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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
c1b48d34-db75-4ef8-8f68-55bdf89537b7 | training-temporal-word-embeddings-with-a | 1906.02376 | null | https://arxiv.org/abs/1906.02376v1 | https://arxiv.org/pdf/1906.02376v1.pdf | Training Temporal Word Embeddings with a Compass | Temporal word embeddings have been proposed to support the analysis of word meaning shifts during time and to study the evolution of languages. Different approaches have been proposed to generate vector representations of words that embed their meaning during a specific time interval. However, the training process used in these approaches is complex, may be inefficient or it may require large text corpora. As a consequence, these approaches may be difficult to apply in resource-scarce domains or by scientists with limited in-depth knowledge of embedding models. In this paper, we propose a new heuristic to train temporal word embeddings based on the Word2vec model. The heuristic consists in using atemporal vectors as a reference, i.e., as a compass, when training the representations specific to a given time interval. The use of the compass simplifies the training process and makes it more efficient. Experiments conducted using state-of-the-art datasets and methodologies suggest that our approach outperforms or equals comparable approaches while being more robust in terms of the required corpus size. | ['Matteo Palmonari', 'Federico Bianchi', 'Valerio Di Carlo'] | 2019-06-05 | null | null | null | null | ['diachronic-word-embeddings'] | ['natural-language-processing'] | [-1.12959385e-01 -4.16720569e-01 -3.91689986e-01 -1.23226427e-01
-7.48300552e-02 -4.85712528e-01 1.03534615e+00 7.50089228e-01
-8.64277303e-01 4.03106481e-01 2.32484475e-01 -4.65814620e-01
-1.66088253e-01 -9.20221031e-01 -1.95524648e-01 -6.41565144e-01
1.22904032e-02 9.69318748e-02 2.78149515e-01 -3.31327379e-01
5.58013797e-01 4.10693288e-01 -1.52851665e+00 -2.62483239e-01
6.04129314e-01 4.95038599e-01 4.22389954e-01 3.75259846e-01
-6.99958980e-01 2.30183110e-01 -5.82968235e-01 -4.22582746e-01
-1.92896966e-02 -4.15709108e-01 -5.48296452e-01 -3.45045403e-02
-2.90175021e-01 1.58464566e-01 -3.46037477e-01 8.83360803e-01
3.24912876e-01 5.00401795e-01 5.24724483e-01 -8.85295451e-01
-8.55151951e-01 3.87321800e-01 -4.61392581e-01 5.82586646e-01
2.45916575e-01 -1.92364767e-01 1.00153410e+00 -7.98923016e-01
6.32312179e-01 9.45752859e-01 5.02953291e-01 5.29176891e-01
-1.23069608e+00 -3.45295042e-01 3.64503801e-01 4.18400913e-01
-1.42593491e+00 -2.25460708e-01 9.95535254e-01 -3.98704469e-01
1.10011721e+00 6.40459210e-02 9.43254054e-01 1.12136602e+00
3.64026099e-01 4.35620338e-01 7.99094141e-01 -7.38924563e-01
4.01814044e-01 2.51062900e-01 3.00402164e-01 3.08854252e-01
4.77735907e-01 -1.17603637e-01 -4.47573155e-01 -2.57152885e-01
4.58523184e-01 2.51707286e-01 -1.70194298e-01 -4.46051091e-01
-9.54780698e-01 1.27878690e+00 1.87401146e-01 1.10287297e+00
-5.15557408e-01 5.98071590e-02 5.64448535e-01 2.46947497e-01
8.31122279e-01 5.05768538e-01 -2.78159171e-01 -4.13925439e-01
-9.00402844e-01 3.21174651e-01 5.43248832e-01 4.56273824e-01
5.94775975e-01 6.82858899e-02 2.83127457e-01 1.02409577e+00
2.72949070e-01 5.20803407e-02 1.01995707e+00 -1.41637981e-01
2.69140601e-01 6.27634645e-01 1.24154024e-01 -1.24096668e+00
-2.21005335e-01 -2.63978448e-02 -4.30430561e-01 -1.32861614e-01
3.40724677e-01 8.06173831e-02 -8.72003317e-01 1.64674699e+00
4.08235729e-01 4.38474029e-01 3.34658206e-01 6.65824652e-01
3.36059690e-01 1.00055051e+00 9.77203026e-02 -2.01645821e-01
1.52461278e+00 -6.14346862e-01 -8.80215406e-01 -3.09448034e-01
8.06316376e-01 -7.81569242e-01 1.14127171e+00 2.96951413e-01
-5.76796234e-01 -4.83462423e-01 -1.09412956e+00 9.11000147e-02
-9.86461818e-01 -3.57415587e-01 7.32470095e-01 6.13735437e-01
-7.39218354e-01 6.35671079e-01 -9.26136017e-01 -9.26632404e-01
-7.36178309e-02 -2.54808441e-02 -2.39090562e-01 2.48300713e-02
-1.31614935e+00 1.09382081e+00 5.30431211e-01 -7.68876821e-02
-3.23689818e-01 -4.10666704e-01 -1.04061162e+00 -1.02722242e-01
-1.01347454e-02 -6.52052313e-02 1.01214910e+00 -7.37216711e-01
-1.28508389e+00 5.81095338e-01 -1.78460389e-01 -5.74159622e-01
9.71548483e-02 -2.03589246e-01 -6.71004832e-01 -9.95677188e-02
-7.73275271e-02 2.61339694e-01 6.36643052e-01 -9.15209651e-01
-3.38838577e-01 -1.90961584e-01 2.48387083e-01 -6.53696880e-02
-9.30229962e-01 5.99485449e-02 -4.29441303e-01 -8.69153857e-01
-5.24449497e-02 -8.40959549e-01 -4.37271059e-01 -2.42406115e-01
2.30386704e-01 -5.41322470e-01 8.66785586e-01 -5.02533197e-01
1.77031171e+00 -2.40974355e+00 1.23465776e-01 1.19295083e-01
-1.31463066e-01 4.73009020e-01 -1.56115234e-01 9.74021852e-01
-5.67003340e-02 2.88689673e-01 -2.14005560e-01 -2.04883724e-01
-1.19269498e-01 6.83081269e-01 -3.11271399e-01 5.24258554e-01
1.49100259e-01 3.92819136e-01 -1.23544812e+00 -4.32448685e-01
4.75268930e-01 7.44837046e-01 -3.71415526e-01 2.02718422e-01
2.73709297e-02 -2.45236419e-02 -5.41698039e-01 2.37572491e-01
2.52155691e-01 1.20283984e-01 4.96213913e-01 1.22889280e-01
-5.58021486e-01 3.58684450e-01 -1.00241554e+00 1.70789945e+00
-8.23516905e-01 8.34435284e-01 -6.61996543e-01 -1.18413651e+00
1.11473536e+00 4.65135306e-01 3.29812020e-01 -6.84176981e-01
2.26274401e-01 2.02623621e-01 1.11120179e-01 -6.25240624e-01
9.21949148e-01 -3.45960528e-01 -1.52781293e-01 5.68291128e-01
2.29223538e-03 -6.49319589e-02 5.04659832e-01 -2.03470334e-01
9.77875054e-01 -4.26812880e-02 6.84302926e-01 -2.38763422e-01
5.40356338e-01 -1.69205397e-01 6.09651804e-01 1.83780208e-01
-2.91526839e-02 2.91897684e-01 3.51932317e-01 -6.59843504e-01
-1.18333280e+00 -7.49159336e-01 -1.69368416e-01 8.89697611e-01
2.84927003e-02 -6.69232309e-01 -3.07205141e-01 -4.50469166e-01
-1.29672885e-01 1.11978698e+00 -8.26256335e-01 -2.22663209e-01
-6.59839630e-01 -7.54489481e-01 2.56445080e-01 5.71938813e-01
-6.08365834e-02 -1.05570996e+00 -1.05844188e+00 5.49021661e-01
8.96044075e-02 -7.84808517e-01 -2.23662660e-01 5.85829914e-02
-9.99978483e-01 -7.76795387e-01 -7.30204642e-01 -7.01645553e-01
6.91857040e-01 3.48016441e-01 7.63863027e-01 -8.47297627e-03
-1.95884332e-01 2.62550533e-01 -9.27540600e-01 -5.38578033e-01
-8.77592713e-02 4.73712608e-02 1.87275782e-01 1.36370108e-01
8.28315079e-01 -5.03311574e-01 -4.68433321e-01 -6.70453459e-02
-1.19603193e+00 -4.73272920e-01 2.39170015e-01 9.12700057e-01
3.32610935e-01 6.37350976e-02 4.34416234e-01 -6.17417336e-01
1.05327451e+00 -7.82413065e-01 -5.17223537e-01 2.52454311e-01
-1.00111127e+00 2.83444971e-01 6.51931822e-01 -8.26781034e-01
-7.50694692e-01 -4.06992584e-01 -4.06965688e-02 -2.87986547e-01
1.86282232e-01 8.09741676e-01 3.46527368e-01 2.47706234e-01
4.19448853e-01 2.42102504e-01 -1.35712564e-01 -4.96204793e-01
4.24876899e-01 7.65982509e-01 -1.01902977e-01 -4.25195485e-01
6.59698129e-01 2.78407097e-01 -3.23330641e-01 -1.10666335e+00
-4.32737857e-01 -5.97447395e-01 -6.03675783e-01 -1.12066604e-01
8.70498657e-01 -3.72499228e-01 -1.85059067e-02 2.81334966e-02
-1.23784566e+00 4.80635986e-02 -2.41728365e-01 7.63185203e-01
-1.30262196e-01 3.95246565e-01 -1.04697421e-01 -8.58567238e-01
-2.19497621e-01 -7.93966532e-01 4.52645659e-01 3.16769153e-01
-6.15260661e-01 -1.56434953e+00 4.48830426e-01 -3.08867037e-01
5.30985951e-01 3.64815414e-01 1.01126826e+00 -7.23271966e-01
-7.24074198e-03 -4.53927457e-01 2.53380537e-01 2.30789155e-01
4.82199848e-01 1.79615244e-01 -7.73185909e-01 -1.71720162e-01
-3.11000682e-02 4.10026647e-02 5.46782434e-01 -6.63525425e-03
1.04747188e+00 -1.67136893e-01 -3.30386788e-01 1.87814251e-01
1.54757810e+00 5.55254579e-01 5.46072245e-01 6.84959650e-01
1.28970817e-01 6.31586015e-01 8.02946985e-01 6.85429096e-01
2.61414170e-01 6.55446947e-01 1.22596761e-02 2.13222861e-01
2.87450701e-01 -1.25001475e-01 3.00539374e-01 1.33922267e+00
-1.54639989e-01 -4.54957753e-01 -1.05290985e+00 1.23319697e+00
-1.74714339e+00 -9.26474512e-01 1.71004415e-01 2.26145530e+00
7.36303270e-01 1.86549410e-01 -1.95346400e-01 3.80917013e-01
3.83763224e-01 7.43380666e-01 -1.35777444e-01 -9.99503136e-01
2.57746100e-01 4.86859649e-01 3.12414408e-01 3.94625157e-01
-7.41529763e-01 9.66399431e-01 5.46569490e+00 4.64942813e-01
-1.32751477e+00 4.08316404e-01 -3.10642775e-02 1.58266816e-02
-4.79160190e-01 1.44976825e-01 -4.85692322e-01 4.32651043e-01
1.27880001e+00 -5.89827478e-01 2.57273853e-01 5.71866989e-01
2.17603400e-01 -5.25893159e-02 -9.24254060e-01 9.53945577e-01
3.28995109e-01 -1.03938043e+00 -1.11627560e-02 -7.09827691e-02
2.78201550e-01 -2.51313657e-01 -3.83756049e-02 1.76367119e-01
-2.72117108e-02 -6.73490882e-01 5.87097764e-01 4.66935545e-01
4.83845770e-01 -9.04933035e-01 8.20849180e-01 1.89232647e-01
-1.29866612e+00 7.67268166e-02 -6.22958302e-01 -2.10829422e-01
2.61386603e-01 3.59290808e-01 -7.91465819e-01 5.44636846e-01
5.52660286e-01 6.35377467e-01 -4.04953748e-01 7.69284904e-01
-3.51814121e-01 6.87249482e-01 -1.50738090e-01 -6.34685695e-01
5.63478053e-01 -3.14184308e-01 3.05720389e-01 1.23393106e+00
7.01190770e-01 -4.33544852e-02 -1.08323142e-01 4.43064392e-01
2.83672392e-01 5.79370618e-01 -1.00006211e+00 -7.28899240e-01
5.22109270e-01 1.00225687e+00 -8.09946418e-01 -1.57171637e-01
-7.09822059e-01 9.47543919e-01 2.21408293e-01 2.42496774e-01
-7.36265182e-01 -7.26472735e-01 8.94502819e-01 1.27784073e-01
5.15222788e-01 -6.72982574e-01 4.77009229e-02 -1.01235414e+00
-1.43547617e-02 -3.95834267e-01 2.53695220e-01 -3.72006744e-01
-1.10026813e+00 7.83447385e-01 2.18445897e-01 -1.33416390e+00
-4.56286818e-01 -5.90299666e-01 -6.88118398e-01 6.74253643e-01
-1.39168358e+00 -8.14724863e-01 5.30758537e-02 3.03697079e-01
7.41228223e-01 -9.57668573e-02 1.03637087e+00 2.74907589e-01
-5.04772067e-01 4.90115792e-01 1.54996172e-01 -7.10913092e-02
4.31321353e-01 -1.03921854e+00 3.53103906e-01 9.87334013e-01
4.15181577e-01 9.89689767e-01 8.69156539e-01 -4.25933987e-01
-1.29727483e+00 -7.62073278e-01 1.49919713e+00 -1.24751039e-01
1.13652587e+00 -4.90975291e-01 -1.07562828e+00 3.72501493e-01
4.82824832e-01 -6.37284666e-02 1.20837998e+00 4.44713622e-01
-3.07047576e-01 -1.25368178e-01 -9.32425439e-01 8.16403866e-01
6.48484588e-01 -6.15908861e-01 -1.13218701e+00 1.63346633e-01
6.35831237e-01 2.41882682e-01 -9.20524359e-01 -1.19446136e-01
6.22669935e-01 -3.65440965e-01 7.60743320e-01 -5.85170925e-01
1.60563961e-01 -3.59079450e-01 -2.08359659e-01 -1.47302866e+00
-3.40859771e-01 -2.88618326e-01 -4.21544649e-02 1.47013366e+00
3.58438373e-01 -8.87669265e-01 2.78474808e-01 2.73078650e-01
1.31714121e-01 -8.71413171e-01 -1.08747053e+00 -1.00241888e+00
1.25354931e-01 -6.35995567e-01 7.98968375e-01 1.17251337e+00
-3.19960043e-02 9.09425840e-02 -3.10639828e-01 4.49464843e-02
9.56838131e-02 5.37572987e-02 4.93120074e-01 -1.26740742e+00
1.95222460e-02 -4.42653090e-01 -7.37817764e-01 -4.93542224e-01
2.90446848e-01 -7.66829550e-01 -1.25001788e-01 -1.54325485e+00
-2.48182282e-01 -3.24078292e-01 -6.44075632e-01 3.85793984e-01
-7.32703507e-02 2.66162828e-02 9.58208144e-02 2.43360121e-02
2.45601702e-02 6.49652004e-01 6.89067543e-01 -9.77849066e-02
-2.85908103e-01 -3.72716665e-01 -5.62567055e-01 6.57418132e-01
1.10932279e+00 -7.26077497e-01 -6.86757803e-01 -4.83193487e-01
3.28180224e-01 -5.06447375e-01 -2.08474621e-02 -7.15247333e-01
6.64920211e-02 -3.22245270e-01 2.07220414e-03 -3.55588078e-01
3.37970346e-01 -8.47143710e-01 1.25818908e-01 4.97334391e-01
7.29831727e-03 6.89523458e-01 2.66173124e-01 6.17690027e-01
-3.86476755e-01 -7.02157795e-01 4.64656353e-01 -7.19472766e-02
-9.34439242e-01 -1.72482859e-02 -5.16564846e-01 -1.24320731e-01
1.26989925e+00 -3.67573798e-01 1.98742241e-01 -8.62343609e-02
-2.63356805e-01 -6.20629564e-02 5.37824750e-01 9.22410786e-01
7.06150472e-01 -1.50480843e+00 -3.46888781e-01 1.09223768e-01
5.09540081e-01 -5.51029384e-01 1.18263140e-01 4.94414389e-01
-5.47898650e-01 3.64968568e-01 -1.20662659e-01 -2.35570267e-01
-1.21155739e+00 8.70483279e-01 -8.41028690e-02 -2.45922238e-01
-6.84857190e-01 5.23808122e-01 -1.86015949e-01 -1.51940793e-01
-1.23803899e-01 -3.88111323e-01 -7.28577554e-01 6.05568171e-01
6.30139709e-01 6.56029135e-02 -1.85074896e-01 -6.70539498e-01
-4.87501413e-01 6.86965108e-01 -1.16438106e-01 -2.86347151e-01
1.65311134e+00 1.38024949e-02 -5.48275374e-02 1.10216069e+00
1.05977070e+00 1.68354716e-02 -6.38681054e-01 -2.70119339e-01
4.11839247e-01 -7.62482882e-01 8.72950926e-02 -5.17401844e-02
-7.17140675e-01 9.25319195e-01 7.05371737e-01 4.42646086e-01
1.04644430e+00 -1.95371196e-01 8.54668796e-01 2.11323634e-01
4.07272846e-01 -1.28720796e+00 -8.69197696e-02 5.23139238e-01
7.31319308e-01 -8.85389209e-01 -2.71210279e-02 7.29278624e-02
-3.82782131e-01 1.33581567e+00 2.58532971e-01 -1.52263656e-01
8.01934063e-01 -1.85214326e-01 1.50872260e-01 -1.82889551e-02
-8.40267122e-01 -2.71712333e-01 2.10952967e-01 2.74743587e-01
7.94431746e-01 1.09655648e-01 -1.31283009e+00 2.38054186e-01
-1.13774128e-01 -2.11243525e-01 5.32970130e-01 1.37215817e+00
-2.49965265e-01 -1.63543761e+00 -1.48415908e-01 2.59090036e-01
-4.50758278e-01 -1.15833990e-01 -1.26074702e-01 9.50363696e-01
2.58436292e-01 9.96982098e-01 1.76544622e-01 -4.07680601e-01
3.47586066e-01 3.77486557e-01 2.96421319e-01 -6.75782859e-01
-3.75907093e-01 -2.04933524e-01 4.53866161e-02 -2.90597618e-01
-6.11750305e-01 -8.52494180e-01 -1.11613905e+00 -2.96376944e-01
-2.89648861e-01 3.77900183e-01 8.52052987e-01 8.78413796e-01
2.18410701e-01 3.70737582e-01 5.03571510e-01 -5.65013409e-01
-2.54609793e-01 -1.17465019e+00 -4.36536103e-01 4.13785607e-01
1.92311162e-03 -9.03901577e-01 -2.36095876e-01 -1.57098383e-01] | [10.413277626037598, 8.736639976501465] |
d92c7765-fbea-4851-8f80-3be09f2d7065 | automatic-counterfactual-augmentation-for | 2307.01214 | null | https://arxiv.org/abs/2307.01214v1 | https://arxiv.org/pdf/2307.01214v1.pdf | Automatic Counterfactual Augmentation for Robust Text Classification Based on Word-Group Search | Despite large-scale pre-trained language models have achieved striking results for text classificaion, recent work has raised concerns about the challenge of shortcut learning. In general, a keyword is regarded as a shortcut if it creates a superficial association with the label, resulting in a false prediction. Conversely, shortcut learning can be mitigated if the model relies on robust causal features that help produce sound predictions. To this end, many studies have explored post-hoc interpretable methods to mine shortcuts and causal features for robustness and generalization. However, most existing methods focus only on single word in a sentence and lack consideration of word-group, leading to wrong causal features. To solve this problem, we propose a new Word-Group mining approach, which captures the causal effect of any keyword combination and orders the combinations that most affect the prediction. Our approach bases on effective post-hoc analysis and beam search, which ensures the mining effect and reduces the complexity. Then, we build a counterfactual augmentation method based on the multiple word-groups, and use an adaptive voting mechanism to learn the influence of different augmentated samples on the prediction results, so as to force the model to pay attention to effective causal features. We demonstrate the effectiveness of the proposed method by several tasks on 8 affective review datasets and 4 toxic language datasets, including cross-domain text classificaion, text attack and gender fairness test. | ['Hao Xu', 'Yingji Li', 'Fausto Giunchiglia', 'Rui Song'] | 2023-07-01 | null | null | null | null | ['fairness', 'fairness', 'text-classification'] | ['computer-vision', 'miscellaneous', 'natural-language-processing'] | [ 4.86564547e-01 1.88571706e-01 -8.20321441e-01 -4.84756798e-01
-3.85989130e-01 -2.93807983e-01 7.99506545e-01 4.90538925e-01
-3.25892448e-01 7.89499760e-01 5.01616895e-01 -5.11187375e-01
-4.09202069e-01 -8.80746424e-01 -5.01057565e-01 -6.23893797e-01
6.78015500e-02 6.47421777e-02 -1.03734694e-01 -2.22211748e-01
4.91809636e-01 -3.85847129e-02 -1.29660344e+00 5.20325422e-01
1.12301886e+00 8.50174069e-01 -2.56879121e-01 4.53718565e-02
-4.00647402e-01 6.82454526e-01 -6.21075869e-01 -8.91064703e-01
-3.89314443e-03 -4.89416182e-01 -6.59543216e-01 -2.21769109e-01
1.02745056e-01 -7.54233524e-02 -9.58868042e-02 1.00159395e+00
4.36677128e-01 -1.49265721e-01 8.58799756e-01 -1.37336576e+00
-5.69117427e-01 1.36542022e+00 -8.03651094e-01 1.58248022e-01
3.28312278e-01 5.62885031e-02 1.51195502e+00 -7.72734344e-01
3.63964260e-01 1.55992985e+00 6.21566176e-01 4.67534661e-01
-1.18140435e+00 -1.31050587e+00 6.46674335e-01 2.42020696e-01
-9.80975747e-01 -2.03458369e-01 9.76839185e-01 -2.43873939e-01
5.14193952e-01 4.48889762e-01 5.47409475e-01 1.34833753e+00
4.44795191e-01 8.10082436e-01 1.10496128e+00 -4.13786173e-01
1.28750890e-01 2.64503241e-01 2.09290370e-01 4.85739052e-01
2.76503116e-01 -6.74591959e-02 -5.97588658e-01 -5.37506282e-01
-7.91150704e-02 1.88355193e-01 -2.33773530e-01 -2.34360993e-02
-7.36612678e-01 1.17135715e+00 3.18801373e-01 1.76580846e-01
-1.25908509e-01 -1.66825309e-01 5.38544893e-01 1.92490518e-01
7.72265732e-01 5.18775523e-01 -6.26497269e-01 2.38133445e-01
-8.56654882e-01 4.84210461e-01 5.69678783e-01 3.13399315e-01
5.38461387e-01 -4.08410370e-01 -3.92371982e-01 1.03047800e+00
4.13391262e-01 3.37504089e-01 6.81353033e-01 -2.14726791e-01
5.44089913e-01 9.81882334e-01 -2.48786613e-01 -1.42313099e+00
-3.68293226e-01 -3.97558540e-01 -8.40287149e-01 -1.52507797e-01
3.20046246e-02 -2.72886157e-01 -6.31202400e-01 1.88910556e+00
3.66461426e-01 1.08842403e-01 -2.27759540e-01 7.70613611e-01
6.69141233e-01 3.84796649e-01 4.40155596e-01 -6.00830972e-01
1.22650743e+00 -3.65026981e-01 -8.98143291e-01 -1.76226333e-01
9.02682304e-01 -6.38613760e-01 1.28187263e+00 4.91517603e-01
-4.44020003e-01 -1.87080219e-01 -1.22461772e+00 3.36329609e-01
-2.51593918e-01 -3.57674330e-01 9.07824814e-01 6.95927978e-01
-1.29097819e-01 7.49535322e-01 -9.32647735e-02 -6.20491467e-02
5.97849667e-01 2.36885250e-01 -9.19867605e-02 -1.10537419e-02
-1.75234604e+00 6.70468211e-01 4.23498809e-01 -7.45599419e-02
-5.45047879e-01 -8.68122935e-01 -5.80281317e-01 5.60232438e-02
7.08696842e-01 -5.23256421e-01 8.39708626e-01 -1.14155650e+00
-1.08604026e+00 5.07264912e-01 2.04929058e-02 -4.85435873e-01
5.44291675e-01 -3.12473089e-01 -4.94127125e-01 -2.05408424e-01
1.49736166e-01 3.04899216e-01 1.09466314e+00 -1.11465144e+00
-6.50234342e-01 -4.48925585e-01 1.04458801e-01 1.74686849e-01
-7.58086205e-01 6.04601614e-02 7.87702054e-02 -9.45595324e-01
-1.05131343e-01 -8.10931265e-01 -2.29860574e-01 -4.41351503e-01
-7.22277164e-01 -5.79850972e-01 8.01166654e-01 -2.20800281e-01
1.62104845e+00 -2.09903216e+00 -2.15980783e-01 3.20167661e-01
2.39734873e-01 1.29796103e-01 7.53648654e-02 2.70723015e-01
-3.56608629e-01 6.20224833e-01 -2.71977991e-01 -1.02029108e-01
-2.20313936e-01 3.56996730e-02 -8.94369066e-01 2.75628895e-01
1.73563138e-01 4.99802977e-01 -7.89643764e-01 -6.77296817e-01
-2.32146289e-02 1.50974765e-02 -7.44821966e-01 2.41684690e-02
-3.48644406e-01 -1.56593382e-01 -6.36707366e-01 3.73135358e-01
5.69573939e-01 9.63041708e-02 3.53612870e-01 -1.53616324e-01
6.18932908e-03 4.85778958e-01 -1.15620005e+00 1.02195728e+00
-4.14385647e-01 1.04067557e-01 -3.18577945e-01 -1.05233181e+00
9.81124640e-01 1.88842222e-01 4.55554217e-01 -4.71529961e-01
2.79259950e-01 1.89638436e-01 2.04083070e-01 -4.93493259e-01
1.54531986e-01 -4.89950150e-01 -1.36146858e-01 5.84293604e-01
-3.11806411e-01 1.34528540e-02 -6.95631951e-02 4.71525282e-01
9.30113256e-01 -2.04693601e-01 3.55936348e-01 -2.03277856e-01
5.28291702e-01 -1.56356514e-01 8.20340335e-01 7.98008323e-01
-1.65233985e-02 2.89821327e-01 8.76698077e-01 -4.17117327e-01
-6.08570576e-01 -6.03116572e-01 -2.41302058e-01 1.26280057e+00
8.88175592e-02 -7.56885171e-01 -5.57922661e-01 -1.19722331e+00
1.98351070e-01 1.05840790e+00 -1.01056778e+00 -6.19212747e-01
-2.26927891e-01 -1.16614926e+00 5.66021383e-01 1.09779760e-01
3.26649636e-01 -1.12238228e+00 -3.18779349e-01 7.44149089e-02
-1.66505441e-01 -5.37288129e-01 -3.84477049e-01 2.83629686e-01
-5.92586219e-01 -1.31905091e+00 -1.00950316e-01 -2.49425590e-01
4.65867728e-01 1.47203818e-01 7.77890384e-01 2.21382573e-01
-7.02789426e-02 -2.30545014e-01 -5.65627813e-01 -7.07561553e-01
-2.76123822e-01 5.33550866e-02 2.78465122e-01 3.12789768e-01
4.28768486e-01 -4.62219983e-01 -5.33791125e-01 1.48474231e-01
-8.28693330e-01 1.57215865e-03 5.92841685e-01 1.03439236e+00
2.70699829e-01 1.93049148e-01 9.70776260e-01 -1.45376730e+00
1.05190897e+00 -7.53027380e-01 -4.01344113e-02 2.55539984e-01
-1.08783531e+00 1.25172049e-01 7.85826564e-01 -6.20987535e-01
-1.18462884e+00 -1.70902014e-01 -1.93784498e-02 -2.05212936e-01
2.12070346e-01 7.68624008e-01 -4.15048599e-01 4.43075299e-01
6.93046153e-01 3.68044106e-03 -2.02371448e-01 -1.73976287e-01
4.39986080e-01 7.74349332e-01 -1.10283844e-01 -4.83679950e-01
7.83831120e-01 3.64382476e-01 -2.27756113e-01 -4.59892571e-01
-1.27816617e+00 -2.27898225e-01 -3.32390159e-01 -1.39466837e-01
5.75861514e-01 -4.60645288e-01 -6.63733423e-01 3.70979458e-02
-1.04872572e+00 8.22649524e-02 9.06684697e-02 3.26094955e-01
-4.14900370e-02 2.55858898e-01 -2.07977057e-01 -9.91356373e-01
-4.52140421e-01 -8.21798086e-01 5.97102523e-01 4.52682227e-02
-6.12205148e-01 -7.08863854e-01 -3.10789421e-03 2.37046912e-01
-2.92673279e-02 -3.12227663e-02 1.38584948e+00 -1.16166317e+00
-3.46074179e-02 -2.83535898e-01 -1.66494489e-01 6.58290088e-02
1.10798538e-01 1.20264009e-01 -9.74907517e-01 6.70632394e-03
6.25924841e-02 -4.27960217e-01 1.04577422e+00 3.08707088e-01
1.47021544e+00 -7.23112226e-01 -5.12651742e-01 2.39572182e-01
9.23584819e-01 9.82978493e-02 3.23034674e-01 1.88629732e-01
6.81223154e-01 8.86693776e-01 7.73316681e-01 6.14319503e-01
1.38671577e-01 5.28386652e-01 4.37186629e-01 -4.15656269e-02
3.92364383e-01 -6.79733634e-01 2.66021967e-01 4.60494250e-01
3.08283210e-01 -2.17896223e-01 -6.83882892e-01 3.63459766e-01
-1.89758325e+00 -9.43477213e-01 -1.90082729e-01 2.07874417e+00
1.14344740e+00 6.69909000e-01 2.95913313e-02 5.14091671e-01
7.66525030e-01 3.32261741e-01 -4.56120729e-01 -5.69611251e-01
-1.51932716e-01 -5.23974970e-02 2.47086897e-01 3.87284726e-01
-9.80999589e-01 9.16816235e-01 5.60566616e+00 1.14162290e+00
-1.14669621e+00 4.63582436e-03 1.03621364e+00 -2.30441108e-01
-9.02233481e-01 1.00687414e-01 -7.27339149e-01 5.08230388e-01
4.77329165e-01 -3.53591681e-01 -6.76078126e-02 8.21625471e-01
4.92220044e-01 -3.39359976e-02 -1.00582409e+00 6.56067312e-01
4.26095650e-02 -1.12612581e+00 5.04194677e-01 8.06258097e-02
6.01941168e-01 -5.03192246e-01 5.75307794e-02 4.40603375e-01
2.63810933e-01 -1.05925083e+00 6.18135214e-01 2.31456697e-01
5.45599699e-01 -1.01634550e+00 7.92878091e-01 5.73942125e-01
-7.10062981e-01 -4.59911913e-01 -3.66999120e-01 -3.41739029e-01
-1.82974085e-01 1.01017559e+00 -1.01500964e+00 3.38541120e-01
5.15910208e-01 5.95460355e-01 -4.79375660e-01 6.51315212e-01
-4.07349825e-01 1.07231307e+00 2.68197246e-02 -4.29200828e-01
-3.89134022e-03 1.49009526e-01 5.01358271e-01 1.25446391e+00
1.27354357e-03 3.00254405e-01 1.83268294e-01 6.68991029e-01
-2.58053452e-01 7.13628709e-01 -7.75871813e-01 1.07584270e-02
5.59109330e-01 1.23869205e+00 -6.22016728e-01 -2.39844456e-01
-3.36881489e-01 5.35748303e-01 3.81339908e-01 1.32965147e-01
-8.60929489e-01 -3.09160531e-01 4.02229965e-01 2.21341312e-01
-1.69069543e-01 4.59754556e-01 -8.57106924e-01 -9.78606522e-01
-2.38686144e-01 -9.77439106e-01 8.12719464e-01 -3.24023306e-01
-1.52710235e+00 1.54627308e-01 -1.14351571e-01 -8.86100352e-01
3.44757847e-02 -2.93672949e-01 -8.99825692e-01 6.20417058e-01
-1.24067390e+00 -9.27770197e-01 1.68669850e-01 4.50451881e-01
5.57532251e-01 -9.51487049e-02 6.86740458e-01 1.35723665e-01
-7.44624019e-01 9.00319040e-01 -5.08564651e-01 1.30414758e-02
9.35531974e-01 -1.01905060e+00 -2.21887454e-01 6.91432476e-01
7.82659501e-02 9.20509279e-01 9.78875875e-01 -1.03887987e+00
-9.28846717e-01 -9.75806057e-01 1.19417346e+00 -4.20553684e-01
8.13105226e-01 -4.43345994e-01 -9.36330974e-01 2.92268246e-01
8.22839513e-02 -3.08025569e-01 9.80818987e-01 8.02599013e-01
-5.84938169e-01 -2.73299992e-01 -8.97775352e-01 8.31885278e-01
9.90956604e-01 -2.32703209e-01 -8.66140008e-01 3.06571722e-01
8.47691894e-01 2.75618672e-01 -2.65053034e-01 5.56595623e-01
5.58308005e-01 -7.22779810e-01 6.11886740e-01 -9.70123410e-01
1.03699577e+00 8.94159451e-03 1.19873963e-01 -1.41438985e+00
-3.44452977e-01 -5.06777227e-01 1.09116077e-01 1.52709997e+00
7.53657222e-01 -3.06344748e-01 5.32289445e-01 5.66649556e-01
7.30359703e-02 -9.65849042e-01 -7.18625546e-01 -3.07853878e-01
5.42935953e-02 -8.66404414e-01 5.91917217e-01 1.31889570e+00
4.74821091e-01 6.33761048e-01 -6.70424163e-01 -4.92018312e-02
5.02053201e-01 2.70165682e-01 5.66706359e-01 -1.30884480e+00
-1.38895839e-01 -5.44363916e-01 6.44141138e-02 -4.46987450e-01
4.58931595e-01 -9.41141486e-01 7.62239397e-02 -1.05811977e+00
6.11831069e-01 -5.29009640e-01 -4.46498334e-01 6.97412550e-01
-7.64382064e-01 6.54085949e-02 2.43517607e-02 1.33159384e-01
-6.08976841e-01 7.38452673e-01 1.05726588e+00 -2.85107434e-01
-1.97696418e-01 2.00081915e-01 -1.13813472e+00 9.43485081e-01
7.35836327e-01 -7.64591455e-01 -5.51155508e-01 1.88546553e-01
6.28289819e-01 -2.59499729e-01 2.29971960e-01 -2.90043861e-01
9.51508805e-03 -5.60252845e-01 2.88910538e-01 -2.37990171e-01
-2.27664053e-01 -6.61185682e-01 -3.03948522e-01 6.33456826e-01
-8.80288243e-01 -1.89164251e-01 -8.98589343e-02 7.40638375e-01
-4.22746129e-03 -3.54768693e-01 5.89820147e-01 2.74507124e-02
-3.23246151e-01 2.54135638e-01 -1.36668697e-01 2.23083064e-01
7.59857953e-01 5.86595237e-02 -1.49226457e-01 -2.60936350e-01
-3.77383083e-01 2.56871998e-01 -7.30235800e-02 7.54589140e-01
5.68441331e-01 -1.26404452e+00 -8.95067155e-01 1.59560636e-01
2.85880536e-01 -4.08752561e-01 8.17584321e-02 6.04083359e-01
3.33528042e-01 2.87522525e-01 2.48503685e-01 -2.13428676e-01
-1.36733055e+00 5.81962764e-01 7.95347393e-02 -4.06282485e-01
-3.55214000e-01 8.01140904e-01 4.26904678e-01 -3.34765613e-01
8.24612677e-02 -4.45718803e-02 -5.87409675e-01 5.44347286e-01
5.54978728e-01 3.97029743e-02 -1.79096758e-02 -1.43800080e-01
-4.00358707e-01 4.52260412e-02 -3.67258787e-01 -4.10493976e-03
1.18029571e+00 3.52801010e-03 -1.58540115e-01 6.33262038e-01
9.05151129e-01 4.17273849e-01 -8.80030453e-01 -2.06950411e-01
1.86150283e-01 -5.30514956e-01 8.14192519e-02 -9.97370303e-01
-7.62851119e-01 6.60566151e-01 2.53129542e-01 3.10341150e-01
8.59602094e-01 -1.08762339e-01 5.03551424e-01 1.01191439e-01
-9.52007845e-02 -1.23190951e+00 1.13463573e-01 4.16803777e-01
8.80711496e-01 -1.25710177e+00 1.98734030e-01 -5.43366969e-01
-6.66129589e-01 9.07599986e-01 7.02936471e-01 1.49487346e-01
6.72020733e-01 1.52297929e-01 -3.68041880e-02 -1.38181031e-01
-9.80903864e-01 1.71680868e-01 2.93146014e-01 6.19360544e-02
6.58593714e-01 1.73046380e-01 -9.69034314e-01 1.19269538e+00
-3.13646048e-01 -1.38563186e-01 2.75081009e-01 3.66440415e-01
-4.00671929e-01 -1.02058077e+00 -2.01464400e-01 8.53462875e-01
-7.42895365e-01 -3.85292709e-01 -7.37495899e-01 5.26170909e-01
4.03199196e-01 9.58896101e-01 -1.71736181e-01 -6.38792753e-01
2.48841345e-01 2.39936590e-01 2.83932593e-03 -6.25622272e-01
-5.31660438e-01 -6.98642284e-02 2.33611375e-01 -3.51083457e-01
-2.02066675e-01 -7.22406983e-01 -1.23671532e+00 -2.59521604e-01
-5.93591332e-01 3.63869816e-01 3.09534162e-01 1.39916658e+00
6.65352046e-02 4.93178099e-01 9.05989289e-01 -1.54988794e-02
-7.72603452e-01 -1.11417556e+00 -4.54062104e-01 8.07295799e-01
6.26921207e-02 -7.75704801e-01 -5.81453443e-01 -2.75807351e-01] | [9.811394691467285, 7.798450946807861] |
9d8298fa-aeb7-46c7-9809-b07c081debed | compound-tokens-channel-fusion-for-vision | 2212.01447 | null | https://arxiv.org/abs/2212.01447v1 | https://arxiv.org/pdf/2212.01447v1.pdf | Compound Tokens: Channel Fusion for Vision-Language Representation Learning | We present an effective method for fusing visual-and-language representations for several question answering tasks including visual question answering and visual entailment. In contrast to prior works that concatenate unimodal representations or use only cross-attention, we compose multimodal representations via channel fusion. By fusing on the channels, the model is able to more effectively align the tokens compared to standard methods. These multimodal representations, which we call compound tokens are generated with cross-attention transformer layers. First, vision tokens are used as queries to retrieve compatible text tokens through cross-attention. We then chain the vision tokens and the queried text tokens along the channel dimension. We call the resulting representations compound tokens. A second group of compound tokens are generated using an analogous process where the text tokens serve as queries to the cross-attention layer. We concatenate all the compound tokens for further processing with multimodal encoder. We demonstrate the effectiveness of compound tokens using an encoder-decoder vision-language model trained end-to-end in the open-vocabulary setting. Compound Tokens achieve highly competitive performance across a range of question answering tasks including GQA, VQA2.0, and SNLI-VE. | ['AJ Piergiovanni', 'Maxwell Mbabilla Aladago'] | 2022-12-02 | null | null | null | null | ['visual-entailment'] | ['reasoning'] | [ 2.81079486e-02 3.31377536e-02 1.20594010e-01 -3.19171727e-01
-1.51460195e+00 -6.36561036e-01 9.12127912e-01 2.23396510e-01
-4.71231401e-01 3.37904274e-01 4.47496861e-01 -4.04619157e-01
5.53388953e-01 -4.87559974e-01 -1.05821240e+00 -5.04190266e-01
4.86406714e-01 2.98168242e-01 -7.27085471e-02 -7.39697963e-02
1.08800501e-01 -1.09580830e-01 -1.40809369e+00 8.68880391e-01
5.91448665e-01 1.11120594e+00 2.26075605e-01 8.51642966e-01
-6.91754997e-01 1.00947428e+00 -5.19544780e-01 -8.84001017e-01
-1.30974412e-01 -4.28793997e-01 -1.08567929e+00 1.14468940e-01
8.12613130e-01 -5.84196150e-01 -4.68630612e-01 7.27947950e-01
6.27970099e-01 -1.60667133e-02 6.73920453e-01 -1.34858298e+00
-1.29369521e+00 5.64806640e-01 -6.14639401e-01 -7.17195794e-02
5.76064765e-01 2.11952567e-01 1.46506166e+00 -1.46361732e+00
5.71440697e-01 1.75078797e+00 1.06558494e-01 5.31290114e-01
-1.23521066e+00 -3.72839838e-01 3.53130817e-01 3.84430051e-01
-1.12782347e+00 -4.37715799e-01 5.53626597e-01 -4.02323306e-01
1.16566527e+00 8.98053572e-02 3.37057501e-01 1.13630223e+00
-9.38313305e-02 1.19755304e+00 5.56568205e-01 -4.77694541e-01
-1.53016821e-01 3.39724049e-02 1.84761226e-01 7.78903484e-01
-2.48627156e-01 -3.37107450e-01 -5.44842303e-01 -1.01852745e-01
4.11870420e-01 2.09952891e-01 -1.35393649e-01 -2.53491312e-01
-1.57223094e+00 1.02833295e+00 7.11168170e-01 -5.18692918e-02
-6.73077881e-01 6.62053227e-01 5.30529916e-01 2.40911067e-01
1.35539025e-01 1.73125193e-01 -5.15442342e-02 3.01915437e-01
-7.04519272e-01 2.85377085e-01 3.39875162e-01 1.09952974e+00
9.48241115e-01 -7.77077954e-03 -1.03892076e+00 1.04141688e+00
6.32940054e-01 7.45965600e-01 2.04850823e-01 -8.15065265e-01
8.40811014e-01 4.36043292e-01 2.21655238e-02 -4.10002708e-01
-2.57223262e-03 -6.12891205e-02 -5.58809578e-01 1.40005043e-02
2.44896516e-01 -1.22607157e-01 -1.37383258e+00 1.59839773e+00
2.42139310e-01 -1.98075294e-01 5.55533648e-01 9.82152879e-01
1.34577167e+00 9.32499349e-01 4.12127376e-01 4.50839788e-01
1.96784747e+00 -1.38054180e+00 -7.95574844e-01 -8.79476517e-02
4.77800906e-01 -9.64854717e-01 1.17843735e+00 -1.84274554e-01
-1.43545401e+00 -7.17978537e-01 -7.90981412e-01 -7.82935560e-01
-4.30726200e-01 1.78341344e-01 2.01180249e-01 -8.82337391e-02
-1.19380236e+00 -1.57796532e-01 -4.84769732e-01 -2.08206370e-01
4.92753178e-01 9.61336195e-02 -3.99580419e-01 -2.57881612e-01
-1.14164591e+00 8.73219907e-01 9.10118371e-02 2.00527608e-01
-1.23625326e+00 -3.99289101e-01 -1.34463346e+00 1.79489300e-01
1.85026333e-01 -1.12012482e+00 1.52291644e+00 -9.18813467e-01
-1.08414459e+00 9.65645015e-01 -7.00305045e-01 -5.26164889e-01
9.77111161e-02 -1.22727200e-01 -1.20791540e-01 5.05831361e-01
4.58808541e-01 1.26070559e+00 1.03102982e+00 -1.30018306e+00
-6.15778506e-01 -1.71900943e-01 3.04210365e-01 3.84016037e-01
-3.86573896e-02 -5.29602021e-02 -1.08474636e+00 -5.00513911e-01
-1.22579329e-01 -5.83964050e-01 3.87901142e-02 -7.62919150e-03
-5.93355477e-01 -5.06922007e-01 8.06692362e-01 -7.51899540e-01
7.47067750e-01 -2.20464563e+00 4.21893984e-01 -1.76721606e-02
3.74507546e-01 -2.18889229e-02 -6.37307346e-01 6.46840036e-01
-4.47143912e-02 -3.59232612e-02 -1.91895694e-01 -8.42421174e-01
3.05297017e-01 1.41167209e-01 -6.53296053e-01 1.09106794e-01
7.27561474e-01 1.54839265e+00 -7.66275764e-01 -6.28076077e-01
1.79432496e-01 7.09402084e-01 -5.32945871e-01 4.70596373e-01
-5.14147222e-01 5.88728189e-02 -4.47519422e-01 1.01466799e+00
5.18939734e-01 -4.29242730e-01 -2.83486873e-01 -4.08122569e-01
1.52933016e-01 2.38852426e-01 -4.58267242e-01 1.89274263e+00
-6.63139641e-01 9.16714966e-01 5.25346510e-02 -8.48320544e-01
7.65609622e-01 5.00713170e-01 -1.34853572e-02 -1.03726614e+00
-9.39259771e-03 -7.80163929e-02 -5.00615835e-01 -8.39398623e-01
8.21030200e-01 -9.86908525e-02 -3.18227381e-01 3.34295332e-01
5.20112872e-01 -3.16352062e-02 4.40611504e-02 6.59319162e-01
6.27976537e-01 9.62408409e-02 -1.53902501e-01 5.66138148e-01
4.16194022e-01 -8.78509507e-03 -2.60911345e-01 6.32849395e-01
6.23698439e-03 7.85000920e-01 7.41600335e-01 1.80708002e-02
-1.04653633e+00 -1.37780631e+00 3.17876369e-01 1.42281365e+00
1.96460728e-02 -3.93964350e-01 -5.73092937e-01 -7.85595834e-01
2.08467662e-01 7.39645481e-01 -8.11954260e-01 -2.01188214e-02
-1.03189372e-01 -8.30601156e-02 7.02364981e-01 8.17109227e-01
3.95670146e-01 -1.33457589e+00 -4.21010315e-01 5.78791350e-02
-6.15544438e-01 -1.26875174e+00 -6.99990690e-01 5.88205718e-02
-2.83482045e-01 -7.88706183e-01 -1.27537632e+00 -1.11543083e+00
5.43365121e-01 4.27592993e-01 1.19703734e+00 5.45696430e-02
-1.00406982e-01 8.98742676e-01 -4.02366459e-01 -3.28179538e-01
-1.96479738e-01 -2.15697691e-01 -6.92762315e-01 3.95297855e-01
3.48410666e-01 6.91879317e-02 -7.10976720e-01 -1.88822433e-01
-9.72897768e-01 1.00767529e-02 7.13897467e-01 1.06579447e+00
8.64723027e-01 -1.10775816e+00 6.34803355e-01 -3.68669271e-01
7.25490630e-01 -5.65576375e-01 -4.13794845e-01 5.41673303e-01
1.59589812e-01 3.89174372e-01 2.87627757e-01 -4.64584231e-01
-8.98411989e-01 1.19655415e-01 -2.05805048e-01 -9.88850296e-01
-1.45372272e-01 5.26764452e-01 -1.31615520e-01 4.51011240e-01
2.75209278e-01 1.26590714e-01 -1.22709587e-01 -3.26991409e-01
1.14612997e+00 7.64777303e-01 9.20949280e-01 -3.50877017e-01
4.80420023e-01 5.16681671e-01 -4.81747359e-01 -7.60782838e-01
-5.74179709e-01 -6.46108508e-01 -2.15387493e-01 -1.35106355e-01
1.70570672e+00 -1.06273973e+00 -8.83885384e-01 2.74415284e-01
-1.64916813e+00 -1.52077833e-02 -1.67290017e-01 4.19458002e-01
-3.85947376e-01 4.48804498e-01 -4.74046856e-01 -8.01467359e-01
-5.78152061e-01 -1.36438954e+00 1.66133678e+00 1.44080505e-01
7.54955485e-02 -8.01137328e-01 -7.85628185e-02 7.39867389e-01
2.33096629e-01 -3.08348443e-02 1.03350449e+00 -5.56055903e-01
-5.71428120e-01 -1.71477012e-02 -6.74998164e-01 1.15355015e-01
-3.57496619e-01 -1.65638924e-01 -1.25394487e+00 -1.04632653e-01
-5.44164121e-01 -8.59764755e-01 1.47464108e+00 3.08842450e-01
1.08322608e+00 1.46881621e-02 -1.76364526e-01 3.99262786e-01
1.21511233e+00 1.83177888e-02 6.44854784e-01 1.93533320e-02
1.08490050e+00 7.04505265e-01 2.89538205e-01 1.40708178e-01
9.86103415e-01 5.93219995e-01 8.50225270e-01 -5.50770998e-01
-2.93953329e-01 -3.80939454e-01 4.29380596e-01 5.39597750e-01
3.17612350e-01 -3.17610413e-01 -8.65613580e-01 9.87968922e-01
-1.88456857e+00 -1.06101751e+00 1.97748188e-03 1.75989771e+00
5.50195575e-01 -4.89777446e-01 6.90477118e-02 -4.02126342e-01
6.47003531e-01 2.15537891e-01 -4.56545204e-01 -5.68863869e-01
-4.24158014e-02 2.19117194e-01 2.34112039e-01 6.52355313e-01
-1.03682053e+00 1.06509101e+00 5.56086636e+00 5.07315874e-01
-1.01165199e+00 2.96705872e-01 3.34105343e-01 -1.71157360e-01
-7.04632878e-01 1.72026753e-01 -5.87620080e-01 4.44984883e-02
6.75316870e-01 3.94074827e-01 2.14906648e-01 5.68454027e-01
-2.52338260e-01 -8.65070745e-02 -1.26675987e+00 1.19343948e+00
5.23803592e-01 -1.47494364e+00 5.42310655e-01 -2.41326466e-01
4.66847599e-01 9.53439474e-02 2.71868229e-01 5.56106865e-01
1.47983491e-01 -1.25098717e+00 1.12996566e+00 6.85060561e-01
9.65506375e-01 -5.82832456e-01 6.70475245e-01 -2.07110822e-01
-1.27299643e+00 -5.88427112e-03 -3.00191820e-01 4.22588140e-01
4.70532358e-01 -5.80894761e-02 -8.89502108e-01 7.39694059e-01
6.18070841e-01 4.78915453e-01 -5.75470567e-01 8.82780313e-01
-1.12534165e-01 2.15245381e-01 1.44090325e-01 -3.68738770e-02
5.48769772e-01 2.33169615e-01 1.78652659e-01 1.32080138e+00
1.39502063e-01 -3.66212368e-01 -1.19788706e-01 1.10451019e+00
-4.33990657e-01 -7.74441659e-02 -5.22118807e-01 -1.99776947e-01
3.23968112e-01 1.13350344e+00 -1.46628290e-01 -7.57270634e-01
-9.98696446e-01 1.45345056e+00 4.50741798e-01 9.30861473e-01
-9.62180078e-01 -6.86436832e-01 3.62737060e-01 -3.54235560e-01
8.53693962e-01 -1.33510865e-02 7.49752522e-02 -1.23640084e+00
1.28685445e-01 -8.21338654e-01 4.96610969e-01 -1.27157414e+00
-1.50848007e+00 6.93667948e-01 6.64574429e-02 -1.07670319e+00
-3.12551975e-01 -5.95727980e-01 -5.33019364e-01 1.17254281e+00
-1.88683510e+00 -1.78883338e+00 -3.22307825e-01 1.01409364e+00
6.73813224e-01 -4.24876288e-02 7.47502863e-01 1.92349628e-01
-2.99041182e-01 6.69744670e-01 -2.92944670e-01 3.95908922e-01
8.69759858e-01 -1.23954582e+00 4.80228066e-01 6.50155962e-01
4.45918471e-01 4.28804100e-01 2.71471232e-01 -3.34048003e-01
-1.49052155e+00 -1.11986315e+00 1.06402910e+00 -4.01709646e-01
5.46795666e-01 -6.09999239e-01 -1.00525820e+00 8.79032552e-01
1.02218461e+00 1.15654627e-02 5.07060349e-01 -1.37759179e-01
-8.62276137e-01 1.68107912e-01 -4.98115659e-01 5.33220291e-01
2.85391122e-01 -1.16544151e+00 -9.11589801e-01 1.60862833e-01
9.27554309e-01 -2.63434857e-01 -4.79760200e-01 3.14000212e-02
5.45681834e-01 -4.83312547e-01 1.09251785e+00 -7.39914417e-01
6.21643841e-01 -3.92587841e-01 -4.08211559e-01 -1.08185434e+00
-1.18017122e-02 -3.13562185e-01 9.86190215e-02 1.24178779e+00
7.58036315e-01 -2.72424906e-01 1.08642623e-01 2.85401046e-01
-1.96168542e-01 -5.88994384e-01 -8.79131436e-01 -1.82794914e-01
1.71174988e-01 -3.84043306e-01 5.08086622e-01 6.76658809e-01
9.89185423e-02 8.77481997e-01 -3.88415813e-01 5.78524433e-02
4.95222777e-01 3.24476421e-01 8.12468469e-01 -5.30374944e-01
-2.46991172e-01 -5.54805040e-01 -6.30097836e-02 -1.41497779e+00
2.58749366e-01 -1.13959229e+00 2.15980306e-01 -2.06446815e+00
4.21779066e-01 2.38147289e-01 -1.18851937e-01 7.21128285e-01
-4.54241544e-01 3.44662040e-01 5.12922645e-01 1.19262464e-01
-9.34564769e-01 7.16265738e-01 1.32354832e+00 -5.70984900e-01
-1.21894330e-02 -6.38398767e-01 -7.17289686e-01 2.66436070e-01
4.10801530e-01 6.87198201e-03 -2.70869255e-01 -8.82301629e-01
-7.22651137e-03 3.07731569e-01 9.59680736e-01 -3.23903650e-01
1.57373160e-01 2.72184014e-01 5.00829816e-01 -1.03989673e+00
6.36284828e-01 -5.46199620e-01 -5.01706362e-01 3.00725899e-03
-6.49734795e-01 2.74763077e-01 2.00238064e-01 4.61253703e-01
-6.22946382e-01 8.01311880e-02 3.12337041e-01 4.36031930e-02
-7.47085750e-01 1.81147963e-01 -4.35745180e-01 4.69052531e-02
9.34779048e-01 -2.41790735e-03 -6.87270820e-01 -8.55472326e-01
-8.24489057e-01 8.11855018e-01 7.64658824e-02 6.35445774e-01
1.05207098e+00 -1.55149531e+00 -8.26370299e-01 7.16252625e-02
5.43518960e-01 -5.24882227e-02 4.85116005e-01 9.14558530e-01
-2.58617342e-01 6.37018442e-01 -1.79018587e-01 -8.30213189e-01
-1.29429841e+00 8.08959365e-01 3.53038728e-01 -8.73309746e-03
-3.37168992e-01 8.58090043e-01 6.57689095e-01 -1.24962732e-01
4.46316838e-01 -5.10211945e-01 -3.04601252e-01 3.63877058e-01
5.07349253e-01 -1.80914611e-01 -1.89193010e-01 -8.04121494e-01
-4.47798699e-01 5.92434525e-01 -1.08927421e-01 -5.68899393e-01
7.95152724e-01 -3.28971505e-01 -1.03025958e-02 3.87545437e-01
1.52788222e+00 -1.10934041e-01 -1.18160737e+00 -3.36911529e-01
-3.78274173e-01 -1.05939433e-03 -1.45564094e-01 -6.27359569e-01
-1.07667470e+00 1.52220416e+00 3.71379942e-01 5.23666255e-02
1.02055740e+00 7.52465248e-01 7.36473262e-01 3.12484503e-01
-4.19677436e-01 -6.94136322e-01 4.89622265e-01 5.83581865e-01
1.14996600e+00 -1.30902100e+00 -5.63666463e-01 -2.31198706e-02
-1.24458826e+00 9.20071602e-01 6.11755013e-01 1.08588740e-01
1.10941254e-01 -1.09648243e-01 3.77016038e-01 -3.52355301e-01
-1.01246345e+00 -7.68756330e-01 5.42332351e-01 5.95720112e-01
4.65106070e-01 -1.24760941e-02 1.77959174e-01 5.54569423e-01
2.21143231e-01 -4.16597903e-01 7.29221851e-02 7.35779881e-01
-3.07075471e-01 -7.97580600e-01 -6.13473356e-01 1.41862214e-01
-4.36260968e-01 -4.71894532e-01 -4.66340959e-01 5.73405743e-01
-8.19488093e-02 1.14163351e+00 4.04182374e-01 -3.10238212e-01
4.49680835e-01 4.01947141e-01 3.91173989e-01 -4.89460677e-01
-6.23724699e-01 2.58211792e-01 1.23475060e-01 -5.50053895e-01
-3.09009075e-01 -3.58403653e-01 -1.43467224e+00 2.17092484e-01
-2.54395872e-01 5.69823617e-03 6.24569535e-01 9.88126457e-01
4.73064959e-01 6.82328343e-01 2.03639224e-01 -7.64030099e-01
-3.17895710e-01 -9.52391148e-01 4.78500873e-02 6.04932070e-01
8.46123338e-01 -4.25776869e-01 -8.95314943e-03 2.83228308e-01] | [10.854939460754395, 1.6163209676742554] |
58361009-2784-4e90-996b-bf0218cf2813 | object-goal-visual-navigation-via-effective | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Du_Object-Goal_Visual_Navigation_via_Effective_Exploration_of_Relations_Among_Historical_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Du_Object-Goal_Visual_Navigation_via_Effective_Exploration_of_Relations_Among_Historical_CVPR_2023_paper.pdf | Object-Goal Visual Navigation via Effective Exploration of Relations Among Historical Navigation States | Object-goal visual navigation aims at steering an agent toward an object via a series of moving steps. Previous works mainly focus on learning informative visual representations for navigation, but overlook the impacts of navigation states on the effectiveness and efficiency of navigation. We observe that high relevance among navigation states will cause navigation inefficiency or failure for existing methods. In this paper, we present a History-inspired Navigation Policy Learning (HiNL) framework to estimate navigation states effectively by exploring relationships among historical navigation states. In HiNL, we propose a History-aware State Estimation (HaSE) module to alleviate the impacts of dominant historical states on the current state estimation. Meanwhile, HaSE also encourages an agent to be alert to the current observation changes, thus enabling the agent to make valid actions. Furthermore, we design a History-based State Regularization (HbSR) to explicitly suppress the correlation among navigation states in training. As a result, our agent can update states more effectively while reducing the correlations among navigation states. Experiments on the artificial platform AI2-THOR (i.e.,, iTHOR and RoboTHOR) demonstrate that HiNL significantly outperforms state-of-the-art methods on both Success Rate and SPL in unseen testing environments. | ['Xin Yu', 'Zi Huang', 'Lincheng Li', 'Heming Du'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['visual-navigation'] | ['robots'] | [-1.67896748e-01 -1.84939399e-01 -1.85469180e-01 -3.44087034e-01
-2.10678264e-01 -2.83996701e-01 6.39404416e-01 -2.62842178e-01
-6.55763924e-01 5.36193013e-01 2.41677642e-01 -4.29971546e-01
-1.28458634e-01 -5.87649405e-01 -6.69305384e-01 -7.93271840e-01
-1.62366390e-01 -5.41281234e-03 5.25063217e-01 -4.32345301e-01
3.52333665e-01 3.24084103e-01 -1.44376993e+00 -5.30882835e-01
9.13487792e-01 7.06678033e-01 6.40983641e-01 3.51515383e-01
7.91077837e-02 9.49561656e-01 -3.74255449e-01 4.19477433e-01
2.24927947e-01 -5.75674415e-01 -3.64770234e-01 2.52279304e-02
1.31184399e-01 -4.78446782e-01 -6.84515297e-01 1.07675993e+00
4.22213346e-01 7.17702091e-01 4.25051093e-01 -1.29397476e+00
-4.59937453e-01 3.98065984e-01 -5.68612218e-01 2.54967928e-01
1.66599497e-01 6.70604527e-01 7.13088512e-01 -5.79613268e-01
5.65839529e-01 1.33732712e+00 2.27368146e-01 7.60759950e-01
-1.01853669e+00 -6.59969807e-01 1.07610345e+00 5.21180749e-01
-9.81148005e-01 -3.57112378e-01 7.72563040e-01 -2.05331638e-01
8.07887554e-01 -1.21279806e-01 9.11430955e-01 1.23690784e+00
4.03885454e-01 1.01854324e+00 7.80165732e-01 -5.59160411e-02
6.48780406e-01 -1.72484681e-01 3.16565514e-01 1.06896281e+00
1.59566000e-01 4.90610570e-01 -6.64832652e-01 1.88489124e-01
7.58756220e-01 4.26683575e-02 -3.67732018e-01 -8.51929843e-01
-1.12985611e+00 7.91587472e-01 7.51082122e-01 -2.03926951e-01
-6.15326524e-01 3.43569785e-01 3.40520948e-01 -1.65889002e-02
-3.04569602e-01 3.66597772e-01 -2.34224975e-01 -2.03002959e-01
-1.84139729e-01 1.03927456e-01 2.85960704e-01 8.73965979e-01
7.67336965e-01 5.83540976e-01 -2.38491997e-01 5.61284065e-01
5.62086821e-01 5.31785071e-01 6.20111942e-01 -1.16397989e+00
3.12526107e-01 6.50540054e-01 4.13938105e-01 -7.66397476e-01
-6.20092690e-01 -3.67257118e-01 -5.71320593e-01 5.58518469e-01
-2.79577053e-03 -2.24829450e-01 -1.33152759e+00 2.11473632e+00
3.78677070e-01 2.38055304e-01 2.68395334e-01 1.02508450e+00
4.28104073e-01 8.82287920e-01 2.59676844e-01 -1.77161932e-01
9.41293955e-01 -1.15067053e+00 -9.11861122e-01 -1.02720356e+00
5.07340968e-01 -5.26337065e-02 1.34442985e+00 3.29922616e-01
-6.63998246e-01 -5.21918118e-01 -1.11076486e+00 3.05232823e-01
-1.21134408e-01 1.14394486e-01 8.37647915e-01 1.06822573e-01
-8.85517418e-01 2.95962572e-01 -1.41984761e+00 -4.24796075e-01
2.11951420e-01 1.76349774e-01 -2.05421597e-01 -9.07332543e-03
-1.14795101e+00 1.08578110e+00 4.69960451e-01 2.67367601e-01
-1.34840751e+00 -1.67508379e-01 -1.20534348e+00 -3.65039669e-02
8.34095418e-01 -3.84091437e-01 1.21695840e+00 -5.26105821e-01
-1.70446706e+00 -8.48703682e-02 -3.71974111e-01 -5.52097559e-01
2.49237493e-01 -4.47343141e-01 -2.76517391e-01 -2.25021884e-01
1.91615313e-01 8.74815404e-01 6.56878591e-01 -1.39556885e+00
-8.70011389e-01 -3.32292855e-01 1.32602856e-01 6.90843463e-01
-3.54039699e-01 -7.60097623e-01 -9.65035677e-01 -1.74866259e-01
4.90295589e-01 -1.33375430e+00 -7.32586324e-01 -3.83622795e-02
-2.82412738e-01 -3.06244075e-01 9.42729652e-01 -3.25287908e-01
1.27320158e+00 -2.06122828e+00 3.67954224e-01 1.48328304e-01
-6.40020743e-02 1.34490654e-01 -3.28347713e-01 2.09304631e-01
3.64240110e-01 -4.07822639e-01 -6.81208028e-03 -2.82082498e-01
-1.03564531e-01 5.96283495e-01 -3.58772635e-01 4.03893977e-01
-1.80402666e-01 7.82571197e-01 -1.03483784e+00 -1.58532098e-01
4.48444933e-01 3.11438441e-01 -5.81141829e-01 1.98483691e-01
-1.87227756e-01 5.75175941e-01 -7.38077104e-01 3.78089875e-01
2.51573622e-01 -3.19272012e-01 2.44801939e-01 -4.93035167e-02
-4.33033705e-01 3.60692352e-01 -1.03750312e+00 1.81722248e+00
-6.49238229e-01 4.85115170e-01 -1.56804323e-01 -6.60275161e-01
7.80944824e-01 -2.46261016e-01 1.75085112e-01 -1.11820078e+00
2.70523969e-02 -1.85994223e-01 5.62283210e-02 -3.06475312e-01
4.93484765e-01 4.61301833e-01 -4.32280898e-02 8.88134688e-02
-1.73611537e-01 2.27342591e-01 2.74401665e-01 3.47741991e-01
9.67640102e-01 4.33724880e-01 3.96597058e-01 -4.35262471e-02
4.89393353e-01 -3.62777710e-03 8.96835506e-01 1.03932571e+00
-5.86436570e-01 -9.70185250e-02 3.74452263e-01 -3.57489169e-01
-5.52135885e-01 -1.03496051e+00 3.08933049e-01 1.10878396e+00
1.00395525e+00 -4.62267637e-01 -2.64200449e-01 -7.17874110e-01
-1.79075614e-01 1.04829836e+00 -8.02167296e-01 -7.76354730e-01
-7.69375145e-01 -6.32452071e-01 -5.06735705e-02 7.45147705e-01
6.25084996e-01 -1.34243584e+00 -1.23133218e+00 2.10550755e-01
-9.76925623e-03 -7.55282879e-01 -4.73963618e-01 3.06610376e-01
-7.52605617e-01 -8.55153084e-01 -4.38329250e-01 -7.63392031e-01
7.14480281e-01 3.83295953e-01 6.75439537e-01 -1.12745814e-01
1.13786593e-01 4.18198109e-01 -2.24503115e-01 -2.84415167e-02
-5.50145283e-02 -2.07521450e-02 1.51071995e-01 -3.55241895e-01
3.13446894e-02 -2.67780185e-01 -7.81869352e-01 4.00032818e-01
-4.23253417e-01 3.21318686e-01 6.14703417e-01 1.00844312e+00
5.44275880e-01 6.95589930e-03 3.69890779e-01 -5.05277872e-01
5.47356367e-01 -3.63711566e-01 -1.02108169e+00 1.45561293e-01
-1.02177846e+00 4.10775691e-01 6.30817533e-01 -5.69346786e-01
-1.26382017e+00 1.23195790e-01 6.02222830e-02 -3.10354143e-01
4.75310348e-02 5.31537890e-01 -4.15804982e-02 9.19293463e-02
4.16830957e-01 6.38817966e-01 -1.17145449e-01 -1.12378836e-01
3.09374690e-01 -2.79033813e-03 4.87578034e-01 -2.22860277e-01
5.38306653e-01 4.65304524e-01 3.52696143e-02 -5.23455739e-01
-7.44311273e-01 -3.54386806e-01 7.50913695e-02 -3.85904402e-01
6.86049581e-01 -9.31493521e-01 -9.56894338e-01 4.44285393e-01
-6.12261474e-01 -7.40867794e-01 6.53969049e-02 4.36730087e-01
-4.44826722e-01 3.32915694e-01 -4.70957637e-01 -1.01616192e+00
-7.54899904e-02 -1.43298340e+00 6.15225732e-01 7.32850254e-01
5.98932169e-02 -8.09411108e-01 1.04006663e-01 -1.12774611e-01
3.37757379e-01 -2.37753198e-01 5.56452394e-01 -1.61698401e-01
-7.49898791e-01 1.06724380e-02 3.53260860e-02 -7.88313225e-02
6.51819780e-02 -3.06704491e-01 -4.15096223e-01 -4.95525986e-01
-2.23499864e-01 -3.22467059e-01 9.45519686e-01 4.93608028e-01
7.77022302e-01 -2.89200038e-01 -5.45549691e-01 5.45778990e-01
1.21766853e+00 8.88684213e-01 5.23311973e-01 8.34738791e-01
6.82314456e-01 4.90381390e-01 9.65437293e-01 5.43055534e-01
6.64163232e-01 6.68780208e-01 8.11482370e-01 2.22449809e-01
1.47706032e-01 -4.99927342e-01 6.77426755e-01 3.73990327e-01
2.94273227e-01 -3.86386693e-01 -8.22985530e-01 5.00707746e-01
-2.30390215e+00 -6.48630321e-01 2.83639044e-01 2.13090754e+00
4.38352674e-01 4.71778095e-01 -2.46198207e-01 -4.89214927e-01
4.33719009e-01 3.68292928e-01 -1.22077370e+00 1.20087989e-01
1.33955166e-01 -5.75440168e-01 4.63925093e-01 6.80947542e-01
-9.91176724e-01 1.36550331e+00 5.45751286e+00 4.54489887e-01
-1.13006127e+00 -2.50611931e-01 1.70032829e-01 -1.15519643e-01
-1.35535449e-01 1.07532255e-01 -1.03870189e+00 4.92898971e-01
6.42343760e-01 -1.20586678e-02 5.14817297e-01 1.23986530e+00
3.54675770e-01 -5.38362920e-01 -7.20582247e-01 7.38390803e-01
-6.73571005e-02 -1.01924825e+00 2.52837148e-02 -1.00901954e-01
6.02591395e-01 8.01943466e-02 2.09868923e-01 8.20217311e-01
6.79078400e-01 -4.40561444e-01 8.39272261e-01 4.20691192e-01
7.18593076e-02 -8.34194362e-01 4.68937397e-01 4.17820930e-01
-1.32974005e+00 -5.45314550e-01 -3.04579139e-01 1.43023610e-01
5.89527309e-01 -1.35428667e-01 -6.43453956e-01 2.58921564e-01
7.18271852e-01 8.87024045e-01 -6.46731615e-01 9.62018609e-01
-6.34539664e-01 3.97025406e-01 -2.24749148e-01 -3.21872503e-01
6.60120070e-01 -4.29718256e-01 6.31716311e-01 5.05464137e-01
1.75017968e-01 1.08342238e-01 5.34435868e-01 5.96126020e-01
5.23524523e-01 -4.05574322e-01 -4.49244112e-01 2.17316657e-01
5.81593931e-01 9.59673107e-01 -5.90148866e-01 -3.50096196e-01
-2.53344774e-01 9.28985059e-01 5.91179252e-01 6.31424367e-01
-8.97447228e-01 -1.44586280e-01 7.86135733e-01 -3.46498728e-01
3.40784192e-01 -4.80444431e-01 4.30582799e-02 -1.06360412e+00
-1.18362851e-01 -6.49549901e-01 2.35577643e-01 -8.74174297e-01
-6.83286190e-01 6.25532389e-01 -1.20825380e-01 -1.25630701e+00
-3.46298158e-01 -3.04637104e-01 -5.95707715e-01 6.28540337e-01
-1.53938866e+00 -7.11293519e-01 -4.69976038e-01 4.92782444e-01
8.90287936e-01 -6.79199770e-02 3.90226811e-01 -1.14200674e-01
-1.01972687e+00 1.72618359e-01 -1.55897737e-01 -1.65783152e-01
4.74848241e-01 -1.12488937e+00 5.53004980e-01 1.24483430e+00
-1.31489158e-01 9.84808624e-01 8.96710455e-01 -8.49845469e-01
-1.46327591e+00 -9.50200021e-01 4.47019637e-02 -1.19022891e-01
4.94287014e-01 -3.64597030e-02 -9.23498094e-01 7.93523252e-01
-1.16530068e-01 -2.02327102e-01 1.18806891e-01 3.30285132e-02
-1.06255449e-01 7.21337199e-02 -5.04262924e-01 1.39140546e+00
1.16739476e+00 -1.76127613e-01 -4.82671559e-01 -1.58755574e-02
6.97617710e-01 -4.33736444e-01 -1.73780601e-02 4.35655117e-01
5.17374456e-01 -9.63262081e-01 8.94561172e-01 -5.54973483e-01
-4.71499041e-02 -7.09860325e-01 1.46905705e-01 -1.43278337e+00
-5.64096093e-01 -3.47749531e-01 -3.97623807e-01 6.77979767e-01
3.15914989e-01 -6.38502479e-01 8.61115992e-01 5.46628654e-01
-3.26701462e-01 -7.17544377e-01 -7.89066911e-01 -7.42948711e-01
-5.49462616e-01 -1.92174643e-01 1.54589459e-01 4.19347852e-01
-1.17591873e-01 5.00648618e-01 -6.73394263e-01 5.69372118e-01
5.12170732e-01 1.72639862e-01 8.63171220e-01 -7.32602179e-01
5.58530260e-03 -3.59467149e-01 -1.21353365e-01 -1.53593016e+00
2.82239735e-01 -2.67270982e-01 6.23773694e-01 -1.80033410e+00
-2.91826297e-02 -3.46914828e-01 -5.37711740e-01 6.61451459e-01
-5.43141246e-01 -4.27800506e-01 1.22761935e-01 2.36795902e-01
-1.06247413e+00 1.06341827e+00 1.40490043e+00 7.02454224e-02
-6.36741459e-01 7.28660673e-02 -4.54363972e-01 7.13003814e-01
8.07641447e-01 -3.23748738e-01 -8.08563709e-01 -3.30820262e-01
1.99926391e-01 2.11210191e-01 3.02201778e-01 -1.17399848e+00
4.84343916e-01 -3.27550948e-01 1.99855745e-01 -8.02153170e-01
5.10227144e-01 -7.46942461e-01 -9.64996517e-02 8.95534813e-01
-5.05892396e-01 2.50019282e-01 2.18645766e-01 1.13144886e+00
5.72858229e-02 -6.78468272e-02 7.46572733e-01 -7.18473122e-02
-1.62768090e+00 3.78462523e-01 -7.40611017e-01 -1.40399083e-01
9.11439121e-01 -3.79319564e-02 -3.12744051e-01 -5.20306468e-01
-5.17871439e-01 9.80454504e-01 4.80717182e-01 6.95929945e-01
7.51785994e-01 -1.11104953e+00 5.60154580e-02 2.65602022e-01
2.18653724e-01 -1.22904740e-01 5.09306252e-01 6.31500423e-01
-2.28248641e-01 3.14906329e-01 -3.30494255e-01 -5.90522110e-01
-9.43860114e-01 4.96581316e-01 3.48507166e-01 -2.01952651e-01
-8.01185310e-01 8.23195636e-01 5.87962031e-01 -3.65237355e-01
5.61809123e-01 -1.50362253e-01 -5.73494792e-01 -1.29411295e-01
4.26448166e-01 3.04148406e-01 -3.66965294e-01 -2.30930716e-01
-4.97341067e-01 2.01476678e-01 -3.84916127e-01 -3.06492090e-01
1.01856637e+00 -4.67725635e-01 4.29027498e-01 4.78576183e-01
6.47080183e-01 -3.92446131e-01 -2.10836935e+00 -8.77623633e-02
-2.04652268e-02 -2.68401444e-01 2.93999195e-01 -8.43785465e-01
-1.00724447e+00 6.16739154e-01 8.64107907e-01 -2.29232922e-01
9.31809962e-01 -2.40095153e-01 6.43170714e-01 9.22769964e-01
5.21157503e-01 -1.01556456e+00 4.03447151e-01 8.73667419e-01
6.97967112e-01 -1.40902627e+00 -3.76563482e-02 4.16965187e-02
-1.13769019e+00 7.20959008e-01 1.34528768e+00 7.42539838e-02
2.46016800e-01 -5.37426546e-02 2.80836433e-01 -7.19096810e-02
-9.85883534e-01 -4.45843130e-01 1.37823910e-01 5.96034050e-01
-2.96723396e-01 -2.52246767e-01 3.03179622e-02 2.31393248e-01
1.49648637e-01 -3.44255924e-01 5.50211966e-01 1.34524500e+00
-7.93123245e-01 -6.42840385e-01 8.70025996e-03 1.57194078e-01
3.19921374e-01 3.96609940e-02 7.28299320e-02 8.15074861e-01
-3.45030576e-01 8.00085962e-01 -4.39005112e-03 -3.59146535e-01
5.18663108e-01 -1.97386786e-01 1.25546694e-01 -3.91193271e-01
-1.47039354e-01 1.72719866e-01 -6.34610802e-02 -9.66396153e-01
-7.10878447e-02 -5.62601507e-01 -1.49367511e+00 1.29702300e-01
-4.46734965e-01 6.95502087e-02 5.11749268e-01 7.48303533e-01
4.70558554e-01 8.83069038e-01 4.38360542e-01 -8.44030857e-01
-6.13486826e-01 -6.50158763e-01 -3.39496851e-01 1.48131460e-01
4.79623765e-01 -1.12493229e+00 -4.83803630e-01 -2.98984706e-01] | [4.47498083114624, 0.5094690322875977] |
6a0fef9c-0efb-4b37-ae44-7045d0a36fec | command-driven-articulated-object | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Chu_Command-Driven_Articulated_Object_Understanding_and_Manipulation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Chu_Command-Driven_Articulated_Object_Understanding_and_Manipulation_CVPR_2023_paper.pdf | Command-Driven Articulated Object Understanding and Manipulation | We present Cart, a new approach towards articulated-object manipulations by human commands. Beyond the existing work that focuses on inferring articulation structures, we further support manipulating articulated shapes to align them subject to simple command templates. The key of Cart is to utilize the prediction of object structures to connect visual observations with user commands for effective manipulations. It is achieved by encoding command messages for motion prediction and a test-time adaptation to adjust the amount of movement from only command supervision. For a rich variety of object categories, Cart can accurately manipulate object shapes and outperform the state-of-the-art approaches in understanding the inherent articulation structures. Also, it can well generalize to unseen object categories and real-world objects. We hope Cart could open new directions for instructing machines to operate articulated objects. | ['Jiaya Jia', 'Chi-Wing Fu', 'Xiaojuan Qi', 'Xiao Tan', 'Xiaoqing Ye', 'Zhengzhe Liu', 'Ruihang Chu'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['motion-prediction'] | ['computer-vision'] | [ 2.24078730e-01 7.28557184e-02 -4.98002052e-01 -3.92204374e-01
-1.76095009e-01 -8.62861514e-01 6.57485604e-01 -2.61885464e-01
1.70527980e-01 1.52236670e-01 4.06758636e-01 -4.41887200e-01
4.28235456e-02 -5.52553058e-01 -7.47369826e-01 -2.22767338e-01
-1.63916424e-01 8.59150827e-01 3.99215877e-01 -3.71518344e-01
5.02617478e-01 9.85641479e-01 -1.65671790e+00 5.92422843e-01
4.44161147e-01 5.08957386e-01 5.21678925e-01 8.59894156e-01
-1.47915751e-01 7.03126431e-01 -3.31829995e-01 -9.53207463e-02
1.97529808e-01 -5.91737702e-02 -9.07796681e-01 4.50731784e-01
5.42669833e-01 -5.86411655e-01 -3.00430357e-01 5.49574435e-01
2.66821772e-01 4.44308400e-01 8.96124125e-01 -1.24022245e+00
-9.54048395e-01 5.33953726e-01 -9.67728570e-02 -1.24519384e-02
5.89697003e-01 8.10073256e-01 1.03679919e+00 -8.49723279e-01
8.59638691e-01 1.80951595e+00 2.36749619e-01 8.27078402e-01
-1.23672581e+00 -3.20651054e-01 5.70784330e-01 3.36456388e-01
-9.52677429e-01 -4.83262450e-01 7.17778921e-01 -7.64271498e-01
1.05015683e+00 4.78075713e-01 9.12320912e-01 1.17640913e+00
-2.04822067e-02 1.26021850e+00 5.20448565e-01 -1.75331250e-01
1.16271123e-01 -6.44239664e-01 -1.08590648e-01 9.07559752e-01
-6.70192018e-02 1.87132493e-01 -4.12361860e-01 6.44403100e-02
1.22432375e+00 -5.99813536e-02 -1.69343144e-01 -6.50126517e-01
-1.70539594e+00 5.00187993e-01 5.50294578e-01 7.22844619e-04
-3.16329926e-01 5.36416411e-01 2.01300204e-01 -4.88621462e-03
-1.39796451e-01 9.32620466e-01 -7.17072248e-01 -3.69552314e-01
-4.21901107e-01 5.55871665e-01 7.75114119e-01 1.43303561e+00
3.07612360e-01 5.57540536e-01 -2.08991528e-01 2.99310982e-01
4.95696723e-01 4.92446601e-01 1.04989968e-01 -1.34728074e+00
4.29186553e-01 5.57381988e-01 7.67721236e-02 -6.96120739e-01
-3.53394568e-01 2.75069792e-02 -2.92065710e-01 4.40573633e-01
4.12390232e-01 3.07494074e-01 -1.18569827e+00 1.33075202e+00
4.60149795e-01 -4.94966023e-02 -3.63333762e-01 9.65639710e-01
5.32981753e-01 4.46677536e-01 1.99365422e-01 3.48380864e-01
1.24404359e+00 -1.04891551e+00 -7.18657374e-01 -1.97682336e-01
3.98063987e-01 -6.23203993e-01 1.53822684e+00 4.26474601e-01
-1.12756860e+00 -6.78233504e-01 -7.05581844e-01 -3.97296101e-01
-1.84716985e-01 5.94528974e-04 1.00915396e+00 4.09208722e-02
-5.84441304e-01 7.13056386e-01 -1.43621397e+00 -1.95703179e-01
6.96305990e-01 5.27437866e-01 -2.15044156e-01 1.18831776e-01
-3.97727430e-01 9.85177696e-01 4.00143504e-01 1.46680132e-01
-1.25079250e+00 -5.34503222e-01 -1.05097628e+00 -1.35259911e-01
5.42526245e-01 -1.07671821e+00 1.54566753e+00 -6.26016855e-01
-1.71683145e+00 6.36740625e-01 -2.47301042e-01 -4.51972662e-03
4.09162790e-01 -5.40185094e-01 2.33776212e-01 1.11175515e-01
-1.21579923e-01 1.20628965e+00 1.02009261e+00 -1.61911917e+00
-2.68842459e-01 -3.19215983e-01 1.40255123e-01 3.19154501e-01
1.19191766e-01 -1.44322172e-01 -4.82782513e-01 -9.89210784e-01
2.68941283e-01 -1.15174639e+00 -1.74320728e-01 6.29427195e-01
-7.03700244e-01 -3.81158799e-01 1.41798520e+00 -5.09802341e-01
8.14991713e-01 -1.88630354e+00 8.92227948e-01 -6.80400655e-02
4.53849286e-02 7.48563409e-02 -3.36497456e-01 5.24853110e-01
2.38069445e-01 1.79892272e-01 4.39559966e-02 -2.73139536e-01
1.06759496e-01 6.28004372e-01 -7.79329002e-01 2.34556973e-01
3.59173268e-01 1.38146937e+00 -9.61403131e-01 -3.97381365e-01
6.34468973e-01 2.48232931e-01 -9.24768388e-01 6.17048442e-01
-9.27043021e-01 9.33588207e-01 -6.63627148e-01 9.70483363e-01
2.18297653e-02 -1.14483036e-01 1.04696907e-01 -2.92823642e-01
6.17917441e-03 4.82194155e-01 -9.30034339e-01 1.70501864e+00
-2.00034142e-01 6.06917202e-01 7.06467927e-02 -7.46627212e-01
6.13525987e-01 1.55796036e-01 2.48996109e-01 -2.40663392e-03
6.93479478e-02 -3.29325527e-01 2.26637304e-01 -9.22940314e-01
2.98324764e-01 2.05634683e-01 -7.14555616e-03 2.66719759e-01
-2.17195377e-01 -9.53399301e-01 -1.30188137e-01 -1.17505547e-02
8.41365218e-01 7.76219010e-01 3.24722290e-01 -1.34800496e-02
8.88253562e-03 1.75093904e-01 7.55362809e-02 5.34962535e-01
-1.38883563e-02 4.53511834e-01 4.60781856e-03 -5.91884315e-01
-1.18155348e+00 -1.42007720e+00 1.55325755e-01 1.55065596e+00
6.12881966e-02 -2.64138669e-01 -4.33409870e-01 -5.23288846e-01
3.05198967e-01 9.40531015e-01 -4.66124296e-01 1.11054070e-01
-9.79291260e-01 1.91660434e-01 3.12487245e-01 1.14118934e+00
1.08738549e-01 -1.46138072e+00 -8.50297570e-01 1.95511635e-02
-1.76725595e-03 -1.13158333e+00 -7.72383213e-01 -7.10338308e-03
-1.06835973e+00 -1.11931181e+00 -2.26589471e-01 -8.30240190e-01
8.56050909e-01 1.86556861e-01 1.06441641e+00 4.83485967e-01
-3.92475694e-01 6.91219807e-01 -3.84694546e-01 -5.24413228e-01
-4.43894476e-01 7.77169093e-02 1.32600531e-01 -5.99633753e-01
-2.35909894e-01 -6.37953997e-01 -3.45627844e-01 5.21425664e-01
-6.26396537e-01 6.38003051e-01 5.46006382e-01 5.98885775e-01
4.81199831e-01 -4.87256229e-01 1.24924950e-01 -4.96811062e-01
2.16959581e-01 -2.23180234e-01 -4.37543362e-01 -3.83019745e-02
-7.14341849e-02 4.06740814e-01 6.28757358e-01 -9.64262545e-01
-8.96940410e-01 4.33322549e-01 -1.09387919e-01 -7.27394342e-01
-6.02468431e-01 1.35799170e-01 -3.16689640e-01 1.96914002e-01
3.36366802e-01 1.12889454e-01 -1.25883073e-01 -5.45351326e-01
9.67605889e-01 2.12375790e-01 8.18645179e-01 -1.05781066e+00
1.01413774e+00 5.43833196e-01 8.14200938e-02 -9.17824805e-01
-6.96725667e-01 -2.64085054e-01 -1.27155864e+00 -1.30574897e-01
1.15467942e+00 -4.25824016e-01 -1.20134115e+00 2.45221570e-01
-1.34233332e+00 -7.44905472e-01 -3.00726324e-01 3.14759314e-01
-1.08403003e+00 2.63305306e-01 -6.83800876e-01 -7.17124224e-01
1.15208343e-01 -1.41074979e+00 1.59413707e+00 -1.40081748e-01
-7.08275199e-01 -8.40173662e-01 -4.23224449e-01 3.06313664e-01
2.16611490e-01 1.71130911e-01 1.20864141e+00 -6.07055783e-01
-1.15979159e+00 1.57853201e-01 2.86383420e-01 -1.71739489e-01
3.80397201e-01 3.56739461e-01 -4.72357303e-01 -3.24840963e-01
-3.62911880e-01 -4.40868646e-01 3.27687025e-01 1.37119368e-01
1.71005023e+00 -5.71361840e-01 -5.36300480e-01 6.91460907e-01
6.31279111e-01 4.02963847e-01 5.47274053e-01 1.30069211e-01
1.11643088e+00 5.21303296e-01 6.51298106e-01 3.11108559e-01
4.11161900e-01 1.01456678e+00 8.33933294e-01 2.46217355e-01
-1.61318228e-01 -5.39785147e-01 2.08409742e-01 8.80981624e-01
-3.52751255e-01 -1.41687259e-01 -9.79882002e-01 4.13568944e-01
-1.65161860e+00 -1.11529422e+00 -5.12826256e-02 1.74056852e+00
8.36771667e-01 1.78773150e-01 1.16786383e-01 -1.31135359e-01
2.62487531e-01 1.27028391e-01 -7.50487328e-01 -3.02715749e-01
3.50347459e-01 1.79534510e-01 8.08026642e-02 5.47674000e-01
-1.04172707e+00 1.33606827e+00 7.48311996e+00 2.93002337e-01
-1.00665700e+00 -3.74287546e-01 -1.12664297e-01 6.65841922e-02
-2.62506604e-01 1.19490184e-01 -8.05083632e-01 1.81273267e-01
2.95956552e-01 1.74652457e-01 7.64521360e-01 1.05996358e+00
3.03230315e-01 2.34765053e-01 -1.74539149e+00 5.91049433e-01
4.70888615e-03 -1.32427955e+00 5.60695410e-01 3.33146332e-03
4.34879869e-01 -3.88690203e-01 9.83230919e-02 3.91581923e-01
6.11091852e-01 -1.19529140e+00 9.14769232e-01 6.63158536e-01
6.12737298e-01 -1.77957892e-01 -1.56259328e-01 6.98507130e-01
-1.50616729e+00 -1.20860070e-01 -1.10518880e-01 -2.78366715e-01
3.02796304e-01 -3.21579903e-01 -1.09347677e+00 1.12357840e-01
3.06122869e-01 6.39369190e-01 -4.10669714e-01 7.81685233e-01
-4.91837502e-01 5.61145544e-01 -2.35429868e-01 -9.50813740e-02
-3.02039515e-02 5.45934290e-02 7.56547034e-01 1.14271021e+00
-1.96669605e-02 3.90002787e-01 7.77403414e-01 8.88653457e-01
2.85774231e-01 -2.80359030e-01 -8.74194920e-01 -2.31357977e-01
6.63986921e-01 8.08229446e-01 -5.77710688e-01 -5.08852899e-01
-1.01515956e-01 9.49346602e-01 4.17111248e-01 3.75548273e-01
-8.00535738e-01 1.03681073e-01 7.59105265e-01 2.42597610e-01
5.71208358e-01 -1.13934648e+00 -3.98054272e-01 -1.03872406e+00
-1.82877555e-01 -1.03002965e+00 -4.80479971e-02 -1.20937824e+00
-1.11815464e+00 1.57103077e-01 5.57796419e-01 -1.26013625e+00
-4.79138017e-01 -9.28376794e-01 -5.85516870e-01 2.69435197e-01
-7.87535071e-01 -1.55616236e+00 -2.92261809e-01 3.52137834e-01
1.31783152e+00 1.16758287e-01 8.37782562e-01 -3.51463258e-01
-1.75461724e-01 2.17169613e-01 -5.93997240e-01 2.52457678e-01
2.97904044e-01 -1.31700063e+00 7.89162040e-01 4.17930841e-01
3.53530347e-01 8.75366986e-01 9.19365048e-01 -8.49714458e-01
-1.93562829e+00 -9.44325209e-01 6.75284937e-02 -1.13142371e+00
6.08366132e-01 -4.68048483e-01 -8.90199482e-01 1.34264803e+00
2.30799928e-01 3.37322708e-03 3.22332621e-01 -7.91035146e-02
-3.24147969e-01 6.38751328e-01 -6.87474489e-01 1.12196434e+00
1.67949092e+00 -4.74302620e-01 -1.31779194e+00 3.77401799e-01
8.65170777e-01 -9.73038673e-01 -8.68705511e-01 4.44749951e-01
7.56506503e-01 -3.99263173e-01 1.21994424e+00 -1.38824296e+00
4.96251047e-01 -5.15776217e-01 -2.94055730e-01 -1.28810096e+00
-4.91359204e-01 -6.20816827e-01 -6.14019275e-01 7.33488142e-01
2.35864837e-02 -1.78088602e-02 7.16892481e-01 6.67814374e-01
-5.77807367e-01 -8.89015079e-01 -4.23116624e-01 -7.21469402e-01
-1.10418983e-01 -6.47990406e-01 4.27468359e-01 8.43116045e-01
-9.04021338e-02 2.08797097e-01 -2.59938598e-01 2.60968059e-01
3.57295811e-01 3.35701466e-01 1.27362728e+00 -1.15696025e+00
-4.73983854e-01 -4.76007432e-01 -5.22977233e-01 -1.68751669e+00
5.52124381e-01 -8.87543082e-01 2.53704488e-01 -1.55592513e+00
-1.06835216e-01 -3.58188093e-01 4.82669264e-01 8.91947448e-01
-2.37659499e-01 -1.39198288e-01 6.09838605e-01 3.39838803e-01
-5.98085344e-01 7.58268952e-01 1.94266248e+00 -1.76779911e-01
-2.82947779e-01 1.37216061e-01 -3.05668801e-01 1.00533295e+00
5.90388775e-01 -1.57651305e-01 -4.02941257e-01 -6.42314315e-01
-2.50604331e-01 -4.53944318e-02 6.73309922e-01 -8.52527142e-01
1.07379831e-01 -8.95153105e-01 4.41042721e-01 -9.61173415e-01
5.86789906e-01 -9.02687132e-01 1.65813848e-01 4.47683215e-01
-6.06576145e-01 3.97667795e-01 3.75802457e-01 4.98109549e-01
2.95443028e-01 1.01970717e-01 4.50238824e-01 -2.92715192e-01
-9.28234935e-01 2.71145523e-01 -5.47248781e-01 -9.74377841e-02
1.15139198e+00 -3.59056145e-01 -3.83292377e-01 -4.01286125e-01
-1.10968077e+00 3.61498952e-01 6.05627954e-01 8.13668191e-01
8.86252344e-01 -1.31663537e+00 -2.55135059e-01 1.32724702e-01
-5.37076704e-02 3.21382672e-01 -2.88804829e-01 4.28301096e-01
-4.46067482e-01 3.45581472e-01 -3.30598205e-01 -9.92692471e-01
-1.39633548e+00 7.98579097e-01 9.00625736e-02 3.67202878e-01
-8.16524267e-01 5.67378223e-01 2.97912061e-01 -5.05204201e-01
2.94785857e-01 -7.99781501e-01 8.00978318e-02 -6.02197886e-01
2.19864279e-01 2.29097769e-01 -4.97174203e-01 -5.20821929e-01
-1.25259012e-01 8.12470913e-01 1.30150244e-01 -5.70401363e-02
1.21206498e+00 1.37952596e-01 5.07179648e-02 7.87838578e-01
7.95770824e-01 1.40811130e-01 -1.62041509e+00 1.72489017e-01
3.59190337e-04 -6.24372363e-01 -5.15510917e-01 -6.42791152e-01
-6.41681314e-01 1.11391282e+00 1.68769553e-01 -1.88707739e-01
6.67857289e-01 3.55833024e-01 4.16262984e-01 1.04307032e+00
5.34251153e-01 -6.22703671e-01 6.44754171e-01 7.45488703e-01
1.46767282e+00 -9.97630775e-01 8.53917152e-02 -6.86080635e-01
-8.62706006e-01 1.25154710e+00 1.01009512e+00 -8.09084848e-02
3.36330682e-01 6.56468034e-01 -1.24200821e-01 -3.88500899e-01
-8.72141898e-01 -2.67479122e-01 6.46428943e-01 9.32957709e-01
3.91135991e-01 1.06874071e-01 3.89009595e-01 4.76257354e-02
-4.67738748e-01 -1.15050524e-01 1.97713569e-01 1.15072656e+00
-5.94564557e-01 -1.05529666e+00 -2.86869496e-01 4.67149645e-01
1.88113656e-02 -3.41332215e-03 -4.54314023e-01 1.10540068e+00
-1.56976715e-01 4.09219891e-01 1.98438480e-01 -4.01669472e-01
5.58373332e-01 2.19414428e-01 8.43019664e-01 -1.00918484e+00
-4.29540008e-01 -3.16250958e-02 -1.89168677e-02 -7.46761024e-01
-3.55774999e-01 -7.17458785e-01 -1.64859915e+00 -1.07320651e-01
-1.64911181e-01 -4.54489648e-01 3.86946201e-01 9.67371702e-01
3.94399494e-01 5.71225882e-01 1.95581228e-01 -1.73405433e+00
-8.17800164e-01 -7.82088697e-01 1.53627768e-01 7.28990018e-01
3.65269601e-01 -1.02370298e+00 -3.89603674e-02 5.32832563e-01] | [5.01302433013916, 0.2791096568107605] |
a3489640-31ba-4b5b-aa98-b8972ba3f60c | protocon-pseudo-label-refinement-via-online | 2303.13556 | null | https://arxiv.org/abs/2303.13556v1 | https://arxiv.org/pdf/2303.13556v1.pdf | ProtoCon: Pseudo-label Refinement via Online Clustering and Prototypical Consistency for Efficient Semi-supervised Learning | Confidence-based pseudo-labeling is among the dominant approaches in semi-supervised learning (SSL). It relies on including high-confidence predictions made on unlabeled data as additional targets to train the model. We propose ProtoCon, a novel SSL method aimed at the less-explored label-scarce SSL where such methods usually underperform. ProtoCon refines the pseudo-labels by leveraging their nearest neighbours' information. The neighbours are identified as the training proceeds using an online clustering approach operating in an embedding space trained via a prototypical loss to encourage well-formed clusters. The online nature of ProtoCon allows it to utilise the label history of the entire dataset in one training cycle to refine labels in the following cycle without the need to store image embeddings. Hence, it can seamlessly scale to larger datasets at a low cost. Finally, ProtoCon addresses the poor training signal in the initial phase of training (due to fewer confident predictions) by introducing an auxiliary self-supervised loss. It delivers significant gains and faster convergence over state-of-the-art across 5 datasets, including CIFARs, ImageNet and DomainNet. | ['Gholamreza Haffari', 'Hamid Rezatofighi', 'Ehsan Abbasnejad', 'Munawar Hayat', 'Islam Nassar'] | 2023-03-22 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Nassar_ProtoCon_Pseudo-Label_Refinement_via_Online_Clustering_and_Prototypical_Consistency_for_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Nassar_ProtoCon_Pseudo-Label_Refinement_via_Online_Clustering_and_Prototypical_Consistency_for_CVPR_2023_paper.pdf | cvpr-2023-1 | ['online-clustering', 'pseudo-label'] | ['computer-vision', 'miscellaneous'] | [ 3.08234394e-01 4.70705539e-01 -3.70532900e-01 -6.78728640e-01
-9.41300273e-01 -4.77116823e-01 6.68220162e-01 3.80083412e-01
-7.98983932e-01 8.85214686e-01 -8.74521583e-02 -1.15590543e-01
3.29541676e-02 -3.31417531e-01 -6.97322667e-01 -8.15311730e-01
1.03200473e-01 6.28308237e-01 3.30602676e-01 1.90974027e-01
9.47621390e-02 2.70161092e-01 -1.50058651e+00 2.46430978e-01
8.10110390e-01 1.14420319e+00 2.75842637e-01 1.45031109e-01
-1.07622877e-01 8.99933755e-01 -9.86796692e-02 -3.55918080e-01
2.27633879e-01 -2.54230887e-01 -7.80316949e-01 1.21640995e-01
3.34414423e-01 1.22657791e-01 -8.89714900e-03 9.82376814e-01
4.42748636e-01 1.53460698e-02 7.82503366e-01 -1.24604142e+00
-4.39316988e-01 6.10497475e-01 -3.21587622e-01 -1.30831197e-01
-1.77636206e-01 -8.81507024e-02 1.12696862e+00 -1.18854046e+00
6.06150508e-01 1.02855718e+00 9.40307498e-01 7.33048797e-01
-1.58733606e+00 -6.99961245e-01 3.18483084e-01 3.05144072e-01
-1.37591875e+00 -4.33007687e-01 7.21918762e-01 -3.55978578e-01
6.63235664e-01 -6.19989745e-02 2.72892565e-01 9.90048230e-01
-4.14760441e-01 8.92904639e-01 1.32320106e+00 -6.99900687e-01
4.99119699e-01 6.29291058e-01 1.75871924e-01 7.60413587e-01
-1.81561172e-01 2.36236140e-01 -5.97860336e-01 -7.46066198e-02
2.73509800e-01 -1.04341447e-01 -1.15599833e-01 -6.64285004e-01
-1.12845325e+00 1.02792931e+00 6.76570654e-01 7.23784268e-02
-1.37697518e-01 9.01599508e-03 4.01302457e-01 2.89745301e-01
8.45284939e-01 4.76886272e-01 -8.30084682e-01 8.98828581e-02
-1.19626856e+00 -2.76392668e-01 5.73936403e-01 6.64104283e-01
1.08968675e+00 -2.12494984e-01 -1.70504466e-01 9.89687741e-01
5.12698412e-01 -1.45113215e-01 4.21465665e-01 -9.36424553e-01
3.43339324e-01 7.60165632e-01 1.96058471e-02 -3.09760511e-01
-1.91283464e-01 -7.05896676e-01 -7.10063696e-01 3.61845285e-01
1.39411792e-01 -9.36613791e-03 -1.11431754e+00 1.87573242e+00
5.50588846e-01 6.85723603e-01 -6.63738400e-02 5.99753976e-01
4.32313651e-01 5.37305117e-01 3.53840441e-01 -2.22813278e-01
8.43233824e-01 -1.34326065e+00 -3.60489249e-01 -2.74241358e-01
8.77211988e-01 -4.84062940e-01 9.79894161e-01 5.05065620e-01
-6.04845166e-01 -7.13851035e-01 -1.12882423e+00 1.48286834e-01
-5.64667702e-01 3.90228271e-01 5.05548358e-01 4.91490513e-01
-1.39592338e+00 1.00502503e+00 -7.66167521e-01 -3.30662308e-03
7.29465604e-01 4.88006473e-01 -4.75025207e-01 -3.02972615e-01
-1.17657721e+00 9.45028007e-01 8.04098487e-01 -3.02353054e-02
-7.51892745e-01 -8.99259627e-01 -8.30938816e-01 -1.40009284e-01
2.59643912e-01 -1.96622640e-01 1.00644171e+00 -1.17791009e+00
-1.39314163e+00 1.06739974e+00 2.10624654e-02 -6.69710875e-01
6.69355214e-01 -2.73538172e-01 -2.83234358e-01 1.60445929e-01
3.00338000e-01 1.29939258e+00 1.02198946e+00 -1.30154037e+00
-6.42064631e-01 -3.18261348e-02 -4.10663635e-01 3.68286401e-01
-3.77309978e-01 -2.52801806e-01 -4.72586125e-01 -5.45597196e-01
5.93784191e-02 -1.13617730e+00 -4.00652528e-01 2.37597451e-01
-2.96601564e-01 -6.80553019e-01 7.80073225e-01 -1.05277158e-01
1.06154644e+00 -2.25516510e+00 8.44089836e-02 2.71586180e-01
2.28364348e-01 6.83831215e-01 -2.05280781e-01 2.69911259e-01
-3.16906989e-01 7.03114718e-02 -4.11878735e-01 -9.10984695e-01
4.40049134e-02 3.07957947e-01 -1.43317893e-01 5.75162709e-01
6.11401618e-01 7.74580777e-01 -1.10507548e+00 -7.28739381e-01
4.27067667e-01 5.50813496e-01 -4.32402343e-01 2.94554383e-01
-2.20793873e-01 4.97725904e-01 -2.69500669e-02 3.32850397e-01
5.50923109e-01 -5.60884416e-01 2.73348279e-02 -1.05683111e-01
-1.01267032e-01 3.56622040e-01 -1.17220426e+00 1.83301592e+00
-5.11269689e-01 4.11525548e-01 -7.52633512e-02 -1.26334906e+00
9.08551931e-01 3.13305765e-01 3.37348461e-01 -3.15558344e-01
-4.62613478e-02 2.18027666e-01 -5.49104571e-01 4.10707295e-02
-6.12537414e-02 -6.18239939e-02 1.60979554e-01 4.78771329e-01
4.98558134e-01 1.50711551e-01 1.74040809e-01 3.99552047e-01
1.01422775e+00 2.80465871e-01 1.14999734e-01 -2.77072042e-01
6.72403991e-01 -1.12076938e-01 5.77544034e-01 5.95892012e-01
-2.84706146e-01 4.85523582e-01 4.50144671e-02 -3.97376716e-01
-9.85459864e-01 -1.16649973e+00 -4.53343332e-01 1.28499281e+00
-1.20871916e-01 -4.05508280e-01 -5.05434811e-01 -1.25901294e+00
6.83608651e-03 5.77764809e-01 -8.56072962e-01 -4.69775468e-01
-2.07875073e-01 -5.52651048e-01 2.04299271e-01 3.59700471e-01
4.49624866e-01 -1.11172628e+00 -1.60824597e-01 4.15102392e-01
2.10376456e-02 -9.31564450e-01 -3.31285566e-01 9.36216533e-01
-9.81905282e-01 -9.28175330e-01 -6.41847014e-01 -1.09489226e+00
1.01958966e+00 -1.32641286e-01 1.17157567e+00 -4.31071930e-02
-1.51308715e-01 7.35437348e-02 -5.42563617e-01 -9.92523953e-02
-4.50693995e-01 1.67233676e-01 1.54361531e-01 2.78888106e-01
5.00572979e-01 -4.86300617e-01 -6.46932662e-01 4.11842108e-01
-5.09637952e-01 6.98110238e-02 6.26026869e-01 1.13919687e+00
7.70181894e-01 3.31468843e-02 8.55259836e-01 -1.21963048e+00
1.14351302e-01 -5.81634939e-01 -5.52231014e-01 2.70786464e-01
-1.20816040e+00 3.83124113e-01 8.27651918e-01 -6.13572717e-01
-8.85397017e-01 4.21801835e-01 -1.05931565e-01 -5.84096968e-01
-1.89864874e-01 3.02798361e-01 2.04395503e-01 -1.46109611e-01
6.66420996e-01 1.99133351e-01 -5.66720776e-02 -6.82784796e-01
5.34395516e-01 7.48277664e-01 4.59280252e-01 -4.19679493e-01
7.42472410e-01 4.00543332e-01 -2.50637174e-01 -3.35341632e-01
-1.43422055e+00 -9.09656882e-01 -8.58813763e-01 -1.71484232e-01
4.80396479e-01 -1.11466646e+00 -3.47117692e-01 2.73134291e-01
-7.87698030e-01 -4.16695446e-01 -6.23779297e-01 5.86811006e-01
-3.53291482e-01 3.34890068e-01 -6.30928874e-01 -5.98396719e-01
-3.30671042e-01 -9.19630408e-01 9.51492190e-01 7.32792392e-02
-1.82072014e-01 -1.05183041e+00 2.52731532e-01 2.29964986e-01
2.68762529e-01 -7.18605667e-02 6.30025387e-01 -9.74182367e-01
-2.92593300e-01 -1.68941170e-01 -3.87371838e-01 8.73913825e-01
-3.64313908e-02 -2.69400358e-01 -1.19882464e+00 -5.90107381e-01
-2.48646796e-01 -1.02017319e+00 1.17430067e+00 5.47407381e-02
1.23203135e+00 -1.86363652e-01 -2.99136400e-01 5.47199667e-01
1.44934738e+00 -2.42709935e-01 3.63297969e-01 3.35460603e-01
5.39113879e-01 6.92536712e-01 9.06303048e-01 3.20866674e-01
2.62187332e-01 3.71101409e-01 4.89809364e-01 -2.19862506e-01
-2.54930466e-01 -4.38706428e-01 3.25978220e-01 1.04090738e+00
4.90737051e-01 1.26366138e-01 -7.58285224e-01 5.93251228e-01
-1.80951738e+00 -5.79420865e-01 4.71492559e-02 2.21287060e+00
1.36313879e+00 3.43711048e-01 -2.10858822e-01 3.49871069e-01
7.81696916e-01 2.64817402e-02 -6.63388550e-01 -8.61827880e-02
4.18937169e-02 4.69378829e-01 5.21426678e-01 3.41134727e-01
-1.49134648e+00 1.14862061e+00 5.74301863e+00 1.13890576e+00
-1.12324619e+00 3.85922253e-01 8.61176193e-01 1.51383534e-01
2.09127422e-02 8.40484723e-02 -1.00603127e+00 5.88614702e-01
1.21996164e+00 4.85478997e-01 2.20208228e-01 1.15564704e+00
-1.07665755e-01 -2.26559415e-01 -1.24220371e+00 8.93205643e-01
-7.85157606e-02 -1.41449118e+00 -2.89705694e-01 -3.04229021e-01
8.90419245e-01 5.21337152e-01 8.76420736e-02 4.32885170e-01
5.47857940e-01 -9.25713599e-01 7.63196945e-01 1.32480577e-01
1.21113408e+00 -7.31148243e-01 7.15717673e-01 5.10022283e-01
-1.06108546e+00 -5.76521754e-02 -3.78080368e-01 2.28502542e-01
-1.72476731e-02 8.40794146e-01 -1.08867931e+00 2.32349500e-01
6.35206342e-01 1.02264833e+00 -7.01294184e-01 1.02447569e+00
-4.87447083e-01 1.00649369e+00 -4.57433105e-01 1.48416474e-01
5.45604408e-01 -2.00047657e-01 -4.46007727e-03 1.29436183e+00
-8.72929618e-02 -2.83471406e-01 2.93520689e-01 6.44332826e-01
-2.60208189e-01 4.63825464e-02 -3.18069667e-01 1.79420084e-01
6.19024336e-01 1.25556874e+00 -7.03530252e-01 -4.53210711e-01
-2.30059743e-01 1.08818531e+00 8.27821791e-01 9.86331552e-02
-6.70329094e-01 -2.13847086e-01 8.88531432e-02 -8.86876974e-03
4.12000149e-01 7.76958540e-02 -1.73134655e-01 -9.06254709e-01
-1.87045693e-01 -3.98434222e-01 3.31880748e-01 -4.94590908e-01
-1.44475353e+00 7.60685563e-01 -2.24941909e-01 -1.32085025e+00
-4.23027836e-02 -4.38867718e-01 -4.23570275e-01 7.32481003e-01
-1.95492387e+00 -1.12686145e+00 -2.92803720e-02 6.36036992e-01
5.88307858e-01 -1.70656919e-01 1.06240427e+00 4.35471535e-01
-4.82784152e-01 6.92600608e-01 3.87871563e-01 -1.31530175e-02
1.00021422e+00 -1.48992169e+00 1.20513946e-01 6.49828494e-01
3.16803217e-01 2.60636985e-01 4.04236436e-01 -4.32588369e-01
-3.93843412e-01 -1.52893746e+00 1.13633180e+00 -2.79775620e-01
5.91790378e-01 -5.56554973e-01 -9.13983822e-01 3.53175282e-01
-1.50530478e-02 5.51636517e-01 8.22173357e-01 1.48707349e-02
-5.83362043e-01 -1.41675845e-01 -1.12775099e+00 1.81141987e-01
8.54262412e-01 -6.29067242e-01 -5.21753728e-01 5.15469849e-01
7.67017126e-01 1.03272267e-01 -7.23950863e-01 3.91184628e-01
6.62946403e-02 -7.82384932e-01 8.85508597e-01 -3.81332666e-01
1.82896003e-01 -3.49005580e-01 1.22437783e-01 -1.39187622e+00
-3.58317852e-01 -6.55173242e-01 -5.39252236e-02 1.30547202e+00
6.66841149e-01 -5.94775140e-01 1.15813577e+00 3.05009514e-01
-1.96487486e-01 -9.21232820e-01 -1.03008473e+00 -7.01187611e-01
-1.41264871e-01 -3.47081959e-01 1.33786455e-01 1.16395593e+00
-2.45803952e-01 3.83368373e-01 -2.70958841e-01 -4.22152914e-02
1.00037515e+00 -1.71117559e-01 3.15174550e-01 -1.61288595e+00
-1.35325655e-01 3.70975435e-02 -2.13897675e-01 -9.76725698e-01
4.55048740e-01 -1.15908003e+00 2.54398167e-01 -1.15313578e+00
1.93336885e-02 -1.27246666e+00 -7.69496739e-01 8.00248563e-01
-1.97418183e-01 6.23489618e-01 9.36915204e-02 3.85985076e-01
-1.18967319e+00 7.24090636e-01 7.28717983e-01 3.49025317e-02
-5.01385815e-02 7.87356570e-02 -4.70301390e-01 7.14344800e-01
9.19386029e-01 -9.83648479e-01 -5.72887361e-01 9.58962832e-03
2.20816731e-02 -5.22993207e-01 1.52352422e-01 -1.08071482e+00
3.00046086e-01 2.28823155e-01 3.20596337e-01 -6.40637577e-01
3.36255878e-01 -9.63852108e-01 -4.09186222e-02 3.30182254e-01
-6.61812007e-01 -3.31747293e-01 -1.33445665e-01 9.17198241e-01
-2.76124626e-01 -3.69543433e-01 1.27696335e+00 3.93545218e-02
-7.50573874e-01 3.29849273e-01 -3.00349165e-02 2.47769952e-01
1.16261995e+00 -2.29431465e-01 1.79077223e-01 1.01493433e-01
-9.55108702e-01 4.25289243e-01 3.45493615e-01 1.95743844e-01
4.70830232e-01 -1.26481712e+00 -5.37517428e-01 2.42033735e-01
2.91382968e-01 3.08682024e-01 -6.64974973e-02 6.57330573e-01
-3.36549044e-01 3.83575380e-01 1.08571865e-01 -7.96460450e-01
-1.04662061e+00 4.65063214e-01 7.84141570e-02 -7.11905956e-01
-3.75967264e-01 1.35615742e+00 7.80786946e-02 -7.57530510e-01
6.10863566e-01 9.11569148e-02 -2.48957187e-01 1.47263587e-01
4.19478804e-01 5.87405339e-02 3.05139691e-01 -4.86481756e-01
-3.30293417e-01 3.44864577e-01 -3.95524591e-01 3.85061875e-02
1.52742648e+00 -2.23054141e-01 1.49383470e-01 5.61839342e-01
1.55727339e+00 -4.50183153e-01 -1.87868905e+00 -6.54660463e-01
3.33282888e-01 -2.03547031e-01 3.54864031e-01 -1.00731885e+00
-9.51114774e-01 7.77482629e-01 7.77005315e-01 -3.17062438e-01
9.09065843e-01 1.51560903e-01 4.63859916e-01 2.16116711e-01
3.61690223e-01 -1.44965661e+00 2.57578492e-01 4.20419335e-01
4.13812071e-01 -1.52238703e+00 -6.83992058e-02 -3.17781210e-01
-7.84672856e-01 8.20601583e-01 4.11827177e-01 -9.47873816e-02
8.45245600e-01 -6.04380071e-02 2.65209600e-02 -3.49921025e-02
-8.74527276e-01 -1.96008503e-01 2.09717348e-01 6.14023089e-01
1.48327321e-01 -5.93665466e-02 -5.90216368e-03 2.50645757e-01
2.60162354e-01 -2.15503648e-02 -3.23227490e-03 8.01466286e-01
-3.74519467e-01 -1.38861060e+00 -4.02450748e-03 7.33560979e-01
-3.68024558e-01 -3.19420397e-01 -2.26477176e-01 4.05059785e-01
3.45790893e-01 8.76088798e-01 -1.07982002e-01 -2.88617313e-01
-5.53586856e-02 2.98416674e-01 1.12742878e-01 -9.52520549e-01
-5.36909997e-01 -4.60716747e-02 1.10510748e-03 -5.52364409e-01
-7.00584590e-01 -5.60810149e-01 -1.14202881e+00 6.23771697e-02
-7.02884138e-01 5.16742408e-01 6.91620529e-01 7.95480490e-01
4.95775014e-01 1.23186313e-01 9.60073948e-01 -9.67868030e-01
-7.36598670e-01 -9.03750956e-01 -5.60726166e-01 5.65863788e-01
2.75448203e-01 -8.47986698e-01 -7.30558813e-01 1.05943210e-01] | [9.477921485900879, 3.593045473098755] |
e851bc81-bc21-41b8-af7b-19ee90bc3cdb | lost-in-context-on-the-sense-wise-variance-of | 2208.09669 | null | https://arxiv.org/abs/2208.09669v1 | https://arxiv.org/pdf/2208.09669v1.pdf | Lost in Context? On the Sense-wise Variance of Contextualized Word Embeddings | Contextualized word embeddings in language models have given much advance to NLP. Intuitively, sentential information is integrated into the representation of words, which can help model polysemy. However, context sensitivity also leads to the variance of representations, which may break the semantic consistency for synonyms. We quantify how much the contextualized embeddings of each word sense vary across contexts in typical pre-trained models. Results show that contextualized embeddings can be highly consistent across contexts. In addition, part-of-speech, number of word senses, and sentence length have an influence on the variance of sense representations. Interestingly, we find that word representations are position-biased, where the first words in different contexts tend to be more similar. We analyze such a phenomenon and also propose a simple way to alleviate such bias in distance-based word sense disambiguation settings. | ['Yue Zhang', 'Yile Wang'] | 2022-08-20 | null | null | null | null | ['word-sense-disambiguation'] | ['natural-language-processing'] | [-1.45685613e-01 -4.05974269e-01 -4.53900367e-01 -5.80724120e-01
-3.50665301e-01 -8.05405021e-01 7.48884499e-01 6.72050476e-01
-9.02667105e-01 5.20955324e-01 9.45127904e-01 -2.78076351e-01
-1.15695573e-01 -8.73905420e-01 -1.70898497e-01 -6.09844863e-01
2.44420424e-01 4.23419215e-02 9.99913216e-02 -5.72101653e-01
4.11673695e-01 1.72014594e-01 -1.08868515e+00 9.97826271e-03
7.60111809e-01 3.80309701e-01 5.68638086e-01 7.77814537e-02
-6.59737945e-01 -2.54891813e-02 -7.90133655e-01 -1.98304668e-01
2.96091791e-02 -1.53136402e-01 -6.20656133e-01 -3.22353184e-01
4.00425255e-01 4.27235633e-01 -2.60889977e-01 1.33776867e+00
4.14422333e-01 4.94653344e-01 7.87258387e-01 -6.54871404e-01
-1.35308194e+00 9.44917381e-01 -2.94228971e-01 6.83525026e-01
3.20763350e-01 1.23906881e-02 1.66735256e+00 -9.14212406e-01
5.63630760e-01 1.51657784e+00 4.51162040e-01 4.84789640e-01
-1.29341924e+00 -4.34225053e-01 5.59822798e-01 -2.30719661e-03
-1.47909355e+00 -1.52221946e-02 6.73610091e-01 -3.46850961e-01
1.09765470e+00 5.52240871e-02 7.42841423e-01 1.35121930e+00
2.91184783e-01 3.49374324e-01 8.33090127e-01 -5.25931776e-01
2.24321261e-01 -1.64555833e-02 5.90861797e-01 -1.21263443e-02
9.82581019e-01 -1.34894669e-01 -2.94870287e-01 -3.11539263e-01
3.55225533e-01 5.40397882e-01 -2.96851784e-01 -7.51746520e-02
-1.04418468e+00 1.09139502e+00 4.07883137e-01 9.53000903e-01
-1.63393706e-01 1.78766415e-01 3.69360119e-01 1.01664230e-01
2.94884861e-01 1.16975415e+00 -7.15139151e-01 -8.78366083e-02
-4.53135639e-01 3.23906511e-01 3.58210534e-01 6.51400566e-01
1.06586242e+00 -1.47677362e-01 -3.22377354e-01 1.15715086e+00
3.07074249e-01 4.98038411e-01 1.18226361e+00 -4.66707110e-01
2.96209067e-01 6.19728267e-01 3.21390629e-02 -1.28340912e+00
-3.03415298e-01 -2.80523270e-01 -3.00919771e-01 -5.39479971e-01
2.07441285e-01 -1.17334031e-01 -7.55285323e-01 2.29344034e+00
-5.71969263e-02 5.02650887e-02 1.41540105e-02 9.26796913e-01
4.42019343e-01 4.14465517e-01 5.43174803e-01 -5.26676849e-02
1.65983009e+00 -4.68757719e-01 -7.29830921e-01 -9.79445159e-01
8.02310705e-01 -7.87887752e-01 1.61120284e+00 -1.83479115e-01
-5.75902879e-01 -3.16815048e-01 -1.05595970e+00 -1.04508400e-01
-7.88759828e-01 -5.23163736e-01 6.08847797e-01 5.66583037e-01
-5.99665821e-01 5.41920483e-01 -4.77887273e-01 -5.61507046e-01
3.87101620e-02 -1.91857234e-01 -1.46687835e-01 -3.93852368e-02
-1.71102619e+00 1.20329118e+00 6.22188985e-01 -3.64442617e-01
-2.50607818e-01 -7.19989419e-01 -1.04497063e+00 1.48189917e-01
2.00123727e-01 -5.76758921e-01 7.97382534e-01 -7.71082520e-01
-5.83444178e-01 8.24215293e-01 -4.77338433e-01 -2.17791930e-01
-5.15493453e-01 -3.94321144e-01 -8.85712743e-01 -3.99991512e-01
3.09734046e-01 4.15446430e-01 5.82909524e-01 -8.46545637e-01
-2.79524148e-01 -3.68763059e-01 1.03924952e-01 2.88602620e-01
-5.40767550e-01 -1.00298263e-01 -8.19685608e-02 -8.84411812e-01
8.28905776e-02 -7.88709581e-01 -4.65402424e-01 -3.72071564e-01
-1.96337178e-02 -6.48640156e-01 3.17142010e-01 2.37413868e-02
1.75918663e+00 -2.42253375e+00 -7.16427341e-02 9.87277105e-02
2.95371152e-02 1.86272681e-01 -4.64747131e-01 6.63697243e-01
-1.50619745e-01 7.08535790e-01 -1.47538185e-02 1.32230548e-02
1.82764784e-01 7.98608184e-01 -5.09842873e-01 2.04729542e-01
2.58044660e-01 8.60125244e-01 -1.23709702e+00 -2.42380694e-01
7.38448426e-02 3.84253472e-01 -5.82119346e-01 -2.61601031e-01
-1.21281505e-01 -2.01124191e-01 -5.60034454e-01 2.58324683e-01
6.06756926e-01 -2.15337619e-01 5.95138490e-01 9.69863404e-03
8.02031681e-02 1.09079206e+00 -9.09549117e-01 1.63433123e+00
-8.70585978e-01 7.36554325e-01 -5.03422737e-01 -8.17638040e-01
7.15736508e-01 -2.91155861e-03 3.70025486e-02 -7.31617570e-01
3.95040177e-02 1.03955060e-01 2.91044325e-01 -3.60001355e-01
1.18812931e+00 -4.16927159e-01 -2.64530987e-01 3.29909354e-01
-9.36762020e-02 -1.92766339e-01 1.53680623e-01 2.42168024e-01
8.73397052e-01 -5.21940351e-01 6.51799321e-01 -5.85158765e-01
1.41999498e-02 -2.17803538e-01 8.42483878e-01 5.77365518e-01
-2.39042789e-01 4.16513234e-01 4.60332572e-01 -1.08722098e-01
-7.62815595e-01 -1.24348938e+00 -4.46168244e-01 1.19670677e+00
4.87228096e-01 -9.09496069e-01 -2.32858166e-01 -4.53106403e-01
2.16091067e-01 1.13152397e+00 -6.79876626e-01 -3.68822753e-01
-4.22306716e-01 -7.27910280e-01 3.59126836e-01 7.11578786e-01
-3.97745758e-01 -8.48084211e-01 -2.75442004e-01 2.86589414e-01
1.26982048e-01 -9.74288344e-01 -6.81355536e-01 3.43732357e-01
-7.40404308e-01 -8.44724059e-01 -3.25421989e-01 -5.29107630e-01
3.95096093e-01 5.00850260e-01 1.49275541e+00 3.11197549e-01
-3.31599824e-02 3.38807791e-01 -6.37786448e-01 -4.25478429e-01
7.63834938e-02 1.70801923e-01 5.08883834e-01 -5.64886034e-01
1.10122180e+00 -5.82844198e-01 -6.02012217e-01 5.74177578e-02
-1.20607495e+00 -9.56264555e-01 1.58548161e-01 9.27181244e-01
5.27875662e-01 -2.47663409e-01 4.00325477e-01 -9.27262962e-01
1.11802316e+00 -6.84508145e-01 -2.06554055e-01 2.03892082e-01
-9.42598104e-01 4.31165725e-01 3.88167888e-01 -7.21669674e-01
-6.97119236e-01 -6.50272191e-01 -5.03419936e-02 4.49464098e-02
1.24243321e-02 6.68607652e-01 -2.74627041e-02 4.53126907e-01
8.17913830e-01 -1.08262658e-01 -4.11836833e-01 -4.52612400e-01
7.25457847e-01 5.32259464e-01 -9.47723836e-02 -7.79423714e-01
5.44082344e-01 2.00445369e-01 -5.14516175e-01 -9.69437420e-01
-1.07789242e+00 -6.35977089e-01 -3.96490395e-01 5.76515257e-01
9.80279863e-01 -8.80160689e-01 7.96362758e-02 -2.53331214e-01
-1.23348844e+00 1.22932799e-01 -2.47530609e-01 6.20005608e-01
2.43500248e-01 4.56990987e-01 -2.73670316e-01 -4.78484631e-01
7.63045028e-02 -1.09121525e+00 8.52160633e-01 3.52950394e-01
-8.81498575e-01 -1.44891644e+00 4.02136236e-01 -3.42649996e-01
5.15115678e-01 -1.98825300e-01 1.16933250e+00 -1.07213950e+00
-4.95869219e-02 1.68759882e-01 -2.71220598e-02 2.87762225e-01
5.83069205e-01 -4.05760184e-02 -8.76364768e-01 3.92904412e-03
-2.61294425e-01 1.18121237e-01 1.06032276e+00 3.32422286e-01
8.70591283e-01 -2.46237949e-01 -5.18825471e-01 2.68687367e-01
1.55049944e+00 -2.24305391e-02 4.77847874e-01 2.37573206e-01
7.07780182e-01 2.99941391e-01 4.99864608e-01 2.29659408e-01
9.35887396e-02 5.29243171e-01 -9.38457251e-02 4.06284243e-01
1.41890302e-01 -4.06599700e-01 2.99029738e-01 9.27023411e-01
4.79624063e-01 -2.85775691e-01 -1.00518394e+00 9.68429744e-01
-1.32957029e+00 -9.22467709e-01 -7.89609030e-02 2.07498455e+00
1.02581906e+00 6.10739253e-02 -2.42877334e-01 -1.51926726e-01
6.69018447e-01 8.69652867e-01 -9.07913744e-02 -7.69023299e-01
-3.46122593e-01 3.87295932e-01 4.75234419e-01 6.78426862e-01
-6.25939310e-01 1.30666459e+00 6.86734438e+00 8.66942227e-01
-1.10724318e+00 1.51458815e-01 3.55484858e-02 -1.05091520e-01
-1.19850528e+00 1.53202266e-01 -7.33964920e-01 6.61828220e-01
7.62797058e-01 -6.37115717e-01 1.83231011e-01 9.00885284e-01
6.05311990e-02 -4.49090153e-02 -1.18507755e+00 8.68856907e-01
2.11890191e-02 -1.01567507e+00 3.89697462e-01 -9.15302634e-02
7.58266211e-01 2.63419896e-01 7.10678920e-02 2.76384801e-01
3.58494729e-01 -1.02256703e+00 3.26640934e-01 2.05652416e-01
4.48932201e-01 -7.27711618e-01 5.11666059e-01 5.46963923e-02
-1.14819586e+00 1.97149962e-02 -1.07400012e+00 -3.24047059e-01
1.75440282e-01 1.12540805e+00 -5.32248974e-01 1.41269788e-01
4.52533454e-01 6.70053542e-01 -7.06687748e-01 3.63406628e-01
-6.18998587e-01 5.87077141e-01 -1.16759725e-01 -4.50263590e-01
3.18388909e-01 -2.60848641e-01 5.24437010e-01 1.36201465e+00
2.15121910e-01 7.91600123e-02 3.58602181e-02 8.41605365e-01
-8.82497244e-03 1.70069680e-01 -8.38982701e-01 -5.43670118e-01
1.11871624e+00 8.52162898e-01 -3.72205466e-01 -4.09050435e-01
-6.19740367e-01 8.36007774e-01 4.11615282e-01 5.24522901e-01
-4.85076338e-01 -2.97619522e-01 1.70081556e+00 -9.66778584e-03
3.08959365e-01 -5.20618379e-01 -4.11127299e-01 -1.28269970e+00
2.41722316e-02 -5.20371199e-01 4.59807277e-01 -4.99248385e-01
-1.89420462e+00 3.55169088e-01 -1.28069162e-01 -8.71321559e-01
-8.05837959e-02 -8.16335797e-01 -6.65154874e-01 8.69553149e-01
-1.59003782e+00 -3.30544323e-01 2.96270430e-01 2.18302682e-01
5.63800454e-01 -4.36106734e-02 9.17769790e-01 1.53885692e-01
-3.92617881e-01 6.66567266e-01 3.69662046e-02 2.47768581e-01
1.02901244e+00 -1.23690403e+00 6.03866339e-01 7.67722309e-01
6.43540800e-01 1.56287420e+00 7.64859021e-01 -6.68619752e-01
-1.03925574e+00 -1.00882721e+00 1.43069732e+00 -8.37249696e-01
1.06161606e+00 -1.46468148e-01 -1.12888038e+00 6.51469946e-01
2.60898232e-01 -1.24746442e-01 1.11919248e+00 8.09403718e-01
-1.00566077e+00 6.20359555e-02 -7.60929048e-01 9.86559689e-01
1.27383304e+00 -1.01493287e+00 -1.37983251e+00 1.39739886e-02
1.32970548e+00 8.17638561e-02 -6.00820601e-01 -6.36108667e-02
3.67777258e-01 -6.85997427e-01 9.73667502e-01 -9.00999546e-01
2.96638697e-01 -1.50520727e-01 -6.08291864e-01 -1.80957580e+00
-5.79858959e-01 -2.16757372e-01 3.50780278e-01 1.24053073e+00
7.43365407e-01 -8.97068202e-01 7.29831085e-02 7.61923671e-01
5.79729751e-02 -5.82302988e-01 -7.93658853e-01 -1.02225482e+00
6.68632329e-01 -7.01038361e-01 1.04480243e+00 1.24813139e+00
3.29927742e-01 6.30055428e-01 1.90964252e-01 1.09801561e-01
-7.91714787e-02 2.25042570e-02 -8.07441995e-02 -1.25886214e+00
-5.12052700e-02 -5.65743625e-01 -6.45140469e-01 -1.25175476e+00
5.27494133e-01 -1.00016308e+00 -1.90696001e-01 -1.29390502e+00
1.20441303e-01 -4.60487366e-01 -8.00698578e-01 3.19401413e-01
-8.41471851e-01 -2.64412701e-01 2.00925052e-01 -1.27408415e-01
-4.13366795e-01 4.45193708e-01 8.79214168e-01 -1.11382850e-01
-1.59649849e-01 -5.60239196e-01 -1.12098014e+00 6.92692935e-01
8.44009638e-01 -7.49217987e-01 -5.63115656e-01 -9.72633600e-01
7.77992606e-01 -5.65749884e-01 7.31573626e-02 -5.24522841e-01
-2.09905043e-01 -6.10316455e-01 1.12049825e-01 3.03483359e-03
2.02207610e-01 -8.43557775e-01 -4.28229570e-01 1.16663463e-01
-4.91308123e-01 6.31896913e-01 -6.48745894e-02 7.57408321e-01
-3.04798752e-01 -3.85036558e-01 4.81774271e-01 -1.64047107e-01
-8.77842665e-01 1.81951821e-01 -5.11801958e-01 7.53790915e-01
4.93639112e-01 -1.09882265e-01 -1.95844278e-01 -5.88101288e-03
-3.17041039e-01 3.71389948e-02 7.61332929e-01 9.63922560e-01
3.51495624e-01 -1.59986770e+00 -3.15325677e-01 -6.62050967e-04
5.36741495e-01 -3.75930667e-01 2.58458275e-02 2.61304647e-01
9.33111235e-02 2.26306200e-01 2.25147918e-01 -3.81180257e-01
-8.84485006e-01 7.22616494e-01 1.17288083e-01 1.29158080e-01
-2.18927741e-01 8.55026960e-01 1.30798206e-01 -3.16150665e-01
-1.29111707e-01 -7.26886213e-01 -2.52492815e-01 4.07973349e-01
6.01273179e-01 -1.69739313e-02 -8.53697509e-02 -5.30457854e-01
-7.27331102e-01 7.25924551e-01 -2.13657111e-01 -1.75356597e-01
9.90180254e-01 -2.67612576e-01 -9.12906155e-02 7.89773345e-01
1.23882985e+00 4.16847497e-01 -4.88897979e-01 -3.33507746e-01
3.81360710e-01 -6.67771816e-01 -1.22167058e-01 -3.52152079e-01
-6.93476200e-01 8.99166286e-01 2.92471558e-01 2.77296931e-01
6.18175685e-01 1.63384795e-01 9.30263281e-01 4.48780179e-01
4.36701655e-01 -1.30947447e+00 -5.98685741e-02 9.41548467e-01
4.51135099e-01 -1.11665606e+00 -1.90220345e-02 -1.44602358e-01
-8.65847945e-01 8.26616168e-01 4.46676999e-01 -1.15802586e-01
8.18677127e-01 7.75665268e-02 3.29639465e-01 -1.19997799e-01
-7.18821883e-01 -2.99722880e-01 2.08933204e-01 6.79324269e-01
9.41720724e-01 4.12949413e-01 -1.02319109e+00 7.69272089e-01
-4.37060148e-01 -7.37669706e-01 3.07377517e-01 5.49337149e-01
-6.11165285e-01 -1.44454157e+00 -1.34410456e-01 2.60553330e-01
-4.87919390e-01 -7.15084076e-01 -5.30454159e-01 5.32210410e-01
1.59174457e-01 9.64804113e-01 3.48951906e-01 -3.01104099e-01
1.39721289e-01 3.63912284e-01 2.57476687e-01 -1.09747815e+00
-3.79794627e-01 -3.86284143e-01 5.33166975e-02 -4.64945674e-01
-5.31580187e-02 -3.44522089e-01 -1.45796716e+00 -1.65000454e-01
-2.99152821e-01 3.45623195e-01 3.74312878e-01 1.09802043e+00
5.55766821e-01 2.99734086e-01 3.61104012e-01 -1.76920101e-01
-6.18320942e-01 -1.08629727e+00 -7.36989200e-01 8.83268952e-01
2.92117149e-01 -8.46011937e-01 -5.67609370e-01 -4.96236145e-01] | [10.351179122924805, 8.924288749694824] |
8f3b5761-d43f-49b0-b952-38c1adcc2d3a | deep-learning-for-ps-weakly-dependent | 2302.00333 | null | https://arxiv.org/abs/2302.00333v1 | https://arxiv.org/pdf/2302.00333v1.pdf | Deep learning for $ψ$-weakly dependent processes | In this paper, we perform deep neural networks for learning $\psi$-weakly dependent processes. Such weak-dependence property includes a class of weak dependence conditions such as mixing, association,$\cdots$ and the setting considered here covers many commonly used situations such as: regression estimation, time series prediction, time series classification,$\cdots$ The consistency of the empirical risk minimization algorithm in the class of deep neural networks predictors is established. We achieve the generalization bound and obtain a learning rate, which is less than $\mathcal{O}(n^{-1/\alpha})$, for all $\alpha > 2 $. Applications to binary time series classification and prediction in affine causal models with exogenous covariates are carried out. Some simulation results are provided, as well as an application to the US recession data. | ['Wade Modou', 'William Kengne'] | 2023-02-01 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [ 1.26094818e-01 1.46713346e-01 -5.74718237e-01 -4.89050269e-01
-6.74079657e-01 -1.81442704e-02 2.56653905e-01 1.45856366e-01
-5.65294027e-01 1.09748268e+00 -8.43532011e-02 -7.08958626e-01
-8.54379654e-01 -1.00513875e+00 -7.42847621e-01 -1.19489896e+00
-7.40937531e-01 4.04969007e-01 -5.71628630e-01 -3.59889492e-02
2.79303472e-02 2.67396241e-01 -9.84558463e-01 -5.33677578e-01
8.32904518e-01 1.28803205e+00 -3.32130551e-01 5.38357675e-01
3.05957735e-01 7.09847689e-01 -1.95164055e-01 -6.60066485e-01
3.65694553e-01 -2.52415031e-01 -2.91039139e-01 -4.65463281e-01
7.04255551e-02 -2.12350577e-01 -4.35785115e-01 1.21552992e+00
5.16128659e-01 3.88175339e-01 1.13915527e+00 -1.13189375e+00
-5.59181809e-01 9.99525130e-01 -8.27521086e-01 6.44003749e-01
-6.25103533e-01 -5.41586801e-02 9.88404214e-01 -5.98197520e-01
-1.83446538e-02 1.04865539e+00 1.04478741e+00 2.39667907e-01
-1.25906515e+00 -1.33582199e+00 1.26866907e-01 9.42491964e-02
-1.17487371e+00 -1.55001849e-01 7.54415154e-01 -7.96693504e-01
8.42763543e-01 -7.78540522e-02 1.18148707e-01 1.26065218e+00
6.03927195e-01 3.78833741e-01 1.02206171e+00 -3.25933665e-01
3.01086158e-01 -3.33040208e-01 6.07533813e-01 5.90754092e-01
2.90707082e-01 4.82855439e-01 -2.39492357e-01 -2.79866397e-01
1.02382529e+00 1.45810634e-01 7.41006508e-02 1.63809910e-01
-7.91189015e-01 1.27342689e+00 -6.20627888e-02 2.06721619e-01
-5.56802094e-01 5.44274271e-01 3.70591521e-01 6.62828624e-01
1.14088285e+00 -4.22949605e-02 -8.75168085e-01 9.49112475e-02
-6.26436591e-01 1.96720541e-01 7.24574029e-01 7.38352716e-01
3.70155364e-01 6.38740122e-01 -2.70730779e-02 8.02904069e-01
6.91276416e-02 8.07636082e-01 1.88653558e-01 -1.04488528e+00
7.34777153e-01 -1.56742427e-02 1.59998015e-01 -9.18813527e-01
-7.75496542e-01 -5.94214499e-01 -1.89810503e+00 -2.82291062e-02
5.33412278e-01 -8.34916651e-01 -6.82289243e-01 2.28923750e+00
-1.37289912e-01 4.38638002e-01 -1.37493789e-01 3.73097330e-01
1.53671160e-01 6.96471751e-01 3.33215088e-01 -9.00278091e-01
9.49584007e-01 -7.53732771e-02 -8.12436342e-01 -6.69861659e-02
4.21163619e-01 -4.18129891e-01 7.46436119e-01 4.31788266e-01
-1.20361972e+00 -4.83763754e-01 -6.82332098e-01 2.84348845e-01
-7.18583986e-02 -6.14967663e-04 9.24858689e-01 7.01089919e-01
-8.08603466e-01 9.25726593e-01 -9.50574815e-01 1.13091141e-01
5.86027145e-01 6.76460862e-01 3.91424857e-02 3.44770342e-01
-1.44113588e+00 3.77401888e-01 1.37230158e-01 2.96516836e-01
-9.11464274e-01 -9.40087378e-01 -6.15441620e-01 2.00859517e-01
1.16269656e-01 -5.11969924e-01 1.18466115e+00 -9.75894272e-01
-1.16933119e+00 7.75245190e-01 -2.85556167e-02 -8.09311152e-01
5.89540839e-01 -2.30203927e-01 -5.79390407e-01 -3.58035564e-01
1.90001577e-01 -1.76090300e-01 9.62897837e-01 -4.89904553e-01
-6.85856581e-01 -5.91596544e-01 -1.61003262e-01 -4.05546337e-01
-1.22593634e-01 2.10920125e-01 4.56299096e-01 -7.51042366e-01
-6.13095127e-02 -9.39078093e-01 -5.84937274e-01 -5.60910106e-01
-3.63752335e-01 -4.75740105e-01 2.89087981e-01 -8.06743979e-01
1.21054614e+00 -2.11106920e+00 -2.66040787e-02 4.32365447e-01
5.17812073e-02 -2.34311625e-01 2.30824471e-01 2.32397869e-01
-7.12700367e-01 2.49867901e-01 -6.20408952e-01 -4.87385839e-01
9.87261534e-02 -4.81576985e-03 -4.55115527e-01 9.08783317e-01
4.19954211e-02 5.37702143e-01 -5.31534135e-01 -4.70288768e-02
-1.32626101e-01 8.85324106e-02 -2.65878469e-01 -5.33901453e-02
-2.57465392e-01 3.20491761e-01 -3.66256684e-01 4.50701356e-01
6.89582646e-01 -4.32381690e-01 9.23950002e-02 2.85470158e-01
-2.39150077e-01 -5.03223063e-03 -1.13943326e+00 1.07636261e+00
-5.51938295e-01 6.07865691e-01 8.16221070e-03 -1.83930326e+00
8.34001839e-01 5.80021739e-01 6.67714000e-01 -3.00876498e-01
5.81294835e-01 1.49000749e-01 2.53346145e-01 -4.65927839e-01
-1.37803242e-01 -8.73604059e-01 -2.75973201e-01 4.44767326e-01
-5.44370748e-02 3.15678567e-01 7.83267692e-02 -4.23266381e-01
1.13026989e+00 -2.21262410e-01 4.55725700e-01 -6.47262812e-01
1.50348052e-01 -3.66692275e-01 8.92008543e-01 1.01672983e+00
-2.98005790e-01 8.15367699e-02 9.90692317e-01 -3.96913201e-01
-1.11069751e+00 -1.15800321e+00 -4.23165560e-01 1.22852015e+00
-3.80632818e-01 3.43605220e-01 -3.21745068e-01 -3.29365611e-01
5.09935170e-02 8.28458071e-01 -9.57272768e-01 -2.85531938e-01
-7.99186587e-01 -1.69826984e+00 5.91390371e-01 9.17833626e-01
4.82146502e-01 -1.12731707e+00 -2.49023482e-01 2.63595402e-01
2.97775343e-02 -5.22804677e-01 -5.34210354e-02 8.02571118e-01
-1.06662476e+00 -7.66137600e-01 -7.36667633e-01 -7.68232107e-01
1.77504510e-01 -3.62140596e-01 1.06493211e+00 -5.32204866e-01
-5.68397762e-03 7.98145533e-02 3.25682670e-01 -7.81466544e-01
-2.13034302e-01 5.19483835e-02 4.53999430e-01 -1.16823651e-01
6.59095824e-01 -1.11657894e+00 -7.28535533e-01 -3.92779373e-02
-6.63116515e-01 -4.45643872e-01 4.99737412e-01 1.03917909e+00
5.51629186e-01 3.66249621e-01 1.02973473e+00 -8.13035667e-01
5.79268873e-01 -7.96300352e-01 -8.44520688e-01 -7.59196058e-02
-6.92435205e-01 -1.33618653e-01 7.34928966e-01 -6.49851620e-01
-1.08112299e+00 -3.12719315e-01 -3.02086491e-02 -4.36842382e-01
-1.48537591e-01 9.47087526e-01 5.24349324e-02 6.62962794e-01
6.20345354e-01 -8.35185722e-02 -2.20156103e-01 -5.44247508e-01
1.38943806e-01 1.64793596e-01 4.21307147e-01 -5.03101587e-01
5.89047372e-01 4.90926564e-01 4.98655945e-01 -6.49379730e-01
-8.03956687e-01 -2.15792991e-02 -4.78587419e-01 1.23550810e-01
1.03166687e+00 -9.03872192e-01 -1.10495234e+00 3.56642812e-01
-1.03229141e+00 -6.21570289e-01 -2.54839808e-01 9.74061310e-01
-9.15210247e-01 -1.80184752e-01 -1.06812835e+00 -1.28360701e+00
-2.80261487e-01 -6.94740891e-01 4.40572411e-01 6.35183677e-02
-2.85676774e-02 -1.32759643e+00 1.93129256e-01 -6.05125092e-02
1.04674876e-01 3.19430530e-01 1.54080009e+00 -7.64626026e-01
-1.15635999e-01 -2.75574714e-01 -2.19401449e-01 5.61200380e-01
2.00750772e-02 -7.91500323e-03 -8.22958708e-01 -8.56089517e-02
3.37742895e-01 -1.34296985e-02 1.12548304e+00 1.34902298e+00
1.59325957e+00 -4.35272038e-01 -2.81412154e-01 6.40329182e-01
1.25254190e+00 8.16108644e-01 4.71417367e-01 -7.68540651e-02
4.74819899e-01 6.72733724e-01 1.69542640e-01 6.83881462e-01
-1.36290744e-01 -8.85071754e-02 3.69427532e-01 -1.62603036e-01
7.04675913e-01 1.36558965e-01 3.26160818e-01 9.56206381e-01
-4.31727320e-01 -3.30408782e-01 -8.95326674e-01 4.30989921e-01
-1.94601953e+00 -1.16609883e+00 -4.60781902e-01 2.25925303e+00
8.76498163e-01 1.96346179e-01 8.96582380e-02 8.72753561e-02
8.45159709e-01 8.23071972e-02 -6.28002167e-01 -4.07984525e-01
-2.58167475e-01 8.80825162e-01 8.94977629e-01 5.60309768e-01
-1.51600182e+00 4.83521610e-01 6.33668423e+00 9.66017902e-01
-9.60088551e-01 3.73905241e-01 1.36204970e+00 -3.08411215e-02
-6.38561044e-03 -3.08471292e-01 -5.40936053e-01 4.78930563e-01
1.15752816e+00 -2.44888216e-01 -3.42886783e-02 8.73780251e-01
5.94864070e-01 1.88733429e-01 -9.02187526e-01 9.21561360e-01
-4.71616089e-01 -1.08049846e+00 -5.40857613e-01 2.12498829e-01
9.17237937e-01 -6.83962330e-02 6.75733984e-01 5.85873127e-01
9.55798388e-01 -1.31252611e+00 2.45442554e-01 3.97297949e-01
8.77841115e-01 -1.30489349e+00 8.07506382e-01 2.99703568e-01
-9.31558311e-01 -5.88881314e-01 -3.81848454e-01 -3.98080319e-01
2.32662737e-01 9.80595291e-01 -3.06150764e-01 5.15451789e-01
7.57920921e-01 8.56119871e-01 2.63255388e-01 6.78214431e-01
6.22525848e-02 1.04596901e+00 -2.81896383e-01 1.94123778e-02
1.60138533e-01 -5.16545832e-01 3.91368389e-01 1.05014563e+00
6.45652831e-01 5.10569155e-01 -1.98579445e-01 5.72167695e-01
-1.46873757e-01 1.01519920e-01 -6.28574550e-01 1.59154609e-01
-1.62008591e-02 5.94022870e-01 -6.46734715e-01 -3.24994296e-01
-3.03638160e-01 3.98832470e-01 4.94278260e-02 6.22864068e-01
-1.04935610e+00 -2.24909738e-01 7.99339354e-01 -9.42326561e-02
1.35904923e-01 -2.81009048e-01 -7.07086265e-01 -1.03625762e+00
-2.77815312e-01 -5.06301880e-01 8.05529177e-01 -2.36127317e-01
-1.85437775e+00 7.70446062e-02 3.30734998e-01 -9.15893734e-01
-2.76083052e-01 -7.49615967e-01 -5.87756813e-01 1.02678668e+00
-1.03947115e+00 -6.19391620e-01 3.23536277e-01 5.38129568e-01
4.01488811e-01 -4.07813638e-01 5.94863713e-01 4.92771119e-01
-1.00420141e+00 3.80704582e-01 6.04357958e-01 2.69021112e-02
3.34850192e-01 -1.26919746e+00 1.99660379e-02 8.16372931e-01
-2.54061699e-01 4.79130924e-01 8.71371210e-01 -7.90317595e-01
-8.06629539e-01 -1.08059096e+00 7.91437924e-01 -5.05136475e-02
9.98489439e-01 -1.95027322e-01 -8.48875880e-01 9.76534724e-01
1.77864417e-01 -2.96314955e-01 8.74959350e-01 4.75104183e-01
-1.26161724e-01 -3.59638721e-01 -9.21162307e-01 6.44208550e-01
1.11556554e+00 -2.42557332e-01 -2.16907188e-01 4.99959171e-01
7.06478119e-01 1.20180517e-01 -9.35092926e-01 7.75148511e-01
4.61894035e-01 -7.98708498e-01 1.02626884e+00 -1.21293187e+00
7.64572918e-01 5.92766047e-01 -3.79306257e-01 -1.05124354e+00
-5.34066260e-01 -8.23452532e-01 -1.63460404e-01 9.37751532e-01
5.45264006e-01 -7.74345815e-01 6.81957960e-01 5.12612402e-01
-1.10149058e-02 -5.31778038e-01 -1.55528617e+00 -7.80585170e-01
7.95809031e-01 -8.86318684e-01 3.00540984e-01 1.27620983e+00
-3.14940214e-01 2.56861687e-01 -6.50531411e-01 1.47930920e-01
8.62411261e-01 -1.37630492e-01 2.66944677e-01 -1.57272255e+00
-4.76313323e-01 -7.98164248e-01 1.13711096e-01 -7.37846434e-01
7.53701270e-01 -6.45913422e-01 -5.23582511e-02 -9.13177669e-01
2.78460477e-02 -8.19127977e-01 -8.40330541e-01 1.35222524e-01
-8.16963911e-02 -2.69798160e-01 -3.60346437e-01 -5.01580676e-03
1.11404136e-01 6.27588630e-01 6.36003852e-01 -1.42645016e-01
-9.96025652e-02 7.82165766e-01 -5.67533791e-01 1.02322674e+00
1.09447002e+00 -5.92460930e-01 -4.73801941e-01 -1.45025969e-01
4.60306078e-01 6.57787681e-01 4.21716511e-01 -5.71056545e-01
-3.09824675e-01 -6.10603273e-01 4.21831518e-01 -5.86479604e-01
3.71032685e-01 -6.06348693e-01 4.43755947e-02 7.09587038e-01
-6.75133526e-01 3.36131662e-01 3.30674164e-02 8.21475267e-01
1.08326882e-01 -2.57650048e-01 8.24256301e-01 -6.07261546e-02
-3.76784131e-02 6.18237019e-01 -4.62971807e-01 -1.89491864e-02
7.24762440e-01 4.93532032e-01 -6.10625111e-02 -7.44285464e-01
-9.24161017e-01 -1.67535488e-02 -5.63177645e-01 -2.60064360e-02
1.32442802e-01 -1.24076188e+00 -8.54679346e-01 -8.87811836e-03
-3.88609171e-01 -1.52192771e-01 5.53227246e-01 1.14629984e+00
-1.83294415e-02 3.92784148e-01 1.44420058e-01 -4.74332511e-01
-7.64967382e-01 6.60258949e-01 4.55269992e-01 -4.17346835e-01
-3.80094498e-01 8.67875934e-01 6.21651471e-01 5.46040274e-02
1.97361946e-01 -5.96844077e-01 -1.68166310e-01 1.93358675e-01
2.82947153e-01 7.50956893e-01 -1.84231713e-01 -1.49580613e-01
-5.31536937e-02 1.51839897e-01 3.11462104e-01 -3.94135982e-01
1.54180253e+00 -6.56209663e-02 -2.77758658e-01 9.65683877e-01
1.26404166e+00 -4.25483972e-01 -1.34774399e+00 -2.92783350e-01
1.83972225e-01 7.89201409e-02 -2.18808427e-01 -3.89405847e-01
-1.02850604e+00 1.07840216e+00 9.19819534e-01 6.00702822e-01
1.04852509e+00 -1.63628787e-01 4.40112203e-01 3.52138132e-01
3.49879637e-02 -1.02093434e+00 -2.11788639e-01 7.02328801e-01
5.53890169e-01 -1.17884302e+00 -2.56314844e-01 6.60137087e-02
-1.43735722e-01 8.88272345e-01 4.24278378e-01 -6.10761523e-01
1.34930277e+00 1.70294836e-01 -2.44920462e-01 9.36292782e-02
-7.75430501e-01 5.20969629e-02 7.93513730e-02 4.25090641e-01
5.47538698e-01 3.48718137e-01 -6.61432862e-01 1.08678198e+00
-2.14997888e-01 -1.48458183e-01 3.09973866e-01 4.22776073e-01
-1.33890077e-01 -5.79424620e-01 -1.40026689e-01 8.27868521e-01
-1.02431154e+00 -2.36310616e-01 3.20368081e-01 9.81544614e-01
5.93088306e-02 7.85150409e-01 3.40551823e-01 6.51997477e-02
4.51541953e-02 3.75643075e-01 1.24015234e-01 -8.66021961e-02
-4.69364107e-01 3.84390563e-01 2.60138344e-02 -3.11203450e-02
-5.06528735e-01 -7.77791917e-01 -8.75880063e-01 -5.50172687e-01
1.02800252e-02 -9.89438519e-02 3.55389416e-01 1.07264078e+00
-9.83154997e-02 3.77526402e-01 7.72605658e-01 -5.96847475e-01
-6.45106435e-01 -1.31068468e+00 -8.43720198e-01 1.27566084e-01
5.79028368e-01 -8.18221807e-01 -6.47723436e-01 1.33782715e-01] | [7.100645542144775, 3.875985622406006] |
e3ec78ff-be8e-452b-bfc4-2df9a4a24e41 | context-aware-learning-using-transferable | 1803.00386 | null | http://arxiv.org/abs/1803.00386v2 | http://arxiv.org/pdf/1803.00386v2.pdf | Context-Aware Learning using Transferable Features for Classification of Breast Cancer Histology Images | Convolutional neural networks (CNNs) have been recently used for a variety of
histology image analysis. However, availability of a large dataset is a major
prerequisite for training a CNN which limits its use by the computational
pathology community. In previous studies, CNNs have demonstrated their
potential in terms of feature generalizability and transferability accompanied
with better performance. Considering these traits of CNN, we propose a simple
yet effective method which leverages the strengths of CNN combined with the
advantages of including contextual information, particularly designed for a
small dataset. Our method consists of two main steps: first it uses the
activation features of CNN trained for a patch-based classification and then it
trains a separate classifier using features of overlapping patches to perform
image-based classification using the contextual information. The proposed
framework outperformed the state-of-the-art method for breast cancer
classification. | ['Ruqayya Awan', 'Navid Alemi Koohbanani', 'Muhammad Shaban', 'Nasir Rajpoot', 'Anna Lisowska'] | 2018-02-12 | null | null | null | null | ['classification-of-breast-cancer-histology'] | ['medical'] | [ 4.03166652e-01 -2.18748413e-02 -1.75332561e-01 -3.10742527e-01
-8.97852302e-01 -1.95565492e-01 5.06054997e-01 6.49772704e-01
-6.76463604e-01 6.87844634e-01 -1.34283006e-01 -3.74086797e-01
-2.20418021e-01 -9.09232795e-01 -5.33406734e-01 -1.04111290e+00
2.64399145e-02 5.96632920e-02 4.16627526e-01 -1.89029962e-01
1.63013086e-01 9.06758606e-01 -1.50132239e+00 5.47381818e-01
7.13251829e-01 1.31476235e+00 3.10635835e-01 3.96940351e-01
-8.57659951e-02 5.25087059e-01 -2.95372069e-01 -5.25956899e-02
9.42037255e-02 -2.56764203e-01 -7.71488667e-01 -5.84015027e-02
8.51171315e-02 -6.31232262e-02 -2.20069379e-01 6.19868279e-01
6.24859750e-01 -9.47071761e-02 4.98236597e-01 -7.11047471e-01
-1.79167718e-01 1.22621745e-01 -3.00726473e-01 3.89925599e-01
-1.73407510e-01 8.58871713e-02 9.49718118e-01 -7.00095892e-01
6.29093647e-01 5.15432119e-01 9.54627335e-01 5.24786115e-01
-1.19642413e+00 -5.43893814e-01 -1.44048885e-01 1.20243458e-02
-1.26912093e+00 -2.76650220e-01 6.41700029e-01 -4.22552824e-01
1.04875910e+00 2.60599762e-01 1.00941777e+00 8.48294616e-01
4.11385685e-01 6.52271986e-01 1.23512232e+00 -4.52985615e-01
2.01004162e-01 2.82905221e-01 4.17915396e-02 6.23223662e-01
1.80851266e-01 3.93656716e-02 -1.35657176e-01 -1.73910066e-01
7.61684239e-01 2.52133787e-01 -1.92417875e-01 -2.53372341e-01
-1.18464434e+00 7.86241651e-01 9.16314542e-01 7.80893803e-01
-4.55101550e-01 5.49573787e-02 5.32582164e-01 2.51638126e-02
5.51151454e-01 4.93101299e-01 -3.01736057e-01 2.84585387e-01
-1.11706233e+00 5.28262779e-02 4.87953395e-01 3.83187234e-01
6.43824458e-01 -2.87812680e-01 -1.81021556e-01 8.03813636e-01
9.70776975e-02 -1.23378798e-01 7.28080153e-01 -1.97473988e-01
-2.72589289e-02 9.98060107e-01 -4.36382890e-01 -9.68007028e-01
-8.31858575e-01 -7.38273740e-01 -1.15640771e+00 1.71263441e-01
4.93592143e-01 2.19791263e-01 -9.81994450e-01 1.48296392e+00
3.48528057e-01 1.73146024e-01 4.68571968e-02 6.27078474e-01
1.01697361e+00 9.70314890e-02 1.83747977e-01 1.28301576e-01
1.41798735e+00 -7.28542566e-01 -4.88148391e-01 9.94600579e-02
9.90789771e-01 -5.89275002e-01 6.93810344e-01 2.53345639e-01
-7.45879948e-01 -3.74819756e-01 -1.17444348e+00 1.90018700e-03
-7.33336508e-01 4.02660161e-01 8.62221539e-01 6.39347792e-01
-1.24539590e+00 5.18385530e-01 -1.01187325e+00 -8.77036095e-01
1.04333210e+00 7.31090963e-01 -6.98936045e-01 -5.79799227e-02
-8.16869915e-01 7.91943908e-01 4.52128977e-01 2.26960972e-01
-8.37309718e-01 -7.04005301e-01 -7.03320622e-01 1.02170020e-01
-2.26432458e-02 -6.87758327e-01 8.49946141e-01 -1.05664110e+00
-1.41336966e+00 1.05206287e+00 1.48255184e-01 -4.59536731e-01
4.83272254e-01 2.22066864e-01 -8.29814896e-02 3.38725567e-01
-1.21215105e-01 7.25295126e-01 3.32414746e-01 -8.06763947e-01
-6.12800539e-01 -4.28054988e-01 5.71643524e-02 -9.37175676e-02
-6.32073045e-01 -2.17431098e-01 -3.40681523e-01 -5.85024297e-01
2.39846483e-01 -9.59521353e-01 -4.50091630e-01 2.06744418e-01
-3.49910051e-01 -6.94465414e-02 6.47772789e-01 -3.87068808e-01
9.74973440e-01 -2.21626496e+00 -7.64901489e-02 3.54754388e-01
2.03053787e-01 4.09251422e-01 -5.38372174e-02 5.85993469e-01
-1.17074914e-01 1.47437319e-01 -7.80832842e-02 -1.64353654e-01
-4.98175144e-01 -1.44885257e-02 3.86319965e-01 7.22876906e-01
5.32933474e-01 1.00926387e+00 -8.25133204e-01 -5.70853770e-01
2.62457430e-01 5.79323947e-01 -5.15218616e-01 1.71181574e-01
1.57824203e-01 5.54959655e-01 -4.47318465e-01 7.54710853e-01
5.85715294e-01 -3.33735913e-01 3.85662317e-01 -3.63131166e-01
-1.24118946e-01 -3.64175364e-02 -7.91259170e-01 1.60470736e+00
-4.26339388e-01 5.62334180e-01 1.68356784e-02 -1.30547488e+00
8.58690739e-01 4.32775617e-01 6.35168791e-01 -4.60082382e-01
3.18820119e-01 4.96939063e-01 1.95147887e-01 -5.61853647e-01
-1.78089552e-02 -2.65386760e-01 3.36873382e-01 1.48370847e-01
3.72980684e-01 1.27705440e-01 -5.38828783e-02 -1.15368716e-01
1.32864225e+00 -7.41920173e-02 6.06724501e-01 -5.16504765e-01
6.62131011e-01 8.90016705e-02 4.38439608e-01 6.49505794e-01
-2.33459339e-01 6.89894736e-01 4.20449108e-01 -8.00886214e-01
-9.01954532e-01 -4.72991496e-01 -5.46176851e-01 7.24987149e-01
-1.86613813e-01 -1.32776946e-01 -5.17810225e-01 -9.33570564e-01
-5.75442042e-04 -1.45587802e-01 -1.17712283e+00 -4.51175217e-03
-5.87755680e-01 -1.12893784e+00 6.60859942e-01 7.09026217e-01
5.44155896e-01 -9.78260577e-01 -7.21242189e-01 1.93672091e-01
1.80003062e-01 -9.11249697e-01 1.37586653e-01 5.83746672e-01
-1.10339785e+00 -1.19801211e+00 -8.74558270e-01 -8.98857176e-01
1.08089697e+00 3.34221095e-01 7.43908882e-01 6.23068929e-01
-7.54470050e-01 5.47198653e-02 -4.09998029e-01 -5.10908782e-01
-2.91189492e-01 5.66030920e-01 -4.99846399e-01 2.38708451e-01
1.55001789e-01 -4.86166209e-01 -8.38041604e-01 4.34040278e-02
-1.18886244e+00 1.28242731e-01 1.33738804e+00 1.30618107e+00
6.06839299e-01 -5.51573783e-02 7.31496811e-01 -1.17808795e+00
3.44901115e-01 -5.08715153e-01 -6.92778379e-02 1.15659133e-01
-3.21924031e-01 -2.82062322e-01 6.30679905e-01 -2.45062962e-01
-8.62249732e-01 1.87664047e-01 -4.50355798e-01 1.85470253e-01
-3.48961532e-01 7.71470070e-01 1.37098864e-01 -6.75166428e-01
6.23247504e-01 1.07634023e-01 4.36952859e-01 -2.74750113e-01
-2.72000432e-01 5.57701826e-01 1.31337389e-01 -2.94150114e-01
5.00134170e-01 6.83924258e-01 4.34465289e-01 -7.56534159e-01
-5.97786307e-01 -7.08557367e-01 -9.11498129e-01 -1.69904530e-01
7.93837726e-01 -6.23583138e-01 -5.96869648e-01 7.03798652e-01
-7.28888035e-01 -1.61564201e-01 4.02993560e-02 4.11902487e-01
-2.10876510e-01 1.92410469e-01 -7.04050720e-01 -4.58119214e-01
-4.19454038e-01 -1.17641747e+00 9.88868833e-01 2.88478196e-01
1.34680048e-02 -1.10794294e+00 1.29255699e-03 2.07429200e-01
7.22105801e-01 5.31912684e-01 1.09529817e+00 -9.75915015e-01
-4.37409639e-01 -6.84986949e-01 -3.01520914e-01 3.91428500e-01
3.08512092e-01 3.79909314e-02 -1.18773270e+00 -5.27691185e-01
-3.03730935e-01 -4.48110491e-01 9.53759432e-01 4.58198071e-01
1.43323004e+00 -8.00101608e-02 -7.59435296e-01 7.25436926e-01
1.80498302e+00 -5.27593158e-02 8.02310765e-01 6.68487012e-01
2.44969904e-01 4.86176610e-01 3.50020051e-01 1.12339936e-01
6.34816512e-02 5.13337255e-01 5.43061435e-01 -7.60685802e-01
-3.91398706e-02 1.70349792e-01 -4.37331170e-01 4.20358837e-01
-1.80303380e-01 4.76425812e-02 -1.04603982e+00 8.06292951e-01
-1.80861020e+00 -5.26568413e-01 2.03809142e-01 2.12377810e+00
6.87909424e-01 1.29933655e-01 -6.12187199e-02 2.64772475e-01
3.62684608e-01 -2.11772323e-01 -2.61324227e-01 -2.19386235e-01
-6.68388233e-02 4.65593427e-01 3.39107126e-01 -2.38134816e-01
-1.23549891e+00 5.39821982e-01 6.60023022e+00 7.60470986e-01
-1.52145159e+00 6.44756365e-04 9.48880851e-01 2.17698708e-01
4.32441607e-02 -2.65759081e-01 -5.27956545e-01 1.37660846e-01
7.35841274e-01 4.77455884e-01 -1.62215099e-01 6.23964250e-01
3.84710124e-03 -2.65561789e-01 -1.04124045e+00 4.93687779e-01
-7.78699517e-02 -1.67062366e+00 5.23933396e-02 2.30187640e-01
7.06680179e-01 2.93068495e-02 1.70217708e-01 1.19818725e-01
-2.72738755e-01 -1.29904354e+00 9.55648050e-02 3.36421251e-01
5.86505473e-01 -5.41812837e-01 1.49277353e+00 2.76506066e-01
-9.52209294e-01 -8.64196494e-02 -3.23336244e-01 -1.02724251e-03
-4.90796894e-01 5.87635934e-01 -1.20563006e+00 8.06153953e-01
7.51936615e-01 5.79414368e-01 -7.55721629e-01 1.28203535e+00
2.65472621e-01 5.68556070e-01 -1.92101061e-01 -1.39344573e-01
5.01569092e-01 1.81083396e-01 1.02125918e-02 1.55659711e+00
3.91272128e-01 -1.76997364e-01 1.26417607e-01 5.32523632e-01
7.68630728e-02 5.41052461e-01 -4.43576396e-01 -7.90535584e-02
2.11959213e-01 1.85772347e+00 -1.06647813e+00 -1.07413821e-01
-5.61428070e-01 4.82922882e-01 5.41287959e-01 5.32230698e-02
-3.78080398e-01 -3.75850737e-01 2.58704752e-01 2.44304717e-01
4.48341936e-01 1.21244401e-01 -4.41396743e-01 -7.77073383e-01
-1.77808821e-01 -7.68566489e-01 6.00449979e-01 -2.50602275e-01
-1.25028539e+00 8.38774383e-01 -3.45850348e-01 -1.32661033e+00
2.02183723e-01 -9.28573847e-01 -7.02389300e-01 7.77461171e-01
-1.84845519e+00 -1.78125155e+00 -7.55227685e-01 4.27917093e-01
1.68702736e-01 -2.07470939e-01 1.26413524e+00 1.91573143e-01
-6.55773640e-01 5.97453415e-01 1.25826418e-01 3.27686161e-01
6.41206622e-01 -1.09958816e+00 -3.10343832e-01 4.63407636e-01
-2.49200791e-01 7.25264788e-01 2.66775548e-01 -2.18544990e-01
-1.17837191e+00 -1.03784227e+00 7.10042119e-01 1.80385802e-02
5.18727243e-01 -2.26582348e-01 -7.68336535e-01 4.30654347e-01
1.09110668e-01 4.84514207e-01 1.20402014e+00 8.87257308e-02
-1.00843042e-01 -4.15138245e-01 -1.39678550e+00 3.40241402e-01
4.79262233e-01 -2.68908441e-01 -8.95320997e-02 1.56856403e-01
1.13831334e-01 -4.01930511e-01 -1.06585622e+00 6.77992344e-01
7.07759202e-01 -9.72665727e-01 8.14967990e-01 -5.01927435e-01
4.88428116e-01 -1.28609911e-01 1.30867613e-02 -1.17198277e+00
-4.39728856e-01 -1.63195878e-01 4.49119091e-01 1.04627514e+00
4.75745350e-01 -7.93218255e-01 8.30426514e-01 1.51155755e-01
-3.16328913e-01 -1.41458094e+00 -1.11679196e+00 -4.13241327e-01
2.31037691e-01 -1.38582205e-02 3.87652129e-01 7.09453821e-01
7.13675395e-02 -2.32057899e-01 1.03185423e-01 -2.05986071e-02
3.02327991e-01 1.01418225e-02 5.53755820e-01 -1.20943809e+00
-3.29147130e-02 -4.99586225e-01 -9.58771467e-01 -4.03054655e-01
-2.08551553e-03 -1.09561372e+00 -8.94429162e-02 -1.43410635e+00
5.47917843e-01 -7.49507248e-01 -8.52763295e-01 6.71552181e-01
-1.71984732e-01 7.46007085e-01 -6.69170097e-02 2.40912050e-01
-3.80974174e-01 4.57021594e-02 1.22034192e+00 -1.81696355e-01
-2.76624784e-02 -1.07132494e-02 -8.02925348e-01 2.74184346e-01
7.90623546e-01 -3.05838078e-01 -3.10794152e-02 -1.82270724e-02
-4.30604406e-02 -3.47193390e-01 5.16146004e-01 -1.36490333e+00
3.18503141e-01 1.38625592e-01 8.31537783e-01 -4.05533910e-01
1.82840124e-01 -8.05112779e-01 1.64941013e-01 7.50575721e-01
-3.70011836e-01 -2.94022620e-01 4.34928328e-01 5.08643210e-01
-5.12518525e-01 -2.13242605e-01 9.21947360e-01 -3.24245960e-01
-4.26289171e-01 3.78611326e-01 -5.44504464e-01 -6.33363843e-01
1.25660622e+00 -5.05952477e-01 -1.74830943e-01 1.12044774e-01
-5.30264258e-01 -1.91175580e-01 3.88897628e-01 4.53658588e-02
4.53853488e-01 -1.25947309e+00 -6.58161342e-01 2.98328638e-01
4.00422156e-01 3.57470624e-02 5.28263509e-01 1.33386004e+00
-7.62749135e-01 6.63290918e-01 -5.31025767e-01 -9.58801448e-01
-1.25220227e+00 3.90622497e-01 4.93282378e-01 -6.08668745e-01
-6.07656062e-01 7.79603124e-01 2.80818731e-01 -4.86615896e-01
-1.41493957e-02 -4.37104076e-01 -4.75400150e-01 -7.32718408e-02
3.14979374e-01 -8.77884403e-02 6.46247447e-01 -3.88446242e-01
-4.29459065e-01 3.96503836e-01 -3.37515771e-01 3.66213739e-01
1.72004461e+00 3.69688690e-01 -2.44984403e-01 2.25939795e-01
1.34672284e+00 -3.97421151e-01 -8.80831778e-01 -2.64277250e-01
-3.05628441e-02 -3.20226759e-01 3.00875843e-01 -8.31114650e-01
-1.17587996e+00 8.77002656e-01 6.64287210e-01 8.08547065e-02
1.36352646e+00 -5.02574332e-02 6.01851642e-01 1.40649244e-01
3.29574138e-01 -9.30261910e-01 1.06032118e-01 2.02378392e-01
5.53336143e-01 -1.34022462e+00 -8.58498588e-02 -5.27439475e-01
-2.31641069e-01 1.49602973e+00 4.76273656e-01 -1.53607577e-01
8.60111713e-01 2.69891471e-01 2.58286178e-01 -3.02472591e-01
-6.65110409e-01 -2.56318152e-01 4.59554613e-01 5.00354409e-01
7.81134367e-01 -4.15744670e-02 -4.51516747e-01 6.29088402e-01
2.97581077e-01 2.76968002e-01 2.47042343e-01 1.14259756e+00
-2.33515382e-01 -1.09412956e+00 1.18263215e-02 6.65986061e-01
-8.66489768e-01 2.06965543e-02 -2.65389591e-01 1.00791943e+00
2.70545095e-01 6.24633551e-01 -2.11105570e-01 -3.08409303e-01
6.31925911e-02 -1.36917055e-01 3.97659779e-01 -6.98721349e-01
-1.15586436e+00 9.79982987e-02 -5.35733514e-02 -4.98854578e-01
-6.78264558e-01 -5.58342040e-01 -8.87345135e-01 4.44882177e-02
-4.67913628e-01 -1.11335918e-01 7.04574764e-01 1.03890598e+00
3.92720997e-01 7.17602551e-01 5.48588872e-01 -7.47380018e-01
-3.29814315e-01 -1.09784412e+00 -5.85517406e-01 4.66551393e-01
2.96421230e-01 -6.10686481e-01 -1.21004887e-01 -1.42709658e-01] | [15.067785263061523, -2.8901560306549072] |
36f566a1-9e44-421e-87b2-ba2c15207f2f | collaborative-auto-encoding-for-blind-image | 2305.14684 | null | https://arxiv.org/abs/2305.14684v1 | https://arxiv.org/pdf/2305.14684v1.pdf | Collaborative Auto-encoding for Blind Image Quality Assessment | Blind image quality assessment (BIQA) is a challenging problem with important real-world applications. Recent efforts attempting to exploit powerful representations by deep neural networks (DNN) are hindered by the lack of subjectively annotated data. This paper presents a novel BIQA method which overcomes this fundamental obstacle. Specifically, we design a pair of collaborative autoencoders (COAE) consisting of a content autoencoder (CAE) and a distortion autoencoder (DAE) that work together to extract content and distortion representations, which are shown to be highly descriptive of image quality. While the CAE follows a standard codec procedure, we introduce the CAE-encoded feature as an extra input to the DAE's decoder for reconstructing distorted images, thus effectively forcing DAE's encoder to extract distortion representations. The self-supervised learning framework allows the COAE including two feature extractors to be trained by almost unlimited amount of data, thus leaving limited samples with annotations to finetune a BIQA model. We will show that the proposed BIQA method achieves state-of-the-art performance and has superior generalization capability over other learning based models. The codes are available at: https://github.com/Macro-Zhou/NRIQA-VISOR/. | ['Guoping Qiu', 'Fei Zhou', 'Zehong Zhou'] | 2023-05-24 | null | null | null | null | ['blind-image-quality-assessment', 'image-quality-assessment'] | ['computer-vision', 'computer-vision'] | [-1.08907551e-01 -1.72177434e-01 1.95714071e-01 -3.33445549e-01
-8.65064919e-01 -4.68459755e-01 4.11972076e-01 -3.32266808e-01
-3.07865441e-01 5.09323895e-01 4.60568756e-01 -2.11708277e-01
5.61870895e-02 -7.36508608e-01 -5.98596573e-01 -8.48283172e-01
9.41193849e-02 -7.77214020e-02 -7.85510913e-02 -1.97947294e-01
1.82135612e-01 3.87906790e-01 -1.50222695e+00 2.48572126e-01
1.11916947e+00 1.10310149e+00 2.81155348e-01 9.14700925e-01
2.69821435e-01 1.00778365e+00 -7.55218744e-01 -6.14160836e-01
4.64316130e-01 -6.15816653e-01 -5.53222775e-01 2.65217006e-01
3.13252866e-01 -8.81759286e-01 -7.37187564e-01 1.35661650e+00
8.72035623e-01 2.18451023e-02 6.96264505e-01 -1.07581460e+00
-1.19218242e+00 9.96536091e-02 -1.92017987e-01 3.10734540e-01
1.23432264e-01 3.02239478e-01 1.01649690e+00 -9.84644353e-01
2.31652573e-01 9.19736505e-01 4.24614072e-01 8.45062852e-01
-9.15143549e-01 -6.20346129e-01 -3.41241807e-01 4.68867779e-01
-1.37920940e+00 -7.65404105e-01 9.18094516e-01 -3.87179971e-01
7.68321812e-01 1.27986789e-01 4.75046277e-01 1.08343816e+00
1.10789955e-01 8.68433416e-01 1.02744758e+00 -3.96357000e-01
5.56279600e-01 2.26773560e-01 -3.57249945e-01 7.02654362e-01
-8.01888481e-02 4.15119708e-01 -2.93991357e-01 -6.21969998e-03
1.05131459e+00 2.68444307e-02 -6.28553748e-01 -3.70399177e-01
-8.65400791e-01 9.17751133e-01 6.67073429e-01 3.07838768e-01
-5.86073756e-01 -4.80576754e-02 2.65787721e-01 5.10422349e-01
2.64320850e-01 2.61496991e-01 -2.16198877e-01 -6.65327832e-02
-9.37162101e-01 -1.48912575e-02 4.79993880e-01 7.58026719e-01
6.61149800e-01 3.25234741e-01 -1.38447776e-01 1.19649363e+00
3.39302689e-01 5.34061313e-01 7.91083515e-01 -1.10542834e+00
3.00590545e-01 4.23226476e-01 7.95159712e-02 -1.00516951e+00
1.94593500e-02 -5.48529565e-01 -1.12601888e+00 6.81306899e-01
2.52757192e-01 -1.71070457e-01 -8.96820545e-01 1.49794781e+00
-3.51927578e-02 -1.98650420e-01 2.44296610e-01 1.15760410e+00
5.93460023e-01 8.92548442e-01 -1.56847477e-01 -9.83898938e-02
1.06511438e+00 -1.11731327e+00 -8.80627811e-01 5.49124833e-03
2.57964492e-01 -5.87004721e-01 9.71926570e-01 7.14571059e-01
-1.28912258e+00 -8.59703898e-01 -1.24780786e+00 -2.28288740e-01
-2.50954300e-01 5.01513004e-01 3.63715857e-01 6.64693594e-01
-1.27896845e+00 5.35290182e-01 -5.98186672e-01 -1.00149356e-01
8.03164303e-01 3.71236145e-01 -3.80052924e-01 -2.93359965e-01
-9.75973666e-01 8.88715208e-01 1.54190704e-01 2.81452179e-01
-1.42377865e+00 -2.65812188e-01 -7.87455440e-01 6.54938295e-02
9.95246693e-02 -7.03995645e-01 1.21679580e+00 -1.40389013e+00
-1.79717755e+00 5.66493452e-01 -1.97124518e-02 -3.77210826e-01
2.69477636e-01 -2.06225410e-01 -8.04203689e-01 4.32808459e-01
-1.11180671e-01 6.10001862e-01 1.25243402e+00 -1.41117907e+00
-5.42597115e-01 -2.38812998e-01 2.04506963e-01 2.14072287e-01
-6.58605099e-01 -1.30619099e-02 -4.82634127e-01 -9.73121285e-01
-3.63149233e-02 -4.05394137e-01 5.43499961e-02 3.67608100e-01
-2.26997405e-01 -3.93273123e-02 6.24517858e-01 -9.35200572e-01
1.33355737e+00 -2.19749022e+00 2.37022370e-01 4.38802131e-03
4.39351261e-01 6.09650552e-01 -3.57294500e-01 3.80727172e-01
-9.11421478e-02 -2.27856353e-01 -5.32403827e-01 -3.89105231e-01
-1.19073868e-01 2.44489327e-01 -1.57435238e-01 5.61051846e-01
3.75123143e-01 7.44954824e-01 -7.97525167e-01 -2.16346771e-01
2.66344011e-01 5.94520330e-01 -7.21602440e-01 7.91815102e-01
8.62600058e-02 2.99389929e-01 -2.22846135e-01 7.11956620e-01
7.58550286e-01 -2.04708770e-01 -1.06790543e-01 -3.96960109e-01
4.55859043e-02 7.70648047e-02 -1.19202721e+00 1.61059165e+00
-4.75546032e-01 7.02846289e-01 -3.82813415e-03 -1.00964904e+00
9.20936286e-01 5.38146436e-01 1.29380673e-01 -9.15178299e-01
3.46689820e-01 2.72627741e-01 6.07485436e-02 -6.62559509e-01
1.85942769e-01 -2.82574922e-01 3.98496270e-01 5.02521932e-01
6.51625693e-01 4.57085762e-03 -3.43410596e-02 6.79537281e-02
1.02528203e+00 -1.25458583e-01 3.31226826e-01 6.03807569e-02
7.67326474e-01 -4.95948195e-01 6.91780984e-01 4.48414296e-01
-6.40988648e-01 8.54618430e-01 7.92428851e-02 -4.06743437e-01
-1.40832889e+00 -1.19481480e+00 2.79777814e-02 7.63625145e-01
1.08668007e-01 -2.51670092e-01 -8.51848245e-01 -6.09178662e-01
-2.77302355e-01 4.52819377e-01 -6.01163566e-01 -3.20065439e-01
-2.38986999e-01 -4.46804523e-01 3.90774637e-01 6.45238161e-01
8.20659995e-01 -1.06243694e+00 -3.39831620e-01 5.71812987e-02
-1.41136348e-01 -8.16619039e-01 -5.39475679e-01 -4.17046137e-02
-7.78870523e-01 -8.80799592e-01 -1.06739080e+00 -7.88144827e-01
7.67421067e-01 2.70044178e-01 9.39915121e-01 1.52509019e-01
-1.10421553e-01 3.13359201e-01 -4.61685866e-01 -2.58689195e-01
-5.56621432e-01 -5.07184088e-01 2.02650324e-01 2.73593366e-01
4.01211083e-01 -7.86322236e-01 -8.40767086e-01 2.24063575e-01
-1.14165974e+00 -7.07846284e-02 7.25283265e-01 1.05098748e+00
5.51260233e-01 3.29657346e-01 6.81183219e-01 -4.20035720e-01
7.15148926e-01 -3.30027521e-01 -6.23957098e-01 2.38415357e-02
-5.66403985e-01 -2.54050102e-02 7.53474116e-01 -2.76787996e-01
-1.08015740e+00 -1.36156768e-01 -4.52751309e-01 -6.36833787e-01
-2.14929953e-01 2.20022678e-01 -4.91870016e-01 -2.25659728e-01
8.30000281e-01 3.73709023e-01 3.10672913e-02 -5.83255649e-01
3.47109675e-01 1.15769625e+00 8.08427930e-01 -1.52068853e-01
9.92273986e-01 2.79260159e-01 -5.71722806e-01 -5.52480996e-01
-6.93963647e-01 -3.78421813e-01 -5.29064834e-01 -2.73695707e-01
8.53272200e-01 -1.21804059e+00 -4.85389382e-01 6.97105527e-01
-1.02254891e+00 -3.22125196e-01 -3.25886667e-01 5.17284155e-01
-6.54070258e-01 3.74718726e-01 -7.43910611e-01 -7.05322385e-01
-4.27948684e-01 -1.08713078e+00 6.31165564e-01 3.88006717e-01
2.86956966e-01 -8.63423347e-01 1.14113756e-01 5.15581667e-01
4.63339478e-01 -1.97496891e-01 7.30212390e-01 -3.77595425e-01
-3.96279484e-01 -2.13874757e-01 -3.50041986e-01 1.23694515e+00
3.07070106e-01 -3.17456126e-01 -1.38774920e+00 -4.70710337e-01
4.10664022e-01 -4.96539533e-01 7.36370742e-01 3.88087422e-01
1.42882848e+00 -4.96016592e-01 3.33128572e-01 7.77062416e-01
1.58577764e+00 3.54883999e-01 9.57412183e-01 2.70804167e-01
5.11666358e-01 2.40841106e-01 -1.31741641e-02 3.52862149e-01
4.10502940e-01 4.62021112e-01 5.29636621e-01 -1.70739263e-01
-4.56705302e-01 -2.33015358e-01 5.50929606e-01 1.27131128e+00
-3.18848670e-01 -3.01000893e-01 -6.80415630e-01 7.21831560e-01
-1.45848441e+00 -9.31561291e-01 3.25306267e-01 1.93093228e+00
9.23052847e-01 -6.75462559e-02 3.02686580e-02 4.91091967e-01
5.12521505e-01 4.88136411e-02 -5.92857361e-01 -3.64014953e-01
-1.42002285e-01 1.35556951e-01 2.20451534e-01 5.02515614e-01
-1.11278784e+00 6.41398430e-01 5.91151953e+00 7.71049798e-01
-9.33409512e-01 2.10773483e-01 3.85329574e-01 1.85650364e-01
-1.68558076e-01 -3.58948112e-01 -1.89053208e-01 4.95212257e-01
1.05986714e+00 1.01804800e-01 8.05042684e-01 7.08799362e-01
1.77136362e-01 3.50526646e-02 -9.74054277e-01 1.15191483e+00
3.62193853e-01 -1.07229733e+00 3.00034851e-01 3.03064380e-02
7.68726408e-01 -7.13699386e-02 2.98363656e-01 2.02670649e-01
3.05372894e-01 -1.02484417e+00 6.19441450e-01 6.82394505e-01
8.95754278e-01 -6.27369404e-01 8.53707254e-01 2.07781896e-01
-6.65741265e-01 -6.42786324e-01 -7.11426914e-01 2.11937539e-02
-4.39471193e-02 6.29480243e-01 -4.34260637e-01 5.89616895e-01
7.71087229e-01 8.29273701e-01 -5.73292971e-01 1.24991620e+00
-3.63210201e-01 7.67824590e-01 2.88410932e-01 4.70546216e-01
4.12126742e-02 -1.18785515e-01 5.97874343e-01 9.77494061e-01
3.98181826e-01 3.35395515e-01 -3.51192713e-01 8.85568380e-01
-4.28070486e-01 -1.27720356e-01 -2.65151918e-01 -4.97206636e-02
2.69163996e-01 1.05404377e+00 -4.28663939e-02 -3.84686112e-01
-5.93958259e-01 1.41921544e+00 2.13436171e-01 5.60554266e-01
-4.19708133e-01 -4.98216212e-01 7.16961801e-01 -2.70559490e-01
4.94175971e-01 -6.01499565e-02 -1.11405127e-01 -1.43378055e+00
1.28327161e-02 -1.31049132e+00 1.27028212e-01 -1.20973301e+00
-1.55364335e+00 8.15002739e-01 -5.92750847e-01 -1.69322503e+00
-1.81699455e-01 -8.59744132e-01 -5.07152975e-01 1.05515528e+00
-1.80724716e+00 -8.97761881e-01 -4.67882842e-01 1.00594902e+00
6.57193899e-01 -5.69969594e-01 1.02355957e+00 5.95899582e-01
-5.69792330e-01 7.17580020e-01 3.50053251e-01 5.27684808e-01
7.42798090e-01 -1.30823588e+00 1.63064823e-02 1.04512918e+00
2.03757510e-01 3.77367318e-01 5.86056232e-01 -2.76790053e-01
-1.34090018e+00 -1.07311916e+00 6.26595020e-01 -2.99791157e-01
4.29178268e-01 -2.00787589e-01 -1.02212882e+00 4.58486229e-01
3.52418184e-01 3.84938508e-01 7.80922115e-01 -2.96320707e-01
-5.42396545e-01 -3.91179740e-01 -1.16243637e+00 3.37512016e-01
7.60783732e-01 -1.09225321e+00 -7.22595990e-01 1.02780312e-01
4.29598778e-01 -2.26910099e-01 -8.50073457e-01 -8.99893511e-03
4.29547995e-01 -1.19487190e+00 8.62238705e-01 -2.88051158e-01
8.43006551e-01 -4.64796126e-01 -4.50926363e-01 -1.58737016e+00
-4.43797797e-01 -2.39021897e-01 -4.47493523e-01 1.10364127e+00
1.36782587e-01 -4.23525482e-01 3.39172661e-01 1.89017877e-01
-2.13874266e-01 -6.72366560e-01 -7.98860013e-01 -7.29634106e-01
9.98456478e-02 -3.29318136e-01 5.63030839e-01 7.99216449e-01
-2.30073839e-01 8.87712911e-02 -6.29092932e-01 3.98828715e-01
6.09796822e-01 -2.12279469e-01 4.41058844e-01 -8.55162442e-01
-4.73539829e-01 -3.66886020e-01 -5.00455856e-01 -1.19521832e+00
-2.05533877e-01 -7.35563934e-01 9.73083898e-02 -1.49176443e+00
1.62727475e-01 -1.39300331e-01 -5.16146660e-01 3.14804614e-01
-7.77620822e-02 4.65071261e-01 1.50695398e-01 3.42361212e-01
-4.64912891e-01 9.00326312e-01 1.41818750e+00 -3.23928028e-01
-9.95277092e-02 -3.51421386e-02 -7.54612386e-01 5.60465693e-01
7.87611365e-01 -2.48452067e-01 -4.66748923e-01 -8.54544759e-01
6.47561401e-02 9.15035009e-02 5.30667305e-01 -1.26516306e+00
3.76202166e-01 2.28024542e-01 7.97540188e-01 -1.34760335e-01
1.88786343e-01 -8.41903210e-01 -2.21309662e-01 1.71550930e-01
-3.33401918e-01 -1.69489115e-01 8.16225484e-02 4.68211412e-01
-5.81486881e-01 -3.25723022e-01 8.20171177e-01 -4.76030968e-02
-6.19660795e-01 4.32179093e-01 -3.44069928e-01 -4.00331885e-01
5.70490539e-01 -1.19032659e-01 -1.95602208e-01 -7.17117250e-01
-6.08819127e-01 -1.18179049e-03 3.66157621e-01 2.80815452e-01
9.49482918e-01 -1.53889370e+00 -7.62614608e-01 3.17985386e-01
1.33049011e-01 -1.81712449e-01 5.36070108e-01 3.71334970e-01
-4.06603068e-01 1.30986586e-01 -5.53658664e-01 -2.01228559e-01
-8.36334050e-01 7.54518569e-01 5.99252939e-01 6.72406554e-02
-6.01787627e-01 9.06105518e-01 1.21817172e-01 -2.32680053e-01
3.01204056e-01 5.35297543e-02 -3.48745733e-01 -1.55925408e-01
9.00487125e-01 2.94674784e-01 1.06748402e-01 -6.63080394e-01
-2.98652332e-02 3.53374511e-01 6.47375733e-02 -6.96659237e-02
1.49593389e+00 -4.16213036e-01 7.98434136e-04 1.09610043e-01
1.40709531e+00 -1.00506201e-01 -1.50884640e+00 -3.88437897e-01
-3.51186037e-01 -6.32613778e-01 5.22265434e-01 -1.19178045e+00
-1.32096827e+00 1.27566636e+00 1.15332901e+00 5.23135774e-02
1.80094481e+00 -1.67498469e-01 8.34740877e-01 2.06774682e-01
1.48658931e-01 -1.10726726e+00 3.23354751e-01 1.32018581e-01
1.22615194e+00 -1.47034061e+00 -2.25409836e-01 1.90927088e-01
-6.99497998e-01 1.14852774e+00 3.88537735e-01 -2.65887171e-01
5.92701316e-01 1.26828160e-02 3.26291710e-01 -7.47583807e-02
-5.04249871e-01 -1.49140954e-01 4.67418045e-01 8.31902564e-01
9.20389816e-02 5.62641062e-02 -5.62214199e-03 9.12131846e-01
-2.07620189e-01 1.55176014e-01 6.18937731e-01 6.04546428e-01
-3.68315220e-01 -1.01787448e+00 -2.92250544e-01 3.61745685e-01
-4.32413012e-01 -1.41799957e-01 5.27682230e-02 2.25277662e-01
3.28881621e-01 1.08595574e+00 3.60710993e-02 -5.49679399e-01
1.93267271e-01 -1.64017990e-01 3.34137022e-01 -3.49359930e-01
-2.94069380e-01 5.73838949e-02 -3.05782020e-01 -5.88113546e-01
-5.28503299e-01 -3.45734626e-01 -7.81840682e-01 -1.42915174e-01
-1.86039180e-01 -1.30232237e-02 5.41585445e-01 8.14003587e-01
1.89246327e-01 5.39135337e-01 1.11408448e+00 -7.77961254e-01
-5.99523306e-01 -1.04183018e+00 -6.88791871e-01 3.51736605e-01
1.00540864e+00 -4.78789419e-01 -3.71617168e-01 4.86451238e-01] | [11.866552352905273, -1.8257113695144653] |
2d029c68-81e8-4e90-819d-9fc8000c456c | contrast-and-generation-make-bart-a-good | 2112.11202 | null | https://arxiv.org/abs/2112.11202v2 | https://arxiv.org/pdf/2112.11202v2.pdf | Contrast and Generation Make BART a Good Dialogue Emotion Recognizer | In dialogue systems, utterances with similar semantics may have distinctive emotions under different contexts. Therefore, modeling long-range contextual emotional relationships with speaker dependency plays a crucial part in dialogue emotion recognition. Meanwhile, distinguishing the different emotion categories is non-trivial since they usually have semantically similar sentiments. To this end, we adopt supervised contrastive learning to make different emotions mutually exclusive to identify similar emotions better. Meanwhile, we utilize an auxiliary response generation task to enhance the model's ability of handling context information, thereby forcing the model to recognize emotions with similar semantics in diverse contexts. To achieve these objectives, we use the pre-trained encoder-decoder model BART as our backbone model since it is very suitable for both understanding and generation tasks. The experiments on four datasets demonstrate that our proposed model obtains significantly more favorable results than the state-of-the-art model in dialogue emotion recognition. The ablation study further demonstrates the effectiveness of supervised contrastive loss and generative loss. | ['Xipeng Qiu', 'Hang Yan', 'ShiMin Li'] | 2021-12-21 | null | null | null | null | ['emotion-recognition-in-conversation'] | ['natural-language-processing'] | [ 1.02277227e-01 -6.40660450e-02 -3.87526527e-02 -9.31417465e-01
-4.77226496e-01 -3.02960575e-01 6.19999588e-01 2.58559622e-02
-3.94236892e-01 7.49974906e-01 5.98913014e-01 1.79504827e-01
5.71038008e-01 -4.06460434e-01 -2.17091665e-01 -5.45215189e-01
2.70030141e-01 1.65718243e-01 -3.67041767e-01 -6.16910458e-01
-6.37747813e-03 -1.35216787e-01 -1.30330157e+00 4.82869416e-01
1.13308012e+00 1.29114771e+00 9.69061479e-02 3.03737849e-01
-4.10464585e-01 1.01293695e+00 -7.80936241e-01 -7.80670106e-01
-2.85029411e-01 -9.19574320e-01 -8.34006310e-01 1.36507109e-01
-2.13152736e-01 -1.69996202e-01 -8.42806250e-02 9.56967711e-01
6.20947778e-01 4.10495162e-01 8.17267597e-01 -1.19618201e+00
-6.74682617e-01 7.22671032e-01 -3.82797480e-01 -1.68041527e-01
6.51214302e-01 -7.22975954e-02 1.31617713e+00 -8.77625525e-01
2.07483947e-01 1.42859459e+00 2.88954884e-01 8.69080901e-01
-9.64292049e-01 -8.32754135e-01 4.55463290e-01 2.02204376e-01
-1.00027549e+00 -5.75166106e-01 1.25164759e+00 1.96372569e-01
7.53743052e-01 2.52930403e-01 4.19456840e-01 1.57589424e+00
-2.33632743e-01 1.22452235e+00 1.17585003e+00 -2.65612036e-01
4.18308005e-02 5.77767015e-01 4.90836352e-02 3.60830247e-01
-7.37138271e-01 -3.97222847e-01 -7.10562706e-01 -1.46722067e-02
2.67804086e-01 -5.20195328e-02 -6.09113097e-01 -2.83163367e-03
-9.58805442e-01 9.71653104e-01 3.84714961e-01 3.33777040e-01
-3.54200840e-01 -7.94036835e-02 8.94425333e-01 6.72274411e-01
6.77891731e-01 4.99550939e-01 -5.36643505e-01 -3.14813495e-01
-2.57813543e-01 -1.02520138e-01 8.42687786e-01 7.61826336e-01
5.14142871e-01 5.65031469e-02 -2.24708334e-01 1.40747869e+00
4.17962313e-01 8.11253563e-02 6.64559722e-01 -5.13823926e-01
3.48155588e-01 5.99582195e-01 -1.09425552e-01 -1.05052364e+00
-2.80751348e-01 -2.85394460e-01 -1.14993441e+00 -4.38702703e-01
6.62545785e-02 -4.78324562e-01 -4.25858170e-01 2.23327327e+00
2.12790012e-01 1.88295111e-01 6.64647818e-01 1.00553930e+00
9.70115185e-01 9.40870225e-01 3.48922133e-01 -3.97029668e-01
1.34431171e+00 -1.14290774e+00 -9.55323756e-01 -2.66787142e-01
5.51429212e-01 -7.34036088e-01 1.70527911e+00 1.90355808e-01
-7.84004927e-01 -4.65273291e-01 -8.79731238e-01 -5.94625659e-02
-2.20173076e-01 1.84416011e-01 8.15650761e-01 4.55274999e-01
-4.89967793e-01 4.85606231e-02 -3.22258443e-01 -1.22433282e-01
1.71462089e-01 -7.78629072e-03 -3.64762470e-02 1.48938000e-01
-1.87711799e+00 8.88270378e-01 1.25909820e-01 1.79512903e-01
-5.71721673e-01 -2.26722673e-01 -1.10594440e+00 1.45601764e-01
3.29586327e-01 -4.50393051e-01 1.20575404e+00 -1.34695876e+00
-2.05078483e+00 7.72447348e-01 -3.48747164e-01 -8.25694650e-02
3.88441920e-01 -2.70709962e-01 -5.86145043e-01 2.83324998e-02
-1.79165557e-01 5.62054694e-01 8.25430632e-01 -1.22044766e+00
-4.02324319e-01 -1.50203660e-01 1.63694739e-01 6.42703712e-01
-7.17238963e-01 1.82640806e-01 -3.90126586e-01 -7.43738294e-01
-1.11082457e-01 -7.72185862e-01 -1.26475230e-01 -4.12878007e-01
-3.97586375e-01 -4.67823416e-01 5.91168284e-01 -4.17814672e-01
1.08874643e+00 -2.13912296e+00 2.30679661e-01 -1.72427506e-03
-1.03762500e-01 -1.66750997e-01 -1.64238214e-01 2.69473076e-01
1.08401909e-01 -1.18213994e-02 -3.40995938e-01 -5.10980129e-01
2.46941119e-01 1.20310418e-01 -4.43979561e-01 -4.81926985e-02
4.86619055e-01 7.47428834e-01 -8.49235058e-01 -3.94562274e-01
4.41381820e-02 4.16817844e-01 -4.28555757e-01 5.93827426e-01
-2.13709980e-01 7.50145376e-01 -7.50709951e-01 5.03068209e-01
3.34415078e-01 -2.19838291e-01 3.26880932e-01 -2.39705548e-01
3.79731983e-01 5.42396724e-01 -8.46455753e-01 1.82821274e+00
-1.03165984e+00 3.19701076e-01 8.21456686e-02 -1.12213838e+00
1.20497239e+00 5.70460320e-01 1.72866017e-01 -8.05718660e-01
3.51554781e-01 4.08353545e-02 -1.25528112e-01 -5.47419071e-01
7.18791783e-01 -6.37634099e-01 -6.72780752e-01 5.18126786e-01
2.87822238e-03 -2.23513067e-01 -2.21333653e-01 3.17213163e-02
5.04447937e-01 -1.59132391e-01 1.47309288e-01 -1.34406192e-02
8.87892425e-01 -4.52240944e-01 8.14530134e-01 3.84892166e-01
-4.23881650e-01 2.68286407e-01 6.15242481e-01 -6.31732121e-02
-3.98194462e-01 -7.15362966e-01 -1.15744807e-01 1.37521076e+00
4.82071549e-01 -2.38173217e-01 -4.10268813e-01 -1.03191161e+00
-4.01113123e-01 8.36724341e-01 -5.40602863e-01 -4.47740853e-01
-2.75335878e-01 -7.78645515e-01 6.32161975e-01 3.77145976e-01
7.22135007e-01 -1.31334805e+00 -3.83556247e-01 2.33718917e-01
-7.02744544e-01 -9.50000584e-01 -4.45734501e-01 2.39087403e-01
-3.19628984e-01 -6.99466825e-01 -7.15056896e-01 -9.61689293e-01
4.61079478e-01 7.81737864e-02 1.24610758e+00 -1.17316730e-01
2.71529675e-01 2.16907769e-01 -8.00166309e-01 -4.64622766e-01
-4.29562747e-01 6.25315756e-02 -1.27758354e-01 2.10109845e-01
6.47151768e-01 -4.43179429e-01 -3.93443793e-01 2.21556276e-01
-8.72160435e-01 2.89200246e-02 2.65167505e-01 1.28971004e+00
3.61652285e-01 5.64601831e-02 1.16253257e+00 -9.17285860e-01
1.25376403e+00 -5.71091592e-01 1.16882786e-01 4.77193624e-01
-3.56725216e-01 4.36206758e-02 1.09783483e+00 -5.24270177e-01
-1.58845997e+00 -2.66794771e-01 -2.57404953e-01 -2.06901804e-01
-1.28869995e-01 6.39294028e-01 -6.16551101e-01 4.32187825e-01
1.58119619e-01 2.65266091e-01 -6.25538751e-02 -1.89794853e-01
5.23148358e-01 9.28172946e-01 1.87721491e-01 -7.41933286e-01
7.03194886e-02 1.02989785e-01 -5.52177131e-01 -6.65848017e-01
-9.08488393e-01 -3.14233541e-01 -5.06066643e-02 -1.92577064e-01
7.66672373e-01 -9.64864373e-01 -6.06878638e-01 5.35800040e-01
-1.21697211e+00 -1.67520598e-01 1.89222507e-02 4.91285741e-01
-4.32965308e-01 2.68320829e-01 -9.11654890e-01 -8.62687588e-01
-4.92525756e-01 -1.22134650e+00 9.79111731e-01 5.62747657e-01
-3.56592387e-01 -1.04985917e+00 -4.48695607e-02 3.08734566e-01
4.30448562e-01 -4.50312234e-02 1.00963843e+00 -8.63233805e-01
2.01469421e-01 1.60983428e-01 -1.24984622e-01 5.96411407e-01
2.91593432e-01 -2.72578806e-01 -1.17968714e+00 1.66141406e-01
3.34575504e-01 -1.10568869e+00 9.05230641e-01 -1.23331368e-01
1.16931725e+00 -1.61389917e-01 1.35847509e-01 2.74487257e-01
9.05914843e-01 2.20289454e-01 5.48839033e-01 -1.46972667e-02
5.88053465e-01 1.17603040e+00 1.00982285e+00 5.71893454e-01
6.70238853e-01 5.04524231e-01 1.37568116e-01 -2.22932830e-01
4.53443676e-01 -1.47388726e-01 5.21519244e-01 1.21789443e+00
4.03352320e-01 -3.55859429e-01 -5.33328414e-01 3.42844069e-01
-1.93813515e+00 -9.45468009e-01 2.12470770e-01 1.82739508e+00
1.31694818e+00 5.78754349e-03 -1.98295236e-01 -3.11757237e-01
7.39207387e-01 6.77072763e-01 -6.75622523e-01 -7.54039705e-01
-2.74968654e-01 6.37004003e-02 -4.29848433e-01 3.17511618e-01
-1.10887599e+00 1.27302885e+00 5.34102154e+00 8.61131012e-01
-1.34140265e+00 -6.75288290e-02 1.11680937e+00 -7.41194561e-02
-5.61083555e-01 -2.03196079e-01 -4.51963097e-01 5.79110682e-01
6.40163064e-01 -2.00898454e-01 2.55400747e-01 7.78536856e-01
1.35830477e-01 2.11914659e-01 -8.88119280e-01 1.07620621e+00
2.81125903e-01 -6.17538035e-01 1.13231763e-02 -5.08167148e-01
5.40585935e-01 -4.34275866e-01 -5.04974229e-03 8.32281530e-01
1.58088386e-01 -9.90232468e-01 3.97434026e-01 4.52924192e-01
5.19034803e-01 -1.07751119e+00 8.39578032e-01 1.68259025e-01
-1.02104175e+00 2.21795276e-01 -4.93554294e-01 -7.77675584e-02
2.69927442e-01 4.96614128e-01 -5.72452247e-01 4.54473346e-01
3.76723468e-01 8.25897455e-01 -1.35358736e-01 1.67349517e-01
-5.01592398e-01 6.02317035e-01 -7.63832331e-02 -4.51410830e-01
2.19779894e-01 -4.09977466e-01 2.19019800e-01 1.36794543e+00
2.77204234e-02 1.96502492e-01 2.96291113e-01 6.82513654e-01
-5.46035469e-01 6.66977882e-01 -3.78882915e-01 -8.90429020e-02
5.24833083e-01 1.29850423e+00 -2.41213158e-01 -4.19520736e-01
-5.44840991e-01 1.29599643e+00 4.25276190e-01 3.56141001e-01
-1.01900876e+00 -5.75296044e-01 9.34520185e-01 -8.46414626e-01
-9.19494219e-03 2.72531599e-01 -6.36089891e-02 -1.47349191e+00
2.29496309e-05 -1.17729080e+00 3.08610082e-01 -6.26192868e-01
-1.74872231e+00 9.19259429e-01 -4.89381135e-01 -1.13468647e+00
-4.07213867e-01 -2.94399858e-01 -7.75273442e-01 9.10607100e-01
-1.79615009e+00 -1.09970272e+00 -1.66464418e-01 7.42346168e-01
7.94304967e-01 -1.33124525e-02 1.04087365e+00 8.32036585e-02
-6.97777510e-01 8.46609056e-01 -1.35159910e-01 9.57512707e-02
1.25349975e+00 -1.26296389e+00 -2.24837348e-01 4.38646913e-01
-1.66967064e-01 5.17185628e-01 5.61851263e-01 -2.62099087e-01
-1.09863889e+00 -9.27960634e-01 8.53003979e-01 8.61410648e-02
5.81494808e-01 -3.31262529e-01 -1.12535644e+00 2.25964814e-01
5.92199028e-01 -5.31160653e-01 1.27188456e+00 5.12092650e-01
-3.67198318e-01 -1.19938953e-02 -9.12493706e-01 9.46201265e-01
7.54787326e-01 -8.56315553e-01 -7.39584625e-01 7.53256679e-02
8.14785063e-01 -1.70819864e-01 -8.39251518e-01 5.41306913e-01
4.97512579e-01 -9.18352604e-01 6.20945394e-01 -6.65052831e-01
8.94018888e-01 1.43558249e-01 -1.40555844e-01 -1.62171149e+00
2.35370398e-01 -6.36230707e-01 1.15401953e-01 1.60353041e+00
5.08287489e-01 -5.75178266e-01 4.86087531e-01 8.52891922e-01
-5.77085428e-02 -9.07891035e-01 -6.16868138e-01 -2.45748967e-01
-2.55548004e-02 -3.63772720e-01 7.67695189e-01 1.44916320e+00
4.89882290e-01 1.07107222e+00 -6.89615965e-01 -2.10871860e-01
-4.66028973e-02 5.29460311e-01 7.58562088e-01 -8.91648591e-01
-3.22660238e-01 -5.81926167e-01 3.21702771e-02 -1.45908332e+00
8.83084297e-01 -6.22609258e-01 2.89206654e-01 -1.02052629e+00
8.64930227e-02 -5.83039165e-01 -5.65015554e-01 3.36058944e-01
-7.27761805e-01 -9.97297168e-02 -1.28928535e-02 -2.29473263e-02
-8.72856379e-01 1.44817376e+00 1.30429077e+00 -1.47595257e-01
-2.43018880e-01 -6.50048777e-02 -1.00088930e+00 6.75684035e-01
9.29895997e-01 -2.15009987e-01 -6.91322625e-01 -2.89535433e-01
1.58731401e-01 2.13803634e-01 -8.75839144e-02 -3.45095873e-01
-1.30617130e-03 -2.22081274e-01 1.55224562e-01 -2.27166817e-01
6.39144123e-01 -6.59087360e-01 -5.60498059e-01 -2.41538920e-02
-9.24799800e-01 -8.32929537e-02 3.72901596e-02 4.93092477e-01
-7.87924528e-01 -2.50357360e-01 5.13373792e-01 -2.99433805e-02
-8.28226328e-01 1.77646309e-01 -2.74948031e-01 3.80718976e-01
6.87159419e-01 1.52745411e-01 -1.20887265e-01 -1.00987184e+00
-5.20492733e-01 5.98797917e-01 3.78482580e-01 6.97297812e-01
6.90685451e-01 -1.28210342e+00 -7.61438370e-01 6.64122077e-03
3.97293061e-01 -7.85017684e-02 4.01956469e-01 5.87186098e-01
1.78416848e-01 -1.14499489e-02 -2.37495527e-02 -2.30132759e-01
-1.28248882e+00 2.54702449e-01 3.30646724e-01 -4.17294443e-01
-6.46679997e-02 1.00517523e+00 4.86694485e-01 -8.46034706e-01
3.46637219e-01 -4.47717793e-02 -5.54824531e-01 2.39347219e-01
4.47927326e-01 -1.46007478e-01 -3.53377044e-01 -6.67878628e-01
-2.10546598e-01 1.27167493e-01 -3.29455614e-01 -1.00555457e-01
1.10357893e+00 -5.32254219e-01 -8.39990526e-02 6.90855980e-01
1.37202370e+00 1.11783154e-01 -1.24848151e+00 -4.85271245e-01
-2.26344943e-01 -1.65751502e-01 -3.68563049e-02 -8.64068508e-01
-9.10520911e-01 1.06690359e+00 1.43485203e-01 2.25569427e-01
1.43370736e+00 -5.59516586e-02 9.40128267e-01 5.34122646e-01
1.61163434e-01 -1.32096517e+00 4.02941942e-01 8.98186147e-01
9.09094453e-01 -1.54284477e+00 -4.40725267e-01 -4.50491369e-01
-1.50769639e+00 9.57938433e-01 8.91472161e-01 2.02946454e-01
2.86580920e-01 -9.82277691e-02 5.13644457e-01 -8.80076215e-02
-9.41249967e-01 -1.29175693e-01 1.41444802e-01 2.32909471e-01
8.67780387e-01 1.43489614e-01 -5.17287731e-01 9.78030324e-01
-2.25353405e-01 -4.88718003e-01 2.57081598e-01 6.46203995e-01
-9.52709541e-02 -1.26998496e+00 1.59616604e-01 2.39714146e-01
-5.22849619e-01 -3.24124217e-01 -7.30299890e-01 4.33999479e-01
-3.98407727e-01 1.26840055e+00 -1.52435541e-01 -4.38586801e-01
2.40088001e-01 3.40295136e-01 1.40980333e-01 -5.14590323e-01
-7.47506201e-01 1.32137910e-01 5.09947419e-01 -2.96618968e-01
-6.31226122e-01 -3.30831468e-01 -1.41385615e+00 1.24843884e-02
-5.28068602e-01 4.45686787e-01 4.66031969e-01 9.89272594e-01
3.38839352e-01 6.59965456e-01 1.30903184e+00 -3.72546017e-01
-7.35688150e-01 -1.06094408e+00 -6.48744404e-01 8.64708841e-01
-4.05612737e-02 -5.55941999e-01 -3.52450579e-01 -1.20505974e-01] | [13.01626968383789, 6.094055652618408] |
d6645a16-70e6-4043-9247-4e4a6f09b77f | data-poisoning-attacks-on-eeg-signal-based | 2302.04224 | null | https://arxiv.org/abs/2302.04224v1 | https://arxiv.org/pdf/2302.04224v1.pdf | Data Poisoning Attacks on EEG Signal-based Risk Assessment Systems | Industrial insider risk assessment using electroencephalogram (EEG) signals has consistently attracted a lot of research attention. However, EEG signal-based risk assessment systems, which could evaluate the emotional states of humans, have shown several vulnerabilities to data poison attacks. In this paper, from the attackers' perspective, data poison attacks involving label-flipping occurring in the training stages of different machine learning models intrude on the EEG signal-based risk assessment systems using these machine learning models. This paper aims to propose two categories of label-flipping methods to attack different machine learning classifiers including Adaptive Boosting (AdaBoost), Multilayer Perceptron (MLP), Random Forest, and K-Nearest Neighbors (KNN) dedicated to the classification of 4 different human emotions using EEG signals. This aims to degrade the performance of the aforementioned machine learning models concerning the classification task. The experimental results show that the proposed data poison attacks are model-agnostically effective whereas different models have different resilience to the data poison attacks. | ['Chan Yeob Yeun', 'Ernesto Damiani', 'Sangyoung Yoon', 'Ahmed Y. Al Hammadi', 'Sani Umar', 'Zhibo Zhang'] | 2023-02-08 | null | null | null | null | ['data-poisoning'] | ['adversarial'] | [ 2.31045652e-02 -1.86241493e-01 4.68738198e-01 -3.30509245e-01
-3.16134959e-01 -9.51012969e-01 4.28663343e-01 5.73345542e-01
-5.70332468e-01 8.76013815e-01 -3.35138619e-01 -4.38857943e-01
-2.49584332e-01 -5.75067282e-01 -6.17500782e-01 -8.89659464e-01
-5.40052593e-01 -1.61441818e-01 -1.10586032e-01 1.15094922e-01
6.92406416e-01 8.59462619e-01 -1.43179727e+00 5.35406768e-01
6.48259938e-01 1.29202318e+00 -6.93239868e-01 4.92212504e-01
5.69910944e-01 9.42156613e-01 -1.14316726e+00 -6.87163949e-01
1.97130308e-01 -9.37187299e-02 -5.85505724e-01 -9.04568374e-01
-1.87256679e-01 -1.11045688e-01 -1.24657704e-02 1.41699815e+00
7.55629539e-01 -5.20461440e-01 6.77509308e-01 -1.96285725e+00
-1.42402291e-01 7.38516569e-01 -2.95895398e-01 5.56190848e-01
5.60890019e-01 1.73164710e-01 7.44887516e-02 -3.78864616e-01
-1.60920098e-01 8.58458698e-01 5.99022269e-01 6.98624790e-01
-1.12538970e+00 -1.53131580e+00 -3.40375245e-01 6.86371326e-01
-1.61996579e+00 1.10865682e-01 9.91949737e-01 -4.98592705e-01
6.25709593e-01 4.61757779e-01 6.57012999e-01 1.65234387e+00
1.03623712e+00 1.00022860e-01 1.84986579e+00 -2.15471730e-01
9.57957506e-01 4.96458679e-01 7.02277780e-01 9.84533131e-02
3.71102989e-01 2.91197240e-01 -1.06626940e+00 -1.03302586e+00
-9.19039845e-02 -2.95577139e-01 -3.69927496e-01 2.76059449e-01
-5.69836378e-01 7.60700703e-01 3.30797046e-01 3.74495924e-01
-5.67961991e-01 7.73328319e-02 7.99990654e-01 4.06282634e-01
3.90260130e-01 6.30891263e-01 -6.85053945e-01 -7.81444833e-02
-3.21258932e-01 -1.55540556e-02 8.54245543e-01 4.19043869e-01
2.69926667e-01 -1.90916173e-02 7.51904994e-02 -5.44518642e-02
3.02390873e-01 5.35198212e-01 6.94856822e-01 1.38348132e-01
3.62342954e-01 3.39237750e-01 5.05629368e-02 -1.24165988e+00
-6.89092696e-01 -3.60585928e-01 -8.08308423e-01 4.46416616e-01
2.06555709e-01 -6.11500084e-01 -1.60055324e-01 1.49241614e+00
1.64944947e-01 4.39347863e-01 1.63555101e-01 4.06660140e-01
3.17575097e-01 4.42331463e-01 6.92458212e-01 -7.94261098e-02
1.43531239e+00 -2.07333907e-01 -7.71615446e-01 1.46522209e-01
6.05623424e-01 -2.08662063e-01 7.99745440e-01 1.14986420e+00
-6.42970800e-01 -3.17647517e-01 -1.56202447e+00 9.66226876e-01
-8.95750821e-01 -4.62406635e-01 6.41755641e-01 1.91364920e+00
-6.84630156e-01 4.49440271e-01 -5.91358840e-01 1.71604022e-01
4.52436090e-01 7.29449511e-01 -3.33631247e-01 4.45435971e-01
-1.69288182e+00 1.23058891e+00 2.58995593e-01 -1.08175362e-02
-1.03124940e+00 -6.28163695e-01 -3.31183672e-01 1.55209508e-02
-4.85353529e-01 -5.21336980e-02 5.61497331e-01 -7.09492862e-01
-1.35837328e+00 5.40626764e-01 5.74200571e-01 -9.13212359e-01
2.98585981e-01 -8.18999186e-02 -7.74605751e-01 1.62858590e-01
-4.82370913e-01 4.22022864e-02 1.06845129e+00 -8.22680891e-01
-2.63909161e-01 -7.17599154e-01 -2.99472153e-01 -2.72078753e-01
-6.58644378e-01 5.07471859e-01 1.11192024e+00 -5.49157977e-01
-4.56571072e-01 -6.35128379e-01 1.97773263e-01 -8.90399873e-01
-5.08218586e-01 -1.37151852e-01 7.10144043e-01 -5.11015475e-01
1.20908296e+00 -2.19357705e+00 -3.32138538e-01 5.91064990e-01
-1.24863654e-01 2.05206886e-01 3.49653393e-01 3.16507846e-01
-5.27581751e-01 2.42844269e-01 -1.80287406e-01 1.74509898e-01
5.41500971e-02 -4.33010846e-01 -7.26483643e-01 9.68513012e-01
-2.12399438e-01 5.46442032e-01 -4.31675494e-01 1.11396782e-01
2.70846516e-01 4.23733652e-01 -2.08163727e-02 1.95684612e-01
5.18246055e-01 4.31449652e-01 -4.00460631e-01 3.20709705e-01
6.67230546e-01 6.79512799e-01 -2.78829094e-02 -2.17947736e-01
3.46123993e-01 1.37126610e-01 -9.69020426e-01 8.63148510e-01
-3.14413548e-01 2.95514941e-01 -3.90658885e-01 -7.74251044e-01
1.07359934e+00 6.86020434e-01 5.65532818e-02 -6.18730605e-01
5.70188940e-01 1.56807080e-01 2.01005042e-01 -3.79932672e-01
-2.65964329e-01 -1.44384444e-01 -4.51920152e-01 6.61361516e-01
9.95159615e-03 1.55746996e-01 -6.52792931e-01 -8.42780247e-02
1.53143311e+00 -4.17356074e-01 4.03252423e-01 -6.08236909e-01
6.84588552e-01 -3.13340187e-01 3.00381333e-01 6.61121905e-01
-3.95235777e-01 -2.29650080e-01 3.49089384e-01 -5.36332726e-01
-3.57681870e-01 -1.00892591e+00 -4.89229470e-01 6.04112983e-01
2.31466904e-01 -1.49716958e-01 -1.43415070e+00 -9.12312448e-01
-8.74254107e-02 1.29371047e+00 -7.65681505e-01 -8.69427025e-01
-1.68914832e-02 -1.29305387e+00 1.38885379e+00 3.18940639e-01
6.66110992e-01 -1.23541832e+00 -1.19755018e+00 1.04040608e-01
2.77122170e-01 -7.34278083e-01 3.78938615e-01 8.40632975e-01
-6.75050855e-01 -1.21802723e+00 -1.44557180e-02 -1.01183884e-01
4.71346766e-01 -5.30358434e-01 5.59414744e-01 -1.24816798e-01
-2.72849351e-01 4.41229761e-01 -5.95767319e-01 -1.08070624e+00
-5.11517406e-01 -1.41660541e-01 5.99754810e-01 3.45874906e-01
8.93042982e-01 -9.03806627e-01 -2.78081685e-01 4.03720409e-01
-9.63325620e-01 -8.30205321e-01 2.21124515e-01 1.89623281e-01
-3.28851104e-01 6.17574811e-01 1.15115106e+00 -6.36205375e-01
1.30283880e+00 -6.54945016e-01 -4.71788526e-01 3.39821011e-01
-6.01193309e-01 -3.32696363e-02 8.17281961e-01 -8.06501746e-01
-7.25252211e-01 -1.29223883e-01 -2.13162050e-01 -1.00893825e-01
-6.94565654e-01 2.33384401e-01 -5.09244621e-01 -6.22307837e-01
1.05053127e+00 1.49143517e-01 -7.43931711e-01 -2.92470276e-01
-1.72294468e-01 9.75627959e-01 2.47384578e-01 -3.07547033e-01
6.32953763e-01 1.54509142e-01 -8.76009017e-02 -4.64283049e-01
-2.36701220e-01 -6.81549236e-02 -3.24364424e-01 -5.39516568e-01
9.07515228e-01 -6.40650988e-01 -1.16408515e+00 1.15192676e+00
-1.50150454e+00 9.15567577e-02 4.13393259e-01 5.38506150e-01
-1.56994730e-01 1.67073280e-01 -6.88193798e-01 -1.12179720e+00
-8.73341799e-01 -1.04189682e+00 3.75923574e-01 3.32369506e-02
-5.05551696e-01 -5.99277377e-01 8.55096430e-02 1.13298945e-01
2.98885375e-01 4.28208292e-01 1.02756214e+00 -1.41487503e+00
2.97615975e-01 -7.16491461e-01 4.68002319e-01 5.20927966e-01
-8.87734443e-02 -3.75320822e-01 -1.55375767e+00 -2.38308161e-01
9.69783068e-01 -4.61565524e-01 2.68294793e-02 -4.28576916e-02
1.11060083e+00 -4.67628896e-01 -1.54113069e-01 2.87802011e-01
1.16718459e+00 8.55938375e-01 9.29051161e-01 4.13757771e-01
1.69244006e-01 7.51093507e-01 2.98041105e-01 4.96621460e-01
-2.05158591e-01 3.89946550e-01 7.60914862e-01 3.61487091e-01
9.82050717e-01 6.93383515e-02 7.55743742e-01 3.08528572e-01
3.46777827e-01 -2.72998437e-02 -9.25995052e-01 -2.29877919e-01
-1.32974589e+00 -8.40032876e-01 -2.03302503e-01 2.31020951e+00
6.61689341e-01 3.45026463e-01 9.29010361e-02 8.48691165e-01
8.66139650e-01 -3.36844116e-01 -6.12421095e-01 -7.85949409e-01
4.98211458e-02 4.79527205e-01 6.82644844e-01 -8.82692188e-02
-1.15737784e+00 4.65723872e-01 5.42561150e+00 9.72563028e-01
-1.14097738e+00 3.99172843e-01 6.25114262e-01 5.99525534e-02
2.87205130e-01 -3.73656243e-01 -5.55628061e-01 9.94843602e-01
1.35918164e+00 -1.24375015e-01 4.25673544e-01 8.78860235e-01
4.13768515e-02 -2.11679429e-01 -1.09146917e+00 1.30107081e+00
2.03134671e-01 -7.49177396e-01 -2.10085139e-01 6.14760304e-03
-3.22339274e-02 -1.82164580e-01 2.57182419e-02 1.33649409e-01
2.03296527e-01 -1.08602774e+00 8.27495694e-01 4.33715671e-01
1.85109749e-01 -1.42793632e+00 8.91717494e-01 7.13025451e-01
-5.40855050e-01 -2.81747818e-01 -1.70701176e-01 -4.43773985e-01
-8.44504535e-02 5.95284998e-01 -3.34652752e-01 2.62322009e-01
1.17263377e+00 -1.44047707e-01 -8.46920550e-01 7.66713381e-01
-5.87685645e-01 9.62415874e-01 -2.44360819e-01 -3.24335337e-01
-1.67365015e-01 2.46368542e-01 3.45834374e-01 8.88950765e-01
-2.06370890e-01 2.08667964e-01 -6.25521719e-01 5.95531642e-01
2.59006739e-01 2.20991634e-02 -7.33545899e-01 4.35820699e-01
6.27033889e-01 1.30021536e+00 -7.98249483e-01 6.45220950e-02
1.00886591e-01 7.68226087e-01 1.91536034e-03 -6.60084784e-02
-1.09383857e+00 -5.92074096e-01 3.24923784e-01 -2.88581967e-01
-5.54389477e-01 8.25028792e-02 -5.31608224e-01 -8.60459030e-01
-1.49759397e-01 -1.06050837e+00 3.77328664e-01 -7.72177994e-01
-1.53532851e+00 9.01123941e-01 1.66919351e-01 -9.57592726e-01
1.44260988e-01 -5.80769897e-01 -6.81304753e-01 6.99420214e-01
-8.15255761e-01 -6.65236592e-01 4.92420085e-02 8.83227110e-01
-1.35285988e-01 -4.12488163e-01 1.09533012e+00 8.76330659e-02
-4.65990186e-01 6.81375861e-01 -4.00633126e-01 2.91676491e-01
4.55023527e-01 -7.43525386e-01 1.75924182e-01 8.00759733e-01
1.77508205e-01 6.67354465e-01 5.67565560e-01 -5.97916543e-01
-9.99401987e-01 -7.66219556e-01 5.37352026e-01 -5.23650408e-01
6.39237881e-01 -7.49050319e-01 -7.19015181e-01 3.77867222e-01
4.17813987e-01 -3.48845452e-01 1.24944091e+00 -2.01986149e-01
-5.21373153e-01 -4.76774387e-02 -1.80251575e+00 4.19755191e-01
9.37035456e-02 -7.18452215e-01 -6.75451636e-01 1.01915218e-01
2.52736688e-01 2.53389627e-01 -9.86980855e-01 2.88129747e-01
5.21746039e-01 -1.00768995e+00 8.16924870e-01 -7.60329902e-01
-2.28009522e-01 -1.69036850e-01 -1.52933031e-01 -1.37981105e+00
1.70313213e-02 -5.74395418e-01 5.10138348e-02 1.31114149e+00
2.28812873e-01 -1.19172490e+00 3.67286295e-01 9.35654521e-01
5.04486144e-01 -5.01304507e-01 -1.26365995e+00 -6.16242111e-01
2.10189462e-01 -7.28400946e-01 9.11385894e-01 1.20734715e+00
7.64697433e-01 4.72209901e-01 -3.49138379e-01 6.68297231e-01
7.14570284e-01 -6.91692948e-01 4.18057829e-01 -1.06144524e+00
-5.69508262e-02 -1.02530666e-01 -1.08051908e+00 2.23845229e-01
9.53171477e-02 -7.19261229e-01 -3.94527823e-01 -4.38842058e-01
-1.06654555e-01 -3.38265151e-01 -6.83155954e-01 6.68362617e-01
-6.69696033e-02 3.85533869e-01 2.13647947e-01 -7.28920847e-02
-2.14170218e-01 3.41497779e-01 1.66090336e-02 -2.12983429e-01
1.04589246e-01 2.96233982e-01 -3.52879703e-01 8.93604457e-01
1.22931099e+00 -1.41759408e+00 -3.14743787e-01 1.16299570e-01
5.40779769e-01 -1.85007513e-01 5.89539230e-01 -1.60009897e+00
5.08576214e-01 3.57879788e-01 5.09915531e-01 -2.46854305e-01
-5.70568219e-02 -1.21188390e+00 3.06847751e-01 8.28178585e-01
-3.40231270e-01 4.77887005e-01 5.02911150e-01 5.37711024e-01
6.28769398e-02 -5.45755923e-01 7.87196279e-01 1.30774170e-01
-9.90256388e-03 -1.51909187e-01 -1.13269341e+00 -4.77845967e-01
1.58380258e+00 -1.66063249e-01 -3.25880229e-01 -1.13819085e-01
-6.07990384e-01 -1.97248235e-01 2.23934706e-02 2.15403587e-01
5.79879880e-01 -9.09117579e-01 -3.18360507e-01 2.18838453e-01
6.86305165e-02 -6.85606778e-01 1.27590269e-01 6.63397193e-01
-2.54928797e-01 1.55990839e-01 -7.03520656e-01 -1.90946534e-01
-1.42504096e+00 9.49364245e-01 4.37851936e-01 -1.63269758e-01
-9.03744530e-03 7.63205111e-01 -1.60033673e-01 -3.42029601e-01
5.40142536e-01 9.43330824e-02 -4.01468813e-01 2.43245825e-01
6.62665904e-01 7.69897163e-01 6.37196422e-01 -2.35613778e-01
-7.93643117e-01 4.66586947e-02 5.68778329e-02 -1.78240314e-01
9.89612222e-01 4.08993363e-01 -2.46659860e-01 3.28564942e-01
1.02164173e+00 -9.69684571e-02 -5.08817852e-01 5.00813007e-01
3.30278605e-01 -5.40568940e-02 1.28242195e-01 -1.32992089e+00
-6.70551300e-01 1.13048685e+00 1.01870370e+00 5.96132338e-01
1.35168386e+00 -6.02850378e-01 3.27250957e-01 3.33424598e-01
1.15158379e+00 -7.82653332e-01 -1.25496030e-01 -1.81646079e-01
4.95216489e-01 -5.39076567e-01 -1.50689855e-01 1.91787854e-02
-6.86203659e-01 9.92054582e-01 3.17524195e-01 -1.36970058e-01
1.00954747e+00 7.02068210e-01 5.62892519e-02 -2.07347706e-01
-5.53211927e-01 8.51050079e-01 -7.84657896e-02 9.42877233e-01
-2.50211120e-01 1.54278025e-01 -4.91237313e-01 1.27690685e+00
-1.66269749e-01 -7.54881352e-02 6.52089238e-01 9.14271176e-01
-1.30946070e-01 -1.01391780e+00 -7.63091147e-01 4.41487342e-01
-7.66471505e-01 -2.62157470e-01 -6.92001700e-01 3.97291124e-01
2.23585367e-01 1.41137350e+00 -3.45035493e-01 -9.37670350e-01
2.10161939e-01 4.36068863e-01 2.45423958e-01 -2.36999229e-01
-1.50939000e+00 -6.56253934e-01 -1.89059734e-01 -4.82357472e-01
-2.67853379e-01 -5.18589795e-01 -9.10549045e-01 -1.88835993e-01
-6.00095570e-01 4.66777474e-01 1.10851991e+00 9.87965941e-01
4.26177651e-01 2.45280191e-01 9.45228755e-01 -6.57587886e-01
-7.69217193e-01 -1.14710748e+00 -9.80010033e-01 1.79894015e-01
-1.23663448e-01 -7.72445023e-01 -9.18302655e-01 -1.41679421e-01] | [13.266497611999512, 3.084388017654419] |
578fa9d9-a9b4-44a0-88a2-c06bd6da826b | rtfe-a-recursive-temporal-fact-embedding | 2009.14653 | null | https://arxiv.org/abs/2009.14653v4 | https://arxiv.org/pdf/2009.14653v4.pdf | RTFE: A Recursive Temporal Fact Embedding Framework for Temporal Knowledge Graph Completion | Static knowledge graph (SKG) embedding (SKGE) has been studied intensively in the past years. Recently, temporal knowledge graph (TKG) embedding (TKGE) has emerged. In this paper, we propose a Recursive Temporal Fact Embedding (RTFE) framework to transplant SKGE models to TKGs and to enhance the performance of existing TKGE models for TKG completion. Different from previous work which ignores the continuity of states of TKG in time evolution, we treat the sequence of graphs as a Markov chain, which transitions from the previous state to the next state. RTFE takes the SKGE to initialize the embeddings of TKG. Then it recursively tracks the state transition of TKG by passing updated parameters/features between timestamps. Specifically, at each timestamp, we approximate the state transition as the gradient update process. Since RTFE learns each timestamp recursively, it can naturally transit to future timestamps. Experiments on five TKG datasets show the effectiveness of RTFE. | ['yang jinrui', 'E Haihong', 'wang haotian', 'Xiaodong Lv', 'wenyu song', 'Meina Song', 'Youri Xu'] | 2020-09-30 | null | https://aclanthology.org/2021.naacl-main.451 | https://aclanthology.org/2021.naacl-main.451.pdf | naacl-2021-4 | ['temporal-knowledge-graph-completion'] | ['knowledge-base'] | [-4.16205436e-01 5.77534456e-03 -3.89860541e-01 -9.38573107e-02
-2.40497533e-02 -3.43914866e-01 7.69735873e-01 2.16157570e-01
-3.09297830e-01 5.01497447e-01 4.40689087e-01 -3.75908285e-01
-2.32840955e-01 -1.08243942e+00 -5.85415781e-01 -5.65315604e-01
-4.68407035e-01 2.11183056e-01 3.71767461e-01 3.48460674e-02
-8.45210850e-02 2.49765247e-01 -9.15312827e-01 -1.59769386e-01
3.97804856e-01 7.64840245e-01 1.47260958e-03 7.33237147e-01
-2.15531275e-01 8.91617656e-01 -3.33753467e-01 -4.85494703e-01
1.43995583e-02 -2.58412123e-01 -7.94350684e-01 -3.17468077e-01
8.82907864e-03 -4.10085320e-01 -1.38562346e+00 9.27822053e-01
3.53636518e-02 3.54968518e-01 3.66659045e-01 -1.74414706e+00
-1.14765716e+00 8.28825295e-01 -2.48728067e-01 2.92365760e-01
2.05299526e-01 1.23371936e-01 1.08123255e+00 -7.64643252e-01
7.80341566e-01 1.17699862e+00 6.77478015e-01 3.79580468e-01
-8.24615002e-01 -6.33675039e-01 7.91817486e-01 5.92618644e-01
-1.21001017e+00 1.35320276e-01 7.93894291e-01 -3.30406308e-01
1.10767400e+00 -2.73003936e-01 1.07929754e+00 9.00160670e-01
4.20832455e-01 9.27010477e-01 7.34243035e-01 -2.34823763e-01
2.17908159e-01 -3.30427438e-01 4.95688409e-01 9.29041982e-01
2.94976175e-01 2.51843333e-01 -8.08064103e-01 -3.33552361e-01
7.00374842e-01 3.97205949e-01 -2.71688968e-01 -3.54719579e-01
-1.28429294e+00 6.94731057e-01 4.42642719e-01 1.97725028e-01
-3.63518894e-01 9.24713850e-01 5.27172029e-01 5.64657986e-01
3.72112870e-01 -1.69153601e-01 -4.80679870e-01 -2.74168193e-01
-4.95732278e-01 -6.41230643e-02 9.40021992e-01 9.93448317e-01
8.74431312e-01 -7.74448691e-03 -1.57302052e-01 1.94444612e-01
4.07378167e-01 3.16802204e-01 5.33846974e-01 -4.60240930e-01
2.66926646e-01 6.72560573e-01 1.50885761e-01 -1.06780958e+00
7.35955173e-03 -2.67440617e-01 -5.06753564e-01 -5.67155719e-01
2.08630368e-01 1.50489986e-01 -1.00089848e+00 1.67029357e+00
5.38383126e-01 1.11959243e+00 7.98448250e-02 3.98512453e-01
4.67091441e-01 9.55203533e-01 9.30198506e-02 -8.26857537e-02
1.10913920e+00 -8.80290389e-01 -8.34933639e-01 9.07083899e-02
8.19472909e-01 -5.07286824e-02 9.28502023e-01 1.40029758e-01
-5.25107861e-01 -2.18564868e-01 -8.76509488e-01 2.52345894e-02
-6.46195471e-01 -1.69149891e-01 8.37084651e-01 2.57516354e-01
-1.18117058e+00 9.19670939e-01 -1.52246141e+00 -5.04505217e-01
1.44659013e-01 1.61105722e-01 -2.59647429e-01 -2.56197482e-01
-1.70048904e+00 6.13985658e-01 9.43761885e-01 2.12172076e-01
-1.31109655e+00 -5.66816151e-01 -9.95955706e-01 7.23321214e-02
6.90403461e-01 -7.01496661e-01 1.08010828e+00 -3.94384950e-01
-1.55441737e+00 3.24570388e-01 -2.88248658e-01 -5.93659103e-01
3.42167437e-01 -3.78858566e-01 -9.29025888e-01 -5.86327091e-02
-1.47073478e-01 -8.90140235e-02 1.03229845e+00 -8.50579500e-01
-5.22407472e-01 -2.76289642e-01 1.90628231e-01 -3.52201611e-02
-5.43004692e-01 -3.17721993e-01 -9.44315791e-01 -4.19285357e-01
1.62812769e-01 -1.00556993e+00 -1.05085485e-01 -1.98213905e-01
-2.08597675e-01 -4.90542293e-01 9.12413418e-01 -7.38657713e-01
1.89903569e+00 -2.33128881e+00 2.36455187e-01 1.13793880e-01
4.92807776e-01 1.23114459e-01 4.19780314e-02 9.98176336e-01
5.84865473e-02 6.98236153e-02 -5.15323430e-02 -4.28232789e-01
2.30668336e-01 6.74650788e-01 -4.72262889e-01 5.28572381e-01
-2.70382404e-01 1.23798478e+00 -1.45818269e+00 -4.39066619e-01
9.57599655e-02 3.29073817e-01 -3.54991227e-01 5.94136007e-02
-3.04939955e-01 -1.90783292e-01 -7.28505075e-01 3.53994727e-01
4.55740809e-01 -6.11403406e-01 4.49482322e-01 -1.63064957e-01
-7.23660886e-02 2.71507412e-01 -9.45782304e-01 1.48490834e+00
-4.63164687e-01 2.95476824e-01 -7.02446759e-01 -6.86677039e-01
7.22378373e-01 3.80321413e-01 4.95440662e-01 -2.49412239e-01
-7.18436763e-02 -4.04508300e-02 -3.68761301e-01 -2.55432725e-01
8.51345062e-01 -6.41133115e-02 -1.92723230e-01 6.55575275e-01
2.42336214e-01 3.86280924e-01 -6.27417415e-02 7.63590574e-01
1.21491122e+00 2.39286646e-01 2.94040084e-01 1.94975048e-01
3.15882593e-01 -2.47699812e-01 7.96169281e-01 5.31045556e-01
-3.92615557e-01 -9.31694284e-02 5.63871503e-01 -4.15565908e-01
-7.50866771e-01 -1.26246786e+00 6.02271795e-01 6.95984960e-01
2.52260119e-01 -1.01765919e+00 -2.81214207e-01 -9.89579082e-01
3.86223555e-01 6.84052706e-01 -9.17497277e-01 -6.93825901e-01
-6.79824829e-01 -3.36911172e-01 4.28435147e-01 7.30404198e-01
5.96722782e-01 -9.41006005e-01 -1.04182243e-01 7.25368261e-01
-1.75446812e-02 -9.34394658e-01 -8.70227039e-01 -1.85713083e-01
-1.02160764e+00 -9.35397506e-01 -3.01615864e-01 -7.00821519e-01
4.47723001e-01 2.68203199e-01 7.49427617e-01 -2.90025212e-02
7.65381828e-02 7.85938263e-01 -5.56716919e-01 1.66592859e-02
-1.85299605e-01 6.61352128e-02 1.80544838e-01 1.98594496e-01
2.84693390e-01 -6.66763008e-01 -7.76569664e-01 1.37828425e-01
-1.09650266e+00 -2.85323355e-02 1.65851548e-01 6.52556300e-01
7.63732016e-01 3.30470502e-01 6.34772480e-01 -9.59830225e-01
5.96792936e-01 -7.13560045e-01 -6.00940764e-01 5.49374223e-01
-1.03258955e+00 1.94310755e-01 7.31616497e-01 -5.96106589e-01
-9.47802484e-01 -4.36921299e-01 3.34534407e-01 -9.89059150e-01
3.97008121e-01 1.04590619e+00 1.10874437e-02 3.64887476e-01
6.80801866e-04 8.11440408e-01 -2.19912395e-01 -5.55707693e-01
7.73211658e-01 2.91430026e-01 5.44637978e-01 -6.69329524e-01
8.92776310e-01 5.93160331e-01 -2.48716846e-01 -5.11433303e-01
-6.62446141e-01 -3.97256166e-01 -2.98307538e-01 -2.68621147e-01
4.60915685e-01 -6.92236722e-01 -8.39036405e-01 6.60775781e-01
-8.34815085e-01 -5.70238590e-01 -4.48922515e-01 6.81029797e-01
-1.87585548e-01 7.47910857e-01 -9.03177202e-01 -6.54118180e-01
-1.36720121e-01 -3.98512125e-01 6.73861086e-01 1.85789764e-01
1.79648325e-01 -1.58871424e+00 4.18228686e-01 -2.97531724e-01
1.08264647e-01 2.57918626e-01 8.65698218e-01 -2.56445348e-01
-8.03777993e-01 -4.97930825e-01 2.45470572e-02 2.53539115e-01
3.55136782e-01 1.49360657e-01 -6.00107551e-01 -6.01175129e-01
-1.81815565e-01 1.83873445e-01 9.49250937e-01 2.04141617e-01
8.66697073e-01 -4.87251520e-01 -7.75551736e-01 7.40128994e-01
1.62996006e+00 4.11810488e-01 5.59549868e-01 3.32482457e-01
8.74904156e-01 -4.37453156e-03 7.38185108e-01 3.53186578e-01
8.68327200e-01 3.24810952e-01 1.78440899e-01 3.96812350e-01
-7.57941529e-02 -1.00408387e+00 7.42706001e-01 1.43787777e+00
-8.18377808e-02 -2.96325952e-01 -7.95860767e-01 9.11349773e-01
-2.02872658e+00 -8.12048018e-01 1.93800032e-02 2.27672935e+00
6.08948767e-01 1.92371890e-01 -1.91604972e-01 -6.68805093e-02
6.31934404e-01 5.97666323e-01 -6.88622594e-01 -2.72209674e-01
3.35783243e-01 -6.93887547e-02 7.74339080e-01 5.46412349e-01
-7.83163786e-01 1.38975251e+00 5.69508219e+00 6.02403998e-01
-1.04800498e+00 2.53181487e-01 -1.36724621e-01 2.46265739e-01
-5.10583282e-01 6.53052986e-01 -8.11074376e-01 6.81314647e-01
9.85868394e-01 -7.77291894e-01 6.07521415e-01 7.29692757e-01
-1.15183696e-01 2.45395660e-01 -1.02740443e+00 8.86474431e-01
-2.17779964e-01 -1.15737247e+00 2.27419734e-01 6.95896614e-03
6.63099527e-01 -7.72601506e-03 -1.10899182e-02 7.17514813e-01
8.94954205e-01 -6.04601979e-01 5.40273607e-01 7.16754735e-01
7.93647468e-01 -8.16918135e-01 3.98802638e-01 1.59058049e-01
-1.93785357e+00 2.22585704e-02 -2.91641742e-01 1.23322614e-01
4.32364583e-01 6.67059958e-01 -1.01013219e+00 1.27447307e+00
6.02633715e-01 1.26907265e+00 -5.09776354e-01 7.85896480e-01
-5.43457925e-01 9.79815543e-01 -2.79746562e-01 -2.82515567e-02
3.42601061e-01 -5.00234008e-01 5.27352273e-01 8.23622465e-01
1.74443185e-01 5.10082357e-02 2.63003767e-01 4.60191488e-01
-1.13533996e-01 -2.76837409e-01 -6.06539786e-01 -6.00143433e-01
7.39817679e-01 8.74951482e-01 -6.91050172e-01 -5.42763174e-01
-5.45302093e-01 1.34000385e+00 4.60681379e-01 6.80115938e-01
-1.05134130e+00 -5.48733592e-01 6.43811345e-01 -1.69972211e-01
5.60449362e-01 -6.30937397e-01 5.53033650e-01 -1.45182192e+00
2.27404218e-02 -1.89077035e-01 9.57772493e-01 -5.99565983e-01
-1.24819839e+00 4.43862408e-01 1.27461761e-01 -1.06711996e+00
1.23407193e-01 -8.48853663e-02 -6.42622471e-01 5.77839673e-01
-1.55372012e+00 -1.17403948e+00 -1.21505201e-01 7.46049702e-01
-3.38434568e-03 4.14074183e-01 6.33030713e-01 1.80032879e-01
-6.19918048e-01 5.50781846e-01 3.77488643e-01 3.02726120e-01
4.91574645e-01 -1.25222504e+00 1.00587702e+00 8.91830504e-01
1.11410730e-01 1.03436053e+00 4.03421402e-01 -1.14605272e+00
-1.63247252e+00 -1.36530125e+00 1.04961538e+00 -2.99611896e-01
1.15430737e+00 -1.81567803e-01 -1.14735043e+00 1.45507061e+00
-2.05302924e-01 2.01983914e-01 3.59152883e-01 1.88679039e-01
-5.95209360e-01 -1.59018800e-01 -6.71001077e-01 6.86834693e-01
1.28336859e+00 -8.87647867e-01 -7.86226332e-01 5.87335862e-02
1.17845809e+00 -2.59788930e-01 -1.21097457e+00 2.74115540e-02
4.72686619e-01 -3.35904360e-01 7.60526597e-01 -8.70218515e-01
-7.45015033e-03 -6.96078181e-01 1.33376122e-01 -1.44808042e+00
-3.92716050e-01 -8.32246006e-01 -9.15127099e-01 1.21412468e+00
-4.73501626e-03 -1.17675638e+00 7.55026996e-01 3.43432158e-01
-1.21906050e-01 -7.87011266e-01 -8.81590962e-01 -1.26417148e+00
-2.17298508e-01 -4.67801690e-01 1.02893662e+00 1.13192296e+00
5.52717485e-02 -9.11828205e-02 -5.78179419e-01 3.68406445e-01
5.80237091e-01 3.07840943e-01 7.67966330e-01 -1.03886235e+00
-3.24856490e-01 9.03864503e-02 -6.15109622e-01 -1.10310376e+00
-3.36787850e-02 -1.18534398e+00 -3.05384189e-01 -1.84417939e+00
8.43377337e-02 -2.28485301e-01 -7.52318203e-01 5.16613483e-01
-3.96184087e-01 -6.12950802e-01 -6.58046454e-02 1.50052324e-01
-7.15361178e-01 9.91030753e-01 1.28964221e+00 -5.83920367e-02
-1.67972237e-01 -3.35937202e-01 -2.75748134e-01 3.73817205e-01
6.28528893e-01 -6.34140372e-01 -8.25250566e-01 -1.27004504e-01
3.83768618e-01 2.41600186e-01 2.77413875e-01 -6.73154652e-01
4.84391332e-01 -1.27977967e-01 -1.82108194e-01 -6.54853284e-01
2.70927697e-01 -6.33328319e-01 5.62769711e-01 5.18880606e-01
-1.97908953e-02 1.92803480e-02 -3.32870148e-02 1.46796632e+00
-2.81493753e-01 1.73351288e-01 2.22385257e-01 1.18755586e-01
-9.84739840e-01 1.08840096e+00 -2.25851670e-01 7.56725743e-02
9.69391048e-01 -3.38926405e-01 -1.96329474e-01 -3.84057343e-01
-9.54967678e-01 5.62669158e-01 5.71133792e-01 4.71409023e-01
8.68120551e-01 -1.54915011e+00 -8.05574581e-02 -2.57025898e-01
4.15259391e-01 -3.39141376e-02 4.40874547e-01 7.71294236e-01
-2.66608477e-01 1.94805011e-01 2.59107023e-01 -1.63459703e-01
-9.04720664e-01 8.34718168e-01 2.69009709e-01 -5.79438329e-01
-1.10649812e+00 5.86284041e-01 3.58538143e-03 -2.42655680e-01
6.85763657e-02 -5.79556942e-01 -1.74281776e-01 -1.88803615e-03
3.61276776e-01 5.77188969e-01 -2.32213497e-01 -1.70413688e-01
-3.39893341e-01 3.72846514e-01 -4.72619861e-01 -1.90013602e-01
1.31060755e+00 -1.21904261e-01 -1.42478734e-01 9.06457961e-01
1.41826618e+00 -2.64587969e-01 -1.29683149e+00 -6.03354812e-01
2.70458311e-01 -6.07290030e-01 3.42412367e-02 -2.30325416e-01
-1.16794813e+00 4.11016464e-01 2.91439086e-01 8.43554288e-02
7.65712440e-01 -3.38345319e-02 1.29922259e+00 3.59313279e-01
7.28020549e-01 -1.01959574e+00 1.16878070e-01 7.76666939e-01
3.10468823e-01 -6.02984011e-01 -4.94598970e-02 -1.96404174e-01
-6.89429462e-01 9.46042657e-01 3.38142008e-01 -1.93520159e-01
1.10086870e+00 -2.13651165e-01 -3.86553198e-01 -5.44927359e-01
-1.02440000e+00 -1.30944774e-02 -1.59134760e-01 4.01209027e-01
-1.32775679e-01 3.52713168e-01 -3.01170051e-01 5.03636837e-01
-5.52446842e-02 4.28077936e-01 5.65140009e-01 1.29182553e+00
-1.08633086e-01 -1.20917988e+00 1.58263937e-01 5.19974351e-01
-1.30005300e-01 -7.60849416e-02 -1.83446392e-01 1.01463771e+00
-2.25062415e-01 5.77931285e-01 -1.54128015e-01 -7.26250470e-01
2.09973186e-01 3.95145953e-01 3.76202762e-01 -6.97163820e-01
-3.24622691e-01 -3.23239088e-01 -1.58176348e-02 -6.68034911e-01
-7.28722140e-02 -7.20187128e-01 -1.46490276e+00 -4.75593448e-01
-4.90369350e-01 4.22862172e-01 3.54554832e-01 5.82816780e-01
4.57414627e-01 6.44810736e-01 6.57896578e-01 -1.29416734e-01
-2.68204242e-01 -6.46745443e-01 -9.97194827e-01 2.89737731e-01
1.27263561e-01 -8.57589424e-01 -5.35103023e-01 6.83254302e-02] | [8.548590660095215, 7.874508380889893] |
25c3354c-c72c-4cb5-b84d-dffbc170d2de | confidence-aware-active-feedback-for | 2110.12255 | null | https://arxiv.org/abs/2110.12255v3 | https://arxiv.org/pdf/2110.12255v3.pdf | Confidence-Aware Active Feedback for Interactive Instance Search | Online relevance feedback (RF) is widely utilized in instance search (INS) tasks to further refine imperfect ranking results, but it often has low interaction efficiency. The active learning (AL) technique addresses this problem by selecting valuable feedback candidates. However, mainstream AL methods require an initial labeled set for a cold start and are often computationally complex to solve. Therefore, they cannot fully satisfy the requirements for online RF in interactive INS tasks. To address this issue, we propose a confidence-aware active feedback method (CAAF) that is specifically designed for online RF in interactive INS tasks. Inspired by the explicit difficulty modeling scheme in self-paced learning, CAAF utilizes a pairwise manifold ranking loss to evaluate the ranking confidence of each unlabeled sample. The ranking confidence improves not only the interaction efficiency by indicating valuable feedback candidates but also the ranking quality by modulating the diffusion weights in manifold ranking. In addition, we design two acceleration strategies, an approximate optimization scheme and a top-K search scheme, to reduce the computational complexity of CAAF. Extensive experiments on both image INS tasks and video INS tasks searching for buildings, landscapes, persons, and human behaviors demonstrate the effectiveness of the proposed method. Notably, in the real-world, large-scale video INS task of NIST TRECVID 2021, CAAF uses 25% fewer feedback samples to achieve a performance that is nearly equivalent to the champion solution. Moreover, with the same number of feedback samples, CAAF's mAP is 51.9%, significantly surpassing the champion solution by 5.9%. Code is available at https://github.com/nercms-mmap/caaf. | ['Longxiang Jiang', 'Chao Liang', 'Yue Zhang'] | 2021-10-23 | null | null | null | null | ['instance-search'] | ['computer-vision'] | [ 1.27201483e-01 -1.90944746e-01 -3.28344822e-01 -3.06992710e-01
-1.19380426e+00 -4.57377791e-01 3.07279944e-01 9.81573015e-02
-7.58551180e-01 5.48605800e-01 2.55804867e-01 -3.88088636e-02
-3.58652979e-01 -5.20907521e-01 -6.07915819e-01 -8.01145852e-01
3.17703420e-03 3.23774725e-01 3.86850595e-01 -2.14086443e-01
3.81008923e-01 3.10495379e-03 -1.46214068e+00 1.06854901e-01
1.43016267e+00 1.03347909e+00 4.78119880e-01 4.59525079e-01
-8.46680775e-02 7.60039806e-01 -4.46249843e-01 -4.98494446e-01
3.22910726e-01 -2.51780339e-02 -6.31738901e-01 -2.26132274e-01
4.79107678e-01 -3.79287899e-01 -2.97577143e-01 1.10224974e+00
7.07334876e-01 5.80764472e-01 4.72237140e-01 -9.98777866e-01
-6.04352593e-01 5.14664292e-01 -5.80171466e-01 3.43608320e-01
4.49108064e-01 2.92970929e-02 1.11858344e+00 -1.35251284e+00
3.83119106e-01 9.55303252e-01 2.78636128e-01 6.39359951e-01
-9.38707054e-01 -8.46666396e-01 5.09650648e-01 6.04519546e-01
-1.58811057e+00 -3.85843247e-01 7.12953568e-01 -2.54597247e-01
4.08159584e-01 5.09053826e-01 5.28465152e-01 8.04391682e-01
-4.56193596e-01 1.33331907e+00 8.00185978e-01 -3.25172722e-01
4.16811138e-01 1.67490497e-01 4.01925087e-01 5.86660922e-01
4.44593132e-02 1.08674243e-01 -9.73361552e-01 -5.03054738e-01
6.38731897e-01 1.63252339e-01 -4.96001661e-01 -4.51783389e-01
-1.03362107e+00 7.12900102e-01 6.93295121e-01 2.76554786e-02
-2.64623970e-01 -5.46140857e-02 1.11951113e-01 3.49922895e-01
5.93026817e-01 2.83260435e-01 -2.83650577e-01 -4.28446233e-02
-6.47321105e-01 5.92791401e-02 3.59374911e-01 6.73944175e-01
7.22089946e-01 -3.53765070e-01 -3.62076998e-01 9.99995887e-01
5.49242198e-01 4.55180138e-01 4.15438384e-01 -8.67598295e-01
6.85515344e-01 8.09222400e-01 5.48679292e-01 -1.06020987e+00
-3.06054670e-02 -4.67067033e-01 -5.97395837e-01 -9.31536313e-03
1.89432845e-01 1.10462792e-01 -7.51054645e-01 1.66525269e+00
5.17955899e-01 7.72309899e-02 -3.46322507e-01 1.24835300e+00
7.17210770e-01 5.71157932e-01 -2.48394860e-03 -5.26996911e-01
8.56305897e-01 -1.05428112e+00 -6.58746958e-01 -1.06702045e-01
4.98154193e-01 -6.19771600e-01 1.40200317e+00 4.50735241e-01
-9.24882591e-01 -7.02993751e-01 -1.02966511e+00 3.31116080e-01
-1.17858261e-01 2.15376332e-01 7.88409173e-01 6.57450736e-01
-1.06250823e+00 3.94882023e-01 -7.54224062e-01 3.95553410e-02
2.96275824e-01 6.68274760e-01 -9.04123560e-02 -2.02800751e-01
-1.35682666e+00 4.69042391e-01 5.41108809e-02 3.59615386e-01
-7.66991913e-01 -5.52114308e-01 -4.89745736e-01 6.70220181e-02
6.99501157e-01 -3.86108488e-01 1.17059183e+00 -1.15175378e+00
-1.49561012e+00 2.50162929e-01 -2.14906737e-01 -2.69792348e-01
7.15991259e-01 -6.88269556e-01 -4.15979445e-01 8.07597563e-02
1.82922836e-02 5.63512802e-01 9.22010005e-01 -9.97709394e-01
-4.84686196e-01 -4.22631294e-01 1.69726685e-01 8.59308064e-01
-8.77887130e-01 -3.98757122e-02 -8.64404321e-01 -4.69793469e-01
7.63823017e-02 -8.69820654e-01 -4.36190784e-01 1.02641821e-01
4.63689379e-02 -5.18844962e-01 5.89618564e-01 -4.06634420e-01
1.47671294e+00 -2.30654955e+00 -6.39605243e-03 3.68318886e-01
2.98074245e-01 3.36748928e-01 -1.75385505e-01 7.94032961e-02
2.14206740e-01 2.81621478e-02 -7.31197596e-02 -2.01314360e-01
-1.23953111e-01 -2.36322016e-01 -1.20138526e-01 3.03586781e-01
-3.30474507e-03 9.62746501e-01 -1.20681202e+00 -4.91679281e-01
4.28891256e-02 3.25518876e-01 -4.53610331e-01 2.57832229e-01
7.75670856e-02 3.77589315e-01 -5.38091660e-01 5.44463098e-01
5.68849564e-01 -6.51454806e-01 2.47374028e-02 -8.91061407e-03
-6.23537302e-02 7.82451183e-02 -1.28281736e+00 1.77411640e+00
-3.54998708e-01 3.79903048e-01 -7.64251947e-02 -7.72090077e-01
9.68983173e-01 1.97716251e-01 5.74232399e-01 -9.65125680e-01
-2.23616257e-01 2.46940166e-01 -2.86612391e-01 -2.65925467e-01
7.07252085e-01 6.59424007e-01 1.57206818e-01 3.70785028e-01
-2.81501919e-01 4.63087380e-01 1.93940505e-01 5.39198458e-01
1.19422138e+00 1.98401008e-02 3.45839038e-02 -7.82967061e-02
7.20256925e-01 -1.73675641e-01 6.79098964e-01 9.06754792e-01
-4.30146605e-01 5.51091790e-01 -2.37806812e-01 -2.60233968e-01
-4.64646697e-01 -1.17115188e+00 9.41580385e-02 1.21069872e+00
7.56233096e-01 -5.68228304e-01 -4.83111560e-01 -1.00274408e+00
-1.99550807e-01 2.37105340e-01 -3.99896234e-01 -1.98451594e-01
-6.39934480e-01 -7.77606368e-01 -5.37052453e-02 3.38646710e-01
7.21715033e-01 -9.53570366e-01 -1.62049323e-01 9.52445567e-02
-5.85880399e-01 -5.94839454e-01 -8.98374319e-01 -1.55471489e-01
-8.14812958e-01 -1.04969001e+00 -9.98325288e-01 -5.58898449e-01
9.57752645e-01 5.93614280e-01 8.67319703e-01 3.42586666e-01
-4.33857106e-02 4.47380930e-01 -6.73011303e-01 -1.29473791e-01
1.90738097e-01 6.52967989e-02 2.01638386e-01 9.09884796e-02
3.83263707e-01 -1.00936905e-01 -1.08870363e+00 8.35001588e-01
-6.79650605e-01 -1.44981265e-01 5.31532407e-01 1.00684583e+00
6.11460567e-01 -5.04832678e-02 7.23914802e-01 -5.77996671e-01
6.28575742e-01 -3.39429915e-01 -6.32912040e-01 4.87798631e-01
-9.53501642e-01 -7.64762685e-02 3.99249941e-01 -6.89098835e-01
-1.11845458e+00 2.66173273e-01 6.67247921e-02 -3.45609605e-01
4.21522826e-01 3.15041721e-01 3.38240489e-02 -2.39649802e-01
9.46399808e-01 1.40514165e-01 -3.26428413e-01 -2.35417575e-01
1.16767995e-01 7.43706703e-01 2.98153907e-01 -3.57567072e-01
9.32479858e-01 3.07702512e-01 -7.08274543e-01 -3.54752451e-01
-8.44616532e-01 -9.74373519e-01 -2.53865272e-01 -5.61217189e-01
2.99516290e-01 -9.65026319e-01 -7.39279985e-01 2.54771203e-01
-7.21138537e-01 -2.84003615e-01 -2.19741911e-01 7.32382715e-01
-3.97137284e-01 5.63478887e-01 -3.92677367e-01 -1.14132559e+00
-7.24459469e-01 -9.48822498e-01 8.04265082e-01 4.77585733e-01
-1.39672384e-01 -5.14862001e-01 6.80883825e-02 7.21656799e-01
3.80462199e-01 -3.43452781e-01 4.08120364e-01 -3.10073376e-01
-8.90184104e-01 -2.49838635e-01 -1.72784582e-01 1.29054815e-01
1.46358669e-01 -3.77558351e-01 -8.51538718e-01 -5.25222480e-01
-1.06835432e-01 -3.43183130e-01 9.72095609e-01 2.88240850e-01
1.15247428e+00 -2.02034757e-01 -1.79510310e-01 1.73434809e-01
1.13902879e+00 3.34054202e-01 6.98205590e-01 2.14444369e-01
4.29459959e-01 4.23972845e-01 1.33636808e+00 3.79634917e-01
2.47291565e-01 9.08124328e-01 1.76890194e-01 -1.25663161e-01
4.11303788e-02 -2.49078676e-01 6.79288983e-01 1.09866631e+00
-2.15298533e-01 -1.63674146e-01 -7.89135396e-01 3.09242427e-01
-2.22148013e+00 -8.65169108e-01 8.92007351e-02 2.78348756e+00
9.64786291e-01 3.01222622e-01 1.38358213e-02 2.88308591e-01
8.29661131e-01 -7.48678893e-02 -8.86634350e-01 1.35656610e-01
4.73637134e-02 -1.35399580e-01 3.80902022e-01 4.83694941e-01
-1.11551225e+00 7.26967990e-01 5.26576328e+00 1.16883087e+00
-8.44612479e-01 2.78623492e-01 6.23979926e-01 -1.05564103e-01
-3.03506106e-01 -3.01900525e-02 -8.13590646e-01 5.19825578e-01
5.41122079e-01 -1.40642554e-01 4.14600104e-01 9.90236163e-01
3.00722182e-01 -1.77841187e-01 -6.18205965e-01 1.26816535e+00
-3.07995398e-02 -1.14365590e+00 -1.00717455e-01 -1.13479786e-01
7.21081078e-01 1.14688046e-01 1.71787396e-01 3.19327593e-01
4.20298010e-01 -6.44873857e-01 6.21437967e-01 5.03230393e-01
8.07635486e-01 -6.18111312e-01 5.85539579e-01 4.48949873e-01
-1.34290040e+00 -4.33045030e-01 -3.55061144e-01 1.45115525e-01
6.10501170e-02 5.12175441e-01 -5.73062003e-01 5.44574082e-01
9.28200424e-01 5.17923176e-01 -5.99924445e-01 1.38176858e+00
-2.40803331e-01 6.61332905e-01 -3.68931144e-01 -2.96317667e-01
-1.14774637e-01 -1.01571545e-01 5.34989774e-01 7.35482633e-01
2.09824033e-02 2.09572852e-01 2.71365166e-01 5.66845059e-01
-3.91177423e-02 4.96564835e-01 -3.77486609e-02 -4.07529660e-02
7.46787667e-01 1.22796535e+00 -6.66289330e-01 -2.05496266e-01
-2.17432678e-01 1.21541870e+00 4.96906221e-01 5.50891936e-01
-7.84994066e-01 -3.37022662e-01 1.38354108e-01 6.77745789e-02
1.39819175e-01 -2.26035133e-01 -3.18349376e-02 -1.21053088e+00
1.55470073e-01 -7.20157564e-01 4.49454576e-01 -5.96444964e-01
-1.25908971e+00 5.40610731e-01 -2.12801009e-01 -1.52462769e+00
-1.14844665e-02 -9.49613601e-02 -4.84279126e-01 6.89350069e-01
-1.36231148e+00 -5.30476332e-01 -6.70931220e-01 7.98565686e-01
7.26881623e-01 -6.92849159e-02 5.95533073e-01 5.76967418e-01
-5.21510661e-01 9.24623370e-01 2.17178702e-01 -8.10018852e-02
8.17743659e-01 -1.04034352e+00 4.73528773e-01 7.38621056e-01
3.02122176e-01 8.36833358e-01 2.74337947e-01 -7.56994069e-01
-1.37241745e+00 -7.19514549e-01 5.92822194e-01 -2.73190528e-01
4.22897875e-01 -4.15726334e-01 -8.08108091e-01 -6.99067935e-02
-3.60475957e-01 -5.63843846e-02 7.01011181e-01 2.52691865e-01
-2.80968398e-01 -1.93975613e-01 -9.39037442e-01 6.56295657e-01
1.39051723e+00 -5.18311203e-01 -2.17182994e-01 3.84948879e-01
7.53967762e-01 -2.20571995e-01 -6.11249447e-01 6.65958703e-01
5.93462646e-01 -7.59834170e-01 1.09574676e+00 -6.43682852e-02
-4.97953705e-02 -4.34217960e-01 -5.78352548e-02 -9.23984826e-01
-5.97559631e-01 -6.70053124e-01 -4.33970183e-01 9.89595115e-01
5.48610628e-01 -5.64246297e-01 1.11207891e+00 7.04864562e-01
1.26495466e-01 -9.22048211e-01 -7.43261337e-01 -6.89336598e-01
-5.27270257e-01 -2.82349914e-01 2.62720346e-01 7.47471690e-01
-4.85186912e-02 3.49558920e-01 -3.58825445e-01 -6.10514916e-02
7.60994256e-01 -1.44584268e-01 6.25521302e-01 -1.33382499e+00
-4.23199832e-01 -1.22483313e-01 -1.08453855e-01 -1.54170966e+00
-2.97415018e-01 -5.40474713e-01 4.07932959e-02 -1.46663141e+00
3.37458491e-01 -8.86912644e-01 -7.34239697e-01 2.97332287e-01
-6.37632906e-01 2.36822098e-01 1.68073535e-01 7.13586628e-01
-1.00068355e+00 7.22020507e-01 1.08968151e+00 -2.80196339e-01
-4.77650851e-01 2.66474366e-01 -5.89614987e-01 4.36890453e-01
8.79483938e-01 -3.30620199e-01 -8.11792910e-01 -2.90262878e-01
5.29186130e-01 2.60277502e-02 8.21214616e-02 -7.48317301e-01
5.88546097e-01 -1.38993233e-01 5.14723003e-01 -8.01052034e-01
4.40803021e-01 -6.39340818e-01 4.71975319e-02 3.93296361e-01
-5.11973679e-01 -9.16812867e-02 -2.85095066e-01 9.61707592e-01
-1.64639220e-01 -1.49682760e-01 3.99483591e-01 -1.97656993e-02
-8.88329268e-01 5.57740450e-01 -4.37629670e-02 -1.90285429e-01
6.08191967e-01 -4.24616456e-01 -3.82700980e-01 -4.64851111e-01
-6.15939140e-01 5.29925346e-01 3.01945508e-01 5.49736500e-01
9.29380894e-01 -1.37650645e+00 -5.59700489e-01 1.08847834e-01
1.67876467e-01 2.91325129e-03 4.47989345e-01 1.02807117e+00
-1.12166986e-01 1.44297212e-01 3.90161008e-01 -5.71892321e-01
-1.47652197e+00 3.51354182e-01 8.48789588e-02 -3.48103881e-01
-3.59195918e-01 1.05577707e+00 2.44829714e-01 -3.73179138e-01
5.83116174e-01 3.73196244e-01 -4.50281084e-01 2.02219114e-02
8.01162899e-01 4.58234012e-01 2.04559371e-01 -2.17474550e-01
-3.62379491e-01 4.14439827e-01 -4.01744246e-01 -1.00204445e-01
9.11929429e-01 -3.91844898e-01 3.30491930e-01 2.89044142e-01
9.53521192e-01 9.85687897e-02 -1.27847886e+00 -5.42783201e-01
2.21084237e-01 -7.00347066e-01 1.66342899e-01 -8.24922919e-01
-9.59810317e-01 3.85582030e-01 9.88642097e-01 -5.20754866e-02
1.13157511e+00 -1.85164511e-01 6.50622487e-01 5.93167365e-01
6.54800713e-01 -1.34027815e+00 5.22818565e-01 2.78028727e-01
8.87764335e-01 -1.36407077e+00 8.82275254e-02 -5.14051616e-01
-6.96365178e-01 6.57422483e-01 7.57424057e-01 6.38089031e-02
5.00352144e-01 -1.54639706e-01 1.33702740e-01 -7.77505785e-02
-7.71002173e-01 -7.62958527e-02 4.80233848e-01 2.36718073e-01
2.85594702e-01 3.18880216e-03 -6.38526618e-01 3.77013892e-01
1.48729697e-01 3.08050178e-02 6.83627836e-03 8.67312610e-01
-5.50193667e-01 -9.84954834e-01 -3.43622059e-01 4.33818877e-01
-1.28718019e-01 -1.99103102e-01 -4.44274962e-01 1.25204653e-01
-2.31868625e-01 1.21248949e+00 -1.34929180e-01 -7.20422268e-01
1.03876382e-01 -3.48470062e-01 3.14225763e-01 -4.42075044e-01
-6.83237493e-01 4.44793142e-02 -8.89335275e-02 -6.71817124e-01
-4.88434434e-01 -4.60664093e-01 -1.21804833e+00 -9.24065635e-02
-1.05504966e+00 5.25723934e-01 5.06842732e-01 5.55449545e-01
3.97573352e-01 -4.27822098e-02 1.02819467e+00 -4.40549195e-01
-6.90572679e-01 -8.20109487e-01 -3.43505800e-01 3.88540834e-01
2.76966859e-03 -8.54252696e-01 -5.80394447e-01 -4.01914448e-01] | [10.058030128479004, 5.086414813995361] |
4e1b1844-2b4a-48b5-94c0-fa4c283ee5c5 | general-purpose-deep-point-cloud-feature | null | null | https://ieeexplore.ieee.org/abstract/document/8354322 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8354322 | General-Purpose Deep Point Cloud Feature Extractor | Depth sensors used in autonomous driving and gaming
systems often report back 3D point clouds. The lack of
structure from these sensors does not allow these systems
to take advantage of recent advances in convolutional neural networks which are dependent upon traditional filtering
and pooling operations. Analogous to image based convolutional architectures, recently introduced graph based architectures afford similar filtering and pooling operations
on arbitrary graphs. We adopt these graph based methods
to 3D point clouds to introduce a generic vector representation of 3D graphs, we call graph 3D (G3D). We believe
we are the first to use large scale transfer learning on 3D
point cloud data and demonstrate the discriminant power
of our salient latent representation of 3D point clouds on
unforeseen test sets. By using our G3D network (G3DNet)
as a feature extractor, and then pairing G3D feature vectors
with a standard classifier, we achieve the best accuracy on
ModelNet10 (93.1%) and ModelNet 40 (91.7%) for a graph
network, and comparable performance on the Sydney Urban Objects dataset to other methods. This general-purpose
feature extractor can be used as an off-the-shelf component
in other 3D scene understanding or object tracking works. | ['Raymond Ptucha', 'Shagan Sah', 'Saloni Jain', 'Atir Petkar', 'Rohan Dhamdhere', 'Miguel Dominguez'] | 2018-03-12 | null | null | null | ieee-winter-conference-on-applications-of-2 | ['3d-object-classification'] | ['computer-vision'] | [-2.59920478e-01 1.75212279e-01 2.07756132e-01 -3.65342766e-01
-3.47624391e-01 -5.32252908e-01 6.54720008e-01 1.21635638e-01
-2.68689960e-01 -2.52548270e-02 -1.55027986e-01 -5.91455042e-01
1.13269776e-01 -1.10815573e+00 -8.93760800e-01 -2.56820112e-01
-4.84457970e-01 3.97226512e-01 6.70907378e-01 -4.28904653e-01
3.36701512e-01 1.15646756e+00 -1.74188161e+00 1.33906946e-01
2.22164974e-01 1.15710592e+00 8.19887221e-02 7.39448071e-01
-2.74220824e-01 2.80318171e-01 -4.19158340e-01 -2.13463336e-01
7.27401316e-01 4.36112046e-01 -2.88563430e-01 2.06767604e-01
1.15620279e+00 -4.14107114e-01 -5.59044659e-01 9.93454814e-01
4.51029152e-01 1.45240113e-01 6.28291667e-01 -1.65789008e+00
-7.57438779e-01 -1.25844866e-01 -2.78611988e-01 3.00124615e-01
3.53667170e-01 3.22062403e-01 9.15125608e-01 -1.01440251e+00
6.14089727e-01 1.51857281e+00 9.36252475e-01 4.16433394e-01
-9.37424779e-01 -9.14131641e-01 2.86042422e-01 -6.62698736e-03
-1.27552485e+00 -1.21706948e-01 8.56544018e-01 -7.18756676e-01
1.83387816e+00 -3.99978273e-02 1.16891110e+00 7.39737034e-01
6.61676705e-01 5.79764307e-01 8.60926151e-01 -6.36515720e-03
1.79251984e-01 -2.07931221e-01 2.74307430e-01 1.13167834e+00
5.19533813e-01 3.34684610e-01 -5.93702614e-01 -2.00770386e-02
8.67639780e-01 2.76516050e-01 8.45271796e-02 -8.95674884e-01
-9.81868804e-01 9.57109690e-01 1.07548189e+00 -2.96665251e-01
-2.50802398e-01 6.31506801e-01 5.02567664e-02 4.56768066e-01
7.16491818e-01 3.56517404e-01 -3.00454795e-01 1.94036156e-01
-5.87161958e-01 5.08266509e-01 5.61773598e-01 1.27453852e+00
1.07141936e+00 2.04970881e-01 3.42907667e-01 2.53017008e-01
4.34482008e-01 5.94947278e-01 1.22172654e-01 -8.31390858e-01
5.03473699e-01 1.04797781e+00 -2.41549537e-01 -9.64369655e-01
-6.76946878e-01 -2.25703344e-01 -4.61129814e-01 9.93606210e-01
1.02017179e-01 2.00724766e-01 -1.43186235e+00 1.17661405e+00
1.43726692e-01 3.59695315e-01 -8.24571699e-02 8.86476457e-01
1.12913513e+00 2.72480339e-01 -2.41631269e-01 8.02589953e-01
1.06997240e+00 -3.01741213e-01 4.99746986e-02 -4.08896089e-01
7.00808704e-01 -4.26343739e-01 8.47969413e-01 4.23673317e-02
-7.94761479e-01 -5.00764370e-01 -1.40079057e+00 -2.70757854e-01
-1.05000734e+00 -4.61982936e-01 1.15551662e+00 6.56805813e-01
-1.40968549e+00 5.41712046e-01 -1.01206863e+00 -7.67296433e-01
8.48769903e-01 7.03122079e-01 -7.56787539e-01 -2.19302341e-01
-6.12897456e-01 1.13150609e+00 2.21262217e-01 -5.46752624e-02
-8.06752205e-01 -6.30355954e-01 -1.11972368e+00 -2.44058341e-01
-2.98907552e-02 -8.33775640e-01 8.36597800e-01 -6.97142929e-02
-1.04972577e+00 1.30840898e+00 1.61142930e-01 -5.65768301e-01
2.23498777e-01 -1.96192414e-01 -1.01416230e-01 1.21998535e-02
2.06351444e-01 9.23823357e-01 8.99564981e-01 -1.01254213e+00
-5.78446388e-01 -6.05765879e-01 1.43813074e-01 2.83758491e-01
1.15615338e-01 -4.74161029e-01 -2.81035036e-01 -2.22779736e-02
7.97529757e-01 -1.14404154e+00 -2.11989954e-01 3.25245559e-01
-3.00401151e-01 -3.72116774e-01 1.10841990e+00 -1.25595525e-01
1.69697210e-01 -2.10402560e+00 -6.64542243e-02 2.99952120e-01
7.60111094e-01 1.18199021e-01 -2.02899769e-01 2.31274351e-01
-6.72496781e-02 1.26588508e-01 -3.86506207e-02 -2.57825136e-01
1.50440112e-01 3.70950013e-01 -1.62888780e-01 7.16029286e-01
8.00363123e-01 1.36160350e+00 -8.73595476e-01 -9.04643685e-02
6.90481544e-01 4.77907211e-01 -7.96166301e-01 -1.01848487e-02
-2.22551394e-02 -1.10831283e-01 -5.20171523e-01 8.85329783e-01
7.39859521e-01 -1.25411108e-01 -5.73160291e-01 -2.21411586e-02
5.59785552e-02 3.76507819e-01 -8.83499265e-01 1.81911242e+00
-1.24062128e-01 9.50893760e-01 -1.42904460e-01 -9.17054951e-01
1.29493344e+00 -1.03852279e-01 4.90614086e-01 -5.54016113e-01
1.82450667e-01 2.30218187e-01 -9.86178741e-02 -2.29029328e-01
6.29217267e-01 -4.26421985e-02 -2.48487338e-01 -6.15976192e-03
4.40628648e-01 -9.86473560e-01 -4.53295380e-01 2.96259046e-01
1.41459668e+00 3.80217433e-02 -5.89058548e-02 -3.21971893e-01
-6.83928058e-02 3.01321834e-01 -6.85233325e-02 7.96144903e-01
-2.75643229e-01 7.86552489e-01 2.58998126e-01 -7.28268623e-01
-1.07167649e+00 -1.21562624e+00 3.98464277e-02 6.28668189e-01
2.11575374e-01 -4.40684110e-01 -7.81491250e-02 -6.58680558e-01
8.45256448e-01 3.75012696e-01 -5.52061498e-01 -3.93332273e-01
-3.31436574e-01 -2.54888922e-01 4.25118089e-01 5.99195063e-01
3.71326476e-01 -6.84198797e-01 -5.72730124e-01 1.47224694e-01
6.67984605e-01 -1.25764203e+00 1.14189096e-01 6.64149106e-01
-1.07545650e+00 -1.10243022e+00 -2.72324204e-01 -6.31494761e-01
5.25044262e-01 7.73865640e-01 1.30135500e+00 1.44425601e-01
-1.41970769e-01 6.46485865e-01 -2.24167690e-01 -1.00392044e+00
7.07857916e-03 1.03705101e-01 8.12643617e-02 -4.85193193e-01
1.08068931e+00 -7.01443374e-01 -2.69983768e-01 7.37826526e-02
-4.94571298e-01 -2.43110093e-03 3.76287550e-01 4.02884334e-01
5.39503276e-01 -2.12007985e-01 1.07135419e-02 -5.58642566e-01
5.78956306e-01 -4.14317578e-01 -8.95719707e-01 -5.07625341e-01
-3.43059331e-01 -2.98440792e-02 2.52478779e-03 -1.05464987e-01
-6.76386282e-02 2.49747381e-01 -2.14585483e-01 -1.13718402e+00
-1.62902981e-01 3.45807791e-01 2.06932798e-02 -6.66794419e-01
7.69730568e-01 -1.25669509e-01 2.22629622e-01 -2.60318637e-01
4.53871459e-01 4.14623708e-01 3.52652997e-01 -1.04767255e-01
1.14661217e+00 6.81719542e-01 5.04046023e-01 -9.96073365e-01
-6.06459439e-01 -5.76969326e-01 -9.63868141e-01 -2.71955907e-01
1.16884601e+00 -1.27778196e+00 -7.54222929e-01 4.18899655e-01
-1.02044833e+00 -3.48355383e-01 -3.14658821e-01 5.23108959e-01
-6.46964908e-01 -9.72658694e-02 -1.96853429e-01 -6.65511370e-01
1.11508928e-01 -9.50351179e-01 1.59272659e+00 -4.42587957e-02
-2.11500749e-03 -7.99242616e-01 -2.84565628e-01 5.86002506e-02
2.95108885e-01 6.68552220e-01 7.30896413e-01 -6.12032473e-01
-9.72127199e-01 -7.51395404e-01 -4.31942880e-01 2.51296580e-01
-2.06446536e-02 -1.72003508e-01 -1.18938279e+00 -2.25838006e-01
-3.15362543e-01 -2.09315985e-01 1.11395383e+00 3.47454160e-01
8.66307497e-01 2.89299697e-01 -3.60274106e-01 1.01661265e+00
1.36948001e+00 -1.44932225e-01 4.73225385e-01 3.74349624e-01
1.17872632e+00 3.07090223e-01 1.76219836e-01 2.40720436e-02
5.77212572e-01 3.65721554e-01 1.05220330e+00 -1.00694977e-01
-3.08927417e-01 -4.85161901e-01 1.67478710e-01 5.50864220e-01
-9.10086036e-02 -6.41250461e-02 -1.28912258e+00 5.46161354e-01
-1.73761642e+00 -6.14963114e-01 -1.57560587e-01 1.93035197e+00
-1.37244761e-01 7.71106601e-01 -1.26679733e-01 -2.80864723e-02
3.81685853e-01 2.34347492e-01 -5.99677026e-01 -4.12634581e-01
-1.00354441e-02 5.08643508e-01 1.08868778e+00 3.35108608e-01
-1.23009872e+00 1.10000253e+00 6.45094347e+00 1.89891413e-01
-1.20273757e+00 -8.26225132e-02 -2.11249059e-03 -2.48600766e-01
-1.95653021e-01 -7.10859448e-02 -9.16407764e-01 -3.30051109e-02
9.10659134e-01 2.60974299e-02 2.49339625e-01 1.07049167e+00
-2.29043514e-01 2.76515752e-01 -1.26930976e+00 1.23266375e+00
6.31717816e-02 -1.58523607e+00 1.64883509e-01 4.12768096e-01
4.12429810e-01 9.15635347e-01 -1.22453272e-02 4.30190623e-01
5.82997203e-01 -1.25824785e+00 7.23201334e-01 3.29423755e-01
6.34067178e-01 -5.75239360e-01 5.03825903e-01 1.78890184e-01
-1.27881241e+00 9.56580788e-02 -8.58353734e-01 -4.24186885e-01
-9.14230347e-02 3.33507568e-01 -1.12181270e+00 2.86659688e-01
9.93673027e-01 1.17535448e+00 -6.97596133e-01 1.10250998e+00
-3.45978923e-02 1.99097872e-01 -7.57705927e-01 -2.53347754e-01
6.37982309e-01 6.01663291e-02 8.18432271e-01 9.66147184e-01
3.72676581e-01 4.85210866e-02 1.53020501e-01 9.11250353e-01
-2.32234016e-01 -3.15032303e-01 -1.57641137e+00 2.53322418e-03
2.09605485e-01 1.06801140e+00 -7.89147198e-01 -2.72094846e-01
-6.65110171e-01 6.14341557e-01 4.10745233e-01 2.96634823e-01
-3.72941375e-01 -4.14071411e-01 1.24604571e+00 2.17929557e-01
6.80145442e-01 -1.04840660e+00 -3.15534741e-01 -1.08978486e+00
-1.44249380e-01 -3.56560946e-01 -2.18114466e-03 -1.11475873e+00
-1.39249003e+00 5.52360237e-01 6.57911822e-02 -1.41049409e+00
-6.18149862e-02 -1.22593868e+00 -5.05893588e-01 1.03595722e+00
-1.75281274e+00 -1.24985409e+00 -7.57674754e-01 8.87536466e-01
1.11104958e-01 -4.05231655e-01 6.65178120e-01 -4.64423671e-02
9.26505625e-02 1.74246550e-01 -4.26402599e-01 2.08940461e-01
1.67907238e-01 -1.28128588e+00 1.17794752e+00 7.45677471e-01
4.04289573e-01 3.30912292e-01 3.24038297e-01 -7.24084616e-01
-1.94188333e+00 -1.40232563e+00 4.95746046e-01 -1.08801925e+00
6.37202144e-01 -9.19377565e-01 -8.58593404e-01 9.67574835e-01
-1.82817698e-01 3.96558851e-01 2.99750358e-01 1.13604642e-01
-4.68740016e-01 1.14963472e-01 -1.07671309e+00 4.56460029e-01
1.60769022e+00 -7.48298645e-01 -6.60110474e-01 4.01746005e-01
9.54737365e-01 -7.51028299e-01 -7.40310788e-01 3.75528902e-01
3.57991606e-01 -8.80140543e-01 1.18324757e+00 -8.03545415e-01
1.36336818e-01 -2.62912035e-01 -5.95282435e-01 -1.16752827e+00
-5.16645074e-01 -2.85563767e-01 -7.02034831e-02 4.08882946e-01
2.54009634e-01 -9.19113040e-01 1.27707624e+00 5.76780379e-01
-6.30188465e-01 -4.27097768e-01 -1.18044865e+00 -9.28056657e-01
1.33967713e-01 -1.02870786e+00 7.99443543e-01 7.29426801e-01
-5.19874692e-01 2.04140499e-01 2.40023956e-01 5.17584443e-01
7.43484855e-01 1.82045743e-01 1.21307528e+00 -1.55339897e+00
3.49692881e-01 -5.91219604e-01 -1.64875329e+00 -1.00231826e+00
1.79134890e-01 -1.37586367e+00 -1.08562976e-01 -1.77262473e+00
-5.85051060e-01 -5.66078126e-01 -3.25808167e-01 6.29233181e-01
1.90076262e-01 6.36994481e-01 2.80450791e-01 9.53591838e-02
-2.37457722e-01 5.86784065e-01 1.18033612e+00 -3.35754931e-01
-2.07042783e-01 -1.30015537e-01 -5.45912981e-01 6.60162330e-01
6.57514274e-01 -3.24351996e-01 -3.61737758e-01 -6.09927356e-01
2.00798333e-01 -5.12556493e-01 8.48320425e-01 -1.25447905e+00
1.97716951e-01 7.78393671e-02 5.48005044e-01 -9.17566955e-01
6.99023902e-01 -9.94769514e-01 4.87207808e-03 2.97745585e-01
3.54474485e-01 1.87780187e-01 6.57732129e-01 5.17283142e-01
-1.00238822e-01 3.51711154e-01 3.64049673e-01 -4.08513248e-01
-1.32703197e+00 7.67579019e-01 -2.33584121e-01 -2.86404669e-01
9.22021091e-01 -8.47620666e-01 -2.99539059e-01 -2.81014711e-01
-5.60325205e-01 1.16529353e-01 6.41803265e-01 8.57949793e-01
1.01874185e+00 -1.49737751e+00 -6.49109960e-01 6.91324353e-01
4.57603782e-01 3.07594299e-01 -2.80607849e-01 3.30007374e-01
-8.25625837e-01 4.22927141e-01 -6.31097376e-01 -1.23482335e+00
-1.00149059e+00 3.88299674e-01 4.64760125e-01 3.48468989e-01
-1.11235428e+00 1.15107679e+00 1.00736126e-01 -6.18893206e-01
1.61643606e-02 -8.56154680e-01 1.16729490e-01 -7.94248357e-02
-7.09611475e-02 -2.23567575e-01 4.83369201e-01 -6.16247058e-01
-7.22790301e-01 7.27594614e-01 2.10394561e-01 1.96327329e-01
1.57897663e+00 2.45387733e-01 2.56644130e-01 5.42966962e-01
1.36291456e+00 -4.34314638e-01 -1.37416458e+00 -5.13199754e-02
-2.04116613e-01 -6.55857146e-01 4.00764793e-01 -2.31274236e-02
-1.13097215e+00 1.10545945e+00 7.72453249e-01 4.00488228e-01
5.91678739e-01 2.09758386e-01 4.37774450e-01 5.74621975e-01
6.37712300e-01 -3.96637321e-01 -7.73525983e-02 8.41089964e-01
8.68845582e-01 -1.38417804e+00 8.38017985e-02 -3.58639032e-01
-1.87190428e-01 9.62071061e-01 5.15291333e-01 -9.31946635e-01
9.94373560e-01 1.84546530e-01 1.07789129e-01 -9.65660691e-01
-6.99078977e-01 -6.34543180e-01 5.20181656e-01 1.02411938e+00
3.71087273e-03 1.62792817e-01 6.23899519e-01 -6.38057068e-02
-6.70920193e-01 -1.72843754e-01 4.39937115e-01 1.08713305e+00
-6.49251461e-01 -6.64699316e-01 -1.22747034e-01 7.70408213e-01
6.59161583e-02 -9.78604853e-02 -4.80331451e-01 1.24123728e+00
1.04294144e-01 6.10764384e-01 4.67112601e-01 -9.06040251e-01
7.94998646e-01 2.11200252e-01 6.05031967e-01 -9.43072319e-01
-4.27216470e-01 -4.12807465e-01 -4.43479829e-02 -9.47310925e-01
-3.59679639e-01 -4.79118049e-01 -1.03101134e+00 -4.69462216e-01
-2.76352227e-01 -4.86502081e-01 1.13494945e+00 6.40616953e-01
6.11419857e-01 5.31120121e-01 1.49960294e-01 -1.56205201e+00
1.52269164e-02 -6.74710035e-01 -6.91591740e-01 4.10962611e-01
5.71982920e-01 -1.09903646e+00 -2.66481519e-01 -3.74029607e-01] | [7.943962097167969, -3.6384639739990234] |
413b14d7-4c95-4045-bfe5-7cb6e7542862 | benchmarking-adversarially-robust-quantum | 2211.12681 | null | https://arxiv.org/abs/2211.12681v1 | https://arxiv.org/pdf/2211.12681v1.pdf | Benchmarking Adversarially Robust Quantum Machine Learning at Scale | Machine learning (ML) methods such as artificial neural networks are rapidly becoming ubiquitous in modern science, technology and industry. Despite their accuracy and sophistication, neural networks can be easily fooled by carefully designed malicious inputs known as adversarial attacks. While such vulnerabilities remain a serious challenge for classical neural networks, the extent of their existence is not fully understood in the quantum ML setting. In this work, we benchmark the robustness of quantum ML networks, such as quantum variational classifiers (QVC), at scale by performing rigorous training for both simple and complex image datasets and through a variety of high-end adversarial attacks. Our results show that QVCs offer a notably enhanced robustness against classical adversarial attacks by learning features which are not detected by the classical neural networks, indicating a possible quantum advantage for ML tasks. Contrarily, and remarkably, the converse is not true, with attacks on quantum networks also capable of deceiving classical neural networks. By combining quantum and classical network outcomes, we propose a novel adversarial attack detection technology. Traditionally quantum advantage in ML systems has been sought through increased accuracy or algorithmic speed-up, but our work has revealed the potential for a new kind of quantum advantage through superior robustness of ML models, whose practical realisation will address serious security concerns and reliability issues of ML algorithms employed in a myriad of applications including autonomous vehicles, cybersecurity, and surveillance robotic systems. | ['Muhammad Usman', 'Lloyd C. L. Hollenberg', 'Martin Sevior', 'Christopher Leckie', 'Sarah M. Erfani', 'Maxwell T. West'] | 2022-11-23 | null | null | null | null | ['adversarial-attack-detection', 'adversarial-attack-detection'] | ['computer-vision', 'knowledge-base'] | [ 5.51924646e-01 5.44124186e-01 1.16997935e-01 4.28100303e-02
-8.19804728e-01 -1.01556349e+00 8.28988492e-01 -2.20820114e-01
-5.04082859e-01 7.75688171e-01 -6.00182950e-01 -5.64665794e-01
-3.93965133e-02 -1.04183769e+00 -1.01794469e+00 -1.25598311e+00
-7.48800561e-02 7.95938373e-02 6.56186566e-02 -5.49096227e-01
1.40915617e-01 7.29400575e-01 -1.17779005e+00 -7.63949156e-02
7.98636258e-01 8.81485820e-01 -7.88939118e-01 8.64248991e-01
4.76552516e-01 6.40053809e-01 -5.40660858e-01 -9.82019186e-01
5.03529668e-01 -6.08663738e-01 -5.35016477e-01 -6.49136484e-01
4.57500815e-01 -1.68008670e-01 -1.04651761e+00 1.79663002e+00
4.05493677e-01 -1.19473517e-01 5.96203089e-01 -1.37651253e+00
-7.29032993e-01 6.29893959e-01 3.01439404e-01 4.96086702e-02
-7.03059137e-02 7.58556783e-01 1.17336249e+00 -1.66151181e-01
5.00475883e-01 1.24782538e+00 6.36800706e-01 1.03727221e+00
-1.43503189e+00 -8.53360355e-01 -9.29838240e-01 3.56112480e-01
-1.24304545e+00 -5.52425504e-01 6.70540750e-01 -1.70211390e-01
7.27026045e-01 1.19437471e-01 1.85197935e-01 1.38425601e+00
7.12414980e-01 2.87576616e-01 1.14871216e+00 -2.80460626e-01
3.50526303e-01 2.38940552e-01 -1.73404709e-01 8.89616966e-01
3.28849703e-01 9.01480734e-01 -3.57260704e-01 -3.49897414e-01
4.54800785e-01 -3.50627959e-01 -3.33599240e-01 -2.09764123e-01
-1.18755150e+00 1.17385590e+00 8.12482536e-01 3.85135561e-02
-9.73879024e-02 6.35268033e-01 5.79666495e-01 6.22348011e-01
-7.42593482e-02 6.97413385e-01 -4.65479761e-01 1.66049063e-01
-2.12279171e-01 2.10057661e-01 8.07590783e-01 5.68219364e-01
7.96354711e-01 2.68793285e-01 1.38459250e-01 -1.05971761e-01
9.31636393e-02 9.87259507e-01 8.82690623e-02 -1.21017146e+00
1.61488950e-01 6.37561828e-02 -1.81535915e-01 -8.38914573e-01
-3.53092641e-01 -3.77589047e-01 -9.93945658e-01 7.23312318e-01
5.46544254e-01 -2.71948874e-01 -4.19031203e-01 2.04150867e+00
3.04279447e-01 8.96801278e-02 8.02938342e-01 7.54133344e-01
5.98858833e-01 5.54338872e-01 -2.58684248e-01 -1.38744697e-01
1.09366524e+00 -2.41866365e-01 -4.65712249e-01 5.16957864e-02
7.95706034e-01 -3.64830583e-01 4.70421195e-01 4.53002542e-01
-7.96973109e-01 -2.03841209e-01 -1.41628718e+00 1.34562388e-01
-4.64861333e-01 -7.68258214e-01 7.69313514e-01 1.37656975e+00
-7.59625554e-01 1.14087451e+00 -8.43511343e-01 -1.15272656e-01
6.12618387e-01 6.77606165e-01 -5.35179973e-01 9.78356227e-02
-1.69011045e+00 1.00163865e+00 4.71079648e-01 1.20354980e-01
-9.03992832e-01 -3.70431542e-01 -6.56274021e-01 -1.63439319e-01
1.50663391e-01 -6.42520607e-01 9.89789665e-01 -9.81290877e-01
-1.74918127e+00 7.79855013e-01 2.48525262e-01 -9.33274567e-01
4.25036430e-01 2.59497315e-01 -6.35216355e-01 3.96494895e-01
-3.69141012e-01 4.57147121e-01 1.09938145e+00 -7.03496575e-01
-6.44520447e-02 -5.45794427e-01 4.75700617e-01 -2.79063821e-01
-1.68224871e-01 -1.64919630e-01 6.29591882e-01 -3.99471186e-02
4.70933318e-02 -1.32898784e+00 -1.90287471e-01 3.53334621e-02
-4.65932041e-01 -6.56177029e-02 8.88185322e-01 -3.46751735e-02
3.94997776e-01 -2.11144209e+00 2.26681769e-01 1.38112694e-01
3.70037705e-01 6.04119003e-01 -6.57314435e-02 5.91840029e-01
6.86732680e-02 3.75689894e-01 -3.75112742e-01 2.18672633e-01
3.33901763e-01 3.35077465e-01 -4.61789131e-01 1.04802167e+00
5.33151805e-01 1.31354475e+00 -8.65742564e-01 -1.73478633e-01
1.92212805e-01 5.60208499e-01 -6.92094743e-01 -2.96828032e-01
-2.50409991e-01 7.02853262e-01 -4.86854136e-01 4.74819750e-01
5.20796895e-01 -7.20004290e-02 -3.64499986e-02 -1.00280941e-01
3.05342764e-01 6.20646924e-02 -6.61826432e-01 1.15606320e+00
3.10773980e-02 9.27899778e-01 1.50153503e-01 -1.20914328e+00
3.50395560e-01 4.15100843e-01 -7.59119764e-02 -7.84610391e-01
5.04782915e-01 3.84546876e-01 5.14055312e-01 -4.43139374e-01
3.04221898e-01 -6.49679780e-01 -3.30836982e-01 2.30907857e-01
2.84495175e-01 -3.78963381e-01 -3.57407868e-01 2.64499038e-01
1.22240973e+00 -1.70189098e-01 8.73556174e-03 -5.25659546e-02
5.18057168e-01 -3.29690687e-02 4.16747570e-01 1.17901826e+00
-8.26214015e-01 1.13961153e-01 7.29904473e-01 -9.55857560e-02
-1.20787251e+00 -1.31162477e+00 -6.54070795e-01 6.41572356e-01
2.23966464e-01 -1.21637471e-01 -8.28826427e-01 -6.52227044e-01
-2.57525258e-02 4.15487140e-01 -5.74441552e-01 -8.82378340e-01
-3.47320080e-01 -9.25158203e-01 1.46390057e+00 -1.49007514e-01
5.75292230e-01 -9.24027205e-01 -3.28723252e-01 -8.95368457e-02
4.40725803e-01 -1.48669231e+00 2.68709153e-01 2.82935083e-01
-7.35153377e-01 -1.29553914e+00 -1.35511123e-02 -1.64492384e-01
2.90377975e-01 -2.64007628e-01 6.58227265e-01 1.33032007e-02
-2.03298405e-01 2.36563370e-01 -1.83844045e-01 -3.59014690e-01
-1.41989207e+00 -1.84907570e-01 6.80114985e-01 -6.20072447e-02
1.23938680e-01 -4.95811820e-01 -4.56790030e-01 -2.03555655e-02
-1.17913795e+00 -5.35689533e-01 7.03974426e-01 9.96235967e-01
-3.80123332e-02 2.17662171e-01 6.23253644e-01 -9.23934042e-01
2.99381852e-01 -3.55931878e-01 -9.11812186e-01 4.55935299e-02
-4.48384464e-01 3.84923667e-01 1.24941897e+00 -2.71830201e-01
-5.56797624e-01 -3.23825270e-01 -4.76810724e-01 -2.41968274e-01
-1.93241611e-01 3.18872392e-01 -1.77338738e-02 -9.89770889e-01
1.16955650e+00 4.34268452e-02 3.47747922e-01 3.91548336e-01
6.42028391e-01 4.83919442e-01 5.58267117e-01 -3.93179864e-01
1.45035470e+00 5.55505514e-01 8.61325324e-01 -1.15969062e+00
-7.36233473e-01 2.99493581e-01 -4.43080574e-01 -1.31888734e-02
1.01237714e+00 -5.49928486e-01 -1.20732152e+00 5.53254306e-01
-1.11317766e+00 2.70976610e-02 -6.44307435e-02 4.79551107e-01
-4.01423901e-01 5.44078410e-01 -8.49112809e-01 -9.76000905e-01
-1.74420670e-01 -1.34296167e+00 5.98798037e-01 2.19125688e-01
4.46905375e-01 -1.10182571e+00 -1.72919959e-01 3.60313833e-01
3.23753387e-01 6.10081613e-01 9.86629426e-01 -7.38582492e-01
-9.27233934e-01 -6.26976430e-01 -6.74818158e-02 6.84071779e-01
-4.09811497e-01 1.98903486e-01 -1.45601654e+00 -4.13454920e-01
-3.22506800e-02 -7.47829914e-01 8.79308462e-01 -6.18516281e-02
6.05880022e-01 -4.29843575e-01 5.97834364e-02 6.94751382e-01
1.39116335e+00 4.37259749e-02 6.14548624e-01 -1.07317775e-01
6.27749681e-01 3.84874761e-01 -9.37491581e-02 -1.31805584e-01
-1.47819117e-01 4.74648505e-01 9.63302016e-01 3.48865360e-01
3.94176364e-01 8.61600712e-02 6.62975490e-01 6.31145120e-01
7.00751022e-02 -1.67436653e-03 -5.86277366e-01 -2.37022832e-01
-1.44068158e+00 -1.24804604e+00 -3.43088150e-01 2.38994074e+00
6.31571174e-01 4.79376644e-01 -2.46987224e-01 3.28697830e-01
4.80010271e-01 5.56617603e-02 -8.05655837e-01 -7.38315225e-01
-6.12501979e-01 5.20608366e-01 7.77354300e-01 3.33109409e-01
-1.21003962e+00 9.21163261e-01 6.14549303e+00 8.32308471e-01
-1.26002312e+00 3.81402522e-01 3.97186697e-01 1.46613911e-01
-2.04874605e-01 1.57055184e-01 -4.38753664e-01 1.68892726e-01
1.36464775e+00 -5.90259172e-02 7.39433467e-01 4.25449640e-01
-2.67559439e-01 1.85690328e-01 -1.18684530e+00 8.47025096e-01
-7.66727328e-02 -1.58926868e+00 -1.35374382e-01 3.14084798e-01
7.22000837e-01 5.28636932e-01 3.79601359e-01 4.77934510e-01
3.67576331e-02 -1.22877681e+00 5.55255771e-01 3.07663172e-01
7.51844525e-01 -9.90580976e-01 7.01296568e-01 4.86904293e-01
-3.17728102e-01 -1.67420089e-01 -4.99867946e-01 -2.29606956e-01
-3.13820392e-01 2.40762666e-01 -2.62936205e-01 5.38567007e-01
2.49720320e-01 2.92547047e-01 -4.97649282e-01 3.14212948e-01
-2.19894618e-01 6.29518092e-01 -2.90598899e-01 -3.00565600e-01
5.86875379e-01 -3.84700119e-01 8.31177771e-01 6.08521342e-01
-1.13333486e-01 1.59777477e-01 -4.55621094e-01 1.30068910e+00
-3.40972304e-01 -4.68073010e-01 -1.15184808e+00 -5.74818015e-01
2.24424616e-01 1.27173400e+00 -4.29590136e-01 -1.06957145e-01
-1.82722673e-01 1.01673996e+00 1.44347221e-01 3.20084631e-01
-9.08960223e-01 -6.10175610e-01 7.94455945e-01 -6.57792449e-01
-8.79130512e-02 -2.55119741e-01 -1.90089345e-01 -1.38912547e+00
-2.02769220e-01 -8.67406428e-01 -1.13117717e-01 -2.39456281e-01
-1.30415642e+00 2.85610497e-01 -5.86769104e-01 -9.87343967e-01
-1.07198648e-01 -1.15974450e+00 -5.60489118e-01 5.52427232e-01
-1.18236423e+00 -8.52027595e-01 3.57011795e-01 5.62208116e-01
-4.11084950e-01 -3.39241713e-01 1.06798697e+00 1.31914187e-02
-7.36874998e-01 7.84740627e-01 6.97654605e-01 3.68178338e-01
3.73333067e-01 -1.10672760e+00 3.80000234e-01 9.91524935e-01
4.39573079e-01 6.38131082e-01 1.07585251e+00 -1.24595627e-01
-2.05241942e+00 -9.18768466e-01 3.64594787e-01 -8.71586859e-01
1.09026039e+00 -4.74272549e-01 -6.53668642e-01 4.43455875e-01
6.03675842e-02 2.72117794e-01 4.23457712e-01 -9.57489386e-02
-7.42373705e-01 1.64609790e-01 -1.42764616e+00 6.37088060e-01
5.93307197e-01 -1.31775486e+00 -4.66716588e-01 6.89956367e-01
9.48709011e-01 -7.29422718e-02 -8.14679623e-01 3.36486429e-01
6.58965230e-01 -1.13662255e+00 1.06287527e+00 -7.11587310e-01
4.91853416e-01 1.48528405e-02 -3.62242907e-01 -1.09770107e+00
2.69084930e-01 -9.73084152e-01 -7.36842081e-02 6.20287895e-01
4.09842938e-01 -1.29676008e+00 7.41283715e-01 4.76030380e-01
-6.29679039e-02 -3.46488982e-01 -1.48856485e+00 -9.01240289e-01
8.98539960e-01 -7.22737789e-01 8.06788802e-02 1.05182540e+00
2.20644213e-02 1.62521183e-01 -1.75077051e-01 5.51325142e-01
1.17715549e+00 -4.33364868e-01 6.85182393e-01 -1.07725215e+00
-3.52336019e-01 -7.70544291e-01 -1.05440831e+00 -3.89695734e-01
4.65547502e-01 -1.18261552e+00 -1.48816124e-01 -3.90297443e-01
-1.54372096e-01 -1.73553526e-01 -2.68941075e-01 -1.07214086e-01
1.01792164e-01 6.66097105e-01 1.26999423e-01 1.40511632e-01
-4.28399056e-01 5.12945533e-01 1.00596607e+00 -2.41558999e-01
2.81556040e-01 1.76218785e-02 -2.80105472e-01 6.72887921e-01
7.14591324e-01 -6.70655787e-01 -5.17781526e-02 -1.86451152e-02
6.61906958e-01 2.30694994e-01 1.00312924e+00 -1.18980122e+00
2.71937817e-01 9.03612003e-02 2.34928634e-02 1.88697591e-01
4.85661656e-01 -5.10611773e-01 -1.40704334e-01 8.62301648e-01
-1.97057933e-01 -4.86771345e-01 -9.70887989e-02 7.78311670e-01
2.71747680e-03 -2.97782212e-01 1.21890724e+00 2.58344803e-02
-2.99887121e-01 4.09992784e-01 -3.11374128e-01 6.48770928e-02
1.03863370e+00 -5.29580340e-02 -7.22580552e-01 -2.08021790e-01
-6.27158105e-01 -1.70422956e-01 5.84470928e-01 3.12200356e-02
4.19830799e-01 -1.06701589e+00 -4.89900738e-01 3.50010574e-01
1.78941354e-01 -4.33321506e-01 2.62966454e-01 9.24293220e-01
-6.47614837e-01 5.39254427e-01 -1.16573647e-01 -5.94417870e-01
-7.45986760e-01 6.35367513e-01 7.73392022e-01 2.64437675e-01
-4.24528867e-01 8.31218719e-01 4.54581603e-02 -6.58749223e-01
2.79258093e-04 1.65436864e-02 3.33650351e-01 -3.46494466e-01
2.55020559e-01 1.34708226e-01 2.76580658e-02 -9.21675622e-01
-2.06722751e-01 2.76449144e-01 2.41801694e-01 -1.16814673e-01
7.96978712e-01 2.70621896e-01 -1.74503163e-01 3.89689386e-01
1.42212081e+00 -2.46012151e-01 -1.04402578e+00 -1.57086477e-01
-5.41736819e-02 1.48513556e-01 1.46290302e-01 -3.81373644e-01
-7.97680438e-01 1.19058001e+00 6.40957236e-01 7.92210639e-01
5.96086144e-01 1.15092516e-01 9.04404104e-01 1.09735394e+00
6.69193208e-01 -7.50694275e-01 -1.22348756e-01 5.35624206e-01
2.73752540e-01 -1.48475683e+00 -4.18659508e-01 8.48229378e-02
-6.33364990e-02 1.09217608e+00 -9.82211716e-03 -5.17274618e-01
4.99157250e-01 -2.06799001e-01 -4.54726070e-03 -1.15377598e-01
-5.93225479e-01 -1.33778639e-02 7.46756792e-02 5.21787047e-01
-9.25259069e-02 2.39285067e-01 5.88874407e-02 1.92190304e-01
-3.06617588e-01 -5.81549883e-01 9.51378942e-01 7.80711055e-01
-3.59471798e-01 -1.08698630e+00 -3.91390502e-01 2.09999949e-01
-9.48403716e-01 -1.67287126e-01 -3.94730508e-01 7.34288454e-01
1.38199806e-01 1.05979204e+00 -3.71675313e-01 -5.71847856e-01
-6.84197471e-02 1.10383645e-01 6.64847970e-01 -1.31770447e-01
-5.49713552e-01 -6.52987123e-01 -1.33263871e-01 -6.39870882e-01
-4.09613669e-01 -6.13808155e-01 -1.23512340e+00 -7.91336834e-01
-6.25984013e-01 4.64147702e-02 8.68090510e-01 1.33403313e+00
1.01745889e-01 3.87833625e-01 7.50288010e-01 -8.17721069e-01
-1.33496356e+00 -5.72752416e-01 -5.76905608e-01 2.56301284e-01
8.38191509e-01 -5.01822054e-01 -8.97254765e-01 -4.50669497e-01] | [5.56846284866333, 5.078665256500244] |
fe65f992-186c-4bd9-a590-fbb80e51c670 | the-hardness-of-reasoning-about-probabilities | 2305.09508 | null | https://arxiv.org/abs/2305.09508v1 | https://arxiv.org/pdf/2305.09508v1.pdf | The Hardness of Reasoning about Probabilities and Causality | We study formal languages which are capable of fully expressing quantitative probabilistic reasoning and do-calculus reasoning for causal effects, from a computational complexity perspective. We focus on satisfiability problems whose instance formulas allow expressing many tasks in probabilistic and causal inference. The main contribution of this work is establishing the exact computational complexity of these satisfiability problems. We introduce a new natural complexity class, named succ$\exists$R, which can be viewed as a succinct variant of the well-studied class $\exists$R, and show that the problems we consider are complete for succ$\exists$R. Our results imply even stronger algorithmic limitations than were proven by Fagin, Halpern, and Megiddo (1990) and Moss\'{e}, Ibeling, and Icard (2022) for some variants of the standard languages used commonly in probabilistic and causal inference. | ['Maciej Liśkiewicz', 'Markus Bläser', 'Benito van der Zander'] | 2023-05-16 | null | null | null | null | ['causal-inference', 'causal-inference'] | ['knowledge-base', 'miscellaneous'] | [ 1.41636819e-01 7.97868967e-01 -2.28340104e-01 -3.73047769e-01
-7.36826062e-01 -6.17006123e-01 8.02779615e-01 1.51292562e-01
-2.20863596e-01 1.14300823e+00 3.06317329e-01 -6.06540322e-01
-8.80433917e-01 -1.04823124e+00 -6.88707888e-01 -5.05167723e-01
-4.66945410e-01 6.12610221e-01 3.59067440e-01 -2.45867193e-01
2.43766502e-01 4.99924123e-01 -1.64564383e+00 1.99916400e-02
6.27333522e-01 5.81883490e-01 -1.43831581e-01 6.95495248e-01
-9.31617767e-02 1.14369106e+00 -3.52855235e-01 -4.28068459e-01
-3.36087316e-01 -4.93312985e-01 -1.15553033e+00 -4.47741896e-01
1.34398609e-01 -1.03418110e-02 -3.04273754e-01 1.30097854e+00
5.74441552e-02 1.20812673e-02 8.32627118e-01 -1.69123769e+00
-4.03193653e-01 1.38885820e+00 -4.53153819e-01 2.20826492e-01
8.78989935e-01 -1.58506036e-01 1.26345599e+00 -2.63690293e-01
4.78069842e-01 1.72396052e+00 5.86635828e-01 5.34183621e-01
-1.34104347e+00 -6.78987145e-01 1.59907907e-01 3.07993382e-01
-1.51016486e+00 -2.13607326e-01 4.97606277e-01 -5.07481635e-01
6.78829312e-01 6.28882825e-01 2.09457099e-01 7.41583467e-01
2.84490902e-02 6.75461233e-01 1.39495206e+00 -7.87318707e-01
3.01632494e-01 -2.40564972e-01 6.16229117e-01 8.16074312e-01
5.80716610e-01 4.20941025e-01 -4.59002018e-01 -4.81845379e-01
5.08974016e-01 -4.53358412e-01 -1.06879003e-01 5.94016239e-02
-1.02042150e+00 9.58279908e-01 -1.40588596e-01 4.62078422e-01
-4.05583978e-02 8.09582055e-01 -1.24849342e-02 6.63105026e-03
-6.66417880e-03 2.15915665e-01 -3.55933487e-01 -5.55791259e-02
-4.81735200e-01 8.48474741e-01 1.09222734e+00 1.09246325e+00
3.59157681e-01 -1.02466352e-01 -5.37104830e-02 1.18295662e-01
6.58420682e-01 8.79610896e-01 -2.63825238e-01 -1.37229824e+00
2.67124325e-01 1.96930334e-01 5.37946284e-01 -7.64596105e-01
-4.72231954e-01 2.30591580e-01 -4.54511940e-01 4.26579593e-03
6.80073917e-01 -1.49925984e-02 -2.29960397e-01 2.24898362e+00
1.27153546e-01 7.55482987e-02 -8.13259184e-03 4.07606453e-01
5.15726328e-01 6.00904822e-01 5.06179869e-01 -6.84663415e-01
1.32653987e+00 1.11242145e-01 -9.85976934e-01 1.61937460e-01
6.09330356e-01 -3.71256828e-01 8.95304263e-01 6.16563678e-01
-1.21955371e+00 6.24198578e-02 -1.04719615e+00 7.29036611e-03
-1.48047253e-01 -3.67259830e-01 1.18709505e+00 9.37102437e-01
-9.68329072e-01 4.47848439e-01 -6.23772025e-01 -8.83156732e-02
-3.26255076e-02 2.39943206e-01 -1.68487206e-01 -1.33780748e-01
-1.47375834e+00 8.27097297e-01 3.02267075e-01 8.13470259e-02
-1.06482530e+00 -4.30855900e-01 -8.90119433e-01 -2.20041927e-02
8.08044434e-01 -3.97827834e-01 1.40055013e+00 -2.66193569e-01
-1.18152630e+00 7.09887087e-01 -2.67810076e-01 -4.19150978e-01
4.05271232e-01 -7.43759936e-03 -5.88744342e-01 -5.20226769e-02
1.09665930e-01 1.16579875e-01 6.19659908e-02 -1.14993036e+00
-4.96642649e-01 -6.90920293e-01 7.57897913e-01 -2.36558095e-01
4.15238023e-01 5.55513859e-01 2.55301625e-01 -1.72370628e-01
1.28935665e-01 -7.47128606e-01 -3.97292197e-01 -3.70371997e-01
-7.10201502e-01 -6.96573675e-01 8.27185363e-02 -1.06461141e-02
1.45980072e+00 -1.68103194e+00 2.17649251e-01 3.90011162e-01
2.74767250e-01 -3.08052242e-01 1.80313542e-01 4.55127537e-01
-3.44563752e-01 5.69389999e-01 -4.85738903e-01 2.27908000e-01
6.23334348e-01 5.55237472e-01 -4.41773683e-01 5.42078555e-01
-4.99742329e-02 7.52353609e-01 -8.80469203e-01 -8.13530326e-01
1.85355887e-01 -7.52303451e-02 -3.92986327e-01 -6.53673289e-03
-7.60550022e-01 -1.83767691e-01 -6.61756396e-01 2.63148397e-01
4.85842228e-01 -6.87861964e-02 5.72390497e-01 2.52534717e-01
-4.52851027e-01 4.69281495e-01 -1.70065534e+00 1.18950498e+00
-2.69654334e-01 2.41070077e-01 3.27779390e-02 -8.55739474e-01
3.84784251e-01 5.62053919e-01 1.87660605e-01 -3.16382051e-01
1.77707061e-01 2.27023233e-02 -2.27712363e-01 -5.13092101e-01
3.15871328e-01 -7.72687614e-01 -7.63080716e-01 6.95099533e-01
-2.80287772e-01 -4.85720277e-01 5.02058029e-01 3.29853594e-01
1.08334243e+00 1.71880767e-01 5.75572610e-01 -6.54678941e-01
5.10363281e-01 2.59900708e-02 5.82913220e-01 1.15752661e+00
-9.02580917e-02 6.43651113e-02 1.16515684e+00 -2.16460288e-01
-6.55484438e-01 -1.48218536e+00 -3.44940066e-01 8.97454023e-01
4.27905172e-02 -5.09842157e-01 -5.56956649e-01 -3.56531620e-01
-7.48116449e-02 1.23689890e+00 -8.04617763e-01 8.46536011e-02
-4.80436444e-01 -8.25958610e-01 1.03628528e+00 4.34181899e-01
3.06257904e-02 -7.52743840e-01 -4.85022128e-01 3.53730842e-02
-3.74470621e-01 -8.26632261e-01 2.00718969e-01 2.80213892e-01
-8.62084687e-01 -1.31175292e+00 2.15032220e-01 9.99580249e-02
2.12544069e-01 -3.42491925e-01 1.00770032e+00 -1.11131117e-01
-2.08150283e-01 3.17259461e-01 -7.74562955e-02 -7.30976105e-01
-6.37824535e-01 -5.36114931e-01 4.55618054e-02 -6.73047423e-01
3.59332263e-01 -5.89161456e-01 6.74223527e-02 1.85859591e-01
-1.06488311e+00 -4.12210703e-01 8.15236121e-02 4.68735337e-01
4.91836339e-01 6.27161086e-01 3.03393066e-01 -1.13581419e+00
5.24351001e-01 -5.00200689e-01 -1.05139124e+00 2.49321148e-01
-6.08958602e-01 6.29734218e-01 4.43226129e-01 7.15693235e-02
-1.14467597e+00 -1.67845652e-01 1.31779667e-02 2.97269166e-01
-1.58979833e-01 8.19642961e-01 -5.76569140e-01 4.73647803e-01
7.91532695e-01 -2.02353239e-01 -4.27077144e-01 -2.00964779e-01
4.92067903e-01 3.56916398e-01 5.13184726e-01 -1.16903043e+00
8.84295225e-01 6.05680883e-01 5.80823421e-01 -3.25078636e-01
-1.12632620e+00 -5.63289188e-02 -2.39548832e-01 -5.00430539e-02
7.80325949e-01 -5.21898866e-01 -1.39547241e+00 -6.78411126e-02
-1.04774439e+00 -2.85338163e-01 -3.77078533e-01 3.69380802e-01
-8.85975420e-01 4.13260102e-01 -3.98578495e-01 -1.49698746e+00
3.57782900e-01 -8.95593584e-01 9.04079914e-01 -1.04489408e-01
-4.01953757e-01 -9.66001868e-01 3.31275016e-01 1.37170777e-01
-1.46743536e-01 5.05402267e-01 1.38628554e+00 -2.97853649e-01
-4.13031310e-01 -1.93877369e-01 -3.07770044e-01 -9.98506695e-02
-3.83981287e-01 3.53589118e-01 -8.18037510e-01 4.65815544e-01
-1.41409725e-01 -1.74707428e-01 5.62514365e-01 4.90998685e-01
1.00478876e+00 -5.18197000e-01 -2.82068133e-01 -3.44719201e-01
1.62606776e+00 1.93908930e-01 8.73618424e-01 -6.11442588e-02
2.15480909e-01 6.55545354e-01 4.85165209e-01 4.47882444e-01
5.98482847e-01 4.94092941e-01 3.64476383e-01 6.15413666e-01
1.83148280e-01 -4.08848375e-01 1.85172006e-01 2.83152908e-01
-5.79750717e-01 -3.13848495e-01 -9.28134739e-01 4.84276295e-01
-2.05682039e+00 -1.59838700e+00 -6.24962389e-01 2.15576315e+00
1.29892898e+00 1.39588416e-01 1.99603096e-01 5.67999184e-01
7.42910206e-01 5.58476038e-02 3.45589280e-01 -5.64855218e-01
-5.77346645e-02 6.13013089e-01 5.17533302e-01 1.03154230e+00
-7.24725306e-01 6.09309614e-01 7.53705215e+00 7.34870076e-01
-5.42121157e-02 1.18621968e-01 2.22349301e-01 1.73797309e-01
-8.32401633e-01 5.10216177e-01 -7.21498311e-01 2.51492351e-01
1.42340612e+00 -2.72992700e-01 2.94963300e-01 5.64380705e-01
2.14413822e-01 -6.17379665e-01 -1.31080198e+00 4.45611447e-01
-1.82281777e-01 -1.11038387e+00 -2.45591953e-01 5.39039914e-03
6.05620980e-01 -5.79598010e-01 -2.52488464e-01 2.25111619e-02
1.23257434e+00 -1.31673872e+00 8.76374424e-01 5.18576324e-01
6.08813882e-01 -8.27612281e-01 7.13016689e-01 3.30869496e-01
-1.05525112e+00 -1.41064286e-01 -2.20400333e-01 -4.90394115e-01
4.35946852e-01 1.03966641e+00 -2.68064320e-01 7.52134442e-01
4.46145713e-01 4.72219363e-02 -1.63399465e-02 5.50946355e-01
-7.16220081e-01 6.18571520e-01 -5.97172499e-01 -2.52742350e-01
-3.74642536e-02 8.85257646e-02 4.15575564e-01 1.11437869e+00
1.04432471e-01 6.45281494e-01 -1.44250423e-01 9.99319553e-01
2.95441896e-01 -4.56275523e-01 -5.96759558e-01 -1.58568531e-01
4.83651668e-01 6.29857779e-01 -4.54482764e-01 -1.92069411e-01
-8.11009258e-02 -8.56889263e-02 1.94849204e-02 6.96759447e-02
-1.03632414e+00 -2.85316914e-01 3.50482881e-01 1.52183741e-01
-2.26632401e-01 -1.47845611e-01 -3.44457924e-01 -9.68802273e-01
-2.66021311e-01 -5.85459888e-01 8.21455479e-01 -8.15640748e-01
-1.09711742e+00 7.99041167e-02 9.47245657e-01 -5.52997291e-01
-4.47997540e-01 -7.61508167e-01 -3.15813869e-01 8.45949650e-01
-1.04427075e+00 -7.97890007e-01 2.88907826e-01 6.53416634e-01
-3.06999236e-01 6.61676824e-01 9.68743622e-01 -9.70774367e-02
-4.27032799e-01 1.44405618e-01 -5.21562457e-01 -3.08344007e-01
1.03298157e-01 -1.62500834e+00 -3.04525882e-01 1.02450287e+00
-5.13133667e-02 6.64075375e-01 1.30431926e+00 -5.85547924e-01
-1.69027078e+00 -6.52104616e-01 1.53257179e+00 -6.43204570e-01
9.65772748e-01 -2.20404744e-01 -2.66359478e-01 9.86962259e-01
-6.50935546e-02 -2.69067466e-01 5.78560829e-01 3.91124278e-01
-7.35856235e-01 -3.46835107e-02 -1.28163946e+00 7.61406064e-01
1.24717426e+00 -6.31166518e-01 -1.04489028e+00 3.28997225e-01
6.06959999e-01 -3.79327163e-02 -8.98824871e-01 4.02743161e-01
6.25528574e-01 -9.44219172e-01 7.80430198e-01 -7.40953565e-01
4.35180694e-01 -5.03502548e-01 -4.69691068e-01 -6.01273715e-01
-1.35493502e-01 -5.85123003e-01 1.47208363e-01 1.03046727e+00
6.03026748e-01 -6.54071391e-01 2.28889614e-01 8.38052273e-01
7.84954652e-02 -1.16178058e-01 -1.22611749e+00 -6.55898094e-01
3.20827961e-01 -1.16825783e+00 5.82385361e-01 8.81262898e-01
6.90049350e-01 8.25304687e-02 -7.98940659e-02 3.43715340e-01
9.02510464e-01 2.55585402e-01 3.76945078e-01 -1.51805007e+00
-4.74809617e-01 -4.55945075e-01 -3.17794800e-01 -3.26953322e-01
4.09716487e-01 -7.03345537e-01 2.93751001e-01 -1.55275309e+00
3.82566988e-01 -6.33173347e-01 -2.63019986e-02 6.69159114e-01
6.99750930e-02 -1.31365925e-01 -1.51690811e-01 -2.00701281e-01
-5.61774909e-01 1.33243725e-01 9.87628102e-01 -6.34966418e-02
2.31812999e-01 -9.28665686e-04 -8.54186177e-01 9.04724181e-01
5.20997047e-01 -6.92676246e-01 -5.79882443e-01 -1.20401688e-01
1.05929196e+00 7.56238818e-01 6.17037952e-01 -5.77347994e-01
5.88242263e-02 -9.14018154e-01 -3.21474463e-01 -4.30920482e-01
7.45270168e-04 -5.71316659e-01 6.23008788e-01 5.09327292e-01
-7.38425136e-01 7.35446289e-02 7.06116557e-02 4.75861102e-01
2.81668622e-02 -7.19007254e-01 5.75003147e-01 -1.99680611e-01
-4.37682778e-01 -3.03114038e-02 -5.81268430e-01 2.96171874e-01
9.04334486e-01 3.36439878e-01 -5.58722198e-01 -3.55802059e-01
-8.39780867e-01 7.44124502e-02 -3.71330455e-02 -2.25472674e-01
3.20929080e-01 -1.03205681e+00 -5.57647049e-01 -7.51101375e-01
7.24328831e-02 -1.95860639e-01 2.83859164e-01 9.11481023e-01
-3.85272652e-01 6.52765393e-01 1.58782676e-01 -1.52987778e-01
-7.45477796e-01 6.61096990e-01 9.34975445e-02 -3.27525824e-01
-1.34941056e-01 7.78591216e-01 2.33635589e-01 -4.86505702e-02
1.28665753e-02 -3.99382770e-01 1.70530789e-02 -2.23564282e-01
6.22177482e-01 6.49676859e-01 -5.25935233e-01 -3.20096910e-01
-6.52124166e-01 1.71392918e-01 4.34084892e-01 -6.75702870e-01
1.10841417e+00 -2.53456563e-01 -7.64638722e-01 8.03552210e-01
6.16670668e-01 2.50096291e-01 -4.29142863e-01 -1.04645155e-01
2.12957710e-01 5.07659279e-02 -1.31739885e-01 -8.08254421e-01
-4.52654749e-01 6.62570655e-01 1.65328890e-01 5.54608345e-01
9.03007865e-01 6.00505710e-01 1.82830468e-02 3.19253922e-01
8.41387808e-01 -8.07845652e-01 -6.56830072e-01 1.70888215e-01
8.93790722e-01 -6.24992609e-01 2.33525500e-01 -7.40244627e-01
-8.70103538e-02 9.28444147e-01 1.35988310e-01 1.30511532e-02
6.79331839e-01 7.36695170e-01 -5.81110835e-01 -2.77383685e-01
-1.01220036e+00 -4.72823977e-01 -2.10772485e-01 5.04361987e-01
4.95823026e-01 6.28618062e-01 -9.64697063e-01 9.68189776e-01
-3.82901162e-01 2.84752935e-01 9.30388391e-01 9.14872229e-01
-4.80867654e-01 -1.09134710e+00 -5.45768201e-01 1.80553004e-01
-5.84756672e-01 -1.07712708e-01 -2.54737586e-01 1.05884850e+00
1.11704461e-01 1.35398805e+00 -2.66479760e-01 -1.49311900e-01
1.62715599e-01 1.84081391e-01 1.00014222e+00 -5.62725008e-01
-2.52946764e-01 -1.92384347e-01 6.28375828e-01 -5.21796048e-01
-7.57512927e-01 -7.52712488e-01 -1.37907934e+00 -6.27685606e-01
-2.50879556e-01 5.30891001e-01 3.17774504e-01 1.21848857e+00
-5.36473572e-01 3.57034981e-01 2.86318660e-01 -4.29085679e-02
-8.07625055e-01 -7.69190848e-01 -9.52242017e-01 -2.74880528e-01
4.06869687e-03 -7.88275838e-01 -6.48987770e-01 5.12864627e-02] | [8.321263313293457, 6.07021427154541] |
c67c4bde-9e06-4826-a0f2-f66cc1934e53 | on-the-relation-between-syntactic-divergence | 2110.04644 | null | https://arxiv.org/abs/2110.04644v1 | https://arxiv.org/pdf/2110.04644v1.pdf | On the Relation between Syntactic Divergence and Zero-Shot Performance | We explore the link between the extent to which syntactic relations are preserved in translation and the ease of correctly constructing a parse tree in a zero-shot setting. While previous work suggests such a relation, it tends to focus on the macro level and not on the level of individual edges-a gap we aim to address. As a test case, we take the transfer of Universal Dependencies (UD) parsing from English to a diverse set of languages and conduct two sets of experiments. In one, we analyze zero-shot performance based on the extent to which English source edges are preserved in translation. In another, we apply three linguistically motivated transformations to UD, creating more cross-lingually stable versions of it, and assess their zero-shot parsability. In order to compare parsing performance across different schemes, we perform extrinsic evaluation on the downstream task of cross-lingual relation extraction (RE) using a subset of a popular English RE benchmark translated to Russian and Korean. In both sets of experiments, our results suggest a strong relation between cross-lingual stability and zero-shot parsing performance. | ['Omri Abend', 'Taelin Karidi', 'Dmitry Nikolaev', 'Ofir Arviv'] | 2021-10-09 | null | https://aclanthology.org/2021.emnlp-main.394 | https://aclanthology.org/2021.emnlp-main.394.pdf | emnlp-2021-11 | ['cross-lingual-zero-shot-dependency-parsing'] | ['natural-language-processing'] | [ 5.30181043e-02 2.81113803e-01 -3.17001611e-01 -5.18217385e-01
-1.24079037e+00 -9.64551032e-01 5.10446012e-01 3.40566605e-01
-4.17854935e-01 7.10794210e-01 6.14159644e-01 -7.07132280e-01
1.38526455e-01 -8.08095276e-01 -7.16496408e-01 -2.04111323e-01
1.98610827e-01 2.86129892e-01 2.15752631e-01 -4.91582155e-01
-2.20483825e-01 1.86239481e-01 -8.87327433e-01 3.22287828e-01
8.06291938e-01 2.88529932e-01 -9.85378623e-02 4.23614085e-01
-1.25000745e-01 5.79110920e-01 -2.23478705e-01 -1.06158221e+00
2.16737226e-01 -6.05480492e-01 -1.18650711e+00 -8.48285705e-02
3.72765481e-01 2.98258662e-01 6.36865124e-02 1.17651117e+00
5.15960634e-01 -9.05072838e-02 2.65351593e-01 -6.79617286e-01
-6.93421543e-01 1.18918848e+00 -3.90175939e-01 6.68494701e-01
4.44942564e-01 6.31933007e-03 1.55550921e+00 -7.16403186e-01
1.31432652e+00 1.34669602e+00 7.73038268e-01 3.08041751e-01
-1.58522618e+00 -3.41052115e-01 -1.54156303e-02 -1.66858330e-01
-8.84132266e-01 -9.16672170e-01 4.67242628e-01 -3.48211110e-01
1.34034610e+00 -5.61834089e-02 -1.23200960e-01 1.02484536e+00
4.38158870e-01 2.59063870e-01 1.27351499e+00 -9.34130073e-01
3.08997855e-02 4.12493832e-02 6.14297926e-01 5.98799050e-01
2.56344467e-01 4.41005975e-02 -5.92649221e-01 1.40888885e-01
6.04027063e-02 -9.47299898e-01 -1.85194150e-01 -7.39989057e-02
-1.16866839e+00 8.35651517e-01 1.89662486e-01 7.26098359e-01
-1.70377716e-02 -3.65812510e-01 7.08574772e-01 8.12201202e-01
6.03581369e-01 4.49564159e-01 -6.97772145e-01 -2.32477263e-01
-5.95917761e-01 -2.09935948e-01 1.03055644e+00 8.82807970e-01
8.41249228e-01 -3.04194063e-01 -1.34221584e-01 8.89441073e-01
-2.30196610e-01 2.64491215e-02 5.42266309e-01 -5.74959099e-01
9.24092770e-01 4.94852066e-01 -4.21568543e-01 -6.02961719e-01
-3.49603385e-01 -3.33491385e-01 -2.56609052e-01 -9.17369649e-02
5.80026031e-01 -4.13550436e-01 -5.79609632e-01 2.19875741e+00
3.28286320e-01 -5.43638408e-01 4.51085746e-01 5.42523146e-01
7.08983541e-01 4.64103937e-01 2.13172063e-01 -3.68343294e-01
1.63220215e+00 -9.41505730e-01 -5.89585483e-01 -5.83354414e-01
1.23097289e+00 -1.00282907e+00 1.35349214e+00 -2.08703175e-01
-1.20228493e+00 -5.67346513e-01 -8.82317305e-01 -5.19625187e-01
-3.68574768e-01 7.44524002e-02 4.96386021e-01 5.87737024e-01
-1.01851547e+00 8.23429465e-01 -8.81737053e-01 -7.73761272e-01
-2.04437956e-01 6.92711100e-02 -8.36637974e-01 -4.35905494e-02
-1.36137462e+00 1.30400562e+00 2.82369882e-01 -3.61033082e-01
-1.68985613e-02 -5.76170385e-01 -1.07226884e+00 1.04946658e-01
2.63718903e-01 -4.65347350e-01 1.32344663e+00 -8.35944772e-01
-1.24681735e+00 1.36063147e+00 -2.79635459e-01 -1.76625818e-01
2.31228679e-01 -1.40430510e-01 -3.68219733e-01 -3.36845487e-01
4.65169817e-01 3.11143458e-01 1.40933171e-01 -9.46021914e-01
-7.73229837e-01 -5.62870979e-01 2.99992859e-01 1.30261391e-01
-1.18679203e-01 6.23262644e-01 -4.36308771e-01 -5.53143740e-01
7.12948069e-02 -1.11709249e+00 9.18845385e-02 -8.35571289e-01
-4.21686977e-01 -2.78174996e-01 4.31406021e-01 -9.53728974e-01
1.41003668e+00 -2.27648902e+00 4.87852126e-01 -8.97411406e-02
-3.72331381e-01 -7.16672316e-02 -2.11920962e-01 5.36213636e-01
-3.84434730e-01 3.98576230e-01 -4.46818471e-01 -6.17441893e-01
-8.02753866e-02 4.24756825e-01 -6.23413622e-02 2.57781565e-01
4.70802575e-01 1.15243173e+00 -7.61598229e-01 -5.51887393e-01
-2.98363179e-01 1.79675922e-01 -5.46168506e-01 -1.05579101e-01
4.35174033e-02 4.07166153e-01 -1.55991852e-01 5.73150992e-01
1.95056245e-01 1.40790030e-01 7.40376830e-01 -2.12306768e-01
-3.26446921e-01 1.04862499e+00 -7.77596354e-01 1.86349165e+00
-5.90814352e-01 4.69729185e-01 -3.81581411e-02 -7.56092012e-01
7.08860755e-01 3.06764692e-01 1.11702017e-01 -1.08101201e+00
1.05230242e-01 3.21909815e-01 5.16289294e-01 -1.64679766e-01
6.61787152e-01 -4.81128544e-01 -5.16406536e-01 6.44285738e-01
3.88060898e-01 9.88833830e-02 5.78085661e-01 1.89899266e-01
1.22843504e+00 4.76988137e-01 5.95193088e-01 -6.80331290e-01
2.74628133e-01 2.90148377e-01 7.58091688e-01 2.42166340e-01
-6.62492514e-02 4.39721137e-01 9.38121736e-01 -1.94907844e-01
-1.05194116e+00 -8.41677845e-01 -2.91616529e-01 1.36658597e+00
-4.69318032e-02 -6.65296495e-01 -9.20633554e-01 -9.68409598e-01
-3.88047248e-01 9.45890129e-01 -6.13163114e-01 -1.01581693e-01
-1.12018943e+00 -9.52538073e-01 6.95754468e-01 3.82458806e-01
1.74900368e-01 -9.47662055e-01 -4.53651369e-01 2.51866937e-01
-5.11187315e-01 -1.60932863e+00 -2.96071500e-01 6.04344308e-01
-7.42854357e-01 -9.08420622e-01 -7.64425620e-02 -1.01774132e+00
1.93848103e-01 -5.68246581e-02 1.76551020e+00 -1.40019670e-01
1.98834226e-01 5.35576865e-02 -5.99494040e-01 -3.66984978e-02
-7.74624646e-01 6.13609374e-01 -1.52307590e-02 -4.30295706e-01
5.68221927e-01 -5.84143460e-01 8.46471786e-02 2.02447012e-01
-5.50321817e-01 -2.21030250e-01 5.08876324e-01 5.85945547e-01
6.47418916e-01 -2.39450112e-01 3.81309330e-01 -1.41215634e+00
7.85513043e-01 -4.76932019e-01 -4.11198229e-01 4.27766889e-01
-5.13603270e-01 4.64060426e-01 6.20602250e-01 -5.75051382e-02
-1.11781406e+00 1.20238823e-04 -2.28682965e-01 3.47988904e-01
3.06277443e-03 5.93838930e-01 -3.85575175e-01 2.42953166e-01
8.15641105e-01 -3.16030413e-01 -3.63022834e-01 -5.98220944e-01
7.23376453e-01 5.06650567e-01 6.05939329e-01 -9.78904903e-01
6.75197661e-01 -1.60759613e-01 -1.96893305e-01 -5.34443796e-01
-9.11062896e-01 -1.67477831e-01 -1.02100718e+00 4.15772736e-01
1.13772225e+00 -8.29227090e-01 2.50661988e-02 -1.03361860e-01
-1.15599155e+00 -5.37852347e-01 -3.49759251e-01 2.59883344e-01
-4.26234514e-01 3.49564224e-01 -9.87608612e-01 -5.04915118e-02
-2.98947990e-01 -1.30831325e+00 1.03486598e+00 -2.85094291e-01
-5.14167547e-01 -1.14315701e+00 4.62718397e-01 1.79501072e-01
2.61328835e-02 1.83258355e-01 1.52514124e+00 -6.16621137e-01
-1.63173884e-01 5.41412272e-02 -2.09897920e-01 1.60590053e-01
3.16254139e-01 -9.67967808e-02 -7.84137249e-01 -2.48043299e-01
-1.82612818e-02 -2.83978939e-01 7.20218241e-01 8.16308260e-02
2.35178564e-02 -6.22875616e-02 -2.14561388e-01 4.76719379e-01
1.46853304e+00 -2.05172345e-01 6.28250718e-01 5.58459044e-01
5.72233737e-01 9.12755072e-01 5.88716984e-01 -3.24269950e-01
6.36549532e-01 8.35464299e-01 2.52476381e-03 -1.81290820e-01
-4.12906826e-01 -2.01133162e-01 7.26637840e-01 1.37784123e+00
2.24056467e-02 -2.00778767e-01 -9.91724312e-01 6.03666902e-01
-1.56085038e+00 -4.81910855e-01 -1.32456213e-01 2.12631035e+00
1.08526766e+00 3.35033804e-01 2.45704502e-01 -7.21599236e-02
7.24125266e-01 1.99025407e-01 1.95782229e-01 -8.80942702e-01
-1.99917674e-01 4.99750912e-01 4.32153463e-01 7.72859395e-01
-9.35555756e-01 1.50429523e+00 5.94787455e+00 5.39508760e-01
-1.06788933e+00 4.71569210e-01 6.24621034e-01 1.75006479e-01
-2.77351260e-01 5.22464156e-01 -8.96938801e-01 2.77966559e-01
1.26365697e+00 -1.03197202e-01 2.13861719e-01 6.63811922e-01
-1.36586115e-01 2.49360353e-02 -1.59149396e+00 4.22019124e-01
-2.03113958e-01 -9.42853332e-01 -2.65994102e-01 2.21568979e-02
4.90548104e-01 3.51209134e-01 -5.26255608e-01 4.73889798e-01
6.84134483e-01 -9.14440274e-01 7.92921305e-01 -1.54917046e-01
1.03634882e+00 -6.01486683e-01 7.70281315e-01 1.86620325e-01
-1.24546957e+00 2.89897442e-01 -2.84333825e-01 -8.71096179e-02
3.91358197e-01 3.72452348e-01 -4.46435153e-01 8.48115563e-01
6.45218253e-01 5.31486988e-01 -7.52400219e-01 1.29578620e-01
-5.66394031e-01 6.68357491e-01 -1.21990897e-01 3.80581141e-01
2.14425862e-01 -3.38982552e-01 5.50953269e-01 1.73022485e+00
1.49147555e-01 6.02357686e-02 -2.30029952e-02 4.61970836e-01
-3.14062297e-01 4.72325176e-01 -6.57218099e-01 5.28934188e-02
3.62004012e-01 1.32719815e+00 -7.83972025e-01 -2.38920808e-01
-7.67130435e-01 9.53385770e-01 9.55541968e-01 2.36676987e-02
-6.47666395e-01 -1.10243015e-01 7.97265232e-01 3.38652246e-02
1.48685217e-01 -3.94434214e-01 -3.39250952e-01 -1.45751774e+00
2.74467319e-01 -1.05739963e+00 4.80048507e-01 -3.63393426e-01
-1.14421856e+00 8.82633030e-01 -2.44473338e-01 -8.60319495e-01
-3.35125387e-01 -6.07252598e-01 -5.71251392e-01 9.89637494e-01
-1.62116206e+00 -1.24036598e+00 2.15686455e-01 2.82469064e-01
5.84179282e-01 1.61216170e-01 1.04813707e+00 4.24278319e-01
-7.92355061e-01 8.07229877e-01 -2.85787851e-01 4.30489689e-01
8.50260496e-01 -1.21466386e+00 1.03262424e+00 1.47103214e+00
5.06517529e-01 6.97229445e-01 5.75366199e-01 -6.20020390e-01
-1.23608148e+00 -9.83708024e-01 1.51961148e+00 -6.65908277e-01
9.17288899e-01 -4.14738268e-01 -9.81882393e-01 1.12939167e+00
3.47129971e-01 4.24140655e-02 5.64824104e-01 7.44972229e-01
-4.73236114e-01 3.29447567e-01 -8.03388059e-01 5.22508204e-01
1.30688357e+00 -7.62931824e-01 -8.60024095e-01 9.19501856e-02
8.62278938e-01 -3.03960562e-01 -1.29010022e+00 3.67910832e-01
3.20909768e-01 -1.00335550e+00 5.65669715e-01 -9.22668755e-01
8.50081563e-01 6.03137054e-02 -5.10688066e-01 -1.60736537e+00
-5.74129760e-01 -5.27255297e-01 5.49761415e-01 1.83831978e+00
9.57893372e-01 -5.80298424e-01 2.69815296e-01 5.10731936e-01
-3.68652344e-01 -6.05284393e-01 -1.03876293e+00 -8.48998487e-01
7.10172594e-01 -3.99443984e-01 3.75314653e-01 1.22831190e+00
3.02211076e-01 1.12296915e+00 -2.58721430e-02 3.18872891e-02
2.92040884e-01 2.05162540e-01 7.00485289e-01 -1.00725532e+00
-6.19571328e-01 -2.95067519e-01 -2.09999606e-01 -5.66012084e-01
3.73783499e-01 -1.29630101e+00 1.17559455e-01 -1.42419660e+00
2.23896161e-01 -5.56296587e-01 -1.53152883e-01 6.64533138e-01
-3.42400461e-01 2.12671474e-01 2.77530611e-01 3.73741418e-01
-3.77218634e-01 1.25242144e-01 7.16897428e-01 3.69934589e-01
-2.17879176e-01 -4.33464736e-01 -9.35931802e-01 6.39980137e-01
7.09927201e-01 -6.38233125e-01 -4.43997197e-02 -9.10180867e-01
3.79421651e-01 1.32875144e-01 -2.63122380e-01 -5.55771589e-01
-1.86525375e-01 4.30226661e-02 -2.23259270e-01 1.92531392e-01
-1.69454157e-01 -3.34027737e-01 -9.16400328e-02 1.37553543e-01
-2.19656065e-01 5.71711659e-01 2.43553579e-01 6.42534047e-02
-3.55529010e-01 -2.60161370e-01 8.25578392e-01 -2.38505691e-01
-7.70720541e-01 -1.37408808e-01 5.65615706e-02 5.66504896e-01
6.39327168e-01 -1.38500616e-01 -5.13602614e-01 1.15004607e-01
-7.14473248e-01 -1.66166440e-01 8.07902873e-01 5.50057411e-01
-3.28504503e-01 -1.03127933e+00 -8.30574036e-01 7.46047720e-02
2.89898574e-01 -4.20447111e-01 -4.39219534e-01 9.08807218e-01
-2.52378732e-01 2.42193013e-01 -9.35802609e-02 -2.60882527e-01
-1.21680355e+00 5.35272300e-01 2.26612478e-01 -5.53942204e-01
-6.86399519e-01 7.55889177e-01 1.16550274e-01 -6.83849394e-01
-3.16339314e-01 -3.81424665e-01 7.24840164e-03 8.30568001e-02
2.52055209e-02 8.41102973e-02 5.18272519e-01 -1.00637245e+00
-4.96045142e-01 5.72437286e-01 3.57483588e-02 -1.97465971e-01
1.27532387e+00 -3.91636252e-01 -3.29434752e-01 4.95052695e-01
1.12352812e+00 6.90422058e-01 -8.29745770e-01 -3.02530885e-01
6.57406569e-01 -1.06739238e-01 -2.51125246e-01 -7.59443223e-01
-8.08590710e-01 6.40892982e-01 1.05555274e-01 2.10438311e-01
8.47825348e-01 3.44146371e-01 8.47959340e-01 1.30835444e-01
3.59116316e-01 -1.03588951e+00 -5.07928789e-01 9.02849972e-01
4.54758793e-01 -1.43871236e+00 -2.23557428e-01 -7.85704315e-01
-6.49502397e-01 9.30343449e-01 2.81861037e-01 -5.87249957e-02
3.02926570e-01 5.86316347e-01 2.51533300e-01 -1.43128648e-01
-6.78552926e-01 -4.33663040e-01 9.50199813e-02 2.42501661e-01
1.11289883e+00 1.80424765e-01 -7.23926365e-01 6.74225271e-01
-7.91075945e-01 -3.96365494e-01 3.64040196e-01 8.57355714e-01
-5.85332811e-02 -1.67252505e+00 1.51164696e-01 -4.36328985e-02
-8.86991441e-01 -5.67123413e-01 -5.82157612e-01 1.05098581e+00
5.61114922e-02 1.09009349e+00 1.21851210e-02 -3.55232209e-01
5.98631680e-01 3.89073640e-01 5.66678524e-01 -1.03121352e+00
-6.80384398e-01 -5.55407256e-02 7.98208594e-01 -5.15393198e-01
-3.04035485e-01 -8.29499900e-01 -1.10463822e+00 -3.39280486e-01
-4.23457325e-01 9.00843740e-02 6.02396607e-01 1.13856733e+00
3.73284906e-01 6.01532578e-01 2.78094858e-01 -3.91464710e-01
-4.16158289e-01 -1.06096649e+00 -1.78380519e-01 8.20955157e-01
-1.67177334e-01 -3.62429589e-01 -1.90264478e-01 1.81684151e-01] | [10.515064239501953, 9.79746150970459] |
7bdb29bf-e9fa-4dfe-b0a3-048f5e4e2f9a | movie-visual-model-based-policy-adaptation | 2307.00972 | null | https://arxiv.org/abs/2307.00972v1 | https://arxiv.org/pdf/2307.00972v1.pdf | MoVie: Visual Model-Based Policy Adaptation for View Generalization | Visual Reinforcement Learning (RL) agents trained on limited views face significant challenges in generalizing their learned abilities to unseen views. This inherent difficulty is known as the problem of $\textit{view generalization}$. In this work, we systematically categorize this fundamental problem into four distinct and highly challenging scenarios that closely resemble real-world situations. Subsequently, we propose a straightforward yet effective approach to enable successful adaptation of visual $\textbf{Mo}$del-based policies for $\textbf{Vie}$w generalization ($\textbf{MoVie}$) during test time, without any need for explicit reward signals and any modification during training time. Our method demonstrates substantial advancements across all four scenarios encompassing a total of $\textbf{18}$ tasks sourced from DMControl, xArm, and Adroit, with a relative improvement of $\mathbf{33}$%, $\mathbf{86}$%, and $\mathbf{152}$% respectively. The superior results highlight the immense potential of our approach for real-world robotics applications. Videos are available at https://yangsizhe.github.io/MoVie/ . | ['Huazhe Xu', 'Yanjie Ze', 'Sizhe Yang'] | 2023-07-03 | null | null | null | null | ['reinforcement-learning-1'] | ['methodology'] | [-9.98728946e-02 -9.86717269e-02 -4.88123158e-03 -3.10558379e-01
-6.46362662e-01 -6.74294949e-01 4.84421611e-01 -3.02085280e-01
-6.57285929e-01 9.49655771e-01 -5.32187939e-01 -1.98916554e-01
-1.99892402e-01 -5.54484069e-01 -1.21492100e+00 -6.29306555e-01
-3.18237305e-01 7.64532238e-02 9.28939059e-02 -5.09443641e-01
3.03577065e-01 2.54809350e-01 -1.74960554e+00 2.04754714e-02
6.85832679e-01 1.35143137e+00 6.37557507e-01 3.92873138e-01
4.17272061e-01 7.12514937e-01 -7.66833127e-01 -4.67592627e-01
5.71421206e-01 -2.28781849e-01 -4.30584937e-01 -7.11584091e-02
6.20391190e-01 -6.99723840e-01 -4.28018719e-01 1.24433541e+00
4.21034962e-01 5.96278965e-01 4.78995770e-01 -1.45860577e+00
-8.93301666e-01 1.67363107e-01 -6.49456680e-01 3.55091900e-01
3.50332886e-01 7.82164097e-01 8.30669284e-01 -9.29994524e-01
7.18192220e-01 1.08333313e+00 1.03149801e-01 9.02009249e-01
-9.36877429e-01 -1.04130363e+00 6.92781508e-01 3.75578851e-01
-9.64958191e-01 -3.85932505e-01 6.47331357e-01 -4.47890073e-01
1.12565255e+00 1.46314055e-01 5.79535961e-01 1.45736265e+00
-3.16954777e-02 9.59671736e-01 1.33119464e+00 -8.49611685e-02
4.02951121e-01 -8.26813355e-02 -2.56672978e-01 8.92397344e-01
1.23601168e-01 5.69665968e-01 -5.36358416e-01 4.81611669e-01
1.02324712e+00 1.34954974e-01 -3.85161817e-01 -4.31968808e-01
-1.10715151e+00 7.98592925e-01 6.95337355e-01 -1.69977576e-01
-1.94424421e-01 5.25389433e-01 3.28065395e-01 1.97994843e-01
2.71947458e-02 4.79565799e-01 -3.91683012e-01 -1.88788056e-01
-3.81315380e-01 3.41151655e-01 1.25615507e-01 1.49078727e+00
7.24614263e-01 6.58622563e-01 1.42424449e-01 8.87814224e-01
1.83184654e-01 9.30467069e-01 3.25326353e-01 -1.50943840e+00
6.99532747e-01 3.03515047e-01 3.55672926e-01 -7.11365342e-01
-3.57884288e-01 -3.24443996e-01 -6.03697956e-01 9.18138325e-01
2.71724761e-01 -2.91249990e-01 -1.03164041e+00 2.02686501e+00
1.34937704e-01 -2.19254091e-01 -8.22222605e-02 9.70448315e-01
7.31148005e-01 4.97841686e-01 2.05652386e-01 -1.36726007e-01
9.08335567e-01 -1.18712962e+00 -3.03668618e-01 -5.33680022e-01
1.20638698e-01 -4.25871223e-01 1.61024570e+00 5.79612911e-01
-1.28568017e+00 -6.08050048e-01 -1.17664230e+00 2.54657507e-01
-3.08239758e-01 8.36803764e-02 7.40700245e-01 2.71698773e-01
-1.06791806e+00 5.68800628e-01 -7.06070781e-01 -3.04677874e-01
6.39320493e-01 4.95618761e-01 -3.52948129e-01 -1.86610147e-01
-7.76129723e-01 8.01506937e-01 1.53691500e-01 1.33813113e-01
-1.67520630e+00 -4.58903760e-01 -8.30119669e-01 -1.48213074e-01
8.03802252e-01 -4.77925748e-01 1.32877648e+00 -7.73173153e-01
-1.67147434e+00 7.36439466e-01 1.65626362e-01 -3.17903787e-01
6.00640059e-01 -3.54618490e-01 -2.54559159e-01 2.16706529e-01
7.43566006e-02 9.88248527e-01 9.65022683e-01 -1.59701443e+00
-7.63558328e-01 -4.14021313e-01 7.06858516e-01 4.18654650e-01
-2.77029902e-01 -2.38069654e-01 -2.93689489e-01 -7.98845828e-01
-2.50625044e-01 -1.02481568e+00 -1.40235841e-01 9.35937911e-02
6.76320633e-04 -3.74245733e-01 5.67870855e-01 -3.31148773e-01
8.15028489e-01 -1.93417561e+00 4.66255248e-01 -2.25386187e-01
2.88203657e-01 2.65364826e-01 -4.81706522e-02 3.39194030e-01
2.32296258e-01 9.07168910e-02 -7.63055608e-02 -1.03118569e-02
7.20929652e-02 3.33806783e-01 -3.20120782e-01 2.59559900e-01
-1.21783381e-02 8.58272433e-01 -9.29343462e-01 -9.47365388e-02
3.44145149e-01 1.76514722e-02 -5.55253088e-01 3.64005327e-01
-5.08954406e-01 6.29847050e-01 -5.60517550e-01 6.30316496e-01
4.59595501e-01 -2.53324449e-01 2.52962504e-02 1.87927738e-01
-5.74161671e-02 -2.77423710e-01 -9.90909755e-01 1.89598119e+00
-2.52182007e-01 2.85870880e-01 1.93831041e-01 -1.22744608e+00
9.14279401e-01 -1.06729276e-01 4.43748832e-01 -1.08462596e+00
2.46392876e-01 6.35078624e-02 -1.01642124e-01 -5.85584342e-01
3.53393167e-01 -2.30174690e-01 -1.39157161e-01 2.96967834e-01
3.79704207e-01 -3.23313028e-01 4.53165382e-01 1.68847032e-02
9.56389129e-01 6.34407699e-01 3.38781923e-02 -3.01634699e-01
2.91460693e-01 2.30484009e-02 4.78743792e-01 8.59684944e-01
-6.49428546e-01 1.69163302e-01 2.88858384e-01 -4.33052808e-01
-1.00613630e+00 -1.41914654e+00 2.74946779e-01 1.40606940e+00
5.24206877e-01 -4.80613410e-02 -6.97750390e-01 -9.36163545e-01
-3.52417342e-02 7.45471060e-01 -7.14403689e-01 -3.30747634e-01
-6.75008178e-01 -3.67185205e-01 3.12815130e-01 8.22632730e-01
7.38777220e-01 -1.55073106e+00 -1.08916485e+00 -2.53060192e-01
5.28551787e-02 -1.05902243e+00 -2.39126250e-01 1.59699008e-01
-7.49436736e-01 -8.97057414e-01 -9.91862178e-01 -6.59031630e-01
6.32623434e-01 4.85564947e-01 1.00883341e+00 -3.42285521e-02
-2.71218687e-01 5.97794771e-01 -5.93811870e-01 -5.44240177e-01
1.11565985e-01 -2.87272424e-01 2.74841696e-01 -5.26513278e-01
1.23432443e-01 -5.67133129e-01 -9.68819141e-01 5.58943987e-01
-7.69647002e-01 -1.03673846e-01 5.19701183e-01 9.18073833e-01
6.15027070e-01 -4.54719067e-01 6.83175504e-01 -3.43043119e-01
4.76055741e-01 -1.90093532e-01 -8.56330395e-01 3.98576595e-02
-6.95883989e-01 -1.26360357e-01 9.93551850e-01 -4.66129929e-01
-9.42224562e-01 -2.11287335e-01 3.44339199e-02 -8.03497553e-01
-3.02496731e-01 1.45951942e-01 1.82283595e-01 4.55418862e-02
8.32784116e-01 4.68230903e-01 -1.52580991e-01 -2.00537562e-01
2.92213261e-01 2.45900616e-01 4.18422997e-01 -7.56775975e-01
7.59635746e-01 5.08018017e-01 -2.09437326e-01 -4.76560265e-01
-6.10102296e-01 -1.81745142e-02 -1.43016294e-01 -6.21267259e-01
8.73820782e-01 -8.06475461e-01 -1.39188635e+00 2.21379742e-01
-4.95413154e-01 -8.87208521e-01 -3.68982822e-01 5.43952465e-01
-1.06055272e+00 3.65045756e-01 -3.64568949e-01 -6.73310697e-01
-1.96299180e-01 -1.39159977e+00 6.27451599e-01 5.59841573e-01
1.57382056e-01 -5.19366562e-01 -2.34545812e-01 6.67639613e-01
4.46134597e-01 1.92914546e-01 8.21678638e-01 -1.31030545e-01
-7.59753287e-01 1.55862138e-01 -3.12404156e-01 5.32769978e-01
-1.06159866e-01 -3.30832034e-01 -8.50390494e-01 -8.57360363e-01
-1.50768429e-01 -9.76347089e-01 7.08225667e-01 2.18893692e-01
1.32178485e+00 -1.25283211e-01 -1.40657462e-02 5.64697266e-01
1.42659807e+00 7.59082258e-01 4.16745663e-01 4.51579154e-01
3.86354119e-01 2.39982337e-01 6.51036322e-01 6.34893358e-01
3.85335892e-01 5.72432518e-01 1.09729052e+00 1.74730957e-01
3.99476569e-03 -2.17193529e-01 5.85888147e-01 4.14094776e-01
-5.85757792e-01 -2.78212160e-01 -6.60201073e-01 4.42971289e-01
-1.72119093e+00 -1.01936555e+00 6.45687521e-01 2.16172147e+00
4.53348726e-01 1.37564525e-01 1.63758010e-01 -3.78177375e-01
5.63810945e-01 2.30335817e-01 -1.07310820e+00 -3.96256596e-01
6.37930259e-02 3.12395841e-01 2.55420923e-01 1.96722135e-01
-9.74844337e-01 1.07983005e+00 5.16871738e+00 7.83279359e-01
-1.22376752e+00 -8.31024870e-02 5.86577296e-01 -5.53982317e-01
1.15007207e-01 -2.45085478e-01 -5.42975366e-01 4.73476291e-01
4.43157971e-01 -6.92070834e-03 9.54591274e-01 1.17642236e+00
5.15750945e-02 -3.45844626e-01 -1.11342180e+00 1.05832076e+00
1.54904470e-01 -1.12710559e+00 -1.94805879e-02 -5.81281371e-02
7.12841809e-01 1.71987757e-01 6.59722030e-01 7.02409148e-01
7.28512108e-01 -1.10170734e+00 1.03184462e+00 1.23743564e-01
1.16193283e+00 -5.86725771e-01 2.31887430e-01 3.79879802e-01
-1.02324140e+00 -5.47684133e-01 -4.48728114e-01 -1.33862168e-01
-5.98192699e-02 -2.31483459e-01 -4.04257745e-01 5.68624735e-01
1.42565942e+00 7.03987777e-01 -3.92645746e-01 5.93743324e-01
-4.97147501e-01 6.73428103e-02 -5.41685298e-02 -2.69788951e-01
2.50702381e-01 -7.19534010e-02 4.29373592e-01 7.91614711e-01
4.41449434e-01 2.61981279e-01 2.79451996e-01 6.95162117e-01
-1.88064575e-01 -1.94272146e-01 -6.19698882e-01 1.22881405e-01
1.51703268e-01 9.91812587e-01 -5.91401339e-01 -2.57996202e-01
-2.38963261e-01 8.90560448e-01 7.02013254e-01 7.06518114e-01
-1.21405852e+00 -2.85207182e-01 5.49959362e-01 -1.88149795e-01
6.11546576e-01 -5.71195841e-01 -1.10761849e-02 -1.42916977e+00
3.32082063e-01 -9.60223138e-01 4.01676297e-01 -1.01882732e+00
-1.15370047e+00 6.75006211e-01 1.19540937e-01 -1.53858221e+00
-1.72000334e-01 -9.65874135e-01 -9.92059708e-02 4.52200562e-01
-1.42114973e+00 -9.55145061e-01 -6.10178351e-01 8.41488123e-01
9.12165105e-01 -4.02694046e-01 7.40492225e-01 6.61997944e-02
-6.31809592e-01 6.93266034e-01 1.85456440e-01 -5.36586158e-02
6.84432030e-01 -1.10521078e+00 2.04699375e-02 6.34535372e-01
-1.10194989e-01 4.63795006e-01 6.27952218e-01 -3.08601558e-01
-1.51031506e+00 -7.49947548e-01 -1.70078516e-01 -4.76157963e-01
4.55979794e-01 -4.12673384e-01 -3.21919948e-01 9.51686561e-01
5.34579098e-01 1.42897874e-01 2.54773229e-01 -2.24726140e-01
-3.67335081e-01 -2.65754580e-01 -1.17297709e+00 9.04507816e-01
1.43806720e+00 -1.40402421e-01 -3.98898184e-01 1.00852132e-01
5.96437395e-01 -5.20035148e-01 -8.74405682e-01 4.77515757e-01
5.74024498e-01 -1.18063188e+00 1.03041971e+00 -6.80206716e-01
8.14832926e-01 -1.53445810e-01 -3.92799109e-01 -1.37827885e+00
-3.56993735e-01 -5.50992787e-01 -1.46397069e-01 6.34785831e-01
4.90501434e-01 -5.86825788e-01 7.49589086e-01 3.83490115e-01
-3.20957989e-01 -9.18128610e-01 -1.00971770e+00 -8.80148828e-01
2.41488740e-01 -3.53153884e-01 1.28078803e-01 8.32759023e-01
1.12489745e-01 -5.63476644e-02 -4.22989041e-01 -7.05540776e-02
5.55088699e-01 4.76924963e-02 7.51442909e-01 -6.46573126e-01
-4.74865824e-01 -5.17764032e-01 -1.54813409e-01 -1.29827154e+00
-4.30221483e-02 -5.94082057e-01 3.58754545e-02 -1.40303135e+00
1.95338398e-01 -4.93634075e-01 -5.22866488e-01 5.89448810e-01
5.20421043e-02 1.16510160e-01 6.39410377e-01 1.84584841e-01
-9.90871489e-01 8.75102758e-01 1.57393026e+00 -1.99795350e-01
1.47644699e-01 -1.53065428e-01 -5.71049631e-01 8.33794713e-01
9.52581644e-01 -2.17304438e-01 -7.29348540e-01 -1.05482459e+00
2.59982646e-01 2.07679108e-01 5.66658676e-01 -1.07171106e+00
7.10130557e-02 -5.76857150e-01 5.14733911e-01 -1.10331461e-01
7.17978299e-01 -7.90043652e-01 -3.44603270e-01 4.39286530e-01
-2.40071312e-01 5.38909614e-01 4.81871694e-01 7.13228405e-01
1.07398786e-01 -1.10192344e-01 8.96063924e-01 -5.93295038e-01
-9.16889489e-01 2.45668516e-01 -2.70816505e-01 2.36383587e-01
1.56982255e+00 -2.68084675e-01 -7.44124472e-01 -4.95929599e-01
-8.76685500e-01 5.49700558e-01 4.49648291e-01 5.48990011e-01
8.16095769e-01 -1.07200027e+00 -2.55703390e-01 1.10705160e-01
1.88416377e-01 8.75787288e-02 5.65571845e-01 6.03983343e-01
-3.49680960e-01 3.55529666e-01 -7.35008955e-01 -4.72708881e-01
-7.83161581e-01 8.69487405e-01 3.35548520e-01 -1.27541376e-02
-4.27537441e-01 1.19746292e+00 4.15911108e-01 -3.34667593e-01
3.72692138e-01 -2.08104476e-01 -3.47873680e-02 -4.08169776e-01
2.46834040e-01 6.43576801e-01 -2.84472376e-01 -3.71379197e-01
-1.97778374e-01 6.18194103e-01 -1.84787944e-01 -3.45682561e-01
1.40465343e+00 -1.01819672e-01 4.47327346e-01 2.51354545e-01
8.26987922e-01 -5.00094533e-01 -2.10048652e+00 6.02368824e-02
-4.28269148e-01 -4.45209712e-01 -4.91004050e-01 -1.13662910e+00
-1.17514598e+00 1.05889201e+00 8.49065006e-01 2.88760886e-02
1.14649475e+00 1.29577190e-01 4.74751115e-01 5.33877909e-01
8.57005417e-01 -1.37381983e+00 6.55489922e-01 4.45499063e-01
1.28702843e+00 -1.44784987e+00 -6.32847622e-02 5.73844798e-02
-8.86444807e-01 7.77986646e-01 1.21374464e+00 -5.15749633e-01
2.25773305e-01 -1.25391230e-01 7.78143331e-02 -1.90898389e-01
-6.49596214e-01 -1.57531351e-01 -1.77323610e-01 7.04679847e-01
-1.12500206e-01 -5.06180599e-02 -1.50668874e-01 4.03123647e-01
-1.59523979e-01 -1.49511427e-01 4.01250571e-01 1.20477438e+00
-4.03800577e-01 -7.09424078e-01 6.71629310e-02 2.78226584e-01
-3.46844077e-01 2.29859158e-01 -3.29863057e-02 1.05031574e+00
1.88316524e-01 9.38425183e-01 -2.40667760e-01 -5.16747415e-01
5.04410803e-01 -6.45797029e-02 7.11884141e-01 -3.36319029e-01
-4.41686243e-01 7.83753023e-02 -1.80233762e-01 -8.40947449e-01
-3.14839512e-01 -5.83054543e-01 -1.41283190e+00 -1.25876039e-01
1.83416024e-01 -2.30826482e-01 4.76920724e-01 6.73428416e-01
4.01620954e-01 5.13540804e-01 6.14094734e-01 -1.17839170e+00
-8.09995115e-01 -8.12636137e-01 -2.18762457e-01 5.89799762e-01
1.42683208e-01 -1.07480419e+00 -1.94106221e-01 1.00269608e-01] | [4.418013572692871, 0.8207255005836487] |
0d264bef-ca07-4c66-980b-f6254dfe5555 | deep-cg2real-synthetic-to-real-translation-1 | 2003.12649 | null | https://arxiv.org/abs/2003.12649v1 | https://arxiv.org/pdf/2003.12649v1.pdf | Deep CG2Real: Synthetic-to-Real Translation via Image Disentanglement | We present a method to improve the visual realism of low-quality, synthetic images, e.g. OpenGL renderings. Training an unpaired synthetic-to-real translation network in image space is severely under-constrained and produces visible artifacts. Instead, we propose a semi-supervised approach that operates on the disentangled shading and albedo layers of the image. Our two-stage pipeline first learns to predict accurate shading in a supervised fashion using physically-based renderings as targets, and further increases the realism of the textures and shading with an improved CycleGAN network. Extensive evaluations on the SUNCG indoor scene dataset demonstrate that our approach yields more realistic images compared to other state-of-the-art approaches. Furthermore, networks trained on our generated "real" images predict more accurate depth and normals than domain adaptation approaches, suggesting that improving the visual realism of the images can be more effective than imposing task-specific losses. | ['Vladimir Kim', 'Ravi Ramamoorthi', 'Kalyan Sunkavalli', 'Sai Bi', 'Eli Shechtman', 'Federico Perazzi'] | 2020-03-27 | deep-cg2real-synthetic-to-real-translation | http://openaccess.thecvf.com/content_ICCV_2019/html/Bi_Deep_CG2Real_Synthetic-to-Real_Translation_via_Image_Disentanglement_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Bi_Deep_CG2Real_Synthetic-to-Real_Translation_via_Image_Disentanglement_ICCV_2019_paper.pdf | iccv-2019-10 | ['synthetic-to-real-translation'] | ['computer-vision'] | [ 7.77261615e-01 5.75559378e-01 3.60501260e-01 -6.87883139e-01
-7.59514272e-01 -5.28225362e-01 6.64845347e-01 -8.83674800e-01
1.25644296e-01 8.15125644e-01 1.96715921e-01 -1.78857058e-01
5.08733094e-01 -8.16434860e-01 -1.15647304e+00 -6.23689771e-01
2.03190014e-01 3.47118914e-01 -3.68900411e-02 -3.64575952e-01
1.44498304e-01 4.00336832e-01 -1.66822541e+00 5.23604095e-01
9.81546640e-01 7.53545642e-01 1.54035687e-01 1.01536334e+00
1.31743222e-01 1.09424186e+00 -5.58726966e-01 -2.43798181e-01
7.58282959e-01 -6.76357448e-01 -5.65668106e-01 4.19974446e-01
9.61461782e-01 -8.98926318e-01 -3.28465551e-01 8.15297484e-01
2.20125794e-01 8.78457725e-02 4.03291076e-01 -1.04494464e+00
-7.43657470e-01 1.11511670e-01 -5.67714989e-01 -5.37614048e-01
4.80506182e-01 6.20118082e-01 6.91018224e-01 -7.38533318e-01
6.90631092e-01 1.28582430e+00 5.39301753e-01 6.46209121e-01
-1.72390461e+00 -8.57876420e-01 4.02685348e-03 -3.75980347e-01
-1.13740647e+00 -6.98062778e-01 8.72187912e-01 -4.29638565e-01
5.28297424e-01 3.93123746e-01 6.61009252e-01 1.31263888e+00
1.88651364e-02 3.28718036e-01 1.57222486e+00 -3.80172580e-01
1.11663878e-01 1.58134356e-01 -9.22543645e-01 8.81153464e-01
8.26383159e-02 5.86183965e-01 -8.06959629e-01 7.69983083e-02
1.29929852e+00 -4.74437565e-01 -5.49790740e-01 -7.45221734e-01
-1.25512290e+00 5.34740627e-01 6.46767855e-01 -3.19508314e-01
-1.49835125e-01 5.32181680e-01 -1.81965396e-01 1.87117025e-01
7.71281183e-01 9.07574475e-01 -4.23133135e-01 1.41489059e-01
-1.10618174e+00 1.35227963e-01 6.40910506e-01 8.02897215e-01
1.03466189e+00 4.84484434e-01 1.43821493e-01 4.38604206e-01
2.57085353e-01 7.39440620e-01 -1.21068977e-01 -1.74934304e+00
3.09134513e-01 1.75851941e-01 4.03203934e-01 -7.28859127e-01
1.01617433e-01 -3.36992592e-01 -5.93189895e-01 1.14151490e+00
4.95335877e-01 -1.88605011e-01 -1.14789510e+00 1.75947189e+00
2.49248222e-01 3.49424690e-01 1.16687365e-01 1.02479815e+00
6.09700263e-01 6.31746829e-01 -2.87444741e-01 2.88555712e-01
6.24086082e-01 -1.15139687e+00 -4.35346425e-01 -5.09615541e-01
1.66178688e-01 -9.88847673e-01 1.50468850e+00 5.18416286e-01
-1.26564360e+00 -5.12215614e-01 -1.14490402e+00 -3.37252259e-01
1.25442371e-01 1.82810530e-01 8.07861328e-01 7.10168064e-01
-1.32634604e+00 8.34215164e-01 -8.95133078e-01 1.56862382e-02
5.53657234e-01 9.83484928e-03 -2.23777086e-01 -1.75701335e-01
-7.02525914e-01 6.86461091e-01 5.77351302e-02 -5.56572117e-02
-1.20835054e+00 -1.15478408e+00 -1.07852113e+00 -1.23543590e-01
1.10839605e-01 -8.63153219e-01 9.94339943e-01 -1.52755928e+00
-2.16820288e+00 1.00846756e+00 2.46544443e-02 -1.23283751e-02
9.02708828e-01 -1.17966108e-01 6.28836453e-03 3.02880049e-01
-1.23301737e-01 1.00591946e+00 1.07583928e+00 -2.10923839e+00
-2.47581482e-01 6.26951084e-02 2.82254487e-01 4.92677033e-01
8.71696472e-02 -3.86052430e-01 -1.28046066e-01 -4.03371722e-01
9.97278914e-02 -9.37486768e-01 -3.19164634e-01 7.32095361e-01
-4.91648346e-01 8.50654066e-01 6.67293191e-01 -8.92434359e-01
1.26811504e-01 -1.98695052e+00 4.32510488e-02 9.89287347e-02
2.42518321e-01 -4.51462045e-02 -3.33048642e-01 3.63148451e-02
-2.23601729e-01 -1.53127573e-02 -3.54688048e-01 -7.53564596e-01
-2.22496077e-01 3.11802506e-01 -4.85982150e-01 3.98070157e-01
2.75310755e-01 8.44499767e-01 -9.66743708e-01 -2.21494392e-01
5.86134315e-01 9.13850427e-01 -6.35892332e-01 4.99747843e-01
-5.45294881e-01 1.33184552e+00 7.48397261e-02 1.92587838e-01
8.61956000e-01 -3.59777510e-01 2.66748697e-01 -1.41337395e-01
-3.08801383e-02 5.49519181e-01 -7.37443805e-01 2.12540436e+00
-9.32761431e-01 1.02660966e+00 -3.93939279e-02 -3.10298830e-01
7.19308794e-01 8.88037607e-02 2.56451428e-01 -8.50271761e-01
-2.02932820e-01 8.63698646e-02 -3.57927173e-01 -2.98587710e-01
3.00372630e-01 -1.80555880e-01 4.45217967e-01 3.42569679e-01
-2.32944891e-01 -9.57861006e-01 -3.42866182e-01 8.04850310e-02
8.74445558e-01 1.11646795e+00 -2.49026924e-01 -7.03925192e-02
7.48211592e-02 -2.23609470e-02 3.54907185e-01 4.84999955e-01
3.80405217e-01 1.29153752e+00 3.29507560e-01 -4.15559709e-01
-1.57478857e+00 -1.38832819e+00 5.26376255e-02 7.17147648e-01
3.33486423e-02 2.12884936e-02 -7.73639619e-01 -3.51498246e-01
-1.51413605e-01 1.04012167e+00 -6.99482501e-01 -3.04538645e-02
-5.97817540e-01 -3.83467317e-01 4.42223251e-01 2.50082493e-01
7.51734674e-01 -7.45403707e-01 -7.69837558e-01 -2.77294487e-01
-2.51255751e-01 -1.27899718e+00 -1.65175915e-01 -1.13274425e-01
-6.66355550e-01 -8.72603714e-01 -6.46285117e-01 -3.83883506e-01
1.01360905e+00 2.82477498e-01 1.52265298e+00 1.14269756e-01
-1.50684670e-01 4.78822701e-02 -8.25627297e-02 -2.05075249e-01
-6.97334230e-01 -3.67111444e-01 -3.75513256e-01 9.59265530e-02
-6.22693002e-01 -8.91417265e-01 -9.72036421e-01 2.60626018e-01
-8.85948658e-01 8.45695674e-01 2.99257785e-01 7.48354375e-01
3.94064724e-01 -3.22017312e-01 -2.29795381e-01 -1.15910161e+00
-3.05950064e-02 1.93082541e-01 -8.62044752e-01 7.26594776e-02
-6.48099840e-01 2.23962709e-01 7.13544130e-01 -3.00900012e-01
-1.77044773e+00 1.97474614e-01 1.05624452e-01 -5.24099231e-01
-1.50322959e-01 -2.13219866e-01 -1.33806765e-01 -6.57614529e-01
1.03290844e+00 5.48248403e-02 -5.69125917e-03 -1.98967829e-01
4.55752313e-01 1.46655634e-01 6.66778326e-01 -7.07368016e-01
1.31992424e+00 9.44943070e-01 2.60270298e-01 -7.12233067e-01
-9.03739393e-01 2.06380561e-01 -3.82928014e-01 -2.01971158e-01
7.89798081e-01 -1.21454382e+00 -6.97010577e-01 5.24984241e-01
-1.20311689e+00 -1.21615124e+00 -5.73956370e-01 3.15212905e-01
-8.08954120e-01 1.66828230e-01 -5.02721369e-01 -6.28384233e-01
1.17897719e-01 -1.03075194e+00 1.48765981e+00 6.80007488e-02
-2.49979664e-02 -9.07963097e-01 -6.66038543e-02 6.47809565e-01
4.92262512e-01 7.85650730e-01 7.09552646e-01 5.21979451e-01
-1.18002188e+00 3.45048875e-01 -5.19885659e-01 5.35284400e-01
1.53409317e-01 1.40972093e-01 -1.48688674e+00 -2.12241501e-01
-5.46265393e-02 -6.63455307e-01 7.13500023e-01 2.40769655e-01
1.27749670e+00 -3.34130228e-01 1.42402530e-01 1.32540894e+00
1.59281337e+00 -1.60106838e-01 1.04175413e+00 9.08099338e-02
1.15375543e+00 7.86575973e-01 3.14420104e-01 2.25472778e-01
4.22793657e-01 4.56815928e-01 5.71645260e-01 -8.12609375e-01
-7.56194890e-01 -5.37597001e-01 3.18108380e-01 2.67035067e-01
-1.10695437e-01 -2.86530703e-01 -7.33330965e-01 1.89271167e-01
-1.48432446e+00 -6.82656467e-01 -2.54363984e-01 2.26526999e+00
1.02452934e+00 -4.28444520e-02 -5.28970718e-01 -1.90675229e-01
2.63958931e-01 3.91741425e-01 -6.71803653e-01 -2.82308131e-01
-4.41734195e-01 4.80902344e-01 6.72603071e-01 8.76676381e-01
-4.91737455e-01 1.14218688e+00 6.82948685e+00 3.24998468e-01
-1.34656310e+00 1.05228484e-01 1.02761710e+00 -3.50741863e-01
-8.94118488e-01 2.41607621e-01 -6.38873223e-03 2.48300046e-01
6.58568144e-01 4.29679066e-01 1.03816080e+00 6.37140155e-01
3.88171107e-01 -3.32225889e-01 -1.21911705e+00 9.20948565e-01
2.56151736e-01 -1.46885526e+00 1.73013676e-02 1.68458596e-01
1.32200217e+00 -5.40907215e-03 4.65576142e-01 -2.75381923e-01
8.13712001e-01 -1.27204239e+00 8.37871492e-01 6.47728503e-01
1.32800376e+00 -4.08950239e-01 1.18987113e-01 1.66947752e-01
-5.28638601e-01 5.40225923e-01 -2.17112973e-01 -2.06479564e-01
2.22469345e-01 6.08962715e-01 -7.20044911e-01 3.33198279e-01
6.72258139e-01 4.96694237e-01 -6.57675147e-01 4.37573165e-01
-8.52581799e-01 5.53968728e-01 -3.14947963e-01 5.44055164e-01
-9.17723775e-02 -3.86084467e-01 1.77341655e-01 7.22365201e-01
1.35865465e-01 5.49708121e-02 -3.89943868e-02 1.35597193e+00
-1.36431620e-01 -3.17366958e-01 -7.51669765e-01 2.88049579e-01
-9.94731784e-02 1.07994831e+00 -4.12696689e-01 -3.31968009e-01
-1.26877382e-01 1.38502932e+00 3.05655420e-01 7.90544510e-01
-7.59260237e-01 -4.03540023e-02 7.97218025e-01 2.66356826e-01
6.32420480e-02 -2.42856398e-01 -8.71975362e-01 -1.47331703e+00
9.09447670e-03 -9.91205871e-01 -4.77681011e-01 -1.66712773e+00
-8.47352207e-01 5.81324160e-01 -3.25792313e-01 -1.13712716e+00
-4.09427524e-01 -5.07109582e-01 -5.28175414e-01 1.14160752e+00
-1.85358953e+00 -1.46422982e+00 -8.70983958e-01 4.42568481e-01
3.13775182e-01 3.10694754e-01 9.52730656e-01 -1.19762622e-01
-8.19217786e-03 2.99880743e-01 8.85244682e-02 -1.05795249e-01
8.64351630e-01 -1.30834019e+00 7.31938720e-01 1.06935358e+00
9.39274356e-02 2.52960593e-01 9.99509931e-01 -4.47205544e-01
-1.19738507e+00 -9.89482105e-01 3.15187782e-01 -5.71778417e-01
1.24698319e-01 -6.05215907e-01 -6.40483499e-01 7.82324076e-01
5.32041669e-01 1.70766637e-01 3.48062247e-01 -3.07900190e-01
-7.14181125e-01 -1.79278567e-01 -1.28762102e+00 9.40339744e-01
1.27370954e+00 -6.82049692e-01 8.19640532e-02 5.82297802e-01
8.51531565e-01 -8.17091286e-01 -4.77227718e-01 2.20063016e-01
6.41593516e-01 -1.55931079e+00 1.18881941e+00 -3.18172127e-01
1.10506654e+00 -4.40281510e-01 -1.68702275e-01 -1.65820599e+00
-4.19140533e-02 -8.76777232e-01 2.17000455e-01 8.36298227e-01
4.20235515e-01 -5.78371584e-01 1.04277718e+00 7.98147976e-01
-6.42602146e-02 -2.48526528e-01 -3.78072232e-01 -4.55981433e-01
1.07236234e-02 -2.94086814e-01 6.41449749e-01 1.01105368e+00
-6.82453990e-01 9.92771164e-02 -6.06333792e-01 3.33904415e-01
1.00890410e+00 1.92339599e-01 1.23888290e+00 -8.71487916e-01
-6.23673141e-01 -5.46865351e-02 -1.00299090e-01 -1.06180024e+00
2.55869865e-01 -5.23398519e-01 2.65656590e-01 -1.30872476e+00
-1.04231395e-01 -5.86364806e-01 3.58723521e-01 3.50420743e-01
6.94550052e-02 8.33849370e-01 8.33560377e-02 5.83620109e-02
-1.17666066e-01 6.27089798e-01 1.74176109e+00 4.51736003e-02
-1.24220438e-02 -3.71490628e-01 -5.63396633e-01 7.55722046e-01
6.41686380e-01 -2.64733970e-01 -6.34631217e-01 -9.78905439e-01
4.05997545e-01 7.55317360e-02 7.54452348e-01 -1.00231719e+00
-2.16924042e-01 -3.82027864e-01 8.43639016e-01 -1.14623159e-01
7.67560124e-01 -6.68977380e-01 4.47287112e-01 2.59302795e-01
-4.32543188e-01 -4.58562255e-01 1.54947760e-02 2.86047161e-01
1.82727456e-01 3.00022960e-01 1.05220580e+00 -3.31781983e-01
-2.70161510e-01 1.52669981e-01 1.71276420e-01 2.81555168e-02
6.15627944e-01 -3.39712352e-01 -3.38862062e-01 -8.25345993e-01
-2.14904904e-01 -2.34682783e-01 1.18457496e+00 1.66742966e-01
6.63633704e-01 -1.12040162e+00 -6.76354229e-01 5.05829692e-01
1.47728413e-01 4.97467041e-01 8.26089010e-02 2.19479650e-01
-1.17762697e+00 -1.53853461e-01 -2.28594452e-01 -6.71932518e-01
-9.98279035e-01 1.84996918e-01 5.73151827e-01 -3.61618176e-02
-5.87149978e-01 8.95446122e-01 7.32920945e-01 -6.41303480e-01
-9.66961905e-02 -3.18386674e-01 6.13534093e-01 -8.78193676e-01
2.90496916e-01 4.79279868e-02 -1.29494160e-01 -5.04087150e-01
6.76266626e-02 5.52864313e-01 4.34317350e-01 -5.63512564e-01
1.40582311e+00 -2.13342130e-01 -3.51517759e-02 3.32634896e-01
1.09757721e+00 1.98047251e-01 -2.23242474e+00 1.40343942e-02
-6.88268721e-01 -1.08796108e+00 3.04638594e-01 -1.10270905e+00
-1.20543420e+00 9.14203703e-01 4.50736284e-01 -3.79889876e-01
1.10104322e+00 -4.08708304e-01 6.86194837e-01 2.14541584e-01
4.51595694e-01 -8.13201606e-01 2.51775533e-01 1.51563242e-01
9.86510515e-01 -1.48061848e+00 6.58861548e-02 -8.31687689e-01
-6.48101687e-01 9.74338472e-01 7.03319907e-01 -2.61277080e-01
2.58481860e-01 4.47188854e-01 4.32648093e-01 1.10287564e-02
-5.07134914e-01 6.29401579e-02 2.99553633e-01 8.62392366e-01
3.51308107e-01 -1.56712197e-02 3.87462527e-01 -4.70103621e-01
-6.04395568e-01 -1.41197983e-02 7.55362034e-01 5.72535098e-01
3.84709723e-02 -9.55738902e-01 -4.98159140e-01 6.85968101e-02
-2.45973729e-02 -2.80973613e-01 -3.56063515e-01 5.33160150e-01
1.62512690e-01 9.86702204e-01 5.62213697e-02 -2.38748282e-01
8.99817720e-02 -2.70464092e-01 8.94952953e-01 -6.76808059e-01
-2.94370949e-01 -9.33381543e-02 2.30074510e-01 -1.01965857e+00
-4.66447294e-01 -4.06038374e-01 -9.93195236e-01 -4.60881889e-01
4.19282019e-02 -4.66654032e-01 9.84196723e-01 6.92239404e-01
3.08562487e-01 6.48995399e-01 9.04687166e-01 -1.35054243e+00
-1.70545474e-01 -5.59536874e-01 -3.21733445e-01 7.50581384e-01
4.96104658e-01 -3.86819869e-01 -5.79311311e-01 3.23666424e-01] | [9.558294296264648, -3.102858304977417] |
c28142b6-49e7-4b2c-a0e3-b5263b703d2c | word-embedding-neural-networks-to-advance | 2212.11933 | null | https://arxiv.org/abs/2212.11933v1 | https://arxiv.org/pdf/2212.11933v1.pdf | Word Embedding Neural Networks to Advance Knee Osteoarthritis Research | Osteoarthritis (OA) is the most prevalent chronic joint disease worldwide, where knee OA takes more than 80% of commonly affected joints. Knee OA is not a curable disease yet, and it affects large columns of patients, making it costly to patients and healthcare systems. Etiology, diagnosis, and treatment of knee OA might be argued by variability in its clinical and physical manifestations. Although knee OA carries a list of well-known terminology aiming to standardize the nomenclature of the diagnosis, prognosis, treatment, and clinical outcomes of the chronic joint disease, in practice there is a wide range of terminology associated with knee OA across different data sources, including but not limited to biomedical literature, clinical notes, healthcare literacy, and health-related social media. Among these data sources, the scientific articles published in the biomedical literature usually make a principled pipeline to study disease. Rapid yet, accurate text mining on large-scale scientific literature may discover novel knowledge and terminology to better understand knee OA and to improve the quality of knee OA diagnosis, prevention, and treatment. The present works aim to utilize artificial neural network strategies to automatically extract vocabularies associated with knee OA diseases. Our finding indicates the feasibility of developing word embedding neural networks for autonomous keyword extraction and abstraction of knee OA. | ['Ahmad P. Tafti', 'Johannes F. Plate', 'Hamid R. Arabnia', 'Hilal Maradit Kremers', 'Mehdi Assefi', 'Husam Ghazaleh', 'Soheyla Amirian'] | 2022-12-22 | null | null | null | null | ['keyword-extraction'] | ['natural-language-processing'] | [-2.78317332e-01 -1.10371493e-01 -8.03241313e-01 4.22857642e-01
-5.42806387e-01 -1.24741927e-01 4.20708358e-02 6.93885565e-01
-6.25724196e-01 8.07843626e-01 8.69942009e-01 -1.51758850e-01
-4.00194854e-01 -6.47136569e-01 3.97524349e-02 -4.54641938e-01
-1.82452992e-01 7.30864167e-01 -2.95106843e-02 -1.59956038e-01
3.04725897e-02 5.48488081e-01 -1.21495605e+00 1.05068855e-01
7.75500596e-01 9.05778766e-01 2.43706331e-01 2.22689092e-01
-2.32464582e-01 3.98140341e-01 -7.01123357e-01 -5.26792780e-02
-2.03630432e-01 -1.19672902e-01 -5.62203228e-01 -2.31208056e-01
-1.32986620e-01 -2.35567927e-01 -5.29323399e-01 7.85278738e-01
8.00114870e-01 -5.65941632e-01 7.43851125e-01 -8.05935085e-01
-9.90729868e-01 1.37377501e-01 -2.22630039e-01 7.92064667e-02
2.75006801e-01 2.33095393e-01 1.02402961e+00 -8.53632033e-01
8.98551822e-01 8.40569079e-01 4.10486251e-01 5.41833639e-01
-8.42163086e-01 -4.91691798e-01 -3.15785855e-01 2.22454473e-01
-1.31367743e+00 -1.81999691e-02 4.81357843e-01 -6.82342887e-01
6.55740142e-01 1.99219510e-01 1.41292870e+00 1.26479197e+00
1.01385546e+00 2.90744394e-01 9.63320911e-01 -3.47268879e-01
2.69363165e-01 -1.68529496e-01 2.43805394e-01 4.47257519e-01
9.23989713e-01 -2.64207989e-01 -6.58984303e-01 -4.49197531e-01
4.77817595e-01 4.84348238e-01 -4.07744139e-01 -2.80132234e-01
-1.52405822e+00 8.44539523e-01 2.84947693e-01 3.21628660e-01
-4.96214628e-01 -3.93228531e-02 8.03322911e-01 4.29134965e-02
1.75194457e-01 7.65660703e-01 -4.85310018e-01 -1.43928751e-01
-7.10289061e-01 3.84246886e-01 7.95413911e-01 2.86961496e-01
5.75733557e-02 -1.45130321e-01 -4.27487642e-02 1.13459206e+00
5.58672309e-01 7.09465146e-01 1.20375657e+00 -6.62090540e-01
-5.70212305e-03 1.13530123e+00 -1.92598671e-01 -1.04339802e+00
-3.94888610e-01 -4.47802246e-01 -8.52999985e-01 -4.45647798e-02
6.67398497e-02 1.90817282e-01 -9.83111143e-01 1.38756478e+00
2.30657741e-01 -3.72031093e-01 -1.78242579e-01 1.03327858e+00
7.74126172e-01 -2.13597771e-02 3.52588207e-01 -2.01000139e-01
2.05209947e+00 -6.99160218e-01 -1.08005714e+00 -4.54852909e-01
7.03861654e-01 -6.85065329e-01 9.25846398e-01 2.30147019e-01
-8.43971908e-01 8.43595788e-02 -1.13880122e+00 -3.93998414e-01
-6.04847014e-01 1.18630283e-01 7.18192458e-01 2.47673362e-01
-6.23105764e-01 9.37753767e-02 -7.33935833e-01 -6.39354825e-01
5.43742359e-01 2.19267324e-01 -6.03933692e-01 -2.54862189e-01
-1.50328255e+00 1.34227622e+00 3.06087524e-01 4.47672745e-03
-3.43573540e-01 -8.25197220e-01 -7.25816905e-01 -8.98846164e-02
5.65073490e-01 -1.32604373e+00 8.01531672e-01 -1.39176324e-01
-9.75196958e-01 8.08178842e-01 -2.11913079e-01 -3.54218304e-01
-1.80278510e-01 -7.76255727e-01 -5.60522974e-01 2.35556617e-01
3.19407344e-01 3.21834385e-01 4.26267117e-01 -4.51400399e-01
-2.64347136e-01 -6.87161505e-01 -5.40412843e-01 -6.16426878e-02
-6.39377117e-01 -9.05670971e-02 -3.07810754e-01 -9.17679012e-01
-1.26528889e-02 -9.50522840e-01 -4.42465365e-01 4.33985710e-01
-1.52635187e-01 -2.15816811e-01 6.64194107e-01 -8.70765507e-01
1.67293596e+00 -2.14619112e+00 4.07022566e-01 -9.37956572e-02
7.33276844e-01 2.24031150e-01 3.65478218e-01 8.03900778e-01
9.30195600e-02 4.84340698e-01 -2.69388454e-03 4.79654789e-01
-1.90258294e-01 3.26844811e-01 -1.34140581e-01 3.64206165e-01
1.60302788e-01 1.03676558e+00 -8.45470190e-01 -6.40066028e-01
2.15585366e-01 5.48571050e-01 -4.76721048e-01 -7.45006427e-02
-1.88913032e-01 -7.09546730e-02 -7.96587408e-01 9.87597287e-01
-2.92554110e-01 -7.27761924e-01 2.86212236e-01 -4.26353186e-01
2.85404801e-01 3.96979481e-01 -7.45769620e-01 1.65011370e+00
-3.31332505e-01 6.52997077e-01 -3.26433569e-01 -6.38507187e-01
4.25876230e-01 5.74110508e-01 7.68673003e-01 -4.29453015e-01
1.89891130e-01 5.49722612e-01 3.49296004e-01 -8.72347713e-01
2.23462701e-01 8.93617049e-03 -2.27841571e-01 1.07866481e-01
-1.16981022e-01 1.13332726e-01 3.13076884e-01 1.33083567e-01
1.17052031e+00 -4.09370214e-01 7.11294293e-01 -2.23074984e-02
3.27353686e-01 2.99652129e-01 4.48071331e-01 3.36194843e-01
-2.13390410e-01 4.89840984e-01 3.03461045e-01 -5.36081672e-01
-1.00017762e+00 -1.34332979e+00 -2.81240374e-01 1.38205543e-01
-1.44544825e-01 -6.59812689e-01 -1.65208623e-01 8.94254297e-02
3.48394930e-01 1.10031687e-01 -5.00306606e-01 -4.61731315e-01
-3.53011966e-01 -7.05511868e-01 3.98542136e-01 2.32545286e-01
2.88754970e-01 -8.21628392e-01 -4.90513444e-01 4.93185043e-01
-9.75495502e-02 -8.72864127e-01 -3.73038322e-01 2.42217537e-02
-1.05044925e+00 -1.46406770e+00 -1.05723417e+00 -9.39153731e-01
2.57208914e-01 -1.44732982e-01 7.26640284e-01 9.58131328e-02
-1.00587678e+00 3.99324894e-01 1.77818611e-02 -3.79520029e-01
-3.37138146e-01 2.38477185e-01 3.11549515e-01 -5.70598483e-01
6.56858563e-01 -2.49498263e-01 -1.01257062e+00 -1.81720629e-01
-9.34599638e-01 -1.25947833e-01 1.21637070e+00 7.90212512e-01
7.49804974e-01 -2.20623508e-01 5.94813645e-01 -1.92738354e-01
1.07041204e+00 -7.86019981e-01 2.30074883e-01 2.05667987e-01
-1.13215554e+00 8.81285779e-03 -5.72120436e-02 -4.98674065e-01
-1.91942245e-01 -5.76066554e-01 9.91611108e-02 -5.07728904e-02
-1.81390032e-01 1.14184475e+00 1.46903172e-01 4.14790332e-01
3.98554325e-01 1.50704056e-01 3.67228448e-01 -5.43972194e-01
8.14076439e-02 9.94742453e-01 1.74631566e-01 -2.25712955e-01
2.06978902e-01 3.97236645e-01 1.73986070e-02 -1.04400885e+00
-7.72281587e-01 -6.94599926e-01 3.51436734e-02 -2.16741338e-01
1.32618940e+00 -8.76833320e-01 -6.45969450e-01 1.06437191e-01
-8.51273179e-01 3.23104441e-01 -2.03137308e-01 9.32906091e-01
-8.50342959e-02 3.15101177e-01 -9.32490885e-01 -2.13610858e-01
-6.53708935e-01 -1.31027281e+00 1.07561374e+00 -1.71810854e-02
-9.52229202e-01 -8.29166710e-01 3.78314614e-01 4.40527231e-01
3.92330676e-01 1.09603338e-01 1.66196847e+00 -5.80784142e-01
-2.31297702e-01 -3.29930127e-01 -1.02811471e-01 2.10383892e-01
7.42578685e-01 -3.47708493e-01 -7.63522694e-03 9.21204612e-02
-1.44242523e-02 -2.30169386e-01 6.16673410e-01 4.62288707e-01
8.43460679e-01 -1.73329771e-01 -5.57567835e-01 3.65933497e-03
1.30001295e+00 3.10781986e-01 5.24411023e-01 8.43822718e-01
5.27265906e-01 5.27658105e-01 3.36126596e-01 1.07957937e-01
1.55489683e-01 6.47039115e-01 2.40587816e-01 2.23393887e-01
-2.86515415e-01 1.04105756e-01 8.59658644e-02 1.26644742e+00
6.27462938e-02 -7.12140650e-02 -1.27043509e+00 8.79351258e-01
-1.33868742e+00 -6.38207734e-01 -1.57991782e-01 1.99856508e+00
1.22354078e+00 1.49342924e-01 -2.82203525e-01 -6.10295869e-02
2.64132410e-01 -6.85501397e-02 -3.94605249e-01 -1.64288104e-01
-1.55756483e-02 3.83208513e-01 3.76089185e-01 -2.21573412e-01
-4.62778986e-01 4.45112675e-01 6.43835163e+00 6.32010758e-01
-1.38335145e+00 -4.88173924e-02 1.56350862e-02 -3.51597399e-01
-1.71672165e-01 -1.51423499e-01 -5.24704039e-01 5.39156854e-01
8.70000005e-01 -4.73634928e-01 1.51309557e-02 8.15624535e-01
5.02555490e-01 -6.50687367e-02 -8.09605658e-01 6.66546285e-01
-1.47333637e-01 -1.51152658e+00 3.81326586e-01 2.89166898e-01
2.46228874e-01 9.61038917e-02 5.29400408e-02 -7.46545866e-02
-8.10850635e-02 -1.12787879e+00 1.29614815e-01 6.00667357e-01
7.16509223e-01 -1.30126476e-01 9.68744040e-01 -3.38352591e-01
-7.63842821e-01 1.12099886e-01 1.52926832e-01 -4.24299855e-03
1.39012069e-01 8.76492977e-01 -6.98653042e-01 2.56596267e-01
9.41592991e-01 7.34934032e-01 -5.05243659e-01 1.23752999e+00
-1.09463342e-01 5.89410722e-01 -3.31996419e-02 -2.31801152e-01
-1.00037664e-01 -1.00915581e-01 5.51924765e-01 5.42881191e-01
4.47733939e-01 -1.30983487e-01 1.97914168e-01 4.14830983e-01
8.75697508e-02 4.02294695e-01 -5.97865403e-01 -9.98954654e-01
3.90157223e-01 6.65291250e-01 -5.49480140e-01 -1.38072297e-01
-5.87538600e-01 5.55608213e-01 -1.00037210e-01 3.82757872e-01
-2.84782857e-01 -2.02090979e-01 1.05273175e+00 4.75636423e-01
-7.90075213e-02 -3.22795570e-01 -1.47426412e-01 -9.71536458e-01
1.22232191e-01 -1.12355661e+00 4.25837517e-01 -7.22158432e-01
-1.42224407e+00 1.52318716e-01 -7.89691433e-02 -1.22233558e+00
-1.64937377e-01 -7.63144016e-01 2.18865320e-01 7.03349233e-01
-1.00846064e+00 -9.58595991e-01 4.01702151e-02 -1.88796576e-02
5.82980335e-01 -3.17127615e-01 1.12439799e+00 2.58077085e-01
-3.62085700e-01 -3.12271565e-01 2.19061241e-01 3.18135798e-01
1.13388991e+00 -9.69282150e-01 -3.67867649e-01 -2.37029165e-01
-4.33606833e-01 1.11140192e+00 7.39731610e-01 -1.14415324e+00
-1.56084180e+00 -6.87961519e-01 1.13722932e+00 -3.83225922e-03
1.17637658e+00 1.17419027e-01 -8.26575279e-01 3.45712066e-01
8.40587243e-02 -3.43736142e-01 1.30931735e+00 4.83137965e-02
-1.84278846e-01 7.68535882e-02 -5.45427680e-01 1.35675442e+00
5.74962318e-01 -4.86273050e-01 -1.21340811e+00 4.83566642e-01
7.67584383e-01 -1.34582162e-01 -1.47435236e+00 4.90636528e-01
1.01336217e+00 -2.02802047e-01 1.17923594e+00 -8.61296117e-01
8.18099082e-01 -1.65059984e-01 5.81119321e-02 -1.01080441e+00
7.32771121e-03 1.92307428e-01 -2.34508023e-01 4.56473976e-01
1.93213925e-01 -7.51702487e-01 6.58893108e-01 3.42959106e-01
-1.58528477e-01 -1.30022967e+00 -1.01681030e+00 -5.37982047e-01
7.55795911e-02 -2.62812316e-01 1.86171278e-01 7.71664977e-01
-2.66393442e-02 2.08157152e-01 -1.10379770e-01 -1.16393417e-01
3.73333991e-01 -3.21454220e-02 2.69374430e-01 -1.50668931e+00
9.12996754e-03 -5.02979279e-01 -7.02547431e-01 -3.44293922e-01
-3.09766889e-01 -9.40347731e-01 -6.86073661e-01 -2.19003034e+00
1.60209998e-01 1.23534482e-02 -3.67738158e-01 2.48910770e-01
1.11455200e-02 1.19759344e-01 -2.92485416e-01 4.43530679e-01
7.35639967e-03 5.51721752e-01 1.36378062e+00 -5.96221983e-01
-1.92032337e-01 -5.06985247e-01 -5.61380565e-01 7.01663435e-01
5.93626022e-01 -8.11903656e-01 -3.26290637e-01 -1.93331927e-01
6.17055535e-01 -9.22206938e-02 5.53276718e-01 -6.70823216e-01
3.15133072e-02 -4.19640571e-01 3.29920292e-01 -4.83623922e-01
1.68709397e-01 -9.31616664e-01 3.93775910e-01 1.16686416e+00
-2.36829951e-01 -6.36428744e-02 1.61476418e-01 6.01855516e-01
-4.64105666e-01 -2.02903584e-01 7.82682076e-02 -1.93327174e-01
-5.97154021e-01 1.77054331e-01 -9.70124185e-01 2.68016517e-01
8.19331944e-01 -2.35866874e-01 -2.81946838e-01 -1.37204215e-01
-1.03972566e+00 1.41701132e-01 3.86627436e-01 8.17691922e-01
7.66416669e-01 -1.39256120e+00 -4.59468037e-01 -1.76462293e-01
4.25121665e-01 -2.00692698e-01 2.70944893e-01 1.02899706e+00
-6.57144308e-01 6.89072609e-01 -1.34091154e-01 -4.51985538e-01
-1.15112865e+00 4.14031446e-01 1.07075460e-01 -4.89698708e-01
-7.93594122e-01 1.04477063e-01 -2.22462565e-01 9.00703818e-02
3.10406774e-01 -2.56603330e-01 -1.99639067e-01 6.61165565e-02
3.62473309e-01 1.71820626e-01 -9.47673991e-02 -1.89009055e-01
-4.19742107e-01 4.83604163e-01 -3.34764510e-01 -5.18109202e-02
1.39823282e+00 6.25770390e-02 -5.04967034e-01 8.17179441e-01
1.05999708e+00 2.66634673e-01 1.26228839e-01 7.20292404e-02
1.47996590e-01 -8.78317282e-02 4.56329249e-02 -8.10008585e-01
-8.68212581e-01 5.41106939e-01 8.57042372e-01 -1.71448827e-01
7.81794369e-01 3.56681794e-01 1.14788425e+00 4.97051716e-01
4.44389045e-01 -9.15753543e-01 7.11096376e-02 9.89910290e-02
1.11965597e+00 -1.00607443e+00 4.76300746e-01 -2.28345245e-01
-2.61657238e-01 1.19704294e+00 2.00079337e-01 2.90993694e-02
9.40433741e-01 1.17348485e-01 3.01917166e-01 -6.25435948e-01
-4.75429058e-01 6.59676641e-02 3.74170452e-01 2.52164125e-01
5.54873884e-01 -8.70630965e-02 -1.11113119e+00 8.08631480e-01
-1.10827209e-02 4.42963541e-01 4.18206751e-02 1.08651567e+00
-3.80259365e-01 -1.47962701e+00 -3.09308797e-01 1.04701161e+00
-8.20039868e-01 -3.05638611e-01 -5.27154088e-01 1.19128025e+00
5.69070466e-02 4.94450986e-01 -2.22585887e-01 4.17734385e-02
1.60758063e-01 3.50182146e-01 1.27867654e-01 -7.36249506e-01
-3.21323693e-01 -2.32593734e-02 1.15871534e-01 -3.24592203e-01
-3.63969088e-01 -4.20728564e-01 -1.30171895e+00 2.48146743e-01
-1.52024448e-01 1.25249118e-01 6.92802489e-01 1.12337327e+00
5.28932929e-01 7.66831517e-01 -2.44638786e-01 1.59100406e-02
-2.94242829e-01 -8.76052737e-01 -5.86010098e-01 2.49341488e-01
1.09618478e-01 -9.14673388e-01 -4.82674569e-01 -4.54355441e-02] | [14.559945106506348, -1.7222821712493896] |
424cde99-e253-4e24-abe5-9660ab7a26c0 | anlirika-an-lstm-cnn-flow-twister-for-spoken | null | null | https://aclanthology.org/2021.sigtyp-1.14 | https://aclanthology.org/2021.sigtyp-1.14.pdf | Anlirika: An LSTM–CNN Flow Twister for Spoken Language Identification | The paper presents Anlirika’s submission to SIGTYP 2021 Shared Task on Robust Spoken Language Identification. The task aims at building a robust system that generalizes well across different domains and speakers. The training data is limited to a single domain only with predominantly single speaker per language while the validation and test data samples are derived from diverse dataset and multiple speakers. We experiment with a neural system comprising a combination of dense, convolutional, and recurrent layers that are designed to perform better generalization and obtain speaker-invariant representations. We demonstrate that the task in its constrained form (without making use of external data or augmentation the train set with samples from the validation set) is still challenging. Our best system trained on the data augmented with validation samples achieves 29.9% accuracy on the test data. | ['Ekaterina Vylomova', 'Matthew Coleman', 'Siddharth Singh', 'Ritesh Kumar', 'Liam Whittle', 'Andreas Scherbakov'] | null | null | null | null | naacl-sigtyp-2021-6 | ['spoken-language-identification'] | ['speech'] | [-2.67144479e-02 6.31598234e-02 -1.44099174e-02 -8.86095047e-01
-1.24837017e+00 -8.13385308e-01 7.17792153e-01 -4.89138275e-01
-5.09423912e-01 6.60989165e-01 2.95740277e-01 -1.23310730e-01
5.13218522e-01 -5.54889068e-02 -4.35309350e-01 -3.25478971e-01
-1.25123531e-01 7.49972761e-01 1.49283861e-03 -5.45441449e-01
-2.70458877e-01 4.90828812e-01 -1.38880277e+00 3.03891480e-01
5.60315251e-01 9.66801465e-01 -2.51059562e-01 9.38300908e-01
-6.49968302e-03 4.78918731e-01 -1.07091928e+00 1.85242057e-01
3.50479066e-01 -5.20622671e-01 -8.92001152e-01 1.45762980e-01
6.82746053e-01 -3.26270193e-01 -4.13076967e-01 7.67427623e-01
9.54530895e-01 3.48833829e-01 5.18900335e-01 -1.20069909e+00
-7.98329711e-01 8.64982784e-01 -3.19279805e-02 1.89296082e-01
3.50344092e-01 5.42276613e-02 4.55248356e-01 -1.25707698e+00
3.31581920e-01 1.31815279e+00 4.67889220e-01 1.16183162e+00
-1.09665930e+00 -9.71393943e-01 2.08572626e-01 -3.35292578e-01
-1.81640506e+00 -1.31303275e+00 4.70250040e-01 -6.54478297e-02
1.08550978e+00 1.51709825e-01 -5.53861484e-02 1.53532732e+00
-8.33248317e-01 9.14316058e-01 7.68799722e-01 -4.36826319e-01
3.14435989e-01 7.43985772e-01 4.62442100e-01 2.99898446e-01
-3.36687714e-01 9.58559811e-02 -7.46534586e-01 -1.26829416e-01
2.57529676e-01 -7.09236741e-01 -5.75440288e-01 -5.97570948e-02
-9.30764198e-01 7.45330036e-01 1.23024285e-01 4.31474507e-01
-9.65664014e-02 -4.44292068e-01 6.77250087e-01 5.44388592e-01
4.85867113e-01 2.91290402e-01 -7.58678555e-01 4.46452573e-02
-1.23415112e+00 2.19557375e-01 1.28288817e+00 1.16914058e+00
4.74205405e-01 6.47487164e-01 9.19459239e-02 1.38298345e+00
1.75166920e-01 6.33110881e-01 1.07842004e+00 -5.71216762e-01
4.09610689e-01 3.18799675e-01 7.60767311e-02 -1.47936746e-01
-4.53821838e-01 -4.81091201e-01 -7.58022666e-01 7.51716420e-02
4.29002464e-01 -6.40594065e-01 -1.10892749e+00 1.99560666e+00
5.12398072e-02 1.32248849e-01 6.19118333e-01 8.02720666e-01
1.30244780e+00 7.82517791e-01 -2.97765918e-02 4.21275273e-02
8.33601773e-01 -8.65860939e-01 -5.07196784e-01 -3.99462283e-01
5.80191016e-01 -5.99514604e-01 9.66029167e-01 5.02306104e-01
-1.08371353e+00 -6.59187675e-01 -1.21301520e+00 6.36278018e-02
-7.47339070e-01 2.49995619e-01 7.08129033e-02 7.62155950e-01
-1.52910352e+00 -3.67776565e-02 -2.01712802e-01 -6.47753119e-01
-4.30972949e-02 5.04945815e-01 -6.48455322e-01 1.00434266e-01
-1.42205048e+00 9.94613647e-01 2.07022935e-01 -3.49274054e-02
-1.15035021e+00 -5.58732510e-01 -9.54858184e-01 -2.13909581e-01
4.15500030e-02 -5.49500659e-02 1.65494156e+00 -1.28554344e+00
-1.86646044e+00 1.13484836e+00 -1.23865589e-01 -6.03627622e-01
5.78557611e-01 -1.70300841e-01 -9.08678651e-01 -3.17981571e-01
-7.08475113e-02 6.00129008e-01 7.21504450e-01 -1.05219054e+00
-3.36397588e-01 -3.84276837e-01 -5.55219173e-01 1.63520351e-01
-4.92785156e-01 4.88062859e-01 -3.38835150e-01 -3.61357003e-01
-1.73532337e-01 -8.71150017e-01 2.54995525e-02 -5.38762569e-01
-5.76714396e-01 -4.14197862e-01 8.69592845e-01 -7.06876874e-01
9.31470692e-01 -2.25879526e+00 -4.25195768e-02 1.87842816e-01
-2.68375099e-01 5.65507829e-01 -3.96717459e-01 3.18751276e-01
-3.35125148e-01 2.35577263e-02 -1.87481120e-01 -6.06579304e-01
9.04771537e-02 -2.82285060e-03 -4.48890686e-01 3.56408983e-01
3.07304472e-01 3.63655329e-01 -4.66172069e-01 2.48924978e-02
-6.81072474e-02 4.12346661e-01 -1.86176330e-01 4.54357058e-01
-9.02646803e-04 4.19819593e-01 -4.69303019e-02 5.97864628e-01
7.21216381e-01 6.55978471e-02 -2.92131547e-02 1.94592267e-01
-1.97752461e-01 5.77927291e-01 -1.43806684e+00 1.63473058e+00
-4.90465492e-01 5.78168392e-01 7.30413616e-01 -9.27381337e-01
1.30315185e+00 5.59880972e-01 -4.08855155e-02 -4.92243201e-01
6.32191449e-02 3.21210563e-01 -5.43963611e-02 -1.16870143e-01
4.72978026e-01 9.02925525e-03 -3.58296573e-01 6.89245582e-01
4.61565495e-01 -2.31566995e-01 -1.23900019e-01 1.81276530e-01
7.02787519e-01 -3.30432624e-01 1.13358900e-01 -4.76575375e-01
7.71801353e-01 -2.17885584e-01 4.22629446e-01 8.32793415e-01
-4.26251113e-01 6.59244955e-01 -2.20495075e-01 -3.24259222e-01
-8.93676162e-01 -9.29949462e-01 -4.28247452e-01 1.64863944e+00
-4.66948837e-01 -5.71862869e-02 -7.62781322e-01 -5.73378384e-01
2.27474719e-02 7.73126006e-01 -4.47597831e-01 -2.90364206e-01
-3.91697943e-01 -4.91466135e-01 1.28844035e+00 6.62628055e-01
6.17718697e-01 -8.93405914e-01 7.11009204e-02 1.99211881e-01
4.07435670e-02 -1.05214047e+00 -8.16988766e-01 4.22470391e-01
-1.06720805e-01 -7.62687624e-01 -8.82993042e-01 -1.14759219e+00
1.62857220e-01 -8.85414630e-02 1.03821790e+00 -4.44677025e-01
-1.40070811e-01 5.31873286e-01 -1.49926275e-01 -5.77305615e-01
-7.95720160e-01 3.06918919e-01 5.91151357e-01 1.10078074e-01
7.48609662e-01 -2.28827037e-02 2.05525253e-02 2.99508840e-01
-5.02677262e-01 -4.29105282e-01 1.01392828e-01 9.58548844e-01
-7.11251646e-02 -5.39751649e-01 1.08450246e+00 -5.74727356e-01
9.11319852e-01 -4.51531976e-01 -3.38658422e-01 2.71191269e-01
-1.00390412e-01 -1.29626125e-01 4.74139780e-01 -6.88374221e-01
-8.44283223e-01 2.28562266e-01 -1.90471262e-01 -2.47283667e-01
-5.86476028e-01 3.24844837e-01 -3.16912889e-01 -2.31412537e-02
1.04113364e+00 6.63703680e-01 3.17792416e-01 -5.88978887e-01
2.67634273e-01 1.50269175e+00 7.99647033e-01 -3.95669937e-01
3.18786681e-01 -2.25547791e-01 -8.07180285e-01 -1.38943172e+00
-4.67807233e-01 -5.57257295e-01 -7.91247249e-01 -2.89840288e-02
3.73105824e-01 -1.31838572e+00 -4.06516552e-01 7.17898726e-01
-9.77490067e-01 -6.05820656e-01 -2.03706190e-01 5.24963260e-01
-3.09358180e-01 1.09908003e-02 -4.91311461e-01 -1.10191858e+00
-6.31521881e-01 -8.31375480e-01 1.03635550e+00 1.62947536e-01
-5.79301238e-01 -8.86356533e-01 2.33581379e-01 1.22466788e-01
7.87276387e-01 -2.81165540e-01 1.86954752e-01 -1.73640132e+00
1.02549177e-02 -5.14385462e-01 2.77118146e-01 1.00525546e+00
2.56171644e-01 1.08429596e-01 -1.61633706e+00 -4.24212843e-01
-2.05724984e-01 -9.43431079e-01 8.33656669e-01 1.19029030e-01
9.90224481e-01 -1.98770836e-01 6.70996904e-02 4.37909305e-01
6.06528401e-01 1.62407178e-02 -7.57561345e-03 -1.16142037e-03
3.30127865e-01 7.30250180e-01 -1.57094285e-01 4.27584261e-01
4.36913490e-01 6.64359927e-01 -2.93838888e-01 1.54707907e-02
-1.97402865e-01 -5.30116931e-02 6.70622587e-01 6.74359798e-01
5.29333055e-01 -2.68096924e-01 -1.15093732e+00 7.13760316e-01
-1.45228267e+00 -8.41321588e-01 3.10614526e-01 2.41171837e+00
1.09152246e+00 -2.31321812e-01 6.61541939e-01 2.09410995e-01
8.41303051e-01 -2.00748041e-01 -8.02519321e-01 -6.47907555e-01
-3.67254317e-01 2.26709932e-01 2.35114694e-01 6.98268771e-01
-1.35991406e+00 1.02219725e+00 7.58473349e+00 2.91124344e-01
-1.43673301e+00 -1.71188608e-01 5.74918866e-01 -1.99294537e-01
3.69094424e-02 -8.26070428e-01 -1.29200137e+00 2.13942647e-01
1.66642702e+00 -4.64532942e-01 4.12672520e-01 8.55553985e-01
8.23185034e-03 4.61330235e-01 -1.26925278e+00 9.03576255e-01
4.97435004e-01 -8.69550705e-01 -3.88514400e-02 -3.35896164e-01
5.37753046e-01 5.29220879e-01 3.30303580e-01 7.54546762e-01
6.98018134e-01 -1.35775912e+00 6.95582747e-01 2.74436891e-01
9.06185508e-01 -6.98543191e-01 6.14111423e-01 6.08775318e-01
-6.82036877e-01 8.36883485e-02 -5.81467152e-02 9.10815746e-02
-2.74948210e-01 -8.01550075e-02 -1.39616942e+00 1.23331606e-01
6.44007325e-01 6.53508723e-01 -3.83909523e-01 6.66156650e-01
2.43137881e-01 6.50249779e-01 -6.04808629e-01 -1.46273881e-01
2.68884331e-01 3.38993549e-01 3.75368178e-01 1.63160253e+00
1.03884839e-01 -1.94732308e-01 4.89794314e-01 6.14058495e-01
-2.96580255e-01 1.81348965e-01 -9.61338341e-01 -6.01109751e-02
6.07794106e-01 9.32139575e-01 1.57632843e-01 -5.45918882e-01
-4.17670101e-01 8.43953311e-01 4.30485100e-01 6.88605607e-01
-1.07331239e-01 -3.03461850e-01 8.25256824e-01 -3.38631094e-01
2.38754340e-02 -1.26995534e-01 -1.30586818e-01 -1.19421983e+00
-2.94019938e-01 -1.21012270e+00 4.53465462e-01 -4.08394039e-01
-1.71363199e+00 1.07025039e+00 -1.99474663e-01 -8.41944695e-01
-9.85793233e-01 -6.36634409e-01 -5.23348510e-01 1.64154625e+00
-1.32975078e+00 -1.20472741e+00 -8.15501064e-02 1.05004513e+00
6.20031655e-01 -8.74616265e-01 1.43240929e+00 1.15758061e-01
-8.36781681e-01 1.23892760e+00 2.93563038e-01 5.41184604e-01
1.17388439e+00 -1.13269591e+00 4.29604232e-01 7.02379584e-01
-1.43749919e-02 7.44082093e-01 5.85100055e-01 -1.21470109e-01
-1.14407134e+00 -1.13358796e+00 9.60625827e-01 -6.05107903e-01
5.09255707e-01 -6.37323976e-01 -1.14862418e+00 8.95667195e-01
3.21434081e-01 1.80867702e-01 1.02332115e+00 3.75249386e-01
-6.90335214e-01 -1.36062667e-01 -1.27175248e+00 1.82169631e-01
5.96186399e-01 -9.70637321e-01 -6.60357893e-01 2.48290807e-01
4.48104203e-01 -3.53322804e-01 -6.23505592e-01 2.20178559e-01
3.24074388e-01 -5.05202949e-01 8.04293275e-01 -9.76401985e-01
-3.11403602e-01 -1.52177615e-02 -5.20732284e-01 -1.45168054e+00
1.59095656e-02 -5.26762307e-01 1.88989609e-01 1.53384423e+00
7.16348350e-01 -7.31166065e-01 4.64050353e-01 1.03382993e+00
-2.89759815e-01 -7.22701922e-02 -1.12790465e+00 -9.85478759e-01
4.85094249e-01 -2.77274460e-01 5.64814031e-01 1.07269537e+00
1.48406237e-01 5.95937669e-01 -3.36465061e-01 3.19401085e-01
3.57147872e-01 -1.23751462e-01 1.02551866e+00 -1.10460222e+00
6.96514398e-02 -3.46547484e-01 -1.94351286e-01 -9.61626709e-01
6.23312414e-01 -9.32329774e-01 3.77840191e-01 -8.74103725e-01
-1.94489956e-01 -2.93693215e-01 -3.68712842e-01 6.27706885e-01
1.26566961e-01 1.79510176e-01 -5.82131967e-02 -1.17241025e-01
-4.36442226e-01 5.48875928e-01 5.88223159e-01 -3.07559580e-01
-3.89804900e-01 2.46513039e-01 -8.51129889e-01 5.87073922e-01
9.09896076e-01 1.27674550e-01 -3.67125869e-01 -3.10886919e-01
-6.87121272e-01 -1.45393610e-01 1.72668323e-01 -9.46883857e-01
4.44966614e-01 2.42302105e-01 4.58137035e-01 -3.08921009e-01
5.40870667e-01 -2.93468654e-01 -4.00622368e-01 1.36659950e-01
-8.92993927e-01 -1.72670856e-01 6.02474272e-01 1.34005219e-01
-3.92339855e-01 5.15665393e-03 1.06199932e+00 -3.80137488e-02
-8.85977566e-01 1.84273794e-01 -2.67354757e-01 3.12370598e-01
7.14392543e-01 4.43393774e-02 -1.77494839e-01 -6.57803357e-01
-1.10925925e+00 3.86859030e-01 1.48938954e-01 8.01773548e-01
4.80388492e-01 -1.26110053e+00 -1.27316880e+00 7.30911851e-01
3.35113347e-01 -3.06015879e-01 -1.09794186e-02 -1.63430907e-02
4.33715209e-02 5.88310361e-01 5.02641127e-02 -4.92374241e-01
-1.29689968e+00 2.91059136e-01 7.89565861e-01 3.83352846e-01
-3.75406682e-01 1.25694978e+00 -9.22811776e-02 -1.11402440e+00
7.23496318e-01 -7.51476958e-02 -2.15750262e-01 7.78497308e-02
1.05893540e+00 1.26260787e-01 2.52128541e-01 -9.91072953e-01
-7.66510487e-01 -2.20105983e-02 -3.64514530e-01 -5.30626655e-01
1.16753781e+00 -1.01770997e-01 3.70804310e-01 8.14004779e-01
1.31017184e+00 -1.25097245e-01 -9.46398139e-01 -6.12407029e-01
-4.19802442e-02 1.12219408e-01 8.83015767e-02 -1.21068490e+00
-6.40660167e-01 8.54069531e-01 7.31764913e-01 3.12104344e-01
6.82641327e-01 9.64811444e-02 4.09948528e-01 4.50219780e-01
2.27079198e-01 -1.15689433e+00 -1.64720014e-01 9.77441311e-01
1.36872888e+00 -1.56494355e+00 -5.71276128e-01 2.13085681e-01
-9.42143619e-01 1.11576509e+00 8.48132730e-01 2.74410658e-02
8.71663749e-01 3.78435165e-01 6.10005856e-01 2.35964447e-01
-9.01454628e-01 -1.03564359e-01 3.60165775e-01 8.23312283e-01
7.56523788e-01 1.60160407e-01 4.30479050e-01 9.99616027e-01
-4.71345991e-01 -1.98734224e-01 3.22392046e-01 5.86411536e-01
-4.42564666e-01 -9.54697788e-01 -3.12720448e-01 1.49313241e-01
-2.00285792e-01 -8.96580741e-02 -8.82731259e-01 6.98341489e-01
-2.35788301e-01 1.24023056e+00 6.66524991e-02 -4.05637324e-01
5.53467870e-01 7.13665187e-01 -5.81581593e-02 -9.11978364e-01
-7.81579137e-01 -1.07630312e-01 4.32118118e-01 -3.30267102e-01
-5.78342378e-02 -7.90079176e-01 -1.12356150e+00 -1.66346163e-01
-2.18625441e-02 3.54385227e-01 8.56931567e-01 7.05053866e-01
4.15556252e-01 3.51447552e-01 9.29199517e-01 -8.01866710e-01
-1.16092324e+00 -1.37949538e+00 -8.51825833e-01 1.58036843e-01
8.89911652e-01 -1.42053157e-01 -4.69237328e-01 -1.35804594e-01] | [14.182998657226562, 6.618321895599365] |
33fccbf3-0e21-4e9c-af9f-210b53213b95 | analyzing-the-mono-and-cross-lingual | 2205.11758 | null | https://arxiv.org/abs/2205.11758v2 | https://arxiv.org/pdf/2205.11758v2.pdf | Analyzing the Mono- and Cross-Lingual Pretraining Dynamics of Multilingual Language Models | The emergent cross-lingual transfer seen in multilingual pretrained models has sparked significant interest in studying their behavior. However, because these analyses have focused on fully trained multilingual models, little is known about the dynamics of the multilingual pretraining process. We investigate when these models acquire their in-language and cross-lingual abilities by probing checkpoints taken from throughout XLM-R pretraining, using a suite of linguistic tasks. Our analysis shows that the model achieves high in-language performance early on, with lower-level linguistic skills acquired before more complex ones. In contrast, the point in pretraining when the model learns to transfer cross-lingually differs across language pairs. Interestingly, we also observe that, across many languages and tasks, the final model layer exhibits significant performance degradation over time, while linguistic knowledge propagates to lower layers of the network. Taken together, these insights highlight the complexity of multilingual pretraining and the resulting varied behavior for different languages over time. | ['Luke Zettlemoyer', 'Hila Gonen', 'Terra Blevins'] | 2022-05-24 | null | null | null | null | ['xlm-r'] | ['natural-language-processing'] | [-2.62684345e-01 -9.24227536e-02 -6.71856329e-02 -2.79931903e-01
-9.77998018e-01 -1.10115492e+00 9.16585028e-01 2.97726721e-01
-7.00729191e-01 7.15755820e-01 2.93398201e-01 -5.95767736e-01
8.67352560e-02 -4.09628868e-01 -1.17064893e+00 -3.23449284e-01
-2.23511264e-01 5.43872237e-01 -6.01052567e-02 -4.81353909e-01
-2.64674485e-01 3.21802497e-01 -1.11299396e+00 3.66974473e-01
1.15336204e+00 -3.66751920e-03 3.49338293e-01 4.76117224e-01
3.53505835e-02 7.33366847e-01 -3.51840973e-01 -5.93016446e-01
2.64907062e-01 -3.82749051e-01 -8.78862619e-01 -1.85951412e-01
7.68694699e-01 -5.51547855e-02 -2.36136556e-01 1.04113209e+00
1.53593421e-01 -1.59217402e-01 5.17627120e-01 -6.55772150e-01
-7.09564924e-01 1.35345793e+00 -4.71124113e-01 3.41035396e-01
2.52164394e-01 4.73144263e-01 1.00050640e+00 -8.68343532e-01
8.14292312e-01 1.32744420e+00 7.99299538e-01 3.43550593e-01
-1.70016444e+00 -5.84197223e-01 4.91086662e-01 -2.95694172e-01
-1.23554933e+00 -7.39365518e-01 2.41683409e-01 -8.47102106e-01
1.12964857e+00 -4.52833027e-01 7.75031984e-01 1.04604673e+00
4.24980313e-01 8.38817954e-01 1.94729340e+00 -4.84582871e-01
-4.28908527e-01 4.41891968e-01 1.89902455e-01 8.11425924e-01
2.80529320e-01 3.51595968e-01 -8.40932667e-01 4.11387712e-01
4.18336689e-01 -3.54071051e-01 -2.64604270e-01 -1.38816193e-01
-1.22882068e+00 3.86481524e-01 2.83345729e-01 8.18445265e-01
-2.36190230e-01 -7.79919550e-02 4.83244658e-01 1.01742506e+00
3.66267592e-01 6.25828266e-01 -8.81056547e-01 -4.00067866e-01
-8.54686558e-01 -2.03000888e-01 4.52044576e-01 7.36281157e-01
1.08183312e+00 1.13237493e-01 2.88057983e-01 9.17403579e-01
1.38693795e-01 6.40056014e-01 5.45600295e-01 -5.03470361e-01
7.25476682e-01 5.46265125e-01 -4.08321410e-01 -3.56649935e-01
-2.42358863e-01 -8.78022790e-01 -4.03468013e-01 1.35057494e-01
7.70195305e-01 -4.32822943e-01 -5.64469993e-01 2.24322367e+00
-1.09919354e-01 -3.33273470e-01 3.39369893e-01 5.06461322e-01
1.73596203e-01 5.35231948e-01 4.33776408e-01 -4.83933911e-02
1.08986330e+00 -6.85276866e-01 -1.82017341e-01 -5.52550435e-01
1.13257611e+00 -8.47863972e-01 1.45117736e+00 2.67930597e-01
-1.35779011e+00 -7.70274937e-01 -1.02258074e+00 -2.42340505e-01
-5.35729885e-01 -2.23515853e-01 5.72411537e-01 4.70429987e-01
-1.70500851e+00 5.61266422e-01 -7.59643376e-01 -7.08199680e-01
5.39824069e-02 2.69657224e-01 -5.55457890e-01 -3.88827741e-01
-1.32095647e+00 1.26747155e+00 3.82156134e-01 -1.56245485e-01
-1.32102478e+00 -8.36282969e-01 -5.93221128e-01 4.26829048e-02
6.54700026e-02 -3.42007935e-01 1.41641414e+00 -1.23752022e+00
-1.34154606e+00 1.20328856e+00 2.99576335e-02 -1.00486346e-01
4.77386922e-01 -2.11652219e-01 -2.53140807e-01 -4.84580010e-01
7.68951774e-02 7.21797049e-01 2.28713423e-01 -1.33990407e+00
-8.48623335e-01 -4.70806807e-01 1.38682708e-01 6.23061061e-01
-3.46487761e-01 2.19638199e-01 -2.46056825e-01 -3.78926158e-01
-2.39395127e-01 -9.56092298e-01 1.95631072e-01 -7.63434470e-01
-5.17948195e-02 -2.08802834e-01 -2.16330104e-02 -8.55618119e-01
1.03074157e+00 -2.21384287e+00 3.36341023e-01 1.04225732e-01
8.49609151e-02 -3.45225297e-02 -3.22394252e-01 7.50609457e-01
-1.39712855e-01 3.13086867e-01 -4.81365062e-03 -4.60308075e-01
-1.84209496e-02 1.65522918e-01 -2.08721936e-01 3.95005941e-01
1.54339254e-01 1.13770127e+00 -9.73194957e-01 -2.63358384e-01
-2.46130005e-01 2.64941186e-01 -7.00421512e-01 9.83833894e-02
-2.50296950e-01 8.95310044e-01 -2.11130377e-04 3.72373432e-01
1.97293058e-01 -7.17553720e-02 4.96853918e-01 3.44793916e-01
-6.26773775e-01 6.07938826e-01 -3.41147691e-01 1.93044949e+00
-8.28360319e-01 7.98054636e-01 3.16639900e-01 -8.39849412e-01
2.65013397e-01 3.48432124e-01 1.84629172e-01 -9.64096665e-01
1.32294700e-01 5.94284534e-01 8.64352465e-01 -1.84075549e-01
4.70688254e-01 -3.36876184e-01 -1.93351820e-01 7.36505449e-01
3.37252855e-01 -2.29512639e-02 4.31720972e-01 2.40412727e-01
7.46001899e-01 1.85086116e-01 5.91799011e-03 -8.20186436e-01
3.27975541e-01 2.38077924e-01 2.28633806e-01 7.62934864e-01
-1.32262230e-01 -1.97193965e-01 3.59513938e-01 -3.75600681e-02
-9.29381251e-01 -1.17425883e+00 -3.44265431e-01 1.81713426e+00
-5.46964705e-01 -1.88857287e-01 -9.48707759e-01 -4.20394212e-01
4.29058410e-02 6.95965648e-01 -5.43045819e-01 -4.32911307e-01
-9.07072127e-01 -5.26699841e-01 5.66980362e-01 1.81578249e-01
3.31263185e-01 -1.15871370e+00 3.12731527e-02 2.05008045e-01
1.59086268e-02 -1.01714158e+00 -5.66078961e-01 3.57589036e-01
-1.13142157e+00 -6.66021943e-01 -6.51752651e-01 -1.06755710e+00
7.24959731e-01 -5.62905222e-02 1.62228847e+00 1.26285300e-01
1.31508797e-01 4.85497236e-01 7.29131773e-02 -1.91740081e-01
-7.94068515e-01 7.02918231e-01 3.00860584e-01 -2.87797540e-01
2.35868260e-01 -6.03513360e-01 -6.17384724e-02 -3.71001437e-02
-5.06211817e-01 3.36492099e-02 9.20476437e-01 4.75675493e-01
3.61987889e-01 1.77707337e-02 5.56079090e-01 -1.04456496e+00
8.67777586e-01 -6.22096777e-01 -5.94778419e-01 3.79685760e-01
-6.51838243e-01 2.79423058e-01 8.37281704e-01 -3.86689901e-01
-1.11568999e+00 -4.40997094e-01 1.95414528e-01 2.34153226e-01
-1.21167161e-01 8.98547649e-01 1.22978231e-02 1.31206721e-01
6.73467159e-01 4.57458347e-01 -2.48191372e-01 -6.46821618e-01
4.30287778e-01 4.01185006e-01 4.69164968e-01 -1.30585551e+00
8.16596389e-01 -7.03000650e-02 -6.81020975e-01 -8.47651243e-01
-9.56380904e-01 -1.03296684e-02 -8.90829861e-01 -2.26844668e-01
8.24257970e-01 -1.41567993e+00 -5.22841036e-01 8.02277684e-01
-8.62890482e-01 -1.28921568e+00 -2.60046721e-01 5.54685891e-01
-2.87313223e-01 -3.47682804e-01 -8.90047908e-01 -4.05204654e-01
1.66734040e-01 -1.53081560e+00 5.91212153e-01 4.74721640e-02
-2.82789528e-01 -1.57265353e+00 3.24001014e-01 1.28737226e-01
3.54195237e-01 -4.23964322e-01 1.49947882e+00 -5.27854919e-01
-6.29839003e-01 3.18047017e-01 7.55341500e-02 2.10951656e-01
1.78882480e-01 -8.74579027e-02 -7.92607725e-01 -7.17295706e-01
-2.75013357e-01 -7.60186911e-01 6.45837963e-01 1.99592054e-01
3.87478262e-01 -6.08039200e-02 -2.04930037e-01 6.71221495e-01
1.33638942e+00 -3.95426936e-02 -1.32639065e-01 4.19759303e-01
5.42925060e-01 6.96454823e-01 5.12542911e-02 -4.19718236e-01
6.92973733e-01 5.32450557e-01 -2.97427088e-01 -3.43357548e-02
-2.39062428e-01 -5.28033316e-01 1.08193195e+00 1.48257983e+00
1.08643226e-01 6.60130158e-02 -1.27818096e+00 7.10930169e-01
-1.41899550e+00 -4.87311244e-01 2.02980846e-01 2.41383457e+00
1.62322283e+00 3.24718624e-01 1.73648536e-01 -4.41314608e-01
2.98066705e-01 -3.78948599e-02 -5.70622385e-01 -5.54900467e-01
-2.84316182e-01 1.52641997e-01 4.06594723e-01 9.26805854e-01
-4.67523813e-01 1.74260056e+00 6.99341440e+00 3.49740595e-01
-1.47324920e+00 2.61661083e-01 5.67190468e-01 1.18981034e-01
-4.49854851e-01 -3.96428481e-02 -9.04340982e-01 6.80538192e-02
1.27429950e+00 -4.36638266e-01 9.76912320e-01 2.86488414e-01
5.58476821e-02 -1.86222404e-01 -1.50819623e+00 2.07755029e-01
-1.81236520e-01 -5.76573730e-01 -2.95191053e-02 1.69100955e-01
1.17381513e+00 7.59824455e-01 4.29743916e-01 8.28556061e-01
1.20492375e+00 -1.07934511e+00 9.70242023e-01 3.22216690e-01
9.83306110e-01 -7.50140667e-01 1.10112518e-01 6.04758978e-01
-9.58742678e-01 9.55794305e-02 -1.09359890e-01 -1.97928965e-01
7.56956339e-02 1.94950059e-01 -8.23264003e-01 1.15200900e-01
3.86556804e-01 6.12758696e-01 -1.00666761e+00 3.35365444e-01
-3.60335529e-01 7.95314789e-01 -7.62989745e-02 4.90434110e-01
3.50759506e-01 -1.73516005e-01 2.28185475e-01 1.21572161e+00
1.98474616e-01 -3.11771363e-01 3.82330924e-01 6.00209057e-01
-3.77812117e-01 3.02091658e-01 -9.70363438e-01 -3.57050776e-01
3.43419373e-01 9.21217322e-01 -3.38684678e-01 -4.02237803e-01
-6.87523127e-01 6.61620617e-01 1.15255392e+00 5.18114030e-01
-4.29362178e-01 3.07289183e-01 7.13228583e-01 3.63612592e-01
-3.29851002e-01 -8.56391370e-01 -2.40612373e-01 -1.26622522e+00
-2.38111153e-01 -1.30236876e+00 1.01973191e-01 -5.82404673e-01
-1.03692639e+00 5.61390460e-01 -5.65166250e-02 -4.35310900e-01
-2.93986827e-01 -5.45841932e-01 -3.18933845e-01 1.14035487e+00
-1.46440506e+00 -1.01705110e+00 2.73173600e-01 6.85705662e-01
3.89251858e-01 -3.80556792e-01 6.21710539e-01 3.81423801e-01
-7.88034201e-01 9.09023643e-01 2.82161534e-01 3.31235111e-01
9.63426530e-01 -1.18745816e+00 4.74149168e-01 9.27839220e-01
3.94192129e-01 1.36454332e+00 5.06416142e-01 -7.33336210e-01
-1.29332399e+00 -9.75746989e-01 1.06566310e+00 -5.97784400e-01
1.07829964e+00 -5.17810345e-01 -9.81418788e-01 1.35454679e+00
7.91272104e-01 -4.69311953e-01 5.02291620e-01 6.60237134e-01
-3.79655242e-01 -3.08592599e-02 -4.79483306e-01 9.85944510e-01
1.32013488e+00 -1.01702023e+00 -4.84412909e-01 2.68793881e-01
8.60721052e-01 -1.29437745e-01 -1.06761050e+00 1.79510042e-01
3.43751073e-01 -9.84977901e-01 6.39547765e-01 -7.89348006e-01
4.64917153e-01 1.11809932e-01 -1.24605708e-02 -1.94882870e+00
-3.44769329e-01 -5.55775583e-01 4.81985390e-01 1.35832274e+00
9.62255776e-01 -8.48319709e-01 1.29155695e-01 9.66816694e-02
-4.07832950e-01 -5.55914521e-01 -5.79710305e-01 -7.37930059e-01
9.82013524e-01 -2.31464684e-01 8.66465792e-02 1.31513321e+00
-2.49903444e-02 5.79029083e-01 1.44648135e-01 1.01072729e-01
4.81517375e-01 -5.06415367e-02 6.38033390e-01 -1.19308186e+00
-2.78890163e-01 -7.06305027e-01 2.71774024e-01 -9.19879735e-01
6.73167348e-01 -1.59645629e+00 1.23312399e-01 -1.14803743e+00
2.54138768e-01 -8.10784817e-01 -3.90172958e-01 5.55608928e-01
-2.73178697e-01 -5.39577305e-02 3.26753855e-01 3.18294406e-01
-4.50256586e-01 1.14073865e-01 1.38734710e+00 1.50094718e-01
-2.51994014e-01 -4.39814866e-01 -8.97378862e-01 8.53689432e-01
7.57351398e-01 -4.84933764e-01 -4.81551349e-01 -1.24560583e+00
4.66630429e-01 -1.00496903e-01 -2.49294475e-01 -1.02802086e+00
8.68938193e-02 -8.26109871e-02 4.42102998e-02 2.26394795e-02
-1.78462461e-01 -7.05761015e-01 -1.09189555e-01 5.55361032e-01
-4.75098997e-01 6.08394623e-01 5.44918239e-01 -7.69866556e-02
-2.07257748e-01 1.57588094e-01 9.87789392e-01 -2.76335269e-01
-5.27508497e-01 2.32990012e-01 -6.20282590e-01 6.36110067e-01
6.78861260e-01 8.31649378e-02 -2.61339307e-01 -7.53022730e-02
-8.78938675e-01 3.90373141e-01 8.42667103e-01 6.15380883e-01
-2.44500890e-01 -1.21722698e+00 -9.90505397e-01 2.96102673e-01
9.35114846e-02 -1.77655816e-01 4.58521731e-02 1.04828835e+00
-3.07003796e-01 7.02581644e-01 -2.24154592e-01 -7.35182941e-01
-7.35420823e-01 3.23947132e-01 5.98966539e-01 -6.93305194e-01
-1.02591969e-01 9.15134072e-01 5.36848545e-01 -8.53681207e-01
-5.05810827e-02 -3.21958303e-01 -9.12148952e-02 2.34224245e-01
1.73882321e-02 -8.37585330e-02 -5.11855781e-02 -8.40549231e-01
-1.08316995e-01 7.42616832e-01 -5.11899889e-01 -4.98215258e-01
1.23890150e+00 -2.20697671e-01 -2.60831147e-01 1.06205058e+00
1.14536023e+00 5.10168552e-01 -1.26459932e+00 -5.44698834e-01
1.71136647e-01 2.42655411e-01 -2.83092469e-01 -1.04243612e+00
-1.00854266e+00 1.09660065e+00 1.84942544e-01 -9.86388773e-02
7.11353004e-01 1.04192898e-01 3.79870266e-01 3.40244442e-01
6.40543222e-01 -1.04079723e+00 -6.33093417e-02 1.21025097e+00
8.19630563e-01 -1.06526661e+00 -3.44500124e-01 2.39736781e-01
-4.29077238e-01 4.16359186e-01 9.30250883e-01 3.55725340e-03
5.07971048e-01 2.82706797e-01 3.25535804e-01 -7.70339593e-02
-9.79141951e-01 -1.40576839e-01 9.36921537e-02 3.00144106e-01
1.09752476e+00 1.36932999e-01 -5.64014390e-02 2.82820672e-01
-8.20946932e-01 -4.31004256e-01 3.19029659e-01 8.03822160e-01
-3.89892727e-01 -1.21805286e+00 -1.09442405e-01 -2.81742625e-02
-5.38320363e-01 -8.08371425e-01 -5.26122153e-01 1.05774081e+00
1.99752122e-01 5.12386560e-01 1.22789018e-01 -1.93115756e-01
3.21763068e-01 6.58905208e-01 7.12921679e-01 -1.00497258e+00
-1.03463304e+00 -1.13802373e-01 -1.38587714e-03 -2.98001796e-01
-4.98311818e-02 -9.52799439e-01 -1.16154778e+00 -4.64486957e-01
1.03384361e-01 1.45430043e-01 4.02503878e-01 1.10768569e+00
1.50215626e-01 6.55377805e-01 3.93674999e-01 -5.50096750e-01
-4.18158233e-01 -1.15598965e+00 -4.62155312e-01 2.49449134e-01
5.11339545e-01 -3.36408913e-01 -2.88410068e-01 1.09416820e-01] | [10.875618934631348, 9.978885650634766] |
b26dca88-745c-4d1a-8d77-a6a6f1663d3d | few-labeled-atlases-are-necessary-for-deep | 1908.04466 | null | https://arxiv.org/abs/1908.04466v4 | https://arxiv.org/pdf/1908.04466v4.pdf | Few Labeled Atlases are Necessary for Deep-Learning-Based Segmentation | We tackle biomedical image segmentation in the scenario of only a few labeled brain MR images. This is an important and challenging task in medical applications, where manual annotations are time-consuming. Current multi-atlas based segmentation methods use image registration to warp segments from labeled images onto a new scan. In a different paradigm, supervised learning-based segmentation strategies have gained popularity. These method consistently use relatively large sets of labeled training data, and their behavior in the regime of a few labeled biomedical images has not been thoroughly evaluated. In this work, we provide two important results for segmentation in the scenario where few labeled images are available. First, we propose a straightforward implementation of efficient semi-supervised learning-based registration method, which we showcase in a multi-atlas segmentation framework. Second, through an extensive empirical study, we evaluate the performance of a supervised segmentation approach, where the training images are augmented via random deformations. Surprisingly, we find that in both paradigms, accurate segmentation is generally possible even in the context of few labeled images. | ['Mert R. Sabuncu', 'Adrian V. Dalca', 'Hyeon Woo Lee'] | 2019-08-13 | null | null | null | null | ['deformable-medical-image-registration'] | ['medical'] | [ 5.65522194e-01 3.43681961e-01 -4.37987708e-02 -5.42683125e-01
-1.17876387e+00 -5.97433686e-01 5.63320935e-01 4.46646959e-01
-9.11669612e-01 7.57960439e-01 -2.61367381e-01 -7.75438361e-03
-1.15582876e-01 -4.15693939e-01 -5.94094694e-01 -8.67444336e-01
-2.47769840e-02 9.79252696e-01 4.40358996e-01 -1.32290259e-01
6.08188063e-02 5.21372199e-01 -1.09028900e+00 -1.76634371e-01
7.71303236e-01 6.34946942e-01 2.36102492e-01 4.00704354e-01
4.91436869e-02 3.33490133e-01 -3.38106692e-01 -2.02447012e-01
4.16331202e-01 -5.28852999e-01 -1.25226843e+00 4.16004121e-01
3.61079603e-01 7.75082931e-02 2.01158687e-01 1.06909657e+00
5.07161617e-01 1.22869447e-01 5.35601974e-01 -8.36549163e-01
2.58034170e-01 4.73867238e-01 -4.86421883e-01 3.20435613e-01
5.03138304e-02 -6.08931296e-02 5.87589443e-01 -6.18221164e-01
8.47144544e-01 5.75305939e-01 6.05569124e-01 4.12906677e-01
-1.47730565e+00 -2.33813718e-01 -1.93374693e-01 -1.73259124e-01
-1.34322429e+00 -2.68867016e-01 7.35110939e-01 -7.07479894e-01
2.86420047e-01 3.16098273e-01 5.55358827e-01 5.60002744e-01
2.11177185e-01 5.25404692e-01 1.64112008e+00 -5.56996346e-01
4.46011662e-01 1.29523978e-01 4.53881502e-01 5.56845129e-01
1.40829951e-01 -1.27704173e-01 -8.24714079e-04 -1.59528777e-01
7.23949492e-01 -1.54682651e-01 -2.05643475e-01 -5.77667058e-01
-1.50032961e+00 7.96156049e-01 2.90458679e-01 8.39232683e-01
-4.61996406e-01 -2.03616560e-01 4.15427387e-01 2.20143601e-01
6.92180991e-01 4.55134988e-01 -2.34912634e-01 2.24877402e-01
-1.61168420e+00 1.06505737e-01 6.18907690e-01 6.61751509e-01
8.94501328e-01 -3.91679615e-01 4.07962427e-02 7.80894697e-01
-3.56830377e-03 1.98259994e-01 7.40306675e-01 -6.35488033e-01
1.44299582e-01 3.73777539e-01 -1.64252922e-01 -6.44798219e-01
-6.98505759e-01 -3.65169793e-01 -9.07377243e-01 2.98151579e-02
8.13748419e-01 -7.10660666e-02 -8.48167539e-01 1.55390632e+00
5.86569607e-01 1.45243704e-01 -6.07183240e-02 7.64315546e-01
3.63506198e-01 1.37155503e-02 1.32089421e-01 -6.61278188e-01
1.21786511e+00 -8.33500803e-01 -5.41469634e-01 -6.36343211e-02
8.47512662e-01 -8.02573800e-01 9.52050328e-01 3.76846254e-01
-1.12809324e+00 -1.97931767e-01 -7.57247388e-01 1.76729634e-01
-1.88802853e-01 -6.41586352e-03 4.05080676e-01 7.57077157e-01
-1.21439350e+00 6.85893536e-01 -1.03997779e+00 -5.38071692e-01
4.66845274e-01 7.09058583e-01 -5.63402653e-01 -1.45877928e-01
-7.79835880e-01 9.14439917e-01 4.33201641e-01 8.62947181e-02
-6.64723098e-01 -4.77820456e-01 -6.91056132e-01 -4.05650407e-01
5.15931427e-01 -3.27124417e-01 1.11984813e+00 -9.70647275e-01
-1.26277518e+00 1.43772507e+00 -4.00453284e-02 -4.35601115e-01
8.49664867e-01 8.35372731e-02 -1.89849939e-02 4.03365731e-01
1.73051357e-01 6.90453351e-01 8.07169735e-01 -1.34283626e+00
-1.51752159e-01 -6.67283356e-01 -9.70623344e-02 -3.54023604e-03
1.07332312e-01 1.58797964e-01 -1.37280181e-01 -6.68477297e-01
3.21167737e-01 -1.14305294e+00 -7.33382046e-01 -2.63430357e-01
-4.60634232e-01 5.05266637e-02 4.96896267e-01 -5.05380034e-01
7.59850621e-01 -1.88563311e+00 2.61083543e-01 4.74504709e-01
3.02447081e-01 1.01070389e-01 1.73617914e-01 3.73522602e-02
-1.83563858e-01 -5.17229773e-02 -8.99072409e-01 -3.97679716e-01
-2.58731097e-01 3.58052254e-01 1.45674229e-01 7.83659935e-01
-7.01082423e-02 1.00931382e+00 -7.77249336e-01 -9.52376425e-01
1.49406523e-01 2.57552862e-01 -4.73724633e-01 2.08291531e-01
7.97759071e-02 1.20015812e+00 -3.77517849e-01 4.23337162e-01
4.43781674e-01 -2.97134995e-01 3.17016751e-01 -1.89856470e-01
-1.28181413e-01 -3.44927371e-01 -1.00866199e+00 2.09192181e+00
-3.44120294e-01 2.70848691e-01 7.93517902e-02 -1.56311417e+00
6.20798647e-01 5.51288307e-01 1.13544643e+00 -2.79726446e-01
4.16097879e-01 4.87032205e-01 7.20560551e-02 -4.53573853e-01
1.79443344e-01 -6.20992482e-01 -7.69585073e-02 6.58623457e-01
1.24348454e-01 -4.57556605e-01 3.59227687e-01 -5.40826432e-02
1.19427514e+00 -1.09273396e-01 4.83823687e-01 -4.99327660e-01
5.81661105e-01 1.56724244e-01 2.88715333e-01 5.99086404e-01
-2.36340806e-01 8.56368303e-01 2.61638761e-01 -3.89928401e-01
-9.99591649e-01 -7.48903871e-01 -5.58137238e-01 6.23460710e-01
7.98971653e-02 1.33668445e-02 -1.38206971e+00 -7.81393588e-01
-5.90213954e-01 2.28458971e-01 -6.23993576e-01 2.13898867e-01
-8.74845862e-01 -1.21280897e+00 3.59456301e-01 1.86849788e-01
2.46714771e-01 -1.11808681e+00 -8.89315844e-01 3.04688394e-01
-3.13576870e-02 -1.32203293e+00 -4.12098378e-01 9.84287187e-02
-1.31649804e+00 -1.26196229e+00 -1.25253117e+00 -9.02120888e-01
1.04675281e+00 -1.73033997e-01 1.01502204e+00 2.18873441e-01
-4.21245843e-01 5.38599670e-01 -2.72713363e-01 -2.12660264e-02
-5.09062707e-01 2.97806948e-01 3.32574174e-02 2.07646653e-01
-3.99074107e-02 -7.62935638e-01 -4.72442567e-01 5.47422290e-01
-1.12771642e+00 -5.54335713e-02 6.54567480e-01 8.40246797e-01
8.31960797e-01 -4.66362923e-01 5.04996121e-01 -1.45090532e+00
2.90390044e-01 -3.04530740e-01 -6.15803003e-01 2.64072865e-01
-5.35979331e-01 -9.04472545e-02 4.33496952e-01 -4.66724664e-01
-6.25027597e-01 4.70635504e-01 -2.87465304e-01 -6.95239604e-02
-5.58720887e-01 5.11258125e-01 8.18830207e-02 -5.56556702e-01
6.58291340e-01 1.04064092e-01 1.30133316e-01 -5.53086936e-01
2.95824230e-01 4.01518017e-01 5.69936991e-01 -5.80174088e-01
7.17532456e-01 6.35742128e-01 1.21080704e-01 -7.70505011e-01
-6.42884195e-01 -5.91214359e-01 -1.37646806e+00 -2.94682980e-01
9.70538676e-01 -4.20912504e-01 -1.14694156e-01 2.75362670e-01
-9.11315918e-01 -6.13173246e-01 -5.60733795e-01 7.25481987e-01
-9.21704292e-01 6.59186780e-01 -3.93633127e-01 -3.19261014e-01
-2.28689700e-01 -1.65245891e+00 1.04577816e+00 -2.76670493e-02
-1.99938536e-01 -1.21799839e+00 3.15970689e-01 4.33557689e-01
1.90762058e-01 3.95382673e-01 8.89371037e-01 -1.15475738e+00
-3.80614877e-01 -3.23850095e-01 2.28093192e-02 1.73357978e-01
2.00179264e-01 -6.87876105e-01 -8.23513508e-01 -4.62048411e-01
2.36762822e-01 -2.03418419e-01 4.95560676e-01 4.71054375e-01
1.15199292e+00 7.95586258e-02 -3.41861039e-01 3.26936364e-01
1.41452146e+00 7.87951797e-02 3.70821595e-01 -6.47584815e-03
6.88605130e-01 8.46346915e-01 6.10463142e-01 -5.45424819e-02
1.68770552e-01 7.97813058e-01 1.08273044e-01 -4.10456181e-01
-3.73318465e-03 5.33925295e-01 -2.81450957e-01 1.17819059e+00
-4.08069551e-01 1.18971117e-01 -1.15551209e+00 6.32321835e-01
-1.58276629e+00 -4.22366261e-01 -8.54581594e-02 2.34543061e+00
1.02893710e+00 -1.32958964e-03 3.10302407e-01 2.63381839e-01
7.78638959e-01 -8.34143758e-02 -3.18045050e-01 1.35008425e-01
1.92319587e-01 5.61219394e-01 5.51971078e-01 4.87306058e-01
-1.19330382e+00 7.14501441e-01 6.80195379e+00 6.85031295e-01
-1.17354620e+00 7.65866756e-01 7.24844813e-01 1.07404441e-01
8.95438343e-02 -4.82734106e-02 -3.61163437e-01 4.05013770e-01
9.09359992e-01 1.16716243e-01 8.16854611e-02 4.63210464e-01
1.90989211e-01 -4.74956542e-01 -1.18857610e+00 9.35638309e-01
1.19399518e-01 -1.09551954e+00 -3.72635514e-01 2.13007867e-01
7.80469000e-01 3.91225778e-02 -1.69130370e-01 -1.81993693e-01
-9.90053639e-02 -9.17699277e-01 3.78381699e-01 3.96879286e-01
7.14730680e-01 -4.38824564e-01 6.85479045e-01 4.05816317e-01
-8.46944630e-01 4.22887832e-01 -2.89402101e-02 5.44498980e-01
4.50942338e-01 6.39197886e-01 -5.37209213e-01 6.48461223e-01
2.81275064e-01 4.52082694e-01 -5.08756399e-01 1.31866014e+00
4.80326004e-02 7.02894866e-01 -3.04096282e-01 6.48978770e-01
2.93063760e-01 -3.83691967e-01 4.72628117e-01 9.87236679e-01
1.59990668e-01 1.86755404e-01 5.35819769e-01 5.17928541e-01
-1.80368219e-02 5.91077447e-01 -4.76560205e-01 3.89000624e-01
-2.37669721e-01 1.46580172e+00 -1.41973066e+00 -3.98515850e-01
-3.21728706e-01 7.82979250e-01 1.97558120e-01 1.38017789e-01
-4.79032993e-01 1.76903144e-01 -2.50762790e-01 3.34612727e-01
-1.04695506e-01 -3.02860051e-01 -2.66802788e-01 -1.12771225e+00
-2.03127693e-02 -6.50056541e-01 3.15149784e-01 -2.54575551e-01
-9.97065604e-01 8.14455926e-01 4.24486309e-01 -1.06166577e+00
-3.75826299e-01 -3.78589123e-01 -4.65541095e-01 4.95032579e-01
-1.37965119e+00 -1.14092743e+00 -7.82868788e-02 5.37530422e-01
5.62602341e-01 -1.26789436e-01 8.34284246e-01 5.94455838e-01
-4.20601577e-01 4.14009720e-01 1.36441514e-01 1.81020021e-01
6.57416284e-01 -1.28869200e+00 -1.00413591e-01 7.44021773e-01
3.54860693e-01 4.86022204e-01 6.82631731e-01 -5.66390276e-01
-8.66338491e-01 -7.68359244e-01 7.14017630e-01 -1.41972423e-01
5.39999485e-01 -1.90737069e-01 -1.08799338e+00 7.51013279e-01
1.69590980e-01 5.45959234e-01 8.23869348e-01 -2.26234987e-01
2.89798677e-01 1.03565015e-01 -1.33161402e+00 2.83502072e-01
6.84817851e-01 -3.39158654e-01 -5.85808039e-01 7.80046105e-01
2.08575279e-01 -4.54843283e-01 -1.24326205e+00 3.76126349e-01
2.15184644e-01 -7.19917119e-01 8.64673018e-01 -3.64243448e-01
4.05529439e-02 -1.55374721e-01 9.67439190e-02 -1.29225731e+00
4.47689772e-01 -6.51730597e-01 5.43850482e-01 8.27687144e-01
3.49258751e-01 -6.63635254e-01 7.75062621e-01 6.75470233e-01
-6.98747113e-02 -7.60199189e-01 -1.25890791e+00 -7.21606135e-01
3.02633911e-01 -1.16407908e-01 2.22093225e-01 1.15837944e+00
-1.41679987e-01 -7.56446794e-02 -1.54842481e-01 -1.91272318e-01
7.80164659e-01 1.23711132e-01 4.03963625e-01 -1.32438385e+00
-2.50395149e-01 -1.89139992e-01 -5.84528983e-01 -5.08454263e-01
3.85136992e-01 -1.04824555e+00 2.73329079e-01 -1.14151478e+00
3.66434515e-01 -7.56110430e-01 -2.62759745e-01 4.02342856e-01
2.63616890e-02 7.08309174e-01 3.32348119e-03 5.44815958e-01
-5.14195740e-01 -1.16367573e-02 1.24896264e+00 -9.55218151e-02
-1.20708071e-01 2.23178700e-01 -1.31930068e-01 8.11652005e-01
8.21195841e-01 -7.66103864e-01 -3.15578103e-01 -1.34305790e-01
-2.39307091e-01 1.42492101e-01 3.73123318e-01 -9.59971726e-01
3.40148330e-01 1.13171123e-01 -8.07377994e-02 -2.59894431e-01
1.56550333e-01 -8.49711835e-01 1.58345833e-01 4.58284527e-01
-3.97263765e-01 1.00147828e-01 -8.89368281e-02 3.97281289e-01
-3.09516430e-01 -5.78192174e-01 1.05464756e+00 -3.79998237e-01
-3.67916912e-01 4.33917195e-01 -3.96774620e-01 2.32701689e-01
1.19048035e+00 -1.62037849e-01 4.80706990e-01 1.48786120e-02
-1.30991018e+00 -1.62338853e-01 4.76009637e-01 -1.92986414e-01
2.85934567e-01 -1.09494948e+00 -5.49158335e-01 -1.66772559e-01
-7.71731064e-02 3.08966428e-01 1.35328785e-01 1.53777313e+00
-5.03489137e-01 2.33928114e-01 -3.37195516e-01 -9.41031694e-01
-1.33254611e+00 5.08134604e-01 3.52615714e-01 -5.71980953e-01
-6.69140816e-01 4.04759109e-01 2.20254809e-01 -4.45399553e-01
-1.89216137e-01 -3.42708915e-01 -2.99969137e-01 2.10634500e-01
1.53043732e-01 1.58193529e-01 4.34137255e-01 -1.12015736e+00
-2.19890535e-01 8.46808374e-01 -1.61884472e-01 -2.55294889e-01
1.41795516e+00 -7.62034133e-02 -3.76671582e-01 4.52578694e-01
1.22399545e+00 -1.17158972e-01 -9.20261264e-01 -4.06713516e-01
4.47432578e-01 -1.95413023e-01 4.68441136e-02 -4.56384331e-01
-1.29734766e+00 9.17466879e-01 7.84229457e-01 1.17323503e-01
1.18478131e+00 1.55894220e-01 6.67005301e-01 1.57766387e-01
6.17788196e-01 -9.74057138e-01 -1.55671746e-01 1.03588000e-01
5.51428020e-01 -1.41512823e+00 1.84641376e-01 -5.91296673e-01
-3.77713472e-01 8.78636599e-01 1.01882346e-01 -1.76015630e-01
6.85170054e-01 2.14073405e-01 2.11272314e-01 -3.44880372e-01
4.03805971e-02 -2.43757382e-01 3.48986626e-01 3.47929001e-01
3.62270176e-01 1.42930737e-02 -5.90988278e-01 2.66085625e-01
-3.50518636e-02 -2.86544021e-02 4.51645195e-01 1.02867985e+00
-2.09442139e-01 -1.57653546e+00 -4.03872579e-01 3.90572071e-01
-7.42192268e-01 -5.75804375e-02 -1.19345151e-01 8.80880177e-01
-6.15929775e-02 5.41197777e-01 -1.49023846e-01 2.22912818e-01
1.72957927e-01 1.29399106e-01 7.32707798e-01 -8.44725192e-01
-8.47632587e-01 3.35639387e-01 -3.65086883e-01 -5.08185327e-01
-1.09685934e+00 -8.88005495e-01 -1.21211827e+00 3.49169701e-01
-3.59358996e-01 1.70340464e-01 7.14096427e-01 1.45132577e+00
-2.20522493e-01 2.83472359e-01 5.23733735e-01 -1.05713642e+00
-3.07329237e-01 -7.72026598e-01 -7.87542820e-01 4.61678386e-01
1.49100155e-01 -6.36940181e-01 2.50530206e-02 3.31293434e-01] | [14.352895736694336, -2.389650821685791] |
8d71eb5d-8ff9-43a2-a3c9-e2602aa718c0 | hybridsdf-combining-free-form-shapes-and | 2109.10767 | null | https://arxiv.org/abs/2109.10767v4 | https://arxiv.org/pdf/2109.10767v4.pdf | HybridSDF: Combining Deep Implicit Shapes and Geometric Primitives for 3D Shape Representation and Manipulation | Deep implicit surfaces excel at modeling generic shapes but do not always capture the regularities present in manufactured objects, which is something simple geometric primitives are particularly good at. In this paper, we propose a representation combining latent and explicit parameters that can be decoded into a set of deep implicit and geometric shapes that are consistent with each other. As a result, we can effectively model both complex and highly regular shapes that coexist in manufactured objects. This enables our approach to manipulate 3D shapes in an efficient and precise manner. | ['Pierre Baqué', 'Pascal Fua', 'Jonathan Donier', 'Artem Lukoianov', 'Nicolas Talabot', 'Subeesh Vasu'] | 2021-09-22 | null | null | null | null | ['3d-shape-representation'] | ['computer-vision'] | [-1.82211936e-01 -4.01358008e-02 1.62753016e-01 -2.96188802e-01
-1.78895429e-01 -7.91385174e-01 7.83096194e-01 -6.03002869e-02
4.43150282e-01 3.58624578e-01 5.75500131e-02 -5.57890125e-02
-2.61907518e-01 -1.38630164e+00 -7.84879804e-01 -5.28209567e-01
1.24067381e-01 8.34751308e-01 -5.42331859e-02 -4.03760940e-01
5.77282012e-02 1.15704715e+00 -1.54030657e+00 3.90409529e-01
7.98682153e-01 8.76139343e-01 2.60334671e-01 2.61642814e-01
-5.57351172e-01 3.21261346e-01 -1.93681166e-01 -3.89373809e-01
3.38022411e-01 1.95030183e-01 -2.47129798e-01 2.18545452e-01
4.05708164e-01 -4.52446431e-01 -7.95903727e-02 8.85457218e-01
2.68242836e-01 -9.90360007e-02 1.18960655e+00 -7.77084351e-01
-1.07339346e+00 2.98222095e-01 -4.50121909e-01 -7.13742435e-01
1.05530709e-01 3.52447033e-01 6.86660051e-01 -1.06860971e+00
4.71747249e-01 1.73971581e+00 8.09469521e-01 5.20249903e-01
-1.39629316e+00 -5.01468599e-01 1.41234905e-01 -6.75057411e-01
-1.42459333e+00 -4.81244683e-01 1.27949154e+00 -4.63063300e-01
7.27706313e-01 4.69505936e-01 8.20057690e-01 8.28216851e-01
3.01685601e-01 6.88555837e-01 8.16274107e-01 -1.30205587e-01
-5.51555306e-02 -1.24521792e-01 -3.26591991e-02 6.24669611e-01
3.70039195e-01 1.58946022e-01 -1.14422701e-02 -1.52728558e-01
1.67019761e+00 2.41382003e-01 1.06644124e-01 -6.90661669e-01
-1.11081231e+00 6.42256379e-01 4.51780587e-01 3.64587098e-01
-6.26511574e-01 7.19414711e-01 -6.24696575e-02 9.70079470e-03
4.55097228e-01 5.67697346e-01 -4.71447885e-01 -2.57297438e-02
-6.30963206e-01 5.75629294e-01 6.03861809e-01 1.32917738e+00
9.01051342e-01 4.27242160e-01 4.09620516e-02 7.91588485e-01
6.11075580e-01 8.40075433e-01 -2.06389010e-01 -1.11752450e+00
-9.65420604e-02 8.56489003e-01 2.13501140e-01 -1.37302673e+00
-2.18460515e-01 -4.98610288e-01 -9.60573196e-01 4.79953825e-01
-1.20596096e-01 3.02326947e-01 -1.10288227e+00 1.59298229e+00
1.71697110e-01 -2.27603726e-02 -3.33497941e-01 6.56919956e-01
9.70692456e-01 6.95881605e-01 9.89088789e-02 3.35658312e-01
1.10364676e+00 -5.99090636e-01 -8.31267774e-01 5.15655577e-02
3.38110775e-01 -8.16329360e-01 9.83635008e-01 2.22370848e-01
-1.59442222e+00 -6.16254389e-01 -9.59593534e-01 -4.43272263e-01
-3.93696427e-01 2.90320124e-02 1.07526374e+00 4.85038131e-01
-1.12113118e+00 5.93169749e-01 -8.11477900e-01 4.02401835e-01
6.74524546e-01 4.72409010e-01 -1.17101394e-01 1.92859128e-01
-7.19500780e-01 7.96914577e-01 6.91523850e-02 2.03221992e-01
-6.55811489e-01 -9.17599678e-01 -7.90406764e-01 -5.51476143e-02
2.31931597e-01 -1.10054970e+00 1.16098928e+00 -8.06396842e-01
-1.74594462e+00 9.38752234e-01 -5.33385910e-02 7.18226656e-02
5.00652254e-01 -1.75771177e-01 -8.90207589e-02 -2.28368908e-01
-2.57421315e-01 6.85875893e-01 8.90059710e-01 -1.82322252e+00
1.74326256e-01 -2.83959240e-01 3.67856652e-01 6.03782805e-03
1.61126584e-01 -2.49983445e-01 -3.06519926e-01 -1.00541890e+00
4.84220117e-01 -5.84021688e-01 -3.49444628e-01 7.33116806e-01
-3.56958032e-01 -1.31203622e-01 8.92657816e-01 -2.32420638e-01
7.81454742e-01 -2.05107284e+00 4.78054374e-01 4.09676194e-01
4.59836602e-01 2.93002933e-01 -1.63880184e-01 4.03687268e-01
2.20648646e-01 4.84994918e-01 -3.38701725e-01 -6.32350564e-01
3.41468453e-01 6.17216885e-01 -4.69712406e-01 1.50923133e-01
3.49998176e-01 1.56739724e+00 -8.45760882e-01 -1.85252219e-01
3.26604903e-01 7.42431402e-01 -6.38539851e-01 2.04990536e-01
-7.87431300e-01 6.17305696e-01 -9.15050924e-01 6.27493978e-01
9.58473444e-01 -4.33733910e-02 3.08819078e-02 -3.15824747e-01
-2.27872983e-01 2.52171606e-01 -1.05070579e+00 1.71140707e+00
-5.52606881e-01 2.23689049e-01 4.32538569e-01 -6.02519572e-01
1.32349014e+00 2.96306640e-01 4.54618275e-01 -4.44600731e-01
2.00544611e-01 2.78834373e-01 -3.76627862e-01 -9.27776247e-02
4.45945531e-01 -4.22902972e-01 9.32416767e-02 4.38709855e-01
-2.98762351e-01 -9.33287561e-01 -4.25048620e-01 -2.38731384e-01
5.25194407e-01 3.67244005e-01 -1.57107990e-02 -5.53780079e-01
2.83508956e-01 -3.76773953e-01 2.77252465e-01 5.20433307e-01
6.37743294e-01 6.33010149e-01 2.43389383e-01 -6.95335805e-01
-1.28006244e+00 -1.57009649e+00 -2.95136422e-01 4.03921962e-01
1.63172528e-01 -3.18983048e-01 -4.70818818e-01 -1.44707084e-01
2.51969934e-01 6.25660002e-01 -7.54319847e-01 -7.02958554e-02
-8.25227022e-01 6.70610294e-02 2.25652069e-01 4.69345748e-01
3.04265112e-01 -9.52656448e-01 -5.52663684e-01 2.32458293e-01
2.78992206e-01 -8.81531417e-01 -3.81660461e-01 -2.35619977e-01
-1.15062273e+00 -7.17388213e-01 -6.26612961e-01 -7.79676855e-01
8.16117823e-01 2.23970547e-01 1.34837508e+00 4.73720700e-01
-1.04946904e-01 3.21776569e-01 -5.59234060e-03 -5.11402071e-01
-6.22677028e-01 -2.19383910e-01 -1.97277218e-01 5.75166903e-02
-2.05173865e-01 -1.05755997e+00 -3.45407367e-01 1.39423400e-01
-1.12451887e+00 5.82296193e-01 5.08739948e-01 3.60770971e-01
9.70257759e-01 -5.19222096e-02 2.31955171e-01 -7.96626627e-01
4.36407566e-01 -2.77565449e-01 -5.12898207e-01 1.32407248e-01
4.09898013e-02 2.13127956e-01 5.41708946e-01 -5.46136737e-01
-9.26208615e-01 -4.67845537e-02 -4.28261399e-01 -6.07277811e-01
-2.64088482e-01 5.81278503e-01 -4.28621620e-01 -1.86569750e-01
2.77924120e-01 1.35651631e-02 2.70709340e-02 -9.72398877e-01
7.14184225e-01 2.34445497e-01 3.50581616e-01 -1.09190094e+00
1.01494551e+00 6.01217687e-01 3.76987755e-01 -9.35370147e-01
-4.43584621e-01 1.88488603e-01 -9.37409759e-01 -4.46359627e-02
6.58054709e-01 -4.88614053e-01 -3.59103441e-01 6.96892917e-01
-1.53252518e+00 -4.33104694e-01 -7.25388408e-01 -1.28452897e-01
-9.08721805e-01 2.49639004e-01 -4.68308240e-01 -8.01048994e-01
-1.26329646e-01 -1.13155699e+00 1.37845528e+00 -2.10342482e-01
-3.49528104e-01 -1.05974591e+00 -2.54781306e-01 -3.74962181e-01
6.38669670e-01 7.39914358e-01 1.26874650e+00 2.78469831e-01
-1.12335527e+00 -1.18989572e-01 -1.01627298e-01 2.84113009e-02
7.20582068e-01 3.12692434e-01 -5.76448500e-01 1.70344058e-02
-2.07008477e-02 1.32296234e-01 4.61043388e-01 3.68272603e-01
1.44982862e+00 -3.56750786e-01 -3.88597518e-01 9.63594675e-01
1.30672777e+00 7.94847012e-02 8.15035522e-01 -1.89654842e-01
8.94008994e-01 5.99449515e-01 1.29016280e-01 2.89162934e-01
2.36693248e-01 8.19161952e-01 4.67435598e-01 -1.37277842e-01
-2.58602679e-01 -4.55010056e-01 -7.37637430e-02 1.04576373e+00
-2.63593972e-01 -6.70952490e-03 -9.60480690e-01 3.80214065e-01
-1.58658957e+00 -7.22021282e-01 -5.40385067e-01 1.77619398e+00
9.18405116e-01 8.89040437e-03 -1.95854887e-01 -1.60092805e-02
4.12618756e-01 6.21536896e-02 -3.16433907e-01 -5.75434506e-01
-3.53971601e-01 5.48432052e-01 1.77247867e-01 3.43788445e-01
-8.97510707e-01 1.01495242e+00 7.22156000e+00 8.16616774e-01
-1.20811653e+00 -1.58599541e-01 2.84558028e-01 2.57100224e-01
-1.19478047e+00 -2.12031037e-01 -2.82610983e-01 3.04595053e-01
4.96462971e-01 -2.04058886e-01 4.73839402e-01 4.72674549e-01
3.77307273e-02 5.06425977e-01 -1.14905632e+00 1.09420657e+00
-4.99619782e-01 -1.78988767e+00 8.62569690e-01 1.37115195e-01
1.02524340e+00 -6.13781333e-01 5.95184028e-01 -1.49581805e-01
3.73498648e-01 -1.27470517e+00 1.20325935e+00 1.04774547e+00
1.01263022e+00 -6.26340866e-01 -2.25977786e-02 5.77122629e-01
-1.24848962e+00 2.19680190e-01 -4.58695263e-01 -8.57106373e-02
3.02802026e-01 5.67722797e-01 -2.83694953e-01 6.22238159e-01
3.74617934e-01 7.48239517e-01 -5.61377741e-02 9.23295379e-01
-3.00947517e-01 1.76924989e-01 -4.81777698e-01 7.62815773e-02
2.22905204e-01 -3.86587650e-01 6.62502050e-01 8.72600615e-01
4.68825281e-01 4.58797514e-01 -3.02829184e-02 1.72269273e+00
-3.13180573e-02 -5.22415414e-02 -1.03333020e+00 -2.32388243e-01
4.12925661e-01 8.72483909e-01 -6.74551785e-01 -2.35014468e-01
-1.62157372e-01 5.42837083e-01 2.96321690e-01 3.14960361e-01
-7.95501292e-01 2.57805884e-01 9.34080720e-01 2.33997270e-01
1.57865971e-01 -1.11564112e+00 -7.64894009e-01 -9.94137883e-01
-8.75733718e-02 -6.45733953e-01 -5.98077416e-01 -8.73413146e-01
-1.48898005e+00 2.54044920e-01 1.23315901e-01 -1.15959859e+00
1.76936433e-01 -7.61194468e-01 -4.49643403e-01 9.53034341e-01
-1.27387631e+00 -1.67982745e+00 -1.53545231e-01 3.66384476e-01
2.61454791e-01 3.10984492e-01 9.70823586e-01 2.08067566e-01
3.22974622e-02 3.25064868e-01 -4.57877740e-02 1.90595314e-02
-7.08951280e-02 -1.00485277e+00 8.93029213e-01 3.26355934e-01
1.06371626e-01 1.05838001e+00 5.29367924e-01 -7.59902954e-01
-1.66647875e+00 -1.14858329e+00 2.27599502e-01 -6.52300775e-01
2.40369707e-01 -3.59875917e-01 -1.10279930e+00 7.46227503e-01
-3.40886176e-01 -2.85124004e-01 3.18308115e-01 -1.49739042e-01
-3.88373673e-01 3.69457036e-01 -1.13957071e+00 8.74885619e-01
1.50649369e+00 -6.18627846e-01 -7.00141251e-01 4.08923477e-02
8.34532380e-01 -6.82245791e-01 -7.88142025e-01 7.17323363e-01
6.50173426e-01 -7.46778965e-01 1.47528434e+00 -7.03988254e-01
4.92882222e-01 -7.37103298e-02 -3.23112130e-01 -1.11190689e+00
-6.26170576e-01 -7.53538251e-01 -5.21742344e-01 1.08935380e+00
1.21223286e-01 -7.85902739e-01 5.14880300e-01 7.53635168e-01
-4.17478055e-01 -8.66425276e-01 -6.39203250e-01 -9.00332093e-01
6.15818739e-01 -4.82521653e-01 1.43403566e+00 1.07251275e+00
-9.27939355e-01 -2.34993055e-01 -1.72231987e-01 9.88720357e-03
7.34302998e-01 5.11785805e-01 7.17841089e-01 -1.54934216e+00
-9.13797095e-02 -9.01483476e-01 -3.82484734e-01 -1.54302466e+00
2.34626591e-01 -9.89653647e-01 -8.51698220e-02 -1.60013485e+00
-1.86271533e-01 -1.23745871e+00 1.30697293e-02 3.56661707e-01
2.28011698e-01 2.44003624e-01 5.53573444e-02 8.70450065e-02
5.61477318e-02 8.54689062e-01 1.90738153e+00 -3.07144225e-01
-2.03572065e-01 -2.48839796e-01 -7.25889921e-01 8.30731094e-01
8.61106217e-01 -7.16055781e-02 -1.50686935e-01 -1.05190897e+00
2.88307369e-01 4.00852226e-02 5.27229071e-01 -6.46277964e-01
-1.86970130e-01 -4.12835807e-01 2.79717714e-01 -7.36357629e-01
7.61771321e-01 -1.09610248e+00 7.99865365e-01 1.79997101e-01
-1.89104825e-01 -1.08115464e-01 4.65080589e-01 3.01926523e-01
1.22797810e-01 -5.00910506e-02 6.97549045e-01 -2.57000983e-01
-2.83512354e-01 7.75810003e-01 -1.95405379e-01 -4.21786964e-01
8.09327066e-01 -4.92040068e-01 -9.63023379e-02 -3.42255443e-01
-8.23052406e-01 -5.83598949e-02 1.06051826e+00 6.27953410e-01
8.75747263e-01 -1.80139267e+00 -5.47950625e-01 4.48874474e-01
-3.69634211e-01 4.69517380e-01 2.57362038e-01 2.63119102e-01
-8.20399821e-01 3.53112191e-01 -4.91121173e-01 -6.61956072e-01
-7.68679142e-01 6.09805822e-01 4.84721720e-01 2.33518884e-01
-6.92465186e-01 7.37084746e-01 6.84021711e-01 -7.13685334e-01
-2.68028706e-01 -6.93332732e-01 9.45315436e-02 -3.68496537e-01
6.74131513e-02 1.07495867e-01 -2.81014502e-01 -9.17557657e-01
-1.24910418e-02 1.19512761e+00 5.67520976e-01 2.02684760e-01
1.55956709e+00 1.26281887e-01 -5.11794806e-01 5.93857348e-01
1.13051212e+00 3.26107293e-01 -1.37450325e+00 -1.37265787e-01
-2.47652292e-01 -6.58583462e-01 9.09346268e-02 -5.05736768e-01
-1.13084030e+00 1.05173421e+00 1.12369068e-01 3.10065180e-01
7.70292938e-01 1.95665464e-01 8.75732660e-01 2.01988623e-01
7.77227461e-01 -5.10566056e-01 1.44403726e-01 6.03108943e-01
1.50294256e+00 -4.07820612e-01 2.31841952e-02 -1.00281453e+00
1.06622308e-01 1.28585899e+00 2.57048845e-01 -5.00717819e-01
9.23135936e-01 5.09023547e-01 -2.24534854e-01 -5.95398009e-01
-5.57476580e-01 -1.01548396e-01 6.01070642e-01 6.91650093e-01
2.99677372e-01 3.62304032e-01 5.55353006e-03 5.39702654e-01
-3.04127127e-01 -1.85645476e-01 8.98427665e-02 1.06771743e+00
-2.98693269e-01 -1.42306137e+00 -2.17769116e-01 2.66063988e-01
-8.45418125e-02 2.29998648e-01 -4.49991763e-01 6.54274642e-01
2.13700950e-01 5.19864202e-01 1.97115868e-01 -2.44126558e-01
4.73472089e-01 -1.85037062e-01 9.77691650e-01 -8.21311235e-01
-3.02945465e-01 -2.81160139e-02 -1.24505877e-01 -5.80586731e-01
-2.19624862e-01 -3.98358911e-01 -1.30171752e+00 -4.97775555e-01
-7.96508715e-02 -2.81090736e-01 5.45481026e-01 5.88628232e-01
4.59384173e-01 5.53010046e-01 4.88691390e-01 -1.43321180e+00
-4.70031559e-01 -3.69976759e-01 -5.35395980e-01 4.66821313e-01
3.72016847e-01 -1.10863471e+00 5.06162979e-02 -5.09183519e-02] | [8.76565170288086, -3.635732889175415] |
a7162abf-3190-4dc5-9808-ad590f3691c2 | a-geometry-aware-deep-network-for-depth | 2304.10241 | null | https://arxiv.org/abs/2304.10241v1 | https://arxiv.org/pdf/2304.10241v1.pdf | A geometry-aware deep network for depth estimation in monocular endoscopy | Monocular depth estimation is critical for endoscopists to perform spatial perception and 3D navigation of surgical sites. However, most of the existing methods ignore the important geometric structural consistency, which inevitably leads to performance degradation and distortion of 3D reconstruction. To address this issue, we introduce a gradient loss to penalize edge fluctuations ambiguous around stepped edge structures and a normal loss to explicitly express the sensitivity to frequently small structures, and propose a geometric consistency loss to spreads the spatial information across the sample grids to constrain the global geometric anatomy structures. In addition, we develop a synthetic RGB-Depth dataset that captures the anatomical structures under reflections and illumination variations. The proposed method is extensively validated across different datasets and clinical images and achieves mean RMSE values of 0.066 (stomach), 0.029 (small intestine), and 0.139 (colon) on the EndoSLAM dataset. The generalizability of the proposed method achieves mean RMSE values of 12.604 (T1-L1), 9.930 (T2-L2), and 13.893 (T3-L3) on the ColonDepth dataset. The experimental results show that our method exceeds previous state-of-the-art competitors and generates more consistent depth maps and reasonable anatomical structures. The quality of intraoperative 3D structure perception from endoscopic videos of the proposed method meets the accuracy requirements of video-CT registration algorithms for endoscopic navigation. The dataset and the source code will be available at https://github.com/YYM-SIA/LINGMI-MR. | ['Hao liu', 'Chengdong Wu', 'Zhuo Yang', 'Peng Wang', 'Tao Yang', 'Shuwei Shao', 'Yongming Yang'] | 2023-04-20 | null | null | null | null | ['3d-reconstruction', 'monocular-depth-estimation', 'anatomy'] | ['computer-vision', 'computer-vision', 'miscellaneous'] | [-2.30632469e-01 2.05570981e-01 -1.26697019e-01 -1.93728656e-02
-6.84847236e-01 -7.06138313e-01 4.33363207e-02 1.65315211e-01
-2.89272726e-01 4.09550130e-01 3.07894230e-01 -5.27245164e-01
-1.96593732e-01 -5.19360065e-01 -7.15480864e-01 -8.24715436e-01
-1.83455497e-01 -1.93797946e-01 2.07182214e-01 2.12303195e-02
2.96982676e-01 2.48383835e-01 -8.65225613e-01 -4.18401137e-02
1.18342662e+00 1.16813970e+00 3.65113586e-01 3.85748476e-01
3.03679705e-01 2.15811566e-01 -6.65741116e-02 -8.86819139e-02
6.14304185e-01 -3.15859228e-01 -2.54535437e-01 -1.95347995e-01
1.81673974e-01 -3.30223978e-01 -3.30541313e-01 1.39296031e+00
7.41181731e-01 -2.00205985e-02 3.48389208e-01 -4.97593224e-01
-3.15915048e-01 1.87091790e-02 -7.79169500e-01 -1.70204174e-02
4.16183561e-01 1.85007781e-01 2.26748779e-01 -9.03229058e-01
8.27201664e-01 6.43476307e-01 1.04858518e+00 3.03769410e-01
-7.63176739e-01 -5.59735656e-01 2.37820774e-01 -2.32373357e-01
-1.32106078e+00 7.73837417e-03 4.96282488e-01 -4.15974051e-01
6.39780462e-01 1.78139061e-01 8.77694488e-01 5.83859801e-01
8.74412179e-01 3.38822991e-01 1.12255216e+00 -3.43069851e-01
8.86249542e-03 4.94529009e-02 -1.76637962e-01 1.02235603e+00
7.66942024e-01 5.44294477e-01 -1.48889825e-01 6.67937705e-03
1.30537581e+00 8.42591897e-02 -9.36977029e-01 -6.47214293e-01
-1.35065043e+00 5.93489766e-01 8.51119876e-01 1.53807521e-01
-5.17933249e-01 -2.09600359e-01 3.45056534e-01 1.77075062e-02
-7.99435824e-02 3.75942647e-01 -1.34972453e-01 -1.02037087e-01
-3.77073944e-01 -2.44785026e-01 6.97521329e-01 1.38526082e+00
2.80343115e-01 -2.95290887e-01 1.63064033e-01 5.99842906e-01
3.35464507e-01 1.94327563e-01 6.76190376e-01 -7.76125371e-01
3.34455997e-01 5.68418801e-01 1.92141891e-01 -9.23627377e-01
-7.43384480e-01 -6.93527102e-01 -9.36442196e-01 1.51687950e-01
2.71732390e-01 -1.17104627e-01 -9.49019611e-01 1.23044968e+00
5.47686815e-01 1.04546763e-01 -6.29232172e-03 1.17452776e+00
1.31219792e+00 1.90718263e-01 -3.74743491e-01 -2.69880712e-01
1.23231983e+00 -1.00795043e+00 -6.99732304e-01 -2.66728729e-01
8.49794745e-01 -1.00475430e+00 1.06472492e+00 4.83063698e-01
-1.28908658e+00 -2.94211149e-01 -9.98639762e-01 2.04122700e-02
1.99125171e-01 2.03673005e-01 4.25420761e-01 6.54300392e-01
-9.04625893e-01 2.30793864e-01 -1.14713514e+00 -2.39633679e-01
8.40967149e-02 3.71585727e-01 -4.76406962e-01 -3.43554646e-01
-8.04601312e-01 8.50303471e-01 1.05083629e-01 3.58395994e-01
-5.94633222e-01 -8.98751915e-01 -1.08333361e+00 -4.30923462e-01
2.30349854e-01 -8.70476723e-01 1.05643535e+00 -4.15233046e-01
-1.61177182e+00 9.45263565e-01 -4.11258359e-03 -7.86619112e-02
6.65090501e-01 -1.27674177e-01 -1.64917961e-01 2.81240374e-01
-7.96700045e-02 3.73239219e-01 9.53295976e-02 -1.33004224e+00
-4.43360984e-01 -6.11500382e-01 1.63698718e-01 6.25298381e-01
4.84139323e-02 -4.73649114e-01 -6.57650650e-01 -6.67410493e-01
8.32770348e-01 -1.20057368e+00 -4.98838127e-01 5.33781171e-01
-4.90923792e-01 7.74703622e-01 3.46342325e-02 -7.09057689e-01
1.26711965e+00 -2.33416700e+00 -3.85536999e-02 3.00780624e-01
2.91176327e-02 -9.63399857e-02 2.32320875e-01 7.15439171e-02
7.31249005e-02 -1.56670064e-01 -7.54948854e-02 -1.24049619e-01
-3.58218342e-01 1.04052275e-02 4.80748177e-01 9.52624500e-01
-7.85012603e-01 8.56530368e-01 -9.80853260e-01 -4.03855741e-01
5.27682245e-01 5.94731450e-01 -7.78609395e-01 4.29035276e-02
4.43240404e-01 8.91379654e-01 -5.15374959e-01 8.59588623e-01
1.08911407e+00 -1.87268451e-01 2.22002417e-01 -6.24517322e-01
-4.04990971e-01 3.77600700e-01 -1.15905595e+00 2.38370657e+00
-6.83063626e-01 8.09427574e-02 4.05429780e-01 -4.03146565e-01
7.36672699e-01 3.41861904e-01 6.33289993e-01 -9.09291089e-01
2.86401331e-01 6.61138773e-01 2.61903882e-01 -5.10219455e-01
9.25088450e-02 -2.28014901e-01 2.63189375e-02 -1.57713339e-01
-2.76731342e-01 -4.56024557e-01 -2.34946728e-01 -2.49690250e-01
6.88197136e-01 1.59293879e-02 5.57724357e-01 -6.32826030e-01
4.26421583e-01 1.63723573e-01 6.18389010e-01 4.31188405e-01
-3.12199980e-01 7.46952057e-01 2.49017522e-01 -4.28052664e-01
-7.35213935e-01 -1.21925199e+00 -6.23024166e-01 -1.60995156e-01
1.04767287e+00 -2.57780850e-02 -5.70317626e-01 -4.98900980e-01
-7.64095411e-02 4.62690264e-01 -5.33301294e-01 -2.65583128e-01
-4.30253118e-01 -6.61179900e-01 5.68941012e-02 4.26455021e-01
3.63502234e-01 -5.04409432e-01 -7.23428249e-01 1.06012665e-01
-3.33657354e-01 -1.04311013e+00 -6.01613522e-01 5.82263954e-02
-1.25693858e+00 -1.20644832e+00 -6.78072214e-01 -1.02796566e+00
1.10267138e+00 3.73310387e-01 8.19481850e-01 -3.28809582e-02
-3.82195950e-01 1.99760169e-01 -3.02947134e-01 -2.42665723e-01
-1.37186408e-01 -1.25939429e-01 -5.48459664e-02 -5.60198545e-01
-1.09771654e-01 -3.21233988e-01 -1.38729286e+00 7.49482274e-01
-6.68117166e-01 1.37050346e-01 5.40812194e-01 9.80342686e-01
8.43096137e-01 -1.92095891e-01 -1.94827706e-01 -5.84064424e-01
2.80272722e-01 -2.77248800e-01 -8.87664497e-01 -7.29760379e-02
-6.54609263e-01 -3.38454098e-01 5.05059123e-01 -3.84325445e-01
-1.01131654e+00 5.31952009e-02 -5.21839634e-02 -3.60603184e-01
3.65144499e-02 5.93982160e-01 2.78928638e-01 -4.42357630e-01
7.31384039e-01 1.28250986e-01 1.69997558e-01 -1.82940081e-01
-6.67407960e-02 2.11509943e-01 5.36366463e-01 -2.67630160e-01
4.91935641e-01 7.25695729e-01 6.08956702e-02 -5.67716002e-01
-5.24607837e-01 -5.83609939e-01 -4.46965694e-01 -8.00039619e-02
6.49482787e-01 -1.11961377e+00 -5.06585836e-01 2.45134979e-01
-7.11796641e-01 -1.62124008e-01 -1.42240077e-01 1.23152673e+00
-3.85674030e-01 5.35650790e-01 -8.13380420e-01 -3.17653984e-01
-4.33837950e-01 -1.46376836e+00 8.74031663e-01 3.05375755e-01
-5.47346584e-02 -1.33535504e+00 3.45776677e-02 -3.82031314e-02
3.27800930e-01 6.18304670e-01 6.52029634e-01 2.84416713e-02
-8.32981586e-01 -3.45018715e-01 -2.60720663e-02 6.80159554e-02
4.33570296e-01 -3.22913408e-01 -5.97229481e-01 -5.84956646e-01
3.42055708e-01 1.56486273e-01 4.76319134e-01 1.17359626e+00
9.14537787e-01 -1.11845016e-01 -4.19452667e-01 1.44635522e+00
1.66410708e+00 4.95125681e-01 7.75889277e-01 5.69565594e-01
4.44219977e-01 3.81446302e-01 9.59459662e-01 4.27565962e-01
3.75077188e-01 5.92060328e-01 7.59688258e-01 -3.82319152e-01
-1.86488684e-02 -2.15155035e-01 2.75365170e-02 9.48623002e-01
-4.86829802e-02 3.90182510e-02 -6.97375894e-01 4.08905119e-01
-1.31839907e+00 -4.01186794e-01 -2.44106799e-01 2.56068277e+00
5.89157999e-01 7.70799667e-02 -4.31140721e-01 -3.38369668e-01
3.57653469e-01 -1.61270916e-01 -4.62586403e-01 -1.04600288e-01
2.35391527e-01 -3.02973855e-03 8.33615184e-01 7.20652699e-01
-8.77784848e-01 3.32969368e-01 5.39780569e+00 3.01707357e-01
-1.33527184e+00 -1.19995177e-01 4.33277130e-01 -1.66160271e-01
-3.03762823e-01 -2.55349249e-01 -5.74764907e-01 4.33813483e-01
2.36100763e-01 -2.20920891e-01 5.93202636e-02 5.88280678e-01
3.00026596e-01 -4.48016316e-01 -8.70154023e-01 1.26246035e+00
7.22202584e-02 -1.13080442e+00 -3.58409256e-01 1.57050312e-01
9.94498610e-01 3.48702446e-02 2.62792975e-01 -1.83430642e-01
-3.09194952e-01 -1.00002456e+00 4.39932287e-01 3.96851182e-01
1.03175104e+00 -3.86499703e-01 8.75949025e-01 2.26350158e-01
-1.21334195e+00 5.30035682e-02 -2.86510676e-01 3.45056206e-01
2.42205262e-01 6.27809763e-01 -7.09863782e-01 7.98567951e-01
7.15693772e-01 7.55400479e-01 -4.92490232e-02 1.53988016e+00
-1.50759622e-01 -2.88735032e-01 -4.33335066e-01 1.72801822e-01
1.95544168e-01 -3.15855265e-01 5.26637137e-01 8.34947109e-01
7.88696766e-01 2.78642744e-01 -9.06205736e-03 5.04050851e-01
1.47707269e-01 2.68025666e-01 -4.15759712e-01 6.52685523e-01
4.38829482e-01 1.17962611e+00 -6.71819091e-01 -9.94892605e-03
-5.74627042e-01 8.83556962e-01 -2.96959937e-01 2.34805867e-01
-9.94310737e-01 -2.13984206e-01 4.71098751e-01 3.37104946e-01
1.56604014e-02 -2.01487213e-01 -5.90306282e-01 -1.25876451e+00
4.52172667e-01 -6.64428055e-01 4.06364202e-01 -7.21346796e-01
-8.29967201e-01 7.72242486e-01 -2.23269314e-01 -1.97007406e+00
4.24701087e-02 -9.42799926e-01 -4.22795445e-01 8.38704467e-01
-1.67426527e+00 -7.41772294e-01 -1.04219866e+00 4.47243601e-01
2.47405410e-01 3.72727960e-01 8.02886605e-01 3.06160569e-01
-1.60190910e-01 5.18470407e-01 2.03953817e-01 -2.15004519e-01
8.63700986e-01 -9.90849853e-01 1.81142278e-02 5.45452416e-01
-4.69643116e-01 8.75438213e-01 3.99134576e-01 -5.77622533e-01
-1.45393741e+00 -7.65266776e-01 4.31618124e-01 -3.12045872e-01
1.86265856e-01 1.76006909e-02 -6.37873292e-01 6.92871451e-01
-5.36538288e-02 2.79882491e-01 5.97678661e-01 -2.87762463e-01
8.76366049e-02 3.96316163e-02 -1.53481925e+00 5.78089833e-01
1.08173859e+00 -1.01489805e-01 -1.89301684e-01 1.92358375e-01
3.54145467e-01 -1.57099950e+00 -1.31889760e+00 7.65489101e-01
8.84571671e-01 -1.33505559e+00 1.15219092e+00 4.25378025e-01
2.66421139e-01 -4.67964083e-01 3.27979997e-02 -1.22450805e+00
-8.97761732e-02 -4.83977258e-01 2.82157123e-01 1.48494646e-01
4.92044449e-01 -9.59336579e-01 8.14560890e-01 5.16967237e-01
-8.08386445e-01 -8.92391562e-01 -9.84918952e-01 -6.19617164e-01
5.99748492e-02 -2.33353283e-02 1.06112167e-01 1.02132058e+00
2.73711652e-01 -5.50376236e-01 1.82169050e-01 5.57664990e-01
6.49755657e-01 2.79355139e-01 5.11671424e-01 -8.08494866e-01
-1.94196835e-01 -3.02793145e-01 -2.81722218e-01 -1.40846622e+00
-6.92340732e-01 -8.89802277e-01 -1.13448001e-01 -1.84598958e+00
3.13424170e-02 -6.20233536e-01 -2.79780149e-01 -4.89395410e-02
-4.98896912e-02 9.46974903e-02 -2.45574713e-01 2.59563625e-01
-8.70981440e-02 3.36424500e-01 2.00009608e+00 4.20693308e-01
-4.90711123e-01 3.55977342e-02 -5.85500598e-01 8.56051922e-01
6.69185519e-01 -2.41455153e-01 -4.64439392e-01 -5.02882898e-01
-1.70112289e-02 2.87454188e-01 3.55378747e-01 -8.63118470e-01
4.40741718e-01 2.16468096e-01 5.27794421e-01 -5.36461949e-01
2.98383772e-01 -1.06975663e+00 3.97739708e-01 9.90596592e-01
1.98760867e-01 1.56016231e-01 4.41138446e-01 2.15576097e-01
-4.14465427e-01 -6.39459267e-02 1.01081848e+00 -4.11866665e-01
-4.21845526e-01 4.58267510e-01 9.38465744e-02 1.29614592e-01
1.07499802e+00 -6.75721288e-01 -1.29214838e-01 -2.17373416e-01
-7.23870575e-01 1.20533206e-01 9.43801880e-01 1.19016632e-01
9.36983287e-01 -1.20797694e+00 -4.86144722e-01 5.68509936e-01
2.32500598e-01 5.83584666e-01 6.47183836e-01 1.64230227e+00
-1.26048326e+00 4.36267942e-01 -1.33399874e-01 -9.19610798e-01
-9.69876289e-01 4.06930745e-01 9.16210771e-01 -1.06850669e-01
-8.41708183e-01 9.15516853e-01 7.02058077e-01 -4.44194138e-01
1.80721626e-01 -9.39515412e-01 2.72515148e-01 -5.71171463e-01
1.61367700e-01 7.53384009e-02 2.07386807e-01 -2.82195985e-01
-4.73106474e-01 1.07728577e+00 -1.03836255e-02 3.56830448e-01
9.60919023e-01 -3.86991680e-01 1.70851499e-01 -4.00105119e-02
1.12618518e+00 3.30142260e-01 -1.24193609e+00 4.88230661e-02
-4.66819376e-01 -8.35062206e-01 1.66830033e-01 -7.82070816e-01
-1.05566859e+00 6.85547352e-01 8.23816299e-01 -3.64587843e-01
1.33294165e+00 -2.88944721e-01 7.83220053e-01 -3.35370272e-01
6.50921345e-01 -5.76614738e-01 -2.05225185e-01 3.07991743e-01
8.85243893e-01 -1.22783780e+00 1.62003189e-01 -9.26430881e-01
-5.44430614e-01 1.06617379e+00 5.75214982e-01 -1.15600824e-01
8.35522473e-01 5.08034706e-01 3.82776976e-01 -1.62178069e-01
-1.15433648e-01 3.41477334e-01 4.40879017e-01 3.51709306e-01
6.59729660e-01 -4.56321836e-02 -5.44570625e-01 2.95675665e-01
-2.12745458e-01 -1.48284826e-02 4.42426026e-01 1.01151848e+00
-1.06601618e-01 -7.41843462e-01 -2.13376492e-01 6.15790896e-02
-5.08311152e-01 -3.36322963e-01 4.59122568e-01 1.08415818e+00
-7.79417157e-02 6.39075935e-01 -1.59634724e-01 1.08298659e-01
8.47201705e-01 -6.35752201e-01 6.93902969e-01 -4.67300713e-01
-6.18565500e-01 5.75921118e-01 -7.70212635e-02 -7.00813413e-01
-2.96054929e-01 -3.96905303e-01 -1.57442594e+00 -3.09535656e-02
-3.97464424e-01 7.46377334e-02 7.53345907e-01 3.67124304e-02
3.10978442e-01 5.65041423e-01 5.56208611e-01 -7.08333135e-01
-2.83184826e-01 -6.52859390e-01 -6.00688875e-01 3.26083958e-01
4.74023879e-01 -7.03642666e-01 -6.93712890e-01 -1.65505022e-01] | [13.839280128479004, -3.10856294631958] |
66d0c4b6-ca26-4fab-8d37-73f5c2c7a08e | supervised-distributional-hypernym-discovery | null | null | https://aclanthology.org/D16-1041 | https://aclanthology.org/D16-1041.pdf | Supervised Distributional Hypernym Discovery via Domain Adaptation | null | ['Jose Camacho-Collados', 'Luis Espinosa-Anke', 'Claudio Delli Bovi', 'Horacio Saggion'] | 2016-11-01 | null | null | null | emnlp-2016-11 | ['hypernym-discovery'] | ['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] | [-7.234970569610596, 3.7888295650482178] |
ef34b710-d700-4353-83e9-a0f9d95d72e8 | modeling-radiometric-uncertainty-for-vision | 1311.6887 | null | http://arxiv.org/abs/1311.6887v2 | http://arxiv.org/pdf/1311.6887v2.pdf | Modeling Radiometric Uncertainty for Vision with Tone-mapped Color Images | To produce images that are suitable for display, tone-mapping is widely used
in digital cameras to map linear color measurements into narrow gamuts with
limited dynamic range. This introduces non-linear distortion that must be
undone, through a radiometric calibration process, before computer vision
systems can analyze such photographs radiometrically. This paper considers the
inherent uncertainty of undoing the effects of tone-mapping. We observe that
this uncertainty varies substantially across color space, making some pixels
more reliable than others. We introduce a model for this uncertainty and a
method for fitting it to a given camera or imaging pipeline. Once fit, the
model provides for each pixel in a tone-mapped digital photograph a probability
distribution over linear scene colors that could have induced it. We
demonstrate how these distributions can be useful for visual inference by
incorporating them into estimation algorithms for a representative set of
vision tasks. | ['Kate Saenko', 'Trevor Darrell', 'Baochen Sun', 'Daniel Scharstein', 'Ying Xiong', 'Todd Zickler', 'Ayan Chakrabarti'] | 2013-11-27 | null | null | null | null | ['tone-mapping'] | ['computer-vision'] | [ 7.30030358e-01 -2.15956554e-01 7.95068964e-02 -7.40107417e-01
-8.86391282e-01 -8.83116782e-01 5.27933240e-01 -4.16290373e-01
-4.18132007e-01 6.54355288e-01 -5.07305264e-02 -5.23920834e-01
1.01472192e-01 -6.38044536e-01 -1.03557491e+00 -4.13186371e-01
3.37986767e-01 2.23646060e-01 5.41811943e-01 4.64317203e-01
4.76786077e-01 5.69060564e-01 -1.47093058e+00 -5.03587238e-02
6.10874176e-01 7.05198824e-01 2.78746635e-01 9.46528912e-01
1.60199374e-01 8.50236654e-01 -8.26113164e-01 -5.19029081e-01
5.84540129e-01 -5.02064586e-01 -2.10747749e-01 3.70813817e-01
1.00271523e+00 -9.29488897e-01 -3.27915490e-01 1.49410796e+00
-1.16588503e-01 5.50298207e-02 7.43868947e-01 -1.11959279e+00
-5.91995478e-01 2.14804724e-01 -8.08635771e-01 1.73550770e-02
2.29436144e-01 2.40894467e-01 6.25876665e-01 -4.81547773e-01
5.95261574e-01 1.05352545e+00 4.71175492e-01 7.16097802e-02
-1.55381048e+00 -6.19946599e-01 -2.95203120e-01 -5.00905067e-02
-1.44228625e+00 -5.32981157e-01 6.34082139e-01 -6.23662591e-01
6.04021251e-01 4.19977605e-01 6.41706407e-01 5.14077306e-01
4.93180573e-01 2.67880652e-02 1.62465298e+00 -6.34550989e-01
2.31886834e-01 3.56929749e-01 -2.32120365e-01 6.49694502e-01
4.79317725e-01 4.65001583e-01 -4.81871426e-01 7.32232910e-03
1.12568307e+00 -4.66997594e-01 -3.60967636e-01 -3.17311794e-01
-7.63893366e-01 3.28831553e-01 3.59623641e-01 -4.35618669e-01
1.34174287e-01 4.93880749e-01 -2.16261417e-01 2.05076020e-02
2.68370569e-01 6.14169717e-01 -3.74277448e-03 -7.27326572e-02
-9.42120612e-01 -1.35996819e-01 4.82508272e-01 1.02792096e+00
8.56639206e-01 -3.86034772e-02 1.91180393e-01 7.38053322e-01
2.49340281e-01 6.38154447e-01 -1.13047868e-01 -1.60016596e+00
1.96750499e-02 -2.87416615e-02 3.71833146e-01 -7.81917214e-01
-1.18000291e-01 1.63037941e-01 -1.10535495e-01 1.19872451e+00
8.98589551e-01 -8.37794095e-02 -1.15317857e+00 1.60333669e+00
6.32196814e-02 6.08019456e-02 -3.98146540e-01 1.10776961e+00
-1.52183354e-01 6.17272973e-01 -1.42809510e-01 -2.24907249e-01
1.27545869e+00 -3.17203373e-01 -5.12654364e-01 -3.89988661e-01
-2.17579722e-01 -1.35924804e+00 1.26600957e+00 8.93779039e-01
-1.04746878e+00 -3.99819076e-01 -1.53841078e+00 -4.90755647e-01
1.38041288e-01 1.76792368e-02 4.25209582e-01 1.00941193e+00
-1.04174423e+00 5.33211231e-01 -6.65847778e-01 -1.87611938e-01
1.58638388e-01 6.68900833e-03 -9.75357555e-03 -2.39079684e-01
-6.21508718e-01 1.28817105e+00 -4.58105430e-02 -1.36613831e-01
-6.71496153e-01 -6.14241958e-01 -6.19457304e-01 -1.95481002e-01
3.63656998e-01 -4.53774601e-01 1.40380073e+00 -1.13409317e+00
-1.55407643e+00 1.07606852e+00 -8.84606838e-02 -2.42803648e-01
6.17907882e-01 -8.45000148e-02 -3.68939221e-01 2.27711201e-01
-5.08565530e-02 5.97995698e-01 1.25030553e+00 -1.39894271e+00
-4.29464549e-01 -2.17111886e-01 8.04872587e-02 2.19608128e-01
1.29257634e-01 1.34775698e-01 -9.64942634e-01 -4.12914425e-01
1.02915064e-01 -1.11393106e+00 8.23766515e-02 6.70082510e-01
-5.18214881e-01 5.51902831e-01 3.77809525e-01 -5.40907979e-01
7.49775052e-01 -2.17774844e+00 -3.66395265e-01 2.38155663e-01
1.05125591e-01 -1.53149724e-01 1.00500293e-01 -4.34890762e-02
1.65102065e-01 -2.91183926e-02 -2.56004244e-01 -1.52465791e-01
-1.13972798e-01 -3.86387333e-02 -4.24238950e-01 7.56666601e-01
2.28831410e-01 2.61251092e-01 -6.16542220e-01 -3.63984406e-01
6.24368489e-01 4.77184623e-01 -1.31931633e-01 2.08211124e-01
-3.09872657e-01 -9.15564783e-03 1.21455692e-01 4.71745789e-01
1.10996640e+00 1.53264374e-01 2.99349844e-01 -4.19056654e-01
-3.61576706e-01 5.15502319e-02 -9.71158087e-01 1.16820729e+00
-3.80064219e-01 1.26786137e+00 -1.32971868e-01 9.02795941e-02
9.09695864e-01 -5.15438914e-01 6.05129302e-02 -5.70795655e-01
2.03503072e-01 -1.25702888e-01 2.01520193e-02 -2.42172867e-01
8.93711090e-01 -4.26781297e-01 2.02804580e-01 6.64752960e-01
-2.93736577e-01 -8.85318756e-01 -6.26495779e-02 9.96299982e-02
7.43248582e-01 2.44950384e-01 9.21937376e-02 -2.02944160e-01
-1.14234053e-01 2.03082293e-01 3.88864577e-01 6.69732690e-01
-1.57815143e-01 1.12196398e+00 4.83629972e-01 -7.02075064e-02
-1.59246194e+00 -1.37906027e+00 -4.45901930e-01 6.13894105e-01
3.48582625e-01 -8.17819219e-03 -8.13839793e-01 1.21973418e-01
8.42930563e-03 9.91695106e-01 -4.54922587e-01 -2.54666150e-01
-1.22292414e-01 -5.58085561e-01 4.52384025e-01 4.42031115e-01
2.98786014e-01 -3.97419274e-01 -7.97810912e-01 -1.36377186e-01
2.40324363e-01 -1.03937340e+00 -7.40613580e-01 1.36881992e-01
-5.33140421e-01 -1.26440394e+00 -3.26842427e-01 -1.97142109e-01
8.28678846e-01 7.15251446e-01 1.19916368e+00 -1.29660130e-01
-6.48280978e-01 7.31656909e-01 8.73206009e-04 -4.63103503e-01
-4.95864481e-01 -8.11004400e-01 -1.28500536e-01 -1.12474494e-01
4.23934489e-01 -7.76711255e-02 -4.69281256e-01 4.23355669e-01
-8.14940333e-01 1.20810196e-01 4.60654646e-01 4.24763620e-01
6.76916301e-01 2.76580542e-01 -3.55573654e-01 -9.24665987e-01
5.21150053e-01 -5.68663478e-02 -1.43101490e+00 2.99284995e-01
-6.23007178e-01 1.01739742e-01 4.14944202e-01 -3.51178080e-01
-1.42991364e+00 2.74840474e-01 6.27482712e-01 -4.58520472e-01
4.86711226e-02 7.75427222e-02 -1.76718365e-02 -3.39970738e-01
9.45910096e-01 -1.35714456e-01 6.65907487e-02 -1.12587109e-01
6.54514313e-01 5.39299250e-01 1.01625657e+00 -6.30314291e-01
9.01093423e-01 6.94395721e-01 1.97552055e-01 -9.22587693e-01
-5.99870861e-01 -1.57651514e-01 -3.50450099e-01 -5.34236670e-01
9.68876362e-01 -9.27547753e-01 -4.88882184e-01 4.93357182e-01
-8.37240696e-01 -5.08996189e-01 -6.27867132e-02 7.06740618e-01
-5.09987473e-01 2.35288650e-01 -3.59711409e-01 -9.48959947e-01
5.98106802e-01 -1.12154841e+00 7.36150622e-01 5.03124356e-01
-1.43821478e-01 -6.94726765e-01 5.22759557e-03 2.14395925e-01
1.53061047e-01 -6.30587863e-04 6.93915308e-01 5.22872865e-01
-1.07344258e+00 -1.88716784e-01 -6.08940661e-01 4.71113801e-01
4.16525528e-02 7.41353571e-01 -1.38697410e+00 -9.11371186e-02
4.81418297e-02 -2.41908148e-01 8.15830946e-01 7.37509429e-01
1.14485288e+00 2.36431316e-01 1.86149999e-02 9.13341403e-01
1.74022305e+00 2.80841351e-01 9.44224358e-01 4.06360269e-01
5.10942876e-01 4.82837081e-01 6.24869287e-01 1.95178941e-01
1.40148774e-01 6.64945006e-01 2.88030654e-01 -9.82574821e-02
-2.91363209e-01 -2.82883406e-01 3.55682254e-01 7.15544000e-02
3.42088528e-02 -1.94003880e-01 -5.71610570e-01 1.47666991e-01
-1.20860946e+00 -8.62534106e-01 -2.57824749e-01 2.74816132e+00
9.06051040e-01 4.10815865e-01 3.00542340e-02 -4.02689904e-01
9.69198883e-01 -6.68471158e-02 -8.61193836e-01 -5.04877925e-01
-2.00928643e-01 -6.39139563e-02 1.29659653e+00 9.06041086e-01
-7.82212436e-01 6.68060541e-01 7.86919308e+00 2.90884584e-01
-1.07870138e+00 -3.22391719e-01 7.77016103e-01 -1.71140969e-01
-4.20765698e-01 3.20434481e-01 -5.34488380e-01 4.43445116e-01
7.98084736e-01 -3.49551469e-01 8.15743089e-01 6.17486775e-01
2.98315197e-01 -9.22827423e-01 -1.12503862e+00 9.71113026e-01
2.71934479e-01 -9.46529210e-01 -2.95872599e-01 1.16557941e-01
6.93587899e-01 -1.02422692e-01 4.64335322e-01 -5.11372924e-01
8.67354929e-01 -9.22723889e-01 9.30745482e-01 7.85319626e-01
1.33859813e+00 -5.36465168e-01 1.02493480e-01 -2.62433589e-01
-5.08505642e-01 1.99306428e-01 -8.89513612e-01 -8.54164660e-02
1.79055154e-01 7.27329910e-01 -1.06738043e+00 -1.56090289e-01
4.97282982e-01 1.12460636e-01 -7.75639296e-01 1.22178209e+00
-4.04795885e-01 3.82286906e-01 -4.34345871e-01 2.35940531e-01
-2.98122644e-01 -6.82379723e-01 2.53432482e-01 8.71128082e-01
5.88583529e-01 8.42519999e-02 -3.16047639e-01 1.14900410e+00
-1.17537931e-01 -5.80935717e-01 -6.30376577e-01 1.72380909e-01
9.27563488e-01 1.07115471e+00 -7.92086244e-01 -2.38809556e-01
-5.61211228e-01 1.17631447e+00 -2.82441527e-02 4.88733858e-01
-8.59108865e-01 -4.28762168e-01 6.96721792e-01 1.22694373e-01
-3.62408906e-02 -4.05871421e-01 -6.58479750e-01 -1.02668333e+00
8.65524858e-02 -5.82708418e-01 -3.14233415e-02 -1.69181323e+00
-1.14535689e+00 3.53385001e-01 1.41899332e-01 -1.39273059e+00
-1.67603970e-01 -8.88361096e-01 -3.48063082e-01 1.11333573e+00
-1.32544410e+00 -8.86989415e-01 -3.72332036e-01 4.15962100e-01
2.56424844e-01 1.13874793e-01 4.95803237e-01 -1.13590047e-01
-1.97454199e-01 4.21690404e-01 2.96002805e-01 -8.75944942e-02
1.23614240e+00 -1.55043542e+00 4.89896595e-01 1.30221343e+00
2.07004890e-01 6.77342534e-01 1.06680286e+00 -5.27408242e-01
-1.30760074e+00 -7.40094721e-01 8.58517811e-02 -8.44864130e-01
5.96441209e-01 -2.02487946e-01 -8.26580107e-01 7.90799022e-01
1.78354964e-01 -1.53955847e-01 5.59718311e-01 -9.40260738e-02
-7.40913689e-01 -2.58505374e-01 -1.08891416e+00 6.19965553e-01
5.28608382e-01 -9.37153518e-01 -4.18515891e-01 1.18597433e-01
3.79384667e-01 -5.06228447e-01 -5.64497769e-01 -3.18237662e-01
9.52410698e-01 -1.34651077e+00 8.86440516e-01 -6.00832254e-02
3.49350572e-01 -7.94891894e-01 -2.03803018e-01 -1.40928829e+00
-1.95245624e-01 -5.42975783e-01 6.39316261e-01 1.12398171e+00
3.44064265e-01 -4.34497505e-01 4.46139097e-01 1.27058923e+00
3.08245625e-02 2.48520687e-01 -6.00681841e-01 -7.43161082e-01
4.45424914e-02 -6.68238580e-01 3.40613663e-01 6.78492010e-01
-3.42930079e-01 -1.00014046e-01 -6.61472678e-01 4.04309779e-01
1.20151174e+00 -1.90919533e-01 8.29975665e-01 -9.34551239e-01
-4.14599270e-01 -2.52053410e-01 -5.07593989e-01 -9.24217522e-01
-3.41693610e-01 -2.62501210e-01 3.76255363e-01 -8.80978048e-01
3.28844488e-01 -5.04618943e-01 -1.35180531e-02 -4.40824032e-02
-1.63635209e-01 6.74290359e-01 3.99400413e-01 2.64414936e-01
-1.33582607e-01 -1.75471470e-01 1.00335658e+00 9.28120390e-02
4.70847683e-03 -3.23667556e-01 -8.25346708e-01 8.07476521e-01
5.92395782e-01 -4.72637057e-01 -5.72117746e-01 -7.49766827e-01
4.07390654e-01 -9.77796968e-03 5.10337770e-01 -1.02727783e+00
3.44609976e-01 -4.28575963e-01 8.29416335e-01 -2.88826585e-01
4.34126228e-01 -9.51611936e-01 4.33909118e-01 3.43417749e-02
-3.25794756e-01 -2.50747293e-01 2.72378504e-01 5.50091445e-01
1.00155532e-01 -3.65584731e-01 1.13654053e+00 -3.01928930e-02
-8.03097427e-01 -1.83411255e-01 -5.33209682e-01 -1.52856603e-01
9.87155318e-01 -2.38702223e-01 -6.67265117e-01 -6.10701621e-01
-3.44189942e-01 -2.76741147e-01 1.31408882e+00 1.23831235e-01
5.61743855e-01 -1.18132651e+00 -2.83261508e-01 3.20364475e-01
1.07355997e-01 -4.14673299e-01 1.37662783e-01 2.31254622e-01
-1.11297488e+00 -1.21987127e-01 -4.37868297e-01 -5.55095136e-01
-1.17908633e+00 5.00580370e-01 4.68034148e-01 6.83568835e-01
-3.75559211e-01 7.64096260e-01 1.28502652e-01 3.08349490e-01
-5.60885631e-02 -5.18842995e-01 5.74367344e-01 -4.03063238e-01
7.95853555e-01 2.98710644e-01 -2.95392752e-01 -3.55886012e-01
-6.72500432e-02 9.56480742e-01 5.58697395e-02 -7.12338805e-01
8.25638950e-01 -5.77785730e-01 -8.52873698e-02 6.90489769e-01
9.25179124e-01 3.15640092e-01 -2.04293680e+00 -2.07562596e-02
-3.44870418e-01 -1.22807264e+00 4.00399178e-01 -9.80079353e-01
-6.75235569e-01 7.60015190e-01 6.29299343e-01 1.49463758e-01
1.16957104e+00 -2.30496645e-01 1.58794343e-01 1.28224358e-01
3.62283051e-01 -1.17842507e+00 -2.02655897e-01 -1.19287588e-01
5.70320189e-01 -1.29065061e+00 5.44270217e-01 -5.17747343e-01
-7.24742711e-01 1.38287330e+00 4.56630826e-01 -2.51698554e-01
3.64946157e-01 5.49579442e-01 4.33177799e-01 1.49943829e-01
-4.80412692e-01 -2.80661997e-03 2.75334895e-01 8.46789718e-01
1.90259963e-01 2.11446166e-01 1.65628463e-01 -2.90779829e-01
-2.61654198e-01 6.55463934e-02 1.20434463e+00 5.95870554e-01
-7.12898016e-01 -8.96956444e-01 -8.55121553e-01 4.12617087e-01
-2.04358339e-01 -2.65040770e-02 -4.71935511e-01 7.75948286e-01
1.00613646e-02 8.96147907e-01 4.82943773e-01 -2.59066552e-01
6.37485608e-02 -2.93229848e-01 9.62227285e-01 -4.21766967e-01
3.17776114e-01 3.62366170e-01 1.65459856e-01 -5.97848296e-01
-3.21746379e-01 -7.52529144e-01 -8.93391967e-01 -4.56544548e-01
-3.65173548e-01 -3.97130877e-01 9.10188317e-01 4.54321146e-01
-1.29546031e-01 3.52454185e-01 4.68601435e-01 -5.88576496e-01
-4.69418079e-01 -5.73661864e-01 -1.03370869e+00 2.43472338e-01
3.28374594e-01 -5.58354497e-01 -6.69045925e-01 5.34576476e-01] | [10.519660949707031, -2.546769142150879] |
5c7c9428-c220-449b-b66b-d06f66835fc9 | composer-compositional-learning-of-group | 2112.05892 | null | https://arxiv.org/abs/2112.05892v3 | https://arxiv.org/pdf/2112.05892v3.pdf | COMPOSER: Compositional Reasoning of Group Activity in Videos with Keypoint-Only Modality | Group Activity Recognition detects the activity collectively performed by a group of actors, which requires compositional reasoning of actors and objects. We approach the task by modeling the video as tokens that represent the multi-scale semantic concepts in the video. We propose COMPOSER, a Multiscale Transformer based architecture that performs attention-based reasoning over tokens at each scale and learns group activity compositionally. In addition, prior works suffer from scene biases with privacy and ethical concerns. We only use the keypoint modality which reduces scene biases and prevents acquiring detailed visual data that may contain private or biased information of users. We improve the multiscale representations in COMPOSER by clustering the intermediate scale representations, while maintaining consistent cluster assignments between scales. Finally, we use techniques such as auxiliary prediction and data augmentations tailored to the keypoint signals to aid model training. We demonstrate the model's strength and interpretability on two widely-used datasets (Volleyball and Collective Activity). COMPOSER achieves up to +5.4% improvement with just the keypoint modality. Code is available at https://github.com/hongluzhou/composer | ['Hans Peter Graf', 'Mubbasir Kapadia', 'Ting Liu', 'Long Zhao', 'Farley Lai', 'Shijie Geng', 'Aviv Shamsian', 'Asim Kadav', 'Honglu Zhou'] | 2021-12-11 | null | null | null | null | ['group-activity-recognition', 'relational-reasoning'] | ['computer-vision', 'natural-language-processing'] | [ 1.55563086e-01 -5.48652448e-02 -3.60908866e-01 -3.21731478e-01
-9.25076008e-01 -7.85459161e-01 6.96300626e-01 7.88517669e-02
-1.68545514e-01 1.05710708e-01 9.24532175e-01 2.51106530e-01
1.20873928e-01 -5.05503058e-01 -1.01273561e+00 -4.39265370e-01
-3.44762839e-02 5.94111197e-02 4.92565855e-02 1.09661231e-02
9.47428048e-02 1.25512540e-01 -1.62613046e+00 1.27537632e+00
5.79968154e-01 1.31105995e+00 -1.16848228e-02 7.94159412e-01
1.43443957e-01 1.64888430e+00 -6.14588916e-01 -3.96629095e-01
4.28669602e-01 -6.79178774e-01 -6.81816280e-01 3.40535045e-01
9.61789310e-01 -3.10323328e-01 -6.44721091e-01 8.15198362e-01
2.51811236e-01 2.73588330e-01 3.00639689e-01 -1.54280961e+00
-1.04855597e+00 7.60887206e-01 -5.88899195e-01 2.81648815e-01
4.57209051e-01 2.74576187e-01 1.38719583e+00 -1.05508411e+00
5.47282517e-01 1.21593249e+00 6.04449213e-01 4.12627757e-01
-1.18491364e+00 -8.31303775e-01 5.37690163e-01 5.63667774e-01
-1.42399919e+00 -6.74033165e-01 1.03379333e+00 -6.18503630e-01
7.80922830e-01 4.43080962e-01 9.85047996e-01 1.48101890e+00
-3.63178968e-01 1.17357016e+00 8.97452354e-01 2.48303451e-02
1.72804415e-01 -2.64016747e-01 -2.20283672e-01 1.01168907e+00
-1.14524595e-01 -4.88125056e-01 -1.25835466e+00 1.13732379e-03
8.39420140e-01 4.07195270e-01 -2.68554837e-01 -4.84757841e-01
-1.65805030e+00 7.23456323e-01 6.68066084e-01 1.20445058e-01
-4.25282121e-01 7.60396302e-01 4.11906093e-01 2.02620372e-01
5.14217317e-01 5.24742603e-01 -2.15692371e-01 -2.07834527e-01
-9.12247360e-01 1.26618668e-01 4.15602356e-01 1.02638054e+00
5.95768750e-01 -4.30251695e-02 -3.98158997e-01 6.94902062e-01
1.55497283e-01 3.64200890e-01 4.31802213e-01 -1.52183759e+00
7.19117284e-01 7.63054073e-01 -9.86779407e-02 -1.09253120e+00
4.40698713e-02 -4.11674291e-01 -6.37326479e-01 1.26876384e-01
4.68199134e-01 1.91604137e-01 -7.86239147e-01 1.72316813e+00
2.57054448e-01 4.89951819e-01 -3.41848910e-01 1.01974678e+00
8.94075215e-01 5.65981627e-01 2.18827263e-01 1.81962878e-01
1.62371743e+00 -1.26050508e+00 -6.62306547e-01 -1.90904245e-01
3.53128642e-01 -4.28693026e-01 1.43809807e+00 4.00227219e-01
-1.24502337e+00 -5.55261850e-01 -9.75489318e-01 -5.18029630e-01
-2.46724680e-01 3.32287967e-01 6.38031006e-01 2.36259893e-01
-8.35527837e-01 5.37385404e-01 -9.98796642e-01 -1.69761449e-01
9.43859637e-01 2.93306746e-02 -5.74019670e-01 1.06431842e-01
-8.65946472e-01 3.32136929e-01 -1.12334728e-01 -1.16704650e-01
-1.16565478e+00 -9.77744222e-01 -8.81106257e-01 1.38280332e-01
6.64150238e-01 -5.82545459e-01 1.21393311e+00 -1.50303638e+00
-1.21107090e+00 9.44660068e-01 -2.28465453e-01 -4.36605781e-01
7.47035742e-01 -5.05814552e-01 -1.09850332e-01 5.18025756e-01
2.13871479e-01 6.12330079e-01 8.58810425e-01 -1.13332319e+00
-6.58673584e-01 -2.85904348e-01 3.62290561e-01 4.51996386e-01
-4.92072612e-01 1.17201090e-01 -7.01287329e-01 -1.08524978e+00
1.98784143e-01 -6.51440978e-01 1.27622947e-01 4.22011793e-01
-1.24920212e-01 -1.12857729e-01 5.78066945e-01 -8.98568571e-01
1.06828630e+00 -2.23230553e+00 2.10230216e-01 -2.92234886e-02
5.35503030e-01 -3.84325266e-01 -2.28887677e-01 3.01072448e-01
-1.18311215e-02 9.01770815e-02 1.47909984e-01 -5.95459938e-01
7.87760392e-02 1.04256466e-01 -3.98616105e-01 5.72628736e-01
1.92991212e-01 1.04167879e+00 -9.29103494e-01 -6.03226781e-01
4.82390560e-02 4.91763473e-01 -8.69397819e-01 1.99427396e-01
-2.73870468e-01 3.58087271e-01 -3.43342423e-01 1.01664698e+00
9.75149646e-02 -5.36410749e-01 2.34766558e-01 -7.77931511e-01
1.26324624e-01 2.55685240e-01 -1.09048581e+00 2.14259148e+00
-3.67016256e-01 8.43964279e-01 8.49763677e-02 -9.46093559e-01
5.02958655e-01 2.75777280e-01 7.22836852e-01 -6.40561700e-01
-1.28983008e-02 -2.27660447e-01 -2.98709571e-01 -6.61981165e-01
5.23927450e-01 1.98758662e-01 -3.52156103e-01 5.00170350e-01
2.73851126e-01 3.65164191e-01 4.09025699e-02 4.69715595e-01
1.24960053e+00 3.16984922e-01 2.26301387e-01 4.67819236e-02
5.12205698e-02 -1.27317861e-01 7.52518892e-01 7.16301799e-01
-2.70723462e-01 7.51203179e-01 5.49103200e-01 -6.42471850e-01
-9.62026119e-01 -1.14936268e+00 6.31129920e-01 1.72972786e+00
3.17240834e-01 -8.86741757e-01 -5.38323045e-01 -8.28001142e-01
-6.07905351e-02 2.75084943e-01 -1.09062946e+00 -1.30285189e-01
-7.02138782e-01 -2.40325302e-01 6.09373033e-01 9.77405012e-01
5.66547871e-01 -8.32317591e-01 -6.32439315e-01 -8.35483149e-02
-7.55275667e-01 -1.15078533e+00 -7.68685639e-01 -1.01870365e-01
-5.11235833e-01 -1.26565278e+00 -4.08774346e-01 -5.41031837e-01
6.16658211e-01 3.76004606e-01 1.14073825e+00 9.95443091e-02
-2.31750041e-01 6.81800961e-01 -5.13496518e-01 -2.27914989e-01
-4.32767160e-02 -9.44880396e-02 -5.50282113e-02 5.74585617e-01
2.22688645e-01 -7.00228095e-01 -9.25312579e-01 3.11034203e-01
-5.39635360e-01 4.14754719e-01 3.90268683e-01 4.83490676e-01
5.54931462e-01 -4.30480421e-01 2.10676387e-01 -6.95361733e-01
1.88452676e-01 -4.69729096e-01 1.13482773e-02 3.25695902e-01
-1.67053252e-01 -1.53576523e-01 4.94012773e-01 -7.76283562e-01
-9.20024335e-01 1.19951509e-01 4.73265737e-01 -1.06120276e+00
-1.67483650e-02 -1.58982968e-03 -2.85209835e-01 2.40360618e-01
8.10236573e-01 1.10018067e-01 -2.01321989e-01 -5.75767100e-01
4.24761236e-01 3.43808144e-01 6.67316198e-01 -7.19415486e-01
6.88800812e-01 1.02334404e+00 -2.90049821e-01 -5.20806134e-01
-1.04061759e+00 -4.72381592e-01 -4.49241996e-01 -5.86790085e-01
1.02271724e+00 -1.54254913e+00 -9.81794119e-01 1.74931735e-01
-8.93922806e-01 -5.15050232e-01 -6.23097003e-01 3.59229594e-01
-7.36662209e-01 2.20531330e-01 -7.62429655e-01 -5.78995705e-01
-2.10186079e-01 -7.24642873e-01 1.18266559e+00 -1.37838528e-01
-5.53002000e-01 -5.25112927e-01 -2.16395333e-01 8.87275457e-01
6.51369393e-02 5.30440032e-01 4.31919724e-01 -5.47367930e-01
-9.27005172e-01 -7.12540699e-03 -8.65612179e-03 1.85799047e-01
-1.81636459e-03 -8.85441899e-02 -1.02758813e+00 -1.60754398e-01
-2.47238234e-01 -5.28372884e-01 1.00366223e+00 7.51649439e-02
1.57217872e+00 -6.52356207e-01 -2.86640972e-02 7.66215742e-01
1.04756773e+00 -1.17431998e-01 3.86403680e-01 1.70591310e-01
1.15947926e+00 5.21352768e-01 5.12688041e-01 5.66748381e-01
4.98232186e-01 5.83899140e-01 4.53400195e-01 -9.04702302e-03
-3.60474408e-01 -8.57280254e-01 5.39921761e-01 5.62078595e-01
-5.26233375e-01 1.10963136e-02 -6.58133030e-01 5.98896444e-01
-2.16731715e+00 -1.53395414e+00 2.52114087e-01 1.78787482e+00
6.58423603e-01 -1.98006153e-01 3.15506786e-01 8.72650072e-02
5.85977793e-01 4.76577431e-01 -5.10903120e-01 9.31749493e-02
-1.48737371e-01 -5.44411093e-02 4.37711000e-01 2.43595541e-01
-1.18487525e+00 7.79877126e-01 5.70589876e+00 7.64773190e-01
-8.61780286e-01 4.89260167e-01 6.24887168e-01 -9.09941673e-01
-3.23840886e-01 -8.61715153e-02 -2.25850642e-01 5.11620343e-01
4.53071296e-01 6.26001656e-02 6.67046010e-01 8.32171261e-01
3.00558895e-01 2.06334919e-01 -1.24034727e+00 1.20203519e+00
5.12650728e-01 -1.58442914e+00 9.14146379e-02 -1.20040506e-01
7.56091058e-01 -2.59022322e-02 1.35374159e-01 2.64014632e-01
4.35663134e-01 -9.21015263e-01 1.46019351e+00 5.47926366e-01
6.72072709e-01 -3.95554096e-01 2.88985204e-02 9.24526155e-02
-1.47863543e+00 -4.04354572e-01 3.66842784e-02 -2.18537211e-01
-2.39536464e-02 4.20420691e-02 -4.74770308e-01 1.78074941e-01
1.08047962e+00 1.16709363e+00 -7.66372323e-01 5.65458715e-01
-2.84643084e-01 5.84061921e-01 -1.74204305e-01 1.52808130e-01
-1.47619965e-02 -2.62514260e-02 2.76867986e-01 1.19765925e+00
3.60805035e-01 3.97797748e-02 2.25778818e-01 8.49131823e-01
-5.08385420e-01 -5.44306524e-02 -3.33613485e-01 -8.55318308e-02
4.91681635e-01 1.22832203e+00 -6.59002721e-01 -5.82929492e-01
-5.04851401e-01 1.20664251e+00 4.91196811e-01 4.73656148e-01
-1.09540129e+00 8.64611417e-02 9.68847275e-01 4.37426716e-01
3.88934344e-01 -1.00692935e-01 -3.43932509e-01 -1.56797755e+00
2.18271673e-01 -1.17190516e+00 7.45545387e-01 -1.03192699e+00
-1.10629368e+00 1.42176092e-01 -7.74480477e-02 -1.36124563e+00
5.35256565e-02 -3.89979661e-01 -5.83602548e-01 2.34943569e-01
-8.87156069e-01 -1.69487071e+00 -6.37325764e-01 8.93129766e-01
8.78723025e-01 -5.83926030e-02 5.09054840e-01 3.40802073e-01
-2.92756855e-01 5.74388266e-01 -2.91070700e-01 5.53483367e-01
7.78800130e-01 -1.24812865e+00 2.11484283e-01 7.66900301e-01
6.27867579e-01 4.95378017e-01 5.42828977e-01 -3.76715571e-01
-1.33887422e+00 -1.05784416e+00 6.74031556e-01 -8.35448921e-01
8.64033282e-01 -8.82194459e-01 -6.09534442e-01 1.09961295e+00
1.40746891e-01 3.65910798e-01 1.00012755e+00 2.65200645e-01
-9.63250637e-01 -2.62164265e-01 -8.77531230e-01 7.88966656e-01
1.65928829e+00 -9.97541904e-01 -5.89824677e-01 2.28966326e-01
5.66464245e-01 -3.07526708e-01 -9.65343595e-01 1.09332427e-01
8.33038449e-01 -9.47167277e-01 1.11433291e+00 -7.83056200e-01
6.69795036e-01 -4.89709675e-01 -4.42699343e-01 -1.00498474e+00
-4.68914449e-01 -5.25535524e-01 -5.82555950e-01 1.22735572e+00
9.33340415e-02 -8.93621743e-02 1.01746404e+00 6.65858448e-01
-3.22964531e-03 -4.60733891e-01 -7.66183078e-01 -4.63079304e-01
-4.40117210e-01 -5.19162178e-01 5.43506265e-01 1.19030523e+00
2.12762371e-01 2.61430234e-01 -6.06784523e-01 1.46557078e-01
5.82269728e-01 3.78216445e-01 8.53485525e-01 -7.59600878e-01
-6.05771661e-01 -2.65679836e-01 -5.07207751e-01 -9.89075363e-01
1.18422084e-01 -8.23872566e-01 -1.63514286e-01 -1.38078976e+00
3.67240399e-01 2.18843222e-02 -4.88702506e-01 7.18730748e-01
3.94661054e-02 7.19984472e-01 5.61630070e-01 3.99007142e-01
-1.24123061e+00 4.95221645e-01 1.22800696e+00 -4.24525321e-01
5.23250774e-02 -3.52058619e-01 -9.69334006e-01 9.37517345e-01
7.17485487e-01 -3.26474160e-01 -4.76242483e-01 -6.05456352e-01
3.48561496e-01 -2.04012468e-01 7.69963205e-01 -1.10381544e+00
2.22232908e-01 -2.43447170e-01 7.45351851e-01 -3.09572488e-01
7.09167063e-01 -7.96803296e-01 3.06494147e-01 2.86628902e-01
-8.70638013e-01 1.07899114e-01 -2.95483977e-01 7.70595312e-01
-2.18025863e-01 4.42188084e-01 5.09862781e-01 -3.41420233e-01
-7.99072266e-01 4.32348728e-01 -3.10786217e-01 1.50140226e-01
1.10896015e+00 -1.78038806e-01 -3.50801736e-01 -6.40271306e-01
-9.10963595e-01 1.78956285e-01 5.39728284e-01 6.52037203e-01
5.79212487e-01 -1.62941706e+00 -6.21375680e-01 5.57042584e-02
3.75067711e-01 -1.57666728e-01 5.01470864e-01 8.06930184e-01
-4.60129857e-01 -2.51243293e-01 -2.34850645e-01 -4.99548078e-01
-1.54306209e+00 6.54513836e-01 2.85763025e-01 2.04308629e-01
-6.33330345e-01 9.19316590e-01 4.90893811e-01 1.06278919e-01
2.62927622e-01 -6.24961972e-01 2.70159934e-02 3.32224965e-01
7.24180102e-01 4.65706110e-01 -3.80102783e-01 -8.37083876e-01
-3.34660202e-01 4.02582049e-01 1.44688100e-01 -8.83857310e-02
1.27927268e+00 -1.27945244e-01 1.22303404e-01 6.29466474e-01
1.28893232e+00 1.27058089e-01 -1.78174067e+00 -3.77021611e-01
-3.33939284e-01 -6.74905241e-01 -2.32117906e-01 -6.50245905e-01
-1.11714840e+00 8.61231744e-01 3.20678860e-01 -3.95915546e-02
1.11295474e+00 3.90176207e-01 4.66493189e-01 2.56579131e-01
1.76137015e-01 -1.27412438e+00 6.11564457e-01 1.63245909e-02
1.11457074e+00 -1.20169926e+00 6.28986433e-02 -2.39864260e-01
-9.14777458e-01 6.83804572e-01 6.69204295e-01 -3.05193663e-01
3.31753016e-01 1.71683922e-01 7.25520998e-02 -2.37313390e-01
-1.07557881e+00 -1.78695396e-01 4.34777200e-01 5.10640860e-01
2.67135799e-01 1.36262760e-01 2.64756501e-01 8.71453822e-01
-4.80678305e-02 -9.32829082e-02 1.85493544e-01 8.74577105e-01
-2.97370732e-01 -6.20913386e-01 -4.30411279e-01 4.14548367e-01
-5.97272456e-01 9.42772701e-02 -5.67111313e-01 5.63344657e-01
3.91978353e-01 8.11815858e-01 3.56897116e-01 -5.53795755e-01
3.57697129e-01 2.41498481e-02 5.11184216e-01 -4.28308696e-01
-8.43206644e-01 4.72747125e-02 1.81918919e-01 -1.17647982e+00
-6.75796449e-01 -7.98744678e-01 -1.07840610e+00 -3.19813251e-01
3.27845335e-01 -2.42472161e-02 2.79727310e-01 4.85580415e-01
5.02624333e-01 5.96435189e-01 4.83535826e-01 -8.76989841e-01
-2.00089201e-01 -8.19090247e-01 -4.84261394e-01 9.93611693e-01
3.33490223e-01 -5.49853742e-01 -2.55426556e-01 6.55077636e-01] | [8.86916732788086, 0.6905648112297058] |
8b47e3e8-8790-468d-9f48-0c193e14e8df | online-hyperparameter-optimization-for-class | 2301.05032 | null | https://arxiv.org/abs/2301.05032v2 | https://arxiv.org/pdf/2301.05032v2.pdf | Online Hyperparameter Optimization for Class-Incremental Learning | Class-incremental learning (CIL) aims to train a classification model while the number of classes increases phase-by-phase. An inherent challenge of CIL is the stability-plasticity tradeoff, i.e., CIL models should keep stable to retain old knowledge and keep plastic to absorb new knowledge. However, none of the existing CIL models can achieve the optimal tradeoff in different data-receiving settings--where typically the training-from-half (TFH) setting needs more stability, but the training-from-scratch (TFS) needs more plasticity. To this end, we design an online learning method that can adaptively optimize the tradeoff without knowing the setting as a priori. Specifically, we first introduce the key hyperparameters that influence the trade-off, e.g., knowledge distillation (KD) loss weights, learning rates, and classifier types. Then, we formulate the hyperparameter optimization process as an online Markov Decision Process (MDP) problem and propose a specific algorithm to solve it. We apply local estimated rewards and a classic bandit algorithm Exp3 to address the issues when applying online MDP methods to the CIL protocol. Our method consistently improves top-performing CIL methods in both TFH and TFS settings, e.g., boosting the average accuracy of TFH and TFS by 2.2 percentage points on ImageNet-Full, compared to the state-of-the-art. | ['Qianru Sun', 'Bernt Schiele', 'YingYing Li', 'Yaoyao Liu'] | 2023-01-11 | null | null | null | null | ['class-incremental-learning'] | ['computer-vision'] | [-1.23536088e-01 3.80772278e-02 -8.83502364e-01 -2.99132675e-01
-8.97624195e-01 -4.64311302e-01 3.18999857e-01 -6.77810758e-02
-6.38874471e-01 7.85876215e-01 -2.53657222e-01 -4.30262268e-01
-3.13558161e-01 -5.72524607e-01 -1.12928629e+00 -8.45746517e-01
2.77449284e-03 5.29751599e-01 4.15141284e-01 9.37945992e-02
5.39282784e-02 3.02201420e-01 -1.41370749e+00 2.33348176e-01
1.14611328e+00 1.36261380e+00 1.39326647e-01 5.79575479e-01
-6.18830277e-03 9.32693601e-01 -2.21859932e-01 -5.82282841e-01
3.65983456e-01 -1.82450756e-01 -9.62612569e-01 -2.21506312e-01
4.88538504e-01 -3.88488352e-01 -3.23318094e-01 8.59875262e-01
4.13887650e-01 1.91964000e-01 6.78735912e-01 -1.22155523e+00
-7.09020019e-01 9.82494414e-01 -5.37643969e-01 3.19217116e-01
-3.78793210e-01 3.09816390e-01 9.26898897e-01 -7.64630973e-01
3.74122441e-01 1.00853384e+00 6.86369121e-01 6.72640681e-01
-1.32875025e+00 -6.82326078e-01 6.72580659e-01 5.76558173e-01
-1.26656222e+00 -5.12950778e-01 5.57825387e-01 -6.57878220e-01
7.68305063e-01 -5.81071414e-02 7.32572675e-01 9.42931890e-01
1.92268178e-01 1.17075777e+00 1.09683836e+00 -4.22267020e-01
6.35172188e-01 3.45862389e-01 4.93522853e-01 6.42545521e-01
2.19095901e-01 1.44320130e-01 -6.95196927e-01 1.91764440e-02
6.94645941e-01 -2.41536926e-02 -8.49310383e-02 -4.86746520e-01
-6.21546686e-01 6.77921951e-01 5.16538262e-01 -4.62408672e-04
-3.68745267e-01 4.45009619e-01 2.32691780e-01 4.78798598e-01
4.96675611e-01 2.78796345e-01 -7.86713243e-01 -2.09305391e-01
-9.92557347e-01 1.98398173e-01 6.96890473e-01 7.14867651e-01
9.36404347e-01 -2.05057427e-01 -3.79392922e-01 9.93893504e-01
1.54812604e-01 3.20890486e-01 4.72140878e-01 -9.35209692e-01
5.87729335e-01 2.97900528e-01 1.05330370e-01 -4.55312937e-01
-1.14651337e-01 -7.02336371e-01 -5.69986284e-01 -5.06118611e-02
5.10401666e-01 -2.67685711e-01 -1.09015894e+00 1.82662559e+00
4.21679258e-01 2.49325678e-01 -1.53871581e-01 6.04340374e-01
3.30637485e-01 6.26918256e-01 1.26487061e-01 -4.02625561e-01
1.00176203e+00 -1.22121930e+00 -5.47905505e-01 -5.36853492e-01
4.83653724e-01 -2.81064689e-01 1.25429142e+00 3.76042992e-01
-1.16199350e+00 -3.62446815e-01 -9.93213117e-01 3.48493725e-01
-6.11479580e-02 -6.60662055e-02 4.84051466e-01 6.70381486e-01
-8.67184520e-01 7.55696893e-01 -9.86733079e-01 5.24137588e-03
6.33696318e-01 3.95213008e-01 1.53016716e-01 -1.05469830e-01
-1.12666523e+00 8.54800940e-01 4.66771841e-01 1.34917513e-01
-9.65384305e-01 -1.05212712e+00 -2.50795960e-01 1.59887671e-01
6.08570635e-01 -6.87446237e-01 1.61215210e+00 -1.20463312e+00
-1.98794043e+00 5.29413581e-01 -2.81920005e-02 -8.36085558e-01
7.15997040e-01 -4.81220156e-01 2.59125140e-02 -1.05622008e-01
-2.51200348e-01 8.10882688e-01 1.05704463e+00 -1.17634201e+00
-7.09312379e-01 -1.18882477e-01 1.03163972e-01 2.66952455e-01
-5.17219186e-01 -4.29462612e-01 -6.59142435e-01 -4.28653300e-01
-9.28987190e-02 -1.11680841e+00 -5.59109934e-02 4.14809994e-02
-1.00037441e-01 -2.33578846e-01 4.43092197e-01 -5.19565642e-01
1.44425869e+00 -2.23768759e+00 -3.03955171e-02 1.73317835e-01
-5.51999025e-02 1.98658556e-01 -6.60444945e-02 -1.31777371e-03
1.81972176e-01 1.32472202e-01 -1.00416280e-01 -2.38066554e-01
-8.21432024e-02 3.56399834e-01 -3.52241188e-01 2.76881784e-01
1.14212595e-01 9.46884394e-01 -8.11326385e-01 -5.37821472e-01
-2.36717030e-01 1.35305300e-01 -9.04366255e-01 -6.31530061e-02
-5.05337656e-01 2.13890210e-01 -3.54997605e-01 5.11022508e-01
5.57181537e-01 -5.43350697e-01 7.17734396e-02 -1.02107421e-01
-9.28914174e-02 2.09889412e-01 -1.08405530e+00 1.49988246e+00
-5.37120283e-01 4.36213076e-01 -6.36307299e-02 -1.05180335e+00
5.41636646e-01 5.72803728e-02 5.18617094e-01 -5.35899043e-01
-8.39082673e-02 2.26903901e-01 5.54831102e-02 -3.02933872e-01
3.18919510e-01 -1.45086646e-01 3.36995304e-01 2.51299828e-01
5.38454652e-02 1.26954466e-01 1.12166218e-02 -2.95690391e-02
1.01223373e+00 -8.33194554e-02 9.03223231e-02 -1.54354826e-01
4.90883738e-02 -7.72923529e-02 7.76154995e-01 1.04582238e+00
-3.57959837e-01 1.33627683e-01 5.55912614e-01 -2.67569065e-01
-7.47702181e-01 -1.07063437e+00 -2.34566554e-01 1.38105595e+00
1.06756255e-01 1.11473404e-01 -7.12382078e-01 -8.96763861e-01
3.42833072e-01 7.28262186e-01 -6.83571041e-01 -4.19497520e-01
-4.01148021e-01 -7.92808771e-01 2.02189356e-01 4.82902050e-01
8.20623517e-01 -7.36232221e-01 -4.69000012e-01 3.73766124e-01
-1.17823116e-01 -8.34141910e-01 -5.95921934e-01 3.25213313e-01
-1.07000411e+00 -6.92538381e-01 -5.78029633e-01 -6.40472651e-01
5.29313326e-01 -8.58102553e-03 9.86955881e-01 -2.98636049e-01
2.66468599e-02 3.80668700e-01 -2.35200793e-01 -5.32243669e-01
-2.03927159e-01 4.95669812e-01 3.14440764e-02 9.17828009e-02
-1.00626960e-01 -5.14391482e-01 -8.91017675e-01 4.52778310e-01
-7.19617009e-01 1.53455719e-01 8.79685521e-01 1.01033092e+00
7.74349868e-01 -3.15884203e-02 7.32122660e-01 -9.15872216e-01
4.98894215e-01 -5.19672155e-01 -5.27125120e-01 5.52051127e-01
-1.04700863e+00 2.89080858e-01 5.51944733e-01 -1.00757527e+00
-1.13538730e+00 9.57616791e-02 1.76294118e-01 -7.01132834e-01
4.82186794e-01 7.17679024e-01 4.10526432e-02 -1.24583945e-01
7.53007770e-01 1.38397872e-01 -1.26355663e-01 -4.03466493e-01
6.03298485e-01 5.57281196e-01 4.20693457e-01 -1.02199137e+00
5.07053375e-01 3.18300843e-01 -4.53068912e-01 -2.85273165e-01
-1.28037059e+00 -3.75488877e-01 -4.63673562e-01 -2.71099448e-01
3.66029620e-01 -9.62276101e-01 -6.38442516e-01 7.14619219e-01
-7.81107962e-01 -9.79711592e-01 -6.36164010e-01 4.06205207e-01
-6.10595465e-01 -1.05328910e-01 -7.32972920e-01 -7.97708690e-01
-4.61560279e-01 -1.06023896e+00 5.85125685e-01 3.11560631e-01
1.90571651e-01 -7.86701322e-01 8.68881196e-02 4.53023046e-01
5.14877081e-01 -1.94950730e-01 9.34460938e-01 -3.30704957e-01
-7.30079830e-01 8.70512053e-02 -1.67411193e-01 5.44391096e-01
-6.75641224e-02 -1.82258207e-02 -1.05573404e+00 -5.11379242e-01
-1.32947281e-01 -5.45077562e-01 1.11746943e+00 6.80163622e-01
1.36726964e+00 -7.97421813e-01 -2.60203332e-01 6.29578531e-01
1.30751944e+00 2.60034561e-01 5.05637884e-01 4.90657896e-01
4.05554086e-01 2.69279778e-01 5.81622601e-01 3.86946350e-01
4.20252353e-01 6.66232765e-01 1.92398086e-01 2.87835896e-01
-1.07360624e-01 -4.59144473e-01 5.62647283e-01 8.26495409e-01
2.09156916e-01 8.57031420e-02 -1.00519574e+00 5.33691406e-01
-2.13616705e+00 -8.13048005e-01 5.86575150e-01 2.41794801e+00
1.58687842e+00 4.62325305e-01 1.44600406e-01 -1.27916902e-01
5.92185497e-01 -9.73949134e-02 -1.22489882e+00 -2.24840760e-01
1.73895106e-01 -6.99878251e-03 7.05061913e-01 5.81479847e-01
-1.03781486e+00 1.04575562e+00 6.10040092e+00 1.19125581e+00
-1.33371747e+00 3.53420913e-01 1.01672745e+00 -4.63873863e-01
-5.24743423e-02 1.36567876e-01 -1.09698570e+00 5.69550157e-01
1.14495742e+00 -2.19633430e-01 6.80473924e-01 1.11189568e+00
3.15664671e-02 -2.60719150e-01 -1.22220373e+00 9.24655318e-01
-3.32968771e-01 -1.41589344e+00 -5.59253022e-02 -2.05968335e-01
9.50595558e-01 2.44295612e-01 2.96677023e-01 8.17249417e-01
4.66126442e-01 -6.29616797e-01 1.06088710e+00 6.45794928e-01
8.05003643e-01 -5.66332221e-01 3.46993208e-01 5.09897947e-01
-9.66837645e-01 -5.41694701e-01 -1.91662893e-01 2.01694936e-01
-1.11611076e-01 9.27203178e-01 -7.00467408e-01 1.79108873e-01
8.66282344e-01 5.12520611e-01 -4.23049241e-01 1.34561110e+00
-8.01033229e-02 1.07102573e+00 -5.55076241e-01 -5.61938658e-02
1.57365724e-01 -2.95232292e-02 1.39636263e-01 1.03334892e+00
2.24567633e-02 -6.50035590e-02 1.42807320e-01 7.40649164e-01
-3.46686631e-01 6.53845742e-02 8.10804740e-02 1.95780694e-01
8.97887230e-01 7.49981523e-01 -4.38764960e-01 -3.23246390e-01
-1.02680400e-01 7.23225594e-01 7.49231279e-01 5.32641232e-01
-8.61489654e-01 -3.87313552e-02 5.03527582e-01 1.99687436e-01
4.76945251e-01 -6.20274842e-02 -3.89384270e-01 -1.14891458e+00
1.49869561e-01 -7.53456891e-01 4.61648643e-01 -4.04286087e-01
-1.49703717e+00 2.21650511e-01 2.43410438e-01 -9.20885205e-01
-3.44487000e-03 -2.79567540e-01 -3.79622877e-01 4.89243954e-01
-1.70725083e+00 -9.38213646e-01 -5.40036075e-02 4.58250731e-01
4.85301942e-01 1.25059694e-01 2.29540333e-01 3.86556506e-01
-9.41275001e-01 1.05493915e+00 6.36716723e-01 -8.35449398e-02
7.24676073e-01 -9.47358787e-01 5.40178157e-02 5.96292555e-01
-1.90662164e-02 4.89157736e-01 4.91602451e-01 -4.55307633e-01
-1.22479820e+00 -1.15110493e+00 3.76832575e-01 -1.97358072e-01
7.62454450e-01 -2.29143590e-01 -9.63040531e-01 6.21086419e-01
-3.86373788e-01 1.61629796e-01 4.33349580e-01 3.06207448e-01
-5.65712273e-01 -6.32103264e-01 -9.80705440e-01 6.42019272e-01
1.05361581e+00 -3.71813625e-01 -2.55434155e-01 3.11584860e-01
8.14022779e-01 -5.11463284e-01 -9.22174215e-01 3.64081025e-01
7.42932200e-01 -6.31346047e-01 9.22544897e-01 -4.69449341e-01
1.78680733e-01 5.47858551e-02 -1.33139605e-03 -1.32040143e+00
-3.84076953e-01 -7.14301765e-01 -6.11837864e-01 1.05065131e+00
6.79722846e-01 -8.91548276e-01 1.01890111e+00 8.00909281e-01
5.33631258e-02 -1.25585747e+00 -1.16043663e+00 -1.11689472e+00
3.88898849e-01 -4.66080993e-01 4.15930122e-01 8.03576887e-01
-9.53187868e-02 2.76625723e-01 -3.04436892e-01 -5.34817837e-02
6.08142614e-01 3.61517556e-02 5.27809083e-01 -1.17499638e+00
-5.94331622e-01 -5.49989581e-01 1.46384925e-01 -1.37641490e+00
-8.29060748e-03 -7.50302374e-01 7.64711425e-02 -1.24015474e+00
4.11547482e-01 -1.07216620e+00 -5.89344323e-01 9.09465015e-01
-3.85896683e-01 -5.35968781e-01 3.79345447e-01 7.15189576e-01
-8.32658589e-01 6.68716848e-01 1.13312471e+00 -1.52735651e-01
-6.56293750e-01 1.47966430e-01 -7.05767930e-01 4.66123581e-01
8.21285009e-01 -6.30680323e-01 -6.00273013e-01 -4.47096467e-01
2.85214365e-01 -5.89968152e-02 1.74467877e-01 -9.58690822e-01
5.48423946e-01 -5.59516549e-01 6.30587116e-02 -3.80751491e-01
8.66654292e-02 -6.14425421e-01 -5.06420992e-02 4.94357973e-01
-5.47916830e-01 -3.09337437e-01 2.11458176e-01 9.24543500e-01
7.56224617e-02 -2.23427147e-01 1.04459476e+00 1.75539926e-01
-5.44158041e-01 4.99363720e-01 -2.58720517e-01 3.19109410e-01
9.41535771e-01 -1.70037732e-01 -4.80050951e-01 -2.17710033e-01
-6.62516177e-01 5.16209006e-01 2.39963338e-01 2.75365502e-01
3.04947883e-01 -1.18850803e+00 -4.90652442e-01 -2.82249488e-02
1.94192827e-02 3.06387305e-01 3.53612453e-01 1.00205767e+00
-1.87350556e-01 1.32654697e-01 3.86426561e-02 -6.07341409e-01
-8.36607933e-01 3.14698875e-01 6.16837859e-01 -7.19224632e-01
-3.53585422e-01 1.00546217e+00 -1.07076867e-02 -3.27323228e-01
6.88513398e-01 -4.87505406e-01 1.35319322e-01 1.03693739e-01
2.36766890e-01 3.81560683e-01 2.02328756e-01 3.81925881e-01
-1.26080573e-01 3.42822313e-01 -5.00719845e-01 -1.82930902e-01
1.23514664e+00 -8.73272046e-02 2.96599388e-01 6.90852225e-01
9.73377347e-01 -5.19291282e-01 -1.80922294e+00 -6.47382796e-01
5.08254878e-02 -2.59160221e-01 3.08828473e-01 -1.01498187e+00
-1.09420121e+00 6.11096799e-01 7.82857537e-01 -8.96107852e-02
9.08573806e-01 5.06965220e-02 7.33222246e-01 6.16121054e-01
4.37250584e-01 -1.57771111e+00 2.88252980e-01 6.75518095e-01
6.51131868e-01 -1.12237978e+00 2.55114157e-02 -1.22686386e-01
-6.76085949e-01 8.73711467e-01 6.36817873e-01 1.29489601e-01
9.29477632e-01 1.57430559e-01 -2.91466296e-01 1.53633997e-01
-1.15304875e+00 1.61225088e-02 2.54376650e-01 1.51216194e-01
-8.65538139e-03 1.11609370e-01 -9.60487053e-02 6.78412914e-01
-6.82618394e-02 3.13659608e-01 1.03984408e-01 9.23985183e-01
-5.61060309e-01 -8.90733480e-01 -2.40104735e-01 7.18212724e-01
-2.54498661e-01 -9.55436155e-02 -6.02516234e-02 6.17756188e-01
7.99942315e-02 7.11959541e-01 2.22246312e-02 -3.96785498e-01
3.27013403e-01 1.85435846e-01 6.02350831e-01 -4.56586391e-01
-5.33706665e-01 -1.28263414e-01 -6.94302246e-02 -3.77227634e-01
-2.83192515e-01 -6.67865872e-01 -1.02895355e+00 -3.08935672e-01
-6.91036761e-01 9.11915153e-02 6.12640440e-01 9.20193732e-01
4.58001196e-01 3.82960260e-01 6.55838311e-01 -6.51192963e-01
-1.08592701e+00 -7.69484162e-01 -4.63761330e-01 1.24597587e-01
2.82384634e-01 -8.34564328e-01 -4.80834275e-01 6.66902438e-02] | [9.337628364562988, 3.4648427963256836] |
82771040-31b9-469c-a8c4-8b20a8458cf5 | multivariate-time-series-regression-with | 2201.00818 | null | https://arxiv.org/abs/2201.00818v3 | https://arxiv.org/pdf/2201.00818v3.pdf | Graph Neural Networks for Multivariate Time Series Regression with Application to Seismic Data | Machine learning, with its advances in deep learning has shown great potential in analyzing time series. In many scenarios, however, additional information that can potentially improve the predictions is available. This is crucial for data that arise from e.g., sensor networks that contain information about sensor locations. Then, such spatial information can be exploited by modeling it via graph structures, along with the sequential (time series) information. Recent advances in adapting deep learning to graphs have shown potential in various tasks. However, these methods have not been adapted for time series tasks to a great extent. Most attempts have essentially consolidated around time series forecasting with small sequence lengths. Generally, these architectures are not well suited for regression or classification tasks where the value to be predicted is not strictly depending on the most recent values, but rather on the whole length of the time series. We propose TISER-GCN, a novel graph neural network architecture for processing, in particular, these long time series in a multivariate regression task. Our proposed model is tested on two seismic datasets containing earthquake waveforms, where the goal is to predict maximum intensity measurements of ground shaking at each seismic station. Our findings demonstrate promising results of our approach -- with an average MSE reduction of 16.3% - compared to the best performing baselines. In addition, our approach matches the baseline scores by needing only half the input size. The results are discussed in depth with an additional ablation study. | ['Martin Atzmueller', 'Alberto Michelini', 'Dario Jozinović', 'Jurgen van den Hoogen', 'Stefan Bloemheuvel'] | 2022-01-03 | null | null | null | null | ['time-series-regression'] | ['time-series'] | [ 2.62197495e-01 1.39796525e-01 1.72559902e-01 -2.57356465e-01
-5.52481711e-01 -4.64397669e-01 4.04741794e-01 6.79409027e-01
-3.96331668e-01 5.86557865e-01 1.61751390e-01 -4.31633323e-01
-4.28852648e-01 -9.35030162e-01 -6.96512997e-01 -9.12098527e-01
-9.09600854e-01 1.98215425e-01 3.09498608e-01 -5.16093314e-01
1.96411550e-01 6.68998063e-01 -1.07576036e+00 1.17688946e-01
3.89991611e-01 1.15362144e+00 1.73070461e-01 7.06748307e-01
1.76483691e-01 9.15606499e-01 -6.89504921e-01 -1.58294946e-01
6.69560134e-02 -9.47296545e-02 -4.69174087e-01 -2.60277510e-01
-1.80954188e-01 -1.43076777e-01 -4.94826704e-01 6.50239110e-01
5.24036884e-01 1.65840656e-01 5.43085456e-01 -9.80379403e-01
-5.01653962e-02 9.25032437e-01 -4.99089956e-01 5.06819963e-01
1.17232248e-01 -1.31911173e-01 9.48528826e-01 -5.09505749e-01
2.89236069e-01 9.16201055e-01 9.29830432e-01 3.72751020e-02
-1.05726612e+00 -3.68157536e-01 -3.70720886e-02 3.76620412e-01
-1.18731463e+00 -1.77389547e-01 1.29709530e+00 -4.35513079e-01
1.22594810e+00 1.57975867e-01 5.18877864e-01 9.64261174e-01
4.85270232e-01 5.34285247e-01 6.85021698e-01 -1.25494063e-01
3.03549618e-01 -5.65212727e-01 1.33973852e-01 2.30086863e-01
1.06492914e-01 -8.99188071e-02 -3.99807990e-01 5.23630939e-02
4.22868341e-01 1.28404751e-01 -2.37118423e-01 1.19739480e-01
-1.17552054e+00 8.66109312e-01 6.01317704e-01 7.56100893e-01
-7.13452578e-01 3.31700236e-01 6.07487738e-01 5.29562414e-01
8.29878867e-01 4.70989317e-01 -4.67392594e-01 -2.73205161e-01
-9.84565258e-01 1.75008811e-02 8.29741299e-01 5.11527359e-01
5.52285790e-01 5.48029780e-01 1.55183434e-01 5.12584388e-01
-8.54495317e-02 4.06453580e-01 3.69272262e-01 -4.49404866e-01
8.89032841e-01 4.62507516e-01 -1.54945254e-01 -1.57760131e+00
-1.05501509e+00 -5.75627506e-01 -1.37448561e+00 -1.61372691e-01
3.29370320e-01 -5.23213923e-01 -8.22077930e-01 1.69579780e+00
-6.12220354e-02 4.40777063e-01 8.31029862e-02 7.15309620e-01
4.90637958e-01 9.66636479e-01 -2.72207409e-01 -3.74375612e-01
8.85575593e-01 -3.80755275e-01 -6.66075528e-01 -3.33607227e-01
8.86394978e-01 -2.78281391e-01 5.87864280e-01 4.48187560e-01
-8.45589340e-01 -4.08899754e-01 -1.10168719e+00 3.34501147e-01
-5.49080133e-01 -2.53284842e-01 4.01272327e-01 4.59172428e-01
-1.20242083e+00 1.01710021e+00 -1.04473913e+00 -1.65202945e-01
2.71804184e-01 4.18608040e-01 -2.67282248e-01 1.32776380e-01
-1.48004460e+00 7.46544659e-01 5.03853738e-01 4.08487827e-01
-5.64335823e-01 -4.17857707e-01 -7.99427152e-01 4.21003461e-01
3.56560618e-01 -1.51434153e-01 8.95029008e-01 -4.81183708e-01
-1.14301622e+00 9.66530666e-02 4.64868806e-02 -1.00910187e+00
3.73157322e-01 4.81440909e-02 -5.83981276e-01 9.40684453e-02
-1.99063033e-01 7.63723860e-03 8.00945461e-01 -6.79756641e-01
-3.68673444e-01 -4.52570319e-01 4.83872108e-02 -1.53825805e-01
-5.62279880e-01 -1.87824011e-01 -3.10379956e-02 -8.61137390e-01
2.25178733e-01 -7.74261475e-01 -5.93871415e-01 -5.07648468e-01
-2.63338774e-01 -2.81655461e-01 7.94306695e-01 -8.26314032e-01
1.55206263e+00 -2.11198688e+00 1.74713582e-01 5.13586819e-01
2.52666146e-01 1.17049240e-01 -6.70582876e-02 1.06684256e+00
-2.17689782e-01 1.48389965e-01 -4.81777519e-01 -2.35479578e-01
-2.88650453e-01 1.70048714e-01 -5.27496219e-01 4.42773432e-01
3.17380339e-01 6.78788066e-01 -8.51141393e-01 -1.51443422e-01
1.09425023e-01 5.54355681e-01 -2.07285702e-01 7.52871037e-02
-9.84178856e-02 5.10717154e-01 -4.23317850e-01 2.07287177e-01
3.00836295e-01 -3.41923535e-01 2.28460431e-01 -1.50832519e-01
-1.86302774e-02 8.06176513e-02 -9.96798158e-01 1.56949437e+00
-3.71563464e-01 7.26350665e-01 -3.62908661e-01 -1.67328429e+00
1.07686388e+00 4.37600911e-01 8.63161623e-01 -6.98076010e-01
-3.68016353e-03 2.10094243e-01 2.26209477e-01 -5.75894177e-01
3.34100008e-01 3.88666727e-02 -1.31681725e-01 3.45850438e-01
-1.92540616e-01 -6.01808168e-02 3.29465419e-01 2.95226648e-03
1.36932290e+00 -1.75259158e-01 2.95018256e-01 -5.29350825e-02
2.96598494e-01 -5.32710515e-02 3.23457390e-01 5.81960976e-01
2.54185051e-01 5.86678445e-01 7.98926651e-01 -6.10910475e-01
-1.05577803e+00 -5.09247541e-01 1.83005899e-01 8.87812316e-01
-2.94708937e-01 -3.86899620e-01 -4.28543508e-01 -3.32806826e-01
-2.29401007e-01 6.33295417e-01 -6.08369112e-01 -9.35010910e-02
-9.28828061e-01 -8.41387630e-01 6.18918180e-01 7.22250164e-01
9.03282836e-02 -1.14183915e+00 -7.79274642e-01 7.41419673e-01
-1.01832278e-01 -1.19395816e+00 -5.04097790e-02 4.60585177e-01
-1.24513638e+00 -8.28984976e-01 -6.68526828e-01 -4.59507227e-01
2.71776557e-01 1.30273715e-01 9.22177315e-01 -9.37427860e-03
2.48765592e-02 1.69855192e-01 -5.68079650e-01 -5.81237376e-01
-2.46972337e-01 2.76695520e-01 -1.96055278e-01 2.64146626e-01
2.88081635e-02 -9.93123114e-01 -5.69687128e-01 2.18787789e-02
-1.22777629e+00 -1.38435081e-01 4.35725838e-01 6.83725059e-01
1.50709003e-01 2.05007046e-01 9.82189298e-01 -6.30620718e-01
7.07713723e-01 -8.97576272e-01 -4.56203163e-01 5.09324074e-02
-4.28095102e-01 5.69034517e-02 1.08884573e+00 -3.84166688e-01
-5.39923847e-01 -1.75725624e-01 -2.05078647e-01 -3.49036098e-01
6.57177269e-02 1.36893535e+00 3.01783949e-01 2.61568964e-01
6.42069578e-01 1.80643469e-01 -8.98333341e-02 -3.18293452e-01
-3.27630364e-03 5.05557954e-01 3.86357427e-01 -2.81485945e-01
4.80008930e-01 4.24179018e-01 4.50684428e-01 -1.04204667e+00
-5.13363779e-01 -5.06315887e-01 -6.35163069e-01 -3.75892669e-01
6.74058557e-01 -6.18619919e-01 -4.98925686e-01 6.74587429e-01
-1.09218657e+00 -4.38952178e-01 -6.56841993e-02 6.04736865e-01
-4.22730029e-01 4.01632756e-01 -5.73612392e-01 -9.35963392e-01
-3.22286069e-01 -7.95244932e-01 9.30292368e-01 -1.27195165e-01
-1.10040002e-01 -1.39888048e+00 1.25611192e-02 -2.21126169e-01
6.27945364e-01 7.42828965e-01 8.68435681e-01 -1.03078222e+00
-2.56265998e-01 -5.76147139e-01 7.74111152e-02 1.24883547e-01
9.69118997e-02 -2.77472019e-01 -8.75198781e-01 -5.21272779e-01
1.14179477e-01 -7.18516633e-02 8.88899386e-01 4.42233473e-01
1.32810128e+00 -1.81001827e-01 -1.76221684e-01 4.38836724e-01
1.31716931e+00 3.87215644e-01 5.35263956e-01 2.99133480e-01
7.92056561e-01 6.99045122e-01 3.67422283e-01 7.82494247e-01
3.72519284e-01 5.07984161e-01 8.00846338e-01 -9.27142948e-02
2.51640022e-01 1.82440460e-01 2.83826858e-01 1.19603014e+00
-4.28521931e-01 -6.35266125e-01 -1.37493110e+00 6.96399629e-01
-1.97508287e+00 -1.11772609e+00 -2.31846616e-01 2.06651354e+00
3.32900494e-01 3.30365509e-01 1.70752648e-02 7.98069119e-01
4.65564400e-01 6.61425173e-01 -5.42469680e-01 -2.62824118e-01
-2.27078304e-01 3.02539796e-01 5.51941574e-01 1.78124577e-01
-1.10364890e+00 4.03464049e-01 5.33587360e+00 4.92032707e-01
-1.46918976e+00 -1.80489764e-01 5.49539447e-01 1.80803016e-01
-1.94476947e-01 -1.80814311e-01 -2.10689068e-01 3.44014287e-01
1.40549314e+00 -1.66413873e-01 3.55509311e-01 5.14882803e-01
5.00956237e-01 2.37795442e-01 -1.03037369e+00 8.79038274e-01
-8.38111043e-02 -1.33500242e+00 -2.48920098e-01 -5.11045009e-02
4.53967184e-01 2.48883024e-01 -9.19618905e-02 2.20877081e-01
-1.05056606e-01 -9.51074839e-01 5.40901661e-01 5.83104610e-01
4.25108254e-01 -7.93521464e-01 1.01611090e+00 5.83596289e-01
-1.38981879e+00 -3.39446962e-01 -1.57424346e-01 -4.27987248e-01
3.65626991e-01 8.07661831e-01 -1.11332417e+00 9.47009265e-01
7.76365638e-01 1.07975614e+00 -4.36305732e-01 1.00250053e+00
8.50754678e-02 1.05829656e+00 -5.48194170e-01 -2.11510539e-01
4.76525366e-01 -5.41866198e-02 5.34453928e-01 1.27329016e+00
6.70823216e-01 1.57853305e-01 1.62527904e-01 2.03508109e-01
2.04771549e-01 1.39772356e-01 -9.47028339e-01 -1.31419182e-01
2.74812043e-01 8.88715446e-01 -8.52596581e-01 -2.50282288e-01
-5.49493313e-01 3.98175985e-01 3.15914124e-01 5.04420340e-01
-7.22480536e-01 -5.66077769e-01 6.09556772e-02 2.27543011e-01
3.76442552e-01 -5.19703209e-01 -3.37543905e-01 -9.00866568e-01
2.63211310e-01 -6.37375653e-01 6.46376073e-01 -6.62750483e-01
-1.23175550e+00 7.11384416e-01 1.66040778e-01 -1.39029694e+00
-7.77514160e-01 -5.21129429e-01 -6.21032774e-01 7.17924774e-01
-1.36460388e+00 -9.58845377e-01 -1.14881463e-01 6.88538134e-01
4.97742325e-01 -7.89478794e-02 6.42060459e-01 3.19727570e-01
-2.99275309e-01 2.16422275e-01 1.78221524e-01 3.58291954e-01
2.56973714e-01 -1.10837674e+00 5.73992729e-01 1.05077815e+00
1.48293614e-01 1.67194530e-01 7.47956514e-01 -5.27170300e-01
-1.49421155e+00 -1.10792673e+00 8.33934486e-01 6.19637780e-02
1.07488525e+00 -2.60710895e-01 -1.20051837e+00 5.95339477e-01
8.46244991e-02 -7.27318414e-03 3.90243173e-01 3.53384353e-02
-1.46260723e-01 -3.50498289e-01 -7.13397920e-01 4.10604328e-01
7.53817320e-01 -4.18926835e-01 -4.34922367e-01 2.47854665e-01
6.73991621e-01 -3.31412941e-01 -1.15651035e+00 5.17268479e-01
2.44896486e-01 -7.84171939e-01 7.72927344e-01 -5.35360515e-01
4.25194561e-01 -6.27392754e-02 -1.81549355e-01 -1.47296727e+00
-2.60031372e-01 -6.27820969e-01 -3.21177512e-01 7.93856621e-01
4.72943157e-01 -7.59135306e-01 6.19900525e-01 2.73925245e-01
-4.49104875e-01 -7.19747901e-01 -1.02384913e+00 -7.40768492e-01
-2.95365304e-02 -7.75026381e-01 7.07625389e-01 1.05537403e+00
2.74317618e-02 3.51248860e-01 -6.12487376e-01 4.29342330e-01
4.15501833e-01 9.50493887e-02 5.39405167e-01 -1.40320158e+00
-2.07240045e-01 -4.85076547e-01 -7.29021013e-01 -6.71679378e-01
-7.50072021e-03 -7.26106405e-01 -1.29464820e-01 -1.53168380e+00
-3.73308629e-01 -1.64861411e-01 -4.97340560e-01 4.76319939e-01
-1.92937464e-03 -1.18726112e-01 9.81397927e-02 9.82594714e-02
-1.73227370e-01 3.51839751e-01 9.43901002e-01 -2.62149364e-01
-2.19915256e-01 2.91448712e-01 -6.74913228e-02 5.62635601e-01
1.20746374e+00 -3.97205234e-01 -5.91038346e-01 -5.83478332e-01
5.49374759e-01 7.83312798e-01 3.29531163e-01 -1.26928186e+00
3.99199396e-01 -2.75841989e-02 1.26209110e-01 -7.35030293e-01
2.76210696e-01 -9.69172597e-01 2.70391375e-01 4.90178078e-01
-2.77577013e-01 6.43410861e-01 2.18985930e-01 7.50998437e-01
-4.78649378e-01 5.31533547e-02 2.55489677e-01 1.07112028e-01
-8.53458881e-01 3.69344503e-01 -3.85640264e-01 -1.79097250e-01
7.89858878e-01 -1.51401326e-01 -1.50595292e-01 -7.51727402e-01
-9.25770283e-01 1.50025144e-01 -1.96065396e-01 4.85150248e-01
6.91684663e-01 -1.06526649e+00 -9.77658629e-01 6.69734702e-02
6.46711513e-02 1.15399897e-01 2.31324852e-01 9.63333547e-01
-4.84645069e-01 2.95747519e-01 -4.76480611e-02 -7.13270903e-01
-7.95811772e-01 7.65359521e-01 1.06807970e-01 -6.27167165e-01
-6.94000363e-01 4.42743033e-01 -1.14940986e-01 1.85471009e-02
1.74336433e-01 -7.35763371e-01 -5.88639140e-01 3.05629492e-01
3.68576258e-01 3.55931133e-01 4.29681718e-01 -4.55193520e-01
-2.06990689e-01 5.37704349e-01 2.91624516e-01 -7.31594954e-03
1.95769882e+00 1.70845408e-02 -1.35570899e-01 8.27085555e-01
1.21659112e+00 -1.78697467e-01 -1.01068330e+00 -2.42815182e-01
4.43966776e-01 3.86819430e-02 -8.26158598e-02 -3.20034742e-01
-1.28491211e+00 1.10875809e+00 3.09145749e-01 9.04326141e-01
1.39436460e+00 -1.50070801e-01 6.17886484e-01 5.44036984e-01
5.23532331e-01 -7.11480796e-01 1.14711812e-02 6.62897229e-01
9.86904621e-01 -1.14458549e+00 -1.10721193e-01 2.81021576e-02
-3.31805617e-01 1.47406054e+00 4.16308381e-02 -2.03379750e-01
9.38408375e-01 3.72899294e-01 -1.24047890e-01 -3.53575826e-01
-7.39443541e-01 -5.68222702e-02 3.34408313e-01 4.91416246e-01
4.68686759e-01 1.07704863e-01 -1.96371093e-01 3.61325353e-01
-2.18641251e-01 -2.55549669e-01 5.02582669e-01 8.43101084e-01
-3.51289988e-01 -7.02776134e-01 -3.55375797e-01 5.79955459e-01
-6.02036357e-01 -6.70404881e-02 3.07617486e-02 7.07512736e-01
-4.14730400e-01 1.15594971e+00 5.81749678e-02 -5.45141935e-01
4.40662742e-01 -1.32136881e-01 -9.40168947e-02 -2.50460684e-01
-5.87620556e-01 7.77220502e-02 1.06056802e-01 -5.29798329e-01
-5.75264037e-01 -5.93945026e-01 -1.16761088e+00 -2.36797616e-01
-8.70302841e-02 9.49819237e-02 6.61335886e-01 8.53544831e-01
2.31565833e-01 7.55559444e-01 8.09121847e-01 -1.12360239e+00
-4.87665951e-01 -1.11626351e+00 -5.34572959e-01 3.33186775e-01
6.29462719e-01 -2.28309005e-01 -3.09549868e-01 -5.29773198e-02] | [6.975058555603027, 2.82202410697937] |
21d40707-524b-4c11-8658-b50c28bbe2ce | gaussian-process-regression-with-1 | null | null | https://ieeexplore.ieee.org/abstract/document/9646444 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9646444 | Gaussian Process Regression With Interpretable Sample-Wise Feature Weights | Gaussian process regression (GPR) is a fundamental model used in machine learning (ML). Due to its accurate prediction with uncertainty and versatility in handling various data structures via kernels, GPR has been successfully used in various applications. However, in GPR, how the features of an input contribute to its prediction cannot be interpreted. Here, we propose GPR with local explanation, which reveals the feature contributions to the prediction of each sample while maintaining the predictive performance of GPR. In the proposed model, both the prediction and explanation for each sample are performed using an easy-to-interpret locally linear model. The weight vector of the locally linear model is assumed to be generated from multivariate Gaussian process priors. The hyperparameters of the proposed models are estimated by maximizing the marginal likelihood. For a new test sample, the proposed model can predict the values of its target variable and weight vector, as well as their uncertainties, in a closed form. Experimental results on various benchmark datasets verify that the proposed model can achieve predictive performance comparable to those of GPR and superior to that of existing interpretable models and can achieve higher interpretability than them, both quantitatively and qualitatively. | ['Tomoharu Iwata', 'Yuya Yoshikawa'] | 2021-12-10 | null | null | null | ieee-transactions-on-neural-networks-and-4 | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [ 9.90181789e-02 2.93323845e-01 -8.84473845e-02 -3.25036466e-01
-6.32526934e-01 -3.33915241e-02 4.25745130e-01 3.22721452e-01
2.14578420e-01 7.57731855e-01 -1.37510180e-01 -2.25942016e-01
-4.30704534e-01 -6.57012820e-01 -6.35817230e-01 -1.15001667e+00
7.52580911e-02 6.18393421e-01 1.60214826e-01 4.63387698e-01
2.96330690e-01 3.76198024e-01 -1.31572306e+00 -9.70814228e-02
1.18132007e+00 1.28508663e+00 3.35920572e-01 3.37986380e-01
2.03150208e-03 6.81767881e-01 -2.20663741e-01 -3.47889543e-01
4.05446673e-03 1.89145785e-02 -2.38261715e-01 9.54628289e-02
-2.52251893e-01 1.52962402e-01 -1.05183192e-01 1.07301879e+00
1.02060907e-01 3.20443302e-01 1.13870764e+00 -1.12347627e+00
-8.59310925e-01 4.40188855e-01 -8.17891002e-01 -1.38651863e-01
-7.12281615e-02 -8.02787393e-02 9.32802200e-01 -1.24825013e+00
-1.89527676e-01 1.45709884e+00 6.06480002e-01 1.09949842e-01
-1.46795273e+00 -4.89295483e-01 3.51619124e-01 1.03970006e-01
-1.40615571e+00 -1.29740074e-01 6.06229305e-01 -5.79184115e-01
4.31141555e-01 1.55755952e-01 2.15759575e-01 7.10576773e-01
8.10210109e-01 7.64134049e-01 9.18595254e-01 -1.94596931e-01
5.43211460e-01 2.65801668e-01 4.75300044e-01 6.26543581e-01
2.89276659e-01 1.76511947e-02 -5.37400544e-01 -4.70853448e-01
7.20512748e-01 2.88118720e-01 -5.04868507e-01 -5.11981308e-01
-8.34726870e-01 8.27707767e-01 2.10474104e-01 -3.83650541e-01
-5.55526078e-01 2.36321017e-01 -1.08275652e-01 -3.56086612e-01
8.10269892e-01 3.77460048e-02 -5.18123806e-01 7.62753412e-02
-5.93316913e-01 1.11126095e-01 7.38591731e-01 9.11106169e-01
5.75685978e-01 1.93492055e-01 -3.97928387e-01 7.86768198e-01
9.78669584e-01 8.21494997e-01 3.47992182e-01 -3.26936215e-01
4.88616943e-01 4.43726182e-01 2.65596241e-01 -9.32579458e-01
-2.55200624e-01 -5.02035081e-01 -1.07054448e+00 7.46748298e-02
3.49464983e-01 -1.39522895e-01 -9.27178144e-01 1.32461405e+00
3.97657335e-01 5.71802795e-01 1.73255205e-01 6.65561199e-01
5.89902103e-01 9.53224242e-01 3.34251910e-01 -1.71608657e-01
1.26816928e+00 -7.23998487e-01 -6.37832582e-01 -3.72225970e-01
-5.16613163e-02 -8.50772679e-01 8.59969497e-01 6.00279391e-01
-6.96400762e-01 -6.74413979e-01 -8.29962313e-01 2.91061163e-01
2.17594862e-01 5.32709420e-01 5.52416623e-01 5.03000379e-01
-5.04087865e-01 6.97504222e-01 -1.08297575e+00 -4.04380867e-03
3.82810235e-01 3.67625207e-01 -1.48784190e-01 3.85807976e-02
-8.82282436e-01 7.10575998e-01 4.90397513e-01 5.16123235e-01
-6.11797929e-01 -9.00016963e-01 -7.68219650e-01 3.86254579e-01
2.92667836e-01 -5.34747243e-01 1.15233254e+00 -3.22732121e-01
-1.64902222e+00 -3.72017175e-02 -5.84528089e-01 -3.62067074e-01
2.70934463e-01 -4.90289479e-01 -2.99131960e-01 -7.10737929e-02
-6.81690574e-02 1.52344778e-01 1.26419508e+00 -1.32234228e+00
-6.43504620e-01 -3.40198874e-01 -6.09654844e-01 1.08695254e-01
2.77089596e-01 -2.35248983e-01 -5.72683454e-01 -5.19662559e-01
4.80726212e-01 -9.48214173e-01 -4.04072344e-01 -2.33184826e-02
-5.36503315e-01 -2.49653578e-01 8.83174956e-01 -8.22882891e-01
9.73698795e-01 -2.29288864e+00 -8.51414446e-03 4.25609797e-01
3.26065272e-01 -2.45666709e-02 3.75599951e-01 2.38855511e-01
-5.55769391e-02 6.34472519e-02 -3.83060157e-01 -5.02320111e-01
1.26781583e-01 2.81987995e-01 -5.42759776e-01 5.94902933e-01
6.18716240e-01 7.16854572e-01 -7.71482229e-01 -2.06538036e-01
4.34825242e-01 7.01374412e-01 -1.00227565e-01 3.65810364e-01
4.13825847e-02 4.12666202e-01 -8.03938925e-01 5.65228999e-01
9.25750673e-01 -3.53653342e-01 -1.49803296e-01 -1.32250115e-01
2.23911300e-01 -3.15889329e-01 -1.48944092e+00 5.61603367e-01
-3.48221123e-01 3.66576672e-01 -1.96237788e-01 -9.35797870e-01
1.45170414e+00 2.69027263e-01 1.10759526e-01 5.42739704e-02
-1.00820653e-01 4.72563505e-02 -2.08153743e-02 -2.74851918e-01
3.41534317e-01 -3.02856684e-01 2.21976161e-01 1.51118621e-01
-1.08218797e-01 -9.94056370e-03 -2.88851410e-01 -2.15560153e-01
4.56320822e-01 9.10017863e-02 5.86266041e-01 -4.05929059e-01
6.95709288e-01 -5.31406462e-01 7.30693042e-01 8.31662238e-01
1.39736459e-01 6.54217422e-01 5.46886146e-01 -1.97859779e-01
-8.19022894e-01 -1.38037479e+00 -5.13499737e-01 5.65379977e-01
3.32653433e-01 8.24300200e-03 -3.36374938e-01 -3.38209301e-01
1.42859444e-01 1.02210200e+00 -5.37627697e-01 -2.61777580e-01
-6.47775158e-02 -1.12939465e+00 -2.29835864e-02 7.60719121e-01
4.49235797e-01 -6.57121599e-01 -1.02415740e-01 3.59013528e-01
1.13353863e-01 -9.55078363e-01 -2.89431781e-01 2.78283179e-01
-1.01851559e+00 -1.03249443e+00 -3.60454500e-01 -3.83941948e-01
8.81110668e-01 2.86856517e-02 6.59083068e-01 -3.49386692e-01
4.70211841e-02 2.09155783e-01 -1.15651406e-01 -6.33412361e-01
-2.12930650e-01 -3.52033347e-01 5.07479645e-02 4.72437769e-01
2.48695374e-01 -2.24882960e-01 -3.36615890e-01 5.69125652e-01
-7.04632759e-01 1.12937488e-01 7.06624031e-01 9.87964392e-01
9.28790271e-01 5.51228821e-01 5.17392933e-01 -9.22818005e-01
7.23398805e-01 -7.19666660e-01 -6.69848502e-01 3.59538257e-01
-7.50322342e-01 3.19784045e-01 7.37994850e-01 -5.24156034e-01
-1.50764370e+00 -5.88182434e-02 2.03284875e-01 -4.23370421e-01
-1.04638994e-01 6.32259071e-01 -4.00178522e-01 8.28684047e-02
2.78544575e-01 3.25708687e-01 6.06439672e-02 -6.38938963e-01
1.54790297e-01 3.48906308e-01 3.64072561e-01 -9.45735693e-01
9.14242923e-01 1.83214515e-01 4.35325533e-01 -8.43676865e-01
-7.69457936e-01 -5.04755735e-01 -5.93641520e-01 -9.40959081e-02
7.08341420e-01 -9.28502798e-01 -6.39310479e-01 4.60196406e-01
-9.67369139e-01 1.33477852e-01 -1.82567433e-01 7.25625873e-01
-4.88239527e-01 3.36344093e-01 -4.11266416e-01 -1.34368098e+00
-2.98528701e-01 -1.25232005e+00 1.12728286e+00 4.58041131e-01
-1.53408617e-01 -1.52673078e+00 -4.50300962e-01 1.66094273e-01
2.17844233e-01 1.81073830e-01 1.06759226e+00 -9.55297053e-01
-4.63136971e-01 -4.97008145e-01 -2.94890195e-01 4.53397989e-01
2.11539507e-01 2.80401647e-01 -9.92875338e-01 -2.00200863e-02
4.27558154e-01 2.67754853e-01 6.28123879e-01 6.57031178e-01
1.47604191e+00 -2.79217362e-01 -4.11522597e-01 3.82531941e-01
1.44137919e+00 1.31991223e-01 5.70253253e-01 -8.85951445e-02
5.61648726e-01 5.37401557e-01 7.78017759e-01 6.24353051e-01
-1.52108613e-02 3.82261693e-01 4.47636247e-01 1.33997157e-01
5.31847715e-01 -4.23299521e-01 3.35945100e-01 4.90514666e-01
-1.34716034e-01 -1.28981829e-01 -8.60678673e-01 -6.04451858e-02
-2.37871480e+00 -7.21410096e-01 -6.02585077e-01 2.49185038e+00
4.23551321e-01 2.03853920e-01 -5.84715486e-01 -2.61802841e-02
8.37297738e-01 -1.75880939e-01 -6.41215622e-01 -2.79618770e-01
6.24084519e-03 1.16404697e-01 5.26649117e-01 6.87983096e-01
-1.08727300e+00 5.65397143e-01 6.48263073e+00 1.11802924e+00
-8.04756820e-01 -2.96840966e-01 8.51229012e-01 3.85260850e-01
-2.18076602e-01 4.43973280e-02 -1.16773069e+00 5.56954443e-01
8.47563684e-01 -4.44991559e-01 1.49560854e-01 1.10217142e+00
6.65130377e-01 -3.50015044e-01 -1.16154075e+00 7.84057736e-01
-1.66940048e-01 -8.75672638e-01 1.02329873e-01 4.44803610e-02
5.95747888e-01 -4.68293846e-01 3.92205626e-01 4.26658720e-01
2.67334998e-01 -1.21200967e+00 6.64626718e-01 1.19101179e+00
1.76408589e-01 -8.88718128e-01 1.10088217e+00 5.16948998e-01
-1.26273394e+00 -5.20187989e-02 -8.00718188e-01 4.57338989e-02
-8.92324280e-03 1.01926482e+00 -9.64316010e-01 7.53904104e-01
4.06567365e-01 7.28480577e-01 -4.75458801e-01 1.17335212e+00
-5.26506305e-01 8.75449777e-01 -2.74841934e-01 4.44940142e-02
-6.25867248e-02 -6.44777477e-01 5.61256826e-01 8.98897767e-01
5.89618921e-01 -7.32690766e-02 2.95165688e-01 1.21387517e+00
3.61569256e-01 3.38385731e-01 -2.55955517e-01 1.51552156e-01
4.65623021e-01 1.08251238e+00 -4.21286315e-01 -9.73848253e-02
-2.84828216e-01 4.74387884e-01 1.83357805e-01 6.81566954e-01
-9.26340163e-01 6.18549809e-03 6.16716564e-01 -2.88459975e-02
3.35605681e-01 -6.60412386e-02 -5.31856596e-01 -9.53752279e-01
1.69492587e-02 -4.46996570e-01 2.01860860e-01 -9.00060833e-01
-1.72530091e+00 3.12490195e-01 1.89198375e-01 -1.21614611e+00
-1.22248352e-01 -7.98027873e-01 -9.45832014e-01 1.37580836e+00
-1.29826379e+00 -1.14948606e+00 -1.29646286e-01 2.91397780e-01
4.81615990e-01 -1.87278420e-01 6.86489701e-01 -4.10987824e-01
-8.20881367e-01 2.33445615e-01 5.88474452e-01 -3.55224222e-01
5.91632307e-01 -1.39025891e+00 -2.11474523e-02 7.02390969e-01
-2.31634408e-01 8.68856490e-01 1.06692159e+00 -7.76381731e-01
-1.20189440e+00 -1.27147901e+00 4.04116988e-01 -3.11031550e-01
1.00807965e+00 -9.96454712e-03 -1.27542531e+00 5.78709126e-01
-3.18185806e-01 -1.68827176e-02 8.55048716e-01 3.27579558e-01
4.82186601e-02 -1.14996120e-01 -1.21480942e+00 4.40292895e-01
5.07641286e-02 -1.08795792e-01 -6.71338677e-01 1.77605942e-01
5.90214610e-01 -4.79736388e-01 -1.00012124e+00 4.21697050e-01
3.45516592e-01 -5.41810811e-01 8.67282033e-01 -4.22220886e-01
4.41837221e-01 -5.39058030e-01 -1.48754001e-01 -1.43305826e+00
-4.82385099e-01 -2.95903534e-01 -4.34361935e-01 1.34635210e+00
5.43322444e-01 -1.03641796e+00 5.82131624e-01 1.20194149e+00
-7.71997720e-02 -1.04944301e+00 -8.42077792e-01 -6.89243972e-01
-1.31482318e-01 -6.40801191e-01 6.09963357e-01 2.37100005e-01
-2.34639391e-01 2.27928191e-01 -4.85153168e-01 9.06210005e-01
1.08700407e+00 3.42520624e-02 6.27387404e-01 -1.54992032e+00
-5.13691127e-01 -2.11407334e-01 -4.07529086e-01 -1.07239199e+00
1.22553214e-01 -6.20032191e-01 2.06676796e-01 -1.53198457e+00
3.12761545e-01 -4.10828292e-01 -1.79053321e-01 1.50353208e-01
-7.48888850e-01 -3.45357507e-01 -7.58865699e-02 4.09095168e-01
-1.17161937e-01 8.84718537e-01 1.08274126e+00 6.79267719e-02
-1.87853679e-01 5.92391968e-01 -7.14574754e-01 1.02101719e+00
8.67931545e-01 -4.24766809e-01 -5.59094369e-01 1.45527139e-01
-9.24824178e-02 9.41995252e-03 3.98355603e-01 -8.55901480e-01
2.06839979e-01 -3.58572334e-01 6.86038196e-01 -6.93177879e-01
4.37694967e-01 -9.64920402e-01 4.58898842e-01 1.25548929e-01
-2.08818197e-01 -4.69319433e-01 2.51485705e-01 1.25829017e+00
-2.52265126e-01 -5.46567857e-01 8.21012676e-01 4.15726513e-01
-4.09228861e-01 5.07087648e-01 -4.15822595e-01 -2.97332972e-01
1.14686489e+00 -5.94913922e-02 -1.40870571e-01 -4.25500989e-01
-8.49933684e-01 4.02774483e-01 7.02002719e-02 1.52955770e-01
8.45059872e-01 -1.24433601e+00 -5.17632008e-01 2.03840330e-01
9.57835689e-02 3.57125521e-01 3.08706433e-01 7.19384491e-01
-1.53078273e-01 4.89290684e-01 3.59898567e-01 -7.94522524e-01
-1.08672011e+00 4.39104646e-01 2.16139644e-01 -4.19330239e-01
-5.86776078e-01 4.76514846e-01 6.15292490e-01 -2.31152371e-01
4.59867269e-02 -5.28351963e-01 -2.84775645e-01 -4.21350151e-01
6.11827195e-01 4.14887816e-01 -2.82399774e-01 -4.46815431e-01
6.25230744e-03 5.34596562e-01 6.36348501e-02 2.49735475e-01
1.27380860e+00 -1.42147750e-01 -9.74704847e-02 8.06468189e-01
7.49320626e-01 4.59908601e-03 -1.64834118e+00 -4.44502681e-01
4.03248146e-02 -5.06600857e-01 6.87713325e-02 -4.04043138e-01
-8.65746021e-01 9.77064312e-01 3.07129741e-01 9.37205181e-02
7.42298722e-01 -1.42388001e-01 1.93507358e-01 1.20619744e-01
2.28322461e-01 -8.10631394e-01 -1.12151340e-01 3.28509808e-01
8.99089813e-01 -1.27870035e+00 7.03592449e-02 -8.33142996e-01
-9.18249488e-01 1.38913167e+00 5.03731132e-01 -5.11760227e-02
9.51952934e-01 3.15262936e-02 -4.38831151e-01 6.85110837e-02
-7.92356491e-01 3.40531617e-01 7.29112923e-01 7.48907506e-01
1.93543673e-01 2.36025602e-01 6.39795437e-02 1.16054165e+00
4.73411344e-02 -2.39748016e-01 3.64083111e-01 4.65417445e-01
-6.72271848e-01 -8.29124212e-01 -7.63719976e-01 6.70572460e-01
-3.53152514e-01 -1.10006414e-01 2.39388496e-01 7.25105047e-01
-2.78005481e-01 9.95839953e-01 -1.03857391e-01 -6.73436150e-02
1.49644762e-01 1.76715061e-01 -8.41153562e-02 -6.72464609e-01
1.18310906e-01 1.39865220e-01 -1.64268374e-01 -1.67897999e-01
1.42334953e-01 -7.58078516e-01 -1.41766894e+00 -2.19932467e-01
-4.50072020e-01 2.92728812e-01 6.66317821e-01 1.10926330e+00
1.43122539e-01 5.34388840e-01 6.05345905e-01 -6.68297231e-01
-9.47665811e-01 -9.31165814e-01 -1.04437089e+00 -1.21735490e-03
-1.05195984e-01 -8.84527206e-01 -6.94744945e-01 8.55106786e-02] | [6.738700866699219, 3.626147985458374] |
10bb2a29-0f13-4865-945c-9ff0dd7ed114 | sam-rl-sensing-aware-model-based | 2210.15185 | null | https://arxiv.org/abs/2210.15185v3 | https://arxiv.org/pdf/2210.15185v3.pdf | SAM-RL: Sensing-Aware Model-Based Reinforcement Learning via Differentiable Physics-Based Simulation and Rendering | Model-based reinforcement learning (MBRL) is recognized with the potential to be significantly more sample-efficient than model-free RL. How an accurate model can be developed automatically and efficiently from raw sensory inputs (such as images), especially for complex environments and tasks, is a challenging problem that hinders the broad application of MBRL in the real world. In this work, we propose a sensing-aware model-based reinforcement learning system called SAM-RL. Leveraging the differentiable physics-based simulation and rendering, SAM-RL automatically updates the model by comparing rendered images with real raw images and produces the policy efficiently. With the sensing-aware learning pipeline, SAM-RL allows a robot to select an informative viewpoint to monitor the task process. We apply our framework to real world experiments for accomplishing three manipulation tasks: robotic assembly, tool manipulation, and deformable object manipulation. We demonstrate the effectiveness of SAM-RL via extensive experiments. Videos are available on our project webpage at https://sites.google.com/view/rss-sam-rl. | ['Cewu Lu', 'Lin Shao', 'Shuang Zhao', 'Cheng Zhang', 'Yunhai Feng', 'Jun Lv'] | 2022-10-27 | null | null | null | null | ['deformable-object-manipulation'] | ['robots'] | [ 4.53909785e-02 -7.39167929e-02 -1.24806359e-01 -2.52122104e-01
-7.96588719e-01 -5.16331255e-01 4.59175736e-01 -1.94272593e-01
-3.27569395e-01 7.34836102e-01 -3.57085347e-01 -1.31774589e-01
-3.75431515e-02 -6.64865077e-01 -1.08537626e+00 -5.12999833e-01
6.16036057e-02 6.49845958e-01 2.45221525e-01 -2.66759247e-01
3.45379561e-01 7.01813459e-01 -1.58190858e+00 -5.79006895e-02
7.10774779e-01 7.85555780e-01 9.23724949e-01 8.63783240e-01
-1.33892559e-02 8.46551478e-01 -3.10015559e-01 4.36086327e-01
4.42952067e-01 -2.68862754e-01 -3.79838169e-01 3.60942222e-02
7.58544877e-02 -6.36804938e-01 -2.06573144e-01 8.93219769e-01
3.81819487e-01 5.87799132e-01 3.49743843e-01 -1.38181210e+00
-2.92965263e-01 1.83827743e-01 -2.19783351e-01 -2.27373764e-01
4.95115519e-01 7.72277832e-01 5.37640333e-01 -6.87109590e-01
7.37656772e-01 1.66577625e+00 1.93912297e-01 7.35036492e-01
-1.09500146e+00 -6.17849112e-01 4.53280181e-01 1.52733311e-01
-8.67104053e-01 -3.04524213e-01 8.65316629e-01 -4.49035913e-01
7.77562141e-01 -2.47566085e-02 9.75022256e-01 1.07371998e+00
5.09769082e-01 9.50657904e-01 1.21668994e+00 -3.05253655e-01
5.21267176e-01 -2.90908128e-01 -4.37559813e-01 1.04507732e+00
-4.86866832e-02 6.31457806e-01 -6.42147183e-01 -2.92220917e-02
1.46507525e+00 1.17627522e-02 -1.05507210e-01 -8.99116755e-01
-1.36699820e+00 5.85012019e-01 4.48807150e-01 -2.64559746e-01
-4.53660816e-01 7.42135227e-01 8.02511126e-02 2.67499745e-01
1.08161293e-01 6.97427988e-01 -4.85974908e-01 -3.62394214e-01
-2.83111691e-01 5.70040524e-01 7.47224033e-01 1.11396968e+00
6.27695560e-01 1.94074422e-01 1.38293087e-01 6.54690802e-01
5.56940913e-01 5.68908215e-01 3.32081228e-01 -1.52720594e+00
1.06545798e-01 3.91753614e-01 5.50091624e-01 -6.02024257e-01
-3.02558273e-01 -3.46241072e-02 -3.56372803e-01 9.04075265e-01
1.62656412e-01 -2.26427004e-01 -7.44356751e-01 1.54618788e+00
7.66024232e-01 2.69204199e-01 -2.93116868e-02 1.13970220e+00
3.44629139e-01 7.79872835e-01 2.10290123e-02 -8.16822052e-02
7.80165553e-01 -1.11756527e+00 -6.09660983e-01 -3.28297049e-01
1.55368835e-01 -4.28895444e-01 1.47206414e+00 6.27050281e-01
-1.14706624e+00 -7.37528443e-01 -1.01140070e+00 -3.15657519e-02
-2.10823547e-02 8.81283581e-02 7.19188571e-01 -3.14189762e-01
-9.31058049e-01 9.09331858e-01 -1.37484753e+00 -1.41070813e-01
2.05875888e-01 3.45746309e-01 -1.34891391e-01 -3.10616978e-02
-7.47583628e-01 1.01572728e+00 3.75839919e-01 1.00600101e-01
-1.50169945e+00 -5.93912244e-01 -9.03151631e-01 -3.34007412e-01
6.78412616e-01 -6.73713446e-01 1.88312995e+00 -7.03295827e-01
-2.27366590e+00 2.36416355e-01 -4.59334329e-02 -1.48679405e-01
6.19078398e-01 -4.47058260e-01 1.94804966e-01 1.33684680e-01
-1.03009894e-01 7.26944745e-01 1.06234181e+00 -1.68427527e+00
-5.31481564e-01 -1.23044938e-01 3.52540135e-01 3.45040858e-01
4.50775206e-01 -3.38022590e-01 -2.19147459e-01 -3.67783606e-01
7.80708939e-02 -1.08533084e+00 -5.82840800e-01 5.43717563e-01
8.25469382e-03 -2.40046993e-01 7.95849442e-01 -3.69872421e-01
3.72278512e-01 -1.91517746e+00 2.90545106e-01 -7.86020905e-02
-1.65204808e-01 2.70350389e-02 -2.34890118e-01 5.45428991e-01
4.80629236e-01 -3.78062725e-01 -5.98084815e-02 -1.03475779e-01
2.05539390e-01 4.08928931e-01 -4.11059678e-01 3.45315218e-01
1.20029293e-01 1.06012726e+00 -1.35219586e+00 -5.17549098e-01
6.66672111e-01 3.52114260e-01 -5.09178281e-01 6.22058988e-01
-8.64329457e-01 1.09775293e+00 -9.09947336e-01 6.07585192e-01
2.59969145e-01 -1.32260382e-01 2.42356658e-01 1.02032190e-02
-2.17255607e-01 7.87306800e-02 -1.04759383e+00 2.07776880e+00
-8.17012966e-01 3.32336277e-01 2.45946541e-01 -7.20861018e-01
1.01752138e+00 9.44521371e-03 4.38634962e-01 -7.04900742e-01
5.82838617e-02 8.70984197e-02 -1.45061180e-01 -6.69619381e-01
3.60333502e-01 1.01346232e-01 -3.80782527e-03 4.87367958e-01
7.38965487e-03 -9.65294778e-01 -4.05845291e-04 -3.42956366e-04
9.78629112e-01 1.14901721e+00 2.15756685e-01 -1.77922882e-02
1.76889509e-01 2.77532935e-01 5.41898847e-01 7.55354702e-01
-2.21263587e-01 4.75313999e-02 9.89782140e-02 -3.61525357e-01
-9.90263045e-01 -1.21756649e+00 3.29765588e-01 1.08259046e+00
5.08766830e-01 -4.53209169e-02 -4.61757064e-01 -3.73975635e-01
1.56099156e-01 8.70767653e-01 -2.58500248e-01 -1.38510272e-01
-7.51202941e-01 -3.19074988e-02 -1.53912380e-01 4.40512270e-01
3.79466146e-01 -1.51551020e+00 -1.27554774e+00 4.18758065e-01
2.81959493e-02 -8.98206532e-01 -1.99990496e-01 -4.46190909e-02
-9.84924376e-01 -1.06776822e+00 -3.55937392e-01 -6.08624399e-01
6.91452920e-01 2.62069166e-01 9.45245862e-01 6.65224195e-02
-3.71907860e-01 9.22952771e-01 -3.34879875e-01 -4.96873617e-01
-7.35999703e-01 -4.28229690e-01 4.03560027e-02 -5.06444097e-01
-4.84212399e-01 -5.21799922e-01 -6.08906686e-01 2.67363846e-01
-7.14910507e-01 5.00746608e-01 5.55983603e-01 6.56422615e-01
9.77988541e-01 -2.71046400e-01 4.65612173e-01 -3.97293955e-01
6.19765282e-01 -1.43377870e-01 -1.22345281e+00 1.17579959e-01
-4.16435003e-01 1.43940195e-01 5.61974704e-01 -6.93643272e-01
-1.17104411e+00 5.62679946e-01 -1.38121201e-02 -5.56947649e-01
-2.06502393e-01 2.22250625e-01 -4.83529679e-02 -2.23542035e-01
3.37993294e-01 2.26051748e-01 1.63064212e-01 -3.75201404e-01
3.53650004e-01 2.19717115e-01 5.30956328e-01 -1.01313674e+00
7.59727120e-01 4.08150762e-01 1.35390371e-01 -5.56355774e-01
-6.28187835e-01 -1.26606867e-01 -5.65366805e-01 -6.76886857e-01
6.21073365e-01 -6.44767106e-01 -1.04465723e+00 4.77905542e-01
-9.12512541e-01 -1.22545910e+00 -4.88697648e-01 5.80091417e-01
-1.21639073e+00 4.60978411e-02 -4.78517622e-01 -9.57155168e-01
-1.00920983e-01 -1.37968850e+00 1.11754358e+00 3.60834241e-01
-8.78928322e-03 -6.64246500e-01 1.22966677e-01 5.14155701e-02
3.56576681e-01 4.65523601e-01 6.36249781e-01 -1.16333239e-01
-9.42634821e-01 8.12693015e-02 2.04926044e-01 1.15815504e-02
7.84048736e-02 5.11380397e-02 -7.97749221e-01 -5.02500296e-01
-1.74454510e-01 -7.73439229e-01 4.98003483e-01 4.61724013e-01
1.49773526e+00 -4.22174074e-02 -3.28380346e-01 3.34629893e-01
1.28836238e+00 4.85108972e-01 9.70883965e-02 3.14955264e-01
5.27788162e-01 3.60312879e-01 1.31801021e+00 5.82020879e-01
3.28433126e-01 7.26958156e-01 8.83380830e-01 2.28615895e-01
4.42149676e-02 -6.04661584e-01 5.35619438e-01 7.25081503e-01
2.42799073e-02 -2.09322236e-02 -6.20172262e-01 8.00966024e-02
-2.05099225e+00 -6.25231445e-01 3.07634264e-01 2.05110121e+00
8.51645112e-01 3.36451195e-02 -8.51591900e-02 -3.82138252e-01
4.94367510e-01 -1.24826254e-02 -1.38526368e+00 -3.80557358e-01
5.34694314e-01 1.38212234e-01 1.38485432e-01 6.16587937e-01
-8.43951285e-01 1.10138369e+00 5.36047125e+00 5.22465050e-01
-1.12688017e+00 3.90388370e-02 1.86798379e-01 -1.08495012e-01
2.75915815e-03 7.89939314e-02 -4.72560287e-01 2.81294137e-01
7.07424462e-01 -1.45884842e-01 9.55884755e-01 1.16307807e+00
5.43569028e-01 -4.81564224e-01 -1.36183548e+00 8.82745206e-01
-3.82797480e-01 -1.27097178e+00 -1.62183329e-01 -1.04521543e-01
4.55533147e-01 8.54994953e-02 -8.71895105e-02 5.40719926e-01
9.81832862e-01 -7.41678059e-01 1.07380605e+00 7.39008486e-01
6.16933405e-01 -6.04980409e-01 3.99410613e-02 7.55963802e-01
-9.92555559e-01 -1.61299482e-01 -4.73482400e-01 -9.03754830e-02
3.60717654e-01 2.36769408e-01 -8.49789500e-01 2.32391164e-01
5.42500794e-01 7.76221335e-01 5.96677363e-02 1.00448275e+00
-5.07693172e-01 2.96358854e-01 -1.71828657e-01 -3.41496795e-01
6.00053705e-02 -2.41502300e-01 6.62100911e-01 7.81692803e-01
1.67635992e-01 1.39740497e-01 7.32289314e-01 1.03699458e+00
1.46839276e-01 -3.11312556e-01 -5.68898082e-01 9.97747928e-02
4.92835522e-01 1.27992594e+00 -5.49859941e-01 -2.25244999e-01
7.50510544e-02 8.01523328e-01 4.90549356e-01 3.23178202e-01
-9.14644718e-01 -6.27633110e-02 5.63590765e-01 -2.11277083e-01
1.20683618e-01 -8.29282284e-01 3.99256319e-01 -9.34051812e-01
-1.38259917e-01 -8.28410268e-01 -2.39243925e-01 -1.22913110e+00
-1.06402266e+00 1.87168360e-01 2.45798126e-01 -1.22299016e+00
-4.40930635e-01 -5.78329921e-01 -2.91139185e-01 4.86425430e-01
-1.66156101e+00 -9.80211735e-01 -5.33729434e-01 4.96591061e-01
1.06870270e+00 8.66278186e-02 7.54672825e-01 -3.07654142e-01
-3.13943982e-01 -2.24543154e-01 -4.61755991e-02 -1.50263712e-01
4.90075499e-01 -1.22839940e+00 4.39212352e-01 2.85251886e-01
-3.76738198e-02 1.87351331e-01 6.91241324e-01 -6.79088235e-01
-2.00905585e+00 -1.07486296e+00 -2.80145317e-01 -2.57830203e-01
7.21653223e-01 -3.58800739e-01 -8.38193417e-01 6.18582428e-01
-2.61440204e-04 1.30068660e-01 -5.75684905e-02 -6.36554003e-01
1.68622568e-01 2.09719669e-02 -1.24221718e+00 7.49065638e-01
1.11958253e+00 -2.08921269e-01 -4.54006851e-01 4.66186792e-01
7.79447019e-01 -1.02699542e+00 -7.63522148e-01 3.15435380e-01
7.73771882e-01 -6.78122640e-01 9.31490958e-01 -3.80932719e-01
4.16932195e-01 -4.16114450e-01 -2.72400677e-02 -1.67317629e+00
-2.12515280e-01 -6.43657386e-01 -3.45464230e-01 7.04058468e-01
3.10732462e-02 -8.46480548e-01 3.95782471e-01 4.58221704e-01
-1.82688162e-01 -8.96735251e-01 -7.66292930e-01 -8.62408578e-01
4.09052595e-02 -4.19853777e-01 6.33005679e-01 4.12629932e-01
-2.22732082e-01 -2.37840161e-01 -8.64818469e-02 3.74896258e-01
6.64433181e-01 4.40417290e-01 1.01732159e+00 -1.06220853e+00
-5.48675776e-01 -1.19872436e-01 -3.56883593e-02 -1.29752290e+00
6.22913659e-01 -4.86996084e-01 7.00003088e-01 -1.81946301e+00
2.38150433e-02 -8.23395073e-01 -1.77813843e-01 5.39248228e-01
1.42407328e-01 -3.55383426e-01 4.81811702e-01 3.12537193e-01
-6.98350251e-01 7.58530498e-01 1.97774088e+00 -2.08843350e-02
-3.14709038e-01 2.07466688e-02 -4.62637544e-02 8.65048170e-01
1.47074723e+00 -3.56951922e-01 -6.05685174e-01 -4.60319787e-01
5.56317419e-02 6.27577603e-01 6.54484928e-01 -9.45918560e-01
4.90262471e-02 -7.77045131e-01 3.84990513e-01 -3.75317752e-01
7.53714859e-01 -8.31022203e-01 8.34549740e-02 8.44921291e-01
-7.11802900e-01 1.84632286e-01 3.50575775e-01 8.04259419e-01
3.11529309e-01 -2.41442502e-01 7.06692636e-01 -4.53935742e-01
-1.00326025e+00 3.51202905e-01 -4.05973941e-01 -8.81537721e-02
1.10088325e+00 2.06581354e-01 -1.54968739e-01 -3.57547790e-01
-6.75487757e-01 3.81750971e-01 6.27886772e-01 4.91448849e-01
9.45315719e-01 -1.04051435e+00 -2.91291475e-01 2.27348253e-01
-9.91817936e-02 3.98561805e-01 4.05749418e-02 4.36711311e-01
-5.57641923e-01 4.39310931e-02 -4.10404325e-01 -6.76423788e-01
-9.91119683e-01 4.22482818e-01 4.92215931e-01 -1.30253255e-01
-8.15420985e-01 6.33243382e-01 5.94084673e-02 -7.90322661e-01
1.86989099e-01 -8.41224670e-01 2.88947672e-01 -9.24214661e-01
2.85629809e-01 4.29767787e-01 -4.18128401e-01 -1.55365869e-01
-1.10304072e-01 4.36305612e-01 2.22024649e-01 -4.43284929e-01
1.49563229e+00 -1.49484619e-03 1.59987986e-01 5.97475290e-01
5.76014459e-01 -2.82704592e-01 -2.27547002e+00 3.55281867e-02
-1.42228663e-01 -4.81000006e-01 1.52650522e-02 -8.93328667e-01
-9.44464147e-01 8.22131217e-01 4.75470364e-01 -2.89230287e-01
6.91057324e-01 -9.51803178e-02 5.99552810e-01 7.60082960e-01
1.04109049e+00 -1.14180887e+00 5.28769672e-01 4.43385422e-01
1.38447142e+00 -1.16631699e+00 4.81588505e-02 -8.82984325e-02
-6.49612725e-01 1.06354535e+00 9.81310666e-01 -5.23128271e-01
3.89934868e-01 5.68716407e-01 3.23232003e-02 2.44321823e-02
-9.03182268e-01 5.69369011e-02 -9.41990092e-02 7.26305604e-01
-9.28061455e-02 1.02179945e-01 4.05040145e-01 -2.06828285e-02
-8.03318694e-02 3.88017356e-01 4.71079618e-01 1.44639158e+00
-6.97593451e-01 -1.10123563e+00 -2.73807347e-01 2.20682517e-01
1.85844943e-01 4.36316520e-01 -2.56369319e-02 7.87490129e-01
-2.41404548e-01 8.21705520e-01 -1.36065304e-01 4.98695374e-02
3.46214682e-01 -2.02599406e-01 1.04936516e+00 -6.55807078e-01
-2.29035988e-01 -5.15371375e-02 -1.93884283e-01 -1.10041821e+00
-4.18108165e-01 -6.10211909e-01 -1.79233015e+00 1.27580017e-01
-6.82839304e-02 -1.18171044e-01 1.02411819e+00 6.21963799e-01
4.20331895e-01 6.20270193e-01 5.84859669e-01 -1.58764398e+00
-8.37431669e-01 -5.71556449e-01 -3.20391238e-01 -6.70601241e-03
5.08495629e-01 -1.15440774e+00 -7.82117397e-02 6.39883354e-02] | [4.678863525390625, 0.7682079076766968] |
56dc67e8-b798-4cad-856e-2f6ab691577d | phee-a-dataset-for-pharmacovigilance-event | 2210.12560 | null | https://arxiv.org/abs/2210.12560v1 | https://arxiv.org/pdf/2210.12560v1.pdf | PHEE: A Dataset for Pharmacovigilance Event Extraction from Text | The primary goal of drug safety researchers and regulators is to promptly identify adverse drug reactions. Doing so may in turn prevent or reduce the harm to patients and ultimately improve public health. Evaluating and monitoring drug safety (i.e., pharmacovigilance) involves analyzing an ever growing collection of spontaneous reports from health professionals, physicians, and pharmacists, and information voluntarily submitted by patients. In this scenario, facilitating analysis of such reports via automation has the potential to rapidly identify safety signals. Unfortunately, public resources for developing natural language models for this task are scant. We present PHEE, a novel dataset for pharmacovigilance comprising over 5000 annotated events from medical case reports and biomedical literature, making it the largest such public dataset to date. We describe the hierarchical event schema designed to provide coarse and fine-grained information about patients' demographics, treatments and (side) effects. Along with the discussion of the dataset, we present a thorough experimental evaluation of current state-of-the-art approaches for biomedical event extraction, point out their limitations, and highlight open challenges to foster future research in this area. | ['Yulan He', 'Joseph Kim', 'Nigel Greene', 'Bino John', 'Byron C. Wallace', 'Gabriele Pergola', 'Jiazheng Li', 'Zhaoyue Sun'] | 2022-10-22 | null | null | null | null | ['event-extraction'] | ['natural-language-processing'] | [ 4.66016054e-01 4.60761115e-02 -7.88774014e-01 -3.61062855e-01
-1.12825668e+00 -8.82197261e-01 5.30356705e-01 1.45038021e+00
-3.09354603e-01 1.13936889e+00 5.37572563e-01 -5.37693501e-01
-2.15798751e-01 -6.19663358e-01 -6.87927127e-01 -3.84614885e-01
-2.05586061e-01 5.54068327e-01 -3.99765074e-01 3.68969202e-01
1.50278345e-01 6.29575193e-01 -9.52805638e-01 6.85849965e-01
9.13970470e-01 7.32605517e-01 -3.55880141e-01 1.23576991e-01
5.87205775e-02 8.39082778e-01 -4.22662675e-01 -4.61540908e-01
-2.67471284e-01 -4.77516204e-01 -6.09380126e-01 -3.01700145e-01
-5.56672625e-02 2.21174099e-02 1.38380066e-01 1.01849782e+00
6.69966698e-01 -2.37946957e-01 4.90308911e-01 -9.10737038e-01
-7.69118071e-01 8.10251534e-01 -3.31750780e-01 2.20068887e-01
7.37106204e-01 2.94885755e-01 9.53936040e-01 -4.28962916e-01
9.23103034e-01 8.69432151e-01 5.86454272e-01 6.40054703e-01
-1.23000968e+00 -6.83967650e-01 6.91781268e-02 -1.14981055e-01
-1.44401860e+00 -5.73910177e-01 1.47902310e-01 -7.92850077e-01
1.15715015e+00 3.78782362e-01 4.44592267e-01 1.20556259e+00
5.28413296e-01 5.01770377e-01 8.44066978e-01 -2.21416224e-02
4.99186724e-01 4.66689393e-02 2.36310169e-01 3.34413677e-01
5.92371166e-01 8.31115693e-02 -5.44399142e-01 -1.07270420e+00
2.88816541e-01 1.96013257e-01 -3.55359107e-01 1.61537305e-01
-1.11310971e+00 7.81597435e-01 -1.45448923e-01 1.67763680e-01
-8.77563298e-01 -2.47803211e-01 4.83704567e-01 -2.16798577e-02
3.73793960e-01 4.14476246e-01 -9.25645947e-01 -7.63258412e-02
-6.43172562e-01 4.17465985e-01 8.44157934e-01 8.36323440e-01
-2.57735942e-02 -5.18384159e-01 -2.28091955e-01 5.39939225e-01
3.43932956e-01 2.21956909e-01 1.91167280e-01 -2.17878625e-01
3.59776616e-01 6.96382046e-01 3.08939368e-01 -5.70293605e-01
-6.62989795e-01 -1.67372435e-01 -7.32433200e-01 -3.72286975e-01
2.67194182e-01 -1.97381034e-01 -7.27553964e-01 1.67650843e+00
4.82340097e-01 1.37139931e-01 6.15853593e-02 5.10834098e-01
1.00237143e+00 3.52338076e-01 8.96845043e-01 -7.28316844e-01
1.88306963e+00 6.86707869e-02 -1.08512866e+00 1.13219082e-01
7.02514768e-01 -8.89861226e-01 3.37769240e-01 6.09742939e-01
-1.08935022e+00 2.50434488e-01 -7.84865260e-01 1.74905494e-01
-4.49554741e-01 -2.45667681e-01 6.88026905e-01 5.80357075e-01
-1.91897407e-01 6.92550063e-01 -9.04538095e-01 -2.99020469e-01
9.87897098e-01 4.33154583e-01 -3.88834924e-01 5.91467954e-02
-1.50496769e+00 1.03537738e+00 2.62639821e-01 -1.47817463e-01
-7.26649582e-01 -1.55715752e+00 -9.70165014e-01 -9.49660093e-02
5.26059926e-01 -9.08185184e-01 1.24236333e+00 1.14025138e-01
-8.81246030e-01 7.46403098e-01 -1.95046589e-01 -6.43426657e-01
2.20673725e-01 -6.02447204e-02 -8.03482890e-01 -1.24284498e-01
3.21188062e-01 7.96899572e-02 -1.33623973e-01 -3.74759763e-01
-7.15328813e-01 -7.43268728e-01 -4.78446096e-01 -3.35653991e-01
1.21133849e-01 6.46549702e-01 5.20077571e-02 -8.35935473e-01
-7.06719339e-01 -6.19233251e-01 -6.95399761e-01 -4.16395873e-01
-5.48327982e-01 -4.89321291e-01 1.49770588e-01 -5.45005739e-01
1.66086817e+00 -1.82048142e+00 -4.58820015e-01 7.07199275e-02
4.05372471e-01 1.26383319e-01 1.35334596e-01 7.56014168e-01
-6.85789287e-01 5.28423786e-01 -1.85315624e-01 3.43702614e-01
-3.20379466e-01 -2.77662843e-01 -2.08317280e-01 8.66036594e-01
5.15847325e-01 1.11429250e+00 -1.01821947e+00 -2.27422416e-01
2.43391275e-01 4.44305718e-01 -5.28610229e-01 -6.67522401e-02
-5.06404459e-01 6.57626152e-01 -1.02892733e+00 6.73601389e-01
3.85542393e-01 -6.09484255e-01 3.49897444e-01 -9.86828580e-02
-2.28891790e-01 8.90492737e-01 -9.36827183e-01 1.44420302e+00
-7.20523223e-02 -1.27113357e-01 -2.72825867e-01 -5.29392540e-01
4.20348167e-01 8.18049729e-01 9.97207344e-01 -5.11142969e-01
2.63746887e-01 5.37503995e-02 -1.88448895e-02 -5.85512817e-01
-2.92921752e-01 -5.13130844e-01 -2.51843363e-01 4.46132749e-01
-3.55770081e-01 4.18636918e-01 5.74532077e-02 6.44918159e-03
1.40463626e+00 -3.03036034e-01 1.07330084e+00 -9.31385458e-02
4.73038435e-01 3.08145553e-01 7.98895121e-01 4.27945554e-01
-1.01566203e-01 2.33119309e-01 4.38954830e-01 -6.03364408e-01
-6.73687875e-01 -8.17211449e-01 -6.86914623e-01 2.73859560e-01
-5.95939696e-01 -7.44153321e-01 -5.12412488e-01 -5.48173964e-01
1.16595410e-01 8.30089211e-01 -7.03286648e-01 2.95289345e-02
-1.93222448e-01 -1.47921252e+00 6.18439794e-01 2.70616770e-01
-2.07087860e-01 -9.57371056e-01 -3.59476000e-01 7.56637454e-01
-3.04469913e-02 -1.04245651e+00 -5.60310662e-01 4.36611362e-02
-6.63836777e-01 -1.77472758e+00 -3.91309351e-01 -4.16346759e-01
2.72257924e-01 -5.41254282e-01 1.13186562e+00 -4.75515932e-01
-6.82403564e-01 -1.86896205e-01 -1.15159109e-01 -9.77133811e-01
-6.40330911e-01 -5.31521678e-01 5.68152554e-02 -2.04764843e-01
9.63685274e-01 -2.45864779e-01 -7.33397365e-01 -5.11370562e-02
-1.07032537e+00 -4.57812697e-01 4.33270186e-01 4.21706557e-01
7.90838122e-01 2.98025720e-02 1.00493002e+00 -1.51477110e+00
9.67849553e-01 -9.06370580e-01 -7.91722655e-01 3.21157902e-01
-8.72318625e-01 4.03255112e-02 5.36665499e-01 -2.08524972e-01
-9.16462362e-01 3.19457799e-01 -3.55447680e-01 3.03705692e-01
-5.49916029e-01 9.65758681e-01 -1.25645563e-01 4.57498670e-01
7.01069355e-01 -2.36829787e-01 -1.96222082e-01 -7.03998983e-01
2.88103938e-01 8.19637477e-01 1.39042541e-01 -2.53491908e-01
3.19339223e-02 2.95867324e-01 5.59723340e-02 -5.23018956e-01
-8.54363501e-01 -5.69053590e-01 2.71690041e-02 4.19258177e-01
9.89971042e-01 -1.03944302e+00 -1.37088811e+00 1.26753123e-02
-1.37992883e+00 3.00464571e-01 -2.04403296e-01 7.15625703e-01
-2.28366628e-01 1.50312707e-01 -8.40691686e-01 -7.00678468e-01
-5.40520370e-01 -1.19650888e+00 8.74087572e-01 -1.20550513e-01
-8.93283427e-01 -9.54891801e-01 4.85586762e-01 1.87340066e-01
-4.52817939e-02 6.92384720e-01 1.25663793e+00 -1.15592563e+00
-2.73877621e-01 -3.76265734e-01 9.39582288e-02 -2.85877675e-01
6.28924608e-01 -6.55746982e-02 -6.85195446e-01 2.00248316e-01
2.47728258e-01 -5.49662225e-02 4.93603230e-01 9.04035270e-01
1.02013230e+00 -5.49964845e-01 -5.61465561e-01 -6.93258876e-03
1.28885078e+00 8.05214942e-01 5.23057342e-01 -9.82854143e-03
3.41356516e-01 6.67419553e-01 5.36456764e-01 7.90841281e-01
2.84824878e-01 5.60283124e-01 -5.31907976e-02 -9.10116211e-02
2.94221938e-01 -1.58546016e-01 8.13186616e-02 5.57336099e-02
1.02894187e-01 -3.55442405e-01 -9.40805376e-01 5.60680687e-01
-1.66222703e+00 -8.34578156e-01 -5.06282210e-01 2.27240181e+00
1.45524776e+00 -1.60104901e-01 1.60667494e-01 -3.21028888e-01
8.27001154e-01 -4.07700270e-01 -6.80795193e-01 -3.88644308e-01
-7.97287002e-02 5.62134922e-01 6.76219642e-01 2.08508849e-01
-1.13821304e+00 5.60901225e-01 6.91757679e+00 5.67009985e-01
-5.95074117e-01 -2.48681586e-02 8.99168432e-01 -1.22157611e-01
-2.85795242e-01 -2.22029164e-01 -7.44368076e-01 5.95196366e-01
1.45017266e+00 -5.56588411e-01 8.99247751e-02 4.40720201e-01
8.78968418e-01 7.56726414e-02 -1.58887219e+00 8.89631152e-01
-3.11598569e-01 -1.92260551e+00 5.66444956e-02 2.32176095e-01
5.51497042e-01 2.29215827e-02 -1.46176964e-01 -5.17082810e-01
5.95236659e-01 -1.20769668e+00 3.05748403e-01 4.15064991e-01
7.82766163e-01 -7.10587919e-01 7.32058346e-01 1.21038802e-01
-9.10167754e-01 2.70847648e-01 1.59362525e-01 2.52831936e-01
4.79540110e-01 9.99321163e-01 -1.06467783e+00 7.40331590e-01
4.33653980e-01 1.15993452e+00 -1.34944364e-01 1.08415818e+00
-1.13113277e-01 5.78528821e-01 -9.79882404e-02 9.09377784e-02
-3.20431590e-02 4.00311761e-02 2.67302155e-01 1.18526220e+00
1.34384736e-01 5.65081418e-01 5.43650612e-02 9.85494912e-01
-3.17861736e-01 3.44335169e-01 -7.62160957e-01 -7.44874239e-01
3.78342688e-01 7.80781031e-01 -4.19732094e-01 -3.56834888e-01
-6.56746924e-01 2.59224951e-01 -2.29375571e-01 7.10674897e-02
-7.40826070e-01 -7.34838843e-02 7.82580376e-01 1.90529421e-01
-2.16775462e-01 5.34889579e-01 -4.71208364e-01 -8.69277358e-01
-2.03486785e-01 -1.19439280e+00 1.15198469e+00 -2.31258020e-01
-1.66507554e+00 3.87483656e-01 -1.06747203e-01 -1.17293453e+00
-2.44797230e-01 -4.36787099e-01 -2.51642168e-02 1.10580909e+00
-1.18676078e+00 -6.52912498e-01 5.40320635e-01 3.66189271e-01
3.38294238e-01 1.14742681e-01 1.09745157e+00 7.31978297e-01
-6.82036161e-01 3.09571117e-01 -2.32945397e-01 2.54592951e-02
8.69109988e-01 -9.00897503e-01 5.28704166e-01 2.21561089e-01
-1.47356391e-01 1.12296128e+00 7.76233375e-01 -1.23828697e+00
-1.42196083e+00 -1.38287294e+00 1.27788377e+00 -8.94029081e-01
9.49152768e-01 -1.53532818e-01 -8.92689466e-01 5.93823314e-01
6.65905103e-02 -4.09970999e-01 1.53647602e+00 1.31038740e-01
-2.86645025e-01 2.65958548e-01 -1.31083643e+00 5.63803911e-01
5.96026063e-01 -3.44913185e-01 -7.77231336e-01 7.64116287e-01
5.47331870e-01 -1.25939906e-01 -1.44494951e+00 4.27247077e-01
3.97737473e-01 -2.50856221e-01 9.23364699e-01 -1.48086250e+00
3.58296156e-01 -4.77816224e-01 3.81819934e-01 -1.06596529e+00
-1.91265777e-01 -8.40908349e-01 7.44337514e-02 1.05443990e+00
1.02769053e+00 -6.27023220e-01 6.37261033e-01 1.05200744e+00
7.20286295e-02 -7.10700929e-01 -5.87420821e-01 -3.12023997e-01
-1.28055736e-01 -5.97321272e-01 7.41308510e-01 1.22754216e+00
5.45829117e-01 6.03121102e-01 -1.78610265e-01 3.15328807e-01
5.84782958e-01 1.47340417e-01 4.78914008e-02 -1.31357598e+00
-4.56848554e-02 -2.67096013e-01 -2.55965233e-01 -2.85686493e-01
-1.75114274e-01 -8.71406496e-01 -2.96736509e-01 -1.50874627e+00
7.56812334e-01 -2.24494308e-01 -4.29372102e-01 7.08725452e-01
-2.79441088e-01 -5.99564463e-02 -5.77338278e-01 -1.25841975e-01
-3.48064631e-01 7.48785585e-02 9.06801999e-01 -1.78558886e-01
-4.35925335e-01 -2.41048485e-02 -1.15478170e+00 6.49138927e-01
6.57074511e-01 -9.67508495e-01 -1.74155921e-01 1.85682416e-01
4.22929823e-01 2.45481879e-01 1.03067093e-01 -1.76658891e-02
2.96504885e-01 -6.03182614e-01 4.30874139e-01 -7.43533313e-01
-4.10709828e-01 -6.43238425e-01 7.60953367e-01 6.96629345e-01
-6.77511811e-01 1.20444424e-01 4.54472303e-01 7.74090528e-01
-2.56867439e-01 9.90618393e-02 5.71161032e-01 -2.35609770e-01
-7.01572895e-02 5.23854613e-01 -6.22389853e-01 1.86759442e-01
1.14299297e+00 4.33461607e-01 -3.98383349e-01 9.38729420e-02
-9.63813424e-01 4.03034300e-01 6.48368895e-02 5.62025249e-01
4.37927425e-01 -1.16496968e+00 -9.98456359e-01 -1.16265491e-01
5.46380222e-01 -1.79778829e-01 4.87781823e-01 8.84820163e-01
-2.86040545e-01 8.65538836e-01 3.31637174e-01 -1.61170185e-01
-1.23386002e+00 1.03031647e+00 -2.75250711e-03 -4.51097876e-01
-4.47320104e-01 3.97154301e-01 3.13630700e-01 1.69216380e-01
1.23463668e-01 -4.36126143e-01 -3.40001673e-01 9.65497717e-02
1.11095631e+00 1.31576955e-01 6.27733529e-01 -2.73272604e-01
-8.33972454e-01 -1.01311885e-01 -4.90142733e-01 2.24086553e-01
1.70617247e+00 2.12559372e-01 -3.50150198e-01 6.53736293e-02
1.09337473e+00 1.68323696e-01 -3.29517186e-01 8.83190557e-02
4.38511670e-01 -1.68838017e-02 -1.23426363e-01 -1.25486040e+00
-6.36483133e-01 3.71937901e-01 2.96291292e-01 7.91274607e-02
9.88616645e-01 2.92173088e-01 5.95676005e-01 -4.66594584e-02
2.82086492e-01 -7.87148893e-01 -7.03324437e-01 -1.02725796e-01
7.07964480e-01 -1.06033003e+00 1.77013189e-01 -7.20947742e-01
-6.73685551e-01 6.78119600e-01 -2.01496378e-01 4.38989192e-01
7.55281210e-01 4.16418105e-01 -1.30798638e-01 -6.38945699e-01
-1.09960353e+00 1.42135635e-01 2.69368410e-01 3.18829417e-01
9.24751997e-01 2.01991618e-01 -6.82900786e-01 1.12706590e+00
3.21620882e-01 5.50331175e-01 3.80979240e-01 1.00918138e+00
1.39683232e-01 -1.64435756e+00 -3.19485426e-01 7.55307674e-01
-1.33127069e+00 -5.59293568e-01 -6.29851878e-01 4.89191204e-01
7.98186436e-02 1.28447080e+00 -4.97333288e-01 2.02212870e-01
9.07954276e-01 2.07954030e-02 2.45629326e-01 -9.23338771e-01
-8.46170962e-01 3.92374128e-01 4.12401736e-01 -6.05684817e-01
-5.08022726e-01 -8.28326404e-01 -1.34092546e+00 -9.25403461e-02
-7.53024891e-02 4.42383826e-01 4.11104500e-01 8.24125767e-01
7.71263838e-01 6.00141227e-01 2.63088465e-01 3.34657729e-01
-2.85493255e-01 -4.90926147e-01 -5.67579687e-01 4.17931467e-01
3.83139700e-01 -3.61356854e-01 1.16146125e-01 3.82387251e-01] | [8.384725570678711, 8.661361694335938] |
66bcdac4-f3fd-49dc-b9ab-f8ad121b9f7c | relationship-explainable-multi-objective | 1909.12268 | null | https://arxiv.org/abs/1909.12268v1 | https://arxiv.org/pdf/1909.12268v1.pdf | Relationship Explainable Multi-objective Reinforcement Learning with Semantic Explainability Generation | Solving multi-objective optimization problems is important in various applications where users are interested in obtaining optimal policies subject to multiple, yet often conflicting objectives. A typical approach to obtain optimal policies is to first construct a loss function that is based on the scalarization of individual objectives, and then find the optimal policy that minimizes the loss. However, optimizing the scalarized (and weighted) loss does not necessarily provide guarantee of high performance on each possibly conflicting objective because it is challenging to assign the right weights without knowing the relationship among these objectives. Moreover, the effectiveness of these gradient descent algorithms is limited by the agent's ability to explain their decisions and actions to human users. The purpose of this study is two-fold. First, we propose a vector value function based multi-objective reinforcement learning (V2f-MORL) approach that seeks to quantify the inter-objective relationship via reinforcement learning (RL) when the impact of one objective on others is unknown a prior. In particular, we construct one actor and multiple critics that can co-learn the policy and inter-objective relationship matrix (IORM), quantifying the impact of objectives on each other, in an iterative way. Second, we provide a semantic representation that can uncover the trade-off of decision policies made by users to reconcile conflicting objectives based on the proposed V2f-MORL approach for the explainability of the generated behaviors subject to given optimization objectives. We demonstrate the effectiveness of the proposed approach via a MuJoCo based robotics case study. | ['Huixin Zhan', 'Yongcan Cao'] | 2019-09-26 | null | null | null | null | ['multi-objective-reinforcement-learning'] | ['methodology'] | [ 1.47057772e-01 8.49606842e-02 -2.84915447e-01 -6.72932416e-02
-5.20747066e-01 -2.80403286e-01 1.75923824e-01 3.66901904e-01
-4.63502079e-01 9.53772068e-01 -4.30395082e-02 -6.00436702e-02
-8.54621649e-01 -5.59775352e-01 -7.77990818e-01 -6.74567401e-01
-1.42967507e-01 3.77263933e-01 -2.80638158e-01 -3.49458069e-01
5.00573874e-01 3.01913053e-01 -1.47032177e+00 -3.04194689e-01
1.21558619e+00 9.84762728e-01 5.44508994e-01 5.69618285e-01
2.62373418e-01 6.79297328e-01 -5.79462767e-01 -1.01659872e-01
2.27889389e-01 -2.18392864e-01 -5.98546982e-01 3.41400474e-01
-7.52378330e-02 -2.12766156e-01 2.74034351e-01 1.14147234e+00
2.20803797e-01 4.34021503e-01 7.41783202e-01 -1.54049826e+00
-5.01052558e-01 4.65028465e-01 -5.90507269e-01 -1.07191771e-01
1.60811260e-01 3.60310286e-01 1.37311840e+00 -2.79360414e-01
2.92249203e-01 1.33295512e+00 2.87004337e-02 3.00218582e-01
-1.35462809e+00 -3.12138915e-01 5.11954486e-01 1.79150313e-01
-1.03380537e+00 1.13238124e-02 6.95859551e-01 -5.65314472e-01
4.25883383e-01 2.73065437e-02 5.01391172e-01 8.41482937e-01
3.50064188e-01 7.77363300e-01 1.00532436e+00 -1.75935552e-01
4.10775334e-01 4.24111962e-01 -2.12104106e-03 6.61997557e-01
3.34974259e-01 2.44530573e-01 -2.83047110e-01 -8.15555379e-02
4.89143908e-01 -1.78840496e-02 -2.50072509e-01 -5.29539406e-01
-8.80079687e-01 8.32681119e-01 3.32492292e-01 2.30894968e-01
-6.92114711e-01 3.41848254e-01 4.53837551e-02 3.13859254e-01
7.57191628e-02 9.61207867e-01 -4.48969871e-01 -1.28189102e-02
-4.68276978e-01 4.49850053e-01 4.82129961e-01 5.56920886e-01
8.58885109e-01 2.39971280e-01 -3.65970671e-01 5.36415577e-01
5.41286170e-01 2.49046177e-01 3.56171101e-01 -1.07547212e+00
5.86042941e-01 5.30891776e-01 7.46155798e-01 -1.33670568e+00
-2.66566247e-01 -5.87786794e-01 -4.74114448e-01 7.35566974e-01
2.97045499e-01 -4.40444052e-01 -4.02626336e-01 2.00216866e+00
2.64938802e-01 -7.28595927e-02 1.53392047e-01 1.13176405e+00
-8.07560012e-02 5.09310663e-01 1.15989335e-01 -3.47336322e-01
9.23300922e-01 -7.96140969e-01 -7.81893432e-01 -3.96057189e-01
2.50793844e-01 -3.45591009e-01 1.12996030e+00 2.65075237e-01
-1.06622648e+00 -3.43058407e-01 -1.17019069e+00 6.53953075e-01
-8.24222416e-02 2.88632721e-01 3.08490574e-01 2.89631546e-01
-6.94654584e-01 8.03184092e-01 -5.18061340e-01 -4.68201227e-02
2.24044055e-01 6.57356441e-01 1.65065095e-01 3.48700285e-01
-1.12587416e+00 1.13747334e+00 5.83209157e-01 1.12639405e-01
-1.26470482e+00 -6.57093346e-01 -5.51749408e-01 3.63174379e-01
1.02999794e+00 -5.64068258e-01 1.12162542e+00 -1.25770533e+00
-1.64995503e+00 1.29233748e-01 2.23611414e-01 -4.24771875e-01
7.10698724e-01 -2.14983180e-01 -1.51961938e-01 -9.41550881e-02
5.76280095e-02 3.76641780e-01 1.00445712e+00 -1.57416415e+00
-8.92406166e-01 -3.22804332e-01 5.77697933e-01 5.30262172e-01
-5.25237739e-01 -3.43149513e-01 -7.00321198e-02 -3.41383994e-01
-5.50472260e-01 -7.83604860e-01 -4.03743923e-01 2.98084430e-02
-4.30193931e-01 -9.17168409e-02 7.02640712e-01 -6.33396029e-01
1.24439609e+00 -1.86054111e+00 6.78890705e-01 3.45925272e-01
1.16578966e-01 1.76490813e-01 -3.06357503e-01 2.71951944e-01
3.76300842e-01 1.80685267e-01 -3.02153885e-01 -2.69746512e-01
1.06126800e-01 2.67305911e-01 -8.51475447e-02 3.38847339e-01
2.84986913e-01 3.99346441e-01 -9.78752196e-01 -2.24998355e-01
2.75199652e-01 2.62714982e-01 -5.73130786e-01 5.26408732e-01
-5.66619098e-01 6.43619120e-01 -8.67394567e-01 3.93610597e-01
1.03517279e-01 -2.22116739e-01 4.35861677e-01 -1.92929059e-01
-1.97503090e-01 -3.98040175e-01 -1.48545468e+00 1.14450562e+00
-6.82774782e-01 1.64399683e-01 2.92664886e-01 -1.30616164e+00
8.57485175e-01 2.76175320e-01 8.30317378e-01 -4.83517975e-01
4.17030275e-01 1.85591772e-01 -7.64374807e-02 -5.34227788e-01
3.83327752e-01 -5.84769137e-02 3.26555542e-04 3.72051477e-01
-1.34009466e-01 1.03785224e-01 3.04849714e-01 -1.72552094e-01
7.15099096e-01 9.30624977e-02 5.81637979e-01 -2.13213310e-01
8.09270442e-01 -1.96624711e-01 6.63322270e-01 7.18055964e-01
-2.37605929e-01 9.09523591e-02 7.46071935e-01 -1.23659343e-01
-8.32263112e-01 -6.15776300e-01 3.77651036e-01 8.78445089e-01
5.08854508e-01 1.81458771e-01 -3.84301126e-01 -7.35525668e-01
2.90371120e-01 1.01045156e+00 -6.55362904e-01 -3.20716262e-01
-2.41740614e-01 -5.04005611e-01 -2.84355599e-02 1.26675501e-01
2.47893587e-01 -9.21228051e-01 -1.03178942e+00 4.04928178e-01
-6.96789995e-02 -8.55363011e-01 -5.79106987e-01 6.57194704e-02
-5.91486514e-01 -1.19905972e+00 -5.21957457e-01 -3.48023653e-01
8.19656551e-01 -1.62443947e-02 6.66972160e-01 1.22892402e-01
-2.36872528e-02 6.09823763e-01 -1.20729133e-01 -3.47393960e-01
-3.45098585e-01 -5.73836192e-02 2.66930878e-01 5.82149029e-01
-4.27675426e-01 -5.27695477e-01 -5.02010345e-01 4.91926610e-01
-9.33851063e-01 -1.27273500e-01 6.56652391e-01 6.76025212e-01
5.15086532e-01 2.79482216e-01 7.69198418e-01 -5.61883450e-01
1.09274244e+00 -6.91846371e-01 -1.01513028e+00 5.95206082e-01
-9.23049033e-01 6.42690301e-01 8.84635568e-01 -6.09521866e-01
-9.01667655e-01 -7.60627016e-02 3.66294205e-01 -6.48719728e-01
1.92408651e-01 6.07175469e-01 -1.81806803e-01 7.53052579e-03
1.91269666e-01 5.02526760e-02 9.01156366e-02 -1.87399641e-01
3.82201254e-01 3.21601331e-01 2.29619741e-01 -9.43075418e-01
6.27147555e-01 1.22877218e-01 7.61361122e-02 -4.94321436e-01
-9.59862411e-01 -2.33336359e-01 -1.12996019e-01 -5.97299695e-01
9.17493045e-01 -3.00266474e-01 -1.43657148e+00 -1.36411535e-02
-1.08589268e+00 -1.38497055e-01 -1.94014311e-01 4.99183804e-01
-8.64161372e-01 9.57421884e-02 1.91750526e-01 -1.22314513e+00
-1.41513780e-01 -1.47972918e+00 5.49644530e-01 3.90299737e-01
-3.11739799e-02 -1.05085599e+00 7.66229955e-03 4.08661187e-01
3.73263091e-01 3.94733250e-01 1.09468603e+00 -4.66060072e-01
-5.30767322e-01 4.60657924e-02 -3.49376202e-02 2.98366666e-01
1.84850931e-01 -1.42052040e-01 -3.16840261e-01 -5.18628299e-01
-6.33352473e-02 -4.31779146e-01 2.21248254e-01 5.12424052e-01
1.05730951e+00 -7.92956114e-01 1.44426338e-02 1.69118941e-01
1.74873078e+00 5.23869753e-01 1.99277818e-01 5.55631876e-01
5.01572192e-01 6.81468785e-01 8.43297601e-01 8.48776102e-01
4.98854101e-01 8.41594934e-01 1.08273518e+00 1.53407097e-01
4.54939812e-01 -1.22078113e-01 5.58346927e-01 1.58184588e-01
-1.85972884e-01 -3.35709572e-01 -5.77925920e-01 5.15802860e-01
-2.33161807e+00 -7.48579621e-01 4.40593541e-01 2.39447832e+00
5.22069693e-01 8.25919285e-02 1.97995871e-01 -1.72457829e-01
9.00335848e-01 -2.66186465e-02 -1.03323889e+00 -3.86406630e-01
2.58834779e-01 -3.05501044e-01 5.13859868e-01 7.30613470e-01
-8.36558521e-01 5.56182921e-01 5.29606581e+00 6.39405072e-01
-1.07899702e+00 -1.46993026e-01 5.34072757e-01 -8.83760229e-02
-3.31046909e-01 2.72787246e-03 -6.78274691e-01 4.91080999e-01
5.25045455e-01 -4.51809943e-01 9.27595139e-01 7.60418236e-01
8.13401341e-01 -1.10660143e-01 -1.04593134e+00 6.33257270e-01
-2.10537583e-01 -1.00372291e+00 -1.07744105e-01 2.03894049e-01
8.33709538e-01 -5.90657592e-01 2.61491776e-01 2.08695680e-01
5.64695001e-01 -8.92808378e-01 9.36387420e-01 7.33339190e-01
2.25513577e-01 -9.65695262e-01 5.81397057e-01 5.54284394e-01
-9.69392180e-01 -6.42455459e-01 -8.45125616e-02 6.69437945e-02
7.63036683e-02 1.96833849e-01 -6.50113761e-01 5.62793434e-01
2.52057195e-01 5.28283417e-01 5.82954548e-02 8.67134929e-01
-3.92119229e-01 1.68516442e-01 2.04938799e-02 -3.76329958e-01
4.94007945e-01 -4.75527644e-01 8.78608227e-01 5.18207669e-01
3.82120520e-01 -1.36875004e-01 6.11993849e-01 1.17966330e+00
6.77115396e-02 2.00741902e-01 -4.10453409e-01 -2.21659496e-01
1.89344332e-01 1.18528485e+00 -3.72427106e-01 2.81951781e-02
-2.20033482e-01 6.10792100e-01 6.74988568e-01 4.15629029e-01
-8.62713456e-01 -1.68717653e-01 9.93439615e-01 -1.97818801e-01
2.78314143e-01 -2.31019258e-01 -2.45412469e-01 -9.04773474e-01
-3.66443396e-02 -1.01583731e+00 2.93253243e-01 -4.43696111e-01
-1.05913079e+00 1.76094890e-01 -1.00363903e-01 -1.24030495e+00
-1.65257454e-01 -4.07970339e-01 -4.50406134e-01 6.69820845e-01
-1.74398482e+00 -7.39507616e-01 2.25274935e-02 3.92971426e-01
5.65658450e-01 -1.97395772e-01 4.03162003e-01 1.25172228e-01
-8.49765241e-01 3.30537826e-01 2.10558668e-01 -5.06897509e-01
3.16552341e-01 -1.15032935e+00 -7.52970874e-01 7.15455532e-01
-2.93127418e-01 2.42752224e-01 1.07555568e+00 -5.23087919e-01
-1.50065279e+00 -1.01219082e+00 3.80165756e-01 1.14422530e-01
7.52765775e-01 3.15011919e-01 -6.14678800e-01 3.98494214e-01
3.64605226e-02 -3.87834668e-01 2.66647041e-01 -1.73177660e-01
1.93671808e-01 -2.74961889e-01 -1.19813120e+00 7.75301874e-01
3.62065732e-01 -1.21655138e-02 -2.94278711e-01 1.79835007e-01
6.39556885e-01 -6.79452345e-02 -7.82097399e-01 2.67611206e-01
5.28545797e-01 -6.18998587e-01 8.83111298e-01 -9.65332210e-01
7.19902515e-01 -3.73336226e-01 -2.74753511e-01 -1.67546749e+00
-1.74622640e-01 -5.71453512e-01 -2.23168612e-01 9.40160394e-01
3.86175543e-01 -5.69828629e-01 4.89259809e-01 7.58057833e-01
7.78529644e-02 -1.08581221e+00 -6.86929405e-01 -8.06353927e-01
-3.09943736e-01 -2.02818424e-01 4.83874649e-01 8.01810145e-01
-2.34823883e-01 1.21496409e-01 -9.05286789e-01 5.68630099e-01
8.43693435e-01 1.53029054e-01 5.64683378e-01 -8.84836257e-01
-6.02607369e-01 -6.06192768e-01 -1.17179483e-01 -6.79375470e-01
2.33444557e-01 -6.84297860e-01 3.17619532e-01 -1.41917753e+00
-6.75549805e-02 -4.95179325e-01 -7.05433547e-01 4.81985956e-01
-2.83862680e-01 -5.80180049e-01 5.51552892e-01 1.35538101e-01
-6.23341799e-01 8.97947609e-01 1.41965020e+00 -2.63327748e-01
-4.94505048e-01 2.27118075e-01 -8.38755786e-01 6.82998836e-01
7.32196093e-01 -5.12787163e-01 -5.50079584e-01 -4.82535273e-01
2.33521000e-01 6.06399357e-01 2.72285104e-01 -8.59962225e-01
2.20575914e-01 -9.08751130e-01 -2.10471988e-01 -9.41102207e-02
2.04235196e-01 -1.18339956e+00 4.19263588e-03 6.36445165e-01
-6.76342607e-01 4.97787893e-02 -1.41031697e-01 8.00232530e-01
-1.34701222e-01 -4.97653037e-01 8.67817998e-01 -9.34843197e-02
-6.17592216e-01 2.84089893e-01 -3.27343673e-01 -4.27399995e-03
1.29223609e+00 1.95070982e-01 1.63706586e-01 -5.72310984e-01
-6.40391052e-01 8.08819115e-01 -4.21115980e-02 4.20196474e-01
5.70171356e-01 -1.03416228e+00 -6.41507328e-01 -3.53959739e-01
-1.10714309e-01 -5.10779560e-01 1.01421922e-02 6.51536345e-01
-7.44269490e-02 2.56247640e-01 -4.30648446e-01 -2.31945708e-01
-1.12749600e+00 4.33156043e-01 4.92412150e-01 -8.06787252e-01
-5.31595908e-02 3.21825951e-01 1.26952063e-02 -2.72546262e-01
3.65808487e-01 -2.90806349e-02 -6.13556623e-01 1.01031132e-01
1.10053271e-01 6.34410501e-01 -3.69493037e-01 -2.74638176e-01
-3.54599357e-02 4.81543660e-01 9.49457958e-02 -2.44520694e-01
1.45662963e+00 -2.18650967e-01 2.86985219e-01 2.82438397e-01
9.68793273e-01 -3.83362621e-01 -1.52996886e+00 -1.36326090e-01
2.74438113e-02 -5.04259467e-01 1.41416535e-01 -9.16057706e-01
-1.05898869e+00 5.34376502e-01 5.06276071e-01 1.02811366e-01
1.07362258e+00 -4.62648481e-01 4.71122384e-01 4.73725855e-01
2.40187570e-01 -1.33117235e+00 4.91412699e-01 2.85161376e-01
9.40004230e-01 -1.24005604e+00 -8.09087306e-02 3.62357609e-02
-8.59282851e-01 1.17409682e+00 8.12605202e-01 -2.77112145e-02
3.46596837e-01 -1.36920527e-01 -1.05514467e-01 -1.45541951e-01
-6.10273063e-01 -3.24629426e-01 3.77704769e-01 1.39615133e-01
-1.47721115e-02 1.90616190e-01 -4.94229585e-01 3.56368124e-01
2.94462919e-01 -1.87712848e-01 4.47304130e-01 9.00511086e-01
-5.40996850e-01 -1.21600771e+00 -4.26595092e-01 3.40757221e-01
-2.20532879e-01 5.18287003e-01 -1.64373834e-02 5.60461879e-01
6.21177536e-03 1.05035043e+00 -3.19440901e-01 -2.85394549e-01
4.80228007e-01 -3.04094255e-01 2.31227309e-01 -3.93462300e-01
-4.93568867e-01 -1.61284715e-01 -1.02185354e-01 -5.67676842e-01
-3.46759558e-01 -4.58046407e-01 -1.20418060e+00 7.03677628e-03
-1.84746012e-01 1.37645617e-01 8.16333055e-01 1.26354027e+00
2.60657281e-01 8.22804511e-01 1.07674170e+00 -6.65489733e-01
-1.14446437e+00 -3.93192798e-01 -5.13615310e-01 4.92842972e-01
4.19201612e-01 -9.46785748e-01 -3.50369096e-01 -3.47765714e-01] | [4.348354339599609, 2.420440673828125] |
f359f787-88c0-493e-b8fd-2b7f2d053839 | clevr-math-a-dataset-for-compositional | 2208.05358 | null | https://arxiv.org/abs/2208.05358v1 | https://arxiv.org/pdf/2208.05358v1.pdf | CLEVR-Math: A Dataset for Compositional Language, Visual and Mathematical Reasoning | We introduce CLEVR-Math, a multi-modal math word problems dataset consisting of simple math word problems involving addition/subtraction, represented partly by a textual description and partly by an image illustrating the scenario. The text describes actions performed on the scene that is depicted in the image. Since the question posed may not be about the scene in the image, but about the state of the scene before or after the actions are applied, the solver envision or imagine the state changes due to these actions. Solving these word problems requires a combination of language, visual and mathematical reasoning. We apply state-of-the-art neural and neuro-symbolic models for visual question answering on CLEVR-Math and empirically evaluate their performances. Our results show how neither method generalise to chains of operations. We discuss the limitations of the two in addressing the task of multi-modal word problem solving. | ['Savitha Sam Abraham', 'Adam Dahlgren Lindström'] | 2022-08-10 | null | null | null | null | ['mathematical-reasoning'] | ['natural-language-processing'] | [ 2.91339844e-01 1.69767514e-01 3.71294111e-01 -1.98643655e-01
-3.94128174e-01 -1.02993298e+00 9.25201952e-01 4.45340693e-01
-2.57072747e-01 3.21721524e-01 1.37762010e-01 -6.85730159e-01
-2.39806324e-01 -1.06810057e+00 -9.09795880e-01 -2.69739926e-01
2.56790638e-01 8.19733024e-01 2.34660819e-01 -4.70693946e-01
3.81432325e-01 4.34800386e-01 -1.42635942e+00 1.11709011e+00
2.79411048e-01 7.54066944e-01 9.74778309e-02 9.91064727e-01
-6.49991035e-01 1.63790512e+00 -7.82783210e-01 -5.51248074e-01
5.41396700e-02 -5.09698808e-01 -1.42698097e+00 3.11380118e-01
9.16680396e-01 -2.66134650e-01 -3.93896401e-01 1.11470079e+00
-1.81072518e-01 3.70997548e-01 8.00100923e-01 -1.23889709e+00
-8.20912838e-01 5.98580360e-01 -2.60664701e-01 3.21777612e-01
1.03891134e+00 6.73833787e-01 9.81227875e-01 -4.95049417e-01
8.82159233e-01 1.69982445e+00 3.37121844e-01 3.62069726e-01
-1.52225351e+00 1.11381508e-01 3.27452183e-01 7.13917851e-01
-1.26601470e+00 -4.31086779e-01 4.82365668e-01 -5.81203759e-01
1.35136366e+00 4.92324352e-01 6.59573197e-01 6.88369572e-01
1.51902214e-01 5.67938447e-01 1.07260060e+00 -4.62753564e-01
4.27594811e-01 1.23114174e-03 1.50829449e-01 1.11650658e+00
-2.16068283e-01 -1.37957886e-01 -3.41683418e-01 3.80866416e-03
1.01569521e+00 -4.24881101e-01 -9.92012024e-02 -5.32081902e-01
-1.38981807e+00 6.42055631e-01 4.36635166e-01 3.82742673e-01
-3.23883832e-01 7.05363154e-01 3.04379046e-01 3.58135909e-01
-1.55445337e-01 7.43488550e-01 -1.98617324e-01 3.56023572e-02
-8.86406243e-01 6.35063589e-01 7.85707712e-01 7.58116484e-01
4.94243383e-01 2.95447633e-02 -3.08322847e-01 6.91271052e-02
2.52076030e-01 1.80419117e-01 1.09284192e-01 -1.28178024e+00
6.24119282e-01 9.08957601e-01 4.03621569e-02 -1.02750385e+00
-2.93834239e-01 2.33656630e-01 -2.04698920e-01 5.68308830e-01
9.73547816e-01 3.95386517e-01 -1.09625614e+00 1.30000365e+00
2.29423955e-01 3.17974538e-01 1.12748198e-01 9.26351607e-01
1.37524426e+00 1.04960585e+00 2.67883748e-01 9.36144814e-02
1.65151179e+00 -7.89232135e-01 -6.85994506e-01 -5.10446608e-01
6.49151802e-01 -3.68557781e-01 9.47177708e-01 3.56070429e-01
-1.70716810e+00 -4.49693918e-01 -8.04416478e-01 -5.67159653e-01
-6.54947460e-01 -1.23019673e-01 4.33325738e-01 1.36540174e-01
-1.06096280e+00 2.97804654e-01 -4.85900134e-01 -2.22325534e-01
3.44661355e-01 -2.81477924e-02 -3.81779134e-01 -2.73690969e-01
-9.77596641e-01 1.36187100e+00 5.69729567e-01 3.68081927e-02
-1.07439172e+00 -7.81505883e-01 -1.18599093e+00 4.35780495e-01
7.31693268e-01 -8.35221946e-01 1.49839139e+00 -1.12649214e+00
-1.25389659e+00 1.50069714e+00 -2.46015087e-01 -4.54745710e-01
8.02674174e-01 1.74570307e-01 -1.75063446e-01 4.81479764e-01
-8.88719857e-02 6.43781960e-01 7.07570970e-01 -1.49295616e+00
-3.55914533e-01 -3.69585186e-01 9.71057415e-01 1.33668214e-01
7.50439167e-01 4.53294441e-02 -5.60088515e-01 -3.65577489e-01
1.36952594e-01 -3.97383511e-01 -9.47992355e-02 2.45000124e-01
-1.78177714e-01 -4.60570976e-02 5.49902141e-01 -5.56041896e-01
9.97836709e-01 -2.00113249e+00 4.88941073e-01 1.22732744e-01
1.89106986e-01 1.04540356e-01 -3.08337331e-01 6.73668325e-01
-6.13824368e-01 1.87483743e-01 -5.03400108e-03 -1.91580690e-02
3.10592443e-01 3.12670857e-01 -6.48411036e-01 4.76331681e-01
2.03881338e-01 1.34472668e+00 -7.68625379e-01 -6.53585732e-01
5.51940322e-01 1.25130698e-01 -2.83398449e-01 1.70914248e-01
-1.03611267e+00 1.22530125e-01 -2.24054232e-01 3.62366527e-01
5.36825776e-01 -3.76889706e-01 1.84568033e-01 -3.34919006e-01
-1.06662083e-02 2.93522209e-01 -1.49780810e+00 1.84796691e+00
-3.43773305e-01 7.34963357e-01 -1.30967706e-01 -1.04631877e+00
2.84202337e-01 3.15691501e-01 -3.47552598e-01 -9.45598185e-01
1.56787008e-01 -5.17180979e-01 -5.05734980e-02 -9.76011693e-01
2.02492192e-01 -4.31373566e-01 -4.97359456e-03 4.89056855e-01
9.56506133e-02 -8.89169097e-01 6.41089916e-01 6.33496046e-01
8.94904852e-01 4.16014463e-01 5.31835318e-01 8.00185185e-03
7.91773677e-01 4.98949289e-01 -1.26020938e-01 9.33333814e-01
6.94454461e-02 4.07533735e-01 1.18566656e+00 -7.09234893e-01
-8.05657744e-01 -1.22525632e+00 3.27731431e-01 1.03251433e+00
2.96555459e-01 -5.77874780e-01 -4.53558445e-01 -4.36349005e-01
-9.13347602e-02 1.36741626e+00 -7.75638819e-01 -1.67547800e-02
-7.67062962e-01 1.64115295e-01 5.52476704e-01 2.85578936e-01
4.61060345e-01 -1.24435747e+00 -1.06502461e+00 -9.45165902e-02
-1.71122119e-01 -1.33347869e+00 1.64095491e-01 -2.05456123e-01
-7.64988542e-01 -1.35089517e+00 -1.33901797e-02 -7.64128864e-01
7.64538109e-01 -1.19148687e-01 1.57867479e+00 3.07502717e-01
-5.18307686e-01 9.92543340e-01 -5.16301915e-02 -2.35833272e-01
-4.51950103e-01 -7.36628294e-01 -6.93478107e-01 -2.69554913e-01
-8.80077779e-02 -2.59207547e-01 -1.42944545e-01 -1.76123098e-01
-1.05173075e+00 5.11047900e-01 4.64295112e-02 3.16134810e-01
4.13118511e-01 1.74635828e-01 -2.58982658e-01 -8.61466765e-01
6.39798820e-01 -1.93167433e-01 -7.95406520e-01 7.76659191e-01
3.10504615e-01 2.97447324e-01 5.30725956e-01 -4.02429432e-01
-1.19868302e+00 -5.13928197e-02 3.58356148e-01 -5.11565149e-01
-6.72459543e-01 6.68687046e-01 1.08895078e-01 2.96923220e-01
6.96008444e-01 2.41280422e-01 -3.74353349e-01 -2.12893002e-02
9.59046602e-01 -2.81228900e-01 8.10545087e-01 -1.00335217e+00
8.91575694e-01 6.41906977e-01 4.55976784e-01 -7.57280231e-01
-7.50843644e-01 -7.57446587e-02 -7.16711283e-01 -4.11046684e-01
1.12926483e+00 -7.09192336e-01 -1.30779684e+00 1.37436584e-01
-1.57580101e+00 -6.01242244e-01 -4.29228067e-01 -2.46903092e-01
-9.15561557e-01 2.33679309e-01 -3.80004764e-01 -9.94939446e-01
3.21775913e-01 -1.19582820e+00 1.02615631e+00 -2.35358975e-03
-4.12915558e-01 -1.10095274e+00 -1.41748995e-01 6.98781908e-01
8.35024938e-02 4.30203676e-01 1.76379812e+00 -4.47633833e-01
-8.86942029e-01 8.30958113e-02 -3.84131312e-01 -2.80803263e-01
-2.40660757e-01 -6.51684031e-02 -6.64206088e-01 4.72069867e-02
-1.66532576e-01 -6.60725892e-01 4.62728739e-01 1.26632974e-01
1.39120471e+00 -2.39729285e-01 -2.63363793e-02 1.79510996e-01
1.57337391e+00 9.15223360e-02 9.18843687e-01 2.20089093e-01
5.20184278e-01 6.12379313e-01 3.88581216e-01 -6.62776679e-02
5.56282461e-01 5.50692439e-01 9.11045909e-01 -4.49979976e-02
-1.15030751e-01 -6.86856732e-02 8.91056359e-02 -2.42788672e-01
-3.35812643e-02 -2.80967891e-01 -1.36032772e+00 6.19362593e-01
-1.94146895e+00 -1.32107711e+00 -4.36601102e-01 1.79804671e+00
8.00261676e-01 -2.59545222e-02 -1.35164440e-01 -1.02968417e-01
2.12386787e-01 4.50357646e-01 -4.16232020e-01 -8.35665762e-01
2.04064816e-01 5.44089079e-01 -1.46965176e-01 7.89918303e-01
-9.24225092e-01 1.20376348e+00 6.40061808e+00 7.73430765e-01
-7.30743587e-01 -2.13178948e-01 2.84489721e-01 -2.80789524e-01
-1.20953597e-01 3.25561389e-02 -8.85359496e-02 -3.32035333e-01
6.30988240e-01 4.55455631e-02 8.05604935e-01 3.89177829e-01
-7.97523186e-03 -7.06450462e-01 -1.75334251e+00 1.00021911e+00
4.50901628e-01 -1.64916956e+00 4.56530213e-01 -5.24431109e-01
2.57096678e-01 -7.56294310e-01 7.30171725e-02 3.12066972e-01
4.07836914e-01 -1.56244159e+00 1.09555054e+00 7.46732295e-01
7.30236650e-01 -5.00478446e-01 6.61754981e-02 6.13280416e-01
-1.15082169e+00 -1.06488839e-02 1.52627438e-01 -4.75944132e-01
1.36831388e-01 -2.41809607e-01 -5.92860222e-01 5.47489762e-01
5.27407944e-01 2.64075637e-01 -6.91958070e-01 7.25785077e-01
-6.30230665e-01 9.81114060e-02 7.73367584e-02 -8.50265566e-03
4.04961973e-01 -3.63518655e-01 5.10067165e-01 1.13389194e+00
-1.77333251e-01 6.21700644e-01 -1.13315195e-01 1.49370110e+00
3.73619258e-01 -4.29473191e-01 -5.56783795e-01 -2.59875238e-01
-1.93304583e-01 9.78233993e-01 -6.71454489e-01 -5.27001977e-01
-4.15458709e-01 1.04796076e+00 4.83844131e-01 6.90727174e-01
-1.09087360e+00 -1.26230925e-01 5.67254186e-01 2.34845072e-01
3.07874918e-01 -4.55410630e-01 -3.01742315e-01 -1.19889665e+00
-2.11732343e-01 -1.16481853e+00 5.80320954e-01 -1.87167788e+00
-8.08197379e-01 7.87910968e-02 4.16631877e-01 -5.63713491e-01
-6.06574900e-02 -1.14725995e+00 -7.51790702e-01 7.72116423e-01
-1.05211139e+00 -1.16624403e+00 -3.23402822e-01 6.76493347e-01
5.99087298e-01 1.84657171e-01 9.84907269e-01 -2.23823696e-01
-2.68300563e-01 -2.70830542e-01 -7.99766243e-01 1.44502953e-01
8.98670554e-02 -1.46768105e+00 3.54479849e-01 7.54741251e-01
4.30402249e-01 4.86986190e-01 1.03045774e+00 -3.78755718e-01
-1.60552180e+00 -6.16045058e-01 8.41091156e-01 -8.08700144e-01
9.55265522e-01 -4.28500205e-01 -9.68350530e-01 1.25473976e+00
6.50489390e-01 -2.00141564e-01 2.22267643e-01 -3.99433374e-01
-6.04206562e-01 4.71072882e-01 -1.09361649e+00 1.08582592e+00
1.00850344e+00 -7.00148165e-01 -1.49390721e+00 6.44309580e-01
5.43159425e-01 -1.02618647e+00 -3.84843290e-01 -9.48913470e-02
1.05689503e-01 -9.92371321e-01 1.51593125e+00 -1.51581430e+00
9.36844289e-01 -5.67410231e-01 -2.36361697e-01 -1.19226038e+00
-2.33571500e-01 5.37242144e-02 -8.20718184e-02 8.21824133e-01
1.90040976e-01 -3.20645541e-01 3.77150208e-01 7.23371625e-01
3.70012611e-01 -4.18784976e-01 -1.07442069e+00 -2.42255598e-01
1.07241549e-01 -7.10513413e-01 4.03425127e-01 9.53189373e-01
2.72843480e-01 3.68803650e-01 1.60388380e-01 3.66299957e-01
1.66205540e-01 5.00812829e-01 7.65347421e-01 -9.62405026e-01
-3.97907287e-01 -3.86183411e-01 -4.43938643e-01 -1.00363457e+00
4.54027325e-01 -9.06492233e-01 -2.03382805e-01 -2.16376996e+00
7.35113621e-02 3.37825119e-01 3.60650718e-01 7.41697133e-01
2.89675146e-01 -1.44882739e-01 6.54211640e-01 -2.63287723e-01
-9.04328942e-01 1.39634475e-01 1.44144595e+00 -4.07301486e-01
2.29857519e-01 -6.90625012e-01 -2.12019190e-01 8.66926789e-01
4.18002367e-01 -5.68817034e-02 -3.71530831e-01 -6.64367735e-01
9.58160579e-01 5.76526701e-01 9.63970482e-01 -7.08448350e-01
4.16223824e-01 -5.64062536e-01 5.03552079e-01 -3.52358103e-01
7.04351068e-01 -1.15279508e+00 1.63371786e-01 5.46382666e-01
-5.28802156e-01 4.80387449e-01 5.73677480e-01 2.70715863e-01
-1.29006729e-01 -3.31540644e-01 6.31294310e-01 -7.61264086e-01
-1.05438852e+00 -2.82033116e-01 -6.86463416e-01 2.18009561e-01
1.34784663e+00 -2.19392210e-01 -7.98755884e-01 -6.80400252e-01
-1.18686497e+00 5.32126188e-01 1.44142300e-01 2.65805691e-01
7.32384801e-01 -1.13941336e+00 -4.84737515e-01 -6.85961097e-02
1.17149539e-01 4.15325202e-02 4.21033859e-01 6.47195935e-01
-1.03326058e+00 4.41224098e-01 -2.03095004e-01 -3.65686893e-01
-1.55579746e+00 9.20791447e-01 8.82216096e-01 -2.74865389e-01
-5.25691688e-01 7.66076446e-01 3.69039506e-01 -4.52523023e-01
2.69296765e-01 -7.36992955e-01 -3.51879030e-01 5.04821464e-02
4.67404127e-01 1.98619395e-01 -1.75153390e-01 -7.79935360e-01
-4.63344514e-01 8.02149296e-01 3.95116508e-01 -3.21933627e-01
1.09957874e+00 7.13227540e-02 -5.18889666e-01 7.46466517e-01
6.18034422e-01 -4.04920340e-01 -5.46124399e-01 -1.92162678e-01
-3.17462832e-01 -4.50744122e-01 -1.04426108e-01 -1.19860196e+00
-6.30587220e-01 1.05006957e+00 6.62977248e-02 3.96779209e-01
8.41160119e-01 3.01521182e-01 5.54308295e-03 7.02974260e-01
1.19194441e-01 -8.83873522e-01 4.25528228e-01 8.02758336e-01
1.51431346e+00 -9.28719580e-01 2.93051302e-01 -3.61383736e-01
-7.73214042e-01 1.33239853e+00 7.53790736e-01 -1.60230875e-01
7.05112740e-02 3.37811583e-03 1.87640309e-01 -9.30554926e-01
-7.93788612e-01 -2.98175007e-01 4.28316206e-01 3.74414444e-01
7.56814256e-02 -1.39132589e-01 2.75178552e-01 -1.09752023e-03
-3.03284049e-01 -1.19483754e-01 6.05215013e-01 9.08152759e-01
-2.78093070e-01 -5.75418413e-01 -7.69308746e-01 3.11597455e-02
-7.09769651e-02 -8.69325250e-02 -7.97720253e-01 1.09059584e+00
2.27936521e-01 9.90072846e-01 4.50531870e-01 3.92837137e-01
7.16716170e-01 3.31880212e-01 1.13493741e+00 -6.76125407e-01
-8.24818730e-01 -5.59345007e-01 2.91072339e-01 -8.60365033e-01
-2.32635304e-01 -5.81349730e-01 -1.67411470e+00 -2.93919444e-01
1.75045148e-01 -3.31864476e-01 3.97263765e-01 1.29718518e+00
-1.70345679e-01 9.33789372e-01 -4.39280719e-01 -6.16180956e-01
-2.12391093e-01 -4.00168002e-01 -2.08802953e-01 6.07655525e-01
3.92232299e-01 -2.02995643e-01 -2.12723464e-01 3.51614296e-01] | [10.697480201721191, 2.0321624279022217] |
f9eeca90-0a61-420c-8fc1-c99d23a8bf73 | retrocomposer-discovering-novel-reactions-by | 2112.11225 | null | https://arxiv.org/abs/2112.11225v2 | https://arxiv.org/pdf/2112.11225v2.pdf | RetroComposer: Composing Templates for Template-Based Retrosynthesis Prediction | The main target of retrosynthesis is to recursively decompose desired molecules into available building blocks. Existing template-based retrosynthesis methods follow a template selection stereotype and suffer from limited training templates, which prevents them from discovering novel reactions. To overcome this limitation, we propose an innovative retrosynthesis prediction framework that can compose novel templates beyond training templates. As far as we know, this is the first method that uses machine learning to compose reaction templates for retrosynthesis prediction. Besides, we propose an effective reactant candidate scoring model that can capture atom-level transformations, which helps our method outperform previous methods on the USPTO-50K dataset. Experimental results show that our method can produce novel templates for 15 USPTO-50K test reactions that are not covered by training templates. We have released our source implementation. | ['Junzhou Huang', 'Yang Yu', 'Chan Lu', 'Peilin Zhao', 'Chaochao Yan'] | 2021-12-20 | null | null | null | null | ['retrosynthesis'] | ['medical'] | [ 5.31578660e-01 1.18203694e-02 -8.27180743e-01 5.05159162e-02
-5.92542827e-01 -1.10865855e+00 8.10581744e-01 1.24692678e-01
-1.60667837e-01 1.21718228e+00 2.85695225e-01 -4.98171806e-01
2.34280050e-01 -7.56987274e-01 -5.41528106e-01 -7.04826117e-01
4.53116417e-01 2.48029351e-01 6.65440500e-01 -3.44739288e-01
5.83508611e-01 6.54358923e-01 -1.25429821e+00 4.52340066e-01
8.32119584e-01 8.06187570e-01 3.88553798e-01 3.20895016e-01
-7.82450661e-02 4.56204295e-01 -4.95179057e-01 -5.00566721e-01
3.22175056e-01 -5.53141296e-01 -8.44971359e-01 -5.99221587e-01
1.06496155e-01 3.57558057e-02 -1.77018568e-01 5.82465589e-01
6.20717764e-01 2.35383749e-01 1.14573181e+00 -7.46932209e-01
-2.09966689e-01 8.10016692e-01 1.88574657e-01 -1.24377333e-01
5.58538675e-01 -1.58315264e-02 1.08673954e+00 -1.43820810e+00
8.65002394e-01 8.07265937e-01 4.15556788e-01 7.77131677e-01
-1.16003799e+00 -1.23675430e+00 2.09044099e-01 1.01612270e-01
-1.31862164e+00 -5.81790805e-01 7.28436470e-01 -2.67148137e-01
1.37084758e+00 3.79909903e-01 7.35229194e-01 1.36390436e+00
4.48988646e-01 5.25898755e-01 9.31578279e-01 -8.02365243e-02
4.34915036e-01 -2.46727675e-01 -4.36839908e-01 5.88037550e-01
1.24996021e-01 3.77970934e-01 -6.89585626e-01 -5.62669188e-02
4.38662410e-01 2.01008264e-02 -1.42156839e-01 -1.67792588e-01
-1.24894643e+00 7.95785427e-01 4.11331087e-01 3.35378259e-01
-4.34724018e-02 -2.03315660e-01 4.13301170e-01 -1.04208134e-01
1.22312307e-01 1.37515259e+00 -9.19199467e-01 9.89108831e-02
-9.93944228e-01 1.08908592e-02 6.95734143e-01 1.25744057e+00
8.63858163e-01 -7.87248742e-03 -8.94021988e-03 4.40404952e-01
-1.08496612e-02 -6.33767545e-02 5.26517630e-01 -5.05005836e-01
2.69537807e-01 6.50749087e-01 4.04853113e-02 -3.73379678e-01
-6.59974456e-01 -1.70506909e-01 -5.24450898e-01 -2.44547248e-01
2.56244093e-01 1.21565431e-01 -8.40031385e-01 1.35045063e+00
3.98019016e-01 -8.94482732e-02 1.41609952e-01 4.88481224e-01
1.07127285e+00 1.12706852e+00 2.24324927e-01 -9.98146534e-01
8.29306722e-01 -1.35138547e+00 -5.42874753e-01 3.97350162e-01
6.68132424e-01 -1.10418057e+00 6.84643328e-01 9.69473481e-01
-1.14225543e+00 -3.51638734e-01 -1.26003444e+00 -5.88779636e-02
-8.66069078e-01 -5.51491901e-02 1.09686828e+00 8.28110278e-01
-4.30541515e-01 9.84641016e-01 -4.12190199e-01 -1.56128898e-01
7.61371925e-02 6.14552140e-01 -4.18754667e-01 2.50488788e-01
-1.07313287e+00 1.00511718e+00 1.06028855e+00 -1.40743598e-01
-1.32117617e+00 -9.89443719e-01 -5.22539854e-01 -8.35502073e-02
1.15971887e+00 -5.11933148e-01 1.32657504e+00 -4.68996495e-01
-1.98811960e+00 1.03852116e-01 -1.02507368e-01 2.89603751e-02
1.50296271e-01 4.58039463e-01 -5.55456221e-01 -5.94970845e-02
-6.99968338e-02 7.36334741e-01 6.94477856e-01 -1.10188186e+00
-5.28477907e-01 1.18771940e-01 -5.56610711e-02 -8.39454755e-02
-1.68446243e-01 8.85533728e-03 -2.38382250e-01 -8.91680121e-01
1.30818456e-01 -8.49097848e-01 -5.67880273e-01 -5.80603719e-01
-6.59312189e-01 -4.07606155e-01 2.70102650e-01 6.63847523e-03
1.40330267e+00 -1.35255909e+00 2.98799962e-01 3.55078340e-01
6.63450174e-03 3.84363085e-01 -2.01164767e-01 9.53592956e-01
-5.54867506e-01 4.97557759e-01 6.88901320e-02 3.00899595e-01
-1.42872706e-01 -1.80985630e-01 -3.93667847e-01 2.00961586e-02
8.30698609e-02 7.26642072e-01 -1.12013113e+00 -3.72353822e-01
2.27854982e-01 6.10237606e-02 -5.91737330e-01 -2.59673651e-02
-7.63433278e-01 4.25908506e-01 -6.14010036e-01 1.00589132e+00
5.26406765e-01 4.06825207e-02 5.41266382e-01 -5.48539162e-01
-7.27906048e-01 7.47975647e-01 -9.29948390e-01 1.91484106e+00
-2.28439525e-01 -3.42335135e-01 -8.62228215e-01 -6.29825771e-01
1.04320216e+00 5.24906099e-01 4.90822673e-01 -5.75417697e-01
-5.04453667e-03 5.13356268e-01 -1.97701007e-02 1.00823283e-01
5.87071061e-01 -3.47305834e-01 -9.42106694e-02 1.81366488e-01
2.76318848e-01 -3.28025073e-01 7.41831481e-01 1.49611905e-01
9.15826976e-01 4.33709025e-01 8.95726264e-01 -1.41809061e-01
7.68394768e-01 1.02076635e-01 7.89632559e-01 6.93854332e-01
2.21473694e-01 3.67308944e-01 2.65508652e-01 -7.60842383e-01
-1.24342847e+00 -9.19197261e-01 1.36401271e-02 1.17270577e+00
-1.92208633e-01 -1.19138813e+00 -3.87903273e-01 -1.02497613e+00
-5.79950511e-01 5.10737002e-01 -3.91034126e-01 -1.65308028e-01
-4.39348191e-01 -5.38964093e-01 4.94429260e-01 4.68454540e-01
-3.47640701e-02 -9.51827347e-01 3.95406365e-01 6.59590662e-01
7.16958046e-02 -6.99705541e-01 -9.38217640e-01 5.22132277e-01
-8.21140647e-01 -1.46192348e+00 -3.89819890e-01 -7.06364572e-01
4.83862787e-01 3.99474829e-01 6.53533399e-01 -1.86065748e-01
-3.40385050e-01 -2.89547235e-01 -3.07723463e-01 -5.72288513e-01
-7.00409174e-01 5.43659270e-01 5.84879629e-02 -6.83238745e-01
-1.26134038e-01 -6.51295483e-01 -6.39874518e-01 5.71348190e-01
-7.67796814e-01 2.40268677e-01 7.19114482e-01 7.57855952e-01
8.74972761e-01 -2.57014520e-02 7.95780540e-01 -6.66017413e-01
1.56484798e-01 -3.26376677e-01 -9.47861075e-01 7.03666329e-01
-1.04375088e+00 3.35737646e-01 1.10216510e+00 -6.25262976e-01
-1.18464553e+00 6.70832515e-01 -8.69012177e-02 1.74921751e-01
5.61738349e-02 5.35087705e-01 -5.88490367e-01 -1.20384149e-01
5.80734313e-01 2.95490831e-01 -5.79140246e-01 -4.24560457e-01
4.56033349e-01 5.06638885e-02 1.61994115e-01 -8.76797915e-01
7.15332508e-01 2.41419766e-02 6.51677549e-01 -5.69551826e-01
-7.66035020e-01 -5.97035706e-01 -5.25677860e-01 6.08824864e-02
6.16516948e-01 -7.17809916e-01 -1.12651253e+00 -1.77184418e-01
-1.00320351e+00 -1.78800762e-01 2.46267915e-01 4.35843229e-01
-6.20227158e-01 5.24008274e-01 -3.95475030e-01 -4.32110906e-01
-5.06452501e-01 -1.45785034e+00 8.50979865e-01 -5.02878129e-02
-2.66815126e-01 -5.58874071e-01 2.70278305e-01 2.53464222e-01
-7.47263283e-02 9.82303824e-03 9.83478129e-01 -7.40161419e-01
-8.89082432e-01 1.18840970e-01 1.90153003e-01 -2.22519368e-01
5.36839783e-01 1.69620529e-01 -6.64089024e-01 -5.90774380e-02
-4.54057395e-01 -3.05999815e-01 1.10391855e+00 -1.09337650e-01
1.34378123e+00 -3.83312702e-01 -6.18215024e-01 5.02204478e-01
1.20541704e+00 7.08321571e-01 6.90163672e-01 2.83160776e-01
6.54557467e-01 3.83091480e-01 9.43591654e-01 4.89727080e-01
-5.62574789e-02 7.69898236e-01 4.64335740e-01 1.28166482e-01
1.68558612e-01 -7.08389461e-01 5.04238725e-01 6.16787791e-01
-6.01641238e-01 -3.65289301e-01 -5.54625630e-01 -7.61046335e-02
-1.63147044e+00 -1.05644357e+00 -1.13920592e-01 2.11398530e+00
1.28618109e+00 8.20733979e-02 4.32082713e-01 3.21051478e-01
2.42189810e-01 -2.96473056e-02 -3.73890817e-01 -3.90178263e-01
-1.18775517e-02 1.03755069e+00 5.86641550e-01 3.70647728e-01
-1.06946898e+00 1.60165894e+00 7.20476389e+00 1.24600875e+00
-1.18547440e+00 -2.47735813e-01 1.79783940e-01 -1.71578959e-01
-3.64997357e-01 4.62181658e-01 -1.11788225e+00 2.80859202e-01
7.96705842e-01 -3.15421581e-01 2.77748019e-01 7.32795060e-01
4.08867717e-01 2.56381482e-01 -1.40448582e+00 6.36676788e-01
-2.33192861e-01 -2.07913423e+00 4.03952092e-01 3.24212611e-02
9.12055314e-01 -6.10047817e-01 -1.82498127e-01 1.75414294e-01
1.89058796e-01 -1.00828218e+00 6.95943356e-01 2.13119835e-01
6.06006205e-01 -9.20196891e-01 6.81056976e-02 -4.27727923e-02
-1.39776230e+00 2.61514448e-02 -3.94930542e-01 1.85059428e-01
-1.52229995e-01 3.37806225e-01 -1.47172022e+00 1.04840899e+00
-7.97504932e-02 8.95498395e-01 -3.95659775e-01 9.59576368e-01
-5.53401887e-01 4.95822102e-01 -2.82785445e-02 -3.88990104e-01
2.21043006e-01 -3.35033357e-01 1.53537020e-01 1.05879402e+00
5.00450790e-01 2.07276478e-01 4.86062288e-01 7.04278886e-01
-2.76021242e-01 6.26630425e-01 -5.78244925e-01 -2.98768938e-01
5.18776119e-01 1.10473299e+00 -1.07461119e+00 -3.86112034e-01
-3.37297469e-01 7.20882177e-01 -5.34126535e-02 2.63016313e-01
-8.41794312e-01 1.03916610e-02 2.89293021e-01 -1.78623080e-01
4.21046883e-01 -2.01475710e-01 6.56669438e-02 -1.05907619e+00
-3.97310555e-01 -1.02513576e+00 4.65811849e-01 -5.07348895e-01
-9.62311327e-01 9.58572328e-02 -1.13930948e-01 -9.83479083e-01
3.44522655e-01 -7.81833291e-01 -5.33038318e-01 2.15706080e-01
-1.16233444e+00 -1.23350322e+00 2.76685536e-01 3.32984895e-01
9.05786097e-01 -2.43464336e-01 9.72312272e-01 1.58025607e-01
-7.82354295e-01 5.28631628e-01 1.39213964e-01 -6.95536077e-01
1.23739791e+00 -1.16527855e+00 1.05365776e-01 5.91960371e-01
-2.55717989e-02 9.17170942e-01 5.81239522e-01 -8.47763896e-01
-1.53033137e+00 -1.17422426e+00 8.34042370e-01 -4.70615298e-01
6.14307761e-01 -6.46358728e-01 -3.08453113e-01 2.77873904e-01
-8.10712799e-02 -5.77382207e-01 1.20553994e+00 -5.17181412e-04
-7.16144323e-01 -1.39458969e-01 -7.47294068e-01 1.00933766e+00
1.42326307e+00 -1.21598512e-01 -3.46318483e-01 6.17778897e-01
8.43039513e-01 -2.85958678e-01 -1.29039192e+00 3.82917017e-01
7.30308652e-01 -5.90955079e-01 1.16488624e+00 -7.77526855e-01
5.35669982e-01 -4.36651438e-01 5.40481098e-02 -1.11516953e+00
-3.37871134e-01 -1.10663843e+00 4.05216366e-02 1.04118872e+00
1.06480837e+00 -4.14610445e-01 7.85519361e-01 3.17397445e-01
-6.80616558e-01 -6.37300491e-01 -6.33023322e-01 -1.05340648e+00
1.55740127e-01 -1.96970493e-01 1.06941152e+00 9.25788760e-01
3.99201065e-01 5.93627572e-01 -4.76056874e-01 -2.30852813e-01
1.12593167e-01 1.63783133e-01 6.82503223e-01 -1.07170713e+00
-1.30438879e-01 -6.54873252e-01 1.12936303e-01 -9.14857864e-01
2.00060338e-01 -1.22355270e+00 -1.02481879e-01 -1.27155566e+00
4.36528206e-01 -2.85489172e-01 -4.25686687e-01 6.25409007e-01
-1.10592715e-01 -1.55864086e-03 -4.73542213e-02 3.40233594e-02
-7.73297548e-01 6.71064675e-01 1.46195257e+00 -3.80874574e-01
-3.74709904e-01 -1.13189764e-01 -8.70994866e-01 3.47242177e-01
8.54597867e-01 -5.80240488e-01 -5.36345840e-01 5.64758301e-01
6.10660315e-01 -3.47280018e-02 -4.10985112e-01 -6.42387271e-01
1.96995646e-01 -8.15413415e-01 3.93944055e-01 -1.08124852e+00
2.68354714e-01 -5.58053076e-01 5.56313515e-01 5.40974557e-01
-2.65011787e-01 -3.51468414e-01 5.47771566e-02 4.28228557e-01
1.99874744e-01 -2.76281714e-01 2.99579948e-01 -3.28792393e-01
-3.54065567e-01 8.10492039e-01 -9.26908314e-01 -5.11371255e-01
9.05731201e-01 -3.06089371e-01 -3.30531925e-01 1.86535478e-01
-5.38805246e-01 -8.95751566e-02 5.75698555e-01 3.30581754e-01
6.04999900e-01 -1.27401543e+00 -4.05802540e-02 -4.22895074e-01
3.75020593e-01 -2.09659282e-02 -2.57224321e-01 6.30894601e-01
-3.17347527e-01 9.64910507e-01 8.44518542e-02 7.51228333e-02
-1.27297187e+00 1.37799978e+00 1.71459317e-01 -3.57877225e-01
1.60866842e-01 5.74285388e-01 4.09451157e-01 -5.04792511e-01
3.79359387e-02 -5.93922436e-01 -7.08390623e-02 7.62081593e-02
4.77845550e-01 3.00966620e-01 4.00751233e-01 -8.12332854e-02
-4.60189074e-01 3.02923679e-01 -3.66479397e-01 3.90924573e-01
1.24503827e+00 5.18735170e-01 -3.16577032e-02 -2.18904633e-02
9.11809206e-01 1.50512040e-01 -7.48171985e-01 7.37973377e-02
7.27276057e-02 -2.09412768e-01 -2.65119761e-01 -9.06483054e-01
-3.72458220e-01 4.91195828e-01 8.83929133e-02 -3.61122191e-01
9.91438210e-01 -3.33716646e-02 7.05198526e-01 9.86355841e-01
2.48978078e-01 -1.18674982e+00 5.22134542e-01 4.78306621e-01
9.45813119e-01 -1.04551625e+00 3.32781225e-01 -8.47054541e-01
-1.10702842e-01 1.46883535e+00 6.07861698e-01 4.53182071e-01
1.49839103e-01 -2.05913752e-01 -6.94164395e-01 7.47824693e-03
-9.45567191e-01 -1.36297271e-01 2.03219995e-01 3.93627077e-01
6.75577044e-01 -8.22263062e-02 -9.25596833e-01 7.09356964e-01
-2.18634516e-01 -1.80673912e-01 4.22691286e-01 9.42518651e-01
-5.03429651e-01 -2.10838532e+00 -1.97912619e-01 -4.64281067e-02
-4.66853589e-01 -5.19755483e-01 -1.24313557e+00 7.29342401e-01
3.04699630e-01 9.60708976e-01 -8.67911816e-01 -3.34714919e-01
3.24612319e-01 2.39562705e-01 1.07788444e+00 -7.24250615e-01
-7.44271338e-01 4.61929500e-01 4.92989808e-01 -4.86475617e-01
-4.87865716e-01 -3.08613539e-01 -1.18605638e+00 -1.74122721e-01
-8.27924013e-01 4.86863494e-01 5.19880116e-01 9.02400076e-01
7.40162432e-02 3.45160723e-01 1.07507718e+00 -7.83561409e-01
-1.43635392e-01 -6.27968192e-01 -2.19342373e-02 -2.15125352e-01
-2.32444525e-01 -7.55145788e-01 9.90518630e-02 1.49688190e-02] | [4.489120960235596, 6.1108222007751465] |
650b24de-2423-4e75-b17a-413490753df8 | diversification-quotients-based-on-var-and-es | 2301.03517 | null | https://arxiv.org/abs/2301.03517v3 | https://arxiv.org/pdf/2301.03517v3.pdf | Diversification quotients based on VaR and ES | The diversification quotient (DQ) is recently introduced for quantifying the degree of diversification of a stochastic portfolio model. It has an axiomatic foundation and can be defined through a parametric class of risk measures. Since the Value-at-Risk (VaR) and the Expected Shortfall (ES) are the most prominent risk measures widely used in both banking and insurance, we investigate DQ constructed from VaR and ES in this paper. In particular, for the popular models of elliptical and multivariate regular varying (MRV) distributions, explicit formulas are available. The portfolio optimization problems for the elliptical and MRV models are also studied. Our results further reveal favourable features of DQ, both theoretically and practically, compared to traditional diversification indices based on a single risk measure. | ['Ruodu Wang', 'Liyuan Lin', 'Xia Han'] | 2023-01-09 | null | null | null | null | ['portfolio-optimization'] | ['time-series'] | [-2.96591967e-01 -3.62307690e-02 -2.34886378e-01 -2.42264971e-01
-2.61651218e-01 -7.37128973e-01 5.61422944e-01 8.87433439e-02
-2.63124764e-01 9.10679400e-01 1.16095200e-01 -5.33095479e-01
-7.46311903e-01 -1.24670422e+00 3.64364311e-02 -9.15957630e-01
-1.31523266e-01 3.00536752e-01 5.04025668e-02 -4.25433725e-01
4.51138139e-01 8.20995510e-01 -1.56323004e+00 -4.96673882e-01
1.10302961e+00 1.36129844e+00 -1.29337981e-01 2.14608490e-01
-9.63861048e-02 6.12039626e-01 -6.69914067e-01 -1.14729965e+00
4.57781166e-01 -4.17973965e-01 -4.05674428e-01 -2.78782725e-01
-3.91523510e-01 -1.56684201e-02 9.87359360e-02 1.20497215e+00
3.06146890e-01 8.26492980e-02 1.06514692e+00 -1.09634686e+00
-5.46158135e-01 6.99855447e-01 -4.71709490e-01 6.34615362e-01
1.42973274e-01 -2.39513546e-01 1.22122371e+00 -6.27442241e-01
2.17857823e-01 9.64924753e-01 6.56119585e-01 2.89011374e-02
-1.06771743e+00 -3.39732170e-01 -8.75556618e-02 -3.52970436e-02
-1.04507816e+00 2.03900948e-01 6.54560328e-01 -7.13722765e-01
4.44119304e-01 5.27150750e-01 5.60533226e-01 3.52788121e-01
7.92305708e-01 2.25961357e-01 1.18291843e+00 -4.20947760e-01
2.18933403e-01 3.35121304e-01 5.70310578e-02 7.31824338e-02
7.12488830e-01 5.04298985e-01 -6.41428530e-02 -3.18473548e-01
8.02777946e-01 1.49268061e-01 -3.01487267e-01 -5.21810889e-01
-7.54743874e-01 1.12379968e+00 -1.46712258e-01 3.49109083e-01
-4.12307858e-01 -3.30194086e-01 2.01025233e-01 5.36328197e-01
4.59847569e-01 2.96617448e-01 -2.07383066e-01 -1.88066959e-01
-7.63014853e-01 3.94652426e-01 9.86580610e-01 7.06270754e-01
4.39944267e-01 2.74727672e-01 -2.59779364e-01 5.67966223e-01
3.54451150e-01 5.69786668e-01 3.74616891e-01 -9.74614859e-01
5.80626428e-01 4.81150150e-01 2.06833482e-01 -1.02245152e+00
-3.70706886e-01 -7.59632468e-01 -8.30078721e-01 5.47805190e-01
6.64065003e-01 -4.54525501e-02 2.48949483e-01 1.61303306e+00
1.49525508e-01 -3.30278516e-01 2.57778823e-01 2.75202245e-01
1.62748545e-01 2.77671188e-01 -3.35767716e-01 -7.51094043e-01
9.64637637e-01 -5.15202761e-01 -6.96104527e-01 4.60011095e-01
3.27332407e-01 -6.67284966e-01 8.65614653e-01 6.39832437e-01
-1.35104573e+00 -1.81266218e-01 -8.36384773e-01 6.18922710e-01
-5.45302965e-02 -4.73209947e-01 3.96160990e-01 1.37126565e+00
-1.05606365e+00 9.56236780e-01 -3.23321849e-01 1.75421193e-01
1.57553449e-01 -6.36783317e-02 2.07755357e-01 4.27112013e-01
-1.06783366e+00 9.69781697e-01 1.40942559e-01 -7.91343376e-02
-5.39373040e-01 -6.70757055e-01 -3.72594208e-01 2.35406950e-01
2.15598345e-01 -4.55706388e-01 9.93497133e-01 -7.56999612e-01
-1.62092304e+00 6.34113848e-01 5.92749238e-01 -5.27564585e-01
7.44654834e-01 2.24938244e-02 -4.15116817e-01 1.62047520e-01
-1.81229282e-02 -6.44426405e-01 6.40929639e-01 -5.88791132e-01
-5.54702938e-01 -3.08657587e-01 3.16025853e-01 -3.49561237e-02
-3.77838910e-01 4.42887664e-01 5.80310285e-01 -1.27385533e+00
-1.21617123e-01 -4.71895546e-01 -1.84293166e-01 -2.76461750e-01
-5.39023764e-02 -1.15242921e-01 -7.74885938e-02 -7.62258410e-01
1.65480733e+00 -2.05889058e+00 2.57080998e-02 4.82109874e-01
-1.56371281e-01 -9.00585502e-02 5.09512424e-01 6.41656816e-01
-1.85970664e-01 6.38417527e-02 -5.79072118e-01 2.93699086e-01
3.25448215e-01 -1.46604981e-02 -5.49153566e-01 4.99816477e-01
1.15494803e-03 6.42018199e-01 -6.25202894e-01 -2.23476619e-01
-9.90703553e-02 3.43138166e-02 -4.08388108e-01 1.68448180e-01
1.72255561e-01 1.02092810e-01 -3.00344735e-01 6.51907980e-01
8.50088716e-01 -2.29406916e-03 1.44411877e-01 3.16298664e-01
-3.47994596e-01 2.40542348e-02 -1.33650792e+00 6.55951858e-01
-3.66557926e-01 1.25243232e-01 -1.52405143e-01 -1.18020523e+00
1.24344909e+00 2.83369213e-01 4.31512713e-01 -3.89016509e-01
8.79953727e-02 4.41873878e-01 -1.27618358e-01 -1.52158560e-02
4.93953556e-01 -8.94587517e-01 -1.93981409e-01 6.89365864e-01
-1.36813387e-01 3.02919708e-02 2.32366890e-01 -1.60120577e-01
7.56449580e-01 -1.51598737e-01 7.87854195e-01 -8.43833804e-01
8.14563572e-01 -5.93540967e-01 6.71416700e-01 3.32086951e-01
-2.23145798e-01 4.37546939e-01 1.22387803e+00 -6.40368136e-03
-8.33331823e-01 -1.48061109e+00 -6.48634851e-01 6.00093484e-01
9.97158661e-02 6.32464588e-02 -6.25536323e-01 -5.77238798e-01
5.36492348e-01 8.87134612e-01 -5.03860712e-01 3.89676541e-02
-3.05452615e-01 -1.03307116e+00 2.67329097e-01 5.15926600e-01
2.58941829e-01 -9.68750000e-01 -6.26477242e-01 2.10172608e-01
3.39529127e-01 -4.25397784e-01 -4.32666212e-01 4.83277440e-02
-1.12285507e+00 -1.16920531e+00 -1.08867490e+00 -9.17222500e-02
1.11884765e-01 -5.73632121e-02 1.35938966e+00 -2.51948863e-01
-5.25836200e-02 4.86915022e-01 -3.76396239e-01 -3.77589762e-01
-4.37731683e-01 -5.18992484e-01 1.41567692e-01 3.31406087e-01
4.19596536e-03 -6.93652213e-01 -5.94142973e-01 6.87798202e-01
-9.42154765e-01 -8.37944686e-01 1.88450038e-01 7.09599197e-01
5.45772433e-01 3.87141943e-01 1.22701430e+00 -6.35551691e-01
8.18891168e-01 -6.42997026e-01 -9.45201397e-01 4.53430086e-01
-1.05714071e+00 6.01191781e-02 4.10574317e-01 -5.78285232e-02
-1.36556256e+00 -9.00169253e-01 3.50493453e-02 1.02076940e-02
3.61458957e-01 5.27812004e-01 -4.01283592e-01 -1.48812950e-01
2.24824116e-01 -3.42278741e-02 -4.37908918e-02 -6.47002757e-01
7.54948780e-02 4.78419483e-01 2.70840406e-01 -5.75123966e-01
5.68396628e-01 4.21947420e-01 2.98121154e-01 -5.92111588e-01
-6.87529087e-01 -1.64237469e-01 -4.36210930e-01 -2.07805336e-01
6.21415079e-01 -3.50861341e-01 -8.28583360e-01 5.91916442e-01
-6.55984879e-01 1.03841066e-01 -8.62307310e-01 5.94992101e-01
-8.27456355e-01 7.09564626e-01 -2.98278600e-01 -1.36908531e+00
-3.57352287e-01 -8.77878845e-01 2.56509483e-01 3.15445036e-01
1.97865874e-01 -1.57176936e+00 3.18532109e-01 -1.04423583e-01
5.35117209e-01 6.92874134e-01 1.00766301e+00 -6.65326774e-01
-3.25789183e-01 -1.49563268e-01 4.16156314e-02 7.67580748e-01
7.76510760e-02 -8.22710246e-02 -4.27784622e-01 -2.66771764e-01
7.54284978e-01 3.20648849e-01 6.69444978e-01 4.24029946e-01
7.66079128e-01 -3.18917185e-01 2.50962794e-01 7.10686922e-01
1.68406296e+00 5.41785896e-01 5.63442290e-01 7.50891209e-01
-1.45459548e-01 9.62711096e-01 8.99837017e-01 1.00283241e+00
8.76783505e-02 5.16941965e-01 4.39799100e-01 7.00034976e-01
6.12925768e-01 9.46653262e-02 4.70211953e-01 7.73094773e-01
-2.66973734e-01 -7.57921264e-02 -7.46960878e-01 3.22043240e-01
-1.51822722e+00 -1.28398120e+00 -3.69658545e-02 2.82601571e+00
6.45416081e-01 4.09733623e-01 4.95885223e-01 2.64364064e-01
7.87511349e-01 2.04437122e-01 -3.81380856e-01 -6.70386970e-01
-6.93236113e-01 4.15304959e-01 6.35812998e-01 4.55557138e-01
-7.30639875e-01 8.35209042e-02 7.32514381e+00 9.47937548e-01
-6.38971567e-01 1.26624079e-02 6.85630262e-01 -4.45894524e-02
-9.10158873e-01 7.63736218e-02 -9.88057137e-01 8.31670046e-01
7.72229433e-01 -9.39654768e-01 -2.55332422e-02 8.57661545e-01
-9.38388705e-02 -8.71724188e-02 -6.41668200e-01 7.56589174e-01
-2.13104993e-01 -1.03576767e+00 -3.34652439e-02 3.86513710e-01
7.47112215e-01 -5.54381192e-01 4.14806873e-01 9.57399048e-03
1.13695078e-01 -9.26292837e-01 8.33399296e-01 1.16479695e+00
5.26252270e-01 -1.24919856e+00 1.03272235e+00 2.52504498e-01
-1.04274571e+00 -3.60979199e-01 -4.56001073e-01 1.03438288e-01
6.00311160e-01 9.72428560e-01 1.24057576e-01 8.49343538e-01
6.37462497e-01 5.46765149e-01 -3.38355452e-01 1.10852158e+00
-5.18277586e-02 4.38362986e-01 -1.58434391e-01 -5.81234358e-02
2.32802164e-02 -1.10164547e+00 7.94533610e-01 7.73956895e-01
8.23106289e-01 -2.21353155e-02 -6.27763033e-01 9.04836655e-01
3.85719627e-01 2.79834032e-01 -3.76841933e-01 -3.46275866e-02
2.71271974e-01 9.20105100e-01 -5.15931010e-01 9.06031132e-02
-5.20637214e-01 3.44866067e-01 -2.03883469e-01 8.33413601e-02
-6.65789306e-01 -5.68767369e-01 7.72177935e-01 2.48844504e-01
5.43903172e-01 -7.36313984e-02 -5.01195908e-01 -1.11104989e+00
2.93787181e-01 -6.02783203e-01 6.31791472e-01 1.66147444e-02
-1.56671870e+00 5.88655591e-01 2.56734371e-01 -1.47732294e+00
-3.20149571e-01 -7.59398222e-01 -9.74435925e-01 1.08475959e+00
-1.59251595e+00 -2.70421892e-01 1.06606126e-01 4.00657892e-01
-1.59881622e-01 -8.06326747e-01 5.47902524e-01 9.43636969e-02
-5.69739759e-01 7.99738646e-01 9.40748930e-01 -5.49036145e-01
1.24509811e-01 -1.61530519e+00 2.38761082e-01 6.70022070e-01
-2.34227851e-01 4.67445821e-01 7.58809388e-01 -5.79465032e-01
-8.15795541e-01 -4.37254846e-01 8.36394370e-01 -3.49743813e-01
1.03148234e+00 3.03488314e-01 -8.34820271e-01 2.02905044e-01
-4.88620810e-02 -4.50113535e-01 9.55638349e-01 -2.26537734e-01
-1.12213515e-01 -3.20886672e-01 -1.34625161e+00 2.93804199e-01
8.56636465e-01 -1.59919441e-01 -6.95862114e-01 4.75646257e-02
5.50890326e-01 3.82083446e-01 -1.41728461e+00 3.87056112e-01
6.37741625e-01 -1.78047872e+00 1.10815954e+00 -2.23911494e-01
1.23073988e-01 2.46056363e-01 -4.23028320e-01 -1.01032460e+00
-2.46051937e-01 -9.13297296e-01 -1.14956692e-01 1.38711417e+00
2.75533795e-01 -1.32604873e+00 5.93008101e-01 2.53718674e-01
2.31251374e-01 -1.08389497e+00 -1.24599147e+00 -1.37204850e+00
5.67459941e-01 -1.94350719e-01 1.03964067e+00 6.02440298e-01
9.72158238e-02 -3.40479791e-01 -1.59860030e-01 -3.00309390e-01
1.04663050e+00 3.30745757e-01 2.04630926e-01 -1.52580321e+00
-6.02319300e-01 -9.97362375e-01 -3.84500474e-01 -5.66100299e-01
-5.48682101e-02 -8.88900340e-01 -7.28060305e-01 -1.04961956e+00
-9.85630229e-02 -4.96642977e-01 -6.10674083e-01 -5.69164217e-01
-1.10801049e-01 -2.37297669e-01 1.78886309e-01 2.88350075e-01
1.79019067e-02 8.07955325e-01 8.91988933e-01 3.66130888e-01
-2.84107849e-02 8.08844566e-01 -7.93046713e-01 9.15848494e-01
9.26770806e-01 -2.85929888e-01 -3.25115472e-01 3.49634707e-01
6.36983871e-01 4.54151094e-01 6.97646365e-02 -5.23134828e-01
-3.59616101e-01 -5.73013484e-01 -2.94506580e-01 -7.15313971e-01
-3.27060409e-02 -5.01923621e-01 4.60400641e-01 5.84043145e-01
-7.52181038e-02 4.38131064e-01 -3.43860418e-01 6.23550594e-01
-4.70288098e-01 -9.07390237e-01 9.50800180e-01 6.07399419e-02
-3.02781654e-03 2.38343865e-01 -1.42839283e-01 2.92458653e-01
1.32608962e+00 -3.91899705e-01 -1.62266046e-01 -4.08240110e-01
-5.98518252e-01 1.33463303e-02 6.08701050e-01 -5.84446415e-02
5.89167774e-01 -1.61517942e+00 -7.27885425e-01 6.43631537e-03
-8.49929452e-02 -5.02663195e-01 1.29693419e-01 9.94115889e-01
-8.01082790e-01 4.93949294e-01 -4.10916358e-01 3.56640480e-02
-7.45668948e-01 7.94475138e-01 6.17232084e-01 -7.28721082e-01
-4.37147915e-01 6.27677023e-01 5.28014839e-01 -7.07989857e-02
2.01110631e-01 -1.04336284e-01 -5.22453010e-01 4.68488157e-01
6.17832243e-01 1.06366801e+00 -2.74037361e-01 -6.13363326e-01
-2.53006488e-01 6.67798638e-01 3.79047990e-01 -4.29687858e-01
1.33743465e+00 -2.59667903e-01 -3.67780000e-01 5.06749034e-01
8.52033257e-01 2.92934597e-01 -1.11249352e+00 -1.40090346e-01
5.37330925e-01 -5.38953245e-01 -3.64042014e-01 -3.75035971e-01
-9.96612668e-01 8.01061809e-01 3.26855242e-01 8.60086977e-01
1.20388508e+00 -2.66755432e-01 5.92023849e-01 -1.23306401e-01
5.42391181e-01 -1.04527521e+00 -3.84931602e-02 2.99957365e-01
1.21954083e+00 -5.78585088e-01 -6.78823562e-03 -2.65545487e-01
-8.30712080e-01 1.08102155e+00 -7.83235133e-02 -3.74982238e-01
1.27757764e+00 2.76558131e-01 -2.39255503e-01 2.47117519e-01
-4.05525208e-01 -2.15189029e-02 3.09857011e-01 6.38387501e-01
2.44371593e-01 2.37439886e-01 -9.60266650e-01 1.18847024e+00
-4.99540925e-01 -4.44188982e-01 5.45902967e-01 7.17363656e-01
-6.54442549e-01 -1.24431098e+00 -3.23932201e-01 5.15492022e-01
-8.17437768e-01 -2.98075676e-02 7.59363174e-03 5.85112214e-01
-1.09577037e-01 7.03195810e-01 4.28725407e-02 -1.65365800e-01
5.10347366e-01 -2.33351469e-01 3.07570279e-01 -2.92153805e-01
-8.28432024e-01 -2.19549611e-01 -8.45800415e-02 -3.40497881e-01
-4.23957914e-01 -1.17031479e+00 -6.59017980e-01 -5.04529357e-01
-4.78708893e-01 3.76810193e-01 2.00034738e-01 7.17168391e-01
-2.29839936e-01 1.51125133e-01 1.19462872e+00 -2.88704783e-01
-1.48025107e+00 -6.90464795e-01 -1.58923459e+00 4.69815582e-02
4.66681346e-02 -1.00003791e+00 -9.00560319e-01 -4.57484037e-01] | [4.974542617797852, 3.9708774089813232] |
63de971d-eb72-462a-8c37-cae67eb59ac2 | autoregressive-3d-shape-generation-via | 2204.01955 | null | https://arxiv.org/abs/2204.01955v1 | https://arxiv.org/pdf/2204.01955v1.pdf | Autoregressive 3D Shape Generation via Canonical Mapping | With the capacity of modeling long-range dependencies in sequential data, transformers have shown remarkable performances in a variety of generative tasks such as image, audio, and text generation. Yet, taming them in generating less structured and voluminous data formats such as high-resolution point clouds have seldom been explored due to ambiguous sequentialization processes and infeasible computation burden. In this paper, we aim to further exploit the power of transformers and employ them for the task of 3D point cloud generation. The key idea is to decompose point clouds of one category into semantically aligned sequences of shape compositions, via a learned canonical space. These shape compositions can then be quantized and used to learn a context-rich composition codebook for point cloud generation. Experimental results on point cloud reconstruction and unconditional generation show that our model performs favorably against state-of-the-art approaches. Furthermore, our model can be easily extended to multi-modal shape completion as an application for conditional shape generation. | ['Ming-Hsuan Yang', 'Min Sun', 'Sifei Liu', 'Xueting Li', 'An-Chieh Cheng'] | 2022-04-05 | null | null | null | null | ['point-cloud-reconstruction', '3d-shape-generation', 'point-cloud-generation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 4.36219841e-01 1.51781872e-01 1.81505814e-01 -1.38108447e-01
-1.06504476e+00 -7.54765332e-01 1.00374687e+00 2.83520147e-02
2.31687531e-01 6.15892470e-01 1.04594752e-01 -3.04641932e-01
1.28448442e-01 -1.01135015e+00 -1.10905731e+00 -7.42927074e-01
1.35913059e-01 8.78193498e-01 -1.20453350e-01 -2.45784447e-01
2.29886994e-01 7.44013667e-01 -1.95363212e+00 3.74027580e-01
8.89751852e-01 7.61918843e-01 4.35734987e-01 4.24354941e-01
-6.45454407e-01 2.28197858e-01 -3.55387151e-01 -4.33256924e-01
3.09620440e-01 -8.55533928e-02 -7.17480242e-01 3.80026281e-01
3.92209232e-01 -2.85957038e-01 -2.28447821e-02 7.67866135e-01
4.26499665e-01 -1.16547041e-01 7.00691938e-01 -1.22575879e+00
-6.33969009e-01 4.85835135e-01 -3.30747336e-01 -6.40254915e-01
4.44776267e-01 -9.62817818e-02 9.70156848e-01 -1.21936858e+00
8.09385538e-01 1.42255211e+00 4.87440586e-01 6.50413692e-01
-1.61074817e+00 -7.15206563e-01 -9.86011699e-03 -2.15829805e-01
-1.41364181e+00 -3.81497025e-01 8.89903903e-01 -3.64788324e-01
8.08735967e-01 4.49575633e-01 8.43823314e-01 1.03274179e+00
-1.64174497e-01 9.38727140e-01 8.55888903e-01 -3.08459729e-01
1.89901158e-01 -1.76146597e-01 -7.27849603e-01 5.23755431e-01
5.57083488e-02 -1.78512141e-01 -7.22836971e-01 -5.78292906e-01
1.05264771e+00 5.08566983e-02 -3.21566202e-02 -5.65004349e-01
-1.53807247e+00 7.43740737e-01 4.39307272e-01 2.43970379e-01
-3.19902092e-01 3.82790595e-01 6.94446713e-02 1.98193461e-01
7.71317959e-01 3.62577349e-01 -2.44866773e-01 -6.51716962e-02
-9.38302815e-01 8.24640810e-01 6.62360072e-01 1.44773769e+00
7.74889827e-01 1.03223749e-01 -1.22289956e-01 7.25665450e-01
3.36407304e-01 9.07327890e-01 6.24204017e-02 -9.10077572e-01
6.34063423e-01 4.67207432e-01 2.33864542e-02 -7.21548676e-01
2.00275168e-01 -6.65147826e-02 -1.08932948e+00 2.75148600e-01
-1.06048826e-02 2.71862864e-01 -9.73404169e-01 1.39212787e+00
3.68088663e-01 3.61429185e-01 2.75294706e-02 7.02129900e-01
6.50062144e-01 8.83493066e-01 -5.90806790e-02 -1.41327428e-02
1.13470709e+00 -3.59604836e-01 -2.32899651e-01 1.58744574e-01
2.12282002e-01 -1.06446230e+00 9.66727436e-01 4.52846557e-01
-1.33058453e+00 -5.55975854e-01 -8.22054386e-01 -2.15871707e-01
3.07320263e-02 -5.91156073e-02 8.59767318e-01 4.09736961e-01
-1.11782205e+00 6.61612928e-01 -1.03452528e+00 -5.24602048e-02
6.06775582e-01 2.71721542e-01 -2.91744769e-01 -1.71416521e-01
-6.57360256e-01 2.60330588e-01 1.89863145e-01 -2.56235272e-01
-8.60923827e-01 -9.68035996e-01 -7.99013436e-01 -1.03912257e-01
3.67717929e-02 -1.42147923e+00 1.09582829e+00 -4.49268937e-01
-1.52365005e+00 8.00384998e-01 -3.89426231e-01 -2.73550749e-01
5.42305291e-01 -1.19761631e-01 1.22101262e-01 7.99043179e-02
2.26611823e-01 1.12055528e+00 1.20021558e+00 -1.65059352e+00
-5.39513290e-01 -2.95644432e-01 -1.41939133e-01 4.02155936e-01
-3.46148424e-02 -3.75524074e-01 -4.46442038e-01 -7.76845098e-01
4.72401112e-01 -1.12541831e+00 -3.35629046e-01 1.59556195e-01
-5.13046503e-01 -3.15066695e-01 9.56221282e-01 -3.61976475e-01
6.14591360e-01 -2.03721595e+00 4.34923649e-01 3.01212907e-01
-5.60598038e-02 -8.71855170e-02 -1.45119101e-01 6.15746498e-01
-9.42323804e-02 4.18656349e-01 -4.78278995e-01 -7.97449946e-01
1.66105077e-01 3.84382188e-01 -1.01881361e+00 3.47935371e-02
4.88320351e-01 1.07859993e+00 -8.94805610e-01 -5.10007262e-01
4.19628352e-01 6.58994138e-01 -7.86060154e-01 3.27276975e-01
-6.38203323e-01 8.87264550e-01 -5.73372543e-01 5.91004491e-01
6.71819925e-01 -2.88621783e-01 -1.55264750e-01 -1.44264892e-01
-2.74813958e-02 2.44681448e-01 -9.66365516e-01 2.29803181e+00
-6.55023158e-01 2.33014300e-01 -1.37316540e-01 -6.35753036e-01
9.84095097e-01 4.97144610e-01 7.12257206e-01 -1.02755554e-01
-2.94484943e-01 2.06434786e-01 -5.41864634e-01 -2.18313988e-02
8.14575195e-01 -4.06473368e-01 -1.33859470e-01 4.15046692e-01
-1.26972154e-01 -1.17767870e+00 -1.99697856e-02 3.75189096e-01
7.79617548e-01 3.89094472e-01 -1.77149355e-01 1.34557202e-01
3.86734784e-01 -1.55387327e-01 4.22125030e-03 3.74411166e-01
8.42150211e-01 1.09308946e+00 1.58895299e-01 -1.52354702e-01
-1.57132447e+00 -1.51774800e+00 -8.18248913e-02 5.10724545e-01
-1.32587209e-01 -6.46081448e-01 -4.27027285e-01 -2.99017012e-01
7.30395094e-02 9.09261346e-01 -2.14386925e-01 8.07925463e-02
-5.37631035e-01 -4.49427456e-01 5.21592438e-01 4.00803089e-01
7.64050409e-02 -8.90120804e-01 -2.32004151e-01 2.57747620e-01
-3.59195590e-01 -1.16743505e+00 -1.69233382e-01 -4.21999067e-01
-1.28677928e+00 -6.15181684e-01 -7.82179236e-01 -6.30476594e-01
8.95207524e-01 2.94597536e-01 1.23799920e+00 2.54750773e-02
-2.29576796e-01 6.07428193e-01 -2.84866810e-01 -5.18372953e-01
-7.88090944e-01 1.48502281e-02 -8.88802856e-02 1.38809502e-01
-2.88671881e-01 -1.08506811e+00 -4.04425085e-01 -1.15346070e-02
-1.38678193e+00 4.77605432e-01 6.10437989e-01 7.22164333e-01
1.04058588e+00 -2.27632329e-01 3.43956649e-01 -6.81389511e-01
4.88768458e-01 -3.31909716e-01 -5.99989355e-01 -1.90265067e-02
-2.65824199e-01 3.46593887e-01 5.23990452e-01 -2.09117934e-01
-1.10133231e+00 2.42745012e-01 -4.04154837e-01 -8.16394269e-01
-1.64815187e-01 2.21367270e-01 -1.55341148e-01 8.94905552e-02
3.26545119e-01 5.95913708e-01 4.68836427e-02 -5.43418705e-01
8.34136009e-01 3.22172731e-01 5.79289019e-01 -9.42797303e-01
1.26596904e+00 8.44285250e-01 3.46860737e-01 -9.79194701e-01
-3.21117073e-01 -2.52526283e-01 -5.87256253e-01 7.79088587e-02
8.15311253e-01 -1.07283938e+00 -4.68647122e-01 3.44250441e-01
-1.61672974e+00 4.16188128e-02 -5.22968352e-01 -2.25332398e-02
-1.02910256e+00 5.67206562e-01 -3.37867826e-01 -7.37679541e-01
-5.28676629e-01 -9.47321773e-01 1.85468912e+00 -2.26434842e-01
-2.29544770e-02 -7.59645104e-01 1.00405505e-02 2.67611951e-01
1.40959233e-01 3.71416062e-01 1.10197318e+00 -6.51342422e-02
-1.36676216e+00 -1.36759594e-01 1.60740083e-03 1.85272455e-01
1.83699027e-01 -2.86144055e-02 -8.78613830e-01 -2.94538528e-01
-2.78631806e-01 -3.78513306e-01 6.29171550e-01 1.10866182e-01
1.58539987e+00 -1.67374596e-01 -4.57509190e-01 8.02983224e-01
1.27249801e+00 -2.92949736e-01 7.76021898e-01 -3.47541720e-01
9.98391092e-01 3.64960313e-01 4.37195450e-01 6.47233486e-01
3.46035898e-01 7.95837224e-01 6.13766313e-01 1.87374622e-01
-2.49193981e-01 -8.06084454e-01 1.19178437e-01 1.17879307e+00
-2.39991710e-01 -1.47406250e-01 -8.89256299e-01 5.67391753e-01
-1.64105976e+00 -9.05823827e-01 -1.09637052e-01 2.13713241e+00
7.84285843e-01 -2.50579625e-01 -3.54428411e-01 2.28506885e-02
4.06363100e-01 1.21271417e-01 -1.83740288e-01 7.49683902e-02
-1.46582589e-01 4.42803741e-01 2.52057403e-01 1.95813134e-01
-8.35560977e-01 9.08033073e-01 5.92279387e+00 1.08549023e+00
-1.00295663e+00 1.64320171e-02 4.92231160e-01 -1.79832727e-02
-9.38864946e-01 1.06805548e-01 -5.93383610e-01 3.06199968e-01
5.98508835e-01 -2.18521118e-01 4.79769289e-01 7.43313134e-01
-2.30976962e-03 1.98568240e-01 -1.29075956e+00 1.34379888e+00
-2.16880590e-02 -1.71299481e+00 6.48342013e-01 2.09897578e-01
9.65310574e-01 -7.87542909e-02 1.92165852e-01 -1.85655337e-02
4.00086701e-01 -1.06891525e+00 9.61512625e-01 5.89852214e-01
1.24869406e+00 -7.16369927e-01 7.87929296e-02 4.78824586e-01
-1.32116592e+00 3.56106311e-01 -5.78959286e-01 2.16671854e-01
5.06803155e-01 7.39509523e-01 -1.11265063e+00 1.05929232e+00
5.33787549e-01 7.26585925e-01 -2.19743714e-01 8.44526052e-01
-1.62815109e-01 4.01848942e-01 -5.62259018e-01 2.46382996e-01
4.46593799e-02 -3.29181612e-01 7.13763237e-01 7.56254852e-01
1.00230014e+00 1.62304834e-01 1.95680857e-01 1.28572178e+00
-1.12457931e-01 6.28172159e-02 -1.02650070e+00 -1.04820095e-01
5.02779603e-01 1.12347579e+00 -5.76138735e-01 -3.86914521e-01
-1.48801312e-01 9.73382294e-01 2.08274603e-01 1.92117065e-01
-6.25793278e-01 2.38780137e-02 6.43833220e-01 2.60062873e-01
4.67267185e-01 -5.84566474e-01 -6.23217165e-01 -1.17720604e+00
2.27150619e-02 -6.30971074e-01 -1.46235272e-01 -1.03750432e+00
-1.36978590e+00 5.81957936e-01 2.17887983e-01 -1.62421525e+00
-5.52795947e-01 -2.69320369e-01 -4.76793498e-01 9.17871118e-01
-1.21277893e+00 -1.68634415e+00 -2.76187599e-01 7.48428524e-01
5.91795444e-01 -1.82733089e-01 9.66680288e-01 1.04324654e-01
3.60169202e-01 1.59870043e-01 -7.53288195e-02 -2.41692841e-01
3.67991507e-01 -1.17497408e+00 9.71838236e-01 5.63933372e-01
5.43203473e-01 4.53686208e-01 6.06438875e-01 -7.51930296e-01
-1.72077489e+00 -1.33977008e+00 7.08222210e-01 -6.89656675e-01
3.09814781e-01 -6.34864688e-01 -7.45684028e-01 4.89152640e-01
-8.39134455e-02 -4.44878750e-02 3.66204500e-01 -2.87816525e-01
-3.94432038e-01 3.44104581e-02 -8.60566974e-01 7.30420768e-01
1.31816936e+00 -5.08760035e-01 -4.40726131e-01 4.56266671e-01
9.34386432e-01 -6.87236726e-01 -8.89219701e-01 3.63609225e-01
1.75385833e-01 -6.87952995e-01 1.30865967e+00 -3.12574118e-01
7.41438568e-01 -3.18801612e-01 -2.19735414e-01 -1.32385600e+00
-1.11917853e-01 -8.06621909e-01 1.02281407e-01 1.32187176e+00
2.32633337e-01 -3.57560098e-01 9.19157982e-01 3.22802961e-01
-3.54218483e-01 -4.92227465e-01 -1.03826988e+00 -5.47664583e-01
1.24986805e-01 -7.69825041e-01 1.14333606e+00 6.89296305e-01
-4.33564663e-01 2.99737781e-01 -3.51449132e-01 1.68641016e-01
8.03123891e-01 5.99209845e-01 1.20387053e+00 -1.23078144e+00
-1.87358320e-01 -2.76790082e-01 -3.14533085e-01 -1.28919601e+00
2.66656101e-01 -1.19332063e+00 9.36343819e-02 -1.56719375e+00
6.54156925e-03 -9.60718870e-01 5.10505736e-01 3.09082031e-01
5.11114337e-02 2.81365275e-01 5.32607615e-01 3.80840331e-01
-1.39956817e-01 9.58757162e-01 1.33150637e+00 -2.73213059e-01
-4.81075570e-02 2.20645308e-01 -4.42934513e-01 4.54744369e-01
4.71115321e-01 -3.46515536e-01 -6.00720227e-01 -6.07767880e-01
3.56213689e-01 2.56497860e-01 5.95797539e-01 -1.02688038e+00
3.65631953e-02 -1.52214989e-01 2.53343791e-01 -9.32903171e-01
6.78064466e-01 -9.51978207e-01 6.85793519e-01 8.51188749e-02
-7.60412291e-02 8.14302787e-02 1.12467945e-01 6.17172599e-01
-2.84481198e-01 1.14812702e-01 2.52653003e-01 -2.74157524e-01
-2.64662266e-01 7.89740741e-01 9.16845351e-02 -1.85470045e-01
8.12355518e-01 -2.09777609e-01 1.04481369e-01 -5.03981233e-01
-6.26755238e-01 -8.92367810e-02 7.23129749e-01 6.82569027e-01
1.05927360e+00 -1.68877530e+00 -9.06067908e-01 5.27610958e-01
1.82375148e-01 6.85959339e-01 2.15293586e-01 3.81825954e-01
-4.98049587e-01 4.85752136e-01 -6.94499537e-02 -1.10716820e+00
-1.17673922e+00 4.13709551e-01 -1.29996702e-01 -3.13282013e-01
-7.48595953e-01 6.47843838e-01 4.72232968e-01 -5.68697751e-01
-1.93572983e-01 -4.98328716e-01 3.87071341e-01 -2.76689708e-01
2.86760062e-01 -3.33809033e-02 1.46284461e-01 -6.23830676e-01
4.74150889e-02 6.94269300e-01 3.19314152e-01 -3.55326742e-01
1.42727089e+00 4.23447900e-02 -3.14543486e-01 2.49995515e-01
1.04712069e+00 7.57426620e-02 -1.26725292e+00 -1.15297429e-01
-2.76255816e-01 -6.35623038e-01 -3.46968710e-01 -2.81534791e-01
-8.96646261e-01 1.00413501e+00 1.25785932e-01 1.16275176e-01
8.99696767e-01 3.27268571e-01 8.89659941e-01 2.30673835e-01
8.61880302e-01 -4.83827978e-01 1.51663765e-01 5.14973581e-01
1.30265570e+00 -7.47554481e-01 -4.26421426e-02 -7.97089100e-01
-4.75139737e-01 1.01035368e+00 1.01665363e-01 -1.78732142e-01
5.55923223e-01 2.21586362e-01 -3.49469543e-01 -1.04750879e-01
-9.52094793e-01 3.65914181e-02 4.48709458e-01 8.38091612e-01
2.01967552e-01 1.53733566e-01 1.42888259e-02 2.21853420e-01
-6.36415124e-01 -5.42883128e-02 3.52840215e-01 8.15503240e-01
-1.14935376e-01 -1.53803635e+00 -5.23834169e-01 3.37448269e-01
-2.05784217e-01 -2.02698007e-01 -1.60334975e-01 4.09669578e-01
4.44701575e-02 4.86735791e-01 2.15027973e-01 -2.24835977e-01
2.93934137e-01 1.96266130e-01 6.30258799e-01 -7.74506629e-01
-2.83866405e-01 1.14983097e-01 -9.74441692e-02 -5.15651584e-01
-4.49203730e-01 -8.28569174e-01 -1.17181242e+00 -2.66989917e-01
8.15066621e-02 8.07873756e-02 8.56859326e-01 6.35764539e-01
5.87967753e-01 2.57125020e-01 5.11082888e-01 -1.49041343e+00
-6.26443744e-01 -6.18822932e-01 -2.79239058e-01 8.84405971e-01
1.74471945e-01 -4.78505015e-01 -5.40905632e-03 4.64695066e-01] | [8.894972801208496, -3.6304080486297607] |
8ca9ed0c-30ee-4ac7-9c69-e9de4404d5fa | surfacenet-adversarial-svbrdf-estimation-from | 2107.11298 | null | https://arxiv.org/abs/2107.11298v1 | https://arxiv.org/pdf/2107.11298v1.pdf | SurfaceNet: Adversarial SVBRDF Estimation from a Single Image | In this paper we present SurfaceNet, an approach for estimating spatially-varying bidirectional reflectance distribution function (SVBRDF) material properties from a single image. We pose the problem as an image translation task and propose a novel patch-based generative adversarial network (GAN) that is able to produce high-quality, high-resolution surface reflectance maps. The employment of the GAN paradigm has a twofold objective: 1) allowing the model to recover finer details than standard translation models; 2) reducing the domain shift between synthetic and real data distributions in an unsupervised way. An extensive evaluation, carried out on a public benchmark of synthetic and real images under different illumination conditions, shows that SurfaceNet largely outperforms existing SVBRDF reconstruction methods, both quantitatively and qualitatively. Furthermore, SurfaceNet exhibits a remarkable ability in generating high-quality maps from real samples without any supervision at training time. | ['Concetto Spampinato', 'Simone Palazzo', 'Giuseppe Vecchio'] | 2021-07-23 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Vecchio_SurfaceNet_Adversarial_SVBRDF_Estimation_From_a_Single_Image_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Vecchio_SurfaceNet_Adversarial_SVBRDF_Estimation_From_a_Single_Image_ICCV_2021_paper.pdf | iccv-2021-1 | ['svbrdf-estimation'] | ['computer-vision'] | [ 8.10643017e-01 1.11561514e-01 3.29623431e-01 -1.48006678e-01
-9.88349736e-01 -4.63512599e-01 7.91848958e-01 -6.90965533e-01
8.73004496e-02 1.04562843e+00 -3.67627926e-02 -6.50316104e-02
2.28768602e-01 -1.12776721e+00 -1.16482365e+00 -9.53968585e-01
5.43588698e-01 4.96146113e-01 1.08506054e-01 -1.74344927e-01
-1.88976359e-02 7.12897837e-01 -1.60091352e+00 1.64570272e-01
1.05179489e+00 1.06003773e+00 2.57934123e-01 5.32860160e-01
1.76951904e-02 5.36147952e-01 -4.66181278e-01 -3.46281171e-01
6.08930171e-01 -6.44875467e-01 -6.91526353e-01 2.88386077e-01
6.30997002e-01 -2.79009879e-01 -3.32824746e-03 1.00185156e+00
4.17649060e-01 -8.86228925e-04 9.92237985e-01 -8.19837034e-01
-8.53277564e-01 -6.83196709e-02 -5.81057668e-01 -4.05195683e-01
4.77338791e-01 2.45606795e-01 6.39332831e-01 -8.47635686e-01
8.42611492e-01 1.05089641e+00 5.50264418e-01 6.28464162e-01
-1.87752151e+00 -4.01792735e-01 -3.40397239e-01 -1.53289646e-01
-1.37329257e+00 -2.95659751e-01 1.06397867e+00 -4.49352384e-01
5.10018110e-01 3.60424042e-01 5.52290976e-01 1.35681236e+00
1.97566375e-02 4.56211835e-01 1.80738902e+00 -5.31130552e-01
2.33325824e-01 1.21873923e-01 -7.56299436e-01 3.84602636e-01
9.79970694e-02 1.77572474e-01 -4.10847068e-01 5.59080504e-02
1.23834181e+00 -2.46945292e-01 -6.14637911e-01 -4.05799687e-01
-1.35870063e+00 4.61290002e-01 5.14510751e-01 1.63498521e-01
-6.72069132e-01 2.07984105e-01 -1.87084302e-02 4.61464435e-01
8.56571436e-01 3.35165977e-01 -2.39899859e-01 2.56821096e-01
-7.36657202e-01 8.45558196e-02 5.06620824e-01 8.11558664e-01
1.04761803e+00 3.59821439e-01 -8.91493037e-02 1.02438819e+00
1.95145473e-01 1.08840561e+00 1.54085994e-01 -9.59027112e-01
1.97506294e-01 3.43232542e-01 3.48748088e-01 -8.61131132e-01
3.13405514e-01 -1.70232907e-01 -9.91457164e-01 5.98367333e-01
3.25913936e-01 7.88886547e-02 -1.07104886e+00 1.59435928e+00
4.54050750e-01 9.26545337e-02 1.14842556e-01 8.31164956e-01
7.84140944e-01 6.68595076e-01 -1.48306698e-01 9.53913853e-03
9.95813191e-01 -7.26922691e-01 -5.17565787e-01 -8.25031027e-02
-3.88602763e-01 -8.56872976e-01 1.32437468e+00 4.56391245e-01
-1.30516708e+00 -5.46655238e-01 -9.52684045e-01 -4.76581929e-03
-1.86809361e-01 9.65973362e-02 4.30615783e-01 6.06510222e-01
-1.24108052e+00 5.34255981e-01 -5.70009291e-01 -8.95459875e-02
4.87614453e-01 8.84894654e-02 -5.79942286e-01 -3.01751733e-01
-8.69401932e-01 5.90982854e-01 -2.67862435e-02 -7.45522417e-03
-1.03659773e+00 -8.80378723e-01 -8.27856064e-01 -2.45646730e-01
-8.73776432e-03 -1.00129008e+00 7.72799313e-01 -1.52824223e+00
-2.01159644e+00 1.15182865e+00 -1.29178524e-01 -1.14368111e-01
7.97427118e-01 1.27465725e-01 -3.57557088e-01 2.65181154e-01
-9.03652888e-03 6.92220449e-01 1.39002800e+00 -1.89835477e+00
-7.51967132e-02 -1.36925817e-01 -3.30206603e-01 -7.92913698e-03
9.83569175e-02 -2.03006372e-01 -1.61509767e-01 -8.65750730e-01
-4.15983163e-02 -6.35511160e-01 1.38553709e-01 1.35255009e-01
-5.17793298e-01 2.18363956e-01 5.61067343e-01 -7.75715232e-01
2.56594867e-01 -1.83090901e+00 3.31108123e-01 2.15766951e-01
9.03441291e-03 2.58990973e-01 -2.68334985e-01 4.18776006e-01
-2.20555171e-01 -5.23992777e-02 -6.99211061e-01 -3.69927496e-01
-1.48886651e-01 5.18968105e-02 -3.84381473e-01 4.41582501e-01
3.14676851e-01 1.18519783e+00 -7.17188120e-01 -1.11939169e-01
3.08702111e-01 9.86448824e-01 -3.11836749e-01 4.80233938e-01
-4.13944155e-01 1.03049815e+00 3.00604515e-02 6.45748854e-01
1.02194417e+00 -1.39932513e-01 -8.12300444e-02 -3.02863091e-01
2.36817710e-02 -5.34602143e-02 -7.51726151e-01 1.72135580e+00
-7.01080739e-01 6.90846860e-01 1.16870828e-01 -7.20792592e-01
1.26223040e+00 3.07596773e-01 4.56642240e-01 -1.00879622e+00
-4.46428806e-02 3.39288145e-01 -5.18976986e-01 -2.26257890e-01
1.72773823e-01 -5.19665003e-01 4.40806299e-01 4.36049640e-01
-5.41114174e-02 -6.05109274e-01 -2.29431883e-01 -3.00980210e-01
8.41310441e-01 4.79109287e-01 -2.39058346e-01 -1.29728213e-01
7.42288649e-01 -1.90192357e-01 2.67764777e-01 3.92548859e-01
4.23862010e-01 1.21059299e+00 2.56885409e-01 -4.27451819e-01
-1.43771362e+00 -1.45063448e+00 -1.33433431e-01 3.55893135e-01
8.36075395e-02 3.37024122e-01 -8.90880048e-01 -5.87098837e-01
1.05209090e-02 6.42409682e-01 -8.45383584e-01 8.41107592e-02
-4.49292630e-01 -5.83333910e-01 1.56803831e-01 3.02858710e-01
7.10921168e-01 -1.33260417e+00 -1.45072743e-01 -8.11265036e-02
-3.47151369e-01 -1.28707469e+00 -2.75952160e-01 -3.07368726e-01
-6.61026657e-01 -9.35958326e-01 -1.16165364e+00 -6.44289434e-01
9.18266356e-01 2.10065618e-01 1.52664495e+00 -1.49453014e-01
-4.02304769e-01 5.12631953e-01 -3.38512540e-01 -3.16054076e-01
-7.70054042e-01 -2.05892220e-01 -4.27031517e-01 4.73289371e-01
-1.13269784e-01 -8.53305459e-01 -8.78608882e-01 6.71516299e-01
-1.22100782e+00 2.89734662e-01 5.39651394e-01 8.55334222e-01
9.55818713e-01 -5.68356887e-02 6.72262430e-01 -9.18909311e-01
3.64120305e-01 -2.67465085e-01 -7.04239130e-01 1.71231240e-01
-6.52001321e-01 -2.09404230e-01 6.93908155e-01 -3.56697589e-01
-1.37322688e+00 -1.23223670e-01 -2.30707884e-01 -3.41634154e-01
-3.89349580e-01 -1.65300310e-01 -2.92054296e-01 -4.89082992e-01
8.28857601e-01 6.61992490e-01 2.66362488e-01 -4.56732184e-01
3.27302963e-01 3.19058836e-01 7.01887548e-01 -7.03548253e-01
1.28325629e+00 8.93831253e-01 1.77476540e-01 -9.35109556e-01
-4.96240318e-01 -1.52392760e-01 -5.72356939e-01 -2.21801817e-01
8.78650486e-01 -9.29390311e-01 -5.10164440e-01 7.96754241e-01
-9.52054262e-01 -9.09302652e-01 -6.42722309e-01 -4.16570902e-02
-9.67838466e-01 2.65039057e-01 -4.20877814e-01 -5.12037337e-01
-4.71880555e-01 -9.51796114e-01 1.41336918e+00 1.14431858e-01
3.10378671e-01 -1.11114085e+00 1.52342305e-01 5.23179650e-01
7.83246815e-01 8.12962472e-01 7.87243605e-01 3.35715592e-01
-8.19659591e-01 3.85482349e-02 -4.78645235e-01 8.80413771e-01
4.45974857e-01 -8.16267654e-02 -1.09492540e+00 -3.76199901e-01
-1.11051567e-01 -4.05216962e-01 7.93184459e-01 3.35393727e-01
1.15917468e+00 -1.93954468e-01 1.82199612e-01 9.31986749e-01
1.77026486e+00 -1.59874469e-01 1.26807642e+00 1.98540598e-01
8.05739403e-01 6.09491885e-01 3.76947433e-01 -8.90670519e-04
1.07316874e-01 7.96594977e-01 6.33684099e-01 -7.32936442e-01
-8.00663352e-01 -3.84295970e-01 3.89429182e-01 2.60439515e-01
-4.90297139e-01 -2.92618394e-01 -5.58476031e-01 4.81092483e-01
-1.41075623e+00 -7.92783737e-01 -1.26139969e-01 2.23045516e+00
7.81467080e-01 -3.28368783e-01 6.85986802e-02 1.88021809e-02
5.73376238e-01 3.72223705e-02 -5.96220732e-01 -2.48201609e-01
-3.81249785e-01 4.98682112e-01 5.23766339e-01 4.06276017e-01
-6.26112998e-01 9.35227633e-01 6.71889687e+00 7.36959696e-01
-1.28567994e+00 1.51013315e-01 7.27787375e-01 4.07451510e-01
-8.92318010e-01 -4.12963957e-01 -1.80705249e-01 4.19551760e-01
5.85144639e-01 3.08826923e-01 8.17260385e-01 4.44956869e-01
1.15735166e-01 2.33474304e-03 -7.21651375e-01 1.06040037e+00
1.20787643e-01 -1.26562870e+00 2.36878678e-01 1.63222596e-01
1.15713108e+00 1.41069606e-01 2.12565809e-01 -3.97120982e-01
3.27265918e-01 -1.35286689e+00 6.45904720e-01 9.39010978e-01
1.56755435e+00 -6.58515096e-01 5.16382813e-01 1.03174515e-01
-8.37051034e-01 4.75124329e-01 -3.22941214e-01 5.22139966e-01
1.52208984e-01 7.19071925e-01 -6.79291785e-01 8.05445850e-01
6.35721684e-01 6.56956255e-01 -3.73026669e-01 6.62950754e-01
-6.82773530e-01 5.69649875e-01 -1.46705166e-01 5.39083898e-01
-2.45791301e-01 -7.02983499e-01 5.38728356e-01 8.88773143e-01
3.79456043e-01 -2.49340296e-01 -2.87030786e-01 1.50628865e+00
-1.79908335e-01 5.40203825e-02 -7.36946642e-01 1.48535594e-01
5.10809645e-02 1.22311699e+00 -5.40365458e-01 -3.97952162e-02
-1.88941956e-01 1.41433764e+00 7.29792342e-02 7.12749541e-01
-6.61040545e-01 -1.80288866e-01 6.33387148e-01 4.37247634e-01
3.95138383e-01 -1.73404679e-01 -2.28007272e-01 -1.18982649e+00
2.32182443e-01 -9.02183592e-01 -3.71413708e-01 -1.25164235e+00
-1.44324446e+00 7.34221518e-01 -2.53825873e-01 -1.14214277e+00
-1.05121143e-01 -6.25519812e-01 -5.46878040e-01 1.35201681e+00
-2.12069345e+00 -1.57309139e+00 -8.42766345e-01 7.52710402e-01
3.16554338e-01 -1.38227358e-01 9.68832910e-01 2.60284930e-01
-1.71847373e-01 3.17919612e-01 2.38688320e-01 -1.09677486e-01
8.28857243e-01 -1.45779145e+00 5.75122118e-01 6.64219141e-01
-6.89005181e-02 2.66813606e-01 8.29307377e-01 -3.87552947e-01
-1.34556341e+00 -1.10784233e+00 2.53411502e-01 -3.61849517e-01
2.04959348e-01 -3.79778773e-01 -9.67788517e-01 4.26585883e-01
3.58904302e-01 2.70497710e-01 4.22616988e-01 -5.42027414e-01
-4.46351975e-01 -2.86012143e-01 -1.58034062e+00 3.39866370e-01
9.60302353e-01 -5.46646237e-01 -2.23493055e-01 3.51217061e-01
5.58205843e-01 -3.77494693e-01 -1.12499774e+00 5.12665033e-01
5.03724098e-01 -1.30085230e+00 1.30479527e+00 -6.29998967e-02
7.92442560e-01 -3.90969753e-01 -1.09707505e-01 -1.60112870e+00
-9.87668037e-02 -7.67343163e-01 1.53640181e-01 1.18004692e+00
2.45468959e-01 -1.06332433e+00 6.99112475e-01 1.73457250e-01
4.25838418e-02 -4.69440550e-01 -5.85831225e-01 -8.12451601e-01
1.26467764e-01 -6.37100488e-02 8.26475501e-01 8.28155100e-01
-1.10485566e+00 7.51438458e-03 -4.41666961e-01 -4.24663424e-02
9.11877990e-01 2.70361692e-01 1.11474800e+00 -1.13922083e+00
-2.92191863e-01 -2.34376043e-01 -2.50716418e-01 -9.13924158e-01
3.24322104e-01 -7.35140800e-01 1.32974312e-01 -1.53728831e+00
5.88814616e-02 -5.60920060e-01 3.57507430e-02 3.60671252e-01
1.20698020e-01 9.08069611e-01 -2.88393855e-01 3.01828176e-01
2.46737879e-02 8.05140078e-01 1.81619608e+00 -2.09620386e-01
1.43016949e-01 -6.00794330e-02 -5.32445371e-01 4.13368821e-01
6.89926386e-01 -2.67982692e-01 -4.29296136e-01 -3.40516448e-01
2.93165147e-01 -1.33810326e-01 7.19634354e-01 -8.67259562e-01
-4.93360966e-01 -1.75147638e-01 4.69891667e-01 -1.60818532e-01
3.24406981e-01 -9.28723574e-01 5.83531141e-01 5.86892031e-02
-1.96898244e-02 -3.58163238e-01 6.67412952e-02 5.55097580e-01
-2.92944402e-01 1.78391322e-01 1.13902438e+00 -7.93272853e-02
-2.44273573e-01 3.61601919e-01 7.79234245e-02 -7.56439120e-02
7.70313978e-01 -3.23516369e-01 -2.86926061e-01 -3.45750302e-01
-4.39795405e-01 -4.91799861e-01 1.21184695e+00 1.87396049e-01
5.60653269e-01 -1.48537517e+00 -9.20121431e-01 4.95255858e-01
1.80273235e-01 2.49616534e-01 2.27100715e-01 7.04548240e-01
-8.01092327e-01 -2.16260161e-02 -3.89640540e-01 -6.74238503e-01
-9.40894306e-01 2.39174142e-01 5.64515829e-01 -1.87271491e-01
-8.47874999e-01 6.05192602e-01 3.51375967e-01 -6.00770354e-01
-4.26028520e-01 -2.05273069e-02 1.52348936e-01 -3.62284064e-01
3.46179217e-01 1.34293303e-01 2.58426547e-01 -7.28140771e-01
9.24011692e-02 8.31537068e-01 4.93049890e-01 -5.02544418e-02
1.62434697e+00 -1.00857086e-01 -3.60458493e-01 2.02538237e-01
1.08662021e+00 2.04924881e-01 -1.68509185e+00 -2.44176254e-01
-5.85477054e-01 -8.65919709e-01 9.77402404e-02 -1.05018842e+00
-1.36004174e+00 6.25054717e-01 4.96609956e-01 3.48612905e-01
1.36174452e+00 -8.83345306e-02 9.38103914e-01 -2.26107031e-01
5.25725782e-01 -8.20572913e-01 2.00623766e-01 1.22778505e-01
1.30136096e+00 -1.24292231e+00 -1.49758101e-01 -6.89365923e-01
-4.79139298e-01 1.05383217e+00 2.39130110e-01 -3.35293353e-01
3.85956258e-01 1.03543848e-01 2.26441085e-01 -2.70576119e-01
-3.38006556e-01 -7.81633928e-02 6.50321066e-01 9.81525362e-01
2.97589272e-01 3.06954328e-02 7.75992945e-02 -2.84091860e-01
-2.20477097e-02 -1.42003834e-01 2.61228234e-01 3.72921497e-01
-1.67773105e-02 -1.33774555e+00 -4.70727831e-01 4.08314429e-02
-3.03239316e-01 1.15838699e-01 -4.28939551e-01 6.15129828e-01
-3.12565565e-02 6.59761548e-01 -4.04971503e-02 -1.25833362e-01
3.79628688e-01 -9.61552784e-02 7.78162777e-01 -4.26221162e-01
-2.91145444e-01 5.76895699e-02 -7.97108635e-02 -7.16644526e-01
-6.76131070e-01 -6.30027711e-01 -7.40580797e-01 -1.95610851e-01
-4.79408354e-02 -2.54819512e-01 1.02639532e+00 6.63111210e-01
3.19200665e-01 6.51374578e-01 9.52474535e-01 -1.10952139e+00
-1.30688906e-01 -8.61149549e-01 -7.38740504e-01 8.49757433e-01
4.65932280e-01 -6.18371248e-01 -5.28120875e-01 2.77949989e-01] | [11.658092498779297, -0.7056453227996826] |
e016fb3f-05bc-48f9-bcdb-4a8f6fb10d7e | lesion-inspired-denoising-network-connecting | 2104.08845 | null | https://arxiv.org/abs/2104.08845v1 | https://arxiv.org/pdf/2104.08845v1.pdf | Lesion-Inspired Denoising Network: Connecting Medical Image Denoising and Lesion Detection | Deep learning has achieved notable performance in the denoising task of low-quality medical images and the detection task of lesions, respectively. However, existing low-quality medical image denoising approaches are disconnected from the detection task of lesions. Intuitively, the quality of denoised images will influence the lesion detection accuracy that in turn can be used to affect the denoising performance. To this end, we propose a play-and-plug medical image denoising framework, namely Lesion-Inspired Denoising Network (LIDnet), to collaboratively improve both denoising performance and detection accuracy of denoised medical images. Specifically, we propose to insert the feedback of downstream detection task into existing denoising framework by jointly learning a multi-loss objective. Instead of using perceptual loss calculated on the entire feature map, a novel region-of-interest (ROI) perceptual loss induced by the lesion detection task is proposed to further connect these two tasks. To achieve better optimization for overall framework, we propose a customized collaborative training strategy for LIDnet. On consideration of clinical usability and imaging characteristics, three low-dose CT images datasets are used to evaluate the effectiveness of the proposed LIDnet. Experiments show that, by equipping with LIDnet, both of the denoising and lesion detection performance of baseline methods can be significantly improved. | ['Xiaorong Pu', 'Jiayu Sun', 'Yazhou Ren', 'Kun Long', 'Kecheng Chen'] | 2021-04-18 | null | null | null | null | ['medical-image-denoising'] | ['computer-vision'] | [ 2.66094744e-01 4.49039694e-03 2.31365204e-01 -4.36709315e-01
-1.05264699e+00 -1.62898749e-01 4.16974455e-01 1.74772218e-01
-5.25144875e-01 3.93728435e-01 2.31583402e-01 -2.67553404e-02
-2.38566831e-01 -8.11565757e-01 -4.34844732e-01 -9.98927653e-01
1.52825922e-01 -2.50975430e-01 2.24157423e-01 -1.58631027e-01
-4.17703576e-02 3.88592541e-01 -1.04960907e+00 4.07459140e-01
1.15977824e+00 1.06732774e+00 5.74698925e-01 3.63769233e-01
2.91609406e-01 8.15703928e-01 -3.63543183e-01 -3.34718168e-01
3.03730696e-01 -4.16991293e-01 -4.41003978e-01 1.40571237e-01
1.62228069e-03 -6.26436591e-01 -5.07197499e-01 1.20890141e+00
8.41335118e-01 4.60506864e-02 5.10715783e-01 -6.34542406e-01
-3.34018201e-01 4.48533714e-01 -4.18679625e-01 3.83731455e-01
-1.92438602e-01 5.66213608e-01 6.29636168e-01 -6.65517151e-01
3.27492476e-01 1.00575364e+00 7.33575523e-01 3.94654483e-01
-1.11191225e+00 -3.60850751e-01 5.59478300e-03 2.14290842e-01
-1.06514466e+00 -3.45581621e-01 7.40898311e-01 -2.18759164e-01
2.84713358e-01 2.86996305e-01 4.63382334e-01 9.62849379e-01
4.26186979e-01 7.08592594e-01 1.01027727e+00 -7.22924620e-02
1.46239921e-01 -5.16849943e-02 -1.98910475e-01 4.80547249e-01
1.69252202e-01 1.65068671e-01 -9.73176118e-03 1.70884639e-01
7.94287384e-01 1.12347372e-01 -5.80000818e-01 -3.18249315e-01
-9.67373669e-01 7.23262548e-01 1.05533552e+00 3.99160653e-01
-6.77974224e-01 3.69094491e-01 5.12037873e-01 2.40580831e-02
5.25284350e-01 1.96481332e-01 6.13620095e-02 1.52146772e-01
-9.01943445e-01 -1.47392511e-01 2.61427641e-01 1.20853260e-01
3.00679386e-01 -1.33695766e-01 -6.92315042e-01 9.68948603e-01
2.04749972e-01 5.14279641e-02 5.60954928e-01 -7.94831634e-01
2.69393355e-01 5.83866179e-01 -4.77366485e-02 -9.11127150e-01
-5.33752203e-01 -9.11088586e-01 -1.19815075e+00 2.91510910e-01
3.47720653e-01 -5.37474193e-02 -1.15094447e+00 1.70035458e+00
4.12576139e-01 5.23609333e-02 -5.69809303e-02 1.30856133e+00
8.56645644e-01 4.21163350e-01 2.68851876e-01 -3.06044817e-01
1.32541955e+00 -9.24995840e-01 -8.48891377e-01 -2.66734272e-01
5.46948373e-01 -6.89412892e-01 9.93144333e-01 3.74195904e-01
-1.42137456e+00 -5.53197682e-01 -1.18411493e+00 4.88376692e-02
1.07117929e-01 2.41174012e-01 4.53915775e-01 5.71371019e-01
-8.99199128e-01 6.40999377e-01 -1.12991452e+00 -1.82002764e-02
6.74306333e-01 2.69179970e-01 -1.28883973e-01 -5.72912693e-01
-1.16301453e+00 9.87078607e-01 8.70999321e-02 4.18300956e-01
-1.15525293e+00 -9.39634204e-01 -6.02484167e-01 2.13514604e-02
3.07502151e-01 -1.05188489e+00 1.00634837e+00 -8.75063002e-01
-1.18428040e+00 6.56882346e-01 2.01113567e-01 -4.22450840e-01
1.01675391e+00 -1.26930803e-01 -2.91208625e-01 4.19895768e-01
1.33844629e-01 5.09794474e-01 8.76891136e-01 -1.34578609e+00
-4.84942883e-01 -4.74690884e-01 -2.71793343e-02 2.35039264e-01
-3.32745045e-01 -1.87381998e-01 -5.99351048e-01 -9.24420476e-01
1.43650174e-01 -4.89029974e-01 -6.32087767e-01 1.85376331e-01
-2.38304108e-01 1.05889946e-01 4.84976470e-01 -9.62657988e-01
1.14759409e+00 -2.36974120e+00 1.65144987e-02 1.78028241e-01
3.28659534e-01 2.20504865e-01 -2.32838601e-01 -2.36117765e-01
-1.06548667e-01 1.13856576e-01 -4.12487835e-01 -3.28758478e-01
-4.16104138e-01 1.19581208e-01 4.20537353e-01 6.44317329e-01
2.88706630e-01 8.90056729e-01 -7.97513545e-01 -3.47331434e-01
2.86690116e-01 7.45586455e-01 -5.92541933e-01 2.36446828e-01
1.53374806e-01 7.84546494e-01 -4.84590858e-01 3.98485363e-01
7.96872258e-01 -9.58411917e-02 3.23868282e-02 -7.71579206e-01
1.02319337e-01 3.51434350e-02 -9.42524731e-01 1.69404101e+00
-7.72963047e-01 1.32737741e-01 5.47825277e-01 -1.17233849e+00
6.47250533e-01 3.78107280e-01 7.40364134e-01 -1.06094408e+00
2.06450328e-01 2.45722443e-01 3.03184986e-01 -6.87526762e-01
2.80693583e-02 -4.37244177e-01 2.42977828e-01 1.17535099e-01
-1.38816565e-01 -1.98047496e-02 -3.29133719e-02 1.12200789e-01
1.26751542e+00 -2.12766871e-01 -2.88274251e-02 -1.34809746e-03
5.53201735e-01 -2.22461700e-01 5.92214882e-01 7.84489572e-01
-5.08019924e-01 8.48049521e-01 5.26804268e-01 -1.12390071e-01
-8.45491886e-01 -1.11809850e+00 -3.19556385e-01 8.30994964e-01
3.33461493e-01 8.54967721e-03 -8.26064467e-01 -7.29715645e-01
-1.68621361e-01 4.49486405e-01 -6.90611482e-01 -5.60053885e-01
-7.35228658e-01 -1.30727232e+00 2.89893448e-01 4.76133943e-01
5.34315586e-01 -9.19536889e-01 -5.90889037e-01 3.86887252e-01
-4.16626394e-01 -1.01072943e+00 -5.05500972e-01 3.02005142e-01
-9.44232821e-01 -1.04308164e+00 -8.90891433e-01 -6.97083473e-01
6.72832608e-01 2.57206112e-01 8.71118009e-01 3.73798698e-01
-5.50175548e-01 3.02876234e-01 -3.67052108e-01 -6.86863437e-02
-6.73249006e-01 -1.03490837e-01 -3.65860462e-01 1.11952253e-01
-2.64330745e-01 -6.82252526e-01 -1.27131546e+00 3.93169135e-01
-1.33604944e+00 -1.48748845e-01 1.04508674e+00 1.01214004e+00
5.49843669e-01 3.45161915e-01 7.14246809e-01 -6.42498910e-01
7.13853836e-01 -3.44867349e-01 -2.82303631e-01 2.01555267e-01
-4.63314623e-01 -1.14626586e-01 4.28918809e-01 -1.91121906e-01
-1.12153757e+00 5.06501161e-02 -4.64929044e-01 -2.46708885e-01
5.69612831e-02 6.36815488e-01 -2.51507074e-01 -9.32362899e-02
5.17172635e-01 2.52346516e-01 2.57384658e-01 -3.42899591e-01
1.60382539e-01 4.68359828e-01 6.44136667e-01 -1.20517507e-01
4.82474506e-01 6.02861762e-01 1.31893948e-01 -4.17524397e-01
-8.39484751e-01 -5.03783882e-01 -2.50331849e-01 -3.06018025e-01
9.65977848e-01 -1.06424487e+00 -5.98478198e-01 5.71859419e-01
-9.93475199e-01 -2.36523777e-01 -2.57612586e-01 4.72707480e-01
-3.83652240e-01 4.86744910e-01 -6.60432696e-01 -5.20298779e-01
-5.29675722e-01 -1.66692007e+00 9.71831858e-01 -2.86034984e-03
2.12102905e-01 -9.46983099e-01 -3.53186727e-01 4.68397111e-01
7.24012375e-01 4.27015781e-01 1.08514142e+00 -3.10209900e-01
-4.77437228e-01 -3.06351036e-01 -4.48390156e-01 8.36918652e-01
2.32505247e-01 -4.83290255e-01 -8.27740669e-01 -4.82557952e-01
4.87390816e-01 -1.03662267e-01 1.23997867e+00 9.04459178e-01
1.31476474e+00 1.36796579e-01 -1.28904849e-01 8.49860668e-01
1.57259512e+00 2.17500906e-02 9.93145466e-01 3.55274409e-01
6.82717860e-01 5.51157296e-01 4.68490481e-01 3.87297750e-01
1.42243311e-01 6.21464312e-01 8.38227391e-01 -5.07104814e-01
-5.96062541e-01 1.10451475e-01 2.78926820e-01 6.29908442e-01
4.72183973e-02 -2.20915943e-01 -8.33212793e-01 6.32037699e-01
-1.61015451e+00 -5.56536555e-01 -3.07069719e-01 2.07276487e+00
7.76098073e-01 1.98406965e-01 -7.38624334e-02 7.60453418e-02
7.94841588e-01 5.82693778e-02 -6.64339185e-01 5.09990938e-02
-2.32031848e-02 9.16815251e-02 4.08677220e-01 3.55524540e-01
-1.15151370e+00 3.68262380e-01 5.53743029e+00 1.06349623e+00
-1.26551926e+00 4.62259978e-01 9.28288698e-01 -6.11497276e-02
-1.64235592e-01 -3.12021285e-01 -2.85346389e-01 5.21943808e-01
5.11447787e-01 1.80036351e-01 1.52685270e-01 4.67946261e-01
7.85331845e-01 -2.48908460e-01 -9.46729779e-01 8.03998470e-01
-1.96597502e-01 -1.04472780e+00 -1.38491318e-02 -1.79115161e-02
4.83447939e-01 -1.14452124e-01 2.98000067e-01 7.26052523e-02
6.97888527e-03 -1.03909039e+00 3.95905912e-01 5.79675734e-01
5.46987951e-01 -8.52483571e-01 1.01278925e+00 2.46301964e-01
-7.62403607e-01 -1.32448435e-01 -2.68975228e-01 3.94896775e-01
3.63055676e-01 1.07681501e+00 -5.41390002e-01 6.69405341e-01
7.37307250e-01 6.55766249e-01 -6.20586455e-01 1.31405497e+00
-3.28408629e-01 6.21590436e-01 7.78452354e-03 5.40642262e-01
3.54370236e-01 -6.79092258e-02 6.19550467e-01 1.04538786e+00
2.30521590e-01 2.43776083e-01 5.87571599e-02 9.56645310e-01
-1.38894618e-01 -5.82198203e-02 1.72439422e-02 4.86742705e-01
-5.94176129e-02 1.50025260e+00 -7.29943871e-01 -1.98077053e-01
-1.55594692e-01 9.68262136e-01 -9.84403342e-02 3.55977386e-01
-8.56214285e-01 -1.72555998e-01 4.98152435e-01 4.09112096e-01
4.37678337e-01 1.94482818e-01 -5.85433900e-01 -9.22390819e-01
1.57979175e-01 -8.61352503e-01 2.79208273e-01 -5.57331443e-01
-1.35778379e+00 5.03480673e-01 -4.25023317e-01 -1.19340396e+00
2.46670097e-01 -3.43468726e-01 -7.94666529e-01 9.60812688e-01
-1.71014297e+00 -1.00973368e+00 -6.33004785e-01 4.85444665e-01
4.09505367e-01 2.66997665e-01 2.69227773e-01 6.40799701e-01
-6.65459692e-01 6.80833817e-01 1.52465820e-01 -4.31482941e-02
7.74329901e-01 -1.07146740e+00 -8.82971138e-02 9.61824358e-01
-6.23536289e-01 2.39970461e-01 5.03603101e-01 -5.79000771e-01
-1.01108778e+00 -1.32853913e+00 7.05084428e-02 -5.93861192e-02
4.09148991e-01 -3.27420495e-02 -9.87977982e-01 1.49897654e-02
5.17470576e-02 3.04760277e-01 3.22668433e-01 -4.60786223e-01
1.04442514e-01 -2.89585620e-01 -1.49200952e+00 5.50687492e-01
8.62174273e-01 -1.86094746e-01 -2.78193533e-01 3.04029703e-01
5.70510566e-01 -2.79470325e-01 -1.10718942e+00 7.66609728e-01
2.26165414e-01 -9.93428409e-01 1.23523891e+00 -1.81812376e-01
6.66803896e-01 -1.64580300e-01 -1.38953449e-02 -1.49324691e+00
-4.29257244e-01 -1.27586037e-01 2.40582988e-01 1.01153803e+00
1.18408382e-01 -4.42154408e-01 6.72227859e-01 2.63109595e-01
-5.01492679e-01 -8.50481510e-01 -1.02993429e+00 -4.03440535e-01
1.83404982e-01 -4.05097395e-01 1.76782355e-01 5.91686726e-01
-4.53448325e-01 1.33435801e-01 -2.07341656e-01 3.03249717e-01
7.51432836e-01 -4.34964269e-01 3.21236759e-01 -7.89358556e-01
-3.22838932e-01 -4.83387798e-01 -3.46524328e-01 -1.00259674e+00
-3.55140805e-01 -8.96735132e-01 2.77945101e-01 -1.81502628e+00
3.35552186e-01 -5.66911221e-01 -6.13619328e-01 3.50840658e-01
-6.04331851e-01 4.79240239e-01 1.89116001e-01 2.88142353e-01
-5.06435871e-01 6.30286753e-01 1.83961308e+00 -3.22822720e-01
-8.59516561e-02 8.63223970e-02 -6.82796299e-01 6.23820364e-01
5.86085021e-01 -5.92752218e-01 -2.06616715e-01 -4.73630041e-01
-6.74573109e-02 1.65039837e-01 7.17060506e-01 -1.05041265e+00
1.70671538e-01 2.26987123e-01 4.86003518e-01 -3.29403579e-01
1.53262794e-01 -7.84410000e-01 -1.39981672e-01 8.04825842e-01
-2.47553572e-01 -4.27083284e-01 7.57217482e-02 5.62378585e-01
-4.81624156e-01 -1.78287268e-01 1.27430499e+00 -2.06711501e-01
-4.75947142e-01 1.47412032e-01 -3.75490636e-01 -1.19282462e-01
1.06312108e+00 -2.82366037e-01 -6.14475720e-02 -3.17368805e-01
-1.03317857e+00 2.45382696e-01 8.68756175e-02 2.71212190e-01
7.27222145e-01 -1.08363783e+00 -9.29775596e-01 -1.30625078e-02
-1.01386867e-01 1.47229746e-01 6.84398055e-01 1.51532602e+00
-4.07051831e-01 4.57602963e-02 -8.34132582e-02 -7.32784688e-01
-9.51178372e-01 4.33840215e-01 7.33419657e-01 -7.60892570e-01
-5.59581697e-01 7.74737835e-01 4.92379725e-01 -1.86338827e-01
2.78906673e-01 -4.63979781e-01 -1.77688062e-01 -4.41414453e-02
4.05057400e-01 3.19097638e-01 5.40864825e-01 -2.55608350e-01
-2.60013372e-01 3.22762907e-01 -2.64833152e-01 2.01479554e-01
1.58079684e+00 -3.15822691e-01 -1.42652288e-01 -1.47902980e-01
1.22299206e+00 -3.77819568e-01 -1.40252328e+00 -7.99371377e-02
-2.69991994e-01 -2.73624033e-01 7.25733459e-01 -1.18259144e+00
-1.57507205e+00 7.98102200e-01 1.27283835e+00 4.63418178e-02
1.61798763e+00 -1.09459482e-01 1.00970769e+00 -3.36929142e-01
-7.18090013e-02 -8.53265822e-01 4.47980255e-01 -7.07501173e-02
8.93241167e-01 -1.44458854e+00 -2.86109634e-02 -6.20903730e-01
-4.77405876e-01 9.66408253e-01 4.88100797e-01 -8.51386786e-02
4.94742781e-01 2.29204088e-01 2.48250410e-01 -2.68567353e-01
-3.27687353e-01 -1.33502066e-01 2.84813643e-01 4.84275579e-01
3.40945542e-01 -1.08321905e-01 -6.24166191e-01 8.33915114e-01
3.56042385e-01 1.41694378e-02 4.40928519e-01 6.95428312e-01
-3.87151539e-01 -8.96608949e-01 -4.42310840e-01 5.42441905e-01
-7.03894377e-01 -1.09989673e-01 1.59524396e-01 5.60154140e-01
4.09390628e-01 9.54460025e-01 -1.59908220e-01 -2.19694674e-01
6.04910553e-01 -5.29169142e-01 2.51682162e-01 -3.98716509e-01
-8.65300953e-01 3.47618818e-01 -1.33112103e-01 -5.71524322e-01
-3.71696293e-01 -3.63161474e-01 -9.73361254e-01 -3.49683501e-02
-3.49624038e-01 -1.77526623e-01 7.12814569e-01 8.17990839e-01
7.24206045e-02 1.15195215e+00 7.27272689e-01 -6.27985299e-01
-7.83759058e-01 -8.98561954e-01 -4.81030852e-01 5.57725072e-01
3.07380706e-01 -4.81319964e-01 -3.39171261e-01 -2.02391818e-01] | [13.464948654174805, -2.461362838745117] |
49730753-9176-41d5-be4c-159b927fcbd5 | web-based-visualisation-of-head-pose-and | 1703.03949 | null | http://arxiv.org/abs/1703.03949v2 | http://arxiv.org/pdf/1703.03949v2.pdf | Web-based visualisation of head pose and facial expressions changes: monitoring human activity using depth data | Despite significant recent advances in the field of head pose estimation and
facial expression recognition, raising the cognitive level when analysing human
activity presents serious challenges to current concepts. Motivated by the need
of generating comprehensible visual representations from different sets of
data, we introduce a system capable of monitoring human activity through head
pose and facial expression changes, utilising an affordable 3D sensing
technology (Microsoft Kinect sensor). An approach build on discriminative
random regression forests was selected in order to rapidly and accurately
estimate head pose changes in unconstrained environment. In order to complete
the secondary process of recognising four universal dominant facial expressions
(happiness, anger, sadness and surprise), emotion recognition via facial
expressions (ERFE) was adopted. After that, a lightweight data exchange format
(JavaScript Object Notation-JSON) is employed, in order to manipulate the data
extracted from the two aforementioned settings. Such mechanism can yield a
platform for objective and effortless assessment of human activity within the
context of serious gaming and human-computer interaction. | ['Grigorios Kalliatakis', 'Nikolaos Vidakis', 'Georgios Triantafyllidis'] | 2017-03-11 | null | null | null | null | ['head-pose-estimation'] | ['computer-vision'] | [ 2.72967756e-01 2.02822223e-01 2.85589039e-01 -6.03974998e-01
-2.06257880e-01 -3.20133299e-01 5.97360194e-01 -1.62271574e-01
-4.71478254e-01 5.49357772e-01 2.69825250e-01 1.87617913e-01
8.89538378e-02 -3.57326061e-01 -4.03483063e-02 -6.17282569e-01
-2.36243784e-01 2.70516817e-02 -1.27698049e-01 -2.69435614e-01
2.19722375e-01 9.62440670e-01 -2.37471390e+00 -9.41364542e-02
2.28948444e-01 1.14899230e+00 -9.20753106e-02 7.25876510e-01
6.45285696e-02 6.34096801e-01 -4.40050602e-01 -1.51634708e-01
-2.76163090e-02 -3.03825110e-01 -4.05534863e-01 3.20423752e-01
-5.40321022e-02 -3.59209090e-01 3.35151941e-01 5.58979273e-01
9.99054909e-01 2.23540753e-01 4.57563221e-01 -1.50604475e+00
3.61597180e-01 -3.83492947e-01 -4.21671361e-01 -6.08417206e-02
1.36501622e+00 3.31608981e-01 1.19857654e-01 -8.77780557e-01
4.78553861e-01 1.01846683e+00 4.42703396e-01 7.76938796e-01
-7.66808987e-01 -6.65099621e-01 -1.49225602e-02 8.64696503e-02
-1.47802746e+00 -8.16347003e-01 9.22561824e-01 -4.12692189e-01
9.86403525e-01 7.30316937e-01 1.03118205e+00 1.42501879e+00
-1.03036918e-01 3.81058633e-01 1.50787222e+00 -6.51145399e-01
4.11857724e-01 3.09601128e-01 -1.84034690e-01 5.62750995e-01
-4.83379960e-02 -1.45679355e-01 -1.06981564e+00 -2.38721326e-01
5.32797635e-01 1.47744771e-02 -2.88114130e-01 -3.55933368e-01
-5.51152825e-01 3.81435096e-01 -2.73677539e-02 2.45486319e-01
-6.84046865e-01 -2.32041657e-01 3.97990197e-01 -2.34010033e-02
3.84846479e-01 -1.68071598e-01 -2.84998834e-01 -8.51333082e-01
-6.79065943e-01 2.13800013e-01 7.92633951e-01 5.66502988e-01
5.67915976e-01 3.23457196e-02 1.13359667e-01 5.63693047e-01
8.25154483e-01 3.97690594e-01 6.58286214e-01 -5.33306360e-01
6.63782582e-02 8.61745477e-01 2.73788739e-02 -1.03212023e+00
-7.35720217e-01 1.64377436e-01 -3.54281545e-01 7.69610524e-01
1.53013542e-01 -1.78799763e-01 -7.76215434e-01 1.50606585e+00
1.04502916e+00 -4.30457175e-01 -3.29139471e-01 9.17800605e-01
8.30620348e-01 -4.20882329e-02 4.39311862e-01 -4.53542739e-01
1.55146158e+00 -2.98483800e-02 -7.65133798e-01 -2.30383739e-01
4.16981876e-01 -4.37721401e-01 9.52057481e-01 5.11358678e-01
-8.48172188e-01 -3.22560787e-01 -9.84797239e-01 3.30802426e-02
-6.49166346e-01 -1.78806648e-01 5.67639589e-01 1.21240926e+00
-9.97665107e-01 1.09360263e-01 -8.58588099e-01 -6.21888161e-01
1.11454368e-01 6.60186291e-01 -7.59245396e-01 5.41910708e-01
-8.01846504e-01 1.06916976e+00 -7.94182494e-02 5.60236692e-01
-2.73173094e-01 -8.75001922e-02 -9.62630272e-01 -3.40186059e-01
-9.56926420e-02 -3.84681523e-01 1.01923907e+00 -1.19173765e+00
-2.16456008e+00 1.46470463e+00 -2.48654321e-01 1.28627867e-01
7.89799929e-01 -3.70273024e-01 -3.02062988e-01 -5.21587953e-03
-3.47634792e-01 3.26739728e-01 1.02223170e+00 -9.69804645e-01
-1.28876623e-02 -1.18568146e+00 -3.62952590e-01 4.58264291e-01
-1.57490194e-01 6.09947145e-01 -1.54299885e-01 -1.33892551e-01
1.97589040e-01 -7.59190381e-01 1.08633228e-01 1.37470827e-01
-8.07327591e-03 -9.98889506e-02 8.11809599e-01 -9.39627051e-01
1.19249964e+00 -2.08713698e+00 -6.67367056e-02 4.96042490e-01
2.98031598e-01 3.23476911e-01 6.35005593e-01 1.59307674e-01
-2.34810472e-01 -3.32896292e-01 1.43465037e-02 -4.56806928e-01
1.91959307e-01 1.90289289e-01 2.28052929e-01 6.41112685e-01
-6.18513152e-02 6.03162885e-01 -5.22712708e-01 -6.51190579e-01
4.79585528e-01 7.45418191e-01 -1.37821406e-01 5.90810359e-01
2.63352573e-01 5.83190858e-01 -2.98588037e-01 9.00609136e-01
6.59771919e-01 5.57468653e-01 3.20596099e-01 1.50524855e-01
-2.82652766e-01 1.14093898e-02 -1.28072011e+00 1.45492387e+00
-4.26146269e-01 2.91709095e-01 3.09237957e-01 -5.68100154e-01
1.28758419e+00 5.26032388e-01 4.72539812e-01 -7.35714793e-01
5.66486537e-01 -1.29745891e-02 -6.35223567e-01 -1.00984049e+00
2.62591839e-01 -6.32414818e-02 2.99681164e-02 3.24520558e-01
-4.44414765e-02 -5.04506333e-03 -4.22743320e-01 -4.43761706e-01
7.57695436e-01 7.32913673e-01 7.29149461e-01 2.12378189e-01
5.50382495e-01 -6.24110520e-01 1.47286683e-01 5.68284430e-02
-6.11566424e-01 4.53349173e-01 2.56529331e-01 -5.52618861e-01
-3.23354691e-01 -9.84284341e-01 1.58909306e-01 1.44671285e+00
-3.58050406e-01 -2.47724459e-01 -8.41025412e-01 -3.25484395e-01
-3.21519077e-01 2.02279225e-01 -6.28591657e-01 1.25176236e-01
-3.35845709e-01 -6.35491014e-01 4.42598581e-01 2.15541035e-01
2.92134076e-01 -1.32411969e+00 -1.67028332e+00 -3.04171182e-02
1.27489030e-01 -8.44873250e-01 3.08149815e-01 2.42855743e-01
-7.37831295e-01 -9.01136577e-01 -6.82681501e-01 -2.57626891e-01
4.64511573e-01 -2.00388104e-01 8.72220337e-01 -9.20431092e-02
-5.98103702e-01 7.52866268e-01 -4.79469299e-01 -6.24550045e-01
6.58016503e-02 -2.06158757e-01 1.69911250e-01 2.10934728e-01
7.56906450e-01 -8.81903231e-01 -6.43630385e-01 4.40778546e-02
-6.38076305e-01 -7.67534599e-02 2.25773886e-01 -9.13779289e-02
1.08657099e-01 -8.56099010e-01 1.54054686e-01 -3.12472075e-01
7.15980589e-01 -4.69194144e-01 -2.92232275e-01 7.72123933e-02
-5.35363674e-01 -2.68386453e-01 5.20548634e-02 -3.21290582e-01
-1.04709435e+00 4.54193950e-01 -3.96411777e-01 -7.22104460e-02
-6.71172500e-01 1.70105547e-01 -3.42932761e-01 -1.72104597e-01
5.95543146e-01 1.75899997e-01 3.88155252e-01 -3.98837149e-01
6.45056441e-02 1.24369264e+00 5.30770421e-01 -2.34541833e-01
3.26772571e-01 4.38616574e-01 2.12579016e-02 -9.77805018e-01
-1.67524189e-01 -4.86470044e-01 -9.41902101e-01 -1.11443818e+00
8.38615656e-01 -7.71492243e-01 -1.26282752e+00 7.74950862e-01
-1.02653778e+00 3.98625582e-02 1.64034486e-01 2.46723697e-01
-6.18907094e-01 2.76781350e-01 -2.07719896e-02 -1.74676275e+00
-5.80751121e-01 -6.57328665e-01 1.20168710e+00 3.95826072e-01
-8.41659188e-01 -5.23588836e-01 1.99534535e-01 4.04249489e-01
3.91288131e-01 8.89057934e-01 1.93195835e-01 -9.08593833e-02
3.51520367e-02 -5.54024875e-01 2.12003559e-01 -8.80096555e-02
1.68764349e-02 1.73438430e-01 -1.39810514e+00 2.33535782e-01
2.80155450e-01 -4.38392729e-01 -2.14228913e-01 1.16324546e-02
7.20594525e-01 -1.60710618e-01 -2.04716367e-03 3.67912531e-01
1.08866525e+00 2.93995202e-01 9.77489233e-01 4.17876601e-01
4.15519416e-01 9.36645031e-01 3.10370892e-01 9.97720301e-01
4.85319436e-01 9.14222360e-01 4.80856091e-01 1.43701034e-02
2.84631103e-01 -2.46969480e-02 5.96559286e-01 2.99429476e-01
-6.20972633e-01 3.37461144e-01 -8.09624374e-01 1.99087095e-02
-1.43959296e+00 -8.63101602e-01 -8.86137933e-02 2.39528561e+00
6.00668609e-01 -8.74536186e-02 6.50870621e-01 5.82989991e-01
4.84171242e-01 8.19127262e-02 -3.30260634e-01 -9.80360389e-01
2.92584568e-01 3.86773378e-01 1.69211980e-02 2.20684126e-01
-7.66305566e-01 5.26348174e-01 6.17950821e+00 8.47224444e-02
-1.57294989e+00 8.57192185e-03 3.00323993e-01 -3.08215380e-01
5.69745451e-02 -5.14695287e-01 -5.54599285e-01 4.39621121e-01
9.91697669e-01 2.05671728e-01 2.30025560e-01 9.69205678e-01
5.75083971e-01 -6.12310469e-01 -6.01180851e-01 1.24410760e+00
2.14207292e-01 -2.83893913e-01 -4.73442376e-01 2.49957461e-02
-2.19619587e-01 -1.03221953e-01 -2.65670300e-01 1.32650793e-01
-4.14538622e-01 -1.05113673e+00 7.22593248e-01 9.89680290e-01
8.68637085e-01 -5.37588060e-01 3.56847197e-01 3.90125990e-01
-9.44187462e-01 1.19162790e-01 3.50233495e-01 -6.81530356e-01
8.07003155e-02 2.51677454e-01 -8.81232679e-01 1.00407019e-01
9.65737581e-01 2.71704346e-02 -4.05039519e-01 6.54945076e-01
-9.99958739e-02 1.29208609e-01 -6.04166269e-01 -3.13414335e-01
-4.17618126e-01 -6.97934404e-02 3.65401030e-01 1.23794103e+00
1.91822037e-01 -4.79871519e-02 -3.56786191e-01 5.21718323e-01
4.91998494e-01 2.85664320e-01 -8.06759238e-01 2.46352881e-01
1.36313707e-01 1.48712730e+00 -6.83422089e-01 2.19841942e-01
-2.58602917e-01 1.06963396e+00 1.51298881e-01 1.52362525e-01
-7.05146730e-01 -1.65163606e-01 7.03474760e-01 3.80612612e-01
-1.51693225e-01 -3.64035130e-01 -2.97184438e-01 -7.75383830e-01
3.92854542e-01 -8.31308842e-01 1.04178324e-01 -1.02855301e+00
-4.98809725e-01 6.00106299e-01 4.49059755e-02 -1.08614802e+00
-8.12037885e-01 -6.29305065e-01 -4.14260089e-01 9.38786328e-01
-8.98918331e-01 -1.13307118e+00 -8.53417039e-01 7.16067135e-01
1.80835918e-01 2.57515520e-01 1.23293149e+00 1.42140910e-01
-6.16414070e-01 3.56465369e-01 -7.53390014e-01 -1.71688676e-01
2.86901146e-01 -9.82050717e-01 -1.54363755e-02 5.05918980e-01
-2.30274409e-01 4.48876709e-01 8.79101932e-01 -3.49476904e-01
-1.41831017e+00 -2.35816419e-01 9.70604002e-01 -6.31175816e-01
1.59003764e-01 -6.54688835e-01 -4.32826072e-01 3.96735251e-01
-8.90293047e-02 1.47500066e-02 1.00330353e+00 -5.59517443e-02
-2.22182542e-01 -3.19905318e-02 -1.46640873e+00 5.16195059e-01
1.01857626e+00 -7.15681314e-01 -4.50955927e-01 -3.01328391e-01
-2.99606055e-01 -4.81812656e-01 -7.96355009e-01 4.28452581e-01
1.18480575e+00 -1.39811718e+00 6.85506761e-01 -2.82056421e-01
-5.75166494e-02 -1.78551316e-01 -3.72344963e-02 -6.42341435e-01
3.08373779e-01 -8.92544329e-01 -3.19482625e-01 1.14590228e+00
1.56239076e-02 -5.32925963e-01 1.03018463e+00 1.32086539e+00
3.73379648e-01 -6.41678631e-01 -1.19708002e+00 -1.78061679e-01
-7.72389114e-01 -6.46110594e-01 5.22464097e-01 4.94201958e-01
5.12553036e-01 -3.05070705e-03 -3.07315737e-01 -1.29736528e-01
1.96036503e-01 -3.66423815e-01 1.10549891e+00 -1.39469326e+00
1.46685183e-01 -1.57159135e-01 -1.03787518e+00 -3.39499086e-01
4.59523546e-03 -2.19924495e-01 -3.12596709e-02 -1.00470841e+00
-8.11885223e-02 1.83394969e-01 4.37723361e-02 4.30423111e-01
6.79109320e-02 4.60205227e-01 8.47968012e-02 -5.62619492e-02
-4.36493814e-01 4.61646646e-01 6.96712852e-01 5.31852782e-01
-2.30621979e-01 2.41018683e-01 -4.11434621e-01 8.31040621e-01
5.26811898e-01 -1.54664725e-01 -2.05581367e-01 2.02704489e-01
4.30452913e-01 3.58099565e-02 6.49491847e-01 -1.16665471e+00
9.55846831e-02 -1.22645317e-04 6.42043471e-01 -2.82470107e-01
6.65534496e-01 -1.00471711e+00 3.49988520e-01 3.15760970e-01
9.53073278e-02 2.21705511e-01 8.84968042e-02 9.23092961e-02
-1.17551118e-01 -6.71728924e-02 4.30308163e-01 -2.35502169e-01
-8.14460397e-01 -1.61636695e-01 -7.23404109e-01 -4.19456899e-01
1.22862756e+00 -9.44630623e-01 1.64616197e-01 -5.09101212e-01
-1.08134031e+00 -3.70659411e-01 4.44618523e-01 4.26372051e-01
5.76104939e-01 -1.03254867e+00 -1.92790642e-01 6.85768306e-01
3.23844671e-01 -3.05483192e-01 3.43923271e-01 9.23840046e-01
-5.28427303e-01 9.21390485e-03 -6.21850431e-01 -4.25479770e-01
-1.72829509e+00 -6.48155063e-02 4.48347956e-01 1.23539738e-01
-1.93099931e-01 6.82499588e-01 -6.56292617e-01 -3.27778906e-01
4.00756687e-01 6.41777217e-02 -5.97452223e-01 4.12030131e-01
6.86568737e-01 6.30194068e-01 4.52907234e-01 -9.54388916e-01
-6.73283994e-01 4.35322583e-01 5.56601286e-01 -3.18208307e-01
1.30247211e+00 -4.56311256e-01 -4.45329864e-03 5.92096269e-01
1.08122444e+00 1.93864301e-01 -1.24441183e+00 5.89440644e-01
2.40235999e-01 -4.06789184e-01 -4.62394133e-02 -6.39671862e-01
-5.77326238e-01 6.85628235e-01 1.14849448e+00 1.83495387e-01
1.45072722e+00 -2.53825665e-01 1.86607897e-01 -2.91758515e-02
7.00110555e-01 -1.18912029e+00 -3.83312136e-01 1.11946814e-01
9.66370106e-01 -1.12902534e+00 -5.36068401e-04 -2.58313298e-01
-6.34632647e-01 1.12214649e+00 4.13370848e-01 3.53633642e-01
5.71206510e-01 4.95409042e-01 3.50450367e-01 -4.88566786e-01
-4.10938948e-01 -4.59208757e-01 1.70578539e-01 9.65628088e-01
7.02891827e-01 1.31102189e-01 -4.31990445e-01 4.89522159e-01
-3.34731609e-01 6.17595434e-01 1.75820515e-01 1.25130236e+00
-3.96008879e-01 -8.13218653e-01 -7.25818992e-01 1.43192887e-01
-5.19658089e-01 4.28098053e-01 -5.99489152e-01 6.43228590e-01
1.62135959e-01 9.16528285e-01 -5.43167815e-02 -3.58848661e-01
4.38395917e-01 5.12432694e-01 5.54271936e-01 -3.42393130e-01
-8.17468226e-01 -2.34190628e-01 7.55122155e-02 -8.57088864e-01
-7.24857867e-01 -7.32783437e-01 -8.21690679e-01 -8.94161090e-02
1.04312465e-01 -2.71089673e-01 1.28280604e+00 9.66248155e-01
3.18535537e-01 -8.05099085e-02 6.72243953e-01 -1.47263551e+00
-1.40930697e-01 -1.07640493e+00 -6.27154171e-01 4.77228642e-01
1.40349597e-01 -8.31175685e-01 -4.32451397e-01 2.73991115e-02] | [13.496696472167969, 2.2068915367126465] |
c7a8d0ef-da44-4467-b6c3-b6821b8ed16b | a-unified-objective-for-novel-class-discovery | 2108.08536 | null | https://arxiv.org/abs/2108.08536v4 | https://arxiv.org/pdf/2108.08536v4.pdf | A Unified Objective for Novel Class Discovery | In this paper, we study the problem of Novel Class Discovery (NCD). NCD aims at inferring novel object categories in an unlabeled set by leveraging from prior knowledge of a labeled set containing different, but related classes. Existing approaches tackle this problem by considering multiple objective functions, usually involving specialized loss terms for the labeled and the unlabeled samples respectively, and often requiring auxiliary regularization terms. In this paper, we depart from this traditional scheme and introduce a UNified Objective function (UNO) for discovering novel classes, with the explicit purpose of favoring synergy between supervised and unsupervised learning. Using a multi-view self-labeling strategy, we generate pseudo-labels that can be treated homogeneously with ground truth labels. This leads to a single classification objective operating on both known and unknown classes. Despite its simplicity, UNO outperforms the state of the art by a significant margin on several benchmarks (~+10% on CIFAR-100 and +8% on ImageNet). The project page is available at: https://ncd-uno.github.io. | ['Elisa Ricci', 'Moin Nabi', 'Zhun Zhong', 'Stéphane Lathuilière', 'Enver Sangineto', 'Enrico Fini'] | 2021-08-19 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Fini_A_Unified_Objective_for_Novel_Class_Discovery_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Fini_A_Unified_Objective_for_Novel_Class_Discovery_ICCV_2021_paper.pdf | iccv-2021-1 | ['novel-class-discovery', 'novel-class-discovery'] | ['computer-vision', 'methodology'] | [ 3.33966941e-01 3.05299819e-01 -3.70085865e-01 -6.61570907e-01
-9.90629971e-01 -6.13395751e-01 6.65937304e-01 3.34961444e-01
-4.34854925e-01 9.72906768e-01 -1.05296299e-01 6.17502369e-02
1.12241790e-01 -6.14357352e-01 -7.21841037e-01 -6.90469623e-01
9.27317813e-02 6.38960600e-01 1.45562366e-02 3.69807720e-01
7.19841570e-02 3.93678755e-01 -1.64772296e+00 3.50977361e-01
7.51138449e-01 1.02128029e+00 -7.19294101e-02 2.46274635e-01
9.48631391e-02 5.86987078e-01 -3.19026142e-01 -5.91364801e-01
3.62605482e-01 -4.15724188e-01 -1.07828844e+00 3.33983034e-01
6.71722889e-01 1.03293486e-01 5.03087156e-02 1.30920160e+00
4.35913712e-01 2.17675462e-01 8.10077667e-01 -1.24743760e+00
-5.67565918e-01 7.46719658e-01 -6.44245327e-01 1.51328534e-01
-2.14870628e-02 5.58104506e-03 1.25189459e+00 -1.12939227e+00
6.68470085e-01 1.09935176e+00 5.23456991e-01 7.84029186e-01
-1.72098911e+00 -8.17317903e-01 4.21362162e-01 1.57066360e-01
-1.61069560e+00 -5.37666857e-01 8.11495304e-01 -6.20070159e-01
3.87993962e-01 1.42782301e-01 1.04981862e-01 1.27220154e+00
-2.59079546e-01 1.03760254e+00 1.31715143e+00 -3.87790680e-01
3.84876251e-01 5.25294125e-01 4.97055560e-01 4.62244481e-01
3.03509355e-01 7.26502761e-02 -3.95349056e-01 -5.96869551e-03
1.90709159e-01 -3.56626324e-02 -3.22361082e-01 -6.48020685e-01
-1.19474483e+00 8.28401864e-01 3.37606579e-01 3.02543908e-01
-1.83903500e-01 -2.40824193e-01 3.77141297e-01 2.25022957e-01
8.48635733e-01 5.29662549e-01 -6.51329577e-01 3.34044367e-01
-7.74616599e-01 1.51942328e-01 5.09133160e-01 9.16370809e-01
1.05977869e+00 -1.34372219e-01 -1.97715729e-01 7.37618625e-01
2.05139369e-01 2.90279657e-01 5.39326847e-01 -7.61056006e-01
2.52789021e-01 5.72903693e-01 7.53408074e-02 -7.53355980e-01
-3.77867132e-01 -9.92621839e-01 -7.14470863e-01 1.59413651e-01
3.86075884e-01 -1.83707431e-01 -1.03420568e+00 1.89982963e+00
5.02126575e-01 3.69791657e-01 3.02368522e-01 7.05586612e-01
8.26018512e-01 4.74928409e-01 2.38896441e-02 -3.50432575e-01
9.32843268e-01 -1.21488154e+00 -5.72849810e-01 -2.73195684e-01
6.22161984e-01 -5.41084766e-01 1.09236896e+00 4.79572922e-01
-7.56150484e-01 -5.13958573e-01 -8.67028236e-01 1.46579012e-01
-3.86915863e-01 2.86501765e-01 5.11270881e-01 6.27679706e-01
-8.24313343e-01 3.99941206e-01 -6.82086885e-01 -2.56229192e-01
6.95568562e-01 2.24331990e-01 -3.87802780e-01 -1.65055186e-01
-6.83388472e-01 6.13507092e-01 7.75087059e-01 -1.03025280e-01
-1.15832555e+00 -7.32355833e-01 -6.74809039e-01 -2.41377912e-02
7.17423260e-01 -4.63905782e-01 1.24655080e+00 -1.16716397e+00
-1.25534332e+00 1.04570234e+00 -6.98406324e-02 -4.49511290e-01
5.52322149e-01 -1.62771776e-01 -4.61390048e-01 1.08702123e-01
4.60459471e-01 7.82067478e-01 8.73158872e-01 -1.67119944e+00
-7.55800307e-01 -4.42621589e-01 1.34767845e-01 4.68432456e-02
-2.75048286e-01 -3.83730590e-01 -1.79175988e-01 -8.07482064e-01
1.71856031e-01 -9.50805783e-01 -1.64996609e-01 -1.39131084e-01
-6.44597530e-01 -4.20495689e-01 8.07894826e-01 -1.79200783e-01
8.49645019e-01 -2.18218851e+00 2.03753605e-01 4.61268350e-02
4.79850352e-01 2.02232018e-01 3.52947079e-02 -1.07208602e-01
-2.22341359e-01 7.93313957e-04 -5.26741982e-01 -5.02892494e-01
-6.50700107e-02 8.67194086e-02 -3.14930409e-01 5.30198038e-01
3.27348083e-01 7.60485053e-01 -1.15666652e+00 -2.17457563e-01
-2.51823105e-02 1.90500826e-01 -4.77327585e-01 4.64178510e-02
-4.00065601e-01 7.70069063e-01 -2.98075169e-01 6.87843025e-01
7.24617302e-01 -5.94058990e-01 1.54773563e-01 -1.38548762e-01
1.58279389e-01 7.27731560e-04 -1.26918781e+00 1.72003901e+00
-3.21561396e-01 4.69757944e-01 -2.04578936e-01 -1.42083991e+00
9.14555848e-01 2.44534746e-01 3.26294094e-01 -2.48977020e-01
1.89347520e-01 3.24051976e-01 -2.11606115e-01 -2.79053658e-01
-1.44880905e-03 -4.52383012e-01 1.17556028e-01 4.44684535e-01
5.36372781e-01 3.66443545e-01 1.78860635e-01 4.72796671e-02
7.75724590e-01 2.23310478e-02 2.12667555e-01 -3.55471253e-01
5.12790799e-01 -3.28267217e-02 7.71602809e-01 8.34524095e-01
-1.07725024e-01 5.04233837e-01 2.84962386e-01 -3.29282254e-01
-7.24385560e-01 -1.26828408e+00 -4.02755886e-01 8.99084449e-01
1.81481510e-01 -1.95467234e-01 -4.08141941e-01 -1.26508307e+00
-3.06313038e-02 9.32125151e-01 -7.34969616e-01 -2.36264274e-01
-1.19182274e-01 -8.71952116e-01 3.22368413e-01 3.59766036e-01
4.15296108e-01 -8.04890454e-01 -2.31354311e-01 1.02372184e-01
-1.87237024e-01 -1.15724623e+00 -2.21502796e-01 4.08729076e-01
-9.58368599e-01 -1.05303383e+00 -6.25578284e-01 -8.29089642e-01
9.33170795e-01 2.36925572e-01 1.06826901e+00 -4.29677337e-01
-2.32428849e-01 4.03609306e-01 -5.23850024e-01 -4.02613103e-01
-1.85182929e-01 1.68499097e-01 1.81549385e-01 6.45000875e-01
3.78391206e-01 -6.65518105e-01 -3.18202496e-01 2.74284333e-01
-9.70644176e-01 1.35658056e-01 5.31839669e-01 9.55588043e-01
8.09781909e-01 7.85348192e-02 9.84168589e-01 -1.37097490e+00
-2.11035702e-02 -8.11499655e-01 -6.20082080e-01 2.99161196e-01
-9.65260863e-01 4.22012769e-02 6.15173459e-01 -5.57568133e-01
-1.21444941e+00 2.14596018e-01 2.04304263e-01 -6.49754107e-01
-4.32975680e-01 4.71637249e-01 -3.74788374e-01 8.77945796e-02
7.51087248e-01 2.16921672e-01 -2.30500132e-01 -7.48689115e-01
4.98520166e-01 5.32966197e-01 4.32941526e-01 -5.85494697e-01
8.13786566e-01 7.34380245e-01 -2.30428994e-01 -4.63699073e-01
-1.48603034e+00 -5.89228988e-01 -8.49801481e-01 -1.37395158e-01
6.01584494e-01 -9.87229228e-01 -2.28806004e-01 3.43819290e-01
-8.26678395e-01 -1.11099645e-01 -6.50765479e-01 7.86282122e-01
-3.79760593e-01 2.44630426e-01 -2.05806836e-01 -5.62632501e-01
-6.70605674e-02 -9.28882301e-01 9.82252657e-01 1.98583275e-01
1.45555623e-02 -9.80069816e-01 -6.22928282e-03 4.61279422e-01
-1.06591908e-02 3.51111382e-01 8.15578461e-01 -1.03860366e+00
-4.69001621e-01 -1.86046198e-01 -3.15892309e-01 6.17977440e-01
2.76337296e-01 -4.35656935e-01 -1.40123153e+00 -4.76506084e-01
5.39622903e-02 -5.19503117e-01 1.13973188e+00 2.15664983e-01
1.25297022e+00 -3.40120494e-01 -4.73901033e-01 4.66493875e-01
1.50955176e+00 9.84342918e-02 1.50298968e-01 1.00621827e-01
7.18407631e-01 6.87180340e-01 6.18888438e-01 3.45496505e-01
2.62176871e-01 5.49234927e-01 3.38270694e-01 9.54461321e-02
-1.36805460e-01 -1.30126953e-01 8.51839408e-02 5.15459538e-01
1.58109814e-01 -3.46637160e-01 -9.68518913e-01 7.41539836e-01
-1.88954282e+00 -7.71479547e-01 8.22271183e-02 2.29764080e+00
9.05092359e-01 2.60285854e-01 6.67464063e-02 1.83898062e-02
9.94057238e-01 -6.91547319e-02 -8.23071837e-01 1.61288172e-01
-2.02029869e-01 1.89185068e-01 3.00629050e-01 4.57606673e-01
-1.36197722e+00 7.98143208e-01 5.44026232e+00 9.34863269e-01
-1.17459285e+00 2.93866903e-01 8.69740009e-01 -1.42378986e-01
-2.56008655e-01 1.82283148e-02 -1.00536668e+00 3.23310912e-01
7.83637583e-01 -3.31261933e-01 1.58760980e-01 8.29137266e-01
-1.03427134e-01 7.55236819e-02 -1.26702356e+00 9.55195010e-01
1.79790944e-01 -1.17640686e+00 2.43281454e-01 -1.34721324e-01
1.05451715e+00 1.33692339e-01 1.87097356e-01 3.30810338e-01
1.88488111e-01 -7.14649498e-01 7.11477041e-01 4.15660709e-01
7.83181071e-01 -6.36313438e-01 6.35283053e-01 3.74945045e-01
-8.33470404e-01 1.22838467e-02 -1.47597432e-01 3.77221145e-02
-2.07455561e-01 8.20218861e-01 -7.11179972e-01 6.70283794e-01
8.01491499e-01 1.09439123e+00 -6.83406234e-01 1.04077673e+00
-2.99659729e-01 8.48485589e-01 -1.67275637e-01 3.53237361e-01
3.14187914e-01 -8.10278207e-02 6.47025764e-01 9.54326689e-01
-5.11571988e-02 -7.26904348e-02 4.66501325e-01 9.43001807e-01
-4.78919148e-01 9.80687737e-02 -4.83725309e-01 1.46642402e-01
2.03254744e-01 1.20870125e+00 -9.40472543e-01 -4.81500447e-01
-4.42541152e-01 9.25204754e-01 4.58273381e-01 4.52178270e-01
-7.60223031e-01 -2.50985384e-01 2.23334342e-01 -8.87316745e-03
3.80224586e-01 2.26752684e-01 -1.97416425e-01 -1.51757562e+00
1.21356487e-01 -6.39049888e-01 5.08559883e-01 -3.41301560e-01
-1.60746539e+00 6.80698037e-01 1.47281602e-01 -1.46201289e+00
1.02733813e-01 -4.47323948e-01 -2.16802567e-01 6.75847411e-01
-1.70369601e+00 -1.05959141e+00 -1.57301307e-01 3.99033904e-01
6.78020835e-01 -2.53983706e-01 8.07775319e-01 5.44671416e-01
-6.13488257e-01 7.75759757e-01 4.19513017e-01 -1.52645502e-02
7.85064995e-01 -1.37909544e+00 8.85156691e-02 8.60212147e-01
2.94543177e-01 2.22118601e-01 5.00210524e-01 -5.02712965e-01
-6.70230567e-01 -1.56541169e+00 9.37602520e-01 -4.83927995e-01
5.72723150e-01 -4.50085372e-01 -9.89804089e-01 8.66607368e-01
-1.93557233e-01 3.53572547e-01 9.44031715e-01 1.66852817e-01
-7.18282878e-01 -1.16242908e-01 -1.26582062e+00 3.23766947e-01
1.06496203e+00 -3.70946705e-01 -4.90403831e-01 5.68394542e-01
8.46450388e-01 -1.19990334e-01 -7.32952952e-01 6.01389408e-01
2.56682128e-01 -7.30311394e-01 8.12669814e-01 -9.26216006e-01
3.50420594e-01 -4.72934663e-01 -2.64611006e-01 -1.16641152e+00
-3.76524240e-01 -2.96427429e-01 -1.52410686e-01 1.36555076e+00
6.92732751e-01 -7.71654367e-01 8.87253106e-01 2.16513857e-01
-1.41306177e-01 -7.87912428e-01 -8.80371153e-01 -1.06657970e+00
-1.71083491e-02 -2.35883817e-01 1.62829265e-01 1.39571750e+00
-3.66879851e-01 7.94610202e-01 -3.82690907e-01 3.85488272e-01
1.00248480e+00 3.24973792e-01 5.16963005e-01 -1.41162443e+00
-2.86754131e-01 -3.38354886e-01 -3.72306138e-01 -8.21428776e-01
5.23874640e-01 -1.32821512e+00 -1.03938859e-02 -1.03926575e+00
3.58918339e-01 -5.07172406e-01 -6.17566824e-01 6.98585629e-01
-2.53504604e-01 4.99798596e-01 5.32726347e-02 3.73045683e-01
-8.69034529e-01 7.68047392e-01 9.77689147e-01 -3.98788661e-01
-2.18403101e-01 1.90193191e-01 -8.98151994e-01 8.67356300e-01
1.02329195e+00 -8.15926075e-01 -4.61038500e-01 -4.92705673e-01
-2.08125755e-01 -4.38195467e-01 5.01739979e-01 -9.95464802e-01
4.51658107e-02 1.72991082e-02 2.12515846e-01 -4.81390059e-01
1.87155962e-01 -8.01125407e-01 4.45652269e-02 3.47704530e-01
-5.84114075e-01 -5.00720918e-01 1.51701808e-01 9.34406579e-01
-3.20011079e-01 -3.32387447e-01 1.05416477e+00 -6.71120211e-02
-8.26460242e-01 3.27349365e-01 -7.81170726e-02 2.69654244e-01
1.15105450e+00 -3.31790559e-02 -2.60172397e-01 -5.16507179e-02
-1.11251462e+00 2.94104844e-01 3.03229004e-01 4.44209665e-01
4.75604296e-01 -1.36959672e+00 -8.40095401e-01 -3.33231278e-02
5.67354143e-01 1.77431136e-01 1.62822336e-01 6.74716890e-01
-4.11402807e-02 2.61138022e-01 6.91229403e-02 -7.56074131e-01
-9.82867897e-01 7.52773821e-01 2.82698452e-01 -2.01207832e-01
-4.44698751e-01 1.05978334e+00 4.92226332e-01 -7.64782727e-01
3.35062206e-01 3.56945731e-02 -2.70703971e-01 1.67035177e-01
3.09986264e-01 4.14448649e-01 4.64589772e-04 -6.22076154e-01
-3.14470679e-01 2.89513409e-01 -4.19182032e-01 2.95841277e-01
1.34553254e+00 -2.52960771e-01 6.49609556e-03 8.73962045e-01
1.45556235e+00 -8.60083550e-02 -1.18473339e+00 -7.59802639e-01
2.06536904e-01 -4.47897673e-01 7.28812441e-02 -1.05530536e+00
-1.08197284e+00 6.17937028e-01 8.90464783e-01 -1.11899830e-01
1.15321863e+00 2.89061576e-01 4.86664712e-01 3.80920917e-01
3.14304620e-01 -7.97741413e-01 1.08580954e-01 3.14535290e-01
6.58168375e-01 -1.65359128e+00 -7.58558512e-02 -5.98567545e-01
-4.76876378e-01 8.77301574e-01 6.92096651e-01 -3.62403169e-02
7.07330883e-01 -3.02520454e-01 4.85805497e-02 -1.76923841e-01
-5.49211025e-01 -3.21001858e-01 4.49281454e-01 4.81292546e-01
2.47974679e-01 2.15877712e-01 -3.14881891e-01 6.35730028e-01
2.10061848e-01 -1.17552377e-01 3.65074635e-01 8.28901649e-01
-3.09365511e-01 -1.20228601e+00 -1.63144320e-01 6.77942753e-01
-5.31687081e-01 -1.23336099e-01 -2.51466841e-01 5.40689230e-01
3.48449916e-01 8.83479476e-01 -4.75594439e-02 -3.12385291e-01
2.04266250e-01 5.71379364e-02 2.80054003e-01 -1.14081943e+00
-2.07845390e-01 8.03834647e-02 -4.59741056e-02 -2.54602909e-01
-7.60350466e-01 -7.42600203e-01 -8.88241768e-01 1.61050931e-01
-5.69353998e-01 4.17858064e-01 2.57857621e-01 8.14762056e-01
3.53218168e-01 3.47036868e-01 9.32635367e-01 -6.68910503e-01
-5.20147502e-01 -7.79592335e-01 -7.50610232e-01 4.31182444e-01
3.83148044e-01 -8.70222330e-01 -4.94073927e-01 2.80878037e-01] | [9.584932327270508, 3.1046884059906006] |
ea62985a-29ca-4155-bb0a-31c012b6aa52 | 1st-place-solutions-for-waymo-open-dataset | 2006.15506 | null | https://arxiv.org/abs/2006.15506v1 | https://arxiv.org/pdf/2006.15506v1.pdf | 1st Place Solutions for Waymo Open Dataset Challenges -- 2D and 3D Tracking | This technical report presents the online and real-time 2D and 3D multi-object tracking (MOT) algorithms that reached the 1st places on both Waymo Open Dataset 2D tracking and 3D tracking challenges. An efficient and pragmatic online tracking-by-detection framework named HorizonMOT is proposed for camera-based 2D tracking in the image space and LiDAR-based 3D tracking in the 3D world space. Within the tracking-by-detection paradigm, our trackers leverage our high-performing detectors used in the 2D/3D detection challenges and achieved 45.13% 2D MOTA/L2 and 63.45% 3D MOTA/L2 in the 2D/3D tracking challenges. | ['Yihan Hu', 'Zhuangzhuang Ding', 'Yu Wang', 'Sijia Chen', 'Runzhou Ge', 'Li Huang', 'Jie Liao'] | 2020-06-28 | null | null | null | null | ['3d-multi-object-tracking'] | ['computer-vision'] | [-6.03225946e-01 -4.28744227e-01 -6.44211173e-02 3.04103315e-01
-8.54732096e-01 -1.05883467e+00 6.42141402e-01 -3.06763332e-02
-4.42356139e-01 2.17632651e-01 -3.85476589e-01 -2.06545919e-01
-1.00650325e-01 -4.63806242e-01 -5.52370965e-01 -2.15367705e-01
-2.49976501e-01 8.95655870e-01 1.12316871e+00 -1.47424415e-01
-2.02778071e-01 9.83816803e-01 -1.78709400e+00 -5.14644802e-01
1.46232722e-02 1.10524356e+00 1.28220469e-01 1.26414037e+00
-1.09031983e-01 1.09648749e-01 -4.74323183e-01 -6.46856725e-02
1.02421165e+00 1.87261492e-01 1.10249475e-01 -1.58049941e-01
1.48196840e+00 -4.40347582e-01 -2.78515339e-01 8.03647757e-01
9.32951450e-01 -4.86958921e-02 3.86648417e-01 -1.67970908e+00
4.76159751e-02 -5.06527573e-02 -8.63293469e-01 5.41229546e-01
6.69901788e-01 5.40868998e-01 5.96711814e-01 -1.02133131e+00
7.83927441e-01 1.67402112e+00 1.14042354e+00 5.97262859e-01
-7.85965085e-01 -1.00311410e+00 -2.75931582e-02 -2.99698830e-01
-1.80791998e+00 -2.54007012e-01 2.56446093e-01 -8.68839502e-01
9.78237748e-01 2.85868824e-01 9.73786414e-01 7.12029219e-01
2.80395299e-01 7.21619070e-01 9.41394150e-01 -6.45955727e-02
-2.51980752e-01 3.59062217e-02 1.09096095e-01 7.37361372e-01
8.07761669e-01 8.53723764e-01 -6.04786694e-01 -5.69074266e-02
7.55637765e-01 6.07314408e-02 3.89822155e-01 -6.79256022e-01
-1.26994085e+00 4.06683266e-01 4.67531323e-01 -1.30194917e-01
5.09224832e-02 3.76899630e-01 1.32061765e-01 1.01226866e-01
4.18502778e-01 -1.74287051e-01 -2.07747787e-01 -1.71778351e-01
-8.83370519e-01 4.56078023e-01 1.25875026e-01 1.84882283e+00
3.94097626e-01 1.13215782e-01 -2.81271815e-01 3.96080054e-02
9.55039620e-01 1.43183696e+00 -2.74050742e-01 -7.31936216e-01
3.81384134e-01 7.51576304e-01 7.47402370e-01 -5.81583440e-01
-7.58816123e-01 -3.69845957e-01 2.99051963e-02 8.15322042e-01
4.39772248e-01 -1.37461856e-01 -8.47297251e-01 1.08838499e+00
1.38451374e+00 5.53348735e-02 -3.27039391e-01 1.15211987e+00
9.59838450e-01 7.04010352e-02 -1.45780340e-01 1.39654040e-01
1.55822003e+00 -7.71368027e-01 -6.49079084e-01 -1.93303093e-01
6.30841792e-01 -9.89404857e-01 1.02075070e-01 -1.22327894e-01
-1.04004550e+00 -8.28568101e-01 -9.25470233e-01 1.23438679e-01
-4.62602586e-01 4.25789952e-02 1.15185365e-01 1.05164111e+00
-1.06227648e+00 -1.31288022e-01 -8.61920476e-01 -7.09631681e-01
6.25784457e-01 4.74720240e-01 -9.65078995e-02 3.35809141e-02
-5.67703187e-01 1.22226584e+00 4.82366383e-01 -2.19624683e-01
-1.35461938e+00 -9.36204851e-01 -5.63740492e-01 -5.35730958e-01
5.62245667e-01 -7.43513882e-01 1.16669798e+00 5.60323954e-01
-1.02910078e+00 1.37128913e+00 -2.34577898e-02 -5.81189156e-01
1.07383776e+00 -7.44277120e-01 -6.49265051e-01 -2.13126510e-01
2.90910214e-01 7.61463583e-01 7.01770425e-01 -9.95848477e-01
-1.31999242e+00 -6.46750152e-01 -2.95077890e-01 1.94679558e-01
3.86219174e-01 2.01933578e-01 -5.80765963e-01 -1.01903796e-01
1.76790103e-01 -1.12297106e+00 -2.90209576e-02 8.47649455e-01
-3.83881629e-01 -2.72274584e-01 1.67487490e+00 7.74564827e-03
7.44804561e-01 -1.98986495e+00 -2.71697551e-01 -4.27210838e-01
5.46269596e-01 4.79736328e-01 2.18181640e-01 2.44708434e-01
6.83821917e-01 -2.79170066e-01 7.76331842e-01 -7.31414258e-01
4.28674459e-01 -1.68337330e-01 -9.01599526e-02 8.88604701e-01
-2.05252111e-01 9.65277255e-01 -1.03779602e+00 -7.86671817e-01
7.76168525e-01 4.04649109e-01 -1.68601722e-01 1.00026697e-01
-4.91214283e-02 5.20296395e-01 -6.20337605e-01 1.28866220e+00
1.15744865e+00 3.82435285e-02 -3.53409618e-01 -2.05902725e-01
-1.06651056e+00 1.62427165e-02 -1.54227841e+00 1.51815963e+00
3.68133485e-02 7.15248227e-01 4.33931679e-01 2.36061603e-01
8.28149199e-01 1.17789477e-01 8.59533846e-01 -4.99519587e-01
2.28985518e-01 3.59625697e-01 -4.54848439e-01 3.45437266e-02
8.59098732e-01 9.42265838e-02 -2.25206390e-01 7.83596262e-02
7.93791842e-03 -3.19593191e-01 8.81103128e-02 1.71637118e-01
1.21607840e+00 4.51644540e-01 1.74523517e-01 -2.44109750e-01
2.91948020e-01 5.12566686e-01 4.76835489e-01 1.24394059e+00
-8.61900568e-01 2.63796329e-01 -8.04309130e-01 -5.47513306e-01
-8.94132674e-01 -1.50278640e+00 -4.42513674e-01 8.10065746e-01
7.28892326e-01 -3.44592005e-01 3.39613855e-02 -6.66312337e-01
8.57699156e-01 1.36487752e-01 -2.54321307e-01 2.93644100e-01
-4.78822112e-01 -1.94203734e-01 7.28143692e-01 3.25302005e-01
4.07236993e-01 -7.30218068e-02 -1.17583060e+00 3.35051268e-01
4.54501987e-01 -1.51820076e+00 -6.04025424e-01 9.14217085e-02
-8.05446506e-01 -1.29359770e+00 -5.17644942e-01 -2.17958972e-01
1.23646021e-01 9.60265696e-01 8.19021881e-01 -1.70379952e-01
-6.03064775e-01 9.67309415e-01 -1.41575992e-01 -9.56737995e-01
-1.18372075e-01 -4.26321626e-01 8.29391479e-01 -4.92913246e-01
4.82208669e-01 3.93327065e-02 -3.48960370e-01 9.55213249e-01
1.24353826e-01 -1.86671495e-01 8.56756866e-02 -1.97372492e-02
7.21141458e-01 -4.20613408e-01 -1.67919859e-01 5.87212071e-02
-2.69138157e-01 -2.20885798e-01 -1.79519939e+00 4.58016507e-02
-3.30816418e-01 -5.82867265e-01 -2.39408687e-01 -6.10177577e-01
-3.93422663e-01 6.05124056e-01 2.34773844e-01 -1.23910487e+00
-7.02554807e-02 -6.50672078e-01 7.68656582e-02 -7.74371207e-01
6.82752669e-01 -4.13157120e-02 -3.36156368e-01 -7.67232597e-01
3.63659680e-01 5.31678557e-01 5.37356436e-01 -1.97782129e-01
1.58885443e+00 1.04248214e+00 3.97705555e-01 -6.16075873e-01
-9.69044387e-01 -1.03001523e+00 -9.58293378e-01 -9.59829152e-01
1.13692403e+00 -1.59198833e+00 -1.09027421e+00 4.15950000e-01
-1.00094855e+00 -4.10027383e-03 -5.68902910e-01 6.37619913e-01
-3.62242848e-01 1.36423320e-01 1.63973123e-01 -1.62813687e+00
-3.50731194e-01 -9.26448166e-01 1.70900273e+00 4.22971576e-01
1.96317181e-01 -7.09997952e-01 3.63010943e-01 1.16127253e-01
3.95488054e-01 5.04328549e-01 -5.77858150e-01 -3.44279051e-01
-1.34537733e+00 -6.49272382e-01 -2.46897370e-01 -5.30622244e-01
-1.69873595e-01 -1.67378098e-01 -9.51356173e-01 -6.25651002e-01
-5.45389414e-01 1.04106002e-01 4.82618511e-01 4.96382445e-01
-9.09264311e-02 5.52666247e-01 -1.03187490e+00 6.28537238e-01
1.31862938e+00 5.13616316e-02 -1.57086909e-01 2.56006956e-01
7.52232909e-01 -1.64395705e-01 1.16533172e+00 5.72602272e-01
5.50962389e-01 1.34209239e+00 8.55011463e-01 3.74147922e-01
-8.06927085e-01 -4.32885289e-01 5.30109406e-01 3.17493975e-01
1.50221154e-01 -8.89281482e-02 -9.77397263e-01 7.70061910e-01
-1.75815809e+00 -7.72073507e-01 -8.62289131e-01 2.10603213e+00
-5.40360771e-02 5.17032921e-01 6.70172513e-01 -4.93519306e-01
1.06665790e+00 -3.07370313e-02 -5.93248308e-01 5.09306252e-01
-7.23864734e-02 -4.42632854e-01 1.28381038e+00 3.79795551e-01
-1.36181605e+00 9.16895807e-01 6.48293400e+00 7.00599015e-01
-3.97967726e-01 4.38105613e-01 -9.40570831e-01 -5.48323154e-01
4.00531977e-01 1.00513823e-01 -2.36713624e+00 3.93532991e-01
9.39329207e-01 -2.67157197e-01 -1.00735381e-01 9.92344141e-01
1.07410833e-01 -1.98759988e-01 -8.10962021e-01 1.24677360e+00
-3.87093991e-01 -1.65179324e+00 -3.45103681e-01 4.60568964e-01
4.49981600e-01 8.75268817e-01 -2.66895562e-01 4.96121138e-01
6.50713563e-01 -1.99896961e-01 1.20802510e+00 2.56812036e-01
8.48448813e-01 -2.05211237e-01 1.58733249e-01 7.49836206e-01
-2.21245623e+00 -3.53612788e-02 -2.75846988e-01 1.66667476e-01
6.52525663e-01 3.90848368e-01 -9.55983758e-01 6.81373775e-01
1.16436815e+00 9.07865226e-01 -5.24713695e-01 1.86885500e+00
1.65549546e-01 3.67365628e-02 -1.20007467e+00 -1.01985499e-01
2.13464543e-01 2.04147369e-01 1.48621500e+00 1.14750838e+00
5.03812551e-01 -1.39156699e-01 6.73518777e-01 6.29707873e-01
-3.40763107e-02 -6.55820608e-01 -8.80358577e-01 4.22098339e-01
8.11730206e-01 1.37345517e+00 -5.53103209e-01 -2.83856839e-01
-1.89062327e-01 1.20563522e-01 -2.68954456e-01 -4.37035173e-01
-1.16073465e+00 3.84680592e-02 1.14996862e+00 3.57420295e-01
8.64833474e-01 -7.11372316e-01 1.80822030e-01 -8.57825696e-01
-1.84385344e-01 4.87510227e-02 5.39291799e-01 -7.90554762e-01
-1.08538675e+00 1.99866876e-01 2.16372490e-01 -1.87647700e+00
1.99918076e-01 -7.09379971e-01 -1.66894376e-01 6.75378680e-01
-1.63605845e+00 -1.58454120e+00 -5.01630247e-01 5.66381216e-01
3.80696803e-01 -1.65903002e-01 3.14515680e-01 8.33374441e-01
-2.39656821e-01 3.13500822e-01 1.65478826e-01 -5.87585606e-02
5.15734136e-01 -1.18029153e+00 7.41431653e-01 9.30671513e-01
1.22978158e-01 1.58096686e-01 6.37373388e-01 -1.13766563e+00
-2.16450500e+00 -1.23560715e+00 4.75223273e-01 -1.50735426e+00
7.85627842e-01 -8.10164213e-01 -1.06910251e-01 8.58869016e-01
-1.95664689e-01 6.47234857e-01 1.07706867e-01 -2.06768095e-01
-9.37234014e-02 6.85670897e-02 -1.38573420e+00 2.90861517e-01
1.64058840e+00 -2.98598170e-01 -4.31909174e-01 6.59623027e-01
7.84437358e-01 -1.29526448e+00 -1.14246905e+00 3.06008428e-01
7.51699686e-01 -4.71793681e-01 1.44209743e+00 -1.26442999e-01
-1.23729575e+00 -1.24650252e+00 -3.35131705e-01 -4.40460116e-01
-2.16737047e-01 -9.46595550e-01 -7.63220072e-01 8.59066010e-01
-3.39184642e-01 -3.69152188e-01 8.81358445e-01 -1.87021885e-02
-4.55151379e-01 3.57121527e-01 -1.75711119e+00 -1.55447745e+00
-5.63068867e-01 -6.76930904e-01 4.58423793e-01 3.95951688e-01
-9.44116473e-01 7.42647517e-03 -4.43134487e-01 6.51959062e-01
1.52535832e+00 2.53590763e-01 1.51687062e+00 -1.71687770e+00
3.20408046e-01 -1.23994432e-01 -8.08125138e-01 -1.39635944e+00
-5.46832740e-01 -8.65379274e-01 -9.97333378e-02 -1.19496715e+00
-2.07962945e-01 -4.83479381e-01 2.71610338e-02 1.80149242e-01
2.80647546e-01 3.27853471e-01 7.44922578e-01 5.69789588e-01
-1.34575260e+00 3.19412112e-01 1.04144466e+00 -1.41875129e-02
2.07624048e-01 2.28917956e-01 4.51980047e-02 3.58030319e-01
-2.02408098e-02 -1.02862227e+00 4.19308156e-01 -4.70640868e-01
-1.67452559e-01 -8.32400769e-02 8.36098850e-01 -1.58395183e+00
6.94453478e-01 3.23527120e-02 3.82210523e-01 -1.91904712e+00
7.67183125e-01 -1.32902682e+00 4.59289640e-01 8.74835372e-01
4.31255072e-01 8.07427391e-02 7.51047432e-01 6.52442276e-01
5.55098653e-01 1.71323970e-01 8.82505834e-01 -9.89434794e-02
-1.09284770e+00 5.45108855e-01 -1.36622101e-01 1.07301369e-01
1.52383423e+00 -7.20400095e-01 -5.54176271e-01 1.29483148e-01
-7.99462616e-01 8.09757054e-01 5.42302608e-01 8.22562575e-01
5.13577402e-01 -1.60258973e+00 -6.92925334e-01 1.87772606e-02
2.93111235e-01 -4.11093757e-02 1.22844696e-01 1.10501158e+00
-2.18889073e-01 7.68717825e-01 -2.63149440e-01 -1.41455424e+00
-1.61320794e+00 4.21671003e-01 6.62540555e-01 -7.32599199e-02
-9.14931655e-01 7.13787973e-01 -3.07401657e-01 -6.15054488e-01
2.98193157e-01 -9.99368727e-02 2.92133719e-01 -5.82438223e-02
7.44488657e-01 7.40449190e-01 -1.86230585e-01 -9.77183759e-01
-1.01145339e+00 1.01910210e+00 1.44839033e-01 9.55499783e-02
9.57939327e-01 -5.19018829e-01 7.12422907e-01 4.88989323e-01
5.02784908e-01 5.32990396e-02 -1.66721570e+00 -2.04753429e-01
1.92735031e-01 -7.79805183e-01 3.23517054e-01 -6.12350166e-01
-5.11961162e-01 6.07752442e-01 1.42293155e+00 3.69719893e-01
2.11746290e-01 2.87072331e-01 6.96880758e-01 2.22080261e-01
1.00434411e+00 -9.20265019e-01 -2.76488453e-01 7.73718536e-01
2.53461301e-01 -1.45757127e+00 3.50207359e-01 -6.76909462e-02
-4.53109816e-02 8.06587219e-01 8.45592380e-01 -4.45696227e-02
5.27159512e-01 5.55255532e-01 1.03728324e-01 -6.85904443e-01
-5.74560940e-01 -7.75714934e-01 3.67548019e-01 9.63065326e-01
-3.22111517e-01 -1.14430040e-01 4.81383115e-01 -3.37728351e-01
2.49046102e-01 -2.45241582e-01 1.94649156e-02 1.35147655e+00
-6.79090798e-01 -6.74443960e-01 -1.12099457e+00 -5.38693368e-02
2.41971910e-01 5.13992012e-01 -3.33606929e-01 1.49052203e+00
2.74972677e-01 8.79529119e-01 2.38416865e-02 -3.89454961e-01
7.53569484e-01 -1.76407248e-01 7.31457055e-01 -5.30660927e-01
-9.33234632e-01 2.15886652e-01 3.62694375e-02 -8.08369219e-01
-3.51251155e-01 -9.92832184e-01 -1.14643502e+00 -1.86993882e-01
-8.01556170e-01 -2.76918143e-01 1.21214867e+00 6.98697031e-01
6.32818699e-01 3.37887198e-01 3.43940854e-01 -1.24326897e+00
-4.25893217e-01 -5.63722551e-01 -5.15631795e-01 -1.93108141e-01
7.54098833e-01 -1.36786711e+00 -1.47136793e-01 -5.51619053e-01] | [6.623039722442627, -2.218648672103882] |
b1a2380f-f11e-41a2-b977-befa08abb9b8 | solving-royal-game-of-ur-using-reinforcement | 2208.10669 | null | https://arxiv.org/abs/2208.10669v1 | https://arxiv.org/pdf/2208.10669v1.pdf | Solving Royal Game of Ur Using Reinforcement Learning | Reinforcement Learning has recently surfaced as a very powerful tool to solve complex problems in the domain of board games, wherein an agent is generally required to learn complex strategies and moves based on its own experiences and rewards received. While RL has outperformed existing state-of-the-art methods used for playing simple video games and popular board games, it is yet to demonstrate its capability on ancient games. Here, we solve one such problem, where we train our agents using different methods namely Monte Carlo, Qlearning and Expected Sarsa to learn optimal policy to play the strategic Royal Game of Ur. The state space for our game is complex and large, but our agents show promising results at playing the game and learning important strategic moves. Although it is hard to conclude that when trained with limited resources which algorithm performs better overall, but Expected Sarsa shows promising results when it comes to fastest learning. | ['Girik Malik', 'Sidharth Malhotra'] | 2022-08-23 | null | null | null | null | ['board-games'] | ['playing-games'] | [-1.54948846e-01 8.17067847e-02 -6.70076981e-02 3.05746406e-01
-7.80109644e-01 -7.86449790e-01 6.03150547e-01 -1.36592656e-01
-1.12070560e+00 1.37884855e+00 -5.91062345e-02 -5.48152745e-01
-4.37918812e-01 -7.11569786e-01 -5.15146494e-01 -7.89141238e-01
-5.96403956e-01 9.15718734e-01 4.71066117e-01 -8.81505132e-01
5.41001618e-01 2.71930665e-01 -1.49548006e+00 -1.40485868e-01
6.01232469e-01 6.89221025e-01 4.33575869e-01 9.75636721e-01
1.88259259e-01 1.19431388e+00 -7.67315507e-01 -2.76019156e-01
5.99676311e-01 -6.01789415e-01 -8.48141670e-01 -2.14149818e-01
-4.85026479e-01 -3.11592311e-01 -2.65941948e-01 8.43252242e-01
8.32379818e-01 5.92488706e-01 2.12145314e-01 -8.95253301e-01
3.20249110e-01 8.06555748e-01 -4.37038094e-01 4.43132192e-01
6.11833632e-01 6.01803720e-01 1.20456111e+00 -1.52675174e-02
5.34071922e-01 1.12810540e+00 3.79926592e-01 6.39348030e-01
-9.13564444e-01 -4.61024821e-01 1.84434429e-01 4.28926945e-01
-8.68267417e-01 9.36668962e-02 2.17377901e-01 9.29744542e-02
1.08152080e+00 7.82025307e-02 1.25001574e+00 1.01611447e+00
1.90878600e-01 1.02771199e+00 1.42278016e+00 -2.75815427e-01
7.98327327e-01 -3.57831776e-01 -8.22885811e-01 4.87817079e-01
4.08128388e-02 7.58352518e-01 -2.89548874e-01 -1.36579111e-01
8.36897731e-01 -3.48933190e-01 -8.75767097e-02 -3.01978946e-01
-8.37512195e-01 1.11184907e+00 2.91581184e-01 2.25659981e-01
-8.36124659e-01 4.12441075e-01 3.66545498e-01 6.35820508e-01
-2.15134472e-01 1.20454335e+00 -6.14291131e-01 -9.54480469e-01
-7.23148942e-01 9.12971258e-01 1.12814748e+00 3.23730350e-01
5.94789147e-01 3.61384332e-01 7.85187185e-02 5.45899510e-01
1.91916935e-02 3.64239722e-01 4.26083803e-01 -1.46109617e+00
3.26061934e-01 1.50143296e-01 5.63677669e-01 -4.34556156e-01
-6.24332368e-01 -3.53760540e-01 -2.56634086e-01 9.37545002e-01
6.75722003e-01 -8.15966308e-01 -5.06096959e-01 1.47954071e+00
7.06525594e-02 2.97260046e-01 4.73176569e-01 8.34893942e-01
2.73164004e-01 7.76694596e-01 -1.25289679e-01 -3.30516279e-01
1.02169359e+00 -9.20041740e-01 -3.32602859e-01 -5.99225819e-01
4.40656155e-01 -3.82916927e-01 6.71818137e-01 9.03564095e-01
-1.66338670e+00 -9.25106034e-02 -9.65582013e-01 7.73441076e-01
-6.34630919e-02 -5.73487580e-01 1.03710461e+00 6.47780955e-01
-1.10635591e+00 8.52190971e-01 -7.52955556e-01 -2.05508426e-01
4.06761497e-01 6.35481834e-01 -3.24164294e-02 -9.95934568e-03
-1.38414240e+00 1.42038214e+00 6.09533489e-01 -1.02367096e-01
-1.32851398e+00 -2.64749348e-01 -6.56615674e-01 4.26813699e-02
1.33522582e+00 -4.74123836e-01 1.75524938e+00 -7.91009009e-01
-2.21667576e+00 4.35402513e-01 5.16676843e-01 -7.65095115e-01
7.14474380e-01 4.88905702e-03 1.06594168e-01 -2.21172869e-02
-1.14760578e-01 5.38248479e-01 6.41901851e-01 -9.99881148e-01
-1.18773258e+00 -6.14918657e-02 7.59424686e-01 7.54979134e-01
3.14187706e-01 9.91598666e-02 -1.12187736e-01 -2.31376857e-01
-4.29792941e-01 -9.72814441e-01 -9.34069574e-01 -8.23369145e-01
2.13911250e-01 -3.24122161e-01 -6.13676496e-02 -2.20406786e-01
1.04860210e+00 -1.58008015e+00 4.83952403e-01 3.03185314e-01
5.32439500e-02 4.02652830e-01 -1.99298292e-01 6.97389245e-01
2.35984117e-01 -1.54016495e-01 1.06626853e-01 3.50833476e-01
3.17354828e-01 4.79017049e-01 4.51722182e-02 1.17215902e-01
-1.84896767e-01 8.98566663e-01 -1.27115083e+00 5.11365477e-04
4.77106199e-02 -1.96161777e-01 -7.68190861e-01 3.21622372e-01
-4.09761965e-01 2.01103896e-01 -7.86971688e-01 2.50372201e-01
6.06687851e-02 4.17196341e-02 5.19659340e-01 6.98343337e-01
-2.34521702e-01 2.43726164e-01 -1.46852791e+00 1.52867818e+00
-2.86953390e-01 3.04416150e-01 4.33088616e-02 -1.16990733e+00
6.44831240e-01 2.36339062e-01 5.17866492e-01 -1.05046010e+00
2.88358390e-01 1.69934079e-01 5.43237448e-01 -4.28903371e-01
6.23876274e-01 -4.11554128e-01 -3.30632031e-01 5.72989106e-01
-1.00239120e-01 -6.60288751e-01 6.51932478e-01 8.61973763e-02
1.59329545e+00 3.83802116e-01 5.19775748e-01 -8.69745687e-02
3.27734590e-01 4.31178212e-01 5.08126855e-01 1.28821325e+00
-1.23222820e-01 5.38551062e-02 8.80812943e-01 -6.30278707e-01
-9.37697947e-01 -9.77852523e-01 5.98218799e-01 1.31125164e+00
2.27429181e-01 -3.32673669e-01 -5.20914197e-01 -5.35467088e-01
-1.65103510e-01 7.76317954e-01 -5.58074594e-01 -2.39780694e-02
-6.65265381e-01 -8.29460680e-01 2.80928046e-01 1.68751672e-01
5.09629905e-01 -1.74238658e+00 -1.26249814e+00 7.90642738e-01
2.48067901e-01 -8.04601908e-01 4.90423888e-02 5.30972779e-01
-6.68902934e-01 -1.30318272e+00 -6.99252307e-01 -4.51277554e-01
6.17562830e-02 -2.11283788e-02 1.37033367e+00 1.63530130e-02
-1.74650133e-01 4.99034196e-01 -5.60226083e-01 -6.31367862e-01
-3.93863946e-01 1.22160956e-01 -7.26142479e-03 -8.22174907e-01
8.41135450e-04 -4.74210024e-01 -6.36294782e-01 2.02655137e-01
-7.36683667e-01 -2.07719654e-01 6.92937851e-01 9.41216350e-01
5.68852946e-02 5.29298604e-01 4.64664549e-01 -7.26819158e-01
1.22011697e+00 -3.54990631e-01 -8.97183716e-01 7.52591342e-02
-2.33232707e-01 2.03531563e-01 7.36322701e-01 -2.72645146e-01
-7.54350960e-01 -3.18611681e-01 -1.77656949e-01 1.04946241e-01
1.74702331e-01 4.67545182e-01 3.75473231e-01 2.52633495e-03
7.77891695e-01 2.45070145e-01 1.13739833e-01 2.54404210e-02
1.77674443e-01 4.43483032e-02 1.10313013e-01 -8.10580790e-01
6.70138657e-01 -5.63288592e-02 -7.29484782e-02 -4.73984033e-01
-5.04077375e-01 -1.15199499e-01 -3.55473310e-02 -3.19473445e-01
7.20797539e-01 -8.03652167e-01 -1.42990267e+00 4.79228973e-01
-6.19492233e-01 -9.41430151e-01 -6.03532255e-01 4.31573689e-01
-1.23196042e+00 -4.14897874e-02 -5.01680911e-01 -9.83208716e-01
1.56302303e-01 -1.24466276e+00 3.64627361e-01 6.63149953e-01
-6.87321126e-02 -9.92812932e-01 3.91598254e-01 1.97072968e-01
4.75403845e-01 1.30995616e-01 4.33585525e-01 -5.49949110e-01
-5.40353894e-01 4.54904437e-02 4.15553838e-01 -2.56966297e-02
-1.45855084e-01 -2.12638959e-01 -3.74631345e-01 -6.15710258e-01
-2.31939405e-01 -7.93756604e-01 4.45838094e-01 4.90456402e-01
7.87303150e-01 -1.23999827e-01 6.22113347e-02 1.62476406e-01
1.35485733e+00 7.25732684e-01 8.52798641e-01 1.08762968e+00
-1.93304569e-01 3.71345013e-01 9.82671261e-01 8.72882783e-01
4.71438080e-01 6.59702837e-01 7.51130342e-01 2.09901199e-01
6.21489406e-01 -5.95360771e-02 5.74352205e-01 2.29815677e-01
-6.93739116e-01 -1.42057583e-01 -8.71962845e-01 2.00105757e-01
-2.16333413e+00 -1.49431014e+00 6.31911278e-01 2.12914968e+00
7.57143497e-01 7.39108264e-01 6.88344121e-01 -5.31764030e-02
2.93542087e-01 1.81152359e-01 -7.55927384e-01 -6.17589235e-01
-6.83495551e-02 6.40023232e-01 5.51554501e-01 4.66217816e-01
-7.17577219e-01 1.42152238e+00 7.04376173e+00 9.26758409e-01
-8.05366278e-01 -9.42081586e-02 4.27371144e-01 -5.35534680e-01
4.31631953e-02 -7.72835091e-02 -5.20610273e-01 3.15094113e-01
8.20749640e-01 -2.00629130e-01 1.20255947e+00 7.25901127e-01
4.77880478e-01 -7.90841758e-01 -6.44914389e-01 8.51707101e-01
-3.48973274e-01 -1.34842515e+00 -5.08178353e-01 1.60448477e-01
9.30335999e-01 8.62224549e-02 -2.31230147e-02 9.66658890e-01
1.44576073e+00 -1.31290638e+00 5.45921981e-01 3.23065311e-01
6.92269877e-02 -1.17820048e+00 9.04244125e-01 6.66527450e-01
-7.14498281e-01 -6.08380437e-01 -5.12628853e-01 -6.88561499e-01
-3.84964272e-02 -3.31428170e-01 -8.44415128e-01 3.88074487e-01
6.50776744e-01 2.57587433e-01 -1.91002652e-01 1.30998778e+00
-4.78340447e-01 4.99991298e-01 -3.11540335e-01 -7.18890727e-01
1.13627565e+00 -3.73995960e-01 3.27025712e-01 4.46937829e-01
3.19233865e-01 4.98867035e-01 4.70121354e-01 4.17106897e-01
4.16521043e-01 -1.56080350e-01 -6.21333659e-01 -4.87412140e-02
3.17688793e-01 1.13856554e+00 -7.64235616e-01 -8.70459303e-02
-6.32875636e-02 6.30910575e-01 4.03297603e-01 3.06812167e-01
-7.15438068e-01 -1.27784893e-01 7.51736283e-01 -2.16867834e-01
5.53873420e-01 -2.58075744e-01 2.40524247e-01 -9.29180503e-01
-5.02844155e-01 -1.41253781e+00 5.39879024e-01 -6.82973146e-01
-7.31914699e-01 5.23616254e-01 -1.15639810e-02 -9.40035164e-01
-8.03255796e-01 -6.87112510e-01 -7.95458794e-01 5.47833443e-01
-1.37313402e+00 -2.34383285e-01 1.93855345e-01 6.45521104e-01
6.38498545e-01 -6.90966845e-01 5.53434610e-01 -3.31658870e-01
-4.30602044e-01 1.01925604e-01 3.51047724e-01 -3.21267903e-01
1.72025308e-01 -1.48152566e+00 1.95844188e-01 3.84365678e-01
1.00277074e-01 -1.86994132e-02 1.05042279e+00 -3.44209790e-01
-1.56668866e+00 -9.32110474e-02 -1.80485994e-01 -2.12141588e-01
9.23149765e-01 5.93202515e-03 -2.03482062e-01 3.48835796e-01
3.51287216e-01 -3.33884597e-01 3.80751878e-01 1.79646134e-01
4.47849095e-01 2.05997154e-02 -9.68256295e-01 9.21798706e-01
8.78493607e-01 7.66987503e-02 -6.50841832e-01 1.69946849e-01
9.94361788e-02 -7.05300570e-01 -5.93991995e-01 8.39170218e-02
4.08283830e-01 -1.24470949e+00 9.17353511e-01 -8.76545370e-01
2.82915890e-01 -1.99842066e-01 1.96059216e-02 -2.05930519e+00
-4.16770667e-01 -1.09294808e+00 1.51194826e-01 4.39302742e-01
5.05618215e-01 -5.71328521e-01 1.31277716e+00 2.04350486e-01
2.21241504e-01 -9.13473368e-01 -8.91689241e-01 -7.30799377e-01
2.67019302e-01 -2.90203124e-01 4.86976534e-01 5.18535733e-01
1.76121697e-01 2.58943677e-01 -5.52913666e-01 -1.97575033e-01
4.89030212e-01 -4.71028797e-02 9.08220053e-01 -9.90785062e-01
-7.84324467e-01 -7.82499075e-01 -2.73349315e-01 -9.33959424e-01
1.25738755e-01 -3.36265773e-01 2.45885164e-01 -1.61966431e+00
-5.91728725e-02 -4.42435294e-01 -2.97267288e-01 2.22921580e-01
-2.26155236e-01 -9.41345617e-02 5.51409900e-01 -2.49228343e-01
-1.20991194e+00 4.95554864e-01 1.58322275e+00 -2.97763338e-03
-3.38827699e-01 4.83365238e-01 -7.95461357e-01 7.51027822e-01
9.64666665e-01 -2.96764672e-01 -4.62220520e-01 -3.42211663e-03
8.18733573e-01 6.72529697e-01 -9.43101272e-02 -1.10552251e+00
2.92673796e-01 -6.23283982e-01 1.69014856e-01 -7.67213255e-02
2.76962489e-01 -6.19500875e-01 -1.29684964e-02 1.02217054e+00
-1.34452820e-01 3.62180978e-01 2.81681597e-01 4.58361894e-01
-3.55367139e-02 -6.00750327e-01 5.54938972e-01 -8.13594639e-01
-1.16079760e+00 2.40810543e-01 -1.03778946e+00 6.22737467e-01
1.32552731e+00 -3.34225506e-01 -1.85558130e-03 -9.90383923e-01
-9.12496805e-01 7.67663598e-01 2.44872838e-01 1.92133039e-01
4.06739652e-01 -9.23460841e-01 -8.18442106e-01 -2.83623219e-01
-3.79128277e-01 -1.58743396e-01 1.66266590e-01 4.44146901e-01
-8.55196476e-01 1.99192733e-01 -6.55403376e-01 -1.06631331e-01
-1.06050646e+00 3.24719846e-01 6.31959677e-01 -9.67025816e-01
-4.48652565e-01 7.64096856e-01 -2.57524818e-01 -4.34866071e-01
2.72213310e-01 1.37383409e-04 -4.06670123e-01 8.11406448e-02
5.12939095e-01 4.72610980e-01 -2.37683609e-01 1.08965024e-01
-9.18972939e-02 1.90259591e-01 -1.14593625e-01 -3.91609281e-01
1.76896012e+00 2.55563945e-01 3.19072306e-01 1.69995695e-01
2.25985289e-01 -3.43654037e-01 -1.71527624e+00 -3.27701159e-02
1.08853638e-01 -4.36892271e-01 -4.78989109e-02 -9.69851494e-01
-8.00237358e-01 5.11186361e-01 3.76117527e-01 5.20754218e-01
9.59880054e-01 -3.43883127e-01 4.60191220e-01 8.24933052e-01
1.02138197e+00 -1.40781200e+00 4.19080704e-01 1.04824960e+00
5.50391674e-01 -1.18497765e+00 1.62859365e-01 4.07817334e-01
-1.09358680e+00 9.77863550e-01 6.41551733e-01 -5.52680373e-01
1.20566979e-01 3.81282538e-01 2.01463695e-06 -2.56331325e-01
-1.12427247e+00 -7.15384424e-01 -3.40595424e-01 6.87112451e-01
-2.36077264e-01 5.71297854e-02 -2.29971617e-01 6.36349469e-02
-5.27114928e-01 -1.83126464e-01 7.81927347e-01 1.02723181e+00
-8.14668596e-01 -1.41588843e+00 -3.76035392e-01 4.84665513e-01
-6.19338751e-01 1.39252901e-01 -1.80748522e-01 1.06045318e+00
-1.33517265e-01 7.37892509e-01 -1.71071365e-01 -4.73067760e-02
3.76474559e-01 -3.79619867e-01 9.07238424e-01 -5.80230474e-01
-1.05305839e+00 -6.22878596e-02 2.81791717e-01 -8.03367257e-01
-3.24345738e-01 -8.09763908e-01 -1.26617825e+00 -6.16449296e-01
6.27995357e-02 5.66117883e-01 3.51097852e-01 1.15097237e+00
-2.98659980e-01 6.24882102e-01 6.79036558e-01 -1.17471921e+00
-1.03386676e+00 -4.79216337e-01 -8.41117501e-01 -3.48121002e-02
6.66590631e-02 -8.01768184e-01 -5.73211722e-02 -9.42088127e-01] | [3.572882890701294, 1.4830763339996338] |
32b9ba33-7b53-4549-a39c-370df4ee81ca | unique-class-group-based-multi-label | 2003.08751 | null | https://arxiv.org/abs/2003.08751v1 | https://arxiv.org/pdf/2003.08751v1.pdf | Unique Class Group Based Multi-Label Balancing Optimizer for Action Unit Detection | Balancing methods for single-label data cannot be applied to multi-label problems as they would also resample the samples with high occurrences. We propose to reformulate this problem as an optimization problem in order to balance multi-label data. We apply this balancing algorithm to training datasets for detecting isolated facial movements, so-called Action Units. Several Action Units can describe combined emotions or physical states such as pain. As datasets in this area are limited and mostly imbalanced, we show how optimized balancing and then augmentation can improve Action Unit detection. At the IEEE Conference on Face and Gesture Recognition 2020, we ranked third in the Affective Behavior Analysis in-the-wild (ABAW) challenge for the Action Unit detection task. | ['Jaspar Pahl', 'Dominik Seuss', 'Ines Rieger'] | 2020-03-05 | null | null | null | null | ['action-unit-detection'] | ['computer-vision'] | [ 5.19395411e-01 2.94407368e-01 -8.13550889e-01 -7.34018326e-01
-8.80665839e-01 -3.02445799e-01 2.73244113e-01 -1.80749446e-01
-4.27331150e-01 6.42971218e-01 3.22041184e-01 3.05585057e-01
1.77053764e-01 -1.83425203e-01 -2.07384422e-01 -6.29883051e-01
1.44972235e-01 4.42830741e-01 -5.06531417e-01 9.16037783e-02
-9.36384201e-02 4.68509108e-01 -1.81633830e+00 7.57239401e-01
2.20140904e-01 1.00827396e+00 -9.01182115e-01 5.58582067e-01
1.47789866e-01 8.83554101e-01 -6.31797075e-01 -2.36763656e-01
1.15421921e-01 -8.36645901e-01 -9.62073982e-01 4.25957292e-01
7.66293049e-01 -4.07225668e-01 2.98503906e-01 8.53032887e-01
9.96553898e-01 1.67168915e-01 8.69541049e-01 -2.05502081e+00
-1.30296834e-02 2.16142744e-01 -1.14330971e+00 -1.58667311e-01
5.00940859e-01 -3.90043229e-01 1.08178818e+00 -9.53985751e-01
9.03357506e-01 1.53044617e+00 7.00743914e-01 1.37100029e+00
-1.27718592e+00 -8.81985068e-01 2.15846911e-01 2.01331615e-01
-1.02205408e+00 -7.22175479e-01 6.33044600e-01 -4.15841222e-01
9.23042059e-01 5.60798705e-01 6.11879647e-01 1.38200796e+00
-4.07956213e-01 1.28881741e+00 1.27445698e+00 -4.88594741e-01
1.64857820e-01 -2.45232671e-01 2.19166085e-01 6.56524956e-01
-1.12340167e-01 -2.32543051e-01 -9.04379725e-01 -4.24239933e-01
1.30955547e-01 -1.64541960e-01 1.28342718e-01 -9.66751128e-02
-9.46186304e-01 8.75159442e-01 -2.51377318e-02 1.35879293e-01
-3.07976156e-01 5.13109386e-01 6.48296595e-01 2.29522228e-01
7.74703443e-01 2.12697595e-01 -4.59062546e-01 -2.16101244e-01
-1.05511069e+00 3.08329821e-01 5.92094004e-01 4.22447413e-01
5.42996943e-01 -2.85939425e-01 -3.68687212e-01 1.23260939e+00
2.69126564e-01 1.99956417e-01 6.69652700e-01 -1.25447428e+00
-3.61900665e-02 5.38636386e-01 -1.08437777e-01 -6.18107855e-01
-1.04949653e+00 3.39833230e-01 -7.01153815e-01 5.75503826e-01
3.37712795e-01 -3.57522726e-01 -1.10832465e+00 2.23803854e+00
7.31578588e-01 3.26886438e-02 -4.04085457e-01 7.61783659e-01
9.55384433e-01 1.75254390e-01 3.65340769e-01 -6.37074947e-01
1.46216762e+00 -9.81767833e-01 -1.09687543e+00 -3.53728563e-01
9.87668395e-01 -6.73742115e-01 9.47925329e-01 5.68492830e-01
-9.95265365e-01 -2.58099139e-01 -8.32486749e-01 2.56965333e-03
-2.45031074e-01 2.36003697e-01 6.61580622e-01 8.06734920e-01
-9.78543818e-01 5.10079801e-01 -6.97014987e-01 -5.33378363e-01
9.36824441e-01 5.16523361e-01 -4.37574834e-01 2.35922933e-01
-9.35398281e-01 1.00450659e+00 -1.32161483e-01 -6.24217317e-02
-5.79382598e-01 -5.13679504e-01 -7.02347934e-01 -3.61877501e-01
2.76092798e-01 -2.38510519e-01 1.35046661e+00 -1.43026149e+00
-1.36841321e+00 1.60563612e+00 -2.76199013e-01 1.66894332e-01
5.31671762e-01 -1.21496737e-01 -3.59339863e-01 3.36678252e-02
-8.88599232e-02 1.21532238e+00 1.00946867e+00 -1.09340572e+00
-5.33915699e-01 -6.32430375e-01 -1.47744894e-01 2.75924355e-01
-5.37130713e-01 4.33855146e-01 -1.43303111e-01 -7.37902403e-01
1.32385522e-01 -1.20348227e+00 7.74881095e-02 1.21807061e-01
-3.29772204e-01 -4.19375598e-01 9.16888297e-01 -4.84294832e-01
1.14911604e+00 -1.98950744e+00 2.15960234e-01 1.61402747e-01
2.26644382e-01 5.02692983e-02 -3.43562603e-01 -1.11142226e-01
-5.92071831e-01 3.55676889e-01 1.57261536e-01 -7.98800945e-01
1.27159297e-01 1.91577747e-01 9.07365084e-02 7.48795629e-01
1.94619268e-01 8.76684725e-01 -8.23191941e-01 -6.94903553e-01
-4.24840041e-02 2.38556772e-01 -6.08050942e-01 2.83078793e-02
-8.95734504e-02 3.59040916e-01 -3.82357650e-02 1.15662682e+00
6.28209293e-01 8.61271098e-02 9.86543447e-02 -2.21594095e-01
4.77064162e-01 -2.16029048e-01 -1.10637712e+00 1.41313338e+00
-2.90535659e-01 6.78065360e-01 2.19815090e-01 -8.28621030e-01
6.34004474e-01 5.56311488e-01 1.21626222e+00 -7.14689493e-01
3.29214841e-01 5.81551120e-02 -1.15406588e-01 -7.03056872e-01
4.02889252e-02 -8.03542435e-01 -2.03310147e-01 7.33772099e-01
1.48343354e-01 -9.16792825e-02 2.99928457e-01 -5.56463599e-02
1.08510768e+00 2.95335233e-01 5.46667695e-01 4.48214673e-02
1.32476211e-01 -3.93826306e-01 7.40851939e-01 2.74411768e-01
-8.21178019e-01 5.69879949e-01 7.85099030e-01 -4.17647749e-01
-5.07040143e-01 -5.78716874e-01 -1.03470854e-01 1.81652367e+00
-2.69559473e-01 -5.55355549e-01 -8.98417652e-01 -9.19437349e-01
8.43535885e-02 1.92473069e-01 -1.05512273e+00 -3.50713134e-01
-2.96778679e-01 -1.21643758e+00 8.46587062e-01 5.90032995e-01
3.49290483e-02 -1.32876730e+00 -9.79012489e-01 -1.40948698e-01
-6.29415154e-01 -9.26141322e-01 -2.86790520e-01 3.70387256e-01
-5.01344264e-01 -1.20312035e+00 -5.45774817e-01 -6.76629663e-01
6.42114818e-01 -3.83763850e-01 1.26486230e+00 1.24305949e-01
-7.28030801e-01 4.57123011e-01 -3.21969807e-01 -7.28750587e-01
-2.64688671e-01 -1.30036518e-01 2.42181227e-01 3.15742642e-01
6.26361609e-01 -2.47439399e-01 -4.11414504e-01 3.41816068e-01
-7.41525173e-01 4.35387380e-02 3.09605390e-01 8.26313674e-01
4.05266792e-01 -5.68224609e-01 8.86485994e-01 -8.91430974e-01
6.38438582e-01 -1.37365133e-01 1.54362202e-01 3.62465680e-01
-3.98943126e-01 -3.42696577e-01 -1.53660566e-01 -7.80282438e-01
-6.95986629e-01 5.23349166e-01 -1.37558997e-01 -3.79164517e-01
-2.93639600e-01 5.66967726e-02 -2.54590511e-01 -4.52771336e-02
5.91242909e-01 -6.38845444e-01 -2.80548390e-02 -1.30938411e-01
3.64090085e-01 7.98265934e-01 2.35085472e-01 -4.35475081e-01
-3.07261031e-02 6.30353093e-01 2.74805456e-01 -5.35975635e-01
-1.14156878e+00 -6.06308162e-01 -7.07660377e-01 -8.67074132e-01
1.03658950e+00 -7.12245584e-01 -1.09006500e+00 6.77185774e-01
-8.26319695e-01 -5.77532232e-01 -3.75239521e-01 1.61764234e-01
-8.37303817e-01 -7.20623415e-03 -6.64376855e-01 -1.04534185e+00
-2.81942278e-01 -9.13843989e-01 1.56609297e+00 1.20758556e-01
-9.31709945e-01 -5.67493677e-01 3.45267624e-01 4.84733045e-01
4.83786128e-02 6.69831455e-01 6.87415421e-01 -5.19066930e-01
4.77795631e-01 -2.02921957e-01 8.02265629e-02 3.83438975e-01
2.61584193e-01 1.45904273e-01 -1.50264394e+00 -2.24404946e-01
-1.34223439e-02 -1.40249705e+00 9.93839622e-01 3.93900156e-01
1.29209340e+00 -1.07859530e-01 -3.19851756e-01 2.43123561e-01
8.72171998e-01 -7.16432929e-02 5.21798670e-01 -9.33327898e-02
6.59289181e-01 1.12182403e+00 8.43989849e-01 6.72379375e-01
-4.33742441e-02 9.29616451e-01 5.12169957e-01 -4.33734208e-01
-2.15358093e-01 3.00775230e-01 3.77676517e-01 2.39191025e-01
-1.13692023e-01 -2.09062397e-01 -7.94837296e-01 3.59991550e-01
-1.86538923e+00 -1.03019941e+00 3.79935023e-03 1.78453481e+00
1.14796281e+00 -2.49166355e-01 5.21623373e-01 2.42336959e-01
8.06280434e-01 2.23496214e-01 -6.02227747e-01 -5.92252731e-01
7.17397919e-03 3.51281464e-01 1.49630338e-01 6.13447666e-01
-1.45225537e+00 8.16021740e-01 6.87066555e+00 9.06580269e-01
-1.12803113e+00 4.36563104e-01 1.16315460e+00 -7.62573361e-01
1.06053218e-01 -4.83490348e-01 -7.15808094e-01 2.70688474e-01
8.24270666e-01 3.95251811e-01 3.12533230e-01 7.36430109e-01
1.29410371e-01 -3.20928127e-01 -1.26699102e+00 1.31228459e+00
4.98849541e-01 -5.23393095e-01 -3.47660393e-01 9.42984670e-02
8.60120833e-01 -2.40882561e-01 5.44877015e-02 4.44943368e-01
1.77563787e-01 -1.38007474e+00 5.71072936e-01 2.11930647e-01
1.05798125e+00 -6.65341079e-01 3.48732352e-01 8.81834924e-02
-8.00244868e-01 -2.18477044e-02 -5.61446957e-02 -1.10309787e-01
1.30604476e-01 5.74743271e-01 -5.98866165e-01 -1.85234562e-01
6.00613296e-01 4.98124987e-01 -5.87936819e-01 6.55524552e-01
-1.70791388e-01 5.54662406e-01 -5.46180129e-01 6.69389591e-02
-5.00364229e-02 2.54396051e-02 6.22958168e-02 1.14009774e+00
-4.08693254e-02 2.95458045e-02 1.48263708e-01 4.37682942e-02
-4.18457597e-01 2.57844955e-01 -4.19501334e-01 7.20620751e-02
1.56175746e-02 1.69550371e+00 -9.78892624e-01 -4.06138003e-01
-2.53546149e-01 1.23276150e+00 2.52743989e-01 2.03470320e-01
-1.11571574e+00 1.59174532e-01 8.20375800e-01 -3.15007657e-01
-2.33606368e-01 5.19524693e-01 -3.94043148e-01 -1.04179168e+00
-1.83552667e-01 -1.00920796e+00 7.61026859e-01 -6.70191169e-01
-1.19189882e+00 2.98843473e-01 -1.28373459e-01 -1.07472491e+00
-4.19505388e-01 -6.93497777e-01 -1.51893616e-01 3.12009215e-01
-1.03231835e+00 -1.34154642e+00 -5.91284633e-01 5.59322655e-01
3.79637420e-01 3.51820737e-01 1.02940321e+00 6.41538620e-01
-6.80093348e-01 7.78748214e-01 -3.36948544e-01 2.81815007e-02
1.32764840e+00 -1.17827117e+00 -2.27284908e-01 3.28752130e-01
3.58707160e-01 -4.59289886e-02 5.53402126e-01 -6.19684160e-01
-6.95119262e-01 -9.92813945e-01 8.98170352e-01 -5.62439680e-01
2.73070395e-01 -5.76436460e-01 -5.47121048e-01 7.12824345e-01
4.05404158e-02 3.17282110e-01 1.18002987e+00 2.76412010e-01
-3.99740040e-01 1.39063701e-01 -1.36295593e+00 5.12390792e-01
1.26792192e+00 -4.80349034e-01 -2.10178271e-01 6.48705840e-01
-7.58620948e-02 -4.22078431e-01 -7.15859532e-01 7.56210029e-01
8.98945332e-01 -7.52554417e-01 8.32477331e-01 -1.17229283e+00
5.06378829e-01 2.00110376e-01 -1.88145354e-01 -1.25544119e+00
-5.34189418e-02 -4.93908077e-01 -1.83078974e-01 1.35019267e+00
1.30386159e-01 -9.43326727e-02 1.00762010e+00 7.10064590e-01
3.81668508e-01 -9.90165591e-01 -1.13951421e+00 -4.42505628e-01
-4.67393287e-02 -4.09635037e-01 1.85526356e-01 1.14418519e+00
3.39211464e-01 4.42813933e-01 -7.19589114e-01 -5.87330222e-01
4.18846399e-01 7.67634436e-02 4.93793011e-01 -1.24598539e+00
-3.50907333e-02 -7.56795526e-01 -4.96334285e-01 -2.43615404e-01
5.56215763e-01 -8.44593406e-01 1.30588457e-01 -1.24190569e+00
5.24950743e-01 -5.63445427e-02 -3.77428919e-01 1.17167616e+00
-3.15322369e-01 1.06499755e+00 2.22357914e-01 -1.77409835e-02
-7.48294055e-01 3.95114660e-01 8.58512998e-01 -1.16093881e-01
6.93521202e-02 -6.01305440e-02 -5.38231134e-01 9.70525563e-01
6.96214139e-01 -4.89644706e-01 -2.22821429e-01 1.72626704e-01
3.76036316e-01 -1.66206524e-01 1.45064428e-01 -7.58883655e-01
-1.77338198e-01 -3.10496956e-01 5.03379583e-01 -5.55565059e-01
5.96887171e-01 -7.96210587e-01 -2.10645288e-01 4.34683144e-01
-6.87632322e-01 -2.90817827e-01 1.42127410e-01 1.11183062e-01
-1.43344462e-01 -2.29269758e-01 1.18883073e+00 4.73368205e-02
-6.03450954e-01 1.57060042e-01 -4.48661178e-01 2.14978158e-01
1.36951041e+00 -3.30678746e-02 -3.41784805e-01 -5.69864988e-01
-1.23273194e+00 2.98095912e-01 3.23213369e-01 4.79402453e-01
3.78205419e-01 -1.70863879e+00 -7.22005844e-01 3.59370112e-02
2.07740620e-01 -4.44651872e-01 1.72777161e-01 1.08615637e+00
-1.29084866e-02 -1.27874047e-01 -5.60254753e-01 -4.66912001e-01
-2.01344156e+00 3.33782196e-01 4.70999420e-01 -3.60351235e-01
2.39783406e-01 1.18925703e+00 -6.27480522e-02 -5.88485420e-01
4.85550493e-01 -1.94309399e-01 -2.57617742e-01 1.20879960e+00
6.51488304e-01 5.57884753e-01 2.71057367e-01 -7.49666214e-01
-7.63920009e-01 5.43378770e-01 3.45294416e-01 -3.76794726e-01
1.17599607e+00 5.14476523e-02 -4.12800491e-01 6.51330948e-01
1.47725892e+00 -1.53924167e-01 -9.13509309e-01 3.27303886e-01
1.70539081e-01 -2.68271923e-01 -1.11666225e-01 -8.43185842e-01
-1.10154307e+00 9.75573421e-01 8.71579349e-01 3.06150820e-02
1.33074439e+00 1.02515481e-01 4.41875190e-01 1.38127223e-01
1.14321679e-01 -1.60743022e+00 5.57488441e-01 1.31124869e-01
8.63122523e-01 -1.42229331e+00 8.45255032e-02 -3.94269913e-01
-7.36990333e-01 1.10352862e+00 9.08173978e-01 3.25537354e-01
5.50923824e-01 6.06028020e-01 3.79070818e-01 -2.36518338e-01
-8.18336189e-01 -4.38141972e-01 2.54829258e-01 4.20755118e-01
5.73835790e-01 1.62727982e-01 -4.13657457e-01 3.42884004e-01
1.67214021e-01 2.31869012e-01 2.21135363e-01 9.39508915e-01
-2.01001361e-01 -1.25019848e+00 -4.85112399e-01 7.95803666e-01
-5.74852943e-01 4.10475582e-01 -9.34637010e-01 4.91348326e-01
4.73762393e-01 9.53606367e-01 1.59945846e-01 -5.12600720e-01
2.58391500e-01 6.76384389e-01 7.66993701e-01 -6.68508887e-01
-5.67573845e-01 2.63256341e-01 3.57015997e-01 -1.09617233e+00
-1.04632878e+00 -9.85863030e-01 -1.36899912e+00 -2.78096031e-02
-2.50033170e-01 -4.94915068e-01 5.01557946e-01 8.52390945e-01
-1.51748704e-02 4.65350360e-01 4.58431304e-01 -9.50599372e-01
-3.44829768e-01 -1.25706828e+00 -6.43061101e-01 9.92548048e-01
2.93766081e-01 -8.81487072e-01 -4.07535762e-01 1.76966801e-01] | [13.581450462341309, 1.8647009134292603] |
95501dab-70d1-4107-a332-a75c1e97fe20 | smoa-sparse-mixture-of-adapters-to-mitigate | 2302.14413 | null | https://arxiv.org/abs/2302.14413v1 | https://arxiv.org/pdf/2302.14413v1.pdf | SMoA: Sparse Mixture of Adapters to Mitigate Multiple Dataset Biases | Recent studies reveal that various biases exist in different NLP tasks, and over-reliance on biases results in models' poor generalization ability and low adversarial robustness. To mitigate datasets biases, previous works propose lots of debiasing techniques to tackle specific biases, which perform well on respective adversarial sets but fail to mitigate other biases. In this paper, we propose a new debiasing method Sparse Mixture-of-Adapters (SMoA), which can mitigate multiple dataset biases effectively and efficiently. Experiments on Natural Language Inference and Paraphrase Identification tasks demonstrate that SMoA outperforms full-finetuning, adapter tuning baselines, and prior strong debiasing methods. Further analysis indicates the interpretability of SMoA that sub-adapter can capture specific pattern from the training data and specialize to handle specific bias. | ['Hua Wu', 'Jing Liu', 'Yan Chen', 'Jing Yan', 'Yanchen Liu'] | 2023-02-28 | null | null | null | null | ['paraphrase-identification'] | ['natural-language-processing'] | [ 1.25889048e-01 -1.98850468e-01 -4.61250573e-01 -5.03742814e-01
-5.75819194e-01 -8.36755931e-01 6.57496691e-01 -3.24017435e-01
-1.90824300e-01 9.77990568e-01 6.19711697e-01 -3.41856033e-01
-3.32808495e-02 -7.99555302e-01 -9.02983487e-01 -5.78870952e-01
8.07812214e-01 4.57816720e-01 -1.89329281e-01 -4.16814685e-01
4.68155503e-01 2.14739919e-01 -1.08810627e+00 4.72312003e-01
1.32028997e+00 6.10397458e-01 -1.26546398e-01 2.95979977e-01
-3.58201325e-01 9.28453922e-01 -9.06230509e-01 -8.26866329e-01
2.07516864e-01 -6.65582716e-02 -7.38280416e-01 -7.33621418e-01
8.60898376e-01 -4.54269141e-01 -5.97560406e-01 1.29616714e+00
7.68317997e-01 -5.64105529e-03 8.53409708e-01 -1.39821172e+00
-1.26616812e+00 8.86427701e-01 -3.41754049e-01 5.18151045e-01
2.39579946e-01 4.50025529e-01 7.15859175e-01 -7.62760460e-01
2.89441109e-01 1.61169064e+00 1.05898821e+00 1.00498939e+00
-1.29763353e+00 -1.29952466e+00 3.85136843e-01 2.79260594e-02
-9.72492099e-01 -4.37487632e-01 9.32000875e-01 -3.68773758e-01
7.99735188e-01 3.88428897e-01 5.57177607e-03 2.06040359e+00
2.54992843e-01 9.70283508e-01 1.41300023e+00 5.08757792e-02
1.39448673e-01 3.31822306e-01 5.73650122e-01 3.41357440e-01
5.21918118e-01 2.72181034e-01 -5.07025242e-01 -6.02771878e-01
4.02358085e-01 1.41697213e-01 -2.10975975e-01 2.55496819e-02
-1.19199598e+00 7.58131504e-01 4.03779358e-01 6.34225644e-03
-1.79578483e-01 3.72124426e-02 5.56085646e-01 5.89836299e-01
5.15166700e-01 8.25969577e-01 -5.39088726e-01 6.69273436e-02
-7.18683243e-01 4.81056064e-01 5.50523937e-01 1.15399909e+00
5.82876444e-01 1.18358657e-01 -7.80059457e-01 1.14047408e+00
-1.61913127e-01 8.79677653e-01 9.25924301e-01 -8.19761097e-01
7.58779943e-01 7.74223447e-01 1.06358081e-01 -1.09647107e+00
-1.88791409e-01 -6.24618351e-01 -1.44627523e+00 -2.19876260e-01
3.58070523e-01 -1.15948237e-01 -8.99568677e-01 2.07355785e+00
3.85750905e-02 1.23810656e-01 8.41612443e-02 8.11343253e-01
9.17933285e-01 4.99068946e-01 2.24440038e-01 3.16296697e-01
1.18421006e+00 -1.04094648e+00 -7.40787029e-01 -8.70647728e-01
3.10200095e-01 -6.56181395e-01 1.69266903e+00 4.26451713e-01
-1.04772079e+00 -8.06381047e-01 -1.08575034e+00 -1.97130412e-01
-4.13797855e-01 -2.73185465e-02 6.11806929e-01 6.50561988e-01
-2.57928163e-01 5.77603996e-01 -1.68082073e-01 5.08793332e-02
7.82217205e-01 4.88065705e-02 -1.93356067e-01 -3.02878588e-01
-1.67460263e+00 1.19335294e+00 1.86834410e-01 -1.83626637e-01
-1.10851133e+00 -1.24158454e+00 -5.97007215e-01 3.64617929e-02
1.37460530e-01 -1.12374806e+00 9.74066019e-01 -1.07851708e+00
-1.18798709e+00 9.02801871e-01 -1.82177618e-01 -5.65845370e-01
7.35805392e-01 -6.93141997e-01 -4.02360290e-01 -4.02731985e-01
5.54524548e-03 4.16796803e-01 1.22167742e+00 -1.12912977e+00
-8.84730816e-02 -2.94237912e-01 2.53401607e-01 -6.61907643e-02
-6.68899000e-01 -2.87498329e-02 1.71938315e-01 -1.24738383e+00
-2.74785519e-01 -7.39205956e-01 -1.13145784e-01 -2.65851617e-01
-5.54580688e-01 -6.25639111e-02 6.15127444e-01 -6.58846498e-01
1.55194736e+00 -1.99373090e+00 3.32815260e-01 -1.01098649e-01
8.05761367e-02 7.56802917e-01 -2.90826380e-01 1.91184908e-01
-1.74427167e-01 2.93929815e-01 -2.89984912e-01 -1.76485568e-01
1.81440309e-01 4.39525962e-01 -1.09602296e+00 2.67090708e-01
1.13439560e-01 9.71518636e-01 -1.06069088e+00 -4.08869833e-01
-2.05558613e-02 -3.37306373e-02 -7.17100739e-01 5.45111895e-01
-1.55085996e-01 2.30480790e-01 -2.52020001e-01 7.40451396e-01
9.69153166e-01 -6.51006177e-02 -2.87248462e-01 -2.35999003e-01
6.36536002e-01 4.28476721e-01 -1.03076911e+00 1.55780053e+00
-6.70057476e-01 2.86980033e-01 4.32256944e-02 -7.05129325e-01
9.74041581e-01 -3.77492383e-02 -4.50797021e-01 -4.93628293e-01
1.54210821e-01 2.65634090e-01 -2.21155882e-01 -7.26245761e-01
5.30313313e-01 -3.34476084e-01 -3.35705787e-01 5.13419509e-01
-6.15947060e-02 -1.96127921e-01 -3.66744578e-01 4.46821183e-01
9.68927383e-01 -1.92857519e-01 2.78981924e-01 -4.25746471e-01
8.08727026e-01 -7.57270977e-02 7.02384949e-01 1.30371153e+00
-3.66161525e-01 6.31308675e-01 3.13641429e-01 -4.73616362e-01
-1.01810944e+00 -9.68339682e-01 -2.13262197e-02 1.35555971e+00
2.13729829e-01 -4.94344421e-02 -6.21778071e-01 -1.06450403e+00
5.33918381e-01 1.16173744e+00 -9.13953602e-01 -8.90719891e-01
-4.60505247e-01 -7.66311109e-01 1.05801141e+00 7.05825269e-01
7.10196972e-01 -1.00263917e+00 3.06871235e-01 -8.53737146e-02
-5.36648273e-01 -6.69474959e-01 -7.89028108e-01 -5.53902015e-02
-7.74794042e-01 -8.84728670e-01 -5.09072661e-01 -4.68471080e-01
6.82987690e-01 4.55668449e-01 1.33137870e+00 -4.19290969e-03
1.26972407e-01 -5.09744138e-02 -5.03834635e-02 -5.95435619e-01
-8.82971346e-01 3.10805708e-01 2.61744022e-01 -3.31839502e-01
7.02319145e-01 -5.93008220e-01 -5.23724735e-01 5.43191314e-01
-9.94543433e-01 -2.86724597e-01 6.82469904e-01 1.14594853e+00
3.68566439e-02 -3.59540701e-01 8.70457292e-01 -1.28931415e+00
1.16088367e+00 -9.04538333e-01 -1.56022549e-01 3.80633980e-01
-8.98939729e-01 3.37159604e-01 9.94928777e-01 -9.43301678e-01
-1.27700627e+00 -6.95292711e-01 -2.61994936e-02 -7.42020488e-01
-3.13760698e-01 1.71394944e-01 -4.93288249e-01 -1.98752526e-02
1.03954136e+00 4.01633054e-01 4.91237417e-02 -6.05197251e-01
3.75555903e-01 1.05742180e+00 6.27107978e-01 -9.92903173e-01
9.70001042e-01 3.83037120e-01 -3.92724693e-01 1.89390522e-03
-1.44946837e+00 -1.16953813e-01 -1.71782643e-01 1.47603929e-01
1.11621819e-01 -1.05945909e+00 -1.67096078e-01 7.63561547e-01
-1.11798060e+00 -1.32728443e-01 -1.17563710e-01 3.26002948e-02
-1.88492492e-01 5.52192986e-01 -5.19500852e-01 -3.84680003e-01
-7.17826366e-01 -8.89470637e-01 7.50037313e-01 1.06735952e-01
-6.75802529e-01 -8.15705657e-01 2.19398066e-01 9.07976270e-01
8.24736655e-01 -2.51292527e-01 1.00384951e+00 -7.97011614e-01
-2.00778320e-01 -2.01300338e-01 -1.84709042e-01 8.40006053e-01
1.38358753e-02 -2.46220559e-01 -1.07875705e+00 -3.70158941e-01
4.84798104e-02 -5.46722531e-01 9.79592562e-01 6.46899119e-02
1.55279207e+00 -7.19354212e-01 -2.17872068e-01 8.45079541e-01
1.01290941e+00 -3.66551399e-01 7.81583130e-01 5.31526625e-01
7.43405044e-01 3.66924107e-01 6.82145238e-01 1.77659318e-01
1.62722901e-01 2.24986196e-01 5.52168429e-01 1.30529359e-01
-5.39360270e-02 -6.91233575e-01 4.87074614e-01 5.96851230e-01
2.74828076e-01 -7.25877583e-02 -5.51859736e-01 7.35470891e-01
-1.68298376e+00 -1.07592309e+00 1.95109705e-03 1.94567537e+00
1.35647345e+00 2.16179267e-01 -1.22291885e-01 -2.63094790e-02
7.86626637e-01 4.46758926e-01 -9.91552234e-01 -5.23128033e-01
-5.11181235e-01 1.18111998e-01 5.83776653e-01 4.04750556e-01
-9.45363700e-01 1.02993715e+00 7.11405849e+00 1.13170838e+00
-7.43361533e-01 1.65560097e-01 4.83019263e-01 -3.10186416e-01
-7.76663542e-01 -8.64832327e-02 -8.46498787e-01 1.04014766e+00
5.10642290e-01 -2.23790616e-01 6.15079701e-01 1.14769077e+00
-2.69716412e-01 4.93271619e-01 -1.14359570e+00 8.33860993e-01
1.85598806e-01 -1.24233341e+00 6.26672924e-01 -3.91464323e-01
8.99226904e-01 -9.39340889e-02 3.01358193e-01 7.86473632e-01
6.61149085e-01 -1.20543385e+00 4.74443644e-01 5.40686190e-01
4.98818010e-01 -6.17916107e-01 7.34913170e-01 8.06730151e-01
-1.84713691e-01 -3.12908083e-01 -8.21174264e-01 -1.73777938e-01
-2.23919138e-01 6.58754528e-01 -4.69964415e-01 2.44233713e-01
6.57339096e-01 5.03565967e-01 -7.76366830e-01 4.83019173e-01
-4.66006905e-01 9.02352870e-01 8.19121487e-04 -1.17686994e-01
7.20667243e-02 2.34398199e-03 7.81707644e-01 1.45950615e+00
8.28846619e-02 -1.30768389e-01 -3.57000202e-01 1.09555316e+00
-3.66477668e-01 -2.38037825e-01 -5.85210383e-01 1.91293389e-01
9.70999479e-01 8.51724267e-01 4.03892338e-01 -6.01484835e-01
-1.76136553e-01 8.61994803e-01 5.67893326e-01 3.22080225e-01
-1.15018034e+00 -4.36337292e-01 1.14236212e+00 -1.16576195e-01
-6.79450482e-02 2.43750930e-01 -6.54641509e-01 -1.57982492e+00
-9.98960924e-04 -1.59464931e+00 6.15656495e-01 -8.66556168e-01
-2.06049395e+00 2.06977978e-01 -4.52079764e-03 -1.01133263e+00
2.35761762e-01 -3.57321948e-01 -7.43823707e-01 9.69589233e-01
-1.57841790e+00 -1.05582130e+00 -4.89860207e-01 7.13102341e-01
5.58546364e-01 -4.98466820e-01 5.87771952e-01 2.11007595e-01
-8.54339600e-01 1.18844640e+00 9.12394896e-02 1.33181125e-01
1.33222449e+00 -1.17019069e+00 4.88383740e-01 8.28036666e-01
-3.19455028e-01 1.27511859e+00 9.72233772e-01 -5.81078410e-01
-1.29341292e+00 -1.25426555e+00 7.86339819e-01 -8.45326543e-01
6.29627347e-01 -1.27596512e-01 -1.32835877e+00 8.03631246e-01
2.58146405e-01 -2.41816372e-01 7.14288712e-01 3.29153687e-01
-1.07034230e+00 -3.61759841e-01 -1.46133983e+00 5.93321443e-01
1.21112132e+00 -7.67900348e-01 -1.37159193e+00 4.14086580e-01
7.08602369e-01 -4.85007226e-01 -5.55096567e-01 5.11248112e-01
3.39452028e-01 -8.77338111e-01 1.39897835e+00 -1.31192017e+00
8.20052087e-01 -6.63210079e-02 -4.35890853e-02 -1.61877072e+00
-6.04222476e-01 -5.88279605e-01 -2.10889608e-01 1.60574305e+00
8.07596967e-02 -9.00765121e-01 6.66170955e-01 7.70429432e-01
1.15445726e-01 -3.25804681e-01 -7.81589210e-01 -9.19509768e-01
6.60491049e-01 -1.55944020e-01 1.18597484e+00 1.43772995e+00
-2.18902454e-01 3.78898025e-01 -8.65043759e-01 2.53071457e-01
7.31981456e-01 3.15851659e-01 1.05763280e+00 -8.61847579e-01
-3.18184346e-01 -5.02488911e-01 1.28559321e-01 -1.12616718e+00
5.65931141e-01 -9.71691906e-01 -2.82248527e-01 -9.53767717e-01
3.27743322e-01 -4.15885746e-01 -4.61036026e-01 4.56947774e-01
-9.56521630e-01 -7.78316557e-02 -2.23047376e-01 3.23096663e-01
-2.30383724e-01 5.89748919e-01 1.07581389e+00 -6.53123021e-01
1.39278337e-01 7.38376454e-02 -1.26249599e+00 7.10200012e-01
9.26232696e-01 -6.25495374e-01 -3.07794720e-01 -7.45910466e-01
5.21135032e-01 -4.42263901e-01 4.94436145e-01 -6.56966507e-01
1.38563737e-01 -4.70143139e-01 2.38352299e-01 -4.12997961e-01
7.02703893e-02 -6.30118489e-01 6.25130609e-02 4.35307711e-01
-8.16485465e-01 -1.67900547e-01 2.43196711e-01 6.71172380e-01
-1.12926543e-01 -5.50694585e-01 8.52406800e-01 -2.45949015e-01
-2.77983844e-01 8.67528170e-02 -2.21915573e-01 6.35051310e-01
7.06900179e-01 1.37122735e-01 -8.24447632e-01 -1.40209883e-01
-2.94039607e-01 3.67969692e-01 4.46185321e-01 7.07810283e-01
3.24025393e-01 -1.52673733e+00 -9.61398184e-01 2.60220438e-01
9.76271927e-02 5.44820800e-02 4.47931170e-01 3.27736139e-01
-3.10516134e-02 3.04332674e-01 -3.10344756e-01 -1.98263749e-01
-1.25924432e+00 9.69316781e-01 3.28274012e-01 -2.83482164e-01
-1.08166076e-01 1.23079491e+00 3.18835676e-01 -7.96377003e-01
2.98001528e-01 -3.32971931e-01 1.45652145e-01 -1.50491640e-01
5.59386909e-01 6.27729774e-01 -1.14661127e-01 -1.63022920e-01
-3.28790724e-01 3.87787312e-01 -4.73408997e-01 5.98095655e-01
9.91810083e-01 -9.83210281e-02 -1.81306839e-01 1.08052842e-01
9.70501006e-01 2.30198786e-01 -1.00584829e+00 -6.42017603e-01
-4.82353777e-01 -7.50907123e-01 -1.37609154e-01 -1.07134104e+00
-8.62738550e-01 9.76070523e-01 7.64191896e-02 -6.60491735e-02
9.14407909e-01 -3.13276350e-01 9.69738245e-01 4.72346842e-01
2.04793721e-01 -1.02537572e+00 1.43573299e-01 5.40183127e-01
1.20339668e+00 -1.31413174e+00 1.84224859e-01 -1.68105900e-01
-8.94011557e-01 6.22263372e-01 1.06247807e+00 -4.32166994e-01
2.61258930e-01 2.06807241e-01 3.27699073e-02 2.03650415e-01
-8.97311866e-01 4.78802681e-01 2.02914208e-01 6.06464148e-01
-7.48726428e-02 -1.04002208e-01 -1.74608424e-01 1.19326913e+00
-2.22667307e-01 -6.57146871e-02 2.09758982e-01 5.34450114e-01
-2.32277095e-01 -9.62441683e-01 -6.02165282e-01 5.07939100e-01
-4.44214225e-01 -3.64170134e-01 -7.76794434e-01 5.36877215e-01
4.05839793e-02 6.94700301e-01 -2.83481836e-01 -5.44783711e-01
5.18861592e-01 2.46095866e-01 3.77269179e-01 -3.12409043e-01
-1.20419097e+00 -8.04459929e-01 2.50133555e-02 -5.74616492e-01
-7.89937079e-02 -5.45470357e-01 -4.91783142e-01 -7.20455348e-01
-2.66068041e-01 6.81730434e-02 1.38436049e-01 8.11360896e-01
6.86349869e-01 3.66361171e-01 5.25872111e-01 -2.68140852e-01
-1.46068752e+00 -1.32553601e+00 -1.49097428e-01 9.42207456e-01
4.71485078e-01 -5.79483569e-01 -9.79434073e-01 -3.83537740e-01] | [10.401604652404785, 7.7227935791015625] |
a85af9c5-c38e-43ef-a549-d3dc8ae90890 | multilingual-entity-and-relation-extraction | null | null | https://aclanthology.org/2021.eacl-main.166 | https://aclanthology.org/2021.eacl-main.166.pdf | Multilingual Entity and Relation Extraction Dataset and Model | We present a novel dataset and model for a multilingual setting to approach the task of Joint Entity and Relation Extraction. The SMiLER dataset consists of 1.1 M annotated sentences, representing 36 relations, and 14 languages. To the best of our knowledge, this is currently both the largest and the most comprehensive dataset of this type. We introduce HERBERTa, a pipeline that combines two independent BERT models: one for sequence classification, and the other for entity tagging. The model achieves micro F1 81.49 for English on this dataset, which is close to the current SOTA on CoNLL, SpERT. | ['Piotr Andruszkiewicz', 'Micha{\\l} Sat{\\l}awa', 'Helena Skowronska', 'Klaudia Firl{\\k{a}}g', 'Alessandro Seganti'] | 2021-04-01 | null | null | null | eacl-2021-2 | ['joint-entity-and-relation-extraction'] | ['natural-language-processing'] | [-9.20058638e-02 4.76267815e-01 -6.41781330e-01 -3.07954341e-01
-9.41908121e-01 -8.60179484e-01 6.64727688e-01 4.80928063e-01
-7.00038850e-01 1.10897136e+00 8.24760273e-02 -4.69968766e-01
8.30033273e-02 -2.10620821e-01 -7.56882310e-01 -1.22066252e-01
-1.85102791e-01 1.09184468e+00 3.90972108e-01 -3.09342891e-01
-1.84468284e-01 2.63432324e-01 -8.99034441e-01 5.53553283e-01
8.69027197e-01 7.52993286e-01 1.48219600e-01 7.58420169e-01
-7.16663674e-02 1.38658762e+00 -7.64135182e-01 -8.99591565e-01
-2.59535342e-01 -1.22024722e-01 -1.60331249e+00 -4.87773746e-01
3.88842821e-01 6.74714819e-02 -2.48009250e-01 6.78724587e-01
5.24401069e-01 -6.10960424e-02 5.36010861e-01 -1.03768337e+00
-8.71533692e-01 1.52819800e+00 -3.99922878e-01 4.59923536e-01
6.83749616e-01 -2.85089523e-01 1.32884455e+00 -7.64460921e-01
1.56900024e+00 9.82671261e-01 7.47454643e-01 3.81585658e-01
-9.54231679e-01 -4.98780668e-01 -1.27684712e-01 2.32101306e-01
-1.67204499e+00 -7.89469242e-01 -1.34816729e-02 -4.50078547e-01
1.94789386e+00 4.14961465e-02 3.02524596e-01 1.00372744e+00
2.42934182e-01 9.05847013e-01 1.00205278e+00 -7.69295931e-01
-2.51507044e-01 1.71085000e-01 2.33986601e-01 5.77626348e-01
1.68415040e-01 -3.40427577e-01 -8.02265286e-01 -1.46861337e-02
3.41634490e-02 -8.94044578e-01 -2.55664468e-01 -3.71945016e-02
-1.37444389e+00 3.79133672e-01 4.12178971e-03 5.70047498e-01
-3.25908870e-01 -2.08815336e-01 6.24901950e-01 3.62571478e-01
6.33404732e-01 5.77589512e-01 -1.19257140e+00 -3.23922127e-01
-5.50682962e-01 2.01083347e-01 1.32269490e+00 1.45823503e+00
2.54802316e-01 -5.21530986e-01 -3.08498628e-02 9.07046974e-01
2.23639175e-01 4.10754144e-01 2.35096186e-01 -7.07399070e-01
9.93877053e-01 4.55277413e-01 -7.47808665e-02 -4.81755197e-01
-5.76146007e-01 -2.61318266e-01 -3.85041922e-01 -5.75057149e-01
5.46065569e-01 -3.64217818e-01 -5.74227750e-01 1.60301232e+00
3.62927288e-01 1.43605443e-02 4.82230663e-01 4.16434884e-01
1.41254413e+00 5.20064116e-01 3.76899242e-01 -1.70150042e-01
1.63810837e+00 -1.07594097e+00 -9.52941835e-01 -3.30161184e-01
1.10491908e+00 -1.07173419e+00 2.98982561e-01 1.12099476e-01
-1.06608593e+00 -2.82848775e-01 -8.80314469e-01 -3.90830398e-01
-7.96084106e-01 4.36897635e-01 1.04106712e+00 4.13977712e-01
-8.53812635e-01 2.52431273e-01 -7.45900631e-01 -9.60846186e-01
1.47308968e-02 1.60777435e-01 -8.20968330e-01 1.68559119e-01
-1.59848952e+00 1.47565448e+00 9.79898155e-01 1.44743808e-02
-3.78306031e-01 -5.82919955e-01 -9.42661047e-01 -3.72523874e-01
4.14256454e-01 -4.40510452e-01 1.53483069e+00 -4.79393564e-02
-1.24999273e+00 1.24740791e+00 -2.95933753e-01 -8.83587182e-01
1.27280623e-01 -6.70773685e-01 -9.17954326e-01 -1.02209240e-01
3.50830942e-01 5.12648165e-01 -9.80358720e-02 -9.08281505e-01
-7.16960430e-01 -1.25635311e-01 -9.37248543e-02 8.16882476e-02
1.03577577e-01 7.37340331e-01 -5.52994192e-01 -4.29747075e-01
-4.12174195e-01 -9.42405045e-01 -1.39052540e-01 -8.11308861e-01
-6.64774299e-01 -4.52754855e-01 3.44594985e-01 -1.32735455e+00
1.64018142e+00 -1.79326880e+00 2.01255620e-01 -1.32320642e-01
4.17865217e-02 4.37528312e-01 -1.28943056e-01 8.78841519e-01
-3.22968811e-01 3.74930322e-01 -1.78043370e-03 -4.28768247e-01
1.73060093e-02 2.74992794e-01 2.51051374e-02 2.43131265e-01
5.11924326e-01 1.25483477e+00 -1.10095537e+00 -7.02872097e-01
-2.60987967e-01 2.52100706e-01 1.17295757e-02 2.48656094e-01
-1.15697496e-01 4.04837817e-01 -1.63549393e-01 6.38616025e-01
3.71853679e-01 -1.59823880e-01 8.72604609e-01 -5.09113781e-02
-3.21218580e-01 9.51576173e-01 -1.04456508e+00 1.90374088e+00
-3.25203806e-01 7.71589994e-01 5.45461997e-02 -5.51895261e-01
6.75182998e-01 7.59537816e-01 3.68805230e-01 -1.80254787e-01
1.37942031e-01 5.70966363e-01 1.65821642e-01 -5.91991186e-01
8.14248025e-01 3.98604840e-01 -4.08855349e-01 9.61165801e-02
6.12811267e-01 -1.16509475e-01 7.57846415e-01 3.80227059e-01
1.36147535e+00 5.30358255e-01 9.74623024e-01 -1.30052805e-01
5.47042727e-01 1.87020347e-01 6.58020616e-01 4.85634297e-01
-9.72399786e-02 5.04261702e-02 4.98775870e-01 -2.95705140e-01
-9.69347239e-01 -7.57561743e-01 -3.27798992e-01 9.93737340e-01
-2.44901448e-01 -7.56951213e-01 -7.24626303e-01 -8.68095696e-01
-5.64603657e-02 6.00903809e-01 -2.98342735e-01 2.65071452e-01
-7.37808526e-01 -5.28923929e-01 1.22231722e+00 4.01347935e-01
2.67782509e-01 -1.19104397e+00 -2.55190320e-02 2.86059946e-01
-5.62265873e-01 -1.82483423e+00 -3.17912936e-01 5.00006974e-01
-2.38271549e-01 -1.06596780e+00 -1.39754027e-01 -1.07518649e+00
3.68072577e-02 -6.11313343e-01 1.84822392e+00 -3.42214465e-01
-2.52806731e-02 -2.15313897e-01 -5.62895119e-01 -3.69092166e-01
-7.23888457e-01 8.18607450e-01 -5.16260304e-02 -7.36502945e-01
6.47120357e-01 -4.77061383e-02 1.10848740e-01 -1.40807256e-01
-4.02655363e-01 -9.72183794e-02 8.14445257e-01 6.29232228e-01
4.64873314e-01 -1.59980521e-01 4.92124140e-01 -1.45617044e+00
2.63631225e-01 -6.58662498e-01 -2.70470023e-01 7.18359709e-01
-3.19620073e-01 -6.94925264e-02 4.31641310e-01 -9.56572145e-02
-1.24168479e+00 2.22150832e-01 -4.28699672e-01 3.80758226e-01
-4.57087040e-01 7.24320948e-01 -2.51566589e-01 3.14649969e-01
5.30705571e-01 -1.93724141e-01 -6.76955879e-01 -7.84279644e-01
7.76301026e-01 9.97611940e-01 8.13220322e-01 -6.15395010e-01
3.21333587e-01 -1.66998103e-01 -9.62234661e-02 -8.26052666e-01
-1.14784300e+00 -7.82720208e-01 -1.11707127e+00 1.98455438e-01
1.09080815e+00 -1.00353408e+00 -5.25763929e-01 5.49866855e-01
-1.61789155e+00 -4.58952576e-01 -2.07364440e-01 5.29432714e-01
-3.64807606e-01 3.05868685e-01 -1.33201337e+00 -5.75380683e-01
-4.54687804e-01 -7.33360469e-01 1.05357230e+00 -2.07434624e-01
-6.63337350e-01 -1.02884424e+00 3.88879895e-01 5.01938820e-01
-1.12734847e-01 1.37020350e-01 5.28723359e-01 -1.07075417e+00
-2.29658812e-01 -1.24662660e-01 -2.27753460e-01 -3.98986936e-02
-5.88570237e-02 9.00612622e-02 -7.85350919e-01 -3.39154340e-02
-6.72685981e-01 -6.68726087e-01 6.67593002e-01 -3.59875076e-02
2.56440133e-01 -1.30647182e-01 -6.17490470e-01 1.91587538e-01
1.10017061e+00 2.45645657e-01 4.79428053e-01 5.11452913e-01
7.92213380e-01 4.83747989e-01 8.96351218e-01 2.09823370e-01
9.86450016e-01 7.62180567e-01 -1.04001373e-01 -2.69736070e-02
-3.88998240e-01 -4.52732295e-01 2.67666221e-01 1.31424439e+00
-9.39551368e-03 -4.51187491e-01 -1.43576503e+00 7.60666132e-01
-1.88565445e+00 -7.44152725e-01 -3.78986388e-01 1.58053875e+00
1.44789028e+00 -3.43404114e-02 -1.32659718e-01 -2.64156222e-01
5.71028709e-01 1.07451200e-01 -1.81807261e-02 -4.60988641e-01
-3.99230808e-01 5.08533716e-01 6.74186945e-01 6.07556105e-01
-1.55361259e+00 1.53980601e+00 7.07059240e+00 9.12811637e-01
-5.26728511e-01 3.07200611e-01 2.51523584e-01 3.87134939e-01
8.11306164e-02 4.08380449e-01 -1.43836105e+00 2.01211914e-01
1.58168507e+00 1.01443470e-01 1.11836165e-01 6.31264389e-01
-3.25236857e-01 -3.96450981e-02 -1.04468453e+00 5.36537349e-01
9.91203636e-03 -1.18865287e+00 -3.01865906e-01 -1.82080656e-01
6.32488549e-01 4.25029695e-01 -5.78826368e-01 4.94870007e-01
9.27019775e-01 -9.73929226e-01 9.32823360e-01 6.54168129e-01
7.69268513e-01 -7.27597237e-01 1.22098756e+00 3.13597292e-01
-1.35648239e+00 5.44657826e-01 1.33343160e-01 1.21283695e-01
5.67620695e-01 6.24075532e-01 -8.00307930e-01 8.91806185e-01
6.86755955e-01 1.09856284e+00 -6.79547846e-01 7.42108583e-01
-5.98674536e-01 6.04259133e-01 -1.97832555e-01 -3.46083157e-02
-1.73959300e-01 2.64810801e-01 5.71858823e-01 1.84409487e+00
-1.41916722e-01 -7.18360320e-02 4.36512858e-01 6.02550730e-02
-3.23607028e-01 5.15296578e-01 -6.76398456e-01 -2.86575317e-01
9.76326466e-01 1.37246656e+00 -4.18091565e-01 -4.97940809e-01
-5.48727751e-01 9.73499179e-01 1.16352224e+00 1.61990285e-01
-5.81324100e-01 -5.29681921e-01 2.61798829e-01 -5.42666018e-01
2.76302248e-01 -3.64133656e-01 -4.45351824e-02 -1.23611069e+00
-1.17583692e-01 -1.02092135e+00 7.20585048e-01 -4.68623549e-01
-1.36998510e+00 8.98115218e-01 -1.06323361e-01 -6.79580450e-01
-4.63008881e-01 -7.62809992e-01 2.20421001e-01 8.70745122e-01
-1.45282710e+00 -1.64587069e+00 3.04427862e-01 1.59295481e-02
1.70668796e-01 -1.73554480e-01 1.08916891e+00 7.20436513e-01
-9.08698618e-01 5.56055784e-01 8.98428902e-04 5.32989860e-01
1.09541750e+00 -1.49295712e+00 7.60971189e-01 9.50429976e-01
2.90199280e-01 7.98815429e-01 5.52001953e-01 -8.58857512e-01
-1.06987107e+00 -9.03208852e-01 2.04935145e+00 -1.05203128e+00
1.22099948e+00 -5.34390748e-01 -6.73389792e-01 1.21747112e+00
7.00946927e-01 -3.35797310e-01 8.47448587e-01 4.27895457e-01
-4.83853877e-01 2.56369770e-01 -9.24724460e-01 2.08824217e-01
1.15509272e+00 -6.53615952e-01 -6.16558909e-01 3.97276372e-01
8.24224830e-01 -8.75187039e-01 -1.88583422e+00 4.87334013e-01
5.38808584e-01 -4.35581863e-01 6.57479465e-01 -7.23318696e-01
5.08672178e-01 -2.52043933e-01 -2.91933030e-01 -1.36656594e+00
-1.81060463e-01 -7.27351606e-01 -5.86422384e-01 1.90433788e+00
1.17554832e+00 -6.12777591e-01 2.85199076e-01 2.23870367e-01
-1.77637234e-01 -6.28456831e-01 -9.09236789e-01 -1.00233948e+00
1.66533589e-01 -3.52074176e-01 6.66131854e-01 1.22050548e+00
3.20548832e-01 9.64472830e-01 -2.69321173e-01 1.19286790e-01
2.82088786e-01 7.98676908e-02 6.87554479e-01 -1.00220549e+00
-4.65996772e-01 -1.94125727e-01 -2.59905517e-01 -9.79638278e-01
6.08433604e-01 -1.03987873e+00 2.29124561e-01 -1.60293710e+00
3.15471619e-01 -5.89726746e-01 1.43100545e-01 7.61898994e-01
-2.86871523e-01 1.67826861e-01 -1.18466265e-01 2.14423627e-01
-8.16343904e-01 1.23559318e-01 6.51606679e-01 1.23430312e-01
3.40368785e-02 -4.61004674e-01 -4.59602475e-01 6.45346224e-01
9.78913844e-01 -5.04489124e-01 2.60446556e-02 -5.92679441e-01
3.94587964e-01 -4.66008447e-02 -3.37775439e-01 -5.36771476e-01
1.41181007e-01 -9.47454870e-02 9.77341905e-02 -7.49727428e-01
4.37075570e-02 -4.31937963e-01 2.32888237e-01 2.52823979e-01
-4.61104602e-01 1.77192196e-01 2.41068140e-01 1.31438196e-01
-3.62256885e-01 -2.21009538e-01 3.99856985e-01 -9.42861736e-02
-1.06582975e+00 8.95935819e-02 -3.86169851e-01 4.99229908e-01
8.98165405e-01 6.00696146e-01 -8.09155107e-01 -1.62438471e-02
-7.48686016e-01 2.77541369e-01 1.05185457e-01 6.49048686e-01
-1.16036668e-01 -1.16690397e+00 -9.76800501e-01 -4.18166578e-01
4.38536257e-01 -2.33503014e-01 -2.43383154e-01 7.93803453e-01
-4.74991381e-01 7.62689710e-01 9.12500173e-02 -1.34882078e-01
-1.30971801e+00 5.60532570e-01 3.01392321e-02 -1.02549040e+00
-2.04322472e-01 8.82095933e-01 -6.08175874e-01 -8.82427573e-01
-2.39098936e-01 -1.29858628e-01 -6.77760720e-01 1.26571223e-01
3.69154602e-01 4.59208637e-01 3.91878605e-01 -1.15600634e+00
-8.52072001e-01 1.55136541e-01 -3.53891432e-01 -1.12563588e-01
1.14790261e+00 -2.48191774e-01 -7.93817937e-01 7.07352161e-01
8.84620011e-01 4.20652062e-01 -3.20107400e-01 -3.39871347e-01
8.69115233e-01 1.51334763e-01 -4.17959899e-01 -1.03873980e+00
-5.30859947e-01 2.98537225e-01 1.91155020e-02 1.16319790e-01
6.68972671e-01 4.72294807e-01 1.01255476e+00 4.69596446e-01
4.54194874e-01 -1.09963620e+00 -5.24810076e-01 1.15264261e+00
5.83886325e-01 -9.95006025e-01 1.24505579e-01 -9.62469697e-01
-6.48398876e-01 5.36612630e-01 6.40156567e-01 3.59372854e-01
4.80100572e-01 6.16222024e-01 2.58180678e-01 -2.32199922e-01
-1.06139934e+00 -5.87006569e-01 5.31402767e-01 6.94779277e-01
1.29163516e+00 3.14655334e-01 -6.46845341e-01 6.18731618e-01
-5.38657904e-01 -3.93534005e-02 7.52479136e-02 1.06994700e+00
1.18784029e-02 -1.60164177e+00 2.65828788e-01 1.51620254e-01
-9.65590239e-01 -5.52956760e-01 -7.03527033e-01 8.87063444e-01
2.39624754e-01 1.27886450e+00 -2.10716307e-01 -1.84507400e-01
6.61066055e-01 2.10750625e-01 6.09728038e-01 -9.49688971e-01
-6.76962674e-01 -4.06309038e-01 1.17084348e+00 -3.08381975e-01
-6.28602743e-01 -8.14999640e-01 -1.26027095e+00 -2.75215805e-01
-5.00239313e-01 3.68596911e-01 2.75842637e-01 8.96758497e-01
3.76388609e-01 4.26099360e-01 1.66630581e-01 -3.73462617e-01
-2.62455605e-02 -1.36702180e+00 -4.45361733e-01 2.26352528e-01
-1.14539161e-01 -4.29622471e-01 2.04269141e-01 1.57417879e-01] | [9.713955879211426, 9.153372764587402] |
d86c2c13-d64c-46f3-931a-0f5912f1e01f | a-color-temperature-based-high-speed | 2108.13656 | null | https://arxiv.org/abs/2108.13656v1 | https://arxiv.org/pdf/2108.13656v1.pdf | A color temperature-based high-speed decolorization: an empirical approach for tone mapping applications | Grayscale images are fundamental to many image processing applications like data compression, feature extraction, printing and tone mapping. However, some image information is lost when converting from color to grayscale. In this paper, we propose a light-weight and high-speed image decolorization method based on human perception of color temperatures. Chromatic aberration results from differential refraction of light depending on its wavelength. It causes some rays corresponding to cooler colors (like blue, green) to converge before the warmer colors (like red, orange). This phenomena creates a perception of warm colors "advancing" toward the eye, while the cool colors to be "receding" away. In this proposed color to gray conversion model, we implement a weighted blending function to combine red (perceived warm) and blue (perceived cool) channel. Our main contribution is threefold: First, we implement a high-speed color processing method using exact pixel by pixel processing, and we report a $5.7\times$ speed up when compared to other new algorithms. Second, our optimal color conversion method produces luminance in images that are comparable to other state of the art methods which we quantified using the objective metrics (E-score and C2G-SSIM) and a subjective user study. Third, we demonstrate that an effective luminance distribution can be achieved using our algorithm by using global and local tone mapping applications. | ['Masayuki Ikebe', 'Yafei Ou', 'Prasoon Ambalathankandy'] | 2021-08-31 | null | null | null | null | ['tone-mapping'] | ['computer-vision'] | [ 4.49861526e-01 -7.53790319e-01 3.19400817e-01 -3.92772108e-02
-2.05685541e-01 -6.12367988e-01 2.16489345e-01 -1.39462322e-01
-4.82597262e-01 6.87798560e-01 -1.44308418e-01 -2.32795388e-01
4.03394818e-01 -1.04596579e+00 -3.88990670e-01 -8.08525741e-01
-3.24616488e-03 -5.16448915e-01 3.09765458e-01 -3.05833429e-01
6.37793243e-01 3.66494119e-01 -1.70508265e+00 2.89415449e-01
1.20009077e+00 1.24176395e+00 3.67589332e-02 1.02732110e+00
-1.50283113e-01 4.85012382e-01 -8.82238805e-01 -2.86869228e-01
5.20539284e-01 -9.61692929e-01 -2.54889131e-01 -7.52996877e-02
3.09678823e-01 -5.23338318e-01 1.63857965e-03 1.54793763e+00
3.46153438e-01 -3.51246931e-02 5.84891200e-01 -1.07339120e+00
-1.24035037e+00 -7.54202232e-02 -1.10375261e+00 -1.26119912e-01
2.36010626e-01 2.51320094e-01 3.00462455e-01 -6.44522309e-01
3.73936474e-01 9.91879582e-01 3.12396288e-01 4.70722675e-01
-1.20496523e+00 -8.82181168e-01 -4.45328176e-01 3.47740263e-01
-1.45935988e+00 -1.35812908e-01 9.08712387e-01 -4.64341864e-02
4.45122331e-01 8.07879329e-01 7.34372735e-01 2.57002413e-01
5.73646009e-01 1.53697759e-01 2.05655241e+00 -8.38558018e-01
2.48499528e-01 4.08077091e-01 -1.11826979e-01 7.52904713e-01
3.12658012e-01 3.43320847e-01 -4.43475366e-01 1.77101746e-01
7.61933923e-01 -1.44885629e-01 -4.78956342e-01 2.33591214e-01
-8.90261769e-01 2.39221811e-01 4.56142277e-01 2.49858409e-01
-1.42425731e-01 3.77689838e-01 -1.42041683e-01 4.54136282e-01
2.91551262e-01 4.18307275e-01 1.01579413e-01 -2.56003916e-01
-7.80273259e-01 -3.12885374e-01 4.25618142e-01 6.33070111e-01
8.19132984e-01 -9.42655466e-03 -1.16396815e-01 9.40855384e-01
2.22327918e-01 1.05758309e+00 4.68405396e-01 -8.69766533e-01
1.30374795e-02 3.77846837e-01 3.67068589e-01 -1.12085056e+00
-1.16483562e-01 1.33573189e-01 -8.92878294e-01 1.28720796e+00
2.23632529e-01 -1.35791019e-01 -9.31542993e-01 1.39816105e+00
-4.62868586e-02 -3.00687123e-02 4.78242747e-02 1.10539258e+00
4.76090074e-01 9.48171079e-01 -1.45083427e-01 -4.88840938e-01
1.54671621e+00 -6.45156860e-01 -1.01374221e+00 2.03315541e-01
-8.61010551e-02 -1.41508615e+00 1.36081052e+00 9.12415206e-01
-1.19639909e+00 -8.42076957e-01 -1.34525585e+00 -1.70853183e-01
-4.93563712e-01 1.44351766e-01 3.47660065e-01 1.37776744e+00
-1.17360985e+00 4.28985417e-01 -3.02056789e-01 -2.33911261e-01
9.05652195e-02 -5.61590306e-02 3.92011590e-02 -3.85237224e-02
-9.01634932e-01 7.44229198e-01 -7.80908689e-02 3.31430547e-02
-1.63269743e-01 -5.98220170e-01 -1.30981982e-01 -1.63456038e-01
3.50401439e-02 -1.79235160e-01 5.67050159e-01 -1.36553788e+00
-2.05795431e+00 7.94175446e-01 -1.70235962e-01 -2.73443889e-02
4.94317532e-01 -1.36561051e-01 -9.05900180e-01 4.03976709e-01
-4.96281415e-01 4.97329593e-01 9.01512444e-01 -1.74522460e+00
-6.70935273e-01 -1.76698685e-01 -2.18937516e-01 2.69417524e-01
-3.96421313e-01 1.24464639e-01 -7.05869138e-01 -6.39534712e-01
1.68996602e-01 -7.20021248e-01 1.23648688e-01 4.40254360e-01
-4.41368550e-01 3.94514918e-01 1.02850080e+00 -7.54420698e-01
1.45648754e+00 -2.33618832e+00 -3.92731696e-01 4.22847986e-01
1.76149577e-01 3.65167707e-01 1.31808808e-02 3.89213413e-01
-1.87269431e-02 9.68023688e-02 -2.15599909e-01 8.12253281e-02
-2.73908406e-01 -3.40416193e-01 -2.63697654e-01 5.39492846e-01
-1.44129440e-01 4.01633680e-01 -5.51831663e-01 -3.57521564e-01
2.83593893e-01 5.01350403e-01 -2.88032621e-01 -3.11647132e-02
3.11510891e-01 1.24823630e-01 9.27545205e-02 7.41900921e-01
1.47182024e+00 3.33960474e-01 -5.14162332e-02 -5.48135340e-01
-5.90155602e-01 -5.80967128e-01 -1.15669549e+00 1.11834788e+00
-4.68738586e-01 8.39647532e-01 -3.70165473e-03 -1.74757957e-01
1.10666311e+00 -8.03148076e-02 1.32200241e-01 -1.23405719e+00
2.48347551e-01 2.58484453e-01 -1.94754452e-01 -4.81738478e-01
7.90724695e-01 -2.60375053e-01 1.81238621e-01 5.47515869e-01
-7.07903683e-01 -5.52355409e-01 7.31436610e-02 6.33253679e-02
5.26048005e-01 -1.40146529e-02 -8.71515833e-03 -2.79080570e-01
5.02226889e-01 -1.93213522e-01 1.60849214e-01 4.73665982e-01
-3.24110270e-01 6.95852697e-01 4.67657775e-01 -7.44576156e-02
-9.77207959e-01 -1.07675791e+00 -1.18365437e-01 6.47049367e-01
9.16169643e-01 -1.82180349e-02 -9.21967208e-01 1.33387476e-01
-2.14000314e-01 7.70663202e-01 -3.22600245e-01 -2.90213525e-01
-2.66574740e-01 -7.26498008e-01 3.84194285e-01 4.89696264e-02
1.06034184e+00 -9.40583229e-01 -7.36840367e-01 -1.35987416e-01
4.93751429e-02 -6.86142445e-01 -4.97492880e-01 -2.40884125e-01
-7.42399275e-01 -9.42185819e-01 -7.44772375e-01 -6.22391999e-01
8.04909825e-01 5.37040353e-01 7.53559411e-01 2.54114002e-01
-8.55042934e-01 2.64939547e-01 -5.66157103e-01 -4.38554704e-01
-3.93312871e-01 -9.11162794e-01 -1.19075052e-01 3.63647163e-01
2.23517209e-01 -2.65004277e-01 -1.23403466e+00 2.34550580e-01
-1.14502823e+00 2.84235775e-01 6.48648500e-01 4.12771106e-01
5.13481915e-01 5.00520945e-01 2.92291981e-03 -7.09508240e-01
8.67362380e-01 2.79481113e-01 -8.70721340e-01 2.63657212e-01
-8.64343286e-01 -1.35442615e-01 8.43131423e-01 -2.61105686e-01
-1.35292149e+00 -2.96418756e-01 1.98644757e-01 1.02901785e-02
2.74120420e-02 -1.68556899e-01 1.58199981e-01 -5.69251001e-01
7.49028325e-01 4.00873661e-01 2.54650593e-01 -2.58193970e-01
5.36445796e-01 8.29787731e-01 7.14740038e-01 -1.81205034e-01
1.00257969e+00 9.07293737e-01 1.00461520e-01 -9.73897457e-01
2.76937127e-01 -1.61688000e-01 -5.61003610e-02 -5.30295491e-01
1.01830745e+00 -5.44719875e-01 -1.15690899e+00 9.39654708e-01
-8.37138653e-01 -2.31136963e-01 -9.67832431e-02 3.58650655e-01
-1.26942396e-01 5.89937925e-01 -6.61386669e-01 -9.42619324e-01
-3.79090250e-01 -9.00700033e-01 6.01475835e-01 7.22009838e-01
2.98104882e-01 -7.62447178e-01 -1.02569133e-01 -1.27720414e-03
8.58331978e-01 2.86780030e-01 9.43822801e-01 7.99517393e-01
-6.73261404e-01 -1.02317870e-01 -6.86501205e-01 5.91347635e-01
4.18954223e-01 4.84433860e-01 -8.93879533e-01 -1.96475178e-01
2.02925541e-02 -8.27226974e-03 8.75475943e-01 2.11735040e-01
1.27067614e+00 5.71786277e-02 -1.43536001e-01 8.05677295e-01
2.07669711e+00 6.53001487e-01 1.36638463e+00 2.89211929e-01
3.41816634e-01 2.38006428e-01 5.38816392e-01 2.59340554e-01
-7.90478960e-02 4.85565394e-01 1.95588410e-01 -8.66063774e-01
-7.08177924e-01 3.80725004e-02 3.71586889e-01 5.28934896e-01
-3.32682282e-01 -2.74177700e-01 -3.35995436e-01 -3.31063159e-02
-9.96500373e-01 -9.02407348e-01 -4.33013976e-01 2.46103358e+00
1.10561466e+00 -8.92945528e-02 -2.37531230e-01 8.59402940e-02
8.40491176e-01 -1.76911518e-01 -3.51067305e-01 -6.84102595e-01
-3.33401978e-01 6.78440034e-01 1.01282787e+00 6.64117455e-01
-7.31912732e-01 7.24453986e-01 6.46103191e+00 8.81922901e-01
-1.64672768e+00 -9.68800187e-02 9.41986620e-01 1.76452640e-02
-4.41434324e-01 -2.12541297e-02 -1.77304462e-01 7.02277482e-01
4.29403692e-01 -5.64825870e-02 8.00942063e-01 2.54897624e-01
3.40166092e-01 -7.94108152e-01 -4.41081583e-01 1.37205637e+00
2.69119740e-01 -8.84621859e-01 -8.09183568e-02 -1.11203052e-01
7.97133029e-01 -6.53652966e-01 5.89395046e-01 -3.76721978e-01
4.61792648e-02 -7.61365592e-01 7.46792912e-01 7.39393175e-01
1.40505147e+00 -6.49452507e-01 2.04828382e-01 -3.04321826e-01
-1.09203172e+00 5.03867790e-02 -5.12951851e-01 9.03617442e-02
-4.23366912e-02 6.90389037e-01 -2.89113045e-01 4.30677652e-01
6.73010468e-01 -1.61184352e-02 -6.80030048e-01 1.08842421e+00
-1.62249878e-01 4.01833802e-01 -2.20745847e-01 -2.82982945e-01
-1.09437257e-01 -7.40477562e-01 2.29039833e-01 1.04021525e+00
6.64022744e-01 4.27453786e-01 -3.95154595e-01 1.16866970e+00
6.85444474e-02 1.41599178e-01 -2.19958529e-01 1.05612770e-01
3.44600648e-01 1.36510682e+00 -9.63145852e-01 -4.29003865e-01
-4.48338062e-01 1.52211571e+00 -5.76065063e-01 7.42785871e-01
-8.60814631e-01 -1.30353463e+00 4.60727364e-01 -8.41631740e-02
-2.11954221e-01 -2.38547757e-01 -7.11693704e-01 -7.58400798e-01
-3.57978009e-02 -6.93996549e-01 -7.40568116e-02 -1.15259099e+00
-8.98289859e-01 5.57159781e-01 -4.10657734e-01 -1.60417163e+00
6.35753989e-01 -8.13594162e-01 -6.98866725e-01 1.28463733e+00
-1.79251301e+00 -5.49766779e-01 -6.75140381e-01 7.40524769e-01
-2.26710793e-02 1.05155684e-01 6.64481938e-01 3.27319950e-01
-1.94926113e-01 6.07624233e-01 4.00848061e-01 -1.16002694e-01
1.14277554e+00 -1.09602976e+00 -2.48512346e-02 1.20815647e+00
-2.45254129e-01 4.43427086e-01 6.95076525e-01 -4.87097085e-01
-1.63882875e+00 -5.91568887e-01 3.32263261e-01 1.16893388e-01
1.96990281e-01 -9.92745385e-02 -6.29249692e-01 -3.01506221e-01
5.79463482e-01 -1.76440254e-01 5.51703215e-01 -4.92194682e-01
-4.23394144e-01 -6.68346703e-01 -1.47294331e+00 7.09759831e-01
4.04547364e-01 -4.74354148e-01 1.13001868e-01 -5.96626252e-02
4.20501798e-01 -1.40718430e-01 -4.21397865e-01 -1.35413975e-01
7.92410553e-01 -1.47341776e+00 6.96983695e-01 3.19100499e-01
4.30550635e-01 -8.27420831e-01 -7.32086450e-02 -1.27589631e+00
-2.49322504e-01 -8.18985283e-01 6.51179612e-01 9.34638202e-01
3.64185214e-01 -9.16732430e-01 4.06095684e-01 4.68835920e-01
1.90417096e-02 -1.38379753e-01 -6.27327621e-01 -5.68460584e-01
-1.43686146e-01 -1.18719541e-01 4.66511130e-01 8.39392245e-01
1.07110590e-01 -3.12458396e-01 -5.18712223e-01 2.24034667e-01
8.52440298e-01 2.60950536e-01 4.92366582e-01 -7.20099628e-01
-2.49530375e-01 -5.88669956e-01 -3.45413059e-01 -8.30150425e-01
-9.04957175e-01 -3.64612490e-01 5.50197028e-02 -1.31527174e+00
1.59963563e-01 -4.53406930e-01 -4.00966763e-01 1.76261291e-01
-2.53822386e-01 1.04466224e+00 3.96108836e-01 1.09165400e-01
-1.48780122e-01 1.96384132e-01 1.48731852e+00 -6.42264113e-02
-3.50672722e-01 -4.61385697e-01 -7.83653378e-01 4.35247242e-01
7.39301503e-01 6.33977503e-02 -2.51071036e-01 -3.77647966e-01
2.79958129e-01 -1.34511769e-01 2.58914441e-01 -1.30423999e+00
1.53138995e-01 -2.28902817e-01 6.91535354e-01 -2.04919323e-01
2.33904153e-01 -8.92237008e-01 6.53368682e-02 8.64809453e-01
9.74494517e-02 -6.15678281e-02 6.42982125e-02 3.34650159e-01
-2.29597285e-01 -4.76835072e-02 1.22659028e+00 -2.65098158e-02
-8.92524242e-01 -9.73315164e-02 -4.12322521e-01 -4.20924604e-01
1.26975405e+00 -6.64873362e-01 -7.47434020e-01 -5.29315233e-01
-1.07707120e-01 -4.34525609e-01 8.17129135e-01 2.19322532e-01
8.49915445e-01 -1.39397705e+00 -4.09744710e-01 4.99632269e-01
2.89726164e-02 -8.91837239e-01 5.00109553e-01 7.39672661e-01
-1.43934596e+00 -9.74875987e-02 -5.37471473e-01 -3.04445952e-01
-1.47593057e+00 5.04622281e-01 1.94206610e-01 4.96861696e-01
-4.49057549e-01 8.89978170e-01 6.23832867e-02 4.68489081e-01
-5.67712560e-02 -3.31295788e-01 -6.73689321e-02 -3.27283084e-01
7.48465121e-01 6.61352694e-01 -1.11793742e-01 -2.21935645e-01
-1.14806101e-01 1.07687986e+00 2.59059995e-01 -4.10134852e-01
8.86496663e-01 -3.33614141e-01 -6.20197058e-01 2.24066570e-01
1.21174634e+00 6.35204792e-01 -9.84462559e-01 1.87318072e-01
-7.81592309e-01 -1.16590345e+00 3.91985886e-02 -1.13036144e+00
-1.10993493e+00 8.83046091e-01 1.52335751e+00 6.13385320e-01
1.88262439e+00 -4.06875044e-01 8.11078072e-01 -1.78173691e-01
2.98889369e-01 -1.59626353e+00 2.90482342e-01 -1.46177500e-01
5.71024060e-01 -9.80635583e-01 6.90821856e-02 -5.58960319e-01
-6.74988031e-01 1.31048846e+00 3.83511752e-01 -5.56040183e-02
4.96365160e-01 4.28295583e-01 5.75838983e-01 7.56297261e-02
-2.56018072e-01 -1.54056758e-01 1.51044115e-01 5.70104539e-01
4.96237874e-01 2.62541920e-01 -6.92555189e-01 -5.62048070e-02
-7.61882365e-02 2.40453463e-02 8.98623526e-01 7.10265458e-01
-7.26162195e-01 -1.10908163e+00 -8.62423718e-01 1.61956027e-01
-4.86940205e-01 -2.34933496e-01 -1.71631247e-01 4.01055098e-01
4.06877398e-01 1.19726944e+00 2.21049353e-01 -4.85757530e-01
1.79802775e-01 -3.02710146e-01 6.03626192e-01 5.58532029e-02
-3.56649980e-02 3.10665429e-01 -3.55345368e-01 -6.17950678e-01
-3.94748241e-01 -6.39695451e-02 -1.31732976e+00 -6.12886608e-01
-1.79367766e-01 -1.48810878e-01 9.99677002e-01 2.18506694e-01
1.47524446e-01 6.23050869e-01 9.52731431e-01 -3.73950571e-01
1.14694290e-01 -5.54970205e-01 -1.03379953e+00 7.52660692e-01
3.00285041e-01 -2.22460985e-01 -4.27615196e-01 3.20333898e-01] | [10.83838176727295, -2.4159348011016846] |
ebb684ed-ec00-4610-ade1-74d9ad0a5f7f | volta-vision-language-transformer-with-weakly | 2210.04135 | null | https://arxiv.org/abs/2210.04135v2 | https://arxiv.org/pdf/2210.04135v2.pdf | VoLTA: Vision-Language Transformer with Weakly-Supervised Local-Feature Alignment | Vision-language pre-training (VLP) has recently proven highly effective for various uni- and multi-modal downstream applications. However, most existing end-to-end VLP methods use high-resolution image-text box data to perform well on fine-grained region-level tasks, such as object detection, segmentation, and referring expression comprehension. Unfortunately, such high-resolution images with accurate bounding box annotations are expensive to collect and use for supervision at scale. In this work, we propose VoLTA (Vision-Language Transformer with weakly-supervised local-feature Alignment), a new VLP paradigm that only utilizes image-caption data but achieves fine-grained region-level image understanding, eliminating the use of expensive box annotations. VoLTA adopts graph optimal transport-based weakly-supervised alignment on local image patches and text tokens to germinate an explicit, self-normalized, and interpretable low-level matching criterion. In addition, VoLTA pushes multi-modal fusion deep into the uni-modal backbones during pre-training and removes fusion-specific transformer layers, further reducing memory requirements. Extensive experiments on a wide range of vision- and vision-language downstream tasks demonstrate the effectiveness of VoLTA on fine-grained applications without compromising the coarse-grained downstream performance, often outperforming methods using significantly more caption and box annotations. | ['Rama Chellappa', 'Yann Lecun', 'Hardik Shah', 'Jiachen Zhu', 'Sayan Nag', 'Li Jing', 'Shraman Pramanick'] | 2022-10-09 | null | null | null | null | ['referring-expression'] | ['computer-vision'] | [ 3.70896190e-01 1.70861498e-01 -3.26068640e-01 -5.97796321e-01
-1.41153491e+00 -6.81114674e-01 5.37537575e-01 6.84212372e-02
-3.72001857e-01 3.26422393e-01 2.51462907e-01 -2.41175249e-01
3.77150774e-01 -6.61332786e-01 -1.07377267e+00 -5.90971291e-01
6.61069155e-01 5.43810248e-01 4.75509107e-01 -2.80890226e-01
-1.94506887e-02 3.36943984e-01 -1.48294902e+00 7.98494101e-01
7.34516799e-01 1.22648454e+00 3.77115041e-01 6.33978963e-01
-3.23692024e-01 5.36502779e-01 -1.18482903e-01 -6.26903951e-01
3.44625622e-01 -2.01367095e-01 -8.16331446e-01 2.52868682e-01
1.09267843e+00 -3.67679685e-01 -3.18749174e-02 1.07236624e+00
2.67512351e-01 -1.85620971e-02 5.86456835e-01 -1.18212140e+00
-8.17256212e-01 3.25595796e-01 -9.02709842e-01 -8.54018237e-03
8.65952149e-02 4.82249558e-01 1.23261976e+00 -1.00259662e+00
5.46850324e-01 1.48230994e+00 7.13226497e-01 5.37753999e-01
-1.31254578e+00 -3.86976451e-01 3.29527646e-01 -3.40270475e-02
-1.35521758e+00 -3.88158858e-01 6.37066424e-01 -3.61906648e-01
1.15012896e+00 2.73364950e-02 4.47902441e-01 9.58025217e-01
1.18194349e-01 8.42269897e-01 1.23426008e+00 -4.06528801e-01
-1.40436083e-01 -7.29250088e-02 1.26529768e-01 1.25887179e+00
-6.81816787e-02 -8.02138150e-02 -4.92603838e-01 1.76638633e-01
9.70102251e-01 -1.16640896e-01 -7.35045224e-02 -2.94579387e-01
-1.23951375e+00 7.32739687e-01 6.74312115e-01 2.05182478e-01
-3.13043505e-01 3.70505720e-01 4.48468596e-01 -5.58029376e-02
5.88699281e-01 -3.42754088e-02 -5.22695839e-01 1.84285402e-01
-1.01560628e+00 -4.85206768e-02 2.49183461e-01 1.03189600e+00
1.15913391e+00 -4.76550870e-03 -6.14216328e-01 7.58495569e-01
3.98973167e-01 6.93934143e-01 3.02281469e-01 -1.11232162e+00
6.20537162e-01 7.85019815e-01 -2.07171142e-01 -6.83420539e-01
-2.85219997e-01 -3.23390126e-01 -8.31260443e-01 1.91288814e-01
5.60124159e-01 3.52332503e-01 -1.57176495e+00 1.72639608e+00
4.36547667e-01 -1.48462474e-01 5.96773103e-02 1.11921179e+00
9.10538554e-01 6.34423316e-01 2.94539660e-01 2.17859536e-01
1.79502583e+00 -1.48129082e+00 -3.05427909e-01 -5.19165754e-01
6.41581416e-01 -7.19318807e-01 1.64330661e+00 -1.04563404e-02
-1.10633588e+00 -6.90179884e-01 -6.23686314e-01 -8.94198000e-01
-4.49678302e-01 2.09460631e-02 6.48545802e-01 3.62101793e-01
-1.35943639e+00 2.18409188e-02 -7.57678807e-01 -3.32213134e-01
8.84518564e-01 2.77362645e-01 -6.14250422e-01 -3.98650557e-01
-8.38407457e-01 8.28140259e-01 2.30112344e-01 5.16791232e-02
-1.00012469e+00 -8.65769565e-01 -1.18573248e+00 -5.21131791e-02
2.45784715e-01 -1.18518579e+00 1.21361995e+00 -1.09478164e+00
-1.29004681e+00 1.32513750e+00 -5.10197520e-01 -5.06199360e-01
4.95533973e-01 -1.74474150e-01 2.46869192e-01 2.97763288e-01
4.90211964e-01 1.38112700e+00 1.13544321e+00 -1.12550545e+00
-6.88302875e-01 -5.17736316e-01 2.37156898e-01 3.54469806e-01
-1.34395510e-01 -1.16653442e-01 -8.97505224e-01 -4.44519877e-01
2.34648399e-02 -5.99502742e-01 -2.62992650e-01 4.33922678e-01
-3.45604688e-01 -4.92530674e-01 9.97156680e-01 -6.12411797e-01
5.06127834e-01 -2.08939338e+00 -9.99853201e-03 -1.38802975e-01
2.95522988e-01 2.02109590e-02 -5.22152007e-01 -1.14311203e-01
1.50531009e-01 -7.09058270e-02 -3.14682275e-01 -6.38852417e-01
-1.05965436e-01 4.13725317e-01 -3.41721177e-01 4.58047628e-01
4.72740322e-01 1.49868524e+00 -7.79438376e-01 -9.75511432e-01
6.86399102e-01 5.36992729e-01 -5.33622026e-01 2.02427551e-01
-6.64422631e-01 4.72009152e-01 -4.53952402e-01 9.63071942e-01
4.46027845e-01 -4.39610332e-01 -5.28928459e-01 -8.06428552e-01
-1.08906202e-01 -2.02113762e-01 -4.12389934e-01 2.14487886e+00
-8.40554416e-01 6.59540713e-01 3.07269692e-01 -1.04194319e+00
6.07422292e-01 -9.52308029e-02 3.51741642e-01 -9.63494599e-01
1.74294218e-01 3.75371017e-02 -3.84529084e-01 -4.43287343e-01
3.62578213e-01 -1.16866834e-01 -3.84012938e-01 7.97367692e-02
1.31887630e-01 -5.17334640e-01 1.38219535e-01 3.21234375e-01
8.02578986e-01 3.35779965e-01 1.72886029e-01 -6.72759861e-02
7.02427924e-01 3.25645298e-01 3.27990472e-01 6.07143939e-01
-3.35514009e-01 8.67297411e-01 1.06201440e-01 -1.07603908e-01
-1.11305952e+00 -1.25085497e+00 -6.32487833e-02 1.53429687e+00
3.09081525e-01 -2.90436238e-01 -9.46122050e-01 -6.58612728e-01
-5.12779877e-02 5.70043504e-01 -5.01325428e-01 6.12073950e-02
-4.53558892e-01 -2.85964310e-01 7.80095637e-01 7.51584053e-01
7.55021214e-01 -8.94412756e-01 -3.07440132e-01 -8.47490802e-02
-5.23965478e-01 -1.71731269e+00 -7.82562017e-01 1.76529780e-01
-6.97116017e-01 -7.48736382e-01 -8.87084544e-01 -9.84964967e-01
8.51234555e-01 4.11467195e-01 1.08605373e+00 -2.07971677e-01
-3.82519990e-01 6.11089885e-01 -5.98357096e-02 -1.37152329e-01
-1.99195206e-01 -1.75316811e-01 -3.11178148e-01 -8.92749336e-03
1.83705240e-01 -3.28960270e-01 -5.83725095e-01 2.64798552e-01
-7.70901680e-01 4.51767743e-01 9.01127100e-01 1.01739502e+00
1.14308023e+00 -3.58654797e-01 2.93778956e-01 -5.74789464e-01
1.97549462e-01 9.26634446e-02 -6.44009888e-01 4.42093343e-01
-3.97973418e-01 9.64112431e-02 5.81217527e-01 -2.91236728e-01
-1.10914695e+00 2.21839160e-01 -2.44578451e-01 -8.50895524e-01
-3.03031117e-01 -1.42734572e-02 -3.42290163e-01 -3.27375650e-01
5.28292000e-01 4.31759149e-01 -4.64170575e-02 -1.79317847e-01
1.14584696e+00 5.37632346e-01 1.09822202e+00 -7.50158191e-01
8.74636114e-01 8.92453730e-01 -1.04405932e-01 -7.69574642e-01
-1.48265231e+00 -6.52238309e-01 -6.96489692e-01 -5.55688031e-02
1.42800963e+00 -1.19752562e+00 -6.52550459e-01 4.41360801e-01
-1.19803631e+00 -6.29933178e-01 -4.39397782e-01 2.69897692e-02
-8.68255079e-01 3.50914270e-01 -5.75753212e-01 -4.92414445e-01
-7.11475134e-01 -1.15002358e+00 1.95358467e+00 2.34637782e-01
6.90210313e-02 -8.97341371e-01 -4.02823448e-01 1.15833950e+00
2.14975536e-01 1.51123792e-01 7.96326280e-01 -1.12749293e-01
-8.96010041e-01 2.15111777e-01 -9.30833280e-01 5.29073119e-01
-2.30496541e-01 -3.42541903e-01 -1.20320344e+00 -8.06837231e-02
-3.79032224e-01 -6.84149861e-01 1.06387985e+00 4.95193541e-01
1.33621871e+00 -8.15899521e-02 -2.78416038e-01 9.24592078e-01
1.36706543e+00 -5.03055394e-01 3.74219894e-01 9.24412310e-02
1.26552975e+00 6.39528096e-01 8.63820314e-01 -1.33793175e-01
6.70181453e-01 6.14193678e-01 5.13125837e-01 -5.73430717e-01
-5.79470396e-01 -4.45264786e-01 4.82920080e-01 3.61057222e-01
1.91612989e-01 -1.75966308e-01 -8.39506269e-01 8.28207493e-01
-1.96132445e+00 -6.72195375e-01 -1.95804879e-01 1.56354463e+00
8.65797579e-01 1.11277677e-01 1.32535165e-02 -4.22212094e-01
5.74677169e-01 1.71555102e-01 -7.28071272e-01 -3.14363390e-01
-3.33298504e-01 2.19565064e-01 7.44936168e-01 5.35735011e-01
-1.16259038e+00 1.63368809e+00 5.08999538e+00 1.16651857e+00
-1.12545049e+00 4.73802418e-01 8.10675144e-01 9.45532620e-02
-3.77313673e-01 -7.11533725e-02 -8.88651371e-01 3.68837006e-02
6.14759266e-01 2.64194578e-01 6.42483458e-02 9.28071499e-01
3.19273233e-01 -9.73420665e-02 -1.25258791e+00 1.27784455e+00
1.56760156e-01 -1.57436478e+00 2.84877956e-01 -1.66212589e-01
7.75551796e-01 5.35624444e-01 1.14955321e-01 2.40515262e-01
2.78138101e-01 -1.08378458e+00 9.18387353e-01 3.11704010e-01
1.20747221e+00 -4.27262306e-01 4.38793927e-01 1.93886817e-01
-1.39792812e+00 6.54990301e-02 -2.12901533e-01 2.30208993e-01
3.58680576e-01 5.44656813e-01 -6.67740405e-01 4.46692318e-01
9.04903054e-01 6.31216407e-01 -5.12005568e-01 3.92926812e-01
-2.35739291e-01 3.28562617e-01 -5.78380525e-01 3.29240888e-01
7.02903986e-01 -7.32744187e-02 3.02519381e-01 1.39210618e+00
-6.37972206e-02 7.35551193e-02 3.95505190e-01 1.03690565e+00
-1.97083518e-01 -9.20454413e-02 -5.54433227e-01 1.04719117e-01
9.11725089e-02 1.49855852e+00 -7.41767704e-01 -4.00761515e-01
-6.16834521e-01 1.25188792e+00 3.67220879e-01 3.24226081e-01
-8.36133957e-01 2.60998979e-02 7.22790301e-01 1.93729281e-01
4.50602770e-01 -1.71631694e-01 -4.94027078e-01 -1.12122750e+00
-6.43370450e-02 -6.20836496e-01 4.19306457e-01 -1.08715618e+00
-1.50814939e+00 4.13487762e-01 -4.78938632e-02 -8.83579850e-01
-2.47841686e-01 -7.65766323e-01 -4.50773537e-01 8.35072696e-01
-1.83574820e+00 -2.13304067e+00 -5.18367827e-01 1.13733673e+00
8.38529825e-01 2.49385357e-01 6.10502243e-01 1.11437239e-01
-3.83610964e-01 7.04800725e-01 -3.73190552e-01 2.53441483e-01
7.15615034e-01 -1.15350938e+00 3.27152103e-01 8.53734553e-01
2.85485864e-01 3.83841217e-01 3.83237422e-01 -4.06109214e-01
-1.34067595e+00 -1.59288538e+00 5.80226541e-01 -6.20824695e-01
6.95926070e-01 -6.28237247e-01 -7.95297801e-01 6.65684998e-01
1.45318553e-01 4.56304193e-01 1.81322977e-01 -1.11747697e-01
-6.99299514e-01 -2.20681801e-01 -1.12761879e+00 6.73198402e-01
1.21268237e+00 -9.57739651e-01 -5.64279258e-01 6.98354006e-01
9.76865828e-01 -4.90343899e-01 -9.06990767e-01 4.65096533e-01
2.95740932e-01 -7.86980152e-01 1.22579515e+00 -3.13157529e-01
4.22260076e-01 -4.53256279e-01 -2.48814777e-01 -6.73162103e-01
-9.44230855e-02 -4.07016128e-01 1.42626375e-01 1.34627318e+00
2.71861315e-01 -3.79760951e-01 8.57809126e-01 5.44662535e-01
-3.75304282e-01 -8.91763270e-01 -8.32780302e-01 -5.09292543e-01
5.77880852e-02 -6.26217902e-01 8.12122300e-02 7.31304705e-01
-5.03449798e-01 6.43236995e-01 -8.00671652e-02 2.28703246e-01
8.91683400e-01 4.54291999e-01 8.98123980e-01 -8.29510748e-01
-1.58846065e-01 -5.07306218e-01 -4.69131678e-01 -1.34656417e+00
4.22042876e-01 -9.94225442e-01 2.64595926e-01 -1.82518554e+00
1.59222066e-01 -2.66866356e-01 -2.78369570e-03 8.59501600e-01
7.08277598e-02 7.30491757e-01 9.19464082e-02 6.25958145e-02
-8.17208588e-01 4.86020952e-01 1.48592627e+00 -4.14325327e-01
8.06774572e-02 -3.61329019e-01 -7.08840907e-01 1.01750743e+00
4.63903695e-01 -1.93973586e-01 -5.08660555e-01 -6.64925873e-01
-2.15231091e-01 -1.79687917e-01 7.93703735e-01 -5.88184774e-01
2.37163723e-01 -1.44615129e-01 4.09472167e-01 -8.72007132e-01
5.71154535e-01 -7.02092826e-01 -4.61380303e-01 6.30650371e-02
-2.56635755e-01 -1.74450129e-01 3.22010040e-01 7.02136576e-01
-2.42984921e-01 1.98322892e-01 1.07309186e+00 -1.65266901e-01
-1.19648170e+00 5.21161020e-01 7.58931711e-02 2.38745078e-01
1.07325876e+00 -4.78340119e-01 -4.08380926e-01 -1.24520667e-01
-5.99964738e-01 4.23494220e-01 6.88801587e-01 4.65096056e-01
7.73028195e-01 -8.87269855e-01 -7.06773639e-01 1.92598298e-01
4.41067100e-01 6.94609821e-01 4.82222676e-01 9.42812681e-01
-4.30903435e-01 5.59620559e-01 -1.45410866e-01 -1.24695659e+00
-1.33008444e+00 3.74107450e-01 4.31102604e-01 -2.59780109e-01
-9.30667341e-01 1.14042211e+00 1.02476215e+00 -3.18344235e-01
6.87398091e-02 -6.35634661e-01 1.97877541e-01 -3.00414860e-01
2.86597192e-01 -2.80683786e-01 1.96190458e-02 -8.35272491e-01
-3.95408452e-01 1.06157434e+00 -2.60418385e-01 -2.15094700e-01
1.00532234e+00 -5.00995994e-01 -8.05812031e-02 1.92934051e-01
1.25509417e+00 -4.02083069e-01 -1.47048008e+00 -4.07429844e-01
-2.38912404e-01 -2.42253438e-01 3.21675599e-01 -7.24361598e-01
-1.05296040e+00 1.17983329e+00 4.33403701e-01 -4.28078562e-01
1.22665501e+00 4.83690947e-01 1.08055985e+00 3.28196138e-01
3.66843164e-01 -8.86455894e-01 2.48823524e-01 5.38631856e-01
8.17715645e-01 -1.48802197e+00 -3.20889145e-01 -5.29395282e-01
-6.94672525e-01 8.14282298e-01 9.31828797e-01 1.78449228e-02
2.39153981e-01 3.04760754e-01 1.99929386e-01 -1.32400662e-01
-6.25338316e-01 -6.01543605e-01 4.47641313e-01 9.71042335e-01
1.13555051e-01 -1.68781325e-01 2.18570992e-01 3.32893610e-01
-8.08719397e-02 -2.34118462e-01 1.69786345e-02 4.29386109e-01
-4.21205193e-01 -7.99207211e-01 -3.40914011e-01 2.88289398e-01
-3.28289658e-01 -4.11195457e-01 -3.26390743e-01 7.31966555e-01
2.34243289e-01 7.87372768e-01 2.56596982e-01 -1.80215891e-02
2.08408833e-01 -1.16183229e-01 6.74113333e-01 -7.04139352e-01
-4.85818475e-01 1.68715611e-01 5.21535650e-02 -9.27033663e-01
-5.40603697e-01 -2.75473028e-01 -1.67379093e+00 -6.51401281e-02
-1.59199268e-01 -2.46099725e-01 7.14316547e-01 1.18082607e+00
4.23662752e-01 5.17891228e-01 6.39925599e-02 -9.70804036e-01
-2.58203030e-01 -6.90521002e-01 -2.08377391e-02 4.36975360e-01
4.53574061e-01 -5.57311594e-01 -1.48583213e-02 3.29729259e-01] | [10.381969451904297, 1.3462852239608765] |
6ba1bad4-8db5-46f2-a2cf-c05bf7a3301c | temporal-collaborative-ranking-via | 1908.05435 | null | https://arxiv.org/abs/1908.05435v1 | https://arxiv.org/pdf/1908.05435v1.pdf | Temporal Collaborative Ranking Via Personalized Transformer | The collaborative ranking problem has been an important open research question as most recommendation problems can be naturally formulated as ranking problems. While much of collaborative ranking methodology assumes static ranking data, the importance of temporal information to improving ranking performance is increasingly apparent. Recent advances in deep learning, especially the discovery of various attention mechanisms and newer architectures in addition to widely used RNN and CNN in natural language processing, have allowed us to make better use of the temporal ordering of items that each user has engaged with. In particular, the SASRec model, inspired by the popular Transformer model in natural languages processing, has achieved state-of-art results in the temporal collaborative ranking problem and enjoyed more than 10x speed-up when compared to earlier CNN/RNN-based methods. However, SASRec is inherently an un-personalized model and does not include personalized user embeddings. To overcome this limitation, we propose a Personalized Transformer (SSE-PT) model, outperforming SASRec by almost 5% in terms of NDCG@10 on 5 real-world datasets. Furthermore, after examining some random users' engagement history and corresponding attention heat maps used during the inference stage, we find our model is not only more interpretable but also able to focus on recent engagement patterns for each user. Moreover, our SSE-PT model with a slight modification, which we call SSE-PT++, can handle extremely long sequences and outperform SASRec in ranking results with comparable training speed, striking a balance between performance and speed requirements. Code and data are open sourced at https://github.com/wuliwei9278/SSE-PT. | ['Cho-Jui Hsieh', 'James Sharpnack', 'Shuqing Li', 'Liwei Wu'] | 2019-08-15 | null | null | null | null | ['collaborative-ranking'] | ['graphs'] | [-1.66160807e-01 -5.17197609e-01 -3.41872483e-01 -5.03878295e-01
-6.06151164e-01 -7.34108567e-01 8.41123939e-01 2.58874863e-01
-6.18282318e-01 3.43826771e-01 8.42803955e-01 -3.11085075e-01
-5.34575522e-01 -7.46847034e-01 -5.16134262e-01 -2.28778780e-01
-2.98872828e-01 6.06095433e-01 3.92735079e-02 -5.22574425e-01
3.44156414e-01 1.09817028e-01 -1.57105148e+00 4.86823708e-01
8.28870177e-01 7.83163071e-01 1.66567549e-01 6.80982649e-01
-7.51051959e-03 7.27053702e-01 -3.17380846e-01 -5.64755559e-01
2.43940353e-01 -2.15552822e-02 -9.55465674e-01 -6.50285363e-01
4.31935430e-01 -4.44769055e-01 -6.37929320e-01 6.44650519e-01
7.03239977e-01 5.21725237e-01 3.35869491e-01 -8.52865577e-01
-1.20979154e+00 1.00820529e+00 -3.09480697e-01 3.85182768e-01
3.63143295e-01 -2.60247350e-01 1.70871949e+00 -8.97016764e-01
3.81863117e-01 8.94503891e-01 5.65716147e-01 3.93885374e-01
-1.00608599e+00 -5.28652489e-01 4.33975130e-01 3.26047510e-01
-1.01033056e+00 -1.85699463e-01 6.57395303e-01 -2.10525423e-01
9.42088842e-01 4.05016899e-01 6.31375372e-01 1.27054930e+00
7.11508170e-02 1.12150764e+00 9.13108468e-01 -2.26924673e-01
-8.34590420e-02 -2.23229840e-01 6.55035377e-01 4.61640716e-01
-1.24221314e-02 2.73150280e-02 -5.31350195e-01 -1.25571385e-01
5.57532489e-01 6.40663326e-01 -1.62446573e-01 -1.82783350e-01
-1.30030179e+00 7.60024071e-01 6.22556925e-01 4.23917502e-01
-2.61527210e-01 3.77904266e-01 6.11779094e-01 5.80344260e-01
7.88405716e-01 7.62349606e-01 -7.63618648e-01 -3.62591505e-01
-7.22024322e-01 4.15161967e-01 5.42708695e-01 7.02366590e-01
2.38469869e-01 -3.51280093e-01 -4.87638831e-01 1.15722132e+00
1.71119809e-01 1.70404345e-01 6.86473966e-01 -7.15017736e-01
2.11572170e-01 4.82021421e-01 2.34133780e-01 -9.47445452e-01
-3.84149075e-01 -9.34645355e-01 -6.67072713e-01 -3.21964920e-01
2.61273205e-01 4.02000127e-03 -6.15272522e-01 1.82880819e+00
-2.03531772e-01 1.12535655e-01 -2.45162413e-01 7.92236209e-01
6.34790659e-01 6.94755852e-01 -7.18939602e-02 2.76858490e-02
1.19922376e+00 -1.25460470e+00 -4.57866430e-01 -1.90693185e-01
5.50456464e-01 -5.37338555e-01 1.50825655e+00 3.40916604e-01
-1.01556981e+00 -6.54579341e-01 -8.59212279e-01 -3.23159337e-01
-4.43266720e-01 -3.31288911e-02 1.14292347e+00 3.21373910e-01
-1.22010088e+00 1.00732291e+00 -6.79760218e-01 -4.54657167e-01
3.28467697e-01 4.92495179e-01 -1.07840821e-01 -1.23347044e-01
-1.48969126e+00 8.61786783e-01 -5.46029173e-02 1.87725127e-01
-5.97237408e-01 -8.88400972e-01 -3.95111620e-01 3.19918811e-01
3.85667950e-01 -6.94877565e-01 1.51068747e+00 -8.76680732e-01
-1.67009592e+00 5.29070139e-01 -1.72326848e-01 -3.56444448e-01
4.08690810e-01 -8.96806121e-01 -5.26836932e-01 -4.79682773e-01
-1.65199012e-01 2.29839548e-01 2.66543984e-01 -7.65326917e-01
-6.28158033e-01 -2.56065309e-01 3.50035578e-01 3.02424818e-01
-7.60136366e-01 1.80000767e-01 -7.21726000e-01 -6.98294640e-01
-3.04556906e-01 -9.33099449e-01 -2.63047487e-01 -4.61810380e-01
6.60102591e-02 -6.17826700e-01 2.47165322e-01 -4.66386110e-01
1.62380004e+00 -1.87235141e+00 1.99271813e-01 2.11442243e-02
2.71344095e-01 4.02719349e-01 -4.82965142e-01 7.14197457e-01
-1.35057168e-02 3.14950913e-01 2.03119278e-01 -4.26390409e-01
2.03454778e-01 5.27894050e-02 -4.80516881e-01 1.70850903e-01
-1.49889410e-01 1.29560149e+00 -1.35757351e+00 1.07405819e-01
5.74420653e-02 4.12429452e-01 -8.81582260e-01 1.56703249e-01
-1.88354328e-01 9.77016613e-02 -4.19798851e-01 3.24430138e-01
1.77429125e-01 -4.35140669e-01 4.41350669e-01 -5.86733073e-02
-1.90157101e-01 6.36424303e-01 -6.12284422e-01 1.78923345e+00
-7.54570544e-01 6.38632953e-01 -5.08979380e-01 -7.06460595e-01
6.25661969e-01 2.00904876e-01 6.35589302e-01 -1.04220223e+00
8.08010176e-02 7.40567222e-02 3.72982249e-02 -2.56209493e-01
1.01640403e+00 2.94852495e-01 -1.14123695e-01 6.77418053e-01
-1.78239584e-01 5.44043362e-01 2.79243648e-01 5.29095113e-01
1.31735349e+00 8.09281841e-02 -9.31817740e-02 -2.47731909e-01
7.10326433e-02 -3.68291229e-01 5.10829687e-01 9.79640424e-01
-6.65363222e-02 5.19209385e-01 3.55645537e-01 -4.87554342e-01
-1.01804650e+00 -1.00182664e+00 3.72339487e-02 1.76747739e+00
5.86293004e-02 -6.74936116e-01 -1.89416111e-02 -7.05554307e-01
-1.95505619e-02 5.72994053e-01 -7.51292646e-01 -1.94911033e-01
-6.22211099e-01 -7.90186942e-01 2.32456222e-01 7.22593069e-01
1.89364716e-01 -1.16339278e+00 -8.13194960e-02 4.23116297e-01
-1.75527409e-01 -6.48623645e-01 -8.08181524e-01 9.36071128e-02
-9.32491839e-01 -9.07033443e-01 -6.78296685e-01 -5.47836900e-01
3.16127092e-01 4.87169772e-01 1.50121522e+00 3.26957643e-01
8.37731808e-02 3.82421702e-01 -7.31572151e-01 -2.35246465e-01
3.04395258e-01 2.46445492e-01 1.40280053e-01 -9.28223133e-02
5.13473570e-01 -6.15334690e-01 -8.65231216e-01 2.34608114e-01
-8.43629479e-01 -6.91476315e-02 5.35275042e-01 8.83340478e-01
3.19492817e-01 -1.14022136e-01 7.97575057e-01 -1.24152386e+00
1.02672517e+00 -5.77043951e-01 -1.79881662e-01 1.41208917e-01
-1.07544470e+00 1.10878803e-01 7.79983938e-01 -4.77177739e-01
-7.71928191e-01 -5.03831208e-01 -3.05407315e-01 -2.10738882e-01
3.52928072e-01 9.27268863e-01 2.88352489e-01 5.96479893e-01
5.46178102e-01 3.27347666e-02 -2.51185596e-01 -7.47189462e-01
5.33806801e-01 5.91733634e-01 2.89869040e-01 -5.64701796e-01
5.42198718e-01 2.48873532e-01 -6.30276859e-01 -2.86138088e-01
-1.31080472e+00 -5.77862620e-01 -3.78520578e-01 -5.89804165e-03
4.72657621e-01 -7.61538804e-01 -7.17643380e-01 3.42293769e-01
-8.73878658e-01 -4.62846935e-01 -9.08023492e-02 3.81630689e-01
-1.64529458e-01 3.49399298e-01 -9.17227924e-01 -5.83433926e-01
-5.87000191e-01 -7.39464164e-01 6.38739586e-01 9.43398178e-02
-1.82483375e-01 -1.11245513e+00 3.34883362e-01 3.77956003e-01
8.89850140e-01 -1.33544430e-01 1.06329274e+00 -8.77562106e-01
-3.36626112e-01 -3.71375829e-01 -2.15221033e-01 3.02718610e-01
1.58969834e-02 -3.89097482e-02 -7.85563469e-01 -4.42758054e-01
-6.25588477e-01 -7.96525553e-02 1.06448030e+00 2.74210721e-01
1.48918462e+00 -2.71856695e-01 -1.12441711e-01 2.68481642e-01
1.25932860e+00 7.23109543e-02 6.60937607e-01 4.19497907e-01
6.92374408e-01 3.46984893e-01 5.96944153e-01 3.15375835e-01
5.79229474e-01 8.97501826e-01 4.99100089e-01 -2.78703831e-02
1.47781581e-01 -4.01143730e-01 6.48172259e-01 1.18849206e+00
-3.75839412e-01 -5.79053581e-01 -5.97184896e-01 4.48253930e-01
-2.12348056e+00 -1.22137392e+00 -9.95579511e-02 2.51791835e+00
7.05585063e-01 1.86104476e-01 2.16658726e-01 -5.43462373e-02
4.67790842e-01 3.67559940e-01 -4.74576592e-01 -5.46432614e-01
1.70752332e-01 4.80733931e-01 3.97857040e-01 4.32278663e-01
-1.05753279e+00 9.13233817e-01 6.04382277e+00 6.33155584e-01
-1.00357449e+00 1.50837481e-01 5.07817447e-01 -4.57977742e-01
-5.03429770e-01 -2.60241419e-01 -4.89213496e-01 5.48478484e-01
1.03046489e+00 -2.24859238e-01 6.74536586e-01 7.04301298e-01
2.45655373e-01 4.28403199e-01 -1.35215187e+00 8.55727017e-01
-7.07319081e-02 -1.30465341e+00 8.28648433e-02 5.12677655e-02
7.96258807e-01 5.02681732e-01 3.43369722e-01 7.72301972e-01
6.47879601e-01 -1.01128054e+00 5.68627119e-01 6.25549853e-01
5.71808755e-01 -7.28678465e-01 9.45698023e-01 9.80299115e-02
-1.15261781e+00 -3.88516545e-01 -5.28568089e-01 -2.80903041e-01
-2.72180722e-03 7.42548525e-01 -3.66413265e-01 6.78946614e-01
7.82687724e-01 1.22525275e+00 -6.18976176e-01 1.05294752e+00
-2.20805243e-01 8.26967716e-01 6.03785850e-02 -2.45994434e-01
2.21310243e-01 -2.70634651e-01 2.88221449e-01 1.24492848e+00
2.99836904e-01 2.84686983e-02 -1.17069632e-02 2.52388626e-01
-4.55564946e-01 1.82255879e-01 -3.79656106e-01 -3.50667655e-01
3.96178097e-01 1.33643758e+00 -3.60616922e-01 -3.23935658e-01
-4.85182941e-01 1.01865411e+00 7.74378181e-01 3.73952031e-01
-9.41929042e-01 -3.09743017e-01 8.86409760e-01 1.51611894e-01
2.22584754e-01 -3.74871552e-01 6.65044934e-02 -1.26705694e+00
-1.55366763e-01 -7.94495463e-01 4.84549820e-01 -6.03770077e-01
-1.49852669e+00 7.01506019e-01 -2.85294116e-01 -1.08526397e+00
-6.61202073e-02 -6.61367953e-01 -6.14357889e-01 6.58072650e-01
-1.46091926e+00 -9.37977731e-01 -2.55692303e-01 2.97244638e-01
7.17632532e-01 -6.63946271e-02 8.04725289e-01 6.69364154e-01
-6.15590811e-01 8.05594087e-01 6.57115042e-01 1.05132632e-01
8.73591840e-01 -1.42732644e+00 4.66866881e-01 6.17189884e-01
3.88373882e-01 1.15492606e+00 4.55675542e-01 -2.83537477e-01
-1.59270787e+00 -1.07501554e+00 1.24699676e+00 -9.24961805e-01
7.28863120e-01 -4.10777450e-01 -6.33928895e-01 7.31741369e-01
2.93790638e-01 -1.48605645e-01 8.34356427e-01 1.02384937e+00
-6.10491037e-01 -2.63864189e-01 -5.65252900e-01 8.77803326e-01
1.47294307e+00 -6.47342622e-01 -6.24728084e-01 3.96354377e-01
7.86011279e-01 -1.07364893e-01 -9.48090494e-01 3.80929232e-01
9.95660663e-01 -8.43912423e-01 1.03554273e+00 -8.71160090e-01
7.89782703e-01 -1.25896811e-01 -7.19475597e-02 -1.44367361e+00
-9.38975871e-01 -4.20735180e-01 -2.52105623e-01 1.03573596e+00
5.37716806e-01 -5.90622842e-01 6.99230790e-01 4.93279934e-01
-3.61763805e-01 -1.22741902e+00 -2.65637159e-01 -5.81739187e-01
4.21169810e-02 -4.32642519e-01 5.63793957e-01 1.07674885e+00
6.17758669e-02 4.55522984e-01 -7.62175083e-01 -1.40020743e-01
1.33059442e-01 2.92885900e-01 5.77691197e-01 -1.35794330e+00
-6.27723277e-01 -7.28711843e-01 1.18688829e-01 -1.46384168e+00
-5.47352359e-02 -1.29632664e+00 -2.42866263e-01 -1.88853085e+00
4.32153225e-01 -5.73369026e-01 -9.88107264e-01 6.05622053e-01
-3.04999739e-01 4.79451507e-01 1.69145763e-02 2.74062157e-01
-9.61071670e-01 5.38738906e-01 1.06857491e+00 -5.19943126e-02
-3.16801161e-01 1.26192391e-01 -1.09245551e+00 3.43395144e-01
7.16963649e-01 -3.13879460e-01 -4.63558197e-01 -8.82170439e-01
9.38461781e-01 -2.60605633e-01 4.53009941e-02 -5.31111896e-01
2.68168330e-01 1.04345694e-01 2.85166621e-01 -5.17407894e-01
7.23003596e-02 -5.71386576e-01 1.40058085e-01 3.10208529e-01
-8.00508022e-01 5.02232134e-01 -7.87500218e-02 6.47356272e-01
-1.73902974e-01 -6.19009547e-02 2.76858389e-01 3.01203318e-02
-7.72387683e-01 6.91467464e-01 -1.78322136e-01 -1.95403486e-01
3.95474017e-01 2.42916755e-02 -4.61973011e-01 -4.45553988e-01
-5.27203083e-01 4.36731040e-01 3.06677461e-01 9.33344901e-01
3.88643533e-01 -1.37356317e+00 -6.75943971e-01 -2.28594810e-01
2.33398348e-01 -3.50932121e-01 3.43748868e-01 7.62223601e-01
-2.36663818e-01 5.69936633e-01 1.02739297e-01 -3.47609036e-02
-1.12595510e+00 4.91458207e-01 1.20934077e-01 -7.88060963e-01
-5.64843357e-01 8.60799789e-01 5.98823763e-02 -5.93298197e-01
1.60632864e-01 -2.68827230e-01 -4.80487615e-01 1.52179495e-01
5.15152395e-01 4.20321882e-01 1.70750678e-01 -4.17488962e-02
-2.87653148e-01 3.26830655e-01 -6.28860235e-01 9.73240882e-02
1.68074477e+00 3.80861536e-02 -1.36081725e-01 4.38919663e-01
1.27351928e+00 3.88518274e-01 -8.82234395e-01 -4.89203423e-01
2.17300445e-01 -5.23550272e-01 1.62989467e-01 -1.18343329e+00
-1.06591511e+00 5.06855071e-01 4.71278369e-01 2.68736571e-01
9.08000886e-01 -1.31056845e-01 8.36067855e-01 4.33089972e-01
3.78706485e-01 -9.81828451e-01 3.23605478e-01 1.00824022e+00
8.62587452e-01 -1.22218120e+00 3.52506936e-02 2.87060320e-01
-5.94384372e-01 8.10185015e-01 4.64538306e-01 -3.03582817e-01
6.82817578e-01 -2.97351271e-01 -3.05411331e-02 -2.78042793e-01
-1.07713878e+00 -1.17323227e-01 5.20341814e-01 -3.27409655e-02
1.10515690e+00 1.31899863e-01 -5.72784543e-01 7.65111387e-01
-2.02543169e-01 -6.67932853e-02 2.71460742e-01 6.22779667e-01
-3.18245828e-01 -1.45571411e+00 3.85221303e-01 9.18594956e-01
-6.05293930e-01 -4.08038974e-01 -1.90357789e-01 4.63735104e-01
-1.68950304e-01 8.49077642e-01 1.49167284e-01 -7.92716622e-01
4.16988283e-01 -6.05748817e-02 2.87438691e-01 -7.86325872e-01
-9.33967710e-01 -3.48806649e-01 1.01958297e-01 -7.30727553e-01
-2.16269195e-01 -6.07260287e-01 -9.65718269e-01 -5.31235993e-01
-3.45704436e-01 3.57927561e-01 4.95814353e-01 7.19909072e-01
6.31248236e-01 7.83497095e-01 8.22025478e-01 -7.63552606e-01
-6.24125838e-01 -9.90393877e-01 -5.06395876e-01 4.50876445e-01
1.07404701e-01 -4.71779704e-01 -2.83498645e-01 -6.06327772e-01] | [10.153741836547852, 5.705169677734375] |
7bb8e690-b892-48f7-8713-14151f78c788 | simulated-annealing-for-optimization-of | 2110.01384 | null | https://arxiv.org/abs/2110.01384v1 | https://arxiv.org/pdf/2110.01384v1.pdf | Simulated annealing for optimization of graphs and sequences | Optimization of discrete structures aims at generating a new structure with the better property given an existing one, which is a fundamental problem in machine learning. Different from the continuous optimization, the realistic applications of discrete optimization (e.g., text generation) are very challenging due to the complex and long-range constraints, including both syntax and semantics, in discrete structures. In this work, we present SAGS, a novel Simulated Annealing framework for Graph and Sequence optimization. The key idea is to integrate powerful neural networks into metaheuristics (e.g., simulated annealing, SA) to restrict the search space in discrete optimization. We start by defining a sophisticated objective function, involving the property of interest and pre-defined constraints (e.g., grammar validity). SAGS searches from the discrete space towards this objective by performing a sequence of local edits, where deep generative neural networks propose the editing content and thus can control the quality of editing. We evaluate SAGS on paraphrase generation and molecule generation for sequence optimization and graph optimization, respectively. Extensive results show that our approach achieves state-of-the-art performance compared with existing paraphrase generation methods in terms of both automatic and human evaluations. Further, SAGS also significantly outperforms all the previous methods in molecule generation. | ['Sen Song', 'Lili Mou', 'Jie zhou', 'Huasong Zhong', 'Hao Zhou', 'Fandong Meng', 'Pengyong Li', 'Xianggen Liu'] | 2021-10-01 | null | null | null | null | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 6.88014030e-01 1.12123741e-02 -1.33566141e-01 -1.11896969e-01
-6.92861497e-01 -6.15296602e-01 5.67951918e-01 4.81753290e-01
-1.82568401e-01 1.01956820e+00 1.33694798e-01 -3.30653697e-01
-6.05127104e-02 -1.21057081e+00 -1.00589800e+00 -7.09911704e-01
3.15361261e-01 7.42529392e-01 -5.75811006e-02 -5.67511201e-01
6.08297884e-01 5.33249199e-01 -1.36586154e+00 6.40248731e-02
1.48762155e+00 6.57447100e-01 4.29038584e-01 4.64138508e-01
-3.31478298e-01 3.16231608e-01 -8.99943113e-01 -6.32607102e-01
-9.70947146e-02 -9.23671901e-01 -8.53046656e-01 3.47845480e-02
1.84653550e-01 1.64064944e-01 1.92173719e-02 1.25678360e+00
6.61449850e-01 4.55936223e-01 6.52807117e-01 -1.04997146e+00
-1.00565815e+00 7.07786381e-01 -2.48007819e-01 -1.24827653e-01
7.10762739e-01 2.97598839e-01 1.18339980e+00 -6.99214041e-01
7.92536914e-01 1.28838921e+00 3.40183586e-01 5.75695395e-01
-1.23929846e+00 -3.83757770e-01 2.02658176e-01 2.73001790e-01
-1.28796148e+00 -2.45884567e-01 8.87491226e-01 -1.47093758e-01
1.04521024e+00 4.42446828e-01 8.34071994e-01 1.03929818e+00
3.97357225e-01 6.84994042e-01 5.18227518e-01 -4.89545882e-01
5.68730772e-01 -2.66173095e-01 -2.17990935e-01 8.59911501e-01
1.11467615e-01 7.38306623e-03 -4.27267164e-01 -3.04552257e-01
4.48313862e-01 -1.90105662e-01 -3.00363779e-01 -3.78045499e-01
-1.19845402e+00 1.18295228e+00 4.29008335e-01 2.85245050e-02
-3.39447021e-01 -1.77932605e-02 3.24275285e-01 1.61715195e-01
9.79106501e-02 1.06020761e+00 -1.15913406e-01 -1.13430902e-01
-8.82565975e-01 4.27341402e-01 9.04035807e-01 1.04840672e+00
6.88392639e-01 3.20458338e-02 -6.35865986e-01 9.45989788e-01
1.26006961e-01 1.22727931e-01 4.74675596e-01 -5.18140197e-01
5.44662178e-01 6.66665971e-01 -3.09508163e-02 -1.05222595e+00
-1.75283447e-01 -6.40123248e-01 -1.10025573e+00 -6.52203783e-02
4.34839949e-02 -1.37691215e-01 -9.89386797e-01 1.70223987e+00
4.18260753e-01 9.10725072e-02 -2.97548771e-02 7.02172279e-01
9.34693336e-01 1.01451850e+00 -9.46751162e-02 -5.14561117e-01
9.50398445e-01 -1.33231640e+00 -7.54715443e-01 -1.79200187e-01
3.86665136e-01 -7.30656624e-01 1.08211315e+00 3.53476167e-01
-1.27406096e+00 -4.53922451e-01 -1.21715069e+00 -2.58030556e-03
-4.64905828e-01 -1.17453299e-01 6.83460832e-01 5.06755590e-01
-9.31487560e-01 8.77938330e-01 -4.80767280e-01 -1.14968881e-01
2.12835297e-01 4.84259486e-01 -1.44325299e-02 1.61695257e-01
-1.48589516e+00 7.51144111e-01 7.45847225e-01 2.92762011e-01
-6.71605825e-01 -5.95136881e-01 -9.96822774e-01 1.43149495e-01
7.28538871e-01 -1.13697791e+00 1.03683972e+00 -6.26338780e-01
-1.93057585e+00 5.07459879e-01 -1.35417655e-01 -4.10638928e-01
4.15688843e-01 1.15192153e-01 -7.21834302e-02 -2.29992092e-01
-9.89057422e-02 5.86936533e-01 8.28149676e-01 -8.99409831e-01
-2.78679579e-01 -1.87954586e-02 1.51868880e-01 6.23446941e-01
-1.25479490e-01 -1.62603974e-01 -6.26235425e-01 -7.56299198e-01
-1.25189498e-01 -9.22522843e-01 -5.56403458e-01 -3.52529645e-01
-8.26166511e-01 -3.51381510e-01 3.90891522e-01 -6.58041894e-01
1.55690038e+00 -1.48724997e+00 9.55251336e-01 4.45762724e-01
1.78066269e-01 3.78845423e-01 -1.99267447e-01 7.94734895e-01
3.60703319e-02 3.45887363e-01 -5.23021460e-01 -2.18844131e-01
8.41522664e-02 9.85927880e-02 -4.06601466e-02 -6.29645074e-03
9.13717598e-02 1.27777064e+00 -1.04501843e+00 -4.93520230e-01
1.13501117e-01 9.12993476e-02 -7.41586387e-01 3.51085037e-01
-8.70317996e-01 3.99535418e-01 -5.80442369e-01 5.97795665e-01
4.90290821e-01 -3.65064085e-01 3.24310601e-01 -8.86718184e-02
-3.97158638e-02 2.14283064e-01 -1.04118764e+00 1.92502284e+00
-4.49422598e-01 2.95954756e-02 -3.73905569e-01 -9.42906141e-01
1.10248077e+00 -1.09922253e-01 1.06795453e-01 -6.37593210e-01
-3.14609334e-02 8.55549201e-02 -3.85095477e-02 -2.02952325e-01
6.77634478e-01 9.31116715e-02 -9.86563638e-02 3.55343699e-01
-1.77897722e-01 -6.05533302e-01 5.87687194e-01 1.61313310e-01
8.00223351e-01 2.54873157e-01 5.80747366e-01 1.67118199e-02
7.39929676e-01 -2.26250123e-02 5.18205464e-01 8.83825183e-01
3.50492954e-01 5.04830658e-01 7.04775870e-01 -2.55433261e-01
-1.08032131e+00 -9.04247403e-01 4.17551160e-01 9.50360715e-01
2.78930813e-01 -5.57168424e-01 -9.84103501e-01 -6.71028972e-01
-1.96199372e-01 8.49827349e-01 -4.88500148e-01 -5.83548188e-01
-7.25057662e-01 -9.46221888e-01 3.16388279e-01 1.43705457e-01
3.69224519e-01 -1.48878980e+00 -8.07670802e-02 5.06936371e-01
-1.86000183e-01 -7.94181526e-01 -9.73019958e-01 -7.73281083e-02
-6.61807120e-01 -7.60041714e-01 -6.76214337e-01 -1.00610328e+00
5.61395586e-01 -2.58349210e-01 1.30056918e+00 3.61456692e-01
-3.93427551e-01 -3.15651059e-01 -4.51605171e-01 -1.78773254e-01
-8.11407864e-01 5.82277596e-01 -5.28671503e-01 -1.89755470e-01
-1.28508300e-01 -6.20752573e-01 -5.17444789e-01 1.80279478e-01
-1.18109715e+00 3.36093605e-01 7.73846447e-01 1.20714295e+00
9.27130997e-01 -6.04425222e-02 5.77854991e-01 -1.06197548e+00
1.24029994e+00 -2.59323716e-01 -7.86017418e-01 7.49575257e-01
-5.90010107e-01 4.33024704e-01 9.92929757e-01 -3.79140556e-01
-8.21767151e-01 2.18693465e-02 -3.16011816e-01 -9.93138328e-02
7.89438039e-02 8.71493518e-01 -4.90346104e-01 -1.08629599e-01
6.35204136e-01 7.12158859e-01 -4.52965461e-02 -9.88113135e-02
4.50857818e-01 2.95993090e-01 3.53878975e-01 -6.66557372e-01
7.06856966e-01 -1.27752289e-01 1.62960336e-01 -6.97577596e-01
-6.57698572e-01 5.99747617e-03 -5.17956972e-01 1.26506209e-01
6.97804093e-01 -3.00310969e-01 -7.56586313e-01 6.13876879e-01
-1.35948205e+00 -2.46329993e-01 -1.41830564e-01 -4.26437855e-02
-8.45811486e-01 7.54674554e-01 -3.88672411e-01 -3.85140449e-01
-7.40498424e-01 -1.44113207e+00 1.12096322e+00 4.30632174e-01
-2.29491219e-01 -1.04521143e+00 1.66181967e-01 2.12558389e-01
2.41085887e-01 6.17734969e-01 1.29887724e+00 -5.39227068e-01
-7.68878818e-01 1.19453467e-01 1.50265798e-01 3.98007073e-02
2.09019944e-01 1.11356609e-01 -8.88701975e-02 -3.52152139e-01
-2.80367494e-01 -3.37890089e-01 6.48984313e-01 3.07791710e-01
1.35723436e+00 -5.09992957e-01 -3.87502491e-01 9.33734298e-01
1.30363178e+00 5.50594389e-01 9.19984341e-01 3.38490516e-01
6.73943400e-01 3.03584129e-01 7.13429868e-01 6.00543261e-01
-3.87911052e-02 8.82203221e-01 4.78591919e-01 -1.50674045e-01
5.18493354e-02 -5.17107606e-01 1.66928679e-01 5.71707547e-01
1.12849995e-01 -7.82732725e-01 -6.29196465e-01 1.84687838e-01
-2.00262499e+00 -8.93835545e-01 2.06829548e-01 2.14672637e+00
1.18991184e+00 8.40356797e-02 3.25143598e-02 -9.28610265e-02
9.62399423e-01 3.10796022e-01 -8.95451665e-01 -6.69424832e-01
-1.84828922e-01 4.74298120e-01 1.48695230e-01 6.81007147e-01
-8.45855296e-01 1.21627605e+00 5.53277826e+00 1.15213227e+00
-9.29107606e-01 -5.14101505e-01 6.42999411e-01 6.32609241e-03
-5.30178249e-01 -8.79720673e-02 -8.07415962e-01 6.45694435e-01
5.32983661e-01 -3.63387346e-01 9.47084844e-01 5.51975727e-01
2.13373065e-01 2.85239160e-01 -1.16507435e+00 7.62264669e-01
2.80205518e-01 -1.81906259e+00 6.19493723e-01 -9.46845785e-02
9.54282105e-01 -6.05176091e-01 3.97727042e-02 2.08839998e-01
3.92012805e-01 -1.40621948e+00 5.53326845e-01 5.19444346e-01
7.28718400e-01 -1.15144229e+00 5.46740532e-01 4.05189276e-01
-1.10668111e+00 1.66694999e-01 -3.14211041e-01 3.73973846e-01
3.70913953e-01 4.64569539e-01 -7.69224644e-01 8.02276134e-01
4.07151207e-02 7.78791189e-01 -4.77618456e-01 1.09302390e+00
-5.84872425e-01 2.07793280e-01 -1.30199686e-01 -7.17582941e-01
3.69946122e-01 -8.13203812e-01 6.85341239e-01 1.06677008e+00
2.52739996e-01 -1.06535107e-03 2.90237129e-01 1.21133244e+00
-2.67934293e-01 3.32037807e-01 -4.25617933e-01 -2.42537618e-01
4.85239416e-01 1.02781987e+00 -5.59678674e-01 -2.86039859e-01
1.25557348e-01 1.16554809e+00 4.72181350e-01 3.75357032e-01
-1.07529497e+00 -8.10814381e-01 2.82101214e-01 -1.70719802e-01
4.47272927e-01 -1.14248373e-01 -1.39761999e-01 -9.77414310e-01
1.11849152e-01 -1.37230527e+00 3.64979953e-01 -5.72453022e-01
-1.04359770e+00 7.21992731e-01 -1.51634529e-01 -7.87257612e-01
-3.99959654e-01 -2.96450049e-01 -9.16901410e-01 1.02236974e+00
-1.34421492e+00 -1.08142185e+00 -3.07747036e-01 2.85242498e-01
6.72797441e-01 -1.40195221e-01 6.66487813e-01 1.55908829e-02
-5.99313378e-01 8.53545845e-01 2.84592420e-01 -2.73739785e-01
3.75969917e-01 -1.22492576e+00 8.08612525e-01 6.73644066e-01
8.36496651e-02 7.87550211e-01 8.52540612e-01 -9.89850521e-01
-1.50484443e+00 -9.89546478e-01 8.08003485e-01 5.61357029e-02
6.55717134e-01 -4.79616553e-01 -7.96204448e-01 1.70836091e-01
1.65055469e-01 -6.78174734e-01 4.03308570e-01 3.69495805e-03
1.60535708e-01 3.33522707e-01 -9.38798010e-01 8.84683728e-01
1.49854350e+00 -1.39477536e-01 -3.83668780e-01 7.04632282e-01
1.06710505e+00 -8.01327348e-01 -7.15255082e-01 3.24905574e-01
3.60694110e-01 -7.96241760e-01 1.02431595e+00 -9.13963556e-01
6.55915260e-01 -3.05938274e-01 1.49197191e-01 -1.67010188e+00
-4.57061321e-01 -1.12453306e+00 -1.10621065e-01 1.02954590e+00
7.01482296e-01 -5.47794640e-01 9.85422313e-01 1.61249176e-01
-4.05687839e-01 -1.28186607e+00 -6.42160714e-01 -7.50731707e-01
-5.98142995e-03 3.64138067e-01 9.80771959e-01 6.45800948e-01
-2.32685864e-01 6.52518332e-01 -4.24628377e-01 -6.43886402e-02
4.90839183e-01 5.70276976e-01 7.05960810e-01 -8.62911522e-01
-6.66474938e-01 -5.80800593e-01 -1.53862387e-01 -1.21594012e+00
3.24394256e-01 -9.78610575e-01 -4.36892733e-03 -1.85525465e+00
7.53791183e-02 -2.06112832e-01 8.00184906e-02 1.96460947e-01
-4.50759351e-01 -2.45215937e-01 7.81513229e-02 -8.69649947e-02
-5.06492257e-01 1.11538243e+00 1.57182992e+00 -3.97544503e-01
-6.01335406e-01 1.14971012e-01 -7.85166204e-01 2.91893601e-01
8.65638256e-01 -2.94366241e-01 -4.63618308e-01 -2.36485571e-01
4.45400506e-01 3.19748044e-01 -8.21077600e-02 -6.47113502e-01
2.83111274e-01 -5.41805804e-01 8.36202949e-02 -6.27017558e-01
2.40226924e-01 -2.62453377e-01 3.21138769e-01 6.35066807e-01
-5.24178445e-01 2.69742638e-01 -5.11506014e-02 7.11056590e-01
-2.67524868e-01 -5.81886351e-01 7.06930697e-01 -2.86773175e-01
-4.07021493e-01 4.37242299e-01 1.27794649e-02 2.33107045e-01
9.89369571e-01 -3.79595906e-01 -3.36621910e-01 -4.55318928e-01
-7.23779917e-01 4.16943699e-01 5.62516272e-01 3.56303602e-01
8.06146145e-01 -1.08614099e+00 -7.56264985e-01 8.69372673e-03
5.69804423e-02 2.65436530e-01 1.98237244e-02 3.68555039e-01
-6.99983120e-01 2.90722072e-01 -1.17135018e-01 -3.54678780e-01
-1.29395330e+00 6.24356687e-01 4.06259179e-01 -7.52449751e-01
-2.16117948e-01 8.16068709e-01 8.63025859e-02 -5.13175189e-01
2.48948395e-01 -1.55953512e-01 -2.71036208e-01 -6.00746237e-02
1.75883338e-01 2.33180925e-01 2.07681522e-01 -1.26146391e-01
-1.52079970e-01 5.03644526e-01 -3.85016859e-01 2.23317429e-01
1.24929512e+00 3.09983104e-01 -3.66792589e-01 -1.56138062e-01
1.07073390e+00 -1.05734274e-01 -7.35583186e-01 -6.13593757e-02
-1.51903883e-01 -1.96028143e-01 -3.67325187e-01 -9.27799642e-01
-7.33864367e-01 5.80123067e-01 8.97013173e-02 1.17355799e-02
9.93990958e-01 -1.59316629e-01 1.00133991e+00 6.82542682e-01
1.30804598e-01 -1.11499572e+00 3.09610754e-01 6.53357148e-01
1.14076817e+00 -9.26880479e-01 1.99190214e-01 -6.29281461e-01
-6.29713833e-01 1.05218399e+00 6.42373860e-01 4.40224260e-02
2.07493138e-02 -1.60469577e-01 -4.94250000e-01 -1.63036928e-01
-6.87541246e-01 1.14235535e-01 4.41656470e-01 4.70413744e-01
3.19397449e-01 -9.05842707e-02 -6.27105534e-01 1.86398700e-01
-4.92818654e-01 -1.63372055e-01 3.39108050e-01 8.44211996e-01
-4.96996760e-01 -1.51699662e+00 -2.40084246e-01 6.07146263e-01
-8.28239024e-02 -4.48932081e-01 -7.78649092e-01 4.76356149e-01
7.21424446e-02 7.75164425e-01 -3.26053560e-01 -1.21960871e-01
4.16701227e-01 -1.96323916e-01 5.48396766e-01 -8.07097793e-01
-8.83675575e-01 -1.65405229e-01 2.15141654e-01 -3.25815350e-01
1.28945122e-02 -5.16525328e-01 -1.19812691e+00 -2.36350581e-01
-5.21692276e-01 4.59059864e-01 5.73817670e-01 7.53115952e-01
4.06819254e-01 7.22369730e-01 7.14891076e-01 -5.12452900e-01
-7.35815883e-01 -5.31584799e-01 -1.90384656e-01 4.92745996e-01
-2.61090342e-02 -4.40502167e-01 -1.05710998e-01 -1.36997029e-01] | [4.965558052062988, 5.741209030151367] |
f0fad800-19b0-4cab-9226-eb2f7ba07ada | factorizable-net-an-efficient-subgraph-based | 1806.11538 | null | http://arxiv.org/abs/1806.11538v2 | http://arxiv.org/pdf/1806.11538v2.pdf | Factorizable Net: An Efficient Subgraph-based Framework for Scene Graph Generation | Generating scene graph to describe all the relations inside an image gains
increasing interests these years. However, most of the previous methods use
complicated structures with slow inference speed or rely on the external data,
which limits the usage of the model in real-life scenarios. To improve the
efficiency of scene graph generation, we propose a subgraph-based connection
graph to concisely represent the scene graph during the inference. A bottom-up
clustering method is first used to factorize the entire scene graph into
subgraphs, where each subgraph contains several objects and a subset of their
relationships. By replacing the numerous relationship representations of the
scene graph with fewer subgraph and object features, the computation in the
intermediate stage is significantly reduced. In addition, spatial information
is maintained by the subgraph features, which is leveraged by our proposed
Spatial-weighted Message Passing~(SMP) structure and Spatial-sensitive Relation
Inference~(SRI) module to facilitate the relationship recognition. On the
recent Visual Relationship Detection and Visual Genome datasets, our method
outperforms the state-of-the-art method in both accuracy and speed. | ['Chao Zhang', 'Yikang Li', 'Bolei Zhou', 'Wanli Ouyang', 'Xiaogang Wang', 'Jianping Shi'] | 2018-06-29 | factorizable-net-an-efficient-subgraph-based-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Yikang_LI_Factorizable_Net_An_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Yikang_LI_Factorizable_Net_An_ECCV_2018_paper.pdf | eccv-2018-9 | ['visual-relationship-detection'] | ['computer-vision'] | [ 2.39602968e-01 -8.62076879e-02 -2.00607195e-01 -5.45680583e-01
2.24332903e-02 -4.18352187e-01 4.21881229e-01 4.47835118e-01
-7.07219169e-02 3.01381558e-01 -4.11771275e-02 -2.33549222e-01
-2.62775958e-01 -1.23994339e+00 -5.29491842e-01 -6.37623489e-01
3.72758061e-02 3.37198287e-01 7.70446360e-01 -7.14077801e-02
1.76429644e-01 5.72867453e-01 -1.56956685e+00 1.62586942e-01
8.68581533e-01 8.22157383e-01 4.86309767e-01 4.75551337e-01
-4.91930991e-01 8.61958444e-01 -3.18372160e-01 -2.28704423e-01
-8.58576149e-02 -3.74713182e-01 -7.42137492e-01 3.34321201e-01
1.60023436e-01 -1.60999507e-01 -8.28723609e-01 1.15686274e+00
3.35323185e-01 1.28856182e-01 3.61753434e-01 -1.46909308e+00
-3.74661326e-01 6.48266077e-01 -9.11158144e-01 1.18744344e-01
6.37737513e-01 -5.82593381e-02 9.52621460e-01 -6.65353835e-01
8.33370745e-01 1.38953269e+00 2.58073300e-01 -1.66837290e-01
-1.08129716e+00 -8.10566247e-01 6.08049750e-01 6.82236850e-01
-1.89886141e+00 -1.25254646e-01 8.74000013e-01 -2.52001494e-01
9.96838868e-01 4.27564979e-01 9.98067558e-01 2.96737522e-01
-8.22757259e-02 7.02031970e-01 6.71251237e-01 -1.47676274e-01
7.99276121e-03 -1.17856301e-01 1.70288816e-01 9.35889125e-01
3.42643142e-01 -5.00006676e-01 -5.04638910e-01 8.29302669e-02
8.64458740e-01 3.20504576e-01 -3.46091956e-01 -6.93021357e-01
-1.25690997e+00 5.18589914e-01 9.28677738e-01 1.06681503e-01
-9.61138904e-02 2.07930148e-01 2.33708933e-01 -6.75629452e-02
3.20832849e-01 4.51111346e-02 4.30940464e-02 2.30656967e-01
-6.55622721e-01 1.15923574e-02 5.88377535e-01 1.39254105e+00
1.06574655e+00 -5.32774389e-01 -1.96937531e-01 8.27879965e-01
4.51888800e-01 2.48526186e-01 -1.19416937e-01 -5.99254608e-01
6.58139706e-01 1.50690794e+00 -3.15086812e-01 -1.79312468e+00
-3.17965835e-01 -3.96908969e-01 -1.11473691e+00 -5.05229592e-01
-2.52026841e-02 4.09912616e-01 -7.93983579e-01 1.47879493e+00
8.19896638e-01 4.19461221e-01 -3.55335981e-01 7.78278649e-01
1.07198703e+00 7.84478903e-01 -2.78203618e-02 -2.40614638e-01
1.43730390e+00 -8.22777331e-01 -6.36600912e-01 -9.49884728e-02
5.66770315e-01 -6.23573840e-01 6.65114582e-01 -3.91670950e-02
-7.92497694e-01 -5.70745885e-01 -9.94371831e-01 -2.48504370e-01
-4.34149593e-01 -1.27529656e-03 1.04034054e+00 2.65013248e-01
-8.76453817e-01 2.17166930e-01 -6.67710006e-01 -6.85203433e-01
6.50395274e-01 4.20146018e-01 -4.51117277e-01 -5.22998512e-01
-8.99852753e-01 3.39372694e-01 6.12411797e-01 2.88897127e-01
-6.27399862e-01 -2.76238471e-01 -1.05729353e+00 1.99936330e-01
5.82056582e-01 -6.95583045e-01 5.43647528e-01 -3.04035187e-01
-9.36258852e-01 6.77772403e-01 -4.12583470e-01 1.50814697e-01
2.24906668e-01 4.94461097e-02 -3.04243237e-01 3.65743846e-01
1.44006133e-01 6.41523123e-01 5.04937708e-01 -1.34546447e+00
-8.47747087e-01 -2.48697668e-01 2.92717725e-01 4.40164626e-01
-2.11121023e-01 -1.92264356e-02 -1.51731026e+00 -4.46341395e-01
6.97700083e-01 -7.44031429e-01 -3.58762026e-01 1.00437202e-01
-8.58530462e-01 -2.85423011e-01 9.84452784e-01 -3.81862402e-01
1.71905911e+00 -2.27197361e+00 3.54650885e-01 6.78986430e-01
5.67184150e-01 3.73725481e-02 -1.10436350e-01 5.98634779e-01
-1.54723659e-01 -2.85409708e-02 -2.23866105e-01 -1.06703527e-01
-2.85585374e-01 3.77113074e-01 -4.36312445e-02 5.74140668e-01
5.34118600e-02 8.00016165e-01 -1.04989767e+00 -1.04118502e+00
3.68167937e-01 3.77743542e-01 -4.76013124e-01 3.86069119e-01
-1.82173267e-01 3.95779788e-01 -7.40824103e-01 5.22392333e-01
9.39307749e-01 -5.54931164e-01 4.85610932e-01 -4.65992957e-01
3.30921002e-02 7.89555162e-02 -1.45851207e+00 1.92234659e+00
-2.37841960e-02 2.85553247e-01 -1.00938082e-01 -1.03849661e+00
9.42387819e-01 -1.00426570e-01 4.12003636e-01 -4.56696004e-01
-3.27840783e-02 -2.91396856e-01 -5.26791140e-02 -4.57513094e-01
1.65073216e-01 6.20937467e-01 -4.69923355e-02 3.16030651e-01
-2.49133512e-01 -1.39246240e-01 2.68349141e-01 8.82465661e-01
1.32517290e+00 2.27552772e-01 5.07669747e-01 -1.48937881e-01
5.91333747e-01 9.61245224e-02 7.08970547e-01 4.71516579e-01
1.49770796e-01 3.80159140e-01 6.36642754e-01 -2.41505593e-01
-5.16365528e-01 -9.31832731e-01 1.29579365e-01 8.93674791e-01
8.79367530e-01 -7.41300941e-01 -6.25012577e-01 -6.23018920e-01
-3.49631682e-02 2.04845920e-01 -3.95215064e-01 -2.31732234e-01
-6.37023389e-01 -6.92569852e-01 7.62101039e-02 4.80975479e-01
7.87475646e-01 -7.54365623e-01 -2.69536883e-01 1.26315102e-01
-2.27555692e-01 -1.13643336e+00 -4.20551479e-01 -2.59343028e-01
-6.77179635e-01 -1.19188440e+00 -4.22172695e-02 -9.36495721e-01
1.21260607e+00 7.05185235e-01 1.04673910e+00 5.23936391e-01
-4.78459179e-01 5.59101328e-02 -4.94720966e-01 7.06412941e-02
2.04103395e-01 7.53431246e-02 -2.58477360e-01 1.25542611e-01
2.26602569e-01 -6.79227293e-01 -7.79953778e-01 4.12730187e-01
-6.38426244e-01 6.32670641e-01 6.33235157e-01 4.71242547e-01
8.02595615e-01 5.05007505e-01 -5.76232979e-03 -1.12134218e+00
1.40897110e-01 -5.99914253e-01 -6.46166086e-01 4.63869810e-01
-2.87965328e-01 4.96347845e-02 4.81998861e-01 -2.67476708e-01
-1.15531731e+00 2.49097437e-01 2.79090703e-01 -2.47514829e-01
5.31425811e-02 4.31703717e-01 -6.10761285e-01 -1.33491680e-01
1.71207324e-01 3.44279051e-01 -3.82470459e-01 -4.16001201e-01
6.11345530e-01 4.21023816e-01 5.15322149e-01 -3.76235753e-01
1.03144073e+00 5.75476706e-01 3.55871707e-01 -6.91888154e-01
-8.56560171e-01 -7.31148481e-01 -7.83643425e-01 -2.86647648e-01
8.58269215e-01 -9.18142557e-01 -8.71253669e-01 3.58126968e-01
-1.14625990e+00 -1.12019246e-02 8.92848819e-02 2.99619138e-01
-1.21913664e-01 4.99271661e-01 -6.64647698e-01 -6.16669893e-01
4.30622362e-02 -1.15743577e+00 1.09934330e+00 2.45449916e-01
1.31348539e-02 -6.40260339e-01 -1.52164385e-01 2.19188616e-01
9.31715071e-02 3.18991214e-01 1.20136547e+00 -3.00418735e-01
-1.24913478e+00 -8.83105397e-02 -9.02576566e-01 -2.60376126e-01
1.98082641e-01 6.73210770e-02 -6.07940912e-01 -1.00509129e-01
-4.76168454e-01 1.10017695e-02 7.19695926e-01 3.85068320e-02
1.53896081e+00 -1.16323810e-02 -9.61900651e-01 8.32519650e-01
1.39680028e+00 1.67266279e-01 5.49004793e-01 -3.46246287e-02
1.37628961e+00 6.65610611e-01 8.90405118e-01 4.56487238e-01
8.28073382e-01 6.69490933e-01 5.30004442e-01 -3.58755112e-01
-2.63486683e-01 -3.49237025e-01 -2.25784019e-01 1.08430815e+00
-1.25118615e-02 -4.08546090e-01 -8.23729694e-01 4.89393294e-01
-2.30413985e+00 -8.43802750e-01 -4.81065482e-01 1.93336964e+00
6.16713643e-01 -6.36293218e-02 -1.96054384e-01 8.61982778e-02
1.04371333e+00 3.84095013e-01 -5.05092502e-01 1.37096480e-01
1.00976154e-02 -1.41792953e-01 2.49908343e-01 3.24629396e-01
-9.55397725e-01 1.10547698e+00 6.08865261e+00 9.55950022e-01
-5.92357755e-01 -3.37319940e-01 6.23357832e-01 1.93321615e-01
-4.64333653e-01 3.57439220e-01 -6.54533803e-01 1.60337836e-01
1.12211719e-01 -2.93282010e-02 4.11675215e-01 8.34281206e-01
2.13401932e-02 -3.38918954e-01 -1.09800851e+00 1.18679035e+00
2.65807230e-02 -1.21645665e+00 3.88381928e-01 -1.40464390e-02
3.67242545e-01 -3.54529321e-01 -3.51572037e-01 4.55534458e-02
2.57869154e-01 -8.34708154e-01 3.52468163e-01 5.33064723e-01
9.20235515e-01 -8.98244619e-01 5.76911747e-01 2.58973718e-01
-2.06808066e+00 1.91705912e-01 -5.71929574e-01 -7.22224191e-02
-2.26418097e-02 6.88207686e-01 -6.66090071e-01 8.84611130e-01
7.61156023e-01 9.59764123e-01 -7.10296929e-01 9.10883665e-01
-2.85740316e-01 3.70268077e-01 -4.70825046e-01 -1.72611713e-01
1.39321089e-02 -3.82135808e-01 2.29614258e-01 1.24106419e+00
-1.90595177e-03 4.71451163e-01 5.43389738e-01 6.49776042e-01
-8.38757381e-02 1.00869767e-01 -6.91968560e-01 1.45462006e-01
8.05893242e-01 1.54361105e+00 -1.03631234e+00 -3.64653975e-01
-5.93570948e-01 9.68838930e-01 6.71627760e-01 3.92814577e-01
-9.30782199e-01 -5.62501431e-01 4.87517476e-01 1.65624246e-01
2.80888528e-01 -2.81477034e-01 -6.71589002e-02 -1.09732997e+00
-1.01735517e-02 -5.20697773e-01 5.93645751e-01 -7.40678668e-01
-1.10896361e+00 3.94334584e-01 3.74054193e-01 -1.14424610e+00
1.87812269e-01 -2.06115723e-01 -5.98408580e-01 7.50993967e-01
-1.30288303e+00 -1.30914223e+00 -7.85874546e-01 8.19169164e-01
2.58069754e-01 2.09534854e-01 5.09989500e-01 3.22412431e-01
-9.50307965e-01 3.19810122e-01 -3.96039397e-01 2.47499496e-01
3.80849898e-01 -1.02736914e+00 3.71488154e-01 1.05355346e+00
8.35623890e-02 9.21022475e-01 2.94577777e-01 -8.21869910e-01
-1.60737133e+00 -1.31224883e+00 8.25527191e-01 -7.34537169e-02
5.69658041e-01 -6.08329892e-01 -8.76838148e-01 6.73362195e-01
1.09645361e-02 3.19064945e-01 5.49053848e-01 2.90089846e-02
-4.68459249e-01 -4.04134393e-01 -9.45572376e-01 8.48542094e-01
1.80460262e+00 -4.54347253e-01 -3.36631596e-01 3.36686015e-01
8.74407828e-01 -3.46740097e-01 -8.93018425e-01 4.41877931e-01
3.43554258e-01 -8.34118426e-01 1.09026194e+00 -2.19061255e-01
2.34344870e-01 -9.87010837e-01 -1.08183898e-01 -8.53118658e-01
-6.87427044e-01 -4.00522381e-01 -1.76811785e-01 1.43325555e+00
1.91352144e-01 -5.26332140e-01 6.73745036e-01 3.24336797e-01
8.55402574e-02 -8.05299461e-01 -6.33212030e-01 -3.16349298e-01
-9.71916318e-01 -3.93998116e-01 8.56467485e-01 9.59273040e-01
4.04829420e-02 6.13697886e-01 -1.38127223e-01 4.18233305e-01
6.88699663e-01 7.30307996e-01 1.07440853e+00 -1.04359031e+00
-2.67754018e-01 -6.44379929e-02 -8.09881985e-01 -1.42563403e+00
-6.49938583e-02 -8.75009894e-01 3.56192067e-02 -1.90507483e+00
7.32871950e-01 -6.72729850e-01 -1.64148971e-01 4.74272132e-01
-4.87819612e-01 1.21463127e-01 2.60329079e-02 2.96690702e-01
-1.12300169e+00 4.49504137e-01 1.34844589e+00 -1.24068961e-01
-9.89597738e-02 -3.64201516e-01 -5.42090356e-01 8.11762333e-01
3.92862171e-01 -3.92067760e-01 -8.81575644e-01 -4.60963041e-01
2.63350457e-01 5.06193787e-02 1.37241915e-01 -8.42833340e-01
5.50253451e-01 -3.47908705e-01 3.76883745e-01 -9.50776637e-01
3.65252048e-01 -9.59893107e-01 5.41372657e-01 3.92357439e-01
1.03181340e-02 1.95042342e-02 -1.18810602e-01 8.39306891e-01
-1.90673232e-01 2.88429648e-01 5.37407398e-01 -1.35306865e-01
-8.65472853e-01 6.25924587e-01 1.12711802e-01 -3.09910893e-01
1.16549194e+00 -3.71960461e-01 -4.11508352e-01 -2.45010018e-01
-3.79902214e-01 4.66880769e-01 4.94763404e-01 1.75867662e-01
8.69820356e-01 -1.27431941e+00 -4.58398581e-01 2.10110769e-01
5.73261023e-01 6.63723409e-01 3.83027703e-01 7.38038778e-01
-7.28079617e-01 3.32809612e-02 6.51797205e-02 -8.71760309e-01
-1.42186427e+00 7.31256247e-01 -1.30149737e-01 -2.40738332e-01
-7.69008875e-01 1.08546793e+00 6.52003109e-01 -1.02227956e-01
1.18865050e-01 -2.82597154e-01 -4.32493716e-01 -1.33285031e-01
3.37486535e-01 3.32972348e-01 -3.78414959e-01 -7.68328547e-01
-7.44966745e-01 8.58077884e-01 -2.01548964e-01 2.64007509e-01
1.17934656e+00 -2.59415984e-01 -6.22239292e-01 2.64771700e-01
1.02027452e+00 -3.00353598e-02 -9.76525605e-01 -3.51528108e-01
-1.28897458e-01 -8.69001031e-01 -1.45327404e-01 -1.91392869e-01
-1.13808846e+00 6.28896058e-01 9.52749550e-02 1.53495580e-01
1.35264695e+00 2.73772269e-01 6.19375169e-01 3.54624689e-01
5.55157304e-01 -6.75150037e-01 7.61315301e-02 1.68378234e-01
6.27472758e-01 -9.43831265e-01 4.82794791e-01 -1.47011650e+00
-4.87480670e-01 7.13374674e-01 7.54419804e-01 -7.88877904e-02
7.34828591e-01 1.16042458e-01 -3.77335310e-01 -5.37652731e-01
-5.86341739e-01 -3.38816583e-01 2.93300033e-01 5.59670866e-01
2.31295407e-01 7.35220406e-03 -2.38784000e-01 2.46240452e-01
5.20628132e-02 -3.90564591e-01 -1.54340090e-02 8.83905768e-01
-4.35076028e-01 -1.01435328e+00 -6.26173541e-02 5.49857676e-01
1.01049192e-01 -1.68914929e-01 -3.49501431e-01 7.86447585e-01
3.26805919e-01 1.15932369e+00 1.86260879e-01 -5.53026140e-01
2.87361085e-01 -5.29229343e-01 2.70528615e-01 -8.92315745e-01
-2.62545258e-01 2.97692921e-02 1.94821015e-01 -1.03420532e+00
-5.53576827e-01 -4.33852702e-01 -1.50641429e+00 -4.42596227e-01
-5.51761568e-01 3.83682875e-03 1.69690266e-01 7.70552337e-01
5.42874992e-01 6.40327513e-01 6.80341959e-01 -5.41353881e-01
3.82219374e-01 -6.43954694e-01 -5.69785058e-01 6.04988158e-01
-1.91737816e-01 -7.01251745e-01 2.98866071e-03 6.75252778e-03] | [10.212055206298828, 1.638039469718933] |
efb5eb9d-b9c8-4364-bfb9-ae096dcdfefe | question-rewriting-assessing-its-importance | 2201.09146 | null | https://arxiv.org/abs/2201.09146v2 | https://arxiv.org/pdf/2201.09146v2.pdf | Question rewriting? Assessing its importance for conversational question answering | In conversational question answering, systems must correctly interpret the interconnected interactions and generate knowledgeable answers, which may require the retrieval of relevant information from a background repository. Recent approaches to this problem leverage neural language models, although different alternatives can be considered in terms of modules for (a) representing user questions in context, (b) retrieving the relevant background information, and (c) generating the answer. This work presents a conversational question answering system designed specifically for the Search-Oriented Conversational AI (SCAI) shared task, and reports on a detailed analysis of its question rewriting module. In particular, we considered different variations of the question rewriting module to evaluate the influence on the subsequent components, and performed a careful analysis of the results obtained with the best system configuration. Our system achieved the best performance in the shared task and our analysis emphasizes the importance of the conversation context representation for the overall system performance. | ['Luísa Coheur', 'Bruno Martins', 'Rui Ribeiro', 'Gonçalo Raposo'] | 2022-01-22 | null | null | null | null | ['question-rewriting'] | ['natural-language-processing'] | [ 4.40472275e-01 4.62622583e-01 5.02738178e-01 -3.93203974e-01
-8.22805226e-01 -6.42380238e-01 1.16247761e+00 3.16678584e-01
-3.89843196e-01 6.21248662e-01 5.28868377e-01 -6.63091958e-01
-3.82487416e-01 -7.96664894e-01 -2.42457926e-01 -2.15444416e-01
2.37461969e-01 7.83919632e-01 4.70994800e-01 -6.53391004e-01
4.59111720e-01 8.76760781e-02 -1.63161027e+00 7.44961560e-01
9.16146696e-01 7.42815971e-01 1.84809595e-01 1.06688142e+00
-8.01545858e-01 1.37743580e+00 -1.01598024e+00 -4.24831182e-01
-3.12103242e-01 -8.64513695e-01 -1.47805858e+00 -3.00713390e-01
1.60769492e-01 -2.21436560e-01 -3.51549895e-03 4.69554514e-01
2.93880075e-01 3.61942828e-01 5.12977898e-01 -1.14445865e+00
-3.78612399e-01 6.63638413e-01 5.44194341e-01 3.57076108e-01
9.68646109e-01 4.49596122e-02 1.19465888e+00 -7.64898181e-01
6.11738384e-01 1.28856647e+00 1.01315401e-01 7.43584931e-01
-9.42637682e-01 -1.62466243e-01 3.09575617e-01 5.05857766e-01
-1.02399576e+00 -5.11831999e-01 5.65787852e-01 -2.95559734e-01
1.35771704e+00 6.61441386e-01 4.50597703e-01 9.58718181e-01
7.85209462e-02 6.52211130e-01 7.88104296e-01 -7.59533226e-01
4.80981022e-01 4.18675601e-01 8.88943195e-01 3.16310376e-01
-3.51653211e-02 -1.47425920e-01 -6.23960853e-01 -5.93413413e-01
4.69169654e-02 -5.33172607e-01 -2.51034707e-01 -1.61762885e-03
-7.75013506e-01 8.53456855e-01 1.74949437e-01 8.24190080e-01
-3.78656507e-01 -8.43371600e-02 7.06396699e-02 6.18794143e-01
1.26610100e-01 9.10961509e-01 -3.37463528e-01 -1.55087709e-01
-2.63341129e-01 5.97024322e-01 1.74475372e+00 6.93235040e-01
5.46001494e-01 -4.55333591e-01 -5.64692140e-01 8.28802407e-01
1.89722061e-01 7.33805150e-02 4.60520804e-01 -9.60225344e-01
5.41456044e-01 9.14195538e-01 3.22671831e-01 -7.63214171e-01
-2.94692457e-01 -1.31533831e-01 -1.26502901e-01 -3.55682075e-01
4.93484199e-01 -3.71015519e-01 -3.34953636e-01 1.57247484e+00
1.53250322e-01 -3.42901826e-01 4.89608586e-01 5.66540122e-01
1.27300143e+00 6.69027328e-01 2.99153924e-01 8.94671306e-02
1.59178746e+00 -9.92805064e-01 -7.14672804e-01 -5.41267097e-01
5.07126212e-01 -6.64975941e-01 9.62635577e-01 -1.11029394e-01
-1.04763591e+00 -4.12962735e-01 -8.09174597e-01 -2.05058828e-01
-5.55785894e-01 -6.36683730e-03 3.79197478e-01 3.65709484e-01
-1.36686385e+00 1.28692821e-01 -1.00501388e-01 -6.23501778e-01
-4.23472911e-01 1.51138768e-01 1.76997006e-01 1.60762370e-02
-1.51533389e+00 1.24849796e+00 1.40786380e-01 1.52660474e-01
-6.01717055e-01 -2.95181692e-01 -5.70274770e-01 5.36024094e-01
5.65604389e-01 -9.49225545e-01 1.66208386e+00 -8.63694668e-01
-1.69668818e+00 6.55707300e-01 -4.99872446e-01 -5.28612018e-01
1.14778444e-01 -1.34345219e-01 -1.56411052e-01 2.58372426e-01
-1.89589679e-01 4.66203004e-01 3.74502748e-01 -1.13019633e+00
-6.60191655e-01 -1.98595062e-01 7.28566408e-01 4.48665291e-01
5.39007150e-02 -3.12936166e-03 -4.25826341e-01 -1.30478919e-01
-1.23143665e-01 -9.69868243e-01 -3.54249328e-02 -6.19595468e-01
6.54151961e-02 -7.05757916e-01 4.50406551e-01 -7.29715228e-01
1.25220215e+00 -1.67807472e+00 2.27962375e-01 1.67567655e-01
9.23315808e-02 2.46330142e-01 -2.45669127e-01 8.34337652e-01
4.29065973e-02 1.22825369e-01 2.31191032e-02 -1.70094460e-01
-7.87472799e-02 1.29032671e-01 -3.95101011e-01 -5.73756933e-01
1.98482171e-01 1.01201344e+00 -7.86214650e-01 -2.97265649e-01
-3.31512541e-02 1.43647209e-01 -4.87209946e-01 7.17615843e-01
-7.11944103e-01 1.96232870e-01 -5.66582620e-01 2.05202267e-01
-2.77140528e-01 -2.79011101e-01 3.71718675e-01 -2.67736206e-04
1.54262841e-01 6.32201016e-01 -7.34119773e-01 1.13058460e+00
-7.45427728e-01 8.18902254e-01 2.51781881e-01 -6.55770838e-01
9.38266456e-01 6.57842517e-01 1.10321216e-01 -7.79022038e-01
1.45874277e-01 9.86700691e-03 3.47022802e-01 -7.91329026e-01
5.69060206e-01 1.24390319e-01 5.02466373e-02 8.71231735e-01
-5.68450242e-02 -2.07783625e-01 3.32612664e-01 5.77132940e-01
1.05214679e+00 -3.09436977e-01 3.56219292e-01 -2.56720871e-01
1.04383469e+00 2.78684348e-01 -2.77721405e-01 1.00081515e+00
-1.78861823e-02 1.80202857e-01 6.56099856e-01 -3.67342174e-01
-4.31282043e-01 -6.02058709e-01 4.31609809e-01 1.37648141e+00
1.03636742e-01 -4.07250196e-01 -9.68028963e-01 -4.96336222e-01
-2.35331476e-01 1.20103931e+00 -5.17178655e-01 -2.73910165e-01
-5.75887859e-01 -4.37975287e-01 4.40374345e-01 1.93240747e-01
3.63913178e-01 -1.47477269e+00 -6.95865571e-01 2.71668434e-01
-7.63114154e-01 -9.29409504e-01 -1.68121234e-01 -1.38866259e-02
-6.42721593e-01 -1.29608202e+00 -3.48071724e-01 -6.29088461e-01
3.42361659e-01 3.81014526e-01 1.38834679e+00 5.11433482e-01
-4.99576777e-02 1.05913043e+00 -4.08465385e-01 -4.83832806e-01
-8.36659193e-01 4.93574947e-01 -6.24190688e-01 3.78945097e-02
5.88452041e-01 -3.89822900e-01 -3.77899498e-01 3.65378022e-01
-8.89160812e-01 -4.09232453e-02 2.27144912e-01 5.42571723e-01
-1.66656390e-01 -5.01020312e-01 9.40702915e-01 -7.29860485e-01
1.67993176e+00 -6.35213852e-01 -3.23613524e-01 8.72286975e-01
-6.25461638e-01 3.84824902e-01 2.38512814e-01 -1.47686020e-01
-1.52199340e+00 -4.50092614e-01 -2.10070938e-01 4.30609554e-01
-2.12229386e-01 5.71568191e-01 -5.89230619e-02 8.77867118e-02
9.83520329e-01 7.49535114e-02 6.62353188e-02 -3.35756361e-01
4.29162085e-01 7.86221206e-01 2.25920126e-01 -8.40821087e-01
2.57633090e-01 -4.70743440e-02 -5.39601564e-01 -9.24773395e-01
-6.17642581e-01 -4.82188433e-01 -1.39899060e-01 -5.40786803e-01
8.97630990e-01 -4.73990470e-01 -9.76122677e-01 1.14820965e-01
-1.33142292e+00 -3.55915666e-01 -2.28880748e-01 4.29986343e-02
-4.21564460e-01 2.41126761e-01 -4.59521860e-01 -1.00470150e+00
-6.51232183e-01 -1.05722141e+00 6.33890748e-01 3.76695871e-01
-8.77328992e-01 -9.28881109e-01 1.99751586e-01 9.40013111e-01
9.43423748e-01 -2.87315726e-01 1.55595112e+00 -1.32487750e+00
-6.09850645e-01 -1.20731577e-01 -8.19959342e-02 6.16401508e-02
-1.51832983e-01 -2.69578695e-01 -1.07487535e+00 1.16059124e-01
1.57838047e-01 -2.04742387e-01 4.32774812e-01 -2.43612289e-01
6.42026186e-01 -2.27737620e-01 -3.59755121e-02 -4.59682852e-01
9.47482407e-01 6.01126492e-01 4.05753851e-01 1.71592757e-01
7.30395764e-02 1.15401018e+00 1.53421402e-01 -3.97511274e-02
7.32094049e-01 5.98288894e-01 1.98364213e-01 3.62445831e-01
-1.15318254e-01 -9.23592523e-02 1.91455588e-01 5.93877673e-01
2.23382175e-01 -3.83399874e-01 -1.04389167e+00 7.31079519e-01
-1.98162854e+00 -7.98481703e-01 1.73649907e-01 1.87576830e+00
6.52356982e-01 -7.03018233e-02 -5.57311438e-02 -2.38742173e-01
4.79903072e-01 4.11662906e-02 -5.20040214e-01 -7.74811625e-01
1.91652447e-01 2.41444364e-01 -4.39818025e-01 1.07061613e+00
-3.26357096e-01 7.78059065e-01 6.98449564e+00 3.10902148e-01
-6.02307558e-01 3.28225791e-02 3.61979276e-01 -1.33832261e-01
-5.85621357e-01 1.53891727e-01 -6.33283734e-01 1.07946500e-01
1.25904441e+00 -5.69476008e-01 6.07829273e-01 7.39818454e-01
-8.68712515e-02 -4.86094266e-01 -1.35065782e+00 5.21880507e-01
2.54488528e-01 -1.37074673e+00 4.27827865e-01 -3.59257460e-01
1.89680964e-01 -1.09867617e-01 -6.99900210e-01 7.12360263e-01
3.18265229e-01 -7.79560626e-01 3.28919351e-01 8.59035790e-01
-1.63538203e-01 -5.57715654e-01 7.38538980e-01 5.19079506e-01
-6.14835322e-01 -2.94957519e-01 1.72255918e-01 -2.11715654e-01
8.71605054e-02 5.26039824e-02 -1.09507167e+00 5.11709869e-01
4.46565241e-01 -1.92100182e-01 -7.53095508e-01 6.08129621e-01
-2.15398937e-01 6.20329440e-01 -3.19314867e-01 -6.16096556e-01
1.64855748e-01 -2.16850370e-01 6.10162020e-01 1.09641814e+00
-1.48007691e-01 4.53725964e-01 -2.98665129e-02 8.10827971e-01
9.76243466e-02 2.29267210e-01 -4.09960896e-01 4.09670733e-02
3.99497300e-01 8.91465306e-01 -3.10110837e-01 -5.37662029e-01
-3.46216530e-01 5.34339428e-01 4.13033575e-01 5.97219467e-01
-2.53070325e-01 -4.17428225e-01 4.98771876e-01 9.11647677e-02
-3.30460258e-02 -1.13310754e-01 -3.65094393e-02 -1.03000748e+00
1.85463294e-01 -1.13682401e+00 6.76420450e-01 -1.00232720e+00
-1.01477742e+00 8.53284299e-01 3.39966923e-01 -2.08970293e-01
-9.90617692e-01 -3.76170993e-01 -9.70727205e-01 1.04262650e+00
-1.21988857e+00 -6.18523955e-01 -4.74095851e-01 4.48392600e-01
6.68148816e-01 -2.16840014e-01 1.02786636e+00 1.21985659e-01
-3.56846184e-01 2.48645931e-01 -3.44847739e-01 -2.24324450e-01
4.40201014e-01 -9.46904778e-01 2.92910069e-01 4.29183245e-01
2.19161272e-01 8.68074358e-01 6.79286063e-01 -3.29499543e-01
-1.24983203e+00 -5.10344088e-01 1.41748083e+00 -7.28086352e-01
4.46035624e-01 -1.02408618e-01 -1.15449965e+00 5.47411919e-01
6.45469546e-01 -1.18377531e+00 8.42727184e-01 3.01071554e-01
-6.29458725e-02 5.50420210e-02 -9.31573212e-01 9.25620437e-01
6.30011261e-01 -7.91520059e-01 -1.22553968e+00 1.82076544e-01
7.65713215e-01 -3.08763850e-02 -5.01464188e-01 1.21106319e-01
4.12813634e-01 -9.56553042e-01 7.71558225e-01 -7.57282972e-01
2.10337698e-01 -1.72124252e-01 3.87557149e-02 -9.83214438e-01
-7.80690387e-02 -3.96684527e-01 -7.90771842e-02 9.46078777e-01
7.92409718e-01 -7.29689717e-01 3.44604611e-01 1.33064938e+00
2.50096411e-01 -8.29981804e-01 -7.40483224e-01 -6.11036550e-03
-6.23884611e-02 -2.64217257e-01 6.53573751e-01 4.17715311e-01
2.99327254e-01 1.06028950e+00 -4.51842882e-03 -1.09858543e-01
-4.40438651e-02 3.65630299e-01 7.40880251e-01 -1.16817951e+00
-2.36002028e-01 -5.50240576e-01 3.34062964e-01 -9.81508434e-01
2.52624631e-01 -7.48634934e-01 2.24786438e-02 -1.81116498e+00
-9.35367867e-02 1.03894360e-01 2.37516955e-01 1.83135644e-01
-4.83328044e-01 -4.66189325e-01 3.22485089e-01 1.87071159e-01
-5.24904609e-01 2.93341815e-01 8.30677271e-01 -7.58631080e-02
-3.35632861e-01 2.99370199e-01 -1.02620840e+00 5.43568194e-01
8.40752661e-01 -1.18871070e-01 -7.38919914e-01 -5.17467439e-01
4.26738709e-01 3.19757819e-01 2.94456184e-01 -8.63087773e-01
7.32460439e-01 9.99623328e-04 -2.58360296e-01 -4.38807756e-01
5.05707979e-01 -7.16351628e-01 -1.45581029e-02 3.57770979e-01
-1.01335430e+00 1.42307937e-01 2.03314304e-01 3.88163716e-01
-3.55309576e-01 -6.51764870e-01 3.39147747e-01 -2.49026537e-01
-5.65300643e-01 -3.60078990e-01 -1.05974793e+00 1.36273324e-01
6.63028061e-01 6.30361140e-02 -2.35836282e-01 -1.09880769e+00
-8.93460691e-01 5.48132777e-01 -2.39830494e-01 8.17386985e-01
2.94750750e-01 -7.09134758e-01 -6.90737545e-01 -9.31371152e-02
2.54046828e-01 -3.02785069e-01 9.02586505e-02 1.46067962e-01
-3.23870182e-01 9.37067866e-01 1.52187333e-01 -2.13293761e-01
-1.29175270e+00 2.19938487e-01 6.78258121e-01 -6.28398836e-01
-8.06051269e-02 5.27251840e-01 2.44643036e-02 -5.20009279e-01
4.57067370e-01 -2.93782949e-01 -7.80805469e-01 2.83943295e-01
6.67311907e-01 2.85143346e-01 2.50666440e-01 -1.70453474e-01
-2.80537337e-01 1.03857376e-01 -2.22257257e-01 -4.56381649e-01
7.61431456e-01 -1.94013968e-01 -2.21869171e-01 2.73941785e-01
7.56911635e-01 -3.21777374e-01 -3.12130809e-01 -6.04199409e-01
4.06894058e-01 1.48759425e-01 -3.52651656e-01 -1.12083364e+00
-3.41077983e-01 7.16248453e-01 5.95641769e-02 7.20269144e-01
7.65985608e-01 7.28283003e-02 4.83532250e-01 1.12642419e+00
2.48751879e-01 -8.65617335e-01 5.90173267e-02 9.08403814e-01
1.29438341e+00 -9.06215370e-01 -2.20519409e-01 -2.29678556e-01
-7.22192645e-01 1.06266868e+00 7.69717097e-01 2.57560939e-01
4.75431323e-01 -1.83616728e-01 3.53305727e-01 -6.03020906e-01
-1.29491448e+00 -2.08820552e-01 2.50202149e-01 2.10647240e-01
5.68689466e-01 -2.29324341e-01 -5.33250809e-01 6.63836896e-01
-1.87946573e-01 -1.12138622e-01 3.07719916e-01 1.01431322e+00
-3.89815122e-01 -1.07672989e+00 -2.71519154e-01 2.80289710e-01
-2.05088913e-01 -1.23816587e-01 -1.22211516e+00 6.69200063e-01
-4.55097437e-01 1.65433288e+00 -4.63283025e-02 -1.84688717e-01
7.16682374e-01 8.58414173e-01 7.89945647e-02 -6.69476211e-01
-1.36531675e+00 -6.83328271e-01 7.66350627e-01 -4.40789014e-01
-2.39295140e-01 -3.50435883e-01 -8.42515886e-01 -1.62411463e-02
-3.22510391e-01 8.71221840e-01 6.89876914e-01 1.27856123e+00
7.46661365e-01 3.87944788e-01 2.21037358e-01 -1.92383841e-01
-5.97102582e-01 -1.00115955e+00 3.37828875e-01 3.83961111e-01
3.95414010e-02 -1.41843617e-01 -4.67235178e-01 -1.18860610e-01] | [12.119818687438965, 7.918738842010498] |
fb477c0b-dfb0-4178-888d-be255b0c92bd | towards-smooth-video-composition | 2212.07413 | null | https://arxiv.org/abs/2212.07413v1 | https://arxiv.org/pdf/2212.07413v1.pdf | Towards Smooth Video Composition | Video generation requires synthesizing consistent and persistent frames with dynamic content over time. This work investigates modeling the temporal relations for composing video with arbitrary length, from a few frames to even infinite, using generative adversarial networks (GANs). First, towards composing adjacent frames, we show that the alias-free operation for single image generation, together with adequately pre-learned knowledge, brings a smooth frame transition without compromising the per-frame quality. Second, by incorporating the temporal shift module (TSM), originally designed for video understanding, into the discriminator, we manage to advance the generator in synthesizing more consistent dynamics. Third, we develop a novel B-Spline based motion representation to ensure temporal smoothness to achieve infinite-length video generation. It can go beyond the frame number used in training. A low-rank temporal modulation is also proposed to alleviate repeating contents for long video generation. We evaluate our approach on various datasets and show substantial improvements over video generation baselines. Code and models will be publicly available at https://genforce.github.io/StyleSV. | ['Bolei Zhou', 'Yinghao Xu', 'Yujun Shen', 'Ceyuan Yang', 'Qihang Zhang'] | 2022-12-14 | null | null | null | null | ['video-generation', 'single-image-generation', 'video-understanding'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 3.80055338e-01 2.05881119e-01 -7.53275156e-02 -7.35508231e-03
-9.15049434e-01 -8.21512282e-01 8.17731321e-01 -6.55303180e-01
5.41288033e-02 8.14793587e-01 3.82439882e-01 -1.80847794e-01
2.88935721e-01 -6.02086246e-01 -1.08874488e+00 -6.99714959e-01
-1.93035483e-01 -1.39509499e-01 1.54711023e-01 -9.72000360e-02
-8.69445801e-02 2.79090226e-01 -1.23325121e+00 4.16363776e-01
7.98573077e-01 8.32386553e-01 1.21111974e-01 1.10717988e+00
3.00889403e-01 1.11353695e+00 -7.02401578e-01 -4.69449550e-01
6.00976944e-01 -8.82511675e-01 -5.93317389e-01 3.09475273e-01
6.41892314e-01 -8.34403455e-01 -7.86622524e-01 7.54905164e-01
3.27663660e-01 2.48284131e-01 5.49287021e-01 -1.42648423e+00
-6.98694289e-01 6.25689626e-01 -3.70698243e-01 2.95523386e-02
4.47906494e-01 6.35134161e-01 8.58665347e-01 -6.64526820e-01
7.93834567e-01 1.25396919e+00 5.37023604e-01 8.30755055e-01
-1.16725469e+00 -7.45898783e-01 2.62316972e-01 6.09246492e-02
-1.29475963e+00 -6.99760854e-01 7.46599853e-01 -4.61932868e-01
5.97286224e-01 4.37048435e-01 6.85961068e-01 1.44126379e+00
1.59842879e-01 5.81807137e-01 5.53907275e-01 -1.73789933e-01
-6.77603036e-02 -4.60245460e-01 -5.71538508e-01 5.67627549e-01
-8.87445435e-02 3.31677973e-01 -4.99376625e-01 -2.57778000e-02
1.23073828e+00 -1.18349232e-01 -5.44596672e-01 -1.89378168e-02
-1.28715634e+00 6.74288392e-01 1.21090211e-01 7.37193152e-02
-3.72044623e-01 7.18320191e-01 2.38707989e-01 2.93960065e-01
3.29570472e-01 3.68753910e-01 -8.89847651e-02 -3.18332434e-01
-1.23171020e+00 5.67731321e-01 5.29714525e-01 1.25888503e+00
4.74272788e-01 5.46043932e-01 -7.03827977e-01 5.17513216e-01
-3.66514064e-02 4.74380344e-01 2.62069046e-01 -1.63849258e+00
4.57925946e-01 -1.02976561e-01 2.59964794e-01 -9.30212021e-01
1.92645654e-01 -1.64833516e-01 -1.03753316e+00 1.27336338e-01
4.85337585e-01 -3.73336107e-01 -1.03209054e+00 2.01960802e+00
1.49080649e-01 8.22690904e-01 -6.68146759e-02 8.40958476e-01
4.01158273e-01 9.85238731e-01 5.01468107e-02 -3.42747748e-01
1.11195219e+00 -1.07552588e+00 -7.77534902e-01 8.46004486e-02
2.61538565e-01 -9.08083737e-01 8.29124033e-01 3.40625226e-01
-1.58662391e+00 -7.23611414e-01 -9.79828715e-01 -1.46469489e-01
2.50585258e-01 -5.62371016e-02 4.10704523e-01 5.01885533e-01
-1.26792133e+00 7.11278439e-01 -9.53551590e-01 1.53265938e-01
3.77952784e-01 1.24486610e-01 -2.46219993e-01 -5.57995848e-02
-1.32847297e+00 3.04822117e-01 2.59016752e-01 -8.30863044e-02
-1.27305889e+00 -9.18360233e-01 -9.32003200e-01 -4.54924703e-02
3.82426202e-01 -1.11205399e+00 1.27574766e+00 -1.16865051e+00
-1.62019038e+00 2.79849559e-01 -2.81487793e-01 -6.67119861e-01
8.18230271e-01 -2.91752666e-01 -3.10841054e-01 3.94545227e-01
-2.91578900e-02 1.11899877e+00 1.34747505e+00 -1.37301195e+00
-4.49439257e-01 4.01805937e-01 3.85021627e-01 1.51051447e-01
-1.48452207e-01 -2.12658882e-01 -7.86488950e-01 -1.43827760e+00
-3.09409678e-01 -1.24874651e+00 -3.04685116e-01 -4.51895669e-02
-3.25119317e-01 2.45043740e-01 7.14341581e-01 -1.03738022e+00
1.41460967e+00 -2.07002878e+00 3.52009535e-01 5.74046895e-02
1.47622600e-01 1.59485430e-01 -3.49982888e-01 3.71067733e-01
-8.08690116e-02 2.19468907e-01 -2.36526728e-01 -4.75861609e-01
-1.11814611e-01 2.01485172e-01 -5.52382827e-01 1.00928664e-01
4.28900570e-01 1.11973643e+00 -8.66502941e-01 -2.47880712e-01
3.56983282e-02 5.82405269e-01 -1.00121653e+00 3.49393040e-01
-3.86923730e-01 8.62517476e-01 -1.27590030e-01 3.97492260e-01
5.87621093e-01 -3.01533431e-01 1.60430685e-01 -3.08949620e-01
1.41546085e-01 2.09251627e-01 -1.08099902e+00 1.91961873e+00
-4.85125005e-01 5.63169777e-01 -1.06917299e-01 -5.80573499e-01
5.02021968e-01 6.09741330e-01 5.79935968e-01 -4.34129775e-01
-6.24545198e-03 -4.19482924e-02 -5.58375157e-02 -1.99606657e-01
5.77664971e-01 -2.16633882e-02 -2.43800785e-02 3.60705078e-01
-6.45575821e-02 -5.11792526e-02 4.49123323e-01 3.63330454e-01
1.04791319e+00 4.59598243e-01 -7.00245053e-02 1.11051120e-01
3.99322301e-01 -3.84276062e-01 7.19430506e-01 6.68469965e-01
5.65909408e-02 1.04572332e+00 5.48973143e-01 -2.49309734e-01
-1.42711580e+00 -1.13170242e+00 3.32762003e-01 7.36886621e-01
1.36070132e-01 -5.63139021e-01 -9.02408183e-01 -4.11474526e-01
-3.19508582e-01 6.45963073e-01 -4.47405428e-01 -2.39439145e-01
-9.73161161e-01 -3.43327850e-01 6.06621325e-01 5.67041099e-01
4.56991613e-01 -8.35484326e-01 -3.62015784e-01 3.53730828e-01
-4.63519096e-01 -1.25261986e+00 -1.11130869e+00 -4.81984377e-01
-6.89066350e-01 -7.66803324e-01 -1.13521349e+00 -6.38077736e-01
5.58820665e-01 9.05989036e-02 1.20674908e+00 1.10847391e-01
2.77263727e-02 2.78252333e-01 -3.94351155e-01 1.31890297e-01
-7.27948487e-01 -1.84251279e-01 -9.36529562e-02 -6.15868047e-02
-4.63942260e-01 -8.62666607e-01 -8.74811471e-01 2.35121101e-01
-1.15607429e+00 5.94829381e-01 2.94442803e-01 9.48545814e-01
4.67372179e-01 -4.86431532e-02 4.12917435e-01 -6.31382823e-01
2.63535321e-01 -5.10085464e-01 -4.69830692e-01 1.08472236e-01
-1.43585518e-01 -1.05224857e-02 8.50628018e-01 -8.23420584e-01
-1.10778368e+00 -1.86281011e-01 -1.09680116e-01 -9.13791955e-01
5.49830720e-02 -1.05281752e-02 -1.35279939e-01 1.31606549e-01
4.48767632e-01 3.76580387e-01 -2.11199583e-03 -1.29730567e-01
6.65486932e-01 6.65983483e-02 7.88537323e-01 -6.55850291e-01
1.04784048e+00 4.81639832e-01 -1.02221474e-01 -4.15587664e-01
-4.77690697e-01 1.37526512e-01 -1.96257263e-01 -3.31531048e-01
8.30075800e-01 -1.26228738e+00 -4.77627069e-01 6.37657225e-01
-1.22298443e+00 -8.05362523e-01 -2.23895416e-01 3.17961305e-01
-8.05072904e-01 4.30284828e-01 -9.36468244e-01 -4.78951722e-01
-2.60788947e-01 -1.10399258e+00 1.08972287e+00 1.03713274e-01
-2.81375587e-01 -8.70449245e-01 -3.97497006e-02 2.75402546e-01
3.28231871e-01 5.09768426e-01 4.72944528e-01 6.20423779e-02
-1.08920062e+00 1.87675580e-01 7.82142952e-02 4.66831475e-01
3.62507582e-01 1.25347644e-01 -6.25302017e-01 -5.64602673e-01
-5.09026758e-02 -6.60855994e-02 9.10927594e-01 3.87243479e-01
1.26861751e+00 -8.07128370e-01 -7.25122020e-02 1.01179636e+00
1.19746792e+00 3.97031963e-01 9.20218766e-01 2.25753020e-02
1.04363704e+00 2.37347871e-01 5.03802478e-01 5.66520870e-01
3.82581860e-01 7.30247915e-01 1.46710619e-01 9.06533282e-03
-6.26119971e-01 -4.53703672e-01 6.39925539e-01 7.42488086e-01
-3.87128353e-01 -6.00946546e-01 -5.51732004e-01 3.90744686e-01
-1.76852930e+00 -1.36755240e+00 2.24681213e-01 2.07617188e+00
9.86877739e-01 7.41655678e-02 2.45201305e-01 3.49028520e-02
7.19796598e-01 3.48396450e-01 -4.80861247e-01 -3.69442292e-02
-1.28168210e-01 2.57636249e-01 5.28382421e-01 6.88498616e-01
-1.08105564e+00 9.67068493e-01 6.16916132e+00 1.02230287e+00
-1.12279403e+00 4.59459685e-02 8.55020404e-01 -3.80681932e-01
-6.45739615e-01 7.97246173e-02 -5.87038696e-01 8.99664342e-01
9.14793551e-01 -2.53359795e-01 6.69039726e-01 3.48907709e-01
6.04739189e-01 2.80316353e-01 -9.86340582e-01 8.61395538e-01
-6.50915727e-02 -1.63937747e+00 4.59489316e-01 -8.68539065e-02
1.16140997e+00 -5.18085897e-01 1.68427378e-01 1.76506385e-01
2.80357152e-01 -1.03173661e+00 1.07260776e+00 6.40843093e-01
1.08932936e+00 -7.53268540e-01 2.57753521e-01 3.02680116e-02
-1.33720839e+00 2.15189885e-02 -4.49612066e-02 1.05193362e-01
7.04078794e-01 4.05260116e-01 -4.58359569e-01 6.47480130e-01
3.25066388e-01 7.61824608e-01 -3.28958929e-01 6.54715955e-01
-2.35090047e-01 6.47042930e-01 -1.61897421e-01 7.16986001e-01
2.55388826e-01 -2.27481231e-01 7.09315598e-01 1.00981867e+00
6.40073776e-01 2.41787672e-01 9.92398709e-02 8.53864789e-01
-1.14277497e-01 -4.07210827e-01 -5.64583242e-01 -1.82512015e-01
4.85269129e-01 9.47991073e-01 -4.11284357e-01 -5.21098793e-01
-3.78328472e-01 1.31182837e+00 -9.78006274e-02 6.59095645e-01
-1.31345856e+00 1.23977214e-02 8.76044452e-01 1.87214836e-01
4.78645086e-01 -4.29086417e-01 -4.28883918e-02 -1.29684162e+00
1.89398840e-01 -1.28011680e+00 1.62900403e-01 -6.61699414e-01
-1.01753795e+00 5.34378469e-01 -2.85165571e-02 -1.48117411e+00
-7.82056153e-01 -4.82559539e-02 -6.31209493e-01 8.36615205e-01
-1.29408658e+00 -1.34693408e+00 -2.61069655e-01 6.76568389e-01
7.44241595e-01 1.57383680e-02 4.51812953e-01 5.17525554e-01
-4.47561741e-01 8.38869750e-01 -2.74634689e-01 1.21455275e-01
8.05020213e-01 -9.48306501e-01 8.15478981e-01 1.22726238e+00
6.14739470e-02 5.91400027e-01 7.88497925e-01 -7.60799408e-01
-1.30405521e+00 -1.34721339e+00 4.18615580e-01 -2.85391033e-01
4.84964311e-01 -2.38208666e-01 -9.26247418e-01 7.19524562e-01
4.11647946e-01 -1.09628841e-01 4.05285895e-01 -6.73953712e-01
-3.75649780e-01 1.68428540e-01 -7.70467758e-01 9.08127367e-01
1.26185894e+00 -4.36825752e-01 -9.83447507e-02 1.13855585e-01
1.05144751e+00 -6.85769975e-01 -9.51961875e-01 4.12956893e-01
5.40011644e-01 -1.00245798e+00 1.08302414e+00 -4.79351997e-01
9.23711360e-01 -5.19470632e-01 1.93004701e-02 -1.12629974e+00
-3.57863188e-01 -1.31216919e+00 -5.59919894e-01 1.27302527e+00
1.45778194e-01 -2.19485357e-01 7.93546855e-01 6.19162977e-01
-1.39318749e-01 -5.83155692e-01 -5.68601906e-01 -9.52503622e-01
6.96082637e-02 -2.50022262e-01 5.84574163e-01 7.50785172e-01
-4.57811683e-01 3.96523587e-02 -1.05299354e+00 2.44613528e-01
5.30466557e-01 4.87679653e-02 9.32132304e-01 -3.38170856e-01
-7.85656929e-01 -2.49423191e-01 -2.00027749e-01 -1.45368326e+00
1.37257203e-01 -5.82285583e-01 2.98194382e-02 -1.17274296e+00
3.45861875e-02 -3.29018176e-01 -1.57000627e-02 2.74047285e-01
-4.72515911e-01 4.13778484e-01 6.06330335e-01 2.77397811e-01
-2.78120965e-01 6.81429863e-01 1.75051367e+00 1.10803079e-02
-1.30933270e-01 -1.92030162e-01 -5.30538917e-01 4.71304268e-01
8.16328168e-01 -2.36730337e-01 -7.74542153e-01 -5.12712538e-01
-8.84802714e-02 4.88075703e-01 5.52526772e-01 -1.07046020e+00
-9.46220569e-03 -2.41128787e-01 3.39255542e-01 -1.65697262e-01
5.06870806e-01 -4.89890784e-01 8.40045393e-01 3.50981385e-01
-4.16667730e-01 9.59834382e-02 2.44196862e-01 4.99180377e-01
-2.77365297e-01 9.22289714e-02 8.20613861e-01 -6.78673759e-02
-5.59162676e-01 7.00867534e-01 -1.57124013e-01 1.59145936e-01
9.33774889e-01 -1.44461051e-01 -1.78808883e-01 -8.74129593e-01
-7.71717072e-01 1.22405447e-01 7.90958464e-01 5.58487892e-01
4.85756099e-01 -1.63074338e+00 -7.35185921e-01 6.48391023e-02
-3.37746292e-01 -4.35237698e-02 6.21493399e-01 6.11653745e-01
-6.75162017e-01 1.52639553e-01 -2.14996383e-01 -3.53419602e-01
-1.08326495e+00 5.04763484e-01 1.00669727e-01 -2.62197495e-01
-6.65026844e-01 7.65574157e-01 4.22181904e-01 3.50815505e-01
3.13905291e-02 -2.02816993e-01 1.91128626e-01 -1.96032912e-01
6.45691693e-01 2.30549797e-01 -4.00401115e-01 -5.36489248e-01
9.26279798e-02 4.18356568e-01 1.03791609e-01 -3.85840952e-01
1.03984833e+00 -2.99786836e-01 2.38963485e-01 6.34920374e-02
1.12652409e+00 1.11546226e-01 -1.95078778e+00 9.35328081e-02
-5.08154154e-01 -5.83988369e-01 -2.61605740e-01 -3.86826932e-01
-1.24970090e+00 4.94992673e-01 2.87895501e-01 1.07576445e-01
1.23686969e+00 -4.00880873e-01 1.25865662e+00 -1.62430435e-01
3.22319239e-01 -6.64210021e-01 2.91663587e-01 4.33120012e-01
9.90498245e-01 -9.94971573e-01 -2.65536159e-01 -5.13058126e-01
-7.22155452e-01 9.88510549e-01 5.06536782e-01 -2.48419404e-01
2.82132745e-01 4.45290148e-01 -1.26023740e-01 3.67727518e-01
-1.02345312e+00 1.05831131e-01 3.68968904e-01 5.23889363e-01
4.24812853e-01 -1.16357431e-01 -2.38424271e-01 2.71860510e-01
-3.17084640e-01 5.24215028e-02 6.02110207e-01 6.82461917e-01
-1.25330146e-02 -1.37722051e+00 -3.38536590e-01 2.03978270e-01
-6.63085997e-01 -3.18025410e-01 1.39652997e-01 6.05178952e-01
1.96356878e-01 8.78233910e-01 2.28300646e-01 -2.37931341e-01
-1.22460850e-01 -7.81756118e-02 5.79358518e-01 -3.09399694e-01
-2.63585120e-01 3.84869635e-01 1.45848542e-01 -8.51471901e-01
-4.68023539e-01 -6.29241049e-01 -9.85932410e-01 -5.65404594e-01
9.62895527e-02 -1.31118372e-01 1.00977473e-01 5.33866644e-01
4.88723963e-01 8.26754451e-01 7.45894015e-01 -1.04814029e+00
-2.58785695e-01 -7.35254884e-01 -1.47897214e-01 6.99493945e-01
4.80394900e-01 -3.18994194e-01 -3.68928760e-01 8.93411934e-01] | [10.912787437438965, -0.5932552218437195] |
ec885dc7-3508-474b-8534-333a7fd7ffc3 | a-robust-regression-approach-for | 1412.5126 | null | http://arxiv.org/abs/1412.5126v2 | http://arxiv.org/pdf/1412.5126v2.pdf | A Robust Regression Approach for Background/Foreground Segmentation | Background/foreground segmentation has a lot of applications in image and
video processing. In this paper, a segmentation algorithm is proposed which is
mainly designed for text and line extraction in screen content. The proposed
method makes use of the fact that the background in each block is usually
smoothly varying and can be modeled well by a linear combination of a few
smoothly varying basis functions, while the foreground text and graphics create
sharp discontinuity. The algorithm separates the background and foreground
pixels by trying to fit pixel values in the block into a smooth function using
a robust regression method. The inlier pixels that can fit well will be
considered as background, while remaining outlier pixels will be considered
foreground. This algorithm has been extensively tested on several images from
HEVC standard test sequences for screen content coding, and is shown to have
superior performance over other methods, such as the k-means clustering based
segmentation algorithm in DjVu. This background/foreground segmentation can be
used in different applications such as: text extraction, separate coding of
background and foreground for compression of screen content and mixed content
documents, principle line extraction from palmprint and crease detection in
fingerprint images. | ['Yao Wang', 'Haoping Yu', 'Shervin Minaee'] | 2014-12-16 | null | null | null | null | ['foreground-segmentation'] | ['computer-vision'] | [ 6.80457950e-01 -4.29012388e-01 -1.39918417e-01 -1.15955330e-01
-2.16877133e-01 -4.52690899e-01 3.09588432e-01 -1.07781626e-01
-7.91798234e-02 6.47132576e-01 -3.42793703e-01 -3.72524709e-01
8.16886351e-02 -5.78266859e-01 -4.57866818e-01 -9.79324758e-01
3.72227609e-01 4.85456079e-01 7.80728638e-01 5.12530543e-02
5.85586667e-01 7.79816210e-01 -1.28541553e+00 3.42927516e-01
7.43244112e-01 7.47176111e-01 -1.01968581e-04 1.01496220e+00
-6.85486555e-01 9.74495947e-01 -6.14880621e-01 -2.58525103e-01
3.03076088e-01 -7.68188596e-01 -3.93360108e-01 5.85529506e-01
4.07506227e-01 -4.69854325e-01 -1.36267975e-01 1.24661779e+00
3.07038516e-01 -1.55228063e-01 8.41010153e-01 -9.20738161e-01
2.01779325e-03 1.94380060e-01 -1.33722913e+00 2.28741527e-01
2.24928662e-01 -4.40710157e-01 1.36823222e-01 -7.24267364e-01
5.96406519e-01 1.10069776e+00 6.82554483e-01 3.29884380e-01
-1.23682177e+00 -6.41502678e-01 -2.38343254e-01 3.52215245e-02
-1.36609757e+00 -4.48843658e-01 8.05768609e-01 -4.53568667e-01
4.06621724e-01 6.90041006e-01 6.51028693e-01 2.85721153e-01
3.27183694e-01 1.01755416e+00 1.11078954e+00 -7.82984018e-01
1.56764895e-01 4.20621991e-01 4.82042968e-01 5.58964789e-01
3.49378079e-01 -5.03766418e-01 6.37203604e-02 -1.99680954e-01
1.27834046e+00 1.12178616e-01 -4.62187648e-01 -3.20375055e-01
-8.93177450e-01 5.21968842e-01 -1.77152097e-01 6.46359801e-01
-2.06795782e-01 -4.07644846e-02 1.10640548e-01 1.02058738e-01
1.66269913e-01 -2.51567006e-01 -8.43196064e-02 -4.11296710e-02
-1.83577979e+00 1.13007918e-01 8.23334396e-01 1.02580178e+00
5.15209258e-01 -6.36582961e-03 -8.24106261e-02 9.58540142e-01
4.89280075e-01 7.16435194e-01 4.56236511e-01 -7.90753722e-01
3.37101668e-01 5.60191810e-01 2.16899380e-01 -1.21181381e+00
7.43731335e-02 2.12583065e-01 -5.81112087e-01 6.82273746e-01
6.71396017e-01 -1.00636035e-01 -1.23456895e+00 7.38818824e-01
3.79287571e-01 1.53532773e-01 -2.86766678e-01 5.23722053e-01
7.12499619e-01 9.71945047e-01 -5.75595021e-01 -5.33463001e-01
1.04607224e+00 -8.47926497e-01 -1.10162246e+00 1.84586480e-01
-1.36413395e-01 -1.35625553e+00 4.17385608e-01 1.02280724e+00
-1.32582438e+00 -6.41408443e-01 -9.56309438e-01 2.06781223e-01
-8.67484808e-02 4.16401058e-01 1.02381781e-01 1.06716311e+00
-6.85784340e-01 3.82505566e-01 -7.74094760e-01 -3.62908840e-01
2.37060845e-01 5.53970277e-01 8.58518630e-02 4.07297118e-03
-3.40558350e-01 5.19244790e-01 4.48363423e-01 2.82098681e-01
-2.87294745e-01 6.82704970e-02 -2.25446612e-01 -2.55349368e-01
3.88257056e-01 -8.39411542e-02 6.68962359e-01 -1.58542204e+00
-1.45256472e+00 8.60808790e-01 -3.59967679e-01 -2.31324151e-01
9.33380127e-01 -4.83684801e-02 -5.45931935e-01 3.29738557e-01
-3.12680215e-01 8.90946947e-03 1.49475431e+00 -1.27535868e+00
-8.28131199e-01 -3.42366755e-01 -7.62477338e-01 3.77647765e-02
1.36664823e-01 2.37479776e-01 -1.11348033e+00 -7.52848029e-01
5.61719358e-01 -7.82874227e-01 7.39257485e-02 -1.89808115e-01
-4.52116311e-01 2.90149987e-01 1.45727575e+00 -1.38018882e+00
1.41335261e+00 -2.16995716e+00 -1.80328220e-01 9.74418461e-01
3.74251371e-03 3.25994760e-01 5.04490614e-01 1.11392438e-01
7.92590454e-02 -9.77661908e-02 -3.13323975e-01 3.71100545e-01
-4.55952913e-01 -7.68842325e-02 1.94627959e-02 6.51288152e-01
-1.72715783e-01 2.18696669e-01 -2.88181514e-01 -1.00258529e+00
3.20651293e-01 4.47833568e-01 -8.20287615e-02 -7.43516907e-02
-7.60055333e-03 1.56835809e-01 -3.26714188e-01 9.76030767e-01
1.26062536e+00 1.87356487e-01 -9.42326039e-02 -2.37787485e-01
-1.57838061e-01 -5.64780474e-01 -1.92832780e+00 7.71116793e-01
3.76396954e-01 1.08296406e+00 4.56201971e-01 -7.53864706e-01
1.21021175e+00 1.47523716e-01 6.54848993e-01 -3.16054076e-01
4.10270989e-01 3.57140869e-01 1.40966298e-02 -5.00529528e-01
6.35481417e-01 2.98999399e-01 6.25472188e-01 6.05593249e-02
-3.71859819e-01 -2.25513756e-01 3.96243244e-01 -1.02150537e-01
6.01322353e-01 3.02119583e-01 1.86223537e-01 -1.83499202e-01
9.63459551e-01 2.32906286e-02 4.34957266e-01 6.61597967e-01
-1.96892545e-01 7.40652204e-01 4.79165435e-01 1.09642282e-01
-1.17456973e+00 -6.44989610e-01 -4.10150826e-01 6.33508086e-01
3.97117645e-01 2.58358233e-02 -1.27149165e+00 -2.16002777e-01
-1.13984533e-01 4.39081281e-01 -2.54725218e-01 4.77351487e-01
-8.12390029e-01 -6.55202031e-01 3.57820630e-01 7.90269151e-02
8.46488714e-01 -8.98277879e-01 -4.81217921e-01 3.99066806e-01
1.34562969e-01 -8.65086555e-01 -2.72573382e-01 1.59984276e-01
-1.17147934e+00 -1.15148830e+00 -1.26973879e+00 -9.61646199e-01
7.67852485e-01 5.06320059e-01 6.18611753e-01 1.48587778e-01
-5.14620900e-01 3.91941905e-01 -3.91494304e-01 -3.33986193e-01
-5.94164193e-01 -5.30674458e-01 -5.77163100e-01 4.14846271e-01
5.19669056e-01 1.13800816e-01 -3.40624839e-01 5.58347046e-01
-8.84938955e-01 1.46058295e-02 4.22150254e-01 6.76087916e-01
7.60018885e-01 6.37450814e-01 -3.78989995e-01 -1.17064154e+00
4.69137102e-01 -6.38451800e-02 -8.20325971e-01 3.31081986e-01
-2.86092013e-01 -4.58816350e-01 3.44556659e-01 -3.63677770e-01
-1.26814115e+00 1.22448549e-01 1.56497866e-01 -2.71349639e-01
-2.23868370e-01 -5.56412265e-02 -3.35173070e-01 -4.05063570e-01
3.82891148e-01 4.30941880e-01 -1.27526987e-02 -5.49372196e-01
1.16021097e-01 1.07033718e+00 6.51369750e-01 -1.35687038e-01
6.77837312e-01 5.57143033e-01 -8.88149068e-02 -1.59911609e+00
2.27447867e-01 -9.02376890e-01 -5.14393389e-01 -4.38273728e-01
1.00143099e+00 -3.31446707e-01 -1.09675907e-01 8.12313080e-01
-9.43752646e-01 -1.67447835e-01 3.61001678e-02 3.46523255e-01
-3.72021794e-01 8.25927734e-01 -6.29417658e-01 -1.22734606e+00
-1.63119316e-01 -1.07294726e+00 6.97461426e-01 6.55394673e-01
2.47870795e-02 -9.00965154e-01 -1.66576624e-01 3.20974886e-01
-1.87424757e-02 3.61347765e-01 7.57580042e-01 -5.20469606e-01
-5.80879033e-01 -5.18266082e-01 -1.79406762e-01 5.53373873e-01
2.25097135e-01 8.49374652e-01 -8.50666106e-01 -1.51550248e-01
3.04192275e-01 3.02501321e-01 8.19699168e-01 8.85289013e-01
9.36427593e-01 -3.25710066e-02 -5.77652156e-01 8.10535729e-01
1.74907804e+00 8.72788608e-01 1.11607850e+00 4.84522492e-01
8.05196881e-01 5.83493590e-01 7.08404064e-01 2.43905246e-01
-6.19740605e-01 4.01805311e-01 2.27432419e-03 -4.46028620e-01
1.10166492e-02 3.44082534e-01 3.01930964e-01 4.04426843e-01
-2.87883818e-01 -3.14642072e-01 -7.87243903e-01 1.35584101e-01
-1.74149728e+00 -1.20229936e+00 -1.02644718e+00 2.15763068e+00
5.41663945e-01 3.71314496e-01 3.50483567e-01 6.63716316e-01
1.25891006e+00 -2.55937994e-01 -2.53985703e-01 -5.48882186e-01
-3.47075671e-01 2.04062656e-01 9.72195506e-01 4.67115909e-01
-1.29917574e+00 8.02754521e-01 5.72554016e+00 1.17253399e+00
-1.16389847e+00 -2.12593481e-01 8.71782601e-01 3.02234858e-01
7.64208883e-02 -1.76157302e-03 -8.98804367e-01 7.28003263e-01
4.47957486e-01 3.02260220e-01 3.11140358e-01 6.44324243e-01
2.49531686e-01 -8.13278317e-01 -6.58840418e-01 1.33344257e+00
2.28007957e-01 -8.92852902e-01 -1.73132494e-01 8.18793029e-02
1.02290106e+00 -5.98969400e-01 -2.91872919e-02 -3.61997396e-01
-3.14584076e-01 -9.54411685e-01 6.97745800e-01 5.71737468e-01
4.26173270e-01 -6.81767404e-01 7.53162563e-01 2.95437038e-01
-9.08142090e-01 1.22771278e-01 -5.90472937e-01 3.62549663e-01
-1.08075462e-01 4.30363804e-01 -8.31767082e-01 2.01481566e-01
4.80436295e-01 3.14325422e-01 -6.02220058e-01 1.52470875e+00
3.94327343e-01 7.92937338e-01 -4.97153908e-01 -1.53127253e-01
2.30728984e-01 -9.25794303e-01 3.99737924e-01 1.52115536e+00
3.27972144e-01 -4.31101546e-02 -1.45079926e-01 7.93330193e-01
4.73365933e-01 6.65498555e-01 -2.86546201e-01 -1.38490751e-01
7.71139041e-02 1.06509221e+00 -1.61407530e+00 -5.19673645e-01
-5.54067194e-01 1.23038161e+00 -8.08008552e-01 4.71216083e-01
-8.69735897e-01 -6.65057182e-01 -2.43027791e-01 4.06962603e-01
6.20022833e-01 -9.24755335e-02 -5.43122530e-01 -8.13043714e-01
-7.76116103e-02 -1.18048096e+00 1.18531428e-01 -5.74526370e-01
-5.22450805e-01 3.60820532e-01 -9.46296304e-02 -1.25556219e+00
6.56314492e-02 -8.15061212e-01 -7.05788314e-01 9.84648466e-01
-9.10966873e-01 -9.21166301e-01 -4.20328259e-01 8.60207856e-01
7.53815889e-01 -5.78077257e-01 1.78918675e-01 2.46067733e-01
-5.94247520e-01 2.70800024e-01 9.34167147e-01 2.16471389e-01
6.55576169e-01 -1.18485105e+00 -5.86451404e-03 1.12007642e+00
7.53422230e-02 6.21711254e-01 8.63387465e-01 -1.05836201e+00
-9.20339465e-01 -5.54315627e-01 3.29833776e-01 -9.91547182e-02
1.24486066e-01 -1.19139150e-01 -9.53652561e-01 3.69270414e-01
6.75419196e-02 -2.98787326e-01 5.28204381e-01 -6.58388138e-01
4.81371641e-01 -1.61612138e-01 -1.29892397e+00 4.58883405e-01
2.31339317e-02 4.44058282e-03 -4.49865401e-01 1.72162205e-01
-4.22247201e-01 -4.02567267e-01 -3.77619326e-01 5.66596948e-02
7.39001691e-01 -1.35355163e+00 9.24028218e-01 6.84760138e-02
1.01082653e-01 -5.53136230e-01 1.11833550e-01 -3.41037035e-01
-1.81737676e-01 -8.10218632e-01 -7.98916221e-02 1.58169353e+00
1.35127246e-01 -1.21675044e-01 1.08078086e+00 4.57510918e-01
5.49241841e-01 -2.54252374e-01 -5.32880604e-01 -4.61513430e-01
-4.42536920e-01 4.58063483e-02 5.74160479e-02 7.62856007e-01
-5.06709814e-01 -9.15437117e-02 -4.61296886e-01 -2.60759503e-01
9.00122225e-01 -1.53253479e-02 9.97703254e-01 -1.26158726e+00
-4.09427017e-01 -5.84520757e-01 -4.07638073e-01 -9.83039081e-01
-4.73480284e-01 -3.12193066e-01 -9.51120909e-03 -1.50110471e+00
7.87331238e-02 -4.02400792e-01 1.41423896e-01 -3.01461369e-01
-1.70202449e-01 3.10960948e-01 2.44568422e-01 4.66495007e-01
-8.84907618e-02 -3.91299039e-01 1.02122426e+00 -1.29170805e-01
-5.68655789e-01 4.49242771e-01 -1.25068113e-01 1.07501316e+00
5.09005785e-01 -2.30796888e-01 -2.03817725e-01 2.20064238e-01
-1.66544840e-01 2.80210346e-01 -9.37529728e-02 -1.00029671e+00
1.63088217e-01 -1.82115391e-01 1.03695893e+00 -1.00710464e+00
7.13084266e-02 -1.32568073e+00 4.31007028e-01 3.99558574e-01
4.20324057e-02 -2.78138697e-01 8.55890661e-02 5.88204443e-01
-1.78550988e-01 -9.72838283e-01 1.08028758e+00 -2.92462856e-01
-8.51744771e-01 -2.49459490e-01 -6.40338540e-01 -4.52966154e-01
1.25794733e+00 -1.21677792e+00 1.89146668e-01 -4.21293080e-01
-6.07818663e-01 -2.13828936e-01 6.81885421e-01 -2.48504937e-01
6.76596463e-01 -1.06207299e+00 -5.75741172e-01 3.60852450e-01
-4.82870460e-01 -2.53174335e-01 -3.91236283e-02 9.86792326e-01
-1.67672980e+00 1.63659155e-01 -2.99916208e-01 -7.36146092e-01
-1.87979722e+00 4.51114357e-01 2.27698401e-01 1.97960645e-01
-5.19067585e-01 7.85399735e-01 -6.93848431e-02 5.32435715e-01
4.64813977e-01 -5.60996532e-01 -4.36436772e-01 -9.50667728e-03
5.79056263e-01 9.62764144e-01 -1.67560309e-01 -8.04393053e-01
-1.78313956e-01 1.03776085e+00 -2.54126191e-02 -2.30671018e-01
8.42529595e-01 -1.05449758e-01 -4.17017251e-01 6.21364236e-01
1.04488003e+00 7.70368814e-01 -1.13702774e+00 1.31785303e-01
1.77880645e-01 -7.13377416e-01 1.73097759e-01 -5.27012229e-01
-9.36603546e-01 7.82304823e-01 9.23024178e-01 6.57048643e-01
1.26448345e+00 -6.26274824e-01 6.56851828e-01 1.39927864e-01
-1.53285183e-03 -1.63668859e+00 -2.03722045e-01 9.13342610e-02
5.96466720e-01 -9.03131008e-01 3.04075062e-01 -5.36379755e-01
-6.13741994e-01 1.82020664e+00 3.30896616e-01 -4.64508116e-01
3.71124953e-01 6.08588696e-01 -3.19177248e-02 2.54799247e-01
1.81706056e-01 -6.84657842e-02 2.58617073e-01 4.82960194e-01
5.80928981e-01 -1.66241497e-01 -6.05154574e-01 6.97588027e-02
2.37342700e-01 -7.30875833e-03 6.92182958e-01 1.06373513e+00
-9.54221070e-01 -1.01086259e+00 -1.30537295e+00 6.59238219e-01
-9.64500368e-01 1.39148235e-01 -6.60839617e-01 1.02487648e+00
3.67508471e-01 8.19155276e-01 8.80372748e-02 1.34323575e-02
6.64246306e-02 1.14346072e-01 6.89463317e-01 -1.03694573e-02
-5.22625983e-01 1.28121722e+00 -1.58470139e-01 -1.85317248e-01
-5.10646880e-01 -7.56491244e-01 -1.23029888e+00 -2.66398638e-01
-6.63801312e-01 1.26475394e-01 7.84523726e-01 5.16768694e-01
-7.51280367e-01 4.16135073e-01 4.12226379e-01 -6.41206503e-01
-5.56576811e-02 -7.54956782e-01 -1.01586497e+00 3.52520227e-01
3.26606989e-01 -1.93754315e-01 -2.40200385e-01 6.82230294e-01] | [8.988982200622559, -0.8355847001075745] |
e26212ff-33af-40f0-bae9-95d8c7988cc2 | transformative-machine-learning | 1811.03392 | null | http://arxiv.org/abs/1811.03392v1 | http://arxiv.org/pdf/1811.03392v1.pdf | Transformative Machine Learning | The key to success in machine learning (ML) is the use of effective data
representations. Traditionally, data representations were hand-crafted.
Recently it has been demonstrated that, given sufficient data, deep neural
networks can learn effective implicit representations from simple input
representations. However, for most scientific problems, the use of deep
learning is not appropriate as the amount of available data is limited, and/or
the output models must be explainable. Nevertheless, many scientific problems
do have significant amounts of data available on related tasks, which makes
them amenable to multi-task learning, i.e. learning many related problems
simultaneously. Here we propose a novel and general representation learning
approach for multi-task learning that works successfully with small amounts of
data. The fundamental new idea is to transform an input intrinsic data
representation (i.e., handcrafted features), to an extrinsic representation
based on what a pre-trained set of models predict about the examples. This
transformation has the dual advantages of producing significantly more accurate
predictions, and providing explainable models. To demonstrate the utility of
this transformative learning approach, we have applied it to three real-world
scientific problems: drug-design (quantitative structure activity relationship
learning), predicting human gene expression (across different tissue types and
drug treatments), and meta-learning for machine learning (predicting which
machine learning methods work best for a given problem). In all three problems,
transformative machine learning significantly outperforms the best intrinsic
representation. | ['Oghenejokpeme I. Orhobor', 'Ross D. King', 'Joaquin Vanschoren', 'Ivan Olier'] | 2018-11-08 | null | null | null | null | ['explainable-models'] | ['computer-vision'] | [ 5.69259048e-01 1.69935539e-01 -5.98710299e-01 -3.62838328e-01
-9.26151276e-01 -3.31228584e-01 6.02363527e-01 4.21449095e-01
-1.99189290e-01 1.03657603e+00 1.32942051e-01 -3.00940067e-01
-2.58652747e-01 -7.60393262e-01 -9.77067351e-01 -7.12909937e-01
1.40679553e-01 5.85642219e-01 -3.37795794e-01 -2.05756560e-01
2.62849122e-01 5.79289913e-01 -1.41681790e+00 5.16638160e-01
7.88300872e-01 8.28499675e-01 1.12666324e-01 4.47575122e-01
-2.50626087e-01 8.58402669e-01 -5.70087612e-01 -2.20104828e-02
-8.43375847e-02 -3.73670489e-01 -9.73091722e-01 -1.06970817e-01
1.73592865e-01 1.59627259e-01 -1.24376053e-02 5.74485779e-01
2.83785313e-01 1.72433317e-01 1.00720966e+00 -1.09402359e+00
-7.69927204e-01 3.10199857e-01 -5.33991933e-01 5.24560511e-02
3.08135729e-02 -1.31561637e-01 1.21252823e+00 -9.50610757e-01
4.52531099e-01 1.23683453e+00 5.46364605e-01 7.78249383e-01
-1.75697851e+00 -6.42951190e-01 -3.09430715e-02 1.83067247e-02
-1.07127702e+00 -3.22405487e-01 5.66536725e-01 -5.56713104e-01
9.83453631e-01 2.20382273e-01 3.86964411e-01 1.09742773e+00
4.09921378e-01 7.73071289e-01 1.06443214e+00 -4.39118147e-01
2.26434112e-01 2.55447567e-01 2.42334545e-01 5.56527317e-01
2.96000332e-01 1.28738195e-01 -3.35336953e-01 -2.94833243e-01
6.68811798e-01 4.96845990e-01 -2.18568712e-01 -2.65521556e-01
-9.81679618e-01 1.05769861e+00 3.99229556e-01 6.19813323e-01
-3.43248606e-01 1.86617479e-01 3.43748420e-01 4.04523164e-01
7.06454635e-01 1.00284898e+00 -8.83203089e-01 2.37657025e-01
-7.83834398e-01 2.47377753e-01 6.29187703e-01 4.09728706e-01
1.06757987e+00 1.50517166e-01 1.23753123e-01 8.33651423e-01
1.75012335e-01 1.46196023e-01 9.51195121e-01 -5.62502384e-01
1.19946823e-01 7.73725808e-01 -9.32828039e-02 -9.50082362e-01
-5.12299895e-01 -5.22999942e-01 -1.17944860e+00 2.44400516e-01
5.10024965e-01 -1.07922498e-02 -9.94726002e-01 1.89982903e+00
-2.36926712e-02 3.57578658e-02 2.41555989e-01 5.67041099e-01
7.27928102e-01 9.49361145e-01 2.78062493e-01 -3.53282928e-01
1.03249776e+00 -7.04223752e-01 -6.56072557e-01 -4.05754060e-01
1.03644514e+00 -4.23196465e-01 7.95629442e-01 5.32552719e-01
-9.00593579e-01 -6.05884194e-01 -9.59225655e-01 -2.19566941e-01
-6.62531197e-01 5.64852357e-02 8.57429683e-01 2.11094037e-01
-6.99098408e-01 8.42407227e-01 -5.13163269e-01 -8.02166089e-02
6.88230693e-01 7.66434848e-01 -6.66364133e-01 -2.43199080e-01
-1.09755516e+00 1.05929065e+00 4.06928688e-01 -1.72294855e-01
-9.21267509e-01 -8.63465786e-01 -8.04419219e-01 2.41282046e-01
1.89683735e-01 -8.34211111e-01 9.61389005e-01 -1.42220342e+00
-1.25014377e+00 9.36933815e-01 -1.79889023e-01 -4.25458252e-01
5.53599652e-03 -1.68709874e-01 -1.39560625e-01 -2.52540082e-01
-1.12997100e-01 5.20077527e-01 8.56738806e-01 -1.19869387e+00
-4.03205335e-01 -5.41761637e-01 8.50200932e-03 -2.42472455e-01
-4.66866940e-01 -1.95275784e-01 1.65673241e-01 -6.76776409e-01
-1.23265557e-01 -9.00918245e-01 -6.02120757e-01 -3.23487744e-02
-2.48754531e-01 -4.45259899e-01 7.43151248e-01 -4.15695965e-01
8.98561597e-01 -2.06881261e+00 5.40658176e-01 2.16476440e-01
5.30017793e-01 5.12024641e-01 -4.07122374e-01 3.13718915e-01
-6.28143966e-01 4.44407851e-01 -1.79289699e-01 -8.31003636e-02
-4.74251539e-01 3.62866968e-01 -3.08467507e-01 3.27521324e-01
3.95587057e-01 1.08385623e+00 -8.53519738e-01 -1.24411754e-01
1.96902335e-01 4.92032200e-01 -4.86621231e-01 3.46579850e-01
-4.12216157e-01 4.65769708e-01 -7.06350088e-01 3.92043233e-01
2.75498986e-01 -5.91718614e-01 3.24892044e-01 -1.08314104e-01
1.32384196e-01 1.77461460e-01 -7.73261070e-01 1.55781817e+00
-8.46006632e-01 6.30943298e-01 -4.39990252e-01 -1.66776526e+00
1.09534550e+00 4.77404445e-01 7.51425922e-01 -6.50905311e-01
4.09231670e-02 4.02723312e-01 -2.87050121e-02 -3.19380134e-01
2.34399568e-02 -4.64560211e-01 1.27192419e-02 5.33520222e-01
1.70708403e-01 -6.50348216e-02 -1.21804051e-01 -8.73808935e-02
1.10072505e+00 -1.36261970e-01 8.25126469e-01 -1.84509411e-01
4.06247497e-01 -3.14957313e-02 7.41434693e-01 5.39984107e-01
2.97142833e-01 5.08442760e-01 5.77325881e-01 -9.61985707e-01
-1.11696136e+00 -5.82481384e-01 -1.69291601e-01 1.18055367e+00
-3.63772869e-01 -4.06421036e-01 -3.78637761e-01 -5.54062605e-01
1.82891637e-01 5.77883542e-01 -1.01474178e+00 -4.43096936e-01
-4.12652284e-01 -9.45898831e-01 1.99310526e-01 5.94690621e-01
-1.07186407e-01 -9.61756527e-01 -2.55044967e-01 4.13382828e-01
1.42152697e-01 -6.69467926e-01 7.05700964e-02 6.62632763e-01
-1.19949079e+00 -1.05856144e+00 -6.11145735e-01 -6.07163548e-01
7.58796990e-01 3.24026436e-01 1.01008058e+00 3.06605637e-01
-5.24432123e-01 -6.46030232e-02 -1.89603984e-01 -6.12862885e-01
-4.87080187e-01 2.08205596e-01 8.75325650e-02 -2.47644167e-03
1.73019916e-01 -5.98958015e-01 -1.68487519e-01 7.00148493e-02
-1.04892099e+00 3.22597891e-01 9.53034461e-01 1.38301969e+00
7.66901731e-01 -2.50256091e-01 1.05535126e+00 -1.24651396e+00
5.28453887e-01 -6.47281945e-01 -1.91199481e-01 3.58691633e-01
-7.10183501e-01 5.47126472e-01 8.69025767e-01 -4.51945245e-01
-6.27097905e-01 1.25049755e-01 -1.68642133e-01 -4.56024528e-01
-7.33530223e-02 8.16075623e-01 -1.06018506e-01 -1.07942438e-02
9.30640817e-01 2.04998642e-01 3.06165695e-01 -5.73663652e-01
3.68562847e-01 6.13741636e-01 1.41335204e-01 -4.86532390e-01
4.25428003e-01 5.70425093e-02 4.04894084e-01 -8.09844017e-01
-1.11450851e+00 -2.50306219e-01 -7.96815336e-01 2.37949207e-01
7.19318569e-01 -7.02289939e-01 -6.26318038e-01 8.40704441e-02
-1.11726868e+00 -3.05464119e-01 -3.60012263e-01 3.95563155e-01
-7.52527535e-01 -7.52363652e-02 -2.69189954e-01 -5.72288573e-01
-3.01583529e-01 -1.14001024e+00 9.05212581e-01 -6.13185875e-02
-4.76355672e-01 -1.31723237e+00 1.48477525e-01 3.05282712e-01
3.71583611e-01 5.44671476e-01 1.51628041e+00 -1.01152289e+00
-3.27428758e-01 -3.57856721e-01 -1.43551201e-01 2.93310463e-01
5.08574903e-01 -1.10720493e-01 -1.13847196e+00 -3.40242445e-01
-1.85657576e-01 -6.57643259e-01 1.08560669e+00 4.83096093e-01
1.81622171e+00 -5.08791924e-01 -5.30995488e-01 5.02875745e-01
1.35094881e+00 8.32836926e-02 5.09324074e-01 7.13800490e-02
6.34134531e-01 5.78297615e-01 3.19917977e-01 3.18213016e-01
-2.72748936e-02 8.77554715e-01 3.17774922e-01 -3.84945244e-01
8.17523897e-02 -1.42084852e-01 1.08499534e-01 5.00928283e-01
-2.52139121e-01 -1.61424458e-01 -8.07967663e-01 2.40269437e-01
-2.01385021e+00 -9.28804636e-01 -1.53945349e-02 2.17209125e+00
1.00942016e+00 -6.62816390e-02 2.78303996e-02 3.01522762e-01
3.73050869e-01 -1.76863670e-01 -7.81549394e-01 -5.95894039e-01
-1.26272544e-01 4.67007577e-01 1.09295852e-01 2.96736509e-01
-9.71460819e-01 7.12822795e-01 6.43510962e+00 9.77616549e-01
-1.31804967e+00 3.69201191e-02 1.07324135e+00 6.79952931e-03
-5.28200388e-01 -2.98617125e-01 -5.44764698e-01 3.79643477e-02
1.27870607e+00 -5.20139575e-01 2.92195559e-01 8.43055367e-01
2.33021647e-01 2.59185493e-01 -1.58901942e+00 1.02531111e+00
-1.54163599e-01 -1.82649970e+00 3.95864993e-01 1.51774496e-01
6.09758139e-01 -2.52718836e-01 1.43928096e-01 3.99838626e-01
1.87475085e-01 -1.57495081e+00 2.17254505e-01 6.23957813e-01
8.34324062e-01 -6.57026112e-01 6.83772922e-01 6.09400511e-01
-6.98610723e-01 -2.12655738e-01 -4.90019411e-01 -2.24855721e-01
-4.47812945e-01 7.85804391e-01 -9.04633105e-01 4.36793894e-01
1.74205974e-01 1.13861537e+00 -3.89289826e-01 7.18953013e-01
-3.12908776e-02 5.03081739e-01 2.40101013e-02 -5.53717688e-02
2.22670376e-01 9.38951522e-02 7.34938607e-02 1.11918914e+00
2.61069149e-01 1.97977602e-01 1.26382619e-01 9.03226376e-01
-2.40941182e-01 3.15941662e-01 -9.57843423e-01 -3.16308528e-01
3.36399004e-02 1.24141312e+00 -3.21681231e-01 -2.86112785e-01
-4.55908567e-01 5.83409607e-01 6.36183918e-01 3.03832442e-01
-2.54243582e-01 -4.70788628e-02 7.42147803e-01 1.49240093e-02
-1.56258881e-01 3.83181125e-02 -5.61126530e-01 -1.00963187e+00
-5.06255686e-01 -1.12670016e+00 5.33053339e-01 -5.61000645e-01
-1.30441189e+00 3.71502221e-01 -2.72743523e-01 -1.00564075e+00
-6.02357805e-01 -8.53437483e-01 -4.61417258e-01 9.58794892e-01
-1.42857945e+00 -1.21106887e+00 6.29910380e-02 3.82787317e-01
6.48799002e-01 -4.48664904e-01 1.31007457e+00 2.48327345e-01
-5.93397856e-01 5.86424589e-01 4.06207263e-01 -1.43012702e-01
5.74832380e-01 -1.21294856e+00 2.22539194e-02 1.09651163e-01
3.04866523e-01 7.28113353e-01 6.66258216e-01 -2.47122467e-01
-1.57730067e+00 -1.19299889e+00 8.30551088e-01 -4.27227706e-01
4.44012165e-01 -2.93815345e-01 -1.23568177e+00 6.97489083e-01
-2.24442706e-01 2.63129354e-01 1.23856139e+00 5.23802519e-01
-4.15187240e-01 -1.93805024e-01 -1.04964256e+00 2.95545191e-01
5.62827885e-01 -5.11199474e-01 -5.89208722e-01 5.49289048e-01
4.98091668e-01 -5.87210990e-02 -9.70679104e-01 3.42178017e-01
6.28369570e-01 -4.95105386e-01 1.04389429e+00 -1.40783715e+00
8.22280586e-01 2.32146727e-03 -1.91870898e-01 -1.60591197e+00
-5.23715675e-01 -1.90952644e-01 -2.61979938e-01 7.85463870e-01
7.96352804e-01 -5.18824697e-01 6.08869255e-01 8.15784156e-01
-1.35383502e-01 -1.31720281e+00 -7.40746856e-01 -5.59597313e-01
4.14872229e-01 -1.94537133e-01 4.59444106e-01 1.14223838e+00
6.08534440e-02 5.97400904e-01 -5.83336234e-01 -2.06165984e-01
3.09385568e-01 3.54044706e-01 7.79329538e-01 -1.63922369e+00
-3.46662849e-01 -3.81976992e-01 -4.77418542e-01 -7.75661886e-01
5.05951226e-01 -1.17420399e+00 -2.68014193e-01 -1.46910000e+00
4.28629547e-01 -5.22065520e-01 -5.00558078e-01 9.04013693e-01
-2.47851998e-01 -1.59050263e-02 -1.32952929e-02 3.53588223e-01
-1.85872883e-01 4.49293971e-01 1.26215851e+00 -3.94611746e-01
-1.71594322e-01 2.61605918e-01 -1.03775549e+00 6.38579488e-01
7.88711011e-01 -5.78272820e-01 -2.69379169e-01 -2.29768082e-01
2.94257283e-01 1.30495071e-01 3.12225699e-01 -6.36133313e-01
-8.83392245e-02 -4.70144659e-01 7.06165075e-01 -4.96957153e-02
2.99539000e-01 -6.98867500e-01 1.64919540e-01 6.04731202e-01
-8.59457612e-01 -2.53977209e-01 3.23833019e-01 5.81641853e-01
-2.61501789e-01 -2.42022887e-01 8.92301917e-01 -2.75829077e-01
-6.07180893e-01 4.61928755e-01 -3.02039415e-01 -4.39953692e-02
9.63140547e-01 -8.61193836e-02 -1.50658146e-01 -2.34999895e-01
-8.99415493e-01 -1.27603859e-01 7.34481290e-02 4.19299722e-01
7.12274611e-01 -1.27040136e+00 -7.65391231e-01 2.28691459e-01
2.29543865e-01 -1.71598807e-01 -6.09003752e-02 4.33430582e-01
5.53311035e-02 5.74009717e-01 -3.01062614e-01 -3.75189513e-01
-1.15751636e+00 6.90944195e-01 3.39631557e-01 -4.42830086e-01
-3.93728703e-01 5.88510811e-01 5.46262562e-01 -3.03982437e-01
-1.35570303e-01 -3.72439027e-01 -3.19616467e-01 7.51431286e-02
5.86892962e-01 2.51946356e-02 7.69637153e-02 -2.94483781e-01
-9.81341302e-02 4.95290220e-01 -3.29978824e-01 3.65838766e-01
1.94013870e+00 4.00090128e-01 -1.29193068e-01 6.74028695e-01
1.44341779e+00 -4.79916364e-01 -8.16667318e-01 -4.22355503e-01
1.09926768e-01 -3.73824894e-01 2.58096308e-01 -6.61918402e-01
-9.77284014e-01 1.27624202e+00 3.44266325e-01 8.52810219e-02
1.05169737e+00 -3.79592320e-03 4.32401925e-01 7.10901916e-01
2.98585594e-01 -8.42136383e-01 3.53132665e-01 3.03498238e-01
1.05668402e+00 -1.43164551e+00 9.13713276e-02 -2.41222292e-01
-4.14488524e-01 1.49272394e+00 4.19553190e-01 1.43781111e-01
5.78826010e-01 6.74463138e-02 -2.92610437e-01 -2.41025567e-01
-1.05497849e+00 3.20647955e-02 4.44018066e-01 5.06050169e-01
8.46198142e-01 -4.01381291e-02 -1.23428792e-01 7.76269853e-01
1.51554525e-01 2.89318234e-01 4.51322407e-01 5.65189958e-01
-5.58402956e-01 -1.27416825e+00 -1.34985998e-01 8.60605240e-01
-5.43354928e-01 -1.79976001e-02 -4.95771110e-01 6.72977388e-01
4.72063012e-02 7.04758048e-01 -8.05804804e-02 -3.10118675e-01
1.05794728e-01 1.57492310e-01 3.78265083e-01 -1.01834023e+00
-4.71528441e-01 -2.55000114e-01 -8.36939290e-02 -4.06135231e-01
-3.75150323e-01 -5.18223703e-01 -1.25696564e+00 -2.99070686e-01
-2.40845710e-01 2.21262559e-01 5.64289033e-01 1.31150889e+00
4.17062998e-01 5.90278268e-01 8.32365930e-01 -8.36612105e-01
-5.03219843e-01 -7.95803487e-01 -5.55813789e-01 4.92856771e-01
5.99542558e-01 -7.47793019e-01 -2.78855711e-01 1.53057709e-01] | [5.351639270782471, 5.64683723449707] |
4e166438-ba5f-4634-9811-ec6af5f5c268 | convgqr-generative-query-reformulation-for | 2305.15645 | null | https://arxiv.org/abs/2305.15645v2 | https://arxiv.org/pdf/2305.15645v2.pdf | ConvGQR: Generative Query Reformulation for Conversational Search | In conversational search, the user's real search intent for the current turn is dependent on the previous conversation history. It is challenging to determine a good search query from the whole conversation context. To avoid the expensive re-training of the query encoder, most existing methods try to learn a rewriting model to de-contextualize the current query by mimicking the manual query rewriting. However, manually rewritten queries are not always the best search queries. Training a rewriting model on them would limit the model's ability to produce good search queries. Another useful hint is the potential answer to the question. In this paper, we propose ConvGQR, a new framework to reformulate conversational queries based on generative pre-trained language models (PLMs), one for query rewriting and another for generating potential answers. By combining both, ConvGQR can produce better search queries. In addition, to relate query reformulation to retrieval performance, we propose a knowledge infusion mechanism to optimize both query reformulation and retrieval. Extensive experiments on four conversational search datasets demonstrate the effectiveness of ConvGQR. | ['Jian-Yun Nie', 'Kaiyu Huang', 'Yihong Wu', 'Yutao Zhu', 'Kelong Mao', 'Fengran Mo'] | 2023-05-25 | null | null | null | null | ['conversational-search'] | ['natural-language-processing'] | [ 1.13871962e-01 1.29053444e-01 -4.65966076e-01 -5.83870053e-01
-1.15445638e+00 -6.22834802e-01 8.09815705e-01 -2.10914403e-01
-2.94285744e-01 5.41425824e-01 6.25965893e-01 -4.10284787e-01
8.75056684e-02 -8.74373078e-01 -5.01433372e-01 -1.22195520e-01
5.29410481e-01 6.84182644e-01 2.15435416e-01 -7.18055546e-01
3.36751908e-01 1.25864998e-01 -1.33029771e+00 5.01499593e-01
1.09536731e+00 7.50631332e-01 7.82968879e-01 9.11490858e-01
-3.55299026e-01 8.40491116e-01 -9.20297503e-01 -4.23531741e-01
-1.27228037e-01 -7.69247174e-01 -1.19819164e+00 -2.29507431e-01
3.35559696e-02 -6.56530082e-01 -5.93896329e-01 5.56234717e-01
5.18259704e-01 3.52013588e-01 3.36550176e-01 -1.04306030e+00
-1.02486300e+00 8.11246812e-01 2.25081697e-01 2.88372904e-01
8.71026635e-01 8.76513124e-02 1.36557078e+00 -6.52456641e-01
6.28278077e-01 1.42492723e+00 -3.21972370e-02 7.76801527e-01
-9.17558491e-01 -5.08468330e-01 3.06004912e-01 3.25821072e-01
-1.35879123e+00 -5.12355745e-01 8.13884974e-01 5.69592901e-02
1.13584554e+00 6.27311945e-01 5.43165386e-01 1.10391271e+00
-1.79478396e-02 1.01425576e+00 4.81773704e-01 -3.38694245e-01
-1.07088856e-01 2.31575638e-01 -7.81369675e-03 4.71244216e-01
-5.07471979e-01 -1.16362303e-01 -5.01456499e-01 -4.16970730e-01
5.77536166e-01 5.92471659e-02 -4.93160933e-01 -2.64838841e-02
-8.95899653e-01 9.30230975e-01 5.72151780e-01 2.72807509e-01
-2.83289403e-01 1.71910122e-01 2.37720668e-01 4.35338974e-01
8.25414062e-02 1.04094481e+00 -4.00834531e-01 -2.79722959e-01
-4.66875285e-01 4.86036122e-01 1.10212886e+00 1.07076132e+00
9.09410477e-01 -4.55847621e-01 -8.01842093e-01 1.15548038e+00
3.86343449e-01 6.01047993e-01 6.16259515e-01 -1.08558989e+00
4.28638548e-01 7.43741274e-01 2.48071954e-01 -9.42175627e-01
1.18680716e-01 -5.23760378e-01 -4.00124460e-01 -9.70193326e-01
-1.09644070e-01 -5.19834831e-02 -6.36856616e-01 1.79502666e+00
-5.67417331e-02 -1.31551340e-01 3.74064118e-01 8.69066238e-01
1.07752752e+00 9.13224339e-01 -2.94767827e-01 -2.60166794e-01
1.19049799e+00 -1.26468074e+00 -8.01892817e-01 -4.44884419e-01
6.90576792e-01 -9.08203244e-01 1.33582640e+00 -1.32219955e-01
-9.33454454e-01 -5.54226398e-01 -6.54785275e-01 -3.71929079e-01
-1.36469603e-01 2.29541093e-01 6.03867888e-01 1.62926897e-01
-1.02887475e+00 -5.58049120e-02 -4.90187019e-01 -1.72051534e-01
-2.48832509e-01 2.97788292e-01 2.67146945e-01 -1.23493128e-01
-1.59780967e+00 7.51559675e-01 2.03648042e-02 1.72540218e-01
-7.73360848e-01 -3.33908349e-01 -6.72465026e-01 5.06275415e-01
5.70989132e-01 -9.09195721e-01 2.01174736e+00 -5.65632880e-01
-1.62553358e+00 5.46155512e-01 -8.62286091e-01 -3.14302385e-01
9.34446678e-02 -1.91663280e-01 -3.66226226e-01 1.13153510e-01
2.01989859e-01 7.35764742e-01 6.06052577e-01 -1.18187416e+00
-5.27013779e-01 -7.76160806e-02 6.95640206e-01 6.42013252e-01
7.45410621e-02 -1.67741999e-01 -1.20829844e+00 -4.36917454e-01
-4.14212048e-02 -9.70276356e-01 -1.97001934e-01 -6.27445579e-01
-4.48274165e-01 -8.13277304e-01 6.88591123e-01 -4.89368469e-01
1.75096202e+00 -1.97683275e+00 4.29654196e-02 1.55206010e-01
-4.80718538e-02 2.83303052e-01 -2.75291979e-01 8.70768666e-01
3.86551589e-01 1.85174853e-01 2.44919971e-01 -2.36398295e-01
1.41290590e-01 4.33143258e-01 -8.97412360e-01 -3.51805896e-01
-1.66287757e-02 1.46266842e+00 -9.34070051e-01 -4.92566913e-01
-2.40071923e-01 2.19218969e-01 -7.18283713e-01 6.93010569e-01
-6.99554205e-01 2.55608499e-01 -8.10296118e-01 3.44549596e-01
1.18882321e-01 -4.71818328e-01 1.08915649e-01 -4.72449549e-02
4.08771962e-01 7.51901150e-01 -6.21379554e-01 1.74689162e+00
-9.56560194e-01 5.41149676e-01 -1.15303673e-01 -6.23775125e-01
1.01145172e+00 3.87911409e-01 1.98101848e-01 -1.19430029e+00
-1.05618112e-01 1.24652743e-01 -1.66204184e-01 -7.63789654e-01
8.04904938e-01 4.66818623e-02 -3.08418453e-01 9.39417899e-01
-2.31399983e-01 -3.68192077e-01 2.02100888e-01 4.71337259e-01
9.21610534e-01 -1.55521721e-01 -1.82075836e-02 3.31823736e-01
8.27549994e-01 -1.07874259e-01 3.15096647e-01 1.11365676e+00
1.38926908e-01 4.61828649e-01 4.05472577e-01 -8.98044482e-02
-4.73961055e-01 -6.18751049e-01 3.45560253e-01 1.57185757e+00
4.76405412e-01 -6.18609905e-01 -7.09717512e-01 -7.30389655e-01
-2.78601080e-01 8.70519102e-01 -2.28031769e-01 -5.94041407e-01
-8.81007910e-01 -2.11120382e-01 5.36848128e-01 2.28230953e-01
4.64632630e-01 -1.26848710e+00 -3.21006119e-01 2.24959359e-01
-8.13040376e-01 -1.04605174e+00 -1.17213905e+00 -2.53209949e-01
-4.85695004e-01 -8.69047165e-01 -5.79308748e-01 -9.63286221e-01
4.57687587e-01 7.21957386e-01 1.39997923e+00 5.52862525e-01
2.11485580e-01 5.89468598e-01 -5.64898014e-01 -1.78686827e-01
-6.69890583e-01 6.16676450e-01 -4.47436780e-01 -2.05930769e-01
4.91208345e-01 -3.34557116e-01 -8.68225574e-01 5.51897764e-01
-9.43431139e-01 -1.42626306e-02 6.68844581e-01 8.99291873e-01
3.80963594e-01 -2.97000140e-01 8.50820422e-01 -7.02523589e-01
1.51924670e+00 -3.35689723e-01 -3.10999423e-01 8.69573593e-01
-6.89485192e-01 4.66239274e-01 4.62202877e-01 -4.47562695e-01
-1.17873585e+00 -3.02524120e-01 -3.46991062e-01 -1.40356794e-01
1.66037545e-01 6.70999348e-01 6.11130483e-02 3.99509817e-01
6.13426447e-01 6.60569489e-01 -1.41574353e-01 -4.54561144e-01
4.99407321e-01 8.11120987e-01 3.76927823e-01 -5.96209526e-01
3.60595375e-01 -3.17794643e-02 -6.45991087e-01 -5.87673604e-01
-7.84208655e-01 -7.02348411e-01 -4.07783955e-04 -4.60020266e-02
7.60527849e-01 -7.01383054e-01 -8.47392201e-01 -9.93930772e-02
-1.36159182e+00 -2.97055304e-01 -2.21872032e-02 2.16055155e-01
-4.77515310e-01 3.31865817e-01 -3.59418720e-01 -7.07123935e-01
-5.61329663e-01 -1.45434248e+00 1.39659250e+00 4.09737289e-01
-3.54910821e-01 -8.79663050e-01 1.98642612e-01 6.32753015e-01
5.42730749e-01 -7.13608265e-01 1.02207136e+00 -8.11780870e-01
-9.04891670e-01 -2.60945976e-01 -1.59479782e-01 6.85755461e-02
2.04214349e-01 -2.24736392e-01 -8.14185083e-01 -1.16114371e-01
-2.63443161e-02 -4.09843862e-01 7.13520527e-01 -2.87759863e-02
1.07112169e+00 -5.90162873e-01 -4.59445447e-01 2.43506968e-01
8.72274578e-01 5.87452710e-01 5.88633657e-01 -1.37228265e-01
3.10652733e-01 4.82557088e-01 5.59774041e-01 9.29563418e-02
8.07559133e-01 1.02034307e+00 1.05115257e-01 1.78960368e-01
-1.88560143e-01 -8.32192183e-01 3.16002697e-01 8.96255970e-01
4.17852104e-01 -6.17569208e-01 -6.48495913e-01 3.07437807e-01
-1.85543060e+00 -9.97395158e-01 4.02677804e-01 1.76891565e+00
1.05554760e+00 -1.31816149e-01 -2.28232846e-01 -4.92728323e-01
4.41982329e-01 2.30737135e-01 -6.83767200e-01 -4.62275267e-01
1.15800709e-01 1.82330653e-01 -2.53081888e-01 1.01676548e+00
-5.22778094e-01 1.10655653e+00 6.54122543e+00 6.73772812e-01
-1.22856605e+00 -1.03247777e-01 3.42614055e-01 1.73504412e-01
-8.61910880e-01 2.84213662e-01 -9.65892971e-01 2.71363884e-01
5.47934830e-01 -5.59638202e-01 7.96747863e-01 8.76287937e-01
1.34956256e-01 -5.77434637e-02 -1.19345224e+00 1.04080832e+00
3.38807821e-01 -1.30255783e+00 3.86062801e-01 -8.62125233e-02
4.09494758e-01 -2.42999017e-01 -1.88184768e-01 1.07062554e+00
2.85114259e-01 -1.03118801e+00 4.72105965e-02 7.67539442e-01
3.30381423e-01 -6.10564232e-01 6.55116379e-01 7.62197196e-01
-1.05377662e+00 -9.94455814e-02 -1.69157907e-01 1.36077568e-01
2.64304310e-01 -2.03552656e-02 -1.38566804e+00 3.74651700e-01
3.16801161e-01 2.83617765e-01 -6.95237100e-01 6.31836772e-01
-3.33859712e-01 3.53244007e-01 -1.91367105e-01 -4.66471642e-01
2.36037254e-01 -1.05055876e-01 4.15042579e-01 1.17895436e+00
1.73785627e-01 2.55685687e-01 3.56133252e-01 9.31330740e-01
-3.56923908e-01 2.00619057e-01 -6.07439280e-01 -3.48279834e-01
7.01527894e-01 9.75966513e-01 -2.00886294e-01 -3.55194241e-01
-3.25293005e-01 1.01227498e+00 2.78095722e-01 7.51989484e-01
-5.42100132e-01 -4.66863841e-01 4.81961101e-01 -8.36975127e-02
1.01093799e-01 -8.29362571e-02 3.19822192e-01 -1.36620331e+00
3.22048217e-01 -1.30899596e+00 3.19908291e-01 -9.56787229e-01
-1.00308907e+00 7.14395165e-01 3.91027927e-02 -8.35893810e-01
-1.15104294e+00 1.66432351e-01 -5.69019318e-01 1.17720997e+00
-1.45150089e+00 -9.70190585e-01 -2.70011663e-01 4.13630158e-01
8.84018242e-01 -1.08773783e-01 9.96604621e-01 2.12714732e-01
-2.35816687e-01 7.69683480e-01 -4.53168660e-01 1.22710750e-01
7.54255593e-01 -9.28504407e-01 3.70320469e-01 4.87427115e-01
2.57964343e-01 1.04917979e+00 5.64681649e-01 -4.61259961e-01
-1.69839489e+00 -5.85720420e-01 1.29742217e+00 -4.65698719e-01
3.56158465e-01 -2.79685080e-01 -1.21528518e+00 5.67943811e-01
4.49462682e-01 -7.70535409e-01 8.23977113e-01 2.83846885e-01
-1.86311871e-01 -2.08262160e-01 -4.89321917e-01 8.70758414e-01
9.83406842e-01 -9.97222483e-01 -8.47066760e-01 4.24962908e-01
1.20929396e+00 -4.46341813e-01 -5.89401901e-01 4.45878476e-01
4.88272190e-01 -7.56436765e-01 1.00121188e+00 -5.06292343e-01
1.35702118e-01 -8.37716088e-02 -1.80315405e-01 -1.20565808e+00
3.81287783e-02 -7.88038433e-01 -2.30417982e-01 1.09549952e+00
6.07111394e-01 -4.58957106e-01 7.04685271e-01 7.61532426e-01
-1.49985045e-01 -1.08527362e+00 -6.71762824e-01 -3.50376189e-01
-1.77636057e-01 -3.56518596e-01 1.02812159e+00 5.60201168e-01
1.10698447e-01 1.07253432e+00 -2.31404379e-01 3.96775547e-03
-2.23319516e-01 6.97631419e-01 1.19307709e+00 -7.93723762e-01
-6.37107372e-01 -6.75376713e-01 3.32002461e-01 -2.15294766e+00
3.71849209e-01 -9.91501510e-01 2.40542859e-01 -1.84848154e+00
1.94570377e-01 -4.09716994e-01 2.10719317e-01 2.18743980e-01
-4.84701753e-01 -4.76918876e-01 7.50056729e-02 3.85514706e-01
-7.26670086e-01 7.75400102e-01 1.82745874e+00 -2.05229193e-01
-6.10594034e-01 4.06809241e-01 -1.06844294e+00 1.54237404e-01
6.44781530e-01 -2.02678815e-01 -1.06357133e+00 -6.82824671e-01
5.64342678e-01 6.52851582e-01 7.85520002e-02 -3.70035529e-01
5.92029750e-01 -3.94730270e-01 -2.78555483e-01 -6.63081706e-01
5.98005831e-01 -3.96239579e-01 -1.85779169e-01 1.61298916e-01
-9.28142786e-01 2.96211839e-01 -3.03198308e-01 6.03706241e-01
-6.79436207e-01 -4.06227499e-01 8.58212188e-02 -2.52508074e-01
-5.96319497e-01 2.87036419e-01 -3.48930389e-01 2.42473662e-01
4.30347264e-01 9.57320929e-02 -2.60062635e-01 -1.21442449e+00
-6.69294178e-01 8.43224883e-01 1.68925002e-01 9.67785478e-01
7.78869748e-01 -1.19680834e+00 -5.28272510e-01 6.06272332e-02
3.12331796e-01 1.34121493e-01 -6.25439640e-03 1.68007001e-01
-1.53857678e-01 1.05538476e+00 5.33831954e-01 -3.90071809e-01
-1.14851499e+00 4.91529793e-01 4.93397117e-01 -5.77723801e-01
-1.26982212e-01 9.02452052e-01 2.35167399e-01 -5.80648124e-01
5.49854457e-01 -3.89906883e-01 -3.91441524e-01 -1.10631853e-01
5.30673027e-01 -1.55611202e-01 -1.41253397e-01 -3.41950029e-01
-2.40327441e-03 2.91467160e-01 -3.38209361e-01 -3.76855999e-01
7.92298675e-01 -4.34974432e-01 -1.57679245e-01 2.01400101e-01
1.34901583e+00 4.08584736e-02 -6.25546515e-01 -6.62290454e-01
6.06430182e-03 -3.89066279e-01 -2.57110238e-01 -8.13843966e-01
-6.97980762e-01 6.97563291e-01 4.83572148e-02 4.29677010e-01
1.16168427e+00 3.68646294e-01 1.18768060e+00 1.26731122e+00
2.92392254e-01 -9.71419990e-01 4.98027831e-01 8.06171358e-01
1.28788209e+00 -1.26537311e+00 -3.14118534e-01 -3.33290368e-01
-7.17872322e-01 1.00419748e+00 8.11579764e-01 3.12093616e-01
4.78266507e-01 -4.15521204e-01 3.20952177e-01 -3.00807595e-01
-1.13448620e+00 -2.50188261e-01 5.28788745e-01 1.91954747e-01
4.84433591e-01 -1.05847329e-01 -3.06816459e-01 4.68542427e-01
-6.21359706e-01 -6.35145558e-03 1.48180023e-01 8.79844129e-01
-4.45372641e-01 -1.45398593e+00 7.16398135e-02 4.32419986e-01
-2.05708563e-01 -1.49161607e-01 -8.31792414e-01 2.69823641e-01
-3.49214345e-01 1.31792569e+00 -1.98140349e-02 -5.35442710e-01
5.15012324e-01 2.55091161e-01 1.16955228e-01 -9.29543972e-01
-6.35694504e-01 6.07713982e-02 3.45084906e-01 -5.22876561e-01
-2.54225701e-01 -2.25988671e-01 -1.21069324e+00 -5.86834401e-02
-5.18862426e-01 8.56260538e-01 4.24405754e-01 9.96510983e-01
7.89545119e-01 3.40794176e-01 8.43253434e-01 -1.14644714e-01
-7.03160465e-01 -1.11358809e+00 1.34840384e-01 3.53145242e-01
3.44665527e-01 -3.74803513e-01 -4.13468368e-02 -1.57483086e-01] | [12.023097038269043, 7.711638450622559] |
9e2b369d-bf37-4b5f-92be-8ae4cb1593fa | towards-spectral-estimation-from-a-single-rgb | 1812.00805 | null | http://arxiv.org/abs/1812.00805v1 | http://arxiv.org/pdf/1812.00805v1.pdf | Towards Spectral Estimation from a Single RGB Image in the Wild | In contrast to the current literature, we address the problem of estimating
the spectrum from a single common trichromatic RGB image obtained under
unconstrained settings (e.g. unknown camera parameters, unknown scene radiance,
unknown scene contents). For this we use a reference spectrum as provided by a
hyperspectral image camera, and propose efficient deep learning solutions for
sensitivity function estimation and spectral reconstruction from a single RGB
image. We further expand the concept of spectral reconstruction such that to
work for RGB images taken in the wild and propose a solution based on a
convolutional network conditioned on the estimated sensitivity function.
Besides the proposed solutions, we study also generic and sensitivity
specialized models and discuss their limitations. We achieve state-of-the-art
competitive results on the standard example-based spectral reconstruction
benchmarks: ICVL, CAVE, NUS and NTIRE. Moreover, our experiments show that, for
the first time, accurate spectral estimation from a single RGB image in the
wild is within our reach. | ['Radu Timofte', 'Yigit Baran Can', 'Berk Kaya'] | 2018-12-03 | null | null | null | null | ['spectral-reconstruction', 'spectral-estimation-from-a-single-rgb-image'] | ['computer-vision', 'computer-vision'] | [ 9.03774559e-01 -4.09740984e-01 3.18791211e-01 -1.23583257e-01
-9.59485590e-01 -9.37785566e-01 2.06780612e-01 -2.65687227e-01
-7.08051860e-01 6.83837831e-01 -2.63026655e-01 -1.17826499e-01
-2.34357730e-01 -7.04027057e-01 -8.97660315e-01 -8.25314760e-01
4.74777877e-01 1.28333732e-01 -4.94771935e-02 -2.24516049e-01
-1.10892840e-01 7.70630836e-01 -1.64598846e+00 -1.11013660e-02
7.31281042e-01 1.24523354e+00 3.94144475e-01 9.14133251e-01
1.59449428e-01 7.20027983e-01 -2.10801810e-01 -7.92468712e-02
8.31771672e-01 -3.01146954e-01 -8.48617077e-01 6.22207522e-01
5.71906745e-01 -5.38750648e-01 -2.63612211e-01 1.28603375e+00
5.17605364e-01 2.19270945e-01 5.07161200e-01 -9.97057438e-01
-3.54539216e-01 1.81615382e-01 -5.92481673e-01 -3.78468961e-01
3.12706470e-01 3.37629437e-01 8.47134948e-01 -4.09472704e-01
4.78461057e-01 4.29679006e-01 9.37181890e-01 1.76997289e-01
-1.63364756e+00 -3.23902756e-01 2.24482007e-02 1.44323185e-01
-1.64823604e+00 -4.52860773e-01 7.96887934e-01 -1.39032423e-01
8.79645884e-01 2.05798969e-01 7.96525776e-01 1.17816508e+00
-5.91219366e-01 4.42255288e-01 1.47646725e+00 -4.93210137e-01
3.23373377e-01 -7.15598091e-02 -4.83096778e-01 5.24807036e-01
5.94256353e-03 5.89688197e-02 -3.35397750e-01 2.11412862e-01
7.43766725e-01 -2.29996055e-01 -8.85736942e-01 -4.74314332e-01
-1.18993473e+00 5.73298812e-01 8.11727703e-01 -1.07187949e-01
-4.04509902e-01 2.04472050e-01 3.22939865e-02 2.38648161e-01
3.78371656e-01 4.48751807e-01 -7.75906086e-01 1.72014982e-01
-9.99977708e-01 -3.61488834e-02 5.35316408e-01 8.11396360e-01
1.15084445e+00 -2.45256126e-02 3.77400935e-01 6.88739240e-01
1.71094522e-01 8.55724275e-01 1.44328801e-02 -1.23418260e+00
2.42602408e-01 2.05245912e-01 6.20025218e-01 -2.16230035e-01
-8.34500194e-01 -6.09943092e-01 -6.05235994e-01 2.20418572e-01
6.02492452e-01 -1.05829537e-01 -8.78390551e-01 1.57812035e+00
1.66854426e-01 4.59939063e-01 3.55317205e-01 1.07229376e+00
6.40105724e-01 6.21160924e-01 -2.90184289e-01 -4.26367104e-01
9.17856038e-01 -7.17473388e-01 -2.00906754e-01 -3.14601153e-01
3.73486876e-01 -7.06129432e-01 1.31548715e+00 8.21326613e-01
-7.11746037e-01 -1.74598262e-01 -1.06871927e+00 -2.22992837e-01
-5.99590003e-01 5.25991321e-01 7.19063222e-01 9.05597925e-01
-1.08460796e+00 7.01029718e-01 -6.39201045e-01 -4.81832594e-01
2.14461014e-01 1.05558395e-01 -3.42445761e-01 -1.35406420e-01
-9.64140773e-01 8.28064263e-01 4.73441482e-01 3.84082079e-01
-7.42604077e-01 -9.21734810e-01 -6.45769298e-01 -1.02560975e-01
6.34232998e-01 -4.70751613e-01 1.41973758e+00 -1.25176203e+00
-1.85610414e+00 9.88688409e-01 2.83729196e-01 -1.41314656e-01
6.43442512e-01 4.42168266e-02 -2.66384393e-01 3.02410156e-01
-3.87289822e-01 3.72247934e-01 6.18591309e-01 -1.57528841e+00
-1.38013825e-01 -1.97492912e-01 3.55134159e-01 3.31060410e-01
-8.53377730e-02 -2.20198527e-01 -3.92488718e-01 2.17878483e-02
2.33354941e-01 -9.70635772e-01 -1.54184520e-01 3.08768809e-01
-5.97592056e-01 7.63785243e-01 2.71193981e-01 -6.45999312e-01
3.71401101e-01 -1.91437793e+00 3.26050550e-01 1.34873942e-01
-2.36067206e-01 1.27113670e-01 -1.65238246e-01 2.76615441e-01
-3.22376251e-01 -1.32434934e-01 -8.70707154e-01 -4.01267856e-01
3.50421406e-02 -1.33748082e-02 -1.03022546e-01 8.65667403e-01
-9.60916206e-02 6.22192860e-01 -8.15007687e-01 3.25542837e-02
5.55002272e-01 7.58292079e-01 -2.30976343e-01 2.08348826e-01
-4.79867816e-01 5.07214308e-01 -7.07803201e-03 7.51605213e-01
1.16073382e+00 -2.90095985e-01 3.38677704e-01 -7.14963317e-01
-4.18115258e-01 -7.42487088e-02 -1.45990801e+00 2.26919842e+00
-1.04279292e+00 5.19000709e-01 4.54362243e-01 -7.17647016e-01
4.61969137e-01 -1.14655858e-02 4.67362940e-01 -7.19869971e-01
4.64025065e-02 4.21406507e-01 -3.34839970e-01 -4.47728604e-01
5.34860909e-01 -3.44193608e-01 3.97282034e-01 3.42271060e-01
1.73437163e-01 -6.02173030e-01 -1.77530915e-01 -1.39808923e-01
7.89170086e-01 7.31906652e-01 4.15752500e-01 -1.09098211e-01
4.98333991e-01 -9.53853875e-02 -1.21682370e-02 6.52675390e-01
1.15988165e-01 9.72011149e-01 3.58558536e-01 -6.99411035e-02
-1.01871645e+00 -1.10086870e+00 -3.07717294e-01 8.30502570e-01
7.44670182e-02 -1.10620268e-01 -6.83202147e-01 -1.37159035e-01
-9.47518423e-02 6.28190339e-01 -5.93451381e-01 2.06725106e-01
3.29954661e-02 -1.19416928e+00 3.47918093e-01 2.56171435e-01
9.50144708e-01 -6.64440691e-01 -9.64857221e-01 1.47113241e-02
-2.49984860e-01 -1.54685152e+00 2.37479791e-01 4.85163003e-01
-5.15124321e-01 -1.28290451e+00 -5.99520922e-01 6.05603196e-02
1.01846151e-01 4.22627360e-01 1.09533703e+00 -2.49922991e-01
-5.93787849e-01 6.88859761e-01 -3.48147601e-01 -1.44286230e-01
-6.45913705e-02 -1.26843145e-02 -2.65346885e-01 2.88947523e-01
-2.20665932e-01 -6.31084740e-01 -7.15275407e-01 -2.14206055e-02
-1.24543655e+00 1.90676481e-01 4.95239884e-01 3.87601256e-01
7.27853656e-01 1.46804377e-01 8.79969597e-02 -7.83779681e-01
-1.68484205e-03 -3.11644882e-01 -1.19391179e+00 3.86247337e-01
-4.46384668e-01 1.25920130e-02 8.08711231e-01 -1.19173303e-01
-1.28929663e+00 7.76522100e-01 -1.38048172e-01 -3.59947085e-01
-4.39600587e-01 3.18712443e-01 -3.08640063e-01 -6.70761406e-01
9.96334791e-01 3.09667647e-01 -3.73848200e-01 -5.27355790e-01
4.99308676e-01 6.63648963e-01 7.85680711e-01 -4.31851864e-01
7.49131024e-01 8.07297170e-01 4.33346212e-01 -1.09525228e+00
-8.04614067e-01 -6.66670740e-01 -6.85986876e-01 -3.79272699e-01
7.09088147e-01 -1.23702252e+00 -9.76837099e-01 6.36645675e-01
-1.05457366e+00 -7.91178405e-01 -5.59542216e-02 3.43387455e-01
-8.21409702e-01 5.73652744e-01 -2.54953593e-01 -1.04682267e+00
-3.99673693e-02 -1.29330027e+00 1.48627305e+00 2.08875641e-01
6.58083141e-01 -8.97110522e-01 2.80731414e-02 1.75609782e-01
5.05985558e-01 4.70892578e-01 3.50297332e-01 3.43282431e-01
-8.46096456e-01 9.10769477e-02 -7.33558655e-01 3.25750351e-01
2.02799648e-01 1.09257892e-01 -1.68148601e+00 -1.81419440e-02
-4.61820401e-02 -6.43231273e-01 1.29920900e+00 4.13557053e-01
1.42462683e+00 2.16616005e-01 1.55013770e-01 1.35857809e+00
2.24186540e+00 -4.46549445e-01 7.10941076e-01 5.27343512e-01
8.20919812e-01 5.19115150e-01 2.57833600e-01 5.54702342e-01
1.10354543e-01 6.71894133e-01 1.28719985e+00 -3.30746084e-01
-9.03137177e-02 2.91387051e-01 8.42229798e-02 -2.42364988e-01
-4.31035012e-01 -4.65441108e-01 -5.69881439e-01 3.59514594e-01
-1.71850955e+00 -6.62507951e-01 -3.57480973e-01 2.44356060e+00
7.07961977e-01 -4.04490262e-01 5.65403439e-02 1.38513699e-01
3.71451080e-01 2.84262300e-01 -7.28833079e-01 2.05005944e-01
-5.94575226e-01 4.27187443e-01 1.28006649e+00 6.17201269e-01
-1.11587107e+00 8.03150594e-01 6.37452650e+00 3.88189435e-01
-1.31794643e+00 1.38527691e-01 3.76488388e-01 -2.95927018e-01
-3.29998374e-01 -8.04904327e-02 -1.33023769e-01 2.31056698e-02
9.33859587e-01 4.68711555e-01 1.33216214e+00 4.12023067e-01
2.59443056e-02 -4.94270861e-01 -7.80573547e-01 1.37184048e+00
-5.70538640e-02 -1.09371996e+00 -5.16951203e-01 4.32081819e-02
6.34025753e-01 4.94655609e-01 1.40691206e-01 -2.70481974e-01
1.81890696e-01 -9.35588896e-01 6.77516878e-01 8.10612977e-01
1.11642826e+00 -6.69458985e-01 4.22666848e-01 1.31291896e-01
-1.00359130e+00 -4.58594821e-02 -6.15689874e-01 1.45477489e-01
-8.79978612e-02 4.84805822e-01 -5.27962208e-01 1.06611264e+00
6.97919130e-01 6.97641015e-01 -6.59973383e-01 1.05013371e+00
-4.10744429e-01 4.25852805e-01 -5.62080860e-01 3.54695886e-01
2.20763341e-01 -5.86259425e-01 4.11558449e-02 1.11719394e+00
4.43144441e-01 8.64346772e-02 -1.77571982e-01 1.19503176e+00
-2.33725548e-01 6.90932795e-02 -4.33639407e-01 1.07780330e-01
-2.11996123e-01 1.71767187e+00 -8.07423472e-01 2.93926895e-01
-5.49262226e-01 1.25878942e+00 1.38461784e-01 6.97919250e-01
-8.11366260e-01 -1.25676781e-01 4.78504181e-01 -1.35080785e-01
2.75896221e-01 -1.24654800e-01 9.97523125e-03 -1.36635506e+00
2.85622142e-02 -5.48969209e-01 8.07757974e-02 -1.59398150e+00
-8.75541627e-01 3.09298694e-01 1.92253646e-02 -1.17652678e+00
1.00988194e-01 -1.25545967e+00 -1.61404595e-01 1.10623610e+00
-2.14430928e+00 -1.24515140e+00 -8.02311718e-01 8.09702218e-01
1.49144828e-01 3.33907157e-01 9.08725321e-01 -8.23822897e-03
-4.39163983e-01 -7.89542198e-02 3.21165174e-01 -3.84047657e-01
5.34699559e-01 -1.39365029e+00 2.30083466e-01 8.76193464e-01
1.28120941e-03 1.78870186e-01 7.39769757e-01 -5.79881631e-02
-1.82559657e+00 -1.01088548e+00 -1.17202878e-01 -1.09199814e-01
9.10115480e-01 -4.75044161e-01 -7.00870991e-01 5.66668153e-01
2.80677769e-02 3.68850082e-01 5.37812769e-01 -2.53544062e-01
-5.94491065e-01 -3.04424137e-01 -1.27950382e+00 3.04829180e-01
1.16433167e+00 -8.17236006e-01 2.72847533e-01 8.88497353e-01
6.69962525e-01 -4.92975950e-01 -7.23814189e-01 2.98690945e-01
6.22831404e-01 -1.40112865e+00 1.24770379e+00 -2.76012808e-01
2.43969485e-01 -5.52844286e-01 -6.71895146e-01 -1.40162778e+00
1.49116188e-01 -5.28476119e-01 3.56776893e-01 7.81279922e-01
4.19090688e-01 -5.15080571e-01 6.65870070e-01 4.55618113e-01
-1.43571913e-01 -5.74902967e-02 -7.69227803e-01 -7.88280547e-01
-2.29555428e-01 -7.74404645e-01 7.01188862e-01 7.74842560e-01
-7.28434682e-01 -6.47959635e-02 -4.99047875e-01 7.29309499e-01
9.27972615e-01 3.55527610e-01 4.73956525e-01 -1.28594637e+00
-6.17771566e-01 -3.00701857e-01 -3.68306264e-02 -6.42694056e-01
4.30756241e-01 -6.70365453e-01 1.87083036e-01 -1.50789320e+00
2.37037301e-01 -2.18319014e-01 -1.91077098e-01 2.94328570e-01
2.01641023e-01 5.20631552e-01 1.14834964e-01 -1.05369702e-01
-4.84705508e-01 3.91829729e-01 1.00678968e+00 -2.18651474e-01
-4.34787385e-02 -1.67600751e-01 -4.87415552e-01 5.68012834e-01
7.72257924e-01 -1.72100276e-01 -4.31277931e-01 -4.46748435e-01
8.76428485e-01 1.46796420e-01 7.71012783e-01 -1.01177633e+00
1.93526208e-01 -2.99511611e-01 4.28579062e-01 -1.08439729e-01
7.77723014e-01 -1.10483181e+00 2.75135964e-01 1.77410189e-02
-6.32722005e-02 -6.95125103e-01 2.84785539e-01 4.45339292e-01
3.91853869e-01 -4.47663903e-01 8.41981828e-01 -3.89190525e-01
-8.97455037e-01 1.84605077e-01 1.58659462e-02 -3.17404479e-01
4.66244042e-01 -1.12729043e-01 -5.50770879e-01 -3.73999953e-01
-5.51522911e-01 -3.06301832e-01 8.92533541e-01 -3.63974661e-01
4.37991112e-01 -9.28302467e-01 -5.40775836e-01 -8.28168914e-02
3.97447169e-01 6.27758428e-02 3.95175934e-01 7.60372221e-01
-8.73308599e-01 4.18238156e-02 -1.84518620e-01 -6.08132422e-01
-8.50872993e-01 5.85568428e-01 1.04741251e+00 2.29434639e-01
-4.00861233e-01 6.69180572e-01 4.65068482e-02 -8.21365058e-01
-6.84043914e-02 -5.41318297e-01 2.04761758e-01 -1.73384160e-01
2.54982591e-01 1.62714362e-01 6.50120437e-01 -7.69847333e-01
-2.74514735e-01 7.04621553e-01 8.06221426e-01 -1.74315467e-01
1.44703078e+00 -1.28689781e-01 -2.12966934e-01 6.17808282e-01
1.20412469e+00 3.76206338e-02 -1.66595423e+00 -2.12873191e-01
-5.02512157e-01 -4.59730864e-01 6.15745604e-01 -1.16387081e+00
-1.29201436e+00 7.68314600e-01 6.87752306e-01 1.08302869e-01
1.79829860e+00 -3.46122593e-01 1.39884144e-01 7.81432867e-01
2.59911865e-01 -1.00606930e+00 -2.68579632e-01 2.54402757e-01
8.07329118e-01 -1.43668628e+00 2.50626385e-01 -4.98241276e-01
-3.18203032e-01 1.41812229e+00 1.06623828e-01 2.07813770e-01
2.74196684e-01 1.27109066e-01 -1.13224117e-02 1.24388494e-01
-1.95022464e-01 -8.78007472e-01 1.08732887e-01 8.38228822e-01
3.32563430e-01 -3.15717496e-02 4.95946348e-01 -6.38881139e-03
1.32217392e-01 -3.46659459e-02 1.04226375e+00 5.08009732e-01
-3.26206803e-01 -7.46349394e-01 -5.35492361e-01 9.53777879e-02
-1.43130809e-01 -3.62789869e-01 -5.15437901e-01 9.46111500e-01
7.66290128e-02 9.85162914e-01 -1.94000036e-01 -1.68542415e-01
4.01867628e-01 1.07190445e-01 9.53834534e-01 -3.27076226e-01
-3.63088816e-01 2.48153418e-01 7.44383857e-02 -8.20202947e-01
-9.86640990e-01 -7.14209974e-01 -9.26921308e-01 -2.92532351e-02
-3.53130341e-01 -4.28897232e-01 1.16517019e+00 8.22380006e-01
-2.85948306e-01 5.57370186e-01 4.57292348e-01 -1.19803810e+00
-3.90441626e-01 -8.28538418e-01 -1.08542991e+00 2.65255664e-02
6.89868510e-01 -6.01453364e-01 -6.15374982e-01 -9.76903066e-02] | [10.230462074279785, -2.4066481590270996] |
5b5dc873-edec-447a-9310-0a3cd470cf72 | what-have-been-learned-what-should-be-learned | 2109.00175 | null | https://arxiv.org/abs/2109.00175v1 | https://arxiv.org/pdf/2109.00175v1.pdf | What Have Been Learned & What Should Be Learned? An Empirical Study of How to Selectively Augment Text for Classification | Text augmentation techniques are widely used in text classification problems to improve the performance of classifiers, especially in low-resource scenarios. Whilst lots of creative text augmentation methods have been designed, they augment the text in a non-selective manner, which means the less important or noisy words have the same chances to be augmented as the informative words, and thereby limits the performance of augmentation. In this work, we systematically summarize three kinds of role keywords, which have different functions for text classification, and design effective methods to extract them from the text. Based on these extracted role keywords, we propose STA (Selective Text Augmentation) to selectively augment the text, where the informative, class-indicating words are emphasized but the irrelevant or noisy words are diminished. Extensive experiments on four English and Chinese text classification benchmark datasets demonstrate that STA can substantially outperform the non-selective text augmentation methods. | ['Hailiang Huang', 'Sonqiao Han', 'Biyang Guo'] | 2021-09-01 | null | null | null | null | ['text-augmentation'] | ['natural-language-processing'] | [ 5.83247721e-01 -2.84593552e-02 -4.26795602e-01 -2.36239851e-01
-2.93126792e-01 -2.09890768e-01 6.17579699e-01 3.53755593e-01
-6.70735240e-01 8.90954137e-01 5.89873374e-01 -3.18521798e-01
2.61320740e-01 -6.51634753e-01 -5.23935445e-03 -9.30448174e-01
4.22647387e-01 3.36598784e-01 1.10107176e-01 -4.98479724e-01
3.44758868e-01 4.98071387e-02 -1.73888421e+00 3.91710937e-01
1.11255622e+00 7.83676684e-01 6.06772862e-02 1.65837124e-01
-8.92911613e-01 5.45994639e-01 -1.01182759e+00 -3.02797228e-01
-3.40577327e-02 -4.73827213e-01 -7.14465499e-01 1.84122950e-01
4.64954637e-02 -2.33235911e-01 -3.61158460e-01 9.97541130e-01
4.15034294e-01 2.56490707e-01 5.78813910e-01 -1.10536301e+00
-5.34762144e-01 1.17604804e+00 -6.58568144e-01 1.35697648e-01
1.71581119e-01 -3.72461230e-01 1.05190825e+00 -1.27701235e+00
1.58443987e-01 1.27392328e+00 5.62781811e-01 4.44115698e-01
-8.36882293e-01 -9.07600343e-01 6.97654486e-01 -1.04599476e-01
-1.13751340e+00 -1.52953714e-01 9.69475806e-01 -2.28085183e-02
6.46048188e-01 4.77480680e-01 6.14872038e-01 9.91632521e-01
-1.55118406e-01 1.24369931e+00 8.85972023e-01 -7.98228979e-01
-1.73057377e-01 3.33932430e-01 4.25677270e-01 3.55993837e-01
3.32268953e-01 -4.99339640e-01 -5.90083241e-01 -2.81139344e-01
2.16706380e-01 3.10407817e-01 -4.21675414e-01 1.66665971e-01
-1.31908345e+00 8.20691407e-01 1.28121898e-01 4.99174446e-01
-2.07711503e-01 -2.04154134e-01 7.31866002e-01 9.23789963e-02
9.41914797e-01 4.23477054e-01 -7.71055520e-01 -3.42261903e-02
-4.74483341e-01 1.09392121e-01 3.20394427e-01 1.06019151e+00
6.24127924e-01 9.54184979e-02 -7.13788450e-01 1.10447490e+00
1.30979002e-01 2.20431313e-01 9.67738688e-01 4.96958047e-02
8.16692352e-01 1.28620017e+00 -5.04266694e-02 -6.89297438e-01
-3.01085442e-01 -5.85745573e-01 -1.13268793e+00 -3.78311187e-01
-6.53003976e-02 -1.90153867e-01 -1.17479253e+00 1.46387899e+00
2.81908154e-01 -2.64704287e-01 -6.92289025e-02 3.13475102e-01
9.22031939e-01 7.36540854e-01 3.17712963e-01 -4.47246581e-01
1.69467306e+00 -1.01441669e+00 -1.33623981e+00 -4.46344942e-01
9.99297917e-01 -9.00196731e-01 1.61502624e+00 3.31104815e-01
-6.47308767e-01 -4.54338282e-01 -1.01607585e+00 9.41793248e-03
-8.49181950e-01 4.26791549e-01 6.89289451e-01 9.09801662e-01
-2.80415714e-01 2.11938232e-01 -3.09587091e-01 1.85293555e-01
5.63091755e-01 2.15215430e-01 -4.69282083e-02 5.40612452e-02
-1.53267181e+00 5.84868848e-01 8.15125406e-01 1.55731710e-03
-7.65422508e-02 -4.82983857e-01 -9.21293139e-01 1.52097344e-01
5.26284218e-01 -1.65141225e-01 1.12093198e+00 -9.04525220e-01
-1.13928139e+00 3.92170757e-01 -3.68179679e-01 -9.55302045e-02
4.48281974e-01 -3.72550696e-01 -5.17493248e-01 -9.71958414e-02
7.91100636e-02 5.96903384e-01 1.01904559e+00 -1.17617476e+00
-8.63816619e-01 -3.11401755e-01 -3.24337214e-01 4.70921338e-01
-1.44842970e+00 1.16066393e-02 -6.46662533e-01 -1.51470554e+00
3.09843749e-01 -7.40134239e-01 -3.16109896e-01 -2.64354467e-01
-6.90538645e-01 -4.23884749e-01 1.50293815e+00 -3.80588412e-01
1.70349228e+00 -2.15275335e+00 -2.61194378e-01 4.43402864e-02
2.90132254e-01 5.01984060e-01 -3.61142196e-02 3.17288220e-01
-1.46394476e-01 5.52632749e-01 -3.89496952e-01 -5.21516204e-01
-3.19782138e-01 2.79639065e-01 -5.93491912e-01 -6.61232769e-02
2.47915581e-01 8.63277555e-01 -8.04019630e-01 -6.66631818e-01
2.07833901e-01 3.43392402e-01 -2.15461493e-01 -1.58292174e-01
-3.07677329e-01 2.20861048e-01 -9.12874877e-01 5.08100450e-01
5.95864952e-01 -1.19275630e-01 -4.41094562e-02 -1.82667971e-02
6.95309937e-02 5.95257521e-01 -1.05918384e+00 9.85392153e-01
-2.78026164e-01 5.96877515e-01 -3.25854868e-01 -8.27524185e-01
1.05421126e+00 2.70011097e-01 3.44612777e-01 -3.99843872e-01
4.29918945e-01 -1.57049283e-01 -8.54600370e-02 -3.04727465e-01
1.04725683e+00 -1.47643154e-02 -2.66547557e-02 6.44999087e-01
-3.72683227e-01 1.57556757e-02 4.11727011e-01 3.13948452e-01
6.30836427e-01 -2.06298187e-01 3.92741561e-01 -1.54413924e-01
7.35106826e-01 1.73352789e-02 5.11384666e-01 7.15747297e-01
-2.14598384e-02 3.39630514e-01 3.99474114e-01 -2.82101810e-01
-9.42609310e-01 -1.71156436e-01 -2.38811225e-01 1.47314203e+00
8.53300840e-02 -7.37733185e-01 -5.81361711e-01 -1.09364057e+00
-1.21816888e-01 8.14105093e-01 -9.44791913e-01 -4.85255688e-01
-5.93090534e-01 -1.10737395e+00 5.31853676e-01 7.64181495e-01
5.76143384e-01 -1.12500393e+00 -5.44526428e-02 2.33877450e-01
-3.05900961e-01 -9.58825588e-01 -7.17022657e-01 3.97768676e-01
-9.81814504e-01 -7.42349088e-01 -7.19761848e-01 -9.63962615e-01
1.12872577e+00 7.45801091e-01 7.83398926e-01 5.72779596e-01
9.27136466e-02 -2.31643632e-01 -9.98800993e-01 -8.21406782e-01
-4.12285596e-01 2.45667294e-01 -7.28317872e-02 -9.60811879e-03
4.75997210e-01 -2.13153195e-02 -1.75166234e-01 4.20945495e-01
-1.16929460e+00 2.77083397e-01 6.08243465e-01 1.35285056e+00
5.20526528e-01 4.05204415e-01 7.69129753e-01 -1.25438535e+00
1.04129553e+00 -8.18808749e-02 -2.07565531e-01 2.29567200e-01
-8.33118081e-01 5.23607098e-02 9.60269809e-01 -9.04106200e-01
-1.09837008e+00 -4.61257920e-02 -2.55884647e-01 -3.36609036e-02
1.28107622e-01 6.92579150e-01 -3.97623718e-01 5.45267276e-02
3.97087663e-01 5.88338852e-01 -3.00721049e-01 -4.84127432e-01
2.32144177e-01 8.83742869e-01 -9.22818333e-02 -4.25389916e-01
7.51957178e-01 2.84214646e-01 -1.70511216e-01 -9.02291536e-01
-1.05547822e+00 -6.04112983e-01 -7.17077374e-01 9.07241553e-03
3.14094216e-01 -5.45145750e-01 -2.63316303e-01 4.43813443e-01
-8.00751805e-01 7.67436996e-02 -5.03718019e-01 3.39866161e-01
2.01834083e-01 6.49759531e-01 -3.47782642e-01 -1.06290948e+00
-6.45588636e-01 -1.08297908e+00 1.05232525e+00 3.37739497e-01
-2.37520188e-01 -7.83870518e-01 -5.67831397e-01 2.59734154e-01
2.06033036e-01 -2.69709915e-01 1.30502200e+00 -1.21959949e+00
1.33354843e-01 -7.22896457e-01 -1.04667179e-01 4.08905119e-01
4.94815320e-01 -7.63999671e-02 -9.19004858e-01 -1.33822471e-01
-1.11145683e-01 -1.45123824e-01 9.89807367e-01 1.52245998e-01
1.80485439e+00 -4.93548542e-01 -4.99446571e-01 2.58080900e-01
5.84730208e-01 4.60376203e-01 6.92314386e-01 3.30030352e-01
8.33699644e-01 5.30935705e-01 1.00451672e+00 5.01882672e-01
-5.20482697e-02 4.12797421e-01 1.97735071e-01 -3.04312825e-01
1.27090678e-01 -2.97753751e-01 7.92355463e-02 9.44763184e-01
2.70757496e-01 -5.84133506e-01 -7.21403480e-01 3.00136089e-01
-1.62435079e+00 -5.56007683e-01 -5.26788235e-01 2.14687276e+00
1.21772206e+00 5.30930638e-01 -7.05321953e-02 8.07713866e-01
8.82295668e-01 1.92139164e-01 -3.68775725e-01 8.40045810e-02
-4.56717700e-01 1.86897591e-01 2.37551913e-01 8.47742260e-02
-1.21775651e+00 1.09233725e+00 6.45206594e+00 1.20073843e+00
-7.53119767e-01 -2.60902848e-02 8.13045084e-01 9.95796546e-02
-3.67743731e-01 -1.32287785e-01 -1.08140910e+00 5.13002574e-01
3.48913729e-01 -2.48485848e-01 -1.62208378e-01 9.62777555e-01
6.91709016e-03 -6.09614626e-02 -6.36565566e-01 7.62037039e-01
1.22493371e-01 -1.13732684e+00 7.39825845e-01 -8.27923715e-02
7.75869548e-01 -6.21718109e-01 1.16142519e-01 5.58215976e-01
4.72417846e-02 -7.06867754e-01 4.92294222e-01 -3.09053957e-02
6.51712656e-01 -9.74125326e-01 9.95646775e-01 4.00569588e-01
-1.23395264e+00 -1.71828732e-01 -2.92839497e-01 5.83534576e-02
-1.61192983e-01 6.77536786e-01 -7.76993930e-01 4.16125864e-01
5.73841333e-01 5.07297397e-01 -8.25177908e-01 6.27512336e-01
-4.65760916e-01 8.42986166e-01 -5.90225607e-02 -5.73012888e-01
2.87055701e-01 -1.49540186e-01 5.55004597e-01 1.27933371e+00
3.03305723e-02 2.16139868e-01 1.82367355e-01 4.02638614e-01
-5.22257328e-01 6.67237222e-01 -2.21972033e-01 -2.51687169e-01
6.28473759e-01 1.26767361e+00 -1.03905261e+00 -7.76774943e-01
-3.59559864e-01 7.92990685e-01 -2.13767439e-01 1.94466010e-01
-5.00216424e-01 -7.42253006e-01 3.19271684e-01 -9.30371135e-02
-4.67922771e-03 9.37853754e-03 -6.72561705e-01 -1.03270125e+00
1.30012825e-01 -9.70895648e-01 5.57998717e-01 -6.92935288e-01
-1.15281045e+00 6.37750089e-01 -5.93539514e-02 -1.41231787e+00
3.04705977e-01 -5.17368197e-01 -5.93768060e-01 9.28558946e-01
-1.27691889e+00 -1.03960586e+00 -4.93343025e-01 3.70393753e-01
1.03775787e+00 -3.48145902e-01 6.46956623e-01 2.10594192e-01
-8.54463100e-01 9.44050491e-01 2.49205083e-01 4.19250131e-01
5.76228738e-01 -1.12764764e+00 4.89176810e-01 8.78874302e-01
-1.64493248e-02 8.60911489e-01 5.77929616e-01 -9.69241560e-01
-1.06404626e+00 -1.16238105e+00 8.64227831e-01 -1.81850880e-01
4.73157227e-01 -5.36801219e-01 -1.25712192e+00 4.12721843e-01
-3.92548442e-02 -2.78071821e-01 7.85128355e-01 1.88233390e-01
-1.76912367e-01 1.44673616e-01 -9.10635889e-01 1.12775552e+00
7.91830599e-01 -2.03643396e-01 -7.06835985e-01 4.62229162e-01
1.07786071e+00 -1.24877952e-01 -3.00804943e-01 4.46771652e-01
2.56711632e-01 -2.33390898e-01 8.28599870e-01 -7.86301613e-01
4.15723860e-01 -2.48472139e-01 1.48941830e-01 -1.28776324e+00
-6.53369427e-02 -4.31173742e-01 -2.82101661e-01 1.61694479e+00
5.35886884e-01 -8.62687230e-01 9.59481120e-01 2.27492630e-01
-2.61169076e-01 -8.87194872e-01 -7.15202391e-01 -5.80949306e-01
7.42568728e-03 -2.75181144e-01 8.26556623e-01 1.32103074e+00
1.04754947e-01 6.11961603e-01 -4.65537220e-01 -4.67074811e-01
6.29625916e-02 -2.71053702e-01 5.21688700e-01 -1.37018418e+00
3.85427922e-01 -7.31165588e-01 4.67361212e-02 -1.18309069e+00
1.02811836e-01 -6.74640000e-01 1.29456773e-01 -1.29191101e+00
3.21606040e-01 -7.09378123e-01 -3.04440945e-01 9.54722404e-01
-9.64833677e-01 1.53865755e-01 -2.51820594e-01 2.30049536e-01
-1.80318892e-01 1.05016673e+00 1.43949330e+00 -3.60701978e-01
-4.45372522e-01 2.14600518e-01 -1.04130280e+00 7.28570759e-01
9.17817771e-01 -5.60210586e-01 -4.63979602e-01 -8.38766173e-02
2.09011227e-01 -5.02464890e-01 -3.50176960e-01 -3.24696213e-01
-3.83662898e-03 -7.00761899e-02 6.58881962e-01 -1.08934379e+00
2.22922578e-01 -8.23917568e-01 -6.73551261e-01 4.63319629e-01
-7.53339410e-01 1.83546081e-01 4.65723038e-01 4.15234953e-01
-2.00962126e-01 -5.37109971e-01 6.37773454e-01 9.10302401e-02
-3.72443348e-01 3.10500830e-01 -6.10325634e-01 -4.93205860e-02
8.18642497e-01 -4.08819199e-01 -3.34556669e-01 -3.24513435e-01
-3.07218879e-01 3.05178732e-01 -1.10464403e-02 5.52034676e-01
6.23832762e-01 -1.34861457e+00 -7.31574237e-01 2.99893856e-01
3.56185019e-01 9.08897743e-02 7.58103356e-02 5.13289690e-01
6.83039650e-02 4.55136329e-01 2.22044647e-01 -2.00128198e-01
-1.56692207e+00 5.51609576e-01 -7.95239210e-03 -3.74608219e-01
-7.85732031e-01 7.32208729e-01 3.52151245e-01 -2.83053249e-01
4.25439030e-01 -3.53973478e-01 -7.33519733e-01 4.06323731e-01
8.33484113e-01 1.45992100e-01 3.66951764e-01 -5.08177876e-01
-2.38692444e-02 1.24794714e-01 -7.25348771e-01 6.39802888e-02
1.05761456e+00 -1.43491074e-01 -2.08615646e-01 3.74530852e-01
7.77712107e-01 2.61583418e-01 -6.72423601e-01 -6.07176304e-01
1.68831304e-01 -5.40798187e-01 2.51273841e-01 -7.13341117e-01
-9.68369663e-01 8.07187021e-01 2.17487022e-01 4.92261976e-01
1.40380085e+00 -3.21681350e-01 6.87859833e-01 5.08580148e-01
-2.01811373e-01 -1.17812085e+00 2.97254056e-01 8.24725986e-01
7.80605137e-01 -1.02988768e+00 3.24697942e-01 -8.27330470e-01
-7.88463771e-01 9.92448330e-01 9.10550237e-01 6.59444571e-01
4.21013534e-01 3.60877812e-01 -9.26021934e-02 9.44160298e-02
-5.86971462e-01 -1.95950300e-01 3.47531676e-01 4.63242978e-01
6.84683681e-01 -2.23347545e-01 -5.48647225e-01 7.47712612e-01
-1.96669370e-01 -6.53686464e-01 4.96371061e-01 1.22154081e+00
-6.17635906e-01 -1.22204149e+00 -5.98177254e-01 1.00862551e+00
-4.93054956e-01 -5.03758132e-01 -7.62343228e-01 8.50676000e-01
3.42306606e-02 9.60469663e-01 -1.13944203e-01 -4.97634917e-01
2.74983138e-01 4.22800839e-01 -1.59354344e-01 -7.80121028e-01
-7.66489983e-01 2.02832684e-01 1.56827107e-01 3.42458397e-01
4.68466878e-02 -5.59273481e-01 -1.32099116e+00 -1.10802300e-01
-1.02530098e+00 4.78317171e-01 5.64577818e-01 1.06668901e+00
3.14810276e-02 8.65750194e-01 7.06600547e-01 -5.56933343e-01
-3.53481770e-01 -1.48149538e+00 -3.25399220e-01 3.57968360e-01
8.93546864e-02 -8.23017120e-01 -2.94885874e-01 -1.29500583e-01] | [10.562234878540039, 7.662924289703369] |
4fcb216e-f502-4fc4-90a8-7eac3358448f | sequence-length-is-a-domain-length-based | 2109.07276 | null | https://arxiv.org/abs/2109.07276v1 | https://arxiv.org/pdf/2109.07276v1.pdf | Sequence Length is a Domain: Length-based Overfitting in Transformer Models | Transformer-based sequence-to-sequence architectures, while achieving state-of-the-art results on a large number of NLP tasks, can still suffer from overfitting during training. In practice, this is usually countered either by applying regularization methods (e.g. dropout, L2-regularization) or by providing huge amounts of training data. Additionally, Transformer and other architectures are known to struggle when generating very long sequences. For example, in machine translation, the neural-based systems perform worse on very long sequences when compared to the preceding phrase-based translation approaches (Koehn and Knowles, 2017). We present results which suggest that the issue might also be in the mismatch between the length distributions of the training and validation data combined with the aforementioned tendency of the neural networks to overfit to the training data. We demonstrate on a simple string editing task and a machine translation task that the Transformer model performance drops significantly when facing sequences of length diverging from the length distribution in the training data. Additionally, we show that the observed drop in performance is due to the hypothesis length corresponding to the lengths seen by the model during training rather than the length of the input sequence. | ['Ondřej Bojar', 'Dušan Variš'] | 2021-09-15 | null | https://aclanthology.org/2021.emnlp-main.650 | https://aclanthology.org/2021.emnlp-main.650.pdf | emnlp-2021-11 | ['l2-regularization'] | ['methodology'] | [ 7.18930840e-01 2.23245740e-01 -4.13708352e-02 -2.52805054e-01
-9.96103108e-01 -7.81086802e-01 5.87573349e-01 3.96983288e-02
-5.04825234e-01 1.04108727e+00 2.61714160e-01 -7.07351863e-01
2.79862881e-01 -5.46377659e-01 -1.08352268e+00 -6.15503013e-01
3.68626952e-01 7.42738187e-01 -2.23196000e-01 -2.53585219e-01
2.81620651e-01 1.39531404e-01 -1.26724589e+00 2.85929680e-01
9.60493743e-01 5.73014796e-01 1.80396080e-01 7.36922264e-01
-2.58795768e-01 3.29850167e-01 -6.84409618e-01 -6.36539698e-01
2.46956378e-01 -6.53496325e-01 -5.74240923e-01 -1.88306794e-02
6.92934930e-01 -8.65671486e-02 -1.75483376e-01 1.09784937e+00
4.25000012e-01 -3.24581005e-02 6.18239939e-01 -8.56699049e-01
-6.46483183e-01 7.28394985e-01 -1.80331096e-01 7.20352083e-02
1.85856462e-01 3.44826430e-01 9.46773946e-01 -1.08891749e+00
6.90922141e-01 1.17844152e+00 8.00837040e-01 6.78797960e-01
-1.54074335e+00 -3.89011860e-01 -8.13238546e-02 -7.22382665e-02
-9.83441234e-01 -6.19421244e-01 4.75400805e-01 -4.41537917e-01
1.19836414e+00 1.78535119e-01 3.52209628e-01 1.50448620e+00
2.74872303e-01 6.68320060e-01 9.98261273e-01 -5.38474321e-01
-8.85729194e-02 2.10182190e-01 -1.41079783e-01 3.57174754e-01
2.63559539e-02 3.56530607e-01 -3.60352725e-01 -2.51163065e-01
6.18346393e-01 -4.58426267e-01 -3.53592843e-01 -5.88073954e-02
-1.16681087e+00 8.16389382e-01 4.15047295e-02 4.52194512e-01
-2.43076175e-01 1.34910762e-01 7.01157570e-01 6.24083519e-01
5.51100850e-01 7.30011106e-01 -8.37352216e-01 -4.27941293e-01
-1.09999466e+00 2.02927843e-01 8.62628102e-01 8.75018239e-01
4.22166586e-01 1.90249905e-01 -1.89413875e-01 7.80342758e-01
-1.76951990e-01 1.58101350e-01 7.41354167e-01 -6.01243317e-01
8.83483648e-01 2.40030095e-01 2.05270544e-01 -4.99139994e-01
-1.14576861e-01 -6.88449740e-01 -7.48338521e-01 -3.19754593e-02
1.03167343e+00 -2.37387016e-01 -9.08524871e-01 2.08680010e+00
-3.23600411e-01 -1.22389086e-01 1.12432808e-01 7.87540793e-01
1.47012636e-01 8.10107648e-01 3.74724045e-02 -3.89980614e-01
9.23335373e-01 -9.06397641e-01 -5.25185943e-01 -5.10319710e-01
9.69509959e-01 -8.42866480e-01 1.51893985e+00 3.44144583e-01
-1.21334279e+00 -5.39453208e-01 -9.57029998e-01 -9.18489397e-02
-1.91777945e-01 1.23751491e-01 3.13856363e-01 5.76594949e-01
-8.17830145e-01 1.02620220e+00 -7.20977426e-01 -4.85054433e-01
1.74766704e-01 2.32390106e-01 -2.12499917e-01 -1.83374749e-03
-1.22933364e+00 1.28247321e+00 5.69773316e-01 1.44438967e-01
-5.40325522e-01 -8.02796066e-01 -5.65465271e-01 2.75726438e-01
4.50192988e-02 -4.95602816e-01 1.34142387e+00 -1.40858674e+00
-1.48466301e+00 7.56683946e-01 -8.97880793e-02 -6.85792506e-01
8.49431396e-01 -3.74323875e-01 -7.18879849e-02 -4.28668380e-01
-2.91813314e-01 4.69027758e-01 6.85877502e-01 -8.61974657e-01
-1.87042192e-01 -3.01252425e-01 -4.05099511e-01 2.24211037e-01
-1.05802849e-01 -5.44839725e-02 2.08377410e-02 -5.65845191e-01
-2.03532115e-01 -1.00655627e+00 -2.45142534e-01 -3.86193842e-01
-2.99753040e-01 -1.96536243e-01 4.55345154e-01 -8.13925743e-01
1.14909995e+00 -2.10851455e+00 3.06417316e-01 7.76383132e-02
-4.12211388e-01 5.16430557e-01 -4.87148732e-01 7.18953609e-01
-3.54450524e-01 1.80769280e-01 -3.80316526e-01 -2.15212613e-01
5.78857725e-03 2.21904352e-01 -4.62139845e-01 4.06920701e-01
4.13736314e-01 8.62383068e-01 -9.72517788e-01 -1.54730037e-01
-2.07412675e-01 3.72481108e-01 -3.70426029e-01 1.79958835e-01
-6.09015346e-01 6.77192807e-01 5.14593497e-02 6.61650151e-02
3.32115591e-01 -8.87307301e-02 2.49497995e-01 2.41150662e-01
-4.65949178e-02 8.34412932e-01 -5.90969861e-01 1.66831398e+00
-5.79284847e-01 9.20278609e-01 -2.22303450e-01 -1.10415661e+00
8.09336662e-01 6.74190998e-01 7.63187334e-02 -6.91620529e-01
2.22557988e-02 6.83083296e-01 5.34014940e-01 -5.14756083e-01
5.31779945e-01 -5.36126018e-01 2.37466201e-01 5.95226109e-01
4.66676801e-02 -1.66983828e-01 2.74330378e-01 -2.12521344e-01
9.85803545e-01 4.20709014e-01 -7.85640925e-02 -1.34800017e-01
2.50113487e-01 8.83316547e-02 5.98843396e-01 7.70979702e-01
1.14141010e-01 6.24758661e-01 7.71758378e-01 -3.42786849e-01
-1.71670043e+00 -7.67846048e-01 -1.64130896e-01 1.08643925e+00
-6.42353356e-01 -3.43174577e-01 -9.58195329e-01 -6.25143826e-01
-1.81577623e-01 1.09201849e+00 -4.66439873e-01 -4.15538877e-01
-8.80005538e-01 -7.56951928e-01 9.21613753e-01 4.36841965e-01
-3.55773047e-02 -1.10724103e+00 -4.02197897e-01 6.09422863e-01
-2.82396644e-01 -9.74725962e-01 -5.42604148e-01 3.81246388e-01
-1.23834288e+00 -5.84876299e-01 -6.70993686e-01 -6.77198887e-01
6.63357556e-01 -3.04162890e-01 1.41665828e+00 1.22262523e-01
1.20629922e-01 -3.45727950e-01 -2.03536108e-01 -3.59609783e-01
-9.68270004e-01 4.66172248e-01 5.17928340e-02 -3.99981350e-01
4.30648118e-01 -7.10076630e-01 -3.64334464e-01 1.88654408e-01
-8.79947245e-01 1.27765864e-01 7.65333831e-01 1.15284133e+00
1.61576748e-01 -3.52624059e-01 7.18020678e-01 -9.96329308e-01
8.41775775e-01 -3.97726566e-01 -5.98039210e-01 1.78348750e-01
-7.79618263e-01 3.36039424e-01 9.30898190e-01 -6.80003464e-01
-6.97589159e-01 -2.36016095e-01 -3.60607982e-01 -2.95005947e-01
-1.88614488e-01 6.54755890e-01 6.54442087e-02 3.09499055e-01
8.59143257e-01 4.10270721e-01 1.00551635e-01 -4.72187221e-01
1.09860517e-01 6.45608485e-01 2.45481282e-01 -6.77580297e-01
6.34333551e-01 -2.66257703e-01 -8.44421089e-02 -6.48425221e-01
-8.68493259e-01 8.26408640e-02 -4.96336758e-01 1.51686937e-01
5.15352488e-01 -5.95468104e-01 -1.71020985e-01 3.96855205e-01
-1.42299592e+00 -6.32092178e-01 -2.30064139e-01 4.61669236e-01
-7.94529080e-01 2.37248555e-01 -8.79990816e-01 -7.59255290e-01
-3.45459282e-01 -1.20819569e+00 8.20306540e-01 -1.39457583e-01
-5.29514730e-01 -1.04479337e+00 7.29420409e-02 1.19988926e-01
6.52116716e-01 1.53608099e-01 1.52633297e+00 -1.02914023e+00
-3.42201799e-01 -3.39018077e-01 -4.02752608e-02 5.51023245e-01
-1.06523395e-01 3.52787897e-02 -7.47687519e-01 -3.91729891e-01
1.76568091e-01 -5.40551543e-01 6.39617324e-01 2.05425546e-01
8.53202581e-01 -4.36405927e-01 -9.26641896e-02 3.49867880e-01
1.35661542e+00 -8.86479318e-02 6.45650923e-01 2.98596054e-01
5.10156989e-01 7.40539789e-01 4.33317840e-01 -4.46972512e-02
-2.79108077e-01 8.10522079e-01 9.38285738e-02 5.30373380e-02
6.63353652e-02 -4.90823954e-01 5.18762946e-01 9.29692864e-01
1.89494997e-01 -3.01721156e-01 -8.39694142e-01 5.26194096e-01
-1.84218013e+00 -8.58362615e-01 -2.72757679e-01 2.58692408e+00
1.16405213e+00 4.68676925e-01 1.71335116e-01 9.90957469e-02
5.61095595e-01 -1.40711695e-01 -6.02157593e-01 -8.76498222e-01
-1.54827356e-01 2.26784274e-01 5.74934661e-01 5.29600978e-01
-4.77323771e-01 9.32209074e-01 6.53732872e+00 7.66230643e-01
-1.33161449e+00 8.67801830e-02 6.08449459e-01 -1.40445024e-01
-2.16487363e-01 1.40831005e-02 -5.66545188e-01 6.37648344e-01
1.51554656e+00 -1.36548594e-01 4.73608941e-01 4.75754827e-01
3.07053804e-01 1.31840184e-01 -1.54854965e+00 6.46802187e-01
-9.07463804e-02 -1.11519289e+00 -7.43090287e-02 2.02191815e-01
6.62883699e-01 2.40584224e-01 7.02632070e-02 4.50024694e-01
8.36028159e-02 -1.35345125e+00 7.09627926e-01 8.34684074e-02
7.89955258e-01 -5.76783836e-01 7.19152451e-01 8.36210907e-01
-4.18779820e-01 4.48495224e-02 -5.69609582e-01 -1.70682490e-01
1.69405445e-01 6.69974148e-01 -9.30677831e-01 2.40908295e-01
7.07854405e-02 3.31194431e-01 -3.46467584e-01 9.08925653e-01
-1.78367138e-01 9.48530376e-01 -2.82168746e-01 2.40696408e-02
4.97776896e-01 -5.05760372e-01 6.24137282e-01 1.33559930e+00
4.91055995e-01 -4.75046575e-01 -2.05024004e-01 9.35384393e-01
-1.78241581e-01 9.28901881e-02 -8.09264898e-01 -5.33979595e-01
2.83247560e-01 7.84795046e-01 -2.35650480e-01 -2.97206759e-01
-5.44352412e-01 8.48623097e-01 5.60697973e-01 5.84711373e-01
-7.01120675e-01 -2.55506396e-01 3.80976439e-01 1.94319159e-01
2.75296539e-01 -3.57367218e-01 -4.54798490e-01 -1.11463916e+00
2.95854747e-01 -1.25396943e+00 -9.67842564e-02 -5.89840591e-01
-1.09758794e+00 6.58019125e-01 -2.88838387e-01 -1.07564843e+00
-4.57764864e-01 -5.86419582e-01 -5.98382533e-01 1.29580748e+00
-1.18365407e+00 -7.41688907e-01 4.18637365e-01 1.05990274e-02
7.90665388e-01 -4.06291112e-02 8.65890026e-01 3.64994675e-01
-3.83254528e-01 7.51099527e-01 3.38352293e-01 2.12452814e-01
7.44895160e-01 -1.13266611e+00 9.10449386e-01 7.13872731e-01
2.25912288e-01 6.79464817e-01 1.21363437e+00 -4.66261655e-01
-1.34600472e+00 -8.97386372e-01 1.22716701e+00 -6.36634171e-01
7.05295026e-01 -5.02970934e-01 -1.31076741e+00 8.26184690e-01
2.91484535e-01 -3.28823537e-01 6.58977866e-01 1.36025414e-01
-4.35607851e-01 4.42701280e-01 -8.79480422e-01 6.69467926e-01
9.73893702e-01 -5.52321613e-01 -6.66075349e-01 5.07899404e-01
5.87028742e-01 -5.36852419e-01 -7.07376420e-01 2.64876932e-01
5.87500870e-01 -9.11254227e-01 4.31498230e-01 -1.12905407e+00
7.78738618e-01 1.16219550e-01 -1.18247651e-01 -1.65458226e+00
-1.22596323e-01 -6.16993368e-01 -2.79466007e-02 1.17623580e+00
8.94383132e-01 -4.57013547e-01 8.78407657e-01 4.70420659e-01
-9.02810618e-02 -9.05396938e-01 -8.90685976e-01 -9.84803259e-01
6.96745336e-01 -2.59089231e-01 3.39373916e-01 7.59996712e-01
1.88182537e-02 6.37392938e-01 -4.25258458e-01 -2.49560624e-01
3.03191125e-01 -1.67087913e-01 5.95822155e-01 -9.72058475e-01
-5.02372980e-01 -5.58718860e-01 -6.80646598e-02 -1.04800236e+00
2.98583418e-01 -9.63189840e-01 3.99435490e-01 -1.13039064e+00
-2.00149082e-02 -4.08923149e-01 1.00851450e-02 9.66063067e-02
-3.10795933e-01 -9.30692926e-02 2.03222960e-01 1.43438563e-01
6.34020045e-02 4.72041726e-01 1.15709591e+00 1.33927882e-01
-1.28647119e-01 1.17027357e-01 -4.06438500e-01 4.25581425e-01
8.80904913e-01 -6.75872147e-01 -2.16765046e-01 -7.38992453e-01
6.54994845e-01 2.28788450e-01 6.81232885e-02 -6.80647850e-01
-1.21494606e-01 -3.91829088e-02 2.50191420e-01 -2.56392539e-01
1.79971769e-01 -6.93190336e-01 1.27920523e-01 4.77126688e-01
-7.32948661e-01 3.53227288e-01 2.36656457e-01 3.70214731e-01
-1.36103094e-01 -4.60298330e-01 7.57978141e-01 -3.10292929e-01
1.44333437e-01 -1.13852315e-01 -5.42749405e-01 1.92470357e-01
3.74160022e-01 -2.09295854e-01 -2.34851480e-01 -3.78361493e-01
-4.99913007e-01 6.73302934e-02 6.23885095e-01 4.16097611e-01
2.15715572e-01 -1.23412299e+00 -9.63930905e-01 2.91306436e-01
-1.02373086e-01 -1.48919538e-01 -2.45665371e-01 1.11863077e+00
-2.78001785e-01 5.21107137e-01 -1.39586821e-01 -4.61887330e-01
-8.97895336e-01 3.59022290e-01 3.94256294e-01 -5.26019096e-01
-5.28629422e-01 7.17682004e-01 -1.10625587e-02 -6.49381459e-01
3.13765079e-01 -2.10837826e-01 4.24025983e-01 -1.70321509e-01
2.86579221e-01 9.29498821e-02 2.93949902e-01 -3.40612203e-01
-7.35466555e-02 3.07811618e-01 -2.92280525e-01 -2.25760177e-01
1.23417521e+00 1.49350584e-01 1.11171409e-01 6.74058914e-01
1.10628581e+00 -1.24285020e-01 -1.22982144e+00 -3.03804874e-01
3.60718668e-01 -3.40400368e-01 -4.68083203e-01 -7.63377726e-01
-6.26750052e-01 1.21208274e+00 3.76842618e-01 1.48771212e-01
6.52914405e-01 -3.51924717e-01 1.06230807e+00 3.32864046e-01
1.96708962e-01 -1.11895049e+00 -2.55994678e-01 8.04346263e-01
8.27562630e-01 -1.10184205e+00 -3.14768165e-01 -3.76983471e-02
-4.94100064e-01 1.30664599e+00 4.35988039e-01 -2.04532802e-01
-9.87696052e-02 3.21521938e-01 2.04888970e-01 2.73162901e-01
-1.17207611e+00 2.19477192e-01 1.09381238e-02 4.49339062e-01
1.00210667e+00 -1.17863759e-01 -5.36170244e-01 1.05297312e-01
-3.99569482e-01 5.75164035e-02 5.99856555e-01 6.68642819e-01
-3.52613717e-01 -1.36772597e+00 -3.66995454e-01 5.62061489e-01
-7.62385666e-01 -3.79840940e-01 -4.81144696e-01 4.97178465e-01
-1.62932828e-01 6.87143922e-01 1.41796038e-01 -4.81552780e-02
2.70908207e-01 7.83134997e-01 6.78783596e-01 -7.65377581e-01
-8.88218105e-01 -2.87476629e-02 2.82799631e-01 -2.39399254e-01
8.74639302e-02 -5.81713557e-01 -9.47016776e-01 -1.70752272e-01
-3.29450935e-01 2.94272900e-01 7.35061765e-01 1.16378236e+00
1.92716554e-01 3.45996380e-01 1.30172133e-01 -7.99361885e-01
-1.18093121e+00 -1.38827515e+00 -1.41630918e-01 6.17590725e-01
3.72818530e-01 -2.52789706e-01 -4.56418186e-01 -2.20776834e-02] | [11.635783195495605, 10.01497745513916] |
a2c2b314-fbd7-4a23-8df4-db5abf6b2570 | go-with-the-flows-mixtures-of-normalizing | 2106.03135 | null | https://arxiv.org/abs/2106.03135v3 | https://arxiv.org/pdf/2106.03135v3.pdf | Go with the Flows: Mixtures of Normalizing Flows for Point Cloud Generation and Reconstruction | Recently normalizing flows (NFs) have demonstrated state-of-the-art performance on modeling 3D point clouds while allowing sampling with arbitrary resolution at inference time. However, these flow-based models still require long training times and large models for representing complicated geometries. This work enhances their representational power by applying mixtures of NFs to point clouds. We show that in this more general framework each component learns to specialize in a particular subregion of an object in a completely unsupervised fashion. By instantiating each mixture component with a comparatively small NF we generate point clouds with improved details compared to single-flow-based models while using fewer parameters and considerably reducing the inference runtime. We further demonstrate that by adding data augmentation, individual mixture components can learn to specialize in a semantically meaningful manner. We evaluate mixtures of NFs on generation, autoencoding and single-view reconstruction based on the ShapeNet dataset. | ['Federico Tombari', 'Luc van Gool', 'Riccardo Spezialetti', 'Mengya Liu', 'Janis Postels'] | 2021-06-06 | null | null | null | null | ['point-cloud-generation'] | ['computer-vision'] | [ 7.81513471e-03 -1.49391154e-02 1.71516240e-02 -2.66829818e-01
-6.01554990e-01 -9.31342781e-01 8.30508888e-01 -1.67762354e-01
1.05418742e-01 5.55361211e-01 5.98497242e-02 -1.83746859e-01
2.12937333e-02 -1.16839397e+00 -1.04082739e+00 -3.37354600e-01
-8.56932849e-02 1.21003914e+00 2.07614750e-01 1.15432218e-01
1.17158599e-01 1.25127530e+00 -1.63495803e+00 2.81499475e-01
9.07579958e-01 7.18388379e-01 1.49062812e-01 9.90887582e-01
-5.07069886e-01 5.98890364e-01 -4.24117416e-01 -2.37809986e-01
6.13578856e-01 1.44409046e-01 -7.44445324e-01 3.89409840e-01
1.33302021e+00 -5.24707794e-01 -3.84630382e-01 5.60056746e-01
2.15698153e-01 3.40861201e-01 9.05326962e-01 -1.20856869e+00
-3.66247624e-01 2.75892794e-01 -2.38938794e-01 -7.51873925e-02
-8.97347648e-03 3.95268828e-01 8.45638335e-01 -1.00487602e+00
7.29922116e-01 1.43886709e+00 5.92488289e-01 6.08569801e-01
-1.85463905e+00 -7.42549181e-01 3.12572628e-01 -3.75498950e-01
-1.30167866e+00 -4.87053573e-01 7.90310979e-01 -5.56882203e-01
1.19458294e+00 1.57425657e-01 7.91765869e-01 7.71303296e-01
-2.34990358e-01 8.13542962e-01 5.88371515e-01 1.56328157e-01
3.32738698e-01 -1.11824758e-01 -2.27901980e-01 8.56495321e-01
2.89904535e-01 -6.71569332e-02 -2.83370584e-01 -3.72678250e-01
1.53100991e+00 -4.61984836e-02 -6.55360371e-02 -8.16670060e-01
-1.21270514e+00 8.16900313e-01 4.71036136e-01 -8.44934061e-02
-2.42252469e-01 6.97911859e-01 9.79520902e-02 -8.22689161e-02
6.31003141e-01 4.54337448e-01 -5.87402165e-01 -7.06606209e-02
-1.11705065e+00 7.49881744e-01 8.63704383e-01 1.33762801e+00
1.11883664e+00 4.08489466e-01 1.23746462e-01 6.58543468e-01
1.90590367e-01 7.07535684e-01 -1.99497998e-01 -1.69033313e+00
4.33810949e-01 5.50539374e-01 1.52433380e-01 -6.06473386e-01
-1.31840736e-01 -7.14963377e-01 -7.47260332e-01 7.01948225e-01
3.88030469e-01 -9.61824600e-03 -1.22783113e+00 1.72765601e+00
4.96883780e-01 7.23742902e-01 -5.28658461e-03 5.44466138e-01
5.31698108e-01 7.92645574e-01 9.18802768e-02 3.65065366e-01
1.01583183e+00 -9.17740107e-01 1.61543526e-02 -1.69162273e-01
4.55855519e-01 -5.80042541e-01 9.16783392e-01 4.78138447e-01
-1.33748424e+00 -6.40141666e-01 -7.63376892e-01 -2.56723613e-01
-7.20022768e-02 -2.42672697e-01 1.10959530e+00 6.34842515e-01
-1.22992694e+00 9.26974952e-01 -1.17197859e+00 -2.15112735e-02
1.02205706e+00 3.91370356e-01 -4.05536085e-01 -1.89266935e-01
-3.93369108e-01 5.94413280e-01 -1.89603567e-01 -4.24151927e-01
-1.10958350e+00 -1.59505236e+00 -9.04675663e-01 2.16048256e-01
7.58099183e-02 -1.58812630e+00 1.24222124e+00 -4.66743499e-01
-1.30102396e+00 4.90240037e-01 -4.92849350e-01 -4.74530667e-01
6.38058603e-01 -1.02149695e-01 7.22212046e-02 4.59424615e-01
1.79259047e-01 1.33150613e+00 8.52830470e-01 -1.54824710e+00
-4.60909814e-01 -3.16958070e-01 2.51640141e-01 1.34371176e-01
1.55013010e-01 -5.59550464e-01 -3.88928026e-01 -5.65397322e-01
3.05882446e-03 -8.84985924e-01 -5.23281991e-01 6.71576977e-01
-1.23094238e-01 -1.65901452e-01 9.86041903e-01 -1.78159222e-01
6.04192376e-01 -1.76852405e+00 3.40520233e-01 2.20959947e-01
4.04905975e-01 1.73650935e-01 -6.50316849e-02 8.53569657e-02
1.43755004e-01 3.88861388e-01 -5.40684283e-01 -9.15754497e-01
-1.02818049e-01 6.52462006e-01 -3.94965619e-01 2.11352527e-01
4.39399630e-01 7.80546904e-01 -9.66062129e-01 -5.10475874e-01
8.00968170e-01 8.14742804e-01 -1.32832539e+00 5.63010275e-02
-5.65725505e-01 7.21844792e-01 -4.40074474e-01 4.75370884e-01
9.50993180e-01 -4.15233999e-01 -1.73984155e-01 -2.04897001e-01
3.50962505e-02 1.95952743e-01 -1.41876841e+00 2.16735911e+00
-8.79623830e-01 2.86929667e-01 1.13490529e-01 -7.41560638e-01
7.17463195e-01 2.11648449e-01 6.92403913e-01 2.74947166e-01
-2.70661980e-01 8.79337117e-02 -3.04335266e-01 -8.99682939e-02
4.87885773e-01 -4.09171969e-01 2.04968110e-01 2.81070143e-01
4.18850094e-01 -6.46172345e-01 1.09227002e-01 3.66379619e-01
9.88840401e-01 4.38719243e-01 -2.13325351e-01 -2.58314252e-01
3.28751147e-01 1.08690541e-02 3.30889225e-01 6.76806808e-01
3.19097966e-01 9.36561882e-01 1.48245022e-01 -5.06460249e-01
-1.39267147e+00 -1.53895307e+00 -2.96293348e-01 4.17301655e-01
-7.20407907e-03 -5.16490161e-01 -4.47721004e-01 -7.17349470e-01
4.90738302e-01 1.01686227e+00 -4.17749137e-01 1.69223845e-01
-8.13599885e-01 -3.88790399e-01 6.02167904e-01 8.36782157e-01
2.66234219e-01 -6.50524139e-01 -6.06598258e-01 2.50810355e-01
1.67592894e-02 -1.28925109e+00 -3.49532992e-01 -1.15087375e-01
-1.42406178e+00 -9.30173457e-01 -6.51280642e-01 -5.12331545e-01
9.24555361e-01 2.79884040e-01 1.40995347e+00 -4.96413112e-02
-2.50022650e-01 6.24434292e-01 2.26252526e-01 -8.93811509e-02
-4.85821486e-01 9.94208530e-02 -1.95686013e-01 -1.73849612e-01
-1.64802715e-01 -1.12467539e+00 -5.50656855e-01 1.16145536e-01
-1.08030176e+00 1.85176373e-01 2.35517099e-01 4.92856562e-01
6.39025271e-01 -1.96298569e-01 1.55807927e-01 -9.87934470e-01
7.79752657e-02 -4.50909972e-01 -5.97817183e-01 -1.32415786e-01
-1.92463487e-01 5.00716805e-01 8.33190322e-01 -3.76224250e-01
-1.21690333e+00 2.38569498e-01 -7.90110007e-02 -1.18559980e+00
-4.98557597e-01 -1.32597849e-01 -1.36800379e-01 -3.20286781e-01
5.90488970e-01 7.33585954e-02 7.66871274e-02 -6.04999185e-01
7.84500718e-01 -3.66132408e-02 4.44008410e-01 -1.19590580e+00
9.83813822e-01 9.27814066e-01 5.23377478e-01 -8.87757361e-01
-5.00639558e-01 -3.72843832e-01 -7.58783221e-01 -1.04648091e-01
6.76888645e-01 -1.00559640e+00 -5.54617822e-01 2.95747340e-01
-1.26274765e+00 -3.75583827e-01 -7.54533589e-01 3.38803768e-01
-8.01380932e-01 3.20022255e-01 -4.36425805e-01 -6.59320772e-01
1.74006037e-02 -1.25484073e+00 1.47267425e+00 -5.58981709e-02
-1.62713870e-01 -1.10834754e+00 6.22447170e-02 2.77778119e-01
3.55227828e-01 2.29279712e-01 9.15504992e-01 -1.84551015e-01
-1.10740566e+00 9.32242796e-02 -6.88009337e-02 3.04783076e-01
6.72928467e-02 2.88485825e-01 -1.02088451e+00 -6.70454949e-02
-2.86592424e-01 3.72051001e-02 8.64261329e-01 3.90508085e-01
1.35893345e+00 -2.64140338e-01 -4.34733123e-01 1.10923481e+00
1.52324855e+00 -9.39219445e-02 4.64041948e-01 -2.10622564e-01
9.87772703e-01 3.28820646e-01 -7.08275586e-02 3.37607920e-01
3.71418029e-01 3.97167861e-01 6.33529067e-01 1.36154637e-01
-3.92431945e-01 -5.01871645e-01 -1.10826701e-01 3.97973746e-01
-4.16400343e-01 -2.00019330e-01 -8.81588638e-01 6.93106413e-01
-1.58612823e+00 -9.44507957e-01 -1.04833111e-01 2.06031585e+00
4.36138660e-01 5.22021987e-02 5.43382876e-02 -8.71782377e-02
4.37639445e-01 1.81569666e-01 -6.68510616e-01 -5.05581535e-02
-7.36385509e-02 7.36188054e-01 5.84396422e-01 9.09213603e-01
-9.30948973e-01 1.02753675e+00 6.56425428e+00 6.07296407e-01
-9.27063644e-01 -2.51780488e-02 5.09585500e-01 -2.40060151e-01
-8.76852393e-01 2.92906165e-01 -8.09209287e-01 5.48389629e-02
7.65401363e-01 2.17004423e-03 4.83462363e-01 8.51975977e-01
-7.83813074e-02 2.55268127e-01 -1.18428433e+00 9.13629413e-01
-3.64257768e-02 -1.93860114e+00 6.96835041e-01 2.68022329e-01
8.58386338e-01 1.66627124e-01 -1.04536921e-01 5.49986213e-02
5.51397860e-01 -1.02076626e+00 7.52776682e-01 5.59769154e-01
7.13377297e-01 -7.50726223e-01 2.35343557e-02 3.96797210e-01
-1.33301020e+00 2.09809765e-01 -5.33051431e-01 5.67614548e-02
4.47093487e-01 5.23565888e-01 -7.97593653e-01 7.17236876e-01
3.58407915e-01 7.58239567e-01 -3.47605914e-01 1.06039405e+00
-7.50757928e-04 5.01861751e-01 -8.16497564e-01 4.56849307e-01
2.31943220e-01 -3.01159084e-01 8.70977879e-01 1.03563583e+00
4.24159437e-01 -1.91958752e-02 2.08706900e-01 1.44771123e+00
-7.49392807e-02 -2.89689660e-01 -7.08959699e-01 3.03746969e-01
6.00787520e-01 1.12500453e+00 -6.42392337e-01 -6.18152380e-01
-2.60547608e-01 6.70722902e-01 4.18855131e-01 3.82712692e-01
-7.50259757e-01 1.83808338e-02 1.11675847e+00 4.30791795e-01
6.29898727e-01 -5.90492606e-01 -5.34583211e-01 -1.40391910e+00
-1.52625069e-01 -2.25505725e-01 -5.06632179e-02 -8.63087833e-01
-1.31466436e+00 4.30817902e-01 4.45711970e-01 -1.33190084e+00
-2.91408092e-01 -6.23133302e-01 -4.49713230e-01 9.39905703e-01
-1.45802367e+00 -1.23459518e+00 -3.23436677e-01 5.03223062e-01
5.43424964e-01 1.02455817e-01 8.73504043e-01 1.94386154e-01
-2.54786592e-02 1.88244924e-01 -3.22708160e-01 -2.26483475e-02
2.73968577e-01 -1.35373962e+00 8.23821545e-01 6.38602376e-01
3.62979293e-01 6.88806415e-01 5.00424445e-01 -6.71369731e-01
-1.24427938e+00 -1.28868055e+00 5.43398857e-01 -5.92478752e-01
2.86255121e-01 -3.56033236e-01 -8.58352304e-01 9.39878523e-01
-9.74123925e-02 3.96763206e-01 4.32659060e-01 -1.04923710e-01
-6.20505750e-01 2.37515010e-02 -1.44293976e+00 5.25600731e-01
1.50362206e+00 -4.08159435e-01 -4.04495001e-01 3.15262437e-01
8.12821865e-01 -4.70752269e-01 -1.07096100e+00 4.41044778e-01
2.91232288e-01 -1.04433465e+00 1.51064646e+00 -7.68265545e-01
5.53274035e-01 -5.61020792e-01 -2.27743179e-01 -1.30521083e+00
-4.15640235e-01 -5.40167272e-01 -4.83729929e-01 1.02287674e+00
1.58160448e-01 -5.87134898e-01 1.16945863e+00 5.76860011e-01
-4.56329435e-01 -6.68381095e-01 -8.07361543e-01 -7.20833838e-01
5.05868614e-01 -6.70500457e-01 9.74809408e-01 8.25049162e-01
-5.98685861e-01 1.38534844e-01 4.19281237e-02 3.48234266e-01
1.00782013e+00 1.68097913e-01 1.07637286e+00 -1.47554779e+00
-3.57210517e-01 -5.86621284e-01 -5.21560311e-01 -1.52971172e+00
2.83058584e-01 -1.16539443e+00 -3.39187294e-01 -1.66536593e+00
-2.23262355e-01 -9.07764554e-01 3.50011051e-01 2.61883587e-01
4.40951958e-02 1.54144272e-01 2.55015194e-01 1.42097369e-01
-8.09425935e-02 4.82496470e-01 1.56292510e+00 -1.19608819e-01
-1.47865135e-02 3.19762453e-02 -5.21845937e-01 7.56211221e-01
6.00950122e-01 -1.95926726e-01 -7.55833507e-01 -7.63802946e-01
-9.71001983e-02 2.19231576e-01 6.23989224e-01 -1.19192004e+00
4.19323444e-02 -1.14057757e-01 6.64291918e-01 -7.30961800e-01
7.84374297e-01 -9.13216114e-01 6.09902620e-01 1.86142907e-01
-2.36026958e-01 -8.01598802e-02 3.91493827e-01 6.10069990e-01
8.24724361e-02 -4.71720211e-02 8.53597701e-01 -6.52775764e-01
-3.42584193e-01 8.04781199e-01 -9.24471542e-02 1.18157320e-01
8.09807777e-01 -4.42446589e-01 -1.19509235e-01 -4.27483469e-01
-8.51597190e-01 5.43735065e-02 8.33327353e-01 1.77639171e-01
5.52971661e-01 -1.30811942e+00 -6.15712225e-01 5.80588162e-01
-9.60993394e-02 1.02776003e+00 4.56079990e-01 1.21694326e-01
-8.51593494e-01 1.90475762e-01 -2.33470723e-01 -1.05427933e+00
-7.13640332e-01 3.13623160e-01 4.43350792e-01 -1.68870211e-01
-8.73582184e-01 9.02303398e-01 5.35837531e-01 -7.26997733e-01
-1.45039275e-01 -5.72406232e-01 1.69925600e-01 -3.19775641e-01
1.32791325e-01 5.63518047e-01 -3.00797056e-02 -6.61076307e-01
-1.46164373e-01 8.51433098e-01 2.10156277e-01 -2.36290440e-01
1.38658631e+00 3.30212951e-01 1.50773972e-01 1.77015945e-01
1.22747195e+00 1.13958545e-01 -1.74936926e+00 1.23324305e-01
-6.17811918e-01 -8.04965973e-01 4.22754772e-02 -4.80811477e-01
-1.36355913e+00 1.07397139e+00 -4.00633886e-02 -1.13123804e-01
5.67395329e-01 1.31320894e-01 6.57405913e-01 1.79783076e-01
6.27220392e-01 -3.76856238e-01 -2.14224681e-01 3.76570702e-01
7.26438582e-01 -6.74795806e-01 -7.79906586e-02 -7.85342097e-01
-3.37361753e-01 1.29939151e+00 6.17534637e-01 -6.90601230e-01
8.60803843e-01 5.05640566e-01 -3.48136514e-01 -1.11420542e-01
-8.54018152e-01 1.32443979e-01 2.83836186e-01 8.13990414e-01
8.54666829e-02 7.50896633e-02 6.30097270e-01 5.81863262e-02
-4.64497477e-01 -1.01705249e-02 5.46743453e-01 8.43971968e-01
-2.10873917e-01 -1.15987992e+00 -1.98380545e-01 6.10361040e-01
-9.06072930e-02 -3.83512266e-02 8.75340477e-02 7.70977557e-01
1.02845632e-01 4.54986662e-01 6.87831640e-01 -1.68286532e-01
3.45310718e-01 1.55103609e-01 1.01354742e+00 -7.22627044e-01
-4.12571400e-01 1.43835530e-01 1.87079273e-02 -7.08452404e-01
-5.47049522e-01 -9.02941346e-01 -1.34118748e+00 -4.66973901e-01
-5.88772744e-02 -1.69758908e-02 5.31211317e-01 6.07896984e-01
7.52632678e-01 4.80771661e-01 3.46025527e-01 -1.39603066e+00
-4.76589113e-01 -6.32924676e-01 -4.12988394e-01 4.67447013e-01
3.89393121e-01 -8.03210497e-01 -5.12747943e-01 2.94239819e-02] | [8.880975723266602, -3.5914013385772705] |
575872d7-8ec9-4425-bd6e-7f590ac65c2d | revisiting-skeleton-based-action-recognition | 2104.13586 | null | https://arxiv.org/abs/2104.13586v2 | https://arxiv.org/pdf/2104.13586v2.pdf | Revisiting Skeleton-based Action Recognition | Human skeleton, as a compact representation of human action, has received increasing attention in recent years. Many skeleton-based action recognition methods adopt graph convolutional networks (GCN) to extract features on top of human skeletons. Despite the positive results shown in previous works, GCN-based methods are subject to limitations in robustness, interoperability, and scalability. In this work, we propose PoseC3D, a new approach to skeleton-based action recognition, which relies on a 3D heatmap stack instead of a graph sequence as the base representation of human skeletons. Compared to GCN-based methods, PoseC3D is more effective in learning spatiotemporal features, more robust against pose estimation noises, and generalizes better in cross-dataset settings. Also, PoseC3D can handle multiple-person scenarios without additional computation cost, and its features can be easily integrated with other modalities at early fusion stages, which provides a great design space to further boost the performance. On four challenging datasets, PoseC3D consistently obtains superior performance, when used alone on skeletons and in combination with the RGB modality. | ['Bo Dai', 'Dahua Lin', 'Kai Chen', 'Yue Zhao', 'Haodong Duan'] | 2021-04-28 | revisiting-skeleton-based-action-recognition-1 | http://openaccess.thecvf.com//content/CVPR2022/html/Duan_Revisiting_Skeleton-Based_Action_Recognition_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Duan_Revisiting_Skeleton-Based_Action_Recognition_CVPR_2022_paper.pdf | cvpr-2022-1 | ['group-activity-recognition'] | ['computer-vision'] | [ 2.96224475e-01 -2.74065048e-01 -2.66805112e-01 -1.18729889e-01
-4.60681945e-01 -3.45016532e-02 5.20954728e-01 -1.13241069e-01
-4.61319387e-01 3.12352687e-01 5.80973268e-01 1.99287832e-01
-9.38781817e-03 -7.11044967e-01 -3.25047761e-01 -6.06948555e-01
6.62284791e-02 1.62014887e-01 6.50343955e-01 -1.83529884e-01
-8.90140086e-02 6.03957474e-01 -1.56249416e+00 2.22284868e-01
4.48529005e-01 1.11505318e+00 -3.34767908e-01 3.74708027e-01
3.53016630e-02 5.98524094e-01 -4.28033769e-01 -4.06874269e-01
3.32441956e-01 -5.40190935e-01 -6.09633029e-01 2.10816294e-01
6.88287795e-01 -3.72556657e-01 -7.41588473e-01 7.35269487e-01
9.28413689e-01 5.72245102e-03 4.07823116e-01 -1.41441607e+00
-2.27412239e-01 2.33095765e-01 -8.23995531e-01 -5.65078631e-02
8.70456815e-01 5.21742642e-01 6.85571671e-01 -6.46130979e-01
5.48866212e-01 1.52857995e+00 8.23049068e-01 7.34097481e-01
-9.80545104e-01 -5.70694268e-01 1.28234968e-01 4.31278020e-01
-1.22074389e+00 -3.05914611e-01 7.26109624e-01 -2.23855257e-01
9.40809965e-01 1.13867328e-01 1.38390410e+00 1.64142954e+00
1.22451536e-01 1.26844561e+00 8.90880406e-01 -1.88240781e-01
1.22991733e-01 -9.72912967e-01 -1.19795300e-01 1.01447427e+00
3.16219687e-01 -8.06168765e-02 -1.08122921e+00 -8.04521367e-02
1.13490474e+00 3.57271522e-01 -4.54854548e-01 -6.55174077e-01
-1.49478555e+00 5.43823600e-01 6.25174284e-01 2.80313957e-02
-4.72866356e-01 6.05496228e-01 5.63636780e-01 -2.65338451e-01
2.35478386e-01 4.63727862e-02 -1.32501170e-01 -5.03799200e-01
-6.59083188e-01 2.48885006e-01 3.34304154e-01 7.85748601e-01
3.14939141e-01 -9.91205871e-02 -3.53650361e-01 5.77507138e-01
3.11766505e-01 6.71085596e-01 5.48796654e-01 -1.00619495e+00
7.02864766e-01 1.22788894e+00 -4.13783848e-01 -1.08702457e+00
-7.94908822e-01 -6.53382391e-02 -9.66990352e-01 2.03997836e-01
5.93365610e-01 1.34854272e-01 -1.15214527e+00 1.55616283e+00
5.07485628e-01 5.82600236e-02 -2.23029509e-01 1.16956401e+00
9.01833177e-01 -4.25282605e-02 1.21082954e-01 3.20570230e-01
1.27366900e+00 -9.91474390e-01 -5.40040076e-01 -1.45343870e-01
5.13177872e-01 -2.86062866e-01 9.92052734e-01 3.96400243e-01
-7.86746323e-01 -5.18456280e-01 -1.06793082e+00 -1.43722743e-01
-4.69479591e-01 2.18130022e-01 8.50598216e-01 6.58812642e-01
-1.04843676e+00 6.22534215e-01 -1.29768062e+00 -8.49592566e-01
6.59358859e-01 2.37519026e-01 -7.63622582e-01 -2.32758790e-01
-9.04376388e-01 7.61284411e-01 4.09372181e-01 1.36053666e-01
-5.80167234e-01 -1.31236374e-01 -1.13694763e+00 -2.84312785e-01
5.34448445e-01 -9.00157094e-01 8.02999437e-01 -3.66417676e-01
-1.62083030e+00 5.45869946e-01 2.19436944e-01 -3.99483889e-01
8.55146766e-01 -3.69376808e-01 -2.71219611e-01 5.42733252e-01
-4.57760543e-02 7.67462790e-01 8.47221673e-01 -4.16349232e-01
-4.45619285e-01 -8.81269276e-01 5.46545349e-02 1.81078225e-01
-2.77744293e-01 -8.73136967e-02 -9.42822576e-01 -7.34050870e-01
5.33289790e-01 -1.08469391e+00 -2.17046961e-01 5.15391171e-01
-3.28273982e-01 -4.10557568e-01 9.08792019e-01 -5.54821610e-01
1.05037463e+00 -1.97752368e+00 3.68449718e-01 4.97068539e-02
2.64859736e-01 6.04105055e-01 -1.94536701e-01 3.16583216e-01
9.05747339e-02 -2.41748288e-01 -2.01693207e-01 -2.94016540e-01
9.09882709e-02 3.43981147e-01 5.19358337e-01 6.55549705e-01
2.33219787e-01 1.34234226e+00 -8.53836119e-01 -7.28621304e-01
4.87168223e-01 6.31969333e-01 -2.35023364e-01 -2.10228637e-01
6.63538501e-02 5.18804371e-01 -5.74386895e-01 1.04809177e+00
3.02077293e-01 -3.62387627e-01 1.17190942e-01 -2.20762581e-01
4.49725389e-01 -2.61270821e-01 -1.25657153e+00 2.34024739e+00
1.10831978e-02 3.65621954e-01 -7.09747821e-02 -9.15570259e-01
8.76689136e-01 2.18887940e-01 8.67877543e-01 -8.38490844e-01
2.09051117e-01 -4.75885384e-02 -3.40257436e-01 -5.44629574e-01
1.52554363e-01 3.01040143e-01 -9.04580057e-02 5.37940860e-01
1.47544727e-01 1.92196518e-01 3.56667787e-01 1.29092440e-01
1.42883766e+00 7.07628369e-01 4.08573866e-01 3.15510631e-01
5.88716626e-01 -3.79575133e-01 7.04831660e-01 3.90324742e-01
-5.75319707e-01 9.13291037e-01 2.65093893e-01 -6.06428623e-01
-4.40786451e-01 -1.01783490e+00 2.51524925e-01 8.17963004e-01
3.52947146e-01 -8.72774303e-01 -6.65648520e-01 -1.08117831e+00
1.96741745e-01 -7.36656189e-02 -6.06409252e-01 -2.75667310e-01
-6.75069332e-01 -4.20575768e-01 8.48761141e-01 1.05384350e+00
1.09537351e+00 -8.39259982e-01 -1.07425499e+00 1.91525355e-01
-4.87241626e-01 -1.22394419e+00 -3.64284366e-01 -3.24836165e-01
-9.19874191e-01 -1.67533648e+00 -1.10178673e+00 -1.39684051e-01
3.67505699e-01 3.03109825e-01 7.82552063e-01 1.22074485e-01
-5.86917520e-01 1.02194476e+00 -6.94122195e-01 -2.72855580e-01
6.77050427e-02 -1.46772206e-01 5.20830862e-02 2.56183863e-01
3.55829686e-01 -6.13167703e-01 -9.21275139e-01 4.11020666e-01
-7.98941612e-01 1.44004658e-01 7.01340079e-01 4.96074229e-01
4.48679835e-01 -1.18061103e-01 9.06897932e-02 -1.74168691e-01
2.47071609e-01 7.56721664e-03 -1.38411194e-01 3.94129038e-01
-3.43994290e-01 -5.05130589e-02 1.47334963e-01 -2.63683617e-01
-8.79805088e-01 4.92681891e-01 1.86926797e-02 -4.91875708e-01
-3.42556447e-01 2.61224329e-01 -3.61019492e-01 -2.32759878e-01
7.26029396e-01 1.24393649e-01 4.96943474e-01 -8.00808072e-01
1.97408631e-01 3.79564971e-01 5.53700805e-01 -4.90054250e-01
4.32142138e-01 6.84056759e-01 4.82235551e-01 -7.33909607e-01
-6.45637155e-01 -6.22671545e-01 -1.14539599e+00 -7.51100779e-01
1.06269562e+00 -9.59874272e-01 -6.42209351e-01 1.06004357e+00
-9.03907180e-01 -2.94853806e-01 -1.39455810e-01 7.05484986e-01
-7.12519765e-01 7.95435250e-01 -5.01847446e-01 -6.34320438e-01
-3.06577891e-01 -1.00465131e+00 1.60821474e+00 2.08694220e-01
-1.76990584e-01 -7.18652904e-01 -1.51905581e-01 6.65446997e-01
1.68135047e-01 9.31785285e-01 3.89890611e-01 -3.08957338e-01
-4.07729298e-01 -5.17313719e-01 -1.97612673e-01 2.89204836e-01
1.80344477e-01 -1.99217379e-01 -5.97416401e-01 -3.00979048e-01
-5.48126876e-01 -4.78981972e-01 1.03945220e+00 1.83894202e-01
1.07665730e+00 7.38156494e-03 -4.38386232e-01 6.03892922e-01
1.04190266e+00 -2.13384897e-01 7.67380595e-01 3.66597652e-01
9.49416816e-01 2.93476254e-01 5.46511471e-01 6.37349784e-01
3.73951256e-01 8.96499693e-01 5.10142207e-01 -1.72847748e-01
-6.19667351e-01 -3.19858372e-01 5.86711109e-01 4.24646974e-01
-7.61670291e-01 3.71157341e-02 -9.97379124e-01 7.10255504e-02
-2.17959023e+00 -8.81593883e-01 -1.62521586e-01 1.96455455e+00
5.36433876e-01 -5.61098419e-02 6.98696554e-01 3.51953149e-01
6.99562252e-01 3.78710389e-01 -6.18410885e-01 4.57284093e-01
-1.81389466e-01 3.04393202e-01 4.23400462e-01 -3.73929799e-01
-1.36323965e+00 6.82960689e-01 6.47225237e+00 7.38466978e-01
-1.04873085e+00 -1.27580047e-01 -8.84021744e-02 -1.22929044e-01
4.52294618e-01 -2.50363082e-01 -5.56285858e-01 3.20426613e-01
4.65906262e-01 3.01291853e-01 6.74512163e-02 7.81521320e-01
-4.26049158e-03 -3.01041514e-01 -1.11442816e+00 1.33288312e+00
2.68355280e-01 -1.02736616e+00 1.33757159e-01 1.42886758e-01
3.10913920e-01 -5.75709082e-02 -2.48794794e-01 1.99711278e-01
4.42494303e-02 -9.83119011e-01 6.19969666e-01 5.74818969e-01
8.34375978e-01 -5.36855578e-01 7.39451289e-01 2.99546778e-01
-1.44845319e+00 -3.73179503e-02 -2.22131476e-01 -2.07932189e-01
3.00928175e-01 2.06925347e-01 -4.19234425e-01 9.40603614e-01
9.04145777e-01 1.33186269e+00 -1.12729001e+00 1.17390740e+00
-5.24229527e-01 3.44600528e-01 -3.47077191e-01 -4.32800036e-03
1.42612040e-01 -1.86035726e-02 2.91124195e-01 1.17056417e+00
2.95810193e-01 1.72342211e-01 5.13134897e-01 4.29650724e-01
3.24956089e-01 8.91219825e-02 -5.58121800e-01 -3.19308937e-02
-6.90057576e-02 1.13395655e+00 -8.82439077e-01 -2.38110065e-01
-4.91550744e-01 1.16864502e+00 2.40027502e-01 3.51896644e-01
-1.02244270e+00 -3.27846289e-01 4.69756663e-01 4.57668342e-02
2.78456986e-01 -4.99799609e-01 1.08360186e-01 -1.33835816e+00
2.24293768e-01 -8.93410325e-01 6.94158435e-01 -7.50674903e-01
-1.04165745e+00 1.24111831e-01 2.49992460e-02 -1.55116618e+00
-1.10956401e-01 -8.87035668e-01 -4.53699261e-01 3.00926119e-01
-9.81915712e-01 -1.51859200e+00 -6.72550559e-01 1.10856831e+00
3.38390380e-01 -3.59717272e-02 7.95324266e-01 1.25680178e-01
-7.31103539e-01 5.27582705e-01 -4.84132618e-01 5.87461829e-01
4.70058382e-01 -1.11755407e+00 4.68701273e-01 9.46241319e-01
2.78678417e-01 3.76991808e-01 7.50411451e-02 -7.66356349e-01
-1.59291160e+00 -1.11455917e+00 3.32824111e-01 -5.28043807e-01
3.06688100e-01 -1.72534093e-01 -5.56208193e-01 4.25622970e-01
-1.40132427e-01 3.06733161e-01 6.25722587e-01 -5.80692887e-02
-5.92776000e-01 -5.01397811e-02 -9.31404471e-01 5.02868474e-01
1.73171532e+00 -3.08509320e-01 -4.35247988e-01 2.10603639e-01
3.26962471e-01 -5.11997044e-01 -1.09193289e+00 6.74465895e-01
7.54373193e-01 -1.22440565e+00 1.08325052e+00 -4.98989940e-01
1.11930639e-01 -4.42311585e-01 -9.44007933e-02 -9.66564894e-01
-2.50665545e-01 -4.32914943e-01 -4.76911008e-01 8.41399610e-01
-1.63118571e-01 -4.80744392e-01 9.81108487e-01 4.77426916e-01
9.11180601e-02 -7.54960060e-01 -1.10892010e+00 -1.09885168e+00
-4.37166870e-01 -5.40433347e-01 6.74692571e-01 4.76762325e-01
1.19433597e-01 1.70216829e-01 -4.34731692e-01 -2.25037754e-01
6.88446879e-01 2.09446754e-02 1.15843844e+00 -1.31903255e+00
-6.65817335e-02 -5.24156988e-01 -1.25323379e+00 -1.10015690e+00
-5.73550016e-02 -8.25532317e-01 -7.72481114e-02 -1.84993589e+00
1.20202102e-01 1.10674277e-01 -2.81898260e-01 9.56567168e-01
-1.04721390e-01 5.42769969e-01 4.70525414e-01 -1.64252482e-02
-9.99292374e-01 9.04698372e-01 1.50213122e+00 -3.95798266e-01
7.65345618e-02 5.19715734e-02 -1.32976636e-01 8.70738804e-01
6.06237113e-01 1.37832211e-02 -3.30473304e-01 -1.66340858e-01
-1.79707006e-01 2.70594880e-02 6.97039902e-01 -1.56832290e+00
3.27102810e-01 -1.03919588e-01 7.94963360e-01 -7.79403627e-01
6.07368410e-01 -8.09512913e-01 2.92213231e-01 5.30002117e-01
6.61787465e-02 -1.33237123e-01 -7.26310909e-02 9.47323263e-01
-1.58637896e-01 3.14592063e-01 5.32804668e-01 -2.87470967e-01
-9.99482810e-01 6.44514561e-01 -2.06204817e-01 -3.08695640e-02
9.84302461e-01 -6.74604595e-01 -4.89538908e-02 -2.95249730e-01
-7.55999684e-01 1.03730731e-01 5.44277608e-01 7.07778096e-01
8.47924948e-01 -1.73771191e+00 -4.99272168e-01 2.33669072e-01
4.76816535e-01 1.85025595e-02 1.82509899e-01 1.25147593e+00
-4.62548673e-01 3.80256146e-01 -4.61238891e-01 -8.84738684e-01
-1.41911519e+00 2.70368993e-01 2.02873707e-01 -1.46360174e-01
-1.07808685e+00 6.25381589e-01 -2.10850477e-01 -3.62031996e-01
5.97490966e-01 -2.58999735e-01 -1.08038388e-01 9.54138488e-02
5.04536510e-01 7.68218696e-01 -5.21333776e-02 -9.17439103e-01
-6.37114286e-01 9.17153478e-01 4.08926636e-01 -4.18362394e-02
1.25284135e+00 7.56964386e-02 8.59724209e-02 1.99551359e-01
1.02711940e+00 -5.37614644e-01 -1.51131499e+00 -4.46232140e-01
8.08056891e-02 -6.90154195e-01 -2.23706082e-01 -6.95762455e-01
-1.53836060e+00 8.83009493e-01 6.25820875e-01 -2.22461119e-01
1.24664581e+00 2.42246110e-02 9.86790836e-01 2.64728040e-01
8.60196114e-01 -1.27321362e+00 6.13147914e-01 3.07894498e-01
1.00924492e+00 -1.15809906e+00 4.06119704e-01 -2.30408221e-01
-6.34462178e-01 1.42014706e+00 7.30372310e-01 2.55815268e-01
5.13604403e-01 -3.01604956e-01 -4.74940352e-02 -5.45896769e-01
-1.59056559e-01 -6.34324849e-01 5.33363700e-01 9.32861865e-01
1.72275513e-01 5.04753180e-02 -2.73679435e-01 4.31145936e-01
1.71307117e-01 8.54343846e-02 1.05507955e-01 1.23746765e+00
-2.47754529e-01 -1.06427574e+00 -3.51063728e-01 2.31936827e-01
-1.46377057e-01 4.56056237e-01 -6.56691670e-01 1.04397225e+00
1.92064136e-01 8.62713158e-01 -3.20761949e-01 -6.50567710e-01
6.32928610e-01 2.40527779e-01 7.89544880e-01 -3.70475262e-01
-2.90936559e-01 2.85642291e-03 8.85999389e-03 -1.52303946e+00
-9.12581384e-01 -8.45597029e-01 -1.23987448e+00 -9.49993879e-02
-2.43645877e-01 -4.57165211e-01 4.71644789e-01 8.80461812e-01
6.36815548e-01 2.61552364e-01 1.39833286e-01 -1.11516190e+00
-4.21414077e-01 -9.37771261e-01 -7.43956804e-01 7.21428275e-01
-8.12441930e-02 -1.19452548e+00 -5.86315729e-02 -5.22750653e-02] | [7.84657096862793, 0.3700287640094757] |
40124800-4d18-4720-a5f5-cf62db2039ef | variable-rate-hierarchical-cpc-leads-to | 2206.02211 | null | https://arxiv.org/abs/2206.02211v3 | https://arxiv.org/pdf/2206.02211v3.pdf | Variable-rate hierarchical CPC leads to acoustic unit discovery in speech | The success of deep learning comes from its ability to capture the hierarchical structure of data by learning high-level representations defined in terms of low-level ones. In this paper we explore self-supervised learning of hierarchical representations of speech by applying multiple levels of Contrastive Predictive Coding (CPC). We observe that simply stacking two CPC models does not yield significant improvements over single-level architectures. Inspired by the fact that speech is often described as a sequence of discrete units unevenly distributed in time, we propose a model in which the output of a low-level CPC module is non-uniformly downsampled to directly minimize the loss of a high-level CPC module. The latter is designed to also enforce a prior of separability and discreteness in its representations by enforcing dissimilarity of successive high-level representations through focused negative sampling, and by quantization of the prediction targets. Accounting for the structure of the speech signal improves upon single-level CPC features and enhances the disentanglement of the learned representations, as measured by downstream speech recognition tasks, while resulting in a meaningful segmentation of the signal that closely resembles phone boundaries. | ['Jan Chorowski', 'Paweł Rychlikowski', 'Ricard Marxer', 'Adrian Łańcucki', 'Santiago Cuervo'] | 2022-06-05 | null | null | null | null | ['acoustic-unit-discovery'] | ['speech'] | [ 6.23956740e-01 5.91852188e-01 -2.43544430e-01 -4.80337113e-01
-6.59292758e-01 -5.50696790e-01 9.57028151e-01 4.49022800e-01
-1.17344812e-01 3.23956937e-01 6.08027816e-01 -1.97130010e-01
-9.08871442e-02 -5.59836090e-01 -6.82886600e-01 -7.95524180e-01
-1.75555483e-01 4.59142238e-01 2.59002745e-01 -5.83240949e-02
1.17172211e-01 6.05425000e-01 -1.81447637e+00 8.87176752e-01
6.55440092e-01 1.01917493e+00 2.47619867e-01 7.13603914e-01
4.86609302e-02 7.05634058e-01 -5.83807468e-01 2.07046345e-01
6.46134615e-02 -7.24997461e-01 -7.92359412e-01 3.18655938e-01
3.21893275e-01 2.73520142e-01 -2.25829348e-01 9.99899566e-01
-8.17533433e-02 1.95129514e-01 9.30421770e-01 -9.41190720e-01
-4.78619516e-01 6.96260571e-01 -1.91797540e-01 1.39044002e-01
-1.80805498e-03 -2.00480178e-01 1.64179635e+00 -6.22269392e-01
2.94022322e-01 1.49510837e+00 5.97952843e-01 3.69274318e-01
-1.90533435e+00 -2.25265995e-01 3.69840175e-01 -3.39400209e-02
-1.20186770e+00 -6.76373720e-01 7.13707447e-01 -6.35797322e-01
1.01506078e+00 2.93313295e-01 5.32460093e-01 9.58355367e-01
8.94707739e-02 5.26838124e-01 1.11136138e+00 -6.64599955e-01
5.55634737e-01 -8.21855888e-02 1.74279898e-01 5.11284947e-01
-2.45146170e-01 3.17695260e-01 -6.75076067e-01 -1.33366540e-01
7.34005690e-01 -1.14757016e-01 -2.45873719e-01 -4.40050602e-01
-9.64082241e-01 8.49977434e-01 4.94928479e-01 7.28304327e-01
-4.90192890e-01 -3.99240106e-02 3.66338074e-01 2.75402278e-01
3.67237777e-01 4.60295230e-01 -5.61269701e-01 2.49215197e-02
-1.30507767e+00 -1.52940676e-01 6.30708396e-01 6.90319717e-01
8.35298717e-01 2.64105916e-01 -3.05484563e-01 7.59357691e-01
2.23612681e-01 -2.54888117e-01 9.38422263e-01 -1.09480476e+00
3.56308848e-01 3.39240283e-01 -3.61678958e-01 -6.90048158e-01
-1.67014807e-01 -6.94069862e-01 -8.75037611e-01 2.35760912e-01
2.24024385e-01 2.62507826e-01 -1.03169644e+00 1.99205518e+00
-2.77886391e-01 8.26406553e-02 9.98723879e-02 4.59868580e-01
5.99942543e-02 1.09362829e+00 1.03431761e-01 -4.49597448e-01
1.04125297e+00 -7.60091960e-01 -4.75770682e-01 -3.13508272e-01
4.15910333e-01 -3.19112748e-01 9.39689457e-01 3.48366290e-01
-1.21798253e+00 -8.69491994e-01 -1.20486832e+00 -1.40248999e-01
-2.97137946e-01 2.98054609e-02 7.84647912e-02 3.61314505e-01
-1.17724192e+00 9.19488013e-01 -7.67361343e-01 -2.34908685e-02
3.04443628e-01 2.94451624e-01 -3.27270716e-01 3.44183534e-01
-1.17228901e+00 7.63041496e-01 5.18681526e-01 -1.46399990e-01
-8.13659072e-01 -7.20654368e-01 -8.08661759e-01 4.61516201e-01
-2.37934560e-01 -2.83123463e-01 1.13578451e+00 -1.27614129e+00
-1.53520966e+00 8.00415695e-01 -3.60309362e-01 -7.88921773e-01
1.44822642e-01 9.29361880e-02 -1.81773350e-01 3.99060398e-01
1.97558235e-02 9.01923001e-01 1.32332623e+00 -1.28585887e+00
-5.11725903e-01 -3.74169528e-01 -3.09388608e-01 2.32965052e-01
-1.97655231e-01 -3.55075121e-01 5.44025376e-02 -8.33228588e-01
5.32290101e-01 -7.23958671e-01 -1.81943536e-01 -2.22939193e-01
-2.19126493e-01 -2.69667834e-01 5.55608869e-01 -7.37415373e-01
1.11272061e+00 -2.60008335e+00 3.42964500e-01 1.81856588e-01
1.30605832e-01 9.72656384e-02 -1.29311219e-01 4.88492221e-01
-2.41223425e-01 2.44830787e-01 -5.38887620e-01 -6.27158046e-01
3.36957984e-02 4.97789055e-01 -6.73374414e-01 5.26348352e-01
5.78590870e-01 6.08946264e-01 -7.38323271e-01 -2.87135094e-01
1.40487507e-01 6.09967530e-01 -6.44069791e-01 9.43836942e-02
-2.42617190e-01 5.05574524e-01 9.79680568e-02 -4.86945212e-02
2.17421427e-01 -5.42737059e-02 2.83521146e-01 -6.03657998e-02
-2.16329172e-01 1.13535690e+00 -7.81629980e-01 1.47919893e+00
-5.14513373e-01 8.87331188e-01 2.18766019e-01 -1.53313673e+00
9.31693256e-01 6.58106327e-01 2.00016886e-01 -7.46924341e-01
-1.59152105e-01 1.10189430e-01 2.18733490e-01 1.13225430e-01
1.86994627e-01 -4.74947572e-01 -7.64692426e-02 6.01602271e-02
3.21972847e-01 -3.39955389e-01 -1.89424843e-01 -2.78117117e-02
7.71212637e-01 -1.46429047e-01 3.62078488e-01 -5.38348913e-01
4.50961918e-01 -4.35871869e-01 6.21833563e-01 7.03434408e-01
-2.23937541e-01 8.94308329e-01 7.43352592e-01 -1.48277227e-02
-1.17077184e+00 -1.34281754e+00 -4.00793284e-01 1.30156732e+00
-2.68175095e-01 -4.41981077e-01 -7.78698266e-01 -3.25630307e-01
-9.64176804e-02 7.65276492e-01 -8.17079008e-01 -2.61604071e-01
-6.10394299e-01 -3.11021715e-01 4.77825075e-01 4.09226924e-01
8.25761110e-02 -8.76021087e-01 -5.49454868e-01 2.72234470e-01
-5.36470152e-02 -1.01983643e+00 -2.92014986e-01 1.02884853e+00
-1.01865637e+00 -5.47910094e-01 -4.68573689e-01 -1.05527532e+00
5.66979110e-01 -4.96717170e-02 1.08372164e+00 -2.38436088e-01
2.32017809e-03 -6.09339261e-03 -3.22429180e-01 2.11960435e-01
-7.78024912e-01 1.65365785e-01 7.70310760e-02 2.47209042e-01
-5.95975062e-03 -9.41073179e-01 -2.62736708e-01 -1.22038886e-01
-8.36993337e-01 4.13090736e-02 6.03998244e-01 9.41485047e-01
6.54517055e-01 2.56147236e-01 6.18848801e-01 -5.07217884e-01
4.39060360e-01 -3.33879143e-01 -3.59518051e-01 -1.70167536e-01
-4.51883435e-01 5.38386941e-01 9.55705941e-01 -4.53021467e-01
-7.42681682e-01 4.92452383e-02 -1.51806951e-01 -3.10401976e-01
-3.57883245e-01 2.76875556e-01 -2.03939676e-02 5.45877516e-01
5.33659637e-01 3.19468766e-01 -4.77599679e-03 -4.80049103e-01
6.02635384e-01 6.42372489e-01 4.94703174e-01 -5.06841004e-01
4.99329329e-01 2.19436556e-01 -2.84238935e-01 -1.24568248e+00
-8.33806336e-01 -2.36123085e-01 -9.85890627e-01 2.46167868e-01
9.77398872e-01 -7.82866299e-01 -3.42471927e-01 -3.82136274e-03
-1.25332999e+00 -3.21482629e-01 -8.40737820e-01 3.60606313e-01
-7.34683871e-01 2.06506640e-01 -7.46108353e-01 -8.50464523e-01
1.10101871e-01 -1.03207207e+00 1.00230420e+00 -2.39449799e-01
-6.45763576e-01 -9.08386707e-01 -9.49535295e-02 -5.73412690e-04
2.92271286e-01 -5.51486872e-02 1.35747457e+00 -8.81696343e-01
-2.85308182e-01 1.32875768e-02 2.07294181e-01 7.49929905e-01
1.33276522e-01 -1.23605274e-01 -1.46643388e+00 -2.63461739e-01
3.00232589e-01 -2.78143525e-01 1.19412994e+00 4.57054377e-01
1.21053505e+00 -6.09597027e-01 -7.43018910e-02 4.09819901e-01
1.04197049e+00 2.70501763e-01 4.60393935e-01 7.48823443e-03
3.47391725e-01 1.13393271e+00 -1.09290443e-01 2.20491290e-01
2.38319091e-03 5.70674002e-01 3.29050243e-01 7.93880373e-02
-2.17379287e-01 -4.75487947e-01 4.42530513e-01 1.11236596e+00
3.03765506e-01 1.71432078e-01 -7.07085907e-01 6.29169285e-01
-1.56378257e+00 -1.11499655e+00 1.61972702e-01 2.29102969e+00
1.03823316e+00 5.53230524e-01 4.44544181e-02 5.23886919e-01
6.93785191e-01 3.67621034e-01 -3.98620397e-01 -7.37044811e-01
-1.25292644e-01 3.46526861e-01 1.85177419e-02 8.01419318e-01
-9.91660833e-01 7.12897778e-01 6.48359060e+00 7.92928100e-01
-1.14013875e+00 -1.87630877e-02 7.63597846e-01 1.27367511e-01
-3.82788390e-01 6.46663606e-02 -5.88655591e-01 5.73319972e-01
1.29298568e+00 -5.93694067e-03 3.59578758e-01 5.60787201e-01
2.20914826e-01 3.16160411e-01 -1.56172431e+00 5.50364494e-01
-2.58548588e-01 -1.42412221e+00 2.00020954e-01 1.91903666e-01
5.33337593e-01 -2.53784545e-02 3.42346728e-01 3.66659760e-01
7.67138693e-03 -1.19751000e+00 1.12277448e+00 2.14266226e-01
5.54910958e-01 -6.19835258e-01 1.65489689e-01 5.89622319e-01
-1.20931387e+00 -2.19506472e-01 -2.45879009e-01 -3.36583436e-01
-2.19141185e-01 3.92074019e-01 -8.30725074e-01 1.71438619e-01
3.85113567e-01 4.98535097e-01 -2.71622419e-01 4.41927880e-01
-2.06917793e-01 8.81607831e-01 -1.35513410e-01 3.34917873e-01
4.07452077e-01 -2.18363509e-01 2.57780761e-01 1.60254443e+00
2.16235425e-02 -1.15513541e-01 1.49229676e-01 1.01066065e+00
-1.27312718e-02 -1.87475756e-01 -5.64812064e-01 -1.04116164e-01
8.24811637e-01 7.44798958e-01 -5.35875738e-01 -3.93688709e-01
-3.19313943e-01 9.81178641e-01 4.52511340e-01 4.33684260e-01
-4.80134070e-01 -2.33002663e-01 8.29256713e-01 2.27289572e-01
6.43699110e-01 -3.76228422e-01 -4.25430298e-01 -8.74028087e-01
-1.12347610e-01 -7.41216481e-01 8.11224356e-02 -3.97753894e-01
-1.26357305e+00 9.05346036e-01 -2.68451780e-01 -1.19903564e+00
-6.88413680e-01 -3.37102115e-01 -5.08979440e-01 1.25316191e+00
-1.38247240e+00 -8.81773114e-01 4.25485909e-01 4.25982773e-01
7.72512674e-01 9.01689157e-02 9.81865764e-01 -2.35963076e-01
-2.32396215e-01 4.32876647e-01 9.09570456e-02 -3.20795886e-02
1.68120220e-01 -1.43145537e+00 4.77194786e-01 7.67304122e-01
5.30655265e-01 6.72095954e-01 7.58610606e-01 -1.41156480e-01
-5.22941530e-01 -1.02374852e+00 1.12976682e+00 -1.20714374e-01
7.19285488e-01 -6.55370116e-01 -1.25815642e+00 6.29901707e-01
1.88051462e-01 3.15545388e-02 5.85967600e-01 4.97293621e-02
-6.89699888e-01 -1.59081876e-01 -8.41975868e-01 3.55382144e-01
5.95402300e-01 -1.26264215e+00 -1.16102028e+00 -6.11709319e-02
1.00698459e+00 1.95050061e-01 -7.74407864e-01 2.58692622e-01
2.86799043e-01 -1.05874527e+00 9.47001398e-01 -5.47540903e-01
4.07250345e-01 -8.42251629e-02 -4.97413546e-01 -1.32839513e+00
-5.42667806e-01 -5.52172065e-01 -1.48301750e-01 1.12849677e+00
5.94740450e-01 -4.80772316e-01 7.28818119e-01 1.33582130e-01
-3.64690274e-01 -6.71938241e-01 -1.17033780e+00 -8.56590331e-01
5.07628858e-01 -4.14758533e-01 7.41216028e-03 7.72479177e-01
3.03883553e-01 6.15119636e-01 3.44551764e-02 4.23001498e-01
4.65728879e-01 8.50768983e-02 -1.52420640e-01 -1.35463905e+00
-5.67275286e-01 -7.34954834e-01 -5.20980358e-01 -1.25255942e+00
3.84256989e-01 -1.02097714e+00 3.38943958e-01 -1.23935962e+00
-2.74433881e-01 -2.53028452e-01 -3.12664092e-01 3.13940078e-01
2.88129866e-01 9.00653005e-02 2.75303394e-01 5.01866400e-01
-8.87865499e-02 6.42010748e-01 7.45988905e-01 -1.66074425e-01
-2.52952397e-01 7.64759779e-02 -5.21255314e-01 8.10198724e-01
7.75788486e-01 -2.20689669e-01 -4.63848382e-01 -2.33117506e-01
-2.58311749e-01 1.51497170e-01 2.34853789e-01 -1.10795534e+00
2.11060047e-01 2.32113749e-01 3.57231051e-01 -4.18725967e-01
7.01486051e-01 -5.99113286e-01 -2.79849678e-01 4.61124450e-01
-1.10124183e+00 -2.88337499e-01 -1.72291677e-02 5.74372828e-01
-6.84042156e-01 -2.24890754e-01 1.17560184e+00 -1.57474607e-01
-4.13843393e-01 -9.83359218e-02 -7.48800278e-01 3.73625718e-02
5.19459128e-01 -2.35669047e-01 6.72858432e-02 -2.74743557e-01
-1.03003955e+00 -2.28458852e-01 2.34196559e-01 3.62168223e-01
4.75917459e-01 -1.13451910e+00 -4.34167594e-01 7.15727925e-01
-9.08887759e-02 -1.64918736e-01 -1.34208947e-01 5.00166774e-01
-3.61668062e-03 6.00778282e-01 -6.55476451e-02 -7.29428530e-01
-9.29784298e-01 4.78896499e-01 4.07979459e-01 -1.99534521e-01
-6.35303080e-01 8.60915124e-01 5.64072311e-01 -3.25766534e-01
5.47138155e-01 -7.14502633e-01 -1.30036831e-01 4.12848562e-01
3.14567536e-01 5.40190637e-02 1.65234864e-01 -7.53635406e-01
-2.87499368e-01 3.42918485e-01 -1.51799649e-01 -4.59902644e-01
1.30033076e+00 -1.98423594e-01 -4.53183278e-02 1.10008538e+00
1.53005517e+00 -1.36962518e-01 -1.72075915e+00 -1.28075078e-01
4.33428377e-01 1.22708060e-01 2.85848733e-02 -5.47623634e-01
-4.94474798e-01 1.35427272e+00 3.01247627e-01 7.87112594e-01
1.04568028e+00 2.79054374e-01 3.68953913e-01 3.50131169e-02
7.60885477e-02 -1.00633931e+00 3.07058156e-01 7.32312202e-01
9.09008622e-01 -8.14785898e-01 -5.12495637e-01 -1.60846308e-01
-4.88487750e-01 1.16756785e+00 7.44858012e-02 -3.20108861e-01
5.93614519e-01 2.70852089e-01 -3.37317139e-01 1.20847516e-01
-9.76208150e-01 -3.05448711e-01 2.67445952e-01 6.14661872e-01
6.32663846e-01 1.86081454e-01 1.89234279e-02 4.97392029e-01
-3.49099249e-01 -6.51147723e-01 3.85089785e-01 6.59432828e-01
-8.58583331e-01 -9.60810483e-01 -2.27005944e-01 2.07914844e-01
-3.27596009e-01 -3.43945801e-01 -2.60659099e-01 4.75227922e-01
1.51305690e-01 9.70935881e-01 5.83475769e-01 -2.75672674e-01
1.38531178e-01 4.05740708e-01 2.59923518e-01 -9.21364665e-01
-4.10061657e-01 3.01105857e-01 -1.27161711e-01 -3.01299989e-01
-2.10605472e-01 -7.45192230e-01 -1.31741893e+00 3.29639047e-01
1.14366144e-01 4.58401412e-01 4.50543523e-01 1.00551462e+00
2.75855035e-01 6.45456433e-01 8.05165052e-01 -1.11482489e+00
-9.20894563e-01 -9.88999605e-01 -7.08102167e-01 4.40320224e-01
1.07395291e+00 -3.44715327e-01 -7.29557693e-01 4.28324670e-01] | [14.810212135314941, 6.511568069458008] |
40f82d16-460b-4a0f-8096-21966df9cef1 | hierarchical-deep-reinforcement-learning-for-1 | 2212.14670 | null | https://arxiv.org/abs/2212.14670v1 | https://arxiv.org/pdf/2212.14670v1.pdf | Hierarchical Deep Reinforcement Learning for VWAP Strategy Optimization | Designing an intelligent volume-weighted average price (VWAP) strategy is a critical concern for brokers, since traditional rule-based strategies are relatively static that cannot achieve a lower transaction cost in a dynamic market. Many studies have tried to minimize the cost via reinforcement learning, but there are bottlenecks in improvement, especially for long-duration strategies such as the VWAP strategy. To address this issue, we propose a deep learning and hierarchical reinforcement learning jointed architecture termed Macro-Meta-Micro Trader (M3T) to capture market patterns and execute orders from different temporal scales. The Macro Trader first allocates a parent order into tranches based on volume profiles as the traditional VWAP strategy does, but a long short-term memory neural network is used to improve the forecasting accuracy. Then the Meta Trader selects a short-term subgoal appropriate to instant liquidity within each tranche to form a mini-tranche. The Micro Trader consequently extracts the instant market state and fulfils the subgoal with the lowest transaction cost. Our experiments over stocks listed on the Shanghai stock exchange demonstrate that our approach outperforms baselines in terms of VWAP slippage, with an average cost saving of 1.16 base points compared to the optimal baseline. | ['Qing Li', 'Chenxin Zou', 'Pangjing Wu', 'XiaoDong Li'] | 2022-12-11 | null | null | null | null | ['hierarchical-reinforcement-learning'] | ['methodology'] | [-9.17647600e-01 -1.77069142e-01 -5.58269143e-01 -2.19132319e-01
-7.97932506e-01 -5.97136438e-01 3.26903373e-01 9.93746594e-02
-4.19739544e-01 8.31137955e-01 3.63997221e-02 -6.30249023e-01
-3.05722505e-01 -1.24477696e+00 -5.91834724e-01 -5.61275303e-01
-4.88440365e-01 1.10234940e+00 1.46313742e-01 -4.73318279e-01
6.91071033e-01 3.42355549e-01 -7.19929695e-01 5.41770220e-01
8.14987123e-01 1.80946743e+00 -2.53979474e-01 4.11589444e-01
-3.38820755e-01 1.06391370e+00 -6.27993941e-01 -5.27885020e-01
1.14454305e+00 -3.76519799e-01 -3.46362442e-01 -1.75878689e-01
-9.04361978e-02 -9.09483254e-01 -2.13358492e-01 5.60824692e-01
2.30042472e-01 1.70601219e-01 5.52645206e-01 -1.05961561e+00
-4.14390028e-01 1.18998349e+00 -7.42806733e-01 6.19400144e-01
-2.24703372e-01 2.74431646e-01 1.64877868e+00 -6.70438409e-01
1.62318662e-01 8.92925739e-01 5.30044615e-01 1.32694200e-01
-1.25015628e+00 -7.05336690e-01 3.53332460e-01 1.69977337e-01
-7.96112597e-01 -7.60926604e-02 5.59432626e-01 -2.84623951e-01
1.32914245e+00 1.07198052e-01 9.69323695e-01 3.64663571e-01
9.93990064e-01 8.80662262e-01 1.15649295e+00 1.25721889e-02
3.88201028e-01 -2.40749314e-01 -1.73625126e-01 1.07202269e-01
1.54939219e-01 4.94332194e-01 -5.45471549e-01 -1.84283704e-01
1.10114670e+00 -8.70762691e-02 3.10667932e-01 -2.10770980e-01
-1.16292846e+00 1.18270934e+00 5.75764775e-01 3.80534351e-01
-9.89250481e-01 6.77615479e-02 3.78999442e-01 1.03056383e+00
4.04176563e-01 8.82210732e-01 -9.83581185e-01 -2.25317881e-01
-1.49644458e+00 8.03638101e-01 1.13413370e+00 4.83848840e-01
5.32094598e-01 5.06039202e-01 -3.71874124e-01 4.61156666e-01
1.51803583e-01 4.53081459e-01 5.77270865e-01 -1.11877155e+00
7.45083869e-01 5.26658654e-01 5.16629279e-01 -7.77319133e-01
-5.08184850e-01 -7.27205515e-01 -6.35603666e-01 4.71209139e-01
5.06373346e-01 -4.48759407e-01 -5.60418308e-01 1.13914680e+00
-7.45632797e-02 -6.24997728e-02 3.94812459e-03 9.78093088e-01
-1.69161290e-01 8.87639284e-01 -2.93941021e-01 -5.67250967e-01
1.28994441e+00 -1.15467846e+00 -6.00869358e-01 -3.72700766e-02
1.68889523e-01 -5.88831365e-01 5.34735262e-01 6.76532388e-01
-1.39694250e+00 -2.83289731e-01 -1.01688170e+00 5.97016096e-01
-3.28118593e-01 -3.04626524e-01 5.54315329e-01 3.19800556e-01
-9.04565156e-01 9.00674582e-01 -8.21125209e-01 6.02422655e-01
2.16843069e-01 4.92248833e-01 5.20292997e-01 1.10987461e+00
-1.56179571e+00 1.05822957e+00 4.11645561e-01 2.53220946e-01
-9.03945506e-01 -8.75399411e-01 -3.59038860e-01 4.09689128e-01
7.06593335e-01 -4.00266439e-01 1.81719029e+00 -1.10707116e+00
-1.73084879e+00 1.54430345e-01 5.55445611e-01 -1.37317550e+00
8.95694673e-01 -2.58084595e-01 -3.90738010e-01 -3.22254896e-02
-4.74743433e-02 1.46038786e-01 7.77857542e-01 -7.23096311e-01
-1.18357515e+00 -7.22402483e-02 1.07098483e-01 4.16427642e-01
5.83840162e-02 -1.87561885e-01 1.23606697e-01 -1.03628922e+00
1.89379379e-02 -5.65282106e-01 -3.08787465e-01 -7.89447904e-01
6.61458150e-02 -3.13150406e-01 1.50042236e-01 -8.92755687e-01
1.62763274e+00 -1.43362248e+00 -1.68210815e-03 4.99551833e-01
1.42959416e-01 -1.77349653e-02 2.69244343e-01 6.49612844e-01
5.33763841e-02 -9.76520702e-02 -8.77381116e-03 3.72219682e-01
3.13014507e-01 -2.45193932e-02 -6.72208428e-01 4.09735274e-03
-1.04321294e-01 1.17134285e+00 -5.68027914e-01 -2.08572716e-01
-1.03066258e-01 -6.13052309e-01 -5.45111120e-01 3.75235945e-01
-8.51707697e-01 5.29315323e-02 -4.71089780e-01 7.23161638e-01
5.95048249e-01 -2.91029006e-01 3.48207027e-01 1.16429433e-01
-4.73387837e-01 4.73649770e-01 -1.34797299e+00 8.95011425e-01
-4.81880814e-01 5.34365140e-02 -2.20415611e-02 -1.07245409e+00
9.47015226e-01 2.23062754e-01 8.64022911e-01 -1.36543202e+00
7.78998733e-02 6.24614000e-01 2.11248130e-01 -1.21765889e-01
6.50041163e-01 -5.04234791e-01 -3.11305135e-01 1.05506575e+00
-4.68876183e-01 2.41863251e-01 2.63233960e-01 -9.50373262e-02
7.52070844e-01 2.09854990e-01 2.69760758e-01 -3.44996393e-01
1.63940251e-01 1.07590659e-02 1.03499556e+00 7.05080509e-01
-2.37605706e-01 -2.34643742e-02 9.01172400e-01 -1.06918955e+00
-1.08002090e+00 -1.10083103e+00 2.47786582e-01 1.22982109e+00
-2.10459195e-02 1.43960223e-01 -4.74253416e-01 -7.12155461e-01
5.66163003e-01 6.56206071e-01 -4.39967275e-01 2.73039132e-01
-9.44941759e-01 -8.92900825e-01 3.74577083e-02 5.71690261e-01
8.33630562e-01 -1.29653764e+00 -1.11311936e+00 8.88446331e-01
1.26663312e-01 -5.23111463e-01 -7.50205040e-01 1.08851269e-01
-9.47700322e-01 -9.14630473e-01 -7.84878731e-01 -3.13412100e-01
-6.95904065e-03 -3.89413089e-01 1.21821034e+00 -9.47198048e-02
5.31023920e-01 -2.12898791e-01 -1.23854607e-01 -3.88713241e-01
-1.67604193e-01 4.34277892e-01 9.09241587e-02 4.01795089e-01
4.42230463e-01 -2.15877280e-01 -1.09981775e+00 2.19960317e-01
-5.69759905e-01 -1.72949299e-01 7.59068906e-01 1.09822154e+00
4.26845163e-01 7.10149556e-02 1.08289194e+00 -8.30298483e-01
9.39099371e-01 -3.18190634e-01 -1.36361611e+00 3.45198065e-01
-1.37733603e+00 2.43508562e-01 5.68925977e-01 -2.51732260e-01
-1.13170350e+00 -5.28848290e-01 1.56898156e-01 -2.37389311e-01
7.63439417e-01 7.70540178e-01 2.20270231e-01 3.63987625e-01
3.74737643e-02 3.61629814e-01 8.10534284e-02 -3.54485929e-01
-6.33572042e-02 4.50763494e-01 1.14910275e-01 -4.04274613e-01
6.56125724e-01 -5.87386936e-02 -4.22214359e-01 -6.48071170e-02
-5.40603459e-01 3.02759428e-02 -3.60521466e-01 -2.18106300e-01
4.30949152e-01 -7.67995954e-01 -1.14017582e+00 6.63682520e-01
-6.03249192e-01 -7.50358403e-01 -3.07398677e-01 5.70707560e-01
-8.98978591e-01 -2.04686522e-01 -1.36384118e+00 -8.95787656e-01
-6.76273227e-01 -9.27392066e-01 4.06182438e-01 2.30615243e-01
-1.47475988e-01 -9.77321327e-01 2.97839612e-01 2.16982812e-01
7.44175255e-01 1.42787680e-01 9.74842727e-01 -1.01476741e+00
-9.52810764e-01 7.16324449e-02 -1.46275848e-01 2.99494386e-01
9.79394093e-03 -3.77334595e-01 -3.28397185e-01 -4.46307391e-01
2.32088536e-01 -6.41218051e-02 8.27123940e-01 5.87271690e-01
6.40031874e-01 -6.65317774e-01 1.86102256e-01 4.01721984e-01
1.54438055e+00 9.28757727e-01 3.40304732e-01 9.99550819e-01
1.65358558e-01 3.12484026e-01 1.02491295e+00 7.67742634e-01
4.94708002e-01 4.41344738e-01 2.33920693e-01 1.98437989e-01
5.04064262e-01 -1.95261538e-01 4.25086796e-01 8.48435402e-01
-2.04642802e-01 1.82562590e-01 -8.58859897e-01 4.24534261e-01
-2.05722451e+00 -1.07707846e+00 5.46942830e-01 2.29504347e+00
1.00507581e+00 7.58153796e-01 5.70284486e-01 -1.39605388e-01
2.60175705e-01 3.58240783e-01 -1.04469168e+00 -8.03238690e-01
-1.99536700e-02 6.35666549e-02 8.64018440e-01 5.05952537e-01
-9.15517926e-01 9.46299791e-01 6.19906044e+00 7.61527538e-01
-1.20370495e+00 -1.59300223e-01 1.02454436e+00 -3.88302952e-01
-2.79520154e-01 9.82146412e-02 -7.92979360e-01 6.90883934e-01
1.30673718e+00 -5.06818056e-01 7.08806336e-01 6.04877353e-01
4.33767438e-01 5.89212514e-02 -1.03180492e+00 6.97009087e-01
-4.14974570e-01 -1.76481104e+00 -1.29533168e-02 3.75562817e-01
7.85730124e-01 -4.78815809e-02 1.67879149e-01 6.92922592e-01
3.40460181e-01 -7.16294587e-01 1.02651262e+00 8.07084620e-01
1.67735070e-01 -1.17355585e+00 9.35374022e-01 2.58897692e-01
-1.27796245e+00 -5.03208995e-01 -1.59866124e-01 -2.54292369e-01
2.15923861e-01 4.39954907e-01 -5.29685497e-01 6.02496922e-01
6.21035039e-01 3.62796873e-01 -1.24810517e-01 8.30950081e-01
1.60138831e-01 4.16147292e-01 -2.41516173e-01 -1.20437689e-01
6.62313282e-01 -6.81779146e-01 2.04533115e-01 6.53870344e-01
2.31189534e-01 7.63237551e-02 3.46760720e-01 7.32255161e-01
-2.99174190e-01 2.09784463e-01 -8.92642811e-02 2.56450707e-03
4.74522173e-01 9.00421798e-01 -9.13424611e-01 -6.15739644e-01
-5.11959851e-01 6.25096381e-01 2.03599498e-01 4.04715568e-01
-8.37853670e-01 -3.96101505e-01 3.82893801e-01 2.90286720e-01
6.95804954e-01 -1.64148495e-01 -6.45914197e-01 -1.26749516e+00
6.65424839e-02 -1.33108091e+00 7.49184728e-01 -5.67023596e-03
-1.34692943e+00 4.83723342e-01 -2.42947936e-01 -1.42640054e+00
-5.34211695e-01 -4.21811074e-01 -5.08879781e-01 8.07845771e-01
-1.61159742e+00 -5.89511454e-01 6.17148638e-01 3.20064813e-01
8.77037942e-01 -6.53065622e-01 2.39475504e-01 -2.66313441e-02
-3.27451646e-01 5.43660462e-01 4.03800637e-01 2.55053848e-01
3.28570992e-01 -1.56850755e+00 4.52709645e-01 4.89892900e-01
6.80961832e-02 3.34074050e-01 5.31715572e-01 -9.84586358e-01
-1.35646045e+00 -6.31322086e-01 8.93682241e-01 -2.12671027e-01
1.15357208e+00 -1.02275848e-01 -9.24375415e-01 7.42580295e-01
8.49208295e-01 -6.32366896e-01 3.46648067e-01 -8.26969370e-02
-1.95189729e-01 -7.90374279e-01 -9.43800151e-01 4.53316242e-01
1.66448116e-01 -1.64027110e-01 -9.73635554e-01 1.08849615e-01
5.65173030e-01 -4.83313948e-01 -1.23946095e+00 1.97872654e-01
8.03452730e-01 -1.10716355e+00 6.26161039e-01 -4.65167135e-01
1.85492858e-01 8.27234387e-02 1.18822642e-01 -1.40962255e+00
-2.81173170e-01 -9.32894886e-01 -5.09809077e-01 7.79966652e-01
7.62340307e-01 -9.77947354e-01 9.23241735e-01 7.83151686e-01
3.18623781e-01 -1.12422276e+00 -1.15543342e+00 -8.86091352e-01
5.61315835e-01 7.69968629e-02 1.23647332e+00 7.66912639e-01
6.97984025e-02 1.40845224e-01 -5.93816638e-01 -1.89507857e-01
8.13074172e-01 1.18319750e+00 2.63128757e-01 -8.95046890e-01
-5.21293044e-01 -1.04070866e+00 3.99558753e-01 -1.02528250e+00
-2.16566131e-01 -5.00586569e-01 -2.18519613e-01 -1.21849298e+00
-1.43959492e-01 -4.23133105e-01 -9.13455546e-01 2.65057057e-01
2.76966959e-01 -3.48954260e-01 3.61548245e-01 4.09597397e-01
-5.87259948e-01 5.44595540e-01 1.16846335e+00 -3.64136368e-01
-5.76468110e-01 3.68774831e-01 -5.47075093e-01 3.65186870e-01
1.09180081e+00 -1.51223913e-01 -8.10221657e-02 -2.57680178e-01
5.19239843e-01 7.67277896e-01 -2.96172500e-01 -4.90029484e-01
3.65716696e-01 -5.94449103e-01 4.79032427e-01 -9.69120860e-01
-2.18026116e-01 -6.39447868e-01 -2.99924240e-02 8.83151054e-01
-5.32883823e-01 7.68581450e-01 -7.81095624e-02 3.58493656e-01
-2.98682272e-01 9.67412964e-02 5.34586787e-01 -3.16889435e-01
-5.23701429e-01 5.59952080e-01 -4.57693219e-01 7.89335221e-02
1.15269840e+00 3.99964936e-02 -2.29075877e-03 -4.22718883e-01
-6.98145330e-01 9.04216111e-01 5.97185493e-02 4.79158849e-01
5.18866599e-01 -1.30944991e+00 -8.29893649e-01 1.35635316e-01
-4.38997239e-01 -2.86336392e-01 6.81630149e-02 7.96549201e-01
-8.62523854e-01 6.34163618e-01 -4.11471367e-01 -9.55538973e-02
-3.02107692e-01 6.47284210e-01 8.31502080e-01 -1.02626002e+00
-6.40633345e-01 4.32151526e-01 -1.54391125e-01 -5.97159043e-02
1.21312156e-01 -5.80715716e-01 -2.14207277e-01 7.87262797e-01
5.70741415e-01 4.64013904e-01 3.16067368e-01 -1.86890125e-01
8.29281751e-03 3.69124800e-01 -3.78109545e-01 -4.08587694e-01
1.60032547e+00 -5.58863170e-02 -1.03671081e-01 5.45544028e-01
5.52024603e-01 -2.44016960e-01 -1.55089760e+00 -2.97297180e-01
7.43104935e-01 -4.14908737e-01 5.94103849e-03 -9.43334699e-01
-1.58756447e+00 4.24561441e-01 2.92633414e-01 7.13031113e-01
1.02692008e+00 -4.81679738e-01 1.32251406e+00 2.96536416e-01
6.44655943e-01 -1.67394722e+00 -1.52752250e-01 5.09836793e-01
9.80050087e-01 -9.34455216e-01 -9.70116556e-02 6.99745953e-01
-9.96074557e-01 1.13804138e+00 6.16024315e-01 -5.86865902e-01
8.29324722e-01 3.43343586e-01 3.69712919e-01 -5.37788942e-02
-1.23624611e+00 2.05061182e-01 -5.19555733e-02 -2.21769497e-01
-7.89531320e-03 4.57283467e-01 -5.18135667e-01 8.47600639e-01
-3.53337049e-01 -5.19066229e-02 4.53490734e-01 8.87771904e-01
-4.78338718e-01 -1.13222957e+00 -2.22356290e-01 9.51426268e-01
-6.56382859e-01 -3.04440297e-02 3.16930003e-02 8.64299059e-01
-2.36655131e-01 5.19963443e-01 5.48419714e-01 -2.83556908e-01
5.06558895e-01 1.08322494e-01 1.33825198e-01 -2.55081892e-01
-1.34903252e+00 7.23818302e-01 -2.77608819e-02 -7.57220864e-01
-1.80322796e-01 -8.47066998e-01 -1.40796471e+00 -5.74481726e-01
4.17619646e-02 5.56269825e-01 5.73102236e-02 8.61157954e-01
1.50878876e-01 4.47518438e-01 1.12680805e+00 -5.28812706e-01
-1.22328544e+00 -8.78476322e-01 -1.04073501e+00 7.82038122e-02
5.95776796e-01 -6.12274110e-01 -3.32964987e-01 -3.93645316e-01] | [4.448153972625732, 3.996840715408325] |
5b8a5c6a-653b-47a4-83be-7d281fab0c73 | avaya-conversational-intelligence-a-real-time | 1909.02851 | null | https://arxiv.org/abs/1909.02851v1 | https://arxiv.org/pdf/1909.02851v1.pdf | Avaya Conversational Intelligence: A Real-Time System for Spoken Language Understanding in Human-Human Call Center Conversations | Avaya Conversational Intelligence(ACI) is an end-to-end, cloud-based solution for real-time Spoken Language Understanding for call centers. It combines large vocabulary, real-time speech recognition, transcript refinement, and entity and intent recognition in order to convert live audio into a rich, actionable stream of structured events. These events can be further leveraged with a business rules engine, thus serving as a foundation for real-time supervision and assistance applications. After the ingestion, calls are enriched with unsupervised keyword extraction, abstractive summarization, and business-defined attributes, enabling offline use cases, such as business intelligence, topic mining, full-text search, quality assurance, and agent training. ACI comes with a pretrained, configurable library of hundreds of intents and a robust intent training environment that allows for efficient, cost-effective creation and customization of customer-specific intents. | ['Marzena Żyła-Hoppe', 'Cezary Kwiatkowski', 'Marcin Baran', 'Bartosz Borowik', 'Adam Wróbel', 'Łukasz Wójciak', 'Mikołaj Morzy', 'Łukasz Augustyniak', 'Piotr Żelasko', 'Piotr Szymański', 'Robert Głowski', 'Adam Artajew', 'Yishay Carmiel', 'Jeff Hodson', 'Jan Mizgajski', 'Daniel Smoczyk', 'Adrian Szymczak'] | 2019-09-02 | null | null | null | null | ['intent-recognition'] | ['natural-language-processing'] | [-1.09502189e-02 3.21807861e-02 3.80943157e-03 -7.24552691e-01
-1.08681107e+00 -7.55291700e-01 6.74460649e-01 5.07867217e-01
2.16857288e-02 3.06920439e-01 8.52204919e-01 -1.95178762e-01
-3.59796375e-01 -4.76098329e-01 -2.35842634e-02 -2.27463081e-01
-2.77276427e-01 1.15410686e+00 -1.68875605e-01 -2.98502386e-01
1.51308149e-01 4.53203797e-01 -1.83643794e+00 6.65127277e-01
6.61458194e-01 1.24352992e+00 1.10140234e-01 9.46088374e-01
-5.21430492e-01 1.12956631e+00 -8.64637315e-01 -9.94956493e-02
-1.93604380e-01 2.20543876e-01 -9.19640660e-01 2.62731896e-03
-4.72012341e-01 -3.13570619e-01 1.22878186e-01 1.92168087e-01
5.71769953e-01 2.23219588e-01 1.02773294e-01 -1.38695931e+00
1.89528018e-01 8.27511966e-01 3.73552799e-01 -1.61518790e-02
1.07518911e+00 3.07477236e-01 1.09467113e+00 -7.67048240e-01
8.52767706e-01 9.45526361e-01 3.80995542e-01 2.32634500e-01
-6.63632870e-01 -6.88439906e-01 1.66989356e-01 2.51930684e-01
-1.13901401e+00 -8.57534528e-01 5.62484086e-01 -1.10343136e-01
1.64650750e+00 9.14780140e-01 7.29457736e-01 1.12043428e+00
-3.98484051e-01 1.08252752e+00 3.43449235e-01 -3.76451671e-01
3.75565737e-01 3.19077224e-01 3.04790676e-01 3.29709828e-01
-5.04426837e-01 -3.87098074e-01 -1.21907187e+00 -4.76085871e-01
-1.21842377e-01 1.67392284e-01 -2.55660504e-01 3.32017273e-01
-1.25216103e+00 5.23141980e-01 -3.61350775e-01 2.11604089e-01
-9.78668809e-01 -4.93883610e-01 6.75722897e-01 1.63148656e-01
3.54385346e-01 7.33924687e-01 -6.40760958e-01 -1.03097832e+00
-9.05376136e-01 3.21523100e-01 1.66146255e+00 1.18055272e+00
6.36297464e-01 7.49056861e-02 -2.84083515e-01 8.72103870e-01
6.76671192e-02 4.06862944e-01 5.04969358e-01 -1.09948444e+00
3.69014740e-01 8.12315404e-01 1.06511816e-01 -4.61133540e-01
-6.26676440e-01 -3.51782441e-01 -4.06559199e-01 -4.75627035e-01
-2.90718496e-01 -2.93774307e-01 -4.85811502e-01 1.31594396e+00
5.42481303e-01 4.24563795e-01 4.13319498e-01 6.97532177e-01
9.28396404e-01 9.78384852e-01 -7.79578090e-02 -5.68505585e-01
1.50049341e+00 -8.16330969e-01 -8.55026901e-01 -1.69346988e-01
5.52757323e-01 -7.40180910e-01 9.52933371e-01 8.24511290e-01
-1.01798391e+00 7.55449980e-02 -6.12603784e-01 2.08728254e-01
-4.21995729e-01 -4.98122394e-01 1.00749552e+00 4.69140410e-01
-9.02732611e-01 -1.50558531e-01 -5.80317676e-01 -5.21526396e-01
5.86179197e-02 3.44488382e-01 -2.39979565e-01 -3.81771401e-02
-9.94161606e-01 4.53582436e-01 3.21334690e-01 -3.00537288e-01
-9.21682894e-01 -1.11823905e+00 -7.89900601e-01 3.96532416e-01
7.18968689e-01 -5.29239714e-01 1.69682384e+00 -4.83575314e-01
-1.87083364e+00 4.81128663e-01 -3.42755139e-01 -6.71507239e-01
-1.84066281e-01 -3.41938734e-02 -7.69574702e-01 2.56964445e-01
4.29740660e-02 2.41919756e-01 7.11381316e-01 -7.43169546e-01
-1.28695762e+00 -5.00114799e-01 -2.32600674e-01 5.66740036e-01
-4.20014083e-01 5.26871800e-01 -7.13004649e-01 -2.13961348e-01
-1.07060492e-01 -5.04984617e-01 5.42269126e-02 -1.01596820e+00
-3.25003833e-01 -4.92118746e-01 1.08599758e+00 -9.49530959e-01
1.37587285e+00 -2.24215722e+00 -2.26357341e-01 4.75023866e-01
1.67137370e-01 5.99313825e-02 -1.15859695e-01 8.21719289e-01
2.39954799e-01 1.17873019e-02 4.75894392e-01 -4.73914027e-01
1.35926440e-01 8.92902091e-02 -6.75083458e-01 -4.90232408e-01
1.82012111e-01 7.10822225e-01 -7.72323549e-01 -4.78750169e-01
3.58462662e-01 2.62335986e-01 -6.41272187e-01 7.05418766e-01
-4.79511589e-01 3.53212506e-01 -4.59781617e-01 7.86416173e-01
2.46979035e-02 -1.18862897e-01 1.79353222e-01 1.95700333e-01
-2.49154940e-01 9.78163660e-01 -1.28941917e+00 1.61781323e+00
-9.54594672e-01 4.85268056e-01 6.62800491e-01 -8.00729573e-01
9.06600773e-01 7.30017722e-01 8.86649370e-01 -4.57066774e-01
-7.14662820e-02 2.87336372e-02 -7.50577033e-01 -6.36781871e-01
7.40366280e-01 3.21824998e-01 -5.33722699e-01 8.53518069e-01
2.65201092e-01 -3.97778988e-01 1.48440063e-01 4.79484230e-01
1.32499301e+00 -7.01263785e-01 1.81271702e-01 5.72690248e-01
3.15305769e-01 2.42369503e-01 4.68324631e-01 5.13488591e-01
1.97773308e-01 3.54952008e-01 8.63375291e-02 -4.90279168e-01
-7.75463641e-01 -5.75491607e-01 2.95297861e-01 1.51215923e+00
-4.03203666e-01 -7.12951064e-01 -5.70564389e-01 -2.92780131e-01
-2.71430850e-01 1.33397973e+00 3.03667814e-01 3.48594002e-02
-3.22781950e-01 -2.17057914e-01 4.68006253e-01 1.65186763e-01
2.49939963e-01 -1.34883392e+00 -3.55591983e-01 6.18262887e-01
-7.84508228e-01 -1.54674900e+00 -4.70910192e-01 4.26180184e-01
-2.83500910e-01 -5.61818302e-01 3.17037225e-01 -5.05244792e-01
-1.75203696e-01 1.85320675e-01 1.04619086e+00 -2.58718550e-01
-4.45107192e-01 8.12376618e-01 -6.40216529e-01 -8.57054353e-01
-7.10093498e-01 2.76587009e-01 1.32348374e-01 2.30380729e-01
3.96352381e-01 -6.88302636e-01 -2.69016892e-01 1.71258539e-01
-6.88389182e-01 1.56295076e-02 2.62811512e-01 5.43936312e-01
3.40636432e-01 4.90843086e-03 9.19342518e-01 -5.90671241e-01
9.70406175e-01 -4.24275130e-01 -4.47388589e-01 4.16782796e-01
-7.24250853e-01 -3.14662725e-01 5.21801710e-01 -2.14559779e-01
-1.19402480e+00 2.40358505e-02 -3.19870174e-01 -1.78406850e-01
-5.88770151e-01 7.92545438e-01 -3.22957486e-01 4.57694113e-01
4.72790956e-01 4.43763167e-01 4.85301279e-02 -2.51678616e-01
5.71475208e-01 1.57528758e+00 7.83628047e-01 -1.92961171e-01
2.59697199e-01 3.55672926e-01 -1.02933049e+00 -1.03279638e+00
-6.42461538e-01 -1.19322872e+00 -1.48054868e-01 -2.60065258e-01
6.68116331e-01 -9.98367667e-01 -1.13748264e+00 1.52569294e-01
-1.04153359e+00 -4.72620159e-01 -7.04011917e-01 2.99605101e-01
-7.04163909e-01 -1.49510562e-01 -3.77087206e-01 -1.17464089e+00
-1.03416216e+00 -7.00442553e-01 1.09578550e+00 4.25939173e-01
-7.95819104e-01 -3.70042831e-01 -3.09476048e-01 9.77369547e-01
3.58616859e-01 -4.34130669e-01 7.92326748e-01 -1.65999281e+00
-7.60572314e-01 -7.49148607e-01 1.64183661e-01 6.06333688e-02
-1.44895673e-01 9.03582945e-02 -9.47997034e-01 8.78637657e-02
6.59721866e-02 -3.51212054e-01 3.48316208e-02 4.01514582e-02
9.93672848e-01 -9.31964993e-01 -3.43747854e-01 6.31769001e-01
6.16041064e-01 5.90162873e-01 3.04119825e-01 1.87855706e-01
2.30244845e-01 6.05177283e-01 8.01655114e-01 1.06470788e+00
8.20542872e-01 6.21351480e-01 2.98525244e-01 2.97930926e-01
3.17981362e-01 8.67977366e-03 1.98061913e-01 1.06631958e+00
3.33222628e-01 -2.26072565e-01 -1.23375404e+00 6.76960588e-01
-1.82185376e+00 -1.29731846e+00 4.55715299e-01 2.03690505e+00
9.03002501e-01 2.78463569e-02 1.87966630e-01 1.59083903e-01
3.26318949e-01 -5.84501959e-02 -4.21619982e-01 -6.22582614e-01
3.44315559e-01 1.34033054e-01 8.45569149e-02 2.02391148e-01
-6.90961719e-01 1.07635760e+00 6.03582811e+00 8.06304038e-01
-1.30994523e+00 2.31073067e-01 4.32032675e-01 -5.19630909e-01
-5.05276978e-01 -1.90703601e-01 -8.73956025e-01 1.69215903e-01
1.45565629e+00 -3.27466398e-01 9.54710662e-01 1.10431004e+00
5.68059981e-01 4.54494655e-02 -8.83695781e-01 1.15739691e+00
-6.85789585e-02 -2.06512022e+00 -1.80522501e-01 -2.02111527e-01
2.27323323e-01 3.65356863e-01 -3.76040637e-01 8.45275223e-01
5.01615644e-01 -7.14422107e-01 6.05409801e-01 4.33340997e-01
6.66340590e-01 -9.93225753e-01 6.00092173e-01 7.69301713e-01
-9.62870419e-01 -6.04161561e-01 4.17010307e-01 7.06387833e-02
4.47248161e-01 4.88767385e-01 -1.67503262e+00 5.48623204e-01
8.67668867e-01 2.23658651e-01 1.66543245e-01 6.38597906e-01
2.70542979e-01 9.49091733e-01 -7.65108824e-01 -2.29961231e-01
3.93690877e-02 -6.21677414e-02 9.27083850e-01 1.32235229e+00
2.21787542e-02 5.34182072e-01 5.62576890e-01 3.56903642e-01
1.80695262e-02 3.23367715e-01 -5.21161139e-01 -3.72021586e-01
9.49427783e-01 1.57208049e+00 -4.88546133e-01 -4.83262181e-01
-2.96100587e-01 7.53561616e-01 -1.21295877e-01 3.13032240e-01
-4.75954115e-01 -5.59994996e-01 9.06936288e-01 3.67774032e-02
2.98765868e-01 -2.07905307e-01 -2.89131463e-01 -9.68328655e-01
-8.52553546e-02 -1.48314250e+00 5.77156961e-01 -8.19187760e-01
-7.50806987e-01 6.44257009e-01 -1.63392931e-01 -8.58036995e-01
-1.16647041e+00 6.41260967e-02 -9.29679275e-01 5.15387475e-01
-1.04793036e+00 -1.24547112e+00 -4.88682330e-01 8.32066834e-01
1.02354729e+00 -6.41777694e-01 1.30561721e+00 1.71839654e-01
-6.07774198e-01 2.87672460e-01 -3.12391043e-01 4.72639175e-03
4.98913765e-01 -8.46779525e-01 2.00262845e-01 3.98819208e-01
3.71004075e-01 3.64708185e-01 6.31293654e-01 -6.05453014e-01
-1.91804647e+00 -1.16331470e+00 1.09871781e+00 -3.31064254e-01
8.08375359e-01 -6.05566800e-01 -7.04448998e-01 6.97917998e-01
7.79363438e-02 -6.45127296e-01 9.79108512e-01 5.78045487e-01
-1.46854222e-01 -5.86980343e-01 -8.57366800e-01 4.98182893e-01
6.60350919e-01 -7.81178653e-01 -4.16469812e-01 7.53604174e-01
1.30009532e+00 -4.08285409e-01 -8.62850070e-01 2.19933808e-01
2.59112775e-01 -6.28707707e-01 7.88431406e-01 -6.35637820e-01
-1.69802189e-01 1.98001355e-01 -1.19383438e-02 -8.25818062e-01
2.55654991e-01 -1.51433921e+00 -7.46274134e-03 1.63123608e+00
7.10370362e-01 -5.10050356e-01 5.60630977e-01 1.07771766e+00
-5.17894804e-01 -4.23791140e-01 -8.64876866e-01 -3.53532106e-01
-1.13700604e+00 -1.38394547e+00 1.02124989e+00 8.97010565e-01
6.58847630e-01 6.18796170e-01 -1.08915539e-02 5.02101839e-01
6.34205192e-02 4.52261895e-01 1.07601273e+00 -1.15126836e+00
-5.89457899e-02 -2.94377327e-01 -2.95551509e-01 -7.37471282e-01
1.02554813e-01 -7.22686052e-01 2.22581193e-01 -1.63566208e+00
-2.34854445e-01 -5.08829176e-01 3.12384546e-01 6.44281149e-01
2.82363266e-01 -4.54484671e-01 -2.92995595e-04 2.85014808e-01
-1.01662183e+00 2.82168537e-01 3.69402200e-01 -1.26029655e-01
-9.85187352e-01 6.36378229e-01 -6.39219046e-01 4.81769383e-01
7.36095965e-01 -1.88277781e-01 -3.46811444e-01 4.70221378e-02
1.87088266e-01 4.28795934e-01 -3.78188770e-03 -8.73849690e-01
6.96301877e-01 -3.29742014e-01 -3.98499608e-01 -9.11822140e-01
7.04989851e-01 -7.44769335e-01 -3.76898050e-02 -2.36009046e-01
-4.36908603e-01 -3.38066161e-01 2.88943648e-02 3.24542135e-01
-5.53235769e-01 -2.19396316e-02 6.12025782e-02 1.92928116e-03
-8.80637288e-01 2.99404681e-01 -8.34454298e-01 1.89807728e-01
1.11717689e+00 -5.19532636e-02 -4.73568507e-04 -1.24934435e+00
-8.83861184e-01 5.16614795e-01 -1.33568525e-01 4.86995488e-01
6.03471518e-01 -6.90414906e-01 -7.61427462e-01 5.19309103e-01
3.49008501e-01 2.79130965e-01 3.08853894e-01 7.18117356e-01
-1.22717284e-01 5.05641758e-01 3.80589962e-01 -6.48840785e-01
-1.28652751e+00 2.95534492e-01 -4.80569713e-02 -3.68113190e-01
-6.60913408e-01 6.34830654e-01 -4.10081506e-01 -5.47146201e-01
7.27821112e-01 -3.23916078e-01 -1.80098057e-01 2.96551555e-01
8.95872951e-01 9.41961259e-02 5.09656012e-01 -1.59466505e-01
-4.68242079e-01 -4.49630648e-01 -6.30460978e-02 -6.21747494e-01
1.75475121e+00 -3.35793704e-01 1.59955621e-02 3.51143181e-01
7.65622854e-01 -4.35523763e-02 -7.43001223e-01 -2.28004575e-01
2.75240332e-01 1.47711560e-01 7.98893198e-02 -1.04772484e+00
-5.00540972e-01 3.11182916e-01 1.15173072e-01 5.37127197e-01
1.28489804e+00 3.20343375e-01 1.01144981e+00 7.31825054e-01
4.50655222e-01 -1.36276793e+00 -2.41171774e-02 7.48143911e-01
1.02928197e+00 -9.99337494e-01 -4.14284736e-01 -2.28161439e-01
-1.21839941e+00 8.21813822e-01 2.35256344e-01 7.04610825e-01
6.22915387e-01 8.27954769e-01 3.94223541e-01 -1.96234673e-01
-1.60332894e+00 -3.29305023e-01 7.75311561e-03 7.90451229e-01
-2.18023032e-01 2.42574349e-01 4.21487302e-01 1.12004960e+00
-6.06471419e-01 -6.74765706e-02 3.11979264e-01 7.06857324e-01
-3.98601413e-01 -6.83911562e-01 -2.31963798e-01 6.21368051e-01
-1.63737923e-01 -3.42614204e-01 -2.74246752e-01 2.35950455e-01
-3.97517145e-01 1.33732116e+00 1.97748035e-01 -4.95448053e-01
5.28091848e-01 7.74304032e-01 -5.45961380e-01 -9.29107189e-01
-1.06193662e+00 1.70855820e-01 8.01803112e-01 -8.41807067e-01
3.22872460e-01 -8.80299211e-01 -1.61915934e+00 -1.62859187e-01
-2.49002308e-01 5.31215787e-01 1.32956600e+00 1.05041111e+00
1.00604451e+00 3.79006833e-01 9.74296808e-01 -5.27071953e-01
-4.06489253e-01 -9.71054792e-01 -1.78200960e-01 4.13207799e-01
3.84289846e-02 2.24779233e-01 -2.56426036e-01 1.04556054e-01] | [12.5247163772583, 7.631338596343994] |
a79fa6d8-ae48-4e51-a916-90ae6ee1dd36 | dialogue-to-video-retrieval | 2303.16761 | null | https://arxiv.org/abs/2303.16761v1 | https://arxiv.org/pdf/2303.16761v1.pdf | Dialogue-to-Video Retrieval | Recent years have witnessed an increasing amount of dialogue/conversation on the web especially on social media. That inspires the development of dialogue-based retrieval, in which retrieving videos based on dialogue is of increasing interest for recommendation systems. Different from other video retrieval tasks, dialogue-to-video retrieval uses structured queries in the form of user-generated dialogue as the search descriptor. We present a novel dialogue-to-video retrieval system, incorporating structured conversational information. Experiments conducted on the AVSD dataset show that our proposed approach using plain-text queries improves over the previous counterpart model by 15.8% on R@1. Furthermore, our approach using dialogue as a query, improves retrieval performance by 4.2%, 6.2%, 8.6% on R@1, R@5 and R@10 and outperforms the state-of-the-art model by 0.7%, 3.6% and 6.0% on R@1, R@5 and R@10 respectively. | ['Jennifer Foster', 'Cathal Gurrin', 'Liting Zhou', 'Van-Tu Ninh', 'Manh-Duy Nguyen', 'Chenyang Lyu'] | 2023-03-23 | null | null | null | null | ['video-retrieval'] | ['computer-vision'] | [-1.19165957e-01 -2.13886932e-01 -2.86415547e-01 4.42360668e-03
-1.02382255e+00 -6.36679590e-01 1.03543031e+00 2.49878302e-01
-5.58060884e-01 6.17838323e-01 4.87929881e-01 3.01505655e-01
-2.54077166e-01 -6.26399517e-01 -1.68912977e-01 -4.16383743e-01
-4.06152494e-02 2.69122511e-01 7.08980978e-01 -5.43038130e-01
6.70096993e-01 1.72666520e-01 -1.47296286e+00 4.30379242e-01
5.10125220e-01 1.20943642e+00 1.20426916e-01 8.23496044e-01
-6.21699356e-02 9.63022411e-01 -6.42363906e-01 -2.28801697e-01
6.01712316e-02 -3.51189584e-01 -7.69865632e-01 9.63175222e-02
2.10657060e-01 -9.03596282e-01 -7.84635842e-01 6.12621605e-01
6.80928469e-01 7.58719742e-01 8.62063468e-01 -1.02622151e+00
-6.32293463e-01 9.72800776e-02 -7.11518466e-01 2.36448124e-01
1.07511115e+00 -3.59028250e-01 9.93354976e-01 -9.54286456e-01
8.64224911e-01 1.36474800e+00 -9.14851725e-02 4.86200511e-01
-4.55490053e-01 -4.27545220e-01 -3.13622087e-01 3.84328455e-01
-1.53564537e+00 -2.73589969e-01 4.92885560e-01 -3.20614666e-01
1.00775850e+00 4.40973371e-01 4.75380629e-01 8.27893198e-01
-2.14502752e-01 9.40629840e-01 7.16573358e-01 -4.10166502e-01
-1.50639772e-01 1.71596274e-01 2.20123127e-01 4.35317487e-01
-3.04483116e-01 -6.44113302e-01 -6.99861467e-01 -1.30514175e-01
7.49659657e-01 1.24829039e-01 -4.00038570e-01 1.85690656e-01
-1.10085189e+00 9.93650973e-01 -4.00890820e-02 7.28816539e-02
-3.39032829e-01 -8.97551477e-02 7.91895390e-01 4.55468744e-01
6.76047862e-01 2.72862971e-01 9.38816741e-02 -5.84260881e-01
-7.53732145e-01 3.44650924e-01 8.75716269e-01 1.12917078e+00
4.04859006e-01 -4.04835790e-01 -3.09629768e-01 1.45927131e+00
3.97177547e-01 5.63930988e-01 3.93418491e-01 -1.17082572e+00
2.94248134e-01 6.10485435e-01 3.34178865e-01 -1.21251428e+00
-4.28820066e-02 1.34090886e-01 -5.89701593e-01 -7.04612374e-01
3.51991728e-02 5.33403363e-03 -4.15431857e-01 9.97227550e-01
2.83570260e-01 -9.66523290e-02 2.59844512e-01 1.00013721e+00
1.58074152e+00 1.17557061e+00 -2.79031843e-01 -6.05282187e-01
1.31835556e+00 -1.42847681e+00 -9.30555463e-01 4.15117174e-01
5.16875565e-01 -1.29816091e+00 9.37661231e-01 3.90074104e-01
-1.27238643e+00 -4.16241854e-01 -6.62056148e-01 8.85950308e-03
-2.43881151e-01 2.61710912e-01 1.67251185e-01 2.89173841e-01
-1.35014248e+00 5.78063577e-02 -1.51999548e-01 -9.36192155e-01
-9.89222527e-02 2.18217567e-01 -4.17089969e-01 -2.01442346e-01
-1.28538978e+00 6.33879006e-01 4.15083431e-02 -2.79778838e-01
-9.06456888e-01 -2.18103006e-01 -3.70835692e-01 -1.81615859e-01
7.45053351e-01 -2.58520007e-01 1.31303632e+00 -4.14826632e-01
-1.69559026e+00 8.86673510e-01 -1.01709194e-01 -3.00281465e-01
3.94911677e-01 -7.08843589e-01 -4.79956508e-01 1.10115123e+00
5.41167781e-02 6.95319295e-01 5.91711938e-01 -8.64653230e-01
-7.31612265e-01 -2.02475101e-01 6.28107429e-01 7.20491290e-01
-4.43839341e-01 5.08639216e-01 -1.20344949e+00 -4.17814285e-01
-3.01686972e-01 -1.09623325e+00 1.36906236e-01 -4.58965711e-02
-6.03459850e-02 -7.96281874e-01 1.07443166e+00 -6.42318785e-01
1.38712299e+00 -1.95818782e+00 2.08248198e-01 -6.22477420e-02
1.26069263e-01 4.33880359e-01 2.42450964e-02 1.14246130e+00
5.19939899e-01 1.37605384e-01 6.29216015e-01 -8.44612066e-03
-2.19618186e-01 -6.54615536e-02 8.77499301e-03 3.98026913e-01
-2.79367089e-01 5.05177021e-01 -8.69676292e-01 -5.40164649e-01
2.22191855e-01 5.27792096e-01 -5.17464221e-01 5.46415508e-01
1.22856675e-02 1.12782419e-01 -7.05241263e-01 4.47402477e-01
2.90722698e-01 -2.40647480e-01 -1.05987072e-01 1.26695577e-02
-1.99450985e-01 4.19390500e-02 -7.86547005e-01 1.70916367e+00
-2.75786251e-01 1.04274619e+00 -1.65120587e-02 -8.10686946e-01
1.02438748e+00 7.29490638e-01 7.11233556e-01 -7.54974008e-01
-7.24479631e-02 3.73506136e-02 -4.26026285e-01 -8.24701130e-01
8.78274381e-01 4.78699327e-01 -1.41724825e-01 6.15687847e-01
-2.87779495e-02 2.22730413e-01 4.32911515e-01 7.16134906e-01
1.20268214e+00 -8.65847096e-02 1.66255713e-01 -8.26905575e-03
9.64414358e-01 -1.09676376e-01 -2.02349946e-01 7.14753807e-01
-9.39840823e-02 6.59241974e-01 4.10807520e-01 -9.09627751e-02
-7.34330118e-01 -5.50763190e-01 1.03779688e-01 1.35191727e+00
5.20613194e-01 -6.27114534e-01 -6.93680406e-01 -4.85044032e-01
-2.38350287e-01 2.46696293e-01 -2.76485652e-01 3.41159105e-02
-4.13785636e-01 -2.83982933e-01 4.06967014e-01 -7.77785704e-02
9.36218500e-01 -1.13095415e+00 -3.50572854e-01 -3.31673920e-02
-5.24286449e-01 -1.37611008e+00 -6.77032471e-01 -6.93869412e-01
-7.22445786e-01 -1.00453174e+00 -1.43360078e+00 -7.21156180e-01
4.83502418e-01 9.35073912e-01 1.04259193e+00 3.66468638e-01
-1.00265577e-01 1.16787350e+00 -1.18329108e+00 7.95902312e-02
-1.35844529e-01 5.76564632e-02 -1.05369329e-01 1.03360109e-01
3.04196090e-01 -8.49858746e-02 -1.08581209e+00 6.80380404e-01
-1.03390360e+00 -2.91811287e-01 2.09792987e-01 6.81243300e-01
1.66879475e-01 -2.57265359e-01 8.61152470e-01 -6.71258748e-01
8.96230996e-01 -7.67651320e-01 -2.70521700e-01 2.68525869e-01
-5.85001647e-01 -3.90786886e-01 8.15107524e-02 -3.89947951e-01
-1.03134787e+00 -5.43173075e-01 1.18280128e-01 -3.87941808e-01
-8.82683694e-02 4.66609359e-01 4.08122271e-01 1.19423948e-01
5.08779645e-01 2.86325216e-01 7.92123079e-02 -5.09587824e-01
3.17776233e-01 1.13275063e+00 2.95076557e-02 -1.54575765e-01
4.19334590e-01 4.06364173e-01 -2.84208059e-01 -1.25825334e+00
-3.34268481e-01 -1.33182549e+00 -1.80582404e-01 -8.56076956e-01
9.75765228e-01 -1.15232193e+00 -8.29946518e-01 4.21675444e-01
-1.16103268e+00 1.01755135e-01 2.11579174e-01 4.52288628e-01
-4.92086649e-01 7.79672086e-01 -7.65501142e-01 -9.86144423e-01
-6.59553409e-01 -1.07770836e+00 1.19774818e+00 2.44545281e-01
-2.29917973e-01 -6.87622964e-01 -6.83364049e-02 9.17892814e-01
4.72176552e-01 -1.13974139e-01 4.08941209e-01 -9.46751773e-01
-7.15608776e-01 -5.28396308e-01 -3.95650089e-01 1.94235608e-01
1.32818341e-01 1.52341183e-02 -5.17886698e-01 -4.09577012e-01
-2.50248700e-01 -5.11433482e-01 5.63480079e-01 5.85639067e-02
7.40901053e-01 -2.84952551e-01 -2.88721979e-01 -2.78703183e-01
1.21017289e+00 6.74452364e-01 8.88836980e-01 3.33602309e-01
1.24788702e-01 6.86993003e-01 1.07014918e+00 7.55514085e-01
3.72455418e-01 9.22670543e-01 3.25904846e-01 2.91647524e-01
-8.27260036e-03 6.13052212e-02 4.36232060e-01 1.09312952e+00
-3.24501604e-01 -5.30519783e-01 -6.33204639e-01 6.01503611e-01
-2.00257683e+00 -1.13298023e+00 4.62652966e-02 2.16843128e+00
6.25079155e-01 -2.46727452e-01 3.62843603e-01 -8.40305015e-02
7.08367169e-01 4.20545399e-01 -1.25271484e-01 -1.70095816e-01
2.57667422e-01 -2.13298783e-01 1.30943537e-01 3.32047075e-01
-9.21036184e-01 8.52850556e-01 5.67870140e+00 9.74794984e-01
-9.68175709e-01 -1.24582313e-02 4.91740137e-01 -2.39881501e-01
1.26109347e-01 -2.96734750e-01 -7.42419481e-01 3.39236706e-01
8.33246648e-01 -4.21990544e-01 5.34822404e-01 7.91358292e-01
4.63483363e-01 -6.31793320e-01 -8.97836387e-01 1.29451036e+00
5.55515587e-01 -1.19818294e+00 3.49920779e-01 -7.42197186e-02
6.39458179e-01 -2.69447803e-01 -5.99826761e-02 4.50856686e-01
-2.65319586e-01 -7.63815165e-01 8.27710703e-03 4.90099400e-01
6.03329778e-01 -7.33138382e-01 8.83148730e-01 3.14353913e-01
-1.14841735e+00 3.50691795e-01 -3.28219056e-01 1.34819984e-01
1.59092262e-01 -7.54373446e-02 -7.96689987e-01 5.68548322e-01
8.24676752e-01 8.95510256e-01 -4.10477877e-01 9.51547801e-01
2.44941339e-01 3.73623461e-01 -2.22473383e-01 -5.36597550e-01
3.66408557e-01 -3.22641790e-01 5.49670160e-01 1.36509836e+00
2.50976712e-01 5.29849112e-01 4.72920984e-02 -3.26258272e-01
-3.30977291e-01 7.12938368e-01 -8.11394930e-01 -1.51632100e-01
3.63881201e-01 1.26834905e+00 -5.64958453e-01 -4.68873739e-01
-6.57047212e-01 1.04226255e+00 -7.90342838e-02 4.44306791e-01
-7.52328277e-01 -6.30389690e-01 7.06874058e-02 1.43694252e-01
4.11473215e-01 -3.55269425e-02 1.05660093e+00 -9.74125266e-01
4.01535071e-02 -9.66687500e-01 3.56453240e-01 -1.06036794e+00
-9.98748720e-01 6.55183434e-01 3.74708652e-01 -1.49281466e+00
-5.76699317e-01 -1.68933883e-01 -2.03456536e-01 4.00692314e-01
-1.30850649e+00 -5.60492694e-01 -5.01519084e-01 6.92296922e-01
1.09115839e+00 -5.06551743e-01 8.13139319e-01 6.55735493e-01
-2.12115750e-01 4.05118853e-01 4.23464239e-01 1.16041876e-01
1.17854285e+00 -6.52256608e-01 -4.21761692e-01 7.62702674e-02
1.48495249e-02 5.39320529e-01 4.98720586e-01 -2.34417677e-01
-1.85154235e+00 -6.77572787e-01 7.23350346e-01 -2.10092187e-01
6.48924470e-01 1.71412513e-01 -5.89601874e-01 2.21909866e-01
8.44502866e-01 -2.67824471e-01 8.20353031e-01 -2.57119417e-01
-1.36806726e-01 -2.76779294e-01 -1.16764688e+00 7.28366673e-01
8.74971688e-01 -7.42729127e-01 -4.91368145e-01 7.04063058e-01
6.64391756e-01 -3.44018459e-01 -1.24482238e+00 1.71232060e-01
7.48803914e-01 -8.99800897e-01 1.05358100e+00 -2.73903638e-01
5.33367634e-01 9.74389762e-02 -3.50601047e-01 -8.48979831e-01
1.50264204e-01 -9.32725132e-01 -6.11513704e-02 1.18016815e+00
2.14589551e-01 -2.80331105e-01 5.62193215e-01 5.14689565e-01
2.23772839e-01 -6.38271689e-01 -7.69915819e-01 -4.41655010e-01
-3.91507566e-01 -7.04784468e-02 -1.81769922e-01 6.37815297e-01
1.94404781e-01 6.38160288e-01 -6.36237621e-01 -4.06687707e-01
2.80168056e-01 -2.29797177e-02 1.03123462e+00 -1.03635359e+00
-9.44678783e-02 -3.27293545e-01 -4.29251552e-01 -1.82448792e+00
-9.43241417e-02 -3.75686824e-01 -2.58386850e-01 -1.76012290e+00
4.26588207e-01 1.74373880e-01 -3.80215198e-02 -2.11210355e-01
1.58351958e-01 5.89164376e-01 3.94064724e-01 5.74741244e-01
-1.43768728e+00 5.40673912e-01 1.40392148e+00 -3.34901176e-02
-9.45754275e-02 -1.11293867e-01 -3.03337783e-01 5.36202312e-01
6.67166889e-01 -2.58324802e-01 -6.04075670e-01 -1.25897512e-01
2.37855852e-01 6.84256136e-01 -2.17406899e-01 -6.51508689e-01
4.61275041e-01 6.15836568e-02 -2.55936474e-01 -8.88651967e-01
8.73513639e-01 -8.01293492e-01 -2.09701415e-02 4.98223342e-02
-6.81742668e-01 1.59498096e-01 -7.43660629e-02 7.87575424e-01
-7.06056356e-01 -4.83885616e-01 1.23963803e-01 -2.85533756e-01
-7.53471017e-01 7.14391023e-02 -7.07347333e-01 1.15488417e-01
1.17118645e+00 -1.38190851e-01 -4.23338711e-01 -1.21875119e+00
-7.96555519e-01 4.76702899e-01 -4.15651221e-03 6.19905353e-01
8.69696736e-01 -1.30371535e+00 -7.10753560e-01 -4.89783406e-01
1.82055414e-01 -2.33903632e-01 2.77919024e-01 7.59606242e-01
-5.45030177e-01 8.71351242e-01 -1.66403723e-03 -6.37984574e-01
-1.98686004e+00 1.69018656e-01 -1.95387140e-01 -2.49365911e-01
-2.18319699e-01 6.34601176e-01 1.97329484e-02 1.54842153e-01
5.96625924e-01 2.26755008e-01 -8.01247418e-01 5.28838575e-01
6.10795498e-01 8.57010424e-01 -2.99639046e-01 -8.34497154e-01
-1.56626508e-01 7.34247327e-01 -4.46894467e-01 -2.91332871e-01
1.19171262e+00 -3.60035539e-01 -1.78429201e-01 5.20917356e-01
1.58193421e+00 -5.42891794e-04 -4.73554224e-01 -4.32992160e-01
-1.09012946e-01 -5.88405013e-01 2.53750738e-02 -5.64354181e-01
-1.03449357e+00 6.36448860e-01 5.19518375e-01 6.70986950e-01
1.16405296e+00 1.21888675e-01 8.14582586e-01 7.33206511e-01
4.60803926e-01 -1.24225032e+00 7.55078435e-01 5.21911621e-01
1.30340254e+00 -1.39739871e+00 2.58251637e-01 -4.17880744e-01
-7.44735181e-01 1.18167746e+00 4.06375229e-01 -2.73197860e-01
7.13652372e-01 -4.55167592e-01 -8.27003550e-03 -2.54888624e-01
-9.63167489e-01 -2.41092555e-02 3.74302030e-01 1.32905066e-01
6.48830175e-01 -3.43644798e-01 -7.95081079e-01 1.27365044e-03
4.55324233e-01 4.60573882e-02 5.56307793e-01 1.01516795e+00
-5.34006655e-01 -1.20345414e+00 -2.43894905e-01 4.65165377e-01
-8.27315271e-01 -2.90541220e-02 -4.49397475e-01 9.01708245e-01
-8.74781132e-01 1.41569448e+00 2.09889491e-03 -3.18132073e-01
3.81777406e-01 5.69820665e-02 1.31158158e-01 -6.01052821e-01
-6.16525769e-01 5.04051268e-01 6.01055920e-01 -5.08744597e-01
-1.01953912e+00 -4.58709210e-01 -9.99398589e-01 -3.31437111e-01
-6.07425272e-01 5.78473985e-01 4.99112993e-01 5.49290597e-01
4.80958402e-01 8.59634876e-02 8.15707207e-01 -6.48902833e-01
-3.96912992e-01 -1.17034602e+00 -4.46817815e-01 3.50377619e-01
1.87117159e-01 -5.82201540e-01 -4.95019823e-01 9.04101282e-02] | [10.442185401916504, 0.7491852641105652] |
7c0159f4-497b-458b-821f-bd812e526d85 | dilated-convolution-with-dilated-gru-for | 1906.01203 | null | https://arxiv.org/abs/1906.01203v1 | https://arxiv.org/pdf/1906.01203v1.pdf | Dilated Convolution with Dilated GRU for Music Source Separation | Stacked dilated convolutions used in Wavenet have been shown effective for generating high-quality audios. By replacing pooling/striding with dilation in convolution layers, they can preserve high-resolution information and still reach distant locations. Producing high-resolution predictions is also crucial in music source separation, whose goal is to separate different sound sources while maintaining the quality of the separated sounds. Therefore, this paper investigates using stacked dilated convolutions as the backbone for music source separation. However, while stacked dilated convolutions can reach wider context than standard convolutions, their effective receptive fields are still fixed and may not be wide enough for complex music audio signals. To reach information at remote locations, we propose to combine dilated convolution with a modified version of gated recurrent units (GRU) called the `Dilated GRU' to form a block. A Dilated GRU unit receives information from k steps before instead of the previous step for a fixed k. This modification allows a GRU unit to reach a location with fewer recurrent steps and run faster because it can execute partially in parallel. We show that the proposed model with a stack of such blocks performs equally well or better than the state-of-the-art models for separating vocals and accompaniments. | ['Yi-Hsuan Yang', 'Jen-Yu Liu'] | 2019-06-04 | null | null | null | null | ['music-source-separation'] | ['music'] | [ 2.27306306e-01 -3.29632431e-01 2.42003441e-01 -5.33689149e-02
-8.72077405e-01 -6.43754005e-01 9.73380916e-03 -1.03283294e-01
-2.33193174e-01 5.47547519e-01 4.13003564e-01 -7.90152326e-03
-2.49741212e-01 -6.83635414e-01 -5.52411199e-01 -8.33742797e-01
-1.83024973e-01 -4.52776670e-01 4.74924713e-01 -1.24818116e-01
1.08231984e-01 4.69951600e-01 -1.66689289e+00 9.28419709e-01
7.32113719e-01 7.57460237e-01 5.81781745e-01 1.17269170e+00
-2.07902685e-01 7.48645306e-01 -8.34795952e-01 1.12945832e-01
1.73044071e-01 -7.62085378e-01 -6.44569218e-01 -3.92105967e-01
6.45616591e-01 -3.19637448e-01 -2.71077037e-01 8.54026258e-01
9.53716993e-01 4.96734381e-01 2.77687848e-01 -7.68320084e-01
-5.06146550e-01 1.46135759e+00 -3.91219407e-01 2.78214484e-01
3.11332699e-02 -4.25700963e-01 1.31733537e+00 -8.62189651e-01
3.30362953e-02 1.06446660e+00 9.03663695e-01 6.62497699e-01
-1.23232162e+00 -9.07630265e-01 1.87522158e-01 -1.16931006e-01
-1.25502908e+00 -7.54829347e-01 6.49076760e-01 1.61320698e-02
1.12002063e+00 7.59664893e-01 4.79412436e-01 8.35922062e-01
-2.48884603e-01 7.34349608e-01 6.32747650e-01 -4.09910679e-01
1.30318850e-01 -2.63001293e-01 4.09743153e-02 2.19686985e-01
-3.75278704e-02 -1.73036262e-01 -8.67363811e-01 -1.60936192e-01
1.15448272e+00 1.19223915e-01 -7.42314398e-01 3.64454448e-01
-1.52906895e+00 3.84451449e-01 7.75872409e-01 6.69805706e-01
-3.76455724e-01 5.29439032e-01 2.33661979e-01 3.89964640e-01
3.74776334e-01 6.79228127e-01 -4.55058128e-01 -2.20472645e-02
-1.56125987e+00 3.74390692e-01 5.66396952e-01 5.92182219e-01
4.88645643e-01 4.54953551e-01 -4.22351748e-01 1.03887832e+00
-1.72480240e-01 1.72145888e-01 6.19117916e-01 -7.78261781e-01
5.09186447e-01 1.64148986e-01 -5.91985025e-02 -5.78907967e-01
-3.10142040e-01 -8.57676744e-01 -1.03987014e+00 3.57060254e-01
3.87105018e-01 -3.96540761e-01 -8.78957152e-01 1.73164046e+00
-7.13755414e-02 8.28226507e-01 1.97058007e-01 1.02985525e+00
9.69670892e-01 1.03712678e+00 -2.84223437e-01 -6.23288117e-02
1.14181972e+00 -1.15236187e+00 -5.77741444e-01 -6.08777180e-02
2.58277744e-01 -1.12721729e+00 7.87631869e-01 4.36233342e-01
-1.45062768e+00 -1.03903365e+00 -1.20881402e+00 -9.20446441e-02
-2.48760059e-01 3.04396659e-01 3.16531718e-01 4.15177673e-01
-1.22059572e+00 1.09437251e+00 -5.94181418e-01 1.11438401e-01
2.56977588e-01 3.67647052e-01 4.48004808e-03 5.09476006e-01
-1.07987702e+00 2.50114381e-01 1.74265608e-01 2.95344234e-01
-7.85367131e-01 -8.39380026e-01 -5.57160616e-01 4.65820819e-01
1.12300031e-01 -6.84711277e-01 1.40099072e+00 -8.41972888e-01
-1.57872534e+00 -2.18142476e-02 -1.73256248e-01 -5.81088006e-01
8.72541517e-02 -4.79325175e-01 -4.24033374e-01 -1.28152519e-01
-2.77619690e-01 6.55526876e-01 1.21620297e+00 -9.08552945e-01
-7.93296218e-01 -7.87376165e-02 6.36487873e-03 2.96107322e-01
-3.83055657e-01 1.80507734e-01 -1.42808810e-01 -1.19905281e+00
3.23682457e-01 -6.74425364e-01 -2.54238039e-01 -5.55610895e-01
-3.55198920e-01 1.17698871e-01 6.79064453e-01 -7.59107590e-01
1.61236179e+00 -2.42849708e+00 3.13259274e-01 7.70968013e-03
2.07030222e-01 4.23109084e-01 -3.92728776e-01 3.24080318e-01
-1.65773749e-01 1.66255876e-01 -2.37454548e-01 -4.83745605e-01
-2.04232708e-01 -6.18111715e-02 -6.76834404e-01 -6.80408925e-02
1.80245489e-01 7.61773109e-01 -7.74258792e-01 7.33648613e-03
-4.26969193e-02 7.93258965e-01 -6.72202289e-01 8.33084062e-02
-2.40564737e-02 4.34168130e-01 -1.35373369e-01 1.79116860e-01
6.77206039e-01 1.78409606e-01 -1.00835927e-01 -3.04393228e-02
-5.30332029e-01 7.31875479e-01 -1.55458212e+00 1.81832933e+00
-6.43126547e-01 7.46666729e-01 4.09875363e-01 -5.16057491e-01
9.94040430e-01 7.06592143e-01 1.86054409e-01 -3.90781939e-01
-1.21273302e-01 4.66756314e-01 1.93898022e-01 3.13603617e-02
7.22455025e-01 -1.79548413e-01 -4.18920368e-02 4.58946943e-01
7.83566833e-02 -1.44292712e-01 -8.52122456e-02 -2.28906229e-01
1.16434860e+00 1.61853313e-01 -1.67703718e-01 3.02890558e-02
3.79123926e-01 -6.59829378e-01 6.36802375e-01 8.19566131e-01
5.96324876e-02 1.03687000e+00 1.65959179e-01 -1.77225500e-01
-9.63406801e-01 -1.11859846e+00 1.57848552e-01 1.49976587e+00
-2.02382684e-01 -6.62725270e-01 -8.68876517e-01 -2.09367767e-01
-2.39879310e-01 4.14330870e-01 -2.50154316e-01 3.59644666e-02
-8.97039056e-01 -4.91830647e-01 1.12504768e+00 7.44059920e-01
5.95923662e-01 -1.34357846e+00 -7.80958295e-01 4.20392722e-01
-3.26255053e-01 -6.78065658e-01 -6.39244020e-01 7.15131938e-01
-9.35558140e-01 -5.03981411e-01 -1.07993579e+00 -9.04514790e-01
2.42948279e-01 4.53301787e-01 9.58779395e-01 -2.14675903e-01
1.80697944e-02 -4.15391266e-01 -5.61259627e-01 -4.89452749e-01
-1.59902975e-01 2.96045899e-01 1.05914228e-01 4.84674051e-02
-2.11835593e-01 -9.22182441e-01 -5.93584418e-01 1.77602321e-01
-9.99856234e-01 1.65126666e-01 6.22838914e-01 5.95588028e-01
6.28649652e-01 4.52892929e-01 7.81888306e-01 -4.38875377e-01
8.57823133e-01 3.43461260e-02 -1.91886887e-01 -2.93317977e-02
4.95917164e-02 1.22931525e-02 8.34007561e-01 -6.83049083e-01
-9.70517874e-01 -1.90858468e-01 -2.84129083e-01 -6.10050857e-01
-8.69660601e-02 1.54695630e-01 7.50367641e-02 1.38887048e-01
6.41885996e-01 1.62525885e-02 -4.39645797e-01 -9.93615568e-01
2.54355818e-01 7.93850064e-01 5.63995183e-01 -1.92868397e-01
5.12186468e-01 9.15452689e-02 -3.03300232e-01 -8.97446513e-01
-6.21171355e-01 -4.70901430e-01 -6.21433020e-01 1.21601941e-02
7.92323709e-01 -8.64417017e-01 -4.73538369e-01 4.43456560e-01
-1.17869270e+00 -4.61423576e-01 -5.83972335e-01 6.16539299e-01
-1.88878253e-01 -5.64713255e-02 -7.96335459e-01 -9.22112226e-01
-4.47155744e-01 -9.34549987e-01 1.09679413e+00 4.01434958e-01
-3.11013877e-01 -5.52860856e-01 1.28150716e-01 -2.05223873e-01
6.07813358e-01 -1.09213360e-01 8.04212928e-01 -4.15685862e-01
-3.41436267e-01 1.23545997e-01 1.19267859e-01 7.48555124e-01
2.53098100e-01 -1.19871691e-01 -1.52468812e+00 -8.85871276e-02
-3.28771845e-02 1.48124576e-01 1.40612078e+00 4.82803166e-01
1.21893036e+00 -2.80662209e-01 1.62024628e-02 6.59687579e-01
9.78260458e-01 1.79351375e-01 8.26432168e-01 -1.12479873e-01
6.85596406e-01 2.58699179e-01 6.41947836e-02 3.29370916e-01
-3.93134087e-01 5.42472482e-01 2.16459736e-01 -2.31172681e-01
-6.12551153e-01 -1.39603123e-01 7.00971544e-01 1.10827315e+00
-3.87172312e-01 -1.18221804e-01 -3.51628929e-01 5.11912525e-01
-1.85277534e+00 -1.18894815e+00 -8.52964073e-02 2.42297554e+00
8.01753283e-01 -2.33248714e-02 1.76548541e-01 7.78120637e-01
7.99610496e-01 3.99518102e-01 -1.93135709e-01 -7.21704066e-01
-2.19143197e-01 8.93554389e-01 2.44764775e-01 4.92002934e-01
-9.88772750e-01 8.28288913e-01 6.24495268e+00 1.06920254e+00
-1.28016853e+00 8.07681009e-02 1.62634417e-01 -6.02371812e-01
-2.32935905e-01 -1.12074904e-01 -6.21961594e-01 1.02402255e-01
1.04327297e+00 2.99067050e-01 7.61675358e-01 4.20631766e-01
1.95260778e-01 1.57036170e-01 -1.07661581e+00 9.94481206e-01
-2.34807476e-01 -1.30222452e+00 7.16404170e-02 -1.52586937e-01
8.11672986e-01 -1.21516153e-01 2.21042469e-01 2.24375248e-01
-1.03227101e-01 -1.10415077e+00 8.14847529e-01 5.26610732e-01
5.21088362e-01 -1.06466043e+00 4.88744110e-01 3.43913794e-01
-1.64133072e+00 -1.91210479e-01 -5.73618829e-01 -3.18039060e-01
-1.16015755e-01 6.62152767e-01 -8.78735840e-01 6.09508932e-01
7.99169183e-01 3.77856046e-01 -2.59659439e-01 1.12654281e+00
-1.07408546e-01 6.81850195e-01 -1.71573833e-01 3.12360674e-02
4.26148951e-01 7.80110955e-02 7.46615529e-01 1.41571236e+00
6.81417942e-01 -1.92297176e-01 -1.10573553e-01 8.50467443e-01
-1.53065637e-01 8.58017430e-02 -2.68040359e-01 -2.78385933e-02
3.25673431e-01 1.23406470e+00 -6.12640083e-01 -2.03075275e-01
-2.44375214e-01 1.28003168e+00 1.46056697e-01 5.52291393e-01
-7.89088368e-01 -9.05104637e-01 1.01789618e+00 8.17876011e-02
6.38353288e-01 -2.32311204e-01 -2.83114761e-01 -8.85796487e-01
-1.10594705e-01 -6.84727371e-01 2.11827442e-01 -8.12596917e-01
-9.49084163e-01 1.02018940e+00 -4.61341649e-01 -1.35021222e+00
-1.81525081e-01 -4.71260935e-01 -7.37070203e-01 1.33658147e+00
-1.40274310e+00 -8.51822436e-01 -6.04638942e-02 7.31184006e-01
6.88236237e-01 8.80542099e-02 1.16308153e+00 3.98334801e-01
-4.59079385e-01 4.61654037e-01 8.02025869e-02 2.41383135e-01
9.08482909e-01 -1.32164133e+00 4.23728555e-01 8.59310031e-01
7.27649510e-01 7.28771389e-01 4.41894859e-01 -5.35688519e-01
-7.82384932e-01 -1.22245264e+00 8.66411269e-01 2.20764764e-02
3.58354419e-01 -4.89012808e-01 -9.90476429e-01 4.52011675e-01
2.81255692e-01 -7.81278387e-02 8.53470325e-01 2.63903767e-01
-4.18170810e-01 -3.25493008e-01 -7.19451249e-01 6.87142074e-01
9.40081298e-01 -6.35344863e-01 -6.20448709e-01 -2.83046871e-01
8.00840378e-01 -3.97939146e-01 -4.48640943e-01 3.88959825e-01
6.46935761e-01 -1.27108562e+00 1.07779431e+00 -4.39954519e-01
3.48262370e-01 -5.54127693e-01 -1.99466795e-01 -1.49497187e+00
-6.22731566e-01 -1.00144768e+00 -1.49356097e-01 1.31438482e+00
6.54528260e-01 -4.01290953e-01 4.00615931e-01 -1.11747324e-01
-3.88356328e-01 -6.02087617e-01 -8.41582775e-01 -6.13618433e-01
-5.53717874e-02 -7.94077277e-01 8.04072261e-01 6.46968961e-01
8.19631666e-02 3.42081726e-01 -4.82464224e-01 4.53538299e-01
2.44403794e-01 2.95068562e-01 4.26504970e-01 -1.12987936e+00
-5.71324229e-01 -6.93564355e-01 -1.59502015e-01 -1.06058025e+00
-1.58594713e-01 -8.95785570e-01 1.57385066e-01 -1.53456497e+00
-1.02608874e-01 -5.51890075e-01 -8.73265684e-01 5.30726016e-01
-1.94054410e-01 6.46362901e-01 3.70154709e-01 1.79735497e-01
-1.25329718e-01 1.75524876e-01 1.36645830e+00 2.35162824e-02
-8.48442435e-01 3.36907327e-01 -7.15762675e-01 8.40946496e-01
9.04618979e-01 -2.99149692e-01 -4.50146228e-01 -5.44139445e-01
1.46398425e-01 2.83356793e-02 3.99309039e-01 -1.40823472e+00
2.72588819e-01 3.49217534e-01 5.86120129e-01 -7.17887998e-01
5.38761854e-01 -4.90798086e-01 2.02006325e-01 6.21257760e-02
-6.20833635e-01 -2.05549866e-01 3.85668665e-01 2.01794013e-01
-4.43189174e-01 -2.71702766e-01 5.83884954e-01 -1.48267031e-01
-3.49889964e-01 4.49381918e-02 -2.86895871e-01 -4.39078152e-01
4.91943747e-01 -1.96059972e-01 8.81881714e-02 -3.65597278e-01
-1.07296264e+00 -2.25720093e-01 -1.37358293e-01 4.75874126e-01
7.34038055e-01 -1.34053624e+00 -8.25061381e-01 2.43085161e-01
-4.25973773e-01 1.28444836e-01 5.63150525e-01 6.00675642e-01
-2.93829441e-01 2.68092126e-01 -7.25191757e-02 -3.41244608e-01
-1.45044506e+00 3.31330806e-01 3.97989303e-01 -1.78038999e-01
-8.21502209e-01 1.33375096e+00 3.80796283e-01 -1.82205170e-01
4.91133213e-01 -9.84115422e-01 -2.38064408e-01 3.22368145e-01
1.00847447e+00 4.86851275e-01 1.50842264e-01 -4.83060241e-01
-1.61585212e-01 5.66301823e-01 1.80850908e-01 -3.32810253e-01
1.42098403e+00 -5.96075188e-05 -1.62426248e-01 7.56674528e-01
8.42231214e-01 6.05008066e-01 -1.09341943e+00 -1.09796412e-01
-4.04796004e-01 -3.72261196e-01 2.78543323e-01 -6.79190755e-01
-1.10844755e+00 1.13703525e+00 3.26661557e-01 4.87628311e-01
1.49739468e+00 -1.47154778e-01 9.82044756e-01 7.02296197e-02
9.11495835e-02 -9.03854668e-01 1.62427664e-01 6.30608976e-01
9.56195712e-01 -4.05142426e-01 -5.26137412e-01 -3.67775679e-01
-4.63434309e-01 1.35818648e+00 2.99542904e-01 -2.34029993e-01
4.74549681e-01 7.53309309e-01 4.47887853e-02 3.52511436e-01
-5.26516855e-01 -4.53402519e-01 4.49049562e-01 3.34843338e-01
8.14091444e-01 1.75520882e-01 2.92071998e-01 9.30947781e-01
-5.56539416e-01 -2.56675810e-01 2.55572110e-01 7.05526650e-01
-6.57072365e-01 -1.12271762e+00 -7.04594553e-01 2.96022236e-01
-7.67017066e-01 -5.32275081e-01 -3.90069932e-01 1.76627398e-01
4.52125639e-01 1.14168954e+00 2.96658009e-01 -6.60604656e-01
3.38506371e-01 3.13292712e-01 5.24863720e-01 -7.08508015e-01
-1.08753705e+00 7.88914919e-01 -4.25727107e-02 -3.57157499e-01
-3.16211611e-01 -4.49636579e-01 -1.46898437e+00 -1.72798157e-01
-3.66977572e-01 2.60183752e-01 3.81236106e-01 5.82900465e-01
2.69395590e-01 1.43887627e+00 6.26435578e-01 -1.08405983e+00
-3.62442434e-01 -1.14179766e+00 -8.56015623e-01 -1.06827669e-01
7.69125938e-01 -3.70416194e-02 -2.74209768e-01 4.57221568e-02] | [15.428714752197266, 5.569364547729492] |
ad6406cf-2bb1-4326-bd17-1bf825fa79fd | instance-optimal-cluster-recovery-in-the | 2306.12968 | null | https://arxiv.org/abs/2306.12968v1 | https://arxiv.org/pdf/2306.12968v1.pdf | Instance-Optimal Cluster Recovery in the Labeled Stochastic Block Model | We consider the problem of recovering hidden communities in the Labeled Stochastic Block Model (LSBM) with a finite number of clusters, where cluster sizes grow linearly with the total number $n$ of items. In the LSBM, a label is (independently) observed for each pair of items. Our objective is to devise an efficient algorithm that recovers clusters using the observed labels. To this end, we revisit instance-specific lower bounds on the expected number of misclassified items satisfied by any clustering algorithm. We present Instance-Adaptive Clustering (IAC), the first algorithm whose performance matches these lower bounds both in expectation and with high probability. IAC consists of a one-time spectral clustering algorithm followed by an iterative likelihood-based cluster assignment improvement. This approach is based on the instance-specific lower bound and does not require any model parameters, including the number of clusters. By performing the spectral clustering only once, IAC maintains an overall computational complexity of $\mathcal{O}(n \text{polylog}(n))$. We illustrate the effectiveness of our approach through numerical experiments. | ['Se-Young Yun', 'Alexandre Proutiere', 'Kaito Ariu'] | 2023-06-18 | null | null | null | null | ['stochastic-block-model', 'clustering'] | ['graphs', 'methodology'] | [ 1.92689836e-01 -3.82084697e-02 -3.84278819e-02 -3.16846877e-01
-1.04472005e+00 -6.79682136e-01 1.30048092e-03 4.49132919e-01
-5.05625248e-01 4.73790854e-01 -5.57234526e-01 -4.60152715e-01
-4.28445697e-01 -6.72467947e-01 -9.06954467e-01 -1.10199571e+00
-4.70466971e-01 1.14026272e+00 3.36412042e-01 5.31028330e-01
2.19818607e-01 2.30150595e-01 -1.29023826e+00 1.41642973e-01
8.40332031e-01 6.70122921e-01 3.17148685e-01 8.28197777e-01
-3.40231806e-02 6.22433543e-01 -3.48152518e-01 -2.49374419e-01
3.27786118e-01 -7.47801304e-01 -9.48465943e-01 6.53822422e-01
-2.52782851e-01 2.94869274e-01 1.73003003e-01 1.36316848e+00
1.76554874e-01 -6.34276047e-02 8.75302076e-01 -1.51513124e+00
-1.89872310e-01 1.16802597e+00 -9.98327196e-01 1.13346994e-01
-7.62612596e-02 -3.24690759e-01 1.07769132e+00 -6.71040654e-01
4.43398356e-01 1.02219450e+00 4.92969036e-01 2.29897961e-01
-1.69632971e+00 -8.00997436e-01 6.74343854e-02 2.02485293e-01
-1.87628257e+00 -7.31710196e-02 4.88460064e-01 -5.09825587e-01
4.66743916e-01 3.57280225e-01 2.33869016e-01 1.77827850e-01
-7.79048979e-01 7.61533618e-01 1.04456866e+00 -8.95967185e-01
6.57347679e-01 2.20584616e-01 2.96507239e-01 6.23618782e-01
8.59048843e-01 -3.42904508e-01 -1.52218118e-01 -4.42585945e-01
2.58752167e-01 2.31396720e-01 -1.66678667e-01 -7.06032276e-01
-1.01682580e+00 8.88913810e-01 3.69019732e-02 1.93722799e-01
-7.79500976e-02 5.25678575e-01 1.77486632e-02 2.07047656e-01
3.84377152e-01 -1.37300819e-01 -4.33794707e-01 7.82445446e-02
-1.22904134e+00 -1.71127871e-01 8.19652855e-01 1.35788155e+00
1.02030110e+00 -4.51327831e-01 3.40440243e-01 7.23418474e-01
2.86731273e-01 9.96022105e-01 -1.57036871e-01 -1.02858436e+00
5.60417771e-01 4.39070404e-01 4.20243800e-01 -9.01958823e-01
-2.88616300e-01 -5.15887976e-01 -8.32348645e-01 -2.01961517e-01
5.40649235e-01 -1.22843266e-01 -5.98665357e-01 1.91738188e+00
3.52224410e-01 2.10919008e-01 -2.12808847e-01 5.53295255e-01
-4.18943137e-01 6.92937791e-01 -1.51627302e-01 -6.04645312e-01
1.09124577e+00 -5.95963955e-01 -5.24973691e-01 1.03653520e-01
9.49490428e-01 -5.83440721e-01 5.87257802e-01 4.40462053e-01
-1.21292388e+00 -1.26046568e-01 -9.14233625e-01 4.82552886e-01
3.61876301e-02 2.11321213e-03 2.35713348e-01 1.06623638e+00
-1.05491126e+00 3.30275506e-01 -1.02049279e+00 -2.38125980e-01
1.17019638e-01 6.16499722e-01 1.71685498e-02 -2.03957543e-01
-4.75170493e-01 1.61463425e-01 5.58815420e-01 3.22240479e-02
-6.32587433e-01 -2.78663427e-01 -5.55970967e-01 3.33158791e-01
6.22678220e-01 -2.88092524e-01 9.47516918e-01 -8.60275626e-01
-7.35883832e-01 7.77586699e-01 -6.00919187e-01 -5.36651015e-01
3.66351664e-01 3.81565183e-01 -8.12913552e-02 4.53499615e-01
2.62762964e-01 2.41883367e-01 5.27418733e-01 -1.68969965e+00
-9.79878187e-01 -4.43889499e-01 -2.67148316e-01 -1.59544662e-01
-2.26263136e-01 -3.65487672e-02 -6.86723590e-01 -2.62028784e-01
4.96447593e-01 -1.34803116e+00 -5.26953697e-01 -4.60711807e-01
-3.73681635e-01 -2.37739027e-01 2.34012365e-01 -2.33254850e-01
1.24788737e+00 -2.31980491e+00 1.34819970e-02 8.93724322e-01
2.35038415e-01 -1.29188642e-01 1.81822687e-01 5.05704224e-01
3.19794454e-02 3.62180412e-01 -5.56758106e-01 -7.27075100e-01
9.35271531e-02 8.00968558e-02 2.91311815e-02 8.79878402e-01
-3.85862380e-01 4.02258098e-01 -8.61877263e-01 -7.26973772e-01
-7.01809078e-02 -1.60993621e-01 -7.68547714e-01 -1.35345101e-01
-1.12777911e-01 -5.71144931e-03 -7.67805278e-02 3.03796858e-01
1.05893242e+00 -9.17142510e-01 9.04745519e-01 2.50407040e-01
1.29306056e-02 -1.18705556e-01 -1.82833540e+00 1.08744013e+00
-1.97071433e-01 1.04027078e-01 3.85175854e-01 -1.25345051e+00
4.79247481e-01 3.60895425e-01 5.14292657e-01 -7.44382339e-03
-7.37618357e-02 5.66331089e-01 -2.11636350e-01 -1.48975685e-01
3.36916149e-02 -4.35341865e-01 -2.86452949e-01 9.77931738e-01
-1.99271619e-01 7.35004485e-01 3.61567765e-01 5.64774156e-01
1.15268552e+00 -6.70450509e-01 1.95112154e-01 -3.89571667e-01
4.23186123e-01 8.29743370e-02 7.62541771e-01 1.13765359e+00
-6.38167337e-02 3.40996146e-01 5.26719213e-01 4.81420681e-02
-1.16554403e+00 -9.06981587e-01 1.11065082e-01 8.58151495e-01
3.29115480e-01 -2.47122347e-01 -1.08068573e+00 -3.66542608e-01
-6.56308383e-02 4.02243197e-01 -6.95126235e-01 2.54636526e-01
-3.59685898e-01 -9.53688979e-01 1.63810700e-01 4.09443706e-01
1.18859380e-01 -8.02857935e-01 -4.14265305e-01 2.42325917e-01
-3.04267108e-01 -1.00773501e+00 -5.72911203e-01 5.91968238e-01
-8.81623983e-01 -1.11938643e+00 -4.23571318e-01 -1.00703895e+00
1.12579942e+00 6.67373717e-01 7.61105061e-01 3.48985016e-01
-7.87138194e-02 1.94939837e-01 -5.39122880e-01 -6.24321327e-02
-2.12613374e-01 2.63389230e-01 5.35942204e-02 2.04916894e-01
6.11205459e-01 -4.88588333e-01 -4.73841697e-01 3.85163754e-01
-9.97635841e-01 -9.25653800e-02 6.04537010e-01 6.01883948e-01
6.81695342e-01 5.99215090e-01 3.78593475e-01 -1.26605582e+00
-9.82791837e-03 -7.62610555e-01 -1.01267219e+00 3.81242126e-01
-6.67790413e-01 3.45569514e-02 6.25723958e-01 -2.10724041e-01
-7.08840907e-01 5.54240108e-01 4.49029863e-01 -3.00663471e-01
-5.65744489e-02 4.93402898e-01 -2.51202315e-01 2.87341148e-01
2.80491829e-01 3.24228704e-01 -1.09575517e-01 -5.78356981e-01
2.66541153e-01 8.20193172e-01 5.57518363e-01 -5.36349893e-01
8.93807650e-01 8.09531868e-01 1.61550686e-01 -6.04399025e-01
-8.29182088e-01 -1.03969932e+00 -8.11061382e-01 -9.52522904e-02
4.89264220e-01 -8.55339527e-01 -1.18127680e+00 5.98727837e-02
-6.78916156e-01 -5.37329912e-01 7.38695487e-02 4.87832993e-01
-5.60255766e-01 7.00546980e-01 -4.99406040e-01 -1.35248053e+00
7.18475431e-02 -8.60881269e-01 6.35591030e-01 -3.07011068e-01
1.60121560e-01 -8.25059235e-01 -1.38981491e-01 4.20159101e-01
-3.00606489e-01 2.70186067e-02 1.06304240e+00 -7.98522294e-01
-8.43588412e-01 -3.75948071e-01 -2.59035021e-01 6.25100583e-02
-6.80961758e-02 -3.28800291e-01 -4.73314047e-01 -6.20759726e-01
-2.71221381e-02 1.03968248e-01 8.22784364e-01 3.33168209e-01
1.31284845e+00 -4.61468667e-01 -6.76188886e-01 3.73913884e-01
1.77594924e+00 3.56745154e-01 2.79091418e-01 2.34321821e-02
5.45637190e-01 5.23638368e-01 6.54935777e-01 7.26238310e-01
2.12164044e-01 5.51256180e-01 3.14618886e-01 1.52878864e-02
4.16803181e-01 -2.91879289e-02 1.36672884e-01 9.98946488e-01
2.48761848e-01 -5.09433746e-01 -9.35470641e-01 1.02062082e+00
-2.00174975e+00 -1.06395149e+00 -7.15953290e-01 2.69245696e+00
9.31624711e-01 7.52621964e-02 4.68592644e-01 6.43372774e-01
1.26928008e+00 -5.98022938e-01 -3.11079800e-01 -1.08600460e-01
1.15882671e-02 1.46437064e-01 1.02928388e+00 7.74775028e-01
-8.90890241e-01 7.00845718e-01 5.74498224e+00 1.00982702e+00
-3.55343431e-01 2.85223007e-01 6.15304649e-01 -2.64570653e-01
-2.32459813e-01 2.65938044e-01 -7.96611071e-01 7.47872293e-01
1.27997816e+00 -1.44537061e-01 6.11889124e-01 8.66723001e-01
1.67192817e-01 -3.14036459e-01 -1.07373762e+00 8.78303230e-01
2.93316180e-03 -1.12601948e+00 -3.64831924e-01 5.12764752e-01
1.00722301e+00 -4.29966539e-01 -5.48615307e-02 -7.59338439e-02
7.84018099e-01 -6.33847117e-01 7.22117126e-01 -6.04295731e-03
6.85435355e-01 -1.19123924e+00 7.42500365e-01 8.10845315e-01
-1.31567943e+00 -3.05598259e-01 -5.27370036e-01 2.19588310e-01
1.76865414e-01 8.42991233e-01 -8.20232332e-01 3.76989663e-01
8.21490049e-01 1.48983657e-01 -2.28773654e-01 1.20680106e+00
1.23613030e-01 1.00539601e+00 -7.90765345e-01 -7.48778358e-02
1.95299894e-01 -3.57539117e-01 1.67386264e-01 1.44931901e+00
3.07501465e-01 4.62800503e-01 4.52190727e-01 5.09381533e-01
-1.97066233e-01 1.20547310e-01 -2.26571001e-02 2.39243329e-01
1.09488118e+00 9.80262935e-01 -1.42211664e+00 -6.34342730e-01
-1.83684036e-01 9.32718754e-01 3.13042790e-01 1.28841087e-01
-8.59056294e-01 -3.18809718e-01 6.35833666e-02 2.82188565e-01
8.79773259e-01 -2.55531579e-01 -3.57357085e-01 -6.66549325e-01
4.29955032e-03 -4.81697768e-01 5.81637383e-01 -2.12942943e-01
-1.23029232e+00 1.19942129e-01 1.23764478e-01 -1.11013389e+00
-1.39047474e-01 -2.09205449e-01 -1.48139093e-02 6.33691788e-01
-9.94850755e-01 -6.49711192e-01 -8.46820101e-02 6.34061098e-01
7.60877654e-02 3.49279910e-01 5.86815238e-01 4.00817662e-01
-5.85974872e-01 6.59861445e-01 7.03855932e-01 4.86084931e-02
4.13885266e-01 -1.54136217e+00 -6.95701689e-02 1.02521706e+00
-2.49954425e-02 5.70991099e-01 7.62341261e-01 -5.62678933e-01
-8.74789894e-01 -1.18511724e+00 1.18657374e+00 -1.97737575e-01
4.93279636e-01 -6.59881055e-01 -7.15419412e-01 7.42601037e-01
-1.64384738e-01 -7.58806535e-04 1.06684744e+00 2.30985418e-01
-2.65595466e-01 -1.86511293e-01 -1.11638319e+00 2.51682997e-01
1.00301325e+00 -3.45679790e-01 4.11483273e-03 4.64910865e-01
4.73085850e-01 2.17874095e-01 -6.98374689e-01 9.72031653e-02
6.87616765e-02 -7.58148193e-01 6.38226688e-01 -3.42871636e-01
-1.19820327e-01 -6.63864195e-01 -2.97501773e-01 -5.99568903e-01
-5.15913188e-01 -6.43919230e-01 -4.31939103e-02 1.11504054e+00
6.42201245e-01 -3.92295063e-01 1.00395763e+00 4.27825451e-01
3.87285203e-01 -4.04150128e-01 -8.78962040e-01 -1.06275177e+00
8.43650382e-03 -4.35139984e-01 4.35932785e-01 9.20671523e-01
1.80747613e-01 1.83731154e-01 -4.98245597e-01 5.81929326e-01
1.13681662e+00 4.06479597e-01 6.68697834e-01 -1.35289073e+00
-5.59752464e-01 -9.50767472e-03 -4.58303019e-02 -9.62546647e-01
-5.22417910e-02 -9.35924351e-01 2.58293092e-01 -1.14651871e+00
9.50610757e-01 -1.05612481e+00 -2.89559186e-01 4.35358614e-01
-2.59089351e-01 1.42328769e-01 3.77847195e-01 5.58006346e-01
-1.31587625e+00 -2.23981123e-02 2.70136058e-01 1.76008180e-01
-1.34150550e-01 2.95343876e-01 -5.57565629e-01 6.13467455e-01
8.40540588e-01 -1.12354350e+00 -2.71132618e-01 -1.55889586e-01
4.93675858e-01 3.49821687e-01 3.74554805e-02 -9.18300569e-01
5.64567327e-01 2.39558071e-02 -3.69555689e-02 -1.06434464e+00
-5.34672514e-02 -1.11720347e+00 5.09907544e-01 7.98220396e-01
-5.45018554e-01 1.49252303e-02 -3.73087078e-01 1.15944266e+00
1.72144830e-01 -7.17469215e-01 1.07902646e+00 -6.74533546e-02
-3.72285210e-02 1.31286442e-01 -6.77147806e-01 3.40056000e-03
1.26462173e+00 -3.01015049e-01 1.83085740e-01 -3.87350351e-01
-8.61799777e-01 3.54129821e-01 6.40369117e-01 -4.85931456e-01
1.55261740e-01 -1.06484365e+00 -5.84509611e-01 -1.13485493e-01
1.27816889e-02 1.60097361e-01 2.45001540e-01 9.56456006e-01
-4.62486327e-01 6.07516766e-01 5.43895364e-01 -7.70706117e-01
-1.47574508e+00 1.14921725e+00 -1.09330798e-02 -4.27819818e-01
-2.29252115e-01 8.73389423e-01 4.92572226e-02 -1.65548280e-01
1.96307644e-01 1.43755272e-01 4.03728902e-01 -1.47031337e-01
3.73823285e-01 6.98481083e-01 -8.06215256e-02 -3.03955376e-01
-3.21430653e-01 2.47017533e-01 -1.65339202e-01 -4.53469217e-01
1.25280702e+00 -4.45287973e-01 -3.69727910e-01 5.93742192e-01
1.25891125e+00 7.51532167e-02 -1.16568840e+00 -4.89857644e-01
5.99666893e-01 -3.22570860e-01 -4.15923029e-01 -4.19267565e-01
-1.13506544e+00 6.96561038e-01 4.95117694e-01 5.40684760e-01
1.11969435e+00 2.31841132e-01 5.52567363e-01 3.72589618e-01
7.57670820e-01 -1.10520136e+00 1.37524428e-02 6.96900859e-02
-8.62976536e-03 -7.81698644e-01 -5.68701215e-02 -6.34532332e-01
-3.52115244e-01 5.42958021e-01 1.46899223e-01 -1.71195604e-02
8.15681279e-01 1.84567764e-01 -4.05989081e-01 -4.11124453e-02
-7.96888471e-01 -3.66855115e-01 -3.75668228e-01 2.60915518e-01
1.38194475e-03 4.08813328e-01 -5.06561339e-01 5.10982096e-01
1.05303720e-01 -2.29218170e-01 5.27440548e-01 8.89557481e-01
-8.33224416e-01 -1.24771667e+00 -5.97284436e-01 3.83389086e-01
-7.00403571e-01 -1.14044659e-01 -1.78125441e-01 5.31443596e-01
2.12144285e-01 1.31158257e+00 1.52390614e-01 -5.85423261e-02
-1.50152400e-01 4.91662435e-02 1.96248204e-01 -6.85125828e-01
-1.70215905e-01 3.83192152e-01 -3.48393112e-01 1.34011256e-02
-5.40788174e-01 -8.59021008e-01 -1.38145578e+00 -6.33605838e-01
-6.02385402e-01 7.28276253e-01 5.46972036e-01 8.30536425e-01
2.29456484e-01 -1.44158617e-01 1.16172516e+00 -2.73541689e-01
-5.37875056e-01 -8.15624177e-01 -9.93541539e-01 4.61098284e-01
-3.87839191e-02 -4.33374763e-01 -7.45130837e-01 4.01446700e-01] | [6.840945720672607, 5.0797810554504395] |
60860ca0-372c-4eaa-a575-211d9f8e405a | multi-scale-attention-flow-for-probabilistic | 2205.07493 | null | https://arxiv.org/abs/2205.07493v2 | https://arxiv.org/pdf/2205.07493v2.pdf | Multi-scale Attention Flow for Probabilistic Time Series Forecasting | The probability prediction of multivariate time series is a notoriously challenging but practical task. On the one hand, the challenge is how to effectively capture the cross-series correlations between interacting time series, to achieve accurate distribution modeling. On the other hand, we should consider how to capture the contextual information within time series more accurately to model multivariate temporal dynamics of time series. In this work, we proposed a novel non-autoregressive deep learning model, called Multi-scale Attention Normalizing Flow(MANF), where we integrate multi-scale attention and relative position information and the multivariate data distribution is represented by the conditioned normalizing flow. Additionally, compared with autoregressive modeling methods, our model avoids the influence of cumulative error and does not increase the time complexity. Extensive experiments demonstrate that our model achieves state-of-the-art performance on many popular multivariate datasets. | ['Peilin Zhao', 'Fan Lin', 'Pengcheng Wu', 'Jiaxiang Wu', 'Ke Xu', 'Shibo Feng'] | 2022-05-16 | null | null | null | null | ['probabilistic-time-series-forecasting'] | ['time-series'] | [-8.48878846e-02 -5.71856260e-01 1.50277346e-01 -3.66468370e-01
-6.07831419e-01 -3.29823643e-01 4.92197365e-01 2.10996091e-01
-2.02340811e-01 5.35651684e-01 3.45564067e-01 -2.84854859e-01
-3.58436942e-01 -7.31695235e-01 -7.88454711e-01 -7.43625462e-01
-4.59928572e-01 3.10197562e-01 -6.67314157e-02 -1.30617442e-02
-1.00099720e-01 5.16291559e-01 -1.21286821e+00 -2.23027751e-01
9.41983938e-01 1.11216044e+00 1.65833339e-01 7.92594969e-01
1.64603640e-03 1.09647489e+00 -6.53084397e-01 -2.19052523e-01
-1.58898994e-01 -2.77051300e-01 -3.50474641e-02 -2.19019175e-01
-3.88011634e-02 -4.95126456e-01 -6.04106069e-01 9.94243443e-01
2.99263299e-01 3.98312390e-01 9.34371233e-01 -1.32577908e+00
-6.16053283e-01 6.48927033e-01 -8.06388617e-01 5.80588460e-01
-1.71197966e-01 -4.29203957e-02 8.49851370e-01 -3.37138355e-01
4.97592241e-03 1.30407310e+00 4.75625366e-01 -3.67113166e-02
-1.12902641e+00 -7.29266047e-01 6.09024584e-01 3.38037819e-01
-1.27990675e+00 1.90297961e-01 1.00610602e+00 -4.97191399e-01
9.22739625e-01 1.45535126e-01 3.90643001e-01 1.06450379e+00
6.41125083e-01 7.58248746e-01 5.96209466e-01 4.97345738e-02
2.87367046e-01 -6.25320613e-01 1.66084349e-01 6.90611899e-02
-1.14289001e-01 2.25807889e-03 -9.14125293e-02 -2.05751866e-01
9.35464323e-01 7.22193658e-01 -1.06485002e-01 3.61071143e-04
-1.18746459e+00 6.55751824e-01 2.80765206e-01 4.61778402e-01
-7.81921327e-01 3.62892687e-01 4.12315398e-01 3.35948206e-02
8.29805672e-01 1.12504110e-01 -6.89763665e-01 -6.56673729e-01
-6.90847516e-01 2.66144305e-01 5.57736218e-01 7.00884521e-01
4.32542920e-01 2.46310905e-01 -2.69967139e-01 4.92602676e-01
1.29346862e-01 7.14108944e-01 4.16697055e-01 -7.72051871e-01
5.75590372e-01 2.49035448e-01 6.70480654e-02 -1.25727189e+00
-4.72777933e-01 -6.79299831e-01 -1.31503654e+00 -3.60375762e-01
4.62751567e-01 -2.44436696e-01 -7.34392881e-01 1.93571198e+00
-3.80282365e-02 7.72464097e-01 -2.27903128e-01 6.77678406e-01
1.85929030e-01 9.68828499e-01 3.70101631e-01 -5.85936308e-01
1.06405532e+00 -6.40788257e-01 -1.09667337e+00 3.86796482e-02
1.98258877e-01 -6.37942016e-01 8.72440934e-01 2.70878226e-01
-7.98329532e-01 -5.05795240e-01 -6.88190639e-01 6.10291138e-02
-3.30035478e-01 5.32486327e-02 7.90009260e-01 2.39979979e-02
-4.40392345e-01 8.44709694e-01 -1.38807034e+00 2.29080599e-02
2.47766793e-01 2.21354529e-01 -3.77791114e-02 8.32133666e-02
-1.31868196e+00 4.49780673e-01 1.72537521e-01 1.53114513e-01
-8.75398397e-01 -1.05479884e+00 -8.59342694e-01 5.22630274e-01
2.13140890e-01 -3.56959850e-01 1.19658351e+00 -5.73632121e-01
-1.44255733e+00 -1.30227521e-01 -5.61614752e-01 -4.01032060e-01
4.20995206e-01 -5.72862864e-01 -6.09937131e-01 -1.60011396e-01
-1.09130912e-01 3.56059670e-02 7.65338719e-01 -6.38828516e-01
-4.06301916e-01 -3.16742092e-01 -1.79766774e-01 -9.76171568e-02
-4.92246538e-01 4.03054394e-02 -4.88646388e-01 -1.02602696e+00
-8.01763088e-02 -6.88315511e-01 -4.77781802e-01 -4.27993506e-01
-2.06564859e-01 -4.15438712e-01 8.53907406e-01 -8.34041417e-01
1.72219336e+00 -2.33651710e+00 2.86194772e-01 6.40287474e-02
1.85229748e-01 -1.49131805e-01 -1.86408550e-01 5.43248057e-01
-3.06495637e-01 6.51632771e-02 -2.56482422e-01 -4.95488435e-01
5.81841879e-02 1.74869299e-01 -9.48038459e-01 5.45566440e-01
4.23471630e-01 7.46370554e-01 -1.00487375e+00 -1.32134780e-01
3.40746284e-01 8.07381690e-01 -4.58157539e-01 4.02181596e-01
-3.54959339e-01 6.87422633e-01 -4.70251977e-01 4.93257374e-01
6.27594948e-01 -5.26267886e-01 -5.54698147e-02 -3.09106022e-01
1.03921175e-01 7.19624683e-02 -1.00300872e+00 1.39366686e+00
-3.16326261e-01 5.34126878e-01 -5.89345336e-01 -1.01349354e+00
9.12099838e-01 2.70423234e-01 7.82046199e-01 -6.20189726e-01
2.83079326e-01 8.24235100e-03 1.00091882e-01 -3.83383602e-01
5.49264193e-01 -2.80421488e-02 -2.19883457e-01 3.05185705e-01
-4.54872623e-02 2.06072420e-01 1.15297288e-01 3.29027548e-02
9.02226269e-01 1.15189813e-01 3.68849099e-01 6.24772757e-02
2.76740819e-01 -5.57746291e-01 8.16565394e-01 5.48368752e-01
-1.40298933e-01 4.98844087e-01 8.67994189e-01 -3.90334010e-01
-1.02853429e+00 -1.11316764e+00 1.87410600e-02 1.02469277e+00
-1.57174915e-01 -4.47857529e-01 -3.19399834e-01 -3.15422177e-01
-7.00711785e-03 7.52442837e-01 -7.38233507e-01 -1.79028362e-01
-7.09310949e-01 -1.01668620e+00 1.53833598e-01 9.81948912e-01
4.11281623e-02 -8.15452635e-01 -1.18564323e-01 5.96207440e-01
-1.86413318e-01 -1.00754380e+00 -7.86478758e-01 1.51581625e-02
-8.60099852e-01 -7.28441000e-01 -9.36511636e-01 -1.89833134e-01
3.67930919e-01 1.34852126e-01 1.02884424e+00 -6.11315846e-01
-3.54993120e-02 2.28101894e-01 -3.34587365e-01 -6.80267274e-01
1.27459854e-01 -2.18447438e-03 -2.41458490e-02 2.22375020e-01
5.50552726e-01 -1.12941098e+00 -6.36493027e-01 1.66347489e-01
-9.26660597e-01 -3.70776743e-01 4.24936861e-01 7.24895477e-01
9.28483367e-01 4.28884685e-01 6.64715886e-01 -3.86135221e-01
7.51744509e-01 -8.68405044e-01 -6.21889174e-01 1.38757080e-01
-9.29875895e-02 -1.10452838e-01 1.09705865e+00 -1.00004089e+00
-8.55589151e-01 -2.20392108e-01 1.78701669e-01 -9.85220850e-01
-2.19527539e-02 8.00679564e-01 -1.44602448e-01 6.42134905e-01
7.24487826e-02 4.58176553e-01 -1.05114013e-01 -6.71112478e-01
3.21167380e-01 3.01213145e-01 6.94582582e-01 -6.33764386e-01
5.13008356e-01 5.51287770e-01 2.83085048e-01 -5.84968150e-01
-6.05428278e-01 -3.42026740e-01 -5.05038023e-01 -5.64983785e-02
8.05839598e-01 -9.45815325e-01 -9.20389891e-01 6.34543359e-01
-1.20595682e+00 -4.31297779e-01 1.37730539e-02 8.23506415e-01
-4.99087483e-01 4.25703198e-01 -6.56377673e-01 -1.12016511e+00
-1.64371595e-01 -8.06095719e-01 1.10421872e+00 1.66512489e-01
-1.75289840e-01 -1.22115350e+00 1.24184981e-01 -2.48955965e-01
5.77403784e-01 5.54387689e-01 9.41309690e-01 -6.96936369e-01
-5.44593394e-01 -3.96573365e-01 -2.18141899e-01 1.84173867e-01
1.93668172e-01 3.15274954e-01 -6.25260353e-01 -8.66843164e-02
4.30379584e-02 2.18198583e-01 7.26355612e-01 6.83409452e-01
1.81563735e+00 -5.28121665e-02 -1.60754338e-01 6.51954472e-01
1.01529109e+00 5.25227726e-01 5.42770505e-01 -1.96030829e-02
9.51527476e-01 5.12700677e-01 5.84941328e-01 8.44857991e-01
6.32684648e-01 4.25667554e-01 3.54417562e-01 1.71725228e-01
4.62257594e-01 -4.00551856e-01 1.60154462e-01 1.20082068e+00
-1.63544461e-01 -4.48137701e-01 -8.79256070e-01 5.39935946e-01
-2.08275700e+00 -1.13524556e+00 -1.08292453e-01 1.98570096e+00
4.70780641e-01 5.70669547e-02 1.10629119e-01 4.43207361e-02
5.59142888e-01 3.77495289e-01 -7.69873619e-01 5.16012125e-02
-1.73106715e-01 -8.74090567e-02 3.49447846e-01 2.69498706e-01
-1.43882918e+00 5.29774785e-01 6.04379272e+00 9.25182879e-01
-1.31501544e+00 -6.67327717e-02 7.16309965e-01 -1.95966601e-01
-2.72363454e-01 -3.39719772e-01 -6.20853961e-01 9.12594020e-01
1.11376667e+00 -5.00581443e-01 5.27819753e-01 8.28236282e-01
4.08319980e-01 3.74605477e-01 -1.03159857e+00 9.44277465e-01
-1.30497992e-01 -8.10333014e-01 -6.08739257e-02 3.10565345e-02
5.05590796e-01 4.10001017e-02 2.04690456e-01 6.02040410e-01
3.65314454e-01 -9.31801677e-01 5.39533973e-01 1.02761257e+00
3.29818398e-01 -9.21170115e-01 6.22307479e-01 3.20301950e-01
-1.46025217e+00 -1.33914277e-01 -2.62598872e-01 -1.95153430e-01
5.55826187e-01 7.72856772e-01 -3.01325411e-01 5.31842232e-01
7.52528191e-01 1.24005818e+00 -2.44874597e-01 1.01245308e+00
1.02837935e-01 8.15259397e-01 -3.40632439e-01 -1.00160778e-01
8.15050229e-02 -2.67523915e-01 5.83143890e-01 1.00057459e+00
6.13315940e-01 7.75104165e-02 1.30511135e-01 7.18587279e-01
1.77356228e-01 -6.33556917e-02 -3.32931370e-01 -5.03308654e-01
4.98968393e-01 8.95611644e-01 -4.07623023e-01 -3.48346591e-01
-4.36622590e-01 5.86477816e-01 3.75842273e-01 7.39945710e-01
-1.14740741e+00 -2.36584619e-01 8.40283692e-01 -2.07158700e-01
5.55473864e-01 -5.41367710e-01 -1.51461780e-01 -1.44884348e+00
7.41412118e-02 -6.22624576e-01 5.85990846e-01 -7.36683071e-01
-1.91220522e+00 5.53853214e-01 5.92177622e-02 -1.49478877e+00
-4.41132039e-01 -3.77132356e-01 -7.89400578e-01 1.00224495e+00
-1.45163214e+00 -9.69143391e-01 -3.28063935e-01 5.92268407e-01
3.29122007e-01 -4.24511991e-02 6.35220468e-01 6.15389705e-01
-7.64633238e-01 4.28883553e-01 3.49296540e-01 6.93926364e-02
5.97062230e-01 -1.26996279e+00 3.57895255e-01 8.15154672e-01
-2.27905631e-01 7.95972943e-01 7.97430158e-01 -7.21350610e-01
-1.30105340e+00 -1.43242598e+00 6.74874306e-01 -2.06586033e-01
1.15656042e+00 -2.39491388e-01 -1.40532351e+00 6.56132877e-01
-8.49158764e-02 -6.08317107e-02 8.84822249e-01 1.20138645e-01
-4.34908003e-01 -3.17152947e-01 -5.43897569e-01 6.94434881e-01
6.94543779e-01 -5.85141361e-01 -4.49799210e-01 2.48616517e-01
1.10245919e+00 -2.80559212e-01 -1.16890335e+00 6.26972973e-01
4.63056892e-01 -4.39873576e-01 9.32573795e-01 -7.45561182e-01
6.28333569e-01 -3.97498250e-01 -3.59251440e-01 -1.27814102e+00
-4.95255053e-01 -6.96465969e-01 -6.37325048e-01 1.35426855e+00
8.66885409e-02 -7.94504762e-01 3.05709720e-01 7.55301654e-01
2.42372863e-02 -6.96021140e-01 -8.78989697e-01 -8.63051236e-01
1.25874773e-01 -7.68752992e-01 9.25011277e-01 8.94660294e-01
-2.60583460e-01 -2.33394969e-02 -8.20622087e-01 3.36430371e-01
5.45792758e-01 2.35769525e-01 7.08483398e-01 -1.08291066e+00
-5.79544067e-01 -6.03539467e-01 -4.27961320e-01 -1.21236575e+00
2.79803604e-01 -2.17862055e-01 -1.53679296e-01 -1.13709223e+00
1.51214749e-01 3.79593782e-02 -8.56273472e-01 2.35664230e-02
-5.60044289e-01 -3.51257980e-01 9.22655985e-02 2.95396090e-01
-6.80618107e-01 1.01822567e+00 1.18727505e+00 3.24955164e-03
-2.12997109e-01 6.28683046e-02 -5.65033197e-01 5.55814922e-01
7.18282819e-01 -4.07679260e-01 -4.93219972e-01 -3.49900067e-01
-1.35739455e-02 3.74993682e-01 3.66277397e-01 -6.70715928e-01
3.20867121e-01 -4.27679479e-01 3.56203645e-01 -1.10620546e+00
3.72935116e-01 -8.83717775e-01 1.78251401e-01 9.01629254e-02
-2.35095888e-01 3.97378385e-01 2.88684040e-01 1.12319589e+00
-4.46410686e-01 4.91923392e-01 2.94542819e-01 2.98331648e-01
-5.40780067e-01 7.85402417e-01 -3.22509855e-01 -1.70902297e-01
9.53766584e-01 5.08813143e-01 -3.66719782e-01 -7.47204185e-01
-6.60620451e-01 3.74405622e-01 -8.06596968e-03 6.12704217e-01
3.00526559e-01 -1.42723882e+00 -4.59212154e-01 8.03348571e-02
-1.62639856e-01 7.27256462e-02 8.11462343e-01 9.40271020e-01
-1.57827631e-01 3.85232657e-01 9.89089608e-02 -6.26703799e-01
-8.05466592e-01 9.55825984e-01 3.15135926e-01 -4.75610137e-01
-4.88488078e-01 5.09126365e-01 4.21696037e-01 -1.93551630e-01
3.63423705e-01 -5.91770947e-01 -3.53230953e-01 7.24129900e-02
9.27853882e-01 4.37532455e-01 -3.47362429e-01 -3.89412642e-01
-2.98616022e-01 5.43141067e-01 -1.31647632e-01 1.43093601e-01
1.59008193e+00 -7.88843557e-02 -2.43409544e-01 9.21557367e-01
1.30524373e+00 -2.13465005e-01 -1.59222019e+00 -2.93067276e-01
-1.98385358e-01 -3.69988292e-01 -5.75608052e-02 -4.76869166e-01
-1.02858531e+00 1.10826850e+00 3.05269748e-01 4.87533361e-01
1.42649531e+00 -2.55689323e-01 8.01237941e-01 7.56833032e-02
1.02974668e-01 -6.53851211e-01 5.59611358e-02 7.01987147e-01
9.36863184e-01 -1.10500824e+00 -2.07600236e-01 -2.36736104e-01
-6.39722466e-01 1.13281322e+00 6.05916739e-01 -3.23764801e-01
1.10622644e+00 2.52642244e-01 -1.04049586e-01 1.64999202e-01
-9.57111180e-01 -2.60289479e-02 4.19971079e-01 4.82212186e-01
6.62678599e-01 2.71834791e-01 -7.27189779e-02 1.15145504e+00
-1.36070609e-01 -1.16782729e-02 1.77183270e-01 5.11401951e-01
8.49437416e-02 -7.69562483e-01 -2.42615789e-01 4.52774972e-01
-7.39019096e-01 -3.62421875e-03 2.65467346e-01 5.58762133e-01
-2.80237675e-01 8.40615869e-01 5.18045783e-01 -4.16906774e-01
3.98097843e-01 -1.49449170e-01 -2.16492303e-02 -9.65015739e-02
-2.30446041e-01 6.77739084e-01 -3.97235245e-01 -4.44217920e-01
-2.02948466e-01 -7.21727908e-01 -9.53967690e-01 -4.51753706e-01
6.25974610e-02 -6.15743846e-02 4.56202418e-01 8.96963477e-01
7.02623844e-01 1.16911435e+00 6.71295524e-01 -9.04059112e-01
-5.55425286e-01 -1.22456908e+00 -7.08057046e-01 4.44599837e-01
4.42823231e-01 -6.85400665e-01 -3.94215345e-01 -4.70274948e-02] | [6.984477519989014, 3.1182289123535156] |
621646cf-bdd1-4fde-87f2-76f5b3a8ac57 | suggestive-annotation-of-brain-mr-images-with | 2206.01014 | null | https://arxiv.org/abs/2206.01014v1 | https://arxiv.org/pdf/2206.01014v1.pdf | Suggestive Annotation of Brain MR Images with Gradient-guided Sampling | Machine learning has been widely adopted for medical image analysis in recent years given its promising performance in image segmentation and classification tasks. The success of machine learning, in particular supervised learning, depends on the availability of manually annotated datasets. For medical imaging applications, such annotated datasets are not easy to acquire, it takes a substantial amount of time and resource to curate an annotated medical image set. In this paper, we propose an efficient annotation framework for brain MR images that can suggest informative sample images for human experts to annotate. We evaluate the framework on two different brain image analysis tasks, namely brain tumour segmentation and whole brain segmentation. Experiments show that for brain tumour segmentation task on the BraTS 2019 dataset, training a segmentation model with only 7% suggestively annotated image samples can achieve a performance comparable to that of training on the full dataset. For whole brain segmentation on the MALC dataset, training with 42% suggestively annotated image samples can achieve a comparable performance to training on the full dataset. The proposed framework demonstrates a promising way to save manual annotation cost and improve data efficiency in medical imaging applications. | ['Wenjia Bai', 'Yike Guo', 'Elsa Angelini', 'Yuanhan Mo', 'Shuo Wang', 'Chengliang Dai'] | 2022-06-02 | null | null | null | null | ['brain-segmentation'] | ['medical'] | [ 6.42298102e-01 6.01864517e-01 1.67835262e-02 -7.37084985e-01
-1.11267734e+00 -1.37869850e-01 2.39953846e-01 4.80525047e-01
-1.07158458e+00 6.96311712e-01 -3.13068479e-01 -2.29249433e-01
-5.31885214e-02 -4.03669924e-01 -4.46930617e-01 -7.66712844e-01
1.95027310e-02 1.02934587e+00 5.85053027e-01 4.05496091e-01
1.06773309e-01 2.86674798e-01 -1.26658309e+00 3.82155329e-01
7.56924033e-01 9.23102558e-01 7.07561731e-01 5.10298014e-01
-7.25990683e-02 8.52674961e-01 -5.29479504e-01 -9.74617228e-02
1.37332931e-01 -3.35379511e-01 -1.25172853e+00 5.23438096e-01
2.28266530e-02 -1.00194067e-01 2.40853593e-01 1.17673659e+00
5.95403492e-01 -1.68736115e-01 8.08389187e-01 -9.21104908e-01
1.18872352e-01 7.66642570e-01 -4.99020875e-01 3.50290626e-01
-1.94087178e-01 1.38289027e-03 6.79025531e-01 -7.28194356e-01
7.79715121e-01 5.91795504e-01 3.56564969e-01 4.35621858e-01
-9.44297493e-01 -4.78520244e-01 -2.25070626e-01 2.42613301e-01
-1.35445702e+00 -1.94424585e-01 4.80307490e-01 -7.26534605e-01
4.30293143e-01 3.60517502e-01 6.59874022e-01 5.00164390e-01
2.31916718e-02 1.07082915e+00 1.28287232e+00 -4.93287832e-01
3.38694513e-01 1.67531833e-01 4.06893790e-01 7.20768034e-01
6.75207078e-02 -3.09430361e-01 4.24544588e-02 5.05817309e-03
5.96266031e-01 -2.44141385e-01 -3.49479616e-01 -3.03935319e-01
-1.26887202e+00 8.57052565e-01 4.50788647e-01 5.63211083e-01
-5.66774845e-01 -1.74983114e-01 6.37231290e-01 -7.20909834e-02
6.81964457e-01 5.05931318e-01 -4.11493152e-01 1.40833586e-01
-1.29934049e+00 -1.23367563e-01 4.89591420e-01 6.05691612e-01
5.17658770e-01 -2.87665516e-01 -1.18849985e-01 1.21160340e+00
1.94226727e-02 1.99638247e-01 7.82860518e-01 -4.99416083e-01
6.06334321e-02 6.76553607e-01 -1.71532482e-01 -4.20961648e-01
-9.47194636e-01 -4.44908500e-01 -9.55569863e-01 1.20912507e-01
6.87424123e-01 -1.06831186e-01 -1.09526491e+00 1.05428469e+00
3.24526966e-01 5.33106364e-02 -8.53050351e-02 1.04460716e+00
7.57686079e-01 3.54477763e-01 2.40228817e-01 -3.73578310e-01
1.43351281e+00 -9.84615564e-01 -4.98302817e-01 -1.87671050e-01
9.67965424e-01 -8.45899045e-01 1.03109288e+00 6.06875181e-01
-7.90518463e-01 -4.37732756e-01 -8.09909046e-01 3.42092305e-01
-1.03738056e-02 4.85121638e-01 5.05445659e-01 5.48400104e-01
-9.00526106e-01 2.15001658e-01 -1.04737818e+00 -3.35654080e-01
9.15945888e-01 6.19229972e-01 -5.70306003e-01 -2.89563060e-01
-6.89828455e-01 9.86120045e-01 1.02972305e+00 2.78900508e-02
-8.53944063e-01 -6.53667748e-01 -5.55018485e-01 -1.78672373e-01
5.47130287e-01 -3.15373629e-01 1.20388913e+00 -1.09401035e+00
-1.06131661e+00 1.19703865e+00 2.05160409e-01 -8.37337136e-01
7.42639244e-01 -7.63209257e-03 -2.39712313e-01 5.09918153e-01
1.37182280e-01 9.84559238e-01 7.23276377e-01 -1.00449204e+00
-7.77433634e-01 -3.97137821e-01 -3.61758918e-01 -4.32678126e-02
-4.09194976e-02 6.24701008e-02 -4.25925910e-01 -5.49731016e-01
1.44050747e-01 -1.11464691e+00 -6.06139541e-01 -2.55618364e-01
-4.76544976e-01 -2.86532313e-01 8.52357090e-01 -7.63518572e-01
8.06321859e-01 -1.82340491e+00 -6.24741502e-02 4.11158293e-01
2.54958034e-01 4.40820694e-01 1.34989142e-01 -2.85373390e-01
-1.27403185e-01 1.88518632e-02 -7.47688532e-01 -7.95619339e-02
-4.03486997e-01 3.69231969e-01 3.46910030e-01 5.22531509e-01
2.46695727e-01 7.33700454e-01 -7.21851349e-01 -8.26920867e-01
3.60385954e-01 2.48242006e-01 -5.08407891e-01 1.21148624e-01
-1.60543889e-01 1.19187331e+00 -4.73066002e-01 4.95971859e-01
4.10836965e-01 -4.59063739e-01 1.67866334e-01 -1.93635702e-01
2.27508649e-01 -3.42118442e-01 -8.39005470e-01 1.68782580e+00
-3.98753524e-01 6.27714396e-01 -7.17930272e-02 -1.61599004e+00
7.77296603e-01 5.93259692e-01 7.86835849e-01 -7.00580001e-01
2.78012127e-01 3.76294792e-01 4.33852404e-01 -8.24996471e-01
1.92854367e-02 -1.86089858e-01 9.32781175e-02 5.44503212e-01
1.28001228e-01 -3.60880792e-01 3.83480787e-01 -3.35600078e-02
1.17468786e+00 -2.16158926e-01 3.08819592e-01 -5.11994779e-01
1.00019193e+00 2.60851115e-01 3.93646479e-01 5.41250348e-01
-2.78399497e-01 4.90679264e-01 3.45720947e-01 -8.14104736e-01
-1.28501248e+00 -5.10760427e-01 -6.04458451e-01 8.90331507e-01
-1.35667607e-01 -1.08188011e-01 -1.20841944e+00 -9.96347547e-01
-6.39390051e-01 5.72613657e-01 -5.50919414e-01 2.99718112e-01
-5.88448703e-01 -1.22084033e+00 4.01160270e-01 3.04939061e-01
7.18700290e-01 -1.29315066e+00 -1.00686467e+00 2.33534381e-01
-3.10665429e-01 -1.34395576e+00 -5.65963164e-02 9.38107669e-02
-1.00885379e+00 -1.35589898e+00 -9.56001401e-01 -8.22777450e-01
1.19959605e+00 -3.50345671e-01 1.01774526e+00 3.77798259e-01
-8.24579716e-01 1.67184323e-01 -5.02489865e-01 -4.48060244e-01
-6.23491645e-01 2.13216662e-01 -3.24961245e-01 -2.88643893e-02
9.81300548e-02 -1.56296358e-01 -5.78814983e-01 4.65209275e-01
-1.10286832e+00 3.68540138e-01 8.37980688e-01 1.01779091e+00
8.86401117e-01 2.36006275e-01 6.43987179e-01 -1.44412339e+00
1.81735113e-01 -2.07867235e-01 -5.38211107e-01 2.68598586e-01
-5.39294720e-01 -2.12649908e-02 4.33788538e-01 -1.27949312e-01
-8.72034967e-01 4.99238014e-01 -5.07951975e-01 -1.01758342e-03
-3.75453770e-01 6.16815150e-01 2.30396360e-01 -5.66238463e-02
6.99749708e-01 1.91406995e-01 -2.77436208e-02 -5.03869057e-01
2.69990116e-02 6.37353659e-01 6.45182550e-01 -3.74307513e-01
1.60833225e-01 3.63244206e-01 2.63916329e-02 -8.96507323e-01
-8.44067752e-01 -6.42924607e-01 -1.19006777e+00 -3.91622424e-01
1.06517255e+00 -4.96808022e-01 -1.40783429e-01 1.92310184e-01
-7.61111140e-01 -4.45559740e-01 -1.87811807e-01 7.64633954e-01
-5.44402719e-01 3.45543414e-01 -2.13658020e-01 -5.09118140e-01
-5.55865407e-01 -1.74845457e+00 1.11544764e+00 -2.65966468e-02
-2.31580362e-01 -1.08384919e+00 -3.67292374e-01 7.35066712e-01
2.01780573e-01 2.94620275e-01 1.04758680e+00 -1.15217960e+00
-3.70072156e-01 -4.38982666e-01 -2.32463747e-01 3.46587420e-01
9.26687494e-02 -4.69152391e-01 -7.53160298e-01 -3.38648051e-01
-7.16714934e-02 -4.60011482e-01 8.16406846e-01 5.74461102e-01
1.48719537e+00 3.01248785e-02 -4.28178102e-01 2.42934138e-01
1.28927696e+00 4.03738648e-01 3.02974850e-01 2.32771695e-01
6.40026450e-01 7.17049301e-01 7.63160944e-01 2.81140089e-01
3.88539769e-02 4.95159775e-01 3.07644039e-01 -3.81421268e-01
-1.06272668e-01 3.55439931e-01 -3.52951765e-01 7.91056156e-01
-1.60025939e-01 6.00224324e-02 -1.38148117e+00 6.12930715e-01
-1.66069913e+00 -4.48362708e-01 -2.80279934e-01 2.10684395e+00
8.68353307e-01 1.77033246e-01 2.07663104e-01 4.92851555e-01
7.23624885e-01 -4.35371995e-01 -5.21689355e-01 1.50490701e-01
2.86187202e-01 4.38670427e-01 4.68673229e-01 3.41131568e-01
-1.32741189e+00 8.41754258e-01 6.19746733e+00 1.05343533e+00
-1.19516337e+00 5.88243306e-01 1.08918726e+00 1.32049859e-01
3.71799946e-01 -2.83140838e-01 -4.60582763e-01 5.11525452e-01
9.25868511e-01 2.59180367e-01 -1.72983378e-01 9.45441723e-01
2.51551837e-01 -5.80456078e-01 -1.09049976e+00 1.07589793e+00
4.94161323e-02 -1.32083225e+00 -8.54927674e-02 -8.44487324e-02
6.65587008e-01 1.04723282e-01 -2.33016640e-01 4.51707728e-02
-2.56279167e-02 -1.13522685e+00 4.75340843e-01 3.22728276e-01
5.92127442e-01 -8.65535021e-01 1.21969163e+00 7.36153722e-01
-6.83482051e-01 1.48224086e-01 -3.91730338e-01 3.64526540e-01
1.50490224e-01 7.28282392e-01 -1.45184338e+00 5.03103316e-01
6.08877838e-01 3.82514060e-01 -7.62183964e-01 1.41084588e+00
-2.16330320e-01 9.19089794e-01 -2.20296562e-01 2.38000795e-01
3.83016407e-01 -1.45833082e-02 2.25382447e-01 1.20890403e+00
1.14703111e-01 2.20437855e-01 6.21653140e-01 4.69526410e-01
6.12120852e-02 4.95975584e-01 -1.34133771e-01 1.45814985e-01
1.55543806e-02 1.49948108e+00 -1.55393994e+00 -6.57754004e-01
-1.84618443e-01 7.58749902e-01 2.25080580e-01 -2.24499851e-02
-7.11268723e-01 1.37962684e-01 -4.42177057e-01 1.23066448e-01
6.69070035e-02 9.27786827e-02 -3.84877264e-01 -7.51565516e-01
-3.16696644e-01 -6.41952872e-01 5.54666936e-01 -6.36968017e-01
-9.60660219e-01 9.42551792e-01 1.04983568e-01 -9.95489895e-01
-1.86329842e-01 -5.44981658e-01 -3.34867030e-01 5.24339497e-01
-1.27982759e+00 -1.14813924e+00 -4.37025368e-01 5.14163852e-01
7.57624030e-01 -2.88939297e-01 7.08348632e-01 4.66311425e-01
-5.73454797e-01 2.31283024e-01 -1.25002742e-01 3.30876470e-01
5.14122069e-01 -1.23291004e+00 -2.43501857e-01 6.41274452e-01
2.90052235e-01 1.47409081e-01 5.76096535e-01 -4.34409261e-01
-7.81708777e-01 -1.11369276e+00 3.27214360e-01 -6.00991063e-02
3.17450255e-01 1.50002360e-01 -1.05240965e+00 5.80619872e-01
1.57771274e-01 2.99779981e-01 8.07776392e-01 -3.01293284e-01
3.23212922e-01 1.00661710e-01 -1.30084312e+00 1.52392790e-01
4.77027327e-01 -2.00400397e-01 -4.72900242e-01 9.59410489e-01
1.73811376e-01 -4.85078067e-01 -1.16566408e+00 6.18642688e-01
2.57522315e-01 -6.63515985e-01 7.55167246e-01 -3.82234693e-01
3.13634604e-01 -2.03420177e-01 2.08754495e-01 -1.27493882e+00
1.97402403e-01 -3.04141846e-02 6.38177156e-01 8.24548602e-01
4.81573671e-01 -3.84729475e-01 9.04590786e-01 7.95640469e-01
-2.50373811e-01 -1.02697945e+00 -9.27565753e-01 -4.90567952e-01
3.74300703e-02 -5.82780123e-01 2.23747596e-01 8.69839191e-01
-2.41191879e-01 1.92429423e-01 -8.49992856e-02 6.39291704e-02
6.53376222e-01 -6.76293150e-02 5.62283397e-01 -1.25762081e+00
-1.04846075e-01 -4.19337839e-01 -5.06874204e-01 -3.92813742e-01
1.58179998e-01 -1.25873792e+00 1.30691200e-01 -1.66018975e+00
4.41550791e-01 -8.13915074e-01 -3.03785861e-01 5.34435213e-01
-2.68439025e-01 6.78411543e-01 -3.92836146e-03 4.81834024e-01
-6.63181841e-01 -1.38266191e-01 1.35875773e+00 -2.36653462e-01
2.55564172e-02 1.48671687e-01 -4.48683910e-02 1.02341378e+00
7.40734279e-01 -6.66071534e-01 -5.26409328e-01 -1.88911617e-01
-2.83521175e-01 2.22491458e-01 3.84609103e-01 -1.12001836e+00
2.41174325e-01 2.59034902e-01 5.13642848e-01 -7.46858776e-01
2.02981979e-02 -9.24991369e-01 1.38428688e-01 7.97319710e-01
-4.59264666e-01 -3.03058267e-01 1.17066935e-01 3.03924322e-01
-4.23579216e-01 -6.88108742e-01 1.07444298e+00 -4.37154919e-01
-8.81894886e-01 4.29904699e-01 -5.48208654e-01 7.76057467e-02
1.23627174e+00 -6.53297827e-02 1.83256820e-01 5.98817654e-02
-1.17966187e+00 1.19836919e-01 5.41194603e-02 4.58612442e-02
5.58104217e-01 -9.34144318e-01 -8.47221851e-01 -8.93777162e-02
9.88909230e-02 3.01125139e-01 3.47252101e-01 1.30944538e+00
-8.20549488e-01 4.85298216e-01 -3.53370667e-01 -1.12607002e+00
-1.49249601e+00 4.03510541e-01 2.00913310e-01 -5.50251007e-01
-8.46614361e-01 8.62617612e-01 3.15343529e-01 -2.61693090e-01
3.40767726e-02 -5.77205479e-01 -5.11723399e-01 2.84225866e-02
5.85011661e-01 1.95144787e-02 4.21545446e-01 -7.30427861e-01
-2.96868175e-01 3.91254961e-01 -2.75462627e-01 -8.94905534e-03
1.37028790e+00 1.04395784e-01 -1.98875636e-01 1.85716674e-01
1.08443689e+00 -4.76776481e-01 -9.22610044e-01 -3.76218736e-01
5.70665896e-01 -3.20267767e-01 4.01284665e-01 -1.05995810e+00
-1.32071531e+00 9.63293433e-01 8.52318883e-01 1.05538018e-01
1.16152418e+00 9.39320847e-02 6.15776896e-01 3.31560999e-01
5.17989516e-01 -1.05452645e+00 -1.78165212e-01 8.21713954e-02
6.99688792e-01 -1.60115516e+00 3.49344164e-02 -5.35961509e-01
-9.15156722e-01 1.15087593e+00 3.56761038e-01 -2.77209226e-02
6.17459655e-01 1.43314451e-01 1.99789554e-02 -4.07534212e-01
-3.28558415e-01 -2.10118145e-01 5.06064832e-01 3.71514767e-01
4.57390785e-01 2.40167797e-01 -4.89462107e-01 5.51957190e-01
-5.00875264e-02 7.45247006e-02 2.54961699e-01 7.06787705e-01
-6.05489314e-01 -1.24242914e+00 -3.21339995e-01 8.38976145e-01
-8.68091941e-01 -6.11723913e-03 -1.11767709e-01 9.10562217e-01
3.43204081e-01 7.59684861e-01 1.24615982e-01 1.32070377e-01
5.70539571e-02 2.02867121e-01 5.04268289e-01 -7.12146759e-01
-4.96685058e-01 2.76450366e-01 1.19622298e-01 -1.76784739e-01
-8.66518736e-01 -5.73012352e-01 -1.54390204e+00 3.88227493e-01
-3.34604502e-01 2.95490593e-01 7.05333054e-01 1.37839770e+00
-1.10115372e-01 7.40925789e-01 2.81594664e-01 -6.88035786e-01
-1.19324550e-02 -1.16214502e+00 -4.92613226e-01 4.50022608e-01
-8.39724094e-02 -6.87561035e-01 2.29613021e-01 4.41847712e-01] | [14.661480903625488, -2.3210179805755615] |
87339bd6-fc02-41ee-81c9-62169357d665 | systematic-comparison-of-neural-architectures | null | null | https://aclanthology.org/2020.emnlp-main.690 | https://aclanthology.org/2020.emnlp-main.690.pdf | Systematic Comparison of Neural Architectures and Training Approaches for Open Information Extraction | The goal of open information extraction (OIE) is to extract facts from natural language text, and to represent them as structured triples of the form {\textless}subject,predicate, object{\textgreater}. For example, given the sentence {``}Beethoven composed the Ode to Joy.{''}, we are expected to extract the triple {\textless}Beethoven, composed, Ode to Joy{\textgreater}. In this work, we systematically compare different neural network architectures and training approaches, and improve the performance of the currently best models on the OIE16 benchmark (Stanovsky and Dagan, 2016) by 0.421 F1 score and 0.420 AUC-PR, respectively, in our experiments (i.e., by more than 200{\%} in both cases). Furthermore, we show that appropriate problem and loss formulations often affect the performance more than the network architecture. | ['Thomas Lukasiewicz', 'Vid Kocijan', 'Frank Mtumbuka', 'Patrick Hohenecker'] | null | null | null | null | emnlp-2020-11 | ['open-information-extraction'] | ['natural-language-processing'] | [ 3.14805925e-01 5.59808314e-01 -1.26317546e-01 -2.70574868e-01
-8.66148829e-01 -8.99128914e-01 5.99419534e-01 5.39538383e-01
-7.08098829e-01 1.14348710e+00 1.36984149e-02 -4.35480863e-01
-2.96998501e-01 -8.22263181e-01 -1.08914006e+00 -3.09510052e-01
-1.08023122e-01 5.24666548e-01 -1.75177939e-02 -2.52217472e-01
-1.51604652e-01 2.25248992e-01 -1.59129882e+00 3.92466635e-01
7.78409898e-01 1.33989894e+00 -2.92170703e-01 5.53036690e-01
6.44881725e-02 1.06948543e+00 -7.01746166e-01 -1.12785673e+00
2.02995747e-01 1.32898718e-01 -1.17686045e+00 -5.45416117e-01
5.18805206e-01 -1.96865231e-01 -2.65777320e-01 9.58615243e-01
1.29480124e-01 1.57085776e-01 9.18322086e-01 -1.04566026e+00
-4.85024840e-01 1.09730244e+00 -4.71499652e-01 4.44402471e-02
2.30005816e-01 -1.87436081e-02 1.50864351e+00 -8.12033594e-01
9.20790613e-01 7.88700402e-01 5.48214257e-01 5.47153234e-01
-1.11562431e+00 -6.59524679e-01 1.85530543e-01 2.10607052e-01
-1.29364538e+00 -3.26078176e-01 4.30498064e-01 -2.00341865e-01
1.01963663e+00 5.94424129e-01 2.48312443e-01 1.27088475e+00
1.21609271e-01 1.03442645e+00 8.70944619e-01 -2.85208464e-01
1.34598901e-02 8.82633924e-02 4.52652156e-01 6.59667790e-01
8.03469658e-01 -8.33980069e-02 -4.18217957e-01 -2.62676515e-02
1.59618765e-01 -2.46179566e-01 -1.04472712e-01 8.82141963e-02
-1.06724536e+00 7.54438698e-01 3.55291933e-01 2.45706394e-01
-3.74618620e-01 1.89046651e-01 6.00418091e-01 1.94211125e-01
3.94791424e-01 8.14321876e-01 -8.02437484e-01 -7.12568164e-02
-7.52528846e-01 8.51006150e-01 1.06582201e+00 1.08586991e+00
4.83884543e-01 -2.33884528e-01 -2.43773609e-01 6.57580972e-01
-9.54881683e-02 1.88405380e-01 -2.52809748e-02 -1.00202799e+00
9.89647627e-01 5.21655262e-01 3.13705295e-01 -7.05974281e-01
-6.45331144e-01 -3.56808186e-01 -8.73412311e-01 -7.71884322e-02
7.34291732e-01 -6.42432272e-01 -8.39815855e-01 1.72041070e+00
-2.83448808e-02 -4.17496234e-01 4.79398102e-01 6.15887105e-01
1.43499076e+00 7.03147829e-01 3.28785300e-01 -1.74079746e-01
1.60996056e+00 -6.47417128e-01 -6.93386495e-01 -5.05657971e-01
6.89257443e-01 -5.32758415e-01 9.44464684e-01 5.32632649e-01
-1.23683715e+00 -2.48443801e-02 -1.22271121e+00 -3.34343761e-01
-6.03040457e-01 2.39883825e-01 6.47807956e-01 2.44584888e-01
-5.13315618e-01 7.79051661e-01 -6.16162658e-01 5.94335496e-02
5.88703692e-01 3.19377720e-01 -5.52778721e-01 -5.55911858e-04
-1.44260395e+00 9.01884019e-01 9.33100581e-01 2.71405410e-02
-6.44390166e-01 -9.87629950e-01 -9.52769995e-01 3.98693025e-01
9.13762331e-01 -6.97756231e-01 1.24679482e+00 -5.87671280e-01
-9.35336530e-01 1.10262311e+00 7.88860768e-02 -9.75124955e-01
5.95030487e-01 -5.69822073e-01 -3.19692612e-01 -3.23567949e-02
2.60727182e-02 6.06681883e-01 1.76103741e-01 -1.35618365e+00
-8.87716293e-01 -3.82086009e-01 5.54197013e-01 -7.97125921e-02
-1.97999850e-01 4.57460254e-01 -3.41770738e-01 -4.80791420e-01
-1.39073417e-01 -9.39139187e-01 2.36222129e-02 -4.77002323e-01
-9.01263952e-01 -5.95877469e-01 3.01588476e-01 -9.31978464e-01
1.51526845e+00 -2.01720548e+00 1.58387735e-01 1.26569659e-01
7.52105772e-01 3.29946786e-01 1.86579451e-01 3.77656549e-01
-2.82943279e-01 3.53258371e-01 -5.72101057e-01 -1.95694223e-01
4.04170573e-01 1.73596144e-01 -1.99668020e-01 6.36896864e-02
2.70225972e-01 8.97061765e-01 -6.22944653e-01 -3.65528256e-01
-3.27087551e-01 2.18625054e-01 -4.60486680e-01 -4.56670150e-02
-7.20761240e-01 -7.63853565e-02 -3.36358041e-01 4.80316252e-01
3.79050165e-01 -2.55419850e-01 2.06784189e-01 -2.70717859e-01
-1.03407174e-01 6.01629138e-01 -1.05324233e+00 1.38437796e+00
-4.50940222e-01 8.12007785e-01 1.20165069e-02 -8.20511997e-01
7.27973342e-01 3.59958589e-01 5.72707236e-01 -5.16218126e-01
3.49981695e-01 3.78220916e-01 1.74129736e-02 -3.64718229e-01
6.86093748e-01 1.53067201e-01 -3.54359359e-01 1.62840605e-01
1.16998367e-01 1.04363002e-01 8.88300121e-01 2.56328732e-01
1.15329516e+00 1.56793371e-01 3.32700789e-01 -2.57354051e-01
3.00341427e-01 1.13319188e-01 6.42998695e-01 7.01420784e-01
1.96320385e-01 4.23704207e-01 1.12696803e+00 -6.07491732e-01
-1.24407995e+00 -9.50313866e-01 -2.88855851e-01 9.66286063e-01
-1.66380376e-01 -6.97288871e-01 -7.77915001e-01 -8.44307780e-01
-1.54505417e-01 1.14357662e+00 -5.74090540e-01 -1.18119922e-02
-7.43066251e-01 -8.09170902e-01 8.89707267e-01 4.48698640e-01
3.68408293e-01 -1.06716049e+00 -7.91425705e-01 1.19104356e-01
-5.18094063e-01 -1.25117993e+00 5.57631291e-02 5.74615538e-01
-3.74395698e-01 -9.69343007e-01 -4.47090715e-01 -4.93465751e-01
2.78724372e-01 -6.91462517e-01 1.60077369e+00 -5.92332780e-02
-1.97000653e-01 -3.67645830e-01 -4.41800177e-01 -7.41058588e-01
-8.91925320e-02 2.45262891e-01 -1.57880262e-01 -2.61126488e-01
5.62656760e-01 -4.88785714e-01 -3.07117224e-01 -1.86827287e-01
-1.00555706e+00 1.50325820e-01 4.32914913e-01 7.63562322e-01
4.90319461e-01 1.24811664e-01 3.03300470e-01 -1.25418150e+00
6.18210733e-01 -4.84183013e-01 -6.75836682e-01 2.77109832e-01
-6.55295074e-01 2.50135213e-01 8.46629620e-01 -1.23070799e-01
-8.31055701e-01 -2.70973623e-01 -1.75601155e-01 -1.78223789e-01
-2.82413781e-01 7.06980348e-01 -1.47331417e-01 5.47060072e-01
7.75895774e-01 -5.26743941e-02 -6.33326828e-01 -4.75172758e-01
3.50276351e-01 4.70532060e-01 7.39556849e-01 -8.42491269e-01
5.91249883e-01 2.59986192e-01 1.28850982e-01 -4.38316315e-01
-1.27816784e+00 -1.45171036e-03 -2.39261001e-01 2.46155903e-01
8.53044569e-01 -7.14650035e-01 -1.07723022e+00 6.05302826e-02
-1.40977240e+00 -8.53940099e-02 -3.77754718e-01 2.65323848e-01
-4.54650432e-01 6.01119325e-02 -6.67673528e-01 -7.85437346e-01
-5.98032117e-01 -8.64638269e-01 7.81551778e-01 2.73171902e-01
-4.06893998e-01 -6.74313784e-01 -3.76733750e-01 4.50605333e-01
-1.16938829e-01 7.14124680e-01 1.25296378e+00 -1.26814032e+00
-3.87323201e-01 -2.45414376e-01 -4.95860815e-01 4.30566847e-01
-1.97686359e-01 7.90735111e-02 -8.64099503e-01 5.34367785e-02
-3.76812667e-01 -4.51550364e-01 8.77373099e-01 1.22339927e-01
1.23915589e+00 -8.18804562e-01 -2.08939254e-01 4.87970769e-01
1.51115537e+00 2.60865778e-01 7.18489289e-01 4.59476829e-01
6.29113615e-01 8.31239998e-01 4.20287251e-01 4.43367660e-01
3.48770857e-01 3.47529113e-01 3.21338564e-01 1.38286799e-01
9.85811874e-02 -2.28048354e-01 2.17943698e-01 3.63503359e-02
-3.80733997e-01 -5.16606092e-01 -1.01036429e+00 5.01341224e-01
-1.78228390e+00 -9.44634140e-01 -2.77382344e-01 2.05592847e+00
1.13925707e+00 5.40021300e-01 -3.51240113e-02 1.88740522e-01
2.51627058e-01 2.35818639e-01 -3.07663143e-01 -5.84910274e-01
-2.25150779e-01 4.89903659e-01 6.01427495e-01 2.08606198e-01
-1.15188169e+00 1.01420557e+00 5.38704014e+00 8.68502080e-01
-7.50487387e-01 -1.49588212e-01 6.25690222e-01 -2.22921550e-01
-2.74194032e-01 -6.72829375e-02 -1.16175568e+00 4.39268947e-01
9.94643807e-01 -2.37636626e-01 4.35925663e-01 5.15638292e-01
-2.49955878e-01 -1.94273964e-01 -1.49848330e+00 7.64903724e-01
1.94119271e-02 -1.36674738e+00 1.15658278e-02 -7.49993175e-02
6.84926808e-01 -1.12792775e-01 -1.01912148e-01 4.53916818e-01
6.82623029e-01 -1.20791364e+00 8.39977145e-01 4.77512151e-01
7.51073003e-01 -8.66923273e-01 8.01359534e-01 4.27603334e-01
-8.35164487e-01 -1.70120314e-01 -1.07756909e-02 3.28953676e-02
-5.96489646e-02 7.35321820e-01 -6.88067317e-01 9.40794826e-01
8.56155038e-01 2.09026590e-01 -3.73356193e-01 5.95165133e-01
-4.41408962e-01 6.33530915e-01 -5.56749165e-01 -3.82005006e-01
4.41461712e-01 -1.28161192e-01 6.68357968e-01 1.21082306e+00
-1.01266555e-01 2.05681399e-01 -2.48153303e-02 1.03423035e+00
-6.00501120e-01 -5.11073566e-04 -5.64682364e-01 -1.79309458e-01
3.66294503e-01 1.03981757e+00 -4.04788077e-01 -3.96116734e-01
-3.87100071e-01 5.17429233e-01 5.47444761e-01 3.98582429e-01
-8.69348586e-01 -8.11396480e-01 4.69229996e-01 8.36019069e-02
2.92838901e-01 1.04480013e-01 -4.72551823e-01 -9.81206536e-01
4.57016498e-01 -8.60434592e-01 7.50121295e-01 -6.92077935e-01
-1.15901196e+00 7.34103203e-01 2.38331035e-01 -6.70052230e-01
-4.44532931e-01 -7.66523063e-01 -2.82630026e-01 7.57093549e-01
-1.18697619e+00 -1.02471340e+00 2.21713886e-01 3.22342753e-01
4.48826253e-01 -1.36786429e-02 7.31360495e-01 4.64501351e-01
-7.07176864e-01 6.46254599e-01 8.59194621e-02 5.79700053e-01
1.92048922e-01 -1.24699569e+00 4.57133293e-01 8.59321952e-01
2.36431941e-01 6.77926660e-01 1.00810909e+00 -5.93076289e-01
-1.12240815e+00 -9.45786834e-01 1.40991950e+00 -7.11874902e-01
8.02360058e-01 -5.26921809e-01 -7.25197017e-01 1.00090778e+00
2.33420521e-01 -1.43053994e-01 3.88081908e-01 4.77978796e-01
-5.43740392e-01 -1.04646221e-01 -1.07808590e+00 8.23607028e-01
9.59670961e-01 -2.02774763e-01 -8.34397972e-01 4.22311783e-01
9.87274349e-01 -5.56419849e-01 -1.11659920e+00 7.16278553e-01
4.73186076e-01 -9.10503805e-01 9.81287003e-01 -1.13973820e+00
1.00075233e+00 9.23556238e-02 -1.56432420e-01 -1.08590543e+00
1.32594323e-02 -5.66314876e-01 -4.00709838e-01 1.20110059e+00
1.11583996e+00 -3.97235841e-01 6.39966547e-01 8.60519469e-01
-2.93496788e-01 -1.09180391e+00 -1.00827956e+00 -6.11587882e-01
3.71466100e-01 -4.99206632e-01 6.45965338e-01 6.69295669e-01
-4.60021161e-02 4.56247211e-01 -3.00288528e-01 -4.87653650e-02
5.90279102e-01 3.87111753e-02 3.68225336e-01 -1.34441459e+00
-1.51652768e-01 -4.50506121e-01 9.00439918e-02 -7.04755068e-01
2.29889408e-01 -9.30149257e-01 -1.51137143e-01 -1.64842451e+00
1.41932428e-01 -4.99199629e-01 -3.36760670e-01 7.75076091e-01
-9.45841148e-02 5.94811849e-02 2.86576092e-01 -1.69114232e-01
-5.88165462e-01 1.82630017e-01 1.03357315e+00 -2.50628114e-01
5.11981770e-02 -3.47022451e-02 -9.56363559e-01 8.95515323e-01
6.36875153e-01 -6.30071044e-01 -4.63375002e-02 -5.60777426e-01
8.05023968e-01 1.15193784e-01 5.10303736e-01 -7.86316752e-01
1.00694150e-01 -9.79781449e-02 2.24949598e-01 -5.54249048e-01
3.89955133e-01 -7.41509199e-01 2.58823708e-02 2.98070848e-01
-6.03951693e-01 1.67245064e-02 3.06314319e-01 1.78244293e-01
-1.71309203e-01 -5.17782688e-01 3.80553126e-01 -1.82135314e-01
-3.72651130e-01 2.34529451e-01 1.05124928e-01 5.05755603e-01
7.30502665e-01 2.21957982e-01 -5.67463696e-01 -2.82826815e-02
-6.36418104e-01 2.31827095e-01 -2.87300527e-01 4.03948903e-01
2.97497034e-01 -1.07804668e+00 -8.68081272e-01 -9.58118737e-02
1.69487998e-01 4.58067775e-01 7.87355453e-02 6.55621648e-01
-5.93139112e-01 5.49622893e-01 7.45633543e-02 3.55089433e-03
-1.11591983e+00 4.54683840e-01 8.25332925e-02 -7.17601061e-01
-6.34925783e-01 9.40190136e-01 -1.02376275e-01 -4.32871372e-01
5.05422175e-01 -4.42927539e-01 -2.48629436e-01 1.59138396e-01
3.45015407e-01 3.70871425e-01 3.27810615e-01 -3.07020187e-01
-3.18047196e-01 -3.74542028e-02 -2.96094030e-01 1.12153227e-02
1.35872698e+00 4.54412133e-01 -1.67734116e-01 3.34521264e-01
1.19396305e+00 -6.53007999e-03 -8.52666974e-01 -2.37214252e-01
2.03336418e-01 -1.39320061e-01 -2.74931312e-01 -1.19003356e+00
-7.88334370e-01 4.13760453e-01 1.16059542e-01 4.79954451e-01
9.17500257e-01 2.18215749e-01 9.35591221e-01 8.48010182e-01
2.48918921e-01 -1.13203609e+00 -5.76205671e-01 7.31053114e-01
8.47927868e-01 -1.00823092e+00 1.58151269e-01 -4.20505881e-01
-6.51408136e-01 8.46772492e-01 5.07594764e-01 -2.06720456e-02
2.73119986e-01 3.52367222e-01 -3.45976949e-01 -3.38694096e-01
-1.02093065e+00 -1.89808071e-01 3.51704746e-01 1.60992667e-01
4.54738915e-01 1.03566326e-01 -5.57250202e-01 1.01083529e+00
-5.62269032e-01 -1.12640113e-01 3.91535938e-01 7.81821012e-01
-2.87439615e-01 -8.90120149e-01 -3.01238149e-01 7.73019552e-01
-9.06393766e-01 -3.60636890e-01 -6.18897080e-01 1.02086401e+00
3.39389950e-01 7.98111975e-01 3.61785106e-02 -1.89588189e-01
4.69465166e-01 2.04183742e-01 3.74460965e-01 -3.65446150e-01
-9.08446550e-01 -2.86626279e-01 7.64388263e-01 -2.54738122e-01
-1.22775666e-01 -6.38841093e-01 -1.51159787e+00 -4.18720931e-01
-1.13962345e-01 2.08405614e-01 5.95259726e-01 1.19651508e+00
6.25306740e-02 6.95576251e-01 2.35015929e-01 -6.10282905e-02
-4.63527352e-01 -9.71145272e-01 -3.87891114e-01 5.06037354e-01
1.99528605e-01 -5.89800954e-01 -3.35685372e-01 -3.70360492e-03] | [9.539167404174805, 8.787618637084961] |
715596c5-db79-454b-be0f-3df1935ff0ea | relate-auditory-speech-to-eeg-by-shallow-deep | 2303.10897 | null | https://arxiv.org/abs/2303.10897v1 | https://arxiv.org/pdf/2303.10897v1.pdf | Relate auditory speech to EEG by shallow-deep attention-based network | Electroencephalography (EEG) plays a vital role in detecting how brain responses to different stimulus. In this paper, we propose a novel Shallow-Deep Attention-based Network (SDANet) to classify the correct auditory stimulus evoking the EEG signal. It adopts the Attention-based Correlation Module (ACM) to discover the connection between auditory speech and EEG from global aspect, and the Shallow-Deep Similarity Classification Module (SDSCM) to decide the classification result via the embeddings learned from the shallow and deep layers. Moreover, various training strategies and data augmentation are used to boost the model robustness. Experiments are conducted on the dataset provided by Auditory EEG challenge (ICASSP Signal Processing Grand Challenge 2023). Results show that the proposed model has a significant gain over the baseline on the match-mismatch track. | ['Dongmei Jiang', 'Yujun Wang', 'Ercheng Pei', 'Jiyao Liu', 'Lang He', 'Liyong Guo', 'Fan Cui'] | 2023-03-20 | null | null | null | null | ['deep-attention', 'eeg', 'deep-attention', 'eeg'] | ['computer-vision', 'methodology', 'natural-language-processing', 'time-series'] | [-2.36129574e-02 -5.10708332e-01 7.24877298e-01 -4.69826311e-01
-4.59104389e-01 1.26896054e-01 4.40211207e-01 5.14131561e-02
-4.44707960e-01 2.38776296e-01 5.42626262e-01 4.94645014e-02
-2.19719395e-01 -3.81713063e-01 -3.94303858e-01 -5.87214470e-01
-3.79352212e-01 -1.23630315e-01 -2.61874404e-02 -2.04937428e-01
4.86859083e-01 3.72554839e-01 -1.35011530e+00 5.55996954e-01
6.43589854e-01 1.58570349e+00 3.26771826e-01 5.10747552e-01
3.15854490e-01 6.32859647e-01 -5.67460060e-01 -9.03921872e-02
-2.31628325e-02 -5.80830932e-01 -6.89638734e-01 -5.13152301e-01
-6.69504628e-02 5.45772612e-02 -4.42106515e-01 1.14224982e+00
1.14791214e+00 1.79757774e-01 5.58628201e-01 -1.33015692e+00
-7.02898443e-01 5.60599267e-01 -4.32785302e-01 1.23392975e+00
4.79418665e-01 3.00479174e-01 9.20371115e-01 -1.35064197e+00
-2.87762612e-01 1.00856960e+00 6.06943905e-01 4.63188887e-01
-9.48804975e-01 -9.99678135e-01 1.12904951e-01 1.14410424e+00
-1.54631150e+00 -4.56712872e-01 9.03722107e-01 -3.64187866e-01
1.29403675e+00 2.18207851e-01 8.33835602e-01 1.51723349e+00
6.74081087e-01 6.45985305e-01 1.18071628e+00 -7.50667676e-02
2.71948576e-01 6.35263622e-02 4.97091383e-01 -2.31964998e-02
-4.33532953e-01 8.07853714e-02 -9.44589257e-01 -1.53131858e-01
3.17999363e-01 -1.26749888e-01 -6.80015862e-01 3.90015215e-01
-1.07069111e+00 4.39141959e-01 7.64141738e-01 5.87631345e-01
-9.27616715e-01 -2.44916290e-01 3.98933500e-01 4.66747910e-01
4.02693301e-01 7.57549405e-01 -5.20006537e-01 -6.33049682e-02
-7.60261834e-01 -1.12576686e-01 5.83331048e-01 4.69936728e-01
3.92625391e-01 1.73119962e-01 -3.93238842e-01 9.44414973e-01
2.57309765e-01 6.79931641e-02 1.12125623e+00 -2.34697968e-01
3.99501503e-01 2.89713532e-01 -5.07265270e-01 -1.39612496e+00
-5.67053437e-01 -7.28479981e-01 -9.00859833e-01 -1.50540382e-01
-3.77906233e-01 -3.93690616e-02 -5.86420059e-01 1.71331298e+00
-2.15755433e-01 7.97092736e-01 -1.39843291e-02 1.07919729e+00
8.86621177e-01 4.47354943e-01 1.93745017e-01 8.72984082e-02
1.54150856e+00 -7.54830599e-01 -7.71525562e-01 -7.13849127e-01
2.96079321e-04 -4.09531027e-01 1.04689109e+00 4.30648565e-01
-1.15182686e+00 -8.02709401e-01 -1.36949480e+00 1.17695324e-01
-4.44753587e-01 -2.29096577e-01 2.61146367e-01 1.33340120e-01
-1.07287037e+00 3.71085286e-01 -6.15715683e-01 -2.19486982e-01
4.87194985e-01 5.86256146e-01 -2.30576172e-01 1.28637865e-01
-1.64852488e+00 1.01943481e+00 2.28861287e-01 3.09848487e-01
-9.69693065e-01 -8.76458287e-01 -5.02274096e-01 5.60048521e-01
-2.88051546e-01 -5.12504458e-01 8.34662437e-01 -9.85446095e-01
-1.40100515e+00 3.98047894e-01 -7.11841136e-02 -6.17039382e-01
-1.21753708e-01 -1.85110539e-01 -6.95325494e-01 9.29753929e-02
-1.12760559e-01 6.81383491e-01 6.83021247e-01 -6.14939511e-01
-4.40271109e-01 -5.42378068e-01 -5.73982358e-01 1.43717870e-01
-4.58663940e-01 3.80501658e-01 -3.78178968e-03 -7.72846639e-01
2.01934949e-01 -6.73296690e-01 1.59905151e-01 -6.83182716e-01
-3.01606148e-01 -5.79822361e-01 3.36505592e-01 -9.76087630e-01
1.13600135e+00 -2.56910348e+00 1.58139557e-01 2.26018339e-01
2.60754764e-01 3.17765623e-02 -5.43799341e-01 2.46543884e-01
-7.05748439e-01 -1.90089852e-01 -1.23474978e-01 -2.47180089e-01
8.36179480e-02 -2.91347802e-01 -3.95839393e-01 4.14952129e-01
4.56438303e-01 9.77814555e-01 -6.80853307e-01 1.38999984e-01
-1.73355401e-01 6.20892942e-01 -4.86930430e-01 5.36558211e-01
5.21326005e-01 6.07228398e-01 5.98176010e-02 3.48351896e-01
6.84800625e-01 -1.95850339e-03 -2.10087940e-01 -2.56763816e-01
-3.34094279e-02 7.20451891e-01 -8.26738238e-01 1.79251003e+00
-3.90123487e-01 8.53217781e-01 -2.33214751e-01 -9.45316792e-01
9.49840248e-01 6.14452839e-01 -3.25453060e-04 -1.16227567e+00
6.15188777e-01 -6.41000178e-03 8.14056098e-01 -7.25967586e-01
-1.66579142e-01 1.04742281e-01 1.96333200e-01 4.38258916e-01
4.04505640e-01 2.07849935e-01 -4.00575519e-01 1.00317016e-01
1.31170464e+00 -6.08577132e-01 1.35627508e-01 -4.71479684e-01
5.55804133e-01 -8.72130215e-01 5.95577657e-01 5.94867289e-01
-3.53649259e-01 6.65048242e-01 4.21713054e-01 -3.17700982e-01
-4.64760989e-01 -8.30175400e-01 -1.80215642e-01 1.10327899e+00
-2.44076513e-02 -2.77992636e-01 -6.74246967e-01 -2.78450102e-01
-1.48499772e-01 7.22746968e-01 -9.37420607e-01 -8.60410511e-01
-1.64372265e-01 -9.30019081e-01 3.45868498e-01 9.37909663e-01
6.09401703e-01 -1.29184043e+00 -3.50981474e-01 1.02562137e-01
-2.95787454e-01 -1.11153817e+00 -3.72330189e-01 5.80519915e-01
-4.30572033e-01 -9.31953728e-01 -4.03802037e-01 -7.64841259e-01
2.56936908e-01 9.26888082e-03 6.86229050e-01 -1.57695234e-01
-1.64211348e-01 8.60497057e-02 -2.71686643e-01 -6.18862152e-01
2.48829007e-01 -4.58081327e-02 2.85556823e-01 4.61242437e-01
1.00256479e+00 -1.02722013e+00 -8.63068163e-01 1.81862578e-01
-4.68802929e-01 -3.66105855e-01 6.68913841e-01 6.96859658e-01
1.28370643e-01 3.86880226e-02 1.09038591e+00 -4.28891741e-02
1.17433667e+00 -9.05547142e-01 5.91615029e-03 -7.56277218e-02
-5.45636237e-01 -7.14735612e-02 6.54008150e-01 -4.61220771e-01
-5.96836388e-01 -3.75056416e-01 -2.97827959e-01 -4.12872791e-01
-2.78237760e-01 5.98370612e-01 -3.86552364e-01 -1.20215453e-01
5.69814026e-01 4.28476989e-01 -5.29048681e-01 -5.66186428e-01
-3.05016935e-01 8.77129495e-01 7.00558066e-01 4.00327183e-02
2.59967655e-01 -1.57587484e-01 -3.91279966e-01 -4.26116198e-01
-6.15346730e-01 -3.41603756e-01 -5.57249188e-01 -1.38129033e-02
9.13198769e-01 -1.01262248e+00 -7.25061178e-01 4.75895047e-01
-1.37401736e+00 2.73196921e-02 2.18884006e-01 9.31663930e-01
-9.26660970e-02 -5.69176301e-02 -5.52602172e-01 -6.69102132e-01
-5.67244351e-01 -1.06933379e+00 5.64868271e-01 9.51722115e-02
-4.62399125e-01 -5.74316084e-01 1.88698113e-01 -3.37145850e-02
6.16830111e-01 -3.07229251e-01 9.50275540e-01 -1.32624626e+00
-1.05647817e-01 -1.77453980e-01 -3.19461703e-01 5.12953579e-01
-8.32309350e-02 -5.65416753e-01 -1.45306098e+00 -1.98046088e-01
4.79942501e-01 -1.42351493e-01 9.67239261e-01 1.62957326e-01
1.59132993e+00 -6.25803396e-02 2.46690325e-02 7.87398756e-01
1.03358221e+00 4.90620762e-01 9.03533816e-01 3.63118470e-01
4.02037650e-01 4.55866516e-01 -2.38763586e-01 3.61447185e-01
4.91203547e-01 4.99358535e-01 3.24980140e-01 1.83541834e-01
-1.76691324e-01 5.44448234e-02 3.09870929e-01 1.22314298e+00
1.27333805e-01 4.45805453e-02 -1.15526712e+00 6.51133657e-01
-1.55441546e+00 -7.66990006e-01 -5.44950105e-02 2.03483796e+00
6.55411541e-01 -1.19628906e-02 -2.80138880e-01 5.87463200e-01
5.41211009e-01 -1.57573938e-01 -4.85581279e-01 -5.30890226e-01
-2.43271530e-01 7.16837645e-01 -2.41500273e-01 3.71256620e-02
-8.36223364e-01 3.65528405e-01 5.94839478e+00 5.00225604e-01
-1.35841799e+00 4.58954275e-01 4.40461755e-01 -2.03197181e-01
2.31989101e-01 -6.46121979e-01 -2.86394000e-01 7.83272684e-01
1.42672443e+00 -2.24842042e-01 6.52586937e-01 4.71202254e-01
2.67104834e-01 -2.13041785e-03 -1.27165031e+00 1.32675409e+00
3.43381226e-01 -8.18535268e-01 -1.85452789e-01 -1.71054408e-01
1.40501544e-01 4.09818590e-01 3.66353542e-01 4.71819460e-01
-3.00933003e-01 -1.10006893e+00 5.88356018e-01 7.41061985e-01
3.61152112e-01 -8.57890487e-01 9.34869051e-01 1.56543404e-01
-1.14085448e+00 -5.57098031e-01 -4.76703852e-01 -2.04357311e-01
-3.60104233e-01 2.96161532e-01 -8.71685266e-01 2.37884775e-01
1.10779786e+00 8.69662046e-01 -7.57406712e-01 1.40533185e+00
-5.07678092e-01 7.85870910e-01 2.08348423e-01 -6.82261139e-02
-5.57940565e-02 2.20787749e-01 6.06915593e-01 1.34098756e+00
4.15452927e-01 3.12137902e-01 -4.67781305e-01 1.02025938e+00
-2.08084717e-01 -1.28201572e-02 -2.69916505e-01 2.55670667e-01
6.54782712e-01 1.30242288e+00 -3.81034672e-01 -8.99794698e-02
-2.81429231e-01 1.12587953e+00 3.53612602e-01 4.44726288e-01
-7.88704157e-01 -7.61963129e-01 8.53477240e-01 -3.62716764e-01
2.51698583e-01 7.84706175e-02 -3.50376070e-01 -9.97813225e-01
-4.95185629e-02 -7.49709129e-01 1.46188274e-01 -1.27516651e+00
-1.64001536e+00 1.21305048e+00 -3.84465098e-01 -9.77422357e-01
1.65853538e-02 -4.56155330e-01 -1.15489483e+00 1.40008521e+00
-1.62376690e+00 -3.70825648e-01 -3.81439358e-01 9.62871909e-01
5.26902556e-01 -1.72638074e-01 8.97325218e-01 5.51243961e-01
-8.55559170e-01 7.61587560e-01 -3.72450501e-01 2.56910086e-01
6.40070438e-01 -1.00076401e+00 4.09807831e-01 8.75731289e-01
9.71298069e-02 7.44081020e-01 3.76344681e-01 -1.91183656e-01
-1.16095507e+00 -1.06144261e+00 1.03711557e+00 -1.63248867e-01
7.02259958e-01 -6.44854069e-01 -1.23167741e+00 3.27309102e-01
8.05720210e-01 -2.79165693e-02 9.50605989e-01 2.17396487e-03
-4.57818955e-01 -4.28138405e-01 -7.76216924e-01 3.07397515e-01
7.17146397e-01 -1.08784938e+00 -1.07212150e+00 -2.89531164e-02
5.98110914e-01 2.04121217e-01 -7.65991390e-01 3.54394972e-01
4.33939755e-01 -8.09334993e-01 7.87997186e-01 -6.49058878e-01
3.08469951e-01 6.85443208e-02 -3.14373821e-01 -1.98308802e+00
-8.08458090e-01 -4.39076334e-01 1.64889097e-01 1.03299105e+00
3.54843229e-01 -7.88676083e-01 5.84220551e-02 4.51680928e-01
-4.89713848e-01 -7.40951121e-01 -1.09602582e+00 -6.51409984e-01
-2.33987704e-01 -7.35970497e-01 8.91825080e-01 1.04882443e+00
3.88680726e-01 7.16336608e-01 -5.40447012e-02 5.66275597e-01
9.73702446e-02 -4.57408965e-01 -1.12714414e-02 -1.38258958e+00
-3.28239240e-02 -5.02023399e-01 -8.56408417e-01 -5.66091895e-01
2.92909145e-01 -1.14701307e+00 1.20557025e-01 -1.23095810e+00
3.53773206e-01 2.32802063e-01 -1.29344022e+00 4.69921827e-01
-2.65451550e-01 1.50086612e-01 5.60181141e-02 -1.00555323e-01
-4.38364983e-01 8.77844870e-01 5.01109362e-01 -1.20463178e-01
-2.82872822e-02 -2.04304755e-01 -8.50775778e-01 6.03054285e-01
9.69609559e-01 -6.07026339e-01 -3.13216060e-01 -5.96570551e-01
1.24126479e-01 -1.57184556e-01 5.69767952e-01 -1.60871935e+00
7.06107795e-01 5.69322288e-01 7.72263885e-01 -3.16474736e-01
3.78409296e-01 -6.72780633e-01 -2.83813030e-01 2.32769087e-01
-6.78227603e-01 4.19968784e-01 4.22972083e-01 5.23376584e-01
-4.47144955e-01 1.21104129e-01 7.69444406e-01 1.37989834e-01
-5.02705514e-01 4.91955996e-01 -5.39105773e-01 5.88917062e-02
6.11798644e-01 -2.91496515e-02 -2.94169754e-01 -1.37499288e-01
-7.96659589e-01 4.62467819e-02 -5.61824441e-01 6.06278896e-01
1.12617505e+00 -1.65093470e+00 -8.47289383e-01 8.44357729e-01
2.31556058e-01 -7.58498192e-01 2.60781080e-01 1.36517286e+00
3.65319014e-01 4.83347803e-01 -4.27245259e-01 -4.34729874e-01
-1.00718057e+00 3.47923130e-01 4.77894694e-01 2.71399200e-01
-3.67546707e-01 1.41443396e+00 3.53395343e-01 2.37080362e-02
4.03714418e-01 -3.70928437e-01 -6.24674201e-01 2.06672907e-01
1.07223034e+00 2.03374997e-01 7.01419115e-01 -5.43547630e-01
-7.55188584e-01 7.57307261e-02 -7.35629126e-02 -1.16883703e-01
1.72728336e+00 -1.20205440e-01 -3.15539032e-01 4.66060251e-01
1.48591018e+00 -4.49444771e-01 -8.84268463e-01 -2.95513928e-01
4.06108908e-02 -2.81499743e-01 5.91535687e-01 -1.17046332e+00
-1.30208266e+00 1.48270309e+00 9.71652567e-01 2.25558251e-01
1.39764082e+00 -1.34206399e-01 7.40917802e-01 2.77591497e-01
1.22564048e-01 -8.70346725e-01 2.30829164e-01 6.92779362e-01
1.35445046e+00 -9.72409487e-01 -4.53105479e-01 4.23860788e-01
-6.65543020e-01 1.13838708e+00 7.14488924e-01 -3.26165617e-01
9.59330857e-01 2.05174193e-01 -1.93553522e-01 -2.06039265e-01
-1.13889396e+00 -5.19752689e-02 6.38493359e-01 6.13674343e-01
3.54075551e-01 -1.33869931e-01 8.87803808e-02 1.54387009e+00
-1.95003614e-01 -1.66530773e-01 1.79043069e-01 5.39689064e-01
-3.65761161e-01 -5.14615893e-01 -1.26669928e-01 4.37664241e-01
-4.83455062e-01 -6.90911949e-01 -5.48127055e-01 3.83443505e-01
2.42210235e-02 1.15125370e+00 3.68022472e-01 -1.06901753e+00
5.53471088e-01 4.34079021e-01 2.29546979e-01 -5.79446614e-01
-1.08606327e+00 -1.70154553e-02 -3.62794667e-01 -7.81166553e-01
-1.28925681e-01 -5.54006457e-01 -1.10193658e+00 1.49152994e-01
-2.22093478e-01 2.72168845e-01 6.34230793e-01 1.04257905e+00
8.55988026e-01 9.53136027e-01 1.00911999e+00 -8.66272986e-01
-3.85244578e-01 -1.35242581e+00 -6.98566139e-01 7.41867572e-02
5.35181403e-01 -5.90588510e-01 -6.64809167e-01 -3.48713070e-01] | [13.20036792755127, 3.4640486240386963] |
4f0b986b-78d3-48b8-b39c-eedb706db359 | to-answer-or-not-to-answer-improving-machine-1 | 2208.01299 | null | https://arxiv.org/abs/2208.01299v1 | https://arxiv.org/pdf/2208.01299v1.pdf | To Answer or Not to Answer? Improving Machine Reading Comprehension Model with Span-based Contrastive Learning | Machine Reading Comprehension with Unanswerable Questions is a difficult NLP task, challenged by the questions which can not be answered from passages. It is observed that subtle literal changes often make an answerable question unanswerable, however, most MRC models fail to recognize such changes. To address this problem, in this paper, we propose a span-based method of Contrastive Learning (spanCL) which explicitly contrast answerable questions with their answerable and unanswerable counterparts at the answer span level. With spanCL, MRC models are forced to perceive crucial semantic changes from slight literal differences. Experiments on SQuAD 2.0 dataset show that spanCL can improve baselines significantly, yielding 0.86-2.14 absolute EM improvements. Additional experiments also show that spanCL is an effective way to utilize generated questions. | ['Xiangang Li', 'Baochang Ma', 'Chenxiao Dou', 'Liangyu Chen', 'Yunjie Ji'] | 2022-08-02 | to-answer-or-not-to-answer-improving-machine | https://aclanthology.org/2022.findings-naacl.96 | https://aclanthology.org/2022.findings-naacl.96.pdf | findings-naacl-2022-7 | ['machine-reading-comprehension'] | ['natural-language-processing'] | [ 4.36357260e-01 2.79712826e-01 -5.39969504e-02 -4.65488672e-01
-1.32524645e+00 -1.06228006e+00 2.23195225e-01 4.01086837e-01
-3.32560718e-01 1.00385535e+00 5.40883422e-01 -5.86733222e-01
1.32919755e-02 -9.15852904e-01 -8.91246021e-01 5.23032360e-02
5.40462315e-01 2.94378817e-01 6.85888886e-01 -5.71016610e-01
4.48234618e-01 4.09132754e-03 -1.41392529e+00 8.84966850e-01
1.41520560e+00 7.02650964e-01 1.96918800e-01 9.63549912e-01
-4.68626380e-01 1.42514658e+00 -7.93565631e-01 -5.51991045e-01
-1.16600037e-01 -9.51266706e-01 -1.63060963e+00 -6.16052747e-01
1.28161681e+00 -3.51958334e-01 -6.32478669e-02 9.08881128e-01
2.98936754e-01 3.81179541e-01 4.20538038e-01 -8.82101774e-01
-1.13564265e+00 5.62738836e-01 -1.53524548e-01 7.49678135e-01
1.22423780e+00 1.90204874e-01 1.42719305e+00 -8.75893056e-01
3.62970471e-01 1.56305110e+00 4.73373324e-01 6.48549914e-01
-1.06560445e+00 -3.80937308e-01 3.44043404e-01 9.80412841e-01
-6.25557184e-01 -2.26725087e-01 6.76381469e-01 3.40086371e-02
1.14564240e+00 8.69569242e-01 1.43517330e-01 9.81923640e-01
2.42895588e-01 8.72776628e-01 1.49233270e+00 -6.06781662e-01
2.77038813e-02 -2.47821420e-01 6.66439712e-01 7.14024186e-01
9.86508131e-02 -3.02509189e-01 -5.39392292e-01 1.82260558e-01
1.29368231e-01 -4.43316489e-01 -6.92331016e-01 2.47184590e-01
-9.72318590e-01 7.76229560e-01 5.25267124e-01 1.00754745e-01
-8.90571922e-02 3.38647771e-03 2.00396180e-01 1.10437143e+00
1.72400028e-01 1.28009772e+00 -8.70084167e-01 -2.74388820e-01
-6.15985155e-01 5.47853768e-01 1.00740099e+00 7.70874023e-01
6.16319001e-01 -3.37847978e-01 -7.79723048e-01 6.92720592e-01
-1.51257172e-01 5.63638210e-01 4.59517092e-01 -1.28185046e+00
7.54720867e-01 7.68555939e-01 1.65699720e-01 -1.14120054e+00
-3.44046414e-01 -3.37765247e-01 -3.97233456e-01 -3.43799829e-01
6.69223726e-01 -4.43934463e-02 -5.51091671e-01 1.81569183e+00
2.09590822e-01 -2.25615315e-02 1.37220398e-01 5.93484759e-01
1.36888778e+00 8.03608179e-01 1.10896274e-01 -1.90624401e-01
1.42846322e+00 -1.38272488e+00 -9.43840146e-01 -4.22805727e-01
5.95099330e-01 -6.28281355e-01 1.99288595e+00 3.57618511e-01
-1.52590919e+00 -7.59477139e-01 -8.61179173e-01 -6.94002450e-01
-1.83901727e-01 -4.35797781e-01 6.68915659e-02 1.66513070e-01
-1.03718960e+00 3.76019061e-01 -2.57455885e-01 -1.02114953e-01
2.12097257e-01 -1.49488747e-01 2.48356145e-02 -4.04169023e-01
-1.56525016e+00 1.31731451e+00 4.03403461e-01 -3.17153245e-01
-4.51070398e-01 -9.86139238e-01 -9.21330988e-01 4.04328734e-01
6.43548071e-01 -8.43019903e-01 1.88089383e+00 -8.23466122e-01
-1.35274184e+00 9.48955834e-01 -5.57792783e-01 -4.74865973e-01
4.29864287e-01 -5.56560695e-01 -4.01839852e-01 5.60193241e-01
8.62176940e-02 8.11020374e-01 6.84416533e-01 -9.78615820e-01
-5.92017353e-01 -4.06893454e-02 7.01118410e-01 3.40837687e-01
-2.15245038e-01 -7.09259361e-02 1.10894434e-01 -6.05071425e-01
2.34203458e-01 -4.47513551e-01 2.90264189e-01 -7.57136289e-03
-2.10562989e-01 -7.70047903e-01 4.71934110e-01 -1.24017584e+00
1.36071682e+00 -1.56098211e+00 8.62107351e-02 -4.68277603e-01
3.65727574e-01 2.65244067e-01 -5.81854463e-01 3.92592520e-01
1.09793410e-01 1.11594908e-01 -3.13763171e-01 2.56511033e-01
1.17926560e-01 1.14713870e-01 -6.92404568e-01 -3.34490448e-01
4.12984192e-01 1.38091612e+00 -1.15676200e+00 -3.32714140e-01
-2.10665658e-01 -3.94337595e-01 -7.35694349e-01 5.40529847e-01
-8.44599962e-01 4.12748069e-01 -1.29471168e-01 4.13643777e-01
6.62502110e-01 -3.77752870e-01 -2.13565439e-01 -8.83532502e-03
4.14391965e-01 1.04913580e+00 -6.41750693e-01 1.53314030e+00
-5.69129705e-01 6.49397254e-01 -3.91342521e-01 -7.50345647e-01
8.83871794e-01 1.30365500e-02 -5.12297392e-01 -1.28845346e+00
-2.59558648e-01 1.49106398e-01 1.35367021e-01 -9.64638591e-01
6.76584303e-01 -2.49147758e-01 -4.30770330e-02 4.49760586e-01
-1.15847595e-01 -5.51254451e-01 2.89247215e-01 1.93133116e-01
1.39064372e+00 -2.64953263e-02 3.84518862e-01 -2.50520080e-01
9.20426250e-01 1.19885042e-01 2.92106301e-01 9.83118832e-01
-3.83274883e-01 6.18923724e-01 5.38400412e-01 -1.70704320e-01
-4.62974489e-01 -1.49122632e+00 1.96670800e-01 1.34963608e+00
3.13051045e-01 -2.35258728e-01 -7.63684332e-01 -1.00235677e+00
-1.50362253e-01 1.31444657e+00 -3.27243060e-01 -5.07707119e-01
-1.01231658e+00 -9.76030678e-02 4.88135606e-01 6.30271137e-01
8.04219186e-01 -1.21220970e+00 -5.73548019e-01 1.99298784e-01
-8.94557357e-01 -1.10377073e+00 -5.02250969e-01 -1.68717012e-01
-9.79984283e-01 -1.15150023e+00 -3.78813475e-01 -1.16747546e+00
3.70977432e-01 4.11365300e-01 1.82593715e+00 4.32325423e-01
1.81507185e-01 4.45792198e-01 -6.61110699e-01 -3.00435364e-01
-5.55834055e-01 3.16582799e-01 -5.04498601e-01 -5.94339669e-01
6.82016253e-01 -3.38257223e-01 -6.74932420e-01 2.03022823e-01
-8.03263366e-01 3.57753299e-02 2.24172056e-01 6.88905180e-01
3.85670394e-01 -3.57310861e-01 1.34298313e+00 -8.36537302e-01
1.19267261e+00 -5.69951534e-01 -5.91382124e-02 6.13279998e-01
-5.33445179e-01 2.09067523e-01 9.45514917e-01 -4.16242510e-01
-1.30667138e+00 -7.50582874e-01 -4.34416831e-01 3.76762718e-01
-2.59564430e-01 5.16499698e-01 -1.88157037e-01 2.53110439e-01
9.69186485e-01 2.18612626e-01 -2.61153191e-01 -4.80462372e-01
6.11331761e-01 5.23840249e-01 7.64191568e-01 -7.96913087e-01
8.03766549e-01 2.58605164e-02 -4.01845455e-01 -4.08339769e-01
-1.79866636e+00 -4.62452829e-01 -2.43405744e-01 2.26145051e-02
6.93878412e-01 -6.56235933e-01 -5.09852529e-01 1.75633088e-01
-1.22262955e+00 -4.46127504e-01 -4.70319301e-01 3.79231088e-02
-4.79240656e-01 6.05128050e-01 -8.43393564e-01 -1.44788653e-01
-5.16396582e-01 -5.40353060e-01 5.53738773e-01 4.20285702e-01
-9.35864329e-01 -1.16458106e+00 1.56874344e-01 1.05573797e+00
5.53343058e-01 1.99266404e-01 1.53130865e+00 -7.18575954e-01
-4.77757663e-01 1.62848756e-01 -2.21995115e-01 6.20731533e-01
2.30072886e-01 -4.54812348e-01 -8.48272383e-01 -2.33995065e-01
4.39947397e-01 -8.94780278e-01 8.77243876e-01 -1.38260886e-01
1.37128568e+00 -6.88018799e-01 3.21034491e-01 1.00904405e-01
1.08427262e+00 -9.68496352e-02 8.38038146e-01 2.61173844e-01
4.03471082e-01 6.72046661e-01 6.36450410e-01 -2.14111060e-01
7.19692707e-01 1.47294134e-01 2.96658397e-01 1.78390592e-01
-6.19867027e-01 -6.21208668e-01 3.05179536e-01 1.08451402e+00
4.79312748e-01 -3.00454557e-01 -7.33836412e-01 6.59490287e-01
-1.36544967e+00 -1.04954171e+00 -7.41106451e-01 1.83658671e+00
1.57567787e+00 1.90555841e-01 -3.58366698e-01 2.25282252e-01
4.20451880e-01 2.80839652e-01 -8.15818369e-01 -7.91436553e-01
-2.38129705e-01 7.21761465e-01 -3.87469590e-01 8.63111496e-01
-6.82507813e-01 1.03176212e+00 6.06103086e+00 6.91144705e-01
-6.33450329e-01 8.23381320e-02 4.49881285e-01 1.18650854e-01
-6.90983951e-01 -3.80414128e-02 -4.75851029e-01 3.50030988e-01
8.69312346e-01 -2.72606432e-01 2.59460092e-01 4.41692203e-01
-3.46588567e-02 -2.51107246e-01 -1.50343704e+00 7.03872979e-01
3.69303614e-01 -9.64685321e-01 4.58106041e-01 -9.42111790e-01
8.77103567e-01 -4.73775595e-01 1.18568376e-01 8.05409729e-01
8.70274529e-02 -1.35456777e+00 3.80302668e-01 5.93942881e-01
6.77543342e-01 -5.53229809e-01 5.99585891e-01 5.68881452e-01
-7.09847867e-01 -7.50478134e-02 -4.76241261e-01 -8.72993767e-01
1.12214148e-01 1.55181408e-01 -6.06693208e-01 2.35133335e-01
6.34015620e-01 3.68764430e-01 -1.10397995e+00 9.89642084e-01
-1.05140269e+00 1.07454848e+00 1.82521150e-01 -3.26910943e-01
1.41826689e-01 2.25348249e-01 2.90891021e-01 9.60462391e-01
8.43026936e-02 4.04763073e-01 -1.93992585e-01 9.27465141e-01
-4.84650820e-01 5.66619001e-02 -1.23722464e-01 2.88993329e-01
6.70239687e-01 8.73139858e-01 -9.81097445e-02 -4.36206847e-01
-4.28295255e-01 1.27175152e+00 7.49853075e-01 3.88256282e-01
-7.48599052e-01 -6.56959713e-01 3.65098089e-01 2.58206483e-02
4.04474176e-02 1.19938076e-01 -4.28660154e-01 -1.30803311e+00
4.41077024e-01 -1.42960477e+00 6.68165326e-01 -1.14012742e+00
-1.56944895e+00 2.52824426e-01 -1.10698573e-01 -7.27245152e-01
-7.04151094e-02 -6.16632342e-01 -8.76679361e-01 8.23594153e-01
-1.89071929e+00 -5.96926093e-01 -5.21687150e-01 3.25876027e-01
1.00160873e+00 4.35447276e-01 7.85379171e-01 -1.84330016e-01
-5.77367358e-02 7.81519055e-01 -1.92004547e-01 -1.44394562e-01
8.82855833e-01 -1.62683070e+00 4.68086749e-01 8.42043161e-01
2.28602849e-02 5.39201140e-01 8.78949523e-01 -3.66447389e-01
-1.06808257e+00 -9.93758440e-01 1.48182583e+00 -9.24172044e-01
5.97722471e-01 -1.25675313e-02 -1.58414602e+00 6.86752796e-01
7.24706709e-01 -5.62948763e-01 7.60597527e-01 8.14344510e-02
-7.42300272e-01 -1.25928104e-01 -1.20390046e+00 8.46408486e-01
1.11429882e+00 -7.44421005e-01 -1.81932974e+00 5.18909872e-01
1.36650872e+00 -4.71646011e-01 -6.29398286e-01 4.91504878e-01
1.62171796e-02 -8.89145017e-01 8.63664627e-01 -1.09322989e+00
8.04530382e-01 -7.46453479e-02 -8.86430591e-02 -1.37381506e+00
-3.89041692e-01 -4.25795346e-01 -5.32605827e-01 1.05302000e+00
4.91611898e-01 -6.00255907e-01 4.24172997e-01 6.02063179e-01
-1.96665660e-01 -6.86499000e-01 -8.61938179e-01 -8.51469398e-01
9.19793546e-01 -1.13954678e-01 6.15625858e-01 9.77386594e-01
2.52278060e-01 9.25217509e-01 1.84835389e-01 -1.01954127e-02
1.90740332e-01 3.65323871e-01 4.55279917e-01 -1.08791625e+00
-3.57232988e-01 -3.79184306e-01 2.81306326e-01 -1.74061263e+00
3.71409506e-01 -1.01927626e+00 6.68989792e-02 -1.86522079e+00
1.73531771e-01 6.34884015e-02 -1.32851422e-01 2.00905621e-01
-9.92008030e-01 -3.27264331e-02 2.53812969e-01 -1.79673985e-01
-9.80440497e-01 4.84176099e-01 1.75882530e+00 -2.54950583e-01
1.35731399e-01 -9.65767503e-02 -1.05520797e+00 7.45103776e-01
1.26272845e+00 -1.58439517e-01 -6.98014975e-01 -8.78877342e-01
5.47656178e-01 6.34620190e-02 3.33524048e-01 -8.71771276e-01
7.76739046e-02 -2.59245366e-01 1.59584016e-01 -6.17054284e-01
1.59855075e-02 -2.96228230e-01 -7.50204384e-01 6.20481789e-01
-9.74978268e-01 4.86627012e-01 2.83625692e-01 3.65213871e-01
-3.86264324e-01 -6.47129238e-01 5.97002566e-01 -2.59064317e-01
-5.75210869e-01 -2.93942571e-01 -2.92538524e-01 1.17017651e+00
6.29945993e-01 8.38914812e-02 -8.14637959e-01 -6.69277966e-01
-4.41349357e-01 6.38975084e-01 1.14974141e-01 5.01847506e-01
8.21607709e-01 -1.11432910e+00 -1.09771526e+00 -3.40150177e-01
1.54854432e-01 1.00396082e-01 4.25726771e-01 4.11640823e-01
-6.96305037e-01 6.27499938e-01 5.93596064e-02 -2.19185621e-01
-1.25276315e+00 6.44761145e-01 4.33647007e-01 -4.57041085e-01
-3.58091295e-01 1.06920028e+00 2.14795247e-02 -8.98282230e-01
3.07846785e-01 -9.20212269e-01 -3.62736583e-01 3.00934277e-02
9.15278137e-01 5.93059897e-01 1.03556514e-01 7.53307566e-02
7.09846765e-02 5.84165037e-01 -2.45551750e-01 2.08877265e-01
7.68903673e-01 -4.14678544e-01 -2.62796015e-01 3.69087487e-01
1.16691458e+00 2.12062765e-02 -9.78757799e-01 -4.43486005e-01
2.51340508e-01 -2.85745680e-01 -6.01109326e-01 -1.46689725e+00
-2.79039413e-01 9.92207646e-01 1.37675494e-01 2.48012513e-01
1.25261879e+00 1.31214797e-01 1.40873659e+00 1.06066310e+00
1.73430949e-01 -9.53873754e-01 5.99180162e-01 9.61342335e-01
1.39654076e+00 -1.32938170e+00 -3.84656370e-01 -5.93676329e-01
-2.71998614e-01 1.00123870e+00 1.26872623e+00 4.55375724e-02
-1.95523456e-01 -3.45687509e-01 1.78902492e-01 -5.68053219e-03
-1.09758723e+00 2.87735146e-02 3.45817655e-01 3.67490768e-01
3.87803257e-01 -1.49039924e-01 -5.35046935e-01 6.14455402e-01
-7.29778409e-01 -2.66029954e-01 7.08570421e-01 8.77484560e-01
-9.37290370e-01 -7.68441260e-01 -4.64455962e-01 5.19208491e-01
-3.83293152e-01 -4.69307899e-01 -6.45174801e-01 4.91178304e-01
-2.33336404e-01 1.48992455e+00 5.47735281e-02 4.50413004e-02
5.25775969e-01 3.67282391e-01 7.92203724e-01 -7.57383287e-01
-7.63084829e-01 -9.85270500e-01 2.77818948e-01 -4.11555171e-01
-4.70674746e-02 -2.97270268e-01 -1.53422523e+00 -2.53050983e-01
-1.21751614e-01 2.61454195e-01 -2.59949535e-01 1.05488670e+00
2.88947284e-01 6.46883011e-01 3.87696117e-01 4.31495905e-01
-1.15965426e+00 -1.03964758e+00 8.09714496e-02 7.57739186e-01
5.78082025e-01 -9.63997021e-02 -6.18104219e-01 3.16110365e-02] | [11.314651489257812, 8.076757431030273] |
9051f57a-8bb1-420d-bfb3-aca581595907 | focal-visual-text-attention-for-memex | null | null | https://ieeexplore.ieee.org/document/8603827 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8603827 | Focal Visual-Text Attention for Memex Question Answering | Recent insights on language and vision with neural networks have been successfully applied to simple single-image visual question answering. However, to tackle real-life question answering problems on multimedia collections such as personal photo albums, we have to look at whole collections with sequences of photos. This paper proposes a new multimodal MemexQA task: given a sequence of photos from a user, the goal is to automatically answer questions that help users recover their memory about an event captured in these photos. In addition to a text answer, a few grounding photos are also given to justify the answer. The grounding photos are necessary as they help users quickly verifying the answer. Towards solving the task, we 1) present the MemexQA dataset, the first publicly available multimodal question answering dataset consisting of real personal photo albums; 2) propose an end-to-end trainable network that makes use of a hierarchical process to dynamically determine what media and what time to focus on in the sequential data to answer the question. Experimental results on the MemexQA dataset demonstrate that our model outperforms strong baselines and yields the most relevant grounding photos on this challenging task. | ['Li-Jia Li', 'Yannis Kalantidis', 'Junwei Liang', 'and Alexander Hauptmann', 'Lu Jiang', 'Liangliang Cao'] | 2018-12-14 | null | null | null | ieee-transactions-on-pattern-analysis-and | ['memex-question-answering'] | ['natural-language-processing'] | [ 2.36788496e-01 3.23381759e-02 1.86514556e-01 -5.34390211e-01
-1.22351527e+00 -7.05088139e-01 6.01773739e-01 4.22799796e-01
-6.69143736e-01 5.10409176e-01 3.45644921e-01 -2.30052799e-01
2.09389910e-01 -4.81295437e-01 -1.07228851e+00 -2.55317003e-01
2.78671026e-01 7.33169377e-01 3.96183550e-01 -3.42684835e-01
2.90623248e-01 -1.35143921e-01 -1.69252920e+00 9.61922169e-01
4.50731218e-01 1.24307263e+00 3.67247492e-01 1.02928984e+00
-1.69669449e-01 1.31665730e+00 -3.97537023e-01 -6.57397568e-01
-9.73363989e-04 -5.00259995e-01 -1.40541613e+00 4.80286926e-01
1.30165529e+00 -9.12679791e-01 -3.45887035e-01 8.29337001e-01
4.04544652e-01 3.20836723e-01 4.74333882e-01 -1.42404950e+00
-9.98199463e-01 4.79095668e-01 -3.04816842e-01 4.74674255e-01
1.04268420e+00 4.17077452e-01 1.18611431e+00 -1.07377088e+00
5.91768980e-01 1.25830591e+00 3.00211668e-01 7.42883742e-01
-8.71487319e-01 -2.32674062e-01 2.01208696e-01 7.24055231e-01
-1.27087116e+00 -5.62197924e-01 5.71343422e-01 -2.55957782e-01
7.10642874e-01 4.02793854e-01 3.40476811e-01 1.14820814e+00
-4.16291118e-01 1.14864230e+00 8.24805200e-01 -3.48600328e-01
9.27356854e-02 1.88076288e-01 2.61778265e-01 1.04760265e+00
-3.95950228e-01 -6.76038206e-01 -8.10862243e-01 -1.98014170e-01
2.27546379e-01 7.22523332e-02 -4.37975168e-01 -6.20163493e-02
-1.34647667e+00 5.01407385e-01 7.49059200e-01 1.76031739e-01
-5.42624414e-01 2.83256710e-01 1.84081718e-01 2.24342972e-01
5.11831567e-02 3.46096158e-01 -1.25656873e-01 1.92947581e-01
-8.41167390e-01 3.97142619e-01 7.60044158e-01 8.55748177e-01
1.02630234e+00 -6.91230357e-01 -4.81005520e-01 6.69099033e-01
1.81427240e-01 6.07959747e-01 1.78358555e-01 -1.14768803e+00
8.59063804e-01 9.57742929e-01 5.50851047e-01 -1.05489111e+00
-1.32641077e-01 5.37196398e-01 -6.43963575e-01 -5.55631042e-01
7.02471077e-01 1.07544065e-01 -8.93816113e-01 1.42170537e+00
4.59326863e-01 2.02583775e-01 2.43020873e-03 1.21346056e+00
1.31012642e+00 8.80595148e-01 2.18753353e-01 2.01805845e-01
1.80770314e+00 -1.13659930e+00 -5.24000883e-01 -5.06817102e-01
2.75141329e-01 -7.62998402e-01 1.45995009e+00 1.37048304e-01
-1.36251926e+00 -6.84142292e-01 -6.43926203e-01 -6.57859921e-01
-4.81497973e-01 2.70273387e-01 1.75935263e-03 4.40069623e-02
-1.41333830e+00 9.76053327e-02 -2.64719546e-01 -7.35361695e-01
4.03732985e-01 5.98064587e-02 -3.56150806e-01 -6.07945442e-01
-1.07261157e+00 6.89455509e-01 3.17755640e-01 3.74734402e-01
-1.32862151e+00 -4.20746803e-01 -8.05021524e-01 1.37791291e-01
6.10942602e-01 -8.77358556e-01 1.55308998e+00 -1.25243306e+00
-9.41558242e-01 1.42968071e+00 -4.25755739e-01 -4.40114528e-01
3.08261305e-01 -2.63384312e-01 -7.80060589e-02 9.66052711e-01
2.77382880e-01 1.18860424e+00 1.02105975e+00 -1.38798094e+00
-7.12950528e-01 -3.44219327e-01 7.64260352e-01 1.88717633e-01
-7.28977680e-01 4.90185544e-02 -8.92310858e-01 8.95077586e-02
-2.22608000e-01 -9.57914531e-01 7.93541595e-02 1.10954307e-02
-4.31337416e-01 -5.47283947e-01 7.51894891e-01 -1.02222848e+00
8.39208722e-01 -1.79307973e+00 2.64512300e-01 -1.81619480e-01
2.03685209e-01 1.99895263e-01 -5.20379901e-01 7.30665982e-01
2.62859017e-01 -1.64419726e-01 -1.54533848e-01 -5.15362561e-01
2.19168626e-02 1.55240968e-01 -7.40029752e-01 2.25614812e-02
3.28163326e-01 1.24388671e+00 -1.08509910e+00 -8.10364306e-01
-9.28668678e-02 2.34911337e-01 -1.83322608e-01 5.81686318e-01
-8.96388531e-01 1.89522192e-01 -4.43158776e-01 8.04237425e-01
4.53474969e-01 -8.34647715e-01 -1.86192930e-01 -3.08214128e-01
2.51905650e-01 -1.24815121e-01 -8.32099795e-01 1.76951098e+00
-2.00339824e-01 7.74195969e-01 4.85944450e-02 -6.58125579e-01
4.88156438e-01 2.80997545e-01 1.54469043e-01 -8.79332721e-01
8.55062902e-02 -1.62763119e-01 -7.76010334e-01 -1.18437934e+00
8.01809728e-01 2.20203206e-01 -1.03389591e-01 7.65180409e-01
2.67380215e-02 -1.40870884e-01 5.03014982e-01 7.41659343e-01
9.85472679e-01 -2.50771921e-02 -1.30203828e-01 2.96257257e-01
8.99247229e-01 2.71650463e-01 -2.61211663e-01 8.05032611e-01
-2.35636264e-01 8.26048791e-01 2.84616768e-01 -7.31051445e-01
-9.85487461e-01 -7.80611098e-01 6.24724507e-01 1.40446460e+00
4.03125018e-01 -2.65914977e-01 -8.60551655e-01 -7.32555985e-01
-2.25182503e-01 5.80693126e-01 -7.58599699e-01 2.74528742e-01
-6.02775395e-01 -2.48007774e-01 4.17914182e-01 3.56998622e-01
9.01771128e-01 -1.22691154e+00 -6.28982008e-01 -7.97183886e-02
-1.15475118e+00 -1.55117762e+00 -8.71447682e-01 -7.12838709e-01
-3.91174257e-01 -1.49779069e+00 -9.37341273e-01 -1.08236396e+00
6.77296400e-01 8.13514054e-01 1.53041208e+00 7.02624023e-01
-2.05803260e-01 1.47832298e+00 -4.85422999e-01 -3.02338481e-01
-2.38595128e-01 -1.00232542e-01 -4.22095656e-01 6.69422984e-01
4.99824762e-01 -1.12989545e-01 -1.01181269e+00 4.44738388e-01
-1.22411501e+00 -6.11727089e-02 2.91024536e-01 4.69110310e-01
5.67492604e-01 -4.74694520e-01 4.56954747e-01 -3.81253868e-01
6.49933934e-01 -5.67309082e-01 -4.34823066e-01 9.11671042e-01
-5.14216311e-02 -1.43633217e-01 3.00093681e-01 -5.18667698e-01
-9.12637711e-01 2.61087269e-01 3.20543908e-02 -3.84013116e-01
-7.72818848e-02 2.76371419e-01 2.57416904e-01 1.42635822e-01
6.58278823e-01 4.54411358e-01 -1.88927025e-01 -6.24326020e-02
5.19396722e-01 5.58231235e-01 8.17700148e-01 -5.46208501e-01
6.20510221e-01 8.00287127e-01 -3.10133487e-01 -8.95149946e-01
-1.46959233e+00 -1.00870442e+00 -5.77469230e-01 -7.21281886e-01
1.24495292e+00 -9.89946306e-01 -1.34111834e+00 1.14628926e-01
-1.37574744e+00 -4.36765045e-01 7.32425153e-02 -3.57453585e-01
-6.74016178e-01 6.50013030e-01 -4.78406101e-01 -9.32890713e-01
-5.00705838e-01 -8.31050754e-01 1.55373394e+00 3.13523799e-01
-6.95075318e-02 -7.29654074e-01 -2.92156544e-02 1.15515471e+00
7.03412062e-03 1.19247362e-01 7.43785322e-01 -3.56926143e-01
-1.09982407e+00 -1.73367113e-01 -4.95872587e-01 1.95180491e-01
-4.69242722e-01 -3.23655635e-01 -1.10045481e+00 -3.06535333e-01
-2.98902929e-01 -1.06890535e+00 1.01598191e+00 -3.88158374e-02
1.30340290e+00 -4.51555252e-01 -1.23718344e-01 -6.53810427e-02
1.27614498e+00 -3.59055698e-01 7.44116724e-01 2.10338965e-01
5.83602428e-01 9.55290973e-01 5.27248442e-01 1.43387213e-01
1.10136795e+00 3.74029905e-01 8.43505681e-01 -1.19430222e-01
-1.89200595e-01 -4.51500624e-01 3.05726528e-01 3.19651812e-01
1.68219924e-01 -4.90904123e-01 -1.02037191e+00 9.50470626e-01
-2.15279174e+00 -1.14412403e+00 -2.07237110e-01 1.93133092e+00
6.04206979e-01 -5.59627593e-01 2.03473330e-01 -1.53594688e-01
5.69321156e-01 2.61506379e-01 -4.95134056e-01 2.22478986e-01
-1.85504884e-01 -3.94332945e-01 -5.28372526e-02 4.26542342e-01
-1.07657623e+00 8.60805631e-01 5.60419512e+00 2.35356182e-01
-6.90932393e-01 2.20766038e-01 6.85401201e-01 -4.19622734e-02
-2.03276187e-01 -9.01637226e-02 -8.31566870e-01 2.29030013e-01
7.82045066e-01 2.60559797e-01 5.03723979e-01 3.32548857e-01
2.47263480e-02 -5.35083055e-01 -1.50647235e+00 1.30010998e+00
7.91514158e-01 -1.47332156e+00 4.58412915e-01 -4.73014325e-01
7.24608183e-01 -1.09057367e-01 8.91237408e-02 2.23941952e-01
1.01160277e-02 -9.40261066e-01 7.80609310e-01 7.93572843e-01
5.26764691e-01 -3.28361422e-01 4.76541698e-01 5.07524848e-01
-1.06598771e+00 -3.15766007e-01 -3.61333072e-01 4.90536354e-02
4.68484521e-01 -8.92691966e-03 -1.30031908e+00 1.32593215e-01
1.06699431e+00 3.89597058e-01 -1.05241311e+00 8.65155697e-01
-3.29447538e-01 3.54309797e-01 -9.13868472e-02 -2.92078853e-01
4.36618894e-01 2.53384531e-01 1.95091501e-01 8.84441912e-01
2.14295447e-01 5.16568959e-01 1.03638545e-01 8.30795288e-01
-5.53892672e-01 5.23115769e-02 -4.06208873e-01 -1.28133446e-01
1.69820279e-01 1.34797883e+00 -5.55016220e-01 -4.77225959e-01
-4.54794943e-01 1.43266428e+00 5.42510152e-01 7.13728249e-01
-5.04391611e-01 -4.50487901e-03 8.48298892e-02 2.61948258e-01
2.41675511e-01 -1.04767740e-01 5.13752043e-01 -1.22300100e+00
1.80237621e-01 -9.50239301e-01 1.01913321e+00 -1.72998893e+00
-1.32165432e+00 6.28209174e-01 -1.67733893e-01 -8.98478985e-01
-2.33488157e-01 -6.12041652e-01 -4.35453236e-01 5.19505143e-01
-1.80952597e+00 -1.50896239e+00 -9.16789532e-01 1.15392971e+00
8.16860378e-01 1.39688209e-01 6.55113459e-01 3.82320136e-01
-2.36580923e-01 2.29056031e-01 -5.19211650e-01 3.22525561e-01
8.70954216e-01 -1.15711343e+00 5.06065972e-02 8.09437037e-01
4.64044511e-01 4.28038895e-01 7.04372108e-01 -4.82858449e-01
-1.57255304e+00 -9.64714408e-01 1.25348055e+00 -1.23243320e+00
6.63522184e-01 -3.07263672e-01 -9.88597035e-01 9.00313199e-01
8.18641722e-01 -2.49596968e-01 5.26304424e-01 -1.37056187e-01
-4.70767170e-01 -1.89785466e-01 -8.19794655e-01 6.83633804e-01
6.20888174e-01 -8.97709846e-01 -8.52318704e-01 9.02302027e-01
7.26105273e-01 -3.46462905e-01 -4.13338929e-01 -3.46405990e-03
3.19606423e-01 -7.21489847e-01 1.18235648e+00 -8.35059702e-01
6.49113357e-01 -4.77173388e-01 -1.46954358e-01 -7.52131104e-01
4.06397641e-01 -5.38096786e-01 -3.23407799e-02 1.22431147e+00
3.93195540e-01 1.21083461e-01 6.80581927e-01 8.79974961e-01
2.73060292e-01 -2.59466499e-01 -7.77106285e-01 -6.95381314e-02
-5.44160187e-01 -2.10430652e-01 5.24590254e-01 6.56848967e-01
-2.17280567e-01 7.22874641e-01 -6.64557576e-01 4.98373598e-01
5.74538469e-01 2.78429836e-01 1.04753220e+00 -9.47184920e-01
-3.09389122e-02 -1.47272344e-03 8.80605076e-03 -1.28192401e+00
1.33352697e-01 -6.62491560e-01 3.30122054e-01 -1.86117506e+00
7.75685906e-01 2.36878455e-01 3.13177565e-03 4.80169445e-01
-4.25150543e-01 6.42003715e-01 3.30543101e-01 2.81471759e-01
-1.44905066e+00 2.86039233e-01 9.89812136e-01 -5.65157950e-01
4.68485355e-02 -2.01574251e-01 -5.33563197e-01 6.10650539e-01
3.57183188e-01 -8.95470977e-02 -4.80012387e-01 -8.69448602e-01
9.00593042e-01 3.55772287e-01 1.10131407e+00 -7.34669864e-01
6.65795922e-01 1.65625848e-02 2.79349476e-01 -1.01075387e+00
6.29892409e-01 -8.09858561e-01 -5.08337915e-01 6.70068040e-02
-6.62840307e-01 3.44647020e-01 1.74896106e-01 8.30762327e-01
-3.69350940e-01 -2.83850610e-01 3.31375420e-01 -3.87741417e-01
-1.15147090e+00 4.85634238e-01 -2.24062398e-01 1.84950233e-01
9.95056570e-01 1.75413117e-01 -5.85590184e-01 -9.97545183e-01
-7.25927353e-01 8.75707269e-01 1.33853540e-01 5.37746668e-01
1.11578453e+00 -1.27536047e+00 -8.76640260e-01 -6.32255256e-01
6.36806071e-01 -1.69459790e-01 6.37421012e-01 4.99570847e-01
-4.66388136e-01 4.28448081e-01 2.95615811e-02 -7.56571054e-01
-1.50248659e+00 8.61715317e-01 2.97138035e-01 1.10949412e-01
-3.02144289e-01 1.08100247e+00 1.47478849e-01 -2.12607116e-01
3.83907646e-01 -1.19256027e-01 -4.37518209e-01 4.69526380e-01
1.10342073e+00 1.09720930e-01 -4.41904925e-02 -7.10769057e-01
-2.18411162e-01 5.52481234e-01 1.48810744e-01 -2.34742954e-01
1.17972791e+00 -7.79162645e-01 -2.42765084e-01 3.08486611e-01
1.06706047e+00 -4.84214425e-01 -1.27081096e+00 -5.63796163e-01
-3.67906578e-02 -2.90415466e-01 -3.60913873e-01 -8.27225983e-01
-5.78412473e-01 1.15448713e+00 4.43991989e-01 3.37165207e-01
1.21202397e+00 4.35693443e-01 1.19874966e+00 8.50478113e-01
7.28014484e-02 -1.00277543e+00 9.74037111e-01 3.49628597e-01
1.23203564e+00 -1.71178603e+00 -2.39281654e-01 1.33986056e-01
-7.96141326e-01 1.06630504e+00 6.98686361e-01 1.34503573e-01
1.38478488e-01 -6.86155796e-01 2.26105183e-01 -4.65431631e-01
-7.86023021e-01 -5.27610898e-01 6.50171280e-01 2.99652129e-01
-2.14213818e-01 -2.98107535e-01 3.20133567e-01 1.26544744e-01
1.57907277e-01 -7.82946348e-02 5.88307321e-01 5.99044621e-01
-6.14955485e-01 -6.07311785e-01 -5.86217284e-01 6.23851046e-02
-1.98898122e-01 -1.41931325e-01 -8.94745588e-01 5.54553807e-01
-2.35408843e-01 1.42723548e+00 4.73667234e-02 -1.02057196e-01
3.48328441e-01 2.63777584e-01 4.39235806e-01 -4.02107298e-01
-5.56295574e-01 -5.68701267e-01 1.12000637e-01 -6.37635231e-01
-7.08821654e-01 -4.56030339e-01 -1.03278887e+00 -1.77823469e-01
2.40711197e-01 6.65028319e-02 4.87296581e-01 1.03418636e+00
2.75251508e-01 1.54696316e-01 3.19024682e-01 -7.87943244e-01
-2.33590633e-01 -7.75317967e-01 1.70953259e-01 8.42267454e-01
7.15481102e-01 3.57470065e-02 -7.27477744e-02 6.86819255e-01] | [10.632328987121582, 1.3471585512161255] |
245235fc-92b9-4643-812a-c91ce8be6a02 | generating-large-labeled-data-sets-for | 1907.02882 | null | https://arxiv.org/abs/1907.02882v1 | https://arxiv.org/pdf/1907.02882v1.pdf | Generating large labeled data sets for laparoscopic image processing tasks using unpaired image-to-image translation | In the medical domain, the lack of large training data sets and benchmarks is often a limiting factor for training deep neural networks. In contrast to expensive manual labeling, computer simulations can generate large and fully labeled data sets with a minimum of manual effort. However, models that are trained on simulated data usually do not translate well to real scenarios. To bridge the domain gap between simulated and real laparoscopic images, we exploit recent advances in unpaired image-to-image translation. We extent an image-to-image translation method to generate a diverse multitude of realistically looking synthetic images based on images from a simple laparoscopy simulation. By incorporating means to ensure that the image content is preserved during the translation process, we ensure that the labels given for the simulated images remain valid for their realistically looking translations. This way, we are able to generate a large, fully labeled synthetic data set of laparoscopic images with realistic appearance. We show that this data set can be used to train models for the task of liver segmentation of laparoscopic images. We achieve average dice scores of up to 0.89 in some patients without manually labeling a single laparoscopic image and show that using our synthetic data to pre-train models can greatly improve their performance. The synthetic data set will be made publicly available, fully labeled with segmentation maps, depth maps, normal maps, and positions of tools and camera (http://opencas.dkfz.de/image2image). | ['Stefanie Speidel', 'Jürgen Weitz', 'Lena Maier-Hein', 'Kurinchi Gurusamy', 'Tobias Roß', 'Sandy Engelhardt', 'Matthew J. Clarkson', 'Carina Riediger', 'Micha Pfeiffer', 'Sebastian Bodenstedt', 'Maria R. Robu', 'Isabel Funke', 'Thilo Welsch', 'Leon Strenger', 'Brian R. Davidson'] | 2019-07-05 | null | null | null | null | ['liver-segmentation'] | ['medical'] | [ 1.78048924e-01 3.96517545e-01 1.11500651e-01 -5.80592930e-01
-7.97256231e-01 -8.60526443e-01 3.53925824e-01 9.00195241e-02
-3.81790161e-01 7.01269805e-01 -6.29146472e-02 -4.12085772e-01
3.61714333e-01 -6.18127227e-01 -1.02408731e+00 -4.45758462e-01
1.35616064e-01 7.88291693e-01 -6.86702430e-02 3.38901132e-02
-1.76402450e-01 5.16385972e-01 -1.02444506e+00 2.10487962e-01
8.12186122e-01 5.83437145e-01 2.72933483e-01 7.51141846e-01
2.92861372e-01 2.83783704e-01 -3.70070636e-01 -2.95704305e-01
7.43196547e-01 -8.80190730e-01 -6.99263155e-01 2.32506275e-01
4.00928557e-01 -5.57841539e-01 -1.14471540e-01 9.07941103e-01
6.27313614e-01 -1.43088490e-01 5.71614683e-01 -9.96619463e-01
-1.14404865e-01 4.15724337e-01 -3.48956436e-01 -2.19120473e-01
1.32499173e-01 4.41933572e-01 9.16966721e-02 -5.06359100e-01
1.13668346e+00 7.77860165e-01 7.60044992e-01 6.10237658e-01
-1.33747697e+00 -6.08453572e-01 -6.88647330e-01 -6.51296139e-01
-1.02870858e+00 -3.12797874e-01 5.34806848e-01 -5.57026803e-01
2.81408638e-01 1.26789376e-01 1.02851319e+00 9.61849034e-01
4.41334099e-01 4.38551664e-01 1.14701903e+00 -4.48949695e-01
-1.15459831e-03 3.06844801e-01 -5.38052738e-01 1.01848745e+00
3.38840127e-01 4.05122012e-01 1.19066328e-01 1.10989638e-01
1.18730378e+00 1.97749957e-02 -4.81355846e-01 -9.99525070e-01
-1.70489526e+00 7.33884215e-01 6.60079122e-01 1.83353350e-01
-1.93628147e-01 2.37801954e-01 4.19470638e-01 1.98622048e-01
2.29394920e-02 9.48971689e-01 -2.28249252e-01 8.52809176e-02
-9.15906489e-01 1.53120965e-01 8.77463996e-01 9.17647779e-01
5.82286596e-01 -7.49245957e-02 5.42948805e-02 5.57728171e-01
-4.75236885e-02 3.39011401e-01 6.75651193e-01 -1.50593913e+00
-2.54902970e-02 4.96578097e-01 2.94335991e-01 -8.61914098e-01
-4.88124162e-01 -6.24445856e-01 -7.13870347e-01 4.36707854e-01
7.81514883e-01 -2.40913630e-01 -1.13814414e+00 1.58815587e+00
3.72028947e-01 -8.15489069e-02 1.43404096e-01 9.20716286e-01
7.90604293e-01 3.01922411e-01 -8.16071555e-02 -4.38909270e-02
9.82663393e-01 -1.01685178e+00 -2.94271469e-01 -1.38805419e-01
1.08307350e+00 -8.65561128e-01 1.03905344e+00 3.07608724e-01
-1.47708380e+00 -3.40883434e-01 -1.05812371e+00 8.68272334e-02
-6.41071275e-02 1.82684183e-01 5.04187167e-01 5.07615149e-01
-1.25102866e+00 6.18497670e-01 -1.06763589e+00 -3.05327386e-01
4.70674604e-01 4.64149356e-01 -6.94743872e-01 -2.46244252e-01
-8.02814543e-01 1.03045022e+00 4.44949776e-01 -1.54916150e-02
-1.28161478e+00 -7.82182693e-01 -1.17461956e+00 -1.87544793e-01
-8.66992958e-03 -7.94734597e-01 1.56103170e+00 -1.26214349e+00
-1.24072981e+00 1.31032801e+00 3.24927449e-01 -3.48078519e-01
1.10003316e+00 5.36235154e-01 2.43212029e-01 3.59629840e-01
1.40866548e-01 1.22475564e+00 4.15853530e-01 -1.65207291e+00
-1.27388477e-01 -5.65131865e-02 -1.49380714e-01 2.02584043e-01
1.73659176e-01 -2.05890656e-01 -5.31726897e-01 -4.38261777e-01
1.62923858e-02 -1.27190030e+00 -5.56036413e-01 4.87866610e-01
-1.33695528e-01 8.05515051e-01 3.85902286e-01 -7.14447021e-01
2.23235980e-01 -1.96147323e+00 -1.30229607e-01 6.40428811e-02
1.08721770e-01 2.65608758e-01 -2.52768666e-01 5.06933406e-02
-2.44297817e-01 -3.09786461e-02 -5.31515479e-01 -3.45989555e-01
-3.81870389e-01 2.62305290e-01 1.28535673e-01 6.95313334e-01
-2.34192938e-01 1.23600018e+00 -1.20807493e+00 -7.51862049e-01
6.14099145e-01 5.28766870e-01 -6.77523911e-01 3.16080987e-01
-4.06147428e-02 1.04472899e+00 -1.81276966e-02 2.18239024e-01
5.76845706e-01 -2.99667895e-01 2.47534558e-01 -1.88027874e-01
2.21886441e-01 -1.65049896e-01 -8.10968816e-01 2.03864074e+00
-8.06697130e-01 6.00473046e-01 2.32270718e-01 -8.16311896e-01
5.92805624e-01 5.00678301e-01 6.57464623e-01 -6.58520222e-01
3.28599006e-01 4.96936440e-01 3.26786160e-01 -3.29382539e-01
5.07280463e-04 -5.74346960e-01 -9.72335637e-02 5.57150722e-01
1.79990813e-01 -1.02164745e+00 3.58914405e-01 -4.46558511e-03
5.93712926e-01 8.72572660e-02 -4.17771004e-02 -2.91390985e-01
8.58100206e-02 6.04866922e-01 2.70922035e-01 5.28062820e-01
-2.59015471e-01 1.18118596e+00 4.70247239e-01 -6.35317504e-01
-1.74247575e+00 -1.05318618e+00 -1.23936795e-01 1.69440255e-01
3.51306140e-01 1.40959293e-01 -1.01772988e+00 -7.07388878e-01
-2.74129540e-01 5.32354653e-01 -6.47187412e-01 -1.90231994e-01
-5.81676781e-01 -7.40412831e-01 5.27930915e-01 2.09290877e-01
2.21196532e-01 -1.14525890e+00 -8.51746559e-01 1.95919663e-01
-1.83035478e-01 -1.06040144e+00 -2.91148603e-01 1.50505126e-01
-9.33488488e-01 -1.13640392e+00 -1.12495768e+00 -1.19689560e+00
1.25041056e+00 5.33697270e-02 1.37167215e+00 1.64723828e-01
-6.08775795e-01 7.88382664e-02 -4.11600759e-03 -2.75134206e-01
-1.08896792e+00 -1.19013615e-01 -2.30749413e-01 -6.47751570e-01
-5.89308560e-01 -2.98760027e-01 -1.13142502e+00 4.56685930e-01
-1.04179502e+00 6.48084521e-01 6.60513461e-01 1.01421392e+00
6.85982406e-01 -2.88760811e-01 8.43268484e-02 -9.80729640e-01
3.64844382e-01 -1.25826702e-01 -6.52818322e-01 -4.88877483e-03
-4.36620444e-01 2.03016289e-02 7.10902393e-01 -4.25065368e-01
-8.75122428e-01 3.65361989e-01 -1.06446758e-01 -4.58481520e-01
-2.72900701e-01 3.19479138e-01 3.55976522e-01 -3.80272180e-01
8.61867905e-01 1.18704863e-01 5.71275055e-01 -2.68215407e-02
3.12906176e-01 2.79407322e-01 6.80128336e-01 -4.99690682e-01
4.23798561e-01 6.87186956e-01 -1.06275886e-01 -2.07775295e-01
-4.97398704e-01 1.50272809e-02 -5.94169497e-01 -1.86184436e-01
6.70475483e-01 -8.48132610e-01 -4.52771813e-01 2.57106036e-01
-7.14436650e-01 -9.60413694e-01 -5.24423957e-01 7.36338019e-01
-9.26494360e-01 1.56442642e-01 -7.75215209e-01 -4.30129506e-02
-2.43141726e-01 -1.68015540e+00 1.15528011e+00 2.10263520e-01
-3.15682828e-01 -1.16514480e+00 1.59449782e-02 2.63851374e-01
4.76278186e-01 7.31460690e-01 7.55165935e-01 -3.98041964e-01
-5.03392756e-01 -4.41345096e-01 -3.41269858e-02 3.73558313e-01
2.17010006e-01 -9.03761461e-02 -4.91122425e-01 -4.34676081e-01
5.71941733e-02 -5.76848090e-01 4.22516406e-01 6.37102008e-01
1.09245169e+00 -3.85225825e-02 -4.13496256e-01 9.33418155e-01
1.42517984e+00 1.10525012e-01 4.87324387e-01 1.94864079e-01
5.24835646e-01 6.14592850e-01 5.01074433e-01 -1.46222413e-01
1.03353828e-01 4.78184909e-01 3.39466035e-01 -7.65414059e-01
-3.57959062e-01 -3.49478722e-01 -1.73556983e-01 5.10996521e-01
4.05257761e-01 -1.60468474e-01 -1.19651830e+00 8.00738275e-01
-1.26876831e+00 -4.88148510e-01 6.22797608e-02 2.24532366e+00
9.31458056e-01 -8.25918745e-03 -2.00210512e-01 -2.86307454e-01
5.89745700e-01 -4.36160266e-01 -3.85093272e-01 -3.31461877e-01
2.58149385e-01 4.92864326e-02 6.83306098e-01 6.02419078e-01
-9.55717623e-01 6.97040796e-01 6.55535984e+00 4.22555625e-01
-1.53284466e+00 1.20814651e-01 1.07938933e+00 -8.04448426e-02
-2.32058510e-01 -7.76487216e-02 -3.39356847e-02 4.83225882e-01
9.40737247e-01 -2.14353099e-01 2.07561731e-01 6.49352789e-01
4.38423038e-01 -2.94932276e-01 -1.17307830e+00 8.84726107e-01
-1.21603105e-02 -1.48691690e+00 1.57755464e-02 1.01828054e-01
1.17721939e+00 1.61033913e-01 4.96394001e-02 -4.93607335e-02
4.65403885e-01 -1.33688653e+00 4.73834425e-01 3.05156648e-01
1.11925435e+00 -5.31813085e-01 9.45928097e-01 3.49999934e-01
-4.53814566e-01 4.41981912e-01 -1.13659039e-01 4.50131863e-01
2.41054907e-01 4.20380205e-01 -1.24926972e+00 3.18390369e-01
3.47456813e-01 4.06681716e-01 -4.83235806e-01 1.32571840e+00
-9.83189493e-02 2.18641937e-01 -3.81758690e-01 3.03223133e-01
3.35887223e-01 -2.85117477e-01 3.63649912e-02 9.83146548e-01
4.98222679e-01 -2.40364239e-01 -3.35472659e-03 8.61694634e-01
-1.23425342e-01 -1.52995521e-02 -7.91851640e-01 9.04393122e-02
-8.49133804e-02 1.31589019e+00 -1.09491932e+00 -4.06832576e-01
-5.21509349e-02 1.01537454e+00 -1.43849149e-01 2.26413891e-01
-7.50629663e-01 -1.51905239e-01 1.06240653e-01 2.31026351e-01
-1.20271318e-01 -2.05476731e-01 -3.29320401e-01 -1.02814448e+00
-1.39840111e-01 -9.33117568e-01 7.19476864e-02 -1.05000901e+00
-7.48023450e-01 7.61007428e-01 -5.58300018e-02 -1.38403201e+00
-6.19364381e-01 -5.38620532e-01 -4.21276420e-01 8.98273647e-01
-1.26762915e+00 -1.05760610e+00 -7.59803653e-01 1.58266082e-01
3.04692715e-01 3.36767524e-01 8.95165980e-01 2.31154636e-01
-1.27431080e-02 3.83835047e-01 2.76702940e-01 3.44364733e-01
8.17784667e-01 -1.13701117e+00 3.55987221e-01 4.27497745e-01
-6.94040284e-02 5.26645541e-01 8.61999810e-01 -3.99431735e-01
-1.00196457e+00 -1.00658369e+00 2.80237973e-01 -4.56324965e-01
1.66701853e-01 -2.11815462e-01 -5.61510265e-01 8.35858881e-01
1.62842304e-01 3.31969559e-01 4.74795192e-01 -7.85938025e-01
3.37801427e-01 1.86805099e-01 -1.48616493e+00 6.17268443e-01
6.61309958e-01 -1.65846255e-02 -2.64557838e-01 5.71112871e-01
4.36618119e-01 -1.00623798e+00 -9.57192004e-01 5.99415123e-01
6.32886052e-01 -9.94892180e-01 9.60319638e-01 -1.69257537e-01
7.50651479e-01 -2.15606034e-01 4.06655103e-01 -1.64225829e+00
4.92256165e-01 -5.08989871e-01 6.36411667e-01 4.65767026e-01
5.66965997e-01 -4.84838814e-01 1.02096629e+00 6.54651284e-01
-1.66005149e-01 -7.16185153e-01 -6.19778335e-01 -5.74352562e-01
3.56185555e-01 -4.83170561e-02 2.97593623e-01 9.81781244e-01
-9.90267545e-02 -1.74739525e-01 2.79457271e-02 -2.72099793e-01
6.34314179e-01 3.04516137e-01 7.90758014e-01 -8.90718758e-01
-1.44635603e-01 -3.26163024e-01 -3.76182437e-01 -6.37174785e-01
4.03315686e-02 -1.09404182e+00 1.43495798e-01 -1.74474776e+00
3.30134392e-01 -8.23928654e-01 2.94040620e-01 3.12289059e-01
1.80319786e-01 8.77406180e-01 -5.00346394e-03 2.20739201e-01
-2.13255391e-01 2.49225721e-01 1.95437753e+00 1.62280817e-02
-7.54146725e-02 -2.71256953e-01 -3.97660375e-01 7.21453726e-01
8.88548017e-01 -5.39448202e-01 -5.15399456e-01 -4.64606196e-01
-1.03211053e-01 4.06253964e-01 3.12123686e-01 -1.05710876e+00
-9.55186561e-02 1.56300947e-01 7.65218079e-01 -4.28182855e-02
2.01012239e-01 -8.46352756e-01 6.13684118e-01 1.00068438e+00
-4.10295606e-01 -9.95550081e-02 3.70434552e-01 -4.51829396e-02
-2.33621359e-01 -2.70746380e-01 1.08211565e+00 -7.27672815e-01
-1.91425532e-01 2.63803661e-01 -1.30210459e-01 1.08458847e-01
1.22680378e+00 -2.55834818e-01 5.14845178e-02 -6.03876710e-01
-9.10370588e-01 2.17747107e-01 1.27118731e+00 9.80402678e-02
5.00711739e-01 -1.06116700e+00 -7.39143610e-01 4.78635699e-01
9.97100472e-02 4.77100104e-01 1.90660164e-01 9.60354030e-01
-1.52916515e+00 3.40285510e-01 -5.24028659e-01 -8.65537465e-01
-9.28066611e-01 5.93335629e-01 9.68792737e-01 -2.45382547e-01
-6.88021004e-01 5.23562193e-01 4.71610755e-01 -9.85686302e-01
-1.07299462e-01 -3.70326668e-01 3.92835289e-01 -5.93723536e-01
1.91984966e-01 -4.08761054e-01 1.49053916e-01 -5.08800268e-01
-2.21787822e-02 4.52515215e-01 1.29867032e-01 -2.86757618e-01
1.33530664e+00 1.02386931e-02 1.24913320e-01 6.79568872e-02
1.27766418e+00 -1.25851378e-01 -1.50332892e+00 3.51354778e-01
-5.45701921e-01 -5.37154973e-01 -1.06566392e-01 -9.96985853e-01
-1.24957967e+00 8.17111313e-01 7.56683350e-01 -2.88558006e-01
9.01112378e-01 -6.02974184e-02 8.08478177e-01 -2.84759086e-02
5.41348159e-01 -5.40759385e-01 5.15581705e-02 2.16383711e-02
7.62041688e-01 -1.39978600e+00 -1.74851477e-01 -3.78048152e-01
-6.92681849e-01 9.28172886e-01 5.51382363e-01 -2.13548824e-01
2.37036481e-01 6.48137331e-01 6.17254615e-01 -1.20478906e-01
-2.36138955e-01 3.52131754e-01 -8.61667469e-03 6.15015984e-01
5.02767920e-01 6.84019998e-02 -2.87033975e-01 -9.64380130e-02
-4.59159225e-01 2.66750216e-01 9.17494655e-01 8.62037897e-01
5.85651323e-02 -1.08844447e+00 -2.04964533e-01 2.91833341e-01
-5.14328122e-01 1.60502438e-02 -1.33445740e-01 1.03269970e+00
-2.10207142e-02 4.31422710e-01 1.29799902e-01 2.14247331e-01
1.17609270e-01 -3.01489115e-01 8.29246938e-01 -6.58977687e-01
-5.78380346e-01 -9.73516703e-02 -1.17224589e-01 -4.18461412e-01
-1.70149341e-01 -2.83574730e-01 -1.30827761e+00 -1.98784664e-01
3.52668613e-02 2.41689369e-01 9.75567222e-01 5.45467675e-01
1.54110298e-01 4.95579302e-01 3.60138893e-01 -1.26116240e+00
-2.36017570e-01 -6.28494084e-01 -3.25175345e-01 7.58631170e-01
3.04294854e-01 -3.31038803e-01 -2.61835396e-01 3.18903744e-01] | [14.316856384277344, -2.2092673778533936] |
dde794de-c60f-464d-bfb7-13d0655c4e79 | a-unified-probabilistic-model-for-learning | 1805.09567 | null | http://arxiv.org/abs/1805.09567v1 | http://arxiv.org/pdf/1805.09567v1.pdf | A Unified Probabilistic Model for Learning Latent Factors and Their Connectivities from High-Dimensional Data | Connectivity estimation is challenging in the context of high-dimensional
data. A useful preprocessing step is to group variables into clusters, however,
it is not always clear how to do so from the perspective of connectivity
estimation. Another practical challenge is that we may have data from multiple
related classes (e.g., multiple subjects or conditions) and wish to incorporate
constraints on the similarities across classes. We propose a probabilistic
model which simultaneously performs both a grouping of variables (i.e.,
detecting community structure) and estimation of connectivities between the
groups which correspond to latent variables. The model is essentially a factor
analysis model where the factors are allowed to have arbitrary correlations,
while the factor loading matrix is constrained to express a community
structure. The model can be applied on multiple classes so that the
connectivities can be different between the classes, while the community
structure is the same for all classes. We propose an efficient estimation
algorithm based on score matching, and prove the identifiability of the model.
Finally, we present an extension to directed (causal) connectivities over
latent variables. Simulations and experiments on fMRI data validate the
practical utility of the method. | ['Aapo Hyvärinen', 'Ricardo Pio Monti'] | 2018-05-24 | null | null | null | null | ['connectivity-estimation'] | ['graphs'] | [ 7.00302944e-02 7.48422742e-02 -1.37687236e-01 -3.34198684e-01
-3.82199697e-02 -6.58763111e-01 2.62458503e-01 6.39407486e-02
-1.70314521e-01 5.74283242e-01 2.80247867e-01 -2.25696415e-01
-6.25022292e-01 -7.64996409e-01 -2.81537145e-01 -9.04577374e-01
-4.63173091e-01 7.21354246e-01 -5.06826751e-02 3.63700598e-01
6.19866177e-02 4.80199814e-01 -9.10619974e-01 -2.35522036e-02
9.14808631e-01 3.09407145e-01 1.80873811e-01 3.69641006e-01
1.01387024e-01 1.48951203e-01 -2.89548486e-01 -1.22492187e-01
1.69272631e-01 -4.45050418e-01 -7.46684670e-01 4.46971893e-01
2.75358558e-01 -9.60539356e-02 -1.30124271e-01 1.37168229e+00
3.18463653e-01 -1.74708907e-02 9.70883787e-01 -1.57383680e+00
-1.88222498e-01 6.83621466e-01 -8.87859285e-01 4.72478867e-02
8.10946822e-02 -5.40553510e-01 1.34246373e+00 -6.56794727e-01
6.71308458e-01 1.29118955e+00 2.73783147e-01 2.09084451e-01
-1.95838916e+00 -6.35363460e-01 4.07120347e-01 1.12358406e-01
-1.50662315e+00 -2.06778064e-01 7.83834517e-01 -1.06715000e+00
3.35152000e-01 3.19941938e-02 6.45982265e-01 9.08990681e-01
1.19113766e-01 4.23331290e-01 9.63551104e-01 -1.20692588e-01
5.13222158e-01 -1.61166921e-01 5.35434425e-01 6.61828935e-01
5.05880475e-01 -2.44553432e-01 -3.83960336e-01 -6.48079813e-01
8.32431376e-01 1.23132311e-01 -3.24546933e-01 -1.00104594e+00
-1.48368704e+00 1.11452103e+00 2.79464722e-01 2.94359118e-01
-5.00069320e-01 -1.27444103e-01 1.51160657e-01 2.52040148e-01
2.75171310e-01 2.67485231e-01 -2.04627395e-01 4.49158221e-01
-1.13903189e+00 1.52834460e-01 7.73981214e-01 7.69042790e-01
7.74944484e-01 -2.78566301e-01 6.30915165e-02 7.02075183e-01
3.71229798e-01 3.25732619e-01 1.65249601e-01 -9.63308513e-01
3.81259471e-01 3.76147509e-01 -2.64937431e-02 -1.50992954e+00
-5.34809351e-01 -3.09758425e-01 -1.31112361e+00 4.80749533e-02
4.85124230e-01 -3.42188120e-01 -5.65786839e-01 2.33050728e+00
1.95344463e-01 2.04618841e-01 -3.92236769e-01 9.32458997e-01
2.56441414e-01 3.62294286e-01 5.89496940e-02 -6.42820835e-01
1.16591859e+00 -2.82416075e-01 -7.50394642e-01 -4.49667126e-02
4.13845122e-01 -3.82261485e-01 4.21580672e-01 2.75076509e-01
-1.02702546e+00 -2.14369342e-01 -8.34614456e-01 3.48342150e-01
2.08221748e-01 1.00363381e-01 8.40109706e-01 4.16631460e-01
-1.07363665e+00 3.72441053e-01 -1.14480376e+00 -4.00845140e-01
5.37112132e-02 5.90557337e-01 -5.46048462e-01 -4.99462001e-02
-9.80446458e-01 4.08574849e-01 2.99720883e-01 -4.89106327e-02
-6.64596736e-01 -2.32099012e-01 -4.75963354e-01 3.24759394e-01
2.42747605e-01 -7.69576848e-01 3.54076177e-01 -1.04840636e+00
-6.70986772e-01 5.57376087e-01 -4.45497066e-01 3.05664018e-02
4.40296412e-01 3.95579904e-01 -1.51422441e-01 3.19229990e-01
4.46840674e-01 4.75207210e-01 6.96057439e-01 -1.03280878e+00
-1.17912605e-01 -4.78310823e-01 -8.95044878e-02 2.52593160e-01
-3.45506728e-01 8.21164548e-02 -4.46314722e-01 -3.88612658e-01
9.02266562e-01 -1.08429539e+00 -3.35953444e-01 3.63494605e-01
-7.91722357e-01 2.70340685e-02 3.24737012e-01 -6.82385027e-01
1.05526578e+00 -2.31246591e+00 7.63866186e-01 9.31295872e-01
9.32791650e-01 -5.81907332e-01 -1.41164392e-01 2.92441070e-01
-5.06454825e-01 1.56370729e-01 -5.28896749e-01 -8.65301266e-02
-2.22592175e-01 1.45449117e-01 2.05533262e-08 9.51342106e-01
3.24975923e-02 4.81079787e-01 -8.96150112e-01 -5.82833529e-01
-2.04753697e-01 3.43637854e-01 -7.94036090e-01 3.89571227e-02
2.91739643e-01 5.83570480e-01 -6.24654591e-01 2.49345452e-01
8.00683618e-01 -5.37069917e-01 9.05586600e-01 -2.14940071e-01
7.12279603e-02 1.23774081e-01 -1.66835785e+00 1.28015661e+00
9.25451443e-02 6.00763977e-01 2.31785387e-01 -1.36960566e+00
5.87589025e-01 4.55290377e-01 1.00296938e+00 4.29775380e-02
-5.15402257e-02 -7.40900934e-02 5.67975581e-01 -2.19372422e-01
-8.20141882e-02 -1.43033773e-01 7.90558830e-02 7.13481426e-01
4.70912494e-02 5.19759059e-01 2.25424126e-01 5.84418356e-01
1.22877395e+00 -4.96423602e-01 2.46977597e-01 -5.60585678e-01
1.18904330e-01 -3.29214543e-01 8.51765096e-01 5.01406968e-01
-1.32664099e-01 3.60571414e-01 1.12956333e+00 1.16950803e-01
-1.06223822e+00 -1.39476097e+00 -3.88576061e-01 5.21883368e-01
9.47727337e-02 -4.51860696e-01 -6.95549786e-01 -4.76815820e-01
-1.00689754e-01 1.38758630e-01 -6.95929766e-01 5.37197180e-02
-1.37429535e-01 -9.51319337e-01 3.49375047e-02 2.78321683e-01
9.79590490e-02 -5.21404028e-01 -9.50608775e-02 2.16286287e-01
-6.23773098e-01 -9.08207715e-01 -3.89722049e-01 7.92296678e-02
-1.15469813e+00 -1.03985167e+00 -5.28055668e-01 -7.59540975e-01
1.16417480e+00 3.54037225e-01 8.33444476e-01 1.14615820e-01
-1.26626581e-01 2.70699829e-01 9.88735235e-04 3.98883820e-01
3.74820381e-02 -3.81518640e-02 3.18640023e-01 4.16665763e-01
2.25634396e-01 -9.93723631e-01 -5.70519269e-01 5.89459598e-01
-7.11385787e-01 1.30077139e-01 3.78320694e-01 9.25786853e-01
5.79546034e-01 3.47338617e-01 4.09010231e-01 -9.04587269e-01
6.17772460e-01 -8.67012143e-01 -4.63960648e-01 3.61313075e-01
-4.03197050e-01 4.69014905e-02 1.47014245e-01 -7.69243360e-01
-5.32263696e-01 1.50899842e-01 4.45225149e-01 -3.49128306e-01
3.01665021e-03 9.95271266e-01 -6.02329016e-01 1.81688964e-01
4.30666417e-01 -1.17357679e-01 1.07617499e-02 -5.10135233e-01
4.47675973e-01 4.96818215e-01 8.56222883e-02 -6.90749288e-01
6.47443295e-01 5.76910615e-01 2.24567249e-01 -9.82149839e-01
-2.05934912e-01 -4.59379464e-01 -1.27257526e+00 -1.99421257e-01
7.75671482e-01 -9.13668275e-01 -5.45720577e-01 -2.10128981e-03
-1.00246871e+00 -3.61156948e-02 1.99258223e-01 8.32807183e-01
-4.11568522e-01 6.71296358e-01 -5.29302955e-01 -6.21918142e-01
3.30356926e-01 -1.06184721e+00 5.20651102e-01 -3.54061991e-01
-5.39343715e-01 -1.28439474e+00 1.70840234e-01 9.39358547e-02
-1.88219491e-02 1.94024742e-01 1.26650262e+00 -4.46370453e-01
-4.64071155e-01 -1.71127290e-01 -2.05192924e-01 -1.60546362e-01
1.44386917e-01 7.66203552e-02 -3.92979652e-01 -4.90353853e-01
-6.34476840e-02 9.30242985e-02 7.40527451e-01 7.94623852e-01
9.05869246e-01 -2.71141022e-01 -7.65794933e-01 3.71852279e-01
1.06522465e+00 -7.71378577e-02 3.92895311e-01 -3.41583669e-01
7.48490393e-01 9.12650168e-01 7.98506886e-02 3.04842770e-01
1.81366295e-01 8.40692997e-01 1.19441397e-01 -3.20647843e-02
2.16756299e-01 -7.47602209e-02 1.42953560e-01 1.16872096e+00
1.90609712e-02 -1.73387364e-01 -9.54036951e-01 5.85695505e-01
-1.96763158e+00 -1.07425451e+00 -5.90869129e-01 2.40322256e+00
6.23247087e-01 -2.12146133e-01 2.31306300e-01 1.76377222e-02
1.16893256e+00 -2.39657566e-01 -6.19165480e-01 8.65330398e-02
-8.71110484e-02 -1.48921698e-01 1.54022947e-01 6.82052851e-01
-8.84752333e-01 5.52702963e-01 6.51367283e+00 3.48627836e-01
-8.16370845e-01 1.56020820e-01 4.33669627e-01 -1.14972897e-01
-3.42081070e-01 3.32638234e-01 -4.74881589e-01 3.83161664e-01
6.33651078e-01 -2.54685193e-01 5.08487582e-01 4.83381510e-01
3.49637210e-01 -2.98398361e-02 -1.07651663e+00 6.17523193e-01
-1.14338249e-01 -6.39152288e-01 -2.68301517e-01 4.31436658e-01
7.02070773e-01 -2.86329966e-02 -2.07733974e-01 -2.76102334e-01
5.44408619e-01 -6.94638431e-01 2.93978453e-01 5.50951779e-01
6.91907287e-01 -5.27217329e-01 2.29919538e-01 4.90433663e-01
-1.11444700e+00 1.47355571e-01 -6.33425891e-01 -2.12802403e-02
3.07711631e-01 1.05044198e+00 -4.48878050e-01 2.33860150e-01
3.86298478e-01 9.17992473e-01 -3.29484880e-01 1.20217097e+00
-3.54796827e-01 6.00718200e-01 -4.03944135e-01 3.47812593e-01
-2.45713815e-01 -7.32672513e-01 5.88067532e-01 7.75834441e-01
3.33085805e-01 5.09346165e-02 4.50307786e-01 1.10415304e+00
2.00237721e-01 2.70710826e-01 -4.84309286e-01 -6.60253987e-02
4.40674484e-01 1.40956700e+00 -1.06902945e+00 -1.62349522e-01
-5.54078043e-01 9.99862432e-01 5.34025192e-01 7.06068754e-01
-4.22362149e-01 -3.79639082e-02 7.98993230e-01 2.39171192e-01
1.43356130e-01 -6.69848263e-01 -2.81071484e-01 -1.56719363e+00
-2.58395430e-02 -5.35991192e-01 3.57340395e-01 -5.38215816e-01
-1.73364663e+00 3.43434572e-01 1.90045133e-01 -1.08078396e+00
-1.49855971e-01 -3.23821038e-01 -5.56735456e-01 1.19634247e+00
-8.18113387e-01 -7.49028444e-01 -5.71483858e-02 8.48390937e-01
-1.93216369e-01 8.04955661e-02 7.32679009e-01 3.47296178e-01
-7.61066020e-01 2.04823852e-01 2.87453949e-01 2.81645477e-01
5.00798762e-01 -1.17792678e+00 -1.64917558e-01 8.89362097e-01
2.41559595e-01 1.13461018e+00 4.41810340e-01 -1.01642656e+00
-9.08688843e-01 -9.14986908e-01 1.03944921e+00 -4.13748063e-02
1.03896093e+00 -7.68550932e-01 -9.88923788e-01 8.81657124e-01
-1.18410058e-01 -9.39192250e-02 1.05170131e+00 6.72662139e-01
-2.81294614e-01 2.55626291e-01 -8.23983729e-01 6.52771175e-01
1.06686175e+00 -4.74210411e-01 -2.59508550e-01 4.74724263e-01
2.08689541e-01 2.59279341e-01 -9.26792979e-01 8.41273591e-02
5.69857478e-01 -6.97777450e-01 8.70305419e-01 -8.70661557e-01
1.72293127e-01 -4.39529121e-01 -1.82090953e-01 -1.37177992e+00
-8.81826937e-01 -1.40454277e-01 1.20957410e-02 1.26924849e+00
4.68158633e-01 -5.53162158e-01 6.38390481e-01 8.78751576e-01
4.39842522e-01 -1.82973176e-01 -1.10444093e+00 -6.47468328e-01
1.74183950e-01 -1.97242603e-01 1.38903245e-01 1.58521378e+00
2.89610744e-01 7.47568607e-01 -4.75579917e-01 5.30089319e-01
9.91076946e-01 2.76191473e-01 3.56905252e-01 -1.68509376e+00
-5.95284700e-01 -4.12752688e-01 -6.41699910e-01 -9.41898882e-01
3.75375777e-01 -1.05340290e+00 -1.72965616e-01 -1.51208007e+00
8.07155609e-01 -5.89929819e-01 -2.16375515e-01 5.68691611e-01
-1.31682113e-01 -2.26978078e-01 1.79817546e-02 5.27263999e-01
-3.14929336e-01 3.91616642e-01 1.08499682e+00 -1.10866942e-01
-2.98996747e-01 1.13923155e-01 -6.65862501e-01 5.39898992e-01
7.01409280e-01 -7.02688336e-01 -6.76190436e-01 -1.52479619e-01
4.43845630e-01 3.78691554e-01 3.98281336e-01 -6.95860445e-01
3.65101784e-01 -2.91843235e-01 4.11805987e-01 -3.60037684e-01
1.83754444e-01 -8.67941260e-01 5.06248653e-01 3.84174019e-01
-4.48102027e-01 -4.71350402e-02 -3.03794861e-01 6.67563319e-01
-1.25298603e-02 -1.94047675e-01 6.19129300e-01 2.03402787e-01
3.09376884e-03 5.49364984e-01 -5.00150919e-01 -6.13087043e-02
8.96864116e-01 2.04124171e-02 2.55181324e-02 -4.78603959e-01
-1.30820251e+00 2.75567383e-01 2.83232957e-01 7.80281723e-02
5.77306926e-01 -1.48969817e+00 -7.09132075e-01 2.81327754e-01
4.19322727e-03 -4.68980372e-01 2.21466050e-01 1.08258736e+00
1.23779230e-01 4.60596569e-02 -3.44672501e-01 -7.58559525e-01
-1.49898481e+00 7.18922377e-01 2.56063968e-01 -2.13026837e-01
-5.16176581e-01 3.89181077e-01 5.92688739e-01 -3.87435913e-01
1.13853198e-02 1.35700792e-01 -3.73527795e-01 4.46402848e-01
2.88930297e-01 2.30236486e-01 -3.41903746e-01 -7.10390806e-01
-3.31174642e-01 5.17031312e-01 1.20751653e-02 -3.71856391e-01
1.41092098e+00 -4.07584637e-01 -5.98424911e-01 5.71979642e-01
1.15307856e+00 -6.86405972e-03 -1.15402818e+00 -4.21267331e-01
-6.45548478e-02 -4.05876428e-01 1.55102670e-01 -3.76201808e-01
-1.37507629e+00 1.12987149e+00 4.60807413e-01 3.22992742e-01
8.70990098e-01 1.22231334e-01 -2.53265761e-02 1.17200084e-01
4.56314325e-01 -7.08073199e-01 -1.53980553e-01 1.92239270e-01
6.33481920e-01 -7.89571762e-01 9.92191583e-02 -6.89318955e-01
-4.81308639e-01 9.13221002e-01 1.72937259e-01 -1.07563049e-01
9.87177789e-01 -5.73787689e-02 -4.87163544e-01 -5.17203569e-01
-5.48063815e-01 -1.55135253e-02 4.67177749e-01 5.46576798e-01
5.34044087e-01 4.88895178e-01 -5.52471161e-01 4.95094389e-01
-3.18756793e-03 -4.41467226e-01 5.85787296e-01 4.73941147e-01
-2.65858710e-01 -1.22185206e+00 -3.49462241e-01 7.21695006e-01
-2.87956595e-01 -2.06440300e-01 -4.21245307e-01 4.39116389e-01
-9.61540490e-02 9.69488084e-01 1.47745252e-01 -1.43339053e-01
-1.73932277e-02 -8.52609333e-03 3.61876190e-01 -7.59044111e-01
1.39873564e-01 5.58576703e-01 -1.56975970e-01 -3.93529922e-01
-5.66732049e-01 -1.13164699e+00 -9.72394109e-01 -4.35226023e-01
-5.18099070e-01 3.70428205e-01 5.96895874e-01 8.56835186e-01
1.93508938e-01 4.31541353e-01 7.10822880e-01 -4.43761647e-01
-2.34148633e-02 -8.39856744e-01 -1.17487156e+00 4.49486315e-01
1.14044592e-01 -1.01718152e+00 -4.57484335e-01 2.78643310e-01] | [7.154681205749512, 5.102121353149414] |
d9eaaf98-c037-4646-bbbd-103ff72a728d | real-time-document-image-classification-using | 1711.05862 | null | http://arxiv.org/abs/1711.05862v1 | http://arxiv.org/pdf/1711.05862v1.pdf | Real-Time Document Image Classification using Deep CNN and Extreme Learning Machines | This paper presents an approach for real-time training and testing for
document image classification. In production environments, it is crucial to
perform accurate and (time-)efficient training. Existing deep learning
approaches for classifying documents do not meet these requirements, as they
require much time for training and fine-tuning the deep architectures.
Motivated from Computer Vision, we propose a two-stage approach. The first
stage trains a deep network that works as feature extractor and in the second
stage, Extreme Learning Machines (ELMs) are used for classification. The
proposed approach outperforms all previously reported structural and deep
learning based methods with a final accuracy of 83.24% on Tobacco-3482 dataset,
leading to a relative error reduction of 25% when compared to a previous
Convolutional Neural Network (CNN) based approach (DeepDocClassifier). More
importantly, the training time of the ELM is only 1.176 seconds and the overall
prediction time for 2,482 images is 3.066 seconds. As such, this novel approach
makes deep learning-based document classification suitable for large-scale
real-time applications. | ['Andreas Kölsch', 'Muhammad Zeshan Afzal', 'Marcus Liwicki', 'Markus Ebbecke'] | 2017-11-03 | null | null | null | null | ['document-image-classification'] | ['computer-vision'] | [ 1.88215360e-01 -2.04276085e-01 -3.17444801e-02 -5.74726880e-01
-6.42679870e-01 -5.55547535e-01 7.22598195e-01 4.42065924e-01
-6.28961086e-01 3.04284602e-01 -5.46644330e-01 -6.78212285e-01
2.04717033e-02 -9.33348119e-01 -5.33594131e-01 -7.71692932e-01
2.12629050e-01 6.57667518e-01 -1.41071230e-02 1.63956627e-01
5.14014602e-01 7.97228932e-01 -1.56428778e+00 5.86493790e-01
8.59736562e-01 1.41307914e+00 2.96464771e-01 1.28390801e+00
-5.15959740e-01 1.11705756e+00 -9.48929489e-01 -5.57202935e-01
-3.53595279e-02 -1.69458494e-01 -7.96108782e-01 2.07644314e-01
5.39193034e-01 -6.67528689e-01 -1.81080177e-01 6.98589385e-01
6.54622197e-01 1.26495823e-01 9.03263271e-01 -9.26006138e-01
-7.33002067e-01 4.57049608e-01 -5.92631638e-01 1.12379538e-02
-7.20968843e-02 -1.21453524e-01 8.39121819e-01 -9.88075674e-01
2.08126381e-01 8.04184794e-01 7.12258041e-01 4.92840052e-01
-8.65849614e-01 -3.86206210e-01 8.81209373e-02 3.44216973e-01
-1.24150574e+00 -3.83189768e-01 5.11966944e-01 -6.08139038e-01
1.20743740e+00 1.56775922e-01 3.87233645e-01 1.07993758e+00
3.14794153e-01 1.18062365e+00 8.84160578e-01 -7.42682338e-01
3.42567861e-01 3.55201304e-01 2.63054520e-01 8.14261496e-01
2.90504932e-01 -2.14508086e-01 -3.82000841e-02 1.27846643e-01
3.25773180e-01 3.33617508e-01 1.48620009e-01 1.05046168e-01
-7.17783332e-01 9.16491508e-01 5.53805947e-01 5.89361489e-01
-5.66128254e-01 1.59364432e-01 7.33479679e-01 3.98370743e-01
4.58401442e-01 2.15681821e-01 -4.13208514e-01 -1.23902395e-01
-1.33666408e+00 -1.14353232e-01 8.03706706e-01 7.18540728e-01
4.36047018e-01 2.11433068e-01 -2.45400831e-01 9.59581971e-01
1.29988179e-01 8.82008448e-02 8.35807681e-01 -4.82181221e-01
3.19977432e-01 7.23802745e-01 -2.45892927e-01 -9.58671272e-01
-4.48011339e-01 -8.88011634e-01 -1.09731448e+00 2.55138129e-01
4.20869291e-01 -9.30413008e-02 -1.13248193e+00 9.06533599e-01
7.36316592e-02 -1.99678868e-01 2.31083348e-01 5.21707714e-01
7.90608287e-01 8.66034567e-01 -5.08477092e-02 1.24921136e-01
1.27461874e+00 -1.36398995e+00 -5.07134318e-01 -8.58426094e-02
8.90473425e-01 -7.55153596e-01 1.01066041e+00 7.75900245e-01
-1.03250277e+00 -7.68403947e-01 -1.19121349e+00 -1.49005458e-01
-7.98639596e-01 5.84543884e-01 5.85114896e-01 8.25720608e-01
-9.69665825e-01 5.88434041e-01 -5.48365712e-01 -2.36865357e-01
5.16120076e-01 5.31474173e-01 -8.89785215e-02 -1.02720991e-01
-6.85514152e-01 6.35039628e-01 5.00598490e-01 3.72525007e-01
-7.83922613e-01 -2.30675325e-01 -3.88489276e-01 5.81348360e-01
-3.21493633e-02 -1.10760160e-01 1.53970695e+00 -1.18687475e+00
-1.76508915e+00 8.82952392e-01 1.28047794e-01 -6.37823701e-01
7.10132718e-01 -1.92379877e-01 -3.28617305e-01 2.24389777e-01
-4.00316924e-01 5.68181574e-01 8.60130668e-01 -1.10618961e+00
-8.39836955e-01 -2.74192601e-01 -7.73194283e-02 -1.24099679e-01
-9.70715642e-01 1.36995809e-02 -5.45790613e-01 -4.31797564e-01
-1.98709592e-01 -8.02159131e-01 2.35990491e-02 1.84249077e-02
-4.43344742e-01 -3.72652471e-01 9.48464692e-01 -7.12696910e-01
1.00688839e+00 -2.10527396e+00 -3.11797857e-01 1.08409502e-01
1.79787800e-01 8.15309584e-01 -2.13810518e-01 2.94911325e-01
-1.53704369e-02 -1.55730382e-01 3.52207720e-02 -5.23493111e-01
1.19781889e-01 -2.80625988e-02 -6.10467903e-02 3.46569002e-01
1.39503092e-01 9.18569684e-01 -4.18988734e-01 -4.74937409e-01
3.40128839e-01 6.78667128e-01 -3.77010673e-01 3.02190840e-01
-2.62404293e-01 -4.74311151e-02 -2.68988192e-01 7.02105403e-01
7.60882437e-01 -2.34456733e-01 2.22314239e-01 1.22757755e-01
-1.48004934e-01 2.18274109e-02 -9.14423883e-01 1.37403214e+00
-7.77648926e-01 1.06291771e+00 -1.14683151e-01 -1.55496418e+00
1.23604095e+00 3.22747201e-01 3.25810969e-01 -9.00510430e-01
5.00373662e-01 3.94949913e-01 -1.09389089e-01 -4.61191952e-01
5.67532182e-01 3.54066640e-01 1.88183859e-01 4.55375999e-01
1.90881923e-01 2.68007576e-01 3.46056223e-01 -8.76297802e-02
9.95859325e-01 -1.87196836e-01 4.75782752e-02 -4.77904864e-02
7.99796760e-01 -7.08849579e-02 1.42115369e-01 8.58155251e-01
2.96725868e-03 3.62625182e-01 3.47623020e-01 -9.64165270e-01
-1.06689012e+00 -4.77776051e-01 -5.22406101e-02 1.30731654e+00
-4.77771074e-01 -2.37248257e-01 -9.95386481e-01 -7.87999809e-01
-1.07617058e-01 7.37255394e-01 -5.30197680e-01 -1.55276060e-01
-4.42713052e-01 -5.99716544e-01 4.71311778e-01 8.58364284e-01
7.61083186e-01 -1.10518837e+00 -4.58129734e-01 4.04907078e-01
1.71849415e-01 -1.07426488e+00 2.66844425e-02 5.90175092e-01
-8.80527377e-01 -7.98583865e-01 -9.12904084e-01 -1.00735486e+00
6.79948688e-01 -7.05298260e-02 1.10588932e+00 4.91247803e-01
-4.88052696e-01 -6.43322021e-02 -4.10097331e-01 -5.85077584e-01
-5.22016346e-01 4.09877241e-01 -3.56640309e-01 1.64896362e-02
6.62778497e-01 -9.59612280e-02 -5.50711811e-01 -1.92234918e-01
-8.53674591e-01 4.94620278e-02 9.50747848e-01 1.11899173e+00
3.60548794e-01 3.84188324e-01 4.77658004e-01 -9.57984924e-01
6.89699590e-01 -2.49671087e-01 -8.40726435e-01 4.39780444e-01
-8.64656389e-01 -9.93573815e-02 9.79974806e-01 -5.43160379e-01
-1.06013584e+00 2.81156987e-01 -3.58496577e-01 -3.89285423e-02
-3.72377694e-01 6.34045541e-01 3.01879913e-01 6.05796762e-02
6.33314431e-01 4.36372340e-01 -2.52671808e-01 -6.15397036e-01
-8.49525109e-02 1.37250006e+00 4.28262979e-01 -2.21884593e-01
3.91125053e-01 1.39957175e-01 -1.60296991e-01 -8.35742831e-01
-9.43142354e-01 -2.90051997e-01 -8.97376835e-01 -1.49077773e-01
7.70861506e-01 -8.13947558e-01 -9.45976377e-01 7.87443459e-01
-1.11445260e+00 -4.12031710e-01 -1.42246056e-02 4.43012983e-01
-4.59002435e-01 9.41904560e-02 -8.25045168e-01 -8.76381218e-01
-9.40029919e-01 -1.03970063e+00 1.04286528e+00 1.82058409e-01
-8.79232120e-03 -1.03534913e+00 -2.16143340e-01 5.19665241e-01
5.57147205e-01 1.57100990e-01 9.40646708e-01 -8.11425328e-01
-1.71398774e-01 -7.15945899e-01 -4.43113953e-01 5.92979133e-01
-1.82629943e-01 3.14282000e-01 -1.29468024e+00 -4.82225209e-01
-2.31098190e-01 -5.76590180e-01 9.94931579e-01 4.55971152e-01
1.73857880e+00 -1.60995722e-01 -2.00798675e-01 5.52103519e-01
1.44534874e+00 4.65027064e-01 5.59574127e-01 6.12074733e-01
4.88220960e-01 2.78984874e-01 3.81029934e-01 4.57829952e-01
3.65376286e-02 4.29371625e-01 3.97608936e-01 -5.70719302e-01
-2.07928885e-02 1.36085972e-01 1.75852925e-01 7.00435042e-01
9.62136611e-02 -6.94944501e-01 -1.10810208e+00 3.95019263e-01
-1.64864540e+00 -7.95853019e-01 -1.65095329e-01 1.98311698e+00
6.36415601e-01 3.27421516e-01 -1.58807501e-01 6.53024733e-01
5.83735526e-01 -9.15532708e-02 -4.67321187e-01 -1.00914216e+00
3.03407729e-01 3.31762642e-01 1.93553895e-01 2.56254047e-01
-1.23919654e+00 7.81401336e-01 5.86394739e+00 7.21399009e-01
-1.37381518e+00 -8.88108686e-02 8.35461676e-01 -1.09745085e-01
4.18294877e-01 -5.01603544e-01 -8.98210645e-01 3.99297714e-01
1.37514436e+00 3.39807928e-01 1.62615374e-01 1.21958745e+00
-2.03790084e-01 3.11902761e-02 -1.02938390e+00 9.15992200e-01
1.64027125e-01 -1.28659463e+00 6.07419126e-02 1.51441678e-01
6.37250900e-01 2.45166812e-02 1.45145684e-01 6.00675702e-01
5.28796278e-02 -1.08804715e+00 6.58648849e-01 3.92891735e-01
7.60161281e-01 -1.22498882e+00 1.03615654e+00 5.49456000e-01
-8.71701896e-01 -4.95487928e-01 -6.39990687e-01 1.13213377e-03
-3.84597123e-01 9.31417584e-01 -1.08637071e+00 2.86291361e-01
7.84002602e-01 2.77397007e-01 -6.29816353e-01 8.71551573e-01
2.04678942e-02 5.96187890e-01 -1.10652149e-01 -4.60146874e-01
5.87600648e-01 1.46524057e-01 -2.42968574e-01 1.65031993e+00
3.71526450e-01 -3.98135364e-01 1.05566129e-01 4.71717626e-01
-3.49922240e-01 9.86941084e-02 -3.28957319e-01 -4.53495473e-01
1.07241169e-01 1.59206283e+00 -9.88949656e-01 -6.61461532e-01
-5.22835314e-01 1.34707510e+00 4.77574676e-01 1.51831657e-01
-7.01797605e-01 -9.75549579e-01 2.04433009e-01 -2.90520459e-01
6.00097835e-01 -2.45178610e-01 -1.71968862e-01 -8.70602608e-01
-5.26543669e-02 -8.41700673e-01 3.29454422e-01 -4.73966807e-01
-9.24786031e-01 9.57554996e-01 -7.25975871e-01 -8.66917074e-01
-5.74225605e-01 -1.16319823e+00 -4.99557287e-01 6.23106122e-01
-1.61294556e+00 -1.15697312e+00 -5.78502774e-01 3.82618695e-01
7.25371540e-01 -5.08765757e-01 1.19141078e+00 4.18133974e-01
-8.10950160e-01 8.43032837e-01 7.76379406e-01 3.48467976e-01
3.42965811e-01 -1.46868598e+00 3.97524357e-01 5.74401915e-01
4.03738022e-01 3.09781104e-01 2.60486394e-01 -7.53894001e-02
-1.31195009e+00 -1.12686861e+00 1.07023680e+00 -1.12166576e-01
2.34961435e-01 -6.37158692e-01 -7.21209407e-01 4.12444085e-01
3.25144917e-01 1.78883933e-02 8.95809233e-01 7.72712054e-03
-2.63284177e-01 -2.61907846e-01 -1.07645631e+00 2.85281450e-01
3.48914564e-01 -4.24413472e-01 -9.98524055e-02 5.51687837e-01
1.72558919e-01 -3.03764671e-01 -7.02882469e-01 1.29079312e-01
7.29048193e-01 -9.90884483e-01 6.23937905e-01 -5.10415256e-01
5.99228442e-01 1.98846683e-01 1.60352383e-02 -1.01572764e+00
-4.35104102e-01 -1.55272752e-01 -4.45923895e-01 1.15439451e+00
4.20577556e-01 -3.32338393e-01 9.58015859e-01 2.29831681e-01
-4.73630950e-02 -1.08264482e+00 -3.55432361e-01 -7.72485793e-01
1.48923531e-01 -3.30086887e-01 4.47144806e-01 9.18234527e-01
-5.69972575e-01 2.02785388e-01 -1.66231953e-02 -1.60603017e-01
2.70932645e-01 4.26608086e-01 6.56425893e-01 -1.44712663e+00
-3.25309038e-01 -4.56849575e-01 -4.18445438e-01 -8.92031848e-01
2.04534426e-01 -8.30806792e-01 -8.33126828e-02 -1.71846223e+00
3.91712874e-01 -2.36282989e-01 -5.23355782e-01 5.58212221e-01
2.06543654e-01 2.51983225e-01 -4.66470839e-03 -2.03486774e-02
-4.76221144e-01 4.06030118e-01 8.29543650e-01 -3.91111910e-01
6.28165230e-02 2.21968204e-01 -5.38424194e-01 7.01480985e-01
9.21245992e-01 -3.48679274e-01 -2.61368275e-01 -7.37736762e-01
-2.66753472e-02 -4.73343223e-01 9.97028276e-02 -1.13150442e+00
2.98869431e-01 1.38083950e-01 1.15677941e+00 -9.54240918e-01
2.24680290e-01 -9.55056131e-01 -4.27576870e-01 8.33524883e-01
-4.92711276e-01 -1.49171606e-01 2.12410018e-01 2.25180939e-01
-3.00832063e-01 -7.67953455e-01 8.04360271e-01 -6.26754016e-02
-5.28204203e-01 1.17618762e-01 -7.82370687e-01 -5.99061549e-01
1.15626359e+00 -3.03601295e-01 -2.53038883e-01 -2.27287993e-01
-4.56309855e-01 -1.43792987e-01 1.49495900e-03 3.27339858e-01
5.46755850e-01 -1.02636135e+00 -6.44110382e-01 2.63259649e-01
7.50175561e-04 -1.27196625e-01 1.75514340e-01 4.67480034e-01
-8.99944961e-01 7.04221725e-01 -1.53258219e-01 -6.71569824e-01
-1.43946826e+00 4.40324455e-01 4.27614212e-01 -3.86833668e-01
-6.19454920e-01 1.02589428e+00 -2.40496397e-01 -5.39498746e-01
6.82116389e-01 -1.82047799e-01 -2.96790212e-01 1.01792514e-01
7.84349024e-01 3.24733943e-01 7.01344490e-01 -2.12891340e-01
-2.52246082e-01 2.57194579e-01 -6.23560369e-01 4.83934171e-02
1.62890983e+00 4.39558446e-01 1.04829587e-01 1.54689237e-01
1.54680288e+00 -4.07478064e-01 -9.76128101e-01 -2.07404822e-01
1.95247680e-01 -1.77346438e-01 6.83869541e-01 -9.45350826e-01
-1.25923419e+00 1.33360183e+00 8.76075327e-01 2.87594795e-01
1.42172503e+00 -4.73251760e-01 7.36906528e-01 1.09484649e+00
-2.39566013e-01 -1.48098922e+00 2.90630609e-01 6.34469867e-01
8.36099327e-01 -1.20502722e+00 -3.61726210e-02 1.81539699e-01
-2.80007929e-01 1.51574838e+00 6.96176410e-01 5.46297505e-02
4.41563457e-01 5.16482234e-01 3.58219743e-02 -2.14444742e-01
-8.95421922e-01 1.63246229e-01 1.55631289e-01 3.47652018e-01
9.06673133e-01 -6.36604577e-02 -3.53278406e-02 4.36367512e-01
-9.01221037e-02 3.00932694e-02 2.19054446e-01 1.27024138e+00
-4.76560444e-01 -9.43939269e-01 -2.44097099e-01 5.86723626e-01
-6.61661565e-01 -9.10978317e-02 -5.75722396e-01 4.34880584e-01
-9.74085554e-02 8.89951229e-01 4.42054808e-01 -4.01074618e-01
1.56983182e-01 1.80463478e-01 3.52290332e-01 -4.52593118e-01
-7.50294864e-01 -9.31404978e-02 -5.56070469e-02 -2.37776279e-01
-1.05834670e-01 -2.32428759e-01 -1.17725110e+00 -5.68093657e-01
-7.08626866e-01 6.27430901e-02 1.18545794e+00 8.60402882e-01
4.67310876e-01 7.44012892e-01 8.28933597e-01 -7.85145760e-01
-7.23095894e-01 -1.04726994e+00 -5.15455782e-01 4.94533591e-02
2.85140812e-01 -2.60798723e-01 1.30712390e-02 1.47168085e-01] | [11.440877914428711, 2.6288514137268066] |
9bdca7df-c1b5-4e3d-9841-b0fc6862d301 | bridging-anaphora-resolution-as-question | 2004.07898 | null | https://arxiv.org/abs/2004.07898v3 | https://arxiv.org/pdf/2004.07898v3.pdf | Bridging Anaphora Resolution as Question Answering | Most previous studies on bridging anaphora resolution (Poesio et al., 2004; Hou et al., 2013b; Hou, 2018a) use the pairwise model to tackle the problem and assume that the gold mention information is given. In this paper, we cast bridging anaphora resolution as question answering based on context. This allows us to find the antecedent for a given anaphor without knowing any gold mention information (except the anaphor itself). We present a question answering framework (BARQA) for this task, which leverages the power of transfer learning. Furthermore, we propose a novel method to generate a large amount of "quasi-bridging" training data. We show that our model pre-trained on this dataset and fine-tuned on a small amount of in-domain dataset achieves new state-of-the-art results for bridging anaphora resolution on two bridging corpora (ISNotes (Markert et al., 2012) and BASHI (Roesiger, 2018)). | ['Yufang Hou'] | 2020-04-16 | bridging-anaphora-resolution-as-question-1 | https://aclanthology.org/2020.acl-main.132 | https://aclanthology.org/2020.acl-main.132.pdf | acl-2020-6 | ['bridging-anaphora-resolution'] | ['natural-language-processing'] | [ 8.12960323e-03 7.39156067e-01 -5.46526968e-01 -3.66057485e-01
-1.50543821e+00 -7.14093745e-01 6.87119961e-01 1.63340032e-01
-3.83437812e-01 1.07178497e+00 5.88046551e-01 -1.93188295e-01
-3.81790936e-01 -9.19525981e-01 -8.47132921e-01 -1.24257416e-01
1.45280585e-01 1.31172943e+00 4.41956669e-01 -8.24490309e-01
2.54530758e-01 2.94012427e-01 -1.19665003e+00 7.03464687e-01
8.05584371e-01 5.00375390e-01 -3.32958996e-02 1.72164977e-01
-6.06188625e-02 1.01786172e+00 -6.16111338e-01 -9.23335254e-01
9.16550960e-03 -3.35522532e-01 -1.77083158e+00 -4.76726174e-01
8.16384315e-01 -1.12813078e-01 -4.79613841e-01 8.54353309e-01
4.35064465e-01 1.24936216e-02 6.51026130e-01 -9.59250391e-01
-6.32303536e-01 1.17077565e+00 -5.48275530e-01 4.25491452e-01
6.08725786e-01 -3.78375322e-01 1.51368701e+00 -8.34517539e-01
1.18528533e+00 1.34176767e+00 8.25679719e-01 7.95017302e-01
-1.10272706e+00 -6.30800426e-01 -1.72683105e-01 9.42419052e-01
-1.18158972e+00 -3.31703305e-01 8.30165684e-01 -6.39565513e-02
1.07086456e+00 1.49739638e-01 2.94921219e-01 1.25278556e+00
-1.49905547e-01 8.56839240e-01 7.95051813e-01 -8.09931755e-01
-2.17344500e-02 -4.34809476e-01 4.17324394e-01 2.62810528e-01
1.45732284e-01 -2.69933820e-01 -6.91240430e-01 -3.07885319e-01
4.54992890e-01 -5.51399648e-01 -3.21081996e-01 -4.24459696e-01
-8.80430222e-01 1.08089209e+00 5.97929657e-01 5.40643692e-01
-3.06460351e-01 -1.94995269e-01 4.37257499e-01 4.26060826e-01
1.65654480e-01 6.58467233e-01 -6.05508745e-01 -1.82715371e-01
-7.49886096e-01 8.21780324e-01 1.18213689e+00 9.94235694e-01
6.51681125e-01 -9.53400254e-01 -4.22014147e-02 7.72565246e-01
-4.26500142e-02 8.99659023e-02 2.48249516e-01 -1.48863721e+00
1.04942846e+00 5.14013469e-01 4.85484213e-01 -6.61559045e-01
-4.50052202e-01 1.78013295e-01 -2.90500615e-02 -3.57996941e-01
1.00985312e+00 4.55271453e-03 -2.16268107e-01 2.15815473e+00
4.85156476e-01 1.08670361e-01 5.29741287e-01 1.16579366e+00
9.98788595e-01 3.91964644e-01 1.48920357e-01 -3.05482805e-01
1.91597152e+00 -1.15461087e+00 -9.67147470e-01 -4.83217478e-01
5.71956813e-01 -5.90454996e-01 1.08693075e+00 2.44082183e-01
-1.42010057e+00 -6.73245415e-02 -1.04864931e+00 -6.35678649e-01
-1.55214295e-01 -1.83824003e-01 6.82008147e-01 5.66600896e-02
-6.22899294e-01 5.74553311e-01 -6.63951159e-01 -7.66649783e-01
3.81455034e-01 2.04469144e-01 -6.03952348e-01 -2.87780404e-01
-1.87851536e+00 1.36042559e+00 3.71234953e-01 -1.41995072e-01
-6.10508025e-01 -6.59357786e-01 -8.08421314e-01 1.39816731e-01
6.74581707e-01 -7.98945308e-01 1.54333782e+00 -6.32879615e-01
-1.23409367e+00 1.41468501e+00 -1.48004442e-01 -7.92259395e-01
3.30767423e-01 -7.00869322e-01 -5.76582365e-02 5.33485055e-01
5.70216656e-01 7.84832478e-01 3.19458604e-01 -1.12363696e+00
-5.10935783e-01 -6.89472497e-01 6.59379900e-01 3.20304245e-01
-9.02349725e-02 4.04540509e-01 -9.39247534e-02 -5.20370245e-01
1.95205584e-01 -7.51428843e-01 4.29613769e-01 -3.79528135e-01
-2.68888742e-01 -8.18879783e-01 6.31829023e-01 -7.51066506e-01
1.01612747e+00 -1.81058705e+00 5.29127121e-01 -4.01401430e-01
-5.30641973e-02 1.80120632e-01 -1.99581385e-01 7.33372152e-01
-1.61904141e-01 -5.52282855e-02 -5.73428392e-01 -2.40108356e-01
1.29040003e-01 4.84830528e-01 -7.26961851e-01 3.21052074e-01
3.09063643e-01 1.00270832e+00 -9.46388364e-01 -7.66642034e-01
-3.92692983e-01 2.10403666e-01 -6.46147788e-01 4.73967582e-01
-3.85047644e-01 4.46596324e-01 -2.03319177e-01 4.82821494e-01
5.76685190e-01 2.05224846e-02 5.79909325e-01 -5.27644932e-01
8.24950710e-02 8.96527171e-01 -1.06348038e+00 2.11933112e+00
-4.19173300e-01 4.20343697e-01 1.87132299e-01 -1.07705855e+00
8.15880001e-01 7.63643265e-01 2.35247284e-01 -7.05920994e-01
2.67361641e-01 4.26212668e-01 -4.47294526e-02 -6.24791145e-01
2.90525228e-01 -2.75378495e-01 -4.38387007e-01 8.36988926e-01
3.06506395e-01 -1.28899083e-01 5.41420400e-01 4.46571976e-01
1.20667922e+00 2.78794855e-01 3.55906874e-01 -3.64227355e-01
5.88859916e-01 4.41062987e-01 7.44318247e-01 6.23975933e-01
-2.35473812e-01 5.88706911e-01 6.67224944e-01 -4.55581963e-01
-7.82379389e-01 -1.19678724e+00 -4.43436712e-01 1.12274098e+00
3.22417468e-01 -4.53506917e-01 -1.14034247e+00 -8.31739604e-01
-1.09261267e-01 6.59126699e-01 -9.33697939e-01 -2.56191399e-02
-1.28147900e+00 -4.05671954e-01 8.82905602e-01 6.02890432e-01
3.63733321e-01 -1.49134374e+00 -7.51198351e-01 2.12822333e-01
-8.00192654e-01 -1.18702447e+00 -1.87342241e-01 1.48655042e-01
-6.71578467e-01 -1.49597847e+00 -8.78974944e-02 -9.15628135e-01
4.97049391e-02 -1.38995796e-01 1.46335244e+00 2.25668494e-02
-1.44688353e-01 1.07706062e-01 -2.76447088e-01 -1.11366533e-01
-2.12405354e-01 4.44865197e-01 -1.78775057e-01 -4.51504827e-01
8.28975260e-01 -6.57647550e-01 -2.65433609e-01 4.81778868e-02
-6.78078830e-01 -3.37746292e-01 2.06005678e-01 1.01411271e+00
4.38826770e-01 -6.16557002e-01 9.39841211e-01 -1.27846289e+00
6.03543103e-01 -6.38314724e-01 -2.82561839e-01 2.28597373e-01
-3.36405784e-01 -1.47291288e-01 3.22307110e-01 -3.85456145e-01
-1.40347469e+00 -4.62844849e-01 9.11387131e-02 -1.46531701e-01
-5.46183884e-02 4.02157098e-01 -5.36285937e-01 4.63350981e-01
8.51841450e-01 -6.45783305e-01 -3.67378533e-01 -8.57553482e-01
8.06499481e-01 6.59263194e-01 1.10924208e+00 -1.23189938e+00
5.60169756e-01 5.85377336e-01 -1.94914445e-01 1.39974117e-01
-1.69191527e+00 -4.05714989e-01 -1.04724860e+00 3.04985553e-01
8.72364163e-01 -9.08414960e-01 -5.57875991e-01 -4.10845391e-02
-1.62015057e+00 -3.30762953e-01 -2.37688705e-01 1.57283619e-01
-8.80317926e-01 2.06667379e-01 -1.28339422e+00 -2.25904331e-01
-5.07227480e-01 -7.43336022e-01 9.29404378e-01 2.01923132e-01
-7.66075611e-01 -7.37129569e-01 4.08433110e-01 9.03189600e-01
1.70637771e-01 8.52881745e-02 1.17032743e+00 -1.02729523e+00
-2.99033195e-01 1.33240268e-01 -4.82101470e-01 -3.34045708e-01
1.24006113e-02 -6.73146605e-01 -9.25590575e-01 -1.23194151e-01
1.55986443e-01 -7.32668936e-01 6.12102032e-01 2.80805794e-03
4.32302475e-01 -4.18504894e-01 -4.01275933e-01 2.28262365e-01
1.01516867e+00 -2.17324197e-01 7.71530986e-01 1.03794968e+00
3.82900894e-01 9.11966920e-01 1.09436393e+00 -1.07421502e-01
8.27877283e-01 1.02409887e+00 3.01116139e-01 4.17461485e-01
-3.41528028e-01 -1.00280285e-01 -4.02430296e-02 4.78117704e-01
-5.29991575e-02 3.05174757e-02 -9.60007131e-01 9.81234789e-01
-2.26263213e+00 -1.13835037e+00 -2.46379673e-01 1.71447086e+00
1.65473986e+00 -1.69876069e-02 -9.41452682e-02 1.16327167e-01
8.32838833e-01 -2.34002516e-01 -1.25299051e-01 -4.18172330e-01
-3.64071280e-01 7.33045757e-01 -1.67722076e-01 7.94589579e-01
-1.24442863e+00 1.44875276e+00 5.33010864e+00 5.67734778e-01
-4.73678440e-01 4.55654323e-01 1.63831394e-02 1.17197327e-01
-1.47819608e-01 4.77563739e-01 -6.79386556e-01 7.25398064e-02
8.50290418e-01 -4.69545163e-02 4.20051157e-01 3.85436684e-01
-5.47366500e-01 -3.78566943e-02 -1.45117748e+00 4.11393821e-01
2.25898117e-01 -1.26806164e+00 -7.21178129e-02 -3.96627873e-01
5.02266586e-01 -1.30437687e-01 -4.98326451e-01 4.98311013e-01
1.86444908e-01 -9.75533247e-01 6.85716271e-01 2.25414827e-01
4.12179083e-01 -6.37071252e-01 9.98360097e-01 2.32992902e-01
-6.75268829e-01 -2.19785273e-02 -2.58825898e-01 -1.62010565e-01
4.09500003e-01 1.32346049e-01 -5.29346049e-01 6.74593747e-01
8.00747097e-01 5.02872825e-01 -4.51147050e-01 4.88830507e-01
-9.91192281e-01 6.32154882e-01 -2.10894108e-01 5.01542807e-01
-4.99345884e-02 -2.78663654e-02 6.68631196e-01 1.10790098e+00
-1.56797156e-01 4.64219689e-01 -2.66675930e-02 1.08206201e+00
-4.22701955e-01 9.19548795e-02 -9.83439535e-02 4.58354145e-01
1.02908456e+00 1.31499028e+00 -2.82385170e-01 -2.05547348e-01
-3.42669636e-01 7.25749671e-01 9.48158324e-01 5.14521524e-02
-6.86167598e-01 -5.09052873e-01 1.99672192e-01 3.20877768e-02
3.95210087e-01 3.60577643e-01 -2.65889168e-01 -9.71125484e-01
5.97844571e-02 -1.05000353e+00 1.13039052e+00 -7.98204541e-01
-1.57951987e+00 3.22239429e-01 2.69944489e-01 -6.27450168e-01
-4.79399294e-01 -3.55736792e-01 -6.73371255e-01 7.44274914e-01
-1.60221100e+00 -1.62912071e+00 -1.54625669e-01 5.30399919e-01
3.26617956e-01 1.45773560e-01 9.98388052e-01 4.55577046e-01
-2.78079331e-01 5.24894357e-01 -6.15836740e-01 4.45040882e-01
1.27816880e+00 -1.23639619e+00 2.16035962e-01 4.76607919e-01
1.49946436e-01 9.16566849e-01 7.75063872e-01 -5.31663656e-01
-1.32119465e+00 -6.34651721e-01 1.46279931e+00 -9.23855722e-01
1.05078077e+00 -1.53747723e-01 -1.43034673e+00 1.20287228e+00
6.35924578e-01 -1.13068089e-01 5.34226894e-01 5.02774954e-01
-8.74515593e-01 1.03250705e-01 -1.33183312e+00 4.40034598e-01
1.23393428e+00 -6.32554948e-01 -1.73815978e+00 1.81822285e-01
7.77069211e-01 -7.24091291e-01 -1.16797221e+00 8.08409154e-01
8.91197920e-02 -8.21786284e-01 9.61319327e-01 -1.07501841e+00
6.73955023e-01 -1.64895073e-01 -3.34512562e-01 -7.96770871e-01
1.16172604e-01 -5.84309280e-01 -4.32018071e-01 1.51442873e+00
3.41409743e-01 -3.34572047e-01 5.18366158e-01 5.14429629e-01
-1.61679164e-01 -4.91268039e-01 -1.44427311e+00 -3.48773688e-01
7.36895442e-01 3.48271802e-02 6.73564851e-01 1.34884179e+00
8.05207908e-01 8.91085207e-01 1.80131227e-01 2.38688067e-01
5.80286086e-01 7.34699965e-01 6.91434860e-01 -1.46413147e+00
-1.97633773e-01 -1.47593975e-01 9.54252854e-02 -7.22872972e-01
8.25616479e-01 -9.99461114e-01 -1.09745286e-01 -1.53390455e+00
5.21909118e-01 -3.38083744e-01 -1.02864981e-01 6.71447337e-01
-3.17150027e-01 3.49849731e-01 6.94794133e-02 4.71835971e-01
-7.13492692e-01 3.16148311e-01 8.87991428e-01 -1.78397164e-01
1.89688370e-01 -5.49130261e-01 -7.22932398e-01 6.18712723e-01
6.65064812e-01 -5.29900014e-01 1.03690647e-01 -5.70134282e-01
2.53092706e-01 2.68324047e-01 3.57901067e-01 -4.80057508e-01
4.35298204e-01 -4.27718125e-02 -3.20311844e-01 -2.32568845e-01
3.84966999e-01 -4.15713578e-01 -3.15062821e-01 1.60631090e-01
-4.50171947e-01 -2.93189995e-02 2.69849688e-01 2.15940386e-01
-3.66104096e-01 -7.05668390e-01 4.55556691e-01 7.18747675e-02
-4.55380499e-01 -2.36472130e-01 -2.83714719e-02 7.87457228e-01
7.81511843e-01 4.02849287e-01 -8.65167081e-01 -2.60472506e-01
-8.15817714e-01 3.61500740e-01 3.86015534e-01 3.70601743e-01
8.01428705e-02 -1.36083782e+00 -8.84792149e-01 -5.76231897e-01
2.03139529e-01 3.01661909e-01 -1.60196063e-03 8.30028832e-01
-1.83924258e-01 3.99666399e-01 -2.78952599e-01 -2.08064079e-01
-1.01948643e+00 5.79636574e-01 1.88731417e-01 -5.62623680e-01
-5.90339899e-01 6.08212888e-01 -2.05341101e-01 -6.25438392e-01
1.57230534e-02 2.19291836e-01 -4.84626740e-01 3.66027832e-01
4.59709972e-01 1.87636703e-01 1.50084272e-01 -6.18072271e-01
-4.80669856e-01 4.15242344e-01 -2.87469059e-01 -1.48487926e-01
1.42739379e+00 -2.33324207e-02 -5.53098917e-01 2.15952113e-01
7.91457176e-01 3.20495754e-01 -7.00267434e-01 -4.38957661e-01
7.08498180e-01 -1.62815854e-01 -3.95394683e-01 -9.07204986e-01
-7.30705798e-01 8.87795627e-01 1.32807344e-02 -3.75088267e-02
7.39986300e-01 7.29635596e-01 9.38629270e-01 5.92856050e-01
5.27430594e-01 -9.88797963e-01 1.14752479e-01 7.01276064e-01
1.00170100e+00 -1.01565623e+00 -2.00230926e-01 -7.47787535e-01
-3.57505381e-01 1.15428638e+00 9.49174881e-01 -1.90171108e-01
6.38632625e-02 4.29959506e-01 2.32130036e-01 -5.78074813e-01
-7.21179664e-01 -2.70113677e-01 -1.13928221e-01 5.89923143e-01
5.14382660e-01 -1.61248401e-01 -8.39768767e-01 1.29232979e+00
-3.29045057e-01 -1.36415347e-01 5.30931652e-01 9.15584087e-01
-3.09576541e-01 -1.36333513e+00 -4.21325892e-01 -1.78897649e-01
-6.75700188e-01 -2.28295550e-01 -7.29101121e-01 1.03634703e+00
-5.93428090e-02 9.75168824e-01 2.76987553e-01 2.65424401e-01
6.05754733e-01 2.99100727e-01 1.03656328e+00 -7.81494796e-01
-6.17283046e-01 -3.88780236e-01 6.14991784e-01 -4.87491965e-01
-7.72419572e-01 -7.58905530e-01 -1.68177032e+00 -5.87863103e-02
-4.89752561e-01 7.23940670e-01 -6.07362874e-02 1.33599830e+00
2.69170851e-01 1.26343533e-01 4.19260468e-03 -4.83309329e-01
-6.11307442e-01 -1.29188383e+00 -2.19523847e-01 8.83924484e-01
-1.19838789e-01 -9.25453842e-01 -3.38774979e-01 8.41402560e-02] | [9.363304138183594, 9.49085521697998] |
fcbcc5c0-6c15-4278-95c2-27ac832412aa | fisheyesuperpoint-keypoint-detection-and | 2103.00191 | null | https://arxiv.org/abs/2103.00191v2 | https://arxiv.org/pdf/2103.00191v2.pdf | FisheyeSuperPoint: Keypoint Detection and Description Network for Fisheye Images | Keypoint detection and description is a commonly used building block in computer vision systems particularly for robotics and autonomous driving. However, the majority of techniques to date have focused on standard cameras with little consideration given to fisheye cameras which are commonly used in urban driving and automated parking. In this paper, we propose a novel training and evaluation pipeline for fisheye images. We make use of SuperPoint as our baseline which is a self-supervised keypoint detector and descriptor that has achieved state-of-the-art results on homography estimation. We introduce a fisheye adaptation pipeline to enable training on undistorted fisheye images. We evaluate the performance on the HPatches benchmark, and, by introducing a fisheye based evaluation method for detection repeatability and descriptor matching correctness, on the Oxford RobotCar dataset. | ['Senthil Yogamani', 'Rudi Villing', 'John McDonald', 'Ganesh Sistu', 'Ciarán Eising', 'Anna Konrad'] | 2021-02-27 | null | null | null | null | ['homography-estimation'] | ['computer-vision'] | [-1.35875314e-01 -8.80584493e-02 -1.70519233e-01 -4.57651526e-01
-6.77937806e-01 -5.36588550e-01 1.05849946e+00 -2.27777194e-02
-8.21878135e-01 1.04754247e-01 -1.40725195e-01 -1.88707665e-01
1.35845646e-01 -4.97915477e-01 -8.75997722e-01 -3.09265763e-01
6.38812855e-02 3.28032762e-01 8.06888998e-01 -4.52956587e-01
6.86501086e-01 7.14543700e-01 -1.78957605e+00 -2.19653726e-01
4.43051189e-01 8.93562555e-01 3.11930746e-01 7.17040062e-01
6.07969999e-01 4.89622474e-01 -1.57157928e-01 -3.81859899e-01
6.46436512e-01 -3.69015783e-02 -2.17545345e-01 -4.35324945e-02
1.17636788e+00 -5.93764067e-01 -5.19657552e-01 9.11900461e-01
2.87328392e-01 2.59417780e-02 5.93194067e-01 -1.46637809e+00
-2.80164778e-01 -1.05846263e-02 -3.47452074e-01 1.73943296e-01
2.19495803e-01 2.66623855e-01 9.01478946e-01 -1.11596096e+00
9.02457476e-01 1.07972193e+00 9.03423369e-01 1.50401294e-01
-8.46779644e-01 -7.05462217e-01 -4.28362787e-01 6.49187684e-01
-1.29652166e+00 -6.97805941e-01 4.81073290e-01 -3.91181022e-01
1.26854849e+00 -4.16326642e-01 4.80670661e-01 6.13093436e-01
6.78768814e-01 7.35170901e-01 1.02282429e+00 -4.39770758e-01
1.84663525e-03 3.03465098e-01 4.77143228e-02 8.35373938e-01
2.28632808e-01 6.40766144e-01 -4.06489134e-01 4.49063629e-01
5.54028273e-01 -2.29976568e-02 2.81463321e-02 -1.28438723e+00
-1.38189149e+00 1.03751767e+00 5.08072853e-01 -3.14865708e-01
-1.29556373e-01 2.63277978e-01 3.09290469e-01 3.37907135e-01
-2.81824693e-02 4.41731513e-01 -1.68776616e-01 -3.41651440e-01
-8.83096993e-01 5.28015435e-01 7.43049741e-01 1.37106216e+00
1.00439239e+00 -3.35294634e-01 4.00038093e-01 5.68907082e-01
4.61813062e-01 1.02191496e+00 1.44168034e-01 -1.22204626e+00
3.10195446e-01 4.98152196e-01 1.48144990e-01 -1.02615047e+00
-4.13181037e-01 1.13102779e-01 -1.30378246e-01 7.95220137e-01
2.18937948e-01 2.85469204e-01 -9.27691340e-01 7.88912833e-01
2.30022073e-01 -6.13505840e-02 1.10264502e-01 9.70240116e-01
8.19803059e-01 2.82673091e-01 -4.72437561e-01 3.50876242e-01
1.13261557e+00 -1.23033249e+00 -2.66315848e-01 -5.50477087e-01
7.49472618e-01 -9.40437019e-01 7.87825704e-01 4.63672161e-01
-6.90894485e-01 -5.01151562e-01 -1.48618531e+00 -3.67325515e-01
-6.07369721e-01 2.85284370e-01 1.85232118e-01 6.60383046e-01
-1.00877190e+00 3.72017950e-01 -7.98611343e-01 -7.96779513e-01
-5.25876507e-02 2.87345290e-01 -8.59755933e-01 -2.06623405e-01
-1.00718570e+00 1.71401346e+00 2.35216945e-01 -3.40170592e-01
-8.55985284e-01 -5.70176840e-01 -1.20864677e+00 -2.04368681e-01
4.95964378e-01 -2.95681030e-01 1.48665333e+00 -3.23350057e-02
-1.60263491e+00 1.07879519e+00 1.87304109e-01 -1.09930563e+00
4.97664660e-01 -4.36766744e-01 -2.33077809e-01 2.23294124e-01
1.93841800e-01 1.26092708e+00 9.05246854e-01 -1.04131556e+00
-1.16279757e+00 -2.10772902e-01 -9.98049304e-02 2.44248554e-01
3.28752726e-01 -7.39075765e-02 -4.41395193e-01 7.36175552e-02
-4.07049432e-02 -1.42873859e+00 5.92612736e-02 1.15165420e-01
-9.24085006e-02 -9.21908766e-02 1.28396583e+00 -3.01943094e-01
6.48508847e-01 -2.14958191e+00 -1.07018642e-01 9.68324542e-02
-3.72822136e-02 2.00481057e-01 1.93716675e-01 5.43676257e-01
3.69110614e-01 -6.09412313e-01 -5.20358048e-02 -3.78318757e-01
1.89244315e-01 1.66554719e-01 -2.79147953e-01 7.62314320e-01
2.06767589e-01 7.44365871e-01 -8.28014135e-01 -3.82540554e-01
1.01158977e+00 3.54758948e-01 -4.75955188e-01 2.51526069e-02
2.13107988e-01 -1.12653123e-02 9.82707441e-02 7.32645214e-01
6.59979999e-01 3.24449569e-01 -5.21626055e-01 -1.48740083e-01
-7.40752399e-01 1.29818723e-01 -1.07748282e+00 1.60236418e+00
-4.79479462e-01 1.13521695e+00 -1.09621584e-01 -5.06726980e-01
1.27011800e+00 -4.34188277e-01 1.04067326e-01 -7.60687351e-01
1.69637501e-01 4.61158931e-01 -1.48784444e-01 -1.51584834e-01
1.18606162e+00 2.89645761e-01 -3.45966399e-01 -1.47707045e-01
1.46997511e-01 -8.54852021e-01 3.31795722e-01 2.19526693e-01
1.09340882e+00 3.45745832e-01 5.20832062e-01 -2.01910943e-01
5.23004234e-01 6.82680488e-01 1.61918513e-02 7.30599344e-01
-4.94675666e-01 8.96697819e-01 1.80612758e-01 -3.49938422e-01
-1.57495248e+00 -7.86393225e-01 -2.53681034e-01 6.77875161e-01
6.08920157e-01 -6.11509383e-01 -5.53127170e-01 -5.61445713e-01
4.21762586e-01 6.72661424e-01 -5.11772394e-01 -1.63717613e-01
-4.81946379e-01 1.21369675e-01 5.06211519e-01 3.89376342e-01
8.59792054e-01 -8.59303117e-01 -1.47092164e+00 1.12609742e-02
5.03633440e-01 -1.52424634e+00 -3.63283604e-01 2.70525634e-01
-3.97219837e-01 -1.32467413e+00 -5.75272799e-01 -5.19232512e-01
1.99778721e-01 9.46577549e-01 9.45201278e-01 -4.10530359e-01
-1.65723041e-01 7.37995148e-01 -3.84601384e-01 -5.90505004e-01
-4.23070461e-01 7.74838701e-02 9.71953571e-02 -5.32803178e-01
6.82208955e-01 8.69097188e-02 -7.69658685e-01 7.49818742e-01
-4.74385887e-01 -2.32031107e-01 1.06274748e+00 7.61889338e-01
4.78785217e-01 -4.43679333e-01 -3.56615067e-01 -5.57553470e-01
1.29066437e-01 -1.33823231e-01 -1.14515793e+00 -1.69484690e-01
-1.07776904e+00 1.50435388e-01 3.96175265e-01 1.18722528e-01
-5.62863052e-01 6.92922056e-01 7.28467256e-02 -6.58142090e-01
-3.80826861e-01 2.26173356e-01 1.12270288e-01 -7.35076547e-01
6.33804917e-01 1.21097423e-01 3.97279561e-01 -2.13992164e-01
5.96844733e-01 7.36325860e-01 1.05277145e+00 1.18739329e-01
8.42485666e-01 6.11618340e-01 2.74122745e-01 -1.14590657e+00
-1.49106175e-01 -1.13056481e+00 -1.02435327e+00 -1.66956678e-01
7.53234863e-01 -1.06812310e+00 -5.58090329e-01 6.49714410e-01
-8.86102974e-01 -1.98282212e-01 9.62204412e-02 6.22736275e-01
-7.60857344e-01 4.82609481e-01 2.40657553e-01 -4.70617086e-01
-4.58307452e-02 -1.37086403e+00 1.54725266e+00 2.23470166e-01
1.30022496e-01 -7.46476114e-01 4.17898446e-01 1.47537962e-01
2.87165642e-01 3.14362645e-02 3.35024372e-02 -4.99403596e-01
-8.62430215e-01 -5.64603567e-01 -4.15413886e-01 2.74919242e-01
-4.47999507e-01 -3.44906747e-02 -8.37662399e-01 -2.06727624e-01
-6.04504704e-01 -2.61917055e-01 1.21723342e+00 1.59242004e-01
9.03454274e-02 3.10205102e-01 -5.06896496e-01 7.19526231e-01
1.32021916e+00 -3.75551684e-03 8.12915802e-01 1.06264293e+00
5.44300675e-01 6.25108004e-01 1.11454320e+00 1.41028345e-01
8.38749349e-01 9.07163680e-01 6.33710206e-01 1.37362078e-01
1.63455252e-02 -4.07602966e-01 5.66270292e-01 1.84268609e-01
1.65089935e-01 2.36620143e-01 -1.17275238e+00 6.76856756e-01
-1.82488859e+00 -8.24487984e-01 -2.90789038e-01 2.29193068e+00
2.64508098e-01 3.74248117e-01 1.38548732e-01 4.93628159e-02
4.85321462e-01 -5.15020406e-03 -3.68186891e-01 -3.97433043e-01
1.28796533e-01 -9.16170329e-02 1.45894372e+00 4.25512522e-01
-1.47402275e+00 1.22326946e+00 6.17984819e+00 3.00898373e-01
-1.20197356e+00 -1.02239817e-01 -2.01124221e-01 3.67831916e-01
4.05999601e-01 3.53908837e-01 -1.54351830e+00 6.63465634e-02
1.02012134e+00 1.47551699e-02 6.09097518e-02 1.31543911e+00
8.61130729e-02 -6.61796451e-01 -1.17079043e+00 1.14242411e+00
3.97613376e-01 -1.14905834e+00 -5.73301494e-01 1.50278881e-01
5.85576892e-01 5.14721334e-01 1.08265929e-01 3.19491059e-01
2.43508652e-01 -6.51514649e-01 6.43919766e-01 2.27629825e-01
6.00463688e-01 -7.31696308e-01 8.50723565e-01 2.20899284e-01
-1.11051536e+00 8.48529935e-02 -5.82830787e-01 9.43924114e-02
8.82429779e-02 -8.61801431e-02 -1.39960015e+00 2.84009367e-01
9.35502112e-01 1.04837358e+00 -1.01485848e+00 1.38785625e+00
-3.16389471e-01 -2.69120112e-02 -5.41941881e-01 -1.27212599e-01
5.68551838e-01 -4.22624834e-02 6.18975163e-01 1.36509264e+00
4.26349670e-01 -3.12611014e-01 2.14712247e-01 3.93184811e-01
1.68235496e-01 -6.72760531e-02 -1.18183625e+00 3.15439463e-01
7.52472103e-01 1.41291428e+00 -6.37228251e-01 -1.27991334e-01
-7.73896039e-01 7.01158464e-01 5.01884036e-02 -2.91096747e-01
-6.24535680e-01 -9.05533135e-01 6.49646163e-01 2.40247756e-01
7.10903406e-01 -7.06882000e-01 -2.42023822e-02 -7.96209395e-01
-2.12682948e-01 -4.73971546e-01 -5.92842512e-03 -8.68898273e-01
-5.30476272e-01 1.53344542e-01 4.93190289e-01 -1.55041504e+00
-6.29829943e-01 -1.02190602e+00 -3.73144895e-01 5.96282840e-01
-1.84483361e+00 -1.31784081e+00 -7.27641463e-01 3.28410178e-01
5.23755193e-01 -4.03279036e-01 3.84830236e-01 -4.22812216e-02
1.98584329e-02 5.21525323e-01 3.40765774e-01 -1.22015968e-01
1.27148569e+00 -1.42073143e+00 8.89374554e-01 1.02209234e+00
3.90028842e-02 5.42322099e-01 1.04179633e+00 -3.98916781e-01
-1.83384359e+00 -9.68918800e-01 5.80225766e-01 -8.85816634e-01
8.65582824e-01 -2.55960286e-01 -5.43313026e-01 6.03850007e-01
1.89583600e-01 1.19313307e-01 -8.25966448e-02 -3.90569389e-01
-2.20016882e-01 -1.77624166e-01 -9.89416420e-01 1.61886722e-01
7.30923057e-01 -4.30804580e-01 -6.57810926e-01 1.16001312e-02
2.34305069e-01 -7.93019414e-01 -7.55416870e-01 4.07142043e-01
8.83284032e-01 -1.25150168e+00 9.11317110e-01 2.81202108e-01
3.24118316e-01 -5.24482191e-01 -1.18002929e-01 -1.38433886e+00
-1.65275007e-01 -3.62258375e-01 1.91082984e-01 7.56949246e-01
1.68891609e-01 -5.99316418e-01 8.98280859e-01 2.37787381e-01
-5.54972291e-01 -1.41829059e-01 -9.30068672e-01 -8.86978149e-01
-2.33658731e-01 -1.20960467e-01 7.64285997e-02 3.61271113e-01
-2.80937329e-02 3.40562552e-01 -5.55128694e-01 1.86412632e-01
7.50629425e-01 -1.21715426e-01 1.61233354e+00 -1.20147204e+00
1.61891162e-01 -3.72302413e-01 -1.20935977e+00 -1.14411032e+00
6.06926046e-02 -6.34069860e-01 5.80789030e-01 -1.34340429e+00
-6.97867051e-02 -5.92618585e-02 2.79439777e-01 1.28776893e-01
1.73679858e-01 5.83663464e-01 2.57800788e-01 5.28284311e-01
-8.08510721e-01 3.35850358e-01 7.69325972e-01 3.29693104e-03
-1.35186970e-01 -2.24427864e-01 -1.06600158e-01 5.65761566e-01
5.91743231e-01 -1.97483689e-01 -2.37300426e-01 -2.12748259e-01
8.35944042e-02 -3.48749161e-01 6.94466650e-01 -1.31940079e+00
6.55169725e-01 2.53670216e-01 1.13682680e-01 -1.16387260e+00
3.62910241e-01 -8.69902611e-01 -1.73932880e-01 5.14333367e-01
7.09051341e-02 4.60279316e-01 1.31384641e-01 6.10029519e-01
-2.25982055e-01 -1.95218742e-01 9.76800442e-01 -6.34199530e-02
-1.77083802e+00 4.30695191e-02 -2.57762939e-01 -2.01575339e-01
1.23326111e+00 -5.83544731e-01 -6.00083172e-01 -3.46483439e-01
6.06410839e-02 4.12949443e-01 1.08617449e+00 6.31458044e-01
9.92689908e-01 -1.09562266e+00 -5.54812908e-01 4.33152378e-01
7.76535690e-01 -9.14271995e-02 -3.67704272e-01 9.07629371e-01
-1.07985640e+00 1.01007676e+00 -5.87908208e-01 -9.77552056e-01
-1.41031408e+00 4.97320056e-01 1.97906956e-01 2.81765014e-01
-7.77074873e-01 5.13881028e-01 -1.24857090e-01 -5.57150066e-01
6.30498901e-02 -4.73417491e-01 -2.97061533e-01 -2.00483590e-01
1.77416727e-01 6.13311589e-01 3.24396044e-01 -1.15002084e+00
-5.43106318e-01 7.90755212e-01 -2.22871646e-01 -2.18677163e-01
1.06724715e+00 -2.80576706e-01 2.99129128e-01 2.67577201e-01
1.11188197e+00 -1.15390427e-01 -1.41544557e+00 -4.01584953e-02
2.92710572e-01 -6.41651750e-01 3.41450185e-01 -3.35823298e-01
-2.89317459e-01 9.89668310e-01 1.03969967e+00 1.71723571e-02
3.89310271e-01 -4.44479473e-02 7.75289178e-01 9.29974139e-01
5.86066246e-01 -1.16486049e+00 -2.17694938e-01 8.75742733e-01
7.46548533e-01 -1.65529716e+00 2.24109948e-01 -2.22499803e-01
-7.83302367e-01 1.36863303e+00 5.00625372e-01 -5.03709674e-01
4.28625852e-01 2.71052778e-01 1.91299900e-01 -1.64193138e-01
-5.68279088e-01 -4.31189239e-01 3.19752932e-01 8.65229607e-01
1.10745039e-02 -2.10256428e-01 -8.05423558e-02 -2.73136705e-01
-4.28104520e-01 -9.36349854e-02 6.78019941e-01 1.19339705e+00
-1.01581824e+00 -1.00117290e+00 -4.90995437e-01 9.10696164e-02
3.18611711e-02 9.32919886e-03 -6.17960453e-01 1.37563515e+00
-1.39417335e-01 7.37461209e-01 1.23654664e-01 -6.61044717e-01
8.04896533e-01 -1.70755953e-01 6.60842597e-01 -5.18634856e-01
-1.73651174e-01 -2.91111618e-01 1.99071050e-01 -7.51194239e-01
-3.16932946e-01 -8.85026813e-01 -9.90383506e-01 -2.96853960e-01
-5.10092735e-01 -5.47226220e-02 1.11321533e+00 7.31854975e-01
3.98324311e-01 -1.23938136e-01 5.49594343e-01 -1.38677061e+00
-7.44446278e-01 -8.37527394e-01 -4.18684781e-01 2.25191474e-01
4.73013520e-01 -1.03271616e+00 -2.51863480e-01 3.16866185e-03] | [7.563199996948242, -2.0980358123779297] |
ad1786e6-519d-493b-94d8-6215d49c5e89 | refvsr-exploiting-reference-inputs-for | 2307.02897 | null | https://arxiv.org/abs/2307.02897v1 | https://arxiv.org/pdf/2307.02897v1.pdf | RefVSR++: Exploiting Reference Inputs for Reference-based Video Super-resolution | Smartphones equipped with a multi-camera system comprising multiple cameras with different field-of-view (FoVs) are becoming more prevalent. These camera configurations are compatible with reference-based SR and video SR, which can be executed simultaneously while recording video on the device. Thus, combining these two SR methods can improve image quality. Recently, Lee et al. have presented such a method, RefVSR. In this paper, we consider how to optimally utilize the observations obtained, including input low-resolution (LR) video and reference (Ref) video. RefVSR extends conventional video SR quite simply, aggregating the LR and Ref inputs over time in a single bidirectional stream. However, considering the content difference between LR and Ref images due to their FoVs, we can derive the maximum information from the two image sequences by aggregating them independently in the temporal direction. Then, we propose an improved method, RefVSR++, which can aggregate two features in parallel in the temporal direction, one for aggregating the fused LR and Ref inputs and the other for Ref inputs over time. Furthermore, we equip RefVSR++ with enhanced mechanisms to align image features over time, which is the key to the success of video SR. We experimentally show that RefVSR++ outperforms RefVSR by over 1dB in PSNR, achieving the new state-of-the-art. | ['Takayuki Okatani', 'Masanori Suganuma', 'Han Zou'] | 2023-07-06 | null | null | null | null | ['video-super-resolution', 'reference-based-video-super-resolution', 'super-resolution'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 5.90605676e-01 -6.07104123e-01 -1.67954385e-01 -8.26827139e-02
-8.52393508e-01 -4.54049051e-01 2.61993140e-01 -4.83728409e-01
-3.34215373e-01 3.88428748e-01 2.92135179e-01 -1.00350179e-01
-5.14595062e-02 -5.09835482e-01 -6.90862298e-01 -7.02858329e-01
2.85536736e-01 -5.20493805e-01 6.10148787e-01 -1.55191779e-01
2.39132345e-01 4.08967704e-01 -1.78295708e+00 5.43945312e-01
6.81692362e-01 1.25046349e+00 7.89572179e-01 8.75203907e-01
1.37866229e-01 9.68585134e-01 -3.93763393e-01 -2.00194657e-01
5.03052950e-01 -3.05459619e-01 -2.97588646e-01 2.47570485e-01
6.78842902e-01 -7.57994413e-01 -6.95443332e-01 9.30434346e-01
4.56032157e-01 2.41483867e-01 -1.39522152e-02 -9.72623229e-01
-5.34977794e-01 2.14491576e-01 -9.28675592e-01 4.32717144e-01
9.67546225e-01 2.00057805e-01 7.88951755e-01 -8.27025950e-01
7.29421377e-01 1.18004513e+00 3.75661880e-01 3.77252311e-01
-1.08750987e+00 -7.11439431e-01 1.11941755e-01 4.49769825e-01
-1.51213646e+00 -7.46561587e-01 6.84913635e-01 -1.68411359e-01
6.78803444e-01 4.18315619e-01 5.83899200e-01 9.54773366e-01
1.47707343e-01 6.36436343e-01 1.08699763e+00 -2.02303991e-01
-6.08172193e-02 -2.51933813e-01 -1.13542318e-01 5.18017337e-02
1.45596966e-01 1.71847418e-01 -9.28704381e-01 1.53623046e-02
9.78402913e-01 4.62108821e-01 -8.50055933e-01 8.43120068e-02
-1.53274786e+00 3.07025820e-01 9.15558822e-03 4.19554502e-01
-3.05706173e-01 8.89689475e-02 9.35873091e-02 4.57465827e-01
1.76889017e-01 -7.30990171e-02 -2.00573340e-01 -1.39605299e-01
-1.12347436e+00 -1.70741349e-01 1.57310665e-01 1.22482896e+00
5.75591266e-01 3.00460514e-02 -1.13869347e-01 9.09942508e-01
2.96086729e-01 9.26250637e-01 4.61901605e-01 -1.44862473e+00
6.59512818e-01 1.53656080e-01 2.11139321e-01 -1.05064547e+00
7.30290934e-02 -1.75082043e-01 -9.64341342e-01 -8.16268548e-02
7.04996139e-02 1.71884760e-01 -3.72213364e-01 1.40274167e+00
9.54976603e-02 5.93510866e-01 3.92166853e-01 9.65141177e-01
8.35524082e-01 8.33849669e-01 -5.65307319e-01 -7.24937320e-01
1.15138340e+00 -8.66659343e-01 -9.06953096e-01 3.55952941e-02
1.47145063e-01 -1.00254273e+00 9.05172408e-01 7.91246653e-01
-1.36074841e+00 -9.15805757e-01 -1.20514715e+00 -2.52155662e-02
3.12969357e-01 3.05743158e-01 -7.59777278e-02 3.82437974e-01
-1.38066101e+00 5.04873335e-01 -4.63975996e-01 -8.16239119e-02
3.15299910e-03 3.88100445e-01 -5.20674407e-01 -5.30592084e-01
-1.02089942e+00 4.29410994e-01 -4.17019427e-02 1.55980378e-01
-6.12726212e-01 -4.74272400e-01 -6.19270980e-01 -3.99305001e-02
6.55512393e-01 -5.55406749e-01 1.14487076e+00 -8.42126727e-01
-1.58478141e+00 4.90466177e-01 -5.50120473e-01 -2.27018073e-01
4.48179781e-01 -3.48791003e-01 -8.40893626e-01 5.79169095e-01
-7.79359266e-02 2.12246701e-01 1.10972261e+00 -1.37403142e+00
-9.41747010e-01 -2.85526544e-01 1.91144288e-01 4.18215573e-01
-2.98787028e-01 2.22620666e-01 -9.19917345e-01 -5.62801480e-01
3.38597506e-01 -7.60024667e-01 -4.38014045e-02 -7.94761777e-02
-3.12682800e-02 2.58823454e-01 1.05999243e+00 -6.69086039e-01
1.40284395e+00 -2.54024434e+00 3.03354859e-01 -8.63936990e-02
3.72086257e-01 3.11574459e-01 -2.27187946e-01 2.06652939e-01
1.19535722e-01 -1.64855033e-01 6.19250946e-02 -3.25027585e-01
-7.04769969e-01 1.80237621e-01 -5.06221890e-01 5.24259448e-01
-1.72993869e-01 4.38254654e-01 -1.01758802e+00 -4.98968661e-01
5.15565634e-01 4.47481722e-01 -2.74710417e-01 3.03633630e-01
3.84960920e-01 8.12256396e-01 -1.48871735e-01 5.46624124e-01
1.00543964e+00 -3.34834039e-01 3.48510474e-01 -5.97226262e-01
-2.32636869e-01 -1.68320999e-01 -1.51015472e+00 1.54872072e+00
-6.62762761e-01 7.00192213e-01 1.37424126e-01 -5.10261595e-01
9.78616655e-01 3.20500135e-01 6.57820523e-01 -9.05000210e-01
-2.37268165e-01 3.52526546e-01 -3.80852163e-01 -4.81637478e-01
7.97787130e-01 1.78789333e-01 2.63091922e-01 2.66587913e-01
-2.69365404e-02 6.48623332e-02 2.73012847e-01 3.19536746e-01
1.05439210e+00 4.22170348e-02 4.64670777e-01 2.72976607e-01
1.03264189e+00 -6.71351254e-01 5.99935591e-01 6.18782580e-01
-1.80325303e-02 9.63575840e-01 -4.84845750e-02 -1.56964839e-01
-9.81773973e-01 -1.26365232e+00 -2.04676930e-02 6.87778652e-01
9.03937161e-01 -5.27153194e-01 -2.60028005e-01 -4.07946378e-01
-4.32540655e-01 3.41321826e-01 -1.84563939e-02 1.49583206e-01
-8.41912031e-01 -2.28652447e-01 2.46107444e-01 3.96971583e-01
7.00506508e-01 -5.21827698e-01 -8.18822145e-01 5.52997403e-02
-5.89471698e-01 -1.67563570e+00 -8.37397993e-01 -4.86826867e-01
-9.17451024e-01 -1.06202078e+00 -9.24729407e-01 -1.21395573e-01
3.92624378e-01 1.31534040e+00 8.85694563e-01 2.48735160e-01
1.80665612e-01 6.85112834e-01 -6.27154171e-01 3.16537648e-01
-2.49910638e-01 -5.25996745e-01 1.80619985e-01 4.91725534e-01
-9.57587808e-02 -6.54762208e-01 -7.67073810e-01 6.46049976e-01
-1.23374832e+00 2.96571910e-01 6.06522441e-01 6.50742531e-01
7.24892557e-01 2.90499143e-02 3.74588311e-01 -3.58287305e-01
1.09055445e-01 -2.61564612e-01 -6.28893077e-01 2.33441621e-01
-3.47223014e-01 -3.52848619e-01 8.12613308e-01 -5.29023826e-01
-1.10118139e+00 -5.39438874e-02 -1.65576354e-01 -8.45901310e-01
6.89496994e-02 -8.76421034e-02 -1.91467643e-01 -1.19555064e-01
1.03410080e-01 6.35473490e-01 -7.83272274e-03 -3.81214678e-01
3.45705777e-01 9.97860193e-01 8.10868919e-01 -1.82113320e-01
7.55565464e-01 7.67790914e-01 6.94068521e-02 -1.11229157e+00
-4.98877972e-01 -9.15238798e-01 -3.73317510e-01 -5.54606318e-01
7.11285353e-01 -1.39158630e+00 -7.35295892e-01 4.97639269e-01
-1.14381301e+00 7.22809508e-02 -2.72367388e-01 7.34796286e-01
-6.71260893e-01 8.71522307e-01 -5.79575717e-01 -7.34827340e-01
-2.52821356e-01 -1.48942459e+00 1.35738218e+00 4.28522676e-01
2.60805309e-01 -4.53874797e-01 -2.43462875e-01 4.94486302e-01
3.70254099e-01 -2.78041869e-01 1.79920942e-02 -7.12466687e-02
-1.00194526e+00 2.52398103e-01 -4.68353480e-01 4.89752442e-01
1.88479096e-01 -7.30936453e-02 -9.05392706e-01 -5.41314781e-01
3.52266788e-01 3.08184415e-01 6.99064255e-01 4.22294557e-01
9.55730081e-01 -1.42577186e-01 -1.49615794e-01 8.19144011e-01
1.74477708e+00 4.95726556e-01 9.75655615e-01 3.03701609e-01
7.16916561e-01 2.56980807e-01 1.01314557e+00 5.28588712e-01
2.54450291e-01 1.32703042e+00 3.48092109e-01 3.32831405e-02
-3.53988916e-01 8.64089057e-02 9.94210064e-01 1.03292012e+00
-4.04555887e-01 -3.68928939e-01 -2.32969895e-01 2.36251369e-01
-1.78909302e+00 -1.18950212e+00 -2.97732502e-01 2.54568362e+00
4.37364012e-01 -3.14585447e-01 -4.68959250e-02 2.04655543e-01
8.62893283e-01 4.46485877e-01 -4.43716645e-01 -1.34373624e-02
-4.29854631e-01 -1.25104010e-01 6.46414816e-01 2.85056025e-01
-6.18000865e-01 2.97527939e-01 6.17281103e+00 1.03798413e+00
-1.26854646e+00 2.40156949e-01 3.14114749e-01 -3.12080532e-01
-3.23667288e-01 2.46558301e-02 -9.59675372e-01 8.54371667e-01
9.09216344e-01 -2.35869333e-01 6.69507921e-01 3.67489368e-01
5.64814031e-01 -4.40702587e-01 -8.79928350e-01 1.47153175e+00
2.87920058e-01 -1.36500144e+00 2.62415316e-02 1.18072391e-01
7.07286179e-01 -2.00759456e-01 1.17591240e-01 -2.22229809e-01
-1.69598684e-01 -6.56715989e-01 8.82382691e-01 6.51277065e-01
1.21218061e+00 -4.66652811e-01 7.75693119e-01 1.82800934e-01
-1.60355961e+00 -3.58972192e-01 -2.13053748e-01 2.83171326e-01
6.58858657e-01 8.13573241e-01 1.84548050e-02 1.24819958e+00
1.00836122e+00 1.19866276e+00 -5.96982479e-01 6.23964608e-01
-1.37618296e-02 9.77548063e-02 -1.24846026e-01 6.81268156e-01
-2.90480375e-01 -4.76579964e-01 7.56198823e-01 8.18804860e-01
9.43817258e-01 3.96272749e-01 1.19405091e-01 3.62650663e-01
1.71351507e-01 -1.96053550e-01 -7.02398717e-01 4.43006366e-01
5.72189391e-01 1.17401206e+00 -3.68908226e-01 -4.89723235e-01
-9.39718843e-01 1.03468430e+00 -4.25748944e-01 3.12965363e-01
-8.26650381e-01 -1.36463478e-01 6.18122280e-01 1.60574943e-01
4.90470201e-01 -3.17157924e-01 1.43115118e-01 -1.53568292e+00
2.63276994e-01 -9.53429759e-01 2.22496808e-01 -1.24094248e+00
-9.54406440e-01 6.84609473e-01 1.62268579e-02 -1.95573163e+00
-1.17822252e-01 -2.11302146e-01 -2.00059488e-01 6.97854578e-01
-1.72525704e+00 -9.38435495e-01 -5.54591656e-01 9.30666029e-01
8.51330757e-01 3.43587436e-02 7.36076310e-02 5.59061944e-01
-3.83576691e-01 3.84972394e-01 3.82978469e-01 -3.28004479e-01
9.05742109e-01 -7.13584244e-01 -2.61628255e-02 1.26266170e+00
1.19799897e-01 4.81505364e-01 5.43204606e-01 -4.80167031e-01
-1.77211940e+00 -9.00862813e-01 3.99309993e-01 -2.05105189e-02
2.73478091e-01 1.09305590e-01 -7.64088988e-01 4.36280370e-01
1.33115664e-01 3.10351431e-01 3.95830721e-01 -6.26117706e-01
-2.21955702e-01 -5.69309890e-01 -1.07066512e+00 5.80323696e-01
1.26509917e+00 -5.32698393e-01 -3.53961200e-01 -1.22182056e-01
1.02093029e+00 -5.51724076e-01 -9.80166674e-01 4.87422764e-01
6.24637783e-01 -1.60194683e+00 1.25843358e+00 4.25449371e-01
3.70317936e-01 -8.38426292e-01 -7.32766092e-01 -9.58907604e-01
5.30347116e-02 -7.33594477e-01 -4.66980964e-01 1.16771042e+00
-2.24226758e-01 -7.46905863e-01 2.76031017e-01 -6.76043928e-02
-1.45972539e-02 -5.81128538e-01 -9.69909012e-01 -1.07436216e+00
-8.23639333e-01 -3.73555154e-01 5.59507430e-01 4.75015849e-01
-3.78721088e-01 9.72309709e-02 -8.30420732e-01 3.45470905e-01
6.67611480e-01 2.50275970e-01 8.55932772e-01 -9.54334676e-01
-5.65001428e-01 -2.03779563e-02 -4.25101817e-01 -1.56644642e+00
-2.74473131e-01 -2.63018072e-01 -1.12545110e-01 -1.27190197e+00
4.31450278e-01 -1.71415046e-01 -1.56309649e-01 -9.27354544e-02
-1.54676765e-01 4.11060512e-01 7.16627121e-01 5.79619408e-01
-8.06861043e-01 2.40237281e-01 1.24542677e+00 1.28059104e-01
-2.32707798e-01 -1.02041438e-01 -5.19330144e-01 5.70499837e-01
2.87697315e-01 -4.74487394e-02 -4.26721841e-01 -4.87424821e-01
1.01720929e-01 5.97710431e-01 2.57542819e-01 -1.15693128e+00
3.10125053e-01 -2.18621060e-01 2.41507515e-01 -7.82504916e-01
6.98371530e-01 -1.08767867e+00 7.18130589e-01 1.65711015e-01
-8.55437517e-02 3.23061079e-01 -1.88287154e-01 8.38135421e-01
-5.24906635e-01 -2.85172611e-02 9.03330207e-01 -2.09075101e-02
-8.77256930e-01 2.58277804e-01 -2.11081594e-01 -2.59160578e-01
1.14202178e+00 -6.13817573e-01 -5.31896651e-01 -6.06802046e-01
-3.95650655e-01 -3.32527701e-03 7.80029297e-01 3.45804930e-01
1.16638529e+00 -1.31650758e+00 -5.13610423e-01 3.71779799e-01
-9.24145952e-02 -9.62072089e-02 9.69605327e-01 1.17684066e+00
-2.45907620e-01 2.53931552e-01 -1.46906167e-01 -9.54972565e-01
-1.74266863e+00 7.75700748e-01 -8.13506469e-02 -4.31405872e-01
-7.85920620e-01 2.84300476e-01 3.04575354e-01 2.86474109e-01
-1.42969042e-01 -2.69466806e-02 -3.20665032e-01 -1.19186983e-01
1.06938958e+00 6.56276584e-01 -9.36863050e-02 -1.05691469e+00
-1.18177459e-01 1.10831976e+00 4.65458184e-02 -2.05584794e-01
1.22399974e+00 -9.81398106e-01 6.85257688e-02 3.75390202e-01
1.21710443e+00 4.85584766e-01 -1.21972060e+00 -3.67116570e-01
-4.99684036e-01 -1.21227109e+00 6.70737494e-03 -1.67362943e-01
-1.39296567e+00 5.94183147e-01 6.64452910e-01 1.54123697e-02
1.82954180e+00 -1.53133973e-01 1.05696690e+00 -1.74311206e-01
7.45557964e-01 -6.78154171e-01 2.69199491e-01 1.46516860e-01
6.48675680e-01 -9.86346543e-01 2.92159617e-01 -6.49334311e-01
-5.84628403e-01 1.25866389e+00 2.76947170e-01 1.42630696e-01
1.55392766e-01 3.06503862e-01 -8.43648836e-02 2.21138716e-01
-7.92263985e-01 -1.40211985e-01 1.71531379e-01 5.95988870e-01
1.03775404e-01 -2.33426183e-01 -2.85645247e-01 2.58745402e-01
3.60708505e-01 3.16263527e-01 9.64723170e-01 8.02688897e-01
-2.41312355e-01 -1.17736804e+00 -7.85723150e-01 3.68310332e-01
-5.16655922e-01 -6.93376213e-02 3.74670446e-01 3.20000470e-01
4.46462743e-02 1.32674170e+00 -1.13762431e-01 -6.52248204e-01
3.61224174e-01 -7.02234685e-01 5.13210118e-01 -2.08880320e-01
-1.94995567e-01 2.52084523e-01 -1.43371075e-01 -1.06767285e+00
-8.06986749e-01 -6.58485353e-01 -8.59499872e-01 -3.75991195e-01
-3.20933253e-01 -2.41620228e-01 3.71051103e-01 7.69972742e-01
5.19885719e-01 4.35306817e-01 1.10855722e+00 -1.13831198e+00
-4.22080100e-01 -4.52940971e-01 -7.61734486e-01 3.46072018e-01
5.86402893e-01 -5.02104878e-01 -5.88807166e-01 3.78953576e-01] | [10.992293357849121, -2.0365986824035645] |
5fc618f6-f069-40a0-a848-5ea01ebced70 | differentiable-multi-granularity-human | 2103.04570 | null | https://arxiv.org/abs/2103.04570v1 | https://arxiv.org/pdf/2103.04570v1.pdf | Differentiable Multi-Granularity Human Representation Learning for Instance-Aware Human Semantic Parsing | To address the challenging task of instance-aware human part parsing, a new bottom-up regime is proposed to learn category-level human semantic segmentation as well as multi-person pose estimation in a joint and end-to-end manner. It is a compact, efficient and powerful framework that exploits structural information over different human granularities and eases the difficulty of person partitioning. Specifically, a dense-to-sparse projection field, which allows explicitly associating dense human semantics with sparse keypoints, is learnt and progressively improved over the network feature pyramid for robustness. Then, the difficult pixel grouping problem is cast as an easier, multi-person joint assembling task. By formulating joint association as maximum-weight bipartite matching, a differentiable solution is developed to exploit projected gradient descent and Dykstra's cyclic projection algorithm. This makes our method end-to-end trainable and allows back-propagating the grouping error to directly supervise multi-granularity human representation learning. This is distinguished from current bottom-up human parsers or pose estimators which require sophisticated post-processing or heuristic greedy algorithms. Experiments on three instance-aware human parsing datasets show that our model outperforms other bottom-up alternatives with much more efficient inference. | ['Luc van Gool', 'Yi Yang', 'Si Liu', 'Wenguan Wang', 'Tianfei Zhou'] | 2021-03-08 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Zhou_Differentiable_Multi-Granularity_Human_Representation_Learning_for_Instance-Aware_Human_Semantic_Parsing_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Zhou_Differentiable_Multi-Granularity_Human_Representation_Learning_for_Instance-Aware_Human_Semantic_Parsing_CVPR_2021_paper.pdf | cvpr-2021-1 | ['human-parsing'] | ['computer-vision'] | [ 3.65879059e-01 5.38097262e-01 -8.81048851e-03 -6.51976228e-01
-1.03271246e+00 -3.32300425e-01 3.53065878e-01 -7.78933018e-02
-5.24732649e-01 4.00829196e-01 3.69529963e-01 4.11194801e-01
-1.63558483e-01 -5.50059140e-01 -8.67834568e-01 -4.54091758e-01
6.86229169e-02 1.06218576e+00 2.82746583e-01 5.80276363e-02
-1.92134246e-01 1.11153804e-01 -1.60629177e+00 2.44333237e-01
6.91536844e-01 8.34507108e-01 3.59741151e-01 6.73320770e-01
1.33255228e-01 4.68541235e-01 -1.58329144e-01 -8.38429511e-01
3.92180949e-01 -2.10695967e-01 -8.76156688e-01 6.02624536e-01
9.61335957e-01 -1.34169176e-01 -9.79598090e-02 9.83072102e-01
6.09872460e-01 2.30100527e-01 4.01060998e-01 -8.76568139e-01
-3.13776374e-01 5.71904063e-01 -9.78138268e-01 -1.19635627e-01
5.27189493e-01 1.29646748e-01 1.25186300e+00 -8.31124425e-01
5.91630220e-01 1.67678142e+00 1.02136099e+00 4.88672167e-01
-1.33715916e+00 -3.19297075e-01 5.78157961e-01 -4.76641208e-02
-1.09071887e+00 -4.27601710e-02 8.02030146e-01 -4.26201284e-01
7.04683602e-01 3.32935035e-01 9.01799440e-01 1.03762305e+00
-3.14660043e-01 1.31966603e+00 1.09390295e+00 -3.14070672e-01
-2.26862170e-02 -9.89829674e-02 3.62786502e-01 1.26932526e+00
3.40711594e-01 -1.25532255e-01 -6.88173771e-01 -1.96173951e-01
8.24806571e-01 1.12456724e-01 -1.28856478e-02 -8.61688972e-01
-1.19906688e+00 8.18915725e-01 7.54066408e-01 -7.31121525e-02
-5.27428269e-01 1.97627217e-01 3.57740641e-01 -3.60013276e-01
2.07119539e-01 2.56633669e-01 -6.49875820e-01 2.64713913e-01
-1.18293214e+00 6.01478577e-01 7.21497834e-01 1.01448774e+00
9.54955459e-01 -4.83916163e-01 -3.48400861e-01 9.92932498e-01
4.23206210e-01 3.34108979e-01 1.97648659e-01 -1.18748784e+00
7.07206666e-01 7.57830441e-01 -1.30435303e-01 -9.77096736e-01
-8.06371033e-01 -5.23433685e-01 -7.74321139e-01 -2.35131815e-01
7.47981548e-01 -1.47534639e-03 -1.17809713e+00 1.79068804e+00
7.79125333e-01 6.85140565e-02 -4.06778574e-01 1.11835551e+00
4.91431832e-01 2.38073781e-01 5.14920235e-01 1.78184956e-01
2.15705419e+00 -1.30576670e+00 -3.41602296e-01 -6.60003483e-01
2.32559130e-01 -1.52424291e-01 9.09958065e-01 2.19858572e-01
-1.21157348e+00 -7.81772673e-01 -6.98969364e-01 -4.59168822e-01
-2.23613799e-01 2.33391121e-01 1.12169051e+00 6.36007965e-01
-6.98238075e-01 6.41682208e-01 -1.05040383e+00 -4.76437837e-01
7.63797760e-01 3.46597493e-01 -5.17251313e-01 -3.64437103e-01
-8.75786364e-01 5.10832310e-01 6.67912364e-01 2.34046727e-01
-4.05088603e-01 -8.25005710e-01 -1.22588181e+00 -3.22193503e-02
7.35583127e-01 -1.49507034e+00 1.04183030e+00 -7.58073866e-01
-1.42604506e+00 1.09711254e+00 -9.16400105e-02 -4.67696190e-01
7.04519808e-01 -5.24548948e-01 2.15786085e-01 4.50801283e-01
5.67328274e-01 1.02945936e+00 8.65332961e-01 -1.04960525e+00
-7.92741597e-01 -8.84849072e-01 -2.32923567e-01 4.93166536e-01
-6.44852296e-02 7.30766309e-03 -1.22345090e+00 -6.55057430e-01
5.55927396e-01 -1.14435935e+00 -5.82152843e-01 -8.05973727e-03
-6.91028237e-01 -2.69134492e-01 2.49599293e-01 -9.33919609e-01
8.98466170e-01 -1.84909379e+00 6.38423681e-01 4.26579326e-01
3.52517277e-01 -2.27541730e-01 -1.18290875e-02 8.66945460e-02
1.97307780e-01 -4.24321949e-01 -7.35755146e-01 -7.91689098e-01
3.45108747e-01 2.45679915e-01 3.40833753e-01 5.39178789e-01
1.49627537e-01 1.19283855e+00 -8.57387483e-01 -7.62589097e-01
2.20220849e-01 4.06928450e-01 -9.21739459e-01 2.47101158e-01
2.28095446e-02 4.05477494e-01 -5.67914665e-01 9.62058425e-01
4.89704967e-01 -4.23164099e-01 2.41036162e-01 -3.66710246e-01
2.97175020e-01 -7.51372799e-02 -1.38018596e+00 2.29301882e+00
-1.46392822e-01 -1.18730523e-01 3.60062867e-01 -8.80842984e-01
3.94685179e-01 -4.83720340e-02 4.76741910e-01 -4.24523115e-01
1.65090710e-02 -6.43258449e-03 -4.83852953e-01 -1.80583447e-01
3.36855054e-01 -3.67073677e-02 -6.49942756e-01 2.02250481e-02
3.43168020e-01 1.10557556e-01 3.56108457e-01 3.71785820e-01
9.80793476e-01 6.20705009e-01 3.33330393e-01 -4.06391144e-01
3.97065818e-01 -1.99053437e-02 8.35777402e-01 9.02946711e-01
-2.60133386e-01 8.37355852e-01 2.48692513e-01 -3.99009407e-01
-7.05433011e-01 -1.12828434e+00 -3.01367119e-02 1.56529844e+00
1.49497509e-01 -4.98546600e-01 -1.03474069e+00 -9.57714081e-01
2.29615018e-01 1.72424406e-01 -7.09130168e-01 2.09828719e-01
-8.85460973e-01 -8.53259385e-01 2.69202679e-01 8.01235557e-01
5.89630246e-01 -1.05973840e+00 -7.38475561e-01 3.23053062e-01
-3.57703865e-01 -1.29261136e+00 -5.51551163e-01 3.73947203e-01
-7.18667924e-01 -1.06217456e+00 -1.14812839e+00 -8.78858984e-01
8.29257905e-01 -1.09042041e-01 1.28193867e+00 -1.67006999e-01
-6.54817402e-01 7.34809518e-01 -1.29121706e-01 5.09845354e-02
2.83935785e-01 7.99567103e-02 -8.35953206e-02 -4.59355228e-02
3.62975806e-01 -4.97662097e-01 -7.28310883e-01 8.62600878e-02
-2.43924990e-01 2.04006284e-01 9.94122148e-01 9.40271437e-01
9.25906241e-01 -9.45435241e-02 -5.70164695e-02 -1.26343858e+00
-4.82637295e-03 -1.24738418e-01 -4.44617420e-01 2.96507925e-01
-1.61406532e-01 2.40224987e-01 2.80301005e-01 -2.29073375e-01
-1.09124899e+00 8.11147511e-01 -7.89182112e-02 -3.64985645e-01
-3.92813087e-01 1.00904122e-01 -5.71421206e-01 1.63054958e-01
3.06304246e-01 1.77747473e-01 -2.70210832e-01 -7.78033793e-01
7.02553391e-01 1.47869542e-01 1.03503978e+00 -9.11357999e-01
9.48031068e-01 5.38211942e-01 -2.14925468e-01 -5.94895124e-01
-1.21887112e+00 -9.31517661e-01 -1.08418894e+00 2.11634673e-02
1.42339838e+00 -1.26015031e+00 -5.44931054e-01 4.54360902e-01
-9.89277005e-01 -2.11051390e-01 -3.33303303e-01 2.00801626e-01
-7.07050443e-01 6.37227416e-01 -9.02455151e-01 -6.60332441e-01
-2.45168895e-01 -8.83308768e-01 1.71348536e+00 1.37303695e-01
-3.30640435e-01 -6.51884377e-01 -2.49589775e-02 9.93818879e-01
-2.83017427e-01 3.03306162e-01 4.08044070e-01 -3.88736099e-01
-6.20415747e-01 -3.08097094e-01 -3.37414742e-01 -5.85221611e-02
-3.87623489e-01 -7.65805364e-01 -9.26943600e-01 -3.30552399e-01
-3.20236206e-01 -3.51923138e-01 1.17733347e+00 3.81431401e-01
1.16750193e+00 -2.88082987e-01 -6.09987736e-01 8.33564878e-01
1.17922425e+00 -7.36070096e-01 2.00049043e-01 2.38982797e-01
1.31630611e+00 8.90657485e-01 3.97405267e-01 3.69619370e-01
8.28554809e-01 6.84927523e-01 7.64728636e-02 -2.11574614e-01
-2.24788263e-01 -6.48510218e-01 4.31573838e-02 3.90794903e-01
-8.42106268e-02 2.72636414e-01 -6.77683294e-01 4.03948575e-01
-2.06130528e+00 -9.54169154e-01 6.18516915e-02 1.81408668e+00
7.12945163e-01 1.46035060e-01 6.22834861e-01 -7.90456310e-02
8.33138585e-01 8.91899019e-02 -4.77568239e-01 2.78612554e-01
8.68924782e-02 1.21559240e-01 5.76785147e-01 3.17084402e-01
-1.56792581e+00 1.19950485e+00 5.57598066e+00 6.88878238e-01
-2.71312535e-01 1.71776712e-01 5.14119983e-01 -3.48021314e-02
5.54837994e-02 -2.31249362e-01 -1.04403937e+00 3.81057680e-01
3.92171204e-01 6.77250266e-01 2.79962182e-01 9.97958362e-01
-2.81020880e-01 -1.37556329e-01 -1.16278756e+00 1.20920694e+00
2.85381041e-02 -8.93856704e-01 -1.00817598e-01 9.68641788e-02
4.99163955e-01 -2.59673506e-01 -2.25010753e-01 5.92477441e-01
4.52121347e-01 -5.79783022e-01 9.38860595e-01 3.90810370e-01
3.51677924e-01 -8.29469800e-01 3.81155789e-01 5.14754951e-01
-1.59138262e+00 -3.96968573e-01 -4.14105743e-01 1.04173042e-01
5.14806211e-01 6.47304296e-01 -3.16226721e-01 5.46012759e-01
1.02459276e+00 5.59848607e-01 -7.01623738e-01 7.08788991e-01
-3.66736323e-01 1.71543732e-01 -6.50884986e-01 3.73432338e-01
2.17634872e-01 -8.16270337e-02 4.62007642e-01 1.56492567e+00
-4.12348621e-02 3.50857913e-01 6.97265983e-01 7.35175252e-01
-8.01125262e-03 8.44267663e-03 -3.54850963e-02 3.30927819e-01
1.75695390e-01 1.56796539e+00 -1.08353317e+00 -4.45032179e-01
-3.82988989e-01 1.52260709e+00 7.82194197e-01 2.04461098e-01
-7.60494351e-01 3.43937576e-02 4.10453051e-01 2.75160074e-01
7.68948078e-01 -3.56236584e-02 -2.82199919e-01 -1.19279683e+00
1.46451578e-01 -7.23845422e-01 9.13242519e-01 -3.73603106e-01
-1.50403178e+00 2.15598568e-01 9.28726941e-02 -5.95939159e-01
-3.30608487e-01 -6.55975819e-01 -2.82464862e-01 5.96737504e-01
-1.31987906e+00 -1.74666584e+00 -1.63091809e-01 8.03226650e-01
5.45462310e-01 1.28278881e-01 6.52424216e-01 3.00585359e-01
-7.92116940e-01 8.06479335e-01 -6.95148051e-01 3.44181061e-01
3.00852120e-01 -1.64117956e+00 2.71817654e-01 1.05925608e+00
2.47296914e-01 5.87311864e-01 6.05836749e-01 -8.33398640e-01
-1.30822659e+00 -1.04382074e+00 8.23218048e-01 -6.28687143e-01
4.51636523e-01 -7.38321185e-01 -6.89106584e-01 7.07524657e-01
-1.14667691e-01 2.49576777e-01 5.77775657e-01 5.86912692e-01
-5.02843320e-01 8.92241970e-02 -1.09024560e+00 3.77307713e-01
1.50890470e+00 -4.14019197e-01 -7.27287114e-01 4.25747693e-01
5.21374583e-01 -5.71903229e-01 -6.98847175e-01 2.63069004e-01
4.18023765e-01 -9.90370333e-01 1.30902123e+00 -5.26301503e-01
1.49235085e-01 -2.79489279e-01 -2.21805871e-01 -7.83045113e-01
-7.71874249e-01 -6.18417740e-01 -4.57049370e-01 1.12474370e+00
3.42775285e-01 -9.10512134e-02 1.10461593e+00 8.74689281e-01
-8.60644132e-02 -7.41855562e-01 -8.42146516e-01 -5.23515046e-01
-2.99205095e-01 -3.63640517e-01 1.87674403e-01 6.07548356e-01
-9.08286050e-02 4.22772527e-01 -5.28586090e-01 5.36093950e-01
1.30561209e+00 1.50300905e-01 8.33458304e-01 -1.26755667e+00
-8.44499648e-01 -2.38657802e-01 -5.34783721e-01 -1.41637683e+00
2.16292322e-01 -9.28989589e-01 3.06047142e-01 -1.57363725e+00
5.91797411e-01 -1.59544304e-01 -2.02025488e-01 5.51172078e-01
-6.95462644e-01 4.13148105e-01 3.26600343e-01 1.51891381e-01
-1.13187039e+00 3.08419883e-01 1.10987258e+00 -8.06500688e-02
-1.21639788e-01 1.32165909e-01 -9.00542855e-01 1.03677785e+00
2.41168797e-01 -4.90993649e-01 -1.49754733e-01 -3.34504068e-01
-1.60533004e-02 3.88640948e-02 7.30326474e-01 -1.19053924e+00
3.72441620e-01 3.29972863e-01 7.71242201e-01 -7.03105688e-01
5.74638546e-01 -6.94918871e-01 -2.03839585e-01 3.41038138e-01
-2.67250597e-01 -1.71282858e-01 -2.43944779e-01 9.39277589e-01
6.63398206e-02 2.10048005e-01 8.52111459e-01 -6.49558544e-01
-9.44604754e-01 5.49759746e-01 1.72539458e-01 2.91989863e-01
8.41789842e-01 -4.76401150e-01 6.11231267e-01 2.51132637e-01
-1.13838530e+00 4.24212068e-01 1.62468657e-01 2.34796539e-01
3.01540881e-01 -1.14956105e+00 -6.96526170e-01 1.56748310e-01
6.00907020e-02 3.23706239e-01 5.78405797e-01 8.62454593e-01
-5.07902652e-02 1.48372784e-01 3.96839380e-02 -8.96478117e-01
-1.21502936e+00 5.29310703e-01 2.49339834e-01 -6.31767154e-01
-1.18626058e+00 1.37718296e+00 4.28212762e-01 -6.80598915e-01
3.53796184e-01 5.02959900e-02 -9.50290710e-02 1.96220726e-01
4.63034481e-01 3.55561048e-01 -2.03883976e-01 -7.40830541e-01
-4.95469332e-01 8.24476898e-01 -1.04700044e-01 -1.12034604e-01
1.39392340e+00 -2.86386460e-01 6.92295432e-02 6.08544126e-02
1.13859081e+00 -1.44077703e-01 -1.75552583e+00 -3.07613134e-01
1.66742504e-01 -2.72574157e-01 -2.21379802e-01 -7.78430462e-01
-1.05418241e+00 6.33359373e-01 4.14488375e-01 -1.92137405e-01
8.22549760e-01 4.58422273e-01 9.90635157e-01 4.05346900e-01
4.77171361e-01 -1.47932410e+00 2.00790558e-02 2.69907743e-01
5.91902196e-01 -1.30070913e+00 1.33070424e-01 -7.57202506e-01
-7.23833859e-01 6.75311685e-01 7.83403635e-01 -2.09666267e-01
4.21080530e-01 2.15025134e-02 -1.55883729e-01 -3.93718958e-01
-1.83727115e-01 -5.07900119e-01 5.67177534e-01 6.27050221e-01
1.20730430e-01 2.05960065e-01 -3.85080576e-02 9.67023790e-01
-4.63887990e-01 -2.93249577e-01 -2.81139821e-01 8.13166916e-01
-5.12404621e-01 -7.86157191e-01 -5.42731643e-01 4.08277124e-01
-3.44843209e-01 8.42933208e-02 -1.80728853e-01 9.05107558e-01
3.51189613e-01 4.90085959e-01 1.34999249e-02 1.15898423e-01
4.82520998e-01 1.71381578e-01 6.67493284e-01 -7.57555842e-01
-5.60692132e-01 2.69405961e-01 1.00886792e-01 -1.07016265e+00
-5.13868928e-01 -1.01052880e+00 -1.33333635e+00 1.27541184e-01
-1.27947226e-01 -6.93258271e-02 2.78600574e-01 1.11526883e+00
1.79332614e-01 5.12538671e-01 1.02942497e-01 -1.40589237e+00
-5.13596117e-01 -6.68002546e-01 -6.10670865e-01 7.48663962e-01
-3.66637334e-02 -7.76253581e-01 -1.03505656e-01 -2.46534999e-02] | [8.163106918334961, -0.238815039396286] |
53dab514-34d2-465a-9dcf-545292e59d1d | a-multi-task-learning-framework-for-sound | 2305.10729 | null | https://arxiv.org/abs/2305.10729v1 | https://arxiv.org/pdf/2305.10729v1.pdf | A Multi-Task Learning Framework for Sound Event Detection using High-level Acoustic Characteristics of Sounds | Sound event detection (SED) entails identifying the type of sound and estimating its temporal boundaries from acoustic signals. These events are uniquely characterized by their spatio-temporal features, which are determined by the way they are produced. In this study, we leverage some distinctive high-level acoustic characteristics of various sound events to assist the SED model training, without requiring additional labeled data. Specifically, we use the DCASE Task 4 2022 dataset and categorize the 10 classes into four subcategories based on their high-level acoustic characteristics. We then introduce a novel multi-task learning framework that jointly trains the SED and high-level acoustic characteristics classification tasks, using shared layers and weighted loss. Our method significantly improves the performance of the SED system, achieving a 36.3% improvement in terms of the polyphonic sound event detection score compared to the baseline on the DCASE 2022 Task 4 validation set. | ['Rohan Kumar Das', 'Tanmay Khandelwal'] | 2023-05-18 | null | null | null | null | ['sound-event-detection'] | ['audio'] | [-3.39633077e-02 -6.78618670e-01 3.90964746e-01 -2.17807963e-01
-1.59590614e+00 -7.40562320e-01 3.98451746e-01 2.32531682e-01
-6.67202771e-01 1.57776222e-01 3.25246900e-01 -8.73925444e-03
2.11572219e-02 -4.41561878e-01 -5.78890264e-01 -6.84900463e-01
-4.07194346e-01 -9.29965898e-02 4.65760291e-01 3.72711480e-01
-1.28690168e-01 1.22337826e-01 -1.72513664e+00 6.67507112e-01
3.04474294e-01 1.44272423e+00 4.82336730e-01 1.03727007e+00
1.46762341e-01 5.96676111e-01 -6.77727461e-01 8.36623162e-02
-2.29591876e-01 -5.29953659e-01 -4.39936101e-01 -5.73539615e-01
4.39699411e-01 -4.30902839e-02 -2.22444981e-01 5.72398543e-01
8.39585960e-01 3.40855002e-01 6.01924241e-01 -1.14811110e+00
7.17006996e-02 5.74836493e-01 -1.43036500e-01 5.71219623e-01
2.80750215e-01 -1.85293004e-01 1.40184164e+00 -1.25507498e+00
-5.71224317e-02 1.00861156e+00 8.61801147e-01 3.23727518e-01
-1.07140863e+00 -1.09795702e+00 5.73075637e-02 5.57800651e-01
-1.33548999e+00 -7.33556867e-01 8.52192760e-01 -6.09852552e-01
1.03642738e+00 1.62220046e-01 1.38830736e-01 1.24807584e+00
-8.69158432e-02 7.84909666e-01 8.60009313e-01 -3.92273158e-01
3.10363770e-01 -3.68431568e-01 2.25343317e-01 2.29962468e-01
-5.15050471e-01 1.88239679e-01 -9.78545964e-01 -1.43478125e-01
2.00937048e-01 -4.06358063e-01 -6.32183328e-02 4.07574356e-01
-9.56223369e-01 4.26031530e-01 1.57672629e-01 2.71472812e-01
-2.95007944e-01 4.38939542e-01 6.85344398e-01 1.04815744e-01
5.83037734e-01 1.96662605e-01 -6.58339798e-01 -4.47485566e-01
-8.86733055e-01 2.04204053e-01 4.97688532e-01 3.55086267e-01
4.67118412e-01 1.62218526e-01 -3.22141975e-01 1.38429534e+00
3.54263812e-01 6.16822779e-01 4.24237907e-01 -6.56737506e-01
4.51168597e-01 -2.13404343e-01 -5.09766676e-02 -5.88601172e-01
-5.94915271e-01 -5.49738348e-01 -3.63458127e-01 -1.35914043e-01
2.60714442e-01 -3.89381886e-01 -8.05937529e-01 2.01557970e+00
1.79224744e-01 8.45228493e-01 -2.10851640e-01 5.61804771e-01
9.63123798e-01 1.02322221e+00 3.64298105e-01 3.19440551e-02
1.43857861e+00 -6.97268009e-01 -7.62782156e-01 -1.02396421e-01
3.36690336e-01 -8.25659752e-01 1.05687833e+00 4.23547387e-01
-8.67974639e-01 -8.87968361e-01 -8.09522450e-01 3.56388927e-01
-2.60395795e-01 3.93059701e-01 1.98468789e-01 5.28084040e-01
-6.28770769e-01 3.05267304e-01 -9.47498322e-01 2.74061412e-01
2.39652142e-01 -7.39070922e-02 5.69782294e-02 4.46082890e-01
-1.51051486e+00 2.22677350e-01 3.63279134e-01 -5.89991286e-02
-1.39927399e+00 -1.19745600e+00 -9.26503539e-01 2.29926661e-01
8.87033418e-02 -3.99632081e-02 1.77051544e+00 -9.48572308e-02
-1.26172221e+00 6.61226571e-01 -1.84691027e-01 -5.56104124e-01
1.89012513e-01 -5.21194279e-01 -8.28344703e-01 2.31660247e-01
1.63377568e-01 2.61643648e-01 8.75341594e-01 -1.03552496e+00
-1.28608203e+00 1.34236395e-01 -2.96102196e-01 -1.66831762e-01
-4.40907717e-01 4.43776250e-01 -4.04853821e-01 -9.53017235e-01
-1.72166124e-01 -8.60044777e-01 1.19629741e-01 -2.74303287e-01
-3.54166776e-01 -5.66055119e-01 8.35878849e-01 -6.88326955e-01
1.53010309e+00 -2.68850803e+00 -4.14457530e-01 -1.65957376e-01
1.31566934e-02 2.04164281e-01 -2.52854407e-01 3.01544815e-01
-1.24750219e-01 -8.60394984e-02 -1.69231370e-01 -7.07261622e-01
2.92727292e-01 -3.46658856e-01 -6.45086944e-01 6.40664548e-02
2.95591652e-01 3.64765346e-01 -1.02054954e+00 -2.36115977e-01
2.13897452e-01 4.28255618e-01 -4.03990179e-01 4.24898565e-01
1.39999799e-02 2.51042038e-01 -3.34537983e-01 4.01240140e-01
4.28250104e-01 1.75928712e-01 -2.52623886e-01 -3.25411558e-01
-1.53523207e-01 8.39362204e-01 -1.06831706e+00 1.52247810e+00
-1.03748941e+00 7.93124437e-01 1.16171591e-01 -6.85985684e-01
7.67122567e-01 8.30067515e-01 6.77685797e-01 -6.66058600e-01
-3.24395269e-01 4.18135583e-01 6.95405106e-05 -4.60931271e-01
1.62320927e-01 -2.79234052e-01 -4.60870802e-01 5.61488390e-01
3.92194204e-02 -1.50132209e-01 3.60659920e-02 -8.61747563e-02
1.35129607e+00 -9.34624225e-02 -1.35694325e-01 3.36399116e-02
4.27227795e-01 -5.90430737e-01 6.88942432e-01 7.27679670e-01
-2.67094493e-01 8.44257057e-01 2.25211427e-01 -5.98380491e-02
-5.80804527e-01 -1.54284501e+00 -4.98060316e-01 1.43844879e+00
-2.96814144e-01 -5.79331875e-01 -4.41964239e-01 -6.62806869e-01
4.39878069e-02 7.93570518e-01 -4.29697543e-01 -2.70432919e-01
-8.10133278e-01 -6.87592506e-01 1.11102641e+00 6.78062320e-01
2.16740832e-01 -1.31564248e+00 -5.58712304e-01 5.74567318e-01
-4.05770987e-01 -1.50185740e+00 -6.64753973e-01 5.94378412e-01
-1.15744926e-01 -8.48302245e-01 -4.32779729e-01 -9.79343116e-01
-1.31468266e-01 7.24792704e-02 1.03171897e+00 -5.90242207e-01
-3.55471015e-01 3.98178697e-01 -6.28799200e-01 -7.14310110e-01
-3.67589086e-01 -3.22895832e-02 1.42918199e-01 3.49876732e-01
1.98980168e-01 -7.35482752e-01 -6.28503799e-01 3.31239969e-01
-5.96238136e-01 -4.28767711e-01 2.76886612e-01 5.70010364e-01
5.37942111e-01 2.60023832e-01 1.29042435e+00 -2.64109164e-01
4.28801656e-01 -5.59372127e-01 -2.16112360e-01 -1.30803466e-01
-6.20454177e-02 -4.48008418e-01 6.95166230e-01 -3.59676719e-01
-1.08817089e+00 9.17415321e-02 -4.63584721e-01 -3.50558400e-01
-3.08508128e-01 3.38146687e-01 -3.27198878e-02 3.17321777e-01
3.68889451e-01 2.42742553e-01 -5.72731972e-01 -9.29004788e-01
-5.21576218e-02 9.87279713e-01 7.11782277e-01 -5.67550480e-01
5.32077491e-01 1.37429267e-01 -2.85121560e-01 -8.80539000e-01
-1.22604322e+00 -7.25940108e-01 -3.95445585e-01 -3.32424968e-01
1.02221143e+00 -1.11974931e+00 -5.22581816e-01 8.59105289e-01
-1.20059216e+00 -3.61519754e-01 -1.33671805e-01 8.52101207e-01
-4.16956782e-01 6.84297755e-02 -7.56488204e-01 -9.98986006e-01
-2.06948549e-01 -8.67526412e-01 1.26144886e+00 -2.81211227e-01
-2.69799054e-01 -7.88154542e-01 3.60470355e-01 1.19837195e-01
2.74940401e-01 1.74162865e-01 9.18902934e-01 -7.79274046e-01
-6.21979237e-02 -1.06968723e-01 1.06869146e-01 6.89743400e-01
3.06141943e-01 -1.24696352e-01 -1.61134267e+00 -9.12663788e-02
-1.02333978e-01 -4.61820662e-01 1.23067701e+00 5.10760367e-01
1.44126952e+00 2.18462229e-01 -2.33288676e-01 4.15058672e-01
9.61356163e-01 5.43414772e-01 -1.85935851e-02 6.66130334e-03
5.86793482e-01 4.84296769e-01 5.93201518e-01 5.89230180e-01
2.98078477e-01 9.40893829e-01 2.38630027e-01 8.42523724e-02
-5.35545588e-01 -3.21763307e-01 6.87868059e-01 1.15933478e+00
4.70284760e-01 -2.66787618e-01 -9.47386682e-01 9.24542487e-01
-1.48161185e+00 -9.09015656e-01 -8.75502229e-02 2.10498548e+00
1.08435786e+00 1.99643999e-01 2.31565297e-01 4.59815085e-01
9.92516696e-01 5.00818849e-01 -2.97791123e-01 -1.48163915e-01
2.63741482e-02 4.96873707e-01 -9.66940075e-02 2.04922929e-01
-1.61520755e+00 6.22845829e-01 6.62291193e+00 1.07472849e+00
-1.13322294e+00 3.19580555e-01 3.79635632e-01 -4.77653295e-01
8.21385235e-02 -4.98491079e-01 -9.11938250e-01 7.87362278e-01
1.42653036e+00 2.16450263e-02 9.49039981e-02 3.53030264e-01
4.41608995e-01 1.99069709e-01 -1.26521254e+00 1.03815448e+00
-4.85390984e-02 -9.88508642e-01 -3.42074096e-01 -3.53670627e-01
5.14280260e-01 2.36833736e-01 2.29407281e-01 5.92935860e-01
-4.91207391e-02 -7.70804882e-01 1.08413625e+00 2.79438615e-01
8.63793492e-01 -7.31909931e-01 2.64925241e-01 4.10644524e-02
-1.71971917e+00 -1.27736047e-01 8.85705799e-02 1.08987562e-01
4.78265405e-01 7.16448963e-01 -7.88069665e-01 4.10175204e-01
1.08969474e+00 6.77569389e-01 -2.71500558e-01 1.37446344e+00
-1.02083281e-01 1.55330205e+00 -4.56097215e-01 1.79174304e-01
1.90733984e-01 3.04836810e-01 7.11233735e-01 1.64193368e+00
5.00715971e-01 -2.78930277e-01 1.38603970e-01 6.29107475e-01
-1.89309195e-01 -1.90225150e-02 -1.64725915e-01 9.32223052e-02
8.49808991e-01 1.10729289e+00 -4.67263758e-01 -1.06644712e-01
-4.90866393e-01 4.98489290e-01 -3.69385444e-02 2.91933000e-01
-1.10751271e+00 -8.36669922e-01 9.12110388e-01 -2.82422423e-01
6.64084315e-01 -9.13121998e-02 1.29157063e-02 -6.76205933e-01
9.14122686e-02 -5.38221896e-01 4.63465065e-01 -5.18276751e-01
-1.37716317e+00 6.64082170e-01 -7.29661360e-02 -1.27469921e+00
-3.01124364e-01 -4.28609341e-01 -9.31051850e-01 7.82375097e-01
-1.59403074e+00 -9.99534309e-01 -7.33829439e-02 3.90814245e-01
7.15565205e-01 -2.99329963e-02 7.34412253e-01 8.65965366e-01
-7.16783583e-01 8.59212220e-01 1.41750783e-01 3.01165640e-01
9.46328580e-01 -1.33794343e+00 6.36104822e-01 7.00325966e-01
5.08629322e-01 -2.14145258e-02 3.51999521e-01 -2.60445654e-01
-7.01152265e-01 -1.43101251e+00 1.15097344e+00 -4.38927740e-01
9.77015793e-01 -8.05474997e-01 -9.07928824e-01 3.15964609e-01
-3.14845592e-01 2.73655713e-01 8.45035255e-01 3.27940762e-01
-4.14349049e-01 -4.07980621e-01 -6.00242376e-01 1.46834418e-01
9.25076723e-01 -1.09061790e+00 -6.32009625e-01 2.52292305e-01
9.68338609e-01 2.70369612e-02 -8.99113715e-01 5.15014410e-01
4.74437565e-01 -6.24478579e-01 1.10093415e+00 -3.47113997e-01
2.22666666e-01 -1.94163412e-01 -3.12716186e-01 -1.33369052e+00
-3.00595939e-01 -5.87466896e-01 -7.80448914e-02 1.59343660e+00
5.98934472e-01 -3.88299078e-01 4.22453821e-01 -1.54884577e-01
-6.83055937e-01 -4.26939875e-01 -1.24013460e+00 -1.05499172e+00
-1.39534092e-02 -1.30129898e+00 4.59029973e-01 5.57561815e-01
-3.40416789e-01 2.75638670e-01 -3.53623271e-01 3.65669936e-01
4.08148408e-01 1.78657293e-01 2.21901149e-01 -1.16834617e+00
-3.64299834e-01 -4.09334630e-01 -2.09602594e-01 -9.24179494e-01
3.77343923e-01 -9.00821686e-01 6.82643592e-01 -1.04858792e+00
-1.36600614e-01 -6.14237309e-01 -1.17848623e+00 5.88340044e-01
-2.94785172e-01 4.32980925e-01 6.25168011e-02 6.75574914e-02
-7.41134882e-01 7.09170341e-01 6.68345332e-01 7.47220591e-02
-8.94555897e-02 4.62545753e-01 -1.01624861e-01 7.12790728e-01
6.44219816e-01 -6.96580827e-01 -2.27399662e-01 -3.60252142e-01
-5.18892966e-02 1.09094389e-01 5.30064166e-01 -1.42839432e+00
1.93701267e-01 2.27291092e-01 -3.08511555e-02 -7.42378771e-01
5.90258539e-01 -4.80160922e-01 -1.03617989e-01 3.37331712e-01
-7.11584091e-01 -3.02125692e-01 4.79281634e-01 6.70273364e-01
-5.55187225e-01 -1.03537247e-01 6.94051266e-01 4.42161441e-01
-7.13405013e-01 1.86280817e-01 -6.41172528e-01 3.70804518e-01
7.82841802e-01 3.52895766e-01 5.55951446e-02 -2.17378065e-01
-8.76871884e-01 -1.83118489e-02 -5.54632962e-01 6.21225536e-01
4.15587336e-01 -1.52700365e+00 -9.38685238e-01 1.86432183e-01
2.39786431e-01 -2.09954828e-01 4.74382102e-01 5.38707733e-01
-4.60220641e-03 3.66042763e-01 1.49334073e-01 -6.38313174e-01
-1.18097520e+00 -1.36243608e-02 4.84395236e-01 1.50744687e-03
-6.22394145e-01 1.16841269e+00 5.13810933e-01 -3.19866538e-01
5.97930551e-01 -5.23374796e-01 -3.15510452e-01 3.78784925e-01
5.99184752e-01 6.24789834e-01 3.36419761e-01 -5.21995425e-01
-7.02593505e-01 4.88619387e-01 1.68753773e-01 -3.50733072e-01
1.35253084e+00 -4.37178165e-02 2.89656550e-01 9.14252341e-01
1.59203970e+00 1.71903417e-01 -1.45241487e+00 -4.31620568e-01
1.59885570e-01 -1.08739018e-01 3.03900272e-01 -9.80850041e-01
-1.03396010e+00 1.03082550e+00 6.39610708e-01 3.61448050e-01
1.25829339e+00 2.61508256e-01 1.29881084e+00 1.00912176e-01
5.46265487e-03 -1.02281463e+00 3.36826742e-01 7.08442569e-01
8.12853634e-01 -1.06457245e+00 -6.84827745e-01 -2.68817633e-01
-4.42355990e-01 8.80291104e-01 3.20138991e-01 -1.11117281e-01
1.04064727e+00 4.74593103e-01 2.60120749e-01 -6.35180473e-02
-8.81738007e-01 -3.12870979e-01 5.08133709e-01 4.58598047e-01
4.57677186e-01 9.90623385e-02 1.04704507e-01 1.11049199e+00
-3.32014918e-01 -6.11913860e-01 -1.08403727e-01 6.85008705e-01
-5.33083200e-01 -8.54818761e-01 -1.46313176e-01 3.91840994e-01
-8.22983801e-01 -2.37141281e-01 -2.71924194e-02 2.64515072e-01
1.94209486e-01 1.41923869e+00 3.74757767e-01 -6.12178147e-01
6.18199408e-01 2.98562795e-01 5.57067841e-02 -7.01702178e-01
-7.49831736e-01 3.24264467e-01 3.67139280e-01 -4.37550247e-01
-5.07686250e-02 -1.00721395e+00 -1.35066426e+00 5.40601969e-01
-9.76415053e-02 3.47575366e-01 6.27748787e-01 8.28265607e-01
3.09826285e-01 1.02727795e+00 1.09020007e+00 -8.60119998e-01
-6.28084302e-01 -9.46413100e-01 -5.49018860e-01 4.03754503e-01
7.74700165e-01 -6.17859662e-01 -6.87739670e-01 1.87911734e-01] | [15.19951057434082, 5.182312965393066] |
e8fb4f30-b0da-4d54-b304-19eafaec0ce0 | universal-successor-representations-for | 1804.03758 | null | http://arxiv.org/abs/1804.03758v1 | http://arxiv.org/pdf/1804.03758v1.pdf | Universal Successor Representations for Transfer Reinforcement Learning | The objective of transfer reinforcement learning is to generalize from a set
of previous tasks to unseen new tasks. In this work, we focus on the transfer
scenario where the dynamics among tasks are the same, but their goals differ.
Although general value function (Sutton et al., 2011) has been shown to be
useful for knowledge transfer, learning a universal value function can be
challenging in practice. To attack this, we propose (1) to use universal
successor representations (USR) to represent the transferable knowledge and (2)
a USR approximator (USRA) that can be trained by interacting with the
environment. Our experiments show that USR can be effectively applied to new
tasks, and the agent initialized by the trained USRA can achieve the goal
considerably faster than random initialization. | ['Yoshua Bengio', 'Junfeng Wen', 'Chen Ma'] | 2018-04-11 | null | null | null | null | ['transfer-reinforcement-learning'] | ['methodology'] | [ 3.58961225e-01 2.61442333e-01 -5.56637766e-03 -5.23185730e-02
-3.96834403e-01 -6.94599271e-01 8.00267398e-01 -5.46529368e-02
-5.13234735e-01 1.30197716e+00 -8.60631987e-02 -8.33319649e-02
-1.08254127e-01 -8.46398950e-01 -1.09550679e+00 -7.35480070e-01
-1.44540995e-01 5.90671480e-01 4.40597773e-01 -5.99458277e-01
5.99110778e-03 3.99001777e-01 -1.35472918e+00 2.19910275e-02
8.56418967e-01 5.77216268e-01 6.90658927e-01 7.25676835e-01
1.08908378e-01 8.58832777e-01 -7.91179895e-01 -6.08985983e-02
3.15122187e-01 -4.82984275e-01 -1.31445861e+00 -1.93263263e-01
-1.49163470e-01 -4.46276695e-01 -3.20600152e-01 8.82613778e-01
2.80539989e-01 8.29252362e-01 8.61434102e-01 -1.20093536e+00
-9.76658702e-01 6.28540158e-01 -1.31673858e-01 1.66140780e-01
2.06157610e-01 1.69188127e-01 6.53375864e-01 -3.86536837e-01
7.17409194e-01 1.24219656e+00 2.84279943e-01 1.15089381e+00
-1.41195667e+00 -6.21463180e-01 3.29754889e-01 1.31261438e-01
-9.18660820e-01 -1.53420389e-01 4.51309353e-01 -2.11575717e-01
8.84309173e-01 -2.99795210e-01 5.38212836e-01 1.51049757e+00
1.16510419e-02 8.28164697e-01 1.26556659e+00 -4.93243098e-01
5.47580302e-01 5.47603220e-02 2.26435787e-03 3.45377535e-01
1.62033483e-01 4.65386063e-01 -4.53674316e-01 -1.17686719e-01
1.02860034e+00 -8.69807228e-02 -3.74146730e-01 -6.72660530e-01
-1.14301717e+00 9.14137363e-01 7.80573845e-01 4.21460778e-01
-5.90319991e-01 5.15218139e-01 3.94867241e-01 8.42039287e-01
1.73239768e-01 8.88156652e-01 -6.67123556e-01 -1.53706208e-01
-8.66178721e-02 2.97209710e-01 6.73195243e-01 6.24907970e-01
9.85841334e-01 2.83217460e-01 -1.84624717e-01 7.55364478e-01
-1.74051985e-01 5.41493297e-01 8.35684240e-01 -1.07913303e+00
2.32717425e-01 5.76341636e-02 5.63456953e-01 -2.14453086e-01
-2.07021862e-01 -2.74255723e-01 -3.82475913e-01 4.10029829e-01
4.38041598e-01 -5.44334352e-01 -1.19601607e+00 2.13196254e+00
1.42192557e-01 5.45203388e-01 6.81579173e-01 6.38533711e-01
2.70417601e-01 8.73290420e-01 2.93117344e-01 -3.25197838e-02
7.24669635e-01 -9.04122114e-01 -4.63267684e-01 -5.76563656e-01
6.35378182e-01 -1.55940846e-01 8.58627915e-01 2.15107977e-01
-1.01317441e+00 -6.97333097e-01 -9.40494239e-01 1.82166860e-01
-5.73647141e-01 -4.90229696e-01 6.42029643e-01 2.54274458e-01
-1.39061058e+00 8.48237038e-01 -9.48081911e-01 -6.47323966e-01
2.62748539e-01 4.62789506e-01 -3.91102582e-01 -8.80152509e-02
-1.55740571e+00 1.37717807e+00 9.20772552e-01 -2.11027324e-01
-1.52664399e+00 -3.97721857e-01 -7.79485762e-01 1.32517755e-01
4.94365871e-01 -8.80826950e-01 1.65555716e+00 -1.24752140e+00
-1.94308317e+00 5.05230963e-01 1.62586614e-01 -6.67000532e-01
1.29326969e-01 -2.02774718e-01 1.72631159e-01 8.52087513e-02
-1.74433276e-01 7.65908301e-01 1.11142194e+00 -1.54544866e+00
-7.05797732e-01 -1.87233970e-01 4.91989315e-01 5.53461969e-01
-4.90572870e-01 -3.16313028e-01 1.01486690e-01 -5.78420103e-01
-3.14640731e-01 -1.04175663e+00 -2.71418422e-01 -4.06454235e-01
2.41353408e-01 -3.42001975e-01 6.04119062e-01 -5.29606342e-01
7.20326185e-01 -1.95950961e+00 6.07375860e-01 1.42472899e-02
-7.98739344e-02 5.09092450e-01 -4.93504643e-01 4.26579773e-01
-2.94746552e-02 -1.19216159e-01 -2.19811901e-01 -5.90079874e-02
-1.23657852e-01 6.54850662e-01 -5.20103097e-01 2.41687195e-03
1.51945487e-01 1.23332489e+00 -1.35757005e+00 6.54096156e-02
3.53670530e-02 3.03468227e-01 -3.34477782e-01 5.02291918e-01
-4.23232853e-01 6.97836936e-01 -7.15655565e-01 6.00258308e-03
2.40456581e-01 -2.02038497e-01 3.50737512e-01 4.23929244e-01
3.95673186e-01 2.44541809e-01 -9.35386539e-01 1.54469919e+00
-8.50006104e-01 4.77234632e-01 -5.38911596e-02 -1.35745358e+00
1.03503764e+00 3.94179434e-01 5.32177351e-02 -5.26749313e-01
3.37672867e-02 4.43682596e-02 2.75000542e-01 -2.99303532e-01
3.76522362e-01 -3.59769881e-01 -1.65032726e-02 6.69633090e-01
3.61252397e-01 -2.98084170e-01 -4.19413531e-03 6.33549783e-03
1.24795532e+00 4.67669576e-01 4.83883888e-01 -6.53180182e-02
3.27992022e-01 -5.83011843e-02 3.60632986e-01 1.03788495e+00
-2.13592291e-01 1.46111831e-01 2.55093902e-01 -3.36941510e-01
-9.43179846e-01 -1.34573328e+00 8.15555453e-02 1.47817659e+00
-5.84906526e-03 9.31189880e-02 -6.03087783e-01 -7.79497206e-01
1.17392771e-01 8.89031172e-01 -9.42274570e-01 -5.65557420e-01
-7.84378648e-01 -3.03108543e-01 2.31392428e-01 7.99965978e-01
5.37712991e-01 -1.56696224e+00 -9.62783933e-01 3.49004984e-01
-3.00792698e-02 -8.80740702e-01 -1.63061112e-01 5.78132153e-01
-9.33860898e-01 -8.19320858e-01 -9.93924916e-01 -9.46548760e-01
6.97066128e-01 2.12280512e-01 1.00787473e+00 2.11753566e-02
1.39488012e-01 7.96608865e-01 -5.47371268e-01 -5.16340673e-01
-7.40301192e-01 2.44219378e-01 1.46411851e-01 -1.93311751e-01
-2.45580599e-02 -6.05321348e-01 -2.93046325e-01 1.57884330e-01
-1.01306581e+00 -2.33127605e-02 5.28555572e-01 1.19173133e+00
2.66561866e-01 3.29469405e-02 1.16705573e+00 -8.29487860e-01
9.72149611e-01 -5.70773482e-01 -5.45883536e-01 2.88570225e-01
-4.52235073e-01 4.44083124e-01 9.48268414e-01 -8.22123826e-01
-1.20776200e+00 2.91772783e-02 2.17596486e-01 -4.29837912e-01
-1.53888181e-01 5.13405442e-01 2.98901111e-01 -2.78212801e-02
8.21661532e-01 3.81757408e-01 1.27591878e-01 -3.36541981e-01
5.10600805e-01 5.24625897e-01 4.32378590e-01 -1.03783309e+00
9.34205830e-01 2.07564667e-01 -1.12297416e-01 -4.71070200e-01
-9.01690006e-01 2.09282693e-02 -5.21644354e-01 -6.25190958e-02
8.63518775e-01 -7.54996419e-01 -6.51018798e-01 3.39378417e-01
-9.68501508e-01 -1.29423320e+00 -5.15196264e-01 4.27123517e-01
-9.51178968e-01 -2.30460316e-02 -4.77871448e-01 -8.00586283e-01
-2.79336944e-02 -9.63305831e-01 6.05977058e-01 2.97714025e-01
-1.09382674e-01 -1.33028734e+00 2.70961881e-01 -1.30679637e-01
6.40177310e-01 1.91046253e-01 9.20330048e-01 -8.23290169e-01
-3.69740307e-01 2.23697066e-01 1.08827807e-01 5.71740389e-01
2.59108484e-01 -6.01878226e-01 -8.56239736e-01 -6.40277982e-01
1.95458811e-03 -9.67560410e-01 9.40450549e-01 2.06985384e-01
1.00029862e+00 -5.58778159e-02 -4.01669413e-01 1.57495856e-01
1.07267690e+00 4.88804877e-01 5.64333022e-01 5.84839940e-01
2.15863109e-01 4.07964289e-01 5.98692775e-01 1.14547975e-01
3.00825775e-01 5.79809308e-01 4.30506259e-01 2.72824317e-01
-1.00560181e-01 -2.83952951e-01 7.08898783e-01 3.33493471e-01
-2.86209673e-01 -7.40512833e-02 -8.95366013e-01 5.38111150e-01
-2.16541767e+00 -8.16500187e-01 6.21736407e-01 2.29965425e+00
9.69743073e-01 2.29621395e-01 4.76200581e-02 -3.54569763e-01
6.59265935e-01 -8.74046981e-02 -8.69606733e-01 -5.88556707e-01
2.85099655e-01 6.62840545e-01 1.76012754e-01 5.52015245e-01
-7.75943279e-01 1.26715899e+00 6.75264311e+00 4.04019505e-01
-9.89769518e-01 1.71862707e-01 3.00544351e-01 1.50289834e-01
5.24314158e-02 1.00792997e-01 -5.58221757e-01 1.96694493e-01
9.47960317e-01 -4.34024125e-01 1.03665543e+00 8.57521176e-01
-3.24261129e-01 -2.04182435e-02 -1.28029394e+00 5.97909570e-01
3.50572467e-02 -8.15759778e-01 7.91682824e-02 -2.15853438e-01
9.68297184e-01 -7.10667521e-02 1.50785953e-01 1.04273236e+00
1.01398540e+00 -1.11259532e+00 2.95328379e-01 4.34490085e-01
5.40984154e-01 -7.71387875e-01 5.29712796e-01 5.77137113e-01
-8.44851017e-01 -3.41847807e-01 -5.54026008e-01 -2.49033391e-01
-1.38568476e-01 -2.33708024e-01 -1.08555984e+00 4.38532114e-01
5.29451430e-01 5.58165491e-01 -4.27623540e-01 6.12276435e-01
-6.05805874e-01 4.37483460e-01 -1.49012342e-01 -2.45532274e-01
3.68556172e-01 -1.11750469e-01 3.18012804e-01 7.31891870e-01
3.70326757e-01 1.48515239e-01 2.66160309e-01 7.27019429e-01
-2.59652287e-01 -2.38252789e-01 -9.41543758e-01 -3.72466706e-02
3.75718713e-01 1.06989813e+00 -4.90409523e-01 -3.52837682e-01
-1.86088651e-01 1.18595612e+00 8.74156117e-01 7.95179307e-01
-6.94789410e-01 -3.67505848e-01 6.49417162e-01 -2.08071128e-01
6.52615190e-01 -1.90223575e-01 2.84667313e-01 -1.12037885e+00
-3.14946145e-01 -7.54877687e-01 3.40036958e-01 -1.07735753e+00
-1.25487018e+00 5.34372389e-01 2.32773721e-01 -8.74608219e-01
-6.44749224e-01 -7.10726082e-01 -5.56684792e-01 9.24543560e-01
-1.70887971e+00 -8.79645765e-01 -1.11606516e-01 8.26250851e-01
5.41973829e-01 -2.80181766e-01 1.02539039e+00 -3.93996477e-01
-2.82202959e-01 3.52137625e-01 3.01007360e-01 2.79288143e-02
6.85719252e-01 -1.44827998e+00 3.82411093e-01 4.39681679e-01
-9.26083978e-03 6.98318660e-01 7.38108158e-01 -6.06569588e-01
-1.30373085e+00 -1.09964120e+00 1.99938133e-01 -6.47030056e-01
8.47378910e-01 -3.17639232e-01 -1.17302477e+00 1.26670909e+00
4.09678489e-01 -1.07105814e-01 2.03641757e-01 2.24423811e-01
-3.38966191e-01 4.01099175e-02 -9.75354135e-01 7.74324000e-01
9.13979650e-01 -3.88472885e-01 -1.09024692e+00 2.98058063e-01
8.03502083e-01 -5.36017954e-01 -8.24875712e-01 1.53263822e-01
1.82377428e-01 -4.90769476e-01 1.02829051e+00 -1.20823181e+00
2.61348188e-01 -1.05774924e-01 1.21636622e-01 -2.13651443e+00
-3.79586995e-01 -6.30504429e-01 -4.16998684e-01 8.34291041e-01
3.06583285e-01 -1.05571282e+00 3.86027038e-01 4.72687393e-01
-1.30924555e-02 -4.59078014e-01 -9.11174893e-01 -1.23688900e+00
5.56958616e-01 5.59777068e-03 5.46479404e-01 8.88297319e-01
1.43544540e-01 6.45182848e-01 -3.77828270e-01 1.35671899e-01
2.50297278e-01 4.15339991e-02 8.31980944e-01 -1.36400497e+00
-6.57653630e-01 -1.68571427e-01 -4.38939855e-02 -1.06174505e+00
6.40238345e-01 -1.13972962e+00 2.03427076e-01 -1.58321154e+00
2.72938281e-01 -3.90762895e-01 -6.23655975e-01 8.53276432e-01
-3.55386198e-01 -2.28465110e-01 4.25584018e-01 -3.93128488e-03
-5.97379863e-01 6.44455075e-01 1.59842467e+00 1.28110889e-02
-3.83265704e-01 1.13485180e-01 -7.26082981e-01 5.69164753e-01
1.09016943e+00 -4.87728894e-01 -8.19669366e-01 -4.76121336e-01
2.39012778e-01 8.46133530e-02 2.08599865e-01 -8.88715088e-01
1.81487240e-02 -2.24487588e-01 4.27997649e-01 2.49770522e-01
4.18511003e-01 -6.48885190e-01 -2.02789143e-01 6.76643729e-01
-5.71260750e-01 -1.35149315e-01 3.09178740e-01 6.35022461e-01
6.52671680e-02 -7.53104568e-01 7.39066184e-01 -2.57552177e-01
-6.70734823e-01 7.79653192e-02 -3.16310495e-01 2.05327690e-01
1.15583444e+00 9.03573707e-02 -4.37450826e-01 -6.23601794e-01
-9.52897370e-01 4.70890224e-01 4.16115046e-01 4.40153420e-01
5.64396143e-01 -1.27527976e+00 -5.99942684e-01 -1.06211968e-01
4.51625697e-02 7.92955756e-02 -8.95359218e-02 1.72042012e-01
-6.47502169e-02 1.67554080e-01 -5.22464156e-01 -2.28653952e-01
-6.25716865e-01 8.25739622e-01 4.03878391e-01 -5.66415608e-01
-4.14499611e-01 4.43198532e-01 4.34722036e-01 -5.54592133e-01
9.91620272e-02 -1.59302816e-01 -2.43072227e-01 -1.44772872e-01
4.70350236e-01 1.89292118e-01 -1.55431122e-01 -1.07065760e-01
8.24906975e-02 2.62796581e-01 -5.03251553e-01 -2.13210583e-01
1.42343974e+00 1.51996911e-02 3.03987116e-01 4.66582775e-01
9.15809512e-01 -7.19098866e-01 -1.53756475e+00 -4.59365457e-01
-1.11774765e-01 -2.07921162e-01 -2.02171564e-01 -9.54663754e-01
-6.44751728e-01 8.00373793e-01 4.87962723e-01 3.99324030e-01
1.04514229e+00 6.95243478e-02 6.74730122e-01 9.98799682e-01
6.51000679e-01 -1.14096534e+00 5.28389692e-01 8.86484265e-01
9.67022061e-01 -1.17212212e+00 -3.61470908e-01 1.85015962e-01
-8.62907290e-01 1.06822896e+00 8.26983988e-01 -3.10774863e-01
3.57279658e-01 4.01975438e-02 -2.39351019e-01 2.54357338e-01
-8.91465247e-01 -6.12400115e-01 -1.62790388e-01 1.24457037e+00
-1.74249653e-02 -1.33511767e-01 5.83630502e-02 3.01547080e-01
4.84126359e-02 2.12149933e-01 5.78426659e-01 1.23382688e+00
-5.77155232e-01 -1.30406952e+00 -3.31847966e-01 3.69938135e-01
-8.64342600e-02 1.47831842e-01 -1.79104477e-01 9.11848366e-01
-2.63871849e-01 6.30167544e-01 -4.24950123e-02 -3.79786864e-02
3.61843139e-01 3.59407306e-01 8.21733475e-01 -9.03566897e-01
-6.21316135e-01 -3.59341025e-01 -1.92518860e-01 -2.81720817e-01
-1.83926821e-01 -4.97702926e-01 -1.43704641e+00 1.04498856e-01
3.46844047e-02 1.48729622e-01 4.07771319e-01 8.90114069e-01
2.34162450e-01 7.02582479e-01 7.12194979e-01 -8.49850595e-01
-1.08250248e+00 -9.82735515e-01 -5.65008581e-01 6.37118876e-01
5.26168644e-01 -1.02316785e+00 -2.99152464e-01 3.19874436e-01] | [4.1726884841918945, 1.4765859842300415] |
548b5aa4-f317-40eb-bb3c-49645f9cf7d3 | st360iq-no-reference-omnidirectional-image | 2303.06907 | null | https://arxiv.org/abs/2303.06907v1 | https://arxiv.org/pdf/2303.06907v1.pdf | ST360IQ: No-Reference Omnidirectional Image Quality Assessment with Spherical Vision Transformers | Omnidirectional images, aka 360 images, can deliver immersive and interactive visual experiences. As their popularity has increased dramatically in recent years, evaluating the quality of 360 images has become a problem of interest since it provides insights for capturing, transmitting, and consuming this new media. However, directly adapting quality assessment methods proposed for standard natural images for omnidirectional data poses certain challenges. These models need to deal with very high-resolution data and implicit distortions due to the spherical form of the images. In this study, we present a method for no-reference 360 image quality assessment. Our proposed ST360IQ model extracts tangent viewports from the salient parts of the input omnidirectional image and employs a vision-transformers based module processing saliency selective patches/tokens that estimates a quality score from each viewport. Then, it aggregates these scores to give a final quality score. Our experiments on two benchmark datasets, namely OIQA and CVIQ datasets, demonstrate that as compared to the state-of-the-art, our approach predicts the quality of an omnidirectional image correlated with the human-perceived image quality. The code has been available on https://github.com/Nafiseh-Tofighi/ST360IQ | ['Aykut Erdem', 'Erkut Erdem', 'Cagri Ozcinar', 'Nevrez Imamoglu', 'Mohamed Hedi Elfkir', 'Nafiseh Jabbari Tofighi'] | 2023-03-13 | null | null | null | null | ['image-quality-assessment'] | ['computer-vision'] | [ 4.62811999e-02 -1.91004202e-01 4.53651249e-02 -2.94283032e-01
-6.36228740e-01 -4.16975290e-01 4.12729532e-01 2.65962481e-02
-1.75755858e-01 3.25143486e-01 4.56590265e-01 -1.48149207e-01
-2.09448606e-01 -7.91155875e-01 -6.86497152e-01 -4.30144042e-01
1.13844415e-02 -1.84403136e-01 3.25979799e-01 -3.27428907e-01
5.68817139e-01 3.18027884e-01 -1.91743708e+00 3.77818227e-01
1.09867287e+00 1.33315873e+00 5.01395583e-01 7.75766611e-01
2.24710912e-01 8.21366370e-01 -4.01394695e-01 -5.02298892e-01
3.54822814e-01 -1.09303400e-01 -6.21791363e-01 -4.91012521e-02
5.68788528e-01 -5.35519838e-01 -3.32107782e-01 1.41155005e+00
6.12150669e-01 -4.37586233e-02 4.07318383e-01 -1.04899955e+00
-7.23075032e-01 1.60445631e-01 -6.55601919e-01 4.63911980e-01
7.27940083e-01 6.55571893e-02 9.60843384e-01 -1.24870849e+00
7.13311255e-01 1.07670426e+00 1.90808102e-01 1.32926002e-01
-7.59037435e-01 -4.07839000e-01 2.57253721e-02 6.44901872e-01
-1.19387329e+00 -5.13145983e-01 7.87876725e-01 -1.46654516e-01
6.22008264e-01 5.37292480e-01 9.84475970e-01 5.66545486e-01
3.19850415e-01 9.57336903e-01 1.05048323e+00 -1.77858353e-01
2.10551918e-01 1.88603714e-01 -3.32742751e-01 6.04592502e-01
-1.28704645e-02 2.11346954e-01 -4.94193554e-01 2.75689214e-01
7.37467170e-01 -9.74093750e-03 -7.13621080e-01 -7.44211674e-01
-1.26119673e+00 2.82979429e-01 6.64932489e-01 1.06355377e-01
-4.13880199e-01 -2.96211183e-01 2.64177173e-01 2.55799055e-01
5.48042297e-01 2.72986770e-01 -1.68590948e-01 -3.28387618e-01
-7.18599141e-01 1.83532760e-01 2.28582263e-01 7.72939265e-01
5.54992676e-01 1.71740521e-02 1.50111606e-02 9.99410391e-01
1.12030767e-01 7.83043444e-01 3.01113278e-01 -1.03535032e+00
4.67037708e-01 4.44280177e-01 4.27892119e-01 -1.45846045e+00
-2.15728462e-01 -5.49027205e-01 -7.74305761e-01 5.34054995e-01
1.32761821e-01 3.43018621e-01 -4.86962587e-01 1.34548736e+00
4.44219619e-01 -8.43263716e-02 8.91565438e-03 1.50185275e+00
9.67150390e-01 8.23985696e-01 -3.93769592e-01 1.89004559e-02
1.20103955e+00 -8.33583236e-01 -6.06195569e-01 3.68134379e-02
-7.31430873e-02 -9.01335895e-01 1.43106127e+00 9.07803535e-01
-1.41383862e+00 -7.66005993e-01 -1.14476788e+00 -1.93217710e-01
-1.71922803e-01 -1.61364786e-02 4.05605108e-01 6.45903587e-01
-1.35949850e+00 2.03453854e-01 -3.02267015e-01 -5.28292321e-02
1.85028344e-01 -9.91098210e-02 -2.26543918e-01 -3.98849845e-01
-1.12204421e+00 7.53388405e-01 -2.20953468e-02 -7.57118538e-02
-9.26427186e-01 -7.12531149e-01 -8.61373663e-01 1.24957718e-01
3.06673825e-01 -5.78623056e-01 9.75355268e-01 -9.13750648e-01
-1.39235592e+00 6.70893610e-01 -1.64075494e-01 -3.41385342e-02
5.28798342e-01 -3.17685336e-01 -6.27863109e-01 3.33440363e-01
1.02608569e-01 5.53484678e-01 7.92499006e-01 -1.53662109e+00
-9.75206673e-01 -4.90920603e-01 5.41502297e-01 7.44896710e-01
-1.32072181e-01 -2.34968718e-02 -7.61879623e-01 -4.36057687e-01
3.02060246e-01 -4.82216299e-01 7.33040497e-02 1.60866082e-01
-2.15999663e-01 2.14772314e-01 6.18861198e-01 -8.03044081e-01
1.10430300e+00 -2.14037991e+00 5.04506864e-02 9.71529111e-02
2.78779775e-01 1.33440956e-01 -1.31729767e-01 8.25814828e-02
1.41340613e-01 -9.98425409e-02 1.29601717e-01 -8.29525664e-02
-3.32813919e-01 -3.65087986e-01 -9.81232226e-02 4.22860831e-01
-9.03470591e-02 6.96071148e-01 -1.17067361e+00 -4.50145245e-01
7.81457067e-01 5.96291125e-01 -6.09736443e-01 3.71807456e-01
2.72262931e-01 4.59163755e-01 -1.08175717e-01 7.28500128e-01
1.12976050e+00 -1.14966661e-01 -1.66714087e-01 -5.53321779e-01
-3.50860000e-01 1.12008937e-01 -1.26430464e+00 1.81786096e+00
-7.92318225e-01 6.47099733e-01 4.44714613e-02 -6.47738636e-01
8.04738224e-01 9.21369940e-02 3.92207623e-01 -1.37551165e+00
-1.56290885e-02 1.66405424e-01 -3.91730338e-01 -6.44138455e-01
8.66980433e-01 3.31841052e-01 5.15576303e-02 -3.00270338e-02
-1.63474634e-01 -4.17544574e-01 1.21149473e-01 2.97571868e-01
6.50834501e-01 2.35710870e-02 4.82127964e-01 -1.06526189e-01
6.33758247e-01 -3.64207566e-01 3.96840394e-01 4.34813887e-01
-1.97303921e-01 1.04369783e+00 8.11621249e-02 -4.93721604e-01
-1.25113142e+00 -1.58621871e+00 -8.50580931e-02 7.28020847e-01
9.07814384e-01 -2.67829716e-01 -6.97146058e-01 -2.52066702e-01
-4.50286388e-01 5.61444461e-01 -3.58538151e-01 -4.95806970e-02
-2.21691400e-01 -2.56448597e-01 -1.18620038e-01 1.38346106e-01
6.71756923e-01 -9.35911596e-01 -9.63702142e-01 2.41918992e-02
-7.90324926e-01 -1.13237822e+00 -5.27195811e-01 -4.46127176e-01
-6.09847665e-01 -8.46099973e-01 -1.02656269e+00 -4.94363964e-01
3.99888068e-01 9.09129500e-01 1.37620234e+00 -1.35347322e-01
-1.90511584e-01 2.98398137e-01 -4.07816499e-01 -2.21404374e-01
6.44385349e-03 -3.35510373e-01 -1.21908136e-01 1.77728087e-01
-1.29601836e-01 -5.72440565e-01 -1.22702718e+00 3.95771027e-01
-1.06440616e+00 5.55386186e-01 6.02028310e-01 5.38030744e-01
7.26108372e-01 2.80990321e-02 2.44106457e-01 -3.73857737e-01
4.70566839e-01 -3.60439628e-01 -5.88871896e-01 7.13130012e-02
-4.47303236e-01 -1.68575078e-01 7.54479766e-01 1.60513390e-02
-1.24723601e+00 -3.89859289e-01 -2.61113554e-01 -2.96496451e-01
-1.22673988e-01 3.34157109e-01 -4.70745146e-01 1.70967057e-01
4.88285482e-01 2.68443882e-01 -3.20770353e-01 -2.22263813e-01
3.91832024e-01 7.89032817e-01 7.01718509e-01 -3.31041664e-02
5.89202762e-01 8.29158902e-01 -2.33471632e-01 -8.30329001e-01
-7.16111422e-01 -6.73681796e-01 -1.09786637e-01 -7.58763433e-01
6.23433888e-01 -1.02284932e+00 -6.93058074e-01 4.22633141e-01
-1.02830565e+00 4.72590104e-02 -2.74041414e-01 3.83913815e-01
-6.75906062e-01 3.81365478e-01 -3.51550132e-01 -7.30100393e-01
-3.19677621e-01 -1.34951222e+00 1.08408451e+00 4.57385808e-01
2.03342795e-01 -6.16130471e-01 -2.81638745e-02 3.90957296e-01
5.16578674e-01 3.09040840e-03 5.55034518e-01 4.04264718e-01
-7.63173699e-01 1.18955582e-01 -5.95645666e-01 3.64485115e-01
1.63859889e-01 -2.73768067e-01 -9.57083762e-01 -1.88083813e-01
1.08693801e-01 -2.58292019e-01 5.13186216e-01 6.49547517e-01
1.26815379e+00 -2.75625110e-01 1.87980816e-01 8.15186620e-01
1.61534810e+00 2.36305878e-01 1.04239988e+00 3.81616086e-01
6.30392253e-01 5.19100368e-01 1.10864556e+00 6.23085022e-01
7.03092039e-01 9.74562705e-01 8.59113097e-01 -3.68499756e-01
-6.72067627e-02 -2.97494322e-01 3.33053470e-01 7.72295296e-01
-2.53461957e-01 -3.27300429e-01 -6.36998117e-01 6.86710060e-01
-1.46844399e+00 -9.95755851e-01 4.32059467e-02 2.30343699e+00
5.06161392e-01 -2.26526451e-03 -7.60530233e-02 5.26981473e-01
5.07901609e-01 3.52951586e-01 -4.97686356e-01 -4.94111925e-01
-1.62727356e-01 -6.22920841e-02 3.60240281e-01 6.12114429e-01
-7.77521014e-01 6.38186574e-01 5.32353783e+00 8.92525733e-01
-1.34284377e+00 1.36233773e-02 7.80099869e-01 -3.42372134e-02
-6.52759373e-01 -2.90805489e-01 -2.94381946e-01 3.41865867e-01
5.76934457e-01 -1.49176314e-01 5.64048588e-01 7.62335777e-01
5.76905370e-01 -4.43671882e-01 -6.51817858e-01 1.41387773e+00
2.59217292e-01 -1.08460820e+00 2.23239765e-01 2.37806328e-02
8.49425256e-01 -2.57616024e-02 6.75251484e-01 -3.02740574e-01
2.28201412e-02 -9.12235856e-01 1.09856570e+00 7.73475945e-01
9.93518353e-01 -1.02975047e+00 6.85853004e-01 2.03875214e-01
-1.24622178e+00 -1.54522480e-02 -5.68830311e-01 2.39905831e-03
2.84498602e-01 9.07201529e-01 -4.26915050e-01 8.44468355e-01
1.30724609e+00 7.77701616e-01 -6.61437750e-01 1.11220634e+00
-7.99053013e-02 1.06278308e-01 3.71213928e-02 1.56643286e-01
-1.85612187e-01 -2.34377906e-01 6.79796875e-01 9.34107363e-01
6.26733899e-01 1.14244543e-01 -2.92659879e-01 6.51900172e-01
3.13164324e-01 2.67261326e-01 -6.63552880e-01 5.07041335e-01
1.94706455e-01 1.21449137e+00 -4.15032953e-01 -2.92779297e-01
-4.34976757e-01 1.18450201e+00 8.73031840e-02 3.82581055e-01
-8.10536861e-01 -3.74532640e-01 6.58425570e-01 2.76439518e-01
2.06616968e-01 -7.79771358e-02 -4.42980230e-01 -1.45524406e+00
3.05085570e-01 -1.05239820e+00 1.04502216e-01 -1.41421402e+00
-7.27845788e-01 7.97067344e-01 -2.16179237e-01 -1.80806136e+00
-7.38846436e-02 -3.11707228e-01 -1.96941808e-01 7.31426597e-01
-1.68422365e+00 -7.46758997e-01 -9.55925822e-01 5.98506093e-01
6.35970175e-01 2.18325481e-01 4.49225962e-01 3.57083619e-01
1.21734731e-01 4.05047834e-01 1.97649673e-01 -1.60219371e-01
5.64217269e-01 -1.10839522e+00 4.26968873e-01 9.53602433e-01
8.30858797e-02 2.11108670e-01 9.33617532e-01 -1.77676275e-01
-1.33464503e+00 -7.86159933e-01 6.23900175e-01 -2.56128162e-01
1.57950036e-02 -1.64740354e-01 -7.19879746e-01 -2.71861460e-02
2.16466621e-01 1.26914412e-01 1.86674654e-01 -4.43624139e-01
-1.55791134e-01 -5.20311058e-01 -1.18722725e+00 7.49810696e-01
1.33171391e+00 -4.35986519e-01 -1.55669764e-01 -4.43252206e-01
4.74712819e-01 -4.80610609e-01 -7.35774398e-01 5.14651239e-01
8.11441481e-01 -1.78108859e+00 1.16943407e+00 2.44187072e-01
7.57260025e-01 -5.72408497e-01 -2.54086822e-01 -1.60800195e+00
-2.59634376e-01 -6.30220249e-02 2.05405027e-01 8.54494870e-01
6.18856996e-02 -3.33800733e-01 4.63479966e-01 3.05058081e-02
-6.76584393e-02 -5.93442976e-01 -8.68223906e-01 -4.41198051e-01
-5.43399751e-01 -5.02794802e-01 5.98639369e-01 5.74608684e-01
4.77804849e-03 3.97991687e-02 -5.63025296e-01 2.11940944e-01
7.74433374e-01 1.26818016e-01 7.63524413e-01 -9.18051362e-01
-1.02177747e-01 -2.90115654e-01 -7.65170872e-01 -1.32436085e+00
-5.31695545e-01 -5.41916668e-01 -6.94076568e-02 -1.82560194e+00
1.89077243e-01 -3.84418726e-01 -3.24215055e-01 -2.74684459e-01
-1.99493498e-01 6.48090661e-01 2.11714044e-01 1.06746733e-01
-8.39249611e-01 7.34848619e-01 1.56467724e+00 -2.29167685e-01
-9.59033668e-02 -1.41611084e-01 -8.49162459e-01 5.78748465e-01
7.88198471e-01 2.66033057e-02 -6.45811856e-01 -5.76313555e-01
5.35970509e-01 3.57485861e-01 3.95084292e-01 -1.41487384e+00
1.33692604e-02 -1.30118296e-01 4.38166142e-01 -6.71922624e-01
4.41791117e-01 -8.56286108e-01 3.06423277e-01 2.80380428e-01
-1.42045125e-01 2.74201542e-01 -4.66185771e-02 3.96055102e-01
-4.97878611e-01 1.28080130e-01 9.52131212e-01 -8.72370154e-02
-9.86697614e-01 2.01063812e-01 -1.42836004e-01 6.26622289e-02
6.86972499e-01 -4.07738596e-01 -2.12093502e-01 -8.86600971e-01
-3.95795554e-01 -7.48591721e-02 9.32790458e-01 6.63675785e-01
1.25845861e+00 -1.36237133e+00 -7.15366602e-01 2.05153659e-01
4.09660399e-01 -1.54901370e-01 7.10582256e-01 6.13936305e-01
-5.88137567e-01 3.19910169e-01 -6.19718373e-01 -8.58725965e-01
-1.31513524e+00 6.80684328e-01 2.54061073e-01 -1.12343736e-01
-5.04919112e-01 3.94501537e-01 7.16886938e-01 -1.40848115e-01
7.55164633e-03 -4.59640652e-01 -5.93455017e-01 -2.63783455e-01
8.05835962e-01 4.59811211e-01 1.84815690e-01 -8.57444644e-01
-2.24617660e-01 7.43666410e-01 1.84074283e-01 -4.37935293e-01
1.15689003e+00 -7.47601092e-01 2.25385621e-01 4.12300199e-01
1.13924384e+00 2.92309761e-01 -1.25843680e+00 -2.00912237e-01
-4.46816415e-01 -1.08451557e+00 1.24688618e-01 -7.97829986e-01
-1.15303755e+00 8.64811778e-01 1.11537325e+00 3.07722706e-02
1.60670972e+00 -7.81604573e-02 5.69792688e-01 -1.69785947e-01
7.11550176e-01 -8.96365583e-01 2.36194268e-01 2.41294369e-01
1.15334678e+00 -1.34708273e+00 -1.41965419e-01 -5.33833146e-01
-7.56564081e-01 7.39413142e-01 5.32380342e-01 1.19972989e-01
4.96214032e-01 7.09677115e-02 3.47428590e-01 2.90821102e-02
-4.62713927e-01 -1.75662190e-01 5.36033511e-01 8.08946550e-01
2.38243461e-01 1.15223587e-01 -2.42554903e-01 2.95716286e-01
-4.31127071e-01 6.84561580e-02 7.58337855e-01 5.73350191e-01
-4.15587634e-01 -5.08309841e-01 -3.98212999e-01 2.37802669e-01
-5.29663444e-01 -1.82523519e-01 1.46838129e-01 2.10562810e-01
8.09927285e-02 1.23267627e+00 2.83991173e-02 -6.06210232e-01
6.00679159e-01 -7.50019848e-01 3.70345056e-01 -2.18407497e-01
-1.75181732e-01 -1.12702556e-01 -1.13919050e-01 -8.86001170e-01
-5.55710614e-01 -4.08303440e-01 -8.91114652e-01 -2.73591638e-01
-3.10589597e-02 8.93961191e-02 8.22999001e-01 4.17819262e-01
3.94362241e-01 4.41544235e-01 8.61746848e-01 -1.11679101e+00
-8.26456919e-02 -7.18134284e-01 -5.21755755e-01 6.02188349e-01
3.85095984e-01 -6.12053454e-01 -3.15351456e-01 -7.85195604e-02] | [11.753669738769531, -1.9467461109161377] |
4509b85b-879c-4960-ab3a-4deca69fa3f1 | tf-coder-program-synthesis-for-tensor | 2003.09040 | null | https://arxiv.org/abs/2003.09040v4 | https://arxiv.org/pdf/2003.09040v4.pdf | TF-Coder: Program Synthesis for Tensor Manipulations | The success and popularity of deep learning is on the rise, partially due to powerful deep learning frameworks such as TensorFlow and PyTorch that make it easier to develop deep learning models. However, these libraries also come with steep learning curves, since programming in these frameworks is quite different from traditional imperative programming with explicit loops and conditionals. In this work, we present a tool called TF-Coder for programming by example in TensorFlow. TF-Coder uses a bottom-up weighted enumerative search, with value-based pruning of equivalent expressions and flexible type- and value-based filtering to ensure that expressions adhere to various requirements imposed by the TensorFlow library. We train models to predict TensorFlow operations from features of the input and output tensors and natural language descriptions of tasks, to prioritize relevant operations during search. TF-Coder solves 63 of 70 real-world tasks within 5 minutes, sometimes finding simpler solutions in less time compared to experienced human programmers. | ['David Bieber', 'Rishabh Singh', 'Kensen Shi'] | 2020-03-19 | null | https://openreview.net/forum?id=nJ5Ij53umw2 | https://openreview.net/pdf?id=nJ5Ij53umw2 | neurips-workshop-cap-2020-12 | ['enumerative-search'] | ['computer-code'] | [-4.17617947e-01 -3.40600997e-01 -2.16669932e-01 -6.75069988e-01
-9.85965803e-02 -6.75541639e-01 1.93881169e-01 3.90296668e-01
-5.30954421e-01 8.67253318e-02 2.12263748e-01 -7.63544500e-01
-3.01110476e-01 -7.15560138e-01 -4.15842295e-01 -8.71491432e-02
-5.49088478e-01 4.62449789e-01 1.68133482e-01 -1.20719083e-01
5.08923531e-01 6.25484467e-01 -1.76487541e+00 8.15548003e-01
7.15191066e-01 1.15378153e+00 1.49102181e-01 1.08030307e+00
-7.49040186e-01 1.41619885e+00 -4.04872596e-01 -6.76937640e-01
3.51744592e-01 4.60002631e-01 -1.18575168e+00 -4.42153066e-01
5.79390466e-01 -4.39040273e-01 -1.02716327e-01 8.19208443e-01
2.11156487e-01 1.05827056e-01 1.51060700e-01 -1.20069540e+00
-2.27658585e-01 8.11276972e-01 -3.88924748e-01 4.70721453e-01
2.37455055e-01 5.30027211e-01 1.47150052e+00 -8.64702940e-01
5.22190511e-01 1.24730074e+00 8.40114534e-01 2.11383045e-01
-1.33759224e+00 -5.01834035e-01 3.04182945e-03 1.43656731e-01
-9.13792491e-01 -3.09495628e-01 1.93091318e-01 -8.44978273e-01
1.79782414e+00 4.31535602e-01 6.78960204e-01 6.49207771e-01
3.25012684e-01 8.73855770e-01 6.19231939e-01 -2.65172154e-01
2.26553231e-01 -1.82275921e-01 6.06785297e-01 1.09382415e+00
4.95323539e-03 -4.33437340e-02 -4.11342859e-01 -5.36782861e-01
5.09356439e-01 1.95641652e-01 1.31238997e-01 -4.50906783e-01
-1.28071809e+00 7.40960419e-01 4.32758480e-01 3.21108788e-01
-3.42214555e-01 4.40296024e-01 1.04923749e+00 6.65755153e-01
2.01703697e-01 9.34813619e-01 -9.10154700e-01 -5.71616173e-01
-9.08974946e-01 6.12707853e-01 1.27701426e+00 9.16693687e-01
8.27317059e-01 -2.95630284e-03 -2.84660816e-01 7.48531878e-01
1.61610544e-01 -1.87565759e-01 4.36721414e-01 -1.23292375e+00
5.82223415e-01 8.38068128e-01 -3.20732862e-01 -1.06715381e+00
-4.28343564e-01 -5.73308945e-01 -3.94674569e-01 3.72170627e-01
5.40157735e-01 -7.16954991e-02 -7.28911519e-01 1.24776065e+00
5.39712310e-02 -4.69249129e-01 -3.60149741e-01 6.85738087e-01
5.38506866e-01 5.49990296e-01 2.23705813e-01 3.40110779e-01
1.24477351e+00 -1.12997591e+00 -2.64636606e-01 -3.36448431e-01
1.37734210e+00 -6.82466149e-01 1.36281288e+00 8.47977757e-01
-1.21500421e+00 -2.79368669e-01 -6.59449339e-01 -4.96642917e-01
-2.87631780e-01 1.53925391e-02 1.35713506e+00 5.08490026e-01
-1.13457334e+00 8.35522950e-01 -1.05372155e+00 3.49344835e-02
4.74110276e-01 5.43476939e-01 -2.78698504e-01 -2.00010091e-02
-7.00006783e-01 9.43652689e-01 4.76396829e-01 2.53254478e-03
-1.04999804e+00 -1.21641243e+00 -7.49855518e-01 4.05104280e-01
3.67364526e-01 -5.50311148e-01 1.63210690e+00 -8.88361275e-01
-1.13146043e+00 8.76003027e-01 6.44195303e-02 -5.19249737e-01
3.72989178e-01 -3.85582328e-01 1.35268226e-01 -2.44139031e-01
-5.48096336e-02 2.69276828e-01 6.75328612e-01 -5.80533028e-01
-6.30122542e-01 -2.52613902e-01 5.39430380e-01 -2.02921569e-01
-3.51522714e-01 4.61747348e-01 -4.80381429e-01 -2.16180727e-01
-2.70935684e-01 -6.67418182e-01 -5.04268885e-01 1.88019037e-01
-3.12458158e-01 -3.82536978e-01 3.89484942e-01 -5.48050761e-01
1.36232805e+00 -2.06372976e+00 2.03810066e-01 2.10349381e-01
8.66676033e-01 2.30575487e-01 -1.98854089e-01 2.81964451e-01
-3.84900779e-01 1.34978533e-01 -6.34149909e-02 -5.19897304e-02
1.63965970e-01 3.16437751e-01 -2.76370972e-01 2.35365823e-01
8.70858729e-02 5.37058711e-01 -1.11284626e+00 -4.62674171e-01
5.12463115e-02 8.82945955e-02 -1.08760953e+00 2.24474341e-01
-4.71003711e-01 -3.28757137e-01 -3.46653879e-01 7.52835929e-01
4.57044333e-01 -2.56826311e-01 1.46321803e-01 -2.13976465e-02
-5.12747288e-01 6.80300653e-01 -9.83480930e-01 1.68484402e+00
-9.29309547e-01 7.66713679e-01 3.27379227e-01 -1.10116720e+00
7.44634926e-01 -1.45778000e-01 4.58472729e-01 -5.60467958e-01
3.55162434e-02 3.58382672e-01 2.06759349e-01 -7.27702916e-01
5.41543365e-01 3.73111248e-01 6.53903410e-02 6.87319279e-01
2.71231413e-01 -3.33867282e-01 5.90687633e-01 2.31268391e-01
1.59419549e+00 1.65019050e-01 2.49731224e-02 -4.41959620e-01
5.07607341e-01 2.13259324e-01 4.66766357e-01 7.53445923e-01
4.62794900e-02 9.85307172e-02 9.84572530e-01 -1.23227119e+00
-9.93504882e-01 -7.31932044e-01 -5.27761504e-02 1.95678818e+00
-6.20555699e-01 -1.04748762e+00 -6.62640572e-01 -6.03890836e-01
2.37418756e-01 6.65484130e-01 -4.70380366e-01 -2.94731315e-02
-7.59634197e-01 -6.04758143e-01 5.95175087e-01 5.61212718e-01
1.49222359e-01 -1.11394584e+00 -9.51421261e-01 3.17490786e-01
3.07392240e-01 -7.35167861e-01 -3.44752848e-01 8.36084008e-01
-1.13024676e+00 -1.10688317e+00 -3.34760517e-01 -6.77705109e-01
6.13505960e-01 -1.21933043e-01 1.81291521e+00 4.65193331e-01
-7.39974976e-01 1.69277743e-01 -1.93471521e-01 -2.01267004e-01
-2.80101240e-01 4.03059512e-01 -3.63527656e-01 -4.61184144e-01
5.28950989e-01 -5.06505609e-01 -3.23275328e-01 -3.05040292e-02
-6.97814465e-01 -3.57880332e-02 5.73133886e-01 7.81863987e-01
1.97358772e-01 7.63545465e-03 -4.94174182e-01 -8.53399336e-01
1.00786269e+00 -5.07143915e-01 -9.06539738e-01 1.43698335e-01
-8.06033075e-01 5.11257648e-01 6.58665895e-01 -1.79316446e-01
-6.85506403e-01 6.12114407e-02 -1.15642734e-01 -5.44289589e-01
1.09343745e-01 8.44991982e-01 4.96699035e-01 -1.35729522e-01
1.07385421e+00 -3.18132825e-02 -5.89839835e-03 -6.00178719e-01
2.17924744e-01 3.34957480e-01 3.15713018e-01 -1.23542678e+00
3.93533677e-01 -3.05402577e-02 -1.68518886e-01 -6.66762054e-01
-6.86410725e-01 -3.06321681e-01 -5.35168707e-01 5.67782903e-03
4.79091436e-01 -5.61002553e-01 -1.05042279e+00 3.45815390e-01
-1.45951164e+00 -5.38860738e-01 -1.74516179e-02 2.28316113e-01
-2.32590571e-01 9.10849273e-02 -9.02620435e-01 -6.14714026e-01
-5.28353691e-01 -1.61598361e+00 8.32030714e-01 -1.79004818e-01
-4.86308396e-01 -9.67956066e-01 2.12830883e-02 2.13117674e-01
7.52547264e-01 -7.61376619e-02 1.23314834e+00 -7.23937809e-01
-7.04617620e-01 -2.61997014e-01 -4.35482681e-01 6.07163787e-01
-4.83851105e-01 4.12238777e-01 -7.53975511e-01 3.62436995e-02
-1.91298738e-01 -4.63244468e-01 7.62108445e-01 1.04500160e-01
1.86884785e+00 -3.15421522e-01 -1.13949135e-01 1.11120892e+00
1.34048975e+00 -1.60990357e-01 4.64492708e-01 5.89149237e-01
7.78820276e-01 7.21669018e-01 2.01484367e-01 6.03347361e-01
3.13418984e-01 3.11214924e-01 6.80880308e-01 2.20391676e-01
2.96305418e-01 2.27955759e-01 2.62163818e-01 7.37524748e-01
-1.78935349e-01 4.13291305e-01 -1.60200155e+00 5.09234071e-01
-1.68547904e+00 -7.42328882e-01 -5.39983869e-01 1.93757582e+00
1.26450908e+00 3.45816970e-01 6.02725372e-02 -4.96253781e-02
8.70759711e-02 1.58941001e-01 -3.84518147e-01 -1.10240066e+00
5.57308793e-01 7.19009399e-01 5.92093110e-01 3.53854239e-01
-9.55347180e-01 7.86933839e-01 6.79426622e+00 4.82832581e-01
-1.33274877e+00 -2.33663581e-02 4.64830697e-01 -2.51288861e-01
-2.35843763e-01 5.93328737e-02 -6.28900707e-01 1.44390211e-01
9.88913476e-01 -2.14961037e-01 8.29815626e-01 1.47987020e+00
-2.73451835e-01 4.67396677e-02 -1.44306910e+00 9.89957869e-01
-4.15482879e-01 -1.43165469e+00 -3.77789944e-01 -4.34070416e-02
2.80161917e-01 4.85905409e-01 -6.03104420e-02 7.10871458e-01
6.71949446e-01 -1.11018920e+00 7.43534923e-01 2.75854856e-01
4.28270578e-01 -4.69121337e-01 5.53874552e-01 2.60218203e-01
-1.00871336e+00 -4.29811358e-01 -3.93867463e-01 -5.04869938e-01
-4.77103084e-01 8.45530450e-01 -7.52527654e-01 -4.76641953e-02
1.09255254e+00 6.11829340e-01 -6.42302573e-01 1.11758339e+00
-5.28645404e-02 4.92736101e-01 -2.11629301e-01 -2.55765945e-01
4.32201147e-01 8.66120979e-02 3.51551205e-01 1.68271184e+00
-5.06137162e-02 -4.90037262e-01 2.30475754e-01 1.24739826e+00
7.60955438e-02 1.89135686e-01 -3.96631122e-01 -3.37640733e-01
2.27533191e-01 1.48544347e+00 -4.92517740e-01 -3.16965461e-01
-5.11903048e-01 4.55223858e-01 6.18718982e-01 1.86573341e-01
-7.18238831e-01 -7.52683818e-01 9.81034338e-01 2.37716019e-01
1.87915817e-01 -5.78248024e-01 -6.42330468e-01 -1.17506969e+00
1.03650153e-01 -1.26023710e+00 3.85407746e-01 -4.73889172e-01
-1.08892226e+00 5.24467826e-01 -5.64343296e-02 -5.73369265e-01
-9.91594046e-02 -1.10387623e+00 -6.28626347e-01 9.29427862e-01
-1.20812654e+00 -4.91236538e-01 -2.66282022e-01 4.35487270e-01
3.60598803e-01 -2.65768290e-01 8.75900447e-01 3.95712674e-01
-5.46130836e-01 4.58856463e-01 -1.95056230e-01 2.74559170e-01
3.90681356e-01 -1.54655862e+00 6.78100586e-01 5.49074054e-01
-9.86235961e-02 1.25213408e+00 7.25952923e-01 -2.98578501e-01
-1.88670194e+00 -8.36333632e-01 1.01343012e+00 -3.82326156e-01
1.14294088e+00 -6.14576876e-01 -9.87179041e-01 7.61603355e-01
1.15127265e-01 4.05479580e-01 5.13325989e-01 6.54381514e-01
-8.65073860e-01 -2.95609087e-01 -7.39863038e-01 5.60741067e-01
1.06709826e+00 -7.09425807e-01 -5.92130661e-01 5.85196495e-01
5.99932671e-01 -6.39927387e-01 -9.06564415e-01 -1.03250518e-02
7.04698503e-01 -1.17151487e+00 1.07665253e+00 -8.92266750e-01
7.91680813e-01 2.33375244e-02 -7.20548108e-02 -1.12081695e+00
-3.75447989e-01 -7.52784729e-01 -2.42654100e-01 6.03550673e-01
4.64473963e-01 -4.63961959e-01 7.50000954e-01 9.99109149e-01
-3.94100398e-01 -9.22876656e-01 -2.93680727e-01 -6.14207745e-01
-4.85572033e-02 -8.21797609e-01 6.63338482e-01 9.78997827e-01
7.66108260e-02 1.24655634e-01 2.30199248e-01 -2.85905033e-01
4.56487775e-01 9.37992111e-02 7.62237549e-01 -1.35137260e+00
-6.67369664e-01 -9.29735899e-01 -3.23413700e-01 -7.98283339e-01
3.07528317e-01 -1.23510337e+00 -9.77044702e-02 -1.15679896e+00
3.23932283e-02 -7.29817927e-01 -1.87000021e-01 9.64007318e-01
1.60418972e-01 -3.36675525e-01 1.38718992e-01 2.38170430e-01
-6.01465046e-01 1.14752807e-01 9.17524040e-01 -2.72576928e-01
-1.57760412e-01 -2.59009033e-01 -6.19966328e-01 8.04647505e-01
4.98226166e-01 -4.53690499e-01 -1.90110564e-01 -1.09568143e+00
8.66959453e-01 -1.00921862e-01 2.40224212e-01 -9.53163087e-01
3.08412731e-01 -1.30461454e-01 1.01501778e-01 -2.97847450e-01
-9.87512544e-02 -6.87272191e-01 -1.53391093e-01 6.54646754e-01
-7.69962788e-01 5.80983639e-01 3.26862961e-01 -5.13036214e-02
-2.04239383e-01 -5.70460439e-01 4.38886225e-01 -5.28034210e-01
-9.24679816e-01 3.06190401e-01 -3.84270251e-01 1.43152475e-01
7.70540357e-01 1.53269442e-02 -3.91612917e-01 1.47389159e-01
-6.61999285e-01 3.48345667e-01 2.06708908e-01 3.41412842e-01
3.92995179e-01 -8.24501634e-01 -4.01574343e-01 3.30528021e-01
1.23427354e-01 1.15756705e-01 -1.83984712e-01 9.84774828e-01
-1.22599912e+00 4.14277673e-01 -2.81104386e-01 -5.06742120e-01
-1.16832209e+00 5.39038777e-01 4.72175747e-01 -5.90614915e-01
-6.16660357e-01 1.25208879e+00 1.97020639e-02 -8.59960139e-01
5.87042332e-01 -8.93068731e-01 2.22692698e-01 -1.10164918e-01
6.86259806e-01 4.18948889e-01 5.65748096e-01 1.74852997e-01
-4.69431192e-01 -2.21915692e-02 -3.96174580e-01 3.06543827e-01
1.62044168e+00 7.37263143e-01 -8.49526703e-01 2.32250839e-01
1.21067488e+00 -2.41075397e-01 -9.16849852e-01 -1.21922083e-01
4.73075837e-01 -5.30475140e-01 4.34979498e-01 -7.61079609e-01
-1.24281895e+00 1.05603600e+00 1.85677662e-01 2.75344908e-01
8.15705776e-01 -3.15809995e-01 6.66119039e-01 8.58353436e-01
3.16231877e-01 -1.08461547e+00 8.78243744e-02 1.15898252e+00
8.29079151e-01 -8.92558336e-01 -3.05642816e-03 -1.95768431e-01
-2.18200833e-01 1.62740564e+00 7.44166493e-01 -2.59264797e-01
7.17680156e-01 7.64891982e-01 -3.11979085e-01 -4.69781160e-01
-1.32327330e+00 1.39707968e-01 2.57401466e-02 3.99234980e-01
9.88550484e-01 -3.63493077e-02 -8.38748962e-02 4.04827148e-01
-3.61170232e-01 2.30473414e-01 2.15748847e-01 1.17706800e+00
-3.34346771e-01 -1.07917702e+00 -4.02489424e-01 8.20613921e-01
-6.36850655e-01 -5.05811274e-01 -6.48291931e-02 4.81306881e-01
7.62523059e-03 4.18835223e-01 2.37906307e-01 -4.92484033e-01
3.56638163e-01 1.42828703e-01 4.61642921e-01 -8.60752821e-01
-1.31392312e+00 -5.89338899e-01 3.05665880e-01 -1.04754126e+00
3.26460809e-01 -5.71828246e-01 -1.16788864e+00 -6.07691646e-01
4.63987365e-02 -1.03978276e-01 8.28800917e-01 7.92036414e-01
5.75923324e-01 5.28430462e-01 9.06119049e-02 -8.36271226e-01
-9.77154553e-01 -6.27647936e-01 -2.91511148e-01 1.46613434e-01
2.46136338e-01 -4.70275253e-01 -2.34350771e-01 -7.40235597e-02] | [8.398140907287598, 3.4931483268737793] |
8ea9ed71-a60f-4bfb-8575-b2314a9a4ccf | blind-estimation-of-room-acoustic-parameters-1 | 2212.13009 | null | https://arxiv.org/abs/2212.13009v1 | https://arxiv.org/pdf/2212.13009v1.pdf | Blind estimation of room acoustic parameters from speech signals based on extended model of room impulse response | The speech transmission index (STI) and room acoustic parameters (RAPs), which are derived from a room impulse response (RIR), such as reverberation time and early decay time, are essential to assess speech transmission and to predict the listening difficulty in a sound field. Since it is difficult to measure RIR in daily occupied spaces, simultaneous blind estimation of STI and RAPs must be resolved as it is an imperative and challenging issue. This paper proposes a deterministic method for blindly estimating STI and five RAPs on the basis of an RIR stochastic model that approximates an unknown RIR. The proposed method formulates a temporal power envelope of a reverberant speech signal to obtain the optimal parameters for the RIR model. Simulations were conducted to evaluate STI and RAPs from observed reverberant speech signals. The root-mean-square errors between the estimated and ground-truth results were used to comparatively evaluate the proposed method with the previous method. The results showed that the proposed method can estimate STI and RAPs effectively without any training. | ['Masashi Unoki', 'Suradej Duangpummet', 'Lijun Wang'] | 2022-12-26 | null | null | null | null | ['room-impulse-response'] | ['audio'] | [ 1.25780210e-01 -9.49755669e-01 1.04916215e+00 -1.76421732e-01
-1.17174220e+00 -4.86931443e-01 1.33110955e-01 -1.74947500e-01
-2.63236165e-01 6.24873698e-01 6.26612186e-01 -4.68628347e-01
-3.58747661e-01 -3.02722275e-01 -5.20966686e-02 -9.72388029e-01
-3.63084018e-01 -3.57142538e-01 3.22541557e-02 -1.03305534e-01
1.57380030e-01 3.90539289e-01 -1.75067019e+00 -1.14939891e-01
1.00617492e+00 1.07336760e+00 6.39174938e-01 1.12289071e+00
2.15864941e-01 5.76662242e-01 -1.14474201e+00 3.41977388e-01
6.40923083e-02 -4.24260467e-01 -8.95117894e-02 -2.25349396e-01
-7.44990334e-02 -3.65098983e-01 -2.54352033e-01 8.64912629e-01
9.17236209e-01 6.50319278e-01 8.81666303e-01 -4.87833947e-01
-1.30129233e-01 1.51095197e-01 -1.45675927e-01 4.76130754e-01
4.97261733e-01 -1.81827813e-01 4.37000543e-01 -6.51417911e-01
-4.86216128e-01 1.00816822e+00 6.83500469e-01 7.05450773e-02
-9.01932716e-01 -7.61500835e-01 -3.90231341e-01 1.32998258e-01
-1.43171751e+00 -6.94888115e-01 9.08312738e-01 -4.12741542e-01
6.16055608e-01 7.24715948e-01 2.80605555e-01 7.28743434e-01
1.30781457e-01 8.64868015e-02 1.48177683e+00 -7.03710973e-01
2.09596157e-01 1.63674831e-01 3.50993901e-01 4.71636921e-01
-1.09099053e-01 5.32236099e-01 -2.11290255e-01 -2.74496347e-01
6.60536647e-01 -3.43198538e-01 -8.82843792e-01 5.14883280e-01
-1.10488725e+00 2.66189367e-01 4.06343788e-01 6.72655106e-01
-2.64748186e-01 -2.04542950e-01 -1.00762300e-01 9.92051512e-02
2.98681140e-01 3.17383856e-01 -1.63321067e-02 -1.57945231e-01
-8.23377788e-01 -1.88901111e-01 8.64159763e-01 4.67514545e-01
3.15150648e-01 5.53638279e-01 -2.28912145e-01 1.32043600e+00
7.19805777e-01 1.19946194e+00 2.13555351e-01 -8.05862010e-01
4.14198965e-01 -5.62396348e-01 5.05251884e-01 -1.14526391e+00
-3.03201377e-01 -6.27013505e-01 -8.33855867e-01 -1.05650969e-01
3.53483349e-01 -2.09141076e-01 -9.03580368e-01 1.27001262e+00
1.89345255e-01 2.54534990e-01 1.48044512e-01 1.04534280e+00
9.09908235e-01 1.11638761e+00 -4.95210886e-01 -6.98627293e-01
1.15247345e+00 -6.97173655e-01 -1.14024091e+00 -4.14414465e-01
-8.51202607e-02 -1.31613708e+00 1.01984298e+00 5.27829766e-01
-8.19923639e-01 -7.86996484e-01 -1.08968520e+00 6.41707599e-01
2.68850803e-01 1.60845369e-01 -3.82542051e-02 1.16133690e+00
-8.37060213e-01 1.74465299e-01 -8.32322598e-01 1.50425211e-01
-5.62914789e-01 -1.56775028e-01 3.64082456e-02 -5.97947426e-02
-1.06431901e+00 6.77488208e-01 -4.02059019e-01 1.00814390e+00
-1.28021979e+00 -6.04183018e-01 -6.48477316e-01 -2.42174476e-01
-2.27389447e-02 -2.11140737e-01 1.39693224e+00 -3.54645550e-01
-1.57692587e+00 2.06901319e-02 -6.07400715e-01 1.64157316e-01
2.06068099e-01 -2.31910810e-01 -1.10829794e+00 -2.15046555e-02
-1.54464215e-01 -8.22751999e-01 1.02408957e+00 -1.76086760e+00
-3.96530703e-02 -1.10075146e-01 -5.21309018e-01 1.38795033e-01
-7.65585825e-02 8.70796368e-02 5.52094541e-02 -6.08844817e-01
6.09128714e-01 -4.52676862e-01 -1.74650162e-01 -5.91310263e-01
-5.62998116e-01 3.80617917e-01 1.93357125e-01 -1.47825253e+00
1.42130339e+00 -2.28894639e+00 -4.05501187e-01 5.23023784e-01
-6.51984811e-02 2.51023144e-01 1.85595527e-02 4.43722785e-01
-1.41218752e-02 -3.69980246e-01 -4.90159005e-01 -2.97365576e-01
-2.29158729e-01 -3.76100928e-01 -3.62002164e-01 7.46335447e-01
-5.78081846e-01 -2.08503470e-01 -7.38907814e-01 -2.10563123e-01
3.49377364e-01 8.39792430e-01 -6.08699434e-02 8.51641297e-01
5.49604058e-01 8.38750064e-01 -4.63254541e-01 5.21264434e-01
8.59112144e-01 6.22696757e-01 -3.19582939e-01 -3.42217445e-01
-4.61548775e-01 2.85671353e-01 -1.28955913e+00 8.48345697e-01
-1.10823047e+00 7.30629086e-01 5.72790027e-01 -5.40372729e-01
1.55012703e+00 5.77879131e-01 -1.52169198e-01 -8.14874053e-01
1.47075802e-01 5.76202154e-01 7.23663643e-02 -7.89371550e-01
3.43679875e-01 -3.76667142e-01 3.87390196e-01 4.09619153e-01
-4.18380320e-01 -3.35605323e-01 -4.54407334e-01 -3.45692813e-01
9.98069942e-01 -3.62511784e-01 -1.16100051e-02 -2.17240244e-01
9.43829954e-01 -7.32398391e-01 3.40764254e-01 5.36073804e-01
-2.17741743e-01 6.69156671e-01 -3.26476872e-01 9.14435275e-03
-7.05835521e-01 -1.30289972e+00 -2.18702972e-01 5.79107523e-01
-3.40738036e-02 1.63226932e-01 -7.29571402e-01 1.73398405e-01
-4.93368596e-01 1.18593097e+00 -1.61440894e-01 8.21382105e-02
-6.32771373e-01 -5.35393953e-01 3.30123127e-01 8.02965313e-02
4.21789020e-01 -8.89071047e-01 -1.88513935e-01 2.93116927e-01
-6.74076915e-01 -9.31230724e-01 -5.22129238e-01 1.29681006e-01
-4.10369366e-01 -6.30655229e-01 -8.46455753e-01 -7.19649494e-01
7.85469890e-01 7.51840949e-01 6.48853660e-01 1.27805658e-02
-4.85950798e-01 6.05983555e-01 -3.93429458e-01 -4.87668544e-01
-6.55427516e-01 -8.12288940e-01 3.21375072e-01 2.08440617e-01
-3.76562923e-01 -6.65270567e-01 -7.77587116e-01 9.99825835e-01
-5.21617770e-01 -4.10844982e-01 1.01097666e-01 3.27718377e-01
2.84509301e-01 5.82372546e-01 4.78791505e-01 4.68395092e-02
9.70224380e-01 -6.83225244e-02 -7.90363014e-01 3.03746432e-01
-2.64519960e-01 -1.66426972e-01 6.84635997e-01 -3.29892367e-01
-1.70470667e+00 -5.24382293e-01 -2.49522746e-01 -4.89132851e-02
-1.77260756e-01 3.33962440e-01 -2.07912877e-01 4.74281050e-02
8.17975998e-01 4.67347980e-01 -2.20808581e-01 -5.98293602e-01
-1.59149081e-01 1.28479159e+00 7.20861614e-01 -4.54427958e-01
9.86594200e-01 1.06358059e-01 -2.45676741e-01 -1.47484267e+00
-6.31644309e-01 -7.52306163e-01 -9.47695151e-02 -4.72020954e-01
7.02090859e-01 -8.04333031e-01 -7.98985064e-01 7.21374929e-01
-1.28697085e+00 -3.00711364e-01 3.65495503e-01 1.24318027e+00
-3.56096148e-01 4.82327610e-01 -4.61681753e-01 -1.76036084e+00
-4.08045888e-01 -1.06618810e+00 3.82658064e-01 1.08596258e-01
-1.92047328e-01 -9.89533901e-01 2.13060394e-01 7.34783053e-01
7.59440124e-01 4.18670811e-02 6.13527954e-01 7.28756562e-02
-2.66118437e-01 -4.19716060e-01 1.76903099e-01 6.70428038e-01
6.00951076e-01 -1.22471385e-01 -1.51425183e+00 -3.34919035e-01
9.14450645e-01 1.84697181e-01 3.54486525e-01 9.07704890e-01
8.91609609e-01 -3.23314548e-01 8.11326355e-02 5.85657775e-01
1.45359838e+00 7.24518120e-01 8.93795192e-01 2.08879977e-01
3.72716337e-01 6.65526867e-01 6.93937540e-01 6.06326699e-01
-1.50370672e-01 3.33949834e-01 4.16339844e-01 3.00857257e-02
-1.55474111e-01 4.91512008e-03 4.88760769e-01 1.43242681e+00
-5.16221046e-01 -4.52435851e-01 -9.09676492e-01 4.37020093e-01
-8.31485569e-01 -9.60908413e-01 -4.08986300e-01 2.75157571e+00
6.46880984e-01 -3.04681480e-01 -3.91464770e-01 5.92597544e-01
7.23501503e-01 3.22357357e-01 -1.62955131e-02 -4.72777665e-01
2.11862192e-01 2.07858160e-01 4.23137397e-01 1.12649107e+00
-6.30013049e-01 7.79404864e-03 6.51699257e+00 3.97220284e-01
-9.38883305e-01 -1.34158477e-01 3.34712863e-01 3.69800121e-01
-5.22089660e-01 -3.96126300e-01 -4.57459033e-01 1.57197297e-01
1.09009659e+00 -8.61879513e-02 1.04804909e+00 2.10712269e-01
7.51234293e-01 -2.67059892e-01 -7.23194957e-01 1.03189218e+00
4.79036532e-02 -2.99915969e-01 -5.64819992e-01 -1.46426231e-01
4.31021452e-01 -4.40112263e-01 2.66995460e-01 -2.29720250e-01
2.78250594e-02 -1.19212878e+00 5.74077845e-01 8.81444931e-01
5.23900449e-01 -6.06160223e-01 7.80098200e-01 4.30343628e-01
-1.28885877e+00 -2.68558580e-02 -2.31408462e-01 1.28945336e-01
3.63034695e-01 8.11653912e-01 -1.12147725e+00 4.94480968e-01
5.68776608e-01 -5.59110492e-02 -9.29005519e-02 1.47672176e+00
-3.37895840e-01 1.14881504e+00 -2.98585951e-01 -1.72812328e-01
-1.44653603e-01 -3.87744397e-01 8.10428381e-01 1.05165851e+00
8.51591229e-01 2.45622307e-01 -4.67052042e-01 8.68460238e-01
4.90058869e-01 3.98788415e-02 -1.19414516e-01 3.26278731e-02
7.63212323e-01 1.05107021e+00 -3.92902970e-01 3.68885368e-01
-7.45515078e-02 6.78618252e-01 -7.37779260e-01 9.75563645e-01
-8.46459329e-01 -7.08206117e-01 4.31386441e-01 1.50784880e-01
1.24855541e-01 -6.24833822e-01 5.23874871e-02 -3.08246106e-01
1.13939844e-01 -8.28937173e-01 -2.51154512e-01 -1.01659131e+00
-1.05539954e+00 9.93452191e-01 -1.93944991e-01 -1.56828845e+00
2.65216678e-02 -3.77327144e-01 -9.34532762e-01 1.58942580e+00
-1.37417614e+00 -4.99628514e-01 -5.25139034e-01 7.02476025e-01
4.17682379e-01 2.88864058e-02 1.12303889e+00 3.56901407e-01
-5.09179354e-01 4.71163899e-01 5.75665414e-01 7.92309735e-03
5.12791991e-01 -8.00786853e-01 1.21847108e-01 8.93631756e-01
-3.51008326e-01 7.57078767e-01 1.29477692e+00 -3.05073917e-01
-1.20686042e+00 -6.84268594e-01 5.03559232e-01 -9.62597132e-02
5.32274723e-01 -1.84845164e-01 -9.52251911e-01 1.17485330e-01
-1.12636007e-01 -6.04248047e-02 8.17996144e-01 -6.30074069e-02
-1.23965964e-01 -4.65032071e-01 -1.10731483e+00 2.42589429e-01
4.00330156e-01 -6.18745089e-01 -5.56618392e-01 -2.32979953e-02
6.78028524e-01 -2.35343188e-01 -5.68834484e-01 2.92071160e-02
4.90695924e-01 -1.10152352e+00 1.24381435e+00 4.81770724e-01
-1.32872671e-01 -6.36450350e-01 -6.74913168e-01 -1.57152033e+00
-1.63153246e-01 -7.13500500e-01 2.44384870e-01 1.48562098e+00
4.44675416e-01 -8.07958245e-01 -7.93078914e-03 4.51070517e-01
-3.12682271e-01 -3.68996076e-02 -9.09225166e-01 -1.00097406e+00
-6.57158494e-01 -6.55975223e-01 3.84852320e-01 3.59144360e-01
-5.05309224e-01 2.55108595e-01 -4.08420056e-01 9.47944045e-01
1.17032182e+00 -2.86333084e-01 5.35526276e-01 -1.00691485e+00
-4.00810003e-01 1.80751964e-01 3.44094895e-02 -8.79337788e-01
-2.61576831e-01 -1.94067433e-01 8.98825943e-01 -1.98752594e+00
-4.05208379e-01 -7.21521676e-01 -4.50406224e-01 -3.51913035e-01
-3.16265851e-01 -3.42358112e-01 -1.67851880e-01 4.76400414e-03
3.40573579e-01 7.20040619e-01 1.32608473e+00 -6.35443255e-02
-4.90524858e-01 7.88817227e-01 -1.80763155e-01 6.74199760e-01
5.65043986e-01 -3.56212765e-01 -6.44738257e-01 -5.56802690e-01
-1.70288563e-01 6.37727976e-01 3.03100824e-01 -1.37271786e+00
1.70109347e-01 3.06959711e-02 1.34361625e-01 -6.98475301e-01
5.83244860e-01 -1.00512755e+00 2.34592527e-01 2.61177272e-01
-1.27171680e-01 -4.18029487e-01 2.83035845e-01 6.27679944e-01
-2.41082579e-01 -1.09816521e-01 6.92068875e-01 1.31889507e-01
-8.25970545e-02 -2.60533363e-01 -4.93113458e-01 -4.04732943e-01
4.40879345e-01 -2.24777758e-01 -2.69825816e-01 -8.04522574e-01
-3.53675723e-01 -3.79418015e-01 -2.89027929e-01 -1.54201705e-02
9.09889877e-01 -1.02263749e+00 -8.34619224e-01 2.40070328e-01
-2.03408360e-01 -8.56618360e-02 7.19569027e-01 6.44233286e-01
-6.70813918e-01 3.28459561e-01 2.44925633e-01 -3.96597713e-01
-1.45486498e+00 4.30563420e-01 7.49661565e-01 1.50245786e-01
-2.01981217e-01 1.00476205e+00 2.73591638e-01 -3.66343826e-01
4.43903297e-01 -2.80447274e-01 -3.66390020e-01 -3.88531536e-01
9.37233508e-01 7.96500742e-01 2.35728920e-01 -7.24759519e-01
-3.65842879e-01 7.46979415e-01 5.49868107e-01 -5.25515914e-01
1.04110587e+00 -6.65122032e-01 -2.63832837e-01 7.78659284e-01
1.26508772e+00 6.16826892e-01 -8.95700395e-01 -1.84311330e-01
-4.89332646e-01 -7.08419085e-01 5.28057218e-01 -9.81948555e-01
-5.62565506e-01 9.42292690e-01 1.06483781e+00 3.59567344e-01
1.44801712e+00 -2.56628990e-01 6.98427200e-01 1.82225004e-01
5.83472788e-01 -8.05915356e-01 2.29329303e-01 3.26612204e-01
1.08557928e+00 -7.78956294e-01 -1.64037675e-01 -4.95964050e-01
-1.73377931e-01 1.05720425e+00 2.40931198e-01 4.71972257e-01
8.23874414e-01 3.22443247e-01 5.99779606e-01 4.19978678e-01
3.28642242e-02 1.12300634e-01 3.13869447e-01 7.52061307e-01
7.00036049e-01 2.26923689e-01 1.75050590e-02 5.13266504e-01
-5.53782344e-01 -4.52879816e-01 3.92213911e-01 6.62094831e-01
-8.70187759e-01 -6.44938707e-01 -1.47569144e+00 1.45276010e-01
-5.16078770e-01 -1.85066015e-01 3.58532101e-01 -1.35058329e-01
-4.06442165e-01 1.83435261e+00 -1.58874810e-01 -5.93759537e-01
7.37372458e-01 -2.77427733e-01 3.47878814e-01 -1.14959158e-01
-1.82533115e-01 4.27506924e-01 2.50855058e-01 -5.27133495e-02
-3.65052402e-01 -3.91616881e-01 -1.16931880e+00 -2.73728877e-01
-5.62198579e-01 5.68698704e-01 1.18736577e+00 6.38817430e-01
-4.02392834e-01 1.00224805e+00 1.39648426e+00 -7.31339157e-01
-4.71317351e-01 -9.64309454e-01 -1.12013829e+00 -2.34292656e-01
1.00743198e+00 -3.34241778e-01 -1.13070655e+00 -3.76141742e-02] | [15.136185646057129, 5.765960693359375] |
a2c95bb8-1896-49a7-959b-eb01adae402e | combining-representation-learning-with-logic | 1712.09687 | null | http://arxiv.org/abs/1712.09687v1 | http://arxiv.org/pdf/1712.09687v1.pdf | Combining Representation Learning with Logic for Language Processing | The current state-of-the-art in many natural language processing and
automated knowledge base completion tasks is held by representation learning
methods which learn distributed vector representations of symbols via
gradient-based optimization. They require little or no hand-crafted features,
thus avoiding the need for most preprocessing steps and task-specific
assumptions. However, in many cases representation learning requires a large
amount of annotated training data to generalize well to unseen data. Such
labeled training data is provided by human annotators who often use formal
logic as the language for specifying annotations. This thesis investigates
different combinations of representation learning methods with logic for
reducing the need for annotated training data, and for improving
generalization. | ['Tim Rocktäschel'] | 2017-12-27 | null | null | null | null | ['formal-logic'] | ['reasoning'] | [ 1.94804475e-01 4.68366027e-01 -6.39709592e-01 -8.43548179e-01
-5.87218821e-01 -6.02077186e-01 7.00990140e-01 6.61300600e-01
-4.56094563e-01 1.10412943e+00 1.08560938e-02 -6.81028187e-01
-2.06399381e-01 -8.84549975e-01 -6.09276891e-01 -2.54740089e-01
5.25144227e-02 8.46779346e-01 -5.88277839e-02 -3.83521199e-01
1.22151606e-01 5.22096694e-01 -1.61517680e+00 5.59398472e-01
7.58799076e-01 9.65785921e-01 1.55536264e-01 3.40411723e-01
-8.08839142e-01 1.20293260e+00 -6.14350677e-01 -3.32006097e-01
1.56440631e-01 -2.40021005e-01 -8.96578252e-01 3.67696537e-03
3.53104830e-01 1.67784262e-02 -1.40768677e-01 1.00568879e+00
6.11213483e-02 4.08032387e-01 7.61723876e-01 -1.09379900e+00
-8.98109257e-01 7.52645910e-01 -6.32096902e-02 5.73222190e-02
6.15973353e-01 -3.58533934e-02 9.96364713e-01 -8.86695325e-01
7.82703042e-01 1.23932064e+00 6.02979720e-01 6.39711082e-01
-1.17824268e+00 -3.16860825e-01 1.15916118e-01 1.43092945e-01
-1.47131968e+00 -3.50657672e-01 8.25809479e-01 -5.76460004e-01
1.14501238e+00 1.28497601e-01 3.35514605e-01 7.41259933e-01
-3.32692891e-01 7.40777731e-01 7.87162066e-01 -9.60116327e-01
4.86152440e-01 7.73853183e-01 4.59221065e-01 9.57429290e-01
4.76778805e-01 -1.49611726e-01 -4.11112309e-01 -3.60208780e-01
6.17193401e-01 6.49649873e-02 -1.64056018e-01 -5.00979424e-01
-8.11598718e-01 9.90005612e-01 2.05219239e-01 4.82969582e-01
-3.98381621e-01 1.57392919e-01 7.66136467e-01 5.97431242e-01
4.17903453e-01 7.64990091e-01 -7.59003878e-01 -7.83422366e-02
-1.00019872e+00 1.51474550e-01 1.01179373e+00 1.02097201e+00
1.25220573e+00 3.48581463e-01 5.86240105e-02 7.02289581e-01
3.54517817e-01 3.40848386e-01 8.39035392e-01 -8.27560842e-01
5.17454505e-01 1.12530923e+00 1.23873316e-01 -9.08292234e-01
-1.06522776e-01 -3.01010787e-01 -4.05807823e-01 1.90224662e-01
3.21368814e-01 -2.11471334e-01 -8.93222809e-01 1.47914529e+00
1.37457056e-02 -1.15807526e-01 4.12123531e-01 4.94149417e-01
5.12969971e-01 6.68107033e-01 1.48121744e-01 -2.83222884e-01
9.72441196e-01 -5.20625293e-01 -6.56076610e-01 -3.84749293e-01
1.08687282e+00 -5.72444759e-02 1.14255357e+00 4.22687262e-01
-8.64482880e-01 -3.63744587e-01 -9.40736175e-01 -1.13404252e-01
-8.59189153e-01 9.32527781e-02 1.23565698e+00 8.06893051e-01
-7.18809783e-01 6.27720714e-01 -7.19997883e-01 -2.01143950e-01
5.50268829e-01 5.92262864e-01 -6.87678099e-01 -4.73490149e-01
-1.20001221e+00 1.21057844e+00 8.04895580e-01 1.43883914e-01
-7.41713166e-01 -6.45420074e-01 -1.44945168e+00 7.48062879e-02
4.71549600e-01 -3.94758344e-01 1.19742513e+00 -1.17658937e+00
-1.26888227e+00 7.96566784e-01 -2.63367265e-01 -5.58509946e-01
2.32311934e-01 -3.37156266e-01 -1.96738407e-01 -7.74718970e-02
-1.19254455e-01 3.41779768e-01 7.37497985e-01 -1.27289748e+00
-2.36034140e-01 -3.12322408e-01 9.17755067e-02 -4.50574718e-02
-4.60127681e-01 -1.86641857e-01 -1.49479404e-03 -3.10911000e-01
-1.57728679e-02 -5.73609173e-01 -3.53690296e-01 -1.77806959e-01
-8.14293418e-03 -6.77966833e-01 8.98600399e-01 -4.37237203e-01
1.13803697e+00 -1.95812786e+00 5.28845042e-02 6.51970088e-01
-1.03211135e-01 4.65718836e-01 -8.36631879e-02 3.54782075e-01
-9.65593159e-02 2.29259297e-01 -7.25584552e-02 -1.70486510e-01
1.54457375e-01 6.24854803e-01 -6.50231302e-01 3.31428409e-01
1.85501367e-01 6.18458629e-01 -1.16574955e+00 -4.92308289e-01
4.63903278e-01 2.10256472e-01 -4.63426143e-01 2.38219485e-01
-6.69739783e-01 1.64925650e-01 -6.16136074e-01 4.53523695e-01
-6.55653700e-03 -2.04632580e-01 3.22394103e-01 2.98498958e-01
1.73698008e-01 4.55000073e-01 -1.41049302e+00 1.99026644e+00
-7.89260745e-01 7.23649740e-01 -1.67791888e-01 -1.59302890e+00
1.27241063e+00 4.83292639e-01 2.41926044e-01 -1.26649737e-01
2.77095735e-01 2.86145627e-01 -2.78432965e-01 -6.51524365e-01
2.95983434e-01 -3.18323433e-01 -1.98672920e-01 4.52154845e-01
2.20360756e-01 -2.63492376e-01 2.58192122e-01 1.48971990e-01
1.04433465e+00 1.82037070e-01 3.70774090e-01 -1.49546355e-01
6.37955070e-01 4.17980403e-01 7.68987894e-01 7.12483346e-01
4.01502907e-01 2.39503160e-02 2.88721800e-01 -8.21529686e-01
-9.49778736e-01 -4.85246539e-01 -1.04640707e-01 1.42964375e+00
-3.58040482e-01 -5.82542896e-01 -5.42579174e-01 -6.18001282e-01
1.43735871e-01 1.21494794e+00 -5.47776937e-01 -2.07667500e-01
-3.90296221e-01 -4.73307036e-02 5.38464069e-01 8.56005847e-01
5.89942932e-02 -1.29255569e+00 -6.28470004e-01 2.59376079e-01
3.48758578e-01 -8.14248621e-01 2.01922789e-01 5.61336100e-01
-8.97510529e-01 -8.82654428e-01 -1.95203945e-01 -6.39839113e-01
9.13103282e-01 -3.29307765e-01 1.04039407e+00 1.48807496e-01
-3.99330407e-01 4.95581359e-01 -3.46567810e-01 -7.58948743e-01
-2.38813326e-01 -1.42013840e-02 7.83013925e-02 -1.84104607e-01
7.40243256e-01 -3.87939841e-01 3.09861302e-01 -2.34384686e-02
-7.98551321e-01 -2.99823076e-01 4.37671572e-01 9.38609123e-01
4.75260258e-01 1.54087037e-01 4.97170001e-01 -1.14289606e+00
6.96739733e-01 -3.95874768e-01 -6.55836403e-01 5.15372634e-01
-4.38068688e-01 6.44294322e-01 7.36582756e-01 -6.22622550e-01
-1.10252881e+00 2.40123585e-01 3.77534986e-01 -5.74920118e-01
-4.03518945e-01 1.01490700e+00 -1.90602854e-01 8.96951184e-02
1.10028923e+00 -1.94217488e-02 -9.93695483e-02 -3.98947299e-01
6.26015961e-01 7.42578089e-01 4.38652277e-01 -1.01483977e+00
6.68776333e-01 9.11465511e-02 -4.70631756e-03 -9.94228601e-01
-1.03183675e+00 -2.08580896e-01 -8.67321610e-01 2.79977351e-01
6.22921646e-01 -6.67716503e-01 -2.64185369e-01 -3.10631633e-01
-1.10906053e+00 -4.76937622e-01 -8.05920541e-01 5.26469111e-01
-6.66751146e-01 4.83571291e-02 -1.37663692e-01 -1.01011419e+00
-2.33573809e-01 -8.05985749e-01 6.08549476e-01 3.94439772e-02
-6.24392748e-01 -1.26687491e+00 5.00417612e-02 1.88967213e-01
4.45316285e-01 2.20730126e-01 1.15498388e+00 -1.14851952e+00
-3.79018039e-01 -8.14446688e-01 1.78680956e-01 5.22803664e-01
8.43584016e-02 -3.41386706e-01 -1.03844392e+00 -7.83052146e-02
-2.38800406e-01 -8.69160533e-01 6.63182974e-01 -3.25373234e-03
1.28225088e+00 -3.29445958e-01 -3.13400924e-01 3.81314456e-01
1.26140058e+00 5.25761535e-03 3.71956229e-01 1.04542464e-01
5.86428106e-01 8.11713755e-01 5.57967246e-01 4.10026759e-01
7.31895640e-02 2.93005019e-01 3.03724874e-02 1.94533557e-01
3.38916361e-01 -4.07664627e-01 4.74198014e-02 1.03001378e-01
-1.86222970e-01 -1.48117112e-03 -1.22388542e+00 6.33809328e-01
-2.07815957e+00 -1.05040908e+00 4.05164480e-01 2.09645033e+00
1.08731818e+00 5.85642643e-02 -2.56291419e-01 4.20983106e-01
5.45032144e-01 -2.37285137e-01 -5.04157841e-01 -5.26804209e-01
3.32151234e-01 3.90032649e-01 4.68614221e-01 6.75942183e-01
-7.86218226e-01 1.16166723e+00 6.63316011e+00 4.85102803e-01
-1.01036465e+00 -1.87305257e-01 8.73967186e-02 2.25476235e-01
-5.43028891e-01 2.24322498e-01 -7.29345024e-01 -1.29424065e-01
1.07062328e+00 -1.98967367e-01 6.34835184e-01 1.38726759e+00
-2.57365834e-02 4.19342518e-02 -1.48236907e+00 1.09933376e+00
2.40914449e-01 -1.42470801e+00 1.79939091e-01 -1.49532229e-01
5.58875442e-01 -2.22717151e-01 -3.75524729e-01 5.70338368e-01
6.44926727e-01 -1.38952470e+00 3.46276700e-01 6.49732649e-01
7.69007921e-01 -6.38465703e-01 8.47687781e-01 5.19028366e-01
-8.04568946e-01 -3.96968096e-01 -7.23854005e-01 -3.23053032e-01
-1.86563537e-01 5.65410078e-01 -8.37798715e-01 2.44303733e-01
1.99522465e-01 4.32603419e-01 -4.83018875e-01 7.06061542e-01
-5.37737310e-01 5.74013472e-01 -4.92317021e-01 -2.47316182e-01
2.57967442e-01 9.96366963e-02 1.58648416e-01 1.22866249e+00
2.16929633e-02 6.79023936e-02 7.15522051e-01 7.36928999e-01
-8.52266103e-02 1.18925303e-01 -1.10587478e+00 -3.59794825e-01
5.37959218e-01 9.07246947e-01 -4.05161440e-01 -5.43538809e-01
-2.84369856e-01 5.32186210e-01 4.66074109e-01 5.11846542e-01
-1.11914948e-01 -6.76738262e-01 3.18349928e-01 2.51340955e-01
8.81216079e-02 -2.91283488e-01 -4.46946919e-01 -1.08456874e+00
-1.41266957e-01 -8.28141510e-01 4.26416069e-01 -6.87613010e-01
-1.27147508e+00 2.19340056e-01 2.90365070e-01 -6.53812528e-01
-9.09891367e-01 -9.88903165e-01 -6.30124569e-01 8.57233405e-01
-1.20451963e+00 -1.13849235e+00 -1.11720778e-01 8.86082053e-01
3.98840159e-01 -7.28219688e-01 1.20729196e+00 -1.32494614e-01
-7.22527280e-02 4.34260190e-01 -2.48593971e-01 3.77644658e-01
5.18235683e-01 -1.37241662e+00 -4.00343359e-01 2.67915606e-01
2.77090132e-01 1.02630305e+00 7.25358725e-01 -4.41658407e-01
-1.67653167e+00 -8.70662093e-01 1.17355943e+00 -5.95835686e-01
5.27028263e-01 -2.19781861e-01 -1.16633046e+00 1.02345467e+00
-1.21787354e-01 5.04574478e-01 9.92349327e-01 6.09579742e-01
-5.93911767e-01 -1.94674075e-01 -1.20699966e+00 1.46911025e-01
5.33274531e-01 -8.71654451e-01 -1.11499429e+00 4.87590939e-01
3.23493034e-01 -1.52226880e-01 -7.84126222e-01 9.60595384e-02
1.50729805e-01 -8.64807516e-02 6.16790295e-01 -1.18993115e+00
1.76060349e-02 -4.87312376e-01 -3.69193375e-01 -1.02899218e+00
-6.56243414e-02 -5.01803756e-01 -1.45548210e-01 1.02844155e+00
7.37168312e-01 -2.85396159e-01 8.71442556e-01 1.25943816e+00
-1.14424638e-01 -5.87170243e-01 -6.48548007e-01 -8.02772701e-01
1.99800748e-02 -7.22459018e-01 4.67073977e-01 1.41885149e+00
5.68794549e-01 5.82804978e-01 3.69084664e-02 -1.29424959e-01
3.40604365e-01 -3.51071358e-02 8.04599762e-01 -1.45894325e+00
-1.09783590e-01 -6.62703738e-02 -8.27870071e-01 -5.19951582e-01
7.70288289e-01 -1.21677768e+00 -2.11911146e-02 -1.77066171e+00
-3.58637691e-01 -4.71252650e-01 -2.06670165e-01 1.13245785e+00
3.24265301e-01 -2.77113855e-01 -1.05193965e-01 -9.88358185e-02
-6.98214352e-01 3.51134628e-01 6.56935513e-01 -3.06494832e-01
-5.24288416e-01 -1.32800251e-01 -6.65979505e-01 1.04557383e+00
9.03178871e-01 -5.26823163e-01 -9.02493000e-01 -5.18197298e-01
4.47219193e-01 -1.36219889e-01 2.60712057e-01 -7.81079352e-01
3.60476643e-01 -5.08790314e-01 5.92521489e-01 -9.41488147e-02
2.46328250e-01 -1.01008856e+00 -2.33276725e-01 1.42843872e-01
-7.60719597e-01 3.66560370e-02 1.54441297e-01 5.82310736e-01
-3.03630471e-01 -8.99149299e-01 4.57409978e-01 -4.37917113e-01
-9.40033495e-01 -9.07017067e-02 -3.33561867e-01 1.14164658e-01
1.06684661e+00 -2.15935752e-01 -4.44074348e-02 -4.11534488e-01
-9.11320865e-01 1.70344755e-01 3.21173251e-01 1.18077882e-01
6.11681938e-01 -1.03867090e+00 -4.68866378e-01 3.15615475e-01
1.31472319e-01 2.95312494e-01 -3.04558486e-01 2.11546440e-02
-3.33658874e-01 5.15491605e-01 -1.23997904e-01 -2.32749715e-01
-8.34859610e-01 6.76148355e-01 1.15035579e-01 -2.04127714e-01
-6.13668084e-01 7.35423267e-01 -5.50597668e-01 -6.32910609e-01
4.09075469e-01 -1.77915111e-01 -2.31772065e-01 6.40601758e-03
6.27815962e-01 1.49680853e-01 7.74045661e-02 -1.42933980e-01
-2.80935973e-01 1.28572375e-01 -2.46384412e-01 -2.22513169e-01
1.36929452e+00 5.10333657e-01 -4.76708002e-02 7.23189950e-01
9.30172265e-01 -2.86706388e-01 -6.87332571e-01 -3.97558331e-01
5.08690953e-01 -4.45603043e-01 1.94378719e-01 -6.64664626e-01
-6.20017409e-01 1.02662051e+00 4.16447729e-01 -2.87657715e-02
6.32450163e-01 1.11236595e-01 1.20114416e-01 1.51764548e+00
5.49887240e-01 -1.21691191e+00 -1.06464468e-01 6.91971600e-01
7.29348600e-01 -1.16944849e+00 2.92019933e-01 -2.16949999e-01
-7.58083522e-01 1.30646396e+00 4.77535218e-01 -2.12522149e-01
7.43358493e-01 2.47145832e-01 7.36568943e-02 -2.84606427e-01
-7.69336998e-01 5.06769307e-02 2.28385761e-01 9.84023929e-01
7.07905233e-01 -8.31188727e-03 -9.60888341e-02 5.62128782e-01
-2.81248719e-01 1.63986459e-01 2.31190056e-01 1.39133012e+00
-7.06475973e-01 -1.28149545e+00 -2.11339220e-01 5.80823183e-01
-3.24552059e-01 -6.19650856e-02 -3.68339777e-01 7.31436610e-01
3.30959372e-02 8.28085601e-01 4.15528342e-02 -1.58210825e-02
9.86864716e-02 7.29813039e-01 4.01462942e-01 -1.29121721e+00
-4.57598001e-01 -6.42064571e-01 2.22751722e-01 -2.93577313e-01
-2.13158116e-01 -4.17252272e-01 -1.50562167e+00 6.47187978e-02
-4.43498969e-01 5.28536201e-01 8.28121424e-01 1.29867697e+00
2.19396189e-01 2.73857802e-01 9.57428128e-04 -7.53136218e-01
-7.82618821e-01 -9.93246496e-01 -4.71471965e-01 4.59289789e-01
2.66520172e-01 -6.19914711e-01 -7.95079842e-02 3.42545718e-01] | [9.279653549194336, 7.482687950134277] |
2d92ccc5-533f-4f9a-a807-ec640560b72e | survey-on-various-gesture-recognition | 1012.00084 | null | http://arxiv.org/abs/1012.0084v1 | http://arxiv.org/pdf/1012.0084v1.pdf | Survey on Various Gesture Recognition Techniques for Interfacing Machines Based on Ambient Intelligence | Gesture recognition is mainly apprehensive on analyzing the functionality of
human wits. The main goal of gesture recognition is to create a system which
can recognize specific human gestures and use them to convey information or for
device control. Hand gestures provide a separate complementary modality to
speech for expressing ones ideas. Information associated with hand gestures in
a conversation is degree,discourse structure, spatial and temporal structure.
The approaches present can be mainly divided into Data-Glove Based and Vision
Based approaches. An important face feature point is the nose tip. Since nose
is the highest protruding point from the face. Besides that, it is not affected
by facial expressions.Another important function of the nose is that it is able
to indicate the head pose. Knowledge of the nose location will enable us to
align an unknown 3D face with those in a face database. Eye detection is
divided into eye position detection and eye contour detection. Existing works
in eye detection can be classified into two major categories: traditional
image-based passive approaches and the active IR based approaches. The former
uses intensity and shape of eyes for detection and the latter works on the
assumption that eyes have a reflection under near IR illumination and produce
bright/dark pupil effect. The traditional methods can be broadly classified
into three categories: template based methods,appearance based methods and
feature based methods. The purpose of this paper is to compare various human
Gesture recognition systems for interfacing machines directly to human wits
without any corporeal media in an ambient environment. | ['Naveen Lakshmikhanth', 'Karthik R. Shastry', 'Manoj Ravindran', 'Harshith C', 'M. V. V. N. S. Srikanth'] | 2010-12-01 | null | null | null | null | ['contour-detection'] | ['computer-vision'] | [-8.41368064e-02 -1.53042197e-01 -2.24409327e-01 -2.79043198e-01
2.93850064e-01 -6.03203416e-01 5.67019224e-01 -4.36019629e-01
-4.63002056e-01 2.17450336e-01 2.04245389e-01 1.15147326e-02
3.08122896e-02 -4.49079275e-01 1.53837025e-01 -9.73328531e-01
5.20996034e-01 1.78708807e-01 1.86592460e-01 -2.49069810e-01
4.99081671e-01 1.06257927e+00 -1.74232578e+00 -1.03954755e-01
3.41031328e-02 8.12021554e-01 2.91757304e-02 7.18594730e-01
-5.03584087e-01 3.34385246e-01 -6.75565898e-01 5.47788292e-02
1.83509111e-01 -5.08449554e-01 -3.70392889e-01 2.24849671e-01
-2.43161574e-01 -6.73471570e-01 1.36162583e-02 7.63470054e-01
7.59397566e-01 1.94496840e-01 9.08685684e-01 -1.21161163e+00
-8.67216513e-02 -1.29826725e-01 -7.73336291e-01 1.37004018e-01
8.39240968e-01 1.02905020e-01 1.26276240e-01 -8.63824546e-01
4.97511178e-01 1.28265369e+00 2.99661636e-01 9.59837615e-01
-5.60192108e-01 -7.00136721e-01 -1.12390131e-01 1.92279682e-01
-1.54353213e+00 -7.36623168e-01 8.93243074e-01 -4.10842746e-01
8.61925602e-01 5.87925017e-01 5.31456649e-01 7.19301462e-01
-1.08567640e-01 5.04048467e-01 1.19566202e+00 -9.71698523e-01
6.21861890e-02 4.09460157e-01 4.17266577e-01 7.59456336e-01
-1.39454640e-02 1.25577524e-01 -5.83382249e-01 -4.55695465e-02
8.50590646e-01 1.93303645e-01 -4.06871825e-01 3.92663442e-02
-8.01906645e-01 3.30192834e-01 1.32576764e-01 6.29422426e-01
-3.30042213e-01 -2.00108916e-01 -3.72809842e-02 8.99053514e-02
-1.03240170e-01 -1.05316356e-01 -1.39648646e-01 -3.29674423e-01
-6.82313144e-01 -1.65810406e-01 9.74327445e-01 6.89161599e-01
2.10192382e-01 -1.44069090e-01 1.38673499e-01 4.97391969e-01
1.13760889e+00 7.30518699e-01 6.43115759e-01 -2.44474545e-01
1.50177717e-01 8.46902609e-01 1.20390043e-01 -6.69813812e-01
-4.85386074e-01 5.68390787e-01 -1.02357768e-01 8.80506277e-01
6.05357528e-01 -3.63707513e-01 -1.04498279e+00 1.14585233e+00
6.90379798e-01 -2.34125376e-01 -1.26453444e-01 1.06162941e+00
1.58447373e+00 4.05639410e-01 -8.40043947e-02 -4.13065106e-01
1.77524579e+00 -4.62820470e-01 -1.08411527e+00 2.49858633e-01
8.09837803e-02 -1.22501707e+00 7.50607610e-01 2.86789536e-01
-7.26969242e-01 -2.44137809e-01 -9.59004760e-01 4.41572219e-02
-5.86715162e-01 3.81802857e-01 3.11169475e-01 1.22451162e+00
-8.15866470e-01 -8.06368217e-02 -8.44836295e-01 -8.12440097e-01
-1.58930838e-01 7.48255074e-01 -2.70193309e-01 5.62759697e-01
-4.49433714e-01 1.04043400e+00 -2.11765356e-02 3.60297799e-01
-1.73711416e-03 1.62515372e-01 -6.87181175e-01 -2.93579340e-01
2.15255886e-01 -1.66485056e-01 1.09780717e+00 -1.00659287e+00
-2.15861320e+00 1.10490203e+00 -6.59132481e-01 2.09686488e-01
5.03568769e-01 8.14370587e-02 -6.13086581e-01 2.85595834e-01
-6.28817797e-01 3.25827479e-01 1.05101943e+00 -1.15929830e+00
-3.36355209e-01 -7.75143802e-01 -2.28316799e-01 4.77031171e-01
-1.07934132e-01 7.78016865e-01 -5.09482741e-01 -2.79796332e-01
4.20670748e-01 -7.24175453e-01 5.71894586e-01 2.79040486e-01
-3.20653379e-01 -5.71737826e-01 1.53016293e+00 -7.06373990e-01
9.91471350e-01 -1.86635721e+00 -3.89144003e-01 3.70142996e-01
-2.46255677e-02 8.08133185e-01 1.85193405e-01 4.58082050e-01
-2.18397398e-02 -1.69357017e-01 4.32608426e-01 5.36942622e-03
-7.75648281e-02 -5.60850874e-02 -4.99537215e-02 5.59074521e-01
-1.37333706e-01 6.88610554e-01 -3.38064730e-01 -6.82543993e-01
4.56847370e-01 8.16540182e-01 1.52455688e-01 2.98168987e-01
3.60011041e-01 7.08553851e-01 -6.67579710e-01 1.04826105e+00
5.03921330e-01 4.50823426e-01 6.81845322e-02 -2.51978695e-01
-4.07254070e-01 -4.28708233e-02 -1.26831353e+00 9.15733993e-01
-1.57848001e-01 6.72889173e-01 4.33179200e-01 -5.72781980e-01
1.27552962e+00 8.24645281e-01 2.69255638e-01 -6.76716208e-01
7.15076327e-01 1.21534921e-01 -4.76592481e-02 -1.19983888e+00
1.42280251e-01 -2.29861498e-01 6.89218700e-01 6.16441548e-01
-4.25485253e-01 1.24566555e-01 -2.31410831e-01 -4.90756094e-01
3.56261760e-01 4.72450763e-01 6.11406922e-01 1.80616096e-01
6.88196301e-01 -4.98050928e-01 7.38234073e-02 -1.73770413e-02
-3.15937132e-01 3.72135699e-01 1.59116298e-01 -3.34887266e-01
-3.84705782e-01 -8.64659846e-01 -1.70168206e-01 8.52553070e-01
1.48193404e-01 2.19513044e-01 -8.27561319e-01 -5.05505979e-01
-3.00783575e-01 2.33852848e-01 -4.26126480e-01 4.14893270e-01
-4.79553640e-01 -4.12515879e-01 3.75097960e-01 3.90497833e-01
5.81437409e-01 -1.37258589e+00 -1.20511150e+00 -2.70359427e-01
2.59066522e-01 -4.92666364e-01 -1.27304718e-01 -1.48450240e-01
-7.81247079e-01 -1.25221550e+00 -1.01347029e+00 -8.60284567e-01
9.40922678e-01 2.00628161e-01 4.30827409e-01 3.16113353e-01
-3.80496711e-01 6.36955917e-01 -6.24930799e-01 -9.60830152e-01
-1.41037241e-01 -3.05614084e-01 5.58639169e-02 1.93965316e-01
1.03395498e+00 -2.71701902e-01 -6.19534552e-01 5.93119264e-01
-4.83152717e-01 -4.42337126e-01 3.02023232e-01 2.92472273e-01
8.26347917e-02 -4.00485545e-01 -1.72083583e-02 -3.75932872e-01
5.88838220e-01 -2.06820443e-02 -3.40421408e-01 3.88519973e-01
-1.14680119e-01 -9.62691829e-02 -7.28681162e-02 -5.21111846e-01
-1.14731061e+00 3.39082032e-01 -4.19605188e-02 -1.06170744e-01
-6.62505388e-01 -1.23829193e-01 -4.00997251e-01 -3.23439151e-01
4.49447930e-01 1.86903119e-01 4.82601434e-01 -6.39424741e-01
-8.09983388e-02 1.47339272e+00 3.24013770e-01 -2.74167418e-01
5.49380183e-01 5.67955256e-01 5.24763837e-02 -1.55516493e+00
-6.96492568e-02 -8.79878700e-01 -1.04919410e+00 -6.89152837e-01
9.51246202e-01 -1.78286448e-01 -1.51282871e+00 6.52288914e-01
-1.14946747e+00 -2.37731338e-02 1.79038391e-01 6.76746428e-01
-8.83313566e-02 4.58365411e-01 -1.40933067e-01 -1.67391908e+00
-5.12239158e-01 -9.31724727e-01 1.03220510e+00 6.99607432e-01
-2.96253771e-01 -9.34698939e-01 -1.20587908e-01 3.28308135e-01
2.59859383e-01 3.10828239e-01 5.31918526e-01 -4.11105871e-01
-1.30465403e-01 -5.93070030e-01 -2.13529589e-03 -3.17220166e-02
4.80761856e-01 3.61947596e-01 -1.20628035e+00 5.13753556e-02
1.85907826e-01 1.07458048e-02 4.04853374e-01 2.59906322e-01
4.62632507e-01 -1.01139285e-01 -4.96789753e-01 3.32658917e-01
1.30760229e+00 9.30841327e-01 7.70231187e-01 1.60640076e-01
4.16602880e-01 8.11893106e-01 6.81430876e-01 3.10951233e-01
-1.21484473e-01 9.61092830e-01 3.05540770e-01 -2.35475957e-01
-2.83500761e-01 1.45057216e-01 4.40848231e-01 4.39044833e-01
-7.37446904e-01 -2.07586691e-01 -8.65539730e-01 -7.81761557e-02
-1.37551200e+00 -1.00459039e+00 -4.19536203e-01 2.44169331e+00
3.69107127e-01 -3.76224905e-01 4.21392918e-01 5.44958472e-01
8.49532485e-01 -3.53826344e-01 -3.98652256e-01 -7.19466329e-01
3.45763415e-01 5.04114568e-01 2.20858634e-01 5.02946198e-01
-1.01107502e+00 6.73099518e-01 6.29517508e+00 2.31437191e-01
-1.70253932e+00 -7.38204867e-02 2.61255950e-02 -2.14816719e-01
2.95664936e-01 -3.92851710e-01 -9.19876933e-01 3.76191914e-01
3.27502012e-01 2.07188159e-01 2.54138827e-01 4.02800143e-01
4.51243252e-01 -5.27719915e-01 -9.58616912e-01 1.37325108e+00
3.37722123e-01 -5.48348010e-01 -1.00637205e-01 1.58385947e-01
9.18043852e-02 -4.36565518e-01 -5.91563173e-02 -3.05013120e-01
-6.89390540e-01 -1.12232411e+00 4.66548532e-01 7.93316722e-01
8.81174207e-01 -4.18016881e-01 7.53343880e-01 2.49698415e-01
-1.34434974e+00 1.61533743e-01 -6.06080778e-02 -2.90619552e-01
9.09722075e-02 -4.24656093e-01 -1.05663550e+00 5.61989136e-02
3.52550685e-01 3.53219658e-02 -2.67269552e-01 1.15485406e+00
-2.69249022e-01 3.94922942e-01 -5.21520674e-01 -5.72212696e-01
-1.20030321e-01 -3.13382179e-01 6.45895123e-01 9.74527061e-01
2.00878739e-01 5.07860124e-01 -3.54440153e-01 7.32662439e-01
4.95475709e-01 4.05935884e-01 -6.27132654e-01 4.97573763e-02
4.16887939e-01 1.26984501e+00 -1.08018816e+00 2.96786781e-02
-4.91807163e-01 9.96861219e-01 -6.93302929e-01 5.11937201e-01
-3.60571325e-01 -6.48636758e-01 2.85344481e-01 2.94319302e-01
6.17126562e-02 -1.77909419e-01 -1.74733996e-01 -6.13622725e-01
4.92265075e-02 -6.53524041e-01 2.65707582e-01 -8.35007250e-01
-4.56015170e-01 4.24760610e-01 4.32240637e-03 -1.29476488e+00
-2.93159574e-01 -1.04008818e+00 -8.62976074e-01 1.12757540e+00
-1.06586194e+00 -1.35139298e+00 -7.33006954e-01 7.91628718e-01
4.92038637e-01 -1.94825813e-01 1.06916678e+00 -4.00109403e-02
-4.65740830e-01 6.92213535e-01 -1.27338395e-01 5.31252086e-01
6.12189949e-01 -9.09242332e-01 -2.58872598e-01 6.28894150e-01
5.25122248e-02 7.79656649e-01 5.23139417e-01 -4.63375419e-01
-1.49510169e+00 -8.94809365e-02 9.63145316e-01 -5.29420793e-01
6.46073893e-02 -1.23065785e-01 -5.09828925e-01 4.52111989e-01
1.01220489e-01 -1.93720251e-01 6.29943252e-01 -1.13924600e-01
-6.96617141e-02 -9.06894207e-02 -1.55863035e+00 4.53579068e-01
5.87172449e-01 -3.32945585e-01 -8.63742709e-01 4.60831486e-02
-2.74482816e-01 -4.48555350e-01 -4.29886580e-01 -2.09420379e-02
1.18241179e+00 -1.02663660e+00 7.26237893e-01 -3.25973302e-01
-3.52507770e-01 -3.07687044e-01 3.64202708e-01 -5.38972914e-01
3.23459178e-01 -7.24990547e-01 2.17144042e-01 1.30540419e+00
1.31811365e-01 -1.03813982e+00 9.66808975e-01 8.25156689e-01
3.92993689e-01 -3.75638694e-01 -7.46615052e-01 -5.62513769e-01
-5.36731601e-01 -1.76409930e-02 5.16473651e-01 6.84688807e-01
3.76387537e-01 2.79584527e-01 7.16717467e-02 5.37438430e-02
3.67423296e-01 -1.17283665e-01 8.93769920e-01 -1.30328929e+00
9.39973444e-02 -5.00236273e-01 -7.45063484e-01 -7.47892559e-01
-3.33461493e-01 -3.14482421e-01 -1.79557115e-01 -1.45103920e+00
8.04719143e-03 -1.26477361e-01 1.67706430e-01 5.05259573e-01
3.42671335e-01 3.36802691e-01 1.09784558e-01 3.26156884e-01
2.69124806e-02 -1.45789847e-01 1.18093443e+00 2.40036964e-01
-6.57132983e-01 4.02516752e-01 3.10349111e-02 9.33596134e-01
7.63301849e-01 -2.36923192e-02 -3.98648977e-01 -2.30255909e-03
2.19851499e-03 6.65052002e-03 3.15803379e-01 -7.10896611e-01
3.39783341e-01 -1.99059665e-01 4.21653301e-01 -7.23629951e-01
4.51370180e-01 -1.01087511e+00 1.00453794e-02 6.25920534e-01
2.03482613e-01 -3.55759729e-03 -5.03237359e-02 8.89688209e-02
-5.00769764e-02 -5.50434113e-01 6.61821663e-01 -2.37942770e-01
-8.21330190e-01 -1.88374341e-01 -5.61179459e-01 -6.51485860e-01
1.09561670e+00 -9.54684079e-01 -1.56816602e-01 -4.97476876e-01
-8.50019753e-01 -2.91088581e-01 3.78353566e-01 3.79085034e-01
6.01848960e-01 -8.73364329e-01 -2.20866874e-01 5.22006154e-01
-9.09651443e-03 -4.15507197e-01 -2.21863255e-01 1.07198405e+00
-8.26742470e-01 5.27067542e-01 -3.06833863e-01 -5.37903249e-01
-2.27404499e+00 2.86445975e-01 5.09097338e-01 5.67421496e-01
-3.71177107e-01 6.71715260e-01 -7.57971108e-02 -4.93688043e-04
5.91182411e-01 -2.91992784e-01 -8.96916211e-01 1.30050570e-01
9.27218139e-01 6.33431613e-01 -5.70362322e-02 -1.08223200e+00
-6.11309588e-01 1.17642987e+00 5.41441023e-01 -3.45288873e-01
9.53091025e-01 -1.33453622e-01 -1.37031734e-01 5.06905794e-01
9.56846118e-01 2.32624456e-01 -7.43983448e-01 1.84227508e-02
-1.86806589e-01 -4.61334944e-01 -9.68084577e-03 -9.79280829e-01
-8.38034093e-01 1.08410048e+00 1.13138664e+00 2.27400914e-01
1.21707559e+00 -1.00445166e-01 2.84328640e-01 3.42563033e-01
4.26582277e-01 -1.25774312e+00 -6.20876402e-02 1.60410553e-01
1.05759370e+00 -1.12502217e+00 -8.66364241e-02 -5.34954906e-01
-4.59485650e-01 1.63421547e+00 3.52957726e-01 3.09506714e-01
8.87040794e-01 4.03088480e-01 5.43202162e-01 -4.89445269e-01
-2.04552547e-03 -6.25248432e-01 7.43124604e-01 9.66425419e-01
8.37314129e-01 5.14616929e-02 -7.66591191e-01 2.26122096e-01
-1.43248469e-01 1.60195887e-01 1.28877074e-01 1.04747963e+00
-6.90432787e-01 -1.05550241e+00 -9.37069356e-01 1.73114225e-01
-5.10957837e-01 4.89634722e-01 -8.24079990e-01 9.62230086e-01
1.62065208e-01 1.32328057e+00 1.69566676e-01 -1.83231667e-01
4.00400519e-01 4.62429583e-01 8.84720743e-01 -5.44870436e-01
-4.80151176e-01 2.95281708e-01 -8.47150460e-02 -3.27305734e-01
-7.80024111e-01 -6.26529157e-01 -1.53511882e+00 -1.78229377e-01
-4.98996466e-01 -7.51806200e-02 1.26159108e+00 1.10412109e+00
-3.37729037e-01 -1.35163993e-01 4.35592026e-01 -9.59767818e-01
-1.99144885e-01 -1.08196259e+00 -7.90298402e-01 2.65541464e-01
2.98628002e-01 -7.38948047e-01 -2.91349918e-01 8.19585770e-02] | [6.506170272827148, -0.2459171861410141] |
7173c5b5-fcf8-4821-b27c-b5ec2e03c78e | api2com-on-the-improvement-of-automatically | 2103.10668 | null | https://arxiv.org/abs/2103.10668v1 | https://arxiv.org/pdf/2103.10668v1.pdf | API2Com: On the Improvement of Automatically Generated Code Comments Using API Documentations | Code comments can help in program comprehension and are considered as important artifacts to help developers in software maintenance. However, the comments are mostly missing or are outdated, specially in complex software projects. As a result, several automatic comment generation models are developed as a solution. The recent models explore the integration of external knowledge resources such as Unified Modeling Language class diagrams to improve the generated comments. In this paper, we propose API2Com, a model that leverages the Application Programming Interface Documentations (API Docs) as a knowledge resource for comment generation. The API Docs include the description of the methods in more details and therefore, can provide better context in the generated comments. The API Docs are used along with the code snippets and Abstract Syntax Trees in our model. We apply the model on a large Java dataset of over 130,000 methods and evaluate it using both Transformer and RNN-base architectures. Interestingly, when API Docs are used, the performance increase is negligible. We therefore run different experiments to reason about the results. For methods that only contain one API, adding API Docs improves the results by 4% BLEU score on average (BLEU score is an automatic evaluation metric used in machine translation). However, as the number of APIs that are used in a method increases, the performance of the model in generating comments decreases due to long documentations used in the input. Our results confirm that the API Docs can be useful in generating better comments, but, new techniques are required to identify the most informative ones in a method rather than using all documentations simultaneously. | ['Fatemeh H. Fard', 'Rishab Sharma', 'Ramin Shahbazi'] | 2021-03-19 | null | null | null | null | ['comment-generation'] | ['natural-language-processing'] | [ 8.51196125e-02 2.95058578e-01 -1.35703549e-01 -2.17656195e-01
-7.35418797e-01 -7.40834832e-01 5.25510073e-01 2.75056362e-01
-1.15543909e-01 5.23386598e-01 3.36704075e-01 -5.09510040e-01
2.61362255e-01 -7.58956909e-01 -7.59612024e-01 -1.47404626e-01
4.94120747e-01 -4.10733968e-02 2.05686450e-01 -2.80326515e-01
4.72672164e-01 -2.75762111e-01 -1.58775389e+00 6.31245017e-01
1.41099548e+00 2.00084269e-01 6.82819784e-01 5.01626849e-01
-7.87937224e-01 7.07639813e-01 -9.49944079e-01 -8.26389074e-01
-1.98101043e-03 -6.77472889e-01 -8.68094563e-01 -3.89327377e-01
2.14723170e-01 -1.53539613e-01 5.46636999e-01 1.11473954e+00
4.24222976e-01 -2.17645526e-01 4.64126199e-01 -1.08834517e+00
-7.07988143e-01 1.56009245e+00 -4.30683076e-01 -1.17597841e-01
5.87992728e-01 -6.83902502e-02 1.25677586e+00 -9.61114883e-01
8.42734635e-01 7.78007865e-01 5.19229889e-01 7.02847123e-01
-1.08234680e+00 -4.96736407e-01 5.24462163e-02 2.86770940e-01
-1.00458241e+00 -2.07194626e-01 9.09223974e-01 -8.05022240e-01
1.18924344e+00 3.92446458e-01 5.95660985e-01 1.16520250e+00
8.78739208e-02 6.34485602e-01 5.84845543e-01 -9.16790545e-01
-2.36706752e-02 6.45340919e-01 3.20016831e-01 5.52622199e-01
3.95953864e-01 -3.44415516e-01 -3.58731627e-01 -2.45451093e-01
2.07756296e-01 -1.04859293e-01 -3.60738933e-01 5.19164652e-02
-1.01954234e+00 9.08542037e-01 6.11302517e-02 6.25836074e-01
-2.90780157e-01 6.96068779e-02 4.55186784e-01 3.54034066e-01
2.26235703e-01 8.87444675e-01 -6.66951299e-01 -7.76996017e-01
-8.98221314e-01 3.21146399e-01 1.09101844e+00 1.31027913e+00
9.46247637e-01 -2.49962020e-03 -4.26289082e-01 1.09997368e+00
1.74607709e-01 4.13280874e-01 5.93915701e-01 -8.05311620e-01
8.28559041e-01 1.17607319e+00 7.17225075e-02 -8.70977163e-01
-2.45700646e-02 -6.24412060e-01 -2.63056457e-01 6.19522184e-02
3.46326917e-01 -3.98549199e-01 -4.23150241e-01 1.58636856e+00
-9.08286497e-02 -4.44605052e-01 9.26877931e-02 3.41925442e-01
9.21845794e-01 5.03526568e-01 -2.41232172e-01 -2.10291892e-03
1.28441799e+00 -1.29000533e+00 -6.41233742e-01 -3.48422676e-01
1.12930262e+00 -1.14049304e+00 1.47804558e+00 2.25642666e-01
-9.11198080e-01 -6.22787774e-01 -1.01758325e+00 1.32448569e-01
-4.14389163e-01 6.00507379e-01 3.55495721e-01 6.63097560e-01
-9.24315631e-01 6.75206780e-01 -6.34674728e-01 -4.48730856e-01
-5.36938533e-02 -2.17116982e-01 -3.62452306e-03 -1.00930601e-01
-9.74872947e-01 8.60737741e-01 1.84543937e-01 -1.62251383e-01
-3.42015773e-01 -8.83976698e-01 -8.54280651e-01 1.83834612e-01
4.18426692e-01 -6.92677915e-01 1.46557868e+00 -1.22212017e+00
-1.40684879e+00 1.52955174e-01 -1.68319255e-01 -8.28717425e-02
2.51040906e-01 -3.14069569e-01 -2.80260265e-01 -4.40513343e-01
1.76781848e-01 3.61161023e-01 4.05518293e-01 -1.05311882e+00
-5.94737887e-01 1.34555086e-01 4.33172733e-01 -2.16976687e-01
-4.53723848e-01 3.24187040e-01 -4.42816556e-01 -5.65556884e-01
-4.33477610e-01 -9.96450841e-01 -1.04475386e-01 -5.09104967e-01
-4.27561432e-01 -1.81090146e-01 3.32081109e-01 -1.04453754e+00
1.74856722e+00 -1.87062371e+00 1.81254134e-01 1.30859658e-01
4.15436402e-02 1.54450387e-01 -3.76559526e-01 8.11065793e-01
-8.26034769e-02 6.84182048e-01 -1.61460817e-01 -5.42834587e-02
1.22236744e-01 -5.72939515e-02 6.16389932e-03 -4.45432216e-01
9.22747403e-02 9.11306679e-01 -7.85353303e-01 -2.10201353e-01
-2.85674244e-01 2.87006766e-01 -8.09362710e-01 2.57447571e-01
-5.20203531e-01 1.42835990e-01 -4.09820467e-01 3.45083028e-01
5.22146583e-01 -2.33376399e-01 1.43285915e-01 -4.46547978e-02
-2.32837722e-01 6.64185107e-01 -9.91610289e-01 1.83631206e+00
-9.43178236e-01 7.16241896e-01 -4.92937922e-01 -5.24481714e-01
1.06974733e+00 3.28275412e-01 -8.86467919e-02 -5.94902813e-01
-1.97410628e-01 3.54958415e-01 3.96710157e-01 -8.63371968e-01
6.07802868e-01 3.73336941e-01 2.41158120e-02 7.74645329e-01
-1.55401558e-01 -6.77404180e-02 8.65552902e-01 3.33005846e-01
1.27901363e+00 5.38528204e-01 4.02364463e-01 3.04609351e-02
6.25945270e-01 8.54774192e-02 5.64975977e-01 7.31784046e-01
7.55982220e-01 6.25652373e-01 8.25653374e-01 -1.97309226e-01
-1.07597733e+00 -4.44953859e-01 3.37785542e-01 9.56766844e-01
-4.50763881e-01 -1.19500709e+00 -8.78328323e-01 -1.01830709e+00
-4.60658044e-01 1.17085457e+00 -5.03031909e-01 -2.92050447e-02
-6.34775281e-01 -4.40209270e-01 6.10827148e-01 6.93407953e-01
8.94193351e-02 -1.17536819e+00 -7.35630572e-01 3.28484207e-01
-5.40134609e-01 -8.48854184e-01 -4.94923353e-01 -9.48927328e-02
-6.90469265e-01 -1.06232381e+00 -6.19416595e-01 -4.15377527e-01
8.10315251e-01 -2.12474056e-02 1.37020004e+00 5.00770748e-01
2.07847096e-02 1.08855635e-01 -1.00835693e+00 -4.85063761e-01
-1.06284547e+00 3.46303225e-01 -6.52079284e-01 -6.52412593e-01
4.45339441e-01 -5.61376154e-01 -9.39872786e-02 1.15931243e-01
-9.62178051e-01 4.33650732e-01 7.83174217e-01 7.65134275e-01
-3.30147557e-02 -3.47127527e-01 4.01583910e-01 -1.36211944e+00
8.37787986e-01 -5.04169583e-01 -6.78791583e-01 4.91125613e-01
-7.89149940e-01 4.38723117e-01 8.89766335e-01 -3.67781192e-01
-1.13095522e+00 -3.20738852e-01 -2.83398509e-01 1.77410200e-01
4.62582149e-02 1.24410903e+00 1.03196325e-02 2.71565557e-01
9.04288590e-01 2.19631404e-01 -1.59491748e-01 -7.08168030e-01
1.64238021e-01 7.93649554e-01 -2.50378609e-01 -7.61003017e-01
8.91520560e-01 -4.10155028e-01 -6.02192342e-01 -4.37880635e-01
-3.39080811e-01 -1.77129745e-01 -2.85427779e-01 -1.02519341e-01
4.01408762e-01 -5.45411825e-01 -1.29794896e-01 4.22011428e-02
-1.66718042e+00 -2.01499909e-01 -1.28593355e-01 4.00849432e-01
-1.37423083e-01 4.08701062e-01 -4.02091026e-01 -6.02272570e-01
-5.28174102e-01 -1.60970592e+00 6.83098793e-01 2.57330000e-01
-6.92342758e-01 -7.57603467e-01 1.59730241e-01 2.27789238e-01
6.28627121e-01 -3.26317400e-02 1.54103518e+00 -8.54146302e-01
-5.41913748e-01 -9.44905728e-02 -1.01364769e-01 6.76443100e-01
2.76455164e-01 4.57291871e-01 -6.23247862e-01 1.11971773e-01
-3.62711340e-01 1.83660939e-01 2.08116159e-01 -1.96393773e-01
9.82869267e-01 -5.55304289e-01 -1.35916278e-01 1.22070715e-01
1.43687785e+00 2.71990716e-01 7.59542882e-01 4.28143442e-01
6.09504998e-01 7.80966759e-01 5.09674847e-01 3.98027211e-01
5.39320230e-01 7.08495736e-01 3.48123580e-01 1.83132887e-01
-1.93932727e-01 -2.34240592e-01 8.40255260e-01 1.44024980e+00
-7.01683983e-02 -1.80300474e-01 -1.03761828e+00 6.77316129e-01
-1.68953753e+00 -6.26665175e-01 -6.40744328e-01 2.18835592e+00
1.15874922e+00 4.34427634e-02 -1.42514691e-01 -7.03726313e-04
3.42238396e-01 -3.17401886e-01 -1.51365697e-01 -5.98226666e-01
2.00507641e-01 1.81425050e-01 6.78987727e-02 2.35031456e-01
-2.08004400e-01 5.87961257e-01 5.10300779e+00 5.69614649e-01
-1.01977015e+00 5.43152317e-02 -6.69507533e-02 2.16492459e-01
-8.44865382e-01 4.84927148e-01 -9.27623868e-01 6.02093816e-01
1.03639650e+00 -5.19052684e-01 4.76649493e-01 1.15119934e+00
3.01052839e-01 -1.80524305e-01 -1.37172139e+00 7.16210902e-01
1.86941981e-01 -1.50156105e+00 1.61051944e-01 -1.08668268e-01
7.98427641e-01 -6.23052381e-02 -4.22349721e-01 4.19676453e-01
2.81974345e-01 -5.67300797e-01 7.34613597e-01 5.07242382e-01
3.31800431e-01 -5.27629972e-01 1.17972708e+00 5.08671343e-01
-8.70528698e-01 7.97392204e-02 -3.66171926e-01 -1.05418973e-02
-1.14850715e-01 7.32703745e-01 -1.07313335e+00 8.13151598e-01
5.82535625e-01 7.10024118e-01 -1.09338176e+00 1.22887087e+00
-6.31632328e-01 7.19453514e-01 -8.62011760e-02 -4.32919830e-01
-6.51958808e-02 -2.27916673e-01 4.19862121e-01 1.42138648e+00
7.33523130e-01 -5.64378321e-01 -1.58355236e-02 1.34210002e+00
-8.26565623e-02 6.59161150e-01 -5.01025379e-01 -3.49071354e-01
5.07805943e-01 1.32903612e+00 -4.45732206e-01 -3.38056415e-01
-8.64045620e-01 6.93883657e-01 1.44065335e-01 4.69739288e-01
-8.96924973e-01 -9.42410529e-01 4.68042642e-01 1.45841166e-01
2.83437341e-01 9.18609500e-02 -1.82231724e-01 -1.03062367e+00
5.46175539e-01 -1.27010500e+00 -6.49114475e-02 -1.02010357e+00
-7.80453563e-01 8.61225247e-01 3.55669148e-02 -1.38338804e+00
-6.33656263e-01 -1.29472896e-01 -7.11378992e-01 9.88215208e-01
-1.36566186e+00 -9.77189898e-01 -4.82548505e-01 -7.02959448e-02
7.79279709e-01 -5.97317554e-02 8.37406218e-01 5.57810068e-01
-3.99638325e-01 8.15252960e-01 -3.93383987e-02 3.08184266e-01
7.72158206e-01 -1.28156984e+00 5.17377317e-01 1.18879664e+00
2.56705999e-01 1.26705408e+00 8.00473988e-01 -7.63170898e-01
-1.16871595e+00 -9.46256876e-01 1.19480252e+00 -4.85602677e-01
6.82443261e-01 -2.76487291e-01 -9.47755516e-01 5.63020945e-01
5.03447235e-01 -7.78076410e-01 8.71004820e-01 1.82675049e-01
-5.18141031e-01 9.12122279e-02 -6.27768874e-01 6.23110890e-01
6.59853399e-01 -3.37265164e-01 -5.55745900e-01 7.45958984e-02
8.64303410e-01 -2.90844232e-01 -8.44062984e-01 -1.97606236e-02
4.74478841e-01 -8.95689011e-01 2.46242523e-01 -3.28272939e-01
9.47199404e-01 -5.26363671e-01 9.51365829e-02 -1.58189940e+00
-2.97622569e-02 -4.74663347e-01 8.22480917e-02 1.73267162e+00
1.26268423e+00 -4.85044420e-01 4.33334619e-01 5.46134472e-01
-3.06939870e-01 -5.27737737e-01 -3.80101919e-01 -6.51038349e-01
-9.12677348e-02 -5.78478336e-01 8.72542381e-01 8.22452307e-01
3.11528504e-01 5.01719832e-01 -1.15536056e-01 -3.75381649e-01
1.46384090e-02 1.17062606e-01 1.02486491e+00 -1.32741189e+00
-5.82377374e-01 -3.88453066e-01 1.45905629e-01 -9.12245274e-01
6.18744902e-02 -1.10172606e+00 6.51174858e-02 -1.77951884e+00
3.75441819e-01 -5.40001094e-01 3.48094255e-01 7.50018179e-01
-2.88676172e-01 -3.24981987e-01 3.65163296e-01 2.50363760e-02
-9.87126008e-02 1.42246887e-01 8.71959567e-01 -2.25813001e-01
-3.82755905e-01 2.40743309e-01 -9.56634820e-01 7.10294485e-01
7.16066062e-01 -8.49406481e-01 -3.73008817e-01 -6.88508749e-01
9.20180321e-01 -1.02500550e-01 -1.28642961e-01 -8.37448120e-01
1.92559898e-01 -6.52756393e-02 -1.81507364e-01 -3.83173943e-01
-2.14492351e-01 -7.69458413e-01 3.73284101e-01 3.53305638e-01
-4.59947288e-01 3.82932752e-01 1.18743435e-01 3.18080783e-02
-1.51501387e-01 -1.12913799e+00 3.71290356e-01 -2.31027573e-01
-3.94439787e-01 -3.65538985e-01 -6.04976535e-01 2.31336020e-02
5.37918448e-01 -2.44262278e-01 -5.22234499e-01 -2.81506032e-01
-2.12442085e-01 -1.80565253e-01 6.76839352e-01 8.41495872e-01
3.66413355e-01 -1.01009905e+00 -7.10766852e-01 3.06039210e-02
5.28376281e-01 -2.91520536e-01 -1.36465326e-01 7.86503673e-01
-4.48544741e-01 3.23559612e-01 -6.83434978e-02 -2.43419617e-01
-1.38114321e+00 2.31003329e-01 3.81279178e-02 -4.20734257e-01
-4.07931864e-01 5.33226490e-01 -1.54271156e-01 -5.98456621e-01
3.65267135e-03 -6.48352802e-01 -4.52481210e-01 1.27313778e-01
7.32407153e-01 3.48569363e-01 3.26929361e-01 -1.32724971e-01
-1.01660319e-01 5.04562616e-01 -2.71355301e-01 -6.00188114e-02
1.48500359e+00 8.63825604e-02 -4.68604296e-01 5.07377982e-01
9.38751876e-01 6.45564258e-01 -5.52065253e-01 -6.80989027e-02
3.66999984e-01 -3.06073934e-01 -4.18667078e-01 -1.05847919e+00
-1.07778454e+00 8.25302482e-01 1.57192364e-01 1.56666890e-01
8.19345951e-01 2.44333632e-02 5.15011609e-01 4.62548614e-01
4.12605226e-01 -1.03715539e+00 8.20867065e-03 6.58333778e-01
8.82743955e-01 -9.03604388e-01 -2.44905069e-01 -5.25833488e-01
-7.31911004e-01 1.40376186e+00 8.65724623e-01 4.50229406e-01
1.38194680e-01 4.12726969e-01 1.05984733e-01 1.12240553e-01
-9.25364673e-01 5.59722893e-02 2.33955249e-01 6.03049457e-01
1.02504289e+00 -2.23123968e-01 -8.01363051e-01 9.95791793e-01
-3.84379625e-01 8.76759812e-02 1.22934580e+00 9.97299016e-01
-1.85459822e-01 -1.79419684e+00 -1.05036348e-01 6.72763407e-01
-4.49689746e-01 -6.16743386e-01 -4.39015418e-01 5.24605572e-01
1.00223340e-01 1.06595099e+00 -4.25225765e-01 -4.74214166e-01
4.57713097e-01 3.11743647e-01 2.67610580e-01 -1.20729876e+00
-1.10568082e+00 -2.70280510e-01 4.06390429e-01 -3.24980974e-01
-3.39548498e-01 -5.11103213e-01 -1.16727197e+00 -7.16719776e-02
-6.94449365e-01 4.83977050e-01 9.99156058e-01 8.70868444e-01
5.97270429e-01 7.99383640e-01 2.96386033e-01 -1.52165800e-01
-3.68185312e-01 -1.27718151e+00 6.20715246e-02 1.07474789e-01
-2.48619244e-01 -4.02597070e-01 -2.61806637e-01 4.54650968e-01] | [7.709213733673096, 7.897206783294678] |
100b24d9-b211-45a6-9373-a27c95ca9894 | parsing-to-1-endpoint-crossing-pagenumber-2 | null | null | https://aclanthology.org/P17-1193 | https://aclanthology.org/P17-1193.pdf | Parsing to 1-Endpoint-Crossing, Pagenumber-2 Graphs | We study the Maximum Subgraph problem in deep dependency parsing. We consider two restrictions to deep dependency graphs: (a) 1-endpoint-crossing and (b) pagenumber-2. Our main contribution is an exact algorithm that obtains maximum subgraphs satisfying both restrictions simultaneously in time O(n5). Moreover, ignoring one linguistically-rare structure descreases the complexity to O(n4). We also extend our quartic-time algorithm into a practical parser with a discriminative disambiguation model and evaluate its performance on four linguistic data sets used in semantic dependency parsing. | ['Weiwei Sun', 'Sheng Huang', 'Junjie Cao', 'Xiaojun Wan'] | 2017-07-01 | null | null | null | acl-2017-7 | ['semantic-dependency-parsing'] | ['natural-language-processing'] | [-8.17373767e-02 7.52324402e-01 -3.26383412e-01 -5.95123053e-01
-1.05858588e+00 -9.15901303e-01 1.46693110e-01 4.44088280e-01
-5.36856711e-01 8.94988477e-01 -2.70488225e-02 -6.66774809e-01
7.96010252e-03 -8.68844748e-01 -6.08662784e-01 -5.50577462e-01
-4.10241812e-01 7.59262264e-01 6.78221345e-01 -1.98998645e-01
-1.14789069e-01 7.25172579e-01 -9.64453936e-01 2.78806966e-02
9.17184830e-01 2.66221851e-01 1.73202693e-01 9.36024129e-01
-4.95861858e-01 4.69424278e-01 -6.99498653e-01 -7.23427117e-01
2.59928375e-01 -5.27618468e-01 -1.32035327e+00 2.66547706e-02
3.31346303e-01 3.96577977e-02 -9.42944735e-02 1.28271043e+00
2.86888391e-01 -1.20973503e-02 2.11240128e-01 -9.02667642e-01
-3.17257971e-01 1.17883074e+00 -7.84886539e-01 5.78098118e-01
2.46939167e-01 -6.31405771e-01 1.30856919e+00 -2.69382507e-01
7.75949955e-01 1.41749549e+00 4.24242526e-01 5.71979225e-01
-1.23120642e+00 -2.99269587e-01 4.63884592e-01 -1.99053381e-02
-9.24322069e-01 -2.75923491e-01 7.44067132e-01 8.15675184e-02
1.52459431e+00 1.98907688e-01 1.03191309e-01 5.65288901e-01
5.33895977e-02 5.09032965e-01 1.16031885e+00 -8.21261466e-01
1.43882230e-01 -4.47304815e-01 8.04574549e-01 8.95212591e-01
7.27241218e-01 -1.69478714e-01 -1.47500023e-01 -7.81984925e-02
4.86206114e-01 -7.90728867e-01 1.78232089e-01 1.52010143e-01
-4.65950906e-01 1.04815137e+00 1.28943413e-01 6.20329738e-01
1.17853314e-01 2.17254743e-01 4.26395446e-01 5.78162968e-01
3.84084255e-01 1.89803913e-01 -9.56665218e-01 1.30419672e-01
-3.82393658e-01 -4.74657491e-02 1.37457824e+00 1.28012586e+00
8.99797797e-01 -3.54463309e-02 6.75136209e-01 8.59086156e-01
-1.58884451e-02 5.27555585e-01 8.54548886e-02 -9.02330756e-01
9.13944185e-01 2.88538724e-01 -1.23326816e-01 -5.79850674e-01
-9.60263848e-01 -2.22607777e-01 -3.28788817e-01 -1.56446934e-01
7.22202539e-01 -6.06728971e-01 -7.49347806e-01 2.01647449e+00
4.98080075e-01 -3.96921247e-01 1.91743433e-01 4.97002959e-01
5.15931368e-01 4.02113497e-01 4.53710467e-01 -4.82284009e-01
1.80175960e+00 -7.60406077e-01 -7.11830914e-01 -5.08269131e-01
1.38751614e+00 -8.50317299e-01 6.39169037e-01 2.81122506e-01
-1.03018570e+00 -1.53810978e-01 -8.28397155e-01 -4.54668671e-01
-2.23623648e-01 -2.28281423e-01 1.14898515e+00 9.43282664e-01
-1.26454556e+00 6.29068613e-01 -9.69917774e-01 -3.25355828e-01
-1.51758865e-01 7.40989447e-01 -6.37991488e-01 2.10879967e-02
-1.23342264e+00 9.52963531e-01 7.71970689e-01 2.67502461e-02
1.75736416e-02 -1.63237065e-01 -1.09663248e+00 -1.01177573e-01
5.46676755e-01 -4.88409072e-01 1.28360403e+00 -7.74999619e-01
-1.23446965e+00 1.23727787e+00 -2.50212431e-01 -2.83640474e-01
2.40227535e-01 -3.13995212e-01 -3.91950428e-01 2.86583960e-01
2.27750048e-01 5.22289336e-01 -3.66115719e-02 -1.04750896e+00
-7.56168246e-01 -6.96869016e-01 5.96451581e-01 1.17717154e-01
-6.93797739e-03 5.35203099e-01 -6.54659033e-01 -6.36767745e-01
5.47213018e-01 -1.00477827e+00 -4.24775094e-01 -6.78022623e-01
-7.52257586e-01 -5.92879713e-01 3.07861984e-01 -9.40649331e-01
1.19060349e+00 -1.90865076e+00 2.28856742e-01 1.25868171e-01
8.28277767e-02 1.04964159e-01 -1.66315898e-01 4.97855753e-01
-2.27007672e-01 3.05815279e-01 -2.98898369e-01 -3.31667125e-01
6.46685362e-02 9.30261910e-01 2.08459392e-01 4.19832379e-01
2.54349440e-01 7.33555198e-01 -7.94613183e-01 -8.94358575e-01
-1.54668689e-01 -4.75710630e-02 -6.22548163e-01 -6.43842295e-02
-1.68848366e-01 4.79079597e-02 -6.90236688e-01 5.74023366e-01
8.63518476e-01 8.87116864e-02 1.08331072e+00 7.53200576e-02
-2.68594116e-01 6.08445108e-01 -1.00276399e+00 1.74023938e+00
-7.26606667e-01 2.84967989e-01 4.23600078e-01 -8.77632856e-01
7.39940703e-01 8.60160291e-02 1.78198516e-01 -4.37434196e-01
2.45698467e-01 3.47337157e-01 7.64485076e-03 -4.32093501e-01
3.10979694e-01 -3.53200883e-01 -7.22803354e-01 2.29407907e-01
2.67527699e-01 2.00347260e-01 6.37436628e-01 3.14776033e-01
1.28063715e+00 4.45659906e-02 3.54792595e-01 -8.74719203e-01
2.86768079e-01 2.25157887e-01 9.30291414e-01 6.55321717e-01
-2.23019227e-01 1.23321436e-01 1.47344458e+00 -4.13160801e-01
-7.99346507e-01 -7.59962440e-01 -2.57737935e-01 1.39999568e+00
7.09492266e-02 -3.47759128e-01 -9.30929065e-01 -1.03661370e+00
-3.86624724e-01 7.43593216e-01 -5.55004537e-01 3.48779380e-01
-1.44490147e+00 -1.06680214e+00 5.71879983e-01 9.05117095e-01
1.53220505e-01 -8.68469119e-01 -5.62420487e-01 2.68840104e-01
-8.59934762e-02 -1.63855863e+00 -1.96023494e-01 1.01477695e+00
-1.10526490e+00 -1.22497320e+00 -9.28197503e-02 -1.57988095e+00
7.90392876e-01 -1.58844188e-01 1.30996299e+00 1.93074018e-01
1.44393444e-01 -7.95029551e-02 -6.15526259e-01 6.99741840e-02
-4.64508593e-01 3.76468003e-01 -3.74809206e-01 -1.00316739e+00
6.20798230e-01 -4.69151706e-01 5.07032536e-02 -2.03385279e-01
-6.10570669e-01 -2.11186469e-01 5.17862737e-01 6.82058811e-01
4.60450113e-01 6.00380413e-02 2.77996004e-01 -1.55493796e+00
2.97073901e-01 -4.55836236e-01 -9.09179509e-01 3.12707812e-01
-1.60489857e-01 5.42688847e-01 7.88419008e-01 6.77390471e-02
-1.14901721e+00 5.08598208e-01 -6.18106186e-01 3.20179403e-01
-1.74321786e-01 4.25053298e-01 -6.86444283e-01 -1.56023970e-03
2.59353697e-01 -5.16797900e-01 -7.02136636e-01 -8.77265692e-01
4.11688507e-01 1.57737166e-01 6.84270620e-01 -8.98672640e-01
5.27098715e-01 9.06655490e-02 3.63510251e-01 -7.59166300e-01
-9.71933961e-01 -3.49106103e-01 -1.10295618e+00 5.51821053e-01
1.14677382e+00 -6.93328083e-01 -4.20019776e-01 1.41674623e-01
-1.63752031e+00 -2.18736008e-01 1.25825197e-01 4.49539334e-01
-2.25314453e-01 9.09293592e-01 -1.15620172e+00 -5.94961047e-01
-2.26826936e-01 -7.01270461e-01 1.04615283e+00 -2.81913746e-02
-3.42556611e-02 -1.19768643e+00 2.57642388e-01 1.38397021e-02
-4.62027073e-01 3.38729978e-01 1.36071646e+00 -1.09051764e+00
-1.75708383e-01 2.25951206e-02 -2.48044044e-01 -2.71945857e-02
-4.91549373e-02 -3.33283901e-01 -6.05226278e-01 8.42455178e-02
-6.40254393e-02 -7.51485825e-02 9.07255232e-01 6.82104900e-02
7.04348207e-01 -2.42796019e-01 -4.51083362e-01 5.74488819e-01
1.62162137e+00 1.86365530e-01 2.15974972e-01 3.18351358e-01
7.17286348e-01 9.55059946e-01 4.64369774e-01 -1.63331274e-02
2.92585790e-01 4.64055508e-01 2.13175729e-01 -5.85033298e-02
-1.63125917e-01 -7.33642429e-02 3.01550299e-01 1.34289181e+00
-1.21881187e-01 -4.72556591e-01 -9.37294304e-01 6.25961304e-01
-1.51710129e+00 -3.14513624e-01 -7.52117991e-01 1.71902823e+00
8.51375222e-01 4.41348463e-01 -2.75074970e-02 1.52886927e-01
9.29273009e-01 5.42184524e-02 -6.43898919e-02 -1.12125099e+00
-1.60561264e-01 7.70129979e-01 8.21985543e-01 1.08533275e+00
-1.25420547e+00 1.50063825e+00 7.13763857e+00 4.31127369e-01
-4.35459793e-01 3.56456995e-01 3.22671622e-01 3.16627473e-01
-4.67534572e-01 1.91401646e-01 -1.05508363e+00 5.89086600e-02
1.10009897e+00 7.49085918e-02 9.18988660e-02 9.13122177e-01
-5.03565967e-01 -9.83569548e-02 -8.35442066e-01 4.18484926e-01
-1.58830494e-01 -6.49514377e-01 -5.75571544e-02 3.48115340e-02
3.03550452e-01 5.48215434e-02 -6.26364887e-01 5.33182994e-02
8.63959730e-01 -5.21119773e-01 4.08850640e-01 -6.25871897e-01
7.63958395e-01 -1.07010257e+00 9.37063813e-01 2.21923009e-01
-1.26710045e+00 1.18381940e-01 -6.14391625e-01 9.94998142e-02
4.90045428e-01 7.09766746e-01 -4.52961117e-01 8.64990473e-01
4.65927154e-01 1.83494046e-01 -2.87095010e-01 3.26084048e-01
-8.19354951e-01 4.89694625e-01 -7.23496854e-01 2.87629887e-02
4.07696456e-01 -2.47400686e-01 4.08495069e-01 1.66342688e+00
-4.16000150e-02 5.92077136e-01 2.14575320e-01 9.60503966e-02
-1.36006296e-01 1.06124029e-01 -2.99438745e-01 1.93493545e-01
4.68357772e-01 1.05010760e+00 -1.46049106e+00 -1.90379396e-01
-6.92960739e-01 1.04497242e+00 8.02875340e-01 -4.66807671e-02
-6.11017346e-01 -6.94596887e-01 4.88162100e-01 -2.63252586e-01
6.68774843e-01 -6.85469627e-01 -2.59865195e-01 -1.11400008e+00
3.36386114e-02 -4.93838698e-01 8.43038201e-01 -9.74752083e-02
-1.07565343e+00 1.03037000e+00 2.95971394e-01 -3.07771891e-01
-2.54256666e-01 -1.02701247e+00 -2.20331028e-01 6.62402332e-01
-1.63734734e+00 -9.40421999e-01 2.43504971e-01 3.49041075e-01
4.33695793e-01 4.65807289e-01 1.10975003e+00 4.14877355e-01
-6.71755850e-01 5.99152923e-01 -2.98148304e-01 3.39034110e-01
9.78362933e-02 -1.72701943e+00 8.99124146e-01 1.29474449e+00
1.80208057e-01 4.24938142e-01 8.34943354e-01 -6.56395018e-01
-1.30291224e+00 -7.62460470e-01 1.81380212e+00 -2.65004158e-01
9.70262706e-01 -5.39859295e-01 -8.34353089e-01 1.05481613e+00
1.92939341e-01 -6.36862963e-02 5.99139869e-01 5.72786689e-01
-4.37429845e-01 2.16130123e-01 -1.07677484e+00 2.80822963e-01
1.59500015e+00 -2.75397867e-01 -9.38169956e-01 4.73422438e-01
9.06756043e-01 -6.64377093e-01 -9.90056634e-01 2.13889092e-01
1.29481614e-01 -7.13780224e-01 5.59741080e-01 -7.92329550e-01
7.33502358e-02 1.50630161e-01 -4.84780595e-02 -1.00497341e+00
-2.53587633e-01 -9.49127734e-01 1.48998931e-01 1.13389087e+00
6.39638305e-01 -8.34721446e-01 7.04143286e-01 5.19001663e-01
-4.13594365e-01 -4.70108062e-01 -1.23155248e+00 -8.97967100e-01
6.27849281e-01 -2.73897082e-01 4.41574305e-01 9.09672260e-01
1.57671049e-01 6.21061027e-01 -1.42202884e-01 4.42859590e-01
5.27469099e-01 3.85178000e-01 1.23143002e-01 -1.24127340e+00
-5.23684680e-01 -1.09140165e-01 -3.60377640e-01 -1.35079205e+00
5.74285626e-01 -9.52720344e-01 7.05386400e-02 -1.43884230e+00
7.99826682e-02 -5.10606885e-01 -2.35975325e-01 6.43194556e-01
-1.10917673e-01 -1.61443263e-01 -1.63194407e-02 -2.76346296e-01
-5.75969458e-01 -1.26422778e-01 8.86432409e-01 4.41228390e-01
-5.98550551e-02 -2.78409928e-01 -7.70438433e-01 9.36817646e-01
9.37286913e-01 -1.11314332e+00 -1.45338178e-01 -9.55517232e-01
4.57601488e-01 3.81024569e-01 -2.19626904e-01 -4.51107383e-01
-8.79818425e-02 4.44793288e-04 -1.13532610e-01 -4.77526724e-01
-6.39014542e-02 -6.11465335e-01 -2.42299750e-01 6.11645877e-01
-1.68193921e-01 5.76211989e-01 2.48687059e-01 3.75207573e-01
-7.94009119e-02 -7.47379422e-01 5.88607073e-01 -3.03109527e-01
-8.43109667e-01 -3.82045358e-02 -2.15375066e-01 4.20374751e-01
8.95828545e-01 5.26613258e-02 -4.75606352e-01 3.17624420e-01
-1.04977095e+00 1.55306771e-01 2.35458508e-01 2.46836707e-01
-1.12099834e-01 -6.81878269e-01 -6.13636374e-01 5.39589189e-02
-3.40686440e-01 2.41375137e-02 -4.84875962e-02 5.13990283e-01
-1.01456082e+00 4.96934354e-01 9.80404764e-02 -2.03229561e-01
-1.45711267e+00 6.48185909e-01 4.04732712e-02 -5.46021819e-01
-7.14994907e-01 1.06245458e+00 2.88855806e-02 -4.08505201e-01
-7.55925626e-02 -4.64023322e-01 -2.17571035e-01 -1.36375219e-01
1.49892226e-01 2.06871197e-01 1.46787852e-01 -7.21279800e-01
-7.06164360e-01 5.77296317e-01 -1.27016798e-01 -2.54351050e-02
1.33420503e+00 4.58166040e-02 -4.62041080e-01 3.47444601e-02
1.21793127e+00 4.50890690e-01 -7.90177941e-01 5.77458134e-03
7.47144341e-01 -2.00287625e-02 -2.50978291e-01 -3.91639054e-01
-1.05781770e+00 6.68073356e-01 -1.69519708e-02 5.11401892e-01
1.15940678e+00 5.71184754e-01 9.78619754e-01 6.60697401e-01
5.98153114e-01 -1.03410900e+00 -5.64962447e-01 7.46899068e-01
4.72950339e-01 -7.80354321e-01 1.09343804e-01 -1.19977856e+00
-4.78983819e-01 1.14777708e+00 4.89752561e-01 -4.06090170e-01
8.23700130e-01 7.69465744e-01 6.36475235e-02 -1.22059889e-01
-7.48238325e-01 -4.43240315e-01 -1.97226107e-01 4.51523125e-01
5.59297502e-01 2.98495680e-01 -9.90900040e-01 7.06320107e-01
-4.60096955e-01 -6.80962682e-01 3.97229135e-01 1.09854019e+00
-5.63798308e-01 -1.70561993e+00 1.25539541e-01 -1.39455765e-01
-1.12023067e+00 -2.76238322e-01 -6.21136308e-01 1.37038505e+00
1.41748026e-01 1.11530328e+00 1.56966940e-01 9.18967798e-02
4.13430601e-01 -5.85913099e-02 8.68378878e-01 -9.05486763e-01
-6.06213212e-01 1.90753907e-01 8.79921257e-01 -4.70671862e-01
-4.76804435e-01 -5.78901410e-01 -1.86475825e+00 -4.24065411e-01
-6.26650572e-01 4.62791890e-01 5.38782597e-01 1.01489222e+00
8.34054947e-02 2.73833632e-01 3.32235932e-01 -2.94614732e-01
-5.00554264e-01 -7.83654451e-01 -6.58365369e-01 1.58508241e-01
1.02753565e-02 -4.12214220e-01 -4.57477450e-01 -8.36037025e-02] | [10.296298027038574, 9.68644905090332] |
9643d681-9640-456f-8a12-874535b3561d | a-joint-convolutional-neural-networks-and | 1905.01574 | null | https://arxiv.org/abs/1905.01574v1 | https://arxiv.org/pdf/1905.01574v1.pdf | A Joint Convolutional Neural Networks and Context Transfer for Street Scenes Labeling | Street scene understanding is an essential task for autonomous driving. One important step towards this direction is scene labeling, which annotates each pixel in the images with a correct class label. Although many approaches have been developed, there are still some weak points. Firstly, many methods are based on the hand-crafted features whose image representation ability is limited. Secondly, they can not label foreground objects accurately due to the dataset bias. Thirdly, in the refinement stage, the traditional Markov Random Filed (MRF) inference is prone to over smoothness. For improving the above problems, this paper proposes a joint method of priori convolutional neural networks at superpixel level (called as ``priori s-CNNs'') and soft restricted context transfer. Our contributions are threefold: (1) A priori s-CNNs model that learns priori location information at superpixel level is proposed to describe various objects discriminatingly; (2) A hierarchical data augmentation method is presented to alleviate dataset bias in the priori s-CNNs training stage, which improves foreground objects labeling significantly; (3) A soft restricted MRF energy function is defined to improve the priori s-CNNs model's labeling performance and reduce the over smoothness at the same time. The proposed approach is verified on CamVid dataset (11 classes) and SIFT Flow Street dataset (16 classes) and achieves competitive performance. | ['Junyu. Gao', 'Qi. Wang', 'Yuan Yuan'] | 2019-05-05 | null | null | null | null | ['scene-labeling'] | ['computer-vision'] | [ 5.47439992e-01 1.26920134e-01 -5.17637789e-01 -6.74035072e-01
-4.32514757e-01 -1.32910535e-01 5.38377404e-01 -6.32893369e-02
-3.36619526e-01 7.01034009e-01 -4.24236357e-02 -2.22501025e-01
1.13134302e-01 -1.12044585e+00 -8.19242060e-01 -6.92130923e-01
3.23187947e-01 1.39475748e-01 8.43701780e-01 -1.18050063e-02
3.27529848e-01 4.03234214e-01 -1.71329498e+00 2.97191799e-01
1.22612071e+00 1.21034098e+00 5.30765533e-01 3.13986242e-01
-5.61545253e-01 1.09982800e+00 -2.66957253e-01 9.33910906e-03
2.60079026e-01 -2.43940115e-01 -8.76531839e-01 4.35495675e-01
5.59847593e-01 -4.01801080e-01 -8.21701288e-02 1.33853245e+00
-1.11849923e-02 1.69828981e-02 5.57913482e-01 -1.22201562e+00
-4.01880950e-01 3.70803297e-01 -7.71664321e-01 1.70127302e-01
-4.16245788e-01 2.68368512e-01 7.87365615e-01 -7.33032763e-01
3.83447111e-01 1.24282491e+00 5.67070305e-01 3.04579645e-01
-1.00075746e+00 -7.80084848e-01 5.72233915e-01 3.08677346e-01
-1.46228051e+00 -1.75824776e-01 1.05296409e+00 -6.70385540e-01
4.84521717e-01 1.23255067e-02 7.18011022e-01 6.73139155e-01
3.87437060e-04 9.26098049e-01 1.20498168e+00 -1.27892658e-01
2.78964847e-01 1.63360238e-01 3.26078027e-01 7.78904557e-01
2.56840616e-01 2.12730989e-02 -2.49615073e-01 3.83252859e-01
9.16604459e-01 3.58493663e-02 -6.75827935e-02 -3.71649981e-01
-1.06702566e+00 7.28287756e-01 7.45025396e-01 2.96970069e-01
-3.20255578e-01 2.73608118e-01 2.67944157e-01 -3.47381264e-01
5.41434288e-01 -4.77980934e-02 -6.70661449e-01 1.97105825e-01
-1.09454370e+00 4.52681743e-02 2.67943710e-01 9.81560230e-01
1.18900037e+00 1.81443170e-01 -1.90964267e-01 7.41948426e-01
5.10544896e-01 5.40669382e-01 8.72420594e-02 -8.84893775e-01
5.59090793e-01 8.08430910e-01 3.87426582e-03 -1.10897708e+00
-3.22706193e-01 -5.89834213e-01 -9.11024511e-01 4.26673979e-01
3.77019107e-01 2.95757800e-02 -1.31308270e+00 1.45467198e+00
4.31971133e-01 5.38977623e-01 -3.81720029e-02 9.03478324e-01
1.05896759e+00 7.59828448e-01 4.09711748e-01 6.67954385e-02
1.25178528e+00 -1.03902888e+00 -7.56682217e-01 -4.29848254e-01
4.15746957e-01 -5.90926766e-01 1.01344132e+00 2.18972459e-01
-8.04588795e-01 -1.05649221e+00 -1.04157591e+00 -1.32870331e-01
-5.86650550e-01 3.41625661e-01 7.91295469e-01 7.32153118e-01
-7.43520498e-01 2.02788353e-01 -7.19805598e-01 -1.45335540e-01
9.32006836e-01 2.71919161e-01 1.00938238e-01 -1.16153114e-01
-1.16623569e+00 6.53922379e-01 8.28704596e-01 3.76547188e-01
-9.55820322e-01 -5.40733278e-01 -8.83115351e-01 -7.89959449e-03
4.18213665e-01 -4.84909713e-01 8.01770985e-01 -1.19578874e+00
-1.42133558e+00 7.78235316e-01 -1.84956774e-01 -3.56431305e-01
5.16021192e-01 -8.87589380e-02 -2.83227831e-01 -2.43656281e-02
3.06335062e-01 1.09348571e+00 8.71285081e-01 -1.59387064e+00
-1.16737819e+00 -6.66165650e-02 1.75125673e-01 1.49166852e-01
-4.15813588e-02 -3.85443866e-01 -5.27052820e-01 -6.83747232e-01
3.12422216e-01 -6.54674590e-01 -4.81166929e-01 -9.85537842e-03
-4.52904969e-01 -2.48265252e-01 1.07321513e+00 -5.05276203e-01
1.03222036e+00 -2.00374675e+00 -2.11650163e-01 9.41997394e-02
4.41386132e-03 5.62035620e-01 8.05482492e-02 -1.79572970e-01
6.06968403e-02 7.09977150e-02 -6.47328675e-01 -2.08990633e-01
-1.93832308e-01 4.65746820e-01 -3.27107280e-01 3.66364866e-01
5.52237093e-01 8.15191269e-01 -8.78818750e-01 -8.72093379e-01
7.60146141e-01 5.00506163e-01 -5.62674761e-01 -2.31837621e-04
-4.88275766e-01 7.53022969e-01 -5.96271574e-01 7.28245020e-01
1.08940411e+00 -1.86882809e-01 -2.51702666e-01 -5.22101402e-01
-5.16585410e-01 6.13256022e-02 -1.50941670e+00 1.49131238e+00
-2.78473526e-01 5.28538764e-01 -9.71128345e-02 -1.09573770e+00
9.62379277e-01 -4.05665077e-02 4.49425578e-01 -5.86494982e-01
2.08751023e-01 1.52675495e-01 -2.93794096e-01 -5.97864509e-01
5.18000662e-01 2.81985067e-02 2.11074725e-01 -2.20990092e-01
-1.16918191e-01 -2.53207535e-01 7.48515427e-02 -1.72892921e-02
5.83238661e-01 5.66392541e-01 -3.63568007e-03 -5.00145674e-01
8.82518232e-01 3.68663073e-01 1.04185283e+00 5.58157086e-01
-3.49847794e-01 6.11637950e-01 1.45857006e-01 -6.06531799e-01
-8.66502225e-01 -7.80482113e-01 -4.28714216e-01 7.39761889e-01
7.41567850e-01 -3.17135490e-02 -7.73992062e-01 -7.89891183e-01
-1.36980191e-01 6.08228743e-01 -6.38047040e-01 6.56277835e-02
-5.59758127e-01 -7.49448895e-01 4.10141647e-01 7.69285142e-01
1.26159954e+00 -9.36014950e-01 -6.38771117e-01 1.07822366e-01
-2.96603143e-01 -1.22876108e+00 -3.18944663e-01 1.73860207e-01
-9.15258646e-01 -1.12862587e+00 -5.87884843e-01 -9.17017400e-01
8.69717062e-01 5.16095817e-01 7.43103147e-01 1.30985558e-01
-1.53149113e-01 -1.81521952e-01 -3.85303080e-01 -3.80801141e-01
-2.78433058e-02 1.24910727e-01 -3.84225965e-01 4.31602538e-01
3.60020250e-01 -2.80554742e-01 -6.56650245e-01 2.82992095e-01
-9.13911581e-01 5.46720982e-01 8.64952326e-01 6.89516485e-01
8.71135473e-01 4.64226693e-01 3.35070699e-01 -1.09263682e+00
-3.17343444e-01 -3.24902564e-01 -8.07792962e-01 3.28473561e-02
-4.75800872e-01 -3.73805128e-02 4.80788410e-01 -1.72070011e-01
-1.40215504e+00 4.76524174e-01 -3.42703462e-01 -2.68908978e-01
-7.24276006e-01 1.62243813e-01 -4.38414127e-01 -1.43102735e-01
2.95970619e-01 2.80093372e-01 -3.63929808e-01 -4.18371856e-01
4.45541680e-01 6.50363624e-01 5.89162767e-01 -4.50852424e-01
9.75328743e-01 8.29468846e-01 8.93701687e-02 -9.49678719e-01
-1.17094994e+00 -5.88824809e-01 -1.05183351e+00 -4.36748981e-01
1.33129644e+00 -1.05959594e+00 -4.97954458e-01 5.97163260e-01
-1.03793228e+00 -6.32679582e-01 -1.36055678e-01 4.70190138e-01
-4.22282994e-01 3.51458311e-01 -3.12906742e-01 -8.60575199e-01
1.77646545e-03 -1.22057402e+00 1.13841558e+00 6.05534613e-01
3.35163921e-01 -8.52891266e-01 -4.04533595e-01 5.57984173e-01
3.75985563e-01 4.62768883e-01 7.19401240e-01 -5.44892251e-02
-9.97397900e-01 1.41893819e-01 -8.11981618e-01 5.43405652e-01
2.18980700e-01 1.03664167e-01 -1.07291949e+00 1.81889236e-01
-2.28253022e-01 -3.53430733e-02 1.15954030e+00 6.54861212e-01
1.30568838e+00 -5.02425693e-02 -4.11540180e-01 6.95855200e-01
1.58698571e+00 3.01049978e-01 8.30659986e-01 4.69266325e-01
1.10524106e+00 6.90588534e-01 9.59395587e-01 1.67539954e-01
5.40872574e-01 4.60666806e-01 7.23115563e-01 -5.62257409e-01
-5.51267445e-01 -3.34702462e-01 1.71014190e-01 5.21779954e-01
-3.54688615e-02 -6.41227886e-02 -7.43329644e-01 7.49824882e-01
-1.84242260e+00 -8.44117403e-01 -7.70864248e-01 1.80663764e+00
6.03449643e-01 4.09665793e-01 -9.79137123e-02 2.97659844e-01
8.85816753e-01 2.53547728e-01 -5.45431018e-01 8.98200199e-02
-2.35341847e-01 5.69998100e-02 8.88348818e-01 6.74619675e-01
-1.51312339e+00 1.26297677e+00 4.70881605e+00 1.04483140e+00
-1.11849618e+00 9.57380384e-02 7.13278651e-01 6.13908231e-01
4.86813560e-02 1.93470329e-01 -1.16335416e+00 5.88311076e-01
2.07211748e-01 5.28574109e-01 -1.22170210e-01 1.01585054e+00
3.02963197e-01 -3.74774516e-01 -4.17292237e-01 8.74648690e-01
-1.01813242e-01 -1.35023558e+00 6.44399971e-02 4.82225008e-02
9.67643142e-01 2.50304937e-02 -3.33198421e-02 1.86248854e-01
3.46446961e-01 -8.35204244e-01 9.15746808e-01 5.06591976e-01
7.03281641e-01 -8.14445972e-01 7.59164274e-01 3.92622083e-01
-1.58342063e+00 2.88449153e-02 -4.91230249e-01 5.67073422e-03
2.57569909e-01 7.72213638e-01 -5.71929932e-01 6.56509817e-01
7.54679918e-01 1.06885934e+00 -7.40804136e-01 1.07182419e+00
-3.95377874e-01 8.15418005e-01 -1.83291793e-01 2.01148480e-01
5.35229564e-01 -3.46143007e-01 1.98944643e-01 1.28120458e+00
-1.68908194e-01 2.10094631e-01 4.36766028e-01 1.05888748e+00
2.67105788e-01 -3.62321958e-02 -3.52268815e-01 4.05693948e-01
1.76388666e-01 1.32456672e+00 -1.20324075e+00 -5.05020499e-01
-3.71290863e-01 8.56991112e-01 -5.43214865e-02 3.73806804e-01
-9.44997787e-01 -3.55258584e-01 4.76093382e-01 1.69531301e-01
4.11784112e-01 -2.69343138e-01 -6.89799845e-01 -9.82658744e-01
-2.87954330e-01 -2.68739492e-01 2.05132678e-01 -7.22092688e-01
-1.01042914e+00 4.17167813e-01 -7.00563833e-04 -1.32112217e+00
4.32799071e-01 -5.70426464e-01 -4.17332828e-01 7.16121078e-01
-2.20752358e+00 -1.43028033e+00 -7.62418807e-01 5.70839345e-01
9.63278770e-01 2.55918741e-01 2.18316406e-01 6.42721176e-01
-6.74898028e-01 9.71344635e-02 -1.29203826e-01 2.79314548e-01
3.78533125e-01 -1.11453509e+00 1.83036894e-01 1.07216942e+00
-1.48072794e-01 2.96441346e-01 4.57774162e-01 -8.13181579e-01
-8.35305989e-01 -1.82836521e+00 6.95352793e-01 -2.27467597e-01
3.09517890e-01 -1.16614431e-01 -9.01206911e-01 5.36314547e-01
-1.30924493e-01 1.76037446e-01 1.72151476e-01 -3.38948160e-01
1.52848795e-01 -2.37856314e-01 -1.13221622e+00 4.33869988e-01
9.29169774e-01 -2.47557715e-01 -4.60626394e-01 3.29868674e-01
6.58827960e-01 -4.21376735e-01 -6.15772963e-01 7.25139380e-01
1.68669119e-01 -8.98906410e-01 9.52871859e-01 -6.25982210e-02
4.32364374e-01 -1.01528049e+00 -1.58506379e-01 -6.43653572e-01
-3.28754485e-01 -4.40933332e-02 1.02292337e-01 1.51681411e+00
3.42514627e-02 -4.30109173e-01 9.82244909e-01 3.78839493e-01
-3.68196726e-01 -6.47498727e-01 -7.40647554e-01 -7.29552865e-01
-9.79297832e-02 -5.77140331e-01 6.17826700e-01 9.81190622e-01
-8.59802604e-01 1.89105600e-01 -3.36483300e-01 4.81481105e-01
7.10948169e-01 9.73306522e-02 8.90745640e-01 -1.33052230e+00
1.86730206e-01 -4.47985590e-01 -3.26347858e-01 -1.41405499e+00
2.76204720e-02 -7.71609724e-01 3.60626608e-01 -1.71620834e+00
4.08830605e-02 -9.61455643e-01 -1.77349180e-01 4.01602536e-01
-3.22642177e-01 4.05442566e-01 -6.39440939e-02 1.58373654e-01
-6.26981974e-01 6.03323579e-01 1.35343027e+00 -1.98242351e-01
-1.82271823e-01 6.26160577e-02 -3.20139170e-01 9.47664797e-01
8.40463519e-01 -4.30555552e-01 -3.60999912e-01 -3.25193256e-01
-1.68351620e-01 -3.02698374e-01 6.60694897e-01 -1.23306608e+00
1.53954655e-01 -4.62206602e-01 5.01003385e-01 -1.12111378e+00
1.80090234e-01 -9.45038855e-01 -3.39451991e-02 5.59661627e-01
-6.99031726e-02 -5.21726608e-01 1.33968785e-01 6.19051278e-01
-1.80590242e-01 -1.43930316e-01 1.16744161e+00 -1.77944794e-01
-1.27344263e+00 4.53114569e-01 -2.95289606e-01 -2.82008443e-02
1.10845006e+00 -5.43484569e-01 -9.17611420e-02 9.79927257e-02
-2.12976336e-01 3.81013542e-01 3.47719073e-01 3.10598910e-01
4.75629747e-01 -1.17314553e+00 -3.97467971e-01 3.21249962e-01
9.29542929e-02 6.54428184e-01 4.04154181e-01 8.08943510e-01
-6.99230492e-01 5.00127017e-01 -1.29815906e-01 -8.55976284e-01
-9.76841450e-01 5.03700614e-01 2.77700692e-01 -9.99006815e-03
-6.41901016e-01 8.95475566e-01 4.98847842e-01 -3.22209269e-01
1.46996930e-01 -6.44181013e-01 -5.17448425e-01 -2.45806121e-04
2.74908066e-01 2.11585060e-01 -1.48141354e-01 -1.18343461e+00
-2.67800152e-01 9.21179712e-01 1.61096171e-01 1.82083219e-01
1.14528894e+00 -3.83984685e-01 8.58383179e-02 3.11372727e-01
9.00987089e-01 -1.75405174e-01 -1.66802680e+00 -2.56174743e-01
-1.30114317e-01 -7.05698013e-01 3.57016891e-01 -5.73467076e-01
-1.33470058e+00 1.08866322e+00 7.47417927e-01 -3.82250324e-02
1.08487952e+00 -1.93475783e-01 9.67980444e-01 -1.50363753e-02
3.33231479e-01 -1.27592599e+00 -1.50875151e-01 4.55500603e-01
3.08843225e-01 -1.50262272e+00 -1.09099105e-01 -7.77817309e-01
-7.18932629e-01 1.15890849e+00 7.46386588e-01 -1.79477111e-01
6.77106440e-01 -1.94488205e-02 1.23257563e-01 -2.21185923e-01
-4.28763255e-02 -7.92599916e-01 3.10706168e-01 6.23100698e-01
1.07524775e-01 -1.35280877e-01 -1.61139667e-01 4.02000964e-01
2.32621118e-01 1.76251933e-01 2.40752786e-01 7.64595211e-01
-7.05262244e-01 -8.47368777e-01 -3.82525772e-01 3.58043432e-01
-2.68465430e-01 5.11427410e-03 7.10372105e-02 6.96293890e-01
9.47620869e-01 1.07731807e+00 7.29824677e-02 -2.78538972e-01
1.83580324e-01 -2.29860514e-01 8.31162408e-02 -6.47100627e-01
-2.21664622e-01 1.20215456e-03 -2.49975637e-01 -4.79339600e-01
-8.53490531e-01 -6.10428274e-01 -1.39917982e+00 4.16804440e-02
-5.42857409e-01 -7.81363770e-02 6.49127245e-01 1.03226709e+00
4.69214059e-02 7.46893108e-01 5.41201890e-01 -9.89178777e-01
2.44352415e-01 -6.91175640e-01 -5.09378374e-01 3.44421536e-01
3.15983146e-01 -8.95986855e-01 -1.64628267e-01 4.42453146e-01] | [9.50399112701416, -0.4539680778980255] |
ebae8736-3ed7-4eca-ae8e-740147df3ed8 | object-level-targeted-selection-via-deep | 2207.01778 | null | https://arxiv.org/abs/2207.01778v1 | https://arxiv.org/pdf/2207.01778v1.pdf | Object-Level Targeted Selection via Deep Template Matching | Retrieving images with objects that are semantically similar to objects of interest (OOI) in a query image has many practical use cases. A few examples include fixing failures like false negatives/positives of a learned model or mitigating class imbalance in a dataset. The targeted selection task requires finding the relevant data from a large-scale pool of unlabeled data. Manual mining at this scale is infeasible. Further, the OOI are often small and occupy less than 1% of image area, are occluded, and co-exist with many semantically different objects in cluttered scenes. Existing semantic image retrieval methods often focus on mining for larger sized geographical landmarks, and/or require extra labeled data, such as images/image-pairs with similar objects, for mining images with generic objects. We propose a fast and robust template matching algorithm in the DNN feature space, that retrieves semantically similar images at the object-level from a large unlabeled pool of data. We project the region(s) around the OOI in the query image to the DNN feature space for use as the template. This enables our method to focus on the semantics of the OOI without requiring extra labeled data. In the context of autonomous driving, we evaluate our system for targeted selection by using failure cases of object detectors as OOI. We demonstrate its efficacy on a large unlabeled dataset with 2.2M images and show high recall in mining for images with small-sized OOI. We compare our method against a well-known semantic image retrieval method, which also does not require extra labeled data. Lastly, we show that our method is flexible and retrieves images with one or more semantically different co-occurring OOI seamlessly. | ['Christoph Angerer', 'Jose M. Alvarez', 'Elmar Haussmann', 'Michele Fenzi', 'Donna Roy', 'Suraj Kothawade'] | 2022-07-05 | null | null | null | null | ['template-matching'] | ['computer-vision'] | [ 4.22805011e-01 -4.90719117e-02 -4.16240036e-01 -6.25082612e-01
-9.75227594e-01 -6.36830866e-01 2.66429514e-01 1.52516291e-01
-5.26277363e-01 2.73265392e-01 -1.84320942e-01 4.65030149e-02
-4.61660177e-01 -8.61130357e-01 -9.50464785e-01 -4.88090008e-01
3.43797840e-02 7.04601407e-01 5.78594565e-01 3.96578237e-02
2.86028832e-01 5.93958437e-01 -2.36760926e+00 1.95260465e-01
6.13632262e-01 1.18886924e+00 5.53807139e-01 2.01600954e-01
-3.34723592e-01 4.25421476e-01 -7.96911180e-01 1.04981542e-01
7.85845339e-01 -5.05014621e-02 -6.26870275e-01 4.35318857e-01
1.14708781e+00 -1.45663619e-01 -2.48779744e-01 1.09625483e+00
2.70902604e-01 2.07884446e-01 5.57506919e-01 -1.73832774e+00
-2.70631850e-01 -5.23672067e-02 -7.13810921e-01 3.67740363e-01
-5.60354963e-02 -1.62936419e-01 8.26488137e-01 -1.01246083e+00
9.04033422e-01 1.29544878e+00 3.14646214e-01 3.28349203e-01
-1.09020174e+00 -1.06789339e+00 1.03378654e-01 2.91557252e-01
-1.84287190e+00 -4.65564996e-01 5.24130523e-01 -1.61980018e-01
8.94390643e-01 3.34666222e-01 3.50872695e-01 4.20820624e-01
-5.17340302e-02 8.80627394e-01 6.53221130e-01 -3.99147362e-01
3.68181109e-01 3.24879706e-01 1.50492534e-01 5.43663442e-01
3.71977717e-01 1.07402161e-01 -8.09535027e-01 -2.19198629e-01
2.95225471e-01 3.82012516e-01 1.02045588e-01 -5.76036751e-01
-1.07095313e+00 9.01040912e-01 4.62213606e-01 4.19650376e-02
-1.18395559e-01 1.43594027e-01 1.60534874e-01 4.03562933e-01
3.85989636e-01 3.21620524e-01 -5.45412302e-01 3.55834633e-01
-1.01815879e+00 3.94411325e-01 4.80274320e-01 1.51758814e+00
1.53157973e+00 -3.96845222e-01 1.65675417e-01 9.11674321e-01
1.25280231e-01 7.88648963e-01 3.99180472e-01 -1.19752753e+00
5.50401092e-01 8.95198166e-01 3.31520140e-02 -1.04462087e+00
-2.61584103e-01 -1.15257084e-01 -2.87623078e-01 3.35072368e-01
2.51446754e-01 2.41085291e-01 -1.16650987e+00 1.71995425e+00
4.67290610e-01 -3.97472903e-02 2.19399795e-01 9.66638267e-01
7.41245151e-01 3.86371881e-01 7.49149919e-02 9.78904814e-02
1.41923475e+00 -7.16394842e-01 -2.78088361e-01 -6.41157746e-01
6.81874990e-01 -7.68929124e-01 1.12891197e+00 1.26626655e-01
-6.31461799e-01 -6.31249487e-01 -8.88054430e-01 2.69986987e-02
-7.17546940e-01 1.51632920e-01 4.83649254e-01 4.45304543e-01
-8.64566386e-01 1.56288207e-01 -1.88528374e-01 -7.32173026e-01
5.41380763e-01 4.88685489e-01 -5.42915225e-01 -4.69642967e-01
-9.08393741e-01 6.80712163e-01 5.88259697e-01 -3.48642260e-01
-1.00027239e+00 -7.10827053e-01 -9.94113803e-01 -1.15114428e-01
8.36213410e-01 -3.31330597e-01 8.96919847e-01 -1.25916469e+00
-2.51536787e-01 1.11350834e+00 -4.47132051e-01 -4.80746418e-01
1.28700972e-01 -1.84521645e-01 -6.14216387e-01 3.54825377e-01
8.10358047e-01 1.46582067e+00 9.96161401e-01 -1.34547794e+00
-1.31932485e+00 -7.12909698e-01 -3.32810241e-03 3.16246063e-01
-3.52427274e-01 2.87013128e-02 -7.76366353e-01 -5.50182104e-01
4.14509177e-01 -1.08912420e+00 -1.59645602e-01 3.06070238e-01
-1.48801759e-01 -3.53099436e-01 1.33550048e+00 6.54759351e-03
9.19098616e-01 -2.37645006e+00 -4.85974401e-01 6.22861922e-01
1.21286260e-02 -1.39624506e-01 -3.82268012e-01 7.75744347e-03
6.09612241e-02 1.01528764e-01 -1.65509760e-01 2.94714626e-02
-2.53830016e-01 6.56768203e-01 -3.19319010e-01 2.94336468e-01
3.00470918e-01 7.36136198e-01 -9.47249234e-01 -8.12190354e-01
1.68348834e-01 -2.73614645e-01 -2.77293593e-01 -9.96740982e-02
-1.59427255e-01 -3.61991785e-02 -4.40383464e-01 9.20234680e-01
5.34747899e-01 -1.07292458e-01 -1.68988183e-01 -2.87270546e-01
1.18957184e-01 -1.60791069e-01 -1.46877170e+00 1.64406264e+00
-1.38136312e-01 6.47481322e-01 -6.97351024e-02 -1.04893363e+00
1.05846667e+00 -1.01242535e-01 6.44952893e-01 -1.05896640e+00
-3.83702427e-01 3.57821167e-01 -2.30230778e-01 -5.81423879e-01
6.28070414e-01 1.28603399e-01 -2.65463561e-01 5.89664400e-01
-1.38129309e-01 -1.90412462e-01 3.85575384e-01 2.99342662e-01
1.19700444e+00 -2.16108531e-01 3.62422690e-02 -5.02002418e-01
1.65684432e-01 5.89641333e-01 7.97129512e-01 1.23188066e+00
-2.82587528e-01 7.53920674e-01 -1.35749131e-01 -5.52282274e-01
-1.15528393e+00 -1.03929520e+00 -3.69226664e-01 1.08500671e+00
7.99266100e-01 -2.37279519e-01 -4.57536638e-01 -6.98190987e-01
3.12770098e-01 4.66132879e-01 -4.70692217e-01 -1.80858642e-01
-3.67408276e-01 -4.56610769e-01 3.49090397e-01 4.92509872e-01
5.18601000e-01 -1.13643610e+00 -8.56612146e-01 -3.49454135e-02
-2.09995523e-01 -1.04455853e+00 -3.45641851e-01 4.02029574e-01
-6.91199243e-01 -1.29802072e+00 -3.30284327e-01 -1.03399694e+00
9.93929088e-01 1.00690269e+00 1.28154278e+00 1.73437163e-01
-7.34627306e-01 6.04383051e-01 -2.36954123e-01 -6.64110959e-01
-1.14215113e-01 -7.90363997e-02 1.67806447e-01 -1.74461797e-01
9.76915658e-01 -9.49378461e-02 -5.26874065e-01 9.64318395e-01
-1.19769633e+00 -2.32855693e-01 4.90167528e-01 6.05794847e-01
1.02443051e+00 5.23899853e-01 6.76001430e-01 -7.78033733e-01
2.85502076e-02 -5.57935059e-01 -6.97771430e-01 3.10705602e-01
-4.96812493e-01 2.37055086e-02 -2.22812239e-02 -5.59448123e-01
-7.18271613e-01 4.44179773e-01 6.31676555e-01 -7.36352682e-01
-3.63032281e-01 2.87463397e-01 -2.61385173e-01 4.78340872e-02
8.70436668e-01 -6.13003299e-02 4.72196825e-02 -2.46933207e-01
2.44072348e-01 7.42181420e-01 5.65932572e-01 -5.07851124e-01
8.81527007e-01 9.09820914e-01 -1.94440648e-01 -7.24295080e-01
-9.25277948e-01 -1.03570116e+00 -4.76875901e-01 -1.27752900e-01
5.02161026e-01 -1.13143802e+00 -1.05698504e-01 -3.17374319e-02
-5.79110801e-01 -9.78856012e-02 -4.25708503e-01 4.61259961e-01
-5.24393559e-01 9.72142518e-02 1.86593041e-01 -6.02460325e-01
-1.05829068e-01 -1.17762792e+00 1.39963925e+00 2.25129634e-01
-1.34046271e-01 -3.34206820e-01 -4.56273228e-01 4.33435291e-01
1.19586773e-01 -1.78686813e-01 8.25635552e-01 -7.62459397e-01
-8.23889792e-01 -2.65046686e-01 -4.40232724e-01 1.93452761e-01
2.74052680e-01 -2.95600295e-01 -1.03197932e+00 -3.35410833e-01
-2.22424820e-01 -5.56391478e-01 9.26974595e-01 2.12466329e-01
1.37348425e+00 6.24322183e-02 -6.88817561e-01 1.90122947e-01
1.43022048e+00 3.53909403e-01 5.01340270e-01 5.08626580e-01
4.46381330e-01 1.01382160e+00 1.37956643e+00 2.84281045e-01
1.32253408e-01 6.96154892e-01 4.02979791e-01 -3.02977681e-01
5.64804720e-03 -6.29843399e-02 3.83735895e-02 -8.23189840e-02
5.72928309e-01 -9.07110721e-02 -9.78117704e-01 1.08389294e+00
-1.94825935e+00 -7.94478536e-01 1.30118370e-01 2.39332604e+00
5.45891285e-01 2.15538472e-01 -3.69016523e-03 -6.78434670e-02
7.66266882e-01 -3.40776712e-01 -1.00432861e+00 1.77313253e-01
-1.83135733e-01 7.58600682e-02 1.10043597e+00 2.07943872e-01
-1.27872717e+00 9.82621431e-01 5.93201780e+00 1.02968955e+00
-7.94687688e-01 -2.36947369e-02 7.00049698e-01 -4.23192173e-01
-2.23597020e-01 8.80704299e-02 -1.17255902e+00 2.06185281e-01
5.82392037e-01 -4.91467975e-02 1.76733509e-01 1.09889650e+00
2.67296404e-01 -5.71686983e-01 -1.13457167e+00 1.23650539e+00
2.25122586e-01 -1.19684577e+00 7.92545527e-02 2.07139358e-01
8.12845707e-01 3.44487906e-01 -9.53952521e-02 1.13888182e-01
2.13613331e-01 -8.83242249e-01 9.06065822e-01 2.57857382e-01
7.83086777e-01 -8.05881560e-01 5.38285434e-01 3.57334614e-01
-1.23675013e+00 -2.97062814e-01 -7.44314015e-01 3.26996416e-01
-3.16941172e-01 5.17840326e-01 -6.56645358e-01 2.00890064e-01
1.33861828e+00 6.71552777e-01 -8.64286363e-01 9.68750477e-01
1.93454757e-01 3.36555570e-01 -8.55806589e-01 3.49310637e-01
2.75525033e-01 -1.38769264e-03 4.05324101e-01 7.98100293e-01
2.80218750e-01 -3.45024541e-02 5.08068919e-01 8.30582678e-01
-4.16626260e-02 -2.01663915e-02 -1.11516786e+00 3.68023306e-01
8.21365297e-01 1.25319707e+00 -9.60635900e-01 -6.48220897e-01
-4.89315450e-01 5.82930982e-01 -3.55848707e-02 3.60506177e-01
-6.73524916e-01 -3.83677632e-01 8.06676269e-01 1.38663635e-01
4.52451557e-01 1.16974570e-01 -1.36747947e-02 -8.47286820e-01
3.21411312e-01 -8.39652240e-01 6.27713442e-01 -8.39767337e-01
-1.20800316e+00 3.15481156e-01 2.97453731e-01 -1.57756221e+00
-2.05045640e-01 -4.99771744e-01 -1.24214306e-01 5.89181423e-01
-1.63118994e+00 -9.36967492e-01 -6.33919060e-01 8.67325068e-01
7.57814944e-01 -2.76592135e-01 6.11266315e-01 5.42927206e-01
-2.06155732e-01 3.50186467e-01 -1.20430887e-02 4.85598892e-02
8.65525424e-01 -8.90320659e-01 1.91827953e-01 6.55523121e-01
4.71103400e-01 4.94464070e-01 3.72638702e-01 -7.27373242e-01
-1.18972588e+00 -1.49055302e+00 7.66996920e-01 -3.97278696e-01
3.14344287e-01 -3.77249300e-01 -7.97335386e-01 5.04635990e-01
-3.53517324e-01 2.37902686e-01 4.98110116e-01 -2.34523699e-01
-3.97461891e-01 -4.81573731e-01 -1.32243180e+00 4.11131471e-01
1.26025450e+00 -4.79734570e-01 -4.75025207e-01 5.77845514e-01
5.77159345e-01 -9.43677351e-02 -6.27196729e-01 5.44467628e-01
3.67585808e-01 -6.40983403e-01 1.15241110e+00 -4.86802399e-01
-8.01607296e-02 -8.53837550e-01 -4.36264396e-01 -7.60042906e-01
7.30758533e-02 2.87198406e-02 5.32282770e-01 9.83607769e-01
4.15001929e-01 -3.42350721e-01 9.92225647e-01 9.38443601e-01
-5.14694229e-02 -2.24791989e-01 -9.69603062e-01 -1.10779583e+00
-4.79225844e-01 -6.92237556e-01 5.57437479e-01 7.51125455e-01
-5.24971187e-01 7.30242655e-02 2.60332618e-02 4.09659058e-01
6.68926597e-01 4.76003259e-01 1.11220729e+00 -1.40294993e+00
1.95234060e-01 -5.22154868e-02 -7.11937249e-01 -7.83395112e-01
1.68665424e-01 -8.66440594e-01 2.62623072e-01 -1.30965054e+00
3.83563042e-01 -9.16113615e-01 -2.38059700e-01 9.00368094e-01
7.87404627e-02 5.99832118e-01 3.59482318e-02 6.05422080e-01
-7.70825982e-01 1.24901757e-01 6.55362368e-01 -6.56426847e-01
-1.32056758e-01 -1.61793530e-01 -5.64474165e-01 7.29105055e-01
8.23111236e-01 -9.05341566e-01 -6.67238355e-01 -3.85452539e-01
-3.43139246e-02 -4.15881395e-01 4.01382506e-01 -1.07102466e+00
1.62006944e-01 -2.53564686e-01 4.46894050e-01 -8.99534702e-01
2.17754647e-01 -1.19938719e+00 2.33190492e-01 1.90452054e-01
-3.05481732e-01 9.67195109e-02 2.46139795e-01 7.22577512e-01
-4.15425420e-01 -6.06693447e-01 7.04383254e-01 -3.44115943e-01
-1.47905743e+00 3.12473923e-01 -2.43542179e-01 1.44470185e-01
1.24643433e+00 -6.31680012e-01 -1.33620143e-01 -1.61602139e-01
-5.40427089e-01 5.05669057e-01 7.64860690e-01 8.34601521e-01
8.39754641e-01 -1.20548856e+00 -4.12667155e-01 3.25840443e-01
9.43200707e-01 2.76982546e-01 -1.25036404e-01 2.73870409e-01
-2.50227392e-01 1.76835164e-01 -1.74683537e-02 -1.05417645e+00
-1.53512764e+00 7.93278575e-01 2.80818701e-01 3.69281352e-01
-6.88649535e-01 4.76716936e-01 6.23111129e-01 -4.13089961e-01
3.53023469e-01 -2.81923860e-01 1.22464433e-01 1.67776868e-01
6.06675684e-01 -2.85143242e-03 2.23817080e-01 -7.30989039e-01
-5.54801047e-01 6.37088299e-01 -1.18573390e-01 -1.38260070e-02
1.12030804e+00 -3.38644892e-01 -3.57008688e-02 2.69634783e-01
1.34063137e+00 -2.06776589e-01 -1.15237892e+00 -5.45273960e-01
3.29546005e-01 -6.76162302e-01 5.76179586e-02 -4.68320787e-01
-1.29165244e+00 3.87617946e-01 9.95393634e-01 -1.12235740e-01
1.04761744e+00 4.98911828e-01 5.41675329e-01 7.80079424e-01
6.95570588e-01 -1.49885750e+00 1.81379259e-01 6.23348840e-02
5.34448445e-01 -1.63550019e+00 -4.24784087e-02 -2.93394357e-01
-5.20909190e-01 8.75676334e-01 7.94948280e-01 -6.43241331e-02
6.11217976e-01 1.59269422e-01 3.65923762e-01 -6.10025644e-01
-5.13603866e-01 -6.13066971e-01 2.95657367e-01 5.89818358e-01
-3.01180452e-01 -7.83311650e-02 4.24469188e-02 -1.13464311e-01
3.24508846e-02 -1.31988138e-01 1.52068391e-01 1.22587693e+00
-7.97535121e-01 -8.09117317e-01 -5.83693206e-01 7.52100706e-01
-5.78098223e-02 -3.65108773e-02 -3.57037753e-01 9.15124893e-01
3.85256916e-01 1.20325983e+00 5.54513991e-01 -1.62193909e-01
2.86027431e-01 -2.66656559e-02 1.52033970e-01 -6.92771018e-01
-1.68748409e-01 1.47229275e-02 -1.04184307e-01 -7.79262662e-01
-6.99130893e-01 -7.05121279e-01 -1.44837141e+00 1.75467029e-01
-5.20898521e-01 5.23083098e-03 6.64340556e-01 8.68946075e-01
6.39592230e-01 -6.86520562e-02 5.63460588e-01 -5.88942170e-01
-7.80274197e-02 -5.43770194e-01 -6.83212340e-01 9.02441859e-01
1.33674979e-01 -8.67380798e-01 -2.20401868e-01 1.81504890e-01] | [9.556206703186035, 1.6255439519882202] |
103ca730-6880-4354-bd6c-2b1283d04de9 | multi-scale-graphical-models-for-spatio | null | null | http://papers.nips.cc/paper/5473-multi-scale-graphical-models-for-spatio-temporal-processes | http://papers.nips.cc/paper/5473-multi-scale-graphical-models-for-spatio-temporal-processes.pdf | Multi-scale Graphical Models for Spatio-Temporal Processes | Learning the dependency structure between spatially distributed observations of a spatio-temporal process is an important problem in many fields such as geology, geophysics, atmospheric sciences, oceanography, etc. . However, estimation of such systems is complicated by the fact that they exhibit dynamics at multiple scales of space and time arising due to a combination of diffusion and convection/advection. As we show, time-series graphical models based on vector auto-regressive processes are inefficient in capturing such multi-scale structure. In this paper, we present a hierarchical graphical model with physically derived priors that better represents the multi-scale character of these dynamical systems. We also propose algorithms to efficiently estimate the interaction structure from data. We demonstrate results on a general class of problems arising in exploration geophysics by discovering graphical structure that is physically meaningful and provide evidence of its advantages over alternative approaches. | ['Firdaus Janoos', 'Niranjan Subrahmanya', 'Huseyin Denli'] | 2014-12-01 | null | null | null | neurips-2014-12 | ['geophysics'] | ['miscellaneous'] | [-1.09273940e-01 -4.52204853e-01 4.11130279e-01 -1.34267509e-01
-1.47242367e-01 -6.13579690e-01 9.74896431e-01 3.11452806e-01
1.68897659e-02 9.54166234e-01 3.67085189e-01 -5.74917316e-01
-9.19463754e-01 -7.38507450e-01 -4.93957400e-01 -9.60430741e-01
-9.16189075e-01 7.04952717e-01 2.15824962e-01 -1.95035145e-01
1.42674595e-01 1.00350654e+00 -1.06419504e+00 -4.87556636e-01
7.19374716e-01 4.36508954e-01 2.79745698e-01 8.71348441e-01
-3.99237834e-02 4.96888310e-01 -3.13665152e-01 2.53020227e-01
-1.59940179e-02 -5.35200179e-01 -3.95568073e-01 3.61243516e-01
1.06736057e-01 1.47990704e-01 -3.69109333e-01 9.53254223e-01
2.84110643e-02 4.26839173e-01 9.66045201e-01 -1.04331589e+00
-3.00535977e-01 2.42811635e-01 -7.50147521e-01 4.49386150e-01
-2.66811192e-01 7.39920810e-02 9.78010774e-01 -6.94842875e-01
3.71908754e-01 1.69963765e+00 6.46507442e-01 -1.79950148e-01
-1.73069167e+00 -3.84832501e-01 2.86788464e-01 -2.92461216e-01
-1.29430282e+00 -1.49156600e-01 7.96943188e-01 -9.18295622e-01
5.43427289e-01 3.05758893e-01 5.98862290e-01 8.80976558e-01
7.15535820e-01 4.21671808e-01 1.31622601e+00 -1.28792435e-01
4.73997891e-01 -2.13882446e-01 1.26310661e-01 7.21629500e-01
4.16691333e-01 1.72387198e-01 -5.62474191e-01 -7.51349747e-01
1.10748947e+00 2.10698545e-01 -2.20841035e-01 -3.56337398e-01
-8.09523761e-01 1.02081823e+00 3.00243080e-01 3.74909610e-01
-4.19676691e-01 3.63931090e-01 -2.13196680e-01 2.74393439e-01
7.30097234e-01 6.22303724e-01 -5.14235139e-01 3.81939746e-02
-1.08080864e+00 1.34761378e-01 8.99612010e-01 4.82543170e-01
6.88387811e-01 1.75407931e-01 3.34425956e-01 4.93020803e-01
9.11579967e-01 9.29303229e-01 8.50806460e-02 -5.88087916e-01
1.68550909e-01 2.16671973e-01 3.60362560e-01 -1.01696742e+00
-5.97064853e-01 -4.35667038e-01 -1.47039044e+00 4.23948765e-01
5.03929615e-01 -4.49074954e-01 -8.61118317e-01 1.63364577e+00
2.05680266e-01 4.94373947e-01 -6.01597363e-03 5.79494894e-01
1.87202230e-01 1.07407081e+00 8.51994976e-02 -3.28263104e-01
1.21112406e+00 -2.67374426e-01 -7.91120172e-01 -2.52489239e-01
2.48459596e-02 -3.34361166e-01 4.99763221e-01 2.17247054e-01
-1.04028666e+00 -2.42781088e-01 -5.73454559e-01 4.79051501e-01
-8.26065466e-02 -1.39262408e-01 7.59405255e-01 2.37328798e-01
-1.25562358e+00 9.64955866e-01 -1.49610579e+00 -3.08147877e-01
-6.65656552e-02 1.09470740e-01 -1.90885365e-01 5.41635036e-01
-1.16357303e+00 6.93249881e-01 -1.91530868e-01 3.75235379e-01
-8.65500331e-01 -5.22715151e-01 -7.65460014e-01 2.59611636e-01
9.40955207e-02 -5.24835289e-01 8.89095128e-01 -3.92683357e-01
-1.20973742e+00 1.14763685e-01 -4.21595722e-01 -1.70926794e-01
5.58026195e-01 -6.39062971e-02 -3.12017351e-01 -9.38263610e-02
-9.51755419e-02 -1.34133250e-01 9.35057700e-01 -1.28353012e+00
-3.00576419e-01 -3.43932807e-01 -3.81136119e-01 1.42882362e-01
9.36632305e-02 -4.44154255e-02 4.82295938e-02 -4.91047680e-01
3.86421740e-01 -1.19071186e+00 -8.50537002e-01 -6.08605184e-02
-5.62806249e-01 6.75813109e-02 7.51248062e-01 -5.29159367e-01
1.01522517e+00 -2.17928433e+00 5.70820928e-01 6.69686198e-01
3.03925812e-01 -2.30969787e-01 9.12253112e-02 1.15754032e+00
1.18841901e-01 4.56591636e-01 -8.55610728e-01 -4.18322086e-01
-1.23246811e-01 3.04016531e-01 -5.60153186e-01 7.72676826e-01
5.11176825e-01 4.96777207e-01 -9.02438164e-01 -6.98229074e-02
1.79795444e-01 3.92476410e-01 5.09236753e-03 1.66467696e-01
-3.91426943e-02 9.64870274e-01 -6.45963073e-01 2.00190172e-01
4.44725692e-01 -6.92078590e-01 3.48690063e-01 4.95480597e-01
-4.22221631e-01 -8.45077485e-02 -1.44488025e+00 1.13814449e+00
-3.31808895e-01 8.55675459e-01 4.99283701e-01 -9.27056193e-01
7.17627108e-01 3.90484959e-01 5.00295043e-01 3.15205343e-02
-1.83064356e-01 -5.32936677e-02 2.60134846e-01 -5.16251981e-01
3.59578848e-01 -5.48749328e-01 1.13874145e-01 3.79123330e-01
-2.09412187e-01 -3.99651706e-01 -6.44439608e-02 8.44736770e-02
1.28230727e+00 -1.54597282e-01 4.84903425e-01 -7.42139876e-01
2.07526028e-01 -2.61234492e-01 3.30981493e-01 9.65738177e-01
1.56670749e-01 2.75952160e-01 6.98642254e-01 -1.01676278e-01
-9.82482672e-01 -8.34271252e-01 -4.66243029e-01 4.73513693e-01
-1.15870655e-01 -3.14183742e-01 1.31704763e-01 2.45564599e-02
3.14685404e-01 3.81368399e-01 -7.90101171e-01 1.84949100e-01
-3.62126023e-01 -1.28924191e+00 1.92781463e-01 3.46496075e-01
1.24392115e-01 -8.04581285e-01 -4.25000608e-01 5.16347587e-01
1.45808488e-01 -7.85664082e-01 6.43574521e-02 2.94811487e-01
-1.36811244e+00 -6.94070458e-01 -5.23550212e-01 -9.43584517e-02
5.49537063e-01 9.24458727e-02 1.03686869e+00 -1.20514154e-01
-2.11335301e-01 4.30297256e-01 1.64212361e-01 -3.21587473e-01
-4.70850229e-01 -4.69549000e-01 1.15237162e-01 2.17195213e-01
-4.01435554e-01 -1.05983162e+00 -4.27813023e-01 2.91282773e-01
-1.14320683e+00 -1.65953357e-02 6.13206148e-01 6.73118830e-01
1.95035592e-01 3.58942389e-01 1.62220433e-01 -8.04968238e-01
7.64022589e-01 -9.04381812e-01 -1.08176279e+00 2.06012160e-01
-3.42312545e-01 5.77061236e-01 2.95931429e-01 -4.05744195e-01
-1.16133142e+00 -1.52133917e-02 3.63116831e-01 -2.85935611e-01
-2.80224621e-01 1.15405917e+00 2.12695569e-01 8.77149031e-03
3.40408236e-01 9.70782265e-02 2.15960685e-02 -4.54145432e-01
1.48650646e-01 1.27722265e-03 -1.50519339e-02 -6.63843572e-01
9.72696364e-01 6.99342549e-01 8.44686627e-01 -1.55856323e+00
-2.00323313e-01 -4.93544400e-01 -5.90305090e-01 -1.43753141e-01
7.28997648e-01 -8.85815084e-01 -3.65532696e-01 4.56390738e-01
-1.22532499e+00 -4.99797106e-01 -5.34513630e-02 9.09676790e-01
-1.89392090e-01 2.08159462e-01 -7.81511068e-01 -1.38151503e+00
3.77679974e-01 -7.89282918e-01 1.05643106e+00 3.95352654e-02
-1.55699283e-01 -1.95548618e+00 6.74993992e-01 -5.72029769e-01
4.96208191e-01 4.24019963e-01 8.81163001e-01 -2.88649291e-01
-6.97153151e-01 1.03615515e-01 -2.98527125e-02 -1.25891447e-01
2.40342066e-01 5.85182786e-01 -5.99165320e-01 -1.51561737e-01
2.46100113e-01 3.72317255e-01 9.10583317e-01 9.46363688e-01
7.03714967e-01 -1.71512946e-01 -4.65106934e-01 4.24283147e-01
1.43038547e+00 -2.22702362e-02 1.37905508e-01 -2.26884857e-01
7.52830386e-01 1.02405155e+00 5.74548542e-02 6.56918287e-01
1.04899500e-02 3.26051652e-01 2.43325979e-01 -1.28600940e-01
4.66685146e-01 1.05497010e-01 3.15024644e-01 8.28825235e-01
-4.08101290e-01 -4.52992827e-01 -1.29518342e+00 5.02777755e-01
-2.20632672e+00 -1.18397081e+00 -7.54039824e-01 2.23732710e+00
5.04635394e-01 -2.04364687e-01 -2.08457381e-01 -3.74236315e-01
6.77898228e-01 3.11952323e-01 -4.34501350e-01 6.35864586e-02
-2.55962670e-01 1.21763535e-01 7.02454031e-01 8.87924612e-01
-1.02122808e+00 3.97390485e-01 6.55230618e+00 2.62068063e-01
-1.18687463e+00 -2.06035629e-01 4.09639359e-01 4.17313218e-01
-4.78821933e-01 2.29336724e-01 -7.12771893e-01 2.10476726e-01
1.14908373e+00 -2.37285331e-01 6.02025390e-02 2.18868315e-01
7.74931669e-01 -3.03107172e-01 -9.31643248e-01 4.43690270e-01
-4.48733926e-01 -1.04131961e+00 -1.37078926e-01 6.25878274e-01
8.49894583e-01 1.57860309e-01 -1.10912561e-01 -2.62934625e-01
1.00217605e+00 -8.57464254e-01 4.85853702e-01 9.10307467e-01
2.35920832e-01 -3.69135469e-01 4.10428524e-01 4.41144675e-01
-1.28209889e+00 1.84359789e-01 -1.58200219e-01 -3.84766579e-01
5.84056914e-01 1.11917841e+00 -7.00857699e-01 5.12831092e-01
4.74338651e-01 9.05235767e-01 -3.80308449e-01 1.26434016e+00
-3.02213013e-01 1.14777279e+00 -6.69023633e-01 -4.87224348e-02
4.79510248e-01 -8.60489786e-01 7.97049046e-01 9.76349056e-01
7.82432795e-01 4.00903553e-01 9.38527659e-02 9.30565178e-01
2.50663221e-01 -2.31756374e-01 -9.21550930e-01 -9.10205171e-02
1.29598975e-01 1.21392930e+00 -9.13953006e-01 -2.94244230e-01
-3.89907002e-01 3.99724483e-01 1.06755108e-01 7.50008523e-01
-5.54555595e-01 2.02964559e-01 6.85847640e-01 1.18021913e-01
2.57184267e-01 -1.03298998e+00 6.46927506e-02 -1.16176021e+00
-3.70570987e-01 -2.94348836e-01 1.57077581e-01 -6.69958532e-01
-1.59041953e+00 3.44193995e-01 4.39543188e-01 -1.11396635e+00
-2.05974728e-01 -5.15569389e-01 -8.97252023e-01 1.20441628e+00
-1.41309977e+00 -6.69903457e-01 5.75623810e-02 3.51651400e-01
1.69702128e-01 3.10543001e-01 7.24866450e-01 -3.45075756e-01
-4.96976137e-01 -8.49350214e-01 1.00255322e+00 -2.24602535e-01
3.88077259e-01 -1.78809679e+00 6.75119340e-01 1.00567162e+00
3.40281099e-01 6.50889814e-01 1.05243289e+00 -1.11435437e+00
-1.35619223e+00 -9.48805809e-01 7.48619556e-01 -4.04239148e-01
1.47414303e+00 -4.27155405e-01 -1.19971156e+00 6.45280600e-01
3.04321498e-01 2.02900395e-01 7.41918743e-01 2.69526422e-01
1.51557803e-01 1.97875023e-01 -5.47713757e-01 4.61671054e-01
6.17360830e-01 -4.67199236e-01 -5.48795044e-01 5.17204702e-01
-1.60951074e-02 3.18019986e-02 -7.71708727e-01 1.85482085e-01
3.14191282e-01 -6.11887276e-01 6.47307694e-01 -8.38035703e-01
3.07445824e-01 -2.91524678e-01 1.35138139e-01 -1.76479208e+00
-6.56038582e-01 -7.38833189e-01 -2.20083725e-03 8.76252949e-01
4.43685621e-01 -8.93957973e-01 2.24536553e-01 5.57646632e-01
2.20308796e-01 -3.97242345e-02 -8.59242797e-01 -8.52910876e-01
1.14385836e-01 -2.40514606e-01 9.32750758e-03 1.04165244e+00
-1.93637088e-01 4.80318874e-01 -5.51326990e-01 9.33850944e-01
9.24321353e-01 2.13537514e-01 4.67968762e-01 -1.91956890e+00
-6.10435605e-01 -4.66465205e-01 -1.34416699e-01 -8.47190738e-01
1.13900691e-01 -3.14991534e-01 3.78422678e-01 -1.33077550e+00
-5.74130565e-02 -4.98696208e-01 -1.57347947e-01 -9.59065109e-02
-1.32239744e-01 -2.36672863e-01 -1.95708573e-01 6.21820211e-01
2.17426550e-02 6.32543623e-01 1.13984168e+00 -4.66907583e-03
-2.12846443e-01 2.02247605e-01 -1.32406196e-02 6.97489202e-01
6.39013052e-01 -7.04721570e-01 -4.44438219e-01 -2.77402401e-01
4.61777270e-01 8.27110410e-01 4.99181300e-01 -6.98777616e-01
4.17484045e-01 -6.27051830e-01 2.42556736e-01 -4.14613426e-01
3.85271430e-01 -8.81680906e-01 6.01761580e-01 3.89956146e-01
-2.06567168e-01 3.99198920e-01 4.44907039e-01 1.14939797e+00
-3.76193702e-01 -5.95059842e-02 3.55382591e-01 -2.38338694e-01
-3.03686589e-01 4.84411627e-01 -1.02142739e+00 -2.06149861e-01
6.94213450e-01 4.44254935e-01 -3.14466581e-02 -7.76209712e-01
-9.21759009e-01 2.94562101e-01 1.85543671e-01 9.06618908e-02
2.64046520e-01 -7.82733440e-01 -8.29500020e-01 3.79019491e-02
-3.24999511e-01 -9.61818397e-02 1.97086751e-01 1.00327528e+00
-4.04662818e-01 2.49675125e-01 -2.26666108e-02 -5.95031917e-01
-1.19742680e+00 6.14575148e-02 1.96338326e-01 -5.66008329e-01
-5.09857595e-01 6.52319968e-01 4.72821385e-01 -8.91568735e-02
-4.32048261e-01 -5.72515845e-01 -2.79245645e-01 6.36209399e-02
3.06990057e-01 3.70238394e-01 -2.61624873e-01 -5.88099957e-01
-3.17833960e-01 5.15970111e-01 4.69080716e-01 -6.46826982e-01
1.43656337e+00 -3.63038123e-01 -4.06241924e-01 1.25712335e+00
6.89728737e-01 2.06180617e-01 -1.60936666e+00 -2.95833975e-01
3.22477460e-01 -1.19188994e-01 1.33343935e-01 -3.02273691e-01
-6.91761851e-01 1.05889547e+00 1.31741941e-01 9.37276423e-01
4.87432420e-01 8.55963454e-02 -1.79801837e-01 5.15997410e-01
-1.17210988e-02 -4.65506405e-01 -2.36508742e-01 6.23875380e-01
1.05272090e+00 -1.03823495e+00 2.47477636e-01 -2.68743157e-01
-1.97047025e-01 1.33284855e+00 -2.51012500e-02 -3.67650747e-01
1.45331597e+00 3.79989594e-01 -1.82504535e-01 -5.77267289e-01
-8.92714560e-01 -1.29241437e-01 3.96929681e-01 1.05813518e-02
2.69323021e-01 1.75443888e-01 5.00839250e-03 -1.67720526e-01
6.93695471e-02 -6.68615401e-01 6.79156303e-01 1.10565889e+00
-4.70854849e-01 -9.64336991e-01 -5.76856673e-01 3.77679735e-01
-1.92614853e-01 -1.51059553e-01 -2.80940562e-01 7.49542594e-01
-5.52172720e-01 8.70246232e-01 2.07701400e-01 3.29401970e-01
3.77588384e-02 3.24261710e-02 1.11348115e-01 -4.80984509e-01
-1.41675964e-01 4.24116910e-01 5.12132198e-02 -1.79852486e-01
-7.26774275e-01 -1.25780535e+00 -9.28796887e-01 -2.08156288e-01
-3.65410060e-01 6.76993608e-01 7.66471744e-01 1.08171391e+00
2.28993267e-01 7.69682646e-01 4.96189326e-01 -1.06433976e+00
-4.24466848e-01 -1.14699316e+00 -1.23369145e+00 1.47448942e-01
6.57100737e-01 -6.32558942e-01 -8.58318567e-01 6.25067800e-02] | [6.651901721954346, 3.499920129776001] |
37ce6478-8e01-4ec4-ab68-7fa812e068e0 | i2c2w-image-to-character-to-word-transformers | 2105.08383 | null | https://arxiv.org/abs/2105.08383v3 | https://arxiv.org/pdf/2105.08383v3.pdf | I2C2W: Image-to-Character-to-Word Transformers for Accurate Scene Text Recognition | Leveraging the advances of natural language processing, most recent scene text recognizers adopt an encoder-decoder architecture where text images are first converted to representative features and then a sequence of characters via `sequential decoding'. However, scene text images suffer from rich noises of different sources such as complex background and geometric distortions which often confuse the decoder and lead to incorrect alignment of visual features at noisy decoding time steps. This paper presents I2C2W, a novel scene text recognition technique that is tolerant to geometric and photometric degradation by decomposing scene text recognition into two inter-connected tasks. The first task focuses on image-to-character (I2C) mapping which detects a set of character candidates from images based on different alignments of visual features in an non-sequential way. The second task tackles character-to-word (C2W) mapping which recognizes scene text by decoding words from the detected character candidates. The direct learning from character semantics (instead of noisy image features) corrects falsely detected character candidates effectively which improves the final text recognition accuracy greatly. Extensive experiments over nine public datasets show that the proposed I2C2W outperforms the state-of-the-art by large margins for challenging scene text datasets with various curvature and perspective distortions. It also achieves very competitive recognition performance over multiple normal scene text datasets. | ['Song Bai', 'Shijian Lu', 'Wenqing Zhang', 'Jiaxing Huang', 'Changhu Wang', 'Chuhui Xue'] | 2021-05-18 | null | null | null | null | ['scene-text-recognition'] | ['computer-vision'] | [ 1.00844765e+00 -7.85029769e-01 2.87764698e-01 -4.14887607e-01
-9.04018700e-01 -5.89784801e-01 9.83737886e-01 1.55214518e-01
-4.42655414e-01 1.01829678e-01 1.72953695e-01 -5.09835072e-02
4.51577961e-01 -4.46695507e-01 -7.86942065e-01 -8.17983091e-01
6.90403581e-01 4.28392172e-01 6.38648689e-01 -7.85268769e-02
9.14436042e-01 3.29293519e-01 -1.49462664e+00 8.19181681e-01
7.04106390e-01 9.24678206e-01 7.13873744e-01 1.11991596e+00
-6.81194484e-01 8.66505325e-01 -6.04249001e-01 -4.86186385e-01
1.74388260e-01 -3.37355733e-01 -2.34457165e-01 6.42458558e-01
7.04118192e-01 -3.00506145e-01 -3.60279202e-01 1.40144086e+00
4.47120339e-01 -3.12773101e-02 7.36765087e-01 -7.28843391e-01
-9.04450417e-01 2.40470439e-01 -7.72647202e-01 -1.08120209e-02
5.69532514e-01 1.38508216e-01 7.45852888e-01 -1.60719287e+00
4.53847021e-01 1.34367418e+00 5.81517696e-01 3.85963261e-01
-1.13027847e+00 -3.29635590e-01 1.90992072e-01 3.14228535e-01
-1.61348212e+00 -4.77166176e-01 4.72996056e-01 -4.92136568e-01
1.14480197e+00 5.20533562e-01 1.61759257e-01 1.25794685e+00
1.05241872e-01 1.14231372e+00 8.04098606e-01 -6.76485956e-01
4.94643487e-03 1.86455622e-01 2.02944919e-01 5.92212677e-01
1.77436262e-01 -3.71598691e-01 -7.46327281e-01 2.79706001e-01
3.70334119e-01 -1.63758807e-02 -4.16226476e-01 1.06804799e-02
-1.37092710e+00 4.45399582e-01 -1.38612315e-01 1.84761450e-01
-8.67385715e-02 6.83853701e-02 5.29059231e-01 2.54783183e-01
2.60480382e-02 -2.74531320e-02 -7.24119991e-02 -1.39216140e-01
-9.79262054e-01 -7.00008273e-02 5.40911555e-01 1.09709287e+00
4.09992695e-01 3.07377666e-01 -2.98154652e-01 1.09127557e+00
5.22546351e-01 1.19913363e+00 8.19155931e-01 1.40979737e-01
9.78651285e-01 6.59263790e-01 -1.47238016e-01 -1.36942947e+00
5.51252216e-02 -3.57130505e-02 -9.77113605e-01 -1.04447022e-01
2.01023221e-01 2.94605404e-01 -1.04302967e+00 8.92502725e-01
3.67367044e-02 -2.63844486e-02 1.82395786e-01 9.46744859e-01
8.57097030e-01 1.19662797e+00 -2.00823903e-01 1.73919678e-01
1.41286051e+00 -1.06711626e+00 -6.83889866e-01 -5.33393919e-01
4.58464444e-01 -1.52444267e+00 1.04516637e+00 6.29085541e-01
-7.84127414e-01 -6.81328893e-01 -1.16557622e+00 -1.44205272e-01
-4.86411721e-01 7.39012420e-01 -1.57888785e-01 7.23000288e-01
-8.56329322e-01 8.98325164e-03 -4.86134499e-01 -5.10416448e-01
1.90224513e-01 7.96590596e-02 -3.61283571e-01 -4.72456366e-01
-6.09818578e-01 6.72944129e-01 4.40086037e-01 3.00203651e-01
-7.61815190e-01 -5.34241237e-02 -8.03263962e-01 -3.08683459e-02
4.45781112e-01 2.32935157e-02 8.46634150e-01 -1.41419268e+00
-1.55966866e+00 8.84027123e-01 -3.54504913e-01 -1.33322448e-01
7.82252133e-01 -2.69040018e-01 -6.85448527e-01 3.17418039e-01
7.91462064e-02 4.40087229e-01 1.49615860e+00 -1.15193450e+00
-7.32012510e-01 -2.18296111e-01 -9.67060030e-01 6.35819018e-01
-2.25561798e-01 1.48091242e-01 -1.13427293e+00 -8.55155408e-01
3.42065156e-01 -6.24297082e-01 3.04071695e-01 1.04794562e-01
-7.41239786e-01 -1.49351567e-01 1.20846379e+00 -6.09512448e-01
7.94266522e-01 -2.19662166e+00 -7.75266141e-02 1.42100841e-01
-7.40665048e-02 5.39784849e-01 -4.87921208e-01 7.13397205e-01
2.97693033e-02 -2.07743853e-01 -8.48139897e-02 -5.81958413e-01
-1.22159328e-02 -1.59824178e-01 -7.20885098e-01 7.05236793e-01
2.62073785e-01 8.61985922e-01 -4.52677459e-01 -5.89715004e-01
6.37444913e-01 2.43027836e-01 -1.12917773e-01 2.27635115e-01
-1.91797152e-01 -7.98570141e-02 -4.88923818e-01 6.69753790e-01
9.21974599e-01 -1.25662968e-01 -5.81931062e-02 1.74620301e-02
-1.08020715e-01 -8.07950869e-02 -1.19458711e+00 1.35661805e+00
-1.22507410e-02 1.14858508e+00 -1.88361898e-01 -1.30039155e+00
1.26522279e+00 -9.79910046e-03 -4.16104645e-02 -8.16097438e-01
1.57220677e-01 3.17929804e-01 -4.00899351e-01 -7.65225351e-01
8.03494453e-01 3.43156576e-01 -6.82823360e-02 1.53091177e-01
-2.38601774e-01 -2.66061276e-01 -5.23424558e-02 2.41827622e-01
9.26236689e-01 -1.85406506e-02 -4.99974191e-03 -2.10613981e-01
1.03946853e+00 1.53990015e-01 1.57431766e-01 1.06527936e+00
-4.59595546e-02 1.01662982e+00 2.30112895e-01 -3.25481951e-01
-1.37260914e+00 -9.63496447e-01 7.03509385e-03 8.98144186e-01
4.19893444e-01 -2.54791081e-01 -5.64206481e-01 -4.34618354e-01
-1.32607743e-01 6.56526804e-01 -3.23071420e-01 5.38860187e-02
-5.03729165e-01 -6.17680848e-01 7.75324643e-01 3.58925253e-01
7.61246383e-01 -8.50257993e-01 -4.29092228e-01 -2.38969401e-02
-5.28302649e-03 -1.57905233e+00 -8.88218999e-01 9.07223001e-02
-3.72256368e-01 -9.55362499e-01 -8.85295451e-01 -1.26600957e+00
9.00458455e-01 7.46907651e-01 5.20823717e-01 8.89458060e-02
-6.37003720e-01 4.83690739e-01 -7.52124131e-01 -7.13489279e-02
-5.47316492e-01 -4.09524322e-01 -9.94434655e-02 6.44452512e-01
5.12806058e-01 2.52183527e-01 -4.05426353e-01 3.25084120e-01
-1.08939207e+00 5.30487120e-01 7.18660235e-01 9.00023937e-01
4.36616987e-01 5.28035350e-02 -7.61073083e-02 -7.01256514e-01
5.03666162e-01 -6.17052652e-02 -8.09310794e-01 6.20606184e-01
-4.33853775e-01 -1.15939602e-01 1.03230822e+00 -5.00750840e-01
-1.08812940e+00 2.79862523e-01 1.30692214e-01 -3.42033207e-01
-3.07895780e-01 1.98502541e-01 -8.72472525e-02 4.08520875e-03
5.03158033e-01 1.34446824e+00 -3.51165533e-01 -3.19006175e-01
1.51356205e-01 1.11513698e+00 8.24805140e-01 -3.76041710e-01
1.01962256e+00 3.56877685e-01 -2.67528176e-01 -1.51836228e+00
-3.27335030e-01 -7.55263329e-01 -5.78566790e-01 -2.44642615e-01
9.35727060e-01 -9.96475101e-01 -4.72082585e-01 1.12136412e+00
-1.27151799e+00 -1.87474459e-01 1.65447056e-01 3.64588350e-01
-2.22607315e-01 1.04847395e+00 -4.35819387e-01 -7.97365010e-01
-3.32516193e-01 -1.33555412e+00 1.52987790e+00 6.79588094e-02
3.12459528e-01 -6.50305808e-01 -2.67016381e-01 3.52644235e-01
1.82224125e-01 -3.39047521e-01 9.44169760e-01 -7.34951198e-01
-6.61427796e-01 -5.91451466e-01 -6.15280628e-01 3.71774077e-01
-6.48190379e-02 7.46737421e-02 -7.29367733e-01 -3.08609009e-01
-9.93761122e-02 -3.87399524e-01 9.06689703e-01 -2.35788319e-02
1.08779407e+00 -8.63372684e-02 -1.11996122e-01 6.19232535e-01
1.74447477e+00 1.86297297e-01 9.34778631e-01 3.05268794e-01
9.99399602e-01 3.52010578e-01 4.31523293e-01 4.43572760e-01
1.10276945e-01 6.82049870e-01 1.76107734e-01 4.50876988e-02
-2.03589395e-01 -3.34585845e-01 6.51221693e-01 9.65190291e-01
5.62425315e-01 -9.13501859e-01 -1.01593697e+00 3.46435875e-01
-1.91391468e+00 -7.53484726e-01 -6.68724120e-01 2.13184214e+00
5.63164949e-01 8.65901783e-02 -5.64570010e-01 1.20379187e-01
1.23670280e+00 2.41042137e-01 -5.38800716e-01 -4.79694396e-01
-7.59474695e-01 -4.99555394e-02 8.02245975e-01 2.12409407e-01
-9.42608774e-01 1.32107365e+00 5.37385654e+00 1.19923329e+00
-1.20950961e+00 -3.16147476e-01 6.15924537e-01 2.62452006e-01
-4.64744791e-02 -9.99848843e-02 -9.96241033e-01 5.11016846e-01
4.64648485e-01 1.91779479e-01 3.63287777e-01 7.30743110e-01
1.42770886e-01 -3.31235737e-01 -8.55117142e-01 1.57049513e+00
8.32194030e-01 -1.28234482e+00 5.40537894e-01 -3.97193909e-01
8.94806564e-01 9.42809731e-02 1.59257203e-01 -2.56110299e-02
7.29648694e-02 -9.50334132e-01 1.07227647e+00 5.34174621e-01
1.07621825e+00 -3.46873701e-01 4.43852693e-01 3.69277835e-01
-1.32437396e+00 1.22025467e-01 -6.26867533e-01 3.00780654e-01
-1.78851739e-01 4.09636319e-01 -8.63198042e-01 2.08819464e-01
5.47804117e-01 9.11891580e-01 -7.87224352e-01 8.49138796e-01
-8.71625096e-02 5.45580626e-01 -1.97494477e-01 -5.09928167e-01
5.07798374e-01 -1.88752159e-01 5.59920251e-01 1.64517868e+00
3.79480660e-01 4.61600348e-03 1.37658522e-01 6.96335077e-01
-1.29897334e-03 4.96518165e-01 -5.19658744e-01 -1.16159394e-01
3.27378571e-01 1.08533597e+00 -1.07399523e+00 -5.56787014e-01
-4.55050617e-01 1.73645401e+00 -6.85982555e-02 3.66997957e-01
-6.49571419e-01 -6.17339015e-01 7.80736879e-02 -2.55073994e-01
6.91213965e-01 -3.31931740e-01 -3.20182651e-01 -1.41515183e+00
2.46452674e-01 -1.05261469e+00 -1.55277215e-02 -9.63012338e-01
-1.14408624e+00 4.35216278e-01 -5.88454723e-01 -1.32359600e+00
2.85617620e-01 -1.02547741e+00 -6.44991040e-01 7.33011782e-01
-1.42589724e+00 -1.09894657e+00 -4.93534297e-01 8.79501641e-01
1.41502774e+00 -4.77049470e-01 5.13360441e-01 6.69616908e-02
-7.08035707e-01 7.68652141e-01 9.19052839e-01 3.34511310e-01
7.84624815e-01 -1.01683557e+00 8.00731838e-01 1.24003410e+00
1.57428429e-01 7.72292092e-02 4.98620987e-01 -8.74254763e-01
-2.02391911e+00 -1.15275967e+00 8.61815035e-01 -2.08819151e-01
5.32173157e-01 -7.46689737e-01 -8.41163278e-01 2.64827818e-01
1.52509302e-01 -1.79482788e-01 1.72473148e-01 -8.49069774e-01
-4.59874690e-01 1.41079903e-01 -7.07369149e-01 8.60787988e-01
5.64014435e-01 -6.01005197e-01 -4.44041222e-01 4.48214442e-01
2.45614737e-01 -3.18476021e-01 -1.93864536e-02 -6.56839609e-02
3.87116492e-01 -8.99761140e-01 7.98112154e-01 -1.66639715e-01
4.68656629e-01 -4.36683416e-01 -6.50620878e-01 -6.22404337e-01
3.58617082e-02 -5.30803084e-01 4.95464861e-01 1.21728802e+00
2.16066346e-01 -4.39662427e-01 5.89897931e-01 2.26492390e-01
-5.43303937e-02 -2.00145856e-01 -8.31326723e-01 -6.53855264e-01
-2.77923584e-01 -6.47566497e-01 1.18935041e-01 8.06380987e-01
-2.46396244e-01 3.08375895e-01 -6.45825744e-01 2.56823510e-01
5.93622446e-01 1.06013633e-01 7.90368557e-01 -7.25393891e-01
-7.61199370e-02 -5.13234675e-01 -7.54492462e-01 -1.71679163e+00
-7.96032697e-02 -8.34092200e-01 6.64805651e-01 -1.21973288e+00
3.55225474e-01 -8.28224793e-02 3.46797705e-01 5.94724156e-02
-2.15045720e-01 3.88866544e-01 2.95719653e-01 4.67539907e-01
-8.18917036e-01 5.86950183e-01 8.76087189e-01 -4.62106675e-01
1.60972252e-01 -3.02782744e-01 -1.53310612e-01 6.50166214e-01
5.48443019e-01 -2.48906732e-01 -1.09609433e-01 -8.56797278e-01
9.34297740e-02 -7.37585649e-02 2.73400426e-01 -1.02793062e+00
6.40922725e-01 -7.13473856e-02 6.92937434e-01 -1.08385885e+00
2.78647482e-01 -7.71391809e-01 -3.15583140e-01 4.15503889e-01
-4.27583545e-01 1.93793014e-01 1.33034766e-01 9.64774370e-01
-6.46162108e-02 -6.47191942e-01 7.48026013e-01 8.51389244e-02
-1.04077554e+00 -2.01961353e-01 -8.25342834e-01 -1.05157115e-01
9.15443540e-01 -6.86379910e-01 -5.17328322e-01 -3.25266510e-01
-1.19670359e-02 7.95515720e-03 3.98873687e-01 5.43932378e-01
1.14441729e+00 -1.00716603e+00 -9.65149403e-01 4.91655141e-01
3.55319709e-01 -1.40036687e-01 1.80793703e-01 7.25111127e-01
-9.63066399e-01 6.52374208e-01 2.74791479e-01 -1.09277797e+00
-1.59240866e+00 3.20914984e-01 1.47905648e-01 7.89229944e-02
-9.50301766e-01 9.89305496e-01 3.66807044e-01 -9.69349220e-02
4.93478745e-01 -1.50846362e-01 3.18043935e-03 -2.90212750e-01
6.19430780e-01 4.89052981e-02 4.41984497e-02 -1.11383891e+00
-3.14494818e-01 1.04170740e+00 -4.63237107e-01 -8.31374452e-02
9.11271274e-01 -3.72408032e-01 1.06112733e-01 3.26424479e-01
1.29906392e+00 -7.28061572e-02 -1.12270510e+00 -5.72840810e-01
3.84808034e-02 -8.87154341e-01 -2.68005133e-02 -7.11544037e-01
-8.00035238e-01 1.04824424e+00 7.56969750e-01 -4.79721501e-02
1.14140737e+00 -5.31776011e-01 7.35429168e-01 7.65646875e-01
-8.75383317e-02 -1.35078418e+00 5.63574433e-01 7.45192528e-01
7.34065533e-01 -1.41476262e+00 -8.21553823e-03 -2.89612293e-01
-1.01392412e+00 1.64411378e+00 4.30043548e-01 -1.25417978e-01
1.65439203e-01 2.39675671e-01 1.81215897e-01 -5.21462895e-02
-6.43208444e-01 -1.78771436e-01 3.73363525e-01 4.69258606e-01
7.25339651e-02 -1.73620060e-01 1.15555122e-01 6.98693842e-02
1.32445723e-01 -6.26545191e-01 6.31026566e-01 7.70260155e-01
-6.88882768e-01 -9.45346832e-01 -8.05988550e-01 5.05954444e-01
-2.54270703e-01 -4.98137414e-01 -5.87092996e-01 4.35246378e-01
-1.41816482e-01 1.00547409e+00 9.01429057e-02 -3.77117455e-01
1.77092656e-01 -3.28341424e-02 2.00542986e-01 -4.83602673e-01
-3.95534426e-01 4.54154402e-01 -2.99538106e-01 -2.43622050e-01
-3.72927785e-02 -7.93577552e-01 -1.16163623e+00 -2.35813841e-01
-6.18565857e-01 -2.34465703e-01 9.69792366e-01 8.81554186e-01
8.49318504e-02 2.66590387e-01 8.94706666e-01 -6.72467113e-01
-5.57649553e-01 -7.67634630e-01 -5.55930316e-01 6.65063143e-01
2.39432782e-01 -6.64802715e-02 -2.62697279e-01 5.10653436e-01] | [11.92921257019043, 2.2532551288604736] |
29ce0c66-2032-41f4-bd10-67f7cbc77b9b | geo-nav-a-geometric-dataset-of-voltage-gated | 2306.12348 | null | https://arxiv.org/abs/2306.12348v1 | https://arxiv.org/pdf/2306.12348v1.pdf | GEO-Nav: a geometric dataset of voltage-gated sodium channels | Voltage-gated sodium (Nav) channels constitute a prime target for drug design and discovery, given their implication in various diseases such as epilepsy, migraine and ataxia to name a few. In this regard, performing morphological analysis is a crucial step in comprehensively understanding their biological function and mechanism, as well as in uncovering subtle details of their mechanism that may be elusive to experimental observations. Despite their tremendous therapeutic potential, drug design resources are deficient, particularly in terms of accurate and comprehensive geometric information. This paper presents a geometric dataset of molecular surfaces that are representative of Nav channels in mammals. For each structure we provide three representations and a number of geometric measures, including length, volume and straightness of the recognized channels. To demonstrate the effective use of GEO-Nav, we have tested it on two methods belonging to two different categories of approaches: a sphere-based and a tessellation-based method. | ['Silvia Biasotti', 'Ulderico Fugacci', 'Andrea Raffo'] | 2023-06-21 | null | null | null | null | ['morphological-analysis'] | ['natural-language-processing'] | [ 3.25278848e-01 -4.24665719e-01 1.41145751e-01 -1.45116568e-01
-5.76171517e-01 -8.11878860e-01 3.74987930e-01 8.13442349e-01
-5.66689909e-01 1.05912173e+00 -2.10519597e-01 -7.19633937e-01
-1.60745814e-01 -5.51543593e-01 -3.95572037e-01 -9.46843684e-01
-4.75166082e-01 5.14576733e-01 1.82200298e-01 -2.40704045e-01
6.60863578e-01 9.84731674e-01 -1.42189682e+00 -3.49298447e-01
9.84444082e-01 8.87493908e-01 1.21064700e-01 4.37159777e-01
-3.08314502e-01 -2.27823004e-01 -4.21784133e-01 -2.84189910e-01
-9.43809748e-02 -4.71857607e-01 -4.53205943e-01 -3.26665759e-01
2.77518332e-01 1.77911773e-01 2.68728584e-01 8.74007821e-01
5.27310550e-01 -7.56679475e-02 7.94385970e-01 -4.88422275e-01
-1.36874259e-01 1.47804497e-02 -3.09193373e-01 2.81060219e-01
4.35444504e-01 5.60676754e-02 9.59101915e-01 -7.21586287e-01
8.17903280e-01 8.69669139e-01 5.39797366e-01 5.60223997e-01
-1.53238547e+00 -4.29438293e-01 -9.09428224e-02 2.09646687e-01
-1.35553300e+00 -4.05246139e-01 5.23505390e-01 -6.62113845e-01
6.85613155e-01 3.36979657e-01 8.93669069e-01 4.21101213e-01
3.21393788e-01 2.35637709e-01 1.23158598e+00 -2.11158946e-01
7.23184347e-01 -2.72806704e-01 -6.39984459e-02 2.44619474e-01
5.80663621e-01 -3.14615577e-01 -3.87934923e-01 -5.83538413e-01
7.76711285e-01 -5.69784045e-02 -3.65847021e-01 -7.42635548e-01
-9.38923657e-01 6.54467046e-01 1.81377411e-01 3.87443423e-01
-3.31490070e-01 -1.14179462e-01 3.31857622e-01 -1.45890117e-01
2.34409161e-02 5.18793464e-01 -5.06647825e-01 -4.29653406e-01
-6.12989962e-01 6.33481681e-01 6.17867768e-01 3.62773508e-01
5.93763053e-01 3.44726183e-02 5.19716442e-01 3.93271565e-01
2.85623223e-01 4.11726564e-01 1.72881350e-01 -3.21572721e-01
2.96321288e-02 8.85487318e-01 -4.89897504e-02 -8.11690032e-01
-7.51175582e-01 -3.61825973e-01 -8.50164473e-01 2.58082300e-01
7.68402576e-01 4.00272667e-01 -7.73262680e-01 1.44722033e+00
5.12806654e-01 -1.54992759e-01 -2.17028543e-01 7.46654630e-01
6.67456985e-01 1.79449901e-01 3.05071265e-01 -2.56860673e-01
1.37943411e+00 3.82882990e-02 -2.75701791e-01 3.18382710e-01
5.77860653e-01 -5.94794035e-01 6.13031447e-01 6.33986235e-01
-9.27586794e-01 1.17320158e-01 -9.29552913e-01 2.94497848e-01
-4.71388757e-01 -2.74925083e-02 7.10785091e-01 8.63006949e-01
-6.06757462e-01 9.48117375e-01 -1.13573515e+00 -4.54578906e-01
7.40057826e-01 4.78440791e-01 -4.75357562e-01 2.14281693e-01
-7.39442408e-01 8.51149917e-01 3.91679555e-02 6.41374737e-02
-5.43676198e-01 -6.15471125e-01 -6.25456214e-01 -1.82883963e-01
-7.19292462e-02 -4.53434169e-01 8.02429676e-01 3.58837135e-02
-1.21167696e+00 8.09429824e-01 -3.45768422e-01 -5.38160205e-01
3.24079275e-01 1.98748082e-01 -9.50496346e-02 2.29391664e-01
-1.41530678e-01 1.13301568e-01 3.91683459e-01 -1.07803023e+00
-2.74102062e-01 -1.12112081e+00 -3.13788980e-01 -9.13707614e-02
6.77958876e-02 1.47592081e-02 1.34140328e-01 -5.70041656e-01
4.94623214e-01 -7.22856462e-01 -6.06917560e-01 2.62361377e-01
-3.59653592e-01 -6.79747481e-03 5.39962351e-01 -6.39635444e-01
1.18607640e+00 -1.91067207e+00 4.03029501e-01 6.01933181e-01
4.65143263e-01 5.08406460e-01 3.10200334e-01 6.41958475e-01
1.34517968e-01 3.29351902e-01 -6.32918119e-01 1.07047446e-01
-3.91280681e-01 1.04788519e-01 -1.19457766e-01 9.63775277e-01
-9.09656100e-03 6.66248679e-01 -7.42228925e-01 -6.24848306e-02
2.60897458e-01 7.43968725e-01 -2.55943120e-01 -1.25225112e-01
-2.85857975e-01 9.43778753e-01 -8.08959961e-01 9.46936667e-01
6.29023135e-01 7.15923756e-02 3.38440150e-01 7.26589561e-02
-3.78314883e-01 4.08694685e-01 -1.06829858e+00 1.45565057e+00
-4.96337563e-02 2.42265433e-01 2.26843759e-01 -8.73350382e-01
8.81906390e-01 1.64770141e-01 4.64969605e-01 -8.00432742e-01
2.06480294e-01 5.70088387e-01 3.07831705e-01 -1.69310287e-01
1.55461296e-01 -2.66007364e-01 3.46234709e-01 2.68162072e-01
-2.96918333e-01 7.72457197e-02 2.26761997e-01 1.76955741e-02
9.79167819e-01 -3.44122276e-02 5.93981385e-01 -4.34679776e-01
5.84901929e-01 1.03074923e-01 4.54467058e-01 2.34060854e-01
-1.72897905e-01 5.98588526e-01 5.08477449e-01 -5.32597303e-01
-1.09465039e+00 -1.05158496e+00 -7.02117264e-01 3.79277229e-01
1.77287996e-01 -4.75650340e-01 -7.38725781e-01 9.45098419e-03
1.75542086e-01 3.23128760e-01 -5.61201394e-01 6.91728666e-02
-4.03544933e-01 -7.78828859e-01 5.17851174e-01 2.78586656e-01
-1.38743863e-01 -8.13316286e-01 -1.05621290e+00 2.62355059e-01
3.27086985e-01 -7.29313672e-01 1.98565453e-01 3.20157081e-01
-1.17381895e+00 -1.44242406e+00 -6.87715948e-01 -3.07050318e-01
6.59924626e-01 -4.79923710e-02 6.87046945e-01 1.95331022e-01
-8.31205070e-01 -6.81586713e-02 -3.84578198e-01 -6.25595331e-01
-2.21512526e-01 -1.06496392e-02 1.74559787e-01 -2.99529493e-01
2.78374612e-01 -8.99341047e-01 -8.45096231e-01 3.09305936e-01
-8.53597641e-01 -4.57130939e-01 5.16771913e-01 4.14871812e-01
8.92546952e-01 -4.06928778e-01 6.26173556e-01 -6.71128988e-01
5.81680477e-01 -3.14149320e-01 -7.33020484e-01 -5.88908307e-02
-4.01986837e-01 -2.23492011e-01 6.10504210e-01 8.70929472e-03
-4.13576573e-01 2.69821763e-01 -4.45472032e-01 1.57062009e-01
-4.52296913e-01 4.59737599e-01 -3.24822217e-01 -5.27399182e-01
5.04840553e-01 5.01548529e-01 2.19690025e-01 -8.21202517e-01
4.76538017e-02 4.16220039e-01 2.31845126e-01 -5.07355988e-01
5.45224726e-01 7.77063072e-01 4.14357543e-01 -1.35590458e+00
-1.61567777e-01 -7.74442077e-01 -6.54729843e-01 1.17884889e-01
7.23798573e-01 -3.24576586e-01 -9.08349276e-01 4.46973681e-01
-9.87819135e-01 8.16764385e-02 9.57511216e-02 4.06153888e-01
-4.07598138e-01 5.07194340e-01 -2.99843788e-01 -7.73263693e-01
-2.86478281e-01 -1.26147115e+00 9.90997314e-01 2.75587350e-01
-3.25441182e-01 -1.07576239e+00 3.28988850e-01 1.50951013e-01
6.65895879e-01 4.91776615e-01 1.36027753e+00 -8.20825458e-01
-5.32299221e-01 -3.30002457e-01 6.29848316e-02 1.80425625e-02
1.80998892e-01 6.77753314e-02 -7.61577249e-01 -2.59084553e-01
-2.56893039e-01 -4.91645932e-02 7.66495526e-01 5.03665209e-01
1.13540637e+00 1.45012259e-01 -5.09246767e-01 5.66089928e-01
1.43850303e+00 5.36095023e-01 7.23042428e-01 4.82630730e-01
4.92996126e-01 3.94417524e-01 2.64467001e-01 5.35228670e-01
2.09584236e-02 8.01824272e-01 6.56353772e-01 -4.29420248e-02
2.97203183e-01 2.92250663e-02 -2.64686882e-01 2.78072655e-01
-5.29188454e-01 2.82370299e-01 -8.89227927e-01 4.39153224e-01
-1.41618752e+00 -8.20267081e-01 -3.33795279e-01 2.69359446e+00
5.72130203e-01 5.47550842e-02 3.88959706e-01 2.75966555e-01
3.72038543e-01 -2.66088575e-01 -8.50248754e-01 -4.74713326e-01
-4.69773799e-01 4.75272924e-01 3.90927285e-01 1.57240182e-01
-8.44075143e-01 5.97315371e-01 6.74316549e+00 6.15341961e-01
-1.26500428e+00 -5.20325065e-01 3.84469420e-01 3.52685988e-01
-4.48170602e-01 2.29224950e-01 -8.39138329e-01 1.90497637e-01
5.41734695e-01 -1.80407301e-01 1.93328395e-01 5.75520098e-01
2.71959335e-01 -4.24671054e-01 -8.43484640e-01 8.64883423e-01
-1.55492067e-01 -1.67244494e+00 2.51327842e-01 3.98305029e-01
2.23408893e-01 -7.17345923e-02 -1.59411982e-01 -5.40441632e-01
-5.76861501e-01 -1.36898422e+00 4.68420088e-01 3.74092638e-01
8.77147555e-01 -7.23983109e-01 5.51220715e-01 4.88909744e-02
-1.26513970e+00 3.38690549e-01 -1.96288347e-01 -2.46286914e-01
3.59792173e-01 5.36214650e-01 -6.53339326e-01 4.04523998e-01
5.24332583e-01 3.63890916e-01 -4.22141492e-01 1.59152937e+00
-2.48715952e-02 5.60629904e-01 -3.20555836e-01 -1.48857579e-01
-1.25898831e-02 -6.17848158e-01 7.30684638e-01 7.81121314e-01
1.93251684e-01 1.36660174e-01 -8.34059715e-02 9.60348666e-01
2.31424496e-01 5.48445761e-01 -5.48128664e-01 -2.06956804e-01
4.66043830e-01 1.20986772e+00 -1.15248859e+00 1.70478210e-01
-2.16342628e-01 4.14262116e-01 -1.14236295e-03 2.66611129e-01
-3.32538724e-01 -4.62674737e-01 1.11193836e+00 5.52897453e-01
3.70295376e-01 -5.02459586e-01 -4.07497436e-01 -8.10408115e-01
1.74578950e-01 -7.29893446e-01 1.02349669e-01 -3.52868289e-02
-9.20765281e-01 4.72785503e-01 -6.83592930e-02 -9.80210245e-01
1.34156704e-01 -8.00756156e-01 -6.68975770e-01 9.80898023e-01
-1.36558890e+00 -9.33957636e-01 7.66931521e-03 2.60057926e-01
2.34004054e-02 9.85570624e-02 1.07040298e+00 3.33960176e-01
-3.45049024e-01 2.27459118e-01 4.48221385e-01 -2.78803408e-01
2.48301834e-01 -1.21842289e+00 4.84958112e-01 2.28798792e-01
1.62629392e-02 9.84359801e-01 1.06247103e+00 -5.67226171e-01
-1.51187205e+00 -7.22058535e-01 6.06214106e-01 -2.76308596e-01
6.57425582e-01 -3.89217168e-01 -1.16713166e+00 2.99389243e-01
-4.17136967e-01 -2.22385451e-01 9.23292458e-01 3.03795449e-02
-1.60288766e-01 2.28218704e-01 -1.16418052e+00 5.39742351e-01
8.32784772e-01 -2.25008011e-01 -3.13555658e-01 1.72221422e-01
-5.47429658e-02 -2.26742625e-01 -8.75819802e-01 4.06872213e-01
8.71044457e-01 -1.26934409e+00 8.60678196e-01 -5.72825611e-01
-5.11452742e-02 -5.90777695e-01 -5.36353253e-02 -1.25912035e+00
1.04683533e-01 -4.67596740e-01 2.37038955e-01 7.47026026e-01
5.19661784e-01 -7.71191955e-01 1.00746584e+00 3.18816602e-01
-1.09209917e-01 -1.28389931e+00 -1.28950894e+00 -6.55851603e-01
3.27937365e-01 -3.50122511e-01 4.18760628e-01 5.59912026e-01
1.38212979e-01 7.24856630e-02 1.93189472e-01 -2.37366766e-01
7.31208980e-01 2.56885231e-01 7.29841471e-01 -1.58227241e+00
1.52318686e-01 -6.53409362e-01 -1.05361879e+00 -6.92149222e-01
-1.87432125e-01 -6.85068250e-01 -3.91259998e-01 -1.50948513e+00
1.49841398e-01 -6.64352715e-01 1.96466018e-02 1.97052822e-01
2.05224797e-01 2.64510959e-01 -2.79055059e-01 7.45743886e-02
-2.84947604e-01 4.96569216e-01 9.85705435e-01 1.21003464e-01
-3.46133500e-01 9.48014855e-02 -6.12048209e-01 8.79188359e-01
9.65716243e-01 -3.10177386e-01 7.66697675e-02 2.34564498e-01
2.56864488e-01 -1.43530592e-01 2.55309850e-01 -7.46317089e-01
-2.43326519e-02 -1.09788530e-01 2.27859765e-01 -5.99344671e-01
3.37248534e-01 -5.60418606e-01 7.09661618e-02 7.34683037e-01
7.91756436e-02 5.98270819e-02 2.65471041e-01 6.48009241e-01
-1.95859641e-01 6.06283993e-02 7.98769534e-01 -1.08759075e-01
-6.32048905e-01 2.59781718e-01 -6.07744932e-01 -1.51810929e-01
1.11578012e+00 -6.83354139e-01 -6.47371709e-02 -1.62474260e-01
-8.41678977e-01 -1.19552143e-01 8.93097103e-01 -5.90895563e-02
7.83182323e-01 -7.42770672e-01 -4.65841442e-01 2.22329438e-01
4.43487972e-01 5.30211478e-02 3.85226488e-01 1.12317729e+00
-8.64261329e-01 5.35068154e-01 -4.41945016e-01 -7.19377637e-01
-1.52459705e+00 3.01491678e-01 3.04101974e-01 1.09561384e-01
-4.71014708e-01 6.65149331e-01 1.07449561e-01 -2.93007582e-01
1.88633427e-01 -3.27520549e-01 -4.16300774e-01 -1.53080285e-01
6.50247157e-01 5.79802811e-01 3.39644730e-01 -9.19436932e-01
-6.46218002e-01 9.22082901e-01 7.98159912e-02 4.58959311e-01
1.37116289e+00 -6.11570366e-02 -4.91386443e-01 3.59202415e-01
6.78083003e-01 1.22236326e-01 -6.47680819e-01 1.95174918e-01
1.63856134e-01 -4.28779721e-01 -3.64178509e-01 -5.42434096e-01
-6.80093288e-01 1.04940438e+00 5.24642050e-01 1.43392444e-01
8.57636690e-01 1.76748723e-01 6.92105532e-01 1.90883473e-01
5.41776359e-01 -5.97068548e-01 -5.34172118e-01 1.56802014e-01
6.19547427e-01 -8.47482562e-01 1.66541070e-01 -5.64137638e-01
-7.42905885e-02 1.04701293e+00 1.10495828e-01 6.34143725e-02
5.88091195e-01 2.69563440e-02 1.19266026e-01 -2.83211768e-01
-5.31554520e-01 -3.09394658e-01 1.64675623e-01 5.82341850e-01
7.40795612e-01 7.17587918e-02 -8.55579913e-01 3.54374439e-01
-6.65181577e-02 -3.25696617e-01 3.86395961e-01 1.09150004e+00
-6.60721719e-01 -1.41475236e+00 -3.70535642e-01 6.28299415e-01
-6.32072330e-01 -6.38296753e-02 -5.64958870e-01 5.79616129e-01
-1.33896813e-01 4.86715078e-01 -9.11326110e-02 -6.36625886e-02
3.63250315e-01 -1.98504180e-01 6.45251334e-01 -5.24673045e-01
-2.92340547e-01 9.06495526e-02 -1.39995432e-02 -3.11348766e-01
-2.77420729e-01 -7.70071268e-01 -1.65930748e+00 -1.18691824e-01
-3.03223073e-01 3.75447124e-01 1.21254492e+00 1.11088610e+00
3.17642838e-01 8.40077102e-02 1.68071717e-01 -9.06695783e-01
-5.25717497e-01 -5.17944396e-01 -1.06941497e+00 4.96341676e-01
1.68478832e-01 -7.10931480e-01 -1.54059142e-01 -1.00468792e-01] | [13.394696235656738, -3.0392699241638184] |
532a91e3-151e-4f5d-882d-036efd55e901 | a-unifying-view-on-task-oriented-dialogue | null | null | https://aclanthology.org/2022.lrec-1.137 | https://aclanthology.org/2022.lrec-1.137.pdf | A Unifying View On Task-oriented Dialogue Annotation | Every model is only as strong as the data that it is trained on. In this paper, we present a new dataset, obtained by merging four publicly available annotated corpora for task-oriented dialogues in several domains (MultiWOZ 2.2, CamRest676, DSTC2 and Schema-Guided Dialogue Dataset). This way, we assess the feasibility of providing a unified ontology and annotation schema covering several domains with a relatively limited effort. We analyze the characteristics of the resulting dataset along three main dimensions: language, information content and performance. We focus on aspects likely to be pertinent for improving dialogue success, e.g. dialogue consistency. Furthermore, to assess the usability of this new corpus, we thoroughly evaluate dialogue generation performance under various conditions with the help of two prominent recent end-to-end dialogue models: MarCo and GPT-2. These models were selected as popular open implementations representative of the two main dimensions of dialogue modelling. While we did not observe a significant gain for dialogue state tracking performance, we show that using more training data from different sources can improve language modelling capabilities and positively impact dialogue flow (consistency). In addition, we provide the community with one of the largest open dataset for machine learning experiments. | ['Ondřej Dušek', 'Patrick Paroubek', 'Daniel Stancl', 'Leon-paul Schaub', 'Vojtěch Hudeček'] | null | null | null | null | lrec-2022-6 | ['dialogue-state-tracking'] | ['natural-language-processing'] | [-1.51575133e-01 8.59079301e-01 -2.81256363e-02 -2.72434831e-01
-8.31493556e-01 -8.88298333e-01 1.21228480e+00 4.09006178e-01
-5.99854112e-01 1.17881584e+00 7.55889177e-01 -2.51942396e-01
-3.20287831e-02 -4.60368931e-01 2.91191973e-02 -2.77601238e-02
1.88325748e-01 1.05443001e+00 4.23313379e-01 -1.04768574e+00
4.35998321e-01 1.54071022e-02 -1.11805546e+00 5.37438989e-01
1.09541345e+00 4.10754710e-01 -1.24579072e-01 8.21181893e-01
-2.85022080e-01 7.54151464e-01 -9.35429692e-01 -7.89768517e-01
-1.27196550e-01 -5.59591651e-01 -1.64991951e+00 4.91308868e-02
1.12526938e-01 -2.81841695e-01 2.34669130e-02 4.66057003e-01
7.81252801e-01 1.71527460e-01 4.67482775e-01 -1.00952125e+00
-2.96198688e-02 7.74989665e-01 2.24780485e-01 -6.45028055e-02
9.17384982e-01 3.84820580e-01 1.04291987e+00 -3.15921992e-01
1.16691303e+00 1.28129637e+00 5.69651186e-01 8.49779785e-01
-1.25784504e+00 3.72514948e-02 -2.46816874e-01 -2.77759403e-01
-6.90851867e-01 -1.02319109e+00 3.51993948e-01 -3.52786809e-01
1.27386260e+00 5.33456743e-01 3.47732157e-01 1.32776427e+00
-1.82664037e-01 6.83717906e-01 1.25768161e+00 -7.47084856e-01
9.71699506e-03 7.06542253e-01 3.02165419e-01 5.25971651e-01
-2.23773777e-01 -3.33344877e-01 -6.21730328e-01 -5.06832719e-01
1.97036058e-01 -1.16028476e+00 -4.08443123e-01 -3.05637449e-01
-1.23437333e+00 1.00832534e+00 -2.56609321e-01 5.96475720e-01
7.09343851e-02 -5.17569721e-01 1.01968193e+00 5.20548940e-01
5.73846221e-01 8.83648396e-01 -5.88288426e-01 -9.00995851e-01
-3.89908940e-01 6.45689070e-01 1.78335059e+00 9.14990783e-01
3.51938158e-01 -3.84185463e-01 -4.08187747e-01 1.32235110e+00
2.83720762e-01 -1.40794948e-01 5.30843079e-01 -1.23637259e+00
8.06697965e-01 5.85872352e-01 3.25044721e-01 -5.67330182e-01
-6.99816287e-01 1.69468835e-01 -3.97296250e-01 -1.64014816e-01
1.01127064e+00 -5.97295642e-01 3.46638747e-02 1.67787647e+00
2.44216919e-01 -8.18459868e-01 7.17873156e-01 4.90128279e-01
1.18585908e+00 5.01933873e-01 7.98573270e-02 -2.53127813e-01
1.38597834e+00 -1.04299366e+00 -9.54960644e-01 -1.54161423e-01
1.30041373e+00 -9.34716582e-01 1.00718617e+00 2.64000356e-01
-1.31961620e+00 -3.49527091e-01 -7.75288820e-01 -2.76728839e-01
-4.19132948e-01 9.26524680e-03 5.60927272e-01 7.40586758e-01
-1.10664392e+00 5.12407959e-01 -3.88425350e-01 -9.02750015e-01
-4.35153246e-01 -7.42726997e-02 -4.16923702e-01 3.96753252e-01
-1.58923352e+00 1.53833687e+00 5.48852563e-01 -3.73794436e-01
-1.87764034e-01 -2.86720246e-01 -8.75891507e-01 -2.23321646e-01
3.64997357e-01 -4.18407202e-01 1.78180873e+00 -3.48620534e-01
-1.92424345e+00 1.06537354e+00 4.53214049e-02 -6.65755153e-01
8.63668799e-01 -6.26678467e-02 -1.50757849e-01 -6.25134334e-02
1.28192931e-01 7.64316738e-01 -2.01738745e-01 -1.05581892e+00
-7.11563468e-01 -1.84108883e-01 5.38462341e-01 5.39161503e-01
-2.24294454e-01 1.80076420e-01 -2.78726697e-01 -1.42262325e-01
-3.85473728e-01 -8.32318664e-01 -1.15155347e-01 -5.38892925e-01
-4.63597357e-01 -6.22231424e-01 3.68467659e-01 -9.77872670e-01
1.41534829e+00 -1.69809473e+00 2.44324967e-01 -3.06923211e-01
2.12152824e-02 4.31820095e-01 7.54238106e-04 1.15183568e+00
4.18016136e-01 2.61478186e-01 -2.63773620e-01 -4.07116711e-01
2.54920751e-01 1.59029424e-01 -1.74998641e-02 -7.00514298e-03
1.82584807e-01 7.50493050e-01 -9.21773732e-01 -4.72837269e-01
1.71212792e-01 2.76718177e-02 -5.70288777e-01 4.33522642e-01
-5.63558936e-01 6.15005672e-01 -1.46519259e-01 -1.79047778e-03
4.04328294e-02 7.35158697e-02 4.50699270e-01 8.11590403e-02
-2.80642062e-01 9.29208934e-01 -8.25600863e-01 1.99710953e+00
-7.40148008e-01 6.47749662e-01 6.72702789e-02 -4.71249282e-01
1.03872466e+00 8.14055443e-01 3.19713205e-01 -6.88157976e-01
3.98604423e-02 1.93366170e-01 2.69665480e-01 -7.01866090e-01
1.06473684e+00 1.41015202e-01 -3.81531537e-01 8.25572371e-01
3.74897033e-01 -3.11337709e-01 7.62909889e-01 4.18260068e-01
8.80796731e-01 1.44306784e-02 5.00928819e-01 -3.47836852e-01
6.01713955e-01 4.24875677e-01 7.20496103e-02 6.04164183e-01
-3.23093742e-01 2.13482574e-01 9.81814742e-01 -6.25766069e-02
-1.14095592e+00 -4.14973289e-01 -2.83421576e-01 1.20164478e+00
-2.93355644e-01 -7.93356061e-01 -1.00473416e+00 -7.40332067e-01
-3.81656885e-01 1.00471771e+00 -3.68543446e-01 1.58046275e-01
-4.60921854e-01 -6.94666743e-01 1.15089428e+00 -9.21583399e-02
6.85262084e-01 -1.04251254e+00 -6.00086153e-01 3.63034785e-01
-7.53649056e-01 -1.20731080e+00 -1.18464477e-01 3.16673294e-02
-5.91993392e-01 -1.10781503e+00 -5.47372520e-01 -4.64546204e-01
-7.86198601e-02 -2.09398210e-01 1.43477046e+00 7.77975991e-02
1.14648774e-01 5.51783442e-01 -6.93288445e-01 -2.24280611e-01
-1.25442708e+00 6.27647340e-01 -3.26145887e-01 -5.26435137e-01
2.63244450e-01 5.54198697e-02 -5.14791273e-02 3.18710446e-01
-5.14300287e-01 2.96544194e-01 6.73661055e-03 9.79556739e-01
-5.56558847e-01 -4.33978945e-01 7.95538425e-01 -1.30259252e+00
1.54119551e+00 -4.12211031e-01 -1.20657615e-01 2.65284389e-01
-4.79536265e-01 8.81173611e-02 2.46455342e-01 1.48842067e-01
-1.53443730e+00 -3.64942998e-01 -5.93361318e-01 7.28495240e-01
-4.08924699e-01 6.04921639e-01 -1.02904532e-02 2.26322889e-01
9.41658080e-01 1.48153687e-02 3.66587937e-01 -5.92882395e-01
5.47959745e-01 1.00388610e+00 3.04002672e-01 -7.29421616e-01
1.71567127e-01 -1.23666562e-01 -6.34974241e-01 -9.85340476e-01
-4.41235781e-01 -5.88722527e-01 -7.76301980e-01 -9.99361202e-02
7.65903652e-01 -6.62463427e-01 -6.59101784e-01 3.80299002e-01
-1.42052364e+00 -7.59617388e-01 -1.59366608e-01 5.60062006e-02
-7.07328141e-01 5.73734760e-01 -8.82177234e-01 -9.40932631e-01
-3.99339199e-01 -1.00554240e+00 7.46854365e-01 7.46313855e-02
-9.70975637e-01 -1.51190674e+00 4.49978709e-01 7.67127454e-01
3.90632391e-01 1.35114521e-01 8.36657882e-01 -1.31957066e+00
2.54354089e-01 2.53311917e-03 8.96711275e-03 1.49529994e-01
3.33771296e-02 6.37567639e-02 -9.86387074e-01 -2.21335083e-01
-1.34347528e-01 -9.93037105e-01 3.89262617e-01 -3.75270724e-01
1.85801938e-01 -3.40753496e-01 4.12583686e-02 -3.07032108e-01
8.08535635e-01 2.04471216e-01 5.30112922e-01 6.62344873e-01
2.35480994e-01 1.35536206e+00 8.10167193e-01 5.47375679e-01
8.20134878e-01 1.06181228e+00 -3.56218629e-02 6.52080551e-02
-5.60245812e-02 -1.17819369e-01 3.41779292e-01 9.13221240e-01
6.98626861e-02 -4.07015830e-01 -1.06724203e+00 5.77987671e-01
-1.93989134e+00 -7.70164669e-01 -2.00079709e-01 1.97941983e+00
1.39525139e+00 1.54863194e-01 5.91278553e-01 -1.73853303e-03
4.31909055e-01 2.48927772e-01 3.49579230e-02 -9.38683033e-01
-1.21419758e-01 2.19202810e-03 1.67257655e-02 8.91588509e-01
-7.24299312e-01 9.68423069e-01 6.40911627e+00 4.84635204e-01
-5.98314226e-01 1.44533217e-01 5.33853054e-01 1.97758973e-01
-1.95789859e-01 1.63097084e-01 -8.90523612e-01 4.35897797e-01
1.27951384e+00 -3.72662157e-01 1.97092116e-01 5.09457648e-01
2.38200590e-01 -3.99732560e-01 -1.12828541e+00 2.89065152e-01
-1.08994827e-01 -1.26241362e+00 -1.99089527e-01 3.73741835e-02
2.62150764e-01 -7.64456764e-02 -5.03941655e-01 6.87150002e-01
6.03054821e-01 -7.99746513e-01 4.93041813e-01 1.95525497e-01
4.85827506e-01 -3.60392749e-01 9.13437724e-01 5.59229970e-01
-4.45671499e-01 3.29382211e-01 -7.59972557e-02 -2.05419898e-01
3.54228199e-01 -8.38916227e-02 -1.29362440e+00 6.68026984e-01
2.77308762e-01 4.02170420e-01 -7.33842969e-01 7.34891236e-01
-6.04126751e-02 4.08997387e-01 -1.73849791e-01 -2.40987822e-01
3.33137214e-01 -1.54158205e-01 5.82902610e-01 1.46880031e+00
-2.38440186e-01 -7.22624511e-02 1.80552036e-01 5.45667589e-01
3.59426253e-02 3.49475920e-01 -7.19207406e-01 -5.18218167e-02
6.12211943e-01 1.27033377e+00 -2.32238874e-01 -3.41302216e-01
-4.16024506e-01 7.61398554e-01 3.46952170e-01 -6.66417405e-02
-3.79115641e-01 -4.44108069e-01 3.70679468e-01 -1.66164458e-01
-4.67640102e-01 -9.62297171e-02 -2.03971073e-01 -1.12434769e+00
-1.66987792e-01 -1.33078527e+00 5.04953623e-01 -4.88700002e-01
-1.06396949e+00 6.70523226e-01 1.72709078e-01 -7.91508198e-01
-9.42685783e-01 -3.33931506e-01 -4.86837149e-01 1.07280517e+00
-1.24821949e+00 -8.54704916e-01 -8.90694335e-02 3.23940784e-01
8.34108531e-01 -2.37793908e-01 1.28009689e+00 9.52454060e-02
-5.99509656e-01 4.48807299e-01 4.71044406e-02 2.50345886e-01
1.16831839e+00 -1.46047330e+00 6.23055398e-01 2.19016179e-01
-1.66579053e-01 6.12231553e-01 9.07021999e-01 -5.01755297e-01
-9.18778539e-01 -5.11658728e-01 1.18504691e+00 -7.47743249e-01
8.12998295e-01 -3.36769342e-01 -9.74494636e-01 5.58068871e-01
9.14407015e-01 -9.06670928e-01 7.54272223e-01 4.88435507e-01
8.71043950e-02 6.01552606e-01 -1.22067130e+00 5.78338742e-01
8.04523408e-01 -4.15889412e-01 -9.14106250e-01 5.05969346e-01
6.25034392e-01 -8.01805258e-01 -1.33557034e+00 2.02241302e-01
3.52783203e-01 -1.30181348e+00 4.94217485e-01 -7.32358038e-01
5.19441068e-01 3.86593997e-01 6.44093305e-02 -1.62443924e+00
1.86037049e-01 -9.41861987e-01 1.00472912e-01 1.68977070e+00
7.47925639e-01 -7.45758176e-01 5.79499900e-01 1.02237582e+00
-1.94137678e-01 -5.60801029e-01 -8.28801274e-01 -4.91858304e-01
4.78938758e-01 -5.20028137e-02 4.22287256e-01 1.01146460e+00
9.16785538e-01 1.00212419e+00 -3.53564560e-01 -5.88473082e-01
7.16892183e-02 -2.44297847e-01 1.28709137e+00 -1.16407394e+00
-1.93976983e-01 -5.10519385e-01 1.84879050e-01 -9.99517739e-01
5.74854985e-02 -7.20830381e-01 -4.52631824e-02 -1.66408837e+00
-9.66141671e-02 -5.93250096e-01 5.64139187e-01 3.51393402e-01
-1.43358320e-01 -1.79153070e-01 2.26593375e-01 2.35398859e-01
-6.78927243e-01 5.15453041e-01 1.18878520e+00 1.80977345e-01
-4.09745485e-01 -1.03084594e-01 -7.45819569e-01 5.28902590e-01
1.02912498e+00 4.02422361e-02 -3.88825417e-01 -3.58119309e-01
-1.23416051e-01 4.63391453e-01 -8.35910514e-02 -6.87110245e-01
1.60204262e-01 9.61690247e-02 -2.96685785e-01 -1.79317266e-01
4.42100704e-01 -2.19215140e-01 -2.74103731e-01 3.96190137e-01
-8.55814576e-01 -1.11219898e-01 4.22610253e-01 1.77096408e-02
-2.01239958e-01 -7.04459250e-01 6.30304039e-01 -3.67330462e-01
-6.69915259e-01 -4.40600306e-01 -6.88183308e-01 5.36806524e-01
8.20808947e-01 -5.57393767e-02 -7.71759748e-01 -6.97101951e-01
-5.93211055e-01 5.08408904e-01 4.06870931e-01 6.16483271e-01
-1.06768109e-01 -8.66778910e-01 -8.60449731e-01 -2.78317422e-01
4.09125865e-01 -1.84938177e-01 -4.87003336e-03 6.66877031e-01
-5.21043718e-01 1.05983067e+00 -4.13358629e-01 -4.46587056e-01
-1.31803727e+00 -9.22427624e-02 3.34185272e-01 -8.11086893e-01
-3.53377134e-01 4.20901656e-01 -4.46055710e-01 -1.15147412e+00
2.95990527e-01 1.16629146e-01 -6.55741751e-01 3.62516314e-01
3.13091725e-01 4.67997521e-01 2.24066004e-01 -6.60648942e-01
1.31937265e-01 -1.83464602e-01 -2.18432590e-01 -6.88753426e-01
1.05841017e+00 -4.20227200e-01 -1.33934066e-01 6.45193577e-01
7.98551083e-01 6.86236247e-02 -9.08358037e-01 -3.11862350e-01
3.89128953e-01 -2.41495326e-01 -3.88044298e-01 -1.18686461e+00
-2.15216011e-01 7.05990493e-01 2.22351193e-01 9.11014199e-01
4.43468392e-01 -1.56311467e-01 6.64334118e-01 5.46600163e-01
4.62799877e-01 -1.26997483e+00 3.67398486e-02 1.03869188e+00
9.90311861e-01 -1.38250053e+00 -2.10384220e-01 -3.55148911e-01
-1.06677437e+00 1.11555004e+00 8.49070847e-01 5.85968971e-01
-1.79029647e-02 -7.19517889e-03 3.48018110e-01 -2.64471740e-01
-1.15321088e+00 -2.29921997e-01 -9.92020443e-02 6.37666702e-01
1.01486921e+00 -2.58691877e-01 -8.19605708e-01 3.04646313e-01
-5.09980261e-01 -2.31590360e-01 8.33157957e-01 9.53645647e-01
-3.67572218e-01 -1.50052726e+00 -1.43300816e-01 3.07529658e-01
-4.00502115e-01 6.54292479e-03 -1.08278823e+00 1.03391743e+00
-5.67189395e-01 1.41344702e+00 -1.83711365e-01 -1.70609564e-01
6.05665386e-01 4.64232177e-01 3.23564976e-01 -9.44866598e-01
-1.06430244e+00 -2.43812561e-01 1.35787821e+00 -1.72984406e-01
-4.08729553e-01 -6.16010010e-01 -9.08275366e-01 -5.42706668e-01
-4.01701957e-01 6.29339159e-01 5.43396771e-01 9.11317825e-01
2.44758263e-01 3.65076035e-01 2.68481135e-01 -5.21406174e-01
-6.84874833e-01 -1.49420059e+00 -1.57746151e-01 4.54488158e-01
-1.63505703e-01 -5.04847288e-01 -8.46425295e-02 -7.12071806e-02] | [12.77100944519043, 7.983251571655273] |
1975bf9a-f168-420a-b77e-e9fbe005bdf0 | generalised-image-outpainting-with-u | 2201.11403 | null | https://arxiv.org/abs/2201.11403v5 | https://arxiv.org/pdf/2201.11403v5.pdf | Generalised Image Outpainting with U-Transformer | In this paper, we develop a novel transformer-based generative adversarial neural network called U-Transformer for generalised image outpainting problem. Different from most present image outpainting methods conducting horizontal extrapolation, our generalised image outpainting could extrapolate visual context all-side around a given image with plausible structure and details even for complicated scenery, building, and art images. Specifically, we design a generator as an encoder-to-decoder structure embedded with the popular Swin Transformer blocks. As such, our novel neural network can better cope with image long-range dependencies which are crucially important for generalised image outpainting. We propose additionally a U-shaped structure and multi-view Temporal Spatial Predictor (TSP) module to reinforce image self-reconstruction as well as unknown-part prediction smoothly and realistically. By adjusting the predicting step in the TSP module in the testing stage, we can generate arbitrary outpainting size given the input sub-image. We experimentally demonstrate that our proposed method could produce visually appealing results for generalized image outpainting against the state-of-the-art image outpainting approaches. | ['Yujie Geng', 'John Y. Goulermas', 'Yuyao Yan', 'Kaizhu Huang', 'Rui Zhang', 'Xi Yang', 'Penglei Gao'] | 2022-01-27 | null | null | null | null | ['image-outpainting'] | ['computer-vision'] | [ 5.53382099e-01 5.08782923e-01 1.14479966e-01 -2.72204638e-01
-7.86459267e-01 -4.09897119e-01 4.85332757e-01 -5.37696362e-01
8.80732685e-02 9.16114807e-01 1.18345812e-01 -2.08094314e-01
3.67280364e-01 -9.15013671e-01 -1.46398664e+00 -6.25104427e-01
3.51578087e-01 2.41328657e-01 1.62059948e-01 -3.23306561e-01
4.21190746e-02 4.81114298e-01 -1.20124924e+00 5.90641975e-01
9.81031060e-01 8.58796358e-01 4.52280253e-01 8.28324854e-01
-1.09810662e-02 1.14748240e+00 -7.41586924e-01 -7.07034886e-01
3.20723176e-01 -9.24062610e-01 -4.78498191e-01 4.86579955e-01
5.29416084e-01 -6.09163702e-01 -5.32387972e-01 7.65323102e-01
3.93408179e-01 -1.16676398e-01 5.58910966e-01 -1.12356341e+00
-1.06801736e+00 4.97845173e-01 -8.53449225e-01 -2.34258592e-01
4.93297338e-01 4.74993199e-01 4.98117149e-01 -8.04715097e-01
8.56402636e-01 1.25035727e+00 7.32402027e-01 6.90489650e-01
-1.41726041e+00 -5.25367677e-01 3.20790820e-02 1.97526291e-01
-1.15862501e+00 -1.84020415e-01 1.34587407e+00 -3.05974483e-02
5.41763425e-01 5.62834084e-01 7.22189367e-01 1.36150968e+00
5.22805214e-01 7.74869621e-01 1.26796567e+00 -4.43345368e-01
7.29897246e-02 1.34250268e-01 -8.81300688e-01 5.31281650e-01
-2.60444164e-01 3.01640570e-01 -2.13454247e-01 1.07834451e-01
1.31798518e+00 1.73001513e-01 -3.26753289e-01 -2.28169501e-01
-1.11137128e+00 6.34156227e-01 7.10201859e-01 -9.49285179e-02
-3.70128751e-01 2.87269086e-01 3.11138809e-01 4.19162571e-01
5.76860785e-01 3.66401613e-01 -1.43306494e-01 3.40589017e-01
-1.07218432e+00 3.25673044e-01 3.95584702e-01 1.11271381e+00
7.97890246e-01 3.93011987e-01 -5.07308602e-01 8.98299694e-01
-1.58903286e-01 2.66450465e-01 3.52848887e-01 -9.98658180e-01
4.96057183e-01 3.19115758e-01 1.16791673e-01 -7.31235683e-01
1.92214474e-01 -5.14231384e-01 -1.18782604e+00 4.86821026e-01
1.65283024e-01 -1.00640528e-01 -1.02013254e+00 1.64485383e+00
3.66086572e-01 3.24034095e-01 -4.13956866e-02 8.56287062e-01
6.78959250e-01 8.80589962e-01 -2.74471063e-02 -2.25604638e-01
1.27516365e+00 -1.24501491e+00 -7.30678201e-01 -3.76132846e-01
1.25301540e-01 -8.43262374e-01 1.18236172e+00 2.55478144e-01
-1.51477373e+00 -8.81213903e-01 -9.53580737e-01 -4.87963110e-01
3.33287939e-02 -1.00213803e-01 3.36015373e-01 2.29611591e-01
-1.06932330e+00 5.76637864e-01 -3.97132277e-01 1.11418039e-01
3.80677134e-01 -5.87463118e-02 -5.96110642e-01 -2.40744814e-01
-1.14624882e+00 8.13932180e-01 3.66642594e-01 2.04438150e-01
-1.10545492e+00 -7.64766097e-01 -1.02318776e+00 1.17337389e-03
1.64050058e-01 -1.06566846e+00 9.84514654e-01 -1.41758084e+00
-1.49835253e+00 7.89643288e-01 8.84304103e-03 -5.32159746e-01
8.56928766e-01 -6.40278682e-02 -1.63620546e-01 1.71964452e-01
3.80525887e-02 1.06406522e+00 1.49941385e+00 -1.75415254e+00
-1.25856578e-01 -1.15756638e-01 -4.26134951e-02 2.80793101e-01
1.48514524e-01 -2.45222792e-01 -3.47079366e-01 -1.35696983e+00
-2.04300195e-01 -5.51277518e-01 -3.73532116e-01 2.98486054e-01
-5.61358750e-01 4.39810514e-01 1.35658395e+00 -1.15159857e+00
9.47171569e-01 -2.01403141e+00 2.33687922e-01 -1.63141072e-01
5.84361218e-02 1.54282689e-01 -3.70858341e-01 6.21494651e-01
-3.53532702e-01 -7.79870972e-02 -5.36152065e-01 -6.33273482e-01
-3.33825886e-01 4.75471705e-01 -6.57420576e-01 1.68763787e-01
4.90160435e-01 1.33878684e+00 -7.56527603e-01 -4.08574224e-01
4.12551761e-01 7.62031376e-01 -7.72931039e-01 5.97692013e-01
-4.99132484e-01 8.91191602e-01 -1.54822901e-01 6.26310945e-01
9.26700234e-01 3.63600999e-02 -1.65122896e-01 -1.89782441e-01
1.18192390e-01 -2.01131418e-01 -6.32650971e-01 1.94026315e+00
-8.02207947e-01 6.44094825e-01 1.24623716e-01 -9.67324495e-01
1.02124131e+00 1.69515699e-01 4.12919279e-03 -7.73611665e-01
-8.07620585e-02 3.46328914e-02 -3.09411913e-01 -5.25565565e-01
5.22006035e-01 -4.22112614e-01 1.40092447e-01 1.50356710e-01
-5.93368225e-02 -5.47771633e-01 -2.57043749e-01 1.77682847e-01
8.69682908e-01 4.95205790e-01 1.07282430e-01 1.33358285e-01
5.65888643e-01 -3.75161052e-01 2.07719475e-01 4.59752798e-01
2.92986482e-01 1.34922314e+00 5.71455300e-01 -5.19108772e-01
-1.66216075e+00 -1.00905919e+00 1.53314054e-01 5.66076458e-01
1.21580750e-01 -7.23935813e-02 -1.02143812e+00 -5.79976380e-01
-3.46618265e-01 8.53905916e-01 -7.87310243e-01 -2.65838116e-01
-6.84299827e-01 -1.55244723e-01 4.46819097e-01 5.56575596e-01
7.91317165e-01 -1.48495686e+00 -3.00932378e-01 3.93200755e-01
-3.46234083e-01 -1.10429800e+00 -9.15668130e-01 6.28960654e-02
-8.70627284e-01 -7.95119703e-01 -1.27027094e+00 -1.06846571e+00
8.69554877e-01 5.91535009e-02 1.23090982e+00 -1.25640363e-01
-2.62482345e-01 -1.08003564e-01 -3.41603845e-01 -2.14815661e-02
-9.58150744e-01 -2.50254869e-01 -5.88522136e-01 1.83704153e-01
-4.85155731e-01 -1.10871851e+00 -1.07937551e+00 2.93618023e-01
-1.40993106e+00 6.77840769e-01 7.64824927e-01 1.24395132e+00
6.96550250e-01 -4.45507355e-02 4.97362435e-01 -9.75651860e-01
5.07438421e-01 -3.28574240e-01 -2.51323670e-01 1.83491692e-01
-2.72511482e-01 1.35899991e-01 1.18076229e+00 -7.02931464e-01
-1.14548743e+00 -5.47554120e-02 -5.22866428e-01 -9.77263629e-01
-2.81355139e-02 -8.20773616e-02 -3.55184555e-01 -1.88967809e-01
5.74852586e-01 8.34471226e-01 -2.23667491e-02 -3.51776063e-01
5.91190577e-01 3.89893740e-01 8.06938589e-01 -6.47871614e-01
9.27918494e-01 5.31229436e-01 2.56404597e-02 -5.21549821e-01
-4.79859382e-01 2.34947070e-01 -3.29868227e-01 -1.44148290e-01
8.25415611e-01 -1.03081918e+00 -3.94770384e-01 5.25520623e-01
-1.47210908e+00 -4.62771356e-01 -6.52930498e-01 -2.70017177e-01
-8.25705171e-01 5.32231808e-01 -8.36759984e-01 -4.57007468e-01
-4.87338543e-01 -1.30400062e+00 1.47296703e+00 7.10453168e-02
7.19684288e-02 -7.91994929e-01 -1.22136101e-01 3.87128592e-01
4.16660249e-01 7.31220603e-01 8.42605174e-01 1.57103851e-01
-7.76464283e-01 5.69559969e-02 -2.92387515e-01 6.58212602e-01
1.81174595e-02 -2.98784554e-01 -8.25387895e-01 -2.86595404e-01
2.80666053e-01 -3.28635514e-01 8.10969174e-01 2.97503442e-01
1.42662847e+00 -8.05106163e-01 -7.58709460e-02 9.24357653e-01
1.47161055e+00 2.01938033e-01 1.31287098e+00 2.41392702e-01
9.02562618e-01 4.77478147e-01 3.91300827e-01 3.51185054e-01
6.55720904e-02 7.28573203e-01 6.74038172e-01 -4.29638743e-01
-4.20580357e-01 -9.02564883e-01 4.20143127e-01 3.85517597e-01
2.24851128e-02 -3.96108240e-01 -2.53515512e-01 4.60138232e-01
-1.47187591e+00 -1.05669904e+00 6.71430156e-02 1.94599056e+00
1.08669388e+00 -8.08752328e-02 -1.60200283e-01 1.10161878e-01
7.65441418e-01 3.46039355e-01 -6.05314076e-01 -5.81240058e-01
-1.87034979e-01 4.20214534e-01 4.87587541e-01 5.85410535e-01
-5.99778831e-01 8.56778979e-01 5.71930981e+00 1.27467585e+00
-1.10169160e+00 1.97729811e-01 1.20065784e+00 1.64893776e-01
-7.53129900e-01 7.96223357e-02 -2.05665082e-01 6.52020514e-01
4.13958907e-01 2.36786589e-01 4.82473254e-01 7.16246009e-01
2.28697121e-01 6.64399713e-02 -9.19072092e-01 9.69328880e-01
1.44787326e-01 -1.49504745e+00 5.76555312e-01 -1.48740336e-01
9.67252672e-01 -5.32108247e-01 2.31913418e-01 6.59982860e-02
-1.06810555e-01 -1.07590246e+00 1.02740335e+00 4.56694961e-01
1.36699581e+00 -9.38274443e-01 3.01522434e-01 4.56858069e-01
-9.21430707e-01 4.79938872e-02 -3.00403267e-01 4.05384824e-02
4.30591911e-01 5.22915244e-01 -4.87357855e-01 5.84456444e-01
4.06635016e-01 5.17710745e-01 -4.98316556e-01 6.43905222e-01
-4.60879207e-01 2.20646530e-01 3.24260108e-02 5.71191370e-01
3.08629602e-01 -3.13103467e-01 6.42576396e-01 9.95191097e-01
4.42309290e-01 9.95161831e-02 -2.40829781e-01 1.08470869e+00
-2.28072926e-01 -2.01784372e-01 -9.41270888e-01 2.70983338e-01
9.17977765e-02 9.83512282e-01 -5.14572144e-01 -3.48677993e-01
-1.33997127e-01 1.74274027e+00 2.70584643e-01 4.67406631e-01
-1.14391291e+00 -2.72056401e-01 3.35239977e-01 5.07422388e-01
4.64841604e-01 3.78992595e-02 -2.28520706e-01 -1.15805459e+00
2.70514041e-01 -9.76592004e-01 -1.90161347e-01 -1.35135520e+00
-1.13644207e+00 9.56610739e-01 -1.99052542e-01 -1.47359991e+00
-3.50821286e-01 -2.53689975e-01 -8.38771939e-01 8.10044110e-01
-1.40503812e+00 -1.69228971e+00 -2.76828796e-01 7.05765903e-01
9.77559924e-01 1.26844682e-02 5.88322043e-01 1.06920972e-01
-1.22985318e-01 7.19298840e-01 1.37340456e-01 -4.41197753e-02
7.46450067e-01 -9.49385703e-01 5.94009042e-01 8.52420986e-01
1.75241381e-02 2.87880182e-01 9.92961526e-01 -6.68351412e-01
-1.31536841e+00 -1.39759517e+00 3.16661894e-01 -1.47836700e-01
2.39283994e-01 -5.27417898e-01 -8.34061027e-01 7.28545249e-01
7.06589282e-01 1.93657324e-01 4.90313349e-03 -7.38299072e-01
-3.89588535e-01 -1.10032402e-01 -1.44351661e+00 7.86032438e-01
1.03501272e+00 -4.59602356e-01 -3.18629950e-01 2.55625874e-01
9.70551312e-01 -7.40442097e-01 -7.21252084e-01 3.36065620e-01
5.14503360e-01 -1.30466688e+00 1.22974813e+00 -1.67437240e-01
1.20454407e+00 -3.57057273e-01 1.20689124e-01 -1.31307960e+00
-2.35462606e-01 -1.06645727e+00 -1.62241399e-01 1.09674990e+00
7.05167698e-03 -3.81256044e-01 9.05888379e-01 1.38335064e-01
-2.86030799e-01 -9.14962888e-01 -7.84349680e-01 -5.82333267e-01
-1.81338971e-03 -2.61228085e-01 5.30455112e-01 6.94978952e-01
-5.48441887e-01 1.95232779e-01 -9.94685769e-01 1.54446363e-02
7.42789567e-01 1.36837780e-01 8.44375134e-01 -2.92094141e-01
-8.07155013e-01 -9.68913510e-02 -2.90177256e-01 -1.25450993e+00
5.77267446e-03 -5.28033555e-01 -1.17678642e-01 -1.28516400e+00
1.63668856e-01 -2.02338636e-01 1.71280175e-01 3.28154176e-01
-2.61951983e-01 7.58336842e-01 2.76147604e-01 1.41373679e-01
1.70599669e-02 8.63677323e-01 2.15286946e+00 -3.57465297e-01
6.21061660e-02 -1.45419985e-01 -6.81918502e-01 3.56490701e-01
5.37979364e-01 -3.30288470e-01 -7.19332516e-01 -4.00382936e-01
1.32450536e-02 6.97209656e-01 8.30057621e-01 -8.94686759e-01
-6.52850717e-02 -8.29465762e-02 7.20641971e-01 -5.45399666e-01
4.89738792e-01 -8.87865245e-01 6.91983283e-01 3.58156323e-01
-2.46977374e-01 1.64310053e-01 3.39278013e-01 6.18338645e-01
-4.93620217e-01 4.69324104e-02 9.59079266e-01 -3.38899106e-01
-4.97688949e-01 3.39622110e-01 1.63398758e-01 -5.64581417e-02
1.13865077e+00 -4.34800208e-01 -9.36511606e-02 -7.03231573e-01
-6.76866353e-01 -1.20095916e-01 7.73369968e-01 4.19421405e-01
9.92155373e-01 -1.40756655e+00 -8.01338136e-01 5.61700404e-01
-5.06957956e-02 3.15112710e-01 7.11808860e-01 4.37509060e-01
-9.06838179e-01 -5.20390794e-02 -5.98413587e-01 -4.74143475e-01
-8.66871059e-01 8.89221966e-01 1.84177026e-01 -4.42427605e-01
-8.56102049e-01 7.77544439e-01 8.26620162e-01 -1.16911605e-01
-1.06660411e-01 -2.42187187e-01 3.29202682e-01 -5.00892878e-01
4.49866325e-01 -1.65510044e-01 -2.69261390e-01 -4.34604913e-01
2.08375379e-01 4.21986520e-01 -2.22292751e-01 -1.45298332e-01
1.23538280e+00 -2.11905167e-01 -1.80031970e-01 8.24292600e-02
1.18969464e+00 -2.37616841e-02 -1.83674514e+00 1.63033798e-01
-8.34400892e-01 -5.37293732e-01 -2.38207921e-01 -7.71871209e-01
-1.23062718e+00 7.40266323e-01 3.98549020e-01 1.89707857e-02
1.52898610e+00 -1.29196182e-01 1.17018807e+00 -3.02069545e-01
3.35421890e-01 -5.49680769e-01 2.29523793e-01 -2.93762572e-02
1.48150373e+00 -1.05201077e+00 -1.69388056e-01 -3.80428523e-01
-7.86907256e-01 9.22147930e-01 5.88984966e-01 -5.81939101e-01
1.56139925e-01 4.32353050e-01 -1.05176978e-01 1.53025210e-01
-6.98990881e-01 3.28785866e-01 1.67154714e-01 8.10935795e-01
1.92039624e-01 -1.47052646e-01 -2.10373640e-01 1.71858415e-01
-1.20734647e-01 -5.96158323e-04 5.76519191e-01 5.68197906e-01
7.53158331e-03 -1.23221123e+00 -5.76249301e-01 3.16641442e-02
-3.29769611e-01 -3.32461119e-01 -9.47069004e-02 8.76128078e-01
3.03372562e-01 3.96114081e-01 -2.72016171e-02 -3.60053927e-01
1.33501545e-01 -2.45520413e-01 7.28224099e-01 -2.52992660e-01
-6.39097333e-01 8.86187404e-02 -2.63832092e-01 -6.28114283e-01
-1.90087944e-01 -1.49080530e-01 -7.05919743e-01 -2.14531332e-01
8.80011171e-03 -8.17422718e-02 3.69374156e-01 6.89366639e-01
4.65441197e-01 7.73104072e-01 6.98538542e-01 -1.25517118e+00
-3.20950955e-01 -1.06075156e+00 -5.57749629e-01 5.41212976e-01
5.20392120e-01 -2.92562395e-01 -1.12313621e-01 4.11698729e-01] | [11.497297286987305, -0.9492271542549133] |
089f0c9d-9ecd-4043-b824-e6bfda0cf1a7 | multitask-learning-for-low-resource-spoken | 2211.13703 | null | https://arxiv.org/abs/2211.13703v1 | https://arxiv.org/pdf/2211.13703v1.pdf | Multitask Learning for Low Resource Spoken Language Understanding | We explore the benefits that multitask learning offer to speech processing as we train models on dual objectives with automatic speech recognition and intent classification or sentiment classification. Our models, although being of modest size, show improvements over models trained end-to-end on intent classification. We compare different settings to find the optimal disposition of each task module compared to one another. Finally, we study the performance of the models in low-resource scenario by training the models with as few as one example per class. We show that multitask learning in these scenarios compete with a baseline model trained on text features and performs considerably better than a pipeline model. On sentiment classification, we match the performance of an end-to-end model with ten times as many parameters. We consider 4 tasks and 4 datasets in Dutch and English. | ['Hugo Van hamme', 'Marie-Francine Moens', 'Quentin Meeus'] | 2022-11-24 | null | null | null | null | ['spoken-language-understanding', 'intent-classification', 'spoken-language-understanding'] | ['natural-language-processing', 'natural-language-processing', 'speech'] | [ 1.69168770e-01 3.17561775e-01 -1.87364951e-01 -9.52607334e-01
-1.58452010e+00 -5.50203323e-01 7.41541803e-01 -7.02561662e-02
-1.00215173e+00 3.47392201e-01 6.86992407e-01 -4.68450665e-01
2.96002746e-01 -2.32256763e-02 -4.15205866e-01 -4.66666996e-01
9.30654705e-02 6.26832545e-01 -8.11145604e-02 -8.69382545e-02
4.66705002e-02 -3.61302495e-01 -1.24414325e+00 1.08441842e+00
5.01813710e-01 8.87493253e-01 3.29463243e-01 1.05749083e+00
-1.15149677e-01 9.13579702e-01 -8.25957477e-01 -5.70808768e-01
1.87908188e-02 2.57391334e-01 -1.09160066e+00 1.39243066e-01
5.74368119e-01 -1.51986316e-01 -5.14519289e-02 3.98833245e-01
7.72793055e-01 2.12256402e-01 5.54883897e-01 -1.05063760e+00
-4.60233092e-01 7.69522846e-01 -1.04014963e-01 3.55849147e-01
4.03546751e-01 1.07396938e-01 1.22305131e+00 -1.00280070e+00
7.06484616e-02 1.30033076e+00 7.57934213e-01 5.54712296e-01
-1.06589663e+00 -4.14828181e-01 4.99130726e-01 -8.05172771e-02
-7.55453587e-01 -1.44645202e+00 3.15180004e-01 -1.62697792e-01
1.80245447e+00 2.88945943e-01 -1.44760698e-01 1.56108880e+00
1.85088798e-01 1.22104716e+00 9.66531098e-01 -4.58951890e-01
6.53508976e-02 4.20959651e-01 4.20081377e-01 6.63021207e-01
-4.16454822e-01 -2.18698025e-01 -1.00150418e+00 -2.50586390e-01
-3.20109606e-01 -3.73080373e-01 -1.04966834e-01 4.12919283e-01
-1.24891996e+00 9.48363721e-01 -2.23700672e-01 2.32808009e-01
-2.27110922e-01 2.14328632e-01 8.35428715e-01 5.81051886e-01
1.10894775e+00 1.43217281e-01 -1.17623508e+00 -4.90229338e-01
-1.01877522e+00 -2.56272912e-01 1.25868738e+00 9.39675629e-01
4.38373148e-01 7.84620121e-02 -3.75916988e-01 1.32114530e+00
4.26869512e-01 2.92593569e-01 9.44850802e-01 -5.42421341e-01
8.81973267e-01 -2.63412595e-02 -2.63760567e-01 -6.18739910e-02
-7.25824356e-01 -4.58604217e-01 -4.05926973e-01 -1.80809230e-01
3.44221562e-01 -7.81491876e-01 -9.72836137e-01 1.69020820e+00
-1.17993228e-01 -1.61789320e-02 5.71604311e-01 2.70868927e-01
8.82182002e-01 8.47651064e-01 4.91192847e-01 -1.13897063e-01
1.69030011e+00 -1.52381968e+00 -8.38849843e-01 -9.46672916e-01
1.23233354e+00 -1.10015380e+00 1.20415175e+00 4.46411222e-01
-1.07531917e+00 -4.28256154e-01 -6.42957151e-01 -4.99020144e-02
-4.14690316e-01 5.23702264e-01 7.05183864e-01 1.20507336e+00
-1.22560310e+00 2.38218635e-01 -8.16636086e-01 -3.97612780e-01
1.68593183e-01 3.16591799e-01 -1.65236831e-01 3.92233670e-01
-9.96092737e-01 1.10412633e+00 1.96285442e-01 -2.72771657e-01
-9.81342435e-01 -6.13288045e-01 -1.03670537e+00 9.98399183e-02
2.79927045e-01 -6.77926123e-01 1.96232283e+00 -1.19005072e+00
-1.83581388e+00 9.86556828e-01 -5.79974949e-01 -5.55928946e-01
1.44529596e-01 -2.80157119e-01 -3.19035470e-01 -2.31567591e-01
1.88961625e-01 4.96391475e-01 8.27630103e-01 -7.82347500e-01
-9.67049479e-01 -2.78569967e-01 1.47908270e-01 5.06369293e-01
-6.67714894e-01 5.65679610e-01 -1.43934563e-01 -5.31299233e-01
-3.47366065e-01 -7.38188088e-01 -1.11986138e-01 -6.69627249e-01
-4.18916523e-01 -5.36708713e-01 7.21846640e-01 -6.51896000e-01
9.28189397e-01 -2.27008915e+00 -5.45254089e-02 -5.17002463e-01
-1.54768080e-01 1.44531772e-01 -3.38035226e-01 5.03413737e-01
-1.52990013e-01 4.14524913e-01 -6.27361089e-02 -1.44654012e+00
1.12224273e-01 2.08233133e-01 -2.12157503e-01 3.42418015e-01
3.45330834e-01 7.94622719e-01 -4.52398509e-01 -2.25126296e-01
-7.43941590e-02 2.48897314e-01 -4.61843759e-01 2.54044622e-01
-8.67707059e-02 5.31359315e-02 -3.44622940e-01 5.34161866e-01
3.15750510e-01 -2.06338018e-01 1.43823354e-03 1.64640453e-02
-1.12648807e-01 1.07763588e+00 -6.85014367e-01 1.84663677e+00
-1.21681511e+00 7.91163683e-01 3.76505584e-01 -1.13111722e+00
4.65695053e-01 8.77934933e-01 6.61581382e-02 -5.89292049e-01
-1.12982318e-01 1.86685428e-01 1.28875047e-01 -5.66652536e-01
2.38023758e-01 -4.90433872e-01 -3.13976228e-01 6.93651795e-01
6.75633132e-01 -9.82846841e-02 -9.91451442e-02 5.99087179e-02
1.15360880e+00 -2.06796005e-01 1.08578764e-01 -3.33809614e-01
4.47479337e-01 -9.76728871e-02 2.02211738e-01 8.97493541e-01
-2.22223818e-01 3.76116902e-01 4.78454858e-01 -3.55765730e-01
-7.77747750e-01 -4.11460698e-01 -1.71222314e-01 2.04856849e+00
-6.65388107e-01 -5.68310142e-01 -6.54285669e-01 -1.06818259e+00
-2.46929407e-01 9.63758945e-01 -4.54514295e-01 5.12990542e-02
-3.87481511e-01 -1.16419530e+00 9.29398537e-01 3.82225037e-01
3.24612767e-01 -1.05243528e+00 -3.51745754e-01 1.45594344e-01
-3.99220824e-01 -1.51162422e+00 -5.63206732e-01 9.49453950e-01
-6.30125165e-01 -3.41721386e-01 -5.44133246e-01 -9.21382368e-01
3.51218833e-03 1.42767116e-01 1.39472985e+00 -1.22382596e-01
2.69477665e-01 5.73981762e-01 -5.04501998e-01 -8.74093711e-01
-6.34025335e-01 7.38980234e-01 1.16189674e-01 1.35393992e-01
3.57701391e-01 -2.63501853e-01 -1.32836148e-01 1.00453660e-01
-6.42882645e-01 3.77666652e-02 6.78203464e-01 7.95126617e-01
-1.64345294e-01 -2.89257377e-01 8.83869588e-01 -1.01835370e+00
7.72264123e-01 -5.72252929e-01 2.46877111e-02 2.34283969e-01
-3.66682112e-01 1.48629725e-01 5.94333589e-01 -2.62531787e-01
-1.18785596e+00 2.32471436e-01 -6.23988390e-01 3.00018042e-02
-5.34678340e-01 5.19943237e-01 8.43961835e-02 1.35178030e-01
3.79244983e-01 2.03326762e-01 -3.96503359e-02 -5.17227888e-01
3.62266481e-01 1.17704117e+00 -1.08175121e-01 -4.19568986e-01
3.25807333e-01 2.76402682e-01 -8.28501284e-01 -1.22697091e+00
-1.32268918e+00 -6.58670008e-01 -5.70260346e-01 1.64303377e-01
9.25050914e-01 -1.37360179e+00 -6.35333657e-01 4.49687690e-01
-1.46784544e+00 -6.81025863e-01 -5.94785484e-03 4.95818198e-01
-6.61573887e-01 1.64817050e-01 -7.87288785e-01 -1.10054111e+00
-5.08987606e-01 -1.19080567e+00 1.54616797e+00 -2.76938081e-01
-2.24269450e-01 -1.46317208e+00 2.18391861e-03 6.08648539e-01
4.41218704e-01 -7.31368840e-01 7.75175154e-01 -1.44368231e+00
9.15982574e-02 -3.69727835e-02 7.75720850e-02 4.16264296e-01
1.56427026e-02 -3.61097515e-01 -1.70770562e+00 -2.88047940e-01
5.18883407e-01 -7.96321094e-01 1.18334413e+00 4.01201785e-01
8.88274014e-01 -4.50276405e-01 -2.32654646e-01 4.36414540e-01
1.02147102e+00 1.20733909e-01 1.56285197e-01 3.62222403e-01
4.45941687e-01 9.55847919e-01 3.76403898e-01 1.15452543e-01
6.81120217e-01 6.76116109e-01 8.45541432e-02 -1.76506132e-01
-1.39010295e-01 1.29368663e-01 9.13191140e-01 1.03975105e+00
5.30924678e-01 -5.25602162e-01 -1.05841267e+00 6.04140699e-01
-1.75143957e+00 -7.87965298e-01 1.85019290e-03 1.82220793e+00
7.87409425e-01 3.53025079e-01 3.78778547e-01 -1.80531591e-01
3.06690931e-01 4.80085552e-01 -1.04171574e-01 -7.78426111e-01
-1.35967508e-01 2.05674678e-01 5.19501746e-01 9.15470541e-01
-1.44890440e+00 1.20720196e+00 7.52660084e+00 8.81234646e-01
-1.09452903e+00 7.03044176e-01 9.16429102e-01 -4.16388184e-01
-1.17533460e-01 -1.57930896e-01 -1.11040640e+00 3.83053094e-01
1.68851101e+00 1.18438169e-01 1.28282011e-01 8.31453860e-01
2.26238504e-01 -1.12776920e-01 -1.29608178e+00 7.65672386e-01
3.68200243e-01 -9.42583978e-01 -1.23100281e-01 -1.70735158e-02
3.52688104e-01 5.14105618e-01 1.26404062e-01 7.86601007e-01
3.79727483e-01 -9.97945428e-01 8.30655873e-01 -2.95070726e-02
4.35097128e-01 -4.59666789e-01 6.78168714e-01 6.89067185e-01
-9.46927309e-01 -3.10245603e-01 -1.64274260e-01 -1.49374738e-01
3.80130142e-01 3.66495848e-01 -9.47000384e-01 3.69272053e-01
5.58614612e-01 3.85340095e-01 -5.47488451e-01 6.24490857e-01
-1.42535204e-02 9.92360055e-01 -3.89530808e-01 -2.77417749e-01
6.27473533e-01 2.07846135e-01 4.59765762e-01 1.88036847e+00
2.18362398e-02 -2.34184459e-01 4.26805615e-01 2.99564619e-02
-7.66657759e-03 3.35646182e-01 -7.44482934e-01 1.08213507e-01
1.01975091e-01 1.40514338e+00 -4.65680450e-01 -4.84431952e-01
-9.45536673e-01 1.02315378e+00 5.92897832e-01 3.04046482e-01
-5.56797922e-01 -1.48745701e-01 7.50172853e-01 -4.14307475e-01
1.53713942e-01 -2.94378519e-01 -5.08602679e-01 -1.20190787e+00
6.60272166e-02 -1.06736910e+00 4.44148600e-01 -5.39234102e-01
-1.21137571e+00 7.71601081e-01 -3.23908448e-01 -5.82547069e-01
-4.68949556e-01 -8.68460119e-01 -8.71853650e-01 8.94102752e-01
-1.75291419e+00 -1.29043579e+00 2.56897479e-01 4.06151950e-01
1.51736939e+00 -4.00458634e-01 1.11689270e+00 3.85429144e-01
-7.64442980e-01 7.04081059e-01 -4.30956818e-02 1.80898607e-01
9.33588266e-01 -1.38258612e+00 8.27434182e-01 6.05813384e-01
3.78369242e-01 3.28586906e-01 4.39093143e-01 -3.44969451e-01
-1.23757780e+00 -8.52160394e-01 1.29965687e+00 -8.96686018e-01
7.01209605e-01 -8.23695719e-01 -6.07672453e-01 1.14145815e+00
7.42163777e-01 -4.85701799e-01 1.10943317e+00 9.19613600e-01
-1.19246401e-01 1.56588182e-01 -6.74887180e-01 3.59739214e-01
8.35344851e-01 -8.05446982e-01 -5.88335693e-01 7.55921185e-01
1.02930450e+00 -2.42016360e-01 -6.11210823e-01 2.65048355e-01
5.34962952e-01 -7.54490435e-01 6.31026685e-01 -9.79072213e-01
2.37811849e-01 4.80126232e-01 -2.64300793e-01 -1.68759942e+00
7.09148347e-02 -7.90410936e-01 2.47639522e-01 1.18473566e+00
1.20072484e+00 -7.64420629e-01 6.23174429e-01 4.89976346e-01
-6.19105220e-01 -7.11234272e-01 -1.13212514e+00 -7.44152486e-01
2.83874184e-01 -8.59454513e-01 4.67509665e-02 7.57336915e-01
1.14679359e-01 1.00173879e+00 -3.24433208e-01 1.33729830e-01
2.68632650e-01 -3.39544684e-01 5.20266771e-01 -9.05501187e-01
-4.19839174e-01 -4.94668931e-01 1.36105850e-01 -1.17530131e+00
7.03931093e-01 -9.05981719e-01 1.53463304e-01 -1.33159983e+00
4.72122878e-02 -3.57295036e-01 -4.39589918e-02 8.13384891e-01
-2.98392296e-01 -5.30232303e-03 1.84236273e-01 1.08171545e-01
-7.40562260e-01 4.52377588e-01 4.62364286e-01 -1.06791981e-01
-1.21526413e-01 3.37040901e-01 -9.19812024e-01 8.23809981e-01
8.58411252e-01 -6.21919811e-01 -2.64782995e-01 -8.85957897e-01
1.41805992e-01 5.63588664e-02 -2.12179497e-01 -7.75033772e-01
4.43805963e-01 3.36694211e-01 6.50362018e-03 -2.79419363e-01
6.91368282e-01 -5.06143928e-01 -6.42348289e-01 1.49089724e-01
-7.44604886e-01 2.71849111e-02 3.93577069e-01 2.89704204e-01
-2.27690756e-01 -6.34561419e-01 5.47410488e-01 -1.98051214e-01
-4.48290586e-01 5.82167879e-02 -9.62236583e-01 1.43758237e-01
6.60020590e-01 1.81099340e-01 -3.94884884e-01 -6.00699067e-01
-9.52920020e-01 3.66943300e-01 -3.43553811e-01 6.71519041e-01
2.18891591e-01 -8.72298837e-01 -9.46878493e-01 1.87284544e-01
2.29923368e-01 -3.25960606e-01 -3.68649722e-03 9.86100852e-01
2.07289234e-01 8.12972844e-01 5.15566349e-01 -5.62627137e-01
-1.39162064e+00 2.81317979e-01 4.63869154e-01 -4.90683287e-01
-5.50488830e-02 9.73406017e-01 3.37852210e-01 -9.00099874e-01
5.56352258e-01 -2.53896385e-01 -2.36530542e-01 4.07055289e-01
4.55628097e-01 1.51473001e-01 5.59938967e-01 -3.89169663e-01
-4.84939426e-01 3.27657074e-01 -4.42589134e-01 -6.61692202e-01
1.29491806e+00 -2.27038920e-01 3.12599450e-01 7.41735935e-01
1.30484390e+00 4.74293493e-02 -9.59084392e-01 -2.48331800e-01
2.72843540e-01 -2.42343061e-02 3.16791266e-01 -9.78645980e-01
-6.81983054e-01 8.98897052e-01 3.47351283e-01 4.88246083e-01
8.83227289e-01 1.60588801e-01 5.46435237e-01 7.69644678e-01
7.39693921e-03 -1.19811225e+00 2.91876405e-01 9.48933184e-01
7.00274885e-01 -1.61205173e+00 -4.17037874e-01 -3.03682715e-01
-1.08336437e+00 9.38130975e-01 3.81769270e-01 1.45183116e-01
7.73582339e-01 5.95831156e-01 3.93567234e-01 -1.48937926e-01
-1.37850654e+00 -4.83529359e-01 1.62640437e-01 5.14712036e-01
9.97078061e-01 -1.00165248e-01 1.71941686e-02 4.61167097e-01
-3.23982328e-01 -4.46181625e-01 2.70500720e-01 8.26616943e-01
-4.43607062e-01 -1.09583950e+00 -9.96555835e-02 5.15189946e-01
-1.15309215e+00 -6.30973220e-01 -2.91064292e-01 4.69285905e-01
-4.19699073e-01 1.60341573e+00 1.27167627e-01 -2.85454929e-01
4.74302262e-01 7.23310709e-01 -6.70269504e-02 -1.15447736e+00
-1.02409482e+00 1.97963312e-01 9.05298293e-01 -4.71286833e-01
-2.98851430e-01 -7.04917550e-01 -7.26035953e-01 1.40828371e-01
-4.12686467e-01 1.45947263e-01 1.03754961e+00 1.30753446e+00
3.19930911e-01 6.28172398e-01 8.50248635e-01 -7.79949069e-01
-8.63619030e-01 -1.42550635e+00 -2.53698617e-01 -1.00829110e-01
6.08626723e-01 -1.44822136e-01 -6.01610899e-01 3.04859690e-02] | [14.096991539001465, 6.956182479858398] |
35c83885-532a-4097-9a6a-47f58b8bf8db | multi-person-implicit-reconstruction-from-a | 2104.09283 | null | https://arxiv.org/abs/2104.09283v1 | https://arxiv.org/pdf/2104.09283v1.pdf | Multi-person Implicit Reconstruction from a Single Image | We present a new end-to-end learning framework to obtain detailed and spatially coherent reconstructions of multiple people from a single image. Existing multi-person methods suffer from two main drawbacks: they are often model-based and therefore cannot capture accurate 3D models of people with loose clothing and hair; or they require manual intervention to resolve occlusions or interactions. Our method addresses both limitations by introducing the first end-to-end learning approach to perform model-free implicit reconstruction for realistic 3D capture of multiple clothed people in arbitrary poses (with occlusions) from a single image. Our network simultaneously estimates the 3D geometry of each person and their 6DOF spatial locations, to obtain a coherent multi-human reconstruction. In addition, we introduce a new synthetic dataset that depicts images with a varying number of inter-occluded humans and a variety of clothing and hair styles. We demonstrate robust, high-resolution reconstructions on images of multiple humans with complex occlusions, loose clothing and a large variety of poses and scenes. Our quantitative evaluation on both synthetic and real-world datasets demonstrates state-of-the-art performance with significant improvements in the accuracy and completeness of the reconstructions over competing approaches. | ['Adrian Hilton', 'Lourdes Agapito', 'Akin Caliskan', 'Armin Mustafa'] | 2021-04-19 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Mustafa_Multi-Person_Implicit_Reconstruction_From_a_Single_Image_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Mustafa_Multi-Person_Implicit_Reconstruction_From_a_Single_Image_CVPR_2021_paper.pdf | cvpr-2021-1 | ['3d-human-reconstruction'] | ['computer-vision'] | [-1.11868002e-01 -3.50628793e-01 4.99092937e-01 -3.51956367e-01
-8.21551800e-01 -4.02236462e-01 2.78968960e-01 -4.34391886e-01
-1.60907537e-01 6.54518604e-01 3.78858328e-01 5.88035762e-01
1.73363104e-01 -3.74110699e-01 -7.16036737e-01 -3.90502214e-01
9.52054933e-02 1.12679362e+00 1.65962443e-01 -1.29781708e-01
-5.31523824e-01 5.07336020e-01 -1.41518152e+00 1.41063109e-01
3.78382504e-01 6.09116018e-01 -7.17036277e-02 8.02119792e-01
5.39459467e-01 2.74508506e-01 -4.77707058e-01 -6.96868122e-01
5.47052205e-01 -2.42072389e-01 -2.74974525e-01 9.05188084e-01
1.18045926e+00 -6.85594976e-01 -5.46551466e-01 5.77133596e-01
8.49339664e-01 3.12691554e-02 4.68414426e-01 -7.30197012e-01
-2.68677086e-01 -1.52050942e-01 -9.98268843e-01 -4.27667439e-01
1.10784793e+00 2.42216557e-01 5.07732034e-01 -1.08350360e+00
7.64552593e-01 1.68461680e+00 1.24477017e+00 5.31824648e-01
-1.73352754e+00 -3.16553622e-01 1.68953866e-01 -2.88844675e-01
-1.61127853e+00 -6.43292069e-01 7.51661181e-01 -4.16133195e-01
4.93478537e-01 3.26124012e-01 1.23208225e+00 1.49199164e+00
-6.28420040e-02 6.39183462e-01 1.25848162e+00 -4.01411772e-01
-8.07807595e-02 -3.91199328e-02 -3.02956760e-01 9.27751243e-01
3.51239711e-01 2.04106316e-01 -5.81993163e-01 -3.55703324e-01
1.39477718e+00 2.31471911e-01 -7.19475374e-02 -9.04170275e-01
-1.46812093e+00 4.39574450e-01 3.19448888e-01 -3.32849413e-01
-5.46651065e-01 3.34421247e-01 6.33213446e-02 -1.91441804e-01
4.55050647e-01 -1.01130284e-01 -2.76095450e-01 2.54113555e-01
-8.95690143e-01 9.32932138e-01 8.64571095e-01 1.13479853e+00
4.35932845e-01 1.39285818e-01 8.67031589e-02 8.78462434e-01
5.02152324e-01 9.36345756e-01 -3.17618400e-01 -1.37395108e+00
2.93036610e-01 2.86828637e-01 6.60058856e-01 -1.19020557e+00
-4.39832538e-01 -5.99909246e-01 -1.00970602e+00 2.44587630e-01
6.33107901e-01 -1.02660671e-01 -9.09852505e-01 1.62106204e+00
7.97663927e-01 1.10396057e-01 -4.13350135e-01 1.40737998e+00
9.48329031e-01 2.85275161e-01 -7.02467710e-02 -7.61287883e-02
1.44134045e+00 -1.07611334e+00 -6.78719878e-01 -6.12179101e-01
-3.17094356e-01 -7.83396721e-01 9.21308100e-01 4.10817564e-01
-1.61340261e+00 -7.00375915e-01 -6.73857868e-01 -1.45161808e-01
1.48234159e-01 -2.60478724e-02 4.26093578e-01 3.26906770e-01
-8.45649123e-01 3.91672373e-01 -7.49655068e-01 -5.62910914e-01
2.76915878e-01 1.74681172e-01 -5.33774078e-01 -4.54236180e-01
-6.68644488e-01 8.67818475e-01 -3.67839575e-01 2.69911647e-01
-1.06642938e+00 -5.63248932e-01 -9.83318806e-01 -4.48921621e-01
3.88173133e-01 -1.47149670e+00 9.50422585e-01 -8.86778057e-01
-1.05404377e+00 1.19839191e+00 -1.01280309e-01 -3.51162776e-02
1.23084307e+00 -5.61951637e-01 -2.07501262e-01 1.13939993e-01
1.54976696e-01 5.26354432e-01 8.13419342e-01 -1.93718648e+00
-5.65697032e-04 -6.65918171e-01 -1.65348724e-01 5.16974211e-01
3.06221604e-01 5.04106805e-02 -8.00641596e-01 -8.63875031e-01
3.73849392e-01 -9.81461465e-01 -3.85350198e-01 7.06389010e-01
-5.86846292e-01 4.96528864e-01 5.62447906e-01 -9.73520935e-01
5.75962245e-01 -1.69971478e+00 7.21660316e-01 -5.98395802e-03
4.16473746e-01 -9.32246596e-02 -2.84983800e-03 2.35372305e-01
3.24360758e-01 -3.28721434e-01 -9.22008157e-02 -1.28188288e+00
1.66358292e-01 3.35785091e-01 1.03589930e-01 8.69122982e-01
-4.25671279e-01 8.83548677e-01 -7.13028789e-01 -8.09529960e-01
5.49205720e-01 1.05663729e+00 -4.69413728e-01 4.39898133e-01
-1.27058953e-01 9.17728245e-01 -2.07163647e-01 9.87665474e-01
6.55364931e-01 -3.17695856e-01 3.39346021e-01 -5.42064667e-01
3.74158710e-01 -4.24502105e-01 -1.57034993e+00 2.01957870e+00
-1.08608752e-01 2.08902638e-02 5.68448901e-01 -3.10376048e-01
5.11189878e-01 3.89828682e-01 6.43152893e-01 -3.13781619e-01
4.06366140e-02 4.79885526e-02 -5.76755106e-01 -3.95237148e-01
2.41206914e-01 -3.53134692e-01 -2.28559285e-01 2.59474874e-01
-2.28665963e-01 -5.18125482e-02 -2.15322480e-01 -1.67846810e-02
6.86349213e-01 6.31198108e-01 1.81171596e-01 -7.58464038e-02
4.59747612e-02 -2.90284842e-01 5.60114920e-01 5.95967412e-01
-2.20925316e-01 1.18051803e+00 -2.67103374e-01 -1.24300504e+00
-1.46750367e+00 -1.56310713e+00 6.36577141e-03 8.00091386e-01
3.25636238e-01 -2.95565903e-01 -6.30549669e-01 -1.83578745e-01
1.92518324e-01 8.93740505e-02 -5.43514609e-01 5.12617707e-01
-7.85070896e-01 -5.94243586e-01 3.22192013e-01 4.82509702e-01
6.46526158e-01 -7.47173965e-01 -6.78551197e-01 1.10723801e-01
-7.63431191e-01 -1.53590631e+00 -5.93697369e-01 -6.03482187e-01
-6.63279116e-01 -1.05594540e+00 -1.01241314e+00 -7.17204630e-01
9.58087146e-01 2.84856439e-01 1.58696473e+00 5.79836182e-02
-5.11177480e-01 6.79302335e-01 1.47561789e-01 1.17976874e-01
-1.00450732e-01 -5.48726559e-01 4.40427661e-01 1.52793348e-01
-1.60112858e-01 -7.80650616e-01 -8.89139771e-01 6.47993684e-01
-2.16194183e-01 4.60434049e-01 2.61910588e-01 5.65209091e-01
9.86275375e-01 -1.49865881e-01 -3.56255332e-03 -5.33793092e-01
2.52077878e-01 -4.16098908e-02 -3.40858370e-01 3.06673646e-01
1.49500579e-01 -5.98216593e-01 2.73377597e-01 -6.07305884e-01
-9.39672947e-01 4.66722280e-01 2.04053652e-02 -7.82791853e-01
-3.48689318e-01 -2.74346888e-01 -2.76657343e-01 -1.33169487e-01
6.92004085e-01 1.85642615e-01 4.97212633e-02 -6.89239025e-01
3.23904574e-01 6.25850111e-02 8.94946396e-01 -8.51490080e-01
8.61944497e-01 7.59454250e-01 7.07757697e-02 -7.89509356e-01
-9.70672488e-01 -1.81209043e-01 -1.20697188e+00 -5.69214225e-01
8.92583847e-01 -1.45927465e+00 -8.22281301e-01 7.52947152e-01
-1.10584390e+00 -3.30694079e-01 -2.14468405e-01 3.18496317e-01
-7.46918619e-01 5.05983174e-01 -8.08061898e-01 -8.17013085e-01
-3.04468066e-01 -9.83382463e-01 1.58280516e+00 -8.23937058e-02
-5.47400475e-01 -1.09106398e+00 -7.79196341e-03 7.36642420e-01
8.42247605e-02 1.05879462e+00 2.24581987e-01 2.88339108e-01
-7.15521693e-01 -3.93326700e-01 9.28206295e-02 -1.01711370e-01
1.81903597e-02 -3.88353884e-01 -7.87121415e-01 -7.61062682e-01
-2.26329654e-01 -4.46484923e-01 4.22024101e-01 5.84298313e-01
7.17535436e-01 -4.66442913e-01 -3.34749758e-01 8.09878290e-01
1.42987120e+00 -7.22643435e-01 4.93477404e-01 -1.26350716e-01
1.18860507e+00 6.16765440e-01 3.52624685e-01 6.29278421e-01
6.68142974e-01 1.12284887e+00 2.93942273e-01 -4.35177803e-01
-4.47394818e-01 -4.76771563e-01 5.81589490e-02 5.14726996e-01
-5.09907067e-01 -3.82694006e-02 -6.54957891e-01 4.56702590e-01
-1.81538785e+00 -8.49183142e-01 -1.27478689e-01 2.35473633e+00
5.98826170e-01 -9.18327346e-02 7.23659158e-01 -1.80399790e-01
6.15732670e-01 1.27574056e-01 -6.92283034e-01 4.87102211e-01
-2.55685121e-01 -3.26932430e-01 2.63133436e-01 6.62051141e-01
-1.02565622e+00 7.73430407e-01 7.13471413e+00 1.83656737e-01
-3.89915943e-01 1.64045811e-01 4.16482240e-01 -5.20226121e-01
-1.71209946e-01 -3.76797140e-01 -6.79843903e-01 2.44473726e-01
2.03211516e-01 5.20487547e-01 6.60218000e-01 5.16540051e-01
2.09389076e-01 6.53523505e-02 -1.03715634e+00 1.45497167e+00
4.05021250e-01 -1.06544459e+00 -6.14047982e-02 2.13225335e-01
8.66494358e-01 -3.32000971e-01 -9.88038257e-02 -1.74115449e-01
4.07911271e-01 -9.49979424e-01 1.18437958e+00 9.46976542e-01
9.55639958e-01 -5.76481462e-01 3.68174136e-01 4.14604157e-01
-1.27847457e+00 2.58377165e-01 -2.48379529e-01 -1.27816260e-01
6.96799636e-01 4.32520121e-01 -2.02769011e-01 2.71999687e-01
8.55697274e-01 6.20330572e-01 -4.63082790e-01 7.98361361e-01
-2.44560279e-02 -1.25341281e-01 -5.25726557e-01 5.30259013e-01
-3.28053206e-01 -1.29734203e-01 7.07073569e-01 1.19131184e+00
1.23577967e-01 2.19985873e-01 7.63453245e-01 8.98364484e-01
9.83778909e-02 -2.91972697e-01 -5.08470774e-01 5.98806739e-01
4.42142099e-01 1.17277837e+00 -5.81019223e-01 -2.71195143e-01
-1.48510695e-01 1.35948157e+00 3.64349335e-01 5.58673501e-01
-9.82589483e-01 5.15419185e-01 6.35555327e-01 9.33195412e-01
2.32908621e-01 -5.81646144e-01 -1.26988098e-01 -1.34237027e+00
2.07204834e-01 -9.86716688e-01 1.86581627e-01 -9.81842399e-01
-1.56589043e+00 5.42779863e-01 2.11121812e-01 -1.09467971e+00
-1.91120133e-01 -2.61652112e-01 -1.34937659e-01 8.55484724e-01
-8.41319621e-01 -1.85516596e+00 -7.78526247e-01 7.94837534e-01
5.88516653e-01 1.67153776e-02 9.74219739e-01 4.54135627e-01
-3.67609769e-01 4.28780198e-01 -2.58868515e-01 9.39031690e-02
7.58857489e-01 -1.13767123e+00 5.97519219e-01 6.31663263e-01
-7.55412728e-02 5.59779227e-01 9.17268634e-01 -8.60254467e-01
-1.70407391e+00 -1.05943990e+00 5.47998965e-01 -8.87862027e-01
-2.13505983e-01 -6.84355497e-01 -5.75538874e-01 1.07436943e+00
-7.31548248e-03 4.09239501e-01 4.40642148e-01 1.81034848e-01
-3.61130893e-01 -1.33427978e-01 -1.44060981e+00 7.44695365e-01
1.43372869e+00 -1.36854112e-01 -3.36576372e-01 4.86552715e-01
3.89275044e-01 -8.85707736e-01 -9.65104461e-01 2.51039952e-01
1.09354711e+00 -1.16171455e+00 1.74238229e+00 -2.50291139e-01
2.14255661e-01 -4.14224088e-01 -4.58415747e-01 -1.06766856e+00
-5.83637238e-01 -6.24077678e-01 -5.39936960e-01 8.24835420e-01
-4.07261968e-01 -1.96437061e-01 7.63683379e-01 8.09303761e-01
3.37563604e-01 -7.81334579e-01 -7.15061188e-01 -4.83672708e-01
-3.68179888e-01 -2.11378768e-01 4.34787214e-01 6.69504464e-01
-6.49963796e-01 2.73199946e-01 -1.29772365e+00 2.37800926e-01
1.58313179e+00 1.06836945e-01 1.34394658e+00 -1.23462451e+00
-4.42663997e-01 2.12743059e-01 -2.19277456e-01 -1.22178614e+00
-1.15079321e-01 -2.39703879e-01 -2.20822021e-02 -1.69033194e+00
5.52690566e-01 -3.46794277e-01 5.30031443e-01 2.12727785e-01
-1.17480510e-03 7.67497122e-01 2.28965223e-01 4.24732983e-01
-8.29157054e-01 4.53499258e-01 1.52508616e+00 1.32734388e-01
8.10388923e-02 5.16773835e-02 -4.96391892e-01 1.17951262e+00
1.64954603e-01 -1.86796546e-01 -2.02367544e-01 -7.41492331e-01
7.86733031e-02 2.64275432e-01 1.16378033e+00 -9.88567173e-01
5.38477732e-04 -2.96056997e-02 1.29218364e+00 -8.30570459e-01
1.15315747e+00 -9.16576207e-01 9.81953859e-01 3.32435906e-01
-5.27576506e-02 2.40893945e-01 8.38263556e-02 8.28678846e-01
5.13871968e-01 4.97907579e-01 9.46782529e-01 -8.48893523e-01
-3.78137559e-01 5.05805552e-01 2.17887878e-01 3.13246213e-02
8.38963568e-01 -4.96726274e-01 2.78473467e-01 -5.96123815e-01
-1.16901863e+00 1.05146714e-01 1.01870143e+00 3.01811010e-01
8.91950309e-01 -1.54049671e+00 -1.04626822e+00 3.70009720e-01
-2.37483382e-01 2.18782157e-01 4.85730976e-01 5.01756847e-01
-7.72554517e-01 -1.58171818e-01 -2.80096143e-01 -9.16996837e-01
-1.54558241e+00 3.76184165e-01 7.48622656e-01 -1.63798124e-01
-1.07316506e+00 5.35699368e-01 2.80113161e-01 -8.16324949e-01
4.07309949e-01 2.15732038e-01 4.28332329e-01 -4.51206446e-01
4.96120572e-01 6.05496049e-01 -4.97436941e-01 -1.21967483e+00
-3.14609230e-01 1.22746921e+00 2.72167534e-01 -1.96185917e-01
1.32143760e+00 -5.20446122e-01 1.51556790e-01 3.27682137e-01
7.69868851e-01 1.18740901e-01 -1.78900981e+00 -5.26652038e-01
-8.79046500e-01 -9.61037159e-01 -4.63121146e-01 -9.51365769e-01
-1.09146821e+00 5.53743362e-01 5.53556502e-01 -3.78160059e-01
8.31848860e-01 9.15493816e-02 9.24183369e-01 1.88227021e-03
8.33027601e-01 -7.85690427e-01 3.52057815e-01 2.13714987e-02
1.26771939e+00 -1.16944528e+00 5.95450401e-01 -5.55157244e-01
-5.51693380e-01 6.70102358e-01 6.96773052e-01 -1.85025007e-01
4.16124642e-01 1.83541402e-01 1.00583099e-01 -4.46596563e-01
-2.76343971e-01 4.39223684e-02 4.10360813e-01 7.85589337e-01
8.00134614e-02 2.29032531e-01 2.58547008e-01 3.62475246e-01
-1.83537945e-01 -2.67347008e-01 9.94434953e-02 7.62233853e-01
-4.94774394e-02 -7.34595180e-01 -8.54636073e-01 -3.66289988e-02
-2.31072202e-01 4.56907332e-01 -3.58820677e-01 7.99774647e-01
2.71731436e-01 7.24435687e-01 -1.67223603e-01 -1.99925035e-01
7.48016238e-01 -2.59081483e-01 1.14417517e+00 -4.36649501e-01
-3.74523103e-01 6.54500127e-01 1.16861627e-01 -6.38848245e-01
-6.78120494e-01 -8.28851581e-01 -7.68173814e-01 -6.51765347e-01
1.30086869e-01 -5.50579906e-01 3.39179963e-01 6.78727627e-01
3.34180176e-01 4.31418449e-01 2.33404309e-01 -1.60457778e+00
-3.19878846e-01 -8.28078628e-01 -6.23229086e-01 9.30220127e-01
4.61962372e-01 -8.36228013e-01 -9.65396017e-02 2.64226168e-01] | [7.1546630859375, -1.1513675451278687] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.