-01 9.35171604e-01 9.32670534e-01 9.23478782e-01
-8.66459608e-01 -6.84933603e-01 7.68955871e-02 3.79369408e-01
-1.37369835e+00 -2.77082860e-01 3.70226055e-01 -2.33474061e-01
1.00205112e+00 1.57085076e-01 7.74279535e-01 1.07037818e+00
2.33669624e-01 9.85385120e-01 8.90321553e-01 -7.90235326e-02
8.28843713e-02 -2.44094819e-01 4.58393216e-01 7.31174648e-01
2.65657961e-01 -1.01333082e-01 2.34628886e-01 2.66896576e-01
1.01521146e+00 2.45395824e-01 -4.02753323e-01 1.99707761e-01
-1.01624238e+00 8.19868326e-01 1.23389733e+00 7.05179632e-01
-5.20961046e-01 5.86685717e-01 5.67569792e-01 -6.04021885e-02
6.23435140e-01 2.27722600e-01 -4.50415909e-01 2.95748003e-02
-9.60278511e-01 -1.30222246e-01 5.66172779e-01 8.20630908e-01
4.29989249e-01 3.54560524e-01 -4.61147189e-01 1.23329377e+00
8.52738619e-02 9.86846834e-02 8.99130762e-01 -6.40466452e-01
1.97328135e-01 5.80478787e-01 -6.25099003e-01 -5.87268353e-01
-6.98973417e-01 -6.85518682e-01 -1.26572037e+00 -2.11859509e-01
6.71054944e-02 -4.60724980e-01 -1.48466575e+00 1.48940575e+00
1.74207818e-02 4.58350033e-01 -1.92765728e-01 1.00127769e+00
1.31148779e+00 3.77179563e-01 3.74897569e-01 3.31900209e-01
1.42053246e+00 -1.36593151e+00 -6.21271133e-01 -3.16543281e-01
8.25908482e-01 -6.20835721e-01 6.48955762e-01 1.11419164e-01
-1.20386803e+00 -4.65614289e-01 -8.86878133e-01 -3.61326635e-01
-3.19110274e-01 2.11858854e-01 8.01808417e-01 5.33795536e-01
-1.50284088e+00 6.29840612e-01 -1.18612683e+00 -4.00211155e-01
1.06372976e+00 6.71309352e-01 -1.61426052e-01 -1.92188069e-01
-1.13440597e+00 1.05367804e+00 6.17267132e-01 3.61483932e-01
-1.02867842e+00 -9.94771302e-01 -8.40392947e-01 8.03669542e-02
2.75989950e-01 -6.19082510e-01 1.40467155e+00 -1.14070022e+00
-1.38563228e+00 7.60671794e-01 1.62217557e-01 -9.67453241e-01
3.64574343e-01 -2.32973352e-01 -1.03235111e-01 4.69037265e-01
-8.09097141e-02 1.20809019e+00 3.03018481e-01 -6.23981237e-01
-3.03767592e-01 -1.88524071e-02 1.01005383e-01 -2.44243927e-02
-2.92760860e-02 8.78157094e-03 -4.80683088e-01 -7.63961375e-01
2.36152802e-02 -7.90047169e-01 -5.87046564e-01 -1.62089795e-01
-7.16555417e-01 -2.06352204e-01 8.04738343e-01 -8.39910507e-01
1.03465784e+00 -1.91419482e+00 -1.15946587e-02 -1.40205815e-01
5.98116279e-01 7.74213135e-01 -3.10082734e-01 -6.93071187e-02
-4.93971109e-01 3.06545943e-01 -5.96770883e-01 -2.91238010e-01
-4.21803445e-01 1.12120844e-01 4.00894523e-01 5.63279927e-01
5.07856719e-02 1.31473291e+00 -6.06415689e-01 -4.35128838e-01
4.35895562e-01 7.74556160e-01 -7.05646813e-01 -7.81129748e-02
-1.36295587e-01 5.03290176e-01 -3.15989226e-01 5.64559221e-01
8.38416219e-01 -3.65834653e-01 -4.21406282e-03 -4.12606061e-01
1.11416824e-01 1.97008714e-01 -5.56458414e-01 1.82593012e+00
-6.23716950e-01 7.73567438e-01 7.89369494e-02 -1.31735241e+00
7.09401906e-01 5.21846175e-01 6.51310086e-01 -8.88681829e-01
6.71301007e-01 1.34048134e-01 3.23690265e-01 -4.57694560e-01
1.33549631e-01 -4.74526621e-02 1.84606835e-01 5.54937050e-02
5.69847167e-01 3.36068928e-01 2.26263016e-01 3.04566950e-01
1.32097280e+00 -1.50214151e-01 1.32517859e-01 -4.90924865e-01
6.28321946e-01 -1.33730575e-01 6.58810318e-01 5.49376190e-01
-2.98673660e-01 7.68180788e-01 5.27917325e-01 -3.63071352e-01
-1.03620338e+00 -7.06639767e-01 -4.53478277e-01 5.57105899e-01
4.65268604e-02 -4.47293550e-01 -1.07547736e+00 -6.10219836e-01
-3.54203761e-01 5.40674984e-01 -8.01954448e-01 -9.59706679e-03
-7.38615334e-01 -8.02358389e-01 8.64570856e-01 6.99340940e-01
1.05306947e+00 -1.33070099e+00 -4.75104362e-01 1.50423035e-01
-2.10097209e-01 -1.08247423e+00 -5.34891903e-01 1.15272507e-01
-9.31869566e-01 -1.18779242e+00 -1.06173921e+00 -1.04445446e+00
9.99703705e-01 -1.40820026e-01 9.58417118e-01 3.70234013e-01
-7.73405254e-01 1.76529467e-01 -4.17759061e-01 -2.17168674e-01
-1.56523123e-01 5.44922173e-01 -6.35518849e-01 -2.18077913e-01
3.40766281e-01 -4.55953121e-01 -1.04563439e+00 1.12045072e-01
-9.68021870e-01 4.46099281e-01 5.82084298e-01 8.22806060e-01
4.59497124e-01 -3.24018210e-01 7.42817104e-01 -1.25971603e+00
4.99863684e-01 -6.93161368e-01 -4.89947528e-01 -8.37355331e-02
-5.59602797e-01 -2.57096976e-01 5.73410749e-01 -1.91158593e-01
-8.82042646e-01 1.47751346e-02 -7.88517416e-01 -3.96757424e-01
-2.14943469e-01 4.48607951e-01 2.80980319e-01 -1.11230515e-01
4.21963751e-01 6.80537522e-02 3.72234024e-02 -8.18835974e-01
2.01960772e-01 5.94031274e-01 4.69667971e-01 -8.07764679e-02
2.02520788e-01 2.47984409e-01 -2.97531068e-01 -3.09493214e-01
-7.74980485e-01 -2.96333671e-01 -7.75093317e-01 -1.05112076e-01
1.28369057e+00 -9.47998464e-01 -5.92365801e-01 6.27300382e-01
-1.02842486e+00 -7.27966607e-01 -3.43273848e-01 6.26648009e-01
-2.29431778e-01 6.99251071e-02 -9.30637121e-01 -1.68168515e-01
-8.43669415e-01 -1.34617245e+00 5.85535467e-01 4.51293588e-01
1.25963651e-02 -1.33256233e+00 -1.87859178e-01 2.94519484e-01
9.13172960e-01 4.47086841e-01 7.67871976e-01 -7.52382636e-01
-3.56906205e-01 -2.68898845e-01 -5.94648540e-01 6.34652853e-01
1.39524996e-01 -5.03481150e-01 -8.15613985e-01 -3.37082386e-01
-1.28951222e-01 -2.51518548e-01 1.46267426e+00 9.41280603e-01
1.68585932e+00 -3.30486983e-01 -3.16699207e-01 9.72460270e-01
1.73476279e+00 1.82038143e-01 9.88045990e-01 2.06839859e-01
1.00736499e+00 2.46604886e-02 -2.84446359e-01 -2.71579828e-02
3.35714012e-01 2.80559182e-01 5.21677136e-01 -6.77700520e-01
-7.64678657e-01 2.21975118e-01 3.40104103e-03 8.64923477e-01
-2.32374087e-01 -1.60429299e-01 -1.12057626e+00 8.27732623e-01
-1.71549964e+00 -7.16521442e-01 -2.42210612e-01 1.69711757e+00
8.68785381e-01 -9.26833004e-02 5.01830094e-02 -1.84970111e-01
9.23917055e-01 2.00579047e-01 -6.66274667e-01 -4.97197926e-01
1.97231025e-01 7.19794214e-01 9.44288313e-01 4.87159133e-01
-1.17610502e+00 1.03799129e+00 5.66430044e+00 9.95823443e-01
-1.40980744e+00 7.88507640e-01 9.10183489e-01 -1.93312481e-01
7.43798465e-02 -3.56790870e-01 -6.12309158e-01 4.68881339e-01
1.09342051e+00 6.66640773e-02 2.38643453e-01 8.42089653e-01
-1.01046003e-01 -3.84267233e-02 -7.63323307e-01 7.65794218e-01
-2.60585379e-02 -1.66015065e+00 -3.52212340e-01 -1.20656349e-01
6.97748005e-01 7.79261529e-01 1.03579514e-01 3.94353479e-01
3.78691971e-01 -1.35898042e+00 3.33617002e-01 2.32332110e-01
1.07130086e+00 -6.34641945e-01 1.04629970e+00 -4.03328761e-02
-9.80908453e-01 7.59838447e-02 -5.67747891e-01 2.41038606e-01
6.00122251e-02 5.91617167e-01 -7.16988623e-01 4.27731782e-01
7.05439031e-01 1.12324786e+00 -7.32757688e-01 1.42517626e+00
-1.61363482e-01 9.85784352e-01 -6.33652359e-02 2.01948926e-01
5.67355931e-01 1.08073622e-01 3.01196605e-01 1.35573614e+00
-1.64190680e-02 1.54074714e-01 -2.10067574e-02 1.00648344e+00
-4.79093522e-01 -1.55123360e-02 -2.04377532e-01 2.42453143e-01
1.60371840e-01 1.54903781e+00 -1.00228369e+00 -3.43090147e-01
-3.64189535e-01 8.70093524e-01 3.51586014e-01 2.38735661e-01
-9.73684669e-01 -5.12569726e-01 2.07249224e-01 1.10719688e-01
3.63121301e-01 9.01628956e-02 -3.84838521e-01 -1.05546558e+00
-4.32111889e-01 -6.10310435e-01 2.92602330e-01 -5.87824285e-01
-1.10303986e+00 1.10628390e+00 -9.90036204e-02 -9.82165694e-01
2.03678995e-01 -5.56314766e-01 -7.11712241e-01 7.97237515e-01
-1.70868850e+00 -1.10509241e+00 -5.08498788e-01 7.36602902e-01
6.67880416e-01 -1.20716952e-01 7.23604798e-01 7.06631958e-01
-8.56713057e-01 4.76436704e-01 3.37025709e-02 7.10537314e-01
3.10660839e-01 -1.24888790e+00 5.15598774e-01 7.56577492e-01
-4.20769811e-01 4.68026519e-01 -1.09573364e-01 -6.61468863e-01
-8.45310330e-01 -1.64547455e+00 4.89381641e-01 1.41586745e-02
4.45738167e-01 -3.28458428e-01 -8.23919713e-01 9.24347043e-01
6.41961932e-01 3.32136452e-01 5.48207998e-01 -4.94359583e-02
-3.09851486e-02 9.07912925e-02 -1.41404092e+00 3.34717751e-01
8.67020249e-01 -1.40951678e-01 -2.74541020e-01 5.99347174e-01
9.32675302e-01 -6.98765934e-01 -9.34072971e-01 4.89444494e-01
2.29286820e-01 -6.56858206e-01 8.33711982e-01 -2.83497959e-01
6.84223294e-01 6.07272536e-02 1.48059100e-01 -1.21768701e+00
-5.75555205e-01 -2.93354779e-01 2.01045498e-01 7.81825900e-01
4.36004996e-01 -8.64435434e-01 6.73231006e-01 4.25442278e-01
-6.65722549e-01 -1.26005471e+00 -9.70568299e-01 -5.44675291e-01
3.86404455e-01 -2.14886278e-01 3.86856705e-01 7.77976274e-01
-3.29333007e-01 1.29246324e-01 -3.03953856e-01 -1.72235921e-01
5.27508438e-01 -5.67384064e-01 -1.07443020e-01 -9.59577262e-01
9.78655219e-02 -5.21447420e-01 -5.32392561e-01 -7.64499068e-01
3.77869494e-02 -1.50277925e+00 -2.43591785e-01 -2.05717468e+00
2.86818206e-01 -4.81197625e-01 -6.15978539e-01 1.02079690e+00
1.19442716e-01 7.53031194e-01 3.84527981e-01 1.28758594e-01
-6.66899920e-01 2.44312346e-01 1.46685994e+00 -2.23762766e-01
-1.41937872e-02 -1.85007572e-01 -7.43535995e-01 7.41117358e-01
1.14169657e+00 -5.34539700e-01 -2.73968756e-01 -7.74136066e-01
-3.45156878e-01 -7.70807564e-02 5.87515712e-01 -1.14486206e+00
3.50282222e-01 3.63134354e-01 6.43818855e-01 -4.45094287e-01
1.02924146e-02 -6.30251706e-01 -2.36194916e-02 9.41748083e-01
-5.15944600e-01 -1.67988241e-01 4.49225307e-01 4.96305972e-02
-1.98055178e-01 -2.85353869e-01 1.18773401e+00 -3.61928463e-01
-6.15930498e-01 6.27841830e-01 -4.99568403e-01 4.08176407e-02
8.94320130e-01 -7.39031062e-02 -5.68824053e-01 -2.40635797e-02
-1.15492582e+00 2.97731578e-01 -1.63524792e-01 3.95774215e-01
6.90152228e-01 -1.26952684e+00 -7.33644545e-01 5.19150011e-02
-4.63214695e-01 2.58289665e-01 5.80818355e-01 1.34106970e+00
-9.85588193e-01 7.91299582e-01 -4.62117463e-01 -4.77649957e-01
-1.00722396e+00 1.05863929e-01 8.74745369e-01 -3.50931883e-01
-9.73411322e-01 1.24292982e+00 3.59530061e-01 -5.21973670e-01
3.19611609e-01 -4.99908745e-01 -2.54795700e-01 -3.95082355e-01
4.18371141e-01 3.12595636e-01 1.27224788e-01 -4.63845491e-01
-4.51656133e-01 2.21828803e-01 -5.49447000e-01 4.85921621e-01
1.56924009e+00 1.97548941e-01 -2.88903415e-01 -2.59829849e-01
1.46363044e+00 -7.87904024e-01 -1.04098046e+00 -2.50834942e-01
-2.00399294e-01 -9.22517031e-02 7.20831931e-01 -1.15865505e+00
-1.88849223e+00 8.53856266e-01 1.01217592e+00 -1.61890417e-01
1.24430740e+00 9.30406302e-02 1.14373136e+00 -3.00561693e-02
-3.13443914e-02 -9.00792181e-01 -1.15970440e-01 4.28008646e-01
8.64014626e-01 -1.13342369e+00 -1.37131006e-01 -3.58544677e-01
-6.34002328e-01 9.77243900e-01 9.18105423e-01 -5.62594533e-01
9.77417767e-01 2.95464486e-01 -1.25620037e-01 -4.09123868e-01
-5.59766948e-01 -3.70299488e-01 1.72322750e-01 4.66676772e-01
6.87401712e-01 -1.13128200e-02 -3.97015244e-01 5.91397762e-01
-9.17855054e-02 3.44171911e-01 5.09998381e-01 6.73773348e-01
-2.21569255e-01 -9.44949508e-01 1.93643704e-01 8.53814840e-01
-9.04679716e-01 -4.99889195e-01 8.71739164e-03 6.87085629e-01
3.29307437e-01 5.96005976e-01 3.36774439e-01 -1.49535313e-01
5.93340844e-02 -2.47422025e-01 4.90989864e-01 -7.61036098e-01
-9.34240520e-01 9.29101706e-02 3.52176689e-02 -5.80000758e-01
-4.11048949e-01 -2.72628516e-01 -1.42178583e+00 -3.58760476e-01
-3.19571525e-01 -2.63647549e-02 7.23848045e-01 7.65132070e-01
4.15611744e-01 1.13084030e+00 2.88645864e-01 -6.59060001e-01
5.51270740e-03 -1.20954084e+00 -5.19283593e-01 3.78317907e-02
2.12959066e-01 -3.57722789e-01 5.03500886e-02 -1.52975721e-02] | [14.68786334991455, -2.588212251663208] |
a2c3330e-7136-4965-b5bb-643be4b2475f | foga-flag-optimization-with-genetic-algorithm | 2105.07202 | null | https://arxiv.org/abs/2105.07202v1 | https://arxiv.org/pdf/2105.07202v1.pdf | FOGA: Flag Optimization with Genetic Algorithm | Recently, program autotuning has become very popular especially in embedded systems, when we have limited resources such as computing power and memory where these systems run generally time-critical applications. Compiler optimization space gradually expands with the renewed compiler options and inclusion of new architectures. These advancements bring autotuning even more important position. In this paper, we introduced Flag Optimization with Genetic Algorithm (FOGA) as an autotuning solution for GCC flag optimization. FOGA has two main advantages over the other autotuning approaches: the first one is the hyperparameter tuning of the genetic algorithm (GA), the second one is the maximum iteration parameter to stop when no further improvement occurs. We demonstrated remarkable speedup in the execution time of C++ source codes with the help of optimization flags provided by FOGA when compared to the state of the art framework OpenTuner. | ['Mahiye Uluyağmur Öztürk', 'Mert Kutay Sezer', 'Berkan Höke', 'Burak Tağtekin'] | 2021-05-15 | null | null | null | null | ['compiler-optimization'] | ['computer-code'] | [-2.13286549e-01 -2.90172458e-01 -2.79234231e-01 -5.92073239e-02
7.19344392e-02 -6.25630975e-01 2.71045655e-01 2.82717824e-01
-3.63239795e-01 8.17417622e-01 -6.78799823e-02 -6.91640019e-01
3.66291441e-02 -9.77965295e-01 -3.94973278e-01 -6.73520505e-01
-3.52443568e-02 2.89280385e-01 3.16858202e-01 -5.71247280e-01
4.53074187e-01 1.54833123e-01 -2.00136542e+00 -1.87267497e-01
1.38583195e+00 6.38248324e-01 2.84023046e-01 7.29407370e-01
-2.66318142e-01 6.21253960e-02 -4.56573308e-01 8.27319454e-03
2.70728767e-01 -2.68030375e-01 -4.53531653e-01 -2.11132586e-01
-3.03779781e-01 2.16400489e-01 2.14528352e-01 1.09101892e+00
4.40714091e-01 6.51583895e-02 5.24661620e-04 -1.14559686e+00
-2.01425537e-01 8.07959795e-01 -5.87195337e-01 2.80801296e-01
-6.49939477e-02 5.57614982e-01 4.87901837e-01 -3.48637730e-01
3.60364377e-01 1.13448858e+00 2.98794329e-01 2.83405900e-01
-9.57132638e-01 -4.59231824e-01 -2.11822420e-01 1.65146679e-01
-1.47905231e+00 -6.91291243e-02 3.01324993e-01 -4.13537830e-01
1.33833313e+00 5.06403267e-01 6.96143985e-01 4.24827546e-01
6.28648818e-01 6.51607588e-02 1.06615365e+00 -8.04913282e-01
5.96796691e-01 2.39563078e-01 2.49565214e-01 8.08431923e-01
3.89907151e-01 4.98022467e-01 -1.69581890e-01 -2.16427028e-01
4.94512558e-01 -4.71387327e-01 -1.79938599e-01 -2.02186108e-01
-1.38317907e+00 8.03492665e-01 1.99921295e-01 4.21980679e-01
-7.43615776e-02 3.66787225e-01 7.75151730e-01 3.51013750e-01
8.79734103e-03 7.70867229e-01 -6.55220509e-01 -7.21764147e-01
-9.61055875e-01 1.80563584e-01 7.22408831e-01 7.99699366e-01
8.97146642e-01 6.03133738e-01 -1.10062726e-01 4.86843050e-01
1.00162186e-01 4.51592952e-01 1.03253770e+00 -4.85780954e-01
1.23966649e-01 1.11355424e+00 -1.84086218e-01 -7.23786235e-01
-5.81003785e-01 -8.51314485e-01 -7.21531987e-01 5.57028770e-01
1.05743282e-01 -4.40459341e-01 -8.21077168e-01 1.37156177e+00
4.18396413e-01 -1.39399618e-01 8.94535482e-02 7.39579797e-01
3.38348776e-01 6.77683592e-01 -7.64904097e-02 -3.30092013e-01
1.47405267e+00 -1.02209902e+00 -6.61893785e-01 1.73117649e-02
7.69196570e-01 -1.15144372e+00 1.16081345e+00 3.32685739e-01
-6.28683627e-01 -6.77656353e-01 -1.51559329e+00 4.50488508e-01
-6.44471705e-01 1.74071118e-01 9.95251477e-01 1.26777327e+00
-1.13576245e+00 5.27742207e-01 -9.17778492e-01 -3.21242481e-01
-1.46440476e-01 8.15106034e-01 8.62175003e-02 5.12146056e-01
-8.10739100e-01 6.17808282e-01 9.78720367e-01 -2.74562985e-01
-4.63912457e-01 -8.29993486e-01 -5.74540555e-01 2.22497582e-01
5.36331177e-01 -6.72214448e-01 8.85735214e-01 -1.11147332e+00
-1.96617067e+00 6.79265141e-01 1.49702519e-01 -5.91232359e-01
3.93433809e-01 -5.64537644e-02 -6.61721826e-01 -6.66795909e-01
-3.02798301e-01 4.16221231e-01 9.20202911e-01 -6.90012038e-01
-5.16154170e-01 -1.49457693e-01 -1.36147261e-01 -8.39517862e-02
-4.41564023e-01 -2.38235388e-02 -2.61991531e-01 -5.76741338e-01
-4.51632559e-01 -1.16242826e+00 -2.78794885e-01 -6.89560831e-01
-1.52714655e-01 1.07626766e-01 1.11139810e+00 -2.02899694e-01
1.76957536e+00 -2.17735481e+00 3.35173994e-01 2.27350742e-01
-1.16990216e-01 6.05238855e-01 2.33107805e-01 1.73294842e-01
-2.82708079e-01 4.05346900e-02 -1.41685933e-01 4.24930423e-01
1.38479263e-01 2.16652513e-01 1.89500693e-02 1.99738756e-01
-2.27465436e-01 6.04954481e-01 -8.27431560e-01 -4.92699653e-01
1.93722650e-01 1.43804446e-01 -8.19166362e-01 4.52320389e-02
-5.52930772e-01 4.31042820e-01 -3.30614239e-01 6.96678579e-01
6.82536662e-01 2.90906020e-02 1.36100501e-01 -5.47256581e-02
-8.65628004e-01 -1.53381020e-01 -1.18009984e+00 1.51783264e+00
-5.24668217e-01 4.48154986e-01 -5.14332093e-02 -6.02325201e-01
1.14326239e+00 1.86770990e-01 8.22638795e-02 -6.01892531e-01
5.34286201e-01 4.50488329e-01 5.00133514e-01 -3.66115808e-01
9.14509892e-01 3.44885796e-01 -2.31282935e-01 4.37530220e-01
-2.04275295e-01 -1.90229163e-01 6.22875988e-01 -1.03732288e-01
9.44770694e-01 2.54024953e-01 6.26666546e-01 -9.65473056e-01
9.22588825e-01 2.56174296e-01 6.03276670e-01 2.64273912e-01
-1.91120040e-02 7.00048655e-02 6.60679996e-01 -4.32917178e-01
-1.13268006e+00 -4.80784744e-01 -2.63942510e-01 1.07181168e+00
-2.94459611e-01 -7.11918533e-01 -9.40534115e-01 -4.76188689e-01
-1.39559597e-01 8.02623510e-01 -5.64457119e-01 -2.86198199e-01
-7.86334217e-01 -1.16464901e+00 2.76278943e-01 1.83906466e-01
7.84823835e-01 -8.29706848e-01 -1.18486416e+00 1.42011225e-01
6.00741744e-01 -5.77209353e-01 -4.19291049e-01 5.38157284e-01
-1.11411178e+00 -7.35938311e-01 -2.36435115e-01 -5.58347821e-01
5.23565352e-01 -1.13504939e-01 1.24066055e+00 4.81982142e-01
-6.31471813e-01 -8.97504389e-02 -4.75902140e-01 -3.82131457e-01
-8.44857872e-01 4.57111627e-01 -8.46508294e-02 -4.18552160e-01
2.16636397e-02 -5.94131351e-01 -3.98235291e-01 3.66521239e-01
-8.46218109e-01 5.53081743e-02 5.52734196e-01 8.69573891e-01
4.68646497e-01 4.18854266e-01 4.31535840e-01 -1.02417815e+00
5.59768498e-01 -3.88889164e-01 -1.41659224e+00 1.75139353e-01
-1.01311696e+00 7.44382620e-01 8.07360291e-01 -1.52291253e-01
-9.94451642e-01 -7.40696862e-03 -1.28235236e-01 -1.05639637e-01
2.03176826e-01 3.95410746e-01 -2.68152326e-01 -4.32569176e-01
7.59284079e-01 -6.75141066e-02 1.32006332e-01 -2.77219355e-01
5.77468127e-02 5.67395747e-01 4.59404856e-01 -7.48450696e-01
7.13597476e-01 -1.58769324e-01 1.87602594e-01 -5.73423088e-01
9.71195623e-02 -1.70698911e-01 -3.54553789e-01 -1.41483098e-01
5.46605349e-01 -3.83801639e-01 -8.88734996e-01 3.11527759e-01
-7.32286632e-01 -4.06269491e-01 -1.09243423e-01 2.42837638e-01
-2.74116546e-01 2.93178052e-01 6.73700944e-02 -4.86116052e-01
-5.77586055e-01 -1.70420980e+00 6.07345343e-01 7.14329243e-01
-2.13932529e-01 -9.06246960e-01 1.02251656e-01 -1.67555794e-01
8.96418452e-01 4.46030617e-01 1.26665711e+00 -1.39841631e-01
-5.37696779e-01 4.04181294e-02 7.68770948e-02 4.64470498e-02
8.14477205e-02 3.79901260e-01 -5.16745448e-01 -3.68610173e-01
-1.17734015e-01 3.29848975e-01 3.66388917e-01 2.49768361e-01
1.07049763e+00 -1.70195028e-01 -4.37717110e-01 1.00102806e+00
1.82105410e+00 4.40419763e-01 5.71799040e-01 7.99246430e-01
3.76099050e-01 -1.12913840e-01 9.86887097e-01 7.91750312e-01
-3.28486860e-01 9.40636218e-01 6.83273137e-01 1.40089065e-01
-2.63453983e-02 2.18530536e-01 2.66239345e-01 9.19908106e-01
-1.16068803e-01 -1.20035358e-01 -1.14906526e+00 3.24020594e-01
-1.80401564e+00 -2.25149572e-01 -2.98476875e-01 2.63650942e+00
7.88070083e-01 2.65818059e-01 1.61882609e-01 1.47124782e-01
6.29253685e-01 -2.21298367e-01 -5.51663399e-01 -9.42002475e-01
-1.98618155e-02 4.98150378e-01 8.56041253e-01 4.86010373e-01
-8.21858704e-01 1.00282955e+00 5.67835760e+00 8.42781901e-01
-1.46079528e+00 1.10051341e-01 3.22400898e-01 -4.08564731e-02
-1.52877405e-01 2.89768636e-01 -9.13923383e-01 7.80038595e-01
1.33722281e+00 -5.68939984e-01 4.21497107e-01 1.05407727e+00
1.48866788e-01 -4.40556169e-01 -8.39703202e-01 8.10001194e-01
-2.47860298e-01 -1.31769443e+00 -2.97441870e-01 1.01916924e-01
7.78586507e-01 -2.21719563e-01 1.29071683e-01 3.26474875e-01
2.96274066e-01 -8.98265779e-01 6.03476822e-01 1.62632130e-02
7.36436248e-01 -1.19353676e+00 1.00229621e+00 1.33799054e-02
-8.79785895e-01 -1.10689431e-01 -3.84054929e-01 -7.81500414e-02
-7.73198754e-02 6.25951588e-01 -9.65517521e-01 4.33641553e-01
7.38147616e-01 8.50896984e-02 -8.68460953e-01 1.28078783e+00
3.29938345e-02 6.53560221e-01 -1.81023806e-01 -3.59761029e-01
1.36945248e-01 -3.41397852e-01 7.55319655e-01 1.07014298e+00
4.49163020e-01 -2.64695808e-02 -7.76346177e-02 7.43823230e-01
3.47176552e-01 2.73204476e-01 1.49636082e-02 -5.05916588e-02
4.05366540e-01 1.46302819e+00 -1.08980811e+00 -2.55866647e-01
-5.16108349e-02 6.25205278e-01 -1.33034738e-03 -1.21010043e-01
-1.09421670e+00 -4.57550347e-01 7.15508580e-01 1.31515220e-01
3.13347369e-01 -5.08110344e-01 -5.06846607e-01 -7.16951966e-01
-3.54470909e-01 -1.07777560e+00 2.94468522e-01 -2.40487754e-01
-2.05918908e-01 7.59352684e-01 -1.01855464e-01 -1.14070046e+00
-2.95042843e-01 -7.19710350e-01 -5.43996155e-01 6.06263697e-01
-1.16744447e+00 -5.68151951e-01 -3.79023582e-01 2.20448986e-01
4.60658967e-01 -4.79351759e-01 9.41681504e-01 3.63878191e-01
-7.56232500e-01 9.56623495e-01 4.28615898e-01 -5.49877048e-01
7.13585198e-01 -1.20316851e+00 1.44766182e-01 1.07728922e+00
-6.25407398e-01 7.16952384e-01 1.19303393e+00 -5.22312462e-01
-1.80176640e+00 -7.56876945e-01 3.04644942e-01 2.45546564e-01
6.81258500e-01 -4.20538008e-01 -7.51719058e-01 3.02894175e-01
5.21074235e-01 -1.84108198e-01 6.16958618e-01 8.74543414e-02
1.10680051e-02 -2.53220886e-01 -9.28584874e-01 5.72538793e-01
3.90585780e-01 3.36939335e-01 6.75129816e-02 2.80230343e-01
8.75115097e-01 -8.36583555e-01 -8.58889818e-01 2.07745954e-01
2.54919112e-01 -1.20308554e+00 6.07400239e-01 -1.03570540e-02
1.29384086e-01 -6.47760272e-01 -2.11284578e-01 -1.37164998e+00
-2.21109420e-01 -1.01158428e+00 4.03421111e-02 1.22256982e+00
3.98759663e-01 -8.61063302e-01 5.56075275e-01 4.13869768e-01
-5.40805042e-01 -6.57499373e-01 -7.52702773e-01 -7.74332523e-01
-1.19707406e-01 1.32972389e-01 1.15844631e+00 6.79253161e-01
2.54262239e-01 1.95244402e-01 -1.54875830e-01 -1.10931881e-01
3.06229711e-01 2.88845867e-01 7.03306854e-01 -8.80831599e-01
-8.73923719e-01 -5.13879180e-01 -4.80471343e-01 -1.89784199e-01
-1.79506868e-01 -7.64988542e-01 -2.47642577e-01 -4.79617029e-01
-5.97128039e-03 -6.40947700e-01 -1.02223307e-01 5.16100943e-01
-2.85009712e-01 -2.36370966e-01 9.20098051e-02 -2.02670708e-01
-3.47094387e-01 2.89884925e-01 9.36465681e-01 1.39209300e-01
-5.23645222e-01 1.48576275e-02 -4.97053862e-01 3.95237684e-01
9.44164038e-01 -2.43108884e-01 -2.65556723e-01 -2.82079011e-01
5.24134398e-01 -8.60336423e-03 -1.39231265e-01 -1.32228303e+00
6.64214194e-02 -4.66052182e-02 -1.15681060e-01 -3.07913601e-01
-3.38167161e-01 -7.66117156e-01 4.77830261e-01 9.61899936e-01
2.67372787e-01 6.37480319e-01 7.90535629e-01 2.86064744e-01
-1.97198644e-01 -3.99184674e-01 1.02646208e+00 5.38187735e-02
-1.10296857e+00 -2.64113337e-01 -5.39315820e-01 -2.82003194e-01
1.26279664e+00 -3.80655289e-01 -3.09375942e-01 3.70084941e-01
-4.48127002e-01 4.86781225e-02 8.27732146e-01 4.76245821e-01
3.86617333e-02 -9.56041753e-01 -4.84795570e-01 5.11408508e-01
-2.28163823e-02 -3.00643563e-01 1.74954325e-01 8.31906796e-01
-9.81198072e-01 6.64047480e-01 -6.50427580e-01 -5.03602564e-01
-1.57766986e+00 8.87453854e-01 1.87387019e-01 -4.97801602e-01
-3.36517930e-01 6.11077130e-01 -7.72899017e-02 -2.08804443e-01
-2.67158717e-01 -4.08166051e-01 -6.31973892e-02 -3.09641898e-01
3.54301989e-01 6.05400622e-01 5.10585427e-01 -3.09539527e-01
-5.05822420e-01 4.32666570e-01 -4.04387452e-02 1.84313700e-01
1.17237902e+00 1.68874577e-01 -6.22079134e-01 1.40885085e-01
9.05647933e-01 1.42902225e-01 -7.06803203e-01 2.60452420e-01
-3.48639823e-02 -5.41122019e-01 5.48069298e-01 -8.46004486e-01
-1.13536835e+00 3.92038971e-01 9.75544214e-01 3.96981798e-02
1.33233428e+00 -6.36983693e-01 5.30869126e-01 -3.17582712e-02
5.79666555e-01 -1.12266636e+00 -2.52193809e-01 6.38187826e-01
3.97592157e-01 -8.99199665e-01 1.54230729e-01 -3.76189411e-01
-2.00290084e-01 1.37014532e+00 8.08583498e-01 -9.50572416e-02
4.11042035e-01 7.93880999e-01 -2.79566735e-01 -1.03123607e-02
-6.78569078e-01 5.20891510e-02 -2.77718753e-02 4.23604757e-01
6.31959498e-01 2.77235657e-01 -8.70998859e-01 3.41566205e-01
-6.21237218e-01 -5.07118516e-02 6.33209825e-01 1.00538003e+00
-5.60821354e-01 -1.66671276e+00 -7.93326497e-01 2.96381384e-01
-4.23971146e-01 -1.28773734e-01 1.30447581e-01 9.57069337e-01
3.05729061e-01 6.66603804e-01 -1.09837666e-01 -3.62667322e-01
1.51838884e-02 -1.14570722e-01 4.52998757e-01 -4.57392126e-01
-1.09235477e+00 -1.00743718e-01 -9.05945972e-02 -5.01461744e-01
1.77365094e-01 -7.02916145e-01 -1.30019534e+00 -5.11857092e-01
-4.84285831e-01 5.00613689e-01 1.08373201e+00 6.19624257e-01
7.17976093e-01 1.18031883e+00 3.39243174e-01 -5.06615520e-01
-3.43312263e-01 -5.42453468e-01 -3.64011258e-01 -3.88919413e-01
-9.27396417e-02 -7.61822522e-01 -2.09622663e-02 -2.12089345e-03] | [5.901954174041748, 3.5982601642608643] |
79e83de1-5603-4ca9-bfcf-9b88fd984692 | lever-learning-to-verify-language-to-code | 2302.08468 | null | https://arxiv.org/abs/2302.08468v2 | https://arxiv.org/pdf/2302.08468v2.pdf | LEVER: Learning to Verify Language-to-Code Generation with Execution | The advent of large language models trained on code (code LLMs) has led to significant progress in language-to-code generation. State-of-the-art approaches in this area combine LLM decoding with sample pruning and reranking using test cases or heuristics based on the execution results. However, it is challenging to obtain test cases for many real-world language-to-code applications, and heuristics cannot well capture the semantic features of the execution results, such as data type and value range, which often indicates the correctness of the program. In this work, we propose LEVER, a simple approach to improve language-to-code generation by learning to verify the generated programs with their execution results. Specifically, we train verifiers to determine whether a program sampled from the LLMs is correct or not based on the natural language input, the program itself and its execution results. The sampled programs are reranked by combining the verification score with the LLM generation probability, and marginalizing over programs with the same execution results. On four datasets across the domains of table QA, math QA and basic Python programming, LEVER consistently improves over the base code LLMs(4.6% to 10.9% with code-davinci-002) and achieves new state-of-the-art results on all of them. | ['Xi Victoria Lin', 'Sida I. Wang', 'Wen-tau Yih', 'Ves Stoyanov', 'Dragomir Radev', 'Srini Iyer', 'Ansong Ni'] | 2023-02-16 | null | null | null | null | ['text-to-sql', 'semantic-parsing', 'arithmetic-reasoning'] | ['computer-code', 'natural-language-processing', 'reasoning'] | [ 1.52495429e-01 -1.94279134e-01 -7.07844615e-01 -5.19143641e-01
-1.35795236e+00 -7.56384134e-01 3.84216547e-01 5.62912166e-01
1.53953761e-01 3.49280298e-01 -4.18104567e-02 -1.01665759e+00
4.61043358e-01 -8.54973078e-01 -1.32328892e+00 -2.08498165e-02
-2.80991107e-01 2.99299300e-01 2.30709806e-01 9.76845697e-02
6.36000216e-01 -1.97335526e-01 -1.68912923e+00 1.00010026e+00
1.31067574e+00 7.10737765e-01 2.85049558e-01 1.15430415e+00
-4.13025290e-01 1.30041885e+00 -8.64696860e-01 -4.30540234e-01
-1.15736537e-02 -4.30322647e-01 -6.60553813e-01 -3.19595575e-01
4.61112589e-01 -3.92512947e-01 1.92673445e-01 1.22120535e+00
3.80550418e-03 -5.36154926e-01 3.88209850e-01 -1.50963569e+00
-5.37832081e-01 1.20314312e+00 -6.43424451e-01 -3.30569834e-01
7.02775478e-01 3.36082011e-01 1.22323549e+00 -6.77918553e-01
5.98352194e-01 1.22503173e+00 5.71900666e-01 4.28805113e-01
-1.42399108e+00 -3.31102431e-01 -3.21822852e-01 4.69562225e-02
-1.32080865e+00 -3.14738989e-01 1.00040168e-01 -8.78614545e-01
1.35283935e+00 3.88535917e-01 8.35681781e-02 7.08184361e-01
3.65475416e-01 8.42349112e-01 1.01993871e+00 -7.72712052e-01
2.78761148e-01 3.85107517e-01 1.38964549e-01 1.06897557e+00
2.45252073e-01 2.19582822e-02 -3.11296970e-01 -7.09489942e-01
-1.80731833e-01 -3.23516548e-01 2.84841806e-02 -3.55499089e-01
-1.25388205e+00 8.03506494e-01 5.26776910e-02 1.94562227e-01
1.13736987e-01 4.84589636e-01 6.15718365e-01 3.03997278e-01
-7.53081143e-02 6.16870999e-01 -8.55597675e-01 -5.64336300e-01
-1.51359153e+00 5.08687496e-01 1.06914699e+00 1.25774336e+00
9.72069263e-01 -2.54316330e-02 -5.77599943e-01 3.51624936e-01
4.42136884e-01 7.21675932e-01 5.09463549e-01 -6.44348264e-01
9.29523945e-01 1.01376164e+00 -2.20400408e-01 -4.85152513e-01
1.71011358e-01 -1.00534983e-01 -1.02810808e-01 3.85766089e-01
2.80607432e-01 -5.60334697e-02 -7.93637514e-01 1.52208197e+00
-1.75478160e-01 -2.67654985e-01 2.06288964e-01 4.10803795e-01
5.49104571e-01 8.38936150e-01 -2.60417789e-01 2.76138067e-01
1.00520217e+00 -8.48325670e-01 1.03567652e-01 -5.43225229e-01
1.44478917e+00 -8.73386800e-01 1.36875248e+00 3.92868906e-01
-8.82295847e-01 -4.70129520e-01 -1.20153427e+00 2.77228743e-01
1.13927210e-02 6.29038334e-01 7.04523087e-01 5.86217284e-01
-1.14827943e+00 4.97433156e-01 -7.81148136e-01 2.19386548e-01
2.08148316e-01 1.39366001e-01 1.54080214e-02 -2.59817004e-01
-6.94325030e-01 6.08482003e-01 4.11892176e-01 -6.05418682e-01
-1.32711947e+00 -9.40775275e-01 -1.16550922e+00 5.26093319e-02
2.03326315e-01 -1.13369912e-01 1.66052938e+00 -7.88607121e-01
-1.10423994e+00 9.56217825e-01 -4.99681473e-01 -5.66309512e-01
4.23084497e-01 -1.18057773e-01 -3.27684850e-01 -4.39735025e-01
3.47046614e-01 2.97076941e-01 5.66661656e-01 -1.07352543e+00
-7.74183154e-01 1.42032981e-01 3.07506382e-01 -6.10673368e-01
1.05725536e-02 1.99885398e-01 -3.17059457e-01 -1.26836032e-01
-3.49050641e-01 -9.94238913e-01 -4.98807915e-02 -3.67487907e-01
-5.45553207e-01 -1.25286922e-01 4.05656457e-01 -8.45044672e-01
1.58416247e+00 -2.49561262e+00 -9.45206955e-02 2.91019082e-01
5.14747538e-02 9.37300697e-02 -2.46131077e-01 3.25366437e-01
1.41419051e-02 4.16683406e-01 -4.68328923e-01 4.16386761e-02
4.31763947e-01 -2.12064102e-01 -6.92326725e-01 4.23310965e-01
3.88872743e-01 9.27884698e-01 -1.02191985e+00 -5.58761537e-01
-1.61146641e-01 -7.75002614e-02 -1.01707065e+00 5.28826594e-01
-9.96799707e-01 -5.74291162e-02 -1.19011462e-01 7.41732657e-01
4.79057312e-01 -3.29832494e-01 1.59418091e-01 3.68053198e-01
-1.61753744e-01 8.18813920e-01 -9.71476734e-01 1.59303379e+00
-9.16403055e-01 8.87619972e-01 -2.12897763e-01 -5.24607956e-01
8.11482608e-01 1.01686843e-01 -2.56785452e-01 -6.12694561e-01
-3.91157538e-01 6.44582152e-01 3.05220544e-01 -5.80605030e-01
4.87652838e-01 4.24844831e-01 -6.65911436e-01 6.38890445e-01
-9.58985984e-02 -5.17575741e-01 7.34470010e-01 4.16065872e-01
1.35938454e+00 4.38750505e-01 3.40552121e-01 -7.99339116e-02
6.35268450e-01 3.60015243e-01 3.02401781e-01 9.59984839e-01
3.95840764e-01 5.02675891e-01 1.08237040e+00 -5.34249470e-02
-1.13438249e+00 -8.72814298e-01 -8.09575245e-02 1.30144536e+00
-3.79624754e-01 -1.11800361e+00 -9.34369981e-01 -9.31994557e-01
1.02227047e-01 1.30810249e+00 -4.38185841e-01 -3.86673987e-01
-6.37456775e-01 -3.91423792e-01 8.20398569e-01 4.17091459e-01
1.52711138e-01 -8.99890423e-01 -5.96675634e-01 1.67435735e-01
-1.54095396e-01 -8.12448263e-01 -5.73138356e-01 3.16059947e-01
-5.98905563e-01 -1.04398894e+00 -2.04387736e-02 -6.81817234e-01
8.98509920e-01 -2.83173144e-01 1.63123345e+00 4.76609290e-01
-3.18799078e-01 -1.93644583e-01 -3.55136961e-01 -9.71709192e-02
-1.57607687e+00 7.86597431e-02 -5.07149398e-01 -5.81122339e-01
2.02894956e-01 -5.24173155e-02 1.00335702e-01 1.77075509e-02
-7.41853297e-01 1.14253052e-01 7.91138530e-01 6.98502064e-01
4.15840805e-01 1.44090235e-01 -6.87124580e-02 -1.14969611e+00
3.98643196e-01 -5.78722060e-01 -1.16300333e+00 6.16897047e-01
-6.89953208e-01 6.35373175e-01 1.04401910e+00 -4.55293089e-01
-6.42812192e-01 -8.89437199e-02 -1.19013868e-01 -2.37568319e-02
-2.57240161e-02 8.70690525e-01 -1.57192379e-01 1.41432032e-01
1.00059247e+00 5.99024355e-01 -4.28812295e-01 -3.95826660e-02
2.96665728e-01 7.53354073e-01 5.30054688e-01 -1.23445344e+00
7.53997564e-01 -1.72242746e-01 -3.73107761e-01 -3.80169079e-02
-4.61663902e-01 -1.19896665e-01 -1.27004012e-01 2.71914810e-01
3.76892149e-01 -8.26833487e-01 -5.87687969e-01 3.98150265e-01
-1.28483748e+00 -8.21660876e-01 6.22990541e-02 1.40180469e-01
-6.32245064e-01 2.83596605e-01 -5.40593147e-01 -8.42199445e-01
-2.77689815e-01 -1.84447229e+00 1.30613899e+00 -1.35453805e-01
-5.69273889e-01 -5.08003116e-01 -4.18239534e-02 1.02197416e-02
4.98990387e-01 3.69081832e-02 1.67264569e+00 -4.87086564e-01
-9.38934565e-01 -4.92673188e-01 -7.81590790e-02 4.86345619e-01
-2.21410632e-01 5.16836464e-01 -7.44095683e-01 -2.18041107e-01
-2.98726231e-01 -4.10559982e-01 4.83898938e-01 4.44568257e-04
1.21263468e+00 -4.90151018e-01 -3.68021280e-01 6.34827912e-01
1.50227141e+00 -7.65297934e-02 6.72020018e-01 2.02447072e-01
5.05165875e-01 1.14877321e-01 7.04491436e-01 5.80889404e-01
2.85109758e-01 4.32680964e-01 4.27123785e-01 4.28795338e-01
-6.65448979e-02 -5.96308529e-01 1.08486247e+00 6.10396028e-01
6.17165029e-01 8.85733515e-02 -1.17506230e+00 6.74212873e-01
-1.46896195e+00 -6.54719949e-01 -2.15990901e-01 2.53543949e+00
1.35760629e+00 4.94203329e-01 -1.47862121e-01 8.52028877e-02
4.83048260e-01 -1.58446938e-01 -3.09454381e-01 -6.86390340e-01
1.38779864e-01 2.53583491e-01 5.41590452e-01 4.89548177e-01
-8.51703763e-01 7.39643574e-01 5.77892113e+00 9.17221308e-01
-1.30928981e+00 -1.13639101e-01 5.57204068e-01 2.46862695e-01
-6.30587757e-01 4.76319045e-01 -1.21127522e+00 5.63144803e-01
1.52165079e+00 -6.52966797e-01 7.94630051e-01 1.44497859e+00
-2.24429563e-01 -3.90501469e-01 -1.70278072e+00 6.05849981e-01
1.64349526e-01 -1.23981273e+00 -2.36893684e-01 -2.66436130e-01
1.01619399e+00 1.40842974e-01 -4.30664383e-02 9.30572152e-01
5.65702736e-01 -1.11878562e+00 1.42418051e+00 1.25258163e-01
1.12666452e+00 -5.76632917e-01 7.38112271e-01 6.79389477e-01
-1.17240548e+00 -5.72580434e-02 -9.12462249e-02 -8.36218894e-02
-3.60143989e-01 7.98446238e-01 -1.49474955e+00 1.00913733e-01
4.93761510e-01 4.97341722e-01 -9.64385986e-01 1.02921319e+00
-5.88787556e-01 8.79129171e-01 -4.36068922e-02 -4.99743015e-01
-3.97098996e-02 4.58964586e-01 1.99650034e-01 1.57797897e+00
5.54032326e-01 -8.38301897e-01 2.45773762e-01 1.48406971e+00
-2.47067004e-01 2.98132412e-02 -4.55862731e-01 -5.14068842e-01
3.74966085e-01 9.86690402e-01 -4.34468508e-01 -7.72076905e-01
-6.68305874e-01 5.58972657e-01 3.18907350e-01 7.05998540e-02
-1.10225034e+00 -5.60908914e-01 5.16702294e-01 1.52444750e-01
3.46997380e-01 -2.39344716e-01 -4.01301056e-01 -1.13512671e+00
5.03992736e-01 -1.55283487e+00 1.10723533e-01 -5.11236310e-01
-7.09171832e-01 6.08279407e-01 -5.84237278e-03 -1.07996547e+00
-6.56032264e-01 -5.78527331e-01 -5.00012577e-01 1.10967922e+00
-1.27051687e+00 -5.79751968e-01 -1.07306009e-02 -1.36541545e-01
3.39651555e-01 -1.83497757e-01 9.11286414e-01 3.02197635e-01
-1.01810835e-01 1.02144861e+00 -1.38969822e-02 2.03479052e-01
5.89982986e-01 -1.43200278e+00 8.79760265e-01 1.15636015e+00
5.15219476e-03 9.85571563e-01 6.84169292e-01 -7.22032130e-01
-1.90868473e+00 -1.28274226e+00 1.06391859e+00 -6.35476232e-01
9.03719783e-01 -7.03177810e-01 -8.38441014e-01 6.13467395e-01
-1.07376121e-01 -3.03483997e-02 3.86790544e-01 -6.55634552e-02
-8.09946299e-01 6.02389500e-02 -9.17673469e-01 4.89790648e-01
4.81963664e-01 -8.27908039e-01 -3.40819031e-01 5.13181627e-01
7.12363183e-01 -9.16243434e-01 -7.19909787e-01 2.77130187e-01
3.11521888e-01 -8.84256423e-01 5.61611235e-01 -5.37051022e-01
1.08694458e+00 -7.22299933e-01 -6.97079360e-01 -1.13117874e+00
1.59886107e-01 -5.04425168e-01 -3.95519704e-01 1.26844621e+00
9.08701599e-01 -3.76687050e-01 5.37707031e-01 6.15962505e-01
-1.83748975e-01 -7.72370458e-01 -2.95468569e-01 -6.94364309e-01
-4.41864738e-03 -8.65965247e-01 1.01832831e+00 5.35926402e-01
2.90118486e-01 2.55160220e-02 -6.44674525e-02 1.61337018e-01
3.29071075e-01 7.86417902e-01 1.10398400e+00 -6.10439837e-01
-9.99555051e-01 -6.90896273e-01 -4.48259085e-01 -9.74528611e-01
4.95059967e-01 -1.35271120e+00 4.98161077e-01 -1.26237643e+00
5.31437457e-01 -4.19587940e-01 2.03110889e-01 6.38181210e-01
-2.55893797e-01 -2.29561105e-01 6.51689395e-02 -7.86515251e-02
-6.89816892e-01 -9.37797874e-02 4.92646664e-01 -5.47418416e-01
-2.56365109e-02 1.06504433e-01 -5.75281680e-01 4.62044448e-01
5.43123603e-01 -6.80575848e-01 -1.35423407e-01 -4.10016745e-01
6.89842403e-01 4.81867105e-01 2.40899533e-01 -1.33613157e+00
4.45268750e-02 -2.46752858e-01 -8.97818282e-02 -5.16523778e-01
-5.22992849e-01 -3.91053408e-01 -3.33324969e-02 7.51374185e-01
-7.30443895e-01 2.75031894e-01 3.75524819e-01 3.97988632e-02
-3.01340550e-01 -7.57934690e-01 6.34845972e-01 -1.38695210e-01
-9.54618096e-01 -8.41504559e-02 -2.85017461e-01 4.82505679e-01
7.80579269e-01 2.51916766e-01 -5.32508433e-01 8.96583777e-03
2.39078999e-02 1.37539044e-01 7.85254180e-01 4.53550369e-01
3.98346514e-01 -9.95434940e-01 -9.58548844e-01 5.43911159e-01
6.86461389e-01 -2.79915690e-01 -2.63686061e-01 6.58757448e-01
-7.46246755e-01 5.10671079e-01 3.65731210e-01 -8.77313375e-01
-1.30649996e+00 6.09076202e-01 1.00251697e-01 -4.10531402e-01
-8.19237158e-02 7.57330656e-01 1.61958057e-02 -7.09212363e-01
8.27009007e-02 -8.69161487e-01 5.70858240e-01 -7.05114841e-01
6.70320630e-01 -3.75869172e-03 4.04517680e-01 -1.69067487e-01
-5.93884766e-01 1.13869414e-01 5.92777170e-02 1.01619788e-01
1.07086122e+00 8.01735699e-01 -6.77335799e-01 5.38006723e-01
1.49223793e+00 5.89566529e-01 -8.72346520e-01 -1.07933976e-01
1.38078243e-01 -5.85948408e-01 -3.80707026e-01 -1.02722275e+00
-7.93870926e-01 9.82313275e-01 5.19408621e-02 -7.79921608e-03
7.57784843e-01 6.33874238e-02 6.27083659e-01 3.60428035e-01
8.00815403e-01 -5.96193135e-01 -1.01576932e-01 6.72574520e-01
5.34843087e-01 -1.21455812e+00 -1.04083836e-01 -7.83954337e-02
-2.67612875e-01 1.11757708e+00 8.83469999e-01 8.65461156e-02
1.41783491e-01 8.73538315e-01 -1.99274421e-01 3.81534904e-01
-8.81161630e-01 2.68754780e-01 2.68859923e-01 3.33889663e-01
9.86748517e-01 4.26572233e-01 -1.16199143e-01 6.27049506e-01
-4.70192522e-01 1.33586526e-01 6.67067051e-01 1.15361547e+00
-3.38849247e-01 -1.28961849e+00 -5.58434665e-01 8.61675501e-01
-4.96797681e-01 -5.90202093e-01 -4.90556322e-02 5.13151348e-01
1.34444207e-01 7.69683301e-01 -2.76349843e-01 -5.30888140e-01
8.95877853e-02 1.14539430e-01 5.35606682e-01 -1.05410087e+00
-5.48339307e-01 -4.57228512e-01 2.68252581e-01 -4.76276845e-01
4.84059274e-01 -6.49144471e-01 -1.60446513e+00 -5.11177540e-01
-2.66691148e-01 2.85722733e-01 7.57443070e-01 6.98957860e-01
4.81948614e-01 5.24693370e-01 4.24511969e-01 -3.41188669e-01
-8.71164322e-01 -7.42199183e-01 -1.23143002e-01 4.45058733e-01
4.96947229e-01 -1.80318385e-01 -3.46249580e-01 3.50070119e-01] | [7.8607892990112305, 7.712847709655762] |
e94e3294-c9c8-4038-a605-b91b7725c383 | animating-arbitrary-objects-via-deep-motion | 1812.08861 | null | https://arxiv.org/abs/1812.08861v3 | https://arxiv.org/pdf/1812.08861v3.pdf | Animating Arbitrary Objects via Deep Motion Transfer | This paper introduces a novel deep learning framework for image animation. Given an input image with a target object and a driving video sequence depicting a moving object, our framework generates a video in which the target object is animated according to the driving sequence. This is achieved through a deep architecture that decouples appearance and motion information. Our framework consists of three main modules: (i) a Keypoint Detector unsupervisely trained to extract object keypoints, (ii) a Dense Motion prediction network for generating dense heatmaps from sparse keypoints, in order to better encode motion information and (iii) a Motion Transfer Network, which uses the motion heatmaps and appearance information extracted from the input image to synthesize the output frames. We demonstrate the effectiveness of our method on several benchmark datasets, spanning a wide variety of object appearances, and show that our approach outperforms state-of-the-art image animation and video generation methods. Our source code is publicly available. | ['Stéphane Lathuilière', 'Elisa Ricci', 'Aliaksandr Siarohin', 'Sergey Tulyakov', 'Nicu Sebe'] | 2018-12-20 | animating-arbitrary-objects-via-deep-motion-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Siarohin_Animating_Arbitrary_Objects_via_Deep_Motion_Transfer_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Siarohin_Animating_Arbitrary_Objects_via_Deep_Motion_Transfer_CVPR_2019_paper.pdf | cvpr-2019-6 | ['image-animation'] | ['computer-vision'] | [ 3.15487653e-01 -1.45186735e-02 -7.77130798e-02 -2.57891640e-02
-6.15850151e-01 -3.74586105e-01 7.80375838e-01 -5.49337149e-01
-1.75829172e-01 3.71566236e-01 1.35471091e-01 8.17530137e-03
6.24709547e-01 -7.75692523e-01 -1.20167005e+00 -8.11258852e-01
-1.23344004e-01 4.06684995e-01 4.32045639e-01 -2.20482070e-02
2.62242824e-01 4.44770038e-01 -1.60161984e+00 3.70890111e-01
4.68549699e-01 1.00284255e+00 3.49248886e-01 1.03265321e+00
9.68425870e-02 1.15436327e+00 -4.47828889e-01 -1.36499748e-01
3.47310334e-01 -6.83258355e-01 -8.81974041e-01 3.73837739e-01
6.40770435e-01 -8.83399904e-01 -7.16573954e-01 8.70084882e-01
1.63578749e-01 2.62521595e-01 6.75476611e-01 -1.47140622e+00
-5.81283510e-01 4.16618496e-01 -7.45917022e-01 -2.73368750e-02
3.68634075e-01 7.96539009e-01 7.46408701e-01 -8.63819420e-01
1.10736108e+00 1.34360361e+00 2.34024435e-01 7.84437001e-01
-1.34615147e+00 -5.85884213e-01 8.76346231e-02 2.00457931e-01
-1.14338934e+00 -4.32332307e-01 1.02969587e+00 -6.05144083e-01
6.55927539e-01 -3.15411463e-02 1.03359306e+00 1.26292026e+00
3.43457282e-01 9.90358174e-01 5.52422464e-01 -3.20730418e-01
3.56636614e-01 -2.29473487e-01 -3.89031500e-01 8.47298086e-01
-2.19292477e-01 2.99897939e-01 -3.09507489e-01 -1.24986157e-01
1.18894649e+00 -1.19216233e-01 -3.20266187e-01 -6.99110091e-01
-1.46760190e+00 7.52392352e-01 4.89351988e-01 3.53480391e-02
-6.68245316e-01 6.84394896e-01 1.35687575e-01 -7.27420971e-02
1.01821892e-01 1.46791056e-01 -4.37758006e-02 -1.06615268e-01
-9.69807446e-01 6.05431557e-01 7.29535699e-01 7.32371867e-01
8.53273988e-01 3.75535756e-01 -2.88431019e-01 5.10025859e-01
3.40361744e-01 5.98305941e-01 2.79997528e-01 -1.52013493e+00
3.31900507e-01 4.36136305e-01 3.74885738e-01 -1.21324301e+00
-2.46872150e-05 2.29757011e-01 -7.61683345e-01 6.46561325e-01
3.21419537e-01 -1.41530067e-01 -1.07454777e+00 1.81354630e+00
5.65370262e-01 6.81319654e-01 -8.18976015e-02 1.21859550e+00
8.20920050e-01 1.07574272e+00 2.19320416e-01 1.03915609e-01
1.12956643e+00 -1.29962039e+00 -5.24425089e-01 -1.84426054e-01
2.72631526e-01 -6.52175069e-01 9.83238161e-01 1.09487407e-01
-1.48318052e+00 -7.90826261e-01 -8.46425295e-01 -1.86057702e-01
8.00965503e-02 1.88696221e-01 3.71089488e-01 -8.26512743e-03
-1.18773663e+00 4.94091064e-01 -9.38734829e-01 -1.09963782e-01
3.47699523e-01 2.16153249e-01 -4.30599630e-01 1.57491174e-02
-9.67283905e-01 5.40294886e-01 4.48861301e-01 4.98984680e-02
-1.49176443e+00 -6.94287002e-01 -1.28441668e+00 5.82687706e-02
3.93067561e-02 -1.26862097e+00 1.23297000e+00 -1.52823257e+00
-1.76047599e+00 7.74954379e-01 -3.77241038e-02 -3.87118548e-01
7.40864098e-01 -2.26626366e-01 -4.06768806e-02 5.03672540e-01
6.16738424e-02 1.38260543e+00 1.41628516e+00 -1.42761636e+00
-7.94923425e-01 9.90825221e-02 -1.29866138e-01 1.55706674e-01
1.50113449e-01 -1.13521442e-02 -9.02365923e-01 -9.17556047e-01
-5.34084678e-01 -1.14066005e+00 -2.93996155e-01 2.08693817e-01
-5.02690077e-01 6.24048859e-02 1.27753448e+00 -6.14223838e-01
1.11169052e+00 -2.00185871e+00 6.97459280e-01 -9.42112654e-02
1.36936322e-01 3.47448140e-01 -5.47626555e-01 2.73701310e-01
-2.01605543e-01 -3.23369294e-01 -3.12563777e-01 -3.55857670e-01
-2.09225878e-01 1.16948159e-02 -4.02187556e-01 3.84911656e-01
5.03346682e-01 1.27480054e+00 -1.13944912e+00 -3.84801328e-01
6.75890684e-01 7.14015901e-01 -5.94810724e-01 6.29583418e-01
-5.80477118e-01 6.10695481e-01 -2.76320934e-01 4.33180898e-01
4.28324342e-01 -3.98538023e-01 -1.12881407e-01 -2.61991352e-01
-2.24967226e-02 -7.47022852e-02 -1.17299855e+00 1.85656738e+00
-1.92591593e-01 7.33570158e-01 7.34155327e-02 -7.50197768e-01
6.93803608e-01 1.34498149e-01 5.94833851e-01 -3.13010007e-01
2.27663606e-01 -2.04443023e-01 -1.26149252e-01 -6.47926629e-01
5.50523460e-01 3.01161408e-01 7.61298910e-02 5.75536847e-01
4.36223075e-02 -2.44483352e-01 2.74776965e-01 4.07923669e-01
1.19592607e+00 6.00865066e-01 -6.09913543e-02 6.71717338e-03
5.54841459e-01 1.29491761e-01 3.96492124e-01 2.66261578e-01
8.14834703e-03 8.28892291e-01 4.05008346e-01 -9.01900470e-01
-1.52376652e+00 -1.01123083e+00 4.03796583e-01 7.21952617e-01
3.90999138e-01 -3.17320645e-01 -8.27712536e-01 -5.22961199e-01
4.34021652e-02 5.22737682e-01 -9.88732517e-01 -1.24974817e-01
-9.30006325e-01 -2.33905226e-01 7.56623819e-02 6.33023202e-01
4.06315923e-01 -1.53083825e+00 -9.46427405e-01 2.32670754e-01
-3.06604534e-01 -1.10730934e+00 -8.32076788e-01 -4.45040315e-01
-6.37034059e-01 -1.03292477e+00 -9.96445060e-01 -9.53570068e-01
8.07982147e-01 3.24608207e-01 1.13730538e+00 2.10180566e-01
-3.63031000e-01 4.06004310e-01 -7.34682456e-02 3.15782614e-02
-8.66616368e-01 -1.71172276e-01 -3.32288146e-01 2.81951755e-01
-7.10456222e-02 -3.45155001e-01 -1.06878877e+00 1.29716188e-01
-1.12146473e+00 6.06264353e-01 6.53887808e-01 7.00400233e-01
4.34436440e-01 -5.03081322e-01 1.57741860e-01 -5.33750892e-01
-3.56213301e-02 -4.89361435e-01 -7.76202798e-01 1.48618938e-02
2.21367270e-01 8.94010812e-02 5.22676706e-01 -5.47594845e-01
-8.66343856e-01 7.14278340e-01 -6.14159964e-02 -8.70444655e-01
-2.25771025e-01 -5.08520938e-02 1.63453724e-03 -8.20010453e-02
4.30696875e-01 3.99205208e-01 2.84753531e-01 -1.69185773e-01
6.93702340e-01 3.23115230e-01 9.61663663e-01 -4.20701921e-01
9.99308169e-01 7.11222231e-01 -4.47396934e-02 -7.12971330e-01
-3.01470697e-01 -1.64412528e-01 -7.54789114e-01 -5.88055432e-01
1.16901016e+00 -9.26526248e-01 -6.95042670e-01 7.36506402e-01
-1.44155216e+00 -8.03442776e-01 -1.63289100e-01 2.66881138e-01
-9.46271718e-01 2.64076352e-01 -7.73926914e-01 -3.65900844e-01
-4.26762074e-01 -1.43085277e+00 1.51743472e+00 2.72203565e-01
-2.74533510e-01 -8.52638543e-01 2.76319116e-01 1.19131900e-01
2.52146482e-01 7.07214355e-01 8.14964771e-01 9.03416127e-02
-1.16173959e+00 1.65260565e-02 -1.42903656e-01 1.28345966e-01
4.71415557e-02 6.59291506e-01 -6.52013779e-01 -3.56643200e-01
-4.58033383e-01 -3.77102226e-01 8.09328496e-01 5.88294208e-01
1.04497993e+00 -5.02999187e-01 -4.75334644e-01 8.47448647e-01
1.32861984e+00 2.83230633e-01 8.73193264e-01 3.85951787e-01
1.05874491e+00 4.83757645e-01 4.47834313e-01 3.33610713e-01
3.36111814e-01 9.25712705e-01 7.14676142e-01 -2.89519221e-01
-3.56941968e-01 -2.91718185e-01 5.00389934e-01 5.51471710e-01
-6.32093996e-02 -2.88115572e-02 -7.20388830e-01 6.46974504e-01
-2.03275466e+00 -1.22306132e+00 1.53945833e-01 1.91935825e+00
6.20153666e-01 -1.80225939e-01 4.74876136e-01 -2.52336174e-01
7.21875787e-01 3.34693789e-01 -6.68614745e-01 -1.76093087e-01
2.45452508e-01 -3.61542664e-02 -1.79734007e-02 4.50502038e-01
-1.19903183e+00 1.17724025e+00 6.02684736e+00 3.52406412e-01
-1.37021101e+00 -2.37273902e-01 7.69662082e-01 9.18103233e-02
-2.81640232e-01 -9.69957486e-02 -4.46432739e-01 5.69714487e-01
6.14116371e-01 -1.08820282e-01 3.53787869e-01 9.38856006e-01
2.68693268e-01 7.58699700e-02 -1.04769468e+00 8.04755807e-01
1.46712884e-01 -1.74973571e+00 4.29450482e-01 -6.63090050e-02
9.51529264e-01 -1.75033763e-01 8.23636129e-02 5.13565056e-02
5.01104355e-01 -8.90853822e-01 9.70421314e-01 4.94878083e-01
6.95315421e-01 -6.79984450e-01 2.02991933e-01 9.44258720e-02
-1.32000422e+00 1.90399569e-02 -1.47180185e-01 1.58980817e-01
3.72012466e-01 5.37194610e-02 -5.50112545e-01 3.90106559e-01
6.11534953e-01 9.69439685e-01 -2.60447383e-01 9.93415117e-01
-3.01667333e-01 2.59781122e-01 2.61611119e-02 3.45031112e-01
5.16038835e-01 -2.62282908e-01 5.75097978e-01 1.26775289e+00
2.34224349e-01 9.77278799e-02 1.51293680e-01 1.14522386e+00
-8.85594040e-02 -1.57617971e-01 -7.59754777e-01 8.15201402e-02
3.89969945e-01 1.54711151e+00 -7.48810768e-01 -6.87758267e-01
-4.97786313e-01 1.24798167e+00 2.50107259e-01 5.32498837e-01
-1.03867698e+00 -3.35757256e-01 8.46146941e-01 1.59254253e-01
6.56218886e-01 -2.29036018e-01 4.15018141e-01 -1.19376242e+00
-2.79985279e-01 -8.94084036e-01 1.71333119e-01 -1.14516425e+00
-7.28659987e-01 7.28417397e-01 -8.14961642e-02 -1.48142672e+00
-7.82137513e-01 -3.68573368e-01 -9.21405256e-01 7.34065115e-01
-1.18310320e+00 -1.23458278e+00 -7.28330195e-01 5.94771087e-01
7.52641380e-01 1.97646618e-02 5.04278243e-01 1.29163861e-01
-4.61352855e-01 6.96536005e-02 -2.54078239e-01 4.12272960e-01
4.71839786e-01 -1.07012260e+00 9.85793233e-01 8.65832150e-01
1.54566824e-01 2.37050474e-01 5.53316593e-01 -5.93840301e-01
-1.67520547e+00 -1.50829148e+00 2.14665771e-01 -5.35968065e-01
5.39220333e-01 -3.01959932e-01 -9.65710044e-01 7.66021430e-01
4.75049436e-01 3.10694635e-01 1.22068105e-02 -9.99199748e-01
-6.42472208e-02 1.23180628e-01 -6.67852581e-01 8.30721319e-01
8.60185027e-01 -1.41782805e-01 -4.70498204e-01 1.44572958e-01
6.92062855e-01 -7.36637950e-01 -5.74999273e-01 2.83072233e-01
7.53242791e-01 -7.34619617e-01 1.09579456e+00 -7.18197405e-01
1.03324771e+00 -5.76743424e-01 2.71637827e-01 -1.28064978e+00
-4.97877330e-01 -7.61676788e-01 -4.47868466e-01 7.73068190e-01
5.85993864e-02 1.14244170e-01 8.96803796e-01 5.74322224e-01
-6.84523880e-02 -6.88396633e-01 -5.13460577e-01 -3.85707229e-01
-1.70196325e-01 -5.68135753e-02 3.59059066e-01 8.25495243e-01
-4.01436985e-01 3.12332720e-01 -6.34546518e-01 -3.20321396e-02
6.09559834e-01 4.10041332e-01 1.41207540e+00 -6.95766866e-01
-2.77429253e-01 -3.76916498e-01 -5.42252243e-01 -1.18846262e+00
4.19375896e-01 -6.38527095e-01 1.98895454e-01 -1.46979499e+00
3.36279571e-01 3.50363478e-02 1.72491580e-01 3.99431676e-01
-4.45437431e-01 4.26699102e-01 4.73750710e-01 3.20072412e-01
-5.27152419e-01 6.58470273e-01 1.57808614e+00 -2.54139304e-01
-3.55811566e-01 -3.12615901e-01 -2.46680781e-01 8.25057626e-01
3.69525284e-01 -3.33207846e-01 -4.38425988e-01 -4.56108272e-01
-2.62480021e-01 4.37951833e-01 7.70941257e-01 -9.98179555e-01
7.28514493e-02 -2.67018944e-01 7.86811471e-01 -4.96235251e-01
4.37805951e-01 -7.06688404e-01 5.64175963e-01 8.55937660e-01
-3.45381290e-01 2.96463460e-01 2.61318833e-01 6.02388084e-01
-1.18662916e-01 1.96056738e-01 9.60930228e-01 -1.31200805e-01
-9.68842149e-01 6.16891325e-01 -2.26357490e-01 -9.59772319e-02
1.29601324e+00 -7.46382698e-02 -3.15637082e-01 -4.98605281e-01
-6.19605660e-01 1.34775653e-01 8.41082275e-01 7.92877734e-01
8.21917236e-01 -1.65467799e+00 -7.83964694e-01 3.35352898e-01
-8.98511112e-02 2.44123973e-02 2.37803251e-01 5.70833445e-01
-9.25794959e-01 1.49511427e-01 -4.88835990e-01 -9.04808164e-01
-1.16366196e+00 9.06361699e-01 2.43313506e-01 1.50756896e-01
-1.03392720e+00 6.57881558e-01 7.37130105e-01 9.35888737e-02
1.53734669e-01 -4.42883849e-01 3.43222842e-02 -3.86369646e-01
8.02713752e-01 1.68380812e-01 -3.77151012e-01 -1.02989268e+00
-1.47023916e-01 7.81329870e-01 -3.30218561e-02 -1.76327080e-01
1.12785149e+00 3.00530512e-02 1.08727932e-01 2.56676197e-01
1.41697657e+00 -4.02910739e-01 -1.98706365e+00 8.37826636e-03
-3.41041803e-01 -7.16100216e-01 -2.65315950e-01 -3.20620418e-01
-1.33403039e+00 7.69734502e-01 2.58874267e-01 -1.82145953e-01
1.12922966e+00 -2.81365290e-02 1.10174835e+00 -7.30237365e-02
3.27860177e-01 -7.43254304e-01 5.59532881e-01 2.13683382e-01
1.01038074e+00 -1.14527822e+00 -3.33417565e-01 -1.01934113e-01
-8.39759409e-01 1.16389906e+00 7.64345109e-01 -4.44404781e-01
3.24361295e-01 3.38766485e-01 1.34442881e-01 -6.56107664e-02
-1.07897067e+00 -2.00813934e-02 3.15934628e-01 6.37526631e-01
1.54558241e-01 -2.63664007e-01 1.11058213e-01 -1.03138136e-02
1.07707359e-01 1.53313741e-01 4.70409185e-01 8.13938379e-01
-3.51852179e-01 -8.71848643e-01 -3.40340316e-01 7.93856848e-03
-2.77033269e-01 1.26636088e-01 -4.71883953e-01 9.47358549e-01
-1.85147479e-01 3.41851175e-01 3.22906703e-01 -3.44892055e-01
1.34840414e-01 -2.10714608e-01 4.61017787e-01 -4.83913660e-01
-3.71611863e-01 1.03163153e-01 -3.09433609e-01 -1.04984951e+00
-5.70228159e-01 -5.91847360e-01 -1.24330986e+00 -3.19292903e-01
2.38701925e-01 -5.13203479e-02 4.88726377e-01 5.07089555e-01
4.98633057e-01 5.91032088e-01 7.11641729e-01 -1.72898269e+00
-1.03162892e-01 -6.52824104e-01 -2.04143852e-01 7.62478769e-01
5.16641974e-01 -5.39434731e-01 -1.90270528e-01 7.71862268e-01] | [10.832925796508789, -0.7998821139335632] |
fbf2a185-42c6-4462-9647-4dda6ebee7cd | softmatch-addressing-the-quantity-quality | 2301.10921 | null | https://arxiv.org/abs/2301.10921v2 | https://arxiv.org/pdf/2301.10921v2.pdf | SoftMatch: Addressing the Quantity-Quality Trade-off in Semi-supervised Learning | The critical challenge of Semi-Supervised Learning (SSL) is how to effectively leverage the limited labeled data and massive unlabeled data to improve the model's generalization performance. In this paper, we first revisit the popular pseudo-labeling methods via a unified sample weighting formulation and demonstrate the inherent quantity-quality trade-off problem of pseudo-labeling with thresholding, which may prohibit learning. To this end, we propose SoftMatch to overcome the trade-off by maintaining both high quantity and high quality of pseudo-labels during training, effectively exploiting the unlabeled data. We derive a truncated Gaussian function to weight samples based on their confidence, which can be viewed as a soft version of the confidence threshold. We further enhance the utilization of weakly-learned classes by proposing a uniform alignment approach. In experiments, SoftMatch shows substantial improvements across a wide variety of benchmarks, including image, text, and imbalanced classification. | ['Marios Savvides', 'Bhiksha Raj', 'Xing Xie', 'Bernt Schiele', 'Jindong Wang', 'Yidong Wang', 'Yue Fan', 'Ran Tao', 'Hao Chen'] | 2023-01-26 | null | null | null | null | ['imbalanced-classification'] | ['miscellaneous'] | [ 3.73794109e-01 4.52518575e-02 -8.11250091e-01 -8.69087875e-01
-1.17480731e+00 -6.50759399e-01 4.61555153e-01 3.05449396e-01
-4.40643817e-01 8.52047980e-01 -1.65286288e-02 -1.82997748e-01
1.34871885e-01 -4.33036774e-01 -5.98486543e-01 -8.79978538e-01
5.30155241e-01 3.08921009e-01 4.35027294e-02 2.20705077e-01
2.36171886e-01 1.15406387e-01 -1.48692787e+00 2.24447459e-01
1.24519992e+00 1.15117240e+00 -1.92252114e-01 4.39949520e-03
-2.22346395e-01 8.27592850e-01 -4.19912964e-01 -2.88147062e-01
3.89614105e-01 -3.55248004e-01 -6.80892587e-01 5.30732751e-01
6.07096195e-01 -3.43989789e-01 8.67662430e-02 1.30669534e+00
3.23613256e-01 -3.99150029e-02 8.65577519e-01 -1.37405217e+00
-6.19500339e-01 6.07342362e-01 -8.70702684e-01 -7.59367421e-02
-1.04467370e-01 -9.64610744e-03 1.24795806e+00 -1.14647591e+00
3.49912941e-01 1.07396996e+00 5.85177898e-01 4.17238474e-01
-1.28026128e+00 -8.31193030e-01 2.92335749e-01 4.36032303e-02
-1.52060032e+00 -4.34347063e-01 7.60021389e-01 -4.39028770e-01
3.60138506e-01 4.20096815e-02 1.52922124e-01 8.64145100e-01
-3.61009032e-01 1.09002554e+00 1.44465518e+00 -6.63449764e-01
5.47703862e-01 2.86518186e-01 5.79607844e-01 6.35001302e-01
4.42280382e-01 -1.49774179e-01 -6.20456755e-01 -3.27773184e-01
3.64649951e-01 9.46341082e-02 -1.79949760e-01 -8.06284547e-01
-8.99046957e-01 8.28982532e-01 1.64959416e-01 -1.78138793e-01
-3.75233265e-03 -2.34263793e-01 4.39298719e-01 2.11491153e-01
8.43102098e-01 2.55380750e-01 -6.73678279e-01 7.55616650e-02
-1.05201375e+00 -3.98416743e-02 5.04622936e-01 1.02672827e+00
9.46990967e-01 -8.13649520e-02 -3.06118608e-01 1.09716785e+00
3.82069260e-01 4.79358166e-01 4.17835861e-01 -8.93934667e-01
5.50150335e-01 7.54392684e-01 2.04327121e-01 -4.22333509e-01
-1.23956785e-01 -6.25387669e-01 -6.71820760e-01 1.04679957e-01
5.90640068e-01 4.33669500e-02 -1.13826895e+00 1.89630437e+00
6.35016501e-01 2.32089996e-01 -5.17179724e-03 7.73513317e-01
3.43409061e-01 2.60383934e-01 2.07726762e-01 -3.86519045e-01
1.10316753e+00 -1.35915923e+00 -7.48046696e-01 -3.18106413e-01
8.51073980e-01 -5.53138673e-01 1.34344709e+00 3.51864994e-01
-8.16997945e-01 -4.72508430e-01 -1.36924076e+00 -8.83137435e-02
-2.13200197e-01 1.88669294e-01 3.96533191e-01 8.23834896e-01
-5.31177700e-01 6.16964579e-01 -9.36903059e-01 1.78816076e-02
7.36401081e-01 2.59451479e-01 -2.69913316e-01 -4.20251727e-01
-8.98550868e-01 4.99208838e-01 3.70472908e-01 -1.99157879e-01
-6.77836239e-01 -7.78343618e-01 -7.94731498e-01 -5.80199324e-02
5.58942437e-01 -2.59892762e-01 1.26488888e+00 -8.57631266e-01
-1.35545099e+00 1.04642701e+00 -1.02280989e-01 -1.83407992e-01
5.65634549e-01 -3.52102518e-01 -2.31098130e-01 9.27526429e-02
2.27951795e-01 5.61579168e-01 8.34704518e-01 -1.37004268e+00
-7.53004849e-01 -5.16914606e-01 -1.95237756e-01 3.74549210e-01
-8.02172303e-01 -1.84551358e-01 -4.38863665e-01 -7.64532685e-01
3.28752398e-01 -9.11890090e-01 -2.20472932e-01 2.28036027e-02
-4.03532863e-01 -2.83599705e-01 6.35559559e-01 -1.41259566e-01
1.09093797e+00 -2.25867963e+00 -1.66522384e-01 2.10499555e-01
4.44775462e-01 3.34230185e-01 -1.23880342e-01 -1.70065220e-02
3.01104542e-02 4.26513851e-02 -3.75187755e-01 -6.52256787e-01
5.55895455e-02 4.01328385e-01 -3.28021914e-01 6.54852867e-01
2.70048827e-01 8.38758290e-01 -1.13792503e+00 -6.48479640e-01
-6.01536706e-02 1.34380013e-01 -3.71254534e-01 2.70655811e-01
-2.32631788e-01 3.16698432e-01 -5.02164125e-01 8.94152045e-01
8.48990917e-01 -6.24207139e-01 3.23608220e-01 -2.62511671e-01
2.68990219e-01 2.90348023e-01 -1.18229830e+00 1.58718944e+00
-1.82770744e-01 7.14121237e-02 1.18887179e-01 -1.17781222e+00
8.71867657e-01 2.29499787e-02 4.52007622e-01 -5.04739404e-01
1.07379854e-01 3.85307431e-01 -3.65880340e-01 -1.85107440e-01
3.05265337e-01 -1.42908812e-01 3.13525740e-03 7.46220410e-01
1.54144034e-01 1.07965730e-01 6.87770620e-02 1.24367110e-01
7.82029927e-01 2.51194865e-01 2.86322683e-01 -4.81326610e-01
3.29502970e-01 -1.68550342e-01 8.55370998e-01 7.32819259e-01
-4.71752077e-01 7.04447031e-01 4.44528341e-01 -1.34960130e-01
-1.04261529e+00 -1.05465698e+00 -3.56753677e-01 1.36075163e+00
1.08328559e-01 -3.34380478e-01 -7.22790122e-01 -1.16989374e+00
9.44736302e-02 4.49505091e-01 -5.81090152e-01 -3.61292541e-01
-3.95903051e-01 -1.19693553e+00 3.56874794e-01 7.05126226e-01
2.96250433e-01 -3.18834424e-01 -2.06586674e-01 -1.49068413e-02
-9.32267681e-02 -1.18833065e+00 -6.23315990e-01 5.58955252e-01
-9.09147680e-01 -8.88106287e-01 -5.26174963e-01 -7.39378870e-01
1.00447428e+00 5.40200293e-01 1.15864325e+00 -6.15839027e-02
-5.47264796e-03 1.36686279e-03 -4.70491618e-01 -2.40615144e-01
-2.77307719e-01 3.07185709e-01 2.48359680e-01 1.49530217e-01
4.97337431e-01 -5.28346777e-01 -5.77795923e-01 4.88913119e-01
-9.67860520e-01 1.01205697e-02 5.62394023e-01 1.21155357e+00
8.58085692e-01 -1.30183503e-01 7.76318014e-01 -1.16295826e+00
2.73659378e-01 -4.81248111e-01 -5.52268207e-01 5.23934662e-01
-1.16448700e+00 1.96528062e-01 5.92232227e-01 -6.63871050e-01
-9.61328089e-01 1.61245733e-01 1.47688523e-01 -4.33295637e-01
2.01088816e-01 3.92970502e-01 -3.57719213e-01 -2.01400667e-01
5.69625080e-01 -6.72909468e-02 -2.88355406e-02 -5.70498228e-01
5.91756105e-01 9.64946568e-01 4.86905128e-01 -8.89273942e-01
7.77418554e-01 6.31131887e-01 -1.60369202e-01 -2.04210445e-01
-1.70663357e+00 -6.60728335e-01 -6.76599860e-01 8.80189985e-02
2.80706435e-01 -1.09802091e+00 -2.10100025e-01 3.66979420e-01
-4.77286041e-01 -4.41813499e-01 -4.30282623e-01 5.72825789e-01
-3.09767455e-01 6.14488542e-01 -8.05382252e-01 -7.14362085e-01
-3.40273112e-01 -1.02031326e+00 1.17035091e+00 2.23603472e-01
8.49044994e-02 -7.58231461e-01 -1.03310771e-01 5.68606436e-01
2.63013154e-01 3.07208556e-03 8.02855015e-01 -9.55773413e-01
-2.79458284e-01 -2.65760213e-01 -5.34850359e-01 7.22047389e-01
2.09554508e-01 -2.15409070e-01 -1.12614262e+00 -5.43919384e-01
8.12757239e-02 -1.02900028e+00 9.34431851e-01 7.26496205e-02
1.24991119e+00 4.13171276e-02 -2.09831282e-01 6.52910531e-01
1.27525485e+00 -2.20366314e-01 1.49357617e-01 1.50905862e-01
8.36722970e-01 5.76625884e-01 1.13272130e+00 5.13517797e-01
4.30410057e-01 4.43767220e-01 2.57129371e-01 -9.47867259e-02
-5.95177375e-02 -3.73553276e-01 1.00133970e-01 1.06368685e+00
4.20290977e-01 -1.54551808e-02 -7.03454554e-01 3.36549729e-01
-1.93111622e+00 -3.58732015e-01 2.06260562e-01 2.33994246e+00
1.32762599e+00 3.83601427e-01 -5.21397963e-03 3.00821960e-01
7.69750535e-01 4.78678122e-02 -7.50907958e-01 2.13748679e-01
-1.43641546e-01 6.69373199e-02 6.86888158e-01 3.46454591e-01
-1.32221723e+00 8.68582606e-01 6.75249577e+00 1.28346503e+00
-1.06764507e+00 2.50360459e-01 9.73302782e-01 -7.34366775e-02
-2.60215044e-01 -1.17444910e-01 -1.07508636e+00 4.77511376e-01
7.65842915e-01 -7.08154216e-03 1.04132861e-01 1.01205027e+00
-2.22453084e-02 1.83402561e-02 -1.23040640e+00 9.35470939e-01
9.63510945e-02 -9.50441957e-01 -2.07392961e-01 -3.22337560e-02
9.84609604e-01 1.02903850e-01 1.22101538e-01 4.36853588e-01
4.20096427e-01 -7.71151423e-01 7.71651745e-01 -4.31561172e-02
1.02877927e+00 -5.34176230e-01 5.60437322e-01 3.62555474e-01
-9.72534001e-01 -9.03154686e-02 -4.36742753e-01 8.41054246e-02
-5.98767679e-03 9.93015110e-01 -6.12962365e-01 4.94415283e-01
3.89040917e-01 7.11801946e-01 -5.47368407e-01 8.49236786e-01
-3.05115372e-01 9.05778706e-01 -3.48810107e-01 2.49734089e-01
1.60037339e-01 -3.46929789e-01 -5.45975901e-02 1.10718262e+00
8.62914976e-03 -2.05949340e-02 6.42235875e-01 6.25407577e-01
-4.34794933e-01 2.20669582e-01 -7.32986331e-02 8.31130818e-02
7.17145681e-01 1.34429455e+00 -8.61752987e-01 -5.82865775e-01
-4.37566936e-01 7.55460799e-01 6.52688920e-01 3.10949564e-01
-8.10288906e-01 -5.82009181e-02 2.63950855e-01 -1.82512552e-01
1.19788222e-01 -7.42552504e-02 -5.76953948e-01 -1.37570620e+00
2.33635738e-01 -8.39863300e-01 4.48786438e-01 -3.37926745e-01
-1.51478279e+00 3.31782371e-01 1.51738403e-02 -1.38626587e+00
7.65997618e-02 -5.98896384e-01 -1.66278154e-01 5.86521626e-01
-1.82784438e+00 -1.17941165e+00 -2.27447495e-01 2.98698545e-01
4.94008690e-01 -3.89122292e-02 6.02251887e-01 6.07243180e-01
-7.36838520e-01 9.16321933e-01 3.86143267e-01 -4.01918404e-02
1.21940112e+00 -1.36449170e+00 1.13879070e-01 6.66537344e-01
1.36551544e-01 5.75331509e-01 4.19753790e-01 -4.51732606e-01
-1.03918850e+00 -1.24562144e+00 5.20837188e-01 -3.53097081e-01
5.58358729e-01 -4.48575795e-01 -1.03446114e+00 4.75114316e-01
-2.34860137e-01 4.83824968e-01 1.10724604e+00 1.32906765e-01
-9.28062499e-01 -2.17394248e-01 -1.29967368e+00 3.67587954e-01
9.46106136e-01 -5.94768703e-01 -4.71853018e-01 5.73660672e-01
8.01714182e-01 -2.55811602e-01 -8.93601596e-01 7.86072135e-01
5.16888261e-01 -7.57906199e-01 8.84520352e-01 -5.80148280e-01
2.54610538e-01 -2.62419343e-01 -3.21305066e-01 -1.19737446e+00
-1.25644565e-01 -3.56378466e-01 -2.68962294e-01 1.49338651e+00
2.74546713e-01 -5.31052530e-01 1.20644009e+00 7.25165904e-01
-7.45377988e-02 -1.07101023e+00 -7.70650268e-01 -8.60571384e-01
1.00595124e-01 -2.57569134e-01 4.47389424e-01 1.29613197e+00
1.02907181e-01 3.50831568e-01 -4.51950580e-01 -2.31071785e-02
1.05683696e+00 2.08377853e-01 5.62070847e-01 -1.12221658e+00
-3.87678802e-01 -1.99617222e-01 -5.60206659e-02 -1.31531596e+00
2.96381712e-01 -8.91267002e-01 7.12483227e-02 -9.02989864e-01
5.33785522e-01 -7.92771518e-01 -6.38377488e-01 6.93033040e-01
-6.06685281e-01 5.88992715e-01 -1.71792075e-01 5.38615763e-01
-1.01875889e+00 5.77533245e-01 1.07027316e+00 -5.51731959e-02
9.08632763e-03 -9.08093974e-02 -8.41640651e-01 6.51984036e-01
7.67167807e-01 -6.17749512e-01 -5.59583008e-01 -4.18481201e-01
5.69891818e-02 -3.42996269e-01 -1.50268614e-01 -7.15864956e-01
9.68812481e-02 -3.17668140e-01 1.13244802e-01 -5.37544847e-01
1.61125675e-01 -6.63576424e-01 -4.78367686e-01 2.67019182e-01
-6.72767162e-01 -3.11512172e-01 -2.41129711e-01 7.55475402e-01
-2.93569982e-01 -1.77212968e-01 1.03490829e+00 2.97741205e-01
-3.31239462e-01 3.14089805e-01 1.96470574e-01 2.86912769e-01
9.68629777e-01 -1.99719537e-02 -3.04774940e-01 -6.95157647e-02
-3.58348936e-01 5.75074017e-01 6.08564079e-01 1.89298168e-01
2.65774786e-01 -1.47674513e+00 -5.62018871e-01 1.85055748e-01
4.80408788e-01 1.58897683e-01 4.83991997e-03 7.97192693e-01
-1.37190282e-01 1.45866483e-01 5.13162911e-02 -6.31659806e-01
-9.49106276e-01 6.70408666e-01 1.93072483e-02 -3.53563845e-01
-2.51664847e-01 7.89901078e-01 1.50342241e-01 -6.10653937e-01
5.10061324e-01 -1.51499048e-01 5.63126579e-02 2.11962443e-02
5.86636245e-01 3.03869337e-01 2.52587736e-01 -3.62135917e-01
-2.26813361e-01 4.40030754e-01 -2.87654579e-01 1.43169304e-02
1.12719584e+00 -3.90055448e-01 3.00346017e-02 5.39084673e-01
1.33491230e+00 5.30446209e-02 -1.57619107e+00 -7.72482336e-01
2.74947137e-01 -5.88391066e-01 1.76208559e-03 -6.79632127e-01
-1.01368237e+00 7.31175482e-01 4.68053311e-01 -9.73213837e-03
9.63534474e-01 -8.12560245e-02 7.95770466e-01 3.29893112e-01
2.55955607e-01 -1.38965523e+00 2.66766012e-01 1.97622150e-01
5.76322451e-02 -1.60501933e+00 1.46476895e-01 -6.93248510e-01
-6.07027292e-01 8.42047572e-01 6.61569118e-01 7.04075322e-02
5.82842648e-01 4.12368745e-01 2.27684975e-01 4.89475094e-02
-6.76212788e-01 -6.91761747e-02 2.18073845e-01 2.33348429e-01
3.88009101e-01 1.17177546e-01 -5.04561841e-01 7.16295242e-01
3.32479447e-01 2.32609827e-02 2.25320771e-01 1.24183345e+00
-5.61415493e-01 -1.38333428e+00 -3.29810709e-01 6.30303741e-01
-5.97532034e-01 -4.16805968e-02 -8.12351108e-02 2.64396816e-01
6.65266290e-02 9.22827542e-01 -2.08030030e-01 -3.77505302e-01
-3.31365429e-02 2.43937686e-01 3.05891514e-01 -7.55581379e-01
-7.29069486e-02 3.03265333e-01 -1.16883829e-01 -2.83950150e-01
-6.47855580e-01 -4.47997838e-01 -1.05451751e+00 1.06909655e-01
-6.79327071e-01 2.31497601e-01 4.98016685e-01 1.00049281e+00
2.49012619e-01 1.15279391e-01 9.84949172e-01 -5.99764287e-01
-1.41435409e+00 -8.89593601e-01 -7.41477370e-01 7.23229468e-01
2.03921095e-01 -8.47583115e-01 -6.09448850e-01 4.08396944e-02] | [9.413321495056152, 3.80349063873291] |
253bc2a1-02a6-4578-bd95-22591c5d6cc3 | advancing-3d-medical-image-analysis-with | 2201.01426 | null | https://arxiv.org/abs/2201.01426v1 | https://arxiv.org/pdf/2201.01426v1.pdf | Advancing 3D Medical Image Analysis with Variable Dimension Transform based Supervised 3D Pre-training | The difficulties in both data acquisition and annotation substantially restrict the sample sizes of training datasets for 3D medical imaging applications. As a result, constructing high-performance 3D convolutional neural networks from scratch remains a difficult task in the absence of a sufficient pre-training parameter. Previous efforts on 3D pre-training have frequently relied on self-supervised approaches, which use either predictive or contrastive learning on unlabeled data to build invariant 3D representations. However, because of the unavailability of large-scale supervision information, obtaining semantically invariant and discriminative representations from these learning frameworks remains problematic. In this paper, we revisit an innovative yet simple fully-supervised 3D network pre-training framework to take advantage of semantic supervisions from large-scale 2D natural image datasets. With a redesigned 3D network architecture, reformulated natural images are used to address the problem of data scarcity and develop powerful 3D representations. Comprehensive experiments on four benchmark datasets demonstrate that the proposed pre-trained models can effectively accelerate convergence while also improving accuracy for a variety of 3D medical imaging tasks such as classification, segmentation and detection. In addition, as compared to training from scratch, it can save up to 60% of annotation efforts. On the NIH DeepLesion dataset, it likewise achieves state-of-the-art detection performance, outperforming earlier self-supervised and fully-supervised pre-training approaches, as well as methods that do training from scratch. To facilitate further development of 3D medical models, our code and pre-trained model weights are publicly available at https://github.com/urmagicsmine/CSPR. | ['Yizhou Yu', 'Jiechao Ma', 'Hong-Yu Zhou', 'Zihao Li', 'Shu Zhang'] | 2022-01-05 | null | null | null | null | ['medical-object-detection'] | ['computer-vision'] | [ 3.41202348e-01 2.21857280e-01 -2.44504258e-01 -5.31110644e-01
-9.85084236e-01 -3.62978131e-01 2.59573936e-01 1.89059138e-01
-5.09926558e-01 4.35202688e-01 3.68575081e-02 -4.83874023e-01
2.91046739e-01 -4.67191070e-01 -6.22125208e-01 -5.71233630e-01
-8.16293061e-03 4.64002818e-01 2.90372252e-01 2.75121201e-02
-2.09902182e-01 6.16422832e-01 -1.13537753e+00 2.84013543e-02
7.19055712e-01 1.01153433e+00 4.47704971e-01 3.06083113e-01
-1.57183751e-01 3.90821248e-01 -2.50860959e-01 9.88727137e-02
3.22413027e-01 -3.19232792e-01 -8.46387565e-01 2.74446547e-01
3.35899442e-01 -5.64075053e-01 -4.14819092e-01 9.85261202e-01
9.10114884e-01 -2.34430194e-01 5.11348188e-01 -7.10325241e-01
-7.22164631e-01 1.05030760e-01 -5.93797803e-01 5.34992635e-01
-2.81510260e-02 7.70832151e-02 5.98636806e-01 -8.49628687e-01
7.22050190e-01 7.15220511e-01 8.48129272e-01 9.02369022e-01
-1.20250201e+00 -6.81578457e-01 -1.11984462e-01 -2.66009122e-01
-1.21887994e+00 -3.32343996e-01 8.35973918e-01 -4.79418665e-01
9.31180477e-01 -2.92754006e-02 5.83887756e-01 1.13925242e+00
-1.34867933e-02 1.06574893e+00 1.07497466e+00 -4.08024698e-01
9.26435832e-03 -1.07535951e-01 1.32876217e-01 1.00926030e+00
2.28888780e-01 2.28153169e-02 -9.74158347e-02 -1.34256884e-01
1.17557812e+00 1.57193989e-01 -3.41636807e-01 -7.74465442e-01
-1.19610703e+00 6.70099676e-01 7.65066385e-01 3.68993729e-01
-2.47803703e-01 -1.49428830e-01 4.51701522e-01 1.03419088e-02
7.50280380e-01 4.21548724e-01 -4.67970788e-01 2.41505176e-01
-8.61910522e-01 -1.86630577e-01 2.73011506e-01 8.11896980e-01
6.96046412e-01 -1.62008945e-02 -8.91087651e-02 8.85002971e-01
1.19409926e-01 3.60213369e-01 6.25925243e-01 -5.75410664e-01
1.46550611e-01 8.75084281e-01 -2.45083317e-01 -6.35322869e-01
-6.98450744e-01 -7.54966736e-01 -9.81548250e-01 2.29709968e-01
5.82051575e-01 -5.79207763e-02 -1.46580374e+00 1.66523337e+00
5.28406143e-01 2.40916833e-01 -5.97420186e-02 1.06443119e+00
1.08137190e+00 1.55652672e-01 4.72185574e-02 1.06982384e-02
1.13395810e+00 -8.40044498e-01 -3.86122286e-01 -1.38350934e-01
9.28324580e-01 -5.63940287e-01 1.05724967e+00 7.91040529e-03
-8.17396879e-01 -4.84377563e-01 -1.03335333e+00 -2.31933430e-01
-2.54336268e-01 1.83807597e-01 8.72936487e-01 5.66244662e-01
-1.02462602e+00 3.39937836e-01 -1.28827024e+00 -3.56518120e-01
1.07095814e+00 3.90214294e-01 -5.85950434e-01 -3.89611572e-01
-8.97645712e-01 1.00950992e+00 2.79203624e-01 9.92672220e-02
-1.12687218e+00 -9.10067618e-01 -9.21397269e-01 -3.59951794e-01
3.42767358e-01 -7.68275976e-01 1.13614500e+00 -7.98617065e-01
-1.15874648e+00 1.48036397e+00 2.19690576e-01 -3.69272023e-01
4.79941547e-01 -1.21276855e-01 -9.22654271e-02 4.86955404e-01
3.60014319e-01 8.64726543e-01 6.44158065e-01 -1.02392495e+00
-2.14678019e-01 -6.22614205e-01 -6.88856393e-02 8.74675587e-02
-3.48403066e-01 -1.93312243e-01 -8.23487997e-01 -6.03905439e-01
2.71562576e-01 -9.03598070e-01 -6.86079562e-01 4.49014008e-01
-4.32295203e-01 -6.54032975e-02 7.26225317e-01 -5.05945385e-01
6.28000259e-01 -2.17300200e+00 -8.09513330e-02 7.01206364e-03
5.17314553e-01 4.68126953e-01 -1.84360132e-01 -2.76468128e-01
-3.11630666e-01 1.76219977e-02 -4.83303159e-01 -2.68556267e-01
-4.50696290e-01 2.59059846e-01 1.74338236e-01 6.74590707e-01
5.17419398e-01 1.05132210e+00 -1.05020046e+00 -6.88299477e-01
3.56204391e-01 5.69503725e-01 -6.38698101e-01 2.27243632e-01
4.95345145e-02 8.25668514e-01 -7.54255831e-01 9.06240642e-01
4.59042460e-01 -7.65450895e-01 2.41022743e-02 -3.05871755e-01
2.11705491e-01 1.48244426e-01 -5.95089376e-01 2.41772652e+00
-4.09650862e-01 3.22223008e-01 -1.19501380e-02 -1.57463205e+00
8.85767937e-01 2.82762557e-01 8.48641276e-01 -6.88355327e-01
3.11596185e-01 4.06643689e-01 -1.56742990e-01 -4.78377640e-01
-1.22562155e-01 -2.36021116e-01 8.94415472e-03 4.48055416e-01
4.62256700e-01 -2.84918278e-01 1.85937341e-03 1.57667488e-01
1.21337569e+00 2.30843768e-01 2.68286228e-01 -2.06801593e-01
3.42417657e-01 2.02965930e-01 6.54514790e-01 6.67190671e-01
-4.73454028e-01 9.36522722e-01 3.45168263e-01 -6.67465866e-01
-1.03303123e+00 -9.88759995e-01 -5.88352025e-01 5.31642199e-01
1.49985284e-01 -1.99753940e-01 -5.94209075e-01 -1.06362271e+00
1.76154636e-03 2.05621645e-01 -8.16959620e-01 -1.19584508e-01
-4.79013830e-01 -1.03333056e+00 6.38885617e-01 5.32163799e-01
4.60784346e-01 -6.99170113e-01 -6.34364069e-01 2.16734409e-01
4.41904739e-03 -1.27352345e+00 -2.82196134e-01 4.10804480e-01
-1.20592272e+00 -1.20937335e+00 -1.09678245e+00 -9.55392301e-01
1.16078854e+00 4.76480424e-01 1.03021383e+00 3.48077327e-01
-8.36646557e-01 3.28354925e-01 -2.89142013e-01 -4.39987630e-01
-3.15459251e-01 1.49387941e-01 -6.93884119e-02 -4.82492805e-01
3.44335228e-01 -5.85200131e-01 -7.20594943e-01 2.19185755e-01
-8.64046037e-01 2.81628907e-01 8.39128792e-01 1.13518775e+00
1.01761889e+00 -3.23476911e-01 6.99222088e-01 -1.01894653e+00
1.79542273e-01 -3.43897074e-01 -3.92988682e-01 1.42463341e-01
-5.05202591e-01 -2.79249791e-02 3.42838466e-01 -3.48490328e-01
-8.70504916e-01 4.79505837e-01 -3.60890865e-01 -6.20956600e-01
-3.30327183e-01 3.93807471e-01 1.16689771e-01 -1.60524324e-01
9.00843978e-01 1.56670690e-01 5.05530000e-01 -5.59691191e-01
1.84573054e-01 5.40658772e-01 5.11836112e-01 -5.05381644e-01
7.85611689e-01 6.70405567e-01 -1.16278671e-01 -7.19666660e-01
-1.24383450e+00 -5.52598178e-01 -8.99362743e-01 -2.03882121e-02
1.01488638e+00 -1.06324208e+00 1.96902435e-02 3.95060748e-01
-7.93705583e-01 -4.04404521e-01 -3.40047777e-01 5.78086436e-01
-4.03609216e-01 5.58194816e-01 -6.12244189e-01 -2.15950474e-01
-5.18450856e-01 -1.37351882e+00 1.21071649e+00 1.24682091e-01
2.70855566e-03 -9.52524364e-01 -2.54254825e-02 5.79902172e-01
4.02805269e-01 4.82583046e-01 9.57092464e-01 -6.81153476e-01
-4.42373246e-01 -4.82287943e-01 -3.67695272e-01 5.29459000e-01
4.96334910e-01 -5.06391406e-01 -8.51523161e-01 -3.02977920e-01
-1.33850545e-01 -7.23572493e-01 9.02456820e-01 4.86827254e-01
1.38524950e+00 3.51005971e-01 -4.82697725e-01 8.98977458e-01
1.15806067e+00 -2.52727062e-01 2.75266707e-01 2.57608324e-01
6.87449038e-01 2.48771995e-01 4.20993954e-01 2.75943339e-01
2.36974373e-01 3.57959360e-01 4.57477421e-01 -7.90272355e-01
-3.48272860e-01 -1.69484064e-01 -2.54689366e-01 6.63888335e-01
-7.40350634e-02 2.62165815e-01 -1.11951065e+00 5.69752216e-01
-1.57662666e+00 -5.75754583e-01 2.33307719e-01 1.95902443e+00
1.08311200e+00 2.59860247e-01 -6.46006018e-02 -7.29394406e-02
6.15520418e-01 -1.34230275e-02 -9.16759789e-01 3.33245188e-01
1.34460721e-02 4.77691144e-01 5.29266059e-01 1.15105093e-01
-1.50555313e+00 9.13618147e-01 5.87896204e+00 7.42688358e-01
-1.40202510e+00 3.15070689e-01 9.68925178e-01 -1.27115786e-01
-4.34777923e-02 -2.74689466e-01 -6.43720865e-01 6.44388422e-02
4.54543889e-01 1.25153422e-01 -1.57421783e-01 1.09660304e+00
1.98498413e-01 1.15060188e-01 -1.04995966e+00 1.05600011e+00
3.98088396e-02 -1.60594881e+00 5.90296499e-02 -2.34516636e-02
6.53910339e-01 4.50996965e-01 -1.33460909e-01 2.65313804e-01
5.89302033e-02 -1.07345510e+00 2.14485392e-01 1.59627479e-02
1.12563515e+00 -3.96430969e-01 8.41566861e-01 3.91842127e-01
-8.26802611e-01 3.68648589e-01 -4.06694263e-01 2.48428643e-01
1.72903895e-01 8.60729337e-01 -9.68006253e-01 5.64091444e-01
8.51181865e-01 9.60514307e-01 -6.01242721e-01 1.07059777e+00
-2.71416575e-01 4.79702771e-01 -4.42649394e-01 2.87978619e-01
3.41394067e-01 2.39544332e-01 2.90506393e-01 1.08039176e+00
1.05753936e-01 2.71347880e-01 3.25204432e-01 7.97756493e-01
-2.39055336e-01 4.04816121e-02 -5.92869222e-01 7.17658401e-02
9.06838998e-02 1.36923790e+00 -9.10824060e-01 -3.44882458e-01
-5.36625803e-01 8.86687338e-01 4.42027003e-01 2.41448328e-01
-7.38294959e-01 -7.18320161e-02 4.58402395e-01 1.59033403e-01
1.71881750e-01 -3.76416117e-01 -2.80714929e-01 -1.32299256e+00
-1.37326762e-01 -6.21605754e-01 4.28851426e-01 -4.86266643e-01
-1.41729558e+00 7.44220853e-01 -1.08404249e-01 -1.26022160e+00
-3.71807441e-02 -7.73749292e-01 -3.98323059e-01 7.00300753e-01
-1.86003828e+00 -1.19207466e+00 -4.93645042e-01 7.14854479e-01
4.33948457e-01 -1.33567184e-01 1.01744306e+00 4.61269140e-01
-6.08874619e-01 5.44330359e-01 -1.52755827e-01 4.05901313e-01
7.21954286e-01 -1.10256827e+00 2.17497379e-01 6.65440738e-01
8.55844542e-02 5.47290385e-01 1.04585536e-01 -5.00200570e-01
-1.31564355e+00 -1.15950191e+00 2.95836240e-01 -4.80480790e-01
4.66684997e-01 -3.28377634e-01 -9.32775497e-01 5.25049031e-01
-2.40840450e-01 7.61854053e-01 9.56441700e-01 -2.90524010e-02
-2.19680563e-01 1.36239454e-01 -1.14476669e+00 3.03507149e-01
1.28222060e+00 -4.96926874e-01 -6.09725118e-01 6.10042036e-01
6.52913034e-01 -7.36363232e-01 -9.88917172e-01 6.46989822e-01
3.25833797e-01 -7.05551684e-01 1.22717047e+00 -6.03009880e-01
3.76705080e-01 -1.66541055e-01 5.19456044e-02 -9.17586446e-01
-1.41708091e-01 -2.93930709e-01 1.25568613e-01 7.14937508e-01
3.78459483e-01 -3.71674448e-01 1.06973815e+00 4.53019857e-01
-5.41898131e-01 -1.07049489e+00 -1.07664073e+00 -5.14928997e-01
1.83667004e-01 -4.29883391e-01 1.01029269e-01 1.10990047e+00
-2.38858044e-01 3.43409449e-01 -4.62798662e-02 1.96485907e-01
7.97584832e-01 2.42690235e-01 7.04709411e-01 -1.08886218e+00
-1.63649619e-01 -2.69630581e-01 -5.38488209e-01 -1.24892044e+00
1.45112470e-01 -1.31175697e+00 -6.36267290e-02 -1.47097981e+00
3.23133826e-01 -8.60429525e-01 -4.91551608e-01 8.67221117e-01
-2.04894423e-01 6.58393145e-01 -2.45877743e-01 3.95712763e-01
-5.74617326e-01 5.57023406e-01 1.65507889e+00 -1.86046004e-01
-1.04012705e-01 -6.45942315e-02 -6.88992500e-01 7.47889340e-01
8.01919937e-01 -4.60671127e-01 -4.65797275e-01 -6.25940919e-01
-4.22246277e-01 2.42317305e-03 4.80733007e-01 -7.74597049e-01
-4.11001928e-02 2.41056576e-01 7.33161211e-01 -6.19033456e-01
1.54488146e-01 -6.20273948e-01 -3.43117654e-01 4.81176734e-01
-1.92135140e-01 -3.51973027e-01 3.78924072e-01 4.63449240e-01
-2.19833508e-01 -2.84770373e-02 9.29644465e-01 -4.92769718e-01
-7.35153317e-01 6.98966324e-01 -1.02314986e-01 1.13217592e-01
9.56740439e-01 -2.53288686e-01 -8.04911554e-02 4.58243191e-02
-8.97189379e-01 2.57893771e-01 5.17697394e-01 3.12591314e-01
7.87683964e-01 -1.17439532e+00 -5.57907283e-01 2.73219734e-01
2.00248078e-01 5.21870375e-01 3.55017304e-01 9.20837760e-01
-6.67813241e-01 3.26416373e-01 -4.20007557e-01 -9.98476148e-01
-8.76464844e-01 4.57193762e-01 4.76585209e-01 -4.07827765e-01
-1.00135875e+00 7.91676164e-01 4.22418803e-01 -7.14833081e-01
2.00020552e-01 -5.24150312e-01 2.20146291e-02 -3.52682978e-01
4.12707776e-01 -3.84888530e-01 2.63510257e-01 -4.12473381e-01
-5.42980134e-01 5.06702304e-01 -2.84412533e-01 4.92228895e-01
1.71881962e+00 6.57191649e-02 2.62536645e-01 3.88675593e-02
1.37585127e+00 -5.21874309e-01 -1.29478633e+00 -4.77806270e-01
-1.59554362e-01 -3.42426360e-01 3.74727696e-01 -6.40495598e-01
-1.28854883e+00 9.92701709e-01 8.08498681e-01 -2.57825971e-01
1.04114151e+00 3.36101443e-01 8.51829410e-01 3.03229660e-01
3.70926976e-01 -8.01944315e-01 3.09844851e-01 2.19276384e-01
6.06697798e-01 -1.65475595e+00 5.78130074e-02 -5.01937866e-01
-3.84491444e-01 9.22859669e-01 7.38524735e-01 -1.44089982e-01
6.88868165e-01 1.79557517e-01 2.54880458e-01 -3.95290732e-01
-3.29629689e-01 -2.70341247e-01 2.62346834e-01 7.17695355e-01
6.30769432e-01 -2.49842629e-01 4.12416086e-02 5.95667005e-01
2.04117805e-01 2.21905828e-01 2.62012035e-01 1.07043779e+00
-1.65483564e-01 -1.02662790e+00 -1.40163582e-02 5.81400692e-01
-4.94347304e-01 -7.95368031e-02 -9.08101201e-02 8.88508320e-01
2.63574179e-02 4.23817605e-01 -8.14796910e-02 -1.01158231e-01
2.51816213e-01 -1.30924433e-01 5.86294174e-01 -8.99354339e-01
-2.44326264e-01 1.97438121e-01 -9.00750905e-02 -4.63605523e-01
-6.78570032e-01 -4.52325970e-01 -1.47097206e+00 2.05657333e-01
-4.19495821e-01 -2.00568944e-01 4.96937215e-01 8.90273690e-01
4.93014634e-01 5.19882679e-01 4.30227160e-01 -1.05352509e+00
-5.76083004e-01 -8.25239956e-01 -3.55751961e-01 5.40744781e-01
2.18076780e-01 -9.69951391e-01 -6.25074804e-02 -1.48237543e-03] | [14.681035995483398, -2.264152765274048] |
83c9df50-64fb-4aab-b51a-13190a78752d | in-context-learning-for-attention-scheme-from | 2307.02419 | null | https://arxiv.org/abs/2307.02419v1 | https://arxiv.org/pdf/2307.02419v1.pdf | In-Context Learning for Attention Scheme: from Single Softmax Regression to Multiple Softmax Regression via a Tensor Trick | Large language models (LLMs) have brought significant and transformative changes in human society. These models have demonstrated remarkable capabilities in natural language understanding and generation, leading to various advancements and impacts across several domains. We consider the in-context learning under two formulation for attention related regression in this work. Given matrices $A_1 \in \mathbb{R}^{n \times d}$, and $A_2 \in \mathbb{R}^{n \times d}$ and $B \in \mathbb{R}^{n \times n}$, the purpose is to solve some certain optimization problems: Normalized version $\min_{X} \| D(X)^{-1} \exp(A_1 X A_2^\top) - B \|_F^2$ and Rescaled version $\| \exp(A_1 X A_2^\top) - D(X) \cdot B \|_F^2$. Here $D(X) := \mathrm{diag}( \exp(A_1 X A_2^\top) {\bf 1}_n )$. Our regression problem shares similarities with previous studies on softmax-related regression. Prior research has extensively investigated regression techniques related to softmax regression: Normalized version $\| \langle \exp(Ax) , {\bf 1}_n \rangle^{-1} \exp(Ax) - b \|_2^2$ and Resscaled version $\| \exp(Ax) - \langle \exp(Ax), {\bf 1}_n \rangle b \|_2^2 $ In contrast to previous approaches, we adopt a vectorization technique to address the regression problem in matrix formulation. This approach expands the dimension from $d$ to $d^2$, resembling the formulation of the regression problem mentioned earlier. Upon completing the lipschitz analysis of our regression function, we have derived our main result concerning in-context learning. | ['Shenghao Xie', 'Zhao Song', 'Yeqi Gao'] | 2023-07-05 | null | null | null | null | ['natural-language-understanding'] | ['natural-language-processing'] | [ 5.44836462e-01 6.29847273e-02 -2.38140702e-01 -5.38360536e-01
-8.94782662e-01 -2.89946377e-01 2.94436187e-01 -3.95042636e-03
-8.20259035e-01 1.04938388e+00 -1.59265459e-01 -6.05812430e-01
-5.98591387e-01 -8.87955070e-01 -7.17525125e-01 -7.52836406e-01
-6.46415949e-01 3.80490571e-02 -3.77693743e-01 -7.22687066e-01
1.87578708e-01 1.56086698e-01 -1.44995821e+00 -1.07963562e-01
6.76687539e-01 1.43914068e+00 3.53729248e-01 6.09180808e-01
-2.61886865e-01 7.99245775e-01 -7.28995502e-01 -5.06131113e-01
6.00598633e-01 -5.68094432e-01 -7.09751666e-01 -4.32120800e-01
4.34944510e-01 2.00596333e-01 -4.67522979e-01 1.36075938e+00
3.24586153e-01 7.37129271e-01 8.73209774e-01 -9.59197342e-01
-8.09290767e-01 6.76784754e-01 -1.12819302e+00 3.75612855e-01
2.19988212e-01 -2.17939243e-01 1.23559463e+00 -8.32271755e-01
1.88965783e-01 1.23020363e+00 4.44338828e-01 4.97692883e-01
-1.14333308e+00 -1.25757587e+00 7.65404165e-01 -1.60739020e-01
-1.62998962e+00 -9.67933889e-03 5.24661779e-01 -4.76563185e-01
1.03488696e+00 5.31887710e-01 1.41164556e-01 7.37788975e-01
8.90931189e-02 6.61712110e-01 1.25433445e+00 -7.06178963e-01
5.51224165e-02 8.48328173e-02 2.06276968e-01 6.80325031e-01
-4.85481434e-02 1.02300152e-01 -3.95935893e-01 2.19460741e-01
1.16792786e+00 -3.56587931e-03 9.24007073e-02 7.58307636e-01
-9.44077492e-01 1.20883739e+00 4.52850670e-01 6.06538892e-01
-1.54086217e-01 6.32880390e-01 1.47940647e-02 6.16811514e-01
4.20701087e-01 5.23294985e-01 -7.61087596e-01 -1.06105566e-01
-1.01657009e+00 3.49080652e-01 4.23166513e-01 1.43124449e+00
9.15771604e-01 5.12484074e-01 -3.00288260e-01 1.26103902e+00
2.97442675e-01 8.19344401e-01 4.42871690e-01 -1.06228828e+00
6.63337529e-01 4.92119670e-01 -1.39333114e-01 -7.86925137e-01
-6.50903046e-01 -3.99827540e-01 -1.35896182e+00 1.89409196e-01
3.96070808e-01 -4.14812595e-01 -6.02765739e-01 2.32612205e+00
-3.81148428e-01 -2.16798544e-01 -3.45807016e-01 6.49082303e-01
6.44288898e-01 8.35494339e-01 3.35962117e-01 -7.31675267e-01
1.34114993e+00 -7.08809078e-01 -4.48196441e-01 -6.72636330e-01
3.59451979e-01 -8.53058398e-01 1.45902002e+00 6.39318109e-01
-1.35995710e+00 -8.60118151e-01 -8.55008423e-01 -7.68664703e-02
-5.83182812e-01 8.93925503e-02 8.09336066e-01 8.65900218e-01
-9.63528931e-01 3.80517632e-01 -2.43033022e-01 1.76419243e-01
2.15886101e-01 8.74646366e-01 -4.34948266e-01 -4.50705085e-03
-1.38422179e+00 7.84312904e-01 9.21603292e-02 1.88079208e-01
-4.98297572e-01 -4.57680702e-01 -1.03889942e+00 -2.20404342e-01
5.45050144e-01 -2.64411777e-01 8.19496810e-01 -8.22070003e-01
-1.12419462e+00 9.18825924e-01 -2.07227662e-01 -3.61114353e-01
1.68584615e-01 -1.94836646e-01 -5.81210256e-01 -6.21564686e-01
3.63673627e-01 6.79338753e-01 7.97890961e-01 -8.89266789e-01
-6.99383795e-01 -6.80492282e-01 2.33228132e-01 6.14439286e-02
-3.85253906e-01 3.35215837e-01 -5.18863976e-01 -1.33494389e+00
1.58787698e-01 -8.27886701e-01 -5.28068781e-01 -4.63786393e-01
-4.53756034e-01 1.59658670e-01 1.75593019e-01 -6.72021091e-01
2.07313418e+00 -1.89938450e+00 2.36496031e-01 5.91005564e-01
2.61448920e-01 1.73682109e-01 1.43327946e-02 1.94376647e-01
-5.01130879e-01 2.77763575e-01 -4.33817744e-01 -1.51961893e-01
7.84277171e-02 1.50420144e-01 -2.77743429e-01 2.17290059e-01
-1.99451625e-01 4.22209352e-01 -5.44451356e-01 -3.12256347e-02
-5.12552969e-02 4.06410754e-01 -4.83328044e-01 -3.68592143e-02
-3.76588762e-01 3.57707560e-01 -5.45638561e-01 4.70127821e-01
7.74135411e-01 -2.64547318e-01 3.22148986e-02 7.67072216e-02
-2.35386238e-01 1.36726713e-02 -1.73523760e+00 1.57450688e+00
-5.85600436e-01 4.41644937e-01 4.50241268e-01 -1.51906955e+00
1.14728999e+00 3.85600738e-02 7.73258567e-01 -1.00659585e+00
2.91684806e-01 -5.33805788e-02 -3.69400606e-02 -2.77327746e-01
4.52148080e-01 -4.34772789e-01 -4.83209282e-01 4.25516605e-01
-1.72860757e-01 -2.72840023e-01 3.73795986e-01 -1.74202189e-01
1.01199496e+00 2.43099835e-02 2.97859281e-01 -2.70498186e-01
8.25139999e-01 -4.19550776e-01 4.77164894e-01 1.09378254e+00
-1.27489388e-01 2.18081251e-01 9.01274860e-01 -9.93504748e-02
-7.89934754e-01 -8.14177692e-01 -3.53312701e-01 2.01413298e+00
5.64055033e-02 -4.94870454e-01 -6.20328367e-01 -5.24163365e-01
-1.21012926e-01 8.65854383e-01 -8.20898890e-01 7.15279579e-02
-8.64897490e-01 -1.23660266e+00 6.89789474e-01 8.06129992e-01
5.27285635e-01 -1.20283520e+00 -2.22875997e-01 2.01784849e-01
1.28066465e-01 -5.61463952e-01 -7.35912442e-01 7.29088426e-01
-5.38355231e-01 -6.54334009e-01 -7.73434162e-01 -6.67472243e-01
8.08887601e-01 -3.22324365e-01 9.32954371e-01 -8.08598474e-02
-5.95985115e-01 1.45248517e-01 -1.75575063e-01 -7.28403389e-01
1.13567404e-01 -2.69907802e-01 3.18706036e-01 -1.92185551e-01
4.02762920e-01 -5.22751629e-01 -6.80251777e-01 3.59866709e-01
-1.04955101e+00 -3.66649151e-01 5.43523550e-01 1.00078321e+00
9.23282743e-01 -5.21336123e-02 6.79490447e-01 -9.56189334e-01
9.42013741e-01 -2.68812031e-01 -8.23557019e-01 9.71044749e-02
-8.40689182e-01 1.50167197e-02 7.00037181e-01 -6.30968034e-01
-5.32911479e-01 -2.98783988e-01 -3.58479738e-01 -4.69689190e-01
5.57732470e-02 7.62310505e-01 1.61792673e-02 3.68657708e-01
9.14405406e-01 1.77616686e-01 -2.37140715e-01 -3.86956125e-01
7.55228519e-01 5.06516397e-01 4.72029239e-01 -8.33504736e-01
6.51071370e-01 4.00254466e-02 6.82040229e-02 -5.52879095e-01
-6.09693348e-01 1.01831760e-02 -2.94467807e-01 3.00289482e-01
9.46153522e-01 -1.00793755e+00 -1.01513827e+00 -5.35567328e-02
-3.54335099e-01 -2.37272382e-01 -5.72815597e-01 7.74791956e-01
-6.44313395e-01 9.61127058e-02 -2.49027237e-01 -1.04942083e+00
-4.24169540e-01 -1.19641244e+00 5.82246542e-01 3.31420690e-01
-2.65997738e-01 -8.84641051e-01 -4.17151362e-01 4.35367346e-01
5.84041953e-01 -1.29930824e-02 1.22835505e+00 -7.06161618e-01
-2.43686840e-01 -2.35747129e-01 -4.16925400e-01 8.40055764e-01
9.09970254e-02 -5.62820673e-01 -5.14489710e-01 -2.56134212e-01
6.98512867e-02 -1.63210221e-02 6.93437338e-01 6.23421490e-01
1.40067172e+00 -5.43507576e-01 -2.62435913e-01 4.94189382e-01
1.08954847e+00 5.80929339e-01 4.49653536e-01 1.06946848e-01
2.67254829e-01 1.73516855e-01 5.64216971e-01 7.09506571e-01
9.55258682e-02 6.92642152e-01 2.06438363e-01 -1.44354209e-01
1.22824401e-01 2.02567428e-01 4.27398294e-01 5.00691712e-01
-2.53365189e-01 -1.63319424e-01 -7.28243768e-01 3.80622298e-01
-1.50035822e+00 -7.30908930e-01 4.87569980e-02 2.46631312e+00
1.08950591e+00 4.13795382e-01 -7.97828808e-02 -3.88633274e-02
5.83794475e-01 1.62402138e-01 -3.06820095e-01 -6.87269807e-01
-2.27556348e-01 1.19763803e+00 5.65978944e-01 5.77233553e-01
-1.08348286e+00 1.07551587e+00 4.64077234e+00 1.33643949e+00
-1.26506591e+00 8.87742266e-03 7.83707201e-01 -5.91446102e-01
-2.07103878e-01 7.13575333e-02 -1.06137860e+00 6.51838124e-01
6.72693372e-01 -4.15331163e-02 7.98450768e-01 9.67035234e-01
1.41055137e-01 -3.00542027e-01 -1.34296191e+00 1.19635057e+00
2.67993629e-01 -1.04367542e+00 -2.54782259e-01 -6.89884797e-02
7.06312358e-01 -2.89834261e-01 5.25527775e-01 1.00547302e+00
4.81099814e-01 -1.61057186e+00 5.82761765e-01 1.32816032e-01
1.47285974e+00 -6.35129809e-01 2.68558025e-01 4.17543530e-01
-1.42110431e+00 -4.69626695e-01 -3.92433286e-01 -2.56114751e-01
9.62392464e-02 5.28799891e-01 -3.61616403e-01 4.18770581e-01
9.26410794e-01 2.54958838e-01 -1.29141938e-02 2.82380134e-01
-1.56191900e-01 3.52189064e-01 -3.24380487e-01 -4.86990400e-02
2.61807829e-01 -4.98481035e-01 4.88256842e-01 1.31331980e+00
2.46623814e-01 4.65152621e-01 4.47680593e-01 9.12990153e-01
-3.51881295e-01 4.65010017e-01 -3.74876946e-01 1.25390336e-01
2.40395978e-01 7.43199706e-01 -5.45535326e-01 -1.93879753e-02
-6.13939106e-01 3.90690476e-01 3.09417158e-01 5.47070801e-01
-9.99902189e-01 -9.55867827e-01 8.17816436e-01 3.27892125e-01
2.97486752e-01 -3.23552042e-01 -4.66419965e-01 -8.48963201e-01
-8.73658657e-02 -1.03898919e+00 6.28164053e-01 -5.86890697e-01
-1.25475407e+00 7.02858090e-01 4.01454628e-01 -1.11590958e+00
-1.24915831e-01 -8.47317278e-01 -7.33119249e-02 1.47116506e+00
-1.05161893e+00 -9.85180199e-01 2.08650678e-01 1.08472288e+00
4.49939162e-01 -7.32322097e-01 9.56473529e-01 6.11606359e-01
-5.10700941e-01 1.04897332e+00 1.17602825e-01 2.58085161e-01
7.07996368e-01 -9.21205103e-01 -1.73409849e-01 5.38123608e-01
1.06889404e-01 1.07588840e+00 5.16562521e-01 -5.99531591e-01
-1.04688621e+00 -8.31643224e-01 8.39583278e-01 -1.71107933e-01
4.73944277e-01 -4.32526052e-01 -5.98890543e-01 1.12547493e+00
-6.95084129e-03 -5.19512221e-03 9.99189615e-01 7.05530420e-02
-3.36607277e-01 -3.24934751e-01 -1.19748187e+00 7.46716559e-01
9.12068725e-01 -3.95343244e-01 -3.42126489e-01 3.21527690e-01
4.57130611e-01 -4.34536099e-01 -1.13232636e+00 4.93652552e-01
4.09636021e-01 -5.82638681e-01 1.14479804e+00 -6.60145938e-01
2.46841505e-01 1.29301056e-01 -7.77878582e-01 -6.66345119e-01
-3.23949546e-01 -8.94306898e-01 1.81414008e-01 7.94385135e-01
7.75806546e-01 -5.45215547e-01 6.16518557e-01 7.55621791e-01
-1.04164347e-01 -7.81941891e-01 -1.34408545e+00 -3.42969686e-01
7.82533228e-01 -7.40445554e-01 1.64626062e-01 8.17335486e-01
-7.00987875e-02 2.37114560e-02 -5.19675255e-01 -2.13989466e-01
3.76532525e-02 -1.69019952e-01 4.18051600e-01 -9.59542334e-01
-6.31457388e-01 -8.42720926e-01 2.33833715e-02 -1.31738222e+00
-1.34301651e-02 -1.05929613e+00 -2.69524574e-01 -1.12661767e+00
3.26414444e-02 -8.37687254e-01 -6.46608412e-01 8.43044043e-01
1.68446880e-02 1.58904478e-01 1.18413717e-01 -1.02717660e-01
-1.02984220e-01 1.68331400e-01 1.12741947e+00 -4.06687677e-01
-4.19940412e-01 2.14056835e-01 -1.16186619e+00 7.35179365e-01
2.32455745e-01 -5.06478012e-01 -4.93492454e-01 -5.50668538e-01
6.86314821e-01 4.06341225e-01 -1.97098330e-02 -5.80981195e-01
2.12166026e-01 -4.51561838e-01 4.83970523e-01 -2.19827101e-01
8.09340835e-01 -4.62192357e-01 -1.79592907e-01 -1.26706839e-01
-5.56715131e-01 1.89856768e-01 3.56316388e-01 3.58347535e-01
-5.28001897e-02 -2.36588553e-01 7.63273180e-01 -4.95377958e-01
-4.20654178e-01 3.05093020e-01 -2.97688723e-01 1.89162791e-01
9.66121018e-01 -1.78231537e-01 -4.11292091e-02 -6.21157289e-01
-1.15200281e+00 1.40874162e-01 -4.62229699e-01 4.41292256e-01
2.74129272e-01 -1.20063806e+00 -8.25937092e-01 6.07703984e-01
-2.12116018e-01 2.69550979e-01 3.61128479e-01 8.53638530e-01
-1.82500124e-01 2.26448596e-01 -4.41675261e-02 -3.35057735e-01
-9.19022679e-01 4.78841901e-01 1.40926674e-01 -5.68008900e-01
7.17550591e-02 1.75751197e+00 5.94403222e-02 -2.21661255e-01
5.77737570e-01 -4.85086292e-01 -2.48113498e-01 1.44476742e-01
3.13739210e-01 4.07366574e-01 -2.87082139e-02 -6.95916295e-01
-3.07151258e-01 6.53727472e-01 -4.09149945e-01 -3.71172845e-01
1.28487253e+00 -4.86103781e-02 -3.72364938e-01 3.08163285e-01
1.10283184e+00 1.53298199e-01 -8.13469410e-01 -2.91059315e-01
-4.08281684e-01 -3.06341082e-01 -2.32121140e-01 -9.13932204e-01
-1.07961810e+00 7.53384173e-01 8.28352451e-01 2.02436239e-01
1.12402093e+00 -5.26656657e-02 2.14119807e-01 2.62729704e-01
3.78791451e-01 -1.60628331e+00 1.56027704e-01 7.27501512e-01
9.72883999e-01 -1.25435233e+00 3.66560251e-01 -1.55465631e-02
-6.72683418e-01 7.66891778e-01 7.23961174e-01 1.52974268e-02
1.19706786e+00 2.63749242e-01 3.81408371e-02 1.12851355e-02
-2.05529645e-01 -7.03627914e-02 3.18244457e-01 1.82846546e-01
9.96204078e-01 4.29530814e-02 -6.38018906e-01 1.41579437e+00
-5.84611893e-01 -2.15277612e-01 7.19870850e-02 1.11265850e+00
-2.83834875e-01 -1.60903835e+00 -4.29389447e-01 7.19339252e-01
-7.49835074e-01 -4.22543079e-01 2.20044583e-01 8.13631475e-01
5.94847679e-01 1.12784398e+00 -1.88611429e-02 -2.56047279e-01
3.52652609e-01 3.46456707e-01 5.38954437e-01 -5.85086703e-01
-7.97456026e-01 4.48081791e-01 -2.92301655e-01 -2.07782179e-01
-4.10255715e-02 -5.98415673e-01 -1.68566597e+00 -1.39587417e-01
-2.79520392e-01 2.00101994e-02 8.11829329e-01 7.55116463e-01
-8.13652482e-03 5.01574039e-01 3.91316652e-01 -3.64138871e-01
-5.61785936e-01 -1.17339706e+00 -1.01386261e+00 4.73379463e-01
1.21186532e-01 -7.26894200e-01 -1.70755237e-01 8.50864276e-02] | [6.696108818054199, 4.502172946929932] |
02dc6632-471d-4432-adcf-b515fe3849ff | ericson-an-interactive-open-domain | 2304.02233 | null | https://arxiv.org/abs/2304.02233v1 | https://arxiv.org/pdf/2304.02233v1.pdf | Ericson: An Interactive Open-Domain Conversational Search Agent | Open-domain conversational search (ODCS) aims to provide valuable, up-to-date information, while maintaining natural conversations to help users refine and ultimately answer information needs. However, creating an effective and robust ODCS agent is challenging. In this paper, we present a fully functional ODCS system, Ericson, which includes state-of-the-art question answering and information retrieval components, as well as intent inference and dialogue management models for proactive question refinement and recommendations. Our system was stress-tested in the Amazon Alexa Prize, by engaging in live conversations with thousands of Alexa users, thus providing empirical basis for the analysis of the ODCS system in real settings. Our interaction data analysis revealed that accurate intent classification, encouraging user engagement, and careful proactive recommendations contribute most to the users satisfaction. Our study further identifies limitations of the existing search techniques, and can serve as a building block for the next generation of ODCS agents. | ['Eugene Agichtein', 'Payam Karisani', 'Jason Choi', 'Ali Ahmadvand', 'ZiHao Wang'] | 2023-04-05 | null | null | null | null | ['conversational-search', 'intent-classification', 'dialogue-management'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-5.45984268e-01 3.07320595e-01 -1.57127738e-01 -4.37240601e-01
-7.85990179e-01 -6.73083484e-01 7.69186795e-01 2.78499871e-01
-2.98123419e-01 5.81480742e-01 7.58206666e-01 -3.94253314e-01
-2.62699246e-01 -3.27924907e-01 2.93443888e-01 -5.72994677e-03
6.88303187e-02 7.96755970e-01 3.75738114e-01 -1.04165566e+00
7.54305124e-01 8.29926059e-02 -1.74429500e+00 5.12730837e-01
1.23726642e+00 7.56309569e-01 3.83354783e-01 1.10284710e+00
-5.91503203e-01 1.34942591e+00 -7.42411256e-01 -4.17271346e-01
-2.01190501e-01 -1.50560647e-01 -1.44371545e+00 -4.75566268e-01
2.62814343e-01 -9.02731776e-01 -2.51164436e-01 1.59370542e-01
6.65208220e-01 5.69738209e-01 3.18555623e-01 -1.39267159e+00
-8.17961574e-01 2.99191356e-01 4.69649464e-01 9.44301784e-02
1.22623491e+00 2.94415861e-01 1.32869387e+00 -5.75812638e-01
6.08702183e-01 1.00906312e+00 3.11799616e-01 8.18955898e-01
-8.65761042e-01 -3.81213307e-01 2.19622906e-02 3.41417104e-01
-8.52421820e-01 -7.95316458e-01 4.33374822e-01 -2.88860768e-01
1.29483080e+00 8.61888289e-01 5.42855442e-01 1.16630566e+00
-2.53431052e-01 9.02986646e-01 7.42985964e-01 -3.98489267e-01
3.68794262e-01 7.46035993e-01 8.04233670e-01 5.92011869e-01
-1.93058655e-01 -3.90342951e-01 -7.19913840e-01 -8.68369579e-01
4.41974625e-02 -1.41551614e-01 -2.18973577e-01 1.63392156e-01
-3.09264362e-01 8.74230206e-01 1.45134449e-01 4.43411738e-01
-5.69222867e-01 -3.98198485e-01 1.70900524e-01 3.88914496e-01
5.21551430e-01 9.69024122e-01 -3.75813991e-01 -1.04447365e+00
-1.71637923e-01 5.80199361e-01 1.87713671e+00 8.71718228e-01
6.24782026e-01 -7.44179428e-01 -4.78480756e-01 1.30445135e+00
4.54423934e-01 3.38847846e-01 4.37449872e-01 -1.57782114e+00
1.20954417e-01 1.03324854e+00 6.14758730e-01 -9.72574234e-01
-3.69877219e-01 -1.51870865e-02 1.76599681e-01 -4.65529144e-01
2.30220228e-01 -2.66044021e-01 -2.95330528e-02 1.45675600e+00
2.35787198e-01 -4.43019420e-01 -4.63393852e-02 7.33461380e-01
1.08206069e+00 4.97386605e-01 5.47564514e-02 -7.11767077e-02
1.62626970e+00 -9.47293401e-01 -9.05712247e-01 -3.50708723e-01
8.34520698e-01 -8.70495915e-01 1.51390135e+00 7.33617097e-02
-1.02990031e+00 -1.24857143e-01 -4.08908248e-01 -3.56524259e-01
-2.02871948e-01 -5.05736649e-01 7.74012208e-01 7.62447178e-01
-1.30201721e+00 -6.78308122e-03 -2.65566528e-01 -7.71154284e-01
-2.33901173e-01 1.18208595e-01 2.04467416e-01 -1.10448614e-01
-1.36244357e+00 1.02019179e+00 -4.73353326e-01 -3.49561483e-01
-3.83500189e-01 -5.83445430e-01 -3.61242920e-01 1.75643757e-01
3.86431634e-01 -5.97983897e-01 2.12754321e+00 -2.08416313e-01
-1.73825157e+00 6.35651588e-01 -3.78362238e-01 -3.41584682e-02
-3.74023877e-02 -3.57582659e-01 -3.37222278e-01 1.70092374e-01
3.54496151e-01 4.10768241e-01 1.88154578e-01 -1.14826441e+00
-7.93599010e-01 -3.50997150e-01 5.70104420e-01 6.67881370e-01
-5.32119334e-01 2.17566401e-01 -5.92733979e-01 2.09728584e-01
-4.79802758e-01 -7.35338867e-01 -1.28269896e-01 -3.81865412e-01
1.11307405e-01 -9.77849722e-01 5.97019076e-01 -8.36828053e-01
1.71996272e+00 -1.90756381e+00 -3.11117113e-01 -3.03227711e-03
3.16455007e-01 3.08829099e-01 -6.67434111e-02 1.15816462e+00
6.58786416e-01 3.95858914e-01 5.36746085e-01 -5.11013687e-01
2.73460329e-01 7.17029199e-02 1.94796436e-02 -2.01484486e-01
-4.03463036e-01 8.06122541e-01 -8.43857706e-01 -5.15020847e-01
-1.73524562e-02 1.67961996e-02 -7.71917105e-01 7.98076570e-01
-5.49604297e-01 -3.29709059e-04 -6.77854061e-01 4.79534417e-01
5.93887419e-02 -3.35382044e-01 -1.61056012e-01 3.25126857e-01
-3.38246673e-01 9.33091104e-01 -5.31133413e-01 1.43657899e+00
-7.71638572e-01 6.21372998e-01 5.15590131e-01 -2.00560957e-01
7.33030796e-01 3.00153613e-01 6.46667719e-01 -1.24946630e+00
6.01161867e-02 5.04473671e-02 -3.59656632e-01 -9.76974010e-01
9.31333065e-01 5.76225042e-01 -1.59581929e-01 1.03567219e+00
-2.85955340e-01 -1.18202612e-01 1.86739415e-01 7.89140821e-01
1.33952856e+00 -6.23175502e-01 1.47955239e-01 -2.47807920e-01
4.92162079e-01 2.85190791e-01 -1.43137842e-01 1.05352998e+00
-5.01066566e-01 -4.65220958e-02 3.03679526e-01 -1.23638913e-01
-4.70353544e-01 -4.57032055e-01 1.32968768e-01 1.59386420e+00
9.63262022e-02 -7.36045599e-01 -1.06999671e+00 -5.77864468e-01
-7.54838511e-02 1.21505880e+00 -7.38769248e-02 -1.25310272e-01
-2.97522038e-01 5.68264276e-02 5.62677741e-01 -4.74063270e-02
7.19819546e-01 -1.14755952e+00 -3.65342945e-01 3.00497442e-01
-1.05588639e+00 -9.79833007e-01 -8.50416481e-01 -3.97868037e-01
-3.43858808e-01 -1.25425804e+00 -2.54909188e-01 -7.87649512e-01
5.33427894e-02 8.08988333e-01 1.37748122e+00 6.62622869e-01
-1.45605728e-01 1.09830856e+00 -6.75628245e-01 -1.04215644e-01
-3.69615734e-01 3.89264792e-01 1.19068311e-03 -3.87881190e-01
6.60242915e-01 -3.85595739e-01 -1.00178850e+00 7.78532505e-01
-4.79751438e-01 -2.84508079e-01 -4.06569690e-02 5.79942346e-01
-6.92714095e-01 -3.56370032e-01 8.30789924e-01 -9.05935049e-01
1.80428720e+00 -7.57069886e-01 -1.14986584e-01 2.26093292e-01
-1.02940631e+00 -1.55239627e-01 2.23857120e-01 8.07786435e-02
-1.41212642e+00 -6.74616873e-01 -6.59884214e-01 7.11431742e-01
-1.76109940e-01 4.24360573e-01 4.14075583e-01 -1.13524057e-01
1.08974421e+00 -2.67545376e-02 2.97367513e-01 -5.25311410e-01
4.01898503e-01 1.54149520e+00 1.04068980e-01 -5.58267117e-01
1.05070718e-01 3.06797829e-02 -1.01065052e+00 -1.13394451e+00
-6.52008176e-01 -1.23938370e+00 1.21993639e-01 -4.58997458e-01
5.54425955e-01 -5.74362457e-01 -1.68969369e+00 2.10998178e-01
-1.18516994e+00 -5.57707965e-01 5.80647811e-02 -1.58545420e-01
-3.26923728e-01 3.07565391e-01 -8.35478604e-01 -1.45669293e+00
-6.84731245e-01 -8.61079276e-01 8.04111660e-01 4.66625422e-01
-9.68708098e-01 -9.39602315e-01 4.43474591e-01 1.38141608e+00
8.33044171e-01 -9.07135785e-01 7.93020844e-01 -1.09714675e+00
-5.21546006e-01 -4.21942383e-01 -6.79658502e-02 -2.16831006e-02
1.37142524e-01 -1.63249344e-01 -9.06820655e-01 -9.98555869e-02
1.54304281e-01 -5.78495443e-01 -4.36304748e-04 -1.52152210e-01
6.72805846e-01 -6.97217047e-01 -2.45926231e-01 -4.25751209e-01
8.02435637e-01 4.64074224e-01 4.92844760e-01 3.25869620e-01
-1.05389833e-01 8.70122075e-01 6.68906868e-01 7.23488331e-01
1.09043646e+00 8.63439322e-01 1.95208132e-01 4.11550343e-01
1.59376994e-01 -2.39564091e-01 2.70166725e-01 6.71755672e-01
1.53254479e-01 -5.06808341e-01 -7.00853825e-01 3.88332903e-01
-1.84168220e+00 -1.00162816e+00 1.42181944e-02 1.83745420e+00
8.69687617e-01 -2.50525236e-01 3.24581563e-01 -3.89573842e-01
3.48306328e-01 -2.22244244e-02 -4.35851902e-01 -5.47843993e-01
5.80177486e-01 3.37439328e-02 -1.60705224e-01 9.47822869e-01
-2.14585081e-01 7.66791761e-01 6.99293137e+00 3.36294383e-01
-3.58958840e-01 1.64731085e-01 4.11111981e-01 -1.22726941e-02
-5.14751256e-01 -4.76028584e-03 -8.58736038e-01 4.95512336e-01
1.09097648e+00 -3.61712128e-01 9.44908082e-01 1.09711814e+00
2.99189061e-01 -3.78190607e-01 -9.84981954e-01 8.37543368e-01
1.39597163e-01 -1.38362598e+00 -4.56407338e-01 1.27276331e-02
2.70174980e-01 -3.84261645e-02 -3.98272008e-01 8.77716362e-01
6.73300564e-01 -5.71957827e-01 3.52056162e-03 5.13682067e-01
9.33036283e-02 -2.91738153e-01 6.23790085e-01 8.62000823e-01
-6.85831070e-01 -3.27435851e-01 2.86155730e-01 -3.87651742e-01
2.31317416e-01 -8.50256979e-02 -1.27553761e+00 -2.76461393e-02
6.90484464e-01 1.67015731e-01 -4.48973089e-01 6.54077649e-01
3.54253262e-01 3.88225436e-01 -3.22442293e-01 -9.29266810e-01
7.26875588e-02 -1.11277029e-01 4.32096153e-01 9.73441541e-01
-1.98765129e-01 6.20391488e-01 3.97536516e-01 4.29516286e-01
-1.16314754e-01 2.81955481e-01 -4.94438738e-01 -2.05051348e-01
9.08109784e-01 1.20788097e+00 -1.11227401e-01 -2.76948005e-01
-3.79638910e-01 8.91952693e-01 5.29677927e-01 2.85667539e-01
-4.48428422e-01 -2.52353907e-01 1.06006491e+00 2.21545786e-01
-5.38083315e-01 -2.24333659e-01 -2.91296899e-01 -8.84653032e-01
2.05853116e-03 -1.28330958e+00 4.41455811e-01 -7.08483398e-01
-1.20717204e+00 5.27172208e-01 -1.46630794e-01 -4.18984413e-01
-7.64887094e-01 -1.05470292e-01 -7.47861683e-01 8.74894559e-01
-1.36911595e+00 -4.67181236e-01 -8.30879927e-01 5.07239997e-01
9.41247880e-01 -1.11527301e-01 1.11456895e+00 5.04468203e-01
-4.77484733e-01 3.61667603e-01 -1.18462138e-01 -4.01007891e-01
8.82018566e-01 -9.45827842e-01 3.03642958e-01 5.49988300e-02
-2.95048028e-01 1.21286929e+00 6.44079745e-01 -6.73709571e-01
-1.72726214e+00 -2.31342912e-01 1.25492573e+00 -7.45810926e-01
5.88414907e-01 -4.29775506e-01 -8.99692953e-01 3.82111847e-01
6.85880423e-01 -9.37718093e-01 1.03352535e+00 6.72604084e-01
1.84966605e-02 8.14305544e-02 -1.08616650e+00 9.97993529e-01
1.08435953e+00 -9.46079612e-01 -6.21880293e-01 5.48015893e-01
9.61300254e-01 -2.14343980e-01 -6.80890322e-01 -1.65497154e-01
6.84745252e-01 -1.11960769e+00 9.01009142e-01 -6.53738916e-01
1.00741208e-01 4.96144980e-01 -5.71203418e-02 -1.14812958e+00
-2.48299241e-01 -1.11062932e+00 -5.49317859e-02 1.44376707e+00
5.62841773e-01 -8.10836792e-01 8.41405928e-01 1.87305105e+00
-3.49917442e-01 -4.94375885e-01 -5.16725183e-01 -1.05032891e-01
-4.96730924e-01 -4.94619459e-01 4.22114551e-01 6.63197100e-01
8.17044973e-01 6.69985354e-01 -1.91074044e-01 -2.07862571e-01
1.62207156e-01 -2.77971953e-01 9.97539341e-01 -1.29494047e+00
-1.37611002e-01 -5.48465848e-01 4.66836721e-01 -1.57093823e+00
2.34173071e-02 -4.45830047e-01 1.54872879e-01 -1.83993530e+00
-2.04854086e-02 -7.24776566e-01 4.69478518e-01 2.51880050e-01
-2.07924739e-01 -2.87571549e-01 -2.59543471e-02 3.95351529e-01
-9.35033381e-01 6.40316844e-01 9.14759040e-01 1.66327834e-01
-6.36064947e-01 4.44620162e-01 -1.38603449e+00 3.82598042e-01
7.85975337e-01 -1.57732461e-02 -5.90916574e-01 -1.89473815e-02
3.61600369e-01 3.85550261e-01 -3.82321291e-02 -3.83080721e-01
8.01290572e-01 -3.26279640e-01 -4.94831115e-01 -4.76341397e-01
7.64084518e-01 -7.73591995e-01 -2.85488844e-01 2.20206395e-01
-7.16920257e-01 -1.15297720e-01 2.29808763e-02 2.05013469e-01
-4.39692698e-02 -5.88928759e-01 1.08387075e-01 -1.72368780e-01
-7.50844538e-01 -7.12472796e-02 -9.19775844e-01 2.74241865e-01
6.92511141e-01 4.82182018e-02 -8.37844789e-01 -1.25576723e+00
-1.70784727e-01 9.06817615e-01 1.68104410e-01 6.51428163e-01
4.56601441e-01 -9.87079740e-01 -3.07764888e-01 1.10505737e-01
4.02232766e-01 -5.02995372e-01 1.84493378e-01 2.74015576e-01
-3.54653865e-01 9.56387401e-01 2.15275779e-01 -2.66228825e-01
-1.41239417e+00 -4.52424400e-02 1.02613166e-01 -1.41391098e-01
-2.10503474e-01 6.10909998e-01 -3.59095842e-01 -6.86303496e-01
7.47397184e-01 1.39719486e-01 -5.46456337e-01 1.42612696e-01
8.06935012e-01 6.61956906e-01 1.73343465e-01 1.17458746e-01
-2.10051283e-01 -2.18957081e-01 -2.92964041e-01 -4.86077547e-01
8.06371689e-01 -6.69820547e-01 -2.19304115e-01 2.32129782e-01
1.08103430e+00 2.14916486e-02 -4.32284296e-01 -2.82596231e-01
2.20957890e-01 -5.65021634e-01 -4.11010049e-02 -1.16051638e+00
-1.24688439e-01 1.40044034e-01 1.51833922e-01 1.02170849e+00
7.35154986e-01 1.78339347e-01 1.12278616e+00 9.18879449e-01
3.50565940e-01 -1.23765337e+00 5.68024457e-01 7.44386554e-01
1.01281452e+00 -1.39323413e+00 -4.58694309e-01 -3.22563857e-01
-9.21305954e-01 7.75112748e-01 9.48734343e-01 5.08143067e-01
5.90416849e-01 -8.89428556e-02 2.77484983e-01 -5.25751352e-01
-1.30053782e+00 -1.22618474e-01 3.99952345e-02 3.45108300e-01
7.63950288e-01 -2.27220595e-01 -4.46800619e-01 6.28592730e-01
-2.43326530e-01 7.24455789e-02 2.69625664e-01 1.09032512e+00
-7.37985373e-01 -1.14629030e+00 7.84583576e-03 5.65524995e-01
-1.19661734e-01 -2.88157701e-01 -8.14004898e-01 4.17956531e-01
-8.16015124e-01 1.90056634e+00 -1.47409439e-01 -5.03567517e-01
7.24875987e-01 4.32780206e-01 -2.82801211e-01 -6.54749572e-01
-1.08551490e+00 -5.34239650e-01 1.05410469e+00 -9.13132370e-01
-2.41208281e-02 -5.62994957e-01 -1.00998676e+00 -7.18391955e-01
-4.94247049e-01 8.72022986e-01 8.58773649e-01 8.65111172e-01
9.43024814e-01 2.32637655e-02 9.96087670e-01 -2.56331801e-01
-6.22059166e-01 -1.14546335e+00 1.16681503e-02 4.66214150e-01
2.64649957e-01 -3.69746715e-01 -5.58003306e-01 -5.27840555e-01] | [12.273353576660156, 7.785153865814209] |
f9c0b865-3a59-40f5-80ae-2cb4e647792d | digging-into-self-supervised-learning-of | 2110.04773 | null | https://arxiv.org/abs/2110.04773v1 | https://arxiv.org/pdf/2110.04773v1.pdf | Digging Into Self-Supervised Learning of Feature Descriptors | Fully-supervised CNN-based approaches for learning local image descriptors have shown remarkable results in a wide range of geometric tasks. However, most of them require per-pixel ground-truth keypoint correspondence data which is difficult to acquire at scale. To address this challenge, recent weakly- and self-supervised methods can learn feature descriptors from relative camera poses or using only synthetic rigid transformations such as homographies. In this work, we focus on understanding the limitations of existing self-supervised approaches and propose a set of improvements that combined lead to powerful feature descriptors. We show that increasing the search space from in-pair to in-batch for hard negative mining brings consistent improvement. To enhance the discriminativeness of feature descriptors, we propose a coarse-to-fine method for mining local hard negatives from a wider search space by using global visual image descriptors. We demonstrate that a combination of synthetic homography transformation, color augmentation, and photorealistic image stylization produces useful representations that are viewpoint and illumination invariant. The feature descriptors learned by the proposed approach perform competitively and surpass their fully- and weakly-supervised counterparts on various geometric benchmarks such as image-based localization, sparse feature matching, and image retrieval. | ['Juho Kannala', 'Shuzhe Wang', 'Xiaotian Li', 'Zakaria Laskar', 'Iaroslav Melekhov'] | 2021-10-10 | null | null | null | null | ['image-based-localization', 'image-stylization'] | ['computer-vision', 'computer-vision'] | [ 2.73437172e-01 -3.78650904e-01 -6.27936244e-01 -5.15858114e-01
-1.17427719e+00 -6.17540121e-01 6.97489858e-01 7.29024708e-02
-2.98956662e-01 4.37795341e-01 8.79036933e-02 2.81740367e-01
-3.26803148e-01 -6.54617965e-01 -9.40062344e-01 -6.39296293e-01
1.86380997e-01 3.34170580e-01 4.77607697e-02 -2.93365657e-01
5.34670353e-01 8.36749494e-01 -1.77546561e+00 7.71988630e-02
5.34317136e-01 1.06829977e+00 -1.06604293e-01 2.72482306e-01
7.51232430e-02 4.37369049e-01 -1.75252303e-01 -7.26106092e-02
7.80661523e-01 -2.57997334e-01 -7.03688085e-01 2.20364451e-01
1.07545805e+00 -1.84010297e-01 -4.53975946e-01 1.07910490e+00
3.75280857e-01 2.72083223e-01 5.17150581e-01 -1.30652881e+00
-9.29830551e-01 -4.44673114e-02 -7.30320930e-01 -5.00294790e-02
4.73505735e-01 4.63198731e-03 1.27574122e+00 -1.12858367e+00
8.87980402e-01 1.11717045e+00 6.73183620e-01 2.27854699e-01
-1.33674777e+00 -5.98634720e-01 1.19953290e-01 1.95418730e-01
-1.75993907e+00 -5.00623167e-01 1.13739526e+00 -2.46016204e-01
1.03251445e+00 2.28682622e-01 4.77569640e-01 7.69373536e-01
-1.69402704e-01 6.28852546e-01 1.16556931e+00 -6.77387118e-01
7.11932257e-02 5.43786678e-04 -1.82168439e-01 1.11535823e+00
7.80245662e-02 3.85824591e-01 -8.44902575e-01 -3.17962646e-01
1.10843587e+00 4.70621794e-01 -1.57930627e-01 -1.04531312e+00
-1.39867246e+00 1.08877945e+00 8.08960736e-01 1.89335257e-01
-1.58147067e-01 1.58818215e-01 1.47048548e-01 5.19665420e-01
5.57025254e-01 8.59832287e-01 -5.41168571e-01 6.53078258e-02
-9.79950964e-01 -3.10331099e-02 6.56355619e-01 1.17212009e+00
1.27046120e+00 -1.62151143e-01 3.63595933e-02 7.78956056e-01
-2.73282565e-02 7.30660915e-01 4.63659286e-01 -1.02026033e+00
1.33073315e-01 7.64424026e-01 -1.08112946e-01 -1.43731892e+00
-2.05289140e-01 -3.06885839e-01 -8.11791956e-01 1.16162524e-01
3.14761400e-01 4.95494157e-01 -9.61543083e-01 1.56190121e+00
3.01136255e-01 -9.27942544e-02 -2.37863824e-01 9.70204651e-01
5.37077427e-01 3.40188026e-01 -3.73414725e-01 1.53276727e-01
9.03986156e-01 -1.14699578e+00 -1.97151497e-01 -1.89696699e-01
5.73532403e-01 -9.01027858e-01 1.04898047e+00 3.05432975e-01
-8.64492238e-01 -5.97548127e-01 -1.00356209e+00 -2.91900814e-01
-4.25397158e-01 2.29861349e-01 7.14448929e-01 3.35536838e-01
-1.04711080e+00 5.87562263e-01 -7.34044850e-01 -6.46568120e-01
4.40126508e-01 5.24444818e-01 -8.30734611e-01 -3.87532353e-01
-6.11951053e-01 7.15598285e-01 -3.01662497e-02 -1.39940143e-01
-6.72476709e-01 -5.68811357e-01 -1.14844036e+00 -1.50399983e-01
4.21877563e-01 -5.37258089e-01 7.05447137e-01 -9.12808239e-01
-1.25061560e+00 1.15638435e+00 -1.32928416e-01 -1.67474434e-01
3.16463947e-01 -2.58349001e-01 5.15622869e-02 2.88286179e-01
2.50902236e-01 9.33193862e-01 1.11342645e+00 -1.17188203e+00
-4.85395074e-01 -5.94663262e-01 3.05028213e-03 2.03063175e-01
-5.18492997e-01 -5.72877303e-02 -6.43362343e-01 -6.47465348e-01
4.01198268e-01 -1.08214378e+00 -2.42338553e-01 5.03109813e-01
-3.00464272e-01 -8.37942958e-02 8.25033545e-01 -1.52229473e-01
5.28305233e-01 -2.09878159e+00 1.36804834e-01 5.06567895e-01
1.67020649e-01 3.99843417e-02 -4.12151784e-01 4.27754343e-01
-6.25722706e-02 -1.61876723e-01 8.70397538e-02 -3.07123065e-01
-4.83804941e-02 2.84029990e-01 -3.44732285e-01 8.79766107e-01
4.02887553e-01 1.18375361e+00 -8.18229020e-01 -4.84721333e-01
5.17264962e-01 4.09217000e-01 -4.71552968e-01 4.54188734e-02
-5.88686578e-02 3.90901536e-01 -4.37579811e-01 9.00525868e-01
5.29341817e-01 -4.63851601e-01 -3.20131481e-01 -4.14632708e-01
-4.13580276e-02 -2.46529952e-02 -1.17729223e+00 2.28261042e+00
-5.25098085e-01 6.58509195e-01 -3.45422953e-01 -1.26097381e+00
1.18267298e+00 -3.02180111e-01 4.62262183e-01 -8.51873040e-01
-2.49794289e-03 2.36443490e-01 -5.56727111e-01 -1.33484930e-01
3.29337716e-01 1.07288398e-01 1.91794448e-02 3.55139852e-01
2.26189166e-01 -2.54233599e-01 2.32346542e-02 4.43517417e-02
1.07284796e+00 2.25647315e-01 5.31516731e-01 -2.01487482e-01
5.59351265e-01 1.07980475e-01 4.30213302e-01 8.96117330e-01
1.03168547e-01 9.17981148e-01 8.56550261e-02 -6.78508878e-01
-1.09473431e+00 -9.00234044e-01 -6.07854910e-02 1.11385632e+00
4.42492723e-01 -4.01569754e-01 -2.68314183e-01 -6.36517048e-01
1.69695869e-01 -1.33144900e-01 -7.15688944e-01 -3.07726949e-01
-4.25703138e-01 -2.76353240e-01 3.38990331e-01 6.88579440e-01
5.55826902e-01 -7.42105186e-01 -3.61134917e-01 -1.98211551e-01
2.35652000e-01 -1.36150718e+00 -6.52263522e-01 2.86633939e-01
-7.63349533e-01 -1.09589708e+00 -7.29900777e-01 -1.01432776e+00
1.04876256e+00 6.91690624e-01 9.22693431e-01 1.67419806e-01
-6.81794107e-01 7.15645969e-01 -4.98865932e-01 9.92272869e-02
7.33583421e-02 1.70983523e-01 8.43053982e-02 1.06948793e-01
3.81664753e-01 -5.97421408e-01 -6.02006376e-01 5.18900990e-01
-7.76423156e-01 -1.60708964e-01 9.07788932e-01 1.13833773e+00
1.01149380e+00 -4.00768131e-01 4.51026112e-02 -6.69012725e-01
1.61425531e-01 2.96125188e-03 -7.80702531e-01 3.92240584e-01
-7.20408618e-01 5.09449184e-01 5.68021417e-01 -4.52469260e-01
-5.53182781e-01 6.11351550e-01 1.89605504e-01 -8.21352959e-01
-1.69555560e-01 2.60541558e-01 1.66788504e-01 -9.87093210e-01
7.87917972e-01 4.70753074e-01 1.33841470e-01 -2.75688201e-01
6.21243298e-01 2.18722463e-01 5.43349445e-01 -5.78645289e-01
1.45529830e+00 9.38489735e-01 2.98477203e-01 -7.44694948e-01
-1.02034676e+00 -1.04823732e+00 -9.60731685e-01 2.39134446e-01
6.05126381e-01 -1.12593365e+00 -3.91545892e-01 2.82747984e-01
-1.00030732e+00 -5.84842376e-02 -5.39711297e-01 3.73644799e-01
-6.51894629e-01 4.81163830e-01 -1.79847971e-01 -2.02534899e-01
-2.71141529e-01 -9.95944262e-01 1.51009321e+00 1.10814393e-01
5.24736233e-02 -8.90234411e-01 2.93726146e-01 2.64034122e-01
4.86026287e-01 3.05754572e-01 4.85096216e-01 -5.21472871e-01
-9.32593048e-01 -4.82327789e-01 -4.34830934e-01 4.95944291e-01
2.48641908e-01 -2.99580336e-01 -8.62238646e-01 -4.84763712e-01
-3.50269198e-01 -7.93350756e-01 9.49347198e-01 1.77370496e-02
1.14563668e+00 -1.15897596e-01 -3.07093620e-01 1.17330718e+00
1.55537021e+00 -3.73222381e-01 3.32303822e-01 4.41117525e-01
8.59122097e-01 3.86692315e-01 7.03701973e-01 3.81446987e-01
2.53170967e-01 8.49686980e-01 3.88666034e-01 -6.05715930e-01
-4.34625186e-02 -3.36749494e-01 9.63848904e-02 4.67965394e-01
-3.27208415e-02 3.43822956e-01 -6.83617294e-01 5.27611911e-01
-1.81514549e+00 -6.99342906e-01 3.20799589e-01 2.16865611e+00
7.56264210e-01 -1.64125651e-01 -1.28193498e-01 -5.09047173e-02
4.76039797e-01 2.31768161e-01 -5.76402664e-01 4.46927622e-02
-4.54395115e-01 3.81929725e-01 7.53325462e-01 2.23031938e-01
-1.26598179e+00 1.16113687e+00 5.45884943e+00 6.85852170e-01
-1.29409277e+00 -4.99207992e-03 4.43695575e-01 1.44031823e-01
-1.76960200e-01 1.37801364e-01 -7.34243035e-01 -1.89218372e-01
1.74306795e-01 9.37044621e-02 5.14783621e-01 1.16824782e+00
-3.48973840e-01 4.71615372e-03 -1.36323690e+00 1.51165044e+00
7.42842317e-01 -1.60558832e+00 1.87859014e-01 1.51371166e-01
1.21878910e+00 2.85273552e-01 1.83586910e-01 -6.18267767e-02
3.48094553e-02 -1.06312823e+00 4.05678958e-01 5.18203855e-01
9.55313206e-01 -6.23601437e-01 5.16041756e-01 1.02812640e-01
-1.16970968e+00 1.92879457e-02 -4.97388244e-01 1.45346522e-02
-2.92465895e-01 3.24447364e-01 -5.02465069e-01 5.52732229e-01
7.48064339e-01 9.61270928e-01 -9.30380344e-01 9.68271792e-01
-1.06326684e-01 3.53487916e-02 -5.65829456e-01 3.43910865e-02
2.50297576e-01 -1.51958000e-02 2.40629643e-01 9.91839588e-01
1.39017895e-01 -9.85537395e-02 4.21819001e-01 9.28491652e-01
-1.25230134e-01 1.93205819e-01 -9.55136895e-01 8.67504403e-02
3.53859156e-01 1.40036070e+00 -7.00557768e-01 2.67842021e-02
-6.06356263e-01 1.22387826e+00 4.68966514e-01 2.08131343e-01
-3.79779339e-01 -4.37613606e-01 5.97368240e-01 1.31889224e-01
4.33242679e-01 -5.59014916e-01 -8.21127221e-02 -1.33104467e+00
1.09818116e-01 -8.10786724e-01 1.66377798e-01 -5.78544855e-01
-1.31575215e+00 4.38839376e-01 -4.41656113e-02 -1.39316523e+00
-3.47655922e-01 -7.18473256e-01 -4.81297731e-01 5.05299389e-01
-1.85525620e+00 -1.57558608e+00 -7.64706790e-01 1.07320547e+00
3.43770534e-01 -2.77523220e-01 9.14313257e-01 1.66280955e-01
-7.74478912e-03 8.60784829e-01 3.15359354e-01 3.17935646e-01
1.05483699e+00 -1.06428015e+00 3.67542654e-01 5.83641231e-01
7.54491985e-01 7.23455846e-01 1.82682648e-01 -1.71588972e-01
-1.76468515e+00 -1.06893528e+00 6.31022632e-01 -4.72669035e-01
5.29962718e-01 -5.48281372e-01 -5.62449098e-01 3.94467384e-01
-2.68419206e-01 7.21099615e-01 4.41232175e-01 7.66520873e-02
-7.68462002e-01 -3.84888649e-01 -9.18616295e-01 3.52005571e-01
1.24671042e+00 -9.40554738e-01 -2.82152981e-01 6.84656322e-01
4.01778489e-01 -3.66307527e-01 -7.80244470e-01 3.43285352e-01
5.05241334e-01 -7.94788837e-01 1.24860585e+00 -5.29850602e-01
2.09022298e-01 -3.29619199e-01 -3.73181492e-01 -1.02943134e+00
-2.62786001e-01 -7.68376350e-01 2.92385638e-01 9.83941913e-01
9.53320861e-02 -4.99734491e-01 8.46835852e-01 1.48482934e-01
1.18770413e-01 -6.61561430e-01 -8.61829281e-01 -1.04602933e+00
-7.52345249e-02 -9.26737860e-02 3.75629604e-01 1.01578319e+00
-3.02191406e-01 1.32902324e-01 -3.37184548e-01 8.18413496e-02
7.54617691e-01 5.17930686e-01 1.06174171e+00 -1.24758303e+00
-1.37720153e-01 -1.41331464e-01 -1.06907976e+00 -1.24053097e+00
3.62693518e-01 -9.30852950e-01 2.15984136e-02 -1.07650220e+00
4.14245367e-01 -4.31222379e-01 -2.61656672e-01 6.69032037e-01
2.04703346e-01 7.38381207e-01 1.84050631e-02 4.70756263e-01
-9.50653076e-01 6.30586684e-01 1.10431147e+00 -2.25781232e-01
2.73830700e-03 -3.32847714e-01 -5.86637735e-01 6.05096161e-01
6.09634578e-01 -4.09280866e-01 -2.15750292e-01 -4.54418421e-01
2.46564418e-01 -2.58222818e-01 7.25514770e-01 -9.07141209e-01
6.00687146e-01 -1.83738306e-01 4.91715133e-01 -3.13505530e-01
4.09801364e-01 -6.32172287e-01 -3.79601508e-01 1.28564775e-01
-4.12818342e-01 1.91343427e-01 -1.13950014e-01 7.90769100e-01
-4.93219793e-01 2.88696066e-02 7.43848085e-01 -1.60027236e-01
-8.77937555e-01 5.18752754e-01 4.12462324e-01 1.15227416e-01
9.28818107e-01 -4.34752822e-01 -1.55946016e-01 -5.23817182e-01
-3.81075263e-01 -9.39479694e-02 8.02653253e-01 5.37544310e-01
9.10080791e-01 -1.36450350e+00 -4.83961642e-01 5.57322979e-01
6.40302241e-01 1.37422420e-02 -1.56749114e-01 7.77998149e-01
-5.50078154e-01 6.01464033e-01 -4.02260661e-01 -8.93285871e-01
-1.25277174e+00 5.13103545e-01 1.97249368e-01 -8.48965570e-02
-5.76373816e-01 8.28006327e-01 2.56430179e-01 -5.85765839e-01
3.57291520e-01 -2.81448752e-01 3.29414994e-01 -2.65478551e-01
1.90435931e-01 5.40131032e-02 2.16459379e-01 -7.96925306e-01
-4.24444586e-01 1.20226169e+00 -1.74459100e-01 1.21768847e-01
1.43206489e+00 -1.06162742e-01 -1.08057693e-01 1.15101971e-01
1.74682426e+00 2.47494387e-03 -1.22492576e+00 -7.33946502e-01
4.44210209e-02 -7.69626200e-01 2.02436224e-01 -4.99534577e-01
-1.13940954e+00 7.67592549e-01 7.09073782e-01 -2.85407245e-01
1.13433290e+00 3.73670787e-01 6.16850436e-01 9.81352568e-01
5.51386952e-01 -1.07199347e+00 4.27819282e-01 5.35293221e-01
8.88607323e-01 -1.75437009e+00 2.09548458e-01 -3.71989191e-01
-4.00445789e-01 1.21122062e+00 4.96318877e-01 -5.30417264e-01
6.90274477e-01 -1.15045838e-01 -4.01101634e-02 -4.05148327e-01
-4.14802790e-01 -3.90023649e-01 7.49077857e-01 6.55763865e-01
6.78219870e-02 -3.26441765e-01 1.78487733e-01 -5.81015460e-02
-6.66730404e-02 -2.98501164e-01 8.10613332e-04 1.00643945e+00
-2.40034580e-01 -1.23097324e+00 -2.89332569e-01 2.62531340e-01
-2.69865617e-02 -1.66924268e-01 -6.01643384e-01 9.55738544e-01
-7.28567764e-02 6.24153793e-01 2.62575764e-02 -3.57059568e-01
2.91884959e-01 -3.24977398e-01 8.02685142e-01 -6.32031739e-01
-2.40912795e-01 -4.08761986e-02 -3.26488167e-01 -9.01495099e-01
-6.35176778e-01 -5.92564583e-01 -9.22420084e-01 2.91173439e-02
-5.71662962e-01 -1.46599963e-01 6.08074903e-01 9.74603593e-01
3.81994188e-01 -1.45330012e-01 9.01303947e-01 -1.07741368e+00
-8.26973736e-01 -6.21812582e-01 -4.28753823e-01 6.83729768e-01
4.33132082e-01 -7.72661686e-01 -4.96910125e-01 7.74709880e-02] | [8.07407283782959, -2.062380313873291] |
3d3a91fb-5619-457e-8ca0-5ced17f8f28d | interactivism-in-spoken-dialogue-systems | 2209.13547 | null | https://arxiv.org/abs/2209.13547v2 | https://arxiv.org/pdf/2209.13547v2.pdf | Interactivism in Spoken Dialogue Systems | The interactivism model introduces a dynamic approach to language, communication and cognition. In this work, we explore this fundamental theory in the context of dialogue modelling for spoken dialogue systems (SDS). To extend such a theoretical framework, we present a set of design principles which adhere to central psycholinguistic and communication theories to achieve interactivism in SDS. From these, key ideas are linked to constitute the basis of our proposed design principles. | ['R. K. Moore', 'G. Huang', 'E. Ip', 'T. Rodríguez Muñoz'] | 2022-09-27 | null | null | null | null | ['spoken-dialogue-systems'] | ['speech'] | [-4.35784101e-01 9.96462822e-01 3.12895000e-01 -4.38103437e-01
4.43911105e-01 -6.11032426e-01 1.27102768e+00 1.66066647e-01
-2.31939822e-01 5.99676728e-01 8.25654089e-01 -6.65613115e-01
-3.30214113e-01 -6.36764348e-01 1.35686725e-01 1.80538997e-01
-8.11689943e-02 2.35375106e-01 -7.68050104e-02 -1.07767820e+00
5.39644003e-01 3.83695722e-01 -1.67745972e+00 3.96046191e-01
9.51766133e-01 7.84669966e-02 2.73222953e-01 9.03228402e-01
-8.22864771e-01 1.34057629e+00 -6.56548202e-01 -1.30135641e-01
-3.88212204e-01 -9.56369340e-01 -1.65890050e+00 3.45717609e-01
-2.84034759e-01 -3.32512677e-01 -3.54536235e-01 6.07841551e-01
4.66264009e-01 3.98675948e-01 6.69610918e-01 -1.07502759e+00
-7.08916545e-01 1.03157091e+00 1.34002551e-01 -9.98258144e-02
1.15387392e+00 1.17976211e-01 8.41142714e-01 -7.73055851e-01
5.27719736e-01 1.74900854e+00 1.66916460e-01 1.02805126e+00
-9.76877868e-01 1.49589881e-01 2.25150168e-01 -7.39443004e-02
-9.31071997e-01 -6.17233634e-01 4.18673575e-01 -3.20793182e-01
1.04243362e+00 7.05497324e-01 1.03013217e+00 5.71685195e-01
2.02567592e-01 9.63937044e-01 1.23449695e+00 -1.20493269e+00
1.84460282e-01 8.39419901e-01 1.00477517e+00 3.76618117e-01
-2.86653221e-01 -2.90577840e-02 -8.74637067e-01 -1.78940937e-01
6.83769226e-01 -5.70032001e-01 -1.55610934e-01 -2.65980303e-01
-7.01204240e-01 1.07855248e+00 -2.72873849e-01 7.42517710e-01
-2.82287419e-01 -4.36371535e-01 5.72938800e-01 8.13585937e-01
4.25768524e-01 4.23415542e-01 -1.13320775e-01 -4.70573246e-01
8.24205577e-03 3.60989988e-01 1.81715286e+00 6.59789801e-01
3.01501334e-01 -2.59450059e-02 -1.23712435e-01 9.68291044e-01
1.18023670e+00 2.30380073e-01 2.76132703e-01 -8.02652121e-01
-3.95914912e-01 6.91488028e-01 2.17594042e-01 -7.21377552e-01
-6.08548582e-01 5.07004499e-01 -2.03511920e-02 5.32734543e-02
1.33240089e-01 -3.38264942e-01 -1.92626804e-01 1.24764180e+00
7.27429032e-01 -7.33010352e-01 7.00655937e-01 7.36642122e-01
1.44095242e+00 6.72177732e-01 1.90369546e-01 -4.00872469e-01
1.32286322e+00 -7.93225527e-01 -1.36470306e+00 2.61290759e-01
1.04371488e+00 -7.25440443e-01 1.22687018e+00 4.40382272e-01
-1.21403670e+00 -4.62454826e-01 -7.28339791e-01 -6.00826107e-02
-3.87693286e-01 -3.86362463e-01 8.12330306e-01 1.11891484e+00
-1.58894312e+00 -1.30300388e-01 -3.53203624e-01 -8.01409245e-01
-9.43319082e-01 1.45682395e-01 1.67267963e-01 6.66448653e-01
-1.48207569e+00 1.44139981e+00 5.62019408e-01 1.39857277e-01
-2.18459994e-01 4.75477576e-02 -8.58044326e-01 -2.90080816e-01
6.91017434e-02 -5.27960539e-01 1.74994576e+00 -9.06236947e-01
-2.41564441e+00 9.45546031e-01 -4.71890457e-02 -3.78135085e-01
2.60250509e-01 -5.79005837e-01 -3.36288482e-01 -9.83671620e-02
-4.20571476e-01 2.86207497e-01 -1.39142141e-01 -1.64321089e+00
-6.20145619e-01 -2.04262689e-01 4.87872899e-01 7.86932349e-01
-3.18365186e-01 4.19959754e-01 -1.72509506e-01 2.48576538e-03
-1.34152234e-01 -5.92453539e-01 -4.84925918e-02 -4.02470350e-01
-7.42345825e-02 -7.02192187e-01 5.49481094e-01 -4.07212734e-01
1.45845580e+00 -1.90245664e+00 1.87499613e-01 1.36849463e-01
6.91070318e-01 4.52206165e-01 2.61649430e-01 1.64871550e+00
3.16256136e-01 -5.95345162e-02 2.85971224e-01 -1.01979688e-01
5.97511709e-01 4.34301853e-01 -6.40853196e-02 -7.63901621e-02
-4.40414160e-01 6.26338124e-01 -7.57614493e-01 -4.03032720e-01
8.97929966e-01 4.24551725e-01 -4.82856840e-01 7.70220459e-01
-3.02353412e-01 4.63343203e-01 -6.77987635e-01 -1.20507769e-01
4.26302880e-01 4.12366956e-01 5.73068380e-01 5.89574575e-01
-8.13302338e-01 6.77140653e-01 -1.05586898e+00 1.40446293e+00
-5.14047682e-01 4.22925621e-01 4.55542594e-01 -7.19648600e-01
1.20855451e+00 6.79223657e-01 1.57757491e-01 -5.89417040e-01
2.53474414e-01 1.78259641e-01 3.69570047e-01 -1.09490359e+00
7.16394961e-01 -3.15023540e-03 -7.16449842e-02 1.04788888e+00
1.02019288e-01 -6.19630098e-01 -3.38304713e-02 5.25257528e-01
3.63090992e-01 -7.93037489e-02 8.84419084e-01 -6.80366516e-01
1.08254576e+00 -3.21706086e-02 -1.99431926e-01 8.38004410e-01
-3.79569679e-01 -4.47895944e-01 4.73570168e-01 -4.83429164e-01
-6.48694158e-01 -7.71019995e-01 -1.62559152e-01 1.24457121e+00
3.82593833e-02 -5.30243754e-01 -1.32725036e+00 -4.02166188e-01
-4.94389862e-01 1.21536541e+00 -1.82489738e-01 2.01674804e-01
-1.45725355e-01 -1.70810968e-01 6.03207588e-01 -2.13464022e-01
5.26122510e-01 -1.32373714e+00 -7.57518232e-01 1.63122728e-01
5.93976229e-02 -6.54217005e-01 1.90130934e-01 -9.59679037e-02
-4.19021606e-01 -8.46250772e-01 -6.79243803e-02 -7.57993758e-01
7.88175389e-02 2.51871884e-01 1.05563378e+00 4.92820054e-01
9.28360000e-02 1.24353266e+00 -9.87902582e-01 -7.54709303e-01
-1.16118491e+00 -3.00297111e-01 5.22126779e-02 -3.86328369e-01
6.62505031e-01 -2.82207608e-01 -1.95955426e-01 4.24459018e-02
-9.63422835e-01 1.89588189e-01 -2.27012217e-01 7.57526457e-01
-5.05322039e-01 -2.40326568e-01 4.41550314e-01 -8.93095732e-01
1.73336434e+00 -3.68937135e-01 -4.36924174e-02 3.30832660e-01
-5.41715026e-01 -1.42613679e-01 1.47864684e-01 1.11111067e-01
-1.44597471e+00 -5.03199518e-01 -4.04389799e-01 8.42719674e-01
-6.60524845e-01 7.97404289e-01 -2.57590830e-01 -3.37885380e-01
7.25374162e-01 3.62046808e-01 6.92745268e-01 -1.94546655e-01
8.43457520e-01 1.15627980e+00 -1.58868760e-01 -8.24664116e-01
-3.00463349e-01 9.60645005e-02 -6.59869492e-01 -1.48110211e+00
-2.52424460e-02 -2.88421571e-01 -7.63055086e-01 -8.12083840e-01
6.36331201e-01 -5.51307142e-01 -1.29087961e+00 4.27374512e-01
-1.34078884e+00 -5.16855121e-01 -3.66191596e-01 6.08448923e-01
-7.61535823e-01 5.31347930e-01 -6.92702830e-01 -1.80533719e+00
-2.76730418e-01 -9.25095260e-01 4.90473270e-01 4.57371950e-01
-8.97429347e-01 -1.73120630e+00 2.00830504e-01 3.05796385e-01
2.42803439e-01 -6.65085912e-02 8.25121880e-01 -1.01344371e+00
2.64762402e-01 4.44810033e-01 2.55208373e-01 1.25608787e-01
1.92980111e-01 3.31852287e-01 -7.16096401e-01 3.13668609e-01
7.02771425e-01 -5.26561975e-01 -9.12905410e-02 1.85546484e-02
1.96333956e-02 -3.55725318e-01 -1.03077181e-01 -4.89828825e-01
1.03082466e+00 6.42207861e-01 2.95509547e-01 1.89916909e-01
2.25612849e-01 1.68740535e+00 7.02307045e-01 7.10417092e-01
9.85077918e-01 6.98454261e-01 -2.49399945e-01 -6.78928569e-02
2.02822030e-01 8.17565247e-02 3.58117074e-01 1.55711198e+00
1.49718523e-01 -3.49160850e-01 -1.40410697e+00 2.68676192e-01
-2.03281260e+00 -5.95167398e-01 -6.28697991e-01 1.67114568e+00
8.78818095e-01 -2.62149960e-01 4.69147116e-01 8.62932485e-03
4.32093561e-01 2.12242931e-01 3.35513026e-01 -1.46456337e+00
2.35438451e-01 -7.30002746e-02 -4.21779543e-01 1.20336139e+00
-3.29311907e-01 1.25933647e+00 7.62487364e+00 2.41768271e-01
-6.67180777e-01 3.94930094e-02 1.12462990e-01 4.24780697e-01
-5.82364798e-01 -4.74756816e-03 -5.89392006e-01 -2.28737714e-03
1.26988018e+00 -4.36692655e-01 3.75397652e-01 2.52627403e-01
9.32566822e-01 -5.24003625e-01 -9.25699711e-01 4.42646235e-01
1.84797376e-01 -8.97790372e-01 1.92906648e-01 -2.90532783e-03
1.00798182e-01 -5.93798339e-01 -5.05645424e-02 1.71197742e-01
7.88907170e-01 -9.07710791e-01 5.04212797e-01 6.12507105e-01
-1.49938643e-01 -7.66876221e-01 4.59527493e-01 7.25171089e-01
-6.21395469e-01 5.70306629e-02 -1.25555560e-01 -7.27421463e-01
4.84529883e-01 -1.24760374e-01 -7.95434713e-01 5.73911011e-01
3.42721432e-01 4.39243279e-02 1.77653953e-01 5.43423533e-01
-1.47635370e-01 3.60501647e-01 -4.47468236e-02 -6.31359160e-01
3.94222260e-01 -5.01736820e-01 6.37354970e-01 1.30960011e+00
-3.72482121e-01 6.71084821e-01 1.92327589e-01 8.05679560e-01
8.14910054e-01 6.84897006e-01 -9.38039482e-01 7.78690204e-02
7.17080355e-01 8.43214869e-01 -6.46037221e-01 -2.04598591e-01
-6.23679757e-01 6.60189390e-01 9.31400433e-02 5.48871793e-02
-3.01836997e-01 7.72139477e-03 4.13089752e-01 -2.18568504e-01
-3.88160616e-01 -5.17724752e-01 -4.70668703e-01 -6.97554171e-01
-5.64013004e-01 -1.31035018e+00 -1.03261404e-01 -2.15715319e-01
-1.00883317e+00 5.18845856e-01 4.64610964e-01 -4.37905878e-01
-6.45129740e-01 -5.09370029e-01 -6.85870588e-01 7.54120231e-01
-7.08268821e-01 -1.32228220e+00 -2.56499294e-02 4.21801597e-01
6.36753321e-01 -4.58241887e-02 1.36199021e+00 -2.49922067e-01
-2.47247785e-01 -2.17862846e-03 -8.35131388e-03 -2.96440631e-01
3.81689548e-01 -1.54549539e+00 6.30478442e-01 1.45430014e-01
5.49169118e-03 1.04281580e+00 1.09299684e+00 -4.68467504e-01
-1.46449244e+00 1.53467683e-02 1.19937122e+00 -4.27293301e-01
7.19677031e-01 -3.03781092e-01 -9.06626046e-01 3.16067517e-01
9.84815955e-01 -1.07623458e+00 1.20918560e+00 3.15292746e-01
3.59331101e-01 7.59281576e-01 -9.13832605e-01 8.75784278e-01
8.74686003e-01 -7.84138978e-01 -1.19420111e+00 1.05590716e-01
7.28371203e-01 -4.83774215e-01 -1.03889823e+00 -1.15404725e-01
5.88702202e-01 -1.26512301e+00 6.28245950e-01 -3.17211032e-01
-2.11638864e-02 1.92957714e-01 5.02753742e-02 -1.29303730e+00
2.92131722e-01 -1.10370183e+00 -1.16168978e-02 1.29784322e+00
2.80440778e-01 -9.63509858e-01 1.76136240e-01 1.34966087e+00
-2.27709219e-01 -2.62077481e-01 -5.09609640e-01 -2.35776260e-01
4.59960252e-01 -4.73196328e-01 4.50838208e-01 8.71196926e-01
1.25138807e+00 7.52426982e-01 -4.33574706e-01 -3.00039232e-01
-3.74802649e-02 -5.65807700e-01 1.10578263e+00 -1.30694199e+00
-2.06550047e-01 -4.09992725e-01 -1.41427249e-01 -1.45005274e+00
2.23045334e-01 -4.83561307e-02 1.56558335e-01 -1.78337502e+00
-4.31618333e-01 1.99069045e-02 3.74106169e-01 -1.74575031e-01
1.28463954e-01 -7.00143456e-01 3.31065834e-01 9.47043002e-02
-4.99502718e-01 7.41430342e-01 1.12481892e+00 4.73937482e-01
-9.57282603e-01 -1.72763392e-01 -7.95408010e-01 7.80209839e-01
8.54108095e-01 1.22157142e-01 -8.81927907e-01 -1.39652520e-01
1.10430336e-02 4.73874927e-01 9.68244331e-06 -3.82501334e-01
4.36769694e-01 -2.95689255e-01 -5.55643916e-01 -2.79153347e-01
8.12621936e-02 -5.92709422e-01 -1.15736157e-01 6.69739842e-01
-7.46460736e-01 -1.69142425e-01 3.25352997e-01 1.85019001e-01
-4.14849102e-01 -3.23276639e-01 3.66053551e-01 2.66374610e-02
-7.10544348e-01 -8.31043839e-01 -1.46586740e+00 -2.91231334e-01
1.19116688e+00 -6.21979713e-01 1.84352584e-02 -8.53442073e-01
-9.39397871e-01 4.39636558e-01 2.87155628e-01 4.81428772e-01
5.81316113e-01 -6.23731971e-01 -5.41030884e-01 5.10501228e-02
-9.98077616e-02 -4.47813421e-01 2.55607933e-01 4.68949199e-01
-7.25969970e-01 9.89531219e-01 -3.52318525e-01 -2.97789097e-01
-1.49270737e+00 1.01246476e-01 3.79621029e-01 2.43836418e-01
-4.33176428e-01 6.12982750e-01 3.73843759e-01 -9.52169955e-01
2.91394264e-01 1.57074124e-01 -1.14757729e+00 -8.07763413e-02
5.42274654e-01 4.36907709e-01 -4.54050690e-01 -8.62062573e-01
-1.49037717e-02 -2.76658218e-02 -2.97010750e-01 -1.08166409e+00
7.85291016e-01 -1.01552629e+00 -5.70374429e-01 1.02575469e+00
3.61509323e-01 -1.04779266e-01 -5.36907136e-01 -2.45760471e-01
4.40924466e-01 -1.51640624e-01 -1.84091985e-01 -6.63896024e-01
5.78239709e-02 7.24000216e-01 5.38585663e-01 1.23398149e+00
5.09719014e-01 -1.03041694e-01 1.61994800e-01 5.83820105e-01
2.78392762e-01 -1.46056855e+00 -1.22865334e-01 9.04888213e-01
9.90503967e-01 -7.75053144e-01 -3.16318810e-01 -7.14796305e-01
-1.17992842e+00 1.32393479e+00 8.39404643e-01 3.06944877e-01
9.37943518e-01 3.10001403e-01 4.26952958e-01 -7.87909210e-01
-9.25297797e-01 -2.04462945e-01 -5.20252399e-02 9.35500562e-01
1.31520271e+00 1.61081225e-01 -1.43509245e+00 3.71303439e-01
-3.94942433e-01 -1.94452375e-01 6.38980210e-01 1.20687258e+00
-7.63951182e-01 -1.45789528e+00 -5.04594624e-01 -2.69873977e-01
-1.58520788e-01 1.64570123e-01 -1.41399586e+00 1.03622067e+00
-3.21769744e-01 1.72027147e+00 -1.97363153e-01 -4.24503058e-01
5.07638097e-01 8.04088190e-02 1.09453499e-01 -7.99849927e-01
-1.28193986e+00 7.79536441e-02 6.33335173e-01 -1.72906354e-01
-8.07948112e-01 -4.99286920e-01 -1.43199801e+00 -7.41330147e-01
-4.75691468e-01 7.02194571e-01 5.52155554e-01 1.21438730e+00
4.64034602e-02 2.72542119e-01 6.48261368e-01 -2.66047120e-01
-4.02020514e-01 -1.14197433e+00 -5.78382850e-01 -1.62704289e-01
-1.19830564e-01 -2.78940588e-01 2.30389042e-03 -4.42036949e-02] | [12.948294639587402, 7.931819915771484] |
6f939c37-a5e2-451c-83a9-762e40a81876 | from-explanation-to-synthesis-compositional | 1902.10657 | null | https://arxiv.org/abs/1902.10657v2 | https://arxiv.org/pdf/1902.10657v2.pdf | From explanation to synthesis: Compositional program induction for learning from demonstration | Hybrid systems are a compact and natural mechanism with which to address problems in robotics. This work introduces an approach to learning hybrid systems from demonstrations, with an emphasis on extracting models that are explicitly verifiable and easily interpreted by robot operators. We fit a sequence of controllers using sequential importance sampling under a generative switching proportional controller task model. Here, we parameterise controllers using a proportional gain and a visually verifiable joint angle goal. Inference under this model is challenging, but we address this by introducing an attribution prior extracted from a neural end-to-end visuomotor control model. Given the sequence of controllers comprising a task, we simplify the trace using grammar parsing strategies, taking advantage of the sequence compositionality, before grounding the controllers by training perception networks to predict goals given images. Using this approach, we are successfully able to induce a program for a visuomotor reaching task involving loops and conditionals from a single demonstration and a neural end-to-end model. In addition, we are able to discover the program used for a tower building task. We argue that computer program-like control systems are more interpretable than alternative end-to-end learning approaches, and that hybrid systems inherently allow for better generalisation across task configurations. | ['Svetlin Penkov', 'Michael Burke', 'Subramanian Ramamoorthy'] | 2019-02-27 | null | null | null | null | ['program-induction'] | ['computer-code'] | [ 5.51859796e-01 6.93930387e-01 1.69581264e-01 -1.75490588e-01
-4.30803508e-01 -9.28699553e-01 9.47912574e-01 -1.04089685e-01
-2.53178060e-01 7.09715307e-01 -2.43734811e-02 -5.01767814e-01
-2.73652285e-01 -4.10037726e-01 -1.29561853e+00 -5.10213077e-01
-8.57394189e-02 6.54428542e-01 3.18725824e-01 -4.81733024e-01
4.11257356e-01 4.86301005e-01 -1.71186602e+00 2.68963784e-01
9.21189606e-01 3.72542471e-01 8.33967865e-01 1.16211057e+00
3.06211203e-01 1.18554711e+00 -3.67579281e-01 2.07085073e-01
1.47509158e-01 -7.12086141e-01 -8.68945181e-01 5.46981357e-02
2.82397211e-01 -1.76560789e-01 -4.51644510e-02 8.61211479e-01
-1.09906234e-02 1.13350727e-01 1.05383706e+00 -1.57522869e+00
-5.57911158e-01 8.28289509e-01 1.31784916e-01 -4.29281771e-01
7.02396154e-01 8.77466142e-01 1.19869649e+00 -1.76241726e-01
8.38745177e-01 1.54114115e+00 3.56844306e-01 6.29157007e-01
-1.77731919e+00 -3.19938242e-01 2.80386209e-01 -5.98899685e-02
-1.09301949e+00 -3.07961106e-01 4.38561052e-01 -9.27091718e-01
1.18439674e+00 -9.97849181e-03 7.73446918e-01 1.25852561e+00
2.96665132e-01 9.00643110e-01 1.15200925e+00 -7.22962797e-01
2.80928522e-01 3.57026197e-02 -1.16821878e-01 1.13406646e+00
-1.47868365e-01 7.85560489e-01 -3.13697189e-01 1.83486696e-02
1.22505307e+00 -3.31772357e-01 -2.86333650e-01 -9.79319870e-01
-1.44684052e+00 6.03287697e-01 4.36974555e-01 1.86078772e-01
-4.20420110e-01 6.03834987e-01 2.08087966e-01 5.98850071e-01
-5.62088132e-01 1.04083431e+00 -7.22600341e-01 -3.79756931e-03
-5.12437224e-01 5.11418760e-01 9.82046425e-01 1.34694803e+00
5.53630710e-01 1.02889277e-01 -1.38393611e-01 1.66216284e-01
3.52120668e-01 3.90082926e-01 3.46350998e-01 -1.50718808e+00
7.38995429e-03 3.40312600e-01 1.97428644e-01 -3.70146126e-01
-3.98204029e-01 3.98316905e-02 -1.51790336e-01 1.05313027e+00
5.71997821e-01 -3.50735374e-02 -1.05722833e+00 2.09146047e+00
1.37357358e-02 -2.31599435e-01 2.29643270e-01 7.30077207e-01
-9.66598541e-02 4.57871944e-01 2.01996505e-01 -1.35499313e-01
1.25531256e+00 -8.04112494e-01 -3.56309950e-01 -1.29543349e-01
7.76149213e-01 -1.28008172e-01 1.47757602e+00 8.18862379e-01
-9.61578846e-01 -8.80581081e-01 -1.12215340e+00 1.86454679e-03
-6.88445494e-02 8.88101459e-02 5.65694988e-01 1.20592006e-01
-1.11944449e+00 8.81911039e-01 -1.17590404e+00 -4.62034345e-01
1.08738750e-01 5.94287992e-01 -2.46558800e-01 3.44037324e-01
-7.18994141e-01 1.27772391e+00 9.24774885e-01 -2.95358896e-01
-1.51384902e+00 -3.09859514e-01 -9.97747362e-01 -1.08842075e-01
5.27259171e-01 -9.56816792e-01 1.74038100e+00 -1.15089047e+00
-2.08053493e+00 6.95363820e-01 2.28420913e-01 -5.63902259e-01
3.69893849e-01 3.11882123e-02 1.62421033e-01 7.11300224e-02
7.29902163e-02 1.02590024e+00 1.19299293e+00 -1.33741724e+00
-5.97681284e-01 -1.47320256e-01 5.22242069e-01 9.65965390e-02
3.69311273e-01 -1.53984696e-01 1.40376285e-01 -2.74575830e-01
-1.55934006e-01 -1.46820700e+00 -1.30530044e-01 1.90278992e-01
-3.65506470e-01 -3.41618478e-01 3.43651086e-01 -5.18290997e-01
5.61038554e-01 -2.14785218e+00 1.06111670e+00 3.78341228e-02
1.67047709e-01 -4.31954861e-02 -1.34370103e-01 3.02821070e-01
-1.74632311e-01 -7.54432827e-02 -4.18992758e-01 8.48303586e-02
4.90152180e-01 4.12422001e-01 -3.47041428e-01 1.56035200e-01
4.88835216e-01 8.36384833e-01 -9.42670524e-01 -4.97335225e-01
3.14804435e-01 1.92238465e-01 -8.51027191e-01 7.12378144e-01
-1.08131921e+00 7.94527113e-01 -4.71898973e-01 3.38912904e-02
-2.34086469e-01 7.69025907e-02 4.75536555e-01 2.74989158e-01
-1.36447132e-01 3.00341696e-01 -1.02374244e+00 1.89815819e+00
-7.10750043e-01 5.32471657e-01 1.15052626e-01 -1.06447244e+00
7.48463213e-01 3.06518197e-01 -1.84452280e-01 -1.69450074e-01
1.88806996e-01 2.98595190e-01 4.64051723e-01 -8.24435532e-01
9.78720337e-02 -3.12346965e-01 -2.82283992e-01 3.32588881e-01
4.17403579e-01 -9.78899181e-01 1.26227587e-01 1.12471431e-01
1.02657866e+00 1.33798981e+00 5.50392926e-01 -1.67274997e-01
2.67266184e-01 4.28671777e-01 -6.78921863e-02 7.66950727e-01
-1.74009874e-02 3.27384830e-01 6.93883836e-01 -2.50128537e-01
-1.16130018e+00 -1.15444911e+00 4.09652233e-01 1.22343111e+00
-3.51441026e-01 -1.73035920e-01 -8.02818060e-01 -3.99410307e-01
-1.88759997e-01 1.25911045e+00 -6.97839856e-01 -1.93871617e-01
-7.45913148e-01 2.01242775e-01 4.68406975e-01 4.70462888e-01
-4.60895710e-02 -1.48936558e+00 -1.19232464e+00 2.51929790e-01
4.49288487e-02 -8.81999016e-01 -3.35777551e-01 6.60552502e-01
-7.93801486e-01 -1.12110925e+00 -4.03625041e-01 -1.04619277e+00
5.51040471e-01 -4.23983246e-01 9.81477141e-01 -2.89844722e-02
-1.62283197e-01 6.24247551e-01 -9.60703753e-03 -4.42041278e-01
-1.06386530e+00 -2.76024677e-02 8.80547538e-02 -5.89673698e-01
-2.14943260e-01 -6.29337966e-01 2.18439981e-01 -1.16552450e-01
-7.06210792e-01 4.44242507e-01 7.05511570e-01 9.16969717e-01
3.45940799e-01 -1.70829311e-01 6.90098032e-02 -5.73549151e-01
6.54824495e-01 -3.91851254e-02 -9.95574713e-01 2.32448757e-01
-4.19545203e-01 7.34028935e-01 8.06544721e-01 -8.06950152e-01
-9.24295366e-01 7.90093601e-01 2.73378134e-01 -3.94562304e-01
-6.76487327e-01 3.39154035e-01 -1.95096985e-01 1.97166055e-01
9.88321900e-01 2.50235707e-01 4.11494195e-01 4.59232777e-02
8.40082526e-01 3.58089238e-01 6.26962900e-01 -9.22470689e-01
7.14815497e-01 -2.74815820e-02 1.55716300e-01 -8.62189472e-01
-2.65397549e-01 9.95357260e-02 -9.15437698e-01 -6.53094724e-02
1.05903566e+00 -6.89037204e-01 -1.10558581e+00 2.24019408e-01
-1.27513158e+00 -1.11123097e+00 -3.28073770e-01 5.64992547e-01
-1.65998971e+00 -1.49602396e-02 -4.63813394e-01 -8.66573393e-01
4.65786517e-01 -1.45253587e+00 1.09780240e+00 -1.65965989e-01
-8.58742476e-01 -7.34423220e-01 -2.65162550e-02 -2.64020115e-01
1.57202221e-02 2.73724824e-01 1.29732907e+00 -5.13179481e-01
-8.11739087e-01 1.79470226e-01 1.34767801e-01 1.78267807e-01
-1.55799747e-01 1.35315806e-01 -6.87473476e-01 -1.57749251e-01
5.49710952e-02 -6.14534438e-01 3.46522093e-01 2.68817753e-01
8.14592600e-01 -2.28317022e-01 -4.88983512e-01 2.14757964e-01
1.28638756e+00 2.23919868e-01 5.42609036e-01 2.62139231e-01
7.15705216e-01 9.34100509e-01 4.54211086e-01 -1.60825014e-01
2.87616640e-01 7.32061684e-01 5.54696918e-01 5.92563927e-01
-1.04582824e-01 -5.41815519e-01 7.85317421e-01 3.36226851e-01
-8.58865827e-02 7.57114515e-02 -8.30057681e-01 4.74577904e-01
-1.84192562e+00 -9.14020240e-01 -3.61505710e-02 1.96801436e+00
9.69111860e-01 4.61870402e-01 3.69639456e-01 -8.34260136e-03
5.52288175e-01 -4.08141524e-01 -5.62581420e-01 -5.41040123e-01
3.39196235e-01 2.48369604e-01 2.63719887e-01 7.61154115e-01
-7.04446256e-01 1.15896237e+00 6.30584192e+00 3.00551534e-01
-8.02073598e-01 -2.63544619e-01 -1.88681245e-01 3.42814028e-01
-2.57235710e-02 4.13759679e-01 -5.23911595e-01 1.67120501e-01
1.12972605e+00 1.04297929e-01 9.65380073e-01 8.01188111e-01
1.25948548e-01 -1.35681152e-01 -1.88687134e+00 4.65279430e-01
-7.15617910e-02 -8.87524307e-01 -1.07964136e-01 -1.34054711e-03
2.91433334e-01 -2.11676002e-01 -1.25552952e-01 7.23156691e-01
9.74358082e-01 -1.12010849e+00 1.25588787e+00 6.24654651e-01
5.61362267e-01 -2.05951318e-01 1.02299908e-02 8.69142592e-01
-7.47622848e-01 -2.95315921e-01 1.30964234e-01 -5.71289897e-01
8.76715407e-03 -4.30275232e-01 -1.00743210e+00 1.16821304e-01
2.04006061e-01 3.44595522e-01 -2.51412541e-01 5.91295063e-01
-8.55308950e-01 2.64070123e-01 -3.04786056e-01 -4.49026912e-01
8.43417346e-02 -6.25139847e-02 5.73730946e-01 9.69436705e-01
2.27590710e-01 5.89817949e-02 3.01597744e-01 1.31632304e+00
5.01585305e-01 -5.34852028e-01 -1.05902231e+00 -1.60846725e-01
1.05504533e-02 7.05770850e-01 -5.70310235e-01 -4.45674241e-01
5.22230007e-02 1.00896049e+00 5.25615215e-01 3.63507718e-01
-8.31309199e-01 -4.83256608e-01 3.16463411e-01 -2.85097122e-01
6.37248874e-01 -6.51535153e-01 -3.10316421e-02 -8.80419850e-01
-2.52800018e-01 -1.21073282e+00 -2.10824206e-01 -1.31336844e+00
-6.47271335e-01 3.59225541e-01 4.00157094e-01 -1.07597411e+00
-9.72301781e-01 -9.93026376e-01 -3.19836646e-01 7.94220030e-01
-8.92381072e-01 -1.12391078e+00 -1.88431926e-02 3.23775411e-01
5.14827788e-01 -1.93051454e-02 1.11031711e+00 -5.93907118e-01
-4.50514853e-02 -3.89180914e-03 -6.04483664e-01 -1.91440448e-01
2.65386641e-01 -1.67626655e+00 2.83858448e-01 5.79841137e-01
1.10480629e-01 8.88689756e-01 1.24051523e+00 -4.36691225e-01
-1.62650216e+00 -6.83385134e-01 5.51060557e-01 -9.31382775e-01
9.05458927e-01 -5.29759824e-01 -7.21373439e-01 1.18615508e+00
4.97280955e-01 -4.13390428e-01 -1.88816246e-02 -9.32806134e-02
-3.16905707e-01 5.07439613e-01 -8.81075859e-01 1.00287163e+00
1.20040107e+00 -7.89349198e-01 -1.35137749e+00 3.97965789e-01
9.97888029e-01 -5.11341333e-01 -6.03028893e-01 -3.63518447e-02
7.18045354e-01 -6.56789064e-01 6.51622534e-01 -8.29970002e-01
5.37040293e-01 -6.35860503e-01 -7.28141740e-02 -1.72674656e+00
-4.08022791e-01 -9.22535181e-01 -4.78860140e-02 7.69842684e-01
6.04965925e-01 -4.82596278e-01 3.02493453e-01 3.40896010e-01
-3.53691220e-01 -8.87628794e-02 -4.42240298e-01 -9.73018944e-01
2.85876244e-01 -3.93279850e-01 2.06080064e-01 5.72811544e-01
4.32634085e-01 7.73645699e-01 -1.25132814e-01 2.34058306e-01
4.78984654e-01 6.55241162e-02 1.07389653e+00 -1.17996490e+00
-9.09238696e-01 -6.85402274e-01 -1.44191489e-01 -1.24779034e+00
5.75008810e-01 -1.02488399e+00 7.78640568e-01 -1.29774570e+00
-1.69237316e-01 -2.04766572e-01 1.50974527e-01 5.53007603e-01
2.12651834e-01 -3.67340982e-01 3.86454582e-01 2.40416422e-01
-2.92555422e-01 3.85556191e-01 1.31328177e+00 -1.25668824e-01
-4.18106437e-01 -1.91731602e-01 -4.55089331e-01 8.70650828e-01
7.93552577e-01 -4.84118253e-01 -6.40848815e-01 -4.62075323e-01
2.03662917e-01 3.00485969e-01 6.93565547e-01 -1.02300894e+00
2.58701235e-01 -4.07896906e-01 4.47624214e-02 7.69251585e-02
2.78568864e-01 -1.05428243e+00 2.14207858e-01 8.69970620e-01
-8.18796694e-01 -2.90254857e-02 2.93799251e-01 6.54606402e-01
2.49217078e-01 -3.88927847e-01 5.69912493e-01 -4.46788937e-01
-8.87005329e-01 -4.22802478e-01 -7.60293722e-01 -2.23582730e-01
1.09561145e+00 -1.60704836e-01 4.80260588e-02 -2.76014268e-01
-1.18364680e+00 2.23578840e-01 5.66794276e-01 3.60716939e-01
3.87938738e-01 -8.76445055e-01 -4.63451773e-01 2.68345773e-01
3.28703374e-01 -1.40752256e-01 -3.72055411e-01 6.73173666e-01
-5.59228361e-01 3.52767229e-01 -6.54318213e-01 -7.84202158e-01
-8.70637178e-01 9.43510711e-01 4.00275499e-01 1.53848127e-01
-5.41069925e-01 6.84902966e-01 2.52024829e-01 -8.78904283e-01
2.25311115e-01 -8.55646789e-01 -3.64806838e-02 -5.00713110e-01
1.41087234e-01 -2.48586103e-01 -5.90517163e-01 -3.24905187e-01
-4.98270243e-02 5.13441265e-01 2.07649842e-01 -5.85439920e-01
1.10203826e+00 1.39729559e-01 1.70139875e-02 8.01329255e-01
7.61472464e-01 -4.40299749e-01 -1.73707056e+00 2.12874219e-01
1.49107262e-01 9.64269564e-02 -4.06097531e-01 -8.09574187e-01
-1.56978294e-01 9.67005491e-01 2.72675186e-01 3.14266026e-01
7.72202909e-01 8.13479796e-02 8.00421536e-02 7.46086180e-01
6.72282338e-01 -9.72466886e-01 3.34217310e-01 8.26292634e-01
1.27321327e+00 -1.06226480e+00 -2.43537813e-01 -2.22673625e-01
-6.95394874e-01 1.24756193e+00 5.71723819e-01 -4.21816885e-01
9.48959589e-02 3.80010605e-01 -3.02524358e-01 -1.32248551e-01
-8.78478527e-01 -3.98917735e-01 -4.96326648e-02 1.14267290e+00
1.54441267e-01 1.74899071e-01 1.01863138e-01 1.81752995e-01
-3.69890809e-01 2.06813991e-01 6.04776502e-01 1.18698323e+00
-5.16050696e-01 -1.04050899e+00 -1.60349682e-01 7.38201141e-02
5.99876568e-02 1.24179028e-01 -5.04163682e-01 1.11692011e+00
1.28286436e-01 6.31805599e-01 -1.12959877e-01 -3.87785256e-01
5.22443712e-01 3.49198252e-01 1.18277562e+00 -1.02833307e+00
-3.23571146e-01 -1.25294566e-01 6.83959648e-02 -5.02853513e-01
-3.50703597e-01 -8.66348267e-01 -1.57837832e+00 2.32501686e-01
-1.95698008e-01 -1.15019545e-01 6.92812681e-01 1.01760566e+00
1.26291797e-01 9.83892441e-01 1.18377432e-01 -1.24723876e+00
-8.79380286e-01 -9.64995146e-01 -2.40708157e-01 2.36898646e-01
5.35388052e-01 -7.81889677e-01 -3.29134911e-01 5.96104681e-01] | [4.485483169555664, 0.7441084384918213] |
7e188f8c-4b1b-4af3-bea1-32060f905eb3 | intriguing-properties-of-text-guided | 2306.00974 | null | https://arxiv.org/abs/2306.00974v3 | https://arxiv.org/pdf/2306.00974v3.pdf | Intriguing Properties of Text-guided Diffusion Models | Text-guided diffusion models (TDMs) are widely applied but can fail unexpectedly. Common failures include: (i) natural-looking text prompts generating images with the wrong content, or (ii) different random samples of the latent variables that generate vastly different, and even unrelated, outputs despite being conditioned on the same text prompt. In this work, we aim to study and understand the failure modes of TDMs in more detail. To achieve this, we propose SAGE, an adversarial attack on TDMs that uses image classifiers as surrogate loss functions, to search over the discrete prompt space and the high-dimensional latent space of TDMs to automatically discover unexpected behaviors and failure cases in the image generation. We make several technical contributions to ensure that SAGE finds failure cases of the diffusion model, rather than the classifier, and verify this in a human study. Our study reveals four intriguing properties of TDMs that have not been systematically studied before: (1) We find a variety of natural text prompts producing images that fail to capture the semantics of input texts. We categorize these failures into ten distinct types based on the underlying causes. (2) We find samples in the latent space (which are not outliers) that lead to distorted images independent of the text prompt, suggesting that parts of the latent space are not well-structured. (3) We also find latent samples that lead to natural-looking images which are unrelated to the text prompt, implying a potential misalignment between the latent and prompt spaces. (4) By appending a single adversarial token embedding to an input prompt we can generate a variety of specified target objects, while only minimally affecting the CLIP score. This demonstrates the fragility of language representations and raises potential safety concerns. | ['Alan Yuille', 'Song Bai', 'Yutong Bai', 'Adam Kortylewski', 'Qihao Liu'] | 2023-06-01 | null | null | null | null | ['adversarial-attack'] | ['adversarial'] | [ 8.06831956e-01 1.39701784e-01 2.38443151e-01 -1.28767341e-01
-7.79841721e-01 -1.01516771e+00 1.12574935e+00 -3.69390339e-01
1.27169520e-01 2.80860573e-01 4.43608701e-01 -2.45632499e-01
3.30850631e-02 -4.68475789e-01 -9.75959837e-01 -9.35355961e-01
2.10312698e-02 3.37742239e-01 3.61322127e-02 -2.66862437e-02
2.13590175e-01 2.25663766e-01 -1.20844805e+00 5.88064790e-01
7.54586518e-01 5.36550105e-01 -4.54640687e-02 8.86825860e-01
1.65717900e-02 1.07800901e+00 -9.65645015e-01 -5.16945958e-01
4.38351005e-01 -7.60715067e-01 -5.15704274e-01 3.54204655e-01
7.90314496e-01 -5.49456000e-01 -5.33249557e-01 1.18200564e+00
3.40129524e-01 -2.76223928e-01 1.10833097e+00 -1.56029773e+00
-1.00107825e+00 4.94469434e-01 -5.14528513e-01 -2.58426312e-02
4.25951809e-01 5.22187591e-01 9.55336630e-01 -7.04115927e-01
8.29010427e-01 1.49173033e+00 5.75253665e-01 7.80795932e-01
-1.58382010e+00 -7.39511549e-01 3.32105160e-02 -1.13896333e-01
-1.18402994e+00 -7.04190254e-01 7.18815684e-01 -7.34892130e-01
6.03363752e-01 2.86719203e-01 1.60492405e-01 2.00830674e+00
4.39520031e-01 6.78826451e-01 1.30198884e+00 -4.82190371e-01
3.26044321e-01 2.15042591e-01 -2.56705344e-01 5.33165157e-01
4.30574454e-02 4.10048187e-01 -6.45066798e-01 -4.85621989e-01
7.08846867e-01 -1.84850141e-01 -3.09825718e-01 -3.12450647e-01
-1.48352695e+00 1.07818377e+00 7.63128325e-02 1.58430874e-01
-3.10066968e-01 3.85797590e-01 1.75672784e-01 4.63073641e-01
2.72985369e-01 5.24343014e-01 -5.89139797e-02 5.92175648e-02
-8.63693118e-01 3.43336076e-01 7.47280419e-01 1.00199687e+00
4.13679600e-01 3.09288234e-01 -3.44586045e-01 3.88088405e-01
-6.53243205e-03 7.08295941e-01 6.02665186e-01 -9.03026104e-01
5.72182178e-01 1.01844713e-01 1.87757820e-01 -1.30293119e+00
-8.10073018e-02 -5.90504371e-02 -7.88035095e-01 3.59027803e-01
8.44199359e-01 -1.45572767e-01 -9.95287418e-01 2.09144402e+00
-2.40350783e-01 7.70635828e-02 5.84825240e-02 7.10951626e-01
3.21286023e-01 6.38970375e-01 1.34438993e-02 -9.82817113e-02
1.15735722e+00 -5.77514827e-01 -6.02846742e-01 -2.91158497e-01
4.10248011e-01 -8.60785902e-01 1.46598220e+00 3.03306550e-01
-1.00083554e+00 -2.54684627e-01 -1.04168880e+00 2.67409772e-01
-2.14765593e-01 -1.79979414e-01 2.45313153e-01 7.03571200e-01
-9.81932044e-01 4.41149831e-01 -5.38271785e-01 -2.35198453e-01
2.66379654e-01 1.08054336e-02 -2.02709988e-01 -1.57358170e-01
-1.17317104e+00 7.02902317e-01 -2.67485201e-01 -3.40594560e-01
-1.21510136e+00 -4.82424855e-01 -5.52272022e-01 -9.41589624e-02
3.74226630e-01 -6.07533455e-01 1.09098232e+00 -1.49989939e+00
-1.10901141e+00 7.25706339e-01 -2.76280314e-01 -4.58108872e-01
9.19074833e-01 -3.77213061e-02 -4.25783068e-01 2.98652112e-01
3.43506396e-01 6.77794158e-01 1.70031500e+00 -1.48557532e+00
-1.13020211e-01 -1.18506692e-01 -3.40271384e-01 3.88800390e-02
-2.42977783e-01 -1.72428668e-01 -1.01356909e-01 -1.23874581e+00
-1.90577477e-01 -8.51322711e-01 -7.36968368e-02 -6.57070801e-02
-9.05053616e-01 2.28977948e-01 8.15729141e-01 -3.41797829e-01
8.90162468e-01 -2.32645416e+00 3.16581652e-02 1.32978186e-01
4.79936808e-01 -1.45414427e-01 -4.82362747e-01 6.24600649e-01
-2.81729728e-01 5.13480067e-01 -2.11258024e-01 -3.12988460e-01
1.80108935e-01 1.37738243e-01 -1.17687500e+00 5.37866712e-01
2.22036436e-01 1.09638381e+00 -7.17538357e-01 -1.43055171e-01
-5.30127920e-02 2.00165242e-01 -3.76134992e-01 6.46262020e-02
-4.10972357e-01 4.08466876e-01 -1.73146695e-01 3.83893311e-01
5.48023403e-01 -1.20451733e-01 -1.86890319e-01 -2.01402873e-01
2.20752746e-01 1.20739721e-01 -8.80186021e-01 1.24657214e+00
4.71656397e-02 8.82999599e-01 -1.70765772e-01 -8.19473207e-01
5.70028961e-01 3.72554839e-01 2.56499767e-01 -7.97123015e-01
-8.61899853e-02 2.36228570e-01 -3.47501859e-02 -5.71503103e-01
4.70361322e-01 -5.46159327e-01 -2.00222179e-01 9.32595730e-01
6.00157566e-02 -2.92382147e-02 -2.27350116e-01 4.73436385e-01
1.46730626e+00 -3.53902072e-01 -1.37174889e-01 -6.86991736e-02
-2.21691549e-01 -6.24755546e-02 3.06769192e-01 1.36447871e+00
-2.05709383e-01 8.89087796e-01 9.77530360e-01 -2.18331665e-01
-1.57813466e+00 -1.41341567e+00 1.19743235e-01 7.06541479e-01
8.99870768e-02 -2.56359339e-01 -8.58766735e-01 -7.52580881e-01
-5.78158870e-02 9.95705068e-01 -6.93951488e-01 -4.33230042e-01
-2.68331200e-01 -6.48548424e-01 1.10212374e+00 2.00841114e-01
8.39026421e-02 -9.53566968e-01 -5.23314118e-01 2.08338778e-02
-2.36405179e-01 -1.18219995e+00 -7.18594313e-01 2.67050471e-02
-5.90529442e-01 -1.03073025e+00 -8.17765355e-01 -4.73962873e-01
1.08798051e+00 1.43808007e-01 1.04050815e+00 -1.73760146e-01
-3.28474641e-01 7.30998456e-01 -2.53412515e-01 -2.07961977e-01
-9.71204400e-01 -3.67441505e-01 2.37562254e-01 2.22526878e-01
1.14768527e-01 -4.39690530e-01 -4.83513236e-01 4.04819518e-01
-1.34899604e+00 8.74704029e-03 4.97923195e-01 9.28420544e-01
1.78442746e-01 4.09168094e-01 5.71906686e-01 -9.27614570e-01
9.67201114e-01 -6.68456256e-01 -4.28330690e-01 5.70902079e-02
-5.00836194e-01 2.02399388e-01 7.79657543e-01 -9.51451659e-01
-7.32367039e-01 -1.73005480e-02 1.97715700e-01 -6.92819655e-01
-3.21669787e-01 7.49086887e-02 1.31783523e-02 4.34286922e-01
1.00305057e+00 5.20312607e-01 1.65739775e-01 -3.42620015e-01
5.36081493e-01 4.48666662e-01 5.38505077e-01 -6.52910888e-01
1.14665413e+00 6.98254049e-01 -3.18362713e-01 -7.58366644e-01
-5.31378984e-01 1.91170320e-01 -1.50511831e-01 -1.31268099e-01
8.69809270e-01 -7.43525863e-01 -2.63095111e-01 6.81528330e-01
-1.27285171e+00 -6.21886671e-01 -4.28427160e-01 1.64820611e-01
-7.27259159e-01 3.70389611e-01 -6.23104036e-01 -7.88464427e-01
1.72537602e-02 -1.26534760e+00 9.63076830e-01 -2.05639586e-01
-6.64499402e-01 -9.31052744e-01 -2.28267595e-01 1.24427350e-02
3.61154675e-01 2.18272746e-01 1.20407081e+00 -8.53375316e-01
-6.87578857e-01 -2.25650176e-01 -4.13838290e-02 3.61774981e-01
2.36451745e-01 1.01901479e-01 -9.63032246e-01 -3.74381900e-01
3.79645169e-01 -3.82785648e-01 8.24461937e-01 3.91666234e-01
7.47484148e-01 -8.86660337e-01 -1.59023076e-01 5.46127021e-01
1.18974376e+00 2.76849270e-01 6.80237949e-01 1.46503627e-01
4.69145298e-01 7.23524570e-01 6.00178763e-02 1.93220183e-01
-5.73255420e-02 5.32763362e-01 2.53383428e-01 -1.84626654e-01
-1.70355756e-02 -7.10136592e-01 9.06504691e-01 3.80778968e-01
6.73744023e-01 -6.08299434e-01 -7.37872005e-01 6.46761477e-01
-1.63167429e+00 -1.27357316e+00 -9.46054757e-02 2.28335977e+00
8.39727640e-01 3.16041321e-01 2.88316812e-02 3.07811927e-02
7.33044684e-01 3.10860097e-01 -6.83660924e-01 -3.94585460e-01
-3.06540221e-01 1.24307498e-02 6.35578215e-01 5.71631014e-01
-8.52254450e-01 8.13072324e-01 7.09339523e+00 8.85719597e-01
-1.27386391e+00 5.14470711e-02 6.76288307e-01 -2.38090470e-01
-7.56265342e-01 7.98747540e-02 -3.85525018e-01 6.94364190e-01
6.76514685e-01 -2.67781168e-01 6.59888864e-01 5.71302056e-01
1.33444220e-01 1.69676572e-01 -1.25757635e+00 8.01743448e-01
1.93462774e-01 -1.26939702e+00 3.44110817e-01 2.11519063e-01
8.26036334e-01 -1.66577980e-01 5.48932970e-01 -4.17147726e-02
6.01974785e-01 -1.33329952e+00 1.18696785e+00 5.78489840e-01
1.01196873e+00 -4.38709974e-01 1.20169654e-01 5.50410688e-01
-4.24531162e-01 -1.99776385e-02 -2.66363680e-01 1.75309509e-01
3.21574360e-02 6.36158824e-01 -8.12295258e-01 -9.82916504e-02
4.12135094e-01 4.12281752e-01 -5.83016515e-01 4.06080037e-01
-1.70033261e-01 7.78680742e-01 -1.93142161e-01 2.91416883e-01
6.11274131e-02 2.02864110e-01 1.01387632e+00 1.05494940e+00
3.64888489e-01 -2.46211931e-01 -2.08653614e-01 1.62081921e+00
-4.04482558e-02 -3.44786078e-01 -1.13129437e+00 -4.59712565e-01
3.64475340e-01 7.81217515e-01 -7.06599593e-01 -1.80856004e-01
-3.05935860e-01 1.55951977e+00 -1.41639888e-01 8.39943171e-01
-9.28477526e-01 -2.36231163e-01 5.68809330e-01 9.17397738e-02
2.81288654e-01 -2.44014621e-01 -4.36233312e-01 -1.26140225e+00
1.81787789e-01 -1.16809344e+00 -2.26424765e-02 -8.53543520e-01
-1.81175864e+00 6.02461874e-01 -9.53275412e-02 -1.24294770e+00
-4.35345948e-01 -3.65574300e-01 -7.25816488e-01 9.63322520e-01
-9.19671357e-01 -8.15814912e-01 -1.65055087e-03 7.48405993e-01
6.37429595e-01 -3.56181413e-01 7.66153455e-01 -1.03467494e-01
-2.81734377e-01 7.90210307e-01 1.48049489e-01 7.94080049e-02
8.81324470e-01 -1.21223354e+00 7.86663949e-01 9.30000007e-01
3.22033376e-01 7.28055000e-01 9.15833533e-01 -7.67630875e-01
-1.26821303e+00 -1.06839323e+00 7.99415112e-01 -8.61660182e-01
6.54847682e-01 -6.21996820e-01 -7.50592172e-01 7.70353854e-01
1.68332130e-01 -3.00543278e-01 3.83616805e-01 -3.92371535e-01
-7.04759240e-01 3.01825076e-01 -1.05169690e+00 1.01115179e+00
7.23560274e-01 -8.66221726e-01 -4.08151418e-01 2.50315964e-01
7.62667418e-01 -1.46479204e-01 -2.33976409e-01 2.19682464e-03
4.67485219e-01 -1.05600226e+00 9.71392751e-01 -6.46808922e-01
8.90069783e-01 -1.06732443e-01 -2.53806263e-01 -1.26600707e+00
-1.22379959e-01 -8.33948255e-01 -1.63607597e-02 1.38411987e+00
3.78648698e-01 -6.37220263e-01 4.82724637e-01 6.78760290e-01
2.69898981e-01 -5.07490396e-01 -9.14623141e-01 -9.08000410e-01
3.01469177e-01 -5.96683860e-01 3.48965943e-01 1.00093806e+00
-1.99728057e-01 4.68008071e-01 -6.42037213e-01 2.48927586e-02
6.81691289e-01 -2.09824458e-01 7.10657716e-01 -6.92374885e-01
-3.94158542e-01 -6.56839073e-01 -1.94579795e-01 -1.21264362e+00
1.92556214e-02 -7.21897900e-01 1.73200026e-01 -1.07879245e+00
2.36788630e-01 -4.99718457e-01 1.02235079e-01 4.21509922e-01
-6.98039383e-02 1.94758773e-01 2.07341701e-01 6.95057452e-01
-1.88126087e-01 4.58010763e-01 1.06738234e+00 -2.76125550e-01
4.52151597e-02 -1.48423329e-01 -9.93387461e-01 6.79534197e-01
5.74666917e-01 -6.80843651e-01 -6.38423741e-01 -4.63941485e-01
4.27401781e-01 3.85707454e-03 8.86959672e-01 -6.12763703e-01
1.53415268e-02 -2.21241727e-01 4.11414564e-01 -1.54897839e-01
1.60470888e-01 -7.59635866e-01 9.21704918e-02 3.03503752e-01
-7.10584283e-01 1.68258674e-03 -1.03520006e-02 8.23172510e-01
6.03784733e-02 -1.86572701e-01 7.14037836e-01 -1.92423359e-01
-2.42731020e-01 7.78172985e-02 -8.00279319e-01 2.33623818e-01
9.58935440e-01 -1.63285375e-01 -5.79180598e-01 -8.69032383e-01
-5.20402014e-01 -2.97874123e-01 8.11002314e-01 5.56733489e-01
5.74976563e-01 -1.38789320e+00 -6.64382100e-01 4.39689517e-01
-5.39895892e-02 -2.72499889e-01 1.24031305e-01 6.92558646e-01
-1.94390073e-01 7.13029280e-02 5.66421486e-02 -6.40478015e-01
-8.19887877e-01 6.07901931e-01 3.84838581e-01 -4.77319621e-02
-8.16530228e-01 7.58646071e-01 7.18264461e-01 -1.07166730e-01
2.66170830e-01 -1.50519088e-01 4.88135964e-01 1.13796934e-01
4.15754229e-01 -1.86826214e-02 -3.09627235e-01 -5.84940672e-01
-8.71938020e-02 2.66668916e-01 -1.57612175e-01 -5.29339731e-01
9.34902132e-01 -1.41842499e-01 1.93061873e-01 4.91975248e-01
1.25073969e+00 3.76289517e-01 -1.42180121e+00 -1.14867881e-01
-1.21285565e-01 -5.84689558e-01 -3.11326444e-01 -8.35814834e-01
-8.76717091e-01 8.36184561e-01 5.12827694e-01 6.22298062e-01
1.02498209e+00 1.42020375e-01 8.76287460e-01 -8.88999850e-02
1.82795431e-02 -8.54549825e-01 5.17195702e-01 2.68395036e-01
1.12186229e+00 -8.80980730e-01 -4.44766998e-01 -9.09239799e-03
-1.00009203e+00 9.43242311e-01 2.92571843e-01 -1.05832778e-01
1.62742808e-01 3.30517530e-01 2.48448774e-01 -3.18729550e-01
-9.79722500e-01 2.49013945e-01 1.06416613e-01 7.97727346e-01
9.86017510e-02 -6.06174320e-02 1.81215525e-01 6.64553642e-01
-3.13027173e-01 -3.16828012e-01 6.63620591e-01 6.36122465e-01
-7.67907575e-02 -9.66384172e-01 -6.53510392e-01 4.35969353e-01
-4.96557206e-01 -2.29904264e-01 -7.62187123e-01 3.91594529e-01
-4.20373902e-02 1.01000774e+00 -2.27704202e-03 -6.32040501e-01
1.63240626e-01 2.94995278e-01 3.58486593e-01 -3.25174034e-01
-9.38757211e-02 4.70335148e-02 -2.05645919e-01 -4.97929722e-01
1.22576900e-01 -7.41421044e-01 -9.20466363e-01 -3.66973758e-01
-1.25438988e-01 -2.67742902e-01 4.42017972e-01 8.78224373e-01
3.70509773e-01 3.36502254e-01 7.78903186e-01 -7.48552263e-01
-9.21708345e-01 -7.26370573e-01 -4.90133911e-01 8.89771461e-01
6.97318494e-01 -3.84142429e-01 -8.45906973e-01 4.56324846e-01] | [11.588644027709961, -0.1167781725525856] |
0bbf9650-a818-4bfb-b064-12ac967b9a67 | geometric-algebra-attention-networks-for | 2110.02393 | null | https://arxiv.org/abs/2110.02393v2 | https://arxiv.org/pdf/2110.02393v2.pdf | Geometric Algebra Attention Networks for Small Point Clouds | Much of the success of deep learning is drawn from building architectures that properly respect underlying symmetry and structure in the data on which they operate - a set of considerations that have been united under the banner of geometric deep learning. Often problems in the physical sciences deal with relatively small sets of points in two- or three-dimensional space wherein translation, rotation, and permutation equivariance are important or even vital for models to be useful in practice. In this work, we present rotation- and permutation-equivariant architectures for deep learning on these small point clouds, composed of a set of products of terms from the geometric algebra and reductions over those products using an attention mechanism. The geometric algebra provides valuable mathematical structure by which to combine vector, scalar, and other types of geometric inputs in a systematic way to account for rotation invariance or covariance, while attention yields a powerful way to impose permutation equivariance. We demonstrate the usefulness of these architectures by training models to solve sample problems relevant to physics, chemistry, and biology. | ['Matthew Spellings'] | 2021-10-05 | geometric-algebra-attention-networks-for-1 | https://openreview.net/forum?id=nLb60uXd6Np | https://openreview.net/pdf?id=nLb60uXd6Np | null | ['classify-3d-point-clouds', 'generating-3d-point-clouds'] | ['computer-vision', 'computer-vision'] | [ 1.32860705e-01 -1.03876159e-01 5.93118854e-02 -4.50286776e-01
-2.83256829e-01 -6.77500129e-01 1.13179100e+00 5.41258976e-02
-2.63621837e-01 3.71144533e-01 2.20461264e-01 -4.66249049e-01
-4.43991721e-01 -8.44316959e-01 -9.70202208e-01 -8.03849459e-01
-2.08117723e-01 6.61753953e-01 -3.26556593e-01 -5.03426731e-01
5.01282215e-01 1.19495702e+00 -1.27498341e+00 -3.02452389e-02
4.21068907e-01 5.92134356e-01 -1.71606883e-01 5.61172068e-01
7.12306947e-02 2.44664431e-01 -1.43154562e-01 -1.24066867e-01
6.19855523e-01 -1.51293039e-01 -7.85960317e-01 -1.34635866e-01
6.88015461e-01 -6.74301013e-02 -4.14820045e-01 8.89411330e-01
1.97914422e-01 4.25155908e-01 9.82252836e-01 -1.00020993e+00
-9.95188177e-01 2.17906579e-01 -4.61668402e-01 7.72470981e-02
-4.44717109e-02 7.23658577e-02 1.30029190e+00 -8.56518865e-01
5.43291926e-01 1.40627587e+00 5.33316135e-01 2.13698700e-01
-1.43350375e+00 -2.93412894e-01 -9.66403782e-02 7.81711340e-02
-1.31904006e+00 -3.24666739e-01 8.59766781e-01 -5.74027061e-01
1.04581106e+00 2.40729466e-01 6.26475751e-01 7.26388991e-01
4.61779684e-01 2.11748064e-01 4.39853340e-01 -3.73729944e-01
1.75231218e-01 -2.64519066e-01 1.07974172e-01 5.95982671e-01
3.84499758e-01 4.72085103e-02 -3.09697110e-02 1.52823208e-02
1.14627039e+00 2.56145507e-01 -6.18673768e-03 -1.01453209e+00
-1.37389243e+00 9.17768717e-01 8.26138973e-01 1.55697584e-01
-1.71683893e-01 1.14955284e-01 1.09366074e-01 4.22378145e-02
7.55846426e-02 1.19084656e+00 -5.01721442e-01 3.28343481e-01
-4.64215100e-01 4.58761185e-01 5.55855930e-01 8.32640529e-01
6.87376738e-01 -4.04950837e-03 2.65545100e-01 4.02121156e-01
2.25153968e-01 5.00836551e-01 2.20998302e-01 -1.10406327e+00
1.46673560e-01 5.70537269e-01 -4.96146828e-02 -1.11016083e+00
-7.28684366e-01 -5.84686935e-01 -1.02330947e+00 3.37001115e-01
4.88668323e-01 4.19451833e-01 -1.01468825e+00 1.87484086e+00
1.39485285e-01 -9.71396938e-02 -3.97761017e-02 7.95122862e-01
4.19184744e-01 5.12943864e-01 -1.66975513e-01 1.94629505e-01
1.17423379e+00 -4.04902339e-01 -1.42114878e-01 1.74198672e-01
7.92986035e-01 -6.25814199e-01 1.05757606e+00 2.18190312e-01
-1.40322816e+00 -5.12771010e-01 -1.13883460e+00 -7.15972424e-01
-6.64411366e-01 -2.28494480e-01 7.10102141e-01 2.67986149e-01
-1.16174281e+00 1.03569221e+00 -8.77121568e-01 -2.70378381e-01
5.57261765e-01 6.23595238e-01 -6.01745427e-01 -1.08127971e-03
-7.91393220e-01 1.06314385e+00 -9.70886797e-02 9.58419293e-02
-4.73687947e-01 -9.76980925e-01 -9.02666271e-01 3.52654964e-01
-5.53304739e-02 -8.96686018e-01 1.09265363e+00 -5.60222566e-01
-1.20235550e+00 9.19093788e-01 -1.94927573e-01 -3.42060387e-01
2.52886593e-01 -8.36932585e-02 5.36563434e-02 -3.22693437e-02
5.40156923e-02 6.10418379e-01 7.05119550e-01 -9.70470369e-01
-1.09353559e-02 -7.07114220e-01 2.63755918e-01 2.48635426e-01
5.93512431e-02 -2.20564261e-01 -1.06176108e-01 -3.83305788e-01
5.07523715e-01 -1.13197720e+00 -3.23269129e-01 -8.58795494e-02
-4.78902519e-01 -2.35713840e-01 9.11917448e-01 -3.10238391e-01
4.17026371e-01 -2.03951669e+00 5.52994847e-01 4.24574822e-01
4.21915025e-01 2.42814422e-01 -2.71171480e-01 3.32720190e-01
-6.52469873e-01 5.12120426e-01 -7.52545372e-02 -3.19010485e-03
8.03495795e-02 9.83786359e-02 -4.83785778e-01 6.60477877e-01
6.74155891e-01 1.12606001e+00 -7.32626677e-01 1.16016425e-01
4.13004935e-01 7.08314776e-01 -8.55226994e-01 -2.27852449e-01
-3.29445928e-01 5.53199708e-01 -3.99712592e-01 2.46710092e-01
5.51382601e-01 -3.94491941e-01 6.96090981e-02 -2.99066693e-01
-2.01485053e-01 7.06859946e-01 -9.63291526e-01 1.58550632e+00
-2.51354307e-01 5.21221638e-01 -9.37235802e-02 -1.27323818e+00
7.64174819e-01 -4.01158929e-02 8.15961123e-01 -6.30281687e-01
3.67285401e-01 -8.16088095e-02 5.85978508e-01 -1.79451644e-01
5.11860967e-01 -9.37943384e-02 5.19677885e-02 5.87112963e-01
1.79103781e-02 -5.38698614e-01 1.45838693e-01 1.35554045e-01
8.36478412e-01 1.13445573e-01 1.45860076e-01 -5.32491922e-01
5.17471492e-01 -2.25447148e-01 2.62195647e-01 6.19891286e-01
1.09152466e-01 6.97710276e-01 5.29351950e-01 -9.30319548e-01
-1.64783669e+00 -1.27112281e+00 -2.69388884e-01 8.63100111e-01
-1.79607257e-01 -2.20498770e-01 -4.19969887e-01 -8.32626522e-02
3.41374911e-02 5.70673525e-01 -6.76313937e-01 -4.89969999e-01
-6.77378356e-01 -6.62257612e-01 2.09692240e-01 4.91112024e-01
2.72121251e-01 -7.75416493e-01 -2.51000524e-01 -8.29462633e-02
2.98474580e-01 -7.99649358e-01 -2.30507627e-01 1.86432526e-01
-8.52518141e-01 -1.18269873e+00 -5.01584589e-01 -6.35384560e-01
6.99259043e-01 4.13085610e-01 9.81566191e-01 -8.38192478e-02
-4.43862319e-01 3.89275521e-01 1.57609135e-01 -7.11309493e-01
-2.15655267e-01 2.56080359e-01 3.07348341e-01 -1.13458157e-01
3.25435787e-01 -8.72816443e-01 -3.18164140e-01 7.45475069e-02
-9.12897289e-01 -6.03149869e-02 4.06392753e-01 6.02001607e-01
4.66940612e-01 -1.31295502e-01 2.11733431e-01 -5.07185459e-01
3.18786353e-01 -2.64973730e-01 -6.78493261e-01 -2.79593915e-02
-9.02643576e-02 6.24111891e-01 7.23767459e-01 -3.72254550e-02
-5.01489580e-01 2.83477120e-02 1.16017222e-01 -4.17608112e-01
-7.05073699e-02 4.31658179e-01 -3.70248824e-01 -2.04036519e-01
6.92944527e-01 -7.91594908e-02 1.35512277e-01 -4.07161951e-01
6.33725941e-01 2.64053303e-03 5.62118828e-01 -8.82439196e-01
8.19126368e-01 7.74897277e-01 8.25630367e-01 -9.93274093e-01
-6.46474183e-01 -2.78300911e-01 -1.19403875e+00 4.54205841e-01
9.15247440e-01 -6.79296613e-01 -7.69732475e-01 2.57923841e-01
-1.21629643e+00 -7.09041804e-02 -4.43332046e-01 5.00823915e-01
-6.11960232e-01 2.27061734e-01 -1.52449831e-01 -3.39018553e-01
3.37314680e-02 -1.29978549e+00 1.06621623e+00 4.24251519e-02
-3.09186369e-01 -1.11442125e+00 8.08090866e-02 -1.06327593e-01
4.22157884e-01 1.61452949e-01 1.39593899e+00 -7.05820501e-01
-8.47291529e-01 -1.51208997e-01 -2.49115884e-01 1.09133698e-01
9.48219672e-02 3.50418180e-01 -7.95225203e-01 -1.50548100e-01
-1.37046736e-03 7.48227239e-02 9.09402430e-01 5.07997155e-01
1.50910437e+00 -9.42052305e-02 -9.69738513e-02 1.04432678e+00
1.15185344e+00 -1.61261261e-02 5.35025597e-01 2.14010194e-01
1.04324985e+00 5.59927762e-01 -2.36698672e-01 9.15018842e-02
3.40907961e-01 5.69598019e-01 5.24412811e-01 -1.39933214e-01
1.27972007e-01 -1.30491629e-01 -1.09910809e-01 7.49469697e-01
-3.86882037e-01 2.05053598e-01 -9.48889792e-01 4.04146880e-01
-1.53095186e+00 -1.10144722e+00 -7.02892691e-02 2.36684632e+00
4.79588628e-01 3.09163402e-03 -1.34118572e-01 1.78338856e-01
4.61389244e-01 1.50459722e-01 -7.41064727e-01 -6.75417304e-01
-1.14542045e-01 5.19578576e-01 5.38061678e-01 5.10882139e-01
-1.09326005e+00 7.53647149e-01 6.72031784e+00 1.07738435e-01
-1.45078945e+00 -4.36086923e-01 4.56065893e-01 7.53679052e-02
-4.14389640e-01 -8.20561200e-02 -6.08804047e-01 2.35539917e-02
8.38395357e-01 -1.33821175e-01 6.41202390e-01 6.87804937e-01
3.09840977e-01 4.25372303e-01 -1.63200331e+00 8.03187966e-01
-4.19986285e-02 -1.73002446e+00 2.85608858e-01 3.62569928e-01
7.58946836e-01 1.97786242e-01 5.06960094e-01 8.76445398e-02
5.29035449e-01 -1.31257844e+00 5.00837862e-01 4.02868450e-01
6.95580900e-01 -4.99048382e-01 2.45644435e-01 1.32855950e-02
-7.87869632e-01 9.52915400e-02 -5.68875849e-01 -3.15965623e-01
-2.02258319e-01 2.73391575e-01 -8.38690937e-01 4.21905369e-01
3.90134513e-01 8.24678183e-01 -2.51752108e-01 7.06104517e-01
-7.16222674e-02 4.60251160e-02 -4.29073095e-01 1.51200846e-01
3.91794592e-01 -6.12076938e-01 4.20988232e-01 7.90747464e-01
4.07454163e-01 1.62296250e-01 -3.12054515e-01 1.08457983e+00
-3.19429696e-01 -8.16351548e-02 -9.81042743e-01 -1.37203500e-01
2.52520859e-01 1.31689548e+00 -6.58069968e-01 -1.08140163e-01
-3.68161082e-01 2.66102105e-01 3.54473263e-01 4.22071695e-01
-5.46159506e-01 -2.40721703e-01 1.17426836e+00 2.50490129e-01
5.16977966e-01 -8.93248618e-01 -6.39356494e-01 -1.21053076e+00
-1.17897615e-01 -6.82638884e-01 -1.18280783e-01 -8.91942024e-01
-1.05693793e+00 1.01302147e-01 -2.23562345e-02 -8.34028602e-01
-1.80489987e-01 -1.12277448e+00 -8.57859969e-01 1.06820977e+00
-1.26072741e+00 -1.18218434e+00 2.30138954e-02 4.63647664e-01
-7.89522678e-02 -9.89464223e-02 8.67571533e-01 3.84438783e-02
-3.94042820e-01 2.65980244e-01 3.55641097e-01 2.32633084e-01
5.17479599e-01 -1.31543446e+00 8.35186362e-01 6.25845313e-01
4.17822719e-01 1.08668804e+00 6.39157891e-01 -1.53035775e-01
-1.79596543e+00 -9.85445380e-01 8.20686162e-01 -8.74100506e-01
6.40637040e-01 -4.73428845e-01 -9.45614100e-01 9.79246974e-01
-2.10085109e-01 7.37860650e-02 5.51489413e-01 2.97356784e-01
-8.75834584e-01 -6.80784881e-02 -8.63968551e-01 8.46676528e-01
9.94339466e-01 -6.16037369e-01 -6.79250956e-01 3.91356587e-01
7.61530161e-01 -2.49706686e-01 -7.40354061e-01 3.15176636e-01
6.69202626e-01 -7.21945822e-01 1.37197995e+00 -1.31478608e+00
5.69375932e-01 -2.46381089e-01 -1.87173501e-01 -1.43169880e+00
-7.64808059e-01 -5.29773593e-01 3.75998378e-01 4.45090652e-01
2.70474225e-01 -5.53775370e-01 6.31705642e-01 7.26785243e-01
-4.43044841e-01 -5.03091037e-01 -6.46792233e-01 -5.16529143e-01
7.31418133e-01 -2.48506486e-01 9.35669124e-01 1.19434679e+00
-1.69081420e-01 4.87263858e-01 1.17508106e-01 1.58409774e-01
4.62151259e-01 6.54879957e-02 9.48516548e-01 -1.52429557e+00
-3.75382602e-02 -8.91853809e-01 -6.81126297e-01 -8.47620785e-01
1.74633965e-01 -1.16900480e+00 -3.30768585e-01 -1.40720606e+00
-4.43353988e-02 -3.90039980e-01 -2.98951387e-01 3.87608767e-01
2.15931222e-01 9.51800272e-02 2.50932693e-01 2.99689949e-01
-3.21148813e-01 5.19255877e-01 1.22388732e+00 -1.13093115e-01
-1.71121329e-01 -2.18397856e-01 -1.01989067e+00 8.35756183e-01
7.89204180e-01 -2.15921402e-02 -2.30425492e-01 -7.68996835e-01
4.65683758e-01 -4.24656808e-01 4.60567325e-01 -8.56242597e-01
5.69679178e-02 -2.58331150e-01 6.53826416e-01 -3.83848578e-01
5.63161016e-01 -6.77936792e-01 -5.22436947e-02 2.63105392e-01
-5.32254338e-01 2.46849999e-01 7.55666569e-02 2.77225614e-01
7.36706704e-02 2.05253720e-01 8.84492338e-01 -2.32203558e-01
-2.99097210e-01 6.50734723e-01 3.36025953e-02 4.20952439e-02
8.16652477e-01 5.65443188e-02 -4.38883007e-01 -3.01813304e-01
-7.38153934e-01 -5.48360310e-02 6.78563893e-01 3.81830782e-01
2.60753006e-01 -1.45263517e+00 -5.53032100e-01 7.40147471e-01
-9.21804737e-03 3.73820633e-01 -6.20459951e-02 7.38765538e-01
-7.21548021e-01 7.37386644e-01 -6.61229312e-01 -6.67231143e-01
-8.28163147e-01 7.42058694e-01 3.80349338e-01 -1.32830098e-01
-3.95817339e-01 5.64984620e-01 6.34705722e-01 -6.94762051e-01
-9.91024002e-02 -8.00149798e-01 1.82260089e-02 -3.37971419e-01
3.11572015e-01 1.51318207e-01 3.06246758e-01 -8.57833028e-01
-2.17503190e-01 7.71518648e-01 -1.23023279e-01 -3.33134234e-02
1.53258407e+00 2.78929949e-01 -2.02536494e-01 2.83913225e-01
1.37633264e+00 -2.77434438e-02 -1.15836179e+00 -2.46809676e-01
-2.20907867e-01 -1.49139732e-01 -6.58091456e-02 -2.67771751e-01
-7.55826831e-01 1.39436162e+00 -4.29255255e-02 3.70782405e-01
4.75360125e-01 2.18723733e-02 3.43442142e-01 7.05013633e-01
-1.01826221e-01 -6.96897447e-01 -5.45021035e-02 9.33354974e-01
1.02411079e+00 -1.15371156e+00 1.33450314e-01 -4.76951376e-02
-4.04252484e-02 1.23737347e+00 2.71421313e-01 -5.76589763e-01
8.99353325e-01 -8.05355683e-02 -3.17664415e-01 -2.69841135e-01
-4.57878798e-01 -9.21062380e-02 5.54159641e-01 6.25009596e-01
5.24168372e-01 4.27220482e-03 2.05958411e-02 1.87174723e-01
-5.96885145e-01 -5.30540526e-01 5.14650702e-01 6.18786871e-01
-4.72592175e-01 -1.12547660e+00 -3.84662956e-01 3.18318397e-01
-3.10019225e-01 -1.95452683e-02 -5.31149447e-01 8.68538380e-01
1.33589935e-02 3.48640710e-01 5.81149757e-01 -2.28585035e-01
1.86124653e-01 8.38006735e-02 8.05361331e-01 -6.59844339e-01
-8.13423023e-02 -1.20878577e-01 -3.39954346e-01 -4.70731705e-01
-2.42941707e-01 -7.79867709e-01 -1.20598376e+00 -5.38617671e-01
1.73670605e-01 -3.26479375e-02 1.04793131e+00 1.02426517e+00
6.48687243e-01 4.56912458e-01 5.04677296e-01 -1.22025371e+00
-6.68384492e-01 -7.28285253e-01 -5.36260545e-01 6.16269886e-01
5.85142553e-01 -6.59807682e-01 -3.09120357e-01 -1.74449831e-02] | [8.840508460998535, 2.461977005004883] |
f01840f5-f1d7-4bfc-ab5d-3cbbe4f1dfe1 | vision-transformer-with-convolutional-encoder | 2209.05032 | null | https://arxiv.org/abs/2209.05032v1 | https://arxiv.org/pdf/2209.05032v1.pdf | Vision Transformer with Convolutional Encoder-Decoder for Hand Gesture Recognition using 24 GHz Doppler Radar | Transformers combined with convolutional encoders have been recently used for hand gesture recognition (HGR) using micro-Doppler signatures. We propose a vision-transformer-based architecture for HGR with multi-antenna continuous-wave Doppler radar receivers. The proposed architecture consists of three modules: a convolutional encoderdecoder, an attention module with three transformer layers, and a multi-layer perceptron. The novel convolutional decoder helps to feed patches with larger sizes to the attention module for improved feature extraction. Experimental results obtained with a dataset corresponding to a two-antenna continuous-wave Doppler radar receiver operating at 24 GHz (published by Skaria et al.) confirm that the proposed architecture achieves an accuracy of 98.3% which substantially surpasses the state-of-the-art on the used dataset. | ['Chamira U. S. Edussooriya', 'Ranga Rodrigo', 'Arjuna Madanayake', 'Viduneth Ariyarathna', 'Nisal Kariyawasam', 'Dhanuka Marasinghe', 'Gayangana Leelarathne', 'Kavinda Kehelella'] | 2022-09-12 | null | null | null | null | ['hand-gesture-recognition', 'hand-gesture-recognition-1', 'gesture-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 2.17373818e-01 5.33084050e-02 2.89955616e-01 -1.61109775e-01
-7.42329180e-01 -9.81891602e-02 6.66134179e-01 -7.37068295e-01
-6.51971459e-01 3.64881873e-01 1.37557119e-01 -3.81904185e-01
-3.04886967e-01 -6.13707006e-01 -4.56850976e-01 -8.45985770e-01
-4.23709273e-01 4.85258579e-01 -5.48195234e-03 -4.79833260e-02
1.95452750e-01 8.08447063e-01 -1.39294624e+00 2.84038812e-01
3.45290840e-01 1.16190529e+00 1.45293728e-01 1.09772956e+00
3.93684208e-01 9.60439146e-01 -5.05803406e-01 -3.08557808e-01
5.65207660e-01 -2.87077036e-02 -2.78471321e-01 -3.11570257e-01
7.07249761e-01 -1.03138483e+00 -1.17550135e+00 5.51000774e-01
9.16642249e-01 -1.73740149e-01 7.02449203e-01 -6.85836077e-01
-6.72682106e-01 3.41721922e-01 -5.17502964e-01 4.28615928e-01
1.10454433e-01 1.08135104e-01 8.95209670e-01 -7.57881582e-01
2.90605754e-01 1.07950115e+00 7.53816426e-01 5.03447831e-01
-5.59930861e-01 -6.04228258e-01 -4.54257458e-01 5.12632847e-01
-1.31116354e+00 -2.26254061e-01 7.03091621e-01 -4.52365458e-01
1.63916790e+00 1.00843668e-01 5.43551624e-01 1.22585130e+00
4.88394141e-01 5.17749310e-01 9.93246734e-01 -4.75732267e-01
1.39740761e-02 -3.07105124e-01 2.21255660e-01 9.15460527e-01
4.44317639e-01 8.30935180e-01 -5.46726465e-01 3.10068130e-02
9.07068014e-01 -1.66855589e-01 -1.53700352e-01 8.93182009e-02
-9.58780706e-01 6.46039188e-01 6.74611449e-01 5.63499570e-01
-9.95178044e-01 7.54040182e-01 6.83013424e-02 5.06155193e-01
4.58703488e-02 1.48518652e-01 -2.16913432e-01 -2.64249414e-01
-9.69635487e-01 3.13596070e-01 8.32000852e-01 9.30193007e-01
2.39750102e-01 5.12428761e-01 -1.38414875e-01 2.42793366e-01
8.10058594e-01 1.10262573e+00 2.52854854e-01 -2.02189073e-01
5.16626358e-01 1.28835201e-01 -3.69961299e-02 -5.31061232e-01
-6.14004612e-01 -8.56470883e-01 -8.04696620e-01 5.63941121e-01
2.79231191e-01 -4.56981480e-01 -1.42632794e+00 9.42161322e-01
-2.38450468e-01 3.33431542e-01 8.45777839e-02 1.51200366e+00
8.27736259e-01 1.21969923e-01 1.43571988e-01 3.73847097e-01
1.60849857e+00 -7.39310980e-01 -4.70348477e-01 -1.87188074e-01
1.47019774e-01 -5.62903941e-01 3.36989284e-01 4.21067417e-01
-4.36379105e-01 -5.35687745e-01 -1.26160204e+00 4.14679796e-02
-1.30668923e-01 5.74933708e-01 5.57672441e-01 1.23491025e+00
-8.88823688e-01 3.86097282e-01 -8.74552071e-01 -4.31885511e-01
4.89761591e-01 3.95668209e-01 -1.07601851e-01 1.97915241e-01
-1.07597780e+00 1.18437493e+00 -1.07173724e-02 4.95978922e-01
-1.06759608e+00 -5.84278107e-01 -4.48615968e-01 1.45813391e-01
-3.51090133e-01 -6.18139446e-01 1.15884054e+00 -5.92208326e-01
-1.76264226e+00 5.28836012e-01 1.74423799e-01 -8.33085597e-01
6.59385443e-01 -4.33658659e-01 -6.03565574e-01 4.19687510e-01
-4.66903687e-01 1.41246676e-01 1.23242474e+00 -7.13599443e-01
-6.05464578e-01 -6.68631136e-01 1.37075916e-01 -3.76389548e-02
1.69177894e-02 -1.17727384e-01 1.74272910e-01 -5.96606731e-01
2.22733915e-01 -7.13570297e-01 4.89566810e-02 -2.34188676e-01
-5.51885627e-02 1.93086118e-01 1.02857435e+00 -9.88792181e-01
9.17325437e-01 -1.88044047e+00 -3.29735018e-02 2.20781744e-01
2.16789976e-01 4.61998224e-01 -1.69113040e-01 4.13087964e-01
-1.15380302e-01 -5.86377263e-01 -1.33443892e-01 -9.08712819e-02
4.68060933e-02 -6.40534684e-02 -3.64083499e-01 6.52707279e-01
4.47150677e-01 9.39200521e-01 -6.49286389e-01 1.21436134e-01
5.50291419e-01 8.01231265e-01 -3.20219159e-01 2.72875875e-01
1.20859846e-01 3.95835936e-01 -6.50891423e-01 9.11464214e-01
8.33409667e-01 1.38891220e-01 1.76706284e-01 -3.15617234e-01
-3.99545044e-01 5.73127121e-02 -9.22088504e-01 1.43675435e+00
-3.06444854e-01 1.16865373e+00 -1.23136885e-01 -8.73536468e-01
1.28674340e+00 4.35286999e-01 4.16811407e-01 -1.05145943e+00
4.27891374e-01 3.09007376e-01 2.81954050e-01 -1.02208829e+00
5.59052646e-01 -6.22842848e-01 1.64597109e-01 3.78342658e-01
4.00663137e-01 5.60403883e-01 -3.44054997e-01 -4.60060269e-01
1.69783425e+00 2.53548563e-01 -8.99391323e-02 -3.31670567e-02
5.00747621e-01 -6.97062835e-02 4.19478193e-02 9.48684037e-01
-2.70166099e-01 4.40008134e-01 1.19918585e-01 -4.88026351e-01
-1.11083496e+00 -9.98768806e-01 -1.97134763e-01 7.76969612e-01
-2.28457808e-01 1.45516738e-01 -3.44866484e-01 -4.48583424e-01
2.91998476e-01 3.01900059e-01 -6.12517357e-01 2.35580489e-01
-8.09957922e-01 -4.59036589e-01 1.23822761e+00 9.75049794e-01
9.12378430e-01 -8.04658115e-01 -1.49937022e+00 3.58465910e-01
2.53347456e-01 -1.03036785e+00 3.05928916e-01 3.11081886e-01
-7.41327882e-01 -1.14795887e+00 -1.11560667e+00 -7.10813403e-01
1.51830211e-01 -7.18617216e-02 5.97065806e-01 -2.16441944e-01
-6.96641922e-01 6.26093149e-01 -6.64017320e-01 -5.47288775e-01
1.76526412e-01 5.08609042e-02 -2.25350842e-01 9.08080488e-02
6.84297562e-01 -4.48217362e-01 -5.82998574e-01 -2.93580234e-01
-2.87462980e-01 -3.54134291e-01 1.31301594e+00 8.54712069e-01
-2.31587529e-01 -4.82940972e-01 5.44623494e-01 -4.19843882e-01
3.45936000e-01 -7.35431984e-02 -6.74456954e-01 1.92737043e-01
-3.19265187e-01 3.89809608e-01 3.73196006e-01 -9.95003805e-02
-1.06106687e+00 1.31734103e-01 -4.23385680e-01 -4.87147450e-01
-2.10811123e-01 3.93555045e-01 1.45056024e-01 -6.63935483e-01
6.07498050e-01 3.89118612e-01 -1.52173281e-01 -7.11487293e-01
2.66611665e-01 1.25424910e+00 5.98990500e-01 -8.82383808e-02
8.87861550e-01 3.24719340e-01 1.47243276e-01 -1.21122372e+00
6.49403259e-02 -4.05046582e-01 -6.68858469e-01 -3.64970893e-01
7.31728911e-01 -9.91623104e-01 -9.70475852e-01 8.96607935e-01
-1.11934566e+00 -2.32493475e-01 6.30502105e-02 9.49804962e-01
-4.88798440e-01 3.25057596e-01 -4.20331150e-01 -1.24336553e+00
-7.55043387e-01 -5.04961193e-01 1.15906799e+00 4.11867797e-01
2.84228295e-01 -7.46571004e-01 1.73164010e-01 1.81081966e-01
9.31686521e-01 8.26841369e-02 6.18845522e-01 -3.76542658e-01
-8.70245993e-01 -2.13913247e-01 -3.62489492e-01 4.25966904e-02
-1.45869851e-01 -3.90060782e-01 -1.58170891e+00 -3.85197908e-01
-2.43653983e-01 -7.48807564e-02 1.20145512e+00 5.18129706e-01
6.79674745e-01 -4.91930079e-03 -2.84659535e-01 6.01577163e-01
1.62589431e+00 3.81364107e-01 9.88081276e-01 3.27941298e-01
4.90104169e-01 -8.99499953e-02 3.96375179e-01 8.67114305e-01
2.97636718e-01 5.23396075e-01 4.07032371e-01 1.42395049e-01
-3.24975401e-01 1.74772754e-01 2.75281638e-01 9.55724195e-02
-6.02980614e-01 -1.44517735e-01 -9.41800296e-01 3.77305627e-01
-1.57106793e+00 -1.17701447e+00 -1.14800103e-01 1.99106419e+00
-8.43919143e-02 -2.55406976e-01 4.78676101e-03 2.71744639e-01
2.82465070e-01 3.39524597e-01 -3.60013276e-01 -3.76367420e-01
1.80490404e-01 1.06178772e+00 8.90310824e-01 4.99095976e-01
-1.54620481e+00 7.62144744e-01 6.07041407e+00 1.10930569e-01
-1.53722751e+00 4.82937135e-02 -4.74753708e-01 6.95127249e-02
9.29801911e-02 -4.89297301e-01 -4.66263056e-01 1.58678636e-01
1.00958228e+00 2.06536666e-01 4.61769968e-01 6.79384768e-01
-1.49683103e-01 1.68386802e-01 -7.44003892e-01 1.12867820e+00
2.56009489e-01 -1.14202476e+00 -1.14655323e-01 5.13846539e-02
-2.09066030e-02 3.40765476e-01 8.55195746e-02 3.69339794e-01
-3.59690994e-01 -1.10414290e+00 6.52125478e-01 8.12606692e-01
9.63034093e-01 -4.42701340e-01 1.03559208e+00 6.90020919e-02
-1.50709128e+00 -4.10475820e-01 -4.17211533e-01 -2.81625271e-01
-6.96751401e-02 2.63939619e-01 -9.93743122e-01 7.57326126e-01
5.61399698e-01 4.30701703e-01 -2.68539995e-01 1.15360677e+00
-1.92435116e-01 5.15980721e-01 -2.61745870e-01 -3.36608142e-01
4.65449542e-01 2.27241844e-01 7.97139227e-01 1.49812198e+00
5.37100196e-01 4.64553446e-01 -4.72145647e-01 6.37539983e-01
3.15072715e-01 -5.87439299e-01 -7.03667581e-01 -1.56509355e-01
2.26215884e-01 1.13784313e+00 -4.31199595e-02 -2.04669714e-01
-3.99319142e-01 9.33785319e-01 -4.44312766e-02 3.25478792e-01
-9.84581590e-01 -9.63330567e-01 6.89385414e-01 -4.28288989e-02
9.39326704e-01 -6.19005561e-01 -3.51538986e-01 -8.42533350e-01
2.58070249e-02 -3.96864921e-01 3.11613858e-01 -6.69045925e-01
-1.00919926e+00 9.91066217e-01 -4.56677347e-01 -1.32703280e+00
-4.50545967e-01 -1.24942470e+00 -3.74445260e-01 1.14092720e+00
-1.86567521e+00 -1.62420785e+00 -8.04677725e-01 4.09571379e-01
2.33679593e-01 -5.99004984e-01 1.11808062e+00 6.37604356e-01
-1.68817550e-01 8.09655607e-01 1.81260221e-02 5.05637228e-01
3.34964871e-01 -1.20257592e+00 6.75298870e-01 7.86562026e-01
2.57111955e-02 4.38692361e-01 4.74615455e-01 -6.21866703e-01
-2.15022492e+00 -1.08648062e+00 9.40921664e-01 -7.95174167e-02
4.50354636e-01 -1.26291007e-01 -5.68748593e-01 6.62816048e-01
2.76111066e-01 -1.93459451e-01 5.03478169e-01 -3.01682334e-02
-4.31719482e-01 -1.33763775e-01 -1.24223208e+00 -7.72341266e-02
1.07099938e+00 -6.27157629e-01 -9.55864012e-01 -9.51950252e-02
-2.77225316e-01 -6.23783767e-01 -7.10366309e-01 1.93406805e-01
1.34878576e+00 -8.05358708e-01 8.82131934e-01 -6.03314102e-01
8.51029232e-02 -1.61883742e-01 -2.69899696e-01 -1.03782392e+00
-7.11527407e-01 -2.93714195e-01 -5.27896941e-01 3.81381303e-01
-1.26120169e-02 -7.75312364e-01 9.54702914e-01 2.02062577e-01
-1.95559084e-01 -4.00773704e-01 -1.50858867e+00 -8.61812592e-01
-3.22475165e-01 -4.32151467e-01 3.26323658e-01 3.10735285e-01
-1.66433781e-01 2.66540140e-01 -5.13784170e-01 5.23193240e-01
1.08191788e+00 2.11500749e-03 4.83248234e-01 -1.35328043e+00
-3.75631005e-01 -1.63845778e-01 -9.59973097e-01 -1.12753868e+00
-1.38767362e-01 -7.11567163e-01 4.61773830e-04 -1.53203416e+00
-1.09333657e-01 -3.81038040e-01 -2.05027238e-01 6.17630601e-01
5.33916473e-01 5.44719219e-01 2.05427229e-01 -1.67216122e-01
9.20033976e-02 3.95265639e-01 8.04661334e-01 -2.55597293e-01
7.52006918e-02 -8.53300840e-02 -3.37538451e-01 3.97332370e-01
6.01351440e-01 -2.35331893e-01 1.62902161e-01 -5.51345706e-01
-3.64434242e-01 7.28675202e-02 8.71197879e-01 -1.53309929e+00
4.16399926e-01 4.11910921e-01 6.79691851e-01 -7.03943670e-01
3.96108598e-01 -9.94496763e-01 9.81086940e-02 1.16883576e+00
2.07754076e-02 -1.95260614e-01 1.89641491e-01 4.55970585e-01
7.96131492e-02 -5.80452895e-03 6.95359588e-01 3.39809597e-01
-1.05548227e+00 7.69724771e-02 -8.00048709e-01 -5.70377469e-01
7.29190648e-01 -3.20096314e-01 -6.74550474e-01 -2.31956676e-01
-5.47145545e-01 -2.56003439e-01 -5.13396740e-01 4.92890447e-01
1.01517105e+00 -1.27617729e+00 -1.10952580e+00 8.71970594e-01
-6.93182573e-02 -8.86038899e-01 3.54368985e-01 8.39546740e-01
-8.09706628e-01 9.60297406e-01 -9.02543545e-01 -4.55889314e-01
-1.22713995e+00 -1.38913214e-01 7.13969409e-01 -8.40408877e-02
-1.13843906e+00 8.01532924e-01 -6.23482049e-01 3.07052210e-02
3.06357235e-01 -5.81351399e-01 -1.35099068e-01 -4.92692515e-02
6.40808702e-01 4.77286160e-01 3.49523336e-01 -4.61774886e-01
-9.05150354e-01 7.68198431e-01 4.04980779e-02 -3.24314982e-01
1.40665090e+00 5.61369777e-01 5.72852850e-01 -1.27984643e-01
9.11157846e-01 -4.57285613e-01 -1.16632235e+00 -1.38196319e-01
-6.49729371e-02 -6.58469319e-01 3.74937415e-01 -1.16300297e+00
-9.80154097e-01 1.23492074e+00 1.40526736e+00 -1.06779203e-01
1.18100846e+00 -3.05190772e-01 6.37499154e-01 8.86703849e-01
7.41385579e-01 -8.13594639e-01 -1.14553772e-01 9.91883099e-01
1.15582275e+00 -8.34661067e-01 -1.79957092e-01 2.61970550e-01
-4.60308850e-01 1.53234136e+00 3.27934444e-01 -6.05592906e-01
7.88879156e-01 6.80322349e-01 3.37837040e-01 -2.88810641e-01
-4.62050766e-01 -7.09645271e-01 5.19674003e-01 9.12720025e-01
4.85160649e-01 2.67005295e-01 -3.67663234e-01 6.68463945e-01
-8.03142190e-02 3.15940231e-01 1.55591816e-01 9.56585526e-01
-5.60536742e-01 -7.39043415e-01 -3.38482946e-01 6.58779860e-01
7.59085128e-03 6.12840131e-02 -4.07185823e-01 8.39087367e-01
8.16523805e-02 7.01088667e-01 1.60533756e-01 -9.34171855e-01
4.17273134e-01 4.09840234e-02 1.18342662e+00 -5.38347922e-02
-9.76183474e-01 -3.08580548e-02 2.31717959e-01 -4.77949440e-01
-3.70415479e-01 -4.89437401e-01 -1.07520020e+00 -2.78834224e-01
-2.93834388e-01 -3.24550211e-01 6.91969991e-01 1.07056761e+00
3.45519334e-01 7.99968958e-01 4.78581309e-01 -8.25768054e-01
-8.72535467e-01 -1.43668425e+00 -9.83121514e-01 -4.39364642e-01
6.84855342e-01 -8.69061887e-01 8.77925307e-02 -3.35843503e-01] | [6.775153160095215, 0.19853460788726807] |
ec4da963-f521-48af-b34b-a43ce5555319 | face2ppg-an-unsupervised-pipeline-for-blood | 2202.04101 | null | https://arxiv.org/abs/2202.04101v3 | https://arxiv.org/pdf/2202.04101v3.pdf | Face2PPG: An unsupervised pipeline for blood volume pulse extraction from faces | Photoplethysmography (PPG) signals have become a key technology in many fields, such as medicine, well-being, or sports. Our work proposes a set of pipelines to extract remote PPG signals (rPPG) from the face robustly, reliably, and configurable. We identify and evaluate the possible choices in the critical steps of unsupervised rPPG methodologies. We assess a state-of-the-art processing pipeline in six different datasets, incorporating important corrections in the methodology that ensure reproducible and fair comparisons. In addition, we extend the pipeline by proposing three novel ideas; 1) a new method to stabilize the detected face based on a rigid mesh normalization; 2) a new method to dynamically select the different regions in the face that provide the best raw signals, and 3) a new RGB to rPPG transformation method, called Orthogonal Matrix Image Transformation (OMIT) based on QR decomposition, that increases robustness against compression artifacts. We show that all three changes introduce noticeable improvements in retrieving rPPG signals from faces, obtaining state-of-the-art results compared with unsupervised, non-learning-based methodologies and, in some databases, very close to supervised, learning-based methods. We perform a comparative study to quantify the contribution of each proposed idea. In addition, we depict a series of observations that could help in future implementations. | ['Miguel Bordallo López', 'Constantino Álvarez Casado'] | 2022-02-08 | null | null | null | null | ['photoplethysmography-ppg'] | ['medical'] | [ 4.57895041e-01 1.62094146e-01 2.74603188e-01 -3.70566428e-01
-7.11285889e-01 -3.05414617e-01 4.92967367e-01 -4.34661098e-02
-1.80654511e-01 3.71034712e-01 4.79842573e-01 3.82940918e-01
-2.57819414e-01 -5.47932208e-01 -4.78994876e-01 -9.06155109e-01
-1.05408102e-01 2.30793327e-01 4.49272580e-02 -1.50006011e-01
5.04747152e-01 8.98503900e-01 -1.78871846e+00 2.98412770e-01
5.13157725e-01 1.00504971e+00 -3.92493010e-01 3.98127109e-01
7.53632188e-02 2.94279516e-01 -3.03916782e-01 -5.03167927e-01
5.02267420e-01 -5.56935728e-01 -5.18551707e-01 1.21973410e-01
6.09075963e-01 -1.61274910e-01 1.06346227e-01 7.53901660e-01
9.61995244e-01 3.49899977e-02 5.54574013e-01 -8.30132127e-01
-2.07004324e-01 2.10612580e-01 -7.64670491e-01 -1.13075644e-01
6.17925584e-01 1.42709419e-01 3.98820311e-01 -1.01029551e+00
6.79960012e-01 1.07141089e+00 7.76523173e-01 6.12379313e-01
-1.40624452e+00 -5.86366534e-01 -5.97491443e-01 2.05334365e-01
-1.33651316e+00 -1.04532647e+00 1.01721787e+00 -2.52524942e-01
6.91274703e-01 4.65085715e-01 6.52131677e-01 8.98209870e-01
2.66563762e-02 1.23806670e-01 1.52027977e+00 -7.36121297e-01
2.92416930e-01 2.30541736e-01 -2.57969588e-01 4.96532142e-01
2.09932819e-01 4.78483475e-04 -1.06623495e+00 -3.20915341e-01
5.73207378e-01 -4.48717147e-01 -2.93879569e-01 -5.81968427e-01
-8.86740565e-01 5.05849123e-01 -3.62965256e-01 5.16745269e-01
-5.09673655e-01 -4.83613787e-03 3.86006534e-01 1.07497402e-01
5.47099233e-01 3.22226375e-01 -2.92732626e-01 -2.77310729e-01
-1.21854115e+00 -2.44583488e-02 1.05434299e+00 5.35609126e-01
7.87821472e-01 -8.26968700e-02 -1.75637126e-01 8.95048618e-01
3.96450520e-01 3.86931449e-01 4.74550575e-01 -1.01592767e+00
3.40177901e-02 3.08111370e-01 -2.08085865e-01 -1.25872004e+00
-4.58543926e-01 -1.38423011e-01 -5.08327305e-01 1.10420175e-01
2.69694090e-01 8.11970830e-02 -5.28543293e-01 1.39750051e+00
4.91432965e-01 4.44854915e-01 -1.95000947e-01 8.77182007e-01
8.89755368e-01 2.02282652e-01 -1.68134674e-01 -6.71299815e-01
1.37808800e+00 -3.41526568e-01 -6.80648625e-01 2.16097534e-01
1.47333860e-01 -9.73482251e-01 6.78412080e-01 7.73826063e-01
-1.09055936e+00 -4.81773496e-01 -8.99371386e-01 6.76001981e-02
-2.41685629e-01 2.50801146e-01 3.85828197e-01 1.27876675e+00
-1.29196012e+00 9.45854902e-01 -6.33435667e-01 -6.84997201e-01
6.33480623e-02 5.80860496e-01 -6.40775204e-01 2.52853185e-01
-7.82307506e-01 8.88856411e-01 -1.38638809e-01 3.56611639e-01
-2.41825074e-01 -6.01173818e-01 -7.99444973e-01 -9.24345404e-02
2.07572997e-01 -3.35754037e-01 7.39217639e-01 -9.10341561e-01
-2.20145369e+00 1.25916505e+00 -3.49164665e-01 -2.98596174e-01
4.63868797e-01 -6.95080757e-02 -3.81249607e-01 6.52336836e-01
-3.44554812e-01 4.41682130e-01 1.36143327e+00 -1.11480653e+00
2.42298041e-02 -5.17604053e-01 -7.70743072e-01 7.05800802e-02
-3.95062566e-01 3.93548965e-01 -6.57007635e-01 -4.30550039e-01
2.43076190e-01 -8.40488613e-01 5.16696274e-02 3.23696062e-02
-2.87491959e-02 -4.48017381e-02 4.76387799e-01 -8.77187014e-01
8.94605577e-01 -2.15691590e+00 1.60626754e-01 6.15851402e-01
1.68408290e-01 3.23207259e-01 -2.45893616e-02 5.43335676e-01
-2.36728862e-01 -1.41445383e-01 -2.23173857e-01 -7.12066889e-01
-7.62602985e-02 -1.66092832e-02 -1.26083493e-01 8.58338952e-01
1.12948500e-01 6.64235473e-01 -5.73677003e-01 -4.73768860e-01
5.00399232e-01 7.84115851e-01 -2.96174079e-01 -6.96952716e-02
4.38895464e-01 6.08305633e-01 1.94086898e-02 6.97042882e-01
9.90441859e-01 4.12463367e-01 3.71467143e-01 -6.07198656e-01
-2.06644133e-01 1.05335057e-01 -1.38995886e+00 1.46792579e+00
-1.39071777e-01 5.30864477e-01 9.47986692e-02 -8.74065936e-01
1.33860922e+00 4.60167646e-01 9.15289402e-01 -5.64597785e-01
2.26157606e-01 3.13621312e-01 -2.61775047e-01 -7.91994870e-01
3.49630356e-01 -2.81823158e-01 4.38732058e-01 3.86478692e-01
3.64017814e-01 -1.00285150e-01 1.65934235e-01 -1.74272642e-01
6.81605041e-01 3.12955379e-01 4.06794012e-01 -4.85703737e-01
7.96573639e-01 -7.10515916e-01 4.52002913e-01 3.48102033e-01
-3.38283867e-01 1.13295615e+00 5.56046844e-01 -1.52513281e-01
-6.26309156e-01 -6.04101479e-01 -1.35743320e-01 6.09571338e-01
-1.35931119e-01 -5.23232043e-01 -8.57959390e-01 -1.94427401e-01
4.44147289e-02 1.81039125e-01 -7.45169163e-01 3.24501246e-02
-5.44986904e-01 -9.55042243e-01 4.45321977e-01 1.50573179e-01
3.39959651e-01 -1.03538954e+00 -5.88258982e-01 3.85619551e-02
-9.99327078e-02 -1.05394351e+00 -2.52491415e-01 -2.46157989e-01
-1.06557167e+00 -1.07190537e+00 -6.92583859e-01 -3.49001795e-01
6.13709927e-01 4.79462408e-02 8.54766071e-01 -3.59793678e-02
-4.34070230e-01 8.82535040e-01 -3.93434703e-01 -2.32839361e-01
-3.04531246e-01 -2.82526344e-01 3.05746645e-01 7.18658447e-01
6.15522385e-01 -7.23406613e-01 -7.31519938e-01 2.31752768e-01
-6.22341394e-01 -4.16135371e-01 5.25498748e-01 2.95393407e-01
6.93007588e-01 -3.11906308e-01 2.52143532e-01 -9.09157634e-01
6.50636137e-01 -1.19543701e-01 -5.12193024e-01 1.97492808e-01
-7.40522385e-01 -2.80196662e-04 4.01808619e-01 -2.45813906e-01
-9.00588095e-01 4.35247779e-01 -1.67374685e-01 -4.62741256e-01
-2.03572750e-01 1.28928423e-01 -1.11296378e-01 -8.00899565e-01
8.98169994e-01 4.51110274e-01 4.10142452e-01 -5.76949716e-01
3.91533852e-01 5.85342824e-01 6.27022743e-01 -5.02118051e-01
8.72203827e-01 6.57311440e-01 3.79587770e-01 -1.28052688e+00
8.39686692e-02 -7.72183061e-01 -9.30683732e-01 -5.10026753e-01
4.22385573e-01 -5.00512898e-01 -8.89712989e-01 6.24504030e-01
-8.23246837e-01 6.43452480e-02 -2.74661660e-01 5.79589963e-01
-5.97959280e-01 8.56116295e-01 -3.48743558e-01 -9.29968834e-01
-5.54500163e-01 -8.54743481e-01 1.11462760e+00 4.37749982e-01
-1.60883442e-01 -7.41231441e-01 4.06260461e-01 4.86471921e-01
4.92899090e-01 4.72745121e-01 2.01020479e-01 -2.83294708e-01
-1.90788969e-01 -8.87695625e-02 6.13255203e-02 5.30579507e-01
2.39764780e-01 3.85003090e-01 -1.36328685e+00 -3.61273170e-01
3.31355870e-01 7.99928904e-02 6.56564057e-01 4.06612903e-01
8.11574578e-01 -6.11848757e-02 -1.07233211e-01 8.19190562e-01
1.31315529e+00 1.68144349e-02 1.26896858e+00 9.46198180e-02
3.33137244e-01 9.32053506e-01 3.00436825e-01 3.80282611e-01
-3.52909677e-02 7.55162477e-01 1.27156824e-01 -1.90891623e-01
-2.39274412e-01 3.40612456e-02 4.17733967e-01 7.90814936e-01
-5.79546988e-01 4.68469262e-01 -4.21631247e-01 3.92124802e-01
-1.47626162e+00 -8.10938299e-01 -2.36878604e-01 2.50935745e+00
6.41409874e-01 -2.96739966e-01 3.31761926e-01 4.22426641e-01
5.43237388e-01 6.78943172e-02 -1.26549765e-01 -5.99722326e-01
-9.13077667e-02 7.91621089e-01 4.61586326e-01 3.54841620e-01
-9.22550082e-01 5.76687455e-01 6.62254000e+00 3.80905658e-01
-1.59685135e+00 -1.91764962e-02 3.24524134e-01 7.45481923e-02
-1.91788655e-02 -5.46373837e-02 -6.41011059e-01 5.09384274e-01
9.34936643e-01 -3.08701564e-02 5.27961433e-01 5.56591749e-01
4.54881877e-01 -2.49498889e-01 -9.03934956e-01 1.29869831e+00
7.82799840e-01 -1.01870978e+00 -2.68209875e-01 9.64463055e-02
3.13255519e-01 -2.99527168e-01 -1.01436362e-01 -4.13601995e-01
-6.74321532e-01 -7.14949906e-01 4.02251035e-01 7.37865150e-01
8.78547430e-01 -4.87067163e-01 4.80313838e-01 -3.24972659e-01
-1.04499209e+00 2.55864531e-01 -4.20835108e-01 2.31710792e-01
-1.68533638e-01 7.78638005e-01 -6.39905632e-01 8.28587890e-01
7.17779100e-01 5.78782916e-01 -7.19985306e-01 1.20442963e+00
-4.30855244e-01 6.32953286e-01 -3.64888489e-01 7.54588172e-02
-4.87021685e-01 -4.01795417e-01 6.91055298e-01 1.18760431e+00
3.76227260e-01 2.43328348e-01 -6.48021042e-01 6.59381211e-01
1.01976603e-01 6.28881037e-01 -4.68198359e-01 1.70459241e-01
2.14854226e-01 1.61019623e+00 -9.58291352e-01 6.51713535e-02
-3.77027363e-01 8.72966647e-01 -1.18049540e-01 2.26478249e-01
-3.61208797e-01 -4.69549477e-01 3.41379225e-01 3.16989809e-01
3.28084946e-01 -2.55036861e-01 -2.30786651e-01 -1.13691056e+00
2.60681778e-01 -8.89780879e-01 4.12245750e-01 -6.83368981e-01
-1.07561493e+00 4.52157915e-01 3.70998122e-02 -1.29681444e+00
-3.40661377e-01 -5.96215725e-01 -4.66010511e-01 8.30517650e-01
-1.54229259e+00 -9.40679193e-01 -4.53986943e-01 5.76005220e-01
-1.57748953e-01 6.12480892e-03 9.32870328e-01 3.99131417e-01
-3.92372400e-01 6.27980351e-01 -7.64922425e-02 -1.72392562e-01
1.17112136e+00 -9.71026540e-01 9.71123576e-02 9.21970129e-01
2.57144451e-01 8.22321117e-01 6.98566198e-01 -3.77088547e-01
-1.71330404e+00 -5.10768890e-01 8.93790007e-01 -3.54485393e-01
1.94375783e-01 -4.93831992e-01 -8.12134862e-01 3.24595779e-01
9.80721042e-02 -7.08430707e-02 8.44876528e-01 -3.50092398e-03
-2.09439859e-01 -6.69309497e-01 -1.47631025e+00 2.12151051e-01
5.85590601e-01 -4.53582048e-01 -6.62023783e-01 5.68000749e-02
-1.43326670e-01 -4.07649606e-01 -8.79246712e-01 2.86224365e-01
9.43433523e-01 -1.35077035e+00 8.50559115e-01 1.08925134e-01
3.88023444e-02 -3.93499047e-01 2.96885550e-01 -1.01511776e+00
-1.29524201e-01 -1.19377923e+00 -4.33401801e-02 1.36401117e+00
6.91000670e-02 -1.03609133e+00 9.32778478e-01 7.68020391e-01
2.06778929e-01 -5.31137764e-01 -9.44353342e-01 -4.74112839e-01
-4.23155129e-01 -1.87488452e-01 3.76795262e-01 8.58088613e-01
1.96402445e-01 -2.03506395e-01 -5.41757286e-01 -3.35887563e-03
7.85875559e-01 3.27589922e-02 9.54786062e-01 -1.21494770e+00
-1.95246473e-01 -2.02413857e-01 -8.61042559e-01 -4.13620025e-01
-2.08161101e-01 -5.01750767e-01 -8.46506581e-02 -1.22295070e+00
9.48998556e-02 6.39723316e-02 -2.94482231e-01 5.64779937e-01
1.77182049e-01 8.07720780e-01 9.26040113e-03 1.16090007e-01
5.38708083e-02 1.91136852e-01 7.59090662e-01 3.33751142e-01
-4.76470888e-01 -2.12737337e-01 -8.05021048e-01 4.43417519e-01
6.04192078e-01 -4.08637524e-01 -2.56277144e-01 1.57416061e-01
8.40991363e-02 -1.63627654e-01 1.12842880e-01 -1.06739020e+00
3.82302791e-01 1.68441460e-01 4.46078777e-01 -3.56638223e-01
4.48441803e-01 -6.78764343e-01 3.42473269e-01 3.85857075e-01
2.12255970e-01 -4.64730382e-01 2.71007389e-01 7.77878193e-03
-1.22346908e-01 -2.34414756e-01 1.05072403e+00 -6.16332702e-02
-4.73371863e-01 -5.89922667e-02 -2.63180405e-01 -4.17213619e-01
7.70279586e-01 -5.12457073e-01 -4.83360142e-02 -5.27946055e-01
-6.45361602e-01 -2.44376183e-01 4.37372714e-01 1.84034944e-01
5.87180614e-01 -9.49393392e-01 -8.21802616e-01 6.20481074e-01
-7.36657083e-02 -6.28683507e-01 2.43314117e-01 1.26584101e+00
-7.07923710e-01 1.69885799e-01 -4.18897629e-01 -6.03029549e-01
-1.63765121e+00 2.49914736e-01 4.08013225e-01 1.43880099e-01
-5.28778553e-01 6.34208560e-01 -3.30713481e-01 -2.96666861e-01
-1.91188287e-02 -1.73739940e-01 -4.66485918e-01 3.19065332e-01
5.54580092e-01 6.35909379e-01 7.02938974e-01 -9.23346162e-01
-4.46510226e-01 1.10539031e+00 2.68363744e-01 -1.65093675e-01
1.41713738e+00 -1.12420820e-01 -5.39360881e-01 1.45572856e-01
9.51777339e-01 3.99608493e-01 -8.85062993e-01 1.04178026e-01
-5.16234450e-02 -6.80366158e-01 -1.05628915e-01 -5.27145386e-01
-1.24319100e+00 6.79416656e-01 8.86818767e-01 4.67968397e-02
1.55412364e+00 -3.67635161e-01 2.98646241e-01 1.35043263e-02
4.45346981e-01 -1.04496241e+00 -1.44508019e-01 -1.33555695e-01
8.98643017e-01 -7.78305590e-01 4.60683107e-01 -8.21755469e-01
-3.26772660e-01 1.51327717e+00 -2.48287013e-03 -9.41254795e-02
7.20014811e-01 2.08721355e-01 2.08482891e-01 -3.01461458e-01
-2.42977232e-01 -1.16789140e-01 4.96094137e-01 7.03075528e-01
4.76912290e-01 -2.58187294e-01 -8.07773054e-01 2.86063075e-01
-1.64350078e-01 3.88274878e-01 4.74204838e-01 8.02760839e-01
-3.40824164e-02 -1.40968406e+00 -6.18296504e-01 3.63159299e-01
-6.36002719e-01 9.32534039e-02 -5.86026251e-01 6.60949707e-01
2.06260588e-02 9.00712609e-01 -3.18620265e-01 -2.58635700e-01
5.20686448e-01 2.72943437e-01 8.88121307e-01 -3.97504449e-01
-6.52300358e-01 3.49934608e-01 -1.40153738e-02 -8.56473684e-01
-8.63153100e-01 -9.12819862e-01 -7.58436382e-01 -1.13190033e-01
-2.76796490e-01 3.43299173e-02 8.48410368e-01 8.05188835e-01
5.61783969e-01 -1.60828695e-01 8.45995903e-01 -1.05735123e+00
-5.16067259e-02 -9.57572103e-01 -7.82084048e-01 4.22746241e-01
-3.56972180e-02 -5.65871537e-01 -5.69140971e-01 2.95465022e-01] | [13.835180282592773, 2.634943723678589] |
206fff5c-70ea-4cc3-9a56-3495e63a154a | tugas-exploiting-unlabelled-data-for-twitter | null | null | https://aclanthology.org/S14-2120 | https://aclanthology.org/S14-2120.pdf | TUGAS: Exploiting unlabelled data for Twitter sentiment analysis | null | ["M{\\'a}rio J. Silva", 'Jo{\\~a}o Filgueiras', 'Miguel B. Almeida', 'Silvio Amir', 'Bruno Martins'] | 2014-08-01 | null | null | null | semeval-2014-8 | ['twitter-sentiment-analysis'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.206260681152344, 3.6059210300445557] |
ab8d473d-2fbc-4a48-9203-b7f845493666 | back-to-patterns-efficient-japanese | 2305.19045 | null | https://arxiv.org/abs/2305.19045v1 | https://arxiv.org/pdf/2305.19045v1.pdf | Back to Patterns: Efficient Japanese Morphological Analysis with Feature-Sequence Trie | Accurate neural models are much less efficient than non-neural models and are useless for processing billions of social media posts or handling user queries in real time with a limited budget. This study revisits the fastest pattern-based NLP methods to make them as accurate as possible, thus yielding a strikingly simple yet surprisingly accurate morphological analyzer for Japanese. The proposed method induces reliable patterns from a morphological dictionary and annotated data. Experimental results on two standard datasets confirm that the method exhibits comparable accuracy to learning-based baselines, while boasting a remarkable throughput of over 1,000,000 sentences per second on a single modern CPU. The source code is available at https://www.tkl.iis.u-tokyo.ac.jp/~ynaga/jagger/ | ['Naoki Yoshinaga'] | 2023-05-30 | null | null | null | null | ['morphological-analysis'] | ['natural-language-processing'] | [ 9.28008035e-02 -8.67479667e-02 -1.62371039e-01 -2.45558560e-01
-1.15052629e+00 -7.05552161e-01 -7.01760640e-03 2.39894181e-01
-1.04129708e+00 7.81766117e-01 7.55877467e-03 -7.43367732e-01
4.11525190e-01 -8.39304686e-01 -6.50033534e-01 -5.63838661e-01
3.06347013e-01 7.59364843e-01 2.81796098e-01 -2.88894605e-02
1.99133784e-01 3.83468121e-01 -1.09434688e+00 4.40818548e-01
6.56316161e-01 7.19363093e-01 5.11128783e-01 7.96143711e-01
-1.93529308e-01 5.08935809e-01 -6.56177878e-01 -9.31706309e-01
2.87799120e-01 -1.15867913e-01 -9.78626907e-01 -4.33514804e-01
6.09334767e-01 -1.61636561e-01 -4.07501608e-01 1.54033232e+00
5.12361825e-01 9.31711122e-02 2.39315584e-01 -3.85939002e-01
-8.07520032e-01 1.08891094e+00 -2.65342772e-01 7.64410675e-01
3.19835991e-01 1.33645058e-01 1.43145204e+00 -1.06189251e+00
5.81765234e-01 8.89619231e-01 6.78678334e-01 5.17904222e-01
-9.41018641e-01 -5.57728529e-01 1.25935199e-02 2.91870147e-01
-1.32901669e+00 -6.77249670e-01 6.17050469e-01 4.26166467e-02
1.59249306e+00 4.07444596e-01 4.92543280e-01 1.06583023e+00
8.25831145e-02 9.92095590e-01 1.01004839e+00 -6.99556172e-01
-1.17998794e-02 -1.74121931e-01 6.71559572e-01 1.16030395e+00
5.90724468e-01 -3.77489507e-01 -5.91285169e-01 -3.36260498e-01
5.34299672e-01 -3.24325860e-01 -2.03433558e-01 5.63505828e-01
-1.41244161e+00 8.31197977e-01 -1.26039878e-01 6.36772275e-01
-3.64088774e-01 -1.32292271e-01 4.89062726e-01 4.93857235e-01
2.74438709e-01 7.88858593e-01 -1.03204095e+00 -4.42641169e-01
-7.21032262e-01 1.59337282e-01 1.04864693e+00 1.12429070e+00
6.39485061e-01 3.66442412e-01 3.54172945e-01 9.80211139e-01
-3.02895814e-01 8.20911169e-01 8.55416358e-01 -9.12268996e-01
8.14185739e-01 2.95422941e-01 -1.72152206e-01 -1.25987589e+00
-5.94763279e-01 -2.40885392e-01 -8.09939563e-01 -6.50996149e-01
7.35198617e-01 -1.75492570e-01 -5.83863735e-01 1.58445632e+00
1.65756822e-01 -3.14953357e-01 -1.45044550e-01 4.02336627e-01
6.63124621e-01 8.90210748e-01 -2.16412023e-01 -4.93512630e-01
1.55045688e+00 -9.82537150e-01 -7.37240911e-01 -3.86864245e-01
8.32078457e-01 -8.99614811e-01 1.61166215e+00 8.37704182e-01
-1.40405750e+00 -3.71508211e-01 -8.32921028e-01 -1.85648993e-01
-4.30766791e-01 4.34527487e-01 9.24077928e-01 8.25003326e-01
-7.90232956e-01 6.66774213e-01 -9.98616874e-01 -3.69942367e-01
2.58486867e-01 4.45398569e-01 -2.62415171e-01 2.08211035e-01
-9.97351348e-01 6.89954221e-01 7.08134413e-01 1.27463222e-01
-1.03244334e-01 -3.21964949e-01 -7.46491313e-01 -1.87591925e-01
4.86806273e-01 -4.96886551e-01 1.68586588e+00 -7.16725051e-01
-1.59539866e+00 9.31954920e-01 -4.58376020e-01 -5.89115143e-01
6.21238276e-02 -3.74351352e-01 -5.62379301e-01 2.26107866e-01
-2.37988919e-01 3.29519242e-01 4.22861695e-01 -4.41913486e-01
-5.78171670e-01 -2.92930037e-01 -4.70693648e-01 -4.68122251e-02
-6.62681520e-01 5.03702700e-01 -5.53569138e-01 -8.93260896e-01
3.00248712e-02 -9.47494984e-01 -2.01413527e-01 -4.56688017e-01
-6.58598840e-01 -3.55498195e-01 4.27303091e-02 -1.05768847e+00
1.48274374e+00 -1.82526207e+00 -1.78757951e-01 2.48721819e-02
-9.15621221e-03 4.08618420e-01 -2.81361550e-01 4.04734731e-01
2.42866755e-01 2.78843433e-01 -5.56377649e-01 -2.08647683e-01
2.57410239e-02 3.93742710e-01 -3.22501838e-01 4.11078602e-01
-8.49314407e-02 1.04533398e+00 -8.61395955e-01 -6.35597944e-01
-3.17905784e-01 -1.45915747e-01 -5.29336810e-01 4.24946211e-02
-1.26971915e-01 -7.62551948e-02 -9.43051949e-02 9.84023690e-01
2.57212758e-01 -2.54295260e-01 5.06997168e-01 3.38666141e-02
1.86573360e-02 8.79708350e-01 -8.98230672e-01 1.55788147e+00
-2.24276423e-01 5.60532451e-01 -1.85429811e-01 -9.63557243e-01
7.43323565e-01 1.29051790e-01 -2.77635660e-02 -6.83803499e-01
2.69746542e-01 4.50576127e-01 1.93849802e-01 -4.79683816e-01
7.47052908e-01 6.09437227e-02 -4.97473150e-01 5.85851729e-01
2.32654139e-01 -8.79947916e-02 7.72512138e-01 1.32156923e-01
1.25988317e+00 -3.16971034e-01 7.39915848e-01 -2.66641706e-01
3.03728133e-01 2.20637798e-01 1.06794369e+00 8.93830597e-01
-1.36562020e-01 3.46048236e-01 3.30710858e-01 -6.66687369e-01
-1.30594194e+00 -9.14629281e-01 1.42754287e-01 1.26384711e+00
-3.14601660e-01 -4.38271195e-01 -9.02655661e-01 -5.35985827e-01
-3.14175814e-01 7.31300533e-01 -1.00010917e-01 2.71098673e-01
-1.22836983e+00 -1.22220409e+00 1.02913654e+00 5.15934348e-01
3.58480573e-01 -1.46053553e+00 -1.51001990e-01 3.23607057e-01
-4.84096169e-01 -1.32177055e+00 -5.39123535e-01 3.00388545e-01
-1.08659208e+00 -8.12774777e-01 -4.01000321e-01 -1.17264748e+00
6.97903752e-01 -3.91146131e-02 1.23541427e+00 1.64548650e-01
-1.33624479e-01 -3.97874206e-01 -2.83521861e-01 -3.66575599e-01
-5.60411453e-01 6.27884388e-01 5.35156071e-01 -3.97561342e-01
7.80979037e-01 -5.91519296e-01 -4.84681176e-03 -1.43093273e-01
-6.17777646e-01 -4.13162820e-02 8.06330323e-01 9.45719004e-01
6.83573067e-01 9.68488008e-02 4.63055074e-01 -1.33044589e+00
6.90470159e-01 -2.64017701e-01 -7.68776655e-01 1.24862053e-01
-5.28268933e-01 -1.38728291e-01 9.68448818e-01 -7.39802241e-01
-8.80484164e-01 1.04496449e-01 -6.71411097e-01 2.75499195e-01
-4.54261392e-01 5.89551091e-01 -2.25686371e-01 3.07028621e-01
6.22021973e-01 5.77131748e-01 -3.15573961e-01 -9.47238684e-01
1.47336155e-01 7.49355733e-01 7.44746387e-01 -5.42189717e-01
8.27542603e-01 1.38319358e-01 -5.77746511e-01 -1.07124853e+00
-8.99761975e-01 -3.60191077e-01 -7.71970332e-01 3.85425061e-01
3.59430432e-01 -5.34009159e-01 -5.64391434e-01 7.66577542e-01
-1.31646585e+00 -5.74625432e-01 -2.42287070e-02 4.19989169e-01
-3.41086656e-01 7.30413795e-01 -1.45570934e+00 -3.46954107e-01
-7.12901711e-01 -5.78873992e-01 5.44035614e-01 6.79667443e-02
-5.72713315e-01 -6.68206215e-01 8.13582838e-02 4.48516458e-01
1.08079970e-01 -5.02477884e-01 1.28596699e+00 -1.16386235e+00
-3.09413373e-01 -2.43216306e-01 -5.44446148e-02 4.17406648e-01
-5.57807349e-02 2.60394253e-03 -7.77979434e-01 2.39664316e-02
1.15492687e-01 -3.35366249e-01 7.20729113e-01 1.25590321e-02
1.23787558e+00 -8.75800490e-01 -3.43678780e-02 4.58426625e-01
1.24521303e+00 1.92571506e-01 3.53855550e-01 2.87605792e-01
6.07616901e-01 3.10111016e-01 4.01110053e-01 5.93701601e-02
3.09658915e-01 3.68132800e-01 -2.59396076e-01 2.59308040e-01
-1.69657301e-02 -2.92509586e-01 4.54618841e-01 2.10048246e+00
-3.20002325e-02 -3.35111409e-01 -1.10137391e+00 7.06319988e-01
-1.63232613e+00 -9.32238460e-01 -2.31383190e-01 1.99422860e+00
1.47508669e+00 3.58550698e-01 1.77078053e-01 1.39441431e-01
5.54183185e-01 4.21345979e-02 -2.50076145e-01 -6.07439995e-01
-3.76912802e-01 2.76975632e-01 5.86867511e-01 6.37851238e-01
-1.03807056e+00 1.34038663e+00 6.50874281e+00 1.14608753e+00
-8.80415857e-01 3.51674557e-01 5.94815850e-01 -3.04846376e-01
-1.48233593e-01 -3.81425619e-01 -1.26893377e+00 5.13008177e-01
1.34812593e+00 -2.82543693e-02 5.70461035e-01 9.29378748e-01
4.51237597e-02 1.22266509e-01 -6.71487153e-01 1.02099109e+00
1.83340445e-01 -1.50401735e+00 1.36352137e-01 8.35269541e-02
3.01208198e-01 5.38414240e-01 -2.33005449e-01 2.51410186e-01
3.71121198e-01 -5.48133910e-01 8.21811199e-01 2.31754512e-01
6.24931753e-01 -7.39593923e-01 8.11226606e-01 6.05707407e-01
-8.42854559e-01 -5.56154065e-02 -7.96607137e-01 -2.88222581e-01
1.85205132e-01 8.68939579e-01 -9.02894676e-01 8.81454796e-02
5.99598885e-01 2.59413868e-01 -7.75526941e-01 7.84901798e-01
-4.11696583e-01 1.25059080e+00 -5.85421026e-01 -6.26894295e-01
1.67029411e-01 -1.07804090e-02 5.72927415e-01 1.82951653e+00
2.10627869e-01 1.12644732e-01 7.63984025e-02 3.72374028e-01
-4.60145712e-01 5.26474357e-01 -6.21968448e-01 -2.47482389e-01
6.83222234e-01 1.20913243e+00 -9.35112834e-01 -6.12128675e-01
-4.04680520e-01 9.16891634e-01 9.57134366e-01 2.35376909e-01
-6.96701407e-01 -7.69949675e-01 4.05639827e-01 -3.44333559e-01
3.22747827e-01 -6.18370712e-01 -4.16405469e-01 -1.42689300e+00
1.97339565e-01 -1.25130308e+00 5.52091241e-01 -4.46653247e-01
-1.42132819e+00 8.57568383e-01 -3.11698020e-01 -6.14408910e-01
-1.83273688e-01 -1.16165066e+00 -2.30559945e-01 4.63810474e-01
-1.17767847e+00 -8.29463124e-01 4.22267169e-01 4.07029301e-01
7.01892138e-01 -2.86629468e-01 1.13042581e+00 3.21722090e-01
-9.37363923e-01 8.29566538e-01 -7.61819771e-03 5.65787792e-01
3.75467390e-01 -1.34201491e+00 9.06398237e-01 1.20227814e+00
7.90092707e-01 8.33271861e-01 4.61590499e-01 -6.83870852e-01
-1.55027473e+00 -7.41990626e-01 1.74814677e+00 -4.78943437e-01
1.03460073e+00 -5.42604625e-01 -1.08214295e+00 6.24876738e-01
1.69246182e-01 -2.17372090e-01 8.91091406e-01 3.36894661e-01
-4.92937595e-01 -4.53242175e-02 -7.35748708e-01 9.11145687e-01
1.35255671e+00 -6.18883908e-01 -9.41341996e-01 5.18446267e-01
8.26489091e-01 -3.55566561e-01 -6.68449819e-01 1.36084899e-01
4.28664297e-01 -4.77109790e-01 6.29267275e-01 -5.45576692e-01
-3.24554183e-02 -7.77660403e-03 -3.04686040e-01 -8.90623868e-01
-4.28560495e-01 -9.86640990e-01 -3.10013801e-01 8.87761295e-01
9.74761307e-01 -8.31931353e-01 8.03575575e-01 1.32082641e-01
-2.05669999e-01 -8.25209916e-01 -9.33172464e-01 -9.96962249e-01
3.05307526e-02 -7.92978823e-01 4.20554399e-01 9.94618297e-01
1.91934153e-01 4.10389960e-01 -2.81398922e-01 1.30829901e-01
4.32873040e-01 2.14564264e-01 5.35972714e-01 -8.53064120e-01
-6.73150837e-01 -5.20405650e-01 -6.14226013e-02 -1.17435575e+00
2.28476435e-01 -1.04176188e+00 1.54338285e-01 -1.05767059e+00
2.03162253e-01 -3.85966897e-01 -1.75117657e-01 8.75365257e-01
-1.84925348e-01 6.15195870e-01 4.01058979e-02 2.79886842e-01
-4.95958358e-01 1.12665392e-01 7.97540724e-01 -3.34611870e-02
-2.81293560e-02 -5.56919090e-02 -8.60551715e-01 1.19265437e+00
1.23200870e+00 -9.26510811e-01 1.13797516e-01 -8.69853795e-01
6.92805529e-01 -4.51909423e-01 -2.99872369e-01 -7.68554091e-01
4.66014415e-01 -1.87555224e-01 9.96744260e-02 -5.56735814e-01
2.32875109e-01 -2.75917768e-01 -3.80892277e-01 4.96133655e-01
-1.95196509e-01 5.80544114e-01 2.39101887e-01 2.75421202e-01
8.99282936e-03 -5.56250036e-01 5.60097456e-01 -4.21760619e-01
-5.88967741e-01 1.39672980e-01 -8.75411689e-01 2.36751556e-01
2.16404364e-01 -4.18828465e-02 -3.59694332e-01 -3.35470699e-02
-4.68363166e-01 -4.06944036e-01 2.93447942e-01 1.02307364e-01
4.41021740e-01 -8.47681820e-01 -5.70510983e-01 2.06488714e-01
-5.50193861e-02 -2.79760480e-01 -1.34045780e-01 6.43568933e-01
-8.63571286e-01 6.52826726e-01 2.02439353e-02 -7.67083168e-02
-1.44978082e+00 4.95774359e-01 -2.33094655e-02 -5.68464100e-01
-6.83486938e-01 1.04634142e+00 -4.14171398e-01 -7.69132912e-01
2.28818148e-01 -6.71083212e-01 1.18656032e-01 -4.64224033e-02
7.92580485e-01 2.65321374e-01 3.51701498e-01 -3.68035227e-01
-1.21357180e-01 1.57593250e-01 -1.81433618e-01 -1.98298320e-01
1.35512710e+00 1.15722381e-01 -4.81622070e-01 5.89932144e-01
9.35873270e-01 4.92821455e-01 -3.76301587e-01 -4.88106817e-01
5.16002417e-01 -1.19048364e-01 -1.23301074e-01 -5.64145744e-01
-7.49245524e-01 7.48724282e-01 -2.26792678e-01 1.80786148e-01
1.04601490e+00 -4.18417156e-02 1.61744165e+00 1.38183570e+00
5.49711943e-01 -1.30147290e+00 2.23251954e-02 1.01453590e+00
5.44556201e-01 -8.61009896e-01 -1.13821134e-01 -4.82739836e-01
-4.91787136e-01 1.04084432e+00 3.33440304e-01 -9.86448824e-02
5.33263564e-01 6.13047779e-01 1.16885215e-01 9.48796272e-02
-8.97365034e-01 -5.29364049e-02 2.33197752e-02 4.05292094e-01
2.42422402e-01 1.91739649e-01 -4.55948621e-01 9.10005748e-01
-6.81217670e-01 -3.96685392e-01 5.83187521e-01 8.94053042e-01
-5.75514317e-01 -1.21995604e+00 -8.95022303e-02 6.22797072e-01
-1.02818787e+00 -7.52316535e-01 -4.44254726e-01 6.12984478e-01
-2.49990761e-01 6.97994232e-01 1.23256840e-01 -3.30887288e-01
4.77073453e-02 5.58443844e-01 5.52147746e-01 -9.35516357e-01
-6.78063095e-01 5.71056567e-02 6.35497689e-01 -4.65694696e-01
-6.57485127e-02 -8.81600440e-01 -1.21165812e+00 -5.01243174e-01
-1.51335016e-01 1.58909574e-01 5.08067608e-01 8.11010063e-01
3.17262411e-01 -1.55591547e-01 2.54042774e-01 -5.04477084e-01
-7.22197890e-01 -1.23205733e+00 -4.38616276e-01 -3.60597894e-02
-1.25698864e-01 4.09527719e-02 -2.24295497e-01 2.89349437e-01] | [10.445728302001953, 10.078166007995605] |
71858018-562f-43f9-b84f-37b752a3c639 | pixelsteganalysis-pixel-wise-hidden | 1902.10905 | null | https://arxiv.org/abs/1902.10905v3 | https://arxiv.org/pdf/1902.10905v3.pdf | PixelSteganalysis: Pixel-wise Hidden Information Removal with Low Visual Degradation | Recently, the field of steganography has experienced rapid developments based on deep learning (DL). DL based steganography distributes secret information over all the available bits of the cover image, thereby posing difficulties in using conventional steganalysis methods to detect, extract or remove hidden secret images. However, our proposed framework is the first to effectively disable covert communications and transactions that use DL based steganography. We propose a DL based steganalysis technique that effectively removes secret images by restoring the distribution of the original images. We formulate a problem and address it by exploiting sophisticated pixel distributions and an edge distribution of images by using a deep neural network. Based on the given information, we remove the hidden secret information at the pixel level. We evaluate our technique by comparing it with conventional steganalysis methods using three public benchmarks. As the decoding method of DL based steganography is approximate (lossy) and is different from the decoding method of conventional steganography, we also introduce a new quantitative metric called the destruction rate (DT). The experimental results demonstrate performance improvements of 10-20% in both the decoded rate and the DT. | ['Hyun-Soo Choi', 'Dahuin Jung', 'Sungroh Yoon', 'Ho Bae'] | 2019-02-28 | null | null | null | null | ['steganalysis'] | ['computer-vision'] | [ 1.07006311e+00 1.72187150e-01 1.63652584e-01 2.72885650e-01
-3.15668017e-01 -3.07841957e-01 4.59900290e-01 -3.17425728e-01
-3.36560130e-01 6.70429468e-01 -1.72184572e-01 -5.59430361e-01
4.08333927e-01 -1.14559460e+00 -8.53063941e-01 -1.33145308e+00
-2.17722610e-01 9.52630788e-02 2.37562731e-01 -7.12005615e-01
5.34506857e-01 1.63862020e-01 -1.25643969e+00 3.15851510e-01
5.48517466e-01 9.67919648e-01 4.11213785e-02 7.05154419e-01
1.49002925e-01 7.85500526e-01 -7.84715414e-01 -1.91026181e-01
5.86427450e-01 -1.04179597e+00 -4.01197761e-01 1.92989156e-01
-1.24574341e-01 -3.56702417e-01 -5.51167667e-01 1.38361561e+00
5.70146024e-01 -4.99349803e-01 4.95395511e-01 -1.11774349e+00
-4.86082405e-01 7.93258548e-01 -7.46509016e-01 1.04465209e-01
1.19873635e-01 2.86276281e-01 4.27349657e-01 -2.90437102e-01
6.06606781e-01 1.22675908e+00 7.39924133e-01 5.21292448e-01
-8.98370922e-01 -1.07904017e+00 -4.11631465e-01 4.03209329e-01
-1.27930069e+00 -4.20789152e-01 9.73645091e-01 4.33211625e-02
7.64452815e-01 4.42705542e-01 6.66240335e-01 8.66310060e-01
9.60187614e-01 6.08395278e-01 1.51463401e+00 -7.40871191e-01
-2.07533419e-01 -1.13727655e-02 -6.06163561e-01 7.84793794e-01
6.19549692e-01 5.31561434e-01 3.74321267e-02 2.18587339e-01
4.96933311e-01 -1.65728386e-02 -4.92732376e-01 -3.98300648e-01
-1.30943489e+00 1.03078175e+00 1.52415171e-01 4.73457605e-01
4.66099335e-03 4.98651475e-01 4.79576379e-01 1.09622920e+00
2.88829684e-01 2.03682005e-01 -6.80309385e-02 3.52811277e-01
-9.79911268e-01 -2.56484270e-01 1.18209374e+00 5.57180822e-01
7.48070478e-01 4.19494182e-01 7.17265680e-02 8.34937915e-02
3.23929459e-01 8.62294495e-01 5.10611892e-01 -5.01076221e-01
4.07691866e-01 1.49463475e-01 -3.37592036e-01 -1.84138680e+00
1.05727680e-01 -2.57686496e-01 -1.42875373e+00 5.67448676e-01
-1.02975231e-03 -1.01765819e-01 -1.06429040e+00 1.25879085e+00
-1.03795588e-01 2.59226620e-01 2.05995485e-01 3.71035606e-01
5.02676547e-01 9.00993884e-01 -5.01277506e-01 -3.98629934e-01
8.05878103e-01 -8.93446922e-01 -9.67216969e-01 -1.62617773e-01
6.69885635e-01 -7.49742568e-01 2.99216628e-01 4.77926105e-01
-9.27018583e-01 -2.41524100e-01 -1.56672931e+00 5.16253412e-01
-4.99481857e-01 -6.51281238e-01 5.16919717e-02 1.22570944e+00
-1.18345976e+00 4.90950853e-01 -4.39755082e-01 3.71401035e-03
2.67480195e-01 7.25964129e-01 -3.35953683e-01 -1.90391049e-01
-1.66458702e+00 8.42366517e-01 9.22614276e-01 -1.64506882e-02
-1.21041715e+00 1.58763617e-01 -1.11304700e+00 1.63211957e-01
3.12372148e-01 -1.64412796e-01 3.20495307e-01 -1.25137782e+00
-1.43180656e+00 9.91267204e-01 4.64858472e-01 -7.30021894e-01
6.76501274e-01 5.05103588e-01 -6.84269726e-01 1.87451318e-01
-4.44066942e-01 2.87267148e-01 1.57877314e+00 -1.46864378e+00
-6.52842700e-01 4.57344130e-02 -2.67353863e-01 -1.37633562e-01
-2.74071366e-01 -2.96508223e-01 -2.44527027e-01 -8.07999194e-01
1.60239711e-01 -1.01056194e+00 -9.41678733e-02 -3.29712152e-01
-4.65652913e-01 5.42731762e-01 1.46732640e+00 -1.00142860e+00
1.22577870e+00 -2.20456433e+00 1.57548234e-01 6.51382625e-01
5.21466315e-01 7.09293842e-01 -1.60768881e-01 5.91777146e-01
-1.15443327e-01 2.73388147e-01 -5.67612588e-01 -1.83595300e-01
-8.65319967e-02 1.78880155e-01 -6.49241805e-02 1.03788769e+00
-4.97346014e-01 9.63628113e-01 -7.37349331e-01 -4.80479628e-01
4.54227388e-01 6.56005740e-01 -2.10321650e-01 -2.05060109e-01
1.00343131e-01 5.47078073e-01 -1.15751669e-01 5.81119239e-01
1.17430544e+00 -1.82010353e-01 6.62483454e-01 5.08936606e-02
2.99981207e-01 -1.92232698e-01 -8.07822764e-01 1.05731261e+00
-3.27860892e-01 9.57753360e-01 9.56583917e-02 -1.38963354e+00
1.07843280e+00 3.57166141e-01 3.67139518e-01 -9.07752812e-01
6.34953558e-01 5.37059367e-01 5.41123040e-02 -4.88395721e-01
2.10545376e-01 -1.23944260e-01 -3.04318160e-01 7.87055671e-01
-2.19044030e-01 2.34925989e-02 -2.47204527e-01 -2.17282221e-01
1.16438329e+00 -3.39461803e-01 3.86955291e-01 1.61585864e-02
8.68411541e-01 -2.72381991e-01 3.01426858e-01 1.05616641e+00
-1.05897263e-01 2.55428553e-01 5.11680722e-01 -4.19599801e-01
-1.41319311e+00 -3.33766162e-01 3.82629722e-01 5.26883066e-01
6.97639942e-01 1.62299886e-01 -8.27277184e-01 -7.92460501e-01
-1.01565674e-01 4.09426123e-01 -4.71801162e-01 -6.16677940e-01
-8.50670695e-01 -8.26383352e-01 9.67892647e-01 -3.63957047e-01
1.21637213e+00 -1.34540677e+00 -5.12277126e-01 2.08472684e-01
-2.60693878e-01 -9.50177133e-01 -2.12343141e-01 -6.63146228e-02
-6.60347164e-01 -1.08030903e+00 -7.38093436e-01 -1.11191452e+00
7.50083327e-01 3.81886542e-01 7.46817827e-01 7.21178949e-01
-5.74498549e-02 -8.24646726e-02 -6.30024433e-01 -9.11899433e-02
-1.21201730e+00 1.42310858e-01 -3.84548664e-01 1.63408086e-01
2.25585073e-01 -5.13898611e-01 -5.65454066e-01 2.67249674e-01
-1.31002474e+00 1.50389716e-01 1.13348508e+00 8.50124598e-01
2.73938626e-01 9.68276322e-01 1.00509681e-01 -1.20762157e+00
4.21089172e-01 -5.78016937e-01 -5.73141634e-01 1.05449237e-01
-1.02706695e+00 1.34172037e-01 4.14256603e-01 -2.85123646e-01
-6.00190759e-01 -3.73078436e-01 -2.41258934e-01 -1.25078872e-01
2.17052758e-01 4.19197679e-01 -2.93791771e-01 -8.91269147e-01
5.56234643e-03 9.05827343e-01 4.43586677e-01 -1.43756926e-01
-1.86517075e-01 7.31815159e-01 3.43863189e-01 3.43232155e-01
1.29806232e+00 7.27045536e-01 1.94267750e-01 -5.24961770e-01
-1.24389812e-01 -8.35573673e-02 -2.34940544e-01 -1.96573529e-02
4.43816930e-01 -4.98792022e-01 -8.00200641e-01 1.15794730e+00
-9.06230450e-01 -1.41208664e-01 1.47932380e-01 1.61409602e-01
-5.19783139e-01 9.35346305e-01 -5.16088426e-01 -4.33032423e-01
-4.65863585e-01 -1.30412519e+00 6.99845850e-01 -5.41944981e-01
3.71142536e-01 -1.34352303e+00 -5.96538819e-02 1.60100743e-01
6.26278877e-01 1.09380448e+00 8.91792178e-01 -6.02078438e-01
-6.96543872e-01 -6.27004325e-01 -2.11286202e-01 6.79363072e-01
2.86308020e-01 -7.90084720e-01 -4.79755014e-01 -1.02990270e+00
6.23097003e-01 1.57309964e-01 1.41515183e+00 2.65487373e-01
1.12665212e+00 -8.09197724e-01 -4.81457442e-01 1.12158203e+00
1.88353205e+00 4.31851178e-01 1.48946846e+00 7.93555915e-01
8.90916884e-01 4.29734051e-01 1.42080754e-01 3.30120295e-01
4.77040224e-02 2.20477298e-01 7.57404864e-01 -2.81023532e-01
-2.22238898e-01 -9.11085308e-02 6.76030993e-01 9.23676610e-01
-2.40988284e-02 -1.05289745e+00 -6.19597316e-01 4.38037634e-01
-1.53731990e+00 -1.15505517e+00 -4.55208123e-02 2.07180548e+00
7.89396942e-01 3.68244082e-01 -4.81606364e-01 8.94543350e-01
1.13385379e+00 7.55761683e-01 -3.70702982e-01 -5.53417623e-01
-3.99700463e-01 1.37058482e-01 1.42416203e+00 6.31254911e-01
-1.19188225e+00 7.91926444e-01 6.27180386e+00 1.17938542e+00
-1.37913299e+00 2.55952805e-01 4.49067205e-01 4.64978248e-01
-2.94172883e-01 -6.85165301e-02 -5.46465695e-01 8.29158008e-01
7.93493211e-01 1.22580335e-01 5.70247471e-01 2.17891663e-01
-9.89659801e-02 1.12883635e-01 -5.36161304e-01 1.04717040e+00
7.53879488e-01 -1.34459078e+00 1.51034042e-01 6.08335555e-01
1.05288744e+00 -4.26945478e-01 4.19733256e-01 -1.34297358e-02
1.43744633e-01 -1.22047460e+00 4.03490812e-01 3.41464758e-01
1.19858229e+00 -9.89249110e-01 1.25797701e+00 2.36519679e-01
-7.76152432e-01 -1.15307868e-01 -1.95495307e-01 3.13521884e-02
-3.89621109e-02 4.88739163e-01 -9.01513755e-01 3.80337894e-01
4.20922905e-01 8.04740846e-01 -1.80944517e-01 6.79545999e-01
-3.37965757e-01 7.02889979e-01 -4.41548079e-02 1.28027275e-01
4.77518141e-01 1.81828931e-01 1.00551736e+00 1.15677202e+00
7.18648255e-01 -3.50288123e-01 -2.07073670e-02 4.14298058e-01
-2.48838738e-01 -2.88652152e-01 -9.08705771e-01 -1.55949190e-01
2.78253257e-01 5.63832402e-01 -9.66482401e-01 -2.52944171e-01
-1.79760959e-02 1.22404242e+00 -7.68566668e-01 1.64802596e-01
-6.68232322e-01 -8.08120847e-01 -5.01761846e-02 2.42468297e-01
9.85904396e-01 -9.10784006e-02 -1.61032468e-01 -1.04933774e+00
-3.20629716e-01 -1.36950529e+00 -6.42478140e-03 -2.02561282e-02
-6.22039974e-01 3.78334463e-01 -2.60708451e-01 -1.42105150e+00
-1.60589904e-01 -3.11942935e-01 -2.37188935e-01 5.28229892e-01
-2.03771496e+00 -1.34358048e+00 -2.63723999e-01 6.34859681e-01
3.10561746e-01 -6.83145523e-01 6.23455703e-01 2.52846450e-01
-1.27842471e-01 7.04828501e-01 6.97859168e-01 2.98315912e-01
3.99905473e-01 -7.26591349e-01 5.90969741e-01 9.63605642e-01
-5.62306583e-01 -7.65259489e-02 9.97494936e-01 -7.77705133e-01
-1.41476440e+00 -9.52781439e-01 8.22358489e-01 1.12874091e-01
2.19310656e-01 -2.96696514e-01 -6.50777936e-01 8.68906140e-01
5.01871824e-01 -2.76617587e-01 3.01881164e-01 -1.04887843e+00
-1.41590238e-01 3.47924270e-02 -1.68437767e+00 1.09513909e-01
7.10661054e-01 -1.32007971e-01 -2.24484742e-01 5.95373549e-02
6.35330081e-01 -2.98694074e-01 -2.59977907e-01 3.77861142e-01
5.66609323e-01 -1.16201544e+00 9.61328864e-01 3.02823007e-01
3.91113997e-01 -2.16816962e-01 -2.19918609e-01 -1.22026074e+00
-2.35310063e-01 -9.17413056e-01 -3.50425124e-01 6.17153525e-01
-7.44281039e-02 -9.22473371e-01 7.19517231e-01 -6.77304149e-01
2.94995815e-01 -2.72778183e-01 -7.61169434e-01 -8.07869256e-01
-3.13797295e-02 9.27814841e-03 9.13079679e-01 1.14717507e+00
-6.17394447e-01 -4.36621457e-01 -1.16859627e+00 1.58531293e-01
1.24666083e+00 -1.28729343e-01 5.61883390e-01 -1.08102393e+00
-1.89780429e-01 -3.08491796e-01 -7.93251514e-01 -7.86538124e-01
1.78413674e-01 -7.50897646e-01 1.23613648e-01 -1.25231481e+00
-5.05274013e-02 -3.40051711e-01 -4.88749117e-01 3.62515956e-01
4.57660526e-01 8.49422872e-01 -7.44061098e-02 4.49995905e-01
-3.70148271e-01 2.32924625e-01 1.57987309e+00 -6.51518345e-01
1.21333554e-01 -2.07127169e-01 -5.86549103e-01 6.01943016e-01
9.82814491e-01 -9.78520453e-01 -1.61753237e-01 -3.08882415e-01
3.94686311e-01 -4.68906052e-02 4.42585647e-01 -1.22875023e+00
1.49565727e-01 -6.97994903e-02 1.51672721e-01 -3.82876992e-01
5.98923936e-02 -1.11288881e+00 3.23024064e-01 1.32711339e+00
-1.41088083e-01 -4.81352240e-01 -1.38956875e-01 6.70267761e-01
-2.47402579e-01 -1.61178648e-01 9.55124617e-01 -3.85665357e-01
-1.03144181e+00 -5.90732209e-02 -5.21686375e-01 -3.90226930e-01
1.24744403e+00 -7.99678028e-01 -1.00642540e-01 -7.17729568e-01
-5.38393438e-01 -2.84676790e-01 7.33786285e-01 6.87803030e-02
1.07946765e+00 -1.08294559e+00 -8.70528460e-01 7.59381354e-01
-1.92199185e-01 -6.21646583e-01 2.11385190e-02 7.96674013e-01
-1.21139979e+00 4.69203323e-01 -5.76557755e-01 -1.97214752e-01
-1.34716570e+00 6.77919865e-01 4.71071273e-01 -6.36336684e-01
-7.34845400e-01 6.03740096e-01 2.25097556e-02 -1.10474907e-01
6.82442337e-02 1.27338260e-01 -2.36250401e-01 -2.36257225e-01
4.95799720e-01 3.63578439e-01 -1.74358830e-01 -8.39382708e-01
-1.56283304e-02 5.32820165e-01 -6.77800700e-02 8.27744156e-02
1.31574786e+00 -7.22126782e-01 -5.84260464e-01 -1.04642525e-01
1.47422898e+00 -2.22570449e-02 -8.61646175e-01 -1.44765347e-01
-4.46030647e-01 -6.82169855e-01 4.33384210e-01 -5.63214123e-01
-1.57929444e+00 7.25537956e-01 8.69903088e-01 6.43747449e-01
1.29021609e+00 -5.92078686e-01 1.69069600e+00 2.89487034e-01
5.04667222e-01 -8.91536474e-01 -2.24739350e-02 4.78406519e-01
4.70926106e-01 -1.31912804e+00 1.45837236e-02 -2.22487360e-01
-2.21694201e-01 1.00247800e+00 -1.14355840e-01 -3.42379630e-01
7.87568986e-01 3.26847106e-01 -1.45755082e-01 -2.99328536e-01
-1.93394125e-01 1.68244198e-01 -1.93797156e-01 7.78543949e-01
-2.36354664e-01 -6.89070299e-02 -5.12606800e-01 -3.79962772e-01
-9.11437720e-03 -6.74503669e-02 8.14140022e-01 1.22042656e+00
-7.40985990e-01 -1.26652491e+00 -8.18579912e-01 1.29592270e-01
-8.58568788e-01 -2.34698236e-01 -1.03856370e-01 7.70551920e-01
3.73861790e-01 8.93628478e-01 -1.81109071e-01 -8.64484072e-01
-4.02141541e-01 -4.57458854e-01 3.05705339e-01 -3.31648350e-01
-1.36366442e-01 2.94261780e-02 -2.41622433e-01 -1.78311288e-01
-7.23513782e-01 -7.73156881e-02 -8.51863086e-01 -1.15487826e+00
-3.76135111e-01 -8.16847570e-03 6.89130545e-01 9.86312389e-01
-5.17027974e-02 6.54394507e-01 1.15229356e+00 -7.25889146e-01
-3.30112159e-01 -6.51961088e-01 -9.04945433e-01 2.57201463e-01
8.96832347e-01 -3.14099044e-01 -9.51868296e-01 2.47420877e-01] | [4.31631326675415, 8.057793617248535] |
0b899e9d-7d09-4841-857e-b3572e8884aa | usip-unsupervised-stable-interest-point | 1904.00229 | null | http://arxiv.org/abs/1904.00229v1 | http://arxiv.org/pdf/1904.00229v1.pdf | USIP: Unsupervised Stable Interest Point Detection from 3D Point Clouds | In this paper, we propose the USIP detector: an Unsupervised Stable Interest
Point detector that can detect highly repeatable and accurately localized
keypoints from 3D point clouds under arbitrary transformations without the need
for any ground truth training data. Our USIP detector consists of a feature
proposal network that learns stable keypoints from input 3D point clouds and
their respective transformed pairs from randomly generated transformations. We
provide degeneracy analysis of our USIP detector and suggest solutions to
prevent it. We encourage high repeatability and accurate localization of the
keypoints with a probabilistic chamfer loss that minimizes the distances
between the detected keypoints from the training point cloud pairs. Extensive
experimental results of repeatability tests on several simulated and real-world
3D point cloud datasets from Lidar, RGB-D and CAD models show that our USIP
detector significantly outperforms existing hand-crafted and deep
learning-based 3D keypoint detectors. Our code is available at the project
website. https://github.com/lijx10/USIP | ['Gim Hee Lee', 'Jiaxin Li'] | 2019-03-30 | usip-unsupervised-stable-interest-point-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Li_USIP_Unsupervised_Stable_Interest_Point_Detection_From_3D_Point_Clouds_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Li_USIP_Unsupervised_Stable_Interest_Point_Detection_From_3D_Point_Clouds_ICCV_2019_paper.pdf | iccv-2019-10 | ['interest-point-detection'] | ['computer-vision'] | [-4.87953484e-01 -7.47875273e-02 -5.54160140e-02 -3.39528859e-01
-8.55186582e-01 -6.73156857e-01 7.58503854e-01 4.21805345e-02
-2.20830098e-01 1.54939547e-01 -4.97017026e-01 -1.82212964e-01
-2.15171337e-01 -5.95720768e-01 -1.18147910e+00 -1.88985854e-01
-2.58395612e-01 9.87207890e-01 4.95531946e-01 -2.90427450e-03
4.55657005e-01 1.36601281e+00 -1.49924195e+00 -7.17266142e-01
5.38649499e-01 1.01915240e+00 2.14858308e-01 5.47831893e-01
9.79780033e-02 -8.16711038e-02 -1.50421828e-01 1.63373157e-01
9.71942246e-01 3.07682544e-01 -2.69141555e-01 -3.24752361e-01
1.00058961e+00 -3.86592925e-01 -3.34587365e-01 9.19359684e-01
4.43179697e-01 -1.18272670e-01 6.49668992e-01 -1.87438381e+00
-6.01801693e-01 1.17125185e-02 -8.17953527e-01 -2.26667281e-02
3.59711289e-01 3.79560962e-02 8.53356779e-01 -1.28676558e+00
7.47294307e-01 1.14464879e+00 1.13104641e+00 2.07649410e-01
-1.08568227e+00 -9.20512021e-01 -3.65324587e-01 -1.06575124e-01
-1.92147493e+00 -2.64714420e-01 8.77290726e-01 -5.33852756e-01
1.04226863e+00 9.21425074e-02 6.48600280e-01 5.34961879e-01
3.08500201e-01 5.92121005e-01 3.60920191e-01 -4.30945605e-01
1.55858666e-01 -7.66844451e-02 1.07375253e-02 8.56386364e-01
3.80745977e-01 5.23444116e-01 -3.62120450e-01 -4.48910892e-01
1.28957236e+00 4.18607891e-01 -4.98007499e-02 -1.47923422e+00
-1.52516723e+00 6.68877006e-01 9.86236215e-01 3.36264670e-02
-3.59153122e-01 4.04266357e-01 -2.07275942e-01 1.21071242e-01
3.26379299e-01 3.11045170e-01 -6.96841598e-01 1.25609532e-01
-8.59808683e-01 3.70225221e-01 4.14745867e-01 1.64636230e+00
1.20674121e+00 -2.79717177e-01 3.60530257e-01 2.17261672e-01
6.08234107e-01 1.10496032e+00 1.02578461e-01 -1.05515218e+00
9.79239047e-02 8.88608992e-01 4.69217688e-01 -1.12604129e+00
-3.82968247e-01 4.77802642e-02 -5.20209193e-01 4.98636574e-01
-3.69746014e-02 2.34233320e-01 -1.02321517e+00 1.00862563e+00
6.43274426e-01 5.72736025e-01 -4.19981331e-01 9.19530034e-01
7.24094391e-01 3.02507937e-01 -5.80946326e-01 4.12171930e-01
6.60756588e-01 -4.06129986e-01 4.86965999e-02 -7.45887086e-02
4.40571427e-01 -7.45386839e-01 8.20040762e-01 -1.17566556e-01
-8.57952833e-01 -5.71646810e-01 -1.13653469e+00 -2.05601782e-01
-2.44628057e-01 2.38544166e-01 3.69839787e-01 1.14270687e-01
-1.22329438e+00 6.23791397e-01 -1.18713737e+00 -5.01767874e-01
3.37616771e-01 6.20259225e-01 -5.09643137e-01 2.70936906e-01
-5.34782112e-01 1.00885010e+00 2.15501606e-01 9.17433873e-02
-6.65304363e-01 -9.38609004e-01 -9.05922771e-01 -3.11860234e-01
9.17361453e-02 -7.13880599e-01 1.31792665e+00 -2.19087765e-01
-1.11464560e+00 1.18936658e+00 -1.95157051e-01 -4.86340910e-01
6.55676067e-01 -6.45461440e-01 -6.04923209e-03 -7.44906813e-02
4.72741991e-01 9.26600039e-01 8.94093513e-01 -1.43085361e+00
-7.12447762e-01 -4.70953166e-01 -4.67317820e-01 -1.32761830e-02
4.57307339e-01 -1.85443774e-01 -2.85453796e-01 -7.44391605e-02
8.64768505e-01 -1.32476270e+00 -1.84991077e-01 5.35123110e-01
-6.48204684e-01 -5.16200840e-01 1.13173461e+00 1.02067709e-01
1.71230659e-01 -2.11129737e+00 -2.42147148e-01 3.99164617e-01
3.50092560e-01 2.07608547e-02 -1.16973877e-01 3.06080759e-01
-2.49924034e-01 -4.94683087e-02 -5.19056283e-02 -2.23509103e-01
1.95573002e-01 2.57569309e-02 -5.21372974e-01 9.06660676e-01
1.94603994e-01 9.68913376e-01 -9.65729654e-01 -2.50587344e-01
8.94032657e-01 6.47328258e-01 -1.37291446e-01 1.54786155e-01
-2.58596167e-02 1.29832223e-01 -4.88026857e-01 8.42144668e-01
1.11141336e+00 -1.74745172e-02 -6.79041743e-01 -1.65645435e-01
-2.64478177e-01 1.08583182e-01 -1.31157577e+00 1.79211116e+00
-1.14650741e-01 6.13584936e-01 -5.28606176e-02 -4.76593614e-01
1.43334317e+00 -7.30417296e-02 7.84935653e-01 -2.52150714e-01
1.13490753e-01 4.15573806e-01 -5.13013363e-01 2.09856942e-01
6.64270639e-01 5.24042472e-02 -4.89407778e-02 3.60599637e-01
1.45138472e-01 -9.35237646e-01 -4.80144173e-01 1.48447052e-01
9.32936013e-01 2.31783047e-01 3.24951470e-01 -2.17423886e-01
3.00246149e-01 2.32547000e-01 4.18857455e-01 8.07507217e-01
-3.38016689e-01 9.77518916e-01 -1.68666929e-01 -7.67078936e-01
-1.19852579e+00 -1.56683326e+00 -4.20098752e-01 2.69077033e-01
5.53583145e-01 -2.84480602e-01 -7.67757148e-02 -5.56349516e-01
8.24022532e-01 4.49190587e-01 -3.27301502e-01 -6.67044893e-02
-5.17194629e-01 2.91978627e-01 2.37450957e-01 5.47295392e-01
1.93383500e-01 -6.90960824e-01 -7.62312710e-01 -2.45277718e-01
3.96829665e-01 -8.70710433e-01 -2.35743985e-01 2.49085560e-01
-9.90363598e-01 -1.25529253e+00 -4.67086941e-01 -7.88924813e-01
9.06007767e-01 7.05239236e-01 9.83818769e-01 2.73195282e-02
-2.33579844e-01 7.59524167e-01 -1.64796069e-01 -7.93898880e-01
-1.84447095e-01 2.31306683e-02 5.53649068e-01 -6.39269471e-01
9.99274671e-01 -6.10530376e-01 -4.74612206e-01 5.61781466e-01
-3.17997515e-01 -3.72029990e-01 2.85843581e-01 3.88800144e-01
9.88072813e-01 -3.12874317e-01 -2.45223373e-01 5.45476750e-02
1.77981392e-01 -1.54630214e-01 -1.27642500e+00 -6.77007511e-02
-4.74730790e-01 2.70195231e-02 -2.18876842e-02 -1.70950666e-01
-6.92381039e-02 6.98210180e-01 2.83210635e-01 -1.11686993e+00
-3.41099024e-01 -1.86590478e-01 1.27540141e-01 -6.51808023e-01
9.33817089e-01 1.43369943e-01 1.00300387e-01 -4.30801839e-01
6.85421586e-01 6.06355250e-01 6.51608944e-01 -5.36674440e-01
1.65373051e+00 7.41050899e-01 1.70326605e-01 -7.32637405e-01
-2.83518314e-01 -7.36261487e-01 -1.53280759e+00 -4.63997535e-02
3.32624197e-01 -1.11049604e+00 -6.42143071e-01 2.69497573e-01
-1.37928939e+00 4.79922742e-02 -5.71183443e-01 4.25518304e-01
-8.30250621e-01 2.67916143e-01 -1.36173576e-01 -4.83002007e-01
-4.87021625e-01 -1.07586682e+00 1.67671597e+00 2.90605694e-01
-3.21508765e-01 -5.64758360e-01 3.91712219e-01 -4.06599194e-01
-5.94118536e-02 4.69711900e-01 2.72978365e-01 -4.09685731e-01
-1.13012898e+00 -8.32541764e-01 -1.55702680e-01 1.29832223e-01
2.88207501e-01 5.29551208e-01 -6.96059167e-01 -4.09030706e-01
-7.07712695e-02 -5.92491403e-02 5.44356287e-01 6.38405204e-01
5.29653311e-01 2.51586258e-01 -8.39446008e-01 8.54254961e-01
1.44537234e+00 8.36225040e-03 2.88123488e-01 5.97531497e-01
7.42224753e-01 6.53007329e-02 7.33738840e-01 2.83103257e-01
3.90105546e-01 4.25722867e-01 6.87509775e-01 3.59084047e-02
2.84465283e-01 -7.49442637e-01 -4.27930616e-02 3.40170473e-01
1.49826263e-03 3.42895806e-01 -1.37544966e+00 7.39240766e-01
-1.81666660e+00 -4.89279121e-01 -2.94446975e-01 2.30408740e+00
2.21741125e-01 3.13160792e-02 -8.92759413e-02 -1.48460567e-01
8.89567673e-01 -1.75259635e-01 -7.97318459e-01 6.66938275e-02
2.38143861e-01 2.91879416e-01 8.66893530e-01 5.05564809e-01
-1.28260159e+00 1.07676291e+00 5.90087891e+00 1.01301469e-01
-1.13674819e+00 -1.45984933e-01 -1.81789860e-01 -1.39252514e-01
-1.29678342e-02 1.85553983e-01 -9.22803998e-01 9.96816903e-02
4.62256283e-01 -4.80401456e-01 -8.65639597e-02 1.31854880e+00
-4.27460410e-02 5.86190186e-02 -1.29768908e+00 1.22892189e+00
-2.78240561e-01 -1.39605689e+00 -2.44492128e-01 1.08159855e-01
5.68241000e-01 8.75907838e-01 3.94870453e-02 -1.47811651e-01
6.63810253e-01 -5.55397987e-01 8.96679580e-01 3.90436143e-01
6.82354152e-01 -6.94231927e-01 3.89189452e-01 4.58888352e-01
-1.22005999e+00 3.10656846e-01 -8.28086495e-01 1.03718661e-01
1.98752526e-03 5.22923291e-01 -1.32025325e+00 4.46335346e-01
8.19065094e-01 9.47317719e-01 -5.65422237e-01 1.36656737e+00
-4.76122677e-01 1.72434431e-02 -1.01263607e+00 -9.78583843e-03
1.17316976e-01 -1.52471453e-01 9.44420636e-01 6.63602650e-01
7.38804936e-01 -9.83900875e-02 9.67749432e-02 1.17762685e+00
4.04581800e-02 -2.66794652e-01 -1.12776387e+00 4.13510829e-01
1.03728366e+00 1.18962336e+00 -7.87900627e-01 3.00489087e-02
3.35941315e-02 7.11813569e-01 2.03311011e-01 -3.69453840e-02
-5.84576011e-01 -5.16207576e-01 9.37754333e-01 4.31109011e-01
4.92210865e-01 -8.24962497e-01 -1.80288851e-01 -1.08959198e+00
2.26790458e-01 -3.56657878e-02 -5.74320629e-02 -1.21181452e+00
-1.12798905e+00 2.55862921e-01 1.39458373e-01 -1.77185047e+00
-3.03882331e-01 -5.87320387e-01 -6.07168496e-01 8.14263165e-01
-1.35500860e+00 -1.29184711e+00 -4.44247395e-01 5.50624132e-01
3.80765982e-02 -3.92747000e-02 4.66121078e-01 -2.10827634e-01
1.10522874e-01 3.61671895e-01 2.08940849e-01 1.81240216e-01
5.50910056e-01 -1.18999755e+00 1.31879175e+00 7.37469077e-01
4.85994786e-01 7.81303585e-01 5.60695410e-01 -7.31847048e-01
-1.57827580e+00 -1.07690108e+00 6.17906749e-01 -1.12586284e+00
6.68710530e-01 -4.81377125e-01 -7.44093955e-01 1.14730167e+00
-4.32059944e-01 5.73104739e-01 -2.04723347e-02 -1.68048099e-01
-2.77233571e-01 -1.69412702e-01 -1.30711067e+00 3.24864566e-01
1.21161938e+00 -4.71608579e-01 -8.27307463e-01 5.32290101e-01
8.66872013e-01 -7.68196166e-01 -6.48378968e-01 5.57694197e-01
3.85560542e-01 -8.49074364e-01 1.44865894e+00 -1.62904963e-01
-2.04026341e-01 -6.94431484e-01 -1.09848060e-01 -1.23110402e+00
-5.74682474e-01 -4.71044451e-01 -7.34799579e-02 7.91948795e-01
1.09517416e-02 -6.74512208e-01 1.02864897e+00 6.15237474e-01
-7.20767677e-02 -1.57461971e-01 -1.30517042e+00 -1.04188704e+00
1.18324603e-03 -5.22313952e-01 7.99865723e-01 8.65977108e-01
-3.14807683e-01 1.74440760e-02 7.32302591e-02 7.22742438e-01
1.01166797e+00 3.34149122e-01 1.31887186e+00 -1.58172524e+00
6.40102804e-01 -2.48631388e-01 -1.14881325e+00 -1.03460670e+00
2.59532064e-01 -9.52440202e-01 2.13447496e-01 -1.36293292e+00
-4.20789629e-01 -7.89237916e-01 2.66267415e-02 5.44419169e-01
3.36595654e-01 -2.20411345e-02 2.12874413e-01 5.38292587e-01
-3.31423253e-01 6.42165899e-01 7.32131124e-01 -4.49979119e-02
-4.08576459e-01 2.08952710e-01 -2.40752101e-01 7.88469195e-01
7.27214277e-01 -7.03568399e-01 -3.66444401e-02 -5.85289478e-01
-6.24830462e-02 -2.82723784e-01 8.63208711e-01 -1.23665690e+00
5.63444793e-01 -1.56052500e-01 6.25953853e-01 -1.38209081e+00
3.49550337e-01 -1.12243617e+00 2.15509400e-01 4.70497936e-01
2.02498391e-01 2.62562454e-01 2.36999750e-01 4.31529313e-01
2.02192962e-01 -1.96958985e-02 7.27629125e-01 -1.54481128e-01
-8.90420973e-01 6.13484502e-01 3.82884085e-01 -5.08053958e-01
1.29779160e+00 -3.99578154e-01 -1.10698126e-01 -6.92769364e-02
-3.34921956e-01 2.89117992e-01 1.06112540e+00 7.88014293e-01
1.13262010e+00 -1.60934055e+00 -5.64049959e-01 6.45310581e-01
2.56911010e-01 7.16053665e-01 -3.33503991e-01 3.71046424e-01
-8.01922500e-01 5.82755327e-01 -1.56673357e-01 -1.46413922e+00
-1.11370897e+00 5.63697457e-01 5.72482347e-01 7.02681661e-01
-8.69146526e-01 9.01820242e-01 -2.84314245e-01 -1.21879232e+00
6.26836494e-02 -9.14578378e-01 6.35063887e-01 -4.47891235e-01
2.30280831e-01 3.35566431e-01 1.71852991e-01 -8.99687409e-01
-7.92941689e-01 1.11862195e+00 -9.73669230e-04 1.53280824e-01
1.46301126e+00 8.52802470e-02 -7.83652440e-02 3.16723675e-01
1.25632143e+00 -2.69262314e-01 -1.34413362e+00 -3.56705964e-01
2.26417243e-01 -8.06602895e-01 -6.38091471e-03 -1.93424851e-01
-7.11848438e-01 6.98799253e-01 9.97716367e-01 -7.67804608e-02
4.39556181e-01 4.07560289e-01 5.91913342e-01 6.84888899e-01
8.76449466e-01 -6.08101666e-01 -3.47251117e-01 5.91291845e-01
9.72232044e-01 -1.30301702e+00 2.27494434e-01 -1.48972198e-01
-7.87025467e-02 1.16419244e+00 5.55401146e-01 -7.78476298e-01
9.17148590e-01 2.91331828e-01 2.21310377e-01 -4.07553077e-01
-1.48882717e-01 -1.29239848e-02 2.51104355e-01 1.02690363e+00
-2.07666770e-01 7.42383748e-02 4.08254504e-01 -9.73048210e-02
-5.93744040e-01 1.45272523e-01 3.14028531e-01 1.20071316e+00
-7.58012712e-01 -7.66011119e-01 -6.90720499e-01 1.56766981e-01
4.42894012e-01 2.21996278e-01 -5.72784185e-01 1.10146546e+00
-1.82996958e-01 3.37599814e-01 5.03697634e-01 -5.37451804e-01
6.95547342e-01 -8.67451280e-02 5.49238920e-01 -5.42268813e-01
3.37516107e-02 -1.95309028e-01 -7.61922657e-01 -8.51296127e-01
-2.47921154e-01 -7.47462094e-01 -1.39108074e+00 -3.50372761e-01
-5.34193158e-01 2.58776769e-02 9.41422522e-01 4.71844316e-01
8.51791441e-01 -3.89079928e-01 7.82006025e-01 -1.54924476e+00
-5.59594274e-01 -6.50259554e-01 -5.60700297e-01 4.78880033e-02
7.57258177e-01 -8.51460338e-01 -5.23830116e-01 -2.71054059e-01] | [7.595898151397705, -2.5901124477386475] |
ba9a7ec3-b811-45bb-960e-060f5b941930 | a-multidimensional-graph-fourier | 2305.07416 | null | https://arxiv.org/abs/2305.07416v1 | https://arxiv.org/pdf/2305.07416v1.pdf | A Multidimensional Graph Fourier Transformation Neural Network for Vehicle Trajectory Prediction | This work introduces the multidimensional Graph Fourier Transformation Neural Network (GFTNN) for long-term trajectory predictions on highways. Similar to Graph Neural Networks (GNNs), the GFTNN is a novel network architecture that operates on graph structures. While several GNNs lack discriminative power due to suboptimal aggregation schemes, the proposed model aggregates scenario properties through a powerful operation: the multidimensional Graph Fourier Transformation (GFT). The spatio-temporal vehicle interaction graph of a scenario is converted into a spectral scenario representation using the GFT. This beneficial representation is input to the prediction framework composed of a neural network and a descriptive decoder. Even though the proposed GFTNN does not include any recurrent element, it outperforms state-of-the-art models in the task of highway trajectory prediction. For experiments and evaluation, the publicly available datasets highD and NGSIM are used | ['Wolfgang Utschick', 'Michael Botsch', 'Andreas Tollkühn', 'Marion Neumeier'] | 2023-05-12 | null | null | null | null | ['trajectory-prediction'] | ['computer-vision'] | [ 1.74051464e-01 3.16380382e-01 -4.25173402e-01 -3.73249799e-01
-1.32571980e-01 -1.21080369e-01 9.49405193e-01 8.15969110e-02
1.91059764e-02 6.73405766e-01 5.06093621e-01 -1.04006219e+00
-2.35008031e-01 -1.26140702e+00 -9.86694694e-01 -6.55828834e-01
-4.89765763e-01 4.65521783e-01 2.75960684e-01 -3.59848499e-01
-1.69842377e-01 6.52491689e-01 -1.57937670e+00 1.22248806e-01
8.92033696e-01 8.32507491e-01 2.67902687e-02 5.64589500e-01
-8.64502937e-02 9.10874248e-01 6.80900142e-02 -4.69964415e-01
2.38505006e-01 -1.19348571e-01 -3.45772803e-01 -3.89668018e-01
6.13567472e-01 -4.10557985e-01 -1.32132959e+00 7.75225639e-01
4.19006079e-01 6.21582389e-01 1.00662088e+00 -1.63653421e+00
-4.88584906e-01 8.68145287e-01 -2.28894636e-01 2.73849696e-01
8.65398198e-02 8.81171972e-02 8.87373686e-01 -7.31944382e-01
8.68501306e-01 1.28914666e+00 1.09002256e+00 4.71202284e-02
-1.12803018e+00 -6.27578378e-01 2.82168686e-01 5.39116740e-01
-1.25437570e+00 -5.97313866e-02 1.03664804e+00 -6.09733820e-01
1.46748710e+00 -1.18947089e-01 9.15011108e-01 1.35081387e+00
4.45664257e-01 7.46479332e-01 3.73653382e-01 1.48827761e-01
-8.95452574e-02 -3.58906269e-01 4.53216136e-01 7.86984265e-01
8.37172717e-02 6.23400390e-01 -2.91871578e-01 -9.64372680e-02
5.03463507e-01 7.37832487e-02 8.72587319e-03 -3.40424567e-01
-1.11501622e+00 7.94658422e-01 8.09942722e-01 -1.47031099e-01
-6.86242223e-01 4.33524489e-01 5.95598876e-01 2.70291656e-01
8.39668393e-01 -3.27480108e-01 1.63985863e-01 -1.36339292e-01
-1.10406041e+00 4.99773741e-01 7.06116855e-01 1.05198407e+00
8.16411257e-01 4.21263188e-01 -4.00028884e-01 3.85651827e-01
1.36381254e-01 5.94730079e-01 -7.41753355e-02 -5.00903845e-01
8.57431352e-01 7.89515078e-01 -3.81353945e-01 -1.41015995e+00
-8.73545647e-01 -5.51368833e-01 -1.23735917e+00 -1.17149547e-01
1.02407597e-01 -2.45723978e-01 -1.01170611e+00 1.75298274e+00
1.65092483e-01 9.32179928e-01 -2.55164457e-04 7.20622182e-01
1.29886079e+00 9.10487950e-01 2.96470761e-01 1.13436535e-01
5.83522618e-01 -9.60300863e-01 -7.55017996e-01 1.55848518e-01
9.28787947e-01 -1.80369526e-01 6.43296659e-01 -1.40243813e-01
-7.73722827e-01 -6.42662108e-01 -9.15734172e-01 9.46244411e-03
-8.12718272e-01 1.16171809e-02 6.22730672e-01 4.34306681e-01
-1.32960141e+00 5.91499448e-01 -5.44839799e-01 -7.08022118e-01
2.33536541e-01 7.68766478e-02 -2.77225196e-01 -5.41825369e-02
-1.43285871e+00 1.06061125e+00 6.77440464e-01 2.00274795e-01
-8.87849391e-01 -8.55827987e-01 -1.11837661e+00 2.96346936e-02
1.67621598e-01 -8.03670228e-01 5.98008335e-01 -2.74962574e-01
-1.17163789e+00 2.02007666e-01 -1.85788646e-01 -1.15230584e+00
5.96923649e-01 -4.22118697e-03 -9.74187970e-01 5.68787707e-03
1.31845148e-02 6.47146642e-01 7.49446213e-01 -9.73051965e-01
-5.94115376e-01 1.54144809e-01 -1.01734214e-02 1.74778089e-01
-1.62725955e-01 -3.42689991e-01 -2.22710177e-01 -6.55332029e-01
-7.96376988e-02 -9.66784000e-01 -2.39773706e-01 -5.06263971e-01
-5.86530626e-01 -3.91170621e-01 1.32325447e+00 -8.97983253e-01
1.69063663e+00 -1.90709937e+00 2.04072341e-01 5.80071628e-01
4.94888574e-01 3.30867618e-01 -3.14685225e-01 8.48521352e-01
-3.91894549e-01 -1.13811366e-01 -1.29492968e-01 -2.67671824e-01
1.97157696e-01 2.28344977e-01 -4.83866215e-01 4.77579653e-01
1.62744317e-02 1.30748582e+00 -9.23257113e-01 -4.15787905e-01
6.71301126e-01 5.48782110e-01 -4.48793560e-01 8.84594321e-02
-3.64047289e-01 1.06149361e-01 -2.96682268e-01 2.15716586e-01
7.99065709e-01 -2.14471072e-01 1.21895157e-01 -3.51573825e-01
-1.55463651e-01 2.27027759e-01 -8.07617903e-01 1.26906359e+00
-2.30470344e-01 7.78018236e-01 -7.88504362e-01 -1.23148048e+00
1.02469921e+00 1.48678392e-01 6.97724223e-01 -7.91201949e-01
-7.29012489e-02 -1.03720054e-01 -2.32569978e-01 -5.46470225e-01
7.81534851e-01 3.07695538e-01 -1.06234662e-01 4.27946866e-01
1.85999662e-01 4.91214007e-01 4.37005609e-01 4.31334943e-01
1.26728475e+00 1.85715497e-01 2.27113497e-02 -9.86717343e-02
5.17873824e-01 8.79027098e-02 2.04205275e-01 6.20348573e-01
1.15355819e-01 3.53151828e-01 5.19480407e-01 -8.44633996e-01
-1.29198563e+00 -1.18980503e+00 1.83864683e-01 9.23648834e-01
4.07034345e-02 -6.56293869e-01 -4.77003813e-01 -8.11667919e-01
1.69602200e-01 1.18275821e+00 -6.86764836e-01 -3.40837777e-01
-9.04386103e-01 -6.47051215e-01 9.79376078e-01 6.05315387e-01
7.57882833e-01 -7.21462607e-01 4.96664420e-02 4.02335048e-01
-2.64146566e-01 -1.30804169e+00 -3.69566202e-01 -1.72086850e-01
-6.00928187e-01 -9.55361903e-01 -2.88873345e-01 -6.26603246e-01
2.85207570e-01 4.58835810e-01 1.08386505e+00 -1.49532497e-01
3.93974096e-01 9.07465592e-02 -2.52118349e-01 -1.98868468e-01
-5.92986643e-01 3.92124802e-01 -2.87892725e-02 9.30970907e-02
4.28003430e-01 -1.02518260e+00 -4.65405315e-01 2.92684317e-01
-4.16471541e-01 4.32197541e-01 5.34101844e-01 5.01403511e-01
4.62509662e-01 2.08504498e-01 6.68819487e-01 -7.06394732e-01
7.00936675e-01 -9.49217618e-01 -5.10372043e-01 5.43794483e-02
-5.77417016e-01 6.25046417e-02 8.18172812e-01 -1.76058561e-01
-9.98977721e-01 -6.21495582e-02 -2.53280342e-01 -7.76721001e-01
-2.59636134e-01 1.05900991e+00 1.71473309e-01 -1.18295893e-01
5.99519491e-01 3.04900438e-01 7.82065094e-03 -1.89870223e-01
7.15400517e-01 2.85334706e-01 4.68174130e-01 -1.02941617e-01
1.04269564e+00 2.07383499e-01 4.85399067e-01 -1.04135060e+00
-4.36800867e-01 -3.75454158e-01 -7.26637781e-01 -9.45820510e-01
9.74812210e-01 -9.45602655e-01 -6.86328590e-01 2.73582637e-01
-1.03938806e+00 -6.24749422e-01 -9.26376954e-02 5.05762815e-01
-7.03051567e-01 4.11640406e-01 -4.26411420e-01 -7.84242272e-01
-4.45385203e-02 -7.02019155e-01 9.30466115e-01 -1.18412085e-01
6.87915683e-02 -1.24577165e+00 1.09335385e-01 2.86663510e-02
4.60873634e-01 8.39310646e-01 1.22834790e+00 -5.24114907e-01
-6.57667577e-01 -2.40101382e-01 -4.76023585e-01 -5.37398122e-02
-1.65385947e-01 4.80947234e-02 -5.02382636e-01 -6.17971681e-02
-6.03023529e-01 1.35691717e-01 1.30073369e+00 7.46279895e-01
9.77610588e-01 -5.04991472e-01 -6.72183990e-01 8.31706762e-01
1.32698929e+00 -5.37794307e-02 7.84367204e-01 2.71535009e-01
1.14414215e+00 5.09663463e-01 3.79613936e-01 8.51732492e-02
8.60168815e-01 5.66756487e-01 6.39206350e-01 -2.10429113e-02
-5.63537598e-01 -7.59446144e-01 4.75190163e-01 7.64149547e-01
-4.99636322e-01 -7.52359331e-01 -1.21596742e+00 5.28180361e-01
-2.45860744e+00 -1.44175780e+00 -5.60570478e-01 1.69879878e+00
-2.53526449e-01 3.10912490e-01 5.25591075e-01 2.27507576e-02
6.92757905e-01 8.35452855e-01 -4.32480544e-01 -3.95183444e-01
-3.75676334e-01 -4.12051320e-01 7.72620499e-01 2.65209079e-01
-1.28047526e+00 8.86898398e-01 6.38852692e+00 1.02268493e+00
-1.07525015e+00 -7.39901140e-02 2.98053682e-01 2.55425453e-01
-4.27221060e-01 -3.15983087e-01 -7.60115445e-01 3.29900265e-01
1.56242514e+00 -2.48483151e-01 4.21107262e-01 7.09287047e-01
5.34695566e-01 4.48409617e-01 -9.03157890e-01 6.34522200e-01
-2.49544401e-02 -1.83946979e+00 6.06527746e-01 1.39837474e-01
4.95016843e-01 6.27610981e-01 2.06377968e-01 5.78247428e-01
4.86229986e-01 -1.21364284e+00 6.26641631e-01 8.28670502e-01
7.73544252e-01 -8.80807996e-01 6.96301758e-01 3.49764854e-01
-1.81294310e+00 -2.83773482e-01 -2.20974103e-01 2.77675055e-02
6.66000485e-01 2.48851150e-01 -1.14120042e+00 1.20769632e+00
3.77667010e-01 1.42350888e+00 -6.10671580e-01 9.21451092e-01
1.45931616e-01 1.08394003e+00 -3.97604138e-01 1.96661614e-02
7.49439657e-01 -4.76646990e-01 9.36097145e-01 1.41736996e+00
5.65554976e-01 -1.20557472e-01 1.63822114e-01 7.72775054e-01
1.33564770e-01 -9.38334614e-02 -1.28653979e+00 -7.66065791e-02
1.85743153e-01 8.93776953e-01 -4.87253398e-01 -2.64639944e-01
-4.75774556e-01 2.15790361e-01 5.29228091e-01 9.16163206e-01
-1.26503551e+00 2.10384224e-02 5.42406917e-01 3.51477891e-01
4.93324190e-01 -4.20116097e-01 -1.00347355e-01 -7.47897565e-01
-2.48233885e-01 -4.04432744e-01 5.66856205e-01 -7.21407294e-01
-1.21055400e+00 9.34748054e-01 4.59647238e-01 -1.67902493e+00
-5.98024249e-01 -4.30992544e-01 -7.94289052e-01 6.19652987e-01
-1.54324806e+00 -1.90505064e+00 -3.19470316e-01 7.77325153e-01
2.28347719e-01 -5.46350837e-01 5.35157144e-01 5.81927299e-01
-5.95212281e-01 5.23636699e-01 1.04336143e-01 1.71563998e-01
5.19552268e-02 -1.13341248e+00 9.39278483e-01 1.01781714e+00
-1.02498502e-01 1.81633621e-01 8.96599889e-01 -1.21659362e+00
-1.20306027e+00 -2.16358590e+00 9.19124782e-01 -1.73561215e-01
8.68923366e-01 -2.93340445e-01 -7.23462403e-01 1.04002464e+00
1.53840512e-01 -4.96358387e-02 3.35274398e-01 2.23127473e-03
-3.53478521e-01 6.75415322e-02 -6.88327610e-01 6.79609299e-01
1.58421850e+00 -6.20193720e-01 -2.81045377e-01 4.47083801e-01
9.76830065e-01 -3.08673918e-01 -8.16170037e-01 5.04714251e-01
4.76553708e-01 -8.47692728e-01 1.10572720e+00 -7.79903769e-01
2.59029120e-01 -2.82062083e-01 -2.45132074e-01 -1.31298745e+00
-6.22870207e-01 -7.24941730e-01 -5.75793505e-01 8.03363025e-01
5.02293706e-01 -8.17289174e-01 8.25623095e-01 1.15062803e-01
-6.19046032e-01 -6.82495177e-01 -1.15446460e+00 -8.75236392e-01
-8.81341845e-02 -8.16479206e-01 1.01801658e+00 6.48892045e-01
-3.13983262e-01 4.08611417e-01 -7.62173414e-01 3.78538519e-01
8.32830548e-01 -3.30686709e-03 9.84064937e-01 -1.19636047e+00
2.35039517e-01 -5.15675068e-01 -8.23227167e-01 -1.12730706e+00
5.55539727e-01 -1.28154707e+00 -4.71940547e-01 -1.71804416e+00
-1.77460343e-01 -1.64540052e-01 -1.91744700e-01 1.81571364e-01
2.53047824e-01 3.45437676e-02 1.78352177e-01 -5.90788312e-02
-5.68574369e-01 8.12462747e-01 9.17704582e-01 -3.09978217e-01
-2.29614645e-01 2.88549513e-01 2.31100935e-02 6.45317614e-01
7.27462351e-01 -2.70744354e-01 -7.07747221e-01 -1.22095130e-01
4.02127206e-01 3.30091149e-01 6.25046790e-01 -1.17270374e+00
4.12448347e-01 1.17879897e-01 7.08174985e-03 -1.54627860e+00
3.06459904e-01 -6.91594005e-01 7.48998761e-01 2.17938825e-01
-1.83350608e-01 2.53701031e-01 1.27390429e-01 1.06652224e+00
-1.99529395e-01 6.36183500e-01 1.69335365e-01 4.18829620e-01
-9.45592761e-01 8.49212587e-01 -6.56363368e-01 -4.60564166e-01
9.45254326e-01 -6.10675812e-01 -5.33925116e-01 -6.78256273e-01
-7.89147317e-01 4.43738401e-01 1.11070968e-01 5.53732514e-01
7.39212811e-01 -1.79540467e+00 -6.71647131e-01 1.11802049e-01
1.20083392e-01 -2.02710986e-01 6.12224996e-01 8.84716749e-01
-2.93811053e-01 6.70947254e-01 -2.85733610e-01 -7.01866865e-01
-1.00860178e+00 6.15568101e-01 4.54112083e-01 -4.66409773e-01
-1.12163293e+00 2.31206819e-01 1.20237224e-01 -7.47456193e-01
1.58563312e-02 -4.11304116e-01 -7.26406813e-01 1.66881576e-01
1.07200809e-01 7.16008008e-01 7.19356537e-02 -1.33699548e+00
-1.32444009e-01 4.21629608e-01 3.18502188e-01 2.26870194e-01
1.57499564e+00 -2.61017710e-01 5.61152622e-02 3.68066043e-01
1.29116058e+00 -7.56996691e-01 -1.12373209e+00 -3.13068777e-01
-1.24050207e-01 4.04770859e-02 6.13766164e-02 -3.35040152e-01
-9.96630669e-01 7.06497073e-01 4.72183228e-01 3.64412040e-01
9.20156360e-01 -2.49425128e-01 1.11953878e+00 5.95795214e-01
2.13360593e-01 -9.02556181e-01 -1.53948009e-01 9.60531652e-01
1.04532385e+00 -8.15917253e-01 -1.89191982e-01 -4.28332150e-01
-4.26298767e-01 1.23444080e+00 4.84602243e-01 -4.24306184e-01
1.21978903e+00 -7.18321428e-02 -4.37853754e-01 -4.75653917e-01
-7.98759341e-01 -4.67266798e-01 5.72711945e-01 8.58256459e-01
-2.35218734e-01 1.81249991e-01 6.50370121e-02 3.68590832e-01
-5.12130558e-01 -1.10250466e-01 3.10388535e-01 1.83917865e-01
-1.96475744e-01 -4.61697370e-01 2.82717854e-01 8.84212196e-01
2.62788743e-01 -5.01574948e-02 7.50797335e-03 8.95621717e-01
1.26334019e-02 8.18334818e-01 2.84234658e-02 -1.00345647e+00
3.70345563e-01 5.53050973e-02 -9.43726972e-02 -1.79334819e-01
-5.05916953e-01 -3.64946842e-01 6.66530728e-01 -8.58852744e-01
-5.84890842e-01 -6.77945912e-01 -9.65449989e-01 -7.12441683e-01
1.29681472e-02 -1.50587946e-01 3.06348026e-01 9.35176551e-01
4.29801285e-01 7.70778358e-01 3.73319477e-01 -1.05439401e+00
9.96461660e-02 -8.63530159e-01 -4.91275072e-01 3.15646708e-01
4.92499918e-01 -8.98361802e-01 -8.97060111e-02 -2.25443736e-01] | [6.476225852966309, 2.067124366760254] |
eaccd41d-9f39-45e2-9d2a-983d246a960c | implicit-function-theorem-estimates-on-the | 2205.12661 | null | https://arxiv.org/abs/2205.12661v3 | https://arxiv.org/pdf/2205.12661v3.pdf | Implicit Function Theorem: Estimates on the size of the domain | In this article, we present explicit estimates of the size of the domain on which the Implicit Function Theorem and the Inverse Function Theorem are valid. For maps that are twice continuously differentiable, these estimates depend upon the magnitude of the first-order derivatives evaluated at the point of interest, and a bound on the second-order derivatives over a region of interest. One of the key contributions of this article is that the estimates presented require minimal numerical computation. In particular, these estimates are arrived at without any intermediate optimization procedures. We then present three applications in optimization and systems and control theory where the computation of such bounds turns out to be important. First, in electrical networks, the power flow operations can be written as Quadratically Constrained Quadratic Programs (QCQPs), and we utilize our bounds to compute the size of permissible power variations to ensure stable operations of the power system network. Second, the robustness margin of positive definite solutions to the Algebraic Riccati Equation (frequently encountered in control problems) subject to perturbations in the system matrices are computed with the aid of our bounds. Finally, we employ these bounds to provide quantitative estimates of the size of the domains for feedback linearization of discrete-time control systems. | ['Ashutosh Jindal', 'Ravi Banavar', 'Debasish Chatterjee'] | 2022-05-25 | null | null | null | null | ['numerical-integration'] | ['miscellaneous'] | [ 5.98866343e-02 5.31816065e-01 -3.01687330e-01 3.29868108e-01
-3.92190605e-01 -9.95763898e-01 6.57769889e-02 1.48499608e-01
-1.45357999e-03 1.07465792e+00 -6.28082514e-01 -4.44719106e-01
-6.51721835e-01 -3.66898984e-01 -8.52489531e-01 -9.09214973e-01
-4.92883205e-01 -1.05106518e-01 -1.70244008e-01 -6.90795004e-01
1.70632318e-01 8.36220622e-01 -7.45940506e-01 -7.57530928e-01
9.90166366e-01 1.31308615e+00 -2.92381406e-01 6.43054187e-01
7.87054420e-01 4.85797346e-01 -5.49222529e-01 4.23137009e-01
4.98384327e-01 -3.52593720e-01 -5.88334382e-01 4.47992563e-01
-1.81581348e-01 -3.05164754e-01 -2.69008338e-01 1.30433619e+00
4.36534911e-01 3.77861142e-01 7.44674444e-01 -1.66700387e+00
-3.67201805e-01 2.53713280e-01 -4.17170733e-01 2.74484038e-01
4.81168367e-02 6.24302514e-02 6.56388998e-01 -7.08454192e-01
2.01916099e-01 6.94363236e-01 5.16269028e-01 2.08612800e-01
-1.43531895e+00 -3.68429005e-01 5.57807758e-02 -1.50875852e-01
-1.48059916e+00 -5.09571552e-01 6.96666956e-01 -6.78834856e-01
7.05453515e-01 6.47954047e-01 7.25947499e-01 -2.44307131e-01
3.66664469e-01 2.06128776e-01 5.62952936e-01 -3.97788703e-01
2.96168596e-01 3.77904952e-01 1.27321303e-01 7.97835529e-01
3.82135957e-01 3.48352529e-02 1.86264142e-01 -2.59122282e-01
1.02686369e+00 -3.61385375e-01 -7.40947127e-01 -6.26551926e-01
-1.04133463e+00 9.77619588e-01 3.87641460e-01 2.73157746e-01
-1.94806278e-01 -5.59594557e-02 1.82351291e-01 4.16253656e-01
5.81535637e-01 7.02026784e-01 -1.15589805e-01 1.52864173e-01
-3.73999059e-01 3.06308270e-02 9.35851097e-01 1.25504053e+00
4.39286739e-01 4.51346099e-01 9.46326777e-02 4.74824667e-01
1.45133883e-01 6.99812412e-01 -2.39054441e-01 -1.01627398e+00
4.40907508e-01 3.70822281e-01 4.51591551e-01 -1.17103839e+00
-3.17751884e-01 -3.16531718e-01 -8.05684745e-01 2.53429264e-01
5.12860060e-01 -7.19896257e-01 -2.02229559e-01 1.49837482e+00
2.71356463e-01 -3.29325497e-01 -3.87379192e-02 8.42513740e-01
-4.17212307e-01 1.14686847e+00 -5.80800354e-01 -9.21300948e-01
8.22870076e-01 -5.38284123e-01 -1.01981556e+00 1.98592156e-01
6.05149984e-01 -3.88915151e-01 5.56705832e-01 1.89390004e-01
-1.23235118e+00 1.74679056e-01 -1.44811511e+00 4.00166333e-01
-2.01695472e-01 3.75590205e-01 -2.08140351e-02 4.61436361e-01
-1.14871788e+00 7.78308332e-01 -6.00459754e-01 -6.92327768e-02
-2.97901660e-01 6.60162032e-01 -1.37346476e-01 8.16970170e-01
-1.19222426e+00 1.12713468e+00 3.52779478e-01 1.00223017e+00
-4.32914227e-01 -1.01666427e+00 -6.61481500e-01 1.36229619e-01
5.96435487e-01 -5.71633726e-02 9.77954268e-01 -9.40069914e-01
-1.72484243e+00 8.46577063e-02 4.18018065e-02 -1.66103974e-01
4.47113246e-01 2.95954317e-01 -1.11321770e-01 4.12088871e-01
-2.13727523e-02 -3.56494486e-01 7.59808719e-01 -7.47787595e-01
-3.55733395e-01 -7.39308000e-02 2.72041261e-01 1.84612527e-01
-2.77411431e-01 -1.64615780e-01 1.44744813e-01 -2.76164174e-01
-4.60435338e-02 -9.75946605e-01 -5.55373907e-01 2.14433149e-01
-3.32066208e-01 -8.93833041e-02 8.67999971e-01 -7.96491623e-01
1.15446711e+00 -2.17949271e+00 5.30120075e-01 6.94881082e-01
-9.20090824e-02 7.72211030e-02 4.19150472e-01 7.42024601e-01
-1.54551163e-01 2.46895596e-01 -2.43667439e-01 5.22742867e-01
1.29954115e-01 -3.46823931e-02 -6.52216017e-01 1.13677728e+00
4.90259647e-01 4.11755830e-01 -5.74825883e-01 -5.52759282e-02
3.05695415e-01 1.45190075e-01 -4.14178818e-01 -6.64836317e-02
1.98444724e-01 2.74161577e-01 -8.21232975e-01 -5.79173528e-02
3.63221288e-01 -1.22046024e-01 2.04291701e-01 -3.48173648e-01
-5.38969636e-01 -1.68135554e-01 -1.32115710e+00 6.82915032e-01
-5.51789224e-01 6.57465518e-01 9.37475145e-01 -1.41684771e+00
4.86124963e-01 5.28833628e-01 8.05745125e-01 -1.43121213e-01
2.93626457e-01 2.74953604e-01 -2.11778536e-01 -8.01737458e-02
1.50329573e-02 -2.45511562e-01 5.71704507e-02 8.30124095e-02
-2.98198640e-01 -5.91164768e-01 5.80062389e-01 2.68979669e-01
6.55364513e-01 -8.13387930e-01 4.79501843e-01 -1.13442862e+00
9.67758238e-01 1.29060984e-01 6.83225751e-01 -2.13233039e-01
-2.56043315e-01 -2.62425244e-01 1.20597208e+00 9.50532258e-02
-1.15639186e+00 -5.77603400e-01 -4.14894521e-01 4.21292514e-01
1.89230442e-01 4.46290001e-02 -5.82624674e-01 -9.59843174e-02
1.84132934e-01 6.14920855e-01 -4.92046207e-01 -2.79845715e-01
-5.17740488e-01 -5.40281296e-01 2.73991819e-03 3.61395210e-01
4.96808365e-02 -8.65745917e-02 -3.90182883e-01 8.60288814e-02
2.27357641e-01 -9.06682014e-01 -7.77369797e-01 2.76163071e-01
-8.11859727e-01 -1.36011899e+00 -8.63569677e-01 -8.23706508e-01
1.20805299e+00 -2.86804944e-01 3.46069008e-01 -1.29098505e-01
-1.70899853e-01 6.84720099e-01 3.52634132e-01 -1.76657841e-01
-2.85921842e-01 -1.90140903e-01 3.06674868e-01 9.42520332e-03
-7.99009740e-01 -2.31927127e-01 -1.65127605e-01 7.17379689e-01
-6.81278169e-01 -2.63290763e-01 4.81804386e-02 7.64436483e-01
5.19457459e-01 3.19636434e-01 6.71654642e-01 -3.32416445e-01
8.82442236e-01 -3.66041034e-01 -1.69798255e+00 3.59444618e-01
-4.68316436e-01 8.38835761e-02 1.13952863e+00 -3.55517566e-01
-4.62875724e-01 1.88873768e-01 5.80776572e-01 -3.61569047e-01
7.66038477e-01 3.80181760e-01 -1.63609713e-01 -7.78667331e-01
2.82128692e-01 3.16982530e-02 3.19293052e-01 7.70602701e-03
2.45430022e-01 4.67872202e-01 1.87712550e-01 -5.30860484e-01
9.66624856e-01 1.92630515e-01 6.22537911e-01 -1.23252702e+00
-4.50700611e-01 -4.48746324e-01 -5.97211719e-01 -4.58936125e-01
4.72256273e-01 -5.35654426e-01 -1.29804325e+00 1.55868977e-01
-1.08146548e+00 -3.38534713e-01 -4.20187026e-01 2.56776750e-01
-9.54445899e-01 1.64587229e-01 -6.41023099e-01 -1.33473802e+00
-1.15322486e-01 -1.09483051e+00 4.60926563e-01 1.55796349e-01
1.50809377e-01 -1.29761446e+00 -1.36842400e-01 -3.73872608e-01
3.98732096e-01 6.22639298e-01 9.05668497e-01 7.78548941e-02
-4.85185415e-01 -1.72326326e-01 8.18490982e-02 4.60515678e-01
4.71546985e-02 2.69430816e-01 -2.46038273e-01 -7.64893591e-01
3.81921470e-01 2.80504972e-01 -3.08721792e-02 6.11839116e-01
5.69823503e-01 -9.25271571e-01 -4.87101674e-01 4.25909191e-01
1.71781135e+00 4.41787094e-01 1.46147445e-01 -8.52075443e-02
4.14031744e-01 5.41207254e-01 4.86475855e-01 3.67333144e-01
-1.35664850e-01 5.24538517e-01 3.57709080e-01 -1.94578588e-01
8.59362423e-01 3.58320832e-01 4.46714222e-01 1.03622162e+00
2.46925112e-02 1.84312180e-01 -5.85856497e-01 7.37893879e-01
-1.71012485e+00 -6.08730793e-01 -2.12977126e-01 2.35504961e+00
7.03845322e-01 -2.53208876e-01 2.71859080e-01 4.27674115e-01
1.08889520e+00 -1.97310925e-01 -7.51235723e-01 -8.34316552e-01
1.28154755e-01 -1.91700369e-01 1.32343745e+00 9.29247916e-01
-1.02713990e+00 -3.40896547e-02 7.09563589e+00 6.30059719e-01
-1.24561465e+00 -2.59839505e-01 4.63683248e-01 9.53571871e-02
8.48265439e-02 5.82629396e-03 -5.56423962e-01 3.74047101e-01
9.78360415e-01 -1.12886226e+00 7.70199537e-01 9.34582949e-01
5.63180208e-01 -1.78977087e-01 -1.00579202e+00 5.39358199e-01
-4.66889113e-01 -8.42663467e-01 -7.99482167e-01 2.17806727e-01
1.00027287e+00 -6.45488501e-01 2.04517990e-01 -4.27707434e-02
-1.75438553e-01 -6.92307889e-01 6.04861557e-01 2.54962564e-01
6.47178471e-01 -1.09521389e+00 3.96437883e-01 3.47565025e-01
-1.24518633e+00 -3.37358356e-01 -3.70739907e-01 -2.01769948e-01
3.95069867e-01 8.94064307e-01 -6.40697718e-01 5.08070469e-01
-7.83429816e-02 5.68687439e-01 9.05456692e-02 1.04524100e+00
-1.79356083e-01 2.95331001e-01 -7.17869639e-01 -5.41847050e-01
9.01731998e-02 -7.09851742e-01 7.18242526e-01 6.45555437e-01
1.34119913e-01 4.62111533e-01 1.80259898e-01 1.05369973e+00
2.41953209e-01 1.41681477e-01 -4.69114512e-01 -2.30316862e-01
2.25281879e-01 1.15769613e+00 -4.45929706e-01 -1.77574545e-01
-4.47733998e-02 4.49803144e-01 -9.91339982e-02 7.20956028e-01
-8.70512009e-01 -1.12522376e+00 7.66867042e-01 7.70754740e-03
7.27703348e-02 -5.79701304e-01 -3.76371652e-01 -7.46125162e-01
4.92048711e-01 -3.85262966e-01 1.43707722e-01 -3.31081867e-01
-7.75513530e-01 1.70411542e-01 1.83138892e-01 -1.27575839e+00
-7.25912213e-01 -5.34230113e-01 -5.18395782e-01 1.11749887e+00
-1.15795314e+00 -1.32195875e-01 5.41497827e-01 4.20309007e-01
1.03595061e-02 3.39242458e-01 4.47991729e-01 4.18015540e-01
-9.83438551e-01 1.63078547e-01 6.36165202e-01 1.11188516e-01
-4.49414141e-02 -1.25829160e+00 -2.97088236e-01 1.12510395e+00
-8.13752055e-01 5.00525832e-01 9.37624216e-01 -2.06552893e-01
-1.92097723e+00 -8.15410495e-01 4.57005799e-01 1.48870423e-01
1.19519186e+00 -2.13113278e-01 -6.51626945e-01 8.40927601e-01
1.57033861e-01 1.40209138e-01 -1.78092539e-01 -4.20192957e-01
5.30426323e-01 -3.98433208e-01 -1.14660084e+00 4.22308296e-01
1.46692723e-01 -5.16298056e-01 6.85462262e-04 5.02418339e-01
4.93328393e-01 -4.36285168e-01 -1.23686719e+00 2.86555529e-01
2.27917001e-01 -7.46168792e-02 6.48810267e-01 -5.50477982e-01
-9.29826647e-02 -5.33203781e-01 1.01372287e-01 -1.65633142e+00
-1.70742005e-01 -1.35647333e+00 -1.49035364e-01 8.83516669e-01
6.16288066e-01 -1.02340639e+00 2.56723017e-01 8.86336207e-01
-2.24017166e-02 -8.64819586e-01 -1.01306760e+00 -1.13172722e+00
2.53410697e-01 2.98243433e-01 -6.43053502e-02 1.01277590e+00
1.04076219e+00 1.92012608e-01 -9.80397090e-02 6.94494486e-01
6.10163450e-01 -2.00184852e-01 1.45710722e-01 -6.97061121e-01
-1.11474253e-01 -3.16854179e-01 -3.32526207e-01 -1.00126588e+00
9.65308025e-02 -5.05807698e-01 2.28776768e-01 -1.25290287e+00
-5.46858311e-01 -5.38337290e-01 -5.28021790e-02 2.83932149e-01
2.17635721e-01 -1.70914322e-01 2.22506315e-01 5.40036745e-02
8.75369906e-02 6.04762435e-01 1.21019876e+00 -2.94114888e-01
-3.53325218e-01 2.80643642e-01 -1.83559701e-01 5.15907526e-01
8.96736979e-01 3.29287536e-02 -6.12241566e-01 -1.06757477e-01
2.01285809e-01 5.76664507e-01 2.03792065e-01 -8.85387003e-01
4.19354588e-01 -3.81337374e-01 -2.84839422e-01 -4.02326465e-01
2.94384927e-01 -1.17522466e+00 1.12255260e-01 8.02869916e-01
-2.87437618e-01 1.00578539e-01 2.69304097e-01 4.18643713e-01
-2.04973802e-01 -3.53880435e-01 1.02820182e+00 6.46349847e-01
-2.98509657e-01 2.59109330e-03 -7.93540061e-01 1.24529101e-01
1.37999809e+00 1.38421863e-01 -1.78161472e-01 -6.53824329e-01
-7.13297904e-01 6.04036927e-01 -5.05758189e-02 -1.94716658e-02
1.10311337e-01 -1.23408341e+00 -2.87724763e-01 2.41427809e-01
-5.86242616e-01 -2.35356480e-01 4.02065553e-02 1.43809807e+00
-5.51194787e-01 9.67938781e-01 -1.37071788e-01 -5.67568421e-01
-8.65238428e-01 6.88218057e-01 7.93035686e-01 7.02479184e-02
7.14442357e-02 2.46185333e-01 -2.93378402e-02 6.57645911e-02
4.57122214e-02 -8.07150364e-01 1.91731855e-01 -1.60290390e-01
2.54158288e-01 7.86492050e-01 4.45739068e-02 -4.03847754e-01
-2.70979643e-01 7.65871048e-01 5.37965357e-01 -1.23289458e-01
1.02797449e+00 -3.04824442e-01 -4.56375778e-01 3.92471671e-01
1.49303114e+00 -5.03492169e-03 -1.21435857e+00 2.53656536e-01
-2.89654940e-01 -9.17157009e-02 1.54742867e-01 -2.84466386e-01
-1.36776924e+00 5.29655159e-01 2.13640422e-01 7.83035994e-01
1.31121874e+00 -5.30317307e-01 2.48069450e-01 4.97776061e-01
4.27274883e-01 -1.27382398e+00 -6.96547329e-01 5.89322329e-01
1.14750504e+00 -4.97781515e-01 9.77772027e-02 -1.17531168e+00
-1.15875199e-01 1.41227055e+00 2.50520229e-01 -4.59890544e-01
9.30911541e-01 3.79643261e-01 -3.35294604e-01 4.09415066e-01
-6.14721954e-01 2.50372648e-01 4.06237483e-01 1.82644591e-01
3.15614641e-01 -7.05454964e-03 -8.83949101e-01 6.29775077e-02
1.53117046e-01 -2.04488829e-01 8.44691575e-01 7.88032949e-01
-5.26943982e-01 -5.58900177e-01 -5.69451749e-01 7.04611912e-02
-2.83706218e-01 4.60170925e-01 -2.03541905e-01 9.14742947e-01
-6.63125515e-01 1.01715279e+00 -3.59806418e-02 2.99290091e-01
8.17933142e-01 -3.09633106e-01 4.18667704e-01 -1.13271825e-01
-2.04982638e-01 -1.73715040e-01 1.01329818e-01 -4.79184687e-01
4.88699935e-02 -2.37443224e-01 -1.22950983e+00 -4.75144684e-01
-8.45203519e-01 6.78479910e-01 7.28787899e-01 7.53937125e-01
8.79128501e-02 5.12926161e-01 1.04279673e+00 -6.41180992e-01
-1.07796788e+00 -5.27497292e-01 -1.01587772e+00 -3.78692299e-01
6.11665249e-01 -4.72174764e-01 -7.16131330e-01 -9.80995689e-03] | [5.447600364685059, 2.619988441467285] |
987caf40-4b60-47a9-a399-f85f805f211c | selective-supervised-contrastive-learning | 2203.04181 | null | https://arxiv.org/abs/2203.04181v1 | https://arxiv.org/pdf/2203.04181v1.pdf | Selective-Supervised Contrastive Learning with Noisy Labels | Deep networks have strong capacities of embedding data into latent representations and finishing following tasks. However, the capacities largely come from high-quality annotated labels, which are expensive to collect. Noisy labels are more affordable, but result in corrupted representations, leading to poor generalization performance. To learn robust representations and handle noisy labels, we propose selective-supervised contrastive learning (Sel-CL) in this paper. Specifically, Sel-CL extend supervised contrastive learning (Sup-CL), which is powerful in representation learning, but is degraded when there are noisy labels. Sel-CL tackles the direct cause of the problem of Sup-CL. That is, as Sup-CL works in a \textit{pair-wise} manner, noisy pairs built by noisy labels mislead representation learning. To alleviate the issue, we select confident pairs out of noisy ones for Sup-CL without knowing noise rates. In the selection process, by measuring the agreement between learned representations and given labels, we first identify confident examples that are exploited to build confident pairs. Then, the representation similarity distribution in the built confident pairs is exploited to identify more confident pairs out of noisy pairs. All obtained confident pairs are finally used for Sup-CL to enhance representations. Experiments on multiple noisy datasets demonstrate the robustness of the learned representations by our method, following the state-of-the-art performance. Source codes are available at https://github.com/ShikunLi/Sel-CL | ['Tongliang Liu', 'Shiming Ge', 'Xiaobo Xia', 'Shikun Li'] | 2022-03-08 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Li_Selective-Supervised_Contrastive_Learning_With_Noisy_Labels_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Li_Selective-Supervised_Contrastive_Learning_With_Noisy_Labels_CVPR_2022_paper.pdf | cvpr-2022-1 | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 1.72016546e-01 -3.63816842e-02 -3.02042961e-01 -5.04866958e-01
-1.35800517e+00 -4.38669115e-01 3.76361758e-01 3.46304834e-01
-3.46967787e-01 7.56666183e-01 2.50149280e-01 2.16287434e-01
-2.75026560e-01 -8.01885605e-01 -6.18839622e-01 -1.01692235e+00
2.00974196e-01 3.49418074e-01 -2.01528937e-01 5.28699681e-02
-3.36646009e-03 1.48737445e-01 -1.52763140e+00 4.96057242e-01
9.08277869e-01 1.08088207e+00 2.98049003e-01 -3.24685909e-02
-2.26676211e-01 6.88058853e-01 -5.88417053e-01 -1.69246003e-01
1.19514138e-01 -4.78208363e-01 -5.34880459e-01 -1.41346768e-01
2.81925015e-02 -7.66189843e-02 -2.61079133e-01 1.16282439e+00
4.26123738e-01 2.01466560e-01 7.12489724e-01 -1.21627319e+00
-6.66857004e-01 8.62385273e-01 -4.02928740e-01 6.51220754e-02
9.30540562e-02 -1.06330238e-01 1.25654626e+00 -1.32900977e+00
3.22698474e-01 1.34871316e+00 5.11802852e-01 6.93268001e-01
-1.31311035e+00 -1.16908610e+00 3.58757317e-01 1.22867823e-02
-1.40684152e+00 -5.37657678e-01 9.78983641e-01 -2.95845956e-01
4.87240642e-01 1.76903591e-01 1.33129165e-01 1.35327065e+00
-3.13819230e-01 9.71384704e-01 1.17742610e+00 -3.10544491e-01
2.04074457e-01 2.98809260e-01 3.00160676e-01 4.79818672e-01
1.52563840e-01 -5.23661915e-03 -6.63711488e-01 -1.02535225e-01
4.79172260e-01 2.46124238e-01 -4.41860646e-01 -7.87795242e-03
-1.18885922e+00 8.93557131e-01 7.95541704e-01 4.80272561e-01
-3.25292617e-01 7.87903275e-03 5.08035302e-01 4.13876176e-01
7.72046030e-01 5.05401194e-01 -3.50497544e-01 4.97949608e-02
-7.89855599e-01 -1.65568739e-02 2.29254350e-01 8.78958821e-01
9.24298882e-01 -4.64993976e-02 -3.44814986e-01 1.19014287e+00
3.16627353e-01 1.36383772e-01 6.82505190e-01 -7.73326397e-01
7.81501234e-01 6.53820634e-01 -1.26069263e-01 -1.05987811e+00
-2.04022020e-01 -7.08157241e-01 -1.22076941e+00 -1.28822371e-01
2.20893815e-01 8.15984383e-02 -8.33148301e-01 1.77110398e+00
-8.66806582e-02 2.64048815e-01 8.91147703e-02 7.70132542e-01
8.94361496e-01 9.16007698e-01 9.92476344e-02 -3.72248143e-01
9.64204311e-01 -9.41082954e-01 -5.79236269e-01 -1.95073515e-01
7.28928149e-01 -5.76448858e-01 9.65424597e-01 4.39729512e-01
-6.97983325e-01 -7.37244248e-01 -8.73660684e-01 1.45198435e-01
-2.69850492e-01 4.29151088e-01 3.65631789e-01 2.20937699e-01
-6.49177611e-01 7.37299383e-01 -6.13314271e-01 2.76748568e-01
7.09649742e-01 1.15216903e-01 -5.14908433e-01 -4.68547374e-01
-1.28035438e+00 4.67710048e-01 6.43201709e-01 2.24720463e-01
-9.45759296e-01 -5.36823094e-01 -9.95240390e-01 1.59182265e-01
4.87919629e-01 -1.11242622e-01 9.05814528e-01 -8.74967337e-01
-1.11357629e+00 6.00113750e-01 -2.00269997e-01 -1.29606470e-01
5.36994636e-01 -2.77685523e-01 -4.02786791e-01 -1.82982069e-02
4.04071867e-01 6.43504143e-01 7.25392282e-01 -1.58127320e+00
-2.64801025e-01 -2.10247174e-01 -2.23851502e-01 1.19896136e-01
-5.30547500e-01 -3.00147802e-01 -3.26227099e-01 -8.23870301e-01
6.92881167e-01 -7.27778316e-01 -1.65496111e-01 8.57668146e-02
-4.43161607e-01 -6.09482646e-01 6.67371750e-01 -4.17532057e-01
9.97102022e-01 -2.56561446e+00 -1.77079830e-02 2.34511718e-01
3.08309644e-01 3.86497796e-01 -4.99845326e-01 2.31332779e-01
-2.09659845e-01 3.21754009e-01 -2.05576479e-01 -7.65905678e-01
-2.92981584e-02 3.77253264e-01 -3.78835648e-01 2.98999637e-01
4.36910808e-01 6.46067798e-01 -1.17039084e+00 -3.85492742e-01
6.63960800e-02 4.09829021e-01 -9.16408002e-02 3.61348510e-01
-1.27020702e-01 5.89118838e-01 -5.70020437e-01 6.72674775e-01
6.78571045e-01 -2.51775563e-01 3.98165509e-02 -2.59024709e-01
2.49552503e-01 5.44800997e-01 -1.24881816e+00 1.49918616e+00
-6.14329219e-01 2.42535651e-01 -2.02931777e-01 -1.34093714e+00
1.46746707e+00 3.19472104e-01 2.58966058e-01 -6.55870974e-01
6.45931959e-02 5.31278968e-01 -3.56723130e-01 -2.54979104e-01
3.05441350e-01 -4.29233573e-02 -1.56859726e-01 3.09793502e-01
2.04534784e-01 4.76962179e-02 1.21476799e-01 3.06084812e-01
6.96019471e-01 6.67731613e-02 1.07204728e-01 8.87707993e-02
2.66817242e-01 -5.01966059e-01 9.73494768e-01 7.06099927e-01
-5.37362918e-02 8.18115950e-01 4.68181700e-01 -2.34805956e-01
-8.02757084e-01 -9.92009401e-01 -2.18357593e-01 9.82957959e-01
2.29511470e-01 -5.10494053e-01 -1.90269783e-01 -9.41583335e-01
-1.50282994e-01 6.71622932e-01 -6.20618582e-01 -4.39039320e-01
-2.52794534e-01 -6.17024243e-01 4.22186524e-01 5.42966783e-01
4.41661894e-01 -1.23167932e+00 2.49915989e-03 2.99745202e-01
-3.22565705e-01 -7.37985551e-01 -2.45561779e-01 6.19001985e-01
-7.69103706e-01 -7.84225762e-01 -7.73887396e-01 -7.95260966e-01
8.79529417e-01 4.35036868e-01 9.33364689e-01 1.33883014e-01
1.17730431e-01 -3.49089652e-01 -7.00952888e-01 -1.15851797e-01
-4.04547334e-01 -7.53553063e-02 1.45253122e-01 8.68233666e-02
3.44840884e-01 -5.97110927e-01 -3.95967036e-01 4.02632207e-01
-8.93429160e-01 -9.33364704e-02 7.49611199e-01 1.33759463e+00
9.29455876e-01 2.18361616e-01 9.38036323e-01 -9.64576006e-01
6.05477929e-01 -8.81425977e-01 -3.66643161e-01 2.99716089e-02
-6.40612662e-01 2.03192517e-01 7.66090691e-01 -5.81689119e-01
-7.87429452e-01 -3.86757664e-02 -1.36442229e-01 -6.16055429e-01
-2.25826547e-01 6.20655596e-01 -2.64497787e-01 4.74246114e-01
6.25424147e-01 2.85819560e-01 -3.29701900e-01 -7.02445805e-01
1.90596342e-01 8.01368356e-01 2.86675364e-01 -6.25600100e-01
6.23062432e-01 8.96211341e-02 -4.20191169e-01 -3.84576380e-01
-1.30213428e+00 -5.40606260e-01 -4.16917622e-01 2.29679625e-02
3.45337301e-01 -1.02087188e+00 -2.64824065e-03 3.03884745e-01
-9.92679834e-01 -5.38532622e-02 -4.56535846e-01 4.95744795e-01
-1.91426769e-01 1.91017658e-01 -5.05685449e-01 -9.25976574e-01
-2.32322186e-01 -1.25956762e+00 1.10786700e+00 2.89293289e-01
-1.67363495e-01 -6.81289554e-01 -9.57917422e-02 4.05457705e-01
1.65458649e-01 5.10745347e-02 9.27631617e-01 -8.68622959e-01
-3.16032946e-01 -3.07771325e-01 -3.88123840e-01 7.58956730e-01
2.37906262e-01 -3.34281832e-01 -1.29284179e+00 -4.33799326e-01
-2.57745266e-01 -7.72776186e-01 1.29944301e+00 1.49342656e-01
1.54850745e+00 -2.78524846e-01 -2.95235813e-01 4.34452146e-01
1.22477961e+00 -3.59322429e-02 4.85955954e-01 2.54369944e-01
6.28316879e-01 5.51365376e-01 8.26902747e-01 4.51644093e-01
8.35291073e-02 5.21846175e-01 3.74190181e-01 3.39689516e-02
-4.07054648e-02 -4.32140708e-01 3.08637708e-01 1.07079089e+00
1.31257683e-01 -2.20439151e-01 -6.36010766e-01 5.92248440e-01
-1.81357467e+00 -7.37726152e-01 3.05407997e-02 2.14542484e+00
1.09508681e+00 2.57859856e-01 -2.12763518e-01 5.63152969e-01
7.40846574e-01 1.97947741e-01 -4.97389436e-01 1.25697091e-01
-1.69330209e-01 1.76915184e-01 9.30685923e-02 2.29939014e-01
-1.09240544e+00 7.64044166e-01 4.61552906e+00 1.24502444e+00
-9.21033502e-01 2.44775906e-01 9.29205835e-01 3.50480601e-02
-5.76299608e-01 -1.37216344e-01 -7.49960124e-01 7.20671058e-01
4.59124565e-01 1.36755228e-01 1.43819660e-01 9.23724890e-01
5.64129241e-02 5.59228435e-02 -1.01458776e+00 1.16707802e+00
-4.20036465e-02 -1.16266620e+00 2.87322681e-02 -2.17777297e-01
9.23200488e-01 -2.05757953e-02 2.45315954e-01 6.28412664e-01
2.74890810e-01 -1.00230551e+00 5.18280208e-01 4.39310849e-01
7.73027897e-01 -8.48891377e-01 1.12456119e+00 5.29962897e-01
-1.19317770e+00 -1.57077968e-01 -7.98050642e-01 2.02365309e-01
-1.55964166e-01 1.15723550e+00 -5.92078090e-01 6.59730494e-01
6.69984818e-01 9.45960701e-01 -4.25973356e-01 9.46384609e-01
-6.17322206e-01 8.11641335e-01 -1.74313501e-01 1.37988999e-01
1.55994952e-01 -6.56870976e-02 3.30812275e-01 1.15191066e+00
4.62413222e-01 -1.17672019e-01 3.16650897e-01 1.13228905e+00
-5.28656602e-01 5.49114235e-02 -5.42209864e-01 -4.01386656e-02
9.14158225e-01 1.25354588e+00 -4.64683175e-01 -3.71852577e-01
-6.19220035e-03 8.82082522e-01 5.69059670e-01 3.24510425e-01
-4.44031894e-01 -2.00036138e-01 2.90468663e-01 -1.83814868e-01
5.27326576e-02 -1.46776601e-03 -2.62560159e-01 -1.07894182e+00
2.01772109e-01 -7.34245777e-01 4.03736264e-01 -5.56323230e-01
-1.78845096e+00 7.22473502e-01 -5.14166877e-02 -1.57617700e+00
-6.52422979e-02 -1.86069861e-01 -4.99873042e-01 8.85234356e-01
-1.67958438e+00 -9.54038262e-01 -4.68421668e-01 4.12059993e-01
7.49738157e-01 -2.91943192e-01 9.04476523e-01 4.14703965e-01
-5.71830571e-01 9.25940633e-01 3.53237242e-01 2.21433967e-01
6.94327950e-01 -1.07061887e+00 1.29423529e-01 6.05850279e-01
4.23761755e-01 5.23570001e-01 2.56924480e-01 -4.46218759e-01
-7.17908442e-01 -1.39015532e+00 8.66254270e-01 5.51833734e-02
3.86634856e-01 -4.00939524e-01 -1.25678098e+00 3.84370029e-01
-4.10327703e-01 4.82899129e-01 6.75452054e-01 5.99873737e-02
-8.07740331e-01 -1.48157805e-01 -1.07944071e+00 2.54201144e-01
8.13272297e-01 -6.59180999e-01 -4.59513813e-01 4.03842211e-01
7.74305999e-01 -8.60377774e-02 -6.67039216e-01 2.27724358e-01
2.39741981e-01 -7.20178545e-01 7.54524529e-01 -1.35409594e-01
5.25060892e-01 -2.48380944e-01 -2.34381184e-01 -1.60257268e+00
-3.51512998e-01 -1.12756014e-01 1.61115780e-01 1.60134339e+00
5.70195019e-01 -6.73230112e-01 6.28853738e-01 1.44389570e-01
-1.78245395e-01 -8.46733332e-01 -9.89590049e-01 -9.32601750e-01
2.85550244e-02 -4.05008435e-01 4.44875956e-01 1.21969628e+00
-1.95382386e-01 2.23339334e-01 -4.44737643e-01 9.93062183e-02
5.85734129e-01 1.00414585e-02 3.16160262e-01 -1.31090343e+00
-3.27580422e-01 -1.15881816e-01 -2.34913841e-01 -9.45518374e-01
2.71931648e-01 -1.12725794e+00 3.52754861e-01 -1.47238123e+00
2.62856126e-01 -1.13230610e+00 -7.96864748e-01 9.09795344e-01
-4.78213966e-01 4.29459780e-01 2.82479405e-01 5.61488688e-01
-7.21001983e-01 8.94493997e-01 1.00068009e+00 -3.57551634e-01
-1.36565372e-01 7.23531246e-02 -7.58709490e-01 5.61307907e-01
9.70467746e-01 -8.92820656e-01 -3.39548916e-01 -2.66401261e-01
5.18272072e-02 -2.53817379e-01 1.41671941e-01 -1.02984107e+00
-1.08396597e-01 9.57826227e-02 4.68509942e-01 -3.88506770e-01
5.88575244e-01 -6.39609873e-01 -1.93851069e-01 2.80757844e-01
-7.84105957e-01 -3.45008135e-01 -7.58193880e-02 5.87348402e-01
-4.73602206e-01 -5.85749567e-01 8.83854151e-01 -1.07915469e-01
-4.80501443e-01 2.04453856e-01 -5.74510135e-02 1.20627418e-01
7.02390134e-01 1.15285710e-01 -3.93931717e-01 -3.73411804e-01
-7.73107886e-01 2.44963378e-01 2.75964998e-02 4.80877280e-01
8.29507947e-01 -1.50881112e+00 -8.88951242e-01 2.60717869e-01
5.19115627e-01 4.37993765e-01 1.89420477e-01 6.04032516e-01
1.22341514e-01 -1.10173365e-02 1.02012634e-01 -6.03850484e-01
-1.05766559e+00 3.93855900e-01 4.82989103e-02 -3.37527812e-01
-5.29711485e-01 1.12274146e+00 1.86606273e-01 -5.30510008e-01
3.12883109e-01 -1.16800793e-01 -2.91809827e-01 3.27913314e-01
6.00391746e-01 2.02832580e-01 1.26151904e-01 -5.43290615e-01
-2.23430887e-01 4.31008697e-01 -3.37955683e-01 1.85630336e-01
1.54131234e+00 2.29165200e-02 3.21450010e-02 6.23943508e-01
1.45387590e+00 -1.54120922e-01 -1.31152952e+00 -7.58796394e-01
-9.93791297e-02 -5.23155928e-01 1.49829522e-01 -6.00464046e-01
-1.32005310e+00 9.26995754e-01 6.78565025e-01 -3.98076363e-02
1.08098984e+00 1.95121557e-01 6.40338242e-01 3.85948747e-01
3.02196801e-01 -9.40835178e-01 5.51220834e-01 3.45509201e-01
8.44850361e-01 -1.43259323e+00 -6.38721064e-02 -3.53371978e-01
-6.08716011e-01 8.60317171e-01 6.25966787e-01 -1.58006310e-01
4.36187238e-01 4.46593873e-02 4.66627330e-02 -1.41464183e-02
-6.38959587e-01 -8.99617001e-02 1.58877075e-01 5.91559827e-01
3.17639977e-01 1.51841953e-01 -9.99456272e-02 8.19599807e-01
3.44731137e-02 -2.37309322e-01 1.95927918e-01 7.91170239e-01
-3.18519175e-01 -1.16791034e+00 -2.78649569e-01 5.62861860e-01
-1.50891006e-01 -1.52866751e-01 -2.21927643e-01 3.60723197e-01
4.32286590e-01 1.05712211e+00 -2.75966004e-02 -5.30791044e-01
1.46887496e-01 -7.97426421e-03 2.49108523e-02 -7.62813747e-01
-3.44959527e-01 2.63618100e-02 -4.92961816e-02 -3.41324300e-01
-2.76384771e-01 -4.18144971e-01 -1.14512908e+00 1.51816875e-01
-5.48241854e-01 3.69588941e-01 4.16322440e-01 8.74708951e-01
2.28037670e-01 6.28101110e-01 9.67093110e-01 -8.48712921e-01
-7.28306770e-01 -1.19694483e+00 -6.52311027e-01 6.34116769e-01
3.40718448e-01 -8.30892682e-01 -6.92616343e-01 -2.25508645e-01] | [9.406689643859863, 3.842633008956909] |
42a0c403-90a6-4fdb-8b0d-d951f19db2d0 | g4-grounding-guided-goal-oriented-dialogues | null | null | https://aclanthology.org/2022.dialdoc-1.11 | https://aclanthology.org/2022.dialdoc-1.11.pdf | G4: Grounding-guided Goal-oriented Dialogues Generation with Multiple Documents | Goal-oriented dialogues generation grounded in multiple documents(MultiDoc2Dial) is a challenging and realistic task. Unlike previous works which treat document-grounded dialogue modeling as a machine reading comprehension task from single document, MultiDoc2Dial task faces challenges of both seeking information from multiple documents and generating conversation response simultaneously. This paper summarizes our entries to agent response generation subtask in MultiDoc2Dial dataset. We propose a three-stage solution, Grounding-guided goal-oriented dialogues generation(G4), which predicts groundings from retrieved passages to guide the generation of the final response. Our experiments show that G4 achieves SacreBLEU score of 31.24 and F1 score of 44.6 which is 60.7% higher than the baseline model. | ['Yunbo Cao', 'Zhao Yan', 'Guanzhong Liu', 'Yiyang Du', 'Shiwei Zhang'] | null | null | null | null | dialdoc-acl-2022-5 | ['machine-reading-comprehension'] | ['natural-language-processing'] | [ 1.51814267e-01 9.45063889e-01 2.14427829e-01 -3.37408364e-01
-1.61088276e+00 -5.56259334e-01 1.22752190e+00 4.41796072e-02
-5.78049161e-02 1.41031921e+00 9.81378913e-01 -8.62498507e-02
3.03424239e-01 -7.71037519e-01 -7.58877993e-02 -3.17126870e-01
2.46787101e-01 1.23946083e+00 -1.10863730e-01 -1.03872836e+00
5.49698830e-01 -3.15745443e-01 -8.59623134e-01 9.03219402e-01
1.07696784e+00 3.33822638e-01 3.99285376e-01 1.43620944e+00
-3.99870843e-01 1.22876310e+00 -1.14514446e+00 -4.27864373e-01
-2.68224001e-01 -1.18214536e+00 -1.70541286e+00 1.18953757e-01
1.75071567e-01 -7.54392982e-01 -1.15208112e-01 6.14805102e-01
8.74119461e-01 5.23916721e-01 8.27121019e-01 -1.15699196e+00
-9.65798140e-01 8.81045222e-01 -8.16811025e-02 -3.04150395e-02
1.12875283e+00 1.48906022e-01 9.31988001e-01 -8.09456408e-01
9.61700916e-01 1.77858949e+00 -4.82614571e-03 1.16159523e+00
-9.79378402e-01 -3.25842053e-02 8.78103003e-02 -2.05605626e-01
-6.70861900e-01 -5.68152428e-01 4.61908281e-01 -4.43228662e-01
1.19244409e+00 3.45957190e-01 1.23407938e-01 1.37651372e+00
8.02850574e-02 9.81483042e-01 1.07840550e+00 -6.74574077e-01
3.11229639e-02 4.04749531e-03 3.63733441e-01 6.73743784e-01
-3.38817179e-01 -3.12657118e-01 -9.38467622e-01 -3.62700790e-01
3.67431670e-01 -6.91521347e-01 -3.49842340e-01 7.62940109e-01
-1.24703312e+00 1.40802085e+00 -9.45654586e-02 -2.33468711e-01
-7.31296599e-01 -1.90307319e-01 2.94897765e-01 6.43207312e-01
6.62616372e-01 9.90628481e-01 -6.05483688e-02 -2.80015677e-01
-2.89352566e-01 1.22287905e+00 1.38860059e+00 1.39953208e+00
4.76468086e-01 -4.51034009e-02 -1.02315116e+00 9.81579542e-01
5.35845697e-01 6.62826121e-01 6.71940386e-01 -1.00604117e+00
9.60906684e-01 6.52908444e-01 6.27552271e-01 -7.73431718e-01
-3.60870391e-01 1.01558439e-01 -8.35884154e-01 -4.64863390e-01
3.38534325e-01 -1.05564058e+00 -3.93475264e-01 1.40946567e+00
2.96772271e-01 -6.55224621e-01 8.60112965e-01 9.72623229e-01
1.51582098e+00 1.25095463e+00 2.05289587e-01 -3.36699426e-01
1.46155071e+00 -1.52986038e+00 -8.15134406e-01 -2.76774317e-01
8.55788291e-01 -9.26384389e-01 1.07216036e+00 2.49437168e-01
-1.40557301e+00 -3.74523610e-01 -7.20594347e-01 -3.27334106e-01
6.76158369e-02 -1.97936781e-02 2.46400893e-01 1.86361626e-01
-1.16886091e+00 -1.06253885e-01 5.97204082e-03 -4.38560724e-01
-5.02329290e-01 -2.03880474e-01 -3.88636068e-02 4.50682379e-02
-1.49178219e+00 8.70670378e-01 3.08619142e-01 -1.94840297e-01
-1.20099688e+00 -3.02243799e-01 -6.63830101e-01 -2.24202394e-01
2.12274119e-01 -9.55903471e-01 2.09051681e+00 -7.60831535e-02
-2.07291961e+00 7.03221619e-01 -2.44755372e-01 -5.55382967e-01
6.83343589e-01 -4.17907715e-01 -3.95087227e-02 2.52541512e-01
3.07790637e-01 8.66906822e-01 3.23582321e-01 -1.11276603e+00
-7.62684643e-01 -1.21457614e-01 4.59595501e-01 1.02403605e+00
1.62202969e-01 -1.37198893e-02 -6.13872260e-02 -2.97673374e-01
-1.71538293e-01 -9.46770132e-01 -2.87673861e-01 -1.11386395e+00
-8.44805181e-01 -9.66306150e-01 2.00898275e-01 -8.73704672e-01
1.07365334e+00 -1.15485978e+00 4.22403425e-01 -5.61962068e-01
1.87983558e-01 2.88188830e-02 -4.58745003e-01 1.40438902e+00
7.12091506e-01 1.52778476e-01 2.63805211e-01 -3.39126527e-01
3.50462228e-01 -4.00642186e-01 -5.59434056e-01 -4.24214214e-01
2.85045981e-01 1.00343931e+00 -1.18276560e+00 -3.32679689e-01
-2.86341727e-01 -7.59388730e-02 -3.96291256e-01 1.04826736e+00
-9.69947934e-01 5.01106322e-01 -8.34770739e-01 2.31752530e-01
2.50913233e-01 -1.75611719e-01 2.65815943e-01 4.72778469e-01
3.20748873e-02 6.90967143e-01 -4.60700154e-01 1.96211934e+00
-5.21009922e-01 4.66668427e-01 -1.57278582e-01 -2.13245511e-01
1.38811982e+00 7.57038891e-01 -1.77671179e-01 -9.24423516e-01
-9.39953998e-02 -3.89595442e-02 -2.14941248e-01 -7.74568617e-01
1.24324214e+00 -5.16227670e-02 -5.73299587e-01 1.08967221e+00
3.15186270e-02 -6.29073918e-01 4.46324050e-01 8.52771223e-01
1.00470674e+00 -3.90752554e-02 7.35882521e-02 -2.29685500e-01
5.35438299e-01 4.84158754e-01 -1.62242070e-01 1.22389317e+00
1.46859944e-01 5.46209276e-01 6.32211208e-01 -1.05047785e-01
-7.12708414e-01 -4.79392618e-01 7.59653568e-01 1.32434726e+00
-7.30180591e-02 -5.24012208e-01 -1.02091980e+00 -7.34051406e-01
-4.76950824e-01 1.16800177e+00 -3.79633367e-01 6.84845671e-02
-7.27763772e-01 -6.84715748e-01 6.03805840e-01 1.59750536e-01
8.02804172e-01 -1.27701652e+00 -2.27119282e-01 4.94034410e-01
-1.06295800e+00 -1.03366053e+00 -4.54419762e-01 -4.42855537e-01
-6.03229702e-01 -9.30341601e-01 -8.87455761e-01 -8.63470197e-01
3.60482335e-01 3.84469718e-01 1.50078118e+00 2.41362095e-01
3.02882344e-01 2.94486821e-01 -7.73658097e-01 -5.02327025e-01
-1.05793095e+00 5.85015535e-01 -4.76706922e-01 -4.51760650e-01
3.00301760e-01 1.18107192e-01 -4.31310266e-01 1.25082612e-01
-3.72685224e-01 7.16854632e-01 2.91202009e-01 9.74425495e-01
-1.16716832e-01 -6.82380915e-01 1.30926180e+00 -9.76661444e-01
2.00688982e+00 -8.07212830e-01 -2.31058016e-01 4.42919582e-01
-4.44606036e-01 -5.15600014e-03 6.26785517e-01 4.52661887e-02
-1.46794784e+00 -5.52055717e-01 -2.27533281e-01 7.75413454e-01
-2.37503618e-01 5.69733322e-01 1.19785771e-01 8.08101296e-01
1.03293002e+00 4.12027776e-01 7.32569769e-02 -4.14501071e-01
3.91900003e-01 1.06428218e+00 3.40354472e-01 -8.87245715e-01
1.85354412e-01 -4.90152925e-01 -6.01037920e-01 -8.84925961e-01
-9.62591469e-01 -3.09841603e-01 -2.12588072e-01 -4.81676221e-01
8.99902940e-01 -9.68008041e-01 -7.18848407e-01 4.18221831e-01
-1.74288607e+00 -8.23279679e-01 1.82924122e-01 1.29207879e-01
-8.15675020e-01 4.71083596e-02 -7.86403596e-01 -1.00101388e+00
-1.09040487e+00 -8.43620062e-01 1.04045129e+00 4.51755494e-01
-6.96825624e-01 -1.05821466e+00 4.71351385e-01 9.30329323e-01
2.50605226e-01 2.95580626e-01 9.27966475e-01 -1.05641270e+00
-4.82004821e-01 -1.73384607e-01 -4.97177523e-03 -1.77878335e-01
-5.81832464e-05 -4.55431700e-01 -7.27824032e-01 -8.82652551e-02
-9.31883231e-02 -1.16393816e+00 3.01367491e-01 -7.51122385e-02
1.29369482e-01 -8.11937928e-01 1.71504393e-01 -3.87168229e-01
9.27364767e-01 4.60344940e-01 5.57085514e-01 2.27891952e-01
2.71389544e-01 1.04840589e+00 1.04076087e+00 6.09510124e-01
9.83745694e-01 6.48974299e-01 1.57378450e-01 2.49919891e-01
-1.96397349e-01 -5.96778214e-01 4.44246322e-01 9.24323380e-01
8.02000016e-02 -1.17127240e+00 -7.84203649e-01 5.60169697e-01
-2.17371964e+00 -9.67244029e-01 -4.93912578e-01 1.46524131e+00
1.08533084e+00 -2.94914544e-01 3.42589557e-01 -6.01659298e-01
6.39526904e-01 2.00450018e-01 -1.88412532e-01 -8.07019651e-01
-2.84084734e-02 -2.54719108e-01 -4.94775176e-01 1.20661855e+00
-4.41877335e-01 1.21188307e+00 6.46415472e+00 3.18684578e-01
-4.09558326e-01 6.67514354e-02 6.50917172e-01 2.91046739e-01
-4.01135892e-01 -1.45317063e-01 -1.25896060e+00 2.15547591e-01
1.17307699e+00 -7.92365611e-01 2.12104604e-01 5.68243325e-01
5.78998446e-01 -8.23527575e-02 -9.99364793e-01 4.25145626e-01
2.58536994e-01 -1.38799310e+00 5.66845477e-01 -6.04608618e-02
7.71749973e-01 -3.11894715e-01 -3.05418789e-01 7.13386595e-01
1.22921693e+00 -1.08088803e+00 3.52192104e-01 4.62437928e-01
1.89813703e-01 -6.77092552e-01 6.35926306e-01 7.58724213e-01
-3.65139335e-01 2.79500842e-01 -4.59923089e-01 -2.31367409e-01
5.90002120e-01 -3.57115939e-02 -1.70527911e+00 6.59434319e-01
-6.13753274e-02 1.01303600e-01 -9.31481645e-02 3.64382714e-01
-4.63656664e-01 5.15544891e-01 3.08390468e-01 -6.62256241e-01
7.35000670e-01 -5.01750112e-02 5.50849915e-01 1.21752369e+00
1.40845835e-01 6.67234719e-01 5.20988882e-01 6.05216920e-01
-2.60529965e-01 3.94370407e-01 -5.09356856e-01 -2.67207205e-01
5.76455832e-01 1.23690104e+00 -5.71555123e-02 -5.86320400e-01
1.48947269e-01 1.05876791e+00 4.24880505e-01 3.04264069e-01
-4.78265852e-01 -5.06150246e-01 1.31104723e-01 -2.86196798e-01
-3.01184595e-01 -1.98004141e-01 6.89288005e-02 -9.58606720e-01
-3.18013668e-01 -1.51921463e+00 4.30129796e-01 -9.76806223e-01
-1.06012547e+00 1.11828375e+00 -3.13607641e-02 -7.82970846e-01
-1.21895909e+00 -1.21471420e-01 -7.52562463e-01 1.22874677e+00
-1.37774253e+00 -1.14877450e+00 -6.01581454e-01 4.06486779e-01
1.57566547e+00 -4.34529006e-01 1.39420378e+00 -2.65657753e-01
-4.78184491e-01 3.62735301e-01 -1.75257355e-01 4.81950864e-02
8.33738863e-01 -1.39027834e+00 6.75211549e-01 4.22968715e-01
-3.35468799e-01 6.21477246e-01 8.56739819e-01 -8.73536110e-01
-1.36852682e+00 -1.03408241e+00 1.49313462e+00 -5.30576169e-01
3.25569153e-01 -2.52376616e-01 -6.33162677e-01 4.93576109e-01
1.19785213e+00 -1.36796999e+00 7.83141255e-01 -8.60234536e-03
1.75905913e-01 6.00521684e-01 -1.04312122e+00 8.57561350e-01
7.05815792e-01 -2.32336059e-01 -6.55676365e-01 9.75821137e-01
9.24309909e-01 -5.78078032e-01 -8.56747746e-01 -2.21050218e-01
1.41618818e-01 -7.20351219e-01 5.16091287e-01 -1.03026450e+00
8.73485804e-01 1.22660950e-01 -2.27831319e-01 -1.69552052e+00
-9.60942060e-02 -1.38372684e+00 -1.86148450e-01 1.36593699e+00
6.64183199e-01 -3.16925138e-01 5.46366811e-01 8.05879653e-01
-4.42959815e-01 -4.18456554e-01 -2.82548100e-01 -3.13755035e-01
4.54596460e-01 4.15177256e-01 7.55836487e-01 7.16693640e-01
4.72171038e-01 1.28923547e+00 -6.34343624e-01 -1.67128801e-01
3.25467944e-01 1.38008341e-01 1.37659156e+00 -9.39900219e-01
-1.49282470e-01 -9.79280770e-02 5.42200744e-01 -1.63916779e+00
1.80731252e-01 -6.09203756e-01 4.55418497e-01 -2.10848117e+00
5.53304963e-02 -9.87902507e-02 7.24119365e-01 1.89701453e-01
-3.87747198e-01 -6.70627654e-02 2.57745594e-01 1.64706245e-01
-7.72068381e-01 7.24646211e-01 1.54265523e+00 -2.53355294e-01
-4.53167140e-01 -4.70372252e-02 -1.15558779e+00 2.17724502e-01
1.01579893e+00 -2.38087550e-01 -9.38029408e-01 -6.32209539e-01
1.92749482e-02 9.51834738e-01 -9.25859958e-02 -4.10562694e-01
2.88100213e-01 -5.47333241e-01 -1.64203942e-01 -7.37698197e-01
4.47306484e-01 4.10245419e-01 -3.05033386e-01 2.60508984e-01
-1.13670695e+00 3.92877579e-01 -2.59359349e-02 3.81021380e-01
-2.13741958e-01 -5.07334709e-01 1.87658206e-01 -6.41519785e-01
-5.50736666e-01 -1.74382120e-01 -1.00002098e+00 7.07789242e-01
6.10926628e-01 2.70941228e-01 -1.03299510e+00 -1.10933995e+00
-6.03410721e-01 7.80051827e-01 -8.52524489e-02 6.86875045e-01
8.34165156e-01 -1.17159510e+00 -1.67602849e+00 -4.23383862e-01
8.15138593e-02 1.55173436e-01 3.09051931e-01 2.23315373e-01
-6.13316894e-01 9.54868317e-01 -4.13941853e-02 -2.08278537e-01
-1.23867655e+00 -3.43594342e-01 1.52001783e-01 -7.14425743e-01
-5.48657000e-01 1.00716174e+00 3.66304368e-02 -6.03883028e-01
5.03150374e-03 3.96530330e-01 -6.08260214e-01 2.76452839e-01
9.72025275e-01 5.18179655e-01 5.60594350e-02 -4.18462873e-01
3.11912090e-01 -1.69600189e-01 -7.39117861e-01 -7.58817613e-01
9.42522287e-01 -4.14222926e-01 1.77786802e-03 1.61791310e-01
8.00419331e-01 -2.70665646e-01 -8.59854043e-01 -1.90461442e-01
9.14270803e-02 -1.42708778e-01 -3.76483947e-01 -1.33141506e+00
-2.05140114e-01 6.49058044e-01 -1.90818176e-01 5.22652447e-01
4.45930541e-01 -2.27128193e-01 9.88702595e-01 9.22821045e-01
2.03556284e-01 -1.12844145e+00 6.91654444e-01 1.03895354e+00
1.68542671e+00 -1.22913492e+00 -2.36130252e-01 -1.27149805e-01
-1.28022170e+00 1.02954328e+00 1.24406755e+00 4.99738082e-02
-2.42627934e-01 -2.16906875e-01 5.13302147e-01 -3.09172601e-01
-1.50255799e+00 1.92820683e-01 1.31214142e-01 6.62963450e-01
7.97642827e-01 -8.59033242e-02 -5.38415909e-01 5.44577718e-01
-6.31195843e-01 -3.39532167e-01 1.03883231e+00 9.90648210e-01
-7.15216875e-01 -1.05258262e+00 -1.49481833e-01 1.36640608e-01
-1.63991109e-01 -2.04337180e-01 -1.17502570e+00 4.75664079e-01
-1.03077471e+00 1.87286186e+00 -2.41629004e-01 -3.66328210e-01
4.13819730e-01 4.93365526e-01 1.59034416e-01 -1.09786093e+00
-9.90788937e-01 -1.16773993e-01 1.08818817e+00 -2.69407798e-02
-2.05842286e-01 -2.86387116e-01 -1.26191986e+00 -4.34612304e-01
-3.42093974e-01 8.50301862e-01 5.37948310e-01 7.69965291e-01
5.61699927e-01 3.51583600e-01 7.33795762e-01 -5.03775716e-01
-8.11309457e-01 -1.55864155e+00 -5.83389774e-02 2.76289493e-01
2.09496245e-01 -4.64005806e-02 1.10246874e-01 -2.41236500e-02] | [12.51427173614502, 8.157846450805664] |
e8fd3a5b-d94d-4f56-937f-9aaa4e1f7988 | full-frame-scene-coordinate-regression-for | 1802.03237 | null | http://arxiv.org/abs/1802.03237v2 | http://arxiv.org/pdf/1802.03237v2.pdf | Full-Frame Scene Coordinate Regression for Image-Based Localization | Image-based localization, or camera relocalization, is a fundamental problem
in computer vision and robotics, and it refers to estimating camera pose from
an image. Recent state-of-the-art approaches use learning based methods, such
as Random Forests (RFs) and Convolutional Neural Networks (CNNs), to regress
for each pixel in the image its corresponding position in the scene's world
coordinate frame, and solve the final pose via a RANSAC-based optimization
scheme using the predicted correspondences. In this paper, instead of in a
patch-based manner, we propose to perform the scene coordinate regression in a
full-frame manner to make the computation efficient at test time and, more
importantly, to add more global context to the regression process to improve
the robustness. To do so, we adopt a fully convolutional encoder-decoder neural
network architecture which accepts a whole image as input and produces scene
coordinate predictions for all pixels in the image. However, using more global
context is prone to overfitting. To alleviate this issue, we propose to use
data augmentation to generate more data for training. In addition to the data
augmentation in 2D image space, we also augment the data in 3D space. We
evaluate our approach on the publicly available 7-Scenes dataset, and
experiments show that it has better scene coordinate predictions and achieves
state-of-the-art results in localization with improved robustness on the
hardest frames (e.g., frames with repeated structures). | ['Juha Ylioinas', 'Xiaotian Li', 'Juho Kannala'] | 2018-02-09 | null | null | null | null | ['image-based-localization', 'camera-relocalization'] | ['computer-vision', 'computer-vision'] | [ 2.73196995e-01 -2.17314571e-01 -9.52986255e-02 -4.52427506e-01
-6.91618264e-01 -4.65418369e-01 4.11108643e-01 -2.36770228e-01
-6.26943290e-01 4.79930848e-01 5.87552264e-02 -2.60201871e-01
2.05469057e-01 -6.89534485e-01 -1.15898180e+00 -7.28326499e-01
4.62354302e-01 1.00857526e-01 3.34792465e-01 6.20370694e-02
3.66839349e-01 6.60477459e-01 -1.22750592e+00 -5.19300252e-02
6.48428559e-01 1.01946402e+00 3.59720945e-01 4.76056904e-01
1.52527601e-01 9.25950825e-01 -2.03584835e-01 -2.40652099e-01
3.63001823e-01 -2.25046575e-01 -6.44127786e-01 3.65701675e-01
7.39142120e-01 -3.43253791e-01 -4.69985455e-01 1.04861462e+00
2.79833734e-01 1.26909032e-01 3.40864509e-01 -9.60323691e-01
-4.42738831e-01 -1.02181092e-01 -8.06406498e-01 -1.73042759e-01
5.31286359e-01 8.17757919e-02 6.04870796e-01 -1.11281717e+00
5.47585130e-01 1.06828642e+00 8.60150456e-01 3.47643048e-01
-1.28854334e+00 -5.94827175e-01 2.34411284e-01 1.48102954e-01
-1.58709598e+00 -3.88352990e-01 1.05841935e+00 -4.50836331e-01
7.60537446e-01 -1.05364367e-01 6.70451224e-01 1.00312996e+00
2.42547825e-01 6.36416554e-01 9.73561347e-01 -5.29713571e-01
1.83799341e-01 -4.07860547e-01 -3.42972249e-01 9.39387381e-01
-6.83423579e-02 1.74349584e-02 -3.78310442e-01 2.07682967e-01
1.05831444e+00 3.48833412e-01 -3.43007356e-01 -7.87528932e-01
-1.52885330e+00 7.87592411e-01 9.55416262e-01 -1.21568948e-01
-4.82146293e-01 2.86868721e-01 -4.98433560e-02 -2.16302142e-01
4.57485944e-01 4.93475944e-01 -5.54558337e-01 1.21714048e-01
-8.27607095e-01 1.29202098e-01 4.80756521e-01 8.88693273e-01
1.18255222e+00 -1.82011440e-01 1.49449438e-01 7.86520541e-01
3.17750543e-01 4.97298002e-01 4.29902256e-01 -1.14357007e+00
5.91703713e-01 5.58775187e-01 1.22628652e-01 -1.16501784e+00
-4.17875469e-01 -3.36726516e-01 -8.67878735e-01 1.90631866e-01
4.06437516e-01 -1.67643219e-01 -1.13724840e+00 1.63447738e+00
4.31947887e-01 4.44549561e-01 1.34756742e-03 1.17075861e+00
4.86652285e-01 6.72761619e-01 -2.10231915e-01 3.70328128e-03
1.06629384e+00 -1.27468419e+00 -3.04129541e-01 -5.85996449e-01
5.77455997e-01 -9.22766984e-01 8.83254170e-01 2.95453012e-01
-5.57444453e-01 -7.55010128e-01 -9.60689902e-01 -1.61985993e-01
-2.88864493e-01 4.13809031e-01 3.93222690e-01 1.40422806e-01
-9.66371238e-01 3.14097762e-01 -9.88056362e-01 -4.82054740e-01
2.55470186e-01 3.01133513e-01 -8.07304740e-01 -3.91341448e-01
-7.32238591e-01 9.40065444e-01 3.68487626e-01 1.71908185e-01
-7.22835541e-01 -3.71804148e-01 -1.33331263e+00 -1.26938537e-01
5.03179491e-01 -7.61216640e-01 1.07420933e+00 -8.39612424e-01
-1.58440387e+00 6.35824621e-01 -3.50776076e-01 -3.39818507e-01
3.87369871e-01 -3.62602502e-01 3.59927639e-02 -3.83917801e-02
2.62413293e-01 1.09958851e+00 8.68269145e-01 -1.31726599e+00
-6.62346840e-01 -2.96618968e-01 1.97268188e-01 2.82501787e-01
3.28343548e-02 -2.32438102e-01 -8.65932405e-01 -5.78407049e-01
7.67140210e-01 -1.30189192e+00 -5.17490029e-01 1.57425255e-01
-5.72448432e-01 9.32509303e-02 7.59369731e-01 -6.27286077e-01
7.26688683e-01 -2.11227489e+00 2.63963789e-01 9.32938755e-02
-4.52754945e-02 -2.52979025e-02 -1.82531863e-01 3.32468450e-02
-1.52323365e-01 -2.35287279e-01 -1.63251474e-01 -6.67599738e-01
-4.68012244e-01 3.59468251e-01 -1.39895692e-01 7.34219193e-01
3.32739502e-01 6.67105258e-01 -7.99388885e-01 -3.90329301e-01
6.79754317e-01 5.84405124e-01 -7.88273752e-01 3.41868967e-01
-1.98955074e-01 9.84483778e-01 -2.64337271e-01 5.32411635e-01
6.79973602e-01 -1.22012362e-01 -2.20750123e-01 -3.82045984e-01
-3.71341079e-01 1.87333375e-01 -1.18804681e+00 2.20002198e+00
-6.38005018e-01 6.29663229e-01 -2.20207989e-01 -9.32475686e-01
9.69775915e-01 -7.99237192e-02 5.74783325e-01 -4.87132996e-01
-1.19928215e-02 3.46919522e-02 -2.58450598e-01 -4.15528268e-01
5.13047636e-01 2.89051622e-01 -5.65471463e-02 -4.32819538e-02
1.04758844e-01 -2.15305045e-01 -8.35406259e-02 -2.56261140e-01
9.75334406e-01 5.59018850e-01 2.68016011e-01 1.22219779e-01
6.17677093e-01 1.93185404e-01 6.46861911e-01 4.55092669e-01
2.20945984e-01 1.07085621e+00 1.09906547e-01 -6.93522751e-01
-1.16563380e+00 -6.77455425e-01 8.30676407e-02 6.93175793e-01
5.14859080e-01 -4.26997900e-01 -7.54426003e-01 -6.60530508e-01
-2.26309046e-01 4.10429001e-01 -3.94476473e-01 -5.71838915e-02
-7.48741210e-01 -4.14404690e-01 9.21887904e-02 5.90683460e-01
8.97871375e-01 -7.99028099e-01 -7.24810481e-01 1.80131644e-01
-4.26538110e-01 -1.42033899e+00 -4.64160651e-01 2.89724678e-01
-6.96073413e-01 -1.16373503e+00 -7.46684015e-01 -7.78727829e-01
1.02016890e+00 5.63479245e-01 6.90759778e-01 -2.01981634e-01
-7.23689273e-02 1.95262983e-01 -4.28767711e-01 -3.31786126e-02
-2.02557631e-02 1.15427300e-01 6.15471639e-02 8.40910748e-02
1.58741307e-02 -3.28963548e-01 -7.37106085e-01 5.45575261e-01
-7.20294416e-01 4.22002703e-01 6.97961271e-01 1.06288552e+00
9.43373799e-01 -7.93859512e-02 -1.88803077e-01 -6.57140970e-01
-1.68115757e-02 -1.80917010e-01 -9.48908567e-01 1.04344428e-01
-3.33531827e-01 1.02428593e-01 6.33702219e-01 -3.71762127e-01
-8.16863477e-01 1.01132572e+00 -2.18808323e-01 -8.38008821e-01
-2.78508782e-01 5.58554232e-01 -8.18906650e-02 -4.87848729e-01
5.79866886e-01 2.60453045e-01 -7.97218829e-02 -4.70344961e-01
3.69047821e-01 3.90042841e-01 7.88254917e-01 -3.42577636e-01
9.83069718e-01 4.10499871e-01 1.55053154e-01 -5.64248204e-01
-1.03604901e+00 -6.00733697e-01 -1.04847109e+00 -2.62222797e-01
1.17543435e+00 -1.28579628e+00 -5.22098064e-01 3.83588612e-01
-1.47581601e+00 -3.34739357e-01 7.00179860e-02 8.12576473e-01
-6.55616820e-01 2.94768453e-01 -2.47768268e-01 -5.39089322e-01
2.42438093e-02 -1.45758820e+00 1.41230261e+00 2.90762305e-01
9.00833085e-02 -6.14094377e-01 6.05779029e-02 2.30020866e-01
6.22864999e-02 3.42302710e-01 4.74856377e-01 -1.52665451e-01
-8.53998780e-01 -3.53325903e-01 -2.36373305e-01 3.99074852e-01
3.63461040e-02 -1.31019279e-01 -9.17047501e-01 -3.24284762e-01
-4.60083876e-03 -1.70263991e-01 8.31158042e-01 3.65751117e-01
1.31433725e+00 -1.60922721e-01 -3.13956201e-01 1.05768979e+00
1.55234814e+00 7.21512735e-02 6.33106589e-01 5.84565103e-01
1.00614166e+00 3.85267735e-01 8.21483254e-01 2.77747035e-01
5.73565483e-01 8.60836685e-01 7.20484614e-01 -3.45483124e-01
-1.37568131e-01 -5.48360348e-01 3.03442180e-01 4.90409404e-01
-6.83608949e-02 -9.89179388e-02 -9.13044989e-01 2.03922838e-01
-2.01016140e+00 -5.95710576e-01 1.43263163e-02 2.35348511e+00
5.36693633e-01 -1.41716436e-01 -2.38327593e-01 -9.60265547e-02
7.79988945e-01 1.35881245e-01 -5.23059309e-01 1.81386352e-01
-1.15865208e-01 -1.41499192e-01 8.95312011e-01 4.98353690e-01
-1.36514628e+00 1.24146163e+00 5.59927845e+00 4.50006127e-01
-1.49296021e+00 -2.98710316e-02 6.80045784e-01 2.01511338e-01
2.98595190e-01 2.35475913e-01 -7.47907996e-01 3.55669200e-01
2.98229933e-01 5.37056148e-01 6.26995087e-01 1.09245670e+00
2.02370018e-01 -3.95059317e-01 -1.09828222e+00 1.26045191e+00
2.65910655e-01 -1.31957209e+00 -3.59347850e-01 4.83989976e-02
8.59574497e-01 3.75994831e-01 -2.19973132e-01 6.00084960e-02
1.78474173e-01 -7.87613809e-01 9.32595074e-01 5.05419970e-01
8.09443653e-01 -5.29258072e-01 8.25875700e-01 5.74918509e-01
-1.08137560e+00 -1.32559702e-01 -5.16005397e-01 -1.09091505e-01
1.16418228e-01 4.88150686e-01 -7.84986556e-01 4.35582727e-01
8.49830806e-01 9.05940294e-01 -7.88075566e-01 1.26214933e+00
-5.75193286e-01 2.06590146e-01 -4.74669218e-01 2.65825093e-01
1.94408029e-01 -1.87303141e-01 3.10133606e-01 8.13368142e-01
5.49347758e-01 -5.44133596e-02 5.00130296e-01 6.52601600e-01
-2.95977946e-02 -1.48023203e-01 -8.50677371e-01 6.16678655e-01
3.95385832e-01 1.21261084e+00 -6.27281427e-01 -7.23363236e-02
-5.53135633e-01 1.19328189e+00 4.60364163e-01 4.39553499e-01
-8.25940430e-01 -1.61763310e-01 5.31983972e-01 3.64884995e-02
3.68418485e-01 -6.63144350e-01 -1.00957550e-01 -1.33749163e+00
1.76878512e-01 -4.66342688e-01 1.94213446e-02 -1.13761520e+00
-1.01829791e+00 5.44490755e-01 -1.42441019e-01 -1.56922758e+00
-3.84638697e-01 -5.89211762e-01 -3.68892580e-01 7.87902236e-01
-1.64910090e+00 -1.29358399e+00 -6.21385336e-01 7.55644917e-01
5.80808103e-01 1.03299379e-01 6.83348298e-01 2.02027321e-01
-5.17783880e-01 3.49523991e-01 -4.31123264e-02 4.38699156e-01
7.94775307e-01 -9.86430109e-01 3.03439200e-01 1.00019312e+00
3.22228760e-01 4.81633037e-01 4.75874931e-01 -3.96164089e-01
-1.54447865e+00 -1.39725280e+00 7.17110038e-01 -4.58723783e-01
3.93312216e-01 -4.04277772e-01 -4.75002259e-01 7.50018954e-01
-1.03780314e-01 4.73933935e-01 1.88353032e-01 -7.53675103e-02
-1.92209259e-01 -1.99407235e-01 -9.54476595e-01 5.97536385e-01
1.02410603e+00 -5.84690928e-01 -3.20627570e-01 3.13866466e-01
7.28697300e-01 -8.59070659e-01 -6.78592443e-01 5.68143368e-01
4.16989058e-01 -8.82842183e-01 1.07389069e+00 -5.03679924e-02
4.73859429e-01 -8.45829427e-01 -2.98006356e-01 -1.35770762e+00
-3.52618843e-01 -3.05058420e-01 2.04328135e-01 1.03928971e+00
2.25829974e-01 -5.15253961e-01 9.49691772e-01 4.41379994e-01
-2.82946557e-01 -7.94348419e-01 -1.04770613e+00 -4.77378428e-01
-3.37092310e-01 -4.32181865e-01 4.24625486e-01 7.01031208e-01
-4.25624937e-01 3.97316903e-01 -5.24429142e-01 5.61369598e-01
3.18104416e-01 1.60104707e-01 1.21177304e+00 -9.88686860e-01
-4.43585478e-02 1.54358363e-02 -6.69991076e-01 -1.50879669e+00
3.29991341e-01 -4.74603087e-01 5.28023005e-01 -1.44177210e+00
4.29236796e-03 -4.78813469e-01 -3.18463370e-02 6.17845595e-01
-5.71758524e-02 6.09505653e-01 3.25751781e-01 4.29005295e-01
-7.38408208e-01 5.30259430e-01 1.16860712e+00 -1.43667638e-01
-1.11736707e-01 5.91630600e-02 -2.75015652e-01 8.64127636e-01
5.46830833e-01 -3.97328734e-01 -1.20563030e-01 -7.89149940e-01
-1.79303195e-02 4.53434810e-02 6.83168292e-01 -1.32248998e+00
4.95758355e-01 -7.13675022e-02 8.58413219e-01 -7.17338145e-01
6.35724962e-01 -1.12646711e+00 1.62121311e-01 3.92336190e-01
-1.61900207e-01 2.31871307e-01 -1.29676595e-01 5.95699430e-01
-4.38044608e-01 -8.07418674e-02 7.86940277e-01 -1.32322356e-01
-8.95128906e-01 4.97296929e-01 -8.42145383e-02 -5.22426248e-01
9.94063020e-01 -2.85238266e-01 -1.13897771e-01 -3.91652852e-01
-4.57047135e-01 1.27992153e-01 8.77825439e-01 3.56630504e-01
7.21131682e-01 -1.36770141e+00 -3.23424429e-01 4.57535923e-01
3.82194817e-01 7.08115220e-01 -1.35795074e-02 9.54735041e-01
-8.26345086e-01 3.43130887e-01 -1.91274211e-02 -1.05160773e+00
-1.01822293e+00 5.91252446e-01 3.15338045e-01 3.39298546e-02
-5.23383141e-01 7.18832850e-01 3.52016956e-01 -6.23422682e-01
2.74513811e-01 -4.71754044e-01 -1.02001853e-01 -4.00773138e-01
1.89288944e-01 -1.45618752e-01 5.86262755e-02 -9.41167116e-01
-3.48877698e-01 1.05979657e+00 2.13457331e-01 6.09385082e-03
1.24170172e+00 -1.73147306e-01 -1.08150892e-01 1.29477188e-01
1.35020101e+00 -5.68553805e-02 -1.75686705e+00 -4.25410956e-01
-7.89627060e-02 -7.16048539e-01 2.51712054e-01 -5.75560510e-01
-1.22579360e+00 7.56598532e-01 6.33117378e-01 -2.63273507e-01
1.12901390e+00 -3.74133810e-02 5.69978833e-01 5.34907520e-01
5.38101733e-01 -8.42351079e-01 8.96490440e-02 7.68489480e-01
7.61984110e-01 -1.48275793e+00 5.27431257e-02 -4.95902747e-01
-4.57218766e-01 1.37197840e+00 7.83086896e-01 -2.24853367e-01
5.04799724e-01 6.12751953e-02 1.15862392e-01 1.25876039e-01
-2.23672062e-01 -2.26517230e-01 1.78307548e-01 5.21149337e-01
2.58073181e-01 -1.82182550e-01 3.33167195e-01 1.12723343e-01
-1.51991487e-01 -2.45328974e-02 2.41671115e-01 8.49785924e-01
-2.69138247e-01 -1.04828167e+00 -5.86592197e-01 7.61029646e-02
-1.03096254e-01 -1.09621249e-01 -2.16654480e-01 6.02135241e-01
3.24691683e-01 8.12291920e-01 1.65912405e-01 -6.19766176e-01
3.33347589e-01 -1.75457671e-01 3.96656185e-01 -6.29534304e-01
-8.65140706e-02 2.52434552e-01 -2.22775608e-01 -8.00449252e-01
-6.53088212e-01 -5.51178873e-01 -1.04721808e+00 -9.35660079e-02
-5.64455271e-01 -1.95913404e-01 1.05905569e+00 9.50429022e-01
3.78698528e-01 3.47456545e-01 8.76814246e-01 -1.47754455e+00
-1.88156858e-01 -9.26023126e-01 -7.19171539e-02 1.93487287e-01
5.03496289e-01 -7.39850938e-01 -4.01480496e-02 1.95117697e-01] | [7.863763809204102, -2.225550413131714] |
e995ddcb-72b4-4161-a798-99c4e4cb8513 | exploring-intra-class-variation-factors-with | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhang_Exploring_Intra-Class_Variation_Factors_With_Learnable_Cluster_Prompts_for_Semi-Supervised_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhang_Exploring_Intra-Class_Variation_Factors_With_Learnable_Cluster_Prompts_for_Semi-Supervised_CVPR_2023_paper.pdf | Exploring Intra-Class Variation Factors With Learnable Cluster Prompts for Semi-Supervised Image Synthesis | Semi-supervised class-conditional image synthesis is typically performed by inferring and injecting class labels into a conditional Generative Adversarial Network (GAN). The supervision in the form of class identity may be inadequate to model classes with diverse visual appearances. In this paper, we propose a Learnable Cluster Prompt-based GAN (LCP-GAN) to capture class-wise characteristics and intra-class variation factors with a broader source of supervision. To exploit partially labeled data, we perform soft partitioning on each class, and explore the possibility of associating intra-class clusters with learnable visual concepts in the feature space of a pre-trained language-vision model, e.g., CLIP. For class-conditional image generation, we design a cluster-conditional generator by injecting a combination of intra-class cluster label embeddings, and further incorporate a real-fake classification head on top of CLIP to distinguish real instances from the synthesized ones, conditioned on the learnable cluster prompts. This significantly strengthens the generator with more semantic language supervision. LCP-GAN not only possesses superior generation capability but also matches the performance of the fully supervised version of the base models: BigGAN and StyleGAN2-ADA, on multiple standard benchmarks. | ['Hau San Wong', 'Si Wu', 'Tianyi Chen', 'Xiaoyang Huo', 'Yunfei Zhang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['conditional-image-generation'] | ['computer-vision'] | [ 5.20251393e-01 3.79093528e-01 -2.16909081e-01 -5.17222703e-01
-8.16205978e-01 -9.54050720e-01 9.38733101e-01 -5.22409081e-01
8.34214017e-02 7.22678900e-01 5.02572954e-03 -7.40847588e-02
4.85888809e-01 -9.57723737e-01 -1.13680089e+00 -1.03012633e+00
4.99863535e-01 4.50942814e-01 -2.28094041e-01 4.38450128e-02
-2.48883620e-01 3.41011256e-01 -1.50117970e+00 7.89219379e-01
8.87015581e-01 9.32484686e-01 6.95909709e-02 5.87015152e-01
-1.36174828e-01 1.03620207e+00 -8.53812635e-01 -5.33270359e-01
4.14578259e-01 -1.03747714e+00 -2.69841015e-01 5.26439011e-01
5.03649294e-01 -1.38497561e-01 -1.12554289e-01 1.21831048e+00
1.94584996e-01 -2.36483052e-01 9.98977840e-01 -1.74501669e+00
-1.00115514e+00 5.07691264e-01 -5.75705469e-01 -5.67079484e-01
2.01928869e-01 4.14566904e-01 7.75849581e-01 -8.40061545e-01
7.18186080e-01 1.34153533e+00 2.96477765e-01 1.12765539e+00
-1.38895214e+00 -1.08404040e+00 1.15619764e-01 -1.78589895e-01
-1.41355264e+00 -3.13221931e-01 1.16496921e+00 -6.63690507e-01
1.39362663e-01 2.75640428e-01 4.79470551e-01 1.68614209e+00
-4.12411429e-02 7.70521164e-01 1.50930619e+00 -5.04088581e-01
4.60272789e-01 7.87542164e-01 -5.87538481e-01 6.14074528e-01
-7.51790255e-02 4.60971296e-01 -3.35748166e-01 3.33851166e-02
6.68344378e-01 1.36616632e-01 -3.13486993e-01 -5.23720145e-01
-1.04066944e+00 1.05197060e+00 6.30540848e-01 7.76307210e-02
-1.37756586e-01 8.59953836e-02 3.13345119e-02 1.10221982e-01
5.31215668e-01 2.52575636e-01 -2.26479948e-01 4.61734146e-01
-1.02549589e+00 -5.86558506e-02 6.02236211e-01 1.37136352e+00
1.05586350e+00 5.16544819e-01 -5.55277705e-01 5.69243550e-01
2.38921657e-01 7.26830900e-01 4.80224431e-01 -7.80946314e-01
3.86267900e-01 6.32537067e-01 -3.00864339e-01 -7.43022978e-01
2.81712025e-01 -5.13990879e-01 -1.04713094e+00 7.54912734e-01
1.58026502e-01 -2.24214569e-01 -1.46769810e+00 1.97255826e+00
2.19054163e-01 3.18858147e-01 2.80903578e-01 7.80679464e-01
6.47924721e-01 8.92767668e-01 1.12995483e-01 -3.05699017e-02
1.08297443e+00 -1.26055908e+00 -6.63910687e-01 -2.27143779e-01
3.20601851e-01 -6.88678920e-01 1.10437012e+00 1.26024604e-01
-6.64504111e-01 -8.72397900e-01 -1.11446261e+00 2.67649621e-01
-5.21707296e-01 4.61693197e-01 3.72885406e-01 8.66434932e-01
-9.75253582e-01 5.02079129e-02 -5.44097304e-01 4.95083816e-02
6.67708695e-01 -3.75918262e-02 -4.65054989e-01 -5.14496624e-01
-1.00257444e+00 3.94768000e-01 5.08531928e-01 -1.38377726e-01
-1.55304205e+00 -5.88921905e-01 -1.06159425e+00 -1.74316406e-01
2.70442307e-01 -4.95624125e-01 4.83561188e-01 -1.83268023e+00
-1.51208091e+00 1.06286550e+00 1.68397859e-01 -2.74224997e-01
6.75874531e-01 3.11356246e-01 -4.79594886e-01 2.15025455e-01
1.97513714e-01 1.29790628e+00 1.66530967e+00 -2.03224134e+00
-3.48151833e-01 -8.99095312e-02 5.24863787e-02 4.95620538e-03
-2.35571429e-01 -2.00033486e-01 -1.81523487e-01 -9.40309703e-01
-3.80994380e-01 -1.22413206e+00 -7.46043250e-02 8.33517984e-02
-8.17744017e-01 6.51615337e-02 1.05381107e+00 -6.38201237e-01
6.61492825e-01 -2.47171998e+00 1.05764702e-01 1.98231727e-01
1.61487535e-01 3.01377922e-01 -3.03247869e-01 1.45381689e-01
-3.08533758e-01 7.19999000e-02 -6.45670652e-01 -5.37161231e-01
7.97026604e-03 4.25694555e-01 -5.78770757e-01 3.12432617e-01
6.93283021e-01 1.12949777e+00 -7.79057741e-01 -3.79584074e-01
2.17805997e-01 5.68577170e-01 -5.77387571e-01 6.08501494e-01
-5.52921593e-01 6.91244245e-01 -2.44661927e-01 7.80380428e-01
7.47101784e-01 -1.53961822e-01 -1.04927709e-02 -1.77981451e-01
4.17448699e-01 -4.18880165e-01 -8.48352790e-01 1.66566336e+00
-4.37757611e-01 5.20477295e-01 -8.35592300e-02 -1.02421260e+00
1.05786943e+00 3.00318837e-01 -4.89302725e-02 -5.00157654e-01
-1.15299458e-02 2.13667229e-02 -9.66165587e-02 -1.32568479e-01
4.32492495e-02 -3.82220596e-01 -1.00306973e-01 2.81435519e-01
4.12335366e-01 -2.14491695e-01 -1.88123837e-01 3.22346121e-01
7.66619921e-01 3.75691056e-01 -1.12313606e-01 -1.32899523e-01
5.00395715e-01 -2.14415222e-01 5.39697111e-01 5.63326776e-01
2.46710684e-02 1.05843914e+00 4.58094388e-01 -5.52312247e-02
-1.10567856e+00 -1.23262012e+00 -2.19618645e-03 9.32441890e-01
-3.83596644e-02 -9.71581638e-02 -1.04063332e+00 -1.12543643e+00
-6.03985898e-02 8.56983244e-01 -9.84538674e-01 -4.75461662e-01
-2.36826718e-01 -5.97352207e-01 6.37965202e-01 5.85835397e-01
4.12171513e-01 -1.32240808e+00 -7.93658495e-02 -6.76531270e-02
1.63228065e-02 -1.27466893e+00 -4.29604471e-01 3.69885027e-01
-2.68561602e-01 -8.94368052e-01 -7.14998603e-01 -9.76747394e-01
1.18187034e+00 -2.03659073e-01 1.03564632e+00 -2.48579711e-01
-2.09016174e-01 4.39122081e-01 -6.10573769e-01 -2.60073483e-01
-8.83753717e-01 -1.98941752e-01 -1.90308198e-01 7.89189100e-01
2.54468191e-02 -4.29559052e-01 -4.20988262e-01 1.12181239e-01
-1.20253837e+00 4.05555576e-01 7.21110642e-01 1.06216168e+00
5.94087780e-01 1.59830391e-01 5.28375804e-01 -1.30736852e+00
7.32355267e-02 -5.32508135e-01 -3.39655191e-01 2.81656712e-01
-4.70091790e-01 -5.56620620e-02 9.44179475e-01 -7.19416320e-01
-1.22045076e+00 3.15511674e-01 1.22144632e-01 -9.68914628e-01
-4.79591370e-01 9.37583148e-02 -7.98128188e-01 -2.89228540e-02
6.93560362e-01 4.83159959e-01 -3.73775959e-02 -8.21415707e-02
8.49619448e-01 5.40134549e-01 6.36445224e-01 -7.15936124e-01
1.30006611e+00 5.88216782e-01 -2.34115198e-01 -4.87884253e-01
-7.39220560e-01 4.28373814e-02 -5.12521148e-01 -1.10931255e-01
1.29153788e+00 -1.32254183e+00 1.94225147e-01 5.66368639e-01
-1.08508813e+00 -6.46098077e-01 -6.29294157e-01 1.73129782e-01
-6.18997931e-01 -1.88490793e-01 -5.32971025e-01 -5.58463991e-01
5.81000298e-02 -1.09146512e+00 1.18870366e+00 1.64945707e-01
2.11250514e-01 -1.05573463e+00 -2.26686835e-01 4.71325070e-01
1.70833394e-01 8.25756907e-01 8.94449949e-01 -6.12080574e-01
-7.37100124e-01 -3.33277166e-01 -3.30021620e-01 9.97184217e-01
2.67911524e-01 1.92501899e-02 -1.29535449e+00 -4.80994970e-01
-1.65800285e-02 -5.96184909e-01 8.15590620e-01 8.44252333e-02
1.27860236e+00 -5.09416401e-01 -2.55788565e-01 8.00744355e-01
1.53658128e+00 2.08555415e-01 7.84082890e-01 -2.53601313e-01
1.21756446e+00 6.08236790e-01 3.35028529e-01 1.78164929e-01
1.81574509e-01 5.40657878e-01 5.31680405e-01 -4.51970100e-01
-6.25363529e-01 -7.51918554e-01 5.99083006e-01 6.95019186e-01
1.96858421e-01 -6.45316124e-01 -4.35159385e-01 4.20772076e-01
-1.42367101e+00 -7.72191823e-01 3.31347585e-01 1.81882441e+00
1.03465426e+00 5.83151504e-02 -1.75297633e-01 3.75468396e-02
9.94862318e-01 2.02099875e-01 -5.27352512e-01 -1.01412043e-01
-4.44678962e-01 3.63624364e-01 4.02716815e-01 4.37801480e-01
-1.01663864e+00 1.04888165e+00 5.09562492e+00 1.22614002e+00
-1.19526899e+00 2.48190239e-01 9.73080099e-01 1.89795032e-01
-5.17798901e-01 1.18453093e-01 -6.68633044e-01 7.55919039e-01
6.82397664e-01 2.79229522e-01 5.47392011e-01 9.83965755e-01
-2.80670911e-01 3.57986063e-01 -1.28012347e+00 8.14612985e-01
6.19007170e-01 -1.14326262e+00 5.35649300e-01 3.20499718e-01
1.26277781e+00 -4.73399699e-01 3.53669465e-01 3.15328747e-01
4.42015141e-01 -1.18557513e+00 8.90895426e-01 3.57309759e-01
1.41893494e+00 -6.95330322e-01 4.73992258e-01 3.63733679e-01
-8.74579847e-01 -3.27804126e-02 -1.88986972e-01 2.76872009e-01
-6.82614446e-02 4.74448711e-01 -7.29508221e-01 4.72231746e-01
3.51594210e-01 8.26265812e-01 -7.90635467e-01 3.05314094e-01
-6.74000800e-01 9.46917892e-01 7.10925087e-02 4.69035357e-01
2.73120314e-01 -2.85401374e-01 2.80414760e-01 1.15904927e+00
2.67328709e-01 -1.51092485e-01 3.24630052e-01 1.29738975e+00
-2.78072178e-01 -3.75006318e-01 -6.38362825e-01 -2.23324925e-01
3.35696965e-01 1.30426586e+00 -6.45369828e-01 -6.12051249e-01
-3.30456257e-01 1.42323899e+00 1.94508806e-01 4.76076871e-01
-9.94696259e-01 -2.66561478e-01 4.04813975e-01 1.87293887e-01
3.68781388e-01 1.43158078e-01 -1.92177281e-01 -1.21900737e+00
1.98253477e-03 -1.11905885e+00 1.55723289e-01 -9.23297465e-01
-1.34567440e+00 7.80019939e-01 -2.47462213e-01 -1.44132733e+00
-2.81680256e-01 -6.10153913e-01 -7.80258954e-01 7.19133079e-01
-1.48279548e+00 -1.86337829e+00 -4.28441674e-01 9.85367000e-01
5.39529264e-01 -5.53996265e-01 9.89576578e-01 3.39436322e-01
-3.52492571e-01 1.07644546e+00 -8.76416862e-02 4.34361786e-01
8.24830115e-01 -1.29668117e+00 -2.34200917e-02 9.72130775e-01
3.54464591e-01 2.19370246e-01 4.77061272e-01 -5.26329398e-01
-1.09343326e+00 -1.73041761e+00 3.75208914e-01 -5.39601028e-01
3.50157112e-01 -9.08386171e-01 -7.90786326e-01 8.47326219e-01
3.87082398e-01 2.95170605e-01 6.61602020e-01 -5.65131962e-01
-7.24080324e-01 -1.96187779e-01 -1.35348439e+00 3.62127930e-01
8.78127813e-01 -8.35120857e-01 -1.63323700e-01 4.73621815e-01
1.13787854e+00 -2.56019533e-01 -4.95414346e-01 3.92688841e-01
1.68857470e-01 -7.34246850e-01 9.09332454e-01 -4.61458057e-01
9.19170320e-01 -5.78683794e-01 -2.83039510e-01 -1.47566044e+00
-2.33272016e-01 -3.33997875e-01 1.08958907e-01 1.76445341e+00
3.56230110e-01 -5.65640330e-01 8.45485032e-01 3.26616108e-01
-2.36875728e-01 -3.25800031e-01 -6.63402379e-01 -7.35743046e-01
3.07098120e-01 -8.54633525e-02 6.28113031e-01 1.33671868e+00
-6.51056826e-01 1.67922065e-01 -6.73055887e-01 2.27783144e-01
7.85852373e-01 1.01982780e-01 8.50407004e-01 -6.60054624e-01
-5.55368900e-01 -1.22852892e-01 -5.61091006e-01 -5.46086490e-01
5.65530241e-01 -1.10546577e+00 8.99424106e-02 -9.54197168e-01
1.04615547e-01 -7.28828371e-01 -4.28915799e-01 6.07640445e-01
-1.21940680e-01 7.35283732e-01 2.92982876e-01 3.88928615e-02
-3.61629009e-01 7.53606856e-01 1.46008730e+00 -5.45150340e-01
1.33260265e-01 -1.74237624e-01 -7.53429770e-01 5.13095737e-01
6.58664048e-01 -6.56033158e-01 -6.33309484e-01 -2.67106503e-01
-9.82149020e-02 -3.46916392e-02 6.19565308e-01 -1.10034668e+00
2.23994050e-02 -1.59346312e-01 6.45878673e-01 -2.46849597e-01
4.03415084e-01 -9.74214971e-01 3.32429826e-01 2.03506649e-01
-5.13097465e-01 -5.26480436e-01 -8.22770447e-02 7.31167197e-01
-4.52691525e-01 1.53642250e-02 1.00102961e+00 -2.27240413e-01
-5.19246340e-01 4.50375766e-01 -2.36608479e-02 1.18472502e-01
1.24630702e+00 -1.41893476e-01 -2.89205760e-01 -3.39497507e-01
-6.41225040e-01 -7.27167800e-02 6.65298641e-01 6.17717445e-01
5.77486277e-01 -1.72713876e+00 -7.99374759e-01 6.42489791e-01
5.15999258e-01 1.59087062e-01 2.07282245e-01 2.60912478e-02
-2.54424632e-01 -1.11209918e-02 -3.34319681e-01 -6.92857385e-01
-9.01598811e-01 9.61283624e-01 1.33961335e-01 -1.01044970e-02
-2.58501738e-01 9.57188487e-01 8.90318632e-01 -7.26768136e-01
-9.52578895e-03 5.29092923e-02 7.98874944e-02 -7.04710409e-02
2.68548131e-01 -2.64177084e-01 -3.64956111e-01 -8.08089674e-01
-7.57588893e-02 4.23930466e-01 1.27575934e-01 1.53847039e-02
9.06630993e-01 1.38261512e-01 -1.40654683e-01 3.85410786e-01
1.45238888e+00 7.33284354e-02 -1.62730432e+00 -2.13044778e-01
-5.83112121e-01 -4.06285644e-01 -2.66608953e-01 -1.00590599e+00
-1.47822285e+00 1.02874041e+00 7.04645872e-01 -1.30763233e-01
1.16176033e+00 1.15608320e-01 5.14289618e-01 -2.80653477e-01
2.30757579e-01 -7.98001587e-01 3.65941495e-01 -1.42922282e-01
9.01763320e-01 -1.36128199e+00 -3.34818095e-01 -4.92091775e-01
-9.24577057e-01 7.26398647e-01 8.60658348e-01 -2.73727357e-01
6.16385341e-01 1.43929973e-01 2.62004256e-01 4.96812277e-02
-5.51010013e-01 4.65552416e-03 3.50495338e-01 9.86125708e-01
4.92814463e-03 6.34944439e-01 3.25510710e-01 5.27494609e-01
-2.45404869e-01 -3.55027527e-01 4.45290625e-01 6.33841097e-01
3.88396271e-02 -1.16131270e+00 -3.28234702e-01 2.18233138e-01
-1.02237500e-01 -1.55070812e-01 -4.75764006e-01 5.08991122e-01
7.54329801e-01 9.04316068e-01 1.48227587e-01 -6.80205464e-01
-1.27597183e-01 1.63710237e-01 4.44525689e-01 -7.69314647e-01
-3.26400518e-01 -6.46009073e-02 -2.95990348e-01 -2.64594108e-01
-4.56858695e-01 -4.19438273e-01 -8.39005530e-01 6.34340346e-02
-2.44953111e-01 3.34984809e-02 5.96478224e-01 8.17180932e-01
2.83831209e-01 6.59908056e-01 1.07154202e+00 -7.96066463e-01
-2.72317588e-01 -7.82912076e-01 -6.90174103e-01 8.30718517e-01
3.71269166e-01 -5.59515655e-01 -6.66903496e-01 5.70835233e-01] | [11.599396705627441, -0.23775401711463928] |
24d14c17-7450-4c59-b62c-594c71d4b56e | rule-augmented-unsupervised-constituency | 2105.10193 | null | https://arxiv.org/abs/2105.10193v1 | https://arxiv.org/pdf/2105.10193v1.pdf | Rule Augmented Unsupervised Constituency Parsing | Recently, unsupervised parsing of syntactic trees has gained considerable attention. A prototypical approach to such unsupervised parsing employs reinforcement learning and auto-encoders. However, no mechanism ensures that the learnt model leverages the well-understood language grammar. We propose an approach that utilizes very generic linguistic knowledge of the language present in the form of syntactic rules, thus inducing better syntactic structures. We introduce a novel formulation that takes advantage of the syntactic grammar rules and is independent of the base system. We achieve new state-of-the-art results on two benchmarks datasets, MNLI and WSJ. The source code of the paper is available at https://github.com/anshuln/Diora_with_rules. | ['Rishabh Iyer', 'Ganesh Ramakrishnan', 'Ayush Maheshwari', 'Anshul Nasery', 'Atul Sahay'] | 2021-05-21 | null | https://aclanthology.org/2021.findings-acl.436 | https://aclanthology.org/2021.findings-acl.436.pdf | findings-acl-2021-8 | ['constituency-parsing'] | ['natural-language-processing'] | [-1.43994940e-02 7.25981534e-01 -4.07650232e-01 -7.12913752e-01
-7.45779812e-01 -8.02127957e-01 3.44389260e-01 1.09068811e-01
-2.83757597e-01 6.88131571e-01 3.71293694e-01 -6.50843203e-01
1.00946926e-01 -1.00687718e+00 -8.88784170e-01 -3.81874532e-01
-1.33171873e-02 2.86220640e-01 1.73441634e-01 -2.85097957e-01
-4.44030687e-02 2.56736390e-02 -1.42799854e+00 4.53959882e-01
7.01542139e-01 6.03940845e-01 3.51433963e-01 8.81481826e-01
-6.28212154e-01 1.29745114e+00 -4.04030502e-01 -5.88288188e-01
1.58654422e-01 -5.12750566e-01 -1.08342898e+00 -5.70328645e-02
1.33248597e-01 -1.29209682e-01 -2.24091202e-01 1.18782556e+00
-4.14618216e-02 -9.96560976e-02 8.46676826e-02 -7.60811031e-01
-8.02982807e-01 1.43881857e+00 -6.34757131e-02 2.52240121e-01
2.64957190e-01 1.59212630e-02 1.57670224e+00 -4.98931825e-01
7.88768530e-01 1.29230440e+00 3.07608515e-01 8.17583621e-01
-1.13496625e+00 -4.67901886e-01 3.64391059e-01 -7.05742016e-02
-7.66668677e-01 -4.09769058e-01 9.91375327e-01 -1.86226189e-01
1.27511537e+00 -8.43744501e-02 2.44786933e-01 1.05755877e+00
2.16779754e-01 9.85758066e-01 1.26138043e+00 -1.01167893e+00
2.55377829e-01 -1.01906575e-01 6.21087193e-01 9.42336559e-01
1.94830701e-01 3.92801613e-01 -4.32118595e-01 6.17917031e-02
4.05741841e-01 -3.57661635e-01 2.26004750e-01 -2.96197265e-01
-7.29624510e-01 1.12471843e+00 1.87253177e-01 6.25255942e-01
-1.35073125e-01 2.33141512e-01 5.62052965e-01 3.81178945e-01
3.83676201e-01 4.31756437e-01 -9.41341579e-01 -2.84831166e-01
-5.53186476e-01 5.92974424e-02 1.09953463e+00 1.03257394e+00
9.73713636e-01 2.84699947e-02 4.42110687e-01 7.60876596e-01
3.73765767e-01 3.79904300e-01 4.83146727e-01 -1.29137945e+00
4.01944607e-01 5.91566682e-01 -4.09583509e-01 -4.23606724e-01
-3.58013719e-01 -1.11325331e-01 -2.61584729e-01 7.68343881e-02
5.84064603e-01 -3.90480518e-01 -7.81581759e-01 2.00906110e+00
3.42129856e-01 -2.36869022e-01 4.37280625e-01 4.11155134e-01
6.91280663e-01 6.32704675e-01 3.65177244e-01 -1.61678195e-01
1.27714229e+00 -8.90532434e-01 -6.80323124e-01 -4.98366803e-01
8.07716846e-01 -4.25064832e-01 1.02475274e+00 2.77554393e-01
-1.14790916e+00 -4.01413530e-01 -8.16530585e-01 -1.92710236e-01
-5.20430446e-01 -5.28205112e-02 1.15755546e+00 6.31778538e-01
-1.21633780e+00 5.47598600e-01 -9.48489785e-01 -4.61496711e-01
2.66663611e-01 4.03934270e-01 -3.80498052e-01 1.06619373e-01
-1.13073611e+00 5.76413453e-01 7.90614069e-01 -1.19159244e-01
-5.89296579e-01 -2.18724161e-01 -1.17913294e+00 -1.22274511e-01
7.19650269e-01 -5.07052898e-01 1.95934570e+00 -1.27907169e+00
-1.90029645e+00 9.92356479e-01 -2.21977532e-01 -6.04437113e-01
6.36039302e-02 -5.12872577e-01 -2.11313382e-01 1.41318470e-01
1.36387303e-01 5.63413560e-01 6.55367136e-01 -1.15167725e+00
-7.83578932e-01 -4.16743815e-01 5.15863299e-01 -1.61199942e-01
-1.06887966e-01 4.73479569e-01 -1.96780011e-01 -5.10121047e-01
-7.58561790e-02 -8.06545973e-01 -2.88045526e-01 -7.08538592e-01
-5.83382845e-01 -4.41570461e-01 5.26001453e-01 -6.56432509e-01
1.23686290e+00 -2.10006332e+00 1.16745681e-01 -5.08345999e-02
1.15066148e-01 4.35714200e-02 -1.00481562e-01 6.17410839e-01
-1.47175550e-01 3.44500870e-01 -5.77482939e-01 -3.48716795e-01
1.04937956e-01 8.35182786e-01 -3.17727715e-01 -6.21597357e-02
3.70227188e-01 9.24288094e-01 -1.05087566e+00 -5.28296292e-01
1.28689796e-01 1.59783307e-02 -8.14160049e-01 4.76555735e-01
-5.68441868e-01 4.07611459e-01 -7.23603904e-01 5.41779876e-01
2.94161797e-01 -3.75029482e-02 7.45990276e-01 1.42675117e-01
-3.08560967e-01 8.43818724e-01 -8.63020539e-01 1.80479538e+00
-5.01295507e-01 1.34084910e-01 8.45368654e-02 -1.16909373e+00
7.42385805e-01 2.21326664e-01 2.96682388e-01 -6.75009727e-01
2.27732345e-01 2.20450997e-01 3.04148883e-01 -4.43609565e-01
1.65608048e-01 -6.45547211e-02 -4.42367554e-01 4.56064284e-01
5.24468303e-01 7.04653561e-02 6.65012300e-01 3.27814430e-01
1.31198514e+00 6.42547965e-01 6.74649596e-01 -5.79889894e-01
5.28183341e-01 7.83536807e-02 7.36178398e-01 9.40137446e-01
-1.19598918e-01 -6.21392839e-02 6.61043465e-01 -4.80548322e-01
-9.03304636e-01 -9.81526077e-01 3.46333422e-02 1.60190785e+00
-4.93408442e-01 -6.66671157e-01 -9.97764647e-01 -9.90955591e-01
-1.51364341e-01 9.36424017e-01 -6.14091456e-01 8.61161277e-02
-1.08007956e+00 -4.19018835e-01 6.57017887e-01 7.26111352e-01
2.48877063e-01 -1.54136670e+00 -7.84578562e-01 3.44550163e-01
1.96716264e-01 -1.42550850e+00 1.93167676e-03 6.92154646e-01
-1.07041037e+00 -1.04541051e+00 8.80260244e-02 -9.96592939e-01
5.72276711e-01 -3.21360588e-01 1.44587672e+00 2.71915317e-01
-3.93898301e-02 4.67125297e-01 -7.36622989e-01 -3.86608988e-01
-1.00723040e+00 3.38568866e-01 -2.58052200e-01 -3.02499592e-01
5.06459117e-01 -5.24351895e-01 -1.45523474e-02 -3.60446364e-01
-7.90821075e-01 -4.83636046e-04 4.74513382e-01 8.00675631e-01
5.05606234e-01 -1.27687216e-01 4.43614990e-01 -1.68125165e+00
3.28746378e-01 -3.72432947e-01 -7.30423391e-01 9.20284167e-02
-4.32687610e-01 5.42007267e-01 9.46243405e-01 1.08043656e-01
-1.31826341e+00 2.92600304e-01 -4.59209323e-01 4.05567735e-02
-6.57055497e-01 5.60775936e-01 -3.88642788e-01 4.21192110e-01
3.60161036e-01 1.09342411e-01 -1.98805198e-01 -7.02218115e-01
7.32166231e-01 4.12194788e-01 5.01159370e-01 -1.07894266e+00
6.72269404e-01 2.78553545e-01 -3.89926940e-01 -5.84792972e-01
-1.09998190e+00 -2.27189973e-01 -8.49944472e-01 2.32340395e-01
8.54942083e-01 -6.36093199e-01 -3.33577275e-01 2.14690626e-01
-1.28745306e+00 -7.70097733e-01 -5.30985355e-01 2.19111845e-01
-7.94704735e-01 4.53019202e-01 -1.09205616e+00 -7.39302456e-01
-3.55558425e-01 -1.02768540e+00 8.98693740e-01 1.12007953e-01
-2.56425828e-01 -1.04727530e+00 2.40978241e-01 7.28937760e-02
2.00968549e-01 -1.00580044e-01 1.24874938e+00 -9.30678844e-01
-5.10264695e-01 2.52529472e-01 9.73003060e-02 5.04927814e-01
2.29971573e-01 1.76473826e-01 -1.16899943e+00 1.03275895e-01
2.50393269e-03 -4.54804063e-01 7.20746875e-01 3.56238365e-01
1.08789301e+00 -4.03854936e-01 5.69118485e-02 4.59435672e-01
1.45263433e+00 3.08093131e-01 4.10272479e-01 5.24688721e-01
5.72122157e-01 8.78005087e-01 3.57441455e-01 9.72419530e-02
6.68532491e-01 2.20637009e-01 3.42037618e-01 1.95725575e-01
-4.47383784e-02 -3.48077774e-01 6.55645251e-01 1.20736814e+00
-5.63935982e-03 -1.87990218e-02 -1.06669104e+00 5.34581542e-01
-1.76695669e+00 -8.01674843e-01 2.55387455e-01 1.63899732e+00
1.03630257e+00 4.76084828e-01 1.06327601e-01 8.81005675e-02
5.55469692e-01 2.70378321e-01 -2.64681488e-01 -1.04690206e+00
-6.78556710e-02 7.11233497e-01 4.60732073e-01 8.64809453e-01
-1.22481382e+00 1.55339968e+00 6.41251993e+00 3.13014418e-01
-8.23757589e-01 1.51104704e-01 5.04601657e-01 4.68272537e-01
-2.80302137e-01 4.51776981e-01 -9.92791474e-01 1.81626722e-01
1.34409511e+00 -8.32800046e-02 4.88445669e-01 1.01300907e+00
4.18885276e-02 3.98659445e-02 -1.27704060e+00 2.24594116e-01
-3.01332861e-01 -1.23346651e+00 -5.52971959e-02 -1.11134812e-01
1.99167415e-01 3.38837326e-01 -3.18334967e-01 6.25875533e-01
1.02164316e+00 -6.45971954e-01 7.42862046e-01 1.23907402e-01
4.87006575e-01 -6.58860862e-01 4.82099771e-01 4.77745503e-01
-1.06836271e+00 -1.97145268e-01 -2.98614711e-01 -3.28776896e-01
-1.08591430e-02 6.72559083e-01 -5.08173883e-01 6.30174339e-01
7.22946405e-01 7.57882714e-01 -6.24444783e-01 2.63247102e-01
-9.19603944e-01 1.08843303e+00 -3.03903073e-01 -2.37206165e-02
4.03938115e-01 -9.99913067e-02 4.21794772e-01 1.57754195e+00
-2.01356947e-01 3.79898548e-02 5.25998831e-01 7.61350572e-01
-1.63954318e-01 3.53739113e-01 -8.76356542e-01 -1.52116939e-01
3.16532701e-01 1.15671241e+00 -5.97345293e-01 -5.13102472e-01
-8.04799616e-01 4.93176997e-01 8.86421025e-01 1.05794959e-01
-4.98900563e-01 -2.37595793e-02 4.58841294e-01 -3.12100202e-01
5.07512808e-01 -4.00853902e-01 -1.90898314e-01 -1.24583244e+00
2.02451840e-01 -1.17207932e+00 7.08877385e-01 -3.30047041e-01
-1.26239240e+00 7.29131222e-01 1.74320474e-01 -5.27078092e-01
-6.54038548e-01 -1.08255172e+00 -5.52071393e-01 4.28548545e-01
-1.80274594e+00 -1.03196180e+00 2.33735502e-01 3.37025225e-01
7.98366070e-01 -1.51953712e-01 1.20369387e+00 -8.72472301e-02
-6.52326584e-01 4.46182966e-01 -1.09559953e-01 5.83752036e-01
2.55635142e-01 -1.63653886e+00 6.21781528e-01 1.05068433e+00
5.19104481e-01 6.42438769e-01 5.44851363e-01 -5.09836495e-01
-1.66749644e+00 -9.52576995e-01 1.18644929e+00 -6.20254934e-01
1.07308972e+00 -4.83730704e-01 -9.20261025e-01 1.02461803e+00
6.62335813e-01 -1.07989378e-01 8.13595176e-01 3.23806554e-01
-6.08036995e-01 1.65747553e-02 -7.70218492e-01 4.10000414e-01
1.13922083e+00 -5.79559863e-01 -9.67142701e-01 1.44698352e-01
8.14818621e-01 -3.57066214e-01 -8.50305200e-01 3.39973927e-01
3.03955495e-01 -9.24051166e-01 5.50789654e-01 -8.81571412e-01
7.61924982e-01 5.89810312e-02 -3.92914087e-01 -1.12493932e+00
-3.38966310e-01 -5.66056132e-01 -3.06086570e-01 1.21353996e+00
7.46372879e-01 -7.35765040e-01 8.26912999e-01 4.54495966e-01
-2.35028252e-01 -6.72184348e-01 -6.58316374e-01 -8.38317037e-01
4.96999383e-01 -7.10928380e-01 4.10565823e-01 7.72306204e-01
2.86844105e-01 5.11377394e-01 9.02518108e-02 1.76244870e-01
6.28386617e-01 3.80521983e-01 6.54594243e-01 -9.85006869e-01
-7.31344700e-01 -2.77543873e-01 -1.16263017e-01 -8.18101823e-01
7.12648988e-01 -1.20993316e+00 2.12609649e-01 -1.30407667e+00
-6.62215874e-02 -2.94893593e-01 -5.15028179e-01 8.70329618e-01
-1.95302535e-02 -3.29858750e-01 3.44131917e-01 -7.89142251e-02
-6.73190415e-01 2.32982755e-01 7.89561510e-01 9.10086706e-02
-2.72005126e-02 -4.10323441e-01 -9.66526687e-01 1.10755301e+00
1.38679206e+00 -7.18072116e-01 -2.54814863e-01 -8.22174370e-01
1.23013705e-01 -1.76951915e-01 -3.63735408e-02 -7.36588836e-01
1.53636903e-01 -1.44984722e-01 -1.15405165e-01 -1.20709702e-01
-2.55137950e-01 -5.69837272e-01 -3.68860692e-01 4.77624238e-01
-5.17936349e-01 3.58731449e-01 3.27092260e-01 2.90471256e-01
-2.66904533e-01 -4.68710899e-01 6.67157948e-01 -5.70700347e-01
-1.01481020e+00 -7.19477683e-02 -4.69783753e-01 3.34350735e-01
6.16721690e-01 2.07552850e-01 -2.49168471e-01 -8.18251073e-02
-8.30572665e-01 5.27089797e-02 2.66997635e-01 5.38639069e-01
2.99072266e-01 -7.50130594e-01 -7.38335788e-01 2.23946735e-01
6.30270690e-02 -7.19378889e-02 -3.39935631e-01 2.51831263e-01
-4.22936052e-01 3.83490145e-01 -7.17148334e-02 -3.88105780e-01
-8.71898115e-01 6.94192052e-01 2.83770949e-01 -6.57044590e-01
-8.29602063e-01 6.76003158e-01 3.37211043e-01 -7.54580140e-01
-4.89617418e-03 -5.90734243e-01 -8.48984197e-02 -4.01768416e-01
4.07213390e-01 -2.83912569e-01 -4.06602584e-02 -5.11653900e-01
-3.89389187e-01 3.63805652e-01 -3.72222453e-01 4.22740653e-02
1.37985170e+00 -5.47994599e-02 -2.66357481e-01 5.11034369e-01
9.77329791e-01 2.44890273e-01 -1.03161585e+00 -4.84751225e-01
8.23352218e-01 2.14343946e-02 -2.29613394e-01 -7.32951701e-01
-1.01702213e+00 7.68538594e-01 1.11262463e-01 2.58211792e-01
1.13743412e+00 3.26103061e-01 6.46441042e-01 7.66598105e-01
6.19761825e-01 -1.14135742e+00 -2.11011052e-01 9.78159189e-01
5.04017770e-01 -1.21708786e+00 -3.57940584e-01 -7.48771608e-01
-5.11128783e-01 1.17019212e+00 5.16518116e-01 -6.09964967e-01
7.68984437e-01 8.28216374e-01 3.58225912e-01 -1.78183079e-01
-9.57187891e-01 -4.59038079e-01 -2.75438160e-01 4.30887401e-01
7.85509408e-01 3.28217208e-01 -3.69773149e-01 1.02416265e+00
-4.90877062e-01 -8.74220580e-02 4.03879583e-01 1.33910155e+00
-4.87589359e-01 -1.81217706e+00 -2.71228347e-02 1.12269953e-01
-7.91961432e-01 -3.90920192e-01 -5.49455464e-01 9.61081386e-01
8.01757164e-03 8.83799493e-01 -2.14740649e-01 -5.50596938e-02
3.63782883e-01 6.08590305e-01 5.91743231e-01 -1.14121532e+00
-6.83073342e-01 8.84194523e-02 4.17519212e-01 -7.86352396e-01
-6.51345432e-01 -5.92741072e-01 -1.59192836e+00 6.67599589e-02
-4.72334102e-02 2.52616316e-01 4.39821512e-01 9.88667488e-01
2.40955785e-01 4.28851187e-01 6.58873022e-01 -3.68872017e-01
-5.48878074e-01 -7.64877558e-01 -2.73760259e-01 4.26344216e-01
2.44352198e-03 -3.70300829e-01 -7.80953765e-02 3.48221749e-01] | [10.38149642944336, 9.476972579956055] |
3ca47812-96cc-445e-abdb-90eb65e133ce | accurate-and-scalable-version-identification | 1910.12551 | null | https://arxiv.org/abs/1910.12551v2 | https://arxiv.org/pdf/1910.12551v2.pdf | Accurate and Scalable Version Identification Using Musically-Motivated Embeddings | The version identification (VI) task deals with the automatic detection of recordings that correspond to the same underlying musical piece. Despite many efforts, VI is still an open problem, with much room for improvement, specially with regard to combining accuracy and scalability. In this paper, we present MOVE, a musically-motivated method for accurate and scalable version identification. MOVE achieves state-of-the-art performance on two publicly-available benchmark sets by learning scalable embeddings in an Euclidean distance space, using a triplet loss and a hard triplet mining strategy. It improves over previous work by employing an alternative input representation, and introducing a novel technique for temporal content summarization, a standardized latent space, and a data augmentation strategy specifically designed for VI. In addition to the main results, we perform an ablation study to highlight the importance of our design choices, and study the relation between embedding dimensionality and model performance. | ['Emilia Gómez', 'Joan Serrà', 'Furkan Yesiler'] | 2019-10-28 | null | null | null | null | ['cover-song-identification'] | ['music'] | [ 5.12678087e-01 -2.00021252e-01 -1.34996086e-01 1.64914921e-01
-1.13195658e+00 -1.06175876e+00 5.26317418e-01 3.58128011e-01
-3.69509846e-01 1.90569937e-01 7.28126705e-01 2.00803354e-01
-4.74891663e-01 -2.52919078e-01 -2.94587374e-01 -6.95156932e-01
-2.13395670e-01 2.65802205e-01 5.22549683e-03 -6.93196878e-02
4.84315276e-01 2.04460904e-01 -1.66226852e+00 2.78708518e-01
4.11584258e-01 1.00932205e+00 -1.56491697e-01 7.76460290e-01
2.29084909e-01 4.47676599e-01 -7.32833028e-01 -7.08253980e-01
2.25108802e-01 -5.51490545e-01 -7.67173946e-01 9.15369019e-02
6.40845895e-01 1.87009811e-01 -5.98377325e-02 5.51012278e-01
8.49229455e-01 1.30285949e-01 7.09221482e-01 -1.20683730e+00
-4.46106076e-01 6.99632704e-01 -3.72981697e-01 2.95639616e-02
5.64335108e-01 -1.51508674e-01 1.62951338e+00 -6.16879582e-01
6.36309922e-01 6.78986609e-01 1.15668166e+00 3.16133767e-01
-1.67452800e+00 -3.57926250e-01 -1.45184726e-01 2.14594796e-01
-1.44752634e+00 -5.27547002e-01 1.01315618e+00 -5.52532554e-01
1.01697779e+00 6.90636933e-01 7.54561424e-01 1.33981371e+00
-1.45276010e-01 8.54448259e-01 8.33327174e-01 -4.74192560e-01
2.85800308e-01 -3.78086441e-03 8.30588415e-02 3.27309936e-01
1.72533646e-01 -1.24712363e-01 -9.67179418e-01 -5.69287658e-01
3.99460167e-01 -4.38034892e-01 -3.05735886e-01 -8.18230331e-01
-1.36118090e+00 8.07181776e-01 -5.64812012e-02 2.89395452e-01
-3.14174175e-01 -8.63561872e-03 8.08646083e-01 3.09766084e-01
4.48912740e-01 1.04837716e+00 -4.78410125e-01 -7.11149752e-01
-1.45608556e+00 5.04524827e-01 8.96101534e-01 5.80075920e-01
1.81212291e-01 -1.23105593e-01 -4.37073052e-01 8.91714454e-01
-1.72163948e-01 -1.33589745e-01 6.49474025e-01 -1.08328652e+00
5.26788771e-01 6.44138098e-01 -9.40649062e-02 -8.60077441e-01
-3.09314102e-01 -7.75542557e-01 -6.02689505e-01 6.21910729e-02
2.64422029e-01 1.07516341e-01 -2.27356046e-01 1.81481528e+00
-8.30109231e-03 1.88171431e-01 -4.97091189e-02 6.49461627e-01
5.53062797e-01 1.72934502e-01 -4.00767446e-01 -4.23032373e-01
1.28607535e+00 -8.97664368e-01 -5.18765628e-01 2.99609788e-02
5.99283338e-01 -9.80923116e-01 1.31959248e+00 7.46093392e-01
-1.08742082e+00 -2.48881176e-01 -1.21303177e+00 -1.70345247e-01
1.39929140e-02 4.58121479e-01 5.51580191e-01 6.69789970e-01
-7.86393583e-01 1.04055429e+00 -6.31578505e-01 -4.39235091e-01
2.59374082e-01 2.19795987e-01 -4.18285400e-01 4.43802148e-01
-7.10634470e-01 4.26196426e-01 3.02174836e-01 -2.71652848e-01
-3.62829477e-01 -7.93252707e-01 -6.75622106e-01 1.40894726e-01
3.01157773e-01 -7.94639051e-01 1.10985756e+00 -6.88017845e-01
-1.54408813e+00 8.93977046e-01 -1.63278878e-02 -6.02490544e-01
6.26563489e-01 -3.92532945e-01 -1.71720892e-01 2.02477559e-01
7.59436414e-02 1.91225648e-01 9.08042312e-01 -8.21204185e-01
-2.38001719e-01 -3.17744315e-01 -2.53845543e-01 7.77062103e-02
-8.45217407e-01 1.54315859e-01 -6.23659194e-01 -1.21365225e+00
1.52190179e-01 -1.06255424e+00 8.92404318e-02 -2.58318126e-01
-5.46250641e-01 -2.84667760e-01 3.96152496e-01 -7.78984904e-01
1.86122370e+00 -2.43989229e+00 7.80182064e-01 4.97598015e-02
9.82261226e-02 7.02254623e-02 -1.81006819e-01 7.95344055e-01
-1.49682865e-01 1.55231088e-01 -5.14269352e-01 -7.99459934e-01
2.67668307e-01 -1.72913760e-01 -4.78235662e-01 4.15924370e-01
5.11192121e-02 7.25113571e-01 -6.09345078e-01 -4.46813405e-01
-7.46883154e-02 2.57778198e-01 -6.44199371e-01 -9.12285410e-03
3.56085524e-02 9.79293212e-02 1.73893310e-02 4.71890926e-01
1.30222589e-01 -1.63421798e-02 2.28087276e-01 -4.78878438e-01
-1.36463672e-01 4.48970377e-01 -1.40370619e+00 2.14327264e+00
-1.68689042e-01 7.56169438e-01 -1.14783555e-01 -8.05703044e-01
8.28047156e-01 3.62910122e-01 7.23034501e-01 -2.55870610e-01
8.92039910e-02 1.00496739e-01 -2.04660326e-01 -4.72940564e-01
8.59418154e-01 3.52505893e-02 -4.10882920e-01 5.67902446e-01
7.95703009e-02 6.36246502e-02 2.53181845e-01 1.94739506e-01
1.31951904e+00 1.93817556e-01 3.42594683e-01 4.87388819e-02
6.08061515e-02 -7.91986436e-02 5.16543388e-01 6.50235772e-01
-1.17283612e-02 1.04198444e+00 7.69450903e-01 -7.93158337e-02
-1.06987751e+00 -1.05652201e+00 -1.38327047e-01 1.01379383e+00
4.20154585e-03 -1.24993503e+00 -5.09294569e-01 -5.36479414e-01
1.46584660e-01 6.54490888e-01 -8.18757653e-01 -2.57156402e-01
-5.60710132e-01 -5.44263303e-01 9.92467999e-01 5.58640063e-01
-5.54607138e-02 -8.85524511e-01 -6.50361955e-01 3.33551198e-01
-3.38144660e-01 -9.24893081e-01 -5.40397644e-01 2.16692820e-01
-9.71452832e-01 -1.06327987e+00 -5.57694077e-01 -6.13028824e-01
-1.14884272e-01 1.22975506e-01 1.07906079e+00 -3.44371796e-01
-6.52263463e-01 4.83866066e-01 -5.77704012e-01 -2.78751701e-01
-1.95953235e-01 5.45676351e-01 8.96807536e-02 3.19345109e-02
1.47020236e-01 -7.57752538e-01 -5.21199703e-01 1.79056853e-01
-8.22795928e-01 -7.21412897e-02 5.12497306e-01 6.95547402e-01
4.79261279e-01 -9.86901596e-02 3.28098983e-01 -4.93755430e-01
9.26036835e-01 -1.41010001e-01 -1.12096444e-01 1.31037176e-01
-7.30347931e-01 2.24537313e-01 3.94681126e-01 -5.60369194e-01
-4.37598675e-01 1.11188546e-01 6.06820472e-02 -4.38795179e-01
1.60853088e-01 4.60504055e-01 -3.65960933e-02 2.41318196e-01
7.12572038e-01 2.90163010e-01 3.86759490e-02 -1.05777812e+00
4.64232594e-01 6.92253470e-01 7.24493086e-01 -4.61092383e-01
6.99427009e-01 3.87262017e-01 -1.13728605e-01 -8.49448800e-01
-8.62878561e-01 -7.09373891e-01 -6.80601537e-01 1.93696886e-01
6.40734255e-01 -7.04035640e-01 -5.53241193e-01 6.37052134e-02
-8.35930526e-01 -6.89831004e-02 -6.16977751e-01 5.31336665e-01
-8.77322733e-01 6.85000122e-01 -4.77361143e-01 -6.78362727e-01
-5.43607354e-01 -8.05450022e-01 1.23856997e+00 -5.08892238e-01
-8.89701486e-01 -6.11243963e-01 5.97212911e-01 4.24788654e-01
4.73934263e-02 3.90422642e-01 8.64454985e-01 -7.36081719e-01
-1.63076401e-01 -3.25145543e-01 1.31139144e-01 4.40897107e-01
7.79352039e-02 -4.48402315e-02 -1.04737008e+00 -3.71970117e-01
-1.82031393e-01 -2.88484633e-01 1.26281404e+00 1.52931884e-01
1.10000563e+00 -2.69672722e-01 -1.22360878e-01 7.30605662e-01
1.30941224e+00 -4.50087667e-01 4.19810057e-01 6.56972945e-01
5.20829856e-01 3.38084251e-01 5.33421516e-01 7.36723483e-01
8.04757252e-02 1.22292173e+00 3.09727430e-01 1.71436131e-01
-1.98578939e-01 -1.98001117e-01 4.87448066e-01 9.39452052e-01
-4.09325771e-02 -9.49712992e-02 -5.09739757e-01 7.92538702e-01
-1.87013078e+00 -1.26739895e+00 3.88186388e-02 2.46216464e+00
8.54360700e-01 1.28833860e-01 6.92195058e-01 8.41600180e-01
3.54451835e-01 3.02271277e-01 -3.79294425e-01 -4.38414723e-01
-3.16156596e-01 3.04236948e-01 3.36802334e-01 8.71742219e-02
-1.20807099e+00 6.07831478e-01 6.60734606e+00 8.32794964e-01
-1.01903081e+00 1.38109371e-01 -1.08140774e-01 -6.26449823e-01
-2.82706290e-01 -2.94776354e-02 -5.33499122e-01 3.00218850e-01
6.22731268e-01 -4.21222411e-02 5.03213644e-01 6.23152912e-01
1.72368914e-01 3.33148897e-01 -1.18796289e+00 1.24011064e+00
5.46363950e-01 -1.39777696e+00 -9.03807506e-02 1.77918673e-01
5.33172548e-01 -8.57523605e-02 6.95945248e-02 1.98355436e-01
-4.74286139e-01 -7.31793523e-01 9.91483927e-01 2.59576291e-01
6.78483725e-01 -6.74604714e-01 4.84347880e-01 -1.22182432e-03
-1.25513232e+00 -2.99738258e-01 -1.96646359e-02 -1.33769572e-01
4.01932783e-02 3.62339288e-01 -5.36237657e-01 6.50877714e-01
6.85051739e-01 7.79711008e-01 -8.58079433e-01 1.31733704e+00
-6.17562942e-02 9.58735466e-01 -1.98053017e-01 7.10615441e-02
-1.20583028e-01 -5.62897883e-02 1.00861895e+00 1.19651043e+00
3.49844486e-01 -4.50345814e-01 1.13578849e-02 6.42060697e-01
1.25023574e-02 3.97018403e-01 -5.23198783e-01 -1.89938858e-01
4.03696865e-01 1.15475106e+00 -5.45215368e-01 -1.33646071e-01
-6.26230538e-02 1.33021116e+00 3.31131458e-01 4.02471200e-02
-6.49905801e-01 -6.03786290e-01 7.04677522e-01 1.28393263e-01
7.13118494e-01 -3.72230053e-01 -5.40091217e-01 -1.28048658e+00
4.79152769e-01 -8.78892839e-01 4.68535185e-01 -3.82986039e-01
-1.12531972e+00 4.16893840e-01 -1.31792516e-01 -1.62965643e+00
-2.49651998e-01 -3.22998434e-01 -6.32393301e-01 3.83356482e-01
-9.77742314e-01 -1.05569220e+00 -4.75042388e-02 1.56606764e-01
4.61655766e-01 -1.83466166e-01 1.17321229e+00 2.14354381e-01
-6.53611124e-01 1.05226409e+00 4.44272816e-01 -2.78592901e-03
9.78823602e-01 -1.34705722e+00 4.18786794e-01 6.52868330e-01
7.09491551e-01 5.94477057e-01 9.41746175e-01 -3.29442710e-01
-1.44920897e+00 -9.58324373e-01 8.48406315e-01 -7.20523238e-01
7.14760780e-01 -6.12919927e-01 -6.13007963e-01 4.87209141e-01
1.33848518e-01 -5.26278675e-01 1.24109340e+00 5.79899609e-01
-6.42214715e-01 -3.01894452e-02 -7.18200684e-01 5.54699183e-01
1.23948014e+00 -5.69709361e-01 -8.22275460e-01 6.67114928e-02
6.42399311e-01 -1.00892626e-01 -1.13707078e+00 2.31701806e-01
9.49692309e-01 -9.61799085e-01 1.04736054e+00 -5.27257860e-01
3.16152871e-01 -2.11069748e-01 -3.21674883e-01 -1.11946642e+00
-5.44191122e-01 -1.03383434e+00 -2.81333536e-01 1.63242817e+00
3.13232630e-01 -1.45145252e-01 8.15443516e-01 6.30965456e-02
-8.15448612e-02 -7.20331728e-01 -1.00124633e+00 -1.14872706e+00
-1.44787982e-01 -6.09962821e-01 5.04176855e-01 8.57952893e-01
1.20639332e-01 6.40919149e-01 -6.42079055e-01 -6.23467192e-02
5.53671122e-01 5.74138403e-01 9.50348914e-01 -1.34946728e+00
-7.35924602e-01 -7.89109826e-01 -7.13971138e-01 -7.52986312e-01
-2.52794158e-02 -9.30749536e-01 -2.93681085e-01 -1.43560421e+00
2.94004112e-01 1.03913648e-02 -4.21874672e-01 2.89795458e-01
-6.06889725e-02 7.01132178e-01 3.65496486e-01 3.84321541e-01
-8.00927222e-01 7.10694075e-01 6.18956566e-01 -2.13074535e-01
-4.94039953e-01 1.49318099e-01 -9.59242582e-01 4.68715549e-01
8.72802556e-01 -5.16119540e-01 -2.62298614e-01 -3.31053853e-01
5.49024105e-01 -2.37992525e-01 4.86548543e-01 -1.04183602e+00
-3.96828502e-02 3.59648764e-01 7.52394125e-02 -5.70227623e-01
6.44993007e-01 -4.27742481e-01 7.62757733e-02 2.11276859e-01
-5.89221954e-01 1.59316063e-01 2.56948531e-01 6.09610558e-01
-1.32000610e-01 -2.81027257e-01 2.98025250e-01 3.05353791e-01
-4.35413748e-01 -7.88611919e-02 -1.48117006e-01 1.60004362e-01
8.30325842e-01 -4.24961716e-01 1.55480191e-01 -1.67892098e-01
-9.18357849e-01 -1.78304031e-01 7.71981120e-01 6.53814971e-01
2.61222124e-01 -1.36487901e+00 -7.97337532e-01 1.19364634e-01
5.93597353e-01 -5.79336405e-01 6.52823783e-03 9.23318386e-01
-1.01732150e-01 3.56754847e-02 1.05946667e-01 -4.71489131e-01
-1.72368503e+00 5.23496091e-01 -1.84071302e-01 -2.82516509e-01
-8.08524728e-01 6.61785781e-01 -3.70540023e-01 -3.13332707e-01
5.37037969e-01 -3.01194012e-01 -1.38286561e-01 5.29968262e-01
5.28355002e-01 5.71119964e-01 3.11115682e-01 -4.87689286e-01
-3.94705445e-01 6.75233901e-01 2.10060209e-01 -3.19697350e-01
1.41033471e+00 -3.57672647e-02 1.56275839e-01 8.67362618e-01
1.33137178e+00 4.75867510e-01 -9.18163478e-01 -1.49757653e-01
6.43302575e-02 -3.74715179e-01 -3.70981134e-02 -7.41833568e-01
-4.71809715e-01 7.32977867e-01 6.55563951e-01 4.76236314e-01
1.03271031e+00 1.46235421e-01 7.62739718e-01 1.08121641e-01
-7.10356385e-02 -1.10933280e+00 2.19822988e-01 2.49096632e-01
9.79368508e-01 -8.28019381e-01 2.66059637e-01 -1.04835026e-01
-7.03875661e-01 9.22480881e-01 -7.40741864e-02 -9.45497826e-02
2.79349893e-01 4.06566299e-02 -1.49329141e-01 -1.80835262e-01
-7.82088399e-01 -1.70854583e-01 4.55619723e-01 2.82173634e-01
5.34170866e-01 9.70284566e-02 -3.75290841e-01 9.05377865e-01
-6.60254717e-01 -2.38983169e-01 3.36724401e-01 7.69795537e-01
-1.66472435e-01 -1.41761315e+00 -2.76226163e-01 5.59937298e-01
-5.90595663e-01 -6.32909760e-02 -8.78826201e-01 5.40602028e-01
-1.06306709e-02 8.82188559e-01 -1.52879134e-02 -6.16446614e-01
4.46869165e-01 1.62748232e-01 7.13924050e-01 -5.97944498e-01
-7.95359373e-01 1.30376831e-01 1.51517674e-01 -5.03694296e-01
-2.61909515e-01 -1.09604979e+00 -6.37758076e-01 -5.04300632e-02
-3.80212992e-01 1.43958256e-01 7.21457601e-01 6.92176044e-01
6.74525678e-01 5.49850643e-01 6.60695970e-01 -8.57135952e-01
-7.91786849e-01 -9.94407296e-01 -6.67571485e-01 7.14594364e-01
2.02306658e-01 -5.55755377e-01 -4.24689025e-01 -9.78457034e-02] | [15.79300594329834, 5.316997528076172] |
f44c23ac-9e2e-427e-bdb6-e4ba7615f473 | towards-an-approach-based-on-knowledge-graph | null | null | https://ceur-ws.org/Vol-3320/paper12.pdf | https://ceur-ws.org/Vol-3320/paper12.pdf | Towards an Approach based on Knowledge Graph Refinement for Tabular Data to Knowledge Graph Matching | This paper presents our contribution to the Accuracy Track of Semantic Web Challenge on Tabular Data to Knowledge Graph Matching (SemTab). This contribution consists of the proposition of an approach based on knowledge graph refinement for tabular data annotation. Internal methods were used to predict the links between cells in the table and external methods were used to predict missing entities and relations. This approach was applied to the annotation of HardTables and ToughTables using DBpedia and Wikidata; and GitTables and BiodivTab using DBpedia and Schema.org. During Round 3 of the competition, we were ranked third and second position respectively for the annotation of GitTables and BiodivTab. | ['Brice Foko', 'Azanzi Jiomekong'] | 2022-10-25 | null | null | null | semtab-iswc-2022-10 | ['graph-matching', 'column-type-annotation', 'cell-entity-annotation'] | ['graphs', 'natural-language-processing', 'natural-language-processing'] | [-7.14726567e-01 1.18965292e+00 -1.99944898e-01 -4.09163088e-02
-4.90948081e-01 -7.75421500e-01 6.33362234e-01 9.47526932e-01
3.10877291e-03 1.22548771e+00 3.60307842e-01 1.38142332e-02
-9.14793134e-01 -1.22290814e+00 -7.07344115e-01 3.10905516e-01
-2.01568067e-01 1.49080729e+00 8.51769686e-01 -7.71541536e-01
-7.36397505e-02 3.63102742e-02 -1.57861960e+00 9.01145875e-01
5.04723072e-01 1.13208318e+00 -4.06892270e-01 1.90812498e-01
-5.70700228e-01 1.47938728e+00 -1.04766473e-01 -1.14482892e+00
2.05332205e-01 8.71394854e-03 -1.53659761e+00 -5.03075421e-01
6.66528642e-01 7.32612610e-01 -2.98859745e-01 9.93006587e-01
3.00643355e-01 -6.55720010e-02 2.91061044e-01 -1.52890015e+00
-2.12881193e-01 1.18519497e+00 -3.04487888e-02 -9.56807956e-02
9.89175975e-01 -6.41563535e-01 1.34393501e+00 -6.87155187e-01
1.58102286e+00 1.24399257e+00 1.04226243e+00 2.49621958e-01
-8.58280957e-01 -2.98559606e-01 -2.71127466e-02 7.20247149e-01
-1.50199544e+00 -5.02853930e-01 -7.05627501e-02 -5.43820500e-01
1.38685393e+00 5.04927039e-01 7.99567223e-01 1.75637290e-01
-2.67686218e-01 -1.10719882e-01 1.05115211e+00 -4.75431174e-01
9.33732614e-02 5.33203244e-01 2.16605827e-01 1.07928514e+00
7.30186045e-01 -2.90236801e-01 -8.96382272e-01 -1.79557323e-01
2.47679770e-01 -9.14343178e-01 5.32395989e-02 -6.45784378e-01
-8.32065821e-01 3.49810511e-01 6.99373484e-02 1.79485679e-01
-1.57977507e-01 -1.30783558e-01 9.37897503e-01 2.64026076e-01
2.60470390e-01 6.24392509e-01 -8.26799273e-01 -3.22720297e-02
-4.30895299e-01 6.00184202e-01 1.44605851e+00 1.62962735e+00
6.62133992e-01 -6.09273851e-01 1.34446695e-01 8.44707727e-01
2.63038605e-01 1.19264442e-02 -3.44569609e-02 -9.50579762e-01
8.08272779e-01 1.32537127e+00 4.16836321e-01 -1.14168286e+00
-8.06259394e-01 1.15016224e-02 -1.08295813e-01 -9.81593654e-02
6.75336480e-01 2.05326274e-01 -7.45618701e-01 1.03192627e+00
6.30193532e-01 -4.54696625e-01 2.69795954e-01 5.75836480e-01
1.76856589e+00 2.08793968e-01 2.10462272e-01 -1.90417185e-01
1.65930378e+00 -6.38198733e-01 -1.06464040e+00 3.03696901e-01
9.91506875e-01 -5.82753658e-01 5.00373244e-01 2.80194521e-01
-1.19021249e+00 -3.64764810e-01 -8.41723263e-01 -2.31693074e-01
-1.40418720e+00 -6.30049348e-01 7.13669717e-01 5.20825028e-01
-8.52090597e-01 4.00231272e-01 -4.52643782e-01 -8.89959991e-01
2.59362727e-01 3.36802661e-01 -8.92170370e-01 -1.10449769e-01
-1.76298738e+00 1.47026658e+00 1.27543676e+00 -2.69614339e-01
-2.25539073e-01 -9.35999990e-01 -8.55259478e-01 -8.53261575e-02
9.42778766e-01 -7.10785508e-01 8.52595091e-01 -1.27049208e-01
-7.02788651e-01 1.22289324e+00 2.93219447e-01 -8.51323128e-01
4.30832237e-01 2.21920118e-01 -1.07398260e+00 -1.62388846e-01
4.31404531e-01 4.09902930e-01 -3.18451047e-01 -1.18290341e+00
-9.41989899e-01 -4.49869365e-01 3.37146014e-01 1.78534836e-01
1.40972421e-01 6.98537156e-02 -7.28429079e-01 5.79261072e-02
2.71075606e-01 -3.01522851e-01 2.93460160e-01 -6.44730926e-01
-4.24085468e-01 -6.28162086e-01 3.35977286e-01 -9.41599488e-01
1.50980353e+00 -1.24701679e+00 9.95247811e-02 6.14636719e-01
3.94498050e-01 -1.93812042e-01 4.91897523e-01 1.18044817e+00
-1.26406863e-01 3.70359838e-01 4.47791606e-01 5.23141444e-01
4.79869545e-01 3.86671603e-01 -9.31042433e-02 -1.42371997e-01
-4.34673369e-01 9.42939699e-01 -7.51269460e-01 -8.87257993e-01
-2.33650178e-01 6.78012446e-02 -3.26618344e-01 -4.10511225e-01
-7.81743646e-01 -3.72959465e-01 -1.28089637e-01 7.19610035e-01
3.71847153e-01 -1.66263044e-01 8.77037883e-01 -7.09501743e-01
-3.57993573e-01 8.00336838e-01 -1.40173864e+00 1.55107892e+00
-1.38855036e-02 8.69546309e-02 -2.07220912e-01 -4.52892661e-01
9.50095654e-01 5.23139358e-01 7.78050661e-01 -1.09360337e+00
-2.10122541e-01 5.43800652e-01 -4.46106166e-01 -7.27638602e-01
7.37344205e-01 1.22810729e-01 -2.59358704e-01 -1.38304606e-01
2.01287851e-01 -1.43182710e-01 9.45190668e-01 6.15007281e-01
1.00528955e+00 6.85197175e-01 4.85956401e-01 -5.16540110e-01
7.27615714e-01 8.82783651e-01 4.71431583e-01 3.67709905e-01
4.89906371e-01 5.95649593e-02 8.35113645e-01 -1.03026271e+00
-9.98761237e-01 -7.41331100e-01 -3.17957029e-02 8.69451523e-01
5.22392010e-03 -1.32703400e+00 -5.51050544e-01 -1.06682932e+00
1.55687779e-01 6.44702315e-01 -8.35676193e-01 5.02743185e-01
-3.31423640e-01 -3.17462564e-01 6.43479407e-01 2.81106889e-01
3.15541804e-01 -7.66250372e-01 -1.87175095e-01 1.09249555e-01
-4.64288056e-01 -1.34964657e+00 1.81885183e-01 1.79727226e-01
-4.05500323e-01 -1.78212535e+00 5.13903081e-01 -6.28609180e-01
2.81466633e-01 -6.86631143e-01 1.60549998e+00 1.52377546e-01
3.69383059e-02 3.70101094e-01 -4.13813502e-01 -6.33906662e-01
-2.69406438e-01 2.19499573e-01 -3.10501933e-01 -7.52336502e-01
5.17448843e-01 -1.08281754e-01 -5.36539108e-02 5.88940084e-01
-6.93449020e-01 2.08345786e-01 -3.63256902e-01 5.75940013e-01
6.01662755e-01 3.52972537e-01 4.80153620e-01 -1.58476841e+00
2.72275776e-01 -5.87056994e-01 -9.85039234e-01 6.74171686e-01
-1.19410944e+00 7.40074068e-02 2.32360110e-01 6.83208287e-01
-8.90691817e-01 1.36216775e-01 -1.16512872e-01 2.29128435e-01
2.99215108e-01 1.22174251e+00 -3.40648890e-01 -3.91787551e-02
7.04489589e-01 -4.87684548e-01 -2.26319820e-01 -6.32505655e-01
4.30061430e-01 1.61793083e-01 5.80364108e-01 -7.59297252e-01
7.12322712e-01 3.15120161e-01 3.32602292e-01 1.16115680e-03
-7.15241015e-01 -5.03973246e-01 -9.13537323e-01 -3.50297511e-01
6.68317556e-01 -9.71828401e-01 -9.84659016e-01 -8.97092074e-02
-9.74465489e-01 -1.98378354e-01 -5.27493358e-01 -1.11700535e-01
-4.80003983e-01 4.10737656e-02 -1.33228242e-01 -3.87468427e-01
-1.50741220e-01 -3.01981658e-01 5.67354441e-01 -2.24378541e-01
-4.04485881e-01 -1.35788655e+00 2.74455011e-01 8.41659606e-01
1.42717093e-01 4.79454309e-01 1.20899713e+00 -1.18116367e+00
-3.91942978e-01 -2.59235293e-01 -4.53753054e-01 -4.82920855e-01
-1.60081655e-01 -1.76274166e-01 -4.84186709e-01 1.50478035e-01
-1.17043185e+00 -3.43597263e-01 1.26898855e-01 -5.08261681e-01
7.58388221e-01 -6.53184950e-01 -6.59449935e-01 2.15358570e-01
1.80862510e+00 2.01867327e-01 1.00711501e+00 1.08899939e+00
6.95982218e-01 1.04758978e+00 8.86151075e-01 2.55134374e-01
1.09555304e+00 1.27140391e+00 6.18567884e-01 3.50789309e-01
-4.47895467e-01 -5.72400689e-01 -3.09181869e-01 6.74036741e-01
-4.55796659e-01 -2.44186461e-01 -1.46023166e+00 7.72126496e-01
-2.10296321e+00 -9.78878140e-01 -8.01579475e-01 2.14108038e+00
9.37464237e-01 3.21879774e-01 3.33565146e-01 6.05716854e-02
4.56173390e-01 -6.19783401e-01 6.73120841e-02 -3.57131034e-01
-2.99987525e-01 4.20238286e-01 6.84734285e-01 6.34276330e-01
-6.81918859e-01 9.27345812e-01 5.96058846e+00 7.88819611e-01
-6.97078407e-02 1.81764379e-01 1.34760737e-01 2.59889543e-01
-3.56482476e-01 4.31744635e-01 -1.28153133e+00 1.60640359e-01
1.15941656e+00 -3.79224569e-01 4.26813751e-01 4.20273513e-01
-2.64244914e-01 -9.21501741e-02 -9.41365480e-01 5.15035033e-01
-2.07306162e-01 -2.01882553e+00 -1.71493515e-02 -1.10238872e-01
5.17738223e-01 -1.35640696e-01 -7.10512936e-01 4.54744488e-01
6.63063586e-01 -1.21051311e+00 8.48390400e-01 8.42870414e-01
7.52665162e-01 -6.94471896e-01 1.06571925e+00 -1.56172320e-01
-1.28118002e+00 2.09446847e-01 -2.32454032e-01 1.99864879e-01
-2.64172584e-01 2.09584415e-01 -1.11714387e+00 1.45759368e+00
1.16180372e+00 5.64431608e-01 -6.40898824e-01 1.15197098e+00
2.98679452e-02 6.42469749e-02 -2.19125226e-01 2.22976074e-01
-1.52431548e-01 7.27743730e-02 3.97269577e-01 1.09068763e+00
-7.30745634e-03 -6.65722862e-02 -2.04882175e-01 4.48610336e-01
-3.40868682e-01 4.51280385e-01 -3.99579227e-01 1.17911078e-01
7.11122155e-01 1.09185708e+00 -5.96545458e-01 -4.65338826e-01
-4.37422663e-01 1.38732404e-01 4.00781512e-01 -1.62547007e-01
-5.82951546e-01 -5.46722770e-01 1.70053050e-01 7.59811640e-01
2.88946211e-01 4.27492052e-01 -3.78618717e-01 -7.50594437e-01
-6.53392449e-02 -9.75406647e-01 1.39446449e+00 -1.14590704e+00
-1.21872723e+00 8.34791303e-01 5.23223221e-01 -9.21420574e-01
-1.74571484e-01 -5.65489650e-01 3.23676854e-01 7.27666616e-01
-1.09015512e+00 -1.34686744e+00 -4.91255373e-01 6.81872308e-01
-1.15042835e-01 -1.81366548e-01 1.11057127e+00 7.22957194e-01
-6.02096207e-02 2.38794014e-01 -2.05939189e-01 6.46338686e-02
7.09579051e-01 -1.46811998e+00 2.58461356e-01 3.20785314e-01
1.05523527e-01 3.93462181e-01 9.02577639e-01 -1.14567089e+00
-1.31540692e+00 -1.10804093e+00 1.54263973e+00 -9.31872010e-01
1.05220068e+00 -2.95502931e-01 -8.04153979e-01 9.77289259e-01
5.77961266e-01 -3.43262106e-02 2.87837297e-01 3.79477829e-01
-5.75357556e-01 -3.19075525e-01 -1.22964072e+00 8.55767280e-02
1.23079801e+00 -3.33706081e-01 -5.08701324e-01 6.69635475e-01
4.17686999e-01 -1.00744033e+00 -1.72590435e+00 4.57036495e-01
6.58309817e-01 -5.69926798e-01 8.34556580e-01 -9.76155281e-01
1.26261741e-01 -6.55468345e-01 -2.10052609e-01 -1.08433151e+00
-7.81930890e-03 -5.80660522e-01 -2.93458462e-01 1.40035295e+00
1.02147877e+00 -5.98916352e-01 8.67405236e-01 6.44840539e-01
-2.66186148e-01 -2.63452023e-01 -1.06605172e+00 -8.93227696e-01
-2.14029744e-01 -2.88785398e-01 8.56916666e-01 1.24471450e+00
9.01817560e-01 2.41800517e-01 -1.34761110e-01 3.09661608e-02
4.63507265e-01 -2.03787744e-01 5.74736238e-01 -1.69045687e+00
8.58607367e-02 -3.40287015e-03 -7.35086203e-01 5.37009649e-02
-2.97953755e-01 -1.28937745e+00 -5.07529616e-01 -2.38848734e+00
7.59664224e-03 -6.99347019e-01 -1.25439288e-02 9.38701332e-01
5.78951240e-01 3.82626355e-01 2.10589170e-02 6.93002939e-02
-1.11628807e+00 -3.35770041e-01 7.78367400e-01 -5.90547100e-02
2.03569517e-01 -4.22046155e-01 -7.94844270e-01 3.52994442e-01
6.55271709e-01 -7.14011848e-01 -3.09701681e-01 -1.19517803e-01
1.43770468e+00 3.31162244e-01 2.29978383e-01 -8.14117968e-01
6.23620093e-01 -1.29457423e-02 -5.30842394e-02 -7.40882158e-01
1.48116872e-01 -1.07290673e+00 1.13192201e+00 3.06981444e-01
-1.85084656e-01 1.35841012e-01 5.89381337e-01 1.76533103e-01
-4.03866321e-01 -2.43519858e-01 1.78214282e-01 -4.46379006e-01
-8.86871815e-01 2.65175384e-02 -1.23728178e-02 4.44229335e-01
8.42224479e-01 -2.54223734e-01 -1.02191091e+00 -2.07575578e-02
-1.34903347e+00 6.27523065e-01 2.43749276e-01 4.33033437e-01
3.17440897e-01 -1.46417856e+00 -5.82927644e-01 -4.23444986e-01
5.33701599e-01 -1.52338341e-01 -1.24667160e-01 1.18349886e+00
-8.12608182e-01 9.65607107e-01 -5.71989298e-01 2.25136474e-01
-1.42857909e+00 5.06081879e-01 4.96729493e-01 -6.87008023e-01
-2.89420009e-01 6.23947799e-01 -7.32194006e-01 -8.27122569e-01
2.27150798e-01 -6.78659359e-04 -6.27938926e-01 3.99616092e-01
1.15316220e-01 7.13853836e-01 9.11236048e-01 -4.95872825e-01
-6.61337316e-01 1.32431448e-01 1.34050593e-01 2.27281272e-01
1.42187989e+00 -1.80897489e-01 -7.86310494e-01 2.81765312e-01
4.33799297e-01 3.16337734e-01 -2.54523307e-01 3.60164791e-02
8.87030959e-01 -2.01302350e-01 -2.78073192e-01 -1.70270371e+00
-7.64340937e-01 -5.32406103e-03 2.17702821e-01 9.64438915e-01
6.04837179e-01 2.95881033e-01 3.64880979e-01 2.37825602e-01
7.07655430e-01 -1.34620214e+00 -6.98987007e-01 7.26813197e-01
1.01050639e+00 -9.54541922e-01 2.42617726e-01 -1.19196844e+00
-4.71963257e-01 1.32541943e+00 7.66988695e-01 5.72117031e-01
5.41611850e-01 3.72520715e-01 -2.44400412e-01 -1.18030572e+00
-1.16441584e+00 -3.64473790e-01 5.54501414e-01 8.62321079e-01
5.22415578e-01 -1.49210185e-01 -6.60508752e-01 7.50943542e-01
-2.91074276e-01 2.66152978e-01 3.11642647e-01 7.05969155e-01
-2.74381012e-01 -1.38782704e+00 -2.57041663e-01 5.11543572e-01
-4.64480788e-01 -3.44524920e-01 -8.90910387e-01 1.46374249e+00
4.71018255e-01 6.69282317e-01 -5.11061475e-02 -5.02966702e-01
9.33443844e-01 3.43505085e-01 5.64883649e-01 -6.62750065e-01
-1.16470718e+00 -3.78526002e-01 1.34667611e+00 -6.46536529e-01
-4.76970583e-01 -7.92253435e-01 -1.51196682e+00 -5.63908637e-01
-2.23693416e-01 8.63442063e-01 4.68050092e-01 7.05704272e-01
2.77335197e-01 4.42285776e-01 -2.60250896e-01 3.86827886e-01
3.87187570e-01 -7.50065267e-01 -5.54628193e-01 3.40000123e-01
-5.79692662e-01 -8.83399248e-01 2.08727866e-01 2.06357643e-01] | [9.286625862121582, 8.046356201171875] |
a6a08782-a4c5-4292-a817-465197a95cec | spatio-temporal-self-supervised-learning-for | 2212.04475 | null | https://arxiv.org/abs/2212.04475v1 | https://arxiv.org/pdf/2212.04475v1.pdf | Spatio-Temporal Self-Supervised Learning for Traffic Flow Prediction | Robust prediction of citywide traffic flows at different time periods plays a crucial role in intelligent transportation systems. While previous work has made great efforts to model spatio-temporal correlations, existing methods still suffer from two key limitations: i) Most models collectively predict all regions' flows without accounting for spatial heterogeneity, i.e., different regions may have skewed traffic flow distributions. ii) These models fail to capture the temporal heterogeneity induced by time-varying traffic patterns, as they typically model temporal correlations with a shared parameterized space for all time periods. To tackle these challenges, we propose a novel Spatio-Temporal Self-Supervised Learning (ST-SSL) traffic prediction framework which enhances the traffic pattern representations to be reflective of both spatial and temporal heterogeneity, with auxiliary self-supervised learning paradigms. Specifically, our ST-SSL is built over an integrated module with temporal and spatial convolutions for encoding the information across space and time. To achieve the adaptive spatio-temporal self-supervised learning, our ST-SSL first performs the adaptive augmentation over the traffic flow graph data at both attribute- and structure-levels. On top of the augmented traffic graph, two SSL auxiliary tasks are constructed to supplement the main traffic prediction task with spatial and temporal heterogeneity-aware augmentation. Experiments on four benchmark datasets demonstrate that ST-SSL consistently outperforms various state-of-the-art baselines. Since spatio-temporal heterogeneity widely exists in practical datasets, the proposed framework may also cast light on other spatial-temporal applications. Model implementation is available at https://github.com/Echo-Ji/ST-SSL. | ['Yu Zheng', 'Junbo Zhang', 'Zhenhe Wu', 'Boren Xu', 'Junjie Wu', 'Chao Huang', 'Jingyuan Wang', 'Jiahao Ji'] | 2022-12-07 | null | null | null | null | ['robust-traffic-prediction', 'spatio-temporal-forecasting'] | ['time-series', 'time-series'] | [-9.50992629e-02 -5.35702229e-01 -8.48461449e-01 -5.67623615e-01
-4.18675780e-01 -2.50114471e-01 7.44706154e-01 7.89657235e-02
2.02854238e-02 7.76438177e-01 4.16795820e-01 -6.78223848e-01
-3.63033712e-01 -1.19758594e+00 -6.64770782e-01 -4.38959897e-01
-4.00131375e-01 4.29557115e-01 5.88605046e-01 -2.58670539e-01
-9.13283452e-02 5.92333376e-01 -1.48687136e+00 1.75465956e-01
1.15913403e+00 1.01789141e+00 9.62174032e-03 3.90071571e-01
-3.85322154e-01 8.45694363e-01 -7.23169595e-02 -1.36896580e-01
3.48771900e-01 -4.36273068e-02 -5.31381428e-01 -1.33911083e-02
5.12218356e-01 -1.07770212e-01 -1.19081068e+00 6.66299522e-01
-2.45135706e-02 3.35716367e-01 5.57875872e-01 -2.01586533e+00
-6.16965234e-01 5.27607322e-01 -6.74657166e-01 7.66879141e-01
-1.55975774e-01 6.42400622e-01 9.37440753e-01 -4.93060559e-01
-5.32189989e-03 1.22699177e+00 6.87704504e-01 7.50062540e-02
-1.30117118e+00 -9.56828237e-01 7.37426400e-01 4.64729249e-01
-1.48840404e+00 -3.36431086e-01 9.67026055e-01 -5.09657800e-01
8.18472087e-01 1.61704868e-01 5.22354364e-01 1.18113923e+00
2.63488311e-02 9.22074795e-01 9.08256412e-01 3.95707756e-01
-1.00718789e-01 -3.73591892e-02 2.62947559e-01 5.39645016e-01
7.95589015e-02 2.37203062e-01 -3.21355402e-01 -6.78141564e-02
5.46949685e-01 3.16885710e-01 2.65408218e-01 -1.60056621e-01
-1.09686470e+00 5.04752994e-01 6.97173238e-01 8.58396888e-02
-4.39142168e-01 3.60061377e-01 5.91694415e-01 1.03582352e-01
7.99151957e-01 -4.14978772e-01 -5.87532818e-01 -2.18619153e-01
-9.02745247e-01 2.33188599e-01 2.55473226e-01 1.27659929e+00
1.20188820e+00 2.79869556e-01 -3.98458213e-01 8.14962327e-01
1.45381004e-01 7.26063132e-01 9.40887854e-02 -7.28315234e-01
1.12872422e+00 7.04053462e-01 -1.94493607e-01 -1.11236155e+00
-6.02755666e-01 -6.19390666e-01 -1.19642007e+00 -2.90553480e-01
4.72243845e-01 -1.53669760e-01 -7.53113329e-01 1.93503988e+00
3.70175749e-01 1.02916121e+00 -2.58138955e-01 6.00020468e-01
7.35199988e-01 7.94012070e-01 5.06342351e-01 1.24812022e-01
1.06369889e+00 -1.14462364e+00 -7.39848554e-01 -2.15030059e-01
8.15202057e-01 -3.25343221e-01 1.06277406e+00 -5.39939463e-01
-8.75956595e-01 -6.86191022e-01 -4.65148747e-01 1.55936733e-01
-8.32992494e-01 6.59717917e-02 8.05932462e-01 4.49817985e-01
-7.57348478e-01 1.42215699e-01 -7.14637518e-01 -2.96924204e-01
7.47311592e-01 1.39795199e-01 -1.18729211e-01 4.72618230e-02
-1.49403214e+00 4.79591519e-01 1.20779216e-01 1.20830290e-01
-4.65741068e-01 -1.33278334e+00 -9.81099606e-01 1.02043405e-01
4.51852113e-01 -6.46816671e-01 7.80391335e-01 -5.33747137e-01
-1.00139427e+00 3.14547747e-01 -5.71407616e-01 -3.37273985e-01
5.11139929e-01 1.77206188e-01 -1.19835854e+00 -2.81357497e-01
6.33041680e-01 2.77196586e-01 3.96849662e-01 -1.07539964e+00
-9.48190570e-01 -2.17398658e-01 -7.21066371e-02 -2.55883992e-01
-3.01828831e-01 -2.82329857e-01 -6.58583403e-01 -9.19175267e-01
-2.06796765e-01 -9.06518221e-01 -2.44148999e-01 -1.00115143e-01
-4.08824384e-01 -3.00865024e-01 1.11970055e+00 -3.06639165e-01
1.64055777e+00 -1.88683629e+00 -6.18690550e-01 3.95459741e-01
1.31419659e-01 4.73899245e-01 -3.91084075e-01 5.71799457e-01
-1.05499551e-01 2.97592208e-02 -2.86506444e-01 -4.22657758e-01
1.30461484e-01 4.65110838e-01 -3.86910021e-01 3.62168700e-01
4.17492092e-01 1.19162345e+00 -1.09318495e+00 -5.49083710e-01
4.54518080e-01 4.39088911e-01 -4.07822162e-01 -1.12648383e-01
-1.67068690e-01 7.49553204e-01 -6.41426682e-01 6.56895757e-01
9.81184542e-01 -4.23227102e-01 -1.98319510e-01 -2.04862222e-01
-3.79217923e-01 5.23000658e-01 -1.08062255e+00 1.23101938e+00
-6.28900111e-01 6.78354144e-01 -3.57794166e-01 -1.20800126e+00
7.62234747e-01 1.44901067e-01 9.19095278e-01 -1.31094885e+00
-3.26369017e-01 2.13676877e-02 -1.88162416e-01 -5.89337051e-01
4.48874831e-01 2.27865353e-01 1.45374220e-02 4.47558165e-01
-2.81753153e-01 7.31388688e-01 3.05198520e-01 3.06611836e-01
1.11850750e+00 -7.80622289e-02 -2.05976710e-01 -6.57585561e-02
8.18185389e-01 4.15104665e-02 8.89348090e-01 5.46438098e-01
-5.24714589e-01 1.78582013e-01 4.35343295e-01 -7.66608417e-01
-9.00273204e-01 -1.21217501e+00 -1.79460585e-01 1.19647312e+00
2.41449460e-01 -2.68219918e-01 -6.07930273e-02 -8.18562627e-01
3.57021093e-01 5.66734016e-01 -7.62867391e-01 -1.15748063e-01
-8.59357655e-01 -8.29239309e-01 6.10596180e-01 7.52305686e-01
8.21494401e-01 -6.08866513e-01 1.20551668e-01 2.06219718e-01
-2.98015773e-01 -1.58217680e+00 -7.84829438e-01 -4.53522384e-01
-7.13600874e-01 -9.32514071e-01 -3.44095320e-01 -3.86056930e-01
3.73658448e-01 6.99417651e-01 1.06634879e+00 3.16503905e-02
-1.05883002e-01 2.39691779e-01 -1.02638513e-01 -1.63260102e-01
-8.66937190e-02 4.63736892e-01 -8.77720788e-02 5.42076826e-01
5.75183153e-01 -1.07891786e+00 -8.33023012e-01 7.53208876e-01
-6.90308690e-01 1.23291180e-01 4.40447032e-01 4.51515138e-01
4.71641809e-01 2.79818475e-01 8.71329427e-01 -8.73810470e-01
3.00615460e-01 -1.11291337e+00 -3.73955041e-01 8.07493404e-02
-5.73573232e-01 -1.39675081e-01 7.08711863e-01 -4.31857049e-01
-1.06565547e+00 -2.73241848e-01 5.12970611e-02 -3.83081168e-01
-3.94018650e-01 4.87650931e-01 -2.63675153e-01 2.66384214e-01
2.58442461e-01 4.65620369e-01 -7.77096897e-02 -3.24457824e-01
2.93052614e-01 5.16537249e-01 3.28120649e-01 -7.04395890e-01
1.20440233e+00 6.89750075e-01 2.84090519e-01 -8.17970574e-01
-7.75455296e-01 -6.93162918e-01 -7.33355999e-01 -3.36386710e-01
5.25348186e-01 -1.06896770e+00 -6.34687781e-01 4.03750271e-01
-7.80045390e-01 -6.82435632e-01 -2.15441227e-01 5.22760212e-01
-4.75539356e-01 2.36758962e-01 -3.64994049e-01 -7.57935405e-01
1.58012569e-01 -9.59937036e-01 9.51638520e-01 1.14680091e-02
1.83887839e-01 -1.50187087e+00 2.97696292e-02 3.98810893e-01
7.98958063e-01 1.08875379e-01 1.04593456e+00 -3.75657320e-01
-8.47270250e-01 1.12182833e-02 -6.98842287e-01 1.05474340e-02
5.07770240e-01 -2.01625209e-02 -7.12869227e-01 7.53572434e-02
-1.02312303e+00 2.43649215e-01 1.05172968e+00 4.44155127e-01
1.45023513e+00 -4.89826918e-01 -5.81772983e-01 7.21771538e-01
1.10691106e+00 5.10440953e-02 5.72801590e-01 3.31233948e-01
9.86314058e-01 6.76277816e-01 5.51479816e-01 4.03847426e-01
1.14537454e+00 8.31202686e-01 3.50512266e-01 -2.85260856e-01
-3.93485904e-01 -5.39803624e-01 1.01908728e-01 5.84149480e-01
-3.43247056e-02 -3.18316519e-01 -1.14274776e+00 8.60177040e-01
-2.28439164e+00 -1.34520006e+00 -6.12097859e-01 2.10238647e+00
2.86243021e-01 1.26505673e-01 7.64958441e-01 -7.89476838e-03
6.82677567e-01 6.77033961e-01 -7.23720253e-01 -2.00061947e-02
-2.71690458e-01 -2.14353457e-01 9.33716953e-01 4.49078470e-01
-1.25398242e+00 1.11930895e+00 5.62294865e+00 8.24580848e-01
-1.16100121e+00 1.12998471e-01 6.08118117e-01 1.06541691e-02
-3.54726821e-01 -1.22931942e-01 -7.51452923e-01 9.50005949e-01
1.11374116e+00 -1.53693482e-01 2.90651977e-01 4.49515343e-01
7.10409164e-01 3.75086546e-01 -6.95988655e-01 7.26919651e-01
-5.58505237e-01 -1.48469520e+00 3.19764853e-01 1.93795472e-01
7.41296887e-01 4.37996536e-01 3.50990236e-01 5.60985208e-01
3.93636435e-01 -9.31394517e-01 3.16950798e-01 6.73852503e-01
6.05506063e-01 -6.62229002e-01 4.01017994e-01 2.27661550e-01
-2.08743238e+00 -2.83313006e-01 2.56383903e-02 2.02708542e-02
5.36176980e-01 5.29936552e-01 -3.09074163e-01 7.18908906e-01
5.21722615e-01 1.31334364e+00 -6.63787663e-01 9.96916592e-01
1.23254105e-01 8.92179072e-01 -2.59958893e-01 4.20536906e-01
5.68717122e-01 -3.45730573e-01 5.08902729e-01 1.50700855e+00
2.59271473e-01 -6.60846084e-02 3.59297812e-01 8.53926659e-01
5.76334372e-02 -1.38182983e-01 -6.50247812e-01 1.93267822e-01
7.50305176e-01 8.88979554e-01 -2.43537694e-01 -4.25614864e-01
-8.89506876e-01 4.60489154e-01 2.68802315e-01 8.28429997e-01
-1.24533892e+00 -3.67900357e-02 1.24473286e+00 6.67116225e-01
2.33504251e-01 -4.37408358e-01 -4.31063414e-01 -1.04010296e+00
1.47761360e-01 -3.14209908e-01 6.11020863e-01 -4.97856855e-01
-1.60178649e+00 4.89565641e-01 1.65236309e-01 -1.43288040e+00
1.87781990e-01 -3.20958883e-01 -9.66360688e-01 8.32983255e-01
-2.09227753e+00 -1.68515956e+00 -4.01880473e-01 8.97410095e-01
6.18171275e-01 -3.89240652e-01 1.96844980e-01 7.86205769e-01
-1.11264181e+00 7.35249519e-01 -1.85796142e-01 2.20964044e-01
4.82253522e-01 -9.84807014e-01 7.12085009e-01 8.41186643e-01
-1.13838293e-01 3.94032985e-01 2.89854497e-01 -6.80999100e-01
-1.10536420e+00 -1.97551727e+00 1.13171136e+00 -3.74857903e-01
1.16095626e+00 -1.51067033e-01 -1.07821250e+00 8.62873614e-01
-1.73153654e-01 6.68635726e-01 6.26707077e-01 1.95835710e-01
-6.32626474e-01 -7.01597750e-01 -7.74799705e-01 6.04974329e-01
1.47905791e+00 -7.06537664e-01 4.49553039e-03 4.39511538e-01
6.07950211e-01 9.99947190e-02 -7.95411527e-01 5.14109254e-01
3.92708272e-01 -6.63236022e-01 1.11281264e+00 -9.04444993e-01
1.88459039e-01 -4.91620719e-01 -2.24417001e-01 -1.14407718e+00
-6.07335508e-01 -4.43104029e-01 -2.83777565e-01 1.33949041e+00
5.20035326e-01 -1.02557647e+00 8.69595110e-01 6.80984080e-01
-2.17532098e-01 -7.17693269e-01 -1.14136064e+00 -1.08498788e+00
1.31106555e-01 -8.36010456e-01 1.21409416e+00 1.07821381e+00
-3.20998102e-01 2.60033697e-01 -5.13435364e-01 5.34585476e-01
7.08065212e-01 1.18948795e-01 1.05133665e+00 -1.19830513e+00
3.37035418e-01 -8.23762596e-01 -5.39463997e-01 -1.32145143e+00
6.28226817e-01 -9.61795807e-01 -4.63519692e-01 -1.39102197e+00
4.82033640e-02 -9.95748818e-01 -5.37916839e-01 4.92112011e-01
-2.39290237e-01 3.95456366e-02 -9.64545608e-02 2.04611614e-01
-7.06596255e-01 9.04940426e-01 1.13141143e+00 -3.00346494e-01
-3.06473821e-01 3.49858075e-01 -4.55898732e-01 3.32929581e-01
1.05417383e+00 -2.43155524e-01 -7.24681854e-01 -3.97531122e-01
-6.00946285e-02 -1.34834304e-01 6.39022112e-01 -1.06829643e+00
5.10927379e-01 -6.61501050e-01 -5.40378764e-02 -9.89163280e-01
2.08128527e-01 -9.74625230e-01 1.08594932e-01 6.26851916e-02
-3.16067904e-01 3.03887427e-01 3.26539963e-01 9.86860335e-01
-1.80702239e-01 8.15153658e-01 4.79185134e-01 1.79671064e-01
-8.29969943e-01 1.22709620e+00 -5.89829981e-01 1.28878608e-01
1.00571573e+00 -3.00760984e-01 -6.02377892e-01 -3.62972379e-01
-4.57427144e-01 7.99345315e-01 5.53211607e-02 7.02330232e-01
3.14020663e-01 -1.73086894e+00 -6.30963922e-01 3.06896210e-01
5.03675938e-01 -1.99490100e-01 6.91899896e-01 1.09137297e+00
-3.03126685e-02 7.10333228e-01 -4.78123426e-02 -6.38832927e-01
-7.93138027e-01 6.92828655e-01 2.73493022e-01 -3.10424775e-01
-7.07069397e-01 1.75320536e-01 2.46769801e-01 -6.26518071e-01
6.51711822e-02 -3.10705066e-01 -1.03384621e-01 -1.49121091e-01
3.38843286e-01 7.37348437e-01 -1.53096169e-01 -1.22831535e+00
-4.13123786e-01 6.23135746e-01 2.73617595e-01 2.02254832e-01
1.29996264e+00 -5.30687451e-01 2.49917775e-01 5.83453715e-01
1.26891708e+00 -2.00765401e-01 -1.31219709e+00 -8.78913164e-01
-3.66233699e-02 -6.81747317e-01 5.54077253e-02 -6.15409553e-01
-1.52371061e+00 8.82506371e-01 3.23934466e-01 2.37089202e-01
1.10177314e+00 -1.27173916e-01 1.21882856e+00 1.44075140e-01
2.00235963e-01 -9.93752182e-01 -1.29017562e-01 6.80868685e-01
5.07704496e-01 -1.50304854e+00 -2.35207960e-01 -7.55716145e-01
-7.00931549e-01 7.39527345e-01 8.65755320e-01 5.15695922e-02
1.22415233e+00 1.42648267e-02 -1.18536644e-01 -2.26652145e-01
-9.64500129e-01 -7.84780025e-01 4.62686330e-01 5.98695517e-01
2.88879961e-01 2.49576837e-01 1.46754920e-01 2.52283812e-01
8.25320184e-02 -1.26996726e-01 -2.35546213e-02 5.86654663e-01
-6.58108294e-02 -1.03834498e+00 9.83781740e-02 5.54118097e-01
1.29670620e-01 1.01463437e-01 1.55670613e-01 9.05858338e-01
2.44414210e-01 1.01080906e+00 5.26150882e-01 -5.20336866e-01
5.06233692e-01 -1.61366314e-01 -2.07892552e-01 -1.74656823e-01
-3.12912077e-01 -3.59962314e-01 9.66632888e-02 -9.06621635e-01
-4.62688953e-01 -7.62871623e-01 -1.03061640e+00 -7.82447219e-01
3.63703489e-01 5.69250260e-04 2.51155734e-01 8.96972120e-01
7.46553659e-01 7.09599495e-01 9.65246379e-01 -5.23428857e-01
2.22597599e-01 -7.77891636e-01 -5.73127747e-01 6.78472161e-01
6.40314460e-01 -9.26830709e-01 -2.24545121e-01 -3.01546723e-01] | [6.482376575469971, 2.048733711242676] |
1ae15080-94e4-4283-8491-64a7afdbe0b7 | 3d-point-positional-encoding-for-multi-camera | 2211.14710 | null | https://arxiv.org/abs/2211.14710v2 | https://arxiv.org/pdf/2211.14710v2.pdf | 3DPPE: 3D Point Positional Encoding for Multi-Camera 3D Object Detection Transformers | Transformer-based methods have swept the benchmarks on 2D and 3D detection on images. Because tokenization before the attention mechanism drops the spatial information, positional encoding becomes critical for those methods. Recent works found that encodings based on samples of the 3D viewing rays can significantly improve the quality of multi-camera 3D object detection. We hypothesize that 3D point locations can provide more information than rays. Therefore, we introduce 3D point positional encoding, 3DPPE, to the 3D detection Transformer decoder. Although 3D measurements are not available at the inference time of monocular 3D object detection, 3DPPE uses predicted depth to approximate the real point positions. Our hybriddepth module combines direct and categorical depth to estimate the refined depth of each pixel. Despite the approximation, 3DPPE achieves 46.0 mAP and 51.4 NDS on the competitive nuScenes dataset, significantly outperforming encodings based on ray samples. We will make the codes available for further investigation. | ['Yifan Liu', 'Jiajun Deng', 'Fisher Yu', 'Changyong Shu'] | 2022-11-27 | null | null | null | null | ['monocular-3d-object-detection'] | ['computer-vision'] | [ 2.55608469e-01 -9.89159346e-02 -3.98933113e-01 -3.09647083e-01
-1.09983253e+00 -7.43213594e-01 6.19756341e-01 9.16193873e-02
-3.98106575e-01 6.94640949e-02 1.62735879e-01 -3.66976231e-01
5.21920502e-01 -9.85173583e-01 -1.25626111e+00 -5.58072329e-01
5.02112582e-02 5.33722103e-01 9.44639146e-01 1.50383338e-01
6.11728311e-01 7.61390090e-01 -1.69981968e+00 5.53069592e-01
3.07931870e-01 1.32611835e+00 3.90119046e-01 1.06859922e+00
-2.51755714e-01 5.57551324e-01 -4.36127156e-01 -8.87225494e-02
7.38569796e-01 -1.83201432e-02 -2.53559351e-01 1.17940456e-01
9.98037875e-01 -1.23504901e+00 -4.44158643e-01 9.57789123e-01
4.86832261e-01 -4.43151481e-02 7.25441337e-01 -9.40606296e-01
-3.21247190e-01 2.58520395e-01 -1.00867951e+00 2.25405604e-01
6.12849414e-01 2.26941004e-01 1.03574109e+00 -1.22682905e+00
5.94574034e-01 1.44495738e+00 6.65353298e-01 2.23102868e-01
-7.78446674e-01 -5.83717823e-01 2.36567542e-01 2.50755161e-01
-1.40000558e+00 -3.62305522e-01 6.01348579e-01 -2.67355353e-01
1.36950839e+00 2.21435890e-01 6.03226900e-01 7.05275118e-01
6.93882406e-02 7.61253595e-01 1.00226831e+00 -3.57150286e-01
1.27009496e-01 -5.35466187e-02 -1.65372528e-02 6.96134031e-01
2.09608316e-01 3.92432511e-01 -8.18618953e-01 1.33488199e-03
9.44829822e-01 -9.68856812e-02 -2.41109446e-01 -3.11478525e-01
-1.09675050e+00 5.00074208e-01 7.76452065e-01 -5.03035724e-01
-2.92613149e-01 4.51147169e-01 4.11720462e-02 -1.74512696e-02
7.08917081e-01 2.62925565e-01 -3.87125522e-01 -1.84289917e-01
-9.68999267e-01 1.53250456e-01 2.74847209e-01 1.38381946e+00
8.36232603e-01 -4.90980476e-01 -3.16408426e-02 4.07552838e-01
2.45707333e-01 8.86505723e-01 -3.34599376e-01 -1.32430971e+00
5.42476237e-01 7.26226330e-01 1.19955726e-01 -8.25455606e-01
-2.46589810e-01 -2.90711015e-01 -3.05497885e-01 4.86959934e-01
4.43488628e-01 3.55564386e-01 -1.01058662e+00 1.02557278e+00
6.03263021e-01 -6.39362633e-02 -2.96955347e-01 1.14755797e+00
8.14801276e-01 9.62599635e-01 -6.74837291e-01 1.57169551e-01
1.16597509e+00 -8.14462364e-01 -8.45612437e-02 -3.64668429e-01
5.38488686e-01 -8.03367019e-01 8.10452580e-01 5.74867308e-01
-1.40118074e+00 -4.72413212e-01 -1.04368150e+00 -6.32657290e-01
-1.08886965e-01 -8.54049549e-02 3.99826050e-01 5.74361920e-01
-1.08988988e+00 1.90891817e-01 -9.99907374e-01 -1.98823232e-02
6.10078335e-01 1.93300724e-01 3.71852443e-02 -6.27159476e-01
-6.50021851e-01 8.53974164e-01 1.14720955e-01 -2.47328982e-01
-8.53427410e-01 -8.15046430e-01 -7.97578096e-01 -6.44548908e-02
5.28521597e-01 -4.98859525e-01 1.32879293e+00 -1.22943655e-01
-9.24501956e-01 1.09983528e+00 -5.29137850e-01 -3.78989190e-01
5.99778295e-01 -3.04573566e-01 3.29817891e-01 5.13665199e-01
2.22300425e-01 1.08093727e+00 6.03834331e-01 -1.22169816e+00
-1.14969289e+00 -6.33230269e-01 3.22958887e-01 4.39576030e-01
7.77926892e-02 -2.37893298e-01 -1.00935984e+00 3.66672911e-02
7.05451131e-01 -7.25654960e-01 -5.70811369e-02 6.98364854e-01
-5.65172374e-01 -1.24614790e-01 7.27875233e-01 -2.72524118e-01
5.36681771e-01 -2.20015430e+00 -5.29676070e-03 -1.40019702e-02
3.67602974e-01 -2.12650910e-01 1.12738013e-01 1.35411441e-01
4.00499433e-01 -9.00430903e-02 -5.17204963e-03 -5.82765639e-01
-2.64544636e-02 2.07779020e-01 -2.88193196e-01 6.03370726e-01
1.47506222e-01 6.21075094e-01 -8.54106426e-01 -3.97939384e-01
5.96788585e-01 4.47867125e-01 -1.03325224e+00 -2.58457474e-02
-4.46935356e-01 -3.44600752e-02 -3.13058615e-01 1.06611001e+00
1.27873456e+00 -2.85228163e-01 -3.96662414e-01 -4.53454435e-01
-4.79224861e-01 5.70071578e-01 -9.29514945e-01 1.68748176e+00
-2.40191147e-01 8.73302400e-01 -1.44457847e-01 -1.95861742e-01
5.86229861e-01 -1.68925717e-01 2.97978044e-01 -9.64586794e-01
-7.18630990e-03 2.36882061e-01 -3.49034816e-01 9.53195093e-04
8.85681331e-01 4.03038353e-01 9.85980481e-02 1.81485236e-01
-3.92064154e-01 -7.26696372e-01 -1.16574295e-01 3.27918321e-01
1.34668291e+00 2.45376736e-01 9.71425250e-02 1.29903734e-01
1.59597639e-02 4.37130958e-01 2.32071117e-01 1.02765369e+00
-1.79699957e-02 1.02692688e+00 5.26501596e-01 -1.78926796e-01
-1.43613362e+00 -1.31420946e+00 -4.04184312e-01 6.86047971e-01
5.85132658e-01 -4.28335816e-01 -4.08151001e-01 -5.07930696e-01
3.76332402e-02 8.07400525e-01 -3.51343125e-01 1.01640217e-01
-4.84935611e-01 -6.71325862e-01 2.82882035e-01 6.36217833e-01
5.93559742e-01 -2.99415682e-02 -1.11452889e+00 -1.15263291e-01
-1.71608552e-01 -1.40859771e+00 -3.55071604e-01 4.75282580e-01
-8.12622845e-01 -1.01165140e+00 -6.16733313e-01 -1.75136194e-01
5.24818122e-01 8.34497988e-01 1.15111029e+00 -3.92422639e-02
-1.91153765e-01 3.46246779e-01 -6.01944149e-01 -4.87994045e-01
-1.25983298e-01 2.14223983e-03 -3.09231848e-01 -6.13562465e-01
5.22063076e-01 -2.57708639e-01 -9.42316473e-01 3.97509128e-01
-5.08088470e-01 3.99078548e-01 3.78340244e-01 4.74553704e-01
9.20926869e-01 -6.17579371e-02 -4.39703614e-01 -4.95991707e-01
-3.96674275e-01 -1.68030411e-01 -9.81749713e-01 -4.01870430e-01
-3.66185606e-01 1.05240906e-03 -2.09949519e-02 -1.03977315e-01
-8.50398660e-01 2.21847653e-01 -3.09775412e-01 -6.10016704e-01
-9.41670686e-03 -2.37568811e-01 -2.50425898e-02 -2.11115360e-01
6.86889350e-01 2.42231742e-01 -5.80415428e-01 -4.05548275e-01
2.22348198e-01 6.51112020e-01 4.93055016e-01 -3.34248841e-01
5.39002955e-01 1.02572370e+00 2.70078480e-01 -8.71123075e-01
-1.04683244e+00 -6.96840048e-01 -4.94037449e-01 -3.96650255e-01
9.56783533e-01 -1.27253282e+00 -5.58538318e-01 4.12688762e-01
-1.67772555e+00 -3.18961561e-01 -1.62267134e-01 4.74639088e-01
-4.46801037e-01 3.01107347e-01 -6.07063830e-01 -9.17970359e-01
1.33395061e-01 -1.41088450e+00 1.72030294e+00 -3.18860769e-01
1.08748682e-01 -3.36400539e-01 -4.01830018e-01 1.94428250e-01
-7.54689947e-02 1.14359491e-01 7.29748130e-01 1.16556264e-01
-1.44218969e+00 -3.28126132e-01 -7.00144529e-01 1.24376617e-01
-2.11662099e-01 -1.50567427e-01 -1.45407677e+00 1.53560266e-01
2.82883570e-02 -1.04666777e-01 1.10409844e+00 5.37877977e-01
1.19786978e+00 1.05429694e-01 -4.40266758e-01 9.92544472e-01
1.44956326e+00 1.28772810e-01 7.80227900e-01 3.04863870e-01
8.17796588e-01 2.59641767e-01 7.71147728e-01 6.01511955e-01
5.87674499e-01 9.20778453e-01 1.03516686e+00 6.41307533e-02
-5.51595509e-01 -3.01474810e-01 2.36201063e-01 2.42120460e-01
-9.09668580e-02 -6.66083455e-01 -9.85784113e-01 2.48458520e-01
-1.26616263e+00 -7.37379014e-01 -6.75008178e-01 2.19541240e+00
6.08056605e-01 6.08825028e-01 -1.00121416e-01 2.04468608e-01
4.74948078e-01 2.08223850e-01 -7.19280362e-01 -3.47438417e-02
-3.16285312e-01 -1.41301349e-01 1.32152534e+00 8.23142350e-01
-7.31516719e-01 8.09296608e-01 6.07720995e+00 6.37762249e-01
-6.97634399e-01 1.10400476e-01 5.19658744e-01 -5.80869496e-01
-3.57666850e-01 1.94847696e-02 -1.49476242e+00 4.16249931e-01
4.34073716e-01 5.33804119e-01 1.87558800e-01 7.43243098e-01
1.28029436e-01 -6.88722610e-01 -1.31702232e+00 1.34561670e+00
9.88561586e-02 -1.31355178e+00 2.05508694e-01 3.59096646e-01
7.39860058e-01 5.98129928e-01 1.79837465e-01 -2.66985148e-01
7.27561414e-02 -4.93072629e-01 1.13330114e+00 2.57754952e-01
9.90391791e-01 -6.99269712e-01 4.06177938e-01 5.78208029e-01
-1.01757717e+00 1.11823924e-01 -5.93742311e-01 -1.64865553e-01
3.15917999e-01 7.83543050e-01 -1.13238084e+00 2.22552642e-01
8.74844074e-01 7.82351255e-01 -6.44318104e-01 1.12030709e+00
-2.17519358e-01 2.83699125e-01 -8.01804483e-01 -5.46582304e-02
4.95489776e-01 2.98482031e-01 7.65097976e-01 9.94549513e-01
5.76223791e-01 1.76262349e-01 -1.14660226e-01 8.14125896e-01
-1.74443964e-02 -5.70164800e-01 -5.99532962e-01 4.40287292e-01
6.60648525e-01 6.00929558e-01 -9.02443409e-01 -3.24266940e-01
-4.46632266e-01 1.17194581e+00 -6.38329536e-02 6.99743032e-02
-8.95680070e-01 -1.78375363e-01 6.61767662e-01 4.94237781e-01
5.53390086e-01 -4.90835428e-01 -5.53993881e-01 -9.80755568e-01
2.60274522e-02 -5.29493153e-01 3.46035250e-02 -1.25282156e+00
-7.83520401e-01 3.03784430e-01 1.46449551e-01 -1.23224282e+00
-1.89051181e-01 -9.50419486e-01 4.69268896e-02 6.55774176e-01
-1.63896215e+00 -7.14241266e-01 -5.59614718e-01 1.71867713e-01
7.21237779e-01 5.82066178e-01 3.64608377e-01 3.68189603e-01
2.80922968e-02 3.37869018e-01 2.15359293e-02 -8.96164179e-02
5.74009240e-01 -1.11467755e+00 7.80266345e-01 8.69021773e-01
9.27485432e-03 -8.05177316e-02 6.95912778e-01 -5.92994630e-01
-1.78221476e+00 -8.47591877e-01 6.36918783e-01 -9.99364734e-01
1.28264233e-01 -5.88036180e-01 -7.43946850e-01 5.22334576e-01
-2.05036610e-01 9.90408938e-03 7.13395625e-02 -2.29933560e-01
-5.28630853e-01 1.11544393e-01 -1.12870800e+00 4.58762676e-01
1.52650237e+00 -5.81727028e-01 -1.64102763e-01 3.41276407e-01
9.67498779e-01 -1.04112005e+00 -4.61652637e-01 4.10086453e-01
4.77932632e-01 -1.42671108e+00 1.46689808e+00 2.99318016e-01
7.11904526e-01 -3.92643064e-01 -7.07870424e-01 -7.79401660e-01
1.19848579e-01 1.54282749e-02 -3.40683311e-01 5.88843107e-01
2.57238686e-01 -2.63399005e-01 1.06332362e+00 4.14282650e-01
-5.18865943e-01 -5.11802018e-01 -1.03958166e+00 -6.19984567e-01
-2.75955677e-01 -9.38010335e-01 5.85430503e-01 3.68932217e-01
-6.38136923e-01 1.81351930e-01 -3.51358801e-02 6.45594597e-01
1.00758672e+00 1.75374836e-01 8.26618612e-01 -8.71057510e-01
-3.53117466e-01 -5.25089085e-01 -5.67939520e-01 -2.10923195e+00
-4.66416448e-01 -7.45667994e-01 8.24239030e-02 -1.51605344e+00
7.81170130e-02 -3.40240777e-01 2.47113824e-01 3.01289316e-02
2.10401520e-01 6.61593020e-01 7.15831220e-02 6.17189407e-02
-6.86379135e-01 3.44991595e-01 1.30768549e+00 -9.13200602e-02
2.60449313e-02 -1.90815315e-01 -2.67804623e-01 8.77517700e-01
5.24218678e-01 -4.51460123e-01 -2.64726728e-01 -1.09736562e+00
5.17643452e-01 1.10219166e-01 8.45232666e-01 -1.04570794e+00
3.80408257e-01 1.56139717e-01 8.71735811e-01 -1.69227087e+00
1.10189247e+00 -7.48212934e-01 -3.66138309e-01 3.89860094e-01
-1.65528968e-01 -1.77588850e-01 1.41623795e-01 6.46533787e-01
2.41641924e-01 -2.07824215e-01 7.73071229e-01 -3.50044101e-01
-9.02181625e-01 3.24994564e-01 -1.75049394e-01 -7.04981014e-02
7.63871014e-01 -6.49767458e-01 -3.37985814e-01 -1.79578871e-01
-2.70756006e-01 8.18493217e-02 7.96389282e-01 9.69293788e-02
1.00538683e+00 -1.06582677e+00 -5.31737745e-01 2.03627244e-01
2.38254994e-01 6.09195471e-01 1.87233314e-01 5.13670504e-01
-9.40500557e-01 4.53257412e-01 2.24198282e-01 -1.16858351e+00
-1.17671835e+00 3.34265977e-01 2.73862034e-01 5.64696789e-01
-7.42795408e-01 1.11628914e+00 4.09388900e-01 4.71781790e-02
5.17648518e-01 -8.03768516e-01 5.16313851e-01 -3.70531306e-02
6.81810439e-01 5.30990839e-01 1.76212668e-01 -2.93451816e-01
-4.72009033e-01 7.46208549e-01 -1.34420097e-01 -2.51755089e-01
1.12425697e+00 -3.45701277e-01 1.23176917e-01 5.16767561e-01
1.20764840e+00 7.18030334e-02 -1.73125434e+00 -1.95416972e-01
-4.70700830e-01 -1.08945251e+00 6.77957833e-01 -5.39419532e-01
-6.30189538e-01 1.15276468e+00 6.31156385e-01 5.51018938e-02
9.34077501e-01 3.20120096e-01 7.41799712e-01 4.87044692e-01
6.49228394e-01 -8.01733196e-01 -5.43455705e-02 6.76988780e-01
5.81119299e-01 -1.44738352e+00 1.86531693e-01 -7.39831984e-01
3.18317935e-02 1.00309002e+00 7.41662860e-01 -4.89424355e-02
3.82735372e-01 5.24285018e-01 -3.85890812e-01 -1.92568794e-01
-7.33179927e-01 -1.62061483e-01 1.58501223e-01 5.65129399e-01
8.31171200e-02 -2.49532685e-01 4.28209126e-01 -1.76501021e-01
-1.35579124e-01 -4.00336236e-01 5.57451427e-01 7.47199118e-01
-7.13886380e-01 -6.88633204e-01 -6.88712001e-01 3.68105143e-01
-5.51061146e-02 -3.30525994e-01 -2.81568497e-01 7.19464004e-01
2.36211479e-01 6.99849904e-01 5.06080151e-01 -3.26252311e-01
3.31194371e-01 -6.00611269e-01 9.41482604e-01 -7.06630886e-01
6.81789294e-02 1.36438370e-01 -7.10900500e-02 -7.74901748e-01
-1.59820110e-01 -6.65717006e-01 -1.20852029e+00 -4.36225981e-01
-4.19268548e-01 -3.59340936e-01 8.47602844e-01 3.97555768e-01
2.84830898e-01 2.47007519e-01 5.61163127e-01 -1.39349318e+00
-4.55224097e-01 -4.56470639e-01 -3.30947727e-01 -3.84246975e-01
6.49157584e-01 -6.44725561e-01 -5.68022847e-01 -2.16331929e-01] | [7.853899002075195, -2.6714186668395996] |
651f84c2-1dea-4df9-b1cf-afa0348447ca | copymtl-copy-mechanism-for-joint-extraction | 1911.10438 | null | https://arxiv.org/abs/1911.10438v2 | https://arxiv.org/pdf/1911.10438v2.pdf | CopyMTL: Copy Mechanism for Joint Extraction of Entities and Relations with Multi-Task Learning | Joint extraction of entities and relations has received significant attention due to its potential of providing higher performance for both tasks. Among existing methods, CopyRE is effective and novel, which uses a sequence-to-sequence framework and copy mechanism to directly generate the relation triplets. However, it suffers from two fatal problems. The model is extremely weak at differing the head and tail entity, resulting in inaccurate entity extraction. It also cannot predict multi-token entities (e.g. \textit{Steven Jobs}). To address these problems, we give a detailed analysis of the reasons behind the inaccurate entity extraction problem, and then propose a simple but extremely effective model structure to solve this problem. In addition, we propose a multi-task learning framework equipped with copy mechanism, called CopyMTL, to allow the model to predict multi-token entities. Experiments reveal the problems of CopyRE and show that our model achieves significant improvement over the current state-of-the-art method by 9% in NYT and 16% in WebNLG (F1 score). Our code is available at https://github.com/WindChimeRan/CopyMTL | ['Daojian Zeng', 'Ranran Haoran Zhang', 'Qianying Liu'] | 2019-11-24 | null | null | null | null | ['entity-extraction'] | ['natural-language-processing'] | [-3.29130143e-02 2.81061172e-01 -2.15073690e-01 -2.89407104e-01
-1.12105644e+00 -4.78880107e-01 4.33079392e-01 1.06989928e-01
-4.06569541e-01 1.06963491e+00 1.20376684e-01 -3.64388525e-01
-5.14130481e-03 -8.47587168e-01 -9.03365552e-01 -3.75697196e-01
6.78678136e-03 7.41102278e-01 2.06793547e-01 -2.64126986e-01
-1.06448017e-01 6.36527166e-02 -1.16001964e+00 4.02457565e-01
1.27518463e+00 6.34964049e-01 2.33698785e-01 2.40577772e-01
-5.26053667e-01 9.27129984e-01 -6.29511118e-01 -9.28079247e-01
1.12624168e-01 -2.94999748e-01 -1.13345134e+00 -4.80945021e-01
8.72728676e-02 -2.12649107e-01 -3.50194275e-01 8.65098357e-01
6.96259558e-01 -4.56714965e-02 6.39022827e-01 -1.52511406e+00
-6.70361459e-01 1.18860757e+00 -7.38531709e-01 -5.90802170e-02
4.66193706e-01 -3.19140226e-01 1.38105440e+00 -1.13384497e+00
6.49245381e-01 1.15864766e+00 6.67545855e-01 3.37886930e-01
-7.77123094e-01 -9.91640210e-01 1.30622778e-02 5.28531790e-01
-1.55182576e+00 -4.42755997e-01 3.01366746e-01 -2.02675983e-01
1.30194688e+00 2.83092380e-01 1.92490608e-01 1.09437096e+00
1.07312456e-01 1.10890090e+00 9.53692853e-01 -5.12620151e-01
-2.73717940e-01 -1.41953304e-01 2.39353627e-01 7.08148181e-01
3.63697499e-01 -2.33976841e-01 -5.25017500e-01 -1.77422747e-01
3.85380238e-01 -2.43025720e-01 -2.93877795e-02 8.69295076e-02
-1.23192847e+00 5.05811989e-01 3.44456345e-01 3.18104267e-01
-3.13787133e-01 3.06761730e-02 2.75581419e-01 8.56273994e-02
5.06255567e-01 3.37433815e-01 -7.65153348e-01 -2.51413047e-01
-6.11873627e-01 4.30781692e-01 8.93781006e-01 1.42905271e+00
6.54592335e-01 -3.70807141e-01 -3.62637013e-01 9.67800021e-01
6.82475343e-02 2.15330184e-01 2.44282529e-01 -5.96982002e-01
1.05285561e+00 5.50100684e-01 1.38002679e-01 -9.00530815e-01
-4.52567935e-01 -3.28510731e-01 -8.85888934e-01 -4.60536808e-01
2.51082242e-01 -4.68664736e-01 -7.68716037e-01 1.65508270e+00
4.59235191e-01 3.56984884e-01 8.59513432e-02 5.54920495e-01
1.19660521e+00 8.02420914e-01 1.98937699e-01 -1.17339030e-01
1.57285583e+00 -1.27144241e+00 -1.00363660e+00 -3.87373716e-01
9.30225432e-01 -8.34557831e-01 6.69545531e-01 -5.36653511e-02
-9.81981814e-01 -3.41565192e-01 -7.04916239e-01 -3.07483286e-01
-6.68552399e-01 4.07678276e-01 8.17786574e-01 2.37070754e-01
-6.04280710e-01 5.05999625e-01 -7.48459220e-01 -3.27293903e-01
1.97461203e-01 3.17248732e-01 -5.49691021e-01 -9.90817919e-02
-1.64455938e+00 1.11985457e+00 8.66323352e-01 2.55459160e-01
-1.23830795e-01 -6.34821475e-01 -1.10504222e+00 2.08010629e-01
8.76622975e-01 -7.80057549e-01 1.33980715e+00 -1.97924793e-01
-1.17008340e+00 4.67421979e-01 -5.88230312e-01 -3.78463924e-01
4.96938974e-01 -6.79017127e-01 -5.32019258e-01 -2.01367810e-01
3.50814044e-01 3.96678835e-01 2.79485554e-01 -1.02711785e+00
-8.55240464e-01 -8.55065659e-02 -1.73744962e-01 1.59734234e-01
-2.70211101e-01 3.53616536e-01 -7.78108358e-01 -6.22552991e-01
-6.12558313e-02 -1.02717233e+00 -1.64608344e-01 -6.87661409e-01
-8.94749463e-01 -7.16539145e-01 5.30418932e-01 -8.54772031e-01
1.63392341e+00 -1.62143660e+00 -4.26515676e-02 -6.58142269e-02
3.20136636e-01 5.48366547e-01 -1.51653513e-01 9.30144250e-01
-2.38985479e-01 4.79526818e-01 -1.65968835e-01 -2.56183386e-01
5.92153072e-02 7.30580986e-02 -1.21552706e-01 -8.22294578e-02
4.35214102e-01 1.25868261e+00 -1.01958942e+00 -7.95274794e-01
-1.23126730e-01 3.56584221e-01 -1.43024579e-01 2.45309427e-01
3.86142725e-04 3.01358588e-02 -6.88779414e-01 6.81582510e-01
6.89760327e-01 -4.35278326e-01 4.92947578e-01 -2.54007429e-01
-1.25476182e-01 7.16559291e-01 -1.12184930e+00 1.37098098e+00
-1.94143936e-01 1.88627213e-01 -3.28425735e-01 -7.80051172e-01
7.40654767e-01 4.96014595e-01 4.51387316e-01 -3.78844082e-01
1.98736358e-02 4.97905880e-01 -2.06693225e-02 -7.56093740e-01
6.69754744e-01 1.18668824e-01 -2.51390427e-01 2.73825198e-01
-2.44269259e-02 3.05054396e-01 4.75346059e-01 4.53408211e-01
1.29169738e+00 4.69285339e-01 5.12584865e-01 2.69306809e-01
2.91532964e-01 -1.24877006e-01 1.02783620e+00 6.46416306e-01
1.53625757e-01 3.40159208e-01 4.51508105e-01 -2.72980392e-01
-1.00460172e+00 -6.48609936e-01 1.34777859e-01 9.08258021e-01
1.53643951e-01 -9.02680755e-01 -5.60100019e-01 -9.54137802e-01
3.97626795e-02 8.10626984e-01 -2.71877944e-01 -2.90550776e-02
-8.31224322e-01 -9.94850039e-01 7.18953848e-01 6.28426373e-01
6.43538058e-01 -1.08803689e+00 -1.65463034e-02 4.31154549e-01
-9.73441780e-01 -1.57706857e+00 -3.68092984e-01 9.91179571e-02
-4.08682466e-01 -8.93280387e-01 -5.58252990e-01 -8.41579080e-01
3.64453584e-01 1.09746821e-01 1.27293217e+00 2.55247001e-02
-5.37537187e-02 -4.45682913e-01 -6.49755836e-01 -4.04070407e-01
-2.98797816e-01 6.42035902e-01 -1.40059248e-01 -3.07658762e-01
5.76062620e-01 -3.66113991e-01 -2.38927633e-01 1.65271416e-01
-6.78431571e-01 3.20344537e-01 9.12249386e-01 8.17786872e-01
4.27225113e-01 1.66848004e-01 7.43039966e-01 -1.39810288e+00
6.07795060e-01 -6.73181653e-01 -1.93147689e-01 6.68857813e-01
-7.29814172e-01 6.36755303e-02 4.98526663e-01 -1.07438348e-01
-1.08000708e+00 3.52049130e-03 -3.49158764e-01 -9.22464654e-02
7.47813936e-03 8.46375108e-01 -1.88972220e-01 4.23640639e-01
1.67711452e-01 1.30798504e-01 -4.87199008e-01 -7.26376235e-01
3.02786350e-01 8.11327279e-01 4.21264857e-01 -7.37036407e-01
9.75234210e-01 -1.60190582e-01 -1.72153801e-01 -4.03894871e-01
-1.14000428e+00 -4.71247792e-01 -7.20224738e-01 1.18661560e-01
6.83713675e-01 -1.10864234e+00 -6.49026811e-01 5.95397830e-01
-1.47404456e+00 -7.06561059e-02 2.01915741e-01 3.58626038e-01
-1.30834982e-01 4.34210569e-01 -1.06293082e+00 -8.29279125e-01
-6.09255075e-01 -7.89047003e-01 1.15327704e+00 1.59777150e-01
-2.06102505e-01 -6.37464106e-01 -1.94578990e-01 6.30022585e-01
1.03635460e-01 3.38165611e-01 1.17017424e+00 -8.88720810e-01
-6.56230807e-01 -1.53113633e-01 -4.18826252e-01 1.82795841e-02
2.06505120e-01 6.19397387e-02 -6.13648832e-01 -7.75475651e-02
-6.01124346e-01 -3.25325906e-01 7.58660674e-01 -6.61741644e-02
1.01787138e+00 -5.17257571e-01 -8.14878345e-01 3.61630321e-01
1.23455656e+00 7.10590929e-02 7.03575432e-01 4.60087717e-01
9.42206800e-01 6.69703305e-01 9.37655747e-01 3.24291855e-01
1.04348922e+00 8.36068571e-01 1.04354464e-01 -2.49905080e-01
-1.34590223e-01 -4.52099621e-01 2.45851740e-01 1.09272182e+00
-1.63503617e-01 -4.62452590e-01 -1.05611157e+00 6.49718285e-01
-2.23935056e+00 -9.41178501e-01 -4.80639726e-01 1.83197522e+00
1.18283534e+00 1.20359935e-01 5.46675287e-02 -1.05048448e-01
8.44475210e-01 -4.35206741e-02 -3.08033377e-01 -1.55825421e-01
-1.43940851e-01 1.49798229e-01 4.38750386e-01 3.39315444e-01
-1.16353536e+00 1.41873336e+00 5.88901567e+00 1.16224825e+00
-5.40865600e-01 1.05092442e-02 3.30273002e-01 1.64195627e-01
-1.84535950e-01 9.46733505e-02 -1.36200392e+00 6.04446054e-01
8.65885913e-01 -4.04547006e-01 6.37698621e-02 6.13012850e-01
5.44271804e-02 1.77084923e-01 -9.31615472e-01 7.57573187e-01
-1.45117536e-01 -1.12171912e+00 7.48659372e-02 5.54865934e-02
4.82250392e-01 2.15056036e-02 -3.82958889e-01 7.27782190e-01
4.72265065e-01 -9.25913393e-01 5.33988893e-01 2.54449338e-01
7.40835845e-01 -8.35683167e-01 1.09227002e+00 5.44406176e-01
-1.49219596e+00 1.33461133e-01 -2.64982790e-01 -1.68745685e-02
3.73468578e-01 9.23721910e-01 -1.02780235e+00 1.19182718e+00
6.61667526e-01 6.83265626e-01 -4.49673831e-01 9.41397846e-01
-4.80095178e-01 4.99843419e-01 -3.73701572e-01 3.08068600e-02
1.11394800e-01 -1.73260406e-01 3.54297251e-01 1.50379026e+00
5.32733202e-01 2.53299981e-01 3.34881276e-01 6.64544940e-01
-4.30121243e-01 3.01081270e-01 -5.97271204e-01 -9.25708115e-02
9.01590884e-01 1.42516840e+00 -4.20072466e-01 -5.10631263e-01
-5.29666722e-01 1.01441598e+00 8.52115631e-01 3.14533770e-01
-9.46472287e-01 -8.72999072e-01 3.83520186e-01 -1.29670694e-01
4.33748573e-01 -2.34295338e-01 -8.50876868e-02 -1.32651508e+00
3.49859387e-01 -9.49240923e-01 3.94960433e-01 -6.71438992e-01
-1.28524399e+00 7.77795672e-01 4.87262122e-02 -9.64624941e-01
-4.01093036e-01 -3.15718085e-01 -3.62677276e-01 9.54598308e-01
-1.71285594e+00 -1.37289858e+00 -2.11525131e-02 3.13689858e-01
4.34795707e-01 3.33471708e-02 6.84227705e-01 6.49260104e-01
-9.34702873e-01 8.67626846e-01 -1.43319825e-02 5.36643445e-01
9.16845560e-01 -1.30770886e+00 8.74784827e-01 9.25718546e-01
1.68233871e-01 7.25692928e-01 5.53197563e-01 -9.14531589e-01
-1.26626599e+00 -1.20124209e+00 1.84177125e+00 -5.94766259e-01
6.54535294e-01 -3.81051004e-01 -8.68897319e-01 9.75744963e-01
1.83888182e-01 -2.82017380e-01 7.26093829e-01 4.22236890e-01
-3.68330091e-01 1.28185481e-01 -8.30806196e-01 7.34323680e-01
1.18334687e+00 -2.55466074e-01 -6.06658161e-01 5.34174204e-01
9.84546900e-01 -6.19590521e-01 -1.08569860e+00 7.57243991e-01
5.30777276e-01 -7.10956335e-01 8.36428583e-01 -5.58766782e-01
6.70468628e-01 -2.44053200e-01 2.18928903e-01 -1.25607347e+00
-4.62170660e-01 -6.61733210e-01 -5.63167691e-01 1.74695206e+00
9.78834033e-01 -7.23078609e-01 5.76905370e-01 6.79244936e-01
-3.20198052e-02 -1.09086967e+00 -7.02809155e-01 -1.04565382e+00
-7.23828152e-02 -2.43283048e-01 9.27151382e-01 1.08444846e+00
9.15194973e-02 8.61846745e-01 -6.20211542e-01 2.26195171e-01
3.98238450e-01 3.56922776e-01 6.98096097e-01 -9.92355764e-01
-2.23124474e-01 -1.33627176e-01 1.77434683e-01 -1.12899935e+00
3.70139986e-01 -9.40542459e-01 6.97097331e-02 -1.88140190e+00
3.52317691e-01 -6.30087793e-01 -2.32823581e-01 1.00906420e+00
-6.95214152e-01 -6.42573163e-02 3.14401060e-01 2.98553348e-01
-6.89992785e-01 5.71406305e-01 1.12072718e+00 1.06919065e-01
6.64954334e-02 7.00358525e-02 -8.22534204e-01 4.67989385e-01
7.53152847e-01 -7.61638284e-01 -9.72246975e-02 -6.25828087e-01
4.08072919e-01 3.93724591e-02 -5.71574941e-02 -5.60538709e-01
3.44900519e-01 3.72998673e-03 2.15277702e-01 -7.98393846e-01
2.37014160e-01 -6.07595026e-01 1.43526390e-01 1.31627142e-01
-3.22591037e-01 3.09709340e-01 -4.62702336e-03 3.93536478e-01
-3.03732902e-01 -2.51870453e-01 1.05445944e-01 -1.28872767e-01
-6.54448092e-01 4.73604262e-01 -7.97107443e-02 1.50649160e-01
8.66571307e-01 2.40447462e-01 -5.52274942e-01 -2.18773633e-01
-3.81945282e-01 6.72902942e-01 -1.00274663e-02 5.85013270e-01
3.62534851e-01 -1.38075686e+00 -1.07433903e+00 -2.42180884e-01
1.39831796e-01 1.32443488e-01 -1.44925015e-02 7.62002945e-01
-2.22465530e-01 5.30380428e-01 2.45309249e-01 -4.65155020e-02
-1.38115418e+00 4.32271153e-01 7.83768483e-03 -8.03178787e-01
-5.48312783e-01 8.85802984e-01 -1.43100038e-01 -6.60480380e-01
-3.63568217e-02 -1.93621382e-01 -2.00004309e-01 4.93031032e-02
3.10306430e-01 3.70310336e-01 1.94556937e-01 -5.10058522e-01
-5.05720437e-01 3.59508961e-01 -5.49275219e-01 1.89061224e-01
1.32980168e+00 -1.35241121e-01 -3.15015674e-01 2.02158943e-01
1.02412629e+00 5.30825369e-02 -6.35941148e-01 -4.31939125e-01
4.69209284e-01 -2.05194518e-01 -3.89058292e-01 -9.80695426e-01
-7.76164472e-01 5.00632942e-01 -1.40072748e-01 2.70927459e-01
8.07545602e-01 8.35790671e-03 1.46637380e+00 5.92205405e-01
5.35445809e-01 -9.92915571e-01 -4.22312796e-01 8.17745447e-01
6.21117651e-01 -1.29036641e+00 1.32756397e-01 -1.02185392e+00
-6.63099051e-01 7.87502646e-01 7.94604659e-01 2.50984311e-01
3.97429198e-01 4.35240328e-01 -2.58268472e-02 -1.40489474e-01
-9.46276426e-01 -4.11445558e-01 2.37824410e-01 3.52584302e-01
7.57934570e-01 1.23198375e-01 -5.56392550e-01 7.47622430e-01
-3.37881744e-01 -2.46575437e-02 2.31504366e-01 9.42858279e-01
-1.80517584e-01 -1.61918390e+00 7.46973231e-02 5.19016862e-01
-7.26137161e-01 -5.47150910e-01 -4.44729269e-01 8.56076598e-01
4.97653969e-02 9.50878263e-01 -4.94688928e-01 -5.25475562e-01
4.41886842e-01 3.08844745e-01 2.42348284e-01 -7.38247454e-01
-6.69592917e-01 3.98310758e-02 7.05125093e-01 -4.14149851e-01
-2.77893007e-01 -6.28555536e-01 -1.31739604e+00 -3.78712535e-01
-6.34344757e-01 4.11390334e-01 3.59759957e-01 1.11155999e+00
6.50707126e-01 5.98113000e-01 4.62923884e-01 -3.16505104e-01
-4.46216881e-01 -1.14557779e+00 -4.66291994e-01 3.07548672e-01
2.60338001e-02 -5.78752160e-01 -4.43967758e-03 -1.32107705e-01] | [9.429718971252441, 8.76710319519043] |
7c0cf51a-917b-4b86-b10e-828f2cd29d8f | weakly-supervised-fine-grained-image-1 | null | null | http://openaccess.thecvf.com/content_CVPR_2020/html/Wang_Weakly_Supervised_Fine-Grained_Image_Classification_via_Guassian_Mixture_Model_Oriented_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Wang_Weakly_Supervised_Fine-Grained_Image_Classification_via_Guassian_Mixture_Model_Oriented_CVPR_2020_paper.pdf | Weakly Supervised Fine-Grained Image Classification via Guassian Mixture Model Oriented Discriminative Learning | Existing weakly supervised fine-grained image recognition (WFGIR) methods usually pick out the discriminative regions from the high-level feature maps directly. We discover that due to the operation of stacking local receptive filed, Convolutional Neural Network causes the discriminative region diffusion in high-level feature maps, which leads to inaccurate discriminative region localization. In this paper, we propose an end-to-end Discriminative Feature-oriented Gaussian Mixture Model (DF-GMM), to address the problem of discriminative region diffusion and find better fine-grained details. Specifically, DF-GMM consists of 1) a low-rank representation mechanism (LRM), which learns a set of low-rank discriminative bases by Gaussian Mixture Model (GMM) in high-level semantic feature maps to improve discriminative ability of feature representation, 2) a low-rank representation reorganization mechanism (LR ^2 M) which resumes the space information corresponding to low-rank discriminative bases to reconstruct the low-rank feature maps. It alleviates the discriminative region diffusion problem and locate discriminative regions more precisely. Extensive experiments verify that DF-GMM yields the best performance under the same settings with the most competitive approaches, in CUB-Bird, Stanford-Cars datasets, and FGVC Aircraft.
| [' Zezhou Li', ' Jianjun Li', ' Haojie Li', ' Shuhui Yang', ' Shijie Wang', 'Zhihui Wang'] | 2020-06-01 | null | null | null | cvpr-2020-6 | ['fine-grained-image-recognition'] | ['computer-vision'] | [-4.11216795e-01 -4.73838240e-01 -4.16532308e-01 -3.63854468e-01
-9.43772972e-01 -5.71109235e-01 5.33688903e-01 -4.76903558e-01
-1.69525698e-01 3.84921521e-01 5.40994346e-01 2.20974430e-01
-2.87489146e-01 -6.98481023e-01 -7.13191330e-01 -8.10886741e-01
8.17755163e-02 2.66023189e-01 5.35331666e-01 -6.24837242e-02
4.63220179e-01 9.43353713e-01 -1.46615493e+00 8.71592462e-01
5.11250138e-01 1.15058517e+00 5.83596706e-01 3.73350382e-01
-2.90472627e-01 8.56241643e-01 -4.13520604e-01 2.55554140e-01
3.17885011e-01 -8.87527838e-02 -7.73609519e-01 2.00721532e-01
5.88789940e-01 -5.57571709e-01 -6.90311491e-01 1.17717481e+00
1.80247217e-01 4.20171708e-01 8.19519401e-01 -8.81989539e-01
-7.93791533e-01 -1.29127391e-02 -7.96387494e-01 5.33031285e-01
-7.94948563e-02 -6.71403259e-02 8.20597112e-01 -1.43069732e+00
6.08133018e-01 1.67403102e+00 4.81436849e-01 3.99337769e-01
-8.82887602e-01 -6.68841422e-01 5.03020346e-01 1.70547202e-01
-1.93481767e+00 -4.22214955e-01 9.07942235e-01 -4.17350799e-01
8.42136979e-01 1.88260689e-01 1.32423133e-01 7.27612138e-01
3.77942622e-01 7.82665968e-01 1.14472973e+00 -1.17160052e-01
-6.14069253e-02 -1.05028629e-01 2.44943053e-01 9.84789133e-01
7.80575946e-02 2.38739684e-01 -6.13277495e-01 -1.86533689e-01
1.31738770e+00 6.46715999e-01 -1.10513061e-01 -4.52974021e-01
-1.20735395e+00 7.74020612e-01 7.62418747e-01 6.25554919e-01
-3.80236685e-01 1.46862373e-01 6.61767051e-02 1.11905739e-01
3.63721311e-01 -2.49361992e-02 -4.76506293e-01 2.60424346e-01
-9.99225855e-01 1.59922421e-01 3.64347905e-01 1.01507187e+00
1.19092846e+00 -5.61106801e-02 -5.21200001e-01 9.60874379e-01
7.17733562e-01 2.78719485e-01 8.31110835e-01 -7.85811245e-01
5.85217953e-01 6.28646553e-01 -5.40130585e-02 -1.23007727e+00
-2.18207642e-01 -7.30616510e-01 -8.79161000e-01 -7.12021440e-02
1.21370211e-01 3.83799136e-01 -1.00096786e+00 1.48969817e+00
2.63853818e-01 3.37438345e-01 -4.82027382e-02 1.22302198e+00
1.08785975e+00 8.64959717e-01 -1.59163609e-01 1.97900862e-01
1.39213967e+00 -1.37362731e+00 -2.49442652e-01 -1.06195353e-01
4.93973374e-01 -7.89325297e-01 1.01761222e+00 6.51447549e-02
-5.97743750e-01 -1.06772447e+00 -9.06932592e-01 -1.90682441e-01
-5.39315343e-01 4.40246135e-01 5.15800655e-01 1.21705361e-01
-9.74139929e-01 4.20635074e-01 -5.52291811e-01 -5.29981144e-02
5.92971623e-01 1.89968616e-01 -7.97356129e-01 -5.56392014e-01
-7.95802653e-01 3.59907001e-01 3.64214987e-01 3.44235539e-01
-1.42018449e+00 -5.29973626e-01 -7.54134178e-01 -2.66427640e-02
2.57921904e-01 -4.44660872e-01 5.86263895e-01 -6.98551893e-01
-1.05934370e+00 7.80386984e-01 -3.89924854e-01 5.60181439e-02
1.40098929e-01 -3.19175452e-01 -4.02842373e-01 1.98316395e-01
4.83265311e-01 7.38336265e-01 1.11260545e+00 -1.32020450e+00
-8.16370130e-01 -5.83518565e-01 -1.14395857e-01 9.80475023e-02
-1.34470696e-02 -1.07612133e-01 -7.41198659e-01 -1.11239588e+00
4.85773712e-01 -6.66415036e-01 -3.13352436e-01 -3.40660095e-01
-3.74402761e-01 -2.96734780e-01 1.03712392e+00 -5.86920023e-01
1.25973356e+00 -2.58885121e+00 1.84572041e-01 2.79165447e-01
4.08892274e-01 -1.30328685e-01 -3.94594967e-01 1.37148112e-01
-9.41836759e-02 -1.49565831e-01 9.27327052e-02 -8.00257400e-02
-2.17786267e-01 3.70540708e-01 -5.19855142e-01 6.21040046e-01
2.44280636e-01 1.11403358e+00 -7.92168558e-01 -5.88316262e-01
3.60769987e-01 2.25554258e-01 -4.79903668e-01 2.58480430e-01
2.99435258e-01 4.20353442e-01 -1.05888712e+00 9.07684624e-01
9.02745903e-01 -2.81974912e-01 -3.54856491e-01 -6.52511537e-01
-1.75933167e-01 -2.31102109e-01 -1.23525131e+00 1.92115974e+00
-2.69215494e-01 2.13789195e-01 1.09700441e-01 -9.31378484e-01
1.05719018e+00 -3.45230043e-01 2.28183910e-01 -6.41910553e-01
-2.69927293e-01 7.96054974e-02 -3.11785340e-01 -2.66799003e-01
4.79363322e-01 -2.03067437e-01 -3.85590196e-01 -5.59230745e-02
2.89347321e-01 2.40767941e-01 -3.51204515e-01 2.55393475e-01
1.02074921e+00 1.36494964e-01 1.86177567e-01 -5.67940295e-01
7.31608689e-01 4.28837910e-03 6.93135679e-01 7.47786462e-01
-2.86043316e-01 7.56624401e-01 -7.46995583e-02 -6.31354451e-01
-6.70970619e-01 -1.26040494e+00 -1.82741508e-02 1.39251840e+00
5.87608397e-01 -4.21951622e-01 -5.44674933e-01 -1.01359189e+00
1.75797403e-01 3.25554103e-01 -6.50333405e-01 -4.08892542e-01
-6.85959637e-01 -4.17685121e-01 3.45281094e-01 5.88117242e-01
6.67410433e-01 -1.01973987e+00 4.59397286e-02 1.02267161e-01
-5.61923981e-02 -8.52721453e-01 -8.24530303e-01 3.42608005e-01
-9.93863404e-01 -7.93908119e-01 -8.10521126e-01 -1.02940357e+00
9.15404558e-01 7.60869145e-01 8.98066759e-01 -1.55765623e-01
-3.76555234e-01 1.17099710e-01 -3.73181820e-01 3.99974644e-01
1.69490859e-01 -1.96634516e-01 5.98139726e-02 1.68034062e-01
3.55877429e-01 -2.88541496e-01 -6.74199224e-01 6.61776900e-01
-9.56142843e-01 -2.44226888e-01 1.12489378e+00 1.14111888e+00
1.19969988e+00 3.95258188e-01 4.69818175e-01 -7.17736244e-01
1.58000663e-01 -4.31851923e-01 -4.95457679e-01 1.77489132e-01
-2.55340725e-01 2.10045919e-01 7.62827337e-01 -2.76132315e-01
-8.81450534e-01 6.08982779e-02 -7.80215487e-02 -1.00584459e+00
-5.39833009e-01 2.37388030e-01 -4.61951613e-01 -2.41554499e-01
4.01037574e-01 6.84146583e-01 -5.04219413e-01 -9.43343282e-01
5.60767591e-01 4.79018688e-01 5.59660137e-01 -6.68776214e-01
9.69435871e-01 5.65577149e-01 -1.38109371e-01 -6.73818171e-01
-1.18635356e+00 -9.07896519e-01 -8.60788882e-01 1.33177951e-01
9.15575564e-01 -1.38716376e+00 -8.71780217e-02 2.51916945e-01
-8.02483082e-01 2.46089231e-02 -3.85792822e-01 5.23236871e-01
-4.32768196e-01 3.38977247e-01 -6.51670039e-01 -3.54017347e-01
-2.47455575e-02 -1.24827540e+00 1.62025559e+00 2.85620123e-01
2.63494134e-01 -6.44110560e-01 -6.16241172e-02 3.23588997e-01
2.43266106e-01 -1.65350828e-02 7.78415799e-01 -3.82840812e-01
-9.73231554e-01 -1.77825198e-01 -6.14290714e-01 3.78145218e-01
1.75505772e-01 -5.05908191e-01 -9.88296688e-01 -5.31225681e-01
-1.59143060e-01 -9.61771160e-02 1.31559455e+00 4.62586880e-01
1.55013168e+00 -3.60308111e-01 -6.48631036e-01 8.94780040e-01
1.39698899e+00 -6.03125617e-02 5.18344820e-01 3.76358554e-02
1.20485318e+00 3.78075957e-01 1.05917656e+00 3.81513238e-01
2.49486715e-01 6.81418657e-01 2.75631964e-01 -1.89828783e-01
-4.72285837e-01 -6.10610068e-01 4.42981809e-01 9.29770768e-01
-1.60211548e-02 1.96635604e-01 -2.84280062e-01 5.53652823e-01
-1.93073237e+00 -7.78118730e-01 2.29741350e-01 1.79183805e+00
4.38021749e-01 2.09432952e-02 7.66287148e-02 -4.65595067e-01
8.03079665e-01 4.52432036e-01 -5.08997798e-01 1.71782672e-01
-2.45710909e-01 -3.45189460e-02 5.63941717e-01 2.85225540e-01
-1.12494147e+00 1.32649577e+00 5.69515228e+00 1.63453853e+00
-8.40176284e-01 2.76758701e-01 6.80626512e-01 1.24717876e-01
-3.16553563e-01 -1.37562826e-02 -1.18066323e+00 4.65848953e-01
3.47686112e-01 5.83705068e-01 3.74272257e-01 1.20968473e+00
-6.56123832e-02 2.75885731e-01 -8.27606320e-01 1.27693641e+00
8.84086825e-03 -1.51848197e+00 5.05855858e-01 2.30138391e-01
8.11418653e-01 1.22822747e-01 1.08580254e-01 5.15958965e-01
1.46984905e-01 -1.10962307e+00 8.65926445e-01 9.13438737e-01
9.78311062e-01 -9.42532182e-01 6.34854972e-01 4.45121318e-01
-1.74481642e+00 -1.61601678e-01 -1.04794562e+00 4.23849910e-01
-3.92949313e-01 6.90574348e-01 -4.24986452e-01 6.63416862e-01
8.91001463e-01 9.34053779e-01 -6.03827715e-01 7.07184911e-01
6.61812723e-02 4.56506073e-01 -8.86563137e-02 3.42301220e-01
6.28727913e-01 -1.37601301e-01 4.06968564e-01 1.36225319e+00
1.26456618e-01 1.69430643e-01 5.30430377e-01 9.16800678e-01
-7.38254115e-02 -2.84939036e-02 -5.32586277e-01 4.58893441e-02
3.45032156e-01 1.53785753e+00 -8.11086833e-01 -3.01900864e-01
-3.92964244e-01 1.25994635e+00 4.72380638e-01 4.22016948e-01
-5.98864198e-01 -3.00460666e-01 6.33176506e-01 1.34797335e-01
6.92214906e-01 -3.48977417e-01 1.74029887e-01 -1.23473036e+00
-7.02521354e-02 -7.52445757e-01 5.12016475e-01 -4.91696686e-01
-1.51133275e+00 5.72671771e-01 -1.65279180e-01 -1.30490041e+00
-3.72275226e-02 -5.77293098e-01 -4.62976038e-01 9.83965874e-01
-1.70964384e+00 -1.44281065e+00 -3.86257768e-01 1.03512776e+00
9.20534074e-01 -4.59734946e-01 5.63717782e-01 3.48281473e-01
-4.88125801e-01 5.95214248e-01 3.22707951e-01 2.01308653e-01
6.89494431e-01 -9.43886936e-01 1.10971183e-01 7.78147042e-01
3.18742245e-01 6.95342302e-01 1.39091864e-01 -7.26083100e-01
-1.49938691e+00 -1.52891338e+00 3.40083688e-01 -4.71271127e-01
4.44401413e-01 -2.87905455e-01 -8.35370660e-01 4.09060240e-01
-5.24657726e-01 6.21833384e-01 4.36721265e-01 -3.31827924e-02
-6.51445806e-01 -3.52071583e-01 -1.18063366e+00 7.79862255e-02
1.21195853e+00 -9.39380765e-01 -7.52778530e-01 2.18774840e-01
8.70635867e-01 -4.77979966e-02 -8.42499554e-01 4.72130030e-01
4.63036656e-01 -8.22874129e-01 1.33701575e+00 -5.81101716e-01
1.12541132e-01 -9.16141212e-01 -6.67771101e-01 -9.71209049e-01
-1.13809538e+00 -1.25335798e-01 -2.73068577e-01 1.28607810e+00
-6.19863197e-02 -4.13884252e-01 6.99520469e-01 -2.65993655e-01
-3.71574849e-01 -7.98763931e-01 -9.90792096e-01 -7.96361804e-01
-1.64061978e-01 -1.84867680e-01 5.02899945e-01 8.56891990e-01
-6.92640841e-01 2.41903022e-01 -3.32161665e-01 4.85170960e-01
7.06805885e-01 6.39629006e-01 6.00442231e-01 -1.00741303e+00
-3.57386500e-01 -3.67146164e-01 -5.22143781e-01 -1.68127298e+00
2.66680747e-01 -9.06416833e-01 2.74798751e-01 -1.24860036e+00
4.81330007e-01 -5.87201536e-01 -8.72548699e-01 4.58691865e-01
-9.35568735e-02 3.90063554e-01 2.21604411e-03 6.97010756e-01
-1.10915613e+00 6.32869601e-01 1.54667747e+00 -2.17362389e-01
5.55387959e-02 -2.26430848e-01 -8.59526277e-01 6.99448109e-01
3.18477929e-01 -4.58959818e-01 -1.02404676e-01 -1.79548249e-01
-4.30862308e-01 -4.29852232e-02 5.77670217e-01 -9.47832227e-01
1.46293074e-01 -1.94598675e-01 1.12059832e+00 -1.01606059e+00
2.88517177e-01 -6.75125062e-01 -1.20860539e-01 1.75509319e-01
-1.64056510e-01 -1.04269706e-01 7.84165934e-02 8.27354014e-01
-6.13979101e-01 -1.23655081e-01 9.29527879e-01 -3.37924778e-01
-1.47571063e+00 6.36091650e-01 -1.08181164e-01 -6.84610084e-02
8.93119335e-01 -2.48711079e-01 -1.92045912e-01 8.24871212e-02
-7.17028856e-01 -7.59861395e-02 3.75372976e-01 6.49972618e-01
1.00692594e+00 -1.72641456e+00 -5.85792065e-01 5.66331327e-01
3.97276849e-01 2.28946656e-01 6.59739375e-01 5.83921909e-01
-2.41193056e-01 5.51648378e-01 -1.72956124e-01 -6.88343883e-01
-9.28606510e-01 8.28283906e-01 1.67647824e-01 -3.93115520e-01
-6.85528398e-01 1.26071692e+00 1.26493025e+00 -2.51904488e-01
-1.09418362e-01 -4.38467801e-01 -2.31373250e-01 -2.13064343e-01
6.85597837e-01 3.39452177e-01 -1.07090130e-01 -1.18038285e+00
-7.04121351e-01 1.09011304e+00 -3.81948799e-01 3.35586369e-01
1.15437949e+00 -4.71621335e-01 -1.75088763e-01 1.86217561e-01
1.61737037e+00 2.10025191e-01 -1.39162123e+00 -3.92042100e-01
-1.66325018e-01 -7.44125485e-01 4.83938873e-01 -4.97405469e-01
-1.07088518e+00 6.87033474e-01 7.35785663e-01 -3.20926398e-01
1.06381333e+00 4.58887726e-01 6.49434090e-01 3.56582820e-01
6.58698559e-01 -8.85140657e-01 3.85057181e-01 5.63110232e-01
8.91655564e-01 -1.12179387e+00 -1.19228467e-01 -3.70858043e-01
-5.93599617e-01 9.76556897e-01 6.48826838e-01 -6.80754304e-01
9.51562583e-01 8.28976557e-02 -1.73619896e-01 -4.00796831e-01
-5.72549522e-01 -2.33540684e-01 9.19354737e-01 5.62305033e-01
3.61535847e-02 1.62668675e-01 3.41365069e-01 1.07774985e+00
2.09536150e-01 -3.30286950e-01 -3.61197233e-01 6.36113644e-01
-7.98583269e-01 -8.07061315e-01 -5.91347873e-01 4.83810455e-01
-2.48197973e-01 -2.16111198e-01 -1.74425498e-01 6.29505634e-01
5.52308440e-01 7.71602094e-01 8.30256343e-02 -6.57614291e-01
2.49527797e-01 -3.59228075e-01 4.21024263e-01 -6.30973935e-01
-2.08471030e-01 4.85990077e-01 -3.17573637e-01 -1.11981153e+00
-1.08506955e-01 -4.35419619e-01 -1.34190309e+00 7.78411627e-02
-3.42387617e-01 1.80988058e-01 1.95187032e-01 1.02809370e+00
5.07804155e-01 5.57558775e-01 7.79441953e-01 -8.57497454e-01
-4.11688685e-01 -9.03451800e-01 -1.08814430e+00 3.07351053e-01
4.97258395e-01 -9.36229229e-01 -2.12714389e-01 -1.49542242e-01] | [9.657719612121582, 2.0094854831695557] |
cd725a91-fd0a-42f7-b415-460d719658d7 | translating-radiology-reports-into-plain | 2303.09038 | null | https://arxiv.org/abs/2303.09038v3 | https://arxiv.org/pdf/2303.09038v3.pdf | Translating Radiology Reports into Plain Language using ChatGPT and GPT-4 with Prompt Learning: Promising Results, Limitations, and Potential | The large language model called ChatGPT has drawn extensively attention because of its human-like expression and reasoning abilities. In this study, we investigate the feasibility of using ChatGPT in experiments on using ChatGPT to translate radiology reports into plain language for patients and healthcare providers so that they are educated for improved healthcare. Radiology reports from 62 low-dose chest CT lung cancer screening scans and 76 brain MRI metastases screening scans were collected in the first half of February for this study. According to the evaluation by radiologists, ChatGPT can successfully translate radiology reports into plain language with an average score of 4.27 in the five-point system with 0.08 places of information missing and 0.07 places of misinformation. In terms of the suggestions provided by ChatGPT, they are general relevant such as keeping following-up with doctors and closely monitoring any symptoms, and for about 37% of 138 cases in total ChatGPT offers specific suggestions based on findings in the report. ChatGPT also presents some randomness in its responses with occasionally over-simplified or neglected information, which can be mitigated using a more detailed prompt. Furthermore, ChatGPT results are compared with a newly released large model GPT-4, showing that GPT-4 can significantly improve the quality of translated reports. Our results show that it is feasible to utilize large language models in clinical education, and further efforts are needed to address limitations and maximize their potential. | ['Kyle J. Myers', 'Janardhana Ponnatapura', 'Michael E. Zapadka', 'Christopher T. Whitlow', 'Ge Wang', 'Chuang Niu', 'Josh Tan', 'Qing Lyu'] | 2023-03-16 | null | null | null | null | ['misinformation'] | ['miscellaneous'] | [-3.57379347e-01 7.66138673e-01 -4.16059822e-01 -4.89020556e-01
-1.45302343e+00 -4.46044564e-01 -1.67073030e-02 6.76070631e-01
-4.82606918e-01 8.70562732e-01 8.06147933e-01 -9.07536745e-01
-3.26075941e-01 -6.34386063e-01 -5.46689332e-01 -1.98114231e-01
1.53217822e-01 6.86421216e-01 4.53193858e-02 -1.81131735e-01
1.60514280e-01 3.67952883e-01 -3.76583308e-01 7.81530380e-01
8.32792938e-01 1.40941963e-01 5.04613042e-01 7.02051938e-01
-3.17883909e-01 1.44956565e+00 -9.14039910e-01 -4.71605182e-01
-2.76324183e-01 -4.90234911e-01 -1.05223143e+00 -1.02937251e-01
-2.38044575e-01 -6.82982564e-01 -1.54587269e-01 4.79794532e-01
8.36405635e-01 -7.55871385e-02 2.25562140e-01 -5.44225097e-01
-7.28027761e-01 8.86896193e-01 -2.73304671e-01 3.09659719e-01
1.05780959e+00 1.74761768e-02 3.47330719e-01 -5.78777909e-01
8.47992241e-01 9.62030470e-01 7.07538724e-01 6.36207342e-01
-7.08155096e-01 -6.68848515e-01 -5.00180349e-02 -4.11673523e-02
-1.23987615e+00 1.29639149e-01 9.10255387e-02 -3.08250248e-01
1.02360332e+00 7.79029310e-01 8.70525897e-01 1.05887091e+00
9.90611434e-01 1.64827168e-01 1.33411872e+00 -2.08830774e-01
1.32930517e-01 5.41445494e-01 1.43795013e-01 6.76554739e-01
1.60534143e-01 -3.61417055e-01 -3.57760549e-01 -4.96310145e-01
9.14543867e-01 -4.56310697e-02 -2.63981193e-01 6.72767282e-01
-1.26229548e+00 9.32756484e-01 4.79461789e-01 4.28278983e-01
-4.43361819e-01 -2.37650856e-01 1.55839875e-01 3.94025266e-01
4.66487139e-01 3.96756887e-01 -9.89547223e-02 -4.57725734e-01
-5.31999826e-01 -4.65127006e-02 1.10879278e+00 8.56624424e-01
6.00922853e-02 -4.13208127e-01 -4.18817520e-01 8.64602685e-01
1.31457657e-01 4.54663098e-01 5.73651969e-01 -1.12805176e+00
3.77955079e-01 3.76372665e-01 4.47910018e-02 -1.11559343e+00
-6.68222427e-01 -6.17966533e-01 -6.08997047e-01 -4.21200067e-01
-4.72270250e-02 -4.86173868e-01 -4.60661471e-01 1.28967392e+00
2.67197993e-02 -3.73993456e-01 1.91136405e-01 9.11861479e-01
1.41398156e+00 5.60810149e-01 3.52636993e-01 -3.79745275e-01
1.78263140e+00 -8.56426597e-01 -9.23312247e-01 -2.57256508e-01
1.33974087e+00 -1.20841181e+00 1.24766231e+00 3.31365705e-01
-1.26150239e+00 -9.70321521e-02 -3.57740372e-01 4.27826755e-02
2.31881082e-01 2.89864004e-01 5.04251897e-01 7.03669071e-01
-1.23903430e+00 1.15673706e-01 -1.00317872e+00 -6.52478576e-01
2.42512878e-02 2.87849575e-01 -4.60975856e-01 -4.08479869e-01
-1.04259622e+00 1.08572578e+00 -1.99790135e-01 -7.79201612e-02
-4.08059448e-01 -7.72285342e-01 -7.49973476e-01 -3.25546451e-02
2.94935763e-01 -1.01760948e+00 1.85417509e+00 -2.67623127e-01
-1.37234282e+00 5.73545039e-01 -1.71294540e-01 -1.70032501e-01
6.54835641e-01 6.82089180e-02 -3.56387138e-01 6.47787869e-01
3.97146970e-01 6.33319557e-01 -1.83636412e-01 -7.04461217e-01
-3.22797120e-01 1.18686743e-01 1.72469839e-01 2.99799889e-01
-1.47377431e-01 4.54023242e-01 -2.54197419e-01 -6.19553983e-01
2.78780162e-02 -8.79823327e-01 -6.29066527e-01 3.74106690e-02
-8.69683921e-02 1.23884268e-01 2.62579918e-01 -1.01586962e+00
1.21236348e+00 -1.83042049e+00 -6.90271437e-01 3.84393558e-02
3.69441599e-01 -2.60486663e-03 2.79464331e-02 8.73032033e-01
-1.38269067e-01 6.25259638e-01 9.70963612e-02 7.24414587e-02
-5.18767297e-01 3.71045738e-01 1.09020419e-01 1.20376550e-01
1.91074327e-01 1.01434159e+00 -8.84923339e-01 -7.48123825e-01
-7.70877004e-02 5.51623821e-01 -5.67193449e-01 8.57980698e-02
3.23936790e-01 7.10338712e-01 -8.06907773e-01 3.48034501e-01
4.22645390e-01 -8.42157543e-01 2.12511823e-01 3.81370872e-01
-1.96804225e-01 8.43216777e-01 -4.33343619e-01 1.32192743e+00
-6.12321854e-01 2.61174649e-01 -8.19361061e-02 -3.81449461e-01
6.98290527e-01 7.82659411e-01 5.19041300e-01 -6.34718716e-01
4.23937887e-02 6.82056174e-02 2.45627701e-01 -8.39029610e-01
2.45901346e-01 -4.28907901e-01 -1.16458341e-01 9.69160020e-01
-5.85945606e-01 -5.02709389e-01 -3.62818353e-02 6.62205040e-01
1.37652969e+00 -6.60373986e-01 5.00382185e-01 -8.04821309e-03
2.80259460e-01 4.20927376e-01 1.65847361e-01 1.09139848e+00
6.79884031e-02 6.59396410e-01 5.52816391e-01 -3.53288241e-02
-4.46279436e-01 -4.94406670e-01 1.14338547e-02 6.19340599e-01
-5.34741044e-01 -5.64041376e-01 -6.41705930e-01 -6.64436877e-01
-5.73792458e-01 1.22151601e+00 -3.21881294e-01 -2.12547854e-02
-3.90616804e-01 -4.97605056e-01 4.72960323e-01 5.74927866e-01
1.60822362e-01 -9.24200296e-01 -7.52197742e-01 4.47016835e-01
-6.86251521e-01 -1.32614875e+00 -6.72263324e-01 -5.10223210e-02
-1.02052701e+00 -7.83624768e-01 -1.18807781e+00 -8.19435775e-01
9.66667473e-01 3.62926751e-01 8.75776649e-01 4.06694800e-01
-8.10749233e-02 9.37447667e-01 -5.68117678e-01 -4.07203943e-01
-9.02813315e-01 -1.39439330e-01 -4.99559939e-01 -1.05430007e+00
9.46674049e-02 -4.50042225e-02 -4.39723194e-01 3.40267777e-01
-7.75081038e-01 4.44787651e-01 9.77936208e-01 7.16458678e-01
2.40840435e-01 -3.96185428e-01 4.12497312e-01 -1.22209704e+00
1.17918885e+00 -6.29306316e-01 1.53246626e-01 1.45202950e-01
-4.03128296e-01 -2.88276136e-01 3.60184669e-01 -3.96147490e-01
-1.14586222e+00 -6.75052524e-01 -5.31806409e-01 3.44220579e-01
-9.85129476e-02 1.11868513e+00 8.31561446e-01 -4.41245615e-01
7.25288749e-01 1.33706555e-01 1.31461620e-01 -8.56058970e-02
-3.48977745e-01 6.95971549e-01 8.79053771e-02 -4.60257530e-01
2.75239825e-01 1.15693212e-01 -5.11275947e-01 -7.61895955e-01
-8.32743526e-01 -3.59443307e-01 2.26597205e-01 -2.48420879e-01
6.36607111e-01 -1.00382245e+00 -8.02988768e-01 -4.04289305e-01
-1.12540674e+00 -2.33356267e-01 -2.41097227e-01 1.25090241e+00
-2.82730818e-01 4.57374036e-01 -1.21751678e+00 -7.11419821e-01
-3.46197426e-01 -1.38109565e+00 8.26886654e-01 1.38458326e-01
-8.12151432e-01 -1.05434251e+00 -1.79891065e-01 6.24193430e-01
5.28600752e-01 -1.68331489e-01 1.17139137e+00 -7.25444436e-01
-3.44412595e-01 -4.18110847e-01 -2.53004044e-01 -3.21735404e-02
3.32817823e-01 -3.49337727e-01 -3.29832494e-01 -4.06801403e-02
5.43369949e-01 -2.82903910e-01 -3.39021459e-02 4.43529248e-01
1.05897033e+00 -4.93707359e-01 -2.02154040e-01 6.60696477e-02
9.44258809e-01 4.86534536e-01 3.93416554e-01 2.70393014e-01
8.87541473e-02 5.42113006e-01 7.93553352e-01 4.92238492e-01
6.74767315e-01 1.88574493e-01 -5.47280982e-02 -2.38914967e-01
-1.45177051e-01 -1.55731097e-01 2.26458848e-01 1.48050988e+00
7.21279830e-02 -1.03794686e-01 -1.23040462e+00 1.93468541e-01
-1.42382896e+00 -4.59100604e-01 -2.82141954e-01 1.77393436e+00
9.51324165e-01 3.09956551e-01 -4.01290476e-01 -5.24008274e-01
2.93578357e-01 -4.13385898e-01 -5.06151281e-02 -7.31275916e-01
2.00085118e-01 3.47693741e-01 3.71773422e-01 5.15236974e-01
-1.50839016e-01 3.35347891e-01 7.03376913e+00 4.69884902e-01
-1.18104804e+00 2.89583862e-01 1.01446748e+00 1.16158538e-01
-5.41470170e-01 -8.43473002e-02 -4.54722613e-01 2.29867801e-01
1.15857303e+00 -6.27724230e-01 -2.64114797e-01 5.52034140e-01
8.20936561e-01 -5.25269270e-01 -7.45097339e-01 6.40724301e-01
-7.50368601e-03 -1.48915589e+00 1.11461189e-02 -1.85482930e-02
4.55476254e-01 -2.32112199e-01 -3.81503552e-02 3.93853754e-01
3.06977957e-01 -9.58164573e-01 4.29353565e-01 5.27180374e-01
7.43245840e-01 -4.15336072e-01 1.25151944e+00 6.68235838e-01
-6.86111212e-01 2.74733305e-01 -2.69673526e-01 -3.29406172e-01
9.00583565e-02 3.30915570e-01 -1.76249278e+00 8.79574299e-01
5.06374717e-01 3.68491143e-01 -3.75376225e-01 1.09543478e+00
-4.14168060e-01 1.01304758e+00 -3.04655105e-01 -3.98821801e-01
4.45286125e-01 1.25168815e-01 3.03850710e-01 1.27734232e+00
6.94158137e-01 7.96256065e-01 2.82126695e-01 5.48281133e-01
1.84435472e-01 5.24199784e-01 -4.34374243e-01 -3.78705300e-02
2.32446358e-01 1.08704638e+00 -8.43149602e-01 -5.14287591e-01
-5.23316979e-01 3.99959415e-01 5.94251379e-02 2.75586396e-01
-7.49502540e-01 1.03274427e-01 -6.06483314e-03 4.75543410e-01
-2.29752272e-01 1.06110722e-02 -4.45765495e-01 -5.94589770e-01
-7.64699951e-02 -1.09002125e+00 4.03756022e-01 -1.22028685e+00
-9.97337043e-01 8.23351920e-01 1.14145629e-01 -1.33706629e+00
-4.53886986e-01 -2.29024962e-01 -3.74947399e-01 8.59922886e-01
-1.25328875e+00 -7.48645425e-01 -2.36204520e-01 3.36274803e-01
5.30879915e-01 3.26041460e-01 1.21922457e+00 1.52962044e-01
-1.59523726e-01 4.87142950e-01 -4.20709699e-01 7.73498788e-02
1.08495951e+00 -7.31326520e-01 -7.35378265e-02 2.24486306e-01
-5.11838675e-01 9.74817753e-01 6.84333086e-01 -8.99863482e-01
-9.42025483e-01 -8.72114122e-01 1.35866177e+00 -2.62990028e-01
5.30917585e-01 2.53416777e-01 -1.09057701e+00 6.48347080e-01
3.30622792e-01 -4.26798403e-01 1.11676025e+00 -4.02491242e-01
2.94096380e-01 3.79817754e-01 -1.18440723e+00 7.55346298e-01
5.60658693e-01 -4.50190723e-01 -6.56554043e-01 1.04184425e+00
9.55205441e-01 -9.65712249e-01 -1.16414869e+00 4.81086187e-02
1.35135785e-01 -5.77494800e-01 6.52756512e-01 -5.70999384e-01
7.48762488e-01 2.77720928e-01 3.44446033e-01 -1.22603118e+00
-2.62342155e-01 -5.85827470e-01 7.46991515e-01 4.88220543e-01
6.80056393e-01 -9.32426751e-01 6.83259726e-01 9.84276772e-01
-3.57741952e-01 -8.91093194e-01 -7.50578463e-01 -2.74974495e-01
9.11666080e-03 -5.93524396e-01 1.21699817e-01 9.01586056e-01
4.42048013e-01 8.30876306e-02 1.15820346e-02 1.01375148e-01
3.33714597e-02 -2.97950357e-01 3.48863482e-01 -6.40927255e-01
-3.68055522e-01 -9.26121324e-02 4.07434180e-02 -9.01827872e-01
-3.68756235e-01 -7.43691862e-01 -2.98231453e-01 -2.18936110e+00
5.99579811e-01 -2.80497372e-01 2.46000737e-01 8.37439120e-01
-7.47726858e-02 -6.45752996e-02 1.67921036e-01 1.84616402e-01
-2.96746463e-01 4.48582992e-02 1.75860727e+00 -4.23205607e-02
-6.30549192e-02 2.81976402e-01 -1.28270900e+00 6.24509931e-01
7.20417261e-01 -9.66970384e-01 -2.71619260e-01 -4.46038961e-01
3.36357802e-01 1.06353760e+00 2.21474111e-01 -6.41286373e-01
5.96484482e-01 -1.74893871e-01 1.59013018e-01 -4.95345026e-01
2.29450405e-01 -7.65456617e-01 3.58860523e-01 8.32818449e-01
-6.04116678e-01 4.02275801e-01 5.16635954e-01 1.34990364e-01
-4.50943559e-01 -3.91405433e-01 1.52669981e-01 -5.92719853e-01
6.69783447e-03 -3.20834070e-01 -1.20882070e+00 -1.15768509e-02
9.86098409e-01 -3.67191210e-02 -5.19331932e-01 -1.21880317e+00
-1.01212978e+00 2.87245780e-01 2.86017498e-03 6.53466582e-02
8.48013639e-01 -8.67851794e-01 -9.61595118e-01 -1.28066629e-01
-1.02039427e-01 -1.15574978e-01 6.02068424e-01 1.53421879e+00
-9.33009505e-01 7.63692975e-01 4.70954888e-02 -4.98370647e-01
-1.42412019e+00 7.87545219e-02 3.62416096e-02 -7.51517117e-01
-9.01316941e-01 6.10778928e-01 2.20355570e-01 -4.36513543e-01
3.71792354e-02 -7.41973102e-01 -8.67110193e-02 -3.80492091e-01
6.72457874e-01 -8.81766453e-02 3.92336428e-01 -2.04904601e-01
-4.20009613e-01 2.48752534e-01 -5.35717070e-01 -3.09453309e-01
1.30510747e+00 -2.04457447e-01 -2.05087870e-01 3.48021716e-01
8.57486486e-01 3.72130930e-01 -1.53191969e-01 -6.47515878e-02
-1.31978527e-01 -1.95931941e-01 -1.42838851e-01 -1.26230216e+00
-7.17946947e-01 5.61922073e-01 1.85640648e-01 2.87477877e-02
1.03249264e+00 2.35130206e-01 6.00383997e-01 6.23722315e-01
4.34277087e-01 -5.67923546e-01 1.00485213e-01 3.26505899e-01
9.77582753e-01 -9.02227402e-01 5.30038849e-02 -6.44601941e-01
-1.10938239e+00 1.13468432e+00 4.12216663e-01 2.98550606e-01
4.90999460e-01 4.74617422e-01 6.81500375e-01 -3.57146531e-01
-1.07849228e+00 5.31152308e-01 -4.29333486e-02 3.35861206e-01
9.66766238e-01 1.95651382e-01 -9.36487257e-01 8.84073734e-01
-4.48075593e-01 2.94711590e-01 1.11397243e+00 1.33526659e+00
-2.01922044e-01 -1.07814240e+00 -7.47117579e-01 4.41821009e-01
-7.64567256e-01 -1.04832351e-01 -3.44341695e-01 8.96188796e-01
-5.67216098e-01 1.08653855e+00 -4.61212903e-01 -1.51935536e-02
5.28017461e-01 -7.37816244e-02 3.41775790e-02 -9.96432245e-01
-1.15809155e+00 2.94900745e-01 5.18798411e-01 -5.54702997e-01
-3.29930812e-01 -4.25031990e-01 -1.52056420e+00 -2.65368998e-01
-1.25639394e-01 6.94352686e-01 6.42043531e-01 8.37156236e-01
3.29625815e-01 1.02736616e+00 1.54171035e-01 7.70446211e-02
-6.05218768e-01 -9.29942489e-01 -2.02829286e-01 -2.21675038e-01
5.89878932e-02 -6.09575287e-02 -2.57274248e-02 -2.84310430e-01] | [8.749312400817871, 8.329926490783691] |
ef556f23-a592-40a5-9211-8442011857c2 | multifit-efficient-multi-lingual-language | 1909.04761 | null | https://arxiv.org/abs/1909.04761v2 | https://arxiv.org/pdf/1909.04761v2.pdf | MultiFiT: Efficient Multi-lingual Language Model Fine-tuning | Pretrained language models are promising particularly for low-resource languages as they only require unlabelled data. However, training existing models requires huge amounts of compute, while pretrained cross-lingual models often underperform on low-resource languages. We propose Multi-lingual language model Fine-Tuning (MultiFiT) to enable practitioners to train and fine-tune language models efficiently in their own language. In addition, we propose a zero-shot method using an existing pretrained cross-lingual model. We evaluate our methods on two widely used cross-lingual classification datasets where they outperform models pretrained on orders of magnitude more data and compute. We release all models and code. | ['Julian Martin Eisenschlos', 'Jeremy Howard', 'Sebastian Ruder', 'Piotr Czapla', 'Sylvain Gugger', 'Marcin Kardas'] | 2019-09-10 | multifit-efficient-multi-lingual-language-1 | https://aclanthology.org/D19-1572 | https://aclanthology.org/D19-1572.pdf | ijcnlp-2019-11 | ['cross-lingual-document-classification'] | ['natural-language-processing'] | [-4.83286470e-01 -4.57328439e-01 -7.97929108e-01 -5.44890583e-01
-1.55313301e+00 -8.21792364e-01 4.77632582e-01 -9.06383768e-02
-8.64388108e-01 7.81196535e-01 1.92976613e-02 -5.06148517e-01
5.33703446e-01 -5.50098419e-01 -7.60032356e-01 -9.19917375e-02
2.17828840e-01 8.59866023e-01 -8.54957104e-02 -2.76223719e-01
-2.74196744e-01 2.02739686e-01 -1.00295234e+00 3.64763260e-01
9.60268855e-01 3.16069335e-01 1.83882713e-01 5.18909037e-01
-4.37796861e-01 4.75315124e-01 -1.99683875e-01 -6.14327669e-01
3.07363927e-01 -4.21787007e-03 -7.56626070e-01 -4.22554091e-02
5.49206316e-01 8.00231248e-02 1.45945564e-01 8.21639359e-01
4.26958650e-01 1.29381657e-01 3.75420988e-01 -8.18124592e-01
-1.00609803e+00 9.49122369e-01 -4.68823999e-01 2.10873336e-01
-1.68195188e-01 1.62840068e-01 9.86379385e-01 -1.02174580e+00
5.51640570e-01 1.30779791e+00 9.36739504e-01 5.51371932e-01
-1.29008687e+00 -7.37640023e-01 3.27060491e-01 -2.47904435e-01
-1.76888204e+00 -8.04819882e-01 4.76660430e-01 -3.91995668e-01
1.54951727e+00 -3.00571412e-01 5.77303171e-02 1.02002478e+00
8.58175848e-03 7.63711810e-01 1.20352149e+00 -8.09686661e-01
-1.49640068e-01 5.80159128e-01 -1.78962611e-02 9.07382667e-01
1.47568822e-01 -1.04103349e-01 -4.24829334e-01 -1.55960873e-01
4.46642429e-01 -2.02101558e-01 3.36580545e-01 -7.28351846e-02
-1.17067170e+00 1.00714684e+00 7.42492825e-02 6.66456640e-01
-1.81848794e-01 6.78353459e-02 7.01752126e-01 3.56142938e-01
1.00838137e+00 5.78661501e-01 -1.12641394e+00 -2.90375739e-01
-1.06461108e+00 -2.46349275e-01 6.60820723e-01 1.12130260e+00
1.24369442e+00 2.22142085e-01 1.85649171e-01 1.38073170e+00
1.90320909e-02 4.35840845e-01 7.99697995e-01 -5.87072730e-01
5.47090769e-01 5.22251368e-01 -2.46322304e-01 -1.74993619e-01
-1.58055648e-01 -5.41769981e-01 -7.46352911e-01 -6.08478785e-01
3.07099801e-02 -2.50677526e-01 -1.04673338e+00 1.58451962e+00
5.00926338e-02 4.06082213e-01 2.72206306e-01 3.18484038e-01
5.31191051e-01 7.29114890e-01 4.48055625e-01 -1.02123156e-01
1.30405390e+00 -1.43932426e+00 -4.98091668e-01 -7.05670118e-01
1.48187983e+00 -9.91565049e-01 1.70580363e+00 2.48855263e-01
-9.05939400e-01 -5.70572913e-01 -6.22209668e-01 -4.08224136e-01
-7.40884244e-01 5.02421141e-01 8.87082696e-01 6.89075172e-01
-1.17154968e+00 4.85865101e-02 -8.43357205e-01 -4.47008640e-01
1.30374402e-01 1.74046218e-01 -3.06836963e-01 -4.07338500e-01
-1.43984878e+00 1.03610015e+00 6.43006265e-01 -2.95209527e-01
-9.03767347e-01 -8.82974863e-01 -1.02867889e+00 -2.78435677e-01
3.69995475e-01 -4.30325478e-01 1.40612662e+00 -7.84340262e-01
-1.46772707e+00 1.34317231e+00 -2.87159741e-01 -4.26509857e-01
2.58371115e-01 -1.88843235e-01 -5.27816713e-01 -4.96417254e-01
2.43059725e-01 6.64156795e-01 3.80165815e-01 -1.00741136e+00
-5.40107369e-01 6.42800424e-03 4.07824405e-02 1.74084917e-01
-6.90915227e-01 2.90838361e-01 -1.08351588e+00 -6.41445875e-01
-4.44757044e-01 -8.15737009e-01 -4.44001317e-01 -5.14139235e-01
-1.39482632e-01 -5.42642236e-01 5.33374488e-01 -5.24258196e-01
1.53524053e+00 -1.97621298e+00 -3.78358901e-01 -1.16310313e-01
-2.73368210e-01 3.81282240e-01 -4.80456412e-01 3.05983484e-01
1.52427733e-01 3.90783578e-01 -7.93884844e-02 -6.62757158e-01
-3.00131243e-04 6.05004013e-01 -1.03048824e-01 1.82426140e-01
-4.28848295e-03 1.05778849e+00 -7.47944772e-01 -6.76135480e-01
4.13609706e-02 5.64456522e-01 -8.26725125e-01 6.71826676e-02
-4.15865302e-01 2.26306338e-02 -3.41019005e-01 7.16155648e-01
2.95638502e-01 -3.75088960e-01 2.67887831e-01 -1.32661946e-02
-2.33689398e-01 6.14738762e-01 -7.05182612e-01 2.00799704e+00
-1.24537861e+00 2.06296548e-01 -1.62907556e-01 -9.05767679e-01
9.45819438e-01 2.96927482e-01 2.17259213e-01 -6.50357604e-01
-2.05364481e-01 4.93197978e-01 -2.35879973e-01 -2.99573302e-01
6.52845204e-01 -3.64345372e-01 -5.04555643e-01 6.85781777e-01
2.62248635e-01 -2.29162276e-01 4.34259951e-01 7.08346218e-02
5.41663468e-01 1.88970268e-01 3.82042587e-01 -4.66883063e-01
5.45571983e-01 8.38737786e-02 4.34360117e-01 5.83546400e-01
-2.27906376e-01 -1.12978391e-01 -7.76741877e-02 -6.22370064e-01
-1.01266217e+00 -6.66580856e-01 -2.22808227e-01 1.98613453e+00
-5.57669580e-01 -6.77080095e-01 -6.78788602e-01 -8.05235267e-01
-8.80182311e-02 7.80369282e-01 -3.55268598e-01 9.57720056e-02
-5.78360379e-01 -8.99657130e-01 9.37776327e-01 4.87899244e-01
2.50733703e-01 -8.12655210e-01 2.28043750e-01 2.32465789e-01
-1.26765803e-01 -1.40119112e+00 -8.96841764e-01 4.27929282e-01
-5.47035813e-01 -5.13821781e-01 -5.06418109e-01 -1.19031906e+00
5.27358472e-01 -7.04882108e-03 1.69379127e+00 1.33246053e-02
-1.51101828e-01 3.00200194e-01 -7.23442584e-02 -2.69080251e-01
-5.00664473e-01 7.62673438e-01 4.29198623e-01 -2.82343835e-01
9.45176840e-01 -2.07079455e-01 2.04020187e-01 1.47327170e-01
-6.48610413e-01 9.85385925e-02 3.70269507e-01 7.59354830e-01
9.14331973e-01 6.96648881e-02 5.62219083e-01 -1.22272122e+00
6.76650286e-01 -4.80093271e-01 -7.72420347e-01 7.54753470e-01
-6.90510392e-01 3.08764219e-01 8.06184471e-01 -4.41959679e-01
-1.04805422e+00 -8.92541837e-03 -3.54118645e-02 -4.26402450e-01
-3.67297046e-02 8.11709702e-01 1.81345075e-01 -2.99227592e-02
6.91929102e-01 1.56872869e-01 -7.14155555e-01 -7.78305769e-01
6.16976440e-01 6.05383396e-01 3.81395191e-01 -1.09508705e+00
7.27953076e-01 9.87662226e-02 -7.20053256e-01 -6.69099391e-01
-1.29889750e+00 -5.01038015e-01 -9.75932717e-01 3.71799558e-01
6.20841742e-01 -1.55910003e+00 -1.47451386e-01 1.47031575e-01
-1.01775265e+00 -8.09096754e-01 2.94777881e-02 6.28901720e-01
-1.55336693e-01 5.51716052e-02 -7.97332823e-01 -5.17520189e-01
-4.31975454e-01 -1.06701910e+00 1.14602840e+00 -2.02447742e-01
6.22704299e-03 -1.89721334e+00 3.59744787e-01 4.12829846e-01
5.31166017e-01 -4.95779842e-01 9.37762797e-01 -7.80092537e-01
-3.04804891e-01 -1.56136602e-01 -5.69598749e-02 4.26570147e-01
2.40500063e-01 2.82472977e-03 -9.13358092e-01 -5.12425661e-01
-6.14229321e-01 -1.00401735e+00 7.48484731e-01 2.63447613e-01
1.16580319e+00 -2.30638415e-01 -2.87677795e-01 8.60071659e-01
1.40816426e+00 -3.48096997e-01 -1.46803886e-01 3.73148531e-01
9.37571704e-01 3.86151046e-01 4.24962103e-01 1.30549699e-01
9.35871243e-01 6.85371459e-01 -5.54251015e-01 -3.58830184e-01
8.80583301e-02 -4.99542624e-01 5.83210111e-01 1.60504699e+00
1.60468906e-01 -1.44126147e-01 -1.36096847e+00 7.73113549e-01
-1.44976389e+00 -4.65663284e-01 1.73351720e-01 2.01027322e+00
1.46768010e+00 2.40348205e-01 -4.26282398e-02 -6.86857700e-01
4.43997324e-01 -1.16473317e-01 -5.59156120e-01 -6.84475482e-01
-3.90131444e-01 5.07497251e-01 6.31214142e-01 9.69724476e-01
-1.04992056e+00 1.98171496e+00 7.15532160e+00 1.08273494e+00
-1.39691544e+00 5.47060370e-01 8.21598530e-01 -1.90151244e-01
-4.51891929e-01 6.94647282e-02 -1.35930562e+00 1.63702965e-01
1.43376029e+00 -5.84297001e-01 6.07865810e-01 1.23156106e+00
2.02297896e-01 1.39008254e-01 -9.88144577e-01 9.62379515e-01
1.89980432e-01 -1.18951178e+00 1.53911650e-01 1.56511292e-02
1.19489145e+00 7.48835623e-01 -7.73391053e-02 1.08985293e+00
8.91382754e-01 -1.05289555e+00 1.98322743e-01 1.14420585e-01
1.10768867e+00 -8.39882493e-01 4.38556224e-01 5.81992507e-01
-1.38180304e+00 5.39241076e-01 -5.14835358e-01 7.89735392e-02
1.31113231e-01 5.24610996e-01 -9.19710994e-01 3.15047920e-01
6.76525414e-01 8.10628891e-01 -8.62553298e-01 3.08506876e-01
-1.25695214e-01 8.72006118e-01 -4.36971754e-01 2.75151730e-01
5.83775938e-01 9.80127417e-03 -2.68260479e-01 1.65607786e+00
2.21409649e-01 -2.35751152e-01 8.32775593e-01 5.21384835e-01
-4.86978114e-01 4.04235959e-01 -6.27117038e-01 -4.30836439e-01
5.65046012e-01 1.29952490e+00 -2.50992864e-01 -9.12979364e-01
-6.59029841e-01 8.26682329e-01 7.84206033e-01 2.54684657e-01
-6.19184971e-01 4.92702499e-02 6.36112571e-01 -8.63411054e-02
-6.17356785e-03 -4.78021204e-01 -1.87949613e-01 -1.71422064e+00
-2.07646951e-01 -1.01465571e+00 5.74899852e-01 -4.97183055e-01
-1.65593028e+00 8.27820420e-01 -1.68128192e-01 -8.54721963e-01
-6.57637417e-01 -7.24934340e-01 -8.84765387e-03 1.17194223e+00
-1.97327578e+00 -1.67974937e+00 1.82297066e-01 8.39393973e-01
8.08158755e-01 -4.71947312e-01 1.22677398e+00 7.13312387e-01
-6.78920388e-01 1.08857274e+00 1.10549174e-01 2.81311393e-01
1.15757620e+00 -9.76372719e-01 7.08010554e-01 7.51443148e-01
3.15511972e-01 1.00891042e+00 2.19397783e-01 -5.51516652e-01
-1.36613679e+00 -1.38558888e+00 1.47395515e+00 -6.95412278e-01
1.09782946e+00 -6.55570030e-01 -1.12126780e+00 1.15424609e+00
2.64955729e-01 3.49353313e-01 9.74458516e-01 7.02519059e-01
-6.41382813e-01 -1.73636479e-04 -7.14027762e-01 4.24878120e-01
8.52549255e-01 -1.05270827e+00 -3.69611770e-01 6.58433080e-01
7.36915946e-01 -2.36757234e-01 -1.03984690e+00 2.01384947e-01
2.97257155e-01 -2.82767594e-01 7.38547921e-01 -8.69016171e-01
1.13507815e-01 -4.18047793e-02 -2.63359606e-01 -1.41100395e+00
-2.57311821e-01 -5.68773448e-01 2.26939246e-01 1.28528810e+00
8.93586278e-01 -6.98340178e-01 5.41407168e-01 4.04887527e-01
-5.41393906e-02 -6.28099501e-01 -5.81612587e-01 -9.78009999e-01
6.23037815e-01 -8.73440504e-01 4.35888469e-01 1.65023994e+00
-2.97017038e-01 5.83282232e-01 -4.98954117e-01 -7.89032429e-02
5.32263696e-01 -1.33260652e-01 8.47409248e-01 -8.08350921e-01
-4.37211186e-01 -2.26714849e-01 2.19959408e-01 -9.09957945e-01
7.89274037e-01 -1.39338076e+00 1.23117566e-02 -1.16513526e+00
3.65676075e-01 -9.90946174e-01 -4.04571593e-01 1.08356690e+00
-2.79361397e-01 4.48580652e-01 -1.56786993e-01 2.09355861e-01
-7.07373977e-01 4.06885654e-01 8.26279342e-01 -6.97361827e-02
-2.24706635e-01 -3.16456020e-01 -7.56650329e-01 8.27506661e-01
8.03612173e-01 -5.35906851e-01 -3.42481226e-01 -9.97661650e-01
2.20418319e-01 -4.05455381e-01 -2.18222618e-01 -5.91991723e-01
1.36344075e-01 -4.21027452e-01 6.80667236e-02 -3.97782147e-01
2.06292376e-01 -3.76682937e-01 -2.64574766e-01 7.05113709e-02
-4.00565892e-01 3.58114421e-01 6.45302236e-01 -1.79367978e-02
-2.24525258e-01 -3.54356281e-02 9.20359969e-01 -4.64724094e-01
-8.59432101e-01 5.27531385e-01 -1.65370584e-01 4.85606998e-01
7.11400151e-01 3.18534553e-01 -1.11928731e-01 -8.67406949e-02
-4.83304918e-01 2.87605256e-01 5.20820916e-01 8.44897807e-01
-7.44725764e-02 -1.35347891e+00 -9.06094968e-01 2.38886192e-01
4.29956496e-01 -2.30763003e-01 8.82548392e-02 5.67620575e-01
-3.53820592e-01 8.86645257e-01 2.74330586e-01 -5.84487796e-01
-1.03580976e+00 6.34328842e-01 5.04870951e-01 -7.69319713e-01
-1.09873287e-01 9.66744423e-01 2.91161120e-01 -1.35440385e+00
-2.56384742e-02 -3.49840283e-01 2.19639838e-01 -1.33896545e-01
3.81456733e-01 -2.64107257e-01 2.46654779e-01 -6.43532693e-01
-4.36757207e-01 6.50573909e-01 -3.46561134e-01 -2.53159434e-01
1.20261371e+00 -5.29653616e-02 -1.97580546e-01 7.73192048e-01
1.34040952e+00 1.92826301e-01 -9.53914940e-01 -8.59038770e-01
2.67541379e-01 -1.10401148e-02 2.79679179e-01 -9.58270192e-01
-7.89457798e-01 9.44452107e-01 1.92521051e-01 -3.99087220e-01
1.01769972e+00 7.64481202e-02 7.33484089e-01 7.24660754e-01
7.08067119e-01 -1.34633744e+00 -1.37863845e-01 9.57954228e-01
3.54594380e-01 -1.49349225e+00 8.59986618e-03 -8.00028667e-02
-7.25881875e-01 7.61738718e-01 8.83819580e-01 8.65315646e-02
7.48851836e-01 6.32756710e-01 4.80058193e-01 8.40391815e-02
-1.28066945e+00 -1.82726294e-01 4.62595344e-01 2.10574761e-01
1.18743813e+00 3.99744034e-01 -8.60670805e-02 6.67048931e-01
-4.61063743e-01 1.27409413e-01 2.24069655e-02 7.51436949e-01
-2.33351097e-01 -1.51572776e+00 -1.79955661e-01 7.49390423e-02
-4.74660099e-01 -8.05125237e-01 -1.55399263e-01 7.00663805e-01
2.73288965e-01 6.69992089e-01 -1.60115473e-02 -6.38004839e-02
9.07672271e-02 4.31215316e-01 2.96414524e-01 -1.02420557e+00
-6.80356264e-01 2.96227038e-01 2.59858817e-01 -3.35108310e-01
-3.14016044e-01 -3.90904486e-01 -1.04304779e+00 -4.20709193e-01
-1.17061064e-01 2.66132653e-01 7.87776053e-01 1.10081530e+00
4.46967155e-01 2.91229635e-01 4.39217091e-01 -5.38575411e-01
-3.44957948e-01 -1.13279891e+00 -2.19944313e-01 1.07974611e-01
-3.32216136e-02 -2.87792325e-01 -2.24382311e-01 4.44312721e-01] | [10.958582878112793, 9.908451080322266] |
da93894b-14c0-4165-9a34-be2fa28d1948 | rethinking-log-odds-linear-probability | 2211.06360 | null | https://arxiv.org/abs/2211.06360v1 | https://arxiv.org/pdf/2211.06360v1.pdf | Rethinking Log Odds: Linear Probability Modelling and Expert Advice in Interpretable Machine Learning | We introduce a family of interpretable machine learning models, with two broad additions: Linearised Additive Models (LAMs) which replace the ubiquitous logistic link function in General Additive Models (GAMs); and SubscaleHedge, an expert advice algorithm for combining base models trained on subsets of features called subscales. LAMs can augment any additive binary classification model equipped with a sigmoid link function. Moreover, they afford direct global and local attributions of additive components to the model output in probability space. We argue that LAMs and SubscaleHedge improve the interpretability of their base algorithms. Using rigorous null-hypothesis significance testing on a broad suite of financial modelling data, we show that our algorithms do not suffer from large performance penalties in terms of ROC-AUC and calibration. | ['Daniele Magazzeni', 'Freddy Lecue', 'Nicolas Marchesotti', 'Danial Dervovic'] | 2022-11-11 | null | null | null | null | ['additive-models', 'interpretable-machine-learning'] | ['methodology', 'methodology'] | [ 1.04618430e-01 6.77583635e-01 -8.35504457e-02 -6.29794598e-01
-6.09539032e-01 -6.32534981e-01 7.56155789e-01 1.68061882e-01
-1.67508870e-01 8.22989285e-01 -1.20714299e-01 -8.89202893e-01
-6.89232707e-01 -7.73528039e-01 -8.95039141e-01 -4.14689600e-01
-1.73461884e-01 4.44330513e-01 -9.17586982e-02 4.84798290e-02
2.17621371e-01 1.30720153e-01 -1.02634621e+00 2.60221779e-01
1.26911914e+00 1.07081294e+00 -3.43931913e-01 8.17610443e-01
2.94310033e-01 1.12110126e+00 -3.61464173e-01 -9.27610517e-01
2.26400658e-01 -3.13727334e-02 -7.37967864e-02 -1.09511174e-01
1.90695897e-01 -5.32725751e-01 -1.38973430e-01 5.60973525e-01
8.56989101e-02 -2.74763014e-02 1.44851446e+00 -1.82332718e+00
-1.01714838e+00 8.91275346e-01 -4.72763598e-01 -1.10759355e-01
8.38445947e-02 1.89734653e-01 1.21811581e+00 -8.83402586e-01
-2.41586447e-01 1.44998765e+00 1.34257007e+00 -6.45935675e-03
-1.64358556e+00 -3.14677149e-01 1.56052485e-01 -2.75849164e-01
-9.93322253e-01 -1.07243866e-01 3.13104361e-01 -6.00337148e-01
9.09085393e-01 6.99243248e-01 2.18943223e-01 8.52959216e-01
3.73502970e-01 6.18268311e-01 1.42186654e+00 -6.73987627e-01
2.69130021e-01 5.18810809e-01 6.44545913e-01 6.95772350e-01
6.92030787e-01 3.07306170e-01 -3.00100386e-01 -8.89787614e-01
1.00903046e+00 1.06339842e-01 2.72530735e-01 -3.26307863e-01
-7.63981462e-01 1.06924987e+00 8.48120302e-02 -3.72350633e-01
-4.78453517e-01 6.28534436e-01 -1.07757658e-01 3.70540828e-01
8.21380138e-01 5.38442194e-01 -8.41329157e-01 3.00266832e-01
-4.35671210e-01 5.39759755e-01 8.25176001e-01 7.06169009e-01
5.14517963e-01 -6.94275945e-02 7.16994517e-03 6.65733099e-01
5.57962835e-01 2.60850579e-01 1.60051778e-01 -1.17238569e+00
2.83794224e-01 6.67452157e-01 4.47756171e-01 -8.26227963e-01
-7.28404403e-01 -6.89066708e-01 -4.85235989e-01 4.11085308e-01
7.38150656e-01 -2.55911380e-01 -6.25446856e-01 1.66355610e+00
-1.32954540e-02 8.84159282e-02 -1.28473371e-01 1.56757057e-01
2.49413207e-01 3.46984655e-01 4.85608399e-01 -1.82746705e-02
1.07092953e+00 -6.22007489e-01 -3.52228671e-01 -3.67565215e-01
8.15920115e-01 -2.97148079e-01 1.09032667e+00 6.10807717e-01
-1.16590226e+00 -2.78430879e-01 -9.54047918e-01 2.65819542e-02
-2.68935800e-01 1.15085453e-01 1.23905051e+00 1.07459772e+00
-9.77236748e-01 8.37234378e-01 -8.14462185e-01 4.80186075e-01
2.91192114e-01 5.84554374e-01 -1.08801961e-01 1.91000447e-01
-1.11600602e+00 1.09346759e+00 1.95309564e-01 -1.50520891e-01
-1.22995242e-01 -1.17827702e+00 -7.09576070e-01 1.17620751e-01
1.49170145e-01 -1.04247463e+00 1.40222597e+00 -1.09264040e+00
-1.25801194e+00 4.69007045e-01 2.44621918e-01 -7.46582091e-01
9.35373783e-01 -4.91760939e-01 4.20858301e-02 -2.81263202e-01
-1.60640463e-01 3.98633391e-01 7.92205930e-01 -1.11170292e+00
-4.91066426e-01 -3.40683430e-01 1.88367411e-01 1.46192357e-01
-1.96615607e-01 1.45852834e-01 3.15145254e-01 -9.78311062e-01
5.93988597e-02 -7.45339096e-01 -5.17286777e-01 1.84618488e-01
-3.94413769e-01 -1.88753307e-01 9.54558477e-02 -8.14768255e-01
1.34691906e+00 -1.90559077e+00 -3.31760913e-01 2.67098010e-01
3.71172458e-01 -4.07119513e-01 2.42509067e-01 -2.47751176e-02
-3.51101935e-01 3.93073976e-01 -4.85199034e-01 -3.25488031e-01
6.23123169e-01 4.25342098e-02 -4.14623141e-01 3.30077380e-01
3.58095229e-01 9.48892832e-01 -7.52379179e-01 -1.15614980e-01
1.86607018e-01 2.89192826e-01 -7.25738347e-01 2.77605373e-02
3.54100056e-02 -2.89424211e-01 -3.56416442e-02 2.96847403e-01
6.89667344e-01 -2.94301242e-01 1.49894610e-01 3.44701886e-01
2.31569499e-01 4.04008985e-01 -1.26031911e+00 5.86620092e-01
-3.33660662e-01 1.11127630e-01 -1.94972530e-01 -1.02762210e+00
8.12835872e-01 1.58621088e-01 1.99901327e-01 -1.01936623e-01
-1.47547930e-01 4.85752732e-01 1.64101675e-01 1.95320304e-02
2.62516826e-01 -3.45044732e-01 -2.48516649e-01 6.17816210e-01
9.81507376e-02 3.48748378e-02 -4.06767786e-01 1.92363724e-01
9.74903941e-01 2.27349997e-01 6.72089517e-01 -4.95016575e-01
-6.04008846e-02 -3.04429770e-01 3.81509870e-01 1.14372516e+00
1.74023241e-01 4.00623530e-01 9.71697211e-01 -3.95976961e-01
-1.01878071e+00 -1.41725707e+00 -4.32468355e-01 1.31905699e+00
-4.16366279e-01 -1.62228122e-01 -5.27603030e-01 -7.21813083e-01
6.14117622e-01 1.39222479e+00 -8.95805478e-01 -2.16300890e-01
-1.22524917e-01 -1.61596453e+00 6.19675756e-01 9.67887998e-01
-1.51816010e-01 -3.70569736e-01 -4.95787233e-01 1.16413176e-01
5.07387400e-01 -5.82822263e-01 -1.65780246e-01 7.77984440e-01
-9.45939898e-01 -1.09515047e+00 -3.93612534e-01 -9.25617218e-02
6.69453561e-01 -1.18164189e-01 1.16491854e+00 -1.28187954e-01
4.84557599e-02 5.92126667e-01 -7.44119734e-02 -7.36195624e-01
-4.99552965e-01 -4.34101075e-01 5.33751510e-02 -1.52145013e-01
3.37713599e-01 -5.64951301e-01 -4.36185867e-01 4.76190478e-01
-4.57266003e-01 1.01505376e-01 4.83950764e-01 9.57264364e-01
3.25706340e-02 -1.94777131e-01 6.98312640e-01 -1.03947353e+00
7.46896446e-01 -8.98589075e-01 -4.66449171e-01 2.65114188e-01
-9.44179416e-01 7.39439204e-02 5.62849283e-01 -5.95844030e-01
-1.06204677e+00 -2.62476683e-01 3.64505976e-01 -4.31687683e-02
1.98280782e-01 6.90905333e-01 -2.41876803e-02 3.68027352e-02
9.67565298e-01 -3.88457254e-02 4.22578454e-02 -5.46971858e-01
4.92915452e-01 6.07258618e-01 6.39320910e-01 -7.64439344e-01
5.42092621e-01 -2.27541607e-02 2.13767141e-01 -3.23251009e-01
-6.21662378e-01 1.37650713e-01 -5.86327016e-01 3.70569266e-02
4.09851491e-01 -7.03865111e-01 -5.44750750e-01 2.98587352e-01
-7.21249461e-01 -4.82205778e-01 -2.15017706e-01 4.87882942e-01
-9.30343211e-01 2.36824036e-01 -5.18175900e-01 -1.46499062e+00
7.36278296e-02 -7.99326003e-01 6.26743376e-01 -1.01679474e-01
-5.66371977e-01 -1.45917618e+00 -3.11961353e-01 1.84272230e-01
3.33734870e-01 4.62385505e-01 1.26138604e+00 -1.29191840e+00
-2.24029645e-01 -4.67508435e-01 -2.62455314e-01 4.51001287e-01
-2.63408333e-01 2.64471680e-01 -9.30812120e-01 1.27690554e-01
-6.67121867e-03 -1.72451138e-01 8.59204113e-01 8.32442939e-01
1.18239987e+00 -7.37157106e-01 -1.90461501e-01 5.48661113e-01
9.89163101e-01 2.54620016e-01 3.88987839e-01 3.58524293e-01
2.84001797e-01 6.93749130e-01 3.15755010e-01 4.27311808e-01
3.96712542e-01 5.05978465e-01 3.38453442e-01 -5.08433171e-02
4.88932103e-01 -3.24146986e-01 5.82229197e-01 1.03902631e-01
-2.44905442e-01 -1.59751073e-01 -7.81001806e-01 6.87634423e-02
-2.05737734e+00 -7.43335545e-01 -6.77749991e-01 2.43786263e+00
7.01150715e-01 4.64012980e-01 3.88578087e-01 1.83331490e-01
5.74305534e-01 -3.88819844e-01 -5.21247983e-01 -5.97047448e-01
-1.92427635e-01 1.98313296e-01 8.86426747e-01 5.77751458e-01
-1.39650619e+00 1.23698093e-01 8.14506531e+00 8.24428856e-01
-2.78504401e-01 1.44749492e-01 1.27259064e+00 5.88167757e-02
-4.45936620e-01 1.43899933e-01 -4.76005465e-01 6.51802897e-01
1.26576400e+00 -4.65706468e-01 2.89128512e-01 1.16057742e+00
-2.19032844e-03 -5.40316664e-03 -1.32964468e+00 4.71795470e-01
-4.12577569e-01 -9.67567563e-01 -4.26299691e-01 2.16888309e-01
6.92236304e-01 -3.67825359e-01 5.93942225e-01 5.68011463e-01
1.03551435e+00 -1.30885637e+00 9.12485838e-01 7.15352893e-01
5.86956441e-01 -6.91787899e-01 7.89993584e-01 3.92311454e-01
-3.94692749e-01 -3.49212706e-01 -3.88080776e-01 -4.25962597e-01
-1.67656183e-01 6.20802760e-01 -7.95425534e-01 1.87410384e-01
1.76534176e-01 1.56741068e-01 -6.66193247e-01 1.01615500e+00
-1.12438299e-01 8.88957381e-01 -6.37803972e-01 2.31628269e-01
5.92277236e-02 -2.96558589e-01 4.38418269e-01 1.15194488e+00
3.41451377e-01 1.26160130e-01 -2.03903154e-01 1.13809621e+00
3.05848658e-01 -6.27835542e-02 -3.36443841e-01 1.73373044e-01
4.37966317e-01 9.66996431e-01 -4.43846434e-01 -4.13583130e-01
-4.95280027e-01 4.51098442e-01 5.71889021e-02 2.17503384e-01
-1.07086360e+00 -9.47172046e-02 4.78176892e-01 1.94870323e-01
-4.01871577e-02 1.04259700e-01 -9.68430340e-01 -8.91147435e-01
-5.83079569e-02 -5.33322752e-01 5.88040531e-01 -8.88161063e-01
-1.64608967e+00 1.43228685e-02 3.10530841e-01 -9.57415104e-01
-4.23190296e-01 -1.08506930e+00 -7.68784821e-01 9.50196505e-01
-8.15576911e-01 -1.15085721e+00 1.17699631e-01 1.23150513e-01
9.32947248e-02 -8.57168883e-02 8.22693586e-01 -2.94312954e-01
-4.29236531e-01 9.11542654e-01 4.58249211e-01 -3.18022490e-01
5.79736829e-01 -1.86262250e+00 4.96082425e-01 4.07652318e-01
-1.97253153e-01 8.71960044e-01 9.53705013e-01 -5.09943724e-01
-9.13752198e-01 -1.00437057e+00 6.28283501e-01 -1.03951395e+00
1.04472888e+00 -2.01605305e-01 -8.81723404e-01 1.06332099e+00
-3.18602771e-01 -3.85222107e-01 8.43133748e-01 5.12194991e-01
-5.64605296e-01 1.79687038e-01 -1.29956388e+00 4.42892700e-01
7.95468092e-01 -1.02589838e-01 -8.21169555e-01 5.11503041e-01
7.42222548e-01 -4.83508259e-02 -1.04852927e+00 3.95347953e-01
8.64366770e-01 -1.20266747e+00 1.09171855e+00 -1.14577973e+00
6.21365964e-01 1.74437806e-01 -2.89130330e-01 -1.29423904e+00
-4.40732062e-01 -6.50460124e-01 -3.08762670e-01 9.99099016e-01
8.30095351e-01 -1.22145796e+00 3.01135778e-01 1.50391781e+00
-1.66327596e-01 -9.47329640e-01 -7.00603902e-01 -9.00622547e-01
4.27226990e-01 -9.75261450e-01 5.13126910e-01 8.80670965e-01
3.11570704e-01 -4.42041177e-03 -3.64664972e-01 -7.52973780e-02
8.48692954e-01 -4.55105782e-01 6.06848955e-01 -1.53146517e+00
-8.83185208e-01 -7.35083222e-01 -4.07364488e-01 -7.36188710e-01
-1.86533257e-01 -7.69788444e-01 -3.73038799e-01 -9.52274263e-01
3.36724460e-01 -8.12284291e-01 -5.16208887e-01 7.74693966e-01
-5.42000175e-01 7.47414529e-02 8.05433840e-02 7.69491568e-02
-3.02156866e-01 2.04510078e-01 4.07933235e-01 1.72016814e-01
-2.24515408e-01 4.25980806e-01 -1.31083572e+00 1.15023661e+00
6.48987293e-01 -5.43485999e-01 -4.59972322e-01 -1.16746619e-01
6.46415710e-01 4.88868877e-02 8.73895049e-01 -2.40558743e-01
-1.58470184e-01 -3.85752052e-01 6.11494422e-01 -9.45545658e-02
2.08669752e-01 -4.31874245e-01 3.77516806e-01 3.41189921e-01
-5.72781444e-01 7.19039664e-02 1.30571812e-01 5.41933417e-01
4.43128377e-01 -2.98148870e-01 5.98050594e-01 9.98354405e-02
3.29778045e-02 -1.97334573e-01 -2.44906887e-01 -2.40609691e-01
9.75150824e-01 -3.33071560e-01 -5.65010965e-01 -5.41970611e-01
-9.69803333e-01 6.03889152e-02 4.09762472e-01 2.56631207e-02
2.21374825e-01 -1.11803985e+00 -8.03113997e-01 3.18619274e-02
-2.78319925e-01 -4.01079178e-01 -3.61443236e-02 1.18384385e+00
-4.02084351e-01 4.88850743e-01 8.76046494e-02 -2.35297736e-02
-7.69715607e-01 5.58955252e-01 3.71209055e-01 -3.03226858e-01
-2.85108447e-01 1.01405168e+00 6.90717757e-01 -5.87517440e-01
-7.10943043e-02 -4.11463499e-01 2.52588391e-01 -2.18097374e-01
5.59047520e-01 8.33974719e-01 -3.15614700e-01 -2.09971353e-01
-2.13123888e-01 -7.55600333e-02 1.29633889e-01 -2.04374224e-01
1.54422426e+00 2.32211035e-02 -3.59343216e-02 5.71165204e-01
5.65011144e-01 -1.38220921e-01 -1.46160519e+00 3.14057432e-02
1.43128887e-01 -3.19044799e-01 -3.08959186e-02 -1.16857684e+00
-2.73505777e-01 6.29526556e-01 1.28649920e-01 5.40301919e-01
8.89360368e-01 -1.45934671e-01 -2.52405584e-01 2.90308148e-01
1.64070651e-01 -8.67862403e-01 -3.99437159e-01 2.44226400e-02
9.44269419e-01 -1.04235411e+00 2.68448919e-01 -4.49303508e-01
-8.25969577e-01 1.06525469e+00 2.92574883e-01 -4.11376178e-01
7.70107687e-01 4.67298836e-01 -3.49455267e-01 1.64953753e-01
-1.13520789e+00 3.45104039e-01 4.22622770e-01 6.99202776e-01
4.98405993e-01 3.86673421e-01 -4.71203208e-01 1.71583056e+00
-5.02740324e-01 -2.10084036e-01 4.98099655e-01 3.80913705e-01
-5.75622916e-01 -7.50397801e-01 -6.58983469e-01 1.02363527e+00
-4.45934087e-01 -2.88140684e-01 -4.74828720e-01 1.03701079e+00
2.18360499e-03 8.90812039e-01 3.73229444e-01 -3.46070379e-01
1.07238255e-01 5.19625247e-01 2.92925864e-01 -3.08040977e-01
-3.58351171e-01 1.02882631e-01 6.04402237e-02 -1.72959700e-01
1.88974097e-01 -9.80055869e-01 -7.41968870e-01 -5.10812759e-01
-5.46779215e-01 -1.19215421e-01 3.45599264e-01 8.95943046e-01
1.62905276e-01 3.93541336e-01 6.47654712e-01 -5.45898914e-01
-1.38348949e+00 -1.17777300e+00 -8.29901516e-01 2.93278638e-02
-1.79444011e-02 -8.40820491e-01 -6.86112821e-01 7.13724792e-02] | [8.624741554260254, 5.446321964263916] |
e8689a19-1483-4ff5-939f-a50cd197f268 | the-2021-nist-speaker-recognition-evaluation | 2204.10242 | null | https://arxiv.org/abs/2204.10242v1 | https://arxiv.org/pdf/2204.10242v1.pdf | The 2021 NIST Speaker Recognition Evaluation | The 2021 Speaker Recognition Evaluation (SRE21) was the latest cycle of the ongoing evaluation series conducted by the U.S. National Institute of Standards and Technology (NIST) since 1996. It was the second large-scale multimodal speaker/person recognition evaluation organized by NIST (the first one being SRE19). Similar to SRE19, it featured two core evaluation tracks, namely audio and audio-visual, as well as an optional visual track. In addition to offering fixed and open training conditions, it also introduced new challenges for the community, thanks to a new multimodal (i.e., audio, video, and selfie images) and multilingual (i.e., with multilingual speakers) corpus, termed WeCanTalk, collected outside North America by the Linguistic Data Consortium (LDC). These challenges included: 1) trials (target and non-target) with enrollment and test segments originating from different domains (i.e., telephony versus video), and 2) trials (target and non-target) with enrollment and test segments spoken in different languages (i.e., cross-lingual trials). This paper presents an overview of SRE21 including the tasks, performance metric, data, evaluation protocol, results and system performance analyses. A total of 23 organizations (forming 15 teams) from academia and industry participated in SRE21 and submitted 158 valid system outputs. Evaluation results indicate: audio-visual fusion produce substantial gains in performance over audio-only or visual-only systems; top performing speaker and face recognition systems exhibited comparable performance under the matched domain conditions present in this evaluation; and, the use of complex neural network architectures (e.g., ResNet) along with angular losses with margin, data augmentation, as well as long duration fine-tuning contributed to notable performance improvements for the audio-only speaker recognition task. | ['Douglas Reynolds', 'Lisa Mason', 'Elliot Singer', 'Craig Greenberg', 'Seyed Omid Sadjadi'] | 2022-04-21 | null | null | null | null | ['person-recognition'] | ['computer-vision'] | [ 1.67265624e-01 -2.15634599e-01 3.75612192e-02 -6.81829333e-01
-1.48680639e+00 -7.04264343e-01 7.84452677e-01 -3.06684285e-01
-5.31799018e-01 4.73912597e-01 3.91312689e-01 -3.57166529e-01
1.59468025e-01 -1.03869967e-01 -4.69044387e-01 -6.05282009e-01
-5.41810878e-02 4.31278199e-01 -2.42064089e-01 -1.46136612e-01
-1.70816369e-02 5.46158552e-01 -1.90052760e+00 5.53028822e-01
4.68141586e-01 9.84977484e-01 -1.37025550e-01 8.30681980e-01
1.97393540e-02 3.85535389e-01 -8.21028292e-01 -3.07604194e-01
-3.73232551e-02 -7.24173337e-02 -7.36970425e-01 1.03254281e-02
9.95869696e-01 -3.41291338e-01 -2.90566653e-01 6.34667814e-01
1.19056320e+00 1.16354853e-01 2.68406510e-01 -1.50822115e+00
-6.47021949e-01 5.08138895e-01 -3.44686002e-01 1.23835817e-01
7.69242465e-01 2.99171180e-01 7.83232093e-01 -1.38450801e+00
4.95417029e-01 1.64376736e+00 8.43633652e-01 7.23469079e-01
-1.32785368e+00 -9.85647082e-01 8.11398327e-02 3.52707028e-01
-1.68835497e+00 -1.39253020e+00 4.78632718e-01 -3.24191183e-01
1.24486089e+00 3.55116516e-01 1.75936684e-01 1.52370632e+00
-3.71936262e-01 7.09112465e-01 1.11503041e+00 -6.90797031e-01
1.34941921e-01 5.81009388e-01 2.51049936e-01 9.49662849e-02
-2.87484974e-01 3.90524626e-01 -7.46376276e-01 -3.84571224e-01
1.81558624e-01 -7.40469277e-01 -4.57906455e-01 -6.67217895e-02
-1.00650203e+00 6.84682250e-01 -1.38534799e-01 3.58499199e-01
-2.47584656e-01 -3.00622284e-01 7.78230071e-01 6.08702719e-01
1.67651504e-01 -2.43672922e-01 -4.68053073e-01 -2.29360133e-01
-1.04054272e+00 3.39474417e-02 8.88896048e-01 7.19479859e-01
4.33917224e-01 7.12102771e-01 -2.54638523e-01 1.43546033e+00
7.77944148e-01 7.34599888e-01 5.95652103e-01 -6.52873456e-01
6.60932779e-01 1.26177035e-02 -6.88397735e-02 -5.89926958e-01
-3.98702413e-01 -2.04744816e-01 -4.23186690e-01 2.48507887e-01
4.18856978e-01 -3.50535899e-01 -1.09152889e+00 1.92191470e+00
8.39106664e-02 7.98059534e-03 3.87881696e-01 6.61683023e-01
1.42365503e+00 5.92115998e-01 1.77858129e-01 -2.37372130e-01
1.50522459e+00 -6.41655087e-01 -8.11390996e-01 -4.79498729e-02
2.84503579e-01 -1.01712728e+00 1.13121355e+00 4.53286171e-01
-1.17214966e+00 -7.69736290e-01 -1.04791999e+00 3.02014798e-01
-4.82038319e-01 2.13232398e-01 4.46115285e-02 1.49056113e+00
-1.57536876e+00 -1.89219549e-01 -4.03278232e-01 -6.29252136e-01
2.42384826e-03 4.81230795e-01 -5.87609410e-01 -2.30532840e-01
-1.38692594e+00 9.03582692e-01 2.34143212e-02 1.24326885e-01
-1.06498575e+00 -4.64511186e-01 -1.03257966e+00 -1.61462516e-01
4.41328287e-02 2.07382888e-02 1.39258873e+00 -1.11759472e+00
-1.57277191e+00 9.76123691e-01 -3.87150019e-01 -2.68155456e-01
3.35725337e-01 1.35361059e-02 -1.31078207e+00 -3.59094255e-02
-1.03774816e-01 8.84115219e-01 6.29557908e-01 -1.20198667e+00
-6.82316661e-01 -4.57736194e-01 -3.25799644e-01 2.59436905e-01
-3.95484358e-01 7.49442637e-01 -6.08767688e-01 -4.78970498e-01
-1.52243689e-01 -8.39853406e-01 5.19952297e-01 -5.48669934e-01
-2.44352922e-01 -3.16750109e-01 1.04226971e+00 -1.11247683e+00
9.99366581e-01 -2.53164387e+00 -3.85065824e-01 3.35272938e-01
-2.75967300e-01 3.17360193e-01 -4.29059803e-01 3.93981457e-01
-4.36405301e-01 1.65484995e-01 1.98543854e-02 -5.29704809e-01
3.18134993e-01 -6.09458312e-02 -3.77442166e-02 4.34859425e-01
-6.72394633e-02 5.16797185e-01 -2.25186601e-01 -2.63870507e-01
2.30930090e-01 8.73648524e-01 -1.51254594e-01 -5.26242107e-02
3.55068177e-01 2.64429361e-01 2.58746862e-01 1.09079432e+00
8.20997238e-01 3.34609151e-01 3.92009951e-02 -3.99265051e-01
-2.83321470e-01 3.29347938e-01 -1.46849144e+00 1.32542062e+00
-3.17487180e-01 9.19831932e-01 8.29673231e-01 -6.15721047e-01
9.43086028e-01 1.05774188e+00 2.72463113e-01 -7.19428778e-01
-7.91275222e-03 4.18196172e-01 -1.43887907e-01 -5.37180007e-01
5.42973101e-01 -4.57960628e-02 4.32054661e-02 3.18567336e-01
3.08773816e-01 1.69074222e-01 6.91424459e-02 4.35413308e-02
5.94150007e-01 -3.16839159e-01 -1.56057209e-01 -7.23579824e-02
9.19429660e-01 -2.85067320e-01 5.99582553e-01 6.52776361e-01
-6.83408380e-01 5.19614398e-01 -1.67847112e-01 4.49558869e-02
-8.65012586e-01 -1.10392666e+00 -3.96626443e-01 1.40485811e+00
-4.79512215e-01 -2.82469809e-01 -8.50478888e-01 -3.13653231e-01
-8.87979846e-03 6.10806942e-01 -1.81418687e-01 1.98141724e-01
-4.00627941e-01 -4.26268876e-01 1.30013287e+00 4.10696089e-01
5.99101543e-01 -1.10111141e+00 1.65863171e-01 1.19977191e-01
-2.78585225e-01 -1.23050761e+00 -7.19063044e-01 5.19330092e-02
-3.60032409e-01 -8.79940152e-01 -9.56891000e-01 -1.11531305e+00
3.14627476e-02 -2.76530087e-02 9.27967846e-01 -5.74844897e-01
-3.53972137e-01 1.06842875e+00 -5.71222305e-02 -2.67970711e-01
-5.08738399e-01 -2.41277054e-01 4.44313139e-01 2.04875752e-01
6.77582741e-01 -8.05451572e-02 -1.46680236e-01 8.04662704e-01
-6.28195941e-01 -5.22817850e-01 5.23884892e-01 9.26270485e-01
2.26089805e-01 -2.53614277e-01 1.02069330e+00 -7.46609196e-02
9.27336514e-01 -1.00619920e-01 -5.11634827e-01 5.83396196e-01
-4.98469949e-01 -6.02981567e-01 -9.59457234e-02 -4.40457433e-01
-1.23709381e+00 4.25951108e-02 -4.47856933e-01 -2.73665071e-01
-6.05837047e-01 6.02277696e-01 -4.91758645e-01 -2.78060436e-01
6.85212255e-01 3.35923433e-01 2.85790890e-01 -5.11690855e-01
1.16539553e-01 1.47250581e+00 6.86336935e-01 -5.52458107e-01
4.45143849e-01 -4.85441312e-02 -8.59057486e-01 -1.28037667e+00
2.82561839e-01 -4.21682805e-01 -2.39053264e-01 -5.05176604e-01
8.53749633e-01 -1.44709074e+00 -8.17546427e-01 9.60806191e-01
-1.02390468e+00 -3.31682488e-02 3.91628928e-02 8.55823219e-01
-2.44359282e-04 2.36720964e-01 -6.69021964e-01 -1.15949392e+00
-3.41257185e-01 -1.39267480e+00 9.49053645e-01 1.70988023e-01
-2.76791096e-01 -7.13120818e-01 2.48190202e-02 7.46497810e-01
7.04429686e-01 -9.02021825e-02 5.11454761e-01 -9.09193397e-01
-8.38190168e-02 -2.59575605e-01 -2.39848986e-01 6.58799350e-01
7.21000955e-02 1.41045138e-01 -1.68209362e+00 -6.95084393e-01
-1.79026395e-01 -6.81591809e-01 4.81716484e-01 3.97042781e-01
3.88603598e-01 -1.14565879e-01 -4.30398546e-02 1.36469677e-01
8.23478639e-01 6.34915948e-01 5.98804116e-01 7.52023160e-02
2.67758340e-01 1.00506008e+00 1.68790191e-01 1.46478757e-01
3.69936079e-01 1.09895492e+00 -7.12774694e-02 6.55369237e-02
-2.65392631e-01 5.85685745e-02 1.04154706e+00 7.88649738e-01
8.63623172e-02 -1.25404373e-01 -1.12395465e+00 4.73611057e-01
-1.20394862e+00 -9.12354648e-01 -3.45857698e-03 2.42889190e+00
6.21195316e-01 -2.15383187e-01 4.67022032e-01 3.69478762e-01
1.05851018e+00 -1.10373005e-01 -5.00062466e-01 -5.35400391e-01
-6.45453513e-01 1.43314034e-01 5.72604835e-02 6.42628193e-01
-1.00045991e+00 6.21300876e-01 6.78800154e+00 9.73131061e-01
-1.30184853e+00 2.46871829e-01 5.81846297e-01 -3.06371629e-01
-1.08376771e-01 -6.42525792e-01 -1.12548733e+00 1.31461814e-01
1.55305946e+00 -1.44959375e-01 4.57289875e-01 6.57409191e-01
1.40258431e-01 2.21896961e-01 -1.03968799e+00 1.26231909e+00
4.37099665e-01 -9.82082784e-01 -2.08663300e-01 4.38356809e-02
3.63546580e-01 5.27139008e-01 1.75491035e-01 6.69401765e-01
7.77068213e-02 -1.16933370e+00 9.81895685e-01 -4.02352475e-02
1.24759316e+00 -7.02631593e-01 7.59811461e-01 -1.43194556e-01
-1.40760720e+00 -1.98283508e-01 3.31965953e-01 4.62964833e-01
2.09526628e-01 8.72207507e-02 -6.69281721e-01 6.26066566e-01
9.57711160e-01 3.03502262e-01 -4.65258598e-01 7.77394354e-01
2.05702156e-01 7.17442214e-01 -3.09994996e-01 1.87370986e-01
-8.84023681e-02 2.96122611e-01 6.85400009e-01 1.52944434e+00
2.48389348e-01 -4.03362066e-01 1.91016182e-01 2.35913321e-01
-4.59995382e-02 1.05255969e-01 -3.88528973e-01 9.08090696e-02
8.33681941e-01 1.00712025e+00 -4.72379141e-02 -2.07190201e-01
-5.88952541e-01 4.84944224e-01 -2.90929496e-01 7.80112505e-01
-6.62976861e-01 -4.70000148e-01 6.47745967e-01 -1.94309473e-01
6.17589569e-03 6.42294958e-02 -6.61573261e-02 -7.28352904e-01
1.13360055e-01 -1.42615259e+00 5.73419452e-01 -6.22880340e-01
-1.26691985e+00 9.20042038e-01 -4.73587252e-02 -1.01997852e+00
-5.01168668e-01 -5.63313723e-01 -4.03114766e-01 1.25529957e+00
-1.22985470e+00 -1.11116219e+00 -2.53745317e-01 9.51512396e-01
5.26704431e-01 -6.92772150e-01 1.07513571e+00 9.17764425e-01
-7.28023946e-01 1.32489443e+00 1.35615543e-01 1.66570589e-01
1.05063963e+00 -7.45537937e-01 9.31536630e-02 4.02248830e-01
-7.16845840e-02 6.12857521e-01 3.74477595e-01 -2.60858983e-01
-1.41759324e+00 -7.95615971e-01 1.03507757e+00 -1.15015268e-01
3.63622010e-01 -5.02498090e-01 -7.07219660e-01 6.51684105e-01
5.39951384e-01 -5.13534844e-01 9.34935391e-01 1.36793360e-01
-6.27134442e-01 -2.98931837e-01 -1.54179299e+00 3.65846187e-01
4.86574203e-01 -1.12279797e+00 -3.79666030e-01 1.13025343e-03
5.28257489e-01 -2.69025952e-01 -8.54468107e-01 4.53553408e-01
9.63462412e-01 -7.26100087e-01 1.07648408e+00 -2.42128506e-01
-2.28810355e-01 -3.10954273e-01 -7.33860791e-01 -9.40087318e-01
5.69382459e-02 -6.11438453e-01 1.60876498e-01 1.82218242e+00
7.51066208e-01 -9.00380969e-01 5.16278327e-01 7.64429986e-01
-4.58632886e-01 -1.99345514e-01 -1.36950123e+00 -1.03860486e+00
-1.33114576e-01 -7.60525525e-01 4.91014451e-01 9.44177985e-01
-1.72647804e-01 2.52316326e-01 -3.33866358e-01 1.78257972e-01
5.22868812e-01 -6.31855428e-01 7.24927962e-01 -9.71251726e-01
-1.01623587e-01 -6.34571373e-01 -4.72951800e-01 -7.25922942e-01
4.16020639e-02 -7.48132646e-01 4.42685336e-02 -1.19401598e+00
-1.24971613e-01 -1.20731503e-01 -2.63547242e-01 6.66419327e-01
2.93162525e-01 3.61951083e-01 1.49787515e-01 2.30760708e-01
-7.49754086e-02 3.93483102e-01 4.05879140e-01 -4.55067575e-01
-3.13419670e-01 -1.23123899e-02 -6.29971147e-01 2.97558397e-01
6.49285376e-01 4.94696982e-02 -7.76637048e-02 -2.77683645e-01
-4.90516275e-01 4.70579229e-02 4.62260544e-01 -1.03773952e+00
3.27993184e-01 3.91296625e-01 3.71087253e-01 -6.52851462e-01
6.85213685e-01 -6.17578506e-01 3.94080758e-01 2.18892172e-01
-3.71153593e-01 9.01319608e-02 7.64050961e-01 1.39862463e-01
-6.84435904e-01 6.90582395e-02 8.17888021e-01 4.23267603e-01
-7.59798408e-01 -1.89783886e-01 -5.69748580e-01 -1.67208984e-01
8.39133799e-01 -4.59556967e-01 -5.95343947e-01 -5.89813650e-01
-1.12606370e+00 2.59534121e-01 2.46665552e-02 7.05437422e-01
5.26338220e-01 -1.47340393e+00 -1.10321379e+00 3.99069548e-01
3.13808322e-01 -7.11862445e-01 5.06141603e-01 9.60024476e-01
-4.71834801e-02 6.83561921e-01 -6.21111915e-02 -8.61664295e-01
-1.98071468e+00 -1.37335151e-01 3.75115514e-01 8.82551596e-02
2.90891062e-02 8.62121403e-01 -5.24704084e-02 -9.08354998e-01
9.64588761e-01 2.38555148e-01 -3.75736564e-01 4.62556094e-01
8.30560327e-01 4.34260964e-01 4.65702981e-01 -1.19024324e+00
-7.39192665e-01 4.11275446e-01 -1.47627816e-01 -7.35900879e-01
1.02055681e+00 -2.19886050e-01 2.29202554e-01 6.25319779e-01
1.28783023e+00 1.27378376e-02 -5.94604909e-01 -2.59861946e-01
-5.73626757e-02 -7.95507580e-02 8.90383273e-02 -1.20125937e+00
-1.00091219e+00 7.81182766e-01 1.30680120e+00 2.00747713e-01
1.10406506e+00 -1.90809906e-01 3.33564907e-01 2.12115154e-01
1.55519500e-01 -1.29582357e+00 -3.18904638e-01 5.92776418e-01
1.18275023e+00 -1.14125252e+00 -5.10465264e-01 -1.04534581e-01
-7.04829633e-01 8.92610073e-01 5.23093522e-01 8.58252347e-01
5.80412030e-01 2.06375405e-01 4.89474177e-01 8.61648992e-02
-5.15734375e-01 1.16730750e-01 4.13347304e-01 8.16243768e-01
8.15995574e-01 4.11999412e-02 1.78473741e-01 5.77633142e-01
-1.16849534e-01 2.65006861e-03 -3.46598215e-02 9.29335415e-01
-1.89053431e-01 -1.11005485e+00 -1.01544428e+00 2.22498164e-01
-3.68880510e-01 -1.64757177e-01 -4.95435059e-01 8.16935301e-01
-4.67448197e-02 1.56155121e+00 -8.40852633e-02 -6.48646235e-01
5.79163671e-01 4.89490986e-01 1.13958940e-01 -2.33321667e-01
-1.01744413e+00 3.30553591e-01 6.89106405e-01 -4.17308122e-01
-3.69936734e-01 -9.63323832e-01 -8.67241859e-01 -4.75179076e-01
-3.14468563e-01 3.26350361e-01 1.17529500e+00 6.12347722e-01
4.36182350e-01 5.09539068e-01 7.22615778e-01 -1.00312495e+00
-5.97821355e-01 -1.38626575e+00 -6.04329348e-01 2.19666123e-01
3.02076876e-01 -3.97646546e-01 -6.54825091e-01 8.84089842e-02] | [14.285845756530762, 6.119856834411621] |
74219203-403c-4e37-ac4b-9e8f0d337d4d | a-fully-first-order-method-for-stochastic | 2301.10945 | null | https://arxiv.org/abs/2301.10945v1 | https://arxiv.org/pdf/2301.10945v1.pdf | A Fully First-Order Method for Stochastic Bilevel Optimization | We consider stochastic unconstrained bilevel optimization problems when only the first-order gradient oracles are available. While numerous optimization methods have been proposed for tackling bilevel problems, existing methods either tend to require possibly expensive calculations regarding Hessians of lower-level objectives, or lack rigorous finite-time performance guarantees. In this work, we propose a Fully First-order Stochastic Approximation (F2SA) method, and study its non-asymptotic convergence properties. Specifically, we show that F2SA converges to an $\epsilon$-stationary solution of the bilevel problem after $\epsilon^{-7/2}, \epsilon^{-5/2}$, and $\epsilon^{-3/2}$ iterations (each iteration using $O(1)$ samples) when stochastic noises are in both level objectives, only in the upper-level objective, and not present (deterministic settings), respectively. We further show that if we employ momentum-assisted gradient estimators, the iteration complexities can be improved to $\epsilon^{-5/2}, \epsilon^{-4/2}$, and $\epsilon^{-3/2}$, respectively. We demonstrate even superior practical performance of the proposed method over existing second-order based approaches on MNIST data-hypercleaning experiments. | ['Robert Nowak', 'Stephen Wright', 'Dohyun Kwon', 'Jeongyeol Kwon'] | 2023-01-26 | null | null | null | null | ['bilevel-optimization'] | ['methodology'] | [-2.13184599e-02 4.43891808e-02 4.22571376e-02 -1.88874483e-01
-1.16839349e+00 -5.87880611e-01 3.45588736e-02 3.04675907e-01
-8.26747358e-01 9.15216744e-01 -3.96070272e-01 -5.78915894e-01
-8.78209949e-01 -6.25453234e-01 -9.37523961e-01 -9.77695048e-01
-3.14100832e-01 4.46566463e-01 -7.48636248e-03 -6.94716498e-02
3.55219215e-01 6.86363950e-02 -1.30083895e+00 -3.49715114e-01
1.18451452e+00 1.24131346e+00 9.21336412e-02 7.58795321e-01
-3.63565795e-02 4.90789831e-01 -3.75951648e-01 -7.52347410e-01
4.67932701e-01 -5.78089654e-01 -3.68574321e-01 -9.40776765e-02
3.87831360e-01 8.63684267e-02 -1.38864622e-01 1.76642716e+00
3.92921746e-01 3.17667216e-01 2.17757165e-01 -9.55979884e-01
-3.83588493e-01 5.33679068e-01 -6.20977700e-01 1.11330830e-01
-8.44888762e-02 2.60384083e-01 1.06462348e+00 -8.60719204e-01
2.93297946e-01 8.73049438e-01 8.11841726e-01 2.26733685e-01
-1.22779739e+00 -7.01107025e-01 3.02130342e-01 -3.77216446e-03
-1.13949752e+00 -2.58708805e-01 6.73418343e-01 -2.24369109e-01
6.78560734e-01 5.61058462e-01 3.48674506e-01 6.59004688e-01
-1.04336329e-01 8.23777497e-01 1.24099648e+00 -3.51609886e-01
2.71993756e-01 2.64865816e-01 3.08710188e-01 9.20342982e-01
6.93218231e-01 1.33578598e-01 -3.69384766e-01 -3.31170410e-01
5.10403991e-01 1.83084346e-02 -3.68331730e-01 -1.02193989e-01
-9.63266134e-01 9.15610075e-01 5.53608537e-02 5.20846993e-02
-5.30016661e-01 5.08139431e-01 2.61935502e-01 1.94269896e-01
4.70469892e-01 5.13165236e-01 -6.46919310e-01 -6.85436249e-01
-1.03820002e+00 3.03043902e-01 8.87453020e-01 9.85790968e-01
4.74607259e-01 5.73281907e-02 -2.55514562e-01 8.97580683e-01
3.30164105e-01 8.82776916e-01 2.11634561e-01 -9.12591815e-01
9.11540329e-01 2.16795057e-01 4.51756120e-01 -9.91407394e-01
-3.33331645e-01 -6.94436193e-01 -8.89473617e-01 -1.16201095e-01
7.62700319e-01 -3.87301594e-01 -7.29170442e-01 1.78616774e+00
3.71091485e-01 9.90488380e-02 -1.02973253e-01 7.79021025e-01
1.45142823e-01 3.76541615e-01 -2.26550400e-01 -7.00395167e-01
1.29982483e+00 -8.11785936e-01 -5.82711399e-01 -3.34922791e-01
3.20447564e-01 -8.42939138e-01 1.31564140e+00 5.10492563e-01
-1.53763402e+00 -9.91853476e-02 -8.59495699e-01 2.94268191e-01
1.83458388e-01 1.90536186e-01 7.37552643e-01 9.24103618e-01
-5.52223146e-01 7.53156662e-01 -9.16908503e-01 3.98918211e-01
4.75995183e-01 3.11295390e-01 1.64369438e-02 3.30063212e-03
-7.87964106e-01 3.16373378e-01 1.49359599e-01 4.85395432e-01
-8.64577651e-01 -8.35456133e-01 -6.20000601e-01 8.73744041e-02
8.40366662e-01 -2.27270231e-01 9.50511396e-01 -1.98574752e-01
-1.36979115e+00 3.61045092e-01 -3.28171670e-01 -4.00209188e-01
7.70021021e-01 -3.26965660e-01 -1.17822036e-01 -8.27349722e-02
1.75028145e-02 -3.30289721e-01 5.60181081e-01 -1.08952010e+00
-8.26441109e-01 -5.71787953e-01 2.24057913e-01 1.93286147e-02
-4.02166039e-01 -1.57442883e-01 -6.51889443e-01 -7.02023327e-01
4.50707644e-01 -1.03475964e+00 -3.97527844e-01 -3.63757282e-01
-5.64421892e-01 1.36378303e-01 -4.20301706e-02 -7.79840648e-01
1.37918961e+00 -2.09932756e+00 5.16427904e-02 4.52734798e-01
-2.22098976e-02 3.05039406e-01 1.49815723e-01 1.82992816e-01
2.66194463e-01 2.76344359e-01 -5.14432371e-01 -5.19628942e-01
2.30673149e-01 2.27797687e-01 -5.25947064e-02 5.88529646e-01
-3.76270026e-01 4.39950407e-01 -1.12215304e+00 -2.64560655e-02
-7.64504261e-03 2.58743256e-01 -5.73768318e-01 -2.34781981e-01
-4.38813530e-02 1.35853933e-02 -5.64208686e-01 6.24568403e-01
8.63276839e-01 -3.33730847e-01 3.71365726e-01 -1.34648621e-01
-5.04750609e-02 1.01691894e-01 -1.79214692e+00 1.23332834e+00
-7.30851710e-01 2.80821443e-01 4.40812111e-01 -1.32052314e+00
5.52789927e-01 9.81724784e-02 5.42905092e-01 -5.08786917e-01
-3.22151333e-02 5.02828538e-01 -1.25704929e-01 -3.94079536e-01
2.33105347e-01 -5.65663099e-01 8.26942250e-02 2.86167562e-01
-1.98549032e-01 -2.83108316e-02 6.45609438e-01 -2.73902088e-01
1.18766463e+00 -2.05352649e-01 -9.91766155e-02 -3.70971859e-01
4.26917702e-01 -2.28249788e-01 8.81137431e-01 1.30428970e+00
-1.19706817e-01 5.05854607e-01 9.67494428e-01 -2.21605569e-01
-8.39258611e-01 -8.61852407e-01 -1.24186791e-01 1.00908113e+00
2.92552203e-01 -2.84061790e-01 -7.00589597e-01 -9.16319132e-01
9.48024690e-02 1.06061983e+00 -5.29605150e-01 -7.41354376e-02
-6.08252048e-01 -1.52031577e+00 5.79652846e-01 4.46307927e-01
6.34146154e-01 -7.20035434e-01 -5.19097388e-01 4.20052260e-01
1.88485160e-02 -7.50789940e-01 -6.07009172e-01 5.74128687e-01
-9.11249042e-01 -1.09671998e+00 -5.65931976e-01 -2.79554307e-01
9.47381139e-01 5.71673065e-02 8.73363137e-01 -3.08928383e-03
-1.07914083e-01 1.07551575e-01 -1.73531950e-01 -2.87804961e-01
1.05827838e-01 -2.52136946e-01 8.65956247e-02 3.84223498e-02
1.60966635e-01 -5.76299667e-01 -6.49816632e-01 2.20121503e-01
-7.79928446e-01 -4.59037036e-01 5.60157478e-01 1.15449667e+00
9.81260955e-01 2.84163952e-01 1.94363758e-01 -1.19778383e+00
7.27695048e-01 -3.20391715e-01 -1.12255228e+00 2.45451450e-01
-1.00993204e+00 3.76083016e-01 8.66592169e-01 -5.19694805e-01
-7.67416358e-01 -2.39349738e-01 -1.33472711e-01 -5.48290253e-01
4.15174603e-01 8.01723361e-01 2.92490255e-02 1.40949756e-01
6.60166085e-01 5.25443494e-01 -3.05898517e-01 -7.17126191e-01
1.83650821e-01 2.68239617e-01 5.56323767e-01 -7.60633707e-01
9.60435629e-01 4.84813482e-01 8.54181722e-02 -4.49432969e-01
-1.11259413e+00 -2.67625362e-01 1.72897786e-01 3.90757859e-01
3.64450812e-01 -5.28752923e-01 -1.32801747e+00 8.06507282e-03
-4.58445072e-01 -1.60049841e-01 -4.61403102e-01 7.00801611e-01
-4.35291350e-01 4.99235034e-01 -3.02841067e-01 -1.32958913e+00
-4.17620778e-01 -1.27531970e+00 5.16810894e-01 3.09202313e-01
1.31007060e-01 -8.67873311e-01 -2.13491276e-01 5.33625841e-01
3.70665878e-01 1.04273096e-01 7.45407224e-01 -6.77400410e-01
-4.07491088e-01 -5.27580380e-01 -2.34684125e-01 6.87183619e-01
-8.98323581e-02 -3.43723655e-01 -3.41580987e-01 -3.06217730e-01
2.02254847e-01 -4.37763147e-02 6.94060743e-01 4.29189354e-01
1.25546777e+00 -8.81466568e-01 -9.02740136e-02 7.53475964e-01
1.33192170e+00 2.35509127e-01 2.93487906e-01 2.93871075e-01
4.27414477e-01 1.67984888e-01 6.37624443e-01 8.33465695e-01
2.88814336e-01 2.89340377e-01 5.45541108e-01 1.11521550e-01
2.52406090e-01 -6.41833171e-02 2.53148973e-01 6.74971342e-01
-1.58873245e-01 -1.37984604e-01 -6.36533618e-01 8.18394899e-01
-1.77210391e+00 -8.33624005e-01 -2.35443056e-01 2.69205117e+00
1.19425905e+00 3.57594132e-01 -1.25321686e-01 1.51614547e-01
5.09353220e-01 2.97952667e-02 -7.92144060e-01 -2.40748242e-01
-1.09839022e-01 5.13286769e-01 8.01425457e-01 5.84608376e-01
-9.42637444e-01 6.46614313e-01 4.53072596e+00 1.07045138e+00
-7.12731361e-01 1.62019238e-01 4.83627856e-01 -6.06614769e-01
-4.92512167e-01 3.39247912e-01 -7.07352757e-01 8.71844828e-01
7.65206218e-01 -3.93499769e-02 9.28850591e-01 1.06299698e+00
2.30302401e-02 -3.32543969e-01 -8.78257275e-01 1.08083582e+00
-1.41887605e-01 -1.18248558e+00 -5.54535210e-01 2.35393018e-01
8.46418202e-01 -1.37615040e-01 2.59186268e-01 3.32967907e-01
5.21530032e-01 -8.42153847e-01 6.24381244e-01 2.29202479e-01
5.58722734e-01 -1.03741026e+00 8.75105441e-01 4.58863050e-01
-1.00376928e+00 -4.10997331e-01 -3.02644044e-01 2.77542531e-01
2.40899831e-01 1.24577391e+00 -3.67161185e-01 6.32925451e-01
9.55298543e-01 -1.15524143e-01 -9.07670707e-03 9.68047440e-01
-3.00842017e-01 6.62959993e-01 -8.54026675e-01 -2.29313195e-01
3.19611609e-01 -5.78617334e-01 6.55364990e-01 9.92053866e-01
6.02514505e-01 2.61845022e-01 1.77173138e-01 5.05657971e-01
-3.01018476e-01 1.38179198e-01 2.67559111e-01 -2.09784359e-01
4.36316133e-01 1.00803840e+00 -6.20163381e-01 -2.83001959e-01
-1.54510602e-01 8.00349593e-01 2.25162908e-01 4.77783352e-01
-1.01468432e+00 -1.00322163e+00 7.91541457e-01 -5.15277535e-02
6.30374312e-01 -2.68389136e-01 -5.14444947e-01 -1.23367345e+00
5.97076714e-01 -1.04123521e+00 5.19812584e-01 -4.94179782e-03
-1.07780170e+00 3.75935644e-01 -1.40723258e-01 -5.89630961e-01
-1.85650960e-01 -4.88201469e-01 -2.94356823e-01 7.96090186e-01
-1.28174376e+00 -4.65515018e-01 -1.13785848e-01 3.70604247e-01
2.16758698e-01 -7.46224970e-02 4.19495255e-01 4.68649805e-01
-7.08117068e-01 8.88094664e-01 7.73777425e-01 -1.30475968e-01
1.38507396e-01 -1.14215171e+00 -9.88330841e-02 9.26226318e-01
2.51649171e-01 7.76586056e-01 1.13906801e+00 -5.18589854e-01
-1.69680178e+00 -7.49275744e-01 8.15979064e-01 -1.15900621e-01
5.16013741e-01 -3.25183719e-01 -6.36391401e-01 5.45522451e-01
-3.26293290e-01 2.83060759e-01 4.28600520e-01 2.59977460e-01
-1.92551538e-02 -1.95101902e-01 -1.36285996e+00 7.17576981e-01
1.10560501e+00 -3.50983977e-01 -1.63129359e-01 6.13585711e-01
3.60761970e-01 -6.44599974e-01 -1.01858711e+00 5.55394053e-01
4.06231195e-01 -9.78238225e-01 9.50428963e-01 -6.41037405e-01
5.19098230e-02 -1.55593365e-01 -4.96633381e-01 -9.54440415e-01
8.72765109e-02 -1.08036053e+00 -4.04628456e-01 7.43015587e-01
6.84858680e-01 -7.83574939e-01 9.34219480e-01 5.84320903e-01
-2.51299851e-02 -9.43609357e-01 -1.28051674e+00 -8.75620723e-01
6.19669557e-02 -7.82330155e-01 1.00344181e-01 6.90178394e-01
-2.22619236e-01 -2.41996661e-01 -5.75693369e-01 3.05478185e-01
7.68497467e-01 1.47484615e-01 5.78472257e-01 -8.94707918e-01
-8.94215167e-01 -4.10044402e-01 1.72901303e-01 -1.09407949e+00
-2.03828990e-01 -6.35154903e-01 2.75194168e-01 -1.23216653e+00
8.96255448e-02 -7.81995177e-01 -6.92718983e-01 4.65309560e-01
-5.20373523e-01 -3.14068869e-02 1.38324335e-01 -1.87815085e-01
-4.77077156e-01 6.19270205e-01 8.05112422e-01 9.08424705e-02
-3.36181283e-01 2.19387636e-01 -8.02108288e-01 8.49397480e-01
4.49037611e-01 -6.91426277e-01 -4.31901664e-01 -4.89682525e-01
6.71365559e-01 2.30406344e-01 8.00463110e-02 -7.51163304e-01
1.39179096e-01 -3.15011859e-01 1.51677085e-02 -4.49172020e-01
3.31621647e-01 -4.72094655e-01 8.66811275e-02 4.75663394e-01
-2.45414078e-01 5.00046788e-03 1.04813181e-01 7.11012304e-01
-1.99502446e-02 -7.37162530e-01 7.24191725e-01 -1.41292900e-01
1.05581522e-01 1.32036060e-01 8.84505361e-02 3.83981764e-01
9.08523798e-01 1.29264668e-01 -2.29702927e-02 -2.92380482e-01
-8.04037869e-01 3.44531626e-01 7.28704631e-02 -1.00600727e-01
4.34978217e-01 -1.03404868e+00 -6.63899779e-01 -1.84668507e-02
-4.05444980e-01 2.73061156e-01 3.25751454e-01 1.26877332e+00
-3.62848550e-01 2.41435796e-01 6.56114340e-01 -3.50630134e-01
-8.39726686e-01 3.98140311e-01 1.95916116e-01 -6.66088283e-01
-2.60511190e-01 1.40905893e+00 -1.12098299e-01 -4.45854872e-01
3.14704746e-01 -2.32099369e-01 5.75389028e-01 -1.65543295e-02
3.82658869e-01 8.79898727e-01 1.56750649e-01 1.10455044e-01
-3.94267410e-01 3.55869919e-01 -6.54063821e-02 -3.78483921e-01
1.38896263e+00 3.56471539e-03 -4.78108972e-02 1.95789561e-01
1.15016103e+00 3.53682578e-01 -1.29183018e+00 -1.98985055e-01
1.58535451e-01 -5.43215990e-01 2.83509530e-02 -8.34335744e-01
-1.02255273e+00 7.97490239e-01 6.24368906e-01 1.83530971e-01
1.11742067e+00 -3.09307933e-01 7.72211790e-01 4.40074801e-01
3.44084114e-01 -1.46540046e+00 -6.14354759e-02 4.10023332e-01
5.33532679e-01 -1.14058423e+00 1.62025556e-01 -2.34157860e-01
-5.35864949e-01 8.18479776e-01 2.05115974e-01 -2.39966139e-02
5.95110595e-01 1.40416687e-02 -3.33131462e-01 -5.12913615e-02
-5.13533294e-01 5.10511734e-02 1.99644089e-01 -1.40004372e-02
1.95518285e-01 -2.77091693e-02 -5.62413454e-01 8.48572969e-01
-2.15825588e-01 -1.27291843e-01 3.38647485e-01 1.03325129e+00
-3.13274562e-01 -8.89002562e-01 -2.61106431e-01 6.84179902e-01
-7.83838451e-01 -4.32994932e-01 3.31394523e-01 6.76384270e-01
1.13742184e-02 9.79315460e-01 -2.88011611e-01 4.69210418e-03
4.42426413e-01 1.19141646e-01 4.32893395e-01 -2.37433299e-01
-4.87522453e-01 2.94780672e-01 1.98853269e-01 -6.97189331e-01
8.61553475e-02 -8.28194797e-01 -1.30768716e+00 -3.12316895e-01
-3.51543844e-01 4.79677618e-01 8.25469792e-01 9.86689270e-01
3.99911046e-01 2.83886820e-01 6.31572604e-01 -3.20693493e-01
-1.22787035e+00 -8.89445722e-01 -4.63374823e-01 2.76658475e-01
1.47887230e-01 -5.29073834e-01 -7.39071667e-01 -2.64690131e-01] | [6.5250701904296875, 4.529043674468994] |
1ce30cb9-1ad7-4451-8eaa-070c9d9ae2ab | a-self-supervised-learning-based-approach-to | 2302.13457 | null | https://arxiv.org/abs/2302.13457v2 | https://arxiv.org/pdf/2302.13457v2.pdf | A Self-Supervised Learning-based Approach to Clustering Multivariate Time-Series Data with Missing Values (SLAC-Time): An Application to TBI Phenotyping | Self-supervised learning approaches provide a promising direction for clustering multivariate time-series data. However, real-world time-series data often include missing values, and the existing approaches require imputing missing values before clustering, which may cause extensive computations and noise and result in invalid interpretations. To address these challenges, we present a Self-supervised Learning-based Approach to Clustering multivariate Time-series data with missing values (SLAC-Time). SLAC-Time is a Transformer-based clustering method that uses time-series forecasting as a proxy task for leveraging unlabeled data and learning more robust time-series representations. This method jointly learns the neural network parameters and the cluster assignments of the learned representations. It iteratively clusters the learned representations with the K-means method and then utilizes the subsequent cluster assignments as pseudo-labels to update the model parameters. To evaluate our proposed approach, we applied it to clustering and phenotyping Traumatic Brain Injury (TBI) patients in the Transforming Research and Clinical Knowledge in Traumatic Brain Injury (TRACK-TBI) study. Our experiments demonstrate that SLAC-Time outperforms the baseline K-means clustering algorithm in terms of silhouette coefficient, Calinski Harabasz index, Dunn index, and Davies Bouldin index. We identified three TBI phenotypes that are distinct from one another in terms of clinically significant variables as well as clinical outcomes, including the Extended Glasgow Outcome Scale (GOSE) score, Intensive Care Unit (ICU) length of stay, and mortality rate. The experiments show that the TBI phenotypes identified by SLAC-Time can be potentially used for developing targeted clinical trials and therapeutic strategies. | ['Vignesh Subbian', 'Chandan K. Reddy', 'Sindhu Tipirneni', 'Amin Nayebi', 'Brandon Foreman', 'Hamid Ghaderi'] | 2023-02-27 | null | null | null | null | ['clinical-knowledge', 'clustering-multivariate-time-series'] | ['miscellaneous', 'time-series'] | [ 2.74210684e-02 -5.07717967e-01 -1.99155480e-01 -5.48877597e-01
-9.33528960e-01 -3.42246681e-01 7.34841730e-03 5.24385989e-01
-3.93249840e-01 6.43676460e-01 3.81557345e-01 -3.37748498e-01
-7.66712725e-01 -2.29092285e-01 -2.83091992e-01 -1.23449087e+00
-6.22184396e-01 8.04755211e-01 -2.30793640e-01 2.67912418e-01
1.21181101e-01 6.49533391e-01 -1.45685637e+00 6.38559282e-01
8.59500170e-01 8.29670429e-01 5.55037856e-02 1.20890342e-01
1.11401025e-02 9.49619710e-01 -5.07819831e-01 5.09296477e-01
1.19097561e-01 -3.90078604e-01 -3.63180786e-01 7.23744109e-02
-4.65564072e-01 1.85225084e-02 -2.81330556e-01 3.93002659e-01
4.57509011e-01 2.18125939e-01 9.89487052e-01 -1.68101275e+00
-1.16159648e-01 5.26642799e-01 -4.96305972e-01 1.49636462e-01
-3.17790091e-01 -8.23385939e-02 2.43233055e-01 -7.75830269e-01
8.76089036e-02 8.60559165e-01 8.87216687e-01 3.90524268e-01
-1.04720914e+00 -6.73969448e-01 5.24667539e-02 5.91600180e-01
-1.52973151e+00 -8.11278373e-02 8.76387656e-01 -8.32143843e-01
6.65143788e-01 1.62439436e-01 5.53332686e-01 7.77379394e-01
2.76455730e-01 4.03582394e-01 1.10980332e+00 -2.62342453e-01
7.04957187e-01 -4.75683510e-01 2.35529363e-01 2.47995555e-01
9.64300428e-03 1.59924343e-01 -1.47853732e-01 -4.14957941e-01
4.88231391e-01 9.46687758e-01 -1.92642301e-01 -3.71924788e-01
-1.60395336e+00 8.45661700e-01 2.03166530e-01 2.59908259e-01
-7.22794950e-01 -8.16690549e-02 6.42882705e-01 1.99294150e-01
4.10651058e-01 2.81665213e-02 -7.11457074e-01 1.08445259e-02
-1.07928574e+00 -3.37645769e-01 3.04140776e-01 6.33848011e-01
5.20326674e-01 5.39243668e-02 -7.34978542e-02 9.83394682e-01
-1.07845202e-01 2.39983007e-01 1.06072879e+00 -1.05302894e+00
1.64751366e-01 6.35114908e-01 -9.84865129e-02 -7.57740140e-01
-9.40314591e-01 -1.86656728e-01 -1.21713233e+00 4.47380207e-02
2.34524369e-01 -4.10169572e-01 -7.18469322e-01 1.62534785e+00
2.22128600e-01 4.23623860e-01 6.92113303e-03 7.44561374e-01
3.80611867e-01 2.97461301e-01 1.28962398e-01 -8.22303534e-01
8.41921926e-01 -5.28601289e-01 -5.75769901e-01 2.90117174e-01
1.02026343e+00 -3.39410067e-01 6.35107517e-01 4.56814528e-01
-4.94767398e-01 -2.20752612e-01 -5.62607229e-01 6.75545752e-01
-2.67959028e-01 2.14912057e-01 5.01433671e-01 3.63738716e-01
-8.89682055e-01 6.10934854e-01 -1.26814282e+00 -1.89164594e-01
3.79267097e-01 3.43269527e-01 -5.06994605e-01 -9.81190875e-02
-7.71561027e-01 6.70430899e-01 5.57115376e-01 4.68073376e-02
-6.93421841e-01 -7.89544225e-01 -5.07089198e-01 -1.61416858e-01
3.23484875e-02 -4.64712620e-01 5.45581877e-01 -4.62016165e-01
-9.60372806e-01 6.05063379e-01 -3.11041057e-01 -3.39644104e-01
2.28783950e-01 9.94052738e-02 -5.42755365e-01 2.88745075e-01
1.20265670e-01 4.08294022e-01 7.22406030e-01 -1.09463871e+00
-5.56892514e-01 -7.32877553e-01 -7.91452467e-01 1.67435985e-02
-3.35427254e-01 -7.30881328e-03 1.06341034e-01 -8.40462267e-01
6.25192523e-01 -9.08372998e-01 -4.42063272e-01 -3.69183213e-01
-9.62843299e-02 -3.37592781e-01 9.03510809e-01 -6.03554904e-01
1.12442625e+00 -2.33971262e+00 -1.29171893e-01 3.63844991e-01
3.20370257e-01 -1.79843411e-01 1.92730241e-02 4.89998072e-01
-8.19128811e-01 -1.32673368e-01 -5.20371795e-01 -2.87113607e-01
-5.25429428e-01 4.31448221e-01 -2.01181009e-01 7.84877300e-01
-1.20457150e-01 3.87897491e-01 -7.35060155e-01 -6.36733651e-01
4.08275932e-01 3.93735349e-01 -3.60522777e-01 2.63104200e-01
2.79957741e-01 9.31849301e-01 -8.93855020e-02 5.08445740e-01
3.76849294e-01 -1.94827929e-01 5.46914749e-02 -1.16364442e-01
-9.36749578e-02 -4.35838848e-01 -8.57938528e-01 1.36351991e+00
3.56732532e-02 2.97481060e-01 -4.09822047e-01 -1.62267423e+00
9.28426206e-01 5.19909263e-01 1.48887098e+00 -3.47264558e-01
2.53368258e-01 8.10514987e-02 -2.00201258e-01 -7.48576164e-01
-3.80058974e-01 -3.28576148e-01 2.84776371e-02 5.91349602e-01
-2.84992516e-01 1.94175482e-01 -6.71680793e-02 -5.99817932e-03
1.14053619e+00 -3.21185321e-01 1.25751033e-01 -1.83193594e-01
3.76650333e-01 2.61694342e-01 8.61415625e-01 3.23863089e-01
-2.45449409e-01 8.54279041e-01 3.34165215e-01 -5.94557524e-01
-8.58114660e-01 -1.10688174e+00 -2.04477653e-01 9.31912184e-01
-3.92501652e-01 -1.12781502e-01 -6.49901330e-01 -3.95679325e-01
-5.71060665e-02 7.47345924e-01 -6.54991865e-01 -4.55651432e-01
-5.17219543e-01 -1.03620517e+00 3.71199101e-01 7.39125252e-01
-3.03754300e-01 -9.97110009e-01 -4.84411895e-01 2.80326068e-01
-5.42933464e-01 -7.41823196e-01 -2.69800037e-01 5.32521665e-01
-1.32020056e+00 -1.40989733e+00 -6.42965913e-01 -8.75376523e-01
1.00760257e+00 3.82456988e-01 3.79332602e-01 1.11784883e-01
-4.23051089e-01 3.32278609e-01 -5.77436924e-01 -4.02549624e-01
-1.89032093e-01 -3.83323342e-01 4.07104999e-01 2.77842790e-01
4.96518642e-01 -8.67337167e-01 -8.15331459e-01 3.49359125e-01
-1.06002045e+00 -1.49171963e-01 3.61737639e-01 8.82836342e-01
8.10424268e-01 2.49294922e-01 1.03985786e+00 -5.42298019e-01
5.73484361e-01 -8.89177084e-01 -2.37171963e-01 2.46182516e-01
-8.47543120e-01 -2.97355860e-01 7.18273580e-01 -6.72429800e-01
-6.11222506e-01 2.78941453e-01 3.72649878e-01 -9.23358977e-01
-4.63904649e-01 8.10234725e-01 1.14270747e-01 4.74894077e-01
6.18351161e-01 3.30113769e-01 3.23737323e-01 -4.86538410e-01
2.50427425e-02 9.43845868e-01 9.15016532e-01 -6.08655512e-01
4.26755100e-01 7.60717988e-01 -8.91964734e-02 -4.72127765e-01
-6.10000193e-01 -1.02645504e+00 -1.08130836e+00 -3.29810739e-01
7.84855366e-01 -1.08887577e+00 -6.58457100e-01 4.94278759e-01
-7.48486400e-01 -2.96570957e-01 -1.68633476e-01 1.08787799e+00
-7.21146226e-01 4.06683296e-01 -5.26997983e-01 -6.74379408e-01
-2.46107653e-01 -9.51025963e-01 7.54145682e-01 -2.63127387e-01
-3.28374475e-01 -1.01928186e+00 1.40885636e-01 3.25083017e-01
2.47986421e-01 8.29131782e-01 1.55421197e+00 -9.32473123e-01
1.71745539e-01 -4.22727615e-01 2.81422958e-02 3.51481795e-01
4.37130749e-01 -2.29906797e-01 -5.66505253e-01 -4.42709595e-01
3.91852230e-01 7.92440102e-02 5.49858749e-01 6.89058721e-01
1.66521931e+00 -8.89103934e-02 -2.83445984e-01 6.35945141e-01
1.31739271e+00 5.78828216e-01 4.84029621e-01 2.10132211e-01
7.55071759e-01 7.35747933e-01 3.96489859e-01 9.05034661e-01
5.88129818e-01 5.39624505e-02 2.84900367e-01 8.80271476e-03
2.17973486e-01 9.31897759e-02 1.90410763e-01 1.35129774e+00
-1.15943439e-01 2.46032298e-01 -1.26522994e+00 8.29909861e-01
-2.07132316e+00 -1.06363261e+00 -4.32178110e-01 2.32849765e+00
7.18438506e-01 -3.78330499e-01 2.73243189e-01 6.27227426e-01
1.05729973e+00 -5.70827484e-01 -7.30499446e-01 7.42616951e-02
-1.36434555e-01 1.18253857e-01 2.71799892e-01 -1.25163898e-01
-8.80480587e-01 4.54610169e-01 6.34732342e+00 3.88005555e-01
-1.19369042e+00 2.30708003e-01 6.78318977e-01 -2.82142729e-01
2.41571158e-01 -4.46768291e-02 5.61083853e-02 6.34201825e-01
1.17465758e+00 -1.55764401e-01 4.34229642e-01 7.39925861e-01
7.35114813e-01 9.53669846e-02 -1.10755825e+00 1.25062633e+00
1.37652665e-01 -1.16651464e+00 -2.14238212e-01 -1.69194400e-01
6.36756659e-01 1.70172215e-01 -1.68021634e-01 2.29107991e-01
2.35336795e-01 -8.70611012e-01 4.12028939e-01 7.00313568e-01
8.24576974e-01 -7.52104044e-01 7.59355307e-01 5.19014120e-01
-8.55498254e-01 -4.30769145e-01 -2.19068959e-01 -3.03823501e-02
1.78324431e-01 7.69886494e-01 -7.82490730e-01 3.09312314e-01
8.26604962e-01 7.17767417e-01 -3.75901163e-01 1.41879392e+00
1.74584776e-01 8.77945185e-01 -2.64131814e-01 4.64780658e-01
-6.36782944e-02 -2.28876337e-01 8.65792111e-02 8.39119196e-01
4.74832803e-01 4.26896363e-01 3.92789304e-01 4.56538558e-01
4.07970995e-01 1.28105968e-01 -1.84237599e-01 2.50781655e-01
6.52666032e-01 9.37687695e-01 -9.64733601e-01 -4.63265091e-01
-2.12222129e-01 5.29150724e-01 7.10581467e-02 5.25133073e-01
-7.95795143e-01 -1.28713965e-01 3.48708957e-01 6.74990118e-02
3.10690068e-02 -2.70005107e-01 -6.38988793e-01 -1.10588717e+00
-1.42563030e-01 -6.49194419e-01 8.96052599e-01 -6.87145293e-01
-1.68946254e+00 4.67548937e-01 2.68312395e-01 -1.76909542e+00
-3.21138918e-01 -2.50949532e-01 -6.37862265e-01 5.64111769e-01
-1.18090630e+00 -7.41571069e-01 -4.31743026e-01 1.28797019e+00
2.54624307e-01 -4.37483370e-01 9.48898673e-01 2.15729326e-01
-7.81961083e-01 2.85453498e-01 6.74048185e-01 2.38325715e-01
8.37904692e-01 -7.78267443e-01 -4.57360268e-01 4.31539208e-01
-4.10524279e-01 5.82245886e-01 3.56243998e-01 -6.99056864e-01
-1.05333424e+00 -1.48007405e+00 5.71544170e-01 -2.36393079e-01
7.49616504e-01 1.07431032e-01 -1.05791414e+00 5.33951223e-01
-4.28135902e-01 7.75943547e-02 1.30901670e+00 -2.32895717e-01
-3.30511630e-01 -1.78164974e-01 -1.18653166e+00 2.22669393e-01
6.52193487e-01 -4.32416946e-01 -8.68405759e-01 7.07141101e-01
5.87015688e-01 1.25690520e-01 -1.25883484e+00 5.20311534e-01
3.36901665e-01 -7.64701426e-01 8.58246684e-01 -8.23130071e-01
1.40431393e-02 -4.24675673e-01 -1.76004842e-01 -1.35396564e+00
-4.46592748e-01 -2.86974877e-01 1.53665632e-01 8.83909941e-01
5.55518530e-02 -6.31317556e-01 7.43783891e-01 8.56944859e-01
-4.39042598e-01 -5.71870148e-01 -1.07908571e+00 -8.50297272e-01
4.31288213e-01 -6.40880764e-01 7.02616811e-01 1.30185354e+00
4.52313125e-01 -1.42054349e-01 -1.85084134e-01 2.34749421e-01
8.73007715e-01 6.66288882e-02 2.51980811e-01 -1.58403683e+00
2.58537829e-01 -2.99555153e-01 -4.16892737e-01 6.77357987e-02
2.30423942e-01 -1.00246692e+00 7.44802952e-02 -1.45357156e+00
1.66046113e-01 -6.60813689e-01 -7.89162874e-01 7.01772571e-01
3.69130797e-03 -2.35659063e-01 -7.77141601e-02 8.00599456e-01
-4.22433704e-01 4.26626593e-01 4.76361364e-01 -6.10387959e-02
-4.60513204e-01 1.44771095e-02 -3.54460269e-01 7.07925379e-01
1.13200271e+00 -8.47516835e-01 -5.07820129e-01 -3.16697210e-01
-3.89458895e-01 6.07091725e-01 2.67120183e-01 -1.22650599e+00
5.16222000e-01 -5.24991930e-01 5.69494843e-01 -8.99524868e-01
7.61361839e-03 -1.04888368e+00 3.55715007e-01 4.06739116e-01
-3.00083369e-01 3.28788728e-01 -1.30394503e-01 5.99000335e-01
-2.72882223e-01 1.17695615e-01 6.51909769e-01 5.95667809e-02
-3.60257864e-01 4.82428074e-01 -7.27775156e-01 -1.26761824e-01
1.32444000e+00 -3.71783882e-01 -1.80247858e-01 -2.75492966e-01
-1.08212173e+00 2.12638289e-01 2.06877723e-01 4.23831224e-01
9.25555348e-01 -1.43768954e+00 -7.96351492e-01 3.48400921e-01
3.94908339e-01 -6.49851933e-02 5.22039771e-01 1.41673195e+00
-3.38720083e-01 1.77281499e-01 -3.23496222e-01 -8.67924213e-01
-9.83149529e-01 9.06274617e-01 1.85489450e-02 1.07900664e-01
-7.48080015e-01 2.11126357e-01 6.69202134e-02 -4.20839280e-01
5.36567926e-01 -2.09726587e-01 -2.74379224e-01 2.27441475e-01
5.76038063e-01 5.78592539e-01 2.75823593e-01 -7.22074151e-01
-3.58593404e-01 3.18124563e-01 1.85424998e-01 1.65426701e-01
1.72762811e+00 -1.25204802e-01 -4.54038322e-01 7.23469675e-01
1.47596633e+00 -8.35946262e-01 -9.69309390e-01 -2.09774569e-01
2.83648223e-01 -2.84658559e-02 -1.67568892e-01 -6.30681038e-01
-1.18553960e+00 8.62649500e-01 9.13805425e-01 -1.60711005e-01
1.33789885e+00 1.37972990e-02 7.34173656e-01 3.53936851e-01
6.06663227e-01 -1.04611945e+00 -2.93431934e-02 4.03204143e-01
7.24932134e-01 -1.01721966e+00 -2.66221792e-01 8.22731573e-03
-7.18178272e-01 1.12419057e+00 3.26846600e-01 -1.14533879e-01
9.92156863e-01 1.76512241e-01 3.57781172e-01 -1.59216449e-01
-8.48817408e-01 3.54387388e-02 4.81408928e-03 9.62871253e-01
2.46188283e-01 4.61143523e-01 -2.40560591e-01 8.62323403e-01
-1.82016760e-01 1.60729751e-01 3.84643734e-01 9.62891459e-01
-3.44443023e-01 -8.27685654e-01 -7.46467650e-01 8.48944366e-01
-2.21140131e-01 7.48337731e-02 -1.00336753e-01 3.21962148e-01
1.36830658e-01 1.19241536e+00 1.71631295e-02 -6.25026584e-01
1.81779921e-01 3.79942566e-01 -7.85798486e-03 -3.55472505e-01
-4.01309967e-01 3.62534791e-01 -4.88452911e-01 -3.62499475e-01
-5.97900569e-01 -8.32828104e-01 -1.79429936e+00 -3.96895073e-02
-8.63668025e-02 4.06208038e-01 5.92319310e-01 1.01922679e+00
5.34421504e-01 4.31712657e-01 1.00749838e+00 -6.22169316e-01
-2.61523098e-01 -9.31964219e-01 -6.72378004e-01 7.40272284e-01
2.87237793e-01 -7.62444794e-01 -5.69124579e-01 5.68217278e-01] | [7.887799263000488, 5.897399425506592] |
47446515-194c-447c-8382-483197e339d4 | domain-adaptation-for-visual-applications-a | 1702.05374 | null | http://arxiv.org/abs/1702.05374v2 | http://arxiv.org/pdf/1702.05374v2.pdf | Domain Adaptation for Visual Applications: A Comprehensive Survey | The aim of this paper is to give an overview of domain adaptation and
transfer learning with a specific view on visual applications. After a general
motivation, we first position domain adaptation in the larger transfer learning
problem. Second, we try to address and analyze briefly the state-of-the-art
methods for different types of scenarios, first describing the historical
shallow methods, addressing both the homogeneous and the heterogeneous domain
adaptation methods. Third, we discuss the effect of the success of deep
convolutional architectures which led to new type of domain adaptation methods
that integrate the adaptation within the deep architecture. Fourth, we overview
the methods that go beyond image categorization, such as object detection or
image segmentation, video analyses or learning visual attributes. Finally, we
conclude the paper with a section where we relate domain adaptation to other
machine learning solutions. | ['Gabriela Csurka'] | 2017-02-17 | null | null | null | null | ['image-categorization'] | ['computer-vision'] | [ 2.14477867e-01 6.47574663e-02 -4.41738814e-01 -5.37178457e-01
-4.89018172e-01 -6.61276877e-01 8.20506811e-01 -3.90199050e-02
-4.71838742e-01 7.09249794e-01 -3.95710431e-02 -9.25386325e-02
-8.34249146e-03 -7.33117104e-01 -5.63016295e-01 -6.48693621e-01
1.52276817e-03 5.34244776e-01 3.15699786e-01 -2.46914372e-01
3.56490791e-01 6.69415057e-01 -1.58672774e+00 7.13001788e-01
4.10045147e-01 1.19271493e+00 7.76626989e-02 8.18872929e-01
-3.03960890e-01 8.55405390e-01 -8.14155817e-01 -2.67008364e-01
6.96091428e-02 -3.83814156e-01 -1.18879247e+00 3.27657223e-01
7.39325941e-01 -3.96584302e-01 -2.63136595e-01 9.62250113e-01
6.63723767e-01 3.39658558e-01 1.34029496e+00 -1.46881843e+00
-9.45333302e-01 2.18383253e-01 -3.57329100e-01 3.87945175e-01
1.56715587e-01 -1.92684475e-02 5.69052577e-01 -1.02700377e+00
9.32647526e-01 1.06642294e+00 7.85802066e-01 1.05976331e+00
-9.94036138e-01 -5.09979546e-01 4.51436222e-01 8.94985080e-01
-9.90935087e-01 -1.48809090e-01 8.54633927e-01 -9.36206758e-01
1.14741921e+00 -3.22572678e-01 3.68633717e-01 1.42644548e+00
-1.91013306e-01 1.09866822e+00 1.00356543e+00 -7.17306197e-01
2.92444289e-01 3.80090922e-01 2.77812660e-01 2.95085609e-01
-2.51908675e-02 1.59395456e-01 -1.37293220e-01 1.88438684e-01
7.20689178e-01 -4.33948249e-01 2.83468395e-01 -9.22586560e-01
-1.01150715e+00 1.12836897e+00 6.04915142e-01 4.90905672e-01
-1.22208670e-01 -3.01042616e-01 9.30587649e-01 2.79283494e-01
3.33235770e-01 3.27786833e-01 -7.37450898e-01 2.89666861e-01
-7.56010532e-01 1.00649223e-01 4.92315650e-01 1.19543099e+00
8.04204941e-01 3.16488922e-01 -2.61092097e-01 1.11936808e+00
-1.51930545e-02 2.57504314e-01 5.51907122e-01 -8.32439005e-01
1.50407314e-01 2.99915254e-01 -2.41441593e-01 -4.83718783e-01
-5.99734664e-01 -8.10894594e-02 -7.94341683e-01 7.00290680e-01
5.83058894e-01 -2.64931023e-01 -1.19354999e+00 1.34603786e+00
-1.64159779e-02 6.77211359e-02 1.03963703e-01 7.20412731e-01
1.49632406e+00 4.70997244e-01 6.87746465e-01 1.42486438e-01
1.36495519e+00 -1.21108639e+00 -4.76126969e-01 -3.28416586e-01
3.03281665e-01 -7.15450644e-01 8.17609847e-01 2.52776772e-01
-8.86431158e-01 -1.17462778e+00 -9.53824580e-01 -2.99507201e-01
-9.84995246e-01 4.39060688e-01 5.09755313e-01 5.68269253e-01
-9.24900711e-01 5.15073180e-01 -4.78033721e-01 -1.12291694e+00
7.10669875e-01 3.63145947e-01 -4.53287899e-01 1.99771419e-01
-1.18618715e+00 1.09469056e+00 6.41625345e-01 -5.91534019e-01
-1.00321949e+00 -6.84671760e-01 -8.64884853e-01 -2.87594765e-01
7.31429383e-02 -8.92557144e-01 1.61516547e+00 -1.48950219e+00
-1.49102783e+00 1.64693749e+00 7.66952708e-02 -5.58797538e-01
4.20778930e-01 -1.02747314e-01 -4.92159337e-01 2.74792403e-01
-9.23113078e-02 1.03459013e+00 1.08202076e+00 -1.33861411e+00
-1.14215124e+00 -2.15074360e-01 2.05026224e-01 9.16421190e-02
-3.28094155e-01 1.02520712e-01 -1.72821939e-01 -7.67110944e-01
-4.78336215e-01 -6.88310206e-01 2.62805633e-02 4.56182286e-02
6.40067682e-02 -5.19060671e-01 1.13164854e+00 -4.79649365e-01
9.53985512e-01 -2.30002785e+00 3.60406905e-01 -2.20276743e-01
-7.49793882e-03 6.15831077e-01 -2.62071699e-01 4.50368077e-01
-6.64314210e-01 -1.82130784e-01 -3.56273890e-01 -1.84218183e-01
-9.84434485e-02 2.62764156e-01 -3.36471349e-01 3.34029704e-01
3.71267408e-01 1.01537097e+00 -8.37846696e-01 -5.69403112e-01
5.78694046e-01 3.07745934e-01 -2.23815963e-01 3.59183252e-01
-3.25705297e-02 6.70897007e-01 -2.38651350e-01 8.24074090e-01
7.85311997e-01 -9.96061116e-02 -5.42197414e-02 -3.35497022e-01
-1.70357004e-01 -3.14764082e-02 -9.57111359e-01 1.66259706e+00
-4.16974962e-01 1.04061210e+00 2.14566424e-01 -1.75556540e+00
8.40092599e-01 2.20441744e-01 5.66871703e-01 -6.94975555e-01
8.20649322e-03 2.15432763e-01 -1.64689511e-01 -5.58054090e-01
3.06047320e-01 -1.85159594e-01 -5.98480478e-02 1.43926144e-01
7.69072413e-01 -1.04155712e-01 1.73414096e-01 -1.72061712e-01
5.98295093e-01 6.07787311e-01 9.19014573e-01 -5.46865575e-02
1.04502189e+00 3.55752140e-01 2.49895766e-01 6.14698172e-01
-7.29485810e-01 7.60995567e-01 3.44415575e-01 -8.74114215e-01
-1.26719582e+00 -1.13808966e+00 -1.84966132e-01 1.80121934e+00
-1.12257719e-01 1.25632748e-01 -9.16458428e-01 -1.20447803e+00
1.12226196e-01 4.84320551e-01 -1.23473430e+00 -1.42494008e-01
-6.21076226e-01 -7.16232121e-01 4.69219983e-01 1.14418268e+00
8.75594616e-01 -1.49095047e+00 -5.16771972e-01 -1.31850585e-01
-1.77780762e-01 -1.00965118e+00 2.25420520e-01 4.17829365e-01
-1.09016800e+00 -1.04179859e+00 -1.10238838e+00 -1.29495585e+00
3.23471338e-01 -1.02301359e-01 1.50178981e+00 -3.00948530e-01
-2.07124725e-01 7.78532147e-01 -5.38910329e-01 -6.74176276e-01
-4.36605662e-01 4.00635034e-01 -3.03463638e-01 -1.64629906e-01
1.00212729e+00 -3.78221393e-01 -5.19419193e-01 2.47129098e-01
-7.27607012e-01 -4.70564216e-01 5.99223197e-01 8.59967589e-01
4.31696802e-01 -6.06528878e-01 5.28976321e-01 -1.02117193e+00
3.76683444e-01 -4.67626184e-01 -3.08593243e-01 3.19854528e-01
-3.57420921e-01 -3.42178851e-01 2.85427064e-01 -5.85008442e-01
-1.16493344e+00 4.90836114e-01 -1.16226852e-01 -3.19616884e-01
-9.58078206e-01 -6.59711733e-02 -9.98754427e-02 -1.56421840e-01
1.10844243e+00 1.45324633e-01 -4.24340032e-02 -5.19921899e-01
6.16889060e-01 7.39213884e-01 6.33456945e-01 -4.13133353e-01
5.50203025e-01 5.98759413e-01 -1.00279801e-01 -1.12516296e+00
-6.72142744e-01 -7.75185168e-01 -1.38944852e+00 -1.90612465e-01
1.18971539e+00 -7.49347448e-01 -1.93658888e-01 7.08002448e-01
-1.16534722e+00 -7.30667889e-01 -4.83757168e-01 2.25098550e-01
-1.21700156e+00 3.54480356e-01 -6.11483812e-01 -3.28434706e-01
-2.30703857e-02 -8.88235152e-01 9.22973335e-01 1.86586559e-01
-2.11315602e-01 -1.51605511e+00 2.10188180e-01 1.09769084e-01
4.20356929e-01 2.54374176e-01 8.93334508e-01 -1.06627178e+00
-9.07888263e-02 1.54763848e-01 -4.11176860e-01 5.95481992e-01
1.46112040e-01 -9.95559916e-02 -1.43463242e+00 -1.74861610e-01
-4.02182370e-01 -3.10266197e-01 9.30430949e-01 6.64477468e-01
1.27019060e+00 2.57333994e-01 -6.61780238e-01 8.59001458e-01
1.27127314e+00 3.25510561e-01 8.16925704e-01 9.01457787e-01
5.13757288e-01 8.74594748e-01 8.57221365e-01 1.56858787e-01
-1.13031582e-03 7.28940547e-01 3.02643239e-01 -4.56676632e-01
-7.73018181e-01 1.04724333e-01 2.63434127e-02 3.00516456e-01
-4.19947386e-01 -2.43236683e-02 -1.00153470e+00 7.27606058e-01
-1.92211616e+00 -8.17207575e-01 8.68700072e-02 1.82281852e+00
4.99563068e-01 -2.80415118e-01 7.14422464e-01 -2.16603242e-02
1.01016128e+00 -2.93593943e-01 -6.31600857e-01 -7.16877401e-01
-1.87312186e-01 2.64532834e-01 4.24953401e-01 1.48924828e-01
-1.93276846e+00 1.22556913e+00 7.83700705e+00 8.01431298e-01
-1.04513907e+00 1.28973246e-01 4.29032117e-01 4.03269112e-01
6.19061470e-01 -4.81027842e-01 -6.90006375e-01 1.27360404e-01
8.21841776e-01 1.09730735e-01 -5.32123493e-03 1.31784678e+00
-5.00882030e-01 1.06658317e-01 -1.32418084e+00 1.03749633e+00
2.68964976e-01 -1.05908620e+00 1.91954255e-01 -2.11582065e-01
7.90735424e-01 4.22052033e-02 3.25762033e-01 7.40439832e-01
8.66981894e-02 -9.31040823e-01 4.56249774e-01 2.01977924e-01
9.21579719e-01 -6.54373169e-01 7.53122985e-01 -8.20552781e-02
-1.04283869e+00 -4.21928227e-01 -5.90822875e-01 -1.62100829e-02
-1.75041080e-01 1.26492409e-02 -7.33219087e-01 4.34687376e-01
1.08353686e+00 1.05305171e+00 -7.11702526e-01 1.15803444e+00
-2.90708631e-01 3.92208755e-01 2.68571079e-01 2.16730192e-01
4.02461410e-01 -3.97540955e-03 4.40860450e-01 1.94858682e+00
-1.85941383e-02 -5.39367199e-02 8.85774419e-02 5.16642928e-01
2.00916082e-01 1.20941140e-01 -8.50763440e-01 2.18248233e-01
9.80174392e-02 1.35127151e+00 -7.01481640e-01 -7.49874771e-01
-8.27948093e-01 1.09102595e+00 2.93493658e-01 6.00018919e-01
-6.86381400e-01 -6.41300797e-01 8.29174757e-01 8.87604505e-02
6.64235175e-01 -2.95894369e-02 -6.22020483e-01 -9.56256151e-01
-5.11139810e-01 -9.20114994e-01 9.82632875e-01 -7.27027297e-01
-1.65395689e+00 7.03379452e-01 4.47504997e-01 -1.52055597e+00
-6.08993053e-01 -1.32080686e+00 -3.41592878e-01 5.63728213e-01
-1.74047089e+00 -1.42111242e+00 -5.09565473e-01 6.96324050e-01
8.33692670e-01 -7.52743185e-01 9.97669160e-01 3.53509098e-01
-1.57592162e-01 5.64237058e-01 1.32809222e-01 5.61177075e-01
1.32161009e+00 -1.51927674e+00 4.77146775e-01 3.78247887e-01
-5.81464954e-02 2.31026262e-01 5.43038011e-01 -2.85384238e-01
-3.85727733e-01 -1.11719394e+00 5.27914524e-01 -7.52453685e-01
5.19090831e-01 -2.32955307e-01 -1.04329455e+00 8.28269422e-01
6.33943796e-01 1.28862843e-01 7.32473195e-01 -5.01677431e-02
-5.31847239e-01 2.08393042e-03 -1.39244473e+00 2.71684229e-01
8.93669665e-01 -3.78132820e-01 -9.81593966e-01 2.04798266e-01
1.40031919e-01 -3.13353926e-01 -7.40457356e-01 5.36138713e-01
3.94421756e-01 -9.97294784e-01 1.44179761e+00 -1.11486495e+00
4.87213403e-01 -1.56571582e-01 -1.16541155e-01 -1.38689375e+00
-8.51201355e-01 -2.43272334e-02 -1.00913793e-01 1.20594335e+00
1.62775725e-01 -3.61267567e-01 7.79410124e-01 1.28831267e-01
-1.17761403e-01 -2.22815908e-02 -7.43141413e-01 -8.94104004e-01
8.69547844e-01 -1.45840257e-01 2.81889975e-01 1.16675901e+00
-1.85519040e-01 4.57912236e-01 -2.14863196e-02 -2.84155965e-01
5.15649617e-01 6.44575804e-02 7.99379885e-01 -1.43555057e+00
4.14671339e-02 -7.63391018e-01 -6.77270234e-01 -8.22303474e-01
4.19872731e-01 -7.99328387e-01 -4.90907058e-02 -1.66543221e+00
2.69352347e-01 1.05793796e-01 -4.07796562e-01 5.34162581e-01
-4.02554534e-02 5.80577672e-01 1.22836970e-01 8.94014761e-02
-6.10987008e-01 1.30890131e-01 9.32040751e-01 -3.41425002e-01
-7.18739852e-02 2.89613187e-01 -3.58251631e-01 1.00254214e+00
1.03984547e+00 -1.55539349e-01 -2.34361380e-01 -4.25741464e-01
-2.27871805e-01 -6.72532558e-01 3.55533510e-01 -1.10278213e+00
-2.55642347e-02 -3.66600463e-03 8.43712747e-01 -4.56697345e-01
1.89154506e-01 -9.72607315e-01 -6.72182143e-01 3.52761805e-01
-2.92910993e-01 -2.00654879e-01 6.86888635e-01 4.07190293e-01
-4.88624245e-01 -4.94942844e-01 1.37127686e+00 -3.21197838e-01
-1.92623699e+00 2.60555651e-02 -6.45560563e-01 -3.51867452e-02
1.48514736e+00 -5.23316920e-01 -1.52535424e-01 -3.20796967e-01
-1.28931689e+00 5.68791144e-02 3.71780246e-01 7.13645220e-01
4.11299020e-01 -1.35103607e+00 -7.99533308e-01 5.28495535e-02
6.78697348e-01 -5.00776708e-01 3.00858498e-01 4.53537792e-01
-3.46516430e-01 5.97404540e-01 -9.17811871e-01 -8.70433390e-01
-1.46747458e+00 1.19472933e+00 5.54524064e-01 8.87428150e-02
-3.66825759e-01 7.00782239e-01 5.16072273e-01 -5.91252387e-01
3.62765342e-01 -4.96133082e-02 -8.17183673e-01 3.02794158e-01
5.50555468e-01 5.88112831e-01 2.75383741e-01 -6.31890595e-01
-6.90808356e-01 1.03043413e+00 1.04854032e-01 1.70763612e-01
1.12497282e+00 -2.64144301e-01 9.71437097e-02 4.93280321e-01
1.21181130e+00 -7.17740893e-01 -1.42557585e+00 -2.31378332e-01
1.18746638e-01 -1.22471102e-01 -3.14577281e-01 -9.95845199e-01
-9.64933336e-01 1.54274571e+00 9.61467505e-01 1.26522794e-01
1.42083848e+00 2.79569626e-01 2.24181026e-01 3.79951984e-01
-4.21039481e-03 -1.45876312e+00 2.76303709e-01 7.66169846e-01
8.80344689e-01 -1.63151133e+00 1.04211628e-01 -3.59601617e-01
-1.00671029e+00 1.47045875e+00 9.53465939e-01 -3.04398894e-01
7.19705284e-01 6.07592426e-03 3.70673090e-01 1.75263524e-01
-3.58013242e-01 -5.96697330e-01 6.94845915e-01 1.70082927e+00
6.09732568e-01 -1.85898557e-01 -2.24188939e-02 3.23170513e-01
2.64268428e-01 1.88852493e-02 2.01448008e-01 8.18459272e-01
-4.74374682e-01 -1.45569348e+00 -3.62360120e-01 -1.30621180e-01
-4.40274656e-01 2.29492053e-01 -8.78803849e-01 1.21871853e+00
4.15126324e-01 7.70114064e-01 4.38617080e-01 -1.76584110e-01
5.77548265e-01 3.39110494e-01 6.95795059e-01 -8.03530991e-01
-6.09632075e-01 -1.69642463e-01 -1.39438182e-01 -3.11849475e-01
-8.55696440e-01 -6.35646045e-01 -8.39253068e-01 6.78789318e-02
3.75149757e-01 -2.19281062e-01 5.37313163e-01 6.38838291e-01
1.86762929e-01 4.71686810e-01 1.05756655e-01 -1.01274943e+00
-2.15906203e-01 -1.24770176e+00 -3.99123967e-01 5.06355405e-01
5.19273460e-01 -8.16750169e-01 3.61487903e-02 6.29189312e-01] | [9.952799797058105, 2.4618685245513916] |
b48678b3-0d04-43b2-9f04-e6db1e4f61a8 | characterization-of-neighborhood-behaviours | 1603.06459 | null | http://arxiv.org/abs/1603.06459v1 | http://arxiv.org/pdf/1603.06459v1.pdf | Characterization of neighborhood behaviours in a multi-neighborhood local search algorithm | We consider a multi-neighborhood local search algorithm with a large number
of possible neighborhoods. Each neighborhood is accompanied by a weight value
which represents the probability of being chosen at each iteration. These
weights are fixed before the algorithm runs, and are considered as parameters
of the algorithm. Given a set of instances, off-line tuning of the algorithm's
parameters can be done by automated algorithm configuration tools (e.g., SMAC).
However, the large number of neighborhoods can make the tuning expensive and
difficult even when the number of parameters has been reduced by some
intuition. In this work, we propose a systematic method to characterize each
neighborhood's behaviours, representing them as a feature vector, and using
cluster analysis to form similar groups of neighborhoods. The novelty of our
characterization method is the ability of reflecting changes of behaviours
according to hardness of different solution quality regions. We show that using
neighborhood clusters instead of individual neighborhoods helps to reduce the
parameter configuration space without misleading the search of the tuning
procedure. Moreover, this method is problem-independent and potentially can be
applied in similar contexts. | ['Nguyen Thi Thanh Dang', 'Patrick De Causmaecker'] | 2016-03-12 | null | null | null | null | ['smac-1', 'smac'] | ['playing-games', 'playing-games'] | [ 1.53119579e-01 -2.87935287e-01 -2.98492640e-01 -8.90404284e-02
-5.04345715e-01 -9.82129633e-01 3.41195583e-01 5.44437349e-01
-3.19826156e-01 6.07691109e-01 -8.37348551e-02 -4.24443007e-01
-7.13678002e-01 -1.11646175e+00 -4.65249717e-01 -1.01086974e+00
3.39520499e-02 7.81997561e-01 5.30150115e-01 -3.05099636e-01
6.12245023e-01 6.44477606e-01 -1.86447012e+00 1.70475170e-01
9.22230124e-01 7.02748418e-01 2.39354029e-01 7.32633114e-01
8.04810319e-03 1.27554879e-01 -6.76330447e-01 1.19020931e-01
3.35376710e-01 -6.12401843e-01 -9.08361733e-01 1.26713082e-01
-1.39302254e-01 3.78388196e-01 5.11667073e-01 1.13410509e+00
4.02267933e-01 3.44029188e-01 5.16984761e-01 -1.13456857e+00
-1.34159982e-01 7.37533689e-01 -4.58800197e-01 1.52077422e-01
2.19947055e-01 7.27015585e-02 1.04139066e+00 -2.52441257e-01
6.82671547e-01 9.38144028e-01 5.36428869e-01 2.11316571e-01
-1.68959391e+00 -3.01571012e-01 -3.53894057e-03 2.90767789e-01
-1.75327635e+00 -3.11512768e-01 7.47970581e-01 -3.92701000e-01
6.40124738e-01 5.37280202e-01 6.67474985e-01 3.57004136e-01
-7.07208663e-02 2.47798756e-01 9.72462714e-01 -7.68785119e-01
8.70281458e-01 3.50411713e-01 1.42435759e-01 4.48423028e-01
4.43433940e-01 -1.12521291e-01 -1.11665368e-01 -5.40021241e-01
5.79087555e-01 -3.67728025e-01 -3.45365345e-01 -1.06465173e+00
-9.65981424e-01 9.61902618e-01 2.53577828e-01 5.94455361e-01
-1.46045849e-01 7.40440264e-02 2.27395505e-01 5.28264880e-01
1.99494034e-01 7.85556376e-01 -5.47714174e-01 -2.67253339e-01
-7.92623281e-01 3.90171230e-01 6.81770265e-01 4.64597613e-01
1.12192571e+00 -7.17357218e-01 9.72625837e-02 8.22225094e-01
-1.61875650e-01 -2.91394275e-02 5.55087686e-01 -8.77991021e-01
3.74146134e-01 9.03194129e-01 3.89022022e-01 -1.30231833e+00
-6.24973774e-01 -3.61547321e-01 -3.68635356e-01 3.65821093e-01
7.38517821e-01 8.57997388e-02 -5.36578476e-01 1.70967197e+00
5.77844381e-01 -1.08658709e-01 -1.25724092e-01 6.88821316e-01
-1.41160518e-01 5.33610106e-01 -3.80895644e-01 -4.60481673e-01
9.15886462e-01 -8.97426665e-01 -2.81944752e-01 1.31062835e-01
1.07743645e+00 -5.19512951e-01 1.08322322e+00 4.35362458e-01
-1.07038736e+00 -3.37120503e-01 -9.16361928e-01 5.01492381e-01
-5.05660355e-01 -9.96678025e-02 3.27281147e-01 9.27199781e-01
-1.24590206e+00 9.08044696e-01 -9.85339463e-01 -4.84769732e-01
-1.34465203e-01 7.07860827e-01 7.47983083e-02 1.54488444e-01
-7.47829676e-01 8.61610711e-01 6.90706432e-01 -7.34163672e-02
-2.30174780e-01 -3.37149292e-01 -5.20322204e-01 2.21575484e-01
7.01948345e-01 -5.53354144e-01 7.95367360e-01 -9.67722714e-01
-1.31598544e+00 4.97648358e-01 -2.20246300e-01 -1.13723345e-01
5.01363873e-01 4.53300595e-01 -2.80371606e-01 -2.84374785e-03
3.86973610e-03 2.61698872e-01 7.69177318e-01 -1.19831991e+00
-7.58260012e-01 -3.01207751e-01 2.09952340e-01 2.11141810e-01
-4.79067087e-01 -6.62003905e-02 -5.11411786e-01 -2.29026929e-01
2.25148916e-01 -1.07080877e+00 -7.22617328e-01 -4.14651394e-01
-3.08071166e-01 -2.74218976e-01 4.52063471e-01 1.28181562e-01
1.69609964e+00 -2.14513397e+00 4.61142629e-01 1.02520168e+00
3.69569436e-02 -3.47130001e-02 -2.09470212e-01 5.75186729e-01
-1.06743470e-01 4.36010212e-01 -1.96537435e-01 7.03190938e-02
-6.38463050e-02 2.13191405e-01 2.25044802e-01 6.75631642e-01
-1.55141652e-01 5.16653061e-01 -9.03649032e-01 -4.60475653e-01
2.28547156e-01 1.16806537e-01 -7.76567221e-01 -2.02348769e-01
-1.51785761e-01 1.38885409e-01 -7.33257532e-01 2.65619874e-01
5.45221984e-01 -3.34092438e-01 4.35719192e-01 9.67623740e-02
-3.36927891e-01 1.95049509e-01 -1.85534644e+00 1.14927781e+00
-5.02063274e-01 7.50459805e-02 2.79670656e-02 -1.23171818e+00
9.15955424e-01 -1.41206294e-01 3.62306625e-01 -2.22123802e-01
2.26298109e-01 3.27829242e-01 5.15476875e-02 -2.65573800e-01
4.77851242e-01 1.01243474e-01 -1.11253001e-01 8.24592650e-01
-4.32499439e-01 7.69821256e-02 6.30265653e-01 -3.48543167e-01
1.24598503e+00 -4.10949707e-01 6.12356544e-01 -5.80594778e-01
7.31318951e-01 2.13169426e-01 5.50336242e-01 8.70867848e-01
7.23588094e-02 5.17198741e-01 5.99862695e-01 -4.51929808e-01
-1.13562202e+00 -7.30550826e-01 -3.37397248e-01 9.12161171e-01
2.78481305e-01 -6.52970552e-01 -7.20655978e-01 -5.18859088e-01
-1.09395109e-01 4.46040392e-01 -8.22679579e-01 -2.86053419e-01
-7.38314331e-01 -7.68497050e-01 -4.10790034e-02 3.06329459e-01
-3.27173099e-02 -1.03016722e+00 -1.01631176e+00 2.74406284e-01
1.95843983e-03 -4.74782914e-01 -3.52583438e-01 4.20267433e-01
-9.39339399e-01 -1.14914370e+00 -2.41617978e-01 -6.88714266e-01
9.07717228e-01 2.04385519e-01 9.74896967e-01 4.26141500e-01
-2.86034912e-01 1.96096033e-01 -6.30978823e-01 2.25026950e-01
-5.79661071e-01 3.52602601e-01 2.96292547e-02 4.70296200e-03
7.17725679e-02 -6.34824276e-01 -4.68764186e-01 6.91826046e-01
-8.60470355e-01 -3.98587495e-01 3.90604109e-01 8.73781383e-01
7.98927188e-01 8.72264683e-01 2.25258306e-01 -8.37258399e-01
9.05957222e-01 -4.69389975e-01 -9.02959704e-01 5.38769066e-01
-7.44149387e-01 4.35485691e-01 8.70147645e-01 -6.09589934e-01
-4.66239542e-01 2.61012852e-01 4.37295288e-01 -1.19247690e-01
-1.45916075e-01 5.43260574e-01 -1.41569689e-01 -3.17084312e-01
7.16328323e-01 1.37785831e-02 1.51700350e-02 -3.80858064e-01
4.80661124e-01 3.09951603e-01 2.47500226e-01 -6.16973162e-01
8.41338813e-01 1.55010492e-01 8.38162079e-02 -4.96574223e-01
-3.29959124e-01 -5.73442221e-01 -6.29065394e-01 -8.51916411e-05
4.07436460e-01 -1.19540617e-01 -8.01487327e-01 3.45678860e-03
-6.11316144e-01 -2.42030099e-01 -3.13240319e-01 6.21408643e-03
-5.78550160e-01 3.61277848e-01 4.15555276e-02 -7.81950653e-01
2.05011129e-01 -1.47742009e+00 5.46353638e-01 1.75294086e-01
-5.54267645e-01 -1.14606476e+00 3.13326001e-01 1.71394140e-01
3.63239229e-01 2.54169643e-01 1.07167637e+00 -7.43761957e-01
-3.83202702e-01 -1.47549301e-01 2.75822610e-01 -5.48988357e-02
3.08835536e-01 1.04131177e-01 -4.24995780e-01 -5.28901935e-01
1.65210694e-01 6.55133575e-02 4.50023562e-01 5.06189883e-01
1.37301540e+00 -4.08799589e-01 -6.07225120e-01 5.08485675e-01
1.64393997e+00 2.96063602e-01 3.65749836e-01 9.39395905e-01
2.53228486e-01 6.34016812e-01 5.92995644e-01 4.86739427e-01
-1.36605576e-02 1.02440083e+00 3.28590959e-01 6.41898438e-02
3.67881119e-01 2.21322447e-01 3.25293243e-02 2.79689312e-01
-6.43997416e-02 -2.05792740e-01 -1.06561315e+00 6.00309312e-01
-2.05658841e+00 -8.75924289e-01 6.30121976e-02 2.56605291e+00
7.64779687e-01 2.52931267e-01 6.23704076e-01 5.11609972e-01
1.05312133e+00 -1.42133012e-01 -3.96019131e-01 -6.60486341e-01
5.90282343e-02 1.54795989e-01 5.79668760e-01 5.94303131e-01
-7.09242940e-01 6.25813663e-01 6.37057638e+00 9.78120267e-01
-9.45923805e-01 -3.36497158e-01 5.81214249e-01 -1.32399350e-01
-3.06896746e-01 1.91359222e-01 -6.10967934e-01 5.02225697e-01
8.81013632e-01 -3.25854152e-01 7.86349237e-01 8.91029119e-01
1.83755144e-01 -3.72404397e-01 -1.15232074e+00 7.08427966e-01
-1.86523825e-01 -1.36485445e+00 -7.77399018e-02 1.86327025e-01
8.46791148e-01 -3.75028282e-01 9.30344760e-02 -2.61005759e-01
2.26404235e-01 -9.86356735e-01 5.40421724e-01 1.75467245e-02
3.29604566e-01 -1.39443088e+00 5.79619110e-01 4.59230185e-01
-1.26014602e+00 -6.45796239e-01 -4.05010551e-01 -4.42258529e-02
-1.60851389e-01 5.16857266e-01 -7.52063811e-01 4.30247515e-01
8.19342196e-01 1.67173639e-01 -7.02234268e-01 1.33034384e+00
3.25254470e-01 4.38327193e-01 -6.27567530e-01 -3.00723553e-01
2.37058520e-01 -6.26706779e-01 5.42445302e-01 7.74664462e-01
3.34309906e-01 -1.80747628e-03 3.06250304e-01 8.90945315e-01
3.63861442e-01 3.80032778e-01 -3.51261914e-01 1.12092249e-01
8.25404644e-01 1.14617610e+00 -1.29795039e+00 -7.55689815e-02
-4.65249456e-02 5.52763104e-01 4.04821098e-01 1.79379135e-01
-6.29327893e-01 -4.96165633e-01 4.63653088e-01 1.71182960e-01
5.73894083e-01 -5.66234328e-02 -5.97476482e-01 -7.21145689e-01
1.70852974e-01 -8.38801622e-01 4.47991282e-01 -2.70696461e-01
-8.97070825e-01 5.93432546e-01 7.81635568e-02 -1.22224712e+00
-3.13052565e-01 -3.23318690e-01 -7.30275214e-01 5.68980098e-01
-1.01343524e+00 -5.06419241e-01 -1.85915194e-02 5.87552428e-01
1.26161456e-01 1.28449321e-01 6.94457471e-01 -5.16994260e-02
-6.40915811e-01 7.48713195e-01 4.89678860e-01 -4.15933341e-01
4.17178929e-01 -1.22025609e+00 -3.97361480e-02 7.25050151e-01
-8.83558244e-02 7.13444769e-01 1.05092466e+00 -2.83370852e-01
-1.07881570e+00 -7.49498904e-01 8.23573887e-01 -4.60699201e-01
9.04253185e-01 -2.76938081e-01 -8.76205683e-01 1.97619960e-01
-2.75497586e-01 -1.86491057e-01 5.86210430e-01 4.91399407e-01
3.36647481e-02 -1.88888520e-01 -1.21251965e+00 7.78508008e-01
8.57333660e-01 -1.22297481e-01 -2.54721671e-01 2.46432036e-01
4.00949985e-01 -2.16233149e-01 -8.65514040e-01 2.45637164e-01
1.59315720e-01 -1.12535036e+00 9.78945792e-01 -3.67371082e-01
1.77221727e-02 -6.02580667e-01 5.14029600e-02 -1.31354034e+00
-6.24263704e-01 -6.83632076e-01 1.65032357e-01 1.26261461e+00
6.36180580e-01 -6.62924170e-01 8.38311613e-01 4.67762768e-01
1.88028917e-01 -1.02992606e+00 -9.75315154e-01 -9.26164210e-01
6.71850741e-02 -2.54569769e-01 1.01870346e+00 8.02345991e-01
3.06824505e-01 3.97767033e-03 -1.68435685e-02 2.62757331e-01
3.20004702e-01 5.20497620e-01 6.31979764e-01 -1.27173352e+00
-5.33255994e-01 -9.12514210e-01 -3.55936289e-01 -6.27472758e-01
-2.12144300e-01 -6.52545035e-01 -8.43879655e-02 -9.96437848e-01
1.87490761e-01 -9.56146359e-01 -2.99359828e-01 5.17901063e-01
-5.98067529e-02 -1.45764742e-02 -5.70884161e-02 2.60503531e-01
-6.44743085e-01 1.71187907e-01 9.31783140e-01 1.29364178e-01
-8.15574169e-01 2.37395301e-01 -6.42652452e-01 4.74231422e-01
9.88012850e-01 -5.86599469e-01 -4.63129669e-01 8.01394135e-02
6.59385204e-01 2.37815678e-02 3.45709547e-02 -9.57146406e-01
2.53108650e-01 -3.58714968e-01 4.67614196e-02 -2.75834173e-01
-6.97311759e-02 -8.82231534e-01 4.01833236e-01 3.98327768e-01
-5.12798548e-01 2.83657193e-01 -7.26028830e-02 5.62431037e-01
1.68160237e-02 -5.35252392e-01 8.75549376e-01 9.50124934e-02
-3.42172563e-01 -2.05096245e-01 -4.04937685e-01 -2.58469164e-01
1.16582894e+00 -6.62953556e-01 4.17277776e-02 -1.66043147e-01
-7.67166615e-01 3.74064893e-01 1.11217737e+00 1.08708471e-01
1.36816904e-01 -1.24001980e+00 -2.29010642e-01 1.12137042e-01
2.37081409e-01 1.56538319e-02 3.68062295e-02 7.51520455e-01
-3.84932071e-01 1.51633039e-01 1.85572391e-03 -6.47663772e-01
-1.43085110e+00 8.87716591e-01 3.00779521e-01 -3.65058899e-01
-5.87613642e-01 5.09888589e-01 -2.25402758e-01 -3.25973719e-01
1.59385160e-01 -3.50870341e-01 -2.98353046e-01 -4.96712178e-02
4.13101405e-01 5.64874828e-01 1.56230137e-01 -3.29112500e-01
-4.82901573e-01 1.01127136e+00 7.14771301e-02 3.65476459e-02
1.38661432e+00 -3.60204130e-01 -3.42416376e-01 3.77460212e-01
1.10625446e+00 6.50570914e-02 -8.17453504e-01 9.56769958e-02
1.89483702e-01 -5.28247416e-01 -1.25164926e-01 -5.68899214e-01
-1.03281868e+00 1.97766483e-01 4.61924314e-01 5.08721888e-01
1.46201777e+00 9.43931192e-02 3.62544447e-01 5.45761526e-01
4.02858824e-01 -1.34614217e+00 -1.72876835e-01 2.86501527e-01
5.26609540e-01 -9.40520942e-01 -1.17563047e-01 -2.94540972e-01
-3.16799492e-01 1.26419914e+00 3.03565353e-01 -8.11364874e-02
5.64045370e-01 1.86515212e-01 -2.91561969e-02 -1.07032344e-01
-6.63417757e-01 -1.10358894e-02 2.32817739e-01 3.51226628e-01
-5.13141826e-02 7.36979023e-03 -6.62108481e-01 3.22845727e-01
-4.79738027e-01 -3.86257440e-01 6.27085030e-01 1.02521122e+00
-6.79314733e-01 -1.42884195e+00 -6.42566502e-01 2.12821096e-01
2.86655407e-02 1.86634049e-01 -3.47572297e-01 7.64619410e-01
2.30639145e-01 1.17285490e+00 -3.27228829e-02 -4.01685923e-01
1.75071269e-01 -1.16352171e-01 3.13568950e-01 -5.00411034e-01
-6.02377713e-01 8.32705945e-03 -2.54535917e-02 -6.41179323e-01
-2.23052591e-01 -7.85863042e-01 -1.16993022e+00 -3.31514955e-01
-5.91585338e-01 6.53268814e-01 4.57076490e-01 9.32216823e-01
3.80246162e-01 1.37561932e-01 8.81739795e-01 -5.65918505e-01
-5.71551740e-01 -2.89161891e-01 -4.86519456e-01 3.17749798e-01
1.65254608e-01 -8.30942094e-01 -5.86151242e-01 -2.47002289e-01] | [7.15424108505249, 4.378809452056885] |
7805e6cc-f321-4f34-be65-cc77ae89c836 | brainclip-bridging-brain-and-visual | 2302.12971 | null | https://arxiv.org/abs/2302.12971v3 | https://arxiv.org/pdf/2302.12971v3.pdf | BrainCLIP: Bridging Brain and Visual-Linguistic Representation Via CLIP for Generic Natural Visual Stimulus Decoding | Due to the lack of paired samples and the low signal-to-noise ratio of functional MRI (fMRI) signals, reconstructing perceived natural images or decoding their semantic contents from fMRI data are challenging tasks. In this work, we propose, for the first time, a task-agnostic fMRI-based brain decoding model, BrainCLIP, which leverages CLIP's cross-modal generalization ability to bridge the modality gap between brain activity, image, and text. Our experiments demonstrate that CLIP can act as a pivot for generic brain decoding tasks, including zero-shot visual categories decoding, fMRI-image/text matching, and fMRI-to-image generation. Specifically, BrainCLIP aims to train a mapping network that transforms fMRI patterns into a well-aligned CLIP embedding space by combining visual and textual supervision. Our experiments show that this combination can boost the decoding model's performance on certain tasks like fMRI-text matching and fMRI-to-image generation. On the zero-shot visual category decoding task, BrainCLIP achieves significantly better performance than BraVL, a recently proposed multi-modal method specifically designed for this task. BrainCLIP can also reconstruct visual stimuli with high semantic fidelity and establishes a new state-of-the-art for fMRI-based natural image reconstruction in terms of high-level semantic features. | ['Nanning Zheng', 'Guibo Zhu', 'Wei Zhou', 'Yongqiang Ma', 'Yulong Liu'] | 2023-02-25 | null | null | null | null | ['image-reconstruction', 'brain-decoding', 'brain-decoding', 'text-matching'] | ['computer-vision', 'medical', 'miscellaneous', 'natural-language-processing'] | [ 5.59907913e-01 3.84194553e-02 -1.20010629e-01 -6.32790744e-01
-9.46528971e-01 -3.46984833e-01 7.87471294e-01 -2.09451482e-01
-1.62694633e-01 3.88163924e-01 5.41708231e-01 -1.35165183e-02
6.60659298e-02 -4.18880612e-01 -9.93743539e-01 -4.25631344e-01
1.73540398e-01 2.91752130e-01 -5.40343821e-02 4.32105437e-02
2.43451968e-02 9.85644460e-02 -1.62305772e+00 8.43387306e-01
6.78949296e-01 1.21507025e+00 6.40990317e-01 2.80207127e-01
-2.29894996e-01 9.75440264e-01 4.55281660e-02 -2.63350755e-01
1.01440735e-01 -7.12191582e-01 -8.92345667e-01 -9.90078747e-02
7.21084714e-01 -2.84326673e-01 -6.23380005e-01 1.06625962e+00
6.59945190e-01 1.28353819e-01 8.71386528e-01 -1.15063155e+00
-9.86743927e-01 7.63882816e-01 -5.39306164e-01 3.06094021e-01
3.79396886e-01 4.92853284e-01 8.75801206e-01 -1.23935676e+00
8.50985944e-01 1.37016416e+00 4.86538559e-01 8.36723924e-01
-1.61956310e+00 -8.19963813e-01 -1.21877566e-01 3.26657891e-01
-1.18359971e+00 -7.16083229e-01 6.49411142e-01 -9.33080614e-01
7.99793363e-01 1.59956679e-01 5.88611603e-01 1.73508251e+00
2.05806777e-01 7.70689726e-01 1.55714297e+00 -1.10183448e-01
6.80719987e-02 -1.84483409e-01 5.30370250e-02 7.50563502e-01
-3.27238530e-01 1.92614242e-01 -8.71540308e-01 2.69764900e-01
8.08631420e-01 1.74543243e-02 -4.84116226e-01 -1.16113484e-01
-1.82797492e+00 7.78439999e-01 8.92932534e-01 5.09975255e-01
-3.32187444e-01 3.01028490e-01 4.70283717e-01 9.46921334e-02
4.49620247e-01 4.57751274e-01 1.02363646e-01 2.50672519e-01
-1.20540774e+00 5.12613207e-02 3.02617103e-01 7.29040205e-01
5.53950071e-01 1.41615257e-01 -8.43306422e-01 9.97441769e-01
-1.42148472e-02 5.02440035e-01 7.81767309e-01 -7.10625887e-01
3.28640074e-01 8.43211487e-02 -3.98331821e-01 -9.61863220e-01
-3.90499115e-01 -3.24905962e-01 -1.07281363e+00 -2.07974374e-01
5.93340136e-02 1.79058626e-01 -1.07252336e+00 2.10574555e+00
-1.01728559e-01 5.03035367e-01 -2.96261162e-01 1.29520917e+00
1.18687129e+00 5.62847495e-01 2.97974646e-01 -1.32335171e-01
1.58147418e+00 -1.04277563e+00 -7.18406379e-01 -6.75720990e-01
-7.41071440e-03 -2.04086885e-01 1.36472118e+00 -1.24749467e-02
-1.07350588e+00 -7.98799455e-01 -8.43815267e-01 -2.30391860e-01
-1.71161339e-01 -1.08228721e-01 5.53001285e-01 1.03450455e-01
-1.01219487e+00 3.07209164e-01 -7.51970351e-01 -1.77848846e-01
9.58583117e-01 -1.03555456e-01 -7.62549460e-01 -3.22874546e-01
-1.18381226e+00 7.42552340e-01 4.78946388e-01 2.09057471e-03
-1.32010961e+00 -1.20720851e+00 -1.02691531e+00 1.92510247e-01
2.44313583e-01 -9.26457882e-01 7.93927789e-01 -1.13017488e+00
-1.12831748e+00 1.24483836e+00 -1.66295633e-01 -3.98156971e-01
3.91523093e-01 3.33241969e-01 -4.49002475e-01 5.99546850e-01
4.12808627e-01 1.21571648e+00 1.18081355e+00 -1.22277808e+00
1.17720522e-01 -5.22213459e-01 -5.07272542e-01 -1.07625820e-01
-3.66182297e-01 -1.53629914e-01 -2.81839401e-01 -9.13523316e-01
-6.25260398e-02 -4.88062531e-01 1.32142141e-01 2.97767252e-01
-3.92977327e-01 1.07769154e-01 2.38410011e-01 -8.47745776e-01
7.97079206e-01 -2.21745944e+00 4.96054053e-01 -1.06518596e-01
5.78067482e-01 -1.69015735e-01 -6.04798138e-01 1.67792693e-01
-4.33955818e-01 -1.73555817e-02 -6.85807049e-01 -3.31072181e-01
5.49924150e-02 -7.97079131e-02 -3.97105306e-01 5.83355904e-01
3.76265973e-01 1.58833826e+00 -1.09047747e+00 -4.49915737e-01
1.68254048e-01 5.51005185e-01 -6.81155205e-01 4.38025206e-01
-1.59396246e-01 8.70087087e-01 -1.93101680e-03 5.39330840e-01
4.70480561e-01 -3.85481358e-01 1.29429594e-01 -6.45379364e-01
3.03676099e-01 1.55291734e-02 -2.29824364e-01 2.51710129e+00
-6.54082119e-01 7.58605421e-01 5.29509857e-02 -1.46048558e+00
5.44076562e-01 3.05541158e-01 5.58950126e-01 -1.24378896e+00
2.33201385e-01 1.30791470e-01 -1.48430720e-01 -7.26125300e-01
-1.54743746e-01 -6.69715643e-01 -1.21670522e-01 4.54714745e-01
6.93906903e-01 -3.78495380e-02 -1.85174823e-01 2.84996986e-01
1.03590822e+00 -3.85462828e-02 6.98997080e-02 -3.96125495e-01
2.04878345e-01 -3.17540377e-01 1.37363166e-01 8.16223085e-01
-1.08148485e-01 8.02558064e-01 2.19836041e-01 -8.43727291e-02
-9.88786340e-01 -1.38152826e+00 -3.83398861e-01 1.13948095e+00
5.68565242e-02 -1.78240731e-01 -7.96975493e-01 -5.22335470e-01
4.70904149e-02 6.28412426e-01 -9.49728787e-01 -4.69455600e-01
-2.86232978e-01 -5.41750908e-01 5.27317643e-01 4.22742695e-01
2.71510512e-01 -1.29292464e+00 -3.95208657e-01 1.80957898e-01
-7.14223146e-01 -1.46982145e+00 -6.11364543e-01 4.33993004e-02
-6.54631972e-01 -9.60389793e-01 -7.82610893e-01 -9.77683246e-01
6.76869094e-01 3.42697352e-01 1.16191983e+00 -8.27045739e-02
-6.79676354e-01 3.38952214e-01 -2.95122534e-01 2.35992089e-01
-3.32609028e-01 -2.26001889e-01 -2.52169251e-01 3.49794537e-01
-1.07019432e-01 -7.99436271e-01 -8.07335496e-01 9.18115824e-02
-1.21787047e+00 5.73902130e-01 6.10643566e-01 1.27063823e+00
5.46262920e-01 -5.56679904e-01 7.81960666e-01 -7.88854778e-01
5.58639467e-01 -8.59412193e-01 -6.60087261e-03 2.71479398e-01
-1.44597471e-01 2.16919348e-01 6.89002037e-01 -5.59910119e-01
-9.18871462e-01 -1.00808874e-01 -2.43475795e-01 -8.33060086e-01
-1.58330217e-01 5.91810048e-01 -4.47020940e-02 -6.25200421e-02
8.42272520e-01 4.49934095e-01 2.00044349e-01 -4.92677361e-01
6.63069487e-01 6.41689718e-01 1.15377617e+00 -4.11414415e-01
4.20181960e-01 5.92042744e-01 -2.08146363e-01 -5.54832935e-01
-1.09985161e+00 -3.69665772e-01 -6.21080935e-01 -3.84164006e-01
1.23468876e+00 -1.17424250e+00 -6.23926520e-01 3.11289996e-01
-1.03241491e+00 -5.10962844e-01 -1.14638638e-02 2.35037342e-01
-7.64090836e-01 2.74636179e-01 -6.46204472e-01 -1.85018599e-01
-5.49998701e-01 -1.21783066e+00 1.31455815e+00 -3.63423407e-01
8.13820679e-03 -8.11218679e-01 -2.28413895e-01 5.95920861e-01
4.95706260e-01 3.91638041e-01 1.04413128e+00 -2.67787725e-01
-3.20120335e-01 2.32097104e-01 -6.87055588e-01 2.72921860e-01
-1.94555283e-01 -9.11126375e-01 -1.21560264e+00 -4.32761252e-01
1.28096715e-01 -7.84701288e-01 1.32756710e+00 3.04024726e-01
1.77839863e+00 -2.64861763e-01 -2.01491281e-01 9.65998530e-01
1.45046031e+00 -3.33638877e-01 6.68040991e-01 -2.77195334e-01
9.47384298e-01 6.44064248e-01 1.03916608e-01 2.50389278e-01
4.06313032e-01 7.65796542e-01 3.74635667e-01 -2.00375542e-01
-6.87453151e-01 -4.59677368e-01 4.51183915e-01 9.91772294e-01
5.58366716e-01 4.54984792e-02 -8.58146667e-01 6.58528268e-01
-1.66062367e+00 -1.16835773e+00 8.04716423e-02 1.87349248e+00
1.05384648e+00 -2.66433120e-01 -8.29368755e-02 -3.85963589e-01
6.06083333e-01 3.47187996e-01 -6.49385154e-01 6.84510618e-02
-1.63682356e-01 5.08196175e-01 1.14485011e-01 2.66066432e-01
-1.02887666e+00 9.19513166e-01 5.81409407e+00 1.00232363e+00
-1.31602526e+00 9.10483658e-01 7.07432866e-01 -1.72591567e-01
-4.98994112e-01 -2.59647250e-01 -1.00977063e-01 6.97195411e-01
9.61184800e-01 -9.95282922e-03 9.19101298e-01 3.92412603e-01
-3.60717997e-02 2.07935154e-01 -1.34757662e+00 1.45359063e+00
5.77072740e-01 -1.57174194e+00 2.45153263e-01 -2.72252142e-01
6.32155180e-01 1.96090192e-01 5.61639778e-02 3.58997792e-01
-2.44300023e-01 -1.36510813e+00 1.16654265e+00 8.10989261e-01
1.55031466e+00 -1.81489959e-01 8.66808295e-02 2.08247900e-01
-1.04855895e+00 -2.13480636e-01 -3.13600630e-01 3.85601729e-01
2.97586620e-01 5.89185178e-01 -3.89453530e-01 3.62462103e-01
5.24377406e-01 1.13221741e+00 -5.27556300e-01 7.77360678e-01
-3.58303860e-02 4.89531159e-01 2.86008507e-01 4.32838589e-01
1.54323727e-01 1.26233831e-01 4.28710163e-01 1.36003566e+00
3.09981227e-01 8.27877969e-02 3.72264743e-01 1.64099193e+00
-3.62535715e-01 6.27991334e-02 -5.44252992e-01 -2.00173363e-01
2.71480680e-01 1.27905500e+00 -5.33152819e-01 -4.79047686e-01
-2.62563378e-01 1.34433150e+00 5.63540399e-01 3.29486549e-01
-9.63653803e-01 -1.07436523e-01 5.09848893e-01 1.78194270e-01
8.45970511e-02 1.87944807e-02 -3.71885717e-01 -1.49939716e+00
-3.65348943e-02 -8.18348467e-01 -3.32016754e-03 -1.10798228e+00
-1.70117986e+00 7.81553030e-01 4.04471382e-02 -1.00005054e+00
-9.46779102e-02 -5.30484736e-01 -3.85580212e-01 7.49052823e-01
-1.51788080e+00 -1.42627525e+00 -6.30207956e-01 9.02418256e-01
6.53830171e-01 -2.63997763e-01 7.91846514e-01 5.31888485e-01
-3.79544228e-01 4.67495382e-01 -2.06132933e-01 1.39403746e-01
6.83637142e-01 -8.65127087e-01 2.84292281e-01 6.96615577e-01
2.13922575e-01 4.24155146e-01 2.86405712e-01 -5.85247278e-01
-1.66211784e+00 -1.13440883e+00 3.29759389e-01 -2.53726929e-01
8.31239402e-01 -8.96885097e-01 -1.01952708e+00 5.11527836e-01
1.03457287e-01 3.41488421e-01 5.68094969e-01 -3.23020846e-01
-7.73477793e-01 -5.72650731e-02 -9.35619891e-01 5.28074145e-01
1.43701530e+00 -1.19289804e+00 -6.58560514e-01 6.06605828e-01
8.32904100e-01 -1.41085565e-01 -9.39513206e-01 2.37893581e-01
6.72802925e-01 -7.39414513e-01 1.22394800e+00 -6.11293852e-01
9.55931842e-01 1.00074656e-01 -1.67295769e-01 -1.47479773e+00
-5.02891183e-01 -2.16779143e-01 2.52184093e-01 9.79165435e-01
6.38391748e-02 -3.25511873e-01 1.25833645e-01 1.53207779e-01
-2.97228903e-01 -4.89289641e-01 -9.85946059e-01 -5.44372678e-01
1.30010039e-01 -5.74320078e-01 4.19657260e-01 1.17157555e+00
-1.96692012e-02 5.31189740e-01 -6.63788736e-01 -3.80550444e-01
8.06332529e-01 2.41751805e-01 2.72441357e-01 -1.03859103e+00
-4.63506818e-01 -6.66196525e-01 -3.48751575e-01 -7.23182201e-01
8.35591257e-01 -1.94088662e+00 3.32049191e-01 -1.42275012e+00
6.45960152e-01 -1.59596041e-01 -5.64118266e-01 7.45468080e-01
-5.92988431e-02 6.06983542e-01 4.63981092e-01 2.13155001e-01
-5.09049654e-01 6.40786886e-01 1.53081107e+00 -5.25294960e-01
3.73978794e-01 -7.58630335e-01 -8.18481684e-01 1.05593190e-01
2.85635978e-01 -3.91579568e-01 -2.59223789e-01 -6.57475889e-01
-8.33744854e-02 5.23243487e-01 8.70944202e-01 -1.01738954e+00
3.73492837e-02 -2.21779756e-02 4.27091986e-01 -7.40790889e-02
3.79449248e-01 -6.04849219e-01 2.25056201e-01 3.30885708e-01
-7.11167693e-01 -2.83289045e-01 5.56431450e-02 6.01851225e-01
-3.31228316e-01 1.32073298e-01 9.45294142e-01 -3.09195131e-01
-8.63688409e-01 5.65314293e-01 -1.14746988e-01 3.92698318e-01
7.23578513e-01 -3.95021513e-02 -4.52693313e-01 -1.25564188e-01
-7.56562293e-01 4.90004569e-02 2.03073695e-01 6.46859467e-01
9.89700258e-01 -1.57080781e+00 -8.62459302e-01 4.02997166e-01
4.25328076e-01 -6.55124426e-01 6.84360266e-01 1.32834959e+00
-1.10669779e-02 3.65285844e-01 -7.56488681e-01 -7.56867051e-01
-6.83901012e-01 7.38388062e-01 3.00572366e-01 7.63856396e-02
-9.26010072e-01 7.84964919e-01 7.88740396e-01 -3.52002949e-01
6.75267950e-02 -1.15872949e-01 7.91206136e-02 1.96391925e-01
7.49632776e-01 -1.43724814e-01 8.69126692e-02 -8.35602939e-01
-3.28011513e-01 4.00839478e-01 2.51715183e-01 -2.66950935e-01
1.59099126e+00 -2.62481906e-02 -3.78855109e-01 5.10198772e-01
1.58367467e+00 -5.17622590e-01 -1.24419916e+00 -3.16746652e-01
-2.48225644e-01 -5.76092243e-01 3.95436674e-01 -1.00011313e+00
-1.41746461e+00 1.18080878e+00 7.40679860e-01 -4.18184578e-01
1.15988743e+00 1.81463510e-01 1.05172098e+00 -8.37420225e-02
4.76252615e-01 -7.00979948e-01 5.43933094e-01 2.28706524e-01
1.27683878e+00 -1.41584146e+00 -4.50822234e-01 -1.36716172e-01
-8.31497490e-01 7.32062936e-01 4.99913901e-01 -3.19549367e-02
6.05263591e-01 2.91543696e-02 -4.68098015e-01 -3.90328467e-01
-8.11174154e-01 -2.54187167e-01 7.84076095e-01 6.65394008e-01
1.88808054e-01 3.32373291e-01 -1.17321797e-01 8.51031661e-01
-8.69385079e-02 9.38005894e-02 3.43693681e-02 3.81459981e-01
-3.17529321e-01 -6.45801365e-01 -1.43476576e-01 8.31510186e-01
-2.37846911e-01 -4.30301547e-01 -1.46758690e-01 2.11425334e-01
2.81272307e-02 6.89113081e-01 2.11169168e-01 -3.92324716e-01
8.28300565e-02 2.48260900e-01 7.44868279e-01 -6.98419631e-01
-6.53660119e-01 2.94779371e-02 -1.85065523e-01 -9.43239093e-01
-3.74941409e-01 -3.50773841e-01 -1.20769274e+00 -6.21131845e-02
2.28994653e-01 -4.14250880e-01 6.74460471e-01 9.79334056e-01
5.21574736e-01 6.07631683e-01 4.99670893e-01 -1.12318873e+00
-2.76315808e-01 -9.10946310e-01 -5.93249440e-01 1.06922233e+00
2.97993988e-01 -8.89309227e-01 -1.07005415e-02 1.38729498e-01] | [10.777883529663086, 2.467440366744995] |
2e964072-e9dd-44f8-b797-f7880fbc2450 | memetic-eda-based-approaches-to-comprehensive | 1906.07900 | null | https://arxiv.org/abs/1906.07900v1 | https://arxiv.org/pdf/1906.07900v1.pdf | Memetic EDA-Based Approaches to Comprehensive Quality-Aware Automated Semantic Web Service Composition | Comprehensive quality-aware automated semantic web service composition is an NP-hard problem, where service composition workflows are unknown, and comprehensive quality, i.e., Quality of services (QoS) and Quality of semantic matchmaking (QoSM) are simultaneously optimized. The objective of this problem is to find a solution with optimized or near-optimized overall QoS and QoSM within polynomial time over a service request. In this paper, we proposed novel memetic EDA-based approaches to tackle this problem. The proposed method investigates the effectiveness of several neighborhood structures of composite services by proposing domain-dependent local search operators. Apart from that, a joint strategy of the local search procedure is proposed to integrate with a modified EDA to reduce the overall computation time of our memetic approach. To better demonstrate the effectiveness and scalability of our approach, we create a more challenging, augmented version of the service composition benchmark based on WSC-08 \cite{bansal2008wsc} and WSC-09 \cite{kona2009wsc}. Experimental results on this benchmark show that one of our proposed memetic EDA-based approach (i.e., MEEDA-LOP) significantly outperforms existing state-of-the-art algorithms. | ['Sven Hartmann', 'Gang Chen', 'Hui Ma', 'Chen Wang'] | 2019-06-19 | null | null | null | null | ['service-composition'] | ['miscellaneous'] | [ 1.53623521e-01 -3.65787745e-01 9.42179263e-02 -3.90304148e-01
-6.23156190e-01 -4.91669863e-01 2.60597497e-01 9.38874856e-03
-2.27800608e-01 6.29637599e-01 -7.47211054e-02 -9.40943658e-02
-8.44082355e-01 -8.37707579e-01 -5.24841666e-01 -7.93581247e-01
-2.10261047e-01 7.28691578e-01 4.93720770e-01 -4.97185588e-01
4.25063908e-01 4.16371107e-01 -1.99365234e+00 3.60300660e-01
1.19999230e+00 1.18775582e+00 2.83479810e-01 1.70116886e-01
-5.40646553e-01 1.39872819e-01 -5.98740041e-01 -3.16940188e-01
3.71638924e-01 -5.96663952e-01 -1.00410569e+00 3.47961724e-01
-4.34019268e-01 2.60904700e-01 1.89749509e-01 1.26533008e+00
5.22962093e-01 5.17399192e-01 1.42433256e-01 -2.00290632e+00
-4.76392508e-01 8.60841751e-01 -4.99670297e-01 4.40539978e-02
4.98100817e-01 1.15968622e-01 1.14589381e+00 -4.11478341e-01
7.13333845e-01 1.26015401e+00 2.32664660e-01 5.87855756e-01
-1.13282132e+00 -4.28680539e-01 5.57926059e-01 8.02497804e-01
-1.47968769e+00 -3.83149534e-01 5.98641276e-01 3.81709963e-01
1.02613544e+00 7.05097914e-01 6.42954469e-01 4.26527768e-01
-2.36089647e-01 4.86151189e-01 1.09687185e+00 -5.40234327e-01
8.99724603e-01 1.71112437e-02 -1.59821153e-01 5.88404059e-01
3.53320241e-01 -1.24610335e-01 -5.14170647e-01 -7.12292969e-01
-1.57855660e-01 2.53168792e-02 -1.88332930e-01 -3.45686704e-01
-8.16828847e-01 5.96937537e-01 -2.27878820e-02 4.47420925e-01
-3.94164294e-01 -2.54890304e-02 2.30783969e-01 4.41127002e-01
2.29834229e-01 3.68102044e-01 -6.27119839e-01 -3.92750382e-01
-6.58262432e-01 5.10602117e-01 1.29095340e+00 1.13892221e+00
4.09208536e-01 -7.42036849e-02 1.03111476e-01 8.10578763e-01
1.51030511e-01 2.53532618e-01 2.73876607e-01 -1.09997141e+00
2.26903856e-01 8.98500085e-01 3.59861374e-01 -7.95960605e-01
-4.85475481e-01 -2.97255278e-01 -3.10321838e-01 3.92669551e-02
-5.96824195e-03 1.71489030e-01 -5.66483617e-01 1.58237028e+00
6.14655256e-01 3.90107155e-01 1.09546527e-01 1.27603507e+00
4.59626555e-01 7.16466665e-01 -5.94364777e-02 -7.18970895e-01
1.23483992e+00 -1.29570019e+00 -6.70686841e-01 1.99413925e-01
3.76528978e-01 -7.42802620e-01 9.79624033e-01 3.59903395e-01
-1.27434790e+00 3.32184196e-01 -1.04729331e+00 6.29108667e-01
-3.83143336e-01 -6.73456788e-01 7.43891001e-01 7.32893884e-01
-9.57625985e-01 5.16772509e-01 -5.32797098e-01 -6.08271480e-01
8.73320699e-02 4.09498781e-01 2.70190947e-02 -5.43191791e-01
-9.43153262e-01 6.59095287e-01 3.38296592e-01 -1.93500757e-01
-7.82157660e-01 -4.96737868e-01 -3.84711862e-01 3.66714656e-01
1.18988323e+00 -7.24751472e-01 1.07935572e+00 -9.55153763e-01
-1.51888204e+00 4.33126807e-01 -2.66706407e-01 -2.31113583e-01
4.06541973e-01 6.41185045e-01 -9.73298550e-01 1.39254257e-01
-1.54842004e-01 -2.89259385e-03 2.84403205e-01 -1.61201823e+00
-1.27990901e+00 -3.98346961e-01 4.77012247e-01 2.00126976e-01
-4.65175480e-01 5.06062388e-01 -4.58847225e-01 -2.11125880e-01
3.98063868e-01 -9.21023667e-01 -6.17974520e-01 -3.09079349e-01
8.25653747e-02 -4.29732621e-01 4.39715087e-01 -3.50802690e-01
1.32270253e+00 -1.77684546e+00 2.15394378e-01 7.11369455e-01
-1.45967185e-01 -1.07975438e-01 -5.73936284e-01 7.64750004e-01
4.20142561e-01 2.94160485e-01 -6.28628016e-01 -4.10633236e-02
4.25589859e-01 6.44755244e-01 4.83857691e-01 4.35204983e-01
-7.41634816e-02 3.66642267e-01 -7.53777683e-01 -4.78302956e-01
-3.06912780e-01 1.92779362e-01 -7.97077537e-01 8.74798745e-03
-7.12725401e-01 6.92908615e-02 -3.69647473e-01 1.09414887e+00
9.19097364e-01 -3.20230067e-01 7.61346698e-01 1.29704088e-01
-1.73178494e-01 1.78211749e-01 -1.62927210e+00 1.91304672e+00
-5.46325684e-01 -5.08257866e-01 3.14935625e-01 -1.10822153e+00
7.29808509e-01 4.30810452e-01 1.12911344e+00 -1.14951444e+00
1.63264409e-01 5.34178853e-01 1.78243101e-01 -5.04150271e-01
1.96723372e-01 7.42776394e-02 6.11935295e-02 6.95305705e-01
-3.17326427e-01 2.77338296e-01 8.22355926e-01 -1.11460119e-01
1.29726577e+00 1.13705732e-02 7.51531869e-02 -5.79789877e-01
8.77658606e-01 1.16594806e-01 1.12585580e+00 5.40059268e-01
-1.65833861e-01 1.36118829e-01 3.74780536e-01 -4.17196244e-01
-8.74695063e-01 -4.93602097e-01 2.27906227e-01 1.36428213e+00
7.72509098e-01 -6.42421126e-01 -7.27825582e-01 -6.75690174e-01
4.26590703e-02 9.41441298e-01 -1.36893243e-01 -1.34865969e-01
-5.80639958e-01 -9.88384008e-01 7.25962222e-02 2.96793599e-03
1.20012157e-01 -1.00871396e+00 -6.53185487e-01 5.78615189e-01
-3.33431423e-01 -1.36342990e+00 -3.30279291e-01 -1.25369236e-01
-6.18883610e-01 -9.93778110e-01 -1.79799020e-01 -6.38682246e-01
5.71762919e-01 3.08575213e-01 1.24817932e+00 5.24088979e-01
-2.21642599e-01 1.35962620e-01 -7.73597658e-01 -5.24741970e-03
-3.35304350e-01 5.69613427e-02 -1.35555174e-02 1.72913224e-01
2.81863362e-01 -1.01818967e+00 -5.00273168e-01 7.26040184e-01
-1.07460630e+00 -1.34116054e-01 3.72430801e-01 6.37026191e-01
7.22669661e-01 7.28544831e-01 6.40506625e-01 -6.39348328e-01
7.94458568e-01 -6.29688025e-01 -8.18698168e-01 6.41330957e-01
-1.26184165e+00 1.08471282e-01 7.38264203e-01 -2.49978423e-01
-1.03981519e+00 -2.10429266e-01 -1.97131813e-01 1.40942752e-01
8.68108124e-02 4.60874051e-01 -6.92803919e-01 -1.65126771e-01
4.46834899e-02 3.80396128e-01 -2.49146193e-01 -7.66986966e-01
5.85654229e-02 5.30467153e-01 2.66558945e-01 -9.81955409e-01
6.86539173e-01 5.88106036e-01 1.49275184e-01 -1.15872063e-01
-1.97892785e-01 -5.37519455e-01 2.50078559e-01 -1.40786231e-01
2.65855968e-01 -2.20497012e-01 -1.35656750e+00 -2.77296212e-02
-8.80234480e-01 3.25815320e-01 -1.12956159e-01 -1.83169413e-02
-5.61173916e-01 2.02044442e-01 -3.17846596e-01 -1.04915226e+00
-4.20379788e-01 -1.38242769e+00 6.13612890e-01 1.45888492e-01
2.59464443e-01 -4.69918311e-01 -2.88385302e-01 3.43950719e-01
7.09555387e-01 2.40300328e-01 1.05793214e+00 -8.96645308e-01
-6.51180625e-01 3.72818708e-02 -1.51034981e-01 -1.46203071e-01
-6.50108652e-03 -3.15589160e-01 -3.06054235e-01 -4.23766941e-01
-1.58822581e-01 2.39772454e-01 1.59617066e-01 -1.80641025e-01
1.22189546e+00 -6.81462049e-01 -1.07118681e-01 4.66930479e-01
1.87131858e+00 6.51076317e-01 3.70553076e-01 8.92226756e-01
4.23202664e-03 6.03086591e-01 7.78972149e-01 1.15230525e+00
5.37961721e-01 9.69553590e-01 8.04317236e-01 3.93429607e-01
1.48995191e-01 3.06931704e-01 3.15002948e-02 9.41118538e-01
-2.92130589e-01 -5.97471833e-01 -7.26573825e-01 3.98641288e-01
-2.11189485e+00 -6.24376476e-01 8.62679705e-02 1.78653812e+00
5.31395555e-01 -8.59509557e-02 2.73401707e-01 5.87102294e-01
7.28685498e-01 -2.49576390e-01 -6.90756917e-01 -6.88170791e-01
-4.13784176e-01 1.70525685e-01 3.97635370e-01 1.81016967e-01
-4.14531291e-01 6.52200520e-01 4.62116337e+00 7.52119362e-01
-6.29168332e-01 5.37260771e-01 5.02330624e-03 -4.44761813e-01
-7.34299421e-01 2.46273607e-01 -3.69919211e-01 7.14594841e-01
1.01350546e+00 -6.62072778e-01 1.41738629e+00 6.76000774e-01
2.78079510e-01 1.29858721e-02 -8.39497268e-01 7.43474722e-01
9.22742784e-02 -1.32533109e+00 -1.12346143e-01 1.94231197e-01
1.06387925e+00 -3.50788593e-01 -2.96769559e-01 4.50822227e-02
3.43598604e-01 -5.59578001e-01 8.64918947e-01 2.74170756e-01
3.52179945e-01 -1.19587111e+00 8.27142417e-01 2.46947974e-01
-1.18707108e+00 -6.14231706e-01 -1.59663558e-01 2.51471639e-01
3.71818721e-01 4.33384001e-01 -1.63153365e-01 1.05891728e+00
1.13609791e+00 6.92285411e-03 -9.32130069e-02 1.26154304e+00
2.13424534e-01 2.02328488e-01 -3.63111615e-01 -3.15838277e-01
3.96902919e-01 -2.67456055e-01 7.87450075e-01 9.00354147e-01
4.41916645e-01 3.55376720e-01 4.11793768e-01 7.03834951e-01
-6.76989853e-02 5.28610885e-01 2.84545183e-01 -1.79439843e-01
7.55889654e-01 1.12263799e+00 -9.00464833e-01 -2.03506142e-01
-3.13650042e-01 7.99852729e-01 -5.27581573e-02 3.71783853e-01
-9.78101015e-01 -2.12308556e-01 1.09400856e+00 -4.27251868e-03
3.01267177e-01 1.08668163e-01 -2.95542568e-01 -9.26635742e-01
3.55082452e-01 -1.22730184e+00 7.71849453e-01 -3.32317352e-01
-1.14637601e+00 7.60639310e-01 -4.03037846e-01 -1.00227749e+00
-3.40225175e-02 1.77728832e-02 -3.28196287e-01 6.43889487e-01
-1.64826262e+00 -9.02068377e-01 -3.44187707e-01 6.48083806e-01
4.61639106e-01 -3.18828523e-01 8.51123631e-01 7.07142234e-01
-4.27545488e-01 5.42948127e-01 2.77129203e-01 -9.90423203e-01
2.06496477e-01 -7.56345153e-01 1.14369385e-01 8.97154689e-01
-3.24287415e-01 2.59294569e-01 1.04168367e+00 -2.57675409e-01
-2.02430081e+00 -7.80773878e-01 8.51163030e-01 3.89777899e-01
4.69612271e-01 -8.39779153e-02 -5.92248380e-01 -6.43911660e-02
-1.15840565e-02 -5.46491668e-02 5.46241522e-01 -1.44344702e-01
-3.95304486e-02 -4.57241178e-01 -1.73338234e+00 6.14282131e-01
1.79165816e+00 1.61013246e-01 6.58412129e-02 4.00488079e-01
1.11727834e+00 4.95855883e-03 -9.74682689e-01 5.39953232e-01
4.40169811e-01 -9.26065922e-01 8.19636464e-01 -8.22013080e-01
8.79371986e-02 -5.74132383e-01 -7.53329813e-01 -1.13502645e+00
-3.19240272e-01 -6.34935260e-01 -2.62916148e-01 1.29817796e+00
3.68534803e-01 -8.16579163e-01 6.48545504e-01 6.55452371e-01
-1.99114442e-01 -8.84312093e-01 -1.06900871e+00 -1.24709678e+00
-4.63587642e-01 -2.18307897e-01 1.64975536e+00 7.77456045e-01
-8.38400424e-02 -3.63944948e-01 -1.15433551e-01 4.66368973e-01
6.57862782e-01 6.09110832e-01 2.10553557e-01 -1.23667789e+00
-4.24978554e-01 -6.70959830e-01 -2.07467705e-01 -8.90121311e-02
1.89359143e-01 -9.55514371e-01 -9.72582996e-02 -1.85649109e+00
1.47990465e-01 -6.39759779e-01 -6.90703511e-01 4.86558139e-01
2.62989998e-01 -1.55660450e-01 4.41045880e-01 1.15019269e-01
-1.06935287e+00 4.73861784e-01 1.09410357e+00 -1.67913496e-01
-1.83209270e-01 8.67197961e-02 -6.83391571e-01 3.06533337e-01
6.72162116e-01 -9.60314512e-01 -2.55214274e-01 -3.32940698e-01
3.51477534e-01 2.97183692e-01 -1.44597173e-01 -7.80890763e-01
4.12896752e-01 -8.53546381e-01 -5.99745750e-01 -2.06705838e-01
2.68175662e-01 -1.24562645e+00 5.28331399e-01 5.30841231e-01
-2.74011821e-01 4.27312791e-01 -1.63319632e-01 3.54651839e-01
-1.48809969e-01 -4.23641056e-01 5.63102245e-01 -2.28276044e-01
-7.93554902e-01 3.20785165e-01 -3.22934985e-01 -6.02477640e-02
1.09733033e+00 4.44583036e-03 -2.07776219e-01 -1.36429638e-01
-5.39172590e-01 4.72029507e-01 5.93338251e-01 5.37657797e-01
5.08940101e-01 -1.18176663e+00 -7.07104564e-01 -3.30386221e-01
3.14235479e-01 -6.01571321e-01 2.91896164e-01 7.42819011e-01
-3.38472009e-01 3.05784881e-01 -3.82575631e-01 -1.18243292e-01
-1.19179440e+00 7.88414478e-01 1.71725899e-01 -2.27134317e-01
-5.12949288e-01 5.73578715e-01 -5.84211707e-01 -6.33987188e-01
4.45981383e-01 1.29690439e-01 2.13562697e-01 -4.15649295e-01
4.53320920e-01 8.39729667e-01 4.32364285e-01 -3.45682025e-01
-8.53919208e-01 1.75191611e-02 4.15739387e-01 -2.77763605e-01
1.65442705e+00 -2.56429702e-01 -6.03943586e-01 -2.44218737e-01
9.01094913e-01 -2.36624271e-01 -6.01928830e-01 -4.03814018e-01
5.33228993e-01 -6.69340849e-01 -1.70694470e-01 -1.32717443e+00
-1.43069541e+00 2.86347102e-02 7.75281549e-01 2.51739383e-01
1.71806264e+00 -1.99457616e-01 1.16584349e+00 2.60213137e-01
6.91146612e-01 -1.31319547e+00 -4.50351238e-01 6.62513226e-02
7.32828021e-01 -7.85602808e-01 -2.19385758e-01 -4.80881900e-01
-3.29751790e-01 8.21191967e-01 6.61704957e-01 4.19024348e-01
2.91477442e-01 4.65073675e-01 -3.30481291e-01 -2.02770308e-02
-1.20712781e+00 -4.56024051e-01 -2.60121115e-02 3.58434558e-01
-1.40896842e-01 2.64427364e-01 -1.18397129e+00 1.05102944e+00
1.02591313e-01 -1.78163603e-01 2.98086524e-01 1.42501342e+00
-3.87155384e-01 -1.43426561e+00 -3.31123501e-01 6.93921000e-02
-4.00023758e-01 -3.42981890e-04 -2.05307350e-01 2.85686135e-01
4.10325974e-01 1.49287784e+00 -2.56063581e-01 -3.51272523e-01
5.56427181e-01 1.51165035e-02 3.79883736e-01 7.67611712e-02
-7.22605824e-01 1.51988506e-01 4.35751736e-01 -8.56076837e-01
-3.66793841e-01 -6.28819406e-01 -1.64051437e+00 -6.00243688e-01
-2.67740011e-01 6.08699203e-01 1.33264446e+00 9.87763703e-01
4.79008496e-01 7.40496755e-01 7.73589253e-01 -2.45253727e-01
-7.07331300e-01 -4.09886450e-01 -5.76227903e-01 4.49545443e-01
-4.90196168e-01 -6.93977594e-01 -1.73000529e-01 -2.72402465e-01] | [8.573087692260742, 6.961820125579834] |
7d438ae5-0c9c-47f3-a3bd-8a7e92843cf5 | snap-self-supervised-neural-maps-for-visual | 2306.05407 | null | https://arxiv.org/abs/2306.05407v1 | https://arxiv.org/pdf/2306.05407v1.pdf | SNAP: Self-Supervised Neural Maps for Visual Positioning and Semantic Understanding | Semantic 2D maps are commonly used by humans and machines for navigation purposes, whether it's walking or driving. However, these maps have limitations: they lack detail, often contain inaccuracies, and are difficult to create and maintain, especially in an automated fashion. Can we use raw imagery to automatically create better maps that can be easily interpreted by both humans and machines? We introduce SNAP, a deep network that learns rich neural 2D maps from ground-level and overhead images. We train our model to align neural maps estimated from different inputs, supervised only with camera poses over tens of millions of StreetView images. SNAP can resolve the location of challenging image queries beyond the reach of traditional methods, outperforming the state of the art in localization by a large margin. Moreover, our neural maps encode not only geometry and appearance but also high-level semantics, discovered without explicit supervision. This enables effective pre-training for data-efficient semantic scene understanding, with the potential to unlock cost-efficient creation of more detailed maps. | ['Simon Lynen', 'Jan Hosang', 'Marc Pollefeys', 'Eduard Trulls', 'Paul-Edouard Sarlin'] | 2023-06-08 | null | null | null | null | ['scene-understanding'] | ['computer-vision'] | [ 1.57934695e-01 2.92636901e-01 9.59431380e-02 -7.13625908e-01
-7.38348424e-01 -9.77665901e-01 5.78709900e-01 2.98157990e-01
-5.00786424e-01 5.24126649e-01 1.49110585e-01 -2.37500235e-01
5.26066422e-02 -1.06113338e+00 -1.06757796e+00 8.61463894e-04
-2.65003163e-02 9.02770996e-01 6.54191017e-01 -4.63739038e-01
3.57459992e-01 4.89872724e-01 -1.80978084e+00 7.90109113e-03
8.28807652e-01 1.04703188e+00 6.69078887e-01 4.64335024e-01
-1.62405550e-01 7.30792105e-01 -1.48476511e-01 -1.41661733e-01
4.76839662e-01 1.86857805e-01 -6.33495390e-01 -9.06373635e-02
9.26401615e-01 -8.09419036e-01 -5.03647268e-01 9.09879804e-01
1.55540869e-01 4.66074720e-02 3.73522133e-01 -1.04131687e+00
-5.40738642e-01 -1.39861807e-01 -3.50149930e-01 2.92416825e-03
3.92486364e-01 3.17186445e-01 6.42759860e-01 -5.06804109e-01
6.22595906e-01 1.08620179e+00 9.69920933e-01 -5.14675900e-02
-1.11529827e+00 -5.01277268e-01 1.53402686e-01 -6.48319162e-03
-1.44202101e+00 -3.79763752e-01 6.11648917e-01 -6.02934182e-01
1.14512241e+00 -3.26846056e-02 5.94287574e-01 8.70996058e-01
-1.73238799e-01 3.95292610e-01 8.87698412e-01 -6.44113198e-02
2.46220231e-01 -3.79948281e-02 -2.26393804e-01 8.90589535e-01
3.93215984e-01 -7.96933547e-02 -7.01076388e-01 1.34723857e-01
1.10192895e+00 2.32791826e-02 -2.65050847e-02 -8.04436147e-01
-1.45292008e+00 7.30782926e-01 8.42091501e-01 -1.91194728e-01
-5.26290476e-01 4.33250904e-01 -5.12375608e-02 4.76212017e-02
3.61808002e-01 8.23257983e-01 -4.29104060e-01 -2.13815436e-01
-7.59044528e-01 4.21958506e-01 4.51005429e-01 1.10978818e+00
1.23095930e+00 -2.23159999e-01 6.81305468e-01 4.45715934e-01
-2.59903669e-02 9.79241252e-01 1.22251719e-01 -1.44333434e+00
6.80636644e-01 6.96101546e-01 5.29257238e-01 -1.42171192e+00
-7.16204286e-01 -2.12949976e-01 -6.44216061e-01 2.91370481e-01
4.06902701e-01 2.31850520e-01 -1.00904834e+00 1.52908242e+00
7.73019344e-02 8.21165219e-02 -2.53110170e-01 1.04402995e+00
5.54192483e-01 4.72356319e-01 -1.36082962e-01 8.63998950e-01
1.11368489e+00 -8.25616062e-01 -2.03345641e-01 -1.04811788e+00
6.32940710e-01 -5.32899678e-01 1.05864787e+00 1.30691931e-01
-7.94119895e-01 -5.30656576e-01 -1.24891794e+00 -5.90147614e-01
-8.43017936e-01 4.03659381e-02 6.88618958e-01 2.96113133e-01
-1.34360409e+00 4.07222688e-01 -8.84715080e-01 -5.82632005e-01
5.31912863e-01 2.02683225e-01 -8.00875723e-01 -2.48727441e-01
-1.04550159e+00 1.26412797e+00 4.02071685e-01 3.05012427e-02
-6.82977378e-01 -6.82190061e-01 -1.36135304e+00 -1.79100260e-01
3.37222099e-01 -6.66405439e-01 1.13883436e+00 -6.55642986e-01
-9.23773289e-01 1.03371298e+00 -1.84564531e-01 -5.65863192e-01
5.45813441e-01 -3.83659184e-01 -4.75410372e-02 2.65573889e-01
7.50761926e-01 1.23637915e+00 5.51803350e-01 -1.30581164e+00
-7.78490901e-01 -4.94438022e-01 3.92060429e-01 3.29540968e-01
2.29522109e-01 -6.68187261e-01 -5.41019142e-01 -1.29137337e-01
5.55278957e-01 -9.47513282e-01 -3.40063632e-01 4.14093077e-01
-2.53311634e-01 2.55056292e-01 7.18259513e-01 -9.67112243e-01
5.81205368e-01 -2.09144163e+00 -1.51809573e-01 1.64378092e-01
2.77332753e-01 -8.86332765e-02 -1.91775084e-01 2.77731776e-01
2.83142447e-01 1.34970352e-01 -3.22111338e-01 -1.14071004e-01
1.58369094e-01 2.90836006e-01 -4.72453177e-01 3.67298514e-01
2.42946282e-01 1.06314480e+00 -1.10216248e+00 -1.10409446e-01
6.68162763e-01 4.05123204e-01 -4.31693196e-01 8.39732513e-02
-3.60989928e-01 4.53783989e-01 -4.64892060e-01 5.40308058e-01
7.77011752e-01 -4.39478934e-01 -9.36557651e-02 -1.94108635e-01
-2.98559815e-01 5.78305960e-01 -1.04506469e+00 2.21089864e+00
-5.94207466e-01 8.81059289e-01 -1.92186713e-01 -7.72940218e-01
9.17603135e-01 -3.89939964e-01 2.33248726e-01 -1.19765377e+00
-1.92228854e-01 2.68872529e-01 -6.11466110e-01 -3.10234398e-01
8.52154553e-01 2.47913882e-01 -3.32418114e-01 4.54688907e-01
-1.34420678e-01 -7.50487566e-01 -1.34616241e-01 1.34766459e-01
1.09095466e+00 2.18688264e-01 7.53989145e-02 -5.68772480e-02
-9.65605117e-03 6.45788670e-01 1.15949176e-01 8.95985901e-01
5.54141179e-02 9.32334125e-01 2.18174115e-01 -9.48240280e-01
-1.54919982e+00 -1.19468582e+00 1.09150447e-01 7.70137906e-01
7.49160111e-01 -2.26755664e-01 -4.78518784e-01 -4.76550102e-01
2.35044956e-01 6.56394064e-01 -4.94450361e-01 9.54200886e-03
-3.66097927e-01 -3.51968825e-01 5.29229522e-01 6.89801574e-01
9.91160810e-01 -5.03760815e-01 -1.07134843e+00 1.13233164e-01
-4.93418932e-01 -1.49583304e+00 -9.41766724e-02 5.73481955e-02
-6.56515241e-01 -1.09886432e+00 -3.28602821e-01 -5.22958934e-01
8.08561563e-01 7.16328204e-01 1.22020626e+00 -2.34094616e-02
-1.36337206e-01 2.31690794e-01 -1.81385607e-01 -4.00164545e-01
-9.51939300e-02 4.06752795e-01 5.36916293e-02 -4.62756753e-01
2.87080824e-01 -7.16954052e-01 -6.58309579e-01 5.22720695e-01
-6.19952798e-01 5.62049925e-01 7.81240344e-01 4.80148613e-01
5.82870722e-01 3.07588652e-02 1.59459978e-01 -6.44720256e-01
6.85167760e-02 -3.70944917e-01 -8.15288246e-01 8.12052861e-02
-4.92835850e-01 2.91172445e-01 1.79397136e-01 2.79394358e-01
-9.01888371e-01 3.21547806e-01 -3.25644091e-02 -2.51433432e-01
-3.71550053e-01 3.62800598e-01 1.33400634e-01 -3.23340803e-01
9.11209106e-01 1.10282548e-01 2.33308366e-03 -4.88826215e-01
5.99030077e-01 4.52958107e-01 1.07321894e+00 -3.41040999e-01
9.37069356e-01 8.55035365e-01 2.58362126e-02 -6.79650366e-01
-1.05266023e+00 -4.05087322e-01 -1.07001066e+00 9.19545144e-02
9.56364870e-01 -1.29862142e+00 -2.55084038e-01 4.52207118e-01
-1.22461557e+00 -5.91985285e-01 9.33916494e-02 2.79010326e-01
-5.90477645e-01 4.15873528e-03 -1.02799051e-01 -3.56143147e-01
1.66209295e-01 -9.48367059e-01 1.61394095e+00 1.91259116e-01
-2.44943276e-01 -7.34924316e-01 -1.23103134e-01 5.15323162e-01
5.50282836e-01 4.47590560e-01 6.72155499e-01 -1.22020200e-01
-1.22709763e+00 -3.80399704e-01 -7.91681051e-01 -6.18764162e-02
1.67618632e-01 -5.00989199e-01 -1.07026172e+00 5.66752627e-02
-4.47119981e-01 -4.55254674e-01 7.48462796e-01 1.40366644e-01
1.05124211e+00 -1.80219159e-01 -4.23424274e-01 8.06729496e-01
1.42838085e+00 -1.30856320e-01 5.60787559e-01 7.62614667e-01
8.61348927e-01 7.31696188e-01 4.83938515e-01 6.79784641e-02
1.02737415e+00 5.64006448e-01 8.07000935e-01 -2.59309381e-01
-2.87708221e-03 -6.40224159e-01 -3.32314104e-01 2.63004214e-01
-9.03873742e-02 4.90775071e-02 -1.32710433e+00 6.89684212e-01
-1.93288016e+00 -7.42714822e-01 -2.42721871e-03 2.30167484e+00
3.43750298e-01 3.09513897e-01 -4.18837845e-01 -4.07603055e-01
4.87388134e-01 3.28316957e-01 -8.41561258e-01 2.66118832e-02
-3.06882381e-01 -4.27060872e-02 1.33690155e+00 5.66511095e-01
-1.32745743e+00 1.16361117e+00 6.70845795e+00 1.75499767e-01
-1.15570891e+00 5.51657788e-02 4.60180849e-01 -7.33116716e-02
-4.12044644e-01 1.71903953e-01 -3.85601491e-01 2.81712323e-01
7.90766120e-01 2.98492134e-01 5.77064872e-01 1.06878102e+00
8.23694766e-02 -6.19614959e-01 -9.30937886e-01 1.29996359e+00
1.79260865e-01 -1.76751006e+00 -2.27046549e-01 1.10964663e-01
7.81868458e-01 6.48216665e-01 -1.73973158e-01 1.00046247e-01
5.85214853e-01 -1.19452596e+00 9.90364969e-01 6.08045459e-01
8.46002698e-01 -5.44504762e-01 7.27265954e-01 4.25440669e-01
-1.01405537e+00 1.00398205e-01 -4.10122037e-01 -3.89782518e-01
2.65480012e-01 3.28475505e-01 -8.67079735e-01 4.01503474e-01
9.08158183e-01 8.51937175e-01 -9.51416433e-01 8.55796933e-01
-3.68140668e-01 -7.72304684e-02 -7.58633554e-01 1.69518232e-01
4.89777803e-01 -2.09975708e-02 1.36657283e-01 6.91256225e-01
4.48203057e-01 -1.30078778e-01 2.89948612e-01 8.57948184e-01
1.20298311e-01 -4.82466757e-01 -1.15179491e+00 4.93722074e-02
7.22929299e-01 9.14344847e-01 -7.62139142e-01 -2.52729326e-01
-2.19611794e-01 1.07793069e+00 5.32037795e-01 2.14958340e-01
-6.58513308e-01 -4.20674533e-01 8.12067568e-01 4.72765893e-01
1.58392847e-01 -8.39035749e-01 -6.23510242e-01 -1.01778936e+00
1.76433012e-01 -3.03283870e-01 -1.59965724e-01 -1.40879750e+00
-8.54536295e-01 5.33458233e-01 9.01727602e-02 -1.15299463e+00
-4.84857738e-01 -7.14775920e-01 -1.56616777e-01 9.05223906e-01
-1.67233825e+00 -1.37288201e+00 -9.39548016e-01 2.13389337e-01
2.17889115e-01 1.79323286e-01 7.18732297e-01 1.82845160e-01
6.09202161e-02 -9.23796594e-02 -8.11517164e-02 2.50954330e-01
6.46046937e-01 -1.16457200e+00 1.05736327e+00 8.48137319e-01
1.43834561e-01 4.25567329e-01 6.28833652e-01 -5.57141721e-01
-1.33748221e+00 -1.11749315e+00 8.09872329e-01 -9.38764870e-01
5.14794707e-01 -5.82086802e-01 -7.00100243e-01 7.78476417e-01
-3.93347651e-01 -5.02002910e-02 1.18449032e-01 1.63908720e-01
-6.80642366e-01 -1.53152987e-01 -9.73946691e-01 5.55645168e-01
1.34181857e+00 -9.89233017e-01 -5.81038594e-01 3.41015846e-01
8.86999726e-01 -7.39214540e-01 -4.02235836e-01 3.00690949e-01
6.01081133e-01 -1.20825243e+00 1.29036748e+00 -1.37394667e-01
3.62194180e-01 -5.33936501e-01 -4.26453143e-01 -1.21380067e+00
-2.80357391e-01 -9.45418626e-02 2.84312755e-01 5.86269498e-01
4.34093237e-01 -7.75447011e-01 8.67648780e-01 8.95091712e-01
-1.43999308e-01 -1.87480062e-01 -8.42307568e-01 -6.98538542e-01
-2.96749800e-01 -6.96125925e-01 1.00860965e+00 8.39256704e-01
-4.88158226e-01 2.54727453e-01 -2.86135107e-01 5.94641924e-01
6.02955401e-01 1.08222850e-01 1.08144307e+00 -1.37000751e+00
2.09384069e-01 -4.35992450e-01 -9.30823386e-01 -1.24049222e+00
6.29778020e-03 -6.46154225e-01 2.99176276e-01 -1.85252964e+00
-3.30756038e-01 -7.70600259e-01 1.80160999e-01 8.76928210e-01
2.02862173e-01 5.14643252e-01 -6.55446723e-02 3.25504035e-01
-6.64680362e-01 3.56125772e-01 8.17946494e-01 -1.39425799e-01
9.13053155e-02 -5.07250071e-01 -6.53906941e-01 8.02225530e-01
7.53617585e-01 -2.60076910e-01 -5.05332291e-01 -1.20888495e+00
5.81432045e-01 -2.94725955e-01 8.67307365e-01 -1.28688705e+00
3.45253050e-01 -9.62968022e-02 5.74807763e-01 -8.59905601e-01
5.59360385e-01 -7.27492332e-01 1.93017215e-01 2.98723299e-02
-5.21033518e-02 1.85502052e-01 3.01416457e-01 5.47934711e-01
-1.76874831e-01 1.27628073e-01 3.70909601e-01 -4.78708863e-01
-1.36608779e+00 3.80459785e-01 -4.40893434e-02 -5.77653423e-02
7.64138043e-01 -4.47839111e-01 -5.18297374e-01 -6.21779263e-01
-1.65770993e-01 4.40253705e-01 1.01968217e+00 5.91046989e-01
4.33846146e-01 -1.14630818e+00 -2.93507755e-01 3.79344434e-01
5.22585154e-01 8.62575054e-01 2.97878206e-01 5.52091300e-01
-1.26760972e+00 6.10098958e-01 -4.59530741e-01 -8.79907727e-01
-4.33155507e-01 2.20547736e-01 2.88176060e-01 1.55541271e-01
-8.39583516e-01 6.12188935e-01 3.67220610e-01 -1.01277018e+00
-1.57020658e-01 -4.24687862e-01 3.86396259e-01 -2.84508407e-01
5.46551049e-01 1.29530326e-01 1.54268935e-01 -7.45384634e-01
-3.77745926e-01 7.38076568e-01 3.88859838e-01 -2.14323953e-01
1.38041711e+00 -4.84541237e-01 -4.59315702e-02 3.73249143e-01
1.00151789e+00 -3.06022018e-01 -1.72738636e+00 -1.96417361e-01
-1.39638837e-02 -7.33745754e-01 1.74095750e-01 -8.10372770e-01
-6.88874960e-01 9.62654769e-01 5.60968459e-01 -7.50414282e-02
6.58656776e-01 9.14022848e-02 9.49488699e-01 7.86976993e-01
1.01403761e+00 -1.02181196e+00 -5.15559763e-02 6.00952506e-01
7.73474991e-01 -1.74210179e+00 -1.14949487e-01 -2.19368681e-01
-4.63751495e-01 1.00544322e+00 6.84106231e-01 -1.26741990e-01
5.18457055e-01 1.26932502e-01 1.69524312e-01 -3.29085350e-01
-2.05034047e-01 -2.67458111e-01 2.70574331e-01 9.04393017e-01
-2.91097581e-01 1.86440498e-01 8.36224556e-01 1.38725609e-01
-5.77274323e-01 -1.61496684e-01 3.17874491e-01 9.97973084e-01
-7.58373618e-01 -6.92561269e-01 -2.70763725e-01 3.81445765e-01
4.18525368e-01 -3.42044793e-02 -3.46187145e-01 8.34045768e-01
2.44653746e-01 6.84605181e-01 4.83012587e-01 -5.24713159e-01
3.57894450e-01 -1.66837126e-01 3.71820211e-01 -6.56877995e-01
4.35646236e-01 -6.27031088e-01 2.02018261e-01 -9.91460741e-01
-1.98396459e-01 -4.94144768e-01 -1.14130914e+00 -3.16762358e-01
3.00354928e-01 -3.41971070e-01 1.05071890e+00 1.14622045e+00
7.85053790e-01 4.50388342e-02 2.39428461e-01 -1.29163206e+00
-2.25969553e-02 -6.52093112e-01 -2.16746435e-01 4.05157089e-01
4.87606645e-01 -9.21994388e-01 -9.13893729e-02 3.54587547e-02] | [7.720203399658203, -2.0476114749908447] |
30811b6b-67bb-41a6-a976-44451fe986db | personalized-entity-resolution-with-dynamic-1 | null | null | https://aclanthology.org/2021.ecnlp-1.6 | https://aclanthology.org/2021.ecnlp-1.6.pdf | Personalized Entity Resolution with Dynamic Heterogeneous KnowledgeGraph Representations | The growing popularity of Virtual Assistants poses new challenges for Entity Resolution, the task of linking mentions in text to their referent entities in a knowledge base. Specifically, in the shopping domain, customers tend to mention the entities implicitly (e.g., “organic milk”) rather than use the entity names explicitly, leading to a large number of candidate products. Meanwhile, for the same query, different customers may expect different results. For example, with “add milk to my cart”, a customer may refer to a certain product from his/her favorite brand, while some customers may want to re-order products they regularly purchase. Moreover, new customers may lack persistent shopping history, which requires us to enrich the connections between customers through products and their attributes. To address these issues, we propose a new framework that leverages personalized features to improve the accuracy of product ranking. We first build a cross-source heterogeneous knowledge graph from customer purchase history and product knowledge graph to jointly learn customer and product embeddings. After that, we incorporate product, customer, and history representations into a neural reranking model to predict which candidate is most likely to be purchased by a specific customer. Experiment results show that our model substantially improves the accuracy of the top ranked candidates by 24.6% compared to the state-of-the-art product search model. | ['Premkumar Natarajan', 'Yang Liu', 'Heng Ji', 'Yue Liu', 'Tong Wang', 'Jiangning Chen', 'Han Wang', 'Ying Lin'] | null | null | null | null | acl-ecnlp-2021-8 | ['entity-resolution'] | ['natural-language-processing'] | [-3.48630905e-01 1.89518303e-01 -7.52082705e-01 -6.13721669e-01
-4.22311246e-01 -6.62389398e-01 2.97551244e-01 7.32584715e-01
-4.15504605e-01 5.62497437e-01 4.47005242e-01 4.36480530e-02
-1.54363409e-01 -1.17958176e+00 -7.13390768e-01 -1.92058235e-02
7.03412816e-02 9.16576922e-01 1.85394678e-02 -3.50847363e-01
-2.39521295e-01 1.52649134e-01 -1.31397510e+00 1.49757758e-01
9.88937676e-01 8.86223733e-01 3.59314919e-01 -1.75167710e-01
-5.01896083e-01 5.04078865e-01 -3.09441358e-01 -1.04451811e+00
8.20521116e-02 2.69756675e-01 -7.44738996e-01 4.22048033e-04
4.23923045e-01 -4.20869678e-01 -5.12473166e-01 1.14810932e+00
3.98225486e-02 2.73740619e-01 4.84919339e-01 -1.12950099e+00
-1.35798061e+00 1.25418437e+00 -4.41930652e-01 -9.10179690e-02
4.35366124e-01 -4.46161538e-01 1.80118430e+00 -1.07260966e+00
8.02764773e-01 1.22726071e+00 3.05905730e-01 9.21100304e-02
-1.34831274e+00 -7.62764931e-01 6.96454227e-01 2.93543041e-01
-1.48764157e+00 -4.93444838e-02 6.77541316e-01 -2.47843593e-01
8.32691014e-01 1.79586694e-01 6.41954780e-01 9.33543086e-01
-4.95895445e-01 1.01109171e+00 1.48234442e-01 1.49874073e-02
-1.07350774e-01 7.56569266e-01 5.69275379e-01 4.81832385e-01
7.64514506e-01 -1.90951139e-01 -5.29766142e-01 -1.71519712e-01
5.05509198e-01 3.94542336e-01 -1.39643908e-01 -6.23711169e-01
-9.47724819e-01 1.01369584e+00 6.84033215e-01 2.76469529e-01
-5.67749381e-01 -2.37912804e-01 2.68485218e-01 -1.12311915e-02
2.61164278e-01 8.83714259e-01 -8.48885894e-01 4.44597960e-01
-4.73298281e-01 5.86117268e-01 9.74376976e-01 1.39162457e+00
9.87867892e-01 -6.21531367e-01 2.50788536e-02 9.39260066e-01
6.02828383e-01 1.18866920e-01 4.12984669e-01 -5.12947083e-01
5.57164490e-01 9.54698741e-01 4.44251329e-01 -1.21243203e+00
-3.77592206e-01 -8.38521302e-01 -5.08552194e-01 -6.75779819e-01
6.03141002e-02 1.43954039e-01 -7.19323575e-01 1.43045175e+00
3.51256371e-01 1.85893089e-01 1.12914570e-01 1.04530072e+00
1.08131123e+00 5.84136844e-01 3.06813836e-01 -5.37156649e-02
1.55506420e+00 -1.11853516e+00 -6.79639220e-01 -3.54799539e-01
5.44363439e-01 -5.94445169e-01 7.82933593e-01 8.96877944e-02
-6.17646396e-01 -4.86046374e-01 -9.68216777e-01 -1.35349035e-01
-6.33914828e-01 1.21440262e-01 1.16440344e+00 1.76251486e-01
-2.54857212e-01 7.24390030e-01 -6.09954178e-01 -5.23549736e-01
1.70028567e-01 2.67643511e-01 -1.73156202e-01 -4.11953032e-01
-1.63180840e+00 7.72140324e-01 6.98464513e-01 -1.40979439e-01
-4.42607224e-01 -9.34536934e-01 -1.12987256e+00 3.38452399e-01
6.23624086e-01 -6.80900037e-01 1.24749303e+00 -5.65297067e-01
-8.87206554e-01 5.51065683e-01 -2.94071853e-01 -3.48446637e-01
5.85895292e-02 -5.57642341e-01 -1.10763097e+00 -3.69548619e-01
5.85397065e-01 3.63467276e-01 3.58856440e-01 -1.34505773e+00
-1.19645321e+00 -5.06024897e-01 3.75050575e-01 2.41484791e-01
-3.76668185e-01 -3.40342581e-01 -9.67873394e-01 -3.47271532e-01
4.51292455e-01 -8.08585465e-01 -1.45046696e-01 -4.32377577e-01
-6.24716341e-01 -6.24180973e-01 3.93542022e-01 -3.67997348e-01
1.26713192e+00 -2.16898561e+00 -3.30108963e-02 3.13613266e-01
4.34601694e-01 -5.35715967e-02 -1.24505907e-01 4.02843773e-01
1.50880441e-01 1.64290845e-01 3.61769050e-01 -1.93099812e-01
2.18164250e-01 2.98477292e-01 -4.05819148e-01 -5.66847771e-02
9.86540690e-02 1.13720548e+00 -1.19582796e+00 -3.47273111e-01
-1.57470927e-01 4.55011338e-01 -3.61652046e-01 -6.65953085e-02
-2.36711159e-01 -6.00364804e-02 -9.72849846e-01 9.05707717e-01
6.36928856e-01 -6.19224370e-01 5.54511011e-01 -7.21380234e-01
1.63616776e-01 6.81309581e-01 -1.20235872e+00 1.43454099e+00
-4.35621202e-01 -1.53049445e-02 2.03580130e-03 -7.37359822e-01
8.54414165e-01 6.88309371e-02 5.68384767e-01 -5.85524559e-01
-1.53091177e-01 3.68126839e-01 -3.63798767e-01 -2.93808818e-01
1.13459146e+00 2.34147325e-01 -1.62596285e-01 7.09206015e-02
7.01379851e-02 4.23151940e-01 3.73310328e-01 4.51458156e-01
7.90305316e-01 -1.07296392e-01 1.55855179e-01 1.66216314e-01
2.70157576e-01 1.60653293e-01 8.89090240e-01 7.24888504e-01
1.07716881e-01 1.64067373e-01 1.56606898e-01 -4.05101031e-01
-7.53523946e-01 -1.14632833e+00 -1.36887908e-01 1.45674992e+00
6.85715437e-01 -5.57854176e-01 1.08409896e-01 -8.22825313e-01
8.85187864e-01 8.71399522e-01 -5.09993553e-01 -1.98834822e-01
-4.26286966e-01 -5.34333050e-01 -1.17335841e-01 5.50288916e-01
1.58424944e-01 -7.92690158e-01 2.41930008e-01 6.20825112e-01
-2.52515197e-01 -1.05157089e+00 -7.49639809e-01 1.33793712e-01
-5.14179289e-01 -1.05622542e+00 -4.65606958e-01 -9.64037776e-01
5.54618597e-01 3.80471706e-01 1.41655731e+00 -2.47761095e-03
8.75577927e-02 1.65369257e-01 -5.58654308e-01 -9.44928452e-02
-4.14098725e-02 5.42677224e-01 3.90862644e-01 1.02619663e-01
9.64146435e-01 -4.57733840e-01 -6.37489200e-01 3.75994205e-01
-6.28972292e-01 -3.68183017e-01 5.94731390e-01 8.04944575e-01
5.90123057e-01 2.18478084e-01 6.54244661e-01 -1.30960464e+00
5.98686874e-01 -8.28137159e-01 -5.14818668e-01 4.30440247e-01
-1.20948195e+00 1.04106247e-01 5.02009690e-01 -6.06162727e-01
-7.78379440e-01 2.32192189e-01 1.39639169e-01 -3.04968327e-01
2.82549828e-01 1.02189708e+00 -3.35449338e-01 5.38908005e-01
3.85131985e-01 1.51142357e-02 -6.12195790e-01 -8.01933289e-01
6.39532983e-01 4.87343580e-01 6.20507598e-01 -4.25388098e-01
7.45309472e-01 -3.81964222e-02 -5.02095103e-01 -1.93173587e-01
-1.20590377e+00 -8.73742402e-01 -3.47237617e-01 4.02694851e-01
5.65526187e-01 -1.30574334e+00 -8.32637787e-01 -9.32963341e-02
-1.01923394e+00 4.85694200e-01 -2.45713264e-01 7.04591274e-01
2.20043555e-01 1.43448025e-01 -7.50934958e-01 -4.22326058e-01
-2.61847734e-01 -9.32557762e-01 8.92800152e-01 4.28759694e-01
-3.65526050e-01 -6.72954261e-01 -2.41361991e-01 4.88295734e-01
2.97680020e-01 -3.03906947e-01 1.17821956e+00 -1.15887904e+00
-8.45510542e-01 -4.31032896e-01 -3.49926174e-01 -7.82807693e-02
4.54195023e-01 -2.61345834e-01 -4.17113870e-01 -2.84755975e-01
-6.51532888e-01 3.11384760e-02 1.08312941e+00 7.75737539e-02
5.79872131e-01 -3.76981854e-01 -8.96311641e-01 3.56177598e-01
1.22231019e+00 9.08594131e-02 1.34969819e-02 4.73173827e-01
9.69297945e-01 6.68259382e-01 8.93290222e-01 5.32952189e-01
9.51275945e-01 8.51113260e-01 2.96343774e-01 2.54515976e-01
3.70708793e-01 -9.28146720e-01 5.66858239e-03 5.98669231e-01
4.18562770e-01 -3.12372863e-01 -3.39282125e-01 7.87241638e-01
-1.88601589e+00 -6.87322021e-01 3.52397412e-02 2.24864149e+00
9.06745732e-01 1.25584394e-01 -6.65212981e-03 -4.45630968e-01
8.19451988e-01 4.63417210e-02 -9.66460168e-01 2.68342644e-01
-9.12275240e-02 -1.14144638e-01 8.70410621e-01 4.24043983e-01
-1.24861991e+00 1.22103059e+00 5.11411524e+00 5.37325263e-01
-7.07964182e-01 -5.99140413e-02 3.09281290e-01 -7.95903280e-02
-8.07557762e-01 2.23207269e-02 -1.47193921e+00 4.46767122e-01
4.49531764e-01 -6.15772843e-01 3.21268886e-01 1.18043661e+00
-3.79849136e-01 3.67443204e-01 -1.34613264e+00 8.25421810e-01
-1.26009434e-01 -1.26105917e+00 2.32776240e-01 1.18906654e-01
5.07834733e-01 -3.20289060e-02 -3.60198249e-03 8.79101217e-01
7.92327762e-01 -7.98526287e-01 5.69662452e-01 3.00957382e-01
3.04074913e-01 -8.54964137e-01 7.07564473e-01 7.19923675e-02
-1.55032647e+00 -2.38607213e-01 -5.60581088e-01 3.87549758e-01
4.16488796e-01 6.65103912e-01 -9.97696757e-01 6.24549806e-01
7.89826632e-01 9.02853012e-01 -3.24221879e-01 7.69143879e-01
-1.61430955e-01 2.40114346e-01 -2.64590710e-01 -1.79512039e-01
1.47640053e-03 -1.95122838e-01 3.14763248e-01 7.81405807e-01
4.22390014e-01 1.45381033e-01 4.27911907e-01 1.03339076e+00
-6.41098619e-01 4.70103949e-01 -3.76450211e-01 -5.07585049e-01
9.15932655e-01 1.44955409e+00 -5.36004566e-02 -4.02501822e-01
-7.02727616e-01 1.19041836e+00 5.74956298e-01 4.01821584e-01
-5.68550169e-01 -3.48123431e-01 1.17096937e+00 1.93045765e-01
5.90633929e-01 5.43036424e-02 2.72592008e-01 -1.17467189e+00
1.91252053e-01 -3.91351193e-01 5.40665567e-01 -4.44562644e-01
-1.73922765e+00 4.50672150e-01 -4.10583407e-01 -9.36943829e-01
-1.63265496e-01 -2.59836882e-01 -1.02438211e-01 8.18310857e-01
-1.63355815e+00 -1.14665651e+00 -7.18315467e-02 1.67105198e-01
3.75265241e-01 -1.79707706e-01 7.09625840e-01 7.34155536e-01
-5.18118024e-01 8.69985998e-01 1.22674152e-01 3.52980614e-01
8.12312901e-01 -1.05197179e+00 5.69859326e-01 2.24221841e-01
4.37870920e-01 1.13095117e+00 6.08126044e-01 -9.24194694e-01
-1.56621706e+00 -1.36683702e+00 1.45842564e+00 -5.28774559e-01
6.87630355e-01 -1.79302260e-01 -1.09749997e+00 9.24106836e-01
-2.37699986e-01 -2.80268025e-02 7.35248625e-01 7.90858567e-01
-7.10832238e-01 -2.78625280e-01 -9.60736871e-01 5.19121349e-01
1.01297188e+00 -4.82599050e-01 -6.90276623e-01 4.18677419e-01
1.25299180e+00 -2.15681046e-01 -1.11161983e+00 4.23553914e-01
5.68994522e-01 -3.54428858e-01 1.17213416e+00 -8.29970598e-01
1.12490989e-01 -4.09803420e-01 -2.50739843e-01 -1.35359073e+00
-7.72201955e-01 -8.93627629e-02 -5.13668716e-01 1.58114851e+00
1.03625214e+00 -5.79924643e-01 9.29837525e-01 1.18333280e+00
7.25406334e-02 -6.74037516e-01 -5.37840247e-01 -7.43378162e-01
-3.73214841e-01 -9.78708640e-02 1.18219483e+00 1.16602302e+00
1.48915857e-01 7.58712113e-01 -3.73695135e-01 5.02307951e-01
3.14691782e-01 5.54612398e-01 4.61852074e-01 -1.73056293e+00
-3.12616438e-01 -3.73463750e-01 -2.62663215e-01 -1.52499282e+00
1.47881255e-01 -1.22028828e+00 -1.76950976e-01 -1.72767127e+00
2.90892094e-01 -9.05579388e-01 -7.90630639e-01 4.88196641e-01
-3.51251245e-01 -1.92432359e-01 1.86483860e-02 2.40739420e-01
-8.01035941e-01 4.59712505e-01 1.03269851e+00 -4.52288449e-01
-4.86984968e-01 1.97203562e-01 -1.29770780e+00 3.20145100e-01
3.70586634e-01 -4.82378662e-01 -2.31796011e-01 -5.14414787e-01
7.47836530e-01 -6.53514341e-02 -3.88074331e-02 -3.00984263e-01
3.59857619e-01 -2.02478506e-02 3.01569104e-01 -6.85065269e-01
4.43350613e-01 -1.07699978e+00 2.22695187e-01 -3.75721529e-02
-4.92732853e-01 -3.81739698e-02 -1.84174269e-01 9.18726921e-01
-2.98338234e-01 -2.37387910e-01 1.21391453e-01 -1.19738579e-01
-9.59890485e-01 5.58732331e-01 1.39917761e-01 -2.33706877e-01
6.62458360e-01 2.85201013e-01 -2.84151912e-01 -2.05479175e-01
-8.64300847e-01 6.84256613e-01 4.20865536e-01 1.09676135e+00
3.73564214e-01 -1.59212065e+00 -6.01277530e-01 7.30248913e-02
6.31344914e-01 1.13803610e-01 7.19020739e-02 4.83009309e-01
3.05248946e-01 5.38369775e-01 3.44678193e-01 -2.40741462e-01
-8.38563800e-01 8.38054597e-01 -5.39610796e-02 -3.26648921e-01
-4.71852511e-01 1.04930830e+00 9.92366001e-02 -5.94007790e-01
2.96102762e-01 -2.31201112e-01 -5.58308482e-01 3.82195979e-01
4.43453580e-01 7.91809633e-02 5.19800968e-02 -7.10050941e-01
-4.42173660e-01 1.17480017e-01 -9.14713979e-01 2.77119875e-01
1.24772823e+00 -2.11273357e-01 -3.75908725e-02 2.65557438e-01
1.30830228e+00 1.84951693e-01 -8.16952884e-01 -9.00082529e-01
2.76732117e-01 -5.30356467e-01 -2.31452975e-02 -8.07027340e-01
-1.39897740e+00 -2.22131703e-03 3.81756485e-01 2.65399039e-01
6.27204418e-01 5.33498704e-01 1.11873412e+00 6.69970930e-01
6.33532405e-01 -1.03434491e+00 -4.17830348e-01 4.43387717e-01
4.93469685e-01 -1.37538505e+00 9.92788672e-02 -7.13304937e-01
-9.15084541e-01 6.45570993e-01 7.42000818e-01 2.01137006e-01
7.31236517e-01 -2.95898199e-01 -4.79579978e-02 -3.61530632e-01
-7.12802708e-01 -3.51465195e-01 4.47117120e-01 3.12471837e-01
5.86992621e-01 4.06477034e-01 -2.72220373e-01 1.19396317e+00
-1.90244451e-01 -1.91523015e-01 1.19114079e-01 5.66420615e-01
-3.60429794e-01 -1.23971760e+00 3.07588577e-01 8.16710532e-01
-3.27964038e-01 -3.32500488e-01 -2.65545905e-01 4.91581351e-01
1.00795366e-01 9.22325730e-01 7.93729946e-02 -6.46520078e-01
7.31355965e-01 1.27510473e-01 -7.88873818e-04 -8.25576484e-01
-4.57903564e-01 -2.08575979e-01 2.21607387e-01 -3.86772007e-01
-3.44751514e-02 -7.03265190e-01 -1.38523018e+00 -2.83365786e-01
-6.85051799e-01 3.99187416e-01 5.05460083e-01 6.44743025e-01
7.86967695e-01 3.54290694e-01 5.86262763e-01 -2.34194160e-01
-4.25825596e-01 -8.16100419e-01 -9.93160486e-01 6.17137074e-01
1.51657939e-01 -9.40459669e-01 -9.21813473e-02 -3.21271956e-01] | [10.068492889404297, 6.163257122039795] |
80042760-c24c-4e03-864c-54fd5b294fe7 | unveiling-transformers-with-lego-a-synthetic | 2206.04301 | null | https://arxiv.org/abs/2206.04301v3 | https://arxiv.org/pdf/2206.04301v3.pdf | Unveiling Transformers with LEGO: a synthetic reasoning task | We propose a synthetic reasoning task, LEGO (Learning Equality and Group Operations), that encapsulates the problem of following a chain of reasoning, and we study how the Transformer architectures learn this task. We pay special attention to data effects such as pretraining (on seemingly unrelated NLP tasks) and dataset composition (e.g., differing chain length at training and test time), as well as architectural variants such as weight-tied layers or adding convolutional components. We study how the trained models eventually succeed at the task, and in particular, we manage to understand some of the attention heads as well as how the information flows in the network. In particular, we have identified a novel \emph{association} pattern that globally attends only to identical tokens. Based on these observations we propose a hypothesis that here pretraining helps for LEGO tasks due to certain structured attention patterns, and we experimentally verify this hypothesis. We also observe that in some data regime the trained transformer finds ``shortcut" solutions to follow the chain of reasoning, which impedes the model's robustness, and moreover we propose ways to prevent it. Motivated by our findings on structured attention patterns, we propose the LEGO attention module, a drop-in replacement for vanilla attention heads. This architectural change significantly reduces Flops and maintains or even \emph{improves} the model's performance at large-scale pretraining. | ['Tal Wagner', 'Suriya Gunasekar', 'Ronen Eldan', 'Sébastien Bubeck', 'Arturs Backurs', 'Yi Zhang'] | 2022-06-09 | null | null | null | null | ['learning-to-execute'] | ['computer-code'] | [ 1.80820882e-01 6.93687260e-01 9.86850560e-02 -2.78461546e-01
-4.28122282e-01 -8.34060311e-01 6.52992964e-01 1.50863037e-01
-3.60061526e-01 4.19397086e-01 4.57462609e-01 -6.31916761e-01
-3.61945540e-01 -5.87146223e-01 -1.08550549e+00 -5.69262981e-01
7.18525276e-02 7.08627343e-01 3.58689278e-01 -4.69826132e-01
2.64298141e-01 4.75867838e-01 -1.48134363e+00 5.48383772e-01
8.29539239e-01 7.82894313e-01 6.87438250e-02 6.10128641e-01
-1.75082311e-01 1.16442144e+00 -6.52727544e-01 -9.51034367e-01
2.97339827e-01 -3.44310492e-01 -1.20302081e+00 -1.49669290e-01
6.26672149e-01 -2.29178444e-01 -1.28165290e-01 9.00388896e-01
2.47339025e-01 -4.44948561e-02 6.27130747e-01 -1.30113804e+00
-9.48906481e-01 1.25970244e+00 -2.78006881e-01 3.35220873e-01
1.40778989e-01 3.33229989e-01 1.51466691e+00 -8.84039283e-01
6.85549796e-01 1.03816950e+00 9.83635068e-01 5.81756830e-01
-1.26281512e+00 -3.11683625e-01 3.70098561e-01 2.21451625e-01
-1.09302425e+00 -3.88385862e-01 5.80484688e-01 -3.26307654e-01
1.08675027e+00 2.51691222e-01 4.05291140e-01 1.08129203e+00
2.07998142e-01 1.07773399e+00 8.68076384e-01 -4.72140014e-01
9.77141112e-02 1.45958185e-01 4.10203218e-01 6.86324835e-01
3.11181188e-01 -2.83138275e-01 -4.17339802e-01 2.31690198e-01
3.27993125e-01 -2.32505128e-01 -2.61715770e-01 -3.72307032e-01
-1.04815853e+00 6.85387671e-01 5.53700209e-01 5.22285998e-01
-6.58704489e-02 2.24696353e-01 3.11214298e-01 5.56656897e-01
8.56469013e-03 8.31051290e-01 -9.50562835e-01 1.96345627e-01
-7.03328252e-01 3.40123504e-01 9.03272152e-01 9.85648155e-01
7.83137143e-01 -2.83805877e-01 -4.15398628e-01 6.76268101e-01
-3.80548611e-02 2.61719488e-02 4.50633466e-01 -9.28182065e-01
7.33354986e-01 6.66199625e-01 -1.50075331e-01 -6.25167668e-01
-6.21654093e-01 -6.20295167e-01 -6.33853734e-01 9.08580944e-02
8.10798228e-01 -1.37462676e-01 -7.95216620e-01 2.13024640e+00
-2.39061266e-01 -1.89504460e-01 1.29827246e-01 6.65821731e-01
6.36186659e-01 1.62400275e-01 -8.11169669e-02 5.09225838e-02
1.57443309e+00 -1.12598515e+00 -5.81668258e-01 -4.51379120e-01
9.05143559e-01 -5.91852903e-01 1.54680860e+00 3.22792143e-01
-1.39654863e+00 -5.38838685e-01 -8.95252287e-01 -4.65571880e-01
-5.99891722e-01 -1.21800087e-01 6.11098886e-01 3.60481709e-01
-1.12733877e+00 1.07848132e+00 -5.37017465e-01 -4.64449912e-01
5.61381161e-01 4.26603854e-01 -1.40502900e-01 -4.11232524e-02
-1.23484492e+00 9.85277057e-01 2.94548541e-01 3.08053762e-01
-5.27384043e-01 -8.97291303e-01 -6.51636839e-01 4.99729574e-01
4.40471739e-01 -8.20771039e-01 1.35240996e+00 -1.07256818e+00
-1.21363831e+00 9.73102868e-01 -9.15813223e-02 -5.46721935e-01
6.95194900e-01 -2.71100491e-01 -1.13367237e-01 -1.49856284e-01
1.07801050e-01 6.91900611e-01 6.30088985e-01 -1.05358207e+00
-4.23622787e-01 -1.57980233e-01 4.14126456e-01 -1.27770379e-01
-2.16457501e-01 8.66370723e-02 -2.61628687e-01 -7.26213276e-01
5.38614094e-02 -9.70912337e-01 6.22283146e-02 -1.20294549e-01
-6.42326653e-01 -3.60245615e-01 5.50267637e-01 -4.18850511e-01
1.00449932e+00 -2.10097790e+00 2.62190223e-01 1.60519496e-01
2.71727860e-01 2.14061990e-01 -3.11199605e-01 5.38601756e-01
-3.29921961e-01 5.50092995e-01 -2.83140898e-01 -5.40701628e-01
2.90150642e-01 5.52860320e-01 -4.94073659e-01 1.10510558e-01
5.63643396e-01 1.10970902e+00 -6.56189978e-01 -2.32026160e-01
-3.71484816e-01 1.69898003e-01 -1.02017415e+00 1.91749483e-01
-5.30574977e-01 3.73601735e-01 -3.46474089e-02 3.12288970e-01
5.98417044e-01 -4.41098928e-01 3.85209680e-01 -2.73134798e-01
-1.31830290e-01 7.86619067e-01 -9.87594843e-01 1.45981276e+00
-5.02536058e-01 7.27600753e-01 -1.02396458e-01 -1.04971647e+00
5.30152917e-01 8.65854472e-02 -1.39521688e-01 -6.97405159e-01
1.97095290e-01 2.16165379e-01 5.29193997e-01 -5.88754714e-01
3.69253337e-01 -7.05500990e-02 3.11852507e-02 5.13087511e-01
2.27396116e-01 1.39402673e-01 3.86720181e-01 3.03135782e-01
1.42355502e+00 5.19140624e-02 8.16242844e-02 -5.18501222e-01
2.98384279e-01 -2.73265261e-02 4.77404147e-01 1.11141717e+00
1.27158090e-01 7.25423753e-01 1.13130927e+00 -4.11362618e-01
-9.83686566e-01 -8.99826109e-01 2.52775960e-02 1.40296972e+00
-1.22809093e-02 -5.05982041e-01 -6.74385905e-01 -9.17086840e-01
-6.52442873e-02 8.07443082e-01 -9.73148465e-01 -2.69173980e-01
-9.30910110e-01 -7.28714466e-01 6.61967933e-01 7.72803605e-01
4.18813318e-01 -1.42095459e+00 -5.61825395e-01 1.54403485e-02
-1.52895576e-03 -1.12505794e+00 -2.28801906e-01 7.41748273e-01
-6.36176944e-01 -1.15002680e+00 -3.44726562e-01 -6.85362399e-01
5.95934570e-01 -1.18307412e-01 1.32513022e+00 4.80858684e-01
1.84381723e-01 -4.66971546e-02 -3.90375465e-01 -2.48754084e-01
-4.41157192e-01 5.00260711e-01 -3.31389636e-01 -6.16663620e-02
3.05064442e-03 -7.53900588e-01 -4.03218210e-01 2.21324161e-01
-8.78812551e-01 5.56838736e-02 7.67060816e-01 7.48660564e-01
1.08800858e-01 1.49265885e-01 1.87521517e-01 -1.30063164e+00
6.68009937e-01 -3.33545804e-01 -4.45624650e-01 4.44858253e-01
-5.62879145e-01 6.92856550e-01 9.73810673e-01 -3.87148947e-01
-8.73678207e-01 -4.09542143e-01 -3.01750749e-01 -4.43142384e-01
5.70848361e-02 4.04304117e-01 -3.80053133e-01 2.36005098e-01
5.66938877e-01 -6.36994392e-02 -1.97573245e-01 -5.64866543e-01
4.14797544e-01 6.28156662e-02 4.75514293e-01 -8.18314135e-01
9.02603924e-01 3.30155820e-01 -2.30662644e-01 -3.13609302e-01
-1.18023372e+00 1.66151375e-01 -6.34189904e-01 2.45828435e-01
9.36043918e-01 -6.09692156e-01 -1.03296661e+00 2.52767086e-01
-1.40991628e+00 -8.36005688e-01 -5.26618719e-01 -2.17072200e-02
-4.57914025e-01 -7.54299387e-02 -8.61641884e-01 -4.36439067e-01
-5.16303554e-02 -1.23842835e+00 8.34670663e-01 4.62460965e-02
-5.49127877e-01 -1.06818867e+00 -4.36556302e-02 2.11153209e-01
4.87784952e-01 -1.70813933e-01 1.48410857e+00 -1.10801935e+00
-7.88357496e-01 4.86861914e-01 -4.29043025e-01 2.75088847e-01
-2.84590065e-01 4.50372398e-02 -1.15009248e+00 -7.07153305e-02
-2.85303183e-02 -2.34852657e-01 1.06107843e+00 -2.44289357e-02
1.18888533e+00 -5.08673429e-01 -8.09403509e-02 6.57812476e-01
1.23573828e+00 -1.00637451e-01 7.74149001e-01 4.14189160e-01
7.63595343e-01 8.35394740e-01 2.15100907e-02 -7.13334754e-02
2.63957143e-01 6.79223299e-01 4.95066106e-01 -4.52200770e-02
-2.63097644e-01 -3.91466409e-01 3.27475041e-01 6.77184761e-01
1.05239585e-01 -4.12017941e-01 -1.00915015e+00 4.16223347e-01
-1.84397829e+00 -9.15765166e-01 -1.35933921e-01 1.85728586e+00
7.64497161e-01 5.73412955e-01 -2.72954330e-02 4.28183734e-01
4.00271982e-01 1.57451749e-01 -2.03005612e-01 -7.91405797e-01
-1.70357943e-01 5.78974485e-01 5.39342642e-01 5.74107707e-01
-6.61834836e-01 8.98243368e-01 6.16588402e+00 5.52527130e-01
-1.13815296e+00 1.55537561e-01 3.17652464e-01 1.23575069e-01
-6.89575195e-01 7.35211223e-02 -7.91645169e-01 3.19378942e-01
6.78609729e-01 1.95421740e-01 5.75255334e-01 5.31768680e-01
-2.09696159e-01 1.55965731e-01 -1.54550016e+00 3.61126244e-01
9.02003571e-02 -1.39272404e+00 1.73145846e-01 -5.13794497e-02
4.14467216e-01 1.51117388e-02 9.28304717e-02 5.71320355e-01
6.63579226e-01 -9.15272415e-01 1.18056381e+00 1.72661752e-01
2.81617433e-01 -4.72591668e-01 7.34996200e-01 2.46476129e-01
-9.68149364e-01 -3.74289274e-01 -6.17053919e-02 -1.79205373e-01
-1.06689654e-01 5.20419061e-01 -8.48060966e-01 3.98551404e-01
6.74466252e-01 3.94165158e-01 -9.86067712e-01 7.28675067e-01
-7.98621893e-01 5.31214714e-01 -1.70458585e-01 7.10359365e-02
2.98109502e-01 1.50502503e-01 3.32859010e-01 1.24016547e+00
1.00770898e-01 -2.05543742e-01 -3.64618450e-01 1.13217115e+00
-3.13233048e-01 -2.57122487e-01 -4.56667274e-01 1.30081221e-01
3.07659388e-01 9.15063500e-01 -8.05792511e-01 -2.20165133e-01
-3.71348768e-01 9.49637651e-01 8.57650697e-01 3.38871151e-01
-1.08133733e+00 -1.92588046e-01 5.45549929e-01 2.43410930e-01
6.98152483e-01 1.79304972e-01 -2.68175781e-01 -1.15475965e+00
1.59951448e-01 -9.31277514e-01 4.09425914e-01 -1.05897272e+00
-1.18665111e+00 5.13410091e-01 -1.78428993e-01 -6.52241707e-01
6.35828972e-02 -7.08712220e-01 -9.69605267e-01 6.55289590e-01
-1.66085029e+00 -1.03575528e+00 -4.59796265e-02 4.95461345e-01
4.02817369e-01 1.97350383e-01 4.51492637e-01 3.34656447e-01
-8.34036708e-01 8.81623030e-01 -2.21367031e-01 3.64174545e-01
5.90697944e-01 -1.43457365e+00 5.91836095e-01 8.62286329e-01
3.30087185e-01 9.90850508e-01 8.80401850e-01 -2.44681939e-01
-1.16177547e+00 -1.02124321e+00 1.32667232e+00 -7.36215770e-01
9.59804595e-01 -6.43814147e-01 -1.10424495e+00 1.17230618e+00
3.22945595e-01 5.26409298e-02 3.77232552e-01 5.52226067e-01
-7.52153039e-01 -1.04390621e-01 -9.35385704e-01 7.89839387e-01
1.50115085e+00 -5.68749785e-01 -1.02096689e+00 3.77366781e-01
9.55217421e-01 -3.20649415e-01 -4.39782828e-01 3.45427960e-01
4.90285009e-01 -1.26373148e+00 7.44403780e-01 -9.48884785e-01
8.45954418e-01 -2.95578204e-02 -3.89972478e-02 -1.06605053e+00
-5.01872182e-01 -6.64101601e-01 4.05152887e-02 1.51724720e+00
8.46541047e-01 -6.83501244e-01 8.01591396e-01 5.56282818e-01
-3.37349564e-01 -8.62671435e-01 -6.61175311e-01 -7.68723786e-01
3.69657874e-01 -3.53556395e-01 5.94045997e-01 9.48451221e-01
-1.46638110e-01 7.21306503e-01 -1.24738753e-01 3.24934870e-01
1.39476597e-01 1.89931154e-01 6.06091619e-01 -1.18542838e+00
-5.75121939e-01 -6.67600155e-01 1.66746169e-01 -1.03678596e+00
3.26575875e-01 -1.13461959e+00 -7.04021603e-02 -1.29222047e+00
-4.64898571e-02 -5.50927281e-01 -2.41758540e-01 7.62785137e-01
-2.43116364e-01 8.97069797e-02 4.55805421e-01 1.96192294e-01
-6.39539421e-01 1.31781280e-01 1.05971849e+00 -1.16253637e-01
-4.26487140e-02 -6.70170560e-02 -9.97484267e-01 7.09714651e-01
6.80154741e-01 -6.64541423e-01 -3.54425669e-01 -8.67277622e-01
8.23188305e-01 -3.16347688e-01 4.92003977e-01 -8.53067577e-01
2.46377245e-01 1.47242323e-01 7.26822689e-02 -2.10715771e-01
-8.44848529e-02 -9.51286554e-01 -3.17276902e-02 5.83804667e-01
-6.37420714e-01 3.36680174e-01 2.73922265e-01 2.03255825e-02
1.32132610e-02 -4.28538531e-01 7.41618097e-01 -1.68814883e-01
-4.14329410e-01 -1.11907467e-01 -3.45973969e-01 4.43807185e-01
5.30823827e-01 2.86525600e-02 -6.45697236e-01 -4.97854315e-03
-8.68810415e-01 3.89213890e-01 3.53997558e-01 3.00402105e-01
-1.02855481e-01 -1.12165082e+00 -5.07841349e-01 3.77844810e-01
4.52397838e-02 1.18744068e-01 -3.69967078e-03 1.00427115e+00
-3.85888487e-01 3.32099348e-01 -1.06664598e-01 -3.43388319e-01
-9.21673894e-01 7.38907218e-01 4.35978979e-01 -6.46418333e-01
-4.23797995e-01 9.79448259e-01 2.75324464e-01 -5.75396061e-01
1.88507706e-01 -8.92775476e-01 -5.25267124e-02 3.11509877e-01
1.80056736e-01 -3.46142016e-02 3.39738250e-01 -7.44382739e-02
-3.60597938e-01 5.53791583e-01 -2.52624780e-01 1.62694141e-01
1.36477923e+00 8.19039941e-02 -1.92054473e-02 3.25473398e-01
1.11776769e+00 3.31025392e-01 -1.07394481e+00 -1.55535161e-01
1.35227218e-01 3.79291289e-02 -5.82888067e-01 -8.10153544e-01
-1.13761628e+00 1.05419385e+00 -9.86144245e-02 5.53009868e-01
9.33210611e-01 2.41434619e-01 4.91432369e-01 5.02656579e-01
2.75439583e-02 -7.20809400e-01 1.62845805e-01 7.69712985e-01
9.79550779e-01 -9.44535315e-01 -1.17368542e-01 -3.69812876e-01
-3.64263475e-01 8.96053970e-01 5.24461806e-01 -1.58708677e-01
4.53197658e-01 4.49431419e-01 -1.17550656e-01 -2.91663826e-01
-1.06794333e+00 -2.98042566e-01 -3.27637270e-02 4.25940931e-01
2.32728019e-01 -4.83279794e-01 -7.88014978e-02 6.27718091e-01
-5.73238730e-01 -1.02587715e-01 3.70241731e-01 7.45319247e-01
-1.62621424e-01 -1.22576666e+00 -1.29283369e-01 2.78849274e-01
-5.16543448e-01 -4.13414389e-01 -6.09050751e-01 1.24556971e+00
4.92112219e-01 4.50153619e-01 2.48009965e-01 -2.11048797e-01
5.54016531e-01 3.17188919e-01 4.81654346e-01 -6.20885551e-01
-1.11613810e+00 -4.64114100e-01 1.63502395e-01 -4.18303609e-01
-1.67745382e-01 -5.27707934e-01 -1.17666256e+00 -4.39328402e-01
-1.33762553e-01 1.10777751e-01 3.08832843e-02 1.22049022e+00
3.61263275e-01 9.23257411e-01 2.58504808e-01 -5.95337570e-01
-4.89494443e-01 -8.19676280e-01 -2.72011310e-01 7.10778058e-01
4.27582651e-01 -5.16053081e-01 -6.61940575e-01 -7.54123554e-02] | [9.695064544677734, 7.4947590827941895] |
69470ecc-97c6-4edf-8e7c-b71b4c62340f | evaluating-coherence-in-dialogue-systems | 1904.03371 | null | https://arxiv.org/abs/1904.03371v2 | https://arxiv.org/pdf/1904.03371v2.pdf | Evaluating Coherence in Dialogue Systems using Entailment | Evaluating open-domain dialogue systems is difficult due to the diversity of possible correct answers. Automatic metrics such as BLEU correlate weakly with human annotations, resulting in a significant bias across different models and datasets. Some researchers resort to human judgment experimentation for assessing response quality, which is expensive, time consuming, and not scalable. Moreover, judges tend to evaluate a small number of dialogues, meaning that minor differences in evaluation configuration may lead to dissimilar results. In this paper, we present interpretable metrics for evaluating topic coherence by making use of distributed sentence representations. Furthermore, we introduce calculable approximations of human judgment based on conversational coherence by adopting state-of-the-art entailment techniques. Results show that our metrics can be used as a surrogate for human judgment, making it easy to evaluate dialogue systems on large-scale datasets and allowing an unbiased estimate for the quality of the responses. | ['Osmar Zaiane', 'Nouha Dziri', 'Ehsan Kamalloo', 'Kory W. Mathewson'] | 2019-04-06 | evaluating-coherence-in-dialogue-systems-2 | https://aclanthology.org/N19-1381 | https://aclanthology.org/N19-1381.pdf | naacl-2019-6 | ['dialogue-evaluation', 'open-domain-dialog'] | ['natural-language-processing', 'natural-language-processing'] | [-9.94235054e-02 3.71767253e-01 5.49756251e-02 -6.92432642e-01
-1.12974679e+00 -9.77677941e-01 8.31793070e-01 5.01046896e-01
-6.30243480e-01 9.86669898e-01 8.19732487e-01 -2.69189537e-01
9.71767008e-02 -6.86619043e-01 7.58197382e-02 -2.40971267e-01
3.62695307e-01 7.01812565e-01 2.86401063e-01 -4.17661160e-01
4.73843396e-01 -1.34972036e-01 -1.23384142e+00 4.82775062e-01
9.23918426e-01 5.92340112e-01 -1.87112298e-02 9.34259653e-01
-2.31350705e-01 1.08245087e+00 -1.20258164e+00 -8.09506834e-01
-1.83202848e-01 -8.42129827e-01 -1.32119858e+00 -8.10977146e-02
3.30074847e-01 -4.31192100e-01 3.12460642e-02 8.15496206e-01
5.28237104e-01 2.78404057e-01 8.81228447e-01 -9.41422045e-01
-5.04071414e-01 6.69103742e-01 2.43392941e-02 7.28634000e-02
1.05538595e+00 2.48234451e-01 1.39780343e+00 -4.10374463e-01
4.64203656e-01 1.39740503e+00 7.35370815e-01 5.87385952e-01
-1.39288127e+00 -2.93473434e-02 -1.98046178e-01 3.00938133e-02
-9.21382844e-01 -6.44142330e-01 3.81453872e-01 -4.16500390e-01
8.81334543e-01 5.97560883e-01 3.24617356e-01 9.36445117e-01
-6.63166568e-02 5.30081153e-01 1.35539067e+00 -4.84829962e-01
3.15770417e-01 5.14825344e-01 4.19806629e-01 5.47843456e-01
1.98723987e-01 -5.15854418e-01 -4.32376027e-01 -5.97903609e-01
3.81775409e-01 -5.00680923e-01 -4.04298663e-01 -7.54736271e-03
-1.21843207e+00 1.00421286e+00 1.91282690e-01 2.72101492e-01
-3.18164319e-01 -2.87869841e-01 6.27107084e-01 6.43467903e-01
7.09538579e-01 9.91767228e-01 -3.43538135e-01 -6.86081707e-01
-7.70260632e-01 6.40177906e-01 1.59468162e+00 4.06742424e-01
6.08545363e-01 -5.28234720e-01 -4.83828396e-01 1.12712550e+00
1.77294478e-01 2.71004319e-01 5.87989390e-01 -1.55077791e+00
4.57624525e-01 5.81238329e-01 5.73808312e-01 -1.16140556e+00
-4.27751750e-01 1.35457292e-01 -3.52682143e-01 5.59313409e-02
1.10385323e+00 -3.94506067e-01 -8.96929428e-02 1.72424185e+00
1.24447845e-01 -8.89018834e-01 7.21487477e-02 1.01216769e+00
7.46570587e-01 4.79150563e-01 -1.31546473e-02 -3.84790629e-01
1.44015539e+00 -9.05999660e-01 -7.22091913e-01 -1.16071701e-01
8.30034018e-01 -9.72096562e-01 1.49606633e+00 4.27083671e-01
-1.20707643e+00 -2.81840742e-01 -7.46662557e-01 -2.19784021e-01
1.52688194e-02 -3.30227405e-01 5.56509852e-01 7.38788605e-01
-1.00750566e+00 6.36779010e-01 -5.00493169e-01 -5.62424302e-01
-3.01998973e-01 -1.03264385e-04 -8.17354321e-02 1.27033457e-01
-1.02708209e+00 1.39844871e+00 -1.20746806e-01 -1.72163248e-01
-3.48170638e-01 -1.80153906e-01 -5.40415287e-01 1.05473764e-01
2.84050465e-01 -5.62046707e-01 1.86941195e+00 -8.77893925e-01
-1.74650979e+00 7.71893144e-01 -3.25075388e-01 -1.99192643e-01
6.43439710e-01 -7.15163127e-02 1.03359567e-02 9.65239182e-02
1.27576098e-01 3.53932887e-01 1.78727552e-01 -8.97654414e-01
-2.99370825e-01 -1.85908735e-01 4.35522616e-01 5.58974981e-01
-2.91248441e-01 2.54942626e-01 -6.83718473e-02 -1.12655684e-01
-1.49032176e-01 -7.79104888e-01 -1.92551374e-01 -2.80640215e-01
-1.83576539e-01 -5.66996634e-01 1.11770235e-01 -6.78673446e-01
1.50665712e+00 -1.66176927e+00 1.87100973e-02 -1.02881387e-01
3.10398608e-01 1.34738907e-01 -1.38107792e-01 7.69672930e-01
6.21464133e-01 3.05794388e-01 -9.05129984e-02 -2.46273279e-01
3.48696887e-01 4.98898998e-02 -7.66477957e-02 1.68790281e-01
6.85732113e-03 6.82739615e-01 -1.06570482e+00 -6.43679619e-01
1.00907050e-02 -1.00082099e-01 -5.28393507e-01 4.27237928e-01
-3.06993395e-01 2.66485810e-01 -3.81307870e-01 1.73771039e-01
4.71326187e-02 -3.73477519e-01 3.92130196e-01 9.29318294e-02
-8.60910397e-03 1.02064824e+00 -8.10754836e-01 1.55782330e+00
-6.90692246e-01 9.01509404e-01 4.26992327e-02 -4.76411819e-01
8.00326049e-01 5.08541465e-01 -5.97885549e-02 -6.32286012e-01
2.31578737e-01 9.27048698e-02 1.92380488e-01 -4.54486132e-01
9.07154441e-01 -1.02092423e-01 -2.60077357e-01 1.12947893e+00
-2.16207393e-02 -5.77689469e-01 4.81409371e-01 4.09648508e-01
1.23397303e+00 -3.37949574e-01 4.38778847e-01 -3.27244282e-01
3.28499675e-01 8.24566185e-02 2.69898117e-01 9.06171679e-01
-2.77104825e-01 4.86523449e-01 9.12355900e-01 -4.36277509e-01
-1.14652956e+00 -8.24954808e-01 -1.92931026e-01 1.10734892e+00
-1.38843700e-01 -5.50066650e-01 -1.03397620e+00 -5.67789435e-01
-2.39166602e-01 8.49522173e-01 -3.59989196e-01 -3.22169699e-02
-2.37785473e-01 -6.86515391e-01 7.32596517e-01 2.22734004e-01
2.03521878e-01 -8.66883755e-01 -7.18210340e-01 3.35601330e-01
-8.29048932e-01 -8.50753129e-01 -3.34942013e-01 -7.32214600e-02
-7.48499811e-01 -1.05991864e+00 -7.20227063e-01 -3.00056994e-01
2.60207504e-01 3.00546974e-01 1.70080078e+00 2.03861296e-01
2.21423939e-01 3.24219733e-01 -4.25737679e-01 -1.86934859e-01
-8.64624679e-01 1.43958300e-01 -1.07796704e-02 -3.93231452e-01
5.77463627e-01 -2.05470935e-01 -6.50922775e-01 6.14278674e-01
-4.83056128e-01 -2.17721686e-01 1.87660262e-01 9.37891960e-01
-1.32392421e-01 -5.65836489e-01 6.90323710e-01 -1.13230705e+00
1.60954177e+00 -4.08716679e-01 -1.70404047e-01 3.96497011e-01
-9.78781044e-01 2.03207701e-01 5.33548474e-01 -2.74420083e-01
-1.16865540e+00 -6.68679893e-01 3.56042981e-02 3.27067554e-01
-8.90578553e-02 5.26956499e-01 3.14326525e-01 2.31455043e-01
1.23370147e+00 -3.23060513e-01 2.87227094e-01 -3.04133594e-01
4.77942318e-01 1.11361361e+00 2.32400507e-01 -7.38681853e-01
1.23388134e-01 -5.95404468e-02 -6.09461188e-01 -7.92955160e-01
-1.03794527e+00 -4.16251153e-01 -2.69625545e-01 -2.88008362e-01
6.22346997e-01 -6.87435627e-01 -8.18992615e-01 1.92018971e-01
-1.34857047e+00 -4.50346470e-01 -1.48921654e-01 4.11544710e-01
-5.11011779e-01 6.69396639e-01 -9.73769307e-01 -1.04023147e+00
-3.22147340e-01 -9.41043854e-01 8.97685409e-01 1.86314270e-01
-1.30466211e+00 -1.20706284e+00 4.43123668e-01 7.98444271e-01
5.24052382e-01 -5.96294403e-02 7.55851388e-01 -9.06576216e-01
-1.56890824e-01 -3.47331107e-01 -2.13718370e-01 3.76723558e-01
-1.55639770e-02 8.64503458e-02 -1.02108335e+00 -1.88099854e-02
1.27620101e-01 -9.56366539e-01 4.92637604e-01 1.58371687e-01
4.96250838e-01 -5.67770243e-01 1.95859313e-01 -3.17173570e-01
9.15946543e-01 -2.64685154e-01 5.20692766e-01 2.52244473e-01
1.94180548e-01 1.00567794e+00 6.08678699e-01 5.30859888e-01
7.21076608e-01 6.58299446e-01 -2.78222531e-01 2.90084869e-01
1.23889074e-01 -6.75648302e-02 2.01122090e-01 1.10059369e+00
1.12743527e-02 -4.53551024e-01 -9.20842290e-01 5.93515515e-01
-1.87659407e+00 -9.75810051e-01 -3.50773603e-01 2.26381350e+00
1.22537982e+00 7.28822500e-02 2.98131883e-01 4.78647985e-02
5.60331702e-01 2.41541505e-01 1.03168944e-02 -9.30504203e-01
7.48515427e-02 1.01694003e-01 5.79263568e-02 8.52980018e-01
-4.17264670e-01 5.85525334e-01 7.36329746e+00 4.54216510e-01
-7.83254862e-01 1.02296807e-01 6.05518401e-01 -1.60181612e-01
-4.55379397e-01 1.80056334e-01 -2.94520438e-01 5.41116059e-01
1.34982574e+00 -3.80006909e-01 4.53358471e-01 5.08852839e-01
2.53383428e-01 -5.85651219e-01 -1.20264077e+00 5.79802036e-01
5.28636463e-02 -8.21339071e-01 -2.91283667e-01 -1.49043977e-01
6.72421455e-01 -9.24418867e-02 -4.87322122e-01 3.79385680e-01
6.76036596e-01 -8.20710599e-01 4.43019956e-01 5.02453923e-01
4.26593184e-01 -4.70196396e-01 8.76754224e-01 5.08664787e-01
-4.87726003e-01 1.40359625e-01 -5.39614856e-01 -5.72548985e-01
1.18192077e-01 5.47029138e-01 -1.02066243e+00 1.30147003e-02
3.17270815e-01 -3.27635221e-02 -5.03271818e-01 6.95090532e-01
-2.49742493e-01 5.94755113e-01 -3.91734481e-01 -7.26803422e-01
1.92817688e-01 -1.54491842e-01 2.55440950e-01 1.21814978e+00
-1.74777396e-02 6.25064075e-02 1.94350690e-01 7.71217942e-01
-1.17932044e-01 3.37751746e-01 -6.15278602e-01 -1.13159522e-01
7.38274574e-01 1.30897450e+00 -5.18966436e-01 -3.88208389e-01
-4.77396548e-01 8.85796130e-01 6.03247404e-01 1.13854714e-01
-4.32189345e-01 -4.08634871e-01 3.83574665e-01 -2.02459976e-01
-3.22100192e-01 -1.25039890e-01 -5.38743675e-01 -1.15137589e+00
1.16348691e-01 -1.27313232e+00 2.38278747e-01 -5.58346629e-01
-1.26998889e+00 7.82501042e-01 -1.26769617e-01 -9.85666931e-01
-7.54261076e-01 -2.83528477e-01 -7.40816236e-01 9.59914804e-01
-1.03166604e+00 -3.80381316e-01 -3.77036542e-01 1.80661038e-01
4.93705779e-01 1.37951672e-01 1.15556896e+00 -1.21043734e-02
-3.09926301e-01 5.17624676e-01 -7.10887983e-02 -2.98175942e-02
1.17921579e+00 -1.42677546e+00 2.90113002e-01 3.12385470e-01
1.61023304e-01 8.26730967e-01 1.10031128e+00 -2.87774056e-01
-8.13849807e-01 -4.50998574e-01 1.25494015e+00 -9.66299593e-01
7.00239718e-01 -6.52775317e-02 -9.86169696e-01 2.06836239e-01
5.46852589e-01 -7.09996521e-01 1.00934529e+00 8.58752966e-01
-4.18863624e-01 1.97049856e-01 -1.02374637e+00 5.96033990e-01
6.45245671e-01 -8.53699744e-01 -8.69394422e-01 5.20595074e-01
3.88890594e-01 -2.54932195e-01 -1.01990879e+00 -1.40008882e-01
6.60223663e-01 -1.20052409e+00 4.14410323e-01 -4.71841782e-01
5.66882968e-01 1.34948179e-01 -1.17804356e-01 -1.57125723e+00
-1.91415787e-01 -5.35490692e-01 2.31654823e-01 1.29589438e+00
7.17541158e-01 -4.71709937e-01 5.81455648e-01 1.38205206e+00
1.74355611e-01 -5.99057257e-01 -5.61402202e-01 -4.68865454e-01
1.78099275e-01 -1.63651675e-01 4.55707967e-01 9.03016686e-01
7.57960737e-01 8.96309078e-01 -2.31815293e-01 -4.14184362e-01
2.73615420e-01 4.57077995e-02 9.43354487e-01 -1.32133090e+00
-4.50863272e-01 -6.08193636e-01 -1.83240578e-01 -1.03386474e+00
5.11340201e-02 -4.33144331e-01 2.24234030e-01 -1.45309818e+00
3.23940992e-01 -3.03218722e-01 1.58885613e-01 9.33496282e-02
-4.08803642e-01 2.29030207e-01 1.12241164e-01 3.49992394e-01
-9.19441760e-01 3.09766859e-01 1.05256844e+00 1.14802709e-02
-1.20329112e-01 -3.15953754e-02 -9.13303435e-01 7.53696203e-01
7.90568531e-01 -3.46309364e-01 -4.42184150e-01 -5.43547213e-01
4.71321285e-01 3.79101425e-01 1.30203411e-01 -7.93679476e-01
3.51535946e-01 -8.43830854e-02 -1.17633464e-02 -1.62703590e-03
2.31487632e-01 -2.32903019e-01 -2.73788482e-01 5.76177686e-02
-8.39149833e-01 1.71606675e-01 -3.12263310e-01 4.42435890e-01
-4.84962642e-01 -6.52198434e-01 5.21093488e-01 -4.95578676e-01
-1.24044769e-01 -4.44586098e-01 -6.35886788e-01 5.36907732e-01
4.47636545e-01 -3.90282273e-02 -4.83440578e-01 -9.67890620e-01
-2.53510207e-01 2.43476734e-01 6.68808341e-01 1.12757511e-01
1.03295349e-01 -8.84647608e-01 -8.85565042e-01 -4.52207953e-01
1.32059872e-01 -2.81332046e-01 -7.72838891e-02 6.87927902e-01
-6.42116189e-01 5.33542097e-01 6.16989136e-02 -5.42542160e-01
-1.22535145e+00 7.62933493e-02 1.69480488e-01 -5.12980103e-01
-7.89004713e-02 6.40693843e-01 -1.57651201e-01 -6.45720601e-01
2.41431877e-01 -1.75334096e-01 -1.12779595e-01 2.95239508e-01
6.48900509e-01 5.27381003e-01 1.01501718e-01 -2.01550543e-01
-7.84804523e-02 1.21434994e-01 -1.41779989e-01 -4.69036996e-01
7.89115965e-01 -4.62042451e-01 -1.26695618e-01 7.29076326e-01
9.79937553e-01 7.43418410e-02 -8.61859977e-01 -2.79662251e-01
1.62012637e-01 -4.74261731e-01 -1.33393660e-01 -9.67722893e-01
-2.21315980e-01 7.66413450e-01 2.80799009e-02 8.17815483e-01
5.88643551e-01 -2.18469962e-01 6.35714769e-01 9.18082833e-01
5.00019908e-01 -1.32079244e+00 4.97412346e-02 5.58738112e-01
7.75995314e-01 -1.47901845e+00 1.29109025e-01 -7.94866979e-02
-9.84010458e-01 9.79654074e-01 3.96136731e-01 1.61868855e-01
-1.07327849e-02 -1.89426504e-02 5.55833161e-01 -9.34559703e-02
-1.26551044e+00 7.51770288e-02 1.86361119e-01 3.03358734e-01
1.10771251e+00 1.85438097e-01 -8.88984859e-01 3.66420656e-01
-5.30953407e-01 -2.21113667e-01 8.33858430e-01 6.48418784e-01
-5.01946032e-01 -1.23521876e+00 -2.43810311e-01 5.77177107e-01
-5.32405794e-01 1.75256804e-02 -8.99087667e-01 4.76618499e-01
-7.08836317e-01 1.60312939e+00 -8.20254087e-02 -1.94548443e-01
3.82928818e-01 3.86465877e-01 5.17691433e-01 -7.41486788e-01
-7.76283085e-01 -3.17791998e-01 9.53634799e-01 -4.56710041e-01
-3.58656138e-01 -7.04079330e-01 -7.65206039e-01 -7.66482532e-01
-5.52444994e-01 6.13379061e-01 4.31161582e-01 8.96720827e-01
2.03498572e-01 1.36405453e-01 6.36905432e-01 -4.64733899e-01
-1.20732164e+00 -1.44660270e+00 -2.92069912e-01 7.88712740e-01
4.35987562e-02 -3.70470762e-01 -4.38350856e-01 -2.10850224e-01] | [12.664900779724121, 8.2196683883667] |
7ceff17c-9228-4b8c-a918-82be53283ded | learning-debiased-representations-via | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhang_Learning_Debiased_Representations_via_Conditional_Attribute_Interpolation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhang_Learning_Debiased_Representations_via_Conditional_Attribute_Interpolation_CVPR_2023_paper.pdf | Learning Debiased Representations via Conditional Attribute Interpolation | An image is usually described by more than one attribute like "shape" and "color". When a dataset is biased, i.e., most samples have attributes spuriously correlated with the target label, a Deep Neural Network (DNN) is prone to make predictions by the "unintended" attribute, especially if it is easier to learn. To improve the generalization ability when training on such a biased dataset, we propose a chi^2-model to learn debiased representations. First, we design a chi-shape pattern to match the training dynamics of a DNN and find Intermediate Attribute Samples (IASs) --- samples near the attribute decision boundaries, which indicate how the value of an attribute changes from one extreme to another. Then we rectify the representation with a chi-structured metric learning objective. Conditional interpolation among IASs eliminates the negative effect of peripheral attributes and facilitates retaining the intra-class compactness. Experiments show that chi^2-model learns debiased representation effectively and achieves remarkable improvements on various datasets. | ['Han-Jia Ye', 'De-Chuan Zhan', 'Qi-Wei Wang', 'Yi-Kai Zhang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['metric-learning', 'metric-learning'] | ['computer-vision', 'methodology'] | [ 2.47213423e-01 1.70287564e-02 -2.49586180e-01 -9.59578335e-01
-4.51417804e-01 -4.58013564e-01 4.30670321e-01 -1.05567239e-01
-8.33284706e-02 7.95182586e-01 1.70427782e-03 6.85373843e-02
-1.30873799e-01 -7.97125280e-01 -8.96171093e-01 -1.06799996e+00
1.79656167e-02 6.07260287e-01 1.48883676e-02 1.13232344e-01
3.68695587e-01 3.98765892e-01 -1.60296118e+00 5.06677747e-01
1.04808748e+00 1.54400349e+00 8.99678618e-02 3.82404365e-02
-2.74864703e-01 8.17051232e-01 -7.09473372e-01 -4.54957962e-01
2.64676988e-01 -5.36567092e-01 -5.96661687e-01 -4.30863835e-02
8.39286089e-01 -1.99614212e-01 -2.04969123e-01 1.36220133e+00
2.85509348e-01 3.94432135e-02 1.15963328e+00 -1.43898368e+00
-1.18115520e+00 2.68495917e-01 -7.52807736e-01 -3.18980478e-02
-2.80322343e-01 9.96640027e-02 7.96893716e-01 -8.73628497e-01
5.65213203e-01 1.37078500e+00 7.65169799e-01 7.28536129e-01
-1.60912108e+00 -9.56329107e-01 4.19127703e-01 3.65471750e-01
-1.57446456e+00 -1.22304343e-01 1.08588219e+00 -5.85175514e-01
1.10785536e-01 1.25191033e-01 4.81841713e-01 1.33879292e+00
2.15984613e-01 1.01514995e+00 1.02763569e+00 8.98680761e-02
3.08804929e-01 3.53087723e-01 -6.93258364e-03 3.34938467e-01
2.30331317e-01 1.16444163e-01 -4.44728494e-01 1.34981632e-01
4.04578954e-01 2.37249836e-01 -2.32065096e-01 -7.27043629e-01
-1.20435560e+00 7.61338413e-01 7.30505645e-01 -2.60473881e-02
-1.38430595e-01 8.80899578e-02 3.41725796e-01 4.26949054e-01
5.53870559e-01 4.00748491e-01 -5.61544120e-01 1.81135163e-01
-8.91389370e-01 1.82999820e-01 1.67257637e-01 1.10308015e+00
1.01853418e+00 -5.03660850e-02 -3.57943565e-01 1.10245121e+00
-6.60782233e-02 4.06112611e-01 6.97692275e-01 -8.61003637e-01
2.67802387e-01 8.42222154e-01 6.76512495e-02 -1.03593719e+00
-1.85870290e-01 -5.56247771e-01 -1.35368860e+00 3.75289887e-01
5.91766953e-01 1.51755273e-01 -1.04575908e+00 1.84008873e+00
1.96485162e-01 -1.34266406e-01 -1.86095327e-01 9.75298762e-01
6.09523296e-01 6.98048115e-01 1.38959423e-01 -1.81842186e-02
9.55284238e-01 -6.20770454e-01 -4.82334107e-01 -2.14558482e-01
6.05488479e-01 -4.16954845e-01 1.47551060e+00 4.34639037e-01
-7.35355437e-01 -7.49641001e-01 -1.05179060e+00 -2.62659267e-02
-4.62622374e-01 1.66972324e-01 2.89899200e-01 3.12635124e-01
-6.31479323e-01 8.79970610e-01 -3.68033707e-01 7.31556788e-02
9.50382173e-01 2.53950655e-01 -1.80801034e-01 -2.28551343e-01
-9.26142216e-01 6.53938651e-01 2.33291209e-01 -1.15493573e-01
-8.77018929e-01 -1.11298692e+00 -6.64255559e-01 -1.07911207e-01
-1.96354799e-02 -3.41621548e-01 8.17130029e-01 -1.72097611e+00
-8.50107729e-01 9.93526399e-01 -1.35918319e-01 -6.64952844e-02
7.71736920e-01 -1.69658378e-01 -6.14648521e-01 -2.38959491e-01
1.84585407e-01 1.00624585e+00 1.25278044e+00 -1.55086005e+00
-5.29841542e-01 -6.54466569e-01 -4.59250182e-01 -6.12844042e-02
-4.58718717e-01 -4.31297809e-01 1.89806879e-01 -9.64740813e-01
4.08788651e-01 -8.11777771e-01 1.18752174e-01 6.26267791e-01
-4.55033273e-01 -3.48883390e-01 1.13191724e+00 -1.78779483e-01
1.14803505e+00 -2.53057289e+00 -1.00616805e-01 9.88206342e-02
2.18755096e-01 8.56781006e-02 -2.37559862e-02 -2.26167098e-01
-5.05406082e-01 1.21776149e-01 -3.72097492e-01 -7.59005034e-03
-1.02526918e-01 4.09390479e-01 -5.16645193e-01 6.23373568e-01
3.97168458e-01 5.59626758e-01 -7.81645715e-01 -4.27474111e-01
-2.17997462e-01 2.57449269e-01 -3.61468732e-01 2.68220901e-01
-2.45106503e-01 3.89097899e-01 -2.52422303e-01 7.84094632e-01
1.08336651e+00 -2.29644567e-01 -3.05129081e-01 -4.85886276e-01
1.34344414e-01 -2.02395883e-03 -9.81750846e-01 1.32768333e+00
-1.10991515e-01 7.56860197e-01 -3.29051584e-01 -1.01896238e+00
1.50051248e+00 -1.98488221e-01 2.86000133e-01 -6.10212624e-01
-5.46301762e-03 1.97797194e-01 -6.03686739e-03 -2.90931731e-01
3.78974408e-01 -2.73295313e-01 9.39572379e-02 3.34717363e-01
-1.88481078e-01 9.19442717e-03 -3.09061527e-01 -4.74988483e-02
5.56485474e-01 1.09851807e-01 -2.79510918e-04 -7.29128301e-01
3.31653178e-01 -1.09039485e-01 9.15094793e-01 5.70347667e-01
-3.64779055e-01 9.98769343e-01 7.77967930e-01 -9.46009278e-01
-1.30930305e+00 -1.14081061e+00 -4.73955840e-01 1.12071919e+00
4.50657994e-01 2.40581602e-01 -5.60154259e-01 -9.89747822e-01
3.45223993e-01 8.76444399e-01 -1.10519469e+00 -7.82900691e-01
-3.11754048e-01 -7.32231975e-01 3.37272435e-01 7.65161991e-01
6.62427247e-01 -8.87149870e-01 -2.74964005e-01 -6.40655460e-04
-1.14423089e-01 -3.99340779e-01 -7.23967195e-01 3.97786707e-01
-9.05671477e-01 -7.72523940e-01 -8.07174087e-01 -9.14272666e-01
1.04040480e+00 1.08525708e-01 1.24949527e+00 -1.27114892e-01
-1.89838663e-01 -4.02547151e-01 -2.12598443e-01 -3.06920469e-01
-2.06436992e-01 -1.13517560e-01 1.72157541e-01 3.63861561e-01
6.97850406e-01 -4.16048765e-01 -9.54603136e-01 7.37505078e-01
-7.38999844e-01 -1.03419393e-01 5.28669178e-01 9.57973599e-01
6.95317984e-01 -8.03177133e-02 6.95584297e-01 -9.00254250e-01
2.75739670e-01 -5.79994857e-01 -2.58355856e-01 1.73791453e-01
-9.74678516e-01 3.15036118e-01 8.10654283e-01 -9.60431933e-01
-9.14994240e-01 8.05556402e-02 3.37341964e-01 -6.27649844e-01
-1.87341154e-01 -1.22024953e-01 -5.41167557e-01 3.22534889e-01
7.38554597e-01 3.51420283e-01 7.88072199e-02 -5.04972100e-01
1.19563840e-01 6.69800937e-01 6.35402918e-01 -8.05947125e-01
5.64287066e-01 4.67193991e-01 1.26394285e-02 -3.63348007e-01
-1.05935264e+00 -5.00310697e-02 -7.16254056e-01 -1.96918145e-01
5.15504360e-01 -7.83348203e-01 -4.91315454e-01 6.92750514e-01
-9.45123971e-01 -3.16958308e-01 -4.99109089e-01 2.69950867e-01
-5.75256944e-01 8.16139802e-02 -2.91132003e-01 -4.71061140e-01
2.16504801e-02 -9.25081134e-01 6.89775586e-01 3.65824789e-01
-4.62568402e-01 -7.15842247e-01 -2.23150894e-01 -4.66921031e-02
3.08254242e-01 1.65132642e-01 1.36298990e+00 -7.61071622e-01
-3.40835363e-01 -1.12375557e-01 -5.87602854e-01 6.78966761e-01
1.47405073e-01 2.05132976e-01 -1.11766911e+00 -3.29853714e-01
-1.39076352e-01 -5.40051579e-01 9.51564133e-01 3.06457222e-01
1.92855883e+00 -5.29593587e-01 -3.29548568e-01 9.48266387e-01
1.24511433e+00 3.11325014e-01 7.72118449e-01 4.50281173e-01
5.98782837e-01 5.96029699e-01 7.23884344e-01 5.02747357e-01
2.24518806e-01 3.67614865e-01 6.30169272e-01 3.59732620e-02
-2.74145037e-01 -4.27658468e-01 8.59331787e-02 2.54214168e-01
3.36180568e-01 -4.78822403e-02 -6.98495030e-01 6.59776211e-01
-1.57195437e+00 -8.59156907e-01 -1.24653265e-01 2.35079145e+00
1.07560635e+00 1.39683306e-01 -2.82471394e-03 -2.66258735e-02
9.93143499e-01 3.05201709e-02 -1.16512799e+00 -1.21763483e-01
-5.10509789e-01 -3.03643495e-01 4.19530392e-01 4.84946370e-02
-1.07333636e+00 6.69369876e-01 5.80440474e+00 8.63084614e-01
-1.22981405e+00 -3.53789270e-01 1.38729596e+00 -6.22775778e-02
-5.57730377e-01 -3.89728129e-01 -6.79585874e-01 1.00345111e+00
3.59013647e-01 9.17936340e-02 1.74586639e-01 1.17852664e+00
-2.39497647e-02 5.43056428e-02 -1.42718697e+00 1.02039552e+00
-3.43052819e-02 -1.17791522e+00 4.01913524e-01 -3.15034389e-03
8.57885182e-01 -3.59993696e-01 5.99165916e-01 2.78067142e-01
1.99014306e-01 -1.18920386e+00 1.05490482e+00 6.32310688e-01
1.11938405e+00 -8.75864446e-01 5.46197653e-01 2.37462223e-01
-8.74651074e-01 -2.59107620e-01 -7.67346382e-01 2.72141367e-01
-5.40981054e-01 5.86476386e-01 -5.49174488e-01 -1.31111249e-01
1.02049637e+00 8.99474978e-01 -7.22868741e-01 9.78461802e-01
1.15938531e-02 4.24530029e-01 9.36769620e-02 -2.61472613e-02
1.71423838e-01 -4.89199579e-01 2.30768353e-01 9.81149614e-01
6.32890224e-01 -1.10963523e-01 -2.84986287e-01 1.08643544e+00
-2.49607548e-01 -2.24024937e-01 -5.51849961e-01 2.74260372e-01
6.72726274e-01 9.25925493e-01 -5.26406586e-01 -4.06963646e-01
-1.20630987e-01 1.12523150e+00 3.80549371e-01 4.03160751e-01
-6.89969063e-01 -4.51017708e-01 9.93819475e-01 2.81918317e-01
4.79513735e-01 2.89099574e-01 -9.11260486e-01 -8.47028196e-01
7.51662254e-02 -7.63889074e-01 4.90761101e-01 -9.84506428e-01
-1.72067332e+00 5.94651878e-01 -2.44419366e-01 -1.66705525e+00
-2.57420577e-02 -6.38180435e-01 -5.59236228e-01 7.49342561e-01
-1.29680789e+00 -9.79528964e-01 -4.83062744e-01 7.11431444e-01
4.73447174e-01 -2.00606585e-01 4.74299878e-01 3.42559129e-01
-4.17505413e-01 8.60718906e-01 5.85233569e-01 2.04368174e-01
1.17569375e+00 -1.32973313e+00 -3.37114185e-02 3.32598299e-01
-2.51821190e-01 4.48353171e-01 8.30134213e-01 -4.22060341e-01
-7.46432781e-01 -1.35352957e+00 7.73679256e-01 -3.93858701e-01
5.16364813e-01 -1.61026284e-01 -1.34013212e+00 4.56091285e-01
-3.33301604e-01 4.43634838e-01 7.11892784e-01 1.25461444e-02
-8.21577013e-01 -8.67804885e-01 -1.31478655e+00 3.95536542e-01
1.06019235e+00 -4.00347233e-01 -5.04210353e-01 9.56981257e-02
4.62378293e-01 8.78534466e-02 -6.92068338e-01 5.28563678e-01
5.90844691e-01 -1.09593153e+00 9.40000772e-01 -7.04695582e-01
6.39480472e-01 -4.55763489e-01 -3.29526871e-01 -1.64146721e+00
-6.01415157e-01 -1.86280534e-01 -6.10021614e-02 1.33984876e+00
2.65365899e-01 -3.12196612e-01 1.00663495e+00 6.99537158e-01
-1.65073395e-01 -7.65535176e-01 -1.01732242e+00 -9.96976256e-01
4.16831553e-01 -5.02886586e-02 9.63191211e-01 1.14029729e+00
-3.74485552e-01 1.14133388e-01 -4.11204815e-01 -1.05982549e-01
6.96375847e-01 3.51702034e-01 4.36490119e-01 -1.49831438e+00
2.40017116e-01 -7.49826610e-01 -4.44368899e-01 -1.31085968e+00
-2.16032919e-02 -7.42487133e-01 6.44804388e-02 -8.18830371e-01
3.40224117e-01 -8.72756004e-01 -4.10080165e-01 3.70250851e-01
-1.61948383e-01 2.62430429e-01 -1.30259097e-01 4.97808993e-01
-3.80032450e-01 7.74172664e-01 1.51411319e+00 -4.41302240e-01
8.00203681e-02 5.47577478e-02 -7.77127147e-01 7.58027554e-01
7.39738345e-01 -6.77031875e-01 -3.26696038e-01 -2.87360400e-01
2.45222166e-01 -4.60802197e-01 4.20121461e-01 -8.46579075e-01
2.95119956e-02 -3.23013514e-01 1.08653057e+00 -6.53914154e-01
1.14427552e-01 -9.09300268e-01 -1.73737109e-02 3.12498897e-01
-7.12406993e-01 -4.80550826e-02 -2.77487844e-01 6.45282984e-01
-2.02141449e-01 -2.20620170e-01 1.34218538e+00 2.38403920e-02
-5.91257215e-01 5.09466469e-01 -8.94774497e-02 2.65462846e-01
1.06071734e+00 -4.83710051e-01 -4.76199239e-01 -3.02982539e-01
-4.31636781e-01 1.45925581e-01 6.86303854e-01 3.01654071e-01
6.50346994e-01 -1.69118917e+00 -4.68116701e-01 5.43184042e-01
3.84238571e-01 3.17017436e-01 1.39079168e-01 4.46986139e-01
-2.21426845e-01 -2.53535867e-01 -4.67961192e-01 -8.20892632e-01
-8.34611416e-01 7.95338333e-01 4.31798577e-01 4.42559451e-01
-5.06411254e-01 1.03650498e+00 5.93152881e-01 -2.55572259e-01
4.14222062e-01 -1.73252776e-01 2.90478617e-02 3.80261689e-01
6.03550076e-01 3.66678208e-01 -1.47823364e-01 -5.41325212e-01
-2.61985600e-01 4.92099553e-01 -3.24523956e-01 5.62443435e-01
1.23566306e+00 -1.93333969e-01 -7.46068060e-02 6.09562337e-01
1.46355450e+00 -5.84092498e-01 -2.07294512e+00 -2.75000304e-01
4.11750600e-02 -6.83667183e-01 -6.75755292e-02 -9.33805943e-01
-1.20916915e+00 1.05335212e+00 8.27422023e-01 1.02195054e-01
1.13401103e+00 1.66589946e-01 5.47558188e-01 3.32098037e-01
-3.82840447e-02 -1.44839001e+00 3.97270232e-01 3.16399813e-01
8.98683369e-01 -1.28769433e+00 -2.26274639e-01 2.58021355e-02
-1.02228940e+00 1.25630903e+00 9.84147549e-01 -2.28355333e-01
7.02565849e-01 2.97259353e-02 1.79902390e-01 -4.35303478e-03
-5.51207662e-01 3.62351030e-01 2.46327683e-01 7.90629566e-01
1.35856301e-01 1.10217690e-01 2.34114408e-01 6.37381077e-01
-1.58913523e-01 -2.74473935e-01 2.98355043e-01 4.56143290e-01
-5.48294008e-01 -6.06707275e-01 -2.82795459e-01 6.04658186e-01
-8.35306570e-02 -3.05993296e-02 -3.98703724e-01 6.22874141e-01
4.38379586e-01 3.63438517e-01 6.91946685e-01 -5.69889784e-01
1.53371960e-01 1.46095783e-01 1.90000683e-01 -1.67706907e-01
-7.49106482e-02 -2.02154666e-01 -4.30215031e-01 -4.31542188e-01
-1.50660232e-01 -7.90033877e-01 -1.06897187e+00 -4.83724356e-01
2.70712506e-02 -1.99201375e-01 2.41429210e-01 6.99945986e-01
3.74808609e-01 2.68444896e-01 9.82610285e-01 -4.65903312e-01
-9.26921785e-01 -8.64040494e-01 -8.78141284e-01 8.35666776e-01
6.43829048e-01 -8.00948679e-01 -6.55412495e-01 1.81141406e-01] | [9.551665306091309, 3.269925355911255] |
45351456-10e5-4635-9f58-526684a5fa3c | sub-graph-learning-for-spatiotemporal | 2211.09740 | null | https://arxiv.org/abs/2211.09740v1 | https://arxiv.org/pdf/2211.09740v1.pdf | Sub-Graph Learning for Spatiotemporal Forecasting via Knowledge Distillation | One of the challenges in studying the interactions in large graphs is to learn their diverse pattern and various interaction types. Hence, considering only one distribution and model to study all nodes and ignoring their diversity and local features in their neighborhoods, might severely affect the overall performance. Based on the structural information of the nodes in the graph and the interactions between them, the main graph can be divided into multiple sub-graphs. This graph partitioning can tremendously affect the learning process, however the overall performance is highly dependent on the clustering method to avoid misleading the model. In this work, we present a new framework called KD-SGL to effectively learn the sub-graphs, where we define one global model to learn the overall structure of the graph and multiple local models for each sub-graph. We assess the performance of the proposed framework and evaluate it on public datasets. Based on the achieved results, it can improve the performance of the state-of-the-arts spatiotemporal models with comparable results compared to ensemble of models with less complexity. | ['Yingxue Zhang', 'Mehrtash Mehrabi'] | 2022-11-17 | null | null | null | null | ['graph-partitioning'] | ['graphs'] | [-3.11407119e-01 -5.49249575e-02 -2.17527479e-01 -9.69047025e-02
6.67930115e-03 -6.80003941e-01 6.75559640e-01 4.16755795e-01
7.66240656e-02 6.94482505e-01 1.40241116e-01 -2.03627706e-01
-5.98854780e-01 -1.15420401e+00 -8.40857148e-01 -1.00825572e+00
-6.39838278e-01 4.69492972e-01 6.84987068e-01 -1.51580617e-01
-6.34752661e-02 3.69978011e-01 -1.53313410e+00 2.32590511e-01
1.16442263e+00 7.34385312e-01 1.95733160e-01 3.73915106e-01
-2.60471515e-02 9.86282051e-01 -2.97149837e-01 3.61471623e-02
1.42390981e-01 -2.12690994e-01 -6.60244703e-01 3.00598163e-02
3.44698340e-01 2.04759017e-01 -7.06107080e-01 1.04518104e+00
2.47958466e-01 1.26991987e-01 8.69188607e-01 -1.29486394e+00
-2.14527756e-01 6.95027173e-01 -7.73766100e-01 2.10052401e-01
1.36757568e-01 -1.04228161e-01 1.20364332e+00 -4.11680728e-01
6.77583992e-01 1.23030698e+00 5.62790036e-01 -1.97931319e-01
-9.74657655e-01 -7.09952593e-01 7.17708409e-01 3.68417948e-01
-1.74844134e+00 -1.04197472e-01 8.22141647e-01 -4.82943445e-01
5.62547505e-01 2.90312618e-01 7.03915656e-01 9.41016912e-01
2.50326544e-01 5.39773345e-01 8.57365429e-01 -1.16347037e-02
1.94727182e-01 -1.49421856e-01 3.62135589e-01 9.52269316e-01
5.14837265e-01 -2.80322373e-01 -3.98997784e-01 -2.01137364e-01
2.98109740e-01 6.02487326e-02 -2.61022449e-01 -6.37525856e-01
-7.29496658e-01 3.89876574e-01 8.72258008e-01 2.76644647e-01
-1.74179092e-01 4.52580415e-02 1.51373148e-01 2.87347636e-03
6.08165324e-01 -2.16920093e-01 -3.08771402e-01 2.27378696e-01
-8.06005895e-01 1.55440137e-01 8.59357715e-01 8.48911226e-01
1.01435018e+00 -4.91691440e-01 -1.96214840e-02 6.85251653e-01
4.63463336e-01 2.87823111e-01 -1.74814761e-01 -1.73846349e-01
6.90051079e-01 1.18809032e+00 -1.89071551e-01 -1.77443707e+00
-6.29682362e-01 -7.98061311e-01 -1.19483507e+00 -3.74947339e-01
3.84204894e-01 -4.10137028e-01 -8.69207442e-01 1.70386851e+00
6.42621219e-01 6.79406762e-01 -4.45986450e-01 6.20036900e-01
1.01870382e+00 7.93117464e-01 4.83910441e-02 -5.72483316e-02
8.36772084e-01 -1.03631544e+00 -5.49936950e-01 -2.57927060e-01
7.80393898e-01 -2.61430323e-01 6.83867276e-01 1.57076970e-01
-5.21074533e-01 -6.06856763e-01 -1.05543184e+00 4.01056081e-01
-6.26167059e-01 1.35758162e-01 6.42988026e-01 3.33711207e-01
-1.10943127e+00 5.92407465e-01 -8.88188183e-01 -7.84092128e-01
2.75217712e-01 3.34258765e-01 -2.70688891e-01 -1.68528214e-01
-1.12259328e+00 3.52248490e-01 5.87878048e-01 1.08925521e-01
-7.76171923e-01 -6.73699081e-01 -6.61096275e-01 2.33605281e-01
5.82086742e-01 -5.88728666e-01 2.32392401e-01 -4.30174142e-01
-6.54561341e-01 4.92922723e-01 -1.50293499e-01 -9.87540483e-02
4.84305590e-01 2.17069656e-01 -5.03535509e-01 -1.64214626e-01
8.72920528e-02 2.51059055e-01 4.67957467e-01 -1.40207517e+00
-9.11657810e-01 -5.45413256e-01 1.45354629e-01 3.99760038e-01
-4.33890939e-01 -4.36337590e-01 -1.01515126e+00 -4.25556153e-01
1.08020408e-02 -1.17499185e+00 -1.94111452e-01 -5.52201569e-01
-5.23764908e-01 -3.24538529e-01 9.95875657e-01 -4.56815481e-01
1.76362312e+00 -2.07567120e+00 2.78499484e-01 6.21984541e-01
6.64315403e-01 2.79403850e-02 -8.25556666e-02 7.83460319e-01
2.70696551e-01 1.45369932e-01 1.40259424e-02 -1.82393119e-01
-2.45592296e-01 2.93023020e-01 2.51671504e-02 6.48695648e-01
-2.52395362e-01 8.29426765e-01 -9.45496619e-01 -6.62916541e-01
2.41838276e-01 3.61307055e-01 -4.45082277e-01 2.58707255e-01
-2.01862305e-01 6.46957517e-01 -7.80235410e-01 3.73606265e-01
9.26754475e-01 -5.90185940e-01 6.90774143e-01 -1.46051168e-01
1.61625952e-01 3.15991193e-02 -1.44537711e+00 1.33257437e+00
-6.26898184e-02 2.34580219e-01 1.95765704e-01 -1.19729912e+00
8.65543246e-01 -1.65047199e-01 6.48760438e-01 -3.10846835e-01
-1.72506705e-01 2.46296506e-02 1.76628485e-01 -4.70572650e-01
-1.37788951e-01 6.34243786e-01 1.55724853e-01 3.28973681e-01
3.59997340e-02 6.09442472e-01 3.28079671e-01 4.66867983e-01
1.31016099e+00 -8.60863253e-02 1.35699853e-01 -6.77208245e-01
5.38798273e-01 -1.96954444e-01 3.21881413e-01 8.03863108e-01
-5.05488962e-02 2.39572525e-01 6.76392615e-01 -5.77354908e-01
-4.43798989e-01 -1.12047923e+00 1.02215610e-01 1.03592789e+00
4.97958541e-01 -7.07721889e-01 -4.89807755e-01 -9.45952654e-01
5.91866300e-02 1.83752686e-01 -7.33502388e-01 -2.30967179e-01
-4.59097534e-01 -1.23315799e+00 2.62244552e-01 1.63157865e-01
5.59385598e-01 -6.20987236e-01 2.07120717e-01 -1.44920290e-01
-2.60011017e-01 -1.25195181e+00 -2.69449323e-01 -9.59702432e-02
-6.39830649e-01 -1.38349223e+00 -7.66519755e-02 -7.90441573e-01
7.73367107e-01 5.22069633e-01 1.23538673e+00 2.73906559e-01
2.86226403e-02 1.01353951e-01 -4.20622855e-01 -1.11031353e-01
3.87993120e-02 6.68310344e-01 -1.66837141e-01 5.00310957e-01
1.31150916e-01 -1.02570605e+00 -6.51585698e-01 4.06378895e-01
-6.71453357e-01 -1.18281087e-02 5.34728289e-01 2.24838153e-01
3.98169577e-01 7.58113682e-01 4.64196295e-01 -1.23638308e+00
5.81966281e-01 -1.00316787e+00 -3.90421599e-01 5.43027461e-01
-5.60680032e-01 6.71699196e-02 8.27312410e-01 -1.90137774e-01
-8.71361375e-01 7.86558986e-02 3.95002574e-01 -2.65300483e-01
2.49237176e-02 8.98287475e-01 -4.46886092e-01 -4.07193094e-01
4.33264643e-01 3.48556250e-01 -3.88623655e-01 -4.02808875e-01
3.31513196e-01 3.80311459e-01 1.05799250e-01 -5.89794695e-01
8.70124936e-01 4.91096914e-01 3.75414371e-01 -1.07750070e+00
-7.50857592e-01 -6.57390714e-01 -8.10169339e-01 -5.36791086e-01
5.04255176e-01 -1.00211465e+00 -4.60226268e-01 4.04434711e-01
-6.76771164e-01 -1.89863905e-01 4.88090515e-01 2.64191836e-01
4.97961566e-02 5.65260828e-01 -2.62837559e-01 -6.93167508e-01
6.17244095e-02 -8.89231265e-01 9.71535265e-01 1.48135453e-01
1.28916755e-01 -1.32240212e+00 3.33301812e-01 2.48928234e-01
2.67278999e-01 6.39409840e-01 1.02643716e+00 -7.01941848e-01
-8.89925599e-01 -9.49565396e-02 -2.96529084e-01 -9.89698172e-02
4.58716393e-01 1.78673178e-01 -6.62246704e-01 -4.46840644e-01
-4.50341374e-01 4.80689742e-02 1.07027709e+00 4.97120023e-01
1.08096337e+00 -2.56632984e-01 -1.01551390e+00 6.18701458e-01
1.43017209e+00 -7.28688836e-02 3.40957135e-01 -3.34625319e-02
1.20113444e+00 6.63988411e-01 4.87893701e-01 3.90093535e-01
8.28938305e-01 4.99708891e-01 6.50872767e-01 -8.37913901e-02
-1.13555893e-01 -4.54911143e-01 2.48043686e-01 8.75104308e-01
-2.75220364e-01 -7.15628624e-01 -1.13700128e+00 7.43317306e-01
-2.31621432e+00 -8.93093050e-01 -4.98868436e-01 2.02603912e+00
2.08798543e-01 3.80844660e-02 2.46579662e-01 -1.58864543e-01
9.08185422e-01 8.13746333e-01 -4.50144202e-01 2.22412184e-01
-1.46608606e-01 -6.73398674e-02 5.42882502e-01 4.70213294e-01
-1.29027581e+00 9.55495417e-01 6.01982260e+00 1.20146096e+00
-9.78033423e-01 -1.89466164e-01 6.60681844e-01 4.64227796e-02
-9.70225409e-02 8.55075344e-02 -7.76100099e-01 5.86208344e-01
8.88204455e-01 -2.22973034e-01 6.23880088e-01 5.86706042e-01
3.03036809e-01 2.55625155e-02 -9.38239396e-01 7.70100713e-01
-1.09636381e-01 -1.11830831e+00 3.55939656e-01 2.25284144e-01
9.89412487e-01 3.14425170e-01 -1.41359344e-01 3.73607039e-01
5.17674863e-01 -1.12945044e+00 5.64947054e-02 4.90442991e-01
2.69503534e-01 -8.63200128e-01 6.57650948e-01 8.13023567e-01
-1.91549599e+00 2.91667450e-02 -4.08405125e-01 -1.43560171e-01
-3.22622806e-01 6.70846462e-01 -5.23748457e-01 1.22518456e+00
9.25355017e-01 1.21340072e+00 -9.51372147e-01 1.01011920e+00
-6.12744736e-03 7.72981524e-01 -5.01536846e-01 -1.18174672e-01
3.37317318e-01 -6.64673150e-01 4.40242440e-01 1.10969985e+00
2.22889543e-01 7.92814866e-02 6.10042334e-01 4.02291209e-01
-1.70362905e-01 3.69389296e-01 -8.91034603e-01 1.91651322e-02
6.59370780e-01 1.34458494e+00 -6.28521383e-01 -1.05625242e-01
-6.28309250e-01 3.72114241e-01 8.80443990e-01 5.17308414e-01
-1.10495508e+00 -1.45392865e-01 4.48261410e-01 5.24291456e-01
3.28117430e-01 -3.33178908e-01 1.13439687e-01 -1.08937538e+00
1.81785211e-01 -7.00759172e-01 8.48669469e-01 -2.88881034e-01
-1.40117002e+00 5.58970690e-01 1.68709576e-01 -1.03206563e+00
3.06109637e-01 -2.58207440e-01 -8.01527023e-01 5.42007864e-01
-1.22426462e+00 -1.48267651e+00 -6.98400378e-01 7.72965372e-01
3.94739443e-03 3.39355096e-02 2.95396417e-01 4.86711293e-01
-6.75344348e-01 3.81060123e-01 3.23786616e-01 1.62079215e-01
5.03770828e-01 -1.18180549e+00 1.22376338e-01 1.00771487e+00
2.68005967e-01 4.14627969e-01 4.77916837e-01 -8.55380476e-01
-1.36068583e+00 -1.65502548e+00 6.47489607e-01 -3.58561873e-01
8.92995894e-01 -6.72144890e-01 -7.64591396e-01 8.11152577e-01
1.17688365e-01 2.74953902e-01 3.68643165e-01 5.64856887e-01
-9.63171646e-02 -3.88193965e-01 -8.42595100e-01 6.07679844e-01
1.58185256e+00 -1.98651284e-01 -4.62954864e-02 4.50133324e-01
5.75568020e-01 -1.96422905e-01 -8.38585734e-01 7.19673812e-01
4.01554435e-01 -1.11693633e+00 8.06319773e-01 -6.04705155e-01
4.73369583e-02 -5.06967664e-01 -5.72810769e-02 -1.45829463e+00
-7.66606152e-01 -2.48521343e-01 -3.71005386e-01 1.31618285e+00
4.09756571e-01 -8.58833253e-01 1.02427661e+00 1.91583171e-01
1.91943333e-01 -8.32401931e-01 -7.55290806e-01 -5.79995513e-01
-1.79166958e-01 6.17006235e-03 7.12455690e-01 9.72157001e-01
-1.41578957e-01 6.52669787e-01 -4.39456522e-01 8.10178220e-01
8.65391552e-01 3.45827788e-01 1.11245596e+00 -1.37264180e+00
-2.70936817e-01 -2.71470279e-01 -6.59032881e-01 -1.10549068e+00
1.97244268e-02 -1.00440741e+00 -5.68082988e-01 -1.69710553e+00
4.52028215e-01 -6.46844745e-01 -4.09141898e-01 2.49996886e-01
-4.54327077e-01 -1.83831081e-01 -2.20636547e-01 2.80113220e-01
-9.98270214e-01 6.57664001e-01 1.15637696e+00 -3.68163019e-01
-3.49696994e-01 6.97041601e-02 -6.19031072e-01 4.97698307e-01
7.04381943e-01 -3.84036154e-01 -9.37198102e-01 -2.92774856e-01
2.75822163e-01 -3.04714411e-01 5.40811978e-02 -1.20247090e+00
5.55206895e-01 -2.50614196e-01 1.51551142e-01 -7.22204566e-01
-5.00595495e-02 -9.46419418e-01 6.31825626e-01 2.51767337e-01
-8.52060392e-02 -1.11717768e-01 9.76151004e-02 1.07886481e+00
-3.07759732e-01 4.43389624e-01 4.88315523e-01 -1.06941082e-01
-5.18945277e-01 7.62663543e-01 9.03420895e-02 1.67020410e-01
1.30117190e+00 4.87281056e-03 -4.99052167e-01 -5.28294921e-01
-8.18962634e-01 7.41908491e-01 4.74404782e-01 4.62708414e-01
1.98679507e-01 -1.28581250e+00 -8.42562616e-01 3.86475883e-02
2.54806310e-01 2.02892855e-01 5.51310420e-01 8.21940005e-01
-5.40443838e-01 3.08264822e-01 1.22201189e-01 -8.00280690e-01
-1.37179911e+00 6.62159920e-01 4.37970579e-01 -7.72634983e-01
-4.05122101e-01 4.93121147e-01 6.56266570e-01 -6.47905052e-01
2.10748360e-01 -9.39976424e-02 -5.82151532e-01 -6.29480034e-02
-9.20716207e-03 4.23849523e-01 -2.09046155e-03 -6.83066726e-01
-5.30778527e-01 7.80230403e-01 -9.56543814e-03 6.99039698e-01
1.47949934e+00 -2.83947825e-01 -4.82109368e-01 4.77267444e-01
1.09884405e+00 2.20538467e-01 -8.93335462e-01 -2.69764453e-01
-5.02052940e-02 -3.85424048e-01 -1.13753766e-01 -4.17481691e-01
-1.34924340e+00 5.02797484e-01 2.26011962e-01 5.11784613e-01
1.09691179e+00 2.23504081e-01 4.26387161e-01 3.28216344e-01
6.28373206e-01 -8.07581365e-01 -1.90428317e-01 3.59785587e-01
5.90781152e-01 -1.19173694e+00 1.77573636e-01 -9.85065401e-01
-3.88325512e-01 7.92881966e-01 7.06466079e-01 -4.15728748e-01
1.34900880e+00 -1.19861327e-01 -3.68199646e-01 -6.06081784e-01
-8.65492284e-01 -2.18037859e-01 5.95402181e-01 4.45229322e-01
1.87205508e-01 2.10959241e-01 -1.97255120e-01 4.27077621e-01
-3.11843399e-02 -3.84262562e-01 6.97264224e-02 4.70248848e-01
-3.39534879e-01 -1.03014934e+00 1.00124359e-01 6.26205027e-01
-1.10207550e-01 1.54050454e-01 -7.15068936e-01 9.15632665e-01
2.79149681e-01 1.19946218e+00 -2.02136174e-01 -6.75936341e-01
2.68591464e-01 -4.00628179e-01 3.22446615e-01 -5.26795864e-01
-3.24234277e-01 -2.45729268e-01 2.69524097e-01 -6.63832605e-01
-4.19414192e-01 -4.86009777e-01 -9.14394915e-01 -5.15239656e-01
-6.86147884e-02 2.97562003e-01 1.10531375e-01 8.07881534e-01
7.15122283e-01 5.00585377e-01 8.33520412e-01 -6.51385128e-01
-2.01886371e-01 -9.21906352e-01 -9.42009628e-01 4.35046971e-01
3.16345729e-02 -8.16202700e-01 -5.38558543e-01 -3.73632938e-01] | [7.305108547210693, 6.103133678436279] |
143ea740-8f2a-4e1e-945f-f02403df0fe6 | metafusion-infrared-and-visible-image-fusion | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhao_MetaFusion_Infrared_and_Visible_Image_Fusion_via_Meta-Feature_Embedding_From_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhao_MetaFusion_Infrared_and_Visible_Image_Fusion_via_Meta-Feature_Embedding_From_CVPR_2023_paper.pdf | MetaFusion: Infrared and Visible Image Fusion via Meta-Feature Embedding From Object Detection | Fusing infrared and visible images can provide more texture details for subsequent object detection task. Conversely, detection task furnishes object semantic information to improve the infrared and visible image fusion. Thus, a joint fusion and detection learning to use their mutual promotion is attracting more attention. However, the feature gap between these two different-level tasks hinders the progress. Addressing this issue, this paper proposes an infrared and visible image fusion via meta-feature embedding from object detection. The core idea is that meta-feature embedding model is designed to generate object semantic features according to fusion network ability, and thus the semantic features are naturally compatible with fusion features. It is optimized by simulating a meta learning. Moreover, we further implement a mutual promotion learning between fusion and detection tasks to improve their performances. Comprehensive experiments on three public datasets demonstrate the effectiveness of our method. Code and model are available at: https://github.com/wdzhao123/MetaFusion. | ['Huchuan Lu', 'You He', 'Fan Zhao', 'Shigeng Xie', 'Wenda Zhao'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['infrared-and-visible-image-fusion'] | ['computer-vision'] | [ 1.49115101e-02 -3.13575089e-01 -2.84295499e-01 -3.26511174e-01
-6.82883680e-01 -2.91413628e-02 7.02254832e-01 -2.74611443e-01
-2.47396290e-01 2.42049485e-01 2.21595213e-01 1.24598918e-02
2.97845639e-02 -1.05891025e+00 -4.56407219e-01 -8.82916629e-01
4.81958091e-01 -4.91292328e-01 1.77017838e-01 -2.42913589e-01
4.24397141e-02 1.27192095e-01 -1.71803975e+00 4.86039877e-01
8.61836255e-01 1.18921137e+00 6.05783224e-01 2.44555414e-01
-2.55604535e-01 5.98407269e-01 -2.40641654e-01 -2.95295715e-01
4.89633113e-01 -1.72248572e-01 -4.13386226e-01 7.96015710e-02
2.57383734e-01 -4.04279888e-01 -5.71445227e-01 1.36256623e+00
6.72467053e-01 -2.73405816e-02 4.66985941e-01 -1.48727715e+00
-1.25337505e+00 4.30338472e-01 -8.31271887e-01 3.39815140e-01
3.38166296e-01 4.44125295e-01 8.79255712e-01 -1.10819435e+00
-8.94138124e-03 1.43500650e+00 4.67753172e-01 4.54756200e-01
-7.59194851e-01 -9.27855134e-01 2.24671543e-01 3.35686058e-01
-1.44530904e+00 -3.76527131e-01 1.01323807e+00 -2.37723872e-01
2.55765229e-01 2.70620644e-01 5.69530070e-01 1.07394266e+00
-7.19245002e-02 1.20416498e+00 1.10474038e+00 -2.36974061e-01
-3.61955762e-01 5.01921594e-01 1.32486641e-01 8.05870354e-01
3.14297676e-01 3.66106153e-01 -2.46885702e-01 -1.65541004e-02
6.79693758e-01 6.86198413e-01 -4.31049615e-01 1.03058256e-02
-1.33165371e+00 6.57362342e-01 8.83333206e-01 4.99431968e-01
-2.29754597e-01 8.31827968e-02 9.29359570e-02 2.81077266e-01
6.43624902e-01 -1.58359483e-01 -2.75614232e-01 6.50235057e-01
-3.21167141e-01 -1.62181519e-02 7.67014250e-02 9.23148811e-01
1.20984793e+00 -1.79558605e-01 -4.37678725e-01 8.26752782e-01
7.67314911e-01 6.01072133e-01 3.74242604e-01 -8.78913462e-01
4.22425836e-01 9.18480635e-01 8.97182152e-03 -1.11705673e+00
-2.84125686e-01 -6.03450537e-01 -9.79040861e-01 3.89304459e-01
-4.11526440e-03 -1.74628258e-01 -6.24161899e-01 1.62811124e+00
5.89147389e-01 3.60994697e-01 6.89392090e-02 1.09800994e+00
1.39048731e+00 7.79333413e-01 1.73277870e-01 -1.17214836e-01
1.74759626e+00 -1.09568417e+00 -6.87703133e-01 -1.90847352e-01
5.81600189e-01 -8.89285743e-01 7.88262308e-01 -1.89444900e-03
-8.14411759e-01 -1.18202019e+00 -1.00841343e+00 -1.53066114e-01
-4.40404207e-01 5.91448605e-01 8.50190938e-01 3.90808105e-01
-8.33023787e-01 1.58687174e-01 -4.64107722e-01 -1.69936553e-01
4.46336120e-01 1.66939661e-01 -4.19162214e-01 -8.53142887e-02
-1.30495465e+00 6.37486696e-01 8.88569653e-01 3.33580762e-01
-2.94046581e-01 -5.84651172e-01 -7.94701695e-01 -1.07521176e-01
4.34176326e-01 -1.08983386e+00 1.05018795e+00 -8.63421500e-01
-1.10221398e+00 8.17022443e-01 6.90926760e-02 7.81807899e-02
4.59523350e-01 -1.77494809e-01 -5.74139535e-01 2.91721076e-02
1.82985827e-01 6.66528344e-01 9.41696644e-01 -1.31425142e+00
-1.24934828e+00 -4.28676158e-01 1.79236412e-01 4.03357238e-01
-5.79248071e-01 9.94401425e-03 -6.48663223e-01 -6.41549706e-01
2.88985163e-01 -4.93077129e-01 -6.33987188e-02 3.33479196e-01
-3.50275248e-01 -7.32520163e-01 9.96512413e-01 -4.76159781e-01
8.90064657e-01 -2.41977525e+00 -1.48219123e-01 -1.94110945e-01
3.95029366e-01 1.50137112e-01 -2.64619499e-01 1.34995267e-01
-1.75665393e-01 8.44048336e-03 1.76682454e-02 -3.47482502e-01
-2.11214930e-01 -2.43631020e-01 -1.65499404e-01 6.24804378e-01
1.92020833e-01 1.13605404e+00 -9.05336022e-01 -6.44614518e-01
6.25110626e-01 6.57177448e-01 -1.18557192e-01 3.12645495e-01
9.81240273e-02 4.67848480e-01 -1.04425740e+00 9.32304323e-01
9.43984449e-01 -2.89398521e-01 -4.48680848e-01 -8.58877897e-01
-1.71674281e-01 -2.72131830e-01 -1.13425493e+00 1.82920206e+00
-4.43494201e-01 1.01597890e-01 2.62412149e-02 -1.12899172e+00
1.12831438e+00 3.39749232e-02 5.91875494e-01 -9.40763950e-01
2.84158945e-01 2.90033836e-02 -5.14043570e-01 -6.12449586e-01
3.14150482e-01 -1.13772564e-01 6.67550936e-02 1.79541335e-01
-1.73504259e-02 2.92196423e-01 -2.69398034e-01 5.92362024e-02
5.23169875e-01 3.13650906e-01 3.98944579e-02 8.36962834e-02
8.83597791e-01 -2.45735556e-01 5.95368743e-01 5.96863031e-01
-1.57919869e-01 5.10927498e-01 -3.27087879e-01 -3.47413450e-01
-6.81701362e-01 -1.08282471e+00 -1.18585326e-01 1.13214302e+00
8.08127522e-01 -2.18808770e-01 -1.95657000e-01 -6.55846775e-01
9.63936746e-02 4.12863225e-01 -6.91363215e-01 -5.19259751e-01
-2.13605851e-01 -8.11873853e-01 3.58944535e-02 5.17823100e-01
9.48963821e-01 -1.04067349e+00 -2.09376872e-01 -4.19187881e-02
-4.39139962e-01 -9.69804049e-01 -3.32551152e-01 -3.73117030e-01
-8.06188345e-01 -1.05303192e+00 -8.09896946e-01 -8.76005650e-01
3.62181962e-01 1.28020382e+00 6.76391780e-01 5.09292305e-01
-4.77119654e-01 5.46016037e-01 -5.68319917e-01 -3.66089731e-01
-1.57556474e-01 -1.65802255e-01 -1.37761146e-01 3.11160505e-01
5.46617866e-01 -4.02819514e-01 -9.54253674e-01 1.54268086e-01
-1.06159985e+00 4.48421538e-01 8.68978024e-01 6.91395521e-01
3.67662936e-01 6.51329756e-02 3.90210003e-01 -2.41177991e-01
2.82569617e-01 -4.98683363e-01 -5.86867094e-01 3.97094488e-01
-4.04807538e-01 -9.48335677e-02 1.50237799e-01 -1.17380090e-01
-1.50611091e+00 6.81678504e-02 -6.04207143e-02 -4.87484306e-01
-2.12433740e-01 1.14556476e-01 -3.45999330e-01 5.32116275e-03
3.36320639e-01 5.11661708e-01 1.87115207e-01 -7.04007328e-01
5.61822057e-01 1.01309955e+00 4.73868102e-01 -2.98928499e-01
1.14609158e+00 6.86130643e-01 -3.12517017e-01 -4.45070148e-01
-1.12431085e+00 -8.05919647e-01 -2.14279965e-01 -1.57605097e-01
8.91395748e-01 -1.39200413e+00 -9.47495162e-01 3.98796469e-01
-1.17319715e+00 3.27667147e-01 -9.91394445e-02 7.38910675e-01
-5.15539289e-01 5.69489360e-01 -5.07590830e-01 -9.34610069e-01
-5.10985374e-01 -1.00282121e+00 1.48497999e+00 4.82893944e-01
8.27030182e-01 -6.89485610e-01 -1.78029105e-01 4.63280350e-01
2.04868838e-01 1.07028686e-01 1.87707931e-01 -4.10829969e-02
-7.32371271e-01 -1.38166323e-01 -8.49162579e-01 4.17788386e-01
4.40212429e-01 3.41689736e-02 -1.20338118e+00 -2.19909474e-01
3.49384636e-01 -1.99776962e-01 1.43878865e+00 1.99581280e-01
1.34071410e+00 -1.41857192e-01 -6.16967797e-01 6.86124802e-01
1.50753391e+00 -1.29870057e-01 5.42388082e-01 5.17717719e-01
8.82826269e-01 6.71190441e-01 9.51298654e-01 3.61640573e-01
3.75612706e-01 7.27329731e-01 5.73613822e-01 -3.30974638e-01
-3.62158269e-01 -1.25435397e-01 4.59812462e-01 4.73621696e-01
-2.48461738e-01 1.12598948e-01 -6.42223537e-01 1.65719897e-01
-2.07355714e+00 -1.26071477e+00 -2.07013920e-01 1.85402548e+00
6.00867569e-01 -1.68938354e-01 1.17780022e-01 3.29305306e-02
1.09167969e+00 1.78321138e-01 -2.72590786e-01 6.42136157e-01
-2.33185619e-01 -3.63983244e-01 3.19007605e-01 2.05612183e-01
-1.32909143e+00 7.11848319e-01 4.82978582e+00 1.24161625e+00
-9.29368258e-01 4.88974661e-01 6.27065778e-01 2.09524870e-01
-4.14615184e-01 -1.00002050e-01 -8.79420102e-01 7.03347385e-01
1.63477093e-01 -7.69734457e-02 1.26075655e-01 7.31685758e-01
6.33814186e-02 2.00071931e-01 -6.39380097e-01 1.36064482e+00
1.00543005e-02 -1.22606158e+00 3.41279358e-02 -7.51207024e-02
4.48401898e-01 1.15628034e-01 2.70003051e-01 3.67507398e-01
8.39900449e-02 -6.11129582e-01 7.21845448e-01 7.71335840e-01
6.26216114e-01 -6.34255826e-01 7.60064006e-01 4.46508318e-01
-1.68053687e+00 -4.67905045e-01 -7.36620545e-01 2.08818570e-01
-4.40481864e-03 6.85093522e-01 -1.63312495e-01 1.18949127e+00
7.91359842e-01 1.07282686e+00 -8.54477465e-01 1.02289855e+00
-1.60230368e-01 -2.19375212e-02 -1.89900368e-01 2.80357271e-01
5.40420823e-02 -3.09152782e-01 3.59836847e-01 9.55932438e-01
3.77226293e-01 1.40707955e-01 4.02242422e-01 1.03752303e+00
1.11254647e-01 9.70087498e-02 -6.67294681e-01 2.27996439e-01
3.72909963e-01 1.75346315e+00 -3.13308060e-01 -3.43535870e-01
-7.34543502e-01 8.94690573e-01 2.00382337e-01 4.33664531e-01
-1.06822205e+00 -2.32761875e-01 5.70325971e-01 -2.70200133e-01
5.00155315e-02 1.06463803e-03 -1.30909890e-01 -1.41875160e+00
1.28454596e-01 -3.01971525e-01 4.82604355e-01 -8.97319913e-01
-1.52633905e+00 3.09064180e-01 -1.47251621e-01 -1.66678119e+00
5.45057118e-01 -6.70121431e-01 -6.85310543e-01 8.67200434e-01
-1.85619211e+00 -1.55721581e+00 -8.33687484e-01 7.13185310e-01
4.47110623e-01 -1.85599923e-02 2.59220093e-01 3.75874043e-01
-5.47929883e-01 4.90196198e-01 4.14008684e-02 1.22363515e-01
5.39739609e-01 -7.77192712e-01 5.14305197e-02 6.05498552e-01
7.02479556e-02 3.48815680e-01 3.21806490e-01 -3.85992825e-01
-1.56478584e+00 -1.34731698e+00 2.58295327e-01 -4.36188340e-01
3.96220267e-01 -1.31443635e-01 -9.31581378e-01 4.19626325e-01
1.82087779e-01 2.89656460e-01 4.69189286e-01 -1.11379966e-01
-4.99346286e-01 -4.14565951e-01 -1.00648868e+00 4.20234948e-01
1.33621860e+00 -5.24311006e-01 -6.98845565e-01 6.16539240e-01
1.14620852e+00 6.38496131e-02 -8.05701256e-01 7.93637216e-01
4.69446301e-01 -1.15974724e+00 1.42858303e+00 -1.64925888e-01
2.81007648e-01 -7.38729775e-01 -1.61450163e-01 -8.50333214e-01
-6.42738700e-01 -6.11272901e-02 -1.95639893e-01 1.30679023e+00
-4.29480635e-02 -8.11309636e-01 4.38607484e-01 6.96008205e-02
-1.00946631e-02 -5.35642982e-01 -6.17973268e-01 -7.29465365e-01
-3.40056509e-01 -2.86759824e-01 6.48264647e-01 9.03216422e-01
-3.84963900e-01 3.55518907e-01 -2.89698899e-01 3.62510413e-01
8.74782920e-01 5.17258704e-01 6.61942065e-01 -1.25896525e+00
-2.27253482e-01 -6.66115999e-01 -4.66266483e-01 -9.36308384e-01
5.97076379e-02 -1.00633502e+00 -1.20791197e-01 -1.66174924e+00
5.53476691e-01 -4.26322669e-01 -6.96367025e-01 6.73614800e-01
-7.36499786e-01 4.80822176e-01 2.31840521e-01 4.26327825e-01
-7.87432373e-01 1.05760121e+00 1.39871621e+00 -4.43657875e-01
6.34340644e-02 -8.23425874e-02 -1.13736641e+00 5.79606175e-01
1.07405603e+00 -2.33729675e-01 -2.60002673e-01 -4.79371697e-01
-4.25049700e-02 -1.30735904e-01 7.06911862e-01 -1.00784290e+00
3.99130881e-01 -1.51827455e-01 7.08519638e-01 -6.69565678e-01
4.15046901e-01 -8.59644949e-01 -3.74640338e-02 6.56925559e-01
-7.17161736e-03 -3.51486444e-01 3.91439162e-02 8.92725885e-01
-1.73892245e-01 2.32327189e-02 6.38603926e-01 -1.82169974e-01
-9.77147043e-01 6.50490344e-01 2.80890465e-01 -5.03763199e-01
1.16871953e+00 -4.73253071e-01 -3.45638305e-01 9.21567716e-03
-4.83303785e-01 3.35696250e-01 3.13712835e-01 8.63677561e-01
8.01128626e-01 -1.71277046e+00 -9.09146011e-01 1.97387338e-01
5.45802593e-01 -3.26432213e-02 5.65482378e-01 9.74309862e-01
5.60990572e-02 9.02482718e-02 -1.04897134e-01 -7.42465436e-01
-1.24865115e+00 8.67363691e-01 2.31285423e-01 3.16321522e-01
-6.71261072e-01 8.57281387e-01 7.30514109e-01 -5.06646156e-01
1.58024840e-02 -1.15718646e-02 -4.53856289e-01 -1.11226305e-01
8.74302924e-01 4.23610121e-01 -2.04782471e-01 -6.38583720e-01
-2.68472016e-01 7.66739130e-01 -2.13806536e-02 2.65628994e-01
1.14706755e+00 -5.21476328e-01 -1.64310902e-01 2.35133916e-01
1.34687996e+00 -2.33767763e-01 -1.14455616e+00 -7.57214665e-01
-4.36180949e-01 -7.32896984e-01 3.86432439e-01 -3.04850012e-01
-1.31934750e+00 8.31494749e-01 9.86576319e-01 2.47152165e-01
1.35494256e+00 2.11674616e-01 6.18127882e-01 3.18858564e-01
3.49760920e-01 -6.51416421e-01 2.62326449e-01 8.18032026e-03
7.31181622e-01 -1.64946795e+00 -4.73480038e-02 -7.28588879e-01
-2.95836359e-01 1.10575843e+00 8.66594315e-01 -8.20495188e-02
6.58664644e-01 -1.27247065e-01 -3.43349949e-02 -1.80279270e-01
-4.17087585e-01 -6.55854821e-01 3.91645759e-01 5.45637071e-01
3.13617796e-01 -5.35682254e-02 -2.67624795e-01 4.83009815e-01
3.86045784e-01 -2.98846886e-02 -9.62211490e-02 8.41185868e-01
-8.62422585e-01 -9.48499143e-01 -4.90795910e-01 4.31045175e-01
-2.89075255e-01 -1.10803753e-01 -4.34192270e-02 4.00087357e-01
4.37347621e-01 1.13044512e+00 -7.52952397e-02 -6.28962338e-01
2.14083993e-04 -3.32861483e-01 3.84625643e-01 -5.68827093e-01
-2.74368674e-01 7.63899386e-02 -4.69910145e-01 -5.63373983e-01
-7.57881701e-01 -1.77314013e-01 -9.97394323e-01 -2.04934418e-01
-8.17915916e-01 1.19898833e-01 5.56369483e-01 6.69647992e-01
3.54011059e-01 5.49567103e-01 1.09720790e+00 -7.35900819e-01
-6.52161956e-01 -9.10071254e-01 -5.99957168e-01 3.95749956e-01
2.22586066e-01 -8.79859030e-01 -3.89221102e-01 -2.65612692e-01] | [10.286189079284668, -1.6819655895233154] |
b2ae4711-7823-4fe6-be29-d60bd6f63433 | low-light-image-and-video-enhancement-a | 2212.10772 | null | https://arxiv.org/abs/2212.10772v4 | https://arxiv.org/pdf/2212.10772v4.pdf | Low-Light Image and Video Enhancement: A Comprehensive Survey and Beyond | This paper presents a comprehensive survey of low-light image and video enhancement. We begin with the challenging mixed over-/under-exposed images, which are under-performed by existing methods. To this end, we propose two variants of the SICE dataset named SICE\_Grad and SICE\_Mix. Next, we introduce Night Wenzhou, a large-scale, high-resolution video dataset, to address the lack of low-light video datasets that discourages the use of low-light image enhancement (LLIE) methods in videos. Our Night Wenzhou dataset is challenging since it consists of fast-moving aerial scenes and streetscapes with varying illuminations and degradation. We then construct a hierarchical taxonomy, conduct extensive key technique analysis, and performs experimental comparisons for representative LLIE approaches using our proposed datasets and the current benchmark datasets. Finally, we identify emerging applications, address unresolved challenges, and propose future research topics for the LLIE community. Our datasets are available at https://github.com/ShenZheng2000/LLIE_Survey. | ['Gaurav Gupta', 'Changjie Lu', 'Jinqian Pan', 'Yiling Ma', 'Shen Zheng'] | 2022-12-21 | null | null | null | null | ['low-light-image-enhancement', 'video-enhancement'] | ['computer-vision', 'computer-vision'] | [ 4.37970847e-01 -8.87727797e-01 1.43629059e-01 -1.10733844e-01
-7.82761574e-01 -5.48131168e-01 4.45453554e-01 -5.00986040e-01
-1.72709420e-01 7.52375364e-01 1.69032127e-01 -1.32812887e-01
-5.40309399e-02 -7.12067902e-01 -5.45291185e-01 -9.97466922e-01
4.15683649e-02 -8.62788737e-01 1.39567494e-01 -3.29862863e-01
5.42575195e-02 2.07212761e-01 -1.86382949e+00 4.14297134e-01
8.66145372e-01 9.36698079e-01 5.35957336e-01 8.76642466e-01
5.12331188e-01 8.37217569e-01 -2.69644439e-01 -2.71416843e-01
8.66644859e-01 -4.57548529e-01 -4.96555358e-01 5.51018596e-01
9.66479480e-01 -6.39625251e-01 -5.43672085e-01 1.25125229e+00
7.82697737e-01 1.41180977e-01 3.26760828e-01 -1.30462849e+00
-8.85551453e-01 4.88024466e-02 -8.01799655e-01 6.27443135e-01
2.85996497e-01 6.10006273e-01 6.88686550e-01 -1.13931465e+00
7.21471429e-01 9.21217978e-01 6.98043466e-01 2.51240194e-01
-8.34296882e-01 -8.19017708e-01 1.96439505e-01 4.03592467e-01
-1.22103322e+00 -7.11241901e-01 6.12722337e-01 -2.61839420e-01
6.39656603e-01 2.60170221e-01 5.05032539e-01 9.93142843e-01
1.86291337e-01 6.48443997e-01 1.55196619e+00 -2.67434597e-01
-5.36353588e-02 -2.97242194e-01 5.23967855e-02 6.95648730e-01
3.54886085e-01 4.39547569e-01 -5.27472198e-01 1.34604976e-01
3.54325891e-01 2.36542404e-01 -7.85355151e-01 1.76431090e-01
-1.11720836e+00 1.88830391e-01 3.33176643e-01 8.27996805e-03
-2.52398849e-01 -6.28210828e-02 1.35982752e-01 3.31883878e-01
5.65253079e-01 1.26744866e-01 -4.31075037e-01 1.40266657e-01
-8.44784558e-01 6.31804690e-02 3.17819007e-02 1.27477491e+00
8.08459044e-01 2.01468349e-01 -2.70725012e-01 9.63100374e-01
4.94127870e-02 6.89781606e-01 -1.40445337e-01 -1.20350158e+00
4.19547766e-01 4.95242216e-02 3.15596193e-01 -9.39431429e-01
-1.63550332e-01 -2.85013855e-01 -1.05485785e+00 1.79728404e-01
-3.68509404e-02 -2.39954770e-01 -4.85430688e-01 1.41901469e+00
2.57281363e-01 4.03928965e-01 2.25506037e-01 9.72431302e-01
1.16299546e+00 8.82656991e-01 -3.36695090e-02 -7.02102065e-01
1.35456812e+00 -1.31372631e+00 -7.80646741e-01 -2.34270710e-02
2.19189629e-01 -1.13216770e+00 1.28111124e+00 6.50272012e-01
-1.27980733e+00 -7.88781881e-01 -8.51815641e-01 -2.15264782e-01
-2.32649788e-01 5.88252962e-01 3.78934413e-01 4.23876613e-01
-1.13014865e+00 4.02655959e-01 -5.49068809e-01 -3.68865043e-01
5.06588519e-01 -1.29383057e-01 -1.01131089e-01 -5.57846069e-01
-1.05416906e+00 3.40625703e-01 6.95209354e-02 2.76354432e-01
-1.07345080e+00 -7.74823487e-01 -7.36256480e-01 -3.96779746e-01
5.51419497e-01 -5.02875865e-01 9.37084973e-01 -4.58718568e-01
-1.21519887e+00 8.35146308e-01 -1.65867567e-01 -6.89775720e-02
3.25026363e-01 -3.83117586e-01 -7.81787276e-01 3.42879176e-01
-5.44448420e-02 5.27738869e-01 9.20852959e-01 -1.51712322e+00
-9.00368035e-01 -1.20478414e-01 3.96959394e-01 1.53638080e-01
-3.20633501e-01 3.24533552e-01 -8.10268044e-01 -8.81266117e-01
-2.82332063e-01 -9.43343282e-01 -5.52352220e-02 1.06520131e-01
-3.84171605e-01 2.16754451e-01 1.20534611e+00 -6.11749470e-01
1.40065873e+00 -2.39230680e+00 -3.94434690e-01 -4.04337734e-01
3.50147516e-01 3.33283484e-01 -4.64529604e-01 3.85264575e-01
-7.74113536e-02 -2.85891257e-02 -3.31443220e-01 -2.70258427e-01
-3.00147861e-01 3.27210091e-02 -2.97802180e-01 7.88197279e-01
3.69643830e-02 5.57027280e-01 -1.02151561e+00 -5.34345686e-01
5.06639779e-01 4.63061988e-01 -5.45241296e-01 2.83968538e-01
2.04535380e-01 4.18144077e-01 -5.67102917e-02 1.25530076e+00
1.17045188e+00 -6.91617355e-02 -2.02036500e-01 -8.01403105e-01
-6.20031714e-01 -2.99485564e-01 -1.08062625e+00 1.36287427e+00
-5.49809694e-01 8.32767069e-01 2.91618317e-01 -4.29757982e-01
7.70196140e-01 -1.25009105e-01 5.37114084e-01 -6.29188657e-01
1.28276587e-01 1.22250535e-01 -5.33034682e-01 -8.11352909e-01
7.41469920e-01 2.05129832e-01 3.27271998e-01 2.29533657e-01
-1.61167666e-01 -1.97574854e-01 6.35970473e-01 1.63326055e-01
9.77345347e-01 1.40143394e-01 2.86051542e-01 -2.87850052e-01
5.96434593e-01 -2.96468079e-01 7.41642356e-01 5.88228524e-01
-5.96944511e-01 7.63823450e-01 -1.86284691e-01 -3.03873152e-01
-8.25190604e-01 -1.09806919e+00 -4.01947707e-01 9.87825692e-01
6.64972126e-01 -6.70760393e-01 -5.46310365e-01 -3.70609701e-01
-5.31880975e-01 2.59765834e-01 -3.36770087e-01 1.56855255e-01
-2.78477401e-01 -1.19076216e+00 2.27861851e-01 1.84870526e-01
1.34063363e+00 -7.67155409e-01 -6.64558053e-01 -2.43101016e-01
-7.39517510e-01 -1.56236541e+00 -5.21787882e-01 -1.06885895e-01
-5.54872870e-01 -1.31212640e+00 -5.81154764e-01 -6.84718251e-01
4.36485946e-01 1.08369410e+00 1.24300241e+00 1.82732031e-01
-5.71046114e-01 5.54107606e-01 -5.96743524e-01 -2.59086698e-01
-4.21220884e-02 -5.42108297e-01 1.63894236e-01 1.27820343e-01
2.55339026e-01 -5.02529263e-01 -1.19947505e+00 6.20995522e-01
-1.20358634e+00 1.87666252e-01 5.28659225e-01 8.59045804e-01
9.05895770e-01 5.40963769e-01 2.72662919e-02 -6.67400301e-01
2.95979321e-01 -4.13734853e-01 -7.69984901e-01 3.37273508e-01
-4.72260922e-01 -6.56132936e-01 6.44602001e-01 -3.07430416e-01
-1.41652691e+00 -1.03200078e-01 -1.18283421e-01 -3.43333542e-01
-2.74277091e-01 -9.88588575e-03 -3.08920115e-01 -3.47010881e-01
5.30106544e-01 3.28779668e-01 -6.18914366e-01 -4.21131581e-01
1.95446715e-01 7.59081006e-01 9.12342250e-01 -3.99741501e-01
8.69396210e-01 8.51222515e-01 -2.53701024e-02 -1.22913766e+00
-1.03712702e+00 -5.17150462e-01 -4.72582132e-01 -5.67626059e-01
9.19279873e-01 -1.26185918e+00 -3.35398525e-01 7.61600912e-01
-8.15020502e-01 -6.36388183e-01 -1.62838474e-01 3.80387872e-01
-4.94482756e-01 6.25444055e-01 -8.50068033e-01 -6.74930751e-01
-3.42166454e-01 -1.28655720e+00 1.27832961e+00 3.65470320e-01
5.12670279e-01 -6.49856865e-01 -1.08261324e-01 5.01526415e-01
3.68955672e-01 2.94726223e-01 3.76800328e-01 5.71425259e-01
-8.30470979e-01 4.13583845e-01 -6.10796571e-01 6.87050760e-01
2.18029276e-01 3.29221755e-01 -1.08491492e+00 -6.12052500e-01
-6.02467358e-02 -2.86727279e-01 1.03813827e+00 5.86693227e-01
1.51475859e+00 -3.24198827e-02 -3.35856900e-02 1.25272739e+00
1.78308582e+00 1.71353593e-01 7.97683537e-01 4.87350106e-01
5.12934744e-01 2.93570578e-01 1.04827189e+00 8.12966526e-01
3.53342950e-01 5.98782778e-01 6.15400136e-01 -4.84467030e-01
-5.60190618e-01 1.69052988e-01 6.69819415e-01 9.10980046e-01
-5.55127323e-01 -4.38519657e-01 -5.42533576e-01 4.92904365e-01
-1.38514793e+00 -1.04114902e+00 -4.67847645e-01 1.82371712e+00
6.87251985e-01 -4.25099194e-01 -8.75581894e-03 1.61757559e-01
7.58260489e-01 5.32583594e-01 -4.15113688e-01 3.44575077e-01
-6.97783709e-01 -5.45496494e-03 5.92800975e-01 1.09526649e-01
-1.44453371e+00 8.04671645e-01 6.16049433e+00 7.75955915e-01
-9.08789217e-01 3.70962650e-01 5.43602407e-01 -1.03109993e-01
1.85730428e-01 -4.81735803e-02 -8.89530361e-01 6.84070587e-01
5.66420615e-01 -1.10795073e-01 5.46183944e-01 5.35064638e-01
4.95353848e-01 -3.18299532e-01 -6.13472521e-01 1.38225484e+00
2.93579787e-01 -1.07832825e+00 -2.39916936e-01 -1.02753326e-01
1.27501500e+00 2.85624355e-01 3.05748820e-01 4.41152491e-02
2.58477658e-01 -3.70836526e-01 3.64029348e-01 4.13389593e-01
1.24817312e+00 -4.56999689e-01 4.95142192e-01 -6.94469661e-02
-1.54487109e+00 -3.34411263e-01 -7.09141314e-01 1.38106123e-01
2.43497506e-01 5.77659547e-01 3.37539017e-01 8.95479143e-01
1.39230740e+00 1.41011107e+00 -8.78710508e-01 1.12959778e+00
-2.61609107e-01 5.83268881e-01 9.50320996e-03 5.77354372e-01
1.02427043e-01 -3.92530650e-01 5.55523396e-01 1.47042739e+00
5.66901207e-01 5.21821082e-01 4.09137249e-01 4.24546123e-01
-1.13262445e-01 -9.48756561e-02 -6.73571944e-01 2.87610799e-01
2.33507365e-01 1.51864040e+00 -5.06143928e-01 -3.23414117e-01
-7.32711196e-01 9.48876798e-01 -4.56593961e-01 6.52713835e-01
-9.64186788e-01 -4.53983396e-01 8.84055912e-01 -2.22189128e-02
8.27986095e-03 -2.05714434e-01 2.17801243e-01 -1.63966310e+00
-1.41455578e-02 -1.05487883e+00 5.04486084e-01 -1.43225348e+00
-1.18201792e+00 7.31223345e-01 1.40488029e-01 -1.90820003e+00
4.73027974e-01 -5.71678400e-01 -5.24441600e-01 4.28004742e-01
-2.31596208e+00 -1.01437581e+00 -1.18692589e+00 9.58241940e-01
1.10460746e+00 -1.59783751e-01 2.02582970e-01 7.57852495e-01
-7.81871974e-01 1.82483986e-01 3.00173104e-01 -1.58490613e-01
9.98067260e-01 -7.64075756e-01 1.10092431e-01 1.56972992e+00
-1.74013898e-01 3.67373526e-01 6.43059492e-01 -4.76235032e-01
-1.49241209e+00 -1.68814921e+00 4.43348020e-01 -2.18348488e-01
5.86111128e-01 -9.25381407e-02 -6.90687060e-01 4.41954672e-01
5.33870280e-01 4.84836996e-01 4.90681201e-01 -5.80227613e-01
-3.29069756e-02 -3.18876892e-01 -1.06474566e+00 7.79158592e-01
1.62079024e+00 -3.98173511e-01 -3.34756053e-03 4.64785397e-01
5.69407046e-01 -1.97391436e-01 -8.62137258e-01 6.06765330e-01
2.61226833e-01 -1.36356890e+00 1.19961369e+00 6.18396178e-02
7.22268939e-01 -6.33758545e-01 -5.94724417e-01 -1.27816272e+00
-3.93826962e-01 -7.01020837e-01 -1.28779924e-02 1.33034956e+00
-2.36694410e-01 -3.41069549e-01 3.47429365e-01 -1.10539058e-02
-2.61788636e-01 -5.98832965e-01 -1.82580218e-01 -9.22801495e-01
-3.37548554e-01 -4.76522267e-01 5.09734392e-01 9.13722396e-01
-5.98840773e-01 -5.00676110e-02 -8.75254273e-01 3.66483450e-01
1.20408452e+00 3.10960919e-01 8.88384819e-01 -7.63196647e-01
-2.41881251e-01 -1.76452077e-03 -7.64818043e-02 -1.12224734e+00
-8.70570093e-02 -6.51939809e-01 1.46745086e-01 -1.38864732e+00
4.55477893e-01 -1.10380746e-01 -1.81845784e-01 3.61774504e-01
-2.34295517e-01 8.88122499e-01 2.73820490e-01 2.31637359e-01
-9.29116488e-01 6.39021158e-01 1.29272938e+00 -2.79664725e-01
-2.13036370e-02 -2.33700424e-01 -6.61145449e-01 9.02289450e-01
8.12148035e-01 -4.70117778e-02 -3.88914824e-01 -5.99689007e-01
4.50519472e-02 -1.85005888e-01 5.18117547e-01 -1.13401473e+00
5.75191490e-02 -4.29846257e-01 3.84818196e-01 -8.69497597e-01
3.06679457e-01 -9.33759391e-01 1.45988986e-01 2.40256086e-01
-1.16257548e-01 2.80848473e-01 5.95325120e-02 4.34699714e-01
-2.88097441e-01 1.10451214e-01 1.29210925e+00 -7.46213421e-02
-1.19709635e+00 6.56776428e-01 -3.00951511e-01 1.42950654e-01
1.08176732e+00 -1.89833537e-01 -8.04055393e-01 -2.02185586e-01
-4.25163418e-01 2.98879415e-01 5.89451730e-01 2.72362262e-01
8.69462013e-01 -1.27602291e+00 -8.54493082e-01 2.84936100e-01
4.55756545e-01 -3.82395446e-01 8.72953892e-01 1.07630146e+00
-5.46240866e-01 9.46225878e-03 -5.16236424e-01 -4.84013230e-01
-1.84742594e+00 6.56588137e-01 1.37963280e-01 6.93567246e-02
-7.71483421e-01 5.81816494e-01 5.43352246e-01 2.96784379e-02
1.35679031e-03 -3.76484632e-01 -2.27023110e-01 -3.20183396e-01
9.33294952e-01 7.31583595e-01 3.22643854e-02 -7.72274852e-01
-1.66332319e-01 7.78299987e-01 3.08197379e-01 4.79219586e-01
1.59137189e+00 -8.76685798e-01 -2.12768950e-02 3.45995240e-02
1.08425748e+00 1.46874562e-01 -1.33679926e+00 -3.12103957e-01
-6.35040879e-01 -1.13677132e+00 2.71182328e-01 -5.72431087e-01
-1.36082792e+00 8.11736584e-01 8.83700311e-01 -2.12990455e-02
2.17910719e+00 -2.38236770e-01 9.04600918e-01 4.22704875e-01
4.06159848e-01 -1.10488057e+00 1.11119822e-01 4.49690461e-01
8.24545860e-01 -1.47343385e+00 2.36279503e-01 -6.63447142e-01
-5.04805386e-01 8.69111359e-01 7.27958560e-01 1.40738830e-01
6.69314146e-01 4.36563551e-01 1.32968009e-01 -9.88295376e-02
-9.47839856e-01 -6.97061181e-01 -8.86171609e-02 7.89087355e-01
3.58075589e-01 -3.67426306e-01 -2.36722812e-01 2.61417478e-01
4.37919982e-03 2.81964600e-01 8.25209975e-01 8.86318386e-01
-3.84642750e-01 -8.48943114e-01 -5.47829270e-01 2.72279769e-01
-6.27288342e-01 -3.45458776e-01 -8.58525094e-03 5.85817814e-01
5.11961877e-01 1.29243803e+00 -3.00459415e-01 -5.52008390e-01
2.82264411e-01 -6.26119852e-01 3.08946431e-01 -1.80972025e-01
-1.40671447e-01 2.64041036e-01 -3.74699235e-02 -7.60743856e-01
-9.85802174e-01 -6.58106863e-01 -4.96302694e-01 -3.97144705e-01
-1.81212634e-01 -2.79596299e-01 2.94777334e-01 3.53055805e-01
2.66820043e-01 5.81060588e-01 8.22401583e-01 -1.17933047e+00
-1.96739547e-02 -6.22136056e-01 -8.37579012e-01 5.32018185e-01
4.38060403e-01 -9.04991925e-01 -4.62301344e-01 6.14315629e-01] | [10.7743558883667, -2.2242298126220703] |
0f83d7b8-de1e-4562-adf5-3b5bd210580f | false-sense-of-security-leveraging-xai-to | 2307.04358 | null | https://arxiv.org/abs/2307.04358v1 | https://arxiv.org/pdf/2307.04358v1.pdf | False Sense of Security: Leveraging XAI to Analyze the Reasoning and True Performance of Context-less DGA Classifiers | The problem of revealing botnet activity through Domain Generation Algorithm (DGA) detection seems to be solved, considering that available deep learning classifiers achieve accuracies of over 99.9%. However, these classifiers provide a false sense of security as they are heavily biased and allow for trivial detection bypass. In this work, we leverage explainable artificial intelligence (XAI) methods to analyze the reasoning of deep learning classifiers and to systematically reveal such biases. We show that eliminating these biases from DGA classifiers considerably deteriorates their performance. Nevertheless we are able to design a context-aware detection system that is free of the identified biases and maintains the detection rate of state-of-the art deep learning classifiers. In this context, we propose a visual analysis system that helps to better understand a classifier's reasoning, thereby increasing trust in and transparency of detection methods and facilitating decision-making. | ['Ulrike Meyer', 'Arthur Drichel'] | 2023-07-10 | null | null | null | null | ['explainable-artificial-intelligence', 'decision-making'] | ['computer-vision', 'reasoning'] | [ 1.22166857e-01 5.41340649e-01 -1.48809224e-01 -4.55261804e-02
1.22275941e-01 -7.00595796e-01 8.25229824e-01 5.38896695e-02
-2.73277402e-01 6.67037010e-01 -1.86435759e-01 -9.44330096e-01
7.77301118e-02 -9.94264841e-01 -4.59039450e-01 -5.11530817e-01
1.58742517e-02 2.19943315e-01 5.72552323e-01 -8.64812359e-02
3.20779443e-01 7.94990242e-01 -1.66545188e+00 8.13654900e-01
6.75888121e-01 8.45169008e-01 -4.61238891e-01 6.82955444e-01
-1.82721585e-01 1.55195749e+00 -1.27171159e+00 -6.69244647e-01
1.07293859e-01 5.04444819e-03 -9.08495009e-01 -5.48742592e-01
4.81108427e-01 -5.66267610e-01 -5.65952808e-02 1.00920081e+00
-2.41852663e-02 -8.38130295e-01 6.84418738e-01 -1.88936090e+00
-5.27646422e-01 7.01318741e-01 -3.66316855e-01 2.63819337e-01
2.43514534e-02 5.32471716e-01 9.03360665e-01 -3.03931892e-01
7.17800498e-01 1.36019433e+00 7.38531768e-01 8.78408074e-01
-1.32929623e+00 -8.55298579e-01 2.99707264e-01 2.65888244e-01
-9.31605101e-01 -2.47507155e-01 9.95308638e-01 -8.42060506e-01
9.70054328e-01 2.89280027e-01 6.32641971e-01 1.74445379e+00
-1.06864229e-01 6.45551801e-01 1.43819678e+00 -4.78522003e-01
5.95288694e-01 7.62516320e-01 4.28328872e-01 9.52288985e-01
1.00769866e+00 3.14082891e-01 -4.15807426e-01 -3.61511201e-01
4.23292726e-01 -2.03398749e-01 1.39535308e-01 -4.77234066e-01
-9.31078434e-01 9.56336737e-01 6.24910235e-01 3.32990617e-01
-4.48092431e-01 1.44527599e-01 4.76789266e-01 2.68387973e-01
4.15327609e-01 6.02264881e-01 -5.25935352e-01 6.46270439e-02
-5.92015564e-01 2.69749254e-01 9.75264013e-01 5.25406837e-01
7.17902899e-01 2.38444760e-01 -1.48781791e-01 9.28211957e-02
2.58192807e-01 4.86958295e-01 1.18831970e-01 -1.01009154e+00
-6.39807954e-02 1.06006336e+00 1.25040501e-01 -1.34513807e+00
-4.37335074e-01 -3.97476405e-01 -3.97266537e-01 1.09973872e+00
8.74032378e-01 -1.71839789e-01 -5.74147105e-01 1.66782701e+00
1.08555533e-01 -1.69402152e-01 1.25729173e-01 6.22601151e-01
6.14169419e-01 9.77989882e-02 3.27759206e-01 3.42609674e-01
1.54029083e+00 -3.91416937e-01 -3.82195473e-01 -3.80350977e-01
7.87344694e-01 -8.51490572e-02 1.11183894e+00 5.66574812e-01
-5.09656727e-01 -3.04844141e-01 -1.31999898e+00 2.28382945e-01
-8.47133338e-01 1.27207309e-01 7.25716293e-01 1.17824996e+00
-7.89934397e-01 3.71516377e-01 -5.62947273e-01 -5.06313324e-01
1.15532005e+00 5.88902235e-02 -8.09295774e-02 3.57493997e-01
-9.30906117e-01 1.11817479e+00 1.90474555e-01 -5.24613738e-01
-1.10548604e+00 -7.40968943e-01 -1.91092491e-01 5.28060310e-02
3.64346594e-01 -4.52209085e-01 1.09562361e+00 -1.01246500e+00
-9.80654716e-01 1.14942598e+00 1.47223711e-01 -8.58489752e-01
9.41640735e-01 9.59252641e-02 -2.48680025e-01 1.07262768e-01
-5.76083956e-04 6.93184435e-01 1.06690621e+00 -1.58107054e+00
-6.94228768e-01 -4.25552398e-01 5.17976224e-01 -9.14077163e-01
-5.40648699e-01 5.63803166e-02 7.63674140e-01 -3.08390677e-01
-3.23593795e-01 -7.23378778e-01 1.43415239e-02 2.63373345e-01
-4.67771620e-01 1.25195440e-02 1.41230130e+00 -6.59309924e-01
1.06462264e+00 -1.91926122e+00 -5.67264974e-01 1.65861234e-01
8.65908444e-01 7.63427913e-01 3.95505279e-02 8.38100389e-02
1.45605177e-01 6.32203758e-01 1.54414773e-01 5.71523421e-03
9.93821099e-02 1.94459006e-01 -5.16941130e-01 2.11604014e-01
6.31303191e-01 7.28628278e-01 -8.58368158e-01 -4.92058545e-01
3.65654379e-01 4.02619600e-01 -6.86744273e-01 1.23335145e-01
-4.12307054e-01 2.52167732e-01 -3.39583576e-01 7.88283885e-01
7.30169117e-01 -2.11205915e-01 5.31850517e-01 -1.50071830e-01
-2.05629334e-01 5.46976984e-01 -7.62932539e-01 8.94363999e-01
-3.96589816e-01 1.12547195e+00 9.62971002e-02 -9.57226694e-01
1.11009121e+00 -1.99603811e-02 1.44155428e-01 -4.81655061e-01
2.82443106e-01 6.69382736e-02 1.35468751e-01 -4.10461843e-01
1.73851773e-01 3.01402330e-01 2.97710478e-01 6.50922537e-01
-1.30866095e-01 2.55981714e-01 -7.31017441e-03 3.07579547e-01
1.35874581e+00 -1.10638805e-01 4.09630299e-01 -2.42105931e-01
4.55967754e-01 4.59515423e-01 3.80897552e-01 1.11206353e+00
-7.43324935e-01 -6.09911270e-02 1.01064813e+00 -1.05665410e+00
-1.04200912e+00 -8.96654725e-01 2.12256357e-01 9.79532003e-01
-2.65325040e-01 -3.27611148e-01 -1.13947964e+00 -1.24947476e+00
9.05664191e-02 7.99874663e-01 -8.07504416e-01 -2.53074229e-01
-3.11522692e-01 -6.64442658e-01 1.01762712e+00 3.90955120e-01
5.82370877e-01 -1.02037311e+00 -1.44597101e+00 6.41786531e-02
5.55992126e-03 -1.21487129e+00 9.62544918e-01 1.49587646e-01
-6.08932316e-01 -1.48189151e+00 -7.00738654e-02 -1.06667966e-01
6.04932189e-01 3.08769017e-01 1.32063866e+00 5.25813878e-01
-5.04013002e-01 5.19855134e-02 -3.36393207e-01 -9.52946961e-01
-1.19228077e+00 1.82263583e-01 -2.39813879e-01 -3.15685213e-01
8.71626675e-01 -7.47159958e-01 -4.74704117e-01 2.93575495e-01
-6.04226232e-01 -1.74467489e-02 5.52874565e-01 5.54562390e-01
-4.67953861e-01 1.52244028e-02 3.91264588e-01 -1.01637447e+00
7.59772778e-01 -3.58233511e-01 -7.28790462e-01 2.14795217e-01
-9.69082415e-01 2.12362930e-02 6.42669201e-01 -5.05713165e-01
-1.03076184e+00 -2.55786240e-01 3.03557098e-01 -1.43224165e-01
-7.58265793e-01 -2.24024072e-01 4.03761538e-03 -2.82783389e-01
1.42516398e+00 -2.87471652e-01 5.13073541e-02 -3.62430304e-01
4.55243379e-01 7.95283616e-01 2.29728788e-01 -4.92167771e-01
7.44035125e-01 8.56598616e-01 3.15678529e-02 -5.98744094e-01
-7.02980578e-01 1.76556874e-02 -6.21457636e-01 -5.99096775e-01
5.25602758e-01 -6.88918948e-01 -1.17926645e+00 2.95492411e-01
-1.51318133e+00 -2.86231488e-01 -1.56910315e-01 -1.60414696e-01
-1.77538753e-01 2.56523371e-01 -3.12001556e-01 -1.18131053e+00
-1.70310944e-01 -1.05766678e+00 6.22310996e-01 -6.42231554e-02
-5.92642248e-01 -7.01183975e-01 -1.48198567e-02 4.43282038e-01
6.05321705e-01 4.85468984e-01 1.04689908e+00 -9.58489180e-01
-5.21638513e-01 -8.70304778e-02 -6.72744036e-01 4.02667731e-01
-1.18327498e-01 3.97068173e-01 -1.70276701e+00 3.64737771e-02
-3.87244195e-01 2.17602961e-03 6.79234743e-01 1.02199920e-01
9.33933258e-01 -7.20796287e-01 -4.79078650e-01 1.06317423e-01
1.08807135e+00 1.94505796e-01 5.43812096e-01 8.25076520e-01
4.61596042e-01 1.04686916e+00 2.02565074e-01 4.77030009e-01
2.01879725e-01 5.81514120e-01 8.61892462e-01 -1.95540130e-01
-3.40501934e-01 -4.31889325e-01 2.15283588e-01 -4.54273283e-01
-1.40839785e-01 6.87510800e-03 -1.26042211e+00 3.02097738e-01
-1.85938001e+00 -1.19883919e+00 -4.76011992e-01 1.73834372e+00
3.54042739e-01 4.63025719e-01 5.23387671e-01 6.26777411e-01
5.82858443e-01 1.00235075e-01 -3.00820202e-01 -8.21538448e-01
2.97388565e-02 -3.05035152e-02 5.61695755e-01 2.89735705e-01
-1.05224466e+00 1.00606799e+00 6.55560207e+00 3.46279889e-01
-1.12264431e+00 1.84081346e-01 3.83787692e-01 1.90405637e-01
-1.01928256e-01 1.47599921e-01 -4.66758847e-01 4.09419268e-01
9.89032149e-01 2.16739848e-01 4.98046070e-01 1.47564006e+00
1.61284715e-01 -3.48104793e-03 -9.34186697e-01 5.97998261e-01
-2.74720907e-01 -1.54012907e+00 7.99798444e-02 3.28874826e-01
2.54679203e-01 -1.46934330e-01 8.11943784e-02 3.14631194e-01
8.01385760e-01 -8.32425654e-01 9.11270380e-01 1.40456393e-01
2.91603178e-01 -5.81673324e-01 5.46408951e-01 3.49768579e-01
-3.98320854e-01 -6.26049995e-01 -3.52056772e-01 -5.98527014e-01
-5.81775784e-01 7.19993949e-01 -1.50136375e+00 -2.92453598e-02
9.41375792e-01 3.60311300e-01 -1.04902363e+00 4.08688992e-01
-5.19708812e-01 7.99712121e-01 -1.45071939e-01 -3.23455155e-01
8.30182508e-02 3.59102339e-01 5.98043919e-01 1.40629995e+00
-1.15606748e-01 -4.49464947e-01 -2.24990353e-01 1.44121778e+00
1.21921279e-01 -3.86944354e-01 -9.75271821e-01 -1.35538623e-01
5.69632530e-01 1.21941721e+00 -7.24450350e-01 -4.08815712e-01
-2.19200775e-01 5.96105635e-01 3.73077393e-01 1.19559087e-01
-7.42995799e-01 3.87635082e-02 1.17277932e+00 3.37823898e-01
2.33194176e-02 9.72884521e-02 -8.12031388e-01 -1.03619719e+00
-7.62452558e-02 -1.15752852e+00 4.15903121e-01 -7.70962298e-01
-9.80259001e-01 5.84873736e-01 -2.93545481e-02 -9.86805737e-01
-2.35400423e-01 -1.08026218e+00 -6.07711434e-01 3.91831994e-01
-1.55056536e+00 -1.52509713e+00 -5.96844375e-01 2.85575747e-01
-9.51514468e-02 -4.56199616e-01 7.89406717e-01 -1.07407600e-01
-5.40097535e-01 3.78100127e-01 -4.50847656e-01 3.69131327e-01
3.06183368e-01 -1.15377343e+00 5.66087484e-01 9.53495502e-01
2.53899306e-01 5.59892356e-01 8.53574574e-01 -5.18727541e-01
-8.70832264e-01 -7.80920148e-01 7.15786457e-01 -9.00142968e-01
8.60935807e-01 -5.60130894e-01 -8.08698058e-01 4.16220486e-01
-3.58912461e-02 -1.84625760e-01 4.81627554e-01 2.73093283e-01
-1.17439592e+00 -1.79409757e-01 -1.48839331e+00 7.11036444e-01
9.83779490e-01 -6.09169424e-01 -4.51920986e-01 2.51491964e-01
3.69202226e-01 3.53661209e-01 -3.41709375e-01 3.14070523e-01
6.80229723e-01 -1.58234227e+00 1.10748720e+00 -1.04552960e+00
2.62097776e-01 -2.95527101e-01 9.07563493e-02 -9.43023860e-01
-1.90340400e-01 -3.42034191e-01 -3.11716080e-01 1.20736289e+00
4.59142923e-01 -9.30994391e-01 1.12732553e+00 6.82791471e-01
5.00451744e-01 -1.60606951e-01 -7.47315228e-01 -7.54539192e-01
7.09218681e-02 -6.44541681e-01 7.13868737e-01 1.06934512e+00
1.67419568e-01 -1.08122721e-01 -1.36886075e-01 3.60662103e-01
8.63065124e-01 -1.33910656e-01 1.05293906e+00 -1.71977651e+00
9.74346399e-02 -6.36999607e-01 -7.44091809e-01 -9.59382877e-02
2.56241500e-01 -5.93616188e-01 -5.68119109e-01 -1.18133736e+00
1.70574203e-01 -3.94855469e-01 -2.14907587e-01 8.47771406e-01
2.82251120e-01 1.56949818e-01 2.70358980e-01 3.07606876e-01
-1.92227244e-01 -3.17707896e-01 5.70514679e-01 -3.49282831e-01
2.53636211e-01 -1.70633823e-01 -9.71093476e-01 1.02733552e+00
1.19652307e+00 -8.04908633e-01 -1.77181229e-01 -2.40305409e-01
2.59324968e-01 -7.58715868e-01 1.23137534e+00 -1.14575863e+00
-2.43872572e-02 -1.18595988e-01 4.44240779e-01 -2.97440231e-01
3.65935229e-02 -8.70685399e-01 -2.03850523e-01 8.26521337e-01
-3.46173555e-01 -3.33063543e-01 1.60939783e-01 3.88782233e-01
2.19056591e-01 -1.84236884e-01 9.33879614e-01 -2.93378383e-01
-7.19203293e-01 -4.06093180e-01 -8.10851216e-01 -1.17622375e-01
9.76605415e-01 -2.14908078e-01 -1.05597401e+00 -2.46939123e-01
-2.10942894e-01 -3.99043202e-01 7.48224139e-01 2.80164450e-01
2.60146141e-01 -7.10085392e-01 -4.91027713e-01 2.67083585e-01
3.16612422e-01 -6.10828996e-01 -2.08075166e-01 4.15017545e-01
-7.12687790e-01 4.27927941e-01 -5.45781493e-01 -5.87158263e-01
-1.31576812e+00 7.34293044e-01 3.81666094e-01 -2.05101058e-01
-4.65515137e-01 5.38813293e-01 -5.10229282e-02 -2.85637408e-01
3.86690557e-01 -3.30751330e-01 -2.65736312e-01 -9.12575796e-03
6.11144066e-01 6.10137582e-01 5.29647060e-02 -1.84877917e-01
-6.93155527e-01 -5.81343770e-02 -7.38054067e-02 -2.44811526e-03
1.11141849e+00 2.70161450e-01 7.32320771e-02 -1.96046069e-01
6.63130403e-01 1.39826722e-02 -1.13206816e+00 7.13067129e-02
3.56222779e-01 -6.56630456e-01 7.69432485e-02 -1.38271046e+00
-8.48624706e-01 1.31095088e+00 6.43636048e-01 7.94386208e-01
8.57953072e-01 -9.94138345e-02 2.19982147e-01 6.24839842e-01
3.71284395e-01 -9.43643093e-01 1.75189465e-01 1.00208014e-01
4.80439126e-01 -1.37666309e+00 -3.21303010e-02 -3.03814471e-01
-4.60559011e-01 1.30200994e+00 7.67244816e-01 3.94709520e-02
2.64245540e-01 5.38049400e-01 4.00041342e-01 -4.31714773e-01
-8.47934544e-01 -2.80080169e-01 -1.93911597e-01 1.33531439e+00
3.97225060e-02 1.96606442e-01 -1.30441952e-02 4.76854831e-01
-9.07294303e-02 2.09384561e-02 6.54770672e-01 8.86514962e-01
-5.52898943e-01 -8.78538311e-01 -5.21971762e-01 -8.80428925e-02
-4.09343749e-01 1.48687109e-01 -1.24703085e+00 9.22898293e-01
2.90354252e-01 1.36161828e+00 -2.11636901e-01 -6.36209428e-01
2.78239369e-01 1.82838887e-01 1.93314120e-01 -1.56046525e-01
-6.88732326e-01 -7.45573223e-01 4.80492890e-01 -5.54163635e-01
-3.12792331e-01 -3.62874895e-01 -7.08021879e-01 -6.96041584e-01
-6.88974932e-02 -1.16572149e-01 1.00237978e+00 6.60876513e-01
4.89438236e-01 5.26587605e-01 4.65913773e-01 -5.07807672e-01
-4.32321787e-01 -7.13153124e-01 -2.69740988e-02 3.54191393e-01
2.70428836e-01 -8.37856412e-01 -5.74507654e-01 -6.58710524e-02] | [5.310491561889648, 7.228697776794434] |
ae65831c-da00-4b68-b1dc-12a15257dbc5 | action-concept-grounding-network-for | null | null | https://openreview.net/forum?id=4_57x7xhymn | https://openreview.net/pdf?id=4_57x7xhymn | Action Concept Grounding Network for Semantically-Consistent Video Generation | Recent works in self-supervised video prediction have mainly focused on passive forecasting and low-level action-conditional prediction, which sidesteps the problem of semantic learning. We introduce the task of semantic action-conditional video prediction, which can be regarded as an inverse problem of action recognition. The challenge of this new task primarily lies in how to effectively inform the model of semantic action information. To bridge vision and language, we utilize the idea of capsule and propose a novel video prediction model Action Concept Grounding Network (AGCN). Our method is evaluated on two newly designed synthetic datasets, CLEVR-Building-Blocks and Sapien-Kitchen, and experiments show that given different action labels, our ACGN can correctly condition on instructions and generate corresponding future frames without need of bounding boxes. We further demonstrate our trained model can make out-of-distribution predictions for concurrent actions, be quickly adapted to new object categories and exploit its learnt features for object detection. Additional visualizations can be found at https://iclr-acgn.github.io/ACGN/. | ['Animesh Garg', 'Wenxin Chen', 'Wei Yu'] | 2020-09-28 | null | null | null | null | ['video-prediction'] | ['computer-vision'] | [ 3.89779717e-01 9.65589285e-02 -4.95808005e-01 -3.34553838e-01
-4.22515064e-01 -2.56889462e-01 5.89258850e-01 -3.29625487e-01
-9.56095532e-02 5.80019951e-01 4.76646692e-01 -1.16696060e-01
2.31477439e-01 -5.58955848e-01 -1.02589893e+00 -4.26277846e-01
1.24128591e-02 2.49143511e-01 5.64642787e-01 7.12336972e-02
2.36838326e-01 1.36088908e-01 -1.88810277e+00 9.11079288e-01
5.41494548e-01 1.00749266e+00 5.59098721e-01 7.69476414e-01
1.40034249e-02 1.58255780e+00 -2.68797934e-01 -8.44903104e-03
1.57583594e-01 -6.41758680e-01 -7.93726146e-01 4.45188791e-01
2.72868484e-01 -4.44170415e-01 -3.97079110e-01 8.21275294e-01
-1.53287025e-02 3.28269273e-01 5.62024117e-01 -1.60563207e+00
-5.31530976e-01 5.68746269e-01 -1.72144011e-01 1.52043579e-02
5.94661295e-01 3.31605405e-01 8.57641697e-01 -6.72135770e-01
6.53107703e-01 1.27375638e+00 3.84279191e-01 9.69874203e-01
-1.02108526e+00 -6.39596224e-01 5.62850773e-01 7.64775038e-01
-9.20443773e-01 -4.84962225e-01 7.12188303e-01 -6.91250980e-01
6.95217431e-01 1.29083633e-01 8.42971444e-01 1.42451394e+00
-4.71126847e-02 1.37010384e+00 1.02527285e+00 -3.16414833e-01
3.59622985e-01 5.27398475e-02 -2.20754132e-01 7.98879802e-01
-2.29957163e-01 1.97862774e-01 -8.63929808e-01 4.83659774e-01
8.12971413e-01 2.77941287e-01 -3.71098876e-01 -5.81988335e-01
-1.47532916e+00 6.28316164e-01 1.28563389e-01 2.57324308e-01
-2.40794718e-01 4.75473881e-01 2.85131633e-01 9.59080178e-03
4.86265987e-01 -3.76472548e-02 -6.48250163e-01 -4.29865032e-01
-6.84384108e-01 5.16076311e-02 7.11105049e-01 1.12500906e+00
6.70082211e-01 -1.67634878e-02 -4.08247739e-01 4.46375996e-01
2.69907743e-01 5.41077495e-01 6.27680480e-01 -1.24731719e+00
2.43920729e-01 4.78751451e-01 8.53244737e-02 -8.43731999e-01
-1.94438040e-01 2.96145342e-02 -4.39999521e-01 1.09435931e-01
3.30052078e-01 -2.60438118e-02 -8.73842180e-01 1.51992428e+00
2.85851121e-01 8.54785979e-01 8.12204704e-02 8.89923871e-01
6.10220790e-01 7.17267394e-01 4.49753731e-01 -1.91357106e-01
1.04588294e+00 -1.34757829e+00 -6.37129486e-01 -4.17912573e-01
8.92182648e-01 -4.02666926e-01 1.10815430e+00 4.31767195e-01
-8.02947819e-01 -9.11259711e-01 -7.37174809e-01 4.58305143e-02
-2.91071862e-01 5.04536927e-01 7.20865071e-01 2.18685552e-01
-9.80981171e-01 6.34188056e-01 -1.20000422e+00 -5.71600795e-01
8.21409345e-01 1.08450115e-01 -4.08719838e-01 -1.40423149e-01
-7.36099124e-01 7.28296220e-01 7.20745623e-01 -6.27761185e-02
-1.40102875e+00 -5.85355163e-01 -7.95710385e-01 -1.77282393e-02
8.17556322e-01 -4.09898639e-01 1.28691101e+00 -1.62220562e+00
-1.66273403e+00 6.46853864e-01 -1.34227097e-01 -6.92816257e-01
4.98238057e-01 -4.66587484e-01 -3.85484904e-01 3.06214303e-01
1.72305614e-01 9.77448821e-01 1.00463319e+00 -1.24548221e+00
-8.84882867e-01 -1.96495876e-01 1.41107768e-01 1.52416930e-01
-3.37943435e-01 -1.50839895e-01 -4.06585217e-01 -5.42585492e-01
-8.65435526e-02 -1.04201221e+00 -2.13990986e-01 9.98158604e-02
-2.86453664e-01 -4.10505593e-01 9.06672657e-01 -7.27556407e-01
9.73318577e-01 -2.12576723e+00 2.45822817e-01 -3.15420866e-01
-1.26858920e-01 2.93632567e-01 -1.98859721e-01 2.79045105e-01
-2.47003809e-01 -2.06539914e-01 -2.69569661e-02 -2.94700891e-01
-6.43670633e-02 3.31201375e-01 -4.75708544e-01 3.01848680e-01
1.71354339e-01 8.86001885e-01 -1.02772129e+00 -5.51867902e-01
5.48714876e-01 2.52361894e-01 -6.56535923e-01 3.75284374e-01
-9.30954278e-01 7.94963300e-01 -5.70005000e-01 5.85868657e-01
2.42697626e-01 -4.76419270e-01 3.39111626e-01 -1.85034782e-01
-9.61048715e-03 -6.26291893e-03 -9.40585136e-01 2.07792139e+00
-3.55846643e-01 5.47567606e-01 -4.23795432e-01 -1.38288593e+00
6.94190919e-01 1.20548904e-01 5.86514473e-01 -6.55839443e-01
-2.00930629e-02 -1.68324336e-01 -2.73270816e-01 -7.77678907e-01
2.08823577e-01 1.32271141e-01 7.93132633e-02 3.52971226e-01
2.10673317e-01 2.01531604e-01 2.93200463e-01 2.17679337e-01
9.19823945e-01 1.12263489e+00 2.44293585e-01 -2.59275660e-02
7.17750192e-01 1.49133682e-01 6.61219835e-01 5.23676634e-01
-3.44792992e-01 3.82938892e-01 6.26016676e-01 -4.90157127e-01
-6.10444427e-01 -8.04869235e-01 2.66507387e-01 1.47028601e+00
2.37483740e-01 -6.18192434e-01 -8.00809979e-01 -1.08931971e+00
-2.37980023e-01 1.00323617e+00 -7.46948779e-01 -1.06351964e-01
-5.20283103e-01 -1.60102710e-01 -2.93903295e-02 7.77932465e-01
4.09133643e-01 -1.34090686e+00 -8.39973569e-01 6.71278080e-03
-4.36483085e-01 -1.41817713e+00 -2.05327436e-01 -9.76235792e-02
-8.31027567e-01 -1.29699516e+00 -4.38514173e-01 -7.41176307e-01
6.55076861e-01 2.48528898e-01 1.08941138e+00 1.50050566e-01
-2.22851679e-01 1.05008936e+00 -7.03425288e-01 -2.94570655e-01
-5.89752614e-01 -3.72204334e-01 5.44644520e-02 4.26703900e-01
5.88418961e-01 -4.87123847e-01 -7.88015187e-01 5.23784578e-01
-6.28495216e-01 7.12621033e-01 5.13044894e-01 5.90496123e-01
7.53730834e-01 -2.58719474e-01 3.68122637e-01 -8.40273619e-01
-3.12015444e-01 -4.12777007e-01 -5.64264178e-01 3.57642859e-01
-2.38265842e-01 -1.95871014e-02 4.96713310e-01 -4.45394874e-01
-1.22071385e+00 7.00899899e-01 1.84161827e-01 -7.50288248e-01
-5.24219096e-01 1.04656532e-01 -2.65152276e-01 3.26700777e-01
3.55072260e-01 5.16579211e-01 4.48676273e-02 -3.52987081e-01
6.16185665e-01 3.97805691e-01 4.43939745e-01 -5.04820108e-01
4.78171200e-01 6.29723430e-01 -7.19663203e-02 -6.34521544e-01
-1.02843022e+00 -3.76852423e-01 -8.73021603e-01 -6.60251796e-01
1.29610574e+00 -1.18912673e+00 -7.77474880e-01 2.50226766e-01
-1.02407670e+00 -9.01651561e-01 -3.62827897e-01 5.25873840e-01
-1.17707527e+00 4.11667138e-01 -3.33165705e-01 -6.77816153e-01
2.90892452e-01 -1.00484204e+00 9.70833302e-01 1.00163624e-01
-5.98171428e-02 -8.57330501e-01 -1.49253651e-01 6.39346719e-01
7.89106935e-02 4.37095053e-02 6.26702785e-01 -6.55949652e-01
-1.03798485e+00 -6.44189119e-02 -1.23647079e-02 6.26471758e-01
-1.55466115e-02 -1.67009607e-01 -8.71273398e-01 -1.14761472e-01
-2.18457237e-01 -6.23633146e-01 8.85347545e-01 2.66896397e-01
1.75263417e+00 -2.64817476e-01 -4.15428221e-01 4.83623594e-01
1.24166441e+00 1.52358562e-01 7.72951543e-01 1.06961861e-01
8.01973641e-01 7.17412889e-01 1.09768176e+00 5.26642561e-01
4.00470495e-01 7.58862674e-01 5.47813475e-01 2.81343520e-01
-4.19013381e-01 -6.53667152e-01 7.56271005e-01 6.59468234e-01
-3.86205852e-01 -2.10666642e-01 -7.63831556e-01 3.45551163e-01
-2.27223516e+00 -1.12968183e+00 -2.57617459e-02 1.91666663e+00
5.81718504e-01 -3.81864421e-02 1.80270776e-01 -2.02141345e-01
5.77087283e-01 1.27603799e-01 -5.54502010e-01 2.52186090e-01
2.62086481e-01 -1.14468895e-01 3.14507008e-01 2.82016695e-01
-1.39987159e+00 1.25791335e+00 5.12734699e+00 9.08781648e-01
-8.65066946e-01 2.33133912e-01 8.58834982e-01 -9.13902298e-02
1.06927127e-01 8.04146603e-02 -7.64879525e-01 5.84871590e-01
9.61700976e-01 9.06786099e-02 3.10642898e-01 1.14917469e+00
3.48768741e-01 -2.90350229e-01 -1.45903099e+00 9.89384353e-01
4.35410619e-01 -1.40468562e+00 1.98048174e-01 -2.96510309e-01
6.30425274e-01 -1.69660673e-01 -9.15295333e-02 6.13047540e-01
2.32815787e-01 -8.45024467e-01 7.57846653e-01 7.84761012e-01
7.89688706e-01 -2.81754076e-01 2.43006527e-01 4.02762204e-01
-1.11003995e+00 -3.60306859e-01 -3.22708398e-01 -2.64316857e-01
1.91984087e-01 -6.02144115e-02 -6.57830954e-01 2.50194341e-01
6.94120467e-01 1.52803326e+00 -5.16820908e-01 6.99288905e-01
-4.42476898e-01 7.27717221e-01 1.23800613e-01 6.53612018e-02
7.06317648e-02 -1.47007316e-01 2.73998260e-01 9.74529028e-01
3.24702084e-01 1.99938148e-01 5.12796760e-01 6.86115682e-01
1.69381335e-01 -4.25535114e-03 -4.18707520e-01 -1.89235657e-01
-4.99166362e-02 1.04284787e+00 -9.44019914e-01 -5.42793393e-01
-6.31647885e-01 1.20639050e+00 2.50151217e-01 4.52733457e-01
-1.13866556e+00 3.74699146e-01 6.02372229e-01 2.02578709e-01
6.57953620e-01 -6.49278164e-02 1.65019542e-01 -1.46429610e+00
-1.39667034e-01 -8.08801889e-01 1.92727581e-01 -9.37289536e-01
-8.24759662e-01 3.19694996e-01 2.69056447e-02 -1.68839347e+00
-3.91278654e-01 -9.50187564e-01 -3.85896832e-01 -1.85692981e-02
-1.24116063e+00 -1.34264076e+00 -4.88806725e-01 8.63164902e-01
1.00251412e+00 -3.67024064e-01 8.48676443e-01 2.03493416e-01
-4.41997528e-01 1.26862869e-01 -6.07961193e-02 1.55164734e-01
5.37034333e-01 -1.01932001e+00 -1.04104906e-01 9.33783472e-01
4.85965967e-01 -1.31784469e-01 6.03002608e-01 -7.20093787e-01
-1.35518754e+00 -1.38128769e+00 5.33503652e-01 -8.48080993e-01
7.30196893e-01 -4.76787180e-01 -5.40520966e-01 1.18947077e+00
-1.13653075e-02 2.60592610e-01 6.40264690e-01 -2.21337929e-01
-2.12121949e-01 -9.45894793e-02 -5.95113933e-01 5.81128359e-01
1.54498303e+00 -2.85521269e-01 -2.86115825e-01 9.45791900e-01
9.50092614e-01 -2.66577274e-01 -5.07105827e-01 2.82514006e-01
4.76714373e-01 -1.26993513e+00 1.01219916e+00 -9.36733425e-01
8.05186510e-01 -3.06939155e-01 -2.90120244e-01 -9.12545562e-01
-1.91196173e-01 -2.55996972e-01 -3.76146525e-01 9.65881467e-01
1.26677096e-01 -1.35719895e-01 1.06127048e+00 4.48605567e-01
-3.02378505e-01 -6.15669191e-01 -7.42987096e-01 -7.28232145e-01
-4.13540959e-01 -8.81528139e-01 2.46508062e-01 7.03597605e-01
6.88565755e-03 1.00810647e-01 -7.32163429e-01 -3.31214257e-02
4.02263403e-01 1.25073999e-01 9.19819057e-01 -8.32991600e-01
-6.71059966e-01 -1.08768739e-01 -6.66908085e-01 -1.33850694e+00
4.61753905e-01 -7.67933607e-01 2.76729882e-01 -1.42337334e+00
2.87711114e-01 -1.28966063e-01 -3.54118437e-01 6.92895293e-01
1.17305979e-01 2.51652449e-01 5.61664939e-01 1.13273099e-01
-1.32302582e+00 7.62396872e-01 1.24763143e+00 -1.62428036e-01
-6.58457503e-02 1.74250528e-01 -2.27213904e-01 9.64511216e-01
7.21307337e-01 -3.46638530e-01 -7.13078618e-01 -2.54243135e-01
-8.72924626e-02 2.56367713e-01 7.10067034e-01 -1.37084961e+00
1.47607893e-01 -4.27937716e-01 3.92870754e-01 -5.63736916e-01
3.74247134e-01 -9.80991900e-01 1.70421749e-02 5.38787067e-01
-5.66839516e-01 -2.85722971e-01 1.41893387e-01 9.94462252e-01
-1.21518508e-01 -3.15865986e-02 4.23575938e-01 -2.63719052e-01
-1.53080654e+00 4.02080566e-01 -4.55740541e-01 -1.20849490e-01
1.53079879e+00 -1.58269092e-01 -2.55835146e-01 -3.90294045e-01
-1.13329327e+00 2.14116827e-01 2.78128535e-01 6.95184767e-01
7.87684739e-01 -1.25186932e+00 -5.94269753e-01 1.82099700e-01
3.75893950e-01 -3.13790202e-01 5.81353128e-01 8.60747039e-01
-4.42359447e-01 3.39978665e-01 -3.07080239e-01 -7.68764377e-01
-1.25711060e+00 8.99948001e-01 1.43577099e-01 -9.54524428e-03
-8.06926310e-01 9.39779580e-01 5.51150918e-01 -1.40745968e-01
4.36039209e-01 -2.64745772e-01 -3.61920595e-01 -2.13629022e-01
6.04918063e-01 2.20846772e-01 -5.34715652e-01 -7.19368577e-01
-2.67439306e-01 3.54959011e-01 1.49754122e-01 2.09662035e-01
1.28116095e+00 -1.67335272e-01 1.36784958e-02 6.14228129e-01
9.16710496e-01 -5.32067776e-01 -1.84126568e+00 -1.18953429e-01
-1.46107981e-02 -5.63008666e-01 -1.75117046e-01 -8.93564165e-01
-1.09585404e+00 9.89254713e-01 8.00776124e-01 -1.75431162e-01
1.11703873e+00 2.43507251e-01 5.55734992e-01 3.63876134e-01
4.81126696e-01 -1.16403115e+00 6.65442228e-01 3.30010444e-01
8.25104058e-01 -1.45509982e+00 -1.01048179e-01 -3.90189290e-01
-1.01933897e+00 1.10916924e+00 1.07953131e+00 -1.26953155e-01
6.12591743e-01 -8.68152305e-02 -1.00051969e-01 6.34924918e-02
-9.05881047e-01 -2.88050532e-01 2.30929866e-01 7.40197480e-01
2.67305464e-01 2.38071024e-01 6.47808835e-02 6.17857456e-01
1.75954625e-01 4.91347730e-01 5.19873738e-01 8.36542487e-01
-3.99742484e-01 -9.68489349e-01 9.05573834e-03 4.73279238e-01
-1.08748168e-01 6.68741539e-02 1.08600073e-01 5.88783681e-01
3.12049031e-01 7.35594749e-01 2.50803322e-01 -6.49802983e-01
-2.37309411e-02 1.38080820e-01 4.28334773e-01 -7.65490472e-01
1.51472494e-01 -1.89491585e-02 9.35125425e-02 -1.23186755e+00
-9.94890511e-01 -8.47909749e-01 -1.17940378e+00 1.46016747e-01
1.29547536e-01 -9.34441462e-02 3.50607246e-01 1.05102503e+00
4.35423106e-01 5.85491061e-01 2.48588800e-01 -9.65880930e-01
-3.03325683e-01 -8.89024019e-01 -3.73173356e-01 6.48089290e-01
8.49388912e-02 -7.68495142e-01 -1.98502481e-01 6.94770217e-01] | [8.600412368774414, 0.5971205234527588] |
fb319644-b748-4165-affd-2c593e927f66 | coordinated-cyber-attack-detection-model-of | 2103.00133 | null | https://arxiv.org/abs/2103.00133v1 | https://arxiv.org/pdf/2103.00133v1.pdf | Coordinated Cyber-Attack Detection Model of Cyber-Physical Power System Based on the Operating State Data Link | Existing coordinated cyber-attack detection methods have low detection accuracy and efficiency and poor generalization ability due to difficulties dealing with unbalanced attack data samples, high data dimensionality, and noisy data sets. This paper proposes a model for cyber and physical data fusion using a data link for detecting attacks on a Cyber-Physical Power System (CPPS). Two-step principal component analysis (PCA) is used for classifying the system's operating status. An adaptive synthetic sampling algorithm is used to reduce the imbalance in the categories' samples. The loss function is improved according to the feature intensity difference of the attack event, and an integrated classifier is established using a classification algorithm based on the cost-sensitive gradient boosting decision tree (CS-GBDT). The simulation results show that the proposed method provides higher accuracy, recall, and F-Score than comparable algorithms. | ['Yang Li', 'Zhenming Zhang', 'Yunchang Dong', 'Xiaoyong Bo', 'Zhaoyang Qu', 'Pengcheng Xu', 'Lei Wang'] | 2021-02-27 | null | null | null | null | ['cyber-attack-detection'] | ['miscellaneous'] | [-1.74133643e-03 -6.59957051e-01 -3.11612070e-01 -1.28182739e-01
-4.26366299e-01 -4.19423640e-01 1.56988487e-01 6.50291085e-01
4.91249524e-02 5.28925538e-01 -1.88447252e-01 -4.35293496e-01
-4.33704615e-01 -9.63016629e-01 2.51727611e-01 -9.35924768e-01
-2.41843298e-01 1.39579907e-01 3.68219048e-01 -9.12429020e-02
4.94435310e-01 6.81731820e-01 -1.30530035e+00 1.84323773e-01
7.69387305e-01 1.30353820e+00 -3.60941082e-01 3.67346138e-01
2.57610917e-01 4.26557392e-01 -1.10355568e+00 -1.88170206e-02
5.26191056e-01 -4.10112888e-01 -1.49267763e-01 -2.51111835e-01
-8.70190620e-01 -1.26706108e-01 -2.21841112e-01 1.25294149e+00
4.24662918e-01 7.98454881e-02 5.32044530e-01 -1.95130599e+00
2.78756469e-01 1.88526779e-01 -6.55583799e-01 5.34650743e-01
4.80164915e-01 2.89898850e-02 4.94920135e-01 -5.41737676e-01
-1.77043378e-02 1.12487042e+00 5.70391655e-01 5.42864129e-02
-1.04045641e+00 -8.93407643e-01 -3.20166387e-02 6.66631401e-01
-1.20562005e+00 4.46924627e-01 1.21626341e+00 -2.61014879e-01
8.75592768e-01 4.85728681e-01 8.66527796e-01 5.73060095e-01
7.90779412e-01 5.49648345e-01 1.13313842e+00 -5.35193980e-01
7.63529181e-01 8.55351388e-02 6.76258326e-01 1.21649519e-01
6.37282252e-01 3.50045800e-01 -1.37031212e-01 -6.08107746e-01
7.28436783e-02 3.56985062e-01 -4.22649533e-02 -1.05844826e-01
-4.48487014e-01 8.76096487e-01 1.96350947e-01 5.09004295e-01
-6.99623585e-01 -7.58596718e-01 8.49517167e-01 4.09152985e-01
4.05590206e-01 3.62427264e-01 -7.25660324e-01 -2.40235031e-01
-3.68574023e-01 1.14901587e-01 7.41401374e-01 3.76672626e-01
2.04202354e-01 4.95819896e-01 2.10116267e-01 3.72759730e-01
1.18644312e-01 3.88544381e-01 6.93181276e-01 -1.89654410e-01
4.51835006e-01 9.55225527e-01 -2.13018015e-01 -1.26151967e+00
-5.02022088e-01 -4.67836708e-01 -8.53848100e-01 5.34979939e-01
-1.03908479e-01 -3.90634060e-01 -4.53149766e-01 1.23369610e+00
5.38821638e-01 1.67276897e-02 3.11050713e-01 4.59671825e-01
-9.88769755e-02 6.54998600e-01 2.83687174e-01 -7.25073159e-01
1.22164643e+00 -8.31208974e-02 -1.02185261e+00 1.67197868e-01
4.35843527e-01 -3.37971509e-01 5.19195795e-01 7.74443746e-01
-4.05731976e-01 -4.47680831e-01 -1.52311909e+00 1.10460722e+00
-6.77886128e-01 -3.69939357e-02 4.30385739e-01 1.09390604e+00
1.63988471e-02 4.52640802e-01 -6.87437236e-01 -3.09437863e-04
2.54573047e-01 1.87863290e-01 -1.64874047e-01 1.02571085e-01
-1.44867826e+00 1.01469994e+00 7.10346162e-01 -2.19705850e-01
-4.42276150e-01 -6.89205408e-01 -6.20583117e-01 1.73314184e-01
1.06076993e-01 -1.09905280e-01 6.22899473e-01 -6.36260211e-01
-1.27994049e+00 -3.30241807e-02 6.38810277e-01 -5.86680174e-01
2.06333324e-02 1.51524156e-01 -1.07353878e+00 2.34660074e-01
-2.05959007e-01 -5.75253785e-01 8.13633740e-01 -7.55479455e-01
-8.26027930e-01 -1.01412404e+00 -3.87690634e-01 1.87452808e-01
-4.28005278e-01 1.86037734e-01 5.76268256e-01 -5.64363301e-01
3.66981089e-01 -3.51369619e-01 -3.49531412e-01 -7.73047745e-01
-2.43975177e-01 -9.78022665e-02 1.55674803e+00 -1.00552654e+00
1.62945759e+00 -2.29720521e+00 -1.14810854e-01 7.39665508e-01
-2.68565834e-01 6.88339889e-01 3.74610186e-01 6.53627932e-01
-4.79420394e-01 -3.82150620e-01 -2.10827008e-01 4.64749008e-01
-1.46158129e-01 -3.92747782e-02 -4.51816410e-01 4.79284406e-01
-5.31475507e-02 -3.28076482e-02 -6.51093364e-01 -6.44950941e-02
7.10865796e-01 3.38550620e-02 -1.31018564e-01 3.86970073e-01
2.66769469e-01 -4.98790927e-02 -5.74734449e-01 7.45038688e-01
7.43502557e-01 4.75751996e-01 3.40068132e-01 -3.37954521e-01
2.30609223e-01 -6.07019104e-02 -1.59149706e+00 6.21901631e-01
-1.86309293e-01 1.44819379e-01 7.99726844e-02 -1.38410592e+00
1.39840424e+00 4.30323601e-01 6.90097213e-01 -7.22803831e-01
4.18395281e-01 -5.15821278e-02 -8.09431076e-03 -3.04132611e-01
-3.87094431e-02 -7.33599951e-03 -4.46105093e-01 2.44657591e-01
-1.32461965e-01 -2.19190344e-01 1.75559178e-01 1.32601976e-01
1.29636443e+00 -5.48296511e-01 7.72782803e-01 -2.56465971e-01
7.51424193e-01 3.39860059e-02 1.07914984e+00 1.00832522e-01
-2.80969679e-01 -3.73503357e-01 7.20983863e-01 -5.09830296e-01
-5.92716694e-01 -9.65871453e-01 -1.71912163e-01 2.65109420e-01
2.03348771e-01 -2.67333746e-01 -5.28516769e-01 -9.68730986e-01
2.48286277e-01 1.14592791e+00 -2.35538229e-01 -7.66607046e-01
-1.46098152e-01 -1.44073284e+00 2.95382261e-01 4.52014059e-01
4.88686055e-01 -7.05227077e-01 -6.53128505e-01 4.71373737e-01
5.05784094e-01 -7.48237014e-01 3.54298711e-01 2.72450060e-01
-8.45832348e-01 -1.59652209e+00 4.09653604e-01 -3.31121415e-01
7.45903492e-01 3.28796767e-02 3.87273192e-01 -1.56297877e-01
-6.12782121e-01 1.57683745e-01 -4.79741901e-01 -8.63198221e-01
-5.09361267e-01 -4.57606375e-01 3.89450401e-01 -3.95347290e-02
5.92274487e-01 -3.87471557e-01 -2.62061894e-01 9.84983593e-02
-7.56684244e-01 -5.96759140e-01 4.50686365e-01 8.84251833e-01
1.12184847e-03 1.10920739e+00 1.11946452e+00 -3.94825488e-01
1.13436723e+00 -6.22180820e-01 -9.77811575e-01 1.95930660e-01
-1.17267418e+00 -4.43012297e-01 9.58864987e-01 -3.16115737e-01
-1.00318563e+00 -9.57116764e-03 3.32227387e-02 -2.35704675e-01
-2.42215127e-01 3.78514558e-01 -5.47752857e-01 -8.59815404e-02
4.80313331e-01 1.78058490e-01 3.37728590e-01 -3.53038609e-01
-3.30316871e-01 8.64272594e-01 2.09977031e-01 -1.66568443e-01
7.00719953e-01 -2.87614148e-02 3.37501109e-01 -5.19969225e-01
-3.45347404e-01 -4.21385288e-01 -4.33810860e-01 -5.41186750e-01
2.30729356e-01 -6.66679502e-01 -8.96673679e-01 7.31158614e-01
-5.95451236e-01 3.77030849e-01 -1.74625173e-01 9.03165042e-01
2.12120041e-02 5.19940257e-01 -6.93905294e-01 -1.21901357e+00
-6.13464355e-01 -9.17766869e-01 3.59413922e-01 2.52473712e-01
2.74391174e-02 -7.04138577e-01 1.52265146e-01 8.20621550e-02
1.99015737e-01 6.59564674e-01 9.85587239e-01 -1.23549128e+00
1.51558131e-01 -1.24580610e+00 1.78457469e-01 8.48372698e-01
2.77030617e-01 1.26587570e-01 -5.06914377e-01 -5.57819486e-01
4.99680907e-01 5.77311814e-02 -1.47204071e-01 1.43263698e-01
1.13405776e+00 -2.06847250e-01 -4.01061386e-01 2.51321290e-02
1.61160493e+00 1.10588014e+00 4.34718728e-01 2.03410938e-01
2.60251969e-01 4.52979416e-01 1.19971597e+00 7.01033533e-01
-1.71826124e-01 3.61289501e-01 4.40704614e-01 1.72034234e-01
5.71862519e-01 1.44004375e-01 2.97844350e-01 8.61421287e-01
3.98934394e-01 1.47156879e-01 -8.92477214e-01 1.24383166e-01
-1.64931500e+00 -9.59804058e-01 -1.49843961e-01 2.24594760e+00
3.40013862e-01 5.74244618e-01 2.33690262e-01 1.30629516e+00
9.01727140e-01 -1.55770794e-01 -4.33927834e-01 -5.87644815e-01
3.67073566e-02 -1.17198778e-02 3.93507659e-01 2.67418265e-01
-1.07058704e+00 1.15548171e-01 5.88367510e+00 1.01540995e+00
-9.62865472e-01 -7.36418664e-02 5.62401474e-01 2.58681476e-01
4.90305632e-01 7.68259913e-02 -5.78601241e-01 8.51222217e-01
1.09396315e+00 -4.87446874e-01 7.16592148e-02 1.04475379e+00
9.28985476e-02 -2.04546094e-01 -4.92817611e-01 7.57997930e-01
9.39010829e-02 -8.29714179e-01 -3.60125870e-01 2.61682458e-02
3.17175031e-01 -4.89174604e-01 -3.81279588e-01 3.43432397e-01
3.78042966e-01 -7.62545168e-02 1.47126049e-01 3.84954840e-01
1.33641049e-01 -1.16447902e+00 1.00705457e+00 4.32742923e-01
-1.23881853e+00 -8.94407988e-01 1.14210956e-01 -3.35515440e-01
7.10906275e-03 9.31774616e-01 -7.99649060e-01 1.01989925e+00
5.76529741e-01 3.60685259e-01 -2.85822511e-01 9.41840827e-01
-1.52238935e-01 6.70706868e-01 -4.47284937e-01 -3.26850414e-01
-2.84125358e-01 -2.65030891e-01 7.48197317e-01 6.63830280e-01
1.15838110e-01 3.16753566e-01 4.96082544e-01 3.96103859e-02
6.20237827e-01 6.99298084e-02 -4.44674224e-01 2.42641494e-01
1.05618358e+00 1.20692909e+00 -7.56485105e-01 -5.44262350e-01
-1.77225709e-01 1.60628304e-01 -3.46193284e-01 -6.96233660e-02
-6.90389991e-01 -8.36884439e-01 3.75262886e-01 -3.45302731e-01
-1.30790979e-01 1.40780598e-01 -6.63097858e-01 -7.74225414e-01
-6.40795240e-03 -9.07015800e-01 1.00072205e+00 -3.65734786e-01
-1.63671446e+00 3.04904342e-01 1.90194547e-01 -1.62816536e+00
-3.36194396e-01 -6.00829065e-01 -8.53400886e-01 7.05363929e-01
-7.90496707e-01 -6.33310020e-01 -1.04088746e-01 6.60138249e-01
3.55704933e-01 -7.30970383e-01 9.56063330e-01 2.74146438e-01
-1.00312734e+00 2.50373065e-01 9.43879485e-02 2.48623341e-01
3.89295034e-02 -8.99898291e-01 -2.53385156e-01 1.26348901e+00
-6.71320200e-01 1.39708798e-02 8.27689409e-01 -8.49171340e-01
-1.40609574e+00 -7.30422854e-01 2.88836807e-01 9.96786654e-02
7.80018866e-01 -2.12590113e-01 -7.93487251e-01 1.55107692e-01
-5.78475744e-02 -6.21060319e-02 8.98651004e-01 -7.04894438e-02
-3.19514543e-01 -6.38352990e-01 -1.86496043e+00 1.83049515e-01
-7.84494877e-02 -3.29556972e-01 -8.71859312e-01 2.12681338e-01
2.40051106e-01 8.51128995e-02 -1.27279532e+00 6.85863554e-01
1.54812694e-01 -6.10095084e-01 8.93217802e-01 -6.43789828e-01
-5.83066940e-01 -4.50358629e-01 -1.64801002e-01 -1.41354465e+00
-4.37898785e-01 -2.17367381e-01 -1.27319023e-01 9.08184290e-01
4.41750288e-02 -9.72848177e-01 4.97055382e-01 3.91899109e-01
-1.80712286e-02 -7.55048871e-01 -1.23707998e+00 -7.10119247e-01
-4.60413456e-01 -3.08753908e-01 8.61503482e-01 1.34483063e+00
7.76292264e-01 2.88609594e-01 1.09235458e-01 5.42868853e-01
9.33220744e-01 -4.29456384e-04 3.55410010e-01 -1.34871984e+00
5.89052998e-02 -3.06927353e-01 -8.78072560e-01 4.79676098e-01
-8.59837532e-02 -1.55266330e-01 -4.33466554e-01 -9.30339515e-01
-2.00959489e-01 -4.29598331e-01 -6.69757426e-01 4.41632658e-01
-3.25769216e-01 -1.81540519e-01 1.98421210e-01 -1.19329607e-02
-2.03135461e-01 5.54631531e-01 3.48780483e-01 -2.78105959e-02
-2.19733968e-01 3.95084381e-01 -4.88080569e-02 6.08578205e-01
1.25419009e+00 -4.36391950e-01 -3.85118097e-01 6.27035081e-01
-3.46139967e-01 7.20445633e-01 -5.07908389e-02 -1.31314349e+00
2.16827214e-01 -1.91752896e-01 6.08587086e-01 -7.60711849e-01
2.33724028e-01 -1.17748749e+00 2.77282655e-01 1.29599607e+00
9.48366299e-02 4.03388619e-01 2.09916010e-01 6.83131397e-01
-4.37594950e-01 -1.91301137e-01 7.36284673e-01 4.05258000e-01
-4.63638306e-01 9.77891162e-02 -6.46965683e-01 -6.75948024e-01
1.68804789e+00 -1.02038801e-01 -2.70307213e-01 1.42356157e-01
-4.97278422e-01 -5.14682978e-02 -1.03785053e-01 5.79049885e-01
7.24950433e-01 -1.39185238e+00 -4.26072478e-01 5.70482433e-01
2.46782415e-02 -7.05832601e-01 2.74533898e-01 4.74250615e-01
-1.94491714e-01 1.79186091e-01 -4.81341422e-01 -2.13678837e-01
-1.51328588e+00 8.50534260e-01 -1.67300738e-02 -5.48096418e-01
-3.47988844e-01 6.46231472e-02 -7.75078893e-01 -1.19939961e-01
1.36991933e-01 2.67110765e-01 -4.72053617e-01 2.34516278e-01
6.69302523e-01 8.43951762e-01 4.83360440e-01 -3.07488292e-01
-5.20270407e-01 8.27263966e-02 9.40162763e-02 1.93056077e-01
1.14011467e+00 2.96916157e-01 -1.90757304e-01 3.20702255e-01
9.12555993e-01 -3.49082679e-01 -8.13923120e-01 6.55273423e-02
7.21897036e-02 -6.59715116e-01 2.84752399e-01 -9.26508188e-01
-1.01256216e+00 4.83054578e-01 8.54417264e-01 8.68559539e-01
1.63358903e+00 -7.46339858e-01 3.71923596e-01 4.94342595e-02
4.94412810e-01 -1.33135355e+00 -1.41930401e-01 1.12820603e-01
4.45408821e-01 -7.14762926e-01 2.66144693e-01 -3.33450466e-01
-6.04203105e-01 1.25383842e+00 4.68101770e-01 -4.24826682e-01
1.15592432e+00 5.74270368e-01 -2.93111861e-01 2.67979838e-02
-9.20031726e-01 4.98323143e-01 -2.54096687e-01 7.39759028e-01
-4.40205038e-01 4.82897401e-01 -7.71941066e-01 8.79593253e-01
9.99467596e-02 -3.07906777e-01 3.99235487e-01 1.33937252e+00
-4.53151613e-01 -9.44821715e-01 -8.34226608e-01 6.03514254e-01
-3.77279788e-01 5.81311464e-01 -9.80805084e-02 5.94041288e-01
3.51265930e-02 1.23594487e+00 6.86056986e-02 -8.59343767e-01
5.55047631e-01 7.92778209e-02 -1.69432303e-03 -1.37949875e-02
-6.30602062e-01 -1.96625367e-01 -1.43283173e-01 -5.80090165e-01
3.32126729e-02 -8.26219797e-01 -1.06015217e+00 -2.05627665e-01
-5.65879703e-01 7.17183709e-01 1.26362956e+00 7.97038913e-01
3.39539975e-01 8.43676209e-01 1.51747763e+00 -4.56807941e-01
-9.35608804e-01 -9.02559876e-01 -8.47343743e-01 3.52998704e-01
-1.95281446e-01 -7.66687036e-01 -8.66847694e-01 -5.30372918e-01] | [6.157215118408203, 2.5664052963256836] |
fd197b49-b5f7-4d43-b1be-4b2e16da4c01 | longitudinal-performance-of-iris-recognition | 2303.12720 | null | https://arxiv.org/abs/2303.12720v1 | https://arxiv.org/pdf/2303.12720v1.pdf | Longitudinal Performance of Iris Recognition in Children: Time Intervals up to Six years | The temporal stability of iris recognition performance is core to its success as a biometric modality. With the expanding horizon of applications for children, gaps in the knowledge base on the temporal stability of iris recognition performance in children have impacted decision-making during applications at the global scale. This report presents the most extensive analysis of longitudinal iris recognition performance in children with data from the same 230 children over 6.5 years between enrollment and query for ages 4 to 17 years. Assessment of match scores, statistical modelling of variability factors impacting match scores and in-depth assessment of the root causes of the false rejections concludes no impact on iris recognition performance due to aging. | ['Stephanie Schuckers', 'Michael Schuckers', 'Masudul H Imtiaz', 'Laura Holsopple', 'Naveen G Venkataswamy', 'Priyanka Das'] | 2023-03-10 | null | null | null | null | ['iris-recognition'] | ['computer-vision'] | [ 2.1563347e-01 -2.3512866e-01 -6.0090286e-01 -4.1510874e-01
-2.1302201e-01 -3.7097827e-01 2.2846647e-01 6.2738466e-01
-4.4306427e-01 4.2136005e-01 4.9929547e-01 -7.8978467e-01
-5.4897106e-01 -4.9106306e-01 -5.2191126e-01 -4.5884138e-01
-2.1902658e-01 6.3571453e-02 -3.0574816e-01 3.1568873e-01
3.7158775e-01 5.1816314e-01 -1.8671103e+00 -1.3085851e-01
1.1125659e+00 2.9975808e-01 -9.0508372e-01 8.8654160e-01
3.5797924e-01 3.2526881e-01 -5.8308733e-01 -5.6175512e-01
1.2010952e-01 -1.6917795e-01 -4.0596679e-01 1.3993859e-01
1.2503854e+00 -7.2293848e-01 -2.9059255e-01 4.4731250e-01
8.9170045e-01 -1.0502100e-01 5.5119354e-01 -6.3125193e-01
-7.0463264e-01 4.3151104e-01 -6.9467288e-01 3.0520308e-01
6.0848063e-01 4.3274757e-01 1.2127597e-01 -3.9373651e-01
2.5908464e-01 7.9230738e-01 6.1798805e-01 6.8292826e-01
-1.5036343e+00 -7.0325369e-01 -2.1742646e-01 -3.3444658e-01
-1.4036255e+00 -8.1755191e-01 -2.2684368e-01 -7.7790761e-01
8.6437708e-01 4.0530246e-01 9.9137473e-01 1.5965280e-01
-1.7949191e-01 2.0103274e-01 1.2695657e+00 -6.3669115e-01
-3.4328565e-01 -6.4876720e-02 4.3661541e-01 1.0453538e-01
6.1758906e-01 7.8586173e-01 -5.5026382e-01 -4.6227906e-02
7.3545170e-01 -2.8820264e-01 1.4242189e-02 1.7485982e-01
-6.0056973e-01 8.3708562e-02 -1.0382257e-01 3.6297596e-01
-1.5534376e-01 -3.5626072e-01 1.4753033e-01 5.0771648e-01
2.7032086e-01 1.7577575e-01 -3.2431605e-01 -5.3651035e-01
-1.0346582e+00 1.6416147e-01 4.1799590e-01 3.7970814e-01
7.8783512e-02 -5.6678481e-02 -3.6482882e-01 8.0981386e-01
2.1184626e-01 7.9169482e-01 1.5092186e-02 -3.6457008e-01
9.6092276e-02 9.1068661e-01 1.6130032e-01 -3.6490479e-01
-5.9680814e-01 -7.6387781e-01 -3.8536915e-01 3.4079236e-01
9.9475735e-01 -3.1538862e-01 -1.0227172e+00 1.4523220e+00
3.9705643e-01 1.7342052e-01 -6.6435777e-02 3.6543739e-01
7.5356883e-01 -3.7525800e-01 1.3339265e-01 -4.1274667e-01
1.0474837e+00 4.8311293e-04 -2.2296248e-01 2.0879422e-01
4.5623004e-01 -1.1412815e+00 2.4702661e-01 4.8035437e-01
-1.2448417e+00 -2.5070059e-01 -7.3429859e-01 1.8998188e-01
-6.0639817e-02 2.8425482e-01 5.8356702e-01 1.7172163e+00
-1.1699461e+00 4.4129321e-01 -8.7879539e-01 -7.0486635e-01
3.8433662e-01 1.1424608e+00 -3.1185648e-01 6.1940666e-02
-4.5161217e-01 9.8918641e-01 8.5812267e-03 1.9000877e-02
1.1750002e-01 -1.0191350e+00 -5.1295102e-01 -2.9642719e-01
-2.9147655e-01 -3.1159806e-01 8.8581997e-01 -5.9329182e-01
-1.1065710e+00 1.1213845e+00 -5.4154080e-01 -2.0751035e-01
3.1266090e-02 -8.8584153e-03 -6.5276819e-01 -2.9131305e-01
-1.2469379e-01 5.7769772e-02 3.5745904e-01 -4.4996285e-01
-6.8915039e-01 -9.9739122e-01 -4.1428137e-01 2.1095194e-01
-3.6640887e-03 7.6111501e-01 -5.9512325e-02 -5.4589039e-01
2.4338587e-01 -7.1270454e-01 4.6611909e-02 -5.2952158e-01
2.2019546e-01 -2.6285076e-01 -1.1120157e-01 -8.4034646e-01
1.5154543e+00 -2.2208209e+00 -5.8056277e-01 4.6516296e-01
1.8550117e-01 5.0846040e-01 -5.8402214e-02 4.6297112e-01
-5.7189894e-01 3.1067967e-02 5.9479010e-01 8.1408352e-02
-5.6580764e-01 -3.2960558e-01 3.0889053e-02 8.1375521e-01
4.9889907e-02 7.8653800e-01 -5.5742711e-01 1.2722679e-02
5.0802183e-01 4.6809277e-01 -2.9611865e-01 2.5069430e-02
5.4534572e-01 4.6167782e-01 -1.9651638e-02 1.3181366e+00
7.4592060e-01 5.6404758e-02 -1.5680985e-01 3.6077857e-01
-6.6462219e-01 4.2225605e-01 -1.0228235e+00 9.9104649e-01
3.5628724e-01 4.1164780e-01 -8.3883554e-02 -3.5834062e-01
8.2458311e-01 4.6223974e-01 5.1005101e-01 -7.4717152e-01
3.0641274e-03 3.5082984e-01 5.5543470e-01 -7.6916337e-02
2.2511807e-01 -1.6927068e-01 7.5817508e-01 4.5996124e-01
-2.3657998e-01 2.7787530e-01 1.9582769e-01 -2.9339141e-01
5.8599585e-01 -3.1464857e-01 1.7248952e-01 -3.1447756e-01
4.1873622e-01 -9.9001788e-02 4.5684990e-01 7.1545213e-01
-1.7403737e-01 3.9506641e-01 3.0271813e-01 -4.7674358e-01
-6.6092491e-01 -1.0030566e+00 -1.1959175e+00 4.5221099e-01
-5.0323731e-01 -6.2901765e-02 -3.9102021e-01 -3.7190389e-02
4.4648936e-01 2.8168219e-01 -8.0798346e-01 4.6485581e-02
-9.5652103e-02 -1.1406778e+00 7.0652759e-01 2.0775591e-01
-1.2960598e-01 -2.4186003e-01 -4.3103766e-01 1.0930537e-01
5.5562663e-01 -4.3145880e-01 -3.9017156e-01 -5.0175393e-01
-1.3476539e+00 -1.5053462e+00 -1.0176573e+00 -4.2004186e-01
1.0120088e+00 -6.4681463e-02 8.8693517e-01 6.3847184e-01
-4.9385279e-01 7.3937255e-01 3.6084335e-02 -8.9212734e-01
-1.2556970e-01 -1.5048994e-01 5.6115574e-01 -9.8760374e-02
9.9277127e-01 -3.5902056e-01 -7.4697340e-01 1.8512870e-01
-5.0198925e-01 -4.9503914e-01 4.2900050e-01 6.9464672e-01
1.2687023e-01 -1.9521905e-01 3.3705890e-01 -5.4369861e-01
1.4673565e-01 -2.0311905e-01 -7.9299635e-01 4.7808263e-01
-1.4130840e+00 -3.8001782e-01 -6.0686767e-01 -5.2707422e-01
-8.1573951e-01 -4.1646358e-01 3.2360810e-01 2.5441113e-01
-4.6416631e-01 2.7576378e-01 6.3767213e-01 -5.8287013e-01
6.6336143e-01 -1.5567431e-01 3.8826844e-01 -4.1771132e-01
-3.1425977e-01 7.0973569e-01 5.7995605e-01 -7.8456175e-01
5.3908324e-01 5.7563353e-02 2.9196352e-01 -1.0135853e+00
-3.2370353e-01 -5.8536875e-01 -4.3243396e-01 -2.8645569e-01
2.9862532e-01 -1.0319304e+00 -1.1481329e+00 1.2794954e+00
-3.9361194e-01 -1.7879741e-01 -3.2869603e-02 1.0343137e+00
3.1480808e-02 -6.5482654e-02 -2.6852173e-01 -1.0816803e+00
-5.5504382e-01 -1.0507792e+00 3.4024334e-01 9.9556190e-01
-3.0650517e-01 -6.9393653e-01 4.0135792e-01 6.0979050e-01
2.7118507e-01 3.8502112e-01 8.2020128e-01 -4.2089337e-01
-5.6499523e-01 -4.5849872e-01 -2.0947470e-01 -9.2659198e-02
3.1347811e-01 7.8903091e-01 -7.6539737e-01 -7.0969433e-01
-4.8562786e-01 5.6980163e-02 2.6714393e-01 1.1178772e+00
3.4311268e-01 3.7925743e-02 -3.2308370e-01 6.7421782e-01
1.4768633e+00 7.0287985e-01 7.8836721e-01 3.4729043e-01
9.9046268e-02 8.5718137e-01 1.9549753e-01 1.0512295e-01
3.4406036e-01 4.7735798e-01 -1.4401574e-01 -2.1119815e-01
-1.2582181e-01 -2.4392860e-01 -6.9028735e-02 1.1670872e-01
-5.9068054e-01 3.4840101e-01 -1.2964910e+00 9.5026124e-01
-1.0786793e+00 -7.0110369e-01 -2.9865992e-01 3.0488632e+00
7.3032397e-01 -9.7574495e-02 6.9737428e-01 2.6275385e-02
8.1102115e-01 -3.9706987e-01 -6.4276296e-01 -6.4290893e-01
-2.0434552e-01 5.4044330e-01 7.5868654e-01 6.0199136e-01
-4.8701423e-01 4.5972157e-01 8.3691578e+00 2.0009856e-01
-9.0661997e-01 -5.4470849e-01 1.0184944e+00 -4.5820901e-01
4.2536702e-02 2.1177241e-01 -8.9479172e-01 1.4832488e-01
9.9658227e-01 -9.8053806e-02 3.5603082e-01 -2.6338136e-01
2.9261512e-01 -6.4372259e-01 -9.2317885e-01 6.0872281e-01
-1.2427168e-01 -9.0226799e-01 -5.2052259e-01 3.7771749e-01
9.7351116e-01 3.8388610e-02 6.6789210e-01 -3.3919457e-01
7.2641067e-02 -1.3861108e+00 1.3590233e-01 8.2518965e-01
1.5389562e+00 -5.0686401e-01 6.3058788e-01 -2.6177365e-01
-6.6891772e-01 -3.7130271e-03 2.2823484e-01 -6.1400217e-01
-4.2117295e-01 5.1690173e-01 -1.0105523e+00 8.8160746e-02
4.5612523e-01 3.6597204e-01 -7.6562536e-01 1.4231328e+00
1.7224295e-01 9.9369800e-01 -5.2721912e-01 2.9122260e-01
-3.3261797e-01 -1.8395513e-01 4.0397942e-01 7.0119995e-01
1.4710879e-01 6.5056813e-01 -7.8214085e-01 5.0863540e-01
5.6077451e-01 2.9176474e-01 -2.8362828e-01 -3.8651991e-01
5.7114828e-01 3.8816816e-01 -4.4908038e-01 6.0356755e-02
-1.0389357e+00 -1.1402166e-01 -1.6155207e-01 5.2876204e-01
2.7661577e-01 -4.3620691e-02 1.0490876e+00 6.8871421e-01
-1.8491401e-01 9.8925091e-02 -1.0815259e+00 -7.2194457e-01
2.3592624e-03 -1.2889363e+00 6.1652845e-01 -1.0896794e-01
-6.7430288e-01 -3.6425972e-01 -2.7793097e-01 -8.0961996e-01
-3.0420190e-01 -1.8290615e-01 -7.1319824e-01 1.8174329e+00
-8.7338036e-01 -9.0097153e-01 3.3334240e-01 1.3121580e-01
-2.6280198e-01 -3.1923944e-01 7.7572817e-01 1.1213458e-01
-8.6615735e-01 1.4087237e+00 3.5812140e-02 2.7201006e-01
8.4766018e-01 -1.0777702e+00 4.7303301e-01 1.0398721e+00
-2.9431063e-01 1.3898247e+00 4.3194488e-01 -1.0409014e+00
-1.0006926e+00 -1.0607149e-01 1.1562104e+00 -8.6323959e-01
1.5852043e-01 5.6162804e-01 -6.2811160e-01 4.3544903e-01
-5.8419827e-02 -8.2066327e-01 1.1385069e+00 1.0431241e+00
-1.7159498e-01 -3.4330419e-01 -1.1663147e+00 4.4772089e-01
6.0973418e-01 -6.0006541e-01 -3.2140526e-01 -3.4851325e-01
9.7084768e-02 -8.9987254e-01 -1.5718563e+00 8.5027039e-01
1.2913442e+00 -1.1024559e+00 8.5223341e-01 -6.3140380e-01
2.0675575e-02 -2.1568650e-01 4.5730188e-01 -4.9497741e-01
-1.5475757e-01 -8.3801365e-01 2.8487456e-01 1.4053669e+00
5.8959341e-01 -8.9951986e-01 1.0568875e+00 1.7437093e+00
4.9071369e-01 -9.3174738e-01 -9.2067009e-01 -3.3168250e-01
2.8510496e-01 -7.6945134e-02 9.2453796e-01 5.5387378e-01
1.7616974e-01 -4.7532347e-01 -3.1609923e-02 5.4156613e-01
7.5760406e-01 -2.0004671e-02 7.4902624e-01 -1.1857951e+00
-9.5751852e-02 -6.8431121e-01 -9.7207695e-01 -3.7125280e-01
-6.9929421e-01 -2.5371590e-01 -9.5134723e-01 -1.0169793e+00
2.3744409e-01 -5.5617517e-01 -3.5350907e-01 5.5514991e-01
-4.6039432e-01 -1.4801972e-01 -5.8502457e-03 -1.2025953e-02
4.4080088e-01 -5.2022433e-01 8.7391454e-01 4.4560019e-02
-6.5322578e-01 6.7765468e-01 -8.1054980e-01 7.4994624e-02
6.5370840e-01 -5.1433332e-03 -3.1009471e-01 -2.7563989e-01
1.6905929e-01 4.0854225e-01 -1.5769298e-01 -1.0287522e+00
3.6819080e-01 -2.5100330e-01 6.7703295e-01 -6.7627764e-01
-3.3212000e-01 -2.9733536e-01 4.1644952e-01 5.7898885e-01
-1.7429654e-01 2.8577703e-01 6.6525847e-01 -2.8834915e-01
6.1995294e-02 -4.3649718e-02 6.4130443e-01 6.1119097e-01
1.2032309e-02 3.5874122e-01 -1.8705727e-01 5.6971803e-02
5.4265153e-01 -6.5681201e-01 -2.9370946e-01 -7.5347379e-02
-8.5904634e-01 3.5957372e-01 9.9127603e-01 3.4388873e-01
1.3822716e-01 -8.7910277e-01 -8.9537615e-01 8.3044094e-01
2.2551879e-01 -3.3654544e-01 2.5819647e-01 1.1513370e+00
-4.1747689e-01 6.6269749e-01 -8.1429616e-02 -5.3293592e-01
-1.9693923e+00 2.6746285e-01 5.9002614e-01 2.8650498e-01
-1.5437745e-01 9.7345459e-01 -2.5294888e-01 1.5227972e-01
5.7889915e-01 -2.1319093e-01 -2.1543176e-01 6.6683055e-03
9.4140881e-01 8.6595756e-01 6.3744903e-02 -6.2020999e-01
-1.6781835e-01 9.0296072e-01 -3.9955136e-01 -1.8393826e-01
8.5563046e-01 -1.5252347e-01 -5.5679363e-01 2.2469848e-01
4.8989764e-01 2.3653207e-03 -7.7423525e-01 -2.0325726e-01
-2.2879870e-01 -8.0441105e-01 -8.0862321e-02 -1.2174191e+00
-8.5928547e-01 2.4906190e-01 1.2066367e+00 -3.9287806e-01
1.2469965e+00 -2.6931232e-01 4.9152657e-02 -3.6902288e-01
8.6427674e-02 -7.7200770e-01 -9.8953176e-01 -6.1198615e-02
3.9419293e-01 -9.7776890e-01 3.6622456e-01 -6.1664987e-02
2.0282860e-01 8.3975405e-01 5.5012494e-01 4.9473017e-01
5.5607831e-01 -5.9271775e-02 5.2988011e-01 -1.3296185e-01
-4.2753220e-01 -1.6471487e-01 7.2725463e-01 8.0866712e-01
9.9529910e-01 5.6166369e-01 -1.0459689e+00 -1.1397016e-02
-2.0516342e-01 3.1983811e-01 5.9197098e-01 3.9654920e-01
1.6134189e-01 -1.6400729e+00 -6.8490309e-01 1.1264517e+00
-8.6556846e-01 -9.8881863e-02 -3.8439557e-01 5.3360647e-01
4.0882510e-01 1.2221593e+00 9.9472508e-02 -3.0853677e-01
4.6842718e-01 3.9952356e-01 6.2991297e-01 -4.9933967e-01
-7.9666203e-01 4.3910719e-02 2.6499438e-01 -1.4631510e-01
-6.9769627e-01 -1.6031319e+00 -6.1671788e-01 -7.3776865e-01
-5.4340357e-01 -8.5438035e-02 6.8108803e-01 9.0711963e-01
1.7028956e-01 -4.4015363e-02 3.8961914e-01 3.6417130e-01
-3.7657982e-01 -8.5088837e-01 -9.6351862e-01 -1.9811121e-01
9.3698716e-01 -3.6849907e-01 -2.7277157e-01 -3.5783997e-01] | [3.748286247253418, -3.6258020401000977] |
ea0b0874-9889-444f-a283-efce6d2c9ded | molecular-docking-and-binding-mode-analysis | 2004.06447 | null | http://arxiv.org/abs/2004.06447v1 | http://arxiv.org/pdf/2004.06447v1.pdf | Molecular docking and binding mode analysis of selected FDA approved drugs against COVID-19 selected key protein targets: An effort towards drug repurposing to identify the combination therapy to combat COVID-19 | The emergence of COVID-19 has severely compromised the arsenal of antiviral
and antibiotic drugs. Drug discovery is a multistep process with a high failure
rate, high cost and it takes approximately 10-12 years for the development of
new molecules into the clinical candidate. On the other side, drug repurposing
also called old drugs for new uses, is an attractive alternative approach for a
new application of marketed FDA approved or investigational drugs. In the
current pandemic situation raised due to COVID-19, repurposing of existing FDA
approved drugs are emerging as the first line of the treatment. The causative
viral agent of this highly contagious disease and acute respiratory syndrome
coronavirus (SARS-CoV) shares high nucleotide similarity. Therefore, many
existing viral targets are structurally expected to be similar to SARS-CoV and
likely to be inhibited by the same compounds. Here, we selected three viral key
proteins based on their vital role in viral life cycle: ACE2 (helps in entry
into the human host), viral nonstructural proteins RNA-dependent RNA polymerase
(RdRp) NSP12, and NSP16 which helps in replication, and viral latency (invasion
from immunity). The FDA approved drugs chloroquine (CQ), hydroxychloroquine
(HCQ), remdesivir (RDV) and arbidol (ABD) are emerging as promising agents to
combat COVID-19. Our hypothesis behind the docking studies is to determine the
binding affinities of these drugs and identify the key amino acid residues
playing a key role in their mechanism of action. The docking studies were
carried out through Autodock and online COVID-19 docking server. Further
studies on a broad range of FDA approved drugs including few more protein
targets, molecular dynamics studies, in-vitro and in-vivo biological evaluation
are required to identify the combination therapy targeting various stages of
the viral life cycle. | [] | 2020-04-14 | null | null | null | null | ['molecular-docking'] | ['medical'] | [ 1.93461210e-01 -6.03354871e-01 -1.10868707e-01 7.68125728e-02
1.09552278e-03 -1.00880504e+00 1.98085561e-01 4.64286834e-01
-3.20148498e-01 1.23223007e+00 -1.69246960e-02 -6.98642135e-01
2.95836311e-02 -3.89958918e-01 -1.94822341e-01 -8.12073946e-01
-2.39381716e-01 7.21535802e-01 -4.92250286e-02 -1.78513736e-01
1.63886711e-01 9.90832329e-01 -6.49658740e-01 -1.35859802e-01
1.27784383e+00 5.80749214e-02 4.56702739e-01 5.53826272e-01
5.65254912e-02 -1.46778509e-01 -3.43832999e-01 1.11508086e-01
2.12563157e-01 -4.91731167e-01 -2.82103062e-01 -4.83152211e-01
-4.25465822e-01 -3.66894841e-01 5.28412163e-01 6.45187795e-01
3.63007396e-01 -1.26140133e-01 8.93549263e-01 -9.79111731e-01
-3.74361604e-01 -5.78574896e-01 -5.36496282e-01 2.18584672e-01
3.70762289e-01 3.41836423e-01 6.18336678e-01 -1.15286815e+00
8.05318594e-01 9.39317167e-01 5.21614969e-01 7.10905850e-01
-1.01193178e+00 -4.34218168e-01 -4.56557602e-01 1.58053383e-01
-1.38705778e+00 3.62575091e-02 1.44617110e-01 -9.78314400e-01
1.51127136e+00 1.82406560e-01 6.27381861e-01 5.91969252e-01
8.02843213e-01 -8.10080394e-02 1.00170743e+00 2.04352379e-01
2.23319948e-01 1.11948580e-01 4.16927300e-02 5.85335791e-01
7.75381684e-01 1.27473459e-01 8.85966048e-02 -9.71625268e-01
5.33812582e-01 6.51118219e-01 -5.64224780e-01 -3.62407237e-01
-6.75840616e-01 1.01900017e+00 2.05771439e-02 3.15354556e-01
-6.24247670e-01 -4.71689165e-01 3.89681906e-01 -1.50872171e-01
-1.28500640e-01 3.76485497e-01 -1.02740335e+00 8.56413096e-02
-3.10337156e-01 2.35105708e-01 2.79465735e-01 -1.16089679e-01
6.13733590e-01 -1.78075135e-01 1.93735212e-01 4.71967965e-01
7.06458926e-01 8.53657007e-01 -4.74889167e-02 2.69152038e-02
-1.09921522e-01 6.56351328e-01 4.10949975e-01 -6.11580193e-01
-7.29190052e-01 -1.42551720e-01 -7.47239292e-01 5.13080284e-02
3.64810497e-01 -3.97407919e-01 -8.87530744e-01 1.82800162e+00
5.20157874e-01 2.70617217e-01 2.62206256e-01 8.67183566e-01
3.95777911e-01 1.04980528e+00 3.81817639e-01 -1.09999168e+00
1.82384551e+00 -4.19899255e-01 -4.80746686e-01 3.57589811e-01
2.11790994e-01 -8.55729997e-01 4.15331185e-01 2.79936135e-01
-4.54717159e-01 -1.09203957e-01 -1.02091479e+00 6.50137484e-01
-2.83162028e-01 -1.75605342e-01 4.19364005e-01 7.93391705e-01
-6.44815922e-01 4.33307678e-01 -1.02094197e+00 -7.58892596e-01
1.03762798e-01 7.56202996e-01 -3.91479403e-01 -1.35305920e-03
-1.15386236e+00 1.10495830e+00 1.89542621e-01 4.56770137e-02
-8.81161332e-01 -6.48432910e-01 -3.52603287e-01 -2.38768458e-01
-9.83684212e-02 -9.88743901e-01 3.98416251e-01 -1.97034389e-01
-1.48019469e+00 6.18612945e-01 -3.17204267e-01 -1.96613416e-01
-2.05047369e-01 -1.57422602e-01 -2.50522673e-01 1.26942456e-01
-1.12742759e-01 1.43441513e-01 3.81002694e-01 -8.91501546e-01
-3.02548915e-01 -8.20985556e-01 -2.38873705e-01 1.57541752e-01
4.57642049e-01 6.04861021e-01 1.62523612e-01 -5.95256090e-01
-3.47195327e-01 -1.43397868e+00 -4.24511969e-01 -6.01068199e-01
-1.32851779e-01 -3.50388944e-01 8.94038022e-01 -3.21659476e-01
7.62457728e-01 -1.73182607e+00 2.30929613e-01 3.30702484e-01
1.15702987e-01 9.28426385e-01 4.10355181e-02 1.19077229e+00
-5.81586137e-02 3.43370050e-01 -7.38139823e-02 8.02916467e-01
-6.40311182e-01 -2.64928162e-01 -2.56194890e-01 9.16135430e-01
-2.60432884e-02 6.81903958e-01 -8.73438418e-01 2.04484552e-01
-1.56138539e-01 7.91623592e-01 -3.52950782e-01 2.83934683e-01
-5.19276321e-01 9.56518352e-01 -9.03101504e-01 6.61393821e-01
1.08272600e+00 -2.13849232e-01 7.39100277e-01 -1.48538738e-01
-5.53523675e-02 7.01336712e-02 -4.61118042e-01 8.77507150e-01
3.12860042e-01 1.12218045e-01 -1.94194049e-01 -2.17671335e-01
5.91813982e-01 8.45640182e-01 4.38964456e-01 -2.51240820e-01
1.66423783e-01 3.77688587e-01 2.01657712e-01 -3.75257879e-01
-1.34926140e-01 -4.90875930e-01 6.24689162e-01 1.84199110e-01
-6.14647329e-01 5.99152744e-01 1.88101344e-02 -3.46114412e-02
8.31099510e-01 9.62594226e-02 6.73529208e-01 -3.82465750e-01
7.37533808e-01 2.85621077e-01 8.55668962e-01 -1.00145945e-02
-1.47062510e-01 9.06493515e-02 2.42698416e-01 -4.46700186e-01
-1.05804551e+00 -9.08367455e-01 -3.88375640e-01 4.75554466e-01
9.80916023e-02 -2.66471922e-01 -6.53959453e-01 -3.54093730e-01
-3.04444194e-01 3.22738141e-01 -2.42882654e-01 3.17589998e-01
-5.60304105e-01 -8.85970592e-01 5.54984272e-01 -1.56071499e-01
5.94841316e-02 -7.53933549e-01 -6.33858263e-01 4.53229934e-01
1.41585693e-01 -7.21737385e-01 -2.70729959e-01 5.55924289e-02
-9.31242228e-01 -1.65681887e+00 -6.17338479e-01 -6.53947592e-01
5.08931637e-01 2.11578384e-01 1.87707216e-01 3.41600180e-01
-1.80653661e-01 -4.05773669e-01 -2.35268041e-01 -5.98778069e-01
-4.83464301e-01 -4.92423564e-01 6.90250576e-01 -5.23185074e-01
5.89317203e-01 -4.10353333e-01 -1.04243350e+00 2.98389375e-01
-6.13442361e-01 -1.38014406e-01 3.45282197e-01 6.60836816e-01
4.06400561e-01 -1.94140017e-01 7.80330956e-01 -7.97025979e-01
7.29154408e-01 -7.15104640e-01 -6.68196976e-01 6.87989369e-02
-5.74536085e-01 -2.96753854e-01 7.92824209e-01 5.84767899e-03
-9.17738855e-01 -1.55485541e-01 -2.63121307e-01 5.82568571e-02
-6.96084052e-02 5.55491328e-01 -1.28981963e-01 2.76430368e-01
4.48689580e-01 1.59661382e-01 4.35693443e-01 -5.07191896e-01
-2.91116893e-01 5.39794385e-01 -1.67866573e-01 -1.31235331e-01
9.24648762e-01 4.98629779e-01 1.92376986e-01 -8.63209009e-01
-2.75984202e-02 -6.34997189e-01 1.48245424e-01 2.08265260e-01
1.42936492e+00 -8.19857359e-01 -1.29077661e+00 5.68931341e-01
-1.47628987e+00 1.89031512e-01 8.25296938e-01 8.43494594e-01
-2.96580009e-02 4.88950878e-01 -7.57485092e-01 -5.32264829e-01
-8.23943675e-01 -1.38460958e+00 3.03717673e-01 4.02068704e-01
-1.35265693e-01 -7.60887086e-01 1.03994894e+00 4.96127009e-01
3.40974420e-01 5.03781676e-01 1.53430223e+00 -9.87421513e-01
-7.06953764e-01 2.17510924e-01 -1.62456349e-01 1.22959994e-01
5.10419548e-01 3.61017525e-01 -3.48998070e-01 -5.70902467e-01
-9.63804424e-02 -1.68153569e-01 2.99445212e-01 5.78084409e-01
-1.93059698e-01 -3.35978717e-01 -8.57221603e-01 4.55687881e-01
1.75259304e+00 1.27476275e+00 9.25870478e-01 1.67547822e-01
4.76686954e-01 3.67082059e-01 8.19837153e-01 4.98261303e-01
-1.87869489e-01 8.65833879e-01 4.23563421e-01 -1.67864487e-02
6.62842214e-01 2.07514942e-01 2.82672137e-01 3.83276731e-01
-7.07731009e-01 -4.76005465e-01 -9.03345585e-01 4.43820953e-02
-1.35985124e+00 -1.02262914e+00 -4.19239610e-01 2.35919094e+00
9.12797749e-01 -3.80870312e-01 9.66997221e-02 -6.38735473e-01
4.88751143e-01 -4.53464359e-01 -6.03278577e-01 -7.15642095e-01
-3.67861390e-02 3.70690554e-01 2.38856643e-01 4.83887315e-01
-6.89827085e-01 5.41567326e-01 5.76888943e+00 4.50842261e-01
-1.10037577e+00 -1.61581069e-01 2.60687798e-01 4.92732108e-01
-1.98813707e-01 5.06570876e-01 -7.85793066e-01 2.52066046e-01
8.99816394e-01 -1.92143172e-02 2.80749202e-01 4.16946262e-01
8.40843081e-01 -2.06611633e-01 -6.83683038e-01 7.27454007e-01
-3.38506252e-01 -1.62232280e+00 -5.30238077e-02 4.86819327e-01
6.76881313e-01 2.11126894e-01 -3.47724795e-01 -2.75712192e-01
-1.98507383e-01 -1.02650797e+00 -3.27062279e-01 2.52030194e-01
7.25274742e-01 -9.96916950e-01 1.01346064e+00 2.20887512e-01
-1.07056761e+00 3.06967437e-01 -2.98183709e-01 1.66072145e-01
4.20299739e-01 2.24291012e-01 -1.30125439e+00 8.91498923e-02
1.89484820e-01 1.59552947e-01 -7.43915215e-02 8.61913800e-01
-1.18881352e-01 2.25557491e-01 -7.85733089e-02 -6.97604492e-02
1.74962610e-01 -5.55795491e-01 7.37322927e-01 9.14051652e-01
-6.26867861e-02 6.67869210e-01 1.62219867e-01 4.15884912e-01
2.16166794e-01 3.59086037e-01 -5.50927997e-01 -3.38875502e-01
2.38370165e-01 9.81675565e-01 -8.15214157e-01 -2.57872969e-01
-3.65532845e-01 7.92268455e-01 -2.70992607e-01 5.05769551e-01
-7.68829167e-01 -1.89861059e-01 1.48789012e+00 2.75691718e-01
2.54622787e-01 -2.34083787e-01 4.84109998e-01 -5.66305935e-01
-7.02959657e-01 -1.32037687e+00 1.53679594e-01 -3.04641694e-01
-9.53256667e-01 6.65176928e-01 -3.17466334e-02 -1.04581690e+00
-1.37361795e-01 -5.55329382e-01 -6.64265931e-01 1.13669658e+00
-1.17775321e+00 -9.53173757e-01 1.46843463e-01 3.70702893e-01
4.88060862e-01 -2.62181699e-01 1.18864572e+00 -3.13757733e-03
-4.34213042e-01 -2.00947672e-01 6.72990263e-01 -4.10837591e-01
6.15742207e-01 -6.12201452e-01 -9.56402719e-02 4.11278009e-01
-6.85572743e-01 1.41994071e+00 1.08087945e+00 -1.23399007e+00
-1.47014487e+00 -7.98030376e-01 8.06859016e-01 -1.98289335e-01
3.72710139e-01 -5.32949157e-02 -6.50189459e-01 5.01157284e-01
7.10430861e-01 -8.39286447e-01 1.40419841e+00 -4.45608765e-01
-1.94522813e-01 5.87605596e-01 -1.30251014e+00 7.20476866e-01
2.24384740e-01 -1.40584275e-01 -4.09833670e-01 4.67312843e-01
5.72593391e-01 7.30902469e-03 -5.47132373e-01 7.83466995e-01
7.63433754e-01 -9.81139362e-01 7.35285342e-01 -9.12745118e-01
-3.47369075e-01 -1.03785813e+00 9.80474651e-02 -1.10885525e+00
-3.46264690e-01 -7.83847451e-01 4.15335089e-01 6.68183982e-01
3.52270424e-01 -6.89404368e-01 4.73573923e-01 2.16588929e-01
2.61045128e-01 -9.88287270e-01 -6.42171800e-01 -5.97060263e-01
-3.23713422e-01 3.49113196e-01 2.25526616e-01 7.81020105e-01
4.23303731e-02 7.76920199e-01 -5.52998304e-01 1.07513130e-01
2.47499853e-01 4.52162065e-02 7.14630306e-01 -1.36059058e+00
-1.66905716e-01 -3.60525101e-02 -2.22100541e-01 -4.60941494e-01
-4.52010274e-01 -5.25577545e-01 -7.29144752e-01 -1.68753135e+00
6.46139979e-01 -2.69582272e-01 -3.52673270e-02 2.01284602e-01
2.35065781e-02 3.01567689e-02 1.08319089e-01 3.44018757e-01
1.40191615e-01 1.51938736e-01 8.82068574e-01 9.04415622e-02
-7.64676094e-01 1.03623077e-01 -4.24989551e-01 6.45736873e-01
9.07274187e-01 -6.29048705e-01 -7.15637445e-01 4.77715850e-01
6.21872425e-01 1.09065361e-01 -2.52129197e-01 2.47393679e-02
-2.89999485e-01 -8.62286866e-01 1.28780112e-01 -9.98896241e-01
3.30417633e-01 -8.14362228e-01 9.55320656e-01 1.13550258e+00
8.35747659e-01 2.74422437e-01 3.22060645e-01 3.68385434e-01
2.58769125e-01 -1.25896469e-01 1.02539539e+00 -2.83388570e-02
-2.48739347e-01 2.40903571e-01 -1.10334182e+00 -5.70588335e-02
1.31976092e+00 -7.09734440e-01 -3.58814508e-01 1.51193887e-01
-5.75031877e-01 -1.09938256e-01 7.97470152e-01 2.79489160e-01
7.12414622e-01 -6.49572790e-01 -8.37166190e-01 4.47868481e-02
7.86732212e-02 -6.60394788e-01 4.20580953e-01 1.05556631e+00
-1.30175436e+00 8.87338877e-01 -3.37975383e-01 -4.61966962e-01
-1.66641927e+00 9.96656060e-01 3.12964112e-01 -3.17121506e-01
-1.23999618e-01 4.66812521e-01 5.51122427e-01 7.87270591e-02
-1.22394942e-01 2.05180019e-01 -5.20199180e-01 -1.31637320e-01
6.72276199e-01 3.71994406e-01 -2.44776141e-02 -8.13493848e-01
-1.09975910e+00 7.14879155e-01 -2.54931509e-01 7.68727541e-01
1.39329934e+00 9.59911048e-02 -6.08212471e-01 -2.02345237e-01
1.17412567e+00 4.33855742e-01 -5.80477715e-01 1.81692988e-01
-1.70642927e-01 -3.17947984e-01 -6.54379547e-01 -1.00204551e+00
-3.59225601e-01 1.71628669e-01 8.82731438e-01 -3.98099184e-01
9.26186442e-01 -3.06721381e-03 9.30843174e-01 6.84619918e-02
5.40978074e-01 -5.85534036e-01 -1.72695324e-01 4.37321037e-01
7.79543519e-01 -9.69688833e-01 -2.17425060e-02 -4.33118969e-01
-5.76177299e-01 8.60702455e-01 4.54993248e-01 2.14494497e-01
4.78149503e-01 6.25997484e-02 2.31971264e-01 -4.51966375e-01
-9.20216441e-01 1.19513929e-01 1.16749227e-01 6.48705423e-01
7.53988683e-01 1.09394521e-01 -1.33573449e+00 2.11046875e-01
6.88764989e-01 5.58213741e-02 4.45395023e-01 7.39030123e-01
-5.90184391e-01 -1.66318870e+00 -4.77642149e-01 8.58668014e-02
-7.28805780e-01 -3.00881445e-01 -4.21886027e-01 6.79045737e-01
2.47554556e-01 7.33702183e-01 -3.49147290e-01 9.19122174e-02
5.02019264e-02 2.06925958e-01 2.65137345e-01 -6.18910730e-01
-5.68630457e-01 5.80627620e-01 -3.63252349e-02 4.61672693e-02
-4.12988782e-01 -4.07169610e-01 -1.95393062e+00 -4.60519046e-01
-3.77141118e-01 5.57191312e-01 7.62683034e-01 8.79162371e-01
6.33961082e-01 -2.02785313e-01 7.99655616e-01 -2.48093531e-01
-1.15579695e-01 -7.82662451e-01 -7.24238217e-01 2.21906025e-02
2.50537604e-01 -4.60896105e-01 -6.05179742e-02 -1.47332400e-02] | [4.695308685302734, 5.105510711669922] |
1dedb041-9e81-485f-8d94-647153ffd13a | a-discourse-aware-attention-model-for | 1804.05685 | null | http://arxiv.org/abs/1804.05685v2 | http://arxiv.org/pdf/1804.05685v2.pdf | A Discourse-Aware Attention Model for Abstractive Summarization of Long Documents | Neural abstractive summarization models have led to promising results in
summarizing relatively short documents. We propose the first model for
abstractive summarization of single, longer-form documents (e.g., research
papers). Our approach consists of a new hierarchical encoder that models the
discourse structure of a document, and an attentive discourse-aware decoder to
generate the summary. Empirical results on two large-scale datasets of
scientific papers show that our model significantly outperforms
state-of-the-art models. | ['Walter Chang', 'Seokhwan Kim', 'Nazli Goharian', 'Franck Dernoncourt', 'Trung Bui', 'Doo Soon Kim', 'Arman Cohan'] | 2018-04-16 | a-discourse-aware-attention-model-for-1 | https://aclanthology.org/N18-2097 | https://aclanthology.org/N18-2097.pdf | naacl-2018-6 | ['unsupervised-extractive-summarization'] | ['natural-language-processing'] | [ 3.66248190e-01 8.41426015e-01 -4.33018744e-01 -2.96779960e-01
-1.17490232e+00 -4.51598883e-01 8.68189037e-01 6.25048995e-01
-1.61933526e-01 1.23618579e+00 1.12853909e+00 -2.71019340e-01
1.68142229e-01 -4.25638229e-01 -8.67650032e-01 -1.23980559e-01
1.77152410e-01 4.88497406e-01 1.37462825e-01 3.30456980e-02
9.26489294e-01 -1.63897797e-02 -1.23396337e+00 8.03964734e-01
1.05949819e+00 3.36308002e-01 4.21807617e-01 1.13979876e+00
-4.37604070e-01 1.13155937e+00 -1.45396090e+00 -3.76448303e-01
-5.53194404e-01 -8.51542950e-01 -1.18043125e+00 1.75815061e-01
6.97631836e-01 -5.16170144e-01 -4.23229128e-01 8.67083371e-01
5.63137412e-01 1.92357212e-01 7.34501481e-01 -5.47038078e-01
-8.76456082e-01 1.33481872e+00 -4.44310099e-01 5.13277531e-01
5.93211532e-01 -2.33678520e-01 1.22383726e+00 -5.79950750e-01
1.09741843e+00 1.40326536e+00 2.28434876e-01 7.04760969e-01
-1.12625730e+00 -7.83804283e-02 3.47310096e-01 1.48802653e-01
-7.19836652e-01 -7.30899453e-01 8.05657864e-01 -1.44105777e-01
1.42165422e+00 5.07890165e-01 7.61255443e-01 1.34638739e+00
7.23860204e-01 1.18530118e+00 3.31270099e-01 -3.33622426e-01
3.80714118e-01 -2.44938239e-01 6.64544821e-01 3.05492431e-01
7.41484404e-01 -8.08473945e-01 -6.52914166e-01 -1.55316204e-01
2.17064485e-01 -5.90582371e-01 -1.52250156e-01 5.45615613e-01
-1.04496348e+00 7.31095731e-01 8.15870091e-02 2.63171613e-01
-7.29238629e-01 1.05263598e-01 6.88136220e-01 7.98849687e-02
7.58059919e-01 9.31352675e-01 2.98802424e-02 -1.98531792e-01
-1.47526157e+00 8.08252752e-01 1.15382850e+00 1.09056807e+00
-5.10481298e-02 2.14416549e-01 -8.99754822e-01 5.79369426e-01
1.09506212e-01 1.35531038e-01 4.19241577e-01 -1.26591587e+00
7.50326037e-01 4.88470107e-01 2.20807061e-01 -6.82308435e-01
-2.06071794e-01 -4.90360677e-01 -7.17729628e-01 -5.52117884e-01
-3.17878276e-01 -2.72445112e-01 -7.13766634e-01 1.41497290e+00
-2.47616634e-01 -1.56964287e-01 5.03381073e-01 4.83791947e-01
1.67755651e+00 1.35829759e+00 -1.03019737e-03 -7.90200651e-01
1.11279202e+00 -1.21696186e+00 -1.14904547e+00 -3.50326359e-01
3.35605234e-01 -5.47798336e-01 5.21259665e-01 2.66519785e-01
-2.08196831e+00 -2.98577219e-01 -1.28218627e+00 -6.25149250e-01
1.25499731e-02 2.42751062e-01 2.38464296e-01 -7.13771880e-02
-1.15012658e+00 6.52853727e-01 -7.10197210e-01 -3.76936674e-01
6.96910679e-01 -5.28640999e-03 1.14436544e-01 2.35964030e-01
-8.93963516e-01 9.75369930e-01 7.77000487e-01 -1.28924772e-01
-9.15952623e-01 -6.89159334e-01 -7.83425808e-01 4.92212117e-01
2.63928056e-01 -1.22231114e+00 1.86532104e+00 -4.34182107e-01
-1.64671803e+00 6.96265101e-01 -5.68797827e-01 -8.34591269e-01
2.21536472e-01 -4.66959983e-01 1.23945950e-02 6.24448299e-01
2.70074010e-01 6.32144451e-01 5.17249107e-01 -1.32952130e+00
-3.86764616e-01 -1.63624883e-01 4.00175638e-02 4.27042246e-01
-3.67246598e-01 2.86121935e-01 -2.70845145e-01 -6.40734851e-01
-3.84535462e-01 -3.61058027e-01 -2.30659410e-01 -9.24748063e-01
-1.03091586e+00 -6.44297957e-01 5.43405116e-01 -9.84720111e-01
1.75905859e+00 -1.42856050e+00 7.14588165e-01 -6.42992735e-01
2.16892824e-01 2.68640488e-01 -2.17187732e-01 7.27236629e-01
1.37732044e-01 4.06659245e-01 -2.84571767e-01 -6.73900783e-01
-4.59089391e-02 -2.43745465e-02 -6.46347821e-01 -9.95343924e-02
3.37088257e-01 1.14904070e+00 -1.09235048e+00 -8.62332940e-01
-2.58798510e-01 -2.55042668e-02 -3.99539381e-01 3.99009436e-01
-8.30559671e-01 2.56529272e-01 -7.73693144e-01 2.53836751e-01
1.91864744e-01 -4.05351311e-01 9.36394334e-02 1.50661275e-01
-3.80430192e-01 8.91096234e-01 -4.23392415e-01 2.01438880e+00
2.00438164e-02 1.03731823e+00 5.47760986e-02 -9.42361891e-01
6.12254739e-01 4.62851137e-01 -6.76749423e-02 -3.68022531e-01
1.94788739e-01 -6.00320473e-03 -1.75705820e-01 -6.09826744e-01
1.13436437e+00 2.30818883e-01 -2.33428285e-01 5.78312576e-01
1.54504374e-01 -5.07170379e-01 9.98989582e-01 8.21672082e-01
1.19075763e+00 -1.65997267e-01 5.07965326e-01 -3.94940615e-01
4.05256897e-01 1.43915921e-01 2.27993548e-01 1.12876344e+00
2.66889483e-01 7.92436779e-01 9.64299321e-01 -8.72266442e-02
-1.21240771e+00 -8.11160386e-01 1.90543756e-01 7.71553516e-01
-1.74304873e-01 -9.38947797e-01 -1.05834711e+00 -4.47125852e-01
-2.25492254e-01 1.39067411e+00 -5.39429247e-01 -1.15254804e-01
-1.00170422e+00 -5.14256716e-01 5.65432727e-01 4.69723940e-01
2.87914693e-01 -1.27262425e+00 -7.53925860e-01 3.13367575e-01
-3.52257967e-01 -9.77713168e-01 -2.90917635e-01 -8.76027122e-02
-1.09509051e+00 -5.31819284e-01 -8.67177010e-01 -8.06544125e-01
4.94606912e-01 6.50316402e-02 1.41108179e+00 -1.83605943e-02
7.12480769e-02 8.22452158e-02 -1.35448366e-01 -8.82087648e-01
-8.50193083e-01 5.95892191e-01 -4.40889388e-01 -5.50060451e-01
-6.14618212e-02 -3.27605158e-01 -2.64410585e-01 -7.77450562e-01
-9.70655859e-01 3.30647141e-01 6.49904191e-01 6.46027148e-01
2.74395317e-01 -4.61849749e-01 1.09794927e+00 -9.45466936e-01
1.56551397e+00 -4.46616232e-01 -9.24396291e-02 3.91843200e-01
-2.57814020e-01 1.41092896e-01 6.84411466e-01 -1.26766086e-01
-1.43097258e+00 -6.62202597e-01 -8.51846784e-02 6.62312582e-02
2.31944043e-02 1.00690496e+00 1.09190091e-01 8.52617145e-01
5.95322728e-01 3.31810892e-01 -2.79998451e-01 -4.15738672e-01
4.84432608e-01 8.14165473e-01 7.95221984e-01 -4.86404419e-01
8.32436308e-02 5.95056973e-02 -1.82356045e-01 -1.30034328e+00
-1.23663855e+00 -1.48376510e-01 -4.16807413e-01 -1.77352846e-01
9.48318183e-01 -8.61763358e-01 -3.12734038e-01 1.10535771e-01
-2.12425780e+00 -7.54230246e-02 -5.86185277e-01 1.80204123e-01
-5.74991465e-01 5.20076215e-01 -9.40002620e-01 -7.15487301e-01
-1.01226592e+00 -7.61347771e-01 1.24153090e+00 7.13115394e-01
-8.54182601e-01 -8.43489707e-01 2.11355582e-01 2.80164629e-01
2.16369614e-01 2.41435111e-01 9.43427265e-01 -7.65502751e-01
-5.41710734e-01 -9.84946415e-02 -8.08291659e-02 1.90699160e-01
-2.32985288e-01 3.44421417e-01 -5.70469022e-01 -8.68540779e-02
7.50743076e-02 -5.26691675e-01 1.53405428e+00 8.00113022e-01
1.27758312e+00 -8.47145975e-01 -4.04715657e-01 9.44986567e-03
7.22665668e-01 1.55735970e-01 6.78673685e-01 9.55925360e-02
4.50151414e-01 5.79102516e-01 2.96842039e-01 4.55690980e-01
5.61846375e-01 1.08914234e-01 7.80227631e-02 2.53826857e-01
-2.85958052e-01 -2.61080861e-01 3.54585618e-01 1.50475466e+00
-5.58298603e-02 -9.66948628e-01 -6.01251245e-01 8.14004064e-01
-2.12922788e+00 -1.29744172e+00 -1.95008397e-01 1.47827721e+00
1.20338440e+00 3.13047141e-01 -1.58408254e-01 -2.99696445e-01
5.53281248e-01 8.35419595e-01 -5.52153409e-01 -1.05340350e+00
-3.49309623e-01 8.16715285e-02 -1.00335628e-01 6.24927938e-01
-9.54308748e-01 9.51710582e-01 7.56486702e+00 5.72363734e-01
-6.82299018e-01 -2.36105755e-01 5.58456481e-01 -5.01445591e-01
-5.25421500e-01 -1.90013200e-01 -7.87751317e-01 3.21982354e-01
1.30391884e+00 -1.05427623e+00 -1.15752578e-01 5.32890499e-01
3.99530739e-01 -2.38580763e-01 -1.26806355e+00 4.65607584e-01
5.70199192e-01 -2.09031105e+00 7.09739506e-01 -2.53306091e-01
1.04923308e+00 -2.06621051e-01 -3.28156352e-01 2.66053766e-01
3.27400953e-01 -9.21734095e-01 8.21566164e-01 7.40595341e-01
3.99456114e-01 -5.52971005e-01 7.01458514e-01 5.93040466e-01
-3.73512864e-01 8.97035226e-02 -5.19665897e-01 -2.29559362e-01
4.08517122e-01 5.68646133e-01 -7.20699072e-01 8.30633879e-01
2.40838289e-01 1.05821621e+00 -5.13891518e-01 7.97119737e-01
-5.45926690e-01 7.94781864e-01 2.40941957e-01 -5.85257232e-01
3.11813742e-01 8.53126943e-02 1.07381511e+00 1.69021130e+00
1.26754895e-01 5.01152098e-01 -4.75777127e-02 1.12109518e+00
-6.12644851e-01 -5.83417267e-02 -5.47638953e-01 -5.69261909e-01
3.62810731e-01 9.51781631e-01 -3.72424901e-01 -9.77818727e-01
1.85145140e-02 8.70393634e-01 4.39217538e-01 3.72918367e-01
-5.13667107e-01 -5.15349805e-01 -2.66233124e-02 -2.14476511e-01
3.75650227e-01 -1.39168009e-01 -6.46539867e-01 -1.26235211e+00
1.09371297e-01 -8.21143925e-01 2.66528547e-01 -9.74538743e-01
-8.41117859e-01 5.45747519e-01 1.87738046e-01 -5.61829507e-01
-5.22826850e-01 7.47686103e-02 -1.13784480e+00 6.63052082e-01
-1.23237979e+00 -7.49701679e-01 -2.95567587e-02 -3.97679806e-01
1.24268115e+00 -2.15750113e-01 9.18501914e-01 -3.39771599e-01
-8.23398530e-01 3.69138010e-02 2.50318617e-01 -1.45020843e-01
4.48316813e-01 -1.37149763e+00 7.10673630e-01 9.41986322e-01
-1.98539615e-01 8.87886941e-01 1.13745368e+00 -1.00998843e+00
-1.41044676e+00 -9.11646664e-01 1.42833638e+00 -5.20634830e-01
4.50213522e-01 -1.95305616e-01 -9.37633574e-01 7.68485427e-01
1.24107158e+00 -1.05910075e+00 5.60600758e-01 -1.15418725e-01
1.54577449e-01 2.70630270e-01 -5.87040842e-01 8.40405643e-01
8.12488258e-01 -9.85750332e-02 -1.47041714e+00 4.77909565e-01
1.12422705e+00 -6.49357021e-01 -6.54600561e-01 1.59468263e-01
4.65916425e-01 -4.37445760e-01 7.14644492e-01 -1.01133358e+00
1.63177264e+00 1.27157837e-01 3.29980165e-01 -1.59876895e+00
-3.54749054e-01 -8.05237174e-01 -9.19943392e-01 1.17421687e+00
4.56786782e-01 3.01675517e-02 6.12061501e-01 4.12192106e-01
-8.10320914e-01 -6.64332986e-01 -7.09645987e-01 -4.31791544e-01
3.46815646e-01 2.98954159e-01 2.27794275e-01 4.36852664e-01
5.89613795e-01 1.20474148e+00 -1.04840793e-01 -4.07853991e-01
5.47243059e-01 3.46119761e-01 6.27637804e-01 -1.22890878e+00
7.02693537e-02 -8.82933378e-01 1.27737045e-01 -1.33635044e+00
6.72450364e-01 -8.11101556e-01 1.16577998e-01 -2.52955127e+00
6.94781899e-01 6.73216939e-01 1.50306582e-01 5.21266982e-02
-5.24377525e-01 -3.98726285e-01 1.16075270e-01 1.01532945e-02
-1.24469209e+00 8.18126321e-01 1.30978048e+00 -6.26139402e-01
-2.78467089e-01 -2.35533655e-01 -1.21531844e+00 5.55233359e-01
7.83488750e-01 -3.66455078e-01 -3.08599859e-01 -6.77421272e-01
1.24613367e-01 4.68482703e-01 -5.58887646e-02 -6.38570011e-01
5.85234523e-01 -7.22359270e-02 3.04829508e-01 -1.20255888e+00
7.55217150e-02 3.00496221e-01 -4.04698312e-01 3.37843806e-01
-1.20016468e+00 6.13388866e-02 3.96301270e-01 4.41168427e-01
-3.98440599e-01 -5.59686482e-01 3.14959466e-01 -3.86889130e-01
-6.05008379e-02 -2.75750369e-01 -7.62413621e-01 3.47730517e-01
5.70988536e-01 9.98255536e-02 -9.03026044e-01 -5.72646439e-01
-3.81986022e-01 5.57099044e-01 1.70961499e-01 5.31993926e-01
6.90974355e-01 -7.89503217e-01 -1.45192552e+00 -4.46193069e-01
-3.48376334e-01 3.51850092e-01 6.31050859e-03 2.90818393e-01
-5.85384309e-01 1.03836346e+00 -6.62694201e-02 -2.41310209e-01
-1.32578051e+00 2.68180162e-01 -2.78986275e-01 -5.56524098e-01
-7.65049636e-01 7.34600723e-01 7.88530484e-02 3.47519457e-01
3.33154410e-01 -5.28841674e-01 -4.45440322e-01 2.90866792e-01
9.95269597e-01 5.83852410e-01 2.28634533e-02 -2.66209036e-01
-1.48990288e-01 8.18948261e-03 -5.84691226e-01 -2.99164295e-01
1.49232781e+00 -7.50723854e-02 -4.43142772e-01 7.66178608e-01
9.39078093e-01 -8.11510161e-02 -8.04497063e-01 -1.40345376e-02
2.18087688e-01 2.22172052e-01 9.28076506e-02 -8.02917659e-01
-1.89109355e-01 7.70926237e-01 -7.43737817e-01 5.12176692e-01
7.84372091e-01 2.22621098e-01 7.83469558e-01 7.49818265e-01
-3.21586162e-01 -1.26421165e+00 1.76422983e-01 8.10853720e-01
1.43008661e+00 -8.71364951e-01 6.44388139e-01 -1.24974005e-01
-5.45899689e-01 1.25637662e+00 3.84700328e-01 -2.35048667e-01
-1.58851400e-01 1.16313860e-01 -4.34280366e-01 -4.27999496e-01
-1.47866166e+00 4.11069900e-01 3.50898296e-01 1.28664985e-01
7.86473215e-01 -1.78217217e-02 -7.52945840e-01 8.72108757e-01
-4.15044755e-01 2.31408328e-02 1.18225002e+00 9.57640827e-01
-6.83475256e-01 -5.41401327e-01 -2.11344868e-01 6.55721784e-01
-7.81031311e-01 -1.49838343e-01 -1.02814054e+00 4.01484460e-01
-7.93971360e-01 1.08077025e+00 1.99746564e-01 2.35343069e-01
3.99158925e-01 2.17498437e-01 6.99359953e-01 -9.41447258e-01
-8.65431070e-01 5.93956299e-02 7.05396771e-01 -2.15888619e-01
-4.35684443e-01 -8.74049902e-01 -1.18985653e+00 -3.59104455e-01
3.83564067e-04 4.67588544e-01 5.30277789e-01 7.38605797e-01
6.37298286e-01 1.18837929e+00 2.87467986e-01 -9.68268454e-01
-6.78256214e-01 -1.32238996e+00 -1.37257010e-01 1.22778915e-01
6.13293588e-01 1.55934572e-01 -7.50936866e-02 3.08373481e-01] | [12.575874328613281, 9.555704116821289] |
39c5cb88-264f-4b48-a767-4665da7fc4c0 | provable-multi-objective-reinforcement | 2011.10134 | null | https://arxiv.org/abs/2011.10134v2 | https://arxiv.org/pdf/2011.10134v2.pdf | Provable Multi-Objective Reinforcement Learning with Generative Models | Multi-objective reinforcement learning (MORL) is an extension of ordinary, single-objective reinforcement learning (RL) that is applicable to many real-world tasks where multiple objectives exist without known relative costs. We study the problem of single policy MORL, which learns an optimal policy given the preference of objectives. Existing methods require strong assumptions such as exact knowledge of the multi-objective Markov decision process, and are analyzed in the limit of infinite data and time. We propose a new algorithm called model-based envelop value iteration (EVI), which generalizes the enveloped multi-objective $Q$-learning algorithm in Yang et al., 2019. Our method can learn a near-optimal value function with polynomial sample complexity and linear convergence speed. To the best of our knowledge, this is the first finite-sample analysis of MORL algorithms. | ['Quanquan Gu', 'Jiahao Chen', 'Dongruo Zhou'] | 2020-11-19 | null | null | null | null | ['multi-objective-reinforcement-learning'] | ['methodology'] | [-5.74670397e-02 2.91509740e-02 -8.00356209e-01 -1.73073009e-01
-1.26787138e+00 -5.26681721e-01 9.58207995e-02 5.73139414e-02
-9.58028853e-01 1.47416687e+00 -2.75910765e-01 -3.78375530e-01
-7.66967773e-01 -4.37225342e-01 -7.45932758e-01 -6.74410105e-01
-3.77851903e-01 8.81067753e-01 -7.20850602e-02 1.05149969e-02
6.29574299e-01 1.05214022e-01 -1.23226595e+00 -3.04345667e-01
1.04323077e+00 1.04774296e+00 3.38638335e-01 9.71157730e-01
2.94889994e-02 7.74543166e-01 -5.35639465e-01 -4.98304404e-02
2.94158399e-01 -3.75175923e-01 -1.11580813e+00 -1.00244798e-01
-1.54846177e-01 -4.94304359e-01 -3.00362967e-02 1.10397387e+00
6.64526165e-01 5.68908453e-01 7.29005039e-01 -1.48919678e+00
-4.28390414e-01 5.14261007e-01 -5.77542782e-01 1.66406095e-01
2.62396514e-01 4.81524676e-01 1.16718125e+00 -2.86669374e-01
3.72561485e-01 1.51681292e+00 2.73667037e-01 6.27852619e-01
-1.39105344e+00 -4.18478876e-01 2.76059866e-01 3.42863590e-01
-9.75617528e-01 -7.59697407e-02 5.62151432e-01 -1.78268164e-01
9.96083677e-01 -1.08724713e-01 5.09238064e-01 6.62653983e-01
5.46565592e-01 1.18606055e+00 1.65045643e+00 -4.15517271e-01
4.91563469e-01 -1.36525646e-01 -3.88305396e-01 8.34537029e-01
1.67571127e-01 5.98075926e-01 -2.62901545e-01 -2.60005295e-01
4.98213589e-01 -1.32172242e-01 1.43223137e-01 -6.45040452e-01
-1.11543155e+00 1.07114661e+00 -1.99369937e-01 -7.40762278e-02
-5.53246558e-01 7.02096283e-01 6.68215007e-02 7.14366317e-01
2.64348537e-01 6.27073467e-01 -7.58073211e-01 -4.52729106e-01
-7.29789734e-01 5.19687772e-01 7.62501597e-01 6.81298554e-01
7.26573884e-01 3.24665785e-01 -3.13421041e-01 4.16520745e-01
4.43880767e-01 8.99859190e-01 4.86140817e-01 -1.52048075e+00
5.25866032e-01 -8.47255625e-03 8.80511582e-01 7.00987596e-03
-3.11836272e-01 -1.69286907e-01 -1.09046958e-01 7.58230805e-01
5.27269363e-01 -7.28571653e-01 -5.61088204e-01 1.78828919e+00
8.04427117e-02 1.18684992e-02 2.79974014e-01 5.48980176e-01
-1.73809841e-01 6.55969858e-01 -6.46132678e-02 -8.22426975e-01
6.26712024e-01 -7.61568964e-01 -7.84178615e-01 -6.91130161e-02
3.72163653e-01 -6.11408651e-01 1.09268308e+00 5.58830023e-01
-1.39688718e+00 -2.64797479e-01 -8.99790347e-01 6.52261913e-01
-6.77915066e-02 -3.91063035e-01 5.20661950e-01 5.76039433e-01
-1.00903392e+00 8.34064245e-01 -4.83883023e-01 2.23281726e-01
2.69237518e-01 6.52445018e-01 3.23012263e-01 -1.06761649e-01
-1.22692347e+00 1.09406292e+00 7.09443688e-01 -2.65325755e-01
-1.54673433e+00 -4.43295807e-01 -5.55676937e-01 -7.06815198e-02
1.18328381e+00 -3.16136271e-01 2.06699228e+00 -1.19248104e+00
-2.00245762e+00 6.73469231e-02 -3.83953378e-02 -4.09706563e-01
4.74840790e-01 -2.26385221e-01 -2.09641799e-01 2.74499357e-01
-5.10977432e-02 3.42991024e-01 1.09594810e+00 -1.19481838e+00
-9.00889277e-01 -6.03942983e-02 3.85983080e-01 4.39830571e-01
7.94298053e-02 -1.44389495e-01 3.03784996e-01 -3.75819385e-01
-7.84296691e-01 -7.60842860e-01 -6.60817087e-01 -4.44976360e-01
1.75325498e-01 -5.53660214e-01 4.33106929e-01 -3.29198420e-01
1.25113630e+00 -1.63852274e+00 2.78281778e-01 2.39990368e-01
-2.47605294e-01 1.42260030e-01 -4.31930989e-01 5.23655832e-01
1.45662963e-01 2.06977621e-01 -2.81563222e-01 -5.09209596e-02
3.46355915e-01 3.97409886e-01 8.51450041e-02 3.88921410e-01
4.66079405e-03 1.02234495e+00 -1.23164761e+00 -7.10489750e-01
-1.14306649e-02 -4.49243128e-01 -5.52104771e-01 2.97446072e-01
-6.51032269e-01 9.61588472e-02 -7.39782572e-01 5.87161958e-01
1.32246800e-02 -1.94306493e-01 2.62437165e-01 4.35066849e-01
-2.70655930e-01 -2.79126048e-01 -1.57678473e+00 1.34463072e+00
-4.79628205e-01 1.09170407e-01 6.21261150e-02 -1.32572889e+00
5.64314127e-01 6.00714624e-01 8.71064305e-01 -5.78470170e-01
1.84183136e-01 3.44938368e-01 1.17441388e-02 -4.54700679e-01
3.35259229e-01 -6.74831450e-01 -2.19835281e-01 7.11650372e-01
2.50331074e-01 -2.13886440e-01 4.38221425e-01 -2.99551308e-01
7.41362751e-01 2.95145094e-01 6.18214548e-01 -4.71909523e-01
2.83413440e-01 -7.05564246e-02 6.18665516e-01 1.17109823e+00
-4.67188329e-01 -3.37163627e-01 5.54157376e-01 -2.16665976e-02
-7.45754361e-01 -1.20931125e+00 3.20343584e-01 1.14058137e+00
7.65986890e-02 1.42157659e-01 -4.03830290e-01 -7.97961354e-01
3.63479108e-01 8.98474872e-01 -4.32541639e-01 -2.32612132e-03
-4.04877335e-01 -8.02845240e-01 4.35627587e-02 2.37564564e-01
3.27312827e-01 -1.36537075e+00 -8.51622045e-01 7.47564018e-01
-1.21706732e-01 -5.69059372e-01 -7.01454580e-01 2.66091347e-01
-9.79727745e-01 -1.27129030e+00 -7.99797118e-01 -5.42913437e-01
3.50370705e-01 -2.65055239e-01 1.06751573e+00 -4.90838587e-01
-4.55994997e-03 1.04514074e+00 3.65038179e-02 -6.16148829e-01
-3.11496645e-01 -1.02343440e-01 3.18320125e-01 -5.63511997e-02
-8.12777802e-02 -1.14267834e-01 -5.71209848e-01 8.19836780e-02
-7.56027400e-01 -4.05246347e-01 8.11293781e-01 1.09221697e+00
8.89091074e-01 4.38347995e-01 8.82698298e-01 -5.41080892e-01
1.08543527e+00 -4.03085381e-01 -1.14130449e+00 6.59352422e-01
-1.15157187e+00 7.73248792e-01 9.10024524e-01 -6.35386825e-01
-1.05589187e+00 -1.95081487e-01 2.35191569e-01 -3.07865053e-01
9.25549716e-02 5.25318980e-01 -4.43128357e-03 -1.82704985e-01
1.68351650e-01 2.93616325e-01 2.46117428e-01 -3.72857809e-01
2.55156249e-01 3.54739428e-01 2.02372614e-02 -1.05088210e+00
6.29922032e-01 -1.51908621e-02 3.76496673e-01 -3.63805711e-01
-9.44046736e-01 -3.79473597e-01 -1.01887010e-01 -4.35680777e-01
6.23352647e-01 -3.77054840e-01 -1.45100296e+00 2.21131384e-01
-5.99518239e-01 -9.17549729e-01 -6.23039663e-01 8.23242188e-01
-1.55697489e+00 3.21317881e-01 -2.65159965e-01 -1.32862389e+00
-1.81651101e-01 -1.05417216e+00 2.96135664e-01 3.46627563e-01
3.59572142e-01 -1.08465266e+00 3.09300631e-01 -2.31629014e-02
1.79456070e-01 7.24460259e-02 8.10827672e-01 -3.55548948e-01
-5.12354553e-01 2.09415525e-01 2.76215464e-01 3.96513671e-01
-7.66966566e-02 -3.50131720e-01 -2.89684147e-01 -9.01562393e-01
1.35568958e-02 -9.10072565e-01 6.30044460e-01 7.44372129e-01
1.33498561e+00 -6.73080027e-01 8.72274414e-02 7.52507597e-02
1.90754652e+00 8.86228561e-01 6.05055392e-02 3.86228591e-01
-7.99923483e-03 2.90889591e-01 1.19429064e+00 7.70210326e-01
2.77467519e-01 2.32837722e-01 6.65844798e-01 2.37304077e-01
6.69383109e-01 -1.12992980e-01 7.30264008e-01 4.58142608e-01
-3.92038107e-01 -3.42662871e-01 -7.41843939e-01 5.28712571e-01
-2.08725071e+00 -1.38989961e+00 8.96725714e-01 2.42206311e+00
9.04227972e-01 3.02476257e-01 4.87764806e-01 -1.65222600e-01
6.29003704e-01 1.02149077e-01 -1.23608220e+00 -7.37826109e-01
2.59550631e-01 5.82467377e-01 1.02152073e+00 1.03965688e+00
-9.01096761e-01 7.03999281e-01 7.24177361e+00 1.05633461e+00
-5.65929830e-01 1.20204896e-01 5.06433189e-01 -3.13105375e-01
-2.62250960e-01 -3.28763016e-02 -8.37580800e-01 3.46735954e-01
1.24469638e+00 -5.15509069e-01 1.20123637e+00 7.81041980e-01
3.76254737e-01 -3.70694339e-01 -8.30766201e-01 7.61197031e-01
-3.01526278e-01 -9.08618093e-01 -3.74169141e-01 2.06712961e-01
8.64637971e-01 -1.55177236e-01 9.66795534e-02 6.90468371e-01
9.94969428e-01 -1.09976375e+00 5.55577219e-01 8.28785837e-01
7.72735417e-01 -1.47782815e+00 5.98510683e-01 5.37252247e-01
-1.11351359e+00 -7.14831054e-01 -1.11993089e-01 -1.30434142e-04
1.79867849e-01 -1.06627353e-01 -5.72963536e-01 4.73756075e-01
3.53382677e-01 3.27775717e-01 5.02818935e-02 8.97231519e-01
-1.75407916e-01 6.73647344e-01 -1.10464305e-01 -6.63939595e-01
8.49217355e-01 -2.81495005e-01 5.87582946e-01 6.93962991e-01
3.78720850e-01 -4.17354750e-03 6.97558761e-01 4.35952723e-01
2.78580874e-01 7.93482270e-03 -4.28002030e-01 -4.40906912e-01
1.93112209e-01 7.01822758e-01 -3.24697882e-01 -8.44917595e-02
-3.80028129e-01 6.24934375e-01 3.80093813e-01 4.66132909e-01
-7.39549577e-01 -5.04926264e-01 5.80295503e-01 -4.21113998e-01
4.90686536e-01 -3.55049670e-01 3.02315742e-01 -7.60702193e-01
-3.17101330e-01 -9.77684975e-01 6.09959900e-01 -2.26353794e-01
-1.26117086e+00 -1.85675964e-01 2.61979640e-01 -1.10430384e+00
-7.30633318e-01 -6.28802240e-01 -2.88030446e-01 7.21296787e-01
-1.83120739e+00 -4.27418321e-01 8.46436083e-01 7.57880032e-01
7.24139273e-01 -2.88655609e-01 6.61265910e-01 -3.35451275e-01
-2.81923294e-01 4.39137042e-01 9.70229030e-01 -3.42310101e-01
5.45017302e-01 -1.71658850e+00 -3.61025631e-01 4.70483094e-01
-3.61439854e-01 -9.11607370e-02 7.10702717e-01 -3.57438415e-01
-1.53847551e+00 -7.51775980e-01 5.53119838e-01 -6.17730469e-02
8.36601496e-01 2.70718843e-01 -3.61992359e-01 5.04994273e-01
3.44579041e-01 -3.15580606e-01 6.17372274e-01 4.41390686e-02
3.15248638e-01 -3.15127373e-01 -1.21087444e+00 6.52475893e-01
6.47389352e-01 -1.45523816e-01 -6.83766425e-01 2.24644035e-01
6.09954476e-01 -1.59578100e-01 -1.18978024e+00 3.91910076e-01
5.15540302e-01 -4.54991668e-01 1.00668812e+00 -1.14396822e+00
2.76972950e-01 3.57535593e-02 -1.88964576e-01 -1.71385050e+00
-3.15900683e-01 -1.07826126e+00 -5.13021946e-01 5.90480685e-01
3.52357745e-01 -7.63773322e-01 3.57322246e-01 2.26293802e-01
1.06891282e-01 -1.09224033e+00 -1.16807175e+00 -1.36924362e+00
3.62281501e-01 -1.99733511e-01 4.78573322e-01 5.63008785e-01
1.47876218e-02 -4.59476113e-02 -7.31992543e-01 -8.48293751e-02
1.19125676e+00 4.54613179e-01 1.90443456e-01 -8.73155951e-01
-7.89844811e-01 -5.12177944e-01 4.78650808e-01 -9.94009972e-01
4.13965434e-01 -5.73954940e-01 3.16657752e-01 -1.44173717e+00
6.53298795e-02 -3.61708075e-01 -9.64076698e-01 3.38931620e-01
-2.14294299e-01 -5.01192987e-01 4.30800349e-01 -2.28477240e-01
-1.19172919e+00 8.61645758e-01 1.67366672e+00 -6.54598102e-02
-3.11356455e-01 4.31619942e-01 -5.74356258e-01 6.77190065e-01
1.36071956e+00 -6.29897177e-01 -7.16090798e-01 -9.61188376e-02
9.28797573e-02 8.46308231e-01 -5.34625202e-02 -6.82466090e-01
-1.24481015e-01 -1.22239971e+00 6.86723143e-02 -5.46215534e-01
1.28017157e-01 -5.92104971e-01 -2.75975466e-01 1.00757504e+00
-6.15059555e-01 4.06063110e-01 -3.40054519e-02 8.33679199e-01
1.26306694e-02 -8.27185512e-01 9.49920297e-01 -5.65396369e-01
-8.22326601e-01 6.07484043e-01 -6.81210935e-01 6.41394973e-01
1.21242607e+00 2.16728866e-01 -3.97699140e-02 -5.20955622e-01
-6.63051903e-01 7.61943936e-01 -1.38927475e-01 -6.24283170e-03
7.40013659e-01 -1.21091914e+00 -6.23171031e-01 -4.39722180e-01
-3.67114514e-01 -3.42346042e-01 -5.04881442e-02 4.56657290e-01
9.25209448e-02 5.09454072e-01 -2.21822754e-01 -6.86937347e-02
-8.87322068e-01 8.14039588e-01 5.40646970e-01 -9.28489447e-01
5.33703826e-02 3.11140180e-01 -4.27464008e-01 -2.26566404e-01
1.42819181e-01 9.68002155e-02 1.63068827e-02 7.31040388e-02
3.75733256e-01 7.54019558e-01 -6.65984690e-01 -4.62680347e-02
4.64714877e-02 4.30085719e-01 8.17078501e-02 -8.07963431e-01
1.18692529e+00 -1.42204270e-01 3.08940589e-01 8.10903430e-01
9.15357232e-01 -4.85051244e-01 -1.71655750e+00 -3.35228264e-01
2.10700557e-01 -3.45389038e-01 8.71429443e-02 -8.91935527e-01
-5.79435647e-01 6.56474829e-01 8.24382186e-01 -5.59097715e-02
1.03022552e+00 -3.82629365e-01 4.87994224e-01 8.33488107e-01
6.33177161e-01 -1.83816874e+00 5.29161274e-01 6.50386691e-01
6.34226143e-01 -1.33018613e+00 -4.01505940e-02 6.03773177e-01
-6.93862498e-01 9.51723099e-01 6.68343663e-01 -1.38388604e-01
6.35639131e-01 5.38798459e-02 -2.77663350e-01 3.24849218e-01
-1.01890361e+00 -5.96813500e-01 5.68403825e-02 5.37465513e-01
-1.42643616e-01 3.25243711e-01 -4.86462593e-01 1.88511640e-01
3.92809629e-01 2.98185706e-01 3.18700105e-01 1.27296448e+00
-8.31269979e-01 -1.29421282e+00 -3.89611006e-01 6.48705959e-01
-6.50343835e-01 4.69225869e-02 3.34816724e-01 7.40849137e-01
-3.01223636e-01 1.06084955e+00 -2.05504119e-01 1.90759659e-01
1.38906762e-01 3.13334048e-01 8.83649170e-01 -2.51119196e-01
-3.54680330e-01 1.07378431e-01 -9.69366282e-02 -6.59371972e-01
-7.07132697e-01 -7.69509673e-01 -1.12536609e+00 -2.47252569e-01
-5.78832850e-02 4.11903560e-01 3.50960553e-01 1.12055111e+00
-2.75357012e-02 3.89265180e-01 1.14703679e+00 -5.38172245e-01
-1.52337015e+00 -6.21010065e-01 -8.04105103e-01 -2.49229465e-02
4.88950104e-01 -7.81110227e-01 -3.53045493e-01 -3.34327430e-01] | [4.217244625091553, 2.399397373199463] |
8e4dfb2f-70b6-47e6-a0fe-541971fec5e1 | 181201711 | 1812.01711 | null | http://arxiv.org/abs/1812.01711v1 | http://arxiv.org/pdf/1812.01711v1.pdf | A Graph-CNN for 3D Point Cloud Classification | Graph convolutional neural networks (Graph-CNNs) extend traditional CNNs to
handle data that is supported on a graph. Major challenges when working with
data on graphs are that the support set (the vertices of the graph) do not
typically have a natural ordering, and in general, the topology of the graph is
not regular (i.e., vertices do not all have the same number of neighbors).
Thus, Graph-CNNs have huge potential to deal with 3D point cloud data which has
been obtained from sampling a manifold. In this paper, we develop a Graph-CNN
for classifying 3D point cloud data, called PointGCN. The architecture combines
localized graph convolutions with two types of graph downsampling operations
(also known as pooling). By the effective exploration of the point cloud local
structure using the Graph-CNN, the proposed architecture achieves competitive
performance on the 3D object classification benchmark ModelNet, and our
architecture is more stable than competing schemes. | ['Yingxue Zhang', 'Michael Rabbat'] | 2018-11-28 | null | null | null | null | ['3d-object-classification'] | ['computer-vision'] | [-3.72254819e-01 2.78010756e-01 -4.46012914e-02 -2.46889800e-01
3.15568805e-01 -4.98843908e-01 4.42820489e-01 4.23139066e-01
-1.00756630e-01 1.81956127e-01 -2.56502390e-01 -4.51628715e-01
8.21289346e-02 -1.27288556e+00 -1.07435048e+00 -4.01439607e-01
-4.15353537e-01 5.37020445e-01 2.32021853e-01 -2.17135623e-01
1.31756917e-01 1.19869184e+00 -1.44769514e+00 -2.62819580e-04
5.25914371e-01 1.12764788e+00 1.16737582e-01 4.73404735e-01
-5.07308006e-01 3.90225470e-01 -4.66555536e-01 -1.13375366e-01
3.78042787e-01 2.43032929e-02 -6.82906866e-01 1.95995510e-01
8.92876804e-01 -1.38647974e-01 -6.99825525e-01 1.39361608e+00
2.12919518e-01 2.12806970e-01 4.74846393e-01 -1.61147690e+00
-1.12325084e+00 2.35095650e-01 -4.03823733e-01 1.55059546e-01
1.00837134e-01 -2.77582984e-02 1.07034230e+00 -9.21585858e-01
5.82283974e-01 1.50651014e+00 8.66487980e-01 2.47096255e-01
-1.13195062e+00 -5.65522730e-01 4.29657638e-01 1.88111141e-02
-1.57847369e+00 2.33373284e-01 9.73954022e-01 -4.03914869e-01
1.22058511e+00 5.28262407e-02 1.21823573e+00 6.64513767e-01
2.76383847e-01 3.02862942e-01 5.26601672e-01 -1.98577084e-02
3.23136806e-01 -5.76769054e-01 2.46901363e-01 9.78266835e-01
4.65815663e-01 -2.58445650e-01 -2.60065407e-01 -1.09032318e-01
1.12664628e+00 4.65993285e-01 -1.44532502e-01 -8.60843122e-01
-1.23293877e+00 8.37496221e-01 1.36089778e+00 9.48114768e-02
-2.86753386e-01 4.44360733e-01 3.29821587e-01 4.50806141e-01
6.12085819e-01 3.32239211e-01 -1.16714500e-01 4.81304765e-01
-4.60649431e-01 3.74327749e-01 6.68057680e-01 1.59729302e+00
1.09247887e+00 -4.13184538e-02 7.72393197e-02 4.53647584e-01
1.52576968e-01 2.95108885e-01 -5.00732698e-02 -4.18334723e-01
5.56796134e-01 1.26883936e+00 -4.10940498e-01 -1.38210559e+00
-7.12758660e-01 -4.59225804e-01 -1.44341171e+00 3.25886756e-01
1.42074093e-01 2.97124207e-01 -1.22815526e+00 1.53193629e+00
2.67579406e-01 4.73019600e-01 -2.26132736e-01 8.44005287e-01
1.28246963e+00 4.82567787e-01 -2.17042252e-01 3.57642859e-01
1.06023598e+00 -7.43005574e-01 -3.44669014e-01 -1.19329967e-01
6.30665600e-01 -1.48072824e-01 1.15549779e+00 -2.14058816e-01
-8.37814271e-01 -6.35712385e-01 -1.28773904e+00 -3.29173625e-01
-7.68604875e-01 -4.77628052e-01 8.03430617e-01 1.17634922e-01
-1.40809393e+00 8.24572086e-01 -8.50304008e-01 -4.77805048e-01
7.91919827e-01 4.72233176e-01 -6.85543001e-01 -3.55467737e-01
-7.74623632e-01 4.63716000e-01 4.99079317e-01 2.40582079e-01
-6.16523504e-01 -6.18142903e-01 -1.06829488e+00 3.47115219e-01
7.17495978e-02 -7.64813900e-01 6.75757885e-01 -5.20854235e-01
-1.09602654e+00 1.03326690e+00 2.34118655e-01 -2.43226245e-01
2.03185245e-01 2.63034344e-01 -1.74373612e-01 5.67902848e-02
9.97178555e-02 6.34540021e-01 8.72212112e-01 -9.26955104e-01
-1.68910787e-01 -6.43926501e-01 2.40014717e-01 1.53537244e-01
-1.28819183e-01 -3.27216148e-01 -6.40134871e-01 -4.41433370e-01
5.93955636e-01 -1.07208252e+00 -3.51403981e-01 1.69382781e-01
-6.43167377e-01 -5.05638838e-01 1.09579277e+00 4.93503325e-02
7.94642091e-01 -2.34296775e+00 1.68292210e-01 3.77048165e-01
9.90548015e-01 2.56495178e-01 -1.96849704e-01 4.11013812e-01
-4.20003444e-01 3.40816915e-01 -1.59905568e-01 -1.60669535e-01
-1.24540657e-01 3.32101434e-01 -1.42567739e-01 6.45596683e-01
4.96233016e-01 1.31123722e+00 -9.90812182e-01 -1.70900106e-01
2.92100400e-01 4.01259512e-01 -7.05298901e-01 1.54604077e-01
-3.81457269e-01 2.40868181e-01 -5.31299889e-01 6.03334010e-01
1.13261819e+00 -8.68271589e-01 3.02256271e-03 -2.00842395e-01
8.08641911e-02 1.81176662e-01 -1.06630254e+00 1.88884199e+00
-1.39033481e-01 4.72099811e-01 -1.79909989e-01 -9.52122390e-01
1.08757937e+00 -1.13187797e-01 4.19140667e-01 -3.35992098e-01
1.45624459e-01 7.42794275e-02 6.48197830e-02 1.00773901e-01
2.70073026e-01 2.70815283e-01 1.34217978e-01 2.63819396e-01
2.84925461e-01 -4.24142838e-01 -1.30935639e-01 1.91863567e-01
1.33372581e+00 -3.96100670e-01 8.13493878e-03 -5.42169034e-01
2.12507412e-01 5.13474122e-02 2.15869144e-01 7.42930174e-01
2.30320096e-02 6.63269222e-01 7.05953002e-01 -8.99730384e-01
-9.70382333e-01 -1.07082880e+00 -4.00076993e-02 4.33308095e-01
5.04048765e-01 -5.78404307e-01 -4.85030711e-01 -8.21427703e-01
4.14457530e-01 2.01639831e-01 -6.54143929e-01 -4.85912949e-01
-5.94286680e-01 -2.92110234e-01 2.06621483e-01 4.27851260e-01
6.21154308e-01 -9.38421488e-01 -2.60610133e-01 2.06480727e-01
4.34748292e-01 -1.28539717e+00 -4.69257325e-01 2.66349047e-01
-1.24945211e+00 -1.31005704e+00 -4.28123623e-01 -1.18310618e+00
9.60940659e-01 7.33922720e-01 1.40324807e+00 4.71400291e-01
9.59695801e-02 1.69998765e-01 -3.71447384e-01 -4.60012704e-01
4.78215665e-02 3.23014915e-01 -2.03193381e-01 1.99040454e-02
5.95134914e-01 -8.65441024e-01 -4.34140891e-01 9.61485356e-02
-8.55522752e-01 1.16041742e-01 3.25253636e-01 7.53788650e-01
8.88541758e-01 2.22051859e-01 1.61424220e-01 -8.44512165e-01
6.46124959e-01 -4.72014248e-01 -7.61581957e-01 -6.45447243e-03
-2.99638003e-01 -7.15461075e-02 7.69459546e-01 -2.77016699e-01
-1.33459926e-01 -3.10229231e-02 8.98697153e-02 -1.14571381e+00
-8.47716033e-02 6.74200356e-01 -2.40178883e-01 -5.57725608e-01
4.83330905e-01 -2.53164202e-01 3.03986132e-01 -4.60745811e-01
4.24352050e-01 8.61962736e-02 2.42206559e-01 -1.81728944e-01
8.93414021e-01 6.30793571e-01 7.26642072e-01 -9.24967825e-01
-5.26764095e-01 -2.73822755e-01 -9.15303469e-01 -1.86404288e-01
9.67373312e-01 -8.57816458e-01 -7.22192585e-01 6.42101645e-01
-1.25439894e+00 -4.45904344e-01 -2.88701832e-01 4.26346213e-01
-4.74572778e-01 2.79164553e-01 -6.78643703e-01 -2.37093434e-01
-2.09555402e-01 -1.02153397e+00 1.23020685e+00 6.02518506e-02
1.72332227e-01 -9.23680544e-01 -3.97407115e-01 -4.84088808e-01
2.32285470e-01 7.28810668e-01 1.27096033e+00 -6.53048337e-01
-1.07711208e+00 -4.29645330e-01 -5.49548686e-01 2.77449876e-01
1.96344495e-01 -6.95158839e-02 -6.40112221e-01 -4.94136363e-01
-3.17104012e-01 -7.01309890e-02 7.33778954e-01 4.61757302e-01
1.56838822e+00 -9.05384719e-02 -3.41496348e-01 1.04948020e+00
1.34672868e+00 -3.31704080e-01 4.92238462e-01 2.17080768e-02
1.32869947e+00 2.23092690e-01 4.87044007e-02 9.42497998e-02
3.44442457e-01 2.99693316e-01 1.10661101e+00 -2.16683730e-01
-2.83354133e-01 -4.60926861e-01 -2.26173982e-01 9.58103478e-01
-1.29517496e-01 -2.69309521e-01 -1.10084307e+00 4.04212326e-01
-1.85534084e+00 -3.78490865e-01 -4.59713221e-01 1.99712551e+00
2.40091868e-02 2.13595122e-01 -8.78613293e-02 -1.07014939e-01
1.07793593e+00 3.98234963e-01 -6.71535730e-01 -1.92983940e-01
-2.59527028e-01 4.39485341e-01 7.58324623e-01 2.56547153e-01
-1.12937164e+00 9.57182586e-01 6.02535391e+00 2.92631358e-01
-1.17098498e+00 -2.20132783e-01 3.28615963e-01 1.53348520e-01
-1.75122455e-01 -1.00296870e-01 -5.21662056e-01 1.97182730e-01
5.12788236e-01 -9.34726894e-02 4.51183885e-01 9.82863903e-01
-4.05195028e-01 6.13282382e-01 -1.42151713e+00 1.20748603e+00
-7.03712627e-02 -1.73503029e+00 4.88969535e-01 3.62288743e-01
5.41110933e-01 5.63467383e-01 -1.08637482e-01 8.82036835e-02
4.08520103e-01 -1.27080739e+00 7.59054601e-01 2.68902272e-01
1.00045562e+00 -8.03466618e-01 6.50590122e-01 2.76451737e-01
-1.39996636e+00 1.41424716e-01 -8.61510158e-01 -2.06113011e-01
-1.71723068e-01 7.09202170e-01 -7.80440688e-01 6.61461055e-01
1.04243684e+00 1.25606155e+00 -5.71550310e-01 1.08287716e+00
-1.37932524e-01 6.61589429e-02 -3.72421890e-01 -2.13640258e-01
4.66898471e-01 -5.18083334e-01 6.57969534e-01 7.55762994e-01
4.50642943e-01 2.18386576e-02 3.15190494e-01 1.20987105e+00
-5.72741449e-01 7.72815347e-02 -1.19734251e+00 -1.63916752e-01
3.27947080e-01 1.09505725e+00 -9.83860373e-01 -1.55232400e-01
-6.69775307e-01 7.25602508e-01 7.59086549e-01 4.17437315e-01
-4.19640630e-01 -5.20662844e-01 9.47180450e-01 1.34973198e-01
4.65705067e-01 -6.26987457e-01 -2.72740602e-01 -1.10352039e+00
9.80278552e-02 -5.14952779e-01 3.97361934e-01 -9.15149927e-01
-1.53361416e+00 7.54684210e-01 -6.82174265e-02 -1.24579453e+00
4.20243621e-01 -9.69440758e-01 -6.82907879e-01 1.04373753e+00
-1.45482159e+00 -1.22491705e+00 -7.80197501e-01 8.86902153e-01
-1.28348961e-01 -4.13110219e-02 6.73932374e-01 2.72240877e-01
-1.73623621e-01 2.38923371e-01 -4.67125893e-01 4.56019402e-01
-2.76450813e-02 -1.26364875e+00 1.21664965e+00 6.19550884e-01
1.94050282e-01 6.60966456e-01 -1.23985792e-02 -7.68383980e-01
-1.77709031e+00 -1.61711752e+00 5.71740091e-01 -2.89201707e-01
6.10551178e-01 -8.79495978e-01 -1.09294462e+00 9.56386507e-01
-1.90004334e-01 6.72000885e-01 3.00577343e-01 1.96954355e-01
-5.74798703e-01 2.99345255e-02 -1.02900362e+00 6.11994028e-01
1.67137694e+00 -6.64712727e-01 -2.85012037e-01 6.36096954e-01
1.16999102e+00 -6.60387814e-01 -8.50139618e-01 4.65996474e-01
-8.27628747e-02 -7.83255935e-01 9.77280974e-01 -1.06647956e+00
2.75738519e-02 -4.25828487e-01 -1.87966213e-01 -1.39897537e+00
-7.22986817e-01 -4.13293600e-01 -1.39203057e-01 5.71798027e-01
1.20941445e-01 -8.44297945e-01 9.30304348e-01 2.25582644e-01
-6.28025889e-01 -7.07765579e-01 -1.01717913e+00 -9.40249741e-01
4.89035919e-02 -3.14302623e-01 1.29446363e+00 1.19402337e+00
-4.35448438e-01 3.91867399e-01 2.72843689e-01 6.39676571e-01
6.78809822e-01 2.75591195e-01 9.41317558e-01 -1.63262284e+00
2.14841112e-01 -7.83875823e-01 -1.21863532e+00 -1.23251331e+00
2.95654774e-01 -1.53730917e+00 -3.76708299e-01 -1.61957097e+00
-3.95615309e-01 -5.03588319e-01 -1.90489024e-01 5.01701415e-01
5.22219343e-03 1.15200840e-01 1.37015477e-01 8.80915448e-02
-3.49244207e-01 6.51517153e-01 1.67853963e+00 -3.57267588e-01
-1.75308466e-01 -1.36416465e-01 -3.36895019e-01 4.97043669e-01
7.04204142e-01 -2.62786955e-01 -4.35612559e-01 -7.26554453e-01
4.69869018e-01 -1.68807521e-01 7.36617267e-01 -1.00044012e+00
2.78659016e-01 5.20360544e-02 2.29350954e-01 -8.61673951e-01
9.28684548e-02 -9.98034954e-01 3.44453216e-01 2.36357480e-01
8.98137167e-02 4.52742666e-01 2.22627878e-01 7.86728919e-01
-3.37817043e-01 2.48100087e-01 7.99396336e-01 -3.62306595e-01
-5.88258147e-01 1.22013509e+00 2.19842374e-01 -8.77309814e-02
9.06830788e-01 -2.45445341e-01 -2.76692241e-01 -1.70008361e-01
-6.67112112e-01 2.79702932e-01 7.94072628e-01 6.00370705e-01
8.09405625e-01 -1.84697056e+00 -4.10699606e-01 5.98420799e-01
5.21732211e-01 9.21442032e-01 3.48175969e-03 4.52985495e-01
-1.03953815e+00 1.03508405e-01 -3.60813677e-01 -1.02484143e+00
-7.64744878e-01 8.60488117e-01 4.73414838e-01 1.41291529e-01
-1.18059313e+00 1.06710052e+00 4.59846526e-01 -6.93428695e-01
2.62787819e-01 -9.00475025e-01 1.21448159e-01 -2.60627061e-01
1.34626344e-01 2.05162936e-03 4.63239729e-01 -6.08405769e-01
-4.04966235e-01 5.81472933e-01 1.54491544e-01 6.65465593e-01
1.63187933e+00 3.34984869e-01 -6.71345353e-01 4.61797416e-01
1.45921111e+00 -4.37447399e-01 -1.05358922e+00 -3.61725092e-01
-1.27544463e-01 -6.00381553e-01 6.68104291e-02 1.76226079e-01
-1.45035672e+00 9.10179794e-01 1.56053692e-01 7.80043244e-01
7.54192054e-01 2.59244651e-01 6.20550573e-01 6.12118006e-01
4.29605693e-01 -7.35329270e-01 2.66855257e-03 8.28945279e-01
1.12653685e+00 -1.05067265e+00 -1.06801048e-01 -7.47004032e-01
3.00771967e-02 1.20662773e+00 6.50047660e-01 -8.08835864e-01
1.23004842e+00 -2.56153256e-01 -2.91462213e-01 -9.37385380e-01
-4.59503025e-01 -3.35899919e-01 3.78234744e-01 7.79451013e-01
-7.28968978e-02 2.20741972e-01 4.06595141e-01 1.57078132e-01
-4.38624591e-01 -3.95049244e-01 3.23504835e-01 7.37067819e-01
-1.40301332e-01 -6.46452248e-01 -2.04837341e-02 7.90609598e-01
1.44567400e-01 -1.32135570e-01 -6.10630095e-01 1.00416505e+00
-3.30019705e-02 5.73867500e-01 5.51030338e-01 -5.82949698e-01
6.88509107e-01 -2.81354696e-01 3.91809136e-01 -1.02536595e+00
-2.18722671e-01 -3.09460908e-01 -2.95365959e-01 -7.66146958e-01
-2.67128795e-01 -3.22884679e-01 -1.18142498e+00 -5.77462614e-01
-2.71395236e-01 -1.37107775e-01 5.99620163e-01 4.05274212e-01
7.73771465e-01 6.41249180e-01 5.80487430e-01 -1.12011647e+00
-6.73324242e-02 -6.64009273e-01 -9.65453327e-01 5.48068881e-01
5.32145083e-01 -8.42734098e-01 -4.08844620e-01 -5.78151822e-01] | [7.952322483062744, -3.691884756088257] |
7e2b3027-d047-45ab-9363-ee2bcbfe1728 | estimating-soft-labels-for-out-of-domain | 2211.05561 | null | https://arxiv.org/abs/2211.05561v1 | https://arxiv.org/pdf/2211.05561v1.pdf | Estimating Soft Labels for Out-of-Domain Intent Detection | Out-of-Domain (OOD) intent detection is important for practical dialog systems. To alleviate the issue of lacking OOD training samples, some works propose synthesizing pseudo OOD samples and directly assigning one-hot OOD labels to these pseudo samples. However, these one-hot labels introduce noises to the training process because some hard pseudo OOD samples may coincide with In-Domain (IND) intents. In this paper, we propose an adaptive soft pseudo labeling (ASoul) method that can estimate soft labels for pseudo OOD samples when training OOD detectors. Semantic connections between pseudo OOD samples and IND intents are captured using an embedding graph. A co-training framework is further introduced to produce resulting soft labels following the smoothness assumption, i.e., close samples are likely to have similar labels. Extensive experiments on three benchmark datasets show that ASoul consistently improves the OOD detection performance and outperforms various competitive baselines. | ['Yongbin Li', 'Luo Si', 'Fei Huang', 'Jian Sun', 'Yinhe Zheng', 'Hao Lang'] | 2022-11-10 | null | null | null | null | ['intent-detection'] | ['natural-language-processing'] | [ 1.19590849e-01 5.26410758e-01 -5.96865356e-01 -8.96487415e-01
-4.20049906e-01 -5.12721062e-01 8.75047863e-01 -7.20385090e-03
3.84204462e-03 3.40613246e-01 5.42282701e-01 3.75160836e-02
4.12319183e-01 -4.04683203e-01 -2.19780445e-01 -1.80723250e-01
3.99395823e-01 7.24643528e-01 3.49002212e-01 -2.37004217e-02
1.98262855e-01 9.72830970e-03 -1.14162338e+00 3.81620497e-01
1.05758810e+00 7.66952753e-01 2.39873767e-01 4.47527975e-01
-6.59248650e-01 6.31187916e-01 -1.13662183e+00 -3.92156929e-01
4.48633760e-01 -3.00656080e-01 -5.23674250e-01 5.72494507e-01
6.04045689e-01 -6.73306167e-01 -5.87102234e-01 1.20778739e+00
1.31769910e-01 1.35473043e-01 1.05796611e+00 -1.55317152e+00
-1.06668568e+00 2.29844257e-01 -4.42869186e-01 -1.43040359e-01
4.79242235e-01 1.83226511e-01 1.22000682e+00 -1.27089262e+00
7.75994003e-01 1.78905809e+00 4.01954502e-01 8.15170765e-01
-1.13185132e+00 -5.56835353e-01 2.02658162e-01 -2.29323968e-01
-1.02196074e+00 -3.38511825e-01 8.74037087e-01 -2.66119242e-01
8.44190598e-01 1.17155440e-01 1.68895051e-01 1.36422288e+00
1.86392125e-02 1.23523605e+00 1.00073683e+00 -4.51521039e-01
2.91006148e-01 8.45936954e-01 6.98346317e-01 7.07842112e-01
2.33950973e-01 -8.66088942e-02 -6.10502124e-01 -3.06400061e-01
3.77700716e-01 3.27785254e-01 -1.72466055e-01 -2.53114104e-01
-7.28490591e-01 1.04655814e+00 3.51531416e-01 1.43787041e-01
2.38989796e-02 -5.45091987e-01 2.99149334e-01 3.68956327e-01
5.74158192e-01 2.35915065e-01 -1.45962879e-01 1.67462304e-01
-4.75732565e-01 5.84857762e-02 1.09222496e+00 1.44088018e+00
9.04046595e-01 -2.15625882e-01 -4.50650841e-01 1.28135765e+00
8.17115009e-01 2.51009703e-01 6.10551357e-01 -9.57423031e-01
6.96935952e-01 1.04262769e+00 2.59961575e-01 -8.43076527e-01
-2.97481078e-03 8.89446661e-02 -5.31663001e-01 5.38631976e-02
3.99327964e-01 2.66307648e-02 -1.11339045e+00 1.48670328e+00
5.07894933e-01 -3.97568569e-02 3.35939974e-01 9.58213151e-01
9.45880353e-01 8.92958820e-01 2.86299512e-02 -9.21769440e-02
1.41379225e+00 -1.31086695e+00 -1.22897768e+00 -8.73807430e-01
7.36609221e-01 -9.29530680e-01 1.37396955e+00 -5.26458910e-03
-3.73998135e-01 -7.00380564e-01 -1.01423991e+00 -2.05854818e-01
-4.51878101e-01 1.76562309e-01 5.28919876e-01 6.02326870e-01
-6.38710141e-01 1.55907780e-01 -2.95690387e-01 -3.80477697e-01
2.89898723e-01 1.75128460e-01 -1.20706677e-01 -2.92321533e-01
-1.28342390e+00 6.25674307e-01 3.93753856e-01 -1.84063554e-01
-9.32221413e-01 -5.14562905e-01 -1.11726272e+00 -1.27831325e-01
5.12841165e-01 -9.69745591e-02 1.37990391e+00 -7.29750812e-01
-1.41280413e+00 1.11706746e+00 -5.12643397e-01 -2.63654143e-01
5.13644874e-01 -2.89144188e-01 -4.84171718e-01 4.78197820e-04
5.12918293e-01 1.10384166e+00 1.33126867e+00 -1.44992244e+00
-5.88118672e-01 -1.53381243e-01 1.38843982e-02 4.99087304e-01
-7.01516449e-01 -2.59509087e-01 -7.33587563e-01 -6.74429715e-01
2.31743827e-01 -1.02618647e+00 5.48197813e-02 1.70103446e-01
-8.64793241e-01 -9.31955636e-01 1.41606951e+00 -3.17060798e-01
1.02068233e+00 -2.35452628e+00 -2.95089275e-01 -7.51324669e-02
5.48695147e-01 2.94936925e-01 3.36796194e-02 2.44086459e-01
3.22374731e-01 -1.45704433e-01 -2.13906273e-01 -6.64930344e-01
6.08273208e-01 4.14465994e-01 -5.76391816e-01 1.70787498e-02
3.53021622e-01 5.69413960e-01 -9.53218758e-01 -7.74038792e-01
4.43508714e-01 1.16213001e-01 -3.20977479e-01 9.00001287e-01
-4.60826784e-01 -1.44199328e-02 -4.77595538e-01 8.30462635e-01
8.32792282e-01 -3.00512612e-01 2.37235188e-01 -1.76534727e-01
2.73917317e-01 6.69258416e-01 -9.63535011e-01 1.34936190e+00
-4.91483837e-01 6.60572350e-01 -3.57203633e-02 -4.97657418e-01
1.36861944e+00 2.48912051e-01 1.15183396e-02 -3.75476837e-01
1.36551157e-01 2.04336897e-01 -3.63935471e-01 -3.89107466e-01
7.83014119e-01 -6.98527619e-02 -2.90039927e-01 3.39535922e-01
2.48731196e-01 -3.82007033e-01 -2.84590516e-02 3.39941710e-01
7.48920381e-01 -2.06923544e-01 2.82082915e-01 -1.32660970e-01
4.57114369e-01 7.45644644e-02 7.69228280e-01 8.22843254e-01
-7.59059131e-01 5.28047562e-01 6.61107779e-01 -1.34100452e-01
-7.43614435e-01 -1.14267528e+00 -2.80067563e-01 8.58906925e-01
5.63048840e-01 -5.26684999e-01 -7.64386892e-01 -1.38392317e+00
1.20454080e-01 1.06696105e+00 -3.46208543e-01 -3.23890418e-01
-4.26755905e-01 -4.23301816e-01 3.03153068e-01 2.78728694e-01
5.23532093e-01 -1.07072246e+00 1.43284619e-01 2.64857203e-01
-5.71782552e-02 -1.33558691e+00 -1.01674497e+00 4.76500183e-01
-7.04057992e-01 -7.14035988e-01 -7.99941719e-01 -1.23487914e+00
8.45194340e-01 6.92444623e-01 8.69506180e-01 -1.65615857e-01
-2.65557766e-02 8.96559358e-02 -3.36795479e-01 -3.70601416e-01
-7.85700619e-01 -1.04019456e-01 2.37272784e-01 7.40889907e-02
1.14824605e+00 4.89700623e-02 -2.09540322e-01 5.55294216e-01
-6.68693841e-01 -1.65516317e-01 4.30190891e-01 9.97709692e-01
3.78190041e-01 -1.14713252e-01 5.09514749e-01 -1.11450720e+00
9.67394054e-01 -5.08693576e-01 -3.46817017e-01 1.26647189e-01
-7.14471161e-01 2.63824403e-01 8.46296251e-01 -7.15637147e-01
-1.45887029e+00 7.55186146e-03 7.28830323e-02 -7.76628792e-01
-5.62525928e-01 -2.02279329e-01 -2.86018759e-01 3.46667290e-01
3.65111709e-01 4.58015949e-02 -1.56058624e-01 -5.56439102e-01
2.12102666e-01 1.43149602e+00 1.62813246e-01 -1.98885083e-01
7.97919571e-01 3.17643821e-01 -7.00954437e-01 -1.20741653e+00
-1.29737115e+00 -7.78814256e-01 -3.92891377e-01 -6.11255169e-02
9.44618046e-01 -8.87460887e-01 -1.57485798e-01 4.18848127e-01
-1.04231131e+00 -3.59622151e-01 -1.65443674e-01 4.77770835e-01
-1.46898136e-01 6.34946167e-01 -6.42833412e-01 -8.50530565e-01
-7.13705719e-02 -1.24035311e+00 1.30136931e+00 6.71439946e-01
-6.82711899e-01 -1.10309076e+00 -4.13729809e-02 4.42223549e-01
-1.44149140e-01 -6.30104363e-01 7.08298385e-01 -1.14529991e+00
-3.88970017e-01 -3.65816742e-01 -4.48832273e-01 3.21711659e-01
6.23876452e-01 -3.67058516e-01 -1.12305629e+00 1.53736295e-02
3.27200174e-01 -6.60837591e-01 7.04011858e-01 1.65617451e-01
7.60085166e-01 -2.00737074e-01 -5.34695327e-01 1.18849687e-01
8.76424611e-01 2.93594509e-01 1.39982447e-01 -1.14488460e-01
6.76388919e-01 7.83168018e-01 1.19810283e+00 3.65816832e-01
4.45408016e-01 8.82947385e-01 1.72286049e-01 -3.24184564e-03
-3.59492987e-01 -5.73654234e-01 6.26644790e-01 6.71523273e-01
9.36204433e-01 -6.64874792e-01 -4.88390297e-01 3.89059931e-01
-1.55300105e+00 -5.17042339e-01 -1.82949260e-01 1.89760911e+00
1.00099945e+00 7.28871346e-01 -4.44850139e-02 -2.89226741e-01
1.04767811e+00 5.24947882e-01 -7.71363080e-01 -4.19385314e-01
1.35770142e-01 -2.64734656e-01 2.03025624e-01 7.74827063e-01
-1.26547027e+00 1.25092888e+00 5.56796265e+00 7.64057934e-01
-6.50222242e-01 1.17962152e-01 6.25689089e-01 5.48370183e-01
-3.25709581e-01 7.11586513e-03 -1.56345880e+00 5.81740022e-01
3.71451110e-01 2.16554776e-01 7.14540621e-03 1.38727534e+00
9.01874527e-02 -4.28122096e-02 -1.01580548e+00 7.49531329e-01
3.23630691e-01 -7.86027431e-01 -1.04920268e-01 9.28461924e-02
8.25435340e-01 -3.32258314e-01 -1.06176294e-01 5.86508334e-01
6.74283981e-01 -4.93815094e-01 3.97793204e-01 6.42059818e-02
6.10920012e-01 -1.60878018e-01 7.04956532e-01 4.72339720e-01
-1.16500783e+00 1.21794008e-01 -7.46306956e-01 2.72472620e-01
3.73149663e-02 5.95672667e-01 -1.37524188e+00 9.48917866e-02
5.22391498e-01 9.88998234e-01 -4.04485315e-01 4.92115974e-01
-5.37378848e-01 5.24946392e-01 -4.45787758e-01 -3.95332068e-01
4.65400934e-01 -2.97982007e-01 5.93330443e-01 1.15499580e+00
-1.25898765e-02 -1.03682697e-01 4.14500386e-01 1.16692770e+00
-3.39067221e-01 -2.20213816e-01 -8.15125644e-01 -8.22376236e-02
7.22321510e-01 1.05057657e+00 -6.72559500e-01 -6.12327278e-01
-5.80507576e-01 1.37686980e+00 1.62338793e-01 1.56442598e-01
-8.75572860e-01 -6.38439238e-01 9.04859364e-01 -4.31799322e-01
5.66889346e-02 -8.65688268e-03 -2.77638882e-01 -1.29158270e+00
2.50839647e-02 -6.75250590e-01 4.57035422e-01 -4.78430897e-01
-1.76036417e+00 4.04500514e-01 -2.06141114e-01 -1.47633326e+00
-3.58576775e-02 -5.35315692e-01 -8.73001397e-01 6.02908850e-01
-1.61133182e+00 -6.55876338e-01 -5.66681921e-01 6.75924942e-02
1.47271347e+00 -1.53561132e-02 5.92592657e-01 4.92587090e-02
-6.54736042e-01 8.16293061e-01 1.08332187e-03 1.78802088e-01
1.26953614e+00 -1.43639994e+00 4.09923345e-01 6.14675224e-01
6.87212944e-02 6.41774595e-01 6.95941150e-01 -9.48126495e-01
-1.00754428e+00 -1.19216931e+00 1.31017745e+00 -4.43232626e-01
3.76020104e-01 -6.69481337e-01 -1.07457268e+00 6.55156076e-01
2.25472152e-01 -1.59530804e-01 4.87117082e-01 -1.99823856e-01
-3.50423008e-01 1.47529915e-01 -1.25989306e+00 6.85868800e-01
1.04553425e+00 -6.35344923e-01 -1.36622238e+00 7.58313954e-01
9.60442543e-01 -3.19293171e-01 -5.83054125e-01 4.97612767e-02
9.58541036e-02 -9.45417881e-01 1.17702067e+00 -9.35356244e-02
1.91333368e-01 -2.95140952e-01 -1.12809113e-03 -1.00999427e+00
1.82093456e-01 -5.28895438e-01 -4.55600649e-01 1.64930856e+00
1.68984249e-01 -5.17326057e-01 9.22075748e-01 7.36316383e-01
-4.97805685e-01 -3.45164031e-01 -5.81346512e-01 -1.03639472e+00
-4.30708796e-01 -1.24426283e-01 2.94023991e-01 9.84241247e-01
1.97892115e-01 6.33222699e-01 -5.10109961e-01 2.35153213e-01
7.51882255e-01 2.37698928e-01 8.26562405e-01 -1.08685100e+00
-1.96119964e-01 -3.00125536e-02 1.68663412e-02 -1.94430745e+00
4.39133734e-01 -8.61722171e-01 3.94327551e-01 -1.30677211e+00
-1.49946690e-01 -7.37338960e-01 1.21096514e-01 5.14724314e-01
-5.06087959e-01 1.70070887e-01 -1.78900640e-02 5.04967272e-01
-6.30795777e-01 8.86018097e-01 1.06748235e+00 -3.02583307e-01
-6.39808655e-01 3.61583792e-02 -3.68175119e-01 9.18076813e-01
6.77690208e-01 -9.23159778e-01 -5.29576838e-01 -8.52660537e-02
-6.56139076e-01 -6.20470792e-02 4.75057773e-02 -7.99863040e-01
-1.09114468e-01 -1.83475211e-01 1.25468910e-01 -7.73605108e-01
6.69816971e-01 -9.02845085e-01 -6.15141392e-01 6.15338981e-01
-6.65030897e-01 -6.35535717e-01 -2.77753204e-01 8.84381950e-01
-2.07962796e-01 -8.28636706e-01 8.30961466e-01 -1.98606756e-02
-9.73996341e-01 7.45218620e-02 -4.04469877e-01 4.56708759e-01
1.07491100e+00 -3.23655665e-01 -3.37695003e-01 -4.49672371e-01
-3.95156622e-01 5.31120002e-01 4.36333179e-01 8.18266392e-01
5.94581306e-01 -1.18918931e+00 -1.68829516e-01 4.16612715e-01
4.37771678e-01 -6.64628819e-02 -6.43863305e-02 2.69694358e-01
-9.42397341e-02 5.70099533e-01 2.60364920e-01 -6.53702796e-01
-1.22746611e+00 4.73706394e-01 -1.15595302e-02 -2.94806033e-01
-7.17590392e-01 9.53527629e-01 5.09958148e-01 -9.11139488e-01
6.68591082e-01 -2.00203061e-01 -1.75941989e-01 9.01134536e-02
3.19189042e-01 4.42502014e-02 -4.35432106e-01 -3.95978123e-01
-3.68748337e-01 2.31745765e-01 -4.32409763e-01 5.28250784e-02
9.05061066e-01 -3.37633699e-01 3.88415903e-01 5.77232301e-01
1.40007651e+00 4.57545556e-02 -1.51300180e+00 -4.01647717e-01
1.88910604e-01 -7.29041576e-01 -2.43628740e-01 -4.47478056e-01
-5.97312629e-01 9.94314492e-01 4.98199075e-01 4.68742937e-01
5.69553018e-01 3.55723888e-01 1.11605322e+00 5.47969580e-01
-8.89328122e-02 -1.27556598e+00 6.28157794e-01 5.78013241e-01
5.18513262e-01 -1.69459057e+00 -3.44606251e-01 -8.84776890e-01
-1.04474723e+00 1.21802449e+00 1.17326784e+00 -1.67531177e-01
4.56484050e-01 -2.48256475e-02 2.90534854e-01 -2.35077858e-01
-5.76356590e-01 -8.12548492e-03 2.17040136e-01 5.76928854e-01
1.93587825e-01 -1.62602231e-01 -2.56393015e-01 3.84712905e-01
1.36668846e-01 -3.55461359e-01 4.18744653e-01 9.18217957e-01
-5.83738446e-01 -1.26208961e+00 -2.68933982e-01 6.76965475e-01
1.79401949e-01 -3.21178325e-02 -8.62034976e-01 6.05756700e-01
-1.67592645e-01 1.03955615e+00 3.53756398e-02 -5.01075268e-01
4.74141240e-02 5.43681324e-01 -3.03736418e-01 -1.00913179e+00
-3.25465828e-01 2.40017727e-01 2.22935989e-01 -1.69696406e-01
-1.02625370e-01 -3.75657320e-01 -1.40115225e+00 5.12878783e-02
-7.46808052e-01 7.73309842e-02 2.62184381e-01 9.47181046e-01
3.98358196e-01 3.24037880e-01 9.81818378e-01 -4.44446236e-01
-8.91483605e-01 -1.02047622e+00 -4.60105091e-01 7.22344458e-01
1.76243857e-01 -7.31185675e-01 -7.99158096e-01 -5.63626923e-02] | [12.406412124633789, 7.543406009674072] |
8f6ec9ab-6251-4b1c-86e1-56c8ce481a23 | convolutional-neural-network-based-partial | 2206.14350 | null | https://arxiv.org/abs/2206.14350v1 | https://arxiv.org/pdf/2206.14350v1.pdf | Convolutional Neural Network Based Partial Face Detection | Due to the massive explanation of artificial intelligence, machine learning technology is being used in various areas of our day-to-day life. In the world, there are a lot of scenarios where a simple crime can be prevented before it may even happen or find the person responsible for it. A face is one distinctive feature that we have and can differentiate easily among many other species. But not just different species, it also plays a significant role in determining someone from the same species as us, humans. Regarding this critical feature, a single problem occurs most often nowadays. When the camera is pointed, it cannot detect a person's face, and it becomes a poor image. On the other hand, where there was a robbery and a security camera installed, the robber's identity is almost indistinguishable due to the low-quality camera. But just making an excellent algorithm to work and detecting a face reduces the cost of hardware, and it doesn't cost that much to focus on that area. Facial recognition, widget control, and such can be done by detecting the face correctly. This study aims to create and enhance a machine learning model that correctly recognizes faces. Total 627 Data have been collected from different Bangladeshi people's faces on four angels. In this work, CNN, Harr Cascade, Cascaded CNN, Deep CNN & MTCNN are these five machine learning approaches implemented to get the best accuracy of our dataset. After creating and running the model, Multi-Task Convolutional Neural Network (MTCNN) achieved 96.2% best model accuracy with training data rather than other machine learning models. | ['Md. Tarek Habib', 'Md. Sadekur Rahman', 'Taminul Islam', 'A. B. M. Raihanur Rashid', 'Tanzim Ahmed', 'Md. Towfiqul Islam'] | 2022-06-29 | null | null | null | null | ['face-detection'] | ['computer-vision'] | [-1.43840447e-01 -3.66775006e-01 1.46418095e-01 -3.36227149e-01
1.27368748e-01 -3.59587133e-01 2.64664263e-01 -1.80608347e-01
-4.56373632e-01 6.52976573e-01 -2.00612277e-01 -1.74028784e-01
7.71163180e-02 -9.24182117e-01 -3.46901417e-01 -5.85544169e-01
2.39585921e-01 3.63503665e-01 1.56987220e-01 -3.88587266e-01
3.68618399e-01 8.44835341e-01 -1.64079130e+00 2.74928600e-01
3.87389570e-01 9.24185932e-01 5.98279499e-02 5.70231020e-01
-9.59106758e-02 7.24789500e-01 -6.72237337e-01 -6.53954089e-01
4.51538175e-01 -1.99465021e-01 -7.60558844e-01 -1.87212422e-01
3.42456549e-01 -5.79776824e-01 8.01642686e-02 1.01921725e+00
5.66553175e-01 -1.67026401e-01 6.45067513e-01 -1.36076581e+00
-5.18102467e-01 4.89978343e-02 -9.55812693e-01 3.20336789e-01
2.88194895e-01 -3.51833776e-02 1.36244103e-01 -6.53687239e-01
2.02352136e-01 1.58136690e+00 7.76510537e-01 6.53760672e-01
-6.30444050e-01 -1.34090757e+00 -6.77176178e-01 1.94097951e-01
-1.48060763e+00 -4.01665896e-01 5.85977614e-01 -3.93401504e-01
6.79998934e-01 -3.40532437e-02 5.93831182e-01 7.26199746e-01
3.58938873e-01 1.03425063e-01 1.24101138e+00 -4.72161084e-01
-3.21827590e-01 4.44348156e-01 9.53749474e-03 8.73599529e-01
3.50036204e-01 -8.58211890e-02 -1.91290379e-01 1.29813477e-01
6.57641351e-01 5.47017336e-01 1.22139126e-01 5.95435560e-01
-4.63322580e-01 7.73000002e-01 3.61656696e-01 4.75209564e-01
-2.56596982e-01 -2.42050514e-01 3.09482872e-01 2.79947579e-01
-1.36002507e-02 -1.38094142e-01 -3.65045071e-01 -1.14999123e-01
-7.91143596e-01 -1.20152608e-01 7.33584225e-01 3.85054559e-01
8.09763312e-01 9.11045745e-02 4.60830331e-01 8.54788661e-01
1.77211896e-01 6.40673578e-01 6.08394444e-01 -4.28097725e-01
7.54485726e-02 1.04336452e+00 -2.73368865e-01 -1.44877231e+00
-3.89593929e-01 -4.08966243e-02 -1.01138699e+00 5.46178877e-01
5.27686298e-01 -2.52667010e-01 -1.01667070e+00 1.25670290e+00
4.21054542e-01 2.24755913e-01 -1.17452681e-01 7.67326534e-01
1.09338629e+00 8.46848667e-01 1.12470210e-01 -6.32305294e-02
1.72110295e+00 -5.69756985e-01 -5.60340464e-01 -1.63824365e-01
7.42273703e-02 -1.04493225e+00 5.11776268e-01 5.61720252e-01
-5.19901454e-01 -6.89714968e-01 -1.06050372e+00 1.15313314e-01
-6.72873497e-01 2.83563107e-01 8.18563104e-01 1.02866471e+00
-8.44311357e-01 3.95671964e-01 -2.43722677e-01 -7.72084534e-01
5.53034067e-01 7.61329770e-01 -8.81075978e-01 -1.13239095e-01
-1.01413357e+00 1.15773273e+00 2.25379527e-01 3.70115668e-01
-7.82398760e-01 -1.50982648e-01 -3.82200271e-01 -1.08128145e-01
2.09172294e-01 4.90011722e-02 7.96864986e-01 -1.58624458e+00
-9.67208207e-01 1.01454687e+00 -1.58832744e-01 -2.01172754e-01
1.83688238e-01 9.01548415e-02 -7.51955092e-01 2.47520413e-02
4.78656106e-02 5.78335762e-01 8.62287700e-01 -8.80827546e-01
-8.21999133e-01 -6.97910011e-01 -1.92655008e-02 -1.53404027e-01
-3.17076564e-01 7.47301996e-01 8.92068669e-02 -1.21395387e-01
-1.24261960e-01 -7.25538075e-01 3.26921374e-01 -1.51213165e-02
-1.50943026e-01 -4.00532573e-01 1.44922590e+00 -9.18592632e-01
8.99582267e-01 -2.17012429e+00 -4.65041846e-01 2.17100576e-01
1.00014657e-01 8.56220484e-01 1.37767866e-01 2.61934191e-01
-1.43339425e-01 4.37932044e-01 1.53313890e-01 2.18590125e-01
-5.77124357e-01 1.85935050e-01 2.64488399e-01 3.55567634e-01
2.58179039e-01 3.96631002e-01 -3.65900844e-01 -7.03206122e-01
6.00769036e-02 7.19274759e-01 -7.37404153e-02 8.42234790e-02
5.24339557e-01 1.33409321e-01 -3.79230887e-01 1.03988814e+00
9.35098886e-01 9.54695940e-02 -4.35803793e-02 -2.06936121e-01
-5.72515316e-02 -6.86643958e-01 -1.36480129e+00 7.13339210e-01
-2.61432707e-01 6.10892534e-01 1.64151832e-01 -1.09936297e+00
1.11460769e+00 5.04750729e-01 -4.77509759e-02 -6.68592811e-01
4.54602689e-01 3.43808949e-01 2.89565384e-01 -8.82259965e-01
8.43820944e-02 -2.73853153e-01 3.49581987e-01 3.00323904e-01
5.84723055e-02 4.55968320e-01 -4.92297150e-02 -1.57873511e-01
7.49221563e-01 -1.22045994e-01 4.11806285e-01 -7.60852844e-02
7.67502964e-01 -1.82323575e-01 8.35722387e-01 2.23720476e-01
-4.53411043e-01 3.81662786e-01 2.96735793e-01 -9.14540231e-01
-1.03316534e+00 -6.91221714e-01 -1.47616521e-01 9.56718504e-01
-1.09615780e-01 3.74001533e-01 -7.59440482e-01 -4.21886057e-01
1.84041858e-02 -1.10713609e-01 -6.10128164e-01 -9.93799418e-02
-3.59260291e-01 -7.27091491e-01 6.99936748e-01 5.06331064e-02
1.35260630e+00 -1.17971170e+00 -5.23339450e-01 3.42104323e-02
2.18518689e-01 -7.87331343e-01 -2.81655341e-02 -2.48799741e-01
-5.98955870e-01 -1.39150405e+00 -5.54363549e-01 -1.00716722e+00
5.98219872e-01 4.76071715e-01 6.57347679e-01 7.10720658e-01
-6.25136256e-01 -2.70012408e-01 -4.17570174e-01 -8.74094844e-01
-1.41249388e-01 -2.24147126e-01 2.11382911e-01 2.12322116e-01
8.56875956e-01 -5.38620174e-01 -5.11358559e-01 1.74734220e-01
-6.27876818e-01 -2.51959741e-01 7.38191426e-01 5.17271996e-01
-2.69634694e-01 6.02456272e-01 8.12306702e-01 -7.69668341e-01
3.21040034e-01 -6.32592797e-01 -3.02178830e-01 3.13055605e-01
-2.86618412e-01 -4.57248837e-01 6.76779270e-01 -2.78007567e-01
-1.04840457e+00 1.41274840e-01 -1.70221869e-02 -8.28599632e-02
-5.09085774e-01 1.40756369e-01 -1.22671552e-01 -3.18532735e-01
5.12509644e-01 -2.98071131e-02 3.04099381e-01 -4.23880011e-01
-3.70508701e-01 1.25690424e+00 3.28135729e-01 -1.57895342e-01
8.61117184e-01 4.29396987e-01 2.30962783e-01 -1.16538119e+00
-4.70283359e-01 -3.89474720e-01 -6.95544422e-01 -6.16022766e-01
1.19702196e+00 -7.96346605e-01 -1.16804397e+00 9.11930680e-01
-1.32939506e+00 4.23666656e-01 8.14994514e-01 4.01284605e-01
5.52316725e-01 8.64825025e-02 -2.95183599e-01 -1.24047279e+00
-4.08063024e-01 -9.88032460e-01 6.21832669e-01 9.27358091e-01
1.66011900e-01 -7.17044711e-01 -3.53887856e-01 5.60642481e-01
4.04785305e-01 4.53313202e-01 7.23404944e-01 -6.42507076e-01
-2.83855319e-01 -3.22968543e-01 -5.34751713e-01 5.82256138e-01
4.38434303e-01 4.57357615e-01 -1.26210761e+00 -9.02760550e-02
4.15720008e-02 -2.55427808e-01 6.38504684e-01 3.07383575e-02
8.36773932e-01 -3.94860238e-01 -2.42326707e-01 4.35213506e-01
1.67020929e+00 8.74538422e-01 9.19891357e-01 3.17796528e-01
6.08093560e-01 6.72120631e-01 2.86700130e-01 9.79062021e-02
1.82076082e-01 2.31607661e-01 3.55770260e-01 -2.76612550e-01
-1.44959018e-02 -1.36034459e-01 3.23345542e-01 3.69510531e-01
-5.74782312e-01 1.20777987e-01 -1.01065958e+00 2.17957750e-01
-1.26293719e+00 -1.30739844e+00 -2.86736786e-01 2.12327743e+00
4.84034598e-01 -2.03984067e-01 2.88127989e-01 5.21333933e-01
1.06094086e+00 -4.53732729e-01 -1.90107092e-01 -7.19326377e-01
-4.91646789e-02 4.81021762e-01 4.24421191e-01 1.52060539e-01
-1.05543375e+00 8.30607116e-01 5.59999847e+00 7.48410106e-01
-1.71352017e+00 1.23515792e-01 1.06929123e+00 3.87810990e-02
5.77962816e-01 -1.11348078e-01 -9.37534869e-01 7.00481713e-01
7.36297429e-01 2.16571525e-01 4.16068882e-01 8.85267019e-01
2.31401503e-01 -4.88659173e-01 -6.28453016e-01 1.25299585e+00
2.78189033e-01 -9.25257385e-01 -2.72917375e-03 9.09709185e-02
2.88445383e-01 -5.89497209e-01 -1.95382506e-01 1.36834100e-01
-1.00924984e-01 -1.59395266e+00 3.15610707e-01 3.46700460e-01
5.92005789e-01 -1.06952715e+00 1.19920909e+00 4.00934398e-01
-1.10113060e+00 -3.68799001e-01 -5.44367194e-01 -4.58219260e-01
-2.90737927e-01 2.61210084e-01 -6.91811860e-01 1.53204575e-01
1.08977711e+00 2.71365225e-01 -6.91928208e-01 9.14568186e-01
4.34100702e-02 3.89672160e-01 -3.27895522e-01 -3.13459128e-01
1.18531339e-01 -2.02296168e-01 5.09304879e-03 1.01529801e+00
4.38195974e-01 4.23919678e-01 -1.74191862e-01 3.62555265e-01
-3.07766255e-02 2.35387728e-01 -8.39270771e-01 2.10044727e-01
4.73791420e-01 1.69781125e+00 -8.76136363e-01 -7.05873519e-02
-5.39019942e-01 9.00074601e-01 3.95364650e-02 -7.47893155e-02
-7.29190648e-01 -5.89008391e-01 3.62292528e-01 2.72487998e-01
-1.86382279e-01 9.43508819e-02 -3.49051431e-02 -5.93410075e-01
-1.01186924e-01 -1.00805366e+00 2.63534725e-01 -6.90635443e-01
-1.00207317e+00 4.20363367e-01 -2.72466868e-01 -7.84150898e-01
2.77956039e-01 -8.62595439e-01 -9.73985910e-01 9.37572122e-01
-1.20768023e+00 -1.43408477e+00 -4.37807679e-01 6.48195148e-01
3.30401152e-01 -6.26983106e-01 6.41544223e-01 6.86718524e-01
-8.38683426e-01 4.62795198e-01 -2.58336127e-01 6.76610231e-01
6.78710103e-01 -6.42176569e-01 -3.36238503e-01 8.17958415e-01
-2.99975961e-01 7.10064888e-01 2.94655174e-01 -6.36857033e-01
-1.34683514e+00 -7.49358237e-01 7.98521101e-01 -1.18998326e-01
9.47801843e-02 -2.83780783e-01 -7.32571900e-01 4.74396229e-01
2.35040143e-01 -1.40555531e-01 5.41287899e-01 -3.29121239e-02
-2.38638252e-01 -6.55594409e-01 -1.64124203e+00 2.08678648e-01
4.91187900e-01 -3.83323073e-01 -4.05609220e-01 2.17700377e-01
1.10334875e-02 2.43907854e-01 -3.90672535e-01 1.22677729e-01
9.82705593e-01 -1.28527594e+00 8.18433523e-01 -5.77650607e-01
4.01965529e-01 -3.69656295e-01 4.42759618e-02 -8.35419178e-01
-1.57098249e-01 -3.01042467e-01 6.14588022e-01 1.68426371e+00
3.41445357e-01 -7.97001302e-01 7.94748545e-01 5.85465312e-01
3.65406156e-01 -5.34448147e-01 -8.59693408e-01 -5.70632577e-01
-1.03334948e-01 7.21554235e-02 6.69608235e-01 1.01008475e+00
-4.75115001e-01 3.75512064e-01 -5.26771367e-01 1.08857162e-01
5.24013996e-01 -4.68768507e-01 7.23797977e-01 -1.51157427e+00
1.66517258e-01 -6.25426993e-02 -8.48943472e-01 -1.49412841e-01
-1.32031500e-01 -5.23010910e-01 -4.48432177e-01 -1.29400730e+00
6.04261577e-01 -3.92759234e-01 2.03389898e-01 8.82746816e-01
8.22640136e-02 5.48395216e-01 1.02268636e-01 1.43046305e-01
-1.29253939e-02 -2.33879939e-01 1.13060689e+00 5.40299378e-02
1.51681691e-01 -1.35887176e-01 -6.60744071e-01 8.67766857e-01
1.05876398e+00 -3.55670869e-01 -2.07060099e-01 -3.53030443e-01
2.02644020e-01 -3.28584798e-02 5.51231146e-01 -1.21763957e+00
4.48154420e-01 -2.40171745e-01 1.04853797e+00 -2.13708222e-01
3.42917055e-01 -1.05880964e+00 4.31092978e-01 7.05383062e-01
5.96358120e-01 2.33516976e-01 1.85802937e-01 -6.75967559e-02
-1.09240539e-01 -4.24016356e-01 1.25204504e+00 -5.16484499e-01
-9.21962023e-01 2.41592646e-01 -2.90215880e-01 -4.20310289e-01
1.42756295e+00 -7.11967885e-01 -5.44831157e-01 -3.60693842e-01
-1.24028474e-01 -2.48461440e-02 1.68111071e-01 3.68748367e-01
6.24371827e-01 -1.09748757e+00 -7.07170963e-01 3.68982404e-01
-3.61327261e-01 -3.33051324e-01 3.37675303e-01 4.94700015e-01
-9.97984707e-01 2.17315912e-01 -8.86634529e-01 -2.75058895e-01
-1.88011682e+00 2.95434058e-01 4.11649406e-01 3.10458869e-01
-9.12523419e-02 7.39128709e-01 -1.53502077e-01 -1.96698785e-01
-6.99788034e-02 3.18483353e-01 -7.78118968e-01 1.47652015e-01
8.13809693e-01 5.56577384e-01 -6.08658530e-02 -1.06243694e+00
-4.90476757e-01 8.74482334e-01 -1.04001120e-01 2.57210553e-01
1.22025502e+00 2.91980952e-01 -6.15907431e-01 5.05413450e-02
1.20182252e+00 -5.09869307e-03 -4.74393338e-01 4.61242825e-01
-3.50211203e-01 -5.79029977e-01 -2.03489587e-02 -8.72807562e-01
-1.30636966e+00 1.05350423e+00 1.04028702e+00 2.55951524e-01
1.14546299e+00 -3.25937003e-01 6.50001109e-01 3.71961683e-01
5.57709575e-01 -1.20287335e+00 -3.54745574e-02 3.83267999e-01
6.62757754e-01 -1.77886963e+00 2.36815345e-02 -3.17996860e-01
-4.69949901e-01 1.42902625e+00 9.44251716e-01 -1.40182123e-01
8.37811470e-01 1.55117467e-01 1.13134593e-01 -2.02566326e-01
-2.97249198e-01 -6.24018572e-02 -1.56177029e-01 8.25715244e-01
4.92908925e-01 -5.56623526e-02 -1.94244429e-01 4.57877040e-01
-3.31118762e-01 1.38642639e-01 4.85760510e-01 8.26869905e-01
-9.22254145e-01 -1.00584424e+00 -7.94047415e-01 6.94501102e-01
-9.85601068e-01 1.44680619e-01 -5.13105512e-01 9.80090916e-01
7.57840216e-01 1.31876278e+00 1.71676800e-01 -5.92048168e-01
5.88763095e-02 -2.76152864e-02 3.32394898e-01 -3.13907653e-01
-6.84850872e-01 -4.90859836e-01 -4.41131406e-02 -1.77533016e-01
-4.13413346e-01 -3.41578454e-01 -1.15900052e+00 -1.04748809e+00
-2.92748094e-01 8.62793922e-02 9.72649932e-01 8.49259734e-01
-1.19538806e-01 -2.83399466e-02 7.90694475e-01 -5.04844964e-01
4.21564095e-02 -1.05302680e+00 -6.91154182e-01 3.10676247e-01
6.82131648e-02 -8.40060174e-01 -3.44010711e-01 1.05742887e-02] | [13.317132949829102, 0.8842295408248901] |
0f173c1c-d06d-4ed4-868d-2bd37623127a | joint-graph-learning-from-gaussian | 2212.01816 | null | https://arxiv.org/abs/2212.01816v1 | https://arxiv.org/pdf/2212.01816v1.pdf | Joint graph learning from Gaussian observations in the presence of hidden nodes | Graph learning problems are typically approached by focusing on learning the topology of a single graph when signals from all nodes are available. However, many contemporary setups involve multiple related networks and, moreover, it is often the case that only a subset of nodes is observed while the rest remain hidden. Motivated by this, we propose a joint graph learning method that takes into account the presence of hidden (latent) variables. Intuitively, the presence of the hidden nodes renders the inference task ill-posed and challenging to solve, so we overcome this detrimental influence by harnessing the similarity of the estimated graphs. To that end, we assume that the observed signals are drawn from a Gaussian Markov random field with latent variables and we carefully model the graph similarity among hidden (latent) nodes. Then, we exploit the structure resulting from the previous considerations to propose a convex optimization problem that solves the joint graph learning task by providing a regularized maximum likelihood estimator. Finally, we compare the proposed algorithm with different baselines and evaluate its performance over synthetic and real-world graphs. | ['Antonio G. Marques', 'Santiago Segarra', 'Andrei Buciulea', 'Madeline Navarro', 'Samuel Rey'] | 2022-12-04 | null | null | null | null | ['graph-similarity'] | ['graphs'] | [ 2.69574493e-01 4.86972213e-01 -2.18004629e-01 -8.29444006e-02
-5.50495088e-01 -4.30629998e-01 6.43158376e-01 2.45118901e-01
-3.41031514e-02 6.92251027e-01 3.83433923e-02 -6.22809641e-02
-1.92988530e-01 -7.12198138e-01 -9.01972771e-01 -9.03782308e-01
-4.12777483e-01 5.55597544e-01 -7.87368715e-02 4.62648243e-01
-1.63015053e-01 3.22804809e-01 -7.11298645e-01 -5.77573299e-01
6.73433304e-01 6.04336977e-01 2.39231467e-01 5.36971509e-01
3.25003356e-01 7.65253544e-01 -3.56495380e-01 -3.39571714e-01
7.32963905e-02 -3.61147344e-01 -4.84198660e-01 7.32300222e-01
2.23918855e-01 5.83815798e-02 -5.57481229e-01 1.19614780e+00
1.69908687e-01 6.32518157e-02 7.23502219e-01 -1.44275010e+00
-1.09781355e-01 7.46770501e-01 -8.35387588e-01 -8.48146006e-02
1.14751533e-01 -1.01541683e-01 1.39255929e+00 -4.04422641e-01
6.86573207e-01 1.13403428e+00 5.01779377e-01 -6.75569922e-02
-1.72178662e+00 -4.57453668e-01 5.78156471e-01 -8.95413980e-02
-1.40233266e+00 -4.08745199e-01 1.22226012e+00 -5.05622268e-01
-2.41068099e-03 -2.22733300e-02 5.27969003e-01 1.35480988e+00
-2.01168787e-02 6.59088850e-01 8.45144928e-01 -3.20805937e-01
3.42642426e-01 1.49541467e-01 8.60749930e-02 9.65605795e-01
4.13225949e-01 -3.09497178e-01 -4.17481124e-01 -4.70505625e-01
5.36866784e-01 1.16992451e-01 -3.60637635e-01 -8.53306234e-01
-9.75510657e-01 9.77384150e-01 3.01826119e-01 1.27238378e-01
-6.25704110e-01 2.35605761e-01 -1.32169664e-01 1.84557781e-01
3.98759782e-01 -1.29039995e-02 -3.23968604e-02 5.07235706e-01
-1.00804281e+00 -9.17339399e-02 1.01874352e+00 7.82437563e-01
1.11006606e+00 -1.06366184e-02 1.69046313e-01 3.86512429e-01
7.35413790e-01 3.45256478e-01 -3.02597791e-01 -6.53206944e-01
5.09089887e-01 5.19164920e-01 8.85191374e-03 -1.46121001e+00
-1.36092991e-01 -7.39194453e-01 -1.14519012e+00 -1.59870893e-01
6.35537148e-01 -4.09388512e-01 -7.45861232e-01 1.98840070e+00
2.41515517e-01 5.87035596e-01 -2.12814406e-01 6.49332702e-01
2.61958569e-01 4.47111368e-01 -8.63795504e-02 -5.04036665e-01
8.42570305e-01 -6.76327050e-01 -5.90210021e-01 -3.48542809e-01
3.13570499e-01 -4.19418931e-01 3.61299545e-01 1.82089075e-01
-8.18962157e-01 -2.28155777e-01 -9.64509726e-01 5.64224362e-01
1.94304183e-01 1.97902679e-01 5.68539083e-01 4.89308894e-01
-8.52072477e-01 5.48059642e-01 -1.10304213e+00 -2.58160949e-01
1.42083943e-01 2.26913586e-01 -3.77850205e-01 -2.16027275e-01
-8.16522002e-01 2.24049434e-01 9.42981616e-02 2.74814785e-01
-1.22190464e+00 -1.27303764e-01 -8.30716252e-01 1.24244325e-01
9.50866997e-01 -6.46788895e-01 6.83938324e-01 -8.62502038e-01
-1.15446496e+00 3.40345472e-01 -3.54498148e-01 -4.09022331e-01
6.27899587e-01 5.82838133e-02 -1.80935040e-01 2.54687250e-01
1.56967387e-01 -3.49908955e-02 1.17410326e+00 -1.33686173e+00
-9.96711180e-02 -3.20033908e-01 3.31379659e-02 -1.05762862e-01
-1.77399039e-01 -4.80129689e-01 -7.37334669e-01 -4.33850378e-01
3.27649593e-01 -1.23165190e+00 -3.48791927e-01 -1.04769960e-01
-1.04599869e+00 1.10525630e-01 5.84898710e-01 -6.34172857e-01
1.00670934e+00 -2.13975477e+00 4.79844958e-01 7.28798151e-01
6.81802630e-01 -2.42050320e-01 -8.29522386e-02 6.47924364e-01
6.65021837e-02 -6.25018999e-02 -9.54595283e-02 -6.97766542e-01
-1.17220089e-01 2.59333491e-01 -4.53827232e-02 9.24174070e-01
-1.98349446e-01 6.60285175e-01 -8.93295348e-01 -4.66570526e-01
1.71043575e-01 4.01147306e-01 -3.96597505e-01 2.77726203e-01
-2.70021528e-01 7.53946662e-01 -8.34265113e-01 2.71685898e-01
5.80418885e-01 -8.40170503e-01 8.01000416e-01 -1.43648490e-01
3.80625635e-01 1.35243684e-01 -1.56049395e+00 1.38345528e+00
-1.99086756e-01 5.75402081e-01 3.98178875e-01 -1.37838805e+00
6.63797438e-01 4.71854270e-01 6.39646053e-01 2.43933443e-02
1.26914000e-02 -5.19304350e-03 -1.73690710e-02 -3.06201458e-01
-2.64325321e-01 -1.49915013e-02 1.05518304e-01 4.39788967e-01
1.56325385e-01 1.95338935e-01 1.43037468e-01 5.84229589e-01
1.33626437e+00 -2.03469828e-01 3.03413868e-01 1.44929007e-01
3.82030696e-01 -5.34067214e-01 6.37248695e-01 9.20769751e-01
7.58502185e-02 3.18852931e-01 9.64706719e-01 2.98440475e-02
-7.82627046e-01 -1.13294184e+00 3.88167530e-01 5.44410408e-01
2.96118148e-02 -5.47583818e-01 -5.78164399e-01 -8.18621874e-01
-8.81475732e-02 3.09873015e-01 -6.64049685e-01 -3.58500220e-02
-2.33038768e-01 -7.34493554e-01 -8.35604221e-03 9.26071927e-02
1.33314729e-01 -4.93573070e-01 -1.85562321e-03 2.49374732e-01
-2.90322870e-01 -1.41095459e+00 -3.49673957e-01 1.32375032e-01
-7.11514056e-01 -1.19065797e+00 -5.79688132e-01 -4.78886962e-01
8.44136417e-01 2.53876716e-01 1.05459535e+00 9.64926407e-02
-6.06022254e-02 4.55054194e-01 -4.52357605e-02 9.48013887e-02
-3.62108707e-01 1.96315512e-01 -2.07894623e-01 7.37333298e-01
-2.84206271e-01 -9.88088608e-01 -2.17767954e-01 6.84057400e-02
-5.67227602e-01 1.27478838e-01 8.44914198e-01 5.69899082e-01
4.23528910e-01 4.19811875e-01 3.17607820e-01 -1.13452673e+00
3.10549408e-01 -7.26334155e-01 -8.47268820e-01 4.30322587e-01
-4.16900337e-01 4.37380433e-01 5.80582201e-01 -4.90772069e-01
-8.07922781e-01 2.12655038e-01 5.80119491e-01 -5.13911903e-01
-1.13107920e-01 8.17724586e-01 -4.50225979e-01 2.78317221e-02
4.57249302e-03 7.33733773e-02 -2.85574305e-03 -4.95865971e-01
3.11231762e-01 2.25119710e-01 2.65486717e-01 -3.97421420e-01
1.17121935e+00 6.23474121e-01 5.60208857e-01 -1.05645728e+00
-7.52294838e-01 -3.92698169e-01 -5.65556705e-01 -3.30819696e-01
5.84665000e-01 -8.61862540e-01 -6.51247442e-01 3.44670713e-01
-9.44263160e-01 -1.60089731e-01 1.00399546e-01 6.60439074e-01
-2.85251319e-01 6.82906330e-01 -3.68788064e-01 -1.04603565e+00
2.88915932e-01 -1.03874087e+00 9.86702979e-01 -1.75841808e-01
-1.83668941e-01 -1.53900957e+00 2.31922448e-01 2.32244834e-01
-9.50673148e-02 5.14190137e-01 9.39664543e-01 -5.47238886e-01
-1.00832474e+00 -4.21971500e-01 -1.63555175e-01 -2.15599965e-02
2.66776234e-01 -2.94971969e-02 -7.27130294e-01 -3.47421557e-01
-1.25579894e-01 3.44391330e-03 8.03028524e-01 4.97359723e-01
6.46591127e-01 -3.08199286e-01 -5.85704744e-01 4.40026224e-01
1.35838830e+00 -4.54016596e-01 2.71449313e-02 -2.60984808e-01
8.59400332e-01 7.27741063e-01 -7.94905704e-04 5.79517245e-01
5.52433133e-01 5.96823752e-01 5.62165499e-01 -1.14569053e-01
2.15213746e-01 -5.96368432e-01 2.28532851e-01 7.57891119e-01
1.05332278e-01 -6.54098928e-01 -8.34563136e-01 4.80749905e-01
-1.90395188e+00 -7.00697243e-01 -2.65033603e-01 2.27255654e+00
4.64912951e-01 2.47158021e-01 1.14961639e-01 6.36233836e-02
9.96107996e-01 4.30029184e-01 -4.15090084e-01 5.89704633e-01
-8.70351717e-02 -2.35248238e-01 5.18792450e-01 6.53485179e-01
-1.02860367e+00 5.12528539e-01 5.65389252e+00 5.36601126e-01
-9.65876400e-01 -1.38266131e-01 4.98875201e-01 1.59109429e-01
-3.74870598e-01 4.86282498e-01 -5.40264666e-01 3.48144472e-01
8.06588173e-01 -9.21530798e-02 5.23295820e-01 4.80022460e-01
3.18906248e-01 -1.83876216e-01 -1.14106202e+00 7.12143838e-01
-7.52925966e-03 -8.46839190e-01 -1.79746896e-01 5.84687293e-01
7.03028917e-01 1.34593695e-02 -1.06229752e-01 -7.93182403e-02
6.58011138e-01 -9.08079982e-01 5.22113562e-01 6.27266169e-01
3.19672853e-01 -3.58730048e-01 4.03865039e-01 6.51650846e-01
-1.34403920e+00 1.43891364e-01 -8.21312368e-02 4.72399732e-03
3.23030412e-01 1.06399310e+00 -8.42936218e-01 7.71056890e-01
9.75752026e-02 9.69561279e-01 -4.39906806e-01 1.00058043e+00
-5.94581306e-01 9.31890309e-01 -4.61427778e-01 2.62277663e-01
1.23654485e-01 -4.93162632e-01 8.33743811e-01 6.11701846e-01
2.03294605e-01 -3.63820642e-01 6.12035930e-01 9.82341528e-01
-2.74403930e-01 -1.13113947e-01 -6.42657757e-01 -3.46685052e-01
4.06824082e-01 1.33448601e+00 -9.91832197e-01 -1.04177237e-01
-7.68692255e-01 8.52525711e-01 5.97306430e-01 8.23262095e-01
-7.03789711e-01 1.21871546e-01 1.17575422e-01 1.42716229e-01
5.02427161e-01 -4.07479316e-01 2.67686725e-01 -1.31583703e+00
1.91371426e-01 -6.36314750e-01 2.98664808e-01 -3.64029586e-01
-1.27454746e+00 2.32815802e-01 -2.54028365e-02 -7.69949615e-01
-3.65320086e-01 -1.13918550e-01 -7.12617278e-01 7.76458979e-01
-1.45604599e+00 -1.15259135e+00 -3.01307946e-01 4.98414457e-01
6.10501878e-02 2.22550377e-01 3.48178387e-01 1.11496672e-01
-7.65431404e-01 2.39596099e-01 9.46732089e-02 2.78312325e-01
4.18256283e-01 -1.33430600e+00 2.11261526e-01 9.87755835e-01
5.26361644e-01 4.91683751e-01 7.54766762e-01 -8.21175694e-01
-1.42982173e+00 -9.61862266e-01 8.15870941e-01 -7.50407577e-02
9.11326170e-01 -7.42070973e-01 -8.53363633e-01 1.06030726e+00
-1.33485021e-02 2.75107265e-01 3.80356997e-01 2.45581090e-01
-2.20122233e-01 1.24171630e-01 -6.02597594e-01 4.47724998e-01
8.16953361e-01 -6.02625072e-01 6.84464797e-02 3.68023902e-01
2.69310266e-01 1.63527057e-01 -6.27532244e-01 1.88026875e-01
2.44868979e-01 -7.39831626e-01 7.32708573e-01 -3.78950179e-01
-1.10441357e-01 -2.08723560e-01 4.55851518e-02 -1.40270615e+00
-1.72630161e-01 -9.18669343e-01 -2.93638647e-01 1.31545973e+00
4.31859910e-01 -5.14377177e-01 9.68299150e-01 4.05346632e-01
5.22840500e-01 -2.97056317e-01 -7.79560328e-01 -7.37260699e-01
-4.71494198e-01 -2.23579004e-01 1.03473783e-01 8.99234593e-01
-1.19353279e-01 7.72244215e-01 -7.58416474e-01 6.25107229e-01
1.15616286e+00 2.67285377e-01 9.42354500e-01 -1.45545912e+00
-8.16942096e-01 7.58313239e-02 -3.29262137e-01 -1.19605470e+00
5.19793689e-01 -7.12124348e-01 6.67524859e-02 -1.26814461e+00
4.85460848e-01 -4.22198951e-01 -1.14890980e-02 9.73605961e-02
-4.11808848e-01 -1.84964880e-01 7.21030757e-02 2.59605020e-01
-7.20955968e-01 5.21676064e-01 9.79977190e-01 -1.15410171e-01
-8.89273137e-02 4.05714273e-01 -3.62802476e-01 8.02119076e-01
4.86326247e-01 -5.98027349e-01 -5.14275610e-01 -1.24888301e-01
3.58194828e-01 5.27783513e-01 5.68071425e-01 -7.75783241e-01
3.94551963e-01 -8.59493464e-02 -2.98436712e-02 -2.41683483e-01
3.47523838e-01 -9.09415483e-01 4.22934979e-01 1.90076157e-01
-2.54026979e-01 -3.86170119e-01 -5.50728858e-01 1.16208017e+00
-8.67478848e-02 -2.20199138e-01 5.13125598e-01 -1.32052759e-02
-7.50540495e-02 5.30363977e-01 -3.63613784e-01 1.83765158e-01
8.93396020e-01 2.32892752e-01 2.07374707e-01 -9.84296501e-01
-1.15119672e+00 2.56513298e-01 4.56208587e-01 -1.23045750e-01
2.75447130e-01 -8.64961147e-01 -5.27200103e-01 1.99555680e-01
-5.25267087e-02 -7.20441267e-02 -7.09002139e-03 1.18688810e+00
2.12062314e-01 1.65055305e-01 3.68921131e-01 -5.60245037e-01
-1.19703615e+00 6.15717292e-01 1.66047707e-01 -5.57122529e-01
-5.42737722e-01 4.94387418e-01 2.80276537e-01 -2.83775777e-01
3.32577288e-01 4.41780239e-02 -1.94356754e-01 1.13119006e-01
-1.36635214e-01 4.72416461e-01 -3.62571985e-01 -6.16445422e-01
-5.09019569e-03 3.48425061e-01 -1.31489756e-02 -1.84551254e-01
1.36828351e+00 -4.73151982e-01 -3.98304733e-03 6.52061403e-01
1.19864535e+00 4.63530272e-01 -1.40504146e+00 -7.17940569e-01
2.08113611e-01 -2.55450666e-01 5.19085526e-02 -3.59577797e-02
-1.22040451e+00 8.26029778e-01 -1.18648082e-01 5.65885246e-01
7.76568115e-01 2.46267408e-01 3.83958668e-01 2.11947396e-01
2.74213940e-01 -6.03961110e-01 1.79863811e-01 6.73922673e-02
2.67332017e-01 -1.13530004e+00 5.37068732e-02 -8.14271092e-01
-2.57709265e-01 8.60947549e-01 1.09630609e-02 -2.86896735e-01
8.48183274e-01 -2.41951402e-02 -3.19448054e-01 -4.24810976e-01
-7.71288395e-01 -1.41096547e-01 1.59320116e-01 4.19877261e-01
1.80629209e-01 1.22383341e-01 2.06420287e-01 7.67700896e-02
5.45313805e-02 -2.72768885e-01 5.64575732e-01 6.79306567e-01
-7.22927079e-02 -1.17026722e+00 -1.75361767e-01 2.77962506e-01
-4.83767837e-01 -4.25411686e-02 -5.02133310e-01 6.42356336e-01
-2.73497909e-01 1.01519465e+00 -2.80732870e-01 3.83488461e-02
-1.04673401e-01 -8.47355947e-02 3.00833941e-01 -6.82792902e-01
2.25954548e-01 5.71702302e-01 1.08477911e-02 -3.09585214e-01
-4.94525790e-01 -7.88676620e-01 -6.74221992e-01 -2.01484531e-01
-5.78842521e-01 3.87161374e-01 5.31473637e-01 1.07054985e+00
1.60281628e-01 4.84162658e-01 7.59865165e-01 -6.60445571e-01
-5.75055778e-01 -8.52802098e-01 -1.01198280e+00 3.50611955e-01
6.48965895e-01 -7.00865805e-01 -8.45093668e-01 3.02816350e-02] | [7.05478572845459, 5.1385698318481445] |
3be13f1f-8fbb-4ad7-9ead-e36cc987de37 | deep-attention-spatio-temporal-point | 2002.07281 | null | https://arxiv.org/abs/2002.07281v5 | https://arxiv.org/pdf/2002.07281v5.pdf | Deep Fourier Kernel for Self-Attentive Point Processes | We present a novel attention-based model for discrete event data to capture complex non-linear temporal dependence structures. We borrow the idea from the attention mechanism and incorporate it into the point processes' conditional intensity function. We further introduce a novel score function using Fourier kernel embedding, whose spectrum is represented using neural networks, which drastically differs from the traditional dot-product kernel and can capture a more complex similarity structure. We establish our approach's theoretical properties and demonstrate our approach's competitive performance compared to the state-of-the-art for synthetic and real data. | ['Minghe Zhang', 'Yao Xie', 'Shixiang Zhu', 'Ruyi Ding'] | 2020-02-17 | null | null | null | null | ['deep-attention', 'deep-attention'] | ['computer-vision', 'natural-language-processing'] | [-4.23544347e-02 -4.30290610e-01 -1.13970600e-01 -3.51112783e-01
-5.03387034e-01 -9.29079875e-02 8.51637185e-01 4.75003988e-01
-5.03755391e-01 5.65478921e-01 2.21970394e-01 -1.84125617e-01
-4.45718855e-01 -8.00800264e-01 -4.81647074e-01 -6.25143826e-01
-7.32908368e-01 3.71392429e-01 5.79427063e-01 -2.13248417e-01
1.49452806e-01 5.77218890e-01 -1.23508143e+00 4.09786887e-02
6.96156025e-01 8.50172520e-01 -2.36965448e-01 7.35818744e-01
3.63127403e-02 1.06880391e+00 -5.64914644e-01 -2.87886232e-01
-1.95864767e-01 -2.62403965e-01 -6.04807317e-01 -3.82951766e-01
-2.33335540e-01 -6.21551946e-02 -9.65026379e-01 7.15410590e-01
3.66391331e-01 3.97820979e-01 1.27714443e+00 -1.29034030e+00
-1.40209365e+00 -7.65069798e-02 -5.39470136e-01 9.98559117e-01
2.27101266e-01 -5.71264587e-02 1.00027657e+00 -9.47798669e-01
1.67852074e-01 1.21185064e+00 9.32707787e-01 1.32255882e-01
-1.42745602e+00 -3.06462705e-01 1.97715864e-01 5.59328139e-01
-1.30767894e+00 1.99627355e-01 7.84232140e-01 -5.13384640e-01
1.17984736e+00 -1.91541761e-02 5.54291189e-01 1.38287675e+00
8.12120914e-01 9.42233324e-01 9.69406784e-01 -2.29665264e-01
3.98711860e-01 -3.89441371e-01 3.52253437e-01 4.13157940e-01
-2.15291027e-02 4.48041975e-01 -3.57597888e-01 -4.32779521e-01
1.04450059e+00 5.22773683e-01 -2.08129421e-01 -2.33747467e-01
-1.28396153e+00 1.09033537e+00 4.15655553e-01 3.95074874e-01
-6.84365571e-01 4.34547395e-01 2.63082922e-01 2.11520091e-01
6.47679389e-01 -1.10380083e-01 -4.73677546e-01 -2.36821353e-01
-7.31411636e-01 1.15327351e-01 8.47797692e-01 6.54361367e-01
3.72969598e-01 -3.41538377e-02 -6.81179762e-01 5.67437947e-01
1.84004351e-01 3.45282972e-01 5.17370284e-01 -6.25709593e-01
5.34276441e-02 3.03945303e-01 1.32275820e-01 -8.26755226e-01
-3.47628832e-01 -2.35902578e-01 -1.12800217e+00 9.59810838e-02
1.60096988e-01 2.46014018e-02 -7.56684780e-01 1.75605726e+00
1.01604179e-01 8.77528846e-01 -7.17949793e-02 6.22350276e-01
2.55618006e-01 9.38664198e-01 2.38662988e-01 -3.39098364e-01
1.14147317e+00 -8.21428239e-01 -1.07359564e+00 3.29260558e-01
-9.69858021e-02 -2.46651843e-01 9.34563279e-01 -1.24506960e-02
-1.06910992e+00 -6.47963285e-01 -7.54462183e-01 -5.54092117e-02
-6.41002595e-01 -3.03399980e-01 9.55023348e-01 3.35668087e-01
-1.04443789e+00 9.06923592e-01 -1.11388385e+00 -3.93579900e-01
4.28601861e-01 2.26961911e-01 -1.92975983e-01 2.11491525e-01
-1.61060178e+00 7.94626057e-01 1.47187769e-01 -1.52402535e-01
-8.61819446e-01 -7.94063091e-01 -8.36577833e-01 2.96847373e-01
1.22604057e-01 -6.86347008e-01 1.25593293e+00 -3.14408481e-01
-1.66165030e+00 2.99347848e-01 -4.68147360e-02 -4.40147191e-01
1.74610391e-01 -3.23268890e-01 -6.84286594e-01 3.42940778e-01
1.95713211e-02 1.83423012e-02 8.35351050e-01 -7.28152394e-01
-3.41264486e-01 -5.93149886e-02 -1.24964654e-01 -1.67113334e-01
-2.99947500e-01 3.29529732e-01 -2.70328194e-01 -1.01704669e+00
-2.49630973e-01 -5.76451838e-01 -4.38150883e-01 1.47988796e-01
8.75522122e-02 -5.29236674e-01 7.20640361e-01 -4.11527157e-01
1.46775270e+00 -2.10496402e+00 6.36969460e-03 -7.02863038e-02
2.10191309e-01 3.86799797e-02 -3.03617120e-02 1.04852009e+00
-3.01211476e-01 -7.77698457e-02 -6.01191223e-01 -5.20761907e-01
1.88865632e-01 3.53489339e-01 -3.78863990e-01 6.02188349e-01
9.03604448e-01 1.06223142e+00 -1.17908144e+00 -3.11000317e-01
3.04683149e-01 6.75543725e-01 -3.73744458e-01 3.01907152e-01
1.52428612e-01 1.41291305e-01 -3.04393500e-01 4.40563709e-01
4.55501467e-01 -4.60326612e-01 -3.39728653e-01 4.99329977e-02
-1.00099176e-01 1.26276404e-01 -8.96730721e-01 1.45879710e+00
-2.88723081e-01 6.26228690e-01 -4.57906097e-01 -1.21995187e+00
8.05066586e-01 5.37034273e-01 6.30482674e-01 -2.71793991e-01
-1.00026187e-02 -3.26352865e-01 -2.13586167e-02 -4.70477492e-01
2.93590903e-01 -4.73022461e-01 -2.92232215e-01 4.21582043e-01
4.22809511e-01 3.60297501e-01 -2.03843459e-01 1.87647734e-02
1.49309492e+00 -4.75118384e-02 4.98478353e-01 -4.84074146e-01
3.74161243e-01 -5.41031003e-01 3.91212940e-01 9.84288216e-01
-6.18942261e-01 3.65146428e-01 6.18965209e-01 -4.73393261e-01
-9.18709397e-01 -1.73176479e+00 -2.24758714e-01 1.02060902e+00
-1.12845145e-01 -1.14886180e-01 -2.51011521e-01 -7.31536329e-01
2.23097295e-01 7.43340254e-01 -1.21456826e+00 -3.52118731e-01
-3.84495467e-01 -1.04614985e+00 6.87067688e-01 1.10513759e+00
3.06286693e-01 -1.12182426e+00 -3.42875659e-01 5.16431451e-01
1.55954463e-02 -7.37068176e-01 -5.96076548e-01 4.06262308e-01
-7.39288807e-01 -9.00941372e-01 -7.08602548e-01 -5.11178970e-01
-5.49591072e-02 -1.06264755e-01 1.02056277e+00 -7.08416879e-01
-3.67144912e-01 7.11469829e-01 -2.93716103e-01 -6.77074134e-01
-6.45191362e-03 -2.01099738e-01 1.08420596e-01 3.51896018e-01
6.88582838e-01 -8.16920936e-01 -7.67820120e-01 8.86137486e-02
-1.28615594e+00 -6.56871080e-01 5.98526478e-01 9.57084477e-01
3.56163293e-01 1.02180630e-01 9.59162712e-01 -3.22253823e-01
1.00049138e+00 -9.28918481e-01 -3.05857122e-01 8.08587000e-02
-3.55890483e-01 3.00362825e-01 6.04697168e-01 -8.66583467e-01
-1.13581431e+00 -4.76946622e-01 1.38396442e-01 -8.43256652e-01
-1.97059095e-01 4.62836683e-01 3.10338765e-01 2.51803517e-01
3.58487934e-01 3.51528645e-01 -2.14398466e-02 -4.52557832e-01
3.59328330e-01 3.53888005e-01 6.46997511e-01 -5.55382192e-01
4.89785165e-01 9.83880877e-01 1.79649249e-01 -5.80829501e-01
-5.12644589e-01 -6.60749137e-01 -6.49981499e-01 4.37307954e-02
9.24171567e-01 -6.54388547e-01 -9.80613887e-01 4.29323047e-01
-1.52577019e+00 -1.88036308e-01 -6.55067623e-01 9.82337594e-01
-9.17548954e-01 1.66470841e-01 -1.39387476e+00 -1.00138712e+00
-1.13286689e-01 -4.26068455e-01 1.20998812e+00 -1.34976298e-01
3.88167687e-02 -1.55315840e+00 8.75960529e-01 -6.87626421e-01
5.06867409e-01 3.81100208e-01 8.94549251e-01 -8.13658118e-01
-2.27446780e-01 -3.64219993e-01 -3.54918897e-01 3.60254169e-01
1.17403060e-01 6.24897406e-02 -7.37831473e-01 1.67619658e-03
4.90353614e-01 2.58047283e-01 1.00370049e+00 4.90470320e-01
1.10274768e+00 2.30660692e-01 -3.27884048e-01 3.63841951e-01
1.47546375e+00 2.38353193e-01 8.50877643e-01 -6.17328323e-02
3.52896124e-01 5.03422141e-01 1.88114405e-01 5.63559890e-01
4.08726037e-01 4.44527596e-01 1.94933549e-01 -1.49838209e-01
3.91561031e-01 -2.05874309e-01 4.30474967e-01 1.21353304e+00
-3.88499975e-01 -2.48571143e-01 -7.96699584e-01 7.66640842e-01
-2.23040509e+00 -1.48479772e+00 -1.96764171e-01 2.04205585e+00
6.34684563e-01 1.51161864e-01 7.43869543e-02 -2.57599559e-02
8.41655433e-01 2.66143590e-01 -3.29240531e-01 -5.67309380e-01
-2.10619010e-02 7.17826247e-01 5.06383419e-01 2.50032276e-01
-1.34495378e+00 6.61593378e-01 8.18815994e+00 8.99403334e-01
-7.34835267e-01 4.62589651e-01 1.22367337e-01 9.26804617e-02
4.10064943e-02 -2.93696195e-01 -4.07716513e-01 6.34973824e-01
1.42866683e+00 -3.58071029e-01 7.68500939e-02 4.04074728e-01
2.34473214e-01 2.73379564e-01 -1.22274709e+00 9.24998939e-01
1.68242212e-02 -1.07066083e+00 -2.89012901e-02 1.10943936e-01
5.85030854e-01 6.24784231e-02 1.20850004e-01 6.13916218e-01
5.82345426e-01 -1.06341124e+00 4.05566573e-01 1.09295619e+00
3.61542642e-01 -8.78446758e-01 6.57047629e-01 2.06416756e-01
-1.46326470e+00 -1.38100967e-01 -4.38484788e-01 -3.86005580e-01
5.34024119e-01 6.61268532e-01 -3.73019218e-01 7.48710811e-01
6.57683074e-01 9.32816088e-01 -2.51521587e-01 1.00978208e+00
-2.99827661e-02 7.91224658e-01 -3.58947366e-01 -1.54795781e-01
3.80664170e-01 -2.48321742e-02 5.55690527e-01 1.55673909e+00
3.57833803e-01 1.07568882e-01 6.40672520e-02 1.01885176e+00
2.91253716e-01 -9.35526639e-02 -7.90702224e-01 -1.39255375e-02
2.26508245e-01 7.72524118e-01 -6.35840237e-01 -4.78394955e-01
-8.72228086e-01 1.42227137e+00 5.01764059e-01 5.35743177e-01
-1.31646335e+00 -6.25155747e-01 8.86149645e-01 -1.35204017e-01
6.70228183e-01 -4.04066592e-01 1.73721671e-01 -1.23911893e+00
-7.31046498e-02 -9.31806639e-02 6.32751107e-01 -7.69637525e-01
-2.19679666e+00 4.51575398e-01 3.44714433e-01 -1.24345493e+00
-2.17428952e-02 -6.56540692e-01 -9.14192855e-01 9.50766802e-01
-1.47259605e+00 -9.39632893e-01 1.78184256e-01 9.61956680e-01
3.33845228e-01 2.18020454e-01 9.31855857e-01 3.25712442e-01
-3.20461035e-01 2.69505024e-01 4.72832054e-01 3.98258306e-02
5.66170216e-01 -1.57679915e+00 6.85881197e-01 3.76469076e-01
9.85139087e-02 6.38314128e-01 4.09910917e-01 -6.23831451e-01
-1.05036700e+00 -1.10246062e+00 7.81927109e-01 -7.12876976e-01
1.27400756e+00 -2.32714400e-01 -1.07467842e+00 8.49564672e-01
2.94781536e-01 1.09426066e-01 1.00657833e+00 1.62673518e-01
-3.65321457e-01 2.09377602e-01 -9.20444787e-01 6.68447554e-01
9.89721298e-01 -7.44183600e-01 -1.25164664e+00 1.10246375e-01
9.04154003e-01 3.34589481e-01 -1.04321039e+00 6.02960885e-01
3.30268949e-01 -7.05325007e-01 1.17233860e+00 -1.03050363e+00
3.67745250e-01 -1.99499890e-01 -2.02965111e-01 -1.28889227e+00
-1.09713948e+00 -5.67017555e-01 -5.14810562e-01 7.97272265e-01
2.14826576e-02 -9.06616747e-01 2.16775253e-01 2.95173109e-01
-4.10070382e-02 -7.81454623e-01 -1.16280794e+00 -1.10705769e+00
3.49101990e-01 -7.10581660e-01 3.82206231e-01 1.07984567e+00
2.16010749e-01 2.93715447e-01 -4.01788831e-01 3.18833500e-01
6.74933791e-01 -9.20784697e-02 1.18938603e-01 -1.42330086e+00
-4.41662401e-01 -3.73638988e-01 -7.84168422e-01 -9.35221612e-01
1.49415493e-01 -7.01623023e-01 -1.54139504e-01 -1.33072507e+00
1.44595265e-01 1.43269971e-01 -1.05224359e+00 1.43403979e-02
-3.29443634e-01 -1.30570149e-02 -1.41790509e-01 2.58964926e-01
-7.77799547e-01 1.15702224e+00 8.87633741e-01 9.29277241e-02
-1.08487658e-01 -4.77660894e-02 -2.84341574e-01 7.56568730e-01
6.79409802e-01 -5.78091145e-01 -3.28223199e-01 -3.36885899e-02
-1.35259092e-01 1.97832257e-01 7.23709762e-01 -1.05767155e+00
4.47387129e-01 -2.83578634e-01 3.49456310e-01 -6.28113985e-01
4.44574386e-01 -7.12050378e-01 -1.87328793e-02 2.48829171e-01
-3.66080552e-01 3.50211501e-01 3.01984221e-01 1.18587458e+00
-3.57133299e-01 7.30215460e-02 3.85309249e-01 1.45829111e-01
-3.39627653e-01 7.13181734e-01 -7.49017596e-01 2.67722476e-02
1.15128863e+00 1.72363117e-01 -1.35528386e-01 -5.59808612e-01
-1.02520931e+00 -7.71700293e-02 5.41204363e-02 3.96975696e-01
6.29263043e-01 -1.84868515e+00 -6.42883837e-01 1.69068441e-01
1.33329481e-01 -5.99779069e-01 3.73858422e-01 1.04459321e+00
-2.14243710e-01 3.45935285e-01 6.07839935e-02 -6.16988719e-01
-4.90928173e-01 1.14903247e+00 3.31747113e-03 -6.70630813e-01
-6.67743564e-01 4.37399715e-01 5.40043592e-01 -1.69683874e-01
-2.11109325e-01 -6.97154760e-01 -8.19140077e-02 -1.46516755e-01
6.88211977e-01 4.73842204e-01 -3.70884806e-01 -2.87548751e-01
-4.14277047e-01 3.32943678e-01 9.25472602e-02 -4.40281630e-01
1.22189462e+00 6.23625889e-02 -4.72084843e-02 1.23181975e+00
1.42221439e+00 -3.58843386e-01 -1.44491684e+00 -4.42094892e-01
6.57755286e-02 -2.77672768e-01 -1.94617033e-01 -3.86576772e-01
-4.13720608e-01 1.02888930e+00 5.91921628e-01 7.23583460e-01
9.65081275e-01 2.36606851e-01 5.82103670e-01 2.09457263e-01
2.98977226e-01 -1.03947330e+00 3.85174364e-01 7.47248113e-01
9.10168707e-01 -9.84581649e-01 -3.53640020e-01 -2.54355758e-01
-4.88434762e-01 9.31213498e-01 7.79806376e-02 -6.06254935e-01
1.40569723e+00 3.58570278e-01 -4.29565489e-01 -3.12085629e-01
-1.10360038e+00 -4.43560004e-01 2.75702566e-01 7.72994876e-01
2.01766431e-01 9.56438035e-02 -2.06380859e-01 7.59209752e-01
4.51510251e-01 1.94403872e-01 2.68281460e-01 1.02744496e+00
-2.89589614e-01 -6.50174677e-01 -2.82554328e-01 4.36490446e-01
-8.53039086e-01 -3.40656310e-01 1.77826837e-01 8.46122682e-01
-1.17812589e-01 6.84016109e-01 5.15792251e-01 -1.02896735e-01
4.64589715e-01 3.97098541e-01 4.06391323e-01 -5.41978300e-01
-3.06458294e-01 3.36214155e-02 -4.29350287e-01 -5.69230199e-01
-5.97499907e-01 -6.70563757e-01 -8.20284188e-01 -2.83350855e-01
-1.08232811e-01 -7.82032833e-02 3.39963168e-01 7.43719935e-01
4.65456784e-01 9.27793622e-01 5.75695455e-01 -9.39519525e-01
-6.16887867e-01 -1.12218463e+00 -6.79810524e-01 6.12794697e-01
3.74025762e-01 -1.09055269e+00 -5.73806763e-01 1.16595581e-01] | [7.005598068237305, 3.4268219470977783] |
9af88905-9eb5-4a9e-9de4-5483dda7e27a | continual-learning-for-on-device | 2207.07429 | null | https://arxiv.org/abs/2207.07429v2 | https://arxiv.org/pdf/2207.07429v2.pdf | Continual Learning For On-Device Environmental Sound Classification | Continuously learning new classes without catastrophic forgetting is a challenging problem for on-device environmental sound classification given the restrictions on computation resources (e.g., model size, running memory). To address this issue, we propose a simple and efficient continual learning method. Our method selects the historical data for the training by measuring the per-sample classification uncertainty. Specifically, we measure the uncertainty by observing how the classification probability of data fluctuates against the parallel perturbations added to the classifier embedding. In this way, the computation cost can be significantly reduced compared with adding perturbation to the raw data. Experimental results on the DCASE 2019 Task 1 and ESC-50 dataset show that our proposed method outperforms baseline continual learning methods on classification accuracy and computational efficiency, indicating our method can efficiently and incrementally learn new classes without the catastrophic forgetting problem for on-device environmental sound classification. | ['Arshdeep Singh', 'Wenwu Wang', 'Mark D. Plumbley', 'Eng Siong Chng', 'James King', 'Xubo Liu', 'Yang Xiao'] | 2022-07-15 | null | null | null | null | ['environmental-sound-classification', 'sound-classification'] | ['audio', 'audio'] | [ 3.06973636e-01 -2.74631888e-01 -1.18652560e-01 -2.28946090e-01
-9.58687007e-01 -5.02294540e-01 1.19105160e-01 1.44548357e-01
-5.73219538e-01 7.48001277e-01 -2.36993730e-01 -3.43229115e-01
-1.44668341e-01 -6.40290022e-01 -8.43386889e-01 -9.34268594e-01
-1.55300815e-02 1.72147766e-01 4.91662949e-01 4.42597300e-01
1.08033270e-01 1.36071384e-01 -1.79577875e+00 2.50337720e-02
8.15752208e-01 1.21121883e+00 3.04336771e-02 6.22026861e-01
1.96724623e-01 3.92035246e-01 -6.03197157e-01 -2.14156285e-01
1.60884131e-02 -2.39384733e-02 -4.11431462e-01 -3.56435210e-01
3.90903443e-01 -3.59837264e-01 -2.32387930e-01 1.13021946e+00
7.83172607e-01 2.63271213e-01 4.95158941e-01 -1.31225848e+00
1.37586556e-02 8.17282617e-01 -2.87495941e-01 3.30776572e-01
-1.25773191e-01 3.46404552e-01 8.35709631e-01 -9.75681424e-01
9.51576456e-02 7.65802681e-01 9.40441668e-01 7.05125690e-01
-1.24204612e+00 -1.05432022e+00 3.42497021e-01 1.08456619e-01
-1.48990965e+00 -6.52929008e-01 9.81696665e-01 -4.00818825e-01
8.87746394e-01 3.65293533e-01 5.10905504e-01 1.11454964e+00
3.42864335e-01 5.78698635e-01 9.60233033e-01 -4.49079573e-01
1.06306112e+00 3.89770001e-01 4.57599044e-01 5.55864513e-01
4.46529955e-01 3.06765020e-01 -7.86826670e-01 -5.19421697e-01
-1.46282353e-02 -1.45921335e-01 -3.61098349e-02 -1.62109941e-01
-5.11972666e-01 3.58688831e-01 -1.00929156e-01 1.14409342e-01
-1.03338715e-03 4.33775514e-01 5.61919987e-01 3.44504207e-01
7.13998735e-01 3.34217161e-01 -9.17358160e-01 -6.30954981e-01
-9.44620788e-01 2.05880962e-02 4.08673733e-01 7.87395418e-01
4.28804368e-01 3.51265371e-01 -7.93830231e-02 7.24897742e-01
2.86943316e-02 7.59055853e-01 6.77202404e-01 -5.22871017e-01
4.74722445e-01 1.46327510e-01 3.25065292e-02 -6.79061174e-01
-3.29356968e-01 -6.53627872e-01 -7.92418718e-01 1.17300808e-01
9.81454849e-02 -3.85175377e-01 -8.29237103e-01 1.80795598e+00
3.86326045e-01 5.78149617e-01 -6.11933693e-02 1.08975001e-01
2.82830805e-01 4.70772088e-01 9.33283567e-02 -4.72472101e-01
9.44424510e-01 -6.76121294e-01 -7.55049825e-01 -1.33248463e-01
4.92847025e-01 -3.10833335e-01 1.47451782e+00 5.87607801e-01
-5.70590317e-01 -7.19773173e-01 -1.40744126e+00 4.25175101e-01
-3.11263859e-01 1.47813752e-01 5.61378896e-01 1.24966657e+00
-4.48242873e-01 9.01272893e-01 -1.24748814e+00 3.02312315e-01
5.44529617e-01 5.34966946e-01 1.92499146e-01 1.72367126e-01
-1.17143989e+00 1.20587058e-01 4.99303669e-01 -7.93353692e-02
-8.59509945e-01 -1.08632433e+00 -2.99138188e-01 6.53559044e-02
2.86187887e-01 -2.36956611e-01 1.35826802e+00 -3.97656471e-01
-1.83279264e+00 6.82354644e-02 3.29266861e-02 -5.07121384e-01
5.28786004e-01 -4.24891353e-01 -5.95365882e-01 -3.57067913e-01
-4.02905494e-01 1.83370098e-01 1.28424394e+00 -9.99669433e-01
-8.28980505e-01 -3.74517620e-01 -3.73232543e-01 -2.23151863e-01
-1.05757570e+00 -6.45981610e-01 -1.89569578e-01 -6.41093493e-01
7.21061379e-02 -1.24301863e+00 -3.41990963e-03 -1.94686502e-01
-2.38104224e-01 -2.11713403e-01 9.54713464e-01 -4.24980253e-01
1.72279084e+00 -2.65436101e+00 -2.72671938e-01 7.72471428e-02
-3.65415215e-02 2.60843068e-01 1.53002828e-01 -1.43528163e-01
3.03836782e-02 1.15509041e-01 -3.36909056e-01 -4.75232244e-01
-4.00900356e-02 4.11588214e-02 -7.32396603e-01 2.13462919e-01
-4.90475334e-02 4.69950914e-01 -1.04745018e+00 -1.41923547e-01
-7.93155283e-02 3.11412305e-01 -7.47094572e-01 1.09183334e-01
-2.94879705e-01 1.00789405e-01 -2.62221038e-01 5.86020172e-01
6.87306404e-01 -6.33076206e-02 1.36470944e-01 -1.54992774e-01
2.39356890e-01 2.89990544e-01 -1.22508860e+00 1.48855448e+00
-9.14043963e-01 3.51346940e-01 -5.85685313e-01 -5.45166969e-01
6.52663052e-01 2.04720855e-01 3.04220855e-01 -5.07727325e-01
2.82918066e-02 3.31864357e-01 -2.06706956e-01 -2.47537509e-01
4.01518583e-01 -1.47318825e-01 -5.27913570e-01 4.37622756e-01
5.15025482e-02 -3.39883476e-01 -4.19853479e-01 -2.52543867e-01
1.09287024e+00 -1.55151069e-01 2.97129583e-02 -1.23732574e-01
7.40821287e-02 -3.82765532e-01 8.42658043e-01 8.10129404e-01
-2.93708950e-01 1.22891203e-01 1.56893790e-01 -3.24809164e-01
-7.26463735e-01 -1.08967042e+00 -3.79907072e-01 9.84449089e-01
-8.20282996e-02 -4.79729772e-01 -7.45781898e-01 -9.06631410e-01
2.40686625e-01 1.09228408e+00 -3.99302542e-01 -7.55304754e-01
-3.69884491e-01 -1.09428144e+00 4.94702458e-01 6.03009105e-01
3.64117652e-01 -5.70257008e-01 -7.85028338e-01 3.61771226e-01
2.29429737e-01 -8.11098576e-01 -3.89475286e-01 5.29568851e-01
-1.17948294e+00 -6.41854644e-01 -7.65063986e-02 -4.37728286e-01
4.32750046e-01 -7.71695375e-02 6.69306636e-01 -2.52241880e-01
-2.55866796e-01 5.35393536e-01 -4.61013615e-02 -7.53906667e-01
-2.70367652e-01 3.71827126e-01 7.14470088e-01 -3.81136052e-02
1.00282259e-01 -8.30637038e-01 -4.06902730e-01 -5.90041131e-02
-6.99003816e-01 -2.76169866e-01 1.54515579e-01 1.14208746e+00
8.13552082e-01 8.59644115e-01 8.80516648e-01 -9.05245125e-01
7.37140179e-01 -3.25918138e-01 -9.04015362e-01 3.54952425e-01
-1.21434736e+00 2.83026010e-01 8.02637517e-01 -8.50333750e-01
-9.13328350e-01 4.11247551e-01 -7.86064565e-02 -3.59579027e-01
4.34286922e-01 1.12679340e-01 -2.28966922e-01 4.64004017e-02
5.98130882e-01 2.38493115e-01 -5.48338532e-01 -6.11130595e-01
2.61635005e-01 9.87270713e-01 5.06618142e-01 -6.64210916e-01
8.69198918e-01 2.24852189e-01 -2.81394459e-03 -6.14862084e-01
-7.81243682e-01 -1.05713308e-02 -5.80581069e-01 -1.76316679e-01
2.58990794e-01 -9.57636774e-01 -6.65740192e-01 6.78695440e-01
-7.60635376e-01 -3.02096367e-01 -7.02988803e-01 4.52858180e-01
-3.78706306e-01 1.15321308e-01 -3.03120136e-01 -1.19027972e+00
-6.50382459e-01 -8.82099211e-01 9.26427066e-01 2.82707997e-02
-2.44714290e-01 -7.49508023e-01 2.31362283e-01 -9.14706513e-02
3.78411204e-01 -6.25642482e-04 1.23203456e+00 -3.49767804e-01
-3.76976311e-01 -4.68683213e-01 4.39030379e-01 5.76443017e-01
3.48276347e-01 -1.25846297e-01 -1.50970995e+00 -6.33226395e-01
1.67748019e-01 -3.57255757e-01 1.01378536e+00 3.65717262e-02
1.70578384e+00 -4.09324080e-01 -4.11357075e-01 4.11269099e-01
1.19032526e+00 4.93458956e-01 2.12355405e-01 -1.69029817e-01
4.61507142e-01 2.65226252e-02 7.73843646e-01 6.84967756e-01
-1.69531718e-01 6.26788557e-01 7.00119659e-02 4.81249511e-01
-2.01953694e-01 -7.21937537e-01 4.50125366e-01 1.24647844e+00
3.68115574e-01 -2.06047505e-01 -8.44924331e-01 4.13504094e-01
-1.54376733e+00 -6.93709016e-01 5.65575123e-01 2.69872737e+00
1.14691877e+00 4.89206284e-01 -2.10917503e-01 8.25522006e-01
3.58158708e-01 -2.55178660e-02 -1.16135418e+00 4.79535460e-02
2.23323956e-01 2.97012508e-01 3.84024888e-01 5.36808312e-01
-1.13481116e+00 8.33107412e-01 6.44406986e+00 1.19950521e+00
-1.41599286e+00 4.12333250e-01 8.28586876e-01 -5.36838889e-01
-1.14964753e-01 -3.05120289e-01 -1.18550181e+00 8.35092187e-01
1.32234788e+00 -2.36484230e-01 4.31399584e-01 1.10941815e+00
-2.42665753e-01 6.39327541e-02 -1.17814469e+00 1.20741236e+00
-1.03312075e-01 -1.12684035e+00 -1.02980077e-01 -3.37296307e-01
8.20759833e-01 -9.87598747e-02 6.23389959e-01 5.79363644e-01
5.46291210e-02 -4.56713736e-01 7.88007319e-01 4.73327726e-01
1.28227282e+00 -8.61571074e-01 5.60384035e-01 3.21799725e-01
-1.11043024e+00 -5.55189908e-01 -2.71982163e-01 1.90095007e-01
-1.34984344e-01 1.14072764e+00 -9.43504930e-01 3.36158276e-02
9.64511991e-01 2.35987261e-01 -7.62064040e-01 8.61802876e-01
-2.17332458e-03 1.26746666e+00 -5.84717751e-01 -2.30457664e-01
-4.56559598e-01 3.92224699e-01 4.86998826e-01 8.79462242e-01
6.70791447e-01 -9.93450657e-02 -1.29589662e-01 4.11561787e-01
-2.11057425e-01 -1.89049095e-01 -3.49385649e-01 -8.87409225e-02
9.72480953e-01 5.77013493e-01 -5.04023492e-01 -2.86783010e-01
2.56556869e-01 9.31623757e-01 1.51213184e-01 1.54272020e-01
-9.57089543e-01 -5.18194377e-01 4.83529508e-01 3.97199690e-02
2.24156111e-01 -6.39895052e-02 -4.71295267e-01 -9.80961323e-01
3.00506115e-01 -6.28640711e-01 3.32650393e-01 -2.69128978e-01
-1.16069937e+00 6.71241522e-01 -1.51803434e-01 -1.37784410e+00
-1.63567260e-01 -3.00164133e-01 -3.04868847e-01 1.93619892e-01
-1.31304383e+00 -6.62712276e-01 -1.35721207e-01 3.40994298e-01
5.22225082e-01 -3.32243711e-01 9.39844370e-01 3.83296400e-01
-7.62579322e-01 1.38150775e+00 5.87151289e-01 -4.30954695e-01
6.38717353e-01 -9.08537090e-01 4.07675922e-01 5.15039444e-01
2.24133432e-01 4.83435184e-01 6.32341325e-01 -6.97982192e-01
-1.42119431e+00 -1.32275081e+00 5.86134553e-01 -5.26070654e-01
6.61737978e-01 -8.06738675e-01 -9.05599535e-01 1.96756989e-01
-4.72066045e-01 2.94149578e-01 8.57111216e-01 2.33349964e-01
-6.35090411e-01 -7.86324322e-01 -1.26416397e+00 4.66410935e-01
1.14032698e+00 -7.55194783e-01 -2.49438837e-01 4.56711233e-01
1.13878071e+00 -3.69188160e-01 -6.91324711e-01 6.78238869e-01
7.05224037e-01 -4.12360191e-01 7.85619676e-01 -5.01403809e-01
-1.20652169e-01 -2.37399518e-01 -3.73097360e-01 -1.26427948e+00
-8.86096507e-02 -6.84734285e-01 -5.91487944e-01 1.35819638e+00
6.07401490e-01 -6.72490537e-01 7.42096782e-01 7.22334087e-01
1.77261233e-02 -8.36866677e-01 -1.54822397e+00 -1.22601628e+00
1.91230431e-01 -8.89260709e-01 8.01260591e-01 5.50961852e-01
-5.33874556e-02 2.67582655e-01 -5.84413350e-01 3.83477539e-01
7.25311637e-01 -7.60432258e-02 4.95287597e-01 -1.28561902e+00
-6.81151330e-01 -2.74551269e-02 -3.08541209e-01 -8.22786331e-01
9.32004452e-02 -6.81329966e-01 1.79977059e-01 -4.91647273e-01
6.73861727e-02 -9.12797928e-01 -8.11775923e-01 6.04381800e-01
-2.87482172e-01 2.02423967e-02 -7.41587952e-02 1.37301013e-01
-6.53496325e-01 8.27993333e-01 5.51840723e-01 -4.12436843e-01
-5.45815945e-01 3.44202876e-01 -3.69237334e-01 5.59192181e-01
8.49422574e-01 -1.01224697e+00 -6.97758079e-01 -2.04743519e-01
3.22021961e-01 -1.94522604e-01 7.42924213e-03 -1.62786448e+00
3.31911027e-01 8.38460922e-02 8.35793763e-02 -5.05727947e-01
2.62603015e-01 -9.47531223e-01 1.53836757e-01 7.70089924e-01
-3.05261165e-01 -3.26542586e-01 6.43862844e-01 1.19682813e+00
1.14228643e-01 -2.13283107e-01 8.14217389e-01 5.17367601e-01
-3.55352879e-01 2.55356997e-01 -2.53263295e-01 -7.79818073e-02
8.75572443e-01 2.83849537e-01 -1.06321685e-01 1.24184787e-01
-7.40290761e-01 -7.02789724e-02 1.71071198e-02 5.02220094e-01
5.33437908e-01 -1.41429257e+00 -1.48119375e-01 4.03133601e-01
2.10896909e-01 -1.04857892e-01 4.24796283e-01 3.29823613e-01
1.34083524e-01 2.14542940e-01 3.84199083e-01 -6.45834148e-01
-1.37697256e+00 4.56104755e-01 2.90944308e-01 -2.55612552e-01
-4.04584527e-01 1.01397872e+00 -4.29292500e-01 -4.25017536e-01
6.84748292e-01 -6.07103705e-01 2.60884523e-01 1.04758829e-01
5.81975520e-01 5.25709331e-01 4.07457709e-01 3.44515204e-01
-3.42311889e-01 5.23153603e-01 -1.98510289e-01 -1.19087972e-01
1.38965559e+00 -1.47748515e-02 3.17322880e-01 1.17423475e+00
1.17199314e+00 1.58859357e-01 -1.47526038e+00 -3.84095579e-01
6.28911555e-02 -4.31753039e-01 2.64290422e-01 -8.63129079e-01
-1.01805091e+00 1.06279111e+00 1.49348414e+00 -1.47596419e-01
1.15074432e+00 -4.33863789e-01 9.36354816e-01 7.79178262e-01
7.26886392e-01 -1.50656748e+00 2.62469351e-01 3.69885683e-01
6.77439392e-01 -1.13845205e+00 9.16742012e-02 -3.90211679e-02
-2.68772125e-01 9.21842277e-01 5.88504016e-01 3.70807856e-01
1.33598447e+00 6.33367896e-01 -2.33516693e-01 1.91131115e-01
-9.33891356e-01 5.98026395e-01 -8.94599129e-03 4.07293677e-01
-8.58607814e-02 2.85818338e-01 1.00265972e-01 1.06785131e+00
-3.83423239e-01 4.29654866e-02 1.68869019e-01 8.70333731e-01
-3.33131909e-01 -9.30081069e-01 -3.78657994e-03 6.20817304e-01
-1.04992114e-01 -1.49505347e-01 -5.09712137e-02 2.01118946e-01
3.81633639e-01 9.10810292e-01 8.28879401e-02 -1.05356705e+00
3.56296480e-01 4.95362133e-01 2.90906101e-01 -5.44695795e-01
-2.56032944e-01 -2.05426350e-01 -1.14057414e-01 -8.26941952e-02
1.36879057e-01 -8.21684361e-01 -1.16661835e+00 7.89681152e-02
-7.46787667e-01 1.73269257e-01 6.91246212e-01 6.66915059e-01
5.92458308e-01 7.34933913e-01 1.10999918e+00 -3.91145855e-01
-1.19195080e+00 -8.43347073e-01 -5.73773265e-01 4.85217609e-02
4.02136087e-01 -6.89636946e-01 -9.49915230e-01 -9.17187035e-02] | [9.971440315246582, 3.6311111450195312] |
4a426ee4-c0b7-4380-bdef-c23474c79b50 | bootstrapping-text-anonymization-models-with | null | null | https://openreview.net/forum?id=-MoY6seu_x | https://openreview.net/pdf?id=-MoY6seu_x | Bootstrapping Text Anonymization Models with Distant Supervision | We propose a novel method to bootstrap text anonymization models based on distant supervision. Instead of requiring manually labeled training data, the approach relies on a knowledge graph expressing the background information assumed to be publicly available about various individuals. This knowledge graph is employed to automatically annotate text documents including personal data about a subset of those individuals. More precisely, the method determines which text spans ought to be masked in order to guarantee $k$-anonymity, assuming an adversary with access to both the text documents and the background information expressed in the knowledge graph. The resulting collection of labeled documents is then used as training data to fine-tune a pre-trained language model for text anonymization. We illustrate this approach using a knowledge graph extracted from Wikidata and short biographical texts from Wikipedia. Evaluation results with a BERT-based model and a manually annotated collection of 553 summaries showcase the potential of the approach, but also unveil a number of issues that may arise the knowledge graph is noisy or incomplete. The results also illustrate that, contrary to most sequence labeling problems, the text anonymization task may admit several alternative solutions. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['text-anonymization'] | ['natural-language-processing'] | [ 3.07335228e-01 6.87133551e-01 -2.83969402e-01 -5.36046147e-01
-7.53578305e-01 -9.79503036e-01 6.51750326e-01 6.24403238e-01
-6.04925215e-01 1.13171434e+00 4.77202713e-01 -8.41651931e-02
-7.84346238e-02 -8.27146888e-01 -6.37579918e-01 -5.32180130e-01
1.82181582e-01 8.21028471e-01 -1.44131109e-01 -2.77057127e-03
3.77007686e-02 5.70741236e-01 -1.10781503e+00 3.16063970e-01
8.41628969e-01 4.79366034e-01 -4.06219751e-01 5.05332291e-01
-2.64494359e-01 7.00488508e-01 -8.51087451e-01 -1.26156723e+00
3.54016781e-01 -2.80411690e-01 -1.24908400e+00 2.71134704e-01
4.40334678e-01 -9.77695882e-02 -3.10414195e-01 1.18107796e+00
4.25442070e-01 -1.56619355e-01 7.02291787e-01 -1.36132026e+00
-6.53690517e-01 9.36208785e-01 -2.09648371e-01 2.27618422e-02
4.60145026e-01 -9.04721022e-02 1.00042486e+00 -1.39341325e-01
9.72950876e-01 9.52394366e-01 8.57060254e-01 5.31147182e-01
-1.51183057e+00 -3.54559243e-01 -1.05948478e-01 -1.68677688e-01
-1.50779843e+00 -4.98425692e-01 6.04154408e-01 -4.48083431e-01
4.84289169e-01 5.69123745e-01 3.32538813e-01 1.08600199e+00
-2.78540343e-01 5.90096772e-01 9.34340835e-01 -6.14632010e-01
2.60018319e-01 9.07399833e-01 4.69000429e-01 7.72801697e-01
7.18281329e-01 -3.28243971e-01 -5.35897553e-01 -8.44779074e-01
-1.10927196e-02 -5.55200577e-01 -4.71761495e-01 -7.95319021e-01
-9.05557513e-01 6.57973468e-01 -4.18495648e-02 3.18459690e-01
-1.62392035e-02 -3.08546364e-01 7.17007697e-01 3.54561657e-01
5.36419392e-01 5.25594413e-01 -3.93104106e-01 3.12932640e-01
-8.39274287e-01 3.03501785e-01 1.44458592e+00 1.00217700e+00
1.18101203e+00 -3.21295947e-01 -9.77683663e-02 5.35383642e-01
6.68336749e-02 3.14988017e-01 4.88308579e-01 -5.77586949e-01
8.76934528e-01 8.51478815e-01 5.06053448e-01 -1.22994244e+00
-1.47157043e-01 -2.85594934e-03 -5.52569330e-01 -8.81286412e-02
9.27958012e-01 -1.51934355e-01 -5.11959612e-01 1.60234964e+00
4.06339198e-01 -2.82060146e-01 3.08654219e-01 5.26395082e-01
5.26157141e-01 1.78452939e-01 8.18565637e-02 -6.89852536e-02
1.40543878e+00 -4.22375649e-01 -8.36453557e-01 -2.35841237e-02
1.09083545e+00 -4.08611953e-01 8.85590851e-01 2.09703609e-01
-6.34051681e-01 -6.69389069e-02 -9.90032554e-01 -7.46893957e-02
-9.35263574e-01 1.18114531e-01 5.50198555e-01 1.43808568e+00
-9.86736655e-01 5.70976675e-01 -6.55232131e-01 -7.13537276e-01
6.37491405e-01 4.56254989e-01 -8.94616485e-01 -2.83501297e-02
-1.30089605e+00 6.38769507e-01 8.53323817e-01 -1.22628406e-01
-4.83807445e-01 -3.49197775e-01 -9.18496907e-01 9.70618427e-02
3.55252892e-01 -7.11381197e-01 8.04026484e-01 -1.13995659e+00
-1.12443185e+00 1.40473306e+00 -1.14559762e-01 -7.15784729e-01
8.12946677e-01 -3.76329534e-02 -4.97123063e-01 1.74099609e-01
1.49164498e-01 5.28072491e-02 7.58890986e-01 -1.37262285e+00
-4.25667793e-01 -7.09366798e-01 5.47689907e-02 1.07854977e-01
-6.57987773e-01 4.94639296e-03 -3.48016828e-01 -7.30210006e-01
-3.46771777e-01 -9.80469167e-01 -2.56784797e-01 -2.26124212e-01
-1.08559000e+00 1.49772942e-01 8.24251950e-01 -1.02793193e+00
1.37318015e+00 -1.71276498e+00 8.57168343e-03 6.86940134e-01
1.60101339e-01 3.94646257e-01 1.54470876e-01 6.73214853e-01
-1.35591000e-01 6.27618074e-01 -5.10586262e-01 -4.66409057e-01
2.29589209e-01 2.02243626e-01 -5.04134893e-01 7.28103638e-01
-1.75607651e-01 7.17756212e-01 -8.88926208e-01 -5.12159467e-01
-6.74614608e-02 7.50900209e-02 -3.24339479e-01 1.88584328e-02
-3.98491710e-01 3.45049620e-01 -6.06655598e-01 4.01158720e-01
5.38063824e-01 -2.04483345e-02 5.14004529e-01 9.75790098e-02
3.49881142e-01 9.44489986e-02 -1.33397210e+00 1.66860104e+00
2.87367515e-02 4.64588821e-01 1.15812838e-01 -1.04215086e+00
7.18480647e-01 4.45249796e-01 2.37247705e-01 -7.26568326e-02
1.01923496e-01 -5.19140810e-02 -6.20323718e-01 -6.24172926e-01
5.75558901e-01 2.59541124e-02 -3.72967422e-01 9.22871649e-01
8.77119303e-02 6.66972920e-02 1.67924285e-01 6.37433946e-01
1.04710686e+00 -4.60478142e-02 3.22699040e-01 -3.68521333e-01
8.91460240e-01 4.98136207e-02 4.79704261e-01 8.13387156e-01
-8.45894367e-02 2.59410381e-01 7.17713714e-01 -5.27883947e-01
-1.11141944e+00 -6.25799060e-01 4.11964878e-02 7.40118206e-01
-1.91585243e-01 -7.52738297e-01 -1.30649137e+00 -1.25149250e+00
1.61003888e-01 9.29855824e-01 -7.78427243e-01 -1.91975623e-01
-3.83363903e-01 -7.30917215e-01 1.06173956e+00 4.75761779e-02
2.77152777e-01 -7.89105058e-01 -1.60995379e-01 -9.69652906e-02
-4.26156431e-01 -1.04032862e+00 -5.55271029e-01 -1.57069728e-01
-5.45714259e-01 -1.28834176e+00 -1.13876656e-01 -5.51483274e-01
1.03408825e+00 -3.61544102e-01 9.61445928e-01 2.48944797e-02
-2.21953005e-01 6.30834222e-01 -1.79190993e-01 -3.79684508e-01
-8.94621730e-01 3.34396899e-01 2.60114968e-02 4.30724353e-01
7.24272549e-01 -4.77261394e-01 4.89489781e-03 1.47815764e-01
-1.20251679e+00 -4.20839548e-01 4.88688238e-02 6.74861014e-01
2.69390613e-01 2.53980190e-01 5.14925361e-01 -1.77928174e+00
6.43108785e-01 -4.05120730e-01 -6.97391331e-01 6.53098404e-01
-4.70362604e-01 2.93072581e-01 8.81885409e-01 -1.99837565e-01
-1.19319189e+00 3.47032189e-01 2.17793554e-01 -9.09550786e-02
-2.53103048e-01 4.13900346e-01 -8.24236393e-01 4.71710861e-02
8.52325678e-01 1.01475313e-01 3.67323644e-02 -5.17558932e-01
7.04832435e-01 9.42302644e-01 5.88384151e-01 -6.86188281e-01
1.03930473e+00 6.67442203e-01 -2.46357366e-01 -8.16389382e-01
-7.27347553e-01 -2.83248693e-01 -1.03187442e+00 2.41314635e-01
8.20731580e-01 -7.62373030e-01 -6.01238787e-01 3.09900731e-01
-9.82278407e-01 -1.09483087e-02 -5.41191995e-01 1.59178108e-01
-5.59962451e-01 7.13710368e-01 -3.43318075e-01 -7.75491595e-01
-2.60123104e-01 -6.83391929e-01 7.25228131e-01 -1.77970082e-01
-5.92880666e-01 -1.29035306e+00 4.74961102e-02 7.67620504e-01
-1.05475910e-01 4.54719424e-01 1.17400372e+00 -1.46520996e+00
-4.03682232e-01 -8.14443290e-01 6.46436512e-02 2.58666635e-01
3.66074711e-01 -2.20976740e-01 -1.18633878e+00 -4.05747890e-01
-7.78796151e-02 -3.63501191e-01 4.82432723e-01 -2.79319644e-01
1.07214403e+00 -8.47199023e-01 -5.91603339e-01 5.17055511e-01
1.22182882e+00 -3.15136284e-01 6.01496518e-01 2.40249112e-01
9.59546924e-01 1.03971446e+00 1.59308553e-01 3.93946618e-01
4.75144356e-01 6.06743872e-01 1.10970296e-01 1.07666984e-01
2.49539882e-01 -6.84879959e-01 -1.55926207e-02 3.24685335e-01
3.82494420e-01 -5.91558099e-01 -8.07917655e-01 6.73550069e-01
-1.70003438e+00 -9.72909629e-01 1.39485404e-01 2.52916956e+00
9.60902929e-01 -1.86317161e-01 1.74982041e-01 6.60236031e-02
8.98491442e-01 7.73467198e-02 -3.74644667e-01 -4.42306966e-01
-2.70551175e-01 -1.70459554e-01 1.00907743e+00 7.08020270e-01
-1.10098314e+00 9.05488014e-01 5.73689556e+00 5.75405598e-01
-6.35880113e-01 -1.60478219e-01 6.06651127e-01 9.33143348e-02
-5.96148849e-01 3.63495886e-01 -8.35197031e-01 5.72647870e-01
1.13807058e+00 -7.54932880e-01 4.49981719e-01 6.50687516e-01
-4.35495228e-02 6.89514056e-02 -1.45955360e+00 6.69974566e-01
2.45603934e-01 -1.25990498e+00 2.82881051e-01 2.43260473e-01
5.75813413e-01 -4.02363896e-01 -2.11563542e-01 -2.04244498e-02
9.33429658e-01 -9.60742772e-01 5.05979598e-01 5.58191895e-01
7.59757042e-01 -8.68177295e-01 6.67211056e-01 5.69877863e-01
-6.94753945e-01 1.03125443e-04 -2.92562991e-01 2.91427076e-01
-8.84687081e-02 6.22589588e-01 -1.20868778e+00 9.20722783e-01
5.34189939e-01 4.66934413e-01 -7.12014854e-01 6.08938932e-01
-3.80226403e-01 5.17969608e-01 -3.51902902e-01 1.81689486e-01
-1.46921538e-02 -2.77718037e-01 6.69564903e-01 1.26045656e+00
-2.26811282e-02 1.36014745e-01 1.73424467e-01 7.51218379e-01
-5.78624010e-01 4.21151489e-01 -1.08704388e+00 -4.88828659e-01
5.89015782e-01 1.06623650e+00 -6.03585780e-01 -4.17342931e-01
-4.02067095e-01 1.26889837e+00 6.06566250e-01 4.98608261e-01
-2.11797267e-01 -5.15854776e-01 4.44864988e-01 2.62595952e-01
1.57434508e-01 2.57560819e-01 9.40875802e-03 -1.46481836e+00
1.62593782e-01 -9.02648270e-01 8.33733320e-01 -5.77905774e-01
-1.25162816e+00 4.24868822e-01 -1.39451802e-01 -8.57917905e-01
-3.76820952e-01 -2.73221344e-01 -3.48263443e-01 1.07736516e+00
-1.02422178e+00 -1.17810214e+00 -1.19950576e-02 5.98973751e-01
-1.21031396e-01 -2.37646282e-01 1.15506411e+00 9.05904099e-02
-6.57159686e-01 8.93886924e-01 2.86355674e-01 5.31281114e-01
7.78404057e-01 -1.39065397e+00 5.83874881e-01 9.62244809e-01
3.09305012e-01 7.95019984e-01 8.19953263e-01 -1.00220764e+00
-1.17228162e+00 -1.30539155e+00 1.27896249e+00 -1.06278634e+00
6.46148324e-01 -6.87993646e-01 -1.18864918e+00 1.29587460e+00
3.72288097e-03 7.04170093e-02 1.11942232e+00 -7.85311311e-02
-5.67173779e-01 6.41679987e-02 -1.61333811e+00 5.03063202e-01
9.57784414e-01 -7.86236227e-01 -7.14499772e-01 5.76511204e-01
6.11186802e-01 -2.99895167e-01 -8.47121239e-01 -3.51540267e-01
2.66645670e-01 -7.53468037e-01 7.77727902e-01 -1.03365374e+00
-1.77098945e-01 -2.89971203e-01 -5.70842996e-02 -1.23197162e+00
1.97496578e-01 -9.11125064e-01 1.72539964e-01 1.72120214e+00
3.91552776e-01 -9.10909712e-01 1.27264059e+00 1.25867534e+00
4.65160698e-01 1.13960668e-01 -9.66340184e-01 -6.93752944e-01
-1.21903541e-02 -8.09435099e-02 8.31044376e-01 1.54152477e+00
2.75067925e-01 3.50040704e-01 -4.60808098e-01 4.51898813e-01
7.28186488e-01 -1.59951389e-01 9.79174137e-01 -1.41015601e+00
-5.82010262e-02 2.17679441e-01 -4.47088420e-01 -4.13388640e-01
7.53324687e-01 -1.29750085e+00 -4.46006477e-01 -1.20630336e+00
1.42062828e-01 -5.24980426e-01 1.67627707e-01 6.20944321e-01
-6.71600550e-02 -6.73980936e-02 -1.64440289e-01 6.65047616e-02
-5.13369918e-01 3.75918239e-01 3.91147435e-01 -1.04768686e-01
-2.25259453e-01 2.43957520e-01 -1.04196393e+00 7.76926100e-01
7.30707884e-01 -7.76235342e-01 -5.52900195e-01 -2.22968459e-01
1.54421583e-01 -1.07853957e-01 3.99747133e-01 -6.01180673e-01
4.68149811e-01 1.94516957e-01 3.34238648e-01 -3.73154432e-01
4.62541878e-02 -1.19109797e+00 3.63785565e-01 2.90750533e-01
-5.97920477e-01 -1.77180022e-01 1.74061313e-01 9.97117102e-01
-1.37437126e-02 -4.50596690e-01 4.69388783e-01 -2.68169075e-01
-3.26342076e-01 2.18332782e-01 -2.39671320e-01 2.87189156e-01
1.03684056e+00 -3.30012411e-01 -3.91969711e-01 -4.53369021e-01
-9.26341474e-01 2.79580742e-01 1.01206601e+00 1.52536392e-01
4.31746393e-02 -9.75385308e-01 -6.22266114e-01 3.14731270e-01
3.68670434e-01 -2.40107074e-01 3.86004373e-02 9.92228687e-02
-3.23875368e-01 3.81839812e-01 -2.79887971e-02 -8.97087902e-02
-1.37975276e+00 9.25993443e-01 1.67955205e-01 -3.26604545e-01
-6.11224592e-01 5.84306538e-01 -3.80697288e-02 -7.64621019e-01
2.49660596e-01 8.67738053e-02 -1.20159805e-01 1.78706184e-01
5.91750503e-01 3.61412883e-01 2.95154363e-01 -8.77633452e-01
-2.72879839e-01 6.37337416e-02 -3.75113636e-01 -1.11303069e-01
1.11999929e+00 -3.15248966e-01 -4.35720801e-01 1.64468795e-01
1.18015277e+00 5.70817113e-01 -7.84828424e-01 -5.81530750e-01
5.02299130e-01 -5.99060416e-01 -4.82103556e-01 -7.29382515e-01
-8.60972881e-01 4.34992641e-01 1.20800517e-01 3.15679342e-01
8.23572278e-01 -1.67820171e-01 3.75375509e-01 8.06055963e-01
3.04710269e-01 -1.02387834e+00 -5.30814350e-01 -3.59155759e-02
5.35507977e-01 -1.06371081e+00 3.19679588e-01 -6.40543997e-01
-7.07990885e-01 8.42182875e-01 2.64244050e-01 4.29267764e-01
4.01408702e-01 6.72673061e-02 2.05616593e-01 -1.03160001e-01
-3.12532037e-01 2.80613065e-01 -8.90684128e-02 9.43635643e-01
-1.12698087e-02 -1.45744383e-01 -1.07409820e-01 6.38132334e-01
-4.57919270e-01 -7.17956871e-02 8.23978782e-01 8.42155576e-01
-3.90528962e-02 -1.33381402e+00 -3.54330897e-01 5.54308057e-01
-7.74228156e-01 -4.26487252e-02 -8.71475399e-01 7.61962354e-01
-1.25197083e-01 7.54952133e-01 -2.80539900e-01 -6.77112117e-02
3.72241497e-01 3.88925940e-01 3.28343175e-02 -7.47185290e-01
-8.06718230e-01 -7.19541252e-01 7.54948020e-01 -2.89758593e-01
-2.69431621e-01 -7.57907391e-01 -8.02672386e-01 -4.18532282e-01
-2.49665529e-01 6.70199215e-01 4.22426999e-01 8.22990417e-01
2.79205263e-01 -1.53367445e-01 3.80358249e-01 -3.69147718e-01
-4.74061131e-01 -4.12940174e-01 -7.99679518e-01 8.57108951e-01
2.31777415e-01 -1.80633029e-03 -4.69323337e-01 6.73224509e-01] | [6.159811019897461, 7.039047718048096] |
356991e2-f333-4e7b-afb8-84154c74e96a | lightweight-self-knowledge-distillation-with | 2305.09183 | null | https://arxiv.org/abs/2305.09183v1 | https://arxiv.org/pdf/2305.09183v1.pdf | Lightweight Self-Knowledge Distillation with Multi-source Information Fusion | Knowledge Distillation (KD) is a powerful technique for transferring knowledge between neural network models, where a pre-trained teacher model is used to facilitate the training of the target student model. However, the availability of a suitable teacher model is not always guaranteed. To address this challenge, Self-Knowledge Distillation (SKD) attempts to construct a teacher model from itself. Existing SKD methods add Auxiliary Classifiers (AC) to intermediate layers of the model or use the history models and models with different input data within the same class. However, these methods are computationally expensive and only capture time-wise and class-wise features of data. In this paper, we propose a lightweight SKD framework that utilizes multi-source information to construct a more informative teacher. Specifically, we introduce a Distillation with Reverse Guidance (DRG) method that considers different levels of information extracted by the model, including edge, shape, and detail of the input data, to construct a more informative teacher. Additionally, we design a Distillation with Shape-wise Regularization (DSR) method that ensures a consistent shape of ranked model output for all data. We validate the performance of the proposed DRG, DSR, and their combination through comprehensive experiments on various datasets and models. Our results demonstrate the superiority of the proposed methods over baselines (up to 2.87%) and state-of-the-art SKD methods (up to 1.15%), while being computationally efficient and robust. The code is available at https://github.com/xucong-parsifal/LightSKD. | ['Lei Guo', 'Pengchao Han', 'Xucong Wang'] | 2023-05-16 | null | null | null | null | ['self-knowledge-distillation'] | ['computer-vision'] | [ 4.07809988e-02 1.99076794e-02 -2.18083173e-01 -3.45028222e-01
-7.12509334e-01 -6.85957849e-01 5.84148884e-01 1.26558572e-01
-4.30832416e-01 5.63566506e-01 -4.08373214e-02 -3.28280061e-01
-6.88089728e-02 -7.69403100e-01 -8.04984510e-01 -8.13099742e-01
4.23575848e-01 3.42377961e-01 4.41121578e-01 -3.02352989e-03
6.03598617e-02 5.87728739e-01 -1.59119844e+00 9.63053703e-02
1.38543379e+00 1.06915069e+00 4.52669472e-01 3.13005775e-01
-1.30491659e-01 5.64790070e-01 -3.64651740e-01 -2.65082985e-01
2.07311496e-01 -2.52856404e-01 -8.17228854e-01 -1.57601878e-01
5.78892887e-01 -3.43277127e-01 -2.86449909e-01 7.61984646e-01
6.49019480e-01 4.47887778e-01 6.91872835e-01 -9.70885217e-01
-7.07669914e-01 5.41599393e-01 -5.24990797e-01 1.80234134e-01
-5.26774488e-03 1.06903508e-01 6.97339177e-01 -1.17075181e+00
3.54752928e-01 9.61167872e-01 4.45247412e-01 5.32648504e-01
-1.19305253e+00 -9.80640113e-01 3.86602938e-01 2.81342983e-01
-1.36748850e+00 -3.08695525e-01 9.73436475e-01 -3.67810279e-01
5.90666592e-01 1.47501975e-01 6.80270374e-01 9.47041571e-01
-4.70758796e-01 1.27283525e+00 1.21099484e+00 -4.54052806e-01
-1.93928983e-02 3.94562513e-01 3.81406933e-01 8.37919652e-01
1.17822357e-01 7.71484710e-03 -5.79590023e-01 4.53913435e-02
5.71114957e-01 7.03355297e-02 -4.08971310e-01 -4.28984016e-01
-6.92941606e-01 6.07821882e-01 5.06333113e-01 7.24721625e-02
-3.29306751e-01 -2.62496054e-01 -1.14459349e-02 1.23256102e-01
4.73047584e-01 2.65323371e-01 -6.83836877e-01 -6.33457080e-02
-9.47610080e-01 8.92334282e-02 5.67692280e-01 9.45414782e-01
8.97632718e-01 5.83426319e-02 -3.75468045e-01 9.43467498e-01
3.03106248e-01 3.14532667e-01 5.38078964e-01 -5.84470153e-01
3.96215439e-01 9.18345213e-01 -3.73205036e-01 -7.46957421e-01
-6.69691935e-02 -7.16481149e-01 -7.12408483e-01 -7.32848272e-02
1.46478131e-01 -2.72824988e-02 -1.14915681e+00 1.71938217e+00
8.03515851e-01 8.58053505e-01 2.59920806e-01 5.72263718e-01
1.30139029e+00 6.24803603e-01 8.46294686e-02 -5.58835417e-02
9.96361136e-01 -1.28726602e+00 -2.38101259e-01 -1.65744901e-01
7.65754163e-01 -6.30794466e-01 1.03777564e+00 5.30118108e-01
-1.05799794e+00 -7.21998394e-01 -8.15176964e-01 -1.57966569e-01
-3.76762569e-01 4.33716565e-01 3.04243743e-01 3.34086061e-01
-9.78487730e-01 5.91267109e-01 -7.31415749e-01 1.26454800e-01
5.84752679e-01 3.51089716e-01 -2.38986060e-01 -2.84889579e-01
-9.39213455e-01 7.58621335e-01 5.82249999e-01 -2.95648109e-02
-1.12584388e+00 -1.10798407e+00 -7.83034146e-01 2.40737833e-02
4.62215066e-01 -4.49628502e-01 1.29457879e+00 -7.71323025e-01
-1.64791083e+00 5.83411515e-01 4.27256338e-02 -1.67442530e-01
4.98525381e-01 -4.15811598e-01 -1.07549680e-02 1.53142020e-01
-2.61617720e-01 6.49085641e-01 7.34005332e-01 -1.35619748e+00
-7.01545000e-01 -3.17172587e-01 7.75220767e-02 5.47416806e-01
-5.62274396e-01 -3.27499658e-01 -8.99851620e-01 -6.49199307e-01
1.12218834e-01 -9.25929010e-01 -5.19425943e-02 -1.62597775e-01
-4.62596238e-01 -5.66076159e-01 9.90205884e-01 -7.12293088e-01
1.46043801e+00 -2.23045397e+00 1.87018275e-01 1.71207458e-01
2.52683133e-01 7.63050139e-01 -3.02810043e-01 7.08618760e-02
6.76426142e-02 -1.76892951e-02 -1.65250540e-01 -5.54496050e-01
-1.45983189e-01 2.93684423e-01 -2.05086887e-01 6.81180432e-02
2.55182743e-01 8.24158430e-01 -8.98495734e-01 -4.90539730e-01
2.05655724e-01 6.85986757e-01 -4.93822992e-01 4.49702740e-01
-1.19890571e-01 4.82795894e-01 -6.82629168e-01 5.81807792e-01
6.67453825e-01 -2.85041243e-01 7.71541148e-02 -3.48826766e-01
-1.39330745e-01 4.98572707e-01 -1.32295477e+00 1.55702007e+00
-5.08174062e-01 1.59826785e-01 -1.21833466e-01 -1.08387160e+00
1.03223741e+00 2.10527450e-01 2.58554935e-01 -6.63283944e-01
1.57416895e-01 1.86789185e-01 -1.56822115e-01 -2.75207102e-01
3.61104250e-01 2.54780620e-01 3.06908399e-01 4.20686245e-01
1.75522566e-01 -1.28854224e-02 9.29514691e-02 2.77564257e-01
8.47241163e-01 4.24200386e-01 1.32232636e-01 -5.90455122e-02
5.86464107e-01 -2.45410889e-01 6.54245794e-01 5.39267778e-01
-1.14870071e-01 4.02991414e-01 5.78133203e-02 -1.99290991e-01
-4.60230798e-01 -8.55241835e-01 -2.73912121e-02 1.13853514e+00
1.63132951e-01 -2.93093055e-01 -6.31127059e-01 -9.70276713e-01
1.62358023e-02 7.76563406e-01 -5.92311442e-01 -4.40077513e-01
-5.12057066e-01 -4.78308409e-01 4.01366830e-01 6.83964610e-01
7.08526075e-01 -8.15673292e-01 -3.40470344e-01 1.25144303e-01
1.56681780e-02 -9.73218977e-01 -4.06242728e-01 2.41020113e-01
-9.44065213e-01 -1.04018581e+00 -6.05690539e-01 -9.55131173e-01
1.04714775e+00 5.13825357e-01 9.38012183e-01 2.43428633e-01
-7.55429268e-03 2.64680862e-01 -3.07751298e-01 -2.50340492e-01
-1.83745116e-01 1.10618979e-01 4.60646637e-02 -6.17535710e-02
2.85759330e-01 -5.72681308e-01 -5.77006459e-01 2.98669547e-01
-8.19266558e-01 5.06811678e-01 8.29112887e-01 7.24242449e-01
9.46660459e-01 1.01733431e-01 4.79615241e-01 -9.28769648e-01
5.74460745e-01 -5.25520504e-01 -4.74161476e-01 4.05163288e-01
-9.17126536e-01 1.96465611e-01 8.37283731e-01 -7.12052763e-01
-1.20193410e+00 2.11244717e-01 5.68461651e-03 -8.27371895e-01
-2.40207180e-01 5.96746445e-01 -1.51741683e-01 -2.07477614e-01
5.16864538e-01 4.78948772e-01 -2.10186556e-01 -7.60465682e-01
3.45553875e-01 4.77934688e-01 5.06498218e-01 -9.33115900e-01
8.66585135e-01 8.59552175e-02 -2.51447946e-01 -4.97217238e-01
-1.13996792e+00 -4.21954840e-01 -6.74996912e-01 -1.41294990e-02
3.46666604e-01 -1.07932258e+00 -3.70526373e-01 7.37971485e-01
-7.44503975e-01 -5.90767682e-01 -1.55453548e-01 4.77301627e-01
3.19822952e-02 2.26908892e-01 -4.24088180e-01 -6.58021867e-01
-4.68437105e-01 -1.21689248e+00 6.02074623e-01 6.49518430e-01
3.44501257e-01 -1.01353526e+00 -5.68909720e-02 6.43620968e-01
2.63812542e-01 5.81822209e-02 7.75375664e-01 -9.66724753e-01
-5.66531420e-01 9.14395750e-02 -1.97170511e-01 5.47565341e-01
2.14336112e-01 6.04498908e-02 -1.05575454e+00 -2.60170251e-01
-3.53271961e-01 -6.73584640e-01 1.13040638e+00 3.10115725e-01
1.39965916e+00 -4.07769918e-01 -4.06552374e-01 6.10759258e-01
1.24871147e+00 1.77655056e-01 2.41543144e-01 2.61175752e-01
1.00969231e+00 4.35711890e-01 5.28552949e-01 2.87087083e-01
6.92151308e-01 4.73538786e-01 1.99460730e-01 -7.13880211e-02
-3.40925694e-01 -4.70485330e-01 3.12614292e-01 1.13572836e+00
-1.19605154e-01 -1.42949060e-01 -8.89669657e-01 6.45776868e-01
-1.65766180e+00 -5.03226876e-01 4.02650535e-02 2.18774843e+00
1.28071630e+00 1.16639487e-01 -9.82873514e-02 5.99722043e-02
4.71643925e-01 -1.26876533e-01 -7.09208608e-01 -1.57762736e-01
1.61074907e-01 2.61730671e-01 1.43634468e-01 3.49256098e-01
-9.86890912e-01 1.10267246e+00 5.05407000e+00 1.05822682e+00
-1.15117753e+00 -7.73722753e-02 5.65071106e-01 5.54363392e-02
-3.42604220e-01 -9.76565033e-02 -1.21591902e+00 4.09978598e-01
7.38959730e-01 -5.54500893e-02 3.67833287e-01 8.79041076e-01
1.46245912e-01 -1.31817997e-01 -9.49110985e-01 7.75634050e-01
1.80877224e-02 -1.14977729e+00 1.89976081e-01 -1.65958479e-01
1.03016257e+00 -1.20138936e-01 1.98190272e-01 6.85284674e-01
4.66951698e-01 -7.95648694e-01 4.21715468e-01 4.77155447e-01
5.51265001e-01 -8.51798356e-01 4.97807056e-01 5.37215531e-01
-1.32950747e+00 5.56221120e-02 -6.48918971e-02 1.66101128e-01
-3.73473853e-01 5.21187544e-01 -1.09798408e+00 7.48303294e-01
7.52629876e-01 8.09679270e-01 -7.33883440e-01 9.38249528e-01
-6.14389479e-01 1.06431544e+00 -3.64939243e-01 1.97467595e-01
3.23880255e-01 -1.46365508e-01 2.16931999e-01 1.00027442e+00
3.08495253e-01 3.07303995e-01 4.79254097e-01 6.70031786e-01
-2.82679826e-01 1.00123279e-01 -2.52876848e-01 1.61329545e-02
8.32623720e-01 1.30029488e+00 -4.59306210e-01 -3.30914974e-01
-5.55155158e-01 6.86152875e-01 7.34047532e-01 2.68802315e-01
-6.80343449e-01 -2.30918705e-01 4.68261629e-01 -7.54606649e-02
4.30191457e-01 -4.65332307e-02 -1.02084555e-01 -9.59019005e-01
5.34857996e-02 -8.28068078e-01 5.75031459e-01 -6.58557057e-01
-1.15107274e+00 4.71992314e-01 9.86450389e-02 -1.13579631e+00
-5.40640354e-02 -3.15427840e-01 -7.22632706e-01 9.24649119e-01
-2.02510548e+00 -1.24952972e+00 -5.37572265e-01 7.03739643e-01
4.47511315e-01 -1.52471662e-01 5.83168864e-01 2.36425400e-01
-8.96753371e-01 8.19752455e-01 1.82575047e-01 2.33725369e-01
6.62420630e-01 -1.29232717e+00 1.19000517e-01 8.25150013e-01
9.21357796e-02 6.67281449e-01 2.70852625e-01 -5.99374294e-01
-1.37217629e+00 -1.32370031e+00 5.94167709e-01 -1.84995472e-01
3.15173030e-01 1.16251715e-01 -1.31092143e+00 5.31977654e-01
-8.33922923e-02 1.01955645e-01 7.66038418e-01 6.66702986e-02
-3.77013952e-01 -2.78674275e-01 -9.76771116e-01 4.33218151e-01
7.14910090e-01 -2.36970663e-01 -7.12620914e-01 5.93797527e-02
7.76270688e-01 -6.52574062e-01 -8.11645508e-01 5.74240088e-01
3.81630599e-01 -7.28029072e-01 9.17172134e-01 -4.97282445e-01
2.89253891e-01 -4.00304943e-01 2.05748394e-01 -1.44745040e+00
-2.53088266e-01 -3.50805849e-01 -5.69398940e-01 1.47058320e+00
4.07223493e-01 -4.73054945e-01 7.42286205e-01 5.87124228e-01
-5.09153962e-01 -1.29469776e+00 -5.44578373e-01 -7.59557903e-01
1.22393919e-02 -4.08760220e-01 5.24763525e-01 1.23631144e+00
-3.43377560e-01 4.06912088e-01 -1.31825790e-01 3.23379368e-01
4.35281157e-01 2.54786491e-01 7.51132011e-01 -1.30388641e+00
-2.06618056e-01 -4.02115792e-01 -6.56875074e-02 -1.26202011e+00
2.41886675e-01 -1.03768277e+00 -1.14490323e-01 -1.69030488e+00
2.15843707e-01 -8.02965045e-01 -6.21362567e-01 1.08056676e+00
-4.93627489e-01 2.32886747e-02 7.95164555e-02 1.21223725e-01
-4.59480941e-01 7.20886409e-01 1.57210863e+00 -3.60552035e-02
-5.78504801e-01 1.31323636e-01 -8.40183437e-01 6.85556829e-01
9.33410764e-01 -4.77802366e-01 -7.85613000e-01 -4.52532321e-01
-1.61080912e-01 -3.44753861e-01 3.97725910e-01 -8.76482904e-01
4.32789028e-01 -2.57914066e-01 4.59770858e-01 -6.78424835e-01
3.17414254e-01 -6.85586452e-01 -9.09587145e-02 3.23604107e-01
-3.30690533e-01 -3.21190685e-01 4.26550806e-01 4.68026549e-01
-9.82558578e-02 -2.99094260e-01 9.03921545e-01 4.39857543e-02
-8.49120796e-01 5.33810198e-01 3.06139529e-01 2.16737628e-01
1.03264129e+00 -1.49422526e-01 -3.99301291e-01 5.38931554e-03
-4.35977340e-01 6.47798955e-01 1.70305505e-01 4.23824161e-01
7.93748915e-01 -1.26902938e+00 -6.39604032e-01 2.60722518e-01
2.13378053e-02 7.09354639e-01 2.84415931e-01 8.62686276e-01
-1.85256332e-01 1.96059391e-01 -4.45238501e-02 -4.65964019e-01
-1.39140022e+00 3.90003055e-01 1.34005681e-01 -2.98030674e-01
-5.60826242e-01 1.15285051e+00 2.52598912e-01 -7.02312827e-01
4.76564914e-01 -4.36651707e-01 -3.57586384e-01 4.32335623e-02
5.22059202e-01 2.67458409e-01 7.59589374e-02 -4.28126693e-01
-2.13046581e-01 4.53186631e-01 -5.10329962e-01 2.92530179e-01
1.44628251e+00 1.76573083e-01 3.24352264e-01 1.34763405e-01
9.11036134e-01 -6.68752715e-02 -1.50894725e+00 -6.92491949e-01
-2.70161480e-01 -1.96682155e-01 3.19478720e-01 -9.69137251e-01
-1.28583479e+00 8.15315127e-01 3.70618135e-01 -1.84753299e-01
1.40112281e+00 -7.87438452e-03 7.04866707e-01 3.23454916e-01
-3.05818021e-02 -1.07434058e+00 1.86209947e-01 5.55471539e-01
6.61609769e-01 -1.11850965e+00 1.21932611e-01 -3.05427194e-01
-6.64645195e-01 8.57451200e-01 1.16152513e+00 2.37103060e-01
6.84168577e-01 -5.35062067e-02 -5.50949872e-02 1.41216397e-01
-9.07667398e-01 -2.71533698e-01 7.57893980e-01 3.05147111e-01
2.98315853e-01 -8.38304013e-02 -7.97564313e-02 9.17481184e-01
-1.65782511e-01 -5.21036126e-02 1.81667805e-01 1.17159081e+00
-5.15949428e-01 -1.10811150e+00 -1.41151026e-01 3.94882351e-01
-5.48379868e-02 -2.13730901e-01 -3.97086918e-01 6.32908762e-01
3.20284486e-01 7.78430820e-01 -3.48286480e-01 -5.64974904e-01
3.26209813e-01 1.85613841e-01 3.47388417e-01 -7.48490334e-01
-5.80738783e-01 2.26179268e-02 -1.89740628e-01 -2.80664861e-01
-4.56398994e-01 -4.31145191e-01 -1.44052815e+00 -8.60066861e-02
-5.39719820e-01 2.38307253e-01 5.06782353e-01 1.05415988e+00
5.80886126e-01 6.15102232e-01 6.66123271e-01 -5.81665993e-01
-4.80970651e-01 -9.02824759e-01 -3.21047395e-01 1.18786953e-01
1.98258564e-01 -9.28255022e-01 -2.59788871e-01 1.77147657e-01] | [9.50477123260498, 3.3220789432525635] |
255d9154-b11b-4683-a61c-b4cfca77e816 | textit-facialfilmroll-high-resolution-multi | 2110.02124 | null | https://arxiv.org/abs/2110.02124v2 | https://arxiv.org/pdf/2110.02124v2.pdf | FacialFilmroll: High-resolution multi-shot video editing | We present FacialFilmroll, a solution for spatially and temporally consistent editing of faces in one or multiple shots. We build upon unwrap mosaic [Rav-Acha et al. 2008] by specializing it to faces. We leverage recent techniques to fit a 3D face model on monocular videos to (i) improve the quality of the mosaic for edition and (ii) permit the automatic transfer of edits from one shot to other shots of the same actor. We explain how FacialFilmroll is integrated in post-production facility. Finally, we present video editing results using FacialFilmroll on high resolution videos. | ['Pierre Hellier', 'Paul Ghezzo', 'Tim Christensen', 'Junghyun Ahn', 'Cédric Thébault', 'Philippe Henri Gosselin', 'Gilles Puy', 'Emmanuel Jolly', 'Bharath Bhushan Damodaran'] | 2021-10-05 | null | null | null | null | ['face-model'] | ['computer-vision'] | [ 3.35579872e-01 -1.95924640e-01 4.04804409e-01 -3.67550045e-01
-6.95235133e-02 -5.90346456e-01 6.07913375e-01 -5.79702735e-01
1.78898931e-01 4.64232683e-01 2.78530717e-01 1.77025124e-01
7.63885975e-02 -2.23137707e-01 -6.56244040e-01 -1.28295720e-01
-2.74382472e-01 4.92212363e-03 2.50364184e-01 5.02206832e-02
2.46965095e-01 8.71023476e-01 -1.69859433e+00 6.31258130e-01
8.00337791e-02 7.53934443e-01 1.82567656e-01 8.97611141e-01
2.32607350e-01 6.84376240e-01 -1.42816812e-01 -6.03313386e-01
5.92455864e-01 -4.87929642e-01 -5.98313630e-01 4.78398979e-01
1.13152862e+00 -5.37483215e-01 -4.37233776e-01 6.47994697e-01
1.88788608e-01 1.38432104e-02 4.27639306e-01 -1.37843513e+00
-4.34771955e-01 1.30046993e-01 -9.68304515e-01 1.29162371e-01
6.94282889e-01 5.95301464e-02 1.99906528e-01 -1.29470062e+00
1.39941740e+00 1.35660172e+00 6.43919587e-01 4.68931496e-01
-1.16440988e+00 -7.55334198e-01 -6.33691102e-02 2.00748324e-01
-1.61091924e+00 -1.43292975e+00 5.02912998e-01 -6.22304678e-01
8.81882489e-01 4.21004772e-01 8.96544099e-01 6.93293095e-01
2.74917513e-01 1.70377642e-03 8.08927417e-01 -5.31193495e-01
-2.20362052e-01 -4.78373289e-01 -3.36829185e-01 8.19358408e-01
-2.55835891e-01 4.54148427e-02 -1.07414079e+00 -2.70936131e-01
1.29381728e+00 -1.51788369e-01 -1.49879515e-01 -3.74410182e-01
-1.10692525e+00 2.38297507e-01 -1.96958438e-01 6.90974891e-02
-5.76318026e-01 3.65154117e-01 1.27284944e-01 5.16505897e-01
5.72609663e-01 1.00627929e-01 -1.12994254e-01 -3.22235942e-01
-1.39854479e+00 3.29878569e-01 4.58776504e-01 1.20799100e+00
8.88163388e-01 3.07869613e-02 -7.77834430e-02 6.81682348e-01
1.93335444e-01 2.76613951e-01 -1.33657411e-01 -1.99310184e+00
4.13891152e-02 -1.89396217e-02 1.19170673e-01 -1.04880738e+00
2.49999762e-03 4.96721923e-01 -3.94632757e-01 6.91672385e-01
5.35726473e-02 -1.32627487e-01 -7.50091076e-01 1.46901417e+00
5.89846075e-01 6.00465953e-01 -5.20381868e-01 5.60863376e-01
6.60318851e-01 4.90069300e-01 -2.80192494e-01 -6.67269588e-01
1.10241199e+00 -9.16022241e-01 -1.01477575e+00 2.15397656e-01
1.38474908e-03 -1.39733100e+00 4.83110040e-01 5.24583220e-01
-1.63027918e+00 -4.05075520e-01 -8.50424111e-01 -1.91796273e-01
2.11588562e-01 1.94559053e-01 2.99053907e-01 5.34974694e-01
-1.53298092e+00 8.97365928e-01 -8.40707541e-01 -4.42525417e-01
4.35425013e-01 3.22126627e-01 -1.03611159e+00 1.16273738e-01
-5.40108085e-01 6.98444068e-01 -2.29149997e-01 -1.10755078e-02
-8.19610953e-01 -7.36547351e-01 -7.23530889e-01 -2.11634904e-01
4.14407492e-01 -6.58127606e-01 1.42036581e+00 -1.34081554e+00
-1.81285417e+00 1.21682072e+00 -6.31699562e-01 2.70052701e-02
5.22297740e-01 -3.53171259e-01 -3.73952866e-01 6.26711786e-01
-1.26896128e-02 8.78088474e-01 1.47170436e+00 -1.13958800e+00
-2.58630991e-01 -2.26460010e-01 9.20245424e-04 3.46217155e-01
-4.97440547e-02 6.87101483e-01 -9.05643046e-01 -8.55128169e-01
-1.63623661e-01 -8.34514380e-01 3.24783862e-01 7.67318368e-01
-5.10879606e-02 4.17412311e-01 1.33992493e+00 -1.12728596e+00
1.24949622e+00 -2.33488560e+00 3.93755972e-01 1.71211660e-01
1.73955575e-01 2.78447747e-01 -2.86467284e-01 8.04337859e-01
-4.13973033e-01 -4.00879234e-02 -9.28253084e-02 -8.45555246e-01
-3.99527669e-01 7.79819041e-02 -1.77869916e-01 7.78618753e-01
-6.28021210e-02 6.24527931e-01 -3.19444835e-01 -6.02173686e-01
2.54588842e-01 5.41664362e-01 -7.60550678e-01 1.76749766e-01
8.63032416e-02 4.64454710e-01 4.67740804e-01 8.07547629e-01
9.54227448e-01 2.19948292e-01 3.70833367e-01 -2.82462955e-01
-5.17824471e-01 -1.27066821e-01 -1.42181063e+00 1.85079050e+00
-9.03687477e-02 8.67688060e-01 9.15087938e-01 -6.85597733e-02
5.38447618e-01 5.72411299e-01 6.59806371e-01 -7.96633661e-02
-1.48203969e-01 -3.72408718e-01 -5.72134614e-01 -5.57984233e-01
7.04559922e-01 -5.12593538e-02 5.05286574e-01 6.80578828e-01
5.48148602e-02 -2.61014223e-01 3.79974395e-01 4.17783290e-01
9.46164787e-01 7.38635182e-01 5.39135277e-01 -1.26941621e-01
2.69058526e-01 -3.18971336e-01 3.96886289e-01 2.32067823e-01
-9.45282131e-02 1.02144074e+00 4.41925824e-01 -1.86595559e-01
-1.20644510e+00 -1.04556870e+00 -8.77157226e-02 1.10188770e+00
-9.32505801e-02 -8.02646875e-01 -8.47192883e-01 -1.52976304e-01
-1.35310829e-01 2.34019965e-01 -7.42371023e-01 3.68675679e-01
-5.67954779e-01 2.59337485e-01 4.44185436e-01 2.37390816e-01
4.15734090e-02 -9.80283797e-01 -8.30058455e-01 -1.13840155e-01
7.06698596e-02 -9.25794125e-01 -1.06193876e+00 -4.70470071e-01
-5.67548335e-01 -1.19743061e+00 -5.58116436e-01 -6.82868123e-01
6.78384542e-01 6.65576816e-01 9.29998636e-01 6.19257033e-01
-4.11534756e-01 7.74134636e-01 -3.50011945e-01 -3.22482549e-04
-4.06524688e-01 -4.42213058e-01 3.60745519e-01 2.39425242e-01
-2.89964080e-01 -8.12559009e-01 -1.28638566e-01 4.78079319e-01
-9.61201429e-01 5.97616553e-01 -2.49431163e-01 2.94301599e-01
2.95706540e-01 -3.74638081e-01 1.40663832e-01 -5.21192074e-01
3.01870465e-01 -1.61713094e-01 -6.89697981e-01 2.41939560e-01
-3.67243662e-02 -5.44426084e-01 6.06178977e-02 -2.08375543e-01
-1.29415035e+00 3.44100744e-01 9.99121070e-02 -1.05780029e+00
1.58410370e-01 7.41005465e-02 9.28365290e-02 -3.31199706e-01
3.97481978e-01 -3.25412065e-01 3.26565653e-01 -4.97051299e-01
4.90655214e-01 5.75316012e-01 7.40453064e-01 -3.29708695e-01
8.22255611e-01 6.56075358e-01 -7.54046906e-03 -1.10051835e+00
-8.69637821e-03 -1.87807903e-01 -1.27060032e+00 -6.91911936e-01
7.85713434e-01 -1.02890229e+00 -5.94229460e-01 6.60115898e-01
-1.22757435e+00 -4.71245021e-01 -1.33952558e-01 1.48618251e-01
-7.90377736e-01 5.82738400e-01 -4.93759453e-01 -5.57811260e-01
4.18036059e-02 -8.31913710e-01 1.14640284e+00 2.34305486e-01
-5.68715215e-01 -4.90569592e-01 3.98493826e-01 -4.48965505e-02
2.06215113e-01 2.89098233e-01 3.39460105e-01 3.39406908e-01
-5.26966810e-01 1.90481082e-01 -9.66031477e-02 8.46062973e-02
2.69315064e-01 1.09520054e+00 -9.03371453e-01 -3.03815335e-01
-2.48749688e-01 -4.93393391e-02 2.56054878e-01 5.86653531e-01
7.76132882e-01 -2.26894096e-01 -3.59535843e-01 8.26855659e-01
9.61480916e-01 2.31076241e-01 9.36742961e-01 3.06866765e-02
5.04640996e-01 7.66498685e-01 5.96332550e-01 6.32749677e-01
-1.07737603e-02 1.17043424e+00 2.03771979e-01 2.86293447e-01
-3.33337396e-01 -2.04450697e-01 5.81820369e-01 6.83641911e-01
-4.33090091e-01 3.21809083e-01 -4.13722008e-01 4.40316498e-01
-1.84876847e+00 -1.50587904e+00 -1.66133389e-01 2.04411459e+00
7.49494851e-01 -7.12023973e-01 1.75093755e-01 -3.70840490e-01
1.06153512e+00 1.11570425e-01 -9.97449160e-02 -5.25673747e-01
-1.17120147e-01 2.90549159e-01 3.35652791e-02 7.20608771e-01
-9.76973653e-01 9.01176453e-01 7.55159044e+00 2.73987353e-01
-9.36409473e-01 2.81906724e-01 1.17282145e-01 -7.33991385e-01
-2.42720805e-02 2.25419194e-01 -4.43402499e-01 2.84907162e-01
6.81440473e-01 -4.40803349e-01 9.65085685e-01 4.49868500e-01
4.29731488e-01 -4.46245253e-01 -1.12821984e+00 1.06052685e+00
3.06760281e-01 -1.60481322e+00 -1.71557263e-01 2.05830261e-01
8.44033301e-01 -3.39145660e-01 -2.05473199e-01 -2.26999760e-01
-5.74705265e-02 -1.08745444e+00 1.00974131e+00 8.74868989e-01
1.21311164e+00 -6.77088499e-01 -1.38326928e-01 -2.61242725e-02
-1.42516184e+00 2.64346510e-01 -3.16700369e-01 -1.92150399e-01
7.00378299e-01 2.30888098e-01 -5.57747841e-01 5.96158743e-01
8.39044034e-01 9.49653208e-01 -4.70258176e-01 9.03973043e-01
7.13393167e-02 -4.40281667e-02 -3.13387215e-01 9.94754076e-01
-4.42009121e-01 -4.58134115e-01 6.29465342e-01 1.02001214e+00
6.44346118e-01 3.90807450e-01 -1.18029900e-01 4.51256514e-01
-2.22443461e-01 -8.66411626e-02 -8.86774063e-01 -2.09768265e-02
6.14620686e-01 1.37841535e+00 -5.49916506e-01 -3.42417061e-01
-3.15479398e-01 1.29288006e+00 2.67718345e-01 1.65658444e-01
-7.75567174e-01 -1.82412297e-01 9.54400659e-01 4.10098165e-01
6.39734745e-01 -4.96626794e-01 -4.84400988e-02 -1.11168504e+00
-4.17561531e-02 -8.76454830e-01 -3.20109650e-02 -1.34026921e+00
-7.74608970e-01 4.57725883e-01 3.28760922e-01 -1.14901590e+00
-2.63302833e-01 -1.65846601e-01 -6.69487834e-01 8.06803226e-01
-8.95630777e-01 -1.29031038e+00 -3.13852668e-01 7.70138025e-01
5.12091279e-01 -1.00394279e-01 7.15037107e-01 3.80555689e-01
-4.03174430e-01 1.70260102e-01 -1.83161914e-01 -3.83343220e-01
1.19347703e+00 -5.30724883e-01 6.07384741e-01 1.06168294e+00
1.18238300e-01 7.77889967e-01 6.04763091e-01 -9.57711160e-01
-1.40108788e+00 -9.20740545e-01 6.95980370e-01 -4.77635115e-01
2.86576658e-01 -2.77971447e-01 -8.02781999e-01 1.25287080e+00
8.25798631e-01 -1.12766899e-01 4.99038637e-01 -3.35753530e-01
-9.69100595e-02 1.51239842e-01 -1.29740429e+00 7.20645368e-01
1.28930974e+00 -5.46125472e-01 -5.17005086e-01 -1.91038474e-04
3.25305879e-01 -6.69651151e-01 -9.34163928e-01 6.43461421e-02
1.06286907e+00 -1.47111833e+00 8.35085869e-01 -2.93806911e-01
4.28422332e-01 -5.53206086e-01 -6.07848689e-02 -1.08702123e+00
-4.09964472e-01 -1.38732326e+00 -1.72655538e-01 1.27437377e+00
-3.54592502e-01 -1.76220804e-01 3.05061460e-01 5.29290795e-01
-1.70463324e-01 -2.61579812e-01 -1.18968284e+00 -4.64762598e-01
-4.71505105e-01 -1.85358286e-01 5.15323639e-01 9.66122150e-01
1.55062377e-01 -2.81072199e-01 -8.84921551e-01 -3.34545225e-03
4.09320801e-01 -1.87772036e-01 1.25357628e+00 -1.12212634e+00
-2.24438518e-01 3.16191688e-02 -1.62593365e-01 -6.36009455e-01
3.45262140e-02 -6.28892839e-01 -1.93467885e-01 -1.18844843e+00
8.39521587e-02 1.38541594e-01 6.06846809e-01 6.99785352e-01
3.33918452e-01 8.06193411e-01 6.41638041e-01 4.03391719e-01
-2.22206801e-01 3.14774096e-01 9.55274403e-01 4.47465420e-01
-1.12405777e-01 -5.03640473e-01 -4.60563660e-01 7.96104550e-01
3.74070734e-01 -4.04677331e-01 -2.58870959e-01 -3.95833641e-01
1.31218553e-01 2.93894351e-01 4.05898243e-01 -9.06988740e-01
1.10392511e-01 -4.43558335e-01 4.11876231e-01 -5.32998145e-01
7.77618408e-01 -8.86054575e-01 1.24245429e+00 4.43113372e-02
5.66669255e-02 6.36057258e-01 2.68539876e-01 2.85628319e-01
1.67856395e-01 -3.94276381e-01 9.82114673e-01 -1.92788765e-01
-5.78670442e-01 2.53824711e-01 -6.61610901e-01 -4.85963613e-01
1.38704932e+00 -3.38951409e-01 -2.87207514e-01 -6.24459982e-01
-8.97823453e-01 -1.77699685e-01 1.38415051e+00 4.45183933e-01
5.97564876e-01 -1.44998598e+00 -5.07634938e-01 5.79660714e-01
-2.85287589e-01 -5.32543004e-01 4.92403984e-01 9.30485725e-01
-1.13182628e+00 -9.80111808e-02 -6.08241558e-01 -4.36258137e-01
-2.16710591e+00 2.75782317e-01 1.32703587e-01 3.86409760e-01
-7.44666159e-01 5.12766361e-01 1.16692320e-01 8.26061219e-02
-1.59275055e-01 2.61953712e-01 1.64764777e-01 9.48765129e-02
9.99735951e-01 6.09910965e-01 3.94300446e-02 -1.01452744e+00
-5.87341070e-01 8.37982535e-01 2.13410184e-02 -5.41252613e-01
1.43409169e+00 -5.00664592e-01 -4.03061628e-01 2.03901500e-01
8.70162487e-01 5.18089414e-01 -1.71707225e+00 2.39268243e-01
-5.37298560e-01 -9.59309459e-01 -1.55864924e-01 -4.46693331e-01
-1.02029586e+00 5.42117000e-01 2.36476257e-01 -1.87251881e-01
1.07665849e+00 -2.83500850e-01 3.97176474e-01 -1.02689210e-02
3.66573572e-01 -1.04062319e+00 -1.60565421e-01 3.38462234e-01
1.25098944e+00 -3.50234509e-01 5.93831420e-01 -5.60052514e-01
-8.81956458e-01 1.42737889e+00 2.79971898e-01 -3.97739448e-02
7.39514172e-01 6.94247127e-01 6.22094832e-02 -3.93371701e-01
-9.67655063e-01 3.74989212e-01 1.07297182e-01 7.83968568e-01
4.30359930e-01 -3.29289734e-01 -1.16706312e-01 -6.05701990e-02
6.24800585e-02 5.43734372e-01 8.87787879e-01 1.20170712e+00
-2.56401241e-01 -1.04589772e+00 -7.84419656e-01 2.04226211e-01
-2.09455669e-01 1.89245269e-02 -4.36726809e-01 5.79246819e-01
6.43813908e-02 9.05890942e-01 2.65989125e-01 -1.52031958e-01
2.23846287e-01 2.43575722e-01 1.06787753e+00 -7.62432456e-01
-6.62227631e-01 3.31943035e-01 3.02819490e-01 -7.96055198e-01
-8.18058193e-01 -1.13030601e+00 -7.84553468e-01 -7.09542096e-01
3.16057689e-02 -2.94562221e-01 5.30953228e-01 6.18514717e-01
8.08751345e-01 1.68092757e-01 5.70860922e-01 -1.68087792e+00
1.16354413e-01 -8.41931760e-01 -6.87244654e-01 1.61641657e-01
1.39478803e-01 -8.50726247e-01 -4.90536205e-02 9.64935482e-01] | [12.91562271118164, -0.4365653693675995] |
931d6e9c-3bc5-45d8-a921-ef75a2b33012 | progress-and-summary-of-reinforcement | 2211.04001 | null | https://arxiv.org/abs/2211.04001v1 | https://arxiv.org/pdf/2211.04001v1.pdf | Progress and summary of reinforcement learning on energy management of MPS-EV | The high emission and low energy efficiency caused by internal combustion engines (ICE) have become unacceptable under environmental regulations and the energy crisis. As a promising alternative solution, multi-power source electric vehicles (MPS-EVs) introduce different clean energy systems to improve powertrain efficiency. The energy management strategy (EMS) is a critical technology for MPS-EVs to maximize efficiency, fuel economy, and range. Reinforcement learning (RL) has become an effective methodology for the development of EMS. RL has received continuous attention and research, but there is still a lack of systematic analysis of the design elements of RL-based EMS. To this end, this paper presents an in-depth analysis of the current research on RL-based EMS (RL-EMS) and summarizes the design elements of RL-based EMS. This paper first summarizes the previous applications of RL in EMS from five aspects: algorithm, perception scheme, decision scheme, reward function, and innovative training method. The contribution of advanced algorithms to the training effect is shown, the perception and control schemes in the literature are analyzed in detail, different reward function settings are classified, and innovative training methods with their roles are elaborated. Finally, by comparing the development routes of RL and RL-EMS, this paper identifies the gap between advanced RL solutions and existing RL-EMS. Finally, this paper suggests potential development directions for implementing advanced artificial intelligence (AI) solutions in EMS. | ['Yuanjian Zhang', 'Jingjing Jiang', 'Jihan Li', 'Zhuoran Hou', 'Liang Chu', 'Yang Lin', 'Jincheng Hu'] | 2022-11-08 | null | null | null | null | ['energy-management'] | ['time-series'] | [-1.30834743e-01 6.31874725e-02 -6.33572280e-01 2.51607895e-01
-1.36530235e-01 -4.98571664e-01 3.72037590e-01 -3.93767297e-01
-3.25107843e-01 1.01016700e+00 -4.54584181e-01 -2.85674661e-01
-5.25167823e-01 -9.02080536e-01 -3.82550418e-01 -1.11244845e+00
2.25199968e-01 1.26854688e-01 -2.25587323e-01 -3.19630235e-01
3.55062485e-01 8.02451313e-01 -1.85799694e+00 -1.98677838e-01
1.08271754e+00 8.93696547e-01 9.82835174e-01 6.25078321e-01
4.60006930e-02 8.15767765e-01 -8.06604385e-01 3.47395629e-01
1.25301868e-01 -6.92571640e-01 -6.19532108e-01 -6.50850907e-02
-9.01650310e-01 -1.23864561e-02 -1.01843495e-02 9.18451011e-01
4.97750491e-01 7.38983452e-01 8.06880057e-01 -2.16963673e+00
-1.24848641e-01 2.89788932e-01 -1.57064497e-01 1.67844996e-01
-1.93387643e-01 3.97476315e-01 3.27431202e-01 -5.98537982e-01
1.99544877e-01 8.63160670e-01 2.29491755e-01 7.97588885e-01
-5.90589166e-01 -6.55654669e-01 -2.49962360e-01 9.46433604e-01
-1.06451261e+00 -1.49755284e-01 1.03653038e+00 -1.08961925e-01
1.53888619e+00 4.32360023e-01 1.18712056e+00 6.11652255e-01
4.96780723e-01 8.78866971e-01 1.29466593e+00 -5.74259341e-01
6.47535443e-01 4.30552542e-01 -2.94541895e-01 2.92174757e-01
3.32622796e-01 5.55777371e-01 8.75789076e-02 4.19715047e-01
3.05033267e-01 -4.93112504e-01 3.96495938e-01 -3.66409302e-01
-5.19564271e-01 9.03313756e-01 2.52968192e-01 5.96799374e-01
-3.65301579e-01 1.77932516e-01 3.93178821e-01 6.90371096e-02
-9.73498002e-02 7.94374347e-01 -4.74452823e-01 -1.59723759e-01
-5.67269444e-01 1.21297255e-01 5.14393568e-01 1.00491190e+00
6.43950105e-01 7.86623478e-01 -1.06458627e-02 8.86920452e-01
3.42851013e-01 7.14128792e-01 4.45256293e-01 -1.58319092e+00
8.38890448e-02 3.65675390e-01 2.56977856e-01 -5.91521025e-01
-7.46142745e-01 -2.02436209e-01 -7.68706203e-01 7.50952065e-01
-6.30101085e-01 -5.62273204e-01 -2.65944779e-01 1.34052622e+00
6.72158524e-02 -3.83623928e-01 5.92536688e-01 5.34873664e-01
7.29540050e-01 1.33836865e+00 2.22380415e-01 -8.49996150e-01
9.60195005e-01 -1.42660308e+00 -1.30900729e+00 -1.45599261e-01
4.24377561e-01 -1.91983521e-01 3.09126437e-01 4.35099900e-01
-1.23729718e+00 -8.70023191e-01 -1.33397818e+00 4.99832541e-01
-9.55104053e-01 3.16900373e-01 3.75090629e-01 7.31809258e-01
-8.68533611e-01 6.35145426e-01 -5.52840948e-01 -3.64587694e-01
4.97219749e-02 4.46095496e-01 3.46158534e-01 4.95404541e-01
-1.22989786e+00 1.59616220e+00 8.89122486e-01 1.88812807e-01
-7.43520021e-01 -4.83170629e-01 -7.48642743e-01 -2.15557054e-01
5.05082369e-01 -5.27337730e-01 1.38214767e+00 -9.46650803e-01
-2.18247509e+00 1.27295852e-01 1.49354205e-01 -2.67134696e-01
5.96359447e-02 2.65638158e-02 -7.71980464e-01 4.09192741e-01
-1.47493601e-01 6.17505968e-01 5.93940079e-01 -1.56571853e+00
-1.18798018e+00 2.77266175e-01 -2.16376439e-01 3.57807368e-01
-1.50969893e-01 4.46704738e-02 2.93542117e-01 -2.14270175e-01
-7.39021242e-01 -8.13391626e-01 -2.53434479e-01 -6.03138566e-01
1.89782962e-01 -1.02510452e+00 1.28457177e+00 -2.89896190e-01
1.29410994e+00 -1.79610789e+00 1.66339919e-01 1.57385439e-01
-2.64008433e-01 5.17574787e-01 8.61746818e-02 6.50657415e-01
8.65019783e-02 2.60654874e-02 6.72056079e-02 1.09729923e-01
2.53203213e-01 7.56108046e-01 1.39142007e-01 1.64968930e-02
1.47372112e-01 1.02666080e+00 -1.10754240e+00 -4.57488894e-01
1.03393102e+00 1.26630366e-01 -7.02870116e-02 5.72411537e-01
-1.98630586e-01 2.66774386e-01 -6.44443870e-01 5.29428065e-01
4.25942987e-01 3.57579231e-01 1.45416055e-02 -4.10774678e-01
-8.71125698e-01 -4.41909730e-01 -1.07049084e+00 1.11144423e+00
-5.70688963e-01 5.49608171e-01 3.30465764e-01 -1.43486571e+00
9.10914779e-01 4.08342868e-01 1.15147567e+00 -1.13916469e+00
2.96247691e-01 4.54079032e-01 -1.54832184e-01 -9.88675296e-01
6.15270793e-01 -1.78687543e-01 4.14311364e-02 -5.52351773e-02
-1.48037430e-02 -5.08985579e-01 6.69871986e-01 -3.90638977e-01
6.28897071e-01 5.06285548e-01 4.00233865e-01 -3.43901992e-01
9.08753216e-01 3.30499381e-01 8.69004011e-01 2.88613260e-01
-4.01130497e-01 -4.51335698e-01 -9.43742320e-02 -1.34693384e-01
-1.04770947e+00 -8.95465314e-01 1.59199685e-01 7.66475081e-01
5.10204196e-01 -1.46784633e-01 -5.96797466e-01 -3.75157833e-01
-1.74273640e-01 1.43655705e+00 7.55151361e-02 -5.46958506e-01
-6.18892729e-01 -7.87904143e-01 2.33734727e-01 7.05756605e-01
9.66815054e-01 -1.21846437e+00 -1.26682413e+00 3.52792621e-01
-3.45247924e-01 -1.13171625e+00 2.29238838e-01 3.13560724e-01
-6.95752263e-01 -1.09302342e+00 -3.16670179e-01 -7.65341938e-01
5.04189253e-01 3.20614845e-01 8.38462174e-01 -1.54422492e-01
-1.77376777e-01 6.74952686e-01 -3.70918661e-01 -9.75449085e-01
-7.88383603e-01 -3.83985698e-01 2.31643617e-01 -6.67749465e-01
3.57407033e-01 -1.22584023e-01 -6.76572144e-01 4.32234764e-01
-5.76640546e-01 2.62936622e-01 9.19889212e-01 6.38912082e-01
5.67272663e-01 1.07359302e+00 1.05337369e+00 3.51028284e-03
7.29322135e-01 -5.05298913e-01 -7.33038425e-01 1.79723829e-01
-1.18169677e+00 -1.66825682e-01 8.18781495e-01 -2.40731686e-02
-1.43863308e+00 -1.32066563e-01 -2.14596242e-01 -5.37106216e-01
-4.06200320e-01 -1.86834663e-01 -1.35926470e-01 -5.45427948e-02
1.97602212e-02 2.24507973e-01 3.05528104e-01 -1.45617947e-01
1.78956360e-01 6.49396658e-01 3.26989472e-01 -5.29404819e-01
7.80433714e-01 -1.78835586e-01 4.75973666e-01 -9.20136273e-01
-1.53894499e-01 -3.95868510e-01 -1.36273161e-01 -8.79034102e-01
1.14319384e+00 -4.39343989e-01 -1.00462282e+00 4.84766901e-01
-9.32493389e-01 -5.53625405e-01 -5.91597736e-01 6.76311076e-01
-9.73360002e-01 -1.51901161e-02 -3.61184567e-01 -1.20737672e+00
-4.58232462e-01 -1.36245823e+00 2.93974608e-01 7.12128460e-01
1.12475425e-01 -1.01232171e+00 -6.09898083e-02 4.71543044e-01
7.00259566e-01 4.25501138e-01 9.87007439e-01 -3.72807793e-02
-3.24899822e-01 3.58014494e-01 3.08168530e-01 8.23386490e-01
7.54566118e-02 1.87092930e-01 -7.27005363e-01 -4.44714099e-01
6.69683143e-02 -1.77628875e-01 2.54458666e-01 4.29804862e-01
1.19934785e+00 -2.89192945e-01 -5.63568652e-01 1.51129052e-01
1.81061709e+00 1.26711988e+00 5.53273737e-01 7.09792852e-01
1.74577594e-01 5.96685708e-01 1.06912518e+00 4.45309490e-01
2.76416004e-01 4.50244129e-01 6.79309547e-01 -2.64680505e-01
-3.88305858e-02 1.03146270e-01 6.03773773e-01 1.18369043e+00
-5.76251566e-01 -3.72856051e-01 -3.07847649e-01 1.92678183e-01
-1.70723021e+00 -1.16603279e+00 1.40834292e-02 1.65416598e+00
2.85575241e-01 -7.73674175e-02 2.75353611e-01 4.57270682e-01
6.77901328e-01 -2.84496516e-01 -6.45572722e-01 -1.14818645e+00
-5.61803058e-02 8.91072005e-02 5.83206475e-01 1.45065010e-01
-8.04865241e-01 4.29113925e-01 6.85683632e+00 1.03854191e+00
-8.10560524e-01 3.17192674e-02 1.58423513e-01 2.06406578e-01
1.18019385e-02 -1.82941347e-01 -5.59742272e-01 5.83856225e-01
1.07657337e+00 -5.92758060e-01 7.75565803e-01 8.75808775e-01
6.81858957e-01 -5.33838928e-01 -9.25298333e-01 9.60197806e-01
-2.24228367e-01 -9.99799311e-01 -2.54845440e-01 -1.27955258e-01
8.78979325e-01 -1.75609916e-01 -3.84425431e-01 6.34585738e-01
7.57416189e-02 -7.64235735e-01 7.02706456e-01 6.23901367e-01
3.87011170e-01 -1.33683753e+00 8.61017942e-01 3.85716647e-01
-1.48896635e+00 -7.66493917e-01 -3.39601696e-01 -1.85729079e-02
4.18050885e-02 1.78908423e-01 -3.35461736e-01 1.14202058e+00
8.63059700e-01 3.86228263e-01 -2.17821434e-01 7.99095809e-01
-2.04604030e-01 2.44208544e-01 -1.66439071e-01 -6.11188710e-01
1.73213199e-01 -6.38109028e-01 6.02277815e-01 1.08372748e+00
4.57399696e-01 1.23490624e-01 5.53845204e-02 1.05460525e+00
5.68168104e-01 -2.07776949e-01 -6.77804708e-01 -1.76191390e-01
6.40512109e-01 1.65113997e+00 -5.14933825e-01 -3.16140175e-01
-4.67309237e-01 4.33024943e-01 -3.10308188e-01 4.46232677e-01
-1.15036666e+00 -6.97145879e-01 5.96584499e-01 -4.60620105e-01
-6.36387393e-02 -1.26591071e-01 8.09422880e-02 -1.58547759e-01
-4.44267571e-01 -5.28284073e-01 3.02308589e-01 -1.07305121e+00
-1.03114009e+00 4.99053560e-02 6.38628721e-01 -1.22898185e+00
-4.20883566e-01 -8.78354788e-01 -7.67785668e-01 3.54492992e-01
-1.78748369e+00 -5.73411167e-01 -2.80311704e-01 2.39995316e-01
1.00396073e+00 -5.07651329e-01 6.14385009e-01 2.02315554e-01
-9.23398316e-01 -6.58313707e-02 7.00799644e-01 -5.46253383e-01
1.02580726e-01 -1.15779114e+00 -6.06604457e-01 5.59760273e-01
-9.07896876e-01 -2.22866446e-01 8.65325153e-01 -2.50953048e-01
-2.01746249e+00 -9.57759202e-01 2.86914319e-01 3.52051705e-01
3.04544270e-01 2.87774920e-01 -2.87515730e-01 1.65166065e-01
1.11528349e+00 -5.51290572e-01 4.63183492e-01 -9.38086033e-01
9.43920672e-01 -5.19542336e-01 -1.42005873e+00 6.11281216e-01
6.16903007e-01 7.36368001e-02 -4.99199301e-01 2.40848679e-02
3.03676605e-01 -9.59244221e-02 -9.94255006e-01 8.39027047e-01
3.10795397e-01 -7.34827757e-01 8.41314137e-01 -1.53586745e-01
9.19948146e-02 -5.39146423e-01 2.79470116e-01 -1.75923991e+00
-7.10251033e-01 -6.50204599e-01 -5.10786116e-01 1.27290940e+00
-6.96181953e-02 -7.05657065e-01 3.50245804e-01 3.20638329e-01
-7.03027129e-01 -9.96535778e-01 -9.27859247e-01 -1.13788080e+00
1.34654969e-01 -3.47016066e-01 5.80118299e-01 5.08568347e-01
2.85976380e-01 1.23565249e-01 -5.71319321e-03 4.48669642e-02
7.57953405e-01 -8.37893933e-02 4.00423914e-01 -7.43079782e-01
3.85660864e-02 -8.22021067e-01 4.91123050e-02 -3.41012001e-01
2.19758853e-01 -7.75421381e-01 1.85266510e-01 -2.27936673e+00
-5.26944250e-02 -1.80169538e-01 -6.51179075e-01 3.76275122e-01
2.01065898e-01 -3.33240271e-01 5.00752032e-01 -1.24833226e-01
-5.49708962e-01 9.52222943e-01 1.40239644e+00 -2.40543187e-01
-2.11358368e-01 6.62218500e-03 -1.23626493e-01 8.31293046e-01
1.63553536e+00 -2.76923031e-01 -8.21962714e-01 9.43346024e-02
1.51759341e-01 1.35537004e-02 2.11195290e-01 -1.34765291e+00
2.94934005e-01 -7.36665189e-01 4.36546355e-01 -1.06076539e+00
1.28128678e-01 -1.36937582e+00 5.47279596e-01 9.92049992e-01
6.90501556e-02 3.80522221e-01 3.40487838e-01 1.60353273e-01
2.77852342e-02 -4.20917362e-01 1.02700520e+00 -1.43418506e-01
-1.27296054e+00 -3.40993017e-01 -1.33811224e+00 -3.42016220e-01
1.93417537e+00 -5.14762342e-01 -2.95696169e-01 2.64698565e-01
-2.62249798e-01 8.49285662e-01 -8.14543813e-02 7.15469301e-01
6.69163346e-01 -1.42767739e+00 -1.66016176e-01 4.41543162e-02
-3.91262561e-01 -3.22386682e-01 7.61853904e-02 6.75928831e-01
-3.74490738e-01 5.94772935e-01 -7.13911831e-01 -3.43363494e-01
-1.16153300e+00 8.46504211e-01 6.58923268e-01 -1.56126812e-01
-3.68691951e-01 -4.67793867e-02 -1.52862489e-01 -1.68069974e-01
9.80788395e-02 -2.27489583e-02 -5.59363782e-01 -3.05996716e-01
-2.05855519e-02 1.46547186e+00 -1.11062281e-01 -4.05234337e-01
-2.75316894e-01 7.19874382e-01 9.16149259e-01 1.06924348e-01
1.12085056e+00 -3.99535328e-01 -1.08357430e-01 3.94512445e-01
7.99540162e-01 -4.81198400e-01 -1.04958737e+00 6.91450953e-01
-7.82898217e-02 8.00542235e-02 1.91644117e-01 -9.53249216e-01
-1.14854586e+00 5.08066535e-01 7.32993126e-01 2.45820478e-01
1.61191738e+00 -2.41173297e-01 5.91216743e-01 4.92030680e-01
5.55945873e-01 -2.31729412e+00 3.73276681e-01 2.63911307e-01
9.60585833e-01 -8.05599272e-01 5.13967387e-02 -1.45074680e-01
-5.53008139e-01 1.35639203e+00 1.02337909e+00 1.98555335e-01
4.71800059e-01 5.77363551e-01 -3.76987398e-01 1.11135527e-01
-6.59740686e-01 -1.94189206e-01 1.53915035e-02 7.98069715e-01
-5.00617847e-02 6.05157688e-02 -7.41590023e-01 6.21392071e-01
9.29891914e-02 1.02737680e-01 2.15167657e-01 1.20613134e+00
-8.16154838e-01 -1.14508271e+00 -6.21862054e-01 3.71960439e-02
-2.07056984e-01 7.33214498e-01 1.38098136e-01 1.04358900e+00
6.24617100e-01 1.59536791e+00 -1.04841113e-01 -4.20928955e-01
5.32918692e-01 -8.79525319e-02 4.60083216e-01 1.09766051e-01
-4.15860683e-01 -1.87211558e-01 1.93862900e-01 -4.43697155e-01
-8.46282840e-01 -3.58377337e-01 -1.89682531e+00 -2.72026539e-01
-2.63190657e-01 6.07901275e-01 1.06596208e+00 9.99408305e-01
1.42385080e-01 1.16088009e+00 1.23726714e+00 -8.66426170e-01
-3.72983962e-01 -5.54550946e-01 -5.98356783e-01 -1.14108354e-01
1.80249661e-02 -8.26203823e-01 -4.81544048e-01 -3.03225607e-01] | [5.533316135406494, 2.324204683303833] |
4b57eb14-1823-485e-b6fa-4e26235a43ab | performance-evaluation-of-3d-correspondence | 1804.02085 | null | http://arxiv.org/abs/1804.02085v1 | http://arxiv.org/pdf/1804.02085v1.pdf | Performance Evaluation of 3D Correspondence Grouping Algorithms | This paper presents a thorough evaluation of several widely-used 3D
correspondence grouping algorithms, motived by their significance in vision
tasks relying on correct feature correspondences. A good correspondence
grouping algorithm is desired to retrieve as many as inliers from initial
feature matches, giving a rise in both precision and recall. Towards this rule,
we deploy the experiments on three benchmarks respectively addressing shape
retrieval, 3D object recognition and point cloud registration scenarios. The
variety in application context brings a rich category of nuisances including
noise, varying point densities, clutter, occlusion and partial overlaps. It
also results to different ratios of inliers and correspondence distributions
for comprehensive evaluation. Based on the quantitative outcomes, we give a
summarization of the merits/demerits of the evaluated algorithms from both
performance and efficiency perspectives. | ['Zhiguo Cao', 'Ke Xian', 'Jiaqi Yang', 'Yang Xiao'] | 2018-04-06 | null | null | null | null | ['3d-object-recognition'] | ['computer-vision'] | [-8.45235959e-02 -5.19590318e-01 7.55670220e-02 -3.62343848e-01
-8.37797940e-01 -6.50920212e-01 1.09226036e+00 4.61100399e-01
-1.50005028e-01 3.74061048e-01 -6.99034333e-02 3.44196558e-02
-4.96165127e-01 -4.79727656e-01 -3.62906128e-01 -7.33150363e-01
-8.71146470e-02 8.20867419e-01 1.69261098e-01 -5.99843785e-02
6.91729009e-01 1.32646823e+00 -1.94546735e+00 -3.30238283e-01
6.96781516e-01 9.68559742e-01 -7.92121142e-02 2.90948689e-01
-1.27970666e-01 -1.25760153e-01 -5.96286595e-01 -3.49538386e-01
6.06369019e-01 1.07210256e-01 -2.28126332e-01 4.66732085e-01
7.42289543e-01 2.48572454e-02 -4.76675108e-02 1.10808825e+00
5.21202385e-01 -2.04615295e-04 1.19376636e+00 -1.31877160e+00
-2.74693251e-01 -3.97674516e-02 -7.28449702e-01 9.39322561e-02
6.27340913e-01 1.81013256e-01 1.05830479e+00 -1.41520846e+00
5.40718257e-01 1.25916708e+00 6.33642077e-01 -1.44452527e-01
-1.22853887e+00 -3.92551571e-01 -7.97743052e-02 -3.50611545e-02
-1.57712591e+00 -4.75006163e-01 9.09659743e-01 -6.75006986e-01
6.89771771e-01 6.71794832e-01 6.19773924e-01 5.68913937e-01
6.23048060e-02 4.23681825e-01 1.03799403e+00 -4.63666618e-01
2.73054034e-01 6.08958416e-02 3.40743303e-01 1.15458786e-01
5.93159258e-01 2.28946120e-01 -5.56119382e-01 -4.61539745e-01
8.13561618e-01 -2.02010875e-03 -1.74232647e-01 -9.19406414e-01
-1.43290627e+00 5.74470997e-01 4.62231815e-01 1.80902913e-01
-3.07654053e-01 -6.48426488e-02 4.09146398e-01 3.25222194e-01
1.65481880e-01 3.10961902e-01 -2.58043855e-02 2.40560800e-01
-8.88991535e-01 4.06210840e-01 5.78852475e-01 1.30781996e+00
9.62375283e-01 -2.79145688e-01 -2.25296885e-01 7.47821629e-01
3.89585346e-01 7.02009857e-01 6.30944744e-02 -6.21559620e-01
2.74609029e-01 7.99085557e-01 1.59470156e-01 -1.56701827e+00
-2.97709554e-01 -6.56356573e-01 -9.07311082e-01 3.25656503e-01
2.52770960e-01 5.95111251e-01 -8.21271896e-01 1.22497880e+00
4.53759760e-01 1.54987827e-01 -2.59871423e-01 9.97509301e-01
1.00238872e+00 1.05126187e-01 -2.30687618e-01 -2.80216694e-01
1.21120799e+00 -3.42579544e-01 -3.11600000e-01 1.23806611e-01
9.75795686e-02 -1.35195374e+00 7.29949534e-01 1.02753200e-01
-9.70420659e-01 -6.37824535e-01 -8.14419925e-01 2.40487844e-01
-2.34312207e-01 1.30542487e-01 2.89794832e-01 4.36339527e-01
-9.52188015e-01 5.53559780e-01 -6.14964485e-01 -6.33257866e-01
1.90653011e-01 5.26765227e-01 -2.49710947e-01 1.41276151e-01
-3.69002223e-01 9.05170262e-01 2.03390926e-01 6.35311455e-02
-1.78349882e-01 -7.37971902e-01 -4.24920261e-01 -2.31846377e-01
-2.22259872e-02 -9.63746667e-01 7.30489731e-01 -2.83504844e-01
-9.64290917e-01 1.34903073e+00 2.51740357e-03 -2.44566679e-01
8.79748225e-01 -7.97414556e-02 -1.02812216e-01 7.63805723e-03
2.27332741e-01 3.13527763e-01 6.81562722e-01 -1.40356445e+00
-6.38584316e-01 -7.30488062e-01 -4.58394855e-01 2.70952523e-01
3.57149810e-01 1.48919553e-01 -5.27147055e-01 -6.82039201e-01
9.12310123e-01 -8.26330781e-01 -1.77308664e-01 1.03809297e-01
-3.23384225e-01 -3.52802515e-01 7.63529420e-01 3.76895629e-02
7.41076469e-01 -2.22804356e+00 5.74843474e-02 6.96327150e-01
2.19119787e-01 -3.95213207e-03 1.51526064e-01 7.16156602e-01
-3.74743566e-02 -1.50306135e-01 -7.36354887e-02 -3.48538458e-01
1.75816506e-01 7.44235516e-02 -3.97222102e-01 9.36792195e-01
1.65331200e-01 7.75774837e-01 -5.97604930e-01 -4.50538933e-01
7.17650294e-01 4.45197016e-01 -1.61824182e-01 5.83510511e-02
9.21161398e-02 5.31868935e-01 -6.29344463e-01 1.01823580e+00
9.33629215e-01 -5.73380217e-02 -5.22813976e-01 -4.00811106e-01
-2.01432034e-01 -2.24812422e-02 -1.51993906e+00 1.38428378e+00
8.68893713e-02 3.95190328e-01 -1.00250863e-01 -8.22224438e-01
1.46528220e+00 9.53358561e-02 8.00799727e-01 -3.83053005e-01
3.23922783e-01 6.89572632e-01 -3.96078557e-01 -1.95047297e-02
6.48618102e-01 8.01134557e-02 8.51817951e-02 2.06797168e-01
-3.82894501e-02 -4.33872849e-01 -5.88072948e-02 -1.17210388e-01
8.16695035e-01 -1.65964097e-01 6.86525643e-01 -6.79062724e-01
7.13593602e-01 2.12464899e-01 1.69230551e-01 7.20126748e-01
-2.52176464e-01 9.77925181e-01 6.56311363e-02 -3.86634052e-01
-1.13562393e+00 -1.17397785e+00 -5.59019685e-01 2.99192429e-01
4.57941055e-01 -1.39204308e-01 -1.53009504e-01 -1.78930745e-01
2.55775332e-01 1.10216640e-01 -2.79759109e-01 1.80795357e-01
-4.82199669e-01 -4.47124541e-01 9.92803276e-02 7.68584162e-02
3.21817219e-01 -7.48406112e-01 -7.88541734e-01 -9.91459042e-02
3.09698939e-01 -1.18022132e+00 -9.80459526e-02 2.95488797e-02
-1.19755197e+00 -1.33007419e+00 -7.42736697e-01 -7.47931600e-01
7.13278949e-01 7.72495151e-01 1.54299903e+00 2.65343875e-01
-3.08181137e-01 7.50729680e-01 -5.26409984e-01 -3.61814529e-01
-2.69788444e-01 -6.81532621e-02 2.78699845e-01 -1.12945363e-01
7.23605514e-01 -7.15307593e-01 -6.01247370e-01 5.92226863e-01
-5.84813476e-01 -3.90448809e-01 6.05633616e-01 6.73473895e-01
8.52070451e-01 -2.84541845e-01 9.45059806e-02 -3.82152081e-01
5.42064071e-01 -1.50513098e-01 -9.12515998e-01 2.95630783e-01
-4.96623427e-01 -1.24158338e-01 1.00136794e-01 -7.96564370e-02
-2.85903841e-01 2.67832428e-01 -1.32567734e-02 -6.33962274e-01
-4.82246488e-01 1.32532716e-01 -1.00983724e-01 -5.98934770e-01
7.73366868e-01 3.06047440e-01 5.21877110e-02 -4.94387537e-01
3.01286578e-01 6.14590049e-01 7.26891398e-01 -8.05589080e-01
1.24624372e+00 6.41709685e-01 3.29531759e-01 -1.00294387e+00
-2.99044400e-01 -1.03850865e+00 -9.33174729e-01 -2.71088809e-01
2.68493712e-01 -6.63195193e-01 -6.01639986e-01 4.18609321e-01
-1.25145888e+00 7.05053568e-01 -5.00769258e-01 5.40612698e-01
-8.38439345e-01 5.28077006e-01 7.40501508e-02 -8.25886369e-01
-3.11939776e-01 -1.37176681e+00 1.19531572e+00 1.33462265e-01
-1.75398946e-01 -7.99648702e-01 1.30471680e-02 -1.46121517e-01
1.70703977e-01 5.90645671e-01 6.80244148e-01 -8.01852107e-01
-8.08723450e-01 -4.74324793e-01 -3.19988251e-01 -1.32650524e-01
1.62714764e-01 1.59762844e-01 -1.04091215e+00 -4.58085239e-01
5.70359603e-02 1.33095188e-02 4.99660313e-01 4.02432054e-01
4.41121489e-01 2.32865423e-01 -3.62709522e-01 4.32083994e-01
1.64204443e+00 -1.07409060e-03 4.55176681e-01 4.61199135e-01
3.86166036e-01 6.11864746e-01 9.27897811e-01 5.87060511e-01
-2.03104287e-01 9.48256493e-01 6.54446006e-01 -5.98395281e-02
-1.15838423e-01 3.67988134e-03 -2.91064829e-01 6.60170436e-01
-6.87810928e-02 1.59643620e-01 -9.83776689e-01 4.14494991e-01
-1.74660993e+00 -6.11594141e-01 -5.09530365e-01 2.67167974e+00
2.75016695e-01 -1.28385633e-01 3.63076746e-01 3.76750618e-01
9.52002704e-01 -1.00174276e-02 -4.38414156e-01 9.62171406e-02
-3.43548685e-01 2.51210537e-02 5.41467428e-01 3.48799855e-01
-9.32880521e-01 5.90033889e-01 6.52490997e+00 8.61564040e-01
-9.72102523e-01 -3.68176877e-01 3.41656148e-01 4.09990937e-01
-2.49221012e-01 -1.48047404e-02 -9.61697221e-01 2.36016467e-01
2.34193012e-01 -1.56595856e-01 -1.87961057e-01 6.26725554e-01
7.32616261e-02 -1.89354151e-01 -1.17122746e+00 1.56535172e+00
1.18668124e-01 -1.23570180e+00 2.34738782e-01 1.17976248e-01
5.38153827e-01 1.24634489e-01 1.33058488e-01 -4.10565138e-01
-9.00506899e-02 -8.75070214e-01 8.43474746e-01 3.98850739e-01
5.53852916e-01 -6.41289294e-01 6.94493413e-01 2.46768758e-01
-1.12444377e+00 3.20078760e-01 -5.44421673e-01 6.40995279e-02
2.65063886e-02 7.16161847e-01 -7.88637400e-01 9.67847645e-01
4.99948680e-01 7.00556636e-01 -4.93550599e-01 1.74317014e+00
1.52714640e-01 -9.33327675e-02 -7.24008381e-01 -4.86720633e-03
1.31559163e-01 -6.07700944e-01 9.30786908e-01 1.16681564e+00
4.64540094e-01 1.45341396e-01 1.51721433e-01 8.97279561e-01
4.13980007e-01 4.57418054e-01 -7.17082679e-01 5.62011480e-01
9.35193419e-01 1.11498797e+00 -9.89572883e-01 -1.26858205e-01
-3.38887125e-01 3.65917772e-01 -1.72304720e-01 1.52406543e-01
-5.18995106e-01 5.17798252e-02 6.81071162e-01 3.02988589e-01
9.66233686e-02 -6.68942273e-01 -6.68038487e-01 -9.24396336e-01
1.85075060e-01 -6.53500378e-01 1.64942771e-01 -5.09647489e-01
-1.43750882e+00 5.68533242e-01 2.71351039e-01 -1.86483932e+00
-7.05771148e-03 -6.10676646e-01 -5.10398746e-01 8.54219317e-01
-1.45194757e+00 -1.14312172e+00 -6.49224401e-01 4.84622002e-01
4.50964302e-01 -4.25838083e-01 5.46550632e-01 3.50761920e-01
-1.17269136e-01 2.81690776e-01 6.96548447e-02 -1.47919685e-01
5.85871398e-01 -8.70311916e-01 4.75862831e-01 6.31194055e-01
3.63307059e-01 6.15292072e-01 1.11986065e+00 -3.90000880e-01
-1.42302990e+00 -6.99993134e-01 8.54908645e-01 -5.66248298e-01
4.69546407e-01 -1.71403922e-02 -7.24898458e-01 2.48962641e-01
-2.36221403e-01 1.49637442e-02 3.52829129e-01 6.18952885e-02
-2.46364743e-01 -4.61576506e-02 -1.09966254e+00 4.00458157e-01
9.96378899e-01 -1.85912907e-01 -6.74481988e-01 2.57028490e-01
9.75578278e-02 -5.37353158e-01 -8.92723203e-01 6.41255796e-01
3.46562326e-01 -1.46786714e+00 1.35062051e+00 -1.06600180e-01
-5.57666421e-02 -5.85622966e-01 -4.56203520e-01 -8.35594952e-01
-3.25916857e-01 -4.81928051e-01 3.76999408e-01 1.27972960e+00
-1.40054841e-02 -5.23575306e-01 8.84337306e-01 5.63629448e-01
-1.35136336e-01 -4.75246727e-01 -1.04900742e+00 -9.60273445e-01
-9.99162048e-02 -3.72753561e-01 7.36753225e-01 8.29597354e-01
-5.52485943e-01 1.25640109e-01 -2.71248482e-02 2.41044238e-01
9.68029380e-01 4.17484283e-01 1.27703345e+00 -1.77514875e+00
2.37302303e-01 -5.93131244e-01 -1.09107327e+00 -1.12918413e+00
-1.19223401e-01 -7.85980821e-01 -2.29362994e-01 -1.23142672e+00
1.54949993e-01 -9.04076099e-01 1.94537116e-03 -1.11808732e-01
1.08189538e-01 4.87017721e-01 3.38115364e-01 9.23313141e-01
-4.05161291e-01 3.31776798e-01 9.06923652e-01 -3.45712677e-02
-2.77774483e-01 4.33982283e-01 -2.93151945e-01 6.66023135e-01
6.19539142e-01 -3.00259769e-01 -4.62358519e-02 -4.02632147e-01
1.30886957e-01 -9.37907249e-02 6.12945676e-01 -1.15094411e+00
5.21403015e-01 -5.53075783e-02 2.48068094e-01 -1.22683120e+00
5.04107833e-01 -1.25932217e+00 5.22914529e-01 4.12816048e-01
1.73580721e-01 2.66283691e-01 3.58328526e-03 5.34935474e-01
-3.77920926e-01 -1.91270798e-01 9.69453216e-01 -1.31402045e-01
-6.81554198e-01 3.31062049e-01 2.02255875e-01 -2.56603241e-01
1.15253031e+00 -1.13836539e+00 1.40038952e-01 -1.37458041e-01
-6.89766943e-01 -7.53813535e-02 8.33777308e-01 2.93246061e-01
5.98574281e-01 -1.46919620e+00 -8.55462849e-01 4.25661951e-01
4.28913295e-01 -1.31469220e-02 -1.69650719e-01 9.17832732e-01
-5.45458615e-01 5.55174351e-01 -3.33570242e-01 -1.40483475e+00
-1.43490732e+00 3.62342298e-01 2.98795521e-01 2.53319263e-01
-5.98711669e-01 4.91886616e-01 -6.07574657e-02 -3.21890622e-01
3.97430569e-01 -4.29372013e-01 -8.69988725e-02 1.28123686e-01
2.64668707e-02 4.94666010e-01 5.01343906e-01 -9.31067288e-01
-5.30115962e-01 1.39439702e+00 2.17090219e-01 1.29694864e-01
1.04761481e+00 -7.28750005e-02 -8.87187570e-02 3.65594208e-01
1.00281489e+00 5.27124405e-02 -9.84202087e-01 -3.26416820e-01
3.91135931e-01 -1.01711321e+00 -2.52833933e-01 -2.36923784e-01
-7.65946269e-01 6.20456576e-01 6.44910753e-01 3.37740004e-01
8.70415330e-01 2.32246995e-01 1.96745455e-01 2.00105667e-01
6.85955703e-01 -5.61893880e-01 -1.75705075e-01 4.43399519e-01
1.04968464e+00 -1.20652604e+00 4.17443603e-01 -7.44538069e-01
-2.42455006e-01 1.19022155e+00 5.70985153e-02 -6.20918870e-01
7.68139958e-01 2.96144299e-02 4.66307029e-02 -5.08419037e-01
-1.19895160e-01 -3.12264115e-01 4.89046574e-01 7.88863540e-01
2.97769874e-01 -2.26428866e-01 -5.53167284e-01 -1.52512893e-01
-4.18408573e-01 -3.90892804e-01 7.54612237e-02 6.21140242e-01
-5.70240140e-01 -1.05989790e+00 -9.24044967e-01 3.76634657e-01
-6.06914461e-02 4.09516752e-01 -5.34196973e-01 1.03936470e+00
-2.68864542e-01 7.58902431e-01 2.27946371e-01 -1.72681242e-01
6.58973396e-01 -4.35878098e-01 4.07288253e-01 -3.91267031e-01
-6.16858840e-01 3.47733647e-01 -3.25978935e-01 -3.19217682e-01
-7.24815249e-01 -8.53291869e-01 -8.99228096e-01 -4.13729489e-01
-4.72918332e-01 2.28213340e-01 7.58463621e-01 6.91516519e-01
3.94174904e-01 -1.15063135e-02 7.36288846e-01 -1.08444273e+00
-8.12711418e-01 -7.23352313e-01 -6.72898173e-01 5.51717460e-01
3.00305843e-01 -9.81006503e-01 -6.05117798e-01 -2.34925985e-01] | [7.916311264038086, -2.6116199493408203] |
d0388f99-4c87-4045-978d-2a4f0dbd9041 | evaluating-end-to-end-entity-linking-on | 2305.14588 | null | https://arxiv.org/abs/2305.14588v1 | https://arxiv.org/pdf/2305.14588v1.pdf | Evaluating end-to-end entity linking on domain-specific knowledge bases: Learning about ancient technologies from museum collections | To study social, economic, and historical questions, researchers in the social sciences and humanities have started to use increasingly large unstructured textual datasets. While recent advances in NLP provide many tools to efficiently process such data, most existing approaches rely on generic solutions whose performance and suitability for domain-specific tasks is not well understood. This work presents an attempt to bridge this domain gap by exploring the use of modern Entity Linking approaches for the enrichment of museum collection data. We collect a dataset comprising of more than 1700 texts annotated with 7,510 mention-entity pairs, evaluate some off-the-shelf solutions in detail using this dataset and finally fine-tune a recent end-to-end EL model on this data. We show that our fine-tuned model significantly outperforms other approaches currently available in this domain and present a proof-of-concept use case of this model. We release our dataset and our best model. | ['Daniel Simig', 'Danial Lashkari', 'Thomas Chaney', 'Johannes Boehm', 'Rafael Aparecido Martins Frade', 'Khalil Kacem', 'Sebastian Cadavid-Sanchez'] | 2023-05-23 | null | null | null | null | ['entity-linking'] | ['natural-language-processing'] | [-2.44487673e-01 3.75368506e-01 -1.86893284e-01 -3.83046657e-01
-1.00465679e+00 -9.04738784e-01 8.19180489e-01 7.35269666e-01
-1.20057917e+00 1.02854776e+00 6.48559928e-01 -1.85220137e-01
-3.43777776e-01 -5.66234112e-01 -5.07930517e-01 4.14284766e-02
-9.40083042e-02 1.17895162e+00 4.85349834e-01 -4.53561991e-01
2.46852651e-01 2.46934637e-01 -1.26606333e+00 1.57491505e-01
9.15855110e-01 3.06413770e-01 2.76304148e-02 4.34077382e-01
-7.49086201e-01 3.16598296e-01 -4.92669672e-01 -1.09431744e+00
9.48872697e-03 1.04985893e-01 -1.20380855e+00 -6.09123051e-01
6.20806038e-01 1.68194681e-01 -1.45397291e-01 8.31740499e-01
8.32081378e-01 1.45417154e-01 5.19622743e-01 -1.14957809e+00
-9.05092537e-01 1.23294795e+00 -4.27515090e-01 4.10843551e-01
5.39505899e-01 -2.54116207e-01 1.41055286e+00 -6.07068121e-01
1.26629424e+00 1.16649246e+00 9.50421631e-01 4.55355495e-01
-1.16484427e+00 -6.66885197e-01 6.46186173e-02 2.56112456e-01
-1.25612831e+00 -3.88076991e-01 5.06101966e-01 -2.09297597e-01
1.42346931e+00 -3.13836746e-02 5.27310610e-01 1.01924682e+00
-3.95550698e-01 9.44581151e-01 8.85306835e-01 -6.15540445e-01
-1.30501613e-01 1.01186804e-01 2.63964295e-01 5.28279185e-01
2.84915507e-01 -4.19269860e-01 -4.47426200e-01 -3.92151058e-01
1.80401281e-01 -3.71434182e-01 -8.84774178e-02 -2.10796952e-01
-1.42530072e+00 7.52544999e-01 2.47264430e-01 8.30748200e-01
-1.82526544e-01 2.94696260e-03 5.12949705e-01 1.37481794e-01
7.96039462e-01 9.06230927e-01 -7.88136065e-01 -4.02829885e-01
-1.18431842e+00 6.97658896e-01 1.37106895e+00 9.83355045e-01
4.82831627e-01 -8.83932471e-01 -8.57608486e-03 9.74921644e-01
2.69623399e-01 2.70718664e-01 3.70911390e-01 -7.87488639e-01
8.26335669e-01 8.97171736e-01 1.17741143e-02 -7.54234493e-01
-8.73456478e-01 -1.46115839e-01 -2.96058774e-01 -5.55648625e-01
6.35493457e-01 -1.42720923e-01 -5.33635080e-01 1.73771977e+00
5.51692545e-01 1.21234551e-01 -2.03225086e-03 4.92456883e-01
1.25114202e+00 4.68165547e-01 4.28837955e-01 1.30543888e-01
1.45732319e+00 -9.00743067e-01 -7.06101060e-01 -1.89486057e-01
7.99491525e-01 -9.12946224e-01 1.13044643e+00 -3.00857443e-02
-1.13565695e+00 4.08267528e-02 -6.16011500e-01 -7.48960733e-01
-9.17095184e-01 -4.37775590e-02 6.95113778e-01 4.27899957e-01
-8.52891266e-01 5.74808955e-01 -5.53361237e-01 -1.01225853e+00
2.52698362e-01 3.25474828e-01 -5.21748900e-01 1.76753446e-01
-1.39691329e+00 1.20916116e+00 6.16614580e-01 -2.47032046e-01
-1.00155443e-01 -1.07317030e+00 -7.24991858e-01 -2.18911245e-02
4.82642055e-01 -7.48220682e-01 1.15206647e+00 -2.20596135e-01
-9.11924839e-01 1.36933172e+00 3.19156121e-03 -6.02059603e-01
5.51537752e-01 -3.42399091e-01 -5.01825213e-01 1.25967646e-02
4.59762245e-01 7.42835045e-01 -7.97022134e-02 -9.62753356e-01
-7.66780555e-01 -2.70499229e-01 5.41950166e-02 -1.04385754e-02
-6.61153197e-01 5.84017932e-01 -5.41403174e-01 -7.07745910e-01
-4.25302118e-01 -9.06575978e-01 -2.92542070e-01 -1.42136037e-01
-3.16366822e-01 -5.44445395e-01 4.82363313e-01 -7.78819144e-01
1.48154891e+00 -1.77108002e+00 1.54609740e-01 -4.11075525e-05
1.24036796e-01 2.69003689e-01 -2.12737381e-01 9.11862612e-01
1.53452203e-01 4.13723677e-01 -5.52598596e-01 -7.35953033e-01
3.20269942e-01 2.44490400e-01 -1.90694153e-01 2.55800575e-01
2.51446124e-02 1.08591795e+00 -1.25997198e+00 -9.99275804e-01
-9.90668610e-02 4.77941245e-01 -6.80030107e-01 1.49672087e-02
-3.03368658e-01 1.73117463e-02 -5.74868858e-01 5.45020998e-01
2.38306552e-01 -1.63360640e-01 3.58906388e-01 -2.85057146e-02
-2.27972180e-01 5.08209527e-01 -1.09670746e+00 2.01930118e+00
-2.87180811e-01 7.24693239e-01 4.24360670e-02 -4.44606274e-01
6.51405275e-01 1.22796349e-01 5.56045949e-01 -3.90556782e-01
8.25811550e-03 4.04446214e-01 -1.88175514e-01 -7.79533923e-01
1.03229356e+00 -1.16486721e-01 -5.23850918e-01 4.32731718e-01
3.42636824e-01 -6.69058934e-02 6.47403300e-01 4.39827591e-01
1.30582738e+00 5.08851886e-01 4.50080574e-01 -3.85400653e-01
3.98620427e-01 3.05031061e-01 1.94866017e-01 4.11358654e-01
-1.37358934e-01 3.78236115e-01 2.51241386e-01 -3.99767309e-01
-1.30008078e+00 -7.15566099e-01 -2.38501593e-01 1.28041255e+00
-9.84468088e-02 -6.04084671e-01 -5.50961733e-01 -8.85695398e-01
1.88711852e-01 8.46323550e-01 -7.42708921e-01 5.42532206e-01
-8.03171337e-01 -9.33246732e-01 9.96997416e-01 2.72644401e-01
2.84024656e-01 -1.24922967e+00 -3.92313927e-01 3.67735952e-01
-5.33476353e-01 -1.24501455e+00 -2.38206819e-01 -6.40391707e-02
-4.85923409e-01 -1.03737736e+00 -6.47544920e-01 -7.52730966e-01
2.48963654e-01 -3.55054855e-01 1.64954698e+00 -2.15989966e-02
-3.25845659e-01 2.53471434e-01 -4.07635659e-01 -4.35523689e-01
-3.46840113e-01 1.01844370e+00 -1.24093011e-01 -6.16072297e-01
8.37420821e-01 -5.56995511e-01 -2.02417180e-01 -1.06176689e-01
-1.02617276e+00 -3.89387608e-01 3.39904666e-01 5.56286514e-01
2.56666481e-01 -5.19054949e-01 8.35080028e-01 -1.16845918e+00
8.08313131e-01 -8.46797824e-01 -3.57499570e-01 5.46919763e-01
-4.82215434e-01 3.45225662e-01 3.57745886e-01 -3.39654267e-01
-1.05993891e+00 -5.15953079e-02 -6.50894225e-01 1.91393018e-01
-1.16386339e-01 8.93133819e-01 -7.88787454e-02 2.08696082e-01
6.87063336e-01 -4.48423654e-01 -5.60661256e-01 -9.26417828e-01
8.13489258e-01 7.45450795e-01 7.16351628e-01 -8.94985974e-01
7.28694916e-01 2.79695094e-01 -1.32763550e-01 -7.80643880e-01
-1.06914926e+00 -8.04897487e-01 -8.20694804e-01 -4.66727838e-03
7.49102116e-01 -8.05298686e-01 -6.18314564e-01 4.24405839e-03
-1.02146614e+00 -1.73192471e-01 -4.47692096e-01 1.83470190e-01
-1.44489437e-01 6.02783144e-01 -5.63077629e-01 -5.62985957e-01
-4.98672813e-01 -4.37138498e-01 1.21121204e+00 1.15882345e-01
-4.91172135e-01 -1.28045559e+00 6.26987576e-01 4.37752515e-01
3.54338288e-01 4.81179982e-01 8.32939923e-01 -1.12119865e+00
-2.08022803e-01 -1.59360215e-01 -1.77815661e-01 -3.28590512e-01
-2.68983275e-01 2.04270378e-01 -7.72400558e-01 6.18519448e-03
-6.37343109e-01 -3.78730148e-01 9.82370675e-01 -6.61978200e-02
7.08599031e-01 -1.87259421e-01 -7.02018380e-01 2.96166956e-01
1.45521438e+00 -4.94540125e-01 5.91538787e-01 9.31156635e-01
6.22746050e-01 8.95752192e-01 4.93052810e-01 3.22263896e-01
1.10106444e+00 5.55775702e-01 1.58571750e-01 1.07175231e-01
-9.74162761e-03 -5.10150135e-01 -2.64360681e-02 7.41355002e-01
-8.09425041e-02 -4.66718316e-01 -1.28217232e+00 1.23685563e+00
-1.92354929e+00 -1.07044089e+00 -2.11390033e-01 1.73611653e+00
1.39091039e+00 1.43906632e-02 2.14104220e-01 -1.92788631e-01
5.39894044e-01 -1.37080380e-04 -1.15183547e-01 -4.38947007e-02
-3.71692955e-01 3.17453980e-01 6.66332483e-01 3.79807055e-01
-1.23555946e+00 1.22173679e+00 6.65853977e+00 7.96498954e-01
-4.90593255e-01 1.10127360e-01 9.51135680e-02 -1.85612246e-01
-5.44324577e-01 1.14951260e-01 -1.24398899e+00 4.68498528e-01
1.18715858e+00 -2.58472681e-01 6.90803826e-02 6.95386291e-01
-1.78583458e-01 -1.01111002e-01 -1.10343385e+00 5.77926636e-01
1.19114399e-01 -1.32117867e+00 -2.58597642e-01 8.76270160e-02
6.45628989e-01 3.77914429e-01 -3.15026611e-01 6.15783036e-01
7.43896008e-01 -8.37867558e-01 7.55252302e-01 5.35289884e-01
5.08502185e-01 -5.24009645e-01 7.90304422e-01 3.28779459e-01
-9.47236121e-01 1.04883373e-01 -2.08352268e-01 1.67917386e-01
5.47428310e-01 6.07910633e-01 -8.12757313e-01 6.14246368e-01
9.16094542e-01 6.97567225e-01 -9.94188786e-01 1.29576814e+00
-2.59193927e-01 6.44451320e-01 -7.05840588e-01 -2.59412766e-01
3.58536839e-01 1.30924836e-01 6.45763159e-01 1.62446415e+00
2.31009081e-01 1.69076607e-01 -2.16868017e-02 6.25137389e-01
-5.87730110e-01 4.11423236e-01 -3.17826480e-01 -2.22204432e-01
6.73017919e-01 1.54784131e+00 -7.08105803e-01 -2.84795761e-01
-4.13945019e-01 6.31657243e-01 9.26270604e-01 -7.56104589e-02
-7.93939888e-01 -6.41554892e-01 3.55767161e-01 2.41933480e-01
1.81762904e-01 -2.47788072e-01 -1.22805923e-01 -1.26806509e+00
-1.22774310e-01 -5.58951974e-01 8.45404863e-01 -5.93707085e-01
-1.78634393e+00 5.36577821e-01 2.71119416e-01 -6.91092193e-01
-3.39050204e-01 -3.85256618e-01 -1.35373352e-02 5.90566993e-01
-1.77745032e+00 -1.65092576e+00 -2.72165406e-02 2.73378611e-01
8.56944844e-02 4.88740578e-02 8.10746789e-01 8.51091027e-01
-3.03726524e-01 5.58375001e-01 3.23913485e-01 3.57613564e-01
1.27291739e+00 -1.29526281e+00 6.92259669e-01 6.25038862e-01
1.73387676e-01 7.71588027e-01 8.03427756e-01 -8.32491159e-01
-1.02369261e+00 -8.25953662e-01 1.82228243e+00 -9.79061961e-01
1.19714475e+00 -4.67326522e-01 -9.33027089e-01 8.94919872e-01
5.14661193e-01 -3.42989296e-01 9.41123605e-01 4.90905434e-01
-3.35856438e-01 3.79693121e-01 -1.41751099e+00 5.01570761e-01
1.28171980e+00 -2.28479370e-01 -1.30454862e+00 2.31084839e-01
5.50798833e-01 -3.68210286e-01 -1.27860904e+00 2.72884130e-01
4.46572512e-01 -4.92439061e-01 9.17457700e-01 -7.07981527e-01
6.17064595e-01 -2.09192559e-01 6.19530082e-02 -1.12006938e+00
-8.32986236e-02 -5.82254767e-01 -4.99778651e-02 1.80304027e+00
7.85003662e-01 -5.05601406e-01 6.98520124e-01 8.02596450e-01
-1.63090661e-01 -4.61496055e-01 -8.86165142e-01 -5.90438783e-01
4.13948596e-01 -3.02432179e-01 7.23079681e-01 1.30339622e+00
2.26636276e-01 5.32658219e-01 -9.08859223e-02 -1.75881505e-01
5.06024063e-01 1.34133354e-01 6.31100893e-01 -1.60518026e+00
4.03321683e-02 -4.51958954e-01 -1.52173683e-01 -6.50104940e-01
4.99736875e-01 -1.11844778e+00 -1.28620967e-01 -2.04035568e+00
2.32160285e-01 -6.68531477e-01 8.00067782e-02 7.06593752e-01
-4.07415360e-01 3.84366244e-01 1.39447540e-01 1.43006235e-01
-7.87925124e-01 4.14456844e-01 5.25508761e-01 3.49770114e-02
-1.41129017e-01 -6.24647856e-01 -7.00432003e-01 6.29602611e-01
4.71627891e-01 -5.94692528e-01 1.83778048e-01 -6.01346791e-01
8.51225972e-01 -5.66826224e-01 6.81213737e-02 -7.49586344e-01
3.88992548e-01 -4.20446433e-02 1.63362727e-01 -7.00290561e-01
1.00587368e-01 -8.57556343e-01 9.86281503e-03 2.90639861e-03
-5.43285608e-01 1.96078703e-01 2.66593277e-01 4.43527222e-01
-1.16965011e-01 -5.00024796e-01 4.59277004e-01 -1.83526099e-01
-8.20305884e-01 1.30014494e-01 -1.73215866e-02 7.83983409e-01
9.62146223e-01 2.15166494e-01 -6.09256029e-01 -4.79418971e-02
-5.40972054e-01 6.23328686e-01 6.33312762e-01 6.03508890e-01
6.85466751e-02 -1.12286472e+00 -1.02854347e+00 -5.27959168e-01
3.33801031e-01 -1.37388632e-01 3.96589078e-02 4.89793032e-01
-5.35113037e-01 3.35035920e-01 -8.50462988e-02 6.30680798e-03
-1.25165117e+00 6.48082912e-01 9.72503051e-03 -7.84748197e-01
-4.99055445e-01 9.14885640e-01 -5.72489321e-01 -8.41980577e-01
7.89038911e-02 -1.03174798e-01 -5.21182597e-01 5.62114418e-01
3.36074352e-01 4.29418027e-01 -7.42969513e-02 -7.31839418e-01
-5.84280670e-01 4.34778571e-01 1.57517657e-01 -4.17977810e-01
1.94685912e+00 -2.87762731e-01 -3.38538885e-01 3.56784433e-01
1.09907544e+00 3.08357567e-01 -5.77257156e-01 -4.59521562e-01
6.92424655e-01 -1.58298969e-01 -2.56444603e-01 -9.93283153e-01
-7.40554750e-01 5.37999570e-01 6.67192787e-02 2.27171987e-01
8.54177117e-01 4.28437531e-01 9.51530993e-01 5.69328845e-01
2.45390654e-01 -1.31684554e+00 -3.94954443e-01 7.83554673e-01
5.83390117e-01 -1.11255407e+00 3.71974260e-01 -2.75158703e-01
-5.30883551e-01 9.08159256e-01 2.62466609e-01 1.36462031e-02
6.01772189e-01 3.73655170e-01 -1.95427835e-01 -4.55519736e-01
-5.93889654e-01 -5.51248312e-01 2.75593519e-01 3.98503393e-01
7.23413110e-01 -2.70335346e-01 -7.15523064e-01 6.23549700e-01
-6.39906049e-01 1.11673221e-01 4.13923442e-01 1.03640330e+00
-2.31610358e-01 -1.56671786e+00 -3.10043599e-02 3.46286535e-01
-9.16145623e-01 -4.09717798e-01 -6.59004748e-01 1.07809246e+00
1.19735427e-01 6.40644729e-01 -3.75942290e-02 1.87368616e-01
4.73282397e-01 2.37372875e-01 4.65329885e-01 -6.58011019e-01
-1.22807598e+00 -3.48020524e-01 6.58004105e-01 -5.56430109e-02
-8.83478940e-01 -9.64785695e-01 -1.32454169e+00 -5.24056554e-01
-1.87073618e-01 2.67344594e-01 7.88261712e-01 8.24902654e-01
4.81233329e-01 1.20337598e-01 -2.09689289e-01 -5.80199480e-01
-7.83480182e-02 -1.06146109e+00 -2.74812639e-01 5.22387564e-01
-6.63815066e-02 -5.38800538e-01 -1.51373968e-01 6.79614916e-02] | [9.556353569030762, 8.99084758758545] |
17d50b54-e0ae-47ca-91d8-b540758942a7 | a-dataset-free-self-supervised-disentangled | 2112.02869 | null | https://arxiv.org/abs/2112.02869v4 | https://arxiv.org/pdf/2112.02869v4.pdf | Physics Driven Deep Retinex Fusion for Adaptive Infrared and Visible Image Fusion | Convolutional neural networks have turned into an illustrious tool for image fusion and super-resolution. However, their excellent performance cannot work without large fixed-paired datasets; and additionally, these high-demanded ground truth data always cannot be obtained easily in fusion tasks. In this study, we show that, the structures of generative networks capture a great deal of image feature priors, and then these priors are sufficient to reconstruct high-quality fused super-resolution result using only low-resolution inputs. By this way, we propose a novel self-supervised dataset-free method for adaptive infrared (IR) and visible (VIS) image super-resolution fusion named Deep Retinex Fusion (DRF). The key idea of DRF is first generating component priors which are disentangled from physical model using our designed generative networks ZipperNet, LightingNet and AdjustingNet, then combining these priors which captured by networks via adaptive fusion loss functions based on Retinex theory, and finally reconstructing the super-resolution fusion results. Furthermore, in order to verify the effectiveness of our reported DRF, both qualitative and quantitative experiments via comparing with other state-of-the-art methods are performed using different test sets. These results prove that, comparing with large datasets trained methods, DRF which works without any dataset achieves the best super-resolution fusion performance; and more importantly, DRF can adaptively balance IR and VIS information and has good noise immunity. DRF codes are open source available at https://github.com/GuYuanjie/Deep-Retinex-fusion. | ['Liang Xue', 'Cheng Liu', 'Haoran Dai', 'Yinghan Guan', 'Shouyu Wang', 'Zhibo Xiao', 'Yuanjie Gu'] | 2021-12-06 | null | null | null | null | ['infrared-and-visible-image-fusion'] | ['computer-vision'] | [ 1.40360788e-01 -3.88051897e-01 1.02044351e-01 -8.33996087e-02
-8.11640799e-01 -2.60655701e-01 6.15970075e-01 -6.21729136e-01
-2.95639969e-02 9.83174622e-01 3.94538969e-01 2.51167446e-01
-4.42519456e-01 -1.14227438e+00 -7.23886073e-01 -1.02882409e+00
3.67803782e-01 1.73601359e-02 5.72083779e-02 -4.97666329e-01
-1.88687786e-01 4.12221849e-01 -1.80291021e+00 1.81475922e-01
1.29006517e+00 1.03479099e+00 4.32271898e-01 4.50176686e-01
-1.25923520e-02 6.04250848e-01 -2.74107546e-01 -2.69517303e-01
5.08976519e-01 -5.68052530e-01 -3.31016958e-01 -2.53297687e-01
4.69679743e-01 -5.05610824e-01 -5.37419260e-01 1.32598531e+00
7.77652979e-01 1.49445623e-01 5.76425433e-01 -8.29984665e-01
-1.23027897e+00 6.02646112e-01 -7.88006425e-01 2.16730013e-01
2.55958825e-01 2.12141141e-01 5.89706123e-01 -7.64980972e-01
3.40092838e-01 1.25345409e+00 6.35404766e-01 5.35541236e-01
-1.37794662e+00 -1.02124286e+00 -2.59828746e-01 -1.65932164e-01
-1.36073720e+00 -4.71480519e-01 8.38134587e-01 -2.16087058e-01
3.06286365e-01 2.78769761e-01 4.89875853e-01 1.24124074e+00
2.91990131e-01 1.68761387e-01 1.48546851e+00 -1.00492798e-01
-1.31194517e-01 1.27738833e-01 -1.10442996e-01 6.20567858e-01
2.61951774e-01 5.39405704e-01 -5.10268867e-01 1.81028485e-01
1.32373118e+00 2.16782823e-01 -6.19057059e-01 2.17788413e-01
-1.19648778e+00 4.04070139e-01 8.02365422e-01 3.81350517e-01
-3.86036128e-01 -3.46754789e-02 -1.68361858e-01 1.64724112e-01
6.63879097e-01 4.46557999e-02 -6.44081905e-02 4.79962319e-01
-8.03824902e-01 4.98222858e-02 3.66662815e-02 7.12635458e-01
9.58944142e-01 2.13170037e-01 -3.29566181e-01 9.44414616e-01
3.72337610e-01 7.99552202e-01 2.17060640e-01 -9.04626608e-01
1.66699633e-01 3.90907377e-01 2.67775744e-01 -9.91193831e-01
-3.24630320e-01 -6.68418765e-01 -1.49384201e+00 5.44695795e-01
1.08387619e-01 -1.04082227e-01 -9.13813174e-01 1.87517989e+00
2.83799231e-01 3.82789195e-01 2.99754679e-01 1.14575183e+00
1.31131697e+00 6.68819427e-01 -8.44428837e-02 -4.38010782e-01
1.55127037e+00 -5.03722310e-01 -8.51277888e-01 6.17179647e-02
-3.46980870e-01 -9.65446889e-01 1.00920534e+00 4.32163924e-01
-1.22305298e+00 -1.06861424e+00 -1.09393191e+00 -2.39055231e-01
-8.64098296e-02 1.88981026e-01 5.77290297e-01 3.71525824e-01
-1.11852801e+00 5.82669318e-01 -4.42246258e-01 -9.92106274e-02
4.89305407e-01 2.82514771e-03 -3.29215497e-01 -1.72908932e-01
-1.43343747e+00 7.60287881e-01 4.56441134e-01 2.53957987e-01
-8.38412881e-01 -8.55503082e-01 -5.65662384e-01 -7.03546852e-02
2.51712978e-01 -1.14315140e+00 7.39142001e-01 -7.58620918e-01
-1.58541882e+00 5.71254671e-01 9.54642426e-03 -1.19021423e-01
4.40399498e-01 -2.34456435e-01 -7.26673841e-01 1.29699364e-01
-8.66945758e-02 4.28987086e-01 9.36859012e-01 -1.77822220e+00
-5.07163763e-01 -3.71747911e-01 6.92812353e-02 2.28614613e-01
-6.32494316e-02 1.36105381e-02 -1.65604472e-01 -5.20207167e-01
1.92098051e-01 -2.39118859e-01 -6.40950259e-03 -1.75260920e-02
-2.84082532e-01 2.70205233e-02 6.59681380e-01 -8.48147213e-01
1.05398190e+00 -2.09156585e+00 1.27084285e-01 -8.39626193e-02
4.93680120e-01 3.36667955e-01 -3.00456565e-02 1.27385668e-02
-1.13796525e-01 2.52390653e-01 -2.16716677e-01 -2.31246710e-01
-1.71056837e-01 -2.46142745e-02 -4.04107153e-01 3.67084086e-01
4.04573604e-02 8.46582115e-01 -8.11707616e-01 -5.02991796e-01
5.41109264e-01 1.13707781e+00 -1.08454049e-01 2.22089678e-01
-1.07076682e-01 9.04412210e-01 -4.51928377e-01 5.05545676e-01
1.16354716e+00 -2.83397853e-01 -9.78812724e-02 -8.71159494e-01
-2.87199348e-01 -3.03259790e-01 -1.27663839e+00 1.75588524e+00
-5.21656930e-01 7.58120939e-02 1.98243409e-01 -6.04904532e-01
1.24476242e+00 2.47659937e-01 4.93361026e-01 -1.16022480e+00
2.62471288e-01 5.83809428e-02 -4.78965342e-01 -2.65756756e-01
4.15787458e-01 -3.72936845e-01 1.78634122e-01 2.70616651e-01
7.50768259e-02 2.82450020e-02 -1.33210689e-01 9.76719558e-02
6.17880106e-01 3.37160081e-01 5.58854453e-02 -1.52899548e-01
6.51801646e-01 -4.54170763e-01 5.67191243e-01 5.75501263e-01
7.71844164e-02 9.63608623e-01 2.58163828e-02 -1.85673311e-01
-1.19074464e+00 -1.32252240e+00 -3.11416805e-01 5.73616683e-01
4.72179681e-01 -1.50811700e-02 -6.40821040e-01 -9.56298187e-02
-3.05656374e-01 4.46977377e-01 -5.79342127e-01 -6.32513911e-02
-3.21307272e-01 -9.61445630e-01 3.67616147e-01 3.09308857e-01
1.23020256e+00 -7.66835809e-01 -1.48152947e-01 -1.00016117e-01
-4.34131056e-01 -1.09798443e+00 -1.68301642e-01 -3.73603016e-01
-5.36563158e-01 -9.76784348e-01 -7.45948493e-01 -2.80266106e-01
3.30286622e-01 6.63967013e-01 1.12296879e+00 7.17110634e-02
-2.49166161e-01 9.00884569e-02 -2.24677756e-01 -9.30954739e-02
-3.27643484e-01 -3.42166096e-01 -5.87608404e-02 1.48413941e-01
-6.20636232e-02 -9.41538393e-01 -9.62488234e-01 2.43369237e-01
-9.90215957e-01 3.82344425e-01 6.71139777e-01 6.12859368e-01
9.15841341e-01 4.53525573e-01 4.83670264e-01 -4.34172541e-01
4.79395032e-01 -4.52984691e-01 -7.56015599e-01 3.22553039e-01
-4.91554588e-01 -1.53106858e-03 7.02643037e-01 -1.75093308e-01
-1.72606516e+00 -3.44778687e-01 -6.32470548e-02 -6.11478209e-01
-1.65906876e-01 -1.62922516e-02 -3.84946823e-01 -1.07823715e-01
8.56881797e-01 2.73866653e-01 6.07581399e-02 -6.67209387e-01
5.73192537e-01 5.16081393e-01 8.61201048e-01 -5.03419399e-01
1.06766820e+00 7.90494800e-01 6.84226155e-02 -4.74955678e-01
-1.00106311e+00 -7.44182020e-02 -3.56650680e-01 -3.03467095e-01
1.11969721e+00 -1.18458188e+00 -8.16154420e-01 6.04808986e-01
-9.39806104e-01 -7.96687827e-02 -1.86234176e-01 5.11133552e-01
-4.51952189e-01 3.27674299e-01 -5.18925130e-01 -9.10394609e-01
-6.01888776e-01 -9.59548116e-01 1.05081213e+00 8.02448034e-01
6.03914618e-01 -8.43780160e-01 -1.62119381e-02 6.14392579e-01
7.99200177e-01 5.85562766e-01 4.47693288e-01 2.00566798e-01
-8.52176905e-01 4.45967048e-01 -7.34537601e-01 4.97943670e-01
3.25759560e-01 6.03173859e-02 -1.10323465e+00 -1.74043626e-01
1.11303650e-01 -9.79003087e-02 1.06800771e+00 5.08007228e-01
1.16394866e+00 -1.09959893e-01 -6.74058422e-02 1.10242343e+00
1.95059764e+00 -1.22619100e-01 1.11982238e+00 3.35120782e-02
7.05377817e-01 1.60455853e-01 4.44739968e-01 3.43075573e-01
2.87180305e-01 5.77049792e-01 5.38193524e-01 -5.45409381e-01
-6.59065187e-01 -2.63357908e-01 1.52506053e-01 4.33055460e-01
-7.35454619e-01 -3.09149414e-01 -3.98164600e-01 1.35517389e-01
-1.78726733e+00 -1.24656177e+00 -2.63155758e-01 2.32327485e+00
9.66049492e-01 -2.56750375e-01 -9.72621217e-02 -7.70175904e-02
8.27147186e-01 2.52825230e-01 -5.16941249e-01 3.00479710e-01
-6.16702557e-01 4.57919896e-01 4.80762899e-01 5.40805280e-01
-9.62204635e-01 7.09838569e-01 5.68815804e+00 1.08554316e+00
-1.11285281e+00 4.50571656e-01 5.53451657e-01 -7.56806415e-03
-6.80057764e-01 -1.74473301e-01 -7.31578588e-01 5.00245392e-01
5.72985291e-01 1.61484648e-02 8.07983041e-01 2.02775195e-01
3.13073784e-01 -5.18971086e-02 -4.37906235e-01 1.19222462e+00
-5.11526205e-02 -1.26572955e+00 2.71601584e-02 -9.81334001e-02
8.29340696e-01 -5.71754016e-03 2.39430845e-01 -3.82238589e-02
5.57516038e-01 -1.26414168e+00 3.62421274e-01 1.24067545e+00
1.14556265e+00 -7.44803786e-01 6.80730402e-01 1.87179551e-01
-1.23783600e+00 -5.91547936e-02 -6.29768431e-01 3.44503045e-01
1.96887881e-01 9.82329071e-01 8.12519342e-02 1.28803587e+00
8.78022194e-01 6.23492539e-01 -3.34030867e-01 6.72111452e-01
-3.86019677e-01 2.49363229e-01 -1.39315903e-01 6.34912550e-01
-3.09595466e-01 -4.83920872e-01 5.17633855e-01 8.64942551e-01
4.98168647e-01 4.00302768e-01 -6.39268532e-02 1.44548130e+00
7.18546435e-02 -1.73418194e-01 -3.95059198e-01 3.01077902e-01
4.29835290e-01 1.60224390e+00 -4.41113591e-01 -1.73163742e-01
-3.38672847e-01 6.43441319e-01 2.91204993e-02 6.34467483e-01
-1.13975799e+00 -1.89508665e-02 6.70310557e-01 1.32261366e-01
5.06842602e-03 -5.85449971e-02 -3.65843356e-01 -1.36337769e+00
-5.60977496e-02 -7.32115984e-01 1.51415154e-01 -1.35390902e+00
-1.50270557e+00 7.07056642e-01 1.09039448e-01 -1.23616111e+00
2.74305165e-01 -3.40833992e-01 -4.11875874e-01 1.34222353e+00
-1.79360867e+00 -1.48786402e+00 -1.07709706e+00 8.51492643e-01
9.54366848e-02 -1.14343487e-01 5.40237069e-01 3.89106363e-01
-5.32049298e-01 3.08491021e-01 2.86853194e-01 -1.06868424e-01
7.82743335e-01 -8.66968870e-01 -6.30416051e-02 1.00889611e+00
-1.60003737e-01 5.34670770e-01 6.01436317e-01 -6.23659313e-01
-1.27204525e+00 -9.71967697e-01 3.92952189e-03 -1.99157342e-01
3.07479858e-01 -1.71302289e-01 -1.08836639e+00 2.46052340e-01
3.77507061e-01 7.75836185e-02 2.81148314e-01 -1.40932813e-01
-5.36036253e-01 -5.50543845e-01 -1.43378234e+00 3.55978876e-01
1.32540989e+00 -4.28384364e-01 -2.66863942e-01 2.82031864e-01
9.89576399e-01 -4.59184140e-01 -1.17275476e+00 6.60865903e-01
4.75694805e-01 -1.42494047e+00 1.33038175e+00 2.44991314e-02
7.63223946e-01 -7.10133791e-01 -3.30102116e-01 -1.16570616e+00
-7.44495213e-01 -4.31436181e-01 -9.93871223e-03 1.42780960e+00
2.44913511e-02 -8.86192441e-01 1.58123225e-01 2.05274642e-01
-5.14221191e-02 -4.22921509e-01 -6.15756035e-01 -5.05221546e-01
-5.51014841e-02 2.17203125e-02 9.70038652e-01 9.22691107e-01
-8.32535565e-01 3.03818494e-01 -5.83759069e-01 4.22601461e-01
1.11941338e+00 3.68527658e-02 6.37630165e-01 -1.18700671e+00
-2.60960907e-01 -5.03098667e-01 -5.76561689e-02 -5.08287251e-01
-1.69269443e-01 -6.50678933e-01 -2.21752957e-01 -1.82142222e+00
4.21659559e-01 -4.66664523e-01 -4.38774526e-01 3.93152148e-01
-2.58197486e-01 5.92679083e-01 1.09119058e-01 3.80011886e-01
-2.85628706e-01 6.65606797e-01 1.71870470e+00 8.39911252e-02
-1.42390775e-02 -2.75838345e-01 -1.11999929e+00 4.28657830e-01
7.90937245e-01 3.99343222e-02 -5.77983320e-01 -3.64120781e-01
3.04736406e-01 1.76981792e-01 8.04572999e-01 -1.01425064e+00
7.15446696e-02 -3.04971993e-01 7.63640881e-01 -4.95716453e-01
3.56407702e-01 -6.53048992e-01 7.94647098e-01 -6.64817765e-02
8.20179954e-02 -5.62583983e-01 1.68848932e-01 3.93016547e-01
-1.18455745e-01 1.86258480e-01 1.11589217e+00 -2.37206131e-01
-5.77345788e-01 5.01067817e-01 4.57397938e-01 -1.05143063e-01
7.30388641e-01 -1.49819613e-01 -8.51585746e-01 -3.00416291e-01
-4.96241063e-01 -4.31407653e-02 6.76699638e-01 3.01424056e-01
5.62272131e-01 -1.44978094e+00 -1.06318247e+00 2.07762614e-01
-2.35902131e-01 2.77003348e-01 9.61256683e-01 7.57266045e-01
-2.22528636e-01 -1.54280439e-01 -4.32176322e-01 -5.24804175e-01
-1.20121360e+00 7.33957887e-01 5.18792331e-01 -1.07825175e-01
-7.42411971e-01 6.41012132e-01 5.44854581e-01 -2.96577483e-01
-2.68276721e-01 -3.96330617e-02 -3.51949722e-01 -2.00951949e-01
7.83681214e-01 3.16860884e-01 -1.67678908e-01 -5.93697608e-01
-2.55814046e-02 8.99104059e-01 2.87133276e-01 3.37289311e-02
1.37749076e+00 -3.14256489e-01 -2.94591099e-01 8.81187022e-02
8.34807694e-01 8.13616365e-02 -1.36239135e+00 -5.21859586e-01
-1.04241908e+00 -6.34462297e-01 3.22823912e-01 -9.88672197e-01
-1.39744151e+00 7.55020797e-01 9.39434886e-01 2.11012214e-01
1.72558594e+00 -5.45885935e-02 6.31370962e-01 -2.84057677e-01
4.27452505e-01 -6.77085400e-01 5.17487451e-02 1.24188334e-01
1.02186203e+00 -1.13812232e+00 2.18818232e-01 -4.91514564e-01
-3.82505745e-01 8.01843882e-01 5.54969668e-01 -1.32292792e-01
5.57709098e-01 2.88079947e-01 -9.78033245e-02 -1.50589064e-01
-4.80044872e-01 -3.93132240e-01 3.82735103e-01 6.97804391e-01
3.48369747e-01 9.30652469e-02 -2.28658617e-01 7.09556460e-01
-2.34235987e-01 2.41742745e-01 2.12840408e-01 2.48687536e-01
-5.40418804e-01 -8.09924245e-01 -6.24833882e-01 2.20309138e-01
-3.23502123e-01 -1.37368813e-01 2.19537690e-02 5.69979191e-01
4.77024287e-01 8.63947928e-01 -1.10660471e-01 -6.16690874e-01
1.70833141e-01 -3.82685184e-01 6.49573147e-01 -6.88285828e-02
-2.89246798e-01 1.88446939e-01 -2.62144595e-01 -6.19699657e-01
-8.44831944e-01 -1.54199436e-01 -1.00568759e+00 -6.21975720e-01
-2.52185404e-01 -1.22431450e-01 5.26361883e-01 7.33940661e-01
4.56150860e-01 8.82381022e-01 6.95418477e-01 -8.64920735e-01
-2.57569641e-01 -9.83806014e-01 -6.04013503e-01 2.87705958e-01
2.84178317e-01 -7.99049020e-01 -5.23392856e-01 -9.78273079e-02] | [10.883707046508789, -2.0423500537872314] |
c1820e3f-15fa-44ce-8ba4-74f4e65f8b32 | investigating-the-lombard-effect-influence-on | 1906.02112 | null | https://arxiv.org/abs/1906.02112v4 | https://arxiv.org/pdf/1906.02112v4.pdf | Investigating the Lombard Effect Influence on End-to-End Audio-Visual Speech Recognition | Several audio-visual speech recognition models have been recently proposed which aim to improve the robustness over audio-only models in the presence of noise. However, almost all of them ignore the impact of the Lombard effect, i.e., the change in speaking style in noisy environments which aims to make speech more intelligible and affects both the acoustic characteristics of speech and the lip movements. In this paper, we investigate the impact of the Lombard effect in audio-visual speech recognition. To the best of our knowledge, this is the first work which does so using end-to-end deep architectures and presents results on unseen speakers. Our results show that properly modelling Lombard speech is always beneficial. Even if a relatively small amount of Lombard speech is added to the training set then the performance in a real scenario, where noisy Lombard speech is present, can be significantly improved. We also show that the standard approach followed in the literature, where a model is trained and tested on noisy plain speech, provides a correct estimate of the video-only performance and slightly underestimates the audio-visual performance. In case of audio-only approaches, performance is overestimated for SNRs higher than -3dB and underestimated for lower SNRs. | ['Pingchuan Ma', 'Maja Pantic', 'Stavros Petridis'] | 2019-06-05 | null | null | null | null | ['audio-visual-speech-recognition'] | ['speech'] | [ 5.01932502e-02 7.00928718e-02 3.29032511e-01 -7.78438747e-02
-7.24820435e-01 -3.97058249e-01 6.11642420e-01 7.37701654e-02
-4.52867001e-01 5.09294271e-01 3.92442644e-01 -3.63351226e-01
1.51154429e-01 -2.86610335e-01 -8.70824695e-01 -8.77928138e-01
1.40284076e-01 3.09931934e-02 4.37815756e-01 -1.33077770e-01
8.27300921e-03 4.28019404e-01 -2.05839109e+00 3.66943717e-01
3.07112455e-01 7.83666492e-01 5.62388003e-01 1.12054634e+00
4.12626415e-02 6.09699667e-01 -1.01855445e+00 -2.15647444e-01
5.46967089e-02 -3.76238823e-01 -4.47246373e-01 9.71491784e-02
5.01198351e-01 -3.81144494e-01 -3.87902588e-01 1.06082261e+00
8.70685697e-01 -5.09619489e-02 4.18710023e-01 -8.99765432e-01
-9.67534855e-02 4.80452001e-01 -1.24734253e-01 2.23932907e-01
4.40201521e-01 2.99154907e-01 6.06755912e-01 -7.96980321e-01
3.59344393e-01 1.35503244e+00 4.54471678e-01 4.67429727e-01
-1.22016048e+00 -4.45253223e-01 5.40939644e-02 3.82169247e-01
-1.18154788e+00 -1.05732024e+00 7.66745269e-01 -2.35459194e-01
9.64183390e-01 4.24157053e-01 5.43985128e-01 1.21704292e+00
-2.07709312e-01 7.99149215e-01 8.73866260e-01 -6.84254169e-01
1.48779616e-01 4.30272490e-01 -2.91239947e-01 -5.90432771e-02
-1.55424207e-01 4.47785318e-01 -4.64530885e-01 1.39379188e-01
9.04838294e-02 -6.56301975e-01 -6.37702703e-01 -1.59716293e-01
-9.87780452e-01 3.38070095e-01 1.20313242e-01 7.00624049e-01
-2.36495316e-01 1.12219602e-01 5.59068143e-01 4.68060195e-01
2.93973356e-01 -1.37907648e-02 -2.10142344e-01 -5.01186728e-01
-1.24188161e+00 5.42937927e-02 8.23209882e-01 4.31286544e-01
9.68411639e-02 4.94027048e-01 -1.02416994e-02 1.14132071e+00
4.37038749e-01 5.06284356e-01 6.33865178e-01 -7.28071392e-01
3.99863750e-01 -2.28657201e-01 1.76131740e-01 -6.68455005e-01
-2.07824066e-01 -7.25744605e-01 -3.24789792e-01 6.57688677e-01
6.11032128e-01 -7.69040659e-02 -9.71293151e-01 1.66344678e+00
8.62570927e-02 1.26595780e-01 3.19295764e-01 1.04561627e+00
9.30247247e-01 5.94311893e-01 -1.89368844e-01 -3.72811705e-01
1.13422513e+00 -7.14474678e-01 -1.02204478e+00 -1.50327057e-01
2.28918850e-01 -1.38237548e+00 1.10436368e+00 7.10905313e-01
-1.07172596e+00 -6.34737909e-01 -1.18223512e+00 3.01804483e-01
-2.06358820e-01 1.37021288e-01 -7.59046823e-02 1.22665751e+00
-1.30273569e+00 3.80162627e-01 -5.12729406e-01 -3.82573694e-01
-4.81288955e-02 2.36879408e-01 -3.44174564e-01 -5.36877625e-02
-1.11262226e+00 9.19334292e-01 1.81679606e-01 1.74475446e-01
-8.84632945e-01 -2.71171033e-01 -7.55043685e-01 3.91752794e-02
5.41669309e-01 -2.34711975e-01 1.50880682e+00 -1.23722088e+00
-1.77055979e+00 4.54004169e-01 -4.08313096e-01 -5.83236337e-01
7.94894814e-01 -1.39177278e-01 -5.88872075e-01 1.26687646e-01
-4.92294431e-01 5.09804845e-01 9.82408941e-01 -1.48566151e+00
-2.08432361e-01 -1.19160719e-01 -1.52540773e-01 2.17178613e-01
8.43902305e-02 7.72972926e-02 -4.55234110e-01 -7.32894719e-01
-1.85640976e-01 -6.95989728e-01 3.85620654e-01 -1.09863520e-01
-2.24123299e-01 2.37651035e-01 1.10214484e+00 -8.15750420e-01
8.32919240e-01 -2.50772977e+00 -2.24653572e-01 -4.40353937e-02
-2.25239396e-01 8.60987842e-01 -1.57226115e-01 4.69876468e-01
-1.78805098e-01 3.69349658e-03 4.71788280e-06 -5.40780663e-01
-9.74897668e-02 1.34832010e-01 -6.36691973e-02 5.11418045e-01
-8.95363688e-02 4.94525880e-01 -6.00530744e-01 -2.03268319e-01
6.38563156e-01 9.61309195e-01 -3.00612926e-01 1.34167382e-02
1.97401866e-01 2.98359841e-01 3.24714303e-01 3.34968299e-01
7.84459651e-01 5.30476511e-01 -2.29113981e-01 -8.87350962e-02
-2.84321874e-01 5.11148751e-01 -1.40166140e+00 1.05904472e+00
-6.84670091e-01 1.15624952e+00 6.02558732e-01 -8.65155756e-01
8.45688760e-01 7.63948321e-01 -4.75994274e-02 -8.33539963e-01
1.10688590e-01 3.77640337e-01 4.29426908e-01 -5.72573662e-01
2.12221757e-01 -3.61760557e-01 6.69421732e-01 -1.27131507e-01
-9.13047269e-02 -1.65332153e-01 -1.11583687e-01 -8.87314156e-02
8.20192456e-01 -1.77095056e-01 2.14720577e-01 5.99455312e-02
7.39219785e-01 -7.41400003e-01 6.38735220e-02 7.62350976e-01
-3.57140094e-01 8.81226599e-01 3.95313829e-01 1.25889271e-01
-1.01157737e+00 -1.06782293e+00 -2.68203050e-01 7.81990349e-01
-1.23172611e-01 -3.08489889e-01 -8.57517719e-01 -2.10164815e-01
-3.08625102e-01 8.95868421e-01 -2.84285247e-01 -1.12751171e-01
-2.26944149e-01 -6.01168036e-01 7.00464308e-01 1.98531508e-01
3.44823927e-01 -1.04100215e+00 -3.80973607e-01 2.21243307e-01
-2.21590176e-01 -1.22429252e+00 -2.07010001e-01 -2.92630680e-02
-5.88122308e-01 -5.92135131e-01 -1.05328035e+00 -5.27458429e-01
1.43642664e-01 1.51185945e-01 7.24196434e-01 -1.00548357e-01
-7.35131055e-02 4.34893698e-01 -3.84602875e-01 -5.74536264e-01
-8.85798693e-01 -4.09441233e-01 -9.85188186e-02 1.93266958e-01
1.09139472e-01 -4.95893717e-01 -4.33657765e-01 4.77724940e-01
-7.34851301e-01 -2.04612389e-01 5.34867942e-01 8.00713181e-01
8.75410736e-02 1.53501466e-01 7.17757225e-01 -2.12948561e-01
4.42072660e-01 -2.27100104e-02 -3.40660691e-01 -1.99251667e-01
-2.40425870e-01 -8.38900581e-02 6.19459331e-01 -6.20450854e-01
-1.00601017e+00 -3.74518819e-02 -8.07791114e-01 -3.97397459e-01
-7.27117062e-01 1.23888686e-01 -4.64476347e-01 -9.68375206e-02
4.89823997e-01 3.30634981e-01 4.91859317e-02 -6.56919479e-01
-2.39587650e-02 1.15098345e+00 4.96333271e-01 1.25396937e-01
4.31895316e-01 4.81400937e-01 -1.61250934e-01 -1.54174757e+00
-2.20638905e-02 -5.54427207e-01 -1.44011170e-01 -5.00033021e-01
4.96551007e-01 -7.18614578e-01 -6.30728841e-01 7.26154745e-01
-1.07444239e+00 -2.87909865e-01 -1.51740789e-01 8.32408249e-01
-5.27781308e-01 7.00396240e-01 -2.46391803e-01 -1.46269929e+00
1.34887584e-02 -1.59076190e+00 9.36488271e-01 -8.17114115e-02
-7.48852119e-02 -6.29875422e-01 -2.39589170e-01 3.13628137e-01
6.10918701e-01 -2.35187396e-01 5.38029909e-01 -5.88745713e-01
-1.39821067e-01 -1.72051489e-01 6.17231093e-02 7.65816689e-01
1.50887728e-01 4.01363075e-02 -1.63929343e+00 -2.77707100e-01
2.15026230e-01 -2.17136629e-02 8.70470405e-01 5.97689390e-01
6.82597876e-01 -2.96735704e-01 1.92681551e-02 1.71718433e-01
1.08717632e+00 5.09295285e-01 9.17973757e-01 9.43876430e-02
2.30664000e-01 7.70296097e-01 4.37026858e-01 9.89176407e-02
-7.43573308e-02 1.01734281e+00 7.76422024e-01 -2.26586640e-01
-7.10963607e-01 -3.04225367e-02 6.60257041e-01 5.09028018e-01
8.35338980e-02 -6.43617392e-01 -6.55292153e-01 7.20057011e-01
-1.13170671e+00 -9.10496712e-01 -2.34759763e-01 2.55869651e+00
4.49688882e-01 4.31907028e-01 2.43262500e-01 8.15131247e-01
8.31823826e-01 2.03644916e-01 -1.42715245e-01 -7.23842263e-01
-1.90758884e-01 1.32438585e-01 3.61867726e-01 1.02479529e+00
-8.33630502e-01 5.32726228e-01 6.17764235e+00 8.48511279e-01
-1.58025420e+00 9.67704803e-02 4.13900226e-01 -4.18211877e-01
-3.56146619e-02 -2.54912078e-01 -4.07564789e-01 5.27546108e-01
1.09965169e+00 3.79310369e-01 3.30965221e-01 6.56812608e-01
7.70111322e-01 -3.69379640e-01 -7.60641813e-01 1.13274372e+00
2.37997189e-01 -5.26500225e-01 -3.52315336e-01 1.90059334e-01
1.24858031e-02 1.34565920e-01 2.55513519e-01 3.66656631e-02
-4.19505358e-01 -9.91672456e-01 9.13604736e-01 2.73222417e-01
5.31196833e-01 -7.73596048e-01 9.53860879e-01 4.66749966e-01
-8.85128438e-01 -1.78536435e-03 -7.90254399e-02 3.10012661e-02
3.44656259e-01 6.47509634e-01 -1.09908044e+00 2.65547991e-01
6.87958837e-01 1.63331866e-01 -3.75535846e-01 1.38945401e+00
-2.12124705e-01 9.03101206e-01 -5.02405524e-01 -7.62367994e-02
1.97595730e-01 4.48387831e-01 1.08185065e+00 1.50730133e+00
3.36050779e-01 -4.32474762e-01 -4.32431936e-01 3.87128800e-01
2.87301153e-01 2.90306360e-01 -6.91433311e-01 1.52928546e-01
2.33968392e-01 7.83916652e-01 -3.65339339e-01 -1.84143275e-01
-5.30789196e-01 8.93977642e-01 -3.24818581e-01 5.38985431e-01
-5.49378574e-01 -3.50924850e-01 5.02495050e-01 2.67027617e-01
5.39419293e-01 -1.88406438e-01 5.25530614e-02 -7.28638828e-01
2.43405163e-01 -9.37926650e-01 -2.85254903e-02 -9.93558526e-01
-6.50126636e-01 6.34069681e-01 -1.81925759e-01 -1.05861902e+00
-4.91230369e-01 -6.34178102e-01 -4.70643580e-01 9.78959560e-01
-1.48086667e+00 -6.57042861e-01 1.11797899e-02 3.12071741e-01
8.17476332e-01 5.97297810e-02 5.46665132e-01 4.41480190e-01
-8.87122974e-02 7.17606306e-01 3.15517604e-01 -1.00441314e-01
7.11202204e-01 -9.52467024e-01 2.37222657e-01 8.98392975e-01
3.05329740e-01 2.92872459e-01 1.45868123e+00 -2.33590186e-01
-1.00248039e+00 -6.29456639e-01 9.95360374e-01 6.43428341e-02
3.11557412e-01 -5.91045499e-01 -1.13101995e+00 8.72705951e-02
3.13007385e-01 -1.33758277e-01 4.57163095e-01 -2.13180408e-01
-1.35604367e-01 -1.78705618e-01 -1.00850534e+00 4.52469856e-01
4.91138250e-01 -6.36843085e-01 -6.57905459e-01 -4.35390025e-02
3.88473183e-01 -2.70291448e-01 -4.57420915e-01 2.61015177e-01
7.77539551e-01 -1.26517761e+00 9.43288982e-01 1.70822523e-03
-3.37723881e-01 -4.60412413e-01 -2.31696263e-01 -1.50745881e+00
3.63574386e-01 -5.60318887e-01 6.28026798e-02 1.33808780e+00
4.17354941e-01 -6.29857302e-01 4.26020294e-01 -1.00802176e-01
2.88982242e-02 -2.41642788e-01 -1.41402161e+00 -9.87854421e-01
-1.65884465e-01 -8.02111626e-01 2.35156611e-01 4.66960460e-01
-1.21020705e-01 1.43265963e-01 -5.71614683e-01 2.44122863e-01
2.64198035e-01 -5.59054732e-01 6.90452933e-01 -9.14180994e-01
-5.26342392e-01 -4.83151019e-01 -6.69528008e-01 -1.00785327e+00
-1.08969912e-01 -2.97089279e-01 2.71035135e-01 -1.37903023e+00
-4.03261423e-01 1.04973890e-01 -3.49155217e-02 -1.18072079e-02
7.92612359e-02 1.59366652e-01 4.10724312e-01 -1.80684909e-01
1.55368354e-02 3.61088216e-01 1.01962841e+00 -3.53284590e-02
-2.46992081e-01 3.47836673e-01 -2.26645470e-01 6.31047547e-01
7.10587859e-01 -3.91686171e-01 -4.92078006e-01 -1.75953671e-01
-1.71749637e-01 5.45785576e-02 5.63675344e-01 -1.26304495e+00
8.67633447e-02 3.92785311e-01 9.63317305e-02 -5.25562704e-01
7.82047927e-01 -9.52677846e-01 1.76194206e-01 4.73444372e-01
-2.66147494e-01 -3.55696023e-01 4.95226979e-01 5.29933929e-01
-4.45224345e-01 -5.69005430e-01 9.52886164e-01 -1.12835810e-01
-3.31427604e-01 -4.82680082e-01 -8.79276574e-01 -4.50978786e-01
6.27567649e-01 -3.72632116e-01 -1.45513833e-01 -1.01641893e+00
-1.02841437e+00 -3.98341149e-01 2.50403285e-01 4.41554725e-01
4.56047922e-01 -9.14944470e-01 -7.70785928e-01 2.83292651e-01
-1.31129876e-01 -4.59544331e-01 4.17977542e-01 9.73385811e-01
-2.97613174e-01 4.93287086e-01 2.20967159e-01 -7.31934369e-01
-1.86392391e+00 6.46366715e-01 6.80616796e-01 2.79947072e-01
-5.31658351e-01 6.12259090e-01 2.36273363e-01 8.39271471e-02
7.04973459e-01 -3.48889649e-01 -3.14030558e-01 8.30593333e-02
6.31709576e-01 3.92790616e-01 5.28831840e-01 -1.02796245e+00
-3.10897380e-01 3.76440018e-01 8.35588649e-02 -5.76161206e-01
9.55207467e-01 -3.70418012e-01 5.52323282e-01 6.97701156e-01
1.23689520e+00 3.96079510e-01 -1.10607529e+00 1.98643580e-01
-3.85512531e-01 -5.66744804e-01 4.80774641e-01 -9.73276436e-01
-9.95715737e-01 1.42169404e+00 9.75382149e-01 3.27369928e-01
1.23393142e+00 1.12097841e-02 2.77226597e-01 9.80034173e-02
-2.32370310e-02 -1.06674027e+00 -1.13313878e-02 2.54890174e-01
1.03489947e+00 -9.12425578e-01 -4.65633243e-01 -4.33596045e-01
-6.13171518e-01 1.09828258e+00 1.64609939e-01 4.32115555e-01
4.63541478e-01 5.22654235e-01 3.44935417e-01 4.32372868e-01
-5.78479707e-01 -5.39413512e-01 1.77199557e-01 9.17315185e-01
4.18416828e-01 -7.44479001e-02 -2.24346951e-01 1.82671249e-01
-3.85729283e-01 -1.30618379e-01 5.88407040e-01 7.06381023e-01
-6.32309735e-01 -9.43305671e-01 -9.19576466e-01 -9.24709905e-03
-7.37385869e-01 -1.87418461e-01 -5.07737458e-01 8.29219460e-01
1.82891548e-01 1.47057796e+00 -3.69678400e-02 -1.86243027e-01
5.38779318e-01 4.79296982e-01 3.43152046e-01 -5.30099310e-02
-5.10646284e-01 6.15956962e-01 3.47818762e-01 -1.84223518e-01
-4.32542801e-01 -6.59089744e-01 -8.48116517e-01 -1.38720229e-01
-4.63811606e-01 -1.17391115e-02 1.14768159e+00 9.00685430e-01
7.92831853e-02 7.36560404e-01 2.46881530e-01 -8.50307405e-01
-2.82656193e-01 -1.23442495e+00 -6.19259953e-01 2.44645119e-01
9.33514059e-01 -6.44441545e-01 -8.59945595e-01 -9.46097150e-02] | [14.527108192443848, 5.374457359313965] |
6b78e66e-8a82-4203-978b-de50bb596fa0 | reduced-label-complexity-for-tight-ell-2 | 2305.07486 | null | https://arxiv.org/abs/2305.07486v1 | https://arxiv.org/pdf/2305.07486v1.pdf | Reduced Label Complexity For Tight $\ell_2$ Regression | Given data ${\rm X}\in\mathbb{R}^{n\times d}$ and labels $\mathbf{y}\in\mathbb{R}^{n}$ the goal is find $\mathbf{w}\in\mathbb{R}^d$ to minimize $\Vert{\rm X}\mathbf{w}-\mathbf{y}\Vert^2$. We give a polynomial algorithm that, \emph{oblivious to $\mathbf{y}$}, throws out $n/(d+\sqrt{n})$ data points and is a $(1+d/n)$-approximation to optimal in expectation. The motivation is tight approximation with reduced label complexity (number of labels revealed). We reduce label complexity by $\Omega(\sqrt{n})$. Open question: Can label complexity be reduced by $\Omega(n)$ with tight $(1+d/n)$-approximation? | ['Malik Magdon-Ismail', 'Alex Gittens'] | 2023-05-12 | null | null | null | null | ['open-question'] | ['natural-language-processing'] | [ 2.46558473e-01 4.40516442e-01 -9.75028500e-02 -5.76040268e-01
-1.43607867e+00 -9.32081342e-01 -7.92674482e-01 3.20473701e-01
-7.85922825e-01 7.05028653e-01 -8.31235647e-01 -7.58433878e-01
-7.56428063e-01 -1.08893311e+00 -9.47579265e-01 -8.06127429e-01
-6.61797225e-01 7.88955569e-01 2.13821810e-02 -1.49809718e-01
2.87871450e-01 5.31335056e-01 -1.78420258e+00 -7.83750489e-02
3.87513161e-01 1.65248597e+00 -2.99038589e-01 7.59043157e-01
-2.20869899e-01 6.98085427e-01 -7.42470384e-01 -9.41781223e-01
9.23083007e-01 -7.64276326e-01 -1.27432060e+00 -1.69472575e-01
4.16897535e-01 -1.65952384e-01 -5.08553624e-01 1.49775434e+00
3.91061902e-01 3.34338814e-01 4.18437749e-01 -8.65201175e-01
-5.42370796e-01 9.39690530e-01 -1.33179414e+00 2.15557009e-01
-1.12236086e-02 4.64262813e-02 1.10488129e+00 -9.77972895e-02
3.42282236e-01 1.05433822e+00 4.60540056e-01 4.73526001e-01
-1.62177873e+00 -1.32600045e+00 -2.91901343e-02 -5.19134223e-01
-1.83164632e+00 -8.70333090e-02 3.81671965e-01 -1.50469929e-01
6.02402329e-01 6.66719139e-01 5.31554967e-02 -3.48853379e-01
-4.92910385e-01 2.83268452e-01 1.31166995e+00 -7.00107515e-01
2.73599476e-01 -1.49326891e-01 6.09464109e-01 1.19094968e+00
3.18680763e-01 3.81132774e-02 -2.10831225e-01 -8.35529417e-02
7.23552704e-01 2.64781624e-01 6.13749847e-02 2.94091761e-01
-5.02873957e-01 1.01840806e+00 2.65192926e-01 6.82176724e-02
2.98740536e-01 1.10975802e+00 2.97041625e-01 5.79824865e-01
3.00003648e-01 3.82779241e-01 -7.42474556e-01 -1.55781865e-01
-1.19071054e+00 3.65751922e-01 4.48598593e-01 1.72202516e+00
1.27493179e+00 -2.05195323e-01 2.01456901e-02 6.72905684e-01
-2.39971265e-01 1.18933189e+00 -3.60164821e-01 -1.88748908e+00
7.58810818e-01 4.83906269e-01 2.13844180e-01 -4.82050508e-01
-6.31581724e-01 7.84184318e-03 -8.45605552e-01 4.72340792e-01
8.96154702e-01 -2.36458302e-01 -6.72943056e-01 1.99012721e+00
8.15748274e-02 -7.50932753e-01 -4.72949654e-01 6.08047664e-01
3.32175851e-01 5.66193461e-01 -1.96934983e-01 -6.94514990e-01
1.46476412e+00 -4.83683854e-01 -1.53364137e-01 2.83911824e-01
1.30506098e+00 -7.19416797e-01 1.19435954e+00 6.50085151e-01
-2.13607788e+00 -3.15485537e-01 -6.93171561e-01 -3.10593098e-01
-2.62710154e-01 -2.66965330e-01 7.17169940e-01 1.22530663e+00
-9.70502555e-01 7.13193238e-01 -3.30983460e-01 5.46353638e-01
6.09576523e-01 1.17650723e+00 -5.89386374e-02 -2.32058421e-01
-5.63666165e-01 -1.09988069e-02 -2.59839356e-01 -4.86652941e-01
-8.18988502e-01 -5.39495945e-01 -4.76546645e-01 -6.62574619e-02
5.90889692e-01 -2.36463454e-02 1.19799304e+00 -2.83916116e-01
-7.96571314e-01 1.36220884e+00 -1.05238624e-01 -3.73181552e-01
2.36338288e-01 6.26415253e-01 -2.77086087e-02 2.91071206e-01
1.94327399e-01 6.14811122e-01 1.15526043e-01 -1.05462134e+00
-1.11081946e+00 -1.04924226e+00 4.94839042e-01 -4.42913681e-01
-1.26794964e-01 4.25663948e-01 -2.01433167e-01 -3.38053793e-01
5.40227830e-01 -1.02820885e+00 -5.28768122e-01 -2.73814112e-01
-1.51617333e-01 -3.28520656e-01 -4.13781144e-02 -9.89666358e-02
1.49505484e+00 -2.21332645e+00 -2.64961064e-01 8.97093654e-01
7.73620605e-01 -5.24803661e-02 1.25184953e-01 1.50995538e-01
-7.98434094e-02 4.13550287e-01 -1.18588150e-01 -1.78449988e-01
3.80398035e-01 1.90776289e-01 -3.23066674e-02 8.31106663e-01
-9.37175274e-01 3.28595668e-01 -5.59159100e-01 -2.48359650e-01
-3.76408249e-01 -6.60782009e-02 -5.26631415e-01 -8.65694135e-02
-5.97836494e-01 -5.28092086e-02 -3.36727977e-01 6.13597393e-01
1.50424314e+00 -4.87207681e-01 1.92920804e-01 1.22263469e-01
-1.41596824e-01 1.01390779e-01 -1.46079528e+00 1.46455121e+00
1.13826536e-01 -9.82092470e-02 4.43270117e-01 -1.46377635e+00
9.88896847e-01 -6.02120645e-02 9.82655823e-01 -1.31919801e+00
5.41340113e-01 5.38116395e-01 -4.23548460e-01 -3.04328412e-01
3.36082637e-01 -6.44327283e-01 -5.28331935e-01 1.03796816e+00
-3.25910181e-01 -1.13068476e-01 2.65979111e-01 -6.71643019e-02
1.74875224e+00 -3.86866838e-01 -6.39361978e-01 -4.22674239e-01
1.56743094e-01 7.96813145e-03 4.97138947e-01 9.93340015e-01
-1.74832284e-01 1.42524511e-01 9.69272375e-01 -5.73911250e-01
-1.21041691e+00 -7.80677497e-01 1.03935634e-03 1.88551724e+00
4.33287561e-01 -2.76141316e-01 -1.23300862e+00 -6.91704452e-01
-1.81182578e-01 6.41444147e-01 -7.70687699e-01 1.93948597e-01
-7.69973576e-01 -7.00062633e-01 8.48226011e-01 5.52837431e-01
2.39059746e-01 -7.09095180e-01 -7.34252930e-01 4.63187471e-02
1.33843586e-01 -5.43704450e-01 -6.69106483e-01 1.08089602e+00
-9.93510127e-01 -8.23297083e-01 -5.44601977e-01 -6.55935407e-01
1.11555767e+00 2.84522563e-01 9.71651196e-01 2.20993459e-01
-4.60565716e-01 1.66094601e-01 -5.02792835e-01 -6.28309786e-01
1.49389893e-01 -7.76527151e-02 -7.23197833e-02 -9.08699453e-01
7.64439762e-01 -5.93840480e-01 -1.15158355e+00 4.67067868e-01
-9.27453279e-01 -6.92821741e-01 3.91284451e-02 3.99193496e-01
1.35324538e+00 4.15850878e-01 -1.03743173e-01 -1.21326351e+00
5.70233315e-02 -2.64515188e-02 -1.41068685e+00 3.33799869e-02
-9.16128159e-01 1.71981454e-01 9.15340066e-01 -1.08424708e-01
-2.11207554e-01 -2.89875239e-01 -3.80592108e-01 -4.64008957e-01
2.18566507e-01 -1.76831663e-01 2.78695256e-01 -8.86526629e-02
1.04362881e+00 1.01539318e-03 -8.15035626e-02 -7.17497170e-01
4.90214229e-01 4.32062328e-01 2.57307619e-01 -1.01037002e+00
3.38170767e-01 6.67760909e-01 8.06990862e-01 -1.53697198e-02
-1.09878969e+00 -3.65982503e-02 7.84321502e-02 4.95767951e-01
5.53799272e-01 -5.03979504e-01 -2.16951442e+00 -5.51963151e-01
-6.12563908e-01 -5.34087300e-01 -1.22662902e+00 3.74496132e-02
-9.54688013e-01 2.60022640e-01 -4.15954351e-01 -1.30848634e+00
-3.81226301e-01 -1.30136347e+00 8.86879027e-01 1.37646943e-02
-2.98879556e-02 -2.77199268e-01 -5.02395749e-01 6.20764077e-01
5.33050597e-02 -5.50019145e-02 1.26242030e+00 -7.34257460e-01
-7.99920499e-01 -5.11743784e-01 -7.53932357e-01 5.41288376e-01
-5.24909973e-01 -7.20794082e-01 -5.29757440e-01 -3.84116411e-01
7.62252137e-02 -5.07653058e-01 5.19097865e-01 6.42992556e-01
1.97368896e+00 -7.27707386e-01 -2.15351790e-01 5.60676515e-01
1.67284644e+00 4.12498951e-01 7.13298380e-01 -2.08097681e-01
2.85501443e-02 2.71836042e-01 5.62091351e-01 8.65373135e-01
1.19341597e-01 2.80904412e-01 5.66974401e-01 2.25385427e-02
1.62373409e-01 2.40159586e-01 -2.71334648e-02 1.81867719e-01
-1.75249040e-01 -2.63541520e-01 -6.18546784e-01 4.19691563e-01
-1.31526387e+00 -6.79072440e-01 -3.64816695e-01 2.39978695e+00
9.07662868e-01 3.90633345e-01 5.38910091e-01 4.48475778e-01
4.40712333e-01 -3.21732402e-01 -2.66239405e-01 -8.98928046e-01
-2.00734347e-01 1.43161011e+00 1.32798600e+00 4.32768106e-01
-3.56661588e-01 6.51856422e-01 4.88648987e+00 1.57752347e+00
-4.35198396e-01 3.26084286e-01 1.05824685e+00 -7.07800329e-01
-4.91111845e-01 1.47099301e-01 -1.02416301e+00 6.65005863e-01
1.06948471e+00 3.58781129e-01 8.84357810e-01 1.08063769e+00
-1.91818342e-01 -4.22004074e-01 -1.45432639e+00 1.22254741e+00
-2.02647880e-01 -1.40316463e+00 -6.45796537e-01 6.38081491e-01
7.90349901e-01 -4.96958524e-01 4.88974363e-01 1.13050684e-01
1.10611534e+00 -1.33718359e+00 6.30115688e-01 1.26262113e-01
1.74153650e+00 -1.13558567e+00 5.09767532e-01 4.43902969e-01
-1.13193619e+00 -4.03909713e-01 -6.84042215e-01 7.09757151e-04
-1.00149572e-01 7.31238186e-01 -2.94874042e-01 1.35601997e-01
1.29353261e+00 -7.23031044e-01 2.91484863e-01 2.20727190e-01
4.98580664e-01 2.46689379e-01 -8.59519362e-01 -1.98724180e-01
3.80623430e-01 -2.71847010e-01 -2.29467511e-01 9.08754051e-01
6.08976364e-01 1.31152785e+00 4.11195457e-01 5.74520290e-01
-7.41750717e-01 2.37696692e-01 -1.90983102e-01 2.05250397e-01
5.47738552e-01 7.19597280e-01 -1.00993717e+00 -1.30095482e-01
-3.80461775e-02 4.02492911e-01 3.05868149e-01 -1.84762534e-02
-9.36606228e-01 -9.19885576e-01 4.69114482e-01 6.94355965e-01
6.67557597e-01 -2.20587969e-01 -7.50261664e-01 -6.99128136e-02
-5.24413437e-02 -6.56446815e-01 8.87852669e-01 -3.91455024e-01
-1.01231945e+00 3.87575984e-01 -9.38558057e-02 -8.12199712e-01
5.75463176e-01 -8.06267440e-01 6.33997619e-01 1.03180039e+00
-9.54027593e-01 -5.85403323e-01 1.60779193e-01 8.14635098e-01
-2.59414434e-01 -6.84305951e-02 9.69451427e-01 5.00647068e-01
1.51235402e-01 1.50310528e+00 3.56019944e-01 -2.40602121e-02
2.08816156e-01 -1.02120578e+00 -3.76525611e-01 7.67986253e-02
-3.61588538e-01 4.92246747e-01 6.62362516e-01 -2.11922109e-01
-1.31932461e+00 -7.29733765e-01 5.80048561e-01 -2.16093108e-01
1.09864444e-01 -2.03975156e-01 -2.11988375e-01 7.92413533e-01
-1.94089577e-01 2.51066476e-01 1.25131917e+00 -3.09439242e-01
-6.97131991e-01 -6.85056925e-01 -1.95692825e+00 5.32241821e-01
1.54948270e+00 -3.01984340e-01 4.09393728e-01 3.60889256e-01
9.11781549e-01 -4.83853906e-01 -1.35933721e+00 3.76923174e-01
3.10212255e-01 -1.08951652e+00 8.96604657e-01 -4.26077068e-01
8.65111426e-02 -7.28282845e-03 -8.63445759e-01 -8.77840966e-02
-2.62941062e-01 -1.14313173e+00 -3.38137150e-02 7.97096550e-01
6.86664343e-01 -2.32336327e-01 1.34554315e+00 1.13193083e+00
1.16566628e-01 -8.30216408e-01 -1.20821393e+00 -6.39350593e-01
7.74466515e-01 -4.68941033e-01 8.36962402e-01 4.95546639e-01
1.06626593e-01 -1.65824413e-01 -1.49977833e-01 -2.46034130e-01
6.14926338e-01 2.65827388e-01 6.93621039e-01 -9.94933784e-01
-5.42733133e-01 -4.98061061e-01 -5.74817434e-02 -1.54931879e+00
-4.43846107e-01 -1.11651063e+00 -2.30052799e-01 -1.05818820e+00
2.87338018e-01 -1.57130492e+00 -1.93101451e-01 7.68402815e-01
6.96478069e-01 2.29037330e-01 3.55658345e-02 -6.14511780e-02
-1.11330020e+00 -2.08767354e-01 1.24458373e+00 9.36084054e-03
1.81113034e-01 5.32940924e-02 -1.18795335e+00 6.12950504e-01
2.98988819e-01 -9.56523955e-01 -2.63057351e-01 -6.70793235e-01
1.08804250e+00 8.57243001e-01 -2.52519637e-01 -6.06893659e-01
1.57135725e-01 -3.10105652e-01 6.83759078e-02 -8.37266803e-01
4.07430232e-01 -1.04435933e+00 4.84646522e-02 2.51141399e-01
-5.36282182e-01 4.57007885e-01 1.24889184e-02 1.84897020e-01
6.84370935e-01 -8.51694763e-01 1.15322864e+00 -6.75851703e-01
5.00418723e-01 6.57305658e-01 1.08700469e-01 3.69738609e-01
1.35027945e+00 -4.04002726e-01 -6.07548118e-01 -1.37001067e-01
-7.50732124e-01 1.67400122e-01 4.58767533e-01 -9.29320216e-01
9.77991074e-02 -8.99240077e-01 -2.42834687e-01 1.06254689e-01
-8.00969824e-02 7.85862863e-01 7.06200659e-01 6.87158108e-01
-8.88922989e-01 1.41344100e-01 4.67522115e-01 -4.25840974e-01
-8.51372123e-01 9.98471439e-01 2.41903901e-01 -5.70322871e-01
-1.82392344e-01 1.83536184e+00 -3.18466663e-01 -2.67270237e-01
6.72413290e-01 -2.28737280e-01 1.01073432e+00 -1.81817144e-01
4.42123681e-01 1.13416123e+00 1.62738338e-02 -1.31922588e-01
3.14760655e-02 5.91398120e-01 -1.94564670e-01 -3.80534440e-01
1.42028832e+00 -3.14085037e-01 -6.52065456e-01 -9.71296355e-02
1.76456845e+00 1.01063646e-01 -9.97489631e-01 -1.93898648e-01
-1.97618261e-01 -5.37299037e-01 -6.07887626e-01 -6.72261655e-01
-1.18021429e+00 1.11254334e+00 1.08241856e+00 7.66982317e-01
1.30037570e+00 2.89271414e-01 1.18122303e+00 3.68360370e-01
8.38168919e-01 -1.76767826e+00 3.33688170e-01 8.44141468e-02
-2.77891159e-02 -7.39642441e-01 1.17071345e-01 -1.27669081e-01
-3.05183917e-01 9.01093185e-01 3.71575832e-01 -2.46317685e-01
1.13589466e+00 4.92422849e-01 -2.89707333e-01 -7.07732201e-01
-1.51930317e-01 -3.03150654e-01 -5.57210147e-01 2.44511012e-02
2.23383322e-01 1.37783051e-01 -5.56280851e-01 1.11497033e+00
-5.18517852e-01 5.06421551e-02 2.05121174e-01 1.24906051e+00
-1.03356314e+00 -1.47534251e+00 -4.51765865e-01 8.17520320e-01
-9.93221998e-01 1.29518852e-01 3.28636855e-01 2.96962410e-01
9.57589328e-01 9.65654135e-01 2.30751663e-01 -2.80555189e-01
4.29803938e-01 2.31281027e-01 8.53245676e-01 -3.12157482e-01
-5.66365838e-01 5.00738919e-01 -2.39241809e-01 -2.09314153e-01
-2.66218670e-02 -4.93711203e-01 -2.06779933e+00 -1.14590454e+00
-1.49930462e-01 3.88477772e-01 7.12423801e-01 5.98559380e-01
1.80902168e-01 6.77772388e-02 8.66101861e-01 9.75511000e-02
-7.80259073e-01 -5.08707643e-01 -1.36284113e+00 3.18943322e-01
-7.92439282e-02 -2.66640007e-01 -3.34719956e-01 -7.57270828e-02] | [6.379638195037842, 4.614198684692383] |
5fc29dfd-ce55-44bc-b8c4-d0fd5ab196d2 | explainable-data-poison-attacks-on-human | 2301.06923 | null | https://arxiv.org/abs/2301.06923v1 | https://arxiv.org/pdf/2301.06923v1.pdf | Explainable Data Poison Attacks on Human Emotion Evaluation Systems based on EEG Signals | The major aim of this paper is to explain the data poisoning attacks using label-flipping during the training stage of the electroencephalogram (EEG) signal-based human emotion evaluation systems deploying Machine Learning models from the attackers' perspective. Human emotion evaluation using EEG signals has consistently attracted a lot of research attention. The identification of human emotional states based on EEG signals is effective to detect potential internal threats caused by insider individuals. Nevertheless, EEG signal-based human emotion evaluation systems have shown several vulnerabilities to data poison attacks. The findings of the experiments demonstrate that the suggested data poison assaults are model-independently successful, although various models exhibit varying levels of resilience to the attacks. In addition, the data poison attacks on the EEG signal-based human emotion evaluation systems are explained with several Explainable Artificial Intelligence (XAI) methods, including Shapley Additive Explanation (SHAP) values, Local Interpretable Model-agnostic Explanations (LIME), and Generated Decision Trees. And the codes of this paper are publicly available on GitHub. | ['Chan Yeob Yeun', 'Nicola Bena', 'Claudio Agostino Ardagna', 'Ernesto Damiani', 'Sangyoung Yoon', 'Ahmed Y. Al Hammadi', 'Sani Umar', 'Zhibo Zhang'] | 2023-01-17 | null | null | null | null | ['data-poisoning'] | ['adversarial'] | [ 4.26645763e-02 1.88056365e-01 4.26057279e-01 -4.50672865e-01
-2.22174630e-01 -7.58102357e-01 2.96045929e-01 1.05279364e-01
-1.25790790e-01 1.04909742e+00 -1.27367690e-01 -3.29787403e-01
-4.23608035e-01 -2.75413871e-01 -5.03747880e-01 -7.09584236e-01
-6.82802439e-01 -2.98706796e-02 -6.78830802e-01 -6.66807070e-02
4.77321833e-01 5.45022368e-01 -1.06464279e+00 4.21125531e-01
7.26794243e-01 1.09942234e+00 -7.40546525e-01 6.96801722e-01
6.69701755e-01 1.10783482e+00 -1.33289731e+00 -6.87067151e-01
9.98473465e-02 -5.35666287e-01 -6.69018090e-01 -4.79673088e-01
-4.93409425e-01 -3.32865268e-01 -1.91119641e-01 1.33932817e+00
7.70080447e-01 -4.59547877e-01 5.06196916e-01 -2.59022713e+00
-5.77604234e-01 7.89432466e-01 -2.11747110e-01 3.31397861e-01
7.07109749e-01 2.73291856e-01 2.25560993e-01 -5.15604258e-01
1.85018275e-02 1.03348792e+00 6.75940633e-01 9.77843404e-01
-1.04133749e+00 -1.66177106e+00 -3.52195710e-01 5.65588415e-01
-1.63277829e+00 5.56352548e-02 9.66072202e-01 -1.63283244e-01
1.05026126e+00 6.03243530e-01 7.98021257e-01 1.72827470e+00
8.75702918e-01 3.19209188e-01 1.55198348e+00 1.86118074e-02
8.52580369e-01 4.26509917e-01 6.14127100e-01 3.34452003e-01
5.34682870e-01 3.82149398e-01 -1.05443215e+00 -9.51793194e-01
2.15471193e-01 -9.33041573e-02 -3.56326193e-01 3.30457151e-01
-6.07165933e-01 9.18444216e-01 3.07410002e-01 -2.09950414e-02
-8.36883187e-01 2.13729441e-01 6.89290166e-01 3.83943796e-01
4.60127622e-01 1.04263783e+00 -7.26381004e-01 -3.10058802e-01
-4.20745641e-01 1.15035428e-02 7.51098931e-01 6.58740759e-01
2.69775450e-01 4.52377141e-01 2.63864398e-01 -2.84924746e-01
1.37600124e-01 6.42197907e-01 4.27980453e-01 -5.94939768e-01
2.42141560e-02 5.42135417e-01 2.09925603e-02 -1.23092604e+00
-6.86238110e-01 -3.76922250e-01 -8.41002047e-01 1.18428014e-01
-2.53363460e-01 -7.51761317e-01 -2.62020350e-01 1.90076780e+00
-1.83100384e-02 5.90321183e-01 3.84967953e-01 8.93088818e-01
5.92028797e-01 3.91957194e-01 7.20277727e-01 -1.59428284e-01
1.40653729e+00 -2.23630011e-01 -1.10528290e+00 1.12257879e-02
5.12392640e-01 1.91714779e-01 8.11213732e-01 9.15557921e-01
-7.47472882e-01 -1.53041119e-02 -1.30600893e+00 8.05431128e-01
-5.81992507e-01 -5.05357802e-01 1.07429409e+00 1.46209514e+00
-8.17291141e-01 2.09784970e-01 -5.80549538e-01 -1.77214712e-01
4.22379404e-01 8.38733017e-01 -4.65942085e-01 4.39627975e-01
-1.84715211e+00 1.08620334e+00 9.68814641e-02 -3.11057661e-02
-1.20659971e+00 -5.09452403e-01 -6.59624040e-01 2.50628620e-01
-3.16561371e-01 -1.97734147e-01 4.80718493e-01 -8.72717202e-01
-1.02807665e+00 4.17406380e-01 2.73736238e-01 -9.72926319e-01
1.49518967e-01 1.33811399e-01 -8.12828779e-01 5.47343016e-01
-2.30600417e-01 4.81524616e-01 7.32770085e-01 -1.20393872e+00
1.23388814e-02 -5.58667541e-01 -1.86343253e-01 -1.24764383e-01
-6.46751702e-01 5.42849898e-01 9.12440777e-01 -5.24766684e-01
-4.05753195e-01 -7.27218032e-01 -4.04870994e-02 -6.75997019e-01
-6.99533403e-01 -7.80719891e-02 9.51041043e-01 -6.79650664e-01
1.26947165e+00 -1.86901033e+00 -3.23845118e-01 5.52015603e-01
3.18692505e-01 -1.08663561e-02 1.32071152e-01 2.87268668e-01
-6.41966939e-01 6.31060481e-01 -2.88798302e-01 -7.02814083e-04
2.94934988e-01 -2.83760101e-01 -7.55192876e-01 7.72512138e-01
9.34194177e-02 8.41765285e-01 -5.24115980e-01 1.92134734e-02
1.27605796e-01 6.51255608e-01 -2.21166760e-01 2.86541969e-01
6.33381844e-01 4.97093678e-01 -5.06827235e-01 6.49552524e-01
4.91837353e-01 3.43716085e-01 -1.17473386e-01 1.98284183e-02
4.22507823e-01 3.01191723e-03 -6.48330986e-01 8.88769627e-01
1.26667470e-01 6.34358764e-01 -2.75114417e-01 -8.43088984e-01
9.23508823e-01 8.74033034e-01 3.45289528e-01 -5.33697367e-01
5.67972302e-01 -6.07420243e-02 2.24828973e-01 -5.26511729e-01
-2.15961430e-02 -6.78115785e-02 -5.00085175e-01 1.05179787e+00
-9.73104611e-02 3.93247120e-02 -6.19710326e-01 4.53019470e-01
1.66040182e+00 -4.71789807e-01 4.79771018e-01 -4.56410527e-01
2.63746113e-01 -3.50486264e-02 5.12273788e-01 7.60182261e-01
-6.41783059e-01 -4.84881699e-02 3.86622578e-01 -6.44080281e-01
-6.12085462e-01 -9.16423678e-01 -2.92761087e-01 5.13251007e-01
4.11099315e-01 -4.70146686e-01 -1.52027750e+00 -5.81650436e-01
-2.98549771e-01 1.42581058e+00 -9.73257840e-01 -9.08031166e-01
2.39499539e-01 -1.00995350e+00 1.38064778e+00 6.20928943e-01
6.49171650e-01 -1.45758975e+00 -1.39011538e+00 -5.10182567e-02
-2.44105250e-01 -8.42876375e-01 2.39943072e-01 7.56917298e-01
-4.45263922e-01 -1.11927819e+00 -6.07170048e-04 6.82881698e-02
8.01856697e-01 -3.26530099e-01 6.15889072e-01 4.88268226e-01
-3.88666958e-01 7.16060519e-01 -4.83477443e-01 -1.18115747e+00
-1.60853371e-01 -4.73148704e-01 8.61655891e-01 -3.82666625e-02
1.05929780e+00 -4.47252423e-01 -3.20357472e-01 3.18283439e-01
-9.45894182e-01 -3.93168032e-01 2.56196018e-02 5.02754271e-01
-1.39328748e-01 6.42964661e-01 1.15155709e+00 -3.84282678e-01
1.43886864e+00 -7.49087930e-01 -2.60542873e-02 3.57455850e-01
-7.03545928e-01 -2.61354059e-01 6.16042495e-01 -7.96489418e-01
-6.99885249e-01 -2.82559305e-01 1.49066553e-01 -3.79256129e-01
-5.76745331e-01 3.22088540e-01 -4.17957693e-01 -3.75892669e-01
7.79612899e-01 7.31680542e-02 -6.21308982e-01 1.19398676e-01
-1.86671808e-01 8.71973217e-01 5.11454761e-01 -3.75621647e-01
6.33683562e-01 1.16676286e-01 -2.18817204e-01 -5.21851420e-01
-2.57859886e-01 2.03687951e-01 -2.55240470e-01 -6.38708651e-01
1.06757557e+00 -6.90898299e-01 -1.15109646e+00 5.51464319e-01
-1.40733683e+00 -6.91609532e-02 1.71906441e-01 5.61235428e-01
-2.87372619e-01 5.79377972e-02 -7.43500173e-01 -1.44100094e+00
-8.08571160e-01 -1.15826643e+00 5.77964425e-01 9.39683765e-02
-9.34747756e-01 -8.02731812e-01 -1.96558759e-01 2.73516029e-01
2.71196187e-01 6.56251431e-01 1.03004682e+00 -1.55082238e+00
2.68508315e-01 -6.61600769e-01 2.97325671e-01 2.39051785e-02
-1.77889943e-01 -3.51984710e-01 -1.30440271e+00 -1.58670947e-01
9.27027643e-01 -6.50174558e-01 -1.15448311e-01 2.55366653e-01
1.11597764e+00 -6.70271695e-01 -1.06773540e-01 4.56064850e-01
7.79217243e-01 7.02027142e-01 8.15153182e-01 3.37315530e-01
1.87920406e-01 7.78360188e-01 2.11331278e-01 7.63319790e-01
4.33091931e-02 1.87382847e-01 7.47440934e-01 -8.86079296e-02
7.23274112e-01 -9.35996100e-02 7.24330306e-01 2.70159632e-01
2.44940743e-01 -2.72883713e-01 -9.12889183e-01 -1.23917371e-01
-1.49219120e+00 -1.19700074e+00 -2.48338446e-01 1.89511156e+00
3.74025851e-01 1.14535037e-02 -6.05475046e-02 6.41705394e-01
8.63008559e-01 -6.26909673e-01 -8.82237852e-01 -8.71112943e-01
-2.93253094e-01 2.36773506e-01 3.50912958e-01 1.13281071e-01
-8.56707036e-01 6.00299537e-01 6.94119167e+00 4.02029574e-01
-8.36428881e-01 2.11641505e-01 7.80262351e-01 -4.40667458e-02
-1.04002461e-01 -1.94100916e-01 -5.03814630e-02 6.30380154e-01
1.41538286e+00 -5.58775842e-01 6.77518785e-01 7.77368963e-01
2.85887480e-01 2.57605575e-02 -1.05222309e+00 1.17970693e+00
3.59626204e-01 -7.08872020e-01 -2.07274958e-01 1.86151758e-01
2.01655075e-01 -3.73454303e-01 2.28432804e-01 1.75866038e-01
1.63404927e-01 -1.42632914e+00 9.11398709e-01 2.55281091e-01
5.73491216e-01 -1.27395082e+00 1.00228655e+00 3.58787328e-01
-6.37090623e-01 -4.90963817e-01 -3.05980802e-01 -6.08032525e-01
-2.09826618e-01 1.22300453e-01 -6.28013730e-01 1.31321386e-01
9.68620420e-01 8.78408998e-02 -7.51534522e-01 5.66052318e-01
-7.68851399e-01 1.07242930e+00 -1.61895081e-01 -2.10851073e-01
-4.78867106e-02 2.70110607e-01 4.89714831e-01 8.72814536e-01
9.61558521e-02 6.81756377e-01 -3.98198396e-01 1.07665598e+00
2.15759173e-01 -3.48327756e-01 -9.07877386e-01 3.16139124e-02
7.08838701e-01 1.21569598e+00 -7.80107617e-01 -1.76759288e-01
3.33971828e-01 1.02110410e+00 -2.08537191e-01 3.53332847e-01
-1.31792569e+00 -5.40451467e-01 3.78212661e-01 -4.22618747e-01
-8.81733000e-01 3.88012916e-01 -7.88545489e-01 -1.10770655e+00
-4.10807222e-01 -1.24032080e+00 3.19559336e-01 -1.28933716e+00
-1.32431436e+00 1.07271671e+00 2.79007137e-01 -8.62396479e-01
-7.96427131e-02 -3.03151280e-01 -8.85043323e-01 6.45158350e-01
-6.33711457e-01 -8.08066905e-01 -1.25863940e-01 9.19576943e-01
1.67464241e-01 -3.88283253e-01 1.34493387e+00 -1.19819224e-01
-7.14713097e-01 6.73278213e-01 -7.52303958e-01 1.34794250e-01
2.97462314e-01 -9.15038466e-01 2.49551222e-01 9.17560339e-01
4.86305133e-02 8.88718069e-01 1.06623793e+00 -8.12522888e-01
-1.29453576e+00 -6.28935337e-01 6.23342931e-01 -7.26811886e-01
6.63170576e-01 -6.32266164e-01 -7.55461395e-01 6.60439551e-01
6.93453014e-01 -6.01969123e-01 1.15954554e+00 -1.81682557e-01
-3.09081018e-01 4.12691921e-01 -1.78358948e+00 5.70805907e-01
4.24920887e-01 -7.31491566e-01 -6.11974418e-01 2.92127788e-01
4.75071609e-01 3.94025177e-01 -7.18338013e-01 3.28729808e-01
3.52758795e-01 -7.34031737e-01 6.68769002e-01 -1.08473599e+00
-1.30463028e-02 -8.26821029e-02 -1.83022320e-01 -1.45393169e+00
-1.12699807e-01 -8.88582647e-01 -1.82593521e-02 1.11658490e+00
3.72345895e-01 -9.64022636e-01 3.47434938e-01 1.53469920e+00
3.11965525e-01 -4.60052431e-01 -8.91132712e-01 -5.60376644e-01
-1.27897426e-01 -8.64669800e-01 1.03254092e+00 1.33889687e+00
1.11872292e+00 2.86204815e-01 -6.40745640e-01 6.56943858e-01
1.01311576e+00 -7.20311284e-01 4.74246413e-01 -1.04336190e+00
8.79736394e-02 -6.00133575e-02 -8.65981221e-01 4.45849031e-01
5.14985919e-01 -7.50865877e-01 -2.73174047e-01 -7.48748302e-01
4.93954778e-01 1.92832917e-01 -6.28128886e-01 1.05127215e+00
-1.12174861e-01 4.02529806e-01 2.99101084e-01 6.47411868e-02
-5.21365523e-01 4.49760228e-01 8.55853334e-02 -1.65913701e-01
4.80096564e-02 -3.33516955e-01 -9.12836492e-01 1.07737565e+00
1.21494925e+00 -1.23125362e+00 -5.26457906e-01 -1.21952534e-01
3.84068757e-01 1.62391052e-01 7.96170056e-01 -1.23754525e+00
4.70410079e-01 1.70074582e-01 6.74797893e-01 -1.54118657e-01
2.33680695e-01 -1.18315935e+00 5.05828559e-01 8.21780682e-01
-6.53491914e-01 5.74159682e-01 5.18897414e-01 3.50196630e-01
1.00734062e-01 -1.16596207e-01 5.00214100e-01 3.12204987e-01
-1.58733830e-01 3.67866196e-02 -9.78781641e-01 -3.87997746e-01
1.47455466e+00 -2.65082359e-01 -5.44688582e-01 -8.36993217e-01
-5.97578526e-01 2.81087067e-02 1.44858882e-01 1.84833258e-01
9.90261137e-01 -1.19044399e+00 -4.90318805e-01 2.06816688e-01
-8.10111128e-03 -1.17498374e+00 1.91289276e-01 8.04899991e-01
-6.56440407e-02 3.07990789e-01 -6.63483620e-01 -3.61476503e-02
-1.29451048e+00 7.21457601e-01 3.57007951e-01 1.38634294e-01
-3.91404748e-01 7.43277788e-01 2.60826528e-01 -1.54983878e-01
4.73746449e-01 3.87786984e-01 -4.21989225e-02 -3.72262478e-01
9.38757122e-01 4.55509901e-01 -1.56204998e-01 -5.68136096e-01
-8.59440207e-01 -2.53797501e-01 3.50243956e-01 -2.83079088e-01
1.35326397e+00 5.60091063e-03 -1.02842100e-01 1.50330812e-02
8.36664081e-01 -4.24922407e-01 -8.32748473e-01 7.57716835e-01
6.77952468e-02 -3.04957945e-02 3.05317193e-02 -1.57610881e+00
-8.65177155e-01 1.19529819e+00 7.94011831e-01 5.23302138e-01
1.39973474e+00 -4.63984430e-01 3.85183007e-01 4.16707009e-01
9.74014878e-01 -7.07771719e-01 -6.93825185e-02 -1.08457811e-01
7.76168823e-01 -6.04358912e-01 -4.94808778e-02 1.61887810e-01
-1.22279584e+00 8.10635686e-01 5.16246557e-01 -2.81271394e-02
6.42196178e-01 6.85424387e-01 1.73100848e-02 -6.26847088e-01
-8.93582523e-01 7.53841817e-01 -1.61088452e-01 1.07406604e+00
-1.58963695e-01 2.10948244e-01 -1.85472205e-01 1.70482087e+00
-3.84718806e-01 -2.82838911e-01 8.04568291e-01 4.93899018e-01
-5.62589467e-02 -5.55673718e-01 -8.53804946e-01 2.33775020e-01
-7.96231806e-01 -3.94542128e-01 -1.22557414e+00 6.20128334e-01
-7.48848990e-02 1.68636131e+00 -4.04403865e-01 -1.05849230e+00
1.08387740e-02 3.53401482e-01 -6.82744458e-02 1.75143480e-02
-1.29663455e+00 -4.46812868e-01 -1.47904068e-01 -6.91304982e-01
-2.96765536e-01 -3.42668921e-01 -1.54256082e+00 -4.27435160e-01
-4.73798573e-01 6.38742924e-01 9.41107869e-01 9.69298005e-01
4.92520183e-01 3.71394694e-01 7.78818965e-01 -6.15767717e-01
-3.36064667e-01 -9.99398708e-01 -9.77004707e-01 4.75209713e-01
-1.86941549e-01 -6.56310976e-01 -9.17512238e-01 -1.47603033e-02] | [13.294084548950195, 3.028852939605713] |
d49421e7-3806-40dd-86c1-92567533eba9 | scaling-up-open-tagging-from-tens-to | null | null | https://aclanthology.org/P19-1514 | https://aclanthology.org/P19-1514.pdf | Scaling up Open Tagging from Tens to Thousands: Comprehension Empowered Attribute Value Extraction from Product Title | Supplementing product information by extracting attribute values from title is a crucial task in e-Commerce domain. Previous studies treat each attribute only as an entity type and build one set of NER tags (e.g., BIO) for each of them, leading to a scalability issue which unfits to the large sized attribute system in real world e-Commerce. In this work, we propose a novel approach to support value extraction scaling up to thousands of attributes without losing performance: (1) We propose to regard attribute as a query and adopt only one global set of BIO tags for any attributes to reduce the burden of attribute tag or model explosion; (2) We explicitly model the semantic representations for attribute and title, and develop an attention mechanism to capture the interactive semantic relations in-between to enforce our framework to be attribute comprehensive. We conduct extensive experiments in real-life datasets. The results show that our model not only outperforms existing state-of-the-art NER tagging models, but also is robust and generates promising results for up to 8,906 attributes. | ['Man Lan', 'Wenting Wang', 'Huimin Xu', 'Xinyu Jiang', 'Xin Mao'] | 2019-07-01 | null | null | null | acl-2019-7 | ['attribute-value-extraction'] | ['natural-language-processing'] | [-3.73306498e-03 2.28206292e-01 -5.42733431e-01 -7.61875749e-01
-7.68420756e-01 -9.32187736e-01 3.26717854e-01 5.70947468e-01
-5.07224202e-01 6.55929506e-01 2.61405915e-01 -1.98237851e-01
-7.38714784e-02 -1.18285584e+00 -5.84021688e-01 -3.84474993e-01
9.77824703e-02 7.81195164e-01 2.74268031e-01 -3.79104733e-01
-6.62154257e-02 9.21704024e-02 -1.41268897e+00 1.49353102e-01
1.14184046e+00 1.25557065e+00 -1.94251742e-02 7.37307360e-03
-6.01892889e-01 4.72579032e-01 -5.53962648e-01 -9.79165018e-01
5.00006117e-02 -8.81418437e-02 -9.77249622e-01 -1.35919809e-01
1.40211001e-01 -5.24215261e-03 3.63008045e-02 1.36029625e+00
3.57740402e-01 -1.29310405e-02 5.62185884e-01 -1.28237319e+00
-9.71416831e-01 1.05429959e+00 -3.38995397e-01 -2.38366738e-01
3.46929431e-01 -4.53818470e-01 1.39246416e+00 -4.85279351e-01
6.81192815e-01 9.83716786e-01 7.09080994e-01 4.92746562e-01
-1.12047505e+00 -6.45005286e-01 3.30081314e-01 -6.48538023e-02
-1.57006562e+00 -3.10168505e-01 5.09161234e-01 3.30208056e-02
9.44779098e-01 2.73754179e-01 5.27518809e-01 7.78983831e-01
-5.55658303e-02 5.52556396e-01 7.44256377e-01 -2.37409651e-01
3.21431667e-01 4.22979891e-01 3.78880084e-01 4.21759009e-01
5.85767269e-01 -3.28564197e-01 -1.63624108e-01 -4.65023667e-01
4.97478962e-01 -9.84941870e-02 3.09476912e-01 -2.19777986e-01
-1.16859818e+00 7.81462550e-01 2.08590001e-01 5.89915663e-02
-6.79528356e-01 -1.02502309e-01 4.32562739e-01 8.09253082e-02
4.64392573e-01 5.74938595e-01 -9.79213715e-01 1.04249440e-01
-3.19230407e-01 2.90563792e-01 1.02014935e+00 1.72610176e+00
8.10476840e-01 -4.50248480e-01 -3.67667302e-02 9.62228298e-01
3.03645641e-01 3.36330444e-01 2.72170424e-01 -7.44867206e-01
5.63055575e-01 8.99448693e-01 4.09727752e-01 -8.15811098e-01
-4.52805996e-01 -5.54068744e-01 -6.61190033e-01 -6.07822299e-01
2.55626142e-01 -1.13127843e-01 -8.66289973e-01 1.72507215e+00
5.84191024e-01 5.49717210e-02 2.42780462e-01 6.74184680e-01
1.06308460e+00 5.88801205e-01 1.01086545e+00 -1.49238765e-01
1.95887804e+00 -6.95049644e-01 -9.57495213e-01 -2.78003067e-01
1.05457056e+00 -8.33098531e-01 8.70320678e-01 2.89641935e-02
-7.17652738e-01 -4.45954919e-01 -9.93979514e-01 -2.00576797e-01
-9.51780617e-01 8.62666219e-02 1.32008791e+00 5.17527401e-01
-3.77761334e-01 4.66099977e-01 -3.34857404e-01 -4.26791370e-01
2.79068351e-01 6.27068937e-01 -3.72561961e-01 4.19436619e-02
-1.85063362e+00 5.95947504e-01 6.41932786e-01 -1.01947278e-01
-1.02527000e-01 -6.66141331e-01 -1.00996780e+00 3.26397479e-01
6.09458148e-01 -8.91591012e-01 1.02255797e+00 -7.34830081e-01
-9.93626714e-01 7.13828266e-01 -1.84851363e-01 -5.65400198e-02
-2.04371855e-01 -3.74629587e-01 -9.61203337e-01 -2.75346577e-01
4.71943319e-01 6.75577700e-01 -8.23760219e-03 -1.40935445e+00
-9.61311042e-01 -4.87461329e-01 3.23149174e-01 1.94780290e-01
-5.91787338e-01 2.46193424e-01 -4.72936124e-01 -6.20587170e-01
2.84310371e-01 -7.62129962e-01 -3.86528045e-01 -5.37626743e-01
-5.17945766e-01 -6.13978565e-01 1.54773265e-01 -3.55846941e-01
1.49141622e+00 -2.02220631e+00 -5.00640988e-01 1.21792786e-01
2.03721195e-01 6.31931573e-02 -8.09989274e-02 3.23499352e-01
8.68694559e-02 3.96719396e-01 -2.68795667e-03 5.02767153e-02
1.94872707e-01 4.32752579e-01 -2.17241615e-01 -8.34965631e-02
9.11002904e-02 1.09045839e+00 -9.30795074e-01 -8.38677526e-01
-2.69016027e-01 4.59499151e-01 -3.56511742e-01 1.22946538e-01
-3.07538837e-01 -2.21526518e-01 -1.21176267e+00 9.69141603e-01
8.18106413e-01 -3.37960839e-01 3.83570731e-01 -6.18526995e-01
1.79108977e-01 4.74229813e-01 -1.22934532e+00 1.64419258e+00
-3.33852410e-01 -5.71277618e-01 -8.32609385e-02 -6.31483018e-01
1.06911135e+00 2.33560890e-01 6.95043087e-01 -5.78190446e-01
2.30686620e-01 2.52453148e-01 -3.85419130e-01 -3.66538972e-01
6.93827093e-01 -1.50743634e-01 -7.07133532e-01 1.42237991e-01
1.54309049e-01 4.20855284e-01 1.98215097e-01 1.19323112e-01
8.14915836e-01 1.91204295e-01 4.70513344e-01 -4.29111779e-01
5.10767639e-01 2.22523361e-01 9.86280441e-01 5.22393346e-01
-1.48946205e-02 7.78335854e-02 4.69579130e-01 -5.48369348e-01
-1.00854373e+00 -7.23607242e-01 -1.83643460e-01 1.42644358e+00
5.76192200e-01 -6.11760974e-01 -7.73202717e-01 -1.13536942e+00
1.82281226e-01 8.12427342e-01 -4.98903483e-01 -5.73743060e-02
-4.49538618e-01 -9.42494929e-01 4.85518456e-01 8.43823731e-01
6.62957072e-01 -8.03821921e-01 -6.18225560e-02 5.69686413e-01
-4.23382014e-01 -1.29403460e+00 -5.02567708e-01 3.40258569e-01
-7.95368969e-01 -7.85960436e-01 -2.11328104e-01 -8.93464565e-01
5.14208078e-01 -1.35204848e-03 1.52126563e+00 -4.37064134e-02
3.67892772e-01 -9.34133828e-02 -5.38341463e-01 -5.86843431e-01
-1.81649014e-01 6.23835087e-01 -9.00907815e-02 -3.20427418e-02
1.23176408e+00 -3.87026757e-01 -5.15280068e-01 6.58692837e-01
-8.39913726e-01 -2.06720427e-01 8.07031274e-01 6.83369756e-01
8.50713909e-01 1.80329576e-01 9.60199535e-01 -1.52736616e+00
4.47678119e-01 -7.73384869e-01 -4.46039885e-01 5.14547467e-01
-1.21800113e+00 1.62259787e-01 6.92843199e-01 -4.78645295e-01
-1.25802255e+00 3.61514926e-01 -3.44172627e-01 4.27959710e-01
-4.15512145e-01 6.95962012e-01 -8.22696686e-01 4.74869400e-01
3.27128172e-01 -5.75656770e-03 -3.51982176e-01 -7.98065186e-01
5.22441268e-01 9.32107031e-01 5.05951345e-01 -8.74160171e-01
4.62612271e-01 1.57832414e-01 1.30329043e-01 -2.68509239e-01
-1.23524201e+00 -6.74647689e-01 -7.42461383e-01 4.20756698e-01
7.10291266e-01 -1.05928445e+00 -8.48352075e-01 1.00413404e-01
-9.82207000e-01 3.69599342e-01 -8.54416192e-02 4.78721440e-01
-2.79090166e-01 2.12874621e-01 -7.04439104e-01 -3.84911656e-01
-4.36299145e-01 -6.79807365e-01 9.88668025e-01 3.15603167e-01
-2.18790010e-01 -7.70253003e-01 -3.35069776e-01 3.35565329e-01
3.87737930e-01 6.36583120e-02 1.29310811e+00 -1.28005934e+00
-2.99603045e-01 -4.99980509e-01 -3.13985348e-01 -7.67090619e-02
2.49989122e-01 -4.79462534e-01 -6.85657561e-01 1.87338293e-01
-3.24766368e-01 -3.50583978e-02 7.13075161e-01 -9.70320851e-02
1.05803537e+00 -4.29066896e-01 -6.54752254e-01 5.10424972e-01
1.39732981e+00 3.43165874e-01 6.99576557e-01 5.12272060e-01
7.50643909e-01 7.80312419e-01 1.22368419e+00 4.59038645e-01
7.90923774e-01 8.20331931e-01 3.05972248e-01 -3.62576216e-01
2.07063094e-01 -5.66541672e-01 -2.40455732e-01 5.96555650e-01
-2.21791744e-01 -3.73495936e-01 -3.77838820e-01 4.95004386e-01
-1.91179824e+00 -6.54975832e-01 -2.38732949e-01 2.23152423e+00
1.10354233e+00 9.44404975e-02 2.21139237e-01 -2.42591292e-01
7.68603921e-01 -1.79571450e-01 -5.19063771e-01 -1.19907431e-01
-1.13809198e-01 7.66503438e-02 7.13725686e-01 1.42574608e-01
-1.32872999e+00 1.21316457e+00 5.96217012e+00 7.98996031e-01
-4.28020239e-01 -2.53977925e-02 7.46449947e-01 3.21836680e-01
-5.63488960e-01 9.17663053e-02 -1.27237463e+00 5.23201168e-01
8.82858634e-01 -1.84347570e-01 4.90160398e-02 1.04076219e+00
-4.07470286e-01 2.32016191e-01 -1.02157378e+00 6.47825181e-01
-2.82974571e-01 -8.86441588e-01 2.76963413e-01 2.95323223e-01
2.66496897e-01 -5.89861214e-01 -2.57709384e-01 5.47252655e-01
6.03585601e-01 -8.96339357e-01 5.30356407e-01 2.05631062e-01
8.49732995e-01 -1.03278565e+00 1.15770268e+00 9.89639834e-02
-1.44359779e+00 3.59827504e-02 -6.79582059e-01 3.58195990e-01
8.80413353e-02 5.71935654e-01 -3.37947458e-01 8.64496708e-01
7.05347180e-01 5.61783671e-01 -4.26782429e-01 5.73851347e-01
-1.15570329e-01 5.00758529e-01 -2.64952779e-01 -7.92861134e-02
1.86442673e-01 -1.07011616e-01 -5.69453835e-02 1.23758996e+00
4.63832408e-01 5.37198544e-01 1.77282900e-01 6.35952353e-01
-3.29272032e-01 4.36844051e-01 -3.70893955e-01 -1.68696091e-01
9.62252140e-01 1.53373599e+00 -7.41957128e-01 -5.82620502e-01
-7.82926202e-01 9.45317209e-01 1.07356422e-01 2.87153333e-01
-8.06718826e-01 -7.39420414e-01 8.35190892e-01 1.45819366e-01
2.81103671e-01 1.13359556e-01 -2.86192030e-01 -1.00618541e+00
2.63634394e-03 -4.15103704e-01 8.68279099e-01 -4.00130153e-01
-1.46739161e+00 6.82313204e-01 -3.10823619e-02 -1.22434759e+00
-1.91758215e-01 -4.69310999e-01 5.50302118e-02 8.36962044e-01
-1.60539269e+00 -1.41908920e+00 -2.11374819e-01 3.51121277e-01
2.16162369e-01 6.56271577e-02 1.15397322e+00 5.96559346e-01
-2.02554971e-01 8.84345591e-01 -1.10585801e-01 3.12378556e-01
8.55396926e-01 -1.14384770e+00 4.84757543e-01 2.18113497e-01
-1.33758122e-02 1.02806175e+00 6.73495710e-01 -8.09106171e-01
-1.35095310e+00 -1.05872345e+00 1.60892093e+00 -5.60753644e-01
3.70754361e-01 -5.53831816e-01 -1.02955842e+00 7.28485107e-01
1.02304399e-01 1.27176434e-01 1.07846284e+00 6.49460435e-01
-5.45342326e-01 -3.53697389e-01 -1.30533874e+00 9.31567177e-02
1.21383083e+00 2.17823554e-02 -5.34350038e-01 5.67315333e-02
1.03219128e+00 -9.81473103e-02 -1.49024844e+00 4.80055124e-01
6.29924119e-01 -3.57220352e-01 1.13002074e+00 -8.87763560e-01
2.65438378e-01 -4.14332896e-01 -1.41326532e-01 -1.11534762e+00
-6.78421497e-01 -2.72693038e-01 -2.23083600e-01 1.95157802e+00
7.47669280e-01 -7.06936717e-01 7.14650095e-01 1.01144087e+00
-1.96397640e-02 -7.72613525e-01 -5.54759085e-01 -8.13062787e-01
-1.48280889e-01 6.43318817e-02 1.49573100e+00 1.20101345e+00
1.91592611e-02 6.86533630e-01 -1.48768350e-01 2.73857266e-01
5.86207390e-01 2.81963170e-01 2.28838086e-01 -1.62305164e+00
-3.30148377e-02 -9.20945257e-02 -3.18511456e-01 -1.13684654e+00
1.78172976e-01 -6.97074115e-01 -1.50995925e-01 -1.56595278e+00
3.55391324e-01 -1.07907403e+00 -7.64669180e-01 8.64721239e-01
-2.99028754e-01 1.24549247e-01 -2.96483755e-01 2.59002566e-01
-8.29475462e-01 3.42587769e-01 9.02640104e-01 -1.81373376e-02
-4.33251671e-02 1.00642918e-02 -1.45652020e+00 4.70967770e-01
6.65929019e-01 -6.69428468e-01 -3.68980736e-01 -2.72549897e-01
6.37031376e-01 -6.44586757e-02 -3.83413546e-02 -4.14474815e-01
1.47477165e-01 -4.27827418e-01 5.97714603e-01 -3.53422642e-01
8.99827182e-02 -1.05681789e+00 2.16164719e-02 -1.47537723e-01
-4.00381893e-01 5.48521802e-02 -8.47053677e-02 4.62721080e-01
-2.74519324e-01 -4.06385720e-01 3.16962600e-01 -2.10078061e-01
-9.78941143e-01 3.32822770e-01 1.27759263e-01 1.23695442e-02
8.79314601e-01 3.71379331e-02 -4.37517285e-01 -4.58631292e-02
-5.40773809e-01 4.03732359e-01 3.92795146e-01 5.02742946e-01
9.92972031e-02 -1.66648817e+00 -4.90065336e-01 4.44829501e-02
5.84255040e-01 -1.14022426e-01 1.01506449e-01 2.35261440e-01
4.35950756e-02 4.86112058e-01 -1.90603696e-02 -2.87469234e-02
-1.16847646e+00 1.01782310e+00 -1.77426815e-01 -2.92576373e-01
-3.69169533e-01 5.92027485e-01 1.33625522e-01 -4.24513727e-01
8.62846524e-02 9.09555331e-02 -6.02818668e-01 1.71375498e-01
3.51077199e-01 8.95560831e-02 8.81380662e-02 -7.18851507e-01
-4.48397130e-01 4.95161265e-01 -4.10142094e-01 3.16608667e-01
1.35853982e+00 -3.82606268e-01 -1.66941017e-01 -1.59593690e-02
9.26058590e-01 3.87348980e-02 -7.34931409e-01 -4.77145821e-01
3.24285418e-01 -3.30146968e-01 -6.95092380e-02 -1.13123846e+00
-9.61024582e-01 3.03205729e-01 2.97054440e-01 2.78727293e-01
1.17254341e+00 4.56818759e-01 1.14885223e+00 3.73198211e-01
6.13937378e-01 -1.35488164e+00 -7.05024302e-01 2.36541107e-01
1.88187510e-01 -1.03390300e+00 -9.04956460e-03 -1.14263368e+00
-8.87526214e-01 8.73496354e-01 6.28246486e-01 3.78004849e-01
5.57722449e-01 3.16111565e-01 1.59545749e-01 -2.98547447e-01
-6.05096281e-01 -3.99130672e-01 2.43044659e-01 5.34086525e-01
8.75681162e-01 2.06038415e-01 -5.74470937e-01 1.23580265e+00
-2.00858489e-02 -7.82450754e-03 8.72269422e-02 9.26308990e-01
-4.57858056e-01 -1.62727010e+00 2.00177193e-01 6.02196515e-01
-7.00077415e-01 -2.81639099e-01 -2.86028922e-01 6.77200913e-01
3.14513981e-01 9.84472215e-01 -1.71493143e-01 -5.35938323e-01
6.28281713e-01 2.48690590e-01 1.12108640e-01 -5.53122699e-01
-6.69768810e-01 1.91146314e-01 4.72757667e-01 -3.40853870e-01
-2.80170232e-01 -4.65004563e-01 -1.48842919e+00 -2.72051483e-01
-5.04307210e-01 5.58499634e-01 5.24801731e-01 6.99958563e-01
8.10586333e-01 2.93114811e-01 5.82334220e-01 1.56800866e-01
-6.81779742e-01 -9.81859565e-01 -9.73562539e-01 7.59178579e-01
-2.62905240e-01 -8.36276770e-01 5.32460697e-02 4.39026244e-02] | [9.979758262634277, 6.301032543182373] |
35552e89-f02b-4beb-af6b-0669e4fa9b7f | a-practitioner-s-guide-to-bayesian-inference | 2304.04752 | null | https://arxiv.org/abs/2304.04752v1 | https://arxiv.org/pdf/2304.04752v1.pdf | A Practitioner's Guide to Bayesian Inference in Pharmacometrics using Pumas | This paper provides a comprehensive tutorial for Bayesian practitioners in pharmacometrics using Pumas workflows. We start by giving a brief motivation of Bayesian inference for pharmacometrics highlighting limitations in existing software that Pumas addresses. We then follow by a description of all the steps of a standard Bayesian workflow for pharmacometrics using code snippets and examples. This includes: model definition, prior selection, sampling from the posterior, prior and posterior simulations and predictions, counter-factual simulations and predictions, convergence diagnostics, visual predictive checks, and finally model comparison with cross-validation. Finally, the background and intuition behind many advanced concepts in Bayesian statistics are explained in simple language. This includes many important ideas and precautions that users need to keep in mind when performing Bayesian analysis. Many of the algorithms, codes, and ideas presented in this paper are highly applicable to clinical research and statistical learning at large but we chose to focus our discussions on pharmacometrics in this paper to have a narrower scope in mind and given the nature of Pumas as a software primarily for pharmacometricians. | ['Vijay Ivaturi', 'Chris Rackauckas', 'Julius Krumbiegel', 'Chris Elrod', 'Casey Davis', 'Jose Storopoli', 'Mohamed Tarek'] | 2023-03-31 | null | null | null | null | ['bayesian-inference'] | ['methodology'] | [ 2.68414587e-01 2.11573089e-03 -2.64285654e-01 -5.01551390e-01
-5.67598462e-01 -2.69352704e-01 3.48664433e-01 5.39045632e-01
-2.51738936e-01 1.12263060e+00 -3.37717608e-02 -9.08003449e-01
-7.04261124e-01 -2.18965188e-01 -4.12902772e-01 -9.68455255e-01
-1.14992909e-01 1.03949320e+00 -1.11713797e-01 4.40413296e-01
6.38476670e-01 5.23048639e-01 -1.17951703e+00 -4.86829020e-02
7.15076149e-01 4.23965305e-01 2.56231576e-01 6.59951210e-01
1.03472225e-01 3.29374999e-01 -8.17785144e-01 -3.91509622e-01
-1.89175934e-01 -5.29080212e-01 -4.28646952e-01 -2.53037155e-01
-2.41727382e-02 -5.44543639e-02 9.96935926e-03 9.05920088e-01
9.24822330e-01 9.14872661e-02 1.06446171e+00 -8.10527623e-01
-4.62633073e-02 6.48944795e-01 -5.11669874e-01 4.54643130e-01
7.83633113e-01 5.08582413e-01 5.17576277e-01 -4.86776680e-01
1.63372934e-01 1.37114537e+00 6.99965119e-01 1.08876564e-01
-1.62680507e+00 -8.40475917e-01 1.02520637e-01 -5.17970063e-02
-1.56354964e+00 -5.21976292e-01 6.14558943e-02 -8.24734390e-01
1.11329460e+00 1.85579076e-01 8.98176074e-01 1.04084182e+00
7.76322603e-01 4.23668414e-01 1.46562183e+00 -3.48437697e-01
6.25985682e-01 3.03609550e-01 3.43637943e-01 5.30168302e-02
6.37534857e-01 4.57987905e-01 -5.00982761e-01 -7.00657129e-01
7.17417657e-01 -2.23969012e-01 -4.02275398e-02 -2.81075180e-01
-7.99414158e-01 1.00928307e+00 -5.00058413e-01 -1.50773391e-01
-5.04675508e-01 -6.67104498e-02 2.07056209e-01 -1.56029761e-01
1.90164164e-01 2.56431162e-01 -8.02402198e-01 -2.90358573e-01
-1.10814428e+00 5.64098179e-01 1.16851234e+00 1.03026748e+00
3.64628471e-02 -8.76902863e-02 -2.30167404e-01 7.70497680e-01
1.00509226e+00 3.54471177e-01 2.09304780e-01 -7.19024181e-01
-3.00701290e-01 -2.44497329e-01 3.83291394e-01 -4.83352035e-01
-5.28620720e-01 -4.98417407e-01 -2.67076552e-01 3.16875160e-01
5.17753243e-01 -2.42156237e-01 -9.68555033e-01 1.04350483e+00
1.79403350e-01 -2.31262073e-02 -2.07576737e-01 2.54177630e-01
9.87860799e-01 2.99969375e-01 7.00359702e-01 -7.03131497e-01
1.76488233e+00 -3.35443690e-02 -9.25870955e-01 1.89736821e-02
2.19250247e-01 -1.26278126e+00 5.24937391e-01 7.68714607e-01
-1.22494459e+00 -1.06476592e-02 -1.10401320e+00 2.55656540e-01
-3.61616850e-01 -2.81971604e-01 7.37207770e-01 1.21502936e+00
-3.42201442e-01 9.93973553e-01 -1.14521170e+00 -3.96919519e-01
7.32036114e-01 5.95604300e-01 3.07810634e-01 5.97362965e-02
-1.06390047e+00 1.29722428e+00 4.75619823e-01 -2.37428486e-01
-5.40091038e-01 -1.00899100e+00 -8.04806173e-01 -6.61808997e-03
4.66269292e-02 -9.84416366e-01 1.74908984e+00 -1.20049678e-01
-1.85649455e+00 4.65549380e-01 -3.59073669e-01 -3.68930250e-01
6.73042417e-01 1.23622403e-01 -3.30085784e-01 -1.12715043e-01
-2.30407938e-02 3.37465554e-01 2.18488351e-01 -8.29451740e-01
-6.62755311e-01 -4.63208050e-01 -3.69671375e-01 1.82945848e-01
8.91789913e-01 4.54556644e-01 -4.12979484e-01 -5.08884788e-01
-8.38900283e-02 -6.18386269e-01 -4.70530212e-01 -5.34514129e-01
-3.16961318e-01 -2.90897340e-01 6.21325336e-02 -5.06011605e-01
1.30292976e+00 -1.51960981e+00 -6.26719654e-01 5.95520794e-01
1.68384854e-02 -1.39130816e-01 4.06446248e-01 5.28821945e-01
-4.79243338e-01 -1.81695685e-01 -3.83333325e-01 1.01701111e-01
1.57598168e-01 -8.18173364e-02 -4.84179556e-02 8.72144878e-01
-3.30798924e-01 7.28778899e-01 -1.09017098e+00 -3.89886320e-01
7.17647076e-01 6.41098499e-01 -4.06571418e-01 -8.62608179e-02
-2.20826879e-01 6.67750359e-01 -3.53687704e-01 7.07217872e-01
9.42334175e-01 -3.76887381e-01 7.36717582e-01 -2.84813959e-02
-1.96043655e-01 8.03885877e-01 -1.29142308e+00 1.18457437e+00
3.00392020e-03 9.61517617e-02 -2.22987354e-01 -7.58350849e-01
6.01146102e-01 4.47044134e-01 6.61221087e-01 -1.32663667e-01
7.56175071e-02 2.38554060e-01 4.14119303e-01 -1.89398929e-01
-1.75950229e-01 -6.02413356e-01 6.27668083e-01 4.86239493e-01
1.86267868e-03 -3.60682517e-01 4.22073722e-01 1.32106066e-01
5.48564494e-01 6.55200720e-01 1.08645046e+00 -4.54321384e-01
2.03730956e-01 -1.32373855e-01 3.25398624e-01 1.08803511e+00
-1.08388931e-01 1.99643329e-01 7.01683283e-01 -9.57251713e-02
-6.25271797e-01 -1.02601027e+00 -1.19551337e+00 6.20182633e-01
-4.22190279e-01 -7.72962630e-01 -4.01800573e-01 -2.59545147e-01
3.25727403e-01 8.62542570e-01 -5.69216490e-01 4.23889667e-01
2.08054081e-01 -1.70696688e+00 2.70000026e-02 2.72545636e-01
-4.00558263e-01 -7.86822736e-01 -5.35882711e-01 5.51780581e-01
4.60671514e-01 -6.38542831e-01 1.00031912e-01 5.89419186e-01
-1.20217800e+00 -1.36438191e+00 -8.04214060e-01 -2.16440335e-01
1.96267754e-01 -1.54091582e-01 1.08311367e+00 -5.19919693e-01
-3.26203376e-01 2.12370142e-01 2.13247672e-01 -1.23990965e+00
-5.42706311e-01 -4.49405938e-01 2.10158527e-02 -9.38046932e-01
1.13859689e+00 -6.98866904e-01 -8.11216235e-01 3.37582767e-01
-4.58436966e-01 -2.18591332e-01 5.99903286e-01 5.44286251e-01
5.28542578e-01 -1.39966592e-01 3.76475811e-01 -1.34254980e+00
7.85605729e-01 -5.41986227e-01 -1.15269589e+00 5.99087588e-02
-1.15822089e+00 -2.31448904e-01 -1.96883589e-01 -1.34249777e-02
-9.91092086e-01 5.41375764e-02 -4.34280604e-01 2.80373126e-01
-3.78581107e-01 8.51680636e-01 -2.97867861e-02 1.25611618e-01
8.88115644e-01 -2.62600869e-01 3.07740957e-01 -7.66487777e-01
8.58063810e-03 8.45027208e-01 -1.48768470e-01 -4.95914996e-01
-1.34444267e-01 2.78860658e-01 2.53054202e-01 -6.87430084e-01
-5.28090596e-01 -6.75921857e-01 -4.65872407e-01 6.86812699e-02
4.66241509e-01 -7.42049694e-01 -1.50523806e+00 1.71609432e-01
-8.18142176e-01 -2.17640221e-01 -9.86764878e-02 1.13212895e+00
-7.78550088e-01 4.75564480e-01 -4.78071958e-01 -1.06405735e+00
-3.94065026e-03 -1.59649324e+00 5.43870032e-01 2.39619106e-01
-7.64971316e-01 -1.24341738e+00 4.04233664e-01 3.74911755e-01
1.39895186e-01 1.89877212e-01 9.51211810e-01 -7.37450361e-01
-5.02615534e-02 -2.58302599e-01 7.19431192e-02 -1.25288188e-01
1.35332480e-01 2.99181819e-01 -1.25408340e+00 -1.65481776e-01
1.37572298e-02 3.53608318e-02 6.16578221e-01 1.46411550e+00
1.17937756e+00 -8.36028680e-02 -7.90507495e-01 3.68569374e-01
1.07069421e+00 6.70692682e-01 5.53525150e-01 6.00800738e-02
5.53158186e-02 6.96786761e-01 5.04883051e-01 7.29485273e-01
2.55260244e-02 5.86776972e-01 -6.76830858e-03 1.00940034e-01
3.92860413e-01 -3.99274193e-02 8.64666179e-02 2.11038306e-01
-1.67424098e-01 9.78832841e-02 -7.43159771e-01 -3.79623994e-02
-1.80021453e+00 -7.64727771e-01 -4.45130587e-01 2.59321904e+00
1.08525062e+00 8.27911869e-02 5.47421455e-01 -1.91957593e-01
7.28754282e-01 -3.73840749e-01 -6.18544281e-01 -4.13541168e-01
2.18474001e-01 6.79029644e-01 7.91666508e-01 8.02288115e-01
-1.06866515e+00 5.05296469e-01 8.32013321e+00 9.89723742e-01
-4.01192218e-01 -3.01881172e-02 7.90529191e-01 -6.55189455e-02
-7.56221777e-03 2.57812619e-01 -9.49402869e-01 4.87144351e-01
1.50399101e+00 -1.39720410e-01 1.81751847e-01 4.71770734e-01
9.86494243e-01 -8.24082434e-01 -1.20462024e+00 9.57222521e-01
-4.37003523e-01 -1.34816122e+00 -1.84665784e-01 3.05557579e-01
5.27882814e-01 1.12683222e-01 -1.05437562e-01 2.56801508e-02
8.30983400e-01 -1.42058766e+00 3.43648344e-01 4.76370335e-01
6.09744191e-01 -7.44254827e-01 7.53372073e-01 -2.30568260e-01
-3.99754614e-01 2.98844606e-01 -5.81768632e-01 -1.81728527e-01
3.50598514e-01 1.04300141e+00 -9.83706057e-01 8.25389922e-01
5.31627297e-01 9.16911185e-01 -1.45615056e-01 1.81118679e+00
-2.17943326e-01 5.17962575e-01 -3.99668187e-01 1.08654117e-02
-5.27906939e-02 -4.30703074e-01 3.34453285e-01 1.45521760e+00
9.28390995e-02 2.18441263e-02 -1.96421579e-01 7.46082783e-01
6.09433651e-01 2.79886752e-01 -3.00100863e-01 -1.38506994e-01
5.12713492e-01 7.80782938e-01 -8.63110006e-01 -4.28605676e-01
-5.55099249e-01 1.15429908e-02 -5.47330260e-01 4.49808896e-01
-4.66286927e-01 -4.13802534e-01 6.38163328e-01 1.61688760e-01
6.42018616e-02 3.30043525e-01 -7.10238397e-01 -5.48606038e-01
-7.28883684e-01 -1.04207873e+00 9.16235030e-01 -7.56613195e-01
-1.16202617e+00 -3.15856129e-01 8.11378658e-01 -8.01146448e-01
-1.60758227e-01 -1.15587759e+00 -2.60757685e-01 1.29378903e+00
-1.07943082e+00 -4.49797869e-01 3.41808975e-01 6.61659837e-02
2.26912782e-01 -8.33884929e-04 1.00667953e+00 -2.40417738e-02
-4.66877997e-01 1.87930062e-01 6.21876776e-01 -5.05413890e-01
9.39261794e-01 -1.37650251e+00 2.04442039e-01 8.07718188e-02
-3.63970786e-01 1.31807077e+00 1.29317915e+00 -1.03255451e+00
-9.44905221e-01 -3.71875942e-01 5.90326428e-01 -5.35673976e-01
7.87976682e-01 1.28725097e-01 -5.00355124e-01 6.83667004e-01
2.50302047e-01 -9.73048747e-01 1.30761111e+00 5.16709447e-01
1.47973508e-01 2.03734204e-01 -1.14232755e+00 6.63032651e-01
3.52584302e-01 -1.65559035e-02 -5.99523604e-01 9.68349457e-01
-2.52702743e-01 -6.32316887e-01 -1.27400982e+00 2.17273697e-01
1.18426847e+00 -7.89824307e-01 8.46378326e-01 -7.49110639e-01
-2.30272964e-01 -4.86926615e-01 3.46031368e-01 -8.81637275e-01
-1.92560539e-01 -1.11713481e+00 1.03236727e-01 3.38543445e-01
5.71132302e-01 -8.66939425e-01 6.34524465e-01 6.85980797e-01
-9.76958275e-02 -6.21619821e-01 -8.23674977e-01 -5.15246809e-01
2.94208795e-01 -6.66529834e-01 4.06010509e-01 7.06190825e-01
7.25409150e-01 3.34890664e-01 -5.37131354e-03 -2.08747804e-01
8.44874084e-01 -6.40740320e-02 5.47484994e-01 -1.38249779e+00
-5.51293433e-01 -5.65778553e-01 -3.13811421e-01 -1.08746576e+00
-5.19069910e-01 -6.57069564e-01 -1.44668892e-01 -1.58184350e+00
6.38609290e-01 5.72611205e-02 -2.53715634e-01 1.79061770e-01
-1.43355832e-01 -1.73307192e-02 -4.97446656e-01 1.00987993e-01
1.57516062e-01 -6.03577644e-02 1.11105251e+00 -8.79556220e-03
-4.77341205e-01 5.80895901e-01 -1.09145129e+00 6.75699472e-01
8.85944068e-01 -7.91010439e-01 -4.49651539e-01 5.12785316e-01
3.19700181e-01 -2.91938428e-02 1.22822754e-01 -3.44922006e-01
1.04308024e-01 -4.73977178e-01 8.14898074e-01 -8.20927978e-01
2.18175232e-01 -5.13541341e-01 5.14963627e-01 7.04788923e-01
-1.14896921e-02 -2.76600197e-02 5.28499186e-01 6.11618698e-01
3.75939995e-01 -6.37086511e-01 1.06357038e+00 -4.37590539e-01
2.56086528e-01 9.74997059e-02 -1.00408638e+00 -4.01815921e-01
8.32326114e-01 -4.17203069e-01 -2.32417248e-02 -5.98542988e-01
-1.44261217e+00 3.67388070e-01 3.41635704e-01 -2.99665600e-01
1.95642427e-01 -5.98341227e-01 -6.27029359e-01 -1.16434284e-01
1.68102667e-01 -3.31904531e-01 3.55706632e-01 1.47222877e+00
-6.18481398e-01 1.01295567e+00 2.67069433e-02 -5.95317066e-01
-1.33084941e+00 3.43683630e-01 4.50313807e-01 -5.43514565e-02
-4.68372524e-01 4.16058034e-01 2.69545674e-01 -9.11997072e-03
3.54533315e-01 -3.42691869e-01 -2.73810208e-01 -5.13978228e-02
5.80073297e-01 5.12649059e-01 3.49456012e-01 -1.50768012e-01
-5.19992471e-01 1.90634176e-01 -3.30769628e-01 -2.86303580e-01
1.26832163e+00 -1.33619651e-01 -2.72081569e-02 5.52052498e-01
4.26373661e-01 -1.10536315e-01 -1.18668699e+00 2.23659009e-01
8.80292356e-02 -3.78140122e-01 1.67482257e-01 -1.35356092e+00
-2.93455392e-01 7.23459721e-01 5.60514390e-01 -1.77901894e-01
3.19800526e-01 1.58974435e-02 -2.45480940e-01 4.73017991e-02
-1.40937313e-01 -1.11274385e+00 -6.07797027e-01 2.37044379e-01
6.46697462e-01 -7.18212903e-01 9.90860283e-01 -3.56099278e-01
-5.27424216e-01 1.14010394e+00 -1.84964299e-01 3.78210962e-01
1.21477866e+00 1.22793764e-01 2.07465645e-02 -5.37666678e-01
-5.64448237e-01 -2.71423403e-02 2.71217525e-01 9.62861955e-01
9.97878909e-01 1.35654151e-01 -7.11191475e-01 5.00487745e-01
-2.11166799e-01 2.33389899e-01 4.85974640e-01 9.93021011e-01
-4.03034866e-01 -1.49826574e+00 -6.42365694e-01 8.38204086e-01
-8.73432398e-01 -3.36289704e-01 -9.24430117e-02 8.78861129e-01
-2.92862445e-01 1.21686089e+00 -1.10354565e-01 5.05636156e-01
1.50524512e-01 2.25493431e-01 9.21136081e-01 -8.09525192e-01
-3.73338580e-01 1.00130844e+00 3.29282701e-01 -2.16704473e-01
-5.15940309e-01 -1.04064775e+00 -7.71068513e-01 -4.57932621e-01
-4.07322466e-01 5.46773255e-01 9.27288830e-01 9.33461428e-01
1.06322840e-01 5.28350830e-01 -1.36595920e-01 -6.39964163e-01
-4.78569537e-01 -1.37064481e+00 -9.98159528e-01 -4.69496161e-01
-1.65433120e-02 -8.60852897e-01 -3.81140620e-01 3.87709104e-02] | [6.508890151977539, 4.065654277801514] |
dab25d95-af1a-4ad9-a0f4-8f967aa3639a | fedgrad-mitigating-backdoor-attacks-in | 2305.00328 | null | https://arxiv.org/abs/2305.00328v1 | https://arxiv.org/pdf/2305.00328v1.pdf | FedGrad: Mitigating Backdoor Attacks in Federated Learning Through Local Ultimate Gradients Inspection | Federated learning (FL) enables multiple clients to train a model without compromising sensitive data. The decentralized nature of FL makes it susceptible to adversarial attacks, especially backdoor insertion during training. Recently, the edge-case backdoor attack employing the tail of the data distribution has been proposed as a powerful one, raising questions about the shortfall in current defenses' robustness guarantees. Specifically, most existing defenses cannot eliminate edge-case backdoor attacks or suffer from a trade-off between backdoor-defending effectiveness and overall performance on the primary task. To tackle this challenge, we propose FedGrad, a novel backdoor-resistant defense for FL that is resistant to cutting-edge backdoor attacks, including the edge-case attack, and performs effectively under heterogeneous client data and a large number of compromised clients. FedGrad is designed as a two-layer filtering mechanism that thoroughly analyzes the ultimate layer's gradient to identify suspicious local updates and remove them from the aggregation process. We evaluate FedGrad under different attack scenarios and show that it significantly outperforms state-of-the-art defense mechanisms. Notably, FedGrad can almost 100% correctly detect the malicious participants, thus providing a significant reduction in the backdoor effect (e.g., backdoor accuracy is less than 8%) while not reducing the main accuracy on the primary task. | ['Truong Thao Nguyen', 'Phi Le Nguyen', 'Thanh Hung Nguyen', 'Huy Hieu Pham', 'Kok-Seng Wong', 'Anh Duy Nguyen', 'Thuy Dung Nguyen'] | 2023-04-29 | null | null | null | null | ['backdoor-attack'] | ['adversarial'] | [-2.60750860e-01 -2.33094037e-01 -3.08999687e-01 -7.87893757e-02
-1.11333776e+00 -1.26878226e+00 6.52031958e-01 -5.34204952e-02
-3.22878987e-01 3.41026932e-01 -1.97969630e-01 -7.93323994e-01
-1.57312810e-01 -7.08038092e-01 -7.59951651e-01 -9.10567284e-01
-3.58366907e-01 1.53041974e-01 4.36683506e-01 -2.45217964e-01
2.22486928e-01 7.87402451e-01 -9.37569559e-01 5.60055733e-01
7.02439725e-01 1.13052940e+00 -6.79265738e-01 5.33156574e-01
1.31709859e-01 7.07562327e-01 -9.59165037e-01 -9.24520016e-01
6.17494524e-01 2.47470550e-02 -4.57689583e-01 -5.17076135e-01
6.52179182e-01 -7.36838639e-01 -5.72720230e-01 1.07313716e+00
5.46932876e-01 -3.85762393e-01 1.43193811e-01 -1.52421701e+00
-3.40584934e-01 7.79524386e-01 -6.75392151e-01 3.71481687e-01
5.92391454e-02 4.11863178e-01 9.90488350e-01 -6.63910449e-01
4.27245528e-01 1.28471541e+00 6.76156163e-01 6.86591864e-01
-1.22388124e+00 -1.03607750e+00 4.52943861e-01 7.03950599e-02
-9.91675317e-01 -3.55267555e-01 5.85038841e-01 3.48390974e-02
5.57448685e-01 8.05326939e-01 -1.69012800e-01 1.60070086e+00
2.91379273e-01 7.52353072e-01 1.03390872e+00 1.42990857e-01
4.07843679e-01 3.20321113e-01 1.64370626e-01 5.28496385e-01
4.91223454e-01 4.24740702e-01 -5.70074201e-01 -1.03833818e+00
2.33532369e-01 1.76683426e-01 -4.35486495e-01 -2.62687683e-01
-3.94164979e-01 1.06143880e+00 4.54267710e-01 -3.04125231e-02
-1.51495546e-01 -9.29937046e-03 6.97548270e-01 6.03368759e-01
3.23127717e-01 3.02649647e-01 -6.50791407e-01 6.88462555e-02
-6.73845112e-01 3.79474670e-01 1.13455343e+00 3.30211431e-01
4.61666793e-01 -5.45174591e-02 -3.33871007e-01 2.22964838e-01
2.07092583e-01 5.08236170e-01 1.40911639e-01 -6.53196394e-01
8.01094353e-01 6.22596264e-01 -1.59944013e-01 -9.90087926e-01
-5.16253561e-02 -8.76653492e-01 -5.52846730e-01 3.69414568e-01
5.91239035e-01 -5.85664213e-01 -4.62895215e-01 1.98967946e+00
7.18770146e-01 1.39328033e-01 -1.07434586e-01 8.00136685e-01
1.07926868e-01 2.60610402e-01 -1.55656829e-01 -9.04163867e-02
1.27763498e+00 -7.66557097e-01 -3.42803508e-01 -3.12866539e-01
6.62881911e-01 -4.73989964e-01 9.56179202e-01 6.70617878e-01
-7.28000522e-01 2.74914622e-01 -9.86397207e-01 5.46620548e-01
-4.95002657e-01 -6.08482838e-01 8.21329951e-01 1.26513040e+00
-8.00063312e-01 4.32076365e-01 -7.39622056e-01 8.29190090e-02
6.49379194e-01 4.63581532e-01 -4.61982310e-01 -5.49217351e-02
-1.21416831e+00 5.37160933e-01 -5.62165976e-02 -2.34441176e-01
-1.30534565e+00 -1.02072108e+00 -3.30467582e-01 3.32491577e-01
8.06981981e-01 -3.92834574e-01 1.20544231e+00 -3.45453680e-01
-1.18765366e+00 6.01680100e-01 3.01655293e-01 -7.97831416e-01
1.02560294e+00 -4.21407431e-01 -4.40597057e-01 9.46698114e-02
-2.79453725e-01 -4.24951732e-01 1.15147960e+00 -1.26637554e+00
-5.78386247e-01 -7.78445065e-01 2.83891767e-01 -2.82494009e-01
-8.96766782e-01 1.69844851e-01 -2.27033541e-01 -5.28932929e-01
-1.38257876e-01 -8.65658283e-01 -1.34189725e-01 -2.54478324e-02
-6.45613134e-01 -4.46037687e-02 1.54007173e+00 -1.87899277e-01
1.27750647e+00 -2.03673530e+00 -4.66312855e-01 4.34734702e-01
3.80625993e-01 7.06532300e-01 -1.85590222e-01 6.93587422e-01
1.84092730e-01 2.55511373e-01 5.90354167e-02 -4.13752586e-01
1.83384061e-01 -1.59771256e-02 -9.31048989e-01 7.95687139e-01
-1.86656192e-01 5.36382437e-01 -6.43586636e-01 -6.63792118e-02
-1.39050603e-01 5.55553734e-01 -8.25597882e-01 3.32254320e-01
-2.14653105e-01 1.59258917e-01 -6.01740658e-01 7.83041239e-01
1.04600179e+00 -1.52907044e-01 3.50824893e-01 7.24519566e-02
1.74100429e-01 3.34181070e-01 -1.14540684e+00 8.90048206e-01
-3.78569543e-01 2.77282953e-01 7.16572404e-01 -4.78686839e-01
7.18807817e-01 1.69363990e-01 3.26695353e-01 -5.31685233e-01
3.27465415e-01 2.84695834e-01 -2.83471078e-01 -7.45915696e-02
8.96139815e-03 2.67672777e-01 -1.43491253e-01 7.92712271e-01
-2.42391348e-01 6.73708379e-01 -2.26587817e-01 6.04860544e-01
1.53762555e+00 -5.06572902e-01 -2.98641980e-01 -1.09943658e-01
6.20952070e-01 -3.75852495e-01 5.75156868e-01 1.26289475e+00
-3.64874482e-01 1.50828496e-01 8.15370619e-01 -4.46238577e-01
-3.09606344e-01 -1.27305973e+00 5.61016724e-02 1.50346339e+00
1.49414659e-01 -5.76294720e-01 -9.43552792e-01 -1.55061615e+00
5.24039328e-01 5.09229839e-01 -6.88180983e-01 -5.87152123e-01
-6.30411267e-01 -8.24462235e-01 1.14522696e+00 5.87235615e-02
6.05814219e-01 -3.86394352e-01 -3.86403292e-01 -7.17788935e-02
1.14134522e-02 -1.01846755e+00 -6.68843567e-01 1.92749202e-01
-5.98271728e-01 -1.54518914e+00 -1.23750538e-01 -5.24014421e-02
4.80160326e-01 6.18157983e-01 8.25589061e-01 1.09531447e-01
-2.18707666e-01 2.80752450e-01 -7.19891340e-02 -2.54984021e-01
-4.76617664e-01 1.10565424e-01 2.41217777e-01 4.07885164e-01
3.03509831e-01 -6.37565255e-01 -7.25702882e-01 5.28757453e-01
-9.99752522e-01 -9.56714451e-01 5.11612475e-01 6.96936667e-01
7.02413544e-02 1.22550339e-01 6.82837844e-01 -1.22184598e+00
8.26405048e-01 -6.93507135e-01 -7.40225136e-01 3.55410635e-01
-6.10440254e-01 4.12992910e-02 1.15077150e+00 -7.13682532e-01
-9.65660036e-01 -3.87042552e-01 -1.37756705e-01 -6.89631283e-01
1.67599514e-01 -6.06361292e-02 -3.43809903e-01 -5.73706746e-01
9.77789938e-01 2.48232946e-01 1.17263552e-02 -6.78157687e-01
4.51985300e-01 5.83156288e-01 5.64463437e-01 -6.54593527e-01
1.31906402e+00 8.58285069e-01 -7.90796801e-02 -4.58018869e-01
-8.59593213e-01 -1.42349377e-01 2.14811161e-01 4.26169625e-03
1.16385907e-01 -6.01980388e-01 -1.21552145e+00 6.28571033e-01
-9.08715904e-01 -1.01872869e-01 4.32220586e-02 7.10347071e-02
2.72949543e-02 6.26378596e-01 -9.21928108e-01 -8.99784744e-01
-9.09180224e-01 -9.66688573e-01 6.02469742e-01 1.70999750e-01
2.08532929e-01 -9.09483850e-01 8.31895173e-02 4.67648625e-01
7.29477644e-01 4.32268918e-01 7.43570626e-01 -1.34884953e+00
-5.72006702e-01 -6.80150628e-01 -1.66037846e-02 3.04877549e-01
-1.40206441e-02 8.82723555e-03 -1.22985816e+00 -9.84902620e-01
5.40282547e-01 -3.47527027e-01 8.49554121e-01 -2.10177615e-01
1.18298614e+00 -9.08314228e-01 -2.94683993e-01 8.71539533e-01
1.27855086e+00 -1.81430414e-01 3.34556162e-01 4.18363243e-01
3.90954018e-01 4.81239170e-01 3.96190166e-01 7.04693258e-01
6.04534559e-02 4.85563964e-01 1.12941229e+00 5.72854429e-02
3.62499595e-01 -4.12307650e-01 7.42300570e-01 -3.92524451e-02
6.33984387e-01 -2.73628265e-01 -5.11414111e-01 3.80263254e-02
-1.53205633e+00 -8.32372963e-01 3.00676841e-02 2.37009716e+00
6.90407097e-01 5.00139117e-01 3.47368240e-01 1.74270794e-01
6.35183811e-01 4.87035632e-01 -7.25626051e-01 -6.24437690e-01
-9.89636257e-02 -3.82448956e-02 8.30267727e-01 2.87413865e-01
-1.13592589e+00 8.06943119e-01 5.89733839e+00 9.19653952e-01
-1.29224086e+00 2.81372130e-01 4.99298871e-01 -3.93663168e-01
-1.07325107e-01 1.43272104e-02 -1.06735146e+00 6.33927166e-01
9.37371194e-01 -6.98969141e-02 4.30357158e-01 1.22338772e+00
-6.22250885e-02 3.52876037e-01 -9.88776147e-01 4.85091716e-01
-5.62299751e-02 -1.37056434e+00 4.26863320e-02 5.30826807e-01
5.01101017e-01 2.24612564e-01 5.60444772e-01 3.89567673e-01
6.16996825e-01 -6.97358847e-01 3.72209221e-01 -1.41096264e-01
5.37686586e-01 -1.10754263e+00 4.63765353e-01 5.59409201e-01
-8.34874332e-01 -6.80292726e-01 -1.36139199e-01 3.82732004e-01
-1.97011948e-01 5.34021199e-01 -5.33874273e-01 4.05367225e-01
7.65593350e-01 -2.87136078e-01 -5.52828610e-01 7.05729365e-01
-1.39121667e-01 8.59745204e-01 -6.13491833e-01 3.02952141e-01
5.35317957e-01 2.13593990e-01 9.36770737e-01 1.08935964e+00
-1.60321668e-01 -3.35980356e-01 3.92596722e-02 7.67129958e-01
-4.11176831e-01 -1.85363099e-01 -6.09072387e-01 9.55316573e-02
9.25580561e-01 1.29089713e+00 -1.46028489e-01 -3.26413922e-02
-1.38964310e-01 7.43826389e-01 4.00808990e-01 1.97458848e-01
-9.38531101e-01 -3.57366621e-01 1.23125601e+00 1.95394263e-01
1.47129521e-01 9.30047259e-02 -1.63503647e-01 -1.13277888e+00
1.67103812e-01 -1.41198909e+00 9.44045961e-01 2.65346497e-01
-1.50971603e+00 7.32205987e-01 -5.05910099e-01 -9.08724666e-01
-1.27711669e-01 -4.39970016e-01 -1.11837864e+00 4.87051278e-01
-1.39408886e+00 -1.05353117e+00 3.04443520e-02 1.14181244e+00
6.73632771e-02 -2.99551874e-01 6.56031013e-01 4.19428170e-01
-1.05153966e+00 1.49540436e+00 3.57210726e-01 2.09998965e-01
8.56633246e-01 -9.88502681e-01 4.95918185e-01 1.20298910e+00
8.64194483e-02 9.52745914e-01 6.73578441e-01 -5.84190369e-01
-1.88531208e+00 -1.10592115e+00 3.02436590e-01 -6.73928916e-01
8.17981422e-01 -7.28682280e-01 -1.04019022e+00 5.66309333e-01
-3.47174853e-01 3.64116907e-01 6.69622719e-01 -7.95678142e-03
-1.21592939e+00 -4.01651740e-01 -1.67830038e+00 6.85465753e-01
7.24660516e-01 -6.59837544e-01 -9.15133357e-02 5.08934498e-01
8.34076405e-01 -2.49993011e-01 -6.13752306e-01 1.39608517e-01
4.77304369e-01 -1.44195306e+00 1.00006771e+00 -8.98186684e-01
-2.42820576e-01 1.10746577e-01 -8.06993693e-02 -8.79218638e-01
-8.00590515e-02 -1.35866344e+00 -1.04429901e+00 1.40563810e+00
1.78283155e-01 -1.18222058e+00 1.18490458e+00 4.59375113e-01
3.73092145e-01 -8.46623600e-01 -1.12776697e+00 -9.96284902e-01
2.00928763e-01 -4.07031268e-01 8.14145982e-01 8.78609776e-01
-1.33181363e-01 -1.96743071e-01 -4.77289677e-01 6.03433669e-01
1.14317489e+00 -4.44118939e-02 9.84951794e-01 -1.01990557e+00
-5.08859575e-01 -3.63640696e-01 6.80973008e-02 -7.73228347e-01
1.78463757e-02 -6.15066111e-01 -6.89296544e-01 -7.43331432e-01
-1.23515680e-01 -3.83290172e-01 -5.22882998e-01 6.94703817e-01
-8.20827186e-02 2.01853827e-01 4.02141482e-01 3.44157130e-01
-4.35946524e-01 2.12557957e-01 7.26122737e-01 -5.84342442e-02
-1.82280600e-01 4.69792068e-01 -1.11304283e+00 5.72465777e-01
8.00193012e-01 -7.15391040e-01 -3.81771088e-01 -2.03085586e-01
3.19035575e-02 -1.96869090e-01 3.67072016e-01 -6.96678579e-01
3.80520642e-01 -1.98508754e-01 4.16367576e-02 -2.85879850e-01
1.74817324e-01 -8.85705531e-01 -1.83610216e-01 7.73133457e-01
-1.86334193e-01 9.27754343e-02 -1.37112429e-02 7.61523247e-01
2.16470510e-01 3.07606459e-01 9.68486190e-01 1.14827439e-01
1.46139562e-02 4.88141507e-01 -1.42559513e-01 2.67190218e-01
1.21188056e+00 -3.22607495e-02 -1.07577193e+00 -3.10406983e-01
-2.38280579e-01 4.09664541e-01 4.16173488e-01 5.10706604e-01
4.03890967e-01 -8.61683249e-01 -6.62547231e-01 5.55035412e-01
-2.05050826e-01 -5.12409925e-01 3.10592890e-01 8.47610414e-01
-2.02783674e-01 3.12539577e-01 1.47014052e-01 -2.30868071e-01
-1.47043014e+00 7.20366061e-01 6.03023231e-01 -6.43325567e-01
-6.19760156e-01 9.17674065e-01 8.38850737e-02 -4.56512213e-01
6.60443425e-01 4.05719817e-01 2.26041600e-01 -8.79405662e-02
9.47982371e-01 6.41601145e-01 3.08275938e-01 -3.23255628e-01
-7.15814054e-01 -6.44380823e-02 -5.61847925e-01 2.35630184e-01
1.08533525e+00 2.15350747e-01 -6.48734048e-02 -4.11762834e-01
1.37478375e+00 4.67548400e-01 -1.40299773e+00 -4.16593790e-01
-3.97805328e-04 -8.49964082e-01 6.75698221e-02 -9.58657086e-01
-1.48302245e+00 6.12976730e-01 5.17300010e-01 6.11247838e-01
9.88937199e-01 -2.29817435e-01 1.17092681e+00 3.38323772e-01
4.33511704e-01 -7.00059772e-01 1.84794962e-01 3.32240880e-01
5.10577500e-01 -9.78859365e-01 -1.08600676e-01 -3.34892035e-01
-4.33926910e-01 8.06617677e-01 8.26856077e-01 -1.83799013e-01
5.32328010e-01 5.59353292e-01 1.99305132e-01 -2.29292169e-01
-1.16775274e+00 4.76513535e-01 -7.97777176e-02 4.41977412e-01
-3.62929046e-01 -1.64262831e-01 -3.82998474e-02 7.59498894e-01
7.53691420e-03 -6.43865705e-01 2.50937611e-01 9.92683709e-01
-3.10547829e-01 -1.17333794e+00 -6.73291981e-01 2.83150613e-01
-1.05820072e+00 3.55689764e-01 -8.04358065e-01 6.02237582e-01
-1.73808396e-01 1.10769749e+00 -3.65239173e-01 -7.68216312e-01
3.41012090e-01 -1.94669440e-01 -8.10573027e-02 -1.37133837e-01
-1.36923194e+00 -2.56834120e-01 -1.10715613e-01 -9.64951634e-01
6.19933248e-01 -1.65632114e-01 -8.95591438e-01 -9.63560879e-01
-5.73230207e-01 5.09970188e-01 6.42119825e-01 6.37819409e-01
6.41947448e-01 1.16693765e-01 1.41802585e+00 -4.61989462e-01
-1.69684803e+00 -3.07813883e-01 -6.27416372e-01 2.54352003e-01
4.99220729e-01 -4.56391931e-01 -1.02176082e+00 -1.00283587e+00] | [5.6993937492370605, 7.255173206329346] |
abc278ff-309c-40f0-8205-dd6a94cc96d7 | wearing-masks-implies-refuting-trump-towards | 2303.12029 | null | https://arxiv.org/abs/2303.12029v1 | https://arxiv.org/pdf/2303.12029v1.pdf | Wearing Masks Implies Refuting Trump?: Towards Target-specific User Stance Prediction across Events in COVID-19 and US Election 2020 | People who share similar opinions towards controversial topics could form an echo chamber and may share similar political views toward other topics as well. The existence of such connections, which we call connected behavior, gives researchers a unique opportunity to predict how one would behave for a future event given their past behaviors. In this work, we propose a framework to conduct connected behavior analysis. Neural stance detection models are trained on Twitter data collected on three seemingly independent topics, i.e., wearing a mask, racial equality, and Trump, to detect people's stance, which we consider as their online behavior in each topic-related event. Our results reveal a strong connection between the stances toward the three topical events and demonstrate the power of past behaviors in predicting one's future behavior. | ['Jisun An', 'Wei Gao', 'Haewoon Kwak', 'Hong Zhang'] | 2023-03-21 | null | null | null | null | ['stance-detection'] | ['natural-language-processing'] | [-4.50509399e-01 8.09714124e-02 -7.07924306e-01 -8.74449492e-01
-9.31686983e-02 -5.63998699e-01 9.92181540e-01 4.07482117e-01
-2.27221519e-01 6.79762959e-01 9.06784177e-01 -3.73074710e-01
4.66148376e-01 -1.15457678e+00 -4.42321181e-01 -2.07100362e-01
9.26198810e-02 2.04541329e-02 -1.20037273e-02 -3.81508917e-01
2.92074651e-01 -2.92432070e-01 -1.01374340e+00 3.58750850e-01
9.57357824e-01 5.57390630e-01 -7.28680313e-01 -1.27538303e-02
-1.62944481e-01 9.80037212e-01 -8.57609451e-01 -9.59140658e-01
4.22277264e-02 -2.96427101e-01 -5.21362841e-01 -1.57122880e-01
7.42638290e-01 -3.12870890e-01 -4.31590021e-01 1.14922690e+00
2.28085816e-02 -9.72418636e-02 6.11974359e-01 -1.11731935e+00
-6.65817857e-01 1.26053870e+00 -1.07404435e+00 6.60438120e-01
2.21255496e-01 5.20010665e-02 1.38481283e+00 -2.84541726e-01
8.19974244e-01 1.50347078e+00 8.08300018e-01 1.58440128e-01
-1.32577276e+00 -1.02046204e+00 8.71888697e-01 -9.68999788e-03
-8.15100729e-01 -3.23835105e-01 1.22603297e+00 -7.92710304e-01
-6.86004162e-02 1.77760631e-01 1.08049262e+00 1.60635090e+00
4.16149735e-01 7.92994380e-01 1.55203414e+00 2.30546311e-01
2.10011989e-01 4.92962569e-01 1.03176975e+00 3.71273488e-01
6.28927112e-01 -2.60097146e-01 -6.79063082e-01 -6.17044985e-01
-9.31911319e-02 6.34724945e-02 -7.45495558e-02 3.67008507e-01
-1.03742790e+00 1.13811040e+00 8.05973172e-01 4.95228142e-01
-2.78825223e-01 -1.56914264e-01 2.21906081e-01 3.85522187e-01
1.23391807e+00 4.40835863e-01 -6.04590885e-02 9.86336321e-02
-9.42263365e-01 3.66230905e-01 1.06612170e+00 4.43347394e-02
9.15885150e-01 -1.38246953e-01 -2.96194136e-01 6.44412339e-01
2.61194855e-01 5.63970029e-01 4.26551044e-01 -5.66830337e-01
4.34739828e-01 8.17519307e-01 1.76520213e-01 -1.78791869e+00
-5.73032081e-01 -6.03924572e-01 -6.07458770e-01 -3.68859887e-01
5.61729252e-01 -7.76073396e-01 -4.48193520e-01 2.09507990e+00
4.74581093e-01 1.92817956e-01 -2.96040863e-01 6.51225269e-01
6.59848928e-01 6.85124636e-01 2.63441294e-01 -2.83465594e-01
1.35489106e+00 -1.96050137e-01 -7.41640866e-01 -2.44193688e-01
3.64727497e-01 -4.81690526e-01 8.05152297e-01 1.50377676e-01
-8.35122705e-01 -3.77063215e-01 -8.93005192e-01 2.11498499e-01
-4.28508312e-01 -3.46303940e-01 7.95323789e-01 6.48509443e-01
-7.16714382e-01 6.42401576e-01 -4.93085623e-01 -4.09388214e-01
6.22999191e-01 -2.24392861e-01 3.57371330e-01 5.01177728e-01
-1.51399899e+00 6.02992356e-01 -2.81870633e-01 -8.59167874e-02
-5.08228302e-01 -6.36460960e-01 -3.87320697e-01 -2.55878001e-01
3.66965652e-01 -3.68548959e-01 1.06301630e+00 -1.09871924e+00
-1.01314974e+00 1.29034400e+00 -4.96552289e-01 -5.88400722e-01
6.77514493e-01 -3.30521256e-01 -8.45846415e-01 -3.40891033e-01
5.91934085e-01 2.06870049e-01 7.76988864e-01 -9.88196135e-01
-4.35064256e-01 -6.34135544e-01 1.53758273e-01 1.18187711e-01
-7.99464226e-01 5.01502827e-02 3.46203476e-01 -5.87650537e-01
1.77394569e-01 -9.28221703e-01 -1.00289293e-01 -2.80122429e-01
-8.34057748e-01 -5.84079623e-01 8.96719337e-01 -4.55493838e-01
1.44795799e+00 -1.95362127e+00 -1.92906991e-01 4.39335018e-01
6.82958245e-01 -2.02844501e-01 4.14278775e-01 2.43480489e-01
2.68976718e-01 6.33582115e-01 3.11562210e-01 -8.09492543e-02
-1.11452833e-01 -1.22818507e-01 -7.33243287e-01 8.47377121e-01
-1.49646193e-01 8.74240577e-01 -8.87606502e-01 -2.57388860e-01
-3.46831143e-01 2.35044751e-02 -4.14509118e-01 -2.12994382e-01
-3.74951720e-01 4.56241518e-01 -7.73682415e-01 4.43720847e-01
7.47701764e-01 -5.77041209e-01 4.75870878e-01 -1.95660386e-02
-4.50582743e-01 7.62916863e-01 -4.13913816e-01 6.79370463e-01
-1.32570848e-01 1.16568696e+00 1.92739099e-01 -1.02761662e+00
1.09744072e+00 8.49944353e-02 3.73630047e-01 -8.08641791e-01
4.55610812e-01 7.50299767e-02 2.99223959e-01 -3.30426097e-01
5.13010561e-01 -2.45872423e-01 -4.21568602e-01 1.10253716e+00
-6.45325720e-01 4.68358785e-01 3.66583541e-02 3.82271558e-01
6.40944242e-01 -4.79576647e-01 3.32526237e-01 -7.62058675e-01
1.70371771e-01 -1.22565061e-01 8.70765090e-01 8.03827226e-01
-6.03206038e-01 -8.67386162e-03 9.76576090e-01 -7.84209609e-01
-7.26799130e-01 -9.12739277e-01 -2.78089315e-01 1.24588382e+00
4.20596391e-01 -2.63832569e-01 -3.41644913e-01 -7.04215109e-01
3.19023192e-01 8.20875883e-01 -1.00149596e+00 -1.18177384e-02
-9.71786678e-01 -1.01664698e+00 2.10366860e-01 -7.89421275e-02
8.79540265e-01 -5.96833527e-01 -2.66388744e-01 2.15473734e-02
-5.64711630e-01 -9.45134461e-01 -3.23071152e-01 -5.22237062e-01
-6.80469334e-01 -1.03686762e+00 -6.15585208e-01 -3.21877867e-01
3.48583251e-01 5.97673655e-01 1.29641938e+00 7.12089520e-03
3.82997721e-01 -4.55794111e-02 -2.00904645e-02 -6.00789189e-01
-2.30119735e-01 1.34889036e-01 3.31821442e-01 5.87207913e-01
7.03541160e-01 -8.26653779e-01 -9.42337036e-01 5.13596356e-01
-2.48861894e-01 8.68728012e-03 -4.16316204e-02 2.70123601e-01
-1.84855998e-01 -2.01936379e-01 5.00882626e-01 -1.44650269e+00
9.78286147e-01 -1.25145197e+00 -3.71535063e-01 1.38665304e-01
-4.30648327e-01 -4.83996868e-01 4.64897245e-01 -5.13498306e-01
-1.11565650e+00 -8.45480204e-01 3.14757824e-01 3.70080024e-01
2.14251056e-02 5.03481865e-01 3.22844446e-01 5.19646764e-01
9.17445064e-01 -1.94904700e-01 -2.54611969e-01 -3.22476745e-01
2.30527610e-01 6.42371476e-01 1.09412305e-01 -4.94734287e-01
8.37033212e-01 9.64511454e-01 -7.19759643e-01 -8.63468468e-01
-1.74457848e+00 -3.66940111e-01 -1.82457417e-01 -5.71275592e-01
1.00456214e+00 -1.10726273e+00 -9.30134714e-01 5.27058840e-01
-1.04823220e+00 -4.55316901e-02 1.89535066e-01 2.97040254e-01
2.30233416e-01 5.41732311e-02 -6.24323547e-01 -7.51665771e-01
-2.21031532e-01 -5.02770722e-01 4.14556354e-01 6.71954632e-01
-1.05261230e+00 -1.32038617e+00 4.78689522e-01 5.04702449e-01
4.24274504e-01 4.91079956e-01 7.50362694e-01 -8.18292499e-01
-1.58743694e-01 3.56630012e-02 -1.33289203e-01 -2.57321566e-01
2.15776145e-01 2.07756415e-01 -8.86452317e-01 -1.90220371e-01
3.23174536e-01 -1.86135247e-01 1.04765356e+00 6.46509588e-01
8.29669774e-01 -7.59430707e-01 -6.90757871e-01 2.79965401e-01
8.35044384e-01 -1.18406884e-01 6.75968453e-02 2.70340741e-01
3.18399936e-01 8.02235126e-01 1.31396055e-01 4.09600705e-01
5.55678248e-01 5.84376037e-01 1.83018874e-02 -2.32953146e-01
2.06255153e-01 -5.66949725e-01 5.66900611e-01 5.78231931e-01
4.17516939e-02 -3.62626165e-01 -9.74417269e-01 3.69774878e-01
-1.83347023e+00 -1.25040948e+00 -5.06711900e-01 1.80547237e+00
7.26504862e-01 5.38708329e-01 4.33945984e-01 -3.17296892e-01
1.06181252e+00 8.82320702e-01 -6.35541201e-01 -3.78264844e-01
-4.03451979e-01 -3.36454511e-01 2.62479335e-01 5.39828777e-01
-1.16420805e+00 8.67434382e-01 6.67349339e+00 1.83500871e-01
-1.44837308e+00 1.67001799e-01 1.45459342e+00 -1.68109983e-01
-6.33050263e-01 9.62194148e-03 -9.10146832e-01 7.19237387e-01
8.74898851e-01 -5.45346260e-01 -1.57616749e-01 7.44467080e-01
5.37142754e-01 -3.61002982e-01 -7.07499564e-01 4.15947556e-01
7.47627690e-02 -1.49180424e+00 -9.97532979e-02 3.24378759e-01
1.20927131e+00 1.58891290e-01 4.13894206e-01 1.05174936e-01
1.10004747e+00 -6.63373291e-01 7.02271163e-01 3.06618154e-01
2.29511067e-01 -1.92054376e-01 3.37076426e-01 3.59666079e-01
-6.63846016e-01 -2.39893809e-01 -1.60376117e-01 -5.62321961e-01
3.93384367e-01 1.09086895e+00 -6.80966079e-01 -9.14440677e-02
6.20058715e-01 1.25101280e+00 -3.84143591e-01 4.03802514e-01
-2.05379635e-01 1.18836915e+00 -1.97475120e-01 -4.26694930e-01
2.94100136e-01 -4.10703570e-01 7.56862819e-01 8.40005040e-01
5.89585118e-02 1.84203729e-01 1.38126180e-01 1.03291273e+00
-3.14660043e-01 -6.48689568e-02 -8.87500286e-01 -1.41669303e-01
3.65185291e-01 1.20240808e+00 -6.83834374e-01 -4.98113245e-01
-5.13762176e-01 5.13725162e-01 4.61747736e-01 4.02497292e-01
-7.31083810e-01 4.67504978e-01 8.83122385e-01 5.43849468e-01
-3.50467950e-01 -1.05452783e-01 -7.62022138e-01 -1.55720556e+00
2.02100128e-02 -7.77904332e-01 4.88637179e-01 -2.58725166e-01
-1.64156520e+00 2.60855466e-01 -2.82521963e-01 -1.04420650e+00
2.42730990e-01 -2.52729803e-01 -1.46655679e+00 4.71143633e-01
-1.26325190e+00 -8.83906841e-01 -2.78595179e-01 4.28286910e-01
2.35091895e-01 -1.43896908e-01 1.77202716e-01 -1.81773826e-01
-7.94712245e-01 4.35323328e-01 -2.16770902e-01 4.79703963e-01
6.19645119e-01 -1.00246668e+00 4.22568798e-01 6.04159415e-01
3.66424888e-01 9.01424289e-01 9.35996830e-01 -1.02474344e+00
-6.89559937e-01 -7.41674662e-01 1.05926216e+00 -5.28783441e-01
1.26280737e+00 -5.56535840e-01 -7.84319699e-01 8.13791156e-01
3.06360811e-01 -7.19859600e-01 1.20435929e+00 9.96767461e-01
-6.36137128e-01 -4.22151871e-02 -8.22462142e-01 1.01910925e+00
1.05404913e+00 -3.68152857e-01 -8.38093698e-01 6.07650876e-01
4.71249700e-01 1.89776927e-01 -3.99881303e-01 5.59746921e-02
6.39910161e-01 -1.25814748e+00 6.18694544e-01 -7.74073601e-01
8.15078080e-01 2.14448541e-01 5.20577505e-02 -1.43401861e+00
-4.26097959e-01 -7.24947453e-01 2.95090396e-02 1.12867713e+00
6.40802562e-01 -9.03979361e-01 8.95432115e-01 7.26655602e-01
4.24993664e-01 -4.90137726e-01 -8.22686017e-01 -3.30308154e-02
4.63742763e-01 -1.60829965e-02 4.88853544e-01 1.52019382e+00
3.74555647e-01 6.97516322e-01 -5.91080010e-01 -4.77223583e-02
5.36145210e-01 7.14174986e-01 9.65652764e-01 -1.73911655e+00
6.08212389e-02 -7.93380618e-01 4.59527448e-02 -1.04245937e+00
2.42146105e-01 -8.09320569e-01 -6.49125576e-01 -1.02753997e+00
5.30215144e-01 -3.76906961e-01 -1.85538337e-01 2.73521617e-02
-3.59366208e-01 2.14267045e-01 6.36210199e-03 5.77195466e-01
-5.77761352e-01 4.10053879e-01 1.26393187e+00 -3.11186671e-01
-3.85009110e-01 3.40136826e-01 -1.27381992e+00 1.13380027e+00
9.45139050e-01 -3.29740196e-01 6.86292797e-02 -1.22621559e-01
9.02172208e-01 -2.10266694e-01 2.04184487e-01 -7.14425981e-01
6.14781640e-02 -3.41230035e-01 3.58613074e-01 -5.43353319e-01
2.93189496e-01 -3.27843696e-01 -3.39063019e-01 6.17722750e-01
-6.84653521e-01 -9.47219431e-02 -2.18203053e-01 8.27977121e-01
-2.98001021e-01 5.06281555e-01 6.84613943e-01 -2.19990507e-01
-2.73434967e-01 2.51287550e-01 -4.70148712e-01 3.81382316e-01
1.09250653e+00 -6.95829168e-02 -8.89887393e-01 -9.10753489e-01
-9.12421882e-01 5.53484917e-01 3.74759227e-01 5.11851013e-01
1.49156973e-01 -1.12777841e+00 -1.14716506e+00 -1.90546855e-01
-2.10918725e-01 -8.04433703e-01 1.15526237e-01 9.88092542e-01
1.43899098e-01 -1.78772081e-02 -1.55321926e-01 -4.95602041e-01
-1.27900267e+00 1.49743006e-01 3.28655839e-01 -3.22284013e-01
-5.51312745e-01 6.82417035e-01 3.65106732e-01 -2.84680843e-01
-2.75362343e-01 1.19516529e-01 -5.81474423e-01 8.51555049e-01
5.41974366e-01 2.99714833e-01 -6.11925364e-01 -7.45493412e-01
-7.44599849e-02 3.03337842e-01 -4.62740779e-01 -3.39335799e-02
1.21231568e+00 -8.55686143e-02 -2.89594233e-01 9.16716576e-01
1.10986793e+00 4.03199404e-01 -9.20796931e-01 -5.32397747e-01
7.49418810e-02 -6.12676322e-01 -1.43250123e-01 -5.04679739e-01
-9.03973579e-01 7.18825519e-01 -1.29668694e-02 1.00940073e+00
2.21503407e-01 2.52497852e-01 7.75264263e-01 2.01147884e-01
1.44133717e-01 -1.25494289e+00 2.22826198e-01 5.11912525e-01
6.69317663e-01 -1.51974666e+00 2.75578320e-01 -4.57493603e-01
-6.67934060e-01 7.94374704e-01 6.51568472e-01 -2.12587461e-01
1.00064969e+00 -1.95538387e-01 2.84528404e-01 -5.27095735e-01
-7.30574906e-01 1.58688694e-01 1.78481057e-01 6.87148869e-02
6.88110411e-01 5.58967471e-01 -6.42949402e-01 2.87043691e-01
-6.49369657e-01 -5.20388663e-01 6.66822433e-01 3.98462802e-01
-6.35719419e-01 -5.70735812e-01 -3.01448435e-01 7.84367025e-01
-7.30289638e-01 1.50347948e-01 -1.10788548e+00 5.44015229e-01
2.30592377e-02 9.83083308e-01 4.50883776e-01 -5.99336684e-01
1.43816443e-02 5.88400178e-02 -3.70426953e-01 -5.15430450e-01
-9.28011179e-01 -2.49704286e-01 5.37880540e-01 -5.29242098e-01
-5.24908483e-01 -9.25296366e-01 -5.92119336e-01 -1.06201041e+00
8.06799084e-02 1.21139064e-01 1.40378043e-01 8.75229657e-01
2.35882252e-01 -1.97586548e-02 1.02565670e+00 -2.31528506e-01
-3.50050926e-01 -9.51836109e-01 -2.96966851e-01 9.04254436e-01
3.06102425e-01 -5.69865286e-01 -4.12888885e-01 -4.22538459e-01] | [8.843643188476562, 10.114853858947754] |
c60fb7de-667f-446b-a346-37c9d36d9e46 | cross-lingual-genre-classification | null | null | https://aclanthology.org/E12-3002 | https://aclanthology.org/E12-3002.pdf | Cross-Lingual Genre Classification | null | ['Philipp Petrenz'] | 2012-04-01 | null | null | null | eacl-2012-4 | ['genre-classification'] | ['computer-vision'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.281436920166016, 3.826171636581421] |
25e8de21-7c9e-4fcd-b0d7-64d5b857a935 | efficiency-analysis-of-asp-encodings-for | 1711.05090 | null | http://arxiv.org/abs/1711.05090v1 | http://arxiv.org/pdf/1711.05090v1.pdf | Efficiency Analysis of ASP Encodings for Sequential Pattern Mining Tasks | This article presents the use of Answer Set Programming (ASP) to mine
sequential patterns. ASP is a high-level declarative logic programming paradigm
for high level encoding combinatorial and optimization problem solving as well
as knowledge representation and reasoning. Thus, ASP is a good candidate for
implementing pattern mining with background knowledge, which has been a data
mining issue for a long time. We propose encodings of the classical sequential
pattern mining tasks within two representations of embeddings (fill-gaps vs
skip-gaps) and for various kinds of patterns: frequent, constrained and
condensed. We compare the computational performance of these encodings with
each other to get a good insight into the efficiency of ASP encodings. The
results show that the fill-gaps strategy is better on real problems due to
lower memory consumption. Finally, compared to a constraint programming
approach (CPSM), another declarative programming paradigm, our proposal showed
comparable performance. | ['René Quiniou', 'Thomas Guyet', 'Torsten Schaub', 'Yves Moinard'] | 2017-11-14 | null | null | null | null | ['sequential-pattern-mining'] | ['natural-language-processing'] | [ 0.0664973 0.42177644 -0.3062638 -0.51979774 0.00654218 -0.40099996
0.27425286 0.9580745 -0.3657355 0.6973146 -0.1684333 -0.5028159
-0.7327351 -1.6560713 -0.65034205 -0.11389408 -0.6009696 0.7660446
0.77655995 -0.14552438 0.41462007 0.5161084 -2.250572 0.7616262
0.7940897 1.1204172 0.11874977 0.02733724 -0.91587704 0.7150106
-0.25045183 -0.67930174 0.2556521 -0.0273216 -1.1417753 -0.17467332
-0.17860077 0.4820954 0.12877966 1.1310532 -0.2981791 0.07419035
0.17231159 -1.3707681 -0.19822828 0.75396645 -0.24646357 0.1058594
1.0536873 -0.17550468 1.0659556 -0.7413053 0.907537 0.9566913
0.5157642 0.181254 -0.9576394 -0.11527627 0.2384488 0.86286217
-1.317937 0.0456106 0.4855369 -0.28885847 1.2911683 0.7394458
1.1470027 0.30604866 0.03635425 0.7401279 1.0398874 -0.841506
0.60203105 0.71195024 0.7529745 0.8715204 0.8406083 0.15802854
-0.66616744 -0.3422093 0.03072463 -0.14954321 -0.22215784 -0.49432465
-0.794723 0.916838 -0.17069823 0.53506994 -0.40924662 -0.30628917
0.52875084 0.6576588 -0.17818868 0.62694365 -0.5229327 -0.2657684
-0.79114515 0.7930254 1.4384019 1.1649051 0.8127835 -0.14695549
0.03013683 0.4095974 0.27070713 -0.08590861 0.21831955 -0.4543657
0.43628556 1.3360006 0.12060633 -1.338583 -0.29463536 -0.05573383
-0.30837786 0.21279263 0.3349276 0.17912449 -0.3727929 1.2105463
0.3612965 -0.11845177 0.15099838 0.59748256 0.5290481 0.97158736
-0.04629696 -0.5569787 1.6600897 -0.75321484 -0.7221758 0.07190541
0.61047506 -0.21906914 0.6958537 0.8926693 -1.161904 -0.23290823
-1.3236171 0.5061971 -0.9639305 -0.39729702 1.0473999 0.9135354
-0.82807076 0.5501859 -0.61466736 -0.4561873 0.18404855 0.4446384
-0.41739357 -0.16172846 -0.9791353 0.9307616 1.1216321 -0.35836884
-0.02533773 -0.8032251 -0.82926375 0.30808246 0.66921496 -0.4150517
0.800476 -0.5187938 -1.1079872 1.0111656 -0.18232231 -1.008268
0.15547265 0.1949338 -0.6797949 0.04383321 -0.26033065 0.1141003
0.11601132 -0.74896616 -0.7817584 -0.35319012 0.3108639 -0.18966483
-0.2907844 0.26507416 -0.04562695 -0.26955235 0.1314466 -0.3575288
-0.39144525 -0.18614778 0.06091413 -0.5072782 0.6938619 -0.17861792
1.7289686 -1.9072764 0.17028737 0.772514 -0.22407085 0.27581167
0.2969711 0.81733376 -0.1523489 0.27576458 -0.28178304 0.34609705
0.21928796 0.7748791 -0.23261236 -0.09053692 0.13247535 0.6652864
-0.6506479 -0.5236983 0.07908048 -0.44397184 -0.88758004 0.20525911
-0.9275521 -0.41124 -0.21925418 0.7961982 0.6438756 0.11101427
0.55838585 0.1417545 -0.6139508 0.2693822 -1.6937495 1.5035434
-0.23939203 0.11914717 -0.10059306 -1.3252392 1.453103 0.27059388
0.50810564 -0.98182 -0.13425899 0.40782607 -0.06862641 -0.80410004
0.42661506 -0.04470445 -0.0848984 0.4159854 -0.161285 0.09060878
0.8663422 -0.29735172 1.3678124 -0.13204312 0.6820287 -0.71419567
0.87990636 0.7770367 0.881632 0.52001005 0.43266156 0.16211183
0.76001745 -1.2058687 -0.65212893 -0.8209982 -0.04378498 0.33079618
0.18012543 -1.057459 -0.26121575 -0.47548974 0.08860588 0.7173082
-0.25032365 0.09746591 -0.7811713 -0.7743315 0.15593186 0.29463488
0.27545166 -1.3342184 -1.2939367 0.49201483 0.11282329 -0.893941
0.6644315 0.5188321 -0.886063 -1.4567196 0.3586787 -0.82582057
0.38254628 -0.18352604 1.3639957 0.22680174 -0.49214044 0.15437183
-1.0163325 -0.906086 -0.2107833 -0.23183085 -0.1239286 -0.08163442
0.7970397 -0.790084 -0.07593147 0.10615557 -1.0554658 -0.20076562
0.43698108 0.44258407 0.740252 0.7572854 0.17072567 -1.1080527
0.60369885 -0.56954736 -1.0164388 0.42564344 -0.8756851 0.18362059
0.54640657 0.0568947 -0.7700397 0.06073361 -0.20395258 0.0401978
-0.33763272 0.96780413 -0.28978208 0.11847112 0.44820675 0.20959517
-0.07336009 -0.41364005 -0.19282728 0.22274797 0.15752846 -0.9831075
0.45393243 0.31858885 0.13699079 -0.7559501 -0.33133712 -0.39179283
-0.33081064 0.1295724 0.6792833 -0.22133161 -0.6050029 -0.05747308
-1.2469321 0.03832244 -0.87178046 0.18810336 -0.59954596 0.22204305
-0.06540158 -0.9246416 -0.21134622 -0.84299815 0.24915178 0.04723043
-0.41876984 -0.6149759 0.20459591 -0.17014155 0.15573251 0.61406356
1.6414152 -1.0008887 -0.62475646 -0.1881658 0.08323386 -0.17240685
-0.32695663 0.01864709 -0.24789497 0.1468187 0.05190726 0.16491069
0.29559487 -0.11065165 1.4398359 -0.7165826 -0.3964668 0.27161154
1.9089705 0.8336699 1.0835382 0.7190443 -0.24061766 0.9610604
1.0352166 0.7360962 0.17882915 0.7002592 0.639574 0.5045227
0.43859622 -0.09951957 0.03515112 0.47656605 0.0311951 -0.09384276
-1.1548862 0.67450976 -2.0197303 -1.196813 -0.44660145 2.0768008
0.72374237 0.36237925 0.42936283 1.0248252 0.40388307 -0.01193533
0.23366626 -1.3422164 -0.07318559 0.7462126 0.1387979 0.3588756
-0.7703251 0.38718924 5.721673 0.5863043 -0.66769713 -0.18982777
-0.17157473 0.1482731 -0.5440796 0.2044001 -0.7056628 0.5420781
1.0199947 -0.29875726 0.29885325 1.0034119 -0.25645664 -0.38995865
-1.3562607 0.8879668 -0.2580917 -1.965232 0.26928326 0.01289095
0.311302 -0.38246343 -0.60503745 0.20175865 -0.01153658 -0.8802092
0.9088808 0.44331428 0.09116173 -0.93823504 0.8788398 0.47374398
-1.2578654 -0.44195336 -0.38168925 -0.42575863 0.09890109 0.9645864
-0.5796424 1.094822 0.91641957 0.433323 -0.28068238 1.445543
0.05657184 0.22466369 -0.62989944 -0.4172977 0.20564367 -0.4397611
0.6350485 1.4439871 0.30466628 0.20927072 0.1422222 0.97225004
0.6261211 0.2555871 -0.8132176 -0.01270153 0.658446 0.7595723
-0.63220674 -0.05186111 -0.57923365 0.20840766 0.09921163 -0.18639837
-0.63624495 -0.663618 0.5072339 0.52654487 0.4027212 -0.1766482
-0.5693881 -0.7596168 0.64556867 -0.8323822 0.87073255 -0.22861494
-0.7829726 0.52531254 0.44553885 -1.0395116 -0.03723023 -0.8131465
-0.72656304 0.4136479 -1.5099386 -0.6402754 -0.23184283 0.5238854
0.35929036 -0.08438421 1.1291327 0.3612785 -0.31379932 0.21282993
-0.7170221 -0.18299797 -0.2705041 -1.3018272 -0.20499176 0.81063724
0.37711394 0.52818674 0.7328893 -0.32945597 -1.7843242 -0.8423835
1.1671525 0.01571332 0.40972358 -0.1250877 -0.96415925 0.40495402
0.2209054 -0.08277703 0.77632546 0.19258007 -0.366851 -0.40654817
-1.3813231 0.26072448 0.97695535 0.09250772 -1.1528645 0.01774672
0.59417456 -0.02783296 -1.001273 0.7009871 0.54539114 -1.5483868
0.7065869 -0.577601 0.3668814 -0.49526784 -0.3992631 -0.7723329
-0.08367039 -0.49381825 -0.45092198 0.96976876 0.65357214 -0.70447505
1.075621 0.54865813 -0.19893076 -1.3278267 -0.8625011 -1.0050693
-0.36795348 -0.69437945 1.0028445 0.79526985 0.7670457 -0.27306795
0.23259349 0.07650411 0.36650646 0.86207706 0.39966664 -1.6542896
-0.53956455 -0.58098525 -0.72177666 -0.4090059 -0.02495233 -1.0985795
-0.57003075 -1.6103375 -0.25022662 -0.8937476 -0.16396973 0.582694
0.7306278 -0.41706035 -0.01882243 -0.17611444 -0.44481385 0.01546814
0.68397725 -0.1809337 -0.29173562 -0.06951264 -0.44636327 0.733343
0.6988746 -0.651167 -0.41207203 -0.28136712 0.8983765 0.19402237
0.01017698 -1.1493479 0.51489586 -0.6106916 -0.345076 -0.8547297
0.12467807 -1.3372972 0.61142033 0.8598432 -0.04933979 0.26242298
0.2445274 0.4675646 -0.6989945 -0.86228377 0.5541651 -0.4532497
-1.2286187 0.07712062 -0.44736508 -0.13878451 1.5453633 -0.8280891
-0.03397924 0.35759884 -0.89545244 0.36818865 0.2970348 0.21168609
0.8297611 -0.9549875 -0.32849 0.46289092 0.36716282 0.32240674
-0.20298894 0.72939986 -0.9485343 0.6380564 -0.47951558 -0.32441163
-1.1168164 0.81447804 0.05076731 -0.6068709 -0.6582495 0.561077
-0.7062596 -0.24639562 0.41376334 -0.51881915 -0.40136907 0.12350707
0.65687335 0.6062444 0.49857602 0.3491898 -0.76125413 0.42958942
0.27218348 0.31173417 1.5761424 0.40565872 -0.7404665 0.17974918
0.5879186 0.12880272 -0.04943312 0.02626433 0.98717797 -0.449228
-0.48577306 -0.7300765 -0.67152923 0.40445083 0.03783204 0.7591461
1.1481994 0.0617938 0.46896726 0.60171676 1.0946711 -1.0239551
-0.29277372 0.43157506 0.8626462 -0.56952214 0.2168123 -1.0258155
-0.28408816 1.3470458 0.6576189 -0.29962426 0.7837424 0.7683772
-0.78950137 -0.5642492 -1.0809451 -0.27932405 -0.20932625 0.62549007
-0.10254923 0.09631354 -0.92630833 0.7680072 -0.25130758 0.07817607
0.53938854 1.3324546 -0.61749834 -1.671987 -0.29720557 0.5649488
-0.3475195 0.30634627 -0.301412 1.0354698 0.55297524 0.8376931
0.18608895 -0.46928078 0.7343075 0.29078612 0.75222355 -0.9851512
-0.83598065 -0.9129044 0.5987256 -0.71994144 -0.5508166 -0.61767435
-1.339378 -0.12935732 -0.07766297 0.7486578 0.5832764 0.75916225
0.28138524 0.09847605 0.31532556 0.156567 -0.3368309 -0.22504383
-0.59760034 0.14682259 -0.3159407 -0.70644355 0.08289723 -0.20834647] | [8.369091987609863, 6.394146919250488] |
b0ec645f-e046-401c-94c0-e5f2b3cc1e5b | universal-face-restoration-with-memorized | 2110.01033 | null | https://arxiv.org/abs/2110.01033v1 | https://arxiv.org/pdf/2110.01033v1.pdf | Universal Face Restoration With Memorized Modulation | Blind face restoration (BFR) is a challenging problem because of the uncertainty of the degradation patterns. This paper proposes a Restoration with Memorized Modulation (RMM) framework for universal BFR in diverse degraded scenes and heterogeneous domains. We apply random noise as well as unsupervised wavelet memory to adaptively modulate the face-enhancement generator, considering attentional denormalization in both layer and instance levels. Specifically, in the training stage, the low-level spatial feature embedding, the wavelet memory embedding obtained by wavelet transformation of the high-resolution image, as well as the disentangled high-level noise embeddings are integrated, with the guidance of attentional maps generated from layer normalization, instance normalization, and the original feature map. These three embeddings are respectively associated with the spatial content, high-frequency texture details, and a learnable universal prior against other blind image degradation patterns. We store the spatial feature of the low-resolution image and the corresponding wavelet style code as key and value in the memory unit, respectively. In the test stage, the wavelet memory value whose corresponding spatial key is the most matching with that of the inferred image is retrieved to modulate the generator. Moreover, the universal prior learned from the random noise has been memorized by training the modulation network. Experimental results show the superiority of the proposed method compared with the state-of-the-art methods, and a good generalization in the wild. | ['Ran He', 'Xiaofei Jia', 'Huaibo Huang', 'Jia Li'] | 2021-10-03 | null | null | null | null | ['blind-face-restoration'] | ['computer-vision'] | [ 6.19300246e-01 -2.70155400e-01 7.66859427e-02 -9.58885998e-02
-7.74612904e-01 1.36492223e-01 3.19844723e-01 -5.93762755e-01
-4.42263037e-01 7.11654186e-01 5.70948064e-01 3.50128680e-01
-4.29550499e-01 -7.93708742e-01 -6.36658251e-01 -1.45882869e+00
1.30932719e-01 -1.18990064e-01 -5.52330017e-02 -2.16776043e-01
3.67151260e-01 4.53577012e-01 -1.98804593e+00 5.39025366e-01
6.49276793e-01 1.15099323e+00 5.65874338e-01 4.65593934e-01
2.43968174e-01 5.42576969e-01 -6.32424712e-01 -1.64898157e-01
1.10191613e-01 -4.48394328e-01 -2.04007030e-01 3.40669632e-01
3.53722692e-01 -3.58731568e-01 -8.48908484e-01 1.43035221e+00
8.32374513e-01 3.04663479e-01 5.78853011e-01 -6.63996577e-01
-1.36517763e+00 1.21587239e-01 -5.28009355e-01 5.75093389e-01
2.68745571e-01 3.29988211e-01 3.50936949e-01 -1.25018752e+00
5.47122598e-01 1.43939209e+00 1.98313266e-01 7.68619061e-01
-1.32280493e+00 -6.23861551e-01 1.16591215e-01 7.19732583e-01
-1.52013731e+00 -7.79252827e-01 1.00388265e+00 -2.25535005e-01
6.95869327e-01 2.70628035e-01 3.55000973e-01 1.05560005e+00
3.20748240e-01 1.03757851e-01 1.45768774e+00 -3.79850447e-01
2.57020175e-01 5.37952259e-02 -2.85047263e-01 5.82410991e-01
1.20846212e-01 4.38905597e-01 -8.25614214e-01 6.54515922e-02
7.93485403e-01 -2.13660598e-02 -8.77280414e-01 2.50901673e-02
-9.94741440e-01 4.87400949e-01 4.83244866e-01 4.48781163e-01
-6.87122643e-01 -2.58537322e-01 -2.38229573e-01 3.23524117e-01
5.19873261e-01 1.66310724e-02 -2.64211804e-01 5.30570030e-01
-8.97231400e-01 -1.30945116e-01 2.65850693e-01 3.92240912e-01
8.40981483e-01 3.29219192e-01 -5.38576782e-01 9.03832853e-01
4.47398841e-01 6.98797762e-01 9.07389104e-01 -8.87416005e-01
3.05169016e-01 2.39805669e-01 6.89172894e-02 -1.25391161e+00
1.71705768e-01 -4.88788247e-01 -1.36537480e+00 5.17139077e-01
-8.23567063e-02 4.36596692e-01 -1.16081917e+00 1.91398633e+00
1.98344380e-01 6.53369546e-01 3.47314864e-01 1.15037239e+00
6.92114949e-01 6.57206416e-01 -1.36691824e-01 -7.20303357e-01
1.55093575e+00 -4.78855312e-01 -1.01317585e+00 -3.35225761e-01
-4.65854615e-01 -8.92007530e-01 8.94433737e-01 3.19047570e-01
-1.07753062e+00 -1.02495861e+00 -1.33190620e+00 -5.11564650e-02
-1.98477015e-01 4.20531392e-01 -9.99372527e-02 5.68874180e-01
-1.13796866e+00 5.81683397e-01 -5.11506319e-01 -1.63390618e-02
4.45231348e-01 1.52238309e-01 -6.95599496e-01 -6.48087442e-01
-1.23437250e+00 9.07102346e-01 1.39700249e-01 6.10864222e-01
-1.16135871e+00 -3.44487190e-01 -7.94792891e-01 2.49838382e-02
-1.80266127e-01 -7.45012164e-01 3.57047737e-01 -8.72209907e-01
-1.50448167e+00 8.21005404e-01 -3.16459715e-01 -4.66282181e-02
1.58085674e-01 1.30960181e-01 -7.42141485e-01 4.47726429e-01
4.24552187e-02 2.87676752e-01 1.82091057e+00 -1.34361541e+00
-2.47725785e-01 -7.70073771e-01 -2.72195250e-01 2.48992696e-01
-5.98658323e-01 -1.55748948e-01 -2.19600841e-01 -1.03338075e+00
5.13509631e-01 -1.85482994e-01 1.34825170e-01 -7.58536011e-02
-1.16747459e-02 2.34269455e-01 6.04999185e-01 -1.41202390e+00
1.15272212e+00 -2.73826694e+00 6.84380949e-01 1.79501042e-01
1.23513369e-02 6.34156615e-02 -4.17107642e-01 -2.04850212e-01
-4.56566125e-01 -1.20106444e-01 -4.30255204e-01 -3.25591922e-01
-1.29590735e-01 1.69743448e-01 -6.02748573e-01 8.58876288e-01
3.25181127e-01 5.26562512e-01 -7.49688804e-01 -2.77170122e-01
2.21490204e-01 9.56382453e-01 -2.96050489e-01 5.59024513e-01
2.89197341e-02 7.62534142e-01 -1.20690256e-01 4.34907198e-01
1.05392051e+00 1.69772446e-01 1.63943842e-01 -7.66565740e-01
5.69755062e-02 -1.14081860e-01 -1.19830477e+00 1.52333164e+00
-3.33349347e-01 3.04882467e-01 2.42024988e-01 -1.07445180e+00
1.07428408e+00 4.43983197e-01 -1.19651198e-01 -9.82526839e-01
-2.96130925e-02 1.87162295e-01 -2.32012749e-01 -8.14270973e-01
8.65119025e-02 -2.65268564e-01 3.95438731e-01 1.41075432e-01
3.25762957e-01 2.16300413e-01 -3.30546290e-01 -2.06130788e-01
8.72037292e-01 5.00625744e-02 -6.83210641e-02 -9.34639573e-02
9.27017212e-01 -7.94033468e-01 5.33607244e-01 3.62073153e-01
5.83567917e-02 7.73582101e-01 -2.52486989e-02 -1.45230174e-01
-8.06338906e-01 -1.11452723e+00 -2.38382116e-01 9.25171018e-01
2.70001769e-01 6.01197705e-02 -7.86187053e-01 -7.64016360e-02
-1.30515218e-01 6.06930256e-01 -8.92681122e-01 -6.95663512e-01
-7.06427038e-01 -1.00400496e+00 1.57644227e-01 1.71458572e-01
8.38047922e-01 -1.18425298e+00 -8.31170753e-02 -5.71403094e-02
-3.60233158e-01 -8.87826800e-01 -4.58847791e-01 -1.96480229e-01
-7.92265594e-01 -9.15827990e-01 -6.21577621e-01 -9.52041924e-01
9.76205826e-01 4.94844139e-01 5.17341971e-01 9.75375473e-02
-5.03047824e-01 4.23711360e-01 -4.92289849e-02 6.09273374e-01
-4.03520852e-01 -7.59353995e-01 8.73164758e-02 8.78481448e-01
2.28654779e-02 -1.00232756e+00 -8.41619372e-01 3.07877690e-01
-1.13750160e+00 -2.14940667e-01 9.03385162e-01 1.11603868e+00
7.41885960e-01 6.03475332e-01 5.55822253e-01 -3.01487178e-01
5.92203557e-01 -4.02954578e-01 -2.27354228e-01 1.32579565e-01
-4.40706640e-01 1.23568177e-01 2.58332610e-01 -7.95485497e-01
-1.37077034e+00 -4.70254451e-01 6.98970780e-02 -5.58981895e-01
-1.05106495e-01 2.41675571e-01 -8.25987339e-01 -1.39760226e-01
7.78200090e-01 8.36335182e-01 9.88130122e-02 -7.69560575e-01
3.86392862e-01 7.34415591e-01 9.02479708e-01 -5.72061777e-01
1.07605982e+00 5.26548862e-01 -2.13056788e-01 -7.44758904e-01
-5.23109794e-01 8.54323506e-02 -4.39810306e-01 -2.36037076e-01
9.27335918e-01 -1.10665286e+00 -4.14417088e-01 7.15049267e-01
-1.09205663e+00 4.53012362e-02 -3.87899250e-01 4.48824406e-01
-4.67858702e-01 4.63844180e-01 -6.53621554e-01 -7.42678344e-01
-1.35735631e-01 -1.11984777e+00 8.11621130e-01 2.69865632e-01
4.70785707e-01 -4.09895569e-01 -3.74735951e-01 4.93933618e-01
6.37362242e-01 5.25353327e-02 1.31556523e+00 -5.08140326e-02
-8.61491680e-01 -8.69072005e-02 -2.25095466e-01 8.15184236e-01
1.91654712e-01 -6.49125218e-01 -1.12072086e+00 -5.99989474e-01
7.26294935e-01 5.38513670e-03 1.12817204e+00 3.13066423e-01
1.20150888e+00 -4.69240993e-01 9.24792290e-02 6.65901542e-01
1.34927297e+00 -8.56079087e-02 1.06617022e+00 2.53935426e-01
3.29539508e-01 4.91314173e-01 2.98213661e-01 2.28253365e-01
5.95255569e-02 5.48524380e-01 4.19584483e-01 2.01581776e-01
-8.27950716e-01 -2.22383589e-01 6.56421959e-01 9.55761850e-01
-3.26853186e-01 1.24027953e-01 -1.83779880e-01 3.61424118e-01
-1.51473844e+00 -1.10982585e+00 4.51524407e-01 2.42314100e+00
9.99951959e-01 -3.32318172e-02 -7.40304589e-01 3.47042710e-01
1.16464818e+00 4.30814028e-01 -6.16175115e-01 2.64142305e-01
-8.03961813e-01 4.95404601e-01 2.90929954e-02 6.77502871e-01
-7.92691290e-01 6.13359213e-01 5.06680155e+00 1.04718900e+00
-8.69523525e-01 4.99117523e-01 5.93129873e-01 1.33885285e-02
-2.59703338e-01 -1.78997070e-01 -3.94167840e-01 5.29431164e-01
6.42639399e-01 3.85941863e-02 1.24516606e+00 2.34114796e-01
1.37171417e-01 1.29904047e-01 -7.81021297e-01 1.06187177e+00
5.81062376e-01 -9.65272307e-01 3.60007703e-01 -1.43388920e-02
5.43191791e-01 -6.53797746e-01 7.93100476e-01 1.38289183e-01
-3.02511811e-01 -1.13939798e+00 5.68856776e-01 1.12504327e+00
1.22527075e+00 -6.04867816e-01 6.08967781e-01 7.30196834e-02
-1.00493133e+00 -3.96008283e-01 -6.70916319e-01 2.84253418e-01
-2.23532066e-01 6.22855842e-01 2.54060403e-02 5.76129019e-01
8.58664393e-01 5.65499783e-01 -5.35912037e-01 5.50072312e-01
-5.82036555e-01 4.92992848e-01 2.25329876e-01 7.68898368e-01
-5.91622233e-01 -2.52817899e-01 8.70886266e-01 7.61408806e-01
4.13422376e-01 3.86829883e-01 -2.33868271e-01 6.40666008e-01
-2.52579093e-01 -9.78287235e-02 -1.61958322e-01 3.05736631e-01
4.15854663e-01 1.20935404e+00 -2.05202907e-01 -2.13624239e-01
-6.40750900e-02 1.23288786e+00 1.55128747e-01 8.47078681e-01
-5.38915873e-01 -1.59909859e-01 7.10509479e-01 8.01036358e-02
4.46323454e-01 -4.17739227e-02 5.03607169e-02 -1.15880239e+00
3.02009702e-01 -9.30948436e-01 2.00090095e-01 -1.12950540e+00
-1.45073080e+00 7.91988134e-01 -4.14643079e-01 -1.16403401e+00
1.14772812e-01 -5.42653561e-01 -3.68426830e-01 1.33260131e+00
-1.81233275e+00 -1.04230142e+00 -3.53557646e-01 9.52332437e-01
2.50587344e-01 -6.01669431e-01 1.01305056e+00 4.41616625e-01
-6.47202730e-01 5.86492360e-01 8.33228230e-02 -8.26253835e-03
8.25760126e-01 -6.89238071e-01 -2.40295351e-01 1.01682127e+00
-2.20006675e-01 7.03904510e-01 6.65549636e-01 -6.41399741e-01
-1.54595411e+00 -1.18328881e+00 4.91327316e-01 6.91716895e-02
5.17841995e-01 -1.28317505e-01 -1.21475995e+00 3.05429697e-01
4.49803658e-02 3.00572962e-01 5.30884147e-01 -5.68720698e-01
-7.02844799e-01 -4.39034760e-01 -1.45314980e+00 4.98163044e-01
1.02460349e+00 -9.48350728e-01 -8.07632744e-01 1.74667418e-01
9.03141618e-01 -2.33934388e-01 -8.60026896e-01 4.16051358e-01
3.06407660e-01 -8.61528456e-01 1.19616795e+00 -5.00663519e-01
4.02150035e-01 -5.99964261e-01 -7.69947112e-01 -1.17684269e+00
-6.52754545e-01 -3.74587893e-01 -2.63415605e-01 1.37400293e+00
-2.20603198e-02 -7.90197790e-01 2.43890479e-01 1.44280106e-01
1.89415272e-03 -4.49324012e-01 -1.47746599e+00 -3.35966796e-01
-3.93750876e-01 1.73462089e-02 5.97660780e-01 7.68947005e-01
-6.55109465e-01 1.61159094e-02 -5.92256784e-01 8.43322635e-01
9.65543509e-01 1.18842162e-01 1.51562229e-01 -9.26045954e-01
-3.81791770e-01 -7.74030313e-02 -6.24319971e-01 -6.78802729e-01
3.45726013e-01 -9.39027309e-01 1.89829338e-02 -1.23683286e+00
1.76700518e-01 2.52010912e-01 -7.93443263e-01 1.98063985e-01
-2.08011135e-01 4.45110500e-01 -3.01196482e-02 3.69288117e-01
-2.40387283e-02 8.64502072e-01 1.33863604e+00 -3.90308261e-01
5.25894016e-02 -2.55911529e-01 -8.11735868e-01 4.36989576e-01
4.85693306e-01 -3.68373185e-01 -1.40696749e-01 -5.42762697e-01
-2.12407544e-01 1.91332743e-01 8.76824319e-01 -1.07636058e+00
3.81555855e-01 1.04803231e-03 9.53704834e-01 5.08404188e-02
6.95927501e-01 -8.60789776e-01 1.54845282e-01 3.23000997e-01
-2.31408298e-01 -3.13149959e-01 4.15323190e-02 9.34691846e-01
-3.27022582e-01 5.55213448e-03 1.10104847e+00 5.87451197e-02
-4.86323148e-01 3.74610305e-01 -1.07567772e-01 -3.01743656e-01
4.37292010e-01 -3.20610642e-01 -5.55598915e-01 -2.44381782e-02
-1.03864133e+00 -4.85584766e-01 1.62922606e-01 5.14753938e-01
1.29134810e+00 -1.44843614e+00 -1.01624727e+00 8.35122943e-01
-1.30700678e-01 -4.60397542e-01 9.25468266e-01 4.12226826e-01
1.53307527e-01 -2.65637904e-01 -4.28098947e-01 -2.98601955e-01
-9.09159124e-01 8.85235846e-01 4.17477220e-01 1.43319458e-01
-6.45839691e-01 6.03418887e-01 1.94436550e-01 9.16100666e-02
2.77430862e-02 2.48343453e-01 -4.58823144e-01 3.65867540e-02
1.00773954e+00 3.18417221e-01 7.79360384e-02 -9.67018485e-01
-1.78554878e-01 8.01868975e-01 1.55001163e-01 -3.10991794e-01
1.38767242e+00 -2.83262521e-01 -8.52977037e-01 8.98312405e-02
1.24288034e+00 1.14724301e-01 -1.18403530e+00 -5.90358138e-01
-3.14703166e-01 -8.54191422e-01 2.80586958e-01 -7.52215147e-01
-1.32108188e+00 7.70155787e-01 1.30312419e+00 -1.88369393e-01
1.79938436e+00 -1.87194318e-01 3.17205131e-01 -1.88469067e-01
2.44672686e-01 -7.63559341e-01 3.92686963e-01 -5.01630381e-02
1.40188372e+00 -7.60721207e-01 -5.75323850e-02 -1.32371202e-01
-6.47402182e-03 1.02007055e+00 5.10533035e-01 -2.67007351e-01
8.02635074e-01 -4.07488942e-02 -1.08802982e-01 3.77322249e-02
-4.79463071e-01 -6.83326721e-02 2.87028491e-01 8.41186166e-01
-2.36849815e-01 -4.61648963e-02 4.72957864e-02 8.51763308e-01
4.09170538e-02 -1.81571066e-01 4.34348546e-02 3.34629744e-01
-5.94669938e-01 -6.96866333e-01 -8.44419897e-01 1.10822022e-01
-2.34118029e-01 -1.21691681e-01 3.69100839e-01 1.06108800e-01
5.18421412e-01 1.12678599e+00 -6.03644922e-02 -5.68674147e-01
4.27044213e-01 9.86471921e-02 6.63463533e-01 -3.34145576e-01
2.69050896e-02 5.62574565e-02 -4.26071644e-01 -4.80847955e-01
-3.59683603e-01 -3.38094890e-01 -7.71453202e-01 -1.79097243e-03
-4.01481055e-03 6.40293285e-02 5.49799740e-01 7.95396209e-01
4.77963865e-01 6.63919091e-01 9.24268603e-01 -1.14676642e+00
-6.47091389e-01 -1.20953321e+00 -7.23106027e-01 5.71254432e-01
6.08229816e-01 -7.83274412e-01 -8.23237538e-01 3.08386862e-01] | [12.825699806213379, -0.06892905384302139] |
a4f5b13a-a20e-4c15-b8f5-75e6cee5567b | multiview-boosting-by-controlling-the | 1808.05784 | null | http://arxiv.org/abs/1808.05784v2 | http://arxiv.org/pdf/1808.05784v2.pdf | Multiview Boosting by Controlling the Diversity and the Accuracy of View-specific Voters | In this paper we propose a boosting based multiview learning algorithm,
referred to as PB-MVBoost, which iteratively learns i) weights over
view-specific voters capturing view-specific information; and ii) weights over
views by optimizing a PAC-Bayes multiview C-Bound that takes into account the
accuracy of view-specific classifiers and the diversity between the views. We
derive a generalization bound for this strategy following the PAC-Bayes theory
which is a suitable tool to deal with models expressed as weighted combination
over a set of voters. Different experiments on three publicly available
datasets show the efficiency of the proposed approach with respect to
state-of-art models. | ['Massih-Reza Amini', 'Pascal Germain', 'Emilie Morvant', 'Anil Goyal'] | 2018-08-17 | null | null | null | null | ['multiview-learning', 'multilingual-text-classification'] | ['computer-vision', 'miscellaneous'] | [-2.65870929e-01 -8.62352327e-02 -9.86545324e-01 -9.55847621e-01
-1.11829007e+00 -4.93412346e-01 1.09296978e+00 2.70588875e-01
-2.91564673e-01 5.22848010e-01 4.04374421e-01 -1.29150809e-03
-1.52187198e-01 -5.45458913e-01 -5.64685881e-01 -6.14816368e-01
1.59440324e-01 6.72424674e-01 2.72508502e-01 -6.83543161e-02
3.48528743e-01 7.05238730e-02 -1.96906292e+00 6.28244996e-01
5.20352900e-01 1.23639023e+00 -5.13706543e-02 5.69506407e-01
1.36295497e-01 7.38396525e-01 -1.81751132e-01 -8.03913414e-01
3.95981133e-01 -1.43627420e-01 -4.47065324e-01 1.73563406e-01
8.96340370e-01 -1.08146667e-01 3.98949981e-01 1.04257298e+00
9.60453972e-02 -1.21541247e-01 1.16102386e+00 -1.28326464e+00
-4.97668907e-02 4.02872652e-01 -7.98360944e-01 3.27586293e-01
2.43833378e-01 -3.64397138e-01 1.36546147e+00 -9.96239543e-01
6.68969512e-01 1.47000122e+00 4.95049149e-01 4.61888015e-01
-1.28094447e+00 -3.28433216e-01 6.30740166e-01 5.72231352e-01
-1.00148940e+00 -2.52369404e-01 8.46316099e-01 -6.50105059e-01
3.34315509e-01 4.36838806e-01 8.32325816e-01 1.09382665e+00
6.23647809e-01 8.99073422e-01 1.68505943e+00 -4.98843431e-01
5.78653872e-01 6.57816947e-01 4.63892937e-01 5.92195213e-01
3.60029280e-01 2.21103922e-01 -7.81359553e-01 -5.52034974e-01
-5.13429195e-02 1.30723774e-01 -3.07708494e-02 -1.14107203e+00
-8.50476682e-01 1.24680650e+00 3.23372751e-01 -7.04635354e-03
-5.07621288e-01 -2.37492830e-01 5.85016012e-01 -1.26296533e-02
9.44076180e-01 -4.32009518e-01 -4.68287796e-01 7.33538866e-01
-6.31475031e-01 4.20928091e-01 5.95156074e-01 7.42086768e-01
7.67054617e-01 -2.40227774e-01 -1.83451518e-01 8.03792238e-01
6.45638645e-01 4.78546768e-01 2.11114839e-01 -7.92199492e-01
3.66791695e-01 4.98991430e-01 5.42735644e-02 -7.18706250e-01
-9.11848396e-02 -3.56389374e-01 -7.78487921e-01 4.94424105e-01
3.93052846e-02 7.95562267e-02 -6.21548176e-01 1.65184128e+00
7.88234949e-01 -1.85097754e-01 -2.77245641e-02 5.21667898e-01
8.91184628e-01 4.52789634e-01 1.25349373e-01 -4.90853339e-01
1.41121316e+00 -1.02941847e+00 -6.31815612e-01 -2.35210076e-01
-3.91516797e-02 -3.21530670e-01 4.19775635e-01 5.85755527e-01
-8.11453044e-01 -6.15541637e-01 -1.37418544e+00 4.95222747e-01
-3.44502568e-01 -1.45423889e-01 4.70985144e-01 9.71969008e-01
-9.17582273e-01 1.34892702e-01 -3.23970526e-01 -8.42368156e-02
3.80208880e-01 1.99174851e-01 -3.71086508e-01 -8.80533084e-02
-6.79520190e-01 1.09329093e+00 3.28168005e-01 -4.35546249e-01
-1.37045693e+00 -6.86858356e-01 -7.72749841e-01 -1.05642706e-01
3.52915436e-01 -9.81796145e-01 1.26417744e+00 -1.23185909e+00
-1.18648076e+00 1.14731705e+00 -3.73624772e-01 -4.50328112e-01
6.87170684e-01 7.34872445e-02 -3.12889218e-01 -8.77318531e-02
2.79101860e-02 4.42693800e-01 1.24085498e+00 -1.89769280e+00
-1.05424953e+00 -9.39234972e-01 3.16292614e-01 3.41205060e-01
1.01650782e-01 -8.18785727e-02 -1.65232763e-01 -3.15130532e-01
2.22681925e-01 -6.33606076e-01 -2.92785555e-01 -2.38880634e-01
1.70240045e-01 -3.39041919e-01 6.32340372e-01 -5.68182230e-01
8.85987341e-01 -1.67122352e+00 3.60655367e-01 2.96457976e-01
4.54865582e-02 -6.90455660e-02 3.15911591e-01 4.16232228e-01
1.04694925e-01 -2.32820719e-01 8.52276683e-02 -4.55807537e-01
-1.74225554e-01 2.42432103e-01 -1.93065479e-01 7.00792074e-01
-5.02804041e-01 4.17461693e-01 -7.41133749e-01 -8.04706752e-01
4.73732322e-01 1.82810590e-01 -4.39054519e-01 1.87394321e-01
-2.62683600e-01 5.77357560e-02 -1.92289799e-01 4.94229645e-01
8.48366618e-01 -3.26572061e-02 6.78078353e-01 -1.23394273e-01
1.98989868e-01 -3.18692386e-01 -1.29086339e+00 1.33370447e+00
-6.27233267e-01 2.06222206e-01 3.94228905e-01 -9.79948103e-01
7.43689060e-01 3.08192194e-01 1.94897503e-01 -2.23491013e-01
2.75127202e-01 -1.32117480e-01 -6.04198158e-01 -1.91885069e-01
2.01749414e-01 -6.84997320e-01 -3.70937847e-02 2.18147531e-01
6.17249787e-01 -3.68708931e-02 -2.18300909e-01 1.42851353e-01
2.34842196e-01 -2.74320692e-02 1.07320178e+00 -6.17805898e-01
9.20164049e-01 -1.76407427e-01 6.31619811e-01 7.32142627e-01
-1.80050358e-01 3.33999068e-01 2.17646524e-01 -7.97220469e-01
-9.03770745e-01 -1.13444781e+00 -1.77888438e-01 1.30778074e+00
2.05908015e-01 -4.78904366e-01 -4.65241551e-01 -1.22486210e+00
4.51215655e-02 7.37331092e-01 -9.81212437e-01 3.74488622e-01
-1.05855815e-01 -6.51661813e-01 -5.92059731e-01 4.58752275e-01
1.38331190e-01 -2.97706544e-01 -6.22270882e-01 -1.19491696e-01
-1.94115102e-01 -8.06896448e-01 -1.57603994e-01 1.91162914e-01
-1.03859925e+00 -1.06127083e+00 -5.44847071e-01 -4.89477515e-01
4.78031725e-01 5.46615601e-01 1.38761175e+00 -3.77661347e-01
2.26313487e-01 7.91194618e-01 -3.73839229e-01 -8.49364877e-01
-4.51030463e-01 -2.21456856e-01 -2.27713156e-02 3.55756253e-01
6.81137592e-02 -3.95945549e-01 -4.00896370e-01 2.76608676e-01
-5.50574005e-01 3.82170454e-03 1.31715208e-01 1.02937412e+00
7.27347255e-01 -5.14829814e-01 4.02773499e-01 -1.24508023e+00
2.30313554e-01 -5.69623232e-01 -1.02531242e+00 7.50332236e-01
-8.23893428e-01 8.21319222e-02 2.76137382e-01 -8.18183720e-02
-1.30755091e+00 7.13960156e-02 2.21984461e-01 -3.75575334e-01
7.04682022e-02 2.49131456e-01 -3.40580076e-01 7.33094364e-02
6.76316440e-01 -6.50506690e-02 -1.23313442e-01 -1.87656179e-01
6.19691670e-01 5.45312881e-01 2.43460923e-01 -4.65789378e-01
4.14990216e-01 8.97142708e-01 1.60786167e-01 -4.44487512e-01
-1.29643357e+00 -8.25536311e-01 -6.53098881e-01 -8.54045331e-01
9.42439675e-01 -1.26916921e+00 -7.43102789e-01 9.68971103e-03
-8.60438526e-01 4.45875436e-01 -6.61810264e-02 5.60019791e-01
-7.63289630e-01 3.53260994e-01 3.23713683e-02 -1.16416299e+00
-4.31164294e-01 -9.86554623e-01 8.76398861e-01 -4.60173674e-02
2.73451954e-02 -1.13444650e+00 4.69964802e-01 9.07612979e-01
-4.10740301e-02 2.54433572e-01 8.48325014e-01 -7.18707025e-01
-3.72717768e-01 -3.30621988e-01 1.16501942e-01 6.80422902e-01
-3.19026470e-01 -1.24676771e-01 -1.37495315e+00 -3.31868052e-01
2.74137050e-01 -3.47928703e-01 1.12284219e+00 6.47387385e-01
9.56023455e-01 -3.21926713e-01 -3.68489116e-01 2.59286791e-01
1.91559958e+00 -9.11287367e-02 7.59849548e-02 1.28318146e-01
3.24894607e-01 6.71446085e-01 8.80283535e-01 7.08938122e-01
6.47266030e-01 8.25317621e-01 8.16437006e-01 2.85995752e-01
3.83147895e-01 -3.19078356e-01 5.01193255e-02 7.30791628e-01
-2.71797955e-01 -1.97338387e-02 -6.11830175e-01 4.96407986e-01
-2.10986590e+00 -1.24687791e+00 -2.16635540e-02 2.35345960e+00
2.02756226e-01 3.72950621e-02 3.80384624e-01 -1.50011539e-01
7.50774860e-01 7.05224097e-01 -1.83786169e-01 -7.56067872e-01
2.92477366e-02 -1.21917240e-02 6.53440833e-01 7.70644426e-01
-1.23588073e+00 3.27469051e-01 7.23533964e+00 9.50976491e-01
-5.72731495e-01 7.32622445e-01 6.94367886e-01 -5.95870335e-03
-5.24205208e-01 1.55320898e-01 -8.85343850e-01 1.76950634e-01
7.22650349e-01 -2.04001427e-01 1.64132267e-01 1.44866538e+00
-3.82978439e-01 -3.29067230e-01 -1.02863455e+00 9.72824216e-01
6.39884889e-01 -1.42028487e+00 1.71919003e-01 1.74122259e-01
1.22970462e+00 1.39383450e-01 -5.08764461e-02 2.91954607e-01
6.60267055e-01 -4.40891832e-01 1.03282797e+00 4.14980501e-01
2.40004539e-01 -9.24237907e-01 7.72348106e-01 7.46018052e-01
-1.05692971e+00 -2.75336534e-01 -3.87341321e-01 1.48822427e-01
-9.59233269e-02 5.28708994e-01 -5.89036942e-01 9.59152520e-01
7.59941876e-01 6.11728966e-01 -2.92978257e-01 6.94501400e-01
-2.36089602e-01 2.02230006e-01 9.51786637e-02 5.22350669e-02
3.90819907e-02 -2.06529036e-01 4.31698561e-01 8.71805847e-01
1.01013388e-02 -1.14465743e-01 1.26855895e-01 1.76850110e-01
5.77957258e-02 2.39983171e-01 -7.95530617e-01 1.07727671e+00
2.71813631e-01 1.39818192e+00 -4.00313914e-01 -6.97203457e-01
-6.14547491e-01 4.56749767e-01 4.22174871e-01 -6.53440505e-02
-6.35010183e-01 8.46173644e-01 4.90549117e-01 -1.35672152e-01
6.82782054e-01 3.33744824e-01 -3.32418680e-01 -1.45200312e+00
-3.39025036e-02 -1.03190053e+00 9.60977376e-01 -3.48680079e-01
-1.63187957e+00 3.36720824e-01 6.90062582e-01 -1.44238305e+00
-3.65232706e-01 -5.90607643e-01 -4.94590253e-01 6.20666623e-01
-1.14746106e+00 -1.54542291e+00 -1.96412101e-01 4.90407169e-01
8.15959156e-01 -3.20318222e-01 7.14411795e-01 -3.58749360e-01
-1.27910944e-02 1.99376419e-01 2.75091529e-01 -5.95408261e-01
5.86922526e-01 -1.20669746e+00 -1.14792086e-01 4.72926825e-01
2.81540900e-01 7.26303682e-02 8.45845044e-01 -2.47880459e-01
-1.18282390e+00 -6.38088048e-01 8.53260219e-01 -7.48987317e-01
9.41020623e-02 -3.40759784e-01 -3.34334940e-01 8.64808142e-01
5.80807090e-01 1.00218996e-01 1.05676591e+00 6.44237757e-01
-8.24056089e-01 -6.97535217e-01 -1.44452560e+00 1.02706172e-01
5.52302539e-01 -3.35501313e-01 -7.32072890e-01 1.63067743e-01
6.65546507e-02 -1.12006254e-01 -6.95019066e-01 7.02520967e-01
9.98345613e-01 -1.56747389e+00 1.10887480e+00 -5.92105269e-01
3.21397573e-01 -3.40415128e-02 -7.37379849e-01 -1.45475388e+00
-3.56182188e-01 3.66939634e-01 -8.16302747e-02 1.03684556e+00
1.66332006e-01 -4.75176066e-01 5.18921673e-01 2.13525251e-01
3.07448477e-01 -7.61096716e-01 -1.27099681e+00 -6.48971558e-01
-1.16341382e-01 -4.08073843e-01 3.27906221e-01 8.27614188e-01
-3.28531712e-02 4.52671021e-01 -6.18990779e-01 9.04158428e-02
1.18977141e+00 3.36883157e-01 7.90229142e-01 -1.44441926e+00
-4.08078372e-01 -2.05218554e-01 -5.62101066e-01 -3.71889025e-01
4.54493433e-01 -7.87390888e-01 -1.98263794e-01 -1.38931799e+00
9.14959848e-01 1.40229821e-01 -3.66146356e-01 -2.17601463e-01
-2.51290947e-01 2.09146347e-02 3.72611135e-01 9.97050386e-03
-7.87323534e-01 5.79600930e-01 7.96264946e-01 -3.36594522e-01
3.68723541e-01 4.11973506e-01 -5.99074244e-01 1.12310123e+00
5.21849096e-01 -5.69014132e-01 -5.27867734e-01 1.11986719e-01
2.65292883e-01 5.52104056e-01 2.95356870e-01 -8.63553703e-01
-3.56685109e-02 -2.42149383e-01 5.41949928e-01 -1.18810880e+00
5.33598006e-01 -1.07335138e+00 3.39175224e-01 4.69970077e-01
-2.91944116e-01 1.23392008e-01 -2.67588377e-01 1.16059470e+00
-3.77570689e-01 -3.82000118e-01 9.31154191e-01 -4.11220670e-01
-6.50399864e-01 6.32060766e-02 -4.14463580e-01 -1.16904505e-01
1.27836943e+00 4.20712829e-02 -2.50144154e-01 -6.22423053e-01
-5.65237880e-01 2.05567807e-01 5.30097604e-01 2.95655787e-01
4.28148508e-01 -1.53538024e+00 -7.43707538e-01 -9.10871997e-02
6.77594841e-01 -5.82844555e-01 5.83321393e-01 4.98100907e-01
-6.79670796e-02 2.22460613e-01 -3.56775522e-01 -8.18984747e-01
-1.88754594e+00 1.01414990e+00 1.06222048e-01 -5.81490755e-01
-1.95472483e-02 5.72034657e-01 3.62229556e-01 -4.31079954e-01
1.82721302e-01 8.86670351e-02 -5.25769711e-01 6.33638024e-01
5.30004025e-01 6.59744263e-01 1.50722355e-01 -8.59822869e-01
-5.65429330e-01 5.64720631e-01 -1.21562276e-02 -4.74542826e-01
1.19990504e+00 -1.93331048e-01 -5.81323542e-03 8.99115622e-01
9.71122086e-01 -1.59450725e-01 -1.15553701e+00 -4.37140524e-01
-1.82896256e-01 -6.19377315e-01 2.54225105e-01 -6.83804274e-01
-8.87018323e-01 7.14761615e-01 8.58760476e-01 -7.43404254e-02
9.93863702e-01 2.29953572e-01 -3.20136398e-01 -1.04547873e-01
8.24931800e-01 -1.03312421e+00 -1.43260792e-01 1.26890630e-01
9.38294291e-01 -1.59187531e+00 5.37996590e-01 -2.68522859e-01
-9.33696330e-01 7.67889023e-01 3.27962875e-01 -1.61606520e-01
1.08846414e+00 -2.77138114e-01 1.72366813e-01 -3.67517099e-02
-1.27072275e+00 -9.55818817e-02 6.64683342e-01 5.69087207e-01
-3.61419022e-02 5.39010882e-01 -6.27533197e-01 3.58249635e-01
3.71802270e-01 -2.45898828e-01 1.51702911e-01 7.78688371e-01
-4.51701522e-01 -1.08116162e+00 -4.47153658e-01 4.65079188e-01
-4.99864519e-01 3.95381153e-01 -3.49484503e-01 6.27218366e-01
2.65199482e-01 1.10292912e+00 -2.30732486e-01 -4.78841007e-01
1.33148760e-01 1.92499995e-01 7.86219478e-01 -5.18567324e-01
-8.63034904e-01 -8.36831518e-03 1.29320264e-01 -3.35995197e-01
-9.63640034e-01 -1.00877047e+00 -1.22656070e-01 -1.54333591e-01
-4.10895616e-01 -5.04446328e-02 9.94900942e-01 7.91098595e-01
-1.38738707e-01 -6.22966737e-02 1.02379656e+00 -1.09987211e+00
-8.91883016e-01 -9.32153106e-01 -7.22118199e-01 3.55222702e-01
2.60829449e-01 -6.72594845e-01 -3.90969872e-01 -1.49070024e-01] | [8.506999015808105, 4.484723091125488] |
48b663a2-59bf-419a-80e1-61b700a2b8a4 | deid-gpt-zero-shot-medical-text-de | 2303.11032 | null | https://arxiv.org/abs/2303.11032v1 | https://arxiv.org/pdf/2303.11032v1.pdf | DeID-GPT: Zero-shot Medical Text De-Identification by GPT-4 | The digitization of healthcare has facilitated the sharing and re-using of medical data but has also raised concerns about confidentiality and privacy. HIPAA (Health Insurance Portability and Accountability Act) mandates removing re-identifying information before the dissemination of medical records. Thus, effective and efficient solutions for de-identifying medical data, especially those in free-text forms, are highly needed. While various computer-assisted de-identification methods, including both rule-based and learning-based, have been developed and used in prior practice, such solutions still lack generalizability or need to be fine-tuned according to different scenarios, significantly imposing restrictions in wider use. The advancement of large language models (LLM), such as ChatGPT and GPT-4, have shown great potential in processing text data in the medical domain with zero-shot in-context learning, especially in the task of privacy protection, as these models can identify confidential information by their powerful named entity recognition (NER) capability. In this work, we developed a novel GPT4-enabled de-identification framework ("DeID-GPT") to automatically identify and remove the identifying information. Compared to existing commonly used medical text data de-identification methods, our developed DeID-GPT showed the highest accuracy and remarkable reliability in masking private information from the unstructured medical text while preserving the original structure and meaning of the text. This study is one of the earliest to utilize ChatGPT and GPT-4 for medical text data processing and de-identification, which provides insights for further research and solution development on the use of LLMs such as ChatGPT/GPT-4 in healthcare. Codes and benchmarking data information are available at https://github.com/yhydhx/ChatGPT-API. | ['Xiang Li', 'Dajiang Zhu', 'Tianming Liu', 'Quanzheng Li', 'Dinggang Shen', 'Wei Liu', 'Lin Zhao', 'Haixing Dai', 'Chao Cao', 'Zihao Wu', 'Lu Zhang', 'Xiaowei Yu', 'Zhengliang Liu'] | 2023-03-20 | null | null | null | null | ['de-identification'] | ['natural-language-processing'] | [ 2.04747155e-01 1.44547895e-01 -2.78215617e-01 -3.49505305e-01
-1.04131222e+00 -5.19639015e-01 1.23169176e-01 8.63820672e-01
-6.59843922e-01 6.82732821e-01 4.69379097e-01 -6.80973351e-01
-3.55311602e-01 -6.81409001e-01 -2.31790558e-01 -5.56646585e-01
1.52371943e-01 5.89873910e-01 -1.79967299e-01 1.16972178e-01
9.46214944e-02 2.87848473e-01 -1.07454240e+00 7.02684343e-01
1.13344753e+00 4.38906550e-01 8.53449851e-03 4.63301867e-01
-3.94503713e-01 5.65161109e-01 -5.30282259e-01 -7.15736389e-01
2.07024619e-01 -2.43380010e-01 -9.62219894e-01 -6.47213876e-01
-2.51033545e-01 -8.69928151e-02 -3.78144607e-02 1.01991868e+00
9.00670409e-01 -3.95866215e-01 2.99270332e-01 -8.41376245e-01
-8.22360337e-01 5.73058307e-01 -3.71493250e-01 -7.67945498e-02
5.40744364e-01 -2.27882341e-01 5.68584323e-01 -4.80900735e-01
7.14458227e-01 7.82484233e-01 1.01656187e+00 1.03025830e+00
-1.00872517e+00 -1.01213086e+00 -4.96238291e-01 -1.16109930e-01
-1.61044157e+00 -4.47939485e-01 1.52644008e-01 -5.46826422e-01
1.01874197e+00 7.42456496e-01 3.88112038e-01 1.05955386e+00
3.69916081e-01 6.36152685e-01 1.01814687e+00 -6.24247313e-01
8.75886828e-02 4.56213087e-01 1.15459047e-01 3.99607033e-01
5.90735257e-01 3.96278612e-02 -2.83518881e-01 -8.48887086e-01
3.04394990e-01 4.22137231e-01 -2.23960295e-01 -9.54390876e-03
-1.14699006e+00 6.91898525e-01 -2.30377093e-01 5.68164587e-01
-3.92539322e-01 -6.38677776e-01 7.40248322e-01 1.13037460e-01
4.52630132e-01 5.24470210e-01 -6.60189092e-01 -1.51356757e-01
-9.12837088e-01 3.27536487e-03 9.63878870e-01 1.10464048e+00
3.87376815e-01 -7.25459397e-01 -3.20238113e-01 8.21420431e-01
8.05789307e-02 2.12754369e-01 8.75383317e-01 -4.59788442e-01
5.84685504e-01 8.79068315e-01 2.08836064e-01 -1.03937542e+00
-4.92826998e-01 -4.97176088e-02 -1.06469417e+00 -3.52718264e-01
1.88793883e-01 -4.62806165e-01 -9.07841086e-01 1.41015029e+00
4.38691884e-01 -9.31149125e-02 4.83518273e-01 3.57758313e-01
8.54999721e-01 2.51268893e-01 6.14017963e-01 -1.44372106e-01
1.70627689e+00 -2.01938912e-01 -1.12020218e+00 6.37560487e-02
1.26676548e+00 -8.01815212e-01 5.27381182e-01 1.02062844e-01
-5.63989580e-01 -8.11044946e-02 -5.52370250e-01 -7.20388144e-02
-6.60744309e-01 -9.37427208e-02 4.90633667e-01 1.18744159e+00
-7.81764209e-01 3.55664790e-01 -9.90827382e-01 -6.40117288e-01
7.22425103e-01 3.74138921e-01 -7.41706371e-01 -1.27370343e-01
-1.51042080e+00 6.83067381e-01 5.40738165e-01 7.00469092e-02
-3.27103734e-01 -9.34182465e-01 -8.97172749e-01 -3.21631134e-03
4.18166786e-01 -7.54722834e-01 9.97819066e-01 -4.69275564e-01
-9.73321974e-01 1.16163647e+00 -1.22452416e-01 -4.97544914e-01
6.13525987e-01 -1.38835922e-01 -9.36974168e-01 -7.64062926e-02
2.82430768e-01 2.37934679e-01 2.63948619e-01 -9.26629126e-01
-8.51588428e-01 -5.69950402e-01 -6.10848904e-01 -1.54735714e-01
-4.82716322e-01 4.00248975e-01 -2.17276886e-01 -9.39061344e-01
-2.49317333e-01 -8.39913726e-01 -3.56519133e-01 -3.46965373e-01
-5.72395384e-01 1.04362483e-03 5.73171437e-01 -1.23928881e+00
1.91177368e+00 -2.21861291e+00 -5.73527455e-01 2.14119866e-01
3.79604548e-01 9.33585286e-01 3.55007291e-01 7.94824660e-01
-1.09355681e-01 7.11615920e-01 -5.36176622e-01 -2.97318071e-01
-2.54790515e-01 2.97330350e-01 -5.29901497e-02 3.67435664e-01
-1.77121252e-01 9.44054127e-01 -8.17964196e-01 -7.13938832e-01
1.92594364e-01 4.93168056e-01 -4.46870178e-01 5.32389618e-02
1.66036353e-01 4.60795283e-01 -6.49650335e-01 7.48205960e-01
7.94028580e-01 -2.43877932e-01 5.73101699e-01 1.61429316e-01
-2.06549853e-01 2.05274582e-01 -1.00424600e+00 1.51794505e+00
-6.56780973e-02 4.51582149e-02 1.27759740e-01 -6.05450392e-01
7.49933243e-01 7.82975018e-01 7.67517447e-01 -5.41909456e-01
1.28412634e-01 2.50712514e-01 -1.83839574e-01 -1.02683306e+00
4.50611621e-01 -1.64630160e-01 -3.30000222e-01 3.45477432e-01
-3.35127860e-01 6.57883704e-01 -9.73989666e-02 1.01628758e-01
9.92097855e-01 -2.52735019e-01 7.51930475e-01 -1.68789640e-01
6.72420084e-01 7.39612356e-02 8.84845138e-01 6.45304382e-01
-2.44547427e-01 4.36307251e-01 1.01780169e-01 -1.77698389e-01
-6.70884848e-01 -3.76472473e-01 -5.60967386e-01 6.01151645e-01
-2.63064295e-01 -6.90753043e-01 -8.24838400e-01 -8.96433115e-01
2.79962659e-01 9.25651312e-01 -5.11424541e-01 -3.85220610e-02
-4.46353078e-01 -9.62930441e-01 1.04713225e+00 2.67026514e-01
3.23157281e-01 -1.09567356e+00 -6.45995021e-01 4.54986006e-01
-3.66776675e-01 -9.11376715e-01 -6.94938719e-01 7.04004019e-02
-6.29984975e-01 -1.30308115e+00 -7.15007603e-01 -7.61007786e-01
7.80374467e-01 -1.86507896e-01 5.54898798e-01 -7.22026965e-03
-5.92316687e-01 3.28559548e-01 -3.84960949e-01 -7.62317479e-01
-8.74494493e-01 2.56444991e-01 -2.35596091e-01 4.19487990e-02
1.06534100e+00 -1.10547774e-01 -5.65901816e-01 1.78607106e-01
-1.19273829e+00 -1.02459617e-01 6.07250214e-01 7.93251872e-01
5.28030872e-01 1.01656169e-01 5.96632779e-01 -1.65567148e+00
7.33513474e-01 -6.05860233e-01 -2.70032078e-01 6.06957495e-01
-1.19015384e+00 4.52257171e-02 4.57129866e-01 -1.39268607e-01
-1.19989550e+00 -6.87527955e-02 -5.38146734e-01 1.45434216e-02
-3.82245153e-01 6.80932105e-01 -3.19322616e-01 8.74047354e-02
5.63595414e-01 1.56999797e-01 9.53083485e-02 -7.82723904e-01
-1.06038116e-01 1.26258242e+00 4.83718038e-01 -1.17999665e-01
4.13848579e-01 3.80666912e-01 -5.10476172e-01 -5.65619528e-01
-4.81047958e-01 -8.72061670e-01 -5.07339835e-01 4.92502987e-01
1.05938387e+00 -8.74092102e-01 -7.07468092e-01 6.70533836e-01
-6.97285593e-01 2.56657511e-01 -6.20820522e-02 5.33627212e-01
1.28604889e-01 8.62868249e-01 -6.28194988e-01 -9.27263021e-01
-8.21248829e-01 -6.86905801e-01 7.77467668e-01 4.79147062e-02
-7.05946386e-01 -1.03447330e+00 2.57984400e-01 6.94137454e-01
2.91747659e-01 3.77753794e-01 1.06216753e+00 -1.15517509e+00
1.28295161e-02 -5.62847078e-01 7.51780495e-02 1.68964714e-01
5.98262608e-01 -3.85023206e-01 -7.60807037e-01 -3.73723567e-01
2.61289418e-01 2.78171480e-01 5.16671419e-01 2.11012572e-01
1.03280032e+00 -7.03772843e-01 -8.50318551e-01 7.75655150e-01
1.33445346e+00 5.12938142e-01 8.04987848e-01 5.42814612e-01
7.30133235e-01 6.73527241e-01 6.13194466e-01 5.81136107e-01
5.92257023e-01 3.02923679e-01 -8.63929018e-02 -2.54260987e-01
3.91993493e-01 -4.80041176e-01 -5.15820198e-02 5.29654860e-01
2.10291162e-01 -1.83030173e-01 -1.21705818e+00 5.89477718e-01
-1.73209500e+00 -9.01539981e-01 -6.42677918e-02 2.39548254e+00
1.19112861e+00 -2.61702299e-01 -1.43628553e-01 2.42560748e-02
8.67791712e-01 -2.56452173e-01 -4.65457827e-01 -4.06987429e-01
1.96772558e-03 1.96752265e-01 1.01656020e+00 1.88743010e-01
-1.12645900e+00 5.99375129e-01 5.53154993e+00 6.32839143e-01
-9.50296998e-01 1.52324215e-01 5.70089877e-01 2.54341155e-01
-3.49466771e-01 -9.89607349e-02 -9.66133118e-01 7.30951250e-01
1.08165491e+00 -4.28667694e-01 -1.43154457e-01 7.58683264e-01
2.93708593e-01 9.46328342e-02 -8.88894141e-01 1.16606140e+00
-1.48494661e-01 -1.38650823e+00 1.58585310e-02 2.57239759e-01
4.44190145e-01 -2.55968869e-01 -1.00536905e-01 1.01771928e-01
2.32903421e-01 -8.59874547e-01 2.87909716e-01 5.11456788e-01
1.10549235e+00 -6.43763900e-01 1.05925941e+00 4.09016430e-01
-8.36939752e-01 -1.80902466e-01 -2.13911846e-01 4.90655929e-01
1.80662885e-01 5.66936314e-01 -1.14498794e+00 1.08904266e+00
8.83885741e-01 6.15693510e-01 -3.87553245e-01 9.83257055e-01
7.62490705e-02 4.60771471e-01 1.51530085e-02 3.71820778e-01
-3.23196322e-01 1.56563930e-02 2.79360861e-01 1.43150616e+00
4.64563698e-01 3.46531302e-01 -6.25301450e-02 4.68212128e-01
-7.26709440e-02 3.83328259e-01 -3.59871089e-01 -4.30756658e-01
5.46216965e-01 1.07010102e+00 -3.66111010e-01 -2.85836726e-01
-6.23335361e-01 7.67720103e-01 -1.18832670e-01 6.21277355e-02
-4.77245748e-01 -5.13763428e-01 8.50017130e-01 4.32232946e-01
1.83446407e-01 1.71703622e-01 -2.25471988e-01 -1.12917447e+00
-2.48963945e-03 -1.18883824e+00 1.21114469e+00 -1.07666537e-01
-1.30195725e+00 4.79700059e-01 -8.06606561e-03 -1.13349378e+00
-2.92358279e-01 -3.01353365e-01 -8.49513430e-03 1.02688432e+00
-1.33995140e+00 -1.08883512e+00 2.80347951e-02 7.08427608e-01
-4.90953214e-02 -4.37720530e-02 1.33835471e+00 6.46377802e-01
-7.58460522e-01 1.07161677e+00 4.27679300e-01 6.02163613e-01
1.02924812e+00 -8.56594443e-01 3.81244004e-01 6.23697937e-01
-3.05660129e-01 1.07865345e+00 3.34047109e-01 -1.16998768e+00
-1.18117559e+00 -1.10068560e+00 1.55397367e+00 -7.54278123e-01
2.27778852e-01 -6.21040881e-01 -1.35480416e+00 7.77406633e-01
-1.96485832e-01 -3.91982943e-01 1.41613019e+00 -7.85372704e-02
-1.79534107e-01 1.31467013e-02 -1.65393412e+00 2.87030339e-01
7.73906469e-01 -7.86574185e-01 -7.38389730e-01 2.03358665e-01
5.55373073e-01 -4.25328374e-01 -1.23994994e+00 2.58655190e-01
5.96661210e-01 -6.94876075e-01 8.40779364e-01 -6.23578668e-01
-2.42105231e-01 -1.16173059e-01 2.15640128e-01 -6.12923801e-01
-2.47703940e-01 -9.26626146e-01 2.21553266e-01 1.61851895e+00
5.94370902e-01 -1.12874615e+00 7.66589463e-01 1.37588799e+00
1.82119414e-01 -4.14036036e-01 -1.03926396e+00 -4.27049011e-01
-3.98576744e-02 -2.08256945e-01 8.15174103e-01 1.42016184e+00
2.81564415e-01 -2.98577011e-01 -5.50209105e-01 3.73499572e-01
2.65104949e-01 -3.27388078e-01 4.11817878e-01 -1.19128108e+00
5.21434061e-02 8.36046562e-02 -2.57877588e-01 -2.65786141e-01
-1.57143027e-01 -1.05151820e+00 -3.83427233e-01 -1.51519680e+00
2.78720677e-01 -5.86140573e-01 -3.79209429e-01 8.99567008e-01
-3.52243423e-01 -4.39425796e-01 -1.46998659e-01 5.76572359e-01
-1.50295660e-01 -4.03328426e-02 7.03040600e-01 9.94202644e-02
-4.30688411e-01 1.36996523e-01 -1.10154533e+00 2.83833653e-01
8.04829478e-01 -1.08130443e+00 -8.54666010e-02 -2.53873914e-01
6.10523447e-02 2.38402560e-02 3.94113846e-02 -6.37851655e-01
4.44684595e-01 -8.27778429e-02 2.28629157e-01 -4.96066302e-01
-3.76908422e-01 -1.05093014e+00 7.16019809e-01 6.58603609e-01
-3.75286758e-01 5.19576296e-02 3.63465399e-01 5.43357611e-01
-1.86729491e-01 -1.87116981e-01 2.93150604e-01 -3.59693319e-01
-6.70968950e-01 3.37266445e-01 -5.41963637e-01 -9.50005054e-02
1.15103757e+00 -4.17284191e-01 -2.82702476e-01 1.44043162e-01
-7.47751892e-01 4.57758874e-01 4.61116344e-01 3.74350935e-01
4.26209450e-01 -8.50064635e-01 -6.80129409e-01 3.90918553e-01
3.74372900e-01 -4.49119806e-02 8.03766429e-01 8.12231481e-01
-3.58534843e-01 7.21521795e-01 -2.34013274e-02 -1.30324751e-01
-1.68164754e+00 1.03450596e+00 1.76551536e-01 -5.19270301e-01
-8.73977244e-01 3.54582906e-01 1.74018174e-01 -5.67359447e-01
2.07457006e-01 -2.92981327e-01 -2.77992696e-01 2.89431959e-02
7.38363087e-01 2.95504034e-01 2.61096537e-01 -5.00266254e-01
-6.43273056e-01 2.01531261e-01 -4.54418600e-01 2.71761030e-01
1.30282903e+00 -3.98671120e-01 -3.17364126e-01 -4.24012206e-02
1.12436473e+00 2.56873488e-01 -4.33827013e-01 -1.04334690e-01
6.22476816e-01 -5.06154358e-01 -3.28610450e-01 -9.96103704e-01
-7.06100106e-01 3.93612146e-01 6.41621828e-01 1.35258734e-01
1.19512165e+00 -1.67685270e-01 9.23344970e-01 7.34967068e-02
2.81623095e-01 -9.76070046e-01 -9.28228557e-01 1.06540918e-02
3.84316951e-01 -1.18419659e+00 -7.69705921e-02 -3.40981930e-01
-7.76166499e-01 7.90311396e-01 1.19738929e-01 8.45621347e-01
7.96617091e-01 4.48521644e-01 5.23043692e-01 -2.08847150e-01
-3.97659183e-01 2.13471875e-01 1.13744214e-01 8.19262803e-01
5.29731989e-01 1.85429692e-01 -6.15542889e-01 9.79881287e-01
-7.83414543e-02 4.30750012e-01 2.24033877e-01 1.33272564e+00
2.16076598e-01 -1.69337213e+00 -4.12007242e-01 6.46852314e-01
-1.14944959e+00 -3.06961477e-01 -2.72967041e-01 7.18068480e-01
2.01730445e-01 9.90012109e-01 -3.90145004e-01 -2.89615035e-01
4.15479332e-01 4.61090535e-01 -2.46326074e-01 -7.53539622e-01
-1.10475421e+00 -7.20513165e-02 4.56390560e-01 -2.47098058e-01
-4.80416715e-01 -8.67294967e-01 -1.24212527e+00 -3.58979106e-01
-2.07153812e-01 5.68021715e-01 4.80455220e-01 7.11450577e-01
1.00270212e+00 1.83362707e-01 1.21371500e-01 2.79296100e-01
-4.40578490e-01 -6.22232616e-01 -5.74461579e-01 4.40902621e-01
3.78397077e-01 1.23302164e-02 -1.05681419e-01 -2.41740048e-03] | [6.806180953979492, 6.99567985534668] |
15257be6-4ca7-4c59-8546-c1cf85208257 | scalable-mask-annotation-for-video-text | 2305.01443 | null | https://arxiv.org/abs/2305.01443v1 | https://arxiv.org/pdf/2305.01443v1.pdf | Scalable Mask Annotation for Video Text Spotting | Video text spotting refers to localizing, recognizing, and tracking textual elements such as captions, logos, license plates, signs, and other forms of text within consecutive video frames. However, current datasets available for this task rely on quadrilateral ground truth annotations, which may result in including excessive background content and inaccurate text boundaries. Furthermore, methods trained on these datasets often produce prediction results in the form of quadrilateral boxes, which limits their ability to handle complex scenarios such as dense or curved text. To address these issues, we propose a scalable mask annotation pipeline called SAMText for video text spotting. SAMText leverages the SAM model to generate mask annotations for scene text images or video frames at scale. Using SAMText, we have created a large-scale dataset, SAMText-9M, that contains over 2,400 video clips sourced from existing datasets and over 9 million mask annotations. We have also conducted a thorough statistical analysis of the generated masks and their quality, identifying several research topics that could be further explored based on this dataset. The code and dataset will be released at \url{https://github.com/ViTAE-Transformer/SAMText}. | ['DaCheng Tao', 'Bo Du', 'Juhua Liu', 'Mengyang Xu', 'Jing Zhang', 'Haibin He'] | 2023-05-02 | null | null | null | null | ['text-spotting'] | ['computer-vision'] | [ 5.29109776e-01 -3.00097734e-01 -1.49003237e-01 -2.84272671e-01
-8.95282149e-01 -8.52485180e-01 5.74778676e-01 -3.39856267e-01
-4.38456237e-02 2.88740367e-01 3.55937749e-01 -2.64299184e-01
5.02785563e-01 -3.35215807e-01 -7.84746230e-01 -2.04677612e-01
4.44574952e-01 2.35896498e-01 5.56774616e-01 2.14138344e-01
5.07402182e-01 8.39770362e-02 -1.47760177e+00 7.15778768e-01
6.77570343e-01 9.47277069e-01 2.00944394e-01 6.19779348e-01
-2.89942235e-01 7.10200429e-01 -8.39753985e-01 -7.82426119e-01
5.95162809e-01 -2.57578820e-01 -3.11104447e-01 2.82388508e-01
1.14483070e+00 -6.19078100e-01 -6.69720650e-01 1.01181626e+00
1.70384169e-01 -7.32726082e-02 5.15708506e-01 -1.36654544e+00
-4.06877309e-01 6.20071173e-01 -8.05802524e-01 2.08700553e-01
4.63164479e-01 3.33648294e-01 9.34202611e-01 -1.31768644e+00
7.00393796e-01 1.06150997e+00 9.45835531e-01 3.39017272e-01
-9.30395901e-01 -1.02107954e+00 1.16072856e-01 2.31258105e-03
-1.64195001e+00 -8.62149596e-01 5.10259569e-01 -8.05099010e-01
6.51134551e-01 4.41090196e-01 5.01129746e-01 1.28762543e+00
-5.05992882e-02 1.06050968e+00 7.12972045e-01 -3.13699961e-01
-1.97885320e-01 -6.68895468e-02 -5.77963442e-02 7.19923675e-01
2.85486013e-01 -4.78579104e-01 -7.49126375e-01 -6.21038862e-02
8.08198214e-01 -1.05917051e-01 -1.39616475e-01 1.07059181e-01
-1.41984034e+00 4.82785076e-01 -2.71837920e-01 -9.75366868e-03
3.57735939e-02 1.15700297e-01 3.67429972e-01 -3.88733484e-02
6.41355276e-01 2.08091706e-01 -1.02430187e-01 -3.47761005e-01
-1.57722068e+00 2.06900641e-01 4.54905659e-01 1.25882280e+00
7.71942377e-01 5.74027523e-02 -3.16760421e-01 8.14792216e-01
3.04339439e-01 6.14593625e-01 2.10410371e-01 -7.90563405e-01
1.14063573e+00 8.20135236e-01 1.15442418e-01 -1.15156078e+00
-2.00131372e-01 2.94188708e-01 -3.47728461e-01 -7.34103769e-02
3.89270395e-01 -3.18178654e-01 -1.17370975e+00 1.11561775e+00
2.07020670e-01 5.37084103e-01 -5.75855970e-01 9.58538592e-01
1.02565968e+00 7.51329660e-01 -1.82293937e-01 2.47101486e-01
1.36881614e+00 -9.80445504e-01 -8.18687975e-01 -4.67963099e-01
5.56794345e-01 -1.21005094e+00 1.16652322e+00 3.67296964e-01
-8.76306295e-01 -2.83511609e-01 -9.03956115e-01 -2.70315170e-01
-3.38640273e-01 5.23383260e-01 6.05760887e-02 6.90811872e-01
-8.79105866e-01 8.18798468e-02 -7.00021982e-01 -3.58938962e-01
6.50970638e-01 2.71583162e-02 -1.83447063e-01 -1.40535384e-01
-8.68644714e-01 4.99493450e-01 1.96916446e-01 2.85180002e-01
-7.19002664e-01 -6.06346905e-01 -8.87632191e-01 -2.58341551e-01
5.96140742e-01 -9.70607772e-02 1.12784290e+00 -1.10195982e+00
-1.04704750e+00 8.87005270e-01 -3.49361718e-01 -2.08590314e-01
8.35844815e-01 -3.18193406e-01 -5.89746833e-01 3.91472489e-01
2.89091557e-01 8.07160378e-01 1.22389746e+00 -1.03444469e+00
-6.21177852e-01 -1.03839718e-01 -1.39943346e-01 2.64927477e-01
-3.05867374e-01 4.73009378e-01 -1.11505318e+00 -1.27476561e+00
-7.30564892e-02 -9.12978232e-01 2.28740141e-01 -7.28167500e-03
-7.61231780e-01 2.32587382e-01 1.29535532e+00 -1.13044858e+00
1.49753976e+00 -2.28280187e+00 -2.41302609e-01 1.40114293e-01
2.64124304e-01 2.07013056e-01 -1.79727301e-01 3.91045392e-01
2.49250039e-01 4.81762439e-01 -1.61256224e-01 -4.57047731e-01
1.13611393e-01 -1.80853084e-01 -4.45029765e-01 4.83838946e-01
1.26947641e-01 8.40648651e-01 -3.20018888e-01 -8.53665411e-01
4.27059114e-01 2.31250077e-01 -4.64148790e-01 -1.31417394e-01
-4.37317878e-01 2.20225662e-01 -4.51347828e-01 1.12081575e+00
6.84964061e-01 -1.58261195e-01 -3.51888500e-02 -3.54177654e-01
-1.05546176e-01 7.51597211e-02 -1.26247919e+00 1.39961410e+00
1.62818685e-01 1.43142128e+00 1.98860452e-01 -4.08631682e-01
6.61266506e-01 2.80700743e-01 6.77999735e-01 -3.96686196e-01
1.97179750e-01 -1.00656129e-01 -5.41956186e-01 -6.86846673e-01
9.67709303e-01 5.08665085e-01 -5.87964617e-02 3.92367423e-01
-3.43877703e-01 -7.08291084e-02 5.69047749e-01 3.03988606e-01
1.15186226e+00 1.69439718e-01 -2.91318536e-01 1.11396581e-01
1.34062752e-01 2.74744004e-01 4.69143748e-01 7.96981275e-01
-2.12672636e-01 1.08137572e+00 5.06001413e-01 -3.46563876e-01
-1.25452113e+00 -7.34215200e-01 -1.03081860e-01 1.11267209e+00
2.77659059e-01 -8.84729505e-01 -8.42849731e-01 -6.62300050e-01
-1.67257655e-02 4.33298349e-01 -4.54619944e-01 4.35369849e-01
-6.63507402e-01 -4.87517089e-01 1.05473065e+00 4.03211117e-01
5.38795829e-01 -8.59921038e-01 -2.29180127e-01 -1.50974542e-01
-5.58886290e-01 -1.60923791e+00 -9.81131375e-01 -4.15901959e-01
-4.45134580e-01 -1.21122122e+00 -5.37349164e-01 -7.55505621e-01
7.17545152e-01 4.11556780e-01 9.19206917e-01 1.97171852e-01
-3.96245450e-01 4.03968453e-01 -4.58231121e-01 -2.70925105e-01
-3.25010419e-01 -9.51918513e-02 -1.17365256e-01 3.52884144e-01
4.11820173e-01 9.89527851e-02 -5.63737690e-01 7.61908233e-01
-1.13976741e+00 5.09175241e-01 3.27359319e-01 3.13927263e-01
4.15489972e-01 -5.12433089e-02 1.62763596e-01 -7.46243656e-01
2.70087034e-01 -4.57676798e-01 -8.39703739e-01 1.99210882e-01
6.11099266e-02 -5.15718758e-01 2.94144481e-01 -5.14803588e-01
-9.47320759e-01 2.15012655e-01 6.88908696e-02 -6.20214343e-01
-2.27286831e-01 3.63146245e-01 -3.54322270e-02 -2.92414483e-02
4.12814975e-01 1.46359488e-01 -2.53590941e-01 -3.87100875e-01
2.01699615e-01 1.02915573e+00 4.90487456e-01 -4.30847734e-01
1.02447140e+00 5.22812605e-01 -5.77164471e-01 -9.61934924e-01
-7.50854731e-01 -4.76061076e-01 -5.27417421e-01 -4.72375810e-01
8.73564780e-01 -1.09831512e+00 -1.27020061e-01 6.43057883e-01
-8.88483405e-01 -5.93239963e-01 2.34156370e-01 2.48780429e-01
-2.19980359e-01 6.74486816e-01 -6.20065391e-01 -5.62554955e-01
-1.30360842e-01 -1.23733485e+00 1.51457512e+00 1.40727028e-01
-3.65403563e-01 -7.05866933e-01 -4.42286611e-01 8.89225066e-01
1.96934342e-01 1.99766904e-01 4.06234264e-01 -4.96577710e-01
-8.08454573e-01 -3.86976212e-01 -2.93874025e-01 4.55362722e-02
1.20373711e-01 5.22870779e-01 -8.70643735e-01 -2.40599334e-01
-4.57763225e-01 -1.64335400e-01 7.44515479e-01 2.84351110e-01
1.20748055e+00 -4.20652181e-01 -4.02670979e-01 6.23672664e-01
1.01965868e+00 9.54607353e-02 6.82968140e-01 4.05561984e-01
1.07513309e+00 3.02152276e-01 7.52446890e-01 6.05304301e-01
2.70608693e-01 7.70616233e-01 3.14856261e-01 -1.90444496e-02
-2.55685896e-01 -4.16825175e-01 4.96558875e-01 5.40444672e-01
2.85555124e-01 -5.74527025e-01 -1.18643534e+00 5.04860044e-01
-1.60316777e+00 -1.03742433e+00 -3.03980589e-01 1.93943012e+00
7.29788005e-01 1.30008638e-01 5.81959747e-02 -7.04909340e-02
1.27642334e+00 2.86544561e-01 -4.52750891e-01 1.24211729e-01
-3.81677657e-01 -2.95534879e-01 7.23851204e-01 2.50893652e-01
-1.44435096e+00 1.28982067e+00 6.59020805e+00 8.99979234e-01
-1.24847448e+00 -1.12771265e-01 5.70341170e-01 -2.91293293e-01
-1.29101381e-01 4.46997620e-02 -1.06162047e+00 8.55087399e-01
6.23518467e-01 -1.38087533e-02 4.24957961e-01 4.66214329e-01
5.03626943e-01 -2.34806925e-01 -8.17525625e-01 1.15210807e+00
4.01895016e-01 -1.47821200e+00 2.54868120e-01 -9.86026302e-02
9.33600485e-01 2.11081252e-01 1.32837370e-01 1.43550649e-01
3.08674835e-02 -1.04241776e+00 1.15755224e+00 2.59206951e-01
1.27338517e+00 -1.62469730e-01 4.01414126e-01 1.27549078e-02
-1.31954968e+00 3.52038704e-02 -8.91518742e-02 2.04866543e-01
1.22568823e-01 3.29619795e-01 -1.02253830e+00 2.25621790e-01
7.26667821e-01 9.80863214e-01 -1.04508483e+00 1.28545356e+00
2.70156600e-02 9.69933152e-01 -5.01915276e-01 1.89857021e-01
4.98155877e-02 -1.85257196e-01 6.44492507e-01 1.45527720e+00
4.05695021e-01 -2.69520551e-01 2.54704922e-01 6.41232491e-01
-2.90870219e-01 1.44756690e-01 -5.64362407e-01 -3.38308036e-01
6.83992863e-01 1.28316438e+00 -1.18737090e+00 -4.46295440e-01
-7.96254754e-01 8.50708246e-01 -2.78420359e-01 5.97897887e-01
-1.28087032e+00 -3.23957264e-01 7.05184817e-01 3.75026882e-01
4.14876282e-01 -3.61936033e-01 -4.70462352e-01 -1.46998286e+00
3.51700425e-01 -1.17788422e+00 2.82812864e-01 -1.07725525e+00
-9.46228147e-01 2.51374811e-01 -9.52857826e-03 -1.41509831e+00
2.08301425e-01 -5.98501742e-01 -5.15971780e-01 3.44147533e-01
-8.21723044e-01 -1.31506538e+00 -5.81400454e-01 5.99735498e-01
1.04841852e+00 -1.77113563e-01 1.32447854e-01 6.25818729e-01
-9.57308829e-01 6.81823969e-01 1.83034152e-01 6.27367377e-01
1.06956685e+00 -7.75348783e-01 7.46565521e-01 1.24714231e+00
3.70945901e-01 2.83125162e-01 5.36715567e-01 -1.06696200e+00
-1.49899268e+00 -1.37042165e+00 3.76609653e-01 -9.20841634e-01
8.68660986e-01 -7.73640633e-01 -8.22648823e-01 1.12828290e+00
5.98582774e-02 -2.24618137e-01 4.27046657e-01 -2.98630774e-01
-2.47718528e-01 1.33875832e-01 -8.30142975e-01 9.15786028e-01
1.01013005e+00 -4.10389394e-01 -3.89121562e-01 3.64752471e-01
3.90684962e-01 -8.18072975e-01 -6.13361359e-01 2.13054717e-01
6.09259427e-01 -6.81533575e-01 6.83544397e-01 -3.23772617e-02
4.88125384e-01 -5.19304097e-01 -3.21454406e-02 -6.53905809e-01
8.90888721e-02 -7.91574419e-01 -5.63986152e-02 1.51161802e+00
4.75960433e-01 -3.62252444e-01 9.59215581e-01 7.84523904e-01
-3.13244760e-01 -1.70030072e-01 -9.57153738e-01 -4.64358658e-01
-2.76409537e-01 -7.69979239e-01 5.58949232e-01 1.07387710e+00
-1.98484093e-01 -3.93293016e-02 -7.04183340e-01 1.27996221e-01
3.66738796e-01 -1.48051932e-01 1.08645332e+00 -8.35171342e-01
2.33151153e-01 -5.78739583e-01 -5.07573485e-01 -1.24139166e+00
1.18203759e-01 -7.43997693e-01 8.57953876e-02 -1.46206498e+00
1.02175042e-01 -3.52856070e-01 2.23285079e-01 6.91849649e-01
-2.01966017e-01 9.24772084e-01 2.32550323e-01 4.61601198e-01
-9.31376338e-01 1.93654299e-01 1.04264867e+00 -2.73248106e-01
2.03314889e-02 -2.44971842e-01 -5.06601751e-01 8.82178009e-01
9.25386906e-01 -3.32775295e-01 -5.30690588e-02 -7.77256846e-01
2.97655970e-01 -1.44599587e-01 3.91312152e-01 -1.00648010e+00
1.92212686e-01 -2.32435539e-01 6.36182547e-01 -1.00820637e+00
4.29121792e-01 -7.21391797e-01 2.34150335e-01 -2.26765320e-01
-2.19961718e-01 1.73006460e-01 4.50737864e-01 4.11730409e-01
-1.16152115e-01 -2.51965493e-01 4.17207599e-01 1.55192047e-01
-8.57703567e-01 1.58521026e-01 -7.58039951e-01 2.45713755e-01
1.00428689e+00 -5.83151877e-01 -6.31851554e-01 -4.19695199e-01
-2.88858175e-01 3.86692286e-01 8.71655345e-01 7.84720361e-01
5.61554193e-01 -1.14472103e+00 -7.25218177e-01 2.35180154e-01
1.72795206e-01 4.55886573e-02 2.22149745e-01 8.10043573e-01
-9.28656936e-01 3.38051349e-01 6.39730468e-02 -7.29483306e-01
-1.57743979e+00 2.63240308e-01 7.72810429e-02 3.54958147e-01
-8.89658809e-01 4.19524521e-01 2.75124699e-01 -3.53749655e-02
2.44693905e-01 -4.56778586e-01 1.07788686e-02 1.40036374e-01
6.47403955e-01 3.48822773e-01 -2.60281414e-02 -9.21167672e-01
-1.85028106e-01 6.95245862e-01 -1.66351292e-02 -6.72683567e-02
8.65746140e-01 -3.93410534e-01 7.16060027e-02 2.09370688e-01
8.21847498e-01 4.97319877e-01 -1.60539269e+00 -1.34581596e-01
1.16548046e-01 -7.58058906e-01 -2.66196400e-01 -5.89323223e-01
-9.98779714e-01 3.80445004e-01 2.84119368e-01 7.13168383e-02
9.64915931e-01 -9.27972272e-02 7.59314597e-01 2.14077801e-01
1.62265021e-02 -1.28262639e+00 2.22742096e-01 5.73128283e-01
7.01519966e-01 -1.29581237e+00 1.09651253e-01 -5.42001009e-01
-7.60844648e-01 1.16253257e+00 8.07065010e-01 2.77085483e-01
1.57564789e-01 5.45321107e-01 3.41224134e-01 9.37290862e-03
-5.51021159e-01 4.41646203e-02 4.41048950e-01 2.63878256e-01
3.82129967e-01 -1.45919979e-01 9.79108289e-02 2.93521047e-01
-3.46428424e-01 -2.16501325e-01 7.76594460e-01 9.81241465e-01
-2.62661010e-01 -8.43022645e-01 -7.88020790e-01 7.73762822e-01
-8.00191581e-01 -3.73785615e-01 -7.16717958e-01 6.84168041e-01
-3.95881496e-02 1.02667522e+00 2.04811171e-01 -3.83787423e-01
-5.14086969e-02 1.52775496e-01 1.51912093e-01 -5.64563870e-01
-3.05727661e-01 2.84046918e-01 3.07280123e-01 -3.55431378e-01
-3.33593756e-01 -7.46681690e-01 -1.05689228e+00 -2.94867158e-01
-3.10120881e-01 -1.92850694e-01 6.48548186e-01 7.36992002e-01
4.33654219e-01 1.49733439e-01 1.33866102e-01 -1.04119956e+00
2.34416500e-01 -1.07378900e+00 -2.44873077e-01 5.12125552e-01
1.49401844e-01 -5.69818974e-01 -2.60238171e-01 6.36134803e-01] | [11.941200256347656, 2.1904265880584717] |
cf3982ba-409b-4dbe-a085-2fd005d71233 | investigation-of-uncertainty-of-deep-learning | 2106.05870 | null | https://arxiv.org/abs/2106.05870v1 | https://arxiv.org/pdf/2106.05870v1.pdf | Investigation of Uncertainty of Deep Learning-based Object Classification on Radar Spectra | Deep learning (DL) has recently attracted increasing interest to improve object type classification for automotive radar.In addition to high accuracy, it is crucial for decision making in autonomous vehicles to evaluate the reliability of the predictions; however, decisions of DL networks are non-transparent. Current DL research has investigated how uncertainties of predictions can be quantified, and in this article, we evaluate the potential of these methods for safe, automotive radar perception. In particular we evaluate how uncertainty quantification can support radar perception under (1) domain shift, (2) corruptions of input signals, and (3) in the presence of unknown objects. We find that in agreement with phenomena observed in the literature,deep radar classifiers are overly confident, even in their wrong predictions. This raises concerns about the use of the confidence values for decision making under uncertainty, as the model fails to notify when it cannot handle an unknown situation. Accurate confidence values would allow optimal integration of multiple information sources, e.g. via sensor fusion. We show that by applying state-of-the-art post-hoc uncertainty calibration, the quality of confidence measures can be significantly improved,thereby partially resolving the over-confidence problem. Our investigation shows that further research into training and calibrating DL networks is necessary and offers great potential for safe automotive object classification with radar sensors. | ['Bin Yang', 'Michael Pfeiffer', 'Adriana-Eliza Cozma', 'Kilian Rambach', 'William Beluch', 'Kanil Patel'] | 2021-06-01 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [ 2.72502899e-01 1.05642572e-01 8.81322771e-02 -6.75246418e-01
-9.37119961e-01 -5.58647215e-01 6.28749073e-01 1.57085896e-01
-5.10141492e-01 1.14725411e+00 -1.68121159e-01 -6.73469365e-01
-3.94177854e-01 -9.75282848e-01 -8.14876974e-01 -8.17714572e-01
1.31008312e-01 5.75237274e-01 2.77440846e-01 5.40459901e-02
-3.36065353e-03 7.97885716e-01 -1.72299182e+00 9.83525887e-02
8.57584476e-01 1.61084855e+00 -3.42944950e-01 4.78306830e-01
1.18336998e-01 5.62971711e-01 -8.49582613e-01 -6.31190598e-01
2.65139639e-01 -4.57617193e-02 4.85338159e-02 -2.25069374e-01
6.24234378e-01 -5.89236557e-01 7.77817741e-02 1.33197272e+00
2.48290554e-01 -1.63649514e-01 8.34641039e-01 -1.42464113e+00
-2.60017782e-01 4.04637307e-01 -9.97998100e-03 1.49459660e-01
-2.25407973e-01 3.22622925e-01 6.85630739e-01 -6.21563733e-01
1.47944301e-01 1.21020865e+00 6.70150578e-01 2.92975098e-01
-1.12570906e+00 -1.12309802e+00 5.42599633e-02 3.34833831e-01
-1.18327487e+00 -7.42309093e-01 5.60454071e-01 -5.95794201e-01
7.21374333e-01 -1.13566168e-01 3.18557739e-01 1.09875882e+00
8.39970708e-01 9.35685635e-02 1.29173112e+00 -9.19468403e-02
4.84998554e-01 3.74873549e-01 1.56568199e-01 4.36561666e-02
1.09651506e+00 9.84956443e-01 -3.75154942e-01 -4.32371236e-02
1.48058981e-01 -5.16869724e-01 -5.79075776e-02 -2.31765926e-01
-8.17196250e-01 8.20107162e-01 3.89513850e-01 1.35860085e-01
-3.23748976e-01 1.72741026e-01 1.07685931e-01 4.25433040e-01
5.09397626e-01 5.29835939e-01 -4.93637711e-01 9.82783884e-02
-9.96139467e-01 4.28596377e-01 7.43077517e-01 4.49942321e-01
6.25788510e-01 6.06699586e-01 1.50708914e-01 3.71281981e-01
4.46761370e-01 9.80097711e-01 -1.19797505e-01 -1.17763638e+00
2.62994498e-01 1.83609035e-02 3.83011878e-01 -9.37095642e-01
-4.86059010e-01 -7.19708204e-01 -7.54301608e-01 1.00446963e+00
4.48934764e-01 -4.00124043e-01 -8.62012327e-01 1.77943885e+00
-1.12024508e-01 2.88762361e-01 5.60214460e-01 7.78641820e-01
4.86729205e-01 3.04342836e-01 1.02266468e-01 -2.60916233e-01
1.09688866e+00 8.67124349e-02 -7.33439684e-01 -5.99747419e-01
2.99713314e-01 -6.00298226e-01 1.05890930e-01 6.93460405e-01
-5.64073741e-01 -6.60381377e-01 -1.64040768e+00 7.11287320e-01
-4.30352598e-01 -3.23993474e-01 5.20863891e-01 9.51965034e-01
-5.23514450e-01 6.37226939e-01 -7.36287236e-01 5.22149391e-02
3.84472281e-01 2.63185799e-01 -3.39596361e-01 -2.08174601e-01
-1.49066365e+00 1.45616567e+00 3.54399979e-01 4.89281297e-01
-7.39044130e-01 -6.28337443e-01 -9.48074281e-01 -1.59240842e-01
3.41596633e-01 -3.12445998e-01 1.24692035e+00 -8.84970427e-01
-9.05280709e-01 3.39760780e-01 1.61771044e-01 -1.13639140e+00
4.62517500e-01 -2.55246162e-01 -8.65376234e-01 -7.87139609e-02
-1.41722485e-01 7.58122325e-01 9.02528226e-01 -1.42773283e+00
-7.13413656e-01 -4.91096646e-01 -1.53780341e-01 -2.26656958e-01
5.13512552e-01 -2.13413730e-01 5.72158754e-01 -1.47847325e-01
4.15028147e-02 -7.96936810e-01 -1.03926659e-01 -3.67816687e-02
7.94018507e-02 6.37669265e-02 7.97716260e-01 -5.31431377e-01
7.26088166e-01 -1.91160488e+00 -7.17083335e-01 4.23815042e-01
8.41166303e-02 2.86309302e-01 1.58177614e-01 8.25900957e-03
6.85236901e-02 -6.93548052e-03 -3.66518468e-01 6.73096180e-02
1.55978888e-01 3.48747671e-01 -5.93160450e-01 5.20678401e-01
6.54494524e-01 7.01157033e-01 -4.33077067e-01 -2.66080528e-01
3.75293642e-01 4.96583968e-01 -2.45600641e-01 -8.01006556e-02
-3.29603165e-01 1.86799869e-01 -1.08439147e-01 7.21745610e-01
9.50069785e-01 3.43351245e-01 -9.65429023e-02 -4.90837991e-01
-1.34334669e-01 2.21040472e-01 -1.28950047e+00 4.97007102e-01
-1.76236451e-01 1.10002625e+00 2.63250351e-01 -1.07748580e+00
1.05591190e+00 1.22517660e-01 1.31829605e-01 -8.17645013e-01
3.06663871e-01 3.48053038e-01 6.39932871e-01 -6.30344898e-02
4.46145743e-01 -6.61083877e-01 -1.31039828e-01 3.70928459e-02
-7.45196044e-02 -4.12250429e-01 -1.52786657e-01 -1.85320482e-01
7.31217146e-01 -1.74902558e-01 7.48434961e-02 -1.05157875e-01
2.32783958e-01 -3.11725494e-02 7.71039307e-01 8.62180769e-01
-4.00339782e-01 2.22844750e-01 3.83731633e-01 -8.27967003e-02
-7.91708589e-01 -1.30830109e+00 -7.86968887e-01 3.63324702e-01
-2.30431519e-02 3.31900477e-01 -3.55286390e-01 -2.89214015e-01
6.36632979e-01 1.27760732e+00 -4.36708659e-01 -5.41127563e-01
-1.68226451e-01 -7.51604199e-01 5.39556444e-01 7.90645480e-01
3.07915270e-01 -7.32877910e-01 -8.71064603e-01 2.87880093e-01
2.44633570e-01 -1.39752364e+00 5.15830100e-01 5.57196558e-01
-6.04696989e-01 -1.00613093e+00 -7.26771206e-02 8.51197690e-02
3.05509776e-01 1.01347104e-01 9.07770276e-01 -2.50265390e-01
3.07054907e-01 4.93529022e-01 -1.23897478e-01 -8.58543813e-01
-6.83142483e-01 -6.33585870e-01 5.07142127e-01 -1.95600227e-01
6.24304831e-01 -4.05559540e-01 -1.87646404e-01 4.49586481e-01
-7.77152479e-01 -5.68773985e-01 9.37333226e-01 6.54140651e-01
1.78016663e-01 3.89418244e-01 1.01683199e+00 -3.96718919e-01
7.65143156e-01 -3.83090675e-01 -1.11133635e+00 5.26151322e-02
-9.23396051e-01 2.32378066e-01 1.65259629e-01 -2.26118714e-01
-1.05444384e+00 -4.06369984e-01 -1.81391820e-01 -4.45926189e-01
-3.59139830e-01 6.17883921e-01 -2.25415602e-01 -3.19876790e-01
7.74863303e-01 -3.82061213e-01 2.20503896e-01 -8.88024369e-05
3.60695906e-02 7.22455800e-01 6.44344807e-01 -5.41301370e-01
8.73899162e-01 2.09930271e-01 2.45638952e-01 -8.09173226e-01
-8.38737130e-01 1.28083825e-01 -4.00277078e-01 -5.09186447e-01
7.37000763e-01 -9.75028396e-01 -6.02382421e-01 2.26257578e-01
-9.68245745e-01 -7.48109594e-02 -1.74563676e-01 9.51775730e-01
-3.94308150e-01 2.33606979e-01 2.22414397e-02 -1.26539338e+00
9.88241658e-02 -1.19962156e+00 8.16783071e-01 1.25448450e-01
-2.07611889e-01 -8.01950514e-01 -3.39631319e-01 4.07642275e-01
4.15095866e-01 3.43925059e-01 8.06545317e-01 -9.54420924e-01
-6.90143347e-01 -5.69801807e-01 -2.76486427e-01 8.59294951e-01
-1.70563266e-01 6.98616877e-02 -1.24043512e+00 -8.51981863e-02
1.61445513e-01 -4.41367775e-01 1.07374442e+00 5.49593568e-01
6.05650604e-01 -6.89635752e-03 -1.89193517e-01 8.88909027e-02
1.03767526e+00 4.52346623e-01 5.95472038e-01 2.25120723e-01
-2.13136375e-01 7.42179334e-01 9.54484820e-01 1.48673385e-01
1.06848970e-01 4.99518991e-01 7.28503525e-01 3.63367260e-01
9.12332535e-03 1.61283582e-01 3.94693196e-01 -3.26581113e-02
2.31469065e-01 -1.73940778e-01 -7.93308198e-01 2.57086158e-01
-1.56582332e+00 -1.10325408e+00 1.14850625e-01 2.56605864e+00
5.32984793e-01 9.58120704e-01 -2.80140758e-01 2.06053376e-01
3.68150741e-01 -2.95520853e-02 -7.46562243e-01 -4.59592611e-01
-2.56132841e-01 5.35386875e-02 7.57180750e-01 6.69157684e-01
-9.59651411e-01 5.51600397e-01 6.20552635e+00 5.02249062e-01
-1.02406752e+00 -2.32399270e-01 6.85110927e-01 1.78847983e-01
-3.16047519e-01 5.56163043e-02 -9.82886851e-01 3.60042661e-01
1.31420851e+00 -6.96632564e-02 -4.30879630e-02 6.87895596e-01
-2.13645156e-02 -6.21788800e-01 -1.04689682e+00 6.04704559e-01
7.92022049e-02 -1.18604112e+00 -2.90469736e-01 1.43057600e-01
4.23698455e-01 -3.73242423e-02 1.15831703e-01 5.31835973e-01
5.09689867e-01 -1.14577830e+00 7.83970654e-01 8.68971765e-01
4.87323821e-01 -1.04763627e+00 1.37178683e+00 3.89153183e-01
-7.14779973e-01 -1.93473250e-01 -4.42328513e-01 -1.46393657e-01
1.95652574e-01 9.76401806e-01 -9.32941556e-01 4.23709393e-01
4.92424279e-01 1.63772121e-01 -4.94498730e-01 9.58096981e-01
-2.61469811e-01 6.12765312e-01 -5.63475192e-01 2.59601939e-02
1.23423144e-01 -2.23340979e-03 5.97588301e-01 9.96518016e-01
5.06700039e-01 -2.44543273e-02 -4.03092019e-02 9.40175235e-01
4.07646388e-01 -8.20374668e-01 -8.94875884e-01 -1.15723722e-01
6.86715543e-01 9.31553602e-01 -3.90750617e-01 -1.34110108e-01
-2.34109774e-01 -9.21744574e-03 -7.34489933e-02 3.55511099e-01
-7.51896441e-01 -1.36224329e-01 9.93286729e-01 2.69722100e-03
3.47045571e-01 -4.83575821e-01 -5.78130722e-01 -6.01987958e-01
-1.31018415e-01 -7.34520674e-01 6.22235052e-02 -9.64844465e-01
-1.37755120e+00 4.78436202e-01 3.48861068e-01 -1.25469756e+00
-6.24959230e-01 -1.17943263e+00 -3.55788827e-01 7.37940907e-01
-1.59350443e+00 -8.20147693e-01 -3.80323939e-02 6.42925277e-02
2.62809303e-02 -2.64366448e-01 5.35934627e-01 -8.55115578e-02
-1.09730035e-01 4.08285886e-01 -3.76982279e-02 -8.00226852e-02
9.62591708e-01 -7.86787331e-01 -3.32318693e-02 8.33225191e-01
-1.21463276e-01 2.04867363e-01 1.29504716e+00 -9.53274846e-01
-1.03956902e+00 -1.01005137e+00 5.66896141e-01 -5.60635507e-01
7.63237894e-01 -1.55124858e-01 -9.32599545e-01 4.67157096e-01
1.07527733e-01 -1.72424033e-01 6.54006064e-01 9.27188173e-02
-4.88706559e-01 -5.69922745e-01 -1.39508152e+00 3.46724868e-01
4.06991035e-01 -3.56872737e-01 -9.26705539e-01 -9.27440524e-02
6.35573268e-01 -1.44718289e-01 -7.47117877e-01 1.15011883e+00
7.38568902e-01 -1.12150586e+00 8.18686604e-01 -5.07709801e-01
-2.37064175e-02 -4.15553719e-01 -6.87759936e-01 -1.32028210e+00
-2.93485131e-02 2.14306444e-01 9.54181538e-04 1.02757597e+00
6.16131485e-01 -9.53935385e-01 5.35987794e-01 8.88049543e-01
-2.49862894e-01 -3.62201989e-01 -1.34848511e+00 -9.59200382e-01
3.06463808e-01 -1.34481728e+00 3.47321063e-01 4.55342293e-01
-4.47521240e-01 1.65357664e-01 -2.49222711e-01 6.51922762e-01
9.66121733e-01 -7.60507137e-02 5.28329909e-01 -1.79750097e+00
7.90919065e-02 -4.16399360e-01 -8.49995613e-01 -4.19401139e-01
4.61865157e-01 -3.31525713e-01 3.17745447e-01 -1.35402310e+00
-4.82452393e-01 -4.51661408e-01 -2.48465613e-01 2.36834198e-01
2.42384344e-01 1.77242324e-01 2.32978314e-01 -1.84777632e-01
-1.83977351e-01 4.97095406e-01 6.67385161e-01 -2.42889926e-01
3.40030760e-01 4.27775323e-01 -7.80409873e-01 7.69686818e-01
9.52642381e-01 -3.14958721e-01 -2.59870827e-01 8.06149840e-03
4.43059266e-01 2.26142369e-02 7.66399384e-01 -1.55847692e+00
3.44601393e-01 -1.75290078e-01 7.86827445e-01 -8.63558352e-01
6.58170402e-01 -1.32093441e+00 1.85024127e-01 3.01453203e-01
6.82461783e-02 -1.73571542e-01 7.17072725e-01 6.54356003e-01
-2.05004171e-01 -2.95640945e-01 9.39598143e-01 1.51053861e-01
-7.24512815e-01 -2.50605047e-01 -6.79352105e-01 -3.02284330e-01
8.73848021e-01 -3.36163580e-01 -4.70311850e-01 -6.34039700e-01
-5.98725319e-01 3.11605692e-01 1.64088786e-01 3.76210719e-01
7.08279490e-01 -1.34998310e+00 -7.32090414e-01 2.88056165e-01
1.36172459e-01 -3.16502303e-01 1.80088595e-01 6.76344514e-01
1.62147000e-01 7.36146033e-01 -2.56192744e-01 -7.02686787e-01
-8.46572161e-01 3.01558375e-01 5.97512126e-01 1.27370253e-01
-5.45407785e-03 5.26323736e-01 -1.84962954e-02 -1.16133362e-01
3.54422361e-01 -4.18929756e-01 -9.05739442e-02 2.14840114e-01
4.73438948e-01 1.39741525e-01 4.00219113e-01 -5.12836337e-01
-4.27109003e-01 3.64430219e-01 2.38066372e-02 -3.12011719e-01
8.15768242e-01 1.46429166e-01 3.08613777e-01 5.99212170e-01
4.89766657e-01 -6.27570227e-03 -1.48757017e+00 -1.00273304e-02
5.20396903e-02 -2.14239419e-01 3.79295200e-01 -1.17305589e+00
-7.38403916e-01 1.04752076e+00 8.06752384e-01 1.03941113e-01
8.61707985e-01 -1.15353189e-01 2.43467212e-01 6.88596606e-01
4.84339684e-01 -1.11768270e+00 -3.26507747e-01 5.47901273e-01
9.01226580e-01 -1.44310772e+00 1.97532654e-01 7.27253854e-02
-7.33598828e-01 1.13945413e+00 7.32042849e-01 1.19682364e-01
8.96509111e-01 7.28187919e-01 1.08195506e-01 2.08505429e-02
-9.95204628e-01 -3.68486822e-01 4.75777328e-01 8.56756806e-01
1.57270059e-01 1.92933708e-01 6.67536035e-02 8.00776899e-01
-3.29523176e-01 -1.40282974e-01 4.32043880e-01 5.95246494e-01
-9.89917994e-01 -6.65664315e-01 -8.10266554e-01 6.82205617e-01
-2.04613078e-02 1.23401985e-01 -2.96487600e-01 1.04597318e+00
3.23996603e-01 1.36939073e+00 3.34522337e-01 -5.07141948e-01
4.32959467e-01 2.49639273e-01 2.95473695e-01 -1.75596938e-01
7.44965449e-02 -2.54890680e-01 5.37138522e-01 -2.88577914e-01
-3.31987232e-01 -8.03730190e-01 -9.67719018e-01 -7.83643574e-02
-4.00453389e-01 1.44933701e-01 7.51564503e-01 1.20604157e+00
1.87556803e-01 6.05059445e-01 1.35592937e-01 -7.42843330e-01
-9.97234583e-01 -1.00804412e+00 -3.82170379e-01 -4.19428587e-01
5.31196833e-01 -1.23608327e+00 -7.87292719e-01 -4.41897064e-01] | [7.530520915985107, 3.8346385955810547] |
3887619f-e820-42ae-a4fa-323df30171aa | end-to-end-semantics-based-summary-quality | 2005.06377 | null | https://arxiv.org/abs/2005.06377v3 | https://arxiv.org/pdf/2005.06377v3.pdf | SueNes: A Weakly Supervised Approach to Evaluating Single-Document Summarization via Negative Sampling | Canonical automatic summary evaluation metrics, such as ROUGE, focus on lexical similarity which cannot well capture semantics nor linguistic quality and require a reference summary which is costly to obtain. Recently, there have been a growing number of efforts to alleviate either or both of the two drawbacks. In this paper, we present a proof-of-concept study to a weakly supervised summary evaluation approach without the presence of reference summaries. Massive data in existing summarization datasets are transformed for training by pairing documents with corrupted reference summaries. In cross-domain tests, our strategy outperforms baselines with promising improvements, and show a great advantage in gauging linguistic qualities over all metrics. | ['Cen Chen', 'Minghui Qiu', 'Youbiao He', 'Hebi Li', 'Yinfei Yang', 'Ge Luo', 'Forrest Sheng Bao'] | 2020-05-13 | null | https://aclanthology.org/2022.naacl-main.175 | https://aclanthology.org/2022.naacl-main.175.pdf | naacl-2022-7 | ['document-embedding'] | ['methodology'] | [ 1.16831802e-01 -5.52056804e-02 -4.58758742e-01 -4.19524789e-01
-1.64286184e+00 -7.50289023e-01 1.02190042e+00 9.18897331e-01
-4.82392490e-01 1.07925141e+00 1.08401239e+00 6.84198141e-02
-1.55167028e-01 -5.05904794e-01 -2.88006216e-01 -1.18842155e-01
2.98667371e-01 3.45035523e-01 2.73697764e-01 -3.65336746e-01
6.94655418e-01 -1.02275051e-01 -1.42977214e+00 3.15062165e-01
1.33383906e+00 6.49477661e-01 5.72832562e-02 7.47268677e-01
-2.81440139e-01 6.53860688e-01 -1.05109596e+00 -6.85746491e-01
-2.59723991e-01 -5.81522107e-01 -6.99572384e-01 -1.21227078e-01
8.60557735e-01 -2.64812201e-01 -1.04082339e-01 9.34732318e-01
8.68795276e-01 2.51543105e-01 8.10818315e-01 -1.04550803e+00
-7.85653889e-01 1.02839696e+00 -3.10382575e-01 1.50854468e-01
8.13753545e-01 -1.27148926e-01 1.57649064e+00 -6.91528976e-01
6.93259001e-01 1.08873022e+00 6.64125323e-01 3.63302976e-01
-1.30496514e+00 -2.65665025e-01 -1.36982113e-01 -5.91649860e-02
-9.13824677e-01 -7.28626490e-01 6.71266973e-01 -2.13665262e-01
1.17736006e+00 4.23332453e-01 1.64413393e-01 1.11054230e+00
-3.17026675e-02 8.28530371e-01 9.00854588e-01 -4.36397642e-01
1.35936469e-01 2.50702381e-01 5.72159588e-01 2.94918925e-01
5.88871062e-01 -4.04281676e-01 -5.64950168e-01 -3.88190329e-01
1.23889081e-01 -5.47514796e-01 -2.76002675e-01 -5.97918294e-02
-1.35634208e+00 8.59971166e-01 1.26948673e-02 3.39458257e-01
-2.82077134e-01 -3.01572122e-02 8.94324601e-01 4.21649039e-01
7.66755104e-01 8.23699594e-01 -2.97644556e-01 -4.18259710e-01
-1.39037859e+00 5.67989230e-01 9.57325220e-01 8.69121611e-01
3.87915105e-01 -7.56865889e-02 -4.72613305e-01 9.99988914e-01
-5.13398238e-02 4.52772021e-01 8.29587340e-01 -1.10602796e+00
7.36038566e-01 5.99392831e-01 3.85544300e-01 -1.15855098e+00
-3.33162576e-01 -4.24722433e-01 -5.85635364e-01 -4.49440598e-01
1.79310173e-01 7.58296177e-02 -2.76693434e-01 1.68694210e+00
-1.15411170e-01 -2.96224922e-01 4.46546197e-01 5.96238017e-01
1.34565890e+00 5.62973022e-01 -1.08686745e-01 -4.89852220e-01
1.10492361e+00 -9.93615866e-01 -8.22065532e-01 -2.05975637e-01
9.08258379e-01 -1.08315766e+00 1.68894589e+00 2.63721913e-01
-1.50334156e+00 -5.66232741e-01 -1.37110722e+00 -2.41401687e-01
-2.42537335e-01 2.72610962e-01 5.41687071e-01 7.54140615e-01
-1.21296942e+00 8.84232104e-01 -5.19361675e-01 -5.05750477e-01
-1.01154996e-03 1.48267671e-02 -3.66819739e-01 6.24580216e-03
-1.20583904e+00 9.53773558e-01 4.81023610e-01 -6.46814823e-01
-1.80716559e-01 -5.55447519e-01 -1.02628982e+00 7.00985417e-02
2.40076572e-01 -7.80071557e-01 1.33838785e+00 -4.17590380e-01
-1.21319652e+00 7.96016991e-01 -1.60150051e-01 -3.81975591e-01
5.21521628e-01 -3.65644604e-01 -4.76138681e-01 9.56708863e-02
5.86138546e-01 5.14423728e-01 1.47255287e-01 -9.98647213e-01
-5.00429213e-01 -5.60852215e-02 -6.41318634e-02 3.76936078e-01
-5.99655986e-01 2.61527777e-01 -1.67035311e-01 -8.55610967e-01
-2.37764671e-01 -4.13927674e-01 6.22805245e-02 -7.24256217e-01
-4.04259920e-01 -5.65968096e-01 4.06037837e-01 -6.10708416e-01
1.79337239e+00 -1.87346220e+00 -1.57500319e-02 -4.12049741e-01
-3.28217857e-02 3.78164738e-01 -3.34553182e-01 8.21221590e-01
3.96497607e-01 4.20533568e-01 -2.31344104e-01 -6.07944787e-01
3.29792321e-01 -2.17030421e-02 -4.39799190e-01 1.25486597e-01
1.14384227e-01 8.99945080e-01 -1.17882884e+00 -9.07315195e-01
-2.42545083e-01 -9.67346355e-02 -4.82956737e-01 2.37533569e-01
-3.54408085e-01 3.74302901e-02 -3.34064901e-01 3.72877598e-01
2.60556161e-01 -2.17368528e-01 -6.04185164e-02 -9.53209102e-02
-1.53419971e-01 9.02717590e-01 -8.89010429e-01 1.98324609e+00
-3.45351934e-01 4.92644668e-01 -4.26938266e-01 -7.52545953e-01
9.66969490e-01 2.65395761e-01 3.26644816e-02 -7.06713498e-01
1.47969201e-02 5.26435912e-01 -2.78237969e-01 -3.31695408e-01
1.31182563e+00 -1.49061859e-01 -6.70574069e-01 7.21338511e-01
-7.69972354e-02 -5.80143094e-01 7.54595459e-01 6.91076875e-01
1.27450585e+00 8.34388379e-03 5.87870300e-01 -2.37944663e-01
6.03431344e-01 2.62953669e-01 4.24996614e-01 7.96964109e-01
-1.10356815e-01 8.50805640e-01 6.40551865e-01 2.57416159e-01
-1.01929855e+00 -1.03341115e+00 8.29043314e-02 1.02497685e+00
9.55447555e-03 -9.83510494e-01 -6.27100825e-01 -7.23899066e-01
-6.74483925e-02 1.26412940e+00 -1.69641882e-01 -1.55388817e-01
-4.59718049e-01 -3.86123806e-01 8.40782642e-01 6.35640204e-01
4.05891776e-01 -8.22312057e-01 -5.01780927e-01 3.91222268e-01
-6.72585785e-01 -1.16698790e+00 -7.19084740e-01 -2.30733812e-01
-9.00336564e-01 -6.33259177e-01 -7.17047453e-01 -5.66097498e-01
-6.82140216e-02 4.18505132e-01 1.60514510e+00 1.88801326e-02
2.73796111e-01 4.36618894e-01 -5.33361554e-01 -3.22136462e-01
-6.64126456e-01 4.00327951e-01 1.39938220e-01 -7.33201802e-01
3.43750179e-01 -3.71120691e-01 -3.77202690e-01 1.00689873e-01
-9.94650126e-01 -2.09693730e-01 5.56325376e-01 8.43982041e-01
3.82114291e-01 -1.99729204e-01 1.42053151e+00 -8.30463290e-01
1.37824857e+00 -3.96030605e-01 7.62105547e-03 4.37818050e-01
-6.46645367e-01 1.25856176e-01 6.92495465e-01 -1.18602589e-01
-1.09672511e+00 -7.13577271e-01 -8.66992921e-02 1.45528883e-01
5.85444719e-02 9.77319837e-01 -2.61107653e-01 6.44060671e-01
7.97006130e-01 -5.10048382e-02 -2.63482302e-01 -3.78826410e-01
4.57162142e-01 8.60665739e-01 7.57731557e-01 -6.29639328e-01
4.10495669e-01 4.39918451e-02 -3.19550574e-01 -8.61912727e-01
-1.16965175e+00 -6.71272576e-01 -4.49273020e-01 1.50843084e-01
4.22586709e-01 -8.80937934e-01 3.11449766e-02 4.24035788e-02
-1.17798543e+00 -6.20239787e-02 -4.02629763e-01 5.41316748e-01
-7.03167856e-01 8.67154360e-01 -4.79862779e-01 -5.05889177e-01
-6.56078994e-01 -7.42466867e-01 1.09914458e+00 1.26718417e-01
-9.19183314e-01 -1.10157931e+00 3.39481682e-01 3.65882427e-01
5.11108994e-01 2.23895147e-01 8.82125795e-01 -9.92993891e-01
-7.71590602e-03 -4.27626729e-01 -2.95998096e-01 5.72650731e-01
2.87252247e-01 1.28767177e-01 -6.23411357e-01 -2.67745912e-01
-4.49351966e-02 -7.67185211e-01 1.02082634e+00 2.38517314e-01
4.73441392e-01 -4.73136216e-01 1.68958511e-02 6.60163313e-02
1.17538476e+00 -2.67514110e-01 4.29389775e-01 3.51136923e-01
3.16514522e-01 6.30752683e-01 9.99847174e-01 5.54572523e-01
6.88746810e-01 6.04472995e-01 -3.82319689e-01 3.59343052e-01
-3.60819697e-01 -3.99673641e-01 4.26777720e-01 1.36366403e+00
2.26037771e-01 -6.40824974e-01 -6.81464851e-01 7.05316365e-01
-1.89704168e+00 -8.78442228e-01 -8.54403675e-02 2.26719713e+00
1.11404765e+00 4.01915073e-01 3.10259819e-01 2.70631015e-01
4.78399307e-01 5.22602499e-01 -1.86123163e-01 -5.50556540e-01
-5.05327940e-01 -9.73812416e-02 6.01422302e-02 4.50390369e-01
-8.61187100e-01 8.81949663e-01 6.37794638e+00 7.73986220e-01
-8.36195648e-01 -1.25081465e-01 3.11183721e-01 -1.17202163e-01
-8.05636168e-01 -3.02281342e-02 -6.24216735e-01 4.06358570e-01
1.16260874e+00 -7.05359995e-01 -1.78835541e-01 6.52215838e-01
1.21962786e-01 -3.35763663e-01 -1.30097306e+00 7.42924333e-01
3.87480825e-01 -1.15472198e+00 1.22096039e-01 -3.28617245e-01
7.91642845e-01 -4.63006496e-02 -5.11180870e-02 3.28530461e-01
3.90560985e-01 -8.15679371e-01 6.36566639e-01 2.81907290e-01
8.90450180e-01 -7.28900909e-01 9.88941669e-01 3.94727111e-01
-7.65479505e-01 4.59574640e-01 -5.15233636e-01 -1.41786903e-01
2.88428932e-01 5.40872872e-01 -7.33091176e-01 7.73425937e-01
2.56903350e-01 7.81886101e-01 -6.98589981e-01 9.01883006e-01
-2.80362606e-01 7.03917265e-01 -6.88977018e-02 -3.93082947e-01
3.44200671e-01 1.49107268e-02 7.43459582e-01 1.59783161e+00
3.96226376e-01 -3.04062851e-02 2.08419368e-01 5.46961606e-01
-2.61406124e-01 3.61024469e-01 -6.32424355e-01 -3.64963979e-01
6.96558237e-01 1.02931690e+00 -6.90491080e-01 -6.91168070e-01
-5.12297332e-01 7.82373607e-01 3.66935074e-01 1.84456389e-02
-6.19410634e-01 -6.91804290e-01 2.68638015e-01 -8.31606891e-03
5.61061203e-02 -2.88562596e-01 -6.46835506e-01 -1.28799784e+00
2.09077030e-01 -8.85986567e-01 5.00384212e-01 -5.70863724e-01
-1.36606157e+00 5.59018672e-01 1.67272538e-01 -1.17175913e+00
-5.28351724e-01 -3.08148284e-03 -7.72705078e-01 8.15982401e-01
-1.53597724e+00 -8.71406138e-01 -2.94624478e-01 8.87925997e-02
5.40103912e-01 -2.24390835e-01 8.31439078e-01 1.89693078e-01
-3.55020255e-01 7.78999865e-01 7.56671205e-02 -3.52165401e-01
1.26628256e+00 -1.43177629e+00 5.37213504e-01 8.10260594e-01
7.35639259e-02 5.77864230e-01 1.08847976e+00 -8.04352045e-01
-9.08427715e-01 -7.86143363e-01 1.49621153e+00 -7.13183880e-01
8.09345841e-01 3.36855464e-02 -1.05003095e+00 3.85850906e-01
4.97940183e-01 -6.74631774e-01 8.90196502e-01 2.06155300e-01
-4.47578818e-01 1.02304965e-01 -1.03558719e+00 5.62464356e-01
9.26924765e-01 -4.29889381e-01 -1.25748801e+00 2.27571383e-01
8.35586786e-01 -6.78291917e-02 -9.17566061e-01 4.57372695e-01
4.15606380e-01 -9.38400388e-01 7.53971338e-01 -5.49177766e-01
7.64308333e-01 -7.87843391e-02 -4.08454955e-01 -1.66558588e+00
-5.16851582e-02 -6.15466416e-01 6.64166883e-02 1.93038869e+00
2.91669369e-01 -4.73091453e-01 5.53553462e-01 4.70676720e-01
-3.92787576e-01 -3.19196641e-01 -6.26379073e-01 -1.01163554e+00
2.43575484e-01 -1.85745478e-01 6.04051292e-01 8.04405868e-01
6.04482651e-01 8.56891096e-01 -1.02582090e-01 -4.43581015e-01
4.89800602e-01 1.69198036e-01 7.98431396e-01 -1.12600541e+00
-7.81975687e-02 -8.18256974e-01 -1.54747650e-01 -1.03469634e+00
2.46551782e-01 -8.24260831e-01 3.93487774e-02 -1.97894430e+00
3.83714467e-01 -2.58301258e-01 -1.43029973e-01 2.84965169e-02
-5.43856204e-01 3.14850472e-02 -1.22742794e-01 2.14448795e-01
-1.03833568e+00 6.58395529e-01 9.38881338e-01 -7.40204006e-02
-2.17160150e-01 -2.46921197e-01 -1.13704956e+00 6.34119034e-01
9.21839535e-01 -2.83695996e-01 -5.38580000e-01 -5.18696129e-01
3.99715215e-01 2.68591166e-01 -4.14574742e-02 -9.38212812e-01
2.58985281e-01 6.58114478e-02 -5.31315021e-02 -7.88415670e-01
1.41258195e-01 -1.18750446e-01 -3.60402644e-01 1.19917326e-01
-7.73508489e-01 4.65736806e-01 9.40158367e-02 5.12561858e-01
-5.46999097e-01 -5.01307130e-01 4.90637749e-01 -9.95423943e-02
-3.70100051e-01 -3.48864883e-01 3.17980312e-02 5.91743946e-01
4.95269775e-01 -1.28727660e-01 -7.04294562e-01 -5.01887083e-01
-5.99539690e-02 2.52695620e-01 7.39200711e-01 3.46940547e-01
5.17799437e-01 -1.22048581e+00 -1.26310885e+00 -4.46304023e-01
3.51676732e-01 -3.20349276e-01 -1.75130405e-02 6.67697608e-01
-2.48248845e-01 6.97279155e-01 -2.86254156e-02 -2.76473194e-01
-1.35775137e+00 4.89523023e-01 -3.77778739e-01 -5.96863329e-01
-4.08776790e-01 6.11000121e-01 -6.50683865e-02 -3.09595615e-01
2.52294928e-01 -3.34149361e-01 -3.71457845e-01 4.97234076e-01
6.84534907e-01 7.71823227e-01 2.71894783e-01 -6.46569490e-01
-2.32126400e-01 3.01942527e-01 -1.17191486e-01 -3.45467836e-01
1.17475593e+00 -3.57794464e-01 -8.72042868e-03 6.61020041e-01
1.20610034e+00 3.60872567e-01 -5.94226122e-01 -4.30836082e-01
5.76406240e-01 -4.32486743e-01 2.48085819e-02 -9.69115496e-01
-3.08694333e-01 6.73838913e-01 -1.77901208e-01 4.57150400e-01
9.59240615e-01 1.36943847e-01 1.02924824e+00 4.55211759e-01
1.81261659e-01 -1.17802680e+00 3.11413884e-01 5.43131232e-01
1.12806177e+00 -1.40912068e+00 2.88686842e-01 -1.41184181e-01
-9.96388197e-01 8.77719104e-01 3.63925725e-01 -4.26974073e-02
3.57553400e-02 5.47902063e-02 4.26148809e-03 -1.44611925e-01
-9.42520618e-01 -9.69328731e-02 5.65293550e-01 3.82333487e-01
8.74497592e-01 6.99650496e-02 -9.78761017e-01 8.70710373e-01
-6.58318281e-01 -2.71358997e-01 7.37963974e-01 7.23437726e-01
-6.26120627e-01 -1.10402727e+00 -8.81603453e-03 7.11189866e-01
-7.16363966e-01 -2.52756238e-01 -7.05378652e-01 8.03774297e-01
-6.49968266e-01 1.37359476e+00 -1.05032563e-01 -3.21959764e-01
6.57866001e-01 1.58411600e-02 4.55147058e-01 -8.72037351e-01
-6.41198635e-01 2.34633684e-02 7.73798943e-01 -2.27014869e-01
-5.15284300e-01 -1.01964474e+00 -1.02450454e+00 -4.54460025e-01
-4.49009746e-01 5.37604988e-01 4.24393773e-01 7.24406481e-01
1.56158909e-01 3.87614399e-01 4.83994395e-01 -6.16382837e-01
-9.29198861e-01 -1.23512661e+00 -4.49169934e-01 5.10282815e-01
2.19931543e-01 -3.91206592e-01 -4.17722732e-01 -2.02497736e-01] | [12.136645317077637, 9.35391902923584] |
09a3d0cb-bf0f-476a-88f3-bdff64a48985 | fast-key-points-detection-and-matching-for | 2211.03242 | null | https://arxiv.org/abs/2211.03242v2 | https://arxiv.org/pdf/2211.03242v2.pdf | Fast Key Points Detection and Matching for Tree-Structured Images | This paper offers a new authentication algorithm based on image matching of nano-resolution visual identifiers with tree-shaped patterns. The algorithm includes image-to-tree conversion by greedy extraction of the fractal pattern skeleton along with a custom-built graph matching algorithm that is robust against imaging artifacts such as scaling, rotation, scratch, and illumination change. The proposed algorithm is applicable to a variety of tree-structured image matching, but our focus is on dendrites, recently-developed visual identifiers. Dendrites are entropy rich and unclonable with existing 2D and 3D printers due to their natural randomness, nano-resolution granularity, and 3D facets, making them an appropriate choice for security applications such as supply chain trace and tracking. The proposed algorithm improves upon graph matching with standard image descriptors. For instance, image inconsistency due to the camera sensor noise may cause unexpected feature extraction leading to inaccurate tree conversion and authentication failure. Also, previous tree extraction algorithms are prohibitively slow hindering their scalability to large systems. In this paper, we fix the current issues of [1] and accelerate the key points extraction up to 10-times faster by implementing a new skeleton extraction method, a new key points searching algorithm, as well as an optimized key point matching algorithm. Using minimum enclosing circle and center points, make the algorithm robust to the choice of pattern shape. In contrast to [1] our algorithm handles general graphs with loop connections, therefore is applicable to a wider range of applications such as transportation map analysis, fingerprints, and retina vessel imaging. | ['Rahul Amin', 'Abolfazl Razi', 'Xiwen Chen', 'Hao Wang'] | 2022-11-07 | null | null | null | null | ['graph-matching', 'key-point-matching'] | ['graphs', 'natural-language-processing'] | [ 5.78844726e-01 -2.35028133e-01 -3.43034491e-02 4.23231840e-01
-2.17639968e-01 -8.60523760e-01 4.45345700e-01 3.64941180e-01
-1.48052111e-01 1.59458533e-01 -3.58638525e-01 -5.16967475e-01
-3.33067060e-01 -9.47465777e-01 -4.39685673e-01 -6.17688179e-01
-4.17698883e-02 1.56545639e-01 4.52086806e-01 -4.95515056e-02
7.08740234e-01 1.04758406e+00 -1.51460791e+00 -3.01063925e-01
6.64863586e-01 1.32243133e+00 -1.15029350e-01 6.77552342e-01
-2.71465689e-01 3.02333474e-01 -4.01163667e-01 -6.47629499e-01
5.21290839e-01 -1.72271147e-01 -2.28630468e-01 2.98169963e-02
5.71620822e-01 -1.81306839e-01 -2.58548230e-01 1.01896310e+00
5.38900554e-01 -4.82153416e-01 7.12931335e-01 -1.13807237e+00
-5.41665912e-01 2.70024147e-02 -8.79341841e-01 -2.77372360e-01
5.94342411e-01 5.89531735e-02 3.38775873e-01 -4.50323433e-01
7.08330512e-01 9.35311079e-01 1.06365991e+00 4.19660062e-01
-1.10971785e+00 -7.78742731e-01 -4.57349360e-01 -4.40662690e-02
-1.38574588e+00 -2.15972304e-01 9.29423571e-01 -3.86652559e-01
7.46846259e-01 4.30350572e-01 5.98782480e-01 7.70014763e-01
6.23737872e-01 -1.46572724e-01 1.35606074e+00 -6.11656368e-01
2.02401161e-01 1.01750381e-02 -1.46312073e-01 1.26796341e+00
7.00773120e-01 1.07414737e-01 -3.65880698e-01 -4.40012038e-01
1.15789032e+00 -8.78247693e-02 -6.06568269e-02 -6.42043889e-01
-1.13029289e+00 3.00853282e-01 4.13780846e-02 2.77516007e-01
8.52768049e-02 9.53859389e-02 1.30828142e-01 4.28361863e-01
-1.14476241e-01 4.46925223e-01 1.69570878e-01 -1.32192194e-01
-6.73258722e-01 -8.38163272e-02 7.96207964e-01 1.16909301e+00
7.29260802e-01 -7.15735853e-02 1.89677969e-01 4.31510031e-01
2.16817975e-01 7.35039771e-01 2.90372044e-01 -9.14921165e-01
2.90573657e-01 7.06799626e-01 9.08391774e-02 -1.57450747e+00
-5.53626597e-01 -1.95637107e-01 -1.08791959e+00 5.86207211e-01
7.11579382e-01 4.47763473e-01 -8.71909380e-01 1.22557020e+00
3.65838915e-01 -9.20352787e-02 -9.36711580e-02 5.90303302e-01
7.24163234e-01 2.77738515e-02 -3.07049841e-01 -2.51087129e-01
1.60612047e+00 -4.06785280e-01 -4.96316284e-01 4.02491659e-01
2.93056428e-01 -1.05616331e+00 8.82456899e-01 4.14770752e-01
-8.56122196e-01 -4.25291479e-01 -1.25859034e+00 7.79149309e-02
-5.96862614e-01 -5.90505227e-02 5.02638876e-01 1.10906172e+00
-9.22022760e-01 7.24303186e-01 -3.47492665e-01 -7.38190114e-01
4.28645253e-01 6.94876432e-01 -4.13767546e-01 2.28972033e-01
-6.66277111e-01 6.01308048e-01 -2.62347311e-01 -3.13773960e-01
3.73189718e-01 -4.57231641e-01 -5.38271904e-01 -1.25376374e-01
5.29083014e-02 -8.71989489e-01 4.97315258e-01 -3.64659697e-01
-1.80671990e+00 1.03622270e+00 -1.82535779e-02 -4.24697787e-01
6.15788817e-01 2.02484652e-01 -6.14486992e-01 3.45372438e-01
-2.03918442e-02 8.61919597e-02 1.20130646e+00 -9.90909159e-01
-1.27287209e-01 -3.41471672e-01 -4.31848198e-01 -2.94391125e-01
-3.28516334e-01 -3.87786999e-02 -3.93844903e-01 -6.88943505e-01
3.59188855e-01 -6.74677253e-01 -1.10418566e-01 2.56984204e-01
-2.48244494e-01 1.91545263e-01 1.09004343e+00 -2.18285590e-01
1.26690602e+00 -2.09181046e+00 -5.02028048e-01 7.85336554e-01
5.09548724e-01 9.80266929e-02 -1.65091902e-01 6.69613123e-01
2.08480433e-01 2.78126717e-01 -4.17296961e-02 2.11707667e-01
-1.28451318e-01 -2.68301427e-01 -9.16324407e-02 7.33053446e-01
-1.25500783e-01 5.41584432e-01 -5.79891741e-01 -6.54605210e-01
3.83177549e-01 2.65265137e-01 8.43510218e-03 -3.22107285e-01
3.16209018e-01 1.98040083e-01 -5.66661954e-01 9.66851354e-01
9.82162476e-01 -2.63069928e-01 -8.27664584e-02 -5.67827404e-01
-4.57870483e-01 -2.68284827e-01 -1.28951406e+00 1.33483684e+00
-1.39889553e-01 3.10153544e-01 1.59979556e-02 -5.40094197e-01
1.41823030e+00 2.22640969e-02 8.47175300e-01 -9.52928901e-01
1.26055062e-01 3.76379043e-01 -4.43355203e-01 -3.32292646e-01
4.11609590e-01 1.41097397e-01 -4.85941842e-02 6.25879884e-01
-5.01675248e-01 -1.57966062e-01 -8.11479837e-02 1.16618939e-01
1.40211821e+00 2.64825970e-02 2.76732534e-01 -3.06895226e-01
5.76828718e-01 -6.11208491e-02 2.78023452e-01 7.94036329e-01
-4.55078185e-02 6.09252810e-01 3.11238527e-01 -6.29454553e-01
-1.08742440e+00 -9.81609106e-01 -2.21945673e-01 -2.96156630e-02
6.97634399e-01 -7.85671175e-01 -6.79002225e-01 -3.54930252e-01
3.16830665e-01 -2.41563097e-01 -2.77686536e-01 5.68012521e-02
-6.23085916e-01 -4.27614331e-01 5.90671897e-01 -1.58990212e-02
8.11063886e-01 -3.16702574e-01 -9.64248478e-01 3.69867504e-01
3.31456512e-01 -1.27298331e+00 -4.30915296e-01 -3.02051157e-01
-9.43923354e-01 -1.39149868e+00 -4.97129470e-01 -6.75983191e-01
8.75290513e-01 3.95451814e-01 7.78169096e-01 3.35188121e-01
-6.76726520e-01 8.12448204e-01 -5.55628315e-02 -2.85721511e-01
-5.20523250e-01 -1.25054464e-01 3.14974397e-01 1.62493795e-01
1.17315814e-01 -7.39576161e-01 -8.56115878e-01 6.02763593e-01
-5.88047743e-01 -4.99158073e-03 5.25925696e-01 5.97508729e-01
7.27046192e-01 2.10367978e-01 1.11775272e-01 -6.68436348e-01
5.40786147e-01 1.71642870e-01 -1.01582122e+00 1.82371795e-01
-1.04950082e+00 9.88149643e-02 8.03994954e-01 -4.70326751e-01
-5.46662092e-01 2.67590374e-01 3.57209355e-01 -3.22285682e-01
1.24610409e-01 -1.67928845e-01 -8.26722309e-02 -1.05805254e+00
6.19574428e-01 6.06208891e-02 3.72655809e-01 -3.92759115e-01
1.89273804e-01 6.66635692e-01 5.68267107e-01 -4.78716165e-01
1.41851926e+00 7.41056085e-01 8.01468611e-01 -1.04164302e+00
3.06120306e-01 -3.17384362e-01 -4.32603508e-01 -2.87592888e-01
6.30131066e-01 -4.31582898e-01 -1.13592911e+00 7.83500671e-01
-1.03573883e+00 1.28632084e-01 -1.38258010e-01 2.70545214e-01
-5.10052443e-01 9.48322654e-01 -4.72767621e-01 -7.65760839e-01
-5.32156289e-01 -9.18982744e-01 1.01189899e+00 3.39721322e-01
-2.77955204e-01 -7.44748712e-01 -7.76579157e-02 1.15507744e-01
4.88715976e-01 7.27068067e-01 1.25421810e+00 5.99244162e-02
-9.30766821e-01 -3.31623435e-01 -2.94074625e-01 -3.29696864e-01
3.53720188e-01 2.98484892e-01 -6.56643748e-01 -2.74909884e-01
-8.38112310e-02 2.51727521e-01 4.99395996e-01 2.37708837e-02
9.61327374e-01 -3.34058732e-01 -7.18913019e-01 7.93304026e-01
1.59898388e+00 1.90566838e-01 7.89193630e-01 5.89941144e-01
5.65697432e-01 4.62021559e-01 4.27197695e-01 3.52748781e-01
1.37816835e-02 7.82355607e-01 3.01016837e-01 -2.74408907e-01
-2.96531171e-01 -2.06431493e-01 1.73552066e-01 6.93041444e-01
-2.34504372e-01 3.71159650e-02 -7.06197441e-01 2.12659799e-02
-1.39563763e+00 -8.28871906e-01 -3.04054052e-01 2.42577004e+00
2.60810614e-01 7.19827740e-03 2.23597288e-01 2.03551874e-01
8.38793457e-01 -1.29772112e-01 -4.73574251e-01 -5.32622874e-01
-1.85895175e-01 4.35825586e-01 1.17823577e+00 2.07006782e-01
-8.99243116e-01 6.44837737e-01 6.27130222e+00 9.00745332e-01
-1.26665008e+00 -2.93058664e-01 2.10257143e-01 4.22446817e-01
-4.33982402e-01 1.36336729e-01 -6.53524399e-01 3.80826026e-01
2.80832350e-01 -2.01816559e-02 2.48267204e-01 3.73519659e-01
-2.44915262e-02 -2.03678206e-01 -6.07040524e-01 1.59359837e+00
-1.28533086e-02 -1.37957084e+00 2.42043823e-01 3.58066618e-01
2.25553617e-01 -5.64755678e-01 1.22333564e-01 -7.30992198e-01
-2.60599405e-01 -8.56735826e-01 5.40204704e-01 4.00350392e-01
1.26153481e+00 -4.82948214e-01 2.78206974e-01 -1.22262560e-01
-1.49703395e+00 -5.72621487e-02 -3.23371381e-01 1.28332734e-01
-1.66277915e-01 7.12947011e-01 -5.17746806e-01 6.47056580e-01
6.64261818e-01 2.82680899e-01 -6.13339424e-01 9.90405798e-01
3.51298243e-01 4.11270857e-02 -6.92307830e-01 -3.01276952e-01
-1.41351625e-01 -5.53023696e-01 8.10313523e-01 8.99219275e-01
7.99015403e-01 5.83626889e-02 -2.30407417e-01 6.22758329e-01
1.24442473e-01 2.93383658e-01 -1.07445145e+00 5.20006903e-02
5.51471293e-01 1.32938540e+00 -1.27965569e+00 1.43709779e-01
-5.44386804e-01 9.48830724e-01 -1.00594513e-01 2.44961634e-01
-4.98330206e-01 -8.11682940e-01 4.10280198e-01 6.16080463e-01
3.23136777e-01 -6.23063147e-01 -5.58180749e-01 -8.25654089e-01
3.25135961e-02 -7.23960102e-01 2.40861371e-01 -4.65775937e-01
-1.03028440e+00 3.33605081e-01 -2.10954368e-01 -1.61845171e+00
2.56846637e-01 -7.17940867e-01 -4.62408006e-01 5.74520528e-01
-1.34449077e+00 -1.18986011e+00 -2.99535096e-01 5.43190002e-01
-1.94398612e-01 -3.40477705e-01 8.83053541e-01 4.85606976e-02
-6.43779188e-02 8.42086017e-01 1.83391541e-01 -2.45562512e-02
6.80570364e-01 -7.73832738e-01 5.97410083e-01 7.36590326e-01
1.64557740e-01 5.14736116e-01 4.45566654e-01 -7.91269898e-01
-2.02821541e+00 -6.86152279e-01 6.61308050e-01 -4.71004128e-01
5.64961135e-01 -5.77187836e-01 -4.30745512e-01 -1.86664104e-01
-1.18218556e-01 -1.97125077e-01 4.85580683e-01 -5.22576869e-01
-4.14275587e-01 -5.75456262e-01 -1.52900290e+00 7.77525187e-01
1.40697670e+00 -4.63484615e-01 1.66948780e-01 1.88058347e-01
2.53546406e-02 -4.86550510e-01 -1.09607768e+00 4.07335669e-01
1.05051041e+00 -9.95283067e-01 9.35493171e-01 1.24661483e-01
-3.37507784e-01 -6.43882990e-01 2.49089077e-01 -3.00729960e-01
-5.73895931e-01 -1.59333479e+00 -8.23153649e-03 1.20780706e+00
8.10597241e-02 -9.31332469e-01 8.29578638e-01 4.62876767e-01
4.16621774e-01 -5.00029564e-01 -1.10227764e+00 -1.05238008e+00
-5.03415108e-01 1.67813618e-02 8.90463412e-01 7.86282539e-01
1.27647147e-01 1.33545399e-01 -1.58466294e-01 2.70228267e-01
1.12924993e+00 4.87702996e-01 9.58788216e-01 -1.43465090e+00
-6.03231527e-02 -5.84516227e-01 -9.70994830e-01 -8.19506645e-01
-3.32361311e-01 -6.73665524e-01 -6.92032218e-01 -1.09177661e+00
-1.95008248e-01 -8.88718963e-01 8.22511241e-02 2.37257779e-01
4.10437465e-01 5.96734107e-01 5.81342205e-02 2.68769413e-01
-1.24092817e-01 -5.98374829e-02 1.23519611e+00 -1.93076238e-01
-2.84327596e-01 2.55664010e-02 -5.77449262e-01 3.90658349e-01
7.84436285e-01 -3.51555675e-01 -2.36439005e-01 -7.44216740e-02
3.33201915e-01 -1.70687541e-01 4.48873937e-01 -1.20409739e+00
4.41646785e-01 1.60369322e-01 4.78275150e-01 -2.99689293e-01
5.47199436e-02 -1.14150178e+00 6.46054268e-01 7.12635040e-01
3.74306172e-01 3.91157120e-01 7.53739253e-02 6.06785536e-01
2.34045774e-01 -9.69442874e-02 7.80180275e-01 -5.88411279e-02
-4.14595485e-01 3.49275321e-01 -4.89679396e-01 -4.41485614e-01
1.11633980e+00 -1.27333486e+00 -5.68037629e-01 -3.65888178e-01
-1.89221501e-01 -2.98852950e-01 1.21741951e+00 2.21929774e-01
7.45656312e-01 -1.43719518e+00 -2.59557307e-01 6.31434083e-01
6.74005076e-02 -4.14805114e-01 -7.60275498e-02 7.46064544e-01
-8.40312839e-01 2.16717899e-01 -3.24141294e-01 -6.87194347e-01
-1.52896750e+00 6.77643418e-01 1.58873975e-01 -8.19895193e-02
-8.32273602e-01 1.75911561e-01 -2.36245766e-01 2.35791668e-01
4.83719483e-02 -3.74794751e-01 3.72399949e-02 -2.96877533e-01
1.87042862e-01 5.40011108e-01 1.53657541e-01 -5.00067174e-01
-4.78236198e-01 1.77306569e+00 1.06223270e-01 2.21325368e-01
8.06139290e-01 -2.19214946e-01 -1.60219163e-01 -8.38126615e-02
8.80048871e-01 4.85232055e-01 -7.53331542e-01 1.21297687e-01
-1.38385996e-01 -4.77500856e-01 -3.51069242e-01 -3.13515812e-01
-8.10210466e-01 5.47048151e-01 8.43529046e-01 3.80928576e-01
1.10377765e+00 -2.35178545e-01 9.46529150e-01 1.53791770e-01
7.40726352e-01 -1.12864423e+00 -7.53435418e-02 6.92475960e-02
3.84715497e-01 -6.01228476e-01 4.32322472e-01 -7.62144327e-01
1.24523558e-01 1.50189829e+00 2.16483995e-01 1.03808537e-01
6.15326226e-01 4.96288747e-01 1.23262592e-01 -2.95626193e-01
-7.51896724e-02 -1.02450863e-01 1.09064236e-01 9.88000453e-01
-1.67433172e-01 -2.74163663e-01 -4.55644488e-01 1.79048285e-01
-4.78594936e-02 -1.68273766e-02 4.80816126e-01 9.59914148e-01
-2.64796942e-01 -1.50558734e+00 -5.05747736e-01 5.22333145e-01
-4.16882217e-01 2.06313342e-01 -4.71641839e-01 8.48930478e-01
4.99221636e-03 7.28283763e-01 -1.33920699e-01 -6.63721383e-01
5.04039884e-01 -2.03439221e-01 9.37025964e-01 1.14462368e-01
-5.00789225e-01 1.69125143e-02 3.73395346e-02 -6.34359896e-01
-5.28567791e-01 -5.97502470e-01 -1.04131830e+00 -4.86457437e-01
-3.39281887e-01 -6.43882379e-02 6.88949943e-01 4.70292628e-01
7.90586114e-01 -3.09205562e-01 7.24535704e-01 -3.45491797e-01
-1.48642272e-01 -1.85191810e-01 -7.23776758e-01 3.63155484e-01
1.24153607e-01 -7.61193395e-01 -2.63246685e-01 -4.37257774e-02] | [8.511868476867676, -2.2341415882110596] |
a8e8f08f-adbe-416f-b45b-1b8769069f53 | semi-supervised-teacher-student-deep-neural | 2112.06142 | null | https://arxiv.org/abs/2112.06142v1 | https://arxiv.org/pdf/2112.06142v1.pdf | Semi-supervised teacher-student deep neural network for materials discovery | Data driven generative machine learning models have recently emerged as one of the most promising approaches for new materials discovery. While the generator models can generate millions of candidates, it is critical to train fast and accurate machine learning models to filter out stable, synthesizable materials with desired properties. However, such efforts to build supervised regression or classification screening models have been severely hindered by the lack of unstable or unsynthesizable samples, which usually are not collected and deposited in materials databases such as ICSD and Materials Project (MP). At the same time, there are a significant amount of unlabelled data available in these databases. Here we propose a semi-supervised deep neural network (TSDNN) model for high-performance formation energy and synthesizability prediction, which is achieved via its unique teacher-student dual network architecture and its effective exploitation of the large amount of unlabeled data. For formation energy based stability screening, our semi-supervised classifier achieves an absolute 10.3\% accuracy improvement compared to the baseline CGCNN regression model. For synthesizability prediction, our model significantly increases the baseline PU learning's true positive rate from 87.9\% to 97.9\% using 1/49 model parameters. To further prove the effectiveness of our models, we combined our TSDNN-energy and TSDNN-synthesizability models with our CubicGAN generator to discover novel stable cubic structures. Out of 1000 recommended candidate samples by our models, 512 of them have negative formation energies as validated by our DFT formation energy calculations. Our experimental results show that our semi-supervised deep neural networks can significantly improve the screening accuracy in large-scale generative materials design. | ['Jianjun Hu', 'Nihang Fu', 'Yong Zhao', 'Edirisuriya M. Dilanga Siriwardane', 'Daniel Gleaves'] | 2021-12-12 | null | null | null | null | ['formation-energy'] | ['miscellaneous'] | [ 4.31116730e-01 5.74693903e-02 -4.01967525e-01 1.87743045e-02
-1.27769065e+00 -2.78693199e-01 4.71234977e-01 1.97433736e-02
2.45221138e-01 1.20699596e+00 2.63115466e-02 -3.64398003e-01
6.54875860e-03 -1.35714436e+00 -9.66060281e-01 -1.19339120e+00
3.28850210e-01 8.65178227e-01 2.29011759e-01 -5.79014957e-01
2.46976122e-01 4.58224595e-01 -1.68269408e+00 3.05717319e-01
1.38515174e+00 1.18732953e+00 2.09167749e-01 3.03283811e-01
3.57524246e-01 3.84746879e-01 -3.36381018e-01 -1.14566103e-01
1.00375794e-01 -6.11427665e-01 -5.32563448e-01 -4.78957862e-01
6.56896383e-02 -2.01279268e-01 -1.50418416e-01 7.92462587e-01
7.31719077e-01 -1.34419262e-01 1.13430762e+00 -8.97691369e-01
-9.84261036e-01 8.52846742e-01 -2.49822631e-01 -3.78773898e-01
3.60772014e-02 2.78622746e-01 1.03799367e+00 -1.04308879e+00
7.16510415e-01 6.95987344e-01 2.86464095e-01 9.33578014e-01
-1.16215992e+00 -1.09965110e+00 -3.55465412e-01 1.41012529e-03
-1.08330107e+00 -4.84428912e-01 1.11143732e+00 -4.10825640e-01
1.21288776e+00 1.87786043e-01 7.95850575e-01 1.31651223e+00
2.82363385e-01 4.45352644e-01 1.02270365e+00 -4.75682408e-01
2.25573659e-01 1.39815614e-01 -3.42698783e-01 7.66138494e-01
4.78977680e-01 4.23728615e-01 -5.44699013e-01 4.99414541e-02
6.65660918e-01 -2.65078753e-01 -4.54666838e-02 -2.66090959e-01
-6.72170997e-01 8.50088656e-01 6.27832353e-01 1.60495386e-01
1.04687326e-01 1.09571725e-01 -9.79653001e-02 -9.31995362e-02
7.18519151e-01 9.74221885e-01 -4.31720972e-01 -6.09801374e-02
-8.62800479e-01 2.21563593e-01 4.00137633e-01 8.81278336e-01
7.02134728e-01 5.41580856e-01 5.99180609e-02 8.93238306e-01
2.96100318e-01 5.95070660e-01 2.03026175e-01 -4.30620134e-01
2.72764206e-01 8.30838382e-01 -5.69491647e-02 -3.91878486e-01
-1.04278125e-01 -5.55446744e-01 -8.77442479e-01 2.97410619e-02
-2.47546379e-02 -8.41220170e-02 -1.09712636e+00 1.68475270e+00
2.05882058e-01 -4.87705797e-01 5.86096123e-02 5.72096348e-01
9.44965005e-01 1.08950424e+00 -2.96149813e-02 -3.93389136e-01
6.27503335e-01 -7.73643970e-01 -9.77906883e-02 2.07591474e-01
5.53159535e-01 -6.05120838e-01 9.57695067e-01 5.54668009e-01
-1.22246850e+00 -5.24165750e-01 -1.44131422e+00 8.27709883e-02
-4.03377473e-01 1.43465191e-01 9.32716310e-01 8.51295590e-01
-7.54262328e-01 8.85930359e-01 -8.00551891e-01 1.05465047e-01
6.21586382e-01 8.67128372e-01 9.60852727e-02 -4.86560911e-02
-1.14051712e+00 4.82862473e-01 4.24009025e-01 -7.95304589e-03
-1.37848663e+00 -7.27281868e-01 -4.12548006e-01 -1.22974239e-01
3.09328109e-01 -4.27079976e-01 8.99329901e-01 -5.02681911e-01
-1.58926558e+00 4.95203316e-01 9.36663449e-02 -2.12971881e-01
2.87377357e-01 1.05580360e-01 -6.41640186e-01 -1.08256958e-01
6.15098663e-02 8.59359264e-01 6.87693655e-01 -1.28871000e+00
-1.94296569e-01 -3.81011376e-03 -4.35127407e-01 -2.05769628e-01
-6.34809077e-01 -3.43825161e-01 1.68899059e-01 -6.52024150e-01
1.79346085e-01 -9.55260873e-01 -2.66733557e-01 -4.76909846e-01
-5.95766306e-01 -4.25018758e-01 6.58904314e-01 -3.74163955e-01
1.08725190e+00 -1.52699375e+00 1.39340803e-01 4.58184063e-01
4.07813430e-01 1.50133535e-01 2.46504378e-02 5.56774080e-01
-2.72800058e-01 3.81509095e-01 -1.87553942e-01 3.34966123e-01
-2.64763653e-01 -2.63297081e-01 -1.30399778e-01 8.57102424e-02
4.30986673e-01 1.07466304e+00 -7.93734133e-01 -5.30645996e-02
-1.94933098e-02 3.70369852e-01 -7.58722067e-01 1.65929452e-01
-7.96350598e-01 5.13217807e-01 -4.53666270e-01 1.08289444e+00
4.56687033e-01 -4.74140316e-01 1.24648280e-01 -2.24936947e-01
-1.18736356e-01 6.36859179e-01 -5.74094951e-01 1.21625841e+00
-2.45509133e-01 2.45531932e-01 -5.45183599e-01 -9.26430583e-01
1.28123176e+00 1.78862348e-01 7.73337543e-01 -1.02465320e+00
4.61169630e-02 6.98104382e-01 3.12951475e-01 -1.64500535e-01
4.37622696e-01 -5.72965920e-01 -1.08589917e-01 3.81102651e-01
1.11482508e-01 -3.18096906e-01 2.45169550e-01 1.58637375e-01
9.05242205e-01 1.69645637e-01 -4.97916132e-01 -4.59287345e-01
3.38521332e-01 1.03886731e-01 4.81919199e-01 4.49919313e-01
4.10175741e-01 5.74466467e-01 1.90388784e-01 -4.35047090e-01
-1.53652573e+00 -1.03746963e+00 -2.42264703e-01 8.83281291e-01
-1.11522019e-01 -5.47324181e-01 -7.06279397e-01 -4.05302048e-01
-3.48626703e-01 5.90380073e-01 -2.90008128e-01 -7.45346367e-01
-4.95361745e-01 -1.36347556e+00 1.96641684e-01 6.41337156e-01
3.65431666e-01 -9.67851818e-01 3.13122392e-01 4.38035607e-01
5.33960685e-02 -4.81715143e-01 -5.48267141e-02 5.35362303e-01
-7.13742793e-01 -1.03796160e+00 -4.48036611e-01 -9.23048019e-01
7.71823883e-01 -1.91986352e-01 8.86145234e-01 -1.48975272e-02
-3.64063323e-01 -2.03855351e-01 -2.11284474e-01 -3.38772804e-01
-1.20946288e+00 4.63364094e-01 4.10089523e-01 -4.46380287e-01
-4.87452857e-02 -7.32574582e-01 -7.61678040e-01 3.14079612e-01
-4.85350490e-01 4.14760262e-01 7.56345689e-01 9.29055095e-01
9.45535958e-01 3.50909531e-01 1.09284401e+00 -8.49123359e-01
4.19866979e-01 -3.63864273e-01 -6.87232673e-01 3.99662167e-01
-1.34649086e+00 3.76036972e-01 7.43283033e-01 -4.89078850e-01
-1.15649045e+00 -2.31799646e-03 -5.92363238e-01 -5.51210381e-02
1.03985101e-01 3.13444942e-01 -4.42173481e-01 -3.22078764e-01
9.59066331e-01 2.66818941e-01 -1.58055961e-01 -1.95963249e-01
9.91019756e-02 5.47237396e-01 1.34419888e-01 -1.19492555e+00
9.52910721e-01 -2.13635370e-01 2.43975624e-01 -7.45462716e-01
-7.24945605e-01 1.62290782e-01 -2.38573983e-01 -2.24631339e-01
7.30329812e-01 -1.01266980e+00 -7.02437222e-01 5.30625284e-01
-6.17226839e-01 -4.09010589e-01 -2.78725833e-01 1.47183523e-01
-2.94152379e-01 -1.05335666e-02 -6.70100808e-01 -8.37508142e-01
-9.00060177e-01 -1.47021496e+00 9.20415223e-01 2.32663769e-02
-1.15726762e-01 -5.73345304e-01 -1.56636834e-01 6.07261181e-01
6.32746160e-01 5.48788190e-01 1.46977675e+00 -5.48159242e-01
-9.84439015e-01 -2.37978473e-01 7.44950026e-02 3.75420690e-01
4.40191925e-01 1.53928772e-01 -7.71225154e-01 -2.33306915e-01
-1.29171580e-01 -6.74190640e-01 1.08297050e+00 3.23126197e-01
1.38624585e+00 -2.59926379e-01 -3.90779257e-01 4.10959452e-01
1.37606120e+00 6.46962643e-01 6.99102759e-01 -3.07766926e-02
1.01597166e+00 1.77159816e-01 1.74688742e-01 1.32709980e-01
-5.64834625e-02 3.22043061e-01 4.79378015e-01 -4.50074747e-02
-1.49882063e-01 -6.61253512e-01 4.69940543e-01 1.13902378e+00
-5.53053975e-01 -4.54722196e-01 -9.01409984e-01 1.49361104e-01
-1.30585217e+00 -8.33275795e-01 -4.40730929e-01 2.22831774e+00
1.09671700e+00 5.23054302e-01 1.32549286e-01 1.34410933e-01
4.23900038e-01 -2.23692268e-01 -9.94330943e-01 -7.94068128e-02
-3.51955116e-01 7.72287250e-01 4.13170934e-01 1.02534615e-01
-7.97026336e-01 8.84834886e-01 5.89973211e+00 1.32328629e+00
-1.35021842e+00 -5.50281366e-05 1.21503735e+00 -1.75345883e-01
-7.31755435e-01 -8.07640627e-02 -1.11461282e+00 4.95249808e-01
1.11817169e+00 6.82465062e-02 3.10920179e-01 8.59843314e-01
-1.25448525e-01 1.07372724e-01 -1.21795750e+00 7.76765704e-01
-1.25454441e-01 -2.06054163e+00 1.52236417e-01 3.23257774e-01
1.21905863e+00 -6.21125922e-02 2.64277339e-01 2.55464226e-01
3.57288331e-01 -1.19750023e+00 7.03405738e-01 2.26404682e-01
1.34174979e+00 -9.34533656e-01 1.20582312e-01 8.63784030e-02
-9.96673942e-01 6.17227145e-02 -2.72210807e-01 1.93938598e-01
-1.78811118e-01 8.27968419e-01 -1.03649557e+00 4.11130637e-01
4.59897995e-01 5.68696976e-01 -4.19854939e-01 5.14062941e-01
-1.76559120e-01 9.30387437e-01 -4.08428937e-01 -5.69988489e-01
-5.80811575e-02 -4.12466645e-01 1.24790460e-01 4.98063713e-01
6.17181718e-01 -5.77838020e-03 -3.49143036e-02 1.38210034e+00
-5.23986518e-01 -1.73142016e-01 -3.84906650e-01 -3.80929261e-01
3.17503214e-01 1.08195853e+00 -8.69785309e-01 -2.27345169e-01
1.57325044e-01 4.19084460e-01 8.92506465e-02 9.61479545e-02
-8.87731850e-01 2.26565360e-04 1.55612275e-01 6.27773702e-01
7.81362951e-02 -2.11297661e-01 -3.07187200e-01 -9.61005390e-01
-1.58900231e-01 -7.23824024e-01 -4.31941003e-02 -6.75746083e-01
-1.42227900e+00 4.94215846e-01 -1.17704265e-01 -1.05010593e+00
3.40705104e-02 -7.52483487e-01 -5.76688051e-01 6.81881309e-01
-1.00062954e+00 -1.31350458e+00 -3.50804627e-02 7.14421719e-02
4.38160598e-01 -3.70842069e-01 8.11109543e-01 1.85176134e-01
-7.49281168e-01 6.26639664e-01 3.82989705e-01 -2.85814196e-01
5.78112662e-01 -1.09951568e+00 3.36958557e-01 5.91331005e-01
-1.17748618e-01 4.40339386e-01 6.04815423e-01 -1.01718032e+00
-1.62232995e+00 -1.19177449e+00 3.97959411e-01 -5.16111553e-01
5.55170119e-01 -8.55862617e-01 -7.20639944e-01 5.85433245e-02
-1.13338649e-01 -4.98531729e-01 8.26199710e-01 -6.49852157e-02
-3.20688672e-02 -2.76243180e-01 -9.66762185e-01 5.77652633e-01
1.27136803e+00 -2.21211433e-01 2.20178232e-01 9.34568703e-01
9.24075365e-01 -3.79548550e-01 -1.12526667e+00 7.87703812e-01
5.13037682e-01 -7.25736916e-01 1.03682780e+00 -3.93270373e-01
1.13767576e+00 5.84610254e-02 -1.77747428e-01 -9.92756426e-01
-2.42742881e-01 -5.27399719e-01 -2.18289480e-01 1.31025493e+00
1.10441959e+00 -4.04540509e-01 1.32783377e+00 8.24720025e-01
-4.67575788e-01 -1.33071160e+00 -7.09369361e-01 -7.79638350e-01
5.02060413e-01 -3.20497811e-01 7.81816721e-01 6.25182331e-01
-3.31868231e-01 5.85067391e-01 -3.83942604e-01 -2.95036018e-01
5.48802972e-01 2.93573558e-01 2.91223437e-01 -1.40075934e+00
-3.21124196e-01 -3.72802317e-01 2.15351894e-01 -6.90106332e-01
2.46817973e-02 -1.18282723e+00 -5.59541881e-02 -1.55348742e+00
1.97678313e-01 -1.15430999e+00 -2.24344835e-01 3.82480055e-01
2.47642606e-01 1.45539820e-01 -5.15729308e-01 1.69817567e-01
-2.85225064e-01 1.04398572e+00 1.33624768e+00 -4.36171055e-01
-3.64452124e-01 8.31894577e-02 -1.04292464e+00 1.29276320e-01
9.42832351e-01 -3.54005247e-01 -5.00526547e-01 -9.05205682e-03
7.10958958e-01 -3.06715406e-02 1.00006394e-01 -1.18731987e+00
-1.67766973e-01 -2.66738534e-01 8.04596066e-01 -5.98082840e-01
3.63909751e-01 -3.47870022e-01 5.11336207e-01 6.06794477e-01
-1.23327740e-01 -5.15927374e-01 -2.46080533e-02 3.34876716e-01
2.36349732e-01 -5.11918776e-02 6.31878555e-01 -1.87766999e-01
-1.41759425e-01 6.77683592e-01 -2.56220043e-01 -2.28370667e-01
8.12048078e-01 -2.39585042e-01 -4.46590990e-01 -5.30031770e-02
-4.26088959e-01 -1.33639157e-01 5.69575965e-01 1.78915486e-01
7.60740757e-01 -1.35483074e+00 -3.39081466e-01 4.19011682e-01
6.20533414e-02 4.06349331e-01 1.53346762e-01 4.33323503e-01
-5.50900817e-01 2.89283842e-01 -1.18338436e-01 -4.52464074e-01
-6.44767284e-01 5.77707827e-01 2.66967773e-01 -2.06688046e-01
-7.17756599e-02 1.04549849e+00 4.10854965e-02 -3.43393654e-01
-3.49687934e-01 -3.28567356e-01 1.63667828e-01 -1.88204676e-01
-1.62492648e-01 3.84071946e-01 4.69037741e-01 -2.68104225e-01
1.52295148e-02 3.60515147e-01 -2.78057963e-01 3.11769128e-01
1.87110746e+00 7.07109869e-01 -1.65754572e-01 8.68041366e-02
1.11334014e+00 2.89469101e-02 -1.22069395e+00 2.48641193e-01
-2.88424432e-01 3.33740830e-01 -1.64350703e-01 -8.97738099e-01
-1.02360940e+00 7.52020717e-01 6.96473777e-01 -3.68484333e-02
1.06497264e+00 2.10194603e-01 1.08083618e+00 4.67748165e-01
3.93762171e-01 -1.24703288e+00 4.45569187e-01 1.01501755e-01
8.57578099e-01 -1.38902891e+00 3.51606049e-02 -6.29687786e-01
-5.82019351e-02 1.09191167e+00 9.96011615e-01 9.06481445e-02
5.05216718e-01 2.62051463e-01 -7.13338435e-01 -4.26394135e-01
-7.92359769e-01 2.65359014e-01 2.30855912e-01 5.36551058e-01
4.16763842e-01 2.68207192e-01 -7.23088831e-02 6.98455334e-01
-5.16780257e-01 -3.20126772e-01 3.05727720e-01 8.04836810e-01
-6.37838304e-01 -1.60542285e+00 7.77329085e-03 9.32411849e-01
-1.22128755e-01 -2.81322539e-01 -4.53012466e-01 4.03841734e-01
1.41411304e-01 7.21329391e-01 -2.80439168e-01 -6.25523508e-01
-7.16134487e-03 7.60046020e-02 7.28208363e-01 -5.97460568e-01
-4.05382723e-01 2.75810868e-01 2.13745236e-01 5.22346199e-02
-2.52918929e-01 -3.35141301e-01 -1.18642819e+00 -3.52179587e-01
-7.57417023e-01 2.40573451e-01 5.23856103e-01 7.42032111e-01
2.64134824e-01 6.12837851e-01 8.20178688e-01 -8.90331089e-01
-1.97396293e-01 -7.01030195e-01 -3.83497834e-01 9.70829427e-02
-2.39882559e-01 -7.90409267e-01 3.35564315e-02 -1.53008565e-01] | [5.189000606536865, 5.372509956359863] |
eb9a12a1-e234-4ff3-b351-5901836558d9 | multi-grained-knowledge-retrieval-for-end-to | 2305.10149 | null | https://arxiv.org/abs/2305.10149v1 | https://arxiv.org/pdf/2305.10149v1.pdf | Multi-Grained Knowledge Retrieval for End-to-End Task-Oriented Dialog | Retrieving proper domain knowledge from an external database lies at the heart of end-to-end task-oriented dialog systems to generate informative responses. Most existing systems blend knowledge retrieval with response generation and optimize them with direct supervision from reference responses, leading to suboptimal retrieval performance when the knowledge base becomes large-scale. To address this, we propose to decouple knowledge retrieval from response generation and introduce a multi-grained knowledge retriever (MAKER) that includes an entity selector to search for relevant entities and an attribute selector to filter out irrelevant attributes. To train the retriever, we propose a novel distillation objective that derives supervision signals from the response generator. Experiments conducted on three standard benchmarks with both small and large-scale knowledge bases demonstrate that our retriever performs knowledge retrieval more effectively than existing methods. Our code has been made publicly available.\footnote{https://github.com/18907305772/MAKER} | ['Wei Bi', 'Xiaojun Quan', 'Ke Yang', 'Weizhou Shen', 'Fanqi Wan'] | 2023-05-17 | null | null | null | null | ['response-generation'] | ['natural-language-processing'] | [ 6.61701709e-02 2.21516535e-01 -4.01279807e-01 -6.24917805e-01
-1.48105145e+00 -8.04631829e-01 5.84962249e-01 -8.91912207e-02
-5.84992707e-01 1.05921435e+00 5.32586217e-01 -3.80450711e-02
-6.87080547e-02 -7.87970960e-01 -5.64111054e-01 -2.00198457e-01
6.64793968e-01 1.08087373e+00 1.80728927e-01 -4.99376446e-01
8.21499154e-03 -4.52788062e-02 -1.16619086e+00 5.25699675e-01
9.30485189e-01 8.93031895e-01 6.82184622e-02 5.33740461e-01
-5.43666072e-02 7.40323544e-01 -7.61766732e-01 -7.02779531e-01
2.18652710e-01 -3.38529319e-01 -1.23275244e+00 -4.20891345e-01
1.02973357e-01 -4.50328439e-01 -4.24965024e-01 6.72802925e-01
9.40283775e-01 5.05564988e-01 5.24120092e-01 -8.99044633e-01
-1.00185633e+00 9.28266048e-01 7.49961147e-03 1.02066532e-01
6.51767075e-01 2.54573345e-01 1.24689496e+00 -1.10321653e+00
7.24616826e-01 1.27693892e+00 1.09410122e-01 9.15224314e-01
-1.43160176e+00 -6.41587079e-01 1.64531738e-01 1.79811120e-01
-1.61615813e+00 -6.87060654e-01 6.15958869e-01 -7.66981244e-02
1.03436780e+00 4.12841320e-01 -8.18004832e-02 1.25853813e+00
-6.48398757e-01 1.19351280e+00 7.32358396e-01 -2.60918111e-01
7.21846297e-02 5.67143142e-01 3.56139600e-01 3.03651780e-01
-5.32910712e-02 -3.92049775e-02 -7.74467051e-01 -3.92501593e-01
4.03361529e-01 -2.42264599e-01 -5.07782161e-01 -1.14201149e-02
-1.03125203e+00 8.66987586e-01 4.10659462e-01 -1.64234787e-01
-6.18566513e-01 -1.97051130e-02 7.93328658e-02 6.50336802e-01
3.14863175e-01 7.72918463e-01 -5.31182885e-01 -1.28969267e-01
-4.03256714e-01 6.30346119e-01 1.25093997e+00 1.21410704e+00
9.33029354e-01 -3.48426282e-01 -8.81432652e-01 1.17333043e+00
1.33858413e-01 7.52785265e-01 4.60640043e-01 -9.78343010e-01
7.08097398e-01 7.72495568e-01 4.91088659e-01 -5.60829461e-01
-3.02258227e-03 -7.28382990e-02 -3.60284239e-01 -6.11135960e-01
2.92586625e-01 -5.21710098e-01 -6.41617596e-01 1.78682745e+00
4.73408520e-01 -1.51284292e-01 3.88116241e-01 1.24549127e+00
1.21008873e+00 5.42052150e-01 3.26162815e-01 2.37510487e-01
1.43662453e+00 -9.80040550e-01 -5.43110967e-01 -3.78705889e-01
4.45302278e-01 -6.31523907e-01 1.36193371e+00 4.67148498e-02
-1.08334160e+00 -2.75968581e-01 -6.11183465e-01 -2.96809405e-01
-2.92471051e-01 2.92661816e-01 4.36237544e-01 1.34276778e-01
-9.05856907e-01 -1.63036380e-02 -3.13226908e-01 -9.43923742e-02
-3.96371484e-02 3.82863462e-01 -2.38977984e-01 -3.40612642e-02
-1.86430621e+00 8.63925397e-01 3.57190430e-01 -6.48080185e-02
-7.92177320e-01 -5.80842972e-01 -6.03091657e-01 1.50847390e-01
7.53308415e-01 -1.07578588e+00 1.81745327e+00 -7.85062551e-01
-1.76734293e+00 6.35307968e-01 -1.63289279e-01 -2.96157897e-01
2.80235648e-01 -5.75467110e-01 -1.62634000e-01 1.34440765e-01
3.54883708e-02 6.39724493e-01 5.98430753e-01 -1.14706063e+00
-4.93143767e-01 -8.78342092e-02 2.90430158e-01 5.38547873e-01
-2.80546606e-01 7.78690949e-02 -8.25489640e-01 -6.25623345e-01
-5.05101144e-01 -9.04575944e-01 2.39162538e-02 -7.40904212e-01
-6.46958470e-01 -6.47114873e-01 2.37858012e-01 -4.81930733e-01
1.46140945e+00 -1.72724652e+00 2.53389776e-01 1.60984695e-01
-1.95890423e-02 3.98240596e-01 -3.72317910e-01 6.22634828e-01
2.00912833e-01 -1.08776823e-01 1.22252062e-01 -9.85646322e-02
2.64563739e-01 -3.02375276e-02 -6.38555646e-01 -4.20566678e-01
3.33385497e-01 1.29133010e+00 -8.86041522e-01 -2.98460096e-01
-3.51109684e-01 3.84566069e-01 -5.52120566e-01 5.08109510e-01
-6.87592268e-01 2.03229800e-01 -1.21606553e+00 7.05348372e-01
2.25683391e-01 -6.24898314e-01 1.77532092e-01 -7.31564686e-02
5.51061511e-01 6.91534162e-01 -1.07632816e+00 1.67207968e+00
-5.15583634e-01 1.15992695e-01 2.00558066e-01 -6.08631790e-01
8.55839312e-01 4.65840906e-01 1.62517935e-01 -8.61923516e-01
4.98388633e-02 9.23485681e-02 -3.96445483e-01 -4.95258570e-01
8.61366987e-01 1.09001705e-02 -4.92559999e-01 7.39790618e-01
6.52319491e-02 -1.18483923e-01 1.32893682e-01 6.85511410e-01
1.40540004e+00 -7.96335563e-02 3.13319266e-02 2.30990335e-01
5.92725158e-01 1.34760126e-01 4.89466310e-01 8.81488740e-01
2.63521850e-01 3.28149199e-01 2.02421114e-01 -6.85632378e-02
-3.81600261e-01 -1.08632576e+00 3.15413237e-01 1.66067350e+00
1.17001697e-01 -3.93105090e-01 -5.48243046e-01 -8.35145354e-01
4.41866964e-01 8.63756955e-01 -3.06606740e-01 -4.96748865e-01
-5.03620505e-01 -4.57245827e-01 7.00309038e-01 4.97294933e-01
2.86454648e-01 -1.37867951e+00 -1.66156471e-01 2.92867512e-01
-7.35134006e-01 -1.13351917e+00 -6.61427736e-01 -2.94659007e-03
-3.01238149e-01 -8.65601540e-01 -6.43028677e-01 -4.71834004e-01
5.37514091e-01 1.20745257e-01 1.65777969e+00 -5.34529286e-03
1.28896825e-03 6.69094145e-01 -5.18879235e-01 -4.13302630e-01
-3.57560068e-01 5.31144857e-01 -6.94260076e-02 -4.96168174e-02
9.21694398e-01 -2.10001856e-01 -7.44518399e-01 5.72884500e-01
-9.63926971e-01 -5.41143529e-02 7.55430400e-01 8.15009594e-01
5.82092881e-01 -5.87099671e-01 1.17975438e+00 -1.09706008e+00
1.25353003e+00 -6.56442583e-01 -5.73239207e-01 6.42356336e-01
-7.29047060e-01 3.50851029e-01 4.69619483e-01 -3.94501925e-01
-1.39476228e+00 1.25743985e-01 2.84846649e-02 -2.77314961e-01
-6.17281497e-02 5.54018497e-01 -1.75190359e-01 4.03152943e-01
9.68380868e-01 2.97132283e-01 -2.26633027e-01 -6.10321283e-01
7.42894650e-01 9.64211822e-01 4.71322596e-01 -1.07596350e+00
6.82997882e-01 2.40317006e-02 -8.05240095e-01 -2.66159147e-01
-1.04574049e+00 -6.47455454e-01 -1.37085900e-01 9.45970491e-02
4.95242745e-01 -1.21152651e+00 -8.40273023e-01 9.35617238e-02
-1.06870937e+00 -5.80276430e-01 -3.97151291e-01 3.97465110e-01
-5.32270253e-01 -2.64346004e-01 -6.00498796e-01 -5.46804667e-01
-8.99504602e-01 -8.84584248e-01 1.24125707e+00 4.05853033e-01
-4.38528538e-01 -6.67528868e-01 2.33763158e-01 8.45681250e-01
6.31685734e-01 -4.54061925e-01 8.18459749e-01 -1.18472970e+00
-9.03786421e-01 -3.44493359e-01 -9.28135142e-02 1.86269626e-01
-2.99731307e-02 -4.50509459e-01 -9.46415722e-01 2.28637643e-03
-3.44408095e-01 -9.63074386e-01 1.03963041e+00 -2.31298745e-01
8.04345071e-01 -5.38726091e-01 -3.10751736e-01 1.94658756e-01
1.00513172e+00 8.28940794e-02 3.05816919e-01 1.26298308e-01
4.51268852e-01 7.32154131e-01 7.78622568e-01 4.42630410e-01
7.73544788e-01 9.76045609e-01 -1.91399530e-01 2.54318893e-01
-7.97442570e-02 -3.93785328e-01 4.07143533e-01 5.41760743e-01
2.98387825e-01 -4.46826458e-01 -8.17380428e-01 6.28031492e-01
-1.99244547e+00 -7.95881629e-01 4.80632275e-01 2.21055532e+00
1.78359234e+00 -2.97788709e-01 1.35846615e-01 -6.58953846e-01
4.56945270e-01 -1.04207247e-01 -8.41091275e-01 -1.50793996e-02
-5.63929230e-03 3.41408253e-01 1.85000956e-01 7.83946872e-01
-8.58196080e-01 1.38648033e+00 5.11201191e+00 8.02019715e-01
-9.41261470e-01 -3.90597582e-02 4.49428350e-01 -4.68146801e-01
-5.84873259e-01 2.56573018e-02 -1.29399133e+00 2.76301742e-01
9.18692827e-01 -6.44385397e-01 4.91455466e-01 7.94737756e-01
-8.55832174e-02 7.78965876e-02 -1.19081795e+00 6.47675574e-01
-7.12634325e-02 -1.08976293e+00 2.55921066e-01 -4.44041759e-01
4.40975040e-01 8.15379620e-02 -5.39404191e-02 8.17246437e-01
9.33812797e-01 -9.72532213e-01 2.15854079e-01 8.43306720e-01
5.46545625e-01 -4.27924126e-01 5.13443291e-01 3.29599112e-01
-7.99352169e-01 1.21799313e-01 -1.80199325e-01 3.99422884e-01
1.99851111e-01 3.61507386e-01 -1.38615656e+00 5.85686386e-01
3.09018344e-01 2.35755786e-01 -5.17171502e-01 6.66485786e-01
-6.41728997e-01 4.58982676e-01 -4.33227688e-01 -2.19759032e-01
-5.79009168e-02 1.11362830e-01 5.78096688e-01 1.20789468e+00
-1.31663278e-01 4.12264466e-01 1.67246521e-01 1.03743422e+00
-5.99145770e-01 1.63739249e-01 -4.49908316e-01 -1.27777398e-01
1.04033673e+00 1.34165907e+00 -3.82093415e-02 -5.44859171e-01
-3.37061852e-01 8.77955854e-01 4.78825152e-01 8.07729125e-01
-5.79430640e-01 -5.94628036e-01 6.31283820e-01 -4.66166213e-02
2.03109860e-01 2.54043669e-01 1.64215162e-01 -1.30354953e+00
3.14864784e-01 -1.13905418e+00 7.92254567e-01 -6.65609956e-01
-1.51372063e+00 5.68734407e-01 1.62986279e-01 -7.82906592e-01
-8.33427370e-01 -1.96145639e-01 -2.43287861e-01 1.14257073e+00
-1.64675689e+00 -8.13273966e-01 -3.25024039e-01 8.92911911e-01
5.25664985e-01 -1.31917909e-01 1.04654145e+00 3.34753722e-01
-4.08594608e-01 9.01109338e-01 -1.63715780e-01 4.11105990e-01
1.25570655e+00 -1.24414492e+00 2.87869334e-01 2.96823204e-01
2.88198311e-02 1.04862130e+00 4.47287500e-01 -7.00441003e-01
-1.61673951e+00 -1.02945352e+00 9.36198115e-01 -9.54047263e-01
4.47825372e-01 -3.45845968e-01 -1.10842979e+00 6.77915573e-01
2.67223626e-01 -2.79131740e-01 8.55447114e-01 2.74518788e-01
-5.27333856e-01 -2.36602798e-01 -7.55643845e-01 5.15679955e-01
6.83542132e-01 -6.50380552e-01 -6.59956336e-01 4.77784574e-01
8.94194841e-01 -5.27597725e-01 -8.81808102e-01 4.11045521e-01
4.56334144e-01 -2.87165999e-01 1.11409044e+00 -7.50404418e-01
1.85997441e-01 -2.77630597e-01 4.30210903e-02 -1.26403129e+00
-1.18176773e-01 -7.95817912e-01 -3.57168913e-01 1.26041687e+00
9.29188371e-01 -6.09229028e-01 7.81196415e-01 1.32165742e+00
2.08741575e-01 -6.26975119e-01 -6.00504041e-01 -6.50443077e-01
-2.32153386e-02 1.01668656e-01 6.49528921e-01 7.49453366e-01
1.37421802e-01 1.12357366e+00 -2.92578191e-01 -8.51972327e-02
1.61767289e-01 5.02543569e-01 1.12614775e+00 -9.62054074e-01
-5.59982538e-01 -3.45314026e-01 3.90285134e-01 -1.51122761e+00
3.47916514e-01 -9.64928091e-01 1.53841123e-01 -1.45836580e+00
2.23781541e-01 -5.83290935e-01 -3.15860182e-01 8.81696403e-01
-7.25148797e-01 -2.77718663e-01 -3.54108997e-02 2.81458169e-01
-1.05646896e+00 7.93131411e-01 1.04621160e+00 -1.23692170e-01
-4.32158560e-01 2.40180641e-01 -1.23530030e+00 2.44368136e-01
7.47234464e-01 -5.17357945e-01 -7.25339115e-01 -6.95183098e-01
3.93954694e-01 4.12036628e-01 2.35609248e-01 -3.70992005e-01
6.91550672e-01 -2.44302616e-01 3.26845273e-02 -2.51886189e-01
5.37049234e-01 -5.40786088e-01 -3.16578932e-02 -3.07134360e-01
-9.85289454e-01 8.49809349e-02 9.85645726e-02 6.06089652e-01
-4.33483571e-01 -1.40791968e-01 2.65089571e-01 -6.90475851e-02
-6.57490671e-01 2.29714304e-01 -3.37519646e-02 5.52900732e-01
7.13556349e-01 4.04343039e-01 -6.80887461e-01 -7.71305263e-01
-6.01607561e-01 7.73143828e-01 1.59102589e-01 7.82897234e-01
6.65547073e-01 -1.36251616e+00 -8.33131790e-01 -7.89446011e-02
4.22943622e-01 1.03783399e-01 1.28094211e-01 5.68179965e-01
1.71809003e-01 7.97290921e-01 2.55961537e-01 -4.44800816e-02
-1.17348790e+00 2.76871920e-01 2.02904969e-01 -4.25031066e-01
-2.18543351e-01 1.20642269e+00 4.20284681e-02 -7.81062961e-01
4.70814407e-01 1.75695658e-01 -2.20526621e-01 2.16602273e-02
5.97445548e-01 9.61833745e-02 7.02607483e-02 -2.85250425e-01
-3.21179241e-01 -1.03445604e-01 -5.19675910e-01 -3.54140639e-01
9.82503653e-01 -8.93201306e-02 9.68037993e-02 3.70452292e-02
9.27305818e-01 2.67521553e-02 -8.85440648e-01 -1.07967889e+00
2.59189993e-01 -3.63373280e-01 -2.74901390e-01 -1.33492041e+00
-8.93448591e-01 4.81716484e-01 3.29774199e-03 -2.46995986e-02
1.20404446e+00 1.32926136e-01 9.83611882e-01 1.11098623e+00
2.59619683e-01 -1.26334155e+00 2.55470753e-01 6.82790756e-01
1.03054774e+00 -1.36673260e+00 -2.48741657e-01 -2.18293056e-01
-1.17718303e+00 6.79977477e-01 9.95396435e-01 2.38906860e-01
3.32077622e-01 -9.86143388e-03 2.88155556e-01 -3.13489020e-01
-1.40860295e+00 -3.39251250e-01 4.70484793e-01 2.27722362e-01
4.84398842e-01 1.10098317e-01 -1.66564584e-01 1.07334495e+00
-1.89376831e-01 2.19452441e-01 7.72689655e-02 8.98941576e-01
-2.95910239e-01 -1.50225174e+00 2.92704180e-02 3.96604210e-01
-5.83485782e-01 -4.16991234e-01 -9.39080000e-01 4.23005164e-01
-7.03415811e-01 1.13977110e+00 -3.49509716e-01 -5.09447813e-01
6.85691059e-01 5.61233461e-01 1.68054588e-02 -8.95304561e-01
-9.18000340e-01 -2.07811669e-01 5.25101721e-01 -5.92581630e-01
-1.45580247e-01 -2.97384441e-01 -1.19592440e+00 -3.19684371e-02
-3.34633589e-01 6.62733912e-01 2.00667799e-01 5.47262907e-01
1.00872421e+00 2.02083230e-01 5.60111940e-01 -4.35056025e-03
-1.05268693e+00 -1.04503417e+00 -1.69272035e-01 6.93157494e-01
1.10658802e-01 -4.79967952e-01 -2.19917783e-04 -1.68642458e-02] | [11.849284172058105, 8.032310485839844] |
a411622d-1f50-4a70-88ef-a972b6741d56 | motion-planning-for-parabolic-equations-using | 2305.12404 | null | https://arxiv.org/abs/2305.12404v1 | https://arxiv.org/pdf/2305.12404v1.pdf | Motion planning for parabolic equations using flatness and finite-difference approximations | We consider the problem of finding an input signal which transfers a linear boundary controlled 1D parabolic partial differential equation, with spatially-varying coefficients and a non-local term, from a given initial state to a desired final state. The initial and final states have certain smoothness and the transfer must occur over a given time interval. We address this motion planning problem by first discretizing the spatial derivatives in the parabolic equation using the finite-difference approximation to obtain a linear ODE in time. Then using the flatness approach we construct an input signal that transfers this ODE between states determined by the initial and final states of the parabolic equation. We prove that, as the discretization step size converges to zero, this input signal converges to a limiting input signal which can perform the desired transfer for the parabolic equation. While earlier works have applied this motion planning approach to constant coefficient parabolic equations, this is the first work to investigate and establish the efficacy of this approach for parabolic equations with discontinuous spatially-varying coefficients. Using this approach we can construct input signals which transfer the parabolic equation from one steady-state to another. We also show that this approach yields a new proof for the null controllability of 1D linear parabolic equations containing discontinuous coefficients and a non-local integral term. | ['Vivek Natarajan', 'Soham Chatterjee'] | 2023-05-21 | null | null | null | null | ['motion-planning'] | ['robots'] | [ 2.33146235e-01 3.28965247e-01 1.58213466e-01 4.81967330e-01
-4.57007170e-01 -4.62763101e-01 3.33041251e-01 -4.09756042e-03
-3.38376343e-01 9.62670922e-01 -3.26524168e-01 -3.21785539e-01
-1.81026652e-01 -7.87964344e-01 -8.18645656e-01 -1.00879025e+00
-3.02188396e-01 3.64822567e-01 6.20097995e-01 -4.31002229e-01
2.83690244e-01 8.12133074e-01 -9.14333642e-01 -4.54297364e-02
9.84021902e-01 8.23708355e-01 -2.52802938e-01 1.03344190e+00
1.62100419e-01 4.58338439e-01 -2.45640337e-01 6.05493963e-01
5.33543646e-01 -7.05765367e-01 -9.81459975e-01 3.52568984e-01
1.75782945e-02 -6.76111504e-02 2.36549657e-02 9.01603043e-01
2.98651218e-01 6.70961320e-01 1.06248569e+00 -1.21741748e+00
-4.85630810e-01 -2.22925216e-01 -5.77009082e-01 -5.02925441e-02
1.57770202e-01 1.40510887e-01 2.76692897e-01 -7.21753001e-01
7.86198080e-01 8.82888615e-01 1.34854257e+00 6.67686164e-01
-1.55026460e+00 1.07586063e-01 -3.34076732e-01 -6.50712550e-01
-1.27700627e+00 -2.38472745e-02 4.29425627e-01 -7.62762666e-01
6.21161938e-01 5.47552943e-01 7.64661074e-01 1.47127062e-01
7.58838356e-01 2.06729189e-01 9.41190422e-01 -3.62156123e-01
3.14210564e-01 8.41564536e-02 1.22711696e-02 8.76815319e-01
3.15458216e-02 -1.74123179e-02 5.48486352e-01 -4.46195632e-01
1.27172434e+00 -1.35730430e-01 -5.68760574e-01 -3.87144059e-01
-7.56992638e-01 7.68108010e-01 3.02454323e-01 8.03946555e-01
-6.40585423e-01 2.01206356e-02 -1.42912120e-01 2.26813689e-01
4.40058053e-01 6.26665711e-01 -5.77484854e-02 2.61309236e-01
-9.21746373e-01 3.52964550e-01 1.06365418e+00 1.05947328e+00
6.42624259e-01 -4.22158018e-02 -3.06449711e-01 2.80837446e-01
-4.57054786e-02 6.39497042e-01 -8.60767365e-02 -1.45480895e+00
-6.01532720e-02 4.52889204e-01 6.11796737e-01 -7.24338353e-01
-3.89746875e-01 -2.78537739e-02 -8.00580144e-01 6.46337628e-01
9.59963500e-01 -6.46820903e-01 -6.90611362e-01 1.49838436e+00
5.44870436e-01 4.59105708e-02 2.65887260e-01 9.03789222e-01
-3.90206501e-02 1.27551687e+00 -2.47974142e-01 -7.77447343e-01
9.33920801e-01 -5.01694500e-01 -4.64300275e-01 2.64145583e-01
7.69088984e-01 -6.12633705e-01 8.55543911e-01 1.58189595e-01
-1.59762323e+00 -3.17195393e-02 -8.37722600e-01 1.19220428e-01
1.68625966e-01 9.32674408e-02 -3.15702945e-01 -6.62187338e-02
-1.20309150e+00 9.57063496e-01 -8.18569660e-01 -2.28688851e-01
-1.49864554e-01 3.44623506e-01 -1.70466840e-01 3.65405023e-01
-1.06360757e+00 7.68081605e-01 -2.40511894e-01 1.35249585e-01
-2.34220386e-01 -1.29748654e+00 -4.98609453e-01 -5.63311689e-02
1.43356621e-01 -6.37691915e-01 1.19621778e+00 -9.15097117e-01
-1.62641525e+00 5.23563325e-01 -1.50429115e-01 -1.91351324e-01
9.62153554e-01 5.57067931e-01 2.86811203e-01 1.78321421e-01
2.70218074e-01 1.07616790e-01 4.89850491e-01 -1.47040427e+00
-3.79948825e-01 -1.49784014e-01 -7.63504207e-02 1.53416157e-01
7.07455650e-02 -2.32726514e-01 -1.85668826e-01 -2.15358853e-01
1.09910227e-01 -1.45051730e+00 -6.35638237e-01 2.87172407e-01
-2.05926284e-01 1.96795067e-04 8.11055422e-01 -4.77660090e-01
9.64059770e-01 -2.03682232e+00 4.11902487e-01 2.31781676e-01
-1.07135482e-01 1.62846714e-01 4.25342143e-01 5.55090070e-01
1.71228066e-01 1.23389147e-01 -1.10675621e+00 -9.96538624e-02
-4.70364630e-01 1.29474938e-01 -2.72239894e-01 1.02064466e+00
2.78683931e-01 4.47862774e-01 -7.39289403e-01 -4.74311709e-01
-2.52065390e-01 6.34646475e-01 -8.65932524e-01 -1.07317075e-01
-8.37515295e-02 6.90141439e-01 -6.64582133e-01 -2.44845346e-01
9.43855941e-01 1.56112447e-01 -1.71358839e-01 1.90601140e-01
-1.03349197e+00 -5.13287246e-01 -1.26864290e+00 8.23636889e-01
-6.65180266e-01 5.71434736e-01 9.22236860e-01 -1.16403031e+00
8.89994323e-01 5.23800015e-01 8.44985783e-01 -8.69089738e-02
4.31139469e-01 5.22857666e-01 -3.07634652e-01 -5.68385959e-01
3.99374098e-01 -9.67891157e-01 -8.99655521e-02 4.04226640e-03
-4.71033692e-01 -6.35738194e-01 1.29155386e-02 -4.67496999e-02
1.13606930e+00 -2.85310894e-01 -4.41462882e-02 -9.06085253e-01
9.84033346e-01 6.84165657e-01 5.04591405e-01 1.37021035e-01
-7.61643276e-02 7.22186029e-01 9.50119555e-01 -3.64919640e-02
-1.27411389e+00 -5.34633279e-01 -6.12291098e-01 2.61681288e-01
4.50977713e-01 2.54605532e-01 -8.59390557e-01 2.07328964e-02
2.68169224e-01 6.63646698e-01 -9.86694813e-01 -1.68085396e-01
-1.04725254e+00 -2.91068494e-01 2.98613429e-01 3.66604388e-01
7.44552791e-01 -7.00210094e-01 -6.93404853e-01 3.48591566e-01
1.95424005e-01 -7.24601865e-01 -8.21004152e-01 8.15848336e-02
-1.02740383e+00 -9.36928868e-01 -1.42382860e+00 -1.01169479e+00
1.02011597e+00 -1.78347707e-01 2.21876472e-01 -4.85408865e-02
3.35191302e-02 6.88784778e-01 4.18936759e-01 1.20036907e-01
-7.22316146e-01 -2.31662765e-01 -1.46751285e-01 1.67649776e-01
-7.05436349e-01 -1.60880134e-01 -4.19258982e-01 6.94721699e-01
-1.04941249e+00 -2.14264363e-01 -1.84321374e-01 5.04695833e-01
7.52918363e-01 1.43777862e-01 2.09327146e-01 -4.95567709e-01
8.42324197e-01 -4.77197349e-01 -1.00445473e+00 -1.38416618e-01
-1.47689447e-01 2.76897728e-01 1.01635444e+00 -8.97120357e-01
-1.05208862e+00 3.70253265e-01 1.26146644e-01 -4.10654485e-01
2.06223175e-01 2.92992145e-01 2.36347333e-01 -5.29101908e-01
8.92741442e-01 6.50480166e-02 2.87508667e-01 -3.95144939e-01
-3.87389697e-02 3.97598475e-01 5.50715864e-01 -6.91739857e-01
4.48228896e-01 8.09492052e-01 8.82838905e-01 -1.36285949e+00
-2.48020887e-01 -5.09096980e-01 -7.47698486e-01 -3.05745155e-01
9.93394613e-01 4.05850820e-02 -8.63399863e-01 5.60370207e-01
-1.37875068e+00 -8.28152001e-01 -9.25121188e-01 2.51147300e-01
-9.76876140e-01 2.17906192e-01 -7.24252880e-01 -1.19007444e+00
1.21129699e-01 -8.36074531e-01 7.83855200e-01 1.44944459e-01
-1.48055941e-01 -1.37598264e+00 4.53913391e-01 -4.07003224e-01
5.42422354e-01 5.98278999e-01 6.56397402e-01 1.80957422e-01
-2.35388890e-01 -2.49306664e-01 2.82443434e-01 3.53941019e-03
-1.84632316e-01 1.59636289e-01 -1.45841256e-01 -6.60503730e-02
7.24740982e-01 3.15606058e-01 5.32577574e-01 8.05303097e-01
8.43670890e-02 -5.54717839e-01 -7.36397147e-01 4.68015075e-01
1.55451143e+00 3.90593082e-01 4.40614849e-01 9.32624117e-02
3.60079259e-01 6.42187715e-01 3.21471006e-01 2.49300823e-01
-3.30579355e-02 5.89335144e-01 -5.13811298e-02 -2.28727087e-01
1.87370658e-01 2.10906535e-01 2.09366024e-01 1.87439963e-01
-4.25060451e-01 6.89030141e-02 -1.01277947e+00 9.70707953e-01
-1.75425076e+00 -8.48507941e-01 -7.25471139e-01 2.50538230e+00
1.02416801e+00 -2.32856557e-01 2.69999474e-01 2.01018825e-01
8.20040166e-01 -5.97344160e-01 -4.18536127e-01 -8.73299837e-01
2.22800389e-01 -4.19434952e-03 1.04834008e+00 1.19388938e+00
-7.64444888e-01 3.19416076e-01 6.65767956e+00 2.19814569e-01
-1.58981395e+00 -1.91641562e-02 8.74067023e-02 3.03912938e-01
-2.87054300e-01 6.17937706e-02 -6.34388506e-01 2.54739732e-01
9.03143764e-01 -6.07246041e-01 1.83066051e-03 6.69793189e-01
6.16945744e-01 -3.02459776e-01 -6.19547665e-01 -1.76757458e-05
-4.31936353e-01 -1.08295226e+00 -4.81214315e-01 4.10317808e-01
1.15526807e+00 -7.27962911e-01 -1.02472045e-01 -2.16514766e-01
1.02290906e-01 -6.32434368e-01 5.11083007e-01 7.09087551e-01
6.29065514e-01 -6.65553510e-01 2.50481606e-01 8.09364080e-01
-1.30404544e+00 8.52941200e-02 -2.89404482e-01 -1.65469751e-01
6.62317693e-01 3.30478132e-01 -5.63638031e-01 1.42238647e-01
1.27193993e-02 2.61566520e-01 4.70229685e-01 1.26463223e+00
4.68330890e-01 3.85358155e-01 -6.47348583e-01 -5.92121966e-02
4.73590404e-01 -7.60234237e-01 1.03318560e+00 9.08117831e-01
6.91966712e-01 7.92344809e-01 9.81269330e-02 1.17546821e+00
4.49466288e-01 2.10804328e-01 -7.52640784e-01 5.60544074e-01
-3.66455823e-01 7.36190736e-01 -8.25104892e-01 -2.28485018e-01
-9.32028890e-02 8.72349858e-01 -2.07048744e-01 5.74518561e-01
-7.54316628e-01 -6.15697026e-01 6.25595629e-01 7.26408243e-01
2.15516478e-01 -7.07736790e-01 -3.18457484e-01 -6.49107575e-01
-1.52359664e-01 1.66252434e-01 2.34793410e-01 -4.93758261e-01
-7.08393753e-01 5.19411802e-01 5.02636850e-01 -1.15437961e+00
-2.35914513e-01 -3.29670310e-01 -8.24570954e-01 1.22865045e+00
-9.15036142e-01 -5.08676231e-01 9.18035805e-02 5.72392881e-01
1.17664367e-01 6.32006526e-01 3.43282998e-01 6.41656592e-02
-2.15200514e-01 -8.18813071e-02 3.21669161e-01 -2.59526283e-01
2.49558493e-01 -1.07798254e+00 -1.05980739e-01 7.44902253e-01
-1.19052458e+00 3.97075772e-01 1.01308584e+00 -9.36300337e-01
-1.37767577e+00 -8.31001937e-01 1.04461408e+00 -7.88895879e-03
5.11182666e-01 2.15965062e-01 -1.41910970e+00 6.83828294e-01
4.93240803e-02 2.93675512e-01 -2.46480122e-01 -8.67846370e-01
6.99973881e-01 2.68270671e-01 -1.54625535e+00 5.96081138e-01
5.91322362e-01 1.94464624e-01 -2.94201195e-01 -1.70112506e-03
3.34735930e-01 -5.94169438e-01 -9.85858679e-01 4.39987630e-01
1.81342140e-01 -3.73231590e-01 6.69251978e-01 -6.67842031e-01
3.31200987e-01 -4.83260632e-01 3.46654445e-01 -1.38238585e+00
-3.86242330e-01 -1.11478603e+00 5.87444544e-01 8.34956706e-01
6.54972911e-01 -8.99212599e-01 5.28316259e-01 1.07975054e+00
-3.02075803e-01 -1.05156541e+00 -1.04249895e+00 -1.08842921e+00
9.44900930e-01 4.11414914e-03 -1.71395510e-01 8.07995319e-01
4.48101103e-01 -1.49142936e-01 -2.16854259e-01 3.51721168e-01
4.83771473e-01 -1.32748902e-01 2.65285790e-01 -1.03383839e+00
-2.93305889e-02 -3.25144619e-01 -1.87998220e-01 -1.25311792e+00
-1.66358128e-01 -6.00762069e-01 5.71692944e-01 -1.79006362e+00
-5.26883066e-01 -8.34392846e-01 5.09906292e-01 6.04222715e-02
2.34717712e-01 -4.60847728e-02 -7.04860166e-02 4.07998204e-01
3.03791851e-01 2.84844816e-01 1.79268479e+00 1.89120412e-01
-9.94486392e-01 5.86536407e-01 -1.44661948e-01 5.86053371e-01
5.68852663e-01 -3.49266261e-01 -3.27008933e-01 1.02663726e-01
-1.74266070e-01 7.24878311e-01 7.89540708e-02 -1.00463212e+00
5.48610687e-01 -5.62123060e-01 -2.10045516e-01 -8.04818869e-02
3.41607809e-01 -6.95481360e-01 3.01569045e-01 6.74710631e-01
-3.66248250e-01 -1.97649136e-01 5.28325438e-01 2.33734071e-01
1.50067240e-01 -6.74848437e-01 1.40169811e+00 -1.80333778e-02
-5.17116189e-02 6.29287213e-02 -1.08736205e+00 3.02658200e-01
1.36695898e+00 -3.31254870e-01 2.70005792e-01 -4.53178674e-01
-1.00465739e+00 1.50117427e-01 7.38937318e-01 -5.08977830e-01
2.03845501e-01 -1.35975969e+00 -4.41929191e-01 4.15056944e-01
-7.25847006e-01 9.47192609e-02 2.04956323e-01 1.34370959e+00
-8.24059248e-01 1.83797315e-01 -1.80872276e-01 -6.11710966e-01
-1.02608895e+00 5.22003233e-01 1.02887833e+00 3.04858778e-02
-6.95774019e-01 5.22458673e-01 1.32298917e-01 -4.16085757e-02
-2.83137679e-01 -8.12802076e-01 -3.30412537e-02 -1.21295959e-01
1.19875446e-01 7.83933342e-01 -2.37265870e-01 -8.80821943e-01
-1.21532567e-01 1.25072217e+00 9.23886538e-01 -6.53665364e-01
1.15326524e+00 -1.65103208e-02 -4.25750986e-02 4.36802953e-01
1.57893229e+00 2.68019646e-01 -1.60628724e+00 1.99003741e-01
-3.35054994e-01 -2.12854668e-01 -9.94197950e-02 7.76909441e-02
-1.00503254e+00 5.20503700e-01 -5.61198629e-02 9.05924797e-01
9.41186368e-01 3.21988463e-02 8.64331067e-01 -2.24441171e-01
-1.15667760e-01 -8.94690394e-01 -3.24758351e-01 8.75054598e-01
1.22454953e+00 -4.52955931e-01 -2.13364914e-01 -7.71127522e-01
-5.77683508e-01 1.22722018e+00 2.77540535e-01 -6.14065409e-01
9.01740670e-01 7.86234319e-01 -1.65113196e-01 1.23032615e-01
-3.80255133e-01 -6.14741817e-02 3.17258477e-01 2.82087713e-01
4.35972691e-01 -2.30158314e-01 -8.82941961e-01 7.51141682e-02
1.66420698e-01 3.52175802e-01 8.55501294e-01 1.15280867e+00
-6.58492506e-01 -9.54001248e-01 -8.02093804e-01 -1.63696989e-01
-2.09989116e-01 5.63886285e-01 -3.13646853e-01 9.59839344e-01
-6.12638332e-02 5.10903656e-01 2.61055499e-01 3.59221667e-01
9.88013446e-01 2.74828315e-01 5.32506347e-01 -3.38555813e-01
-3.95002753e-01 4.67891544e-02 -1.61997825e-01 -1.53542057e-01
-1.15109414e-01 -7.26641774e-01 -1.95162404e+00 -4.54980642e-01
-1.37064293e-01 5.75590074e-01 4.50396717e-01 7.13621855e-01
1.58675253e-01 2.87628829e-01 5.41035712e-01 -1.18232858e+00
-4.77504462e-01 -5.83512306e-01 -6.65605724e-01 6.29562214e-02
6.12490594e-01 -4.98156369e-01 -7.89045453e-01 1.37429506e-01] | [6.204409599304199, 3.3252112865448] |
f208e44a-b953-41f2-b8cf-330d3340a956 | multi-step-greedy-policies-in-model-free-deep-1 | 1910.02919 | null | https://arxiv.org/abs/1910.02919v3 | https://arxiv.org/pdf/1910.02919v3.pdf | Multi-step Greedy Reinforcement Learning Algorithms | Multi-step greedy policies have been extensively used in model-based reinforcement learning (RL), both when a model of the environment is available (e.g.,~in the game of Go) and when it is learned. In this paper, we explore their benefits in model-free RL, when employed using multi-step dynamic programming algorithms: $\kappa$-Policy Iteration ($\kappa$-PI) and $\kappa$-Value Iteration ($\kappa$-VI). These methods iteratively compute the next policy ($\kappa$-PI) and value function ($\kappa$-VI) by solving a surrogate decision problem with a shaped reward and a smaller discount factor. We derive model-free RL algorithms based on $\kappa$-PI and $\kappa$-VI in which the surrogate problem can be solved by any discrete or continuous action RL method, such as DQN and TRPO. We identify the importance of a hyper-parameter that controls the extent to which the surrogate problem is solved and suggest a way to set this parameter. When evaluated on a range of Atari and MuJoCo benchmark tasks, our results indicate that for the right range of $\kappa$, our algorithms outperform DQN and TRPO. This shows that our multi-step greedy algorithms are general enough to be applied over any existing RL algorithm and can significantly improve its performance. | ['Mohammad Ghavamzadeh', 'Yonathan Efroni', 'Manan Tomar'] | 2019-10-07 | null | https://proceedings.icml.cc/static/paper_files/icml/2020/5786-Paper.pdf | https://proceedings.icml.cc/static/paper_files/icml/2020/5786-Paper.pdf | icml-2020-1 | ['game-of-go'] | ['playing-games'] | [-4.46224213e-03 1.32880583e-01 -5.44517875e-01 -1.19025388e-03
-7.21237540e-01 -5.47212362e-01 4.55776423e-01 4.00840975e-02
-1.09085596e+00 1.17135954e+00 -3.51577491e-01 -5.78209162e-01
-5.55644333e-01 -8.22154462e-01 -6.03425860e-01 -7.39515364e-01
-4.16539788e-01 7.21440196e-01 2.69449145e-01 -5.66515386e-01
4.27631795e-01 2.00733110e-01 -1.47742093e+00 -2.20598355e-01
7.78244555e-01 9.19545591e-01 3.69771987e-01 7.01536596e-01
3.25234719e-02 8.37137878e-01 -4.61109012e-01 1.07425794e-01
4.75733399e-01 -6.13111258e-01 -8.33003163e-01 -3.68987471e-01
-5.53613901e-01 -3.61771464e-01 1.74446762e-01 8.67590249e-01
5.03137589e-01 6.53670549e-01 4.11153793e-01 -1.27548110e+00
1.21046595e-01 7.11944401e-01 -6.96959138e-01 2.80477434e-01
3.80347520e-01 5.87631524e-01 9.19932485e-01 -1.77150667e-01
6.12956345e-01 1.34665000e+00 1.72128499e-01 6.75760210e-01
-1.40913856e+00 -7.51870513e-01 5.98311663e-01 7.61867547e-03
-1.08120430e+00 -2.69235540e-02 5.99832773e-01 -2.38659590e-01
1.29329896e+00 -4.30107489e-02 8.13735783e-01 8.16673517e-01
3.58180583e-01 6.98910952e-01 1.67485285e+00 -5.72348833e-01
8.29240501e-01 -1.77562952e-01 -3.36901546e-01 7.10958183e-01
-1.28225043e-01 6.56402886e-01 -4.04098421e-01 -2.69296765e-01
8.28180015e-01 -5.18221736e-01 1.85757354e-01 -4.08873379e-01
-8.29228342e-01 1.01186991e+00 1.70973852e-01 1.43986970e-01
-4.82431531e-01 7.57790685e-01 2.64150470e-01 7.43267357e-01
1.88654870e-01 7.73891747e-01 -6.00405395e-01 -6.31845593e-01
-6.92205429e-01 6.97601378e-01 7.35337496e-01 5.56907892e-01
9.04993057e-01 2.20261976e-01 -1.16065927e-01 6.92050755e-01
3.33826900e-01 4.18333918e-01 3.76165509e-01 -1.58880651e+00
5.42188466e-01 2.98118860e-01 7.82818675e-01 -3.04433227e-01
-4.84948516e-01 -1.24567181e-01 -6.89387470e-02 8.04269135e-01
5.59014738e-01 -4.32666451e-01 -7.94598341e-01 2.09435105e+00
2.77253687e-01 2.51174197e-02 7.24945217e-02 7.37773836e-01
-4.17855084e-02 5.94698668e-01 3.38794827e-01 -5.71355283e-01
9.53251600e-01 -7.55787492e-01 -3.10782701e-01 -5.32742739e-01
7.10144043e-01 -3.92055541e-01 1.31022513e+00 7.21128464e-01
-1.27682912e+00 -2.00112760e-01 -8.04811478e-01 6.14591837e-01
-1.18868202e-01 -3.39013427e-01 7.32828975e-01 5.89739084e-01
-1.23909855e+00 8.44292939e-01 -8.26653659e-01 -1.25081956e-01
2.44622901e-02 6.62640512e-01 1.78242281e-01 1.01953456e-02
-1.14850247e+00 1.02169621e+00 6.34079814e-01 -1.58605918e-01
-1.25722980e+00 -3.07162881e-01 -5.11721909e-01 -5.16558476e-02
9.55296278e-01 -3.92376363e-01 1.59342504e+00 -1.02551377e+00
-1.92027557e+00 4.85684752e-01 8.87960345e-02 -6.08197510e-01
6.37014091e-01 3.45403515e-02 -7.34350737e-03 1.42935812e-01
-2.65922174e-02 8.04548562e-01 8.32270741e-01 -1.29178190e+00
-8.78308237e-01 1.11745363e-02 7.01479077e-01 6.79116189e-01
3.70469064e-01 8.98542032e-02 -9.41424891e-02 -2.94286668e-01
-2.25377932e-01 -1.18763876e+00 -7.46849358e-01 -4.70987529e-01
8.69261026e-02 -3.45043242e-01 2.86110282e-01 -2.02451229e-01
1.27776659e+00 -1.72081614e+00 2.14668021e-01 3.55377614e-01
-6.45842254e-02 2.46513918e-01 -2.26918459e-01 4.79929656e-01
1.53230533e-01 2.17587650e-01 -1.39580086e-01 -4.03010063e-02
5.10829501e-02 6.46998942e-01 -2.99450420e-02 9.80660021e-02
-3.63443524e-01 6.99660718e-01 -1.19904077e+00 -2.41926193e-01
1.56715840e-01 -8.85984898e-02 -7.27571607e-01 2.87412047e-01
-8.71274471e-01 4.14709955e-01 -7.90115535e-01 3.40897053e-01
1.75815731e-01 5.61406761e-02 5.12513697e-01 5.49186230e-01
-4.08323884e-01 2.59480298e-01 -1.21104670e+00 1.40374565e+00
-5.38785934e-01 2.41380166e-02 2.73853898e-01 -9.69731987e-01
9.13358271e-01 1.45134196e-01 5.89008927e-01 -9.40647542e-01
2.44486168e-01 1.76627919e-01 1.93568334e-01 -1.93886593e-01
4.48271215e-01 -4.28874165e-01 -2.84497112e-01 8.56917500e-01
-1.84404030e-01 -2.46875122e-01 4.89187866e-01 2.05935701e-03
1.32415414e+00 5.49100339e-01 4.29567546e-01 -5.62803447e-01
4.01371479e-01 1.19762182e-01 6.98586226e-01 1.22673035e+00
-4.53991771e-01 -2.10266113e-01 9.62013364e-01 -3.84805739e-01
-7.96479404e-01 -8.48534286e-01 4.10394907e-01 1.65579557e+00
1.04758114e-01 -1.07195206e-01 -6.26652241e-01 -7.70344019e-01
-3.76740247e-02 1.06485498e+00 -6.51932716e-01 -2.43482649e-01
-8.41369748e-01 -7.54438698e-01 2.36977398e-01 4.20569897e-01
3.73198450e-01 -1.58516610e+00 -1.04832697e+00 5.96096456e-01
5.56426197e-02 -4.95763361e-01 -2.85340995e-01 7.63085604e-01
-9.62122619e-01 -8.94557297e-01 -3.07997823e-01 -2.77880043e-01
5.99824429e-01 2.64335722e-02 1.10137427e+00 1.09603554e-01
1.13627821e-01 6.39835477e-01 -3.44411999e-01 -2.84607053e-01
-3.70466769e-01 7.06689572e-03 7.73724616e-02 -5.80575168e-01
-2.06585154e-02 -6.07834518e-01 -6.82968795e-01 4.11067277e-01
-7.01586246e-01 -1.27326161e-01 5.10862708e-01 8.29063654e-01
7.44943380e-01 2.62666047e-02 6.17744982e-01 -9.28498983e-01
1.19296730e+00 -3.59867454e-01 -1.00527740e+00 1.74852550e-01
-1.05285120e+00 6.13429368e-01 7.52741277e-01 -6.94306910e-01
-9.35437202e-01 -8.94042552e-02 -1.17411666e-01 -4.45409775e-01
1.02894515e-01 5.99768400e-01 8.95117819e-02 -1.37435988e-01
6.52473509e-01 8.85792673e-02 4.82457578e-02 -2.21073210e-01
3.93007964e-01 4.00678553e-02 -4.63355258e-02 -1.08930850e+00
4.66835499e-01 -6.61702966e-03 -2.44116280e-02 -1.37578487e-01
-5.48029780e-01 -1.12760358e-01 -1.45538032e-01 -2.46455163e-01
5.22692263e-01 -6.15251064e-01 -1.21537697e+00 1.65564373e-01
-4.41279441e-01 -1.17969608e+00 -5.08999586e-01 4.51508850e-01
-1.11381114e+00 5.94871305e-02 -3.72258484e-01 -1.01143003e+00
-1.12521224e-01 -1.28422189e+00 3.41082871e-01 4.54035640e-01
4.22053896e-02 -9.58166957e-01 3.98463964e-01 1.32219583e-01
4.11721498e-01 1.29804984e-01 9.66927111e-01 -3.16395760e-01
-3.94293845e-01 3.91725332e-01 9.13389847e-02 1.45835683e-01
-2.77409494e-01 -2.12836668e-01 -4.78089690e-01 -6.71473384e-01
-2.54690319e-01 -6.58409953e-01 6.57680511e-01 5.47409236e-01
9.16267395e-01 -5.42986989e-01 -1.54660463e-01 2.55678505e-01
1.54349756e+00 9.43470716e-01 4.31938887e-01 9.39512968e-01
6.13738298e-02 1.73055232e-01 1.05191779e+00 7.33564377e-01
4.11590308e-01 5.91957510e-01 7.27485359e-01 2.99509019e-01
5.04281461e-01 -2.95096964e-01 4.82597768e-01 1.85992777e-01
-3.72030586e-01 -7.75634050e-02 -8.43661785e-01 3.42702985e-01
-2.08776975e+00 -9.21569169e-01 6.60001993e-01 2.32261372e+00
1.04712915e+00 6.10415280e-01 5.16027749e-01 -1.07052289e-01
2.39057332e-01 2.57579207e-01 -8.78514826e-01 -1.15695632e+00
4.16091144e-01 5.90820968e-01 6.69639170e-01 7.74617195e-01
-6.28700674e-01 1.34314203e+00 6.56757832e+00 9.45215464e-01
-1.00851607e+00 3.96333672e-02 5.08378685e-01 -3.94175261e-01
-2.70460725e-01 2.61919856e-01 -7.31003761e-01 5.15688062e-01
1.05948091e+00 -2.58037180e-01 1.03705466e+00 1.04455662e+00
5.24025679e-01 -8.05289924e-01 -9.96069193e-01 6.00108802e-01
-5.24316072e-01 -9.26176786e-01 -5.62194943e-01 2.35226050e-01
6.45705819e-01 1.16841160e-02 -3.73243243e-02 9.09057736e-01
1.22226250e+00 -9.78859782e-01 7.36167014e-01 1.90623939e-01
5.59119940e-01 -1.14548123e+00 4.51024592e-01 7.01644480e-01
-1.14960790e+00 -3.92299324e-01 -2.41770804e-01 -3.02806020e-01
-5.70688993e-02 -9.76749603e-03 -8.12446356e-01 3.07705104e-01
6.26554668e-01 2.56019905e-02 -7.82264173e-02 6.54779553e-01
-4.86370295e-01 5.39872229e-01 -3.62992644e-01 -3.46699297e-01
7.96281219e-01 -3.89896989e-01 3.54394644e-01 7.62730896e-01
1.64134771e-01 2.70995617e-01 5.26463270e-01 7.36994743e-01
2.37450704e-01 -5.26472330e-02 -3.64761740e-01 1.12318091e-01
4.90519941e-01 9.08596277e-01 -6.92437410e-01 -4.10981884e-04
1.61369696e-01 3.66764307e-01 5.55032790e-01 4.22742635e-01
-7.88302124e-01 -7.25632757e-02 6.99000061e-01 -4.47219610e-02
3.00968617e-01 -3.33699107e-01 1.08138911e-01 -7.73696661e-01
-2.93991983e-01 -1.09496164e+00 4.90815997e-01 -5.99244237e-01
-6.06393754e-01 4.08975005e-01 4.39864129e-01 -9.35925782e-01
-7.99882174e-01 -3.67575556e-01 -4.66282994e-01 7.35498726e-01
-1.57294393e+00 -4.85170037e-01 3.51239085e-01 6.86407745e-01
4.17901069e-01 -3.21388394e-02 7.07113862e-01 -3.61915737e-01
-4.83294696e-01 2.81031013e-01 2.72622466e-01 -3.79610538e-01
2.42956936e-01 -1.33734083e+00 -1.26656443e-01 4.72082675e-01
-3.96541595e-01 3.62763196e-01 9.32132959e-01 -4.39916253e-01
-1.29691100e+00 -5.95747530e-01 2.40540817e-01 9.70526636e-02
5.91670215e-01 -8.47228095e-02 -5.18687308e-01 5.91271996e-01
6.05237819e-02 -1.26404062e-01 2.96431452e-01 8.80252123e-02
8.21425021e-02 -2.48594731e-01 -1.43218279e+00 7.06087947e-01
8.56499434e-01 -1.44313976e-01 -3.29542935e-01 -3.59680131e-02
6.05392873e-01 -4.87592161e-01 -7.59036183e-01 1.96364447e-01
5.15216947e-01 -1.05170631e+00 7.86900342e-01 -5.56862295e-01
-4.22640890e-02 -8.93031955e-02 1.20540457e-02 -1.58426452e+00
-2.74868608e-01 -9.60572481e-01 -6.59150183e-02 6.92235649e-01
4.68039185e-01 -7.92704344e-01 7.19456017e-01 6.89151466e-01
1.52010858e-01 -1.08043623e+00 -1.23547506e+00 -8.81771624e-01
3.70279193e-01 -4.73447919e-01 3.45055044e-01 5.11158288e-01
-4.62507037e-03 -1.23377986e-01 -4.67428207e-01 -2.78962880e-01
6.54163539e-01 1.14352979e-01 5.03892601e-01 -7.88104534e-01
-6.96549654e-01 -6.00990951e-01 3.77746254e-01 -1.04337454e+00
2.24938601e-01 -3.85630608e-01 7.45933577e-02 -1.43420720e+00
-6.62254319e-02 -1.11308777e+00 -6.39568567e-01 8.31792593e-01
6.14659116e-03 -4.19991165e-01 4.93591666e-01 2.06830010e-01
-7.63293147e-01 4.34777707e-01 1.27987480e+00 1.45353839e-01
-7.29704797e-01 1.07739158e-02 -6.42147183e-01 7.00398207e-01
1.15148401e+00 -7.91788042e-01 -6.12664461e-01 -1.45273671e-01
5.92715442e-01 6.13007903e-01 -3.27127799e-02 -7.38400340e-01
-2.43321694e-02 -9.75763083e-01 -9.20747221e-02 -1.38678506e-01
4.62295860e-01 -4.56275195e-01 -5.01649231e-02 8.40182722e-01
-6.85100794e-01 1.72077000e-01 1.29298374e-01 5.16511858e-01
2.44460210e-01 -5.79236805e-01 8.62983942e-01 -6.03276789e-01
-7.58550525e-01 4.85924399e-03 -6.85282111e-01 2.91203350e-01
1.22210228e+00 -9.56492499e-02 -1.73295513e-01 -4.53111768e-01
-8.71833205e-01 6.87337637e-01 3.46544862e-01 2.08903819e-01
4.73561406e-01 -9.00006831e-01 -2.36545384e-01 -1.09714516e-01
-1.93536729e-01 -1.92947656e-01 -1.28614530e-01 7.24761784e-01
-3.81286174e-01 3.62217933e-01 -3.90765607e-01 -1.38831913e-01
-1.01952636e+00 5.62734962e-01 6.17601097e-01 -1.00424957e+00
-2.37886488e-01 7.44697452e-01 -1.45406246e-01 -3.92900318e-01
3.16843808e-01 -2.81892180e-01 -9.70144272e-02 -1.79898024e-01
2.24180996e-01 3.87681484e-01 -3.95474702e-01 -7.57110417e-02
-2.42936552e-01 4.08098698e-01 -5.45562282e-02 -8.01579356e-01
1.24849856e+00 -1.60294324e-01 1.63420469e-01 3.29258651e-01
4.98840541e-01 -3.69908124e-01 -1.74422717e+00 -8.56451094e-02
4.15791310e-02 -3.97272974e-01 1.28063738e-01 -1.12544107e+00
-7.98166692e-01 4.63608325e-01 5.79657257e-01 1.10541835e-01
1.10222518e+00 -2.51399368e-01 1.59493968e-01 5.27476847e-01
1.05272043e+00 -1.47700167e+00 1.69901654e-01 6.29342616e-01
4.48506355e-01 -9.92789268e-01 1.92802340e-01 3.55617404e-01
-9.37944829e-01 8.02665830e-01 8.35117817e-01 -2.79992700e-01
3.43317330e-01 2.20334366e-01 -1.21985041e-01 1.00789323e-01
-1.22271430e+00 -5.30504107e-01 -3.30312610e-01 4.87029761e-01
1.51470136e-02 1.75069764e-01 -8.40348363e-01 4.40931946e-01
-2.77376026e-02 1.42695159e-01 4.35829073e-01 1.48059607e+00
-7.98608482e-01 -1.69315624e+00 -2.56529331e-01 3.58321875e-01
-3.09336215e-01 9.17892531e-02 -1.41904250e-01 1.01503301e+00
5.99844195e-02 1.02492380e+00 -2.71244049e-01 -1.86181352e-01
1.42004430e-01 1.87233891e-02 6.44050062e-01 -3.58782709e-01
-9.91575897e-01 2.67331511e-01 2.55278558e-01 -9.06102598e-01
-3.88119251e-01 -5.43026686e-01 -1.58725953e+00 -3.76473784e-01
3.00969612e-02 4.01560724e-01 5.24025083e-01 9.94240344e-01
7.14723691e-02 2.37930357e-01 8.16828012e-01 -7.56021678e-01
-8.75805199e-01 -5.97227812e-01 -5.89614809e-01 -4.59956639e-02
-3.36344540e-02 -9.71629441e-01 -3.81328940e-01 -5.90500176e-01] | [4.114576816558838, 2.1623358726501465] |
e983a3bd-96b4-49ec-8f8b-69b3c63108f7 | motionhint-self-supervised-monocular-visual | 2109.06768 | null | https://arxiv.org/abs/2109.06768v3 | https://arxiv.org/pdf/2109.06768v3.pdf | MotionHint: Self-Supervised Monocular Visual Odometry with Motion Constraints | We present a novel self-supervised algorithm named MotionHint for monocular visual odometry (VO) that takes motion constraints into account. A key aspect of our approach is to use an appropriate motion model that can help existing self-supervised monocular VO (SSM-VO) algorithms to overcome issues related to the local minima within their self-supervised loss functions. The motion model is expressed with a neural network named PPnet. It is trained to coarsely predict the next pose of the camera and the uncertainty of this prediction. Our self-supervised approach combines the original loss and the motion loss, which is the weighted difference between the prediction and the generated ego-motion. Taking two existing SSM-VO systems as our baseline, we evaluate our MotionHint algorithm on the standard KITTI benchmark. Experimental results show that our MotionHint algorithm can be easily applied to existing open-sourced state-of-the-art SSM-VO systems to greatly improve the performance by reducing the resulting ATE by up to 28.73%. | ['Dinesh Manocha', 'Yu-Ping Wang', 'Cong Wang'] | 2021-09-14 | null | null | null | null | ['monocular-visual-odometry'] | ['robots'] | [-3.32161725e-01 1.58991531e-01 -5.79445004e-01 -4.30939227e-01
-4.72394705e-01 -1.26615867e-01 6.19049489e-01 -5.72171450e-01
-4.39788967e-01 6.05782926e-01 3.40720385e-01 8.24156180e-02
2.53186047e-01 -3.84623289e-01 -8.70230973e-01 -4.52001214e-01
2.17424467e-01 6.77723467e-01 5.00909388e-01 2.81396024e-02
2.08397791e-01 3.80345672e-01 -1.55411315e+00 -2.37759098e-01
7.50648201e-01 8.29462171e-01 3.51431310e-01 7.60587871e-01
2.71585613e-01 1.17888916e+00 -1.34380087e-01 1.12616140e-02
5.39190352e-01 -3.71595114e-01 -8.96264255e-01 2.10855052e-01
9.59504247e-01 -3.85842383e-01 -7.22382724e-01 1.01113796e+00
4.99087274e-01 1.03496984e-01 6.05748951e-01 -1.35911822e+00
-1.80243805e-01 1.07170433e-01 -4.56322044e-01 4.81620664e-03
3.89090657e-01 4.48253065e-01 8.99399579e-01 -1.04760253e+00
1.32882571e+00 1.36681771e+00 7.80329525e-01 5.99051416e-01
-1.20532775e+00 -1.74138695e-01 -5.13882786e-02 3.97381157e-01
-1.39796698e+00 -4.84967798e-01 7.73311317e-01 -6.21859252e-01
1.22743928e+00 -2.57248998e-01 7.03769445e-01 8.76001060e-01
4.72443134e-01 1.04375958e+00 7.66158819e-01 -2.47726649e-01
1.16213389e-01 3.08039263e-02 -1.59030765e-01 1.00465000e+00
2.22104818e-01 4.97587174e-01 -6.78131461e-01 1.00303061e-01
7.17050731e-01 -4.13419902e-01 -2.29512438e-01 -1.45811641e+00
-1.26317251e+00 7.31494427e-01 7.63134181e-01 -1.99980617e-01
-1.06609203e-01 4.44265038e-01 2.21716136e-01 1.98735073e-01
3.59948605e-01 3.52712423e-01 -3.51876795e-01 -2.50820220e-01
-1.14324820e+00 3.45426798e-01 7.37528265e-01 9.82315421e-01
1.13073444e+00 2.45836318e-01 3.04467417e-03 5.73526382e-01
6.97643757e-01 4.77823764e-01 5.37751555e-01 -1.36977220e+00
4.49371397e-01 5.07887125e-01 2.04515070e-01 -8.57594907e-01
-4.13303137e-01 -1.49890929e-01 -3.41530561e-01 5.09238720e-01
1.18937150e-01 -1.74043000e-01 -1.01872933e+00 1.60694134e+00
4.13355023e-01 2.83717901e-01 6.89142942e-02 1.17472649e+00
7.50017405e-01 4.07382101e-01 -2.35106647e-01 1.77914083e-01
6.62531018e-01 -1.54162514e+00 -4.74174827e-01 -7.83769071e-01
8.69130731e-01 -6.80229366e-01 7.14524448e-01 1.63655058e-02
-9.64780152e-01 -6.16767704e-01 -1.42757201e+00 -2.52077520e-01
3.43718641e-02 2.83278793e-01 3.96840692e-01 3.05386454e-01
-1.36753714e+00 9.25826788e-01 -9.28988218e-01 -5.35670221e-01
8.95081908e-02 3.98468882e-01 -3.68545473e-01 -1.20488338e-01
-9.32559431e-01 1.25463414e+00 4.24748093e-01 -1.10264659e-01
-1.05551279e+00 -5.24163961e-01 -1.47708547e+00 -4.15007502e-01
2.48464912e-01 -1.21062148e+00 1.41316903e+00 -7.39106297e-01
-1.75393724e+00 1.05204570e+00 -4.33599114e-01 -6.79521561e-01
8.11352730e-01 -4.83062595e-01 2.33529545e-02 1.36669755e-01
4.68237221e-01 1.22266901e+00 9.12349641e-01 -1.26336586e+00
-6.53013408e-01 -1.23284288e-01 -3.43117356e-01 6.43532455e-01
3.69090974e-01 -5.01080215e-01 -8.26696277e-01 -2.24217162e-01
3.63527566e-01 -1.45302582e+00 -3.85361582e-01 2.04062551e-01
-3.85934591e-01 1.28883839e-01 9.09586668e-01 -3.63720179e-01
8.40520442e-01 -1.94774723e+00 4.75921631e-01 -1.53482720e-01
1.19893756e-02 2.74744630e-01 9.06270444e-02 1.36290908e-01
1.71367571e-01 -5.48395276e-01 -2.07653999e-01 -9.38203573e-01
-5.82008325e-02 4.05934840e-01 -6.73765270e-03 7.53052890e-01
5.75540662e-02 1.12198758e+00 -1.08809197e+00 -4.61234719e-01
7.33230233e-01 2.72936940e-01 -6.29429519e-01 2.88613915e-01
-3.56311560e-01 5.70479274e-01 4.97331768e-02 4.59347546e-01
5.56507707e-01 -2.47956499e-01 3.28702442e-02 -9.85566410e-04
-8.60262960e-02 6.36969268e-01 -1.27598238e+00 2.23025751e+00
-2.46598512e-01 8.68604481e-01 -3.29497576e-01 -3.59972388e-01
9.10603702e-01 -8.00532475e-02 5.04877508e-01 -3.96691322e-01
1.08257100e-01 3.48873079e-01 -4.81606781e-01 -2.35619709e-01
8.30254376e-01 1.57692358e-02 9.40972120e-02 3.73276472e-02
3.46972942e-01 -4.61444706e-01 -1.17246188e-01 1.09070264e-01
8.07530642e-01 1.00081110e+00 5.09016156e-01 -1.81578279e-01
6.45877719e-01 5.29693723e-01 8.85927916e-01 4.22262698e-01
-5.59122562e-01 8.71057451e-01 2.08348297e-02 -6.05802536e-01
-1.22183526e+00 -9.51950073e-01 9.56154689e-02 2.40919724e-01
5.67059159e-01 -2.30955109e-01 -4.76163447e-01 -7.09227681e-01
2.38302901e-01 3.49276662e-01 -2.89994240e-01 -3.29813927e-01
-4.85595852e-01 -4.07140404e-01 2.81883210e-01 5.90609133e-01
6.91583991e-01 -9.47371244e-01 -5.91565907e-01 3.05908233e-01
-3.12428355e-01 -1.46667123e+00 -5.79720080e-01 9.08310562e-02
-9.81214225e-01 -9.37075973e-01 -6.20599687e-01 -8.97861004e-01
4.06093836e-01 5.44962764e-01 1.22400904e+00 -3.59814197e-01
-2.63667852e-02 3.62519324e-01 5.56157529e-02 -6.47043735e-02
-4.85002518e-01 1.81013405e-01 3.06224555e-01 -1.51529998e-01
5.48487365e-01 -4.10189211e-01 -7.24027157e-01 4.45708156e-01
-4.03121978e-01 1.40660936e-02 2.80541539e-01 7.50349164e-01
6.28012955e-01 -6.29169583e-01 -5.07615842e-02 -4.77269620e-01
-2.19043896e-01 -3.88972193e-01 -8.84343266e-01 -2.77612537e-01
-8.31949770e-01 5.06215036e-01 1.03658989e-01 -1.56481922e-01
-9.85385180e-01 6.28967762e-01 1.43240094e-02 -8.51626992e-01
7.65404701e-02 1.61830992e-01 -8.75189621e-03 -6.38351798e-01
6.86775744e-01 1.13545790e-01 1.40097558e-01 -2.71750987e-01
5.77159405e-01 4.86371547e-01 8.34664404e-01 1.81290787e-02
1.12967789e+00 8.32870841e-01 3.10300231e-01 -7.89424717e-01
-9.69740152e-01 -1.01159477e+00 -8.58941257e-01 -1.67203829e-01
9.13015902e-01 -1.52710760e+00 -4.77286220e-01 6.41261756e-01
-1.18882227e+00 -5.63285470e-01 -3.10793281e-01 7.56215930e-01
-1.06603074e+00 8.16561341e-01 -6.08735144e-01 -5.79075336e-01
-2.80606627e-01 -1.12734818e+00 1.28974020e+00 6.49076477e-02
-2.81638682e-01 -1.14743567e+00 6.55777276e-01 3.24015260e-01
-3.97397205e-02 2.28077009e-01 1.58221319e-01 1.90473422e-02
-1.08662117e+00 7.41891563e-02 5.11525432e-03 3.83922309e-01
-1.62465125e-01 -3.40803266e-01 -9.90396321e-01 -4.27814543e-01
-1.80290505e-01 -4.71101850e-01 1.05199325e+00 5.43634832e-01
2.24810109e-01 1.11148745e-01 -3.01523119e-01 1.22208333e+00
1.62171447e+00 -1.46486700e-01 8.79482090e-01 8.98855209e-01
1.24761939e+00 3.32444668e-01 8.96910191e-01 1.92208543e-01
8.54757965e-01 1.21797550e+00 7.39149570e-01 1.17704131e-01
-2.75388867e-01 -6.87119305e-01 7.38491178e-01 8.44287694e-01
2.59200901e-01 1.48955643e-01 -7.61971533e-01 8.76573741e-01
-2.25936222e+00 -7.08374679e-01 -1.70940638e-01 2.14739013e+00
5.45362771e-01 9.65408385e-02 -2.22395752e-02 -3.47120762e-02
3.45238388e-01 5.66922128e-01 -6.06480837e-01 -3.30998808e-01
-1.51241243e-01 -3.61636132e-01 8.61386001e-01 8.57050955e-01
-1.49359822e+00 1.42235434e+00 6.39593172e+00 2.86456645e-01
-1.10696256e+00 -4.84937355e-02 -5.80101982e-02 -6.93404749e-02
1.71924353e-01 2.45973751e-01 -1.22334528e+00 1.55764848e-01
8.32947731e-01 -5.09217568e-02 2.66930223e-01 1.28285635e+00
2.07631782e-01 -2.25083351e-01 -1.14703751e+00 1.14374149e+00
3.16835225e-01 -1.29867053e+00 -9.34674591e-02 5.75004369e-02
1.13196242e+00 6.32632971e-01 -2.39633486e-01 1.14432447e-01
4.58976179e-01 -6.72907293e-01 8.67628098e-01 3.24506879e-01
6.10015452e-01 -4.79847997e-01 7.57592261e-01 3.74930739e-01
-1.37840080e+00 1.22834496e-01 -6.83477402e-01 -1.99071422e-01
3.94283235e-01 4.22622144e-01 -1.11106455e+00 8.59498978e-01
5.76934159e-01 1.22631979e+00 -4.39137816e-01 1.32687783e+00
-5.75726271e-01 -4.80877970e-05 -2.86903560e-01 2.83390284e-01
4.99313712e-01 -9.16552991e-02 1.04726195e+00 7.74799168e-01
2.01210082e-01 -6.12242460e-01 3.22462440e-01 6.97654426e-01
7.05651119e-02 -1.72717586e-01 -9.94805574e-01 5.11631072e-01
1.83175609e-01 8.39266658e-01 -3.13160010e-03 -3.18501294e-01
-3.96468729e-01 1.33768821e+00 5.59860229e-01 1.51364937e-01
-5.03131449e-01 -1.22841792e-02 1.07667637e+00 -3.73096881e-03
5.68078220e-01 -4.18636262e-01 -2.26332054e-01 -1.49623299e+00
8.11799243e-02 -3.86438876e-01 2.23977014e-01 -1.03916752e+00
-1.00770724e+00 4.83850718e-01 -1.94680035e-01 -1.76694942e+00
-7.53458202e-01 -5.86658835e-01 -2.52919018e-01 7.45982707e-01
-1.91123712e+00 -9.65017021e-01 -5.82810462e-01 3.88551086e-01
4.64277059e-01 -1.42030224e-01 4.50821310e-01 1.47202805e-01
-1.88559666e-01 2.71513164e-01 2.10331064e-02 -4.26554121e-02
7.85433173e-01 -1.35124326e+00 9.71994460e-01 1.09216225e+00
2.78055489e-01 1.11613683e-01 7.96082139e-01 -7.25172460e-01
-1.30108559e+00 -1.43817842e+00 1.24590397e+00 -9.62954283e-01
4.94947612e-01 -6.36128113e-02 -5.94972789e-01 9.89514828e-01
-8.48284736e-02 3.25661957e-01 -1.30684927e-01 -5.15986204e-01
-1.47534728e-01 7.32819289e-02 -9.74004626e-01 7.66709447e-01
1.30313158e+00 -5.13560832e-01 -7.89084852e-01 9.11524221e-02
8.50809872e-01 -9.16570008e-01 -6.03741944e-01 4.62062627e-01
3.56784403e-01 -1.09265447e+00 1.05813432e+00 -3.33218783e-01
3.36315304e-01 -5.65607250e-01 -1.43985124e-02 -1.49544764e+00
-2.67277151e-01 -8.46410751e-01 -3.81916612e-01 6.87951028e-01
1.17440701e-01 -3.77083600e-01 1.23892200e+00 2.16465294e-01
-1.75514251e-01 -4.33679253e-01 -1.06469119e+00 -1.09839070e+00
-2.49551296e-01 -3.27980787e-01 1.81575224e-01 7.46957600e-01
-1.32480115e-01 3.95487577e-01 -1.02654910e+00 3.01665336e-01
1.05848682e+00 -4.34139408e-02 1.24794948e+00 -1.08597934e+00
-3.22314113e-01 1.08351640e-01 -1.07780862e+00 -1.72332716e+00
2.26081789e-01 -8.45968604e-01 3.73430967e-01 -1.59928107e+00
-2.34825239e-02 7.21962154e-02 2.10684761e-01 1.01949826e-01
-1.13730334e-01 2.54146099e-01 4.00132120e-01 4.09702629e-01
-6.50551319e-01 9.35797095e-01 1.26040268e+00 9.04926378e-03
-6.35193229e-01 -1.26028523e-01 -3.52714807e-02 9.67604160e-01
4.64248091e-01 -5.72573900e-01 -3.91205937e-01 -4.14613992e-01
-7.80881867e-02 1.10154934e-02 4.77161199e-01 -1.39999175e+00
4.60360140e-01 7.32422918e-02 1.27119049e-01 -9.77826059e-01
4.41049725e-01 -7.44629204e-01 1.09885328e-01 7.20298350e-01
2.22238749e-01 -5.82595579e-02 7.03310370e-02 6.48644388e-01
-2.85841286e-01 -1.96887404e-02 9.34367120e-01 -4.94470298e-02
-1.40369785e+00 4.05150980e-01 -1.73166126e-01 1.55304551e-01
9.19878721e-01 -2.95657277e-01 -3.09701741e-01 -6.87549353e-01
-6.23444438e-01 5.21302819e-01 9.19215798e-01 6.03744268e-01
6.90239131e-01 -1.61721575e+00 -2.64203995e-01 1.76398963e-01
4.65812504e-01 2.46012032e-01 -8.45120028e-02 6.69349074e-01
-9.30134237e-01 8.12214255e-01 -2.39605919e-01 -1.06219292e+00
-1.07102954e+00 4.94357139e-01 6.02515996e-01 -2.67634779e-01
-8.33036900e-01 5.64393103e-01 -2.36765761e-02 -7.46780157e-01
3.81408751e-01 -3.51208329e-01 -3.28825191e-02 -4.83700663e-01
1.62205532e-01 4.38077301e-01 -1.21208064e-01 -9.79310095e-01
-5.09708166e-01 9.25828815e-01 3.41860622e-01 -2.47846022e-01
1.10571349e+00 -5.22704661e-01 2.22137287e-01 4.90363181e-01
1.31410062e+00 -2.77680308e-01 -1.88639927e+00 -5.28777599e-01
2.52578080e-01 -4.55413491e-01 1.32317364e-01 -2.90177673e-01
-8.50661457e-01 6.85570657e-01 5.21705091e-01 -6.56778276e-01
6.37765229e-01 -1.56072453e-01 8.77838075e-01 4.58917320e-01
7.91795492e-01 -9.53982115e-01 1.45493731e-01 9.21488881e-01
5.54807842e-01 -1.47212768e+00 1.93244427e-01 -5.32902777e-01
-8.79645884e-01 7.17693031e-01 7.11135507e-01 -3.64820153e-01
4.52295691e-01 -2.28587419e-01 5.11928260e-01 -2.08698690e-01
-5.22710919e-01 -4.52289194e-01 5.78045845e-01 6.22273326e-01
-7.28123561e-02 -2.54776865e-01 -4.29616831e-02 -1.66534260e-01
-1.55695274e-01 2.33265638e-01 5.44584155e-01 1.04127777e+00
-4.19770509e-01 -1.08074749e+00 -2.45209560e-01 -1.87136084e-01
8.52022469e-02 9.99007225e-02 -3.72678995e-01 8.05404902e-01
-1.58485979e-01 6.42188013e-01 -7.40812123e-02 -6.63011193e-01
4.23963219e-01 9.33932737e-02 4.22631502e-01 -7.35682905e-01
-1.18236467e-01 -1.15678474e-01 2.22738340e-01 -1.09998858e+00
-6.93150699e-01 -6.80434942e-01 -1.13594830e+00 -2.77456939e-01
6.60574660e-02 -3.92753631e-01 6.35248125e-01 9.49360311e-01
4.60879415e-01 1.15171462e-01 7.00219393e-01 -1.43026817e+00
-6.13150001e-01 -7.27887034e-01 -3.77457052e-01 4.29203689e-01
5.36502004e-01 -8.79788399e-01 -4.74356681e-01 -7.92777985e-02] | [8.150262832641602, -2.131906747817993] |
301e5a7b-f865-4a2b-a18c-86fc385444ed | working-with-a-small-dataset-semi-supervised | null | null | https://aclanthology.org/W13-4901 | https://aclanthology.org/W13-4901.pdf | Working with a small dataset - semi-supervised dependency parsing for Irish | null | ['Mark Dras', 'Jennifer Foster', 'Teresa Lynn'] | 2013-10-01 | null | null | null | ws-2013-10 | ['transition-based-dependency-parsing'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.462052822113037, 3.651705026626587] |
c0fbc35d-62c6-44d4-b1fa-cca2a989f44f | tag-boosting-text-vqa-via-text-aware-visual | 2208.01813 | null | https://arxiv.org/abs/2208.01813v3 | https://arxiv.org/pdf/2208.01813v3.pdf | TAG: Boosting Text-VQA via Text-aware Visual Question-answer Generation | Text-VQA aims at answering questions that require understanding the textual cues in an image. Despite the great progress of existing Text-VQA methods, their performance suffers from insufficient human-labeled question-answer (QA) pairs. However, we observe that, in general, the scene text is not fully exploited in the existing datasets -- only a small portion of the text in each image participates in the annotated QA activities. This results in a huge waste of useful information. To address this deficiency, we develop a new method to generate high-quality and diverse QA pairs by explicitly utilizing the existing rich text available in the scene context of each image. Specifically, we propose, TAG, a text-aware visual question-answer generation architecture that learns to produce meaningful, and accurate QA samples using a multimodal transformer. The architecture exploits underexplored scene text information and enhances scene understanding of Text-VQA models by combining the generated QA pairs with the initial training data. Extensive experimental results on two well-known Text-VQA benchmarks (TextVQA and ST-VQA) demonstrate that our proposed TAG effectively enlarges the training data that helps improve the Text-VQA performance without extra labeling effort. Moreover, our model outperforms state-of-the-art approaches that are pre-trained with extra large-scale data. Code is available at https://github.com/HenryJunW/TAG. | ['Larry S. Davis', 'Joseph F. JaJa', 'ran Xu', 'Chetan Ramaiah', 'Ramprasaath R. Selvaraju', 'Yuqian Hu', 'Mingfei Gao', 'Jun Wang'] | 2022-08-03 | null | null | null | null | ['question-answer-generation'] | ['natural-language-processing'] | [ 3.06909889e-01 2.53452267e-02 1.80551112e-01 -6.13227129e-01
-1.53993189e+00 -7.62445450e-01 4.75522041e-01 -5.73991388e-02
-6.02963641e-02 4.79129106e-01 4.35364574e-01 -3.96495491e-01
4.22004342e-01 -7.32476175e-01 -7.81688750e-01 -5.33575118e-01
8.77214372e-01 6.98109090e-01 4.87964213e-01 -4.73356783e-01
6.60577118e-02 -1.47956952e-01 -1.47363281e+00 8.41220379e-01
1.31220067e+00 1.11291099e+00 5.13268054e-01 5.57976902e-01
-6.01173103e-01 1.22910631e+00 -5.61284602e-01 -7.72551119e-01
1.61141992e-01 -8.36472094e-01 -9.67321157e-01 3.34320605e-01
1.05987239e+00 -6.39202952e-01 -4.10162330e-01 7.86152482e-01
3.93037081e-01 1.09869577e-01 5.77609062e-01 -1.22207475e+00
-8.26934934e-01 1.57008231e-01 -4.41951722e-01 -1.16444908e-01
5.00103652e-01 5.72803617e-01 1.24776971e+00 -1.11346197e+00
6.62785828e-01 1.30696714e+00 2.49693230e-01 6.85799837e-01
-1.01530910e+00 -4.51841980e-01 3.68822962e-02 3.71356964e-01
-1.17515945e+00 -4.87401813e-01 9.76355255e-01 -2.89137155e-01
5.99716246e-01 3.12714636e-01 4.55946505e-01 1.26331854e+00
9.03737359e-03 1.34869194e+00 9.93580520e-01 -2.89811611e-01
1.37469238e-02 -1.36856511e-02 -2.10150704e-02 9.84206617e-01
-2.04094142e-01 -3.73750836e-01 -5.74684799e-01 1.86564382e-02
5.25758684e-01 -8.48375261e-02 -5.71210742e-01 -6.02785885e-01
-1.21761155e+00 9.35834229e-01 5.60355484e-01 4.91896458e-02
-2.48285189e-01 1.82117119e-01 4.18172836e-01 1.75174236e-01
2.18080506e-01 3.70171130e-01 -3.09817612e-01 -6.69078156e-02
-7.62602508e-01 1.28489897e-01 3.31973404e-01 1.05891633e+00
9.57522273e-01 7.11625814e-02 -6.06472969e-01 6.74832642e-01
2.43105516e-01 9.51167226e-01 1.34625256e-01 -1.21165693e+00
1.06414711e+00 1.04412282e+00 1.98178515e-01 -7.95587540e-01
7.94683695e-02 -2.01896355e-01 -7.02628613e-01 -1.50237292e-01
7.08815038e-01 2.07794815e-01 -1.27846098e+00 1.50673318e+00
4.48405117e-01 -3.14290732e-01 1.63701370e-01 1.08494949e+00
1.19972122e+00 1.06151533e+00 8.14548284e-02 9.54064429e-02
1.48515010e+00 -1.55676889e+00 -9.26240504e-01 -4.59311068e-01
4.23781514e-01 -8.62992585e-01 1.72782874e+00 2.71693796e-01
-1.17654324e+00 -6.35378122e-01 -6.88359976e-01 -6.37981117e-01
-8.00391957e-02 3.32676888e-01 4.04389054e-01 5.53242028e-01
-1.09432828e+00 -2.96606749e-01 -4.81497645e-01 -1.99115068e-01
7.14070797e-01 -7.84288049e-02 -2.42132664e-01 -7.97029793e-01
-9.40116048e-01 5.06246686e-01 -6.27920851e-02 1.52646244e-01
-1.51178038e+00 -5.72622299e-01 -1.13419604e+00 2.77504250e-02
6.78782463e-01 -1.13371050e+00 1.45416498e+00 -1.22010291e+00
-1.26511908e+00 6.87012911e-01 -6.75487161e-01 -2.66074017e-02
5.15760958e-01 -1.96890384e-01 -3.63141559e-02 7.51686752e-01
3.04353923e-01 9.18028533e-01 9.58704531e-01 -1.78899002e+00
-4.66534823e-01 -5.12902021e-01 2.89712220e-01 4.15218323e-01
-1.29673481e-01 -3.05388898e-01 -8.40981007e-01 -3.70125175e-01
6.20947871e-03 -4.99579102e-01 -2.91443497e-01 2.90009439e-01
-2.82004684e-01 -2.28606597e-01 6.91119015e-01 -7.88905680e-01
8.31589639e-01 -1.92008841e+00 1.69043206e-02 -2.49957830e-01
2.75014549e-01 3.19838673e-01 -5.37136018e-01 6.80910230e-01
4.92437154e-01 -2.62406081e-01 -4.31161761e-01 -5.57603300e-01
-1.05164982e-02 4.71706539e-01 -8.15752625e-01 1.21601120e-01
3.07790011e-01 1.50722671e+00 -1.12494409e+00 -8.35044444e-01
3.32010686e-01 2.92474389e-01 -3.41570765e-01 5.83892822e-01
-8.09184432e-01 6.57167196e-01 -7.19485760e-01 8.57385933e-01
6.79062247e-01 -6.20455801e-01 -1.00856878e-01 -3.79850268e-01
2.79459953e-01 1.94157705e-01 -5.67927599e-01 2.07114649e+00
-4.03244078e-01 5.96921623e-01 -1.57339573e-02 -9.14156199e-01
7.40860939e-01 2.18823954e-01 1.76445127e-01 -1.25209057e+00
2.48762295e-01 1.51321575e-01 -4.73156393e-01 -7.64768600e-01
6.16191626e-01 1.51901677e-01 -1.47239417e-02 3.69897425e-01
2.14021519e-01 -4.97314245e-01 2.69349426e-01 7.37867594e-01
8.80568385e-01 1.30897194e-01 2.73204967e-02 2.64058888e-01
6.59357786e-01 4.06377673e-01 2.20357001e-01 8.01733553e-01
-3.25036407e-01 8.49973202e-01 3.94662529e-01 -1.37628064e-01
-1.09854400e+00 -1.35213125e+00 2.65414089e-01 1.08860803e+00
4.07488883e-01 -4.16519970e-01 -7.70084679e-01 -1.04939604e+00
-3.65933269e-01 7.79490530e-01 -7.32486963e-01 1.03468962e-01
-3.43091369e-01 -2.03936584e-02 5.53611815e-01 4.75058138e-01
6.55600905e-01 -1.12752330e+00 -4.24061239e-01 -1.96966261e-01
-1.02157390e+00 -1.29490292e+00 -6.77329183e-01 -3.96115035e-01
-7.62546778e-01 -1.06382132e+00 -7.60782897e-01 -8.62601697e-01
8.73745203e-01 7.30234444e-01 1.57434022e+00 1.06379032e-01
-1.45799935e-01 8.28139424e-01 -6.68394446e-01 -1.67218998e-01
-4.06783462e-01 -2.76975483e-01 -8.02004576e-01 1.29873216e-01
1.97650671e-01 7.09866465e-04 -9.28728342e-01 4.34631944e-01
-1.01220179e+00 3.61905426e-01 6.97991252e-01 9.96841550e-01
6.98481858e-01 -4.67908174e-01 5.10166407e-01 -6.55826330e-01
2.62373984e-01 -2.86340564e-01 -4.71438050e-01 6.00089669e-01
-2.04311952e-01 7.62799084e-02 5.59715569e-01 -1.98547795e-01
-1.39446747e+00 7.13378713e-02 -2.88442194e-01 -4.24899578e-01
-1.19458333e-01 1.78511694e-01 -4.22078431e-01 1.19137205e-01
5.72876215e-01 6.25681221e-01 -1.56772599e-01 -5.45671061e-02
8.30609441e-01 5.35500646e-01 5.87832034e-01 -6.67135775e-01
7.88112879e-01 8.06079149e-01 -2.77584493e-01 -5.80834508e-01
-1.27588379e+00 -7.89351702e-01 -2.90875971e-01 -3.30881536e-01
1.08712280e+00 -1.01713550e+00 -6.16479337e-01 3.25973928e-01
-1.33417618e+00 -5.42261183e-01 -3.27771664e-01 -4.38348912e-02
-6.00911021e-01 7.00548589e-01 -3.62952471e-01 -8.28274310e-01
-6.78223729e-01 -1.27605200e+00 1.55339539e+00 3.83929461e-02
3.35393995e-01 -7.97239065e-01 -3.24286111e-02 1.35055709e+00
1.73215106e-01 -5.38443662e-02 8.05383801e-01 -2.86294580e-01
-1.12797678e+00 1.66284859e-01 -5.85179031e-01 3.06271583e-01
-2.72405408e-02 -4.19847816e-01 -1.06105208e+00 -1.20407857e-01
-3.24733071e-02 -9.86665070e-01 1.06604600e+00 1.33364618e-01
1.26289296e+00 -2.41189614e-01 4.48144861e-02 3.16532552e-01
1.41235864e+00 -1.56606119e-02 7.33465850e-01 -8.86617899e-02
1.05116224e+00 7.35134184e-01 9.70320582e-01 1.96679011e-01
9.63567734e-01 6.04544759e-01 8.72660100e-01 -3.33045810e-01
-4.86953110e-01 -4.88346219e-01 4.20802176e-01 7.42736697e-01
2.74163425e-01 -7.26297915e-01 -1.01790833e+00 1.02361465e+00
-1.87111223e+00 -9.12573934e-01 -6.06755197e-01 1.84674919e+00
9.23272789e-01 -4.72034991e-01 -1.95167124e-01 -1.37709543e-01
2.52753735e-01 2.63854355e-01 -4.98912275e-01 1.04584433e-02
-2.70123035e-01 1.99214563e-01 3.86571474e-02 4.86326486e-01
-7.93966293e-01 1.13611031e+00 5.28125334e+00 9.38596308e-01
-7.10951090e-01 2.92969435e-01 6.33676291e-01 4.06698622e-02
-8.12617123e-01 -7.16672093e-02 -4.34976488e-01 9.44705680e-02
4.58220482e-01 2.16202497e-01 2.34137371e-01 6.77700520e-01
1.46946818e-01 -2.73284167e-01 -9.13918197e-01 1.09617734e+00
5.25407016e-01 -1.34714973e+00 7.05504715e-01 -2.83266902e-01
9.63591039e-01 -3.96466404e-02 2.45709524e-01 3.59701037e-01
3.33976865e-01 -1.00599420e+00 7.68871427e-01 3.28755349e-01
8.56595099e-01 -5.25208116e-01 7.36273885e-01 1.99031919e-01
-1.27927446e+00 -1.19755231e-01 -4.29808527e-01 2.63122886e-01
4.34313327e-01 4.51549858e-01 -8.90945315e-01 8.96759987e-01
6.19015098e-01 5.58803618e-01 -9.48403656e-01 7.61064947e-01
-5.66684246e-01 7.67882586e-01 1.73427016e-01 5.55856675e-02
5.51227093e-01 -1.07880175e-01 2.31339991e-01 7.33839035e-01
2.47766390e-01 2.55313784e-01 1.72801360e-01 9.62757766e-01
-1.79085150e-01 2.99194485e-01 -4.65775907e-01 -1.52758462e-02
1.40745133e-01 1.32692730e+00 -2.60759622e-01 -4.52664018e-01
-6.38798892e-01 1.31579256e+00 3.51207614e-01 6.80227041e-01
-8.03859353e-01 -7.23115876e-02 1.84489831e-01 3.90127711e-02
3.69903296e-01 -9.05369818e-02 -2.68054605e-01 -1.40829229e+00
1.46231413e-01 -1.14132774e+00 6.29574955e-01 -1.48488367e+00
-1.40468848e+00 6.42725825e-01 -3.28652978e-01 -1.17263055e+00
-1.66202232e-01 -5.11330843e-01 -3.04473042e-01 7.63873041e-01
-1.67156518e+00 -1.83131433e+00 -8.98360491e-01 1.12290716e+00
1.05410051e+00 -1.01509653e-02 5.23086607e-01 1.85471982e-01
-5.39905392e-02 5.47008753e-01 -1.21087581e-01 1.30543023e-01
9.65446889e-01 -1.27295566e+00 3.80027652e-01 8.73795092e-01
4.24627513e-01 3.01351726e-01 5.22969961e-01 -5.19852877e-01
-1.54266822e+00 -1.19977701e+00 7.34063268e-01 -9.58914101e-01
4.93056178e-01 -5.20728350e-01 -1.01794231e+00 5.58961153e-01
6.35695696e-01 -9.60355327e-02 4.06399816e-01 -4.15377498e-01
-5.65282226e-01 -1.49528682e-01 -8.88224423e-01 7.36767411e-01
9.51847374e-01 -6.54431939e-01 -6.96193993e-01 2.93350816e-01
9.08658743e-01 -3.47591370e-01 -3.83186758e-01 3.37924421e-01
2.65839845e-01 -9.52404857e-01 1.06865954e+00 -3.35689127e-01
7.90402770e-01 -5.02782881e-01 -2.86914647e-01 -1.04022539e+00
2.00713679e-01 -2.81736374e-01 -5.02617843e-03 1.18881905e+00
2.54929036e-01 -2.12879464e-01 7.55223513e-01 3.75724912e-01
-4.19723749e-01 -6.74142718e-01 -8.06679070e-01 -2.41226986e-01
-7.19588697e-02 -3.66092414e-01 4.58617747e-01 9.18103337e-01
-3.90518069e-01 6.67914271e-01 -5.18469930e-01 1.34658575e-01
6.93732023e-01 3.65418464e-01 1.27778292e+00 -6.82907581e-01
-2.16641322e-01 -6.66581765e-02 -5.61444946e-02 -1.61933184e+00
-5.28762750e-02 -6.54456317e-01 2.52627641e-01 -2.19263697e+00
5.01302958e-01 -2.31505707e-01 2.15975672e-01 5.31917989e-01
-5.68567693e-01 4.39196438e-01 1.82249054e-01 6.06146120e-02
-1.18742943e+00 1.14414740e+00 1.88586342e+00 -4.94205475e-01
1.86368585e-01 -4.77046758e-01 -5.03638446e-01 4.12193567e-01
6.11946166e-01 -1.62530616e-01 -9.08849955e-01 -9.31331635e-01
2.72967696e-01 3.59155893e-01 6.48144424e-01 -6.43693864e-01
1.32737845e-01 -3.13389778e-01 2.97317177e-01 -1.00130260e+00
6.01119220e-01 -7.45103776e-01 -3.47209662e-01 9.93687138e-02
-2.24364862e-01 -8.53675306e-02 1.44913584e-01 6.30457997e-01
-4.70402390e-01 -1.23998113e-01 4.88188535e-01 -1.92203566e-01
-8.35088074e-01 4.21979785e-01 -1.33296967e-01 4.24787939e-01
7.88458943e-01 6.05432456e-03 -8.97816777e-01 -8.10136676e-01
-1.60468817e-01 6.50593996e-01 6.26044512e-01 3.60948980e-01
1.01721978e+00 -1.15375137e+00 -7.59740114e-01 -1.46095440e-01
7.29154468e-01 2.72462994e-01 8.32626045e-01 6.91087961e-01
-4.98513967e-01 4.74765778e-01 -3.29579227e-02 -8.24909747e-01
-1.20213401e+00 5.93018651e-01 3.18006814e-01 -3.18550408e-01
-3.71013463e-01 8.00134957e-01 8.91619921e-01 -4.68177855e-01
2.39423383e-02 -2.06337377e-01 -1.48127794e-01 -1.71154320e-01
4.86052454e-01 -7.69187808e-02 -1.74738556e-01 -7.27920651e-01
-8.18043873e-02 6.58866167e-01 5.87291047e-02 -2.86916316e-01
9.04745758e-01 -4.82503772e-01 -5.29018790e-02 2.97098696e-01
9.05836344e-01 1.37252331e-01 -1.28768373e+00 -3.37130070e-01
-5.26078105e-01 -7.05986440e-01 -2.30703667e-01 -9.75594640e-01
-1.01808333e+00 1.38632786e+00 3.74516487e-01 -2.65069276e-01
1.34952211e+00 4.23022389e-01 1.11854422e+00 5.59481859e-01
3.45719904e-01 -7.22328901e-01 9.28299367e-01 4.23724532e-01
1.10407829e+00 -1.57455802e+00 -2.37436533e-01 -5.01723289e-01
-1.19765794e+00 7.80729890e-01 9.44535136e-01 3.06536198e-01
-2.25290805e-01 -4.78264183e-01 5.47261953e-01 -4.25432950e-01
-8.39573324e-01 -4.93775010e-01 5.66250920e-01 6.98700786e-01
3.07871904e-02 -2.81114250e-01 5.87965399e-02 5.48757434e-01
1.10097565e-01 -2.94222862e-01 5.13639808e-01 6.32697284e-01
-4.87989962e-01 -1.02308273e+00 -4.06233639e-01 2.69578397e-01
-1.02436759e-01 -2.78950334e-01 -5.22102237e-01 6.04833305e-01
-8.85060802e-02 1.29378939e+00 -1.05019361e-01 -7.68035278e-02
2.49388754e-01 -1.53492600e-01 6.25482798e-01 -4.89758819e-01
-3.68704855e-01 -1.26600489e-01 1.53336138e-01 -8.55645418e-01
-4.00786102e-01 -3.97745460e-01 -1.14107466e+00 3.60357724e-02
-1.49525881e-01 1.02561504e-01 4.24636722e-01 8.86853456e-01
3.35276186e-01 4.92688537e-01 4.42309052e-01 -4.07684445e-01
-3.04414213e-01 -8.21551681e-01 -7.72423223e-02 7.41041064e-01
3.86946142e-01 -5.42245269e-01 -1.95080608e-01 3.49963516e-01] | [10.902643203735352, 1.5718796253204346] |
19cd9a63-d4a0-408a-be74-a2052deb5221 | qlib-an-ai-oriented-quantitative-investment | 2009.11189 | null | https://arxiv.org/abs/2009.11189v1 | https://arxiv.org/pdf/2009.11189v1.pdf | Qlib: An AI-oriented Quantitative Investment Platform | Quantitative investment aims to maximize the return and minimize the risk in a sequential trading period over a set of financial instruments. Recently, inspired by rapid development and great potential of AI technologies in generating remarkable innovation in quantitative investment, there has been increasing adoption of AI-driven workflow for quantitative research and practical investment. In the meantime of enriching the quantitative investment methodology, AI technologies have raised new challenges to the quantitative investment system. Particularly, the new learning paradigms for quantitative investment call for an infrastructure upgrade to accommodate the renovated workflow; moreover, the data-driven nature of AI technologies indeed indicates a requirement of the infrastructure with more powerful performance; additionally, there exist some unique challenges for applying AI technologies to solve different tasks in the financial scenarios. To address these challenges and bridge the gap between AI technologies and quantitative investment, we design and develop Qlib that aims to realize the potential, empower the research, and create the value of AI technologies in quantitative investment. | ['Tie-Yan Liu', 'Xiao Yang', 'Weiqing Liu', 'Jiang Bian', 'Dong Zhou'] | 2020-09-22 | null | null | null | null | ['stock-market-prediction'] | ['time-series'] | [-4.12748039e-01 7.01560453e-02 -2.80658202e-03 -3.90402317e-01
-1.72612444e-01 -5.56192279e-01 6.36960208e-01 -8.81617665e-02
-3.74677509e-01 4.29598898e-01 1.75470114e-02 -4.55164194e-01
-5.13331950e-01 -1.17463410e+00 -2.33745828e-01 -1.73124596e-01
-1.41120896e-01 7.40230560e-01 -3.41218114e-01 -3.56061220e-01
7.41085649e-01 7.54315317e-01 -1.31541431e+00 3.08251739e-01
7.46838808e-01 1.52772796e+00 -2.08371803e-01 5.81193455e-02
-6.00908577e-01 1.38943624e+00 -5.62282324e-01 -7.43294537e-01
6.00803137e-01 -1.32047072e-01 -7.65417457e-01 -2.35738128e-01
-1.33814633e-01 -3.26267660e-01 2.34006308e-02 7.87815571e-01
2.94765621e-01 -1.33727893e-01 4.25749809e-01 -1.60967255e+00
-4.82234985e-01 7.83592582e-01 -4.63778555e-01 3.10192198e-01
1.10530537e-02 5.21347582e-01 1.08530891e+00 -8.42643797e-01
3.45183492e-01 1.02914655e+00 4.76620048e-01 8.71674046e-02
-4.86305624e-01 -6.18722141e-01 1.52220652e-01 -6.09452650e-02
-5.74904859e-01 -4.59701456e-02 7.92916954e-01 -8.14138472e-01
1.05091298e+00 2.10060224e-01 1.28169036e+00 5.26737809e-01
2.65897572e-01 9.42511380e-01 1.04504192e+00 -4.36593801e-01
3.85438353e-01 3.99734706e-01 -1.38646856e-01 2.45833635e-01
5.15077114e-01 2.40584582e-01 -4.93263125e-01 4.27028358e-01
5.49934924e-01 4.07305479e-01 4.80364859e-01 -4.03647684e-02
-1.29425013e+00 8.23497236e-01 3.48026544e-01 7.28643715e-01
-6.61884546e-01 3.39775890e-01 3.07470381e-01 4.17430252e-01
4.09355879e-01 1.39634645e+00 -5.29311240e-01 -6.01004660e-01
-1.17316604e+00 3.54924411e-01 8.85365248e-01 6.25992537e-01
7.70640075e-01 3.97163033e-01 3.69732901e-02 -1.88076068e-02
1.88899711e-01 2.96693563e-01 4.29825902e-01 -1.34399903e+00
6.63228869e-01 1.16282856e+00 1.81142375e-01 -9.56699193e-01
-4.40957338e-01 -7.73903549e-01 -7.18357205e-01 3.74840707e-01
3.39859039e-01 -2.83473656e-02 -4.18255538e-01 1.02221859e+00
-2.10173111e-02 -4.52392012e-01 -7.67487064e-02 7.38893092e-01
2.61278421e-01 4.13881004e-01 -8.19205865e-02 -1.08961798e-01
8.62181962e-01 -9.87459779e-01 -7.67744780e-01 -1.35895878e-01
5.80108106e-01 -3.79228830e-01 1.02673447e+00 7.50152767e-01
-1.12105441e+00 -3.69907469e-01 -9.78496969e-01 3.56611967e-01
-7.57033408e-01 -3.59094262e-01 1.30070138e+00 6.20311737e-01
-7.15821028e-01 6.50015473e-01 -5.37010133e-01 3.66574615e-01
5.74496865e-01 3.15102577e-01 2.18939990e-01 2.86024183e-01
-1.34065592e+00 1.18745148e+00 5.53238928e-01 5.15782952e-01
-6.26133204e-01 -1.13759649e+00 -3.37929606e-01 2.81792432e-01
5.03456652e-01 -3.53417426e-01 1.19520950e+00 -1.24690366e+00
-1.54374444e+00 5.48316538e-01 9.77438092e-01 -7.86034703e-01
8.78720641e-01 -2.41673470e-01 -2.14538857e-01 -4.25585918e-02
-2.16777921e-01 5.43766558e-01 3.71764749e-01 -4.78033513e-01
-7.18738794e-01 -2.13850364e-01 1.89310506e-01 -1.17236958e-03
-8.37436497e-01 9.78900567e-02 1.69579655e-01 -5.36680281e-01
-3.16565722e-01 -5.76560676e-01 -4.90545928e-01 -3.01458538e-01
3.80683452e-01 -2.28548255e-02 5.42058110e-01 -5.37564754e-01
1.27539051e+00 -1.83750486e+00 -4.60362621e-02 2.50488162e-01
2.76814938e-01 2.64711410e-01 2.14484453e-01 6.20828211e-01
1.67991087e-01 4.82758999e-01 -8.45471248e-02 1.36645049e-01
2.76019901e-01 -1.37933716e-01 -5.41212916e-01 -1.46063372e-01
6.23918533e-01 1.18654692e+00 -1.00470328e+00 -3.95192862e-01
3.04183155e-01 -1.16550848e-01 -4.43567902e-01 4.12178755e-01
-3.71758252e-01 1.70531332e-01 -7.19500303e-01 1.03121889e+00
5.19489586e-01 -8.11498240e-02 9.53599513e-02 1.09737545e-01
-6.99481130e-01 -3.29277404e-02 -9.96369541e-01 1.24749005e+00
-5.13858616e-01 3.90342146e-01 -2.33645305e-01 -1.13396573e+00
1.15540504e+00 2.25414589e-01 9.53411102e-01 -1.38317049e+00
1.87921766e-02 5.29241204e-01 8.65301117e-02 -5.54922879e-01
6.12619996e-01 -3.26756150e-01 -3.49066943e-01 8.35924983e-01
-1.35138631e-01 -6.96751475e-01 4.93076235e-01 -4.50942479e-02
9.15580153e-01 3.17168713e-01 1.25718594e-01 -2.07110137e-01
2.02061623e-01 1.38026744e-01 6.80042326e-01 3.45581532e-01
-2.42067128e-01 9.20247659e-02 6.96814477e-01 -9.25137043e-01
-8.92490983e-01 -7.58561373e-01 1.34964600e-01 7.42300212e-01
-4.21916991e-01 1.15483299e-01 -4.91777569e-01 -6.84922695e-01
4.19227332e-01 4.76522982e-01 -5.02763987e-01 1.67471185e-01
-5.29143274e-01 -7.37847030e-01 4.31400806e-01 6.40387714e-01
7.57301629e-01 -1.44351113e+00 -1.11975622e+00 5.03381252e-01
2.04391405e-01 -8.53486717e-01 1.06639564e-01 -1.30125750e-02
-9.18676972e-01 -9.86269295e-01 -5.55215538e-01 -1.18471868e-01
3.45991105e-01 -4.71112505e-02 1.39284122e+00 -1.05976043e-02
-4.65327203e-01 1.84180468e-01 -1.55281112e-01 -1.08059657e+00
-1.32445037e-01 2.02337980e-01 -1.10757098e-01 -3.25580150e-01
7.48785511e-02 -2.16398939e-01 -4.95306611e-01 -8.22869316e-02
-9.14342403e-01 1.75432950e-01 7.26022542e-01 5.82223654e-01
9.68309715e-02 4.32724386e-01 1.18177414e+00 -7.47208357e-01
6.41197383e-01 -6.06888711e-01 -1.03334427e+00 4.43317860e-01
-1.27266133e+00 1.33197039e-01 5.11404693e-01 -1.95984393e-01
-1.22367096e+00 -4.54607546e-01 4.66646910e-01 -1.44375414e-01
5.76871336e-01 1.07101226e+00 -7.26127103e-02 2.57841945e-01
1.72789216e-01 -3.56477976e-01 2.09246367e-01 -2.93628424e-01
3.96695435e-01 5.60508549e-01 1.26125172e-01 -6.96242750e-01
6.58389390e-01 2.24915534e-01 2.48503655e-01 -1.74931413e-03
-8.37721407e-01 4.98663308e-03 -4.74943817e-01 -7.40926623e-01
5.45387149e-01 -5.51370203e-01 -9.90026534e-01 5.09123266e-01
-6.78280711e-01 -2.78875798e-01 -6.89852834e-01 4.51992720e-01
-4.05421197e-01 -1.31346151e-01 -3.29300940e-01 -1.08246899e+00
-6.44342065e-01 -1.06982481e+00 2.54038155e-01 4.69696909e-01
-1.96975902e-01 -1.08033204e+00 9.38494224e-03 5.23843646e-01
9.90115643e-01 5.94762802e-01 7.60950804e-01 -4.14529264e-01
-8.09101164e-01 -4.59351212e-01 -2.67879158e-01 6.44214809e-01
1.12974927e-01 4.74128515e-01 -7.43587673e-01 -5.41100604e-03
1.70352593e-01 -4.16145951e-01 4.37605083e-01 -1.70352951e-01
1.14497769e+00 -3.92310649e-01 2.97234386e-01 3.79718006e-01
1.30890775e+00 7.28856564e-01 5.12984872e-01 1.00400805e+00
4.16429758e-01 1.30855846e+00 1.21168005e+00 7.47512043e-01
3.27707887e-01 2.89258897e-01 4.01815653e-01 -1.06784269e-01
4.47330296e-01 -4.41298448e-02 1.10162437e-01 8.58104944e-01
-2.17830494e-01 1.04950182e-01 -1.38015604e+00 4.14923906e-01
-1.94148993e+00 -1.11242104e+00 2.20078126e-01 1.98501360e+00
7.49344289e-01 4.74117666e-01 1.78187117e-01 2.55878299e-01
1.07527405e-01 8.96573532e-03 -5.37899256e-01 -7.45913625e-01
1.35525381e-02 3.17630917e-01 3.74720156e-01 -2.25860223e-01
-6.60537004e-01 7.95693219e-01 6.61198664e+00 3.42097163e-01
-1.36932051e+00 -4.66614991e-01 9.67786312e-01 -1.25987470e-01
-7.15031266e-01 1.31244613e-02 -3.75414938e-01 5.81855536e-01
1.13041723e+00 -6.44850075e-01 4.93793696e-01 9.61736262e-01
1.49012834e-01 1.31232992e-01 -1.02560413e+00 3.55021626e-01
-4.80718791e-01 -1.81707478e+00 -1.07994922e-01 3.05178195e-01
8.31610084e-01 -3.09564769e-01 4.88620490e-01 4.20330614e-01
2.29656026e-01 -1.06527507e+00 1.12304556e+00 1.05065739e+00
4.20822471e-01 -1.03341043e+00 1.07308114e+00 2.13175297e-01
-9.79177892e-01 -5.82929492e-01 -1.81728780e-01 -5.50552130e-01
-1.51894301e-01 7.83134758e-01 -8.22044432e-01 5.77049673e-01
7.35979319e-01 7.41732419e-01 -3.93074632e-01 9.07195091e-01
-3.54819535e-03 2.61503100e-01 1.50653794e-01 -7.76746944e-02
4.55079079e-01 -6.47374392e-01 -2.19908938e-01 1.01720822e+00
2.48964757e-01 -2.57395923e-01 -2.14330390e-01 9.63818431e-01
-1.07941274e-02 2.91949399e-02 -5.67844510e-01 -1.11481190e+00
4.23599750e-01 1.41928911e+00 -6.95596516e-01 -1.92566901e-01
-3.66765589e-01 -1.59968957e-02 1.21338487e-01 3.91043201e-02
-5.98815143e-01 -4.48254317e-01 3.87056112e-01 2.06918806e-01
-1.19954750e-01 -2.53666520e-01 -9.78689730e-01 -9.17481661e-01
2.65570227e-02 -1.11517274e+00 2.61553496e-01 -4.60830957e-01
-1.05657256e+00 1.71794489e-01 -6.42700270e-02 -1.33207870e+00
-5.90775013e-01 -6.88621402e-01 -5.88805735e-01 6.26373768e-01
-1.59055746e+00 -1.08577943e+00 -3.07538837e-01 -9.13919061e-02
1.55181155e-01 -5.51277041e-01 6.33729517e-01 2.35237285e-01
-4.61946785e-01 1.42848432e-01 1.70210935e-02 4.41828817e-02
3.94984573e-01 -1.35047293e+00 4.29232508e-01 6.63645864e-01
-1.61964193e-01 5.38470805e-01 1.51125312e-01 -5.28783143e-01
-1.72108889e+00 -6.66850686e-01 7.38065720e-01 -4.52477932e-01
1.12120974e+00 -1.84845209e-01 -5.58329701e-01 5.98057866e-01
1.04624793e-01 -4.25842017e-01 5.70682585e-01 -6.53176755e-02
-3.04481775e-01 -7.29475021e-01 -1.31284714e+00 2.24004790e-01
4.60520566e-01 -3.62051457e-01 -6.36773646e-01 3.81744057e-02
9.06635284e-01 3.29409949e-02 -1.05770421e+00 4.25611317e-01
6.89015150e-01 -9.13331509e-01 8.77223134e-01 -4.69940424e-01
5.88256061e-01 1.09528862e-01 2.47958243e-01 -8.08923125e-01
-8.59230608e-02 -7.73519576e-01 -2.88254797e-01 1.41449571e+00
2.56389380e-01 -8.41409862e-01 7.72693694e-01 1.05303001e+00
-1.06287830e-01 -9.16455865e-01 -6.91079438e-01 -6.21522129e-01
2.70414382e-01 -2.97348768e-01 1.27364564e+00 1.01414382e+00
3.78040858e-02 -3.74849617e-01 -2.01295286e-01 -6.82456672e-01
6.22503042e-01 3.08804452e-01 6.60844207e-01 -1.47293735e+00
-2.58145243e-01 -9.05672073e-01 -3.02436799e-01 -8.16954896e-02
-8.64490271e-02 -5.67236602e-01 -3.76441211e-01 -1.41256380e+00
1.27464766e-02 -6.97316527e-01 -7.69508421e-01 4.33749408e-01
1.60741359e-01 1.08691745e-01 6.36391044e-01 3.20476800e-01
-3.91050547e-01 2.93633372e-01 1.30982077e+00 -3.05154860e-01
-1.12559877e-01 1.33273795e-01 -1.16027677e+00 6.18673980e-01
6.10104561e-01 -2.99276918e-01 -2.18289822e-01 -3.83766890e-01
1.12504387e+00 2.36747339e-01 -2.49833107e-01 -8.30947697e-01
7.81835988e-02 -7.40939260e-01 1.50103137e-01 -5.72734714e-01
-1.00547612e-01 -9.64793384e-01 3.23534936e-01 7.30666697e-01
-1.81986675e-01 2.59313345e-01 -1.12220004e-01 -1.34877518e-01
-5.48839211e-01 -1.98451579e-01 4.89507884e-01 -2.68008173e-01
-5.14563203e-01 2.53209025e-01 -2.78238475e-01 5.28186262e-02
1.15045345e+00 -2.40735471e-01 -3.55755508e-01 1.58413783e-01
-8.21561813e-02 6.63112819e-01 2.30440170e-01 3.81378978e-01
4.49773759e-01 -1.02489102e+00 -4.75959986e-01 -4.43431772e-02
-1.77031994e-01 1.93867460e-01 -2.64759008e-02 6.73659384e-01
-8.16695452e-01 6.10968053e-01 -8.38203549e-01 1.49222344e-01
-3.92427385e-01 5.07934332e-01 5.30844569e-01 -8.06274891e-01
-1.86328903e-01 7.03362823e-01 -1.62369445e-01 -2.14767024e-01
1.66507155e-01 -5.26680827e-01 -1.46066114e-01 7.73120463e-01
5.04722059e-01 7.95781434e-01 1.87538102e-01 2.15537608e-01
-3.42364520e-01 2.24023119e-01 6.52070567e-02 -4.13509786e-01
1.94596255e+00 4.07478482e-01 -4.08632696e-01 6.76100016e-01
4.82263684e-01 -1.16427489e-01 -1.33860862e+00 1.14874125e-01
8.75254393e-01 -5.81348419e-01 3.92760783e-02 -1.12129521e+00
-1.25763333e+00 1.16826952e+00 1.58505052e-01 4.87968236e-01
1.09771299e+00 -7.34099865e-01 6.30637944e-01 2.39443406e-01
6.75793886e-01 -1.50919008e+00 5.40486038e-01 3.34260374e-01
9.51446533e-01 -1.13384438e+00 3.25112998e-01 2.21204013e-01
-8.19421947e-01 1.48555112e+00 5.69485724e-01 -2.79516112e-02
4.36548412e-01 6.78351343e-01 6.69762343e-02 -3.71740699e-01
-6.05787337e-01 6.62739724e-02 1.45300671e-01 3.05568188e-01
6.36268675e-01 2.07951158e-01 -1.94027290e-01 7.65484035e-01
-2.35738501e-01 4.26154345e-01 4.57147509e-01 1.09890306e+00
-3.99261087e-01 -1.21205449e+00 -1.94276404e-02 3.87001634e-01
-6.01195276e-01 2.84176227e-02 -6.08393908e-01 1.00951648e+00
1.46163210e-01 6.10262990e-01 2.74602503e-01 -2.67244697e-01
4.41352278e-01 -2.00338379e-01 2.24601746e-01 -4.03532356e-01
-1.34400511e+00 -3.68788719e-01 -9.77025628e-02 -5.40101588e-01
-2.18622580e-01 -6.95623219e-01 -1.43005502e+00 -3.34061235e-01
6.49052626e-03 4.19968337e-01 1.02239156e+00 1.01480508e+00
1.83613732e-01 6.50942206e-01 1.06416917e+00 -5.37288547e-01
-9.08550024e-01 -8.29913378e-01 -4.90743935e-01 -6.90059140e-02
-3.15004289e-01 -4.86650378e-01 -4.21091914e-01 -4.14531738e-01] | [4.533057689666748, 4.063995838165283] |
44bb3de1-17ae-4e6b-9f69-2750f3dad64e | representation-learning-on-large-and-small | 1707.09873 | null | http://arxiv.org/abs/1707.09873v1 | http://arxiv.org/pdf/1707.09873v1.pdf | Representation Learning on Large and Small Data | Deep learning owes its success to three key factors: scale of data, enhanced
models to learn representations from data, and scale of computation. This book
chapter presented the importance of the data-driven approach to learn good
representations from both big data and small data. In terms of big data, it has
been widely accepted in the research community that the more data the better
for both representation and classification improvement. The question is then
how to learn representations from big data, and how to perform representation
learning when data is scarce. We addressed the first question by presenting CNN
model enhancements in the aspects of representation, optimization, and
generalization. To address the small data challenge, we showed transfer
representation learning to be effective. Transfer representation learning
transfers the learned representation from a source domain where abundant
training data is available to a target domain where training data is scarce.
Transfer representation learning gave the OM and melanoma diagnosis modules of
our XPRIZE Tricorder device (which finished $2^{nd}$ out of $310$ competing
teams) a significant boost in diagnosis accuracy. | ['Fu-Chieh Chang', 'Chuen-Kai Shie', 'Chun-Nan Chou', 'Jocelyn Chang', 'Edward Y. Chang'] | 2017-07-25 | null | null | null | null | ['melanoma-diagnosis'] | ['computer-vision'] | [ 1.34900033e-01 3.84226382e-01 -2.85614431e-01 -4.83290255e-01
-1.04478717e+00 -3.96966264e-02 1.19948640e-01 2.06634507e-01
-3.29483569e-01 7.72156596e-01 4.82337892e-01 -6.56396970e-02
-3.87579590e-01 -1.02123809e+00 -5.49957573e-01 -4.64451998e-01
6.21503070e-02 7.20039964e-01 -4.55477834e-01 -4.91940826e-01
1.29190624e-01 5.06448627e-01 -1.37629473e+00 7.78569102e-01
8.46581817e-01 1.22607887e+00 7.71090463e-02 3.73662949e-01
-1.31688744e-01 1.24817610e+00 -7.95109451e-01 -7.82380402e-02
2.82204956e-01 -1.93972006e-01 -8.23763847e-01 -1.86421216e-01
2.31478959e-01 -2.17392281e-01 -1.21927336e-01 4.21386212e-01
7.83170164e-01 -9.75890234e-02 6.64723217e-01 -8.87016952e-01
-8.73765707e-01 3.87265831e-01 -3.44581962e-01 3.95074815e-01
-9.16230902e-02 1.00319661e-01 8.27664256e-01 -7.70726562e-01
7.67748952e-01 9.31275725e-01 7.30116665e-01 7.65705466e-01
-8.31636012e-01 -6.42973542e-01 -1.24002151e-01 -1.32938549e-02
-1.04522300e+00 -2.96483129e-01 2.29596198e-01 -6.90236449e-01
1.28216779e+00 1.20323531e-01 7.55412400e-01 1.02168345e+00
3.61472011e-01 5.93826234e-01 9.00066853e-01 -2.31596068e-01
2.45830536e-01 3.20603460e-01 1.76848605e-01 3.93581390e-01
2.51341999e-01 3.03368986e-01 -3.93402159e-01 1.39383078e-01
6.75146639e-01 4.80446011e-01 -5.29833920e-02 5.41043021e-02
-8.25484812e-01 9.56480503e-01 8.50393593e-01 4.50240016e-01
-5.71965873e-01 2.68405765e-01 3.81076455e-01 6.59188151e-01
5.50247252e-01 9.65976238e-01 -6.81632936e-01 -3.98774177e-01
-5.92618465e-01 -1.40016172e-02 6.59032881e-01 7.74831235e-01
3.53867859e-01 3.19910884e-01 9.57262814e-02 9.95915294e-01
-2.21981511e-01 3.02446783e-01 8.30451667e-01 -5.30533433e-01
6.22215152e-01 9.26733315e-01 -3.23517203e-01 -8.26361120e-01
-8.36088657e-01 -8.12430918e-01 -9.85170066e-01 2.28746414e-01
3.00960422e-01 -4.31363732e-01 -9.75538790e-01 1.46883655e+00
-2.20035583e-01 -2.89353967e-01 -6.15902652e-04 8.43481600e-01
9.96720672e-01 5.84713697e-01 -1.10208519e-01 2.35825792e-01
1.09357774e+00 -8.42342257e-01 -4.40858871e-01 -4.69222158e-01
1.16769087e+00 -5.15304983e-01 8.29092801e-01 5.05010545e-01
-1.13802707e+00 -5.52743614e-01 -1.05245471e+00 -2.18429163e-01
-6.65270388e-01 6.00011200e-02 8.21247935e-01 4.31047559e-01
-1.03995585e+00 7.33167887e-01 -6.99045241e-01 -4.61025298e-01
8.57073665e-01 8.73935699e-01 -5.35202265e-01 -5.79947829e-01
-9.83557343e-01 1.11265302e+00 1.67252541e-01 -1.04328908e-01
-8.80331218e-01 -9.74207401e-01 -7.08118737e-01 1.42658785e-01
-3.80837657e-02 -6.38489366e-01 9.11883652e-01 -1.25357306e+00
-1.02951479e+00 8.90069723e-01 4.55397755e-01 -4.88042653e-01
1.32991374e-01 -1.31099299e-01 -4.03530926e-01 2.96501573e-02
-1.88546777e-01 6.81729734e-01 7.92423606e-01 -7.15573847e-01
-7.06639111e-01 -6.29020452e-01 -1.07871011e-01 1.28051698e-01
-6.30116224e-01 -2.39907011e-01 -1.23571403e-01 -4.95775074e-01
-2.12437823e-01 -8.28521013e-01 -2.96885610e-01 1.14030093e-02
1.11409262e-01 -1.01366729e-01 4.91267443e-01 -6.27203047e-01
8.64377797e-01 -2.23564124e+00 2.13755816e-01 2.49136373e-01
6.09016299e-01 1.89526007e-01 -2.80017376e-01 3.82425845e-01
-3.56256276e-01 1.70671836e-01 2.49961019e-01 1.81168362e-01
-5.25808275e-01 2.12220713e-01 -1.20993564e-02 2.62729466e-01
5.47164679e-01 1.03277946e+00 -7.88331509e-01 -1.27690852e-01
7.71893412e-02 7.69304812e-01 -7.32073963e-01 1.77279979e-01
1.48380339e-01 2.86966473e-01 -4.30864751e-01 6.99687839e-01
3.14441383e-01 -6.74195945e-01 -9.79432277e-03 -9.70943719e-02
2.02405170e-01 1.41246125e-01 -8.89884114e-01 1.70803547e+00
-7.06163347e-01 7.60970712e-01 1.03929356e-01 -1.23060465e+00
1.07461798e+00 2.86204189e-01 9.21960652e-01 -9.83902097e-01
2.65360922e-01 4.48025763e-01 4.69847709e-01 -5.79316497e-01
3.07911992e-01 -5.51081955e-01 1.61685735e-01 4.11177576e-01
2.62305409e-01 -3.28585893e-01 -2.74796009e-01 -1.09459748e-02
1.40585518e+00 -2.78609902e-01 4.71579023e-02 -2.13698551e-01
-2.14932352e-01 1.26281098e-01 4.75887299e-01 2.48347461e-01
9.04352888e-02 8.65626216e-01 4.55833584e-01 -8.02217960e-01
-1.03019762e+00 -5.96996427e-01 -1.58851549e-01 1.21123099e+00
-5.02816081e-01 -5.39089441e-01 -4.98224437e-01 -3.72681469e-01
1.45554453e-01 3.21825355e-01 -1.08643222e+00 -5.53272307e-01
-4.67729926e-01 -7.34999895e-01 -2.88265515e-02 9.93148386e-01
1.17177889e-01 -1.18690062e+00 -7.40191162e-01 7.86492676e-02
3.13355297e-01 -6.01530850e-01 -1.23222888e-01 6.00164115e-01
-1.19374979e+00 -8.50927889e-01 -9.05161560e-01 -7.99208760e-01
8.66793811e-01 -7.62548447e-02 1.17945600e+00 2.31021851e-01
-8.97064209e-01 2.36167803e-01 -3.62681091e-01 -8.40443730e-01
-1.52810693e-01 4.14692998e-01 -5.75077236e-02 -3.95385891e-01
4.41557318e-01 -2.21851856e-01 -5.43578148e-01 2.59589434e-01
-7.25732386e-01 -3.04087400e-02 9.68610227e-01 1.14824629e+00
4.86409515e-01 -9.95056033e-02 7.72801936e-01 -1.04045880e+00
8.50496709e-01 -6.73290491e-01 -9.47385803e-02 -7.15265349e-02
-8.03512275e-01 1.08262986e-01 4.78679866e-01 -2.86487460e-01
-4.50874895e-01 -1.89032882e-01 -2.63509274e-01 -4.57804263e-01
2.49782771e-01 5.50099671e-01 1.79572269e-01 -8.28094855e-02
1.08320594e+00 -3.58335316e-01 5.42978168e-01 -3.67964178e-01
2.46853992e-01 8.61149490e-01 1.54548325e-02 -5.92621326e-01
1.74901426e-01 2.77719200e-01 -7.33590052e-02 -8.15233767e-01
-5.69501817e-01 -3.03047895e-01 -6.25110149e-01 1.48523256e-01
8.62044275e-01 -9.86885011e-01 -4.44913983e-01 1.48201361e-01
-6.32550538e-01 -4.46397454e-01 -1.18258727e+00 5.74281514e-01
-5.44761598e-01 -5.63898265e-01 -5.17414331e-01 -3.31033200e-01
-6.79298401e-01 -1.26423907e+00 9.72739398e-01 2.58186132e-01
-3.25789750e-01 -1.05444014e+00 -8.12069327e-02 3.26514840e-01
9.65935469e-01 3.97197574e-01 1.05494547e+00 -7.87257731e-01
-2.33816639e-01 -8.19635332e-01 -3.96331489e-01 5.53673506e-01
4.25134808e-01 -4.47087407e-01 -1.09357607e+00 -4.37200904e-01
-7.04347482e-03 -7.80844688e-01 8.73530269e-01 5.13478339e-01
1.50941181e+00 -6.53286278e-02 -3.96379501e-01 8.50120783e-01
1.25992858e+00 1.53367355e-01 8.41657341e-01 2.18917593e-01
6.47022843e-01 7.08658516e-01 3.10611725e-01 3.89099687e-01
1.32978752e-01 4.40499067e-01 3.05380911e-01 -5.48980474e-01
-2.96770900e-01 -1.04120918e-01 -2.75119003e-02 8.73552859e-01
-2.43467763e-01 1.48812309e-01 -1.19785213e+00 4.89133716e-01
-1.50560021e+00 -7.00896144e-01 4.29474086e-01 1.94658124e+00
5.62021911e-01 1.78609669e-01 1.46155417e-01 5.68993390e-02
2.52513140e-01 -2.06416324e-01 -6.61863208e-01 -6.42198801e-01
-3.27521078e-02 3.93514067e-01 3.14657629e-01 3.88218127e-02
-5.79895616e-01 5.68002939e-01 6.56887722e+00 6.92843914e-01
-1.48718607e+00 1.35674834e-01 9.25531149e-01 -6.32146418e-01
-2.15263516e-01 -5.93767226e-01 -6.95263743e-01 2.46072263e-01
1.37355578e+00 -1.74186081e-02 2.13954210e-01 1.02521145e+00
-1.07299551e-01 3.65212530e-01 -1.41935289e+00 1.27208793e+00
2.49343902e-01 -1.78474164e+00 -5.91629334e-02 4.96269256e-01
6.76658332e-01 2.89655268e-01 2.45668039e-01 6.08247459e-01
1.99364752e-01 -1.85824502e+00 3.19511034e-02 4.94604170e-01
9.55299973e-01 -5.32288253e-01 1.10677278e+00 1.78260788e-01
-7.15030074e-01 -5.59631467e-01 -4.58926648e-01 -2.99344927e-01
-3.44235569e-01 2.63539881e-01 -1.14157474e+00 2.09646136e-01
6.57014608e-01 1.10970545e+00 -6.11818075e-01 8.15693319e-01
1.55660048e-01 4.50743288e-01 -1.97197676e-01 -5.60911298e-02
8.66337866e-02 6.59402013e-02 -1.54909149e-01 8.98291707e-01
3.63095313e-01 1.89724430e-01 -4.86362278e-02 6.05720818e-01
-2.13637173e-01 1.87487513e-01 -7.75254130e-01 -3.42574745e-01
5.69690131e-02 1.40952694e+00 -3.56152594e-01 -6.10674322e-02
-3.25601697e-01 4.30137455e-01 3.57018590e-01 -3.26525345e-02
-1.45281643e-01 -3.09675455e-01 5.62895298e-01 6.59015715e-01
8.96380097e-02 3.69446605e-01 -6.82845473e-01 -1.02097368e+00
-2.21452072e-01 -1.17374647e+00 7.44377732e-01 -7.76042104e-01
-1.37106609e+00 7.14876294e-01 -3.51963520e-01 -1.30179429e+00
-4.33505356e-01 -9.60754275e-01 -4.75247324e-01 9.43925977e-01
-1.26579773e+00 -1.31117392e+00 -4.50659871e-01 5.58383942e-01
6.03655100e-01 -6.14254057e-01 1.14180183e+00 4.99840975e-01
-2.79111236e-01 6.63682640e-01 3.08333129e-01 3.26635569e-01
5.89128733e-01 -1.15130222e+00 -6.71805441e-02 -4.49041948e-02
-2.69879133e-01 5.69686472e-01 2.60263801e-01 -3.11294049e-01
-1.57548583e+00 -8.17381501e-01 4.64715719e-01 -6.80440485e-01
3.63062233e-01 -2.03127787e-01 -7.58946598e-01 5.50954103e-01
-7.71208182e-02 3.53145212e-01 1.25675642e+00 4.94898230e-01
-2.21489280e-01 -5.86202085e-01 -1.39679694e+00 3.92940798e-04
7.37965286e-01 -4.16677982e-01 -4.27913487e-01 5.00833094e-01
5.19538879e-01 -3.98040831e-01 -1.26655638e+00 2.86710024e-01
6.40423894e-01 -4.78596836e-01 9.28077579e-01 -1.12469649e+00
9.32037771e-01 5.73648989e-01 -4.55157667e-01 -1.62447262e+00
-2.43379340e-01 -2.34565914e-01 -6.24261163e-02 8.23355138e-01
6.40321672e-01 -6.90227270e-01 8.03129971e-01 9.65542316e-01
-3.15651655e-01 -1.53755999e+00 -7.12629497e-01 -5.51129878e-01
7.76828825e-01 -3.02590489e-01 6.98756099e-01 7.67899156e-01
2.02194065e-01 3.85066062e-01 -2.52406418e-01 -3.86779457e-01
2.05141574e-01 1.05681643e-01 5.30622900e-01 -1.37848568e+00
-3.40908766e-01 -2.80879766e-01 -9.28191423e-01 -4.46533382e-01
-3.13360780e-01 -1.10356879e+00 -2.90718347e-01 -1.93530011e+00
-1.91739649e-02 -6.56718969e-01 -6.24589324e-01 6.26526117e-01
2.24547222e-01 1.01930149e-01 3.25626373e-01 2.53207058e-01
-3.13048989e-01 1.46709934e-01 1.36823988e+00 -3.69102150e-01
-2.18527123e-01 -1.27856266e-02 -1.34230733e+00 4.18547004e-01
7.11198986e-01 -2.65701383e-01 -4.35958296e-01 -7.42763877e-01
4.46847558e-01 3.08834091e-02 -2.03418225e-01 -1.09945846e+00
-7.63416570e-03 5.26579954e-02 9.68859911e-01 -1.99444711e-01
5.89711308e-01 -8.98044527e-01 -1.65325880e-01 7.24704981e-01
-5.39185047e-01 3.90254147e-02 4.75159317e-01 3.90751660e-02
-2.18751043e-01 -9.96219143e-02 8.96598816e-01 -2.97615111e-01
-6.55602992e-01 3.34753782e-01 4.07560123e-03 4.54270691e-01
1.09549212e+00 -1.45899937e-01 -4.93300319e-01 -4.46139008e-01
-8.42651069e-01 1.49424404e-01 2.20270410e-01 7.09396422e-01
8.60387206e-01 -1.26032174e+00 -7.98917353e-01 5.57938039e-01
1.00422390e-01 1.49806038e-01 1.56900644e-01 8.22467744e-01
-4.85610813e-01 4.49848354e-01 -6.44762158e-01 -5.17116547e-01
-9.60182607e-01 5.58940232e-01 4.77506310e-01 -2.42176235e-01
-6.18528426e-01 1.09444141e+00 1.71683952e-02 -5.55354834e-01
1.75384447e-01 -4.87982899e-01 -6.10326268e-02 3.03181112e-01
7.91719198e-01 4.37671155e-01 5.47287047e-01 -6.09492548e-02
-2.69610792e-01 7.24219441e-01 -5.29378414e-01 1.91327512e-01
1.97319114e+00 4.57357466e-01 2.10262239e-01 3.89275700e-01
1.43214977e+00 -5.37171721e-01 -1.05303037e+00 1.29154533e-01
-1.32622167e-01 -3.72441202e-01 1.83856383e-01 -1.11181200e+00
-1.40538442e+00 1.53746951e+00 1.02296638e+00 5.25860721e-03
1.04165745e+00 1.86020404e-01 6.18097126e-01 3.82633060e-01
3.73321831e-01 -1.34370339e+00 5.08658767e-01 4.04344738e-01
1.11147881e+00 -1.38828886e+00 1.15599588e-01 1.86827853e-01
-8.66056621e-01 1.09622705e+00 7.60496736e-01 -2.98732132e-01
9.02719736e-01 4.14754033e-01 2.76000425e-02 -5.41518688e-01
-7.75407672e-01 1.16530262e-01 2.41348803e-01 9.39556777e-01
7.17375219e-01 2.36290041e-02 1.72171518e-01 7.74089158e-01
-2.80773014e-01 5.32290280e-01 2.58742213e-01 1.01897156e+00
-4.08056080e-01 -9.93366897e-01 -1.98628113e-01 1.35944366e+00
-3.53299171e-01 -7.93928057e-02 -3.68935645e-01 7.40185857e-01
3.22840661e-01 6.91401422e-01 2.42475078e-01 -4.65621948e-01
4.88216758e-01 -1.41534388e-01 4.12482172e-01 -1.07177544e+00
-9.08792615e-01 -1.74654096e-01 -6.55607209e-02 -5.92268050e-01
2.37916976e-01 -4.10061538e-01 -1.30423379e+00 -2.00520605e-01
-1.81469217e-01 2.71028355e-02 9.43306804e-01 7.56376028e-01
6.59434497e-01 8.67560089e-01 4.68501389e-01 -6.93869233e-01
-7.80456901e-01 -1.00080824e+00 -7.68670976e-01 4.32326019e-01
4.61961329e-01 -4.17284697e-01 -7.69890845e-02 -2.52601236e-01] | [15.091412544250488, -2.324871301651001] |
8cdac840-c162-48b4-bba8-3d8fc15fcb6a | spell-checking-for-chinese | null | null | https://aclanthology.org/L12-1423 | https://aclanthology.org/L12-1423.pdf | Spell Checking for Chinese | This paper presents some novel results on Chinese spell checking. In this paper, a concise algorithm based on minimized-path segmentation is proposed to reduce the cost and suit the needs of current Chinese input systems. The proposed algorithm is actually derived from a simple assumption that spelling errors often make the number of segments larger. The experimental results are quite positive and implicitly verify the effectiveness of the proposed assumption. Finally, all approaches work together to output a result much better than the baseline with 12{\%} performance improvement. | ['Bao-liang Lu', 'Xiaolin Wang', 'Shaohua Yang', 'Hai Zhao'] | 2012-05-01 | null | null | null | lrec-2012-5 | ['chinese-spell-checking'] | ['natural-language-processing'] | [ 3.13883841e-01 -1.22056901e-01 -2.15596139e-01 -4.98146921e-01
-8.09000790e-01 -7.97698617e-01 1.64460987e-01 3.00235808e-01
-6.57855988e-01 7.34631956e-01 -7.51348883e-02 -9.54861939e-01
2.45689824e-01 -4.91256535e-01 -5.46259761e-01 -4.21854407e-01
2.17671975e-01 2.59893954e-01 6.72342539e-01 -4.55142468e-01
8.05360556e-01 4.66866910e-01 -9.24093127e-01 5.50140500e-01
1.31462514e+00 5.91944456e-01 3.67587626e-01 9.33678150e-01
-3.27999413e-01 5.09879827e-01 -8.19668949e-01 -6.59055173e-01
3.13115686e-01 -5.63444555e-01 -1.17585826e+00 -1.32503971e-01
1.23339914e-01 -2.24540859e-01 2.79919058e-02 1.33000731e+00
2.06856310e-01 2.46738195e-02 3.42757523e-01 -7.06926584e-01
-4.66813356e-01 1.06543398e+00 -4.07540590e-01 3.13451469e-01
6.23115838e-01 -3.76953870e-01 9.01392639e-01 -5.78710258e-01
5.65553248e-01 8.33314657e-01 6.00168169e-01 4.18542564e-01
-4.64626342e-01 -6.42334700e-01 4.96060938e-01 1.25702932e-01
-1.61330736e+00 -4.46219593e-02 2.13830248e-01 2.56749988e-01
1.18059587e+00 6.44387007e-01 3.96220356e-01 3.60578954e-01
1.78922731e-02 1.17460287e+00 9.86933529e-01 -1.05839825e+00
-1.29853263e-01 2.17673272e-01 6.79408431e-01 5.48317075e-01
5.09622753e-01 -2.62255371e-01 1.67193353e-01 4.35375534e-02
3.59140009e-01 -1.27247140e-01 -3.95429313e-01 1.35123283e-01
-9.57823873e-01 5.30023754e-01 9.95645002e-02 6.69259548e-01
4.52030510e-01 -3.93077880e-02 3.52959692e-01 1.88262403e-01
8.05925801e-02 5.01523376e-01 -5.48692763e-01 -4.36785996e-01
-1.31947494e+00 4.86843586e-01 9.08091068e-01 1.82367301e+00
3.57695818e-01 -4.93861362e-02 -1.12940681e-04 3.88381958e-01
2.24204764e-01 6.59179986e-01 4.29278940e-01 -5.71584463e-01
7.97210217e-01 6.78605378e-01 3.52327466e-01 -7.58064747e-01
-3.25577140e-01 -2.79343843e-01 -3.82522285e-01 -2.56619215e-01
6.22932076e-01 -1.67761400e-01 -9.18261528e-01 1.06303537e+00
-1.72775880e-01 -1.43516392e-01 -8.43439549e-02 6.64659083e-01
2.91696697e-01 8.75688374e-01 -7.55865872e-02 -1.78354725e-01
1.03838992e+00 -1.34304452e+00 -1.07485116e+00 9.73237380e-02
1.15192533e+00 -1.14450026e+00 1.12189960e+00 7.75005817e-01
-1.25984657e+00 -4.41218168e-01 -1.25604677e+00 1.70062631e-02
-5.67443609e-01 2.36738205e-01 5.73292911e-01 1.39798963e+00
-9.48390484e-01 6.39008701e-01 -7.01257944e-01 -1.46897212e-01
-1.74442440e-01 4.39689010e-01 -7.54643651e-03 1.56322405e-01
-1.13481748e+00 6.94179893e-01 9.45777774e-01 2.58419663e-01
-3.06510311e-02 -1.60345584e-01 -5.91314375e-01 6.38428777e-02
4.47410524e-01 1.59737095e-01 1.63359737e+00 -9.12233174e-01
-1.48273110e+00 5.20685196e-01 -4.95940804e-01 -6.72527909e-01
7.01456904e-01 -7.52120316e-01 -7.39831328e-01 1.66050792e-01
-2.80121893e-01 3.12248588e-01 2.13081241e-01 -1.03192639e+00
-9.48705018e-01 1.50720805e-01 -1.77736193e-01 5.25905676e-02
-1.35763973e-01 6.43310785e-01 -1.16274142e+00 -1.03450513e+00
1.45452339e-02 -1.13536131e+00 -4.85661626e-01 -6.36166394e-01
-5.50898194e-01 -5.13564609e-02 7.64219344e-01 -7.25686550e-01
2.23446107e+00 -1.92083681e+00 -4.02475178e-01 5.34924388e-01
-3.97175252e-01 9.60500062e-01 9.07749385e-02 6.52524650e-01
-4.67152819e-02 5.70007622e-01 -4.83168215e-01 -1.23154007e-01
-9.38313305e-02 1.19296592e-02 -5.09404898e-01 4.43557113e-01
7.96023458e-02 9.49577570e-01 -8.71757805e-01 -6.79576099e-01
-1.56942576e-01 -1.97082013e-01 -6.28560543e-01 -3.68595570e-02
-1.98704481e-01 -1.42973155e-01 -3.57280344e-01 6.76787078e-01
1.12235117e+00 4.02819887e-02 3.37791800e-01 4.65214252e-01
-5.30015767e-01 5.22229016e-01 -1.53512406e+00 1.67164195e+00
1.00350000e-01 5.20377696e-01 -2.64187038e-01 -5.67038178e-01
7.95307636e-01 5.71184605e-02 -2.71005958e-01 -5.86602747e-01
1.82155892e-01 4.68759030e-01 8.67260098e-02 -1.97848320e-01
1.11141360e+00 4.16360468e-01 -3.22664678e-01 5.28131068e-01
-5.05340636e-01 -2.86472589e-01 4.74067479e-01 4.95468855e-01
5.41159272e-01 3.60067993e-01 5.41073501e-01 -6.32672608e-01
7.38193870e-01 3.82228702e-01 5.11118889e-01 1.06432104e+00
-3.64580274e-01 7.00271845e-01 4.70932543e-01 -2.93203741e-01
-1.05850911e+00 -7.33036160e-01 -1.09990373e-01 7.73796499e-01
5.14315009e-01 -7.95080721e-01 -1.23602438e+00 -8.61133456e-01
-4.75748390e-01 7.10130870e-01 -2.68732905e-01 3.23672324e-01
-1.14556718e+00 -4.49670494e-01 1.21141469e+00 8.36365581e-01
5.61434031e-01 -8.53911638e-01 -5.37973166e-01 1.33729488e-01
-2.40647629e-01 -1.10927916e+00 -6.79615498e-01 1.41935036e-01
-9.38741028e-01 -7.26202071e-01 -7.42366672e-01 -1.27738547e+00
8.32258046e-01 6.09403193e-01 5.73216856e-01 8.22982728e-01
9.18284506e-02 -3.98673952e-01 -9.74592745e-01 -6.12654984e-01
-4.56200302e-01 4.14451271e-01 -2.98897684e-01 -6.10104620e-01
3.83221745e-01 2.71830440e-01 -3.04581314e-01 4.02972370e-01
-9.30281818e-01 7.24591091e-02 5.47923982e-01 6.22942626e-01
4.71142441e-01 7.22058415e-02 2.84513712e-01 -1.44405115e+00
6.13652945e-01 7.61444196e-02 -7.40895033e-01 5.80117285e-01
-9.38862622e-01 1.77376688e-01 8.85618806e-01 -2.68973440e-01
-1.29503906e+00 6.87761977e-02 -3.47060949e-01 3.96542519e-01
-2.63716638e-01 3.87352794e-01 -3.12898397e-01 -7.55251497e-02
2.99920291e-01 3.37504923e-01 -4.94881094e-01 -7.22197354e-01
3.78818154e-01 9.24837351e-01 5.67941546e-01 -4.02816623e-01
5.61638713e-01 2.40667835e-01 -5.25488377e-01 -5.13622761e-01
-4.27545488e-01 -7.20324695e-01 -9.00162578e-01 2.71457285e-01
3.97492975e-01 -7.14534521e-01 -4.68443155e-01 7.00356662e-01
-1.14283311e+00 -4.80804592e-01 1.73934400e-01 3.42208058e-01
-1.98729351e-01 1.20303833e+00 -8.95214200e-01 -8.03387463e-01
-4.84780848e-01 -1.14866972e+00 7.78757155e-01 2.77927339e-01
-2.63190359e-01 -7.94136167e-01 -1.71428233e-01 -1.32602647e-01
2.08439678e-01 -2.38599598e-01 5.76310217e-01 -9.27283823e-01
-4.49581861e-01 -4.78956550e-01 -3.69529158e-01 1.74353406e-01
-1.35528773e-01 4.03246075e-01 -6.41980171e-01 -1.37685046e-01
-2.39538088e-01 2.46882066e-01 7.13784993e-01 3.33184749e-02
1.24853659e+00 -3.25605720e-01 -4.85336721e-01 6.56371653e-01
1.53677166e+00 5.91043174e-01 8.46000969e-01 5.44594944e-01
4.49093163e-01 2.78512299e-01 1.08514905e+00 3.92784774e-01
2.09763423e-01 5.18527567e-01 6.66030049e-02 -7.31546804e-03
1.08103089e-01 -3.86678159e-01 3.73791963e-01 1.09006405e+00
-1.75645761e-02 -6.88646972e-01 -1.09318805e+00 5.20373642e-01
-1.71927285e+00 -8.28326106e-01 -8.00623477e-01 2.10768461e+00
9.10058081e-01 6.11489832e-01 8.87583662e-03 4.79948372e-01
7.14302778e-01 -1.25128329e-01 2.23115772e-01 -1.25082231e+00
-3.37501019e-01 3.42137635e-01 9.98900056e-01 7.31761575e-01
-1.19663501e+00 1.50231564e+00 7.81163120e+00 1.03610063e+00
-1.00612485e+00 -3.40904713e-01 3.26087534e-01 3.12292129e-01
-4.24326628e-01 2.44407937e-01 -1.33457696e+00 4.53931183e-01
9.96462822e-01 -2.71486223e-01 1.87941995e-02 6.34117901e-01
1.07471988e-01 -2.41058782e-01 -6.22263789e-01 5.47248721e-01
1.35001034e-01 -1.14598942e+00 1.43286869e-01 -2.57624716e-01
9.62770343e-01 -3.90899718e-01 -7.35461246e-03 5.00889719e-02
3.85260969e-01 -7.66117156e-01 1.15957785e+00 8.53030533e-02
6.31273508e-01 -9.36548471e-01 9.81171310e-01 4.91492212e-01
-1.23882449e+00 1.34795383e-01 -2.90838391e-01 -3.53633821e-01
4.33565170e-01 1.13742739e-01 -8.14572632e-01 8.24432075e-01
6.56371653e-01 4.12853420e-01 -8.84485304e-01 1.47869527e+00
-6.45178318e-01 8.77351522e-01 -1.87791824e-01 -6.90556943e-01
6.74146712e-01 -2.09700644e-01 2.43503496e-01 2.13576531e+00
2.96451002e-01 3.61227870e-01 -1.66005254e-01 3.78074855e-01
2.65477240e-01 4.60595220e-01 -3.42778265e-01 6.37651607e-02
4.47297841e-01 7.83312798e-01 -1.22748399e+00 -7.03219891e-01
-4.18029785e-01 1.17167592e+00 1.96444348e-01 2.11138964e-01
-1.06845558e+00 -1.01077664e+00 -7.76055828e-03 -3.17867041e-01
6.65435016e-01 -4.46119815e-01 -6.83494210e-01 -1.06090939e+00
1.07860558e-01 -1.00420094e+00 4.48387533e-01 -2.59531260e-01
-3.26454848e-01 6.99864209e-01 1.74122751e-02 -1.35136187e+00
-8.46560523e-02 -6.92563415e-01 -6.16358578e-01 8.28449726e-01
-1.64916503e+00 -7.44606793e-01 1.68990910e-01 3.57317448e-01
7.93447375e-01 4.00659174e-01 7.55696118e-01 3.87794316e-01
-7.61300743e-01 1.40424609e+00 2.41463214e-01 3.55739951e-01
7.16290295e-01 -1.34266496e+00 8.89588118e-01 1.77744913e+00
2.08195865e-01 1.16806817e+00 6.66915417e-01 -7.75536537e-01
-1.10137248e+00 -8.03559959e-01 1.59341729e+00 -3.60378414e-01
4.71321642e-01 -3.71862739e-01 -8.73007119e-01 5.80408394e-01
2.98257858e-01 -4.63063389e-01 6.36699915e-01 -2.02880859e-01
-1.68846101e-01 2.81457663e-01 -9.48370099e-01 6.81532800e-01
9.43004131e-01 -1.59813643e-01 -7.15712905e-01 -3.65012810e-02
8.80946159e-01 -8.44727933e-01 -4.20600414e-01 1.81810990e-01
5.53270757e-01 -7.38694608e-01 4.42107290e-01 -6.18045032e-01
1.07526273e-01 -4.98581856e-01 5.95015660e-02 -7.06429660e-01
-1.82867229e-01 -1.01159859e+00 1.82862788e-01 1.10951519e+00
1.00162113e+00 -2.41980955e-01 6.71473682e-01 7.75547445e-01
-5.19934356e-01 -4.31990653e-01 -4.67514038e-01 -9.70404327e-01
3.10797215e-01 -7.91914046e-01 7.41230607e-01 8.07345510e-01
6.20006263e-01 -3.15002114e-01 -3.66887063e-01 1.97932452e-01
1.74028948e-02 2.20992148e-01 5.05333006e-01 -5.24487257e-01
-1.04525276e-01 -6.92663789e-01 1.52258491e-02 -1.62533426e+00
-1.20891012e-01 -6.48768365e-01 2.27267340e-01 -1.17964435e+00
2.41262335e-02 -6.31482244e-01 -3.57230574e-01 3.10392201e-01
-5.38945556e-01 2.48653460e-02 5.71030915e-01 -7.85032213e-02
-8.60972226e-01 -2.74365366e-01 1.12453127e+00 5.68545908e-02
-2.96340346e-01 3.45266372e-01 -8.40687394e-01 7.91616261e-01
8.70606124e-01 -3.51976335e-01 -7.37786368e-02 -8.75568330e-01
3.20781201e-01 -3.01024169e-01 -5.64666510e-01 -8.03615093e-01
4.91660714e-01 -1.70195952e-01 -1.21830089e-03 -9.32704568e-01
-2.93665290e-01 -9.19568181e-01 -2.37923786e-01 9.04953837e-01
-3.98975551e-01 5.86734533e-01 3.62636268e-01 3.87135357e-01
-3.03144336e-01 -7.60068774e-01 6.15873396e-01 6.78464025e-02
-1.02939391e+00 -1.09655678e-01 -4.26000208e-01 -1.56408120e-02
1.08681202e+00 -6.36104822e-01 -1.14228539e-01 -1.06106453e-01
-3.99703562e-01 2.09991738e-01 5.62051952e-01 3.44983369e-01
5.20187795e-01 -9.10154462e-01 -4.46854353e-01 2.61998087e-01
1.04679324e-01 -2.32189253e-01 -1.03064224e-01 6.23410046e-01
-1.54145133e+00 1.01909041e+00 4.88809645e-02 -1.10569537e-01
-1.40092385e+00 7.53979981e-01 -1.05224133e-01 -5.28024197e-01
-6.55995667e-01 1.02801979e+00 -4.11713004e-01 -1.85576037e-01
5.10950863e-01 -8.91880095e-01 -2.70199716e-01 -2.99953163e-01
9.03542280e-01 5.16805410e-01 1.78990856e-01 -4.14436191e-01
-4.17596430e-01 3.93996626e-01 -3.97997886e-01 -1.89166561e-01
7.93158531e-01 -2.47598991e-01 1.97955798e-02 2.19980516e-02
9.28542852e-01 5.91257691e-01 -7.30697572e-01 -1.33278564e-01
4.30034786e-01 -5.74710906e-01 -5.15923858e-01 -7.59921670e-01
-6.63540304e-01 9.45170283e-01 2.50262946e-01 7.36798495e-02
1.14879131e+00 -5.65674841e-01 1.14650023e+00 6.00250542e-01
4.15582150e-01 -1.39837515e+00 -6.87888265e-01 8.50208282e-01
2.64773816e-01 -1.02311289e+00 9.48929489e-02 -7.05130756e-01
-7.81376660e-01 1.26810503e+00 5.02072453e-01 -2.03432545e-01
4.68954921e-01 5.63362241e-01 1.18224621e-01 4.19049591e-01
-3.82935554e-01 -1.70431688e-01 1.45576566e-01 3.69485021e-01
5.71511745e-01 1.87305525e-01 -1.20409274e+00 9.01882470e-01
-3.19422960e-01 -1.89345866e-01 7.34974980e-01 1.22154462e+00
-7.62550473e-01 -1.50412250e+00 -3.22438568e-01 -4.43809927e-02
-9.77495253e-01 -4.55070436e-01 -6.15411460e-01 1.11562979e+00
-1.20573252e-01 1.02951908e+00 -1.33762598e-01 -2.51734942e-01
2.57969975e-01 -2.41514385e-01 2.73605436e-01 -5.07979155e-01
-9.23305571e-01 3.03020954e-01 1.36387333e-01 -4.60477769e-01
-2.48216107e-01 -6.69920146e-01 -1.65467715e+00 -4.28688526e-01
-8.78182411e-01 4.65620577e-01 3.89713496e-01 8.24229956e-01
-4.26244810e-02 2.05775462e-02 4.41986412e-01 -6.54237419e-02
-5.05693614e-01 -7.20291018e-01 -4.88694966e-01 3.21268797e-01
1.56504812e-03 2.52768308e-01 -9.57325846e-02 2.15194121e-01] | [10.917216300964355, 10.82865047454834] |
becddd8b-fbd1-42e7-9794-796c22bcad7a | a-decomposition-dynamic-graph-convolutional | null | null | https://www.sciencedirect.com/science/article/abs/pii/S0031320323003710 | https://www.sciencedirect.com/science/article/abs/pii/S0031320323003710 | A Decomposition Dynamic graph convolutional recurrent network for traffic forecasting | Our daily lives are greatly impacted by traffic conditions, making it essential to have accurate predictions of traffic flow within a road network. Traffic signals used for forecasting are usually generated by sensors along roads, which can be represented as nodes on a graph. These sensors typically produce normal signals representing normal traffic flows and abnormal signals indicating unknown traffic disruptions. Graph convolution networks are widely used for traffic prediction due to their ability to capture correlations between network nodes. However, existing approaches use a predefined or adaptive adjacency matrix that does not accurately reflect real-world relationships between signals. To address this issue, we propose a decomposition dynamic graph convolutional recurrent network (DDGCRN) for traffic forecasting. DDGCRN combines a dynamic graph convolution recurrent network with an RNN-based model that generates dynamic graphs based on time-varying traffic signals, allowing for the extraction of both spatial and temporal features. Additionally, DDGCRN separates abnormal signals from normal traffic signals and models them using a data-driven approach to further improve predictions. Results from our analysis of six real-world datasets demonstrate the superiority of DDGCRN compared to the current state-of-the-art. The source codes are available at: https://github.com/wengwenchao123/DDGCRN. | ['Wenchao Weng; Jin Fan; Huifeng Wu; Yujie Hu; Hao Tian; Fu Zhu; Jia Wu'] | 2023-05-01 | null | null | null | pattern-recognition-2023-5 | ['traffic-prediction'] | ['time-series'] | [ 3.49939093e-02 -1.82675615e-01 -1.75209448e-01 -3.59505773e-01
-3.27258324e-03 -1.55645534e-01 4.44872528e-01 -5.39073832e-02
2.53783315e-01 3.56785685e-01 2.40885779e-01 -7.04705775e-01
-1.21255204e-01 -1.33837533e+00 -4.69527036e-01 -3.09703499e-01
-4.04144168e-01 3.18264872e-01 4.57391441e-01 -4.35980380e-01
-2.68308640e-01 7.26955235e-01 -1.54955399e+00 2.08673954e-01
7.01237738e-01 9.44913328e-01 -9.73472595e-02 6.64818466e-01
-3.43661636e-01 7.68657684e-01 -4.65981185e-01 -2.27495581e-01
1.38402507e-01 -2.68959045e-01 -2.81255215e-01 -1.34497598e-01
8.74835066e-03 6.64270148e-02 -1.20246553e+00 8.42696786e-01
1.67517468e-01 1.40926003e-01 2.37991720e-01 -1.53998983e+00
-5.32299936e-01 6.38456106e-01 -2.14964122e-01 5.34104705e-01
1.19174764e-01 4.48485970e-01 9.86298442e-01 -2.56144881e-01
3.53916973e-01 1.25475800e+00 6.18062675e-01 3.32686186e-01
-1.30713463e+00 -9.80199337e-01 4.35839474e-01 4.39874232e-01
-1.29096925e+00 -4.35957968e-01 1.12820840e+00 -4.56930935e-01
9.47690725e-01 3.98138106e-01 8.29162240e-01 1.23933542e+00
2.85981447e-01 5.30924737e-01 4.68219757e-01 2.61550725e-01
-6.59692138e-02 -3.92138958e-01 2.09640950e-01 4.55606461e-01
1.20881088e-01 3.52921009e-01 -1.51378855e-01 6.88491762e-02
6.00856960e-01 5.80197096e-01 -1.40249640e-01 1.76412195e-01
-8.86410534e-01 6.11379206e-01 8.50236475e-01 4.80083048e-01
-6.77348733e-01 5.93699992e-01 5.02998054e-01 4.19183135e-01
6.98863685e-01 -1.23059012e-01 -6.60695806e-02 -3.10602725e-01
-6.29603028e-01 -6.40590489e-02 5.39089501e-01 6.64725900e-01
7.82162189e-01 4.29459900e-01 -1.84213310e-01 8.18899512e-01
3.12432528e-01 8.20869207e-01 2.54185230e-01 -5.21091640e-01
6.18440449e-01 8.92818570e-01 -4.74941164e-01 -1.76504540e+00
-6.61547124e-01 -5.84462821e-01 -1.07684779e+00 -3.02226037e-01
2.47996360e-01 -1.57087877e-01 -8.82208645e-01 1.61704504e+00
8.95621628e-03 9.29995775e-01 -1.71913445e-01 6.74442708e-01
9.36029375e-01 8.82543921e-01 -1.00820372e-02 1.94269463e-01
7.32270360e-01 -5.71332812e-01 -7.25836694e-01 -1.42391533e-01
6.68408692e-01 -1.71338186e-01 8.10205162e-01 -3.68698989e-03
-5.77962518e-01 -4.48707283e-01 -5.33358574e-01 4.61707920e-01
-7.73647368e-01 -1.38559982e-01 6.03515148e-01 4.28543746e-01
-1.19315565e+00 3.68048966e-01 -8.08965623e-01 -2.92696267e-01
3.97050142e-01 2.19847932e-01 -4.98489588e-02 -6.03262298e-02
-1.59859133e+00 4.58856851e-01 7.12249354e-02 7.75352538e-01
-5.91164887e-01 -6.06240451e-01 -9.14170623e-01 1.04502209e-01
2.96787024e-01 -2.12476656e-01 9.00479078e-01 -6.59778357e-01
-1.23999846e+00 1.77675679e-01 -2.37438545e-01 -7.36460388e-01
3.96864116e-01 2.61000961e-01 -1.41292691e+00 -4.82169613e-02
6.67360500e-02 7.54464194e-02 6.75301731e-01 -8.93483758e-01
-3.83102238e-01 -6.80944473e-02 -1.93877146e-01 -3.67512137e-01
-8.50746557e-02 -2.04890504e-01 -5.77642858e-01 -6.10314548e-01
5.12697697e-02 -1.00483370e+00 -4.79036033e-01 -4.67774153e-01
-7.01448917e-01 -1.58724844e-01 1.26147437e+00 -5.39427102e-01
1.70962965e+00 -1.98218334e+00 -5.39001703e-01 7.88839877e-01
5.19365668e-01 4.93882269e-01 -4.20956671e-01 6.38678372e-01
-2.77996868e-01 3.27367261e-02 -1.62172019e-02 -4.85577099e-02
-2.13697746e-01 2.83999354e-01 -5.60273409e-01 2.86352932e-01
1.68680802e-01 1.27966750e+00 -1.07024598e+00 3.59958224e-02
5.19570649e-01 7.37049103e-01 -8.18471313e-02 -7.77450949e-02
-2.81382948e-01 4.83456165e-01 -6.01856232e-01 4.28283483e-01
5.04559517e-01 -3.12718064e-01 1.50246635e-01 -5.37136793e-02
-5.59038259e-02 4.06360120e-01 -8.43465328e-01 8.02349031e-01
-5.96296310e-01 1.12187481e+00 -2.92707890e-01 -1.18323493e+00
1.19944131e+00 2.57104874e-01 6.20500684e-01 -1.17377222e+00
1.68617785e-01 9.20083188e-03 7.25218430e-02 -3.92918259e-01
2.20912203e-01 3.00316662e-01 -9.15709659e-02 3.82925689e-01
-5.16012430e-01 4.38895226e-01 3.53371799e-01 3.05055052e-01
1.61234975e+00 -7.22461760e-01 -3.22212636e-01 3.57981592e-01
5.46342552e-01 -3.01164538e-01 5.78804076e-01 4.12740082e-01
-1.03023186e-01 4.06603247e-01 7.45857298e-01 -5.92902362e-01
-4.42520797e-01 -1.08762729e+00 1.93435609e-01 6.53643906e-01
-6.04648516e-02 -3.32571626e-01 -2.49696866e-01 -5.92809975e-01
2.81104445e-01 7.19301105e-01 -6.44432366e-01 -3.93268019e-01
-7.13635325e-01 -5.31213939e-01 6.08725488e-01 4.59871560e-01
4.58337843e-01 -1.11998463e+00 -1.55823112e-01 4.98242289e-01
-1.94528893e-01 -1.29407144e+00 -4.85208511e-01 -3.85417730e-01
-5.88942349e-01 -1.19202948e+00 -1.68535158e-01 -1.69120193e-01
6.74811602e-01 5.74343204e-01 1.05881989e+00 3.21386069e-01
-2.26601034e-01 4.26348388e-01 -3.41453880e-01 -2.22406089e-01
-4.64656144e-01 2.11968645e-01 -6.92036673e-02 6.39708638e-01
4.84908879e-01 -1.00325000e+00 -4.72292721e-01 4.96103227e-01
-7.79793918e-01 -4.83169127e-03 4.54657614e-01 3.54671150e-01
3.15356582e-01 3.08595181e-01 7.77298987e-01 -1.00512636e+00
8.54932904e-01 -9.33530211e-01 -6.88017368e-01 -4.88686748e-03
-5.00807285e-01 -2.13892415e-01 8.78774703e-01 -3.99282247e-01
-5.84822416e-01 -1.95698932e-01 -5.15344590e-02 -8.09891343e-01
-2.22573578e-01 7.05783486e-01 2.93791573e-03 2.55548567e-01
4.13256615e-01 2.03988612e-01 2.61253603e-02 -1.66275084e-01
2.81808019e-01 5.92078865e-01 3.72855186e-01 -6.75812811e-02
8.22234273e-01 4.17326987e-01 9.05579850e-02 -1.05967200e+00
-4.12228674e-01 -5.07085621e-01 -3.76359731e-01 -8.14114273e-01
4.67926562e-01 -5.93599796e-01 -8.42918694e-01 4.60391581e-01
-9.90028501e-01 -6.10842049e-01 -2.54606724e-01 4.37691748e-01
-9.07151997e-02 2.52534878e-02 -4.65482086e-01 -8.69660616e-01
-1.42748564e-01 -7.43765473e-01 7.18719006e-01 -3.28685269e-02
-1.20749481e-01 -1.33035040e+00 -7.33681917e-02 8.49889293e-02
8.67980540e-01 5.83307564e-01 6.57851577e-01 -4.65339869e-01
-6.65662289e-01 -4.85300958e-01 -4.47696686e-01 1.23995610e-01
4.10080820e-01 2.38311231e-01 -6.06279492e-01 6.57740161e-02
-6.70955658e-01 5.29195726e-01 1.00911522e+00 5.58601201e-01
1.32626390e+00 -3.58974844e-01 -5.87638855e-01 4.72738564e-01
9.97411668e-01 3.28851998e-01 8.37354422e-01 -6.93877190e-02
9.76354182e-01 5.75564384e-01 1.36519849e-01 2.14968681e-01
5.90513349e-01 6.19765460e-01 6.07209384e-01 -2.24150077e-01
-3.64950389e-01 -4.81152028e-01 3.39709640e-01 7.56758273e-01
1.79448202e-01 -5.08981168e-01 -1.39782870e+00 7.40814924e-01
-2.05894780e+00 -1.23818660e+00 -6.84185565e-01 1.96187735e+00
8.47155303e-02 4.42847133e-01 3.73195022e-01 2.64643043e-01
1.01612055e+00 4.65698272e-01 -7.35554516e-01 -4.62420613e-01
6.39717877e-02 -8.08348432e-02 6.85953975e-01 4.14965183e-01
-8.01473081e-01 9.76672232e-01 5.80840540e+00 4.62179154e-01
-1.54651642e+00 -9.62276459e-02 6.27924263e-01 -3.43850441e-02
-4.48678493e-01 -2.42356360e-01 -4.29686576e-01 7.86034942e-01
1.54431927e+00 -1.81083918e-01 5.41091323e-01 4.55668956e-01
8.54895532e-01 3.32137674e-01 -6.10667408e-01 9.70454991e-01
-2.61278421e-01 -1.35566962e+00 7.16231987e-02 1.40631154e-01
4.07293916e-01 5.16866028e-01 2.40113556e-01 1.79124966e-01
6.41115308e-01 -1.25655842e+00 1.40194327e-01 1.00022864e+00
6.91787541e-01 -7.18358338e-01 5.24548829e-01 6.80861622e-02
-1.75266838e+00 -1.40737966e-01 4.83501554e-02 -1.03494011e-01
4.70227450e-01 1.10437322e+00 -9.38430727e-01 5.14748514e-01
4.97244835e-01 1.36637604e+00 -5.84960699e-01 8.76776755e-01
-1.93974748e-01 1.13934851e+00 -2.48919293e-01 -7.52781853e-02
2.57376939e-01 -3.39929581e-01 6.80425227e-01 1.30441523e+00
3.14007342e-01 -5.87811321e-02 7.83741698e-02 1.01215315e+00
-9.09485072e-02 -2.78074354e-01 -1.00959945e+00 -3.25653136e-01
5.69345415e-01 1.07912850e+00 -7.38798261e-01 -1.84805900e-01
-4.70220357e-01 4.00379479e-01 1.16099022e-01 7.48684466e-01
-9.48900461e-01 -5.59717834e-01 1.03035462e+00 5.29335439e-01
1.41325980e-01 -4.33245927e-01 -7.65540898e-02 -9.51770186e-01
-2.83070207e-02 -4.78768438e-01 3.65800261e-01 -5.22009671e-01
-1.26789796e+00 7.47007012e-01 -1.40271321e-01 -1.49256968e+00
-3.38828206e-01 -2.91282296e-01 -1.07657695e+00 6.89774394e-01
-1.40469790e+00 -1.05008423e+00 -3.95478785e-01 7.18304038e-01
2.69193441e-01 -1.47550896e-01 3.26227516e-01 4.99997050e-01
-9.75888431e-01 4.27676350e-01 -1.44224599e-01 4.69695151e-01
4.90861684e-02 -7.51996636e-01 9.98936713e-01 9.84157920e-01
9.24642384e-02 3.21894914e-01 4.94828612e-01 -8.85550380e-01
-1.28548908e+00 -1.79833853e+00 9.44016457e-01 -1.75798014e-01
1.02220082e+00 -3.76853108e-01 -1.02063251e+00 8.98495317e-01
-3.58401746e-01 4.94244188e-01 4.25515324e-01 7.86155760e-02
-3.65392953e-01 -5.78055322e-01 -8.72214615e-01 6.59365773e-01
1.30568016e+00 -5.80339253e-01 9.03078914e-02 1.67248651e-01
6.58001840e-01 -1.30464002e-01 -4.96527255e-01 2.23470613e-01
3.47167194e-01 -7.52183080e-01 7.54534304e-01 -3.74140918e-01
-2.58091441e-03 -3.54255587e-01 1.23073682e-01 -1.54093039e+00
-2.71261275e-01 -5.39783239e-01 -3.58699054e-01 1.01917207e+00
5.02721786e-01 -1.32850015e+00 7.10229456e-01 5.58854640e-01
-5.66976033e-02 -6.49655461e-01 -8.66453290e-01 -7.45245993e-01
-5.22208035e-01 -1.25744903e+00 1.07439494e+00 7.84136653e-01
-1.32935435e-01 3.38203371e-01 -3.12280118e-01 2.27394506e-01
2.94162959e-01 -8.24127905e-03 7.96718419e-01 -1.48473275e+00
2.18893290e-01 -7.22187698e-01 -9.50933933e-01 -9.13901865e-01
4.36696172e-01 -1.02766562e+00 -2.98865318e-01 -1.65917122e+00
-5.29364645e-01 -6.04675949e-01 -3.04838479e-01 5.30757010e-01
3.10108036e-01 1.73315138e-01 -2.22573765e-02 -3.12615260e-02
-4.33122337e-01 5.77192128e-01 9.90559399e-01 -2.58598506e-01
-4.21711564e-01 4.24034357e-01 -5.29224992e-01 3.60467136e-01
1.13980186e+00 -2.05831721e-01 -6.69785917e-01 -3.06201726e-01
2.71531940e-01 1.23805590e-01 5.00436604e-01 -1.12979853e+00
3.72176409e-01 -1.03447072e-01 3.21947015e-03 -7.65350521e-01
1.75708622e-01 -9.37461197e-01 5.35941541e-01 5.00667155e-01
-2.40015581e-01 2.75834471e-01 1.85269356e-01 9.09772754e-01
-3.15225214e-01 8.39142203e-01 3.25803816e-01 2.16533899e-01
-7.07993031e-01 7.51765311e-01 -7.31653512e-01 -2.09406376e-01
9.14295375e-01 -2.19952583e-01 -4.10457522e-01 -9.61375892e-01
-6.82352543e-01 5.38753271e-01 -8.73907059e-02 8.34030509e-01
9.00020361e-01 -1.43436706e+00 -5.69528222e-01 5.06249011e-01
3.24779488e-02 -2.99978703e-01 3.52828562e-01 8.95626128e-01
-2.00786531e-01 6.02960467e-01 1.32621422e-01 -5.57447851e-01
-1.07308793e+00 3.21199149e-01 5.73480070e-01 -2.41630703e-01
-8.88899446e-01 3.07815731e-01 -1.36272147e-01 -6.01297796e-01
6.35830611e-02 -6.57359600e-01 -4.20721889e-01 -4.23619188e-02
3.32087845e-01 6.50340140e-01 1.75176755e-01 -9.89243925e-01
-2.95521826e-01 3.34551305e-01 2.77880043e-01 2.36247078e-01
1.34788883e+00 -2.58571774e-01 -5.87558523e-02 6.20881677e-01
1.26672101e+00 -3.04045200e-01 -9.30748522e-01 -4.39687222e-01
3.74324294e-03 -4.44720298e-01 1.06753960e-01 -4.88536954e-01
-1.82197392e+00 8.73065710e-01 4.76920605e-01 7.26284027e-01
1.16715050e+00 -2.66177971e-02 1.26350486e+00 2.44590476e-01
1.59037933e-01 -7.82675326e-01 -1.19056828e-01 7.19405830e-01
7.79638231e-01 -9.22478795e-01 -6.13779008e-01 -4.92008567e-01
-4.81605738e-01 1.17461288e+00 5.14043450e-01 -9.57871005e-02
1.20642281e+00 1.02598697e-01 1.04353219e-01 -5.65569937e-01
-8.83927941e-01 -5.57892919e-01 3.63564342e-01 5.38098693e-01
1.83969781e-01 4.38266724e-01 1.78342611e-01 2.34889820e-01
-3.32844466e-01 -1.09564207e-01 3.59610200e-01 3.60353112e-01
-1.30694315e-01 -9.01974559e-01 -5.56971133e-02 9.03111279e-01
-8.77843425e-02 1.95849687e-01 -3.25689524e-01 5.19440651e-01
-2.66211331e-01 1.33801508e+00 4.35007691e-01 -9.55181956e-01
6.33605957e-01 -3.00000876e-01 -3.80798548e-01 -3.84038627e-01
-3.10976624e-01 -4.95239466e-01 1.97970659e-01 -9.44738686e-01
-1.39062360e-01 -5.73566854e-01 -1.28883433e+00 -6.21261835e-01
-1.55371306e-02 1.28262281e-01 4.31659967e-01 7.31706262e-01
7.63885498e-01 9.49412882e-01 9.69476879e-01 -6.45796776e-01
3.66101891e-01 -7.59381831e-01 -5.23967743e-01 3.52296919e-01
6.17619276e-01 -6.54019952e-01 -2.98132956e-01 -4.44982708e-01] | [6.468679428100586, 2.046082019805908] |
660e2118-11cf-4a5c-8443-c7e2229b7771 | semantic-reinforced-attention-learning-for | 2108.08443 | null | https://arxiv.org/abs/2108.08443v1 | https://arxiv.org/pdf/2108.08443v1.pdf | Semantic Reinforced Attention Learning for Visual Place Recognition | Large-scale visual place recognition (VPR) is inherently challenging because not all visual cues in the image are beneficial to the task. In order to highlight the task-relevant visual cues in the feature embedding, the existing attention mechanisms are either based on artificial rules or trained in a thorough data-driven manner. To fill the gap between the two types, we propose a novel Semantic Reinforced Attention Learning Network (SRALNet), in which the inferred attention can benefit from both semantic priors and data-driven fine-tuning. The contribution lies in two-folds. (1) To suppress misleading local features, an interpretable local weighting scheme is proposed based on hierarchical feature distribution. (2) By exploiting the interpretability of the local weighting scheme, a semantic constrained initialization is proposed so that the local attention can be reinforced by semantic priors. Experiments demonstrate that our method outperforms state-of-the-art techniques on city-scale VPR benchmark datasets. | ['Danwei Wang', 'Xiaoyu Tang', 'Zhenyu Wu', 'Jun Zhang', 'Yufeng Yue', 'Guohao Peng'] | 2021-08-19 | null | null | null | null | ['visual-place-recognition'] | ['computer-vision'] | [ 3.07099462e-01 1.15772873e-01 -3.44317108e-01 -4.47374284e-01
-5.37341118e-01 -3.20568591e-01 7.81528056e-01 8.68418217e-02
-2.51409888e-01 5.23721457e-01 7.17029810e-01 -1.21612595e-02
-2.95609385e-01 -6.90035164e-01 -7.69914210e-01 -7.36253977e-01
2.57555872e-01 -2.11833101e-02 3.97499591e-01 -4.90293741e-01
6.25189662e-01 4.24411088e-01 -1.54940939e+00 3.50452393e-01
1.11589694e+00 1.10298049e+00 7.56678283e-01 9.43949968e-02
-2.03428775e-01 8.45746160e-01 -2.98526555e-01 1.74431995e-01
2.50073344e-01 -3.75328094e-01 -6.80407286e-01 2.11254492e-01
2.44442634e-02 -7.20171332e-02 -2.23017022e-01 1.09417379e+00
2.54564196e-01 3.49295348e-01 6.66760802e-01 -1.17845905e+00
-1.13920045e+00 3.64810675e-01 -6.95205748e-01 9.91763771e-02
1.90855280e-01 2.97470391e-01 1.53753209e+00 -1.17235589e+00
5.15236259e-01 1.25489104e+00 7.17857629e-02 2.57303745e-01
-1.29073143e+00 -4.09106821e-01 7.62361348e-01 3.82349193e-01
-1.51402354e+00 -3.65584880e-01 1.36928988e+00 -5.64148486e-01
7.64454126e-01 1.35652944e-01 6.27986312e-01 1.18715560e+00
-2.52686471e-01 8.51128697e-01 1.06236434e+00 -4.45951998e-01
3.23222369e-01 7.61806965e-02 -1.98329628e-01 4.74174857e-01
7.64618767e-03 1.93312034e-01 -6.91773415e-01 2.25880090e-02
9.59706604e-01 1.75114840e-01 -4.40401137e-01 -7.96955824e-01
-1.32566166e+00 8.99478436e-01 1.11423230e+00 5.55636168e-01
-4.28714365e-01 1.90003306e-01 1.85014009e-01 -1.98349789e-01
3.95193875e-01 5.65518498e-01 -3.84623706e-01 2.46262506e-01
-7.75328934e-01 -2.84643308e-03 -9.08301175e-02 6.31637990e-01
1.07600522e+00 1.33473575e-01 -5.24906516e-01 1.07215309e+00
9.15674806e-01 2.82837331e-01 4.65533316e-01 -6.27239347e-01
6.37467146e-01 8.34916174e-01 1.07446618e-01 -1.25744319e+00
-1.88203484e-01 -3.13725203e-01 -6.00940704e-01 2.00211897e-01
1.16982102e-01 4.73079264e-01 -1.19016945e+00 1.81540847e+00
1.90104187e-01 1.60554588e-01 -1.13648430e-01 1.14209819e+00
9.19701636e-01 4.93659049e-01 2.87976056e-01 2.50922471e-01
1.47604716e+00 -1.18533313e+00 -7.21341491e-01 -5.68979263e-01
2.25496799e-01 -3.72985005e-01 1.31051588e+00 -1.65508747e-01
-5.66780388e-01 -6.50508046e-01 -1.28202164e+00 -3.88512492e-01
-6.00122690e-01 1.45989031e-01 4.70090002e-01 1.84990659e-01
-9.03001547e-01 2.51991421e-01 -6.04684412e-01 -4.58555877e-01
7.16067314e-01 4.57656607e-02 -2.96327114e-01 -5.40702641e-02
-1.20833802e+00 6.75342441e-01 4.90415335e-01 3.40940446e-01
-9.11393225e-01 -4.35075194e-01 -1.20145154e+00 3.02101672e-01
3.59365195e-01 -5.51826775e-01 6.54645801e-01 -1.22869360e+00
-1.42239976e+00 7.55909741e-01 -3.63255709e-01 -1.02746665e-01
3.28146338e-01 7.42832273e-02 -1.01858839e-01 3.38524133e-01
2.61125118e-01 9.20935214e-01 1.11184776e+00 -1.69118309e+00
-7.21487463e-01 -4.63002324e-01 1.09500609e-01 3.00469577e-01
-3.82879347e-01 -1.22309618e-01 -5.39664924e-01 -8.72887194e-01
2.19986200e-01 -6.24476314e-01 -4.12676036e-01 1.90160144e-02
-4.18228179e-01 -2.22899586e-01 7.16086090e-01 -5.77113867e-01
9.78872478e-01 -2.36510086e+00 2.75660545e-01 2.90258408e-01
2.85975993e-01 6.75144196e-02 -2.41879225e-01 2.14278698e-01
-8.60953927e-02 9.23433751e-02 -3.06573689e-01 -2.99972206e-01
1.62213087e-01 3.09518278e-01 -7.98298001e-01 3.42833340e-01
5.30579150e-01 1.05877197e+00 -1.08002639e+00 -3.65714520e-01
4.73374903e-01 6.01708829e-01 -5.82064390e-01 2.38455087e-01
-2.96631426e-01 7.14501917e-01 -7.35204697e-01 6.80495918e-01
5.26036024e-01 -3.77425790e-01 5.29497229e-02 -2.80529648e-01
-1.67046025e-01 3.35073829e-01 -1.06951749e+00 1.88611078e+00
-2.62132466e-01 4.85068291e-01 -1.20113492e-01 -1.09299016e+00
1.06342220e+00 5.48272245e-02 1.96989432e-01 -9.34752464e-01
9.65808704e-02 1.69660136e-01 -1.10788040e-01 -3.67725104e-01
5.87321877e-01 1.44442365e-01 3.53815556e-02 1.59980401e-01
-2.74562612e-02 3.01335752e-01 -1.55598387e-01 8.42371210e-02
7.16851890e-01 4.63481843e-01 4.49764878e-01 -4.81515765e-01
6.74933970e-01 -1.58239961e-01 1.00517547e+00 6.49223685e-01
-4.70995188e-01 8.53406906e-01 4.55641210e-01 -4.55766737e-01
-8.51292431e-01 -8.39350939e-01 -1.33568451e-01 1.25698376e+00
5.31675458e-01 -3.53385419e-01 -3.94625485e-01 -8.03637385e-01
-1.05959184e-01 5.74655473e-01 -1.00995302e+00 -1.78957880e-01
-4.82935369e-01 -4.43239748e-01 -1.69357739e-03 9.47707415e-01
5.09396017e-01 -1.46092355e+00 -6.09281063e-01 9.23109502e-02
-2.78597742e-01 -1.05830419e+00 -4.60043192e-01 1.87421367e-01
-6.54450834e-01 -9.55035329e-01 -6.96160018e-01 -7.94941664e-01
1.00385606e+00 5.91020703e-01 8.45794380e-01 1.62690014e-01
7.26520792e-02 2.33086601e-01 -6.08148396e-01 -8.26804042e-02
2.98797458e-01 1.12839341e-01 -2.72490650e-01 4.14046139e-01
4.32321697e-01 -6.91282928e-01 -7.86985040e-01 2.23310187e-01
-7.26123333e-01 1.08047344e-01 6.94272697e-01 1.10882545e+00
6.78499818e-01 -3.27708036e-01 6.15249693e-01 -5.78046262e-01
3.09134007e-01 -3.95758510e-01 -6.63579583e-01 2.93225527e-01
-3.26868683e-01 4.06842351e-01 5.73094785e-01 -3.27767879e-01
-1.14651549e+00 2.35928506e-01 -8.87551382e-02 -5.22341251e-01
-2.64162540e-01 3.09547693e-01 -5.03831446e-01 -1.38644636e-01
3.30352187e-01 4.07585144e-01 -3.50942731e-01 -4.22556937e-01
5.97289622e-01 4.67121035e-01 1.80653870e-01 -7.87825823e-01
8.85749280e-01 7.02614427e-01 -2.16447756e-01 -7.13807166e-01
-9.22988951e-01 -4.02119219e-01 -7.52494991e-01 2.02282239e-03
1.09133971e+00 -9.43137050e-01 -3.91200513e-01 5.86342290e-02
-1.00187624e+00 -3.05949301e-01 -3.82039189e-01 3.26328874e-01
-6.05828345e-01 2.11021051e-01 -2.47195825e-01 -7.55419612e-01
-3.52168083e-02 -1.39833820e+00 1.33080566e+00 1.93704128e-01
-2.26138514e-02 -8.30764055e-01 -4.31685038e-02 3.25308263e-01
3.78557116e-01 1.61685944e-01 8.09830606e-01 -6.26946151e-01
-1.02419019e+00 3.51224601e-01 -7.47572660e-01 2.71730814e-02
1.27953768e-01 -4.25912291e-02 -1.14975119e+00 -1.13837861e-01
-3.11914921e-01 -4.18482363e-01 1.14229524e+00 3.01315337e-01
1.41367245e+00 -1.30465537e-01 -3.97027910e-01 6.84550643e-01
1.41103554e+00 3.73300724e-02 8.02488863e-01 7.04592347e-01
9.65722203e-01 6.62346900e-01 6.33234143e-01 4.30916280e-01
7.71443903e-01 6.80749893e-01 7.33077288e-01 -1.99299932e-01
-3.22123356e-02 -6.91022336e-01 9.55468938e-02 3.84135187e-01
-1.32550597e-01 1.07208192e-01 -6.63970172e-01 8.32733572e-01
-2.20394969e+00 -8.26402903e-01 2.58430749e-01 2.17932105e+00
5.38653076e-01 -8.71692412e-03 -1.03907906e-01 8.06261897e-02
6.37215972e-01 8.00176740e-01 -4.73966420e-01 -9.27160978e-02
-9.62235257e-02 -2.26480514e-01 1.67377725e-01 3.89236361e-01
-1.00872850e+00 1.06747818e+00 5.28447104e+00 7.84908056e-01
-1.19017625e+00 2.81285793e-02 4.25031513e-01 2.83048004e-01
-6.46816850e-01 2.42470384e-01 -6.24189258e-01 3.86247516e-01
1.15598217e-01 2.60110289e-01 3.74393731e-01 9.06530797e-01
2.36139938e-01 1.65001959e-01 -8.90044034e-01 9.78968024e-01
1.38959110e-01 -1.20901966e+00 1.37241721e-01 3.30766410e-01
5.85518420e-01 1.24747930e-02 2.25960433e-01 4.27947521e-01
2.97693312e-01 -9.33997571e-01 9.16534603e-01 3.80752951e-01
5.23717761e-01 -6.92465663e-01 3.22391957e-01 1.45264566e-01
-1.51600170e+00 -3.10412109e-01 -4.98624295e-01 5.65865934e-02
7.85638466e-02 3.23859602e-01 -4.32483166e-01 4.96369213e-01
7.36239910e-01 1.06491971e+00 -7.06056118e-01 8.78046453e-01
-8.73795569e-01 3.60142440e-01 -3.97164449e-02 1.63649082e-01
4.27804172e-01 -1.26062691e-01 5.28556764e-01 9.84601676e-01
9.81943682e-02 6.40876666e-02 2.16528639e-01 1.28010762e+00
2.66252141e-02 1.51798114e-01 -5.71212769e-01 1.33242786e-01
5.18286824e-01 1.47651505e+00 -7.24941909e-01 -3.61785740e-02
-4.01508838e-01 1.00803494e+00 7.55656362e-01 6.24878943e-01
-7.32162535e-01 -2.58191884e-01 7.14425921e-01 2.77783386e-02
8.78118098e-01 -1.71056956e-01 -3.94026488e-01 -1.34395146e+00
4.99581732e-02 -4.10593927e-01 2.79753596e-01 -8.96876216e-01
-1.26534414e+00 5.77645481e-01 -3.26020300e-01 -1.16947222e+00
1.44375935e-01 -4.99419659e-01 -5.96283615e-01 8.98609161e-01
-2.23162913e+00 -1.51371539e+00 -3.48377436e-01 6.11299098e-01
4.18989152e-01 5.85183017e-02 6.27690852e-01 1.69330254e-01
-4.76070434e-01 4.44249392e-01 -1.76580399e-01 2.59134859e-01
5.80063760e-01 -1.25765491e+00 1.18837304e-01 9.48487580e-01
2.70645946e-01 7.43709505e-01 4.34265852e-01 -3.90594810e-01
-1.03620744e+00 -1.04959512e+00 9.39636707e-01 -3.94348890e-01
6.41103506e-01 -6.53176844e-01 -1.06172407e+00 7.07729459e-01
2.06484318e-01 5.15201211e-01 6.33316517e-01 2.89294451e-01
-7.51964927e-01 -1.77311406e-01 -1.06416202e+00 7.52249300e-01
1.08108580e+00 -7.10629463e-01 -9.57730770e-01 4.45299894e-02
8.73864055e-01 -1.03089765e-01 -5.72304070e-01 3.89507383e-01
2.80093461e-01 -7.67600596e-01 1.08969450e+00 -5.09866953e-01
5.12229264e-01 -7.43529081e-01 -4.39728022e-01 -1.29288006e+00
-6.84664965e-01 -1.60393804e-01 3.99333378e-03 1.44387770e+00
2.40663156e-01 -6.59858525e-01 4.39274609e-01 4.74353135e-01
-1.53202727e-01 -6.71528637e-01 -8.41360629e-01 -5.01409650e-01
-2.77074337e-01 -3.23383033e-01 7.69952714e-01 1.06283486e+00
-5.54526784e-02 3.71078968e-01 -3.76065940e-01 3.65045339e-01
5.25892079e-01 3.61347646e-01 4.99181330e-01 -1.20925534e+00
-1.09044068e-01 -5.35278976e-01 -3.98504347e-01 -1.24266553e+00
2.20182672e-01 -7.39519358e-01 2.23133400e-01 -1.65052032e+00
3.18612725e-01 -4.35755610e-01 -7.93320179e-01 7.89955080e-01
-3.41925085e-01 2.14409217e-01 4.08609271e-01 2.47471750e-01
-9.31295276e-01 1.14142227e+00 1.26369381e+00 -2.45932490e-02
-2.12192118e-01 -5.31555235e-01 -1.03370702e+00 7.69544840e-01
5.59690475e-01 -2.79494137e-01 -4.60324109e-01 -4.74152952e-01
2.24278182e-01 -2.66813517e-01 5.61467111e-01 -5.84622741e-01
1.03968367e-01 -4.17982101e-01 3.61947089e-01 -6.33546829e-01
2.78156251e-01 -9.43255544e-01 -5.23064435e-01 7.00514913e-02
-3.62670273e-01 -3.90168905e-01 -2.38350958e-01 7.90100038e-01
-3.93917471e-01 1.54967725e-01 8.05858791e-01 -1.11039169e-01
-1.14807332e+00 3.01028311e-01 -1.10126100e-01 7.93322846e-02
9.54384506e-01 -2.74591416e-01 -3.50991607e-01 -4.34435904e-02
-4.47291255e-01 3.17453891e-01 3.75408441e-01 6.96687520e-01
8.35669577e-01 -1.62033284e+00 -3.62833530e-01 4.81884450e-01
6.78180754e-01 1.23602614e-01 3.12859148e-01 6.87161863e-01
-4.28287387e-02 5.97942412e-01 -2.62350619e-01 -6.04255319e-01
-6.67887211e-01 7.77392745e-01 1.49517357e-01 -1.75982773e-01
-6.88662648e-01 7.49214590e-01 6.89124227e-01 -4.57382560e-01
2.13829100e-01 -4.35307741e-01 -6.75113082e-01 1.08142696e-01
7.02218473e-01 -2.31269673e-01 -2.43386224e-01 -9.84707475e-01
-5.48005760e-01 7.14422524e-01 -7.18207136e-02 2.60934923e-02
1.63011181e+00 -4.14953172e-01 -5.97266364e-04 3.27715993e-01
9.30783629e-01 2.44345865e-03 -1.72922337e+00 -4.62866992e-01
1.26943393e-02 -8.47081661e-01 1.69991910e-01 -7.06108749e-01
-1.17234695e+00 9.67292905e-01 3.03411633e-01 1.56493206e-02
1.20984137e+00 1.68292709e-02 5.26878238e-01 1.36375830e-01
4.62022930e-01 -1.10615098e+00 2.75989264e-01 4.85868931e-01
1.05499840e+00 -1.49186862e+00 -2.40452975e-01 -3.79668623e-01
-8.13712656e-01 8.36513281e-01 6.90887034e-01 -2.93989360e-01
4.78315234e-01 -3.28875929e-01 -8.74307230e-02 -2.03166187e-01
-5.31545043e-01 -5.97041368e-01 5.83146155e-01 6.24822259e-01
1.98100701e-01 -1.74754530e-01 -1.82980075e-01 8.69240940e-01
2.43269220e-01 -8.91167969e-02 1.53221022e-02 8.61290157e-01
-5.95588446e-01 -8.18222582e-01 -2.90545106e-01 -4.49000448e-02
-8.06839094e-02 -1.53210819e-01 -2.25716710e-01 4.87935364e-01
1.60615996e-01 7.42446005e-01 -1.39723897e-01 -1.46330401e-01
1.87420070e-01 -1.58170566e-01 1.88915387e-01 -5.50458908e-01
-2.78372735e-01 1.68967098e-01 -2.89216548e-01 -7.58436382e-01
-4.08805251e-01 -4.30889547e-01 -1.24147093e+00 2.93960601e-01
-1.11221798e-01 -1.41878381e-01 5.48496962e-01 1.12861228e+00
5.96746266e-01 7.73042619e-01 5.79215407e-01 -1.14202940e+00
-2.02184930e-01 -7.63905048e-01 -6.55954480e-01 5.06305516e-01
4.66298163e-01 -1.07648909e+00 -4.37218904e-01 -1.35343477e-01] | [9.833759307861328, 0.12119445204734802] |
272aa3fe-8b85-47b3-9818-e29774cfdaa6 | rgb-d-and-thermal-sensor-fusion-a-systematic | 2305.11427 | null | https://arxiv.org/abs/2305.11427v2 | https://arxiv.org/pdf/2305.11427v2.pdf | RGB-D And Thermal Sensor Fusion: A Systematic Literature Review | In the last decade, the computer vision field has seen significant progress in multimodal data fusion and learning, where multiple sensors, including depth, infrared, and visual, are used to capture the environment across diverse spectral ranges. Despite these advancements, there has been no systematic and comprehensive evaluation of fusing RGB-D and thermal modalities to date. While autonomous driving using LiDAR, radar, RGB, and other sensors has garnered substantial research interest, along with the fusion of RGB and depth modalities, the integration of thermal cameras and, specifically, the fusion of RGB-D and thermal data, has received comparatively less attention. This might be partly due to the limited number of publicly available datasets for such applications. This paper provides a comprehensive review of both, state-of-the-art and traditional methods used in fusing RGB-D and thermal camera data for various applications, such as site inspection, human tracking, fault detection, and others. The reviewed literature has been categorised into technical areas, such as 3D reconstruction, segmentation, object detection, available datasets, and other related topics. Following a brief introduction and an overview of the methodology, the study delves into calibration and registration techniques, then examines thermal visualisation and 3D reconstruction, before discussing the application of classic feature-based techniques as well as modern deep learning approaches. The paper concludes with a discourse on current limitations and potential future research directions. It is hoped that this survey will serve as a valuable reference for researchers looking to familiarise themselves with the latest advancements and contribute to the RGB-DT research field. | ['Andre L. C. Barczak', 'Teo Susnjak', 'Napoleon H. Reyes', 'Martin Brenner'] | 2023-05-19 | null | null | null | null | ['3d-reconstruction', 'fault-detection'] | ['computer-vision', 'miscellaneous'] | [ 3.18330944e-01 -4.24055755e-01 6.54789954e-02 -4.18438882e-01
-7.83621430e-01 -3.65134686e-01 3.55574757e-01 2.11898070e-02
-5.76759815e-01 2.94415355e-01 -2.53461361e-01 -3.73675585e-01
-2.44034648e-01 -6.12196565e-01 -1.33865878e-01 -1.17517650e+00
1.98670730e-01 1.09383360e-01 4.99833673e-02 -1.79897472e-01
2.69196004e-01 7.98244417e-01 -2.01509833e+00 8.61055776e-03
4.41444635e-01 1.45848346e+00 3.46850961e-01 4.27121252e-01
-5.95075004e-02 2.75825381e-01 -5.15805006e-01 -2.11317390e-01
2.79610753e-01 -1.44396156e-01 -5.20566404e-01 3.83938134e-01
2.29592472e-01 -5.38934231e-01 -5.01512110e-01 5.93108475e-01
7.79456437e-01 7.51915425e-02 4.88754511e-01 -1.18103540e+00
-4.15343553e-01 -1.55142680e-01 -8.67306232e-01 1.58437654e-01
4.12818760e-01 2.46124610e-01 5.36297202e-01 -7.40215957e-01
2.04727754e-01 9.72904325e-01 8.22137237e-01 3.59115899e-01
-7.07990170e-01 -3.17324102e-01 -1.89047560e-01 5.01483202e-01
-1.09659433e+00 -4.45328355e-02 1.01665187e+00 -3.51503998e-01
1.10462964e+00 2.52352476e-01 1.02234149e+00 8.05566549e-01
4.26574528e-01 8.42100143e-01 1.60027444e+00 -6.28938198e-01
1.59821078e-01 -9.91578698e-02 2.68675126e-02 5.19920230e-01
1.59901455e-01 4.05314595e-01 -7.16926455e-01 -1.06032938e-02
4.85766679e-01 1.16885111e-01 -2.02617571e-01 -4.76828903e-01
-9.22791898e-01 8.14413846e-01 5.00152946e-01 2.67705560e-01
-4.43874627e-01 6.78378195e-02 4.24644470e-01 1.65135458e-01
4.26253170e-01 -7.74813769e-03 -4.11569268e-01 -1.56635597e-01
-9.23290610e-01 5.57025634e-02 2.99401104e-01 4.16580349e-01
1.00478613e+00 2.68435746e-01 6.57671094e-01 8.20398569e-01
7.75075674e-01 9.53001797e-01 2.68610775e-01 -1.02368605e+00
6.81069791e-02 4.72593188e-01 -3.34774315e-01 -8.27304661e-01
-4.93814290e-01 3.04102182e-01 -7.03107178e-01 7.33478367e-01
4.62698489e-02 -2.42139727e-01 -1.15074992e+00 8.33616257e-01
2.27022275e-01 -2.79049367e-01 1.62048146e-01 1.05408645e+00
1.08162892e+00 2.79755801e-01 -1.88390732e-01 -1.05500273e-01
1.40338504e+00 -2.89022505e-01 -7.42714524e-01 -7.33030200e-01
3.20852220e-01 -7.59365618e-01 5.34737647e-01 3.66109014e-01
-6.44989669e-01 -3.96891505e-01 -1.17396986e+00 -6.96283802e-02
-5.44402957e-01 -1.00754961e-01 8.18787396e-01 8.06342363e-01
-9.09824610e-01 2.98928171e-01 -1.29988253e+00 -8.10996652e-01
5.00243723e-01 1.88830227e-01 -5.38279593e-01 -5.41872919e-01
-1.09212792e+00 1.33459365e+00 2.11745739e-01 7.26897836e-01
-6.56164885e-01 -2.57922471e-01 -1.07015407e+00 -1.02241778e+00
3.38616341e-01 -5.71987629e-01 1.04262543e+00 -4.40644175e-01
-1.35963809e+00 1.08919370e+00 -6.37520850e-02 -1.87540293e-01
1.85800686e-01 -3.21622640e-01 -2.90140122e-01 2.91694939e-01
-1.09142907e-01 3.18567634e-01 5.61812758e-01 -1.42882919e+00
-8.44968140e-01 -8.81992579e-01 1.54238977e-02 4.02305812e-01
1.46613754e-02 1.03144713e-01 -3.50228995e-01 -7.66163692e-02
5.43965578e-01 -7.74192214e-01 -3.51049930e-01 1.57427639e-01
-1.81430310e-01 7.29440385e-03 1.19867265e+00 -6.71135545e-01
7.97633827e-01 -2.04229593e+00 4.96076792e-02 1.59346327e-01
-2.12472845e-02 1.80800140e-01 4.83902127e-01 6.49404764e-01
9.66012701e-02 -3.18782210e-01 -8.21325541e-01 -4.50672090e-01
-1.69703186e-01 5.75701416e-01 9.25377011e-02 9.89273250e-01
-1.07452348e-01 8.00485253e-01 -6.80853128e-01 -3.86114091e-01
1.03157187e+00 7.17842460e-01 4.26954389e-01 -1.35065481e-01
4.27112162e-01 4.27129835e-01 -4.09184337e-01 1.26549923e+00
8.32171261e-01 4.72194731e-01 -9.25575942e-02 -5.12208343e-01
-3.87388587e-01 -1.05757445e-01 -1.04488075e+00 1.71486866e+00
-4.28579360e-01 9.49288249e-01 2.95245588e-01 -1.15826488e+00
1.04696095e+00 4.44021314e-01 8.07235897e-01 -1.06490397e+00
1.81002468e-01 5.92054605e-01 -5.28568089e-01 -8.23599577e-01
6.15609646e-01 -4.40540254e-01 -1.78254321e-01 4.79832590e-01
-3.53465468e-01 -5.15579104e-01 -2.37259924e-01 -2.55484611e-01
8.55033636e-01 4.12575006e-01 2.52197266e-01 4.92936999e-01
3.68041426e-01 4.04899359e-01 2.40088433e-01 3.27204853e-01
-4.68966872e-01 5.55203021e-01 -2.29050830e-01 -3.99250329e-01
-7.73663223e-01 -9.43025410e-01 -3.20505559e-01 5.42950571e-01
4.65026915e-01 2.08994314e-01 -3.19980264e-01 -3.04623663e-01
3.28930169e-01 2.73715019e-01 -5.81349671e-01 -1.43561587e-01
-5.93820095e-01 -1.04840374e+00 5.96844971e-01 7.39713669e-01
8.27707946e-01 -1.01748240e+00 -1.45962691e+00 -4.83788587e-02
-2.74282962e-01 -1.00243986e+00 6.33995771e-01 5.57970226e-01
-1.38096452e+00 -1.27299082e+00 -6.39025986e-01 -5.81961989e-01
3.01535368e-01 7.29800820e-01 7.88680255e-01 -1.55126527e-02
-5.37455499e-01 9.64851916e-01 -5.80447376e-01 -4.39675331e-01
-9.43316612e-03 -4.31395799e-01 -2.14910850e-01 -4.08269256e-01
7.94076800e-01 -2.78923094e-01 -5.29356360e-01 1.45972684e-01
-1.09393585e+00 -3.35880190e-01 7.89959431e-01 5.90996206e-01
5.24600744e-01 5.11420742e-02 -4.59260754e-02 -1.51708260e-01
3.26136827e-01 -2.73074210e-01 -2.42871106e-01 1.49431020e-01
-6.38953269e-01 -4.34766233e-01 -3.38254392e-01 2.40272611e-01
-1.17056501e+00 1.98872134e-01 -2.67672271e-01 -3.54397118e-01
-7.93702245e-01 8.72060359e-01 -8.68655741e-02 -3.81767601e-01
6.22271478e-01 1.97019547e-01 4.50783968e-01 -5.11050224e-01
3.03323150e-01 1.16939569e+00 7.61846066e-01 -2.31826201e-01
8.96805584e-01 9.62472975e-01 -6.57815021e-03 -1.24534559e+00
-4.56073344e-01 -8.38186979e-01 -1.01499045e+00 -7.90694773e-01
9.91069973e-01 -7.19074011e-01 -3.62869501e-01 1.04413354e+00
-8.61623704e-01 -1.01384640e-01 -1.88647315e-01 5.36915362e-01
-6.14237309e-01 6.69282138e-01 -5.91304004e-01 -1.17170131e+00
-2.16363624e-01 -1.27316070e+00 1.33788788e+00 3.60980868e-01
1.48037434e-01 -1.11389434e+00 1.02386750e-01 8.64144862e-01
3.41102540e-01 7.78852642e-01 3.60551327e-01 1.91318303e-01
-3.89396638e-01 -6.51837885e-01 4.44873609e-02 5.51988304e-01
3.91674459e-01 -5.70177101e-02 -1.32210338e+00 -2.48918328e-02
2.19082192e-01 -4.20713514e-01 1.00063479e+00 4.77968991e-01
5.11101365e-01 6.72470212e-01 -4.90651339e-01 5.46928108e-01
1.55977285e+00 3.91651124e-01 6.78097606e-01 8.51657212e-01
7.31863141e-01 8.39367747e-01 1.12023008e+00 2.86864877e-01
4.88341361e-01 4.73107100e-01 9.07563627e-01 -4.10902768e-01
1.05705202e-01 6.26872003e-01 2.05441937e-01 6.16724074e-01
-6.12173498e-01 9.87655967e-02 -1.17901456e+00 4.09490138e-01
-1.62006497e+00 -8.97198200e-01 -3.29102665e-01 2.05163693e+00
2.70245016e-01 -2.97342837e-01 -5.97011931e-02 8.61485720e-01
5.44302046e-01 1.35460600e-01 -6.53115034e-01 -2.44096771e-01
-4.12558496e-01 2.03906536e-01 8.97713125e-01 1.37159035e-01
-1.30893505e+00 4.73529309e-01 6.47106075e+00 2.84084618e-01
-1.22179019e+00 -6.13245219e-02 1.75721258e-01 3.10424268e-01
-4.99516092e-02 -1.70254067e-01 -2.57841349e-01 2.82283034e-02
7.26693690e-01 3.29580516e-01 9.11004096e-02 4.79800522e-01
2.55213648e-01 -8.81721377e-01 -6.64763927e-01 1.04478049e+00
4.01707828e-01 -6.95265889e-01 -6.67337418e-01 3.36937964e-01
3.54799986e-01 3.66882086e-01 1.43201575e-02 -1.86557055e-01
-2.14993894e-01 -6.86089396e-01 8.45509946e-01 5.56984365e-01
5.50222218e-01 -7.23656058e-01 1.24118578e+00 9.04723629e-02
-1.23198080e+00 -2.53782988e-01 -2.10902810e-01 -1.54351264e-01
4.60182130e-01 6.10027373e-01 -1.82174534e-01 1.22582448e+00
1.16844821e+00 8.71651888e-01 -4.35963482e-01 1.22455335e+00
-2.30765119e-01 1.79496855e-01 -3.69909167e-01 3.58461261e-01
3.83568674e-01 -2.91896522e-01 2.52425760e-01 1.00171351e+00
3.48907351e-01 7.34730139e-02 2.06127867e-01 3.08578283e-01
7.31033564e-01 -3.60236347e-01 -7.77319491e-01 1.56001940e-01
4.15580124e-01 1.31710649e+00 -8.29123735e-01 -2.13974528e-03
-8.07353556e-01 7.72611201e-01 -4.17468578e-01 4.50776160e-01
-6.87796056e-01 -4.21810120e-01 7.37097681e-01 -1.12860046e-01
4.54188138e-02 -6.97187781e-01 -6.31087184e-01 -4.34389234e-01
-1.73499156e-02 -5.46869218e-01 3.33687067e-01 -1.00904298e+00
-1.24069655e+00 3.12268466e-01 4.07739788e-01 -1.33175337e+00
-4.37408723e-02 -9.80812907e-01 -1.80038482e-01 9.73783374e-01
-1.54696381e+00 -1.28344238e+00 -7.01138437e-01 4.13083524e-01
3.26428115e-01 1.12997718e-01 6.75846577e-01 3.22334141e-01
-5.32857180e-01 -8.61135796e-02 1.55611560e-01 -3.69951688e-02
5.95360577e-01 -1.00926757e+00 1.24776661e-01 7.01316595e-01
-2.19350174e-01 1.46114513e-01 6.24444067e-01 -6.37606442e-01
-1.72060585e+00 -6.37533247e-01 2.45899364e-01 -3.55522066e-01
3.65833402e-01 1.74695089e-01 -7.75280774e-01 5.39157510e-01
2.27314934e-01 -6.05478100e-02 7.05760181e-01 -4.72109288e-01
3.70415188e-02 -2.31476843e-01 -1.36751533e+00 4.68367115e-02
4.47191119e-01 -8.69310319e-01 -6.81676567e-01 -9.52820331e-02
3.41321304e-02 -7.23131359e-01 -9.64392126e-01 5.74137866e-01
7.52761483e-01 -1.34741044e+00 1.08768713e+00 3.61237258e-01
2.45634168e-02 -4.58655268e-01 -4.06870097e-01 -1.01766825e+00
-1.27136186e-02 2.01737791e-01 7.57118538e-02 8.67034137e-01
-1.02947041e-01 -6.35051429e-01 8.36874068e-01 5.76217115e-01
-7.52617419e-01 -6.93626761e-01 -1.17327535e+00 -4.22017038e-01
-1.34976372e-01 -1.00265193e+00 2.23106027e-01 6.68031812e-01
-2.01248139e-01 -1.54672474e-01 -1.25650942e-01 1.51364088e-01
8.14749777e-01 1.04077585e-01 5.10114849e-01 -1.28541589e+00
3.60472739e-01 -3.68904501e-01 -8.04617286e-01 -7.87063718e-01
-1.75487995e-01 -4.62414145e-01 4.23759878e-01 -2.11490321e+00
-8.74651149e-02 -2.18676016e-01 -6.30108081e-03 5.01162946e-01
1.08233437e-01 8.06388974e-01 -8.56850222e-02 3.28258336e-01
-4.45175963e-03 4.68465686e-01 1.25046051e+00 -2.70474136e-01
1.17581971e-02 7.34554883e-03 -4.66870666e-01 7.26104081e-01
8.18232894e-01 -1.88798569e-02 -1.67531133e-01 -5.52357793e-01
6.49578795e-02 3.98030952e-02 5.96021712e-01 -1.07975280e+00
3.27564240e-01 -9.32247937e-02 4.55339253e-01 -1.16914344e+00
6.15398765e-01 -1.13274562e+00 1.80970237e-01 3.66049618e-01
4.73207384e-01 2.48613104e-01 3.43607455e-01 5.23932397e-01
-4.65039700e-01 -1.28930911e-01 8.30660820e-01 -3.70840579e-01
-1.41713810e+00 -4.15893719e-02 -7.41630018e-01 -5.47741294e-01
1.24060583e+00 -1.21325898e+00 -1.33094117e-01 -5.73749430e-02
-6.02450013e-01 1.17731757e-01 5.74917793e-01 3.62688303e-01
1.03987384e+00 -1.12931716e+00 -2.92868793e-01 1.54091150e-01
2.07621202e-01 2.43079364e-01 5.99780262e-01 1.17619324e+00
-7.18413532e-01 4.06109422e-01 -3.40928733e-01 -1.05924964e+00
-1.33756459e+00 2.76275575e-01 6.14276052e-01 3.94184887e-01
-6.35672152e-01 5.91508985e-01 -4.71502781e-01 -3.99853855e-01
2.61164099e-01 -1.74846992e-01 -2.30841964e-01 1.81552842e-01
1.66569650e-01 7.70518482e-01 5.03376424e-01 -1.09529090e+00
-7.45938003e-01 1.17984998e+00 5.68303227e-01 -1.81847721e-01
1.25276113e+00 -6.93504214e-01 -2.33514890e-01 8.17445397e-01
9.06787395e-01 -7.90924847e-01 -1.18614745e+00 -1.23233341e-01
3.56408744e-03 -3.72393668e-01 5.22979915e-01 -6.95381463e-01
-1.43748498e+00 1.17228520e+00 1.19579172e+00 1.92362204e-01
1.52151263e+00 2.06898913e-01 6.88170969e-01 2.26788625e-01
6.18057430e-01 -1.11404467e+00 9.19992700e-02 4.21037406e-01
4.96141881e-01 -1.32410169e+00 3.79335999e-01 -1.38207853e-01
-6.16099536e-01 1.32573628e+00 3.30157667e-01 2.96152622e-01
7.01328039e-01 3.77533644e-01 5.65705895e-01 -5.70138991e-01
-1.29439995e-01 -4.79167670e-01 1.39936227e-02 1.11226165e+00
5.37539780e-01 -1.00441173e-01 1.88184798e-01 -3.36582452e-01
3.72069664e-02 -6.09305874e-02 3.99194717e-01 1.67515588e+00
-7.56061494e-01 -1.05321538e+00 -1.06763327e+00 4.48491693e-01
-2.15349019e-01 3.63212854e-01 -3.07247937e-01 9.77436006e-01
3.08997691e-01 1.24340546e+00 2.32423723e-01 -5.98901510e-01
4.90528226e-01 7.46077970e-02 6.07802272e-01 -1.90615192e-01
-3.79910588e-01 1.22293726e-01 -2.52763815e-02 -5.48325062e-01
-1.22671187e+00 -1.11754990e+00 -1.10439932e+00 -3.49556655e-01
-5.78839064e-01 -2.16593862e-01 1.20261908e+00 1.18246078e+00
-3.30968231e-01 5.58209896e-01 4.53149766e-01 -1.30086517e+00
-2.13905796e-02 -9.19740319e-01 -8.46903265e-01 -2.23115399e-01
5.02736509e-01 -9.56409097e-01 -3.46478790e-01 -4.46883850e-02] | [8.435394287109375, -2.158364772796631] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.