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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
3a438cae-ce5d-4857-a4ab-cf4608d2c2c8 | deep-aesthetic-quality-assessment-with | 1604.04970 | null | http://arxiv.org/abs/1604.04970v3 | http://arxiv.org/pdf/1604.04970v3.pdf | Deep Aesthetic Quality Assessment with Semantic Information | Human beings often assess the aesthetic quality of an image coupled with the
identification of the image's semantic content. This paper addresses the
correlation issue between automatic aesthetic quality assessment and semantic
recognition. We cast the assessment problem as the main task among a multi-task
deep model, and argue that semantic recognition task offers the key to address
this problem. Based on convolutional neural networks, we employ a single and
simple multi-task framework to efficiently utilize the supervision of aesthetic
and semantic labels. A correlation item between these two tasks is further
introduced to the framework by incorporating the inter-task relationship
learning. This item not only provides some useful insight about the correlation
but also improves assessment accuracy of the aesthetic task. Particularly, an
effective strategy is developed to keep a balance between the two tasks, which
facilitates to optimize the parameters of the framework. Extensive experiments
on the challenging AVA dataset and Photo.net dataset validate the importance of
semantic recognition in aesthetic quality assessment, and demonstrate that
multi-task deep models can discover an effective aesthetic representation to
achieve state-of-the-art results. | ['Yueying Kao', 'Kaiqi Huang', 'Ran He'] | 2016-04-18 | null | null | null | null | ['aesthetics-quality-assessment'] | ['computer-vision'] | [ 7.27781877e-02 -1.64252564e-01 3.89591753e-02 -5.73764861e-01
-8.10649514e-01 -2.23111033e-01 4.31336433e-01 2.28701085e-02
-3.44992131e-01 1.86102316e-02 1.97687194e-01 2.85998315e-01
-3.54396731e-01 -7.63848901e-01 -3.18983823e-01 -6.59906566e-01
4.08461541e-01 1.65265173e-01 -1.56876564e-01 -2.25145310e-01
3.36220890e-01 -2.97334138e-03 -1.55886269e+00 3.48218203e-01
1.12824070e+00 1.73385072e+00 1.88882500e-01 1.98655039e-01
-3.56557369e-01 5.91449320e-01 -5.74635744e-01 -8.30564916e-01
2.31118768e-01 -2.84414113e-01 -1.03956032e+00 2.92456508e-01
-1.73797123e-02 -1.24478832e-01 3.92725527e-01 1.18488717e+00
6.73684359e-01 1.77043766e-01 7.90365696e-01 -1.62772667e+00
-9.27903473e-01 3.02055031e-01 -5.71037948e-01 -2.78873593e-01
-5.37126660e-02 2.43819416e-01 1.55551815e+00 -7.87118733e-01
7.60827735e-02 1.27191031e+00 4.89257783e-01 1.78461835e-01
-8.45942438e-01 -4.54786003e-01 -9.19144787e-03 5.79825878e-01
-1.29072356e+00 -2.95825720e-01 1.22645462e+00 -3.59132707e-01
2.63430208e-01 1.47992224e-01 8.86009634e-01 9.30744529e-01
-1.03623278e-01 9.68261421e-01 1.26838124e+00 -3.23803276e-01
2.62608588e-01 8.60951766e-02 -1.66058257e-01 7.27260590e-01
-9.89671424e-02 -1.86046898e-01 -7.46469676e-01 1.09950431e-01
5.95651746e-01 -1.83746979e-01 9.10678599e-03 -3.86711001e-01
-9.32555258e-01 9.41002190e-01 7.00763166e-01 2.57057011e-01
-3.04237664e-01 3.19219083e-01 6.98825061e-01 1.10004604e-01
4.81233239e-01 6.95771456e-01 -1.74260125e-01 -2.02455014e-01
-6.12282813e-01 -2.17409238e-01 3.06034476e-01 5.37004411e-01
9.18638527e-01 -8.37375373e-02 -2.73041964e-01 1.27816939e+00
4.83972430e-01 2.91393727e-01 3.76489758e-01 -1.06240225e+00
8.61593410e-02 1.07642412e+00 -1.11647800e-01 -1.13626242e+00
-2.62264967e-01 -7.15240717e-01 -8.78044069e-01 4.18238223e-01
2.35253498e-01 2.92186141e-01 -4.79605019e-01 1.76614070e+00
2.91360587e-01 -4.01626229e-01 -1.23041786e-01 1.06860328e+00
8.61971855e-01 4.13014978e-01 4.58958894e-01 7.64200017e-02
1.70920718e+00 -1.33648109e+00 -7.90367246e-01 -3.40694904e-01
4.53893542e-01 -7.66825557e-01 1.61563265e+00 3.44438732e-01
-9.84741092e-01 -7.26846099e-01 -1.12211549e+00 -2.83295959e-01
-3.85307699e-01 5.81018329e-01 7.57431388e-01 4.30779576e-01
-1.00872302e+00 5.42378366e-01 -2.47875020e-01 -2.09550798e-01
5.97214222e-01 8.24799836e-02 -1.62722945e-01 1.36283651e-01
-1.15477169e+00 9.79165435e-01 2.43432730e-01 3.32695365e-01
-5.39629161e-01 -2.95154452e-01 -6.46642089e-01 3.27785403e-01
3.46005023e-01 -8.99784982e-01 1.01249731e+00 -1.63262522e+00
-1.50404763e+00 1.16416061e+00 5.85768707e-02 1.67539656e-01
3.87566209e-01 -1.47549599e-01 -3.28266740e-01 2.72149175e-01
2.32932553e-01 6.32588387e-01 8.80257308e-01 -1.41489267e+00
-4.79714692e-01 -3.93070430e-01 2.39267945e-01 4.33590710e-01
-9.88628209e-01 3.83422114e-02 -5.54161668e-01 -5.66271245e-01
-9.67164114e-02 -6.49071634e-01 1.26425987e-02 3.91165525e-01
-2.63005704e-01 -5.50751805e-01 5.04800439e-01 -4.55934554e-01
1.12333894e+00 -2.14102101e+00 3.36059570e-01 2.09738418e-01
2.46854246e-01 4.58441041e-02 -4.30574924e-01 2.28061602e-01
1.96543843e-01 2.18736902e-01 -9.06282291e-02 -6.30782008e-01
3.03344220e-01 1.11061506e-01 1.34857625e-01 8.15310478e-02
3.31981778e-01 1.31702912e+00 -6.77109480e-01 -6.92743778e-01
8.98252055e-02 3.63662213e-01 -2.38801166e-01 4.84838337e-01
-8.34261328e-02 3.31432104e-01 -7.51791477e-01 7.08867252e-01
5.33091843e-01 -7.71171987e-01 -5.70760854e-02 -7.26392567e-01
2.16609791e-01 -3.67426267e-03 -9.39051032e-01 1.98652542e+00
-8.27650726e-01 2.88921326e-01 -3.48984264e-02 -1.02187324e+00
1.25698721e+00 8.80642459e-02 5.38000345e-01 -1.12872136e+00
4.17644888e-01 2.08650276e-01 -1.80654943e-01 -7.30563164e-01
4.30373818e-01 -3.50800455e-01 -2.04410180e-01 6.02548420e-01
-4.95239571e-02 -1.57048211e-01 -1.18669689e-01 -8.41528773e-02
6.54207647e-01 2.35998541e-01 5.43324590e-01 -3.23048443e-01
6.86243415e-01 -3.03413182e-01 4.28884178e-01 1.58876106e-01
-5.31679153e-01 3.96366566e-01 3.90868157e-01 -5.55381238e-01
-9.54518080e-01 -8.14418912e-01 9.96562392e-02 1.36261964e+00
5.30014396e-01 -4.36016798e-01 -6.58356965e-01 -6.92777157e-01
-2.80550629e-01 2.65237868e-01 -9.47157562e-01 -4.52484548e-01
-1.71251029e-01 -8.33877325e-01 2.85977483e-01 4.50057209e-01
7.80014932e-01 -1.19830155e+00 -5.76368332e-01 -9.95432660e-02
-6.61789179e-01 -1.11805618e+00 -3.18689108e-01 4.17420156e-02
-4.71124649e-01 -1.12172771e+00 -6.32593930e-01 -5.43309927e-01
6.75050199e-01 3.07349920e-01 1.21591580e+00 3.62526298e-01
-1.91310063e-01 4.36802864e-01 -6.24380112e-01 -1.76224425e-01
-2.57732958e-01 5.78435184e-03 -4.40320283e-01 3.93175453e-01
2.34401345e-01 -5.94528139e-01 -8.91877353e-01 4.44568515e-01
-1.01001072e+00 2.87237406e-01 8.93769085e-01 8.01547289e-01
5.45876503e-01 3.01402919e-02 6.99156106e-01 -3.39828700e-01
7.66426623e-01 -2.77504884e-02 -1.43526286e-01 6.61191165e-01
-8.24497402e-01 2.00043082e-01 5.34538150e-01 -1.30941316e-01
-9.79514062e-01 -4.74708825e-02 -2.63132602e-01 -4.05313611e-01
6.77225515e-02 5.33648312e-01 -6.19648397e-01 -3.34569037e-01
1.56218752e-01 7.02924356e-02 2.03812271e-02 -3.03521097e-01
5.58825195e-01 5.49300134e-01 3.01297277e-01 -8.25144053e-01
6.79307878e-01 5.02859950e-01 1.14045091e-01 -4.21768397e-01
-1.29984951e+00 -5.07189631e-01 -7.13289857e-01 -6.20346725e-01
9.38166857e-01 -9.04161334e-01 -9.95947182e-01 5.16632140e-01
-1.21287870e+00 -9.17103812e-02 -3.63081008e-01 -1.03885412e-01
-6.86496556e-01 5.64513505e-01 -2.91806340e-01 -8.47513080e-01
-7.00121760e-01 -1.31118011e+00 1.35698020e+00 1.58106625e-01
-1.64298296e-01 -1.03780890e+00 -1.22345999e-01 9.38510180e-01
4.56987679e-01 1.53486878e-01 1.07609320e+00 -3.05913597e-01
-4.73676354e-01 3.01565919e-02 -8.82630050e-01 5.53931475e-01
1.00119755e-01 -1.22508399e-01 -1.20000339e+00 8.95520747e-02
7.43004009e-02 -8.54751706e-01 9.50717986e-01 2.59075940e-01
1.61589754e+00 -2.73981959e-01 3.28774542e-01 7.43970573e-01
1.57022536e+00 -1.81935653e-01 7.64985144e-01 6.93671525e-01
8.27181220e-01 9.05281603e-01 7.65550315e-01 5.45426011e-01
8.61034393e-01 8.84698749e-01 7.34038174e-01 -4.04246420e-01
-2.02493697e-01 -1.07313216e-01 7.79094100e-02 1.01756036e+00
-1.68183371e-01 -6.65046880e-03 -7.62359619e-01 1.81689665e-01
-1.97295940e+00 -7.17381775e-01 4.83447500e-02 1.87362623e+00
8.63310337e-01 -1.74269661e-01 2.61125177e-01 2.39909828e-01
6.69387579e-01 3.16973269e-01 -4.51079130e-01 -3.74880403e-01
-1.57154992e-01 -4.64349687e-02 -6.01693168e-02 1.55108541e-01
-1.30028200e+00 9.30247188e-01 5.79237509e+00 1.26950157e+00
-9.60563004e-01 1.82443127e-01 9.55460489e-01 1.92452580e-01
-5.58406055e-01 -1.98089108e-01 -1.94514036e-01 4.56634700e-01
4.83077317e-01 -2.05311403e-01 4.43712443e-01 8.42159688e-01
2.07507744e-01 -4.58604470e-02 -9.48416233e-01 1.23430192e+00
2.14659795e-01 -7.77803898e-01 3.35408628e-01 4.02046330e-02
4.35099483e-01 -6.46411419e-01 3.12262595e-01 1.60895914e-01
-1.86473299e-02 -1.01300228e+00 9.28380609e-01 4.85459715e-01
8.53776634e-01 -7.79805183e-01 7.20130444e-01 -9.21968594e-02
-1.53282011e+00 -1.67085797e-01 -3.93894643e-01 -8.25941041e-02
2.01945245e-01 7.76255012e-01 -2.43534565e-01 7.07881749e-01
7.58121312e-01 8.37117255e-01 -8.88705969e-01 9.48074043e-01
-3.94173265e-01 -1.80149060e-02 1.76979318e-01 -2.73073614e-02
1.67810321e-01 -4.61816221e-01 4.60116528e-02 8.20738852e-01
3.70289445e-01 -1.25021413e-01 1.38005450e-01 9.98091459e-01
-1.33784130e-01 5.28376162e-01 -1.76811427e-01 -1.58976391e-01
4.48247679e-02 1.72761595e+00 -8.49833190e-01 -9.05621052e-02
-2.24353850e-01 1.14609420e+00 6.35821521e-01 1.51708588e-01
-8.57365370e-01 -8.78182054e-02 5.47610998e-01 -3.15172285e-01
3.41008604e-02 6.55801967e-03 -9.45533812e-01 -1.06853998e+00
1.71066180e-01 -7.65158832e-01 3.18041950e-01 -1.00213587e+00
-1.46426857e+00 5.75228691e-01 -5.24482846e-01 -1.50439000e+00
2.65970886e-01 -5.94562411e-01 -7.75073230e-01 6.14826739e-01
-1.98811734e+00 -1.74429762e+00 -7.14318573e-01 5.31102121e-01
3.79907638e-01 1.19689479e-01 9.31485415e-01 3.37973207e-01
-4.53866899e-01 5.56736052e-01 -1.86532468e-01 4.13065441e-02
6.79666221e-01 -1.23270404e+00 8.11666548e-02 6.23292863e-01
-4.93478850e-02 5.09841554e-02 4.47319925e-01 -2.26261660e-01
-1.07813382e+00 -8.36556494e-01 7.05051601e-01 -1.48270831e-01
7.30548084e-01 -1.81817308e-01 -7.38348544e-01 -2.13279158e-01
2.41171896e-01 -2.87027508e-01 1.05693388e+00 4.00402606e-01
-6.15136147e-01 -2.80350029e-01 -7.02941179e-01 3.41189295e-01
1.09202302e+00 -6.52554691e-01 -3.27522308e-01 1.67799354e-01
5.56958675e-01 2.39697799e-01 -9.56513941e-01 4.12023842e-01
6.99689984e-01 -1.14430797e+00 1.14125180e+00 -2.73108304e-01
1.06895566e+00 -6.87135309e-02 -1.31015956e-01 -1.44847691e+00
-4.36415493e-01 -7.86173791e-02 1.11195676e-01 1.42320418e+00
2.26926297e-01 -1.99798360e-01 5.35671771e-01 6.19898856e-01
2.86540743e-02 -9.13294017e-01 -8.16283584e-01 -6.22126400e-01
-6.43477365e-02 -3.26983362e-01 6.00679040e-01 8.07070434e-01
-1.70085877e-01 6.15558267e-01 -5.95269740e-01 -1.76343426e-01
6.69330359e-01 5.07752597e-01 5.66043317e-01 -1.53846860e+00
-1.14793234e-01 -9.32602286e-01 -3.28762680e-01 -6.83027387e-01
2.50514448e-01 -8.40765357e-01 1.10507784e-02 -1.77163327e+00
6.01761281e-01 -4.93423849e-01 -6.43290937e-01 6.09060884e-01
-5.14171600e-01 3.19520295e-01 2.92949706e-01 4.22106713e-01
-1.17212129e+00 1.25787568e+00 1.57906210e+00 -3.15777451e-01
1.67231321e-01 -8.73199180e-02 -1.26170826e+00 6.70278072e-01
7.45336056e-01 -1.33007765e-01 -4.92991418e-01 -6.74802721e-01
6.98842049e-01 -2.09675923e-01 5.69683611e-01 -7.95342386e-01
1.30813077e-01 -2.80193835e-01 2.98970342e-01 -5.57821691e-02
6.40904009e-01 -9.73056316e-01 -2.59550750e-01 2.38336340e-01
-5.29066622e-01 -1.17450943e-02 -1.93128183e-01 4.85413224e-01
-4.10569727e-01 -8.81447345e-02 1.01726580e+00 -2.35503808e-01
-9.08642948e-01 4.29775298e-01 4.00390267e-01 3.22753415e-02
8.68761480e-01 -1.42824680e-01 -3.44093561e-01 -3.75781476e-01
-6.05407298e-01 3.29933703e-01 4.45098519e-01 5.70052385e-01
7.51301467e-01 -1.69792676e+00 -5.29265881e-01 -3.91558819e-02
5.60710073e-01 -3.11908454e-01 5.13695240e-01 6.45173132e-01
-1.50952309e-01 -1.31454803e-02 -5.96683741e-01 -5.12455881e-01
-1.20808470e+00 3.56133968e-01 3.03028911e-01 -4.05293345e-01
-1.69174165e-01 8.06467891e-01 4.09849137e-01 -2.51887590e-01
2.70838410e-01 1.27619043e-01 -5.82618356e-01 4.05414999e-01
3.43790025e-01 3.35178047e-01 -6.76671937e-02 -7.37141192e-01
-2.67420173e-01 8.83653343e-01 2.83030063e-01 7.74032921e-02
1.43782401e+00 -5.46241581e-01 -2.95178741e-01 4.20730740e-01
1.33661604e+00 -5.39924860e-01 -1.09939039e+00 -2.76992381e-01
-1.62348419e-03 -6.30047202e-01 2.14687288e-01 -9.85432267e-01
-1.36072969e+00 1.21460247e+00 5.63052773e-01 1.57316655e-01
1.59498680e+00 8.15746933e-02 9.16071117e-01 1.69889495e-01
2.58238763e-01 -1.47103345e+00 8.94607186e-01 8.41877162e-02
1.02483845e+00 -1.62419021e+00 4.42442968e-02 -4.51943338e-01
-9.86966908e-01 1.20655143e+00 7.83937931e-01 1.01846613e-01
3.79034281e-01 -2.16353238e-01 1.98342755e-01 -4.24435556e-01
-4.69413549e-01 -4.77925211e-01 7.68674672e-01 3.43130201e-01
2.34456420e-01 1.41591027e-01 -3.37241679e-01 9.94034588e-01
-1.08549342e-01 -1.27028555e-01 1.29192606e-01 4.22099769e-01
-5.86765170e-01 -1.12925649e+00 3.76788825e-02 1.15246668e-01
-3.07509989e-01 -4.99767512e-02 -4.78375673e-01 3.06428343e-01
1.09738640e-01 9.02393341e-01 -2.78570145e-01 -5.81959546e-01
2.67188728e-01 4.02778313e-02 1.93049058e-01 -2.20640287e-01
-6.36487722e-01 1.48104161e-01 -3.33367251e-02 -8.51979613e-01
-7.26904571e-01 -2.75465250e-01 -7.23695099e-01 -1.93821535e-01
-2.19551831e-01 2.21141353e-02 8.93467128e-01 1.15811300e+00
2.67273068e-01 7.20290482e-01 9.46125984e-01 -6.04553461e-01
-5.73031127e-01 -6.99462175e-01 -6.78467095e-01 7.31447041e-01
-4.45247516e-02 -7.44275033e-01 -3.80332142e-01 -9.39023942e-02] | [11.526801109313965, -1.0289933681488037] |
b564fab6-fd0f-488f-ac9f-8219c73b2e41 | enhanced-center-coding-for-cell-detection | 1904.08864 | null | http://arxiv.org/abs/1904.08864v1 | http://arxiv.org/pdf/1904.08864v1.pdf | Enhanced Center Coding for Cell Detection with Convolutional Neural Networks | Cell imaging and analysis are fundamental to biomedical research because
cells are the basic functional units of life. Among different cell-related
analysis, cell counting and detection are widely used. In this paper, we focus
on one common step of learning-based cell counting approaches: coding the raw
dot labels into more suitable maps for learning. Two criteria of coding raw dot
labels are discussed, and a new coding scheme is proposed in this paper. The
two criteria measure how easy it is to train the model with a coding scheme,
and how robust the recovered raw dot labels are when predicting. The most
compelling advantage of the proposed coding scheme is the ability to
distinguish neighboring cells in crowded regions. Cell counting and detection
experiments are conducted for five coding schemes on four types of cells and
two network architectures. The proposed coding scheme improves the counting
accuracy versus the widely-used Gaussian and rectangle kernels up to 12%, and
also improves the detection accuracy versus the common proximity coding up to
14%. | ['Jaideep Kapur', 'Cedric L. Williams', 'Daniel S. Weller', 'Aijaz Naik', 'Haoyi Liang'] | 2019-04-18 | null | null | null | null | ['cell-detection'] | ['computer-vision'] | [-2.70980094e-02 -2.50649571e-01 1.20585091e-01 -8.69232491e-02
-3.93233538e-01 -2.77688622e-01 5.47116816e-01 5.71963072e-01
-8.54006886e-01 1.00611782e+00 -1.86737522e-01 -1.35543838e-01
1.98362157e-01 -7.58971155e-01 -2.64464676e-01 -1.13026035e+00
-8.03760961e-02 4.82507944e-01 5.21806896e-01 2.32326269e-01
5.60788035e-01 8.16822469e-01 -1.22229671e+00 1.30498290e-01
7.70589232e-01 9.86155629e-01 1.45754650e-01 9.59175825e-01
-2.04042688e-01 8.89593601e-01 -5.73582172e-01 -2.89979223e-02
-5.04089557e-02 -1.91661969e-01 -4.98253167e-01 -2.35351831e-01
3.51779126e-02 -5.02863191e-02 -1.00090802e-01 9.55481887e-01
6.44843817e-01 -1.97519347e-01 1.34153402e+00 -7.70193756e-01
-6.71184778e-01 -6.72956044e-03 -6.84970796e-01 7.17103004e-01
1.92630917e-01 -3.77022028e-02 4.22620028e-01 -1.09112358e+00
5.26702583e-01 9.76419270e-01 7.80024171e-01 5.14465809e-01
-1.24988210e+00 -5.36889613e-01 -4.08116430e-01 -1.28809243e-01
-1.70873523e+00 -2.64930487e-01 2.45814309e-01 -8.97842109e-01
7.52622724e-01 1.34756789e-01 5.57485104e-01 3.88838083e-01
5.63042641e-01 5.13820112e-01 1.33542621e+00 -5.73908567e-01
4.31924313e-01 1.61207184e-01 3.90431553e-01 8.83757114e-01
5.13580382e-01 -1.21933430e-01 -1.10659435e-01 -7.09574595e-02
1.14184046e+00 3.20079058e-01 -2.42066383e-01 -7.22853243e-02
-1.41322446e+00 8.19054902e-01 4.73293453e-01 7.24620283e-01
-1.06792673e-02 3.00518066e-01 3.34196091e-01 -2.08776116e-01
5.32576680e-01 4.96616483e-01 -1.73493810e-02 6.16264641e-02
-1.00663590e+00 -4.44997773e-02 6.17474556e-01 6.54644787e-01
5.08728027e-01 -5.00948615e-02 -5.00059843e-01 9.40423012e-01
4.11447100e-02 4.91636246e-01 5.05527973e-01 -8.24355543e-01
-1.42062590e-01 7.58283973e-01 1.48277581e-01 -1.11273611e+00
-7.51882255e-01 -3.27600777e-01 -1.13305354e+00 3.66892755e-01
1.10586405e+00 -1.34761393e-01 -8.20292175e-01 1.25344050e+00
1.00133680e-01 2.86430717e-01 -1.26396805e-01 6.69894755e-01
7.48504162e-01 7.29998827e-01 1.74275581e-02 -2.37704858e-01
1.29644692e+00 -8.03400338e-01 -5.65285563e-01 1.39436409e-01
1.05692279e+00 -5.35340190e-01 6.23761058e-01 1.58768192e-01
-7.62439549e-01 -5.02652645e-01 -8.71119738e-01 -1.59201249e-01
-6.95717156e-01 3.95369977e-01 6.19540393e-01 5.06477296e-01
-1.08772802e+00 5.06739080e-01 -7.30318367e-01 -6.34533882e-01
7.00343549e-01 4.16354597e-01 -3.54123771e-01 -1.28332391e-01
-5.89034319e-01 8.50021541e-01 2.18718082e-01 -2.31811345e-01
-5.56339502e-01 -5.45397103e-01 -4.81880277e-01 1.34336919e-01
-3.48720551e-01 -4.48647797e-01 7.92945445e-01 -3.10180008e-01
-1.10618424e+00 1.17578423e+00 -3.46932709e-01 -4.95743155e-01
5.17477512e-01 2.69430012e-01 -6.19129241e-02 2.98935294e-01
2.18982771e-01 9.11864221e-01 2.40882218e-01 -1.11964524e+00
-7.46412635e-01 -2.63850927e-01 -3.15904945e-01 -2.12037805e-02
-3.04413050e-01 -1.96334288e-01 -3.63160342e-01 -5.74214220e-01
-5.23219556e-02 -7.61885822e-01 -1.73737153e-01 5.56917787e-01
-2.25050479e-01 -1.55253038e-01 6.19455814e-01 -3.87573957e-01
9.60835218e-01 -2.05344534e+00 -1.59823090e-01 5.05161174e-02
6.50136054e-01 2.45496869e-01 1.66445017e-01 2.57247269e-01
3.25138807e-01 -1.13213643e-01 -1.64311111e-01 -3.01312149e-01
-4.84487504e-01 -1.28672682e-02 6.97413981e-02 7.73425639e-01
1.30315632e-01 8.42800379e-01 -8.51140082e-01 -8.30625117e-01
2.80683994e-01 7.62861669e-01 -2.14139432e-01 5.96625358e-03
3.47552717e-01 6.71008527e-01 -2.34641761e-01 8.52899492e-01
5.40511131e-01 -5.79421937e-01 -3.64197582e-01 -8.83599892e-02
-2.10529566e-01 -4.13416654e-01 -7.30821431e-01 9.70366478e-01
-1.13104053e-01 8.19707930e-01 -3.73271018e-01 -9.28830385e-01
1.19440150e+00 -1.58657283e-02 5.29627919e-01 -5.55465639e-01
2.91208565e-01 4.10986692e-01 1.59294549e-02 -2.39148974e-01
2.95876078e-02 -1.95821509e-01 7.97430649e-02 2.55916893e-01
-6.34554327e-02 1.44844517e-01 3.76757473e-01 2.62070242e-02
1.11648250e+00 -4.60901678e-01 6.88631475e-01 -7.24675417e-01
7.55600572e-01 -1.39355466e-01 4.90175873e-01 5.34930885e-01
-5.41851640e-01 6.56323671e-01 5.87882221e-01 -8.43261540e-01
-9.47225153e-01 -8.49243462e-01 -4.81887281e-01 8.92333925e-01
2.05962569e-01 1.70279443e-01 -7.61187017e-01 -3.96155566e-01
1.80256411e-01 2.97883689e-01 -9.24864769e-01 8.14520717e-02
-3.92509609e-01 -8.30775797e-01 8.51630330e-01 7.04874933e-01
5.40281057e-01 -8.72984886e-01 -8.18187714e-01 2.56145187e-02
-7.52530620e-02 -8.65206957e-01 -3.31521779e-01 5.29387176e-01
-8.60353291e-01 -9.41955566e-01 -1.32669175e+00 -1.12167084e+00
9.58497941e-01 5.44890523e-01 6.40930176e-01 3.99509639e-01
-6.45467103e-01 -4.48882133e-02 -1.25261441e-01 -3.92289847e-01
-1.26741871e-01 -6.24599718e-02 7.37952814e-02 -5.13095558e-02
5.96477985e-01 -1.57996625e-01 -6.81915820e-01 3.02809238e-01
-4.04343396e-01 1.16743315e-02 4.81768847e-01 1.04844618e+00
8.11545789e-01 -2.40095064e-01 5.78679323e-01 -1.03444469e+00
4.01392579e-01 -2.22156703e-01 -6.00938082e-01 1.70493335e-01
-4.35250610e-01 6.79057464e-03 7.71042466e-01 -3.77523452e-01
-6.97633266e-01 1.90678775e-01 2.69935429e-02 -1.77838504e-01
-1.86064273e-01 1.38277888e-01 4.49374735e-01 -5.89940310e-01
8.05379272e-01 2.84710437e-01 1.02396585e-01 1.41911851e-02
1.20902527e-02 4.35604364e-01 4.46378559e-01 -2.74480909e-01
3.53753328e-01 9.05736804e-01 2.20197871e-01 -9.89069581e-01
-3.88773710e-01 -8.85115325e-01 -8.12858105e-01 -3.61560971e-01
1.03349662e+00 -6.94924474e-01 -8.26209247e-01 7.25951970e-01
-1.18301833e+00 -3.18597406e-01 -2.01918587e-01 4.77223217e-01
-4.43631619e-01 3.84662598e-01 -1.01354468e+00 -9.05438900e-01
-9.67086107e-02 -1.05689359e+00 1.23686814e+00 4.33406293e-01
-9.43309814e-02 -1.41569805e+00 2.75279880e-01 -3.77428532e-02
4.47526515e-01 3.55369180e-01 9.89819586e-01 -6.20432556e-01
-3.02710056e-01 -3.53675187e-01 -8.04572284e-01 -2.19172731e-01
6.78872541e-02 2.97458991e-02 -1.26297414e+00 -4.43164587e-01
-1.64140597e-01 -1.67058170e-01 1.15230775e+00 8.85538042e-01
1.23984671e+00 1.60508946e-01 -1.04748642e+00 6.60845280e-01
1.55785608e+00 4.85673636e-01 6.64408445e-01 -3.45338285e-02
5.50607324e-01 5.75851500e-01 3.49209934e-01 2.61628866e-01
-2.14866716e-02 4.14017349e-01 2.37176433e-01 -4.66567904e-01
-1.51964009e-01 1.42875969e-01 -2.35797301e-01 1.03493059e+00
-1.92858651e-01 -2.38916159e-01 -1.12322736e+00 4.33361769e-01
-1.68292069e+00 -9.39239502e-01 -2.02089742e-01 2.17474294e+00
5.17085433e-01 1.13127586e-02 8.17177221e-02 2.53330499e-01
9.13148046e-01 -2.57222056e-01 -4.91206735e-01 -1.16432481e-01
-1.50687724e-01 1.50378386e-03 6.95194185e-01 5.91303170e-01
-1.16066039e+00 6.36845887e-01 6.84551620e+00 1.01216114e+00
-1.17266655e+00 2.40572356e-02 1.34667504e+00 1.44439816e-01
3.21092963e-01 -4.63850051e-01 -1.03609538e+00 6.11835897e-01
5.01586080e-01 6.91220090e-02 4.02800143e-02 6.74508870e-01
5.19976020e-02 -5.99487722e-01 -1.26953590e+00 1.42161608e+00
1.11866951e-01 -1.60103059e+00 2.04964820e-02 3.09716433e-01
6.46874249e-01 -1.80431753e-01 3.18672992e-02 1.89742342e-01
1.63749188e-01 -1.52070582e+00 4.79537696e-01 7.75660276e-01
1.09482920e+00 -4.33853686e-01 1.18574357e+00 5.74179888e-01
-1.26168668e+00 -1.10941209e-01 -7.47285008e-01 -2.49507800e-01
1.01465598e-01 7.09838092e-01 -7.39115953e-01 -1.94683984e-01
5.51721752e-01 7.19098747e-01 -8.33282948e-01 1.45661962e+00
4.31996405e-01 4.12425697e-01 -3.23516637e-01 -4.47746277e-01
7.02649802e-02 -2.79959917e-01 1.91652235e-02 1.74429691e+00
5.76235890e-01 7.94671178e-02 5.21651469e-02 7.05443621e-01
-3.52859050e-02 1.04435280e-01 -8.70996416e-01 2.32320204e-01
6.27248943e-01 1.38508701e+00 -1.50306034e+00 -3.29883188e-01
-4.05824035e-01 8.64981651e-01 7.08440185e-01 1.83499932e-01
-8.21900010e-01 -4.93887842e-01 2.09506422e-01 3.16806614e-01
7.91180059e-02 -1.09423734e-01 -6.08967125e-01 -6.36509001e-01
-4.98762101e-01 -1.69538617e-01 1.61543250e-01 -5.93888640e-01
-1.36469471e+00 3.11088771e-01 -2.98093379e-01 -1.15074122e+00
3.76831859e-01 -1.01621950e+00 -6.47097945e-01 6.63809299e-01
-1.36511421e+00 -8.71624708e-01 -5.79228461e-01 3.16112101e-01
4.35822427e-01 -1.39726147e-01 9.38592434e-01 4.25545275e-01
-4.18536484e-01 4.62375939e-01 4.66481119e-01 4.01997983e-01
5.27097046e-01 -1.33831728e+00 -6.33969717e-03 4.40203995e-01
1.51833206e-01 5.23197234e-01 3.12913835e-01 -3.53665948e-01
-6.65695667e-01 -1.08502936e+00 1.04533780e+00 -3.00060064e-01
2.60306180e-01 -3.24304491e-01 -1.04418755e+00 1.77713245e-01
-1.24401346e-01 4.74456549e-01 8.35555434e-01 -4.03721988e-01
-2.82711565e-01 -9.69143957e-02 -1.22726977e+00 5.31076908e-01
6.12764120e-01 -5.31675220e-01 -1.59553230e-01 3.87710810e-01
1.88191056e-01 -1.79323077e-01 -6.31246924e-01 6.67161569e-02
6.23067439e-01 -1.09734082e+00 9.40883577e-01 -1.01729006e-01
1.43947661e-01 -5.19980192e-01 -5.81953414e-02 -1.14820743e+00
-8.78824651e-01 -9.75764543e-03 1.75997406e-01 8.60823393e-01
4.35455084e-01 -7.34832585e-01 9.35166240e-01 8.74655768e-02
1.12111844e-01 -1.00473547e+00 -1.03120434e+00 -7.42461562e-01
4.01869327e-01 2.06560761e-01 5.08046784e-02 9.74659920e-01
3.78311455e-01 4.05589968e-01 2.06193067e-02 -1.88661024e-01
5.35478592e-01 -3.05912256e-01 5.28483629e-01 -1.58202040e+00
-6.57410696e-02 -6.86695516e-01 -8.74250889e-01 -1.17247295e+00
-2.85956055e-01 -8.63480628e-01 6.58531934e-02 -1.54808104e+00
5.05994856e-01 -4.85184222e-01 -3.64538282e-01 1.37623891e-01
-7.08092600e-02 6.55191243e-01 -1.69591215e-02 4.97027248e-01
-6.38488531e-01 -5.62594086e-02 1.12113833e+00 -1.40135512e-01
1.13585979e-01 -2.51608759e-01 -2.37220407e-01 8.21330309e-01
8.38341713e-01 -2.92608589e-01 -1.24634266e-01 -3.44450504e-01
-5.41894417e-03 -3.06852795e-02 4.77871180e-01 -1.57092071e+00
6.22799397e-01 1.83926210e-01 7.48352349e-01 -4.72772360e-01
3.44025970e-01 -6.21749222e-01 7.89896324e-02 8.85118842e-01
-4.50631499e-01 -6.20352291e-02 6.61139563e-02 7.46967494e-01
-1.26937166e-01 -1.70610979e-01 1.43100142e+00 -1.02854833e-01
-4.54368323e-01 -8.14485271e-03 -6.72794521e-01 -1.66447982e-02
1.45045209e+00 -8.48220289e-01 -5.01040220e-01 -1.92580536e-01
-5.63820601e-01 -8.50513503e-02 4.85750556e-01 -3.83818656e-01
5.40779829e-01 -1.42336476e+00 -5.20194709e-01 1.65555447e-01
3.51989448e-01 -2.85811156e-01 -8.86242464e-02 1.18806291e+00
-1.20844638e+00 8.15882146e-01 -4.29665238e-01 -9.17553484e-01
-1.11873770e+00 5.60207367e-01 4.61020917e-01 -3.87022018e-01
-3.35558444e-01 1.06898510e+00 6.04293764e-01 8.09814257e-04
2.36030325e-01 -7.02945232e-01 -5.98122656e-01 -1.18422292e-01
6.48298919e-01 7.95319140e-01 -1.69784039e-01 -8.26662481e-01
-3.43603522e-01 1.01906598e+00 -1.50516674e-01 3.94572586e-01
8.54840815e-01 -3.37932669e-02 -3.77137549e-02 8.50662947e-01
1.11910033e+00 -1.78953856e-01 -1.33740580e+00 -4.07533385e-02
1.03123458e-02 -3.57401013e-01 -1.22383654e-01 -4.51779515e-01
-9.52746391e-01 1.27434635e+00 7.14783967e-01 4.38699871e-01
6.55142009e-01 -7.87649900e-02 4.68717188e-01 3.22287917e-01
6.24712169e-01 -9.47930872e-01 1.31326571e-01 5.47028840e-01
3.60502601e-01 -1.19652879e+00 -5.67832142e-02 -6.48993433e-01
-4.08838868e-01 1.11785316e+00 5.82682371e-01 -3.02005231e-01
7.56674409e-01 5.59572339e-01 9.80407596e-02 -3.79329681e-01
-4.21504110e-01 -1.82581335e-01 -7.52328783e-02 1.00334716e+00
7.47602761e-01 2.35656332e-02 -2.80804992e-01 2.37832651e-01
3.22223634e-01 1.27809182e-01 3.54865164e-01 5.22994041e-01
-9.10223961e-01 -3.30826789e-01 -2.97431797e-01 8.32646728e-01
-5.10585904e-01 3.29743996e-02 -3.23578089e-01 6.75125301e-01
2.54183888e-01 6.17517650e-01 3.01754475e-01 -4.48639728e-02
-1.51555970e-01 -8.68597925e-02 2.65700668e-01 -6.16162539e-01
-2.48155490e-01 1.96302515e-02 -4.70248699e-01 -3.10232759e-01
-3.69501740e-01 -5.02290010e-01 -1.59074843e+00 -4.17655468e-01
-5.95463932e-01 5.00852540e-02 7.41523737e-03 8.28087866e-01
1.04316324e-01 2.97701091e-01 1.88214734e-01 -7.90190876e-01
-1.03083961e-01 -7.91206181e-01 -9.03413475e-01 2.22045660e-01
1.89318925e-01 -8.91048253e-01 -3.78825665e-01 3.24988961e-01] | [14.690690040588379, -3.2046170234680176] |
f66b7e43-fdef-4106-8e76-8bb516635b59 | recent-advances-in-the-applications-of | 1708.07281 | null | http://arxiv.org/abs/1708.07281v1 | http://arxiv.org/pdf/1708.07281v1.pdf | Recent Advances in the Applications of Convolutional Neural Networks to Medical Image Contour Detection | The fast growing deep learning technologies have become the main solution of
many machine learning problems for medical image analysis. Deep convolution
neural networks (CNNs), as one of the most important branch of the deep
learning family, have been widely investigated for various computer-aided
diagnosis tasks including long-term problems and continuously emerging new
problems. Image contour detection is a fundamental but challenging task that
has been studied for more than four decades. Recently, we have witnessed the
significantly improved performance of contour detection thanks to the
development of CNNs. Beyond purusing performance in existing natural image
benchmarks, contour detection plays a particularly important role in medical
image analysis. Segmenting various objects from radiology images or pathology
images requires accurate detection of contours. However, some problems, such as
discontinuity and shape constraints, are insufficiently studied in CNNs. It is
necessary to clarify the challenges to encourage further exploration. The
performance of CNN based contour detection relies on the state-of-the-art CNN
architectures. Careful investigation of their design principles and motivations
is critical and beneficial to contour detection. In this paper, we first review
recent development of medical image contour detection and point out the current
confronting challenges and problems. We discuss the development of general CNNs
and their applications in image contours (or edges) detection. We compare those
methods in detail, clarify their strengthens and weaknesses. Then we review
their recent applications in medical image analysis and point out limitations,
with the goal to light some potential directions in medical image analysis. We
expect the paper to cover comprehensive technical ingredients of advanced CNNs
to enrich the study in the medical image domain. | ['Lin Yang', 'Fuyong Xing', 'Zizhao Zhang', 'Xiaoshuang Shi', 'Hai Su'] | 2017-08-24 | null | null | null | null | ['contour-detection'] | ['computer-vision'] | [ 3.88208270e-01 1.36845931e-01 -2.71054417e-01 -3.34516734e-01
-2.36095637e-01 -2.77230322e-01 1.61398295e-02 3.56305242e-01
-6.19732320e-01 4.72075045e-01 -2.72589177e-01 -5.68182886e-01
2.92931367e-02 -8.28375399e-01 -2.56626248e-01 -8.54232013e-01
-4.34316695e-01 8.36816132e-02 3.48258406e-01 -3.77573937e-01
8.64730850e-02 8.19588244e-01 -1.16182613e+00 1.59894317e-01
7.49549985e-01 1.30205488e+00 2.64327717e-03 6.90451562e-01
-3.47563356e-01 4.09632564e-01 -5.83278596e-01 -4.17431146e-01
5.58135472e-03 -4.68140513e-01 -7.48159051e-01 1.20042726e-01
1.14009351e-01 -8.56116489e-02 -1.01783872e-01 1.32147014e+00
7.27477252e-01 -1.72663614e-01 5.21701217e-01 -8.99832189e-01
-6.18483007e-01 1.99487537e-01 -7.35110462e-01 4.05581683e-01
-3.06264311e-01 -4.08710301e-01 4.61209685e-01 -7.83126354e-01
5.66486001e-01 7.12676406e-01 1.04733813e+00 7.33448505e-01
-6.26856267e-01 -4.48760897e-01 -3.66655029e-02 8.96758810e-02
-1.13458765e+00 1.09224297e-01 9.72118199e-01 -2.54524052e-01
5.43570638e-01 2.79791623e-01 9.60227430e-01 4.04982477e-01
6.99686348e-01 1.10930848e+00 7.09940076e-01 -4.16991949e-01
6.37087598e-02 -1.15364574e-01 2.32607067e-01 1.05787563e+00
4.04161572e-01 1.14614382e-01 8.66400898e-02 1.05724297e-01
1.24435711e+00 3.68382931e-02 -4.89472777e-01 -3.64230812e-01
-1.06402349e+00 1.09122145e+00 8.76943290e-01 6.42191112e-01
-2.04577118e-01 1.26852661e-01 5.10803640e-01 2.25460832e-03
4.89285767e-01 2.61309147e-01 -4.13745582e-01 3.93947840e-01
-1.04180872e+00 9.80728790e-02 3.72315824e-01 8.00819755e-01
3.78913939e-01 2.00601712e-01 -1.26556739e-01 8.41217399e-01
8.49010330e-03 9.92592722e-02 5.98534524e-01 -3.68302524e-01
-2.16372237e-02 6.55109644e-01 -3.83295327e-01 -1.05327439e+00
-9.99419689e-01 -8.19056094e-01 -1.44077456e+00 5.42255998e-01
5.40374458e-01 -3.16345721e-01 -1.22060072e+00 1.07537878e+00
2.18141764e-01 -5.65570556e-02 -7.49992281e-02 8.26225221e-01
1.12825477e+00 4.23819363e-01 7.42612109e-02 -2.32450590e-01
1.72717190e+00 -8.25920701e-01 -1.02179158e+00 -1.48201153e-01
5.71537912e-01 -8.48945200e-01 4.44756120e-01 5.17824352e-01
-1.08914602e+00 -4.87661719e-01 -1.24501109e+00 -1.63879603e-01
-5.49293637e-01 1.74678251e-01 1.16579175e+00 8.12160492e-01
-8.92999291e-01 7.42822230e-01 -1.00149608e+00 -2.92718738e-01
1.08390236e+00 5.07699728e-01 -1.35358036e-01 1.55234858e-01
-1.09351301e+00 7.48437166e-01 3.29214752e-01 6.25024915e-01
-1.84235290e-01 -4.34166849e-01 -9.55731511e-01 -1.11156181e-01
3.21885914e-01 -5.92701495e-01 1.38944113e+00 -9.54338014e-01
-1.08951986e+00 1.21172106e+00 3.15378159e-01 -4.27323818e-01
6.11548722e-01 6.11573756e-02 -4.57519233e-01 2.62566209e-01
-3.26506346e-01 7.04891562e-01 4.74711269e-01 -1.06545019e+00
-8.17939699e-01 -2.31334090e-01 -2.86150247e-01 -6.67730868e-02
-2.81155378e-01 1.96276456e-02 -5.37255883e-01 -1.07700992e+00
4.00509477e-01 -6.14180028e-01 -5.98640382e-01 5.45950592e-01
-4.49530840e-01 -1.92908078e-01 9.63222623e-01 -3.00446719e-01
1.23065937e+00 -1.88089061e+00 -2.58578092e-01 -2.44752527e-03
2.96063036e-01 4.85652298e-01 1.68474346e-01 -4.08083759e-02
-3.18096906e-01 1.60993204e-01 -7.49609411e-01 -1.44085765e-01
-5.74826002e-01 6.03763498e-02 1.91754162e-01 6.43552244e-01
3.60831261e-01 1.34687507e+00 -6.73967302e-01 -8.38907778e-01
3.35465908e-01 5.55299103e-01 -1.65890500e-01 -9.89549607e-03
-5.46520762e-02 3.24367404e-01 -4.55008835e-01 9.39433575e-01
8.97884846e-01 -3.42566252e-01 -5.93270883e-02 -4.63541538e-01
-2.14644611e-01 -4.44187760e-01 -9.11350965e-01 1.56026876e+00
-8.47748369e-02 8.74140382e-01 2.63113171e-01 -1.31685305e+00
9.51383412e-01 5.93172610e-01 7.57286727e-01 -7.99360693e-01
4.17566150e-01 3.86522621e-01 1.36150494e-01 -7.24909127e-01
2.93228656e-01 -5.77938914e-01 3.71278487e-02 8.61129314e-02
-7.00229183e-02 -1.38662785e-01 5.37042208e-02 -2.90053874e-01
4.83772695e-01 -2.27697060e-01 7.22774327e-01 -4.37391013e-01
5.59067726e-01 2.69666791e-01 4.98346359e-01 4.11696821e-01
-6.33857131e-01 9.40602958e-01 3.94033462e-01 -9.38218296e-01
-6.93189800e-01 -7.00445831e-01 -5.63317716e-01 6.04153156e-01
1.93674445e-01 3.05150568e-01 -8.21441829e-01 -6.57844245e-01
-2.39183441e-01 -6.53762445e-02 -1.04966450e+00 6.01606779e-02
-1.06879199e+00 -1.31327188e+00 6.14911556e-01 8.94104958e-01
7.10368931e-01 -1.55860400e+00 -1.07299364e+00 4.16014016e-01
9.03577730e-02 -9.11847413e-01 -3.58207762e-01 4.69837457e-01
-1.32478738e+00 -1.28798747e+00 -1.46203911e+00 -1.43492019e+00
7.60225952e-01 1.85361326e-01 1.08635223e+00 6.98712826e-01
-8.71163905e-01 -1.56765625e-01 -2.82920837e-01 -6.91374123e-01
-2.47354403e-01 6.48993701e-02 -5.97508192e-01 -3.36784929e-01
1.74296364e-01 -1.23448119e-01 -9.69387650e-01 6.06829189e-02
-1.21215451e+00 1.19655192e-01 7.47352064e-01 1.07878244e+00
5.36520183e-01 1.59303010e-01 5.23322642e-01 -1.16687727e+00
6.22014344e-01 -9.18771997e-02 -4.17403698e-01 3.37809712e-01
-5.19793928e-01 -1.09048717e-01 3.05324703e-01 -1.95947498e-01
-1.02862608e+00 1.01076171e-01 -5.99328279e-01 8.20362568e-02
-3.72169971e-01 6.79446101e-01 2.52504110e-01 -2.02761531e-01
5.17033100e-01 7.11494777e-03 4.06585522e-02 -3.45588744e-01
9.07411352e-02 3.92808676e-01 7.93724775e-01 -2.51939088e-01
3.02070737e-01 8.42203379e-01 8.87042806e-02 -8.92458022e-01
-8.29368711e-01 -5.50589979e-01 -6.54207230e-01 -2.46822521e-01
1.29432058e+00 -1.65059090e-01 -6.60641849e-01 7.33764589e-01
-1.27984869e+00 -3.59238356e-01 -9.65190083e-02 1.60702139e-01
-3.49235535e-01 4.88425016e-01 -9.89028275e-01 -6.62826300e-01
-8.07671666e-01 -1.29006135e+00 7.90943384e-01 7.19509065e-01
7.63142407e-02 -1.57754028e+00 -4.99221981e-02 -3.46994475e-02
5.62077045e-01 8.03142667e-01 1.02809775e+00 -2.06833407e-01
-2.21413612e-01 -2.73305237e-01 -4.28100914e-01 1.64609820e-01
3.35596830e-01 1.72418091e-04 -9.54780340e-01 -1.85854271e-01
7.25119933e-02 -1.11111574e-01 1.10850155e+00 1.01005065e+00
1.43794394e+00 3.12858611e-01 -5.65524697e-01 8.87393653e-01
1.55551684e+00 6.07166052e-01 7.21619606e-01 4.07011241e-01
4.52101916e-01 6.47320092e-01 4.32512850e-01 8.49881209e-03
-1.89460427e-01 3.12507808e-01 5.02114356e-01 -1.10847330e+00
-2.73794532e-01 3.31669003e-01 -4.50720519e-01 6.57633603e-01
-2.64133573e-01 -1.27341852e-01 -9.11029696e-01 4.02868032e-01
-1.60134339e+00 -4.85609263e-01 -4.77527082e-01 1.62933612e+00
7.51999915e-01 3.64550531e-01 -6.28303066e-02 2.72967815e-01
7.62629390e-01 -5.50664179e-02 -5.00995457e-01 -3.77159953e-01
-2.62088209e-01 7.50661790e-01 3.79619420e-01 2.67581433e-01
-1.62184811e+00 8.86554122e-01 6.50099945e+00 7.38712013e-01
-1.34151661e+00 -2.41656184e-01 1.06219363e+00 6.60581768e-01
1.64462194e-01 -5.19235194e-01 -3.65319908e-01 -8.04405510e-02
1.63359076e-01 2.66427755e-01 -4.05714244e-01 8.26093376e-01
-7.28875026e-02 -2.64600575e-01 -7.70971239e-01 1.01674712e+00
-6.19164035e-02 -1.73941779e+00 -1.62052184e-01 -9.66800228e-02
6.30189478e-01 -1.63041160e-01 2.16537327e-01 1.90124325e-02
-3.14227521e-01 -1.30278790e+00 4.62583899e-01 1.08209185e-01
1.07325757e+00 -7.98393846e-01 1.11705315e+00 4.35691364e-02
-1.40030396e+00 2.10857838e-01 -4.80951667e-01 6.90963194e-02
4.05426443e-01 6.98706746e-01 -4.04088080e-01 5.17748654e-01
6.73713803e-01 6.00018322e-01 -4.08200085e-01 1.43191493e+00
7.49953091e-02 3.85081649e-01 -3.64118330e-02 -1.21416792e-01
5.33202648e-01 -1.08252071e-01 -3.23580466e-02 1.60211480e+00
1.38695776e-01 3.65459144e-01 -5.19025698e-02 7.79566526e-01
-2.22725440e-02 2.09455594e-01 -2.65676975e-01 1.38581246e-01
-2.11890459e-01 1.47009850e+00 -1.62570000e+00 -3.33014220e-01
-5.35644948e-01 8.62515092e-01 -3.15371938e-02 2.30019495e-01
-6.41003847e-01 -5.95261455e-01 4.79231447e-01 -1.28867686e-01
-2.15071980e-02 -2.10007906e-01 -7.76959598e-01 -7.34530807e-01
-2.58543640e-01 -6.04480445e-01 6.93071365e-01 -2.81693250e-01
-1.06665850e+00 8.74888718e-01 -1.89443052e-01 -1.16108072e+00
3.60734701e-01 -1.02450264e+00 -8.95326972e-01 5.03001451e-01
-1.87522364e+00 -1.00706017e+00 -5.45006812e-01 4.82105076e-01
6.66083038e-01 1.40532568e-01 8.85119498e-01 2.81767935e-01
-5.16601443e-01 5.05097270e-01 -3.99241447e-02 7.82754242e-01
3.97417814e-01 -1.20815670e+00 5.97136736e-01 5.11632144e-01
-8.38438421e-03 4.53673035e-01 4.54359412e-01 -5.80421567e-01
-7.38682747e-01 -8.51447523e-01 5.96099555e-01 2.77345538e-01
2.50063926e-01 -2.22094998e-01 -1.08849800e+00 4.14928973e-01
3.95596862e-01 3.53280395e-01 5.67982554e-01 -3.68454307e-01
2.86897928e-01 2.49747947e-01 -1.14485240e+00 5.55797935e-01
6.83536351e-01 8.96143988e-02 -3.19753468e-01 2.19805509e-01
3.73635918e-01 -7.54192352e-01 -6.07156992e-01 6.58980191e-01
5.91669798e-01 -1.26580322e+00 1.02227867e+00 -4.73655134e-01
2.84508616e-01 7.99192861e-03 5.69413841e-01 -1.03866506e+00
-2.74508566e-01 -2.57441550e-01 2.38183632e-01 5.22637427e-01
2.25092366e-01 -4.93007302e-01 1.31909561e+00 2.48350859e-01
-5.21988750e-01 -1.38654912e+00 -8.08125734e-01 -4.10925478e-01
6.62715316e-01 -4.46319431e-01 1.78979933e-01 9.27013338e-01
2.18518032e-03 -1.26415923e-01 1.68206375e-02 -1.70646101e-01
5.66612840e-01 2.00057015e-01 1.04968317e-01 -1.34274220e+00
2.16044173e-01 -9.60550368e-01 -4.34319049e-01 -1.05735624e+00
-3.61246079e-01 -7.70021141e-01 -1.09328628e-02 -1.85005403e+00
1.29676968e-01 -4.50329423e-01 -2.31794268e-01 4.16628540e-01
-2.66764462e-01 7.07831562e-01 -7.57059604e-02 -1.07088469e-01
-1.85438290e-01 1.99589580e-01 1.83636773e+00 -2.09101260e-01
-7.30207786e-02 3.14800411e-01 -6.26270771e-01 9.16723311e-01
1.21709752e+00 -2.73927838e-01 -7.71979392e-02 -2.35716730e-01
2.08505496e-01 1.11787975e-01 1.75170884e-01 -1.02804112e+00
4.44981724e-01 -2.19684038e-02 7.50379562e-01 -8.91323328e-01
3.14165726e-02 -7.90383041e-01 -3.34327698e-01 9.60470259e-01
-9.06639546e-02 1.30118787e-01 4.82000411e-01 2.72888839e-01
-4.88089770e-01 -3.04846972e-01 1.27499115e+00 -4.74074453e-01
-8.71363282e-01 5.39476871e-01 -4.85555798e-01 -1.82424843e-01
1.16070712e+00 -5.26485920e-01 1.40502425e-02 -2.12984741e-01
-9.61461425e-01 3.03736739e-02 1.03201168e-02 2.24519908e-01
9.35791016e-01 -1.16230559e+00 -5.73178709e-01 2.84406126e-01
-2.40061269e-03 4.23331887e-01 3.76281053e-01 8.14315915e-01
-1.11325932e+00 4.61531848e-01 -2.44844466e-01 -5.77800512e-01
-1.28798449e+00 6.49239182e-01 7.98724532e-01 -2.85619110e-01
-1.01984990e+00 1.09135246e+00 4.49485362e-01 5.21020405e-02
3.43430072e-01 -8.67934167e-01 -5.03022075e-01 -1.39831334e-01
5.12083471e-01 1.20289877e-01 4.50346380e-01 -3.63833368e-01
-3.76358181e-01 8.46274078e-01 -8.62349663e-03 3.75501841e-01
1.31026018e+00 1.61661491e-01 -1.21710248e-01 -1.68053564e-02
1.26721859e+00 -4.29787248e-01 -1.07001412e+00 5.38590103e-02
3.08053732e-01 9.52082351e-02 1.87669501e-01 -7.12129295e-01
-1.64434147e+00 1.30711484e+00 8.67710292e-01 2.70086914e-01
1.27503073e+00 -7.39681423e-02 1.16337085e+00 -1.00920619e-02
1.75160229e-01 -1.10081184e+00 3.27193916e-01 2.71359473e-01
8.03419411e-01 -1.45971942e+00 4.08477224e-02 -7.55806386e-01
-3.53933871e-01 1.69784486e+00 3.79664630e-01 -1.93434149e-01
1.07579553e+00 6.20940566e-01 6.11457646e-01 -6.43364549e-01
5.87249734e-02 -3.65146130e-01 4.29345548e-01 6.81959450e-01
9.86341774e-01 1.23420231e-01 -4.05265749e-01 4.39498812e-01
3.00737526e-02 -7.64805004e-02 3.64218950e-01 1.05819213e+00
-6.00269377e-01 -1.06166840e+00 -4.95351523e-01 4.13033724e-01
-8.17491829e-01 -5.94832748e-02 -1.48709789e-01 1.06692779e+00
3.52944642e-01 6.62656605e-01 1.92097306e-01 1.64331228e-01
2.77400136e-01 -1.91133082e-01 4.65745360e-01 -5.03189921e-01
-6.38628840e-01 3.35812956e-01 -3.44563782e-01 -1.64343029e-01
-4.11348432e-01 -3.82827669e-01 -1.66732347e+00 -4.79512056e-03
-5.57357073e-01 3.85431573e-02 6.81231081e-01 6.82715654e-01
-2.86845297e-01 8.44832122e-01 -8.73690820e-04 -8.34319770e-01
-1.82730824e-01 -7.34770954e-01 -6.63781285e-01 2.24589005e-01
4.87324357e-01 -5.87163448e-01 9.29838941e-02 2.13353932e-01] | [14.59304141998291, -2.572633743286133] |
5539a168-c7df-4c35-b952-fdf5d794fadd | physics-informed-deep-neural-networks-for | null | null | https://ieeexplore.ieee.org/document/9158400 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9158400 | Physics-Informed Deep Neural Networks for Transient Electromagnetic Analysis | In this paper, we propose a deep neural network based model to predict the time evolution of field values in transient electrodynamics. The key component of our model is a recurrent neural network, which learns representations of long-term spatial-temporal dependencies in the sequence of its input data. We develop an encoder-recurrent-decoder architecture, which is trained with finite difference time domain simulations of plane wave scattering from distributed, perfect electric conducting objects. We demonstrate that, the trained network can emulate a transient electrodynamics problem with more than 17 times speed-up in simulation time compared to traditional finite difference time domain solvers. | ['Zhen Peng', 'Christos Christodoulou', 'Shu Wang', 'Oameed Noakoasteen'] | 2020-08-04 | null | null | null | ieee-open-journal-of-antennas-and-propagation | ['physics-informed-machine-learning'] | ['graphs'] | [-1.62456468e-01 -3.25923204e-01 7.49365509e-01 -1.10210799e-01
-7.19406247e-01 -1.82487607e-01 2.52855122e-01 -6.21945262e-01
-1.73576280e-01 8.32695127e-01 9.69931930e-02 -5.68068862e-01
-4.35477406e-01 -8.08740437e-01 -8.22353244e-01 -7.51047671e-01
-5.79881132e-01 4.50534761e-01 -2.82992005e-01 -2.86209792e-01
8.88090581e-02 7.06430614e-01 -9.91847277e-01 4.11803693e-01
3.36247653e-01 1.11477411e+00 -5.86018488e-02 9.54077423e-01
3.75854522e-01 1.41762817e+00 -5.58314264e-01 5.04831374e-01
3.39002572e-02 -5.76326072e-01 -9.42982972e-01 -5.14453113e-01
-4.34656203e-01 -4.72292542e-01 -1.18912411e+00 7.66524255e-01
7.55929887e-01 3.32925498e-01 8.21795821e-01 -6.98510528e-01
-6.69573605e-01 5.45947611e-01 -2.29847312e-01 7.07407236e-01
-6.46159723e-02 2.18599036e-01 5.41850924e-01 -6.32381558e-01
8.02796960e-01 8.75474751e-01 9.47441339e-01 4.07022476e-01
-9.57242966e-01 -5.66988587e-01 -2.37526581e-01 1.63400769e-01
-1.19961715e+00 -2.22229138e-01 9.22806025e-01 -3.27558547e-01
1.63801289e+00 -2.84894437e-01 7.75335133e-01 1.09732699e+00
9.55039203e-01 2.53416479e-01 6.68607056e-01 -2.23319396e-01
2.01458305e-01 -5.32500684e-01 2.42512971e-01 4.18415010e-01
-2.61637479e-01 6.24535739e-01 -3.32446754e-01 -6.21148765e-01
9.76456761e-01 -2.26004422e-01 -4.50761676e-01 5.92092723e-02
-6.78682506e-01 5.42657793e-01 4.41632867e-01 3.98191512e-01
-7.77028143e-01 6.39986873e-01 6.57441199e-01 6.50675058e-01
7.84428477e-01 6.65541410e-01 -4.92311388e-01 -3.45345289e-01
-6.64468765e-01 6.55874550e-01 9.41164076e-01 3.62310410e-01
4.52204980e-02 6.95042789e-01 4.95772809e-02 4.61639404e-01
-1.59405787e-02 7.32926130e-01 3.97606939e-01 -8.85230362e-01
-3.07962745e-02 -1.62683561e-01 3.74429464e-01 -7.99355090e-01
-6.34319365e-01 -5.92431247e-01 -1.06018305e+00 3.43815088e-01
-3.45105022e-01 -1.01178861e+00 -8.62632632e-01 1.61263585e+00
-8.46459903e-03 7.79236555e-01 4.35533822e-01 1.22503996e+00
7.02053070e-01 1.39776123e+00 -1.37949362e-01 -3.22319537e-01
5.30127048e-01 -5.35798430e-01 -7.60061502e-01 9.95810255e-02
7.40988672e-01 -3.58177930e-01 -1.44423088e-02 1.26076415e-01
-1.39843476e+00 -1.32325977e-01 -9.85016584e-01 1.50306344e-01
-1.40914740e-02 -1.27797827e-01 5.74846327e-01 -2.89821684e-01
-1.24435842e+00 1.24902487e+00 -1.19225430e+00 3.97972047e-01
2.60377806e-02 3.69802147e-01 1.38399944e-01 4.44484204e-01
-1.49303067e+00 7.41263866e-01 -3.22873145e-01 3.80314857e-01
-1.04184306e+00 -1.27705336e+00 -6.70564830e-01 1.80406228e-01
-3.96954179e-01 -7.52712011e-01 1.69253516e+00 -6.96663320e-01
-1.81542611e+00 5.99787533e-01 -2.48910800e-01 -7.51394808e-01
2.05238923e-01 -8.09568837e-02 -8.16990614e-01 1.14894740e-01
-7.20685571e-02 -1.38588160e-01 5.31750202e-01 -8.49807441e-01
4.84112427e-02 1.83415264e-01 -2.90536433e-01 -7.98084121e-03
2.58235306e-01 8.79586563e-02 2.16682553e-01 -4.55257952e-01
1.05161533e-01 -9.04925287e-01 -7.23322928e-01 -2.01741174e-01
-1.28475711e-01 1.53516933e-01 8.94531786e-01 -6.86088920e-01
8.98582399e-01 -1.93612993e+00 2.62038827e-01 1.51098505e-01
3.17024291e-01 3.53814006e-01 -4.49214280e-02 6.54681444e-01
-5.48997223e-01 -2.69126683e-01 -2.86521554e-01 2.23945245e-01
-1.44808650e-01 -3.80837649e-01 -1.01053965e+00 6.69959545e-01
2.22749293e-01 9.66398001e-01 -6.34260476e-01 5.24531305e-01
-2.68594492e-02 8.05435896e-01 -6.55766368e-01 4.43703800e-01
-5.81940949e-01 8.20170820e-01 -7.03579545e-01 -4.65207435e-02
9.16635096e-01 -8.46961260e-01 3.72195840e-01 2.29935646e-02
-5.58250844e-01 7.13430822e-01 -7.49396563e-01 1.54438663e+00
-7.05994368e-01 1.05716681e+00 2.03406140e-01 -1.47717488e+00
7.35606790e-01 7.99357533e-01 8.51181746e-01 -1.42616379e+00
5.37876368e-01 1.32330909e-01 1.96119286e-02 -7.25416243e-01
2.21396267e-01 -4.82922196e-01 -1.83518857e-01 1.03667414e+00
-4.23903838e-02 -1.62203729e-01 -2.91327208e-01 -4.91458140e-02
1.61122799e+00 -5.51896319e-02 -1.66475818e-01 -2.50702500e-01
1.63008347e-01 1.65783033e-01 4.38380450e-01 6.91688180e-01
2.88583338e-01 4.29076046e-01 6.60218596e-01 -9.40509200e-01
-1.24642479e+00 -9.22097385e-01 -3.89289886e-01 5.69898784e-01
2.18198486e-02 -8.88047516e-02 -4.01835829e-01 4.05574650e-01
-1.79342642e-01 4.61684883e-01 -5.27700245e-01 -2.94710100e-01
-1.00933266e+00 -8.92211080e-01 5.42310536e-01 5.31804860e-01
3.78958076e-01 -1.48730803e+00 -8.49034846e-01 6.85837388e-01
6.19415790e-02 -1.11410189e+00 3.09886277e-01 4.63306218e-01
-7.86907196e-01 -7.42807209e-01 -7.50664353e-01 -8.25135887e-01
2.24179879e-01 -1.50431484e-01 1.30322134e+00 -9.35592651e-02
-8.20866227e-01 2.38784090e-01 -8.41693580e-02 -1.42439410e-01
-2.71310598e-01 -2.65547365e-01 7.15757012e-02 -4.75901455e-01
-6.45202026e-02 -1.23044467e+00 -5.02781451e-01 -2.28752017e-01
-6.84218347e-01 1.13269195e-01 2.37714887e-01 8.02507758e-01
4.70055789e-01 -3.02678198e-02 5.42884469e-01 -7.35606432e-01
9.05537903e-01 -9.25578356e-01 -8.95562470e-01 -1.21376673e-02
2.28825390e-01 2.16974154e-01 1.09123528e+00 -4.04390752e-01
-1.14229679e+00 -5.60976207e-01 -4.45256442e-01 -6.75146401e-01
2.03614548e-01 8.18715334e-01 6.95060492e-01 -4.15046811e-01
6.06706560e-01 3.94167691e-01 -3.38153213e-01 -4.51794684e-01
-8.43452811e-02 4.06807363e-01 4.93318379e-01 -5.32687128e-01
3.94390315e-01 4.45349753e-01 1.65409043e-01 -8.27840090e-01
-5.33509672e-01 8.64950269e-02 -2.15402782e-01 -6.50773197e-02
4.58488315e-01 -9.22175527e-01 -1.07399058e+00 7.52690136e-01
-1.63780415e+00 -6.99287295e-01 4.26284596e-02 6.64176404e-01
-5.72224498e-01 -1.61562920e-01 -1.03940916e+00 -7.04866827e-01
-5.56082129e-01 -5.15588522e-01 6.88403487e-01 1.42203078e-01
-6.10624589e-02 -1.25156605e+00 6.47990108e-01 -7.67670929e-01
7.71599054e-01 5.20761311e-01 1.12165952e+00 -2.02880912e-02
-5.77751935e-01 -1.11654744e-01 -2.60331452e-01 2.22317483e-02
-3.13502520e-01 -2.80868292e-01 -6.76428616e-01 -3.75785142e-01
6.85674012e-01 -4.11234200e-01 8.63797605e-01 1.12604868e+00
1.51998734e+00 -1.85236096e-01 -5.13588727e-01 1.31091535e+00
1.55157149e+00 4.59334403e-01 3.62126708e-01 -2.25866944e-01
3.48630339e-01 -7.62879923e-02 -6.08351789e-02 8.84165108e-01
-6.86065182e-02 1.69117808e-01 -2.98691802e-02 -1.88292220e-01
1.92334265e-01 2.01689243e-01 2.16542885e-01 9.48201716e-01
6.54549748e-02 -5.24341464e-01 -1.16288137e+00 5.87870657e-01
-1.59780800e+00 -1.13694465e+00 -7.05821589e-02 1.45390809e+00
4.30127740e-01 -2.08071679e-01 -5.67269087e-01 -1.58277974e-01
3.51208240e-01 3.30434889e-01 -1.09317708e+00 -7.82474399e-01
5.36884926e-02 7.66163647e-01 3.41062516e-01 4.40242141e-01
-8.44075978e-01 7.23182082e-01 8.17245960e+00 1.03714362e-01
-1.92523098e+00 -1.10622626e-02 3.94551069e-01 -4.76080477e-01
-3.31638128e-01 -3.19365472e-01 -1.48674607e-01 3.02034765e-01
1.42078447e+00 -7.44669080e-01 5.53554058e-01 5.75733721e-01
3.18139464e-01 4.39730138e-01 -8.90853524e-01 9.46549356e-01
-2.16059431e-01 -1.90987813e+00 -1.69397667e-01 -3.36903602e-01
8.40839624e-01 6.18360341e-01 2.86139786e-01 2.52470732e-01
5.43823481e-01 -1.43023622e+00 2.44461983e-01 7.47416437e-01
8.99390936e-01 -7.73084700e-01 4.07313615e-01 4.91960764e-01
-9.13926423e-01 1.34049561e-02 -3.71792555e-01 -5.70899069e-01
7.57454395e-01 7.46779799e-01 -4.76914376e-01 3.40815485e-01
8.34291160e-01 9.49020147e-01 5.86120963e-01 7.87279308e-01
4.00874227e-01 8.65866244e-01 -3.63055259e-01 -1.01320826e-01
6.32321596e-01 -3.44222605e-01 6.59298956e-01 9.99162674e-01
6.19020164e-01 1.01680064e+00 -2.44389072e-01 1.32839727e+00
-4.51211482e-02 -6.38473511e-01 -8.34699333e-01 -3.36408168e-01
6.92768320e-02 7.81741858e-01 -2.11494848e-01 -9.97491032e-02
-2.20863342e-01 6.34764254e-01 3.68882954e-01 8.73715937e-01
-1.15055883e+00 -6.02632225e-01 9.13317442e-01 5.10238111e-03
7.12085366e-01 -4.75765795e-01 -1.67963490e-01 -9.23895061e-01
3.42410840e-02 -4.19137388e-01 1.84787512e-02 -1.11316657e+00
-1.30776381e+00 1.05095923e+00 -2.43596166e-01 -1.11182892e+00
-6.39763594e-01 -8.13656867e-01 -1.01598120e+00 1.28547299e+00
-1.33219838e+00 -6.29588485e-01 1.92665309e-01 4.42559689e-01
-1.06467925e-01 -1.82474315e-01 9.16465998e-01 2.68476605e-01
-3.23210001e-01 -7.28459144e-03 6.62509501e-01 4.09306973e-01
-1.34101212e-01 -5.30834794e-01 9.42587137e-01 5.23332238e-01
-2.93339133e-01 5.30235529e-01 8.96313548e-01 -5.81302285e-01
-1.53367102e+00 -1.01024330e+00 7.26689756e-01 1.73884794e-01
7.52123415e-01 -1.73889473e-01 -1.19711423e+00 7.57521212e-01
3.88614833e-01 3.89136404e-01 3.87422115e-01 -1.05463475e-01
6.87897578e-02 3.60718638e-01 -8.34244251e-01 3.31793547e-01
9.62691247e-01 -8.02025676e-01 -3.45091283e-01 4.03254598e-01
6.82275951e-01 -8.66494060e-01 -5.22811651e-01 5.65598667e-01
3.01540583e-01 -8.03673565e-01 1.03255033e+00 -9.15668130e-01
7.90962577e-01 2.95276701e-01 2.91660160e-01 -1.49960983e+00
-5.28912306e-01 -9.47093964e-01 -1.61801055e-01 1.60021350e-01
3.51428747e-01 -5.21399319e-01 6.40421987e-01 3.26187700e-01
-2.55785912e-01 -9.10706282e-01 -1.01154077e+00 -3.40650916e-01
7.18295693e-01 -5.04664779e-01 4.19317991e-01 8.25670540e-01
1.36678234e-01 7.65258893e-02 -5.62454224e-01 5.40610552e-01
3.29334676e-01 6.39220297e-01 -2.46711969e-02 -9.24110472e-01
-3.34035754e-01 -8.91762897e-02 -1.76355839e-01 -1.35845852e+00
8.07428479e-01 -8.65394592e-01 1.08685985e-01 -1.32389402e+00
8.97549093e-02 -2.69290000e-01 -4.31430578e-01 -4.21802625e-02
6.34428442e-01 2.62781810e-02 -4.41999674e-01 -2.01338589e-01
-5.12281477e-01 8.60668063e-01 1.10075402e+00 2.00774565e-01
-1.26041353e-01 -1.13371558e-01 -1.24308340e-01 6.05397463e-01
6.75431132e-01 -8.31541419e-01 -1.89491689e-01 -1.06911576e+00
4.70610708e-01 9.70118940e-01 7.36882687e-01 -1.30433035e+00
6.30931020e-01 -4.32111993e-02 4.16042268e-01 -5.17878950e-01
4.98086244e-01 -1.32782415e-01 1.14448997e-03 4.25110102e-01
-4.60273981e-01 1.50005400e-01 6.23153627e-01 2.88683832e-01
-3.57389957e-01 2.29449525e-01 7.02688038e-01 -2.75059968e-01
-5.92124224e-01 6.01719618e-01 -1.07541370e+00 2.53756195e-01
5.47412157e-01 4.51178700e-01 2.75118519e-02 -5.70736289e-01
-6.89699709e-01 -2.70983874e-04 -4.06980477e-02 7.30509609e-02
6.14097714e-01 -1.20764244e+00 -8.02961409e-01 3.75215620e-01
-5.75065553e-01 -2.76800603e-01 8.56905401e-01 4.42399740e-01
-8.36236417e-01 5.41952372e-01 -4.68253583e-01 -4.25422549e-01
-5.95987976e-01 8.60017836e-02 1.12202621e+00 -3.10993761e-01
-1.30159914e+00 9.83781397e-01 7.98732266e-02 -4.62334633e-01
-3.14403683e-01 -3.19534481e-01 -8.68488699e-02 -7.96955824e-01
4.05932397e-01 1.44806877e-01 -1.59082208e-02 -3.29537481e-01
-2.44039610e-01 6.49573028e-01 1.56734928e-01 -4.90756422e-01
1.74756026e+00 3.40390652e-01 -3.60984415e-01 3.16068619e-01
1.60078061e+00 -6.84212267e-01 -1.20021129e+00 -4.12634224e-01
-4.75144595e-01 1.99614525e-01 6.35941803e-01 -7.68866003e-01
-1.06374395e+00 1.05104995e+00 3.72832149e-01 -1.43770250e-02
1.03451121e+00 -2.04560623e-01 1.31663883e+00 8.16293716e-01
2.83893287e-01 -8.46220195e-01 -2.62044877e-01 1.31570482e+00
8.12707961e-01 -7.30724871e-01 -2.97564834e-01 1.92117482e-01
-3.45761746e-01 1.21124852e+00 1.88444689e-01 -8.59357178e-01
1.30720341e+00 1.20910740e+00 2.03218341e-01 -6.47545874e-01
-1.51147592e+00 5.05595028e-01 -5.99322841e-02 3.08086276e-01
6.59152210e-01 -1.99701920e-01 1.19192064e-01 8.81256580e-01
-2.83510000e-01 3.39835882e-01 3.96865904e-01 8.43964279e-01
-1.83226272e-01 -5.55275023e-01 5.32304011e-02 2.43328124e-01
-5.60596466e-01 2.11676359e-02 -3.69106941e-02 3.82284403e-01
-4.62627083e-01 5.46465576e-01 6.30521119e-01 -1.36629745e-01
1.42935961e-01 -2.38979496e-02 4.10232425e-01 -2.70239919e-01
-5.20420790e-01 -3.48285139e-01 -1.44170895e-01 -5.83612025e-01
7.49828247e-03 -2.65786350e-01 -1.71975374e+00 -5.95426738e-01
2.92634964e-01 5.96011877e-01 3.66292119e-01 1.02265215e+00
7.04339921e-01 1.12709975e+00 8.42992127e-01 -1.11456394e+00
-5.36728919e-01 -7.52534211e-01 -7.94022143e-01 9.10952762e-02
8.43761146e-01 -3.48679572e-01 -4.71137762e-01 -3.23617578e-01] | [6.781737804412842, 3.430706739425659] |
aa00f0d2-e512-4ba5-8019-5b607e7d1624 | efficiency-aware-answering-of-compositional | null | null | https://aclanthology.org/I17-2038 | https://aclanthology.org/I17-2038.pdf | Efficiency-aware Answering of Compositional Questions using Answer Type Prediction | This paper investigates the problem of answering compositional factoid questions over knowledge bases (KB) under efficiency constraints. The method, called TIPI, (i) decomposes compositional questions, (ii) predicts answer types for individual sub-questions, (iii) reasons over the compatibility of joint types, and finally, (iv) formulates compositional SPARQL queries respecting type constraints. TIPI{'}s answer type predictor is trained using distant supervision, and exploits lexical, syntactic and embedding-based features to compute context- and hierarchy-aware candidate answer types for an input question. Experiments on a recent benchmark show that TIPI results in state-of-the-art performance under the real-world assumption that only a single SPARQL query can be executed over the KB, and substantial reduction in the number of queries in the more general case. | ['Abdalghani Abujabal', 'Rishiraj Saha Roy', 'Gerhard Weikum', 'David Ziegler'] | 2017-11-01 | efficiency-aware-answering-of-compositional-1 | https://aclanthology.org/I17-2038 | https://aclanthology.org/I17-2038.pdf | ijcnlp-2017-11 | ['type-prediction'] | ['computer-code'] | [-1.68173343e-01 4.43769157e-01 -5.07927895e-01 -4.84316885e-01
-1.16070783e+00 -8.08201909e-01 4.33144182e-01 4.43088889e-01
-3.67481261e-01 8.57504725e-01 4.14257586e-01 -6.50854230e-01
-3.37338030e-01 -1.25360763e+00 -7.75260150e-01 1.52180232e-02
2.11487874e-01 1.09540594e+00 9.56449270e-01 -4.19620156e-01
5.79367578e-02 1.53391466e-01 -1.75264740e+00 1.00957048e+00
9.02883589e-01 1.49733090e+00 -3.53201628e-02 7.92722464e-01
-7.74014473e-01 1.45522237e+00 -4.05351281e-01 -1.00932884e+00
-2.90347580e-02 -1.21323727e-01 -1.66361320e+00 -4.22849447e-01
9.40024793e-01 -6.80055022e-02 -3.06893617e-01 8.74354362e-01
1.23463243e-01 -7.91305527e-02 3.65053147e-01 -1.06561756e+00
-5.16486287e-01 5.04810154e-01 4.73875016e-01 2.32823566e-01
9.05295432e-01 -1.96763217e-01 2.05459094e+00 -8.80673587e-01
9.17018116e-01 1.28124046e+00 5.05966663e-01 5.26831508e-01
-1.31528044e+00 2.41063442e-02 -1.31261870e-01 5.65390766e-01
-1.10253501e+00 -1.79224208e-01 2.58056581e-01 -1.73351616e-01
1.40975916e+00 8.21973205e-01 4.43255126e-01 3.67467314e-01
1.42771512e-01 7.52717674e-01 1.00156033e+00 -4.85828519e-01
3.71936858e-01 4.73465711e-01 6.73739970e-01 1.08252525e+00
1.55734345e-01 -3.09819996e-01 -6.11696541e-01 -8.06992650e-01
-1.21345453e-01 -4.80130255e-01 7.18959793e-02 -3.03933859e-01
-1.01826429e+00 9.66071308e-01 3.73847783e-01 2.27551654e-01
-2.96169102e-01 -2.91106775e-02 3.18131477e-01 7.37757146e-01
6.17292076e-02 8.02703917e-01 -9.75213349e-01 2.89918512e-01
-6.49782062e-01 8.32736790e-01 1.57541549e+00 1.27224815e+00
1.04972935e+00 -5.70373952e-01 -3.76335025e-01 6.70879602e-01
3.88676412e-02 3.97753119e-01 3.01069736e-01 -1.17087698e+00
6.25301123e-01 1.28504789e+00 2.89740086e-01 -4.24492508e-01
-3.64727437e-01 -4.66013178e-02 8.55220556e-02 -5.01572371e-01
5.16428053e-01 5.09462357e-02 -2.19612345e-01 1.57417548e+00
6.78502440e-01 -5.63358068e-01 3.82477731e-01 7.13470578e-01
1.09803724e+00 5.63689113e-01 -1.05803944e-01 1.38969198e-02
1.91676319e+00 -9.46890891e-01 -4.49477524e-01 -3.44690293e-01
7.37801194e-01 -7.24701703e-01 1.20952570e+00 -3.73967141e-02
-1.24251068e+00 -3.02560836e-01 -7.27576613e-01 -5.59725225e-01
-6.27025843e-01 -3.78301181e-02 6.97421908e-01 5.12734771e-01
-1.04225707e+00 -1.25465617e-01 -2.38672838e-01 -3.19762170e-01
-1.34215578e-01 2.83423901e-01 -1.15825556e-01 -2.92010695e-01
-1.38882995e+00 9.17156816e-01 3.13380420e-01 -4.32425529e-01
-6.24951363e-01 -1.15600705e+00 -6.47336125e-01 2.67217159e-01
8.66025686e-01 -1.15363252e+00 1.32077932e+00 -5.25756896e-01
-1.21045423e+00 8.18981946e-01 -6.65854156e-01 -3.54330271e-01
-3.86808738e-02 -8.43407139e-02 -4.89692032e-01 2.04036608e-01
3.17382276e-01 5.11835515e-01 3.36759567e-01 -9.54294801e-01
-1.09040916e+00 -5.62609851e-01 7.25044608e-01 1.45545080e-01
-2.51481354e-01 2.93203712e-01 -5.45622766e-01 -1.47004843e-01
9.60242599e-02 -7.03909516e-01 2.32115611e-01 -1.69389725e-01
-4.26229954e-01 -9.33641493e-01 4.88708317e-01 -8.10443819e-01
1.36933804e+00 -1.71764553e+00 2.28861183e-01 1.91067517e-01
9.39079970e-02 -1.64197057e-01 3.47147160e-03 5.79474270e-01
2.77776599e-01 -7.31916502e-02 -8.12169239e-02 3.59704614e-01
5.22714734e-01 5.54407001e-01 -7.20647812e-01 -2.70497620e-01
6.74533620e-02 1.03609073e+00 -5.85267186e-01 -7.34024942e-01
-3.69267166e-01 -3.94625276e-01 -1.20873308e+00 6.08429372e-01
-1.30785799e+00 -3.93147945e-01 -3.61340970e-01 7.39988208e-01
8.02331939e-02 -3.02750081e-01 5.47185123e-01 -4.07381773e-01
1.09542891e-01 1.00693285e+00 -1.18969083e+00 1.31705880e+00
-6.62437201e-01 8.53722468e-02 -9.85532347e-03 -7.64053762e-01
6.02225661e-01 3.75502020e-01 6.92803115e-02 -7.81326115e-01
-5.32705188e-01 6.07234240e-01 -3.94618899e-01 -9.00411725e-01
5.39177001e-01 -8.48795623e-02 -4.67347533e-01 4.50918466e-01
2.03945383e-01 -2.55940646e-01 5.03742635e-01 3.40266705e-01
1.31010282e+00 -7.35191703e-02 2.02754349e-01 -5.54625332e-01
9.44268405e-01 3.86611193e-01 4.69746977e-01 7.71037877e-01
4.03434932e-01 -1.11410871e-01 7.69206226e-01 -6.87861621e-01
-8.24372411e-01 -1.23217654e+00 1.20722447e-02 1.62620616e+00
2.05773294e-01 -4.75697011e-01 -4.45154846e-01 -1.00273609e+00
3.61852646e-01 1.20421934e+00 -3.24502200e-01 1.76557213e-01
-8.10913265e-01 -3.94101709e-01 5.61658263e-01 3.70280147e-01
2.80929923e-01 -1.08878112e+00 -5.44547141e-01 1.69067279e-01
-8.21035087e-01 -1.10136485e+00 -2.44679168e-01 2.03854233e-01
-7.42966175e-01 -1.38221121e+00 2.40132660e-01 -9.22306418e-01
2.88774967e-01 -2.41269931e-01 1.64631546e+00 2.75060892e-01
-1.40534252e-01 5.35098255e-01 -1.42114416e-01 1.10050021e-02
-2.03897700e-01 3.36305231e-01 -5.34282327e-01 -1.28701955e-01
5.79232216e-01 -3.39656807e-02 -3.21120143e-01 2.77860940e-01
-8.87696564e-01 -5.50781608e-01 2.70095795e-01 9.24337864e-01
6.73874855e-01 3.25518101e-02 6.35220885e-01 -1.21634388e+00
5.41294456e-01 -5.08458912e-01 -6.95862234e-01 1.06864750e+00
-7.32196629e-01 4.76824999e-01 6.17575407e-01 7.49464780e-02
-1.51356435e+00 -4.69336569e-01 -2.60502487e-01 2.34039038e-01
-4.45366129e-02 7.87647426e-01 -4.64591444e-01 1.40256673e-01
8.14595401e-01 2.00751245e-01 -3.04062873e-01 -3.43408227e-01
7.09339678e-01 5.32062352e-01 6.74055994e-01 -1.24465263e+00
6.87575340e-01 3.91378611e-01 -1.56121045e-01 -8.14856768e-01
-1.27401149e+00 -7.54952192e-01 -4.36545789e-01 1.29515305e-01
7.74974227e-01 -6.64903402e-01 -7.78452754e-01 -3.75332505e-01
-1.12850237e+00 -2.52945006e-01 -4.69452918e-01 -9.57959983e-03
-6.17326856e-01 3.01865906e-01 -6.31868660e-01 -6.28337681e-01
-3.48568320e-01 -7.57270575e-01 1.07514238e+00 -1.16104722e-01
-5.81318676e-01 -1.04270613e+00 1.10992730e-01 1.06343722e+00
2.43470132e-01 -5.41941643e-01 2.05029202e+00 -1.00171947e+00
-1.15958548e+00 -3.57077301e-01 -3.20744485e-01 8.28626230e-02
-2.50890255e-01 -4.56197083e-01 -6.25535727e-01 1.07860275e-01
-9.02192816e-02 -6.89933777e-01 5.15718162e-01 -3.19151461e-01
8.70627642e-01 -1.00480950e+00 1.84552953e-01 1.97808325e-01
1.69340742e+00 -1.88559115e-01 4.85437691e-01 2.45209932e-01
2.81573683e-01 9.49553847e-01 5.36413252e-01 9.26614255e-02
9.79621530e-01 7.60893345e-01 2.78399557e-01 4.68813866e-01
-1.76144838e-01 -4.47824270e-01 7.67131373e-02 7.11801469e-01
4.65262204e-01 2.83490717e-02 -6.77400112e-01 8.34232688e-01
-1.74155056e+00 -1.04050875e+00 -1.62844062e-01 1.99166751e+00
1.28407109e+00 -1.89938784e-01 2.91074216e-01 -1.20239548e-01
1.55762315e-01 5.63616678e-03 -4.69279140e-01 -4.32171494e-01
-1.24986462e-01 6.01605177e-01 3.57008368e-01 7.83152997e-01
-6.07115388e-01 1.03207433e+00 6.33315992e+00 5.30083716e-01
-4.99389440e-01 2.69822359e-01 6.53475970e-02 2.00579688e-02
-1.00291705e+00 4.26696390e-01 -1.04216826e+00 1.71193272e-01
8.32512081e-01 -1.00787669e-01 6.27276421e-01 8.33400130e-01
-8.03825974e-01 -1.99401811e-01 -1.26731169e+00 1.65608078e-01
2.16898739e-01 -1.61475480e+00 2.11806282e-01 -2.55605519e-01
4.76504296e-01 -7.92164207e-02 -2.50743896e-01 7.20523059e-01
4.23046380e-01 -5.01270354e-01 8.73931706e-01 5.40689349e-01
3.97550106e-01 -6.54419720e-01 8.23185384e-01 5.38434803e-01
-1.05520105e+00 -3.54389399e-01 -5.83217263e-01 2.47244149e-01
-2.66759306e-01 4.16664928e-01 -6.72268689e-01 8.28379810e-01
7.87786901e-01 -2.10019916e-01 -7.11023510e-01 6.29673541e-01
-3.12822431e-01 6.22448206e-01 -3.24684471e-01 -4.62772161e-01
6.29113093e-02 1.96094513e-02 2.31530815e-01 1.13658404e+00
2.87686531e-02 1.32189527e-01 1.31314367e-01 8.47334564e-01
-1.68946490e-01 2.66196370e-01 -1.03745036e-01 2.19652265e-01
6.46864355e-01 9.62932825e-01 6.48454428e-02 -7.36043453e-01
-6.55481577e-01 5.93884110e-01 6.66849613e-01 4.01769280e-01
-4.08401787e-01 -3.96313280e-01 5.25850952e-01 1.02242738e-01
4.89410222e-01 3.32538724e-01 -2.85573184e-01 -1.36352324e+00
5.17520070e-01 -1.13483429e+00 1.32519305e+00 -7.14389622e-01
-1.31241524e+00 1.89876795e-01 7.51644075e-02 -1.60149202e-01
-1.97670981e-01 -7.29569972e-01 -3.46461058e-01 8.00723970e-01
-1.71306682e+00 -9.86130476e-01 4.60020043e-02 6.81866467e-01
3.07297498e-01 -1.20238826e-01 1.04754269e+00 2.17272177e-01
-4.03888255e-01 5.78846574e-01 -2.52245277e-01 -1.58161312e-01
4.18023437e-01 -1.54679799e+00 -1.38102069e-01 5.00634611e-01
2.95926154e-01 8.68959129e-01 5.94742596e-01 -4.83476937e-01
-1.89436626e+00 -1.17674828e+00 1.97121656e+00 -8.69885981e-01
8.82822514e-01 -1.06226690e-01 -1.04069948e+00 7.89330781e-01
2.38376975e-01 -3.35646905e-02 8.22360992e-01 4.26874131e-01
-1.03701699e+00 -5.62910676e-01 -1.14440954e+00 3.80581349e-01
6.66516364e-01 -1.08621562e+00 -1.27752113e+00 7.43962467e-01
9.84558284e-01 -1.94127709e-01 -1.14844668e+00 3.70563775e-01
5.79257429e-01 -9.43297803e-01 9.97879446e-01 -1.16143405e+00
2.87571967e-01 -1.37566224e-01 -7.63531148e-01 -5.73764563e-01
-1.78600773e-01 -2.73353785e-01 -4.32083696e-01 8.86385024e-01
8.04879904e-01 -7.28638291e-01 8.53003383e-01 9.08254743e-01
-1.83096179e-03 -1.03906047e+00 -1.06585097e+00 -6.66481376e-01
4.53615040e-02 -1.98180780e-01 7.41141915e-01 5.97327650e-01
3.20337385e-01 4.90903467e-01 2.76789099e-01 4.34786052e-01
4.74076331e-01 8.76340449e-01 7.17450500e-01 -9.87441301e-01
-5.88829398e-01 -3.55413377e-01 -9.05858818e-03 -1.07872987e+00
4.25443441e-01 -1.15404725e+00 -1.09854832e-01 -1.56015277e+00
8.04352611e-02 -4.28337365e-01 8.86493623e-02 4.18588668e-01
-2.97572911e-01 -2.90898532e-01 -1.31652847e-01 8.44622627e-02
-9.30996120e-01 1.57847553e-01 6.47427320e-01 -3.26942652e-01
4.49930102e-01 9.16923955e-02 -8.18867147e-01 4.43484724e-01
5.40704966e-01 -3.15701663e-01 -4.11585659e-01 -3.16288292e-01
9.06524122e-01 4.78029370e-01 4.09281313e-01 -6.45189881e-01
5.17766356e-01 -2.14455500e-01 -3.18454087e-01 -6.35690451e-01
2.21619591e-01 -6.76664591e-01 -1.72536746e-01 5.39343894e-01
-7.56753147e-01 1.47332728e-01 -2.08104953e-01 4.41673100e-01
-5.38356781e-01 -6.65255427e-01 5.64532638e-01 -1.42355248e-01
-6.09452426e-01 1.21194728e-01 -1.58289224e-01 8.69216561e-01
4.98537719e-01 3.28579932e-01 -6.49150074e-01 2.82774139e-02
-4.99333173e-01 3.28173548e-01 2.27486551e-01 2.18251035e-01
3.80473465e-01 -1.23807251e+00 -4.24823284e-01 -5.23591489e-02
5.90470374e-01 -1.23731546e-01 2.19104365e-02 5.64817667e-01
-5.76157212e-01 8.77599657e-01 4.24286425e-01 -1.40581086e-01
-1.20470345e+00 6.90943956e-01 3.79640132e-01 -6.88723624e-01
-2.44410232e-01 1.12907767e+00 4.93421592e-02 -1.03318763e+00
3.82318109e-01 -3.05238158e-01 -3.21664624e-02 1.65808365e-01
2.52254277e-01 2.64589936e-01 4.39995885e-01 -1.54811189e-01
-3.87708873e-01 1.76936865e-01 -2.02205833e-02 -1.95902377e-01
9.02931213e-01 1.81141749e-01 -8.14855516e-01 3.08533520e-01
1.20424879e+00 1.49100929e-01 -3.77677292e-01 -9.15946543e-01
6.38495028e-01 -2.87489325e-01 -3.49706888e-01 -1.06488192e+00
-4.99308616e-01 6.13920391e-01 -1.45869657e-01 4.08107668e-01
8.49982142e-01 4.29550469e-01 1.03125358e+00 1.02297294e+00
4.52044606e-01 -1.05528450e+00 1.37660965e-01 6.22769177e-01
9.01655138e-01 -6.45912290e-01 -1.07723691e-01 -4.03332919e-01
-3.95986348e-01 9.71529663e-01 6.65114641e-01 2.02557892e-01
4.27823275e-01 -8.24985094e-03 -1.64110512e-01 -5.00049770e-01
-1.48832273e+00 -1.51140362e-01 4.78632450e-01 1.25519931e-01
3.32262963e-01 7.25815371e-02 -1.93052188e-01 5.48475921e-01
-5.22774875e-01 -2.28551239e-01 -5.70716076e-02 8.15406799e-01
-6.67694032e-01 -1.05936003e+00 -2.38069132e-01 7.71910727e-01
-3.32454324e-01 -2.33058706e-01 -4.81139749e-01 6.12220943e-01
1.91930100e-01 9.90722716e-01 -1.93537816e-01 -1.72376826e-01
5.67230642e-01 6.30308151e-01 6.40531242e-01 -8.90648842e-01
-9.58557308e-01 -5.65879107e-01 7.02491343e-01 -6.96909726e-01
-4.08944814e-03 -5.85056722e-01 -1.02259648e+00 -1.69113934e-01
-2.78824925e-01 6.04316294e-01 3.16319138e-01 1.00660002e+00
4.20097381e-01 2.09734499e-01 3.20267290e-01 5.29466391e-01
-1.04413772e+00 -6.28161132e-01 -1.05839089e-01 4.97293115e-01
8.83476809e-02 -2.68302888e-01 -3.37278575e-01 -7.78493062e-02] | [10.350957870483398, 7.841146469116211] |
ff687aac-24d8-41e1-ba55-ea6ac8899d6d | a-fast-partial-video-copy-detection-using-knn | 2105.01713 | null | https://arxiv.org/abs/2105.01713v2 | https://arxiv.org/pdf/2105.01713v2.pdf | A Fast Partial Video Copy Detection Using KNN and Global Feature Database | We propose a fast partial video copy detection framework in this paper. In this framework all frame features of the reference videos are organized in a KNN searchable database. Instead of scanning all reference videos, the query video segment does a fast KNN search in the global feature database. The returned results are used to generate a short list of candidate videos. A modified temporal network is then used to localize the copy segment in the candidate videos. We evaluate different choice of CNN features on the VCDB dataset. Our benchmark F1 score exceeds the state of the art by a big margin. | ['Rushuai Liu', 'Hongwei Guo', 'Weijun Tan'] | 2021-05-04 | null | null | null | null | ['partial-video-copy-detection'] | ['computer-vision'] | [ 3.17086428e-02 -5.15044272e-01 -7.75070727e-01 -9.19348150e-02
-9.62750435e-01 -6.92142010e-01 2.74738520e-01 -7.89743736e-02
-6.41046882e-01 5.53340733e-01 1.68904543e-01 3.60275596e-01
-6.84079006e-02 -4.70965564e-01 -1.06510675e+00 -6.17060006e-01
-1.18634239e-01 9.74506661e-02 9.14822221e-01 2.95931220e-01
4.72222894e-01 6.83395803e-01 -1.67226923e+00 5.06304383e-01
2.26583064e-01 1.37077129e+00 6.51976168e-01 5.97421944e-01
1.30762026e-01 1.03366256e+00 -5.61249793e-01 -1.01204604e-01
4.96711731e-01 -3.74824256e-01 -7.07447171e-01 1.38327494e-01
7.70650208e-01 -6.67861819e-01 -9.32372808e-01 1.04925537e+00
2.61162817e-01 5.01284361e-01 2.08557233e-01 -1.04348803e+00
-2.70537168e-01 6.56565547e-01 -2.26297215e-01 6.36589825e-01
5.74793160e-01 -3.70201208e-02 1.03991139e+00 -1.22134137e+00
1.28229773e+00 8.02805603e-01 9.84013826e-02 4.85345989e-01
-6.19428098e-01 -7.31096804e-01 1.29719034e-01 9.31024730e-01
-1.75863063e+00 -5.79929531e-01 7.35558331e-01 -2.78253853e-01
1.03066504e+00 1.60625964e-01 7.30608225e-01 7.14656651e-01
1.43651471e-01 7.66481996e-01 2.00981274e-01 -2.35518038e-01
1.63487270e-01 -1.83730468e-01 1.21280298e-01 8.49943042e-01
-3.33874971e-02 2.04358101e-02 -1.01463652e+00 1.10294679e-02
6.93743944e-01 1.89060181e-01 -4.68233019e-01 -3.78334701e-01
-1.32391143e+00 6.76225066e-01 5.51753104e-01 4.02088821e-01
-3.67289513e-01 5.45818448e-01 5.05091548e-01 5.05012453e-01
-4.86128777e-02 1.59454241e-01 -3.90689135e-01 -1.20624900e-01
-1.24703574e+00 2.72692531e-01 6.76897466e-01 9.55834866e-01
9.44533467e-01 -2.94459820e-01 -4.50517505e-01 4.57678854e-01
7.88888633e-02 1.83354229e-01 4.71575797e-01 -1.20063627e+00
5.44394255e-01 5.69894195e-01 7.83990026e-02 -1.22815871e+00
3.92099582e-02 -2.65359938e-01 -4.14464414e-01 -1.01062901e-01
1.56244308e-01 3.68726283e-01 -8.79250586e-01 1.24391770e+00
4.04608369e-01 5.07106602e-01 -3.89159888e-01 1.08980227e+00
7.52657294e-01 8.45083714e-01 -3.66128772e-01 -3.29794645e-01
9.41188276e-01 -1.40042555e+00 -5.06522655e-01 1.69389039e-01
3.50853682e-01 -8.10780525e-01 5.80482006e-01 5.22745609e-01
-9.29044068e-01 -7.54627764e-01 -1.04402494e+00 -1.04241118e-01
-1.47707865e-01 3.14679563e-01 7.34634548e-02 1.88219517e-01
-1.37131846e+00 6.80334330e-01 -8.35436821e-01 -2.43427739e-01
5.29362559e-01 4.40300286e-01 -6.28659666e-01 -4.22645688e-01
-8.18113863e-01 4.69099075e-01 7.68528700e-01 1.61482647e-01
-1.48925364e+00 -4.68633980e-01 -6.44580066e-01 6.71883672e-02
7.97361672e-01 -2.23423779e-01 9.40608144e-01 -1.15067661e+00
-1.13130808e+00 6.93522394e-01 -3.04948151e-01 -5.86627245e-01
5.86761355e-01 -3.94361913e-01 -2.78576404e-01 6.95312321e-01
-4.97733206e-02 7.46853173e-01 1.19621372e+00 -6.86647058e-01
-8.47225666e-01 3.60284969e-02 1.07275121e-01 1.55527219e-01
-6.80010021e-02 3.76367033e-01 -1.34818411e+00 -7.59047806e-01
1.92561727e-02 -9.18356240e-01 8.95201787e-03 1.11458182e-01
-3.75690520e-01 -2.92089105e-01 1.01659203e+00 -6.34302676e-01
1.38462436e+00 -2.21510434e+00 2.40209192e-01 3.55538279e-01
2.99656689e-01 1.16740517e-01 -3.14438105e-01 3.03624421e-01
4.87649404e-02 -4.50141914e-02 4.37028021e-01 -5.51032200e-02
-2.38943785e-01 -2.10912317e-01 -2.06651673e-01 7.15252936e-01
4.93711866e-02 9.38585341e-01 -6.34893477e-01 -6.55636311e-01
1.96107313e-01 2.75616944e-01 -7.44190156e-01 2.98082769e-01
-3.27723473e-01 -2.26533730e-02 -5.00404418e-01 7.81957686e-01
4.27664280e-01 -2.44756535e-01 1.87986437e-02 -3.30299109e-01
-4.01055627e-02 6.08621463e-02 -1.34350228e+00 1.90273345e+00
1.75596595e-01 8.61223459e-01 -2.85022467e-01 -9.79474008e-01
7.36009359e-01 2.01926678e-02 6.10641181e-01 -4.67187971e-01
-1.21617563e-01 1.91095501e-01 -2.14628756e-01 -4.94849831e-01
7.02509761e-01 5.75031221e-01 9.83891264e-02 1.85965687e-01
2.74578869e-01 6.27980769e-01 3.21391284e-01 2.87862837e-01
1.30976558e+00 2.42914364e-01 2.33855948e-01 -1.43504739e-01
7.21931577e-01 1.09571539e-01 4.88256961e-01 8.30971956e-01
-2.65043825e-01 6.14992321e-01 3.69713336e-01 -8.36665273e-01
-9.89420295e-01 -8.27212214e-01 2.51092792e-01 1.30693829e+00
3.60365599e-01 -6.36747956e-01 -6.73098266e-01 -7.33356774e-01
-4.58433181e-02 -7.19752237e-02 -7.30163693e-01 1.03045313e-03
-7.21121728e-01 1.00601785e-01 1.42341599e-01 3.79696131e-01
5.97413421e-01 -1.17134690e+00 -8.02403927e-01 1.78479448e-01
-1.10277154e-01 -1.09568834e+00 -9.50567186e-01 -1.60079785e-02
-6.89381540e-01 -1.31316590e+00 -5.32566130e-01 -1.18250728e+00
4.31864917e-01 4.72943395e-01 8.80993366e-01 4.40988839e-01
-2.43564904e-01 2.38010406e-01 -6.99391007e-01 4.21464950e-01
-2.65306801e-01 1.00317053e-01 -7.06088021e-02 1.08563587e-01
3.14885706e-01 9.39320698e-02 -9.41576421e-01 4.42264497e-01
-9.32927132e-01 -5.39963722e-01 5.21088123e-01 6.30568743e-01
1.08535254e+00 7.24298209e-02 1.51944727e-01 -4.67531830e-01
3.60793658e-02 -3.67466718e-01 -8.72423351e-01 3.68764460e-01
-3.49200927e-02 -3.62179466e-02 4.38667834e-01 -6.99086845e-01
-3.86850238e-01 4.95495558e-01 2.91728765e-01 -1.21017325e+00
2.00356916e-02 6.07515991e-01 -6.82808906e-02 -2.63380110e-01
2.44394854e-01 4.95841086e-01 -3.68446082e-01 -5.11367261e-01
2.15289351e-02 3.66615742e-01 6.16913676e-01 -1.15738180e-03
8.24656188e-01 2.25681961e-01 -1.68923542e-01 -7.72357643e-01
-3.05125415e-01 -5.67174375e-01 -8.00266027e-01 -4.90031034e-01
8.38551044e-01 -1.07447338e+00 -4.42817241e-01 3.00168514e-01
-1.14689124e+00 -2.16417059e-01 -7.61637092e-02 5.33172369e-01
-3.38938057e-01 4.20199722e-01 -5.26640058e-01 -3.14499646e-01
-1.93817154e-01 -1.38929510e+00 1.02224469e+00 2.04699039e-02
1.66604705e-02 -2.45522380e-01 1.56215196e-02 8.70429575e-02
2.04712823e-01 1.02701351e-01 4.84873086e-01 -8.89224350e-01
-1.48158908e+00 -5.16940117e-01 -1.26528978e-01 2.20869452e-01
-3.19224708e-02 2.15583101e-01 -6.47149920e-01 -4.81913924e-01
-1.30886689e-01 -2.85426497e-01 1.27529490e+00 4.03418839e-01
1.33385503e+00 -2.38447890e-01 -6.74209714e-01 6.94161475e-01
1.71906602e+00 3.09367150e-01 6.04301274e-01 3.76421839e-01
6.62819624e-01 1.23856768e-01 8.88762236e-01 4.89986479e-01
1.34594468e-02 7.97514319e-01 4.68536377e-01 4.99846637e-01
-1.14260301e-01 -2.45430693e-01 6.67858720e-01 5.92691958e-01
1.55867949e-01 -4.94591594e-01 -6.60875976e-01 5.40298402e-01
-1.72788703e+00 -1.15827703e+00 4.02382612e-01 2.22001672e+00
3.73920768e-01 1.03242472e-01 1.84648976e-01 -1.50439158e-01
1.01567972e+00 2.96661168e-01 -5.92734337e-01 1.44773364e-01
-1.58317044e-01 2.61498597e-02 5.19168735e-01 2.50498265e-01
-1.42041838e+00 1.01678681e+00 6.15782070e+00 1.03210807e+00
-1.14409745e+00 1.36015313e-02 5.53692698e-01 -4.89075363e-01
4.94241863e-02 2.52250619e-02 -9.17663336e-01 4.23389226e-01
7.32811213e-01 -2.13191256e-01 5.69607735e-01 9.80891228e-01
7.03809038e-02 -2.98231602e-01 -1.19301438e+00 1.26554370e+00
4.36558664e-01 -1.82227170e+00 1.92506284e-01 -5.28711360e-04
8.23753357e-01 3.80755007e-01 -1.15230918e-01 9.14532840e-02
-3.12047750e-01 -9.41528201e-01 9.38372135e-01 6.30220592e-01
9.38443065e-01 -8.62660170e-01 5.63981414e-01 5.06918132e-02
-1.46162653e+00 -3.11659068e-01 -7.23137319e-01 3.87433767e-01
-1.45350978e-01 2.55344361e-01 -5.65190613e-01 3.07716727e-01
9.41184163e-01 8.31202507e-01 -6.49449646e-01 1.45455277e+00
2.35076219e-01 4.29207236e-01 -3.45628202e-01 -1.38715263e-02
3.12577069e-01 1.29826292e-01 4.76871550e-01 1.12489748e+00
4.76368874e-01 -2.42338311e-02 4.30086493e-01 5.15716970e-01
-4.91962224e-01 3.17173272e-01 -4.59470659e-01 -1.39951035e-02
6.23763978e-01 1.18681741e+00 -1.01542759e+00 -3.91974688e-01
-6.14891768e-01 1.22011781e+00 3.31467390e-01 1.94161564e-01
-8.61561656e-01 -4.51774001e-01 3.91435474e-01 2.30468120e-02
1.06732297e+00 -8.43035802e-02 9.39973712e-01 -1.36849499e+00
2.30449662e-01 -8.03733349e-01 5.27279854e-01 -6.08336627e-01
-8.09520245e-01 7.36813843e-01 -2.37314597e-01 -1.56326294e+00
-2.79284328e-01 -4.67824012e-01 -4.29636776e-01 3.58249068e-01
-1.37138081e+00 -8.02877247e-01 -5.04417002e-01 9.36888993e-01
9.46249723e-01 -3.64635408e-01 3.40390503e-01 3.15438211e-01
-5.35783947e-01 6.36925936e-01 3.27272177e-01 5.37903070e-01
7.01097667e-01 -6.68138742e-01 2.89595693e-01 9.71923411e-01
2.56481946e-01 4.86316770e-01 3.82615268e-01 -9.13811922e-01
-1.80684137e+00 -1.31387222e+00 5.84523320e-01 -1.75327286e-01
6.24756873e-01 -3.27720195e-01 -8.37052584e-01 6.38370037e-01
2.10503012e-01 5.48912287e-01 2.85089761e-01 -7.26511121e-01
-4.73950207e-01 -3.49480867e-01 -8.74054730e-01 1.60403788e-01
9.94378150e-01 -6.08076751e-01 -3.05420905e-01 1.75365314e-01
1.06163025e+00 -4.69662160e-01 -6.76999390e-01 1.82852864e-01
8.24099779e-01 -8.03230703e-01 7.59305477e-01 -5.60873508e-01
3.65403771e-01 -4.71364200e-01 -5.88784277e-01 -6.26332343e-01
-3.15968186e-01 -8.95466387e-01 -3.29798043e-01 9.64189768e-01
2.12761760e-01 4.22584713e-02 1.05460083e+00 2.51886159e-01
-1.36556225e-02 -6.11087739e-01 -1.18945575e+00 -7.44450212e-01
-4.60815430e-01 -3.87407988e-01 4.16072786e-01 5.15118182e-01
-1.14047535e-01 -1.52723253e-01 -3.58805627e-01 9.34151337e-02
5.61992526e-01 4.32411224e-01 6.37547791e-01 -7.62449741e-01
-2.00789690e-01 -3.09666157e-01 -7.80331492e-01 -1.21160913e+00
2.33481646e-01 -9.38622177e-01 1.03226513e-01 -1.04939306e+00
5.34024358e-01 -6.98448941e-02 -6.97393894e-01 1.50974885e-01
3.58172432e-02 5.27812719e-01 4.94984567e-01 4.90325928e-01
-1.22255898e+00 2.26764038e-01 8.33045423e-01 -1.78006142e-01
-2.02721581e-01 -1.66999608e-01 -4.54740673e-02 2.83961535e-01
7.33409286e-01 -6.11893713e-01 -1.49276227e-01 -4.11923587e-01
1.02933601e-01 3.25196534e-01 4.21895862e-01 -1.31102240e+00
4.87975359e-01 -1.08631812e-01 6.88955247e-01 -1.11352098e+00
1.96725905e-01 -8.57346058e-01 3.34421396e-01 4.81000364e-01
-6.13554478e-01 2.42819607e-01 -1.36109451e-02 8.47524226e-01
-3.79087865e-01 -2.87301689e-01 6.11411631e-01 -6.97108507e-02
-1.13391113e+00 6.41820490e-01 -1.84263363e-01 -2.48365641e-01
1.16873145e+00 -4.17500734e-01 -1.46156475e-01 -3.07048678e-01
-5.39014637e-01 3.69164608e-02 7.11497545e-01 6.14866078e-01
1.02598023e+00 -1.39511263e+00 -4.27098095e-01 2.07656726e-01
1.32179573e-01 -2.85742670e-01 4.02112514e-01 5.72216153e-01
-8.15552473e-01 5.75997293e-01 -2.73932427e-01 -6.71913564e-01
-1.38679624e+00 1.04935277e+00 3.05688620e-01 3.00327260e-02
-4.75555807e-01 9.20846879e-01 9.59549621e-02 3.93609762e-01
5.78328133e-01 -3.95216584e-01 -2.00775951e-01 1.25407621e-01
7.98306882e-01 4.45814937e-01 5.69636635e-02 -9.51879621e-01
-5.72234213e-01 7.12906659e-01 -1.96154907e-01 1.14586599e-01
1.30936861e+00 -2.42999583e-01 -1.48577258e-01 3.37630250e-02
1.79524660e+00 6.48592263e-02 -1.15161312e+00 -2.93746442e-01
1.57332867e-01 -8.21987569e-01 3.03873364e-02 -3.17962915e-01
-1.50248611e+00 2.74673522e-01 6.94952726e-01 -2.23703504e-01
1.32454991e+00 3.57021779e-01 6.27756655e-01 7.52772987e-01
5.48385739e-01 -1.04814506e+00 3.09244007e-01 3.38817626e-01
8.24270427e-01 -8.80655706e-01 8.38143602e-02 -1.77074388e-01
-5.42556465e-01 1.38802660e+00 6.21303141e-01 -5.86535454e-01
6.05013728e-01 7.30718523e-02 -1.99880242e-01 -2.08410978e-01
-9.81928945e-01 -5.41844256e-02 3.90539497e-01 2.87158132e-01
3.00088488e-02 -4.91663992e-01 -2.46854518e-02 2.52660125e-01
1.42845809e-01 2.88032502e-01 5.55237293e-01 7.01668620e-01
-7.52674162e-01 -8.72137547e-01 -2.21434936e-01 5.82069218e-01
-5.33644617e-01 -2.57651079e-02 -1.79830953e-01 6.94582045e-01
-4.26551178e-02 6.16842389e-01 2.65930414e-01 -7.10501909e-01
1.06691204e-01 -2.52887271e-02 3.14962476e-01 -2.99717396e-01
-4.64100093e-01 4.79655564e-01 -3.22020829e-01 -1.31086183e+00
-5.45242429e-01 -7.21414864e-01 -9.43373978e-01 -9.71421972e-02
-4.22563404e-01 -3.58211249e-02 2.58024216e-01 6.04262888e-01
3.38862985e-01 1.63875017e-02 8.42562377e-01 -9.40043569e-01
-2.55539060e-01 -5.93653321e-01 -4.81056362e-01 2.61626124e-01
5.10452390e-01 -5.54837525e-01 -1.99989617e-01 1.14597961e-01] | [9.199127197265625, 0.2959247827529907] |
6bbcc435-722b-42b3-85b8-0767301c1747 | real-time-scalable-dense-surfel-mapping | 1909.04250 | null | https://arxiv.org/abs/1909.04250v1 | https://arxiv.org/pdf/1909.04250v1.pdf | Real-time Scalable Dense Surfel Mapping | In this paper, we propose a novel dense surfel mapping system that scales well in different environments with only CPU computation. Using a sparse SLAM system to estimate camera poses, the proposed mapping system can fuse intensity images and depth images into a globally consistent model. The system is carefully designed so that it can build from room-scale environments to urban-scale environments using depth images from RGB-D cameras, stereo cameras or even a monocular camera. First, superpixels extracted from both intensity and depth images are used to model surfels in the system. superpixel-based surfels make our method both run-time efficient and memory efficient. Second, surfels are further organized according to the pose graph of the SLAM system to achieve $O(1)$ fusion time regardless of the scale of reconstructed models. Third, a fast map deformation using the optimized pose graph enables the map to achieve global consistency in real-time. The proposed surfel mapping system is compared with other state-of-the-art methods on synthetic datasets. The performances of urban-scale and room-scale reconstruction are demonstrated using the KITTI dataset and autonomous aggressive flights, respectively. The code is available for the benefit of the community. | ['Shaojie Shen', 'Fei Gao', 'Kaixuan Wang'] | 2019-09-10 | null | null | null | null | ['3d-point-cloud-reconstruction'] | ['computer-vision'] | [ 9.12180692e-02 -1.83583856e-01 5.40220253e-02 -5.05922496e-01
-6.24526560e-01 -6.19344711e-01 2.34834552e-01 -1.68961123e-01
-5.27272761e-01 5.90830922e-01 -4.31012630e-01 2.39114746e-01
2.27159169e-02 -1.00980544e+00 -9.36885476e-01 -2.30456918e-01
3.09645385e-01 8.58546674e-01 7.48427391e-01 -3.20301831e-01
2.95368731e-01 7.15050638e-01 -1.60332966e+00 -2.81304181e-01
8.87607932e-01 1.13387465e+00 7.68636703e-01 5.30133665e-01
-3.61458585e-02 4.42348003e-01 -1.66652381e-01 4.31489050e-02
8.08172405e-01 -1.34921283e-01 -2.26703793e-01 2.24381864e-01
1.03241313e+00 -4.15432543e-01 -4.16402489e-01 1.12482226e+00
3.11100155e-01 2.51315683e-01 4.30892408e-02 -9.99359667e-01
1.03217326e-01 -1.70822009e-01 -6.75412357e-01 -4.96038467e-01
5.46792388e-01 -1.71184659e-01 5.43936610e-01 -9.20374513e-01
8.63478482e-01 1.01608598e+00 8.00100863e-01 1.29988240e-02
-9.67664599e-01 -7.77491510e-01 -9.64433700e-02 2.01087013e-01
-1.87113738e+00 -2.62634933e-01 6.68775976e-01 -2.38505691e-01
8.29652607e-01 3.69589508e-01 8.17205369e-01 4.12776977e-01
5.59981823e-01 -5.50055206e-02 1.15990925e+00 -3.15699488e-01
2.50712782e-01 4.26284969e-02 -2.38820881e-01 1.18183947e+00
5.44869304e-01 1.59455940e-01 -1.11581481e+00 4.70595956e-02
9.78481293e-01 1.88728750e-01 -3.55689675e-01 -7.78712273e-01
-1.67455125e+00 5.07428110e-01 8.23971570e-01 -2.04163074e-01
-2.14923367e-01 5.89769900e-01 -2.65415817e-01 -6.55289292e-02
1.21135496e-01 1.42739043e-01 2.94855721e-02 -1.69990156e-02
-1.22664261e+00 1.06227919e-01 3.31669509e-01 1.33230996e+00
1.49535620e+00 9.91575699e-03 7.05151558e-01 5.03551006e-01
3.94148678e-01 1.10211623e+00 1.89708024e-01 -1.39550984e+00
5.19481659e-01 5.27479172e-01 4.79851902e-01 -1.29046261e+00
-5.24613619e-01 -1.41426951e-01 -6.57850862e-01 3.47621262e-01
5.55794425e-02 2.97493935e-01 -6.96621716e-01 1.23096251e+00
5.25354803e-01 3.54807317e-01 -1.07471310e-01 9.20236826e-01
5.76387167e-01 7.23207653e-01 -9.13500726e-01 3.38228121e-02
1.08778870e+00 -1.00501585e+00 -6.71216846e-01 -7.23981917e-01
3.74926269e-01 -8.54540586e-01 8.04077566e-01 3.98424596e-01
-9.16460872e-01 -7.16775417e-01 -1.41657007e+00 -1.86805725e-01
-1.05896533e-01 2.02393368e-01 2.70471156e-01 3.67654830e-01
-1.30248690e+00 1.62617207e-01 -1.03614712e+00 -6.84057891e-01
-2.76282579e-01 3.81558090e-01 -4.88287747e-01 -3.59177679e-01
-8.65465701e-01 9.12813306e-01 4.13934559e-01 1.71343088e-01
-6.99774444e-01 -3.50263774e-01 -9.57425952e-01 -3.56218129e-01
2.69370735e-01 -7.51035333e-01 7.85529196e-01 -4.35645163e-01
-1.28770638e+00 7.42477059e-01 -3.62775087e-01 -3.39361042e-01
6.60782218e-01 -1.90479383e-01 1.94807053e-02 2.30434254e-01
2.82779008e-01 9.16824043e-01 5.62644482e-01 -1.36084592e+00
-7.85176814e-01 -4.01301950e-01 -1.25914082e-01 5.44290423e-01
1.89570233e-01 -5.07520378e-01 -8.13902974e-01 -7.23932087e-02
9.11584795e-01 -1.30356765e+00 -2.39220619e-01 3.12239140e-01
-1.15331925e-01 6.45474255e-01 8.98182809e-01 -5.83763540e-01
7.11973131e-01 -2.02579093e+00 2.07826868e-01 3.56833041e-01
3.32101360e-02 -5.33192754e-01 1.62212044e-01 3.82350624e-01
6.59363985e-01 -5.05525708e-01 -2.14114055e-01 -6.46631300e-01
-1.65051743e-01 4.42018449e-01 -1.24578446e-01 8.48072946e-01
-6.60891294e-01 6.00562274e-01 -6.28664434e-01 -5.90065956e-01
7.11259127e-01 4.21048373e-01 -2.46282622e-01 3.78825106e-02
1.02270320e-01 5.45861304e-01 -1.98285952e-01 7.86617458e-01
1.01423061e+00 1.39884338e-01 3.34366821e-02 -4.61613476e-01
-6.03102326e-01 -3.38541679e-02 -1.51237476e+00 2.27406931e+00
-6.42383516e-01 6.15994394e-01 4.58864659e-01 -2.42586210e-01
1.21608663e+00 -3.03554475e-01 3.95199358e-01 -7.04294801e-01
-3.66551131e-02 5.40137708e-01 -6.57083035e-01 2.29381666e-01
9.34298396e-01 -9.33888461e-03 -2.62804955e-01 2.09339187e-01
-2.58369923e-01 -9.65732932e-01 -2.23294143e-02 2.18407750e-01
9.75703061e-01 3.07786167e-01 1.50394812e-01 -3.93231362e-01
4.48786288e-01 6.17756963e-01 6.57723844e-01 4.92405236e-01
8.33560601e-02 6.53433323e-01 -4.33082819e-01 -5.65848470e-01
-1.12275767e+00 -1.19042075e+00 -1.57477930e-01 3.53186697e-01
1.12754226e+00 -3.88806552e-01 -6.47212923e-01 -1.04959399e-01
6.72240332e-02 3.60912710e-01 -3.20665210e-01 2.33139053e-01
-4.98940498e-01 -2.23552600e-01 7.89923817e-02 3.06245565e-01
1.15377522e+00 -4.74133611e-01 -1.23614156e+00 1.39501542e-01
-3.95745486e-01 -1.37356460e+00 -5.18759131e-01 1.13727404e-02
-9.17784333e-01 -8.56376588e-01 -1.72851101e-01 -6.50896728e-01
7.90130258e-01 8.36871326e-01 7.60641754e-01 -1.03749581e-01
-1.96060464e-01 5.35607874e-01 -2.87807435e-01 -4.58569601e-02
-1.31074106e-02 -2.32612520e-01 3.33106667e-01 -1.04873046e-01
-3.54943871e-01 -3.39082569e-01 -5.83932281e-01 8.51564407e-01
-5.18672585e-01 5.86260080e-01 4.38950986e-01 5.72764039e-01
1.23581505e+00 -2.01964863e-02 -4.06300157e-01 -4.29869741e-01
-1.43675134e-01 2.70734429e-01 -1.29704535e+00 1.20453618e-01
-5.54661036e-01 -1.46523371e-01 2.42634997e-01 9.59579870e-02
-9.33712661e-01 8.09330940e-01 1.84619337e-01 -4.38297987e-01
3.14858347e-01 2.84906834e-01 -1.06828451e-01 -8.22036922e-01
5.46119630e-01 4.52543020e-01 -7.11474419e-02 -1.87960118e-01
2.57307440e-01 3.56030822e-01 8.72433603e-01 -3.31370592e-01
1.10620391e+00 1.00217438e+00 2.94326305e-01 -9.06425595e-01
-6.60672843e-01 -6.15295649e-01 -9.67449605e-01 -6.04943454e-01
1.05205166e+00 -1.19892454e+00 -4.11652029e-01 4.26131219e-01
-1.10233951e+00 -4.48825359e-01 -3.21885385e-03 6.76413953e-01
-7.58433342e-01 1.96706831e-01 -3.62277150e-01 -5.00473380e-01
-1.13418102e-01 -1.29793870e+00 1.40513694e+00 1.86668634e-01
2.55224071e-02 -7.60689557e-01 2.83852130e-01 4.27183002e-01
2.70311028e-01 5.31390071e-01 -6.55728206e-02 5.62025070e-01
-1.35751522e+00 -2.85663903e-01 -1.69311315e-01 -1.38990894e-01
6.20418824e-02 -1.88967660e-01 -7.34993935e-01 -4.78769571e-01
1.07639572e-02 -1.08395129e-01 6.23824179e-01 1.39386237e-01
6.04998469e-01 4.90843840e-02 -4.29802030e-01 1.24319744e+00
1.72887456e+00 1.19280137e-01 4.97582495e-01 6.64216936e-01
9.76775169e-01 4.21215206e-01 1.23921680e+00 4.32128310e-01
7.05928445e-01 1.11244607e+00 7.59449303e-01 -1.89709514e-01
-1.47599921e-01 -3.34850937e-01 5.15347540e-01 1.06954598e+00
-8.18939954e-02 1.35758996e-01 -1.12382221e+00 1.54569611e-01
-1.85572648e+00 -5.64422607e-01 -3.70528072e-01 2.46233082e+00
1.85382590e-01 -7.38266706e-02 -5.19299328e-01 -3.19972157e-01
5.72605491e-01 3.23368847e-01 -4.89117533e-01 -1.03293635e-01
-6.29317313e-02 4.29165037e-03 1.24410260e+00 1.06290293e+00
-7.98433006e-01 1.10986698e+00 5.15605688e+00 3.97218913e-01
-1.13382638e+00 3.20135295e-01 -8.75786468e-02 -3.69266719e-02
-2.36712620e-01 1.82847261e-01 -1.02732778e+00 1.35612726e-01
7.95632243e-01 -8.14986154e-02 5.70178390e-01 9.73249674e-01
5.75564615e-02 -6.37259960e-01 -7.49536276e-01 1.47349501e+00
3.14082295e-01 -1.41878235e+00 -3.68244082e-01 2.70140439e-01
9.22663212e-01 5.79714477e-01 -2.74132580e-01 -2.82102078e-01
1.36949599e-01 -5.57475626e-01 1.34064317e+00 5.50165534e-01
8.94599795e-01 -8.00279319e-01 6.14103079e-01 6.01699412e-01
-1.87703669e+00 2.98868030e-01 -7.54759789e-01 2.88061257e-02
4.05535936e-01 5.16193867e-01 -8.40210974e-01 8.60176802e-01
9.25177515e-01 7.36346960e-01 -7.08480358e-01 7.87073016e-01
-1.83761403e-01 -2.81224728e-01 -6.43928766e-01 3.18691283e-01
1.58160150e-01 -6.46771252e-01 2.38641247e-01 7.27769136e-01
9.18273330e-01 6.83988854e-02 3.72786790e-01 6.71092629e-01
1.72156006e-01 -1.61019400e-01 -7.49221921e-01 5.80192089e-01
5.41578174e-01 1.43370330e+00 -9.15490985e-01 -3.32878232e-01
-2.14173913e-01 1.25414705e+00 8.91729593e-02 -2.60024108e-02
-1.03142595e+00 -1.67272970e-01 5.15973389e-01 4.56588179e-01
6.65416662e-03 -9.26072478e-01 -2.71050870e-01 -1.15125883e+00
3.76207270e-02 -2.60425717e-01 -2.92633981e-01 -1.09667170e+00
-3.57094467e-01 7.85636783e-01 -7.07175434e-02 -1.50138235e+00
-1.52776134e-03 -2.44432226e-01 -6.80449679e-02 7.25326121e-01
-1.47086191e+00 -1.03176808e+00 -1.34759152e+00 6.50143445e-01
4.20060128e-01 2.28391930e-01 6.56996429e-01 9.59911048e-02
-1.60468936e-01 -2.57964265e-02 2.79717475e-01 -3.67961586e-01
9.00089860e-01 -9.39607203e-01 5.92145920e-01 1.15232646e+00
2.94904828e-01 3.34947973e-01 5.49376547e-01 -9.75038469e-01
-1.81261289e+00 -1.24743187e+00 6.23330176e-01 -5.12683272e-01
2.17499614e-01 -7.83001125e-01 -4.62483048e-01 6.64069831e-01
-6.63600937e-02 2.33593017e-01 2.38292236e-02 -5.28954923e-01
4.42370437e-02 -6.86362863e-01 -1.19852817e+00 1.51609167e-01
1.11910284e+00 -4.74076033e-01 -2.09773228e-01 3.61743838e-01
6.41651332e-01 -1.12110782e+00 -7.29990780e-01 4.63158101e-01
6.31785095e-01 -1.28948438e+00 1.03916943e+00 8.60536397e-01
-2.71794885e-01 -9.70560968e-01 -6.12179637e-01 -9.94458616e-01
-1.19003035e-01 -3.49675000e-01 2.61279732e-01 7.86992431e-01
9.56764817e-02 -8.42430294e-01 8.40704858e-01 4.17015791e-01
-4.22777712e-01 -2.17882663e-01 -1.24138284e+00 -8.50402534e-01
-9.22314107e-01 -3.39636445e-01 5.51736712e-01 7.21851766e-01
-5.03723979e-01 -4.25492693e-03 -4.47548807e-01 6.83557332e-01
1.09965098e+00 3.22690070e-01 1.30190980e+00 -1.09747314e+00
-1.68742910e-01 1.55449212e-01 -6.58319533e-01 -1.21755135e+00
-6.21765032e-02 -6.74649954e-01 4.20725763e-01 -1.82466543e+00
-1.88351452e-01 -7.43957162e-01 2.26855680e-01 2.49120489e-01
3.95818919e-01 6.86854005e-01 3.47302049e-01 5.68253338e-01
-6.24610722e-01 5.09989321e-01 1.00448501e+00 -8.00104961e-02
-1.60209671e-01 -4.50758696e-01 1.18805259e-01 6.76499426e-01
5.42019010e-01 -3.92863631e-01 -4.53003228e-01 -6.39540851e-01
4.33465660e-01 3.14174145e-01 2.75026709e-01 -1.67677176e+00
4.43017751e-01 -2.89788157e-01 4.37736809e-01 -1.08593380e+00
9.36444402e-01 -1.14065683e+00 8.77229631e-01 7.48600066e-01
3.67613941e-01 4.39296246e-01 1.25898138e-01 5.40503323e-01
-2.62519330e-01 -6.19955827e-03 9.30515885e-01 -2.24005967e-01
-1.00202215e+00 4.09450233e-01 1.26375407e-01 -3.67214739e-01
1.31637561e+00 -4.95967120e-01 -2.62416124e-01 -4.37096417e-01
-2.01767102e-01 1.48701310e-01 1.43295562e+00 8.48076120e-02
1.00359130e+00 -1.41744196e+00 -4.25178617e-01 5.36694765e-01
2.77142823e-01 5.69384933e-01 3.33195329e-01 9.15747821e-01
-1.59459043e+00 3.09520245e-01 -3.57139677e-01 -1.12609363e+00
-1.19434023e+00 2.12608501e-01 3.33938301e-01 1.66804418e-01
-6.77964807e-01 7.02561200e-01 3.05914074e-01 -6.10931873e-01
-7.27928290e-03 -5.24347901e-01 7.03244686e-01 -3.78051817e-01
4.90553856e-01 4.37359869e-01 2.03797847e-01 -1.01885819e+00
-7.22957313e-01 1.38800192e+00 5.82706749e-01 -3.91830742e-01
1.07391548e+00 -5.92603564e-01 -2.49888524e-01 3.71310651e-01
8.71912360e-01 4.16771501e-01 -1.47223377e+00 1.39532611e-01
-4.86382633e-01 -8.27063799e-01 2.80594349e-01 -2.18287155e-01
-8.97405148e-01 7.57170141e-01 7.94748425e-01 -4.69978660e-01
1.00544691e+00 -1.21804982e-01 7.96767533e-01 4.86767441e-01
1.26730406e+00 -1.02105677e+00 -1.41904533e-01 4.77513194e-01
8.05752754e-01 -1.16752744e+00 3.36093575e-01 -4.68869805e-01
-5.36637783e-01 1.15886247e+00 5.52350938e-01 -3.02088678e-01
2.82334417e-01 4.10305589e-01 8.09618980e-02 -7.05808448e-03
-2.14142248e-01 7.71666095e-02 8.74305591e-02 5.47314763e-01
-2.69027025e-01 1.87418014e-01 1.70033067e-01 -2.58240968e-01
-4.56064790e-01 -2.44664684e-01 6.98023915e-01 8.97647381e-01
-8.68501425e-01 -1.10199523e+00 -8.58600914e-01 -1.05935566e-01
6.03886366e-01 1.47501811e-01 -2.71549970e-01 8.53217959e-01
2.02437520e-01 6.30966783e-01 2.20934138e-01 -7.08619535e-01
3.81892115e-01 -4.14240509e-01 7.35670984e-01 -4.89010006e-01
-2.00003847e-01 9.12908986e-02 -4.71277870e-02 -1.25871778e+00
-3.05430710e-01 -6.51407599e-01 -1.50736058e+00 -3.91267806e-01
-1.98214367e-01 -4.98030931e-02 1.14598775e+00 4.64901656e-01
3.12474936e-01 9.84950811e-02 5.89469969e-01 -1.32316780e+00
7.46629387e-02 -4.94336635e-01 -7.79612064e-01 -1.76256016e-01
1.25950813e-01 -8.21512341e-01 -1.95467189e-01 -4.31783460e-02] | [7.370563507080078, -2.240346670150757] |
7de36431-d445-4362-a49d-6b7ceb88879e | how-machine-deep-learning-helps-us-understand | 1903.03408 | null | http://arxiv.org/abs/1903.03408v2 | http://arxiv.org/pdf/1903.03408v2.pdf | How Machine (Deep) Learning Helps Us Understand Human Learning: the Value of Big Ideas | I use simulation of two multilayer neural networks to gain intuition into the
determinants of human learning. The first network, the teacher, is trained to
achieve a high accuracy in handwritten digit recognition. The second network,
the student, learns to reproduce the output of the first network. I show that
learning from the teacher is more effective than learning from the data under
the appropriate degree of regularization. Regularization allows the teacher to
distinguish the trends and to deliver "big ideas" to the student. I also model
other learning situations such as expert and novice teachers, high- and
low-ability students and biased learning experience due to, e.g., poverty and
trauma. The results from computer simulation accord remarkably well with
finding of the modern psychological literature. The code is written in MATLAB
and will be publicly available from the author's web page. | ['Marc Maliar'] | 2019-02-16 | null | null | null | null | ['handwritten-digit-recognition'] | ['computer-vision'] | [-2.56569535e-01 4.16932911e-01 -3.20349574e-01 -3.57698262e-01
1.29087225e-01 -2.29985222e-01 7.34015480e-02 6.89511672e-02
-5.71582317e-01 5.82330525e-01 2.34740204e-03 -6.48704171e-01
-2.89398819e-01 -6.81949973e-01 -6.76780701e-01 -5.67207456e-01
3.66390407e-01 1.04135245e-01 -2.74284512e-01 -1.42237858e-03
6.22207820e-01 4.30443227e-01 -1.23044109e+00 -2.56000668e-01
1.26021814e+00 5.07601142e-01 3.25435787e-01 3.98549080e-01
2.98372582e-02 1.13731658e+00 -4.09918517e-01 -5.43784082e-01
1.56470016e-01 -6.03674829e-01 -5.07318556e-01 1.54025719e-01
2.79008448e-01 -5.29220462e-01 -5.31422317e-01 1.30296159e+00
3.71963322e-01 1.43848971e-01 7.50940025e-01 -8.92134905e-01
-1.07401967e+00 6.83350682e-01 -3.55004072e-01 1.88825384e-01
1.43619671e-01 1.74310416e-01 3.60351324e-01 -8.71872783e-01
2.28550807e-01 1.44485819e+00 6.02194726e-01 6.07459843e-01
-1.19297230e+00 -8.01653862e-01 1.13783419e-01 -1.20765217e-01
-1.04695189e+00 -3.09038758e-01 6.85429633e-01 -7.82180965e-01
3.73215437e-01 -1.56491086e-01 8.30787361e-01 1.06122851e+00
4.02272403e-01 4.35208589e-01 1.39269519e+00 -5.15654325e-01
1.59810707e-01 8.05406272e-01 5.67346811e-01 6.44502163e-01
5.38586199e-01 5.37996233e-01 -3.14743817e-01 1.92131668e-01
1.23090136e+00 5.17454483e-02 -2.08537534e-01 -2.87500530e-01
-6.56384647e-01 8.24158072e-01 3.85422558e-01 2.94923812e-01
-7.57968843e-01 2.54394021e-02 -1.14056744e-01 9.91024792e-01
1.11220136e-01 4.41114813e-01 -2.32511312e-01 7.72711169e-03
-5.39788485e-01 1.53573513e-01 8.09079528e-01 3.35008472e-01
5.34193039e-01 5.70200562e-01 1.81828618e-01 7.79213667e-01
5.23940563e-01 3.39503825e-01 7.43886173e-01 -1.29697013e+00
1.36599287e-01 5.99666774e-01 3.47627960e-02 -1.04815114e+00
-1.70741662e-01 -8.35438967e-01 -5.91117561e-01 6.92616343e-01
8.16508591e-01 -7.05135643e-01 -9.08129811e-01 1.65974402e+00
-4.84772623e-02 -9.60826203e-02 6.74166381e-02 8.23252141e-01
5.57358444e-01 5.59960425e-01 1.85151175e-01 -3.90121043e-01
9.40311193e-01 -5.63929498e-01 -5.34600973e-01 -4.59826708e-01
3.94042015e-01 -3.96468073e-01 7.91369796e-01 6.38404012e-01
-1.52305114e+00 -6.35479808e-01 -8.29527259e-01 2.87380487e-01
-3.45985711e-01 -7.29294196e-02 6.76804543e-01 7.40275145e-01
-9.51718748e-01 9.07801569e-01 -8.17020357e-01 -3.74077857e-01
5.16538441e-01 3.83775413e-01 -1.96646415e-02 -5.05942628e-02
-9.12818730e-01 1.07446456e+00 3.60218734e-01 -6.08715154e-02
-9.55855787e-01 -8.19789588e-01 -5.27293742e-01 2.83876359e-01
5.40592521e-02 -5.88113606e-01 1.32181752e+00 -1.84381819e+00
-1.44638944e+00 7.81196952e-01 2.37745017e-01 -1.47267044e-01
4.49236780e-01 -6.00746507e-03 1.46207334e-02 -6.75432105e-03
-1.20398350e-01 4.71460313e-01 7.80248821e-01 -1.24855578e+00
-2.87394136e-01 -7.65254021e-01 -2.12743491e-01 2.89532274e-01
-3.75715703e-01 1.73244387e-01 2.54707694e-01 -6.65983915e-01
1.27192721e-01 -5.69645345e-01 -2.79202670e-01 -2.73292005e-01
1.23276794e-02 -3.58092755e-01 3.11696708e-01 -6.62111759e-01
9.24412966e-01 -2.13928938e+00 -3.25589329e-02 4.57659096e-01
3.00915748e-01 2.55866408e-01 8.38983282e-02 2.87817895e-01
-5.10515511e-01 3.17533985e-02 2.88102090e-01 1.41012818e-01
-1.52729824e-01 2.54301071e-01 6.12956099e-02 4.31012601e-01
2.54676566e-02 5.58132052e-01 -6.78723633e-01 -1.24552637e-01
-2.01323882e-01 3.77099246e-01 -4.64345783e-01 4.09021050e-01
3.88681889e-01 3.78065497e-01 -5.78636765e-01 2.90177554e-01
3.63197923e-01 -3.71327043e-01 1.76581457e-01 5.30954897e-01
-1.62643671e-01 4.26176667e-01 -1.09290314e+00 8.22165072e-01
-9.30133164e-02 8.01369429e-01 2.90876597e-01 -1.41473734e+00
9.95383024e-01 5.30417740e-01 -3.07591349e-01 -6.57306015e-01
3.50770712e-01 2.79594660e-01 6.42826378e-01 -8.08013201e-01
7.54113346e-02 -4.40165699e-01 5.76983690e-01 6.69335425e-01
6.98117688e-02 -7.86529332e-02 -4.91898246e-02 2.05742553e-01
6.13387823e-01 -2.18702450e-01 -3.40204723e-02 -4.48602736e-01
1.70032844e-01 -1.75143763e-01 6.86151922e-01 9.99285877e-01
-1.39346913e-01 -2.63147891e-01 7.65489638e-01 -3.46433371e-01
-1.08158183e+00 -9.08041835e-01 -4.32915874e-02 1.16617382e+00
-4.38729674e-01 4.70092058e-01 -7.68414378e-01 -1.78630248e-01
2.59159952e-01 1.01453531e+00 -4.73870128e-01 -1.90410778e-01
-4.04727131e-01 -3.49263489e-01 9.20987427e-02 5.35387397e-01
4.48108166e-01 -1.15657151e+00 -5.52094340e-01 4.85052317e-02
5.33982158e-01 -3.77610117e-01 9.44526345e-02 2.20522374e-01
-1.14920080e+00 -6.89726949e-01 -7.47583270e-01 -1.05347347e+00
1.14017797e+00 -8.24087337e-02 9.11848962e-01 4.54184502e-01
-2.57520713e-02 6.49375618e-01 1.86938256e-01 -8.44680488e-01
-6.02010727e-01 -2.54732430e-01 4.28986549e-01 -5.09145379e-01
6.38764799e-01 -8.27438712e-01 -3.16284090e-01 -2.63651032e-02
-6.73008800e-01 -1.58284546e-03 7.44193196e-01 7.39945531e-01
-2.07020104e-01 3.71245623e-01 8.78068805e-01 -8.02972078e-01
9.39661264e-01 -4.03940529e-01 -6.67104781e-01 -8.12503882e-03
-1.09419036e+00 -1.20050944e-01 5.46480358e-01 -8.72738004e-01
-1.19662738e+00 -1.07720286e-01 4.80625480e-02 -2.59700447e-01
-6.26745164e-01 7.46190786e-01 3.14540863e-01 -2.70411551e-01
6.59915864e-01 1.95345461e-01 2.43558437e-01 -7.08319604e-01
-3.24564368e-01 5.71668744e-01 5.54782450e-01 -6.54805422e-01
8.44547033e-01 -2.87519008e-01 -2.49774843e-01 -8.97650301e-01
-7.11660743e-01 4.99155283e-01 -4.48335677e-01 -3.56436700e-01
4.48268831e-01 -1.06547642e+00 -8.97676349e-01 8.04357171e-01
-7.98145711e-01 -6.45815969e-01 -1.85285911e-01 9.86275911e-01
-2.35722974e-01 -4.70071472e-02 -7.77860940e-01 -1.02342486e+00
2.02570483e-01 -9.52122927e-01 -2.52726793e-01 1.02002120e+00
-5.81183657e-02 -1.37737858e+00 -1.16444893e-01 3.09187561e-01
2.55887836e-01 -1.71477795e-01 1.36350155e+00 -1.04349363e+00
-5.23881555e-01 -8.22621882e-02 7.78884292e-02 6.29180431e-01
-1.79280967e-01 1.63349912e-01 -6.99832857e-01 -3.02745700e-01
4.05328095e-01 -8.08745325e-01 7.32474923e-01 5.04888296e-01
1.06811500e+00 -4.58592504e-01 4.12489995e-02 3.15625846e-01
1.29866719e+00 4.74324167e-01 5.52818477e-01 2.79432982e-01
4.65863496e-01 9.39984739e-01 2.67662052e-02 2.39478990e-01
4.34041560e-01 -1.97549745e-01 3.99852768e-02 5.81264636e-03
3.44377041e-01 -4.63368595e-01 4.11735624e-01 1.09382975e+00
-7.03734756e-02 1.64087787e-01 -1.20173323e+00 5.72342813e-01
-1.48748875e+00 -8.72881711e-01 -6.67324141e-02 2.17891383e+00
9.57763791e-01 3.37029308e-01 1.31457046e-01 4.21015061e-02
4.78987157e-01 -4.17248636e-01 -4.91551220e-01 -6.86362803e-01
3.19091678e-01 1.47517994e-01 3.15042913e-01 6.84964418e-01
-3.82139087e-01 8.10460865e-01 7.43371820e+00 3.57060701e-01
-1.14976084e+00 -2.34231338e-01 1.01601100e+00 1.22044690e-01
-2.34210476e-01 3.72101134e-03 -6.07198715e-01 2.86336005e-01
1.14974022e+00 -2.90138751e-01 5.10123789e-01 6.20018184e-01
5.20291328e-01 -3.76851290e-01 -1.10898054e+00 6.22467339e-01
-1.76510960e-01 -7.11868286e-01 -1.75997183e-01 1.66736960e-01
7.91693628e-01 -4.87903446e-01 5.45488596e-01 4.00457531e-01
6.18229747e-01 -1.39709270e+00 3.87434602e-01 6.89538598e-01
1.91739425e-01 -8.97633076e-01 4.41297054e-01 9.16985810e-01
2.67599546e-03 -4.99426842e-01 -5.55902302e-01 -9.26729798e-01
-6.90113664e-01 3.24878931e-01 -6.47764981e-01 -2.11315513e-01
3.76267016e-01 3.75722647e-01 -6.34490848e-01 8.67505610e-01
-3.31568003e-01 9.45882797e-01 -9.82222240e-03 -7.07097277e-02
1.88149974e-01 -4.30617392e-01 2.09686980e-01 7.82276452e-01
1.73507929e-01 6.64854527e-01 -3.63241695e-02 1.09785974e+00
3.29556540e-02 5.78962341e-02 -5.61937332e-01 -5.62310815e-01
4.37864244e-01 9.79404569e-01 -4.94811386e-01 -4.12654728e-01
-3.39871138e-01 5.21254659e-01 4.63034779e-01 5.86128771e-01
-5.99667989e-02 -5.15008032e-01 4.86284047e-01 1.59816325e-01
1.01172663e-01 -1.26847982e-01 -5.67518413e-01 -8.98662806e-01
-5.22079587e-01 -1.07687569e+00 -1.72803044e-01 -9.15834606e-01
-1.08556807e+00 -1.24292262e-01 -7.20020086e-02 -2.25921720e-01
-3.93894285e-01 -8.60500038e-01 -9.01283801e-01 1.14508522e+00
-1.05474472e+00 -2.62084901e-01 -1.88562516e-02 5.18807352e-01
2.02580124e-01 -3.96936655e-01 5.18186927e-01 2.45640457e-01
-9.64933395e-01 5.39525628e-01 2.96020240e-01 5.26530981e-01
3.61036271e-01 -1.01714444e+00 -8.57728720e-02 5.22005141e-01
-3.29620898e-01 7.81286776e-01 8.96746457e-01 -7.78650105e-01
-1.15317476e+00 -4.29306388e-01 1.10615778e+00 -2.39182666e-01
5.65929651e-01 2.13019345e-02 -1.22972810e+00 8.65282953e-01
3.22788328e-01 -6.87369227e-01 5.20587683e-01 -5.60543612e-02
-2.31545288e-02 -3.74340899e-02 -1.17734766e+00 5.66793084e-01
3.66210163e-01 -3.58021826e-01 -9.19461608e-01 2.05364704e-01
1.91510081e-01 -2.22081214e-01 -9.66752708e-01 -1.14431068e-01
6.56805754e-01 -1.19913363e+00 8.47904682e-01 -9.14236963e-01
7.35217571e-01 5.51603675e-01 4.33022171e-01 -1.54362726e+00
-4.55760539e-01 -2.59741992e-01 2.51413941e-01 1.06420505e+00
3.15755963e-01 -7.86383569e-01 7.85671711e-01 1.01767135e+00
2.53573388e-01 -7.79507637e-01 -3.49880755e-01 -6.31563604e-01
7.70883381e-01 5.92972822e-02 9.66627896e-02 1.24320376e+00
7.08810538e-02 2.15822101e-01 -3.81144807e-02 1.89236179e-01
8.62610042e-01 -1.87351570e-01 4.35193777e-01 -1.67362237e+00
-3.09775263e-01 -5.09878516e-01 -7.16205686e-02 -9.39486325e-01
2.81972259e-01 -5.20305455e-01 -2.47741446e-01 -1.22417533e+00
1.63590148e-01 -3.28284085e-01 -2.52250731e-01 3.89276087e-01
-2.34680623e-01 -1.74653918e-01 1.70415059e-01 1.18134229e-03
2.59170029e-02 1.98337108e-01 1.39559388e+00 4.51109231e-01
-3.91641587e-01 2.96089202e-01 -1.19753742e+00 1.31706309e+00
8.44411492e-01 -6.05820835e-01 -3.40625018e-01 -2.31500149e-01
3.57879788e-01 5.06479442e-01 3.39009613e-01 -7.03720570e-01
2.51399189e-01 -6.26558125e-01 1.09870493e+00 5.35849966e-02
7.08155930e-02 -9.48158562e-01 -1.19950540e-01 8.04444611e-01
-6.32122517e-01 6.09176420e-02 3.33830826e-02 1.97031423e-02
1.12717591e-01 -8.07590365e-01 1.13639426e+00 -4.17620391e-01
-1.42332599e-01 -1.54909387e-01 -7.88408160e-01 1.46515638e-01
6.56681359e-01 2.34589968e-02 -3.57473403e-01 -9.84684587e-01
-6.92222774e-01 2.11517021e-01 3.58995765e-01 6.04097582e-02
7.75766909e-01 -1.24338508e+00 -6.82892501e-01 3.12813073e-01
-7.54858673e-01 -4.24129307e-01 -3.17825079e-01 8.65872443e-01
-3.91130865e-01 3.40565413e-01 -4.97250468e-01 5.05583249e-02
-1.08067846e+00 3.91672462e-01 4.62913781e-01 2.67668366e-01
-2.02449039e-01 8.70134354e-01 2.76725620e-01 -3.60403180e-01
5.79138160e-01 -4.45324071e-02 -1.57629415e-01 -4.81756926e-02
5.17752051e-01 5.20550966e-01 -6.29928410e-01 -2.82252491e-01
1.69557869e-01 3.17629904e-01 -2.58723140e-01 3.14815380e-02
1.39319766e+00 -7.18372539e-02 -5.69305979e-02 7.90361047e-01
7.49628782e-01 -2.35703811e-01 -1.20026648e+00 -2.25297809e-01
-9.56300423e-02 -6.16963089e-01 1.86956912e-01 -8.83188128e-01
-8.82229984e-01 1.02337503e+00 4.82020468e-01 1.58265322e-01
8.56804550e-01 -4.42969233e-01 7.43874609e-02 7.37649083e-01
-1.21722773e-01 -1.31474209e+00 4.30579156e-01 5.34078836e-01
4.79632735e-01 -1.21389019e+00 1.01039603e-01 1.13211632e-01
-6.80578470e-01 9.87422824e-01 9.06902015e-01 -3.78624588e-01
6.48755848e-01 6.78448305e-02 2.72213310e-01 -9.63471904e-02
-8.96698713e-01 2.92543828e-01 2.00840265e-01 4.46974188e-01
8.43679011e-01 -3.98732536e-02 -3.54854673e-01 5.52432358e-01
-2.22738400e-01 2.00703382e-01 8.76799345e-01 6.92261875e-01
-9.67727959e-01 -9.57173049e-01 -6.70850933e-01 4.71417069e-01
-7.64657795e-01 1.41866347e-02 -4.66921538e-01 7.67660499e-01
6.22324608e-02 6.90226972e-01 2.66550779e-01 -2.84512136e-02
2.08632443e-02 5.98686218e-01 6.06451869e-01 -4.75069731e-01
-8.55061769e-01 3.46142873e-02 -2.58382469e-01 -7.39898114e-03
-2.18945086e-01 -5.99154472e-01 -9.85472560e-01 -6.44172370e-01
-9.22803432e-02 1.70136973e-01 5.92382371e-01 8.52740645e-01
-2.15176120e-01 3.65773231e-01 4.77185547e-01 -5.90167224e-01
-1.13634956e+00 -9.90033150e-01 -1.04408634e+00 1.31079480e-01
5.11281013e-01 -2.92260796e-01 -6.48825169e-01 -9.84484702e-02] | [10.197188377380371, 7.800768852233887] |
8c8cd65a-bbe7-4566-8f93-5cf216c6e4d4 | event-based-human-pose-tracking-by-spiking | 2303.09681 | null | https://arxiv.org/abs/2303.09681v3 | https://arxiv.org/pdf/2303.09681v3.pdf | Event-based Human Pose Tracking by Spiking Spatiotemporal Transformer | Event camera, as an emerging biologically-inspired vision sensor for capturing motion dynamics, presents new potential for 3D human pose tracking, or video-based 3D human pose estimation. However, existing works in pose tracking either require the presence of additional gray-scale images to establish a solid starting pose, or ignore the temporal dependencies all together by collapsing segments of event streams to form static event frames. Meanwhile, although the effectiveness of Artificial Neural Networks (ANNs, a.k.a. dense deep learning) has been showcased in many event-based tasks, the use of ANNs tends to neglect the fact that compared to the dense frame-based image sequences, the occurrence of events from an event camera is spatiotemporally much sparser. Motivated by the above mentioned issues, we present in this paper a dedicated end-to-end sparse deep learning approach for event-based pose tracking: 1) to our knowledge this is the first time that 3D human pose tracking is obtained from events only, thus eliminating the need of accessing to any frame-based images as part of input; 2) our approach is based entirely upon the framework of Spiking Neural Networks (SNNs), which consists of Spike-Element-Wise (SEW) ResNet and a novel Spiking Spatiotemporal Transformer; 3) a large-scale synthetic dataset is constructed that features a broad and diverse set of annotated 3D human motions, as well as longer hours of event stream data, named SynEventHPD. Empirical experiments demonstrate that, with superior performance over the state-of-the-art (SOTA) ANNs counterparts, our approach also achieves a significant computation reduction of 80% in FLOPS. Furthermore, our proposed method also outperforms SOTA SNNs in the regression task of human pose tracking. Our implementation is available at https://github.com/JimmyZou/HumanPoseTracking_SNN and dataset will be released upon paper acceptance. | ['Li Cheng', 'Sen Wang', 'Xinxin Zuo', 'Yuxuan Mu', 'Shihao Zou'] | 2023-03-16 | null | null | null | null | ['pose-tracking', '3d-human-pose-tracking', '3d-human-pose-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 1.91701874e-01 -3.55238765e-01 2.15890035e-01 -1.26057984e-02
-5.12513340e-01 -2.38040075e-01 4.45532650e-01 -1.00920960e-01
-7.17061639e-01 7.58202493e-01 1.61640614e-01 3.90835851e-01
3.69464867e-02 -6.38701499e-01 -9.55918074e-01 -6.38831317e-01
-3.04183334e-01 4.79735941e-01 5.21953344e-01 -1.48931354e-01
-1.96285143e-01 7.22592235e-01 -1.74189818e+00 5.88940307e-02
2.14795873e-01 1.20139337e+00 1.48645267e-02 6.18208587e-01
2.93782473e-01 8.77760649e-01 -5.61683655e-01 -6.05275333e-02
3.53327125e-01 -5.01894414e-01 -3.90578270e-01 4.06505093e-02
3.56818825e-01 -4.21928972e-01 -7.91267216e-01 7.63159513e-01
7.29996145e-01 9.87328589e-02 2.50506192e-01 -1.45158517e+00
-1.94902629e-01 1.81870788e-01 -5.01438260e-01 3.70838612e-01
5.87105691e-01 3.98362845e-01 4.23679918e-01 -9.05623019e-01
1.02156997e+00 9.74127829e-01 8.88645291e-01 5.79583943e-01
-9.53646481e-01 -6.99806452e-01 -3.20533058e-04 2.60485351e-01
-1.39563942e+00 -2.60832638e-01 8.74302149e-01 -4.34700757e-01
1.19508183e+00 1.08899558e-02 1.44697833e+00 1.73504412e+00
2.64075220e-01 9.19704020e-01 9.32266712e-01 -1.33642539e-01
3.98388147e-01 -5.92035174e-01 1.19212139e-02 7.18731105e-01
2.71983266e-01 4.53236610e-01 -9.38071907e-01 -5.23028485e-02
1.00832045e+00 3.10912997e-01 -3.34133595e-01 -5.05937457e-01
-1.59695029e+00 4.83418733e-01 4.58405375e-01 2.21770853e-01
-7.26772130e-01 5.73435664e-01 4.62106019e-01 3.56683619e-02
2.12707862e-01 -8.09521452e-02 -3.64850342e-01 -3.56692284e-01
-1.02924156e+00 5.94486177e-01 7.13713765e-01 6.94207132e-01
5.12444913e-01 3.84525329e-01 -1.17422938e-01 2.90439814e-01
7.86285568e-03 7.42017746e-01 5.26920557e-01 -9.37796891e-01
-1.74260698e-02 4.06660497e-01 1.87932283e-01 -1.03720284e+00
-7.18018949e-01 -3.45465899e-01 -8.36536407e-01 1.62103280e-01
7.19011366e-01 -9.47932079e-02 -9.75812495e-01 1.89595807e+00
3.21873605e-01 5.67551315e-01 -2.14741722e-01 1.18206799e+00
7.09897161e-01 6.19761527e-01 1.93886571e-02 -2.70008624e-01
1.43354559e+00 -5.01874030e-01 -6.27171874e-01 -1.14503644e-01
1.09065056e-01 -3.48868579e-01 6.06776059e-01 4.26505357e-01
-1.16770005e+00 -5.14672220e-01 -8.89295936e-01 5.46373650e-02
-3.62966925e-01 -1.40670672e-01 7.79627323e-01 2.58106440e-01
-9.76168513e-01 3.76178890e-01 -1.37770462e+00 -7.35439599e-01
4.78656381e-01 4.20210451e-01 -4.34514433e-01 1.17027327e-01
-1.28096998e+00 7.86932588e-01 3.53614569e-01 1.62677109e-01
-1.05347121e+00 -5.78611612e-01 -9.70963180e-01 -2.04029322e-01
4.40830976e-01 -9.31296110e-01 1.00841188e+00 -7.78903842e-01
-1.23613775e+00 9.65876520e-01 -2.64685154e-01 -9.15897369e-01
5.68879247e-01 -4.00937855e-01 -1.55439243e-01 3.53755087e-01
-2.71749664e-02 9.76620018e-01 7.15637803e-01 -9.14740860e-01
-5.25757313e-01 -4.34781522e-01 -5.81279919e-02 -6.68901112e-03
-3.11296910e-01 9.87470895e-02 -7.04824209e-01 -8.60613167e-01
-5.10772355e-02 -1.10624194e+00 -2.40861133e-01 2.69611388e-01
-2.43367195e-01 -4.13462445e-02 6.91012919e-01 -5.57260871e-01
1.05344272e+00 -1.92265987e+00 3.25895131e-01 -8.25100690e-02
4.02513623e-01 1.64712206e-01 8.12758803e-02 3.61923575e-01
-2.21626312e-02 -5.71477354e-01 -4.13784385e-01 -3.41093421e-01
-1.31196603e-01 7.61258006e-02 -1.10054389e-01 6.67120874e-01
2.16964111e-01 1.16213012e+00 -8.39772165e-01 -5.71753263e-01
5.16965389e-01 8.09971452e-01 -3.06624323e-01 -1.58555973e-02
-7.23841339e-02 6.60248160e-01 -2.28817016e-01 9.05864894e-01
2.46567518e-01 -4.51585174e-01 -2.37517238e-01 -3.06670874e-01
-2.16636911e-01 -1.33549228e-01 -1.17122734e+00 2.18438268e+00
1.30223244e-01 4.66139466e-01 -2.88367778e-01 -9.19302762e-01
7.10960686e-01 3.67318839e-01 1.08269322e+00 -7.15001881e-01
3.35610390e-01 8.41449797e-02 -2.22865254e-01 -3.44520539e-01
2.96747208e-01 1.11244455e-01 -3.49039197e-01 2.43805990e-01
2.84567475e-01 2.97446370e-01 2.76371151e-01 7.50910118e-02
1.49730814e+00 5.52789569e-01 4.31635857e-01 1.38026550e-01
1.61691621e-01 2.39080966e-01 7.60844111e-01 6.61203146e-01
-6.09972119e-01 7.64996529e-01 1.34522453e-01 -7.81544268e-01
-1.16253829e+00 -1.24447811e+00 -3.80461328e-02 6.00676119e-01
3.01003963e-01 -2.37580881e-01 -7.43192792e-01 -2.11856335e-01
2.67282012e-03 1.94908485e-01 -6.16793156e-01 -3.56895849e-02
-9.47945237e-01 -7.32244372e-01 8.06034744e-01 7.21284330e-01
4.84251767e-01 -1.42248046e+00 -1.37426138e+00 4.90621060e-01
-2.56736010e-01 -1.38849390e+00 -2.72629678e-01 2.53199309e-01
-7.04713285e-01 -1.08805478e+00 -9.92365360e-01 -6.51944876e-01
3.60438585e-01 -7.71407187e-02 8.83690596e-01 -2.99468011e-01
-6.96019411e-01 6.31411433e-01 -2.06362247e-01 -4.68192875e-01
1.81765735e-01 -2.62551755e-01 1.69302762e-01 -4.72139306e-02
5.46216846e-01 -7.82037675e-01 -7.74751961e-01 5.57021610e-02
-7.73308277e-01 1.95692971e-01 3.78313035e-01 7.68426955e-01
8.82299185e-01 -2.68582284e-01 4.26031739e-01 -3.75919670e-01
1.14044212e-01 -3.23507935e-01 -6.14548445e-01 -1.03447326e-01
2.18197964e-02 -3.07902843e-01 3.78075153e-01 -6.83195412e-01
-7.33112633e-01 6.52416945e-01 -1.27323121e-01 -9.44813132e-01
-4.34090495e-01 2.60628790e-01 2.37096131e-01 5.91806322e-03
6.96904004e-01 4.39289302e-01 9.66194123e-02 -2.55952943e-02
4.11302336e-02 -5.12998726e-04 9.34699774e-01 -2.79290497e-01
7.94600308e-01 9.33291793e-01 1.48012787e-01 -7.21896291e-01
-5.82876682e-01 -4.04910833e-01 -6.20648742e-01 -6.01777256e-01
1.15797555e+00 -1.21754718e+00 -9.28501725e-01 9.24795449e-01
-1.24030721e+00 -2.85979658e-01 -4.09390926e-01 6.59414053e-01
-7.91148484e-01 1.88045427e-01 -9.40258265e-01 -8.68505359e-01
-2.89448500e-01 -9.04651165e-01 1.27468038e+00 1.52562201e-01
-5.12694955e-01 -6.53481424e-01 1.35982260e-01 1.04147360e-01
1.96107224e-01 1.00510085e+00 1.75051063e-01 -4.22048181e-01
-7.49230146e-01 -4.45516676e-01 1.05341874e-01 -1.98351577e-01
-1.05943352e-01 -3.21153760e-01 -9.58979607e-01 -3.07797670e-01
1.28015220e-01 -4.44699675e-01 7.88110077e-01 8.31797242e-01
9.66065526e-01 2.06161693e-01 -4.17921782e-01 7.12204158e-01
1.24245334e+00 2.03106776e-01 6.34030581e-01 3.42107356e-01
8.41287673e-01 4.16459233e-01 4.46989596e-01 7.92282164e-01
4.55584079e-01 7.94549525e-01 5.45654118e-01 -1.98553130e-01
-3.07389796e-01 -1.79361507e-01 4.43609208e-01 6.65006340e-01
-3.15546274e-01 -3.17767680e-01 -8.97016764e-01 6.55400515e-01
-2.02902818e+00 -1.17533064e+00 8.36685300e-03 2.18735099e+00
5.59515059e-01 1.29497156e-01 4.43300486e-01 3.55231017e-01
8.12826991e-01 1.49258733e-01 -7.74738371e-01 3.72632146e-01
-3.38664800e-01 4.23199087e-01 4.40309882e-01 2.47629751e-02
-1.30544209e+00 7.85453141e-01 4.98407364e+00 5.44024765e-01
-1.20142519e+00 1.45446271e-01 2.71495223e-01 -6.12013042e-01
2.39835307e-01 -2.51256377e-01 -8.34646821e-01 5.57021499e-01
8.55548978e-01 -2.33958173e-03 3.37081462e-01 4.80381817e-01
3.80467802e-01 -6.94572106e-02 -1.09778726e+00 1.28897262e+00
1.51870355e-01 -1.27431619e+00 -1.61668718e-01 -5.95998578e-02
4.29084003e-01 1.51887715e-01 -2.24077195e-01 1.19168326e-01
5.68727292e-02 -8.80113423e-01 1.12182629e+00 6.90878510e-01
6.54893160e-01 -7.48656332e-01 5.04011929e-01 3.41329575e-01
-1.45121861e+00 1.85305867e-02 -5.13588935e-02 -1.18901856e-01
6.66680038e-01 6.45560682e-01 -2.69646645e-01 3.84023666e-01
1.13414299e+00 9.38804030e-01 -3.54504555e-01 1.16008127e+00
5.09996638e-02 5.23871422e-01 -6.54211879e-01 -7.49119371e-02
7.08867535e-02 3.28426749e-01 8.31911922e-01 1.12307703e+00
4.15243447e-01 1.48665801e-01 3.06035340e-01 8.61079216e-01
9.37303230e-02 -3.55259120e-01 -7.35875666e-01 1.23244837e-01
4.23193336e-01 1.04066014e+00 -9.87671077e-01 -2.00456396e-01
-3.82794857e-01 1.12536848e+00 8.95447731e-02 3.02294165e-01
-1.14585698e+00 -2.01371849e-01 4.93337482e-01 2.10666835e-01
5.65356791e-01 -3.15678895e-01 -6.17165007e-02 -1.23461831e+00
2.02765241e-01 -6.97004735e-01 4.20807958e-01 -8.46306086e-01
-1.20363963e+00 5.72259188e-01 9.30168256e-02 -1.37302983e+00
-3.68730366e-01 -4.09559339e-01 -2.80040354e-01 3.85191858e-01
-1.24922454e+00 -1.21647799e+00 -5.12326479e-01 1.01139486e+00
5.24124861e-01 6.22728094e-02 7.02424228e-01 4.36436146e-01
-4.58683282e-01 3.81588072e-01 -3.55501354e-01 3.51225972e-01
5.74043572e-01 -8.53502393e-01 4.76884127e-01 9.45619524e-01
2.25797877e-01 3.53506029e-01 6.49241865e-01 -8.11928153e-01
-1.73205996e+00 -1.10930443e+00 7.77183771e-01 -5.81058681e-01
6.21688783e-01 -4.82541829e-01 -6.79161251e-01 6.93538010e-01
-4.59826225e-03 3.82210910e-01 2.74295926e-01 -3.60399187e-01
1.82372332e-02 4.23468575e-02 -9.19718027e-01 5.22314370e-01
1.50772965e+00 -3.76454532e-01 -6.23213291e-01 1.84762046e-01
5.98027587e-01 -7.66455650e-01 -8.11843693e-01 6.44926488e-01
6.97696269e-01 -9.76115584e-01 1.23411620e+00 -2.61543423e-01
3.63232821e-01 -6.18643939e-01 -4.43935432e-02 -6.87805176e-01
-1.86134547e-01 -4.16499346e-01 -6.03309691e-01 7.02210009e-01
-1.01527035e-01 -3.75459313e-01 9.97584283e-01 1.79758146e-01
-7.93329179e-02 -7.45744765e-01 -1.21301198e+00 -8.51805806e-01
-3.88625205e-01 -6.10045850e-01 1.71803281e-01 7.71951437e-01
-3.80869836e-01 5.95691521e-03 -6.63872778e-01 2.09957361e-01
8.81956398e-01 9.33765806e-03 6.57438040e-01 -1.27558815e+00
-2.20701933e-01 -1.70730382e-01 -8.07286322e-01 -8.99662673e-01
1.66459177e-02 -6.71228647e-01 1.81539789e-01 -1.19076920e+00
2.01624379e-01 8.95021483e-02 -3.78963679e-01 5.03650129e-01
3.77952270e-02 5.07847369e-01 2.18203872e-01 2.45187625e-01
-8.77951801e-01 5.22766829e-01 9.94720519e-01 9.48880836e-02
-9.70106125e-02 -3.11210781e-01 8.38989206e-03 8.61960411e-01
5.16114414e-01 -5.42925119e-01 -1.82344601e-01 -1.98412076e-01
4.72582430e-02 3.94797802e-01 1.09004557e+00 -1.51389945e+00
5.70966840e-01 2.10856959e-01 7.17323184e-01 -8.69481325e-01
6.90707803e-01 -8.12536180e-01 5.46105325e-01 7.93215990e-01
-8.94906670e-02 1.80933684e-01 3.22135270e-01 7.72362947e-01
-1.68450326e-01 2.76946992e-01 5.97099602e-01 -3.36512983e-01
-1.01101172e+00 6.12414181e-01 -5.15428185e-01 5.49504012e-02
1.13165951e+00 -5.60105205e-01 7.50530977e-03 -2.50331670e-01
-7.43094087e-01 -7.91075453e-03 5.47346115e-01 2.97241300e-01
6.61915362e-01 -1.41491842e+00 -5.42042553e-01 1.88618805e-02
1.27896965e-01 3.68695520e-02 3.64407331e-01 9.19521451e-01
-4.95201737e-01 3.83243829e-01 -5.31068742e-01 -9.96350050e-01
-9.40226436e-01 5.58169842e-01 1.59960091e-01 -2.29923069e-01
-1.05458426e+00 8.19117308e-01 1.89237460e-01 -1.33833498e-01
3.98880959e-01 -2.88036138e-01 4.70172381e-03 2.55351097e-05
4.08716530e-01 3.18775982e-01 -9.45895240e-02 -6.81879401e-01
-5.74610472e-01 6.20959520e-01 3.11568975e-01 -1.38444141e-01
1.44545627e+00 1.23461574e-01 1.90566540e-01 6.32764757e-01
8.47065508e-01 -3.91969055e-01 -1.62301087e+00 -1.02682635e-01
-1.58162504e-01 -6.76689520e-02 -2.40660086e-01 -5.84906042e-01
-1.11992574e+00 7.55829990e-01 8.12936902e-01 -1.99878722e-01
1.26912379e+00 -1.36919245e-01 1.31222510e+00 3.97698343e-01
7.50426173e-01 -8.71497333e-01 2.14493126e-01 4.55511838e-01
7.34867811e-01 -1.08629751e+00 -1.33382708e-01 -2.18850181e-01
-4.29827631e-01 9.06082034e-01 6.58937633e-01 -4.46687549e-01
4.93626744e-01 5.86572528e-01 -2.13387176e-01 -3.31498235e-01
-6.98988318e-01 -4.38013583e-01 -1.01533398e-01 6.30555093e-01
1.41995281e-01 -2.85276115e-01 -1.53141484e-01 5.30417323e-01
-6.91348463e-02 5.41719317e-01 1.78745300e-01 1.18818474e+00
-2.26700485e-01 -6.74120784e-01 -4.85834420e-01 2.72128493e-01
-3.99550766e-01 -9.57258418e-03 -1.12845823e-01 9.97322440e-01
2.55647898e-01 4.03297722e-01 1.31406114e-01 -2.72979647e-01
4.99793440e-01 1.18521377e-01 6.92164361e-01 -2.55940676e-01
-6.71642303e-01 1.22489929e-01 -1.54438540e-01 -9.08892035e-01
-7.58820653e-01 -8.49347234e-01 -1.50252056e+00 -3.14400852e-01
5.81763089e-02 -4.72202331e-01 3.81059766e-01 9.34609950e-01
3.48716676e-01 6.44237816e-01 2.23399438e-02 -1.26925266e+00
5.92339747e-02 -6.61633730e-01 -4.39253300e-01 6.12066567e-01
2.00287089e-01 -9.52274323e-01 -2.10530624e-01 4.14934874e-01] | [7.522775650024414, -0.7750266194343567] |
bf259bab-d189-4ffb-a580-28b510a5543a | learning-low-dimensional-temporal | null | null | https://icml.cc/Conferences/2018/Schedule?showEvent=1910 | http://proceedings.mlr.press/v80/su18a/su18a.pdf | Learning Low-Dimensional Temporal Representations |
Low-dimensional discriminative representations enhance machine learning methods in both performance and complexity, motivating supervised dimensionality reduction (DR) that transforms high-dimensional data to a discriminative subspace. Most DR methods require data to be i.i.d., however, in some domains, data naturally come in sequences, where the observations are temporally correlated. We propose a DR method called LT-LDA to learn low-dimensional temporal representations. We construct the separability among sequence classes by lifting the holistic temporal structures, which are established based on temporal alignments and may change in different subspaces. We jointly learn the subspace and the associated alignments by optimizing an objective which favors easily-separable temporal structures, and show that this objective is connected to the inference of alignments, thus allows an iterative solution. We provide both theoretical insight and empirical evaluation on real-world sequence datasets to show the interest of our method.
| ['Bing Su', 'Ying Wu'] | 2018-07-01 | null | null | null | icml-2018-7 | ['supervised-dimensionality-reduction'] | ['computer-vision'] | [ 2.40952253e-01 -4.28609163e-01 -5.18632829e-01 -2.52046704e-01
-6.18900061e-01 -8.70429397e-01 7.77996778e-01 -2.88018107e-01
-2.52555192e-01 5.11683941e-01 5.46542346e-01 -4.56151515e-02
-3.89251441e-01 -2.76903898e-01 -4.97227639e-01 -1.22782624e+00
-5.15198350e-01 4.17225242e-01 -2.47913405e-01 1.98344454e-01
-9.57551040e-03 6.32567823e-01 -1.28522921e+00 2.84574717e-01
4.94977683e-01 5.55803239e-01 7.70187825e-02 6.24960661e-01
9.44869742e-02 3.15057009e-01 -1.48191616e-01 2.73274690e-01
4.12003070e-01 -5.49558461e-01 -8.03186059e-01 3.67670834e-01
2.13618279e-01 -1.24715120e-01 -6.82760775e-01 7.99593687e-01
1.79744661e-01 3.97209078e-01 1.02002144e+00 -1.18047369e+00
-6.33248270e-01 8.16881005e-03 -5.92938423e-01 3.48856956e-01
7.92405605e-02 2.47663036e-02 1.17951596e+00 -8.39134753e-01
9.21654463e-01 1.20090282e+00 5.35467684e-01 5.18846273e-01
-1.95468664e+00 -1.62396893e-01 3.00898671e-01 4.35522437e-01
-1.08740437e+00 -2.53411025e-01 1.07427895e+00 -9.80942309e-01
7.83362985e-01 2.26260886e-01 5.82571268e-01 1.46444094e+00
3.56199890e-02 1.10079503e+00 8.90427530e-01 -3.31439734e-01
2.97972351e-01 -4.30290312e-01 5.71135640e-01 2.69752234e-01
5.93594089e-02 1.73363149e-01 -6.95406199e-01 -3.27974766e-01
6.04590356e-01 4.59324926e-01 -2.68968791e-01 -1.15050805e+00
-1.41063142e+00 8.30983400e-01 -4.95105758e-02 4.55970258e-01
-3.62667322e-01 -5.20932525e-02 5.79556882e-01 3.81908774e-01
4.72953588e-01 1.13812990e-01 -5.25869429e-01 -3.00470382e-01
-6.88997090e-01 1.10304974e-01 4.10952002e-01 7.64340401e-01
8.40835392e-01 -1.88837618e-01 5.67261130e-02 7.72390544e-01
-1.81878963e-03 3.58488530e-01 7.25327134e-01 -1.04249549e+00
4.11650211e-01 3.95757705e-01 1.63005963e-01 -8.81410956e-01
-5.03927827e-01 -2.09764779e-01 -1.02180517e+00 -3.05432230e-01
3.75409245e-01 1.01309679e-01 -6.34702742e-01 2.17193580e+00
2.18627632e-01 3.92643064e-01 1.25428155e-01 9.37567890e-01
-1.52796879e-01 8.59672725e-01 -1.27372742e-01 -7.69029200e-01
9.97355700e-01 -6.31900072e-01 -8.62923563e-01 1.00596435e-01
1.04990315e+00 -3.04583132e-01 1.16961217e+00 4.80203509e-01
-5.91466546e-01 -4.00427818e-01 -9.04212534e-01 -1.62399471e-01
-2.26952154e-02 2.56252974e-01 5.98085940e-01 2.61521131e-01
-7.29471207e-01 7.30189323e-01 -1.29835594e+00 -5.69380522e-01
6.05002306e-02 2.10245386e-01 -5.77616155e-01 3.50310355e-02
-9.63210106e-01 4.04249012e-01 3.40282559e-01 -4.30210494e-02
-1.07313728e+00 -4.02715594e-01 -8.57762933e-01 -2.30478898e-01
2.10629985e-01 -2.59618253e-01 8.71777773e-01 -6.63742781e-01
-1.22540736e+00 8.30473125e-01 -6.33779466e-01 -3.76915067e-01
9.63837206e-02 -5.45632124e-01 -3.18793654e-01 3.19370776e-01
-6.26117215e-02 2.14400634e-01 8.78010392e-01 -1.09242737e+00
-3.05409819e-01 -5.98730326e-01 -2.69441605e-01 6.62862733e-02
-7.86302805e-01 -2.64481574e-01 -4.55045253e-01 -9.02054965e-01
3.16226214e-01 -1.35156715e+00 -3.26923460e-01 -6.09090552e-02
-2.38931805e-01 -3.40490580e-01 1.09769201e+00 -7.32210219e-01
1.42457449e+00 -2.45472860e+00 8.52073133e-01 1.13350064e-01
7.99743533e-02 -2.49096695e-02 -2.71159947e-01 5.30755341e-01
-4.14829701e-01 -1.00637168e-01 -3.69355798e-01 -4.50046092e-01
-9.16827172e-02 6.34926438e-01 -7.46965408e-01 8.64282072e-01
6.09096438e-02 5.88066161e-01 -9.30577040e-01 -2.21541986e-01
1.16259046e-01 2.75282323e-01 -6.49213910e-01 2.69998521e-01
-1.04026228e-01 6.42615378e-01 -5.31916678e-01 1.82583526e-01
4.52226043e-01 -2.93527573e-01 7.44648695e-01 -2.59595096e-01
-9.47546288e-02 5.00363529e-01 -8.75422895e-01 2.06850290e+00
-1.99590638e-01 7.60195792e-01 -3.53920132e-01 -1.60051656e+00
7.69425452e-01 3.18325639e-01 1.00185072e+00 -2.90310174e-01
-3.15685481e-01 -1.23362131e-02 -2.27795303e-01 -5.79502642e-01
2.76473343e-01 -2.42562760e-02 -1.97039217e-01 6.22330725e-01
7.29710907e-02 3.65905762e-01 2.01283455e-01 2.48102576e-01
9.46327448e-01 5.34585416e-01 3.71904880e-01 -4.40332890e-01
3.82425755e-01 -7.16437921e-02 1.00905001e+00 3.53924662e-01
-2.38029778e-01 3.79115880e-01 7.71916628e-01 -6.47052050e-01
-1.26124454e+00 -1.19844401e+00 -1.92889303e-01 9.16496992e-01
1.51258007e-01 -5.67359328e-01 -4.91502494e-01 -8.88344824e-01
-1.14880495e-01 5.36747575e-01 -7.69616187e-01 -3.74085903e-01
-9.41588759e-01 -7.45730340e-01 5.94015568e-02 4.43264872e-01
-2.54449010e-01 -4.62266147e-01 -4.32741195e-01 1.36533409e-01
-3.16096395e-01 -1.00423503e+00 -7.87710249e-01 4.34778184e-01
-1.27641499e+00 -8.50764632e-01 -6.11851752e-01 -5.27642310e-01
5.90562582e-01 4.19290900e-01 7.75038719e-01 -6.38496995e-01
-3.77421737e-01 4.80879903e-01 -2.76757002e-01 2.37893835e-01
-1.01034328e-01 9.83865038e-02 6.95357919e-01 1.99716896e-01
6.62815273e-01 -1.04391468e+00 -4.66083527e-01 4.12086844e-01
-9.80575919e-01 -2.45509800e-02 3.21276218e-01 1.02016723e+00
8.23867202e-01 -2.25323722e-01 3.96285921e-01 -4.07338530e-01
1.58182144e-01 -5.92669010e-01 -4.98338014e-01 2.50394225e-01
-4.75545794e-01 6.69567168e-01 6.33094549e-01 -7.17488945e-01
-9.10618722e-01 2.75978297e-01 4.91244316e-01 -9.03979242e-01
-1.73258677e-01 4.25787240e-01 -2.17570767e-01 5.19567132e-01
5.99792004e-01 5.97359955e-01 8.79716873e-02 -9.33402419e-01
5.27463377e-01 4.88910973e-01 3.33361417e-01 -8.12509894e-01
8.08285117e-01 8.38094354e-01 9.90420058e-02 -9.91220295e-01
-8.10288846e-01 -6.52409494e-01 -1.12068784e+00 1.13086235e-02
6.80885494e-01 -7.89338112e-01 -4.81979817e-01 8.08235705e-02
-9.41901982e-01 -1.87515646e-01 -4.09331918e-01 8.87221396e-01
-9.29967284e-01 9.19634223e-01 -6.92517281e-01 -7.46621668e-01
1.38454556e-01 -7.10446477e-01 1.07160842e+00 -3.20481986e-01
-4.18691546e-01 -8.59685361e-01 6.98214412e-01 -5.32904491e-02
-2.25576162e-01 5.19760959e-02 1.13995969e+00 -6.50942564e-01
-5.02860308e-01 2.39799377e-02 2.62081653e-01 2.84548014e-01
4.70482469e-01 1.93769578e-02 -7.82042325e-01 -5.75662851e-01
1.54205352e-01 -1.41273305e-01 1.03146732e+00 4.83845919e-01
1.28202963e+00 -3.82392228e-01 -6.24344110e-01 6.18974388e-01
1.14479649e+00 3.26146483e-01 2.07303181e-01 6.03984334e-02
7.50009000e-01 8.81763995e-01 6.78555250e-01 6.70835435e-01
-1.19585730e-01 1.07201493e+00 2.16285624e-02 2.98554689e-01
1.84052408e-01 -2.04881534e-01 6.95762038e-01 1.18033969e+00
-5.44054173e-02 -4.67050225e-02 -7.53059447e-01 8.13147724e-01
-2.19547963e+00 -1.02419400e+00 9.27968770e-02 2.25923705e+00
9.99298096e-01 -2.95808434e-01 3.94720942e-01 5.80023974e-02
6.01531744e-01 3.36887300e-01 -7.73455560e-01 -1.12828635e-01
-2.70162076e-01 -1.56995595e-01 1.17947027e-01 2.93453693e-01
-1.21235299e+00 7.17794180e-01 6.58814716e+00 7.46282995e-01
-1.02667689e+00 -1.71341598e-02 3.06461900e-01 -3.92762065e-01
-3.81444246e-01 2.33048876e-03 -6.28115237e-01 3.73310000e-01
8.57802451e-01 -3.31479996e-01 3.17204118e-01 8.97044122e-01
4.40216571e-01 4.83165473e-01 -1.56689823e+00 1.09640503e+00
-1.64003387e-01 -1.34989929e+00 2.98558027e-01 4.07665759e-01
7.58994818e-01 -1.75349265e-01 1.05466038e-01 8.80973041e-02
1.86399505e-01 -5.25124788e-01 3.84382337e-01 5.07443368e-01
5.99898100e-01 -7.23293364e-01 7.85929486e-02 4.40509826e-01
-1.14662409e+00 -2.43930295e-02 -3.95077795e-01 8.97296146e-02
2.62652993e-01 7.12624311e-01 -5.15306592e-01 5.84459066e-01
4.91707712e-01 1.45729339e+00 -8.95907357e-02 4.96182621e-01
6.18599392e-02 6.21372759e-01 -2.74226457e-01 3.57159525e-01
2.58142680e-01 -8.46336007e-01 8.10425401e-01 1.18723524e+00
3.47387910e-01 1.49660766e-01 1.93611726e-01 5.87537408e-01
2.56386161e-01 -6.30824119e-02 -8.93559396e-01 -4.63254422e-01
2.56111115e-01 8.96579921e-01 -3.49084646e-01 -1.85856387e-01
-4.59037602e-01 1.13903248e+00 2.39892945e-01 6.68382645e-01
-5.53354800e-01 1.12785427e-02 1.36761045e+00 -2.63285935e-01
4.74114895e-01 -9.41496491e-01 -7.82906935e-02 -1.57494843e+00
2.40995929e-01 -8.72406244e-01 7.06519544e-01 -8.35180357e-02
-1.46222305e+00 3.34510058e-01 1.10015934e-02 -1.73888814e+00
-6.78938925e-01 -6.38307273e-01 -1.68501716e-02 6.22775316e-01
-1.09376752e+00 -7.73485899e-01 2.19726697e-01 7.51424968e-01
6.22870445e-01 -7.68639371e-02 9.03783262e-01 1.76013142e-01
-7.89836526e-01 4.17688549e-01 8.35194647e-01 2.02405211e-02
6.38970613e-01 -1.19025564e+00 3.22784901e-01 9.04812217e-01
4.50807959e-01 8.87109220e-01 6.34680212e-01 -5.05950212e-01
-1.70565271e+00 -9.54835474e-01 6.96789384e-01 -4.00841743e-01
8.81355643e-01 -5.98758280e-01 -1.05844426e+00 8.87750983e-01
-2.31926188e-01 4.01402228e-02 1.06234741e+00 2.88497567e-01
-7.77389884e-01 -6.35269806e-02 -6.13092542e-01 7.34065115e-01
1.38474798e+00 -1.08432055e+00 -7.38808095e-01 4.39130098e-01
7.40935266e-01 2.09067479e-01 -8.62131536e-01 3.66449624e-01
4.83954161e-01 -6.25938058e-01 9.49273169e-01 -1.30301166e+00
1.55197516e-01 -4.86797631e-01 -4.87719983e-01 -1.12990224e+00
-4.76894975e-01 -7.60951459e-01 -6.19379401e-01 8.69746804e-01
6.76065385e-02 -3.47339392e-01 8.42053354e-01 3.43586683e-01
9.45679694e-02 -4.44787204e-01 -9.87889349e-01 -1.50421381e+00
7.45096197e-03 -4.01019603e-01 2.07624376e-01 1.32115281e+00
2.85093874e-01 3.93875718e-01 -8.28426003e-01 3.05693895e-01
8.39069724e-01 6.33242607e-01 7.10398018e-01 -1.14531362e+00
-4.79066193e-01 -2.78411418e-01 -5.32040834e-01 -1.60338128e+00
4.13954288e-01 -7.14308679e-01 -9.10607278e-02 -7.43287325e-01
5.42297006e-01 -2.15397835e-01 -2.64343530e-01 3.37610215e-01
3.65755847e-03 -3.55597317e-01 -1.73669338e-01 7.81333506e-01
-5.19240856e-01 1.01943409e+00 8.91428769e-01 -5.94810471e-02
-3.14958066e-01 -2.20906153e-01 -1.74643397e-01 4.95152086e-01
5.48418939e-01 -3.99996579e-01 -5.84069490e-01 -3.07807475e-01
-4.03264724e-02 2.23115901e-03 6.19246997e-03 -5.56895316e-01
-1.61916628e-01 -4.97297078e-01 1.45031229e-01 -6.91161394e-01
5.19765079e-01 -7.39784598e-01 1.86284587e-01 5.44552445e-01
-6.55509293e-01 -2.01961085e-01 -1.04630589e-01 1.00637007e+00
-3.53681207e-01 1.48110375e-01 7.18359947e-01 4.46402639e-01
-6.81389987e-01 5.32977819e-01 -3.38991612e-01 -9.34886038e-02
9.97707903e-01 -7.40596652e-02 2.37077903e-02 -3.35454434e-01
-1.00712764e+00 4.01042476e-02 4.98585492e-01 4.49750721e-01
3.49094391e-01 -1.70584750e+00 -4.80067223e-01 2.05993548e-01
2.97780186e-01 -2.87567258e-01 3.31689507e-01 9.34374392e-01
1.11338586e-01 6.45534515e-01 -8.60639513e-02 -1.03951311e+00
-1.26260078e+00 1.08365154e+00 1.76562294e-01 -2.43627742e-01
-8.52525890e-01 3.95920366e-01 8.31242800e-01 -2.44991556e-01
2.05234140e-01 -2.37736076e-01 -1.32993817e-01 3.79817575e-01
4.56306934e-01 3.55363876e-01 -2.99252063e-01 -5.38926363e-01
-4.85935897e-01 6.65237188e-01 -2.51644135e-01 -2.33607292e-01
1.52529371e+00 -3.91099542e-01 -7.95366541e-02 8.61270368e-01
1.69613302e+00 -3.02339524e-01 -1.64991558e+00 -7.87464499e-01
5.23345351e-01 -5.25737584e-01 -2.73668975e-01 1.27389193e-01
-7.45257616e-01 9.94578779e-01 5.35731673e-01 2.44173899e-01
1.11348403e+00 1.77861258e-01 5.81554651e-01 6.85927451e-01
3.26546282e-01 -1.01733887e+00 3.48305672e-01 6.65769041e-01
7.91751146e-01 -7.83802032e-01 -2.15427175e-01 -2.06377894e-01
-5.30454338e-01 1.20872724e+00 1.72417790e-01 -1.40987098e-01
5.58284998e-01 -3.02772354e-02 -2.72808522e-01 -1.45449996e-01
-9.58839357e-01 -1.05986977e-02 3.50588351e-01 4.89113361e-01
3.76676559e-01 1.46100566e-01 -3.44999999e-01 4.88927394e-01
2.20114693e-01 -2.43322805e-01 2.01568171e-01 8.27446342e-01
-1.39764592e-01 -1.35202968e+00 -1.76192418e-01 1.21303707e-01
-2.01443970e-01 8.06090832e-02 -3.11528116e-01 5.46593070e-01
-2.48449713e-01 4.45089519e-01 2.02375159e-01 -3.46236229e-01
8.15263316e-02 3.71943444e-01 4.10590082e-01 -5.18339097e-01
3.22695732e-01 4.48165685e-01 -1.14290141e-01 -7.92110085e-01
-5.41109979e-01 -1.28669178e+00 -1.08991444e+00 -6.55438080e-02
1.46392703e-01 2.59582222e-01 3.18195045e-01 1.04837215e+00
5.40272295e-01 9.76665989e-02 1.06066203e+00 -8.82826388e-01
-7.58431613e-01 -7.00667918e-01 -8.21912766e-01 7.18727112e-01
6.59923017e-01 -6.68884039e-01 -5.55085421e-01 5.68306327e-01] | [7.825881481170654, 3.988450288772583] |
f9910a3e-d6e3-4061-bdd2-92bf4a285ad6 | fcgec-fine-grained-corpus-for-chinese | 2210.12364 | null | https://arxiv.org/abs/2210.12364v1 | https://arxiv.org/pdf/2210.12364v1.pdf | FCGEC: Fine-Grained Corpus for Chinese Grammatical Error Correction | Grammatical Error Correction (GEC) has been broadly applied in automatic correction and proofreading system recently. However, it is still immature in Chinese GEC due to limited high-quality data from native speakers in terms of category and scale. In this paper, we present FCGEC, a fine-grained corpus to detect, identify and correct the grammatical errors. FCGEC is a human-annotated corpus with multiple references, consisting of 41,340 sentences collected mainly from multi-choice questions in public school Chinese examinations. Furthermore, we propose a Switch-Tagger-Generator (STG) baseline model to correct the grammatical errors in low-resource settings. Compared to other GEC benchmark models, experimental results illustrate that STG outperforms them on our FCGEC. However, there exists a significant gap between benchmark models and humans that encourages future models to bridge it. | ['Ming Cai', 'Jiayu Fu', 'Jiawei Peng', 'Jianwang Wu', 'Lvxiaowei Xu'] | 2022-10-22 | null | null | null | null | ['grammatical-error-detection', 'grammatical-error-correction'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.22081959e-01 2.95957088e-01 1.68548256e-01 -4.02470738e-01
-1.26602292e+00 -3.06019545e-01 2.57910013e-01 4.50153023e-01
-4.40075636e-01 8.99165034e-01 4.46678191e-01 -5.76939702e-01
2.59287894e-01 -4.90298241e-01 -8.69350135e-01 -1.30608408e-02
5.85479915e-01 2.52726883e-01 3.84338409e-01 -5.19980073e-01
8.88400435e-01 -5.91493428e-01 -1.54825759e+00 3.57466906e-01
2.15818620e+00 -1.35956937e-02 6.36469841e-01 9.44815934e-01
-1.37430787e-01 6.67581439e-01 -9.01480794e-01 -1.13956010e+00
-5.84492683e-01 -8.20422173e-01 -1.06251073e+00 -6.33015096e-01
8.88764799e-01 1.47770032e-01 3.99060786e-01 1.71691823e+00
6.45889819e-01 -1.46026120e-01 5.62940121e-01 -7.58352518e-01
-1.52596009e+00 1.03186333e+00 -1.35476246e-01 2.07434639e-01
5.00252306e-01 -1.43084630e-01 9.48598921e-01 -1.16674030e+00
9.90712464e-01 1.41197562e+00 8.25300932e-01 1.38550580e+00
-4.09949273e-01 -6.35091245e-01 4.69995588e-02 4.17796731e-01
-9.99071300e-01 -1.86448351e-01 3.92035753e-01 -2.98523128e-01
1.07937515e+00 1.84121907e-01 3.50126415e-01 1.23190475e+00
2.19749972e-01 8.43915880e-01 1.16806412e+00 -1.21992433e+00
-1.54398337e-01 -7.72353336e-02 5.62609911e-01 8.17874551e-01
4.81589288e-01 -2.63888597e-01 -5.95046759e-01 2.63942182e-01
-8.27808399e-03 -6.04282618e-01 -3.83646339e-01 6.45662367e-01
-9.19241786e-01 7.27407515e-01 -3.15073371e-01 2.35020712e-01
1.53735667e-01 2.38827556e-01 3.66997153e-01 5.37408113e-01
5.61305940e-01 4.64166343e-01 -4.11512822e-01 -8.14500093e-01
-5.53847671e-01 6.56904101e-01 5.61343312e-01 1.58062911e+00
1.22207455e-01 -2.56217383e-02 -5.84180951e-01 1.05085909e+00
4.55480456e-01 9.19738770e-01 7.27222741e-01 -7.81216264e-01
1.13711190e+00 7.74984062e-01 -3.91087402e-03 -9.25085545e-01
-3.71297561e-02 -3.44930708e-01 -5.89575052e-01 -3.91421139e-01
4.78044987e-01 1.50426552e-01 -6.00102901e-01 1.57176185e+00
4.16251831e-02 4.97591272e-02 -1.26944527e-01 7.03624904e-01
1.20864809e+00 2.41885558e-01 4.84559447e-01 2.17835203e-01
1.21382010e+00 -1.09467816e+00 -1.08606040e+00 -2.75127113e-01
1.00128269e+00 -1.10003924e+00 1.53895056e+00 3.38003278e-01
-1.12575841e+00 -5.21008968e-01 -8.50727499e-01 -5.78877091e-01
-2.76496351e-01 4.75695074e-01 1.92447990e-01 1.03012121e+00
-7.47654974e-01 5.55526733e-01 -5.82820415e-01 -3.41954708e-01
1.04667850e-01 -5.19203007e-01 2.86944434e-02 -3.29491168e-01
-1.42099166e+00 8.73625517e-01 3.69121015e-01 -3.46995406e-02
-6.08566999e-01 -9.04007018e-01 -9.45642352e-01 -6.82379305e-03
1.29726112e-01 -2.18366593e-01 1.70558727e+00 -3.18284631e-01
-1.29067755e+00 1.14354587e+00 -1.98239386e-01 -2.78596431e-01
5.40644825e-01 -4.19232547e-01 -7.65716493e-01 -2.90818393e-01
4.84496266e-01 4.09952104e-01 4.13485676e-01 -6.05937958e-01
-9.95625198e-01 -1.26831576e-01 -4.01625812e-01 -1.64807942e-02
1.87616557e-01 6.18400097e-01 -2.81220555e-01 -9.27750170e-01
-2.95513384e-02 -7.74790764e-01 2.31695876e-01 -6.72963798e-01
-3.32341820e-01 -9.75540936e-01 1.10141747e-01 -1.09987676e+00
1.99731743e+00 -1.77903843e+00 -2.18434587e-01 -3.12504143e-01
-5.54145761e-02 7.56626308e-01 -1.98862210e-01 2.95431614e-01
9.71225798e-02 6.19965851e-01 -2.85319865e-01 -6.98896229e-01
9.95713770e-02 1.47066429e-01 -8.15002024e-02 -2.82321461e-02
2.22439051e-01 1.05333436e+00 -1.47957909e+00 -7.64480174e-01
-5.14146090e-01 -2.37944260e-01 -8.69717002e-01 2.57569994e-03
-1.41222104e-01 1.51497900e-01 -3.86210769e-01 1.02098441e+00
8.05070758e-01 9.97228473e-02 1.12793676e-03 9.17566776e-01
-1.53501704e-01 1.05050766e+00 -8.45812201e-01 1.80138397e+00
-1.24803349e-01 4.87457871e-01 -2.71697938e-01 -6.07576728e-01
1.02037227e+00 1.82399854e-01 -7.58326113e-01 -8.91359329e-01
1.32446483e-01 8.85728419e-01 9.56370234e-02 -7.99167633e-01
1.12190235e+00 3.43332291e-01 -4.04350102e-01 4.20021713e-01
1.98753014e-01 -8.38573650e-02 6.26956224e-01 4.60586756e-01
1.03943479e+00 3.11927609e-02 2.54675448e-01 -5.05598426e-01
7.53995299e-01 1.51741445e-01 1.04993832e+00 1.24830711e+00
-5.23185074e-01 1.01672506e+00 2.88275003e-01 1.87485307e-01
-6.74020350e-01 -6.03560209e-01 -1.97915323e-02 1.03695977e+00
-2.80634481e-02 -8.82518172e-01 -1.15911698e+00 -1.04184127e+00
-6.23055026e-02 9.72077966e-01 -4.28274989e-01 -1.61197469e-01
-9.34520483e-01 -3.97911668e-01 9.44801748e-01 3.19605500e-01
5.83208442e-01 -1.31558001e+00 -1.16861418e-01 5.10312557e-01
-8.52725625e-01 -8.88914287e-01 -9.24198210e-01 -7.36269414e-01
-7.10642636e-01 -1.37687457e+00 -2.18859270e-01 -1.02649283e+00
6.64273560e-01 2.70398676e-01 1.56858218e+00 8.92827034e-01
-3.43043990e-02 1.25407636e-01 -9.90890443e-01 -1.09015942e+00
-1.04038858e+00 1.57560855e-01 -2.50255793e-01 -8.67747843e-01
8.51836979e-01 3.71286161e-02 -5.44314533e-02 -2.21604079e-01
-2.52982289e-01 1.61440820e-01 2.84258038e-01 1.00002229e+00
3.83752853e-01 -6.63147807e-01 8.66210103e-01 -1.56627595e+00
1.00889504e+00 -3.21382433e-01 -4.53564525e-01 8.23460102e-01
-1.13341284e+00 -2.03718930e-01 3.14842492e-01 -1.26227692e-01
-1.32352853e+00 -7.75378644e-01 -3.34703386e-01 4.38827604e-01
1.02070838e-01 6.33637667e-01 -1.41099215e-01 4.06058952e-02
6.03812635e-01 1.76431328e-01 -2.59143889e-01 -9.17710960e-01
1.30005807e-01 9.67561543e-01 9.65801597e-01 -9.46260512e-01
2.40887657e-01 -7.86899745e-01 -6.38003707e-01 -1.56124204e-01
-1.12392235e+00 -1.32000610e-01 -4.58688170e-01 -3.08227450e-01
5.53155601e-01 -1.01164246e+00 -6.08766913e-01 9.29452956e-01
-1.57018077e+00 -1.14040807e-01 2.26825222e-01 4.54536229e-01
-6.56482130e-02 4.61222410e-01 -9.22427535e-01 -5.11850715e-01
-5.76644301e-01 -8.59319389e-01 9.08004522e-01 5.61145067e-01
-4.70278449e-02 -8.67930889e-01 1.43879697e-01 7.59760737e-01
2.92733580e-01 -2.16392532e-01 1.05729926e+00 -4.73686635e-01
-4.72577184e-01 -7.93722719e-02 -5.16376048e-02 4.63393569e-01
-3.35138470e-01 7.44314343e-02 -3.49973917e-01 -3.31007093e-02
-5.29221773e-01 -3.08750749e-01 8.10849786e-01 -1.32278195e-02
1.42149639e+00 -2.40068898e-01 -1.36018666e-02 2.10009411e-01
1.22117805e+00 4.74095484e-03 8.04868281e-01 3.28868270e-01
7.92241991e-01 3.49771440e-01 9.80768919e-01 -3.24832015e-02
1.05872500e+00 3.02397788e-01 5.65640777e-02 6.56502485e-01
-6.71895683e-01 -9.21768188e-01 4.15649146e-01 1.72026622e+00
-2.05431879e-01 -4.15086836e-01 -1.13599241e+00 1.00433433e+00
-1.77179062e+00 -8.65597129e-01 -1.05744219e+00 1.94445682e+00
1.42496288e+00 4.82234322e-02 -4.97087985e-01 -9.04183164e-02
1.03964233e+00 -4.42628473e-01 1.70316026e-01 -8.56247306e-01
-1.97953612e-01 4.60762441e-01 1.89765632e-01 6.59008503e-01
-8.48627746e-01 1.21865332e+00 6.20113325e+00 1.02893078e+00
-5.25479496e-01 5.97598314e-01 2.45236412e-01 4.60816979e-01
-5.86978436e-01 3.84732299e-02 -1.53222740e+00 1.13403141e+00
1.03193605e+00 -2.70245790e-01 8.56974046e-04 7.19356179e-01
4.83328290e-02 -1.90812618e-01 -7.11119235e-01 7.65284836e-01
5.84450066e-01 -1.26727748e+00 -2.00631917e-01 -3.84270698e-01
1.16712594e+00 -1.59935594e-01 -2.48767897e-01 9.71978426e-01
7.59924948e-01 -9.01482880e-01 1.02365685e+00 5.03698766e-01
8.25753450e-01 -4.79455382e-01 9.10412312e-01 5.87608397e-01
-3.93903792e-01 1.47163570e-01 -7.21922219e-01 -2.98642427e-01
-1.19340703e-01 5.38205147e-01 -5.09958148e-01 4.96231675e-01
9.11408067e-01 7.98130691e-01 -1.29218400e+00 1.07203805e+00
-1.13199902e+00 1.47997189e+00 3.77837718e-01 -5.39218128e-01
-7.82930329e-02 -1.33711204e-01 4.00963217e-01 1.47694910e+00
7.43785620e-01 1.58319876e-01 -2.77345300e-01 1.00864363e+00
-3.00209701e-01 2.03127801e-01 -1.44718084e-02 2.02664435e-01
1.20464051e+00 9.27131593e-01 -6.79852366e-02 -3.83114308e-01
-4.19487417e-01 8.37335706e-01 8.34062278e-01 -2.18857527e-02
-7.05794096e-01 -7.77982891e-01 3.43272924e-01 -1.27805814e-01
-5.16610965e-02 1.34380251e-01 -5.85017562e-01 -1.46168721e+00
1.83928281e-01 -1.09251487e+00 4.55103040e-01 -6.98947787e-01
-1.24141490e+00 1.09882019e-01 -2.95998603e-01 -1.29716051e+00
-2.73981482e-01 -5.14667153e-01 -8.19608331e-01 1.12106705e+00
-1.83780015e+00 -1.14263916e+00 -2.62582302e-01 3.40216696e-01
6.61549509e-01 -1.98415160e-01 5.91782808e-01 6.73359215e-01
-7.91229665e-01 1.37397516e+00 1.89123988e-01 3.80170107e-01
1.22885537e+00 -1.79532254e+00 7.42613137e-01 1.56574905e+00
-3.01058620e-01 8.86738062e-01 5.34148157e-01 -1.09090579e+00
-6.59900010e-01 -1.43700516e+00 2.31305194e+00 -8.31659198e-01
3.61521393e-01 -4.45633113e-01 -1.28931451e+00 5.79686522e-01
2.09155589e-01 -1.52129412e-01 3.44442487e-01 1.90540805e-01
-1.68559358e-01 1.84413582e-01 -1.06413066e+00 3.65703851e-01
1.31449568e+00 -4.61229742e-01 -8.99904728e-01 1.84686407e-01
7.53162563e-01 -1.04589427e+00 -6.13267183e-01 1.10189989e-01
3.14136297e-02 -5.64117968e-01 3.27455819e-01 -7.50850260e-01
1.10424590e+00 -1.77373946e-01 2.97151387e-01 -1.64271712e+00
-4.27948415e-01 -5.20197570e-01 1.15920760e-01 1.96552944e+00
5.42888463e-01 -5.11848211e-01 2.97086716e-01 5.43451190e-01
-1.07490301e+00 -3.50287825e-01 -8.69731665e-01 -7.64518619e-01
6.86370671e-01 -5.02903461e-01 9.22273457e-01 1.02341461e+00
1.62968740e-01 -5.09863831e-02 -2.59788215e-01 6.43023252e-02
4.53218788e-01 -7.65314624e-02 3.51016104e-01 -1.15330589e+00
1.14890710e-01 -2.21489325e-01 8.92003775e-02 -7.27241576e-01
2.95733511e-01 -1.20234859e+00 5.12089312e-01 -1.19348252e+00
3.61591130e-01 -6.39773726e-01 1.59837767e-01 4.92101729e-01
-1.27447391e+00 8.75534043e-02 1.22800298e-01 -3.22046690e-02
-6.88658953e-01 4.72742617e-01 1.32132816e+00 -1.42726889e-02
2.82929152e-01 -3.25035989e-01 -9.21508372e-01 4.77317005e-01
7.79723287e-01 -8.81437659e-01 1.07758924e-01 -8.90784323e-01
5.21761119e-01 -1.04811952e-01 1.52488619e-01 -6.20591223e-01
3.52151603e-01 -3.30211788e-01 -1.98778912e-01 -5.64379692e-01
-9.30895209e-01 1.61974058e-01 -3.81860167e-01 3.40978593e-01
-6.24914527e-01 4.36320931e-01 -2.84195393e-01 3.16510528e-01
-3.66251767e-01 -1.10325897e+00 5.93802392e-01 -3.20985049e-01
-8.40898693e-01 -2.54646480e-01 -5.49413145e-01 7.26236880e-01
2.99636841e-01 1.03557102e-01 -1.05303371e+00 1.85398888e-02
-2.96927411e-02 4.52105045e-01 1.67496964e-01 8.22881043e-01
5.08827507e-01 -1.47589278e+00 -1.19977593e+00 1.83242992e-01
6.03324771e-01 2.42219940e-02 3.47377032e-01 5.77777326e-01
-7.49787092e-01 5.84156096e-01 5.89885190e-02 -1.50838003e-01
-1.57271910e+00 -1.59064129e-01 6.14399128e-02 -3.73336017e-01
-4.27746087e-01 1.17044616e+00 -7.80721962e-01 -1.06646371e+00
4.93676886e-02 -5.45792758e-01 -4.77639019e-01 -4.75763947e-01
8.88205647e-01 5.54652870e-01 4.74062771e-01 -3.07640284e-01
-1.90106019e-01 2.05029801e-01 -2.73975641e-01 3.88229281e-01
8.86741877e-01 -4.21679825e-01 -3.85962307e-01 -3.68387327e-02
5.37450314e-01 3.70025992e-01 -5.22956967e-01 -2.70437807e-01
3.92545789e-01 -6.68316245e-01 -3.17233533e-01 -1.27230680e+00
-5.73046207e-01 8.90354335e-01 3.67292643e-01 -5.19908428e-01
7.85450220e-01 -2.33517602e-01 7.91461170e-01 2.65706331e-01
4.61972117e-01 -1.74848878e+00 -9.65260938e-02 1.25412083e+00
8.91738951e-01 -1.43091273e+00 -3.87394190e-01 -7.60883927e-01
-5.33679187e-01 1.01058352e+00 1.18208385e+00 -1.20310541e-02
2.88175553e-01 -3.42298180e-01 6.49086908e-02 -1.40159605e-02
-7.72218108e-01 6.98405178e-03 3.79853427e-01 6.33705437e-01
9.62942541e-01 3.73414725e-01 -1.47117329e+00 1.19281352e+00
-7.45177209e-01 -4.68921522e-03 1.19067180e+00 6.86057031e-01
-4.40040469e-01 -1.40351510e+00 -1.52051315e-01 2.13893056e-01
-8.23780835e-01 -4.68650162e-01 -2.05502525e-01 8.31958055e-01
4.74689722e-01 1.36873281e+00 -8.47937986e-02 2.36630384e-02
3.06300431e-01 1.57788023e-01 5.36688030e-01 -9.82632399e-01
-8.40184271e-01 -6.02258205e-01 3.38660866e-01 -2.67683417e-01
-2.05028057e-01 -9.22642529e-01 -1.11182952e+00 -3.86018991e-01
-4.67121869e-01 3.47870857e-01 4.95521665e-01 8.94908309e-01
3.46165389e-01 4.10546541e-01 1.06797382e-01 3.47494483e-01
-7.54535377e-01 -1.62750757e+00 -3.96955580e-01 6.42961979e-01
-2.14981027e-02 -4.16296571e-01 -1.25237539e-01 1.94245819e-02] | [11.060613632202148, 10.78318977355957] |
119aedc1-00fe-49fd-8042-97639e350ab2 | category-anchor-guided-unsupervised-domain | 1910.13049 | null | https://arxiv.org/abs/1910.13049v2 | https://arxiv.org/pdf/1910.13049v2.pdf | Category Anchor-Guided Unsupervised Domain Adaptation for Semantic Segmentation | Unsupervised domain adaptation (UDA) aims to enhance the generalization capability of a certain model from a source domain to a target domain. UDA is of particular significance since no extra effort is devoted to annotating target domain samples. However, the different data distributions in the two domains, or \emph{domain shift/discrepancy}, inevitably compromise the UDA performance. Although there has been a progress in matching the marginal distributions between two domains, the classifier favors the source domain features and makes incorrect predictions on the target domain due to category-agnostic feature alignment. In this paper, we propose a novel category anchor-guided (CAG) UDA model for semantic segmentation, which explicitly enforces category-aware feature alignment to learn shared discriminative features and classifiers simultaneously. First, the category-wise centroids of the source domain features are used as guided anchors to identify the active features in the target domain and also assign them pseudo-labels. Then, we leverage an anchor-based pixel-level distance loss and a discriminative loss to drive the intra-category features closer and the inter-category features further apart, respectively. Finally, we devise a stagewise training mechanism to reduce the error accumulation and adapt the proposed model progressively. Experiments on both the GTA5$\rightarrow $Cityscapes and SYNTHIA$\rightarrow $Cityscapes scenarios demonstrate the superiority of our CAG-UDA model over the state-of-the-art methods. The code is available at \url{https://github.com/RogerZhangzz/CAG_UDA}. | ['DaCheng Tao', 'Jing Zhang', 'Wei Liu', 'Qiming Zhang'] | 2019-10-29 | category-anchor-guided-unsupervised-domain-1 | http://papers.nips.cc/paper/8335-category-anchor-guided-unsupervised-domain-adaptation-for-semantic-segmentation | http://papers.nips.cc/paper/8335-category-anchor-guided-unsupervised-domain-adaptation-for-semantic-segmentation.pdf | neurips-2019-12 | ['synthetic-to-real-translation'] | ['computer-vision'] | [ 3.41118246e-01 1.01611987e-02 -3.04479957e-01 -7.63582051e-01
-1.00675428e+00 -5.82048714e-01 3.28086585e-01 4.01157737e-02
-3.58191013e-01 5.34974933e-01 -1.90018266e-01 -3.15919220e-02
-2.69892722e-01 -6.64565206e-01 -5.85741580e-01 -9.66280639e-01
2.57066220e-01 3.82997483e-01 5.10026276e-01 1.71641499e-01
1.99270815e-01 2.50675291e-01 -1.35874450e+00 1.90581307e-01
1.29481590e+00 1.12901497e+00 5.94054997e-01 1.04683213e-01
-1.36878967e-01 5.11713885e-02 -4.34343159e-01 -2.22227320e-01
4.11705047e-01 -4.51710910e-01 -6.97231233e-01 3.77036512e-01
4.43028927e-01 -7.75618702e-02 -9.95220020e-02 1.22939205e+00
3.65054756e-01 3.19112480e-01 8.41344833e-01 -1.27380443e+00
-5.92290163e-01 2.47665644e-01 -9.20609415e-01 2.42983088e-01
-1.40420437e-01 1.03274815e-01 1.08411443e+00 -8.37119699e-01
5.80503762e-01 9.19662893e-01 5.91880560e-01 5.64011693e-01
-1.18547499e+00 -9.03831124e-01 6.17495179e-01 2.40432665e-01
-1.48595297e+00 -2.20435187e-01 1.02778280e+00 -4.70452398e-01
3.05913836e-01 6.63664863e-02 4.01407480e-01 9.64208782e-01
-2.81722844e-01 9.80499387e-01 1.02996004e+00 -3.71876180e-01
3.65284383e-01 2.96518624e-01 1.93622395e-01 4.38984126e-01
9.79607329e-02 -1.66234598e-01 -4.41367000e-01 5.49603440e-02
7.07638741e-01 1.27890194e-02 -1.51908413e-01 -7.63436377e-01
-8.41799140e-01 8.23422074e-01 5.28790832e-01 2.85060674e-01
-2.93896854e-01 -2.46663213e-01 3.04725915e-01 -3.85201909e-02
4.90880489e-01 1.97260886e-01 -6.13765776e-01 7.59826005e-02
-8.38769555e-01 2.55965620e-01 5.31901717e-02 1.04052138e+00
8.93732846e-01 -2.49246404e-01 -9.96318981e-02 1.32294786e+00
3.63709867e-01 4.24605340e-01 5.88737071e-01 -1.01941514e+00
5.56897461e-01 8.02399516e-01 -8.69226009e-02 -8.55219245e-01
-2.03815922e-01 -5.16958594e-01 -5.21479130e-01 -8.97104759e-03
8.78550828e-01 6.11683130e-02 -1.07505929e+00 1.86545467e+00
6.01893008e-01 1.53837562e-01 -1.56634692e-02 1.03992391e+00
4.49244797e-01 4.60658252e-01 2.37547606e-01 6.99215382e-02
1.30183315e+00 -8.36855292e-01 -3.01370800e-01 -4.87924725e-01
5.66132367e-01 -6.51865304e-01 1.37251151e+00 9.52538028e-02
-7.39301622e-01 -7.02722311e-01 -9.40645218e-01 7.74758169e-03
-3.62732202e-01 3.66881579e-01 4.36859787e-01 5.27727842e-01
-4.60830450e-01 4.29475844e-01 -7.84610152e-01 -3.52478862e-01
8.80834043e-01 3.09248835e-01 -1.80380583e-01 -2.03259975e-01
-1.10860384e+00 3.16619962e-01 5.03623903e-01 -1.03320874e-01
-6.70769334e-01 -7.44314969e-01 -8.80339921e-01 -9.74313021e-02
2.88249791e-01 -3.27142328e-01 1.12480879e+00 -1.16294372e+00
-1.18700731e+00 1.07818794e+00 -1.72069922e-01 -3.05405349e-01
5.14419794e-01 2.29788907e-02 -3.44841510e-01 2.69134808e-02
5.63874722e-01 9.77369130e-01 8.54541183e-01 -1.25086653e+00
-1.18244648e+00 -5.98772109e-01 -1.70586616e-01 4.44642156e-01
-4.54135180e-01 -3.16003650e-01 -5.46684206e-01 -9.23688948e-01
5.73815942e-01 -8.38149250e-01 -1.21957988e-01 1.01896740e-01
-4.61823255e-01 -3.78940821e-01 7.59778321e-01 -4.89623308e-01
1.15518343e+00 -2.44885421e+00 -1.10818073e-01 3.55377287e-01
-1.67586450e-02 1.33685485e-01 -6.22078218e-02 -2.05077887e-01
-9.91840214e-02 -2.12484643e-01 -6.47033811e-01 -1.68081388e-01
-3.27940322e-02 2.62809724e-01 -2.44483784e-01 3.87149990e-01
1.85646832e-01 4.20971572e-01 -8.93483341e-01 -5.46564877e-01
1.07661732e-01 1.81421697e-01 -5.64175427e-01 6.67516068e-02
-1.75656497e-01 6.99281871e-01 -7.35283017e-01 7.85127640e-01
9.58300948e-01 -1.71910882e-01 1.69075266e-01 -9.16891247e-02
8.90558735e-02 2.34297246e-01 -1.24663651e+00 1.71581185e+00
-3.05590689e-01 3.40109527e-01 -7.10111931e-02 -1.30643499e+00
1.07785892e+00 -1.58417836e-01 4.76351231e-01 -7.31962681e-01
6.12362735e-02 5.32466412e-01 7.10193142e-02 -2.08702952e-01
3.05451006e-01 -4.51946706e-02 -2.73036182e-01 5.63147850e-02
9.46433749e-03 1.26793757e-02 2.54628863e-02 -1.28661647e-01
6.50113225e-01 3.70364875e-01 2.85399318e-01 -3.82342815e-01
4.56523240e-01 2.19193310e-01 1.02880311e+00 5.33159792e-01
-5.61277926e-01 8.28421056e-01 2.87698716e-01 -8.81375968e-02
-9.07659531e-01 -1.33905852e+00 -3.19709152e-01 1.18928134e+00
5.62720358e-01 2.47027606e-01 -8.27129066e-01 -1.15811980e+00
1.21684194e-01 8.87695849e-01 -5.55759311e-01 -3.58193815e-01
-4.98290539e-01 -6.62629843e-01 2.96804965e-01 7.64193058e-01
8.38066578e-01 -7.77566135e-01 -3.50899696e-01 1.86089810e-03
-2.28703469e-01 -9.72289681e-01 -7.00788796e-01 2.83025742e-01
-9.84915137e-01 -9.29787815e-01 -1.00635755e+00 -1.10127795e+00
9.63487446e-01 2.48711452e-01 7.30852306e-01 -4.08126771e-01
1.39991259e-02 1.63896859e-03 -3.44776869e-01 -3.01537901e-01
-9.63604674e-02 1.25627071e-01 -7.30652884e-02 1.43197209e-01
6.57232881e-01 -5.94341218e-01 -8.39513600e-01 7.70461261e-01
-6.58140779e-01 -1.39322653e-01 4.41645026e-01 8.85271907e-01
1.00556099e+00 1.22920051e-01 8.10100257e-01 -1.10018241e+00
1.78032756e-01 -5.70418298e-01 -5.59580088e-01 -2.63492577e-03
-5.66307306e-01 -2.31693819e-01 5.99464297e-01 -5.80637634e-01
-1.13153136e+00 3.12892228e-01 6.78174663e-03 -5.02213180e-01
-5.26344061e-01 2.11097717e-01 -6.98834956e-01 4.36635643e-01
6.45224035e-01 2.52819300e-01 -2.68876493e-01 -5.27929306e-01
3.11831534e-01 7.86460102e-01 5.30881464e-01 -6.17318332e-01
7.70308673e-01 5.81361294e-01 -4.55037564e-01 -5.01593769e-01
-8.95600319e-01 -7.10289955e-01 -7.68399537e-01 -6.07063435e-03
7.83980310e-01 -9.18571055e-01 1.74939279e-02 5.84212482e-01
-6.82581484e-01 -4.35755134e-01 -5.68836868e-01 4.60003674e-01
-5.32571793e-01 2.63873935e-01 -4.48115468e-02 -5.32449007e-01
-1.36247454e-02 -1.14417648e+00 1.05519307e+00 6.34414613e-01
-2.79236853e-01 -9.17315781e-01 -2.65060216e-01 5.04412174e-01
-1.03391193e-01 9.60484147e-02 9.33567822e-01 -9.73422289e-01
-3.58998924e-01 -2.34441474e-01 -3.78942519e-01 5.06652057e-01
4.64387774e-01 -2.91175663e-01 -1.05071497e+00 -2.21654102e-01
-2.75055319e-01 -7.66971037e-02 7.85955787e-01 6.36760831e-01
1.41406131e+00 2.96833012e-02 -5.84508717e-01 5.27855337e-01
1.14773536e+00 4.83923018e-01 3.34813684e-01 4.79530036e-01
6.51253760e-01 6.42064989e-01 1.14672422e+00 3.38876903e-01
4.04369026e-01 8.39731157e-01 2.59014249e-01 -5.17671444e-02
-3.57869208e-01 -3.90371561e-01 2.07844615e-01 4.84051973e-01
3.12124163e-01 -4.96092625e-02 -8.22227299e-01 1.07739067e+00
-1.65566742e+00 -4.76463348e-01 4.26063221e-03 2.34416103e+00
9.37178552e-01 3.07176560e-01 2.52959490e-01 4.01174097e-04
1.00149846e+00 -9.26119611e-02 -9.99832690e-01 -9.79530886e-02
-6.14388101e-02 5.08212857e-02 5.04595935e-01 1.74469396e-01
-1.38148057e+00 9.51009214e-01 4.59853506e+00 1.33992398e+00
-1.07999754e+00 1.45065799e-01 8.30134749e-01 5.39727025e-02
-1.64739072e-01 -4.74432372e-02 -8.53372872e-01 7.62751520e-01
4.41359878e-01 4.84066941e-02 4.46850210e-02 1.17733264e+00
1.67426094e-01 -1.59764916e-01 -9.12388742e-01 7.98441708e-01
-2.41659775e-01 -9.70093548e-01 -1.70952201e-01 3.17127034e-02
8.59992206e-01 -1.61808893e-01 3.42098862e-01 3.35103273e-01
2.24475279e-01 -4.89534289e-01 8.31576765e-01 1.07614538e-02
8.96999121e-01 -7.04291999e-01 6.00266755e-01 2.19372705e-01
-1.21308780e+00 -2.10096836e-01 -4.61071491e-01 2.89824009e-01
-1.04512095e-01 5.50416708e-01 -8.49855602e-01 4.88180816e-01
9.98415172e-01 6.89665198e-01 -5.10319889e-01 9.13709521e-01
-2.21269414e-01 6.99320376e-01 -2.70545930e-01 2.88236380e-01
2.86969274e-01 -2.63977438e-01 6.31821215e-01 9.62909698e-01
2.99978077e-01 1.27253691e-02 2.69841939e-01 9.14185524e-01
-3.23578715e-02 8.47230852e-02 -3.43917251e-01 3.16879839e-01
7.13362217e-01 1.09381723e+00 -8.94694090e-01 -2.58422464e-01
-3.36764187e-01 1.01981759e+00 2.20045239e-01 3.93064409e-01
-9.94264781e-01 -5.13158262e-01 8.41143906e-01 3.18863332e-01
5.36738694e-01 7.16754347e-02 -7.67389357e-01 -1.00110912e+00
1.66459799e-01 -5.69019794e-01 7.27933705e-01 -5.08331060e-01
-1.42304599e+00 3.64086151e-01 4.81824949e-02 -1.57612848e+00
-3.74335013e-02 -4.72102284e-01 -5.53384006e-01 8.22834730e-01
-1.53854787e+00 -1.23709750e+00 -3.44837010e-01 5.82482815e-01
8.39753211e-01 -1.76622316e-01 4.56709683e-01 4.75560576e-01
-6.18972182e-01 1.00797570e+00 4.40339744e-01 2.57645696e-01
8.97125781e-01 -1.25819874e+00 4.91303131e-02 8.70236635e-01
-8.89523923e-02 4.27709579e-01 5.22119403e-01 -5.84798038e-01
-4.95737106e-01 -1.37105060e+00 5.48031211e-01 -2.42200762e-01
4.74524915e-01 -3.75124127e-01 -1.09491527e+00 5.93797863e-01
-2.82040805e-01 1.37070954e-01 6.02934182e-01 -4.22591530e-02
-3.39897364e-01 -3.12175751e-01 -1.42451024e+00 4.59806621e-01
1.11102426e+00 -2.90173769e-01 -4.27698612e-01 2.13738874e-01
4.89176393e-01 -2.92766631e-01 -8.30048919e-01 4.40632045e-01
4.14086998e-01 -7.76848435e-01 9.02830064e-01 -3.24382365e-01
3.34759206e-01 -5.04948616e-01 -3.32330287e-01 -1.28662395e+00
-3.40904415e-01 -9.26047489e-02 2.67421395e-01 1.58740747e+00
5.03646910e-01 -7.52420306e-01 1.01678360e+00 4.63280499e-01
-3.90470594e-01 -7.88607955e-01 -1.14704657e+00 -9.97175574e-01
4.07344878e-01 -3.90205115e-01 5.81410468e-01 1.07080996e+00
-2.62083650e-01 5.45177385e-02 1.62091330e-01 4.33158904e-01
6.10623419e-01 4.61269259e-01 5.27031183e-01 -1.09041560e+00
-1.14456743e-01 -5.52023947e-01 -3.08906496e-01 -1.24653208e+00
6.03148714e-02 -1.05450904e+00 2.41853759e-01 -1.31474578e+00
1.28918082e-01 -1.08971000e+00 -4.84133482e-01 6.10110164e-01
-3.80786896e-01 2.22352505e-01 4.98817340e-02 4.13521081e-01
-5.23787379e-01 4.74608153e-01 1.28194439e+00 -1.80079594e-01
-4.26589072e-01 1.88953057e-01 -8.11796665e-01 8.04239333e-01
9.86417353e-01 -5.81804276e-01 -4.86183673e-01 -3.35884929e-01
-4.75146204e-01 -4.70291704e-01 3.25348943e-01 -9.81171310e-01
-3.38534750e-02 -2.61076123e-01 5.71301997e-01 -4.77612317e-01
1.80980965e-01 -8.56884181e-01 -2.29414180e-01 1.10602491e-01
-3.09741199e-01 -5.14515221e-01 1.64672688e-01 6.55902684e-01
-2.13561773e-01 -2.91080713e-01 1.27262259e+00 1.12756252e-01
-1.09604168e+00 2.66724288e-01 -6.48999587e-02 2.58625805e-01
1.22554553e+00 -6.48653686e-01 -9.23108496e-03 5.10763973e-02
-8.97351384e-01 3.57070088e-01 5.81336796e-01 3.94533783e-01
4.04445618e-01 -1.27778661e+00 -4.87564683e-01 3.71214777e-01
5.01162052e-01 4.12999630e-01 5.36933005e-01 8.31844747e-01
-1.04075775e-01 1.68016404e-01 -1.35709867e-01 -8.89794350e-01
-1.09229374e+00 3.82792085e-01 3.94424677e-01 -1.68507054e-01
-4.14113075e-01 1.13260794e+00 6.55481696e-01 -8.24176788e-01
1.34014621e-01 -2.08707273e-01 -1.48300141e-01 6.27866164e-02
2.04608321e-01 3.45529586e-01 -3.88135426e-02 -6.00426376e-01
-5.20765603e-01 6.80036783e-01 -3.04226130e-01 2.05789492e-01
1.11843526e+00 -4.38268960e-01 3.94116163e-01 2.85879940e-01
1.06734407e+00 -1.32232532e-01 -1.78046083e+00 -4.43782210e-01
9.93271023e-02 -5.93792915e-01 -1.87470049e-01 -1.02270269e+00
-1.21791077e+00 8.60937059e-01 9.47233617e-01 -1.21372871e-01
1.37286866e+00 3.66340935e-01 8.08465123e-01 -2.59231687e-01
1.76805630e-01 -1.43612540e+00 4.37312275e-02 2.35572278e-01
5.11131048e-01 -1.28849971e+00 -2.32651770e-01 -6.55958951e-01
-9.65798736e-01 7.24088788e-01 9.08599377e-01 -1.59962490e-01
4.62748647e-01 -2.23901510e-01 2.35864401e-01 1.22643717e-01
-1.04028180e-01 -2.19204694e-01 2.66524762e-01 9.31635737e-01
4.50837016e-02 1.77514091e-01 -1.71618670e-01 9.81909454e-01
2.68998137e-03 -1.13892257e-01 1.21245660e-01 7.50039458e-01
-4.03631777e-01 -1.04649532e+00 -4.68393266e-01 4.22651559e-01
-2.12897614e-01 1.39852911e-01 -9.30782333e-02 8.33505273e-01
5.40970862e-01 8.71187627e-01 2.64855117e-01 -2.38531545e-01
5.22441387e-01 1.15592524e-01 2.57059187e-01 -5.35436928e-01
-1.10810421e-01 3.16080481e-01 -2.37225533e-01 -3.24277699e-01
-2.56528825e-01 -1.00458586e+00 -1.67437124e+00 1.55614838e-01
-3.30881149e-01 2.25177798e-02 4.17022765e-01 7.06062853e-01
5.20377696e-01 4.38999206e-01 8.02707493e-01 -5.03526866e-01
-4.63666201e-01 -8.77226353e-01 -6.92731857e-01 5.49443185e-01
6.58504851e-03 -1.00718546e+00 -2.56370604e-01 1.18749730e-01] | [9.66067886352539, 1.3996021747589111] |
30d6b4cf-14e8-4d27-b689-03d30a212926 | why-artificial-intelligence-needs-a-task | 1604.04660 | null | http://arxiv.org/abs/1604.04660v2 | http://arxiv.org/pdf/1604.04660v2.pdf | Why Artificial Intelligence Needs a Task Theory --- And What It Might Look Like | The concept of "task" is at the core of artificial intelligence (AI): Tasks
are used for training and evaluating AI systems, which are built in order to
perform and automatize tasks we deem useful. In other fields of engineering
theoretical foundations allow thorough evaluation of designs by methodical
manipulation of well understood parameters with a known role and importance;
this allows an aeronautics engineer, for instance, to systematically assess the
effects of wind speed on an airplane's performance and stability. No framework
exists in AI that allows this kind of methodical manipulation: Performance
results on the few tasks in current use (cf. board games, question-answering)
cannot be easily compared, however similar or different. The issue is even more
acute with respect to artificial *general* intelligence systems, which must
handle unanticipated tasks whose specifics cannot be known beforehand. A *task
theory* would enable addressing tasks at the *class* level, bypassing their
specifics, providing the appropriate formalization and classification of tasks,
environments, and their parameters, resulting in more rigorous ways of
measuring, comparing, and evaluating intelligent behavior. Even modest
improvements in this direction would surpass the current ad-hoc nature of
machine learning and AI evaluation. Here we discuss the main elements of the
argument for a task theory and present an outline of what it might look like
for physical tasks. | ['Jóna S. Sigurðardóttir', 'Thröstur Thorarensen', 'Kristinn R. Thórisson', 'Jordi Bieger', 'Bas R. Steunebrink'] | 2016-04-15 | null | null | null | null | ['board-games'] | ['playing-games'] | [ 3.67323607e-01 -8.11533555e-02 3.44364345e-02 -1.46663919e-01
8.08126554e-02 -7.21217930e-01 5.82032859e-01 6.50058985e-02
-5.34781218e-01 6.63588405e-01 -2.98612207e-01 -6.94622755e-01
-9.38571811e-01 -7.41337657e-01 -3.16948712e-01 -7.73695588e-01
-2.05811650e-01 4.02563900e-01 2.60203272e-01 -6.62782848e-01
5.34067333e-01 6.18250787e-01 -1.97987318e+00 -1.31373763e-01
6.78200424e-01 9.52862322e-01 3.37461948e-01 7.76666999e-01
2.54664779e-01 6.45545423e-01 -7.37467051e-01 5.25659509e-02
1.33971840e-01 -3.43556970e-01 -9.39960897e-01 1.51166990e-01
-6.10071048e-03 1.17990695e-01 1.65238053e-01 7.43538201e-01
3.79191816e-01 3.47629696e-01 8.87962461e-01 -9.91143703e-01
-3.24850470e-01 -1.97787806e-01 1.29712820e-01 2.73010105e-01
4.53293443e-01 5.13604403e-01 8.68664861e-01 -1.35604635e-01
1.85139090e-01 7.51521766e-01 4.30111408e-01 3.92860055e-01
-1.12164128e+00 -3.49455923e-01 -3.09466403e-02 -3.60646546e-02
-1.09961259e+00 -4.39769089e-01 6.56625807e-01 -7.06706405e-01
9.53349650e-01 5.48098862e-01 6.80263877e-01 7.30476558e-01
4.27460581e-01 1.93904713e-01 9.93470132e-01 -3.94109011e-01
3.09132099e-01 3.84261668e-01 3.92026752e-02 4.87673610e-01
5.59323072e-01 4.69555557e-01 -2.83571631e-01 1.03872657e-01
6.96725368e-01 -6.08818769e-01 -3.58089328e-01 -5.26146829e-01
-1.37605548e+00 4.02876824e-01 9.36113447e-02 7.05794394e-01
-2.07219139e-01 6.60497993e-02 5.64672649e-01 6.12088978e-01
-1.93457827e-02 1.36756194e+00 -6.78019822e-01 -4.79853451e-01
-4.94295001e-01 6.95250154e-01 8.79547536e-01 7.05980122e-01
4.29990381e-01 2.66770631e-01 1.88778162e-01 5.51134348e-01
-3.36119384e-02 4.50867981e-01 3.88292909e-01 -1.02905273e+00
2.51987666e-01 5.65315068e-01 4.71731633e-01 -8.85198295e-01
-6.06826246e-01 -6.11358941e-01 -3.32943290e-01 7.78626263e-01
6.07184172e-01 -3.32029760e-01 -3.85403484e-01 1.41079903e+00
1.90341905e-01 -2.26577833e-01 6.06260449e-02 1.00593495e+00
3.42761308e-01 3.90219986e-01 1.05566196e-01 -2.65652746e-01
1.45623565e+00 -2.86911279e-01 -5.63561022e-01 -2.71656275e-01
8.16394269e-01 -7.05381691e-01 1.43960249e+00 7.33116448e-01
-1.10919428e+00 -6.81803882e-01 -1.26145840e+00 3.98365289e-01
-5.20374715e-01 -2.40591452e-01 9.32679534e-01 1.09980428e+00
-6.94045007e-01 8.05330634e-01 -5.69197953e-01 -1.87436610e-01
-3.59122247e-01 4.91881788e-01 -1.89433455e-01 3.87444407e-01
-1.17075062e+00 1.47396588e+00 5.29205978e-01 3.67996663e-01
-6.60215437e-01 -6.27576232e-01 -7.41312504e-01 -1.97334617e-01
5.03798962e-01 -6.23093128e-01 1.37267911e+00 -7.73597419e-01
-1.45749795e+00 7.48446584e-01 1.83988407e-01 -1.76444486e-01
1.50600404e-01 -1.38611823e-01 -6.37293220e-01 -2.06245810e-01
-1.93572044e-01 4.66108546e-02 7.19465554e-01 -1.23336339e+00
-4.88510281e-01 -2.32872441e-01 4.51627940e-01 3.48214507e-01
-1.81818038e-01 1.47353888e-01 3.90419334e-01 -3.80797029e-01
-1.26245514e-01 -8.92450571e-01 -1.92352667e-01 -1.81630641e-01
-7.75820613e-02 -1.43514678e-01 7.02182472e-01 -1.72225267e-01
1.30471599e+00 -1.96125007e+00 1.09632738e-01 1.75534010e-01
1.11744307e-01 4.66288865e-01 2.44128361e-01 5.34022331e-01
-2.44655371e-01 1.45896366e-02 -1.33175328e-01 5.37096500e-01
2.50630617e-01 6.90391064e-02 -2.35929936e-02 2.63749272e-01
1.20866857e-01 6.14927292e-01 -8.73198509e-01 -3.07680935e-01
3.77660483e-01 -1.42875360e-02 -4.61921245e-01 1.50710031e-01
-1.81164980e-01 2.20219508e-01 -8.04601371e-01 3.87138844e-01
-2.75860161e-01 4.72291000e-02 1.52869299e-01 -4.47436050e-02
-3.47817779e-01 4.10465956e-01 -1.28877902e+00 1.11874211e+00
-5.37935972e-01 6.16057932e-01 2.02844724e-01 -1.32624102e+00
1.07309914e+00 4.77679998e-01 5.06321907e-01 -7.83854127e-01
2.38665879e-01 3.23976517e-01 5.91114640e-01 -9.28688049e-01
5.30268013e-01 -4.40530926e-01 -2.11216673e-01 4.33703572e-01
-3.07879418e-01 -7.77772069e-01 1.99215904e-01 -3.21730524e-01
1.18558991e+00 2.29431212e-01 2.07548633e-01 -7.93095171e-01
5.63570380e-01 3.41619730e-01 1.88572824e-01 6.18821084e-01
-2.67314851e-01 2.08836198e-02 3.27481866e-01 -6.65152848e-01
-9.22474444e-01 -8.95853877e-01 -5.54678977e-01 9.13831770e-01
2.67772019e-01 -1.02804556e-01 -6.68096483e-01 -2.34275222e-01
1.18789949e-01 8.50707710e-01 -4.36922312e-01 -3.12510401e-01
-5.06504953e-01 -7.48227835e-01 2.57303387e-01 3.14673454e-01
2.12082267e-01 -1.29173231e+00 -1.26892591e+00 2.33472303e-01
1.79562211e-01 -9.09524620e-01 2.34356031e-01 3.98381591e-01
-8.65371823e-01 -1.20079374e+00 -1.47610590e-01 -5.05902290e-01
3.46963853e-01 9.00394917e-02 1.05142236e+00 3.53957862e-01
-2.65279889e-01 4.58386451e-01 -2.54670978e-01 -7.90041149e-01
-5.00747800e-01 -1.00457072e-01 3.99897873e-01 -5.17568409e-01
2.55641937e-01 -6.77630305e-01 -5.61104715e-01 6.76158249e-01
-1.00186741e+00 -2.29712948e-01 5.99574327e-01 5.53750813e-01
-5.84772862e-02 5.89222968e-01 6.30030155e-01 -7.83820927e-01
9.22413170e-01 -1.45468250e-01 -5.16503990e-01 2.74636149e-01
-6.95302904e-01 2.12905988e-01 5.98994017e-01 -3.71759027e-01
-7.87673950e-01 -3.41871768e-01 4.62594721e-03 1.92493379e-01
-5.54104507e-01 4.30517852e-01 -2.33569309e-01 -3.54522109e-01
9.25922930e-01 -1.24292567e-01 1.27521515e-01 -1.40078545e-01
-1.51332229e-01 5.26035607e-01 4.44237977e-01 -1.14204323e+00
1.02048433e+00 -9.65252984e-03 3.47231179e-01 -1.03877962e+00
-5.46550512e-01 -1.17517389e-01 -4.78704184e-01 -5.10803759e-01
8.85643303e-01 -2.30996355e-01 -1.10761678e+00 1.46187216e-01
-6.88398302e-01 -5.70700407e-01 -4.42936599e-01 6.42056823e-01
-9.27877784e-01 1.28119066e-01 -2.26433069e-01 -9.93240893e-01
2.74007410e-01 -1.43958390e+00 9.18198287e-01 -2.08741166e-02
-7.23487973e-01 -1.14544153e+00 -1.17867060e-01 5.77758729e-01
6.53199673e-01 2.61883318e-01 1.21241057e+00 -5.74592352e-01
-2.16018215e-01 -2.91711360e-01 2.64363974e-01 3.73326451e-01
1.25969693e-01 4.56039160e-02 -9.04539227e-01 -2.52454549e-01
3.47510219e-01 -3.14305812e-01 1.81705087e-01 9.80988219e-02
1.02687955e+00 7.11698532e-02 -1.68476999e-01 2.67710481e-02
1.29012954e+00 6.61952496e-01 6.54938281e-01 4.92592633e-01
1.58523798e-01 1.02696037e+00 8.55311275e-01 1.60261869e-01
-1.67780951e-01 1.00973761e+00 4.45531845e-01 -4.52510733e-03
2.53965855e-01 2.97334820e-01 1.42441571e-01 6.21370614e-01
-5.69261789e-01 -2.10692093e-01 -1.06148672e+00 3.38271677e-01
-1.42050231e+00 -1.08777773e+00 -2.17835888e-01 2.49879837e+00
4.97615486e-01 6.79369509e-01 1.41687458e-02 6.79472685e-01
4.51196551e-01 1.17374323e-01 -3.46601367e-01 -7.22935021e-01
4.94279712e-01 2.87462652e-01 2.57061511e-01 4.25199449e-01
-9.05842304e-01 4.94009167e-01 7.31767559e+00 5.69692969e-01
-9.07633424e-01 -3.54566038e-01 3.88690412e-01 1.50892660e-01
-2.28723586e-01 -2.65900921e-02 -3.95136565e-01 4.04878765e-01
1.07156003e+00 -2.17293754e-01 5.23599803e-01 7.08207309e-01
4.53006268e-01 -4.79144186e-01 -1.30941033e+00 5.46663225e-01
-2.77969956e-01 -9.50210571e-01 -2.73753017e-01 1.74879119e-01
2.00468197e-01 -5.94231844e-01 -3.35051827e-02 3.85800391e-01
1.37294382e-01 -1.04107904e+00 6.55684173e-01 5.78262448e-01
5.67642510e-01 -6.30983829e-01 6.61283553e-01 5.77128470e-01
-8.69040847e-01 -2.25769252e-01 -2.67121017e-01 -6.17282927e-01
9.49466899e-02 4.51479524e-01 -5.80717385e-01 5.43896317e-01
5.01888573e-01 -1.29122451e-01 -2.49787569e-01 9.05436277e-01
9.11007524e-02 4.79411423e-01 -2.18223661e-01 -4.97942120e-01
2.36985341e-01 -3.45752329e-01 5.93979478e-01 8.70779574e-01
1.37420848e-01 2.72861391e-01 -8.23399797e-02 7.75475979e-01
7.24617004e-01 -1.07048988e-01 -7.39884377e-01 -1.20150022e-01
3.57804060e-01 1.06592035e+00 -7.04926968e-01 -2.04916745e-02
-2.08071142e-01 3.50819707e-01 -2.08242968e-01 2.98456132e-01
-8.89153659e-01 -6.55056298e-01 1.00623322e+00 5.73047221e-01
-3.45914423e-01 -6.28198028e-01 -3.98177266e-01 -7.63312221e-01
1.33670852e-01 -1.05805647e+00 -1.16754137e-01 -7.72004902e-01
-8.66837084e-01 4.29702491e-01 4.16170776e-01 -1.08965755e+00
-5.54484427e-01 -1.15753043e+00 -6.36586845e-01 8.15977395e-01
-1.06549942e+00 -5.63574076e-01 -2.79805839e-01 3.75833273e-01
3.14165890e-01 -2.33529478e-01 8.45324039e-01 1.14278637e-01
-4.59600657e-01 1.27343744e-01 -2.38882124e-01 -1.52921498e-01
3.66001606e-01 -1.24741328e+00 2.55623043e-01 5.21323562e-01
-1.23778589e-01 6.24345481e-01 1.24187553e+00 -3.68043393e-01
-1.72255325e+00 -3.98002774e-01 6.74939692e-01 -8.50300133e-01
9.03955936e-01 -3.08807820e-01 -7.04585195e-01 1.91006258e-01
-8.85904208e-02 -3.82043839e-01 3.21554452e-01 4.47116196e-01
1.41691700e-01 -1.46060064e-01 -8.44682634e-01 6.79143786e-01
1.14648616e+00 -5.97739279e-01 -8.23097110e-01 5.56865990e-01
5.69481671e-01 -2.57823884e-01 -9.67713714e-01 6.56286597e-01
6.92235529e-01 -1.20539737e+00 1.01172316e+00 -8.34125936e-01
2.07972884e-01 -3.86440456e-01 8.94924849e-02 -1.17352593e+00
-4.12822127e-01 -6.50423288e-01 4.07457173e-01 7.19087780e-01
5.89646935e-01 -8.65018785e-01 7.39981532e-01 9.15647149e-01
-4.47540998e-01 -8.02488267e-01 -6.13559842e-01 -1.09962785e+00
1.12804808e-01 -7.94734240e-01 4.89364088e-01 1.00480437e+00
2.71942586e-01 2.90994883e-01 1.22733198e-01 1.68712527e-01
3.26568246e-01 -1.09261900e-01 7.43367016e-01 -1.47909713e+00
-5.73768377e-01 -6.71259761e-01 -6.51926160e-01 -5.64401031e-01
2.07806546e-02 -5.19461751e-01 4.03388850e-02 -1.21731305e+00
-4.36132461e-01 -7.16379523e-01 -8.19609538e-02 1.25979811e-01
1.12848952e-01 -1.23211235e-01 1.75565500e-02 1.23338073e-01
-2.33642220e-01 1.44843414e-01 1.35009325e+00 1.95622414e-01
-1.18432328e-01 2.30736166e-01 -6.45128071e-01 7.55550206e-01
9.84583855e-01 -1.24581009e-01 -7.63495982e-01 -3.06371689e-01
4.08699751e-01 5.91602325e-02 4.69633758e-01 -1.53053284e+00
6.53913021e-02 -4.74410564e-01 2.46755704e-01 2.18881994e-01
3.20422471e-01 -1.10736144e+00 3.74464035e-01 5.25582731e-01
-3.82361978e-01 3.54941338e-01 1.69945359e-01 1.63071603e-01
-1.08329065e-01 -6.45506322e-01 5.86924434e-01 -4.23529372e-02
-7.48906732e-01 -1.64630890e-01 -6.03211403e-01 -7.70352036e-02
1.35092735e+00 -7.46343970e-01 -3.81574929e-01 -2.22695842e-01
-7.64749348e-01 6.55212849e-02 3.95144105e-01 3.38093728e-01
2.07865983e-01 -8.60900760e-01 -2.23312676e-01 2.50893116e-01
1.07954100e-01 -2.01641649e-01 8.72475728e-02 7.29094446e-01
-6.29873216e-01 5.04282057e-01 -2.62926757e-01 -1.96731925e-01
-8.80822957e-01 7.28493035e-01 5.90627134e-01 -4.59259227e-02
-3.00434828e-01 3.70719641e-01 4.06288505e-01 -3.55314910e-01
-1.28666610e-02 -4.02683258e-01 -1.72060400e-01 -2.82999367e-01
3.08216393e-01 1.99923411e-01 4.01298970e-01 -3.22784036e-01
-2.55995601e-01 5.20689964e-01 3.22494388e-01 6.95941783e-03
9.69382048e-01 2.70637959e-01 1.06627792e-01 5.28353214e-01
5.62812924e-01 -2.34359249e-01 -9.75952804e-01 3.25593293e-01
1.95485145e-01 -2.35733524e-01 6.60908297e-02 -8.09180796e-01
-4.93620574e-01 9.02443349e-01 2.48638317e-01 7.19201028e-01
1.14804316e+00 -2.24915206e-01 3.16366822e-01 7.48759806e-01
6.46101356e-01 -1.29429841e+00 -4.49970439e-02 3.49146992e-01
9.31303620e-01 -7.21189797e-01 7.33575895e-02 -3.48392546e-01
-5.94816387e-01 1.14802980e+00 7.52549589e-01 4.85609062e-02
7.40262151e-01 6.06238365e-01 -1.24270111e-01 -5.43244362e-01
-7.55826116e-01 -2.95840949e-01 2.90661037e-01 5.69736958e-01
6.50884330e-01 8.39449987e-02 -3.92447531e-01 2.37554550e-01
-4.44863468e-01 -2.57686168e-01 3.59428465e-01 1.13247693e+00
-9.54143822e-01 -1.18072319e+00 -5.72532117e-01 5.13575196e-01
-1.62191197e-01 3.61047953e-01 -2.57094234e-01 1.21773529e+00
2.59242237e-01 9.92561877e-01 -4.92042340e-02 -6.14364386e-01
7.91375399e-01 1.32251278e-01 5.01037121e-01 -5.83564997e-01
-7.05390930e-01 -4.45445210e-01 6.01691306e-01 -4.51799810e-01
-3.78596216e-01 -6.29465401e-01 -9.32641387e-01 -4.00524616e-01
-3.23808402e-01 5.13680041e-01 8.31015587e-01 1.04630959e+00
-1.61434151e-02 7.27572083e-01 4.11946446e-01 -9.85768139e-01
-7.03678906e-01 -5.31157970e-01 -8.40816736e-01 3.42071533e-01
1.68446009e-03 -1.08577883e+00 -4.03273344e-01 -1.58168167e-01] | [5.615322113037109, 3.6655032634735107] |
fbb25e94-870c-46f2-806b-a92912dffb33 | learning-with-difference-attention-for | 2306.14603 | null | https://arxiv.org/abs/2306.14603v1 | https://arxiv.org/pdf/2306.14603v1.pdf | Learning with Difference Attention for Visually Grounded Self-supervised Representations | Recent works in self-supervised learning have shown impressive results on single-object images, but they struggle to perform well on complex multi-object images as evidenced by their poor visual grounding. To demonstrate this concretely, we propose visual difference attention (VDA) to compute visual attention maps in an unsupervised fashion by comparing an image with its salient-regions-masked-out version. We use VDA to derive attention maps for state-of-the art SSL methods and show they do not highlight all salient regions in an image accurately, suggesting their inability to learn strong representations for downstream tasks like segmentation. Motivated by these limitations, we cast VDA as a differentiable operation and propose a new learning objective, Differentiable Difference Attention (DiDA) loss, which leads to substantial improvements in an SSL model's visually grounding to an image's salient regions. | ['Balaji Vasan Srinivasan', 'Srikrishna Karanam', 'Aishwarya Agarwal'] | 2023-06-26 | null | null | null | null | ['visual-grounding', 'self-supervised-learning'] | ['computer-vision', 'computer-vision'] | [ 5.45844972e-01 3.25837523e-01 -1.16665483e-01 -3.97499561e-01
-7.41855443e-01 -4.23820317e-01 7.56812334e-01 2.94389755e-01
-2.88535297e-01 4.96188998e-01 1.65836200e-01 -1.96823642e-01
1.92519948e-02 -4.02396679e-01 -1.06797874e+00 -3.77536684e-01
8.70109871e-02 7.44753033e-02 5.42285740e-01 -2.06301838e-01
3.33566785e-01 5.38206339e-01 -1.52578700e+00 2.48147771e-01
9.09491479e-01 6.11512601e-01 3.47246647e-01 5.42611063e-01
-1.38535425e-01 8.74630213e-01 -5.97380877e-01 -2.51729101e-01
2.81825334e-01 -6.23105049e-01 -1.12896657e+00 2.13233605e-01
1.07236564e+00 -2.71211356e-01 -6.91585019e-02 1.17794049e+00
1.81174502e-01 1.71013638e-01 7.88407862e-01 -1.53692293e+00
-1.02747977e+00 4.25900370e-01 -8.79178107e-01 4.67061549e-01
7.22096264e-02 2.74174809e-01 1.45222270e+00 -1.03317738e+00
6.77650988e-01 1.48864436e+00 6.69011176e-01 3.80795568e-01
-1.78086472e+00 -2.12627307e-01 5.20529747e-01 -8.56097266e-02
-9.59946811e-01 -3.29403698e-01 8.87981653e-01 -4.38016117e-01
8.94828260e-01 1.90767825e-01 5.09607017e-01 8.60328674e-01
7.82749131e-02 1.21225321e+00 1.26309276e+00 -3.91454101e-01
1.18743487e-01 -8.14874694e-02 1.04648232e-01 7.16833115e-01
2.34968573e-01 4.19336371e-02 -5.54140985e-01 3.85585606e-01
9.95877862e-01 3.57002132e-02 -1.83368355e-01 -5.89629710e-01
-1.22854114e+00 6.93949461e-01 1.01538312e+00 3.03756505e-01
-2.16672838e-01 2.71977544e-01 1.45526215e-01 2.32687697e-01
5.78779101e-01 7.34134257e-01 -3.44399571e-01 3.84915233e-01
-1.04865384e+00 2.38465443e-01 1.53408065e-01 7.11936057e-01
9.52974141e-01 1.22411638e-01 -4.91912276e-01 7.30627596e-01
1.44903600e-01 2.17176393e-01 1.32370770e-01 -1.01076901e+00
2.52958506e-01 7.40384698e-01 -7.84398708e-03 -9.46406126e-01
-4.40357238e-01 -5.75580657e-01 -6.44949198e-01 8.27451289e-01
6.07641280e-01 1.11610025e-01 -1.28121614e+00 1.79142177e+00
1.40539274e-01 -1.68175660e-02 1.51992366e-01 1.03737628e+00
9.72594500e-01 4.83858585e-01 5.12270868e-01 2.03820556e-01
9.95990574e-01 -1.29655635e+00 -4.14266318e-01 -7.51884460e-01
2.99518734e-01 -5.32465816e-01 1.53086126e+00 4.09230739e-02
-1.54265749e+00 -8.62915277e-01 -1.11521304e+00 -5.48072755e-01
-3.89702201e-01 -1.38287190e-02 6.21832788e-01 8.68750513e-02
-1.23319328e+00 6.81254148e-01 -7.68775403e-01 -4.36490506e-01
1.01866949e+00 1.22486591e-01 -2.45994270e-01 2.22849026e-01
-7.70142198e-01 1.05236387e+00 1.95357174e-01 -1.60031691e-01
-1.03709030e+00 -1.04046953e+00 -1.11207020e+00 3.64520848e-01
2.74702072e-01 -7.22763121e-01 1.16035807e+00 -1.28759491e+00
-1.07067859e+00 1.32588720e+00 -1.27463490e-01 -4.71599460e-01
6.06612146e-01 -2.02818960e-01 -1.06430031e-01 2.58247912e-01
4.26916510e-01 1.16196299e+00 1.00833666e+00 -1.73313594e+00
-7.11199641e-01 -2.32177839e-01 3.23237509e-01 1.83409646e-01
-1.69524681e-02 -1.41019732e-01 -3.27169389e-01 -8.52774262e-01
1.08669765e-01 -5.52224398e-01 -4.02529448e-01 3.90626878e-01
-6.47075295e-01 -2.01521873e-01 7.19628334e-01 -4.64650750e-01
7.67886579e-01 -2.11143804e+00 1.31234899e-01 -4.58452068e-02
4.19393986e-01 4.63328540e-01 -5.01884878e-01 7.57420361e-02
-2.38384813e-01 2.84249187e-01 -5.14059067e-01 -6.11181676e-01
-1.13479592e-01 1.70000419e-01 -3.68830115e-01 3.39787632e-01
9.90870535e-01 1.30094004e+00 -1.09933388e+00 -6.10025167e-01
2.84513921e-01 3.51273417e-01 -5.12003183e-01 2.66967177e-01
-3.16345781e-01 5.06480634e-01 -1.89947784e-01 5.20354688e-01
6.31441534e-01 -5.24343073e-01 -1.57626286e-01 -2.72571236e-01
-1.18406154e-01 3.14408913e-02 -7.67056704e-01 1.74851084e+00
-2.70879388e-01 9.39198732e-01 1.12139911e-03 -1.08098435e+00
7.13229179e-01 -2.60903090e-01 2.57064223e-01 -9.25079882e-01
-1.35517359e-01 3.23912054e-02 -4.23444286e-02 -4.00573730e-01
4.04036254e-01 -2.25495726e-01 1.67661309e-01 3.65498185e-01
2.17374846e-01 -2.51081645e-01 1.72191486e-01 4.20257539e-01
8.44139636e-01 2.06566721e-01 2.37304270e-01 -4.51726645e-01
4.55687255e-01 8.21973830e-02 3.67793113e-01 9.90678966e-01
-3.15112710e-01 1.04174685e+00 6.17062211e-01 -2.49914587e-01
-1.12457359e+00 -1.39468515e+00 -1.00743122e-01 1.35782063e+00
4.47115988e-01 -8.22858810e-02 -7.09151089e-01 -9.12160337e-01
3.32174867e-01 4.93333042e-01 -8.10452104e-01 -1.58023223e-01
-4.33920950e-01 -2.82337159e-01 1.65220529e-01 8.93112421e-01
4.76547629e-01 -1.31452477e+00 -7.94549704e-01 4.03609835e-02
1.54090106e-01 -1.02251959e+00 -5.15421450e-01 3.47364455e-01
-5.72208941e-01 -1.05057538e+00 -1.15325248e+00 -1.13517106e+00
9.83899713e-01 4.06560898e-01 1.41515470e+00 1.96585223e-01
-4.64415103e-01 2.35991091e-01 -5.53899296e-02 -5.04349351e-01
-3.83641928e-01 -1.10914232e-02 -3.97095680e-01 4.27548327e-02
-1.55558586e-01 -4.39508319e-01 -5.95717430e-01 1.52159914e-01
-9.31183577e-01 1.96857154e-01 6.85639322e-01 9.16329443e-01
6.60756767e-01 -3.90376538e-01 6.61683619e-01 -9.47037518e-01
4.26241040e-01 -2.78966695e-01 -5.65307021e-01 2.68529028e-01
-5.52769125e-01 2.84754038e-01 5.09327710e-01 -1.77291885e-01
-8.66741180e-01 1.16100185e-01 -5.90162612e-02 -4.35036659e-01
-1.50998682e-01 2.21318677e-01 1.91033110e-01 -2.32057571e-01
4.86663043e-01 1.44110546e-01 1.40306473e-01 -4.39369082e-01
5.84665239e-01 2.46591032e-01 9.07727957e-01 -3.39375138e-01
8.67013216e-01 6.33985758e-01 7.18854889e-02 -6.29864275e-01
-1.47163737e+00 -4.99322802e-01 -8.32107484e-01 -3.21370326e-02
1.06523812e+00 -7.90962994e-01 -6.01548433e-01 2.83255309e-01
-1.00594127e+00 -6.93236291e-01 -5.33570170e-01 1.72661785e-02
-8.17575634e-01 2.64848977e-01 -3.69923800e-01 -6.42810583e-01
-9.77206677e-02 -9.92848575e-01 1.41701782e+00 3.43397826e-01
-1.82085454e-01 -9.85750973e-01 -1.86127856e-01 1.11070894e-01
4.22902673e-01 3.77402961e-01 1.03363049e+00 -3.38906258e-01
-7.04354286e-01 3.58171672e-01 -7.30897546e-01 4.80137825e-01
2.73726359e-02 3.92624438e-02 -1.04095864e+00 -4.50288504e-01
-4.53095436e-01 -4.57595348e-01 1.32956290e+00 5.59386134e-01
1.32030177e+00 -1.25118643e-01 -3.20210367e-01 7.56619871e-01
1.42840767e+00 -1.39411733e-01 4.57631320e-01 2.40816683e-01
8.29377770e-01 5.67221582e-01 5.00138342e-01 8.64027347e-03
1.42144099e-01 5.25617421e-01 6.48810446e-01 -9.03482437e-01
-7.40178049e-01 -4.42376226e-01 9.66001227e-02 1.45209968e-01
1.16471715e-01 -1.02645472e-01 -7.05734849e-01 8.29283655e-01
-1.87468719e+00 -8.29554617e-01 -2.39106134e-01 1.93850851e+00
7.57377446e-01 3.12636882e-01 2.49505773e-01 5.63504919e-02
5.55312395e-01 4.61899400e-01 -7.36575186e-01 -4.05748099e-01
-3.34107935e-01 1.74642608e-01 3.49687010e-01 4.69317168e-01
-1.29193926e+00 1.14054430e+00 7.21424150e+00 5.50664485e-01
-1.16995931e+00 -8.32334068e-03 8.78809929e-01 1.10864073e-01
-5.10337055e-01 -8.21696743e-02 -4.20314699e-01 3.36332262e-01
3.38269502e-01 -2.55971774e-02 1.69672102e-01 6.78994238e-01
9.71347839e-02 -2.04451889e-01 -1.27435482e+00 8.91393125e-01
1.13869876e-01 -1.58716893e+00 2.73711860e-01 -2.18640193e-01
1.01200140e+00 3.52906026e-02 4.57841307e-01 1.50757089e-01
4.56799060e-01 -1.21550393e+00 9.51346815e-01 2.49506339e-01
7.19584465e-01 -5.12363791e-01 2.58298278e-01 -3.55992764e-02
-9.39352572e-01 -4.44927923e-02 -3.49690795e-01 -7.46822134e-02
2.01878682e-01 2.41081521e-01 -7.34171271e-01 1.64448440e-01
7.77248919e-01 9.53362286e-01 -9.84195888e-01 1.08877909e+00
-3.51583511e-01 4.25905913e-01 2.34150831e-02 2.06413984e-01
6.99363768e-01 9.21956971e-02 4.71879274e-01 1.19461167e+00
-1.97652683e-01 -2.46022940e-01 1.86942324e-01 1.49023485e+00
-1.83568195e-01 -9.15061682e-02 -5.38995922e-01 1.21148065e-01
1.39502371e-02 1.05762804e+00 -9.09577787e-01 -3.87871474e-01
-4.41200346e-01 1.23357689e+00 6.31156504e-01 7.19442129e-01
-7.30266988e-01 -3.91100198e-01 7.48480856e-01 1.82031810e-01
4.87763852e-01 -7.24351928e-02 -4.83154982e-01 -9.12570000e-01
-6.38214797e-02 -5.04023910e-01 2.28991687e-01 -1.17941272e+00
-1.47060239e+00 3.94900918e-01 -9.41255838e-02 -1.05859768e+00
8.94865543e-02 -5.16140699e-01 -7.22808301e-01 8.41014147e-01
-1.92758858e+00 -1.35653961e+00 -3.19929183e-01 5.33880413e-01
6.69941545e-01 9.60967764e-02 2.73851037e-01 -3.48480679e-02
-3.64587665e-01 5.55068076e-01 -3.26146744e-02 1.31179318e-01
6.04289293e-01 -1.63214445e+00 7.82766342e-01 1.02292669e+00
4.65724647e-01 3.74445409e-01 6.31222188e-01 -3.95398408e-01
-9.35658038e-01 -1.13923037e+00 7.42753804e-01 -5.58491766e-01
5.30331552e-01 -3.16424400e-01 -1.13787520e+00 7.67461360e-01
2.94810504e-01 5.10859907e-01 1.93223935e-02 -5.24529815e-02
-4.04286116e-01 1.63394138e-01 -1.04001272e+00 7.80726612e-01
1.25445473e+00 -5.27499020e-01 -7.63439775e-01 3.52203310e-01
8.31213951e-01 -2.15124607e-01 -5.99883676e-01 4.30991441e-01
1.61211148e-01 -1.07307243e+00 1.33587301e+00 -9.23289716e-01
7.19711542e-01 -3.54987919e-01 2.44037524e-01 -1.25995338e+00
-4.72769469e-01 -5.21208823e-01 1.65730998e-01 1.24515903e+00
3.78066272e-01 -5.24907351e-01 6.05543911e-01 2.29391560e-01
-3.50502193e-01 -6.46502972e-01 -5.75629473e-01 -7.39983141e-01
2.79681981e-01 -2.11930685e-02 3.55692357e-01 8.87797117e-01
-1.07985452e-01 3.91233802e-01 -1.08922958e-01 1.17125995e-01
7.28117764e-01 4.30054843e-01 5.79987824e-01 -1.35537136e+00
-8.30926597e-02 -7.23064244e-01 -3.64105225e-01 -1.06270349e+00
2.56445438e-01 -1.01771533e+00 2.99130708e-01 -1.71667862e+00
4.47891712e-01 -4.44945127e-01 -5.16206801e-01 6.80752814e-01
-5.49709320e-01 6.08611286e-01 5.05418658e-01 1.21800378e-01
-8.41910601e-01 3.82404894e-01 1.55861962e+00 -4.76800054e-01
-1.21992446e-01 -3.04393888e-01 -8.88183773e-01 7.06575632e-01
4.77958798e-01 -2.13948011e-01 -3.54419529e-01 -4.49200362e-01
-2.29881108e-01 -3.09981078e-01 7.52438784e-01 -9.40493345e-01
1.00069568e-02 -1.95437521e-01 6.25924170e-01 -4.63138223e-01
1.18293069e-01 -4.36250418e-01 -4.66280073e-01 3.14635247e-01
-8.22081208e-01 -1.37771428e-01 2.48573601e-01 4.71308708e-01
-3.17134351e-01 -4.64612581e-02 1.02730036e+00 1.16024045e-02
-1.13575888e+00 1.92737445e-01 -5.46395034e-02 4.95464325e-01
9.25044119e-01 -3.58073741e-01 -4.02724713e-01 -2.08353996e-01
-7.70525038e-01 2.40106300e-01 6.85358465e-01 3.86472523e-01
6.30247355e-01 -1.16227424e+00 -7.50261664e-01 4.53901023e-01
3.01649064e-01 1.99759766e-01 2.46418297e-01 6.99371099e-01
-4.38758731e-01 2.76114821e-01 -5.85000277e-01 -8.13861251e-01
-1.22687256e+00 6.92183256e-01 1.64547548e-01 -1.14512090e-02
-8.85576248e-01 1.02231479e+00 6.45288885e-01 -6.78733140e-02
3.25252712e-01 -7.07085788e-01 -1.77017301e-01 -1.40845522e-01
3.21709603e-01 -5.49726635e-02 -5.90372644e-02 -6.38373971e-01
-3.69539261e-01 8.10481608e-01 -2.31197104e-02 1.14078015e-01
1.24921072e+00 -2.05733076e-01 3.05676490e-01 2.92904824e-01
1.31842232e+00 -2.47524008e-01 -1.94336021e+00 -2.82680631e-01
5.99734671e-02 -6.43391371e-01 1.46805540e-01 -8.24733675e-01
-1.17521870e+00 1.04570806e+00 4.78498816e-01 2.42555901e-01
9.55877066e-01 4.50737357e-01 6.75602317e-01 1.69219002e-02
-3.94302756e-02 -9.74344432e-01 5.24245918e-01 3.52367669e-01
1.08466136e+00 -1.74182963e+00 3.01402491e-02 -4.20734197e-01
-8.66151035e-01 7.81690180e-01 6.20893419e-01 -2.62328625e-01
3.49435657e-01 8.31841454e-02 1.97469369e-01 -2.25806192e-01
-3.92480582e-01 -8.59380901e-01 6.83699787e-01 8.22058678e-01
3.02460164e-01 -2.95874536e-01 9.64964852e-02 8.54078904e-02
1.27344325e-01 -2.41680920e-01 4.01651293e-01 7.99562454e-01
-4.07697856e-01 -6.07308030e-01 -1.29586086e-03 3.36477041e-01
-3.69476587e-01 -8.64031985e-02 -4.36926723e-01 8.99921417e-01
-1.00557119e-01 6.95583701e-01 4.66571122e-01 1.34896825e-03
3.30933481e-01 -2.64686406e-01 5.19637167e-01 -7.51850903e-01
-5.16852140e-01 -1.00286871e-01 -2.75600821e-01 -6.70003891e-01
-5.37160933e-01 -4.81638402e-01 -1.43357670e+00 1.52243465e-01
1.41424671e-01 -2.90117323e-01 4.06424552e-01 8.23960364e-01
2.38517329e-01 8.68051350e-01 3.87131006e-01 -1.21391284e+00
-4.48049366e-01 -5.29696286e-01 -4.31156754e-01 1.06016564e+00
7.94022739e-01 -6.99099779e-01 -3.25552523e-01 1.39484897e-01] | [9.820395469665527, 1.3911646604537964] |
ece75529-9a81-4185-bde5-8c018f777aca | 2d-and-3d-cnn-based-fusion-approach-for-covid | 2303.08740 | null | https://arxiv.org/abs/2303.08740v1 | https://arxiv.org/pdf/2303.08740v1.pdf | 2D and 3D CNN-Based Fusion Approach for COVID-19 Severity Prediction from 3D CT-Scans | Since the appearance of Covid-19 in late 2019, Covid-19 has become an active research topic for the artificial intelligence (AI) community. One of the most interesting AI topics is Covid-19 analysis of medical imaging. CT-scan imaging is the most informative tool about this disease. This work is part of the 3nd COV19D competition for Covid-19 Severity Prediction. In order to deal with the big gap between the validation and test results that were shown in the previous version of this competition, we proposed to combine the prediction of 2D and 3D CNN predictions. For the 2D CNN approach, we propose 2B-InceptResnet architecture which consists of two paths for segmented lungs and infection of all slices of the input CT-scan, respectively. Each path consists of ConvLayer and Inception-ResNet pretrained model on ImageNet. For the 3D CNN approach, we propose hybrid-DeCoVNet architecture which consists of four blocks: Stem, four 3D-ResNet layers, Classification Head and Decision layer. Our proposed approaches outperformed the baseline approach in the validation data of the 3nd COV19D competition for Covid-19 Severity Prediction by 36%. | ['Abdelmalik Taleb-Ahmed', 'Cosimo Distante', 'Amir Nakib', 'Fadi Dornaika', 'Fares Bougourzi'] | 2023-03-15 | null | null | null | null | ['severity-prediction'] | ['computer-vision'] | [ 8.19766670e-02 1.12969242e-01 -8.64278004e-02 -2.48645142e-01
-5.20670235e-01 -2.74636269e-01 3.56593937e-01 1.48627445e-01
-5.81719041e-01 5.75113297e-01 3.14120471e-01 -4.18867320e-01
-1.45605773e-01 -6.81702614e-01 -3.78230184e-01 -5.33447146e-01
-3.23182233e-02 8.68246317e-01 5.90586126e-01 -1.24008432e-02
-2.31244490e-01 7.99257696e-01 -8.58705938e-01 6.64348960e-01
1.90384835e-01 1.35661507e+00 3.07951689e-01 1.05596828e+00
2.80454308e-01 1.02838302e+00 -4.09065276e-01 -1.65401906e-01
1.36250466e-01 -3.36738497e-01 -8.90094876e-01 -4.66175884e-01
3.89214575e-01 -2.34796405e-01 -3.83710206e-01 6.76543117e-01
8.49998534e-01 -3.02351922e-01 8.82671297e-01 -9.79627848e-01
-1.61670640e-01 4.89504129e-01 -5.61793745e-01 9.30416644e-01
-3.91581431e-02 2.51524091e-01 7.93331146e-01 -7.19658256e-01
8.29853535e-01 1.08360803e+00 9.46974933e-01 6.08850181e-01
-5.94181597e-01 -6.19892061e-01 -3.21007282e-01 4.19043630e-01
-1.09014964e+00 4.55864102e-01 2.97814578e-01 -7.87721336e-01
1.17728782e+00 3.54133874e-01 9.21971023e-01 1.27226138e+00
8.67125154e-01 8.24670613e-01 8.14229369e-01 7.06572682e-02
-1.48341760e-01 -1.21190697e-01 3.07431757e-01 7.02775717e-01
1.66343227e-01 1.44751683e-01 3.37348193e-01 -2.03800291e-01
8.74438763e-01 2.81381130e-01 -4.67252098e-02 -1.44763902e-01
-1.22551715e+00 9.15391505e-01 8.43972147e-01 5.47843516e-01
-5.92162907e-01 2.14494258e-01 8.16247880e-01 -4.32598367e-02
4.82255399e-01 3.88978511e-01 -4.00504231e-01 3.80191743e-01
-7.02728748e-01 5.49677193e-01 3.78817141e-01 4.02801156e-01
-2.09642082e-01 7.01227039e-02 -5.71381569e-01 8.00209343e-01
2.86109477e-01 5.32046437e-01 8.84741068e-01 -1.72301337e-01
3.53759319e-01 6.76657259e-01 -4.16149944e-01 -7.34743774e-01
-1.14436924e+00 -7.27637887e-01 -1.41211033e+00 -1.10960290e-01
1.48184389e-01 -2.83294976e-01 -1.40585387e+00 1.38647461e+00
1.99871987e-01 4.18856800e-01 -1.89683959e-01 1.08178699e+00
1.34984851e+00 4.45381731e-01 2.59802848e-01 1.25133887e-01
1.50891352e+00 -1.10605001e+00 -2.36529067e-01 4.53403294e-01
7.51832247e-01 -7.24034190e-01 4.67227638e-01 3.24397236e-01
-1.04025543e+00 -5.86220384e-01 -8.85757565e-01 1.55547932e-01
-4.32377785e-01 -6.79215938e-02 3.29462677e-01 4.41783756e-01
-1.32060885e+00 4.05464739e-01 -7.76028574e-01 -4.34008241e-01
5.49531043e-01 4.11737740e-01 -1.02789767e-01 -1.35737136e-01
-1.37083268e+00 1.25463653e+00 5.78958273e-01 -1.27120137e-01
-1.55207300e+00 -9.46770608e-01 -4.61841792e-01 3.99271138e-02
1.45576686e-01 -1.15109634e+00 1.29382300e+00 -5.53078294e-01
-7.41677582e-01 1.08421850e+00 5.45154691e-01 -8.01843941e-01
6.73405647e-01 2.43532509e-02 -4.18103606e-01 1.46234453e-01
-4.12345082e-02 6.29839003e-01 5.28934658e-01 -7.48400152e-01
-5.81973255e-01 -3.42750996e-01 -9.46992636e-02 2.01823469e-02
1.94404930e-01 8.44133049e-02 -3.64179999e-01 -8.86494100e-01
-2.16212913e-01 -1.07245231e+00 -6.25408411e-01 -2.65878260e-01
-5.05611598e-01 -3.31690758e-01 8.67361665e-01 -6.23996794e-01
1.00980306e+00 -1.87795377e+00 -8.44360236e-03 1.86113760e-01
7.55837977e-01 6.37568355e-01 6.88439161e-02 -1.05059408e-01
-6.65459871e-01 2.48854518e-01 -8.63062516e-02 -2.20691800e-01
-5.61239064e-01 4.11786921e-02 6.10326789e-02 5.11019289e-01
1.10898808e-01 9.95701075e-01 -5.69652200e-01 -7.15618849e-01
3.88704509e-01 4.30555224e-01 -6.10197365e-01 3.26704293e-01
-2.78475016e-01 6.09956801e-01 -5.39643824e-01 5.55494726e-01
8.15088868e-01 -5.94616294e-01 -2.35343143e-01 -1.37826875e-01
7.74569660e-02 -6.39767945e-02 -4.95188385e-01 1.51103497e+00
-3.37090969e-01 3.49857092e-01 -1.83331713e-01 -1.07402909e+00
4.54580277e-01 8.03228796e-01 9.81048465e-01 -5.78973651e-01
5.05477488e-01 1.75424486e-01 4.53619182e-01 -7.00218320e-01
2.85376254e-02 -3.74144524e-01 8.74176994e-03 2.60120124e-01
1.29226476e-01 -2.42287919e-01 -9.78452340e-03 -9.91677865e-03
1.27669275e+00 -5.24575055e-01 4.12919194e-01 -3.21471125e-01
7.99137592e-01 8.03597271e-02 3.52620959e-01 6.97057366e-01
-4.04581755e-01 9.38243389e-01 4.86152500e-01 -9.73037899e-01
-1.00434101e+00 -1.02620113e+00 -3.74331743e-01 4.16666567e-01
-2.27127239e-01 -5.21186367e-02 -6.44216597e-01 -9.94234324e-01
-1.69670388e-01 3.55478495e-01 -9.45836544e-01 2.67253071e-02
-8.94033134e-01 -1.00995493e+00 9.84115422e-01 7.31271267e-01
5.16965687e-01 -1.34196532e+00 -8.39050531e-01 2.48056665e-01
-7.37420246e-02 -1.05033576e+00 -9.78848487e-02 3.04879427e-01
-8.58758688e-01 -1.28853309e+00 -1.11866105e+00 -8.09334993e-01
2.61165053e-01 -3.42307128e-02 1.16990435e+00 4.03815329e-01
-7.25816488e-01 1.51095688e-01 -3.95758837e-01 -8.13130438e-01
-4.08779800e-01 4.25735772e-01 -2.47953549e-01 -3.88137609e-01
1.53010702e-02 -3.73555571e-01 -8.86647105e-01 2.14967772e-01
-7.53794491e-01 2.30659544e-01 6.94023728e-01 8.03749561e-01
7.71311879e-01 -3.36631358e-01 4.63995874e-01 -1.04436076e+00
2.76450098e-01 -8.81383419e-01 -3.03889334e-01 4.26713899e-02
-5.13711274e-01 -3.97978872e-01 7.60403275e-01 1.29116490e-03
-7.12022781e-01 -4.53733243e-02 -6.95370615e-01 -8.38743925e-01
-2.56666720e-01 4.67167795e-01 4.90576804e-01 2.34805718e-01
6.39521956e-01 1.03059433e-01 -2.68034369e-01 -4.60201681e-01
-2.19774961e-01 4.61528748e-01 2.74344921e-01 -8.14790092e-03
4.51849222e-01 3.61950278e-01 2.81382352e-01 -7.21078336e-01
-9.82697010e-01 -5.62773347e-01 -5.77677488e-01 -2.85997391e-01
1.67164207e+00 -8.50720525e-01 -5.26727796e-01 6.00739837e-01
-1.38752460e+00 -7.11538689e-03 -2.51558304e-01 5.70196748e-01
-3.56348604e-01 -6.87246099e-02 -7.34223127e-01 -1.50374115e-01
-8.33961368e-01 -1.76425695e+00 8.95745277e-01 2.79798359e-02
-5.24466634e-02 -1.06643689e+00 3.52240622e-01 2.39872664e-01
6.71315610e-01 6.10388577e-01 1.27774954e+00 -1.21118009e+00
-3.45191240e-01 -3.38073075e-01 -4.23153281e-01 4.24952835e-01
-2.59021431e-01 -4.29831147e-01 -9.41346288e-01 -1.33559555e-01
3.30396034e-02 -4.14860338e-01 1.17456114e+00 8.62340152e-01
1.62924302e+00 1.52440578e-01 -3.93039107e-01 6.92965806e-01
1.51337087e+00 4.27889526e-01 4.32607472e-01 5.92045896e-02
8.76408458e-01 9.51203853e-02 2.15259641e-01 3.65256965e-01
1.90144971e-01 5.41316688e-01 9.92760360e-01 -4.95897859e-01
-5.43277562e-01 2.46522754e-01 -2.61166573e-01 7.83496916e-01
-3.33908409e-01 -5.79064846e-01 -1.35121226e+00 5.48930228e-01
-1.36383057e+00 -7.26224363e-01 -6.29107416e-01 1.52614272e+00
3.47512245e-01 2.18701795e-01 1.46499351e-01 -1.08227290e-01
5.53473234e-01 2.23807409e-01 -4.73306417e-01 -6.99113607e-01
1.27661332e-01 4.75539535e-01 5.18146217e-01 1.03580110e-01
-1.32967722e+00 5.68579674e-01 6.23045444e+00 8.82802963e-01
-1.33489573e+00 5.03098607e-01 9.30289328e-01 4.73460481e-02
5.40442485e-03 -5.29488623e-01 -7.96697736e-01 3.19032937e-01
9.07186031e-01 2.59519368e-01 -3.56826067e-01 9.07574654e-01
1.55866429e-01 2.76749320e-02 -9.72143054e-01 8.04260612e-01
1.30027875e-01 -1.62443709e+00 1.99852079e-01 -1.35704890e-01
8.39363992e-01 6.88570321e-01 1.07582331e-01 3.89390439e-01
1.81759804e-01 -1.42233753e+00 2.40056515e-01 4.25802886e-01
9.76360083e-01 -5.34706235e-01 1.28356993e+00 2.96804845e-01
-1.14664173e+00 3.97366146e-03 -3.90234679e-01 4.50813323e-01
5.21794111e-02 4.98856157e-01 -1.49426794e+00 6.52396619e-01
8.86171162e-01 7.64445961e-01 -4.68757004e-01 1.19367063e+00
8.81360397e-02 8.25254977e-01 -3.15632336e-02 -2.31371690e-02
5.11258006e-01 5.04057229e-01 6.55310869e-01 1.49213731e+00
2.15841666e-01 1.97852254e-01 2.34608456e-01 6.87466383e-01
-2.56345272e-01 8.07040259e-02 -6.95065916e-01 2.80097336e-01
-3.57137740e-01 1.17441571e+00 -7.94334531e-01 -6.49502873e-01
-2.20232561e-01 5.25155067e-01 -1.86941177e-01 -2.12162673e-01
-1.10470176e+00 1.30484626e-01 9.53753963e-02 2.90955812e-01
4.44494665e-01 2.37931117e-01 -4.71097618e-01 -8.46898019e-01
-6.06502235e-01 -7.54351497e-01 8.55377078e-01 -8.36197317e-01
-1.27707434e+00 9.75816309e-01 1.08853787e-01 -1.14607620e+00
-2.46747687e-01 -8.02624226e-01 -7.53186464e-01 8.57416213e-01
-1.42266726e+00 -1.22363687e+00 -4.18228209e-01 6.83064103e-01
8.40551257e-01 -3.44987035e-01 1.01580346e+00 3.63473892e-01
-3.80159378e-01 4.74524260e-01 -4.73039448e-02 2.36183897e-01
5.18883049e-01 -1.20842826e+00 3.47838908e-01 5.24545908e-01
-4.75951791e-01 -8.07713196e-02 2.89269328e-01 -7.28971422e-01
-7.70904779e-01 -1.43597472e+00 7.32419550e-01 -4.03460503e-01
4.77265745e-01 6.35966146e-03 -5.92532575e-01 7.10816920e-01
2.34253243e-01 2.32150853e-01 5.05133629e-01 -5.03184259e-01
-1.33811147e-03 1.43951610e-01 -1.26708651e+00 1.77167624e-01
7.03264356e-01 9.09945294e-02 -6.01895630e-01 4.59609121e-01
1.03498161e+00 -6.62325323e-01 -1.14004135e+00 7.83976257e-01
3.77510160e-01 -7.08104074e-01 1.17997050e+00 -6.63805306e-01
8.60277593e-01 1.68704078e-01 -2.89817542e-01 -1.13689649e+00
-3.81376714e-01 2.17476100e-01 3.13631177e-01 2.65448332e-01
5.27649283e-01 -1.79249257e-01 8.27620566e-01 -1.57755375e-01
-4.70168769e-01 -1.25979698e+00 -9.02103782e-01 -3.90323460e-01
3.88381362e-01 -6.66209042e-01 3.45084667e-01 7.98506320e-01
-7.37620711e-01 3.48873287e-01 -2.90258318e-01 -7.94832632e-02
2.65485764e-01 -1.51830643e-01 3.79694581e-01 -1.28900945e+00
-3.46110672e-01 -5.17442584e-01 -6.28401518e-01 -5.08265853e-01
-3.15284938e-01 -1.26002073e+00 -3.17842901e-01 -1.72959328e+00
6.38627827e-01 -2.92345941e-01 -7.91019738e-01 4.02583629e-01
4.94430959e-02 4.53017205e-01 3.80604059e-01 1.22151680e-01
-4.02051836e-01 5.03453426e-02 1.79952574e+00 -2.50266314e-01
1.12978779e-01 2.68616736e-01 -2.07186833e-01 8.52152407e-01
7.05093622e-01 -7.25566626e-01 -3.95816326e-01 -2.43793994e-01
7.71506429e-02 7.07191586e-01 5.58166027e-01 -1.21820700e+00
-3.24930958e-02 6.79078400e-02 7.03212321e-01 -1.22956514e+00
3.00572425e-01 -9.49240625e-01 1.01851933e-01 1.13716888e+00
-3.12075317e-01 3.88155758e-01 3.12164128e-01 2.98645377e-01
-1.88379720e-01 -1.17025390e-01 1.11848807e+00 -5.46813607e-01
-4.77985889e-01 7.83931613e-01 -5.65535963e-01 2.09944829e-01
1.21751451e+00 9.81658995e-02 -3.01037312e-01 9.40260589e-02
-9.11349654e-01 1.35798141e-01 -2.10688099e-01 3.05922240e-01
7.65247107e-01 -1.00160062e+00 -1.05730569e+00 3.99184152e-02
-1.08428374e-01 4.32256877e-01 7.35794365e-01 1.29663634e+00
-9.06008780e-01 7.83501863e-01 -4.86899972e-01 -9.43908453e-01
-1.33432460e+00 5.92913866e-01 7.86504447e-01 -1.23036981e+00
-7.27034450e-01 1.06585586e+00 6.59430385e-01 -2.97005683e-01
1.19255126e-01 -4.63361025e-01 -6.80872917e-01 -1.60882547e-01
4.01370525e-01 1.68093562e-01 4.33077753e-01 -6.08916640e-01
-5.96286595e-01 4.72404987e-01 -5.44126213e-01 2.09167048e-01
1.60207689e+00 5.31625271e-01 -3.93769108e-02 1.54069006e-01
1.42021859e+00 -6.78026795e-01 -3.97708923e-01 -7.59536847e-02
-1.97186351e-01 1.96447283e-01 1.30520552e-01 -1.30355167e+00
-1.41963124e+00 1.08423102e+00 1.15946877e+00 -5.13330065e-02
1.21828282e+00 6.19849470e-03 1.00317132e+00 -5.87936491e-02
-2.04966813e-01 -5.06956160e-01 3.05933058e-01 7.47710347e-01
1.02560759e+00 -1.12022591e+00 -6.45545349e-02 -1.14563564e-02
-6.85154378e-01 1.04632628e+00 7.52663136e-01 -4.58060414e-01
1.04796100e+00 2.45528862e-01 -5.08622415e-02 -7.22142458e-01
-7.90440917e-01 1.62271991e-01 4.59356368e-01 4.42156911e-01
4.95504797e-01 3.75843078e-01 -3.32121104e-01 8.53736699e-01
-1.14215046e-01 1.71646941e-02 4.09643322e-01 5.75373948e-01
-3.37833166e-01 -7.52429008e-01 -2.00966865e-01 8.34778905e-01
-9.50137377e-01 -2.29415357e-01 -2.18691275e-01 1.16515410e+00
5.16737759e-01 2.84110546e-01 -1.75571978e-01 -6.72738492e-01
4.23168540e-01 -1.11864567e-01 1.82154983e-01 -6.68679059e-01
-1.16363811e+00 7.90536329e-02 -1.07801050e-01 -6.06070578e-01
-3.19142133e-01 -5.10929286e-01 -1.17288256e+00 -1.20895870e-01
-1.39804661e-01 -2.19051093e-01 6.73798203e-01 8.76772285e-01
-8.29144940e-02 1.11246669e+00 4.80179429e-01 -5.23177505e-01
-5.81242919e-01 -1.09422100e+00 -4.60889280e-01 1.99314311e-01
3.66463661e-01 -5.85223675e-01 4.58008237e-02 -2.68249452e-01] | [15.370205879211426, -1.863943099975586] |
cabbdf36-8323-43de-bbe9-cf974c1547a0 | s3c-self-supervised-stochastic-classifiers | 2307.02246 | null | https://arxiv.org/abs/2307.02246v1 | https://arxiv.org/pdf/2307.02246v1.pdf | S3C: Self-Supervised Stochastic Classifiers for Few-Shot Class-Incremental Learning | Few-shot class-incremental learning (FSCIL) aims to learn progressively about new classes with very few labeled samples, without forgetting the knowledge of already learnt classes. FSCIL suffers from two major challenges: (i) over-fitting on the new classes due to limited amount of data, (ii) catastrophically forgetting about the old classes due to unavailability of data from these classes in the incremental stages. In this work, we propose a self-supervised stochastic classifier (S3C) to counter both these challenges in FSCIL. The stochasticity of the classifier weights (or class prototypes) not only mitigates the adverse effect of absence of large number of samples of the new classes, but also the absence of samples from previously learnt classes during the incremental steps. This is complemented by the self-supervision component, which helps to learn features from the base classes which generalize well to unseen classes that are encountered in future, thus reducing catastrophic forgetting. Extensive evaluation on three benchmark datasets using multiple evaluation metrics show the effectiveness of the proposed framework. We also experiment on two additional realistic scenarios of FSCIL, namely where the number of annotated data available for each of the new classes can be different, and also where the number of base classes is much lesser, and show that the proposed S3C performs significantly better than the state-of-the-art for all these challenging scenarios. | ['Soma Biswas', 'Jayateja Kalla'] | 2023-07-05 | null | null | null | null | ['class-incremental-learning', 'few-shot-class-incremental-learning', 'incremental-learning'] | ['computer-vision', 'methodology', 'methodology'] | [ 3.61027420e-01 8.51486772e-02 -2.46619266e-02 -2.33136043e-01
-4.16668743e-01 -3.34518224e-01 4.00473982e-01 4.65915918e-01
-4.31479543e-01 1.15593159e+00 -4.92486432e-02 2.14718208e-01
-1.24144800e-01 -8.23122561e-01 -6.12839997e-01 -8.51261079e-01
-1.36687204e-01 6.77864730e-01 8.72818708e-01 -1.13385297e-01
3.65357324e-02 3.48974407e-01 -2.19032693e+00 3.04344922e-01
9.67898965e-01 1.01038146e+00 2.57798135e-01 3.69196504e-01
-2.41405874e-01 7.60179639e-01 -5.34267783e-01 -1.41833305e-01
1.49875715e-01 -3.18338394e-01 -5.01088023e-01 3.54411662e-01
4.00182217e-01 -1.46047041e-01 -2.17615321e-01 9.34858680e-01
2.87525952e-01 4.52121705e-01 5.09695411e-01 -1.19355559e+00
-4.83215213e-01 4.64268595e-01 -5.14142871e-01 1.86881408e-01
-5.71943559e-02 9.96835902e-02 5.98296285e-01 -1.23944533e+00
5.93047798e-01 1.12851191e+00 6.94811583e-01 7.89854050e-01
-9.36246693e-01 -7.15052962e-01 4.73285973e-01 4.52050775e-01
-1.48871410e+00 -4.99996364e-01 6.80443764e-01 -2.61122197e-01
5.89316726e-01 1.46105677e-01 5.08808136e-01 9.18037176e-01
-1.15820244e-02 8.99709761e-01 9.28041875e-01 -4.51931566e-01
7.44119644e-01 6.44119024e-01 5.39157331e-01 5.93735874e-01
3.31519276e-01 1.53253049e-01 -6.13033533e-01 -1.52807176e-01
4.16690558e-02 5.79748929e-01 -1.25423655e-01 -5.91302514e-01
-8.07047129e-01 6.12897992e-01 2.31354862e-01 2.50664920e-01
-1.91103488e-01 -4.58137184e-01 3.96614552e-01 3.04115027e-01
5.22384226e-01 -1.30450511e-02 -7.63182759e-01 4.88871224e-02
-8.95292163e-01 4.65380354e-03 5.70649981e-01 1.25015581e+00
1.00614691e+00 4.07566354e-02 -3.37238133e-01 8.32523048e-01
-2.36346439e-01 2.19776213e-01 9.02455926e-01 -3.69530708e-01
4.63834822e-01 8.35499763e-01 2.18379140e-01 -5.48141360e-01
-2.10163444e-01 -8.45900238e-01 -7.61074007e-01 8.91657546e-02
3.19054663e-01 -8.83360803e-02 -1.27825880e+00 1.79117537e+00
5.18717885e-01 3.68305564e-01 2.12031707e-01 3.87346268e-01
5.19467175e-01 7.04823852e-01 -1.38594657e-01 -6.03013158e-01
7.86628425e-01 -1.02936733e+00 -4.55757022e-01 -5.07137418e-01
4.37909633e-01 -2.97858715e-01 1.10510433e+00 3.83641064e-01
-5.46364784e-01 -7.99395263e-01 -1.16819000e+00 3.87028754e-01
-5.14950156e-01 -6.38198033e-02 5.44169962e-01 4.96914685e-01
-6.62275791e-01 7.81290531e-01 -7.28195667e-01 -2.82655925e-01
7.65851974e-01 9.49670374e-02 -2.13001907e-01 -5.30363619e-01
-1.08369720e+00 5.40709198e-01 7.20697522e-01 -8.99047330e-02
-1.10836744e+00 -9.00660872e-01 -5.50009549e-01 2.35569760e-01
6.79428756e-01 -1.06231160e-01 9.90276694e-01 -1.10981429e+00
-9.42601085e-01 3.05432200e-01 -2.11829126e-01 -3.98892790e-01
6.58673048e-01 -1.95431277e-01 -4.75355119e-01 -2.64167309e-01
8.77422169e-02 6.27726793e-01 1.09431338e+00 -1.19679403e+00
-1.10333800e+00 -5.62084436e-01 -1.79623604e-01 2.83381939e-01
-8.02090228e-01 -7.88483799e-01 -2.78645128e-01 -5.44074118e-01
1.39049198e-02 -9.35168862e-01 3.49731334e-02 5.22041284e-02
-7.34049752e-02 -2.90919930e-01 1.22791290e+00 -7.24744275e-02
1.24582648e+00 -2.24881172e+00 -2.45675389e-02 -2.83874035e-01
7.41399154e-02 6.10164702e-01 -1.64418399e-01 3.31641585e-01
-1.32999808e-01 -3.04070979e-01 -3.04166406e-01 -3.18811268e-01
-5.00686407e-01 5.27227759e-01 -3.22231174e-01 2.24507824e-02
2.12448955e-01 4.36113358e-01 -1.26134944e+00 -1.49983108e-01
1.35033414e-01 2.68212020e-01 -3.38481128e-01 2.54755877e-02
-1.69379681e-01 4.09089744e-01 -1.10991247e-01 6.62554562e-01
7.10111737e-01 -7.41502941e-02 -1.63285702e-01 1.28293842e-01
1.79210901e-01 -2.50078648e-01 -1.49855983e+00 1.41055846e+00
-4.23247963e-01 1.44577235e-01 -4.43382323e-01 -9.49663460e-01
8.97060394e-01 1.58829659e-01 1.01791896e-01 -3.32819134e-01
-1.10081986e-01 1.55795828e-01 -3.49492654e-02 -2.68259048e-01
1.11728959e-01 -4.21277553e-01 1.22905098e-01 3.56458992e-01
5.06927669e-01 2.96736747e-01 4.47246164e-01 1.91111252e-01
1.06801009e+00 -1.19693622e-01 3.10045272e-01 -1.45979077e-01
5.09570181e-01 -2.16884807e-01 1.19235384e+00 1.03439713e+00
-4.27473933e-01 4.22311544e-01 2.26106383e-02 -7.58587480e-01
-8.76103342e-01 -1.11365902e+00 -1.16736367e-01 1.05532038e+00
2.12908000e-01 -1.29174590e-01 -4.08552289e-01 -1.10353041e+00
6.97388873e-02 1.16180933e+00 -8.02048683e-01 -6.23700500e-01
-1.25491038e-01 -8.56074870e-01 -1.61362186e-01 5.56693792e-01
4.06687558e-01 -1.04717457e+00 -7.57775784e-01 3.13516438e-01
1.49549663e-01 -8.45167518e-01 -3.57587218e-01 5.24453223e-01
-1.19885838e+00 -1.26395619e+00 -5.74541926e-01 -7.05933928e-01
9.10922229e-01 4.81207758e-01 7.53512442e-01 6.34600371e-02
-4.39035296e-01 2.30393246e-01 -6.07512772e-01 -7.22882807e-01
-2.88409233e-01 1.91556200e-01 3.02507699e-01 4.44869936e-01
4.64794278e-01 -5.39365768e-01 -3.65095645e-01 2.82250881e-01
-1.02043426e+00 1.01691172e-01 4.73288924e-01 1.14400947e+00
6.00702345e-01 8.17303061e-01 9.74510849e-01 -1.48966551e+00
7.80107304e-02 -6.39439642e-01 -3.33355755e-01 4.77655083e-01
-8.64383221e-01 4.41275798e-02 9.29097772e-01 -8.86731148e-01
-1.28583801e+00 6.28467724e-02 2.69343734e-01 -4.34170783e-01
-1.10398591e-01 2.22758904e-01 -1.47679448e-01 1.91643924e-01
7.76942372e-01 3.08742344e-01 -1.99783996e-01 -6.42187834e-01
1.92464828e-01 5.52381039e-01 4.79695141e-01 -4.53274637e-01
8.96637201e-01 5.60387850e-01 -9.73842293e-02 -7.28734434e-01
-1.37906957e+00 -3.87640834e-01 -7.88306355e-01 -1.35531649e-01
9.61007848e-02 -8.93428802e-01 -5.44402599e-02 6.81538403e-01
-5.66616297e-01 -1.33752018e-01 -9.60594118e-01 2.05913439e-01
-1.73176125e-01 2.86020130e-01 -4.19402689e-01 -1.00861406e+00
-3.80755991e-01 -9.04431701e-01 6.57272100e-01 5.79421461e-01
1.03077389e-01 -8.08432341e-01 -1.35395661e-01 -1.43833971e-02
4.43527579e-01 1.54202327e-01 1.03585863e+00 -8.70949507e-01
-3.74629855e-01 -7.46996582e-01 1.49148300e-01 5.46392500e-01
4.98355210e-01 -3.88852984e-01 -1.18964946e+00 -8.39231372e-01
1.83342651e-01 -5.82347631e-01 1.05592823e+00 -2.26259381e-01
1.03497744e+00 -3.12741816e-01 -3.56925726e-01 1.97827622e-01
1.47894561e+00 3.93243879e-01 3.55678439e-01 -1.16663026e-02
5.16667962e-01 5.48481286e-01 9.40481484e-01 6.66551352e-01
1.24761000e-01 3.29672247e-01 3.21478605e-01 1.85694039e-01
-3.82280052e-01 -4.55318481e-01 1.18453294e-01 9.48751807e-01
3.58941555e-01 -1.14688903e-01 -7.65615344e-01 8.71711433e-01
-1.94228363e+00 -7.20290244e-01 3.05422544e-01 2.60209489e+00
9.45778668e-01 4.48923409e-01 -1.54584661e-01 5.25802910e-01
8.64068210e-01 -2.44188011e-01 -1.12061048e+00 1.64726317e-01
-1.52878001e-01 3.40037376e-01 2.82560349e-01 3.74787092e-01
-1.09208155e+00 8.49445462e-01 5.44328880e+00 1.08314538e+00
-8.88953269e-01 3.47805321e-01 6.96773767e-01 -3.02662283e-01
1.52973622e-01 1.80530593e-01 -1.21841741e+00 5.84723413e-01
7.40118802e-01 -1.38252333e-01 4.12074178e-01 9.85157728e-01
-4.24185008e-01 -3.41177404e-01 -1.20119393e+00 8.06065559e-01
2.77115196e-01 -1.09506369e+00 1.09781735e-01 -4.78741616e-01
9.65277195e-01 -5.80234267e-02 -3.08507886e-02 7.09009886e-01
1.16789848e-01 -2.28348777e-01 7.91466117e-01 6.29896402e-01
7.85823822e-01 -9.46943164e-01 8.59708130e-01 9.09826994e-01
-9.73436117e-01 -6.80243552e-01 -6.69686556e-01 -5.05434722e-02
-3.36714745e-01 9.33201134e-01 -7.34325409e-01 3.24349910e-01
8.02516222e-01 7.24438012e-01 -9.28210855e-01 1.16556895e+00
-2.16885760e-01 6.78416073e-01 -3.25845480e-01 9.11052227e-02
-3.94411348e-02 2.24405795e-01 4.91443545e-01 8.36692750e-01
3.34571689e-01 8.36505890e-02 2.13116407e-01 4.22760636e-01
6.71612695e-02 -9.28894132e-02 -4.08343703e-01 2.21120104e-01
7.34725714e-01 1.06994748e+00 -8.30891967e-01 -6.06652737e-01
-3.71941358e-01 1.09053946e+00 5.14552474e-01 3.36232394e-01
-4.13158923e-01 -5.69098532e-01 2.49154612e-01 1.67284027e-01
4.57500160e-01 2.13951960e-01 -4.18197684e-04 -1.36591935e+00
1.30807728e-01 -4.89274532e-01 6.74637794e-01 -4.52383846e-01
-1.41058075e+00 6.45803511e-01 -6.45561367e-02 -1.40056551e+00
-2.60814596e-02 -2.13312134e-01 -5.22089005e-01 3.90208066e-01
-1.50298440e+00 -8.73191118e-01 -4.31760788e-01 4.91699576e-01
9.11675096e-01 -4.19964641e-01 7.59066164e-01 3.67864430e-01
-5.35217285e-01 8.47803056e-01 4.21184480e-01 -4.02122170e-01
7.21623838e-01 -1.01821196e+00 6.89585656e-02 7.56488025e-01
5.13918027e-02 2.80243963e-01 6.81109428e-01 -8.31267715e-01
-1.05939031e+00 -1.51639652e+00 8.42051983e-01 -4.14885700e-01
3.33447337e-01 -6.10476136e-01 -1.27313173e+00 4.55864489e-01
-4.15964246e-01 3.44989926e-01 5.38063109e-01 -6.29890934e-02
-3.43020201e-01 -6.12052679e-01 -1.33719659e+00 2.90814102e-01
9.48418975e-01 -1.37211204e-01 -5.42253077e-01 3.71286511e-01
7.24525988e-01 -1.19462267e-01 -3.96248728e-01 5.71307898e-01
3.33108246e-01 -8.94948840e-01 7.40718007e-01 -5.81201613e-01
-3.39372978e-02 -3.81517321e-01 -1.41611457e-01 -1.45029652e+00
-3.13560694e-01 -1.35275796e-01 -4.63564366e-01 1.51087880e+00
2.99814165e-01 -6.12698793e-01 9.21415329e-01 4.78180885e-01
2.09910870e-02 -7.82894731e-01 -1.09080994e+00 -1.05389428e+00
-1.22189187e-01 -1.39549911e-01 5.02569973e-01 9.50320542e-01
-2.84703434e-01 5.58865547e-01 -5.99346519e-01 -8.26017708e-02
8.01772416e-01 1.31689236e-01 4.75160480e-01 -1.51851153e+00
-2.37682447e-01 2.18719766e-01 -4.56029117e-01 -5.76605082e-01
-2.31831759e-01 -8.55480790e-01 1.29879028e-01 -9.64969635e-01
5.98919988e-01 -7.26467609e-01 -5.90079963e-01 8.25209081e-01
-5.82183838e-01 -8.87285173e-02 2.36089990e-01 2.88366884e-01
-8.01268458e-01 8.43195856e-01 9.97282684e-01 -2.05504641e-01
-4.50251073e-01 2.85746306e-01 -6.42636240e-01 7.51276672e-01
5.06788790e-01 -8.43148291e-01 -8.86003375e-01 -1.18017510e-01
-1.70827955e-01 -1.12833709e-01 -8.79890323e-02 -1.50871801e+00
3.68825465e-01 -1.07227281e-01 7.18022168e-01 -8.03014755e-01
2.27131784e-01 -9.53746796e-01 7.64494464e-02 5.05189776e-01
-2.85012662e-01 -4.87828493e-01 2.10286587e-01 1.14979875e+00
1.04586147e-02 -5.47327697e-01 1.17240322e+00 -1.96738750e-01
-7.01053679e-01 4.76483256e-01 -1.09733216e-01 2.58384407e-01
1.24927938e+00 -2.34444976e-01 -2.07516447e-01 -1.68416917e-01
-9.47988272e-01 3.28149945e-01 3.70582372e-01 6.05429292e-01
6.48503602e-01 -1.29656267e+00 -4.94098783e-01 4.46389556e-01
5.02772808e-01 1.72645524e-01 4.09825504e-01 3.77229422e-01
-6.74561635e-02 1.09141417e-01 -4.63856123e-02 -4.50648487e-01
-9.73898649e-01 1.13866484e+00 -9.45860669e-02 -3.05833220e-01
-6.63613915e-01 8.57663095e-01 2.97501683e-01 -4.02718633e-01
6.47746563e-01 -2.19027307e-02 -1.17553145e-01 2.43549168e-01
8.11732769e-01 5.35054505e-01 2.82280177e-01 -2.25177839e-01
-1.76310837e-01 2.02377111e-01 -7.72129118e-01 3.74887675e-01
1.51485276e+00 -2.49214172e-01 3.93129438e-01 9.85414684e-01
8.88401210e-01 -3.99393171e-01 -1.54231143e+00 -7.36341298e-01
1.44119292e-01 -4.94352877e-01 -1.47127524e-01 -1.00011575e+00
-9.22160387e-01 1.03470671e+00 1.08474493e+00 -2.63559371e-01
1.30270445e+00 -2.59623706e-01 9.24897373e-01 3.87359470e-01
7.98060238e-01 -1.29837382e+00 1.47308409e-01 4.52875286e-01
5.54149926e-01 -1.17006779e+00 -5.69524840e-02 -2.81500190e-01
-7.43347406e-01 1.03387094e+00 7.78169990e-01 1.08168319e-01
7.72660375e-01 -6.31373003e-02 -3.10477555e-01 1.48493007e-01
-1.09434521e+00 -1.67314157e-01 -1.44728730e-02 5.82839251e-01
-1.47813186e-01 -6.55999407e-02 -7.56593794e-02 7.10293293e-01
3.53733957e-01 1.29545361e-01 5.24203658e-01 1.24276209e+00
-7.64740407e-01 -1.01731992e+00 -1.68827429e-01 7.23441064e-01
1.05077680e-02 8.38177465e-03 -2.29947761e-01 4.98093814e-01
5.43384194e-01 8.40248883e-01 3.51824611e-02 -3.91014427e-01
4.04211611e-01 3.72225225e-01 3.35825056e-01 -9.95332479e-01
-1.65895179e-01 -4.60219085e-02 -3.53108674e-01 -1.90908313e-01
-3.80796529e-02 -7.12918401e-01 -1.01711690e+00 1.33061081e-01
-5.22684336e-01 2.69899994e-01 1.68414071e-01 8.02345932e-01
4.22443479e-01 4.33691204e-01 8.89741480e-01 -6.36698782e-01
-7.40323901e-01 -1.00481737e+00 -7.78484344e-01 5.32519579e-01
3.99879932e-01 -1.01167023e+00 -5.88763177e-01 -4.75669466e-02] | [9.902300834655762, 3.255535364151001] |
7527bc7e-0ee8-4473-8dba-667d40dac39f | supervised-discriminative-sparse-pca-with | 2001.03103 | null | https://arxiv.org/abs/2001.03103v2 | https://arxiv.org/pdf/2001.03103v2.pdf | Supervised Discriminative Sparse PCA with Adaptive Neighbors for Dimensionality Reduction | Dimensionality reduction is an important operation in information visualization, feature extraction, clustering, regression, and classification, especially for processing noisy high dimensional data. However, most existing approaches preserve either the global or the local structure of the data, but not both. Approaches that preserve only the global data structure, such as principal component analysis (PCA), are usually sensitive to outliers. Approaches that preserve only the local data structure, such as locality preserving projections, are usually unsupervised (and hence cannot use label information) and uses a fixed similarity graph. We propose a novel linear dimensionality reduction approach, supervised discriminative sparse PCA with adaptive neighbors (SDSPCAAN), to integrate neighborhood-free supervised discriminative sparse PCA and projected clustering with adaptive neighbors. As a result, both global and local data structures, as well as the label information, are used for better dimensionality reduction. Classification experiments on nine high-dimensional datasets validated the effectiveness and robustness of our proposed SDSPCAAN. | ['Chin-Teng Lin', 'Yu-Kai Wang', 'Zhenhua Shi', 'Jian Huang', 'Dongrui Wu'] | 2020-01-09 | null | null | null | null | ['supervised-dimensionality-reduction'] | ['computer-vision'] | [-9.33689699e-02 -5.63356578e-01 -8.16082209e-03 -4.09248710e-01
-3.14934582e-01 -5.87895632e-01 3.59919399e-01 3.07989240e-01
-1.76788028e-02 4.34916228e-01 6.08864188e-01 2.35532224e-01
-7.55602121e-01 -6.34895384e-01 -1.10566184e-01 -1.18300188e+00
-7.78407231e-02 2.12862700e-01 2.29725704e-01 2.13105202e-01
3.80347282e-01 8.80154610e-01 -1.62275469e+00 1.43931001e-01
8.89920831e-01 7.26245821e-01 2.93794703e-02 9.16673243e-02
-3.19180667e-01 3.34985554e-01 -1.25646085e-01 3.76779854e-01
3.85006696e-01 -4.65390921e-01 -2.40733325e-01 2.25566208e-01
1.31843075e-01 1.81458354e-01 -4.32204068e-01 1.27279580e+00
4.80609208e-01 2.93759137e-01 4.78663892e-01 -1.27099836e+00
-4.74171489e-01 1.87981755e-01 -8.67660582e-01 -2.98098065e-02
2.68188685e-01 -1.28000855e-01 6.49510682e-01 -9.57978904e-01
7.18845069e-01 1.35816514e+00 4.45368022e-01 1.78294390e-01
-1.45396900e+00 -5.21629453e-01 1.18378066e-01 2.00373068e-01
-1.50375831e+00 -3.23340803e-01 1.50633740e+00 -4.28460568e-01
4.59211528e-01 5.51158011e-01 6.55636668e-01 7.12300956e-01
-4.19826210e-02 5.19003689e-01 1.32645297e+00 -2.90959269e-01
4.55352068e-01 -3.11751738e-02 5.29503942e-01 4.44800466e-01
4.95596051e-01 -2.97326185e-02 -4.09972757e-01 -5.95397532e-01
3.63109320e-01 5.85907042e-01 -3.74386609e-01 -1.23679459e+00
-1.24755752e+00 7.85674870e-01 3.01732302e-01 4.96992737e-01
-3.37740183e-01 -3.57253879e-01 3.41957361e-01 1.12888902e-01
2.95752734e-01 2.46525645e-01 -1.11149117e-01 -9.70061943e-02
-8.34090471e-01 -9.81465578e-02 5.44911742e-01 8.07444453e-01
1.12273073e+00 -6.96827024e-02 1.00471877e-01 7.99532056e-01
4.49026138e-01 5.08577228e-01 8.36408913e-01 -8.85225236e-01
2.44332224e-01 1.14578462e+00 -2.53859103e-01 -1.88055873e+00
-7.66532779e-01 -1.18716232e-01 -1.31161153e+00 1.41994730e-01
4.71615791e-03 1.04510762e-01 -7.49411643e-01 1.49655449e+00
5.76138079e-01 1.69406846e-01 1.07573867e-01 8.90987456e-01
6.13974929e-01 5.81937313e-01 -1.41014606e-01 -6.34228408e-01
9.20769155e-01 -5.43168664e-01 -8.98977816e-01 1.25175253e-01
7.43810534e-01 -6.52175307e-01 1.09726906e+00 3.06146026e-01
-6.02491915e-01 -3.60150725e-01 -9.38042223e-01 -3.25431712e-02
-2.16352090e-01 1.38489336e-01 5.38864315e-01 5.73836148e-01
-6.13918900e-01 6.64467275e-01 -1.07530618e+00 -3.49276721e-01
2.92140305e-01 2.88248122e-01 -6.50720000e-01 -4.90721971e-01
-6.61141932e-01 2.11217254e-01 4.47421849e-01 9.28405672e-04
-2.01608270e-01 -4.18197542e-01 -8.63549531e-01 1.50752455e-01
9.76334587e-02 -1.12512030e-01 8.34228396e-02 -6.73933446e-01
-1.13412738e+00 3.31103861e-01 -4.36210155e-01 -1.49531867e-02
2.90733501e-02 -9.23886597e-02 -4.30739075e-01 2.43274510e-01
2.53995620e-02 2.08754808e-01 8.75570893e-01 -1.44876838e+00
-2.40356609e-01 -7.88240552e-01 -6.65300965e-01 3.09318572e-01
-7.78463304e-01 -2.27027297e-01 -3.05944800e-01 -6.06283605e-01
1.05181277e+00 -1.00712705e+00 -4.08615112e-01 -8.21293518e-02
-4.68723565e-01 -9.69596952e-02 1.68276381e+00 -7.33430803e-01
1.45751989e+00 -2.63240933e+00 3.36677551e-01 9.76696312e-01
3.70422274e-01 6.55382127e-02 -7.46469423e-02 4.08816218e-01
-4.49239135e-01 -1.58746950e-02 -4.70058233e-01 -2.51546979e-01
-3.44480425e-01 2.41570234e-01 4.16068807e-02 8.18012357e-01
-1.61250681e-01 4.04305190e-01 -9.21491683e-01 -5.60127079e-01
4.06775594e-01 5.25685191e-01 -4.08165902e-01 -1.98645648e-02
3.93942952e-01 5.26438534e-01 -5.38031757e-01 5.63417196e-01
8.71428907e-01 -2.25327853e-02 3.19334269e-01 -4.62749392e-01
-2.16042265e-01 -2.38864079e-01 -1.80964983e+00 1.80644333e+00
1.31924093e-01 3.50227863e-01 1.51710227e-01 -1.08892834e+00
1.32946074e+00 1.00537367e-01 1.01981974e+00 -5.83568931e-01
-2.93055296e-01 2.35386267e-01 -4.88288887e-02 -3.07853907e-01
1.82116091e-01 2.33430400e-01 2.02906340e-01 3.87284040e-01
-1.88239619e-01 2.95535088e-01 2.28064910e-01 4.10788745e-01
1.07548463e+00 -6.79297447e-02 3.82498294e-01 -4.71690625e-01
6.11451030e-01 5.03525557e-03 9.28661883e-01 1.34541661e-01
-9.18872207e-02 8.72157693e-01 6.56000376e-01 -4.50081885e-01
-1.02138019e+00 -7.19988942e-01 -2.39741907e-01 5.77515602e-01
3.51637721e-01 -8.47094536e-01 -4.93572921e-01 -7.37595975e-01
-1.04194060e-01 4.12111253e-01 -2.55477995e-01 -4.37298357e-01
-6.00509584e-01 -8.87128115e-01 2.67565693e-03 1.55127302e-01
2.89232373e-01 -5.98447561e-01 -2.73623671e-02 8.84506255e-02
1.49708763e-01 -5.31023502e-01 -3.98889631e-01 1.05655149e-01
-1.25756645e+00 -1.12109470e+00 -4.56234425e-01 -6.02933884e-01
1.09534848e+00 7.19534099e-01 4.40970361e-01 4.34206761e-02
-1.07390329e-01 2.49398202e-01 -4.31178242e-01 3.03584814e-01
-7.49654695e-02 -1.44621640e-01 4.81422931e-01 3.26387554e-01
6.03579640e-01 -9.85260963e-01 -4.90749270e-01 4.19121832e-01
-9.16933239e-01 -2.32185483e-01 6.08916104e-01 9.09532368e-01
9.41957295e-01 4.87532020e-01 1.75877601e-01 -9.74223852e-01
7.03705132e-01 -3.78456146e-01 -4.32572544e-01 9.43408981e-02
-7.83008814e-01 2.10605368e-01 7.97284186e-01 -2.91803807e-01
-9.49870646e-01 6.27057791e-01 2.58320481e-01 -7.99503207e-01
-3.85501206e-01 5.92914402e-01 -8.45756829e-01 -1.90318778e-01
7.32324183e-01 4.00125802e-01 1.70180917e-01 -9.54789102e-01
2.23861620e-01 5.98559737e-01 1.98524103e-01 -3.59138638e-01
8.84592533e-01 7.25588679e-01 5.14679611e-01 -1.00032663e+00
-2.73636580e-01 -8.94337416e-01 -1.10474718e+00 2.42420435e-01
4.39104438e-01 -6.26241326e-01 -2.62072384e-01 1.81277633e-01
-7.19393671e-01 6.45468831e-01 -4.46938425e-01 6.93141937e-01
-1.80084869e-01 9.17958677e-01 -1.49729237e-01 -5.55116057e-01
-3.85996103e-02 -9.75678027e-01 8.21791410e-01 4.84814011e-02
-1.51034310e-01 -8.64907920e-01 3.51557314e-01 -1.40223607e-01
2.29388326e-01 3.84803176e-01 1.11417043e+00 -7.81878650e-01
-2.04537913e-01 -4.16362435e-01 -1.82965040e-01 1.80541471e-01
6.44187748e-01 1.18775532e-01 -6.86462581e-01 -5.36653101e-01
2.19963506e-01 2.07911626e-01 8.22174132e-01 2.25532576e-01
1.09822237e+00 -5.28161943e-01 -6.64639533e-01 7.07202911e-01
1.24632001e+00 1.31816968e-01 5.63880861e-01 1.04869187e-01
1.18739617e+00 8.58130991e-01 5.89342713e-01 4.48976368e-01
-4.33092035e-04 5.17405987e-01 2.09663808e-01 -6.81238025e-02
-4.97651333e-03 -2.77708560e-01 2.51598090e-01 1.28338981e+00
2.92773787e-02 3.05532485e-01 -8.70655179e-01 3.93128812e-01
-2.05978942e+00 -9.59139347e-01 -4.70149010e-01 2.33416104e+00
4.45608318e-01 -1.98372483e-01 -6.77721435e-03 6.33954048e-01
8.85395765e-01 2.60307163e-01 -5.53794324e-01 -1.05053104e-01
-4.21740234e-01 -3.13397825e-01 2.49518752e-01 2.63592988e-01
-1.13517594e+00 4.98567790e-01 5.55051136e+00 8.57192934e-01
-9.68908489e-01 -1.28125697e-01 1.94926292e-01 -6.03733510e-02
-3.57069492e-01 2.65551746e-01 -4.94744778e-01 6.61642730e-01
6.23985052e-01 -6.55545741e-02 3.05639207e-01 1.26988864e+00
5.34803331e-01 7.97479972e-02 -7.27040887e-01 1.21152103e+00
5.88025153e-02 -1.00939369e+00 2.09632993e-01 2.36094326e-01
8.04077148e-01 -4.96123582e-01 -1.50346965e-01 -2.13541642e-01
1.30288377e-02 -6.31656826e-01 -1.84985414e-01 6.47082388e-01
3.44784081e-01 -1.01712728e+00 6.66007936e-01 3.97381157e-01
-1.18553078e+00 -3.48680988e-02 -6.12408638e-01 2.75042564e-01
6.76352382e-02 9.69793022e-01 -2.57532239e-01 4.88262028e-01
9.37263072e-01 1.38308156e+00 -7.78739512e-01 1.18017876e+00
-1.66643456e-01 4.85172600e-01 -5.85736632e-01 3.18876982e-01
2.80436799e-02 -7.61273682e-01 8.66222382e-01 9.24560547e-01
3.71475726e-01 1.40802458e-01 4.58012193e-01 4.74599391e-01
2.13948488e-01 4.63959545e-01 -7.21208990e-01 -2.79599726e-02
5.25826991e-01 1.27039373e+00 -9.06877458e-01 -9.92120951e-02
-5.58115304e-01 8.85212362e-01 1.91099197e-02 3.30265611e-01
-3.19855213e-02 -4.90501761e-01 8.06751728e-01 2.04473570e-01
1.02306828e-01 -7.46409655e-01 -3.44625652e-01 -1.12741172e+00
2.11531669e-01 -9.11829948e-01 6.58403754e-01 -3.08933169e-01
-1.28457177e+00 2.26780221e-01 -5.82092209e-03 -1.79375839e+00
1.89075898e-02 -2.98588455e-01 -4.75356579e-01 5.90621948e-01
-1.24031532e+00 -9.77468133e-01 -4.76684839e-01 9.72175300e-01
7.98006579e-02 -2.06473351e-01 7.93059170e-01 2.81826407e-01
-6.96322381e-01 1.00021049e-01 7.35691488e-01 -6.80086091e-02
7.18795538e-01 -1.18046021e+00 -4.07358289e-01 9.73623037e-01
2.61397570e-01 8.69393647e-01 3.71353239e-01 -8.55990589e-01
-1.67568552e+00 -9.98128772e-01 5.01058102e-01 9.15561169e-02
4.87785667e-01 -3.33840013e-01 -1.35956311e+00 3.92398685e-01
-2.75351942e-01 1.94679067e-01 9.97438610e-01 1.30255774e-01
-2.11450309e-01 -3.47813576e-01 -1.21614122e+00 4.26580131e-01
8.85333300e-01 -3.38031560e-01 -4.07173514e-01 4.18050230e-01
5.63021898e-01 4.74950485e-02 -1.06558120e+00 2.29617015e-01
3.43001455e-01 -1.08919191e+00 9.19882655e-01 -5.43083012e-01
-1.53859407e-01 -7.79076517e-01 -2.33597323e-01 -1.27615869e+00
-7.20113039e-01 -3.98473561e-01 -3.95607740e-01 1.39560497e+00
3.66174504e-02 -5.02857566e-01 9.54780996e-01 6.22669339e-01
-1.60454959e-02 -3.07662755e-01 -1.04336739e+00 -8.24586689e-01
-3.97916138e-01 -1.36057943e-01 6.19300008e-01 1.35148704e+00
1.16073608e-01 2.58835226e-01 -6.39743447e-01 4.13076997e-01
7.93994367e-01 2.96039492e-01 6.33949697e-01 -1.61144912e+00
2.16821164e-01 -2.41210818e-01 -9.68921304e-01 -5.66744924e-01
8.10934007e-02 -7.73119211e-01 -2.28027150e-01 -1.47859728e+00
1.05755560e-01 -5.31446517e-01 -4.96646583e-01 3.68327588e-01
-8.57758224e-02 2.80920994e-02 1.15297874e-02 9.74765301e-01
-4.58076328e-01 8.04740489e-01 9.54517961e-01 3.49781811e-02
-8.26512516e-01 -1.99500605e-01 -5.38425684e-01 7.40961373e-01
6.26418412e-01 -7.72355020e-01 -4.58284438e-01 -8.37960839e-02
-8.20134804e-02 -3.67769420e-01 -1.14487156e-01 -1.08983982e+00
4.85901594e-01 -3.61821324e-01 5.87515473e-01 -5.56108415e-01
3.52892965e-01 -1.29057753e+00 3.28411132e-01 3.79865885e-01
6.62381249e-03 -1.48266796e-02 -1.46492586e-01 8.35962772e-01
-5.67046046e-01 -3.54153216e-02 8.54645073e-01 7.63637200e-02
-7.91380703e-01 3.45466346e-01 1.82764810e-02 -4.48506027e-01
1.13160253e+00 -3.81161153e-01 -1.99492887e-01 -1.44377381e-01
-8.43567550e-01 2.33413801e-01 6.92001402e-01 2.72943914e-01
9.69059110e-01 -1.63517833e+00 -5.29389083e-01 5.21611691e-01
1.69373214e-01 4.43334393e-02 3.69517475e-01 1.02728307e+00
-3.20135713e-01 2.25865975e-01 -2.61029959e-01 -7.39015818e-01
-1.51805270e+00 1.09192443e+00 -1.49090737e-01 -3.36789750e-02
-9.29269850e-01 2.27129638e-01 -3.01848277e-02 -4.61994946e-01
7.94478282e-02 -4.70170081e-02 -5.35432637e-01 2.11887047e-01
5.44363141e-01 7.56728113e-01 -6.52711242e-02 -8.81752849e-01
-4.56549853e-01 1.03167856e+00 1.91294760e-01 2.88877338e-01
1.43189263e+00 -3.59034419e-01 -6.33891404e-01 3.96220356e-01
1.46856236e+00 2.24037409e-01 -1.03884661e+00 -3.85003984e-01
2.98160046e-01 -9.72570062e-01 3.21777672e-01 -5.27264066e-02
-1.19581807e+00 9.29515600e-01 7.08132625e-01 3.62814724e-01
1.32874143e+00 -2.88493127e-01 4.62576419e-01 5.63881755e-01
1.13670997e-01 -9.37565684e-01 -1.05913319e-01 1.11391500e-01
7.84897566e-01 -1.09976768e+00 4.02916074e-01 -5.68551481e-01
-6.16359472e-01 1.18646228e+00 4.79148716e-01 -2.60577977e-01
9.04716372e-01 -1.12621002e-02 -2.01001436e-01 -2.91032881e-01
-3.66429865e-01 -2.19574627e-02 4.11784947e-01 7.31110156e-01
1.83780655e-01 -1.19230583e-01 -4.57037717e-01 4.89296645e-01
2.23290380e-02 -4.22935575e-01 2.08249182e-01 1.20765448e+00
-4.47504014e-01 -1.18379092e+00 -6.49963915e-01 6.08442783e-01
7.40195960e-02 1.42818555e-01 -6.81321621e-01 5.67552149e-01
2.62420089e-03 7.75783062e-01 -4.54358757e-02 -3.42902273e-01
3.37764561e-01 8.79893377e-02 -2.97227263e-01 -4.92300302e-01
-3.75097096e-02 4.93126601e-01 -3.80163461e-01 -8.60174358e-01
-5.93121350e-01 -8.63038063e-01 -1.26447630e+00 -3.16616625e-01
-2.15092003e-01 3.63050461e-01 7.95511603e-01 6.38833046e-01
7.75829196e-01 6.08785301e-02 1.04266810e+00 -6.36611402e-01
-1.79718092e-01 -7.14588761e-01 -1.06795430e+00 6.03817105e-01
2.60386974e-01 -6.34374797e-01 -5.77899456e-01 5.76840388e-03] | [7.860804557800293, 4.3155951499938965] |
13165b0b-a87a-4f42-829a-86d8770394b2 | pedestrian-attribute-recognition-a-survey | 1901.07474 | null | http://arxiv.org/abs/1901.07474v1 | http://arxiv.org/pdf/1901.07474v1.pdf | Pedestrian Attribute Recognition: A Survey | Recognizing pedestrian attributes is an important task in computer vision
community due to it plays an important role in video surveillance. Many
algorithms has been proposed to handle this task. The goal of this paper is to
review existing works using traditional methods or based on deep learning
networks. Firstly, we introduce the background of pedestrian attributes
recognition (PAR, for short), including the fundamental concepts of pedestrian
attributes and corresponding challenges. Secondly, we introduce existing
benchmarks, including popular datasets and evaluation criterion. Thirdly, we
analyse the concept of multi-task learning and multi-label learning, and also
explain the relations between these two learning algorithms and pedestrian
attribute recognition. We also review some popular network architectures which
have widely applied in the deep learning community. Fourthly, we analyse
popular solutions for this task, such as attributes group, part-based,
\emph{etc}. Fifthly, we shown some applications which takes pedestrian
attributes into consideration and achieve better performance. Finally, we
summarized this paper and give several possible research directions for
pedestrian attributes recognition. The project page of this paper can be found
from the following website:
\url{https://sites.google.com/view/ahu-pedestrianattributes/}. | ['Shaofei Zheng', 'Rui Yang', 'Jin Tang', 'Bin Luo', 'Xiao Wang'] | 2019-01-22 | null | null | null | null | ['pedestrian-attribute-recognition'] | ['computer-vision'] | [-1.67146996e-01 -4.64813292e-01 -2.39152506e-01 -8.31267715e-01
-3.92413557e-01 -2.33625665e-01 6.59134150e-01 3.74806494e-01
-4.65268016e-01 8.79826784e-01 2.41613820e-01 5.18293232e-02
-1.61076263e-02 -9.86934245e-01 -5.42504847e-01 -9.21063423e-01
-2.76008751e-02 4.79660571e-01 4.29977886e-02 -1.89871460e-01
3.38780209e-02 5.22137225e-01 -2.04451895e+00 5.05349278e-01
5.97816527e-01 1.31287324e+00 -3.68953824e-01 6.17697299e-01
-1.77853867e-01 7.73136973e-01 -5.79581439e-01 -1.09058368e+00
1.83627799e-01 -1.87532101e-02 -7.86859989e-01 -3.90851311e-02
9.02751744e-01 -2.36252517e-01 -3.24957907e-01 1.02914464e+00
7.70775676e-01 1.76738486e-01 1.06771922e+00 -1.93856013e+00
-6.32072449e-01 3.06001127e-01 -4.19773668e-01 3.15882146e-01
2.76283234e-01 -7.70778060e-02 8.41866314e-01 -7.24887013e-01
-1.23082496e-01 1.28410363e+00 9.24899638e-01 6.09860837e-01
-7.87509620e-01 -6.41326547e-01 2.32045621e-01 9.21431005e-01
-1.51808500e+00 -4.32276368e-01 6.29502833e-01 -5.82402349e-01
6.35287642e-01 5.58765948e-01 5.99538088e-01 1.41280675e+00
1.38450965e-01 1.15526772e+00 1.07229173e+00 -3.13652337e-01
-2.62498498e-01 1.58671364e-01 6.28354192e-01 5.41694283e-01
2.66406953e-01 -1.52357191e-01 -1.35624871e-01 -1.75823584e-01
3.22727144e-01 1.33985691e-02 2.96022534e-01 -2.48380527e-01
-1.10614920e+00 6.46396339e-01 2.36551166e-01 -2.17000376e-02
-3.15328807e-01 -9.40144900e-03 9.42379594e-01 1.54404759e-01
4.42942709e-01 -2.33663961e-01 -2.70519823e-01 5.64265400e-02
-2.88463354e-01 1.84740365e-01 6.52647793e-01 1.31611121e+00
7.29203463e-01 4.01838496e-02 -2.00582638e-01 1.35698080e+00
2.28879735e-01 4.26878452e-01 9.32619348e-02 -8.63125265e-01
1.95060700e-01 3.94428879e-01 -1.87336624e-01 -9.85608339e-01
-4.38740849e-01 -2.69507114e-02 -1.28884089e+00 1.70678422e-01
4.93234217e-01 -2.08627418e-01 -4.86271262e-01 1.35442567e+00
3.39541614e-01 1.15984149e-01 -1.77006423e-01 6.72504842e-01
1.49763191e+00 5.28371930e-01 7.49993742e-01 2.13799581e-01
1.88337374e+00 -9.79295433e-01 -6.14564061e-01 -6.83457479e-02
2.32420549e-01 -8.65239143e-01 6.82084501e-01 2.54994005e-01
-7.38936603e-01 -1.03799999e+00 -8.92420590e-01 2.29063444e-03
-9.20889854e-01 4.32524145e-01 8.10593963e-01 1.03967142e+00
-1.11061466e+00 9.91796851e-02 -3.82979423e-01 -7.84274101e-01
8.25671434e-01 5.18480241e-01 -3.58768225e-01 7.49557316e-02
-1.32877326e+00 8.87650430e-01 4.41509485e-01 -1.06607191e-01
-7.72923350e-01 -2.16467306e-01 -8.15444946e-01 -2.22612843e-01
3.76589596e-01 -1.04920936e+00 1.13782191e+00 -8.53511512e-01
-9.58996296e-01 1.21193206e+00 -2.60534763e-01 -3.03711027e-01
4.37712431e-01 -4.86831367e-01 -5.02396643e-01 -1.80317640e-01
1.90551668e-01 7.19565094e-01 4.12077099e-01 -1.07844949e+00
-1.17286646e+00 -4.71724689e-01 3.47152710e-01 6.93239793e-02
-6.14917219e-01 4.19342816e-01 -2.13224560e-01 -6.05381429e-01
-5.45502961e-01 -7.71516442e-01 -1.44805297e-01 -2.39454489e-02
-4.80229050e-01 -7.90456295e-01 5.95760524e-01 -3.72561634e-01
9.94558454e-01 -1.96427703e+00 -3.32442731e-01 -1.29832298e-01
3.23901772e-01 4.62349176e-01 1.14110619e-01 3.34454179e-01
-1.77348092e-01 6.29181713e-02 2.27417182e-02 -3.63255531e-01
1.43441185e-01 1.97462797e-01 3.55229706e-01 6.25780523e-01
1.50450123e-02 8.29135656e-01 -5.99340737e-01 -9.55050111e-01
5.39576650e-01 6.56103611e-01 1.84117362e-01 1.89345598e-01
1.46979198e-01 2.61575520e-01 -6.48902893e-01 1.11874700e+00
8.26411486e-01 2.45320112e-01 -3.27972889e-01 -5.23475885e-01
-2.90060133e-01 -3.53875637e-01 -1.33211422e+00 8.76106560e-01
-2.46777266e-01 6.66125298e-01 -1.79687470e-01 -1.25135124e+00
1.05033648e+00 4.25823450e-01 7.41654217e-01 -3.73148084e-01
4.50659841e-01 2.44354963e-01 -2.97083706e-01 -6.88093305e-01
3.60813916e-01 2.49965683e-01 -1.38912767e-01 2.24591084e-02
5.39971739e-02 6.10789537e-01 5.02121866e-01 -2.04474807e-01
4.53633159e-01 2.41728485e-01 7.85443902e-01 -2.32397407e-01
1.32786059e+00 -1.68552071e-01 7.06208408e-01 7.56069899e-01
-6.59488738e-01 3.32395464e-01 4.26229060e-01 -1.34762394e+00
-1.28645480e+00 -8.02296162e-01 -3.51968348e-01 1.54972136e+00
-2.16562934e-02 -1.59942195e-01 -8.14487278e-01 -6.09871387e-01
-8.83573368e-02 2.56872982e-01 -6.42750621e-01 1.01952508e-01
-6.27576888e-01 -1.32500792e+00 8.71135950e-01 9.38852072e-01
1.17746401e+00 -1.17959380e+00 -4.12881941e-01 2.66068485e-02
-7.71250308e-01 -1.28351998e+00 -2.35390086e-02 -2.41438732e-01
-3.79202127e-01 -1.15055287e+00 -9.23538387e-01 -9.17413533e-01
2.71423906e-01 4.58052754e-02 1.42574358e+00 2.17397735e-01
-1.65407702e-01 6.21619761e-01 -4.85372335e-01 -7.30292141e-01
-1.57290652e-01 6.92555457e-02 3.35863113e-01 3.69796306e-01
1.16209388e+00 -4.13996637e-01 -5.34640014e-01 5.45123756e-01
-3.02384019e-01 -1.12857975e-01 8.67409527e-01 5.38483799e-01
5.88761151e-01 -2.64399558e-01 4.90144253e-01 -7.70728767e-01
3.56059164e-01 -4.86370802e-01 -2.88802475e-01 3.03192019e-01
-6.94082528e-02 -4.17249888e-01 5.36124706e-01 -1.63947836e-01
-9.19485271e-01 2.60106355e-01 -7.02647865e-01 1.22612659e-02
-1.21090889e+00 -2.38749117e-01 -9.60998178e-01 -8.00767988e-02
4.96552557e-01 2.02641293e-01 -1.39460698e-01 -3.77014339e-01
2.36489162e-01 9.73506927e-01 4.55720931e-01 -9.51873004e-01
1.75611869e-01 5.19725204e-01 2.77430564e-01 -1.08819687e+00
-9.60400462e-01 -8.42147052e-01 -1.30377185e+00 -6.90189958e-01
1.28904152e+00 -8.77636015e-01 -1.22050381e+00 1.06696332e+00
-1.24640191e+00 6.71861544e-02 7.20399022e-02 3.48021209e-01
-8.50697875e-01 4.69915092e-01 -6.22421086e-01 -7.53300250e-01
-3.64885241e-01 -1.21086526e+00 7.30376303e-01 3.87951583e-01
-8.63119587e-03 -1.04564369e+00 -3.97306234e-01 5.64558685e-01
4.00564969e-01 6.96074069e-01 7.99560070e-01 -1.09943068e+00
-2.66726822e-01 -8.31587687e-02 -4.74630475e-01 4.26914454e-01
-3.79841141e-02 7.14966804e-02 -1.21589851e+00 -7.67491683e-02
-4.58615810e-01 -3.54962707e-01 9.87254381e-01 6.57417119e-01
1.38102353e+00 -1.65524408e-01 -5.59460640e-01 6.89499676e-01
1.24166822e+00 2.95274526e-01 7.63709664e-01 9.21062827e-01
9.38605547e-01 9.26612079e-01 7.05257058e-01 4.65584636e-01
8.27579439e-01 7.86826670e-01 4.26230460e-01 -1.08164661e-01
-2.05428362e-01 4.72923338e-01 2.96821203e-02 7.16920555e-01
-8.08185577e-01 -3.46888632e-01 -1.03438950e+00 3.61534715e-01
-2.04680419e+00 -1.14135861e+00 -7.20010221e-01 2.07558560e+00
2.64100879e-01 -3.67327705e-02 8.42379570e-01 2.60429919e-01
1.13094854e+00 1.08734101e-01 -1.97268933e-01 -5.95678806e-01
-4.57688451e-01 -1.00806303e-01 3.65198672e-01 1.36884630e-01
-2.24034190e+00 7.13235974e-01 5.14455843e+00 8.79829526e-01
-4.50651824e-01 1.42528817e-01 1.13833773e+00 3.56782556e-01
6.59777164e-01 -5.00702262e-01 -1.22369468e+00 5.83877146e-01
9.27896619e-01 8.69985670e-02 -3.62659514e-01 9.67701137e-01
-1.39436543e-01 5.22916839e-02 -1.05399096e+00 1.16797078e+00
3.12449276e-01 -8.22544038e-01 4.25450534e-01 -4.59431037e-02
1.13199279e-01 -3.12086076e-01 8.10450390e-02 3.23071420e-01
1.93905085e-01 -8.88059437e-01 4.11512315e-01 7.44617283e-01
8.03030312e-01 -1.05927765e+00 1.25543833e+00 -1.62739400e-02
-1.76399302e+00 -1.56711608e-01 -7.03462660e-01 2.78647617e-02
2.53859106e-02 3.75589222e-01 -5.09669147e-02 8.40132892e-01
1.23625207e+00 9.95012820e-01 -8.34882915e-01 1.60205019e+00
-4.07631025e-02 2.41479591e-01 8.87456313e-02 -2.72389650e-01
2.68196821e-01 -3.60998392e-01 3.89891773e-01 1.71226251e+00
2.18618527e-01 2.56834865e-01 5.22034407e-01 2.37272114e-01
2.53413796e-01 1.60862476e-01 -6.16777003e-01 5.62377095e-01
5.02380073e-01 1.47959673e+00 -5.71510553e-01 -4.58149970e-01
-9.12631929e-01 6.95737302e-01 -3.01608741e-02 8.33043307e-02
-9.75455642e-01 -4.19851035e-01 9.55180407e-01 -3.54745495e-03
-1.74543969e-02 -6.93583712e-02 -1.63448244e-01 -6.79173589e-01
-8.90195295e-02 -9.38749969e-01 8.73708487e-01 -3.85482162e-01
-1.49707448e+00 7.95513988e-01 2.71605581e-01 -1.51499319e+00
1.48851201e-01 -9.98786509e-01 -3.71733993e-01 7.67822742e-01
-1.45892167e+00 -1.67502367e+00 -8.29210639e-01 8.42542529e-01
8.94831419e-01 -7.60403633e-01 8.08044732e-01 9.00146961e-01
-6.24331415e-01 7.85334289e-01 -9.77423787e-02 6.85864210e-01
8.87320936e-01 -1.23288572e+00 5.52085757e-01 4.77977872e-01
-1.15084946e-01 1.33570939e-01 4.70411956e-01 -3.67311269e-01
-6.44758403e-01 -1.25239921e+00 1.01310456e+00 -7.25219488e-01
3.50020438e-01 -1.83238149e-01 -7.39634275e-01 6.36272073e-01
5.60284913e-01 4.78919484e-02 1.20076418e+00 2.19554920e-02
-2.04890564e-01 -2.29014948e-01 -1.12522542e+00 2.77074665e-01
1.05043709e+00 2.07137167e-02 -2.63726532e-01 3.37563455e-01
1.32268012e-01 -8.38076994e-02 -9.38188374e-01 4.75334466e-01
6.93181515e-01 -1.19603789e+00 1.45308363e+00 -7.00838029e-01
2.54921496e-01 -1.81361109e-01 -3.54144633e-01 -1.11138737e+00
-8.03446710e-01 3.42310034e-02 1.80116948e-02 1.54522943e+00
9.63585973e-02 -6.11311376e-01 7.36618400e-01 3.94376040e-01
-3.70114982e-01 -9.11543250e-01 -7.70540476e-01 -8.18953753e-01
1.60495445e-01 -1.95040435e-01 7.18117356e-01 7.82916844e-01
-7.61764705e-01 3.26751292e-01 -5.13370395e-01 1.88081175e-01
1.12851357e+00 -3.51454675e-01 7.98889518e-01 -1.62493396e+00
3.22940201e-01 -7.23169267e-01 -9.95327294e-01 -8.30150068e-01
1.94872960e-01 -6.15121424e-01 -4.46368426e-01 -1.66994214e+00
5.17595530e-01 -6.18700981e-01 -3.86442989e-01 4.76483941e-01
-2.56196082e-01 4.15625870e-01 2.23183557e-01 2.61236250e-01
-9.20265079e-01 3.20921928e-01 7.97834516e-01 -3.71518195e-01
3.75909746e-01 4.74225968e-01 -5.47457218e-01 1.00407422e+00
1.27645457e+00 -1.67336747e-01 2.86820177e-02 -4.21461135e-01
-4.62516397e-02 -5.72582483e-01 5.94328463e-01 -1.47789085e+00
3.69369775e-01 -1.21483669e-01 9.02590156e-01 -6.86666846e-01
6.09800100e-01 -8.03611875e-01 -8.57731402e-02 1.97234198e-01
-5.14739789e-02 6.04034960e-01 1.68917671e-01 2.46683881e-01
-2.68374622e-01 -2.41368264e-01 1.02006245e+00 -3.55286449e-01
-1.44071794e+00 6.70188010e-01 -6.22713566e-01 -8.41969922e-02
1.40974092e+00 -4.77115750e-01 -4.02229995e-01 -2.49871716e-01
-8.86744857e-01 3.91252786e-01 -4.31531221e-02 4.54848498e-01
7.23058462e-01 -1.76560271e+00 -1.09779036e+00 -1.09719187e-01
3.27737004e-01 -4.90641803e-01 2.60060102e-01 8.50338578e-01
-4.28859413e-01 4.13466334e-01 -9.38257635e-01 -5.66356838e-01
-1.76487851e+00 6.74770176e-01 3.93033624e-01 3.42288256e-01
-2.71172762e-01 8.67877126e-01 9.14621428e-02 -3.42411608e-01
4.75706249e-01 5.75511098e-01 -1.14488900e+00 4.69753981e-01
5.37662387e-01 8.40877175e-01 -3.71598676e-02 -1.30026650e+00
-7.09026039e-01 7.90186644e-01 -1.23403460e-01 5.72047949e-01
1.04761708e+00 -2.41607845e-01 -2.71024942e-01 3.53083074e-01
1.07635593e+00 -4.91030514e-01 -8.45174968e-01 -1.40644357e-01
9.03272033e-02 -3.29626918e-01 -6.57662928e-01 -5.15232027e-01
-1.03293681e+00 1.03209972e+00 8.24658036e-01 1.61157638e-01
1.11237490e+00 7.88419098e-02 8.55333924e-01 4.94595230e-01
4.42264915e-01 -1.18248951e+00 -2.60703526e-02 4.86586332e-01
4.74187344e-01 -1.51762426e+00 6.10505305e-02 -6.14530981e-01
-7.53712237e-01 1.18955982e+00 9.99438941e-01 4.40625958e-02
7.17635334e-01 1.52294978e-01 2.30764240e-01 1.29456129e-02
-2.48935446e-01 -5.24077296e-01 2.65472859e-01 1.08073819e+00
6.43349946e-01 2.74428189e-01 -1.90120980e-01 4.55152720e-01
-1.71612203e-01 -3.32676113e-01 4.13025975e-01 5.92326164e-01
-5.02909243e-01 -1.37621951e+00 -6.75193667e-01 6.55115068e-01
-5.18010974e-01 3.56551073e-02 -4.90079045e-01 7.18454719e-01
7.19193399e-01 1.04483330e+00 1.70750707e-01 -6.01487756e-01
5.30135632e-01 7.35319406e-02 2.50404269e-01 -1.60027578e-01
-4.68920529e-01 -3.94641876e-01 4.69568253e-01 -3.50902736e-01
-8.87151182e-01 -7.72978961e-01 -4.18914020e-01 -8.26365352e-01
2.26837382e-01 -3.59524041e-02 2.12632880e-01 6.99501514e-01
3.11761368e-02 3.97369415e-01 3.17072183e-01 -6.51551187e-01
6.05230629e-02 -8.36409926e-01 -2.00221002e-01 4.67135072e-01
1.51043043e-01 -1.01508224e+00 8.06037709e-02 2.64240712e-01] | [14.480571746826172, 0.9678635597229004] |
adf313cc-7b67-4085-a0f5-21a7aa53497e | adversarial-attacks-on-deep-learning-based-3 | 2203.10183 | null | https://arxiv.org/abs/2203.10183v3 | https://arxiv.org/pdf/2203.10183v3.pdf | RoVISQ: Reduction of Video Service Quality via Adversarial Attacks on Deep Learning-based Video Compression | Video compression plays a crucial role in video streaming and classification systems by maximizing the end-user quality of experience (QoE) at a given bandwidth budget. In this paper, we conduct the first systematic study for adversarial attacks on deep learning-based video compression and downstream classification systems. Our attack framework, dubbed RoVISQ, manipulates the Rate-Distortion ($\textit{R}$-$\textit{D}$) relationship of a video compression model to achieve one or both of the following goals: (1) increasing the network bandwidth, (2) degrading the video quality for end-users. We further devise new objectives for targeted and untargeted attacks to a downstream video classification service. Finally, we design an input-invariant perturbation that universally disrupts video compression and classification systems in real time. Unlike previously proposed attacks on video classification, our adversarial perturbations are the first to withstand compression. We empirically show the resilience of RoVISQ attacks against various defenses, i.e., adversarial training, video denoising, and JPEG compression. Our extensive experimental results on various video datasets show RoVISQ attacks deteriorate peak signal-to-noise ratio by up to 5.6dB and the bit-rate by up to $\sim$ 2.4$\times$ while achieving over 90$\%$ attack success rate on a downstream classifier. Our user study further demonstrates the effect of RoVISQ attacks on users' QoE. | ['Farinaz Koushanfar', 'Seira Hidano', 'Mojan Javaheripi', 'Jung-Woo Chang'] | 2022-03-18 | null | null | null | null | ['video-denoising'] | ['computer-vision'] | [ 3.43563527e-01 -1.40477836e-01 -1.31064311e-01 -1.29384905e-01
-9.69309628e-01 -6.81503534e-01 1.92765165e-02 -9.40972492e-02
-2.40479618e-01 2.84813225e-01 -7.06542432e-02 -8.51029813e-01
4.08485867e-02 -8.26248348e-01 -9.31915581e-01 -8.80070508e-01
-7.91527987e-01 -5.85040331e-01 1.40609488e-01 -2.64258534e-01
1.66913062e-01 4.42975163e-01 -1.02156162e+00 2.88502216e-01
2.93381989e-01 1.58213496e+00 -4.28664118e-01 1.14985287e+00
6.33768439e-01 1.17515564e+00 -9.78240252e-01 -7.65412807e-01
6.75793231e-01 -4.17470187e-01 -3.49392056e-01 -5.62694371e-02
3.40618610e-01 -1.00185478e+00 -1.39123940e+00 1.44289553e+00
5.67596734e-01 -1.14505433e-01 3.38813365e-01 -1.54343438e+00
-4.32781130e-01 6.42115414e-01 -4.87313896e-01 6.73663378e-01
1.75539955e-01 6.10082686e-01 8.22100222e-01 -1.31550685e-01
2.66879722e-02 9.32957113e-01 4.68704134e-01 6.29592299e-01
-9.45621490e-01 -1.20160556e+00 3.29226181e-02 3.86438638e-01
-1.47620595e+00 -7.18955994e-01 6.36500895e-01 -1.18736841e-01
7.78233171e-01 6.09451115e-01 2.00227454e-01 1.09609890e+00
3.67190957e-01 4.76766318e-01 4.30099696e-01 -7.14762486e-04
3.56860340e-01 -1.75411761e-01 -4.95661050e-01 5.19115686e-01
-2.52639945e-03 2.84871399e-01 -3.20407212e-01 -4.32760030e-01
8.22668552e-01 -3.06478083e-01 -7.06611037e-01 1.54377699e-01
-5.91642022e-01 8.09086740e-01 1.56099871e-01 -3.07756066e-01
-2.41389289e-01 8.12133729e-01 7.27720857e-01 8.48453820e-01
1.82350367e-01 1.89272314e-01 -2.86068618e-01 -3.12772870e-01
-6.22385085e-01 -2.05587298e-02 6.70762658e-01 1.18840575e+00
4.28339513e-03 5.89487433e-01 -6.24166131e-02 4.38954979e-01
2.60361552e-01 7.23789215e-01 3.05967748e-01 -1.35658801e+00
8.46395016e-01 -4.15668488e-01 -9.64490771e-02 -1.38281631e+00
2.32792288e-01 -4.42582816e-01 -8.13844562e-01 2.38625452e-01
1.75755292e-01 -4.58865911e-01 -4.40698504e-01 1.94096994e+00
-1.52737424e-01 4.89578903e-01 3.26484025e-01 8.61832201e-01
4.05001968e-01 8.19309056e-01 1.70313343e-01 -5.27483940e-01
9.16266143e-01 -5.25859892e-01 -5.68454087e-01 -4.93487045e-02
4.75000262e-01 -7.63355672e-01 6.07739210e-01 5.48531175e-01
-1.48987579e+00 -3.59591424e-01 -1.32105827e+00 7.21819937e-01
4.07975554e-01 -7.17398942e-01 1.13864362e-01 1.16760397e+00
-7.32449591e-01 4.80607420e-01 -6.73917055e-01 4.42277163e-01
7.06707239e-01 5.38246214e-01 -1.31059997e-02 -4.75285342e-03
-1.51164401e+00 5.92920966e-02 -1.12254120e-01 -3.55393767e-01
-1.53386223e+00 -8.44413102e-01 -6.29541218e-01 3.01396161e-01
3.95805538e-01 -4.39191610e-01 1.32482445e+00 -1.03465772e+00
-1.32042968e+00 4.97303993e-01 5.52844167e-01 -9.03285384e-01
7.37086833e-01 -1.98337361e-01 -8.33946586e-01 7.54317760e-01
-3.48063439e-01 1.05463997e-01 1.25544393e+00 -1.07531428e+00
-6.72844350e-01 -1.54568464e-01 3.89101446e-01 -8.45522881e-02
-8.17551315e-01 2.71646410e-01 -4.35162753e-01 -1.10763764e+00
-1.05569430e-01 -6.97159052e-01 -5.49602062e-02 3.79705541e-02
-2.72948623e-01 4.71444041e-01 1.11941147e+00 -5.24421871e-01
1.40125132e+00 -2.47351480e+00 -1.03215955e-01 3.16194683e-01
4.53520715e-01 6.52689040e-01 -3.05767119e-01 3.49301279e-01
-1.14135131e-01 6.13085747e-01 6.87985942e-02 1.73180282e-01
-2.45884523e-01 -4.66457568e-02 -5.86907983e-01 7.60614932e-01
-2.14477584e-01 5.54694831e-01 -7.82063961e-01 -3.10248826e-02
-6.18728623e-02 4.96277750e-01 -9.48853433e-01 4.43899900e-01
4.93843807e-04 1.09120561e-02 -4.12520826e-01 5.58331788e-01
6.98635519e-01 1.15546301e-01 1.05315864e-01 -1.93484649e-01
5.12711465e-01 3.62217613e-02 -9.21243727e-01 8.59525561e-01
-4.44437057e-01 8.54443073e-01 4.00643855e-01 -1.07898593e+00
5.69795728e-01 6.34304285e-01 5.38242817e-01 -7.34623671e-01
5.21969259e-01 6.17898479e-02 8.03684890e-02 -5.84062099e-01
2.66962089e-02 -1.15142623e-02 -1.60073146e-01 3.97379965e-01
-1.59919009e-01 1.83009356e-01 -3.22315723e-01 3.78547847e-01
1.53052914e+00 -6.73960924e-01 -1.32136988e-02 -1.05535410e-01
3.53808343e-01 -6.52178705e-01 6.17473960e-01 1.08530033e+00
-6.53755903e-01 3.55471343e-01 8.36552858e-01 1.04519883e-02
-1.13544524e+00 -1.23953915e+00 1.53923005e-01 1.06635141e+00
3.66121501e-01 -3.43350530e-01 -8.57297421e-01 -5.97210467e-01
1.63862407e-02 5.57348013e-01 -9.93636698e-02 -7.57962108e-01
-5.64042211e-01 -4.66808468e-01 1.37876594e+00 4.35018629e-01
9.12059546e-01 -5.10746062e-01 -5.50551891e-01 1.29827604e-01
-1.71956822e-01 -1.45013165e+00 -6.55214429e-01 -2.62740076e-01
-5.53849399e-01 -9.41307068e-01 -2.93922156e-01 -5.50372422e-01
1.50189161e-01 3.98071140e-01 6.89677358e-01 1.80543467e-01
-1.06044121e-01 5.20873845e-01 -4.81112063e-01 -1.29899591e-01
-9.08553541e-01 -3.13987106e-01 2.53757417e-01 1.75439373e-01
1.49989858e-01 -8.43072712e-01 -7.15085804e-01 4.78954464e-01
-1.27628291e+00 -6.47292733e-01 1.42366558e-01 4.98176455e-01
1.84595257e-01 6.61258459e-01 7.86085665e-01 -4.60944504e-01
6.12559915e-01 -6.41063154e-01 -6.43830061e-01 -1.86378717e-01
-2.66277164e-01 -3.78632247e-01 1.06800973e+00 -6.53404653e-01
-3.71285617e-01 -6.72220409e-01 -4.55267727e-01 -9.81830716e-01
1.06484801e-01 2.61597335e-01 -5.25857270e-01 -3.38168055e-01
7.50311196e-01 2.74442464e-01 1.13305591e-01 7.72890821e-02
1.54135644e-01 9.92384136e-01 6.96529627e-01 -2.51974881e-01
1.15067565e+00 4.96148020e-01 -1.01825362e-02 -9.74721134e-01
-2.76493579e-01 5.16900904e-02 2.60717481e-01 -4.07071263e-01
4.92773294e-01 -1.08343172e+00 -1.02751470e+00 6.58079982e-01
-7.32752860e-01 -3.31803232e-01 4.91718017e-02 3.22792083e-01
-6.69138193e-01 7.68096447e-01 -9.18930888e-01 -5.61056077e-01
-4.60486382e-01 -1.28081214e+00 4.15844768e-01 -1.37674510e-01
3.04415584e-01 -6.38935685e-01 -7.30680645e-01 4.89823133e-01
5.88073432e-01 4.16005015e-01 8.19880188e-01 -5.58573127e-01
-6.67183399e-01 -4.09718692e-01 -5.19128330e-02 9.78774369e-01
-7.14198425e-02 -8.41668397e-02 -9.03408945e-01 -7.67664552e-01
4.98661041e-01 -3.09670210e-01 4.44813967e-01 1.87352046e-01
1.68918622e+00 -8.83067727e-01 9.49277654e-02 1.06990397e+00
1.38829625e+00 6.18884683e-01 1.03113079e+00 1.11986063e-01
4.83485132e-01 7.94199929e-02 4.54159707e-01 7.75682271e-01
-4.75699157e-01 4.90714759e-01 9.82559860e-01 3.44055891e-01
1.60558134e-01 -1.32587656e-01 7.42933333e-01 4.22448575e-01
2.61052638e-01 -8.77220273e-01 -5.59030890e-01 1.85902923e-01
-1.40552998e+00 -1.16870391e+00 1.23651996e-01 2.37288499e+00
4.19669092e-01 6.97624743e-01 3.33797708e-02 5.06625414e-01
6.98566198e-01 4.17027682e-01 -5.68983614e-01 -5.28447688e-01
9.20012370e-02 1.61164373e-01 1.01957190e+00 3.66392940e-01
-1.29733682e+00 7.60622799e-01 5.29536772e+00 9.90760863e-01
-1.30582988e+00 7.62149766e-02 7.88609862e-01 -5.07819891e-01
4.83561791e-02 -3.52726758e-01 -1.92203209e-01 6.91337287e-01
1.28694403e+00 -3.22268218e-01 6.35165393e-01 8.56000841e-01
5.01430154e-01 5.10458052e-01 -9.35081422e-01 1.09468889e+00
9.75700095e-02 -1.33680844e+00 1.70265213e-01 2.93534040e-01
2.97009140e-01 -1.49560645e-01 4.57901239e-01 1.63173988e-01
1.70636460e-01 -9.83468652e-01 6.98051393e-01 -4.82842848e-02
1.25756776e+00 -1.25826311e+00 5.30543864e-01 1.29799217e-01
-8.43018353e-01 -7.48894930e-01 -1.08916551e-01 2.01347470e-01
2.32419848e-01 1.76599652e-01 -2.18732059e-01 2.76347417e-02
6.89298153e-01 9.60558280e-02 -7.50813400e-03 6.65040016e-01
-2.82312900e-01 1.03542399e+00 -6.14552051e-02 2.61284381e-01
3.58724236e-01 4.08212423e-01 9.36317742e-01 1.21165764e+00
6.59032315e-02 5.83954871e-01 1.05576210e-01 2.05760017e-01
-6.63851142e-01 -1.75265670e-01 -4.95285034e-01 1.66019648e-01
7.86611974e-01 5.61449170e-01 -6.60187751e-02 -2.38825157e-01
-1.60911232e-01 1.08639359e+00 -3.82338196e-01 3.91897053e-01
-1.34535050e+00 -7.70307004e-01 1.39517438e+00 2.65673250e-01
4.63412136e-01 -2.93964863e-01 1.16108291e-01 -9.88555908e-01
3.68958601e-04 -1.39516282e+00 3.86305600e-01 -3.98878872e-01
-1.10359001e+00 3.34954381e-01 -1.52245238e-01 -1.29141104e+00
-1.35819867e-01 -2.27281287e-01 -4.05870467e-01 4.51591313e-01
-1.25976181e+00 -4.21441138e-01 -1.62131175e-01 7.98199892e-01
4.63473618e-01 -5.72727978e-01 4.69515741e-01 6.73464239e-01
-6.51542902e-01 1.38813078e+00 2.73224860e-01 6.22311652e-01
2.89215028e-01 -5.45703948e-01 3.94915462e-01 1.25004375e+00
-3.23210627e-01 2.53499269e-01 8.26833129e-01 -3.32090944e-01
-1.72370481e+00 -1.30111551e+00 5.98768890e-02 1.49549142e-01
7.22871482e-01 -3.66305143e-01 -7.45817304e-01 5.18106341e-01
5.50697446e-02 2.33555377e-01 7.63026178e-01 -7.69378364e-01
-7.59614229e-01 -3.13342422e-01 -1.50458562e+00 8.02545249e-01
9.96529400e-01 -7.39235222e-01 9.20565128e-02 1.99268907e-01
1.09581554e+00 -2.89248943e-01 -1.04713762e+00 4.31923777e-01
5.14931977e-01 -1.03532636e+00 1.16594660e+00 -1.02572691e+00
5.15255928e-01 1.06907688e-01 -8.03223848e-01 -9.62099195e-01
-2.39164934e-01 -1.24284792e+00 -4.14317548e-01 1.00530648e+00
2.54002661e-02 -4.01469946e-01 7.83518016e-01 3.55111361e-01
-5.62502444e-02 -5.55533886e-01 -1.08554828e+00 -9.65267301e-01
1.80927947e-01 -8.35335374e-01 3.49145710e-01 7.76170254e-01
-1.46579415e-01 -2.22826391e-01 -6.18146062e-01 7.79658794e-01
7.22159445e-01 -6.90946460e-01 7.34455824e-01 -1.36195242e-01
-7.39852250e-01 -4.91336703e-01 -9.24605191e-01 -1.24840748e+00
-1.28994927e-01 -5.12630403e-01 -3.45050365e-01 -6.04438186e-01
-1.72266707e-01 -2.67001241e-01 -4.52990711e-01 3.81915122e-02
1.02224611e-01 4.73174989e-01 4.51891094e-01 1.18446223e-01
-3.77747327e-01 2.13966593e-01 6.54553533e-01 -1.63443774e-01
7.19855651e-02 1.80444971e-01 -7.50785887e-01 5.51362276e-01
9.98344421e-01 -5.61096787e-01 -6.86823010e-01 -6.25830352e-01
2.30216272e-02 5.82504511e-01 3.28903139e-01 -1.17785227e+00
2.92349560e-03 -2.34190196e-01 -1.53971672e-01 2.98842221e-01
2.86768496e-01 -9.35993731e-01 -2.02153161e-01 7.68530607e-01
-4.93954867e-01 -8.87943283e-02 6.90802634e-02 7.96797931e-01
4.54051867e-02 -2.46399697e-02 1.31011295e+00 6.33819327e-02
-3.52843642e-01 4.30482447e-01 -9.12514925e-01 3.27612758e-01
1.30315876e+00 9.71217267e-03 -3.83178532e-01 -1.10701466e+00
-5.35642505e-01 7.05753416e-02 2.78269827e-01 4.14594918e-01
8.06156814e-01 -9.70120192e-01 -9.29239869e-01 3.21421564e-01
-1.39884770e-01 -7.44836152e-01 3.56410861e-01 3.31199050e-01
-7.82975793e-01 2.34907847e-02 -7.00812712e-02 -4.42000687e-01
-1.58297920e+00 7.42382824e-01 6.93977356e-01 1.58431485e-01
-4.44897175e-01 9.37834799e-01 5.10316063e-03 6.23443067e-01
6.57601953e-01 1.99155107e-01 3.15408587e-01 -5.55044711e-01
7.65566647e-01 5.45784950e-01 1.39841326e-02 -4.69018549e-01
-1.19094573e-01 -9.57474783e-02 -2.59068906e-02 -6.38650209e-02
8.99084091e-01 -2.28317216e-01 3.37404907e-01 -4.58301067e-01
1.72401989e+00 1.26733467e-01 -1.29078329e+00 -6.97274730e-02
-6.62176251e-01 -9.73635733e-01 3.62434775e-01 -3.69135082e-01
-1.55610836e+00 7.63700426e-01 8.08928013e-01 6.24949396e-01
1.46767402e+00 -3.21563989e-01 1.31617630e+00 1.81358874e-01
2.97308713e-01 -7.88491488e-01 4.18141216e-01 5.20100221e-02
4.87455070e-01 -7.89104998e-01 -2.12177545e-01 -3.08277786e-01
-4.40959930e-01 9.67031002e-01 4.82998699e-01 -2.06387699e-01
8.73807907e-01 5.31029642e-01 1.29431024e-01 7.38554150e-02
-6.96508169e-01 5.35894871e-01 -2.71669298e-01 5.09338975e-01
2.29430497e-02 8.30171928e-02 -3.25152352e-02 3.46445262e-01
-6.71526194e-02 -3.00397635e-01 9.58406627e-01 1.01358366e+00
-5.61751544e-01 -8.16460431e-01 -4.73485231e-01 4.80591327e-01
-1.21275449e+00 -6.70120567e-02 9.52168775e-04 3.77955973e-01
-7.20428377e-02 1.41004896e+00 1.11266464e-01 -9.29138362e-01
2.43373483e-01 -4.71922219e-01 8.03225413e-02 3.35212797e-02
-4.46997881e-01 1.38205662e-01 1.01826943e-01 -8.10813904e-01
1.66731820e-01 -3.51983905e-01 -1.01410604e+00 -9.04967487e-01
7.08796158e-02 6.28003627e-02 5.11267602e-01 6.84743226e-01
3.35392594e-01 6.09672964e-01 1.40704906e+00 -3.24215233e-01
-9.99660790e-01 -3.43727380e-01 -5.33338189e-01 5.15994191e-01
6.20902538e-01 -5.73679619e-02 -7.47935712e-01 4.65602688e-02] | [5.406497955322266, 7.89202356338501] |
490d0734-e60a-48ef-a009-98c8e2e97cda | attention-mixtures-for-time-aware-sequential | 2304.08158 | null | https://arxiv.org/abs/2304.08158v2 | https://arxiv.org/pdf/2304.08158v2.pdf | Attention Mixtures for Time-Aware Sequential Recommendation | Transformers emerged as powerful methods for sequential recommendation. However, existing architectures often overlook the complex dependencies between user preferences and the temporal context. In this short paper, we introduce MOJITO, an improved Transformer sequential recommender system that addresses this limitation. MOJITO leverages Gaussian mixtures of attention-based temporal context and item embedding representations for sequential modeling. Such an approach permits to accurately predict which items should be recommended next to users depending on past actions and the temporal context. We demonstrate the relevance of our approach, by empirically outperforming existing Transformers for sequential recommendation on several real-world datasets. | ['Romain Hennequin', 'Bruno Sguerra', 'Guillaume Salha-Galvan', 'Viet-Anh Tran'] | 2023-04-17 | null | null | null | null | ['sequential-recommendation'] | ['miscellaneous'] | [-2.69781679e-01 -5.98010063e-01 -6.70565903e-01 -3.36876273e-01
-4.79649194e-02 -5.71283638e-01 7.17748821e-01 2.60079931e-02
-2.90032804e-01 4.02534425e-01 7.31488466e-01 -4.92840499e-01
-4.97661918e-01 -7.21425056e-01 -2.70636380e-01 -3.52668345e-01
-2.78364629e-01 5.29506326e-01 3.80035043e-01 -3.66964459e-01
4.96626377e-01 2.32781753e-01 -1.40493667e+00 5.41716337e-01
5.09475172e-01 1.01020312e+00 2.68673241e-01 6.44582391e-01
-1.97459355e-01 9.10441339e-01 -4.64960411e-02 -5.27889311e-01
7.36428285e-03 -2.50901341e-01 -6.22465909e-01 -3.58883023e-01
3.29917639e-01 -5.99278033e-01 -6.87381029e-01 3.55299562e-01
1.75993711e-01 8.05170596e-01 6.83214128e-01 -1.08353341e+00
-1.36427236e+00 1.00600421e+00 6.24160580e-02 6.71533525e-01
2.93777019e-01 -2.71313697e-01 1.73240864e+00 -1.09497130e+00
1.55793577e-01 8.86351764e-01 7.72146285e-01 4.26998764e-01
-1.33916521e+00 -4.91738915e-01 1.03528595e+00 5.98585248e-01
-1.01096523e+00 -2.84790516e-01 7.37503171e-01 -4.01629299e-01
1.37524462e+00 3.70476097e-01 8.25573504e-01 1.52846599e+00
3.68113518e-01 9.25713718e-01 5.68047345e-01 -1.90702975e-01
2.24039838e-01 9.69614461e-02 6.74544275e-01 1.59531504e-01
-3.07332166e-02 3.32052797e-01 -6.83471084e-01 -4.18184608e-01
7.37505853e-01 8.00528884e-01 5.46886884e-02 -2.81443328e-01
-9.61279273e-01 9.32040632e-01 1.82258457e-01 5.81524253e-01
-5.63301206e-01 1.64123058e-01 2.23146632e-01 5.23877859e-01
5.93896866e-01 3.80044013e-01 -7.39424109e-01 -1.39001980e-01
-6.65574253e-01 2.17208520e-01 6.64831161e-01 8.07370126e-01
1.61010966e-01 1.12223960e-01 -4.76367921e-01 8.45022082e-01
5.27230322e-01 3.17030877e-01 6.86484098e-01 -5.83417654e-01
6.29972294e-02 2.15623692e-01 5.58168769e-01 -9.27093625e-01
-1.55959308e-01 -5.63014388e-01 -3.24472368e-01 -5.56634784e-01
2.71077663e-01 1.68144658e-01 -4.96082425e-01 1.52577055e+00
7.78225139e-02 4.37894076e-01 -1.62445262e-01 7.88198411e-01
5.14938951e-01 6.06583178e-01 3.32957387e-01 -2.03970730e-01
1.05968261e+00 -1.15444899e+00 -7.70160913e-01 -1.57685786e-01
3.10504109e-01 -7.25599945e-01 1.21772516e+00 5.22739649e-01
-9.10789728e-01 -4.92299408e-01 -7.08865821e-01 -8.07749201e-03
-1.46759272e-01 1.72071293e-01 9.96211827e-01 5.07679164e-01
-1.15969157e+00 9.82670844e-01 -9.97092843e-01 -4.84503597e-01
-6.30052015e-02 4.68745679e-01 1.58757001e-01 1.77150056e-01
-1.18904805e+00 6.94566131e-01 -1.37283042e-01 1.98079869e-02
-5.83504379e-01 -7.64567196e-01 -3.68981957e-01 3.56713414e-01
3.42792213e-01 -6.88769341e-01 1.76095736e+00 -9.02573228e-01
-1.92433441e+00 -1.45190418e-01 -2.88438499e-01 -6.53263867e-01
-9.05312970e-02 -7.64497101e-01 -8.11381936e-01 -3.55186135e-01
-3.32970649e-01 -1.67289227e-01 8.06923270e-01 -4.19982463e-01
-9.05505896e-01 -1.61328256e-01 3.35240543e-01 1.68925792e-01
-8.92726839e-01 1.45748397e-02 -5.05970061e-01 -9.04602230e-01
-2.31756046e-01 -9.01102185e-01 -5.15531123e-01 -2.82673776e-01
7.16354027e-02 -6.97388053e-01 5.73105574e-01 -2.85573989e-01
1.83998489e+00 -2.15564251e+00 2.23560557e-01 -4.30664085e-02
5.35664223e-02 -4.57071252e-02 -2.55959958e-01 7.87268519e-01
1.98011234e-01 -1.40411764e-01 6.51387513e-01 -4.55033362e-01
2.51662910e-01 2.09699541e-01 -6.96326673e-01 2.30348200e-01
-4.32647079e-01 9.39831436e-01 -1.04154694e+00 4.20545004e-02
3.17583978e-01 5.66751480e-01 -9.32922244e-01 4.32887495e-01
-3.79472554e-01 7.39528835e-02 -5.92561603e-01 1.79166004e-01
3.87243032e-02 -4.78080481e-01 6.34074807e-01 -2.76442826e-01
-1.31001994e-01 9.48908389e-01 -8.55454206e-01 1.33079183e+00
-5.91885746e-01 3.37343037e-01 -5.03583848e-01 -7.08380401e-01
6.02666795e-01 5.44093847e-01 7.05381095e-01 -9.33095217e-01
3.99028622e-02 -1.82756335e-01 -2.42230937e-01 -3.49773258e-01
8.67206335e-01 -1.15601078e-01 4.82270904e-02 7.83602417e-01
1.65235296e-01 6.97853327e-01 -7.73119479e-02 3.80319983e-01
8.39424968e-01 1.57905966e-01 3.63567084e-01 -2.39888698e-01
6.89344853e-02 -4.43103522e-01 5.60284436e-01 8.35019708e-01
-7.34188855e-02 2.28837684e-01 3.29354927e-02 -7.91022182e-01
-7.99591243e-01 -9.19142783e-01 3.47535402e-01 1.82978606e+00
5.11741033e-03 -8.77048731e-01 1.95957527e-01 -9.85110819e-01
1.51947245e-01 1.12171173e+00 -9.28671420e-01 -2.23402202e-01
-4.37143207e-01 -4.85641330e-01 -2.06908986e-01 9.55100238e-01
-5.70385635e-01 -8.17541420e-01 -2.15622231e-01 5.66596985e-01
-1.02322452e-01 -6.95585847e-01 -1.12184286e+00 2.20843940e-03
-1.09754038e+00 -8.44516218e-01 -2.99927205e-01 -2.59826928e-01
1.25224395e-02 6.07271910e-01 1.22345161e+00 -8.44062790e-02
4.23071593e-01 8.49921823e-01 -5.98364234e-01 8.87663662e-02
2.68569559e-01 7.93105271e-03 4.68592465e-01 2.58262932e-01
5.40212989e-01 -9.91486013e-01 -9.18729484e-01 5.54011703e-01
-5.82180262e-01 -1.76164970e-01 3.57202262e-01 6.55878186e-01
3.57866853e-01 -3.49734753e-01 4.21723425e-01 -1.08427739e+00
6.73054993e-01 -9.84261692e-01 -3.29805434e-01 2.72519648e-01
-1.10714579e+00 2.28758194e-02 8.85023355e-01 -1.05597305e+00
-1.04253352e+00 -4.87306565e-01 -2.69940346e-01 -3.73715103e-01
1.56523913e-01 5.68566680e-01 1.69433668e-01 3.10121000e-01
3.17830384e-01 4.21187937e-01 -6.42655969e-01 -1.03543019e+00
5.18974006e-01 3.04137498e-01 1.00528128e-01 -5.54556489e-01
2.20790580e-01 2.63424605e-01 -6.87535822e-01 -4.32235032e-01
-9.26323712e-01 -5.37634194e-01 -3.70872229e-01 1.88777484e-02
3.25525790e-01 -5.61220407e-01 -7.47887909e-01 -8.55385512e-02
-6.46045566e-01 -4.48523402e-01 -3.09440941e-01 8.18925679e-01
-2.78337449e-01 3.99012119e-01 -9.36638236e-01 -8.27677786e-01
-4.56626713e-01 -6.70285404e-01 3.68341953e-01 -9.28195193e-02
-3.74025345e-01 -1.29011619e+00 4.62191284e-01 -2.41153836e-02
8.89428914e-01 -7.11492002e-01 6.79019272e-01 -9.87563848e-01
-3.14562172e-01 -1.46077901e-01 1.03991814e-01 -2.17351075e-02
2.14323133e-01 8.40631798e-02 -4.70928371e-01 -3.75108093e-01
-2.68006444e-01 1.89131603e-01 5.65557361e-01 4.48963255e-01
9.51723874e-01 -5.16045034e-01 -3.56817871e-01 4.28222388e-01
1.28931034e+00 4.01354700e-01 2.99841106e-01 1.84259832e-01
6.12066507e-01 7.95381069e-02 5.53424597e-01 7.17947125e-01
6.89353406e-01 1.06517148e+00 1.26106650e-01 6.31347656e-01
1.41342103e-01 -5.79979181e-01 6.02984309e-01 1.01243281e+00
-2.55659252e-01 -5.44538438e-01 -3.16285759e-01 6.40712202e-01
-2.10056067e+00 -1.16380692e+00 1.87618695e-02 2.14628530e+00
3.70123088e-01 5.50107807e-02 4.23161447e-01 -2.89649904e-01
2.14087799e-01 8.18241164e-02 -4.18204993e-01 -4.22525972e-01
3.18248630e-01 9.76225436e-02 2.74396390e-01 4.96050447e-01
-1.02740026e+00 8.27759147e-01 7.56652594e+00 5.78063846e-01
-1.01179814e+00 3.66322875e-01 1.80384770e-01 -5.26800215e-01
-6.26973212e-01 -2.40607839e-02 -9.04869318e-01 5.55482328e-01
1.33685160e+00 -2.67338872e-01 4.41287696e-01 8.89714181e-01
1.75246492e-01 4.67011392e-01 -1.32975852e+00 6.75071359e-01
6.97678551e-02 -1.19044065e+00 1.76454753e-01 1.05981700e-01
5.97489238e-01 6.95675239e-02 4.67954516e-01 6.16482019e-01
8.86651099e-01 -5.84493995e-01 7.46236026e-01 7.15881586e-01
2.49766022e-01 -5.30747592e-01 4.12423253e-01 1.67492524e-01
-1.33383346e+00 -5.12033582e-01 -3.22456688e-01 -3.25394511e-01
2.74259239e-01 4.42043662e-01 -5.22281170e-01 4.99103576e-01
8.06916833e-01 1.18120921e+00 -2.50194192e-01 9.42554235e-01
-1.54915482e-01 1.08342195e+00 -2.21768409e-01 -1.65238127e-01
2.62751639e-01 -2.55756259e-01 3.99549603e-01 1.18451202e+00
5.38000882e-01 3.97188008e-01 2.58551627e-01 1.64364025e-01
1.96290314e-01 4.14639801e-01 -4.56950188e-01 -2.25991324e-01
4.86147225e-01 9.63057399e-01 -4.06696677e-01 -3.17134798e-01
-8.25142145e-01 1.07545662e+00 5.50742865e-01 5.27812064e-01
-8.21308732e-01 3.76601219e-01 9.19987857e-01 1.53323859e-01
9.68504965e-01 -4.09794182e-01 8.40769783e-02 -1.47072494e+00
-3.42122376e-01 -6.82709813e-01 7.88830340e-01 -3.91304672e-01
-1.53567576e+00 5.49639761e-01 -1.51430547e-01 -1.21010423e+00
-2.30143107e-02 -4.08798128e-01 -7.16654778e-01 4.97785479e-01
-1.16121161e+00 -1.19024587e+00 3.68990570e-01 6.94170654e-01
8.08084130e-01 -1.09573983e-01 8.77928197e-01 4.75333899e-01
-5.35059333e-01 7.15301514e-01 4.35659915e-01 -4.48368967e-01
7.66043067e-01 -1.31717789e+00 5.57005644e-01 6.02672935e-01
5.58873117e-01 1.07177460e+00 8.44986975e-01 -3.54012370e-01
-1.46636164e+00 -9.29574430e-01 1.26909506e+00 -7.70767748e-01
9.82198060e-01 -1.93895817e-01 -6.92380965e-01 1.23974884e+00
2.49580249e-01 -2.04744533e-01 1.30841386e+00 8.65350485e-01
-6.01221204e-01 -1.76809445e-01 -6.01944625e-01 7.46832788e-01
1.22308660e+00 -6.54404402e-01 -7.39727139e-01 1.04653664e-01
9.20604169e-01 1.98126763e-01 -9.96586263e-01 -6.07835129e-02
1.05275583e+00 -8.52341712e-01 1.12929463e+00 -1.03942645e+00
1.60158664e-01 3.41506749e-02 -5.09481013e-01 -1.26813436e+00
-1.32468975e+00 -5.98262429e-01 -8.10633898e-01 9.57319915e-01
3.85590076e-01 -4.95434582e-01 6.68358684e-01 7.10517287e-01
-9.03356001e-02 -7.62133241e-01 -5.59330106e-01 -5.35691798e-01
-1.57421455e-01 -6.74135864e-01 8.15027356e-01 9.63696718e-01
3.25777024e-01 4.48081404e-01 -1.08314216e+00 1.29403964e-01
3.06860179e-01 4.98766392e-01 3.73240471e-01 -1.22079933e+00
-9.53337491e-01 -4.68956143e-01 9.11250934e-02 -1.60626554e+00
-1.76645741e-01 -6.10280752e-01 -3.55342984e-01 -1.31086123e+00
1.43979207e-01 -4.29133922e-01 -1.25511169e+00 3.38569760e-01
-1.97540015e-01 2.76615828e-01 -5.88097759e-02 2.15160459e-01
-9.65292811e-01 7.04586148e-01 7.64146268e-01 -2.78168521e-03
-2.66515791e-01 4.26825523e-01 -9.38497186e-01 2.58794427e-01
7.32597411e-01 -3.29304367e-01 -8.16714048e-01 -6.19115591e-01
4.79084134e-01 6.55114204e-02 -7.99454078e-02 -5.26789129e-01
3.68767172e-01 -5.88281631e-01 1.32517099e-01 -6.53869033e-01
3.20888221e-01 -1.00423598e+00 3.89252692e-01 6.44850805e-02
-5.26459336e-01 3.89283866e-01 1.43518254e-01 1.02292633e+00
2.72379130e-01 1.84871867e-01 1.29868329e-01 1.32803291e-01
-8.60067904e-01 6.69194937e-01 -7.87448883e-01 -5.34716845e-01
6.48789823e-01 1.11845918e-01 1.56616718e-01 -4.01240945e-01
-1.10584974e+00 1.00427918e-01 1.62714675e-01 7.90787280e-01
7.60648489e-01 -1.48145366e+00 -2.95533359e-01 1.44291654e-01
1.11743212e-01 -1.19074094e+00 4.95424360e-01 9.30488706e-01
3.72299254e-01 6.49969876e-01 3.91561128e-02 -4.97596338e-02
-1.28662825e+00 1.08647227e+00 1.42805859e-01 -4.88689214e-01
-8.17401230e-01 8.94960701e-01 1.15678921e-01 -2.22575888e-01
2.12258950e-01 -5.53573251e-01 -7.22703278e-01 1.95850302e-02
8.07636321e-01 2.67264545e-01 -1.34175643e-01 -2.01340362e-01
-2.30297118e-01 1.22314170e-01 -5.28113425e-01 -9.04646292e-02
1.46864700e+00 -4.80717152e-01 1.79734871e-01 7.84302890e-01
7.72154391e-01 2.61616439e-01 -1.11261785e+00 -8.07362854e-01
1.23774052e-01 -6.64317310e-01 2.35567719e-01 -7.29215086e-01
-9.58753705e-01 4.88110334e-01 4.02002335e-01 3.87540281e-01
8.94196212e-01 -2.73236275e-01 9.29349184e-01 4.11614597e-01
4.34098452e-01 -8.24812591e-01 9.18487832e-02 5.66204309e-01
5.39229333e-01 -8.49887192e-01 -3.78115661e-02 9.19272751e-02
-6.99831784e-01 1.01353467e+00 4.16859806e-01 -2.82591194e-01
1.12257111e+00 -1.11458808e-01 -1.32099479e-01 -1.37097957e-02
-1.49586225e+00 6.45342246e-02 6.37985826e-01 3.01364273e-01
7.99167275e-01 2.25203604e-01 -3.58474761e-01 1.34376252e+00
1.73011810e-01 1.59977466e-01 1.16961159e-01 7.54165888e-01
-4.08199847e-01 -1.46803677e+00 1.22039430e-01 6.35349631e-01
-5.02061665e-01 -3.30348134e-01 -1.39239207e-01 1.90669060e-01
-2.02722132e-01 9.56117094e-01 1.11552395e-01 -8.28926444e-01
3.71780336e-01 1.59348294e-01 4.27109718e-01 -6.08291268e-01
-8.05833161e-01 1.84684172e-01 9.25198570e-02 -8.22702348e-01
-1.86388008e-02 -7.74277389e-01 -4.78457063e-01 -5.04285395e-01
-3.50335747e-01 5.70358038e-01 1.14057496e-01 9.82303917e-01
5.63553393e-01 4.44689244e-01 7.40707338e-01 -8.01906586e-01
-6.38527870e-01 -9.55823183e-01 -5.98921180e-01 3.48345131e-01
3.20969850e-01 -7.96602130e-01 -8.66806656e-02 -2.09269971e-01] | [10.161585807800293, 5.678708076477051] |
2cff65cc-aa1a-4ffb-a62c-55d5ed0fde7c | analysis-of-constant-q-filterbank-based | 2211.16363 | null | https://arxiv.org/abs/2211.16363v1 | https://arxiv.org/pdf/2211.16363v1.pdf | Analysis of constant-Q filterbank based representations for speech emotion recognition | This work analyzes the constant-Q filterbank-based time-frequency representations for speech emotion recognition (SER). Constant-Q filterbank provides non-linear spectro-temporal representation with higher frequency resolution at low frequencies. Our investigation reveals how the increased low-frequency resolution benefits SER. The time-domain comparative analysis between short-term mel-frequency spectral coefficients (MFSCs) and constant-Q filterbank-based features, namely constant-Q transform (CQT) and continuous wavelet transform (CWT), reveals that constant-Q representations provide higher time-invariance at low-frequencies. This provides increased robustness against emotion irrelevant temporal variations in pitch, especially for low-arousal emotions. The corresponding frequency-domain analysis over different emotion classes shows better resolution of pitch harmonics in constant-Q-based time-frequency representations than MFSC. These advantages of constant-Q representations are further consolidated by SER performance in the extensive evaluation of features over four publicly available databases with six advanced deep neural network architectures as the back-end classifiers. Our inferences in this study hint toward the suitability and potentiality of constant-Q features for SER. | ['Goutam Saha', 'Md Sahidullah', 'Shefali Waldekar', 'Premjeet Singh'] | 2022-11-29 | null | null | null | null | ['speech-emotion-recognition'] | ['speech'] | [-2.67345756e-01 -2.25581706e-01 -9.12167951e-02 -3.65499288e-01
-9.18933511e-01 -4.24530834e-01 1.98101774e-01 4.67753708e-01
-5.34114480e-01 6.74132526e-01 5.26227593e-01 -3.53893377e-02
-2.89873481e-01 -4.39591080e-01 -4.63653244e-02 -6.07565761e-01
-7.14688659e-01 -7.24348187e-01 -5.46299741e-02 -7.34311163e-01
8.58259201e-03 5.81680536e-01 -2.02682495e+00 5.34193814e-01
5.88902831e-01 1.47531509e+00 -4.67905626e-02 9.46123362e-01
6.77668527e-02 2.91991383e-01 -1.22738385e+00 -1.92687720e-01
-2.10821331e-01 -3.87504250e-01 -5.05431354e-01 -3.95478427e-01
-6.45340234e-03 1.82753250e-01 -1.25732124e-01 8.36775601e-01
8.45648885e-01 5.84599376e-01 3.93010408e-01 -1.02785325e+00
-6.03152275e-01 2.44025126e-01 2.11049896e-03 1.02263415e+00
7.61105895e-01 -2.73382604e-01 1.13674343e+00 -9.79897797e-01
1.75151914e-01 1.22474456e+00 1.23583269e+00 1.62763342e-01
-6.90890908e-01 -6.02047384e-01 3.80169824e-02 4.67067242e-01
-1.15774143e+00 -4.02481467e-01 1.09664965e+00 -4.63516749e-02
1.75148904e+00 3.53478104e-01 9.19151306e-01 1.18212211e+00
6.52515471e-01 3.99482161e-01 9.86366987e-01 -6.03820920e-01
5.75532317e-02 -6.80170506e-02 1.30502507e-01 2.58468419e-01
-4.57260996e-01 4.91639107e-01 -1.05894780e+00 -2.00787529e-01
5.01673996e-01 -2.45041803e-01 -4.15405095e-01 7.63712943e-01
-7.86167562e-01 8.19165051e-01 2.44888037e-01 1.01463985e+00
-6.83798909e-01 -2.91613932e-03 1.02047741e+00 7.10632741e-01
9.52341914e-01 3.82010877e-01 -8.93785000e-01 -4.88674253e-01
-7.74546564e-01 8.99224635e-03 6.05483174e-01 2.45408997e-01
1.48896426e-01 9.92569625e-01 -1.29490837e-01 1.15220678e+00
-5.33872657e-02 3.47698897e-01 1.19195366e+00 -6.01993084e-01
-3.68141234e-02 -1.99686974e-01 -9.55407023e-02 -1.10223734e+00
-7.49732494e-01 -6.43246114e-01 -7.07055926e-01 -5.39438277e-02
1.62114501e-01 -2.84251809e-01 -2.56792009e-01 1.80486023e+00
9.53926221e-02 -6.57528043e-02 3.10931176e-01 8.26757669e-01
9.19607997e-01 9.24153268e-01 2.14075565e-01 -7.42801845e-01
1.80280936e+00 -4.71678734e-01 -1.70306778e+00 2.32204899e-01
9.94705632e-02 -7.68348753e-01 1.12970173e+00 5.69147170e-01
-9.93940949e-01 -1.08758938e+00 -1.08305132e+00 9.83018652e-02
-7.64525294e-01 2.19949335e-01 5.32737672e-01 1.18853557e+00
-9.41509724e-01 8.44370782e-01 -4.58382368e-01 -6.95733801e-02
-8.80676433e-02 9.67788622e-02 -4.70700592e-01 1.12629020e+00
-2.05958652e+00 9.08959985e-01 2.62476802e-01 -7.14695677e-02
-1.53856680e-01 -8.61299515e-01 -7.98055112e-01 2.85651058e-01
-3.39924753e-01 3.37634161e-02 1.34423220e+00 -1.25747550e+00
-1.85608625e+00 5.44935465e-01 -3.41646612e-01 -7.59192705e-01
-2.29459226e-01 -3.64110321e-01 -1.37260675e+00 7.74484634e-01
-4.82519269e-01 1.49616137e-01 1.26015389e+00 -4.75208461e-01
-6.03296459e-01 -1.20760873e-01 -2.89813042e-01 1.22525757e-02
-8.29956293e-01 2.68712193e-01 7.61624157e-01 -1.15104127e+00
3.52737345e-02 -2.55543560e-01 3.24296296e-01 -3.79460543e-01
2.83072233e-01 -4.53412145e-01 7.21058071e-01 -9.00144219e-01
1.63013363e+00 -2.40672612e+00 -5.26042879e-01 -2.92439669e-01
-2.04764158e-01 5.47851026e-01 -6.17720410e-02 7.71990597e-01
-7.82132924e-01 6.43152371e-02 3.41003358e-01 7.40156099e-02
5.29617220e-02 -1.64171681e-01 -5.84545076e-01 4.67477977e-01
5.90067327e-01 6.24521255e-01 -6.63296759e-01 -1.33288726e-01
2.92423218e-01 9.75987852e-01 -3.13232213e-01 -9.65284109e-02
4.65374023e-01 2.02477127e-02 7.92353451e-02 3.15677077e-01
4.69393134e-01 4.13879752e-01 -2.61659116e-01 -6.47451103e-01
-3.37233305e-01 6.42954648e-01 -7.60321140e-01 1.38046741e+00
-6.02997422e-01 9.81231868e-01 1.41280591e-01 -8.97393286e-01
1.46365011e+00 1.04585981e+00 5.25738001e-01 -9.13397968e-01
1.86261252e-01 2.93008864e-01 3.49223502e-02 -5.05533636e-01
5.34860015e-01 -6.00808442e-01 7.71870092e-02 -9.28473622e-02
4.53414977e-01 -9.57003534e-02 -8.53658020e-02 -4.61158395e-01
6.00607634e-01 -2.70136982e-01 6.03185236e-01 -3.83321285e-01
7.81354010e-01 -6.81280255e-01 3.89164805e-01 1.89715281e-01
-7.95341909e-01 3.63411546e-01 1.72958732e-01 -4.80819404e-01
-6.13222063e-01 -9.58886027e-01 -5.18668950e-01 1.49028587e+00
-4.61052030e-01 -6.83142602e-01 -5.21522462e-01 -1.37263462e-01
-5.91742657e-02 6.47165954e-01 -5.28013110e-01 -4.64350104e-01
-4.36010212e-01 -5.43076336e-01 9.38010693e-01 5.30783892e-01
7.21801892e-02 -1.20159948e+00 -1.08738792e+00 5.29882014e-01
-4.11147088e-01 -1.01691115e+00 -4.26674873e-01 8.25722992e-01
-8.69188666e-01 -4.27667499e-01 -6.48405015e-01 -7.14018404e-01
-3.52598995e-01 -1.61425829e-01 9.83245432e-01 -4.61855471e-01
-2.44878888e-01 4.74478275e-01 -8.84975135e-01 -8.48307073e-01
-1.38874307e-01 -3.95549804e-01 2.85294175e-01 3.13925855e-02
5.15494347e-01 -7.95939267e-01 -4.97635961e-01 -1.62351914e-02
-6.42179012e-01 -8.95652294e-01 1.15503989e-01 1.00779808e+00
3.93852055e-01 4.47279990e-01 1.41186178e+00 -6.16693981e-02
1.30752885e+00 -1.77388519e-01 4.14838502e-03 -3.21408004e-01
-1.32490605e-01 -4.40859854e-01 9.41413403e-01 -9.35898244e-01
-1.07605994e+00 -6.01695836e-01 -6.70922935e-01 -5.54654717e-01
-1.13061052e-02 5.22281528e-01 2.88408309e-01 1.82151437e-01
9.57929552e-01 2.53249705e-01 -5.22561856e-02 -4.00469840e-01
8.57665986e-02 8.20847690e-01 5.34692526e-01 -4.18906063e-01
2.42773965e-01 1.34919956e-01 -2.89355963e-01 -1.43011820e+00
-6.63672566e-01 -4.41723287e-01 -3.10986400e-01 -3.58298779e-01
7.48206913e-01 -8.65122676e-01 -8.73642921e-01 3.88633996e-01
-1.17144787e+00 2.94248223e-01 -6.13276958e-01 8.23338807e-01
-6.66031837e-01 1.65178031e-01 -9.69872057e-01 -1.63014138e+00
-8.71823490e-01 -6.03533864e-01 1.00322497e+00 4.11168784e-01
-5.67649424e-01 -1.02868295e+00 -7.25716278e-02 -1.13417968e-01
6.62611485e-01 4.70897913e-01 7.42670059e-01 -6.78382695e-01
6.44894660e-01 -2.62825489e-01 5.51514030e-01 6.18238926e-01
2.97433019e-01 5.15364073e-02 -1.61706030e+00 -2.05794290e-01
5.29102564e-01 -2.96828538e-01 5.45702457e-01 6.88116610e-01
1.13999712e+00 -3.58225822e-01 4.20226544e-01 5.70938766e-01
1.00829339e+00 7.58728504e-01 4.18693930e-01 1.40928524e-02
-2.13613093e-01 5.97046316e-01 7.48231769e-01 7.28001416e-01
-1.75339535e-01 5.43304443e-01 8.54742713e-03 -2.76933581e-01
-4.08050045e-03 -1.21439369e-02 5.83925903e-01 1.42623818e+00
-1.40322641e-01 9.48911719e-03 -4.52205896e-01 6.88014686e-01
-1.14072633e+00 -1.21741569e+00 -1.49762198e-01 1.86606789e+00
8.91672313e-01 1.55595601e-01 4.75238681e-01 9.76946652e-01
5.05446136e-01 4.07618374e-01 -3.26954216e-01 -1.15236092e+00
-4.09831285e-01 8.68649185e-01 -2.07458884e-01 2.04648331e-01
-1.10216713e+00 4.76406485e-01 6.87725019e+00 1.16043818e+00
-1.45993185e+00 3.25093091e-01 4.55480635e-01 -1.73795089e-01
1.04352370e-01 -8.03630114e-01 -3.26763391e-01 2.74287134e-01
1.80832040e+00 -5.49549818e-01 2.28521526e-01 7.65763700e-01
4.38827932e-01 1.57713860e-01 -7.74210930e-01 1.20686698e+00
-1.78355440e-01 -9.98241603e-01 -4.24997300e-01 -3.61870587e-01
8.24767947e-02 -1.97009638e-01 3.94051731e-01 5.12162805e-01
-6.61659300e-01 -1.05530131e+00 9.82967794e-01 4.12300736e-01
1.02394366e+00 -1.28151488e+00 7.43912816e-01 -7.52456486e-02
-1.65788853e+00 -2.98065305e-01 -4.45026428e-01 -2.06134617e-01
-4.97270823e-02 7.35375166e-01 -5.94985306e-01 7.37503588e-01
9.23067570e-01 5.62258303e-01 1.23458914e-01 4.55770552e-01
1.82743609e-01 7.24791825e-01 -1.43085957e-01 -8.75858814e-02
2.59816856e-03 1.21397264e-01 6.11693561e-01 1.75965393e+00
2.38435075e-01 3.18895102e-01 -2.84665048e-01 4.14572209e-01
5.46994746e-01 2.18761116e-01 -1.81976020e-01 -5.30705214e-01
5.60464919e-01 1.15558946e+00 -5.08664608e-01 -2.23760940e-02
-3.39254498e-01 6.26186609e-01 -1.77748248e-01 4.07225579e-01
-8.03290963e-01 -1.02231276e+00 1.12292719e+00 -3.01337421e-01
3.01474184e-01 -2.37320542e-01 7.63453394e-02 -6.61886036e-01
-9.38663930e-02 -8.70963633e-01 5.02524793e-01 -8.86724591e-01
-1.39215505e+00 1.07392037e+00 -1.60549372e-01 -1.33090830e+00
-3.38226885e-01 -6.65254116e-01 -5.28081119e-01 1.26902580e+00
-1.71498120e+00 -6.33130431e-01 2.40265980e-01 7.67468274e-01
6.74683452e-01 -7.82467052e-02 1.46837687e+00 3.05684954e-01
-1.43493161e-01 7.83724010e-01 -2.73197353e-01 -7.22592250e-02
4.13368613e-01 -1.22788775e+00 1.88716158e-01 5.09221256e-01
6.44500703e-02 6.36037827e-01 7.29184449e-01 -1.50077462e-01
-1.04594100e+00 -7.35169530e-01 1.20733225e+00 -1.38070658e-01
6.30188167e-01 -2.69837469e-01 -1.11071587e+00 -2.94573959e-02
2.84721971e-01 1.03292368e-01 1.04902399e+00 -2.21826769e-02
-4.31942314e-01 -4.40292329e-01 -1.25956154e+00 1.37109607e-01
4.47725654e-01 -1.19657242e+00 -1.07673514e+00 1.16279997e-01
9.97360647e-01 -2.22975567e-01 -1.37167645e+00 2.94808149e-01
6.96917474e-01 -1.32792497e+00 1.12216818e+00 -4.64129567e-01
1.41800329e-01 1.39534160e-01 -3.22930753e-01 -1.59415030e+00
-3.23295772e-01 -1.00465739e+00 -8.43431279e-02 9.67759252e-01
1.14331953e-01 -8.93312693e-01 4.29966897e-02 -4.30368185e-01
-3.69101524e-01 -6.97049439e-01 -1.30321229e+00 -9.65378225e-01
-4.95741442e-02 -7.05521405e-01 4.82436419e-01 8.73366356e-01
5.82800448e-01 1.46196857e-01 -1.42912760e-01 7.41581321e-02
-1.75949737e-01 -1.91626489e-01 -8.57028440e-02 -1.21179950e+00
-9.37004834e-02 -5.04576385e-01 -5.21716297e-01 -2.43399441e-01
2.14379787e-01 -4.99880552e-01 -2.90297568e-01 -7.90963769e-01
-7.27053821e-01 4.11076367e-01 -7.55142331e-01 3.09877336e-01
-6.46675527e-02 1.42811581e-01 1.15111805e-01 -3.56397122e-01
1.42936215e-01 9.45225835e-01 1.12677193e+00 1.50810108e-01
-3.08590204e-01 -9.47658941e-02 -1.32089600e-01 7.66755939e-01
7.22307861e-01 -2.52807200e-01 -5.56863725e-01 4.55514938e-01
-1.07187219e-01 4.58570987e-01 -7.35051036e-02 -1.12687659e+00
-2.87502676e-01 7.50141367e-02 5.81293941e-01 -6.02842629e-01
9.83468771e-01 -7.82673299e-01 7.44856820e-02 5.47369897e-01
-1.86190039e-01 3.23065281e-01 8.64493489e-01 3.42497379e-01
-7.78064370e-01 7.71607757e-02 8.80920529e-01 -3.46310027e-02
-6.03060961e-01 -2.87681907e-01 -9.06749368e-01 -6.64459914e-02
4.37157512e-01 -5.82999408e-01 -2.97478050e-01 -6.42548680e-01
-9.53191221e-01 -7.25650728e-01 -4.97419536e-01 5.55763066e-01
6.76386714e-01 -1.40527880e+00 -5.40024877e-01 5.64136147e-01
1.62814204e-02 -9.11133349e-01 5.29699504e-01 9.13787723e-01
5.29167987e-02 6.76551998e-01 -5.37656784e-01 -4.16950464e-01
-1.36886513e+00 2.35097170e-01 6.94881320e-01 1.44647695e-02
-4.73878741e-01 1.04111063e+00 -1.03842676e-01 2.63620540e-02
2.40763888e-01 -6.89508617e-01 -7.00459540e-01 6.36325300e-01
7.09786832e-01 3.77458334e-01 6.07857585e-01 -8.92040908e-01
-6.30204022e-01 3.77912670e-01 4.09784585e-01 -3.22586447e-01
1.51859558e+00 -1.06218033e-01 8.68997052e-02 1.06545305e+00
1.33931875e+00 4.50764857e-02 -7.84919739e-01 1.84592009e-01
2.88418680e-02 -9.13910419e-02 3.70111346e-01 -8.88757408e-01
-8.43600214e-01 9.01760757e-01 8.48724723e-01 8.49586844e-01
1.76410306e+00 -6.10088050e-01 7.95362949e-01 -1.57530408e-03
2.41578773e-01 -1.32186019e+00 3.76056105e-01 8.41882706e-01
1.20072913e+00 -5.44719994e-01 -3.80337268e-01 -2.73381203e-01
-4.91611570e-01 1.72101068e+00 1.08583614e-01 -1.37324005e-01
8.98839772e-01 2.32500330e-01 4.00928080e-01 4.98634093e-02
-9.40394640e-01 -3.01341712e-01 4.28987443e-01 8.66576374e-01
1.00418925e+00 1.51269451e-01 -5.90625584e-01 1.02852559e+00
-1.03424156e+00 -4.08521444e-01 2.52246201e-01 6.28329873e-01
-4.75745678e-01 -7.69469559e-01 -7.54803777e-01 1.51307136e-01
-9.04900551e-01 8.35092440e-02 -1.52447805e-01 6.84437335e-01
1.92805883e-02 1.54608500e+00 2.73832947e-01 -4.72292572e-01
6.05038702e-01 7.53293157e-01 3.19554538e-01 -1.08191662e-01
-1.30120456e+00 6.26232922e-01 2.47422054e-01 -6.65511072e-01
-5.45031548e-01 -3.54293346e-01 -1.29959428e+00 1.38872460e-01
-4.29506987e-01 4.11497831e-01 9.17911768e-01 5.62130332e-01
2.69705534e-01 1.12753761e+00 7.36644387e-01 -7.52573013e-01
-5.75671971e-01 -1.29803073e+00 -8.55172455e-01 5.10587990e-01
1.06207597e+00 -4.97721076e-01 -6.28981948e-01 1.51334450e-01] | [15.119505882263184, 5.529289722442627] |
fb39e43d-c38d-4579-bb51-21703243743a | detecting-adversarial-text-attacks-via | null | null | https://openreview.net/forum?id=DZ_7nGQgZgk | https://openreview.net/pdf?id=DZ_7nGQgZgk | Detecting Adversarial Text Attacks via SHapley Additive exPlanations | State-of-the-art machine learning models are prone to adversarial attacks: maliciously crafted inputs to fool the model into making a wrong prediction, often with high confidence. While defense strategies have been extensively explored in the computer vision domain, research in natural language processing still lacks techniques to make models resilient to adversarial text inputs. We propose an adversarial detector leveraging Shapley additive explanations against text attacks. Our approach outperforms the current state-of-the-art detector by around 19% F1-score on the IMDb and 14% on the SST-2 datasets while also showing competitive performance on AG_News and Yelp Polarity. Furthermore, we prove the detector to only require a low amount of training samples and, in some cases, to generalize to different datasets without needing to retrain. | ['Anonymous'] | 2021-05-16 | null | null | null | acl-arr-may-2021-5 | ['adversarial-text'] | ['adversarial'] | [ 3.28808308e-01 5.65717280e-01 -1.77520096e-01 -3.81125093e-01
-8.16748381e-01 -1.08606946e+00 8.55673730e-01 8.37304592e-02
-4.48419094e-01 6.52982712e-01 -3.05142164e-01 -7.93678105e-01
4.73606557e-01 -7.17016160e-01 -1.14594257e+00 -3.70433390e-01
1.19089238e-01 6.30796790e-01 4.72937554e-01 -5.83540574e-02
4.09868313e-03 5.99750280e-01 -7.18906701e-01 4.46224540e-01
5.43918908e-01 8.38908017e-01 -6.10442996e-01 8.81076932e-01
2.53163666e-01 1.00525784e+00 -7.08191156e-01 -1.16688478e+00
6.20972753e-01 -2.72210896e-01 -6.91936374e-01 -1.28603682e-01
7.51554191e-01 -4.24343437e-01 -6.47954047e-01 1.43660140e+00
1.47701129e-01 -4.15349454e-01 5.39813757e-01 -1.43949497e+00
-8.06663394e-01 8.67597938e-01 -4.02997464e-01 2.67275363e-01
5.63601442e-02 4.26719338e-01 1.03928208e+00 -6.14904046e-01
7.10246086e-01 1.22275817e+00 4.34558243e-01 1.08240783e+00
-1.28168213e+00 -9.32720065e-01 2.47469798e-01 2.09668592e-01
-7.43764102e-01 -1.75386593e-01 5.53526878e-01 -1.86125591e-01
6.99600041e-01 3.33445162e-01 3.06165274e-02 1.88615239e+00
3.26725513e-01 9.78779078e-01 1.25313044e+00 -2.24107489e-01
3.26940119e-01 4.19426829e-01 7.49799386e-02 7.49097943e-01
6.05861604e-01 3.65376323e-01 -1.38386294e-01 -4.11022425e-01
2.08656445e-01 -1.26651466e-01 -1.16362654e-01 -1.61128938e-01
-8.25616479e-01 1.20017946e+00 5.36376417e-01 1.64532159e-02
-2.11219519e-01 3.00355494e-01 6.08159959e-01 4.49400067e-01
4.00671750e-01 7.27506697e-01 -6.41914368e-01 2.93098032e-01
-6.49648845e-01 3.10809314e-01 9.54073310e-01 6.67344630e-01
2.25905299e-01 2.73275137e-01 1.07365541e-01 4.81388867e-02
1.65541545e-01 7.53406584e-01 2.67379522e-01 -4.92130280e-01
5.55512726e-01 2.79622018e-01 -1.44787937e-01 -1.04634929e+00
-1.61423668e-01 -6.32988513e-01 -7.75490880e-01 4.21204060e-01
5.31103432e-01 -3.53121400e-01 -9.50722754e-01 1.69732249e+00
1.51860267e-01 3.88825953e-01 2.34174475e-01 8.12334657e-01
3.01722020e-01 5.24047315e-01 5.69756217e-02 2.24170145e-02
1.12773454e+00 -9.06049669e-01 -2.09427774e-01 -7.96145797e-01
5.64104259e-01 -4.40615475e-01 8.26433778e-01 5.60872734e-01
-8.48667026e-01 -1.10958461e-02 -1.23875976e+00 2.37214029e-01
-3.18727583e-01 -5.36656976e-01 6.47774100e-01 1.02006149e+00
-3.48487586e-01 5.46943307e-01 -7.98604071e-01 5.11946566e-02
8.53813827e-01 2.57239729e-01 -6.65089011e-01 -8.37126374e-03
-1.38913059e+00 1.14170372e+00 2.81708986e-01 -4.66145515e-01
-1.36763310e+00 -6.64834738e-01 -5.67853868e-01 4.58702706e-02
6.66401088e-01 -6.05718791e-01 1.29340625e+00 -1.18863046e+00
-1.10056829e+00 1.04792082e+00 2.31689364e-01 -1.44642150e+00
1.12106359e+00 -4.75422591e-01 -5.57477713e-01 2.47222543e-01
-1.29925281e-01 2.77593702e-01 1.32388520e+00 -1.19118810e+00
-2.89814144e-01 -3.38873923e-01 2.86610484e-01 -4.02510345e-01
-4.30543691e-01 1.37362957e-01 9.25724357e-02 -7.13156998e-01
-2.55152315e-01 -1.19546056e+00 -5.49233317e-01 1.42943487e-01
-8.05213928e-01 1.18034445e-01 1.06768334e+00 -2.48064935e-01
7.14470685e-01 -1.93622994e+00 -2.13242427e-01 1.62886903e-01
2.79069960e-01 8.41459394e-01 -9.49902162e-02 2.59052545e-01
-8.47446248e-02 5.89697361e-01 -1.23110965e-01 -2.82789707e-01
1.37146130e-01 2.53975719e-01 -1.16985118e+00 8.43268275e-01
2.05878943e-01 8.13199580e-01 -7.86889136e-01 -1.88853189e-01
1.07805274e-01 3.09310555e-02 -7.86577046e-01 8.52105990e-02
-5.75793743e-01 2.41350919e-01 -5.48109233e-01 3.89559358e-01
6.16096854e-01 -2.40135074e-01 6.00877255e-02 4.21419799e-01
7.27562964e-01 3.83061677e-01 -8.29665422e-01 9.13809359e-01
-2.12050676e-01 7.14309394e-01 1.09165721e-02 -1.08245802e+00
6.33005381e-01 1.72765389e-01 -2.18088225e-01 -2.55354375e-01
4.50136065e-01 1.51920795e-01 1.34760112e-01 -1.85117856e-01
6.90235570e-02 -1.15672484e-01 -3.72291327e-01 5.81153035e-01
5.55347875e-02 1.74511559e-02 -3.99053365e-01 5.28529286e-01
1.40405178e+00 -4.76784468e-01 1.34004816e-01 2.02668250e-01
5.48052967e-01 4.23222147e-02 5.96830010e-01 1.24017847e+00
-3.18071425e-01 3.20379198e-01 6.84886634e-01 -5.68732798e-01
-1.13035643e+00 -1.04878259e+00 1.19363844e-01 1.02582991e+00
-4.23861407e-02 -2.45142564e-01 -9.30204690e-01 -1.43785334e+00
3.64126652e-01 1.08283257e+00 -7.56767988e-01 -4.72284913e-01
-3.88443857e-01 -3.44639719e-01 1.23667514e+00 4.89961296e-01
5.23970306e-01 -9.40308690e-01 -3.83681267e-01 -2.37572510e-02
2.92520255e-01 -1.44078171e+00 -2.93861449e-01 2.13198319e-01
-5.79529643e-01 -1.27225804e+00 -1.51055247e-01 -2.68787295e-01
6.59858048e-01 5.16923852e-02 1.05518115e+00 1.91593483e-01
-1.76842928e-01 7.16101229e-02 -3.77823770e-01 -8.79325330e-01
-1.12102818e+00 6.35996535e-02 2.54710257e-01 -1.38955899e-02
3.93475085e-01 -3.31020206e-01 -1.79799110e-01 3.40951771e-01
-9.92901862e-01 -2.49946982e-01 4.87516701e-01 1.00391054e+00
2.78310537e-01 -7.47141019e-02 5.94392896e-01 -1.52038157e+00
3.51701349e-01 -5.67680478e-01 -7.51407504e-01 1.09870983e-02
-7.08750606e-01 -1.71666518e-02 1.35348725e+00 -8.88483047e-01
-5.98298192e-01 -2.56370846e-02 -1.83838308e-01 -8.15580308e-01
-3.37035537e-01 8.86839777e-02 -2.59505928e-01 -1.89024195e-01
1.20003796e+00 1.37416959e-01 -2.46047616e-01 -1.08890340e-01
5.89112759e-01 3.54916871e-01 6.58437371e-01 -2.81869560e-01
1.54962015e+00 5.81113040e-01 2.85693973e-01 -3.35104764e-01
-1.14151895e+00 7.36425743e-02 -2.56547302e-01 1.58217981e-01
6.50853395e-01 -5.94991684e-01 -7.07570076e-01 4.07667220e-01
-1.08235025e+00 7.02167526e-02 5.92220947e-03 1.87860399e-01
-3.78349900e-01 3.66484582e-01 -8.10217917e-01 -7.24460065e-01
-5.12550652e-01 -9.51762140e-01 4.20963049e-01 4.49920781e-02
-1.29771337e-01 -9.38813090e-01 -2.68247962e-01 5.45074821e-01
3.07120860e-01 4.44163293e-01 8.53458166e-01 -1.63116276e+00
-5.49069285e-01 -9.13360059e-01 1.70785978e-01 5.13855278e-01
-5.08082688e-01 -7.67602026e-02 -1.23222816e+00 -2.79880166e-01
1.40044138e-01 -5.62731504e-01 1.16703415e+00 -1.82946324e-01
1.21181655e+00 -9.40011978e-01 -2.80362040e-01 4.92867321e-01
1.07953048e+00 -5.71296625e-02 5.46338856e-01 5.67800820e-01
4.95082527e-01 3.06196839e-01 4.87588197e-01 1.32640317e-01
-2.29956105e-01 2.70334810e-01 1.13563299e+00 2.33241636e-02
3.86334687e-01 -6.07938409e-01 4.41610187e-01 -2.53762573e-01
5.15255630e-01 -5.09059072e-01 -7.14916110e-01 3.33436698e-01
-1.73414254e+00 -1.23310506e+00 -1.58757448e-01 1.93900645e+00
6.90344393e-01 9.30880368e-01 -6.36566654e-02 1.90242052e-01
5.19004464e-01 3.04756671e-01 -8.90204370e-01 -7.73840308e-01
-2.93287575e-01 2.18369722e-01 1.01322567e+00 3.23974550e-01
-1.49255836e+00 1.40201056e+00 6.20765305e+00 7.27534473e-01
-1.21519268e+00 1.00891419e-01 7.56054580e-01 -1.82227552e-01
-2.76252657e-01 2.54275590e-01 -6.42279148e-01 2.44687110e-01
9.81758237e-01 -3.32395792e-01 4.12667453e-01 1.44539583e+00
-2.91556090e-01 3.54139566e-01 -1.04290116e+00 4.57594216e-01
9.57151502e-02 -1.41076231e+00 3.22500467e-02 1.51682034e-01
6.24758482e-01 2.18960166e-01 3.96124721e-01 4.15009707e-01
8.14226031e-01 -1.24636757e+00 7.82271326e-01 -6.01819977e-02
3.46739858e-01 -9.09806371e-01 7.14775801e-01 8.79047871e-01
-2.47280121e-01 -2.60607094e-01 -6.22026980e-01 6.84556365e-02
-7.13033229e-02 4.21736598e-01 -1.18099451e+00 2.24784881e-01
4.17706847e-01 1.61307365e-01 -6.97174370e-01 4.58935201e-01
-7.91521251e-01 1.19507802e+00 -2.32365817e-01 -1.03960305e-01
5.18893242e-01 4.82731819e-01 8.82105291e-01 1.07032001e+00
-2.11819544e-01 -1.87354051e-02 2.31362805e-01 7.37306654e-01
-6.39899731e-01 -2.44234741e-01 -8.30088735e-01 -2.94291586e-01
4.06610101e-01 1.10079539e+00 -2.97341436e-01 -3.74459296e-01
-1.97595850e-01 1.13959372e+00 2.55855739e-01 4.47365865e-02
-1.04063177e+00 -6.86809327e-03 6.80663347e-01 4.89962287e-02
2.93854892e-01 -8.33299756e-02 -4.70611632e-01 -1.28601396e+00
-7.34949261e-02 -1.26221168e+00 4.22096014e-01 -4.62609053e-01
-1.62417352e+00 6.98310554e-01 -4.30010617e-01 -1.00466323e+00
-4.73838717e-01 -7.55809307e-01 -6.95551336e-01 7.04566061e-01
-1.11362231e+00 -1.15442944e+00 3.24581385e-01 6.67195559e-01
3.48166347e-01 -6.59311414e-01 1.06025374e+00 -2.26363823e-01
-4.95840013e-01 1.09727299e+00 -4.17320430e-02 6.03599250e-01
8.12205672e-01 -1.14494371e+00 1.09970057e+00 1.32953703e+00
5.18314540e-01 5.61876297e-01 1.22991669e+00 -6.75167084e-01
-1.54875386e+00 -1.09356439e+00 7.05268383e-01 -6.38630688e-01
1.01851678e+00 -5.11924863e-01 -1.12340522e+00 9.96750355e-01
3.10491305e-02 4.21186388e-01 5.00294626e-01 3.03814770e-03
-1.24655485e+00 1.76810905e-01 -1.61189032e+00 7.90504336e-01
6.88580811e-01 -7.04411149e-01 -6.98552668e-01 4.96392280e-01
7.64371455e-01 -4.80017453e-01 -3.58903855e-01 3.23582470e-01
4.31451023e-01 -8.48545551e-01 8.84941399e-01 -1.43299341e+00
4.60620016e-01 -5.09160087e-02 -1.25214905e-01 -1.16040087e+00
-2.54803866e-01 -7.79729307e-01 -3.00916195e-01 9.81966376e-01
6.72325134e-01 -9.71670508e-01 9.79232252e-01 6.22418225e-01
2.03802422e-01 -5.10717750e-01 -1.02917790e+00 -8.25430274e-01
4.51464176e-01 -8.43298078e-01 2.30580226e-01 1.14330316e+00
-1.86170921e-01 2.52560019e-01 -7.71824300e-01 6.48876369e-01
7.56338775e-01 -1.50681466e-01 1.10589707e+00 -1.12938952e+00
-6.24441504e-01 -3.65786731e-01 -6.27481520e-01 -6.69783533e-01
6.14609182e-01 -9.64862943e-01 -3.20660502e-01 -8.33159029e-01
3.13494913e-02 -1.42561987e-01 -2.04780132e-01 6.79801047e-01
-2.42584631e-01 3.67599815e-01 5.18984199e-01 8.24381188e-02
-4.65311199e-01 7.91534260e-02 7.42771506e-01 -4.14848089e-01
4.73084837e-01 3.26970875e-01 -9.44100618e-01 1.06332755e+00
8.43952239e-01 -1.06132758e+00 -2.46345565e-01 -2.34658584e-01
1.93887457e-01 -1.17052816e-01 5.98119259e-01 -8.05539191e-01
1.69679254e-01 -2.64542490e-01 4.88292098e-01 -1.74337119e-01
9.53528881e-02 -8.50514293e-01 -1.30135491e-01 7.66432881e-01
-6.18696570e-01 -1.18530206e-01 2.44180158e-01 8.94497871e-01
2.76715487e-01 -2.75751233e-01 1.16586733e+00 -1.67032611e-02
-1.80921182e-01 4.42552447e-01 -4.24656242e-01 1.81665063e-01
1.32963622e+00 3.33270371e-01 -7.71013677e-01 -4.83986080e-01
-6.28887534e-01 9.52196345e-02 6.50605977e-01 5.50633788e-01
4.32695806e-01 -7.41559565e-01 -7.72376716e-01 2.26726696e-01
-6.37722239e-02 -4.46449012e-01 1.29440606e-01 1.68062180e-01
-4.11984563e-01 4.24019754e-01 -4.89801839e-02 -2.34811693e-01
-1.44024205e+00 1.04294240e+00 4.02315348e-01 -5.38765311e-01
-6.33474886e-01 9.85151827e-01 8.95550996e-02 -1.14730656e-01
2.51882434e-01 1.88444138e-01 2.86816001e-01 -3.68128598e-01
6.75479949e-01 6.97726831e-02 -4.73572053e-02 -4.18958873e-01
-5.42843699e-01 -2.21924812e-01 -5.99785924e-01 -1.54901594e-01
9.40046072e-01 5.88961303e-01 4.33951974e-01 -5.55613339e-02
9.99792457e-01 2.73510277e-01 -1.03572083e+00 -3.31904739e-01
2.92966925e-02 -5.24591267e-01 -9.19191614e-02 -1.04814720e+00
-8.38728428e-01 1.04592693e+00 3.47678810e-01 5.13558686e-01
8.30307305e-01 -1.99982780e-03 8.70095730e-01 6.88625634e-01
4.53408092e-01 -6.34979546e-01 1.26385912e-01 6.06628776e-01
7.92476356e-01 -1.27681363e+00 -5.20444103e-02 -3.39227885e-01
-1.04420567e+00 8.76342595e-01 4.76327688e-01 -5.66987813e-01
4.86871421e-01 4.41445351e-01 2.46799245e-01 2.15571284e-01
-1.06672025e+00 3.33512336e-01 1.85829952e-01 4.89976764e-01
-1.87137812e-01 1.65342584e-01 -4.82364558e-02 6.61259353e-01
-3.07524711e-01 -4.39308792e-01 7.05162942e-01 6.77134514e-01
-4.13057834e-01 -1.19427931e+00 -3.50123376e-01 3.68058920e-01
-1.36916721e+00 -1.82286695e-01 -8.76074672e-01 7.02362537e-01
-2.43133977e-01 1.05700302e+00 -4.47539300e-01 -6.13112032e-01
2.20131621e-01 1.12361193e-01 7.05176219e-02 -5.60741544e-01
-9.64933515e-01 -3.03762257e-01 1.59242705e-01 -5.12225270e-01
1.54116228e-01 -3.84281009e-01 -1.16810060e+00 -5.66399038e-01
-4.11640108e-01 -1.04069307e-01 6.53767467e-01 9.80879307e-01
2.84368932e-01 1.39686555e-01 8.37737918e-01 -4.40439612e-01
-1.17614579e+00 -8.69040906e-01 -3.04971933e-01 4.10330206e-01
3.94885242e-01 -3.43184978e-01 -7.83717513e-01 -1.96353614e-01] | [5.862773895263672, 7.986826419830322] |
1140521e-2acc-40a9-b678-cdf6f6aab7a5 | lfkqg-a-controlled-generation-framework-with | null | null | https://aclanthology.org/2022.coling-1.572 | https://aclanthology.org/2022.coling-1.572.pdf | LFKQG: A Controlled Generation Framework with Local Fine-tuning for Question Generation over Knowledge Bases | Question generation over knowledge bases (KBQG) aims at generating natural questions about a subgraph, which can be answered by a given answer entity. Existing KBQG models still face two main challenges: (1) Most models often focus on the most relevant part of the answer entity, while neglecting the rest of the subgraph. (2) There are a large number of out-of-vocabulary (OOV) predicates in real-world scenarios, which are hard to adapt for most KBQG models. To address these challenges, we propose LFKQG, a controlled generation framework for Question Generation over Knowledge Bases. (1) LFKQG employs a simple controlled generation method to generate the questions containing the critical entities in the subgraph, ensuring the question is relevant to the whole subgraph. (2) We propose an optimization strategy called local fine-tuning, which can make good use of the rich information hidden in the pre-trained model to improve the ability of the model to adapt the OOV predicates. Extensive experiments show that our method outperforms existing methods significantly on three widely-used benchmark datasets SimpleQuestion, PathQuestions, and WebQuestions. | ['Xuanjing Huang', 'Qi Zhang', 'Tao Gui', 'Xin Zhou', 'Zichu Fei'] | null | null | null | null | coling-2022-10 | ['natural-questions', 'question-generation'] | ['miscellaneous', 'natural-language-processing'] | [-3.72904688e-02 7.14952826e-01 -1.29881546e-01 -1.16524354e-01
-9.04105902e-01 -7.83342302e-01 4.82056379e-01 2.06886783e-01
2.64406595e-02 1.13652456e+00 5.53353846e-01 -3.47241610e-01
-1.44130290e-01 -1.47304749e+00 -9.84815776e-01 -1.33222446e-01
4.87559974e-01 7.12021530e-01 8.66033256e-01 -8.62366140e-01
2.14199945e-01 -9.88625810e-02 -1.43664467e+00 3.59931231e-01
1.50922453e+00 7.93202043e-01 1.57144353e-01 5.00680029e-01
-6.68673694e-01 8.97779167e-01 -6.91217363e-01 -7.43564367e-01
-5.20944744e-02 -5.98565340e-01 -1.47006691e+00 -3.32873940e-01
5.74615240e-01 -2.08966851e-01 -3.27448517e-01 9.44754660e-01
4.76048499e-01 1.79805264e-01 5.89315474e-01 -1.21213746e+00
-8.18497479e-01 8.08186889e-01 2.02988937e-01 1.17556356e-01
7.01741874e-01 9.11303088e-02 1.55014455e+00 -6.40529931e-01
8.39126110e-01 1.21372008e+00 1.66425601e-01 7.15279162e-01
-8.65267992e-01 -4.41476196e-01 3.02361578e-01 5.39583147e-01
-1.32721734e+00 -1.76705748e-01 6.07258081e-01 -1.53214723e-01
9.35700715e-01 4.68165696e-01 5.71252584e-01 5.82969069e-01
2.45696474e-02 5.29539526e-01 6.97614431e-01 -5.17522216e-01
2.70982474e-01 2.54031837e-01 1.93759888e-01 6.77299440e-01
2.82318056e-01 -3.98057133e-01 -5.56767106e-01 -4.70691890e-01
5.05834639e-01 -5.56352317e-01 -5.37463546e-01 -2.63661653e-01
-1.13199341e+00 1.08173823e+00 4.35524255e-01 6.48777708e-02
-3.66868556e-01 -1.29327737e-03 -1.69966463e-02 3.63591313e-01
1.14087828e-01 1.00208831e+00 -8.29781771e-01 1.20096579e-01
-3.93735081e-01 9.42364216e-01 1.26625609e+00 1.26314509e+00
1.16435444e+00 -5.29140294e-01 -9.01109576e-01 7.25562811e-01
2.38798246e-01 5.45259833e-01 4.25352991e-01 -8.39938581e-01
8.87209833e-01 1.14117265e+00 4.19786721e-01 -8.82272303e-01
8.83301720e-02 -2.12374046e-01 -3.87609392e-01 -6.73247099e-01
4.19384897e-01 -3.68540496e-01 -6.36568248e-01 1.79996634e+00
8.07428658e-01 -3.69667970e-02 2.26209119e-01 5.42088926e-01
1.42423928e+00 9.06001449e-01 6.11717207e-03 2.56952178e-02
1.66406953e+00 -1.03295994e+00 -6.55954421e-01 -3.61429125e-01
6.64307892e-01 -4.39289421e-01 1.27882934e+00 -2.16308221e-01
-9.22399580e-01 -3.52700412e-01 -6.71269536e-01 -1.91278428e-01
-5.50707936e-01 -2.57308900e-01 3.36288005e-01 3.91915470e-01
-1.05059958e+00 3.04321796e-02 1.17274635e-01 -3.11234117e-01
-1.20307477e-02 9.22093317e-02 1.51104443e-02 -4.25280899e-01
-1.98244691e+00 5.38586497e-01 6.94986165e-01 -3.08537424e-01
-7.69514978e-01 -7.92067051e-01 -9.14329648e-01 3.13753873e-01
1.00792706e+00 -1.18296874e+00 1.35245323e+00 -4.49007750e-01
-1.27312458e+00 4.67496186e-01 -4.73342687e-01 -1.39633194e-01
1.76597267e-01 -7.74364397e-02 -2.94572800e-01 3.16154778e-01
5.11354923e-01 7.88145304e-01 7.78437734e-01 -1.08966327e+00
-6.71097040e-01 -9.66406688e-02 6.61730230e-01 4.13848937e-01
-1.94065481e-01 -3.11894447e-01 -6.73273265e-01 -4.86670852e-01
-3.59849446e-02 -6.49467468e-01 -2.30196491e-01 -5.86829603e-01
-6.37399197e-01 -8.90705705e-01 4.84376609e-01 -6.61198556e-01
1.54737902e+00 -1.46595609e+00 -1.05175041e-01 6.63085952e-02
6.98534399e-02 3.88982862e-01 -3.11504453e-01 7.68090308e-01
1.96405470e-01 2.13425711e-01 1.27946317e-01 5.38534284e-01
1.70438349e-01 3.30295682e-01 -8.47051442e-01 -6.58189058e-01
2.60267943e-01 1.43791699e+00 -1.20707190e+00 -7.38006830e-01
-4.32296306e-01 -2.45871723e-01 -8.43179464e-01 5.56621373e-01
-1.06767309e+00 1.99215874e-01 -9.22924876e-01 4.67865855e-01
3.93547058e-01 -5.52395821e-01 -3.14820632e-02 -2.22393349e-01
4.49728936e-01 7.63505220e-01 -1.19613087e+00 1.31261361e+00
-2.13629678e-01 -9.36762914e-02 -5.05943477e-01 -5.28362393e-01
9.63077366e-01 3.24758291e-01 -6.02658354e-02 -6.77267015e-01
-3.44287366e-01 1.91621065e-01 -3.28646898e-01 -8.76205027e-01
6.33266449e-01 -4.23062965e-02 -1.94265500e-01 4.62978244e-01
1.52299196e-01 -5.24962664e-01 6.07337952e-01 7.66654432e-01
1.21707118e+00 -5.89849427e-02 3.11984241e-01 -1.45086840e-01
7.20918775e-01 1.92076683e-01 6.03232324e-01 1.05169773e+00
6.09152690e-02 3.79848927e-01 8.23925138e-01 -1.05837502e-01
-5.22012293e-01 -9.44529712e-01 5.01875758e-01 1.02632749e+00
2.47214943e-01 -5.93456984e-01 -9.32601035e-01 -1.22896135e+00
-8.07651877e-02 1.04846811e+00 -6.06338024e-01 -2.26766989e-01
-7.77712405e-01 -4.33208495e-01 3.16441178e-01 4.04689938e-01
4.81906265e-01 -1.43823481e+00 -6.35165423e-02 4.42386627e-01
-1.01319921e+00 -1.24589050e+00 -6.71071351e-01 -4.49574172e-01
-7.09942937e-01 -1.34373736e+00 -4.60374862e-01 -8.03945780e-01
7.07911789e-01 2.93945312e-01 1.75641012e+00 3.12397748e-01
1.84773520e-01 6.71878159e-01 -8.08979988e-01 -3.59258801e-01
-3.13625455e-01 4.67739761e-01 -5.94416976e-01 -8.42062309e-02
5.18720150e-01 -2.98876107e-01 -5.94853163e-01 3.73756737e-01
-1.06902826e+00 -5.48066683e-02 3.63071889e-01 6.34506881e-01
7.45390832e-01 8.40691850e-02 1.14726269e+00 -1.07792079e+00
1.01005709e+00 -6.12785280e-01 -5.08165896e-01 8.49852085e-01
-5.50618649e-01 3.32384199e-01 7.45874524e-01 -2.28305265e-01
-1.06457388e+00 -5.16813338e-01 -1.24526225e-01 3.11635822e-01
1.07275881e-01 5.69086552e-01 -5.61486602e-01 1.22844137e-01
6.25684679e-01 4.56496745e-01 -5.38530886e-01 -2.21798718e-01
7.20824599e-01 3.18193078e-01 2.70716339e-01 -9.07544971e-01
1.01513731e+00 1.62430733e-01 -2.91208684e-01 -6.15564644e-01
-1.21158135e+00 -5.70257127e-01 -2.38232240e-01 -6.51623756e-02
7.23198354e-01 -6.64611816e-01 -4.05570090e-01 1.88030303e-01
-1.13066053e+00 -4.07269537e-01 -5.85716963e-01 -1.81129631e-02
-3.82800400e-01 3.55063319e-01 -2.05949858e-01 -3.77851486e-01
-6.77652895e-01 -8.16221356e-01 1.01645815e+00 6.48590326e-01
-1.91949040e-01 -9.37318623e-01 3.06512147e-01 8.03704500e-01
3.07849884e-01 -1.81010947e-01 1.55467331e+00 -8.03799629e-01
-1.05635977e+00 -1.24406427e-01 -2.51442671e-01 1.41783088e-01
1.67647868e-01 -2.32441604e-01 -4.98871893e-01 8.16919059e-02
-3.91992092e-01 -5.33999920e-01 6.42397523e-01 -7.66440704e-02
9.97333705e-01 -7.97087550e-01 -2.88991094e-01 1.25773609e-01
1.22085679e+00 -1.05836645e-01 7.77530849e-01 1.98725238e-01
6.43767416e-01 7.92199910e-01 8.47524166e-01 9.86805484e-02
1.17062974e+00 6.01616144e-01 4.41446215e-01 4.08084989e-01
-2.46069983e-01 -8.85974348e-01 1.20698892e-01 8.74689162e-01
3.60899508e-01 -3.75408232e-01 -6.10375404e-01 1.03863180e+00
-1.89147007e+00 -7.32827902e-01 -1.86046734e-01 1.97261310e+00
1.17312896e+00 -1.56127468e-01 -1.27659023e-01 -1.71393931e-01
4.86124039e-01 1.78621143e-01 -5.34343719e-01 -9.23107117e-02
-2.58761439e-02 3.52789283e-01 -4.13276702e-02 4.94410098e-01
-6.32947683e-01 1.19804776e+00 5.70498276e+00 8.51010740e-01
-4.46633935e-01 -8.74682292e-02 2.61967987e-01 3.33924651e-01
-9.96143043e-01 3.98881525e-01 -1.43961811e+00 4.48459089e-01
6.11081719e-01 -6.11357450e-01 2.06588179e-01 6.72568798e-01
-1.63952023e-01 -1.60027929e-02 -9.37348902e-01 3.50397736e-01
1.84377477e-01 -1.27758121e+00 6.86915398e-01 -2.75306433e-01
1.00336218e+00 -3.84232938e-01 -3.61579120e-01 8.75662148e-01
4.51913059e-01 -8.82336080e-01 4.24273282e-01 5.59755445e-01
4.99271750e-01 -6.44029379e-01 6.20378375e-01 6.43546164e-01
-1.22388184e+00 -8.14402197e-03 -5.58058977e-01 3.37200820e-01
1.69425055e-01 6.53705537e-01 -1.22117066e+00 8.18527699e-01
6.27320409e-01 6.66723633e-03 -7.73187935e-01 8.73683453e-01
-1.07808280e+00 8.15631270e-01 -1.14216127e-01 -4.45589542e-01
3.56593102e-01 -7.98698515e-02 4.52804059e-01 8.04933071e-01
3.04409295e-01 4.13157731e-01 2.86791939e-02 9.07703996e-01
-4.62304384e-01 4.27670777e-01 -2.85755992e-01 6.33400865e-04
6.19503856e-01 1.17408943e+00 -3.29025209e-01 -5.12325525e-01
-3.57542574e-01 6.94337487e-01 4.76604342e-01 5.68543375e-01
-6.30005062e-01 -7.13372946e-01 2.62824625e-01 3.21802855e-01
5.43945074e-01 3.16257358e-01 2.67330676e-01 -1.33897674e+00
3.67813885e-01 -1.21742713e+00 8.64327252e-01 -1.01309824e+00
-1.12213266e+00 4.24193561e-01 1.46694155e-02 -7.19031692e-01
-5.44619441e-01 -1.91440985e-01 -6.52084112e-01 9.92623031e-01
-1.94141603e+00 -1.21708786e+00 -5.15805125e-01 6.30717933e-01
3.45423579e-01 1.78582847e-01 7.38285244e-01 2.89341956e-02
-2.59669453e-01 5.24063468e-01 -3.39678496e-01 1.36920869e-01
8.10903907e-01 -1.41670716e+00 5.47294796e-01 8.12198460e-01
6.52002320e-02 6.62505329e-01 4.50975955e-01 -8.38235795e-01
-1.23331821e+00 -1.43836987e+00 1.56894052e+00 -7.12756395e-01
4.69698250e-01 -1.29872158e-01 -1.26188934e+00 6.50128603e-01
1.01922028e-01 -2.89974004e-01 7.36912489e-01 4.91871573e-02
-4.58050579e-01 -1.24133050e-01 -9.18176830e-01 6.77255690e-01
7.10373223e-01 -2.82416642e-01 -1.11930168e+00 5.14051259e-01
1.27162528e+00 -2.75372326e-01 -6.33670628e-01 4.61598605e-01
7.09504262e-02 -6.39284551e-01 8.68938446e-01 -9.29536104e-01
4.44596797e-01 -5.06418705e-01 2.90673357e-02 -1.38029540e+00
-1.52166322e-01 -6.88736975e-01 -7.16054738e-01 1.37755728e+00
6.13526821e-01 -7.30904162e-01 7.62467861e-01 6.94851398e-01
6.28267899e-02 -9.45553005e-01 -8.00459027e-01 -5.28453708e-01
7.18657225e-02 1.77989736e-01 1.25546908e+00 7.85480499e-01
1.01014413e-03 6.53278053e-01 -4.48260494e-02 2.11282715e-01
3.31206501e-01 4.58092272e-01 1.05471861e+00 -1.06825769e+00
-3.62175971e-01 -6.91373497e-02 1.75217405e-01 -1.45694518e+00
-3.73256728e-02 -7.61498272e-01 8.54655653e-02 -2.15102506e+00
1.16719395e-01 -3.75702977e-01 9.71507505e-02 4.62255180e-01
-1.12659597e+00 -2.92844385e-01 -1.28378987e-01 -3.18714529e-02
-8.25875700e-01 7.37051487e-01 1.75823379e+00 -4.61724289e-02
-1.18982203e-01 1.26632527e-01 -1.29152441e+00 4.40662116e-01
6.31021082e-01 -4.68262702e-01 -8.21444273e-01 -3.33344430e-01
8.50689054e-01 1.71097249e-01 2.72095650e-01 -4.64294225e-01
2.74554372e-01 -5.28548777e-01 -1.42526358e-01 -5.21057725e-01
6.72507659e-02 -3.11118275e-01 -3.79871964e-01 2.81100839e-01
-2.27593109e-01 -2.02266499e-01 -7.44343773e-02 5.79083979e-01
-4.19961691e-01 -6.22113764e-01 3.91399145e-01 -3.50378662e-01
-7.80806482e-01 4.18910235e-01 -3.44832912e-02 9.76638258e-01
8.12235475e-01 1.75747484e-01 -6.02215290e-01 -7.87901819e-01
-2.10912451e-01 8.40669572e-01 1.31631345e-01 4.86322701e-01
5.43800890e-01 -1.32288420e+00 -8.96458447e-01 -1.31001666e-01
5.68590701e-01 5.26764631e-01 4.22002852e-01 2.67258048e-01
-4.24855053e-01 7.57655382e-01 3.97943348e-01 6.26002327e-02
-9.03607905e-01 5.35661101e-01 3.47935706e-01 -8.96900892e-01
-2.31387675e-01 9.90778506e-01 3.62400264e-01 -7.57866561e-01
-1.04758561e-01 -3.41205150e-01 -6.02586448e-01 1.36953235e-01
6.31898522e-01 2.51965672e-01 4.10129577e-02 -3.31979901e-01
-1.45489886e-01 4.31131691e-01 -3.44765693e-01 5.90837300e-02
8.64252925e-01 -9.65507925e-02 -2.75621623e-01 -1.33294478e-01
6.50233924e-01 1.08665876e-01 -6.69801593e-01 -5.56988180e-01
-1.20836474e-01 -1.27884135e-01 -4.93210047e-01 -9.42939997e-01
-6.23765111e-01 5.86083949e-01 -3.10359895e-01 4.81314361e-01
1.08521819e+00 2.78255671e-01 1.27186608e+00 7.83937037e-01
4.59045738e-01 -1.08833110e+00 2.76698738e-01 8.33554685e-01
9.32464838e-01 -7.68869877e-01 -2.13804021e-01 -7.72615075e-01
-5.93979597e-01 7.54713655e-01 1.08807659e+00 1.89544544e-01
3.23553026e-01 -6.33400142e-01 5.24306409e-02 -3.71470243e-01
-1.00148594e+00 -3.22340101e-01 3.89040291e-01 4.79585409e-01
-4.49161464e-03 -1.28848761e-01 -3.74507576e-01 7.79906571e-01
-4.89404529e-01 4.17796010e-03 4.59079683e-01 8.37365508e-01
-7.82709301e-01 -1.17873859e+00 -2.11011857e-01 5.87542236e-01
-3.84379834e-01 -2.65891105e-01 -7.63317704e-01 5.75480461e-01
1.85054522e-02 1.23798025e+00 -5.30485272e-01 -4.82203960e-02
6.48453474e-01 4.13934082e-01 3.25602770e-01 -1.00978732e+00
-6.19579375e-01 -5.18413186e-01 3.00866574e-01 -4.33052689e-01
1.90001167e-02 -2.04852879e-01 -1.07908905e+00 1.00647509e-01
-6.31374180e-01 7.96791673e-01 1.48659647e-01 1.08403742e+00
5.68446398e-01 5.41342199e-01 4.81681943e-01 3.63799721e-01
-7.52738178e-01 -9.35092032e-01 -4.66080427e-01 5.72673321e-01
6.09742813e-02 -3.57586175e-01 -2.20156059e-01 2.81084999e-02] | [10.95822525024414, 7.931543350219727] |
53107625-6950-4679-a63b-48ee7c47f761 | a-generative-adversarial-approach-with | 2009.10663 | null | https://arxiv.org/abs/2009.10663v1 | https://arxiv.org/pdf/2009.10663v1.pdf | A Generative Adversarial Approach with Residual Learning for Dust and Scratches Artifacts Removal | Retouching can significantly elevate the visual appeal of photos, but many casual photographers lack the expertise to operate in a professional manner. One particularly challenging task for old photo retouching remains the removal of dust and scratches artifacts. Traditionally, this task has been completed manually with special image enhancement software and represents a tedious task that requires special know-how of photo editing applications. However, recent research utilizing Generative Adversarial Networks (GANs) has been proven to obtain good results in various automated image enhancement tasks compared to traditional methods. This motivated us to explore the use of GANs in the context of film photo editing. In this paper, we present a GAN based method that is able to remove dust and scratches errors from film scans. Specifically, residual learning is utilized to speed up the training process, as well as boost the denoising performance. An extensive evaluation of our model on a community provided dataset shows that it generalizes remarkably well, not being dependent on any particular type of image. Finally, we significantly outperform the state-of-the-art methods and software applications, providing superior results. | ['Ionuţ Mironică'] | 2020-09-22 | null | null | null | null | ['photo-retouching'] | ['computer-vision'] | [ 7.45045364e-01 -1.15756921e-01 5.77755213e-01 -2.95878261e-01
-7.43365049e-01 -4.32078600e-01 3.67171168e-01 -2.56602585e-01
-2.57410944e-01 6.26477480e-01 -4.97716963e-02 1.45652238e-03
1.51880354e-01 -6.58612728e-01 -9.76155281e-01 -8.05064201e-01
5.45063496e-01 -9.23077911e-02 2.67740369e-01 -4.75108802e-01
2.81708717e-01 4.93860006e-01 -1.58881116e+00 1.55535340e-01
1.10971272e+00 7.54096210e-01 2.04244018e-01 7.34652281e-01
1.29905969e-01 8.30707371e-01 -7.96185851e-01 -8.37227941e-01
4.33448285e-01 -4.77932125e-01 -4.63282347e-01 5.35097063e-01
9.30089355e-01 -5.60729146e-01 -1.53394639e-01 1.10386479e+00
6.04923189e-01 6.06953278e-02 3.75678837e-01 -9.85007524e-01
-1.01405489e+00 6.63299561e-02 -6.95659876e-01 -6.77814633e-02
1.52631178e-01 3.84856671e-01 4.51681018e-01 -6.23116612e-01
6.26801729e-01 7.76312888e-01 9.54028428e-01 6.17620051e-01
-1.39101768e+00 -6.47891700e-01 -1.69973388e-01 2.39582822e-01
-1.14564741e+00 -3.91968995e-01 9.70549703e-01 -2.47194201e-01
5.46586573e-01 2.82655001e-01 5.51492095e-01 1.24433219e+00
3.11547041e-01 5.91891050e-01 1.57912791e+00 -5.16150177e-01
2.87207514e-01 2.07321942e-01 -3.39557230e-01 3.87343377e-01
2.22831070e-02 -7.48485923e-02 -4.49097723e-01 2.83266604e-01
8.41917753e-01 9.23503265e-02 -4.46009129e-01 -9.23511609e-02
-6.04684353e-01 4.83465701e-01 4.88422573e-01 3.72779340e-01
-6.03639185e-01 2.19087929e-01 2.00988904e-01 2.77509421e-01
7.04250038e-01 4.77757782e-01 -5.99460714e-02 -2.91582327e-02
-1.22082913e+00 6.98561361e-03 3.98287833e-01 6.18315279e-01
7.97811866e-01 2.35654846e-01 -1.20687626e-01 1.07687378e+00
-3.05824339e-01 4.49758649e-01 9.85516459e-02 -1.08312082e+00
1.48727581e-01 4.59229797e-01 4.71774749e-02 -1.02951992e+00
1.38649330e-01 -3.05328906e-01 -1.03293300e+00 7.72024751e-01
2.15367973e-01 -9.55943465e-02 -1.28711367e+00 1.39167476e+00
3.11315298e-01 2.63583511e-01 -1.79933771e-01 9.17625189e-01
5.02783895e-01 4.96931046e-01 1.07873052e-01 -8.53513256e-02
1.06811047e+00 -9.87567961e-01 -9.69685197e-01 -3.87652397e-01
-1.01440057e-01 -1.07925212e+00 1.27306437e+00 9.14957821e-01
-1.33570480e+00 -6.79934919e-01 -1.18524873e+00 -2.40215436e-01
-1.74792722e-01 1.00567408e-01 5.47810793e-01 9.28507447e-01
-1.20931208e+00 9.04289007e-01 -6.67358875e-01 -3.96115124e-01
5.93166649e-01 2.26844758e-01 -4.94291246e-01 -4.61911142e-01
-7.62905300e-01 9.74726558e-01 -1.28117725e-01 2.88168222e-01
-8.33920181e-01 -8.08368027e-01 -6.91703916e-01 -9.49596241e-02
4.79728192e-01 -8.03336561e-01 1.10925317e+00 -1.38950598e+00
-1.75447941e+00 8.41026485e-01 3.97804193e-02 -4.15724635e-01
7.09871411e-01 -5.31123042e-01 -4.32144493e-01 2.49100134e-01
-2.05279708e-01 3.84135753e-01 1.52039874e+00 -1.60729206e+00
-2.32570753e-01 -2.95576364e-01 8.59168619e-02 3.18844467e-02
-5.43601573e-01 -2.66630296e-02 -5.95973969e-01 -9.56897259e-01
-1.71580940e-01 -7.40294039e-01 -1.62314847e-01 1.90574408e-01
-3.30955178e-01 3.00996184e-01 8.89861405e-01 -1.15940702e+00
1.01231575e+00 -2.21725678e+00 5.91918007e-02 1.27276611e-02
3.99413444e-02 5.48460424e-01 -9.17826816e-02 4.57887709e-01
5.00621609e-02 9.30346474e-02 -6.18016005e-01 -5.97542584e-01
-2.31190592e-01 2.46956900e-01 -1.21769853e-01 4.24538910e-01
4.25152183e-01 7.17699647e-01 -6.34595811e-01 -2.86728978e-01
5.15502632e-01 8.70357931e-01 -3.46158177e-01 3.40487361e-01
-3.01794764e-02 4.73102897e-01 1.10074088e-01 7.35051751e-01
9.42344725e-01 -4.46146056e-02 -3.61023471e-02 -4.52993602e-01
6.72308356e-02 -3.18535358e-01 -1.04893899e+00 1.77012122e+00
-7.54478335e-01 6.60429537e-01 2.82049626e-01 -6.55414343e-01
8.17141473e-01 9.34513658e-02 2.30697900e-01 -8.15108120e-01
-1.44701190e-02 1.74812093e-01 -3.43772441e-01 -8.75979841e-01
6.20966375e-01 -3.15248728e-01 5.29895306e-01 2.31934652e-01
-2.02125758e-01 -5.12819767e-01 3.62004414e-02 1.46020919e-01
1.16034937e+00 3.54020387e-01 2.40150876e-02 2.02231273e-01
3.27383816e-01 -1.20283797e-01 3.72032613e-01 5.27586222e-01
6.83387443e-02 1.03045523e+00 -5.87037504e-02 -1.03530698e-01
-1.32890761e+00 -1.01329255e+00 5.97187504e-02 8.07150543e-01
1.40952244e-01 -6.33418337e-02 -1.18199217e+00 -4.93313223e-01
-2.59415805e-01 6.91658199e-01 -6.93942070e-01 -1.59526810e-01
-5.41899979e-01 -5.54766417e-01 2.14510262e-01 5.70110142e-01
7.33413875e-01 -1.03101897e+00 -3.99841160e-01 7.28991553e-02
-5.16384989e-02 -1.21954238e+00 -5.62559605e-01 -1.26455322e-01
-8.40182424e-01 -9.48838174e-01 -1.00863707e+00 -6.69930816e-01
8.53158891e-01 4.93124515e-01 9.92622554e-01 2.14144036e-01
-5.52188635e-01 5.50112545e-01 -5.88272929e-01 -4.15673435e-01
-5.68892360e-01 -1.94350451e-01 -2.93000966e-01 2.83841580e-01
-5.81857860e-02 -8.14803720e-01 -8.57866764e-01 2.90136516e-01
-1.41394258e+00 -4.94026877e-02 8.82209480e-01 8.75420034e-01
3.81334394e-01 4.80510533e-01 4.23932582e-01 -1.12593997e+00
5.98047554e-01 5.42148165e-02 -3.27006370e-01 2.32078701e-01
-5.92712998e-01 -2.96656102e-01 6.30164206e-01 -5.05083919e-01
-1.45344079e+00 1.01767868e-01 -3.04548144e-01 -4.46051359e-01
-2.03311905e-01 9.35355797e-02 -1.58347741e-01 -4.65219378e-01
7.22874761e-01 1.94860667e-01 9.68229175e-02 -5.82324684e-01
2.72201955e-01 6.78903520e-01 8.60147953e-01 -8.00260529e-02
1.06500816e+00 6.93787634e-01 -6.89230859e-02 -1.01369512e+00
-7.68743217e-01 -2.57461250e-01 -4.94306207e-01 -4.52708632e-01
8.09956849e-01 -7.44716346e-01 -4.70524907e-01 7.78494239e-01
-9.20373261e-01 -5.25678754e-01 -2.09528551e-01 -1.33582959e-02
-3.70925963e-01 7.64638007e-01 -6.30524516e-01 -7.18857229e-01
-4.84116614e-01 -9.21249866e-01 1.06003296e+00 4.25420940e-01
9.72014368e-02 -8.08901072e-01 -1.22176439e-01 7.90260196e-01
6.60542428e-01 4.41318780e-01 5.23522437e-01 2.55837083e-01
-6.08426690e-01 -2.59457588e-01 -9.25326049e-02 1.11197710e+00
2.01618001e-01 1.47185221e-01 -1.25676465e+00 -3.18470657e-01
2.60185212e-01 -1.95193529e-01 8.19002569e-01 3.53211492e-01
1.32864344e+00 -6.24332428e-02 4.72274944e-02 6.25840068e-01
1.51817250e+00 4.68905456e-03 1.29107726e+00 4.63805884e-01
6.42234266e-01 4.80528980e-01 6.08199298e-01 1.85843095e-01
1.23285186e-02 5.53513527e-01 6.20160460e-01 -5.98645568e-01
-5.40089726e-01 -2.15571836e-01 4.01018411e-01 3.70304495e-01
-3.93556476e-01 -1.77687377e-01 -5.11809468e-01 4.32179809e-01
-1.54788816e+00 -8.69418383e-01 -3.03275794e-01 2.03656125e+00
9.06839669e-01 -1.09395176e-01 -1.57107741e-01 2.84770191e-01
7.76921511e-01 -3.25124189e-02 -5.11570871e-01 -3.40309322e-01
-1.78093299e-01 6.95377529e-01 5.10185838e-01 2.22915113e-01
-9.72700477e-01 8.91733289e-01 6.28634977e+00 8.12248111e-01
-1.09629858e+00 2.04888448e-01 6.09540343e-01 7.42167383e-02
-8.12486932e-02 -1.78380623e-01 -2.40580857e-01 5.93498528e-01
4.32775468e-01 3.86597902e-01 6.45652056e-01 7.46131599e-01
3.60172868e-01 -3.30821514e-01 -7.69432962e-01 1.07125616e+00
4.24367875e-01 -1.15289271e+00 -2.04645991e-01 -1.04075998e-01
9.97147381e-01 -4.43397433e-01 2.14112923e-01 -1.04100868e-01
2.58846115e-02 -1.09956539e+00 5.71848512e-01 6.76569164e-01
8.18269312e-01 -6.62353635e-01 8.19666028e-01 4.53789197e-02
-6.85879469e-01 1.40920971e-02 -2.84532279e-01 5.41578941e-02
4.28511441e-01 8.99451554e-01 -6.09351873e-01 5.38379729e-01
9.56734061e-01 5.52732468e-01 -8.54024112e-01 1.14138770e+00
-6.71612501e-01 6.06307149e-01 -4.30884734e-02 4.73757386e-01
-1.44461975e-01 -3.03433716e-01 3.53873134e-01 1.06288612e+00
4.23963875e-01 1.37288347e-02 -2.55409181e-01 8.15554857e-01
-2.12248906e-01 2.00943463e-02 -4.23314601e-01 3.57649173e-03
1.67152546e-02 1.47001624e+00 -5.57455420e-01 -1.77047625e-01
-2.76836783e-01 1.61372483e+00 5.56536838e-02 4.44040924e-01
-8.02382052e-01 -3.55359405e-01 4.14112240e-01 4.91501510e-01
4.10743922e-01 -1.36365011e-01 -3.32440346e-01 -9.12008166e-01
3.55126858e-01 -1.15181732e+00 -3.45944390e-02 -1.22146535e+00
-1.62690914e+00 6.11348212e-01 -5.41646361e-01 -1.14661467e+00
1.65775284e-01 -4.70097184e-01 -7.69194543e-01 7.62496769e-01
-1.61799109e+00 -1.50221539e+00 -8.70509028e-01 6.06131852e-01
6.28651679e-01 1.57944098e-01 4.81260777e-01 5.81286609e-01
-4.62570429e-01 5.16944647e-01 2.30100211e-02 -1.02050491e-01
1.12953568e+00 -1.14158690e+00 2.31936619e-01 1.29553986e+00
-2.99026296e-02 4.76180196e-01 9.29149449e-01 -6.44431293e-01
-1.36181700e+00 -1.01336515e+00 3.42459589e-01 -2.84829617e-01
3.59727234e-01 -1.65353879e-01 -1.01109385e+00 4.70166206e-01
6.92714453e-01 -1.94859400e-01 4.35113251e-01 -1.83012530e-01
-1.69364840e-01 -4.01929647e-01 -1.24961638e+00 5.16193390e-01
8.67943168e-01 -4.87848252e-01 -3.54421794e-01 2.02689961e-01
4.14888144e-01 -4.66984391e-01 -8.25381935e-01 3.42159897e-01
3.82891238e-01 -1.30937147e+00 1.02111554e+00 -5.57933822e-02
7.66152322e-01 -4.87451166e-01 1.66325614e-01 -1.40180266e+00
-1.50260344e-01 -7.08803415e-01 1.20524526e-01 1.65405476e+00
1.63443573e-02 -4.05085862e-01 6.28077030e-01 7.61998713e-01
-2.75951505e-01 -3.16867530e-01 -4.44436163e-01 -6.38864875e-01
-2.95548648e-01 -2.41292074e-01 2.80497044e-01 8.16230714e-01
-6.46609724e-01 1.47035763e-01 -8.03651512e-01 2.68479675e-01
7.59068310e-01 -2.00236946e-01 1.05721152e+00 -8.17095399e-01
-4.54702407e-01 -1.38686180e-01 -3.27194691e-01 -5.98176360e-01
-1.69509441e-01 -3.97754490e-01 1.92436010e-01 -1.59436023e+00
2.25951299e-01 -1.23929560e-01 1.02419280e-01 4.30833459e-01
-4.72249836e-01 8.86270344e-01 2.51196235e-01 -4.88987565e-02
-2.21278816e-01 3.70874554e-01 1.41136336e+00 -2.46073246e-01
-8.01253393e-02 -4.56116647e-02 -8.70908499e-01 6.90719604e-01
8.29136014e-01 -3.42402369e-01 -3.29159737e-01 -6.38284922e-01
2.75071472e-01 -4.77061898e-01 7.19675004e-01 -1.20821142e+00
2.09251210e-01 1.28513128e-01 5.23670137e-01 -2.21320122e-01
4.79607463e-01 -1.00349283e+00 4.89737481e-01 2.35674128e-01
2.81863641e-02 -2.16277704e-01 2.44819790e-01 6.25451386e-01
-2.57057458e-01 -3.82012814e-01 1.04768395e+00 -1.32349119e-01
-6.95300817e-01 2.00580433e-02 -1.54358491e-01 -3.04431230e-01
1.12266755e+00 -4.35443223e-01 -2.76483715e-01 -5.47653913e-01
-6.53364480e-01 -2.57490367e-01 9.68602836e-01 1.72780916e-01
6.70958817e-01 -9.50996697e-01 -5.61429501e-01 7.16584325e-02
-1.09859578e-01 -2.89755724e-02 8.45062673e-01 8.29979241e-01
-7.93472111e-01 -5.15574455e-01 -3.30957502e-01 -5.07594049e-01
-1.48689878e+00 6.60561383e-01 1.84573561e-01 -6.13707975e-02
-8.25452805e-01 7.16910779e-01 -6.33314773e-02 1.01026595e-01
1.23929158e-01 -8.58713612e-02 1.01599723e-01 -1.80341020e-01
6.85018301e-01 4.83984917e-01 3.96019995e-01 -1.52556047e-01
2.75657997e-02 7.03249454e-01 -1.43592939e-01 9.47053879e-02
1.64009678e+00 -2.59601891e-01 -1.78170979e-01 -6.56856671e-02
8.24882865e-01 1.39671728e-01 -1.51137924e+00 6.06965199e-02
-4.35808033e-01 -8.98054600e-01 1.81028321e-01 -1.06922662e+00
-1.34838605e+00 8.67264748e-01 7.49573946e-01 2.51231879e-01
1.73367023e+00 -3.38929117e-01 9.41774547e-01 -4.43356447e-02
3.26272637e-01 -1.27668869e+00 3.67912263e-01 -1.45250738e-01
1.03857803e+00 -1.37201715e+00 2.21383452e-01 -5.91140687e-01
-8.66730809e-01 9.95513082e-01 4.47394937e-01 -2.51427978e-01
2.01287746e-01 3.72420847e-01 3.04522574e-01 -7.98062086e-02
-1.79494292e-01 -2.21551612e-01 2.11864516e-01 8.82553220e-01
2.37373680e-01 -3.54005128e-01 -1.86275363e-01 2.03057900e-01
6.37056082e-02 1.71076730e-01 6.98792994e-01 9.52985764e-01
-1.02501094e-01 -1.19586277e+00 -5.81178904e-01 1.91900834e-01
-6.16336107e-01 -1.14596866e-01 -4.39735860e-01 8.13139915e-01
2.40865946e-01 1.09031618e+00 -2.63422817e-01 -2.56731272e-01
3.95936310e-01 -1.52246714e-01 6.60848677e-01 -3.68689507e-01
-9.10740018e-01 7.98974559e-02 -7.19559416e-02 -6.46736920e-01
-7.82528341e-01 -4.66285050e-01 -6.78894520e-01 -4.56331491e-01
-3.61515582e-01 -3.69838446e-01 9.34219182e-01 6.48718357e-01
3.11141640e-01 7.92098820e-01 5.95261216e-01 -9.25410032e-01
-3.96129817e-01 -9.93458211e-01 -6.86519504e-01 8.71202528e-01
1.60598710e-01 -4.70661372e-01 -3.38821918e-01 6.15717351e-01] | [11.113311767578125, -2.051558017730713] |
9a326b1e-4635-4045-b95b-28da6bc36625 | a-deep-learning-technique-using-low-sampling | 2111.05120 | null | https://arxiv.org/abs/2111.05120v1 | https://arxiv.org/pdf/2111.05120v1.pdf | A Deep Learning Technique using Low Sampling rate for residential Non Intrusive Load Monitoring | Individual device loads and energy consumption feedback is one of the important approaches for pursuing users to save energy in residences. This can help in identifying faulty devices and wasted energy by devices when left On unused. The main challenge is to identity and estimate the energy consumption of individual devices without intrusive sensors on each device. Non-intrusive load monitoring (NILM) or energy disaggregation, is a blind source separation problem which requires a system to estimate the electricity usage of individual appliances from the aggregated household energy consumption. In this paper, we propose a novel deep neural network-based approach for performing load disaggregation on low frequency power data obtained from residential households. We combine a series of one-dimensional Convolutional Neural Networks and Long Short Term Memory (1D CNN-LSTM) to extract features that can identify active appliances and retrieve their power consumption given the aggregated household power value. We used CNNs to extract features from main readings in a given time frame and then used those features to classify if a given appliance is active at that time period or not. Following that, the extracted features are used to model a generation problem using LSTM. We train the LSTM to generate the disaggregated energy consumption of a particular appliance. Our neural network is capable of generating detailed feedback of demand-side, providing vital insights to the end-user about their electricity consumption. The algorithm was designed for low power offline devices such as ESP32. Empirical calculations show that our model outperforms the state-of-the-art on the Reference Energy Disaggregation Dataset (REDD). | ['Raghunath Reddy', 'Vishal Garg', 'Sahil Chilana', 'Ronak Aghera'] | 2021-11-07 | null | null | null | null | ['non-intrusive-load-monitoring', 'non-intrusive-load-monitoring', 'non-intrusive-load-monitoring'] | ['knowledge-base', 'miscellaneous', 'time-series'] | [-3.91629199e-03 -1.70106683e-02 -1.12721100e-01 -2.87153363e-01
-8.30599010e-01 -5.74531913e-01 3.79701734e-01 -2.64439225e-01
1.42835900e-01 8.16119492e-01 4.01053846e-01 -3.15881312e-01
1.15112402e-02 -1.23508763e+00 -5.80749154e-01 -1.01079929e+00
5.51311933e-02 2.28527024e-01 -8.29614222e-01 2.92080015e-01
-3.80335391e-01 6.07507646e-01 -1.52055001e+00 1.56924695e-01
6.10566318e-01 1.53679013e+00 1.40414182e-02 7.02332199e-01
1.70369059e-01 1.03574741e+00 -9.79405224e-01 2.57754594e-01
6.59200922e-02 -3.38811129e-01 -8.05301309e-01 -2.61310250e-01
-2.26136118e-01 -1.14822268e+00 -2.93523669e-01 9.65584815e-01
8.62627208e-01 1.63638163e-02 6.66643262e-01 -1.51161480e+00
-6.51186943e-01 1.08783817e+00 -2.22011924e-01 3.17333877e-01
3.88787240e-01 1.66378736e-01 5.36576927e-01 -7.05679879e-02
-3.95787805e-01 6.93282902e-01 6.67351305e-01 4.36718792e-01
-1.18979871e+00 -8.64517510e-01 -2.16333404e-01 3.72763306e-01
-1.38580608e+00 -6.00234091e-01 1.17883265e+00 -2.14559153e-01
1.83665454e+00 5.13755381e-01 6.77671015e-01 1.53387129e+00
2.33112648e-01 8.84725511e-01 8.20226490e-01 -1.66782796e-01
6.72278225e-01 -2.12897994e-02 4.47265804e-02 4.30943131e-01
2.04308257e-01 8.43243748e-02 -2.07910985e-01 -4.02720511e-01
4.05284353e-02 2.20179453e-01 -1.90939233e-01 4.47416782e-01
-7.59767711e-01 4.84875649e-01 4.23152655e-01 8.17406416e-01
-6.26154602e-01 4.55635786e-01 3.76041710e-01 1.83734559e-02
6.70738757e-01 4.74930778e-02 -6.52907133e-01 -2.66322404e-01
-1.26073086e+00 -2.77041912e-01 1.04250348e+00 6.99477553e-01
6.76881850e-01 3.90107363e-01 -3.14580053e-01 2.91249096e-01
2.89061546e-01 9.42419648e-01 9.62979615e-01 -8.12474549e-01
2.90770799e-01 7.49678433e-01 2.28546917e-01 -5.80154657e-01
-7.94142723e-01 1.33632869e-01 -1.44235075e+00 1.60813943e-01
-3.51049304e-02 -6.52082920e-01 -6.84372485e-01 1.79824865e+00
-6.13923110e-02 1.46613792e-01 1.14746571e-01 4.15468216e-01
5.64060688e-01 9.38211322e-01 1.69096574e-01 -4.34500366e-01
1.31017661e+00 -5.37473261e-01 -1.01731133e+00 9.24721211e-02
3.74231219e-01 -7.76768103e-02 5.45045316e-01 1.19338535e-01
-8.89767051e-01 -5.19737005e-01 -1.33330882e+00 -7.24646598e-02
-1.07366967e+00 4.14414555e-01 2.05942407e-01 7.21096754e-01
-9.90361035e-01 1.00516939e+00 -1.22018182e+00 -4.10813391e-01
6.53950334e-01 7.18931019e-01 2.15048954e-01 8.83045673e-01
-1.24362743e+00 9.19366300e-01 6.15621328e-01 3.00654322e-01
-8.46173584e-01 -5.87595701e-01 -7.90639162e-01 3.63376677e-01
-2.14298010e-01 -5.31915009e-01 1.39765215e+00 -8.76154602e-01
-1.67064667e+00 3.87889266e-01 -6.87092319e-02 -8.13806117e-01
4.03319448e-01 -1.80741161e-01 -9.16306615e-01 -3.61440629e-01
4.58993390e-02 7.78883025e-02 9.96445656e-01 -8.22333872e-01
-8.43872726e-01 -5.58475733e-01 -3.57591152e-01 -2.17749208e-01
-4.56916571e-01 -6.91928029e-01 5.77916443e-01 -3.62831503e-01
-5.67786574e-01 -5.93268275e-01 4.41258460e-01 -7.55322099e-01
-9.35466826e-01 -7.07121432e-01 1.34928334e+00 -1.15730357e+00
1.34403849e+00 -1.82196963e+00 -5.16772985e-01 1.63639620e-01
-3.24649736e-02 3.00583124e-01 3.97571951e-01 2.68075943e-01
-3.15945178e-01 -1.38990236e-02 -1.64031878e-01 -6.54677272e-01
6.35469377e-01 2.56046474e-01 -5.52453697e-01 4.92034823e-01
-2.49575581e-02 1.47553897e+00 -9.66053486e-01 5.80580533e-02
9.26646709e-01 7.78954089e-01 4.93836135e-01 5.98086655e-01
-3.48453373e-01 2.42786244e-01 -3.15499485e-01 5.96541464e-01
6.14901781e-01 -1.90132067e-01 2.47056678e-01 -7.68364370e-01
-3.83668952e-02 5.86232901e-01 -9.50922191e-01 1.50757027e+00
-9.35753286e-01 6.93530917e-01 -2.51428246e-01 -8.73692274e-01
3.98201168e-01 6.91146612e-01 7.88756251e-01 -9.70437288e-01
6.13959193e-01 2.37592444e-01 -6.98292196e-01 -5.51337957e-01
2.32661292e-01 3.06260169e-01 -4.22425270e-01 8.20876241e-01
4.05220911e-02 4.11299407e-01 -5.34110330e-02 -5.24895310e-01
1.39785898e+00 1.97675787e-02 5.28428495e-01 -3.43421608e-01
3.14056814e-01 -6.25218511e-01 1.76740065e-01 4.03729081e-01
-4.12283316e-02 -7.13949129e-02 -8.63113701e-02 -6.08876765e-01
-7.70757616e-01 -1.00068486e+00 5.88278994e-02 9.85722065e-01
-4.69864368e-01 1.20807514e-01 -1.12702477e+00 -6.92159355e-01
3.38118263e-02 1.41302419e+00 -6.49988830e-01 -2.52007216e-01
-6.40301764e-01 -1.10190856e+00 5.92787683e-01 7.95211136e-01
1.01767039e+00 -1.23376942e+00 -1.27509522e+00 4.57563490e-01
-4.27577168e-01 -7.44042158e-01 -4.84083652e-01 8.88755679e-01
-2.95818150e-01 -1.13495851e+00 -4.57526445e-01 -5.51641822e-01
5.11749864e-01 -2.90369272e-01 1.32694149e+00 -4.31394577e-01
7.55563471e-03 2.72707343e-01 2.08844289e-01 -5.21307290e-01
-4.74136680e-01 6.38970613e-01 3.44351083e-01 2.39420589e-02
1.04979014e+00 -1.18235445e+00 -7.14794874e-01 -6.85351044e-02
-5.54259479e-01 -1.18850105e-01 2.48165771e-01 2.09861726e-01
3.31529140e-01 6.95005953e-01 6.74174368e-01 -4.70580578e-01
6.94030881e-01 -8.06222796e-01 -8.24336410e-01 1.53287858e-01
-9.85255182e-01 1.53207377e-01 1.22476792e+00 -2.50091076e-01
-8.23838890e-01 2.46685386e-01 -1.54097930e-01 -3.41689855e-01
-4.13266122e-01 -4.85065639e-01 -7.67100692e-01 6.33653998e-01
1.59370288e-01 2.85623908e-01 -9.21330571e-01 -7.78118432e-01
1.83500618e-01 1.07210624e+00 8.51222575e-01 -1.26403809e-01
5.60561776e-01 3.18522692e-01 -1.23941869e-01 -5.39876461e-01
-6.38276696e-01 8.09612405e-03 -4.97402638e-01 9.40750539e-02
9.69497800e-01 -9.96029079e-01 -1.49288464e+00 8.95677388e-01
-1.28346503e+00 -6.18146956e-01 -7.34065592e-01 -2.10757181e-01
-4.19263065e-01 -1.54650703e-01 -3.81775051e-01 -1.20357597e+00
-1.11216819e+00 -8.02754521e-01 1.46096969e+00 4.24392700e-01
-5.72631001e-01 -1.11841679e+00 5.90446107e-02 1.21032503e-02
5.76941729e-01 6.95352733e-01 8.63924325e-01 -5.81547797e-01
-2.52899706e-01 -1.91674292e-01 1.16500758e-01 5.67375064e-01
8.92599225e-01 -4.51995283e-01 -1.62832010e+00 -4.70550150e-01
4.75954056e-01 1.69229861e-02 5.39707959e-01 6.35218799e-01
1.49771214e+00 -1.01706004e+00 -4.85003591e-01 7.10692227e-01
1.59703541e+00 4.02501762e-01 5.01152873e-01 -1.71795473e-01
7.28664398e-01 -1.85855716e-01 -5.04411638e-01 4.98200417e-01
6.13927126e-01 2.43162692e-01 3.77465308e-01 -1.96111381e-01
1.43536612e-01 -4.30930048e-01 5.98640323e-01 6.10147893e-01
2.03865170e-01 -8.09644282e-01 -3.18911582e-01 7.71058261e-01
-1.87276936e+00 -9.69345450e-01 3.17523986e-01 1.81273305e+00
6.86803460e-01 -1.07718751e-01 2.52379179e-01 6.67255342e-01
5.30971825e-01 2.55988568e-01 -1.22181499e+00 -4.58319843e-01
3.46197076e-02 2.98486978e-01 8.77240956e-01 3.24166626e-01
-1.12099159e+00 -1.07653774e-02 5.86088419e+00 5.97469091e-01
-8.24334562e-01 4.17288065e-01 8.98489118e-01 -3.01050365e-01
7.64712170e-02 -9.19405282e-01 -5.33694267e-01 1.27441871e+00
1.59356320e+00 -2.27848992e-01 1.01954710e+00 8.26043844e-01
5.60608566e-01 -3.43365014e-01 -1.71894956e+00 1.28335011e+00
-2.42013916e-01 -8.12444150e-01 -4.02422458e-01 1.98706806e-01
8.19483638e-01 1.56433553e-01 -1.62361234e-01 -1.66870411e-02
8.06425393e-01 -1.09389699e+00 4.51576769e-01 8.70399714e-01
7.21865535e-01 -9.12391305e-01 9.30217326e-01 3.48532051e-01
-1.59541976e+00 -2.83072501e-01 1.76145464e-01 -1.38722256e-01
1.89419642e-01 9.71633434e-01 -6.77260697e-01 4.69593883e-01
1.04416120e+00 4.65816081e-01 -2.96605617e-01 -9.35254246e-02
-1.89356551e-01 5.41903734e-01 -8.50873470e-01 8.87579843e-03
-1.98700011e-01 2.06506755e-02 -7.68929943e-02 1.08874273e+00
5.84134281e-01 -4.58098613e-02 -2.47870192e-01 1.34907389e+00
-4.61433500e-01 -8.08195770e-01 -6.90528810e-01 1.87144466e-02
6.06597245e-01 1.45701623e+00 -4.08692956e-01 -5.10330021e-01
-1.77621961e-01 1.51476860e+00 -6.95675910e-02 4.23271418e-01
-8.77696574e-01 -3.40825349e-01 8.29365194e-01 -1.73821002e-01
3.11415821e-01 1.99121267e-01 -2.68013239e-01 -9.09516513e-01
1.22143932e-01 -1.76384211e-01 2.82090455e-01 -7.95118630e-01
-1.53793502e+00 2.11924434e-01 -1.06499381e-01 -7.35148191e-01
-1.00142086e+00 -2.47838244e-01 -8.88462663e-01 9.61175859e-01
-1.36665988e+00 -1.18744457e+00 -5.09202361e-01 7.30803668e-01
5.29790044e-01 3.36766802e-02 1.21806848e+00 3.94390404e-01
-9.14426029e-01 4.35896575e-01 5.65769672e-01 4.28322524e-01
-2.29972005e-01 -1.64412200e+00 7.66786158e-01 7.63978481e-01
-1.58708945e-01 -1.25389367e-01 4.57478791e-01 -5.70684731e-01
-1.35001028e+00 -1.66898441e+00 8.16213548e-01 -6.59703493e-01
2.89000511e-01 -5.48727870e-01 -4.04455692e-01 9.04227495e-01
5.99739432e-01 -1.47724271e-01 7.06194997e-01 -6.81385994e-01
3.97064179e-01 -3.76116782e-01 -1.65591061e+00 -2.79593021e-02
8.19133461e-01 -8.72709095e-01 -5.23936450e-01 2.23961458e-01
3.19702089e-01 -7.85121173e-02 -1.08831906e+00 1.69220671e-01
4.83198166e-01 -9.52405989e-01 6.70371473e-01 -1.70910936e-02
-1.42206341e-01 -3.25174868e-01 -4.41438675e-01 -1.55452740e+00
-3.33840221e-01 -8.29831302e-01 -1.36242640e+00 1.77958786e+00
-9.12525728e-02 -7.24313796e-01 4.97319937e-01 9.51045215e-01
3.53765130e-01 -4.42188770e-01 -1.06206787e+00 -4.80670989e-01
-1.96471274e-01 -4.35292363e-01 1.58406806e+00 6.18014932e-01
-6.54441044e-02 3.81123394e-01 -2.14809537e-01 4.55985963e-01
6.64632916e-01 1.71907455e-01 1.52602509e-01 -9.87542093e-01
5.05857058e-02 -2.96682298e-01 7.17714382e-03 -6.25846803e-01
3.86423141e-01 -6.50028348e-01 -8.73181503e-03 -1.67535269e+00
-9.24272090e-02 3.53739083e-01 -5.82708120e-01 9.46989536e-01
3.62073630e-01 2.74759591e-01 -1.94237038e-01 -1.31576598e-01
-2.72440940e-01 5.07808268e-01 -4.68278788e-02 -6.76553667e-01
-3.27638626e-01 1.95360169e-01 -6.40999436e-01 8.34501266e-01
1.21046329e+00 -1.79187983e-01 -3.17797214e-01 -4.04105365e-01
2.11041588e-02 -4.13674831e-01 5.91290534e-01 -1.31477499e+00
6.92977235e-02 2.83568591e-01 1.02810049e+00 -9.85298991e-01
7.41046965e-02 -1.42368126e+00 7.78726280e-01 6.05166137e-01
1.06223645e-02 4.63548154e-02 1.59451321e-01 1.88142732e-01
5.38696826e-01 2.68715590e-01 4.19422567e-01 -2.19793648e-01
-5.23756027e-01 7.39515871e-02 -6.28606617e-01 -5.02908647e-01
8.95211637e-01 2.24385172e-01 -2.73782045e-01 -2.52412230e-01
-6.30933642e-01 3.74444604e-01 2.21752048e-01 5.89452326e-01
-1.05845265e-01 -1.88309193e+00 -1.68234900e-01 5.15924275e-01
-4.46942896e-01 -2.01394446e-02 5.35720475e-02 1.87601313e-01
2.28789374e-01 7.00720549e-01 1.78372264e-01 -4.36927080e-01
-7.43086874e-01 7.02236354e-01 9.98034418e-01 -4.66300666e-01
-3.99801254e-01 -2.64931712e-02 -3.48604500e-01 -2.21249480e-02
3.53827506e-01 -1.17750275e+00 -3.16836834e-02 3.83930862e-01
6.97668791e-01 1.13984561e+00 5.73965549e-01 -5.94597578e-01
-5.15378058e-01 4.94456172e-01 6.72784388e-01 5.29851556e-01
1.38770759e+00 -4.41526443e-01 -8.80619735e-02 8.39661002e-01
1.75281453e+00 -6.52806580e-01 -1.14900136e+00 7.13857338e-02
-1.95060983e-01 3.25068384e-01 4.00005281e-01 -1.23271346e+00
-1.44155025e+00 5.51089525e-01 1.41260743e+00 1.05786645e+00
1.76215518e+00 -1.31366387e-01 1.51082265e+00 3.12615395e-01
4.37036574e-01 -1.47909200e+00 -7.28838086e-01 -7.82081410e-02
3.71866286e-01 -1.07457173e+00 -3.32055300e-01 6.51593208e-01
4.49307323e-01 1.00457001e+00 2.06030175e-01 1.96249142e-01
8.55316341e-01 6.35291338e-01 -3.20892274e-01 -1.08373269e-01
-2.64961809e-01 -2.54436582e-01 5.42726107e-02 6.67000055e-01
-8.30277521e-03 5.18815339e-01 5.32386899e-01 9.71246302e-01
-4.67915565e-01 4.88540828e-01 -5.73834293e-02 7.22931862e-01
-3.40386759e-04 -5.12538493e-01 -3.00844431e-01 8.66383076e-01
-4.42040175e-01 2.32784271e-01 -7.98118487e-02 1.03656471e-01
7.27036119e-01 1.38740683e+00 4.56702650e-01 -4.34974194e-01
5.18328905e-01 4.74956304e-01 1.83887854e-01 1.18653230e-01
-5.63856125e-01 -4.63408083e-01 -1.40430853e-01 -7.44323254e-01
-4.99635994e-01 -3.66058886e-01 -1.09215534e+00 -7.54386008e-01
-1.36611953e-01 -2.90814489e-01 5.60746789e-01 1.17050481e+00
4.56568778e-01 1.04169297e+00 1.18567228e+00 -1.39158750e+00
-3.57162714e-01 -1.15214801e+00 -8.08224738e-01 3.61910731e-01
9.61092412e-01 -1.76391035e-01 -4.43169922e-01 1.08484440e-01] | [16.0660400390625, 7.58033561706543] |
4bcce555-49f2-4dba-936b-ce652afd10f9 | neural-audio-fingerprint-for-high-specific | 2010.11910 | null | https://arxiv.org/abs/2010.11910v4 | https://arxiv.org/pdf/2010.11910v4.pdf | Neural Audio Fingerprint for High-specific Audio Retrieval based on Contrastive Learning | Most of existing audio fingerprinting systems have limitations to be used for high-specific audio retrieval at scale. In this work, we generate a low-dimensional representation from a short unit segment of audio, and couple this fingerprint with a fast maximum inner-product search. To this end, we present a contrastive learning framework that derives from the segment-level search objective. Each update in training uses a batch consisting of a set of pseudo labels, randomly selected original samples, and their augmented replicas. These replicas can simulate the degrading effects on original audio signals by applying small time offsets and various types of distortions, such as background noise and room/microphone impulse responses. In the segment-level search task, where the conventional audio fingerprinting systems used to fail, our system using 10x smaller storage has shown promising results. Our code and dataset are available at \url{https://mimbres.github.io/neural-audio-fp/}. | ['Yoonchang Han', 'Karam Ko', 'Kyogu Lee', 'Hyungui Lim', 'Jeongsoo Park', 'Donmoon Lee', 'Sungkyun Chang'] | 2020-10-22 | null | null | null | null | ['audio-fingerprint'] | ['audio'] | [ 4.90474761e-01 -3.00410807e-01 -2.93859784e-02 -3.43324542e-01
-1.36956942e+00 -6.71781719e-01 2.73561299e-01 -2.86313206e-01
-2.14462012e-01 6.12637401e-01 6.89487383e-02 6.02190010e-02
-5.32847084e-02 -3.80131513e-01 -7.75247335e-01 -6.30006552e-01
-5.71524680e-01 2.57088035e-01 2.50522166e-01 3.23376685e-01
2.05194846e-01 1.76838174e-01 -1.99738264e+00 6.48774505e-01
2.76992500e-01 1.28259706e+00 2.25563392e-01 8.92730236e-01
3.12687069e-01 3.40842545e-01 -9.26295996e-01 -1.06276363e-01
3.17990303e-01 -9.28637534e-02 -4.99589741e-01 -6.97346330e-01
6.30362988e-01 -5.48390567e-01 -5.69520950e-01 1.02356982e+00
1.17018950e+00 2.27003977e-01 3.96614820e-01 -1.32502782e+00
-2.40426660e-01 7.85288990e-01 -1.85464457e-01 -1.11457817e-02
8.11380923e-01 -1.14546604e-01 8.43434751e-01 -1.04851580e+00
2.37704560e-01 1.15996683e+00 9.96202826e-01 6.31991565e-01
-1.18181002e+00 -1.22238457e+00 -4.02032375e-01 2.62069792e-01
-1.76996541e+00 -9.60503101e-01 8.51677775e-01 -2.08878219e-01
6.95108473e-01 6.25719309e-01 2.88877636e-01 1.41905379e+00
-1.71694428e-01 8.50760937e-01 7.89298534e-01 -4.37608778e-01
3.14517021e-01 1.09754600e-01 -1.28532171e-01 3.75468910e-01
-2.57375538e-01 3.50398272e-01 -1.10409796e+00 -7.77444303e-01
4.99107510e-01 -2.89806604e-01 -4.00501311e-01 -1.04821417e-02
-1.20359504e+00 3.37385297e-01 -6.58797994e-02 4.95965108e-02
-6.45418689e-02 6.39419258e-01 5.35321832e-01 6.06841266e-01
3.28067690e-01 1.64227322e-01 -3.98239225e-01 -4.71918166e-01
-1.39682472e+00 2.00967133e-01 7.94818223e-01 8.34166884e-01
6.26427054e-01 -8.64379923e-04 -1.83301166e-01 1.15065742e+00
3.02726388e-01 6.09777868e-01 6.21891439e-01 -1.31463003e+00
4.14940178e-01 -3.38147610e-01 1.95321143e-02 -8.68507266e-01
-1.88432515e-01 -4.06685919e-01 -7.27420270e-01 -4.42075878e-02
2.79454470e-01 -1.46823168e-01 -5.82619429e-01 1.84431517e+00
3.59139055e-01 9.69171226e-01 -3.30548704e-01 9.78499532e-01
4.89846855e-01 6.91214263e-01 -3.68470937e-01 -5.62851280e-02
1.16359913e+00 -8.56576681e-01 -5.63231528e-01 7.97061846e-02
4.64519054e-01 -1.17609048e+00 1.18237722e+00 7.31698930e-01
-9.56682384e-01 -8.46554339e-01 -1.24020576e+00 1.73271805e-01
-3.02782774e-01 2.10201055e-01 4.68912572e-01 9.80821550e-01
-1.28161919e+00 7.50481308e-01 -6.02717459e-01 9.11399350e-02
4.03639898e-02 4.32260782e-01 -1.18479431e-01 1.47712991e-01
-1.42124820e+00 -1.06349468e-01 1.27675191e-01 -1.82684546e-03
-9.88184094e-01 -8.49239230e-01 -4.15700227e-01 -3.94685417e-02
6.48324490e-02 -3.41988653e-01 1.31568837e+00 -5.78364432e-01
-1.57900441e+00 5.88963211e-01 -1.18763804e-01 -3.69864643e-01
4.32901174e-01 -4.12288368e-01 -6.92532897e-01 1.29133299e-01
-8.30866620e-02 5.46036720e-01 1.33684790e+00 -9.11350131e-01
-3.74252439e-01 -1.95544749e-01 -4.21573699e-01 -1.17491484e-01
-5.35460055e-01 1.40282258e-01 -6.72409117e-01 -1.00314891e+00
1.23322733e-01 -1.16754293e+00 9.33045968e-02 1.03529356e-02
-3.58174354e-01 2.32724011e-01 6.81722879e-01 -6.65988982e-01
1.61168563e+00 -2.53981280e+00 -4.40982580e-01 4.11658764e-01
-1.52222350e-01 2.62440294e-01 -4.23564285e-01 6.63039446e-01
-5.18463887e-02 7.52548203e-02 8.92311782e-02 -6.26382589e-01
3.36805940e-01 -3.20411950e-01 -8.10768247e-01 2.89564192e-01
-2.74532974e-01 2.74893969e-01 -9.15587246e-01 -4.50003922e-01
-1.92052528e-01 6.31999433e-01 -5.91945887e-01 2.48364583e-01
-2.87732426e-02 1.37080282e-01 -3.45842004e-01 8.12941790e-01
7.79907405e-01 1.17120773e-01 5.34699224e-02 -2.41537839e-01
1.71324760e-01 5.28726935e-01 -1.70564806e+00 2.06722450e+00
-3.09788048e-01 6.29022181e-01 3.05377990e-01 -6.36162400e-01
8.01850140e-01 5.64412355e-01 5.67876101e-01 -5.55560946e-01
-2.58015662e-01 4.47011262e-01 -5.20043790e-01 -3.50666642e-01
5.52632928e-01 4.61294502e-01 -1.08696632e-01 9.12474513e-01
1.02242358e-01 -2.67971568e-02 -1.61173016e-01 6.43522069e-02
1.31768525e+00 5.96803762e-02 -3.19276959e-01 2.18155727e-01
4.30994809e-01 -6.71364188e-01 4.67853308e-01 1.06745732e+00
-1.49015889e-01 9.76048708e-01 2.86801811e-02 -1.19477943e-01
-9.00984704e-01 -1.18141603e+00 -4.03882951e-01 1.41910768e+00
-7.63721243e-02 -6.35290205e-01 -7.87552059e-01 -1.96095929e-01
1.80670798e-01 3.23986918e-01 -1.72464132e-01 -2.28460342e-01
-5.49756706e-01 -4.04350400e-01 1.46082664e+00 1.38068140e-01
2.62809843e-01 -9.14259851e-01 -2.99496055e-01 3.04625064e-01
-1.60548493e-01 -7.09004462e-01 -9.81551111e-01 1.60485476e-01
-6.80913150e-01 -6.43595874e-01 -7.81322420e-01 -8.07241976e-01
1.30008265e-01 1.02087870e-01 7.86997676e-01 -1.71629116e-01
-4.15986955e-01 3.94976556e-01 -1.97285026e-01 -1.95963174e-01
-3.32225442e-01 1.63891867e-01 5.94960272e-01 1.00202948e-01
4.11616080e-02 -1.05362785e+00 -6.74992740e-01 5.10661185e-01
-8.27063441e-01 -4.91965055e-01 2.67654538e-01 9.20098960e-01
6.93129003e-01 -1.17445908e-01 6.15324080e-01 -3.23719352e-01
9.51390922e-01 -3.96313131e-01 -4.76608962e-01 2.18128264e-01
-6.00501597e-01 -4.05231789e-02 4.24674332e-01 -9.69717562e-01
-6.27235591e-01 1.42359897e-01 -2.23593041e-01 -6.42778933e-01
-4.69782725e-02 2.16226593e-01 -3.90646756e-02 -1.37208730e-01
8.19254994e-01 4.37018603e-01 -9.01847854e-02 -8.76230359e-01
2.92677552e-01 1.15140450e+00 5.98889828e-01 -8.66887987e-01
6.30456150e-01 1.81848675e-01 -3.22665930e-01 -6.74532294e-01
-2.32427314e-01 -5.34375668e-01 -2.15836152e-01 -2.97749341e-01
-6.21793233e-02 -9.21466172e-01 -8.01073432e-01 5.95748186e-01
-9.61633265e-01 -4.09938514e-01 -3.20798218e-01 6.85546160e-01
-6.02153897e-01 4.55154032e-01 -7.06113338e-01 -1.07173407e+00
-5.48884451e-01 -9.93698061e-01 1.33471513e+00 1.85512044e-02
-1.64980009e-01 -3.11116070e-01 5.65798223e-01 2.56695691e-02
4.33898002e-01 -2.40009964e-01 4.32149053e-01 -6.68574095e-01
-5.82824767e-01 -3.98624778e-01 7.94173181e-02 4.18885678e-01
6.12393543e-02 2.18672417e-02 -1.63493049e+00 -5.41733623e-01
-1.93352580e-01 -5.13639510e-01 7.49668241e-01 1.18676484e-01
1.39336002e+00 -4.15053040e-01 -4.81905490e-01 8.77406299e-01
1.02235115e+00 2.43970439e-01 3.73206586e-01 7.33450949e-02
1.63769528e-01 2.40091160e-01 6.88329101e-01 8.65877271e-01
-1.96356952e-01 1.01806915e+00 1.07095882e-01 2.89973110e-01
-3.14296246e-01 -4.53541249e-01 5.72607696e-01 1.04917991e+00
2.52846956e-01 -1.90625846e-01 -7.09121883e-01 3.31237644e-01
-1.57106626e+00 -1.15968287e+00 3.90528172e-01 2.66432977e+00
1.16642249e+00 -1.75763648e-02 1.70955673e-01 3.73573869e-01
7.76505053e-01 3.33070934e-01 -6.90594077e-01 -8.89205337e-02
-7.47401789e-02 3.81714940e-01 2.33653381e-01 3.21273029e-01
-1.02879822e+00 5.24353445e-01 6.55741405e+00 1.15416753e+00
-1.29030275e+00 2.31713116e-01 3.14318478e-01 -5.92539728e-01
-2.58394212e-01 -2.38718107e-01 -7.14960873e-01 8.92562807e-01
1.41221964e+00 -2.90776212e-02 9.47552264e-01 7.64773548e-01
4.57168706e-02 3.62739682e-01 -1.13027680e+00 1.33279276e+00
-6.62743822e-02 -1.08691776e+00 -2.47627214e-01 -4.74267863e-02
4.41083103e-01 1.56134516e-01 5.41380465e-01 2.78581619e-01
-2.27019608e-01 -7.12277710e-01 9.75158513e-01 6.92447424e-01
1.28428662e+00 -4.76403743e-01 3.48382831e-01 -1.47490148e-02
-1.36323512e+00 -1.00210421e-01 -2.99088389e-01 1.80203661e-01
1.00571223e-01 8.14110518e-01 -1.02733505e+00 2.33090788e-01
9.24936533e-01 1.56226680e-01 -3.85505944e-01 1.29412663e+00
2.21721247e-01 9.61196423e-01 -8.89666259e-01 6.92201927e-02
-3.66777837e-01 4.79160845e-02 5.70199668e-01 1.45279968e+00
7.85897493e-01 -4.48353529e-01 -2.41191655e-01 4.99121577e-01
-2.29772732e-01 4.86201532e-02 -4.66239154e-01 8.04301426e-02
1.18936563e+00 8.93012047e-01 -1.88244268e-01 -2.90041536e-01
-4.51531522e-02 1.06092000e+00 6.41834959e-02 5.05601048e-01
-9.67402220e-01 -8.51847231e-01 8.40337455e-01 9.80874673e-02
2.03316242e-01 -1.11682214e-01 9.80941355e-02 -9.94300485e-01
3.57519597e-01 -1.06254900e+00 2.77314335e-01 -6.71150863e-01
-1.02071655e+00 3.94726723e-01 -1.93608165e-01 -1.59791744e+00
-6.13927901e-01 -1.15058355e-01 -4.52731371e-01 7.26976812e-01
-9.94914234e-01 -7.10391581e-01 -6.46272376e-02 6.64692521e-01
1.89007282e-01 -2.40134954e-01 1.15604687e+00 1.00907183e+00
-3.39302421e-01 1.37028444e+00 5.42761385e-01 -8.51692483e-02
1.30372429e+00 -8.16603661e-01 6.10221505e-01 3.55109811e-01
4.80897307e-01 6.75790727e-01 6.18195832e-01 -4.77419943e-01
-1.46896315e+00 -8.18156123e-01 7.95942962e-01 -9.08676311e-02
6.43190742e-01 -7.04395294e-01 -9.08772886e-01 3.21927965e-01
-1.83397248e-01 1.70133442e-01 9.37948287e-01 -7.79573340e-03
-6.77732348e-01 -5.70920289e-01 -1.03071201e+00 3.09045911e-01
1.20769227e+00 -1.25993896e+00 -1.74502596e-01 3.53921533e-01
7.11467922e-01 -3.72195601e-01 -9.60503876e-01 3.62791896e-01
1.23317289e+00 -8.51082802e-01 1.33108974e+00 -3.03710580e-01
-3.57703626e-01 -4.19808567e-01 -4.00521606e-01 -8.11736465e-01
-1.63066432e-01 -1.21530092e+00 -4.35147583e-01 1.58167076e+00
5.22348106e-01 -6.80457711e-01 8.49962175e-01 2.51865923e-01
-4.42747474e-02 -4.83185947e-01 -1.28654718e+00 -1.01228499e+00
-4.83805925e-01 -7.59949565e-01 8.92781734e-01 7.31874287e-01
1.74255043e-01 -1.06586240e-01 -7.42608249e-01 3.34752232e-01
8.25079381e-01 -6.09031878e-02 6.94662333e-01 -1.03736556e+00
-7.29044199e-01 -1.71818540e-01 -3.19210112e-01 -1.21889031e+00
-9.67089739e-03 -6.48040056e-01 5.86240031e-02 -6.23211563e-01
-2.94233829e-01 -7.90624857e-01 -7.42140412e-01 3.08310777e-01
1.76498890e-01 5.23472428e-01 8.64386410e-02 4.59078014e-01
-5.70877969e-01 3.83700252e-01 6.19977236e-01 -2.88608849e-01
-3.36916089e-01 3.47712934e-01 7.52729189e-04 2.80068487e-01
1.03665149e+00 -7.91105449e-01 -4.66734976e-01 -4.59108293e-01
1.71059698e-01 1.28119692e-01 1.71144590e-01 -1.37120616e+00
4.54376251e-01 3.18446428e-01 2.15416804e-01 -4.96174932e-01
9.12699759e-01 -5.69594324e-01 7.82401741e-01 3.54770094e-01
-5.49938321e-01 -4.74429801e-02 1.31645247e-01 5.34829676e-01
-3.22131753e-01 -2.91227013e-01 3.50590974e-01 2.28178069e-01
-2.33365089e-01 1.83275014e-01 -2.41048127e-01 -1.05998702e-01
4.17984307e-01 -2.57816404e-01 -2.29290724e-01 -4.37104911e-01
-6.32456899e-01 -2.55849034e-01 3.36147457e-01 3.63767326e-01
6.57440364e-01 -1.63256907e+00 -4.59354758e-01 4.79202271e-01
-8.71077031e-02 -4.93383288e-01 3.29142332e-01 5.09204626e-01
-1.72200456e-01 4.73885655e-01 4.47417609e-02 -5.12274981e-01
-1.31186414e+00 4.10373747e-01 1.47123039e-01 1.70444578e-01
-1.36028960e-01 1.17546570e+00 -2.88760662e-01 -3.31327766e-01
9.61838901e-01 -1.74552798e-01 3.03575128e-01 1.80658415e-01
8.45556557e-01 4.99656945e-01 1.57647878e-01 -2.56633669e-01
-4.04634356e-01 5.86238980e-01 1.69832110e-01 -5.48113227e-01
9.10121322e-01 -1.09137058e-01 2.97586530e-01 5.55947959e-01
1.55586636e+00 2.20148966e-01 -1.11911786e+00 -3.74571443e-01
-1.06041946e-01 -5.62289774e-01 -5.12370728e-02 -5.18618166e-01
-8.38175356e-01 7.62001634e-01 1.15044117e+00 2.71682113e-01
1.19624829e+00 -9.32551026e-02 9.67082202e-01 4.25819010e-01
6.27798319e-01 -1.17305279e+00 2.19334476e-02 2.36559674e-01
7.57812202e-01 -6.64292216e-01 -2.14079112e-01 9.98345762e-02
1.27511583e-02 1.01756585e+00 1.18105158e-01 1.80990115e-01
8.71715546e-01 5.70859313e-01 1.80289567e-01 3.66345555e-01
-9.11431432e-01 2.18158960e-01 1.86289832e-01 7.81527221e-01
3.15196127e-01 5.34495637e-02 2.59694960e-02 6.81848347e-01
-4.52687383e-01 -1.17055038e-02 -4.51566726e-02 8.08570027e-01
-3.18051577e-01 -1.30877841e+00 -6.21793509e-01 4.02374178e-01
-4.72224206e-01 -2.33723164e-01 -2.28042454e-01 8.97946432e-02
1.05810165e-02 7.84912705e-01 3.79261398e-03 -9.66777503e-01
2.08541080e-01 2.23217741e-01 2.77609915e-01 -1.94162831e-01
-4.65672910e-01 1.52660280e-01 1.93737954e-01 -8.11005116e-01
-4.02838290e-02 -7.48345256e-01 -1.01497483e+00 -2.50752121e-01
-2.80999601e-01 2.72549629e-01 8.88071895e-01 1.21785991e-01
6.37336075e-01 1.96619883e-01 9.03479934e-01 -1.11074114e+00
-9.00601447e-01 -9.92759883e-01 -7.74098217e-01 1.53046697e-01
5.71180880e-01 -4.47628349e-01 -6.46458030e-01 1.17510453e-01] | [15.37159252166748, 5.554564952850342] |
e113187c-7688-437b-92ed-dc7289992768 | self-organization-and-artificial-life-a | 1804.01144 | null | http://arxiv.org/abs/1804.01144v1 | http://arxiv.org/pdf/1804.01144v1.pdf | Self-Organization and Artificial Life: A Review | Self-organization has been an important concept within a number of
disciplines, which Artificial Life (ALife) also has heavily utilized since its
inception. The term and its implications, however, are often confusing or
misinterpreted. In this work, we provide a mini-review of self-organization and
its relationship with ALife, aiming at initiating discussions on this important
topic with the interested audience. We first articulate some fundamental
aspects of self-organization, outline its usage, and review its applications to
ALife within its soft, hard, and wet domains. We also provide perspectives for
further research. | ['Justin Werfel', 'Carlos Gershenson', 'Vito Trianni', 'Hiroki Sayama'] | 2018-04-03 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [-7.76560679e-02 1.85630426e-01 -1.46873906e-01 1.72165126e-01
6.86883926e-01 -8.59675884e-01 4.05256569e-01 5.18829405e-01
4.86902148e-02 1.02318585e+00 1.11105636e-01 -1.11199073e-01
-4.64142829e-01 -6.99791133e-01 -2.79532552e-01 -8.44478369e-01
-2.11793184e-01 3.38245958e-01 -4.06955369e-02 -6.00761235e-01
5.85330307e-01 3.83795738e-01 -1.93896294e+00 -1.82639778e-01
1.14733338e+00 7.26228654e-01 3.91727209e-01 4.01300311e-01
-1.44963503e-01 7.08985090e-01 -1.00548124e+00 6.05085772e-03
-2.49280274e-01 -4.32494521e-01 -1.00088286e+00 5.16186059e-01
-3.71535152e-01 3.67631763e-01 -1.81161448e-01 6.91991448e-01
2.73087651e-01 -5.63803082e-03 7.60847270e-01 -1.36004961e+00
-1.00291967e+00 2.12671384e-01 -1.54113472e-01 3.01860154e-01
4.86715019e-01 1.03942059e-01 7.95320451e-01 -5.41356027e-01
6.26712203e-01 9.01896834e-01 5.82067490e-01 4.61387098e-01
-1.17939889e+00 -2.86467075e-01 1.42910555e-01 2.24068955e-01
-1.15941358e+00 -3.10265660e-01 7.81432748e-01 -6.02368295e-01
1.09859288e+00 1.98817417e-01 1.21202636e+00 7.84595430e-01
8.78305376e-01 5.59188366e-01 1.59663975e+00 -7.91264415e-01
4.94675785e-01 2.19342917e-01 4.62109745e-01 2.46067420e-01
1.04801905e+00 5.37130125e-02 -6.73285007e-01 -2.35319138e-01
9.06655908e-01 -3.04183066e-01 2.32406080e-01 -4.86915797e-01
-1.21469259e+00 6.06953263e-01 7.92348012e-02 9.93968487e-01
-3.89577508e-01 5.81554957e-02 2.46904880e-01 5.40983081e-01
3.35130095e-01 9.77020144e-01 -4.26407307e-02 -2.85701066e-01
-3.97780716e-01 1.87180847e-01 1.03782475e+00 5.84998906e-01
6.12312138e-01 1.11928694e-01 3.28386933e-01 6.93922341e-01
1.39193991e-02 1.66744724e-01 6.33220732e-01 -1.15897179e+00
-3.61684263e-01 6.19550228e-01 3.26049805e-01 -1.07332873e+00
-3.41909349e-01 -5.43163180e-01 -8.97712469e-01 1.12174004e-01
-7.14912564e-02 6.69345781e-02 -2.77628064e-01 1.14313221e+00
2.66995728e-01 -4.77264494e-01 3.13565820e-01 5.55847049e-01
7.41431057e-01 4.93227720e-01 -2.08219975e-01 -5.28198838e-01
1.06183863e+00 -1.30357182e+00 -1.37725484e+00 -2.53607720e-01
-3.60542908e-02 -7.11639285e-01 9.39788938e-01 3.25809985e-01
-1.13742959e+00 -3.52821141e-01 -1.00281608e+00 3.22925478e-01
-6.11885369e-01 -4.88231868e-01 1.04446101e+00 7.14951515e-01
-9.73367214e-01 7.03403592e-01 -1.02831733e+00 -1.11605167e+00
7.67260268e-02 2.95650959e-01 -2.33968347e-01 4.12465155e-01
-9.62884605e-01 1.25013041e+00 2.57948428e-01 -1.95707023e-01
-5.06850839e-01 -1.22911073e-01 -5.66228271e-01 -1.54092625e-01
4.79033649e-01 -8.90625358e-01 9.80362654e-01 -7.61044443e-01
-1.80348408e+00 9.42081988e-01 -1.62280008e-01 -5.09260297e-01
1.82772368e-01 -9.23495367e-02 -3.99330229e-01 1.36891395e-01
3.31630141e-01 -9.89021212e-02 5.54077744e-01 -1.38379252e+00
-3.03360075e-01 -3.00154001e-01 1.07090086e-01 2.17825621e-01
-7.67985284e-01 1.14803754e-01 3.77982825e-01 -8.89580131e-01
2.15426356e-01 -7.22944796e-01 -2.44099855e-01 -3.53807241e-01
-1.53427184e-01 -7.24227071e-01 6.92515731e-01 -4.41733934e-02
1.74608564e+00 -1.85477674e+00 1.10734321e-01 -1.36534363e-01
6.40491903e-01 -1.41684301e-02 2.72861332e-01 1.49850631e+00
7.68342540e-02 2.48130277e-01 -2.64794976e-01 6.25042543e-02
-7.62614906e-02 5.33106208e-01 -7.56952167e-02 4.26028103e-01
7.96643272e-02 8.54610920e-01 -1.08405232e+00 -4.21139657e-01
3.23907673e-01 9.79728550e-02 -8.98156781e-03 7.95825645e-02
2.04886258e-01 3.51712584e-01 -3.18760127e-01 1.02169621e+00
2.89229035e-01 -6.28405750e-01 2.22770467e-01 2.33913213e-01
-8.56377959e-01 4.21823785e-02 -8.62256706e-01 1.16185129e+00
-2.03477532e-01 6.70263827e-01 1.95596039e-01 -1.04798293e+00
9.48689044e-01 2.93412119e-01 1.00851345e+00 -3.43218029e-01
-7.77237266e-02 4.83161807e-01 2.61457682e-01 -8.07092071e-01
2.60415345e-01 -2.04157218e-01 -1.69343531e-01 9.13892865e-01
-2.47894358e-02 -3.86356235e-01 5.60639739e-01 -5.14139012e-02
1.35500979e+00 -1.54964149e-01 9.75553393e-01 -9.83003557e-01
7.45698631e-01 9.94543731e-02 3.10089797e-01 6.71973586e-01
-7.06448019e-01 -1.27390921e-02 3.37001234e-01 -8.82256866e-01
-8.52500141e-01 -8.50120425e-01 -3.60972524e-01 9.59952533e-01
5.79989791e-01 -6.59069359e-01 -7.35097885e-01 -3.21444273e-01
2.38693610e-01 4.32560235e-01 -9.39199448e-01 5.35276942e-02
-4.28146899e-01 -7.24639297e-01 -1.38834296e-02 1.59344718e-01
6.95908964e-01 -1.50866604e+00 -1.17694974e+00 4.70864326e-01
-1.57432497e-01 -8.78770053e-01 1.60850704e-01 2.31921583e-01
-8.62129271e-01 -9.44422305e-01 -5.43419302e-01 -7.89094090e-01
7.41100848e-01 7.25426018e-01 1.28984308e+00 3.84972185e-01
-1.94525942e-01 6.90361440e-01 -3.87840331e-01 -5.91823518e-01
-6.27797902e-01 1.73175603e-01 3.18395048e-01 -3.70275527e-01
1.39959335e-01 -7.17560530e-01 -4.38008964e-01 5.30663967e-01
-1.04996645e+00 -1.12350686e-02 3.81746829e-01 9.08283591e-01
2.45680824e-01 2.39287212e-01 9.96510386e-01 -5.34689665e-01
1.01349449e+00 -2.93656349e-01 3.98697285e-03 3.11383337e-01
-6.80591226e-01 -3.18090349e-01 2.89374977e-01 -5.10264896e-02
-6.64177060e-01 -4.06430900e-01 2.95227528e-01 2.94380903e-01
-1.76030606e-01 5.47912955e-01 1.37295038e-01 -4.29065257e-01
5.73239028e-01 1.65872559e-01 6.78022683e-01 -2.21476525e-01
-2.11988345e-01 8.41730952e-01 1.53054679e-02 -5.37786007e-01
5.59647560e-01 7.77995169e-01 4.55819592e-02 -1.24130595e+00
-8.76696110e-01 -3.36809486e-01 -5.02262712e-01 -5.42536616e-01
4.81260806e-01 -2.44702518e-01 -6.63932562e-01 4.23313588e-01
-1.01455128e+00 -1.54757649e-01 -4.54570293e-01 -1.27831921e-01
-8.52340877e-01 3.79765391e-01 -2.18945757e-01 -8.77598524e-01
-3.24124336e-01 -9.31002021e-01 5.55103004e-01 4.47735637e-01
-5.82489491e-01 -1.34958589e+00 2.14730918e-01 4.27204698e-01
8.64552915e-01 5.96115887e-01 6.28721535e-01 -2.83454061e-01
-8.29545483e-02 -2.32470743e-02 2.82095581e-01 2.49075413e-01
7.70849586e-01 4.63616282e-01 -6.11707151e-01 -2.80211359e-01
3.69221568e-01 -9.54571739e-02 4.32247996e-01 3.77251804e-01
1.00659049e+00 -3.22210759e-01 -5.03768265e-01 -6.61154091e-02
1.35557652e+00 3.86124700e-01 2.64872700e-01 1.06390977e+00
-9.52867270e-02 1.25653958e+00 8.87758851e-01 4.77922112e-01
5.20592690e-01 3.60040843e-01 5.83742559e-01 -1.31490931e-01
2.78487075e-02 3.35122883e-01 2.51757741e-01 1.24789870e+00
-5.15490770e-01 -7.17972592e-02 -1.03467000e+00 5.25836170e-01
-2.03019595e+00 -7.74297297e-01 1.56662371e-02 1.82101321e+00
4.00664479e-01 -9.63759329e-03 5.10789037e-01 2.90150136e-01
1.09567022e+00 -8.06252360e-02 -5.82236111e-01 -5.93626320e-01
-4.73550171e-01 3.49651650e-02 8.34066197e-02 2.39376470e-01
-1.08198678e+00 1.05194724e+00 8.14891434e+00 2.49164656e-01
-9.91763413e-01 -6.66828174e-03 5.27350903e-01 3.81748080e-01
-8.04221928e-02 8.27450305e-02 -2.99034148e-01 5.52515805e-01
5.73854148e-01 -9.61827755e-01 4.49633688e-01 7.22975075e-01
4.48572576e-01 -2.60015965e-01 -7.59087265e-01 6.65142000e-01
-8.93643498e-02 -1.33662426e+00 -2.29023218e-01 1.42379716e-01
8.70196402e-01 -4.09899682e-01 -8.10222924e-02 -2.45404951e-02
1.30166635e-01 -9.64447320e-01 6.66463494e-01 2.74438292e-01
3.44599754e-01 -5.16766727e-01 8.91920090e-01 1.92539632e-01
-9.00922596e-01 -4.72940147e-01 -2.75507569e-01 -8.55651498e-01
2.41577372e-01 7.61239886e-01 -2.09621936e-01 7.02236354e-01
9.40211415e-01 8.47404540e-01 -3.98432136e-01 9.17367578e-01
-6.07525557e-02 2.97705084e-01 -1.38602927e-01 -5.41593015e-01
5.78874908e-02 -3.93319041e-01 8.56524944e-01 6.88598156e-01
2.17972949e-01 7.37673342e-02 -3.03499047e-02 7.91098535e-01
6.66854858e-01 1.73742250e-02 -7.48898089e-01 -6.28542960e-01
5.61603010e-01 1.12266600e+00 -1.12608576e+00 -4.09466594e-01
-2.87074029e-01 6.14936769e-01 1.54790625e-01 5.33482507e-02
-5.40168822e-01 -5.41568935e-01 8.97207975e-01 4.10878211e-01
-1.52124211e-01 -6.02145255e-01 -8.45608294e-01 -1.05239010e+00
-2.55263001e-01 -7.46891260e-01 4.37874906e-02 -4.55984801e-01
-1.46946287e+00 4.69359815e-01 3.52707393e-02 -1.16297054e+00
7.05880150e-02 -3.40476245e-01 -5.91869593e-01 1.64627358e-01
-1.35048509e+00 -8.85784566e-01 -5.54466605e-01 -2.02314675e-01
3.62610966e-01 -2.63748825e-01 8.62612486e-01 -3.09899479e-01
-9.20872033e-01 -1.65514842e-01 4.04872328e-01 -6.93409085e-01
7.61871994e-01 -1.29587924e+00 2.67457664e-01 2.15378240e-01
-5.73333144e-01 9.55543399e-01 1.15350974e+00 -6.09896243e-01
-1.50849116e+00 -6.40953004e-01 1.08079219e+00 -4.45209593e-01
1.04572999e+00 -2.14308843e-01 -5.88498831e-01 2.89456338e-01
6.73033535e-01 -4.69248861e-01 8.67863178e-01 -1.83853328e-01
5.81251085e-01 -6.58897534e-02 -1.26322281e+00 6.90772593e-01
1.28890646e+00 1.22365162e-01 -7.91273952e-01 5.96717417e-01
6.71602428e-01 1.18686430e-01 -1.11377835e+00 3.20298284e-01
7.88723171e-01 -1.32549512e+00 8.11937869e-01 -3.03098440e-01
4.91666585e-01 -3.44385177e-01 2.93510616e-01 -1.19028735e+00
-8.91069353e-01 -1.01900339e+00 -3.14672649e-01 1.07982957e+00
-4.68354851e-01 -1.30734837e+00 6.02841198e-01 2.05429867e-01
-3.14148068e-01 -9.54775274e-01 -7.79067755e-01 -1.17970192e+00
1.00903541e-01 4.47177202e-01 7.56832182e-01 1.08520269e+00
7.10111678e-01 2.89718807e-01 1.64399177e-01 -3.11394215e-01
7.14455187e-01 4.09828514e-01 8.20209682e-01 -1.38837302e+00
5.27925864e-02 -4.76317883e-01 -2.33386427e-01 -4.57322240e-01
-2.29191110e-02 -4.83704209e-01 -2.86238462e-01 -1.78128314e+00
1.02538377e-01 -2.78731585e-01 -3.56025308e-01 2.10288003e-01
1.40306070e-01 1.46056205e-01 5.13602644e-02 7.04120696e-01
-6.51733875e-01 5.78094363e-01 1.60589278e+00 2.73045897e-01
-3.70039135e-01 -9.57976505e-02 -1.17887330e+00 6.20617390e-01
1.15454221e+00 -2.07642153e-01 -9.29672197e-02 -5.62119521e-02
3.49325985e-01 -2.26799145e-01 -5.19722141e-02 -1.15764809e+00
1.68763116e-01 -6.88853025e-01 -1.98810697e-01 -2.91412234e-01
-1.48833618e-01 -9.47383642e-01 2.58340240e-01 9.00687635e-01
3.20790373e-02 4.17397916e-01 1.55781629e-02 3.89080077e-01
-4.33807671e-01 -4.47683722e-01 6.50693357e-01 -4.48699147e-01
-1.87125579e-01 -3.69963765e-01 -1.04256904e+00 -4.36113656e-01
1.72637653e+00 -8.06261897e-01 -3.58877033e-01 -2.57704526e-01
-8.13976705e-01 4.72680360e-01 1.04170310e+00 2.71058351e-01
3.54320526e-01 -1.15141106e+00 -1.86512575e-01 1.22604787e-01
1.44384161e-01 -2.40787059e-01 1.54027700e-01 7.15270519e-01
-8.37445796e-01 6.89108491e-01 -7.09221244e-01 -4.45646912e-01
-1.04969823e+00 5.28914690e-01 8.63532275e-02 -1.90806001e-01
-2.57866770e-01 1.66287810e-01 2.30620757e-01 -7.95298666e-02
-2.28462741e-01 7.60954618e-02 -4.83888179e-01 -1.14701949e-01
2.58873552e-01 1.01950991e+00 -9.01183933e-02 -3.93079787e-01
-4.49168324e-01 7.44273782e-01 2.26675063e-01 4.14955057e-02
1.67966306e+00 -5.93942344e-01 -9.93879080e-01 7.92192042e-01
2.39251524e-01 -1.94143772e-01 -8.24834466e-01 2.47205198e-01
-3.04813050e-02 -3.55996281e-01 -3.11267346e-01 -4.52343732e-01
-4.70443040e-01 5.23427010e-01 1.45169377e-01 1.22150791e+00
1.00760019e+00 2.28803158e-01 4.36671942e-01 2.61375308e-01
5.35956025e-01 -1.50844920e+00 8.52869987e-01 6.74362421e-01
1.42784417e+00 -9.14394557e-01 8.77687261e-02 -8.74046445e-01
-4.12759840e-01 1.27892792e+00 7.48587132e-01 -3.50224614e-01
7.05405295e-01 3.07005256e-01 -2.50559509e-01 -2.68373102e-01
-9.20446694e-01 2.08593402e-02 -3.39445919e-01 9.12213862e-01
7.76274562e-01 3.71574461e-02 -9.29505289e-01 3.82666022e-01
-4.24796611e-01 7.43456651e-03 9.68006849e-01 1.50362051e+00
-7.85589814e-01 -1.30728376e+00 -7.37183690e-01 2.87706256e-01
-3.04296881e-01 4.01908785e-01 -6.92597389e-01 5.27823925e-01
1.82192907e-01 1.12036157e+00 1.70449197e-01 -1.72041491e-01
1.11231878e-01 -2.56084532e-01 4.83816832e-01 -7.94753313e-01
-8.85390997e-01 7.87946861e-03 1.35430232e-01 -2.08909705e-01
-7.47970939e-01 -8.28989506e-01 -8.87307346e-01 -5.89403093e-01
-2.74587065e-01 4.54613715e-01 5.24350107e-01 6.39080703e-01
5.13617337e-01 5.31961560e-01 6.18409693e-01 -8.81032467e-01
-3.05872977e-01 -8.60475600e-01 -5.65283239e-01 -7.57834241e-02
1.95639357e-01 -1.05256736e+00 -4.89682555e-01 2.65857615e-02] | [5.627943992614746, 4.1948347091674805] |
9d350385-f313-47fa-bcff-8b996becda02 | long-tail-learning-with-attributes | 2004.02235 | null | https://arxiv.org/abs/2004.02235v4 | https://arxiv.org/pdf/2004.02235v4.pdf | From Generalized zero-shot learning to long-tail with class descriptors | Real-world data is predominantly unbalanced and long-tailed, but deep models struggle to recognize rare classes in the presence of frequent classes. Often, classes can be accompanied by side information like textual descriptions, but it is not fully clear how to use them for learning with unbalanced long-tail data. Such descriptions have been mostly used in (Generalized) Zero-shot learning (ZSL), suggesting that ZSL with class descriptions may also be useful for long-tail distributions. We describe DRAGON, a late-fusion architecture for long-tail learning with class descriptors. It learns to (1) correct the bias towards head classes on a sample-by-sample basis; and (2) fuse information from class-descriptions to improve the tail-class accuracy. We also introduce new benchmarks CUB-LT, SUN-LT, AWA-LT for long-tail learning with class-descriptions, building on existing learning-with-attributes datasets and a version of Imagenet-LT with class descriptors. DRAGON outperforms state-of-the-art models on the new benchmark. It is also a new SoTA on existing benchmarks for GFSL with class descriptors (GFSL-d) and standard (vision-only) long-tailed learning ImageNet-LT, CIFAR-10, 100, and Places365. | ['Dvir Samuel', 'Gal Chechik', 'Yuval Atzmon'] | 2020-04-05 | null | null | null | null | ['generalized-few-shot-learning', 'long-tail-learning-with-class-descriptors'] | ['methodology', 'methodology'] | [-6.41478822e-02 -2.25480214e-01 -6.48421109e-01 -6.98842049e-01
-1.09871221e+00 -3.39706540e-01 9.42072392e-01 2.51282096e-01
-4.73284960e-01 7.27082253e-01 4.30078506e-01 -5.30148670e-02
-1.68860152e-01 -7.85229921e-01 -6.74843371e-01 -7.50034690e-01
3.77471074e-02 9.05346334e-01 5.56524158e-01 -1.92357346e-01
-1.74658999e-01 1.32848129e-01 -2.02279878e+00 9.31248188e-01
2.62065977e-01 1.45332229e+00 2.66953819e-02 4.09142733e-01
-4.03336644e-01 1.23443258e+00 -5.02545595e-01 -4.56312954e-01
5.44624664e-02 -1.84390560e-01 -5.45467615e-01 -9.30968970e-02
1.19064569e+00 -2.18916878e-01 -2.75631160e-01 8.52468550e-01
8.62928927e-01 2.19276711e-01 1.27418470e+00 -1.45278394e+00
-5.93416929e-01 7.43978858e-01 -6.46917343e-01 2.09824651e-01
1.03472829e-01 2.90771246e-01 1.34495854e+00 -1.38797176e+00
7.57869363e-01 1.51959658e+00 9.67823863e-01 5.65168977e-01
-1.24721956e+00 -1.02023685e+00 2.31456026e-01 6.61774218e-01
-1.47966135e+00 -3.60657871e-01 2.86002070e-01 -6.34285629e-01
1.32245326e+00 3.13861668e-02 4.24796879e-01 1.65079451e+00
2.84354895e-01 1.42334497e+00 8.94116104e-01 -2.31824890e-02
2.09183872e-01 -3.67221907e-02 6.22890890e-01 3.51059645e-01
2.06785753e-01 3.76318485e-01 -6.61107957e-01 -3.15258652e-02
-1.45917594e-01 1.59128368e-01 5.27582504e-02 -5.66537559e-01
-1.12384725e+00 1.04234922e+00 6.03361785e-01 4.33927737e-02
-8.67447183e-02 1.08529754e-01 8.73573303e-01 3.86677176e-01
8.38140070e-01 1.22012869e-01 -7.20775306e-01 -9.98629332e-02
-1.10326827e+00 5.54811656e-01 8.67234230e-01 1.18379378e+00
9.70778644e-01 2.22337708e-01 -7.91501999e-01 1.06253588e+00
-1.07351921e-01 5.08916140e-01 6.56959534e-01 -5.95101655e-01
1.49103880e-01 1.90048501e-01 -4.00663912e-01 -3.17167103e-01
-4.23873514e-01 -8.99638355e-01 -9.19772863e-01 1.64139017e-01
9.44070965e-02 2.08963290e-01 -1.41759479e+00 1.55934608e+00
-2.07413092e-01 2.92218357e-01 1.12608574e-01 5.13874829e-01
1.58182812e+00 6.03485107e-01 3.21834415e-01 2.51186222e-01
1.24643505e+00 -1.18036246e+00 -5.65513670e-01 -4.21349883e-01
8.63942564e-01 -5.96132457e-01 1.14662206e+00 4.75643933e-01
-6.02517724e-01 -6.44560814e-01 -7.60401070e-01 -2.36451894e-01
-8.72177601e-01 6.19624369e-02 5.63407242e-01 3.61280829e-01
-8.66734028e-01 5.28217852e-01 -3.63781333e-01 -3.57957125e-01
9.58012342e-01 1.05196260e-01 -4.23439533e-01 -3.98978174e-01
-1.27919221e+00 9.21504676e-01 6.89088404e-01 -7.09606230e-01
-1.46447849e+00 -1.11750782e+00 -1.18511868e+00 1.74418285e-01
3.11919779e-01 -5.61553895e-01 1.42110944e+00 -8.93035352e-01
-8.00666869e-01 1.20355725e+00 1.95943058e-01 -6.70778990e-01
2.96003163e-01 -1.49258614e-01 -3.23316127e-01 -2.54723996e-01
4.85276222e-01 1.17347229e+00 8.18588495e-01 -1.15328610e+00
-8.30757737e-01 -3.00550044e-01 -1.70012161e-01 -8.48920196e-02
-2.92619854e-01 -3.31002623e-01 -1.07089607e-02 -7.35362589e-01
-5.19036949e-01 -7.45435238e-01 1.51722223e-01 1.28870755e-01
-5.00654101e-01 -7.54631996e-01 9.95358527e-01 -1.89324096e-02
1.02899170e+00 -2.27667689e+00 -3.67318451e-01 -2.87034959e-01
2.84786046e-01 5.94123960e-01 -4.82700258e-01 4.28654939e-01
-4.17769045e-01 -2.97323227e-01 -7.91675001e-02 -6.48387194e-01
1.68792188e-01 4.05727476e-01 -4.68614131e-01 4.15575117e-01
5.18380888e-02 9.25742269e-01 -9.92407203e-01 -5.12013078e-01
4.08954769e-01 3.02169681e-01 -4.86742646e-01 9.04301256e-02
-4.31788623e-01 -1.84763625e-01 -7.60161653e-02 9.56220388e-01
6.27286613e-01 -1.61009446e-01 -6.79018617e-01 -5.36271393e-01
1.06116563e-01 -5.46183847e-02 -8.05742204e-01 1.62266076e+00
-3.73249680e-01 6.90528333e-01 -4.73483175e-01 -8.96368742e-01
7.13955998e-01 1.18580371e-01 2.39265531e-01 -6.55253589e-01
1.83291614e-01 3.71921748e-01 -7.18853623e-02 -3.12766463e-01
2.59110481e-01 -3.00488949e-01 -1.21430323e-01 1.90033410e-02
9.12018359e-01 -2.31893390e-01 5.80597878e-01 4.13528949e-01
1.00209057e+00 -3.35819507e-03 4.52184856e-01 -2.52638161e-01
1.43126994e-01 -1.42822400e-01 4.43551272e-01 1.05158126e+00
-1.94264278e-01 9.14595604e-01 5.24629354e-01 -7.73752749e-01
-1.01460350e+00 -1.23149240e+00 -5.74944496e-01 1.63834822e+00
-1.85818821e-01 -7.75806665e-01 -2.31409952e-01 -9.58896875e-01
3.71038347e-01 1.25021160e+00 -8.59868228e-01 -4.69803303e-01
3.00810933e-01 -7.12593138e-01 6.98103964e-01 7.40269184e-01
3.96483362e-01 -1.01859248e+00 -3.53706867e-01 1.83912426e-01
4.55537811e-02 -1.12595403e+00 -2.88535744e-01 9.86476660e-01
-3.62474710e-01 -1.12250280e+00 -9.87070024e-01 -6.81052446e-01
2.22187832e-01 1.55614421e-01 1.75816357e+00 -3.52320284e-01
-4.95691359e-01 3.34562749e-01 -4.89904433e-01 -7.97493994e-01
1.17542157e-02 7.91999698e-02 -9.41501260e-02 -6.30616099e-02
7.13539183e-01 -2.06393346e-01 -4.12376225e-01 3.19294244e-01
-8.18655789e-01 -7.44467527e-02 5.23120582e-01 1.15579748e+00
6.23887241e-01 -2.89885461e-01 6.70343101e-01 -1.02248096e+00
3.16805393e-01 -7.82173932e-01 -3.28800417e-02 1.33269161e-01
-4.10061926e-01 -7.62104020e-02 4.82348114e-01 -5.77781141e-01
-7.70756125e-01 -1.92591194e-02 -4.77925688e-01 -8.49555790e-01
-2.92974561e-01 3.77052277e-01 -7.68356696e-02 1.17049202e-01
1.08678412e+00 2.61694610e-01 -1.99059799e-01 -5.38952827e-01
4.65383530e-01 7.20618665e-01 5.22477925e-01 -5.09729087e-01
2.69930363e-01 3.85060489e-01 -8.36622994e-03 -8.84390712e-01
-1.74802446e+00 -9.13465202e-01 -4.99704093e-01 1.02071419e-01
7.55872726e-01 -1.43167722e+00 -1.83070540e-01 7.35153258e-01
-8.01449716e-01 -5.38782775e-01 -7.86469758e-01 3.63614053e-01
-8.29537749e-01 -1.16974011e-01 -4.96747971e-01 -3.35554808e-01
-2.78409064e-01 -1.06848693e+00 1.54330993e+00 -6.17136620e-02
-1.95663273e-01 -9.46286678e-01 8.97910669e-02 -8.36973172e-03
5.83465815e-01 1.64233506e-01 8.29600394e-01 -1.26310289e+00
-8.30153227e-02 -2.83569068e-01 -4.29278135e-01 6.33474231e-01
-1.09903559e-01 -5.35608307e-02 -1.35350204e+00 -4.39396352e-01
-6.00452483e-01 -1.24923158e+00 1.78940606e+00 5.25381684e-01
1.52531075e+00 -5.70573285e-03 -3.91924441e-01 8.46565723e-01
1.18328059e+00 -2.05134377e-01 4.88578349e-01 2.02872530e-01
7.03632832e-01 2.55874395e-01 7.51161933e-01 6.32281303e-01
5.66770315e-01 6.13186657e-01 7.84463167e-01 -4.39784788e-02
-8.01192224e-01 -2.08553016e-01 3.04418564e-01 3.64552915e-01
3.80302548e-01 -3.05341721e-01 -9.97853041e-01 7.60643303e-01
-1.74230599e+00 -1.01287568e+00 -1.41330943e-01 1.84999728e+00
8.82124066e-01 1.83498755e-01 9.70147625e-02 7.87979439e-02
4.12281185e-01 4.57277834e-01 -5.94697714e-01 -4.15434241e-02
-4.98525947e-01 2.18220562e-01 5.02819240e-01 1.08354039e-01
-1.61899221e+00 1.03575504e+00 5.93114138e+00 1.65956247e+00
-1.04849720e+00 2.44689882e-01 7.73651600e-01 -2.68071949e-01
-1.47299767e-01 -2.50691295e-01 -1.39968204e+00 4.53727275e-01
9.76641059e-01 -3.83879282e-02 -3.43721926e-01 1.26398754e+00
-6.44620359e-01 -5.13519906e-02 -1.44318831e+00 1.35305595e+00
3.79404187e-01 -1.41042507e+00 3.55553985e-01 -1.42427921e-01
8.59377384e-01 7.17096865e-01 1.67791426e-01 1.20516455e+00
4.98367190e-01 -9.90553558e-01 9.03890073e-01 4.53565717e-01
1.13730788e+00 -8.82943273e-01 1.00503933e+00 4.84456033e-01
-1.12872684e+00 -2.61856943e-01 -7.51758993e-01 1.28468305e-01
-2.21313342e-01 8.08855891e-01 -6.11073256e-01 2.08291769e-01
8.77294123e-01 1.40368545e+00 -9.15443540e-01 1.28087449e+00
-5.69490232e-02 6.63586378e-01 -1.80911675e-01 8.94398689e-02
7.16783881e-01 4.83222336e-01 3.58731538e-01 1.54045868e+00
2.84128666e-01 -2.48633102e-01 6.02186918e-01 5.24964988e-01
-3.96806113e-02 -2.10749611e-01 -8.86460245e-01 8.96608829e-02
9.36371461e-02 1.14838231e+00 -4.51587290e-01 -7.42208958e-01
-6.63455069e-01 4.36662853e-01 2.84389883e-01 2.02703908e-01
-5.59979916e-01 -3.06317151e-01 6.92723453e-01 2.19285309e-01
4.30230170e-01 2.79070675e-01 -5.95978368e-03 -1.40268421e+00
-6.69021547e-01 -9.12717104e-01 8.85771453e-01 -8.18244636e-01
-1.83144581e+00 7.78285921e-01 2.76154131e-01 -1.48266399e+00
-4.00628805e-01 -8.08040679e-01 -2.29934648e-01 3.77801806e-01
-1.80379152e+00 -1.60937572e+00 -3.96965832e-01 1.05122447e+00
9.89440024e-01 -7.05783427e-01 1.00134397e+00 4.58816916e-01
-7.03361481e-02 6.90793991e-01 1.94318637e-01 8.33149776e-02
1.21799016e+00 -1.54088080e+00 2.82873333e-01 1.85663626e-01
3.87072861e-01 5.15472442e-02 6.24216378e-01 -5.66573143e-01
-8.02768409e-01 -1.47569180e+00 7.02614963e-01 -5.19741058e-01
8.11777413e-01 -3.71411592e-01 -8.12728643e-01 7.44291365e-01
1.72542289e-01 7.84768343e-01 7.13996708e-01 1.44838721e-01
-9.70019102e-01 -4.47487772e-01 -9.40819442e-01 3.47376168e-02
9.48555410e-01 -5.45765758e-01 -6.50141954e-01 6.85546875e-01
5.31873703e-01 -2.79257149e-01 -4.60198969e-01 6.49308920e-01
4.98559415e-01 -9.90887463e-01 1.10772896e+00 -6.09833300e-01
4.47386593e-01 4.09814948e-03 -6.04477286e-01 -1.79432213e+00
-4.58928227e-01 6.70275390e-02 -1.94940180e-01 1.02382874e+00
1.85421839e-01 -3.16149563e-01 7.22666025e-01 -1.60703361e-01
-4.76737469e-01 -7.88373053e-01 -8.33738625e-01 -1.17259026e+00
3.03226262e-01 -5.67416728e-01 3.27181965e-01 8.24008763e-01
-7.66637206e-01 4.49923307e-01 -4.44587976e-01 -5.83925366e-01
9.21707869e-01 2.25472450e-03 7.90870249e-01 -1.61423314e+00
1.33217843e-02 -4.74063098e-01 -7.93369412e-01 -6.53726101e-01
5.53675950e-01 -1.26128113e+00 1.88457947e-02 -1.59308064e+00
5.72438896e-01 -2.77620524e-01 -4.70662355e-01 8.80648673e-01
3.27800289e-02 4.84375119e-01 2.65459716e-01 3.49145122e-02
-1.07421005e+00 9.26049829e-01 9.76919889e-01 -5.03833354e-01
5.49943268e-01 1.54382035e-01 -3.92199010e-01 8.16720009e-01
2.60180295e-01 -5.58024883e-01 -7.35422909e-01 -9.14891958e-02
1.64328471e-01 -4.16012526e-01 4.71325576e-01 -1.31535554e+00
2.27459431e-01 1.55447051e-01 5.24533331e-01 -1.25896132e+00
4.72699374e-01 -8.03230047e-01 -2.65439004e-01 2.78824657e-01
-5.68022192e-01 -3.81899983e-01 4.21215966e-02 5.64656258e-01
-5.63734710e-01 -1.94572598e-01 1.08903229e+00 -3.37697357e-01
-1.23593771e+00 7.43290305e-01 -5.91860302e-02 5.70373893e-01
1.05003977e+00 -7.13353679e-02 -6.60393775e-01 -4.96654481e-01
-8.59066665e-01 3.35813701e-01 1.43496364e-01 5.86147904e-01
6.61206007e-01 -1.48319852e+00 -1.01536620e+00 3.17554563e-01
9.57960129e-01 4.96741273e-02 3.16606671e-01 7.51394033e-01
-1.90085739e-01 3.41527402e-01 -3.41112733e-01 -9.49850559e-01
-1.18384624e+00 6.76900923e-01 2.89448142e-01 -2.26556987e-01
-6.08056247e-01 1.23996949e+00 7.25161672e-01 -5.92813671e-01
5.27640581e-01 -2.01609597e-01 -2.42056563e-01 7.66767919e-01
8.31526518e-01 1.80643141e-01 4.31352913e-01 -6.87792420e-01
-5.05353034e-01 4.95302826e-01 -3.99046928e-01 3.69073004e-01
1.56541967e+00 1.85928822e-01 1.75543234e-01 8.81877780e-01
1.52413762e+00 -5.99097669e-01 -1.34521711e+00 -6.06685460e-01
2.89241485e-02 -2.74106979e-01 2.24752828e-01 -1.00705409e+00
-1.15737665e+00 1.30893290e+00 7.47166634e-01 -1.71371818e-01
7.98834860e-01 2.52862126e-01 6.56875253e-01 5.10192394e-01
4.68330979e-01 -9.74642277e-01 3.53760153e-01 1.09130764e+00
9.25029516e-01 -1.53140473e+00 -1.27067044e-02 6.70823306e-02
-8.33150148e-01 1.14618254e+00 6.81688905e-01 -1.35515898e-01
1.04932976e+00 4.30523634e-01 -9.67227668e-02 -2.02487245e-01
-1.37001717e+00 -7.26093233e-01 5.76663017e-01 8.37657511e-01
3.50737453e-01 1.55825958e-01 2.38585413e-01 5.89796305e-01
-9.12632644e-02 2.28660963e-02 3.24643403e-01 6.33406222e-01
-5.69190562e-01 -8.31660390e-01 4.84301969e-02 1.01611805e+00
-4.01514977e-01 -3.72480720e-01 -1.59493819e-01 7.95250297e-01
5.04539132e-01 6.69589698e-01 1.92508265e-01 -2.92692393e-01
3.47156405e-01 1.35042250e-01 2.37061992e-01 -9.42878544e-01
-5.19280016e-01 -1.22422427e-02 2.94296205e-01 -7.31793344e-01
-1.81889534e-01 -4.35704827e-01 -7.33801186e-01 -1.07660674e-01
-1.97668616e-02 -2.00974882e-01 4.29266930e-01 7.46095300e-01
1.57418668e-01 6.65478051e-01 1.81347996e-01 -1.04505110e+00
-7.54641414e-01 -1.02899361e+00 -8.70887339e-01 5.51916420e-01
5.88872850e-01 -1.07703340e+00 -7.00689495e-01 -3.22606921e-01] | [9.872037887573242, 2.6637041568756104] |
fa1628b3-0c7e-4f83-b481-2b742710959e | compilable-neural-code-generation-with | 2203.05132 | null | https://arxiv.org/abs/2203.05132v1 | https://arxiv.org/pdf/2203.05132v1.pdf | Compilable Neural Code Generation with Compiler Feedback | Automatically generating compilable programs with (or without) natural language descriptions has always been a touchstone problem for computational linguistics and automated software engineering. Existing deep-learning approaches model code generation as text generation, either constrained by grammar structures in decoder, or driven by pre-trained language models on large-scale code corpus (e.g., CodeGPT, PLBART, and CodeT5). However, few of them account for compilability of the generated programs. To improve compilability of the generated programs, this paper proposes COMPCODER, a three-stage pipeline utilizing compiler feedback for compilable code generation, including language model fine-tuning, compilability reinforcement, and compilability discrimination. Comprehensive experiments on two code generation tasks demonstrate the effectiveness of our proposed approach, improving the success rate of compilation from 44.18 to 89.18 in code completion on average and from 70.3 to 96.2 in text-to-code generation, respectively, when comparing with the state-of-the-art CodeGPT. | ['Qun Liu', 'Xin Jiang', 'Hao Wu', 'Jin Liu', 'Pingyi Zhou', 'Yitong Li', 'Fei Mi', 'Yao Wan', 'Yasheng Wang', 'Xin Wang'] | 2022-03-10 | null | https://aclanthology.org/2022.findings-acl.2 | https://aclanthology.org/2022.findings-acl.2.pdf | findings-acl-2022-5 | ['text-to-code-generation'] | ['computer-code'] | [ 4.70910855e-02 2.89087027e-01 -2.18706846e-01 -2.69969285e-01
-1.10676873e+00 -6.14024639e-01 3.06870013e-01 1.55088007e-01
1.90557450e-01 5.81627131e-01 1.10612549e-01 -6.77038491e-01
5.09352863e-01 -1.01347816e+00 -9.63779211e-01 3.37320752e-02
1.79959044e-01 4.36861217e-01 -2.13015899e-01 -2.89841145e-01
4.17759627e-01 -3.40757996e-01 -1.30299032e+00 5.26818812e-01
1.63310087e+00 4.06040847e-01 5.62700748e-01 9.89329576e-01
-5.48487425e-01 8.80775094e-01 -5.56091368e-01 -6.16018057e-01
-1.00159571e-01 -5.70312738e-01 -7.60792136e-01 -2.10242331e-01
-6.01375066e-02 -2.46352732e-01 -2.19799932e-02 1.20079851e+00
2.09917903e-01 -4.63214666e-01 6.04397774e-01 -1.05338621e+00
-1.18970203e+00 1.31495583e+00 -4.73333180e-01 -2.90601909e-01
6.14783943e-01 2.71964520e-01 1.12164497e+00 -9.14950371e-01
4.25571680e-01 1.26245439e+00 3.90786976e-01 7.79926062e-01
-1.28209639e+00 -6.69970393e-01 -1.91427127e-01 -4.40645903e-01
-1.25742805e+00 -4.40794528e-01 4.88511533e-01 -8.80158186e-01
1.41899312e+00 1.97066758e-02 4.60830837e-01 8.59319925e-01
5.17308056e-01 7.52756536e-01 7.12335825e-01 -7.44098306e-01
3.59828747e-03 5.21984026e-02 -3.65224332e-02 1.19897676e+00
3.05263102e-01 7.38602951e-02 -2.41117895e-01 -1.88317105e-01
4.93158013e-01 -4.69924986e-01 -1.32771209e-01 -1.07508358e-02
-1.31763458e+00 1.01997769e+00 1.21918209e-01 -1.01751551e-01
-1.09552585e-01 3.41092408e-01 6.27848148e-01 2.32013598e-01
3.37870270e-01 7.40258992e-01 -4.34506387e-01 -5.35523713e-01
-8.95178020e-01 5.94483256e-01 8.23799014e-01 1.75891042e+00
8.84225607e-01 4.36202973e-01 -3.36320192e-01 7.48459458e-01
4.89504755e-01 7.49216974e-01 8.88535619e-01 -2.89996475e-01
1.06650913e+00 1.02488947e+00 -1.43939942e-01 -8.05992186e-01
-1.89412966e-01 -2.16815799e-01 -5.09967804e-01 -8.69986266e-02
5.25786076e-03 -5.22464693e-01 -6.77007973e-01 1.58089542e+00
-1.96970046e-01 -5.14698982e-01 4.07863975e-01 4.75953162e-01
9.40897286e-01 9.49390471e-01 -7.88927972e-02 9.71454754e-02
1.05733323e+00 -1.30414259e+00 -4.05636579e-01 -4.63005424e-01
1.11731458e+00 -8.69709074e-01 1.47421730e+00 9.11332071e-02
-1.08379173e+00 -7.19521403e-01 -1.02817833e+00 -8.78350958e-02
-7.86040202e-02 7.51567781e-01 7.15527058e-01 6.63914680e-01
-1.07381260e+00 2.96124578e-01 -7.95391858e-01 -8.40777308e-02
3.07532012e-01 7.50390068e-02 3.06829456e-02 1.22782690e-02
-8.09573352e-01 5.35324693e-01 7.49679148e-01 -3.89468580e-01
-1.26535010e+00 -5.04348397e-01 -1.19665086e+00 1.91700563e-01
1.02963850e-01 -6.24756694e-01 1.66504276e+00 -8.88286650e-01
-1.55196476e+00 7.56692231e-01 -1.43646672e-01 -4.95807648e-01
3.42855304e-01 -3.99064541e-01 -2.92846441e-01 -3.88066918e-01
3.64874572e-01 8.01303208e-01 5.99123657e-01 -9.19576883e-01
-6.20128572e-01 2.69103020e-01 2.26952821e-01 -2.41199713e-02
-2.68201828e-01 1.17415935e-01 -2.55525887e-01 -6.78283989e-01
-3.24106187e-01 -1.05376124e+00 -1.54484227e-01 -5.00958145e-01
-6.54401183e-01 -3.97090733e-01 3.19539368e-01 -1.05348337e+00
1.38612211e+00 -2.10974884e+00 1.34622693e-01 -1.52027026e-01
7.70800188e-02 7.59644657e-02 -4.81137902e-01 5.37428677e-01
3.61744501e-02 4.60328907e-01 -3.62226874e-01 5.50866164e-02
3.15261632e-01 -1.98307797e-01 -4.20290679e-01 -1.03957811e-03
6.89418197e-01 1.09848189e+00 -1.03347278e+00 -4.13058549e-01
-1.86282754e-01 -1.63435191e-02 -1.00343597e+00 5.70796490e-01
-7.42058873e-01 2.31348485e-01 -5.65395236e-01 6.27103508e-01
2.67699003e-01 -1.35402948e-01 2.73459218e-02 4.69344378e-01
-2.36173987e-01 7.35602856e-01 -5.95024467e-01 1.93752193e+00
-1.04391408e+00 6.38416767e-01 -5.28721631e-01 -3.74406070e-01
1.36069787e+00 2.84979075e-01 -3.57535094e-01 -5.79876006e-01
3.78040858e-02 5.52797318e-01 4.66770053e-01 -8.10649216e-01
7.37407684e-01 2.59101123e-01 -5.58627069e-01 8.11043382e-01
6.38545826e-02 -4.81205970e-01 7.80380905e-01 2.90395707e-01
9.07991111e-01 3.86924028e-01 3.91697854e-01 -4.81030554e-01
5.95794439e-01 1.74992472e-01 2.69277215e-01 6.43910706e-01
4.57280576e-01 6.72475278e-01 8.36457014e-01 -2.40623012e-01
-1.25777340e+00 -5.43372810e-01 2.97376335e-01 1.09355783e+00
-2.81382591e-01 -7.84384608e-01 -1.35337830e+00 -7.04739928e-01
-3.65863293e-01 9.86941516e-01 -3.23151797e-01 -5.29325902e-01
-7.05323696e-01 -6.21283770e-01 7.18702734e-01 5.47086120e-01
3.69512767e-01 -1.36174381e+00 -5.65143645e-01 4.15530920e-01
-3.41860533e-01 -8.92965615e-01 -7.19711602e-01 -9.66169536e-02
-6.41584098e-01 -9.05857146e-01 -3.56126159e-01 -1.12696886e+00
9.57271099e-01 -1.91848576e-01 1.42378902e+00 4.71510828e-01
-1.89752042e-01 -3.64929259e-01 -5.82387984e-01 -2.89511412e-01
-1.42371500e+00 6.08799577e-01 -3.35748672e-01 -4.61377025e-01
3.59345414e-02 -1.23817384e-01 -5.57267927e-02 -3.80906127e-02
-7.44977713e-01 6.59176528e-01 8.88235271e-01 9.79500651e-01
1.89473867e-01 -1.50849655e-01 6.79080725e-01 -1.08055902e+00
9.95471001e-01 -4.23326135e-01 -9.16458011e-01 3.71744841e-01
-5.70522308e-01 4.30198967e-01 1.01102912e+00 -2.25099206e-01
-1.14478290e+00 3.67452912e-02 -3.57556194e-01 1.89162448e-01
7.70881400e-02 9.23047781e-01 -2.25175455e-01 7.48536289e-02
1.00682437e+00 5.70005655e-01 -2.67290533e-01 -6.32181689e-02
4.44363832e-01 8.35008144e-01 4.26883727e-01 -1.00162864e+00
8.78989518e-01 -3.82360578e-01 -6.28251433e-01 -1.39637873e-01
-4.11301732e-01 2.50429958e-01 -3.90738517e-01 1.40178323e-01
7.82532215e-01 -1.10018933e+00 -1.92505926e-01 4.17469382e-01
-1.41638672e+00 -5.53082466e-01 7.45624676e-02 1.71968430e-01
-6.12986803e-01 9.83499289e-02 -6.93294287e-01 -5.99933982e-01
-6.59054399e-01 -1.74101675e+00 1.20178270e+00 1.50029689e-01
-4.70463991e-01 -7.64428556e-01 9.89908129e-02 1.78563744e-01
5.50966680e-01 1.86003312e-01 1.48885620e+00 -3.82392645e-01
-7.18436480e-01 -1.97236404e-01 -3.27355802e-01 2.36778080e-01
1.90430358e-01 2.50572741e-01 -6.58851206e-01 -2.83520162e-01
-4.78931487e-01 -5.06759167e-01 3.62744391e-01 -5.03745787e-02
1.02385855e+00 -5.43223321e-01 -2.20980272e-01 5.85670888e-01
1.42134690e+00 3.94708097e-01 4.56051379e-01 1.40484348e-01
7.82619834e-01 2.72014916e-01 5.43085396e-01 5.87734640e-01
6.77067459e-01 3.52184683e-01 5.06789863e-01 2.00938925e-01
-9.67705324e-02 -7.35041261e-01 7.39602149e-01 1.14976406e+00
3.69190276e-01 -1.98847666e-01 -1.38586688e+00 6.91643059e-01
-1.51048648e+00 -3.46575022e-01 -3.68756413e-01 1.95314431e+00
1.26362765e+00 2.74579465e-01 -1.12087898e-01 -1.01955250e-01
7.29763806e-01 -2.01116920e-01 -4.00836527e-01 -7.55733848e-01
3.44068199e-01 4.05643135e-02 1.19039983e-01 4.17085409e-01
-8.76849473e-01 1.24164712e+00 5.66339779e+00 7.26236045e-01
-1.03373110e+00 -2.81334501e-02 6.54731691e-01 3.22349787e-01
-5.29676795e-01 2.45750874e-01 -8.37316394e-01 5.51446676e-01
1.05729127e+00 -6.84029937e-01 6.89220667e-01 1.32016802e+00
7.91167766e-02 1.66358232e-01 -1.32535994e+00 7.51123548e-01
2.21400604e-01 -1.35745442e+00 2.39326328e-01 -2.10258037e-01
1.18235290e+00 -1.66940436e-01 -2.06456125e-01 8.73214602e-01
5.82848191e-01 -1.05665791e+00 1.21546793e+00 1.57472253e-01
1.06310844e+00 -9.05326545e-01 6.72367811e-01 4.87172842e-01
-1.22607255e+00 -4.77743597e-04 -4.33241069e-01 -1.59089252e-01
-5.04714213e-02 6.79927170e-01 -1.08853805e+00 3.86467785e-01
2.85171360e-01 5.69561303e-01 -8.41714740e-01 7.90165246e-01
-5.55510223e-01 6.58004165e-01 3.21105421e-01 -4.29903835e-01
1.12912394e-01 1.51597857e-01 2.54632741e-01 1.40928149e+00
6.81816399e-01 -3.98627579e-01 3.86648208e-01 1.62399411e+00
-4.17998850e-01 2.65423834e-01 -4.68720168e-01 -6.64932251e-01
3.82495731e-01 1.32681024e+00 -4.17057037e-01 -4.88646954e-01
-4.25490648e-01 6.13665164e-01 5.26279449e-01 1.25505000e-01
-1.03716540e+00 -9.52483296e-01 2.21387103e-01 -9.70853567e-02
2.33970750e-02 -2.53318995e-01 -3.80865604e-01 -1.17289567e+00
2.62398332e-01 -1.25735462e+00 -2.54790455e-01 -8.13258648e-01
-7.78298140e-01 1.00398982e+00 -2.44763777e-01 -1.19514048e+00
-6.40249848e-01 -4.40360606e-01 -7.89041758e-01 1.07539988e+00
-1.33057892e+00 -1.15175712e+00 -3.07558089e-01 -7.39413947e-02
8.78749907e-01 -7.38159418e-01 7.31700063e-01 1.32785276e-01
-4.98760104e-01 8.73201609e-01 -5.17822765e-02 3.42866868e-01
2.49432385e-01 -1.31965554e+00 1.27161527e+00 1.14741361e+00
-5.63863218e-02 8.47765207e-01 3.76257420e-01 -7.71095157e-01
-1.52918804e+00 -1.62299681e+00 1.00578833e+00 -2.21240118e-01
6.78766727e-01 -8.43868315e-01 -7.22526073e-01 5.29076874e-01
3.07763815e-01 -4.37352449e-01 3.76391798e-01 -2.01884598e-01
-4.21981096e-01 2.76875556e-01 -8.35920930e-01 6.42656147e-01
7.59458005e-01 -4.34029341e-01 -2.35982969e-01 4.06746984e-01
1.09747863e+00 -9.37037766e-01 -7.14632988e-01 1.39534757e-01
3.22615415e-01 -6.73203290e-01 2.98714936e-01 -3.48876029e-01
1.33614445e+00 -3.18365425e-01 -1.35149583e-02 -1.49487567e+00
-1.56270042e-01 -7.54268885e-01 4.34393436e-02 1.52378011e+00
9.47382390e-01 -3.80814672e-01 4.37858760e-01 3.81959945e-01
-6.57311201e-01 -6.15862310e-01 -2.10135117e-01 -5.29178083e-01
2.96072066e-01 -2.21445918e-01 1.02738106e+00 8.28845203e-01
3.33237201e-01 2.90355593e-01 -1.25659853e-01 -1.82954997e-01
4.76397388e-02 3.89469087e-01 1.10012698e+00 -7.13873506e-01
-7.44573295e-01 -5.40142596e-01 -6.19453304e-02 -1.01115334e+00
4.85569000e-01 -1.38435876e+00 5.33891916e-01 -1.56672132e+00
4.79763865e-01 -4.52478111e-01 5.61640561e-01 6.19479179e-01
-4.55922842e-01 -2.72607088e-01 -2.43796487e-04 4.85227779e-02
-3.08484346e-01 5.92547774e-01 1.19537580e+00 -3.17046762e-01
-9.58095863e-02 -1.73904389e-01 -1.00291538e+00 4.29333270e-01
1.04957938e+00 -5.75440764e-01 -5.14033437e-01 -8.79269242e-01
6.11960351e-01 3.72863621e-01 -1.18567653e-01 -1.05603778e+00
-2.39606962e-01 -1.65619999e-01 -1.60442516e-01 -2.32157543e-01
-6.04433000e-01 -4.97097522e-02 -6.69285581e-02 6.09617710e-01
-6.65359616e-01 4.72164392e-01 4.12478238e-01 1.82781860e-01
-2.46950433e-01 -6.49897754e-01 7.03770816e-01 -2.65123218e-01
-5.56588054e-01 1.17496409e-01 -4.98867095e-01 4.34536189e-01
7.77043045e-01 2.84503192e-01 -5.61968684e-01 -5.47319166e-02
-6.09061904e-02 1.56754702e-01 4.42706823e-01 8.54053438e-01
6.10193908e-01 -1.33195913e+00 -1.07921159e+00 4.77569520e-01
5.78131676e-01 1.57401860e-01 -1.12291381e-01 3.67575854e-01
-1.01138842e+00 5.29004514e-01 -1.21212997e-01 -4.33689773e-01
-8.86076510e-01 2.40903929e-01 2.26794064e-01 -4.69866872e-01
-3.05696964e-01 8.17077339e-01 2.72564113e-01 -8.08016121e-01
-1.97492242e-01 -6.95178628e-01 1.10230662e-01 -6.42105579e-01
4.14115041e-01 -6.48948550e-02 2.09953904e-01 -3.65261108e-01
3.06412224e-02 2.75867313e-01 -3.39814514e-01 8.42471942e-02
9.79173541e-01 3.61847073e-01 -5.17052114e-01 1.46481320e-01
1.09357059e+00 -7.59339482e-02 -9.45747972e-01 9.56600457e-02
1.37192339e-01 -2.78452367e-01 -3.05827588e-01 -8.56920123e-01
-9.27491844e-01 9.81001079e-01 1.97420374e-01 6.71375394e-02
7.22213686e-01 -1.91872865e-01 8.44881117e-01 5.18745065e-01
5.50504029e-01 -6.79156780e-01 1.44679144e-01 7.59030044e-01
1.04763353e+00 -1.33979475e+00 -4.43917066e-01 -2.88311034e-01
-5.25746047e-01 1.25136614e+00 1.06311023e+00 -2.46647343e-01
4.09080740e-03 4.77414846e-01 -1.29856125e-01 1.81054860e-01
-1.01486862e+00 2.29503781e-01 2.19508618e-01 5.37205219e-01
1.12856197e+00 1.96216241e-01 -3.01561892e-01 7.06277072e-01
-7.70686448e-01 -3.83936092e-02 8.78059626e-01 9.14515793e-01
-4.03328806e-01 -1.21969175e+00 -9.82356146e-02 6.93337202e-01
-2.69750446e-01 -6.93664670e-01 -1.85542986e-01 5.52827358e-01
-8.05472061e-02 9.77777898e-01 -2.43512392e-01 -4.00475562e-01
1.38634071e-01 -2.77865399e-02 2.22070813e-01 -1.32135177e+00
-7.42278278e-01 -1.07103042e-01 2.63496161e-01 -1.35466769e-01
2.16265008e-01 -5.04973710e-01 -1.40993869e+00 -1.27337441e-01
-5.34844220e-01 2.65707046e-01 6.22386813e-01 7.57503569e-01
5.08872747e-01 8.95934105e-01 4.57573563e-01 -5.18645644e-01
-4.68501747e-01 -1.03361177e+00 -1.19955584e-01 9.58152413e-02
1.91573560e-01 -1.69049665e-01 9.72046331e-02 5.74617386e-01] | [7.7608256340026855, 7.8157219886779785] |
0bf0f3aa-f4b5-4f83-a99c-ea0b8209425d | beyond-the-hype-assessing-the-performance | 2306.15887 | null | https://arxiv.org/abs/2306.15887v1 | https://arxiv.org/pdf/2306.15887v1.pdf | Beyond the Hype: Assessing the Performance, Trustworthiness, and Clinical Suitability of GPT3.5 | The use of large language models (LLMs) in healthcare is gaining popularity, but their practicality and safety in clinical settings have not been thoroughly assessed. In high-stakes environments like medical settings, trust and safety are critical issues for LLMs. To address these concerns, we present an approach to evaluate the performance and trustworthiness of a GPT3.5 model for medical image protocol assignment. We compare it with a fine-tuned BERT model and a radiologist. In addition, we have a radiologist review the GPT3.5 output to evaluate its decision-making process. Our evaluation dataset consists of 4,700 physician entries across 11 imaging protocol classes spanning the entire head. Our findings suggest that the GPT3.5 performance falls behind BERT and a radiologist. However, GPT3.5 outperforms BERT in its ability to explain its decision, detect relevant word indicators, and model calibration. Furthermore, by analyzing the explanations of GPT3.5 for misclassifications, we reveal systematic errors that need to be resolved to enhance its safety and suitability for clinical use. | ['Mohammad R. K. Mofrad', 'Elizabeth Tong', 'Salmonn Talebi'] | 2023-06-28 | null | null | null | null | ['decision-making'] | ['reasoning'] | [ 2.86006369e-02 6.25010788e-01 -3.99033368e-01 -4.74205077e-01
-1.30911553e+00 -5.55728376e-01 1.21014705e-03 8.89253676e-01
-6.06687427e-01 4.12531674e-01 5.61075628e-01 -9.41586196e-01
-3.61314505e-01 -1.04710914e-01 -7.24441051e-01 -2.01247811e-01
1.86008990e-01 6.98857069e-01 -8.65953863e-02 4.71464604e-01
2.06839293e-01 2.76427537e-01 -6.95632160e-01 4.65992182e-01
8.12913060e-01 4.94097352e-01 1.67323098e-01 5.80083847e-01
3.59793484e-01 1.30688989e+00 -6.19629025e-01 -7.11455762e-01
-3.23623382e-02 -2.09481299e-01 -7.56072402e-01 -5.17495014e-02
1.76711783e-01 -5.45516789e-01 5.50791733e-02 7.18765914e-01
9.43731844e-01 -4.26289022e-01 5.91715991e-01 -9.75267887e-01
-6.02473080e-01 1.06724358e+00 -2.07675219e-01 3.88137817e-01
5.25021672e-01 5.41724861e-01 8.41658413e-01 -3.95172209e-01
7.31846392e-01 1.04466069e+00 8.37466121e-01 3.20099354e-01
-1.08065414e+00 -7.93634295e-01 4.35565189e-02 -3.18403058e-02
-1.43007815e+00 -3.59177440e-01 -4.18084636e-02 -6.68499351e-01
1.03486919e+00 1.76764727e-01 6.72650397e-01 1.10756695e+00
9.92209613e-01 3.94912660e-01 1.16003716e+00 -3.76715928e-01
3.59346688e-01 4.63816941e-01 3.40473562e-01 6.85873806e-01
7.04026580e-01 3.31505276e-02 -4.17688280e-01 -6.59896195e-01
6.31688952e-01 -1.77498549e-01 -3.73147786e-01 2.88483828e-01
-1.45878685e+00 8.38125229e-01 3.89175922e-01 1.02832437e-01
-5.40311337e-01 9.55232605e-02 2.44004548e-01 1.54687669e-02
1.01548292e-01 8.06890130e-01 -3.98588777e-01 -3.31674248e-01
-9.67504084e-01 1.95209309e-01 9.38460827e-01 8.94610167e-01
-2.22110778e-01 -3.78581434e-01 -5.11225939e-01 5.60549378e-01
5.14903843e-01 4.75223452e-01 6.17769122e-01 -1.03342175e+00
1.37537077e-01 3.67374361e-01 8.00711885e-02 -9.14906859e-01
-7.69982934e-01 -5.00705063e-01 -4.09407675e-01 -2.46095613e-01
1.00628376e-01 -5.26697524e-02 -7.64160275e-01 1.26487815e+00
-3.44251134e-02 -1.51595339e-01 -5.85514568e-02 7.90444016e-01
1.04766393e+00 1.39911607e-01 7.51155078e-01 -1.69636726e-01
1.66759932e+00 -7.78379679e-01 -9.16181982e-01 -5.29009819e-01
1.11639524e+00 -1.01821434e+00 1.08121550e+00 3.33797187e-01
-1.33112180e+00 -2.18908086e-01 -7.52858996e-01 1.29880816e-01
3.07706773e-01 6.57247528e-02 3.01580876e-01 6.73655391e-01
-8.56408954e-01 3.66641313e-01 -9.94462013e-01 -4.61401701e-01
5.32634676e-01 2.97004670e-01 -3.26940924e-01 -8.90209600e-02
-1.02636886e+00 1.38737655e+00 7.45460242e-02 -6.19857199e-02
-9.18119073e-01 -7.77691185e-01 -8.97161067e-01 -6.88481331e-02
1.47230208e-01 -9.54845607e-01 1.84622478e+00 -1.25349343e-01
-9.36031818e-01 8.70658636e-01 -1.24712951e-01 -5.26036084e-01
5.80953717e-01 1.65632814e-02 -4.92686033e-01 2.74151087e-01
5.02749324e-01 4.88684088e-01 1.88692689e-01 -1.01040578e+00
-4.82710898e-01 -1.35638133e-01 -4.00514491e-02 3.37034166e-02
8.36955458e-02 3.18224102e-01 -8.82403478e-02 -6.15368366e-01
6.67593330e-02 -1.13256502e+00 -7.36455917e-01 -2.16391072e-01
-3.18301857e-01 1.89187035e-01 -2.20996961e-01 -7.17425227e-01
1.59156871e+00 -2.07073259e+00 -5.69424748e-01 3.65190625e-01
6.80505216e-01 4.28406298e-02 1.70482993e-01 1.64153054e-01
1.27198873e-02 6.94233000e-01 2.56037600e-02 -1.54239252e-01
-2.02413768e-01 1.39093161e-01 2.95032024e-01 4.94387299e-01
1.08600274e-01 1.27498102e+00 -7.61133075e-01 -9.28504109e-01
1.45413712e-01 4.47833896e-01 -7.56146491e-01 3.45081389e-02
4.12283450e-01 3.56181234e-01 -4.74665344e-01 6.43406391e-01
2.13968962e-01 -9.36711133e-01 1.87889606e-01 3.28565203e-03
1.70647427e-01 6.07012272e-01 -6.56735420e-01 1.35453069e+00
-3.83355975e-01 3.61710995e-01 -1.17522687e-01 -2.04964906e-01
4.86487627e-01 5.71797550e-01 4.99255419e-01 -5.19654751e-01
2.71169037e-01 3.03022474e-01 4.75713104e-01 -8.87948751e-01
8.08183849e-02 -5.05617321e-01 2.50236485e-02 7.16647863e-01
-3.18511963e-01 -2.93514460e-01 -4.40895468e-01 3.84559780e-01
1.32476997e+00 -4.82768327e-01 6.60711765e-01 -3.30162615e-01
-3.88335809e-03 3.31039846e-01 5.28115511e-01 1.06371379e+00
-4.47239459e-01 7.52717197e-01 3.72500837e-01 -1.57578945e-01
-6.40079319e-01 -6.11072838e-01 -3.99272650e-01 3.93001407e-01
-3.90273392e-01 -7.19512999e-01 -8.87206972e-01 -5.83919823e-01
-9.20813829e-02 1.18171120e+00 -5.87650001e-01 -4.13065463e-01
-2.08946928e-01 -6.16486132e-01 6.26628160e-01 5.87901175e-01
-5.92296384e-02 -9.80194211e-01 -1.35095012e+00 4.49205846e-01
-4.93759304e-01 -1.30385637e+00 -6.14194036e-01 1.34425819e-01
-7.90699065e-01 -1.21843207e+00 -4.71079230e-01 -5.15482426e-01
7.46893048e-01 8.86713620e-03 1.26919162e+00 1.53759375e-01
-3.31131250e-01 7.63858438e-01 -3.28953803e-01 -8.46603930e-01
-1.01216185e+00 -8.99766907e-02 -1.81995317e-01 -8.82908344e-01
7.08669066e-01 1.87055692e-01 -7.17480004e-01 4.07451779e-01
-9.37165260e-01 -2.17118412e-02 9.57976758e-01 6.95171654e-01
4.30413991e-01 -2.25617528e-01 2.51894563e-01 -1.26712072e+00
9.76459086e-01 -5.31819284e-01 -1.80993617e-01 3.90309095e-01
-1.28944945e+00 -1.42156020e-01 -7.89274052e-02 -4.03937906e-01
-6.59520984e-01 -2.86518544e-01 -6.98731020e-02 -3.38057317e-02
-7.63595104e-02 9.97886002e-01 3.30588907e-01 -1.58941612e-01
8.57609808e-01 -6.02474391e-01 1.05690069e-01 -1.43668517e-01
-1.20373718e-01 8.63861620e-01 1.64573103e-01 -3.35612625e-01
3.15332919e-01 3.42601128e-02 -5.68127930e-01 -2.01542154e-01
-9.13530469e-01 -2.88238853e-01 -1.75932512e-01 -2.22314045e-01
7.31682956e-01 -1.03247523e+00 -7.20729291e-01 -3.41076463e-01
-1.18160129e+00 -2.49265954e-01 -3.00300449e-01 9.65208769e-01
-3.32782894e-01 2.33591899e-01 -8.24492991e-01 -5.61259210e-01
-3.90168160e-01 -1.87490356e+00 1.04665637e+00 -1.27293169e-01
-1.22837567e+00 -8.54383707e-01 -2.02542350e-01 8.09035599e-01
2.14343265e-01 7.65668377e-02 1.11379480e+00 -1.00673985e+00
-3.01230460e-01 -6.71248019e-01 3.90704125e-02 4.80803847e-02
2.32096970e-01 2.54553501e-02 -7.48430192e-01 -8.98534358e-02
4.77153063e-01 -1.70078322e-01 1.46481961e-01 1.01069498e+00
1.15583384e+00 -4.34133977e-01 -4.52096224e-01 2.19494626e-01
9.91673708e-01 4.71155196e-01 2.47245580e-01 5.60906947e-01
4.40923125e-01 7.64960349e-01 7.15835273e-01 4.03683752e-01
8.69996309e-01 2.29017273e-01 -4.73024063e-02 -2.33177394e-01
-4.68423404e-03 -4.29120064e-01 2.69722492e-01 9.06074464e-01
2.48447910e-01 -2.47152850e-01 -1.56170714e+00 3.80943388e-01
-1.61219501e+00 -2.85348296e-01 -2.06453145e-01 2.12088037e+00
6.23829186e-01 5.78288138e-01 -3.41175139e-01 -1.05435468e-01
3.52427065e-01 -2.61911064e-01 -4.39515620e-01 -5.06791115e-01
-6.03182130e-02 -8.41058716e-02 7.75026441e-01 3.94842058e-01
-3.67288560e-01 5.29251158e-01 7.51351595e+00 3.94764662e-01
-7.27320552e-01 3.50677550e-01 9.82372820e-01 -1.19005859e-01
-6.33571506e-01 -5.24203405e-02 -4.91180658e-01 2.87010729e-01
1.37550592e+00 -1.28322363e-01 -1.60282657e-01 5.44512928e-01
7.84098923e-01 -2.11704582e-01 -1.30454016e+00 8.46596301e-01
8.86922237e-03 -1.27025115e+00 -6.33426616e-03 1.20344155e-01
4.83553588e-01 -6.78885430e-02 1.76590547e-01 1.31537303e-01
4.40315276e-01 -1.46914768e+00 8.84913683e-01 2.18887493e-01
7.04931021e-01 -2.57446200e-01 1.23646533e+00 2.84040987e-01
-4.88973618e-01 -2.35898923e-02 -1.99427828e-01 1.16875865e-01
2.53266752e-01 4.70615983e-01 -1.50536919e+00 2.13640839e-01
6.64445758e-01 4.73747760e-01 -6.32377863e-01 1.00740898e+00
-2.87259012e-01 1.06073403e+00 3.51862572e-02 2.21269414e-01
3.05836916e-01 3.77943784e-01 1.02479652e-01 1.25449598e+00
9.03173238e-02 4.85194236e-01 2.84942845e-03 6.25829577e-01
5.96310943e-03 3.22319478e-01 -4.72607493e-01 -2.98566520e-01
4.26460415e-01 1.00040841e+00 -7.88134515e-01 -4.40063179e-01
-4.34083402e-01 4.28824186e-01 -5.60192466e-02 2.09090859e-01
-8.58282447e-01 3.81735712e-01 2.39476472e-01 5.62618554e-01
-4.04466540e-01 3.26557070e-01 -5.99446893e-01 -7.81025469e-01
1.33491578e-02 -1.39914715e+00 6.49760902e-01 -1.01382720e+00
-1.04144621e+00 7.78835475e-01 4.72846739e-02 -1.25213623e+00
-2.08104447e-01 -4.11329895e-01 1.43207952e-01 6.78315282e-01
-1.33298719e+00 -8.83818507e-01 -2.78637707e-01 2.51509637e-01
3.91218454e-01 1.16074890e-01 9.08308327e-01 2.29246318e-01
-4.54162598e-01 6.75098598e-01 -3.14927131e-01 1.75307002e-02
1.18095016e+00 -8.48687768e-01 3.54115129e-01 3.71163607e-01
-2.34714583e-01 1.04811060e+00 7.98285544e-01 -9.92959499e-01
-8.84965539e-01 -8.06823313e-01 1.22172511e+00 -8.69198382e-01
5.55166245e-01 5.47712259e-02 -8.81751060e-01 1.00020158e+00
-7.18368664e-02 -3.18811655e-01 1.23724961e+00 -1.32991716e-01
-6.62802234e-02 2.48556077e-01 -1.31800616e+00 5.91883838e-01
7.46182263e-01 -6.64898157e-01 -7.61825144e-01 6.31018817e-01
7.19356000e-01 -6.70557976e-01 -1.28636992e+00 4.74319041e-01
6.23988032e-01 -4.96973693e-01 8.19581926e-01 -9.59954858e-01
7.05353200e-01 4.01349477e-02 6.00400381e-02 -1.15726423e+00
-4.55980688e-01 -5.64250052e-01 5.32785058e-01 5.24598777e-01
9.28583980e-01 -6.88371241e-01 3.62537831e-01 1.44372964e+00
-1.99828651e-02 -6.85518026e-01 -6.45828545e-01 -1.84594318e-01
6.86831474e-02 -7.02397525e-01 3.81268322e-01 9.31440294e-01
2.80390322e-01 4.37276028e-02 -1.52207538e-03 4.75172877e-01
3.22203845e-01 -3.01394224e-01 2.08935350e-01 -9.39556301e-01
-2.29104459e-01 -2.92904943e-01 -3.53674203e-01 -4.79809761e-01
-1.17152572e-01 -7.39143193e-01 7.79770911e-02 -2.15990353e+00
6.41473413e-01 -4.39693570e-01 -3.81674409e-01 4.70759273e-01
-3.92484933e-01 -1.56738758e-01 1.78937212e-01 4.75380868e-01
-3.90805393e-01 -5.50184511e-02 1.09686112e+00 -1.08114421e-01
1.68026369e-02 3.43672410e-02 -1.21015704e+00 7.88354218e-01
8.13925624e-01 -9.54694986e-01 -3.39925140e-01 -5.70850670e-01
1.83505639e-01 3.28790426e-01 3.89492244e-01 -6.95547104e-01
4.66617465e-01 -5.36933765e-02 2.94332713e-01 -1.93805829e-01
-9.58470106e-02 -9.19520020e-01 4.12759095e-01 8.22571993e-01
-8.09478104e-01 6.35530174e-01 4.66579318e-01 3.59204382e-01
-1.14600830e-01 -3.64305556e-01 4.63936925e-01 -4.69926566e-01
-4.79928367e-02 -3.60305831e-02 -8.56767714e-01 9.49845612e-02
9.69876528e-01 -1.93683669e-01 -1.45975992e-01 -5.52843273e-01
-7.63040781e-01 3.88348311e-01 4.22137588e-01 2.54381239e-01
7.91419923e-01 -8.48410785e-01 -8.37310314e-01 -8.67030844e-02
4.06051934e-01 6.97425008e-02 3.80123019e-01 1.21910608e+00
-6.76417172e-01 7.53208697e-01 1.49622053e-01 -6.01959527e-01
-1.17122018e+00 3.69773805e-01 2.81516343e-01 -3.38568121e-01
-6.12340808e-01 7.77627230e-01 2.35796273e-01 -3.32393557e-01
3.12202066e-01 -8.71460557e-01 -2.73629017e-02 -1.74395308e-01
5.24049044e-01 1.65880799e-01 3.05463582e-01 -4.63096052e-01
-5.23237467e-01 2.74253905e-01 -2.73444116e-01 -2.69936860e-01
1.15694034e+00 -2.32246801e-01 -5.47962524e-02 5.01092315e-01
7.51754642e-01 4.07914221e-02 -5.62282085e-01 -4.55155782e-02
2.03940690e-01 -2.18146637e-01 2.40610510e-01 -1.30200350e+00
-5.89633644e-01 5.83045065e-01 4.95886624e-01 -2.93715745e-01
7.83240676e-01 1.18225835e-01 3.18398565e-01 2.22987197e-02
3.51433486e-01 -8.60071480e-01 -2.19919100e-01 -2.36845672e-01
6.70346022e-01 -1.51440346e+00 7.50837550e-02 -2.35098675e-01
-1.29978597e+00 8.22785616e-01 3.82570088e-01 6.01453066e-01
7.37720191e-01 4.21826124e-01 6.11337066e-01 -6.35492742e-01
-7.84175634e-01 5.39840519e-01 1.94784492e-01 1.40613720e-01
7.79759526e-01 4.21594948e-01 -4.99076039e-01 9.61386919e-01
-3.51148874e-01 2.58410752e-01 8.42219353e-01 1.09910131e+00
1.05172850e-01 -7.80781984e-01 -4.41183299e-01 6.81812048e-01
-8.14650595e-01 -3.23546112e-01 -1.87880009e-01 7.01088786e-01
-1.89088225e-01 1.26077890e+00 -5.39578140e-01 -3.03699404e-01
6.19556069e-01 -5.82513921e-02 2.26735370e-03 -9.24026847e-01
-1.10178173e+00 2.48180300e-01 2.50570238e-01 -7.08884895e-01
-4.33638513e-01 -7.01081216e-01 -1.20071602e+00 -1.58279389e-01
-3.13000113e-01 3.67729694e-01 6.54023409e-01 8.18126380e-01
4.88275141e-01 7.87811399e-01 -2.71064565e-02 2.73061395e-01
-7.94746518e-01 -9.47593749e-01 -2.52664477e-01 9.14038345e-02
4.13948387e-01 -4.09691662e-01 -2.02605426e-01 -4.17354628e-02] | [8.71349811553955, 8.37412166595459] |
fb494a6e-0278-4d9b-b084-6f99d6dd310d | a-unified-learning-approach-for-hand-gesture | 2101.02047 | null | https://arxiv.org/abs/2101.02047v3 | https://arxiv.org/pdf/2101.02047v3.pdf | Unified Learning Approach for Egocentric Hand Gesture Recognition and Fingertip Detection | Head-mounted device-based human-computer interaction often requires egocentric recognition of hand gestures and fingertips detection. In this paper, a unified approach of egocentric hand gesture recognition and fingertip detection is introduced. The proposed algorithm uses a single convolutional neural network to predict the probabilities of finger class and positions of fingertips in one forward propagation. Instead of directly regressing the positions of fingertips from the fully connected layer, the ensemble of the position of fingertips is regressed from the fully convolutional network. Subsequently, the ensemble average is taken to regress the final position of fingertips. Since the whole pipeline uses a single network, it is significantly fast in computation. Experimental results show that the proposed method outperforms the existing fingertip detection approaches including the Direct Regression and the Heatmap-based framework. The effectiveness of the proposed method is also shown in-the-wild scenario as well as in a use-case of virtual reality. | ['S. M. Mahbubur Rahman', 'Mohammad Tariqul Islam', 'Mohammad Mahmudul Alam'] | 2021-01-06 | null | null | null | null | ['fingertip-detection'] | ['computer-vision'] | [ 1.05475806e-01 -1.60224661e-01 7.69929355e-03 -3.40881437e-01
-3.49532992e-01 -6.03025079e-01 4.96821135e-01 -5.42285025e-01
-7.98495412e-01 3.87675256e-01 -2.79046055e-02 -7.36161619e-02
-6.59466237e-02 -4.60772574e-01 -6.33301318e-01 -8.44400287e-01
8.93379003e-02 3.99876505e-01 3.26646984e-01 3.08691233e-01
4.92411792e-01 1.02685392e+00 -1.56620443e+00 -1.96734909e-02
-5.79994097e-02 1.06639647e+00 1.15454912e-01 1.27990389e+00
7.48354793e-02 4.87705350e-01 -3.36635709e-01 -3.53305452e-02
3.15334439e-01 4.27549109e-02 -1.79129362e-01 -3.51296097e-01
7.47916162e-01 -9.40335393e-01 -5.96817791e-01 7.08987534e-01
1.16198516e+00 1.96425021e-01 5.10164261e-01 -9.83095467e-01
7.99151957e-02 1.33436307e-01 -6.20678902e-01 6.76633716e-02
2.19240010e-01 2.77671337e-01 6.83662951e-01 -1.11240602e+00
5.32369316e-01 1.16393781e+00 4.81321752e-01 5.93698084e-01
-6.83258355e-01 -6.94826126e-01 3.02858412e-01 1.36851341e-01
-1.64468288e+00 -3.01940203e-01 9.46527719e-01 -4.66387808e-01
1.07867169e+00 2.94048935e-01 4.23207551e-01 1.04605627e+00
-1.43448282e-02 1.15216279e+00 8.19896579e-01 -6.30642593e-01
1.82949498e-01 -8.01083893e-02 3.64856452e-01 7.39265680e-01
1.72453910e-01 -2.29364214e-03 -5.62092543e-01 -1.39572650e-01
1.25990987e+00 2.71311969e-01 3.11526865e-01 -2.82207578e-01
-1.25785482e+00 6.24830388e-02 5.27496576e-01 4.41829085e-01
-1.01480925e+00 4.46228981e-01 6.37119114e-02 -3.79737824e-01
2.76480675e-01 -2.16702014e-01 -5.28628111e-01 -3.23570400e-01
-1.07570171e+00 6.45089030e-01 7.90759861e-01 9.66738045e-01
2.66335249e-01 -2.69073129e-01 -7.49602497e-01 3.37628782e-01
5.24790406e-01 4.93366271e-01 1.49786230e-02 -5.70917010e-01
6.63380563e-01 6.95334077e-01 6.72452569e-01 -6.33540690e-01
-6.05288029e-01 -9.59799662e-02 -3.16801608e-01 6.24003053e-01
9.71864581e-01 -5.91612995e-01 -1.14789939e+00 1.35707235e+00
3.22922468e-01 2.93513119e-01 -5.83857834e-01 1.14959323e+00
4.03926581e-01 -2.17909649e-01 2.78945416e-01 3.97535890e-01
1.13051724e+00 -7.68357277e-01 -5.75837255e-01 7.88976103e-02
1.11955982e-02 -7.39104152e-01 8.67971659e-01 2.99099147e-01
-1.05894923e+00 -5.85050106e-01 -8.73893499e-01 1.59743279e-02
-5.01541793e-01 7.34404564e-01 5.76081038e-01 9.37324286e-01
-8.43882442e-01 7.08239675e-01 -1.18868339e+00 -5.67582309e-01
5.43526232e-01 7.07582533e-01 5.72734438e-02 3.16330373e-01
-5.07534623e-01 6.68848157e-01 1.67221069e-01 4.49114114e-01
-2.19775662e-01 -4.80900139e-01 -3.82884473e-01 1.60779089e-01
1.83257777e-02 -6.08028471e-01 1.17224431e+00 -4.47930396e-01
-1.95922267e+00 5.98789513e-01 -7.30312288e-01 5.53999357e-02
1.03604269e+00 -7.94636905e-01 -5.53040951e-02 -1.50907353e-01
-4.29634482e-01 4.99811351e-01 9.54261720e-01 -9.73577440e-01
-9.80644763e-01 -9.83433962e-01 -8.72176364e-02 2.40706831e-01
-1.92884579e-01 2.12293610e-01 -7.58084416e-01 -2.72495776e-01
3.33115757e-01 -1.06458616e+00 1.07376136e-01 -8.64525791e-03
-7.89519191e-01 -6.47384763e-01 8.18459272e-01 -7.52847970e-01
1.14244282e+00 -1.75438321e+00 1.05724577e-02 5.08881509e-01
1.19128101e-01 3.78549486e-01 8.04822743e-02 6.61896020e-02
-1.45199616e-02 -3.60213578e-01 3.23375583e-01 -5.25083184e-01
1.02364630e-01 -1.20313436e-01 -9.99346152e-02 4.86237764e-01
9.46018398e-02 1.07318377e+00 -7.53277719e-01 -4.41492528e-01
8.68624210e-01 9.08002615e-01 -2.52161294e-01 2.86139727e-01
1.49147570e-01 6.98209941e-01 -5.12770355e-01 8.53269041e-01
7.93781877e-01 2.30986640e-01 9.44427624e-02 -2.73060709e-01
-3.64103764e-01 1.42519057e-01 -1.33882785e+00 1.78248048e+00
-3.62800956e-01 6.14943147e-01 -7.85108283e-02 -1.89796060e-01
7.89926589e-01 3.77244234e-01 4.95748192e-01 -3.72223496e-01
3.34217906e-01 7.42856339e-02 -3.78710553e-02 -6.03138149e-01
4.00049180e-01 4.48384553e-01 2.17309326e-01 5.72785556e-01
-9.61579010e-03 6.70693457e-01 -2.38597989e-01 -2.68116385e-01
8.48061383e-01 8.44903350e-01 -1.17473848e-01 3.39382678e-01
5.46021700e-01 -3.10881704e-01 -7.45737040e-03 1.02740538e+00
-4.73823190e-01 6.49471581e-01 1.73419297e-01 -6.15249217e-01
-1.03392601e+00 -1.10406494e+00 1.26178473e-01 1.41785789e+00
7.90525600e-02 3.43201421e-02 -1.07081759e+00 -7.24896669e-01
1.67081594e-01 6.14604235e-01 -7.27834702e-01 4.09284920e-01
-1.03369391e+00 -4.25929666e-01 6.86232746e-01 9.66714263e-01
3.71744066e-01 -1.29466629e+00 -1.11066699e+00 -2.76460517e-02
2.50151217e-01 -9.16472495e-01 -7.17849806e-02 -1.52993724e-01
-8.02638650e-01 -1.19014180e+00 -1.40611231e+00 -6.01352990e-01
5.98634362e-01 2.29839794e-02 4.64744031e-01 -3.59469116e-01
-4.91680354e-01 5.48860312e-01 -2.08900142e-02 -7.14913130e-01
3.92392188e-01 2.55669802e-01 1.53743535e-01 2.83607960e-01
9.03049171e-01 -4.94504601e-01 -1.13615775e+00 4.25694659e-02
-9.65509489e-02 -4.33194101e-01 5.41151404e-01 3.82462054e-01
3.60127330e-01 -5.63030839e-01 -1.55504970e-02 -3.83512408e-01
6.35095119e-01 -1.63989924e-02 -3.29912871e-01 4.02493954e-01
-2.94720709e-01 -7.96832591e-02 1.27903342e-01 -3.60713065e-01
-1.25762689e+00 8.34856749e-01 9.81093794e-02 -6.15090191e-01
-7.06553996e-01 -2.62527436e-01 2.94604078e-02 -7.25847930e-02
6.40250504e-01 9.83897522e-02 -2.40993366e-01 -9.60854888e-01
5.70933044e-01 1.07631397e+00 6.83908939e-01 -1.41731218e-01
5.28113782e-01 5.07874370e-01 3.93238105e-02 -8.55025351e-01
7.89031014e-03 -6.56733394e-01 -1.38482559e+00 -5.30286968e-01
7.36277342e-01 -5.67946911e-01 -1.22944534e+00 8.53670716e-01
-1.49685419e+00 -2.54290570e-02 9.89123881e-02 5.45325577e-01
-3.99383187e-01 2.85651654e-01 -1.82128057e-01 -1.60690522e+00
-6.11198843e-01 -7.98599064e-01 1.38773525e+00 6.09582424e-01
-4.45888519e-01 -6.32899225e-01 5.42720631e-02 -2.39806578e-01
3.64357948e-01 2.27795079e-01 4.38237429e-01 -5.88197589e-01
-4.81965929e-01 -1.14937282e+00 -3.62310469e-01 -2.37551123e-01
2.89680451e-01 -9.01517179e-03 -1.50173259e+00 -1.80346638e-01
-6.31766558e-01 1.91194952e-01 5.99162757e-01 7.13022053e-01
1.28391206e+00 1.82818741e-01 -4.62822825e-01 3.83732498e-01
1.18399918e+00 1.01550601e-01 5.52363634e-01 8.55750218e-02
1.05105579e+00 7.35376596e-01 4.70701247e-01 7.41333067e-01
2.93122649e-01 7.94946909e-01 5.39132833e-01 1.18165933e-01
-4.50663008e-02 -1.05107456e-01 2.93253213e-02 -3.24872248e-02
-9.96401906e-01 -3.92383188e-02 -7.16690540e-01 3.79313052e-01
-1.96820724e+00 -8.20573926e-01 2.81377770e-02 2.53618097e+00
2.18232069e-02 -3.03952485e-01 4.83066291e-01 1.91136897e-01
1.04016602e+00 5.33722807e-03 -9.16424930e-01 -1.87083110e-01
4.03601974e-01 5.58189511e-01 8.15137565e-01 4.81847793e-01
-1.50175619e+00 1.25988662e+00 6.00114775e+00 5.35296425e-02
-1.24319613e+00 -6.95523247e-02 -8.84539098e-04 -4.47791427e-01
7.35768914e-01 -4.51417208e-01 -1.02123570e+00 3.49771649e-01
4.97169137e-01 4.61299181e-01 7.27594256e-01 8.98266494e-01
2.39960223e-01 -3.40858251e-02 -1.13235199e+00 1.26625597e+00
-1.71289116e-01 -7.46398151e-01 -3.63749564e-01 -5.52543700e-02
5.66092357e-02 3.11902583e-01 4.63279597e-02 -1.67765394e-01
-6.51092455e-02 -8.73815775e-01 6.50920689e-01 1.06346929e+00
7.57567585e-01 -7.56793141e-01 5.46524465e-01 4.68746305e-01
-1.26110542e+00 -1.93514287e-01 -1.40746415e-01 -2.24335313e-01
7.94041157e-02 -1.13935746e-01 -1.19838655e+00 1.80138826e-01
7.46784031e-01 2.93658286e-01 -4.00157064e-01 1.06559455e+00
-1.86791629e-01 2.87966847e-01 -5.67566752e-01 -3.91047984e-01
-6.87340572e-02 2.56398976e-01 7.41795778e-01 1.48687339e+00
2.53282756e-01 -8.57249945e-02 -3.29420418e-01 9.99424160e-01
4.21082020e-01 -1.22136371e-02 -4.25149441e-01 -6.92773536e-02
4.19334620e-01 1.49134874e+00 -6.94194376e-01 -3.90492201e-01
-1.06654175e-01 1.39963591e+00 1.30091801e-01 6.52732193e-01
-4.43260521e-01 -9.39585447e-01 7.46849298e-01 -6.69776872e-02
4.30606097e-01 -3.41614425e-01 -6.82387829e-01 -8.41042101e-01
3.62473488e-01 6.76831696e-03 2.98371408e-02 -5.79181135e-01
-8.58423233e-01 3.92881960e-01 -3.29525888e-01 -1.07720995e+00
-8.62480700e-01 -1.11884701e+00 -8.20071578e-01 1.53351271e+00
-1.18678319e+00 -1.23875856e+00 -9.01594162e-01 8.09720516e-01
2.90976286e-01 -2.01751873e-01 9.59689319e-01 1.86494529e-01
-7.60469437e-01 9.30271208e-01 -1.16452023e-01 4.50920254e-01
7.69741833e-01 -1.23893893e+00 9.25078094e-01 6.68611109e-01
-1.04009651e-01 1.09448051e+00 3.82116020e-01 -7.81393766e-01
-1.49594271e+00 -7.99116492e-01 8.51856112e-01 -8.11167300e-01
4.04690206e-02 -2.63989687e-01 -4.01021034e-01 7.08731592e-01
-2.13650376e-01 1.14940360e-01 3.27919483e-01 1.73501581e-01
-1.79603308e-01 1.16363488e-01 -1.20715117e+00 7.51522362e-01
1.16024995e+00 -6.28706336e-01 -2.73608446e-01 2.84291953e-01
-1.35352641e-01 -5.28727174e-01 -6.20442212e-01 1.15863524e-01
1.68423378e+00 -8.48336935e-01 1.14044631e+00 -8.03852022e-01
-2.02977896e-01 -1.35024294e-01 -1.41254649e-01 -6.32060826e-01
-4.07215893e-01 -5.52704155e-01 -6.17655158e-01 9.57471251e-01
-1.01464614e-01 -5.22111058e-01 1.40660834e+00 1.12713075e+00
4.68622029e-01 -4.79923457e-01 -9.74791586e-01 -3.52172703e-01
-2.77520239e-01 -5.94685018e-01 5.68176568e-01 3.05589199e-01
-6.33148029e-02 -1.45445004e-01 -2.79931366e-01 2.08277494e-01
9.47869301e-01 -2.10363165e-01 1.03803706e+00 -1.40630460e+00
-1.58205479e-01 -5.44344008e-01 -5.04836977e-01 -9.61423337e-01
-2.78091848e-01 -4.00231868e-01 2.73166485e-02 -1.26631904e+00
3.67707312e-02 -2.53882080e-01 -8.16403568e-01 3.87812942e-01
-2.11472601e-01 2.66737163e-01 2.77884722e-01 1.77364647e-01
-1.43993601e-01 -3.30475241e-01 1.01524413e+00 1.01575971e-01
-7.54304945e-01 4.68612283e-01 -4.50550169e-02 9.04146969e-01
8.04629147e-01 -1.79398030e-01 8.09061155e-03 -1.65228680e-01
-2.84547597e-01 -2.96132505e-01 8.63200724e-01 -9.91881013e-01
5.54515660e-01 2.12074876e-01 9.36966836e-01 -1.04813027e+00
5.19372940e-01 -7.16318607e-01 -4.62700516e-01 5.62144458e-01
-2.35537022e-01 -2.04908490e-01 -1.32055869e-02 4.36583608e-01
5.75927377e-01 2.09447891e-02 2.61349797e-01 -6.51727170e-02
-6.53761089e-01 2.22134367e-01 -1.75124884e-01 -8.36856246e-01
7.65156209e-01 -5.19924045e-01 1.43225864e-01 -1.71448618e-01
-8.76625776e-01 -2.57395923e-01 -4.05654457e-04 7.84644663e-01
5.87301254e-01 -1.03930187e+00 -4.14931923e-01 3.94654095e-01
-9.30388942e-02 -1.86393186e-01 1.79892674e-01 8.02583992e-01
-3.92622322e-01 9.08438683e-01 -4.24146444e-01 -7.57309914e-01
-1.41722000e+00 1.31279811e-01 4.40434158e-01 1.84019014e-01
-5.78469992e-01 9.82310653e-01 -1.49283871e-01 -4.62990671e-01
9.68094647e-01 -3.16384494e-01 -4.37980115e-01 -1.53369740e-01
7.35266447e-01 8.94114077e-01 -6.96110576e-02 -6.96148098e-01
-6.51552260e-01 8.69479239e-01 -6.34753257e-02 -4.78380144e-01
1.19999051e+00 1.03001773e-01 7.00830221e-02 3.39844525e-01
9.10825372e-01 -2.21706450e-01 -1.68351448e+00 -1.75966427e-01
-1.97827686e-02 -5.70011914e-01 7.35182092e-02 -1.03680813e+00
-8.18226755e-01 1.28539705e+00 1.41371047e+00 -5.23241401e-01
6.70782030e-01 -3.16959381e-01 5.85147560e-01 5.37689388e-01
3.91364217e-01 -1.13069463e+00 -3.98409784e-01 3.76839310e-01
8.04124832e-01 -1.01068246e+00 -1.95524231e-01 -1.09802410e-01
-1.91676036e-01 1.23801875e+00 5.06498277e-01 -4.31044668e-01
7.11306572e-01 2.29199618e-01 1.14308238e-01 -2.15130672e-02
1.82131171e-01 -2.39215925e-01 5.46889782e-01 6.22395575e-01
5.96529305e-01 2.14415476e-01 -2.05423966e-01 5.47648728e-01
1.05057202e-01 5.44555843e-01 -3.95121545e-01 1.23955607e+00
-6.87336251e-02 -9.82569218e-01 -4.77013856e-01 2.99980551e-01
-3.91080439e-01 1.35160178e-01 -7.70592272e-01 5.93788683e-01
1.12284631e-01 6.68672383e-01 3.70722145e-01 -8.17212462e-01
4.56968755e-01 8.15720081e-01 1.02749932e+00 -1.54956192e-01
-6.22056007e-01 6.53937608e-02 -4.40615624e-01 -8.53246450e-01
-2.55784422e-01 -6.54916525e-01 -1.27194059e+00 2.01694760e-02
-2.08612248e-01 -7.18080640e-01 1.35956883e+00 1.03318262e+00
5.56837678e-01 5.05338609e-01 3.71971399e-01 -1.97725904e+00
-6.46180749e-01 -1.34380615e+00 -5.64440548e-01 -7.79274926e-02
2.62780130e-01 -7.43410408e-01 8.50391611e-02 -2.63270944e-01] | [6.480550289154053, -0.42162081599235535] |
6d7f7557-957c-416d-a830-d22840717b2c | idioms-probing-and-dangerous-things-towards | 2304.14333 | null | https://arxiv.org/abs/2304.14333v1 | https://arxiv.org/pdf/2304.14333v1.pdf | Idioms, Probing and Dangerous Things: Towards Structural Probing for Idiomaticity in Vector Space | The goal of this paper is to learn more about how idiomatic information is structurally encoded in embeddings, using a structural probing method. We repurpose an existing English verbal multi-word expression (MWE) dataset to suit the probing framework and perform a comparative probing study of static (GloVe) and contextual (BERT) embeddings. Our experiments indicate that both encode some idiomatic information to varying degrees, but yield conflicting evidence as to whether idiomaticity is encoded in the vector norm, leaving this an open question. We also identify some limitations of the used dataset and highlight important directions for future work in improving its suitability for a probing analysis. | ['John D. Kelleher', 'Vasudevan Nedumpozhimana', 'Filip Klubička'] | 2023-04-27 | null | null | null | null | ['open-question'] | ['natural-language-processing'] | [ 2.63294250e-01 1.51322246e-01 -5.49132884e-01 -6.19088709e-01
-2.53395766e-01 -8.84805083e-01 9.30615425e-01 8.00990313e-03
-6.01180196e-01 5.11778116e-01 1.10030878e+00 -6.03367567e-01
-3.10372651e-01 -8.50221097e-01 -2.70405829e-01 -3.86210144e-01
-1.22086883e-01 5.70516646e-01 4.24925238e-02 -7.35899091e-01
4.90661025e-01 6.54946923e-01 -1.00838745e+00 6.52666748e-01
2.74111509e-01 6.04907930e-01 -1.93152398e-01 2.86219656e-01
-1.88393816e-01 9.80309784e-01 -6.00143909e-01 -7.01909184e-01
1.70868322e-01 -4.00983989e-01 -1.26816523e+00 -1.35133609e-01
3.81684810e-01 4.84643616e-02 -1.01808637e-01 9.57957864e-01
3.10895979e-01 3.62276249e-02 6.78760409e-01 -8.92823637e-01
-9.09938574e-01 9.00233686e-01 -8.88956413e-02 6.08492851e-01
3.25429499e-01 -3.51187319e-01 1.53384614e+00 -1.01753104e+00
1.07041633e+00 1.18203485e+00 6.66829824e-01 4.25671697e-01
-1.22898376e+00 -3.96000266e-01 1.48136690e-01 5.72916031e-01
-1.10831547e+00 -4.99036819e-01 1.38602448e+00 -4.90818113e-01
1.04893470e+00 4.35161412e-01 5.38076043e-01 1.53230953e+00
1.58839330e-01 5.75151145e-01 1.62245858e+00 -4.05223429e-01
3.13041843e-02 3.53466153e-01 3.42094660e-01 7.94014275e-01
-1.90282106e-01 2.91481435e-01 -4.43265736e-01 -3.46316457e-01
6.64639652e-01 -6.55044794e-01 -1.61924347e-01 -4.37849611e-01
-9.87766385e-01 1.57019091e+00 5.33404291e-01 7.55496144e-01
-2.81822324e-01 -2.89973989e-02 8.39478493e-01 7.26422012e-01
4.92167234e-01 6.90976381e-01 -6.66698098e-01 -5.54956079e-01
-5.64033389e-01 2.43292868e-01 8.43732774e-01 2.78587341e-01
5.93359768e-01 -2.88478639e-02 4.23907489e-01 1.04306543e+00
1.40826285e-01 -2.51989454e-01 5.40202856e-01 -9.21500266e-01
3.13592911e-01 3.74770552e-01 -2.36597478e-01 -1.41091454e+00
-4.22638625e-01 4.28954847e-02 -3.71843316e-02 -2.06207614e-02
1.13143735e-01 -1.22003593e-01 -1.67805448e-01 2.01185799e+00
8.58750269e-02 -3.30527693e-01 -3.29080410e-02 9.05984640e-01
7.94805288e-01 5.39767027e-01 1.06334895e-01 -4.36170734e-02
1.78545201e+00 -7.83123076e-01 -7.10229218e-01 -4.25890118e-01
6.53576672e-01 -7.98877537e-01 1.24743950e+00 2.86223948e-01
-8.21704686e-01 -3.26044381e-01 -1.07737720e+00 -1.74954042e-01
-5.69505155e-01 -1.60380617e-01 1.03248167e+00 9.61939752e-01
-7.54243314e-01 3.05673957e-01 -7.68999994e-01 -3.87862176e-01
8.62892866e-02 -1.00815721e-01 -5.85196197e-01 2.38568753e-01
-1.45408273e+00 1.50673044e+00 5.16062975e-01 -8.86536315e-02
-2.16043577e-01 -6.00907564e-01 -1.22404683e+00 -3.83163124e-01
1.33056968e-01 -1.96746424e-01 7.78655469e-01 -9.19454694e-01
-1.52206206e+00 1.19532180e+00 -2.61205193e-02 -2.47357726e-01
2.66072303e-01 1.48479953e-01 -7.47900009e-01 2.15274706e-01
-3.59677710e-02 5.87590098e-01 3.53111386e-01 -1.31590593e+00
-9.89146437e-03 -5.12212098e-01 5.59621930e-01 2.20666267e-02
-3.00819755e-01 6.86648786e-01 2.47453824e-01 -1.14192879e+00
7.62210712e-02 -8.98746490e-01 1.54181331e-01 -1.04497842e-01
5.29170856e-02 -2.17349127e-01 9.01755989e-01 -6.62820578e-01
1.38127506e+00 -1.95455718e+00 6.34788036e-01 -6.30575418e-02
-1.06006935e-01 -7.17515424e-02 -2.06752300e-01 9.32977498e-01
-3.99663657e-01 2.83881545e-01 -3.93570840e-01 -5.57659008e-02
2.09547445e-01 8.52015972e-01 -4.10075665e-01 6.50620937e-01
3.67895275e-01 1.30709326e+00 -5.90790331e-01 -4.56747264e-01
6.32158145e-02 5.11465728e-01 -7.68341303e-01 -1.39751345e-01
4.87153828e-02 1.54102966e-01 -3.58091742e-02 7.62753367e-01
4.09926981e-01 3.52237612e-01 5.20297885e-01 -5.01432359e-01
-2.07314923e-01 7.03688502e-01 -9.84487772e-01 1.53423631e+00
-7.38810480e-01 1.04565239e+00 9.23092142e-02 -1.51054156e+00
8.54733169e-01 4.25299816e-02 9.78944376e-02 -8.57136726e-01
2.08412439e-01 1.46946043e-01 3.89810026e-01 -4.86578315e-01
6.00449085e-01 -9.52907145e-01 -5.60083926e-01 5.88689804e-01
1.10634960e-01 -1.22278199e-01 -1.07195452e-01 -9.88409519e-02
8.89971018e-01 -3.92923839e-02 5.45150757e-01 -7.46631563e-01
4.56498802e-01 4.37807495e-04 4.89002794e-01 1.43947750e-01
-3.24625015e-01 1.92859545e-01 7.14498460e-01 -6.18433297e-01
-9.94606197e-01 -1.26775265e+00 -9.11962867e-01 1.22785592e+00
-1.79453731e-01 -6.34778619e-01 -3.47300261e-01 -8.71330023e-01
-1.79775059e-01 8.77409399e-01 -1.23472786e+00 4.73638289e-02
-9.93895113e-01 -9.81266975e-01 7.17673004e-01 9.08831537e-01
8.45988169e-02 -1.34306622e+00 -1.01575196e+00 2.86566526e-01
-2.40421236e-01 -9.78284895e-01 -4.16714996e-02 5.02815843e-01
-7.04163909e-01 -9.56205308e-01 5.33683226e-02 -8.39025736e-01
2.15296537e-01 -3.01493645e-01 1.16546214e+00 -1.30332932e-01
-1.63094774e-01 1.93499327e-01 -6.96196139e-01 -1.11425929e-01
-4.14456248e-01 -1.42339356e-02 -8.07023793e-02 -3.69517654e-01
6.53475344e-01 -6.12793207e-01 -1.36159807e-01 2.40826085e-01
-1.08230877e+00 -3.07673186e-01 1.53073296e-01 1.16733301e+00
-8.10230523e-02 -5.47449179e-02 3.66221964e-01 -1.09356844e+00
1.04779387e+00 -6.30653024e-01 -7.57801533e-02 -2.25856140e-01
-2.90311307e-01 7.22332299e-02 4.37183768e-01 -4.59430724e-01
-7.94108868e-01 -7.00155914e-01 -6.60146296e-01 1.23139635e-01
5.00047393e-02 8.87164772e-01 -1.85181528e-01 -1.31023899e-01
6.35227501e-01 -6.58171400e-02 6.99419379e-02 -4.58380073e-01
6.10446632e-01 5.40697992e-01 3.98895085e-01 -1.13741314e+00
4.51990038e-01 4.49440062e-01 -3.48892331e-01 -1.01217139e+00
-7.55814254e-01 -1.91535190e-01 -7.64074922e-01 3.39985311e-01
7.45832682e-01 -4.36233699e-01 -2.97109663e-01 -1.09937914e-01
-1.18103242e+00 -3.08860093e-01 -2.41465330e-01 9.49740186e-02
-6.43830538e-01 4.99958128e-01 -9.29268003e-01 -4.43934321e-01
9.23662707e-02 -1.13048637e+00 7.92385101e-01 -5.00449121e-01
-1.11875343e+00 -1.58960569e+00 5.08647442e-01 4.96197015e-01
5.14346421e-01 3.97595108e-01 1.50946963e+00 -8.27977836e-01
3.52535576e-01 1.63328126e-01 -2.37925753e-01 1.78190067e-01
3.24058533e-01 -1.81614742e-01 -8.76004040e-01 -2.09648609e-01
3.43747437e-01 -5.12459219e-01 8.28745186e-01 -2.78306365e-01
7.73945808e-01 -5.45916617e-01 1.62613720e-01 5.01035035e-01
1.47353005e+00 3.53242196e-02 5.84846437e-01 7.39088655e-01
3.80746335e-01 9.45968151e-01 4.70929623e-01 2.65469044e-01
3.60368848e-01 1.00113797e+00 2.46784359e-01 2.66873121e-01
1.55839557e-02 -1.71285152e-01 4.84068066e-01 1.09607005e+00
2.00207025e-01 2.29098126e-02 -8.84704471e-01 8.46528471e-01
-1.47465038e+00 -1.01831985e+00 2.10840970e-01 1.38892996e+00
1.09976447e+00 5.56501485e-02 1.26416862e-01 3.08060616e-01
-2.41877176e-02 8.51278901e-01 7.35893101e-02 -1.23627472e+00
-3.70067447e-01 8.70370865e-01 2.06267908e-02 9.73798931e-01
-9.30379987e-01 1.07539940e+00 7.40244913e+00 4.35670346e-01
-1.09243619e+00 3.24313283e-01 6.05826676e-02 2.75418818e-01
-6.49148762e-01 8.15499425e-02 -2.51298755e-01 3.73695165e-01
1.14782608e+00 1.13413066e-01 4.43936646e-01 6.72448218e-01
-2.39763990e-01 2.63686001e-01 -9.31207836e-01 5.48004031e-01
2.79837132e-01 -1.32300973e+00 1.15016922e-01 3.60563514e-03
2.83946365e-01 1.49243981e-01 -2.41384516e-03 3.56000841e-01
1.53512731e-01 -1.23143446e+00 6.90976858e-01 -7.14362040e-02
6.12837136e-01 -8.52291524e-01 8.76587808e-01 -1.87762454e-03
-9.42065716e-01 -1.81547657e-01 -3.72337133e-01 -4.92060363e-01
4.26730335e-01 -3.39430198e-02 -5.85998893e-01 3.33761483e-01
3.20953280e-01 5.04201055e-01 -7.45129287e-01 6.61472827e-02
-4.78272468e-01 5.98181427e-01 -5.87418154e-02 -7.66783133e-02
5.93359530e-01 -2.84478635e-01 5.07857084e-01 1.63436925e+00
-2.73609936e-01 -9.74891055e-03 1.18239246e-01 1.06085408e+00
2.97346443e-01 2.71149307e-01 -6.62199497e-01 -4.53278273e-01
3.62325132e-01 1.19688690e+00 -7.03038335e-01 -1.01394787e-01
-5.24460554e-01 1.24681377e+00 5.25087655e-01 -4.37914208e-02
-7.96243310e-01 -3.38451862e-01 1.08683228e+00 -6.10437542e-02
2.86587089e-01 -4.88533080e-01 -4.15166169e-01 -1.17793357e+00
-9.27517191e-02 -1.00283134e+00 4.48490053e-01 -5.12732685e-01
-1.40821409e+00 6.34787083e-01 2.08912939e-01 -4.92929190e-01
-3.43357950e-01 -1.13370430e+00 -5.02543032e-01 5.93707979e-01
-1.22970665e+00 -1.22748291e+00 1.18236817e-01 2.15953842e-01
4.70567226e-01 -4.84391004e-02 1.32661927e+00 1.65944219e-01
-3.33838046e-01 5.56276381e-01 -4.49170232e-01 3.78653020e-01
3.80027413e-01 -1.31698716e+00 3.44237804e-01 3.00358683e-01
6.50405526e-01 1.02003658e+00 9.41578388e-01 -2.84004241e-01
-1.23268282e+00 -4.61506158e-01 1.12371731e+00 -8.78189564e-01
1.29457426e+00 -4.70975012e-01 -9.01843369e-01 9.12273169e-01
5.82237840e-01 -5.20395562e-02 1.20251882e+00 4.47997600e-01
-8.63141894e-01 6.55185759e-01 -1.24309039e+00 6.28934443e-01
1.13874602e+00 -1.04740191e+00 -1.47984231e+00 -1.83309719e-01
6.72098339e-01 2.78246142e-02 -1.14722490e+00 5.80681741e-01
6.31630421e-01 -1.01806808e+00 8.27548683e-01 -1.02242923e+00
6.85014665e-01 6.94789514e-02 -8.39947701e-01 -1.53609562e+00
-3.08232248e-01 -4.70388914e-03 1.10224262e-01 1.10556984e+00
2.90370494e-01 -7.28213787e-01 6.57625496e-01 8.07647705e-02
-5.28727882e-02 -9.78042722e-01 -1.15773988e+00 -5.52977860e-01
7.37937331e-01 -5.81220448e-01 2.37900183e-01 1.64904034e+00
5.11417389e-01 6.98040247e-01 -8.80687237e-02 -1.36500955e-01
-4.56745438e-02 2.32216507e-01 3.81363720e-01 -1.08661294e+00
-3.17241341e-01 -7.93511748e-01 -7.91053355e-01 -7.92804062e-01
8.66064668e-01 -1.21193337e+00 -6.14823997e-01 -1.18302858e+00
1.56498909e-01 -1.29814550e-01 -2.26643071e-01 4.04025823e-01
4.19797480e-01 6.25318825e-01 1.07836917e-01 -2.86313832e-01
1.87094230e-02 5.46277702e-01 6.49206519e-01 -9.29416195e-02
1.59115851e-01 -7.88354516e-01 -9.76085961e-01 6.83311880e-01
7.49730706e-01 -2.74572462e-01 -3.40350568e-01 -3.83416355e-01
3.94615531e-01 -1.49187222e-01 3.13211292e-01 -1.66730329e-01
-3.32155228e-01 -2.90032238e-01 5.68298511e-02 -1.53440729e-01
5.82899034e-01 -6.73153579e-01 -6.67877793e-02 4.21353817e-01
-3.64198267e-01 6.77447021e-01 5.29064715e-01 -1.93416961e-02
-3.16051364e-01 -3.89910489e-01 7.36921966e-01 -2.35818028e-01
-9.72884059e-01 -3.47004324e-01 -5.19683003e-01 2.77990341e-01
6.78556740e-01 -2.41234481e-01 -2.80292660e-01 -9.75501835e-02
-5.79030097e-01 -2.50220120e-01 6.80409729e-01 7.34853685e-01
5.48219025e-01 -1.66114473e+00 -5.74460566e-01 3.34218532e-01
4.68666583e-01 -1.03199363e+00 -2.28694692e-01 7.63454437e-01
-4.65667903e-01 4.24512684e-01 -4.66115355e-01 -9.81654599e-02
-1.21932864e+00 6.77396536e-01 1.16892710e-01 1.64138973e-01
-6.38142586e-01 8.15623999e-01 -3.02048475e-01 -8.87696445e-01
-3.18538517e-01 -1.59529328e-01 -1.87115356e-01 7.25831270e-01
4.04638290e-01 1.03225864e-01 -1.07715018e-02 -1.22848344e+00
-5.93179345e-01 5.60078025e-01 -5.20036630e-02 -4.79313046e-01
1.44864643e+00 1.48951575e-01 -5.91815889e-01 9.27116036e-01
1.68614841e+00 7.37971067e-01 -3.76889646e-01 -3.00338417e-02
3.68997931e-01 -5.47136009e-01 -1.43162891e-01 -8.28008175e-01
-6.17352009e-01 8.51610363e-01 2.84330845e-01 1.86861396e-01
7.03811765e-01 3.32039118e-01 4.96422917e-01 1.67578816e-01
4.09781843e-01 -9.83628809e-01 1.36883274e-01 7.44395852e-01
1.07958484e+00 -7.89939582e-01 -1.41881257e-01 -2.06064805e-01
-7.55614579e-01 1.19940460e+00 3.07902843e-01 -3.82963657e-01
5.89146852e-01 4.86302555e-01 4.51812506e-01 -7.22217143e-01
-6.82345450e-01 -1.13040563e-02 1.25735343e-01 4.33745235e-01
8.60887587e-01 1.37986109e-01 -1.00327325e+00 5.49133718e-01
-9.20180082e-01 -6.49549365e-01 4.39770460e-01 7.57480741e-01
-8.92914757e-02 -1.40056157e+00 -4.61329566e-03 7.56473914e-02
-5.31802118e-01 -1.93407893e-01 -9.12547350e-01 1.08863509e+00
-3.48007586e-03 6.63175881e-01 2.83604473e-01 -4.49866354e-01
2.55086333e-01 4.15162534e-01 7.67733395e-01 -6.79530978e-01
-4.91325915e-01 -3.79449576e-01 7.04317510e-01 -5.09879529e-01
-4.79595721e-01 -6.53740585e-01 -9.87911105e-01 -3.03586721e-01
-1.22964278e-01 1.79282725e-01 5.73916912e-01 8.82492840e-01
-9.93944257e-02 3.10369194e-01 3.01043361e-01 -6.30485415e-01
-5.18718958e-01 -1.05355275e+00 -3.34701061e-01 6.50144994e-01
4.40273024e-02 -9.62335646e-01 -3.89475316e-01 -5.63643336e-01] | [10.70042610168457, 9.384377479553223] |
fc1d60b0-5ae9-4109-a4d6-0ed1ed4fa751 | community-question-answering-entity-linking | 2205.11917 | null | https://arxiv.org/abs/2205.11917v1 | https://arxiv.org/pdf/2205.11917v1.pdf | Community Question Answering Entity Linking via Leveraging Auxiliary Data | Community Question Answering (CQA) platforms contain plenty of CQA texts (i.e., questions and answers corresponding to the question) where named entities appear ubiquitously. In this paper, we define a new task of CQA entity linking (CQAEL) as linking the textual entity mentions detected from CQA texts with their corresponding entities in a knowledge base. This task can facilitate many downstream applications including expert finding and knowledge base enrichment. Traditional entity linking methods mainly focus on linking entities in news documents, and are suboptimal over this new task of CQAEL since they cannot effectively leverage various informative auxiliary data involved in the CQA platform to aid entity linking, such as parallel answers and two types of meta-data (i.e., topic tags and users). To remedy this crucial issue, we propose a novel transformer-based framework to effectively harness the knowledge delivered by different kinds of auxiliary data to promote the linking performance. We validate the superiority of our framework through extensive experiments over a newly released CQAEL data set against state-of-the-art entity linking methods. | ['Yadong Wang', 'Jianbo Gao', 'Wei Shen', 'Yuhan Li'] | 2022-05-24 | null | null | null | null | ['community-question-answering', 'community-question-answering'] | ['miscellaneous', 'natural-language-processing'] | [-5.37149608e-01 1.76230073e-01 -2.70928480e-02 -1.89410131e-02
-1.21320105e+00 -8.26846659e-01 7.31174588e-01 8.47287953e-01
-5.73329449e-01 8.10463667e-01 6.43202126e-01 -1.58698574e-01
-3.40407014e-01 -1.08089435e+00 -5.97473562e-01 -2.29373917e-01
3.42467993e-01 6.16392195e-01 8.90268326e-01 -5.28080404e-01
1.00330152e-01 -1.80441707e-01 -1.15729439e+00 3.57994556e-01
1.52576637e+00 6.71371937e-01 -4.59691137e-02 1.05490282e-01
-6.95369542e-01 1.00129163e+00 -7.15803325e-01 -9.49765146e-01
-3.28790098e-01 -1.81855693e-01 -1.15016913e+00 -5.58499515e-01
2.11123735e-01 6.17384799e-02 -4.09003496e-01 9.11823869e-01
6.56937361e-01 1.04287975e-02 3.89288038e-01 -1.37728119e+00
-6.40666962e-01 9.56510127e-01 -3.08825642e-01 3.82183552e-01
6.07201278e-01 -3.25732768e-01 1.48980641e+00 -8.25331211e-01
9.18939710e-01 1.21776676e+00 9.13580418e-01 1.92614779e-01
-7.04452813e-01 -5.91543972e-01 -2.97807995e-03 3.95243526e-01
-1.23641598e+00 -2.99172908e-01 4.36684668e-01 -4.02190596e-01
7.69296288e-01 4.05720830e-01 1.70134813e-01 6.82614207e-01
-5.84407389e-01 8.36274624e-01 7.43612587e-01 -3.82506698e-01
-2.57393178e-02 1.72942758e-01 5.60188949e-01 4.96471137e-01
4.10234660e-01 -7.49150753e-01 -5.49935997e-01 -6.21825874e-01
4.66879681e-02 -2.26591125e-01 -3.83866578e-01 -7.08789229e-02
-1.25432658e+00 8.62530351e-01 3.51527333e-01 3.74588668e-01
-4.59419966e-01 -1.25392646e-01 5.82890153e-01 -3.48867998e-02
2.18146145e-01 8.55040729e-01 -7.07687795e-01 2.30795473e-01
-4.70873833e-01 6.24027669e-01 1.09253871e+00 1.33128059e+00
7.55572081e-01 -9.98402596e-01 -8.84578645e-01 8.51811826e-01
5.66580296e-01 4.21901822e-01 2.84005821e-01 -8.68225932e-01
8.60561609e-01 1.10579312e+00 5.51695585e-01 -1.33435571e+00
-4.08409387e-01 -5.48531771e-01 -4.15631175e-01 -9.34171498e-01
5.83294213e-01 -2.61105388e-01 -1.93124354e-01 1.80386710e+00
9.57250118e-01 1.87596411e-01 2.60722637e-01 6.29471600e-01
1.48551357e+00 6.72212124e-01 3.67991656e-01 -2.18665183e-01
1.99351180e+00 -1.02347755e+00 -1.13170016e+00 1.57941937e-01
9.62820530e-01 -9.84290242e-01 6.94597125e-01 -4.04183179e-01
-7.26003945e-01 -1.77213326e-01 -5.66529751e-01 -4.10320133e-01
-6.42152667e-01 -1.56576429e-02 3.47457319e-01 1.60014734e-01
-4.99874413e-01 1.02873906e-01 -2.66126662e-01 -3.88070643e-01
2.06372842e-01 -1.65002868e-01 -1.97100028e-01 -1.47069469e-01
-2.01716113e+00 7.13578284e-01 5.92626393e-01 2.13816874e-02
-2.57111609e-01 -1.04740214e+00 -6.38254583e-01 2.01805338e-01
9.60908890e-01 -1.05950165e+00 1.38463116e+00 -5.95022477e-02
-7.30414093e-01 6.44638240e-01 -1.43571943e-01 -2.35375583e-01
2.56795585e-01 -5.02594829e-01 -7.75395989e-01 1.77580237e-01
7.93288827e-01 2.43561953e-01 1.68697610e-01 -9.92143452e-01
-1.03770959e+00 -3.08665425e-01 4.76450145e-01 1.37980729e-02
-5.41348815e-01 1.87041685e-01 -8.69937181e-01 -6.04465246e-01
-1.12700917e-01 -5.45118749e-01 7.13528246e-02 -4.31869566e-01
-6.77830279e-01 -8.98632050e-01 7.77361214e-01 -9.44305003e-01
1.86856675e+00 -1.76291764e+00 -1.60236761e-01 2.20545843e-01
6.11332119e-01 3.13246667e-01 -4.27441671e-03 1.01573360e+00
3.88040721e-01 2.32293516e-01 3.29051726e-02 2.05661748e-02
3.41955602e-01 7.43017793e-02 -4.64479268e-01 -1.14137016e-01
1.30682290e-01 1.02590430e+00 -1.43924069e+00 -1.04068017e+00
-7.06510544e-01 1.03844717e-01 -4.74816203e-01 2.38439023e-01
-5.85372746e-01 4.43684638e-01 -1.17743504e+00 6.41837597e-01
4.34899092e-01 -5.95303059e-01 1.32013172e-01 -5.82828224e-01
-7.25055039e-02 5.75735629e-01 -1.07245672e+00 1.51603425e+00
-3.12848389e-01 1.93397775e-01 2.33416613e-02 -5.19385576e-01
6.86520994e-01 5.94794095e-01 4.54651296e-01 -6.22283816e-01
-1.42569482e-01 3.80132407e-01 -2.08556578e-01 -9.21035945e-01
7.07515001e-01 2.89123803e-01 -2.32147738e-01 3.54264319e-01
8.03157687e-02 4.50035125e-01 5.17832875e-01 8.47634733e-01
1.36559153e+00 -1.33052468e-01 3.01370680e-01 -2.73325592e-01
9.36568201e-01 3.92885655e-01 7.70465672e-01 7.92355359e-01
-9.27925631e-02 -2.35666037e-02 5.20541012e-01 1.76230207e-01
-9.76027131e-01 -5.94475150e-01 -1.21744320e-01 1.20981967e+00
1.73630208e-01 -8.91183972e-01 -6.32495522e-01 -8.53217065e-01
1.56857908e-01 5.40708303e-01 -5.53308427e-01 1.09196402e-01
-6.43657744e-01 -6.57908380e-01 7.43321538e-01 4.22925085e-01
5.61030447e-01 -8.72806907e-01 -1.66804697e-02 5.25718987e-01
-9.37447548e-01 -1.39927435e+00 -5.40015459e-01 -4.12734628e-01
-4.17479545e-01 -1.41303623e+00 -5.88027477e-01 -6.80326045e-01
2.67028064e-01 2.57697374e-01 1.57855928e+00 2.82535791e-01
1.46715507e-01 5.19286752e-01 -7.59962678e-01 -1.14639975e-01
-2.97770470e-01 4.73432481e-01 -2.79494554e-01 -1.38240859e-01
5.20503104e-01 -4.40214127e-01 -7.77737558e-01 4.06569183e-01
-8.98028135e-01 -1.18326917e-01 3.77250463e-01 7.25398660e-01
3.67811531e-01 -6.22712262e-02 1.12296593e+00 -1.36035848e+00
9.51348901e-01 -1.04534638e+00 -4.32129383e-01 8.31052065e-01
-4.41016495e-01 1.27393872e-01 5.06432593e-01 -2.16099873e-01
-1.26191866e+00 -4.71643925e-01 -1.83122113e-01 2.93086946e-01
2.39029139e-01 1.16921568e+00 -4.49107826e-01 3.84339899e-01
5.43218017e-01 -2.36722887e-01 -7.39170253e-01 -7.29026854e-01
9.49091017e-01 7.12121189e-01 8.08399618e-01 -8.29389632e-01
9.72283959e-01 2.31034324e-01 -2.35569760e-01 -1.86781019e-01
-1.38612127e+00 -1.23551810e+00 -4.25718963e-01 -8.89523700e-02
8.20890665e-01 -9.84112263e-01 -8.29357743e-01 1.25000417e-01
-1.34443998e+00 4.66978431e-01 -1.82598472e-01 2.61149645e-01
1.66703556e-02 4.85978544e-01 -5.52889168e-01 -5.65487087e-01
-5.96948564e-01 -5.87146938e-01 8.61155629e-01 5.03244758e-01
-5.18926643e-02 -9.74916160e-01 5.52077413e-01 7.14331210e-01
1.99559689e-01 1.34636894e-01 1.17438006e+00 -1.12467098e+00
-6.23078227e-01 -2.69086331e-01 -3.82419705e-01 -3.87017101e-01
-3.62345111e-03 -1.20134242e-01 -7.47001171e-01 -1.18122794e-01
-5.32933891e-01 -3.91354620e-01 6.81257308e-01 -3.35102826e-01
5.69954515e-01 -4.78368074e-01 -6.12368643e-01 -1.11152552e-01
1.11860693e+00 -2.22054765e-01 3.59082848e-01 6.54848218e-01
8.08501065e-01 8.46660793e-01 8.82979512e-01 4.17461216e-01
9.82454002e-01 8.68861258e-01 2.65305519e-01 1.61469683e-01
-6.05402738e-02 -6.36292875e-01 -9.41970944e-02 1.37744522e+00
6.25256225e-02 -4.10870850e-01 -1.00184834e+00 7.47023404e-01
-2.15300679e+00 -1.05867231e+00 -6.13129556e-01 1.78699064e+00
1.40677106e+00 -2.51660079e-01 1.76212728e-01 -1.04797259e-01
9.27278817e-01 -1.33407295e-01 -3.75952363e-01 6.53466940e-01
-1.88779488e-01 -2.06825450e-01 1.72495782e-01 2.89015621e-01
-1.11979532e+00 6.68592274e-01 4.81636000e+00 1.16922116e+00
-1.23769060e-01 4.95047510e-01 -1.02824956e-01 5.65586329e-01
-7.30817676e-01 1.84447572e-01 -1.21580994e+00 6.33396864e-01
8.20455968e-01 -5.79175055e-01 -1.89988568e-01 6.35428190e-01
-1.74525574e-01 1.62928551e-01 -8.99109721e-01 4.78645772e-01
-2.52515137e-01 -1.45152473e+00 1.37074720e-02 -2.67970949e-01
7.39635229e-01 -1.92018703e-01 -3.92425328e-01 7.83283472e-01
5.59928477e-01 -3.75317663e-01 5.96083343e-01 6.42657638e-01
3.44351590e-01 -3.88656348e-01 1.10051394e+00 4.26490754e-01
-1.39567256e+00 -5.10020070e-02 -1.91752225e-01 6.07738912e-01
2.33157665e-01 7.44247615e-01 -6.49452329e-01 1.14341569e+00
8.81719172e-01 5.49116552e-01 -7.80600429e-01 1.35598242e+00
-4.67291236e-01 6.74335241e-01 -2.52988845e-01 -8.28961730e-02
3.38739604e-01 -1.55767351e-02 5.29541790e-01 1.28918862e+00
2.79808730e-01 3.80882710e-01 2.81450003e-02 7.72605598e-01
-6.82787061e-01 6.57281816e-01 -1.53056905e-01 -1.75799057e-02
1.20682478e+00 1.41954505e+00 -4.62262183e-01 -4.79890585e-01
-7.99367845e-01 5.04669905e-01 5.91710627e-01 3.01092505e-01
-7.01810658e-01 -6.35136962e-01 1.97442994e-01 6.12554848e-02
3.59565794e-01 2.49227151e-01 4.92200673e-01 -1.29321253e+00
3.01499277e-01 -6.89632237e-01 9.51287329e-01 -6.43216491e-01
-1.77561045e+00 2.49863639e-01 6.61577471e-03 -1.12357640e+00
-1.43213689e-01 1.16952091e-01 -5.01739621e-01 7.95852959e-01
-1.73757005e+00 -1.27382791e+00 -3.93333077e-01 6.87972963e-01
-1.38573572e-01 1.53129816e-01 6.30621672e-01 9.80755568e-01
-6.71633780e-01 6.71816170e-01 4.05178428e-01 5.51235557e-01
1.08455086e+00 -1.37548137e+00 2.03148991e-01 8.05445194e-01
5.61657958e-02 9.90959764e-01 6.30204976e-01 -7.96265066e-01
-1.26035714e+00 -1.23583186e+00 1.29512405e+00 -8.08904886e-01
9.85669434e-01 -1.99336961e-01 -1.25817907e+00 5.96101224e-01
2.80250937e-01 -1.42304152e-01 8.89938891e-01 5.40578783e-01
-5.65070510e-01 -6.26027435e-02 -8.39257061e-01 3.10738951e-01
7.90860057e-01 -7.88342595e-01 -1.05128944e+00 4.59272951e-01
1.23660266e+00 -3.25735867e-01 -1.13142729e+00 5.16823173e-01
1.36762559e-01 -3.82478178e-01 1.07867527e+00 -9.35207188e-01
2.14105889e-01 -8.01188469e-01 -3.30795534e-04 -9.81125891e-01
-1.41153589e-01 -4.48307633e-01 -5.20126224e-01 2.21483469e+00
5.64853489e-01 -5.74448586e-01 3.39462548e-01 4.97453541e-01
-3.15104514e-01 -5.10749638e-01 -9.00806427e-01 -5.96917868e-01
-1.63744554e-01 3.70958000e-02 8.49321306e-01 1.39051545e+00
2.27052554e-01 6.92038357e-01 5.59046492e-02 3.85612249e-01
3.80450010e-01 4.50756878e-01 6.68884993e-01 -1.50035775e+00
1.87887065e-02 -2.65015483e-01 1.01977833e-01 -8.95856261e-01
2.33522691e-02 -1.04889441e+00 -5.57732880e-02 -1.71668887e+00
5.75937748e-01 -8.66856456e-01 -2.69049704e-01 4.02324915e-01
-8.54901373e-01 -1.79502293e-01 -5.13070896e-02 5.78116179e-01
-1.29451346e+00 6.80725396e-01 1.10739148e+00 -4.14309278e-02
3.78728323e-02 -2.37180501e-01 -8.84674668e-01 6.48367047e-01
3.64558578e-01 -7.78094172e-01 -2.71122336e-01 -3.16047370e-01
8.99591446e-01 2.72348672e-01 2.80078471e-01 -5.28597534e-01
7.79982507e-01 1.10065840e-01 -3.10537100e-01 -7.96024263e-01
-2.76676476e-01 -6.92563653e-01 -3.71382125e-02 5.83706684e-02
-3.78813952e-01 5.36121279e-02 -8.30948651e-02 8.87398005e-01
-5.33657432e-01 -4.65374500e-01 1.22279562e-01 -6.51666597e-02
-7.05230176e-01 2.63762832e-01 -2.44540516e-02 9.36122954e-01
6.82329118e-01 5.71960926e-01 -1.15889823e+00 -2.85946369e-01
-5.32453656e-01 8.68125319e-01 1.23464577e-01 3.34427327e-01
-4.35463618e-03 -1.58109295e+00 -8.58741224e-01 -7.20688224e-01
5.34577906e-01 6.77939579e-02 3.88218999e-01 1.00497448e+00
-9.46926847e-02 7.23292470e-01 2.30116129e-01 -1.93702668e-01
-1.00767994e+00 6.27680600e-01 7.07804412e-02 -7.95347035e-01
-4.08190548e-01 6.73169017e-01 1.59516521e-02 -6.47787869e-01
7.09264353e-02 4.95151319e-02 -8.57797682e-01 4.08208579e-01
5.92149615e-01 4.41223413e-01 1.32483214e-01 -5.19655824e-01
-4.28837985e-01 2.16580838e-01 -1.08345188e-01 -2.29007308e-03
1.08636129e+00 -4.52363789e-01 -4.53579932e-01 7.25000203e-02
8.76540244e-01 5.62193215e-01 -4.64849412e-01 -8.78360271e-01
5.49533963e-01 -1.16205759e-01 -2.06738740e-01 -8.95458102e-01
-8.22592735e-01 6.97879434e-01 5.17656505e-02 5.11330366e-01
8.56968284e-01 4.84022290e-01 1.06294608e+00 6.47466183e-01
4.46219593e-02 -1.07318735e+00 -3.88743803e-02 4.58095491e-01
6.86333001e-01 -1.09483743e+00 -1.10299125e-01 -7.20982313e-01
-2.94407070e-01 7.23737895e-01 6.25900805e-01 6.42145813e-01
4.02103812e-01 -2.94897228e-01 1.18617266e-01 -6.45091891e-01
-5.53303123e-01 -6.10371709e-01 4.33699310e-01 3.58444899e-01
3.65602374e-01 -2.59322762e-01 -6.21326268e-01 1.19872713e+00
4.99941260e-02 -2.21589997e-01 2.22547293e-01 9.20487702e-01
-3.91188413e-01 -1.08901691e+00 -2.80104965e-01 3.86998594e-01
-6.88653290e-01 -3.48623067e-01 -6.94098324e-02 8.37165058e-01
1.85489297e-01 1.14604247e+00 -3.18170995e-01 -1.14624515e-01
5.23437440e-01 1.62345886e-01 -1.61582708e-01 -5.19553661e-01
-8.64250243e-01 -1.53892934e-01 4.75326687e-01 -3.71559352e-01
-8.88597071e-01 -7.46996999e-01 -1.22145808e+00 -2.51313418e-01
-6.53247952e-01 8.01800907e-01 1.48448542e-01 1.11189592e+00
5.34422219e-01 5.12656569e-01 4.84977365e-01 4.47075188e-01
-2.95351684e-01 -1.03600931e+00 -3.88589740e-01 6.05009735e-01
-2.66511012e-02 -6.73025966e-01 3.01353224e-02 -2.81542968e-02] | [9.568812370300293, 8.710436820983887] |
60bfb095-b3c9-4c0c-a70c-b85509164fc4 | evaluating-graph-signal-processing-for | 1703.01842 | null | http://arxiv.org/abs/1703.01842v3 | http://arxiv.org/pdf/1703.01842v3.pdf | Evaluating Graph Signal Processing for Neuroimaging Through Classification and Dimensionality Reduction | Graph Signal Processing (GSP) is a promising framework to analyze
multi-dimensional neuroimaging datasets, while taking into account both the
spatial and functional dependencies between brain signals. In the present work,
we apply dimensionality reduction techniques based on graph representations of
the brain to decode brain activity from real and simulated fMRI datasets. We
introduce seven graphs obtained from a) geometric structure and/or b)
functional connectivity between brain areas at rest, and compare them when
performing dimension reduction for classification. We show that mixed graphs
using both a) and b) offer the best performance. We also show that graph
sampling methods perform better than classical dimension reduction including
Principal Component Analysis (PCA) and Independent Component Analysis (ICA). | ['Vincent Gripon', 'Mathilde Ménoret', 'Nicolas Farrugia', 'Bastien Pasdeloup'] | 2017-03-06 | null | null | null | null | ['graph-sampling'] | ['graphs'] | [ 3.50769520e-01 -4.16985787e-02 2.41801813e-01 -3.23931038e-01
1.16183020e-01 -6.51801705e-01 6.40289068e-01 1.94040984e-01
-3.26440454e-01 4.85515505e-01 4.36716527e-01 -1.11517482e-01
-7.72544026e-01 -8.47342432e-01 -2.97623008e-01 -7.18663216e-01
-8.28132868e-01 4.30570483e-01 -6.79936409e-02 -2.69777961e-02
8.47078040e-02 9.39672172e-01 -1.08857453e+00 -3.20484303e-02
8.44856620e-01 5.50129175e-01 -5.53113222e-02 5.67181647e-01
4.32521058e-03 3.06789398e-01 -2.63000399e-01 -1.08445704e-01
2.69982964e-01 -7.12305665e-01 -5.34523487e-01 2.51535773e-02
8.37991834e-02 2.08445296e-01 -6.52526557e-01 1.32293522e+00
5.31393588e-01 3.40267956e-01 1.00747228e+00 -1.08748567e+00
-6.69863105e-01 6.65713608e-01 -8.64636838e-01 8.39263618e-01
2.62122840e-01 3.77457254e-02 6.77546322e-01 -6.28404737e-01
9.17076051e-01 1.22133946e+00 2.83400118e-01 2.78706282e-01
-1.77095366e+00 -5.21725833e-01 3.80718149e-02 5.50941765e-01
-1.34130466e+00 -3.97361815e-01 1.27804542e+00 -6.94544256e-01
9.80343878e-01 3.37272853e-01 1.24522865e+00 1.05055726e+00
7.74277627e-01 1.56256005e-01 1.36942589e+00 -2.70157158e-01
4.64851767e-01 -5.82774043e-01 9.07071948e-01 6.38425708e-01
6.76537454e-01 -1.78184193e-02 -2.47359306e-01 -3.78591508e-01
5.17289162e-01 -1.28535271e-01 -3.74397278e-01 -7.87271380e-01
-1.38181663e+00 8.84071052e-01 4.84081626e-01 6.38892591e-01
-9.14787471e-01 -8.91263857e-02 4.50618923e-01 2.31677607e-01
3.93600613e-01 4.33606684e-01 1.06684454e-01 3.86660218e-01
-8.53576243e-01 -6.50794134e-02 5.40785849e-01 4.28399295e-01
5.54588020e-01 3.34333420e-01 -1.01980023e-01 6.59619570e-01
3.17133069e-01 3.57116729e-01 7.23568559e-01 -6.33391559e-01
2.52041161e-01 6.21317565e-01 -7.59995699e-01 -1.40832865e+00
-1.14498842e+00 -1.52483672e-01 -1.27133238e+00 -1.49537832e-01
1.74302667e-01 -2.23605543e-01 -6.83760703e-01 1.58013737e+00
2.20711827e-02 6.01617545e-02 -2.39968345e-01 8.31157506e-01
7.55239069e-01 6.84765354e-02 8.78247768e-02 -5.96629024e-01
1.53512979e+00 -2.87521213e-01 -9.65897143e-01 -2.76811630e-01
5.25279164e-01 -2.79350393e-02 4.51181263e-01 4.90825236e-01
-9.62836206e-01 -3.32598805e-01 -1.00329328e+00 2.01878160e-01
-3.28990161e-01 -1.41711503e-01 8.63654494e-01 8.73002172e-01
-1.36702895e+00 5.32473624e-01 -1.13789785e+00 -2.22870603e-01
5.79417229e-01 3.76604825e-01 -9.26660299e-01 -2.94025600e-01
-9.95670080e-01 6.85192049e-01 1.52603418e-01 6.26394600e-02
-6.28995776e-01 -4.44577575e-01 -9.05197322e-01 3.24200653e-02
-4.98049930e-02 -6.64020181e-01 7.82634541e-02 -5.69946229e-01
-1.13563335e+00 7.03955770e-01 -1.67512268e-01 -5.77991486e-01
1.21738978e-01 3.49487931e-01 -4.64718044e-01 7.01389909e-01
-1.57247394e-01 4.04013485e-01 9.43320096e-01 -6.71010852e-01
7.11614132e-01 -1.20366549e+00 -5.46698928e-01 4.81909588e-02
-3.51554424e-01 -1.28860161e-01 2.79906243e-01 -5.01245379e-01
7.30330408e-01 -7.33859539e-01 -3.14225137e-01 -3.95852089e-01
-5.98675966e-01 1.30050004e-01 5.19607246e-01 -9.67496812e-01
1.10454583e+00 -1.95408928e+00 7.36154616e-01 7.86160946e-01
1.01938665e+00 -5.98662086e-02 -3.20448518e-01 3.72302234e-01
-7.40315139e-01 5.68550527e-02 -3.83987159e-01 1.73462272e-01
-3.29796344e-01 7.13333935e-02 4.13742512e-01 1.09840429e+00
-1.70903374e-02 1.19829154e+00 -7.41951048e-01 -2.75772035e-01
2.88478076e-01 4.75726187e-01 -6.23236597e-01 -2.13854343e-01
4.49686587e-01 9.19711709e-01 -2.43614703e-01 9.41320360e-02
7.79943347e-01 1.47010824e-02 7.08960176e-01 -4.21319216e-01
2.03437150e-01 -5.19400686e-02 -9.24599528e-01 1.77232718e+00
6.02226816e-02 7.79450893e-01 2.32005835e-01 -1.55856061e+00
1.07809246e+00 1.78164482e-01 8.62602174e-01 -8.00199807e-01
3.62415612e-01 -3.25247258e-01 9.77549136e-01 -6.20800257e-01
-3.00959826e-01 -6.52287304e-02 2.74817765e-01 5.36565542e-01
4.79537100e-01 1.11539267e-01 5.43346822e-01 4.59127128e-01
1.57901824e+00 -4.53967482e-01 5.08022308e-01 -8.66054296e-01
5.99228740e-01 -3.55143160e-01 1.26929343e-01 3.07328463e-01
-4.56573635e-01 3.22330475e-01 1.14239323e+00 -2.68851578e-01
-6.06482983e-01 -1.03389990e+00 -7.60191083e-02 4.08634365e-01
-2.89239973e-01 -3.50858331e-01 -9.15978372e-01 -4.15296674e-01
-7.73585886e-02 6.19168520e-01 -8.49339247e-01 -3.48533422e-01
-4.34310764e-01 -1.28794575e+00 2.92254359e-01 2.21324578e-01
2.44576171e-01 -6.63718700e-01 -4.35260475e-01 3.09660789e-02
1.52621269e-01 -9.30106282e-01 -1.16767615e-01 1.89753011e-01
-1.23274004e+00 -1.47313201e+00 -6.32059574e-01 -3.40754360e-01
8.47290337e-01 4.11279976e-01 7.92018592e-01 -2.24390537e-01
-6.03481114e-01 7.56320298e-01 -2.41249189e-01 3.00119445e-03
-1.50429994e-01 -2.90939391e-01 2.64552832e-01 2.28262171e-01
2.87251621e-01 -1.07374716e+00 -6.71959817e-01 6.44387156e-02
-9.00560439e-01 -9.90137383e-02 4.27895635e-01 4.50492829e-01
6.14518523e-01 1.67939499e-01 4.20202315e-01 -8.75551462e-01
1.24675190e+00 -5.89746952e-01 -3.95943791e-01 9.26064551e-02
-5.76480865e-01 6.70428723e-02 4.90334004e-01 -2.37824440e-01
-6.17939353e-01 2.12762102e-01 2.57864177e-01 -5.30462384e-01
-3.11014324e-01 5.37066460e-01 -2.98354030e-01 -3.64077568e-01
1.09563565e+00 1.54960230e-01 4.45645332e-01 -4.47449118e-01
5.40767133e-01 1.29078940e-01 1.51734695e-01 -5.36620505e-02
3.41893137e-01 6.75600290e-01 5.52593827e-01 -1.14524269e+00
1.72940806e-01 -1.90530822e-01 -1.28264511e+00 -4.94860172e-01
9.11981702e-01 -4.62872505e-01 -5.69889247e-01 -9.35976859e-03
-8.55336428e-01 9.44107026e-02 -1.03144772e-01 1.03660023e+00
-3.91456455e-01 8.35326374e-01 -4.98666853e-01 -6.15664840e-01
-3.74949813e-01 -1.08685219e+00 4.83779758e-01 -3.20479542e-01
-2.38483071e-01 -1.14020514e+00 3.62601399e-01 5.85284531e-02
3.58045965e-01 6.18571937e-01 1.24736619e+00 -8.01925659e-01
1.48336023e-01 -3.09663802e-01 -1.52848288e-01 -3.06076196e-04
1.91875488e-01 -3.40944290e-01 -6.73225701e-01 -2.73511350e-01
3.70496154e-01 5.72691560e-01 8.26062202e-01 8.08092356e-01
9.98362899e-01 -1.87062606e-01 -4.18367118e-01 2.95697868e-01
1.20674706e+00 1.35654658e-01 6.82616889e-01 -3.63272667e-01
8.25580001e-01 1.00679958e+00 -3.81233692e-01 9.10098329e-02
-1.18437530e-02 3.75330955e-01 1.23905346e-01 3.23856205e-01
-3.88298571e-01 2.52189904e-01 1.14527792e-01 1.18029940e+00
-2.74378836e-01 9.66150034e-03 -1.24863100e+00 2.02609718e-01
-1.56199622e+00 -9.24682200e-01 -8.91515434e-01 2.07664680e+00
-5.00281975e-02 -4.70922999e-02 3.44130248e-01 2.87507027e-01
8.86501670e-01 1.72607407e-01 -4.74924564e-01 -2.30517074e-01
-2.55510688e-01 4.21071351e-01 4.26291406e-01 3.70107800e-01
-7.38432288e-01 3.83359700e-01 7.52818346e+00 2.24021152e-01
-8.41846228e-01 2.43482441e-01 4.62591618e-01 -2.06906691e-01
-2.28914902e-01 -2.17436448e-01 2.08049148e-01 2.41088599e-01
1.02006197e+00 -3.47389340e-01 8.13727498e-01 4.94997382e-01
3.13318074e-01 -1.66127667e-01 -9.26743984e-01 1.18237972e+00
2.72987366e-01 -1.03389537e+00 1.21324107e-01 3.52766216e-01
4.15983260e-01 1.24912426e-01 -2.89414436e-01 -2.04214141e-01
-2.01188967e-01 -8.76915812e-01 1.83341473e-01 7.74777353e-01
5.01734734e-01 -5.47392547e-01 2.47960389e-01 2.79893905e-01
-1.15143549e+00 5.78590818e-02 -3.70074779e-01 -1.78704143e-01
1.61216676e-01 9.27850485e-01 -4.07764733e-01 6.52793705e-01
2.18437210e-01 8.50556731e-01 -8.16181600e-01 1.05046177e+00
-2.83021748e-01 5.15805066e-01 -5.27309664e-02 -3.45141925e-02
-2.23505497e-01 -9.00007725e-01 9.04523790e-01 9.91497934e-01
8.69736075e-02 2.62179703e-01 -3.57530117e-01 1.09611857e+00
2.74122059e-01 3.28211248e-01 -1.00843501e+00 -3.89375329e-01
-6.32470548e-02 1.47933829e+00 -1.57370317e+00 -1.88263804e-01
-3.96439999e-01 7.53395915e-01 3.62543404e-01 4.06786770e-01
-3.74939680e-01 -2.38646835e-01 4.85674500e-01 1.50464743e-01
-1.51941448e-01 -8.32564235e-01 -2.71992862e-01 -1.26322412e+00
-1.07331783e-01 -4.07255143e-01 4.50659245e-01 -5.19860685e-01
-1.36680126e+00 7.35120595e-01 4.39033568e-01 -7.70896375e-01
-1.94455385e-01 -6.46266580e-01 -5.20119786e-01 8.81527185e-01
-1.08107650e+00 -4.74648774e-01 -2.33162194e-01 8.13147068e-01
-7.84648731e-02 -2.38293529e-01 9.36360002e-01 2.17363343e-01
-6.95101738e-01 -2.21703678e-01 -1.23277031e-01 9.78775173e-02
-1.76382121e-02 -1.08244038e+00 4.43144560e-01 9.43654895e-01
4.58100945e-01 6.56999528e-01 3.39263916e-01 -8.98910105e-01
-1.93780994e+00 -6.93351626e-01 3.09568495e-01 -1.10760652e-01
7.91361153e-01 -6.10208392e-01 -9.27331567e-01 5.10892928e-01
8.81830007e-02 1.85256258e-01 9.06118691e-01 -7.70663470e-02
-1.36731386e-01 2.15341941e-01 -1.30735052e+00 3.85426342e-01
1.39847195e+00 -3.22381973e-01 -6.85008168e-01 5.43108463e-01
2.08366960e-01 2.07694322e-01 -1.01863396e+00 1.46374822e-01
3.37760299e-01 -1.07555032e+00 1.10935330e+00 -7.54643321e-01
-1.14368290e-01 6.40454292e-02 1.06045328e-01 -1.63382709e+00
-9.17076826e-01 -3.69567156e-01 -2.40373626e-01 5.30553639e-01
4.51699877e-03 -9.09427941e-01 6.15926981e-01 6.29852831e-01
1.37334377e-01 -2.07155198e-01 -1.09170842e+00 -8.23260605e-01
7.42836222e-02 -4.47787672e-01 3.00760090e-01 1.10338211e+00
4.79345053e-01 6.19858682e-01 1.60405532e-01 -1.46192476e-01
8.09145451e-01 -7.82939270e-02 1.35030925e-01 -1.64417720e+00
9.13591962e-03 -7.12103546e-01 -1.10719943e+00 -1.09054059e-01
4.62906688e-01 -1.49111640e+00 -7.32463002e-01 -1.70852292e+00
1.13065250e-01 2.44476691e-01 -1.03957228e-01 2.10651264e-01
2.65540391e-01 2.15595737e-01 -1.95328668e-02 1.51818752e-01
-2.11128354e-01 4.23166066e-01 1.25609839e+00 -5.26634715e-02
-3.48592907e-01 -3.68622720e-01 -5.02396703e-01 3.87989491e-01
8.42664540e-01 -3.16073596e-01 -5.68372607e-01 -1.19892605e-01
1.03765942e-01 2.79098451e-01 3.10063362e-01 -1.04339302e+00
7.62677565e-02 3.59421313e-01 7.53089190e-01 -1.70293897e-02
3.07322562e-01 -7.59992003e-01 4.18823093e-01 7.29515553e-01
-1.63572758e-01 1.58192188e-01 2.34377220e-01 7.67001748e-01
5.06183095e-02 -4.65711765e-02 6.96867347e-01 -2.17459530e-01
-4.29000348e-01 2.05681846e-01 -6.87179148e-01 -2.19344333e-01
1.04495680e+00 -7.40561709e-02 -2.88852513e-01 -2.43828818e-01
-1.40386438e+00 -9.16450694e-02 -5.68834282e-02 1.75552234e-01
8.42065752e-01 -1.37352324e+00 -5.99631906e-01 4.86918360e-01
-1.97681963e-01 -8.06158066e-01 6.45415306e-01 1.46879482e+00
-5.14116406e-01 4.66874182e-01 -9.39408839e-01 -4.90209579e-01
-1.24276614e+00 9.82468963e-01 2.41894096e-01 -7.08382279e-02
-9.11755741e-01 4.39761370e-01 1.43384859e-01 -1.30894229e-01
-3.20994139e-01 -1.65757164e-01 -7.38858521e-01 5.95996201e-01
5.96004426e-01 7.53004014e-01 3.37794781e-01 -1.02073014e+00
-5.06239116e-01 3.48991096e-01 4.12597805e-01 -1.84085280e-01
1.53827786e+00 -7.39452988e-02 -5.42587817e-01 3.03040802e-01
1.22980189e+00 -2.60757834e-01 -5.51085830e-01 4.70675081e-02
7.81507045e-02 -3.26532662e-01 5.07823408e-01 -3.25767964e-01
-1.42035174e+00 1.02496862e+00 7.40790963e-01 7.83583820e-01
1.13336921e+00 -1.59598812e-01 2.08279248e-02 2.24486068e-01
4.32498157e-01 -7.30602264e-01 -2.42122948e-01 -5.37594557e-02
1.16059804e+00 -6.04410112e-01 3.02137822e-01 -4.85094666e-01
-5.14050961e-01 1.39070833e+00 3.31571400e-02 -7.63262391e-01
1.29313278e+00 -1.08774573e-01 -5.66965342e-01 -1.00446630e+00
-3.30345452e-01 -2.21343860e-01 7.31405497e-01 8.22079301e-01
3.47884893e-01 3.85222167e-01 -8.80381346e-01 6.36233807e-01
-1.49589434e-01 -4.45724636e-01 4.16439831e-01 6.42836928e-01
-1.07543044e-01 -8.51982832e-01 -4.77423042e-01 1.01396620e+00
-6.52781427e-02 -3.31723168e-02 -8.24596763e-01 5.97686887e-01
-4.48041052e-01 8.25424194e-01 2.78773494e-02 -3.45463485e-01
4.22230780e-01 3.65837395e-01 9.37067330e-01 -5.07190466e-01
-3.31992388e-01 2.11161166e-01 -4.36962098e-02 -8.36394608e-01
-5.49987733e-01 -9.24509704e-01 -1.12020206e+00 -3.40058580e-02
-2.49144256e-01 -4.87783700e-02 9.18074489e-01 6.00159705e-01
6.06246293e-01 7.76193619e-01 3.90598893e-01 -1.04605281e+00
2.23436520e-01 -1.11354768e+00 -1.26815617e+00 3.10612440e-01
-9.06312987e-02 -9.37812984e-01 -3.90189856e-01 -2.59043902e-01] | [12.402898788452148, 3.415306568145752] |
1d367439-2dea-47c5-8bed-2c03ce89043f | ala-adversarial-lightness-attack-via | 2201.06070 | null | https://arxiv.org/abs/2201.06070v1 | https://arxiv.org/pdf/2201.06070v1.pdf | ALA: Adversarial Lightness Attack via Naturalness-aware Regularizations | Most researchers have tried to enhance the robustness of deep neural networks (DNNs) by revealing and repairing the vulnerability of DNNs with specialized adversarial examples. Parts of the attack examples have imperceptible perturbations restricted by Lp norm. However, due to their high-frequency property, the adversarial examples usually have poor transferability and can be defensed by denoising methods. To avoid the defects, some works make the perturbations unrestricted to gain better robustness and transferability. However, these examples usually look unnatural and alert the guards. To generate unrestricted adversarial examples with high image quality and good transferability, in this paper, we propose Adversarial Lightness Attack (ALA), a white-box unrestricted adversarial attack that focuses on modifying the lightness of the images. The shape and color of the samples, which are crucial to human perception, are barely influenced. To obtain adversarial examples with high image quality, we craft a naturalness-aware regularization. To achieve stronger transferability, we propose random initialization and non-stop attack strategy in the attack procedure. We verify the effectiveness of ALA on two popular datasets for different tasks (i.e., ImageNet for image classification and Places-365 for scene recognition). The experiments show that the generated adversarial examples have both strong transferability and high image quality. Besides, the adversarial examples can also help to improve the standard trained ResNet50 on defending lightness corruption. | ['Geguang Pu', 'Yang Liu', 'Jincao Feng', 'JiaYi Zhu', 'Qing Guo', 'Yihao Huang', 'Felix Juefei-Xu', 'Liangru Sun'] | 2022-01-16 | null | null | null | null | ['scene-recognition'] | ['computer-vision'] | [ 2.03613728e-01 -3.29509974e-02 5.36566138e-01 -2.51247019e-01
-2.57576525e-01 -1.01818621e+00 4.90216583e-01 -5.42270541e-01
-3.85927886e-01 6.95778131e-01 -8.25473592e-02 -2.67010808e-01
1.72624648e-01 -9.90699589e-01 -1.16037989e+00 -1.07851911e+00
1.29744858e-01 -5.77942848e-01 1.77771360e-01 -4.86410230e-01
-9.35634598e-03 6.37391925e-01 -1.23110068e+00 2.20277458e-01
1.01629674e+00 8.44355166e-01 -6.42360523e-02 4.85290170e-01
2.48354226e-01 7.50811100e-01 -1.09095478e+00 -5.48079431e-01
6.34542763e-01 -3.76669943e-01 -4.26452279e-01 -2.00301439e-01
5.02917945e-01 -4.99209136e-01 -7.46582329e-01 1.78206170e+00
6.45362794e-01 4.55085225e-02 4.78905827e-01 -1.43971837e+00
-1.19546473e+00 4.38919693e-01 -4.42387372e-01 1.68719202e-01
1.19583309e-01 5.75716734e-01 3.54492605e-01 -4.21891451e-01
1.69713482e-01 1.47733021e+00 4.34784859e-01 1.02707314e+00
-1.01529598e+00 -1.25719178e+00 2.62297601e-01 2.67609209e-01
-1.25285661e+00 -2.81254709e-01 1.10215211e+00 -5.80985919e-02
1.40997902e-01 5.79725385e-01 2.68518925e-01 1.70268369e+00
1.34836152e-01 4.68920290e-01 1.30879533e+00 4.57191579e-02
5.74261136e-02 2.55171448e-01 -3.21134090e-01 4.58981931e-01
2.66367853e-01 4.85637546e-01 1.37531742e-01 5.44689596e-02
8.49121273e-01 1.59401670e-01 -7.34991074e-01 8.59131068e-02
-1.04503965e+00 6.13049090e-01 9.60657835e-01 1.78064965e-02
-9.08795372e-02 2.85774060e-02 5.16738057e-01 4.54704553e-01
1.20606095e-01 5.46417356e-01 -2.70209044e-01 3.76511514e-01
-1.37157410e-01 2.97764167e-02 4.43029225e-01 8.07952583e-01
5.61089039e-01 6.14728391e-01 -2.38233238e-01 7.00786173e-01
1.32767990e-01 9.55096781e-01 4.77844656e-01 -5.54288864e-01
5.50494015e-01 3.90473872e-01 -1.20266922e-01 -1.33620811e+00
-1.43256888e-01 -4.24438238e-01 -1.29737592e+00 7.33510733e-01
4.02242184e-01 -4.36277330e-01 -1.06803238e+00 2.02969265e+00
2.81183630e-01 3.76939952e-01 4.53908592e-01 9.78879213e-01
9.86916065e-01 8.86775017e-01 1.65849209e-01 1.51309157e-02
1.11212718e+00 -8.09293389e-01 -7.07926035e-01 -1.52894989e-01
2.37886697e-01 -8.01889837e-01 1.38354754e+00 4.93418545e-01
-6.82711482e-01 -8.44719827e-01 -1.33816671e+00 2.51901746e-01
-4.16714996e-01 -2.23149508e-01 2.57821023e-01 1.01851976e+00
-4.45320815e-01 6.42456651e-01 -5.11006355e-01 2.34210119e-01
6.29530966e-01 1.67732731e-01 -5.67395449e-01 -1.82723686e-01
-1.69752407e+00 7.09898770e-01 3.95070910e-01 2.39560828e-01
-1.26533234e+00 -7.51376927e-01 -7.07791567e-01 -8.61439183e-02
1.44124553e-01 -2.12044165e-01 5.19717336e-01 -1.40191114e+00
-1.63051832e+00 5.85911095e-01 7.25443184e-01 -3.27767253e-01
8.48603606e-01 -2.19434053e-01 -8.03485811e-01 1.92795396e-01
-4.60366547e-01 4.23398077e-01 1.35862601e+00 -1.38876987e+00
-7.54206926e-02 -1.17302693e-01 4.13122028e-01 -7.85773844e-02
-7.81375527e-01 -1.97826661e-02 1.19191043e-01 -1.19190192e+00
-2.41681367e-01 -8.96197677e-01 -1.76027983e-01 7.09856153e-02
-6.56338513e-01 4.41975087e-01 1.17399776e+00 -6.99711919e-01
6.49331033e-01 -2.55832314e+00 -1.36123642e-01 1.25234336e-01
2.35112514e-02 8.69435191e-01 -4.09230351e-01 -2.15843152e-02
-5.56881905e-01 4.30594683e-01 -3.12371910e-01 2.85057783e-01
-3.93641591e-02 4.31654960e-01 -7.70022750e-01 7.61924326e-01
2.20228255e-01 7.87594378e-01 -7.06383109e-01 -1.71466798e-01
2.79721320e-01 6.98548079e-01 -5.34206748e-01 4.25181895e-01
-4.50429656e-02 6.76645279e-01 -5.06054699e-01 2.53484130e-01
1.17356920e+00 5.21433592e-01 -5.32401323e-01 -4.76448894e-01
4.75737333e-01 -3.85452539e-01 -1.30897772e+00 9.11687315e-01
-4.42213863e-01 4.59107906e-01 1.94033995e-01 -1.03229761e+00
1.05530906e+00 2.74639487e-01 -1.19701564e-01 -5.99123895e-01
3.86812001e-01 -3.20956893e-02 2.02687651e-01 -6.19337142e-01
-3.94759513e-02 -1.56903341e-01 1.36011273e-01 6.88867867e-02
-3.32352519e-01 -1.29218116e-01 -5.03756285e-01 3.07089556e-02
9.88858044e-01 -1.20455556e-01 -1.90141812e-01 -2.68937618e-01
8.28609049e-01 -6.13446891e-01 6.91133559e-01 5.76202810e-01
-3.09822112e-01 6.87917829e-01 4.24151897e-01 -4.56527561e-01
-9.78219271e-01 -1.14632452e+00 -1.64693028e-01 6.84820771e-01
4.32399988e-01 1.81927457e-01 -9.56310153e-01 -9.13662970e-01
-2.92849600e-01 5.82479537e-01 -7.73218453e-01 -8.85141373e-01
-5.53723395e-01 -7.28416383e-01 1.12708890e+00 5.52553594e-01
1.23889363e+00 -1.19951856e+00 -6.48270100e-02 -2.15215832e-01
-3.06171812e-02 -1.21714211e+00 -4.22547460e-01 -5.48617616e-02
-4.88397390e-01 -1.02049041e+00 -7.96124876e-01 -7.85592735e-01
9.93695617e-01 1.91996127e-01 5.68504035e-01 3.17235529e-01
-1.55158684e-01 -5.45073710e-02 -5.20346820e-01 -5.11439502e-01
-6.78200483e-01 -5.06076992e-01 3.32188457e-01 2.74477094e-01
-2.11602867e-01 -7.47457385e-01 -5.90493679e-01 6.54246330e-01
-1.46585405e+00 -4.07802224e-01 4.42621022e-01 8.31710100e-01
3.12604070e-01 4.90625918e-01 4.65518981e-01 -6.98780596e-01
5.78029692e-01 -2.70018905e-01 -5.52048385e-01 1.66305184e-01
-1.13386892e-01 -7.21016852e-03 1.56178212e+00 -1.08302093e+00
-1.01978469e+00 -2.28083804e-01 -5.02595901e-01 -8.15692306e-01
-5.16482174e-01 -7.98379928e-02 -8.00246358e-01 -6.78324521e-01
1.07262671e+00 2.95113325e-01 -1.55151099e-01 -2.00311884e-01
3.26584846e-01 2.58727670e-01 6.64526165e-01 -5.99718392e-01
1.63034594e+00 3.83499831e-01 1.51627749e-01 -7.43926466e-01
-7.00706840e-01 4.01337385e-01 1.62495468e-02 -6.16580993e-02
8.19124401e-01 -7.80432820e-01 -7.41116464e-01 1.03384686e+00
-1.07372963e+00 -3.04854929e-01 -9.93857458e-02 3.15588295e-01
-1.89368308e-01 6.25572324e-01 -5.87070405e-01 -5.16440630e-01
-2.38963515e-01 -1.24852204e+00 5.11691988e-01 3.46473396e-01
5.90257585e-01 -7.55503535e-01 -3.59150946e-01 9.39785019e-02
4.55315977e-01 7.52607703e-01 7.38845527e-01 -7.07925320e-01
-4.23018336e-01 -2.38797665e-01 -2.21758187e-01 1.02389312e+00
2.58453190e-01 1.45614922e-01 -1.20662081e+00 -4.44870830e-01
3.79188538e-01 -3.81666690e-01 7.89693534e-01 -3.70016359e-02
1.58239770e+00 -8.67439628e-01 2.08820343e-01 1.19732475e+00
1.33905065e+00 1.97147399e-01 1.13855422e+00 4.23297167e-01
1.00030553e+00 5.62239587e-01 3.99879247e-01 8.64672214e-02
-4.16218340e-01 3.23877871e-01 9.96298015e-01 -3.45602989e-01
5.61089814e-02 -2.15532959e-01 6.06671512e-01 3.45052540e-01
-1.47879019e-01 -3.23460937e-01 -4.41229790e-01 5.41711748e-02
-1.38855112e+00 -1.11661541e+00 -5.03504686e-02 1.91105521e+00
9.01695848e-01 4.49258298e-01 -1.57393306e-01 4.13703620e-01
9.31644022e-01 4.37163413e-01 -7.11361349e-01 -3.98486227e-01
-4.37971145e-01 -2.02096011e-02 7.31197774e-01 2.52767354e-01
-1.19629347e+00 9.07300234e-01 5.11733866e+00 1.09694433e+00
-1.29614544e+00 -8.33284631e-02 7.69694924e-01 1.02610156e-01
-3.18593174e-01 -4.19964582e-01 -4.68637586e-01 6.37525678e-01
4.64527994e-01 8.15840885e-02 5.95642567e-01 9.75361228e-01
4.45930734e-02 6.12903416e-01 -6.98897719e-01 8.95676911e-01
-4.57865708e-02 -1.03719735e+00 3.28843415e-01 -1.16305009e-01
7.83850193e-01 -3.25901359e-01 5.34364223e-01 3.24036896e-01
3.55706245e-01 -1.21355331e+00 6.30809367e-01 2.32961401e-01
7.96849072e-01 -1.15746951e+00 8.32582295e-01 3.29888850e-01
-8.95194232e-01 -1.46976113e-01 -8.14417303e-01 6.51579648e-02
-1.51020303e-01 4.40210313e-01 -1.32646948e-01 5.51174462e-01
7.88220227e-01 4.45078701e-01 -6.10142529e-01 6.38788223e-01
-7.73714840e-01 6.89410985e-01 -1.13511883e-01 1.99707478e-01
3.07179123e-01 -1.70140579e-01 7.00158834e-01 9.17716920e-01
8.10567141e-02 3.01321775e-01 -4.96304296e-02 9.69245970e-01
-3.00597817e-01 -7.12603405e-02 -9.04715002e-01 1.41185030e-01
3.36229950e-01 1.14441609e+00 -4.52760756e-01 -4.43495065e-03
-6.38262257e-02 1.04304934e+00 -8.14223662e-02 6.23307943e-01
-1.30150867e+00 -7.83878326e-01 8.45792651e-01 -2.79122800e-01
1.50388032e-01 9.50114278e-04 -1.35636136e-01 -1.12877738e+00
1.05565481e-01 -1.37528229e+00 1.42179862e-01 -6.31182313e-01
-1.41216779e+00 8.97170067e-01 -2.71058530e-01 -1.44234252e+00
4.61261243e-01 -6.99013412e-01 -1.03013253e+00 8.70622516e-01
-1.41918886e+00 -1.09335494e+00 -4.15247977e-01 1.23231268e+00
2.56402165e-01 -3.12047631e-01 5.48437238e-01 3.32266867e-01
-8.06884110e-01 1.13831460e+00 -6.67935088e-02 5.87364018e-01
7.35597074e-01 -9.49002624e-01 5.18077075e-01 1.26516092e+00
-7.50702918e-02 6.29470527e-01 8.74558032e-01 -3.37393612e-01
-1.20235717e+00 -1.52892637e+00 -2.52183110e-01 -2.57820398e-01
5.84617436e-01 -4.00634408e-01 -1.18698502e+00 5.55877268e-01
1.39091492e-01 4.12749559e-01 2.49159813e-01 -5.33592165e-01
-6.45048738e-01 -4.13643807e-01 -1.57796156e+00 9.65742171e-01
8.42722178e-01 -5.38568079e-01 -6.62891686e-01 3.21867794e-01
1.11686313e+00 -2.48213530e-01 -6.08389437e-01 5.00544667e-01
2.13487044e-01 -9.71010685e-01 1.37882173e+00 -6.57829642e-01
5.08246779e-01 -3.83115023e-01 -2.38029614e-01 -1.55815256e+00
-2.30784982e-01 -7.08268702e-01 3.52857828e-01 1.46174932e+00
8.96126479e-02 -1.02904296e+00 4.92267430e-01 4.21828449e-01
-1.16010576e-01 -3.17470491e-01 -7.09407032e-01 -1.17215228e+00
3.65521133e-01 -2.56406039e-01 8.73603642e-01 1.15884268e+00
-5.66505253e-01 -1.10588878e-01 -6.55166507e-01 8.06640804e-01
6.43293440e-01 -4.50651407e-01 7.31027126e-01 -7.98888206e-01
-1.19920239e-01 -2.88116366e-01 -5.81559658e-01 -6.23500228e-01
3.11239749e-01 -4.65287864e-01 5.20759560e-02 -7.98247159e-01
-3.33530724e-01 -2.87205756e-01 -4.28223848e-01 5.97276926e-01
-4.47979480e-01 5.71541011e-01 1.94580331e-01 -5.39526232e-02
-3.47612202e-02 6.70491993e-01 1.64664221e+00 -3.89261097e-01
9.46423411e-02 3.00619658e-02 -6.54177666e-01 9.19561684e-01
1.00619471e+00 -5.63921690e-01 -6.44127011e-01 -4.86845464e-01
8.19076002e-02 -4.22610492e-01 6.55487061e-01 -1.20985675e+00
-1.22853100e-01 -3.13897640e-01 5.02259791e-01 9.99127328e-03
8.70326757e-02 -1.12949693e+00 4.35581990e-02 6.92435920e-01
-2.64754295e-01 -1.36205330e-01 3.09027642e-01 5.79586029e-01
-1.93537995e-01 -1.61986440e-01 1.40723860e+00 -1.22954242e-01
-5.26126385e-01 5.86760938e-01 -4.44523804e-02 1.91563383e-01
1.22150576e+00 -1.83483928e-01 -6.72911942e-01 -2.58914918e-01
-4.05553818e-01 -6.33814782e-02 4.14910078e-01 4.23172832e-01
7.74028838e-01 -1.41704953e+00 -6.87500358e-01 4.90094960e-01
-1.06819205e-01 -6.63599893e-02 4.57641602e-01 4.10169214e-01
-5.79312384e-01 -2.89393812e-01 -7.71919608e-01 -2.25330636e-01
-1.17313945e+00 1.29657996e+00 5.52719653e-01 2.28819340e-01
-6.65146232e-01 9.24772501e-01 8.03619325e-01 -3.94750088e-01
3.72733504e-01 -1.25177428e-01 -3.22190940e-01 -3.93558055e-01
7.10666776e-01 2.20623672e-01 -1.34750605e-01 -4.47067052e-01
-2.81735212e-01 5.17325401e-01 8.56665000e-02 3.36020380e-01
1.20057094e+00 1.30450115e-01 -9.10666138e-02 -2.19167799e-01
1.33838856e+00 2.33756244e-01 -1.52510059e+00 -2.63395701e-02
-7.97325671e-01 -5.80923557e-01 -7.60301575e-02 -6.33117676e-01
-1.62155402e+00 9.96254206e-01 7.48022377e-01 5.18308818e-01
1.33847070e+00 -5.35428226e-01 8.83409142e-01 4.21595305e-01
9.89670083e-02 -6.49777293e-01 3.12981129e-01 3.17378074e-01
1.12672544e+00 -1.18625164e+00 -2.95274019e-01 -3.64662081e-01
-6.73270345e-01 1.04352891e+00 8.72084618e-01 -4.69536245e-01
5.87657809e-01 2.85445482e-01 3.57651442e-01 2.22954005e-01
-2.28008792e-01 2.88528442e-01 2.95283735e-01 9.23124373e-01
-2.90429056e-01 -2.18671948e-01 7.83831105e-02 6.49262249e-01
-3.13451231e-01 -6.63148403e-01 5.79693019e-01 5.06396472e-01
-3.45847636e-01 -7.94512153e-01 -7.94979513e-01 -4.30291630e-02
-6.20121062e-01 -1.12704404e-01 -3.84231150e-01 6.97497129e-01
3.40665817e-01 1.06434393e+00 -4.97375131e-01 -7.74464071e-01
5.26096582e-01 -4.50426310e-01 2.35938370e-01 -1.96964711e-01
-6.13294959e-01 -3.30078781e-01 -2.58431166e-01 -6.00060105e-01
-1.32813796e-01 7.24338517e-02 -9.62273240e-01 -5.83368480e-01
-2.54760563e-01 1.42851425e-02 3.73025239e-01 8.34582686e-01
-6.35367706e-02 8.08188081e-01 1.25360572e+00 -6.50664449e-01
-8.87622237e-01 -7.72304416e-01 -6.04065478e-01 8.15250337e-01
5.22745550e-01 -5.44781744e-01 -8.28508258e-01 -3.76206078e-02] | [5.489735126495361, 7.9291486740112305] |
505ca667-3c4a-4ae4-9dda-bbc17d5b5076 | leveraging-graph-based-cross-modal | 2211.00526 | null | https://arxiv.org/abs/2211.00526v1 | https://arxiv.org/pdf/2211.00526v1.pdf | Leveraging Graph-based Cross-modal Information Fusion for Neural Sign Language Translation | Sign Language (SL), as the mother tongue of the deaf community, is a special visual language that most hearing people cannot understand. In recent years, neural Sign Language Translation (SLT), as a possible way for bridging communication gap between the deaf and the hearing people, has attracted widespread academic attention. We found that the current mainstream end-to-end neural SLT models, which tries to learning language knowledge in a weakly supervised manner, could not mine enough semantic information under the condition of low data resources. Therefore, we propose to introduce additional word-level semantic knowledge of sign language linguistics to assist in improving current end-to-end neural SLT models. Concretely, we propose a novel neural SLT model with multi-modal feature fusion based on the dynamic graph, in which the cross-modal information, i.e. text and video, is first assembled as a dynamic graph according to their correlation, and then the graph is processed by a multi-modal graph encoder to generate the multi-modal embeddings for further usage in the subsequent neural translation models. To the best of our knowledge, we are the first to introduce graph neural networks, for fusing multi-modal information, into neural sign language translation models. Moreover, we conducted experiments on a publicly available popular SLT dataset RWTH-PHOENIX-Weather-2014T. and the quantitative experiments show that our method can improve the model. | ['Stan Z. Li', 'Yidong Chen', 'Chong Wu', 'Cheng Tan', 'Siyuan Li', 'Jiangbin Zheng'] | 2022-11-01 | null | null | null | null | ['sign-language-translation'] | ['computer-vision'] | [ 1.86842859e-01 1.20319828e-01 -3.01475406e-01 -2.22239390e-01
-7.19758570e-01 -7.39749074e-02 4.39461559e-01 -4.98712391e-01
-5.73624194e-01 4.07397240e-01 7.73191214e-01 -3.96215826e-01
8.53055567e-02 -7.47889280e-01 -6.26445234e-01 -5.57094455e-01
3.34432602e-01 3.17750335e-01 2.03874260e-01 -3.02374244e-01
-2.27641270e-01 -1.29942879e-01 -1.70515227e+00 4.47783887e-01
1.01772547e+00 7.21512020e-01 3.82278770e-01 2.86071241e-01
-5.41103303e-01 7.83180833e-01 -5.23683280e-02 -5.23220479e-01
1.18582301e-01 -8.27957749e-01 -5.58805346e-01 -2.33927071e-01
5.83033025e-01 -3.94888759e-01 -6.36604726e-01 1.18443620e+00
1.03574610e+00 -2.36766204e-01 4.73929137e-01 -1.18204367e+00
-1.21020150e+00 8.62629175e-01 -2.25413159e-01 -3.16009194e-01
2.10511386e-01 3.04623812e-01 1.25236034e+00 -1.14898729e+00
9.75431979e-01 1.45507514e+00 5.24403155e-01 9.40033197e-01
-5.49127162e-01 -7.28192389e-01 2.69691736e-01 6.85082316e-01
-1.11883461e+00 -2.80469626e-01 9.85121310e-01 -2.43252829e-01
8.53014946e-01 -1.16931595e-01 1.12359571e+00 1.17275941e+00
-1.16286702e-01 1.31050408e+00 1.15790033e+00 -5.49468160e-01
-1.34061769e-01 -2.13110238e-01 -3.29363085e-02 8.85924339e-01
1.61205009e-02 1.85780361e-01 -8.85538816e-01 3.43877614e-01
4.57443386e-01 -2.75541365e-01 -4.83492106e-01 -1.32400602e-01
-1.31165743e+00 6.79143250e-01 5.85399389e-01 5.33166707e-01
-2.19893888e-01 1.46918565e-01 4.63205546e-01 4.23084259e-01
4.47009206e-01 -2.98379093e-01 -3.30506772e-01 -5.54235913e-02
-6.39727592e-01 -1.73038542e-01 5.59567630e-01 8.56752992e-01
4.39504683e-01 7.89055526e-02 -3.48939151e-01 1.08499253e+00
8.60163212e-01 8.05670738e-01 8.05753291e-01 -3.79295468e-01
6.56615674e-01 8.45645189e-01 -6.65115833e-01 -5.91656446e-01
-2.59039611e-01 -3.41286778e-01 -7.80692279e-01 2.55909026e-01
3.46866071e-01 -2.03419238e-01 -1.40971899e+00 2.07904196e+00
-4.73110490e-02 4.52001035e-01 2.81220794e-01 1.14581788e+00
1.42009676e+00 2.37939253e-01 1.64521366e-01 -1.51372746e-01
1.31647480e+00 -1.03008449e+00 -9.57697749e-01 -1.13646664e-01
6.98315322e-01 -8.05567086e-01 1.29890525e+00 8.21545906e-03
-9.98690486e-01 -3.88369471e-01 -7.72228181e-01 -3.75509083e-01
-4.97268140e-01 2.91168064e-01 5.38910687e-01 3.22252393e-01
-1.21475828e+00 1.05475768e-01 -7.39710867e-01 -7.51446247e-01
6.71398818e-01 -1.85736027e-02 -4.38758254e-01 -4.02398378e-01
-1.28791273e+00 1.06759071e+00 2.56068319e-01 4.18747783e-01
-3.63655537e-01 -4.01848912e-01 -9.05842483e-01 -4.32886928e-01
8.22364762e-02 -1.01624191e+00 1.09353983e+00 -6.83494985e-01
-1.52312696e+00 9.77949142e-01 -3.89788985e-01 -1.99030623e-01
7.40484893e-01 -1.37155235e-01 -4.61183131e-01 -8.13391805e-03
2.48536840e-02 8.52119029e-01 7.69502640e-01 -1.16137195e+00
-7.42490888e-01 -4.96418834e-01 -2.69854873e-01 4.21332300e-01
-6.10121489e-01 1.72842704e-02 -6.47472084e-01 -7.13619530e-01
4.54121590e-01 -8.92381549e-01 1.57944024e-01 2.57392526e-01
-3.24768066e-01 -5.31539023e-01 9.14472580e-01 -1.00333405e+00
1.07554233e+00 -2.04026818e+00 6.02426410e-01 5.16046882e-02
2.77434140e-01 4.85163897e-01 -4.29533869e-01 2.61204183e-01
1.40176296e-01 -7.67556131e-02 -5.06530344e-01 -5.58139503e-01
1.63081408e-01 4.86728042e-01 -6.80134669e-02 3.25564384e-01
-6.78619444e-02 1.39053488e+00 -9.77529526e-01 -5.54740191e-01
2.11078137e-01 7.54027128e-01 -4.67160761e-01 6.28321096e-02
-2.45969534e-01 4.08417970e-01 -2.16960460e-01 9.31600451e-01
5.03609419e-01 -1.50283671e-03 -4.77310009e-02 -4.06092763e-01
-2.66295001e-02 1.14123374e-01 -7.37270892e-01 2.10192323e+00
-5.56575537e-01 7.20870256e-01 -2.52051294e-01 -8.77133608e-01
6.35837674e-01 3.77455592e-01 4.90776390e-01 -9.24013376e-01
4.08676982e-01 7.06094146e-01 1.33358151e-01 -8.22801352e-01
-7.92762041e-02 -3.87012959e-01 1.58496723e-01 3.35422009e-01
2.52730399e-01 -1.02878608e-01 1.79072231e-01 2.84437872e-02
9.09089208e-01 1.60556287e-01 -2.32223719e-01 3.04443032e-01
4.51476932e-01 -1.17884703e-01 2.90657818e-01 1.49815395e-01
-1.53256610e-01 6.25618398e-01 8.78965110e-02 1.11615099e-02
-7.01390982e-01 -1.16312575e+00 1.26265615e-01 7.92939007e-01
1.65928230e-02 -2.25440472e-01 -5.67770302e-01 -6.29447818e-01
-3.20881680e-02 5.30583501e-01 -2.66091049e-01 -2.26744920e-01
-6.98825419e-01 -4.50948119e-01 7.27511048e-01 5.37114978e-01
7.85394371e-01 -1.34055197e+00 -2.27303773e-01 1.95807487e-01
-4.52664018e-01 -1.32550097e+00 -7.85875618e-01 -3.14422935e-01
-6.49456739e-01 -9.43693280e-01 -1.18754959e+00 -1.54367435e+00
6.53257370e-01 1.67102501e-01 5.44321299e-01 1.11423522e-01
-1.23527206e-01 4.60660726e-01 -6.62185431e-01 -3.97822022e-01
-4.35578346e-01 -1.78297892e-01 -2.64312699e-02 1.03010960e-01
6.01311862e-01 -8.19022298e-01 -3.25327039e-01 1.25176413e-03
-8.52243006e-01 4.99338895e-01 8.76276135e-01 1.09088385e+00
4.74530816e-01 -3.35973591e-01 4.23980564e-01 -3.58285576e-01
7.45921791e-01 -1.58804342e-01 -1.65971726e-01 3.68620694e-01
-4.25610363e-01 1.69429630e-01 2.44323999e-01 -6.28091097e-01
-8.53319883e-01 -5.98370694e-02 -4.88524288e-01 -4.45224971e-01
1.64325923e-01 9.36670065e-01 -3.36445928e-01 -2.51000494e-01
3.57405543e-01 6.06087208e-01 2.95873910e-01 -6.27983928e-01
7.23644018e-01 8.93371284e-01 5.82003236e-01 -1.80637583e-01
9.50867951e-01 3.50090146e-01 -7.13244379e-02 -7.15518236e-01
-5.26470482e-01 -3.16081405e-01 -5.90718508e-01 -4.64344561e-01
9.60633993e-01 -8.44986796e-01 -5.78785598e-01 1.04026783e+00
-1.36044741e+00 -3.70013595e-01 -2.20254257e-01 8.97048712e-01
-5.77860832e-01 5.97492337e-01 -5.43241024e-01 -5.34331441e-01
-3.11116070e-01 -1.06967330e+00 1.04828894e+00 1.59128740e-01
1.64270371e-01 -9.07105327e-01 2.66586542e-02 5.75794816e-01
3.42351139e-01 1.42935617e-02 1.07189882e+00 -2.95881689e-01
-5.61576366e-01 7.66859949e-02 -5.47481298e-01 7.07334161e-01
2.04292610e-01 -4.55746651e-01 -8.16287816e-01 -2.32703626e-01
-4.50809062e-01 -3.05909991e-01 1.20920205e+00 3.65841925e-01
6.75945997e-01 -2.00568095e-01 -1.37115538e-01 5.72349846e-01
1.31760967e+00 -3.40692848e-02 6.52091742e-01 2.59439778e-02
1.10484993e+00 4.14395690e-01 2.72346288e-01 -4.64565307e-02
9.08506393e-01 5.35490870e-01 2.73363978e-01 -3.31133425e-01
-1.16873038e+00 -7.58027315e-01 6.13120914e-01 1.56761003e+00
-4.09300268e-01 -2.22284719e-01 -9.41079974e-01 8.36858153e-01
-2.04155827e+00 -8.17707479e-01 -6.64857477e-02 1.75906658e+00
8.12965989e-01 -1.13023072e-01 -2.55678803e-01 2.36199692e-01
5.73518872e-01 8.60528797e-02 -6.27938271e-01 1.47657067e-01
-5.50208747e-01 3.21634561e-01 4.90136206e-01 5.25447547e-01
-7.84223437e-01 1.36654174e+00 4.88548851e+00 7.05487013e-01
-1.32459199e+00 3.48311543e-01 -3.23989272e-01 1.25355855e-01
-5.75154126e-01 -1.01177566e-01 -5.88923573e-01 3.55838716e-01
5.24912834e-01 -1.07889205e-01 5.24607778e-01 4.18550491e-01
3.94483238e-01 4.69387114e-01 -8.59954417e-01 1.13512671e+00
4.97210920e-01 -1.14435637e+00 4.79651123e-01 2.84555163e-02
5.16398311e-01 4.95021701e-01 -2.20016474e-04 4.69777644e-01
2.71933883e-01 -9.51146364e-01 8.96369100e-01 5.76830924e-01
1.03170180e+00 -2.38905162e-01 8.21564853e-01 3.55469584e-01
-1.40393043e+00 -5.29097468e-02 1.21327452e-01 8.03267062e-02
6.36814952e-01 2.43084058e-01 -6.41514778e-01 5.94762683e-01
5.41851103e-01 1.19457924e+00 -3.86314005e-01 1.13454580e+00
-6.78953588e-01 6.66709185e-01 -3.03864419e-01 -1.88765854e-01
1.83737800e-01 -2.80321147e-02 7.42710829e-01 9.59849894e-01
6.95149541e-01 -1.32002801e-01 1.21960983e-01 5.93486786e-01
-1.16860427e-01 4.79936928e-01 -7.90369034e-01 -9.00276601e-02
5.67512587e-02 4.81319785e-01 -1.93491042e-01 -1.69521511e-01
-8.55100095e-01 1.09824181e+00 2.02732250e-01 6.53692961e-01
-5.82775414e-01 -2.86394387e-01 6.36891544e-01 9.82637051e-03
4.00315672e-02 -3.85522217e-01 -2.85598159e-01 -1.43431306e+00
5.51603079e-01 -6.61264122e-01 3.27648669e-01 -8.99524152e-01
-1.42492139e+00 5.58874965e-01 -3.06561053e-01 -1.57099104e+00
-1.62380874e-01 -9.25418437e-01 -2.65483350e-01 8.96200418e-01
-2.05555558e+00 -2.02367115e+00 -3.02247763e-01 9.72975016e-01
5.11317551e-01 -4.70943242e-01 7.25806892e-01 6.37646556e-01
-1.36238411e-01 8.94526184e-01 -2.71085709e-01 6.11645281e-01
5.33177674e-01 -8.07181120e-01 1.28919274e-01 9.81718361e-01
3.62094045e-01 3.34765255e-01 3.56847793e-01 -7.83994377e-01
-1.54838157e+00 -1.12156689e+00 1.30768573e+00 7.94052705e-03
7.57364154e-01 -1.00667119e-01 -7.52500832e-01 4.81761307e-01
1.65478200e-01 2.24284455e-02 4.90843356e-01 -2.25421607e-01
-4.99311984e-01 1.51513264e-01 -8.79777074e-01 9.19003785e-01
1.82926917e+00 -9.04510617e-01 -6.64778769e-01 3.75586152e-01
1.06687260e+00 -4.78758276e-01 -7.85269737e-01 5.84599614e-01
7.59939432e-01 -4.97538626e-01 7.58016586e-01 -5.24774432e-01
2.79561520e-01 -4.27204967e-01 -3.98266256e-01 -1.53679645e+00
2.55314469e-01 -3.56152743e-01 7.08207255e-03 9.61831391e-01
4.90580142e-01 -7.69496918e-01 5.83787680e-01 3.47627923e-02
-5.02272010e-01 -6.04792118e-01 -1.46192408e+00 -7.67010093e-01
2.09510386e-01 -8.64775360e-01 4.99493480e-01 7.68390119e-01
1.40404701e-01 2.61787444e-01 -4.87443984e-01 2.62977257e-02
6.60632014e-01 -9.86896530e-02 5.65582514e-01 -1.24414361e+00
-1.25778407e-01 -7.69124746e-01 -6.29036009e-01 -1.10290575e+00
4.64062750e-01 -1.50328839e+00 9.06285420e-02 -2.19090724e+00
-4.30883653e-03 -4.96629104e-02 -2.95404911e-01 9.34759796e-01
4.69828397e-02 2.14804560e-01 3.18951100e-01 1.17025323e-01
-1.96456462e-02 8.64321351e-01 1.82590377e+00 -4.81771141e-01
-2.28906423e-01 -3.92223373e-02 -5.99283457e-01 6.53900981e-01
6.50471032e-01 -2.15927914e-01 -4.41190124e-01 -8.97647023e-01
-2.55497433e-02 -3.30974013e-02 4.99414086e-01 -8.08130085e-01
4.73276943e-01 -8.74762330e-03 -4.04566586e-01 -5.03201187e-01
2.28833035e-01 -8.57953489e-01 -4.17497307e-01 4.55651402e-01
-1.77584857e-01 -4.22532737e-01 3.73401567e-02 3.72451842e-01
-5.65315187e-01 2.02390030e-01 6.33388400e-01 -5.23670316e-02
-8.93700480e-01 6.82533383e-01 -1.82830274e-01 1.25106394e-01
6.13403857e-01 -2.33421877e-01 -4.07192558e-01 -4.47227418e-01
-6.89607322e-01 2.12627307e-01 1.99394561e-02 8.31461489e-01
9.10227060e-01 -1.77286696e+00 -1.07238615e+00 2.87926078e-01
5.36875188e-01 -1.89413846e-01 3.05571288e-01 9.14889216e-01
-2.27725267e-01 2.01788723e-01 -1.70902327e-01 -5.27054071e-01
-1.30304408e+00 -3.43644619e-03 2.42817089e-01 7.30337203e-02
-9.88962829e-01 8.48653555e-01 -3.59492511e-01 -5.87046742e-01
4.72435325e-01 -7.40239978e-01 -3.29865783e-01 1.34018183e-01
2.85039365e-01 -5.15178069e-02 -3.15126181e-02 -1.07390749e+00
-2.23024428e-01 1.08548272e+00 2.96902686e-01 -4.83712405e-01
1.37927711e+00 -9.59168305e-04 -2.20145404e-01 3.97341192e-01
1.32889044e+00 -2.75681347e-01 -6.28071666e-01 -7.87616491e-01
-2.18416899e-01 -3.23638320e-01 2.38945663e-01 -8.95396650e-01
-1.34046173e+00 1.17058599e+00 7.80275166e-01 -4.36622083e-01
1.33797705e+00 1.82956204e-01 1.35542691e+00 3.17892790e-01
5.19834220e-01 -1.06237221e+00 3.54152173e-02 6.41125441e-01
1.13374567e+00 -1.27519345e+00 -6.00271344e-01 -2.72957444e-01
-4.88664687e-01 9.89721417e-01 5.03924251e-01 2.29169369e-01
9.27378237e-01 5.13930209e-02 4.85530913e-01 -7.32930303e-02
-3.60267550e-01 -8.02264988e-01 5.27818501e-01 8.12984467e-01
2.70794660e-01 2.23352373e-01 -6.12866461e-01 5.52587688e-01
-3.76842201e-01 3.61988217e-01 1.32610098e-01 7.89900899e-01
-3.95173877e-01 -1.38431096e+00 1.13276253e-02 3.65659595e-01
2.53778309e-01 -3.29956025e-01 -2.55143017e-01 6.18981004e-01
4.43639666e-01 8.79763246e-01 -3.98364186e-01 -6.01915002e-01
6.45624220e-01 3.24915498e-01 4.87262189e-01 -4.34928685e-01
-6.34325668e-02 -8.42931122e-02 3.90623361e-02 -2.69327670e-01
-6.65970147e-01 -6.87233627e-01 -1.54519606e+00 8.64559039e-02
9.61151943e-02 -5.44904828e-01 7.36811101e-01 1.29717231e+00
3.42218429e-02 5.38061380e-01 1.14810176e-01 -6.15832090e-01
-2.24986628e-01 -1.19034898e+00 -3.39236915e-01 4.73982692e-01
3.88482362e-01 -7.20225453e-01 -3.94356191e-01 -8.02723095e-02] | [9.211767196655273, -6.520484447479248] |
9295c4ee-b5e6-4632-9273-c2e2d493f2f0 | 4d-temporally-coherent-light-field-video | 1804.11276 | null | http://arxiv.org/abs/1804.11276v1 | http://arxiv.org/pdf/1804.11276v1.pdf | 4D Temporally Coherent Light-field Video | Light-field video has recently been used in virtual and augmented reality
applications to increase realism and immersion. However, existing light-field
methods are generally limited to static scenes due to the requirement to
acquire a dense scene representation. The large amount of data and the absence
of methods to infer temporal coherence pose major challenges in storage,
compression and editing compared to conventional video. In this paper, we
propose the first method to extract a spatio-temporally coherent light-field
video representation. A novel method to obtain Epipolar Plane Images (EPIs)
from a spare light-field camera array is proposed. EPIs are used to constrain
scene flow estimation to obtain 4D temporally coherent representations of
dynamic light-fields. Temporal coherence is achieved on a variety of
light-field datasets. Evaluation of the proposed light-field scene flow against
existing multi-view dense correspondence approaches demonstrates a significant
improvement in accuracy of temporal coherence. | ['Jean-yves Guillemaut', 'Marco Volino', 'Armin Mustafa', 'Adrian Hilton'] | 2018-04-30 | null | null | null | null | ['scene-flow-estimation'] | ['computer-vision'] | [ 5.23811042e-01 -8.86093795e-01 3.38975132e-01 -2.43910387e-01
-3.60007912e-01 -5.77775896e-01 5.28614879e-01 -2.91203469e-01
-1.66534185e-01 8.95872772e-01 2.33802542e-01 1.22699082e-01
-2.15807259e-01 -6.40897989e-01 -4.60942745e-01 -4.89876270e-01
2.89543241e-01 8.23634267e-02 5.13830125e-01 -1.64727226e-01
5.34016311e-01 5.89829147e-01 -1.62942433e+00 1.97017357e-01
4.17504668e-01 5.57904124e-01 6.97822630e-01 8.46722901e-01
9.33954641e-02 8.28797042e-01 1.54306442e-02 3.55449021e-02
5.11625171e-01 -3.65715861e-01 -6.69434845e-01 1.92380115e-01
1.11767256e+00 -9.59169686e-01 -8.65486503e-01 9.10789132e-01
4.33561355e-01 5.27922630e-01 2.47300547e-02 -1.00023842e+00
-1.02854960e-01 -5.07842183e-01 -7.58386850e-01 7.11720765e-01
1.05419707e+00 7.10576475e-02 5.87529302e-01 -1.02154315e+00
1.21204042e+00 8.14275086e-01 2.78238416e-01 2.84496903e-01
-1.31938207e+00 -4.73267019e-01 -2.70556450e-01 3.50196153e-01
-1.28416586e+00 -6.98690414e-01 9.26203430e-01 -5.64740956e-01
1.12472999e+00 2.03567550e-01 1.07118535e+00 5.07152855e-01
3.41971934e-01 1.09344736e-01 1.26695752e+00 -6.73178554e-01
3.82166682e-03 -1.26806974e-01 9.96257141e-02 6.36976302e-01
2.02087060e-01 5.14135003e-01 -9.85475183e-01 -3.53273563e-02
1.25853002e+00 1.37347639e-01 -7.09389865e-01 -4.33407009e-01
-1.37787950e+00 4.29356426e-01 -1.28916148e-02 2.79601246e-01
-3.37455243e-01 2.00337678e-01 1.66268080e-01 3.67654637e-02
5.30283928e-01 3.02913189e-01 -4.72190644e-04 -1.97684094e-01
-9.44744945e-01 1.25833347e-01 4.30780798e-01 9.72717047e-01
1.11252725e+00 1.65579408e-01 3.25721323e-01 4.69451994e-01
2.08029330e-01 6.26239240e-01 -4.35002409e-02 -1.33627510e+00
2.96936065e-01 2.64353693e-01 2.99148828e-01 -1.31955123e+00
-1.55059949e-01 9.17871073e-02 -5.56803226e-01 3.08572471e-01
3.37983519e-01 3.26269001e-01 -5.85021794e-01 1.16427767e+00
5.21681964e-01 8.85261476e-01 3.20053436e-02 1.02796614e+00
6.89907432e-01 8.60103667e-01 -6.01801753e-01 -6.46555126e-01
9.41328406e-01 -5.21865427e-01 -9.24073935e-01 -6.26009926e-02
1.42349005e-01 -1.25228488e+00 5.83193779e-01 4.41180259e-01
-1.42927754e+00 -4.26350772e-01 -8.60004008e-01 -3.44457120e-01
2.29893237e-01 -2.53401577e-01 6.58231676e-01 3.89209539e-01
-9.76621509e-01 2.51115113e-01 -8.37613225e-01 -3.97466272e-01
6.33527189e-02 2.80235469e-01 -5.83003342e-01 -6.86963320e-01
-6.82697296e-01 7.53412366e-01 1.78160176e-01 5.47082573e-02
-5.86747348e-01 -9.62473035e-01 -8.57540667e-01 -2.43496940e-01
7.81837329e-02 -9.66956496e-01 7.87855089e-01 -4.56821173e-01
-1.47830367e+00 7.40474284e-01 -5.91106057e-01 3.27788517e-02
2.40337163e-01 -1.07069246e-01 -3.07713091e-01 6.87542796e-01
5.60787506e-03 3.39723915e-01 5.88025391e-01 -1.26897264e+00
-7.28598356e-01 -1.21282682e-01 2.81624168e-01 4.64561880e-01
-1.36769060e-02 -4.20849249e-02 -5.85784018e-01 -2.04640999e-01
4.75363851e-01 -9.41782057e-01 -1.03778064e-01 1.58695236e-01
-1.06235512e-01 5.18332362e-01 1.11190391e+00 -4.61957127e-01
1.10139251e+00 -1.92110157e+00 -2.19976883e-02 -1.00304797e-01
3.72385412e-01 6.84770569e-02 1.08289294e-01 4.54733104e-01
-1.91925501e-03 -7.65797257e-01 9.07343328e-02 -2.27688074e-01
-7.51045048e-01 3.52957904e-01 -3.49728644e-01 1.01842570e+00
-3.64628762e-01 3.72521043e-01 -1.05386090e+00 -5.60015321e-01
1.13199162e+00 8.63455057e-01 -9.46966052e-01 1.52705178e-01
-1.61214583e-02 1.03204978e+00 -2.53707230e-01 4.45932299e-01
1.04406238e+00 -2.36643478e-01 1.13291875e-01 -5.29468000e-01
-5.64119101e-01 1.07665986e-01 -1.46452987e+00 2.08047748e+00
-6.74675763e-01 1.30294979e+00 1.06418449e-02 -4.22397107e-01
5.72680473e-01 3.13518107e-01 1.04156494e+00 -8.27697933e-01
7.43673146e-02 1.33033335e-01 -4.29554582e-01 -5.02605259e-01
1.01914656e+00 -2.21758425e-01 6.40601397e-01 4.23437417e-01
-1.42688990e-01 -4.76336747e-01 3.73753130e-01 2.84641773e-01
9.96306717e-01 2.37807378e-01 2.43513167e-01 -1.71365827e-01
5.94317019e-01 1.85026899e-02 4.81401473e-01 1.52763575e-01
-8.65832642e-02 8.42332482e-01 -4.07963902e-01 -6.54255271e-01
-1.13742220e+00 -1.02655160e+00 -2.89794087e-01 3.47670019e-01
7.27640986e-01 -5.63121259e-01 -7.49986321e-02 2.09414542e-01
-3.80043536e-01 4.05262262e-01 -8.20303485e-02 5.39422333e-01
-9.24794316e-01 -2.99633741e-01 -2.38437951e-01 7.48703107e-02
5.80492437e-01 -4.01314378e-01 -1.11878574e+00 1.85373217e-01
-5.89029014e-01 -1.60644484e+00 -5.10867834e-01 -4.40578163e-01
-9.68258500e-01 -1.16708541e+00 -4.44598556e-01 -5.07089496e-01
7.39021838e-01 1.19434106e+00 1.04822576e+00 -4.94319797e-02
-5.60342312e-01 8.20532739e-01 -4.35701646e-02 1.54428676e-01
-1.28099963e-01 -8.48702908e-01 1.54212505e-01 1.01625562e-01
2.36155778e-01 -9.16676402e-01 -9.70271051e-01 4.94987994e-01
-9.44925606e-01 4.06966537e-01 -6.58972710e-02 7.99739242e-01
7.06035197e-01 -5.49985189e-03 -3.13929558e-01 -5.48224986e-01
-1.44647360e-02 -1.81108057e-01 -1.14133883e+00 -7.46054575e-02
-3.69522214e-01 -3.91881675e-01 3.77588302e-01 -1.07446268e-01
-1.48456204e+00 3.41610134e-01 3.12736988e-01 -8.28402936e-01
-2.02750657e-02 7.45635331e-02 3.44992101e-01 -5.10464787e-01
4.84105974e-01 2.10267529e-01 -1.72040746e-01 -1.24617025e-01
1.86936632e-01 1.69490412e-01 7.00824142e-01 -2.18865246e-01
7.03730881e-01 1.17839181e+00 4.99455273e-01 -1.23538780e+00
-5.30715466e-01 -1.09487331e+00 -1.03271222e+00 -5.71686566e-01
8.76763225e-01 -1.01367128e+00 -7.29740083e-01 1.80957556e-01
-1.37660599e+00 2.10888982e-02 -1.62163600e-01 1.10752952e+00
-7.78669417e-01 7.97438323e-01 -4.69343215e-01 -6.10118866e-01
9.32205934e-03 -1.06545937e+00 1.12998116e+00 1.51206955e-01
-1.16401222e-02 -1.14235651e+00 5.15969872e-01 4.70500767e-01
1.27441809e-01 2.30073124e-01 3.28942984e-01 8.41971338e-01
-1.42683327e+00 -8.02065525e-03 -2.84243464e-01 -1.21291921e-01
2.08862647e-01 1.35870934e-01 -1.08064473e+00 -3.40542018e-01
1.10803448e-01 9.89660919e-02 3.49384695e-01 8.57524514e-01
4.51777607e-01 1.14409365e-01 -3.46448660e-01 1.07929242e+00
2.05910730e+00 2.76411980e-01 9.24287736e-01 3.20496827e-01
8.01384985e-01 5.14361501e-01 6.57087982e-01 5.62833071e-01
6.52481318e-02 9.71676648e-01 2.01754272e-01 2.23466605e-02
-3.69941980e-01 -1.33927077e-01 8.45565402e-04 9.73330259e-01
-4.52862889e-01 -1.67031214e-01 -7.15239704e-01 5.23207307e-01
-1.50236881e+00 -1.43918252e+00 -6.79627180e-01 2.30292654e+00
4.37013626e-01 -3.60608160e-01 -2.97014117e-01 2.02042371e-01
4.87764835e-01 2.38031775e-01 -1.62045375e-01 -6.54488131e-02
-2.00825542e-01 2.36491226e-02 4.45049316e-01 1.08965874e+00
-5.80618620e-01 8.06849182e-01 6.58050346e+00 2.97541857e-01
-1.35671306e+00 1.64318353e-01 -1.56109795e-01 -4.28867012e-01
-4.89063352e-01 3.20768774e-01 -6.67602181e-01 1.99646637e-01
3.92578453e-01 -3.32764179e-01 4.00754899e-01 2.49793798e-01
5.45516849e-01 -7.84311593e-01 -1.05906630e+00 1.62142932e+00
8.46209824e-02 -1.59953141e+00 -1.10162668e-01 2.07993239e-01
1.11595321e+00 2.42401920e-02 -1.72191128e-01 -7.65770137e-01
-4.90867607e-02 -5.43213546e-01 2.85951912e-01 6.87981308e-01
1.13985002e+00 -4.12097394e-01 1.39622033e-01 8.73606578e-02
-1.34118450e+00 2.48689964e-01 -4.32446957e-01 -2.55188465e-01
9.71873283e-01 7.84878194e-01 -6.58704519e-01 6.61878467e-01
7.15442598e-01 1.07381487e+00 -7.21543133e-02 1.02888048e+00
4.87352997e-01 1.52590573e-01 -4.58410233e-01 5.27453303e-01
-5.56673780e-02 -7.14381218e-01 8.40944171e-01 8.77754748e-01
4.26986068e-01 6.87965631e-01 1.15604147e-01 6.87199831e-01
5.08904576e-01 -1.96850330e-01 -1.06906044e+00 3.32922548e-01
1.84722707e-01 1.13872349e+00 -7.46087134e-01 -1.29685327e-01
-9.14876819e-01 1.09675407e+00 -2.66501963e-01 5.18360317e-01
-4.25318897e-01 -6.04965575e-02 6.26350403e-01 4.97150600e-01
-9.91975218e-02 -8.05915415e-01 -9.33161601e-02 -1.41854572e+00
-5.89726083e-02 -2.40435749e-01 1.64786637e-01 -1.27214563e+00
-8.91830266e-01 6.29494965e-01 3.31916869e-01 -1.73389757e+00
-2.80900121e-01 -3.89884621e-01 -2.47658998e-01 1.07922471e+00
-1.86418104e+00 -9.40982878e-01 -8.94781351e-01 1.19483447e+00
6.74710751e-01 5.04679903e-02 5.01998484e-01 5.76119661e-01
4.31574211e-02 -2.09321275e-01 3.05486560e-01 -3.80098373e-01
6.46929622e-01 -7.41372705e-01 3.69081832e-02 1.29678357e+00
1.32883906e-01 7.21008539e-01 8.11384678e-01 -6.59401059e-01
-1.85113692e+00 -6.19313598e-01 6.21747494e-01 -5.42577624e-01
2.70926088e-01 -4.65609655e-02 -6.74933136e-01 7.38603532e-01
3.65157396e-01 4.66028184e-01 5.19407749e-01 -4.22197640e-01
6.84765354e-02 -1.86767131e-01 -1.02201521e+00 3.21065634e-01
1.06259406e+00 -7.94599235e-01 -2.37714559e-01 2.76660949e-01
2.09897578e-01 -6.99962378e-01 -7.27011621e-01 1.20942876e-01
7.49036252e-01 -1.43016362e+00 1.25560629e+00 9.92340073e-02
4.75201339e-01 -5.17930329e-01 -4.02670532e-01 -9.88509953e-01
-3.38884681e-01 -9.72609460e-01 -6.61631823e-02 7.89703965e-01
-5.12247205e-01 -6.28333747e-01 7.98473001e-01 5.20425260e-01
-1.31213322e-01 -1.49715215e-01 -8.51855278e-01 -5.10603011e-01
-8.22065592e-01 -3.67230475e-01 -1.10837281e-01 1.21459186e+00
-1.35713935e-01 4.78999376e-01 -5.95781386e-01 3.57663721e-01
1.03487909e+00 3.77582222e-01 9.04167771e-01 -1.20225036e+00
-2.88323253e-01 1.06676221e-01 -5.59386671e-01 -1.39829361e+00
-2.30602473e-02 -6.26484692e-01 -1.42958805e-01 -1.56277740e+00
1.47641242e-01 -4.81446296e-01 3.00439805e-01 -4.62543100e-01
2.32817411e-01 6.07633829e-01 9.22492594e-02 5.29608727e-01
-1.52761847e-01 2.84700513e-01 1.52881980e+00 3.01681638e-01
-3.32226068e-01 -2.46596336e-01 1.52556464e-01 5.74130118e-01
3.53715628e-01 -6.73314482e-02 -8.40270817e-01 -7.14431107e-01
3.89624566e-01 4.91755038e-01 5.07055521e-01 -9.38021541e-01
4.52068597e-01 -4.64509428e-01 2.80509025e-01 -8.97182882e-01
8.97932410e-01 -1.05568039e+00 6.95549726e-01 2.14657247e-01
1.97911203e-01 2.99033493e-01 1.99757680e-01 7.41391003e-01
-2.90852875e-01 -8.93958434e-02 9.13874626e-01 -2.34144077e-01
-1.10444176e+00 4.59163964e-01 -1.75605133e-01 -1.57516763e-01
1.07591355e+00 -7.05290198e-01 -3.18303257e-01 -5.03470421e-01
-3.29556644e-01 -3.26649338e-01 7.50904024e-01 1.16129875e-01
1.08639359e+00 -1.24219918e+00 -4.85465467e-01 6.04072332e-01
8.75980705e-02 -5.53425103e-02 6.13064468e-01 8.55436206e-01
-1.28572094e+00 5.60644269e-01 -5.69182515e-01 -1.14252663e+00
-1.65388203e+00 4.27556872e-01 3.18915635e-01 9.87962112e-02
-1.28755355e+00 5.17602026e-01 5.01057088e-01 3.20579797e-01
-3.04988444e-01 -2.05264583e-01 -1.34124849e-02 -4.38412815e-01
6.43521547e-01 7.40138054e-01 -1.10690789e-02 -9.44078207e-01
-2.72441953e-01 1.24792802e+00 2.68908411e-01 -4.20408845e-01
1.29736722e+00 -7.63741493e-01 -1.62387714e-01 2.22112358e-01
1.37824702e+00 3.32791209e-01 -1.23255086e+00 -2.46830031e-01
-6.29308939e-01 -1.54780447e+00 5.43238044e-01 -1.09318011e-01
-1.06810009e+00 1.04490578e+00 5.90786159e-01 -5.05841002e-02
1.40641153e+00 -4.02583003e-01 9.28634286e-01 6.86845332e-02
7.99181700e-01 -6.75266922e-01 -3.55367665e-03 4.20517802e-01
5.45402169e-01 -1.11797190e+00 3.27491432e-01 -8.57688308e-01
-1.25578940e-01 1.41583085e+00 4.48096812e-01 7.07213059e-02
7.67506063e-01 3.82921010e-01 -1.10601000e-01 -3.64411384e-01
-5.89616239e-01 4.14564228e-03 9.62047353e-02 7.25863159e-01
4.69667315e-01 -5.30221999e-01 -1.38026252e-01 -8.74125659e-01
1.49710968e-01 3.32525581e-01 1.01364028e+00 1.04673731e+00
-2.20521718e-01 -9.16555524e-01 -4.03977722e-01 -3.24546359e-02
-2.10342795e-01 -5.26423454e-02 3.20989996e-01 5.59923053e-01
-1.16929479e-01 9.79022384e-01 2.86112010e-01 -4.37927768e-02
2.25422546e-01 -4.86099333e-01 1.11484289e+00 -5.33654213e-01
-3.79619226e-02 3.28802884e-01 -6.69024810e-02 -9.67128336e-01
-1.16169906e+00 -5.97165048e-01 -9.56433296e-01 -6.48571074e-01
-2.22287536e-01 -1.40581816e-01 6.09974861e-01 5.67207754e-01
2.39460751e-01 1.55495256e-01 6.35152817e-01 -1.10874212e+00
4.17274266e-01 -4.14439827e-01 -6.61260426e-01 5.42638719e-01
7.17806339e-01 -8.17030489e-01 -3.47003400e-01 6.00371063e-01] | [9.558503150939941, -2.5265188217163086] |
ddcfbf17-a382-4f94-93d6-446557a9de18 | domain-adaptation-using-class-similarity-for | 2011.02782 | null | https://arxiv.org/abs/2011.02782v1 | https://arxiv.org/pdf/2011.02782v1.pdf | Domain Adaptation Using Class Similarity for Robust Speech Recognition | When only limited target domain data is available, domain adaptation could be used to promote performance of deep neural network (DNN) acoustic model by leveraging well-trained source model and target domain data. However, suffering from domain mismatch and data sparsity, domain adaptation is very challenging. This paper proposes a novel adaptation method for DNN acoustic model using class similarity. Since the output distribution of DNN model contains the knowledge of similarity among classes, which is applicable to both source and target domain, it could be transferred from source to target model for the performance improvement. In our approach, we first compute the frame level posterior probabilities of source samples using source model. Then, for each class, probabilities of this class are used to compute a mean vector, which we refer to as mean soft labels. During adaptation, these mean soft labels are used in a regularization term to train the target model. Experiments showed that our approach outperforms fine-tuning using one-hot labels on both accent and noise adaptation task, especially when source and target domain are highly mismatched. | ['Pengyuan Zhang', 'Li Wang', 'Yuling Ren', 'Jiangjiang Zhao', 'Han Zhu'] | 2020-11-05 | null | null | null | null | ['robust-speech-recognition'] | ['speech'] | [ 4.71066177e-01 -1.35663480e-01 -1.39484748e-01 -7.35305488e-01
-8.50982368e-01 -5.11175692e-01 4.18467253e-01 -1.16381474e-01
-6.15762651e-01 7.47155309e-01 3.40166211e-01 9.07945111e-02
3.69465172e-01 -6.41143680e-01 -7.44356573e-01 -8.24561059e-01
5.82972169e-01 3.59008133e-01 4.01906669e-01 -6.92059025e-02
-3.05300020e-02 5.70960231e-02 -1.18077493e+00 2.09753245e-01
7.23123670e-01 1.11205256e+00 6.89813554e-01 5.05321145e-01
-3.90284985e-01 5.02824724e-01 -9.21375334e-01 -1.46338254e-01
1.11758644e-02 -7.32331753e-01 -5.18482268e-01 -1.76853448e-01
2.89704740e-01 -3.99645492e-02 -1.34334834e-02 1.25378907e+00
8.09830725e-01 5.37518919e-01 6.88074946e-01 -8.12093556e-01
-6.44037545e-01 6.98301375e-01 -2.61328846e-01 3.04720163e-01
-7.52868950e-02 -1.18797250e-01 7.22939253e-01 -9.28494871e-01
4.20486867e-01 1.17857158e+00 3.57782245e-01 9.33925688e-01
-1.11337161e+00 -1.06921709e+00 4.09246653e-01 3.02801102e-01
-1.11698139e+00 -7.06858635e-01 1.12251270e+00 -1.74255982e-01
7.36284435e-01 -2.33104616e-01 3.13509971e-01 1.32925212e+00
-2.40062743e-01 7.15979218e-01 9.21299636e-01 -5.89648664e-01
4.51491207e-01 3.96142423e-01 -6.10337295e-02 -8.07547662e-03
-3.79743338e-01 1.61746025e-01 -6.39782608e-01 5.07210195e-02
5.03871918e-01 -3.93808842e-01 -1.10203385e-01 -6.28634095e-02
-1.00856233e+00 7.77244270e-01 4.71702702e-02 5.04382432e-01
-4.12820160e-01 -2.12286666e-01 4.27874625e-01 1.89899862e-01
6.04061842e-01 -1.07262395e-01 -7.49138653e-01 -2.94194490e-01
-6.77051842e-01 -1.35097265e-01 6.78975880e-01 7.91649580e-01
8.05422306e-01 4.81419414e-01 -9.33303908e-02 1.53700233e+00
4.31186885e-01 9.22773838e-01 8.35629582e-01 -8.07313859e-01
3.67477000e-01 1.82021305e-01 -2.48511747e-01 -6.76766574e-01
-1.78917617e-01 -4.61734444e-01 -8.23347986e-01 6.33245036e-02
3.79955471e-01 -4.66884822e-01 -1.05639923e+00 2.18586874e+00
4.40716475e-01 6.62763536e-01 4.82842028e-01 8.25075269e-01
8.12380791e-01 9.95933414e-01 4.29080606e-01 -4.20700371e-01
1.02830458e+00 -8.03191662e-01 -8.76352072e-01 -6.79739296e-01
5.98430276e-01 -1.06942689e+00 1.01575184e+00 2.21729964e-01
-9.02445674e-01 -8.55593622e-01 -8.43741655e-01 2.47527435e-01
-1.90246657e-01 1.53303474e-01 -9.04352069e-02 4.52616543e-01
-7.71901846e-01 1.39031857e-01 -6.15197480e-01 -3.30116510e-01
2.43835300e-01 2.91097283e-01 -1.36945382e-01 -4.80184741e-02
-1.43557906e+00 9.83244896e-01 8.05117786e-01 -1.96731836e-01
-1.14590085e+00 -6.43312633e-01 -7.85824180e-01 -6.89305291e-02
1.48845300e-01 -2.04871267e-01 1.46771407e+00 -1.21166527e+00
-1.88899601e+00 5.81219971e-01 -2.57550597e-01 -4.75028098e-01
-1.70210913e-01 -1.43901646e-01 -8.24386597e-01 -1.71086460e-01
-7.24554136e-02 6.87169254e-01 1.07158685e+00 -1.15998828e+00
-9.72941101e-01 -9.97611880e-02 -3.89992148e-01 4.67641950e-01
-5.80802798e-01 8.80490616e-02 -3.73194575e-01 -7.75076389e-01
7.59707540e-02 -8.16722810e-01 2.09148433e-02 -3.97126764e-01
-1.40331507e-01 -4.33841407e-01 1.05746782e+00 -7.97421455e-01
1.33789611e+00 -2.49445319e+00 1.01708844e-01 2.07538068e-01
-4.17017311e-01 6.65480614e-01 -2.19523624e-01 -1.56328619e-01
-2.34220754e-02 -2.80025095e-01 -3.40612024e-01 -1.75378948e-01
2.27769855e-02 4.69991654e-01 -1.95998371e-01 9.58247632e-02
2.67128497e-01 3.40170950e-01 -7.83007860e-01 -6.10087693e-01
2.56401241e-01 5.41294217e-01 -4.17399943e-01 6.17521167e-01
-3.71978551e-01 5.93955755e-01 -3.44835281e-01 2.12552577e-01
7.84891963e-01 3.60072792e-01 9.83530656e-02 -2.85040528e-01
3.08333337e-01 2.90987790e-01 -1.27449942e+00 1.69492006e+00
-8.32857370e-01 3.91410410e-01 1.81415737e-01 -1.35140634e+00
1.44743347e+00 4.40331757e-01 1.49316370e-01 -7.26258934e-01
4.27626744e-02 4.02621180e-01 1.13859430e-01 -2.93840706e-01
7.68895224e-02 -5.18363178e-01 -1.45352170e-01 2.35770762e-01
4.41487193e-01 -1.75911531e-01 -1.87587664e-01 -1.84169918e-01
6.26175046e-01 2.10124001e-01 1.69454664e-01 4.04653996e-02
6.58259213e-01 -2.35398963e-01 8.60902905e-01 3.74664009e-01
-3.10437500e-01 4.59577680e-01 6.85906038e-02 -1.12000167e-01
-9.08012152e-01 -1.18637502e+00 -1.09602503e-01 1.56377244e+00
-1.21462896e-01 -2.24761348e-02 -8.27735662e-01 -7.88416803e-01
-4.32063043e-01 8.67574751e-01 -2.92256057e-01 -5.44551432e-01
-7.99720883e-01 -6.33233428e-01 4.92137611e-01 6.49404228e-01
4.09119040e-01 -1.28478467e+00 -4.78551164e-02 5.88946879e-01
-3.61928463e-01 -1.17669690e+00 -6.21949673e-01 4.00167942e-01
-8.50940645e-01 -5.27789831e-01 -9.99530315e-01 -1.29866374e+00
3.81059110e-01 -1.29267871e-01 1.06359363e+00 -4.61453289e-01
4.73084182e-01 1.07123554e-01 -4.66587931e-01 -6.30322516e-01
-7.88197637e-01 -7.06382394e-02 3.00562620e-01 2.27422312e-01
7.80205190e-01 -6.66237473e-01 -2.36480102e-01 3.92031819e-01
-7.80601442e-01 -3.06085974e-01 5.93445063e-01 8.58060598e-01
8.59669209e-01 4.76675034e-02 9.88733828e-01 -6.88243210e-01
5.91015399e-01 -5.10680914e-01 -5.01763105e-01 2.57202089e-01
-2.82948077e-01 1.90859493e-02 7.82851815e-01 -9.55532730e-01
-1.63039446e+00 1.66556120e-01 -4.10764575e-01 -3.89540613e-01
-4.72671509e-01 4.91963506e-01 -5.09001315e-01 3.71738136e-01
7.67877340e-01 2.66223580e-01 -1.89415812e-01 -6.93090737e-01
2.63058275e-01 9.50789332e-01 6.28734350e-01 -5.03765047e-01
5.23574233e-01 -9.80754495e-02 -5.62020838e-01 -7.63307750e-01
-8.52938354e-01 -3.20930153e-01 -4.64380413e-01 -5.59170395e-02
9.69451070e-01 -9.44019794e-01 3.28024477e-02 7.25113750e-01
-1.24397707e+00 -4.62920219e-01 -3.02344620e-01 9.20605421e-01
-4.47093070e-01 9.13328975e-02 -2.84144878e-01 -7.87843823e-01
-8.48286226e-02 -1.00885618e+00 6.91104472e-01 3.42301577e-01
-1.35355741e-01 -1.33966386e+00 2.91597724e-01 -4.47256155e-02
5.91947913e-01 -3.73455465e-01 8.85890841e-01 -1.30831635e+00
9.24504027e-02 -3.63991670e-02 6.64595217e-02 1.00529885e+00
2.77357072e-01 -2.65888095e-01 -1.21531940e+00 -1.36445478e-01
3.46873224e-01 -3.08016568e-01 5.93435884e-01 6.09321654e-01
1.04667020e+00 -1.67907134e-01 -2.64328063e-01 4.69974726e-01
1.18191099e+00 4.95496750e-01 3.52698237e-01 1.46452531e-01
6.41518235e-01 5.22433758e-01 8.01930130e-01 3.10389459e-01
2.40412831e-01 8.98175538e-01 4.82746661e-02 2.12575123e-02
-3.49104166e-01 -2.41643831e-01 7.34238982e-01 1.30355775e+00
2.56552756e-01 -1.49602622e-01 -8.38914394e-01 6.37237191e-01
-1.56142104e+00 -7.29746938e-01 4.17061418e-01 2.26166368e+00
1.22634971e+00 2.79220343e-01 -1.04330108e-02 -1.61689550e-01
1.06027925e+00 -8.83653294e-03 -6.77527666e-01 -4.27934408e-01
-2.76703119e-01 2.61937380e-01 1.33341640e-01 4.61937159e-01
-1.01507556e+00 1.08506918e+00 5.91041803e+00 1.34965456e+00
-1.16155005e+00 3.45554322e-01 4.42387164e-01 6.38062656e-02
-9.90733355e-02 -5.18768132e-02 -1.16709530e+00 7.51920879e-01
1.45290387e+00 -7.50100911e-02 2.61696756e-01 9.11132932e-01
2.18648836e-01 1.13664985e-01 -9.50129032e-01 9.38163042e-01
6.82692826e-02 -8.48469019e-01 -1.16974436e-01 -1.92158580e-01
7.47738302e-01 -4.52745613e-03 -6.13242351e-02 6.01825297e-01
4.39685881e-01 -5.71287274e-01 5.01760602e-01 3.59560698e-01
6.43181503e-01 -8.54545653e-01 7.94008434e-01 4.59048152e-01
-1.03484178e+00 1.39110833e-01 -7.62553096e-01 3.91839683e-01
2.94007957e-01 5.64449906e-01 -9.87536132e-01 1.68088242e-01
6.38910532e-01 6.57023013e-01 -9.32645947e-02 7.78240919e-01
-2.09766507e-01 1.13661456e+00 -4.64530945e-01 2.01284871e-01
-2.23326273e-02 -1.37904048e-01 4.52017009e-01 1.36481035e+00
5.19133210e-01 -4.18866985e-02 2.92668104e-01 4.74513501e-01
-4.88808751e-02 2.27993041e-01 -4.65112507e-01 1.66882694e-01
8.70603442e-01 1.00446951e+00 -1.84814721e-01 -5.42854548e-01
-4.79031324e-01 8.94422472e-01 2.32637838e-01 5.68590581e-01
-8.97657692e-01 -3.30527753e-01 6.23676419e-01 -4.12736923e-01
3.75399739e-01 4.10028808e-02 -2.75217265e-01 -8.91502500e-01
-6.27932549e-02 -8.13371122e-01 3.82934958e-01 -5.92338920e-01
-1.42455137e+00 7.87980795e-01 -1.00654326e-02 -1.45971191e+00
-4.61180091e-01 -4.39576745e-01 -6.81007743e-01 1.14883852e+00
-1.67851484e+00 -1.00803721e+00 6.58883676e-02 8.69560182e-01
7.08022058e-01 -6.51496410e-01 9.99549508e-01 4.14256841e-01
-4.66958493e-01 9.07816112e-01 3.68980259e-01 2.22240999e-01
1.19524443e+00 -1.01778877e+00 3.10487524e-02 8.22327495e-01
9.87040177e-02 1.15311608e-01 6.42208695e-01 -6.12039506e-01
-5.86689234e-01 -1.13534772e+00 9.45988476e-01 -1.14193838e-02
4.72504586e-01 -1.42780364e-01 -1.23402548e+00 5.27918160e-01
1.56480074e-01 1.78567916e-01 9.22046065e-01 -2.08229618e-03
-2.07408220e-01 -2.99654335e-01 -1.00450826e+00 3.15413773e-01
6.35555923e-01 -5.68085730e-01 -8.05164754e-01 2.33897880e-01
9.64810789e-01 -4.33957428e-01 -8.70120049e-01 2.57058620e-01
1.91358030e-01 -6.00698173e-01 8.71079266e-01 -3.77156168e-01
5.69543466e-02 -2.58350521e-01 -5.14720678e-01 -1.78019738e+00
-1.69850886e-01 -1.65885493e-01 -2.93958504e-02 1.76807559e+00
6.05288684e-01 -4.00527656e-01 5.19087434e-01 4.17173415e-01
-4.57313597e-01 -1.84460536e-01 -1.07601583e+00 -8.14172149e-01
2.81532586e-01 -6.02805257e-01 5.35185158e-01 8.16563010e-01
-2.46984988e-01 4.71480638e-01 -4.32494462e-01 1.99404478e-01
3.69572908e-01 -2.92372614e-01 4.28272963e-01 -1.04397750e+00
-3.20333272e-01 -2.77818114e-01 -2.79393762e-01 -1.13308322e+00
4.52771604e-01 -7.83673823e-01 4.03323442e-01 -1.21559513e+00
-2.11435899e-01 -6.38698637e-01 -7.02211499e-01 5.42021394e-01
-2.93501794e-01 1.95537299e-01 1.63687333e-01 1.71284556e-01
-4.23018157e-01 5.91741681e-01 1.11382484e+00 -6.61461726e-02
-3.98991644e-01 3.02125931e-01 -5.29491842e-01 8.37206066e-01
9.52012241e-01 -7.61784971e-01 -4.78788167e-01 -5.88486373e-01
-3.59368116e-01 -7.57693872e-02 3.27256732e-02 -1.26982749e+00
2.62081474e-01 -3.13481212e-01 3.46295506e-01 -2.89758801e-01
5.21289110e-01 -9.12354946e-01 -1.38204470e-01 -9.65206127e-04
-4.97553974e-01 -5.09235859e-01 3.28753501e-01 5.68157732e-01
-6.08104467e-01 -4.39953268e-01 1.23952496e+00 1.02756917e-01
-1.06351626e+00 9.12673846e-02 -2.89632171e-01 3.15563768e-01
8.93283486e-01 -1.56069905e-01 1.06708836e-02 -4.62523907e-01
-1.03191710e+00 -1.91297233e-02 1.78490654e-01 5.16644180e-01
6.44827187e-01 -1.73059225e+00 -8.96000683e-01 3.81132334e-01
3.34151462e-02 1.49988815e-01 4.55599368e-01 5.07014573e-01
2.19482586e-01 1.17269211e-01 -2.19025254e-01 -7.26185143e-01
-1.08982658e+00 3.91597718e-01 2.93324709e-01 7.44482130e-03
-1.34221107e-01 1.22840071e+00 5.49986780e-01 -5.31506419e-01
4.08999413e-01 -1.47312105e-01 -3.99634153e-01 -6.23210426e-03
6.43830717e-01 2.55065225e-02 1.13030002e-01 -8.99517179e-01
-4.38181311e-01 7.25619197e-01 -2.44887825e-02 -3.37223679e-01
1.15085161e+00 -4.23391491e-01 2.17049077e-01 5.82251370e-01
1.29096496e+00 -6.50639161e-02 -1.49990368e+00 -9.15027678e-01
-4.26653423e-04 -7.29071647e-02 -1.03361700e-02 -1.07641327e+00
-1.06467843e+00 1.15533566e+00 9.21369672e-01 -2.23228559e-01
1.48758304e+00 1.57094195e-01 8.98082852e-01 3.83441858e-02
6.60965592e-03 -1.49785495e+00 5.30450493e-02 8.79364312e-01
6.30193472e-01 -1.28112650e+00 -6.63019121e-01 -3.28517072e-02
-1.01421618e+00 9.82182384e-01 8.80930364e-01 -1.00276217e-01
8.53526115e-01 2.37841636e-01 5.33171177e-01 4.69166517e-01
-5.98605573e-01 -8.41099918e-02 3.85847270e-01 1.07791173e+00
4.03636992e-01 5.01329042e-02 6.16409630e-02 8.70033205e-01
-2.12900043e-01 -5.83105385e-02 1.63021371e-01 6.79520428e-01
-7.74520874e-01 -1.46830034e+00 -5.47767758e-01 2.14952141e-01
-4.51648206e-01 -1.83215201e-01 -1.11354189e-02 2.25098610e-01
3.21089625e-01 1.05087113e+00 7.44572058e-02 -4.88640696e-01
3.70109171e-01 4.69785661e-01 1.10514268e-01 -7.99321711e-01
-1.87064514e-01 4.55195516e-01 -9.97297838e-03 -1.54102594e-01
-5.46376050e-01 -4.01827842e-01 -1.29273665e+00 2.16806337e-01
-3.97199571e-01 2.78626353e-01 7.90813088e-01 1.13309407e+00
1.60451993e-01 4.55946207e-01 5.54513752e-01 -6.23377800e-01
-6.78399563e-01 -1.24584746e+00 -5.69512069e-01 5.08843958e-01
2.37298429e-01 -5.85020661e-01 -2.91627526e-01 4.43439692e-01] | [14.394051551818848, 6.583807945251465] |
d56b29c4-224b-4978-8e8c-b08cfc0f6ba4 | twisty-a-multilingual-twitter-stylometry | null | null | https://aclanthology.org/L16-1258 | https://aclanthology.org/L16-1258.pdf | TwiSty: A Multilingual Twitter Stylometry Corpus for Gender and Personality Profiling | Personality profiling is the task of detecting personality traits of authors based on writing style. Several personality typologies exist, however, the Briggs-Myer Type Indicator (MBTI) is particularly popular in the non-scientific community, and many people use it to analyse their own personality and talk about the results online. Therefore, large amounts of self-assessed data on MBTI are readily available on social-media platforms such as Twitter. We present a novel corpus of tweets annotated with the MBTI personality type and gender of their author for six Western European languages (Dutch, German, French, Italian, Portuguese and Spanish). We outline the corpus creation and annotation, show statistics of the obtained data distributions and present first baselines on Myers-Briggs personality profiling and gender prediction for all six languages. | ['Walter Daelemans', 'Ben Verhoeven', 'Barbara Plank'] | 2016-05-01 | twisty-a-multilingual-twitter-stylometry-1 | https://aclanthology.org/L16-1258 | https://aclanthology.org/L16-1258.pdf | lrec-2016-5 | ['gender-prediction'] | ['computer-vision'] | [-7.15394735e-01 1.56944484e-01 -2.93727338e-01 -4.08234954e-01
-1.42803520e-01 -7.21051872e-01 9.15154636e-01 6.00014091e-01
-6.79849386e-01 7.79706895e-01 4.06476378e-01 1.44584496e-02
-3.00202668e-02 -6.73021913e-01 3.52017879e-01 -5.06244540e-01
-1.55673444e-01 8.57452750e-01 -2.66290158e-01 -1.40031278e-01
5.20001650e-01 1.88267097e-01 -1.47560918e+00 -1.32995680e-01
9.05114055e-01 8.73321533e-01 -2.97727846e-02 6.78995013e-01
-1.48087963e-01 6.63685679e-01 -8.20952833e-01 -1.11513484e+00
-3.40551406e-01 -1.44610539e-01 -9.16829824e-01 2.27811143e-01
3.46314281e-01 1.46401852e-01 3.76077555e-02 1.19857097e+00
5.00620544e-01 -1.17791779e-01 9.36176598e-01 -1.07174194e+00
-4.99995500e-01 6.95824444e-01 -8.71939659e-01 1.15470447e-01
5.34612715e-01 -9.96506885e-02 1.10732305e+00 -6.54645205e-01
7.69857049e-01 1.45524633e+00 5.93250632e-01 4.91758764e-01
-1.07664716e+00 -5.95169604e-01 -1.88916117e-01 9.23498347e-02
-1.37249243e+00 -3.69608521e-01 7.40284443e-01 -7.92877555e-01
3.25305462e-01 4.18479264e-01 4.52005535e-01 1.60676980e+00
2.48945892e-01 5.05288661e-01 1.82022035e+00 -3.21787536e-01
1.53598366e-02 9.12333131e-01 3.23710799e-01 6.98342621e-01
3.95481400e-02 -4.33308721e-01 -8.26392770e-01 -4.69464839e-01
4.40056652e-01 -5.91778755e-01 4.77544188e-01 3.34668189e-01
-1.20766258e+00 6.29215181e-01 -4.02182788e-01 3.07018042e-01
-1.37839988e-01 -6.51898742e-01 1.03817713e+00 3.65965575e-01
9.09520388e-01 5.98810732e-01 -3.25504303e-01 -7.16560841e-01
-9.97610331e-01 5.63930690e-01 1.02392447e+00 7.60115862e-01
5.71183562e-01 -2.10862458e-01 -2.76639730e-01 1.38231599e+00
2.33422861e-01 3.37951392e-01 9.18893516e-01 -9.00922477e-01
3.40668797e-01 5.56249678e-01 2.80713648e-01 -1.33078742e+00
-7.38849998e-01 -3.12077999e-01 -9.24567997e-01 -1.38298750e-01
6.15158260e-01 -2.78364241e-01 1.33831993e-01 1.44979060e+00
8.32997262e-02 -5.89969635e-01 -2.12199286e-01 6.82964921e-01
1.19211209e+00 8.78309011e-02 1.63339734e-01 -4.18703675e-01
1.85043478e+00 -5.00384986e-01 -6.10054255e-01 -1.41975105e-01
3.31993937e-01 -6.92651927e-01 1.13571358e+00 4.57372397e-01
-1.24782801e+00 -7.92153597e-01 -4.90657747e-01 7.98601657e-03
-3.95366848e-01 4.33176219e-01 5.13998747e-01 1.12260079e+00
-9.56509352e-01 6.93925321e-01 -3.72157812e-01 -4.69057590e-01
3.32771361e-01 2.31417209e-01 -5.04196525e-01 6.70016766e-01
-9.84002829e-01 6.30622149e-01 4.51756977e-02 -6.02188230e-01
-1.07846595e-01 -3.26107830e-01 -5.34946144e-01 -1.20921552e-01
-2.93793559e-01 -4.18580204e-01 1.02468872e+00 -9.44752574e-01
-1.74867141e+00 1.81045675e+00 -2.47090131e-01 -1.28802881e-01
7.54190385e-01 2.86283582e-01 -7.31475711e-01 -2.23403230e-01
4.94564980e-01 3.57138306e-01 5.79520464e-01 -7.86256909e-01
-9.16111171e-01 -5.06509423e-01 -4.98498440e-01 -2.69622624e-01
-8.34901512e-01 6.94472551e-01 -1.85764376e-02 -5.70462227e-01
-2.62259841e-01 -7.63573825e-01 1.55866370e-01 -9.59588766e-01
-7.14220107e-01 -9.50917006e-01 1.70453846e-01 -8.46073508e-01
1.39883566e+00 -2.13057971e+00 3.32302717e-03 2.76534498e-01
5.81982970e-01 -2.09690168e-01 3.64081264e-01 4.44278806e-01
2.81515270e-01 2.26592675e-01 4.36781675e-01 -8.66096497e-01
5.16723812e-01 -1.36973456e-01 -2.87831463e-02 5.13570070e-01
-1.40096173e-01 5.36354065e-01 -1.03098726e+00 -8.54595661e-01
-3.31640005e-01 -1.59058467e-01 -3.13104212e-01 2.77835160e-01
1.31649882e-01 7.87933052e-01 -1.64387524e-01 7.20255375e-01
5.73059618e-01 2.38563791e-02 4.99406457e-01 3.76272738e-01
-7.15163708e-01 5.88620245e-01 -5.85114241e-01 6.91502333e-01
-2.52979368e-01 8.40571165e-01 1.11524627e-01 -3.59706074e-01
1.55048740e+00 -2.55366620e-02 4.28828925e-01 -3.45693022e-01
4.18051273e-01 4.06869471e-01 2.14682370e-01 -3.66955429e-01
9.69694734e-01 -1.70659080e-01 -6.78604543e-01 4.33584690e-01
-8.26041028e-02 2.77716249e-01 7.98107266e-01 -8.89447927e-02
7.50755191e-01 -1.98512115e-02 4.50952888e-01 -7.14325786e-01
7.00549603e-01 -2.68179178e-01 7.50026524e-01 5.45093000e-01
-5.63874900e-01 2.17843398e-01 1.17555797e+00 -2.98489690e-01
-1.05144799e+00 -6.77885354e-01 -4.60587233e-01 1.58765614e+00
-5.00319302e-01 -7.55925477e-01 -7.63899088e-01 -5.56193471e-01
1.63049072e-01 2.85408974e-01 -6.18934095e-01 2.79955715e-01
-9.33481753e-02 -5.93892753e-01 9.05929923e-01 1.65640324e-01
6.67434633e-01 -1.16729987e+00 -2.51627564e-01 -3.88685949e-02
-1.94191739e-01 -1.15671051e+00 -3.08990389e-01 -2.32633829e-01
-5.18834770e-01 -6.26629293e-01 -6.69560671e-01 -6.95864260e-01
3.02031785e-01 -4.50327277e-01 1.47911799e+00 -2.11737931e-01
2.31791198e-01 2.49176264e-01 -2.44109780e-01 -4.71635669e-01
-6.98660016e-01 5.05101621e-01 5.99381447e-01 1.33963630e-01
7.31359363e-01 -5.26985526e-01 -1.27319381e-01 3.44319493e-01
1.24902122e-01 -2.77632177e-01 2.38304466e-01 5.09438515e-01
-1.14932001e-01 2.63919622e-01 6.32361829e-01 -1.17336130e+00
7.82652020e-01 -5.83442807e-01 -1.10432729e-01 -3.64275962e-01
-5.22347927e-01 -5.00262678e-01 6.30186856e-01 -3.43112260e-01
-1.05314696e+00 -4.86507267e-01 -3.41751456e-01 2.61980385e-01
-4.71403599e-01 5.52768826e-01 -2.11684063e-01 1.98185250e-01
4.47021544e-01 2.86716484e-02 -5.79995252e-02 -6.28275692e-01
-3.76244158e-01 1.38885629e+00 7.03159451e-01 -8.17890286e-01
4.68320936e-01 8.40185359e-02 -4.59700599e-02 -1.18717468e+00
-9.52061653e-01 -5.61369300e-01 -9.88715589e-01 -4.38328624e-01
6.53106570e-01 -8.20736945e-01 -1.57439065e+00 6.80612445e-01
-8.36946726e-01 -1.04374021e-01 1.52971357e-01 1.74068213e-01
-4.37144756e-01 4.88770753e-01 -8.05180252e-01 -1.02371287e+00
-3.72745425e-01 -7.48446703e-01 8.98631930e-01 2.36574247e-01
-1.25625789e+00 -1.39916515e+00 1.45179167e-01 3.40171993e-01
8.60340719e-04 1.12146214e-01 8.54698360e-01 -5.85788906e-01
6.96741104e-01 -2.29084715e-01 -2.55059928e-01 1.53663769e-01
-2.31518865e-01 1.37365684e-01 -9.25768197e-01 -5.96731156e-02
-1.53372675e-01 -3.39815050e-01 6.34248197e-01 -5.29966801e-02
1.32674468e+00 -5.75895369e-01 -5.76374978e-02 2.87173212e-01
9.95573163e-01 -3.78294468e-01 3.64735007e-01 6.82571292e-01
6.49352849e-01 1.01925361e+00 4.25722182e-01 1.15153241e+00
9.96182680e-01 5.86707652e-01 -1.25264779e-01 3.15983176e-01
4.83007938e-01 -1.14430360e-01 7.19700038e-01 1.13064337e+00
-6.69766009e-01 6.45681322e-02 -9.29542065e-01 4.05477107e-01
-1.58051538e+00 -1.02769232e+00 -5.63161016e-01 2.18284869e+00
1.15121603e+00 2.48130336e-01 1.24887085e+00 2.85850197e-01
6.11885428e-01 1.10986568e-01 3.48683670e-02 -9.02076185e-01
-4.31531310e-01 -6.82729930e-02 4.23645169e-01 1.49923116e-01
-1.07681346e+00 7.76425421e-01 6.34809446e+00 5.90181112e-01
-9.29694772e-01 7.82687590e-02 8.63273740e-01 1.87030241e-01
2.30771318e-01 -4.30105716e-01 -1.22922873e+00 7.70524383e-01
1.29123712e+00 -3.47746283e-01 1.75627261e-01 8.77500415e-01
2.97848314e-01 -3.00172359e-01 -9.42297220e-01 1.19529867e+00
-1.78612042e-02 -6.67710304e-01 -7.26284564e-01 5.13112485e-01
6.42912924e-01 -1.89735666e-01 2.37147301e-01 4.82028395e-01
2.60799468e-01 -7.64632225e-01 9.30372477e-01 4.56891954e-01
8.69565666e-01 -8.76721621e-01 6.93829000e-01 4.84821409e-01
-5.88971972e-01 -3.58848900e-01 -5.79322278e-01 -6.03240073e-01
-3.75446975e-01 7.93355227e-01 -7.82494128e-01 1.31943032e-01
7.84604728e-01 9.40826893e-01 -1.03326356e+00 5.63570082e-01
1.22118287e-03 8.44267964e-01 2.49687687e-01 -4.25123543e-01
-7.79875442e-02 -5.18961847e-01 5.48246682e-01 1.50991416e+00
3.69216025e-01 -5.20957649e-01 2.70522475e-01 7.80581951e-01
-5.09517007e-02 4.76296633e-01 -1.74290180e-01 -4.56878722e-01
3.82033497e-01 1.98766267e+00 -7.78804779e-01 -3.81624728e-01
-4.28829610e-01 8.11460078e-01 6.14768684e-01 -1.98722348e-01
-3.85442525e-01 -1.38907954e-01 7.43485332e-01 4.60051596e-01
-3.89528245e-01 -3.05045426e-01 -8.68996799e-01 -1.02144682e+00
-3.49505275e-01 -8.93790841e-01 4.65165854e-01 -3.12708974e-01
-1.91328073e+00 2.93625563e-01 -4.07515258e-01 -9.88391042e-01
-2.30155602e-01 -9.42552328e-01 -7.01515019e-01 1.08313310e+00
-8.42976451e-01 -1.05801535e+00 -4.31154728e-01 1.20131746e-01
1.79819182e-01 -7.77545512e-01 9.85501349e-01 3.15692008e-01
-8.28036964e-01 7.43171096e-01 1.99831292e-01 3.52995068e-01
1.13850927e+00 -1.72950602e+00 2.79840648e-01 -1.35212928e-01
-4.05948997e-01 7.29237080e-01 9.62165833e-01 -8.55320811e-01
-9.64842081e-01 -7.42067218e-01 1.59070957e+00 -7.94404745e-01
1.01534331e+00 -5.76946557e-01 -4.33198601e-01 4.82647598e-01
3.93705636e-01 -7.10092485e-01 8.44168305e-01 8.78326952e-01
-2.54953980e-01 1.75920740e-01 -1.06773221e+00 5.70454478e-01
9.02166486e-01 -4.92841780e-01 -2.12116525e-01 4.63762522e-01
-4.00235474e-01 -1.58682823e-01 -1.45663750e+00 -2.05009639e-01
8.01686168e-01 -1.37498999e+00 5.63584387e-01 -3.82839650e-01
1.08164072e+00 3.44311655e-01 3.18986565e-01 -1.51920021e+00
-8.03050160e-01 -7.39767373e-01 2.43213698e-01 1.82907069e+00
4.01523001e-02 -7.17486441e-01 7.67458677e-01 5.09441555e-01
-1.54045792e-02 -4.94615316e-01 -7.28791237e-01 -7.74179757e-01
4.63973433e-01 2.52613090e-02 5.55746913e-01 1.17982590e+00
7.89328516e-01 6.18294299e-01 -4.12666172e-01 -5.04607856e-01
5.16827524e-01 -2.07265783e-02 1.08438385e+00 -1.94486856e+00
-1.75282627e-01 -1.00406611e+00 -5.47443569e-01 -3.56640339e-01
7.05618858e-01 -8.91009450e-01 -4.64960098e-01 -1.04564774e+00
5.86718976e-01 -4.81682092e-01 2.90176034e-01 4.99816015e-02
-1.97938830e-01 3.38916838e-01 -2.80677319e-01 2.94457078e-01
-5.71929693e-01 2.91036796e-02 1.06500125e+00 7.05178156e-02
1.15694096e-02 1.49816006e-01 -7.65208483e-01 9.71573234e-01
8.67015958e-01 -6.38016984e-02 2.86786526e-01 4.30753112e-01
6.86476171e-01 -9.15530920e-02 1.52960628e-01 -7.95546770e-01
-5.15222475e-02 -2.73210943e-01 6.66962385e-01 -3.96646559e-01
1.87879696e-01 -8.11754987e-02 -3.74618232e-01 4.10584770e-02
-2.03669578e-01 1.84728920e-01 -2.17343152e-01 -1.07624875e-02
-1.33494005e-01 -4.06424075e-01 9.25924420e-01 -3.87852013e-01
-2.48861775e-01 1.82456508e-01 -6.77239597e-01 1.48423299e-01
5.88451266e-01 -1.67179778e-01 -4.75772262e-01 -2.74365753e-01
-6.64660990e-01 -9.33082402e-02 8.46461236e-01 2.50848144e-01
-3.80276628e-02 -1.24393499e+00 -9.67344165e-01 1.53689861e-01
1.77293122e-01 -9.85744119e-01 1.16407432e-01 1.35484588e+00
1.22412704e-02 2.87790328e-01 -6.50027275e-01 -2.11994544e-01
-1.52509439e+00 3.33792984e-01 -1.68223679e-02 -3.85690510e-01
-2.08123654e-01 6.65907919e-01 -3.77691202e-02 -5.91128767e-01
-1.03932291e-01 6.27643585e-01 -9.92104530e-01 8.46919358e-01
4.73260641e-01 1.00343156e+00 -5.26288413e-02 -1.21298170e+00
-2.84653455e-01 -1.20246015e-01 2.86124535e-02 -3.40124369e-01
1.01708794e+00 -4.50418115e-01 -7.94845939e-01 1.09271944e+00
9.03982937e-01 7.56561816e-01 -3.56191814e-01 1.24660237e-02
3.96719128e-01 -6.02440596e-01 -3.89595032e-01 -3.91720623e-01
-3.01350862e-01 7.56258607e-01 -1.52790500e-02 8.66593421e-01
5.08593738e-01 -4.73099202e-02 5.92149317e-01 -1.60497800e-01
4.45428908e-01 -1.83577633e+00 -2.68689860e-02 1.04703498e+00
6.42268717e-01 -1.03856361e+00 4.98319715e-02 -2.93258816e-01
-1.01794541e+00 1.09276307e+00 6.25914931e-01 1.27031326e-01
6.24267578e-01 -8.13861564e-02 6.79481328e-02 -7.40501806e-02
-6.08888686e-01 -1.61861867e-01 3.73890162e-01 5.92440844e-01
1.17806518e+00 4.96222883e-01 -7.69298673e-01 1.23058343e+00
-1.20821178e+00 -4.80248660e-01 8.34213257e-01 2.67786622e-01
-3.17611754e-01 -1.29000092e+00 -5.67494571e-01 9.57600117e-01
-9.58636284e-01 2.23444074e-01 -9.41962719e-01 4.52761352e-01
1.65882409e-01 1.00777936e+00 3.73385847e-01 -5.47278702e-01
2.59359255e-02 1.48753881e-01 5.68162501e-01 -6.49988115e-01
-9.93785024e-01 1.26608402e-01 7.55978286e-01 3.14215213e-01
-2.86125243e-01 -1.29619193e+00 -6.67132437e-01 -1.07569623e+00
3.18420261e-01 2.83444136e-01 4.19198900e-01 8.67702782e-01
1.36652231e-01 1.32136717e-01 1.01639926e+00 -9.48671281e-01
-1.98370859e-01 -1.48183513e+00 -1.29218483e+00 4.80376452e-01
-2.24020883e-01 -6.97181344e-01 -4.02242810e-01 1.95019934e-02] | [9.388717651367188, 10.326961517333984] |
076b3164-7d49-47dc-b011-c4072a6509dc | a-safe-semi-supervised-graph-convolution | 2207.01960 | null | https://arxiv.org/abs/2207.01960v1 | https://arxiv.org/pdf/2207.01960v1.pdf | A Safe Semi-supervised Graph Convolution Network | In the semi-supervised learning field, Graph Convolution Network (GCN), as a variant model of GNN, has achieved promising results for non-Euclidean data by introducing convolution into GNN. However, GCN and its variant models fail to safely use the information of risk unlabeled data, which will degrade the performance of semi-supervised learning. Therefore, we propose a Safe GCN framework (Safe-GCN) to improve the learning performance. In the Safe-GCN, we design an iterative process to label the unlabeled data. In each iteration, a GCN and its supervised version(S-GCN) are learned to find the unlabeled data with high confidence. The high-confidence unlabeled data and their pseudo labels are then added to the label set. Finally, both added unlabeled data and labeled ones are used to train a S-GCN which can achieve the safe exploration of the risk unlabeled data and enable safe use of large numbers of unlabeled data. The performance of Safe-GCN is evaluated on three well-known citation network datasets and the obtained results demonstrate the effectiveness of the proposed framework over several graph-based semi-supervised learning methods. | ['Zhiwei Ye', 'Jing Zhao', 'Haitao Gan', 'Yadong Yan', 'Zhi Yang'] | 2022-07-05 | null | null | null | null | ['safe-exploration'] | ['robots'] | [-1.20746061e-01 3.47014934e-01 -3.31302106e-01 -5.23137450e-01
-2.47350469e-01 -3.00872684e-01 4.82588291e-01 7.24872798e-02
-3.81126285e-01 7.81021893e-01 -6.48742020e-02 -4.66856480e-01
-3.56864423e-01 -1.06004441e+00 -5.04923820e-01 -7.51278996e-01
-1.11137360e-01 5.15913665e-01 2.19688445e-01 2.90888697e-01
1.42711475e-01 3.23081970e-01 -1.22390234e+00 -2.24939600e-01
1.26477194e+00 9.39479649e-01 1.11516006e-01 1.09280109e-01
-3.55530292e-01 6.40019536e-01 -4.51920219e-02 -2.10608035e-01
4.34640855e-01 -3.81472766e-01 -6.83456957e-01 -9.70293134e-02
4.57886197e-02 -1.15185514e-01 -4.25284505e-01 1.37917495e+00
2.39866838e-01 3.97913992e-01 9.42740500e-01 -1.33066821e+00
-7.25027323e-01 8.73355925e-01 -5.43368697e-01 -3.20597440e-02
-2.70632535e-01 -1.66054413e-01 8.01700354e-01 -9.76634443e-01
5.55618703e-01 1.20933139e+00 5.21931291e-01 4.19343710e-01
-8.94049883e-01 -8.54723513e-01 2.24710077e-01 -4.13334966e-02
-1.50164330e+00 1.41853362e-01 8.69720042e-01 -2.63007015e-01
4.22659248e-01 -1.96524858e-01 5.03705025e-01 7.08755255e-01
-2.47179195e-01 7.90856719e-01 1.13803458e+00 -5.17573416e-01
3.60051751e-01 7.14023933e-02 3.86987478e-01 1.03131950e+00
3.18229795e-01 3.99459958e-01 -2.67458498e-01 -8.02291930e-02
7.89765060e-01 2.06578955e-01 -1.75503686e-01 -3.18008721e-01
-1.00447428e+00 8.85011256e-01 1.08003700e+00 2.58750856e-01
-1.73056483e-01 -1.79948673e-01 2.00242743e-01 1.77045584e-01
6.86707973e-01 1.02806374e-01 -1.82495907e-01 1.81294486e-01
-8.39442492e-01 -2.90701777e-01 7.70438492e-01 1.04749966e+00
9.60184634e-01 -6.82994574e-02 -5.51184155e-02 8.22006285e-01
6.79990053e-01 3.13997954e-01 3.12644750e-01 -3.92856210e-01
4.94138747e-01 1.24872553e+00 -3.25374097e-01 -9.18148756e-01
-4.03378367e-01 -6.35820925e-01 -1.04913306e+00 1.81919277e-01
3.21023911e-01 -3.65606815e-01 -1.28963208e+00 1.47333086e+00
3.31903428e-01 4.83765155e-01 1.80344388e-01 9.75851417e-01
1.00049436e+00 7.18392730e-01 6.70397505e-02 -3.07367653e-01
7.48820603e-01 -1.29058087e+00 -5.54117560e-01 -6.13489747e-02
9.62684691e-01 -3.86441916e-01 7.91002512e-01 4.74607080e-01
-3.52701545e-01 -5.88840544e-01 -1.22754610e+00 1.73029825e-01
-5.17524719e-01 2.83900917e-01 6.80676520e-01 4.54562306e-01
-9.05819476e-01 8.27576518e-01 -6.33389235e-01 -3.33031982e-01
6.65527582e-01 3.66114736e-01 -3.29424590e-01 -4.62004364e-01
-1.34261227e+00 6.25874639e-01 1.05492461e+00 4.41679388e-01
-9.55697000e-01 -2.30664641e-01 -7.85878837e-01 8.18225592e-02
6.91519916e-01 -8.71104598e-02 8.21736276e-01 -8.02463055e-01
-1.19859731e+00 4.50168550e-01 4.01123136e-01 -2.57828772e-01
4.27400440e-01 4.97163869e-02 -5.87665081e-01 1.73736382e-02
8.94680321e-02 5.04748642e-01 6.15771651e-01 -1.20781922e+00
-4.09669429e-01 -2.69831926e-01 -1.61499649e-01 4.98340547e-01
-5.88766217e-01 -4.01929498e-01 -4.96628940e-01 -6.17575705e-01
4.02817070e-01 -8.83785725e-01 -3.00702214e-01 2.23823972e-02
-6.27722085e-01 -5.41983783e-01 1.02281284e+00 -2.30158255e-01
1.23071265e+00 -2.16312528e+00 -8.15662146e-02 7.38153756e-01
5.10351002e-01 6.10586524e-01 -1.49410665e-01 2.03139819e-02
-1.19547844e-01 1.24648549e-01 -3.51869285e-01 -1.36719093e-01
-2.68696427e-01 3.43911886e-01 2.27022231e-01 3.93279701e-01
1.38721079e-01 9.82383847e-01 -1.20315349e+00 -8.49658787e-01
2.77195901e-01 1.31040171e-01 -7.98519999e-02 2.98272818e-01
-3.02171499e-01 3.49229723e-01 -7.72717834e-01 5.28729558e-01
7.55937457e-01 -5.39224625e-01 7.29122013e-02 -3.90244834e-02
1.11767508e-01 -1.28674001e-01 -1.31430781e+00 1.39614868e+00
-1.54751152e-01 1.05659574e-01 -2.44444743e-01 -9.97853577e-01
1.42003083e+00 1.04506023e-01 1.93038136e-01 -3.47429067e-01
2.11176157e-01 4.19414788e-01 9.61446464e-02 -3.58574420e-01
-4.50809896e-02 1.62713341e-02 3.32344949e-01 7.07665205e-01
3.37227553e-01 2.66187459e-01 7.13837668e-02 4.47769195e-01
8.23898435e-01 2.39979550e-01 1.67572379e-01 -4.82329071e-01
7.59915888e-01 -2.18653947e-01 5.85530877e-01 6.98880136e-01
-1.33245856e-01 3.92493755e-01 4.54503328e-01 -4.56806779e-01
-6.91369832e-01 -1.00658870e+00 -9.92730856e-02 8.24042559e-01
3.69820684e-01 -3.09804589e-01 -7.17049479e-01 -1.43860877e+00
-5.20585105e-02 5.95058501e-01 -6.41740501e-01 -4.68561918e-01
-1.26658782e-01 -6.88356817e-01 3.68698567e-01 4.59297687e-01
8.07565868e-01 -1.17809308e+00 2.27748811e-01 9.26985070e-02
1.76084265e-01 -6.61543548e-01 -3.56341153e-01 3.93407464e-01
-9.72293139e-01 -1.25158572e+00 -5.26796818e-01 -9.90150332e-01
1.15818167e+00 1.61391661e-01 5.19757092e-01 3.00791591e-01
2.29296684e-01 -3.21839839e-01 -5.04741907e-01 -1.69549406e-01
-4.33621556e-01 2.00826183e-01 1.18320651e-01 2.10285336e-01
5.57647645e-01 -5.16517580e-01 -4.35534656e-01 5.08081734e-01
-7.48124599e-01 -8.98427854e-04 6.53406978e-01 9.02394712e-01
5.80115259e-01 2.39624232e-01 8.44820976e-01 -1.25544107e+00
7.01295972e-01 -6.80454254e-01 -7.57974386e-01 4.28497940e-01
-1.35389113e+00 2.68869340e-01 9.17123437e-01 -4.53724742e-01
-1.02508152e+00 -9.40871015e-02 1.60093337e-01 -6.67851448e-01
6.92713112e-02 9.84220862e-01 -2.14793980e-01 -2.67416358e-01
7.50199020e-01 5.80144338e-02 2.39730060e-01 -5.81471860e-01
3.98782313e-01 9.04982507e-01 3.90052348e-01 -5.58682024e-01
8.74922395e-01 2.97753680e-02 1.87973455e-01 -3.23752314e-01
-1.07832575e+00 -4.41141456e-01 -7.19798326e-01 -3.47186685e-01
6.69635952e-01 -7.37130940e-01 -3.53582263e-01 5.46903491e-01
-6.78966224e-01 -2.28775829e-01 -8.24536756e-02 7.74444461e-01
-1.10174626e-01 6.15038216e-01 -3.62135828e-01 -8.50258768e-01
-3.61217022e-01 -1.10988867e+00 6.25759184e-01 6.25710070e-01
3.25795501e-01 -1.34054828e+00 -2.43241295e-01 6.18214207e-03
2.66661853e-01 4.91343886e-02 9.23609316e-01 -1.35859263e+00
-3.34353745e-01 -4.79811579e-01 -5.58424830e-01 6.45310640e-01
1.41304567e-01 -4.57276739e-02 -7.71142542e-01 -3.43681276e-01
-1.48308158e-01 -6.66032076e-01 9.08383906e-01 1.46715745e-01
1.27484643e+00 -1.25954628e-01 -5.88700891e-01 6.72853291e-01
1.45186138e+00 1.51723742e-01 4.09312129e-01 1.51865810e-01
9.72378373e-01 3.72511148e-01 8.99219990e-01 8.40629339e-02
5.28040230e-02 6.41566962e-02 5.69005430e-01 -1.55035943e-01
1.75844684e-01 -5.97818673e-01 -1.97112504e-02 1.13262367e+00
-1.92892566e-01 -4.04207855e-01 -9.92252767e-01 1.58202767e-01
-2.10896420e+00 -4.37771887e-01 -3.17896962e-01 2.12908077e+00
6.92285478e-01 5.10397553e-01 -1.83739811e-01 1.58699751e-01
1.07059681e+00 5.85484840e-02 -6.95838749e-01 -1.49542745e-03
-4.08667140e-02 6.24616146e-02 5.21904290e-01 4.03770775e-01
-1.30717027e+00 9.52786326e-01 5.87050295e+00 1.10472262e+00
-9.83704388e-01 -7.36453459e-02 7.20720589e-01 3.74488354e-01
-3.27332914e-01 1.55097425e-01 -7.35206187e-01 4.33408976e-01
7.93429852e-01 -2.14450836e-01 5.00577390e-01 1.19381070e+00
3.94048840e-02 -5.67253008e-02 -9.66382444e-01 8.33015442e-01
-1.99432299e-03 -1.15796363e+00 2.04394713e-01 -5.69806024e-02
9.69629467e-01 1.05671346e-01 -1.89430341e-01 4.67780977e-01
6.01220846e-01 -1.11245644e+00 1.99307099e-01 4.59422231e-01
8.77632022e-01 -7.68389583e-01 9.97798264e-01 6.12962842e-01
-1.16740203e+00 -5.94657399e-02 -5.40818989e-01 3.45531404e-02
-1.05179906e-01 8.76663148e-01 -7.64772475e-01 9.78375852e-01
3.86715919e-01 1.12432134e+00 -7.95818031e-01 1.07944477e+00
-5.21865904e-01 7.93352008e-01 -3.51113409e-01 -2.59810328e-01
6.15835667e-01 -6.93835497e-01 1.68973953e-01 7.30282605e-01
2.12222725e-01 1.44193128e-01 5.34061909e-01 9.84506667e-01
-3.25107604e-01 3.85262698e-01 -6.72092378e-01 -1.91253498e-01
5.61024308e-01 1.44647777e+00 -9.39365625e-01 -3.43997926e-01
-4.58910137e-01 5.74480653e-01 7.53201008e-01 4.62437570e-01
-5.77720582e-01 -4.46352601e-01 -3.83474261e-01 -2.21168220e-01
-4.59292382e-02 5.19054942e-02 -1.76197931e-01 -9.35256660e-01
-1.59067288e-01 -4.49502409e-01 7.08553135e-01 -7.51747012e-01
-1.49942148e+00 7.01136947e-01 -4.92591932e-02 -1.35247672e+00
1.16721742e-01 -7.01838076e-01 -7.84250915e-01 1.04047740e+00
-1.24953139e+00 -1.13319361e+00 -4.14773226e-01 6.86099052e-01
4.82302159e-02 -5.10602355e-01 6.82566047e-01 2.68307120e-01
-6.63738430e-01 5.86666584e-01 3.58789176e-01 4.42701340e-01
6.19855046e-01 -1.29282331e+00 7.59390146e-02 9.14606392e-01
8.14322755e-02 6.75248384e-01 -1.09278429e-02 -1.05075395e+00
-8.55762422e-01 -1.53142536e+00 5.82119823e-01 2.03683779e-01
4.36218649e-01 -2.18090370e-01 -1.07371581e+00 6.91554129e-01
-3.26912969e-01 5.30606210e-01 6.00871205e-01 -1.92886184e-03
-4.37653214e-01 -1.70440264e-02 -1.28590393e+00 5.11691153e-01
1.13335633e+00 -4.20323789e-01 -5.91698229e-01 4.47779417e-01
8.20057511e-01 -1.29639268e-01 -7.65145898e-01 5.27045190e-01
1.17872506e-01 -6.67241156e-01 4.68816221e-01 -4.90588367e-01
3.43177259e-01 -4.94140655e-01 1.88270107e-01 -1.39138639e+00
-2.33966067e-01 -2.82748014e-01 -1.88826680e-01 1.11412704e+00
4.78950232e-01 -7.10080206e-01 9.59671259e-01 4.22513604e-01
-1.98173746e-01 -8.66312265e-01 -6.92200124e-01 -7.61019409e-01
-8.82438943e-02 -1.70458645e-01 4.74061698e-01 1.11878836e+00
1.64203137e-01 3.25628191e-01 -1.55892521e-01 -1.23832570e-02
8.73867571e-01 1.07606903e-01 3.01397622e-01 -1.68442261e+00
8.48843977e-02 -1.58370942e-01 -4.02625173e-01 -9.17837501e-01
2.06191853e-01 -1.47431374e+00 -3.40814069e-02 -1.53578138e+00
1.84642479e-01 -8.33405316e-01 -5.88599265e-01 6.96519554e-01
-4.84896451e-01 1.02164239e-01 -2.89345443e-01 3.05420041e-01
-7.41032302e-01 5.84627986e-01 1.43110490e+00 -8.99473503e-02
-2.40333185e-01 1.14775281e-02 -5.52402318e-01 6.71303689e-01
7.55960643e-01 -4.67916787e-01 -8.41129005e-01 2.09868047e-02
-1.38839200e-01 -9.54534709e-02 5.46811856e-02 -9.03583229e-01
3.43926758e-01 -1.13587365e-01 3.88835907e-01 -6.53745472e-01
-3.25788707e-01 -8.63757551e-01 -2.46732756e-02 4.57535028e-01
-4.21041518e-01 -4.04103756e-01 -2.43715391e-01 8.62431109e-01
-1.14435561e-01 -4.94184524e-01 7.36258507e-01 -5.38695641e-02
-6.57427907e-01 7.79099047e-01 1.27572119e-01 3.54613401e-02
1.13218772e+00 -1.69736803e-01 -2.92582691e-01 -1.85721129e-01
-8.96597326e-01 7.36881673e-01 2.57612795e-01 3.61632824e-01
6.75207376e-01 -1.53040743e+00 -3.85222107e-01 3.46347541e-01
3.31371784e-01 5.83717346e-01 6.38851821e-02 5.68195403e-01
-5.39495170e-01 1.33508176e-01 -3.89762484e-02 -5.21593809e-01
-6.85602427e-01 9.77684736e-01 3.11300546e-01 -3.76977503e-01
-6.17645502e-01 7.49261439e-01 -4.95191738e-02 -8.08611572e-01
5.07273674e-01 -8.43774825e-02 -4.73826528e-01 -2.49540910e-01
1.27108470e-01 2.99410254e-01 -9.40996781e-02 -2.08672062e-01
-1.71989664e-01 3.00115049e-01 -1.89776689e-01 8.11435729e-02
1.21661687e+00 3.82713117e-02 -3.43701899e-01 3.81269366e-01
1.32141435e+00 -3.41084570e-01 -1.16911411e+00 -5.89502931e-01
2.41026521e-01 -1.82776824e-01 1.52537942e-01 -7.67969310e-01
-1.25917029e+00 7.71475673e-01 4.97632623e-01 -2.50860043e-02
9.46373880e-01 -1.39998734e-01 2.84777164e-01 4.78898972e-01
4.01854932e-01 -1.02863908e+00 6.29557744e-02 3.25720727e-01
3.98898065e-01 -1.38528955e+00 6.22315854e-02 -5.53830802e-01
-4.80909497e-01 1.24027526e+00 1.08463955e+00 -1.55058041e-01
1.08468306e+00 4.74033505e-02 8.30787718e-02 -4.02614087e-01
-2.75598228e-01 -9.44291279e-02 2.79552728e-01 6.41748488e-01
1.41781956e-01 -2.90290773e-04 -4.49831039e-01 6.64938092e-01
1.06064871e-01 2.16352269e-01 3.05748492e-01 6.84488773e-01
-3.72846544e-01 -9.87702072e-01 -6.51452318e-02 9.56622243e-01
3.74828279e-02 -1.14707559e-01 -3.11515659e-01 5.84776103e-01
1.22441776e-01 8.39035571e-01 -2.06438527e-01 -6.85345650e-01
-1.11646429e-01 1.50095925e-01 1.34383300e-02 -8.64131093e-01
-4.17172104e-01 -1.03434632e-02 7.63942972e-02 -1.93133369e-01
-4.46323991e-01 -1.41299814e-01 -1.56216621e+00 -1.38073340e-01
-7.99808979e-01 7.42257357e-01 2.84055352e-01 1.01381588e+00
4.64774668e-02 4.17450339e-01 8.33319962e-01 -3.89353216e-01
-7.83644378e-01 -1.12213266e+00 -9.10735190e-01 4.19046283e-01
-1.22377388e-01 -8.39632392e-01 -5.10388613e-01 -4.55195010e-01] | [7.363381862640381, 6.1165595054626465] |
c8a99a03-1497-4cbd-b4df-aa495a29c254 | tempadacos-learning-temporally-structured | 2305.10816 | null | https://arxiv.org/abs/2305.10816v1 | https://arxiv.org/pdf/2305.10816v1.pdf | TempAdaCos: Learning Temporally Structured Embeddings for Few-Shot Keyword Spotting with Dynamic Time Warping | Few-shot keyword spotting (KWS) systems often utilize a sliding window of fixed size. Because of the varying lengths of different keywords or their spoken instances, choosing the right window size is a problem: A window should be long enough to contain all necessary information needed to recognize a keyword but a longer window may contain irrelevant information such as multiple words or noise and thus makes it difficult to reliably detect on- and offsets of keywords. In this work, TempAdaCos, an angular margin loss for obtaining embeddings with temporal structure, that can be used to detect keywords with dynamic time warping is proposed. In experiments conducted on KWS-DailyTalk, a few-shot keyword spotting (KWS) dataset presented in this work, it is shown that using these embeddings outperforms using other representations or a sliding window. Furthermore, it is shown that using time-reversed segments of the keywords while training the system improves the performance. | ['Alessia Cornaggia-Urrigshardt', 'Kevin Wilkinghoff'] | 2023-05-18 | null | null | null | null | ['keyword-spotting', 'dynamic-time-warping'] | ['speech', 'time-series'] | [ 2.71042772e-02 -4.44340169e-01 -4.85459179e-01 -3.17774802e-01
-9.76153731e-01 -6.00118160e-01 6.77668035e-01 4.36264455e-01
-6.20206416e-01 3.26550066e-01 1.58117190e-01 -3.09128165e-01
-1.32441103e-01 -1.65841475e-01 -3.02410871e-01 -6.99967325e-01
-4.11701292e-01 5.04879132e-02 5.81834853e-01 -2.47169703e-01
3.98237824e-01 4.60386813e-01 -1.51943398e+00 1.10362455e-01
3.26240301e-01 9.34891403e-01 4.91876453e-01 8.37989867e-01
-3.52056503e-01 1.12525269e-01 -9.76049066e-01 3.78574044e-01
3.61731887e-01 -4.57706243e-01 -4.07353342e-01 -1.77504867e-01
2.91609108e-01 -2.53881872e-01 -4.74592745e-01 6.44777417e-01
5.01437008e-01 5.87422132e-01 3.96956146e-01 -1.16224480e+00
-2.85189331e-01 6.19623542e-01 -3.43007028e-01 6.78360939e-01
5.53120255e-01 -2.85560042e-01 1.17973197e+00 -1.18628740e+00
6.08329415e-01 9.25875962e-01 7.26963282e-01 3.18448931e-01
-1.22220290e+00 -6.91428185e-01 3.57734114e-01 4.59843993e-01
-1.71540785e+00 -5.17974794e-01 6.52420104e-01 -1.02388263e-01
1.16568995e+00 5.09653389e-01 7.07089126e-01 1.09116375e+00
3.18746567e-01 6.17530644e-01 5.46915114e-01 -8.07663798e-01
2.33416378e-01 1.91790253e-01 4.20305699e-01 1.88135192e-01
-1.37039348e-01 -3.06777246e-02 -8.86449397e-01 -3.89446825e-01
-4.67271246e-02 2.50503514e-02 -4.84186292e-01 -1.43166274e-01
-1.14749408e+00 8.92923236e-01 -1.58768538e-02 8.81601512e-01
-2.78527230e-01 -8.51483643e-02 5.14941692e-01 5.77105165e-01
6.01278126e-01 5.75128555e-01 -3.84618878e-01 -6.71439528e-01
-1.47967291e+00 2.89045632e-01 8.31441998e-01 6.96885109e-01
5.06536424e-01 9.12739784e-02 -7.29641914e-02 9.69749510e-01
-2.21823454e-01 2.07276300e-01 1.05251122e+00 -2.56235600e-01
3.70005995e-01 2.07076088e-01 3.16556722e-01 -9.00357664e-01
-4.50874805e-01 1.29832149e-01 -1.28003865e-01 -1.84764206e-01
4.10464555e-01 -3.03780884e-02 -1.12841165e+00 1.54455888e+00
2.86157161e-01 3.34716916e-01 -4.80138101e-02 8.20978284e-01
4.77080733e-01 1.13761580e+00 -3.23643535e-01 -4.69058782e-01
1.34020209e+00 -7.61808217e-01 -1.12533689e+00 -3.34761441e-01
5.96382558e-01 -1.10381734e+00 1.09316289e+00 2.00718865e-01
-7.07423449e-01 -3.60663146e-01 -1.26274776e+00 3.02014444e-02
-9.13875163e-01 -1.07942805e-01 1.74876437e-01 4.18436795e-01
-8.21930349e-01 6.01614177e-01 -7.82072842e-01 -5.88234425e-01
-4.63528425e-01 1.57717735e-01 -2.61604339e-01 1.90489650e-01
-1.69525337e+00 9.78457212e-01 3.92540455e-01 -8.42383355e-02
-5.23844600e-01 -8.29516232e-01 -9.61495042e-01 1.08356513e-01
4.35459435e-01 3.87421697e-01 1.40061271e+00 -6.36416316e-01
-1.26384771e+00 4.39932227e-01 -4.85048264e-01 -7.02480078e-01
2.59111881e-01 -2.89707512e-01 -8.24468315e-01 4.48512137e-01
2.60627014e-03 3.11882555e-01 1.21667302e+00 -5.46171963e-01
-4.91284192e-01 -2.33951524e-01 -3.14590782e-01 1.31086871e-01
-5.78339696e-01 2.75342345e-01 -5.29453099e-01 -9.92716432e-01
-3.73467989e-03 -8.77044380e-01 8.90915021e-02 -2.67132103e-01
-2.72139370e-01 -4.41443175e-01 1.30714071e+00 -7.82797217e-01
1.80048513e+00 -2.42040014e+00 -1.70365065e-01 1.92301795e-01
-2.83825785e-01 4.47615921e-01 -1.08155176e-01 9.67341483e-01
-1.48940831e-01 1.18824290e-02 6.79632425e-02 -2.80864358e-01
-2.16868013e-01 3.02479088e-01 -6.60514176e-01 6.39063239e-01
-1.27708063e-01 5.16281128e-01 -7.18703568e-01 -4.04642463e-01
2.13847309e-01 3.69752020e-01 1.23943977e-01 7.85691515e-02
-5.12363166e-02 -2.69762516e-01 -1.43697098e-01 3.29947948e-01
3.10800552e-01 1.06671214e-01 -4.43113111e-02 7.78137147e-02
-4.88135159e-01 4.82815772e-01 -1.29133677e+00 1.49969935e+00
-5.98074973e-01 1.04905534e+00 -1.54375270e-01 -1.00978076e+00
9.28347230e-01 7.39270866e-01 5.47542095e-01 -7.57483125e-01
-1.03207871e-01 2.45521545e-01 -1.25943497e-01 -4.34151739e-01
9.30643618e-01 2.06163973e-01 -1.68115184e-01 6.89105570e-01
2.24152431e-02 -1.53354272e-01 3.44929755e-01 1.40339732e-01
1.07933056e+00 -4.94003296e-01 2.94350475e-01 -4.59332354e-02
1.74759433e-01 -3.60532627e-02 3.70212078e-01 8.13750148e-01
-9.73607153e-02 7.80988157e-01 2.40492970e-01 -3.37475359e-01
-1.19219697e+00 -7.21491039e-01 -2.78997719e-01 1.22568119e+00
1.22028887e-01 -8.64586711e-01 -3.42848599e-01 -3.59357595e-01
1.29191324e-01 8.38613808e-01 -6.49383307e-01 -4.21892703e-01
-5.43802977e-01 -3.70617241e-01 6.64466679e-01 3.57574940e-01
-2.16997162e-01 -6.42612219e-01 -9.11360085e-01 5.22345185e-01
-1.89376846e-01 -1.11999655e+00 -1.01581860e+00 5.48991501e-01
-5.94403088e-01 -8.55590522e-01 -9.82468128e-01 -6.73868537e-01
3.48452777e-01 6.61407292e-01 5.66945791e-01 -3.53362381e-01
-7.00581968e-01 4.87340152e-01 -7.79024005e-01 -3.35360497e-01
-2.34557644e-01 7.97423944e-02 1.82007119e-01 -6.68954253e-02
4.90859509e-01 -2.56820381e-01 -4.81260598e-01 2.86860675e-01
-9.04292166e-01 -4.45959449e-01 1.48873031e-01 1.03958178e+00
2.34389886e-01 -4.63803969e-02 4.54845309e-01 -2.69992501e-01
1.05453908e+00 -3.60491455e-01 -5.09858906e-01 3.77305210e-01
-8.48503590e-01 2.53814533e-02 6.05328977e-01 -1.07771027e+00
-4.18689817e-01 -3.56398642e-01 1.96777120e-01 -6.11670315e-01
2.06696585e-01 4.12765265e-01 3.80549937e-01 6.06813096e-03
5.07584453e-01 5.19439101e-01 9.37781855e-02 -5.91842413e-01
2.89394081e-01 9.86686707e-01 1.74560055e-01 6.25782907e-02
5.91307163e-01 2.85245121e-01 -6.67771935e-01 -1.34783518e+00
-4.71310765e-01 -1.10193360e+00 -3.61906826e-01 -1.04904883e-01
4.43818480e-01 -6.11719310e-01 -1.41823366e-01 1.05043776e-01
-1.10061324e+00 -4.67980164e-04 -2.73676723e-01 6.87295496e-01
-2.37867355e-01 3.44066590e-01 -2.52803922e-01 -1.04272020e+00
-3.15096647e-01 -7.89465368e-01 1.05377316e+00 3.36513162e-01
-8.16056788e-01 -8.29161882e-01 2.06089288e-01 -3.23863626e-01
5.28719783e-01 -2.52237409e-01 6.20303392e-01 -1.30778897e+00
1.10569172e-01 -6.57037020e-01 2.88504541e-01 2.82906681e-01
2.85486877e-01 2.38907740e-01 -7.95892298e-01 -3.25058103e-01
-5.14769517e-02 -9.45439637e-02 8.42047691e-01 1.58523574e-01
9.05052423e-01 -2.55305141e-01 -4.40763175e-01 2.03939646e-01
9.74907219e-01 7.46878624e-01 3.73920768e-01 3.00220281e-01
7.91451707e-02 2.62050152e-01 9.47423756e-01 6.95928693e-01
-7.77264163e-02 9.15120721e-01 -2.36252233e-01 2.94661373e-01
2.24333584e-01 -1.77161649e-01 4.25853401e-01 8.53333116e-01
4.31208760e-01 -3.14056247e-01 -8.73219132e-01 9.21016216e-01
-1.58650982e+00 -9.52415049e-01 4.83249813e-01 2.40508938e+00
9.25833106e-01 2.09512562e-01 2.87862029e-02 6.56495333e-01
8.11606050e-01 6.80202067e-01 -4.80359167e-01 -8.16018462e-01
4.47492786e-02 3.08764040e-01 4.91086096e-01 6.20181084e-01
-9.45378065e-01 8.95081878e-01 6.47400665e+00 9.17373955e-01
-1.44310570e+00 5.27594462e-02 5.92837185e-02 -2.69496322e-01
-6.66605532e-02 -3.94863226e-02 -9.49943960e-01 6.88753903e-01
1.22660971e+00 -6.34331584e-01 3.23843241e-01 6.66076064e-01
5.93103290e-01 -5.28475344e-01 -1.00221050e+00 1.04609466e+00
3.41591299e-01 -9.74257350e-01 -3.83460253e-01 -2.11615503e-01
2.21284404e-01 -4.79503609e-02 -7.13016540e-02 1.94495469e-01
-3.14826876e-01 -7.28788435e-01 6.65730834e-01 2.31379330e-01
5.81370175e-01 -7.96274066e-01 4.82327849e-01 3.16560417e-01
-1.30471575e+00 4.84366715e-02 -3.06037247e-01 1.14910178e-01
1.31836087e-01 6.02771163e-01 -1.27422833e+00 2.20066309e-01
6.20773554e-01 4.70275640e-01 -2.70107299e-01 1.15781665e+00
1.01930395e-01 7.82073796e-01 -6.52998924e-01 -5.04436314e-01
4.76802468e-01 4.01200987e-02 7.27323592e-01 1.37077832e+00
4.88629371e-01 -1.13859825e-01 1.88694850e-01 2.53409833e-01
2.29242384e-01 2.50638932e-01 -7.93772459e-01 -5.61826289e-01
8.46989274e-01 1.05239987e+00 -7.65591264e-01 -3.42111349e-01
-3.76283973e-01 1.13367510e+00 -1.54337153e-01 4.91739482e-01
-7.93517232e-01 -1.03347623e+00 8.55079830e-01 2.93636937e-02
7.91101873e-01 -5.61724961e-01 2.43862584e-01 -8.05126488e-01
1.75791353e-01 -5.10407090e-01 3.50115478e-01 -5.53155303e-01
-7.69405127e-01 4.95243222e-01 1.71021804e-01 -1.37843633e+00
-5.09016991e-01 -2.41721824e-01 -6.34264052e-01 8.11836541e-01
-1.13185263e+00 -5.22734404e-01 1.36695892e-01 3.66730064e-01
8.90455186e-01 8.19348916e-02 9.34988320e-01 2.21028328e-01
-2.46874630e-01 6.41025245e-01 4.30799037e-01 -3.88021916e-02
9.36043620e-01 -9.75687563e-01 4.57651913e-01 7.48010516e-01
6.77601576e-01 6.84581161e-01 1.07750821e+00 -4.77288038e-01
-1.42741525e+00 -5.35459697e-01 1.31118286e+00 2.05910236e-01
9.48982894e-01 -6.24635398e-01 -1.28151286e+00 3.25748861e-01
1.43537581e-01 1.08604722e-01 6.06509328e-01 7.65580079e-03
-4.91360456e-01 -3.22701663e-01 -7.19511807e-01 4.79830563e-01
5.30922890e-01 -8.22352886e-01 -9.34887826e-01 3.40779275e-01
8.71521056e-01 -3.16080719e-01 -5.65577567e-01 9.81038809e-02
8.09142590e-01 -6.16318941e-01 7.94807851e-01 -4.55828279e-01
-3.63936722e-01 -1.64370481e-02 -2.98313368e-02 -1.56486809e+00
2.41006643e-01 -8.58702004e-01 -1.70751974e-01 1.05264676e+00
4.63014632e-01 -5.74469805e-01 4.12822753e-01 4.30762440e-01
1.59007519e-01 -7.42548525e-01 -1.37674844e+00 -1.21635294e+00
-3.97753686e-01 -6.72745943e-01 3.04113239e-01 7.99735844e-01
3.59554589e-01 1.42524987e-01 -4.25492078e-01 1.07780136e-02
2.28864327e-03 1.73022121e-01 4.12122607e-01 -9.53858018e-01
1.79472700e-01 -2.52534330e-01 -5.25733411e-01 -1.01822937e+00
8.02304689e-03 -3.67572099e-01 3.81803811e-01 -1.15950072e+00
-4.40547705e-01 -2.82889873e-01 -4.30843771e-01 4.52909172e-01
-4.92724963e-02 -8.63896236e-02 1.41252339e-01 6.11660667e-02
-3.19739848e-01 4.86192256e-01 5.38507283e-01 -1.67523772e-01
-3.64260495e-01 -2.20983080e-03 3.94615121e-02 4.09427404e-01
7.17449486e-01 -6.10618174e-01 -3.98222268e-01 1.65326491e-01
1.16777290e-02 1.50229167e-02 3.07022613e-02 -7.41942942e-01
4.36033100e-01 -4.52651232e-02 5.98498397e-02 -9.05958474e-01
7.72448540e-01 -7.32386351e-01 -2.14918748e-01 3.14750552e-01
-3.53675812e-01 2.90715754e-01 3.48644793e-01 6.63709998e-01
-4.02638614e-01 -5.26532829e-01 5.51844358e-01 4.19628434e-02
-9.27452087e-01 -8.86920989e-02 -7.24871993e-01 1.21903807e-01
1.19200194e+00 -4.19929087e-01 3.77646834e-01 -5.65084338e-01
-6.41712189e-01 2.45553404e-01 1.33273914e-01 9.03215826e-01
8.29507530e-01 -1.17867780e+00 -1.72176316e-01 1.78953737e-01
3.75256062e-01 -5.77980995e-01 7.68114477e-02 8.13418925e-01
-1.36452451e-01 5.30734956e-01 3.96891177e-01 -4.76797342e-01
-1.74189591e+00 3.94417375e-01 -1.51047995e-02 1.19605817e-01
-9.39314961e-01 9.24289465e-01 -5.45020163e-01 8.44565183e-02
7.08579123e-01 -6.40458643e-01 -1.22933939e-01 7.24771857e-01
9.66709197e-01 1.50356039e-01 3.47501785e-01 -4.95289832e-01
-4.84418750e-01 5.76639354e-01 -3.44754308e-01 -4.59429443e-01
1.17383301e+00 -2.61911720e-01 2.98547179e-01 1.30840063e+00
1.35565352e+00 1.79180279e-02 -8.80396664e-01 -3.90891939e-01
5.28187871e-01 -6.44367397e-01 1.35673627e-01 -2.74917811e-01
-4.87263143e-01 6.43127561e-01 5.93226194e-01 6.56939149e-01
9.04227793e-01 2.18746718e-02 1.02742565e+00 4.40463781e-01
2.00369611e-01 -1.54753733e+00 -4.48518917e-02 5.02469778e-01
8.10055375e-01 -1.03831637e+00 -1.26718596e-01 2.41741195e-01
-5.49304128e-01 1.45687640e+00 6.10634908e-02 5.88772446e-02
7.78835237e-01 4.65175658e-01 1.34504557e-01 4.67564501e-02
-9.00783360e-01 -1.41287729e-01 2.92015374e-01 1.79912239e-01
9.10478756e-02 -1.17369927e-01 -5.34160197e-01 2.67516613e-01
-2.38511354e-01 -2.83193797e-01 4.99663085e-01 1.26212049e+00
-6.35780513e-01 -1.09098625e+00 -5.39489150e-01 3.86958212e-01
-4.88516062e-01 -1.03144303e-01 -4.63293403e-01 9.10912514e-01
-9.93518233e-02 1.00163078e+00 3.30321014e-01 -3.82900536e-01
3.26768935e-01 5.30393243e-01 1.49832994e-01 -6.18864536e-01
-4.02015895e-01 1.19409420e-01 1.09514549e-01 -4.29730117e-01
-1.40456185e-01 -5.45667589e-01 -1.24165916e+00 5.28287003e-03
-7.42660820e-01 6.16384625e-01 9.58209157e-01 9.27849114e-01
2.64291734e-01 2.47177839e-01 8.17878127e-01 -6.19308233e-01
-6.43027842e-01 -1.22083855e+00 -7.86738276e-01 9.55332518e-02
6.31192207e-01 -7.46937692e-01 -7.42443800e-01 -1.29474744e-01] | [14.254700660705566, 6.370201110839844] |
64e51f74-3a90-4d37-980c-6539205d727d | plan-execution-for-multi-agent-path-finding | 2207.01752 | null | https://arxiv.org/abs/2207.01752v2 | https://arxiv.org/pdf/2207.01752v2.pdf | Plan Execution for Multi-Agent Path Finding with Indoor Quadcopters | We study the planning and acting phase for the problem of multi-agent path finding (MAPF) in this paper. MAPF is a problem of navigating agents from their start positions to specified individual goal positions so that agents do not collide with each other. Specifically we focus on executing MAPF plans with a group of Crazyflies, small indoor quadcopters . We show how to modify the existing continuous time conflict-based search algorithm (CCBS) to produce plans that are suitable for execution with the quadcopters. The acting phase uses the the Loco positioning system to check if the plan is executed correctly. Our finding is that the CCBS algorithm allows for extensions that can produce safe plans for quadcopters, namely cylindrical protection zone around each quadcopter can be introduced at the planning level. | ['Pavel Surynek', 'Matouš Kulhan'] | 2022-07-05 | null | null | null | null | ['multi-agent-path-finding'] | ['playing-games'] | [ 1.57230273e-01 4.29966420e-01 6.76149577e-02 1.16504580e-02
-3.45222473e-01 -8.72666299e-01 6.43765450e-01 4.36617374e-01
-4.42010760e-01 1.14282477e+00 -2.09167525e-01 -4.09204125e-01
-8.98254693e-01 -1.26598358e+00 -4.44910794e-01 -6.59061372e-01
-7.81749249e-01 1.21522915e+00 8.54631245e-01 -6.95132494e-01
5.54857969e-01 6.21163845e-01 -1.17005193e+00 -1.82535976e-01
9.21071053e-01 1.97098330e-01 5.61050594e-01 1.06397295e+00
3.68529141e-01 3.35096747e-01 -7.00892806e-01 3.52596104e-01
5.61762393e-01 -2.15624183e-01 -1.03287804e+00 -1.22718401e-01
-2.29127765e-01 -2.31798008e-01 2.11411282e-01 7.90555835e-01
3.30896564e-02 4.15891409e-01 5.05552173e-01 -1.98236012e+00
4.49787527e-01 5.00469267e-01 -3.31329674e-01 -1.06026590e-01
8.16741109e-01 5.72197698e-02 4.59502131e-01 2.44197957e-02
8.26928973e-01 1.06263924e+00 4.56507832e-01 1.85775220e-01
-8.15430224e-01 -1.85754016e-01 2.65798718e-01 7.74767846e-02
-1.47314084e+00 -9.64829791e-03 1.23191558e-01 -1.21177353e-01
1.44299531e+00 6.01500869e-01 5.95600307e-01 1.52670652e-01
9.67905819e-01 -2.11806670e-02 8.26345921e-01 -4.67720181e-01
6.18590832e-01 -2.68855661e-01 4.63920906e-02 8.08633685e-01
4.79355961e-01 4.89391029e-01 -6.61359206e-02 -4.70625639e-01
5.94832301e-01 -3.69774133e-01 -2.49544024e-01 -5.67991376e-01
-1.22467411e+00 8.14276218e-01 3.74676675e-01 3.72092009e-01
-5.85845053e-01 2.13165894e-01 -7.13496059e-02 2.85451114e-01
-3.23941410e-01 8.31956863e-01 -3.28432739e-01 -9.95178446e-02
-4.32453960e-01 8.69384944e-01 1.01082158e+00 1.09348321e+00
7.86541879e-01 -6.10790431e-01 3.56607199e-01 -2.63341963e-02
1.88790306e-01 3.03602308e-01 -4.06439692e-01 -1.24531150e+00
7.88637698e-01 5.85972011e-01 7.20805347e-01 -1.10704494e+00
-9.81359601e-01 9.93991271e-04 -2.26738021e-01 1.04540777e+00
4.03828591e-01 -3.41334164e-01 -4.77474540e-01 1.53780723e+00
6.06603563e-01 -2.25540504e-01 4.66462195e-01 5.97374260e-01
1.05617993e-01 8.84506702e-01 -3.59977186e-01 -4.15746897e-01
9.26326334e-01 -1.25343025e+00 -4.38650668e-01 -2.71433294e-01
1.06263506e+00 -5.50653398e-01 3.64853829e-01 5.05917788e-01
-1.16837549e+00 -4.56920899e-02 -1.30291474e+00 4.12100554e-01
-5.54479122e-01 -5.49566388e-01 3.81560832e-01 6.72376454e-01
-1.44723964e+00 5.03337085e-01 -1.02760708e+00 -4.25174057e-01
-3.50638151e-01 5.39401293e-01 -4.74711984e-01 -6.46065548e-02
-8.52291822e-01 1.20554864e+00 5.54254651e-01 -1.00443766e-01
-9.63504076e-01 6.91173449e-02 -8.65621626e-01 1.60082374e-02
9.12320256e-01 -6.63741231e-01 1.08171928e+00 -2.71540910e-01
-1.27992618e+00 7.28274211e-02 -8.72602537e-02 -4.64368761e-01
3.35955203e-01 3.81590694e-01 -5.18546179e-02 6.66965917e-02
6.77940249e-01 4.01958257e-01 5.02403527e-02 -1.45018816e+00
-1.17871642e+00 -3.34781677e-01 4.73546475e-01 6.04655921e-01
6.75434411e-01 -1.67285532e-01 -8.24925676e-02 2.29869470e-01
4.01617080e-01 -1.37708163e+00 -9.68018055e-01 -5.30801654e-01
-7.31024444e-01 -2.65801340e-01 6.99451864e-01 1.25868209e-02
1.07125711e+00 -1.61867499e+00 1.84213400e-01 7.57811666e-01
-2.06241012e-01 -3.21047843e-01 -1.08223349e-01 1.28904247e+00
2.78825045e-01 6.61929697e-02 -1.13346972e-01 -7.67025054e-02
6.32808059e-02 5.41757405e-01 1.25635251e-01 4.81070787e-01
-5.86757243e-01 4.31147784e-01 -1.02119017e+00 -2.94490635e-01
1.15727566e-01 -2.91288137e-01 -5.77530324e-01 -1.59899324e-01
-4.03831720e-01 1.72604263e-01 -7.05012143e-01 4.80116427e-01
8.17771316e-01 2.07375526e-01 2.81548440e-01 7.40010858e-01
-1.09956264e+00 4.67075020e-01 -1.41486597e+00 1.56921494e+00
-1.84299931e-01 1.72809437e-01 4.35316086e-01 -4.00716275e-01
6.73371613e-01 1.15307257e-01 5.37747264e-01 -3.84666592e-01
-1.57078847e-01 3.68167549e-01 -1.06234863e-01 -1.97135761e-01
7.87553251e-01 8.01295862e-02 -5.33285022e-01 7.10070670e-01
-7.95753896e-01 -3.59236062e-01 5.10104537e-01 2.45931119e-01
1.45701134e+00 1.19988518e-02 5.30052960e-01 -5.20573318e-01
5.21291018e-01 9.07288671e-01 5.17503083e-01 9.99267042e-01
-1.94878086e-01 9.48414654e-02 2.68772066e-01 -5.31770945e-01
-6.13648474e-01 -9.58123863e-01 3.18289101e-01 7.40652323e-01
9.99313533e-01 -7.05139875e-01 -7.98255265e-01 -6.06037378e-01
-2.99553812e-01 7.87040830e-01 -4.79124784e-01 4.56946909e-01
-8.90920401e-01 -2.94451952e-01 7.40189254e-02 -6.91795796e-02
6.23162806e-01 -8.02257121e-01 -1.57730901e+00 6.03446901e-01
-2.02732846e-01 -7.53941059e-01 -3.99997860e-01 3.56279463e-01
-2.97744572e-01 -1.48679352e+00 -6.79586753e-02 -7.73029447e-01
9.34074819e-01 7.01445162e-01 5.56130886e-01 5.14662147e-01
3.14324200e-01 3.67707461e-01 -5.67705154e-01 -3.62225056e-01
-3.28445077e-01 1.35991141e-01 -2.33329590e-02 -7.97484994e-01
-2.66109020e-01 -4.95948702e-01 -3.64666969e-01 8.05549324e-01
-3.41052592e-01 3.77934158e-01 1.67333335e-01 3.46169978e-01
4.82273042e-01 1.07713747e+00 2.60193020e-01 -3.56597155e-01
6.64997041e-01 -4.02937651e-01 -9.88456070e-01 3.05909842e-01
-5.78398705e-01 -2.06286818e-01 5.82218587e-01 2.94545352e-01
-6.70373082e-01 1.59235150e-01 -5.50689176e-03 6.59833431e-01
-2.45893836e-01 4.19748813e-01 -2.04603598e-01 -4.89779413e-01
5.55258632e-01 -1.36896759e-01 -2.45332137e-01 5.29758679e-03
6.61029816e-02 5.75375073e-02 3.02468538e-01 -5.50924420e-01
7.95372427e-01 4.62209225e-01 5.44760406e-01 -3.53051156e-01
2.73159772e-01 -2.78327286e-01 -4.51101780e-01 -4.69060838e-01
7.75426090e-01 -2.00637251e-01 -1.04245758e+00 8.21647644e-02
-1.21733069e+00 -6.38998687e-01 1.49829134e-01 7.00009316e-02
-8.26968074e-01 1.57210931e-01 -8.41455832e-02 -1.15700567e+00
7.87067115e-02 -1.26995027e+00 9.16655660e-01 5.10724485e-01
-3.17722619e-01 -8.73775303e-01 6.45482302e-01 6.61486164e-02
2.12515160e-01 8.59799325e-01 6.70618296e-01 -4.08163607e-01
-9.89224195e-01 1.55579776e-01 3.54060322e-01 -1.05462861e+00
1.32208526e-01 -2.91367084e-01 -2.15931423e-02 -5.29318273e-01
-2.02478558e-01 3.07395816e-01 6.46399185e-02 4.72514212e-01
1.24454819e-01 -5.52228212e-01 -1.17031217e+00 1.07271731e-01
1.69674230e+00 1.18090177e+00 6.11842275e-01 1.14209116e+00
-8.48874971e-02 7.51729429e-01 1.18782067e+00 1.81928515e-01
7.59761810e-01 1.04084563e+00 9.37203586e-01 2.56353021e-01
4.81321245e-01 -1.17532350e-01 3.11919749e-01 -2.19877318e-01
-2.58976787e-01 -8.05786610e-01 -1.12310660e+00 6.71627045e-01
-2.23498034e+00 -8.62039924e-01 -4.81560081e-01 2.05340242e+00
1.16358042e-01 2.43889168e-01 2.28652999e-01 2.28071839e-01
6.01006627e-01 -5.72805814e-02 4.24505547e-02 -6.18783414e-01
2.74114609e-01 -2.22889945e-01 1.07307208e+00 1.36167276e+00
-8.81403089e-01 7.04352081e-01 6.23470831e+00 2.36025348e-01
-2.99313247e-01 1.65049657e-02 -3.24692130e-02 2.44424060e-01
-1.81362882e-01 5.19064426e-01 -7.45550036e-01 1.91646367e-01
7.08847940e-01 -7.67970085e-02 5.90376616e-01 6.14751756e-01
5.52139223e-01 -1.00742173e+00 -7.26216197e-01 5.31313475e-03
-3.45186710e-01 -1.32871401e+00 -5.82799375e-01 4.09550458e-01
8.58872831e-01 -3.06628764e-01 -4.94333059e-01 -3.99419852e-02
6.62441969e-01 -8.69848907e-01 7.13631332e-01 1.36961177e-01
-3.25081609e-02 -1.10441697e+00 5.36645293e-01 9.29838896e-01
-1.41326141e+00 -2.40953818e-01 -7.23627210e-03 -4.45383579e-01
9.98879135e-01 -1.03914224e-01 -1.33381128e+00 1.18956196e+00
4.36860144e-01 -2.50803649e-01 9.51978341e-02 1.39178848e+00
-1.93485558e-01 -5.04787743e-01 -6.72951996e-01 -3.05696219e-01
8.22045028e-01 -5.21194577e-01 9.98489261e-01 3.92267913e-01
6.81374967e-01 3.12742770e-01 6.20040059e-01 5.29195070e-01
1.26460290e+00 -2.29672089e-01 -7.85547018e-01 5.55797100e-01
5.57299495e-01 6.88603997e-01 -1.15225363e+00 -3.95267196e-02
1.89007387e-01 6.31016374e-01 -1.55304849e-01 3.24565440e-01
-7.90235519e-01 -6.09003842e-01 4.63419944e-01 2.47981131e-01
-1.11633286e-01 -6.27967954e-01 -2.69033521e-01 -2.41948068e-01
-2.38985509e-01 -5.17360687e-01 4.21833515e-01 -8.39848340e-01
-2.95318663e-01 7.77561665e-01 4.79689360e-01 -8.84209156e-01
-5.56282699e-01 -1.74212724e-01 -8.70042086e-01 6.21660590e-01
-1.27907908e+00 -1.03967249e+00 -3.48268896e-01 6.06052876e-01
5.28586686e-01 -1.05910245e-02 9.42933202e-01 -2.36916646e-01
-1.46551862e-01 -1.85676008e-01 -2.11863577e-01 -5.58399856e-01
-1.42700016e-01 -1.13469756e+00 3.54586035e-01 1.16293812e+00
-2.54543930e-01 5.69966316e-01 1.14599824e+00 -1.20444059e+00
-1.30459177e+00 -7.27400720e-01 1.04628980e+00 -2.07522735e-01
1.63402349e-01 6.35932162e-02 -3.29397589e-01 9.62484598e-01
3.23775917e-01 -8.54076564e-01 6.52181283e-02 -1.76336318e-01
5.17534554e-01 1.61922500e-01 -1.33678603e+00 8.40126932e-01
1.06842422e+00 4.13094014e-01 -2.95846075e-01 5.12911975e-01
4.09411460e-01 -7.74690449e-01 -2.54876465e-01 4.05968994e-01
1.99201018e-01 -1.10910714e+00 7.46664345e-01 -2.67694145e-02
-3.91204923e-01 -1.02793765e+00 7.45285600e-02 -1.63396025e+00
-2.11229041e-01 -1.17529452e+00 8.13025951e-01 5.80978513e-01
5.66487312e-01 -1.01518309e+00 8.83545637e-01 6.25125527e-01
-6.18154287e-01 -4.50406462e-01 -1.52144706e+00 -9.45304990e-01
-2.86119044e-01 -1.12282574e-01 7.58709133e-01 5.76184213e-01
6.47556126e-01 1.96971539e-02 -2.68930852e-01 8.84443700e-01
5.13468742e-01 1.60984844e-01 9.41445708e-01 -1.03913891e+00
-2.68509656e-01 -2.79465735e-01 -9.85603128e-03 -9.38408554e-01
-2.51192659e-01 -4.95286942e-01 4.64251816e-01 -2.16173053e+00
-4.55377311e-01 -1.10689819e+00 5.56009412e-01 6.73330367e-01
5.63639998e-01 -4.80582684e-01 3.81827712e-01 -4.51185070e-02
-5.65530300e-01 -1.15209922e-01 1.29636121e+00 1.50518894e-01
-6.46414816e-01 4.06797796e-01 -2.36786306e-01 3.39217633e-01
1.06639230e+00 -6.58786237e-01 -7.60072351e-01 -9.41364095e-02
5.32324791e-01 8.25961649e-01 -9.98817291e-03 -1.07366920e+00
6.97332501e-01 -8.47334087e-01 -6.20240390e-01 -8.21704328e-01
4.88379657e-01 -1.09716046e+00 9.53707933e-01 1.22886431e+00
-2.10719742e-02 7.11725533e-01 2.09183380e-01 4.67685372e-01
2.23894179e-01 -7.42944956e-01 3.04195940e-01 -2.67122418e-01
-6.29625440e-01 -4.24114615e-02 -9.70250309e-01 -6.51641369e-01
1.82776868e+00 -4.26047742e-01 -8.19307864e-01 -4.48988050e-01
-5.53787947e-01 8.68028820e-01 7.74554968e-01 -1.15672193e-01
5.84841967e-01 -1.02781415e+00 -3.91790688e-01 4.57530431e-02
-1.88134566e-01 2.13605136e-01 9.90438163e-02 8.54474664e-01
-1.05019188e+00 7.00249791e-01 -5.14165342e-01 -9.68951136e-02
-1.47108865e+00 5.84511220e-01 5.87692738e-01 -5.25271595e-01
-7.03442752e-01 6.16096675e-01 1.30766351e-02 -2.21517161e-01
8.16157311e-02 -4.00783092e-01 -3.20682198e-01 -4.32575345e-01
5.96762955e-01 5.76691329e-01 -2.17131019e-01 -4.86910641e-01
-8.15591574e-01 6.91179454e-01 2.79803008e-01 -6.79564714e-01
1.27405584e+00 -4.53192800e-01 -2.20541239e-01 -2.69155502e-01
2.96159059e-01 3.71047109e-01 -1.03503287e+00 6.45751953e-01
3.75929698e-02 -5.12049973e-01 -2.93740183e-01 -9.57279682e-01
-2.96126992e-01 -1.63483500e-01 8.11353251e-02 7.65130639e-01
9.52638268e-01 -1.45308733e-01 5.89789093e-01 3.70798916e-01
1.46819723e+00 -1.03935695e+00 -1.91722870e-01 7.47121990e-01
7.45222151e-01 -4.43811208e-01 1.45660952e-01 -8.49404514e-01
-4.63818580e-01 1.07726943e+00 6.66612864e-01 -2.57309496e-01
1.37256354e-01 4.73687440e-01 -2.16108173e-01 -4.49336529e-01
-6.65439367e-01 -1.93927422e-01 -5.82715034e-01 1.02980566e+00
-8.19298685e-01 1.68793082e-01 -6.75442457e-01 1.25380546e-01
-5.51005244e-01 -3.01120043e-01 1.12019718e+00 1.50844932e+00
-1.09493756e+00 -1.30670905e+00 -8.57692778e-01 -2.28885129e-01
3.31973314e-01 4.13625151e-01 -5.01017809e-01 1.22940445e+00
5.44413209e-01 1.36746216e+00 1.80240441e-02 -3.29850256e-01
3.70976120e-01 -4.45724308e-01 3.99016112e-01 -6.01773679e-01
-6.55786335e-01 -3.74800004e-02 8.42231870e-01 -7.29797065e-01
-3.83997470e-01 -7.27812290e-01 -1.81681705e+00 -3.27918679e-01
-1.77056402e-01 7.49704242e-01 5.53039551e-01 9.93264735e-01
5.72926477e-02 3.05991262e-01 6.53002501e-01 -8.36858153e-01
1.64870501e-01 -2.87883043e-01 -3.70936185e-01 -7.29662359e-01
3.60662699e-01 -8.40121329e-01 -2.09169105e-01 -6.50412798e-01] | [4.955561637878418, 1.67836594581604] |
70521a9e-1be5-44cb-88e9-753a293149b5 | a-quasi-bayesian-perspective-to-online | 1602.00522 | null | http://arxiv.org/abs/1602.00522v3 | http://arxiv.org/pdf/1602.00522v3.pdf | A Quasi-Bayesian Perspective to Online Clustering | When faced with high frequency streams of data, clustering raises theoretical
and algorithmic pitfalls. We introduce a new and adaptive online clustering
algorithm relying on a quasi-Bayesian approach, with a dynamic (i.e.,
time-dependent) estimation of the (unknown and changing) number of clusters. We
prove that our approach is supported by minimax regret bounds. We also provide
an RJMCMC-flavored implementation (called PACBO, see
https://cran.r-project.org/web/packages/PACBO/index.html) for which we give a
convergence guarantee. Finally, numerical experiments illustrate the potential
of our procedure. | ['Sébastien Loustau', 'Le Li', 'Benjamin Guedj'] | 2016-02-01 | null | null | null | null | ['online-clustering'] | ['computer-vision'] | [-3.36029142e-01 -9.11044255e-02 -2.21131906e-01 -4.38292116e-01
-1.05415523e+00 -7.90959060e-01 8.42539296e-02 1.13906145e-01
-5.20314574e-01 5.71391404e-01 6.95157424e-02 -2.83391267e-01
-3.69946837e-01 -5.66251040e-01 -9.44255412e-01 -9.08288062e-01
-8.21345508e-01 7.15427637e-01 2.23352730e-01 3.90307248e-01
-9.92425811e-03 6.49330989e-02 -1.13410270e+00 -7.41196275e-02
4.58223790e-01 9.87129569e-01 -2.45660916e-02 1.00932896e+00
3.12416255e-01 4.41229522e-01 -1.44611835e-01 -4.96899903e-01
5.92619777e-01 -4.91417319e-01 -6.64494753e-01 1.68424591e-01
-4.88241732e-01 -3.32599431e-01 -1.75759450e-01 9.75055516e-01
6.05571330e-01 1.83164120e-01 3.33281696e-01 -1.43256509e+00
-5.41461036e-02 1.05653000e+00 -8.10531676e-01 5.01110017e-01
2.59393215e-01 -2.73768138e-03 1.00991940e+00 -5.95596552e-01
3.96028370e-01 1.11394811e+00 6.77142739e-01 3.77786160e-01
-1.37673461e+00 -6.80210650e-01 3.96438658e-01 1.50691643e-01
-1.71991026e+00 -5.87876856e-01 4.93408263e-01 -2.13592947e-01
2.12643608e-01 4.47128236e-01 6.55612767e-01 9.48592126e-01
-2.41199374e-01 1.04171824e+00 9.08438265e-01 -2.83782095e-01
8.14077795e-01 -7.44076520e-02 4.76444960e-02 2.78793842e-01
4.26140904e-01 6.26586527e-02 -6.14911854e-01 -7.15987682e-01
4.05888051e-01 1.41742066e-01 -2.89887100e-01 -5.93403399e-01
-9.03849244e-01 9.01003122e-01 -2.61446158e-03 -1.03865311e-01
-4.42852318e-01 5.50350845e-01 3.52773726e-01 1.99662626e-01
5.67934692e-01 -2.80604780e-01 -5.63355446e-01 -6.23995066e-01
-7.40542173e-01 1.55487314e-01 1.22332215e+00 1.23117292e+00
3.60231996e-01 -3.07247132e-01 1.09424278e-01 6.47985220e-01
2.72665679e-01 5.88841438e-01 1.74419478e-01 -1.31886697e+00
2.53741264e-01 -3.54913056e-01 5.88460982e-01 -8.21010768e-01
-4.56550777e-01 -3.99951339e-01 -7.53796279e-01 -4.75794613e-01
4.70441550e-01 -5.92806339e-01 -1.84187099e-01 1.83645153e+00
6.91615224e-01 5.88289618e-01 -2.00004011e-01 8.88250351e-01
-2.03199103e-01 5.16129971e-01 -3.34105849e-01 -8.17642152e-01
8.48812521e-01 -7.21009135e-01 -6.90792620e-01 6.33829087e-02
1.25008672e-01 -4.21567202e-01 6.35106206e-01 6.60315812e-01
-1.31798375e+00 2.82652497e-01 -6.75678253e-01 6.06164277e-01
-4.07243632e-02 -6.84976578e-01 8.79076242e-01 9.43563521e-01
-1.06523824e+00 5.14283359e-01 -1.39461505e+00 -3.61260414e-01
5.64929545e-01 1.09258004e-01 3.87250841e-01 -3.46794203e-02
-7.50228703e-01 1.73188135e-01 3.99048299e-01 -1.17905565e-01
-1.04240942e+00 -5.15645027e-01 -3.26114953e-01 -3.74764353e-02
9.09691989e-01 -5.68284571e-01 1.56230247e+00 -9.63865817e-01
-1.53155792e+00 3.74143928e-01 -2.18784273e-01 -7.26119757e-01
9.24751818e-01 -2.35422984e-01 4.03507687e-02 4.28747982e-01
-2.74945945e-02 4.60432544e-02 6.92334890e-01 -1.15778041e+00
-8.73422146e-01 -2.08283812e-01 4.83226366e-02 2.57891506e-01
-3.70107114e-01 1.29083097e-01 -5.37280321e-01 -3.83095741e-01
-9.58008245e-02 -9.46541846e-01 -7.42830098e-01 3.17248702e-02
-2.90465087e-01 3.35479379e-02 1.60226673e-02 -4.13163573e-01
1.33012199e+00 -2.13830519e+00 -1.57203272e-01 6.10903740e-01
9.36576072e-03 -4.64865446e-01 2.29544207e-01 7.28635609e-01
3.19169581e-01 2.87712991e-01 -4.00975615e-01 -2.68458217e-01
2.35708922e-01 1.07324518e-01 -2.76117250e-02 9.83349681e-01
-5.18627822e-01 4.16154712e-01 -1.25462413e+00 -3.32215607e-01
2.52691824e-02 8.78773034e-02 -8.05557549e-01 1.37228891e-01
-2.01967567e-01 2.14520738e-01 -4.81337011e-01 6.35840356e-01
9.14157927e-01 -5.79977930e-01 7.72396505e-01 4.41393942e-01
3.68788801e-02 -1.72612909e-02 -1.68548310e+00 1.53817582e+00
-4.13214177e-01 2.66097844e-01 7.03124642e-01 -9.90155697e-01
1.14739574e-01 2.12295160e-01 7.54642665e-01 5.50302025e-03
2.19491333e-01 4.94109876e-02 -2.57062703e-01 -2.55283445e-01
2.14926481e-01 -1.81007713e-01 -1.73110649e-01 8.66061211e-01
-2.45818973e-01 3.39757711e-01 4.58001606e-02 4.80886489e-01
1.36406350e+00 -4.69770916e-02 3.07386696e-01 -3.71339381e-01
-6.32480159e-02 -1.33083075e-01 6.51524842e-01 1.31622314e+00
-5.49291253e-01 2.86813259e-01 4.78925556e-01 -1.76208645e-01
-1.05812109e+00 -1.05136144e+00 -3.46155912e-01 1.15696836e+00
1.28673404e-01 -4.00655478e-01 -7.83711672e-01 -1.87110752e-01
3.15478712e-01 5.21270514e-01 -5.66128731e-01 1.69285670e-01
-1.85709819e-01 -1.22054207e+00 2.80134976e-01 3.45020771e-01
-1.20966986e-01 -4.55938876e-01 -8.33239198e-01 5.32484949e-01
-2.11897835e-01 -1.02190530e+00 -5.50286412e-01 2.25123525e-01
-9.29132998e-01 -1.10978854e+00 -3.09071600e-01 -8.90431777e-02
4.46526319e-01 6.71591341e-01 9.81577516e-01 3.62400152e-02
-2.41232455e-01 1.12298250e+00 -5.97418129e-01 -3.53959888e-01
-1.65853314e-02 2.84292139e-02 1.23264641e-01 2.81883150e-01
1.55043036e-01 -7.55916536e-01 -9.04349804e-01 2.26229772e-01
-8.52315187e-01 -2.44679525e-01 1.10835001e-01 6.77795827e-01
6.78798139e-01 2.16566876e-01 6.93633139e-01 -9.67340589e-01
6.02666855e-01 -1.08323717e+00 -1.16830647e+00 1.33800730e-01
-5.86857140e-01 -1.63106516e-01 4.96615857e-01 -4.05288517e-01
-8.44846249e-01 2.84291387e-01 4.27285992e-02 -5.12538910e-01
4.35449556e-02 5.25377750e-01 5.44878319e-02 4.18358743e-01
3.77201319e-01 1.96146324e-01 4.96633947e-02 -3.39666694e-01
6.59368038e-01 7.68351197e-01 4.45920050e-01 -8.62422705e-01
8.25392127e-01 9.87196863e-01 -2.74483800e-01 -5.61648071e-01
-9.54154849e-01 -8.33263338e-01 -3.58251482e-01 -4.03934032e-01
2.60810524e-01 -1.12527418e+00 -1.25885916e+00 2.27220729e-01
-7.20444500e-01 -8.30080867e-01 -7.96907172e-02 4.34293449e-01
-8.95246923e-01 5.13687789e-01 -6.69981718e-01 -1.48388159e+00
-1.79875076e-01 -3.66376638e-01 7.07376361e-01 5.32609634e-02
2.03236133e-01 -9.34608877e-01 1.01995878e-01 1.85467452e-01
1.62435785e-01 3.27413529e-01 -7.51751959e-02 -7.54743338e-01
-7.02673852e-01 -1.71880811e-01 6.55920729e-02 -7.08825439e-02
-1.94122165e-01 2.52448529e-01 -8.06052089e-01 -7.02586174e-01
8.33224878e-02 -5.18720299e-02 4.72736686e-01 5.57187200e-01
1.50520277e+00 -8.01460147e-01 -4.43819165e-01 5.55743396e-01
1.64331579e+00 8.37479979e-02 6.49992004e-02 2.34277159e-01
1.30105838e-01 1.86873078e-01 7.96483338e-01 1.65775633e+00
6.08579695e-01 5.13613403e-01 4.25361961e-01 4.70636368e-01
7.91318178e-01 -1.66054383e-01 3.83649468e-01 7.69918621e-01
2.66988993e-01 -2.70769209e-01 -8.47123444e-01 7.92408824e-01
-2.36861086e+00 -1.08617568e+00 -1.66656688e-01 2.64846396e+00
1.13144815e+00 1.04017004e-01 6.07860744e-01 2.57081669e-02
9.56475556e-01 -2.44600669e-01 -7.48022377e-01 -2.41252586e-01
1.25174493e-01 1.27664104e-01 1.08619559e+00 6.50740623e-01
-1.09917009e+00 6.60085976e-01 6.30094671e+00 1.04321611e+00
-4.30017829e-01 4.96986955e-01 6.19473159e-01 -6.01526380e-01
-1.19275235e-01 7.20062107e-02 -3.60513806e-01 8.11529040e-01
1.65257645e+00 -7.94393659e-01 1.01569438e+00 8.81823778e-01
5.40180922e-01 -3.29356730e-01 -1.09018624e+00 9.17462170e-01
-3.22372615e-01 -1.28159046e+00 -8.62628043e-01 2.26216480e-01
5.90973616e-01 2.47299284e-01 -3.51212978e-01 -2.66839005e-03
1.05639577e+00 -3.65886569e-01 1.02620065e+00 3.60271752e-01
4.36081886e-01 -1.28132963e+00 5.19475937e-01 4.31068182e-01
-1.15547049e+00 -4.37170744e-01 -5.59498906e-01 -1.46136358e-01
3.30779135e-01 1.09019351e+00 -5.67013919e-01 6.59503162e-01
1.20288658e+00 6.22432172e-01 -8.48696306e-02 1.34569931e+00
-1.16789892e-01 1.03692222e+00 -9.27871108e-01 -8.58110711e-02
-6.66866824e-02 -3.86930943e-01 5.84420562e-01 1.31222653e+00
2.26270601e-01 4.12741959e-01 3.90609294e-01 5.65499604e-01
-1.89849749e-01 3.28436270e-02 -4.06765312e-01 2.83963174e-01
1.17709780e+00 1.33545291e+00 -1.04911423e+00 -4.13542151e-01
-2.03835204e-01 6.80768967e-01 2.94156224e-01 3.18037838e-01
-9.83716369e-01 -1.36534721e-01 5.17266631e-01 -2.11152092e-01
6.90643072e-01 -2.07185403e-01 -1.20381340e-01 -1.07363963e+00
-3.42509001e-02 -6.02641225e-01 1.07305181e+00 -4.07707900e-01
-1.56067014e+00 -2.08451658e-01 3.52274850e-02 -1.00494611e+00
-7.75551125e-02 -1.57837689e-01 -5.99635422e-01 1.57961056e-01
-1.11771989e+00 -2.23740131e-01 2.76035722e-02 6.29495382e-01
2.87081301e-01 5.82284808e-01 2.49704391e-01 2.49574572e-01
-6.76597357e-01 5.92478991e-01 9.09260035e-01 -2.55134907e-02
4.28424656e-01 -1.40134156e+00 -9.63786989e-02 8.54839742e-01
-1.50890350e-01 2.35824078e-01 9.16637242e-01 -2.95160383e-01
-1.81635141e+00 -1.00495410e+00 3.44638199e-01 -3.00999075e-01
1.20205915e+00 -6.72513962e-01 -4.14995462e-01 9.35180664e-01
1.02842823e-01 -1.18061453e-01 1.16289818e+00 3.06333065e-01
-1.02548912e-01 -3.58634949e-01 -1.19819236e+00 6.49877310e-01
1.15009427e+00 -5.04753552e-02 -7.16193616e-02 7.19666541e-01
7.29109824e-01 -1.43686146e-01 -8.63939285e-01 1.82767391e-01
5.78810155e-01 -1.04633379e+00 6.47301495e-01 -2.81442672e-01
-2.70377606e-01 -1.34736761e-01 -3.06614876e-01 -9.72683132e-01
-1.95379093e-01 -1.17368948e+00 -4.07881826e-01 1.07894576e+00
2.09979117e-01 -8.33964705e-01 5.36272526e-01 5.72470486e-01
2.29935318e-01 -4.75900978e-01 -1.30805767e+00 -1.10495150e+00
7.44061768e-02 -7.48250365e-01 3.77702683e-01 8.97056580e-01
5.31132281e-01 -1.64162695e-01 -4.71134007e-01 4.34646100e-01
1.01432455e+00 3.54991049e-01 6.93298280e-01 -8.36911380e-01
-5.85662305e-01 -4.00874794e-01 1.75669268e-01 -9.80821848e-01
-1.22915708e-01 -6.82032466e-01 3.63618612e-01 -8.50067377e-01
5.76436758e-01 -5.03542781e-01 -4.61055100e-01 2.53978550e-01
-6.34234846e-02 -1.61430046e-01 2.25977197e-01 1.41266987e-01
-1.53941727e+00 6.01150692e-01 4.17033732e-01 3.53553027e-01
-5.19579127e-02 4.87314016e-01 -6.04175568e-01 4.16696042e-01
1.07595849e+00 -9.14706349e-01 -2.01988623e-01 5.55803180e-02
4.16374654e-01 3.70997250e-01 1.56883463e-01 -7.83424854e-01
3.41391891e-01 -3.54640335e-01 -5.21921851e-02 -5.22169769e-01
-3.88717279e-02 -9.23541546e-01 2.99045324e-01 9.14676428e-01
-4.32965279e-01 5.81953563e-02 1.59497489e-03 1.14192796e+00
3.59473854e-01 -2.28422821e-01 7.97969043e-01 -1.85762599e-01
1.33964419e-01 5.46671271e-01 -6.78994834e-01 4.28881794e-01
1.04331148e+00 1.74861759e-01 -9.68898274e-03 -9.10705268e-01
-7.06143856e-01 8.39913249e-01 4.52897668e-01 -3.58103923e-02
2.12788895e-01 -1.17753828e+00 -7.19905019e-01 -5.74333429e-01
-1.69201747e-01 -1.04213826e-01 2.17319265e-01 1.00675797e+00
-2.94237465e-01 1.75812084e-03 3.93524557e-01 -5.59202731e-01
-9.44384336e-01 9.87854362e-01 3.38051736e-01 -3.99415381e-02
-5.52978158e-01 5.76078475e-01 -3.38727772e-01 -1.91023424e-01
5.96493304e-01 2.26624206e-01 5.76060832e-01 3.51512246e-02
4.87254351e-01 8.02360654e-01 -3.66887957e-01 7.92575404e-02
-4.62994784e-01 -9.88194942e-02 2.92737573e-01 -5.78292370e-01
1.57267952e+00 -6.63603365e-01 -1.87300146e-01 7.28537023e-01
7.66669631e-01 2.69380771e-02 -1.54464781e+00 -4.46270734e-01
4.08456594e-01 -5.97767711e-01 -1.28020391e-01 -5.66695631e-01
-1.07011318e+00 4.22638714e-01 5.85331619e-01 4.81071860e-01
1.08202457e+00 8.23074579e-02 4.45869774e-01 4.01447684e-01
8.06285203e-01 -1.33288503e+00 -2.50368267e-01 5.42542748e-02
4.06071782e-01 -1.03193474e+00 5.26136123e-02 -1.30658284e-01
-4.35278863e-01 8.29775810e-01 1.30996391e-01 -1.16494514e-01
1.20415568e+00 4.18130487e-01 -1.93418369e-01 7.07182959e-02
-1.37668490e+00 -3.35891634e-01 -5.40308952e-01 3.76968771e-01
9.16906148e-02 4.78945374e-01 -4.64876711e-01 8.64098608e-01
-8.53890255e-02 6.99570635e-03 1.03350961e+00 1.16892469e+00
-5.00366449e-01 -9.29438889e-01 -5.03836393e-01 4.55736011e-01
-7.67087102e-01 -5.72127365e-02 -3.88459377e-02 4.59645748e-01
-4.81064141e-01 1.42101622e+00 8.13031793e-02 -1.41391810e-02
-2.76031941e-01 -1.28909275e-02 1.88242808e-01 -3.10069740e-01
-3.39484781e-01 4.90624666e-01 -6.22033216e-02 -7.56526411e-01
-4.40861076e-01 -1.08948839e+00 -1.07730377e+00 -6.38774693e-01
-2.21595898e-01 4.76757020e-01 5.84025443e-01 5.50959229e-01
3.81895900e-01 5.26391380e-02 1.42420506e+00 -7.81937897e-01
-8.00220788e-01 -6.77080750e-01 -7.33303010e-01 9.89598930e-02
1.63745925e-01 -2.72768974e-01 -7.97151148e-01 -1.28352940e-02] | [4.620736122131348, 3.426583766937256] |
16d9321b-b9f6-4d93-886d-1090a1b30c35 | keep-your-distance-determining-sampling-and | 2207.05078 | null | https://arxiv.org/abs/2207.05078v1 | https://arxiv.org/pdf/2207.05078v1.pdf | Keep your Distance: Determining Sampling and Distance Thresholds in Machine Learning Monitoring | Machine Learning~(ML) has provided promising results in recent years across different applications and domains. However, in many cases, qualities such as reliability or even safety need to be ensured. To this end, one important aspect is to determine whether or not ML components are deployed in situations that are appropriate for their application scope. For components whose environments are open and variable, for instance those found in autonomous vehicles, it is therefore important to monitor their operational situation to determine its distance from the ML components' trained scope. If that distance is deemed too great, the application may choose to consider the ML component outcome unreliable and switch to alternatives, e.g. using human operator input instead. SafeML is a model-agnostic approach for performing such monitoring, using distance measures based on statistical testing of the training and operational datasets. Limitations in setting SafeML up properly include the lack of a systematic approach for determining, for a given application, how many operational samples are needed to yield reliable distance information as well as to determine an appropriate distance threshold. In this work, we address these limitations by providing a practical approach and demonstrate its use in a well known traffic sign recognition problem, and on an example using the CARLA open-source automotive simulator. | ['Daniel Schneider', 'Koorosh Aslansefat', 'Mohammed Naveed Akram', 'Andreas Schmidt', 'Ioannis Sorokos', 'Al-Harith Farhad'] | 2022-07-11 | null | null | null | null | ['traffic-sign-recognition'] | ['computer-vision'] | [ 9.50687379e-02 -1.95459634e-01 -2.28368178e-01 -4.99594808e-01
-5.39477229e-01 -4.63648856e-01 4.96456832e-01 3.20186108e-01
-5.45098841e-01 6.06929421e-01 -6.88289106e-01 -8.78153741e-01
-2.37440199e-01 -5.45253038e-01 -6.02326095e-01 -8.08270812e-01
-2.46335492e-02 7.20545173e-01 6.98395669e-01 -5.53712361e-02
3.81827652e-01 9.29849029e-01 -2.06055379e+00 4.91815545e-02
7.58861840e-01 1.08053923e+00 5.88860735e-02 7.21472085e-01
-1.83561981e-01 6.06075764e-01 -1.00961483e+00 1.32023141e-01
2.73993969e-01 -3.59546900e-01 -5.33248365e-01 1.88421354e-01
1.65400490e-01 -2.18038902e-01 4.91069108e-01 7.92959034e-01
1.55916423e-01 1.39990523e-01 6.92540944e-01 -1.58599246e+00
4.83254343e-01 1.64212406e-01 -4.84858118e-02 1.84475064e-01
1.58082440e-01 5.39079607e-01 5.82186401e-01 -3.39238614e-01
3.93867850e-01 8.44323575e-01 4.22715545e-01 3.26658487e-01
-1.45653605e+00 -6.37501180e-01 1.91894591e-01 3.35087985e-01
-1.34628356e+00 -6.02104962e-01 6.54116929e-01 -6.52306497e-01
8.09270501e-01 1.56518951e-01 2.89672971e-01 8.76678824e-01
3.01138878e-01 3.08143646e-01 1.27668071e+00 -6.38669014e-01
8.09924126e-01 6.60085917e-01 6.85046166e-02 2.37119436e-01
5.36167145e-01 1.03163712e-01 -8.97187926e-03 -1.50957718e-01
8.27774480e-02 -5.17171860e-01 8.59584063e-02 -5.52079558e-01
-6.82498038e-01 5.99696279e-01 -1.11962587e-01 4.61693138e-01
-3.84231895e-01 7.04276785e-02 5.05865812e-01 4.52734292e-01
-1.40608742e-03 3.60868186e-01 -4.44791555e-01 -4.74339366e-01
-8.96484077e-01 2.17970714e-01 8.73154581e-01 7.19145894e-01
8.45612049e-01 2.39413586e-02 2.41576657e-01 5.31464934e-01
2.42914841e-01 2.00698063e-01 3.16632986e-01 -8.02719533e-01
2.05298409e-01 7.50061393e-01 1.45254374e-01 -7.45241165e-01
-1.84232622e-01 -9.93560106e-02 -1.90768152e-01 1.03788733e+00
5.25985777e-01 -1.37602175e-02 -7.83583999e-01 1.50527799e+00
3.90820295e-01 1.50192499e-01 2.59546906e-01 6.65923238e-01
2.32851923e-01 3.27482253e-01 2.95575976e-01 -1.58477440e-01
1.06387055e+00 -1.30896658e-01 -4.20659900e-01 -3.89451087e-01
9.97756660e-01 -6.59043014e-01 9.84413564e-01 7.47838378e-01
-6.53396010e-01 -6.05687916e-01 -1.40731418e+00 6.56737745e-01
-6.79266274e-01 2.87889391e-02 6.40691370e-02 7.38698959e-01
-6.51560366e-01 5.38701952e-01 -6.03938580e-01 -5.23100853e-01
-5.99193014e-03 5.51098466e-01 -3.15022349e-01 -7.84077346e-02
-9.67311025e-01 1.18475902e+00 4.11742777e-01 2.48104095e-01
-7.48226345e-01 -1.52691633e-01 -7.05628157e-01 -2.81412750e-01
4.41229165e-01 -6.63818493e-02 1.32784235e+00 -9.92000222e-01
-1.30507207e+00 5.85811734e-01 1.20771438e-01 -5.65153718e-01
6.68823600e-01 1.28132328e-01 -7.50405788e-01 -2.51401395e-01
1.25094429e-01 1.95483908e-01 7.07642317e-01 -1.44155645e+00
-9.11369205e-01 -1.97273314e-01 2.77044296e-01 -1.80010408e-01
5.33957444e-02 2.47979403e-01 -2.95381904e-01 1.50032833e-01
-1.39490485e-01 -1.11423922e+00 -2.31463864e-01 -1.66679323e-01
-1.81860372e-01 -1.28217891e-01 1.02494478e+00 -5.11170685e-01
1.29030585e+00 -2.05172229e+00 -3.49955082e-01 6.54647768e-01
-1.66615114e-01 5.10676146e-01 1.60875276e-01 3.72813016e-01
7.49625787e-02 -1.91574788e-03 -2.26467147e-01 6.31956160e-02
3.51145561e-03 4.24988180e-01 1.00332923e-01 5.20759523e-01
3.43454391e-01 1.34222418e-01 -6.63240373e-01 -6.21960342e-01
5.86264551e-01 7.92915523e-02 -1.50311321e-01 2.71680325e-01
-2.56000429e-01 3.27951640e-01 -5.19301951e-01 5.48090339e-01
4.16630119e-01 2.44335920e-01 1.36513472e-01 -4.43330407e-02
-2.86125809e-01 5.30459546e-02 -1.50958776e+00 7.17115521e-01
-7.53328860e-01 7.46081233e-01 1.07990302e-01 -1.11760867e+00
1.03516006e+00 2.82482207e-01 4.21464711e-01 -7.60110855e-01
3.39932680e-01 6.48838222e-01 3.94629538e-01 -7.62100637e-01
2.89194047e-01 -1.30067065e-01 -1.24078788e-01 3.98036629e-01
-4.41921383e-01 -4.58112061e-02 2.54002184e-01 -3.08846176e-01
1.12498951e+00 1.70664221e-01 2.61691302e-01 -1.39553785e-01
8.12803686e-01 2.92424768e-01 4.76625234e-01 3.32493335e-01
-3.22260737e-01 3.12322915e-01 5.32529652e-01 -1.77638695e-01
-7.41756797e-01 -6.43651903e-01 -2.81948179e-01 6.27121687e-01
1.66272357e-01 -1.21432347e-02 -7.24577904e-01 -8.34989130e-01
1.62560761e-01 1.17877436e+00 -3.66816461e-01 -3.87551486e-01
-5.80861509e-01 -2.52059877e-01 4.43602443e-01 4.69399393e-01
2.89308965e-01 -1.01568687e+00 -1.20521581e+00 4.51407373e-01
2.47251362e-01 -1.10438251e+00 2.29661599e-01 4.86596763e-01
-7.33332515e-01 -1.26100743e+00 -1.36277480e-02 -2.65162945e-01
7.69387543e-01 -1.34455003e-02 8.52061927e-01 1.03228517e-01
-2.52887994e-01 1.70217887e-01 -4.38251704e-01 -3.61908525e-01
-1.02220237e+00 -1.27117157e-01 6.00567684e-02 1.33554861e-01
4.99440789e-01 -1.82569534e-01 -2.77038276e-01 8.00999999e-01
-9.86145377e-01 -5.00589311e-01 8.58370364e-01 4.10376042e-01
3.89352411e-01 3.97247940e-01 7.32113600e-01 -6.79190576e-01
7.01671004e-01 -4.51255590e-01 -1.01274943e+00 3.09813946e-01
-9.93991852e-01 1.08463071e-01 7.31853306e-01 -5.43829143e-01
-6.75796092e-01 -2.83955573e-03 -2.98929214e-01 -3.14899921e-01
-8.26900899e-01 4.87936705e-01 -4.45689559e-01 -1.66785456e-02
6.85715020e-01 -1.03880331e-01 4.19935673e-01 -2.55683452e-01
-1.51024714e-01 1.13228941e+00 1.92444950e-01 -5.50333261e-01
8.98885369e-01 1.51401103e-01 9.41415280e-02 -8.11719716e-01
-1.67512760e-01 -6.43112600e-01 -6.10698462e-01 -6.30682647e-01
5.84678054e-01 -2.49537200e-01 -6.32249832e-01 8.48188773e-02
-7.63492644e-01 -5.16741574e-01 -1.91693887e-01 4.85135615e-01
-5.41056275e-01 7.58162588e-02 2.01228738e-01 -1.14400887e+00
1.89010665e-01 -1.48079967e+00 7.82708347e-01 1.00844562e-01
-4.91884112e-01 -9.02486205e-01 -2.41828278e-01 3.64909828e-01
4.87331271e-01 2.80084699e-01 8.23180616e-01 -9.82705116e-01
-4.49868858e-01 -7.62877822e-01 2.03047156e-01 7.49233007e-01
3.13483000e-01 3.60750794e-01 -9.34697747e-01 -3.05594057e-01
-9.90138426e-02 -1.50652632e-01 2.58488238e-01 1.19543495e-02
8.21972549e-01 -1.88374135e-04 -3.91856700e-01 2.14249194e-02
1.31929743e+00 6.61415517e-01 6.17314458e-01 8.16163123e-01
1.45481586e-01 8.61090720e-01 1.30782688e+00 9.27909762e-02
7.07115307e-02 9.41376209e-01 4.97656941e-01 3.48390788e-02
1.16010964e-01 1.77388549e-01 6.19183660e-01 1.36889637e-01
2.93672919e-01 -1.03847548e-01 -9.24691856e-01 3.67450535e-01
-1.69630539e+00 -8.07632744e-01 -1.36142686e-01 2.74837923e+00
3.57478440e-01 7.14314520e-01 3.87243062e-01 6.18570387e-01
4.62761492e-01 -2.52711058e-01 -5.26770711e-01 -8.97302210e-01
1.88081697e-01 -3.21897924e-01 5.56989253e-01 3.34193468e-01
-7.96284914e-01 4.59629625e-01 5.28399611e+00 5.61382711e-01
-1.34962165e+00 -1.73061848e-01 5.50329626e-01 1.00666143e-01
6.89314082e-02 2.75143653e-01 -8.15915167e-01 6.37584746e-01
1.32331669e+00 -5.60814217e-02 6.34679943e-02 8.78657877e-01
5.54387867e-01 -6.49687290e-01 -1.24366200e+00 5.09097040e-01
-5.58354519e-03 -7.62765825e-01 -5.86076736e-01 2.70804524e-01
6.78867102e-02 -6.71394691e-02 -3.16215426e-01 4.08425033e-01
-4.26000878e-02 -8.52353632e-01 8.47470701e-01 3.65212500e-01
6.14220321e-01 -7.61646867e-01 1.01371241e+00 6.03704393e-01
-1.00713682e+00 -2.57501453e-01 -2.12147646e-02 2.03637794e-01
1.63925543e-01 3.66406232e-01 -1.19660389e+00 5.45651078e-01
6.16107643e-01 -8.24797377e-02 -5.84419966e-01 1.24347973e+00
-1.04840575e-02 6.86450899e-01 -5.73461950e-01 -1.88230798e-01
1.76414728e-01 -1.29821002e-01 5.41994631e-01 1.18528020e+00
2.52155244e-01 -4.71651345e-01 8.92937630e-02 7.35005558e-01
6.80600107e-01 7.01724365e-02 -9.33040380e-01 2.36778006e-01
4.41705555e-01 1.10188437e+00 -6.96976364e-01 -7.89542720e-02
-4.83899593e-01 4.01716858e-01 -1.41929910e-01 3.78230929e-01
-9.65731263e-01 -3.35043252e-01 7.61693895e-01 4.27621335e-01
2.78056264e-01 -1.62061602e-01 -1.32343650e-01 -2.07671523e-01
3.06423128e-01 -9.94957745e-01 2.54842579e-01 -4.02552128e-01
-7.29464710e-01 6.06507957e-01 4.31173325e-01 -1.76855898e+00
-5.63263237e-01 -7.98186898e-01 -6.64787710e-01 8.21918249e-01
-1.35073519e+00 -8.16181004e-01 -2.00488940e-01 9.21947956e-02
4.03973937e-01 -1.66550457e-01 4.70170736e-01 3.65845889e-01
-6.42344058e-01 4.58880097e-01 -8.89012441e-02 -3.25227648e-01
6.15859509e-01 -1.07688987e+00 -5.66219129e-02 8.58313143e-01
-1.93908229e-01 2.96204239e-01 1.12672091e+00 -5.62633932e-01
-1.38188577e+00 -9.61368799e-01 7.24948049e-01 -4.21386421e-01
5.15323162e-01 -2.32265890e-01 -1.06229365e+00 4.14541274e-01
-3.13698769e-01 4.07219343e-02 4.16888356e-01 4.59410585e-05
3.05132624e-02 -6.09232187e-01 -1.36219776e+00 6.78357959e-01
4.52763140e-01 -2.77742088e-01 -3.69904965e-01 -4.09485772e-02
-2.69960295e-02 -1.74345031e-01 -7.72287965e-01 6.56633139e-01
4.66010451e-01 -1.18563390e+00 4.54035759e-01 -3.54466498e-01
-3.37277144e-01 -7.57244229e-01 -7.77415037e-02 -1.03665614e+00
1.70761228e-01 -3.65797281e-01 3.23845185e-02 1.46721065e+00
7.20814228e-01 -8.69298458e-01 7.22778201e-01 9.90483642e-01
-6.41473308e-02 -8.58686447e-01 -1.09795272e+00 -1.07645476e+00
-1.87557817e-01 -9.61971343e-01 5.87828338e-01 4.22096640e-01
-2.58714914e-01 4.40296270e-02 2.75154952e-02 4.21887577e-01
3.66729498e-01 -1.43613338e-01 1.18625009e+00 -1.29326797e+00
-8.09903592e-02 -5.07197380e-01 -8.13197732e-01 -4.47231412e-01
1.99439541e-01 -2.83032387e-01 5.06458461e-01 -1.37343049e+00
-3.62696648e-01 -8.03752482e-01 -3.08860630e-01 3.09529871e-01
2.17951685e-01 -1.54980704e-01 -8.78086761e-02 1.07245065e-01
-3.91922742e-01 1.53527139e-02 5.00387132e-01 -7.98542351e-02
-2.73310483e-01 5.09676754e-01 -2.07418814e-01 5.84952354e-01
9.24439430e-01 -3.14486712e-01 -4.66486007e-01 9.48317274e-02
-6.36597872e-02 -1.69126585e-01 4.32926416e-01 -1.39224589e+00
3.29901189e-01 -4.00707722e-01 -1.13544814e-01 -4.92759049e-01
1.84444726e-01 -1.43507302e+00 3.78024548e-01 5.34162879e-01
-1.24641083e-01 7.07430765e-02 4.43314165e-01 3.66433918e-01
-2.22249404e-01 -6.35199189e-01 7.60543466e-01 3.52859199e-01
-1.02392423e+00 -1.58477977e-01 -7.40097940e-01 -4.63324547e-01
1.63265634e+00 -8.86934698e-01 -1.78195089e-02 -1.28458723e-01
-5.02198517e-01 2.62217194e-01 6.98031604e-01 4.51576918e-01
5.23709774e-01 -8.99826109e-01 -3.21113855e-01 3.82121444e-01
6.85342669e-01 -2.11262047e-01 -1.24209017e-01 1.00046587e+00
-5.23512483e-01 3.44799489e-01 -1.96894053e-02 -6.68344140e-01
-1.50414133e+00 5.80189049e-01 5.49609184e-01 1.23946935e-01
-4.83491093e-01 1.13233402e-01 -3.82882297e-01 -1.92275658e-01
3.88173848e-01 -5.44651389e-01 -2.23646119e-01 -6.16426878e-02
3.73574376e-01 3.28083098e-01 5.99573076e-01 -8.57588947e-01
-5.33157170e-01 4.53254193e-01 2.09384277e-01 -1.09876700e-01
8.81856859e-01 -2.35615317e-02 2.47409835e-01 7.21390009e-01
1.02104568e+00 -6.79597482e-02 -1.36338091e+00 2.41743654e-01
4.30939704e-01 -4.10412818e-01 7.24140629e-02 -7.03067422e-01
-7.99745977e-01 6.04108453e-01 8.94864380e-01 6.19144022e-01
1.11643994e+00 2.39413921e-02 3.37816745e-01 3.53101343e-01
6.88777447e-01 -1.51169431e+00 -3.64314735e-01 8.28623548e-02
7.03502297e-01 -1.25719941e+00 -6.50501326e-02 -9.29555148e-02
-5.68184495e-01 1.04120266e+00 7.33049393e-01 2.12429211e-01
5.75875282e-01 4.75268871e-01 2.85944581e-01 -1.55732483e-01
-7.21162498e-01 -3.39531362e-01 1.54576957e-01 6.21969461e-01
1.09647989e-01 7.83461332e-02 -2.43149549e-01 2.62867324e-02
1.82793826e-01 1.31901786e-01 4.18977141e-01 1.33706236e+00
-6.23436868e-01 -1.25504398e+00 -6.40605390e-01 6.10533834e-01
-2.69317389e-01 6.24683321e-01 -4.34063345e-01 1.08255148e+00
4.22571957e-01 1.30887890e+00 -5.64310402e-02 -5.22875488e-01
7.55792320e-01 2.80886471e-01 1.24435306e-01 -5.75754344e-01
-4.33904320e-01 -2.16243967e-01 6.60134077e-01 -4.79235023e-01
-2.13439658e-01 -9.35202301e-01 -1.28363454e+00 -9.70433280e-02
-5.14585376e-01 1.39606297e-01 1.09390724e+00 1.10860145e+00
1.32148892e-01 4.69566494e-01 7.07769215e-01 -6.67680383e-01
-7.33771265e-01 -5.74171424e-01 -4.03478026e-01 3.91168475e-01
2.48015761e-01 -1.00045002e+00 -5.94911397e-01 -2.31377840e-01] | [5.904052734375, 1.1769529581069946] |
33bf3a5c-49ad-41a6-a7d1-1a94f432589c | clip-surgery-for-better-explainability-with | 2304.05653 | null | https://arxiv.org/abs/2304.05653v1 | https://arxiv.org/pdf/2304.05653v1.pdf | CLIP Surgery for Better Explainability with Enhancement in Open-Vocabulary Tasks | Contrastive Language-Image Pre-training (CLIP) is a powerful multimodal large vision model that has demonstrated significant benefits for downstream tasks, including many zero-shot learning and text-guided vision tasks. However, we notice some severe problems regarding the model's explainability, which undermines its credibility and impedes related tasks. Specifically, we find CLIP prefers the background regions than the foregrounds according to the predicted similarity map, which contradicts human understanding. Besides, there are obvious noisy activations on the visualization results at irrelevant positions. To address these two issues, we conduct in-depth analyses and reveal the reasons with new findings and evidences. Based on these insights, we propose the CLIP Surgery, a method that enables surgery-like modifications for the inference architecture and features, for better explainability and enhancement in multiple open-vocabulary tasks. The proposed method has significantly improved the explainability of CLIP for both convolutional networks and vision transformers, surpassing existing methods by large margins. Besides, our approach also demonstrates remarkable improvements in open-vocabulary segmentation and multi-label recognition tasks. For examples, the mAP improvement on NUS-Wide multi-label recognition is 4.41% without any additional training, and our CLIP Surgery surpasses the state-of-the-art method by 8.74% at mIoU on Cityscapes open-vocabulary semantic segmentation. Furthermore, our method benefits other tasks including multimodal visualization and interactive segmentation like Segment Anything Model (SAM). The code is available at https://github.com/xmed-lab/CLIP_Surgery | ['Xiaomeng Li', 'Yiqun Duan', 'Hualiang Wang', 'Yi Li'] | 2023-04-12 | null | null | null | null | ['interactive-segmentation'] | ['computer-vision'] | [ 2.69396394e-01 3.46055627e-01 -1.81263342e-01 -3.02453429e-01
-8.96639168e-01 -5.36108315e-01 5.37826061e-01 -4.54567485e-02
-3.83419156e-01 4.60338622e-01 1.80281788e-01 -4.20297176e-01
-1.91139337e-02 -3.26733857e-01 -8.02214086e-01 -6.62157834e-01
5.84598422e-01 4.23471779e-01 2.40485862e-01 -1.86786443e-01
8.30052570e-02 -6.86234832e-02 -1.73604584e+00 4.88101333e-01
1.21884704e+00 8.33191395e-01 3.81926507e-01 4.25636947e-01
-3.30469698e-01 4.94776666e-01 -2.96921313e-01 -6.75992131e-01
-1.13568582e-01 -2.30723083e-01 -7.50323296e-01 -1.14869185e-01
7.49816537e-01 -2.49876559e-01 -1.89574525e-01 1.32739925e+00
3.80934000e-01 4.47571389e-02 6.90081894e-01 -1.23509693e+00
-1.02536106e+00 7.07141936e-01 -8.65961671e-01 -2.01596208e-02
-1.11774363e-01 3.42949033e-01 1.21580911e+00 -1.01822746e+00
7.08461702e-01 1.13789451e+00 4.88551080e-01 7.37673104e-01
-1.03566957e+00 -7.37233818e-01 2.62883663e-01 3.01570714e-01
-1.33600569e+00 -3.01380664e-01 4.22755033e-01 -4.61858988e-01
7.87073731e-01 4.86722618e-01 5.31961501e-01 1.31287801e+00
-1.34619221e-01 9.56368387e-01 1.05933094e+00 -3.11975896e-01
-9.18940529e-02 2.76183784e-01 4.23526645e-01 9.63854790e-01
2.53155082e-01 1.66684091e-02 -4.26529229e-01 3.46558601e-01
5.35839975e-01 1.01047352e-01 -3.13388705e-01 -2.26537809e-01
-1.35437393e+00 7.37311959e-01 6.29774094e-01 3.31095487e-01
7.49300942e-02 2.62961328e-01 4.55846846e-01 -1.79118752e-01
5.72892010e-01 2.54299283e-01 -4.35571790e-01 8.08379948e-02
-9.48830068e-01 -1.63153350e-01 3.64179760e-01 9.83274579e-01
7.02689826e-01 1.61925983e-02 -4.57269430e-01 9.63769555e-01
3.64185631e-01 5.58064044e-01 3.31803888e-01 -9.08425748e-01
4.52723354e-01 6.32082760e-01 -2.95743585e-01 -9.36697721e-01
-5.15255690e-01 -6.32215142e-01 -1.00520241e+00 3.35719958e-02
5.72799027e-01 -6.18769266e-02 -1.23585200e+00 1.66935968e+00
1.51690021e-01 4.23240811e-01 1.59149512e-03 1.24920225e+00
1.44019902e+00 5.51568985e-01 4.07983899e-01 -5.04561104e-02
1.70153356e+00 -1.45295191e+00 -9.09098864e-01 -4.63516384e-01
7.79770255e-01 -5.92275858e-01 1.50776756e+00 2.78297842e-01
-6.63982570e-01 -6.57222867e-01 -9.24901187e-01 -3.82033885e-01
-6.16634667e-01 2.57759422e-01 8.30546975e-01 5.86809874e-01
-9.42947507e-01 2.46175453e-01 -4.44940746e-01 -5.03600478e-01
8.00008059e-01 1.79450214e-01 -2.85341352e-01 -9.53304619e-02
-1.25551748e+00 6.24171615e-01 3.72948825e-01 5.02749197e-02
-8.55115414e-01 -7.92243958e-01 -9.66498494e-01 2.87358403e-01
6.93708539e-01 -7.16323674e-01 1.04675043e+00 -8.64187658e-01
-1.07230747e+00 1.04948401e+00 -2.56124854e-01 -3.29959124e-01
6.07071698e-01 -1.66091263e-01 -3.15244377e-01 1.72330707e-01
1.30846694e-01 1.21408534e+00 7.06881523e-01 -1.28560221e+00
-5.44776678e-01 -4.74867731e-01 6.77879453e-02 2.29538575e-01
-3.85841131e-01 -2.21064001e-01 -1.04870391e+00 -5.50184906e-01
-3.08012571e-02 -8.68681550e-01 -2.25424916e-01 -4.52062227e-02
-6.82068110e-01 -1.09865606e-01 4.89408791e-01 -7.00195909e-01
1.17067385e+00 -2.31532884e+00 -6.20094687e-02 -1.67048741e-02
5.49765468e-01 2.75193572e-01 -2.34073043e-01 -1.38979480e-01
-6.16207942e-02 4.36367810e-01 -3.18258017e-01 -4.27883029e-01
7.79469982e-02 1.29601657e-01 -2.62757450e-01 2.72030056e-01
5.13865836e-02 1.31043065e+00 -6.65996313e-01 -6.97579443e-01
3.95875126e-01 4.88779813e-01 -5.16510844e-01 -2.49450251e-01
-3.97902310e-01 5.81658661e-01 -2.86659807e-01 7.40560949e-01
6.75964713e-01 -6.12496555e-01 -4.04728390e-02 -3.27278465e-01
-1.06983162e-01 -4.47862148e-01 -9.83693957e-01 1.99902570e+00
-1.94064856e-01 7.67244160e-01 -2.08933860e-01 -8.79881859e-01
6.52603745e-01 2.19293445e-01 1.76168665e-01 -8.58320713e-01
2.98239857e-01 1.26269490e-01 -1.34467959e-01 -7.28235066e-01
4.07806307e-01 7.43857995e-02 -4.83175442e-02 3.16071436e-02
1.10480234e-01 1.35731429e-01 1.04787657e-02 4.72917169e-01
5.16107976e-01 3.62742729e-02 9.88979936e-02 -1.86478719e-01
4.78885829e-01 1.32331923e-01 4.08254683e-01 9.01546896e-01
-4.34497714e-01 7.84153819e-01 6.76574111e-01 -1.71837628e-01
-7.60799885e-01 -8.53392422e-01 -1.60083845e-01 1.32197297e+00
6.65193856e-01 -3.88281226e-01 -9.66845572e-01 -8.21735084e-01
-1.11885332e-01 9.10498500e-01 -7.93162882e-01 -5.56783155e-02
7.53681408e-03 -6.46057963e-01 6.67962134e-01 5.23256421e-01
5.80684602e-01 -9.66944695e-01 -2.48278528e-01 -3.25294256e-01
-5.62554061e-01 -1.27781820e+00 -5.40082157e-01 2.35228632e-02
-6.07039392e-01 -1.01916313e+00 -9.14707243e-01 -8.46004486e-01
6.22292042e-01 5.39983451e-01 7.82696247e-01 1.54770821e-01
-5.02697289e-01 2.35763788e-01 -1.49804488e-01 -4.72187221e-01
-7.47740343e-02 1.37882352e-01 -2.26937473e-01 -4.57610050e-03
3.52831751e-01 -1.41538352e-01 -6.88874125e-01 3.61722678e-01
-7.70358264e-01 5.72307348e-01 5.56372404e-01 9.11272407e-01
8.10526192e-01 -3.64633322e-01 5.21804810e-01 -1.03666139e+00
4.05880392e-01 -3.38080376e-01 -3.93312991e-01 4.50189710e-01
-7.39244044e-01 -4.06962596e-02 1.77535042e-01 -3.44034940e-01
-1.24704421e+00 -2.29208805e-02 -7.87033737e-02 -5.80317140e-01
-4.87256110e-01 2.77486384e-01 -2.00758368e-01 1.67576015e-01
5.70842803e-01 6.64577186e-02 2.56918091e-03 -3.27085495e-01
8.42924893e-01 7.90390968e-01 6.11880541e-01 -2.40700200e-01
5.17297387e-01 6.16369724e-01 -3.80920470e-01 -8.32145333e-01
-1.11823273e+00 -5.43431222e-01 -5.98677576e-01 -3.29246491e-01
1.33625424e+00 -1.09686160e+00 -8.55597079e-01 2.48513401e-01
-1.22610247e+00 -3.46283495e-01 -2.99476832e-02 2.41550133e-01
-3.34097028e-01 3.09232116e-01 -4.94949192e-01 -6.70528769e-01
-3.55196238e-01 -1.35225606e+00 1.21074879e+00 5.81781089e-01
-1.33468017e-01 -8.93073440e-01 -4.01296586e-01 9.85626876e-01
2.68391579e-01 1.41185559e-02 9.95024145e-01 -8.01133990e-01
-6.41550660e-01 2.17949763e-01 -8.51955414e-01 3.48327942e-02
-4.05849040e-01 -2.69955285e-02 -1.41157436e+00 1.29523487e-05
-4.88301992e-01 -3.58448446e-01 1.19784939e+00 6.88030362e-01
1.44264472e+00 1.85700759e-01 -5.37194014e-01 8.41328979e-01
1.32923532e+00 -5.19108735e-02 5.12842476e-01 2.60707974e-01
9.72203314e-01 7.52851009e-01 6.64426029e-01 1.65412217e-01
4.27237481e-01 5.42619705e-01 7.29859412e-01 -5.84263086e-01
-3.27967167e-01 -2.19720736e-01 4.65149432e-02 5.69356024e-01
-1.03402965e-01 -2.11937234e-01 -9.49202240e-01 5.51479101e-01
-2.13796544e+00 -7.12797761e-01 -6.12248003e-01 1.76565838e+00
5.38239658e-01 1.24579221e-01 -2.20169842e-01 -4.03038025e-01
8.38996947e-01 3.52411866e-01 -7.14625776e-01 -3.23083162e-01
-1.74134672e-01 -1.45711064e-01 4.28848207e-01 4.86965120e-01
-1.11547434e+00 1.44541812e+00 5.07246637e+00 1.25646007e+00
-9.50535655e-01 4.79407877e-01 9.42045987e-01 -1.66697145e-01
-5.38363278e-01 -2.19993740e-01 -8.26223433e-01 2.62952775e-01
6.27697945e-01 1.20533176e-01 3.42960387e-01 7.63299286e-01
1.49326846e-01 -9.84922051e-02 -8.43766689e-01 1.19534445e+00
2.82078624e-01 -1.53668308e+00 1.96087912e-01 6.12058416e-02
7.44409919e-01 1.49599373e-01 3.08662176e-01 6.06054723e-01
-1.08416326e-01 -1.13303399e+00 6.44111037e-01 4.68834639e-01
1.12191284e+00 -6.06491923e-01 8.06711614e-01 3.73788238e-01
-9.47134256e-01 4.47355397e-02 -3.38466823e-01 1.80630922e-01
1.83757335e-01 4.10649180e-01 -4.68077898e-01 3.99040163e-01
6.68994963e-01 6.20329082e-01 -7.82783151e-01 8.12209964e-01
-3.19360137e-01 6.30302787e-01 5.31646758e-02 5.34294657e-02
4.94524121e-01 -1.98055908e-01 3.42445672e-01 1.30005980e+00
1.49899155e-01 1.45442992e-01 -6.23026378e-02 1.10654807e+00
-1.79149196e-01 2.89876491e-01 -6.45118117e-01 2.52178889e-02
1.54623866e-01 1.45414484e+00 -8.29290330e-01 -5.52889884e-01
-7.10369825e-01 1.03800714e+00 2.84757942e-01 6.45234406e-01
-1.32694054e+00 -2.54367888e-01 5.32046914e-01 -2.26055905e-01
2.25258321e-01 9.08575803e-02 -7.10209370e-01 -1.28638554e+00
-2.36753404e-01 -7.54853845e-01 4.54599500e-01 -8.73200297e-01
-8.90025377e-01 5.19769788e-01 -3.21206748e-01 -1.00359190e+00
3.08866322e-01 -6.34070754e-01 -4.09614474e-01 6.83451295e-01
-1.58327687e+00 -1.40234208e+00 -4.96999502e-01 6.05628848e-01
9.68723297e-01 -6.42102510e-02 6.90918505e-01 3.62460643e-01
-8.95075619e-01 8.01434994e-01 5.34656048e-02 2.68381629e-02
8.77295256e-01 -1.12810004e+00 1.99526966e-01 7.24187315e-01
4.85268742e-01 4.31954890e-01 6.19789243e-01 -6.15135610e-01
-9.35336471e-01 -1.01422918e+00 6.89866602e-01 -5.22265792e-01
7.29468882e-01 -3.68063033e-01 -9.63305831e-01 5.94613135e-01
3.44397545e-01 2.14630691e-03 7.41122067e-01 2.21658498e-01
-4.36897844e-01 9.91113111e-02 -8.26492190e-01 8.68411779e-01
1.28007424e+00 -4.23245490e-01 -4.98724341e-01 4.36616659e-01
1.03543687e+00 -3.30075085e-01 -5.14805973e-01 3.52297217e-01
6.02733493e-01 -9.43032026e-01 9.41833377e-01 -6.34246528e-01
7.35752940e-01 -1.03193082e-01 -2.33236849e-01 -1.02340817e+00
-2.78332710e-01 -2.21983254e-01 -5.06109446e-02 1.31771505e+00
7.44870722e-01 -5.21491766e-01 6.14322424e-01 6.84800088e-01
-2.51035184e-01 -7.41833925e-01 -7.62880802e-01 -4.47086930e-01
-8.08640346e-02 -7.06388652e-01 2.89590359e-01 1.07823396e+00
-7.43494183e-02 5.76089025e-01 -3.94034803e-01 2.94927716e-01
6.53648734e-01 1.76649123e-01 6.19092643e-01 -1.15834641e+00
-5.59946038e-02 -6.32970691e-01 3.25064026e-02 -1.08482850e+00
2.98754662e-01 -1.17559469e+00 -4.19427454e-02 -1.72763741e+00
6.14073753e-01 -1.61130130e-01 -3.78450096e-01 7.13800371e-01
-3.73373538e-01 5.35276473e-01 4.19205606e-01 2.26469859e-01
-9.56939697e-01 5.57683945e-01 1.64722896e+00 -4.67697620e-01
-4.21218714e-03 -3.21959347e-01 -1.05122459e+00 8.73792946e-01
7.34316468e-01 -2.42780477e-01 -3.98329258e-01 -5.58139980e-01
8.39513391e-02 -1.10107705e-01 4.99952793e-01 -6.96313322e-01
1.83118045e-01 -1.28604909e-02 2.61566281e-01 -5.17760217e-01
2.01122597e-01 -7.56621361e-01 -1.38283700e-01 4.22323227e-01
-4.85168606e-01 -4.33444381e-01 3.94167751e-01 7.62478590e-01
-1.72251523e-01 -1.25553101e-01 6.36621058e-01 -2.46701743e-02
-1.18379211e+00 2.88459599e-01 -1.61191687e-01 1.18040726e-01
1.06767106e+00 -2.68744767e-01 -7.71800756e-01 -2.19631553e-01
-9.70434904e-01 5.60587466e-01 2.80198812e-01 6.13169730e-01
5.27719975e-01 -1.06859934e+00 -5.65090895e-01 2.33712733e-01
4.39188927e-01 1.88119663e-03 7.49773383e-01 1.33966637e+00
-3.60057890e-01 5.97682595e-01 -6.47239201e-03 -8.39759231e-01
-1.42744970e+00 5.63575089e-01 1.81278542e-01 -7.55080208e-02
-8.16566229e-01 9.73945737e-01 8.29073727e-01 -4.97852385e-01
4.29907560e-01 -3.66847008e-01 -4.56900269e-01 2.63253391e-01
4.95961875e-01 2.20227689e-01 -1.49041921e-01 -5.51005661e-01
-2.85014927e-01 7.45955288e-01 -1.80965111e-01 3.28501426e-02
1.03770840e+00 -4.03165728e-01 5.83811216e-02 4.28418756e-01
9.21077549e-01 -1.95649073e-01 -1.35984552e+00 -1.70080528e-01
-2.18269512e-01 -2.41413713e-01 9.22760367e-02 -1.11234879e+00
-1.26395369e+00 1.28955054e+00 6.93637073e-01 -1.47496704e-02
8.38326037e-01 3.75568271e-01 7.31975794e-01 1.01044483e-01
1.28944397e-01 -1.10071695e+00 2.17893571e-02 5.17397702e-01
8.17804575e-01 -1.78869462e+00 -3.36622953e-01 -6.30534768e-01
-1.15672183e+00 8.24415684e-01 7.87081718e-01 4.47202116e-01
3.51278067e-01 -7.15598017e-02 3.26269984e-01 -4.33934838e-01
-5.24456620e-01 -5.38102150e-01 5.47009826e-01 5.61053634e-01
3.01507652e-01 2.05795124e-01 -2.24898815e-01 9.36707675e-01
4.39418070e-02 -2.33317763e-01 1.65970549e-01 -3.08333561e-02
-4.75502759e-01 -4.21507925e-01 -2.62834728e-01 4.23869580e-01
-1.75583065e-01 -4.29390371e-01 -3.26993138e-01 9.67488289e-01
3.35650206e-01 8.18775654e-01 2.26589404e-02 -3.05128664e-01
1.68734953e-01 1.99977428e-01 4.80610505e-02 -4.74615157e-01
-2.94512630e-01 2.99255222e-01 8.84151533e-02 -6.96246922e-01
-2.05174014e-01 -2.90080816e-01 -1.58666873e+00 -1.32028714e-01
-5.31205654e-01 -1.83348715e-01 6.22471571e-01 1.06423426e+00
4.32228029e-01 9.65046763e-01 -8.52633864e-02 -5.71244299e-01
-1.02339633e-01 -8.93903553e-01 -5.09139657e-01 4.18292820e-01
9.39510167e-02 -7.50053287e-01 -3.51520985e-01 1.17357962e-01] | [9.893745422363281, 0.7048428058624268] |
69b02b48-6794-4602-b673-a33d5320568e | off-policy-evaluation-in-embedded-spaces | 2203.02807 | null | https://arxiv.org/abs/2203.02807v2 | https://arxiv.org/pdf/2203.02807v2.pdf | Off-Policy Evaluation in Embedded Spaces | Off-policy evaluation methods are important in recommendation systems and search engines, where data collected under an existing logging policy is used to estimate the performance of a new proposed policy. A common approach to this problem is weighting, where data is weighted by a density ratio between the probability of actions given contexts in the target and logged policies. In practice, two issues often arise. First, many problems have very large action spaces and we may not observe rewards for most actions, and so in finite samples we may encounter a positivity violation. Second, many recommendation systems are not probabilistic and so having access to logging and target policy densities may not be feasible. To address these issues, we introduce the featurized embedded permutation weighting estimator. The estimator computes the density ratio in an action embedding space, which reduces the possibility of positivity violations. The density ratio is computed leveraging recent advances in normalizing flows and density ratio estimation as a classification problem, in order to obtain estimates which are feasible in practice. | ['Georgios Theocharous', 'David Arbour', 'Jaron J. R. Lee'] | 2022-03-05 | null | null | null | null | ['density-ratio-estimation'] | ['methodology'] | [ 2.14254707e-01 -3.55005205e-01 -7.93497384e-01 -1.74937859e-01
-5.81914485e-01 -7.87344098e-01 6.99202001e-01 2.90647596e-01
-6.33199453e-01 8.53429019e-01 6.41122580e-01 -7.51041591e-01
-4.30490136e-01 -9.00810421e-01 -5.06046832e-01 -5.15981138e-01
-2.66099632e-01 3.77474070e-01 8.31267014e-02 3.79646391e-01
4.52848673e-01 3.56843919e-01 -1.45084321e+00 -4.00731824e-02
8.72394741e-01 1.04106832e+00 -2.03146964e-01 8.30596566e-01
-2.01209828e-01 8.56660545e-01 -6.64470196e-01 -3.00225258e-01
4.12520528e-01 -2.29394868e-01 -5.71705937e-01 -2.55663604e-01
4.18894231e-01 -6.63492978e-01 -2.64778286e-01 1.29426992e+00
1.79513797e-01 6.26462221e-01 9.63279784e-01 -1.64934874e+00
-3.05959135e-01 5.97350061e-01 -6.19156122e-01 4.12594348e-01
2.47759804e-01 6.97702821e-03 1.41762197e+00 -3.98384631e-01
2.17818633e-01 1.33257759e+00 9.59097445e-02 3.06652814e-01
-1.68869805e+00 -6.51181638e-01 2.12036178e-01 1.13521174e-01
-9.39862549e-01 -5.63903451e-01 3.70876878e-01 -6.18520319e-01
6.81959391e-01 2.86913037e-01 4.84504670e-01 1.13353169e+00
3.38304974e-02 7.44491518e-01 1.07767510e+00 -2.05277100e-01
6.30833030e-01 4.25312400e-01 1.70752749e-01 2.89029211e-01
5.87532103e-01 4.50138122e-01 -4.75301296e-01 -8.00287426e-01
6.07051373e-01 2.19245866e-01 -1.83927089e-01 -5.84025502e-01
-9.68527615e-01 1.01755524e+00 -1.10283732e-01 1.15168408e-01
-5.07259727e-01 3.06267232e-01 5.18287838e-01 2.85448223e-01
4.24287736e-01 4.30642754e-01 -2.36570567e-01 -7.43752658e-01
-9.97330785e-01 5.68217814e-01 9.78220165e-01 5.84466577e-01
5.32573223e-01 -2.75595456e-01 -4.75651860e-01 5.78634322e-01
3.27662081e-01 4.10082817e-01 1.97562248e-01 -1.23380005e+00
5.63944817e-01 2.92342842e-01 5.39126098e-01 -7.61285722e-01
1.77064508e-01 5.58264367e-02 -3.80096674e-01 3.76201361e-01
9.81815696e-01 -7.44756237e-02 -3.90834451e-01 1.98213196e+00
2.63263404e-01 2.54598796e-01 -2.11130649e-01 6.35999799e-01
-4.71841604e-01 5.61972976e-01 4.09106582e-01 -3.13464075e-01
1.19948590e+00 -4.01323110e-01 -6.58067942e-01 -1.21537507e-01
5.58623254e-01 -5.61633348e-01 1.44087148e+00 3.85036051e-01
-9.27256584e-01 1.48806174e-03 -8.95919979e-01 3.53981256e-01
-2.12077305e-01 -1.99651390e-01 5.89309335e-01 8.38775635e-01
-6.61923945e-01 7.72018373e-01 -8.04227173e-01 -3.25614929e-01
4.64092255e-01 1.89339861e-01 -7.73418322e-02 -9.93881002e-03
-1.09464431e+00 8.54424238e-01 1.95414305e-01 -4.65743929e-01
-6.98019683e-01 -8.01876366e-01 -5.61943412e-01 5.61688423e-01
7.12837219e-01 -1.62653685e-01 1.30643272e+00 -4.85487610e-01
-1.49382651e+00 1.62061248e-02 5.35219833e-02 -4.15563941e-01
6.16170228e-01 -2.19520971e-01 -3.53009939e-01 -2.06534937e-01
-1.21418415e-02 -2.15356462e-02 9.78950262e-01 -6.35803401e-01
-8.62852931e-01 -3.64635170e-01 2.32171267e-01 2.44512767e-01
-6.31456017e-01 -3.07484269e-02 5.03332429e-02 -5.37580788e-01
-5.19788861e-01 -7.80155838e-01 -2.15563536e-01 -2.62828395e-02
-2.43266255e-01 -2.44373664e-01 5.89542985e-01 -4.68422771e-01
1.61771107e+00 -2.25286794e+00 -1.43283755e-01 7.10266590e-01
1.71562657e-01 5.74265569e-02 -7.68557191e-02 2.12157503e-01
2.68117577e-01 2.24557817e-01 -9.14066583e-02 1.59096316e-01
4.57778841e-01 1.06975943e-01 -6.62586093e-01 5.90400219e-01
-1.02815211e-01 1.94421813e-01 -9.27664638e-01 -3.11821133e-01
2.71794558e-01 1.47894830e-01 -9.55010056e-01 2.98787057e-01
-3.18367898e-01 5.69674522e-02 -5.13806403e-01 1.39036372e-01
4.19540018e-01 -1.00699864e-01 5.46437144e-01 -1.90765560e-02
-1.02570988e-01 5.32550454e-01 -1.57996929e+00 9.04904783e-01
-5.84953547e-01 3.28286439e-01 -1.41946092e-01 -1.11893976e+00
4.09169137e-01 2.23703623e-01 7.89672375e-01 -4.58222479e-01
-6.24484988e-03 2.50974949e-02 -4.36948910e-02 -2.00900301e-01
5.84735215e-01 -2.04838604e-01 5.30579016e-02 9.00967360e-01
-2.49078631e-01 3.16887230e-01 4.28820223e-01 3.19400251e-01
1.45138240e+00 -7.73878172e-02 3.22493553e-01 -2.83298135e-01
1.41568020e-01 -2.98374146e-01 6.72688961e-01 1.00259244e+00
-3.96212637e-01 -4.47482578e-02 1.16845489e+00 1.05820298e-01
-9.43987012e-01 -1.19312537e+00 -6.85726553e-02 1.27751434e+00
-2.66352803e-01 -4.88368183e-01 -5.19569218e-01 -9.04201925e-01
4.53370064e-01 9.15131390e-01 -5.57532609e-01 -2.79080987e-01
-1.96498364e-01 -5.62537253e-01 1.73608035e-01 7.09242702e-01
2.98988875e-02 -7.29893684e-01 -5.69115162e-01 3.95320326e-01
9.91466269e-03 -9.68980014e-01 -8.57742667e-01 -1.48334771e-01
-7.26767182e-01 -1.34941947e+00 -5.10072589e-01 1.11257434e-01
6.09564483e-01 1.35098293e-01 9.13442731e-01 -3.97813648e-01
1.15768062e-02 6.53111696e-01 -5.48096858e-02 -2.39525884e-01
-3.66157621e-01 -2.03429639e-01 4.25885588e-01 9.24409926e-02
6.55607581e-01 -5.45414627e-01 -5.83282769e-01 2.83481061e-01
-8.79937589e-01 -6.70079887e-01 3.34592015e-01 7.55276203e-01
2.82612085e-01 3.00005645e-01 6.02772415e-01 -8.74374092e-01
1.15547812e+00 -6.86335385e-01 -1.08621907e+00 4.04772431e-01
-9.55848157e-01 2.63654888e-01 7.37316549e-01 -7.06208825e-01
-8.58941734e-01 -5.22499681e-01 4.06674325e-01 -3.06984752e-01
1.10631712e-01 3.97345304e-01 -1.24631919e-01 3.76197040e-01
6.79967940e-01 -2.47005075e-01 1.53056875e-01 -4.58269864e-01
4.96266454e-01 7.81198204e-01 1.18207470e-01 -8.56516182e-01
4.07945871e-01 2.77509779e-01 -1.54078111e-01 -6.57450259e-01
-7.29413211e-01 -5.32227933e-01 9.96194780e-02 -1.00260302e-01
5.73230267e-01 -3.00398171e-01 -1.14770985e+00 -1.27823457e-01
-4.81875330e-01 -3.28512609e-01 -7.21897542e-01 9.36962485e-01
-7.49494374e-01 2.91057795e-01 -2.90339738e-01 -1.19101787e+00
-2.34422056e-04 -1.13824391e+00 4.21248823e-01 1.50428966e-01
-3.40668947e-01 -9.39074576e-01 2.07057863e-01 -1.98094025e-01
5.20308077e-01 -2.85453796e-01 1.17656386e+00 -7.93361306e-01
-2.53969848e-01 -3.31353009e-01 -2.84236610e-01 5.24249911e-01
1.39454812e-01 1.33270606e-01 -5.82686245e-01 -5.27849674e-01
-9.85326022e-02 -1.36661291e-01 5.11381567e-01 5.31229854e-01
1.37082267e+00 -5.84267914e-01 -9.56125483e-02 -1.65719185e-02
1.13764286e+00 2.17082381e-01 3.41859192e-01 4.76515032e-02
1.47139505e-01 4.27135050e-01 7.34505236e-01 8.91441345e-01
-2.06283852e-02 6.10844314e-01 -1.09117121e-01 5.37964821e-01
3.09144258e-01 -4.87060189e-01 6.97542131e-01 2.91054308e-01
1.77054748e-01 -3.33073765e-01 -5.92022300e-01 3.61483186e-01
-1.89065361e+00 -1.15588903e+00 5.92086852e-01 2.99075294e+00
7.94477701e-01 3.29584450e-01 5.74033082e-01 9.68981981e-02
6.18067503e-01 1.98585093e-01 -7.65524864e-01 -3.48123461e-01
4.87216443e-01 2.83540159e-01 8.59428465e-01 4.60306257e-01
-7.98979223e-01 3.95914108e-01 6.72365522e+00 7.95630276e-01
-8.05936038e-01 3.48558426e-02 3.98397952e-01 -3.27575505e-01
-3.83008540e-01 3.62311691e-01 -6.35970771e-01 8.01365256e-01
1.21169615e+00 -5.96490383e-01 7.52356172e-01 8.67269218e-01
2.11945564e-01 -4.29217786e-01 -1.24266565e+00 7.80707359e-01
-3.90452057e-01 -9.47093308e-01 -2.03420445e-01 6.42506659e-01
2.68932045e-01 -1.97188392e-01 2.47896269e-01 5.25946498e-01
8.64371717e-01 -6.51342094e-01 3.35514247e-01 3.78921688e-01
7.14500666e-01 -8.85182381e-01 3.51693094e-01 2.79048890e-01
-8.59503984e-01 -3.55917811e-01 -5.24500430e-01 1.30710214e-01
1.88084945e-01 7.80967236e-01 -7.23142087e-01 -9.46196467e-02
3.96085054e-01 5.15508056e-01 -1.65464543e-03 1.05374491e+00
-2.33116541e-02 9.31334853e-01 -5.52664936e-01 -1.28669798e-01
-4.88447137e-02 -5.00487030e-01 6.24543905e-01 7.58793473e-01
5.84442616e-01 -2.93183982e-01 2.12651208e-01 7.03327000e-01
-1.94341615e-01 1.76439419e-01 -6.42414033e-01 -5.31960607e-01
6.47883773e-01 9.99386728e-01 -4.04050738e-01 -4.17493105e-01
-6.22700989e-01 3.79095435e-01 3.17642897e-01 5.62410653e-01
-7.66806245e-01 -2.02709243e-01 1.12472570e+00 2.07410902e-01
1.09663434e-01 -1.97514653e-01 8.26376677e-02 -1.21408379e+00
-1.17768720e-01 -1.01846373e+00 6.97638869e-01 -1.14466865e-02
-1.44114673e+00 -1.18418671e-01 3.23236704e-01 -1.16664577e+00
-3.34011078e-01 -5.50274909e-01 -4.15396810e-01 7.08462775e-01
-1.13325286e+00 -1.46166041e-01 3.38164270e-01 4.13508445e-01
1.67266384e-01 -2.22251579e-01 7.56493092e-01 6.01136029e-01
-7.23985612e-01 6.48971736e-01 4.16877836e-01 -3.16585630e-01
7.65822589e-01 -1.33071315e+00 -3.42106782e-02 7.50334382e-01
1.48848489e-01 6.91739082e-01 6.97077572e-01 -5.41258156e-01
-1.32312703e+00 -8.46960068e-01 4.58763808e-01 -3.55045289e-01
9.03028190e-01 -2.55155236e-01 -8.05713296e-01 6.52468383e-01
-3.40743482e-01 1.66476205e-01 9.12261963e-01 4.96844321e-01
-4.65114027e-01 -5.63531630e-02 -1.29056215e+00 7.60644555e-01
7.91153669e-01 -7.13932693e-01 -2.29540244e-01 2.87209332e-01
2.57634491e-01 2.78400719e-01 -9.30005789e-01 8.22002739e-02
8.24838042e-01 -7.31742799e-01 6.91745937e-01 -1.14492893e+00
3.52394709e-04 -1.65175676e-01 -4.64853227e-01 -1.31381476e+00
-2.39259586e-01 -6.75929546e-01 -5.53676188e-01 1.12929893e+00
3.63612711e-01 -9.18146789e-01 6.57244682e-01 1.11028087e+00
7.18237817e-01 -3.68369579e-01 -9.54216182e-01 -1.01868641e+00
-4.56376523e-02 -6.44084692e-01 8.06393564e-01 7.81829238e-01
2.74927080e-01 1.22594468e-01 -4.37608957e-01 -2.11421520e-01
7.56379664e-01 -8.45694095e-02 6.80839598e-01 -1.09344423e+00
-5.26976228e-01 -5.72622597e-01 -3.21269363e-01 -9.75141585e-01
3.45097005e-01 -7.50728965e-01 -1.41654015e-01 -1.09655535e+00
3.21311206e-01 -6.74832523e-01 -7.61684537e-01 2.83017248e-01
-7.69302025e-02 -3.41368884e-01 3.92019786e-02 -3.20240147e-02
-3.87089729e-01 6.43298268e-01 6.62279606e-01 1.41390264e-02
-2.68951297e-01 3.70226651e-01 -7.67209530e-01 6.09839737e-01
7.54853666e-01 -6.15046263e-01 -6.90590620e-01 1.57142535e-01
2.03893870e-01 5.57144731e-02 1.81221202e-01 -6.89409494e-01
6.58266246e-03 -6.34084225e-01 -3.54221999e-03 -2.66532898e-01
4.71690949e-03 -9.06958044e-01 4.18163128e-02 1.57469824e-01
-7.67082453e-01 -1.22044638e-01 -1.94348052e-01 9.76559341e-01
6.36401549e-02 -2.81689167e-01 7.87800729e-01 1.39528915e-01
-2.12867558e-01 4.42471445e-01 -4.48776156e-01 3.81597579e-01
9.77463841e-01 1.97330937e-02 -2.17595071e-01 -5.65089047e-01
-4.24997747e-01 1.87981814e-01 3.87048125e-01 3.05290699e-01
3.70361447e-01 -1.42311895e+00 -3.50480080e-01 1.21190012e-01
6.25674501e-02 -5.81842661e-01 5.19798696e-02 8.18835855e-01
2.91425642e-03 1.38262138e-01 -4.74739708e-02 -7.48122707e-02
-9.09949183e-01 5.02314568e-01 1.04243055e-01 -3.70215207e-01
-3.83590668e-01 4.87375587e-01 5.50592318e-02 -2.22982526e-01
2.98298180e-01 -1.66035444e-01 -2.45508060e-01 1.31548688e-01
8.69517684e-01 7.51853228e-01 -2.34794691e-01 -1.49713919e-01
-1.99317008e-01 -8.90764967e-02 -2.12903053e-01 -6.81099415e-01
1.07002175e+00 2.13003412e-01 6.21511303e-02 3.25188160e-01
1.09805989e+00 2.24313319e-01 -1.28501797e+00 -4.76342350e-01
1.71691433e-01 -1.14276946e+00 2.13240400e-01 -4.24148381e-01
-8.55210841e-01 6.26396656e-01 6.57348096e-01 6.28650546e-01
8.22267652e-01 -4.00074244e-01 6.41066551e-01 2.06563696e-01
1.23881601e-01 -1.50148833e+00 -1.68615535e-01 2.93811977e-01
4.18068022e-01 -9.10530806e-01 1.61093473e-01 4.29160558e-02
-5.43179512e-01 8.09691608e-01 4.67232943e-01 1.92368086e-02
8.43178809e-01 2.61698663e-01 -4.37733889e-01 7.61459321e-02
-7.83246100e-01 -7.78843695e-03 1.56315088e-01 3.52605611e-01
3.31982613e-01 3.82298082e-01 -4.25265580e-01 2.85974383e-01
-4.64175967e-03 1.60331614e-02 5.81384718e-01 1.01240897e+00
-2.44425461e-01 -1.38441885e+00 -1.44789055e-01 1.23384655e+00
-7.22694993e-01 2.90173031e-02 1.20915473e-01 2.43582487e-01
-5.96411526e-01 9.14093018e-01 3.94012332e-01 -2.37124145e-01
5.08911848e-01 1.70688361e-01 3.81702930e-01 -5.11881769e-01
-8.21685195e-02 -1.48516610e-01 1.07992172e-01 -7.17830479e-01
-9.92751494e-02 -7.93119013e-01 -7.70808578e-01 -5.96532464e-01
-4.56849724e-01 3.74355733e-01 7.35667229e-01 7.91584969e-01
3.24465305e-01 3.59510481e-01 7.73004234e-01 -4.24635530e-01
-1.38316715e+00 -7.54949868e-01 -9.42920804e-01 4.70298380e-01
1.34675235e-01 -1.01297247e+00 -6.10078752e-01 -4.21530575e-01] | [4.121892929077148, 2.525270938873291] |
601d6c50-d783-4cf2-baea-1ba151501df3 | improving-retinanet-for-ct-lesion-detection | 1906.02283 | null | https://arxiv.org/abs/1906.02283v1 | https://arxiv.org/pdf/1906.02283v1.pdf | Improving RetinaNet for CT Lesion Detection with Dense Masks from Weak RECIST Labels | Accurate, automated lesion detection in Computed Tomography (CT) is an important yet challenging task due to the large variation of lesion types, sizes, locations and appearances. Recent work on CT lesion detection employs two-stage region proposal based methods trained with centroid or bounding-box annotations. We propose a highly accurate and efficient one-stage lesion detector, by re-designing a RetinaNet to meet the particular challenges in medical imaging. Specifically, we optimize the anchor configurations using a differential evolution search algorithm. For training, we leverage the response evaluation criteria in solid tumors (RECIST) annotation which are measured in clinical routine. We incorporate dense masks from weak RECIST labels, obtained automatically using GrabCut, into the training objective, which in combination with other advancements yields new state-of-the-art performance. We evaluate our method on the public DeepLesion benchmark, consisting of 32,735 lesions across the body. Our one-stage detector achieves a sensitivity of 90.77% at 4 false positives per image, significantly outperforming the best reported methods by over 5%. | ['Qi Dou', 'Martin Zlocha', 'Ben Glocker'] | 2019-06-05 | null | null | null | null | ['medical-object-detection', 'skin-lesion-identification'] | ['computer-vision', 'medical'] | [ 3.79217148e-01 -2.91133523e-02 -5.31515598e-01 -1.83287978e-01
-1.35068882e+00 -5.65662503e-01 3.99242759e-01 3.94199252e-01
-7.75099635e-01 4.25502121e-01 2.48873755e-01 -2.84570843e-01
1.99288040e-01 -2.69596636e-01 -3.54448438e-01 -8.72341514e-01
-1.26271218e-01 7.21272528e-01 5.41438639e-01 3.00948828e-01
-5.89850694e-02 6.24675035e-01 -7.25320458e-01 3.61138552e-01
7.26350248e-01 1.12813425e+00 2.10039482e-01 8.69654834e-01
3.51072758e-01 6.42708838e-01 -2.11788461e-01 -4.38079387e-01
1.65458947e-01 -2.37930685e-01 -7.44965792e-01 7.31455982e-02
4.63177055e-01 -3.86361867e-01 -4.30843234e-01 9.88027632e-01
7.87030220e-01 -2.85991997e-01 1.09567249e+00 -5.20233035e-01
-2.65579224e-01 4.43150222e-01 -1.05791259e+00 7.25377321e-01
-2.29966775e-01 4.36509281e-01 1.00787342e+00 -8.42803299e-01
6.11437500e-01 6.71231508e-01 1.02243531e+00 7.18020737e-01
-1.24386919e+00 -4.53730047e-01 -1.52338400e-01 -9.80196595e-02
-1.54714906e+00 -2.85932630e-01 2.47157365e-02 -6.54720008e-01
8.95305216e-01 4.34560686e-01 6.01572514e-01 9.82201338e-01
4.04404372e-01 7.55453289e-01 7.25683272e-01 -3.41868192e-01
4.31449749e-02 7.20093250e-02 -4.94786687e-02 1.12876618e+00
4.78681594e-01 -4.44435095e-03 1.10668808e-01 -4.01402384e-01
9.84433234e-01 9.01996866e-02 -4.14285630e-01 -5.84356546e-01
-1.18917549e+00 8.82153213e-01 9.25380349e-01 2.33034506e-01
-4.54236478e-01 4.14039254e-01 4.57939535e-01 -5.56092381e-01
4.40710187e-01 3.30245852e-01 3.79826389e-02 3.20747375e-01
-1.00703776e+00 -1.88412860e-01 3.67353559e-01 3.69243473e-01
-2.78398115e-02 -4.01398808e-01 -8.59482586e-01 8.52556467e-01
2.76179999e-01 2.02692717e-01 7.17684805e-01 -2.36423254e-01
3.37547272e-01 7.33010054e-01 -2.20442384e-01 -2.66262323e-01
-9.89654064e-01 -1.01864672e+00 -7.78256476e-01 1.18635952e-01
4.73870188e-01 -1.77429423e-01 -1.69186544e+00 1.36846030e+00
4.70043927e-01 -7.06454273e-03 -3.93911630e-01 1.04892683e+00
9.55902100e-01 -1.58167943e-01 4.38593864e-01 -7.78510943e-02
1.80655539e+00 -1.11980879e+00 -2.43580282e-01 -1.60295725e-01
9.91401434e-01 -5.64061046e-01 8.01273167e-01 2.08127692e-01
-9.41628277e-01 1.24688700e-01 -1.01546192e+00 8.95103067e-02
6.14311993e-02 7.55121768e-01 6.20787680e-01 1.02222097e+00
-9.21063006e-01 2.93666989e-01 -1.24589336e+00 -4.58924323e-01
1.09705520e+00 5.65805733e-01 -9.77062508e-02 -3.65851410e-02
-4.24746722e-01 1.06747591e+00 2.13664427e-01 -2.27825165e-01
-1.21915400e+00 -1.18391073e+00 -5.06442904e-01 -2.05761656e-01
6.81788921e-01 -1.04762948e+00 1.55752075e+00 -5.33585846e-01
-1.18818867e+00 1.20456612e+00 4.18527499e-02 -5.83157957e-01
9.16769087e-01 1.78841159e-01 -9.29107144e-02 4.87343192e-01
1.92849517e-01 7.45172203e-01 5.49100339e-01 -8.38355899e-01
-6.18183434e-01 -3.73665392e-01 -4.89910901e-01 3.06367576e-01
-1.36779800e-01 1.07644312e-01 -8.19667339e-01 -6.15281701e-01
1.50902912e-01 -1.02846193e+00 -7.34889090e-01 4.22710538e-01
-6.93090856e-01 -5.34543246e-02 3.78469378e-01 -6.83954239e-01
1.25455260e+00 -1.70628262e+00 -1.53934062e-01 4.58768874e-01
7.45640814e-01 1.14433013e-01 3.31180185e-01 -3.85946691e-01
-8.38062540e-02 1.31872103e-01 -4.40804183e-01 -3.14611405e-01
-1.97202757e-01 -1.96245939e-01 2.98270911e-01 7.87047446e-01
1.79735154e-01 1.37975395e+00 -7.30893493e-01 -7.48169243e-01
1.47406787e-01 3.03048700e-01 -5.31407177e-01 -8.44800919e-02
2.26155501e-02 4.74935502e-01 -4.54430163e-01 9.17172849e-01
4.33709919e-01 -7.98715830e-01 1.50136754e-01 -3.67052704e-01
8.13009515e-02 4.92449664e-02 -7.02948332e-01 1.93038189e+00
-3.52793574e-01 4.56792235e-01 2.35109292e-02 -5.75900257e-01
2.88237035e-01 4.36431736e-01 1.01201165e+00 -3.19136620e-01
4.57448334e-01 2.68171459e-01 3.32161963e-01 -6.23464108e-01
-2.99298782e-02 -2.40584716e-01 1.35188997e-01 4.03984457e-01
-7.56027326e-02 -2.13170528e-01 1.16020992e-01 1.26274019e-01
1.76867235e+00 -3.15797508e-01 6.09653533e-01 -1.52397126e-01
2.55298704e-01 2.82901168e-01 3.51723820e-01 9.20565665e-01
-4.60188210e-01 1.09572983e+00 4.55843776e-01 -3.39317381e-01
-8.67794275e-01 -1.06652474e+00 -5.23085177e-01 8.89818192e-01
-1.93437770e-01 -1.84038520e-01 -6.46987677e-01 -1.25840390e+00
-3.26777026e-02 3.26474607e-01 -9.87127423e-01 -1.04706436e-01
-6.34839058e-01 -1.45843983e+00 7.46206582e-01 6.78609490e-01
2.82950282e-01 -6.90086305e-01 -9.39198792e-01 3.42580169e-01
6.53508678e-02 -1.10648179e+00 -6.93625689e-01 4.71073180e-01
-7.76675999e-01 -1.18411851e+00 -1.33726692e+00 -6.01919532e-01
8.32923949e-01 -2.79825367e-02 1.16374600e+00 3.21626484e-01
-1.19435215e+00 2.61925131e-01 -9.25406627e-03 -2.51723498e-01
-2.78107971e-01 2.16291040e-01 -4.12765473e-01 -2.00070694e-01
1.92785546e-01 2.24689525e-02 -1.05678856e+00 3.97142589e-01
-7.19379187e-01 3.42321247e-02 1.02653408e+00 8.64351451e-01
6.83613539e-01 -3.89160424e-01 2.96564311e-01 -9.59889650e-01
3.00983936e-01 -4.34768677e-01 -4.05167073e-01 4.50816125e-01
-3.31950575e-01 -8.15948993e-02 3.35807279e-02 -3.70021045e-01
-9.10034597e-01 5.56509733e-01 -7.04660863e-02 -2.39055187e-01
-7.13715628e-02 2.63755500e-01 4.96502817e-01 -5.85683465e-01
1.22634625e+00 -1.12251766e-01 -1.87548667e-01 -1.81708574e-01
2.10336939e-01 4.52462614e-01 6.45382047e-01 -2.16668680e-01
4.95243639e-01 8.38654280e-01 1.38431281e-01 -5.26060760e-01
-9.86248195e-01 -9.13701475e-01 -7.17165232e-01 -1.62426278e-01
9.63675559e-01 -6.98677540e-01 -5.40344179e-01 1.16273314e-01
-9.08492386e-01 -3.10810179e-01 -3.62406701e-01 6.71458006e-01
-3.60410661e-01 2.33858705e-01 -7.48269975e-01 -5.27582943e-01
-7.78249681e-01 -1.45266128e+00 1.41078222e+00 1.76527113e-01
-1.49866253e-01 -7.98608482e-01 1.14981733e-01 1.20157085e-01
5.32506526e-01 4.02464360e-01 8.43224108e-01 -6.34697199e-01
-4.94177490e-01 -1.94442272e-01 -6.59726024e-01 -9.67204794e-02
1.29442558e-01 -7.24511370e-02 -7.37981856e-01 -3.73602122e-01
-3.96343112e-01 -3.19775850e-01 1.40232790e+00 9.55443382e-01
1.29396939e+00 3.86843115e-01 -9.81201947e-01 9.35025573e-01
1.49877656e+00 -1.25542562e-02 2.21519291e-01 2.36837983e-01
6.65743530e-01 8.39213803e-02 1.93068236e-01 5.23919642e-01
2.84259822e-02 7.33404160e-01 6.00183070e-01 -4.46053922e-01
-6.45308554e-01 1.55379251e-01 -2.37957895e-01 -3.83570418e-02
-1.48142666e-01 -2.06351534e-01 -1.38498688e+00 6.07564867e-01
-1.58157682e+00 -4.68541920e-01 -1.46511689e-01 1.91762042e+00
8.87914121e-01 1.51082247e-01 2.35122859e-01 -3.84472758e-01
8.43623042e-01 -2.38675952e-01 -8.58579993e-01 3.34288239e-01
2.27546379e-01 4.26636934e-01 1.13864458e+00 1.07583642e-01
-1.58135962e+00 7.15624511e-01 6.49899340e+00 8.83227229e-01
-1.26479113e+00 4.17338520e-01 7.93265581e-01 -4.65074778e-01
2.54286975e-01 -4.10904050e-01 -8.25023651e-01 1.16671175e-01
3.79564822e-01 5.96894920e-02 -1.40677109e-01 5.71213245e-01
6.55054823e-02 -3.48405659e-01 -1.19510794e+00 9.85516489e-01
1.70600235e-01 -1.53196824e+00 -2.95057178e-01 4.55553047e-02
8.72680247e-01 5.43868840e-01 3.55798811e-01 1.20835312e-01
4.64824289e-01 -1.17354023e+00 4.57968026e-01 2.38275990e-01
1.16155350e+00 -4.13720876e-01 7.13689268e-01 1.17705263e-01
-1.09444284e+00 1.10343313e-02 -1.90558910e-01 8.36998045e-01
2.26898432e-01 5.13982773e-01 -1.54410529e+00 3.38170111e-01
5.28933048e-01 6.11466169e-01 -7.77688920e-01 1.89712572e+00
-5.01458272e-02 6.72237158e-01 -5.91185451e-01 4.87205349e-02
4.00561482e-01 4.18558747e-01 5.30750096e-01 1.48259783e+00
2.60654092e-01 1.99183613e-01 2.26440251e-01 8.18978250e-01
-2.69944668e-01 2.09023550e-01 -2.49799807e-02 3.34399849e-01
1.13192976e-01 1.63071930e+00 -1.18965101e+00 -3.63255411e-01
-1.42332762e-01 6.75059676e-01 2.50800759e-01 3.94061282e-02
-1.10806513e+00 1.10177860e-01 1.75550655e-01 2.07529560e-01
3.38089675e-01 2.07822368e-01 -3.53186131e-01 -8.92148674e-01
-1.47481665e-01 -5.92938662e-01 7.81072617e-01 -4.14039433e-01
-1.24124372e+00 6.25870228e-01 -1.60455927e-01 -1.16515040e+00
-1.96802020e-02 -7.20007479e-01 -6.57140493e-01 5.72817266e-01
-1.53522539e+00 -1.20050383e+00 -5.49702108e-01 2.86125720e-01
5.43248117e-01 4.78159748e-02 6.31662607e-01 4.07864928e-01
-8.20461869e-01 8.50093603e-01 -1.69515491e-01 2.38712996e-01
7.80733526e-01 -1.31273329e+00 2.79508024e-01 5.09741783e-01
-4.28577140e-02 6.65870830e-02 2.66183794e-01 -6.44382060e-01
-9.16730762e-01 -1.24095762e+00 1.54553100e-01 -5.24119973e-01
8.06447923e-01 -1.60804048e-01 -7.17277348e-01 6.37101591e-01
-1.70393467e-01 6.29532635e-01 5.32234073e-01 -3.80335361e-01
-7.84390494e-02 3.29495072e-01 -1.31601334e+00 5.37041724e-01
1.04623842e+00 -6.85018394e-03 -2.41229430e-01 8.09922814e-01
4.02053237e-01 -9.48861778e-01 -6.01596057e-01 7.16256917e-01
4.41358209e-01 -6.49461985e-01 1.11172843e+00 -6.15306437e-01
1.90041944e-01 -6.48141233e-03 2.47853667e-01 -1.10902870e+00
-5.33663332e-01 -3.22697133e-01 1.28761902e-01 3.41401249e-01
7.39258349e-01 -4.28316355e-01 1.31844270e+00 4.83391970e-01
-3.55173469e-01 -1.15363801e+00 -1.09970570e+00 -4.11330640e-01
5.20799533e-02 -3.10996830e-01 6.76174834e-02 5.21882713e-01
-2.56347060e-01 -6.50570095e-02 1.19649068e-01 1.63724840e-01
6.62028253e-01 -2.05395147e-01 2.11246505e-01 -9.28707123e-01
-3.98559034e-01 -8.93400013e-01 -4.80636239e-01 -1.02216923e+00
-4.26362276e-01 -1.21731973e+00 6.83999658e-02 -1.52137172e+00
7.84058034e-01 -6.18808389e-01 -4.26794022e-01 5.34896791e-01
-2.15047076e-01 7.50694394e-01 -1.61704496e-01 1.26751080e-01
-6.46995664e-01 1.89602062e-01 1.34707856e+00 -2.56001741e-01
7.00391037e-03 1.82139099e-01 -4.32486206e-01 1.06107163e+00
5.42090833e-01 -7.43083239e-01 1.20398887e-01 -3.06034952e-01
-2.26631090e-01 -4.83130524e-03 6.57633603e-01 -1.24104679e+00
3.77803028e-01 3.21449786e-02 6.44761026e-01 -7.44667232e-01
4.16908443e-01 -6.47694707e-01 -3.24268907e-01 8.67809057e-01
-3.67650092e-01 -8.94369557e-02 2.02659503e-01 6.29294336e-01
2.42281646e-01 -4.08231765e-01 1.12060928e+00 -3.00471783e-01
-4.33406323e-01 5.81515491e-01 -2.25154817e-01 1.09125510e-01
1.33961213e+00 -1.58639938e-01 -3.76067132e-01 1.27849415e-01
-9.55935538e-01 2.20850900e-01 1.54372513e-01 -4.18750308e-02
4.44208503e-01 -1.17356098e+00 -1.03407264e+00 -7.90431947e-02
2.70523876e-01 -4.23768051e-02 2.71295726e-01 1.50270844e+00
-6.82672203e-01 4.83131230e-01 5.08101396e-02 -1.08121121e+00
-1.37980223e+00 7.14129657e-02 9.20727015e-01 -7.24272966e-01
-8.45852077e-01 1.18123233e+00 4.45818394e-01 -6.20880201e-02
2.83069521e-01 -5.66297293e-01 -1.51685148e-01 -2.06732988e-01
2.67158002e-01 9.23897699e-02 5.47388852e-01 -3.88818949e-01
-5.12329757e-01 6.02630734e-01 -5.33950150e-01 -3.73587310e-02
1.13841200e+00 3.05401593e-01 3.72302234e-01 -2.70849615e-01
1.09525478e+00 -1.57576308e-01 -1.12912869e+00 -4.54007208e-01
1.00640342e-01 -3.17392707e-01 5.04169405e-01 -1.15656686e+00
-1.22714615e+00 5.70559442e-01 9.64208186e-01 -2.34001651e-01
8.93226802e-01 4.45287138e-01 6.83042645e-01 3.50215472e-02
-7.12724999e-02 -7.30341673e-01 2.22327068e-01 8.64711553e-02
7.19880223e-01 -1.54679716e+00 3.11248094e-01 -5.40305376e-01
-6.35231495e-01 9.78683233e-01 5.64740360e-01 -9.77865383e-02
4.23340470e-01 3.79245609e-01 9.05669108e-02 -5.38107455e-01
-7.51578271e-01 -4.13615465e-01 6.76859915e-01 3.58403176e-01
5.30754447e-01 2.58382499e-01 -2.06871137e-01 5.17334044e-01
1.61524698e-01 -2.71626204e-01 3.87729526e-01 9.37821627e-01
-6.59322500e-01 -8.31259966e-01 -3.51621002e-01 9.68495250e-01
-7.38581419e-01 -1.21145032e-01 -2.85768449e-01 9.80899215e-01
1.04506267e-04 3.71347606e-01 -1.02062888e-01 1.07744485e-01
3.08236808e-01 -3.41688395e-01 6.21565580e-01 -1.02507079e+00
-8.23666215e-01 2.61398494e-01 1.39028523e-02 -4.53466952e-01
1.27159916e-02 -7.88715363e-01 -1.43861818e+00 3.10051918e-01
-5.97711146e-01 -2.75248915e-01 7.70614505e-01 7.61325836e-01
-1.91529542e-02 7.91953146e-01 2.26589695e-01 -6.93674862e-01
-8.37693810e-01 -8.15044880e-01 -2.50591487e-01 2.26672798e-01
2.74023920e-01 -8.88823390e-01 -1.77996010e-01 -2.30763435e-01] | [15.05505084991455, -2.3024353981018066] |
2a54610c-9cc1-46f7-b73b-1f85d6b70656 | harrisz-harris-corner-selection-for-next-gen | 2109.12925 | null | https://arxiv.org/abs/2109.12925v6 | https://arxiv.org/pdf/2109.12925v6.pdf | HarrisZ$^+$: Harris Corner Selection for Next-Gen Image Matching Pipelines | Due to its role in many computer vision tasks, image matching has been subjected to an active investigation by researchers, which has lead to better and more discriminant feature descriptors and to more robust matching strategies, also thanks to the advent of the deep learning and the increased computational power of the modern hardware. Despite of these achievements, the keypoint extraction process at the base of the image matching pipeline has not seen equivalent progresses. This paper presents HarrisZ$^+$, an upgrade to the HarrisZ corner detector, optimized to synergically take advance of the recent improvements of the other steps of the image matching pipeline. HarrisZ$^+$ does not only consists of a tuning of the setup parameters, but introduces further refinements to the selection criteria delineated by HarrisZ, so providing more, yet discriminative, keypoints, which are better distributed on the image and with higher localization accuracy. The image matching pipeline including HarrisZ$^+$, together with the other modern components, obtained in different recent matching benchmarks state-of-the-art results among the classic image matching pipelines. These results are quite close to those obtained by the more recent fully deep end-to-end trainable approaches and show that there is still a proper margin of improvement that can be granted by the research in classic image matching methods. | ['Dmytro Mishkin', 'Fabio Bellavia'] | 2021-09-27 | null | null | null | null | ['image-matching'] | ['computer-vision'] | [-1.03435919e-01 -2.61931479e-01 -7.03183189e-02 -3.61542672e-01
-8.13767731e-01 -2.27838814e-01 7.35963464e-01 4.37903225e-01
-6.02070987e-01 1.78534552e-01 -1.89586002e-02 4.01317552e-02
-3.44931692e-01 -6.30153656e-01 -5.44375658e-01 -6.98977113e-01
-2.33195037e-01 7.21818268e-01 5.43076694e-01 -5.63099921e-01
4.99801397e-01 9.47038591e-01 -2.03399062e+00 7.82253444e-02
2.12144479e-01 1.18773735e+00 1.48535103e-01 5.64899862e-01
7.02841729e-02 2.82421827e-01 -1.18947625e-01 -4.88908798e-01
5.02492249e-01 -8.95382762e-02 -6.09423637e-01 -1.44504845e-01
1.07193446e+00 -2.17757300e-01 -4.65913057e-01 9.49989259e-01
7.90551066e-01 -5.43962009e-02 3.17442417e-01 -1.21592772e+00
-2.36903012e-01 2.54402906e-01 -7.50591278e-01 7.23328367e-02
5.76895297e-01 -2.51136906e-02 1.11322308e+00 -8.82968783e-01
9.12204385e-01 1.04271865e+00 1.04794681e+00 1.53474569e-01
-9.55522776e-01 -6.47279084e-01 -3.96146148e-01 4.85675603e-01
-1.40436709e+00 -3.44037056e-01 5.91828883e-01 -2.92756796e-01
1.31672585e+00 9.43707377e-02 7.09897876e-01 5.50563633e-01
2.72341162e-01 7.73187995e-01 9.33523059e-01 -6.77585781e-01
-9.28909555e-02 -2.62969453e-02 1.48578912e-01 8.79079461e-01
2.02782974e-01 5.59863925e-01 -2.93661475e-01 7.35686198e-02
6.35697722e-01 2.07669139e-01 -6.12387545e-02 -8.90759230e-01
-1.24951506e+00 7.44882226e-01 8.06168079e-01 6.62245691e-01
-4.32339668e-01 2.91163236e-01 4.98943239e-01 3.86633337e-01
8.99669379e-02 3.91653150e-01 -2.46110678e-01 -7.61755407e-02
-1.18656456e+00 5.27673006e-01 5.38098335e-01 1.04749906e+00
1.18791950e+00 -3.26689631e-01 -9.60737839e-02 5.17311990e-01
2.32441843e-01 4.94572461e-01 4.39661205e-01 -7.29903162e-01
2.56060094e-01 9.97600019e-01 -6.90047219e-02 -1.28255105e+00
-6.20254159e-01 -4.46624547e-01 -8.03425014e-01 6.19254112e-01
5.68300486e-01 4.01405722e-01 -8.73492241e-01 1.34224069e+00
2.05520719e-01 -1.22289933e-01 -2.92728513e-01 7.69381583e-01
6.56273186e-01 1.67810708e-01 -2.01645404e-01 5.90130925e-01
1.55007589e+00 -7.74853468e-01 -2.56385863e-01 -2.52012145e-02
6.60170794e-01 -1.34844100e+00 7.74665296e-01 2.06773564e-01
-8.16400886e-01 -8.31281662e-01 -1.38632226e+00 -1.74469128e-01
-8.20079267e-01 -7.98584372e-02 9.37742889e-01 6.91708088e-01
-1.48445618e+00 7.41968930e-01 -5.88255286e-01 -7.25411654e-01
3.78771961e-01 7.53481567e-01 -9.20265317e-01 -1.90748960e-01
-9.45460618e-01 1.18275976e+00 1.69830918e-01 5.91571406e-02
-2.34914660e-01 -7.32474566e-01 -1.01972401e+00 8.19407031e-02
1.29849613e-01 -7.73577988e-01 1.10164106e+00 -7.17869222e-01
-1.12852454e+00 1.55208421e+00 -6.87615052e-02 -5.50760508e-01
8.14080119e-01 -8.29504877e-02 9.86830071e-02 1.51853144e-01
1.36112675e-01 1.10550046e+00 6.12420797e-01 -8.08352590e-01
-1.02123499e+00 -3.91799361e-01 -2.00604826e-01 -1.10679835e-01
1.98043510e-01 1.08169459e-01 -5.88458538e-01 -3.98263872e-01
2.00753450e-01 -1.07380438e+00 -4.04034883e-01 2.36943901e-01
-2.30126455e-02 -5.09604633e-01 7.29943633e-01 -2.50461191e-01
6.69670045e-01 -2.10728192e+00 -1.34634441e-02 3.40108842e-01
2.90142447e-01 3.65248144e-01 -3.43125500e-02 8.00862432e-01
-4.00424957e-01 -3.71571422e-01 -1.19238149e-03 -5.90662837e-01
2.74265975e-01 -7.98155293e-02 -7.04609081e-02 8.48502815e-01
8.01117122e-02 9.10150230e-01 -7.31309414e-01 -4.57335562e-01
7.19293952e-01 3.62096727e-01 -3.00619721e-01 4.65925150e-02
3.61959368e-01 -1.42987192e-01 -1.88478485e-01 7.77422547e-01
8.86342645e-01 7.44961277e-02 -5.45754254e-01 -3.21653366e-01
-4.60533470e-01 -2.20234483e-01 -1.34629464e+00 1.99593914e+00
-1.86645731e-01 9.38675642e-01 -7.46247172e-02 -1.20850646e+00
1.12361646e+00 1.20074466e-01 8.00395429e-01 -1.02405953e+00
4.31674123e-01 6.53409362e-01 -3.43035348e-02 -2.79959768e-01
5.09181798e-01 9.04366747e-02 -2.39842981e-02 1.08023308e-01
2.11963147e-01 -2.76211202e-01 2.47101977e-01 3.82563509e-02
8.74832213e-01 2.43649021e-01 4.96500105e-01 -5.14562249e-01
7.10761189e-01 1.77208826e-01 5.12348525e-02 8.61598909e-01
-3.73097003e-01 9.35727775e-01 -2.77600400e-02 -7.80808270e-01
-1.21480453e+00 -6.55777156e-01 -3.60372216e-01 6.45497859e-01
2.76954532e-01 -3.12487215e-01 -6.82087362e-01 -2.36370593e-01
3.84153903e-01 -2.38783196e-01 -7.62388408e-01 -9.07313824e-02
-8.38261902e-01 -4.44188297e-01 6.10462725e-01 5.08578658e-01
7.32230842e-01 -1.09240973e+00 -1.10766578e+00 2.16038674e-01
4.34929162e-01 -1.00089478e+00 -1.98486418e-01 4.11438912e-01
-6.47817612e-01 -1.31245887e+00 -9.96050358e-01 -1.02116501e+00
3.33193243e-01 3.81713480e-01 1.19261289e+00 3.63444209e-01
-8.56231153e-01 4.18092966e-01 -4.27339852e-01 -4.16912109e-01
-9.05856267e-02 3.85444701e-01 -2.74887890e-01 -1.72951609e-01
6.37007058e-01 -4.73947495e-01 -7.83219397e-01 3.38927120e-01
-8.53707194e-01 -3.30549419e-01 8.40395689e-01 8.16089153e-01
3.45035404e-01 -3.44792336e-01 1.22387491e-01 -5.12154460e-01
1.70027055e-02 1.23937532e-01 -8.48015010e-01 1.02759786e-01
-7.85486877e-01 2.30754599e-01 2.92907983e-01 -2.05321163e-02
-4.14202511e-01 4.06857669e-01 -6.63819969e-01 -2.19371021e-01
-3.26556683e-01 5.24484739e-02 -4.93159220e-02 -8.02512228e-01
5.51077604e-01 -8.00946429e-02 1.74501032e-01 -4.23368990e-01
5.03012061e-01 4.86417681e-01 6.98242307e-01 -3.16246718e-01
1.02003372e+00 6.49663508e-01 3.28919321e-01 -6.54399812e-01
-1.70229003e-01 -1.00738573e+00 -8.98047447e-01 2.69093993e-03
8.01846623e-01 -6.47712469e-01 -8.93189967e-01 8.26189756e-01
-1.07399535e+00 9.19533074e-02 -3.61511886e-01 4.01838034e-01
-6.13978922e-01 4.38755572e-01 -4.44865227e-01 -3.15309554e-01
-4.32335287e-01 -1.47316563e+00 1.42536068e+00 3.80470872e-01
-1.25262260e-01 -8.96258414e-01 2.57522374e-01 1.98598448e-02
6.86749399e-01 3.77641708e-01 6.75538719e-01 -6.24324858e-01
-7.95886695e-01 -9.35573161e-01 -4.20264482e-01 4.54946011e-02
-2.24094793e-01 1.63868710e-01 -1.02324069e+00 -3.83272976e-01
-3.08295906e-01 -6.23337887e-02 1.02257311e+00 3.05281967e-01
4.40559745e-01 3.87400866e-01 -5.33886909e-01 9.44999993e-01
1.73695552e+00 2.55245954e-01 9.31450844e-01 1.04081666e+00
3.37488949e-01 6.59537435e-01 5.77451885e-01 1.32188991e-01
2.40626574e-01 1.16077244e+00 5.84905386e-01 -6.84011221e-01
-3.86117220e-01 -8.66134018e-02 -1.28768340e-01 1.33998528e-01
-4.73349281e-02 4.23451185e-01 -9.45488930e-01 4.24144655e-01
-1.82035661e+00 -9.45863068e-01 -2.31213525e-01 2.19260144e+00
2.27832839e-01 1.02658749e-01 1.26633033e-01 4.03130263e-01
5.13895154e-01 1.06413104e-01 -2.26217076e-01 -3.03765059e-01
-3.39096516e-01 7.21350014e-01 7.69312024e-01 4.34028804e-01
-1.31653154e+00 8.97887409e-01 6.67448997e+00 7.99752712e-01
-1.31450164e+00 -3.15957248e-01 2.34043121e-01 5.31363189e-01
2.38215938e-01 7.75559843e-02 -1.01193881e+00 1.31558582e-01
3.34304303e-01 -7.04812333e-02 9.86157879e-02 1.03710496e+00
-2.65878201e-01 -3.08194906e-01 -1.31865418e+00 1.39248109e+00
1.56711698e-01 -1.54504192e+00 -1.90829337e-01 8.20888802e-02
4.40234542e-01 2.74230897e-01 8.34958032e-02 2.78410196e-01
-1.68882430e-01 -8.45457017e-01 7.24747479e-01 3.49613577e-01
3.75395715e-01 -6.80840313e-01 1.14979959e+00 2.18825452e-02
-1.40303135e+00 -4.01573814e-02 -5.80990314e-01 -5.77774970e-03
-8.60018730e-02 4.08623219e-01 -6.67593896e-01 7.42513239e-01
9.38920915e-01 7.67465055e-01 -9.67290759e-01 1.48414707e+00
5.17827012e-02 -2.33844951e-01 -3.98528814e-01 -9.18354765e-02
4.80189472e-01 -1.61765143e-01 3.48323345e-01 1.60890782e+00
3.17633659e-01 -3.76699001e-01 2.11472474e-02 5.69000363e-01
1.72688305e-01 2.96745300e-01 -7.95311868e-01 6.41278505e-01
2.69186378e-01 1.50046909e+00 -7.77521014e-01 -3.08948398e-01
-2.76406407e-01 8.51575792e-01 1.87069893e-01 -1.73148677e-01
-5.49294055e-01 -7.90845573e-01 7.77057469e-01 2.32795283e-01
2.36903772e-01 -1.98130652e-01 -1.16765134e-01 -8.63726735e-01
7.58877471e-02 -8.27540696e-01 3.28714162e-01 -6.25663102e-01
-1.04648602e+00 6.53806627e-01 3.75020429e-02 -1.22739518e+00
-4.77355152e-01 -8.98658574e-01 -5.60170829e-01 9.03563678e-01
-1.78513718e+00 -1.23063850e+00 -4.74684596e-01 6.94249451e-01
3.33153933e-01 -2.30636731e-01 8.45851362e-01 6.05840921e-01
-2.03052014e-01 8.33611071e-01 1.05182923e-01 2.52919585e-01
9.53063190e-01 -1.26128018e+00 5.95267475e-01 7.98730373e-01
2.84006536e-01 5.80447674e-01 7.77308047e-01 -1.01136249e-02
-1.44414639e+00 -3.61801177e-01 1.09711814e+00 -4.56107110e-01
5.51712513e-01 -3.65627646e-01 -6.90271676e-01 3.78923029e-01
3.94424736e-01 3.70944999e-02 2.27597028e-01 -3.68616767e-02
-5.45958817e-01 -5.04867733e-01 -1.14593768e+00 3.61275971e-01
9.25619364e-01 -5.34449637e-01 -7.31134117e-01 5.44560961e-02
2.33016849e-01 -4.69386786e-01 -6.07772052e-01 4.93473440e-01
7.51693487e-01 -1.49026489e+00 1.28395581e+00 -1.70151189e-01
-3.44982147e-02 -2.33975172e-01 -6.17012866e-02 -8.93698573e-01
-5.01371026e-01 -3.97729874e-01 6.37188196e-01 1.15627825e+00
1.36917084e-01 -6.90002799e-01 1.06452084e+00 3.89477253e-01
-3.32646668e-01 -6.04345679e-01 -1.26980066e+00 -6.83131456e-01
-2.34410893e-02 -1.63521588e-01 4.19117868e-01 7.17859507e-01
-1.21265128e-01 2.53873523e-02 -1.35516599e-01 -1.74599722e-01
6.87286735e-01 3.32106739e-01 1.06611180e+00 -1.35578120e+00
-1.12501882e-01 -1.08751678e+00 -1.17528892e+00 -9.62714851e-01
-9.17912796e-02 -8.57601762e-01 -3.06919347e-02 -1.38633537e+00
1.38295554e-02 -4.41955119e-01 -1.64858878e-01 4.81028229e-01
-1.43302500e-01 6.24393642e-01 2.73596585e-01 2.47976035e-01
-3.60286951e-01 9.79884565e-02 8.51565123e-01 -2.01243803e-01
1.50131509e-01 -8.14155340e-02 -2.40322769e-01 4.97343838e-01
3.75303149e-01 -3.21189255e-01 7.99207017e-02 -1.68329313e-01
1.79985106e-01 -3.18195164e-01 3.84283185e-01 -1.35500646e+00
4.94134992e-01 5.87334573e-01 4.20421928e-01 -6.77847564e-01
2.59297460e-01 -1.10152757e+00 1.82905778e-01 8.80258322e-01
-1.86919644e-02 6.26630604e-01 1.78833574e-01 -2.91766096e-02
-5.34668565e-01 -1.95326731e-01 1.06050718e+00 -1.72202721e-01
-1.10941851e+00 4.04402018e-01 -1.04520641e-01 -4.51969415e-01
9.90727723e-01 -8.57012987e-01 -8.87114927e-02 -1.03255130e-01
-3.67498368e-01 6.02686107e-02 6.84268653e-01 5.74887514e-01
3.65818709e-01 -1.19075072e+00 -6.16887152e-01 4.03019488e-01
4.05633390e-01 -2.11663634e-01 6.90947920e-02 9.75398719e-01
-9.54851806e-01 5.80359280e-01 -5.09652615e-01 -8.78887892e-01
-1.39394581e+00 7.20491767e-01 5.40963352e-01 -1.94601268e-01
-9.26240027e-01 6.18688583e-01 -2.73029178e-01 -9.28263813e-02
4.86158490e-01 -5.10287762e-01 -1.43471420e-01 2.21413016e-01
4.70215887e-01 3.32956970e-01 5.20734966e-01 -7.72583067e-01
-5.60301960e-01 1.25445640e+00 -6.27657548e-02 2.17532188e-01
1.31098115e+00 1.05202474e-01 -3.34847830e-02 -9.52435136e-02
1.43370247e+00 -2.29969714e-02 -9.04966295e-01 -1.04229257e-01
4.80207443e-01 -5.93585372e-01 -1.34489790e-01 -4.91064191e-01
-1.09499359e+00 9.94817972e-01 1.11083651e+00 1.11943580e-01
1.01417851e+00 -6.86413422e-02 6.92535102e-01 3.01285744e-01
6.91939116e-01 -6.98087335e-01 -2.13913828e-01 3.51630598e-01
5.86016834e-01 -1.44829750e+00 -3.74340974e-02 -1.75339669e-01
4.91963141e-02 1.47220135e+00 2.48901993e-01 -6.65830016e-01
7.08004236e-01 5.02928972e-01 3.89273345e-01 -3.56375903e-01
-1.74780533e-01 -6.10289335e-01 3.99091214e-01 7.25069702e-01
5.05993009e-01 -2.81692922e-01 -4.09036905e-01 -1.20979995e-01
-3.18513736e-02 -3.95741016e-02 -2.04318345e-01 7.94586420e-01
-5.61479986e-01 -1.55982637e+00 -5.22403240e-01 -8.14936757e-02
-2.44181588e-01 -1.17623713e-04 -3.49531084e-01 1.40734935e+00
3.59548777e-01 5.24801850e-01 -1.66911349e-01 -3.36170465e-01
1.00149620e+00 -2.02423424e-01 7.32488990e-01 -5.20390980e-02
-1.07009172e+00 -1.04788683e-01 -4.55885082e-01 -9.06608582e-01
-4.88007963e-01 -5.99101126e-01 -1.11320961e+00 -4.65531945e-01
-2.19422698e-01 1.60912260e-01 7.73390830e-01 7.91680992e-01
1.91682801e-01 1.38378486e-01 4.95711029e-01 -1.61090064e+00
-6.09620750e-01 -7.23659277e-01 -4.67773944e-01 6.90604508e-01
3.25806320e-01 -7.53243208e-01 -2.68601954e-01 -3.71007204e-01] | [10.482585906982422, 0.2142300009727478] |
e1effbe3-00d0-4407-9386-19bb3966b3e1 | a-conditional-generative-adversarial-network-1 | null | null | https://ieeexplore.ieee.org/document/8519215 | https://ieeexplore.ieee.org/document/8519215 | A Conditional Generative Adversarial Network to Fuse Sar And Multispectral Optical Data For Cloud Removal From Sentinel-2 Images | In this paper, we present the first conditional generative adversarial network (cGAN) architecture that is specifically designed to fuse synthetic aperture radar (SAR) and optical multi-spectral (MS) image data to generate cloud- and haze-free MS optical data from a cloud-corrupted MS input and an auxiliary SAR image. Experiments on Sentinel-2 MS and Sentinel-l SAR data confirm that our extended SAR-Opt-cGAN model utilizes the auxiliary SAR information to better reconstruct MS images than an equivalent model which uses the same architecture but only single-sensor MS data as input. | ['Xiaoxiang Zhu', 'Michael Schmitt', 'Claas Grohnfeldt'] | 2018-11-04 | null | null | null | igarss-2018-11 | ['cloud-removal'] | ['computer-vision'] | [ 6.80831909e-01 -4.68098819e-02 5.70086837e-01 -2.96028972e-01
-1.03960121e+00 -8.28000247e-01 5.86325109e-01 -8.00480187e-01
-3.36989552e-01 1.05327129e+00 -2.30366603e-01 -6.51964366e-01
3.73146348e-02 -9.30072129e-01 -5.94967365e-01 -1.10179448e+00
2.36396879e-01 1.94510520e-01 -2.55200937e-02 -3.43068689e-01
-4.52488929e-01 9.27555740e-01 -1.15899551e+00 3.05375397e-01
9.55806196e-01 1.20960224e+00 4.30188209e-01 9.45697844e-01
6.10333204e-01 9.17038620e-01 -7.21659541e-01 1.32969677e-01
1.09419441e+00 -6.20863557e-01 -3.02370310e-01 2.68402696e-01
7.44217455e-01 -4.29065347e-01 -7.16636479e-01 1.23934066e+00
3.35754037e-01 7.22076790e-03 4.01180238e-01 -9.64406133e-01
-8.32115769e-01 1.40421361e-01 -2.18154937e-01 1.41363725e-01
-3.92527312e-01 7.37585545e-01 2.93941528e-01 -4.91427839e-01
5.03832698e-01 9.24145162e-01 4.19088006e-01 4.85430866e-01
-1.19580555e+00 -5.33314943e-01 -2.71423399e-01 -4.94048983e-01
-1.39969563e+00 -3.78640443e-01 4.83870327e-01 -4.37470466e-01
6.77424908e-01 4.15795386e-01 7.34849632e-01 9.29699302e-01
6.99492395e-01 3.43076319e-01 1.63108683e+00 -1.96528524e-01
1.05522372e-01 8.04746523e-03 -3.29060286e-01 3.99705082e-01
6.63741112e-01 8.06720018e-01 -2.03688309e-01 -1.82231471e-01
7.62851894e-01 2.43762508e-01 -7.14527905e-01 1.13810234e-01
-1.07639909e+00 1.21444631e+00 8.33719313e-01 -1.24046381e-03
-9.24283862e-01 1.70566529e-01 -7.74502099e-01 5.74296832e-01
4.76123661e-01 7.96940267e-01 -3.14834453e-02 1.03051734e+00
-1.47063339e+00 3.14726323e-01 2.51747191e-01 7.76459455e-01
1.15237665e+00 1.40586305e+00 5.46722412e-02 2.63702441e-02
4.63546604e-01 1.85464597e+00 1.33135870e-01 -9.56470668e-01
8.18252861e-02 -7.54548842e-03 3.86242926e-01 -5.05296826e-01
-2.09018966e-04 -8.03085566e-01 -9.43357110e-01 6.02111042e-01
-3.62581581e-01 -6.54678881e-01 -1.73673713e+00 1.30347574e+00
-1.38887167e-01 6.99642003e-02 1.00257754e+00 1.20491385e+00
9.16368008e-01 8.47282350e-01 -2.02880591e-01 -1.20721228e-01
6.86503410e-01 -5.67435265e-01 -6.07591927e-01 -1.33449399e+00
-2.79458821e-01 -7.00869262e-01 2.57464558e-01 5.69331311e-02
-7.00728774e-01 -5.30214310e-01 -1.46390235e+00 6.49866462e-01
-3.58856261e-01 5.58084846e-02 7.99919844e-01 5.35085797e-01
-1.18368459e+00 9.95351672e-02 -6.63337231e-01 4.52853041e-03
3.17427874e-01 -1.99507661e-02 -8.46431628e-02 -3.45671326e-01
-1.34211695e+00 7.42381275e-01 6.78261638e-01 3.68261874e-01
-1.73438406e+00 -7.88893104e-01 -1.00529587e+00 -5.75191915e-01
1.42761514e-01 -1.01296449e+00 7.68312037e-01 -1.45411313e+00
-1.04782236e+00 6.27355039e-01 3.86299640e-01 -8.35600078e-01
-1.07823387e-01 -2.05329061e-01 -1.00466621e+00 6.58972204e-01
-1.48321331e-01 7.21518040e-01 1.32235575e+00 -1.90175569e+00
-1.24404833e-01 -4.74690080e-01 -1.51647657e-01 6.46760091e-02
7.95280218e-01 -1.26610771e-01 4.62143123e-01 -7.60648131e-01
2.49729082e-01 -1.01635361e+00 -6.38751626e-01 -3.33157510e-01
-6.02775335e-01 1.29624081e+00 8.48314703e-01 -6.25000834e-01
3.48161578e-01 -2.04715800e+00 3.91321369e-02 9.64311883e-02
6.77016750e-02 6.22525632e-01 -6.31838620e-01 3.34590226e-01
-7.32273757e-02 -2.23477975e-01 -8.97913814e-01 -9.36880857e-02
-5.90627730e-01 3.74267220e-01 -7.93867350e-01 4.67376828e-01
5.79471350e-01 1.24939322e+00 -4.86943990e-01 1.51589900e-01
7.79210255e-02 5.36632359e-01 3.81001504e-03 5.08472085e-01
-6.69875324e-01 8.58449578e-01 -4.44215417e-01 1.03640032e+00
1.34646058e+00 -6.60269633e-02 3.78381014e-02 -1.39139295e-01
-6.73830658e-02 -5.66210210e-01 -7.29188025e-01 1.24859416e+00
-7.77481794e-02 5.13405383e-01 6.68027699e-01 -4.45309520e-01
1.03628135e+00 1.53108276e-02 1.35991186e-01 -9.43019688e-01
4.40430604e-02 9.59526002e-02 -6.08557835e-02 -2.39480853e-01
7.53859580e-01 -7.53390372e-01 -2.63683330e-02 3.60181957e-01
-6.98862150e-02 -8.80261898e-01 -6.42001033e-01 2.60060847e-01
8.55140388e-01 -1.50063455e-01 -1.12679265e-01 4.06440720e-02
3.56412858e-01 5.55818677e-01 5.94817102e-01 8.50955069e-01
1.20432444e-01 7.76906550e-01 -5.94565153e-01 -4.06034380e-01
-9.46641326e-01 -1.52936816e+00 2.65265644e-01 -7.37343654e-02
-1.71133969e-02 4.23157692e-01 -2.90756196e-01 -4.18829083e-01
4.07807678e-02 9.52619433e-01 -3.81880522e-01 -1.03147283e-01
-2.74344057e-01 -1.14653957e+00 7.02359378e-01 6.32625967e-02
7.93370724e-01 -8.66307557e-01 -7.07352221e-01 1.14345483e-01
-1.56831473e-01 -1.38305104e+00 1.57106429e-01 -7.43177459e-02
-8.02496791e-01 -1.08181477e+00 -4.21557039e-01 -8.84284973e-02
5.72334230e-01 6.78476930e-01 1.01243389e+00 -3.64580631e-01
-4.17499393e-01 4.43912804e-01 -5.07468283e-01 -7.76740849e-01
-7.63807774e-01 -8.92344713e-01 4.68592346e-02 5.21146417e-01
-2.11453289e-02 -3.95646244e-01 -5.29396594e-01 -3.24670762e-01
-1.69020414e+00 7.13054463e-02 9.55621004e-01 7.99509168e-01
6.22702956e-01 -2.98175290e-02 6.46188483e-02 -6.43734872e-01
-1.49300545e-01 -3.88528645e-01 -1.02820551e+00 1.09637804e-01
-4.80413079e-01 -4.54640239e-02 4.24510539e-01 1.99683849e-02
-1.12441814e+00 2.84380555e-01 1.73329636e-01 -9.14230049e-01
-2.28587419e-01 8.18765581e-01 -1.08279299e-03 -7.12353885e-01
7.27195680e-01 8.18140030e-01 2.57713825e-01 -2.71716475e-01
1.50420532e-01 7.27851570e-01 1.05909717e+00 2.57551700e-01
1.60267603e+00 1.07688868e+00 2.15596363e-01 -1.07458651e+00
-1.17535794e+00 -1.54857725e-01 -3.74596655e-01 -1.67016864e-01
1.20635176e+00 -1.47015893e+00 2.99808353e-01 7.87057817e-01
-1.11773002e+00 -5.94768107e-01 -2.69225210e-01 8.27275276e-01
-3.47928137e-01 1.88281611e-01 -1.67346165e-01 -1.08507049e+00
-6.50408983e-01 -7.72092402e-01 1.09004664e+00 1.51305288e-01
7.81898022e-01 -6.96302652e-01 1.99720412e-01 5.20241380e-01
8.17406654e-01 9.21463907e-01 3.58652204e-01 -2.80202590e-02
-1.45865393e+00 -1.34490639e-01 -2.08083138e-01 9.82904196e-01
1.50937393e-01 -2.60554224e-01 -8.63693178e-01 -9.19890225e-01
3.05069417e-01 -4.24870282e-01 1.39009190e+00 4.90930617e-01
4.57412004e-01 -5.38576186e-01 1.91476107e-01 1.30203211e+00
2.30313921e+00 2.63074666e-01 1.07382858e+00 2.71259546e-01
5.14373422e-01 7.70073086e-02 8.18297982e-01 3.16085935e-01
-1.17464758e-01 1.47299796e-01 1.17078841e+00 -4.86651182e-01
-6.89707249e-02 1.28785372e-01 5.11446476e-01 7.25292340e-02
2.06667231e-03 -6.58544123e-01 -6.35624588e-01 4.40636128e-01
-1.27145338e+00 -1.06853640e+00 -2.48578340e-01 2.11674547e+00
2.97332853e-01 -4.17836547e-01 -6.54740453e-01 -5.14941216e-01
1.44537434e-01 8.06052506e-01 -3.60887170e-01 2.64955550e-01
-8.51127267e-01 8.06772411e-01 1.16708398e+00 8.16739380e-01
-1.10711670e+00 1.09261334e+00 6.68365860e+00 7.72867277e-02
-1.44820940e+00 2.25925505e-01 1.20178640e-01 6.94853067e-02
-8.03124189e-01 1.74888641e-01 -4.66917783e-01 2.20464781e-01
1.34720051e+00 2.48237759e-01 5.75750470e-01 2.56823838e-01
1.07513377e-02 -3.26911151e-01 -2.59157538e-01 7.47776687e-01
2.58418590e-01 -1.57695448e+00 5.16478717e-01 1.31410033e-01
9.40798819e-01 5.98075151e-01 1.99048847e-01 -1.95251405e-01
6.87873304e-01 -1.10399938e+00 5.80445826e-01 1.20692968e+00
1.11797071e+00 -6.24269605e-01 9.73941743e-01 3.34579647e-01
-5.54021657e-01 1.47056580e-02 -3.73254478e-01 2.54869223e-01
2.38136768e-01 7.28038609e-01 -6.76955640e-01 1.29193223e+00
5.78304887e-01 2.66842335e-01 -5.79769254e-01 6.20152116e-01
-2.89760590e-01 5.78579068e-01 -1.15253784e-01 7.82570720e-01
7.29348600e-01 -5.29245436e-01 9.57467854e-01 7.40700126e-01
7.35794246e-01 7.05578923e-01 2.77640820e-01 1.20564437e+00
4.65601951e-01 -8.23821366e-01 -9.76306140e-01 -4.71673846e-01
2.26951852e-01 1.12569284e+00 1.27002001e-01 -2.59299725e-01
-8.33833963e-02 9.85674322e-01 -5.84810674e-01 7.67925441e-01
-7.03304946e-01 -1.28279813e-03 1.08083463e+00 3.87012511e-02
6.51870430e-01 -6.60370886e-01 2.10259408e-01 -9.66774762e-01
-5.48687398e-01 -9.31611061e-01 1.08655021e-01 -1.46790707e+00
-1.15456247e+00 1.19729400e+00 -2.27652058e-01 -1.53810179e+00
-3.38659078e-01 -5.10269463e-01 -4.49541301e-01 1.50547409e+00
-2.36315107e+00 -1.83912730e+00 -9.64642525e-01 7.93461800e-01
-5.44994101e-02 -6.38608038e-01 9.46359932e-01 -3.55219424e-01
-2.69678384e-02 -4.99894693e-02 1.01929836e-01 3.19796726e-02
4.17869747e-01 -8.05414557e-01 3.93534362e-01 1.61597705e+00
-1.65742055e-01 1.83027253e-01 8.39632213e-01 -1.05807364e+00
-1.72872436e+00 -1.97498119e+00 3.71654212e-01 -1.62789479e-01
3.48653525e-01 7.05974326e-02 -5.19880533e-01 1.18574977e+00
6.08266234e-01 3.63429278e-01 5.87780237e-01 -1.27241421e+00
-3.89960468e-01 -3.49041134e-01 -1.38095224e+00 4.15363386e-02
2.60814905e-01 -5.10765493e-01 -5.79136014e-01 5.61209559e-01
1.17207515e+00 -4.49202389e-01 -6.65323138e-01 9.19477046e-01
-1.38826713e-01 -1.01297152e+00 1.15713477e+00 -6.68515146e-01
2.18190596e-01 -9.25283313e-01 -9.05427158e-01 -1.72676480e+00
-4.05974358e-01 -6.08674467e-01 1.60955161e-01 3.89381975e-01
3.55634063e-01 -8.69671881e-01 4.15498227e-01 -3.26648116e-01
-6.59207880e-01 2.42079362e-01 -6.74522042e-01 -1.14293754e+00
-4.64033745e-02 -5.91610745e-02 6.47810519e-01 8.22919250e-01
-1.57961512e+00 -2.32214965e-02 -7.23771572e-01 1.30424798e+00
1.19414485e+00 5.40253639e-01 7.65543520e-01 -1.11394966e+00
-4.73167300e-01 4.40488249e-01 -2.35617176e-01 -4.13671285e-01
-8.43877196e-02 -7.51087964e-01 2.60980815e-01 -1.40706658e+00
-1.74673945e-01 -3.05576026e-01 -1.86703101e-01 3.24699730e-01
-8.07407349e-02 7.86185980e-01 3.51455033e-01 1.76898777e-01
6.11801073e-02 6.66055381e-01 1.32480383e+00 -4.85014081e-01
2.88337618e-01 3.44586447e-02 -6.44757926e-01 2.96828449e-02
6.22638762e-01 -4.19296294e-01 -1.71008348e-01 -6.88551307e-01
3.14322151e-02 4.50836271e-01 1.07717586e+00 -1.40734506e+00
7.11555630e-02 -6.65082514e-01 5.70446074e-01 -8.84927928e-01
6.98087692e-01 -7.96703815e-01 8.67933154e-01 5.63434243e-01
4.41898108e-01 -1.12198427e-01 3.26059937e-01 6.46713257e-01
-4.80161518e-01 8.33240971e-02 1.21547949e+00 -5.69307446e-01
-8.10488641e-01 6.93592370e-01 -4.24345732e-01 -3.00150305e-01
9.28750455e-01 -2.62020808e-02 -7.35587478e-01 -5.97336769e-01
-6.32755697e-01 -3.93316373e-02 5.41100979e-01 -9.57440957e-03
1.09770596e+00 -1.05328858e+00 -1.35034394e+00 7.60171473e-01
2.06130937e-01 4.68716510e-02 5.49384058e-01 4.46544141e-01
-7.90306330e-01 5.25622189e-01 -4.13649917e-01 -4.71468955e-01
-8.65790665e-01 4.23176229e-01 7.31689572e-01 -4.32477742e-02
-2.66521186e-01 8.67595375e-01 -1.23734018e-02 -6.40146792e-01
-7.34289885e-01 1.77295238e-01 3.66401345e-01 -4.91192281e-01
6.97443724e-01 -2.62350529e-01 1.39934555e-01 -7.35350907e-01
-2.10795358e-01 2.40644410e-01 5.35244644e-01 -3.74948949e-01
1.53830910e+00 1.07664444e-01 -2.56948918e-01 7.49752447e-02
5.03193557e-01 1.58527777e-01 -1.33797145e+00 -4.02477294e-01
-6.47768319e-01 -6.30504072e-01 6.40243530e-01 -1.17945492e+00
-1.55010450e+00 5.59612215e-01 6.31129682e-01 -6.44730702e-02
1.47584677e+00 -8.25166479e-02 7.63904393e-01 3.95400882e-01
3.15268517e-01 -5.45403004e-01 1.16149588e-02 6.24719501e-01
1.09858406e+00 -1.17756641e+00 1.38293549e-01 -2.79232502e-01
-8.76843512e-01 8.67412210e-01 3.36429477e-01 -2.13535875e-01
4.79960591e-01 3.75279516e-01 4.82293725e-01 -6.35809422e-01
-4.31445301e-01 -6.19309783e-01 1.81792513e-01 9.20069098e-01
-4.66881633e-01 3.67168427e-01 5.90302587e-01 1.13662049e-01
-9.38848928e-02 -2.14354962e-01 8.67984176e-01 1.03945088e+00
-6.02673829e-01 -7.14094162e-01 -6.87948763e-01 4.45509523e-01
2.11740900e-02 -5.11324704e-01 -4.31860209e-01 8.11501265e-01
-3.93503532e-03 1.07150733e+00 3.00059140e-01 -4.97956932e-01
-8.26899260e-02 -8.93649980e-02 4.81439531e-01 -7.00473189e-01
-3.38597536e-01 -3.02488338e-02 -3.67919467e-02 -7.35836267e-01
-7.23194718e-01 -6.38092697e-01 -7.28815556e-01 -3.12134027e-01
-3.79874855e-02 5.95663898e-02 7.10609615e-01 7.54005671e-01
3.49342853e-01 4.77864742e-01 1.12529874e+00 -6.97650194e-01
-4.96572047e-01 -1.00158072e+00 -1.02264023e+00 -2.39694834e-01
1.14391935e+00 -1.57834768e-01 -7.18503535e-01 -2.30376065e-01] | [10.084554672241211, -2.0867834091186523] |
c95cba0f-eae8-421e-bbe1-07b76290a71a | gflownets-for-ai-driven-scientific-discovery | 2302.00615 | null | https://arxiv.org/abs/2302.00615v2 | https://arxiv.org/pdf/2302.00615v2.pdf | GFlowNets for AI-Driven Scientific Discovery | Tackling the most pressing problems for humanity, such as the climate crisis and the threat of global pandemics, requires accelerating the pace of scientific discovery. While science has traditionally relied on trial and error and even serendipity to a large extent, the last few decades have seen a surge of data-driven scientific discoveries. However, in order to truly leverage large-scale data sets and high-throughput experimental setups, machine learning methods will need to be further improved and better integrated in the scientific discovery pipeline. A key challenge for current machine learning methods in this context is the efficient exploration of very large search spaces, which requires techniques for estimating reducible (epistemic) uncertainty and generating sets of diverse and informative experiments to perform. This motivated a new probabilistic machine learning framework called GFlowNets, which can be applied in the modeling, hypotheses generation and experimental design stages of the experimental science loop. GFlowNets learn to sample from a distribution given indirectly by a reward function corresponding to an unnormalized probability, which enables sampling diverse, high-reward candidates. GFlowNets can also be used to form efficient and amortized Bayesian posterior estimators for causal models conditioned on the already acquired experimental data. Having such posterior models can then provide estimators of epistemic uncertainty and information gain that can drive an experimental design policy. Altogether, here we will argue that GFlowNets can become a valuable tool for AI-driven scientific discovery, especially in scenarios of very large candidate spaces where we have access to cheap but inaccurate measurements or to expensive but accurate measurements. This is a common setting in the context of drug and material discovery, which we use as examples throughout the paper. | ['Yoshua Bengio', 'Alex Hernandez-Garcia', 'Cheng-Hao Liu', 'Jason Hartford', 'Tristan Deleu', 'Moksh Jain'] | 2023-02-01 | null | null | null | null | ['experimental-design'] | ['methodology'] | [ 3.48889738e-01 8.54683593e-02 -4.25899476e-01 -1.40656680e-01
-8.21892619e-01 -6.05259299e-01 7.63883948e-01 4.69558775e-01
-5.98493993e-01 1.23759782e+00 -5.77545725e-02 -6.53405428e-01
-4.66221571e-01 -7.82264531e-01 -8.78609896e-01 -8.84350300e-01
-1.17442384e-01 9.70667541e-01 -7.23256767e-02 2.54653752e-01
3.14881772e-01 4.81137842e-01 -1.47874689e+00 -1.93636730e-01
7.93329835e-01 7.80068338e-01 2.14171559e-01 5.21095753e-01
1.09871939e-01 9.87738743e-02 -2.69905657e-01 -1.81883544e-01
-5.91651313e-02 -3.65894914e-01 -4.91991937e-01 -6.23417318e-01
-3.78571481e-01 -6.69292063e-02 3.94367427e-01 1.08591938e+00
6.76926732e-01 5.00756465e-02 7.35031366e-01 -8.76691222e-01
-5.79939149e-02 7.68863618e-01 -4.12185252e-01 -1.56013202e-02
1.68173656e-01 5.38460672e-01 9.64291573e-01 -4.46951270e-01
6.57474458e-01 1.43231821e+00 2.15837240e-01 2.42188126e-01
-1.61529136e+00 -7.72146583e-01 -1.46464512e-01 1.66132495e-01
-1.03635979e+00 -3.27654123e-01 5.24838328e-01 -5.98429024e-01
3.01859587e-01 2.07809240e-01 6.86304688e-01 1.55487406e+00
1.90332562e-01 6.17323518e-01 1.29937124e+00 -3.58645886e-01
9.57178891e-01 9.62516144e-02 -2.90004939e-01 3.33612859e-01
7.10220754e-01 5.79114735e-01 -5.05576551e-01 -3.93521845e-01
6.10952079e-01 3.02169360e-02 -2.61051297e-01 -4.20317143e-01
-1.32338107e+00 1.08777416e+00 9.08506289e-02 2.09237978e-01
-4.79872823e-01 3.25037569e-01 2.27047384e-01 3.19328122e-02
2.51790375e-01 1.06127477e+00 -7.94843912e-01 -3.25335741e-01
-8.99088621e-01 6.11700773e-01 7.69086778e-01 1.95685580e-01
5.58183849e-01 -3.09097618e-01 -4.14512008e-02 3.00015837e-01
4.50363487e-01 6.71615899e-01 1.38933375e-01 -8.31202030e-01
2.06833724e-02 3.09668541e-01 5.59177279e-01 -5.14376044e-01
-4.49398845e-01 -6.45277917e-01 -6.59479678e-01 2.65176356e-01
7.19300210e-01 -2.12078035e-01 -8.07539999e-01 1.97210038e+00
7.45002151e-01 2.63228267e-01 -1.63146242e-01 8.06793213e-01
4.79342043e-02 5.07689655e-01 2.96674520e-01 -3.77645850e-01
1.36642063e+00 8.52764621e-02 -3.38639826e-01 -2.07700711e-02
5.73977947e-01 -4.50060308e-01 1.03321505e+00 7.23611951e-01
-8.36686432e-01 5.59173375e-02 -9.44675088e-01 2.78931588e-01
-4.14477557e-01 -2.01048836e-01 8.68760943e-01 7.30065048e-01
-3.95454019e-01 7.76039362e-01 -8.36080074e-01 -2.34501302e-01
8.83164585e-01 3.84628564e-01 -9.65176001e-02 -2.93944031e-01
-1.24196613e+00 8.45449448e-01 5.10065258e-01 2.05718771e-01
-1.33240902e+00 -1.00058186e+00 -4.49372172e-01 1.25435352e-01
8.12918842e-01 -8.97060275e-01 9.32462871e-01 -3.13787341e-01
-1.52308941e+00 4.73505139e-01 2.66795695e-01 -5.27786672e-01
5.88240921e-01 -6.37170896e-02 6.80818558e-02 -2.44729117e-01
-4.26653661e-02 5.45112610e-01 7.62912631e-01 -8.22135687e-01
-3.69305432e-01 -6.14834428e-01 -2.12975860e-01 -3.79438102e-01
1.37717009e-01 -5.61847165e-02 3.58274668e-01 -2.29143545e-01
-1.49018869e-01 -9.50777650e-01 -5.50503135e-01 -1.16942763e-01
-4.93068248e-01 -2.16757372e-01 3.79406303e-01 -2.04404086e-01
7.67183483e-01 -1.62228858e+00 4.40290302e-01 3.69762003e-01
9.86242890e-02 2.24354014e-01 2.99485009e-02 3.93116266e-01
1.82861865e-01 3.71414125e-01 -2.60271728e-01 1.58639282e-01
1.97807655e-01 -1.53499901e-01 -3.03485483e-01 5.35062015e-01
1.78247973e-01 8.59478951e-01 -1.18390501e+00 -9.14667472e-02
1.41947627e-01 4.83745307e-01 -4.72667605e-01 2.03601599e-01
-1.06437981e+00 9.94434595e-01 -8.89625549e-01 5.24794638e-01
3.02673191e-01 -3.07949185e-01 2.24945068e-01 1.38550863e-01
-1.82911530e-01 1.44350290e-01 -1.06312370e+00 1.67169762e+00
-4.66516912e-01 3.91098820e-02 -5.64019047e-02 -1.13130343e+00
6.79061353e-01 1.96239799e-01 4.17370528e-01 -4.12823141e-01
4.54374939e-01 4.32191789e-01 1.78305551e-01 -3.37552696e-01
-3.96585874e-02 -5.84865630e-01 -1.45384729e-01 5.80834568e-01
-5.73353395e-02 -2.21252054e-01 5.99354021e-02 -4.21307646e-02
1.13687944e+00 3.38820308e-01 3.55828851e-01 -3.03453565e-01
3.02390009e-01 -7.52713671e-03 5.79093874e-01 7.50810981e-01
1.36648426e-02 2.18141265e-02 8.36235344e-01 -5.15347302e-01
-1.11220098e+00 -1.01355648e+00 -4.05685931e-01 6.73874676e-01
-3.59793365e-01 -1.16511315e-01 -4.85993445e-01 -4.23055023e-01
1.14588842e-01 7.70866334e-01 -5.81065476e-01 -3.18354279e-01
-1.41000703e-01 -1.06853282e+00 2.35992104e-01 5.20394593e-02
-1.17624164e-01 -9.14270818e-01 -7.99390793e-01 3.15463036e-01
2.87267894e-01 -7.28312731e-01 2.31411219e-01 3.80579233e-01
-7.29762375e-01 -1.18322492e+00 -7.03792155e-01 6.24947250e-02
2.78557241e-01 -3.77955675e-01 9.08607185e-01 -4.75791216e-01
-6.57730877e-01 5.13721146e-02 -1.38171405e-01 -9.35909629e-01
-4.80082244e-01 -5.15879951e-02 2.29221359e-01 -2.23587155e-01
2.69195229e-01 -8.62742186e-01 -7.45951951e-01 1.29192382e-01
-9.67891037e-01 -1.17539234e-01 7.90862322e-01 1.06133962e+00
7.37813473e-01 5.65317757e-02 1.08169270e+00 -8.09508979e-01
5.48337758e-01 -7.76454628e-01 -1.38947546e+00 2.57375211e-01
-6.62794113e-01 6.59588814e-01 5.29978335e-01 -5.48643708e-01
-8.32957387e-01 -1.15923151e-01 -9.96696874e-02 -2.21403107e-01
-1.94800988e-01 8.79148722e-01 -2.73966253e-01 2.04185262e-01
7.83092141e-01 -3.17357361e-01 8.73312503e-02 -5.13272941e-01
6.41094625e-01 5.17984271e-01 1.19685419e-01 -9.95378196e-01
4.96435672e-01 4.30815965e-01 5.55097103e-01 -6.97153270e-01
-8.71318758e-01 -4.95896600e-02 -2.14346945e-01 3.07900906e-02
6.78244829e-01 -6.84609652e-01 -1.20719659e+00 1.14881266e-02
-8.38009477e-01 -3.53031844e-01 -4.86585408e-01 7.82617629e-01
-5.48020542e-01 -9.87163931e-02 7.96680897e-03 -1.15068030e+00
-4.20024842e-02 -1.27456260e+00 1.01222277e+00 2.66001552e-01
-2.32802525e-01 -9.54725683e-01 4.07888889e-01 3.33640724e-01
4.95322227e-01 4.36954081e-01 1.13556600e+00 -8.05539966e-01
-8.90172005e-01 -2.69450694e-01 4.84153628e-02 -1.00365411e-02
-1.31587654e-01 -1.59743890e-01 -9.80586231e-01 -1.07510231e-01
-5.70969805e-02 -6.22953594e-01 8.49141598e-01 8.08023632e-01
1.24781489e+00 -2.14201137e-01 -4.29254860e-01 3.04742396e-01
1.10704935e+00 1.17116317e-01 3.42393309e-01 3.74703966e-02
2.86493689e-01 7.59306848e-01 5.14683485e-01 5.85550725e-01
4.23049591e-02 6.28330648e-01 4.21022117e-01 4.42552298e-01
4.25922185e-01 -3.33694637e-01 2.25863662e-02 -2.08998732e-02
1.46884426e-01 -1.74625039e-01 -7.91712284e-01 3.09246033e-01
-1.78780472e+00 -9.65859354e-01 2.62084901e-01 2.87012339e+00
1.24849653e+00 2.77281016e-01 2.13246062e-01 -8.37861970e-02
5.61208427e-01 -2.98994422e-01 -9.10581946e-01 -1.16532609e-01
6.61735460e-02 4.16237116e-01 4.59919065e-01 3.75346303e-01
-6.55172467e-01 6.54845595e-01 6.11018133e+00 9.62934792e-01
-1.18232632e+00 -4.68529835e-02 9.77854133e-01 -2.04549611e-01
-6.61639154e-01 2.94741184e-01 -8.28079343e-01 6.09118462e-01
1.48301733e+00 -2.20399097e-01 4.87983495e-01 5.44407547e-01
5.81162453e-01 -5.45965731e-01 -1.31355906e+00 7.92369008e-01
-6.87887132e-01 -1.56810057e+00 -2.95694619e-01 3.20935279e-01
5.33180416e-01 1.56314954e-01 9.56619754e-02 9.15309489e-02
7.85994053e-01 -1.27040410e+00 3.13635260e-01 6.37945116e-01
7.58207619e-01 -7.02004433e-01 6.40024841e-01 5.63370645e-01
-3.38580966e-01 -5.67118973e-02 -4.43865329e-01 -3.04065794e-02
1.59706146e-01 1.37625909e+00 -1.00313187e+00 3.36437792e-01
3.25481117e-01 3.47405165e-01 1.16245873e-01 1.25921655e+00
-3.10199320e-01 7.91426718e-01 -7.23651588e-01 -3.20696026e-01
-4.62072566e-02 -4.18653101e-01 5.76659799e-01 6.26472712e-01
4.70827311e-01 -2.36479655e-01 -1.08974442e-01 1.48050737e+00
-1.11769184e-01 -2.57105201e-01 -5.57316959e-01 -5.68700194e-01
6.31386578e-01 1.23588669e+00 -7.31521130e-01 -3.53983790e-02
1.28918126e-01 9.80354026e-02 9.81949642e-03 3.10766786e-01
-6.74654782e-01 7.99006000e-02 4.60704893e-01 1.96484234e-02
2.20723078e-01 -9.89203155e-02 -1.97411656e-01 -9.87282336e-01
-3.47292662e-01 -9.19036210e-01 3.07959437e-01 -4.23745751e-01
-1.17759120e+00 -1.49101093e-01 3.70862156e-01 -5.10052979e-01
-3.89231741e-01 -5.99000633e-01 -3.90741467e-01 8.93170714e-01
-1.41715324e+00 -7.98291445e-01 4.17866051e-01 4.05526273e-02
7.83649758e-02 8.11165795e-02 7.48332739e-01 -1.18995517e-01
-5.97161174e-01 5.31175844e-02 4.09962237e-01 -6.15615785e-01
5.94723463e-01 -1.07867610e+00 -1.71114486e-02 4.06731308e-01
1.47095546e-01 5.35285294e-01 1.15836859e+00 -7.52344251e-01
-1.70377624e+00 -8.15835655e-01 4.59540933e-01 -4.18176591e-01
9.32394683e-01 -4.99680161e-01 -6.12114906e-01 2.68584877e-01
-4.34038490e-01 -4.22396772e-02 6.53565526e-01 4.89975035e-01
-4.13935572e-01 -4.07060944e-02 -1.18381917e+00 5.40755689e-01
5.26560426e-01 -2.62635499e-01 -3.08779597e-01 5.53963900e-01
5.63489497e-01 -6.08708989e-03 -9.12074387e-01 3.38683456e-01
6.93413258e-01 -6.39897883e-01 8.36601377e-01 -7.80159593e-01
2.26812065e-01 -2.41651937e-01 7.96420127e-02 -1.27426517e+00
9.27433372e-02 -9.40776944e-01 -7.07853166e-03 9.08017278e-01
5.78099191e-01 -6.18680477e-01 6.95390284e-01 6.17100060e-01
2.93976754e-01 -8.99880230e-01 -1.06131077e+00 -7.08273649e-01
3.18167984e-01 -5.45998514e-01 5.92689037e-01 5.80622971e-01
-5.03812805e-02 2.73717284e-01 -2.68073857e-01 -6.76152930e-02
8.56500030e-01 5.98109588e-02 5.77486694e-01 -1.68368435e+00
-5.91993570e-01 -5.18933713e-01 -3.05946559e-01 -4.83284950e-01
-9.96412709e-03 -6.37969851e-01 1.01649702e-01 -1.09730768e+00
3.58715057e-01 -6.62459731e-01 -2.59235471e-01 1.80009544e-01
-1.98221341e-01 -2.72298545e-01 -3.71864110e-01 -1.10351242e-01
-5.17477274e-01 6.00808978e-01 9.69190240e-01 -3.63782346e-02
-1.86940044e-01 1.68110371e-01 -8.86136651e-01 5.36065996e-01
5.98056555e-01 -6.30937815e-01 -5.03169894e-01 2.08330542e-01
9.04133201e-01 1.08335696e-01 5.72293878e-01 -6.97843671e-01
-1.12008592e-02 -5.86237729e-01 2.55435050e-01 -2.00982034e-01
3.35219242e-02 -6.05912864e-01 5.07929683e-01 4.63945091e-01
-4.77819860e-01 -5.79018950e-01 5.60240522e-02 8.67790997e-01
2.59591937e-01 -1.63049027e-01 7.24155426e-01 -1.32639781e-01
-2.92193890e-02 4.00840044e-01 -2.98750132e-01 5.23013948e-03
1.02257788e+00 2.87820071e-01 -2.78261244e-01 -2.07331955e-01
-5.98982871e-01 2.17226505e-01 3.51634443e-01 1.50499985e-01
3.18178415e-01 -8.60951185e-01 -7.44749904e-01 -1.13001084e-02
-4.12629247e-02 2.59653062e-01 1.84643611e-01 9.36015427e-01
8.95770788e-02 5.76944351e-01 -1.27956858e-02 -5.75227499e-01
-5.97064793e-01 6.12068355e-01 1.16681278e-01 -4.27386135e-01
-2.16049656e-01 6.59805298e-01 1.69047013e-01 -2.09567353e-01
7.57447183e-02 -3.17707658e-01 1.28686160e-01 4.94148135e-02
6.38131022e-01 3.11970204e-01 -9.24397856e-02 1.37061805e-01
-2.13949665e-01 1.16073802e-01 -9.97714885e-03 -1.86468050e-01
1.64304078e+00 1.96988314e-01 -6.97733909e-02 5.70002675e-01
7.39516079e-01 -5.87548278e-02 -1.37109792e+00 5.91626540e-02
3.57564121e-01 -3.39000791e-01 3.06062490e-01 -1.13677466e+00
-5.05802035e-01 8.70991826e-01 4.23343182e-01 3.96987461e-02
6.83491290e-01 2.70360768e-01 3.36283237e-01 3.86019140e-01
5.06107688e-01 -9.32151794e-01 -4.51258980e-02 -8.09195861e-02
6.26267970e-01 -1.25914514e+00 2.84557372e-01 2.04717919e-01
-7.97206759e-02 8.60863030e-01 -1.79034859e-01 2.29145452e-01
5.67414343e-01 2.04343274e-01 -6.81364477e-01 -3.10482115e-01
-8.55945051e-01 -8.77366886e-02 1.27179578e-01 5.42726815e-01
3.17308247e-01 9.04886425e-02 -3.27879816e-01 6.28147900e-01
1.89931482e-01 3.10790271e-01 3.82697582e-01 6.86007082e-01
-6.93825960e-01 -1.61826360e+00 -4.66181010e-01 6.51031613e-01
-5.01837075e-01 -5.15058963e-03 -1.26197428e-01 5.78469574e-01
4.20753956e-02 6.86343431e-01 -4.31020886e-01 1.12659810e-02
-1.96853131e-01 1.77680463e-01 5.29542863e-01 -4.69172925e-01
1.09645873e-02 1.44158527e-01 2.49705557e-02 -6.04618609e-01
-3.76234353e-01 -7.78521121e-01 -9.50830877e-01 -1.84510991e-01
-4.78590786e-01 4.30175245e-01 1.12620389e+00 1.17768371e+00
5.11657834e-01 2.98102051e-01 3.39688450e-01 -9.07445014e-01
-9.41835165e-01 -8.09266448e-01 -5.91512501e-01 -6.61957636e-02
1.45877868e-01 -9.34594214e-01 -4.17285085e-01 -3.01879287e-01] | [5.983887672424316, 4.57805061340332] |
d5df277c-4e0d-403e-8df1-51a3e7e462b7 | part-aware-contrastive-learning-for-self | 2305.00666 | null | https://arxiv.org/abs/2305.00666v2 | https://arxiv.org/pdf/2305.00666v2.pdf | Part Aware Contrastive Learning for Self-Supervised Action Recognition | In recent years, remarkable results have been achieved in self-supervised action recognition using skeleton sequences with contrastive learning. It has been observed that the semantic distinction of human action features is often represented by local body parts, such as legs or hands, which are advantageous for skeleton-based action recognition. This paper proposes an attention-based contrastive learning framework for skeleton representation learning, called SkeAttnCLR, which integrates local similarity and global features for skeleton-based action representations. To achieve this, a multi-head attention mask module is employed to learn the soft attention mask features from the skeletons, suppressing non-salient local features while accentuating local salient features, thereby bringing similar local features closer in the feature space. Additionally, ample contrastive pairs are generated by expanding contrastive pairs based on salient and non-salient features with global features, which guide the network to learn the semantic representations of the entire skeleton. Therefore, with the attention mask mechanism, SkeAttnCLR learns local features under different data augmentation views. The experiment results demonstrate that the inclusion of local feature similarity significantly enhances skeleton-based action representation. Our proposed SkeAttnCLR outperforms state-of-the-art methods on NTURGB+D, NTU120-RGB+D, and PKU-MMD datasets. | ['Shiqian Wu', 'Chen Chen', 'Mengyuan Liu', 'Aidong Lu', 'Ce Zheng', 'Wenhan Wu', 'Yilei Hua'] | 2023-05-01 | null | null | null | null | ['skeleton-based-action-recognition', 'action-recognition-in-videos', 'self-supervised-action-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 4.86718833e-01 1.63264945e-02 -6.63109958e-01 -3.57287049e-01
-4.40704972e-01 4.31758225e-01 5.49049973e-01 -3.34912091e-01
-3.72597992e-01 5.00812352e-01 9.58989799e-01 6.81044877e-01
-1.62777096e-01 -5.26006401e-01 -4.88077611e-01 -8.92287850e-01
5.63478321e-02 2.28395671e-01 3.79729837e-01 -4.34895426e-01
2.19008386e-01 4.45188612e-01 -1.85520577e+00 5.82097411e-01
5.42894959e-01 1.15153849e+00 1.37694657e-01 5.01537472e-02
-1.39128536e-01 9.27946270e-01 -4.35177267e-01 1.86103016e-01
3.50629151e-01 -7.93214023e-01 -8.06324482e-01 3.23828012e-01
5.87365568e-01 -3.71252477e-01 -4.95335788e-01 7.37768710e-01
6.21761203e-01 3.01228046e-01 6.46092653e-01 -1.22865295e+00
-7.23555803e-01 3.02764446e-01 -1.12869453e+00 4.90822434e-01
6.57453418e-01 4.86295462e-01 1.02758443e+00 -8.22243094e-01
4.63029683e-01 1.62134922e+00 2.64624268e-01 5.76245070e-01
-8.80314410e-01 -6.86370850e-01 3.60663295e-01 6.43130362e-01
-1.18522358e+00 -2.72880256e-01 1.17274368e+00 -2.01083958e-01
9.01725233e-01 1.17537715e-01 9.77265120e-01 1.19351089e+00
4.51226532e-01 1.30090308e+00 9.37129140e-01 -3.81433845e-01
-1.19187608e-01 -6.41854286e-01 -1.32459164e-01 8.29261005e-01
-2.00705558e-01 1.03356741e-01 -9.30150092e-01 2.09858999e-01
1.16120994e+00 5.50341249e-01 -1.82615504e-01 -5.75953066e-01
-1.23773479e+00 6.32725894e-01 6.77044451e-01 4.49961782e-01
-6.05871379e-01 2.53615677e-01 6.06827796e-01 3.92569527e-02
3.50971758e-01 -8.50734711e-02 -5.26958764e-01 -1.35490075e-01
-3.39052528e-01 -2.27502771e-02 -2.05448255e-01 6.46155179e-01
7.67905951e-01 1.74698740e-01 -4.14512604e-01 8.89761031e-01
4.09069985e-01 3.85785073e-01 1.22825229e+00 -9.54696953e-01
5.11139333e-01 1.01339197e+00 -4.75981563e-01 -9.68765914e-01
-4.87704545e-01 -2.47595608e-01 -8.18665385e-01 3.84238988e-01
2.07596034e-01 3.43893528e-01 -1.09738779e+00 1.70297670e+00
4.32409734e-01 2.56210193e-02 -5.15578352e-02 1.18457937e+00
8.83688450e-01 2.38622561e-01 4.54373777e-01 -1.97213124e-02
1.33809340e+00 -1.21745360e+00 -6.51734352e-01 -3.75822067e-01
5.99090695e-01 -3.34168971e-01 1.35508680e+00 6.84777871e-02
-9.08258379e-01 -1.00965428e+00 -8.80601227e-01 -2.90950775e-01
-1.26500472e-01 1.08419858e-01 6.80122316e-01 6.01252727e-02
-4.49958265e-01 5.86421192e-01 -7.77712762e-01 -4.46757823e-01
8.66504610e-01 2.14015663e-01 -6.06197536e-01 -8.74830112e-02
-1.17191911e+00 7.44512200e-01 4.41882998e-01 -3.98843475e-02
-9.11758363e-01 -2.30203003e-01 -1.16112506e+00 -2.13377792e-02
4.95862544e-01 -5.69957435e-01 8.26363802e-01 -1.25332236e+00
-1.52444029e+00 9.34397519e-01 1.56436846e-01 -1.52109504e-01
3.82292032e-01 -3.93769175e-01 -3.31432730e-01 5.58323801e-01
3.72825235e-01 8.50837171e-01 1.10484517e+00 -7.82932818e-01
-5.58130443e-01 -7.20651925e-01 -4.42882404e-02 6.48313224e-01
-4.03933436e-01 4.32434753e-02 -1.43374756e-01 -1.11889625e+00
3.90302509e-01 -5.15421152e-01 -2.42486760e-01 4.44961905e-01
-2.15704385e-02 -4.49610829e-01 1.07495236e+00 -5.44291079e-01
8.67400885e-01 -2.25907731e+00 4.15020525e-01 -6.84363022e-02
-2.03561317e-02 2.55984545e-01 -4.20965463e-01 1.41832516e-01
-5.06686687e-01 -3.72894228e-01 -3.41541499e-01 4.68973927e-02
-3.58019024e-01 5.91818035e-01 3.33097614e-02 4.89680588e-01
4.16743845e-01 1.09502196e+00 -1.10242486e+00 -8.66761565e-01
6.14768207e-01 2.74843097e-01 -4.25007522e-01 9.38901864e-03
-1.57398004e-02 7.76002347e-01 -7.94076025e-01 8.65238667e-01
2.33775645e-01 9.12657753e-03 -1.86280891e-01 -2.09548891e-01
2.77608991e-01 7.14954883e-02 -1.11612904e+00 2.24896812e+00
-3.26892167e-01 1.23524904e-01 -5.37246764e-02 -1.31536949e+00
9.85694110e-01 7.11768568e-02 8.86043131e-01 -1.04512048e+00
2.66800046e-01 -6.05500378e-02 -9.18203965e-02 -6.54696345e-01
-4.11165357e-02 -3.48915160e-01 5.72493374e-02 5.01868069e-01
1.95925295e-01 1.66627765e-01 2.51377914e-02 -9.93351266e-02
9.06236649e-01 7.05879152e-01 6.31810188e-01 -8.05691555e-02
8.87629747e-01 -4.28666323e-01 8.16427410e-01 1.84185207e-01
-5.42455733e-01 6.50986254e-01 8.53877366e-02 -5.87364733e-01
-3.95284921e-01 -1.11607194e+00 3.97092365e-02 1.44691825e+00
3.10761243e-01 -4.96988147e-01 -3.95537823e-01 -1.16091168e+00
2.68913247e-02 2.77288467e-01 -8.59955668e-01 -7.43874431e-01
-7.40057290e-01 -2.67716527e-01 2.01393902e-01 1.17350423e+00
8.72786939e-01 -1.59695613e+00 -9.27763164e-01 1.13983676e-01
-1.91143543e-01 -5.89829206e-01 -5.41212440e-01 1.51773080e-01
-1.04710543e+00 -1.15850139e+00 -9.63671267e-01 -7.73623049e-01
7.18179882e-01 4.46287721e-01 7.97477365e-01 1.29273281e-01
-6.59386277e-01 5.61056376e-01 -5.76363087e-01 -1.38916686e-01
7.64898285e-02 -3.41907322e-01 3.99732850e-02 2.33253270e-01
4.95792687e-01 -7.55617321e-01 -7.26590276e-01 5.05559623e-01
-9.58423436e-01 6.79856688e-02 9.71873939e-01 9.94346499e-01
7.71254480e-01 -2.97043741e-01 5.26504517e-01 -3.08460087e-01
5.41339964e-02 -3.70643437e-01 3.70079637e-01 5.75091727e-02
-6.88013881e-02 1.77614242e-01 5.78685582e-01 -5.41362584e-01
-1.21190178e+00 2.32774660e-01 -5.73924035e-02 -5.61841428e-01
-3.90964955e-01 2.04757422e-01 -5.59662580e-01 4.80789691e-02
5.90803504e-01 4.69436049e-01 3.40434849e-01 -5.14569521e-01
4.52278435e-01 5.50099969e-01 4.86073256e-01 -7.03306258e-01
4.86111104e-01 7.30999172e-01 7.32205436e-02 -7.70576060e-01
-9.93431866e-01 -5.72564423e-01 -1.00541568e+00 -3.62921476e-01
1.06412280e+00 -8.73290896e-01 -2.02698350e-01 6.79112971e-01
-6.27632499e-01 -1.94503695e-01 -1.00070643e+00 5.82171381e-01
-1.03111327e+00 7.96890497e-01 -5.22492826e-01 -4.41812247e-01
-1.89904809e-01 -1.01699471e+00 1.44653082e+00 2.76603281e-01
-3.78857702e-01 -5.95887303e-01 2.34592855e-02 6.80878818e-01
-5.44314156e-04 3.66886109e-01 7.69272149e-01 -4.23094064e-01
-2.12865964e-01 7.16256094e-04 -9.98036712e-02 5.54249585e-01
6.04337990e-01 -4.44334239e-01 -6.84483349e-01 -1.19812667e-01
2.49669235e-02 -7.76180804e-01 1.06946290e+00 3.09667617e-01
1.43165898e+00 -8.76822621e-02 -2.87072450e-01 4.35380727e-01
8.52148354e-01 2.54275110e-02 8.26457083e-01 1.63193092e-01
8.36738884e-01 5.80363870e-01 9.88069832e-01 5.54137051e-01
3.41044716e-03 6.33911133e-01 5.50261676e-01 -3.25851828e-01
-4.47377473e-01 -2.87679672e-01 4.37833071e-01 4.04148757e-01
-5.60234308e-01 4.92040485e-01 -4.84566629e-01 3.75482410e-01
-1.91498125e+00 -1.02302611e+00 2.02890202e-01 1.83927929e+00
9.13199127e-01 1.11564808e-01 2.50541270e-01 2.40561530e-01
7.23240137e-01 6.68731928e-01 -7.17564523e-01 1.11929923e-02
-1.01091690e-01 4.08069730e-01 2.03803461e-02 -1.58475995e-01
-1.21685505e+00 8.23479354e-01 5.42585230e+00 1.06510270e+00
-8.51056218e-01 3.40891108e-02 3.47061336e-01 -1.87467575e-01
1.55540198e-01 -2.76338249e-01 -2.73873389e-01 4.67994392e-01
1.57263473e-01 2.24486813e-02 -2.31565475e-01 1.06074369e+00
2.29413167e-01 -1.80969521e-01 -1.00770915e+00 1.11814106e+00
4.07646954e-01 -7.49646902e-01 3.51564288e-01 -7.16863126e-02
5.58595121e-01 -3.49343985e-01 -3.47914435e-02 3.87929857e-01
1.39233783e-01 -9.10464108e-01 4.68687624e-01 6.65803611e-01
6.04979753e-01 -7.17321098e-01 5.35341084e-01 1.73025325e-01
-1.45581281e+00 -4.35433507e-01 -4.15025830e-01 -3.68753046e-01
1.06381811e-01 9.02571380e-02 -6.65315911e-02 6.56972587e-01
8.21369231e-01 1.47743607e+00 -5.71465373e-01 7.02280700e-01
-5.63950181e-01 1.89239055e-01 -6.22639470e-02 2.99166262e-01
2.63105005e-01 -1.33617759e-01 4.21788722e-01 8.20113719e-01
-5.26147410e-02 2.99209952e-01 4.45252568e-01 6.16598487e-01
2.35357672e-01 2.21330568e-01 -4.56780344e-01 2.48278722e-01
-3.13938618e-01 1.04229295e+00 -5.87554514e-01 -3.73564094e-01
-4.52670902e-01 1.17537534e+00 2.51393080e-01 1.16427951e-01
-8.57746959e-01 -1.18652470e-01 7.86718905e-01 1.89566582e-01
2.85927951e-01 -1.88989937e-02 7.61184394e-02 -1.00687981e+00
4.31232229e-02 -9.20407832e-01 8.29325736e-01 -8.47866118e-01
-1.17938077e+00 -1.28846681e-02 1.57847181e-01 -1.60664809e+00
-7.89782405e-02 -5.00932932e-01 -7.11715102e-01 5.43000698e-01
-1.28131998e+00 -1.34676087e+00 -4.55097288e-01 8.50590110e-01
9.36698496e-01 -1.68283418e-01 6.04048610e-01 6.56360313e-02
-4.36638147e-01 3.74946564e-01 -3.50968212e-01 2.21846104e-01
7.31782258e-01 -9.55074787e-01 -5.55506237e-02 5.26306927e-01
1.55216560e-01 4.24736053e-01 2.00835571e-01 -7.88452625e-01
-9.46698010e-01 -9.90126491e-01 3.25391710e-01 -2.32229829e-01
3.31377327e-01 2.03049898e-01 -8.52017224e-01 5.35080254e-01
-1.29626393e-01 4.54650521e-01 6.86947107e-01 -1.74674287e-01
-3.51699978e-01 -1.58426106e-01 -9.62512255e-01 5.90166569e-01
1.71519208e+00 -3.61112833e-01 -1.09858346e+00 3.79949749e-01
3.44315439e-01 -8.85112807e-02 -7.99572051e-01 6.21675074e-01
6.73657894e-01 -1.07359684e+00 1.11819720e+00 -9.66646314e-01
6.96079671e-01 -3.62801313e-01 -8.11085999e-02 -1.05469406e+00
-5.24470687e-01 -7.61846546e-03 -2.45313063e-01 9.67618346e-01
-3.98611486e-01 -3.27665359e-01 9.17473733e-01 -1.57057513e-02
-3.21355432e-01 -9.39561486e-01 -1.14755881e+00 -8.09848249e-01
-7.15201497e-02 -1.12894848e-01 3.43771487e-01 7.77137518e-01
7.18153045e-02 2.97771722e-01 -3.33431125e-01 -5.20655394e-01
3.95137370e-01 2.42524415e-01 7.54933774e-01 -1.17197192e+00
-2.21477792e-01 -4.67369556e-01 -9.96052444e-01 -1.22186041e+00
3.58622581e-01 -8.99694264e-01 -1.30743474e-01 -1.52595186e+00
4.41625863e-01 1.36030942e-01 -6.56626105e-01 9.60446060e-01
-3.21982652e-01 4.22219515e-01 1.55997410e-01 1.24181785e-01
-8.21255744e-01 1.23871267e+00 1.80308926e+00 -3.77493620e-01
-3.69238225e-03 -1.30163252e-01 -5.75904548e-01 1.06116641e+00
6.31636381e-01 -1.98370442e-01 -4.14771229e-01 -1.20643511e-01
-5.17033696e-01 -2.44485304e-01 5.22283733e-01 -1.28488743e+00
-1.84293270e-01 -3.47527802e-01 8.94537985e-01 -6.53972924e-01
4.74291921e-01 -8.62654090e-01 -3.25806290e-01 8.29398215e-01
-4.47729141e-01 -3.77605677e-01 -1.75687402e-01 6.81885183e-01
-3.19400698e-01 1.03671476e-01 8.06944549e-01 -4.97571409e-01
-1.24750316e+00 3.80192548e-01 -2.95534462e-01 1.98011592e-01
1.14858031e+00 -7.02876210e-01 4.11976501e-02 -1.70170829e-01
-9.44890201e-01 2.18257084e-01 2.73063153e-01 6.37187243e-01
9.03378785e-01 -1.74297142e+00 -6.64724648e-01 5.55014968e-01
4.42031801e-01 5.37319705e-02 5.76959729e-01 9.82212424e-01
-9.94855762e-02 1.28337905e-01 -9.85032439e-01 -7.29544103e-01
-1.38123953e+00 5.50645828e-01 2.38986671e-01 -1.87426522e-01
-9.42264438e-01 8.19621265e-01 6.27262592e-01 -1.72206566e-01
1.53238013e-01 -3.15964848e-01 -4.61755544e-01 5.76680824e-02
6.03059113e-01 6.28534377e-01 -3.95039290e-01 -1.00594652e+00
-4.77473170e-01 1.05678058e+00 -3.59230191e-02 2.97050565e-01
1.27628648e+00 -4.82093431e-02 1.27348438e-01 2.26930723e-01
1.16382360e+00 -3.29232067e-01 -1.56403053e+00 -5.01042128e-01
-3.13683808e-01 -8.08408260e-01 -2.17318401e-01 -6.28768623e-01
-1.40754402e+00 9.75045919e-01 7.71336913e-01 -5.42291462e-01
1.39201581e+00 3.74980152e-01 8.37437928e-01 1.30752072e-01
5.00167549e-01 -1.39685643e+00 9.16048527e-01 2.27368250e-01
1.25813043e+00 -1.25595391e+00 2.21030474e-01 -3.41348290e-01
-8.11015964e-01 9.82442737e-01 1.02791536e+00 -2.31173590e-01
4.24769759e-01 -2.43936539e-01 -5.44334203e-03 -3.38455677e-01
-1.89511836e-01 -5.36928177e-01 3.61222327e-01 8.05353940e-01
2.21965432e-01 -1.05622046e-01 -6.16456091e-01 7.49300420e-01
2.18306065e-01 -1.35233715e-01 3.95869128e-02 1.31544006e+00
-6.19625390e-01 -9.27720129e-01 -3.43226016e-01 4.54366863e-01
-2.51794606e-01 2.37256259e-01 -5.20165801e-01 8.83534968e-01
3.94393742e-01 5.61081290e-01 8.75446722e-02 -3.71798724e-01
4.01649177e-01 9.70646217e-02 6.64552391e-01 -7.95749664e-01
-2.87830323e-01 1.55592620e-01 -1.82603225e-01 -1.07308912e+00
-8.70819509e-01 -8.11666250e-01 -1.65118992e+00 5.01566470e-01
-1.21178254e-01 -2.29474440e-01 -2.46912129e-02 1.14798057e+00
2.02542171e-01 7.02685475e-01 5.03211677e-01 -1.12553442e+00
-6.96554065e-01 -1.01382840e+00 -8.24591935e-01 7.66143620e-01
6.73041120e-02 -1.30538034e+00 -2.44421259e-01 3.71687822e-02] | [8.060609817504883, 0.6095632314682007] |
f98998a0-e20e-4cfe-b05c-76ddffa3f973 | graph-signal-sampling-for-inductive-one-bit | 2302.03933 | null | https://arxiv.org/abs/2302.03933v1 | https://arxiv.org/pdf/2302.03933v1.pdf | Graph Signal Sampling for Inductive One-Bit Matrix Completion: a Closed-form Solution | Inductive one-bit matrix completion is motivated by modern applications such as recommender systems, where new users would appear at test stage with the ratings consisting of only ones and no zeros. We propose a unified graph signal sampling framework which enjoys the benefits of graph signal analysis and processing. The key idea is to transform each user's ratings on the items to a function (signal) on the vertices of an item-item graph, then learn structural graph properties to recover the function from its values on certain vertices -- the problem of graph signal sampling. We propose a class of regularization functionals that takes into account discrete random label noise in the graph vertex domain, then develop the GS-IMC approach which biases the reconstruction towards functions that vary little between adjacent vertices for noise reduction. Theoretical result shows that accurate reconstructions can be achieved under mild conditions. For the online setting, we develop a Bayesian extension, i.e., BGS-IMC which considers continuous random Gaussian noise in the graph Fourier domain and builds upon a prediction-correction update algorithm to obtain the unbiased and minimum-variance reconstruction. Both GS-IMC and BGS-IMC have closed-form solutions and thus are highly scalable in large data. Experiments show that our methods achieve state-of-the-art performance on public benchmarks. | ['Junchi Yan', 'Xiaokang Yang', 'Hua Chai', 'Zhaobing Han', 'Gang Zeng', 'Haoyu Geng', 'Chao Chen'] | 2023-02-08 | null | null | null | null | ['matrix-completion'] | ['methodology'] | [ 3.07239115e-01 5.63089885e-02 -7.04401433e-02 -2.79316902e-01
-8.68160367e-01 -4.82166141e-01 1.05858840e-01 -1.47157116e-02
7.92451203e-02 4.09055471e-01 3.23997080e-01 -1.62641019e-01
-2.98365623e-01 -7.37295151e-01 -1.06716454e+00 -7.04192162e-01
-3.68979007e-01 2.28139862e-01 -1.08422183e-01 -2.16792017e-01
-1.28704906e-01 1.62525137e-03 -8.28184962e-01 6.49546683e-02
7.02390969e-01 8.85888159e-01 -2.48058997e-02 7.06860900e-01
3.36483210e-01 5.26711643e-01 -1.97584197e-01 -7.58144557e-01
4.89455312e-01 -4.11044925e-01 -3.49372655e-01 3.91605854e-01
3.44739139e-01 -2.56974027e-02 -6.06698871e-01 1.61412156e+00
4.73203421e-01 2.87078738e-01 7.14934707e-01 -1.06278014e+00
-8.31417322e-01 8.55749667e-01 -8.90145898e-01 -1.71402737e-01
5.35665333e-01 -3.00131917e-01 1.50906134e+00 -1.12322998e+00
4.77939129e-01 1.16044259e+00 1.10949683e+00 1.24020912e-01
-1.84636486e+00 -3.98385018e-01 3.08603853e-01 1.07324682e-01
-1.61981714e+00 -3.15117121e-01 9.86827791e-01 -3.73473972e-01
2.39395127e-01 3.11700523e-01 6.07340276e-01 1.21170223e+00
-3.96613963e-02 6.76774383e-01 9.41814661e-01 -2.42847174e-01
3.57381374e-01 -4.27038502e-03 2.32086867e-01 7.66992033e-01
6.39980793e-01 -3.16570044e-01 -6.01360261e-01 -5.30775845e-01
7.14550436e-01 1.05309434e-01 -6.59577668e-01 -5.71401298e-01
-1.11947107e+00 1.03984046e+00 2.83269376e-01 -9.81153026e-02
-6.49857759e-01 4.13384050e-01 1.76812574e-01 6.46193683e-01
6.33699596e-01 -7.31487647e-02 -1.78827226e-01 2.93317139e-01
-7.78602839e-01 9.90166068e-02 1.03081763e+00 1.15002739e+00
8.51485550e-01 9.90610644e-02 -2.70848602e-01 8.86712968e-01
6.44186378e-01 7.02830255e-01 2.71163322e-02 -7.77963400e-01
4.41210598e-01 -3.28991525e-02 -2.72014011e-02 -1.44308174e+00
-2.18713358e-01 -9.71475422e-01 -1.11763930e+00 -4.63191211e-01
2.62911081e-01 -2.35802501e-01 -4.94106621e-01 1.80053198e+00
4.26365316e-01 8.18354607e-01 -3.68984491e-01 8.98670852e-01
5.94694257e-01 5.39087892e-01 -4.71362740e-01 -5.37435651e-01
1.20024085e+00 -4.17781860e-01 -7.26016223e-01 -1.45765349e-01
4.78111714e-01 -5.04319072e-01 1.14734733e+00 8.39844406e-01
-9.01662946e-01 -3.55786383e-01 -1.12814617e+00 2.96642333e-01
3.01207393e-01 1.23822607e-01 5.56549549e-01 9.56062675e-01
-1.01679647e+00 8.02678525e-01 -5.86952686e-01 1.35520950e-01
1.60159856e-01 2.11511612e-01 -3.50420177e-01 -4.11253721e-01
-1.07404757e+00 1.28796622e-01 -3.18379462e-01 2.09965438e-01
-7.55804062e-01 -7.35080838e-01 -6.90304160e-01 -1.40095562e-01
6.19851053e-01 -6.31922066e-01 9.48931396e-01 -8.28719914e-01
-1.42196536e+00 3.23774040e-01 3.20943408e-02 -4.81397957e-01
1.89714909e-01 4.35881726e-02 -5.13334215e-01 1.18794501e-01
8.01275298e-02 -3.08483869e-01 1.43957412e+00 -1.14655375e+00
-1.28891654e-02 -3.87565643e-01 -1.71154052e-01 7.57086203e-02
-2.51120985e-01 -5.53182960e-01 -7.82926500e-01 -9.18925941e-01
5.39424419e-01 -1.11264396e+00 -3.85493875e-01 -2.50897437e-01
-5.30853868e-01 1.23227052e-01 2.31317490e-01 -7.72071779e-01
1.25968397e+00 -2.19071388e+00 2.00296253e-01 8.95515382e-01
5.20889819e-01 -2.58143336e-01 -2.50569284e-01 6.27378702e-01
-1.09290220e-01 -1.38362721e-01 -2.41920701e-03 -4.97827321e-01
1.03938349e-01 1.42973781e-01 -2.38332018e-01 1.08098972e+00
-2.54368007e-01 6.75288618e-01 -1.03853214e+00 3.98563407e-02
-1.95027888e-01 5.38223684e-01 -1.07935154e+00 -4.41839313e-03
5.85054010e-02 2.40438432e-01 -4.32405174e-01 3.30277890e-01
8.74868512e-01 -6.88039482e-01 5.57186127e-01 -5.10436118e-01
5.63431978e-01 2.31309414e-01 -1.78325617e+00 1.73892713e+00
-2.84415841e-01 2.51060724e-01 6.64829731e-01 -1.34727156e+00
6.81734920e-01 2.66170502e-01 5.52105308e-01 -5.84210195e-02
3.63895786e-03 8.16495344e-02 -1.37944788e-01 -1.79642737e-01
1.80547729e-01 -3.59348059e-02 -6.07917607e-02 3.72367531e-01
2.36575350e-01 -8.05919617e-03 8.10393691e-02 6.98978782e-01
1.41695476e+00 -3.11281145e-01 2.39356890e-01 -1.78547204e-01
3.43249083e-01 -6.51286900e-01 5.05304039e-01 9.63027179e-01
2.38788441e-01 6.50599062e-01 6.93447232e-01 2.48339117e-01
-5.94881952e-01 -1.11451125e+00 1.29521685e-02 1.07675076e+00
2.63418201e-02 -7.92536736e-01 -7.13736892e-01 -6.24073744e-01
1.72352299e-01 5.60180068e-01 -5.88327944e-01 -1.62557036e-01
1.09046856e-02 -8.28369737e-01 -3.60365026e-02 3.19770202e-02
3.00385710e-02 -2.04032630e-01 5.28591514e-01 3.87267709e-01
-2.18982890e-01 -1.12503803e+00 -1.05119133e+00 1.85869150e-02
-6.32592022e-01 -8.93648207e-01 -5.80413342e-01 -6.34006500e-01
7.03666449e-01 6.98512673e-01 1.05745983e+00 -1.04844086e-01
9.27755423e-03 1.01324725e+00 -4.19119447e-01 1.01850905e-01
-1.04310736e-01 -4.00772393e-01 1.36649743e-01 9.02989447e-01
-8.82096738e-02 -1.00385177e+00 -6.18491292e-01 2.27286622e-01
-6.63368821e-01 -5.83559871e-02 4.51428324e-01 9.07155216e-01
9.28922951e-01 6.30621240e-02 6.13121033e-01 -1.46394026e+00
9.61402833e-01 -8.25833559e-01 -4.60970640e-01 1.15558162e-01
-8.81587267e-01 8.51047337e-02 6.57745481e-01 -6.67926550e-01
-3.85862917e-01 1.08995244e-01 1.31499231e-01 -6.55764997e-01
6.30889654e-01 8.68953228e-01 -1.56054601e-01 -4.05913830e-01
7.38389671e-01 1.34574622e-01 3.33816782e-02 -5.56029141e-01
7.41596758e-01 4.78578925e-01 5.51689744e-01 -5.74131846e-01
9.18602049e-01 3.06411445e-01 1.73627198e-01 -1.00145006e+00
-8.84068727e-01 -6.33440495e-01 -1.12553656e-01 -3.18959326e-01
3.37395370e-01 -1.25175393e+00 -8.08242619e-01 8.22965726e-02
-7.22940207e-01 -9.56589282e-02 -3.98306310e-01 6.50138676e-01
-3.70815158e-01 7.78951883e-01 -7.55917430e-01 -8.99948597e-01
-2.57797807e-01 -8.20322990e-01 9.07772660e-01 -2.53658623e-01
1.16529837e-01 -1.08054590e+00 1.71562776e-01 4.19846363e-02
1.22845337e-01 3.34684402e-02 5.78351021e-01 -6.21886849e-01
-5.41822314e-01 -5.28212667e-01 -8.41100886e-02 5.81331313e-01
-1.04197539e-01 -4.29766864e-01 -6.07933819e-01 -6.55421317e-01
2.20416427e-01 1.69503782e-02 6.90015316e-01 6.72214389e-01
1.07139170e+00 -4.29548293e-01 -1.49010807e-01 6.86715841e-01
1.43281412e+00 -5.46781003e-01 3.28509510e-01 -6.20081127e-01
9.54830885e-01 1.71185046e-01 2.50719219e-01 7.83242702e-01
2.65234709e-01 6.81373179e-01 3.65062118e-01 2.98407137e-01
-1.10435456e-01 -5.18175364e-01 5.99257231e-01 1.51517892e+00
7.76917115e-02 -3.25679809e-01 -3.00241202e-01 3.29668432e-01
-1.95562601e+00 -6.95059180e-01 -6.68757677e-01 2.59710026e+00
7.32959688e-01 1.80163439e-02 1.33701250e-01 9.94142070e-02
8.29385579e-01 2.10427642e-01 -4.42488104e-01 1.57674581e-01
2.20845453e-02 3.58356655e-01 9.63970602e-01 6.74614251e-01
-8.77057791e-01 2.27762282e-01 6.00615168e+00 8.68725002e-01
-5.56754351e-01 3.36294532e-01 3.24980050e-01 1.99537817e-02
-5.63739598e-01 4.26511876e-02 -5.21080673e-01 5.63475192e-01
1.06898046e+00 -2.44736701e-01 1.12344432e+00 7.04945147e-01
2.07694829e-01 3.45482111e-01 -1.03099859e+00 1.21095514e+00
1.78968549e-01 -1.10480988e+00 -2.68731892e-01 2.58646637e-01
8.90690446e-01 5.04725985e-02 2.81107090e-02 2.48220757e-01
5.18506289e-01 -7.24694252e-01 4.85782355e-01 6.88394785e-01
9.36809778e-01 -5.11547744e-01 3.12586576e-01 1.98388964e-01
-1.36717379e+00 1.30131915e-01 -6.90133572e-01 1.33437291e-01
7.13097826e-02 1.24455416e+00 -7.68511832e-01 7.52183497e-01
3.30461621e-01 1.06963944e+00 -1.53495461e-01 1.04218423e+00
-3.01971614e-01 1.20766556e+00 -4.51984137e-01 -8.11517388e-02
-1.54166669e-01 -8.81488204e-01 7.59086013e-01 9.76203024e-01
5.40176451e-01 6.23039417e-02 5.09646952e-01 5.94367862e-01
-4.60593283e-01 4.78718281e-01 -5.90625465e-01 8.12369958e-02
4.74072546e-01 1.37562406e+00 -4.72255051e-01 -1.33054182e-01
-8.08152854e-01 1.00497055e+00 2.36557245e-01 6.34410083e-01
-7.27926493e-01 -2.80013412e-01 4.65928704e-01 3.06071579e-01
5.71082771e-01 -1.44724473e-01 -5.02255447e-02 -1.27391994e+00
1.37107491e-01 -1.12148881e+00 3.67877275e-01 -4.51999724e-01
-1.63739681e+00 8.58866796e-02 -2.33203530e-01 -1.35554767e+00
-1.25076652e-01 -5.41788757e-01 -5.18885911e-01 7.45346487e-01
-9.07010794e-01 -7.49775887e-01 -1.37928024e-01 8.85342121e-01
7.74477422e-02 -3.59062143e-02 6.25916481e-01 5.79290211e-01
-4.77068484e-01 6.96652472e-01 4.97544259e-01 1.36873021e-03
5.41663170e-01 -1.23871315e+00 4.33857709e-01 9.37405109e-01
6.93758726e-01 6.20212972e-01 7.45810986e-01 -7.37055004e-01
-2.09488988e+00 -1.13514876e+00 2.50621378e-01 -1.12762608e-01
9.38556731e-01 -7.08745241e-01 -7.80256391e-01 7.22211957e-01
-2.26431027e-01 5.67545831e-01 6.78693295e-01 3.98607492e-01
-4.69838828e-01 -2.09740236e-01 -8.99658918e-01 5.43465734e-01
1.29973459e+00 -8.16758752e-01 -1.47570921e-02 8.36609244e-01
5.21008074e-01 -3.00587595e-01 -8.40553284e-01 7.84060266e-03
2.71496207e-01 -7.59492993e-01 1.06790245e+00 -6.20582759e-01
-2.27814615e-01 -3.28805149e-01 -5.52641094e-01 -1.57927489e+00
-7.54877567e-01 -1.35561883e+00 -3.48214269e-01 9.54940975e-01
4.94322926e-01 -7.95746684e-01 8.38996232e-01 2.78472036e-01
-8.46723989e-02 -5.91275752e-01 -7.47883499e-01 -6.65892065e-01
-5.60326159e-01 -7.81124413e-01 1.37532875e-01 7.89167404e-01
3.36104222e-02 7.59361088e-01 -8.79517198e-01 5.34620345e-01
1.11925900e+00 8.28979984e-02 7.67145872e-01 -1.34995258e+00
-1.10631347e+00 9.01544243e-02 -4.13614661e-01 -1.47408676e+00
2.95043048e-02 -1.22158778e+00 -8.43425223e-04 -1.28848994e+00
7.39783272e-02 -1.81805313e-01 -1.99391589e-01 -7.66417235e-02
-1.68216199e-01 3.32596540e-01 -4.38525192e-02 1.23886243e-02
-6.92358196e-01 4.38821644e-01 1.13255215e+00 -5.62688708e-03
-3.46321873e-02 4.42006260e-01 -1.06102657e+00 5.35874486e-01
1.77615181e-01 -5.53992331e-01 -6.87651098e-01 1.03577398e-01
7.44035959e-01 2.81295091e-01 6.25947714e-02 -8.08069825e-01
2.11624339e-01 2.49178633e-01 1.06466576e-01 -3.08896422e-01
4.04487073e-01 -7.87991285e-01 4.76348013e-01 1.58618882e-01
-3.37744355e-01 -3.38730603e-01 -4.66684341e-01 1.45083594e+00
1.96318328e-01 -3.73713791e-01 5.23480773e-01 2.04127327e-01
-2.67039947e-02 5.53343415e-01 -1.78522646e-01 3.20294678e-01
5.35708606e-01 2.48749241e-01 -1.79035086e-02 -1.07186973e+00
-9.00640905e-01 9.47007351e-03 1.13453843e-01 -2.14303225e-01
5.73229432e-01 -1.66116655e+00 -1.00183129e+00 2.85346717e-01
-1.59344777e-01 -4.00126070e-01 2.57821083e-01 1.17418838e+00
-4.10170592e-02 -2.00787947e-01 6.66306257e-01 -4.94479001e-01
-9.90618825e-01 5.31367004e-01 1.56416260e-02 -3.07820827e-01
-6.67389393e-01 1.02449179e+00 2.40281895e-01 -2.92267323e-01
2.82937735e-01 -3.22233945e-01 5.43220490e-02 5.65810874e-02
3.96001250e-01 3.18765283e-01 3.00506800e-01 -6.18773878e-01
-1.01048194e-01 4.22472596e-01 1.41399518e-01 -1.19902305e-01
1.40149939e+00 -2.76586801e-01 6.40232954e-03 5.20160854e-01
1.45413077e+00 7.87212014e-01 -1.17203271e+00 -6.87407136e-01
-2.50012845e-01 -5.00326276e-01 4.28335249e-01 -2.89529711e-01
-1.22431898e+00 3.77030402e-01 3.49725336e-01 7.62659073e-01
1.03200197e+00 -1.60745800e-01 4.78118390e-01 3.77815634e-01
5.33404410e-01 -9.15400386e-01 1.13729220e-02 2.58991420e-01
8.62430334e-01 -7.89961219e-01 3.30231637e-01 -7.84878373e-01
-5.54027259e-01 7.76371956e-01 -3.36200267e-01 -5.74404240e-01
1.08445823e+00 -2.96062771e-02 -5.81403852e-01 -3.12524021e-01
-7.97738671e-01 -1.41595185e-01 7.87034929e-01 5.68354368e-01
3.60822260e-01 4.62710589e-01 -2.00349808e-01 9.68975604e-01
2.91516352e-02 -4.77002054e-01 5.87285459e-01 2.59530127e-01
-2.39573076e-01 -1.01665449e+00 -3.15318584e-01 9.97739732e-01
-4.23892885e-01 -3.20502788e-01 -1.68446437e-01 2.82694191e-01
-4.17647213e-01 1.14347863e+00 -4.19396698e-01 -5.92644393e-01
4.68524367e-01 -4.20963056e-02 4.63482797e-01 -7.76121080e-01
-2.01057985e-01 5.48502088e-01 2.76866257e-01 -7.19671190e-01
-9.84743685e-02 -1.00034356e+00 -8.42699587e-01 -3.69051278e-01
-5.49957752e-01 3.10948014e-01 6.41022861e-01 4.89246100e-01
4.45206404e-01 4.82684106e-01 1.07639444e+00 -8.02457511e-01
-9.40427065e-01 -9.59956408e-01 -1.30665123e+00 5.55551767e-01
2.15418950e-01 -5.05226254e-01 -7.16257453e-01 -1.36593819e-01] | [6.982749938964844, 4.780142784118652] |
0efc69d8-0f12-469f-aafb-78c2e5aa817c | two-stage-mr-image-segmentation-method-for | 2304.08072 | null | https://arxiv.org/abs/2304.08072v1 | https://arxiv.org/pdf/2304.08072v1.pdf | Two-stage MR Image Segmentation Method for Brain Tumors based on Attention Mechanism | Multimodal magnetic resonance imaging (MRI) can reveal different patterns of human tissue and is crucial for clinical diagnosis. However, limited by cost, noise and manual labeling, obtaining diverse and reliable multimodal MR images remains a challenge. For the same lesion, different MRI manifestations have great differences in background information, coarse positioning and fine structure. In order to obtain better generation and segmentation performance, a coordination-spatial attention generation adversarial network (CASP-GAN) based on the cycle-consistent generative adversarial network (CycleGAN) is proposed. The performance of the generator is optimized by introducing the Coordinate Attention (CA) module and the Spatial Attention (SA) module. The two modules can make full use of the captured location information, accurately locating the interested region, and enhancing the generator model network structure. The ability to extract the structure information and the detailed information of the original medical image can help generate the desired image with higher quality. There exist some problems in the original CycleGAN that the training time is long, the parameter amount is too large, and it is difficult to converge. In response to this problem, we introduce the Coordinate Attention (CA) module to replace the Res Block to reduce the number of parameters, and cooperate with the spatial information extraction network above to strengthen the information extraction ability. On the basis of CASP-GAN, an attentional generative cross-modality segmentation (AGCMS) method is further proposed. This method inputs the modalities generated by CASP-GAN and the real modalities into the segmentation network for brain tumor segmentation. Experimental results show that CASP-GAN outperforms CycleGAN and some state-of-the-art methods in PSNR, SSMI and RMSE in most tasks. | ['Jin Li', 'Lin Lu', 'Jiawei Jiang', 'Li Zhu'] | 2023-04-17 | null | null | null | null | ['tumor-segmentation', 'brain-tumor-segmentation'] | ['computer-vision', 'medical'] | [ 3.24153990e-01 1.13790385e-01 1.52075663e-01 -4.02090698e-02
-9.55474555e-01 -3.05951267e-01 1.29107103e-01 -6.09902024e-01
-2.36427084e-01 7.24613249e-01 2.53979951e-01 -1.62412792e-01
1.11815862e-01 -8.80157590e-01 -5.09584069e-01 -1.28393090e+00
3.21775228e-01 2.62713104e-01 3.64014745e-01 -2.01830998e-01
-5.33327125e-02 4.05817688e-01 -7.23952174e-01 2.73596078e-01
1.22794318e+00 8.26956391e-01 7.45356441e-01 2.66515493e-01
-1.22534327e-01 8.00448537e-01 -5.99260747e-01 -1.27496913e-01
8.98095667e-02 -9.53704953e-01 -6.69726431e-01 7.16646686e-02
-3.70929062e-01 -4.32936311e-01 -4.85748529e-01 1.21277678e+00
8.70460927e-01 -2.31910311e-03 5.68669140e-01 -1.08127868e+00
-5.78596592e-01 7.62875319e-01 -7.57178545e-01 2.03488812e-01
-1.46949321e-01 4.47419107e-01 1.23269565e-01 -5.27691424e-01
5.74148536e-01 8.97800684e-01 4.03301388e-01 6.60635769e-01
-8.89983594e-01 -8.58591318e-01 -3.90963163e-03 1.55180484e-01
-1.35433376e+00 1.28429302e-03 1.11571980e+00 -3.33916336e-01
1.25027433e-01 2.92547017e-01 6.32200778e-01 1.07672679e+00
4.15260077e-01 7.56904423e-01 1.05145335e+00 -1.93427518e-01
-4.61580008e-02 -1.87850058e-01 -3.64797652e-01 7.41599381e-01
-6.35318831e-02 9.93883684e-02 1.22111082e-01 2.42364034e-01
1.06724644e+00 2.71598935e-01 -6.18859649e-01 -3.90242189e-02
-1.36773288e+00 6.78525209e-01 1.02372217e+00 7.65148759e-01
-5.86510837e-01 1.31734256e-02 2.55424708e-01 -1.51429564e-01
1.44189790e-01 4.25137669e-01 1.10503219e-01 1.59990236e-01
-9.16170120e-01 -2.02780068e-01 2.27308214e-01 6.82055175e-01
4.37294990e-01 3.36844176e-01 -3.45712483e-01 6.47026896e-01
1.15025923e-01 6.31897867e-01 1.03246903e+00 -7.44796693e-01
4.41278666e-01 6.65466130e-01 -4.50540155e-01 -9.84378994e-01
-3.85559440e-01 -8.49794865e-01 -1.38253319e+00 -4.83626239e-02
1.16935886e-01 -3.76704097e-01 -1.21693385e+00 1.73692036e+00
3.49875182e-01 2.59289980e-01 -7.27849305e-02 1.19120371e+00
1.02515185e+00 7.20592856e-01 8.95675644e-02 -2.88466632e-01
1.15271485e+00 -1.04436898e+00 -7.85418570e-01 -2.51287729e-01
5.12260735e-01 -6.37444854e-01 8.06218803e-01 -3.00593767e-02
-1.08587456e+00 -4.26086694e-01 -1.04999435e+00 2.66888350e-01
-7.15000033e-02 2.40806788e-01 4.68081504e-01 3.82267714e-01
-8.47319722e-01 1.25698611e-01 -8.39358926e-01 2.08903715e-01
6.12468719e-01 3.19884866e-01 -1.97888717e-01 -3.30452979e-01
-1.28860211e+00 6.33145750e-01 5.21575451e-01 3.53777647e-01
-9.96056974e-01 -7.43110180e-01 -7.39443064e-01 -7.10886046e-02
4.06762779e-01 -8.57441902e-01 8.78866196e-01 -1.22929966e+00
-1.44375408e+00 4.29426521e-01 2.16561452e-01 -4.37447242e-02
6.31371737e-01 5.11502028e-01 -3.49420875e-01 4.42730129e-01
1.75398901e-01 8.60821545e-01 6.82363272e-01 -1.19716501e+00
-2.58576155e-01 -4.51854497e-01 -2.12178379e-01 3.03214818e-01
2.07150191e-01 -1.83660954e-01 -5.00208914e-01 -1.00950038e+00
3.13236266e-01 -9.68112469e-01 -4.33635592e-01 -4.84284073e-01
-6.66356146e-01 3.73827666e-01 7.60235250e-01 -1.14813673e+00
1.11670959e+00 -2.20612884e+00 4.27607328e-01 3.16082060e-01
2.50231415e-01 2.01645896e-01 -9.59460437e-02 -1.75003991e-01
-1.34787619e-01 2.50530720e-01 -5.20728290e-01 1.70776516e-01
-4.41145092e-01 1.50753379e-01 -6.26426772e-04 3.01787436e-01
-8.73501822e-02 1.34101915e+00 -8.61402750e-01 -7.74066687e-01
2.78932363e-01 4.97503310e-01 -4.29699600e-01 4.10729527e-01
8.25247541e-03 1.13482523e+00 -7.69157231e-01 8.10658813e-01
6.43027484e-01 -2.45796561e-01 7.66190514e-02 -6.22906029e-01
2.26942152e-01 -3.13520968e-01 -9.16500807e-01 1.75738585e+00
-2.98845232e-01 8.87030512e-02 7.52606690e-02 -9.99464273e-01
5.74777186e-01 4.43032205e-01 6.85314238e-01 -8.69171441e-01
4.09044266e-01 3.66153389e-01 3.16430032e-01 -7.28661776e-01
-3.86101715e-02 -2.38099888e-01 -1.15190214e-02 2.98632115e-01
-9.65702012e-02 -9.63424519e-02 -8.47897902e-02 1.45504907e-01
8.01459491e-01 -9.69576165e-02 -1.53260589e-01 4.99777310e-02
6.41254485e-01 -1.35911733e-01 7.29920089e-01 3.05940092e-01
1.23188719e-02 9.54460144e-01 4.57431614e-01 -1.49231747e-01
-1.04593658e+00 -7.37823665e-01 9.98327360e-02 4.22370702e-01
4.17991012e-01 2.11387604e-01 -1.07274270e+00 -8.14945757e-01
-7.01258838e-01 4.57062751e-01 -7.24320471e-01 -4.01571840e-01
-7.23235250e-01 -1.03084862e+00 4.64340359e-01 7.01957345e-01
9.12691474e-01 -1.31565094e+00 -3.37730289e-01 4.09865826e-01
-6.81175172e-01 -8.84979546e-01 -7.26712346e-01 -3.12713236e-01
-8.62579346e-01 -9.85008895e-01 -1.16807413e+00 -8.05611074e-01
1.01952910e+00 7.80748352e-02 7.13756919e-01 2.39116952e-01
-1.96456552e-01 -1.73268765e-01 -4.06336039e-01 -1.02347232e-01
-5.01031637e-01 3.04724306e-01 -5.28151035e-01 4.64627147e-03
-3.00892383e-01 -6.07137501e-01 -8.72467518e-01 3.98046792e-01
-1.28243077e+00 5.24047971e-01 1.06574118e+00 1.22724199e+00
7.68043578e-01 1.61077216e-01 6.10408247e-01 -8.47864687e-01
5.11414170e-01 -6.01207972e-01 -2.21396267e-01 3.28261465e-01
-3.86995345e-01 -1.42910868e-01 5.46478629e-01 -5.39828062e-01
-1.02427566e+00 5.15514128e-02 -3.19496661e-01 -5.29018044e-01
6.36596084e-02 5.02092600e-01 -6.44493401e-01 -1.96035162e-01
3.23038787e-01 7.14247227e-01 2.43611962e-01 -1.76858939e-02
1.50441706e-01 5.12578249e-01 7.20752180e-01 -2.01733604e-01
4.37004119e-01 2.08747506e-01 9.00701433e-03 -2.63172746e-01
-5.32842457e-01 1.07283406e-01 -3.74128729e-01 -3.31283063e-01
1.21276116e+00 -6.45371318e-01 -3.45350057e-01 7.40883410e-01
-1.04519629e+00 -2.65911490e-01 -1.50563732e-01 5.64615607e-01
-3.79000843e-01 2.75565326e-01 -7.38363028e-01 -2.68824697e-01
-6.80278540e-01 -1.81348872e+00 8.90054107e-01 5.94953418e-01
4.61603194e-01 -8.15721273e-01 -3.70828092e-01 4.55433011e-01
6.44193172e-01 6.33381188e-01 9.52992618e-01 -3.43308866e-01
-8.91321540e-01 -8.59234557e-02 -2.58832783e-01 3.03119779e-01
1.93297938e-01 -3.92802447e-01 -5.83580554e-01 -1.55359492e-01
2.19231397e-01 -6.10922277e-02 6.78381026e-01 7.75069356e-01
1.43978715e+00 -2.81835794e-01 -3.43643457e-01 9.01324093e-01
1.37881613e+00 6.72609150e-01 9.20133591e-01 2.29352191e-01
1.04742563e+00 2.97144920e-01 3.33343238e-01 -9.77175534e-02
2.35690609e-01 5.29597819e-01 6.01746082e-01 -5.93171179e-01
-4.58898813e-01 -2.23159343e-01 1.38830319e-01 9.07469630e-01
-1.14305377e-01 -1.99951902e-01 -7.63480008e-01 4.15542841e-01
-1.58728576e+00 -9.28193867e-01 6.58630207e-02 1.79151928e+00
8.10169637e-01 -1.79012492e-01 -1.88030124e-01 1.25783645e-02
9.41977918e-01 -4.31091860e-02 -7.21427739e-01 2.84263819e-01
-1.01826742e-01 -1.02911226e-01 4.37926978e-01 2.93519706e-01
-7.26712167e-01 5.82962155e-01 5.27282906e+00 1.29315209e+00
-1.51121259e+00 4.52563822e-01 1.07865870e+00 2.03906819e-01
-4.58661079e-01 -2.25907326e-01 -3.32607657e-01 1.04057801e+00
4.17277664e-01 1.44876570e-01 4.79929656e-01 5.34469962e-01
6.86051548e-02 -1.84150860e-02 -4.53910112e-01 1.01724982e+00
2.30929270e-01 -1.21534514e+00 7.90007412e-02 1.89272493e-01
8.40286016e-01 -1.18691735e-01 1.07837416e-01 2.08625928e-01
2.17878502e-02 -1.14539492e+00 4.03564006e-01 7.75450408e-01
1.01156664e+00 -8.80258918e-01 1.18284154e+00 4.00330544e-01
-9.99045014e-01 8.37895572e-02 -2.07243159e-01 9.10671175e-01
3.76329035e-01 5.49708962e-01 -5.39690852e-01 9.23060179e-01
3.54160041e-01 4.07328904e-01 -5.36519110e-01 9.50776219e-01
-4.05456275e-01 4.72177386e-01 -1.37844577e-01 3.29881489e-01
3.45252573e-01 -2.80940264e-01 6.56826735e-01 7.85697937e-01
4.27915812e-01 3.82991076e-01 2.07445905e-01 1.06698120e+00
-3.27537507e-02 4.58553173e-02 -2.75884032e-01 1.55863985e-01
3.39774460e-01 1.40371406e+00 -8.92675459e-01 -3.87941003e-01
4.85718576e-03 9.25494015e-01 -6.00077137e-02 3.82705748e-01
-9.90881920e-01 -2.70918310e-01 -2.40053043e-01 2.04290256e-01
1.65893763e-01 9.39549655e-02 -2.12940320e-01 -1.10091877e+00
-1.11440755e-01 -1.08863306e+00 2.88672745e-01 -1.10432279e+00
-1.03898787e+00 1.07641912e+00 -1.54489785e-01 -1.29088247e+00
-2.80954242e-01 -1.10430144e-01 -7.24875808e-01 1.04731691e+00
-1.37375021e+00 -1.39707673e+00 -6.59723818e-01 7.65088081e-01
4.89222676e-01 -1.67200789e-01 4.53161776e-01 4.91951495e-01
-6.79708660e-01 5.91933787e-01 -1.30667776e-01 2.97313035e-01
5.18424153e-01 -9.08213794e-01 -3.02958727e-01 9.48422730e-01
-4.23934609e-01 3.29361111e-01 1.32341132e-01 -6.90691233e-01
-1.03587449e+00 -1.28222489e+00 1.12062097e-01 1.33574724e-01
2.33067185e-01 1.64913580e-01 -9.53581572e-01 5.01464367e-01
1.56017616e-01 6.76116571e-02 4.59996104e-01 -8.68943214e-01
4.19664919e-01 -1.56141728e-01 -1.32270515e+00 5.55206776e-01
6.70507610e-01 -2.10461721e-01 -3.50886852e-01 2.24064499e-01
8.80199194e-01 -8.37092996e-01 -9.43289399e-01 6.69317484e-01
2.46095836e-01 -7.58485496e-01 8.60499561e-01 -6.39245287e-02
7.00566769e-01 -4.80792046e-01 1.28265500e-01 -1.48678887e+00
-2.62684166e-01 -1.94529667e-01 3.23100656e-01 1.12259233e+00
2.52263010e-01 -7.02821672e-01 5.42435288e-01 4.70100880e-01
-4.11007494e-01 -1.00958192e+00 -7.65951157e-01 -3.36599827e-01
3.19780670e-02 -3.02353390e-02 1.00615168e+00 1.02757740e+00
-4.71302152e-01 2.07167536e-01 -2.44219974e-01 1.13865308e-01
4.86827850e-01 1.89900219e-01 3.62047881e-01 -6.96454942e-01
-2.64520168e-01 -5.48124611e-01 -2.84049064e-01 -8.16488087e-01
-9.06994045e-02 -9.51056600e-01 -1.00824781e-01 -1.62391698e+00
3.27183723e-01 -5.96700609e-01 -3.02835882e-01 4.91083026e-01
-4.00369525e-01 3.41546595e-01 1.93262801e-01 3.73166084e-01
-2.79279768e-01 6.68735206e-01 2.15015888e+00 -3.25599998e-01
3.89084518e-02 -9.30054709e-02 -7.74260283e-01 4.76267368e-01
7.24577844e-01 -3.87907952e-01 -4.93574113e-01 -4.65295047e-01
-1.92037389e-01 5.52367330e-01 4.58283275e-01 -1.01294243e+00
3.42011064e-01 8.38371273e-03 6.85819626e-01 -5.64777434e-01
2.26504635e-02 -8.35584164e-01 4.97675002e-01 6.42351985e-01
-8.01063105e-02 -1.06434189e-02 -5.38665336e-03 2.29452595e-01
-4.46705818e-01 -1.97212085e-01 8.71162951e-01 -3.81584108e-01
-3.37803364e-01 6.34526849e-01 -8.90292749e-02 -5.42644225e-03
1.17877674e+00 -1.83015957e-01 -2.85318464e-01 -2.88390070e-01
-7.96377122e-01 2.56794751e-01 3.28254640e-01 1.88668177e-01
6.53289974e-01 -1.47727859e+00 -7.31884599e-01 4.28943574e-01
-3.38922292e-01 4.27708119e-01 7.96186328e-01 1.27787781e+00
-7.24302292e-01 1.70805544e-01 -3.54788631e-01 -6.27842247e-01
-7.84276605e-01 4.58187282e-01 7.36732483e-01 -6.59133315e-01
-5.57062864e-01 5.00735044e-01 5.79473615e-01 -2.79809743e-01
-1.86878249e-01 -1.67581126e-01 -3.40226352e-01 -3.07880521e-01
4.80394661e-01 1.41100138e-01 -2.28416957e-02 -6.54822767e-01
-1.27276078e-01 5.57759523e-01 -3.48026417e-02 -6.06525987e-02
1.14818001e+00 -6.83242083e-02 -1.69654027e-01 -2.03975007e-01
9.84795094e-01 2.74296198e-02 -1.20394397e+00 -8.20356756e-02
-8.58659625e-01 -2.36458078e-01 3.46784472e-01 -9.86687243e-01
-1.93631899e+00 8.33524168e-01 7.79623985e-01 -3.41585949e-02
1.50316727e+00 -1.23172276e-01 1.13892972e+00 -5.80606937e-01
2.73005366e-01 -5.26980281e-01 8.65588244e-03 4.49255407e-02
9.10209119e-01 -9.67804253e-01 -2.59889811e-01 -4.84259725e-01
-9.03540611e-01 8.99471998e-01 8.51713896e-01 5.77164553e-02
3.55135411e-01 3.32427293e-01 4.48761769e-02 -1.83402658e-01
-1.22307681e-01 1.06615014e-01 3.41872275e-01 5.41259885e-01
9.18819979e-02 9.12155136e-02 -3.72083426e-01 9.45202410e-01
-2.60266718e-02 -1.96232706e-01 2.81486452e-01 6.16037548e-01
-1.39573544e-01 -1.05885255e+00 -3.48606259e-01 3.59984010e-01
-6.56663299e-01 -2.30726629e-01 1.47515675e-02 6.40011847e-01
3.55600417e-01 6.57665968e-01 -1.36352390e-01 -4.69734997e-01
7.83241540e-02 -3.95408571e-01 4.06931132e-01 -2.38892868e-01
-5.46313345e-01 3.56539935e-01 -4.81185675e-01 -4.53279585e-01
-4.00394976e-01 -3.20129395e-01 -1.44283617e+00 -5.75048178e-02
-4.22029853e-01 1.79014862e-01 6.61622643e-01 9.43873644e-01
3.34323227e-01 1.23225856e+00 7.96477973e-01 -6.99925184e-01
-1.80393383e-01 -1.11315894e+00 -3.98888648e-01 3.81095171e-01
-6.12816401e-03 -4.60778117e-01 -1.30658686e-01 -1.76937059e-02] | [14.142081260681152, -2.123483419418335] |
efb4fc7c-7e23-4416-8e4b-5b443e75bd6b | depth-completion-using-geometry-aware | 2203.10912 | null | https://arxiv.org/abs/2203.10912v2 | https://arxiv.org/pdf/2203.10912v2.pdf | Depth Completion using Geometry-Aware Embedding | Exploiting internal spatial geometric constraints of sparse LiDARs is beneficial to depth completion, however, has been not explored well. This paper proposes an efficient method to learn geometry-aware embedding, which encodes the local and global geometric structure information from 3D points, e.g., scene layout, object's sizes and shapes, to guide dense depth estimation. Specifically, we utilize the dynamic graph representation to model generalized geometric relationship from irregular point clouds in a flexible and efficient manner. Further, we joint this embedding and corresponded RGB appearance information to infer missing depths of the scene with well structure-preserved details. The key to our method is to integrate implicit 3D geometric representation into a 2D learning architecture, which leads to a better trade-off between the performance and efficiency. Extensive experiments demonstrate that the proposed method outperforms previous works and could reconstruct fine depths with crisp boundaries in regions that are over-smoothed by them. The ablation study gives more insights into our method that could achieve significant gains with a simple design, while having better generalization capability and stability. The code is available at https://github.com/Wenchao-Du/GAENet. | ['Yi Zhang', 'Hongyu Yang', 'Hu Chen', 'Wenchao Du'] | 2022-03-21 | null | null | null | null | ['depth-completion'] | ['computer-vision'] | [-8.29827115e-02 7.82276019e-02 -2.57357266e-02 -3.77933174e-01
-3.73895228e-01 -4.57415372e-01 2.41362751e-01 3.40922624e-02
2.72182319e-02 3.66214871e-01 5.19721694e-02 -8.76498297e-02
-1.88286006e-01 -1.11516094e+00 -6.12456858e-01 -6.14741445e-01
-3.61779854e-02 3.46960098e-01 2.41903514e-01 5.41052371e-02
3.69363546e-01 8.89160693e-01 -1.57039344e+00 -1.72738761e-01
1.06672013e+00 1.08086717e+00 5.21358013e-01 3.12897384e-01
-3.51838320e-01 3.12152594e-01 9.83134285e-02 -1.81856975e-01
3.80458981e-01 1.69739291e-01 -3.09949964e-01 2.06328377e-01
5.56851268e-01 -5.89218676e-01 -6.62976682e-01 9.05252516e-01
2.93574214e-01 8.20573606e-03 6.88388109e-01 -9.10330117e-01
-6.46497548e-01 -1.72039881e-01 -8.54901731e-01 -3.45089167e-01
2.21055672e-01 2.31454551e-01 9.44294512e-01 -1.06356895e+00
3.87187093e-01 1.29924059e+00 5.76753914e-01 2.81171173e-01
-1.08433914e+00 -6.20369732e-01 3.14671785e-01 2.28173971e-01
-1.61541092e+00 -2.71642387e-01 1.35661423e+00 -3.66785705e-01
6.72760904e-01 2.28232071e-01 8.11598837e-01 6.29108906e-01
4.40055914e-02 6.42096043e-01 8.37626696e-01 -2.75559485e-01
2.04806447e-01 2.32877879e-04 -6.50735795e-02 1.02591980e+00
4.37089175e-01 1.43813446e-01 -4.75441694e-01 4.37417552e-02
1.24573255e+00 4.18365091e-01 -4.54709738e-01 -8.52168620e-01
-1.03306580e+00 8.40350270e-01 8.97020221e-01 4.59149554e-02
-2.07124963e-01 2.08066240e-01 -6.69011548e-02 -3.27635892e-02
7.07300007e-01 2.14744017e-01 -3.74640256e-01 1.15609571e-01
-6.61854565e-01 1.49836466e-01 4.64944988e-01 1.05757105e+00
1.34658062e+00 6.24469370e-02 3.25337827e-01 7.19842792e-01
5.87313533e-01 5.99062383e-01 -3.65651660e-02 -1.19599628e+00
4.60886478e-01 9.49617028e-01 -3.99230681e-02 -1.26455903e+00
-2.60240167e-01 -1.84813797e-01 -8.35148036e-01 3.97566229e-01
2.25710884e-01 3.01666021e-01 -8.83810103e-01 1.35111225e+00
5.67010283e-01 3.72471124e-01 -2.27732301e-01 8.84625733e-01
8.11151028e-01 5.16856253e-01 -4.06727165e-01 8.50277990e-02
1.21507442e+00 -6.69195592e-01 -4.24858034e-01 -4.70855445e-01
3.28321218e-01 -6.43176794e-01 1.04868042e+00 1.95234239e-01
-1.03810513e+00 -4.94271100e-01 -1.06205881e+00 -5.90705276e-01
-1.37183383e-01 2.44183391e-01 1.10083485e+00 3.73647153e-01
-9.56582487e-01 5.44396758e-01 -1.09306037e+00 -2.63142645e-01
6.28672004e-01 2.68275470e-01 -2.33298898e-01 -4.08783853e-01
-6.37036681e-01 4.62113589e-01 1.62887916e-01 2.66450167e-01
-5.57838678e-01 -8.37947547e-01 -1.17562926e+00 -1.41697302e-01
4.56475616e-01 -9.92196500e-01 8.76539946e-01 -2.34107614e-01
-1.42355323e+00 7.34787762e-01 -4.52726424e-01 6.93053380e-02
4.20734435e-01 -1.00727871e-01 1.80647448e-01 3.33919108e-01
5.07870056e-02 6.39071047e-01 7.03683138e-01 -1.55033374e+00
-2.55873054e-01 -8.63748789e-01 1.92684099e-01 3.00055861e-01
-2.27231041e-01 -8.07223439e-01 -7.67139256e-01 -4.76685405e-01
8.15524876e-01 -6.75828218e-01 -3.96825105e-01 7.08727241e-01
-2.83898175e-01 -8.84681642e-02 8.21300685e-01 -4.43128288e-01
9.79771435e-01 -2.16756296e+00 1.11978710e-01 1.05059750e-01
4.42741096e-01 -1.11276619e-01 -6.28566742e-02 5.02391756e-01
2.84669518e-01 4.73953635e-02 -4.75296348e-01 -6.28998697e-01
-7.69669339e-02 3.25807661e-01 -2.74081945e-01 6.32539392e-01
2.50451833e-01 9.25548673e-01 -8.20630729e-01 -3.54091048e-01
7.75355577e-01 8.42670619e-01 -6.90129042e-01 3.52755070e-01
-1.99564353e-01 4.55218881e-01 -9.17885005e-01 9.89669919e-01
1.18295801e+00 -1.56323701e-01 -9.72683579e-02 -5.13194859e-01
-2.27366671e-01 1.96360126e-01 -1.27666378e+00 2.18606305e+00
-6.91781580e-01 3.54938596e-01 3.08238834e-01 -7.80216813e-01
1.07750297e+00 -1.90365911e-01 4.75248873e-01 -5.84224641e-01
-4.45603766e-02 1.17822632e-01 -6.48851156e-01 -4.57532465e-01
3.76730978e-01 3.42902765e-02 2.69150704e-01 1.78433299e-01
-3.68862242e-01 -7.02698648e-01 -3.87328267e-01 1.99173927e-01
7.67730951e-01 3.31981927e-01 2.92576253e-01 -3.15674432e-02
4.06729668e-01 -2.04663649e-01 5.89613199e-01 3.94725084e-01
1.22268103e-01 7.94674337e-01 1.55411854e-01 -3.07709128e-01
-8.55617523e-01 -1.24960387e+00 -3.25632304e-01 2.13384241e-01
6.12842798e-01 -4.12689865e-01 -2.86405236e-01 -3.61250073e-01
3.16012442e-01 3.81734103e-01 -5.72328985e-01 -5.80281951e-02
-5.79537153e-01 -2.61440724e-01 -1.08176649e-01 6.44395113e-01
5.44117570e-01 -5.45369744e-01 -5.58751643e-01 -6.79929554e-02
6.43762350e-02 -1.08915401e+00 -4.30840045e-01 -3.20298187e-02
-1.45194185e+00 -1.09677947e+00 -4.73855168e-01 -7.44802773e-01
8.69897425e-01 7.48083532e-01 8.54925275e-01 2.17774466e-01
-3.92779469e-01 4.64550793e-01 -2.85778433e-01 -1.71611100e-01
4.13082719e-01 -8.48278031e-02 -2.13459074e-01 -8.30804110e-02
2.26280659e-01 -1.15483296e+00 -9.83404756e-01 2.22810909e-01
-8.27124953e-01 3.77855361e-01 6.97743773e-01 5.43865204e-01
8.75443161e-01 1.05090946e-01 -5.71067780e-02 -7.78340459e-01
5.24161309e-02 -4.27144259e-01 -6.83510125e-01 -1.36209488e-01
-5.79504490e-01 1.26593411e-01 4.53944147e-01 -1.92135647e-02
-1.00126016e+00 2.88413495e-01 -1.93410814e-01 -7.83168197e-01
-2.11068079e-01 2.33594626e-01 -5.73689520e-01 -3.00702691e-01
1.64927945e-01 4.20045763e-01 4.80915010e-02 -7.93483138e-01
5.01776695e-01 4.21261221e-01 3.53944927e-01 -7.16456354e-01
1.07933772e+00 1.00886536e+00 1.89751118e-01 -9.46961999e-01
-7.47714043e-01 -6.40508950e-01 -8.71269763e-01 1.26744527e-03
4.59431410e-01 -1.12213337e+00 -6.11786783e-01 3.91613781e-01
-1.17620862e+00 -3.44817042e-01 -1.44142643e-01 4.87156242e-01
-5.20301878e-01 6.86469018e-01 -5.83129168e-01 -8.98762584e-01
-1.01111799e-01 -9.20252681e-01 1.43667316e+00 1.87599614e-01
2.54537702e-01 -1.07795429e+00 6.64796308e-03 3.86662811e-01
1.19057715e-01 4.60296303e-01 8.44250143e-01 3.98682237e-01
-1.38087571e+00 1.36651829e-01 -4.66527939e-01 1.99650347e-01
3.87747884e-01 3.18433009e-02 -1.10057926e+00 -3.26096028e-01
1.00495391e-01 -3.26870941e-02 9.35974121e-01 3.55267256e-01
1.52431345e+00 -1.68198660e-01 -4.00497705e-01 1.16943634e+00
1.81273973e+00 -2.01945320e-01 7.80870497e-01 2.45494902e-01
1.11031306e+00 6.90656781e-01 6.98927820e-01 6.13954961e-01
7.69066870e-01 6.12347722e-01 8.83334935e-01 -4.76579592e-02
-2.91084498e-01 -5.22912145e-01 1.09029405e-01 9.42764103e-01
-1.77293196e-01 2.44852342e-02 -7.24960864e-01 5.22602856e-01
-1.71413100e+00 -5.80871701e-01 -2.27940515e-01 2.13932276e+00
5.21010518e-01 -4.96337488e-02 -3.63745719e-01 6.99066073e-02
4.43806469e-01 4.36559618e-01 -7.79730022e-01 -8.96253064e-02
1.49177166e-03 2.40551144e-01 4.50605541e-01 7.88900197e-01
-7.39013016e-01 9.68840659e-01 5.13355064e+00 8.21239531e-01
-9.69780385e-01 -8.47756490e-02 2.75016725e-01 -4.96415161e-02
-9.09417868e-01 1.83393419e-01 -7.22254038e-01 3.07096958e-01
3.03205639e-01 -1.31150141e-01 4.25707579e-01 8.32330585e-01
4.12325203e-01 -1.11823715e-01 -9.92622316e-01 1.22404242e+00
6.43028170e-02 -1.35996211e+00 1.56730354e-01 3.83467585e-01
6.55115306e-01 -1.28120765e-01 4.08046618e-02 1.89644862e-02
2.14158036e-02 -8.23544621e-01 6.32670105e-01 6.73983634e-01
9.78968203e-01 -8.16288531e-01 3.20921868e-01 3.93935770e-01
-1.61851287e+00 7.26374090e-02 -6.46313906e-01 -2.03075737e-01
1.34149566e-01 9.83450055e-01 -5.55431485e-01 8.36848319e-01
6.90234661e-01 1.16905463e+00 -5.48693717e-01 9.85677540e-01
-4.37173605e-01 2.59904474e-01 -4.24250901e-01 2.30476871e-01
5.97982071e-02 -6.32090926e-01 3.69726509e-01 8.15885484e-01
5.11299074e-01 3.13201368e-01 1.21147387e-01 1.09314203e+00
2.53505763e-02 1.82110984e-02 -8.55812848e-01 1.74846306e-01
6.70968175e-01 1.22027826e+00 -4.96844471e-01 1.11794196e-01
-5.55098951e-01 8.79241943e-01 5.30898392e-01 4.28087085e-01
-5.92381477e-01 -3.03633928e-01 9.05274987e-01 4.98026222e-01
3.42759550e-01 -7.14330316e-01 -7.08441257e-01 -1.29743719e+00
3.22426558e-01 -2.42046967e-01 5.70250526e-02 -8.41377020e-01
-1.26888883e+00 1.74185336e-01 -1.23818994e-01 -1.35302377e+00
1.72314331e-01 -6.70074046e-01 -6.16159856e-01 8.76620293e-01
-1.95206153e+00 -1.35582650e+00 -8.23759556e-01 5.86434960e-01
5.94336629e-01 3.64048719e-01 5.77680647e-01 2.58807600e-01
-4.78899360e-01 3.26043069e-01 3.88463680e-03 -8.47058147e-02
3.03772628e-01 -1.22271156e+00 1.84442595e-01 7.17717469e-01
1.13198504e-01 5.88649809e-01 3.00560653e-01 -5.70465386e-01
-1.82907832e+00 -1.15394676e+00 4.45154071e-01 -4.42481160e-01
4.28978950e-01 -4.16953832e-01 -1.08648801e+00 4.72705871e-01
-3.32717448e-01 6.79457486e-02 3.85650873e-01 1.68669559e-02
-4.01534945e-01 -3.11130285e-01 -1.08695436e+00 5.30090809e-01
1.37612605e+00 -6.13107026e-01 -3.86290669e-01 6.22365586e-02
8.74169827e-01 -5.90046346e-01 -8.36702108e-01 5.69085836e-01
4.73636836e-01 -1.14494526e+00 1.27167380e+00 1.80349737e-01
4.60445851e-01 -5.96266389e-01 -3.47275555e-01 -1.06191993e+00
-3.25922161e-01 -2.81302780e-01 -4.72139210e-01 1.13778818e+00
-8.79279971e-02 -7.60756552e-01 1.00380671e+00 5.50492406e-01
-4.07401353e-01 -1.06767821e+00 -8.87880087e-01 -5.97509980e-01
8.24668631e-02 -6.02095544e-01 8.01883578e-01 8.09133351e-01
-6.00692570e-01 4.61606346e-02 -1.75104782e-01 6.44353032e-01
8.82951915e-01 5.77469826e-01 9.26297724e-01 -1.31290245e+00
-2.18056422e-02 -2.73496270e-01 -6.01724446e-01 -1.67593622e+00
1.12430841e-01 -7.60079741e-01 -1.39213830e-01 -1.78936756e+00
1.21097252e-01 -7.38136411e-01 -1.77134722e-02 3.57958972e-01
-9.17698517e-02 2.61329263e-01 -8.99612978e-02 2.56490469e-01
-2.01008946e-01 1.06862080e+00 1.56838202e+00 -9.14835706e-02
-2.19204053e-01 -1.27236918e-01 -8.21277440e-01 7.30033219e-01
7.67640173e-01 -1.83351353e-01 -5.66403747e-01 -9.11656916e-01
-3.09152566e-02 -5.39237112e-02 4.80641812e-01 -8.49087596e-01
2.15397656e-01 -3.07915270e-01 4.23592985e-01 -8.29084635e-01
7.33919740e-01 -9.75098729e-01 4.71165627e-02 1.32439360e-01
2.47266114e-01 -2.32946262e-01 2.07734078e-01 8.56869221e-01
-1.64427057e-01 -7.91946203e-02 6.00002885e-01 -2.03294039e-01
-8.84960294e-01 9.26513851e-01 3.94931734e-01 -2.94075876e-01
9.37288821e-01 -5.54838300e-01 -4.87185903e-02 -2.20371544e-01
-3.42069119e-01 3.20946366e-01 1.10494387e+00 2.14480549e-01
1.09230065e+00 -1.42635334e+00 -4.52443421e-01 3.95844787e-01
2.12033093e-01 8.01074564e-01 4.91262019e-01 5.18501699e-01
-7.79010832e-01 3.33170950e-01 3.05331238e-02 -8.93633962e-01
-9.19394732e-01 3.49257588e-01 1.51312366e-01 1.54006109e-01
-9.83324945e-01 7.88078725e-01 4.33302671e-01 -4.86125261e-01
1.67782024e-01 -5.48648000e-01 1.05137065e-01 -1.21172719e-01
3.76865983e-01 2.94550747e-01 3.74035574e-02 -4.33203936e-01
-4.09688026e-01 1.27271986e+00 1.40206993e-01 2.29342222e-01
1.48245740e+00 -4.35254902e-01 -1.11917138e-01 4.90785897e-01
1.25548458e+00 3.19766670e-01 -1.74139452e+00 -2.57490516e-01
-4.12555814e-01 -1.17732704e+00 4.16248202e-01 -2.12171406e-01
-1.32675016e+00 1.19682419e+00 4.62337047e-01 -1.20772272e-01
1.09597325e+00 1.58712134e-01 8.32177520e-01 9.80077386e-02
7.09385753e-01 -5.86023510e-01 2.17648134e-01 2.98840374e-01
9.41727996e-01 -1.37446547e+00 2.53586888e-01 -8.53460252e-01
-1.94810033e-01 1.22020233e+00 6.42762721e-01 -1.94187686e-01
6.78282022e-01 1.71875339e-02 -1.68471679e-01 -4.04948145e-01
-4.47289705e-01 -3.86104167e-01 3.22878391e-01 7.74865806e-01
1.30898669e-01 1.20761484e-01 1.45424187e-01 2.87131637e-01
-1.56624839e-01 -2.82981902e-01 3.17110270e-01 7.27271616e-01
-5.27965665e-01 -1.00354266e+00 -2.87144303e-01 2.75536656e-01
2.88095385e-01 1.60461999e-02 -1.40807524e-01 7.04189837e-01
1.84496149e-01 6.78605795e-01 1.46770537e-01 -2.84732163e-01
4.17039275e-01 -4.90566909e-01 6.91979647e-01 -9.38723147e-01
3.79443109e-01 6.11240268e-02 -3.51575106e-01 -9.10534799e-01
-2.84785539e-01 -6.28888786e-01 -1.22140944e+00 -3.62927645e-01
-1.82058334e-01 -1.85803413e-01 7.24327862e-01 4.68143910e-01
4.29605395e-01 2.03184679e-01 9.10157025e-01 -1.17822218e+00
-1.08054772e-01 -5.46548665e-01 -7.27646351e-01 1.42560035e-01
4.49355751e-01 -8.58116150e-01 -6.82725191e-01 -2.97789425e-01] | [8.626337051391602, -2.9159913063049316] |
206789a2-85f1-4841-b49f-f0ef7044638b | unsupervised-melody-to-lyric-generation | 2305.19228 | null | https://arxiv.org/abs/2305.19228v1 | https://arxiv.org/pdf/2305.19228v1.pdf | Unsupervised Melody-to-Lyric Generation | Automatic melody-to-lyric generation is a task in which song lyrics are generated to go with a given melody. It is of significant practical interest and more challenging than unconstrained lyric generation as the music imposes additional constraints onto the lyrics. The training data is limited as most songs are copyrighted, resulting in models that underfit the complicated cross-modal relationship between melody and lyrics. In this work, we propose a method for generating high-quality lyrics without training on any aligned melody-lyric data. Specifically, we design a hierarchical lyric generation framework that first generates a song outline and second the complete lyrics. The framework enables disentanglement of training (based purely on text) from inference (melody-guided text generation) to circumvent the shortage of parallel data. We leverage the segmentation and rhythm alignment between melody and lyrics to compile the given melody into decoding constraints as guidance during inference. The two-step hierarchical design also enables content control via the lyric outline, a much-desired feature for democratizing collaborative song creation. Experimental results show that our model can generate high-quality lyrics that are more on-topic, singable, intelligible, and coherent than strong baselines, for example SongMASS, a SOTA model trained on a parallel dataset, with a 24% relative overall quality improvement based on human ratings. O | ['Nanyun Peng', 'Jing Huang', 'Tagyoung Chung', 'Wenbo Zhao', 'Chenyang Tao', 'Gunnar Sigurdsson', 'Alessandra Cervone', 'Shereen Oraby', 'Anjali Narayan-Chen', 'Yufei Tian'] | 2023-05-30 | null | null | null | null | ['disentanglement'] | ['methodology'] | [ 2.12916389e-01 4.39634621e-02 -1.24057382e-01 -1.07709251e-01
-1.11881959e+00 -1.19464099e+00 6.32966101e-01 -6.64255679e-01
2.66611695e-01 6.80141509e-01 8.57463002e-01 2.13769563e-02
1.57224596e-01 -5.84627748e-01 -6.05111063e-01 -6.32075310e-01
4.17581201e-01 6.21884465e-01 -4.10695016e-01 -3.75390232e-01
1.79581828e-02 3.69995758e-02 -1.38487530e+00 2.97440916e-01
9.79274213e-01 7.41515160e-01 3.68856937e-01 1.07339561e+00
8.34463760e-02 7.36730993e-01 -9.85199571e-01 -2.61202246e-01
3.97624940e-01 -1.07632029e+00 -3.86414617e-01 1.04186438e-01
6.30682707e-01 -3.61595482e-01 -8.15805048e-02 6.01870239e-01
7.72248268e-01 1.77488312e-01 7.73878992e-01 -7.61212409e-01
-3.51933539e-01 1.00480223e+00 -1.84820462e-02 -3.91342223e-01
4.43809658e-01 3.32558244e-01 1.73844981e+00 -8.38726044e-01
5.89282453e-01 8.56822729e-01 4.88134176e-01 6.08035386e-01
-1.48867607e+00 -8.27162743e-01 -3.77239555e-01 -2.31047288e-01
-1.26926386e+00 -8.73761654e-01 1.01471865e+00 -5.70205510e-01
5.14342129e-01 8.05594742e-01 9.28155005e-01 1.08545792e+00
-1.54102907e-01 9.70330596e-01 7.48661995e-01 -5.03786325e-01
4.98525314e-02 -3.04210782e-01 -6.01868987e-01 1.84309229e-01
-5.04226983e-01 1.50296181e-01 -8.96787465e-01 5.65979220e-02
8.37943375e-01 -5.82463205e-01 -3.19249243e-01 7.55189881e-02
-1.32944512e+00 8.10386896e-01 -5.73928514e-03 2.75764823e-01
-2.78790832e-01 2.14127243e-01 3.10410649e-01 2.23664060e-01
7.71040097e-02 1.34042716e+00 -2.83898503e-01 -6.16861641e-01
-1.66143179e+00 7.65529454e-01 9.77659285e-01 8.89598191e-01
1.94666773e-01 6.65251732e-01 -7.29192436e-01 1.16379058e+00
6.36156276e-02 6.31564498e-01 6.08004570e-01 -1.10430670e+00
3.69387329e-01 -4.93247360e-02 3.55282496e-03 -5.96622050e-01
5.48749268e-02 -7.78918207e-01 -6.37183368e-01 -1.56742156e-01
3.76020521e-01 -1.77093282e-01 -7.05223441e-01 2.03550363e+00
3.20564653e-03 1.04055636e-01 -1.54023811e-01 8.94732714e-01
8.44343841e-01 1.01774633e+00 -3.55965942e-01 -4.56558853e-01
1.27717364e+00 -1.17047727e+00 -9.18856680e-01 -3.86990011e-02
1.46571591e-01 -1.22948813e+00 1.56734359e+00 5.50592661e-01
-1.49720681e+00 -7.57795036e-01 -1.01697838e+00 -2.47859254e-01
4.86036509e-01 3.91288400e-01 4.36610013e-01 4.99185175e-01
-6.32126629e-01 5.40415645e-01 -3.57056767e-01 2.64489740e-01
1.95778701e-02 3.02536897e-02 1.46070257e-01 4.49784786e-01
-1.23985004e+00 3.65254700e-01 3.44134510e-01 -1.50152966e-01
-1.12482512e+00 -9.97463942e-01 -7.49735713e-01 7.88602754e-02
2.67671645e-01 -7.13754177e-01 1.64616859e+00 -8.35878849e-01
-2.01303124e+00 5.14609694e-01 4.60535334e-03 -3.04048359e-01
4.91200358e-01 -3.40113431e-01 -2.00037852e-01 -3.20045054e-02
1.93531454e-01 7.04420030e-01 1.04193997e+00 -1.08893549e+00
-4.04050618e-01 2.41993338e-01 -2.61513174e-01 7.93716550e-01
-4.37458128e-01 -1.75075501e-01 -8.08367789e-01 -1.34679592e+00
-2.00371891e-01 -1.10804069e+00 1.35176405e-01 -5.81445575e-01
-8.04342568e-01 -3.74432653e-02 4.66287315e-01 -8.44245315e-01
1.84135795e+00 -1.94346941e+00 4.59451795e-01 2.49351398e-03
-5.36509976e-02 -2.70697009e-02 -1.91888854e-01 4.98544097e-01
1.07691139e-01 -8.23558494e-02 -1.43618807e-01 -3.53710085e-01
3.16407114e-01 -1.47409700e-02 -1.04334414e+00 -1.78262174e-01
-1.09654903e-01 1.21063375e+00 -7.53384352e-01 -4.18263286e-01
-2.52292138e-02 3.52373302e-01 -9.69656587e-01 7.79312253e-01
-6.22149050e-01 8.38785708e-01 -4.18917090e-02 4.89232957e-01
-1.20770536e-01 4.49554957e-02 2.57495493e-01 -3.70616466e-01
-2.49232501e-02 9.68553603e-01 -8.92518878e-01 2.00671530e+00
-7.04915762e-01 6.14663422e-01 -2.30997205e-01 -1.73921883e-01
1.03697455e+00 6.75138354e-01 3.19033593e-01 -2.64893472e-01
-8.99395440e-03 1.73781186e-01 1.92292124e-01 -2.26728663e-01
8.42908561e-01 -6.03202939e-01 -4.95491713e-01 6.99209094e-01
6.13868237e-02 -1.31723511e+00 3.14779788e-01 1.18904047e-01
8.25501859e-01 4.49167997e-01 2.84126788e-01 3.64018120e-02
5.44954427e-02 -1.12536542e-01 5.44966877e-01 6.65455818e-01
4.26275283e-01 1.23183227e+00 2.89215773e-01 1.51675075e-01
-1.34957993e+00 -1.26681578e+00 8.56166407e-02 1.18924391e+00
-4.89009947e-01 -9.08738554e-01 -9.77916598e-01 -1.74722224e-01
-3.49545300e-01 1.03192580e+00 -1.65714815e-01 1.42439425e-01
-8.50236237e-01 -4.28362668e-01 1.02122116e+00 2.06979737e-01
2.50137337e-02 -1.42012632e+00 -1.93772972e-01 3.66113156e-01
-6.53574347e-01 -8.82692277e-01 -1.17375147e+00 1.04854591e-02
-6.20853066e-01 -6.29058838e-01 -7.93690383e-01 -7.36343741e-01
1.15357161e-01 -3.24366808e-01 1.41026771e+00 -2.24995226e-01
-1.63410142e-01 -4.37471345e-02 -3.08735937e-01 -2.41842359e-01
-7.27767169e-01 4.16146189e-01 8.72028321e-02 -1.28950939e-01
-2.47651026e-01 -8.08874369e-01 -3.42719316e-01 1.66996315e-01
-9.53758180e-01 7.57124901e-01 3.23204875e-01 1.04540181e+00
6.09577835e-01 5.82645833e-02 7.46853292e-01 -8.41905117e-01
8.29359651e-01 -1.38035357e-01 -4.41313446e-01 -7.52221718e-02
-3.06749731e-01 -1.12899549e-01 1.11170602e+00 -4.90092516e-01
-9.78020310e-01 1.87209062e-02 -2.76545912e-01 -3.37448001e-01
4.12497595e-02 5.78945756e-01 -3.84942681e-01 6.88174367e-01
7.29461253e-01 4.13698822e-01 -2.60287941e-01 -5.40324807e-01
8.32887113e-01 6.99841321e-01 1.00382078e+00 -9.43876088e-01
1.19188190e+00 -4.58225310e-02 -3.90732795e-01 -8.77020299e-01
-1.04474390e+00 -1.11000903e-01 -4.61574227e-01 -2.44282439e-01
7.37283826e-01 -1.13683081e+00 -4.10439849e-01 3.18058044e-01
-1.02626920e+00 -6.13790631e-01 -7.27638245e-01 3.83069813e-01
-1.01910889e+00 8.72526914e-02 -9.10407007e-01 -5.80393016e-01
-6.75146163e-01 -9.03908670e-01 1.08913195e+00 5.06692156e-02
-9.43492889e-01 -6.77039385e-01 3.77033502e-01 6.62108421e-01
1.93003699e-01 2.44423866e-01 9.06187415e-01 -2.16294542e-01
-6.33586884e-01 3.22437249e-02 2.50708401e-01 2.26374328e-01
4.78229642e-01 8.23202636e-03 -1.09759271e+00 -2.13082090e-01
9.00942311e-02 -6.58697963e-01 4.62751180e-01 3.70187551e-01
8.70288551e-01 -6.63165271e-01 3.94726127e-01 9.49555278e-01
7.77727485e-01 5.15685901e-02 5.34091532e-01 -1.46192953e-01
9.82271552e-01 6.08973384e-01 5.88750362e-01 4.62665945e-01
2.44022295e-01 1.15596282e+00 -1.22406796e-01 4.81354855e-02
-6.47463500e-01 -9.23824191e-01 7.54810095e-01 1.77454853e+00
1.28784567e-01 -3.13429624e-01 -5.25483966e-01 7.37853944e-01
-1.60220909e+00 -1.10536015e+00 6.19347766e-02 2.13014793e+00
1.78503478e+00 -2.15319581e-02 2.19244510e-01 2.31121093e-01
4.36465502e-01 4.86032277e-01 -4.70984101e-01 -2.01603428e-01
-2.33819589e-01 4.96887207e-01 -2.06033260e-01 6.30509257e-01
-9.05003130e-01 1.34932709e+00 6.12013912e+00 1.32714629e+00
-1.07910943e+00 -1.02214284e-01 4.78063166e-01 -7.23690033e-01
-8.04923058e-01 3.72789204e-02 -8.06479514e-01 5.40963650e-01
7.95087516e-01 -2.94179946e-01 1.07738721e+00 6.23364687e-01
6.76951826e-01 4.07453597e-01 -1.29075623e+00 1.00325358e+00
2.03069404e-01 -1.54002941e+00 2.78046131e-01 -1.53013602e-01
1.10519052e+00 -3.98854762e-01 2.72237271e-01 3.68890047e-01
3.79117131e-01 -1.37512004e+00 1.28821051e+00 2.81668514e-01
1.34283566e+00 -6.72532380e-01 -1.52860940e-01 2.09906071e-01
-1.37310326e+00 3.21190804e-01 1.70472950e-01 -1.60474479e-01
5.62534571e-01 4.44634378e-01 -1.28492188e+00 3.57127547e-01
1.37627691e-01 7.42263615e-01 -2.10508704e-01 7.81168640e-01
-7.69818664e-01 1.30270910e+00 -1.37897223e-01 2.06167728e-01
4.51533198e-02 -1.93002120e-01 9.12799537e-01 1.28202224e+00
4.13903832e-01 -1.54474983e-02 3.14209461e-01 1.23327184e+00
-1.34162769e-01 3.38426888e-01 -5.28582752e-01 -5.20782769e-01
5.94801009e-01 1.29992318e+00 -3.47256571e-01 -2.08967760e-01
2.57377118e-01 1.09890831e+00 6.78047026e-03 3.80560189e-01
-7.38635242e-01 -2.89073974e-01 5.76995969e-01 1.35309428e-01
2.41696492e-01 -3.19741964e-01 -6.61579251e-01 -1.19012570e+00
-2.01445699e-01 -1.38315129e+00 6.33163229e-02 -9.28891361e-01
-1.24417269e+00 7.59871960e-01 -3.42937589e-01 -1.43950152e+00
-1.01502073e+00 3.26683000e-02 -7.01056659e-01 1.02847040e+00
-1.06901193e+00 -1.36480272e+00 7.02042058e-02 2.12502986e-01
1.03308380e+00 -4.26776379e-01 9.99071002e-01 1.73012733e-01
-1.85958162e-01 6.82516754e-01 -1.25034332e-01 1.09292760e-01
9.12614763e-01 -1.43130100e+00 5.25310099e-01 7.72126317e-01
8.49819481e-01 6.14403784e-01 8.59342396e-01 -6.75987601e-01
-1.09911442e+00 -1.03523695e+00 9.99702632e-01 -6.25834048e-01
5.62613666e-01 -7.36098945e-01 -5.22455871e-01 3.71700317e-01
2.97378927e-01 -8.21732581e-01 1.08646953e+00 2.35657096e-01
-2.12498054e-01 -3.76908435e-03 -2.93417186e-01 1.15923071e+00
7.95942843e-01 -8.53996813e-01 -6.76334381e-01 1.65265799e-01
9.23284411e-01 -5.94904780e-01 -8.12621415e-01 2.66268969e-01
8.44117224e-01 -7.44430661e-01 8.30143929e-01 -3.39129657e-01
7.72875607e-01 -6.74055874e-01 -2.14558572e-01 -1.46431184e+00
-2.52437025e-01 -1.37622750e+00 -1.79058120e-01 1.47176290e+00
4.76924807e-01 4.28759694e-01 7.19808877e-01 2.35281572e-01
-3.04162741e-01 -3.56273949e-01 -4.53556448e-01 -7.41667688e-01
-1.15131117e-01 -7.39970386e-01 6.06756806e-01 1.02605677e+00
-6.23714812e-02 1.01382077e+00 -9.21112359e-01 -3.12725455e-01
3.33803117e-01 5.29690266e-01 1.21860242e+00 -8.81933928e-01
-1.04860592e+00 -6.84313416e-01 4.37674552e-01 -1.40801382e+00
2.09952220e-02 -1.16860437e+00 4.59667385e-01 -1.18929911e+00
-2.03536116e-02 -6.72970414e-01 4.23518978e-02 4.82760638e-01
-2.94932991e-01 6.43616974e-01 7.14534938e-01 5.84760368e-01
-2.71280944e-01 9.47330713e-01 1.46169519e+00 -7.05363154e-02
-7.74822891e-01 1.55698910e-01 -7.59234190e-01 6.41430318e-01
5.43983698e-01 -3.53859931e-01 -5.41401446e-01 -2.17009202e-01
1.88179955e-01 4.09204125e-01 -1.18123785e-01 -9.08234358e-01
1.38669044e-01 -1.84315369e-01 1.55825868e-01 -5.58897495e-01
6.29770219e-01 -1.24053270e-01 5.43240309e-01 8.07169080e-02
-6.80744410e-01 -2.83085704e-01 8.61710086e-02 4.72712554e-02
-3.34619135e-01 -3.00933540e-01 5.49191415e-01 -1.95234343e-02
1.39548518e-02 6.60647675e-02 -2.18495324e-01 4.52774048e-01
3.15010965e-01 -2.43382417e-02 1.11580178e-01 -9.74587381e-01
-6.07371807e-01 -1.97001383e-01 4.47277039e-01 5.58910906e-01
2.96236783e-01 -1.71299779e+00 -9.95826781e-01 2.30117232e-01
4.50946167e-02 1.13766000e-01 -7.66078308e-02 3.90143037e-01
-4.25954580e-01 3.20564359e-01 7.00653568e-02 -3.80870789e-01
-1.16333640e+00 -8.47207382e-02 -2.67909411e-02 -5.00477254e-01
-4.65187758e-01 8.11525643e-01 5.84812045e-01 -6.34619713e-01
2.23287046e-01 2.71561686e-02 -9.52609777e-02 1.68174475e-01
3.94745290e-01 4.11846079e-02 -3.24151605e-01 -5.76033235e-01
3.08496803e-01 4.60366607e-01 1.74673006e-01 -7.41764188e-01
1.21674669e+00 1.77254435e-02 -9.12464485e-02 8.08285475e-01
6.91934407e-01 8.53118181e-01 -1.44078922e+00 8.83939937e-02
-2.44601712e-01 -4.23400581e-01 -8.75217095e-02 -1.13832712e+00
-5.91443777e-01 6.63782001e-01 -2.62772501e-01 1.64026752e-01
1.01194286e+00 -2.41432503e-01 1.25138617e+00 1.21800117e-01
1.33207208e-02 -1.19119179e+00 4.19013917e-01 7.55957365e-01
1.27262378e+00 -5.56229115e-01 -1.67356551e-01 -2.05984131e-01
-1.01614380e+00 9.91748035e-01 4.76410955e-01 -6.79146126e-02
1.41931921e-01 3.00681382e-01 2.19466507e-01 2.19617888e-01
-8.00973594e-01 7.40017695e-03 8.15072179e-01 2.59246111e-01
7.56908298e-01 3.01332146e-01 -2.73920357e-01 1.00013208e+00
-1.30496144e+00 -2.37299606e-01 3.75846148e-01 3.01552080e-02
-2.96494037e-01 -1.36295855e+00 -3.89516413e-01 2.42119789e-01
-4.28921819e-01 -7.99404323e-01 -5.19526601e-01 1.71971038e-01
1.43808410e-01 9.36550558e-01 -1.70548279e-02 -5.02213120e-01
4.56401892e-02 1.67774409e-01 4.60347801e-01 -9.80105460e-01
-8.79464924e-01 9.03575659e-01 3.29543233e-01 -1.57687083e-01
2.16531977e-01 -5.54566026e-01 -1.07291853e+00 1.88270826e-02
-3.47780794e-01 3.18383604e-01 3.98825139e-01 7.42418468e-01
1.71095118e-01 8.40571821e-01 8.58084679e-01 -9.71433580e-01
-3.59062970e-01 -1.02289236e+00 -6.19153500e-01 4.92517680e-01
3.28510217e-02 -4.72218245e-02 -2.26094455e-01 4.33595449e-01] | [15.890233993530273, 5.64668083190918] |
a0bc599f-803b-4494-8579-e490c45263b2 | lpf-defense-3d-adversarial-defense-based-on | 2202.11287 | null | https://arxiv.org/abs/2202.11287v2 | https://arxiv.org/pdf/2202.11287v2.pdf | LPF-Defense: 3D Adversarial Defense based on Frequency Analysis | Although 3D point cloud classification has recently been widely deployed in different application scenarios, it is still very vulnerable to adversarial attacks. This increases the importance of robust training of 3D models in the face of adversarial attacks. Based on our analysis on the performance of existing adversarial attacks, more adversarial perturbations are found in the mid and high-frequency components of input data. Therefore, by suppressing the high-frequency content in the training phase, the models robustness against adversarial examples is improved. Experiments showed that the proposed defense method decreases the success rate of six attacks on PointNet, PointNet++ ,, and DGCNN models. In particular, improvements are achieved with an average increase of classification accuracy by 3.8 % on drop100 attack and 4.26 % on drop200 attack compared to the state-of-the-art methods. The method also improves models accuracy on the original dataset compared to other available methods. | ['Kimia Noorbakhsh', 'Shohreh Kasaei', 'Arian Etemadi', 'Hanieh Naderi'] | 2022-02-23 | null | null | null | null | ['adversarial-defense', 'point-cloud-classification'] | ['adversarial', 'computer-vision'] | [-1.45600051e-01 -1.40492976e-01 2.20206678e-01 9.11772400e-02
-4.81234998e-01 -1.08095753e+00 7.75172234e-01 7.78883472e-02
-1.67721570e-01 4.66944188e-01 -3.89930695e-01 -4.57546920e-01
-9.87168029e-03 -9.79888618e-01 -8.61836076e-01 -8.09369624e-01
-3.06871533e-01 3.33680451e-01 5.35991430e-01 -4.19861406e-01
4.12274033e-01 1.29607034e+00 -1.17146122e+00 -9.35558043e-03
7.29667068e-01 1.05698228e+00 -5.88452160e-01 4.62717026e-01
-9.72072873e-03 3.65817875e-01 -1.21143699e+00 -6.37972116e-01
7.26298332e-01 2.18336955e-01 -2.95331270e-01 -4.25433546e-01
5.48882067e-01 -2.60087729e-01 -7.97639310e-01 1.38918424e+00
7.35375345e-01 -9.22203958e-02 6.65171623e-01 -1.62529910e+00
-4.23196197e-01 3.07648748e-01 -6.39412105e-01 2.16914445e-01
2.95882523e-01 2.68084288e-01 2.28621483e-01 -6.40024424e-01
4.04667437e-01 1.26218545e+00 6.93132162e-01 7.09788680e-01
-1.06082344e+00 -1.34930480e+00 -3.39619536e-03 1.21786445e-01
-1.38263321e+00 1.51711464e-01 9.78967488e-01 -1.35548383e-01
8.19153309e-01 4.96583223e-01 3.01100671e-01 1.37352991e+00
4.05284375e-01 2.78206497e-01 7.31400788e-01 -1.03321493e-01
1.90390021e-01 2.28249412e-02 -2.14483604e-01 1.95000410e-01
3.95875692e-01 6.91141844e-01 -9.93045419e-02 -6.63561821e-01
5.47864795e-01 -8.55162442e-02 -1.01256788e-01 -2.12043598e-01
-6.73972666e-01 8.10233235e-01 6.86109841e-01 1.86212271e-01
-2.04637185e-01 9.23812240e-02 4.84089583e-01 1.80792645e-01
4.50353146e-01 3.98155898e-01 -4.35125232e-01 9.08104181e-02
-5.63916087e-01 5.42485952e-01 6.68293834e-01 1.02760208e+00
2.94055581e-01 2.95086205e-01 2.14719877e-01 5.82941413e-01
2.46549681e-01 9.82557058e-01 -1.55370966e-01 -4.69159067e-01
7.54246354e-01 5.12655139e-01 -1.97670296e-01 -1.41193676e+00
-3.36630255e-01 -5.12192309e-01 -1.13853097e+00 6.96692348e-01
3.77411038e-01 -2.35261291e-01 -1.09395456e+00 1.50386596e+00
4.41169828e-01 5.91676652e-01 1.36323437e-01 5.58622003e-01
6.61651313e-01 6.21899247e-01 2.40426064e-01 3.08668196e-01
7.43049204e-01 -2.13325456e-01 -2.98428267e-01 2.83700079e-02
1.93053201e-01 -1.05279136e+00 5.76615810e-01 3.81904274e-01
-7.66534209e-01 -5.78726888e-01 -1.19458961e+00 6.31409347e-01
-4.96095002e-01 -5.92090964e-01 4.23577756e-01 1.21274781e+00
-3.56273711e-01 7.25666940e-01 -5.89115977e-01 -1.75960101e-02
6.56281412e-01 6.60176873e-01 -5.76120198e-01 4.23823856e-02
-1.48489439e+00 1.04181278e+00 1.01974137e-01 -1.40746474e-01
-9.56543624e-01 -1.14019668e+00 -4.62155521e-01 -2.07044214e-01
-7.29646683e-02 -2.30832726e-01 7.99499512e-01 -2.84153879e-01
-1.27058041e+00 6.77022874e-01 4.43971455e-01 -8.38261604e-01
8.73363972e-01 -2.14265853e-01 -7.85899282e-01 2.90921062e-01
-3.25064689e-01 3.88958067e-01 9.58509028e-01 -1.40132511e+00
-5.45287311e-01 -3.63116473e-01 3.02201390e-01 -1.54668450e-01
-2.21348226e-01 1.05584092e-01 -2.18346208e-01 -9.25780356e-01
7.93999881e-02 -1.25164139e+00 -3.25698942e-01 9.45832133e-02
-5.19356191e-01 1.29841268e-01 1.40248680e+00 -2.21855804e-01
7.29571760e-01 -2.47935438e+00 -1.95888951e-01 6.27358317e-01
1.45706072e-01 7.71669567e-01 1.02407932e-02 5.68476677e-01
-4.37507749e-01 5.82171321e-01 -7.88023695e-02 -8.39791447e-02
-6.37140796e-02 2.13323504e-01 -8.36583972e-01 7.91588426e-01
2.08356783e-01 3.80222887e-01 -6.64368212e-01 -6.08784556e-02
5.81906080e-01 3.95237625e-01 -6.40734732e-01 9.11445394e-02
1.43181473e-01 5.57793200e-01 -4.82041448e-01 7.40998745e-01
1.37350869e+00 5.71741819e-01 -4.79237795e-01 8.48241076e-02
3.75066161e-01 -5.69207892e-02 -1.16242552e+00 1.16854393e+00
-1.59505442e-01 4.66439933e-01 -1.45702228e-01 -6.64288878e-01
1.22469509e+00 3.90678436e-01 5.58586657e-01 -3.67990613e-01
2.93830991e-01 1.65143117e-01 2.22568825e-01 -9.36880037e-02
1.49466014e-02 -8.48998651e-02 -2.80508697e-01 -4.16655727e-02
-1.16257712e-01 -4.42349494e-01 -3.70502412e-01 2.56658554e-01
1.21046948e+00 -3.47985864e-01 -1.49774686e-01 -1.88669980e-01
8.22499156e-01 -1.56172842e-01 4.56601053e-01 7.95808256e-01
-2.94278800e-01 6.10491991e-01 3.82904142e-01 -3.91716212e-01
-8.93155158e-01 -1.44891834e+00 -2.69941241e-01 2.15573490e-01
5.44082046e-01 -3.58169168e-01 -5.81505477e-01 -1.10032094e+00
5.78516722e-01 8.51446450e-01 -5.32542229e-01 -7.25639880e-01
-6.85700238e-01 -6.33820295e-01 1.31211722e+00 3.85046601e-01
6.90600872e-01 -7.60232091e-01 8.36173818e-02 -4.91107404e-02
2.84955919e-01 -1.41528916e+00 1.40323425e-02 2.35937769e-03
-8.34172428e-01 -1.11432862e+00 -2.55751014e-01 -4.67648417e-01
5.84913492e-01 2.09860981e-01 9.12398279e-01 2.67050385e-01
-7.94433206e-02 -4.33045113e-03 -5.43761313e-01 -9.77281690e-01
-7.61067152e-01 4.47507314e-02 4.50462669e-01 -2.15460658e-01
3.84373635e-01 -8.74932230e-01 -1.89393148e-01 6.43038690e-01
-9.17851448e-01 -8.35759103e-01 2.60234028e-01 4.76557612e-01
2.88181096e-01 4.46066737e-01 5.32915115e-01 -8.48456919e-01
4.41642404e-01 -5.68496287e-01 -7.88592815e-01 -3.68097603e-01
-3.42835069e-01 -3.26938748e-01 1.03005612e+00 -7.49642253e-01
-5.27465165e-01 -1.33843273e-01 -6.41754389e-01 -9.09722269e-01
-4.77090955e-01 -8.60214457e-02 -3.89365703e-01 -8.68466020e-01
1.03509557e+00 -5.46497032e-02 -6.73801675e-02 -3.79152119e-01
7.05217123e-02 4.50089544e-01 4.75811213e-01 -4.94971633e-01
1.91479766e+00 4.23002869e-01 4.60894346e-01 -9.20732915e-01
-2.31315449e-01 -7.27948025e-02 -4.73847389e-01 -2.50110298e-01
4.32926238e-01 -6.21635377e-01 -7.78665423e-01 8.47616792e-01
-1.16190290e+00 1.85511172e-01 7.32357893e-03 2.78423995e-01
-2.66102821e-01 6.08095050e-01 -3.53042990e-01 -5.86106062e-01
-2.80088484e-01 -1.21413267e+00 6.93040967e-01 2.48870556e-03
-6.94961175e-02 -7.47693121e-01 -1.87368065e-01 2.00295579e-02
4.51778233e-01 8.76253963e-01 1.02474177e+00 -1.01034975e+00
-4.88560408e-01 -9.03364599e-01 1.63405389e-01 5.70615292e-01
-5.97412186e-03 3.82272482e-01 -9.69086170e-01 -3.63526523e-01
2.32847467e-01 3.77496295e-02 1.90903217e-01 -1.02539770e-01
1.32969403e+00 -2.35285416e-01 -3.87373745e-01 8.77553105e-01
1.36019993e+00 3.50180417e-01 9.87282336e-01 3.70839417e-01
5.67096412e-01 9.65950266e-02 6.02112353e-01 1.63973391e-01
-5.65807700e-01 6.71909988e-01 1.16522741e+00 -3.63980681e-02
1.24993369e-01 -2.01770887e-01 1.49799705e-01 2.75066285e-03
-4.00845893e-02 -2.96042711e-01 -1.07279551e+00 2.26422057e-01
-1.26851821e+00 -1.16029489e+00 -2.79630691e-01 2.07900739e+00
3.43370408e-01 7.82239497e-01 3.57258469e-02 6.22539818e-01
8.18492174e-01 3.26926500e-01 -5.37432671e-01 -3.93407524e-01
-7.81492367e-02 5.03210187e-01 9.18917358e-01 2.87355185e-01
-1.32153726e+00 9.62323546e-01 6.44090891e+00 9.18931186e-01
-1.19519806e+00 -3.29020202e-01 2.15261549e-01 -1.40765011e-01
2.51797915e-01 -1.95905879e-01 -7.77083457e-01 6.67107940e-01
7.68267095e-01 -2.57178873e-01 6.19564950e-02 1.09180820e+00
-9.70192552e-02 3.95911127e-01 -7.98754692e-01 8.48111391e-01
3.70679311e-02 -1.30557728e+00 2.52543151e-01 3.93426239e-01
5.51767051e-01 7.77228102e-02 2.43741021e-01 1.89158767e-01
4.02251512e-01 -1.12524843e+00 5.83958745e-01 4.30800766e-02
5.74576974e-01 -1.41606736e+00 9.95130718e-01 6.57458723e-01
-8.42819750e-01 2.27782857e-02 -6.00218236e-01 1.46007344e-01
6.84960112e-02 4.60249096e-01 -8.18731070e-01 8.11965644e-01
8.12011540e-01 3.74596953e-01 -5.41167796e-01 1.14823687e+00
-5.97539544e-01 7.19416916e-01 -6.41491771e-01 4.05163690e-02
3.08126092e-01 1.87558323e-01 1.09845853e+00 8.96828532e-01
6.99125752e-02 1.50625065e-01 -3.11246850e-02 4.01583523e-01
-6.64480031e-02 -2.67536223e-01 -9.37905967e-01 3.86847705e-01
7.46400774e-01 9.15564597e-01 -3.45278680e-01 2.13715285e-01
-1.39189675e-01 6.06034219e-01 -1.27158508e-01 1.55714050e-01
-1.24851120e+00 -5.43919802e-01 1.25015390e+00 8.41686353e-02
2.74954289e-01 -4.30496454e-01 -3.71420324e-01 -7.52298057e-01
-1.99142308e-03 -1.18656707e+00 2.26089552e-01 -3.23785067e-01
-1.59630609e+00 8.14841390e-01 7.22543523e-02 -1.79018664e+00
-1.48945823e-01 -6.57881558e-01 -8.78405213e-01 9.14899230e-01
-1.06004345e+00 -1.11330342e+00 -2.67860383e-01 8.31101477e-01
4.01484706e-02 -5.96828043e-01 9.89117265e-01 4.56995368e-01
-3.64126861e-01 1.00215495e+00 -5.01608215e-02 3.95791262e-01
5.90626895e-01 -8.35400462e-01 1.02528882e+00 1.16301262e+00
9.61057469e-02 4.34997052e-01 8.29462588e-01 -6.71679378e-01
-1.07578254e+00 -1.17526078e+00 3.77467215e-01 -6.74205363e-01
6.31388307e-01 -6.44578815e-01 -9.30143833e-01 3.90563577e-01
-1.42688751e-01 9.78500471e-02 6.87033594e-01 -3.25329036e-01
-5.68512201e-01 -6.21885844e-02 -1.78290343e+00 8.18785489e-01
9.36691523e-01 -3.92724752e-01 -5.86800396e-01 3.02262008e-01
7.96627343e-01 -7.07462907e-01 -9.03429270e-01 6.91988647e-01
1.51403338e-01 -8.25636387e-01 1.53158617e+00 -7.56751895e-01
1.25675336e-01 -4.72713411e-01 -3.35491717e-01 -1.15051901e+00
-3.74041319e-01 -6.77556872e-01 -2.01863870e-01 1.17020226e+00
3.12110037e-01 -8.23814213e-01 1.02371347e+00 3.58517587e-01
-2.84653232e-02 -4.32003945e-01 -1.27257931e+00 -8.62196684e-01
5.16688585e-01 -6.70739472e-01 8.63809288e-01 9.56806660e-01
-5.54010212e-01 -1.69515759e-01 -3.10749918e-01 9.75403070e-01
8.94349575e-01 -5.13892233e-01 1.34933388e+00 -1.29574287e+00
2.96614598e-02 -2.36585572e-01 -1.21706879e+00 -5.31833410e-01
1.07511729e-01 -7.99863458e-01 -5.58902860e-01 -8.81307244e-01
-6.33958399e-01 -6.24289572e-01 -1.61981940e-01 3.63478005e-01
-1.88179016e-01 5.75441360e-01 6.38771832e-01 9.42738503e-02
2.33957425e-01 3.27885181e-01 9.30743456e-01 -3.41220409e-01
9.48889554e-02 6.40471160e-01 -4.77554590e-01 7.93200254e-01
9.51690435e-01 -9.88297880e-01 -3.07351798e-01 -2.40079626e-01
-2.12831572e-01 -4.67348665e-01 4.86550093e-01 -1.55455732e+00
6.63011745e-02 -1.19489588e-01 5.90333760e-01 -9.29883599e-01
3.52500081e-01 -1.34061003e+00 3.19013357e-01 6.90219402e-01
1.71380773e-01 7.67954141e-02 7.96854556e-01 5.56886137e-01
-1.30833194e-01 -1.93407729e-01 1.12731755e+00 1.67975724e-01
-2.96528131e-01 5.37538409e-01 -1.71852887e-01 -7.04837441e-02
1.44733381e+00 -2.57041574e-01 -4.27712858e-01 -2.35520542e-01
-5.89902282e-01 2.73960584e-04 5.90314627e-01 6.78322852e-01
5.45258403e-01 -1.35877109e+00 -7.49145329e-01 5.55550575e-01
-1.07842892e-01 6.75896928e-02 2.39532173e-01 1.43794671e-01
-8.05615246e-01 3.34510170e-02 -5.22942126e-01 -6.76920295e-01
-1.36218500e+00 7.37185717e-01 4.03022736e-01 -1.38219565e-01
-5.66017687e-01 8.17081630e-01 -1.89141273e-01 -4.74649370e-01
4.94660139e-01 7.93168545e-02 3.70205520e-03 -4.75581139e-01
3.50621164e-01 4.40526783e-01 3.18843246e-01 -7.04657197e-01
-7.41813600e-01 7.02425599e-01 -1.79249808e-01 2.13797599e-01
1.23829639e+00 5.86753070e-01 6.31018430e-02 -2.04312503e-01
1.22297060e+00 5.25164545e-01 -8.80787015e-01 2.09953502e-01
-3.37707669e-01 -8.73223186e-01 -3.24158072e-01 -6.94647253e-01
-1.12467945e+00 8.05056453e-01 6.35855019e-01 5.38599372e-01
1.04123688e+00 -3.27204168e-01 7.25799561e-01 1.80867225e-01
4.88174677e-01 -4.02147055e-01 -1.46345198e-01 7.38093197e-01
8.99722934e-01 -8.69258225e-01 9.93170291e-02 -8.24744880e-01
-3.37940514e-01 9.23711240e-01 6.65700853e-01 -8.52030337e-01
1.10306144e+00 4.84865427e-01 2.29616016e-01 8.19333196e-02
-3.09012532e-01 5.20476401e-01 2.65914321e-01 1.07404029e+00
-3.55389982e-01 -2.10353836e-01 1.52105957e-01 3.10207397e-01
-5.44967651e-01 -6.09555542e-01 3.06789815e-01 8.29974890e-01
-6.60617724e-02 -1.30614614e+00 -7.38244712e-01 1.46089777e-01
-7.08117604e-01 1.68309763e-01 -7.50111282e-01 1.22761714e+00
1.10975772e-01 9.74484026e-01 -1.73341125e-01 -8.50862920e-01
8.95926416e-01 -1.39110193e-01 1.73438087e-01 -3.04557830e-01
-1.07356131e+00 -3.22216779e-01 -1.71230823e-01 -6.08881116e-01
8.69326945e-03 -4.32184339e-01 -1.01583171e+00 -8.54319811e-01
-3.98430884e-01 9.98016298e-02 7.44428813e-01 6.41827762e-01
4.82062787e-01 5.11114717e-01 1.04450786e+00 -1.03871357e+00
-9.69134927e-01 -7.77731597e-01 -4.61131543e-01 6.74182832e-01
1.85296670e-01 -8.70406389e-01 -9.87430573e-01 -5.09669662e-01] | [7.696038246154785, -4.481686592102051] |
334e33e0-0f20-47af-9041-f3f9fc468bd7 | answering-ambiguous-questions-through | 2011.13137 | null | https://arxiv.org/abs/2011.13137v2 | https://arxiv.org/pdf/2011.13137v2.pdf | Answering Ambiguous Questions through Generative Evidence Fusion and Round-Trip Prediction | In open-domain question answering, questions are highly likely to be ambiguous because users may not know the scope of relevant topics when formulating them. Therefore, a system needs to find possible interpretations of the question, and predict one or multiple plausible answers. When multiple plausible answers are found, the system should rewrite the question for each answer to resolve the ambiguity. In this paper, we present a model that aggregates and combines evidence from multiple passages to adaptively predict a single answer or a set of question-answer pairs for ambiguous questions. In addition, we propose a novel round-trip prediction approach to iteratively generate additional interpretations that our model fails to find in the first pass, and then verify and filter out the incorrect question-answer pairs to arrive at the final disambiguated output. Our model, named Refuel, achieves a new state-of-the-art performance on the AmbigQA dataset, and shows competitive performance on NQ-Open and TriviaQA. The proposed round-trip prediction is a model-agnostic general approach for answering ambiguous open-domain questions, which improves our Refuel as well as several baseline models. We release source code for our models and experiments at https://github.com/amzn/refuel-open-domain-qa. | ['Bing Xiang', 'Andrew O. Arnold', 'Ramesh Nallapati', 'Dejiao Zhang', 'Feng Nan', 'Zhiguo Wang', 'Cicero Nogueira dos santos', 'Patrick Ng', 'Henghui Zhu', 'Yifan Gao'] | 2020-11-26 | null | https://aclanthology.org/2021.acl-long.253 | https://aclanthology.org/2021.acl-long.253.pdf | acl-2021-5 | ['triviaqa'] | ['miscellaneous'] | [ 1.14138322e-02 4.71691072e-01 2.42953405e-01 -5.77078581e-01
-1.75275433e+00 -1.15250349e+00 2.10051641e-01 2.11994097e-01
-1.94497809e-01 9.63697851e-01 3.38968933e-01 -7.57345676e-01
-1.66225910e-01 -8.91511559e-01 -6.52617812e-01 -7.15158507e-02
6.89922154e-01 1.25314510e+00 8.28272045e-01 -7.82735407e-01
1.17549993e-01 -2.73472071e-01 -1.26177251e+00 7.76137531e-01
1.63406789e+00 7.31799543e-01 4.36874211e-01 9.19635773e-01
-6.62629843e-01 7.55637407e-01 -7.14892626e-01 -7.80776978e-01
7.21628591e-03 -6.32433891e-01 -1.63019598e+00 -2.66115874e-01
6.50171936e-01 -3.74968350e-01 1.85489282e-02 1.01658535e+00
3.55252475e-01 2.21840501e-01 4.47628170e-01 -9.32330251e-01
-9.72348571e-01 6.15988016e-01 1.74077488e-02 4.44307715e-01
8.86784256e-01 1.94220319e-01 1.54443944e+00 -1.08159137e+00
3.61879081e-01 1.39431059e+00 2.12889209e-01 6.31485522e-01
-7.80163705e-01 -4.69962835e-01 3.28737020e-01 6.05378091e-01
-1.01000822e+00 -2.12649867e-01 3.57131124e-01 3.19468230e-02
7.56236136e-01 4.75510120e-01 1.80447519e-01 7.26384580e-01
2.99875997e-02 8.06165278e-01 8.69522989e-01 -5.48033059e-01
1.53360099e-01 -5.33518232e-02 5.87180734e-01 5.86930573e-01
-7.63939768e-02 -5.97348034e-01 -2.12413535e-01 -4.91042107e-01
1.74601287e-01 -5.03903151e-01 -3.96708667e-01 2.75644958e-01
-1.03095353e+00 8.96635890e-01 4.05966163e-01 7.87317157e-02
-4.01184678e-01 -3.73712182e-01 -1.29847126e-02 5.58575153e-01
1.14872083e-01 8.33264589e-01 -7.66145229e-01 -1.60467133e-01
-4.73841667e-01 8.51444185e-01 1.14298177e+00 8.00958812e-01
8.73520017e-01 -7.62030184e-01 -5.92628181e-01 1.03257346e+00
2.57718056e-01 4.50173616e-01 3.44812840e-01 -1.34581065e+00
8.32976282e-01 7.19099939e-01 5.00703752e-01 -7.43290246e-01
-2.19965518e-01 -1.65082514e-01 -3.07793647e-01 -5.84677815e-01
6.90793216e-01 -2.33826593e-01 -7.22828329e-01 1.59107089e+00
6.41791046e-01 -8.23740065e-02 3.25837940e-01 1.10307193e+00
1.21647990e+00 1.08738267e+00 -3.49493958e-02 2.43101530e-02
1.76907146e+00 -1.29512286e+00 -8.59279931e-01 -5.07643640e-01
4.52857137e-01 -1.00422299e+00 1.35721338e+00 2.70048052e-01
-1.13118744e+00 -5.16533077e-01 -6.03428662e-01 -6.46139979e-01
-9.68317017e-02 -6.43571541e-02 -2.96387984e-03 2.88076788e-01
-8.76869440e-01 -7.78177753e-02 -2.78272420e-01 -1.16216742e-01
-7.45518208e-02 6.62722662e-02 1.55118287e-01 -4.98075753e-01
-1.79533565e+00 8.96732807e-01 2.96454102e-01 1.55004850e-02
-5.52864313e-01 -5.65072298e-01 -7.71440089e-01 1.63946882e-01
7.35140145e-01 -1.11425781e+00 2.00786304e+00 -5.13044536e-01
-1.29453492e+00 5.11936605e-01 -8.95478487e-01 -3.82965505e-01
1.83025643e-01 -4.04294163e-01 -5.50184667e-01 3.22217047e-01
5.76792598e-01 8.47395837e-01 5.03005743e-01 -1.15233696e+00
-9.26196992e-01 -2.53762156e-01 7.94608533e-01 3.81437331e-01
2.01729253e-01 -2.72621326e-02 -3.93361181e-01 -1.76107317e-01
4.36558604e-01 -7.34055817e-01 -1.99426070e-01 -3.66326243e-01
-2.48720735e-01 -8.95361364e-01 5.03261864e-01 -9.01378393e-01
1.44297421e+00 -1.69426715e+00 -9.60008055e-02 -1.33635327e-01
2.42099628e-01 3.54687721e-01 -5.03605485e-01 4.68508482e-01
3.63408208e-01 1.77969053e-01 -4.70444173e-01 1.14738896e-01
5.81703223e-02 5.27616024e-01 -8.15730512e-01 -4.07336682e-01
4.78072524e-01 9.76214170e-01 -1.14311242e+00 -5.05703151e-01
-2.93964177e-01 -2.56810844e-01 -7.07624257e-01 5.31418562e-01
-1.07102752e+00 4.20770466e-01 -6.60717666e-01 4.37397718e-01
7.90236652e-01 -4.85559106e-01 -4.48065698e-02 6.86198249e-02
2.12002963e-01 1.08117020e+00 -1.15133119e+00 1.39808631e+00
-5.70849299e-01 2.90626407e-01 -2.31805611e-02 -5.15815854e-01
7.87640810e-01 3.87280226e-01 -2.69483477e-01 -8.62950802e-01
-2.07175300e-01 6.30819201e-01 9.60621312e-02 -8.49194169e-01
7.73770154e-01 2.53756158e-02 -2.20339715e-01 5.35455465e-01
4.84915860e-02 -4.62492079e-01 4.79636490e-01 4.69035566e-01
1.05966258e+00 -6.85349107e-02 1.78384185e-01 -1.01447619e-01
9.44179356e-01 3.23372692e-01 5.57955027e-01 9.50274348e-01
-1.50632620e-01 7.11197436e-01 6.26350701e-01 -3.12814295e-01
-5.33851624e-01 -1.28272843e+00 1.09598652e-01 1.27965248e+00
3.63689870e-01 -5.26176751e-01 -8.05858254e-01 -9.98641312e-01
-2.66223788e-01 1.13599765e+00 -2.52719253e-01 1.61497191e-01
-7.90166199e-01 -1.26613706e-01 4.61632997e-01 9.65676904e-02
4.59574759e-01 -9.55816209e-01 -1.59230351e-01 4.04726833e-01
-1.22121668e+00 -9.90948677e-01 -5.81682801e-01 -1.36867240e-01
-6.39645755e-01 -1.33696806e+00 -4.21025574e-01 -1.01969457e+00
4.34518188e-01 3.06553066e-01 1.47871351e+00 2.57605940e-01
6.00939751e-01 1.51666299e-01 -6.62554085e-01 -3.48014176e-01
-6.42742038e-01 3.25487584e-01 -4.24863368e-01 -1.92503601e-01
4.97383505e-01 -1.15377136e-01 -6.07348859e-01 6.12210214e-01
-9.99466479e-01 -2.16841862e-01 2.60891587e-01 9.20265198e-01
6.47155881e-01 -3.82039636e-01 1.15262282e+00 -9.13589954e-01
1.22946382e+00 -7.78592587e-01 -4.06711787e-01 6.94728613e-01
-1.78942889e-01 2.87977666e-01 9.09746110e-01 -1.21102877e-01
-1.26778936e+00 -4.30587411e-01 -7.38743842e-01 2.74042739e-03
-2.11735293e-01 7.42210805e-01 -2.39845425e-01 4.81666654e-01
8.28343332e-01 4.42754962e-02 -2.87430346e-01 -5.09569168e-01
6.58652127e-01 8.57427835e-01 5.18810272e-01 -8.30562770e-01
7.57063270e-01 -8.73019267e-03 -7.01564133e-01 -4.26055968e-01
-1.58804691e+00 -6.83204234e-01 -3.49886030e-01 6.94088265e-02
6.87632799e-01 -8.14821482e-01 -3.13993841e-01 -3.04328036e-02
-1.69355667e+00 -9.57518220e-02 -1.04667835e-01 2.37514582e-02
-2.88250715e-01 5.40503442e-01 -4.91417527e-01 -5.61870337e-01
-3.31076950e-01 -1.08416653e+00 8.39450419e-01 6.15340769e-01
-5.41847527e-01 -7.83222795e-01 1.28283665e-01 1.05141377e+00
3.33416224e-01 -5.64413786e-01 1.21194720e+00 -1.15484917e+00
-6.71959579e-01 1.47838369e-01 -4.58684191e-02 2.69896686e-01
1.38213679e-01 -3.08141857e-01 -6.79175675e-01 1.11634806e-01
8.03895518e-02 -5.81849337e-01 7.25317478e-01 -2.69133337e-02
1.00997305e+00 -6.62245214e-01 1.00587487e-01 -1.49727628e-01
8.89906287e-01 5.67130521e-02 5.88351667e-01 2.43835956e-01
1.75396711e-01 6.92889214e-01 1.03189301e+00 -9.37540270e-03
1.08994818e+00 3.65542620e-01 2.55532235e-01 3.68093669e-01
2.01047361e-02 -4.21379417e-01 2.02504218e-01 9.48674858e-01
5.60397804e-01 -6.00846350e-01 -9.12035704e-01 1.04429495e+00
-1.80230808e+00 -8.00807834e-01 -5.53571582e-01 1.86615324e+00
1.09834230e+00 7.37283304e-02 -7.96447322e-02 -1.34876087e-01
6.87699497e-01 -6.58956096e-02 -5.12515962e-01 -6.40316546e-01
-1.20492473e-01 4.52918738e-01 -3.61269534e-01 1.07872486e+00
-6.15388751e-01 1.04176188e+00 5.59531307e+00 7.43495047e-01
-4.55037296e-01 1.98388740e-01 5.92651844e-01 3.25754225e-01
-9.35665250e-01 3.03513348e-01 -9.33284760e-01 2.47471228e-01
9.47789669e-01 -2.69908130e-01 2.13794276e-01 4.62660432e-01
1.43050358e-01 -3.15390080e-01 -9.82702494e-01 3.49010050e-01
1.26299575e-01 -1.19945443e+00 4.55981344e-01 -4.88664657e-01
6.66399181e-01 -1.46828264e-01 -1.30340084e-01 6.45728052e-01
4.04849201e-01 -8.15683007e-01 5.58405340e-01 2.93523192e-01
1.35322765e-01 -7.10910022e-01 8.97246301e-01 7.66227484e-01
-8.59569192e-01 -2.41184130e-01 -4.03649002e-01 -3.14723492e-01
3.85052770e-01 3.50013793e-01 -8.99267495e-01 6.63932025e-01
6.01579487e-01 -6.88236803e-02 -6.24995708e-01 9.59318578e-01
-8.38103294e-01 9.17921007e-01 -3.07749450e-01 -3.69080365e-01
4.79029357e-01 -6.14812784e-02 5.57395577e-01 8.09425235e-01
2.91344494e-01 7.49469101e-01 8.93136263e-02 9.48821962e-01
-1.68017387e-01 1.78644538e-01 5.69429919e-02 3.33347768e-01
8.10442030e-01 1.03011894e+00 -4.00747396e-02 -5.06439567e-01
-4.96762365e-01 8.87824714e-01 5.83400548e-01 2.13342950e-01
-7.59980679e-01 -5.49918890e-01 4.60420400e-01 3.19289230e-02
2.27597401e-01 8.33132789e-02 -1.03360698e-01 -1.23373806e+00
4.40381914e-01 -1.37806571e+00 1.00661147e+00 -1.09874582e+00
-1.47570908e+00 7.33404994e-01 -1.69434920e-01 -9.87896681e-01
-5.03131628e-01 -3.80916059e-01 -6.11720741e-01 1.15785420e+00
-1.86155808e+00 -6.55647397e-01 -1.35890305e-01 4.26711917e-01
9.94482577e-01 3.72974873e-01 7.94631660e-01 9.85943675e-02
-1.69192195e-01 4.24132288e-01 -1.85065389e-01 8.60740915e-02
9.00137246e-01 -1.36570430e+00 6.16611123e-01 9.03437674e-01
2.38373309e-01 7.68207073e-01 7.66301513e-01 -5.37629306e-01
-1.05719531e+00 -9.39313054e-01 1.61616170e+00 -7.33669639e-01
5.85649073e-01 -1.72289871e-02 -1.52148044e+00 5.86786509e-01
5.19767463e-01 -3.63213569e-01 7.71629751e-01 1.03493169e-01
-2.12362111e-01 2.24878099e-02 -1.04052281e+00 6.54569030e-01
5.96174598e-01 -5.14750004e-01 -1.59210968e+00 4.43226248e-01
1.24532044e+00 -7.06075907e-01 -5.21321893e-01 3.52114707e-01
1.39994234e-01 -8.28100264e-01 6.59300566e-01 -6.30654335e-01
4.61759061e-01 -5.73752820e-01 -1.51891395e-01 -1.27701199e+00
-2.78278351e-01 -6.15974903e-01 -2.03145355e-01 1.11933362e+00
1.10615182e+00 -6.71315312e-01 4.25015122e-01 7.55775869e-01
-3.91587317e-01 -8.06491315e-01 -1.16338193e+00 -3.49956661e-01
3.14613998e-01 -3.73643637e-01 8.32318902e-01 5.80167770e-01
-3.77882505e-03 7.59460449e-01 2.29211405e-01 6.32940948e-01
1.03689983e-01 5.01315415e-01 6.39800549e-01 -8.64021122e-01
-2.45796233e-01 -6.18032515e-02 3.99280667e-01 -1.96162760e+00
1.54191270e-01 -6.68027759e-01 3.09808135e-01 -2.11171174e+00
-1.88627452e-01 -4.21621799e-01 4.29168306e-02 3.10575008e-01
-8.34394813e-01 -6.45808652e-02 1.04569219e-01 -3.80170019e-03
-1.01489568e+00 5.85698664e-01 1.53412664e+00 -1.07088596e-01
-2.48345152e-01 2.64038146e-01 -1.14490139e+00 5.43989122e-01
8.38976085e-01 -4.09044564e-01 -4.02974784e-01 -8.60649467e-01
4.63460922e-01 3.98929954e-01 3.23427647e-01 -7.05782473e-01
4.02972072e-01 -3.06065947e-01 -5.18584512e-02 -8.24960291e-01
2.86313713e-01 -4.57128376e-01 -5.19843042e-01 2.88322419e-01
-5.75709462e-01 3.79982352e-01 9.75035578e-02 4.06027824e-01
-5.47554135e-01 -6.34319127e-01 4.06971991e-01 -3.19100291e-01
-6.37385786e-01 6.71336949e-02 -4.60452408e-01 8.07055175e-01
4.84841585e-01 1.61676243e-01 -6.45017982e-01 -7.57433712e-01
-6.92058921e-01 1.05256879e+00 5.70193641e-02 6.91155612e-01
7.14823544e-01 -1.03708398e+00 -1.05499780e+00 -2.34617755e-01
2.79232949e-01 4.62997496e-01 4.39826220e-01 3.34328502e-01
-4.35907900e-01 7.66219079e-01 3.59164953e-01 -4.90624070e-01
-1.03011227e+00 2.00715199e-01 5.06502748e-01 -5.47726452e-01
-4.70484085e-02 9.80651736e-01 1.30362257e-01 -9.73117471e-01
-1.72333181e-01 -5.51359057e-01 -4.23518687e-01 -8.03526640e-02
7.13961303e-01 1.73484713e-01 2.27841824e-01 -4.12520647e-01
-1.60993278e-01 2.16925681e-01 -1.94971591e-01 -1.87399760e-01
7.09802151e-01 -5.85527956e-01 -3.75293225e-01 2.63295650e-01
7.09251404e-01 1.72519132e-01 -7.93192685e-01 -5.23100495e-01
2.95516383e-02 -4.07406956e-01 -5.20485997e-01 -1.24219382e+00
-2.91492164e-01 7.99691498e-01 1.10667702e-02 4.02513653e-01
1.12339401e+00 4.77156371e-01 1.26791060e+00 8.12275589e-01
2.19424278e-01 -8.55756283e-01 2.54908204e-01 1.20464861e+00
1.13126850e+00 -1.13342428e+00 -5.48590302e-01 -5.67208648e-01
-6.27316952e-01 1.00582612e+00 1.19992125e+00 2.83335745e-01
8.35189149e-02 -3.88722718e-01 4.71036732e-01 -9.78549197e-02
-1.16487038e+00 -2.05124065e-01 3.39902848e-01 3.57194006e-01
3.31139535e-01 6.38312427e-03 -5.09365380e-01 1.05151594e+00
-6.52471185e-01 -3.37441295e-01 6.79367006e-01 7.31904566e-01
-9.44821060e-01 -1.24165976e+00 -5.86330950e-01 4.42285389e-01
-4.49415863e-01 -2.58517385e-01 -5.99061668e-01 2.58463621e-01
3.92592661e-02 1.63482010e+00 -1.95552874e-02 -1.24472920e-02
5.42476177e-01 4.00684237e-01 2.91583806e-01 -8.15692842e-01
-6.36191368e-01 -4.10557985e-01 3.90953690e-01 -1.87901393e-01
-1.48374438e-01 -4.01490897e-01 -1.47152126e+00 -3.61532122e-02
-4.24015582e-01 7.95899987e-01 5.66341057e-02 1.28286684e+00
5.76939046e-01 1.86236307e-01 5.14174759e-01 1.72715411e-01
-7.54672945e-01 -9.53885853e-01 1.33747354e-01 2.54415572e-01
5.52288830e-01 -2.09923968e-01 -2.39458412e-01 -1.91619545e-01] | [11.431597709655762, 8.03747844696045] |
f9d4248b-d28c-41ef-b855-819f1160ba14 | learning-localized-generative-models-for-3d | null | null | https://openreview.net/forum?id=SJeXSo09FQ | https://openreview.net/pdf?id=SJeXSo09FQ | Learning Localized Generative Models for 3D Point Clouds via Graph Convolution | Point clouds are an important type of geometric data and have widespread use in computer graphics and vision. However, learning representations for point clouds is particularly challenging due to their nature as being an unordered collection of points irregularly distributed in 3D space. Graph convolution, a generalization of the convolution operation for data defined over graphs, has been recently shown to be very successful at extracting localized features from point clouds in supervised or semi-supervised tasks such as classification or segmentation. This paper studies the unsupervised problem of a generative model exploiting graph convolution. We focus on the generator of a GAN and define methods for graph convolution when the graph is not known in advance as it is the very output of the generator. The proposed architecture learns to generate localized features that approximate graph embeddings of the output geometry. We also study the problem of defining an upsampling layer in the graph-convolutional generator, such that it learns to exploit a self-similarity prior on the data distribution to sample more effectively. | ['Giulia Fracastoro', 'Enrico Magli', 'Diego Valsesia'] | 2019-05-01 | null | null | null | iclr-2019-5 | ['point-cloud-generation'] | ['computer-vision'] | [ 2.64542013e-01 4.98963982e-01 1.68333456e-01 -4.27164555e-01
-3.59703898e-01 -5.72348297e-01 7.59535730e-01 2.06905529e-01
-8.33402015e-03 2.92911977e-01 -1.10607229e-01 -2.43470311e-01
2.39413530e-01 -1.40830672e+00 -1.09673488e+00 -7.28572190e-01
-1.50462151e-01 9.90440965e-01 -1.87813044e-01 5.46802506e-02
1.43490687e-01 1.10386741e+00 -1.51953590e+00 -2.35777639e-04
7.42551804e-01 7.00537086e-01 2.40992323e-01 8.20455551e-01
-4.45992619e-01 2.65834868e-01 -6.35681152e-01 -1.67722687e-01
3.08040202e-01 -5.32890439e-01 -5.58796108e-01 4.78346199e-01
7.16259360e-01 9.98855457e-02 -1.92002639e-01 1.10701072e+00
3.11772287e-01 1.73250228e-01 1.07161891e+00 -1.41248167e+00
-9.05827343e-01 2.37390071e-01 -2.84974188e-01 -1.84482008e-01
3.10349539e-02 2.07576286e-02 9.99671221e-01 -7.35914886e-01
6.38257802e-01 1.19907856e+00 6.06435180e-01 5.02053142e-01
-1.40668690e+00 -4.09801334e-01 -1.47036746e-01 -3.10104340e-01
-1.27304327e+00 2.03543827e-01 1.08068764e+00 -6.95446491e-01
8.21970463e-01 8.26673582e-02 9.14801478e-01 7.98798263e-01
8.40286091e-02 5.73646903e-01 7.96760917e-01 -3.64205509e-01
5.56748450e-01 -1.07115604e-01 -3.22344810e-01 8.30368757e-01
2.07967624e-01 2.24996563e-02 -5.29827848e-02 -3.37136418e-01
1.38728368e+00 2.74182916e-01 -2.71390468e-01 -8.52579296e-01
-8.06373477e-01 1.33184254e+00 1.06828725e+00 2.46741429e-01
-5.90010762e-01 5.69163442e-01 -8.15528110e-02 2.23130792e-01
6.77785397e-01 4.12329018e-01 2.50567216e-02 2.10623637e-01
-8.29394758e-01 3.14733058e-01 7.34342933e-01 1.36163521e+00
1.25882149e+00 2.85576671e-01 1.16562378e-02 5.62565207e-01
2.35711277e-01 4.46949482e-01 2.18426570e-01 -5.48265040e-01
2.74838716e-01 8.92135382e-01 -2.10819945e-01 -7.82254875e-01
-5.76913916e-02 -5.29425621e-01 -1.03825521e+00 4.26887989e-01
2.40834832e-01 4.07313323e-03 -1.26121008e+00 1.49760520e+00
3.14541489e-01 5.64992189e-01 -1.25761673e-01 7.01326191e-01
6.77925289e-01 7.15816796e-01 -1.30976334e-01 3.76902431e-01
7.59989560e-01 -4.72024173e-01 -3.76472250e-02 -1.86968774e-01
3.80094230e-01 -3.76158267e-01 9.38395083e-01 2.57673245e-02
-8.62779975e-01 -6.40647948e-01 -8.81982446e-01 -5.00910357e-02
-4.82858032e-01 -3.04381344e-02 7.20515251e-01 4.13618952e-01
-1.27548134e+00 8.00670505e-01 -7.93983996e-01 -3.97834569e-01
8.93579781e-01 4.07682568e-01 -2.13665709e-01 -2.04737037e-01
-5.12901187e-01 4.57229704e-01 2.06497073e-01 -1.19753718e-01
-7.26432979e-01 -6.63861394e-01 -1.07344246e+00 2.56743789e-01
-1.59115329e-01 -1.00475764e+00 8.10180902e-01 -9.78932023e-01
-1.28943181e+00 1.12541962e+00 1.13325380e-01 -5.09829700e-01
3.54287535e-01 1.43797517e-01 1.05095029e-01 -2.93292459e-02
1.40300199e-01 6.45322025e-01 1.22571862e+00 -1.22736931e+00
-3.61958474e-01 -5.26561379e-01 1.92183759e-02 1.85618982e-01
2.15085939e-01 -7.14945853e-01 -1.71101034e-01 -7.97060430e-01
2.46424556e-01 -1.04647982e+00 -5.29897034e-01 -2.06371211e-02
-4.59343702e-01 -3.93014133e-01 9.27013874e-01 -2.45733842e-01
3.28011930e-01 -2.11228871e+00 1.62457496e-01 5.34177780e-01
3.60619575e-01 4.28175330e-02 1.02049317e-02 5.50874233e-01
-1.92572549e-01 1.52300417e-01 -4.50081706e-01 -3.56755227e-01
-6.51771855e-03 4.99196649e-01 -3.20704728e-01 6.27355933e-01
5.97354650e-01 1.05725658e+00 -1.06918180e+00 6.82121068e-02
4.16822106e-01 6.56748116e-01 -5.40850759e-01 4.95930374e-01
-6.95074379e-01 6.72913909e-01 -6.00869238e-01 1.51289135e-01
7.50659287e-01 -4.25194770e-01 -2.76706874e-01 1.06991623e-02
9.28918198e-02 5.63996471e-02 -8.13251197e-01 1.88120174e+00
-5.66128671e-01 4.86057639e-01 -1.79347575e-01 -1.21060145e+00
1.27872694e+00 2.20553100e-01 4.20216322e-01 -1.17531002e-01
1.36510521e-01 8.47334564e-02 -2.05310091e-01 1.93658993e-02
2.05764204e-01 -3.89957756e-01 6.11024313e-02 3.71016562e-01
3.41147959e-01 -8.71851683e-01 -2.46699676e-01 4.62001413e-02
1.15314496e+00 -8.25886428e-02 1.31373554e-01 -2.68157631e-01
1.17992401e-01 4.04781848e-02 -3.91636044e-02 5.35538137e-01
7.13132858e-01 1.06295824e+00 4.00755495e-01 -4.50653940e-01
-1.15994918e+00 -1.17647648e+00 1.03396595e-01 3.21226746e-01
-1.25472546e-01 -2.02720702e-01 -8.56686294e-01 -7.79605329e-01
1.97564483e-01 8.13035369e-01 -6.49297774e-01 -2.19167233e-01
-4.25501257e-01 -2.41096109e-01 4.00313288e-02 6.78304374e-01
3.38324338e-01 -1.08317077e+00 -5.21138012e-01 8.21437165e-02
5.24024487e-01 -1.08404255e+00 -4.15123105e-01 2.79374331e-01
-1.15170276e+00 -1.05805635e+00 -7.43334651e-01 -9.80013728e-01
1.28947866e+00 1.37777403e-02 1.43343949e+00 1.75894648e-01
-2.89703459e-01 6.55921102e-01 -1.98554367e-01 -5.71414590e-01
-5.60458660e-01 9.52992365e-02 -4.89998877e-01 2.62814671e-01
3.78914028e-01 -9.07599330e-01 -3.17943275e-01 -4.94381338e-02
-1.07969129e+00 1.40424699e-01 4.45263982e-01 8.59982193e-01
8.81893277e-01 1.36083245e-01 1.00996949e-01 -1.29225826e+00
6.51751935e-01 -5.62569082e-01 -7.41894066e-01 -1.30622864e-01
-1.57026321e-01 3.95487458e-01 8.10542047e-01 -2.08058417e-01
-3.33571404e-01 2.38706931e-01 -1.98339641e-01 -9.91442382e-01
-4.11953270e-01 3.88983697e-01 -1.17300756e-01 -3.08584481e-01
6.43078744e-01 6.67434707e-02 1.78495675e-01 -4.02011275e-01
5.51730633e-01 2.56076664e-01 3.60362232e-01 -5.93328059e-01
1.14833760e+00 4.66210306e-01 4.74973828e-01 -1.12799931e+00
-3.48545283e-01 -5.32557189e-01 -6.80890381e-01 -1.07684648e-02
9.60380971e-01 -6.70633614e-01 -3.42722684e-01 2.61695445e-01
-1.21604252e+00 -5.44304013e-01 -9.20483887e-01 2.69030243e-01
-9.24598992e-01 -1.71508864e-01 -1.13102876e-01 -5.86674213e-01
-2.13649184e-01 -8.16722095e-01 1.42694354e+00 2.30885997e-01
4.48291339e-02 -1.36889791e+00 6.72904104e-02 -4.36883181e-01
2.94861227e-01 8.39406013e-01 1.12214243e+00 -5.42106569e-01
-8.55547369e-01 -3.74637306e-01 -2.44498905e-02 5.57608485e-01
3.30983311e-01 -2.11222723e-01 -9.35229361e-01 -3.27883899e-01
7.03436509e-02 -6.58646002e-02 6.06542706e-01 4.99272674e-01
1.39392972e+00 -3.17630738e-01 -2.86935627e-01 9.11439002e-01
1.74626994e+00 -1.82395190e-01 5.99239111e-01 -3.54048967e-01
1.09438443e+00 5.00379443e-01 5.10365516e-02 2.67452419e-01
8.45300332e-02 3.88216019e-01 7.68216491e-01 -3.01336080e-01
-1.53631866e-01 -7.14858472e-01 -1.76634565e-01 4.98730868e-01
-2.43715107e-01 -1.53962612e-01 -9.76067007e-01 5.10195673e-01
-1.60146761e+00 -6.62525415e-01 -1.07630715e-01 2.26205492e+00
2.46336639e-01 -6.26951233e-02 6.50552362e-02 -7.07944250e-03
8.20451021e-01 2.64407080e-02 -5.41227937e-01 -4.70833331e-01
1.33533463e-01 1.00946105e+00 5.73237896e-01 3.98369491e-01
-9.25597966e-01 8.80154788e-01 5.14122581e+00 4.70359385e-01
-1.41557443e+00 -1.43126771e-01 5.14183104e-01 4.75256979e-01
-6.07642591e-01 -1.27368020e-02 -4.06549275e-01 4.37359393e-01
5.31801522e-01 -1.40430197e-01 4.35836822e-01 9.81497526e-01
-3.43605787e-01 2.68256545e-01 -1.49597752e+00 1.10748494e+00
4.02077101e-02 -1.47170615e+00 4.48578507e-01 3.45664203e-01
8.98887515e-01 8.89844447e-02 2.39047691e-01 4.67937067e-02
6.10912442e-01 -1.46328020e+00 5.86992145e-01 5.35230815e-01
9.34978068e-01 -8.73254657e-01 4.03378844e-01 5.19646883e-01
-1.12346458e+00 3.80360872e-01 -5.72330058e-01 1.32299839e-02
-4.52271104e-02 7.00307369e-01 -1.37651384e+00 5.10555148e-01
2.74629474e-01 8.38102698e-01 -4.18398708e-01 1.07504010e+00
-4.43049580e-01 5.59805632e-01 -5.10439515e-01 -8.22494105e-02
3.96047533e-01 -6.25481486e-01 5.43236494e-01 8.32568824e-01
6.98363483e-01 -6.93055764e-02 2.96816736e-01 1.53106248e+00
-3.27076048e-01 -9.44409519e-02 -1.23494065e+00 -2.06748068e-01
1.68853447e-01 1.03169107e+00 -9.45942700e-01 -3.37386318e-02
-2.50308782e-01 1.00603700e+00 4.51161981e-01 4.44875747e-01
-4.35408682e-01 -2.34857351e-01 5.94708085e-01 4.50871617e-01
6.16752923e-01 -5.37581205e-01 -2.13555902e-01 -7.86945343e-01
-1.47056594e-01 -2.23946944e-01 7.55620524e-02 -7.58315980e-01
-1.50773549e+00 6.19195163e-01 -6.63905591e-02 -1.29383671e+00
-5.73604941e-01 -6.04734540e-01 -8.92659187e-01 1.34201467e+00
-1.21511197e+00 -1.44379127e+00 -5.81465006e-01 5.85975587e-01
2.35588998e-01 -8.64768773e-02 9.85502124e-01 -2.21169576e-01
3.07545990e-01 1.52335018e-01 -1.63711652e-01 3.18329573e-01
-1.46352723e-01 -1.74072015e+00 9.93171275e-01 6.20724738e-01
6.19882882e-01 4.42390561e-01 4.95930672e-01 -6.30186021e-01
-1.41892016e+00 -1.54113615e+00 3.15977186e-01 -3.22752982e-01
2.49229312e-01 -6.97314441e-01 -8.58856142e-01 8.05345118e-01
-1.59107402e-01 4.06213522e-01 3.08867902e-01 -2.22910210e-01
-6.80862740e-02 2.41680354e-01 -1.28789115e+00 3.13375980e-01
1.15476966e+00 -5.25731981e-01 -1.85093373e-01 5.00415742e-01
4.78089720e-01 -6.40503705e-01 -6.79572165e-01 1.61670074e-01
-2.00663134e-01 -8.59622657e-01 9.18347716e-01 -5.73418498e-01
3.63084763e-01 -2.69622564e-01 -5.06636314e-02 -1.85771298e+00
-2.52537191e-01 -5.28471708e-01 1.52334431e-02 1.02778602e+00
6.14824286e-03 -7.55072892e-01 1.31305873e+00 3.85407582e-02
-2.71072865e-01 -7.87955880e-01 -7.61482835e-01 -6.39742732e-01
1.87849537e-01 -1.59382626e-01 8.43887985e-01 8.69038820e-01
-7.86549091e-01 3.87576371e-01 2.77331322e-01 4.64027435e-01
7.78988242e-01 4.30023670e-01 1.02253711e+00 -1.50341964e+00
-3.33548456e-01 -3.95256519e-01 -1.11542499e+00 -1.09480250e+00
2.96949148e-01 -1.31620061e+00 -4.83727679e-02 -1.75929332e+00
-4.69224274e-01 -8.92296672e-01 3.13389868e-01 8.78639743e-02
1.02261908e-01 9.42022800e-02 -2.58349534e-02 -1.01547889e-01
2.22799748e-01 6.04681373e-01 1.38499725e+00 -2.91437000e-01
-2.07478013e-02 4.36109185e-01 -4.21779722e-01 3.47048402e-01
6.79352760e-01 -3.20986986e-01 -6.68708742e-01 -5.12992680e-01
1.13906451e-01 -1.23511799e-01 6.35789514e-01 -1.11156809e+00
9.12970155e-02 6.92394152e-02 3.10110867e-01 -5.35570323e-01
3.00851643e-01 -9.74244237e-01 5.16221642e-01 1.63184568e-01
-1.57596190e-02 5.21396771e-02 -8.91424790e-02 6.64200366e-01
-2.48758793e-01 -2.10088328e-01 7.61226892e-01 -3.97706240e-01
-4.33166116e-01 8.21295321e-01 2.47185364e-01 1.75784737e-01
9.07602727e-01 -3.08577657e-01 2.10563779e-01 -5.32363594e-01
-7.45073855e-01 -7.99595639e-02 9.40262258e-01 2.84717202e-01
8.44964683e-01 -1.41458392e+00 -5.89522123e-01 6.51211083e-01
7.02600554e-02 9.41084623e-01 -1.42063126e-01 6.47570714e-02
-8.48984599e-01 4.85504232e-02 -1.31604761e-01 -1.05282199e+00
-7.40207613e-01 4.57038671e-01 3.95116806e-01 -4.59131598e-02
-1.00036216e+00 9.90443468e-01 3.60259295e-01 -4.18247402e-01
5.41194528e-02 -6.96607947e-01 1.39431804e-01 -3.96643698e-01
-6.44454211e-02 -7.01738447e-02 2.77394086e-01 -5.43836653e-01
1.39847741e-01 6.09011352e-01 3.86440873e-01 1.76106766e-01
1.57169473e+00 5.71018398e-01 -1.64828002e-01 6.44595027e-01
1.46084797e+00 -2.46270046e-01 -1.25762522e+00 -1.91472650e-01
-1.47612974e-01 -5.63644648e-01 -5.24754077e-02 -6.93719834e-02
-1.23644960e+00 1.07065225e+00 3.51250350e-01 5.91943800e-01
7.21935093e-01 4.35510963e-01 5.19119084e-01 -8.99006985e-03
6.33472443e-01 -5.34303308e-01 2.82452069e-02 4.91337299e-01
1.05335915e+00 -8.69897127e-01 -1.66816920e-01 -6.50690496e-01
-1.68209255e-01 1.25051844e+00 3.21463376e-01 -9.19830799e-01
8.63413990e-01 2.12253153e-01 -4.04405892e-01 -6.57921433e-01
-1.99955314e-01 -2.86994219e-01 4.41420406e-01 1.09554946e+00
2.41061360e-01 3.81799698e-01 4.18614268e-01 4.83370060e-03
-6.75041497e-01 -2.30605960e-01 2.53723115e-01 9.35205698e-01
-1.17674343e-01 -1.00431311e+00 -1.45764723e-01 8.54499996e-01
7.03712599e-03 8.52971002e-02 -4.37928259e-01 6.62232518e-01
4.90310751e-02 2.43206397e-01 6.35203302e-01 -2.28439957e-01
3.24142218e-01 -8.80930871e-02 7.36670792e-01 -1.24087632e+00
-3.42907250e-01 -2.74006575e-01 -3.11994970e-01 -3.23430836e-01
-3.54499459e-01 -5.12211859e-01 -1.33268809e+00 -1.03002209e-02
-1.13809749e-01 4.00245398e-01 9.25979793e-01 5.69716930e-01
3.39033008e-01 5.33658206e-01 8.31566393e-01 -1.27641594e+00
-3.80657643e-01 -7.83173859e-01 -9.13210928e-01 7.18577445e-01
4.07816559e-01 -6.76204979e-01 -4.02679950e-01 -3.24223712e-02] | [8.760618209838867, -3.7198987007141113] |
3a253734-e6f6-441e-b2df-17628707c8d5 | deepquarantine-for-suspicious-mail | 2001.04168 | null | https://arxiv.org/abs/2001.04168v1 | https://arxiv.org/pdf/2001.04168v1.pdf | DeepQuarantine for Suspicious Mail | In this paper, we introduce DeepQuarantine (DQ), a cloud technology to detect and quarantine potential spam messages. Spam attacks are becoming more diverse and can potentially be harmful to email users. Despite the high quality and performance of spam filtering systems, detection of a spam campaign can take some time. Unfortunately, in this case some unwanted messages get delivered to users. To solve this problem, we created DQ, which detects potential spam and keeps it in a special Quarantine folder for a while. The time gained allows us to double-check the messages to improve the reliability of the anti-spam solution. Due to high precision of the technology, most of the quarantined mail is spam, which allows clients to use email without delay. Our solution is based on applying Convolutional Neural Networks on MIME headers to extract deep features from large-scale historical data. We evaluated the proposed method on real-world data and showed that DQ enhances the quality of spam detection. | ['Nikita Benkovich', 'Dmitry Golubev', 'Roman Dedenok'] | 2020-01-13 | null | null | null | null | ['spam-detection'] | ['natural-language-processing'] | [-2.58740336e-01 -7.61298180e-01 2.73486108e-01 -2.19879270e-01
-4.10287321e-01 -8.86385322e-01 6.57508492e-01 4.25059617e-01
-4.53239560e-01 4.46029007e-01 -1.63519934e-01 -6.75977170e-01
8.68327990e-02 -1.31751871e+00 -3.37873876e-01 -4.71572548e-01
3.72780822e-02 4.24244732e-01 7.70389557e-01 -4.64559227e-01
6.80737853e-01 9.14059937e-01 -1.34886515e+00 7.02433288e-01
7.86104620e-01 9.45801079e-01 1.81339029e-02 6.24261677e-01
-4.56320494e-01 5.77000320e-01 -1.23363054e+00 -6.37488723e-01
4.24256086e-01 -9.40920711e-02 -7.38416612e-01 -1.44089177e-01
6.96547151e-01 -7.65511036e-01 -7.12568760e-01 1.08798099e+00
3.77087414e-01 -1.65558577e-01 8.60539451e-03 -1.37355089e+00
-2.53080547e-01 2.49400884e-01 -2.12694228e-01 4.69065279e-01
2.06963688e-01 2.98012704e-01 7.63275325e-01 -5.45090020e-01
4.65118021e-01 1.26403296e+00 5.97758532e-01 4.24281180e-01
-9.33154106e-01 -9.80553031e-01 -2.88462937e-01 4.84900326e-02
-7.90961862e-01 -2.71935642e-01 3.47423673e-01 -7.61023685e-02
6.19426489e-01 4.65708613e-01 4.79970515e-01 1.05564678e+00
5.59882879e-01 7.30055988e-01 9.79134619e-01 1.04981579e-01
2.42961034e-01 2.03309789e-01 4.86284882e-01 4.82740253e-01
3.57404411e-01 7.21073002e-02 -2.25302592e-01 -6.88781977e-01
4.13644820e-01 5.46276748e-01 -6.05479889e-02 4.78534028e-02
-6.77188694e-01 8.96814585e-01 6.29348576e-01 3.75040472e-01
-3.45104456e-01 -4.68730787e-03 8.67478549e-01 6.51549816e-01
4.47163284e-01 4.79434729e-01 -3.08869988e-01 -2.86324263e-01
-9.51691151e-01 2.49122500e-01 1.04871511e+00 4.65379238e-01
5.87153018e-01 -2.19709873e-01 1.97285041e-02 6.17570698e-01
-1.05966523e-01 5.29294491e-01 4.55982834e-01 -4.05858427e-01
4.07193810e-01 1.01964867e+00 1.46775201e-01 -1.22813594e+00
-2.08904251e-01 -3.91690940e-01 -7.55306363e-01 1.63608983e-01
5.34516215e-01 -4.32063714e-02 -6.95344210e-01 1.03295779e+00
1.96780235e-01 2.34027311e-01 -4.85636324e-01 1.07236683e+00
3.77679765e-01 7.01547205e-01 -2.20715404e-02 1.34588957e-01
1.30948651e+00 -5.64259946e-01 -4.64273870e-01 -8.12328383e-02
3.59047800e-01 -9.87730563e-01 8.81912291e-01 5.05590200e-01
-5.70595324e-01 -2.36480638e-01 -7.34922171e-01 4.10977781e-01
-6.03575766e-01 -4.01457906e-01 7.03599870e-01 8.80088449e-01
-7.54801989e-01 8.73518884e-01 -5.93071043e-01 -3.71495903e-01
4.33304101e-01 5.82584679e-01 -4.09892082e-01 8.19072723e-02
-1.46078765e+00 7.98888266e-01 -6.43627569e-02 -1.38129771e-01
-8.59038711e-01 -5.67042410e-01 -1.59056768e-01 3.78351390e-01
1.58542722e-01 -8.50876123e-02 1.48934090e+00 -1.10407674e+00
-9.17327404e-01 5.72979510e-01 -3.05226212e-03 -6.70025110e-01
7.19165266e-01 -3.04878712e-01 -5.45038700e-01 3.69961798e-01
6.56667128e-02 -2.08227277e-01 1.26299644e+00 -1.00605500e+00
-9.49675679e-01 -4.36967313e-01 -2.63653323e-02 -6.81119621e-01
-6.07941687e-01 3.28362852e-01 -2.89029002e-01 -3.98310393e-01
-5.83017841e-02 -7.31364131e-01 -1.34279028e-01 -4.05460566e-01
9.16885287e-02 -2.56127894e-01 1.44346035e+00 -6.84706450e-01
1.22547650e+00 -2.08584857e+00 -8.36001575e-01 3.72477174e-01
4.63611424e-01 9.48902786e-01 -3.04765970e-01 6.54625535e-01
8.85924697e-02 8.66365954e-02 1.31885648e-01 2.22020999e-01
-3.56241435e-01 -7.28436187e-02 -8.19231808e-01 6.23801291e-01
2.80276716e-01 7.76955366e-01 -9.91727471e-01 -1.50213718e-01
2.16915220e-01 3.35057259e-01 -2.68919319e-01 2.98368990e-01
-2.58192062e-01 -9.04708579e-02 -4.95576978e-01 5.93233585e-01
1.14982307e+00 -2.06331164e-01 2.28133962e-01 1.31040588e-01
2.09012792e-01 7.60323286e-01 -9.11643565e-01 6.51207924e-01
-6.92043185e-01 6.65000021e-01 4.66778755e-01 -8.15514565e-01
8.59392941e-01 1.56316742e-01 3.69632244e-01 -7.69614935e-01
3.97752941e-01 4.30812091e-01 -1.85078755e-01 -3.64099979e-01
5.87257564e-01 2.27287367e-01 6.14333618e-03 9.37377751e-01
-3.66021633e-01 1.97829932e-01 3.46905380e-01 7.33179331e-01
1.67351866e+00 -5.93196630e-01 -3.77906531e-01 -6.68290481e-02
6.22248352e-01 -2.70546339e-02 2.91617632e-01 9.31563795e-01
-4.23890084e-01 1.83425680e-01 6.36356235e-01 -4.64228988e-01
-1.00310516e+00 -1.24766302e+00 -2.54254099e-02 1.07544017e+00
2.15562031e-01 -4.33808208e-01 -7.71184146e-01 -1.31646121e+00
5.59068739e-01 3.85919452e-01 -1.10547371e-01 -2.19051987e-01
-9.54868436e-01 -8.10490072e-01 3.60693067e-01 1.49946690e-01
5.57965159e-01 -1.02111685e+00 -3.74374926e-01 6.38481498e-01
-1.29799366e-01 -6.57882214e-01 -5.00059605e-01 -3.52425203e-02
-8.87717783e-01 -1.41374302e+00 -2.74280488e-01 -7.14315414e-01
6.24330461e-01 9.39942896e-01 1.09120548e+00 8.92002821e-01
-3.68944854e-01 -5.00307120e-02 -5.16846359e-01 -7.87776560e-02
-8.99577260e-01 1.99656621e-01 2.63027959e-02 1.19012728e-01
8.65601182e-01 -4.57207590e-01 -7.26853013e-01 5.13143182e-01
-1.00446129e+00 -7.26011693e-01 4.05559182e-01 7.95291424e-01
-5.21764874e-01 4.71004814e-01 1.12153471e+00 -1.05005574e+00
1.11453187e+00 -5.45374751e-01 -9.47372198e-01 -1.11541286e-01
-6.24611318e-01 -1.76267952e-01 9.71171975e-01 -4.18688178e-01
-7.74405360e-01 -2.00107440e-01 -1.51244894e-01 1.54494941e-01
8.37836191e-02 -1.55327961e-01 -1.05343643e-03 -5.83939999e-02
6.21463716e-01 2.11336046e-01 3.35375577e-01 -6.50779784e-01
1.06902458e-01 1.63784504e+00 2.88976997e-01 -1.14226781e-01
8.20503175e-01 6.02004528e-01 -3.13590974e-01 -7.53866613e-01
-4.32040930e-01 -8.45953166e-01 -9.30098891e-02 -8.35757107e-02
1.34276750e-03 -4.10691708e-01 -1.40495586e+00 7.05238521e-01
-1.16232014e+00 2.62345612e-01 4.77222264e-01 -6.36358783e-02
1.29013360e-01 7.87205994e-01 -1.05889130e+00 -8.17150414e-01
-7.00859725e-01 -8.78937006e-01 9.33721125e-01 7.95468967e-03
-1.84805065e-01 -5.51346898e-01 -2.45712891e-01 3.95729482e-01
7.81301618e-01 -4.51243758e-01 7.54549384e-01 -9.57930088e-01
-9.48726773e-01 -1.05704904e+00 -6.50861740e-01 5.36774755e-01
2.75668323e-01 -1.87884327e-02 -7.43278861e-01 -5.22130847e-01
2.58686572e-01 1.24423534e-01 8.31970513e-01 -3.21216941e-01
1.22245705e+00 -2.77527839e-01 -3.52148324e-01 9.91402045e-02
1.22060597e+00 2.20406264e-01 7.59974599e-01 6.09100580e-01
9.43520516e-02 5.23628175e-01 5.51309705e-01 3.14124405e-01
-3.37228626e-01 2.85521269e-01 7.86823690e-01 1.97736859e-01
2.24507391e-01 -5.69066778e-02 4.11299407e-01 5.06417334e-01
6.38669908e-01 -3.11599582e-01 -6.50096297e-01 5.84709644e-01
-1.65949261e+00 -1.07008779e+00 -6.37790382e-01 2.11643291e+00
7.35777497e-01 3.51787359e-01 1.69255346e-01 1.19037524e-01
9.91204500e-01 2.70522200e-02 -1.41337007e-01 -5.38245678e-01
6.48811311e-02 4.06116694e-01 6.85297251e-01 3.39046568e-01
-1.18844676e+00 9.11515892e-01 5.91504383e+00 7.49882817e-01
-1.37953210e+00 3.07570957e-02 2.21749902e-01 -8.70148391e-02
8.85187760e-02 8.49132091e-02 -8.57702315e-01 1.08064449e+00
1.24576259e+00 -1.10724911e-01 3.91592443e-01 9.77846563e-01
3.42838734e-01 -2.91766226e-01 -6.91265762e-01 5.73627353e-01
-1.02015465e-01 -1.55406165e+00 -9.51329898e-03 -4.98773269e-02
3.18331093e-01 2.22665519e-02 -2.35119224e-01 2.58182466e-01
3.72521400e-01 -6.50866628e-01 2.07505524e-01 1.06995784e-01
1.83595389e-01 -9.97843266e-01 1.13405168e+00 3.84355485e-01
-5.59736133e-01 -3.90354127e-01 -3.84080052e-01 -8.50313306e-02
1.86115906e-01 9.89589572e-01 -1.13239181e+00 1.14994675e-01
7.22055793e-01 1.55594349e-01 -8.74330938e-01 1.17058158e+00
-3.56309675e-02 6.50597036e-01 -2.74443477e-01 -5.92518449e-01
2.78616101e-01 -1.78902730e-01 4.25411165e-01 1.21774530e+00
8.60058218e-02 -3.59593213e-01 -3.71249504e-02 5.65179646e-01
-3.98128152e-01 -9.08972695e-02 -4.51758772e-01 -2.28240967e-01
3.32431018e-01 1.40824568e+00 -7.37421811e-01 -5.40818930e-01
-2.22721100e-01 1.10057485e+00 4.80771111e-03 -3.62288356e-02
-5.54989636e-01 -9.12379622e-01 7.89018750e-01 6.04155242e-01
7.01607391e-02 -2.55258352e-01 -2.63520092e-01 -1.06095088e+00
9.87023711e-02 -1.25605893e+00 2.39452079e-01 -2.73733675e-01
-1.73583353e+00 4.77102906e-01 -7.84042418e-01 -1.03864014e+00
1.62096739e-01 -6.56788707e-01 -6.88366055e-01 9.61498559e-01
-9.83026147e-01 -6.88613951e-01 -1.25902772e-01 2.88401395e-01
4.66587156e-01 -3.67749631e-01 4.66494173e-01 6.91773355e-01
-1.40884846e-01 3.85858178e-01 3.67288589e-01 4.14344609e-01
8.93117487e-01 -1.01393330e+00 9.41986442e-01 7.83902228e-01
-1.87279582e-01 1.09945524e+00 5.80191135e-01 -1.04687631e+00
-1.48823583e+00 -9.70526457e-01 1.18368268e+00 -7.25171924e-01
9.78366852e-01 -6.96294785e-01 -1.24164081e+00 1.72397465e-01
9.52297915e-03 -4.17343974e-01 2.14243039e-01 -4.90519814e-02
-4.32676464e-01 -3.94103408e-01 -1.32025754e+00 5.81658602e-01
5.00579894e-01 -7.23105431e-01 -5.42988241e-01 5.96094549e-01
5.01579523e-01 1.88880384e-01 -1.94603890e-01 1.49757357e-03
5.03951728e-01 -1.17367554e+00 6.47713125e-01 -8.45809877e-01
1.89540192e-01 -2.17762128e-01 3.67427915e-01 -1.03162122e+00
-1.27197936e-01 -6.84878528e-01 -5.13962880e-02 1.14207876e+00
1.95112862e-02 -8.08277667e-01 1.07596278e+00 3.99652630e-01
4.00470614e-01 -3.34142953e-01 -8.06769848e-01 -7.72738934e-01
-5.26429638e-02 -3.50518793e-01 9.35388088e-01 6.96961522e-01
-3.01974043e-02 2.34458312e-01 -2.90054917e-01 1.85318917e-01
5.26640475e-01 3.80382597e-01 9.90527689e-01 -1.32575893e+00
4.13941331e-02 -2.03082472e-01 -4.27078426e-01 -1.03026795e+00
1.29242197e-01 -8.20100784e-01 -1.97070867e-01 -1.16708136e+00
1.92371476e-02 -4.02221441e-01 -1.13319911e-01 2.05766737e-01
-1.44572735e-01 1.27883673e-01 9.33655873e-02 1.57821640e-01
-3.80727410e-01 2.09628865e-02 8.33472490e-01 -3.28030020e-01
-5.69534451e-02 6.07327998e-01 -3.93835604e-01 3.40771228e-01
1.07139659e+00 -6.86838150e-01 1.85714945e-01 -1.59803480e-01
2.40019843e-01 -2.60177583e-01 4.96626824e-01 -6.51459813e-01
1.31917194e-01 -2.05858685e-02 1.90385848e-01 -6.82554424e-01
2.30987892e-02 -1.03417325e+00 -4.57782954e-01 8.78814399e-01
7.44412988e-02 9.71454009e-02 -6.46798238e-02 5.41784942e-01
-1.33356512e-01 -3.23755682e-01 1.04217422e+00 -1.19673125e-01
-4.47173566e-01 7.49434829e-02 -8.12918842e-01 -3.61598521e-01
7.99760938e-01 6.79361969e-02 -8.59946728e-01 -4.97220010e-01
-1.20367226e-03 3.37581873e-01 4.85921592e-01 7.54661739e-01
5.03924370e-01 -1.05969751e+00 -5.17656147e-01 4.82999802e-01
-2.30974808e-01 -7.61148036e-01 9.49864909e-02 4.59554344e-01
-1.03976178e+00 4.18621302e-01 -5.98759241e-02 -3.04190457e-01
-1.43805492e+00 7.55447209e-01 2.94705302e-01 9.75505561e-02
-6.20878696e-01 5.40757537e-01 -4.41433549e-01 -4.25259769e-01
4.04540867e-01 3.54790539e-01 1.93636999e-01 3.76071893e-02
8.44281316e-01 6.31526232e-01 5.65405667e-01 -2.84814447e-01
-4.83049273e-01 -2.91940629e-01 -5.26868403e-01 2.32623264e-01
1.25863826e+00 1.35699138e-01 -7.34395325e-01 -3.48201543e-01
1.34350014e+00 2.92631745e-01 -6.77269638e-01 -1.30767941e-01
3.29073489e-01 -1.08382547e+00 -2.35046148e-02 -6.88417494e-01
-1.03400493e+00 9.36217725e-01 6.52463078e-01 8.23530495e-01
8.16340983e-01 -4.00560111e-01 1.60898197e+00 4.76895273e-01
4.51040149e-01 -1.28865314e+00 1.79231480e-01 7.27467120e-01
3.58710647e-01 -1.31345284e+00 -2.21919551e-01 -3.08567137e-01
-1.40853241e-01 1.20068312e+00 6.53376579e-01 -3.05274129e-01
5.60214043e-01 2.53261626e-01 2.97647178e-01 -1.86009675e-01
-7.09752619e-01 2.71192193e-01 -4.00359869e-01 4.43008542e-01
1.81176454e-01 -5.47164902e-02 -5.42482138e-01 2.00572908e-01
-1.80791412e-02 -1.46770284e-01 6.08258665e-01 1.10653293e+00
-8.91786516e-01 -1.42813647e+00 -6.60796225e-01 8.77093613e-01
-7.11911500e-01 -8.08973014e-02 -7.24947512e-01 3.23808104e-01
-1.92505792e-01 1.30156422e+00 2.44792923e-01 -4.28929299e-01
3.19744200e-01 9.99674350e-02 -1.03696518e-01 -4.61046219e-01
-1.38175488e+00 -2.37278685e-01 8.03927779e-02 -4.30458933e-01
3.90728921e-01 -1.30587742e-01 -1.11921620e+00 -1.11372495e+00
-6.15648508e-01 4.50797349e-01 8.58158410e-01 4.99347746e-01
4.59076166e-01 1.06744528e-01 1.18167555e+00 -2.86530256e-01
-1.23711431e+00 -7.39614308e-01 -6.52862906e-01 8.16426992e-01
4.06922489e-01 -2.81554520e-01 -6.73646033e-01 -4.56645936e-01] | [7.815227031707764, 10.001577377319336] |
4780458f-f5e6-4bd0-9445-af0a49deb8a0 | filtering-distillation-and-hard-negatives-for | 2301.02280 | null | https://arxiv.org/abs/2301.02280v2 | https://arxiv.org/pdf/2301.02280v2.pdf | Filtering, Distillation, and Hard Negatives for Vision-Language Pre-Training | Vision-language models trained with contrastive learning on large-scale noisy data are becoming increasingly popular for zero-shot recognition problems. In this paper we improve the following three aspects of the contrastive pre-training pipeline: dataset noise, model initialization and the training objective. First, we propose a straightforward filtering strategy titled Complexity, Action, and Text-spotting (CAT) that significantly reduces dataset size, while achieving improved performance across zero-shot vision-language tasks. Next, we propose an approach titled Concept Distillation to leverage strong unimodal representations for contrastive training that does not increase training complexity while outperforming prior work. Finally, we modify the traditional contrastive alignment objective, and propose an importance-sampling approach to up-sample the importance of hard-negatives without adding additional complexity. On an extensive zero-shot benchmark of 29 tasks, our Distilled and Hard-negative Training (DiHT) approach improves on 20 tasks compared to the baseline. Furthermore, for few-shot linear probing, we propose a novel approach that bridges the gap between zero-shot and few-shot performance, substantially improving over prior work. Models are available at https://github.com/facebookresearch/diht. | ['Dhruv Mahajan', 'Vignesh Ramanathan', 'Yi Wen', 'Yash Patel', 'Simon Vandenhende', 'Todor Mihaylov', 'Abhishek Kadian', 'Abhimanyu Dubey', 'Filip Radenovic'] | 2023-01-05 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Radenovic_Filtering_Distillation_and_Hard_Negatives_for_Vision-Language_Pre-Training_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Radenovic_Filtering_Distillation_and_Hard_Negatives_for_Vision-Language_Pre-Training_CVPR_2023_paper.pdf | cvpr-2023-1 | ['text-spotting'] | ['computer-vision'] | [ 4.74511206e-01 -1.92522302e-01 -2.37052575e-01 -4.82732534e-01
-1.49566305e+00 -2.37107366e-01 9.84284759e-01 6.17650896e-02
-8.16589177e-01 3.64585996e-01 3.80388170e-01 -2.75027156e-01
3.06921929e-01 -2.99890995e-01 -6.19843721e-01 -5.23447990e-01
4.39108491e-01 4.38562602e-01 3.14744771e-01 -2.22869843e-01
2.65336186e-01 -1.88166142e-01 -1.65824819e+00 4.01305914e-01
6.15953863e-01 9.76482272e-01 3.64168644e-01 8.91859770e-01
-7.02743279e-03 8.68395329e-01 -4.11502600e-01 -4.72677857e-01
3.07788461e-01 -4.14396912e-01 -6.92024648e-01 -5.83250895e-02
9.45542157e-01 -4.83736694e-01 -1.38378054e-01 1.04998243e+00
9.32647884e-01 4.03779775e-01 5.43250799e-01 -1.16049111e+00
-8.53735685e-01 5.48684299e-01 -6.73689961e-01 4.26560134e-01
-9.30461958e-02 5.95190167e-01 1.33956158e+00 -1.64189565e+00
6.23691022e-01 1.36468458e+00 6.11553073e-01 8.65884781e-01
-1.30370951e+00 -5.21431148e-01 4.54554856e-01 5.22922873e-01
-1.18857718e+00 -7.91197836e-01 5.12959301e-01 -4.63461131e-01
1.31000388e+00 1.28338769e-01 4.18659478e-01 1.60100615e+00
-3.95258605e-01 9.66520846e-01 1.05491328e+00 -7.16761351e-01
3.45782012e-01 -1.33424744e-01 6.97500467e-01 6.35879040e-01
9.94633585e-02 4.85167615e-02 -5.36893487e-01 -1.58244967e-01
1.43818751e-01 1.39153317e-01 -1.22242406e-01 -3.73771161e-01
-8.62986684e-01 8.95961702e-01 9.05630812e-02 1.33280188e-01
-6.22914508e-02 1.65147558e-01 5.29552400e-01 2.08762750e-01
7.43750334e-01 4.77017939e-01 -4.70627308e-01 -3.82119387e-01
-1.11816430e+00 1.30247921e-01 7.11358130e-01 8.23174119e-01
7.90683687e-01 2.17235833e-02 -8.02927732e-01 1.15120232e+00
5.70746548e-02 4.95255858e-01 4.37451065e-01 -9.61695552e-01
4.71540093e-01 1.64190069e-01 -3.63639295e-02 -3.43212843e-01
-1.36412635e-01 -4.72595602e-01 -5.35180807e-01 5.60951680e-02
2.47982711e-01 -5.04390150e-02 -1.49810350e+00 1.66680360e+00
1.81158319e-01 4.20043528e-01 -1.26871347e-01 7.63281643e-01
9.91695821e-01 6.12684011e-01 1.20274059e-01 -1.61394104e-01
1.33292413e+00 -1.57906199e+00 -6.00789785e-01 -7.06999302e-01
6.86769426e-01 -7.11627483e-01 1.73613799e+00 3.61779779e-01
-1.08043742e+00 -4.35318649e-01 -1.06146288e+00 -5.20660043e-01
-2.80357718e-01 4.55958061e-02 3.05960029e-01 4.43631887e-01
-9.34118211e-01 5.60953140e-01 -8.99778605e-01 -4.36189711e-01
6.50362790e-01 -7.55581185e-02 8.75412524e-02 -5.03647625e-01
-9.61231351e-01 9.67221200e-01 3.04891169e-01 -2.94460565e-01
-1.15651178e+00 -9.45238829e-01 -8.76314819e-01 1.09977737e-01
9.47975576e-01 -6.30503416e-01 1.76916516e+00 -6.02504611e-01
-1.39706159e+00 9.71624613e-01 -3.01906973e-01 -6.23737276e-01
5.21893322e-01 -6.76269293e-01 -1.82934459e-02 8.31004698e-03
6.15256354e-02 9.76887703e-01 1.03240287e+00 -1.10205793e+00
-5.19210219e-01 -1.81035891e-01 7.24359080e-02 2.79178649e-01
-4.61647898e-01 1.14363223e-01 -9.63610649e-01 -5.79728603e-01
-3.11242104e-01 -7.05683112e-01 -3.49876165e-01 1.54725760e-01
-2.56226033e-01 -2.77079403e-01 5.55353582e-01 -5.60446680e-01
1.13570726e+00 -2.15186000e+00 6.68515638e-02 -3.23221445e-01
3.52539539e-01 5.30041635e-01 -6.18630886e-01 1.70468628e-01
7.36386105e-02 5.46204224e-02 -5.59883058e-01 -8.99823427e-01
7.87074044e-02 2.64528245e-01 -3.45402330e-01 2.09467322e-01
5.78039169e-01 1.06831133e+00 -1.05972314e+00 -4.34673280e-01
2.20762610e-01 2.80168831e-01 -7.79399812e-01 1.01258136e-01
-3.83216351e-01 -2.25961884e-03 2.07759917e-01 7.87447751e-01
5.30986190e-01 -4.93721873e-01 -1.56033054e-01 -8.79259184e-02
-9.35950801e-02 2.96086192e-01 -9.61878419e-01 1.78257811e+00
-4.58334684e-01 6.13390565e-01 -9.55377519e-02 -9.67126012e-01
6.26107037e-01 -1.36128455e-01 4.65264209e-02 -9.79651392e-01
2.05102935e-01 5.08432463e-03 -8.24924260e-02 -4.03317094e-01
5.92773199e-01 2.56295968e-03 8.80116746e-02 3.10547441e-01
4.43955481e-01 -1.72228888e-01 4.36555296e-01 4.02581006e-01
1.11887670e+00 9.69931930e-02 4.88344699e-01 2.29533389e-02
1.89945418e-02 -5.76729216e-02 3.98760945e-01 1.26214695e+00
-4.51602995e-01 9.49805021e-01 4.20151949e-01 -2.42981799e-02
-1.15085721e+00 -9.86328363e-01 7.36301243e-02 1.70060027e+00
-1.17733642e-01 -5.84685445e-01 -5.83157122e-01 -7.05493748e-01
-2.15120539e-01 9.79357719e-01 -7.58672059e-01 -2.87372380e-01
-3.71475101e-01 -8.21704507e-01 3.50255847e-01 5.17507315e-01
1.14479721e-01 -6.26346767e-01 -4.67227727e-01 -5.45616001e-02
-3.59400511e-01 -1.20899892e+00 -6.47646189e-01 4.50829536e-01
-5.54372609e-01 -7.50477731e-01 -8.72286916e-01 -7.17923582e-01
3.19999784e-01 8.34771574e-01 1.32794523e+00 -4.32808474e-02
-6.16534054e-01 5.62042832e-01 -4.60133255e-01 -6.56814814e-01
-6.22018240e-02 1.99493989e-02 -1.04848161e-01 -2.77663589e-01
6.25149310e-01 -3.01406682e-01 -5.93243420e-01 6.96256012e-02
-5.98711133e-01 6.44143224e-02 6.78201735e-01 1.24899685e+00
6.56947970e-01 -9.25823510e-01 2.36675754e-01 -7.59837568e-01
5.69711208e-01 -6.28453195e-01 -3.95772427e-01 2.92635888e-01
-8.67047012e-01 8.68381560e-02 3.28552604e-01 -6.30360126e-01
-9.57654059e-01 -2.27546543e-01 -9.87885743e-02 -9.10448015e-01
5.40598370e-02 2.83895433e-01 2.01096699e-01 1.53518021e-01
1.01065159e+00 2.58407593e-01 -1.21491283e-01 -5.87740362e-01
7.84322262e-01 6.39302015e-01 5.29887140e-01 -4.42006826e-01
6.07021630e-01 5.67250609e-01 -5.02807736e-01 -9.94902432e-01
-1.34381831e+00 -9.38195646e-01 -4.32884514e-01 1.39770418e-01
6.60924613e-01 -1.21532810e+00 -1.57933921e-01 4.11135793e-01
-1.09896994e+00 -5.64309597e-01 -6.36703253e-01 3.28721553e-01
-3.44626695e-01 5.13379514e-01 -5.53126574e-01 -9.06012714e-01
-6.78263545e-01 -8.88598025e-01 1.20717502e+00 -5.38989455e-02
-7.08687678e-02 -5.86171806e-01 1.49578020e-01 4.18292046e-01
4.29149240e-01 -3.39907587e-01 5.63413382e-01 -8.82278204e-01
-3.42281520e-01 5.47956340e-02 -5.22138536e-01 5.03202915e-01
-3.68629396e-01 -1.42906293e-01 -1.33557773e+00 -4.57104295e-01
-3.44440737e-03 -9.13302124e-01 1.53839850e+00 4.31420952e-01
1.08076346e+00 -2.24728853e-01 -3.41580361e-02 6.69155896e-01
1.30682528e+00 -1.99852571e-01 4.87879008e-01 3.37752372e-01
6.74743712e-01 4.35706407e-01 8.61847937e-01 7.01745272e-01
3.34858507e-01 6.54347777e-01 3.35106373e-01 -1.80151090e-01
-3.85829926e-01 -1.12326890e-01 4.05384868e-01 7.97189176e-01
7.07910433e-02 -2.13484198e-01 -1.06099427e+00 7.55288780e-01
-1.85440791e+00 -1.01152396e+00 1.89524442e-01 2.12764263e+00
9.95321274e-01 4.84092653e-01 1.43592611e-01 -1.91139832e-01
5.64276457e-01 4.17736024e-01 -7.84880757e-01 -1.51003286e-01
-1.35989534e-02 4.63448852e-01 3.21982116e-01 6.94074810e-01
-1.17294800e+00 1.18106318e+00 5.85901499e+00 9.34274137e-01
-1.00277495e+00 6.70663595e-01 6.12463236e-01 -7.49104142e-01
-2.32694671e-01 1.09201141e-01 -1.23087239e+00 2.87976533e-01
9.06177282e-01 -1.72879204e-01 3.61380130e-01 1.09014785e+00
8.59066620e-02 -1.02665640e-01 -1.11568820e+00 1.18257475e+00
6.10598326e-01 -1.28781319e+00 2.36210264e-02 -2.73059458e-01
8.52795303e-01 6.80879474e-01 1.41091943e-01 8.35296988e-01
4.85661209e-01 -6.75691664e-01 7.16763258e-01 4.24221426e-01
8.69392872e-01 -2.10985631e-01 5.05735636e-01 3.13235164e-01
-8.70122194e-01 -2.10874349e-01 -4.53361988e-01 -1.05405927e-01
6.26961514e-02 5.47312319e-01 -8.49390447e-01 1.36083558e-01
7.66485989e-01 5.99228382e-01 -7.57748902e-01 1.05539596e+00
-2.12815762e-01 8.80536973e-01 -2.05442995e-01 -1.12862729e-01
4.29829597e-01 1.61265507e-01 6.07125700e-01 1.54238772e+00
6.21852949e-02 -4.57043834e-02 3.33080292e-01 7.57979870e-01
-2.12854773e-01 -5.56565868e-03 -4.42273796e-01 -7.71428570e-02
5.27096212e-01 1.17855954e+00 -3.58670473e-01 -7.56756365e-01
-7.55529165e-01 1.10100484e+00 5.52216887e-01 3.65204602e-01
-8.13386738e-01 -2.92372644e-01 6.19070053e-01 -1.63072631e-01
6.07245445e-01 -1.43187821e-01 -1.74160168e-01 -1.36035013e+00
6.93416744e-02 -9.09399092e-01 4.10583675e-01 -5.91639519e-01
-1.45418262e+00 3.00426930e-01 -3.14437263e-02 -1.20573556e+00
-2.99783260e-01 -5.69975376e-01 -5.99740088e-01 6.18593395e-01
-1.82021284e+00 -1.14984453e+00 -4.26408589e-01 3.99319768e-01
1.22442997e+00 -4.48678918e-02 6.97060525e-01 3.05522144e-01
-6.22016013e-01 7.76987374e-01 1.67835057e-01 -1.36416733e-01
1.08678782e+00 -1.19239211e+00 8.49437058e-01 1.09654129e+00
1.99275687e-01 5.74479103e-01 6.02907240e-01 -5.50378442e-01
-1.14392030e+00 -1.13815236e+00 7.90594578e-01 -4.87256229e-01
9.11954463e-01 -6.65744901e-01 -1.00455546e+00 5.45875132e-01
2.11508080e-01 2.92319268e-01 6.19771421e-01 3.84360731e-01
-8.06754887e-01 5.04426882e-02 -6.68819070e-01 7.04433858e-01
1.19754982e+00 -7.07868874e-01 -7.99663961e-01 5.06274104e-01
1.03429782e+00 -1.51981130e-01 -3.21806937e-01 1.38846308e-01
3.95145446e-01 -7.23475218e-01 1.00195920e+00 -7.01173484e-01
7.25448072e-01 1.82060823e-01 -4.27035779e-01 -1.08546674e+00
-5.02569437e-01 -5.64833641e-01 -5.87824941e-01 9.28794444e-01
4.49988455e-01 -3.03441465e-01 6.68086290e-01 3.32056344e-01
-2.88776189e-01 -7.68233299e-01 -8.42828989e-01 -1.00239384e+00
-3.99096459e-02 -6.07993066e-01 -1.79735318e-01 9.61431265e-01
-2.33031809e-01 7.91302025e-01 -7.73412704e-01 -3.22780341e-01
8.73397052e-01 -3.00855160e-01 7.77960777e-01 -1.04527998e+00
-7.91465461e-01 -4.49443519e-01 -6.34641945e-02 -1.20775151e+00
-2.26691350e-01 -8.73738468e-01 3.14450949e-01 -1.59174800e+00
6.41954005e-01 2.05848590e-01 -5.68209529e-01 8.12604368e-01
-5.77191293e-01 3.47123593e-01 5.46337843e-01 2.78387398e-01
-9.80219722e-01 7.73833573e-01 8.53742957e-01 -3.07882488e-01
-1.29566148e-01 -2.18914375e-01 -7.38403022e-01 6.54976428e-01
5.77212155e-01 -4.24463660e-01 -3.47511172e-01 -5.79357624e-01
2.52933148e-02 -4.23305362e-01 4.37608272e-01 -9.76940691e-01
2.79710084e-01 1.74828507e-02 3.94718125e-02 -6.71706736e-01
7.05802202e-01 -2.69453883e-01 -7.71543622e-01 4.91594732e-01
-4.91555125e-01 -2.75390267e-01 1.40204668e-01 7.00008929e-01
4.06354889e-02 -3.92726511e-01 1.05332363e+00 -2.72340804e-01
-1.02974224e+00 2.42533594e-01 -2.53942851e-02 5.68064213e-01
7.71140277e-01 -8.06997940e-02 -7.32901990e-01 -2.11288854e-01
-7.70121157e-01 3.74322385e-01 2.80828714e-01 6.17177010e-01
6.24592066e-01 -9.80825603e-01 -8.28445554e-01 -5.40740676e-02
5.48313975e-01 -2.13386938e-01 3.28681201e-01 1.06496775e+00
1.31189972e-01 2.22009331e-01 2.01373361e-02 -6.97877467e-01
-1.39315224e+00 5.66486776e-01 4.32971120e-02 -2.85460919e-01
-6.07251346e-01 1.23623371e+00 2.67010450e-01 -3.98703694e-01
5.96592546e-01 -2.32279390e-01 -4.97014634e-02 3.58085155e-01
8.22375119e-01 5.54360926e-01 9.39975604e-02 -9.49123800e-02
-2.44585261e-01 3.75642031e-01 -6.37862623e-01 -3.07407528e-01
1.40712285e+00 -1.33112609e-01 3.13603401e-01 7.79638112e-01
1.21625638e+00 -3.97522360e-01 -1.52581525e+00 -6.91735208e-01
3.56765352e-02 -3.53878379e-01 2.47734010e-01 -9.18137610e-01
-5.16141415e-01 1.24233270e+00 8.61564934e-01 -1.94882601e-01
8.72970760e-01 1.15982734e-01 7.08587468e-01 7.85604239e-01
-1.35490343e-01 -1.44162476e+00 5.57165325e-01 9.06651556e-01
7.19585955e-01 -1.75858092e+00 -6.24673255e-02 -1.31166294e-01
-8.49154890e-01 7.70199537e-01 8.17193866e-01 3.86199281e-02
4.18480664e-01 3.63859415e-01 -1.97831035e-01 -1.16177417e-01
-1.22500944e+00 -7.32160270e-01 3.06150526e-01 4.35584038e-01
1.70036599e-01 -1.47807196e-01 -2.63160288e-01 5.97077429e-01
2.23790824e-01 9.41504538e-02 2.98865050e-01 1.06222773e+00
-8.07574511e-01 -7.08968878e-01 -1.25486910e-01 5.99551499e-01
-2.85903335e-01 -7.74082541e-01 -4.98256981e-01 4.71718580e-01
-5.60548007e-02 8.62528682e-01 1.39186531e-01 -2.90763974e-01
4.03805971e-01 2.92522818e-01 3.41575980e-01 -9.60521400e-01
-2.03817695e-01 2.07915187e-01 2.23961949e-01 -6.01587653e-01
-1.30815461e-01 -4.87353623e-01 -8.08093488e-01 1.26814231e-01
-4.56179619e-01 -3.24107558e-01 5.50657570e-01 1.03115702e+00
5.08871019e-01 4.46834087e-01 4.23443049e-01 -8.93344998e-01
-1.12798095e+00 -1.23594093e+00 6.33470155e-03 3.06479543e-01
3.76425654e-01 -6.26490057e-01 -5.33343017e-01 -5.52638695e-02] | [10.041793823242188, 2.2284281253814697] |
fd06f8e1-0d5f-4b3d-a071-3f9b897949c2 | learning-a-sensor-invariant-embedding-of | 2107.09092 | null | https://arxiv.org/abs/2107.09092v2 | https://arxiv.org/pdf/2107.09092v2.pdf | Learning a Joint Embedding of Multiple Satellite Sensors: A Case Study for Lake Ice Monitoring | Fusing satellite imagery acquired with different sensors has been a long-standing challenge of Earth observation, particularly across different modalities such as optical and Synthetic Aperture Radar (SAR) images. Here, we explore the joint analysis of imagery from different sensors in the light of representation learning: we propose to learn a joint embedding of multiple satellite sensors within a deep neural network. Our application problem is the monitoring of lake ice on Alpine lakes. To reach the temporal resolution requirement of the Swiss Global Climate Observing System (GCOS) office, we combine three image sources: Sentinel-1 SAR (S1-SAR), Terra MODIS, and Suomi-NPP VIIRS. The large gaps between the optical and SAR domains and between the sensor resolutions make this a challenging instance of the sensor fusion problem. Our approach can be classified as a late fusion that is learned in a data-driven manner. The proposed network architecture has separate encoding branches for each image sensor, which feed into a single latent embedding. I.e., a common feature representation shared by all inputs, such that subsequent processing steps deliver comparable output irrespective of which sort of input image was used. By fusing satellite data, we map lake ice at a temporal resolution of < 1.5 days. The network produces spatially explicit lake ice maps with pixel-wise accuracies > 91% (respectively, mIoU scores > 60%) and generalises well across different lakes and winters. Moreover, it sets a new state-of-the-art for determining the important ice-on and ice-off dates for the target lakes, in many cases meeting the GCOS requirement. | ['Konrad Schindler', 'Emmanuel Baltsavias', 'Yuchang Jiang', 'Manu Tom'] | 2021-07-19 | null | null | null | null | ['lake-ice-detection', 'lake-ice-detection'] | ['computer-vision', 'miscellaneous'] | [ 5.34108758e-01 -1.87802002e-01 7.06559196e-02 -5.97912848e-01
-9.92256224e-01 -7.46641517e-01 6.96575344e-01 1.31508663e-01
-6.95706904e-01 7.27615058e-01 1.76847473e-01 -3.18996876e-01
-2.17498273e-01 -1.02202725e+00 -6.55601859e-01 -1.07585132e+00
-4.28846657e-01 1.74597487e-01 -1.75878808e-01 -4.84587371e-01
-5.19531131e-01 7.46552825e-01 -1.70090568e+00 2.05438331e-01
9.53709543e-01 1.02825975e+00 3.78584564e-01 8.09230983e-01
2.33847141e-01 1.97905898e-01 -1.13678880e-01 2.90200055e-01
6.11196160e-01 -2.53879279e-01 -3.63120258e-01 1.97879717e-01
6.10014498e-01 -2.91688204e-01 -2.30098858e-01 1.26775622e+00
1.82022586e-01 -3.24711412e-01 4.86959010e-01 -8.71463656e-01
4.47835401e-02 5.88998675e-01 -5.53035736e-01 8.97977874e-02
-3.40598434e-01 3.50186199e-01 9.99539554e-01 -4.97880429e-01
3.07433426e-01 9.74916220e-01 6.87834084e-01 4.37936420e-03
-1.56043851e+00 -4.41636413e-01 3.08094829e-01 -3.17533195e-01
-1.34192693e+00 -4.85581368e-01 2.22213328e-01 -7.92792976e-01
7.38221705e-01 5.61126411e-01 6.72801971e-01 5.36614537e-01
2.62151003e-01 4.71726656e-01 1.46803415e+00 1.79485010e-03
4.66638319e-02 -1.55779749e-01 8.08082893e-02 4.59682196e-03
5.93560994e-01 4.36584145e-01 -4.23757046e-01 -1.38083007e-02
5.21607518e-01 4.10379201e-01 -5.20613670e-01 -4.88351360e-02
-1.21917999e+00 9.23476458e-01 9.61349368e-01 3.61583799e-01
-7.96336889e-01 1.80471748e-01 1.41676545e-01 6.33750379e-01
7.11872637e-01 2.95975506e-01 -6.39582753e-01 5.60365915e-01
-1.39964247e+00 4.19277042e-01 4.24246401e-01 2.52108634e-01
1.06160760e+00 2.81561106e-01 3.70819211e-01 2.87216336e-01
8.01451623e-01 1.27778769e+00 6.87157065e-02 -5.76858401e-01
5.29902518e-01 7.08484054e-01 2.55622953e-01 -7.80951560e-01
-5.52540362e-01 -5.34296155e-01 -1.30920136e+00 7.01640368e-01
1.46633849e-01 -4.19304043e-01 -1.08960176e+00 1.64921391e+00
1.31570930e-02 -3.54708284e-02 6.65519416e-01 1.16706765e+00
6.56431973e-01 1.06622982e+00 1.24937430e-01 -2.55338401e-01
1.58203900e+00 -3.36114913e-01 -6.77358747e-01 -9.71793056e-01
5.78656733e-01 -2.80482143e-01 2.40047172e-01 -9.72882565e-03
-5.72142720e-01 -5.55237055e-01 -1.42091238e+00 3.17340195e-01
-7.19227374e-01 2.51129419e-01 4.77846503e-01 2.29468629e-01
-1.19387841e+00 6.47581995e-01 -1.00235558e+00 -3.72551143e-01
2.16707081e-01 3.35499197e-01 -5.80702007e-01 -4.30391356e-02
-1.32617438e+00 9.03792500e-01 6.09783947e-01 8.68023098e-01
-8.75568807e-01 -6.00122273e-01 -1.18941414e+00 1.34000450e-01
3.39187235e-02 -4.45668161e-01 6.36693716e-01 -1.23454213e+00
-8.60001624e-01 9.91213918e-01 9.30031948e-03 -7.91675627e-01
3.08515459e-01 -4.53698546e-01 -5.55950522e-01 -1.07693739e-01
1.61290884e-01 7.53377557e-01 8.53083372e-01 -1.33109152e+00
-9.08011675e-01 -8.06656718e-01 -2.50982165e-01 3.13927472e-01
7.56768659e-02 -6.49510249e-02 2.77520299e-01 -3.87535989e-01
4.72928286e-01 -8.92130554e-01 -5.86462021e-01 8.21566805e-02
1.10621296e-01 3.46777052e-01 7.16695786e-01 -8.48800004e-01
9.32441831e-01 -2.40228677e+00 4.50596631e-01 1.44779310e-01
2.33035967e-01 3.74651223e-01 -2.48993158e-01 5.05324662e-01
-2.56222248e-01 9.63122174e-02 -1.07249641e+00 -1.74161270e-01
-1.51612595e-01 5.26661515e-01 -6.25233531e-01 7.58424878e-01
4.55583364e-01 8.76927853e-01 -7.53758967e-01 5.62409172e-04
2.03995541e-01 4.86806840e-01 3.18859547e-01 3.76660556e-01
-2.82698721e-01 3.80094260e-01 -3.04954171e-01 6.58161223e-01
1.19397974e+00 4.69904281e-02 3.52747470e-01 -2.78780192e-01
-6.92241311e-01 5.67868538e-02 -1.35527515e+00 1.58987737e+00
-1.53993532e-01 6.93401515e-01 6.46111667e-01 -6.90286517e-01
1.18825328e+00 3.73350859e-01 2.35317603e-01 -6.84207857e-01
-1.89437091e-01 5.72798967e-01 1.18868880e-01 -4.22709703e-01
6.53899848e-01 -4.11242574e-01 -1.74038559e-01 3.82111967e-01
-1.15814209e-01 -2.26456568e-01 4.82530110e-02 -2.14343727e-01
6.28402174e-01 2.47473240e-01 3.27450871e-01 -5.56428373e-01
2.85427272e-01 2.65750080e-01 6.50455832e-01 4.96195555e-01
-1.01625405e-01 5.43312430e-01 2.53316790e-01 -8.94666910e-01
-1.04979432e+00 -9.07215118e-01 -3.92046750e-01 6.80432022e-01
-2.21258551e-01 3.65228765e-02 -1.13471836e-01 -3.30292016e-01
3.04136276e-01 1.00099631e-01 -6.70307040e-01 2.53527910e-01
-4.48971361e-01 -1.42371142e+00 6.22256935e-01 4.19171512e-01
6.34423852e-01 -8.52965117e-01 -1.13231349e+00 3.78006756e-01
-1.47155553e-01 -9.98598039e-01 4.03391600e-01 7.35228539e-01
-1.25482249e+00 -9.33245838e-01 -6.45811379e-01 -3.12584281e-01
3.05475146e-01 2.38745719e-01 1.16911495e+00 -4.00900453e-01
2.90235598e-03 -9.65529084e-02 -3.58958960e-01 -2.98598826e-01
-2.89062917e-01 -2.72071152e-03 3.86625566e-02 3.45440388e-01
3.82071644e-01 -6.32987857e-01 -4.42107052e-01 4.77561206e-02
-1.31730199e+00 -3.22733857e-02 8.65641356e-01 5.84241033e-01
4.72390652e-01 -1.01188488e-01 1.08672053e-01 -2.70579278e-01
2.64509916e-02 -4.91167337e-01 -1.03849220e+00 1.17716402e-01
-1.87387854e-01 1.52753338e-01 -1.68777462e-02 3.06622326e-01
-6.89684153e-01 6.61307514e-01 1.46797210e-01 -3.65736969e-02
-3.68409485e-01 1.19430840e+00 -1.77474499e-01 7.79253319e-02
6.70034230e-01 3.68160039e-01 2.73350984e-01 -6.55898571e-01
1.40252113e-01 9.47453737e-01 6.85630560e-01 6.88126907e-02
1.07384753e+00 8.06810975e-01 1.18019685e-01 -1.17260313e+00
-6.61379457e-01 -5.49091578e-01 -8.02441597e-01 -8.87145996e-02
1.01348710e+00 -1.68564343e+00 -2.04135299e-01 7.69662559e-01
-9.98114407e-01 -3.16198349e-01 -1.89496934e-01 6.36774540e-01
-8.55478048e-02 1.23763047e-01 -9.71535221e-02 -8.96966755e-01
-4.58471209e-01 -1.11584318e+00 1.13655794e+00 2.58846581e-01
9.92693603e-02 -7.35491276e-01 4.02890712e-01 -1.03420213e-01
5.01659930e-01 8.01751494e-01 2.38281578e-01 -1.63464636e-01
-7.57088661e-01 -4.99004722e-02 -5.80085874e-01 4.81566966e-01
3.95123243e-01 -5.76415546e-02 -1.29583669e+00 -5.96149266e-01
3.70425433e-02 -1.96642876e-01 1.72098947e+00 5.61837792e-01
2.47547776e-01 -9.38498676e-02 -2.05366984e-01 8.67060423e-01
1.87758636e+00 -1.08442657e-01 7.33025551e-01 5.19291282e-01
4.31628108e-01 8.02238226e-01 4.02774155e-01 2.42843598e-01
1.73894852e-01 3.61026704e-01 9.41185772e-01 -4.99666125e-01
3.11517924e-01 3.85715753e-01 5.86564422e-01 2.58738130e-01
-1.11919463e-01 -1.12071484e-01 -1.14160049e+00 8.71589065e-01
-2.00339270e+00 -8.47856760e-01 -3.57164949e-01 2.34023690e+00
4.18558240e-01 -2.88610607e-01 -3.88206631e-01 1.96900710e-01
3.01749527e-01 9.80199337e-01 -4.88588423e-01 -5.36400005e-02
-6.80101216e-01 5.91685213e-02 1.22599018e+00 6.96857631e-01
-1.62878525e+00 7.17493951e-01 5.37707758e+00 -1.24307433e-02
-1.43225610e+00 2.18812772e-03 3.43602836e-01 2.26336122e-01
-2.96684593e-01 3.59964781e-02 -8.00591230e-01 1.00248754e-01
1.25094008e+00 2.75567889e-01 1.28178895e-01 2.91573673e-01
3.48650724e-01 -3.05437922e-01 -7.58360445e-01 6.74901128e-01
-2.03374267e-01 -1.25898182e+00 -6.78625777e-02 3.01940978e-01
6.28334641e-01 8.84219825e-01 -1.12475030e-01 -1.00717172e-01
3.74741435e-01 -1.14137316e+00 4.11711872e-01 7.47208178e-01
8.47678006e-01 -3.32310379e-01 1.02962518e+00 2.19562039e-01
-1.47220755e+00 -1.11855201e-01 -2.03051642e-01 -3.59761685e-01
2.75517404e-01 8.32475305e-01 -2.44766355e-01 9.51274514e-01
9.85857785e-01 1.17444730e+00 -3.34156662e-01 8.52876723e-01
-1.94908842e-01 4.12823111e-01 -8.14162016e-01 6.56447291e-01
6.99616134e-01 -3.39966536e-01 6.25166297e-01 9.97149289e-01
4.71243292e-01 2.30677515e-01 1.94000885e-01 7.17171967e-01
1.39721915e-01 -2.63737291e-01 -9.58252907e-01 -1.16634227e-01
1.75750032e-02 1.18260920e+00 -2.87883490e-01 -3.07212502e-01
-2.51216412e-01 5.89535356e-01 -2.29638830e-01 4.16387051e-01
-2.14412257e-01 -6.37052879e-02 1.28390121e+00 -9.38726738e-02
5.02972662e-01 -4.17881995e-01 -2.26400226e-01 -1.12334752e+00
1.20397277e-01 -7.34884977e-01 2.17455909e-01 -7.00553954e-01
-9.72170949e-01 8.49846780e-01 1.56548530e-01 -1.70892298e+00
-1.61495388e-01 -6.93962395e-01 -4.68136042e-01 1.38370764e+00
-2.34971476e+00 -1.38982666e+00 -5.59892595e-01 7.42497295e-02
1.65055157e-03 1.10459968e-01 1.00432944e+00 6.81645935e-04
-2.29157507e-01 -3.40217382e-01 5.42419136e-01 1.32127076e-01
4.62701887e-01 -1.09921825e+00 5.65833926e-01 1.17542076e+00
2.59585120e-02 -4.92968969e-03 7.09122241e-01 -5.92576087e-01
-1.38096142e+00 -1.47214818e+00 1.20306003e+00 7.14937374e-02
6.86178863e-01 -9.17816386e-02 -1.08135343e+00 8.58096302e-01
2.63699353e-01 2.34812692e-01 5.56772053e-01 -1.07291102e-01
-4.07101244e-01 -7.29177594e-01 -9.64989424e-01 5.44025637e-02
2.84533083e-01 -4.34216201e-01 -5.27327776e-01 3.00957978e-01
6.24546409e-01 -2.21310005e-01 -8.41531456e-01 8.14509332e-01
5.76665163e-01 -9.46105957e-01 6.97420120e-01 -3.93689066e-01
2.93002099e-01 -8.19564342e-01 -7.71527350e-01 -1.53585017e+00
-3.28730345e-01 -2.96785161e-02 4.44936305e-01 7.32003808e-01
4.40038234e-01 -7.81225562e-01 2.57558048e-01 8.01102445e-02
-6.87769204e-02 4.24285382e-02 -1.16224921e+00 -6.01931810e-01
-9.13935248e-03 -3.09506685e-01 6.44076288e-01 7.84891367e-01
-6.43612087e-01 5.63823394e-02 -4.67609614e-01 1.12739050e+00
7.81727016e-01 4.45600063e-01 5.97923756e-01 -1.75192153e+00
-1.01534026e-02 -2.30562866e-01 -4.81039673e-01 -7.30343640e-01
-2.14948937e-01 -6.58851922e-01 1.22266270e-01 -1.39641023e+00
-1.37464911e-01 -2.94712514e-01 -2.74763495e-01 8.03095400e-01
2.42291507e-03 5.02444565e-01 2.90061563e-01 3.78818661e-01
2.59931296e-01 5.62799692e-01 6.26483321e-01 -4.27608371e-01
-2.04963341e-01 -2.39885777e-01 -2.52874702e-01 5.53466082e-01
6.86036825e-01 -5.69017351e-01 3.00210446e-01 -9.31078911e-01
6.36663914e-01 1.61293134e-01 7.10325420e-01 -1.23709953e+00
-3.92832607e-02 -9.05744582e-02 2.02180058e-01 -8.76737654e-01
3.41371298e-01 -1.14269316e+00 5.65854371e-01 6.79676414e-01
1.10482708e-01 -8.77073258e-02 4.94119674e-01 3.71553063e-01
-5.64153433e-01 1.14524595e-01 8.70478332e-01 -2.35022530e-01
-9.11899984e-01 2.84414202e-01 -4.81143564e-01 -6.01132929e-01
6.55200541e-01 -2.73760222e-03 -2.17447758e-01 -7.86636919e-02
-6.94103420e-01 6.22450829e-01 2.44992927e-01 3.07678044e-01
3.97574037e-01 -9.46781576e-01 -1.47914600e+00 6.47555351e-01
5.43193042e-01 2.91835636e-01 4.11375940e-01 7.96183705e-01
-5.11404395e-01 3.82939041e-01 -3.56898278e-01 -1.02662325e+00
-1.11913526e+00 -2.40924656e-02 6.15889549e-01 -4.18911010e-01
-4.64668125e-01 4.99891698e-01 -1.95693187e-02 -5.73909223e-01
-3.87657017e-01 -7.14521587e-01 -3.38653058e-01 7.20706761e-01
7.87243307e-01 -3.49825740e-01 1.62361816e-01 -1.05941510e+00
-3.41584533e-01 6.45304859e-01 2.58470744e-01 -1.67681322e-01
1.78238690e+00 -1.48163348e-01 -3.28728974e-01 7.75556028e-01
8.96123350e-01 -5.79933584e-01 -1.42153871e+00 -5.06713629e-01
-1.02925725e-01 -3.19983423e-01 6.14970148e-01 -8.06542754e-01
-1.27316320e+00 9.63974297e-01 1.06777859e+00 1.97228536e-01
1.20921350e+00 -8.83210599e-02 3.35564524e-01 3.78438443e-01
-6.74064085e-02 -5.07443607e-01 -9.35528219e-01 4.69589412e-01
1.06260443e+00 -1.68264043e+00 2.29103014e-01 1.04543515e-01
-3.36016834e-01 1.14237535e+00 -9.00586993e-02 -1.10682309e-01
7.20640957e-01 2.57226169e-01 4.44266289e-01 -3.76513988e-01
-4.97252434e-01 -7.52913117e-01 1.79380193e-01 3.40977192e-01
1.82967901e-01 6.68012500e-01 -9.26778987e-02 5.92695065e-02
1.24709435e-01 9.91034061e-02 3.77432346e-01 1.00220907e+00
-7.22305596e-01 -8.58520150e-01 -7.87084401e-01 4.65753257e-01
1.22980282e-01 -2.36619845e-01 -2.29926184e-02 6.33979678e-01
-1.48763945e-02 6.22212350e-01 3.38709772e-01 -2.44140580e-01
6.06590621e-02 7.62935951e-02 -2.11873814e-01 -4.82558697e-01
-6.05144382e-01 1.41020313e-01 -5.39794117e-02 -4.93448526e-01
-1.08723104e+00 -9.57253933e-01 -6.21597290e-01 -1.07147068e-01
4.36399095e-02 2.10986659e-01 8.72214019e-01 1.02467239e+00
2.42603764e-01 2.76767462e-01 7.18028545e-01 -1.38420141e+00
-3.99195552e-01 -1.36672807e+00 -1.05618930e+00 -9.32463333e-02
1.17380977e+00 -3.24849099e-01 -6.16866052e-01 -5.57396747e-02] | [9.68061351776123, -1.6738847494125366] |
9c2de8b8-ddb9-43dd-804e-52e7f7fb6216 | explore-faster-localization-learning-for | 2207.01342 | null | https://arxiv.org/abs/2207.01342v1 | https://arxiv.org/pdf/2207.01342v1.pdf | Explore Faster Localization Learning For Scene Text Detection | Generally pre-training and long-time training computation are necessary for obtaining a good-performance text detector based on deep networks. In this paper, we present a new scene text detection network (called FANet) with a Fast convergence speed and Accurate text localization. The proposed FANet is an end-to-end text detector based on transformer feature learning and normalized Fourier descriptor modeling, where the Fourier Descriptor Proposal Network and Iterative Text Decoding Network are designed to efficiently and accurately identify text proposals. Additionally, a Dense Matching Strategy and a well-designed loss function are also proposed for optimizing the network performance. Extensive experiments are carried out to demonstrate that the proposed FANet can achieve the SOTA performance with fewer training epochs and no pre-training. When we introduce additional data for pre-training, the proposed FANet can achieve SOTA performance on MSRATD500, CTW1500 and TotalText. The ablation experiments also verify the effectiveness of our contributions. | ['Weiqiang Wang', 'Weijia Wu', 'Yuanqiang Cai', 'Yuzhong Zhao'] | 2022-07-04 | null | null | null | null | ['scene-text-detection'] | ['computer-vision'] | [-7.11600259e-02 -6.65627122e-01 -8.11074525e-02 -4.72047448e-01
-5.39007008e-01 -2.01370474e-02 7.49621630e-01 -1.13994770e-01
-5.83433151e-01 4.76372009e-03 7.35122040e-02 -9.23015103e-02
2.02195123e-01 -5.36235332e-01 -5.31440556e-01 -3.66586864e-01
4.80294406e-01 4.68234360e-01 5.71178734e-01 7.08307773e-02
3.64359558e-01 2.14739248e-01 -1.19349706e+00 -3.67051666e-03
9.11981583e-01 1.22499073e+00 6.70581102e-01 7.12226510e-01
-3.07633698e-01 7.47626722e-01 -6.27912223e-01 -3.76163691e-01
3.11535925e-01 -7.76189491e-02 -3.99670810e-01 2.35238388e-01
5.99634409e-01 -9.05629456e-01 -8.70012522e-01 8.20695579e-01
9.23097908e-01 1.57672346e-01 7.23938048e-01 -7.68253863e-01
-4.23916996e-01 7.04642832e-01 -7.95967400e-01 1.62752867e-02
1.74251780e-01 8.90472457e-02 8.94468963e-01 -1.39210379e+00
1.75456941e-01 1.18553388e+00 9.33082700e-01 2.21670538e-01
-5.61668336e-01 -8.73792291e-01 1.57756954e-01 8.22395459e-02
-1.79136956e+00 -4.46766287e-01 5.60147524e-01 -7.85987750e-02
1.04063892e+00 -3.77572775e-02 3.82662088e-01 1.02430129e+00
1.85660914e-01 1.41901660e+00 2.91558057e-01 -5.05272329e-01
-6.16218671e-02 7.43681565e-02 1.42540991e-01 1.06238174e+00
4.96371269e-01 -2.74317950e-01 -7.85575569e-01 9.09420848e-02
8.42434645e-01 3.58920008e-01 -1.60095453e-01 -8.50194693e-03
-1.17037427e+00 6.82364106e-01 3.26658428e-01 2.63848960e-01
-5.20194992e-02 4.32311565e-01 8.13082337e-01 -1.46745546e-02
4.28727686e-01 -1.28953591e-01 -1.03123255e-01 -2.11681977e-01
-1.43819547e+00 2.46563200e-02 6.06472969e-01 1.26231396e+00
3.48508000e-01 1.97309777e-01 -4.61841196e-01 9.88324583e-01
5.19829452e-01 1.05683553e+00 8.14283907e-01 -7.32081234e-02
7.73660481e-01 6.58243716e-01 -1.81775168e-01 -1.05166757e+00
-4.52193648e-01 -5.44438958e-01 -9.78156209e-01 -4.07477409e-01
4.48822863e-02 -1.03708185e-01 -1.15329623e+00 9.76816595e-01
1.06631264e-01 2.25960761e-01 -2.15365797e-01 9.90434706e-01
7.53321409e-01 6.58208966e-01 -2.50878692e-01 1.97054982e-01
1.29785037e+00 -1.13081372e+00 -8.62038910e-01 -3.42315912e-01
9.27993536e-01 -1.05877817e+00 1.33976853e+00 3.91276300e-01
-8.31909955e-01 -5.20955086e-01 -1.23883426e+00 -5.19363701e-01
-1.66442707e-01 8.90642405e-01 5.37001312e-01 5.65744400e-01
-7.48036742e-01 4.31649834e-01 -1.09376502e+00 -6.11558735e-01
5.77394426e-01 6.02139950e-01 2.32413501e-01 -3.82050984e-02
-8.79087627e-01 4.43261385e-01 4.87993479e-01 2.32780620e-01
-7.52424181e-01 -1.98546290e-01 -6.37173295e-01 4.00922596e-01
2.58185238e-01 -4.44259971e-01 1.28424764e+00 -6.55816495e-01
-1.79255104e+00 6.82394147e-01 -9.91610214e-02 -5.83214462e-01
7.14180589e-01 -4.35414046e-01 -3.89993489e-01 2.91636109e-01
8.98479298e-02 5.77653170e-01 1.10012996e+00 -4.69008386e-01
-7.33534575e-01 -2.90547699e-01 -6.44076407e-01 3.58496636e-01
-1.05415273e+00 3.12743932e-02 -1.18478739e+00 -9.26195383e-01
2.59743243e-01 -6.45716786e-01 1.91487953e-01 4.09168661e-01
-7.48622119e-01 -4.19205964e-01 1.23997521e+00 -5.32266557e-01
1.40390384e+00 -2.15607619e+00 -3.81577015e-01 2.16379642e-01
4.45637405e-01 4.76686478e-01 -2.92567164e-02 3.11218143e-01
4.72516894e-01 -1.47082329e-01 1.17649890e-01 -7.50253320e-01
3.32202762e-01 -3.24101061e-01 -4.24849212e-01 8.09601009e-01
-1.03314355e-01 8.32146049e-01 -2.01060638e-01 -8.00145864e-01
5.63060164e-01 4.70014572e-01 -3.02593052e-01 -1.01242634e-02
-2.60757118e-01 -4.24558967e-01 -7.71475971e-01 6.45398319e-01
9.49164987e-01 -6.25542581e-01 -1.91010073e-01 -3.09440523e-01
4.83536758e-02 1.89983472e-01 -1.14714980e+00 1.81881571e+00
-2.69067585e-01 1.09122181e+00 -1.42267086e-02 -6.30784750e-01
1.01382017e+00 -7.58006126e-02 1.87136233e-01 -1.07633138e+00
7.52928972e-01 2.49875560e-01 -5.45574546e-01 -3.03244114e-01
9.44440126e-01 3.23486209e-01 1.32760525e-01 5.99457920e-01
-7.22710714e-02 1.92617387e-01 7.79133961e-02 4.29571867e-01
9.81166065e-01 -6.68383762e-02 -2.25238234e-01 -2.71562576e-01
6.49688303e-01 -6.54092133e-02 2.58687973e-01 8.27498972e-01
-4.18784022e-02 5.22208810e-01 1.00358777e-01 -4.10577178e-01
-1.03933442e+00 -5.60341179e-01 -4.17774588e-01 1.13773131e+00
5.36210060e-01 -6.86298251e-01 -7.01933861e-01 -7.53765821e-01
-1.15379151e-02 3.34051579e-01 -3.00191700e-01 -1.03607297e-01
-5.85093260e-01 -5.85780144e-01 9.36379075e-01 7.35594928e-01
1.16175103e+00 -5.98216891e-01 -3.22685629e-01 -5.00411578e-02
1.88391767e-02 -1.37119782e+00 -1.06189919e+00 4.31552865e-02
-7.53537118e-01 -7.67740965e-01 -9.55846548e-01 -1.16536927e+00
7.03848064e-01 7.09178329e-01 5.52754760e-01 3.20868492e-01
-3.84524494e-01 2.47967839e-01 -4.57078129e-01 -3.36706132e-01
-3.41090523e-02 3.94197702e-01 -2.39741169e-02 -1.02006085e-01
5.97450137e-01 -1.66646987e-01 -7.77251720e-01 5.10485232e-01
-9.65235591e-01 2.63277203e-01 7.86043644e-01 9.31627572e-01
3.74566078e-01 1.23578690e-01 8.83808434e-02 -4.90403593e-01
6.55445874e-01 1.27754346e-01 -8.87077630e-01 2.25931436e-01
-7.03100026e-01 1.26701459e-01 8.46508145e-01 -5.03700912e-01
-9.32707012e-01 1.25340566e-01 -1.35567307e-01 -7.32315481e-01
3.40528339e-01 4.10428941e-01 9.34109241e-02 -2.30509117e-01
7.16757715e-01 6.87809765e-01 -1.89798936e-01 -5.47548234e-01
8.06692615e-02 1.09378469e+00 4.21340108e-01 -3.38681430e-01
9.70223725e-01 6.16407812e-01 -2.43983418e-01 -8.86417389e-01
-6.31051123e-01 -7.05377758e-01 -4.87222195e-01 1.55493006e-01
5.32178581e-01 -1.20326972e+00 -9.29140449e-01 9.45152223e-01
-9.76816416e-01 -3.94187182e-01 3.61848265e-01 6.48686886e-01
-3.15498918e-01 7.77452946e-01 -6.83084905e-01 -6.39286041e-01
-1.07887793e+00 -9.73094106e-01 1.50015593e+00 2.50642717e-01
4.18724328e-01 -9.79625881e-01 -2.16549888e-01 3.57708745e-02
3.33781123e-01 -5.79307914e-01 3.46206516e-01 -7.10788012e-01
-5.59302986e-01 -4.61463064e-01 -7.55639136e-01 5.80489337e-02
-4.65704016e-02 -1.98907685e-02 -8.74756038e-01 -5.00220954e-01
-2.48231605e-01 -4.23278242e-01 1.14648294e+00 4.00805742e-01
1.17612278e+00 1.01190777e-02 -5.27329624e-01 9.76381421e-01
1.46090412e+00 -1.20974239e-02 3.57671261e-01 3.85504335e-01
9.69248056e-01 -3.08701564e-02 6.09928012e-01 8.25410545e-01
3.09394687e-01 6.35350704e-01 3.11743394e-02 -2.15136468e-01
-1.38835713e-01 -3.74201328e-01 3.79236609e-01 9.18133914e-01
7.23299325e-01 -6.43500388e-01 -9.11513388e-01 2.94458747e-01
-1.85088062e+00 -6.72189593e-01 -1.36087522e-01 2.07906842e+00
5.39605498e-01 3.92438501e-01 -3.74921709e-02 1.95779115e-01
9.01803136e-01 1.41810685e-01 -8.35267007e-01 9.60384607e-02
-1.03471115e-01 1.21951938e-01 7.06411719e-01 2.72440523e-01
-1.34359372e+00 1.41657233e+00 5.90875769e+00 1.38227141e+00
-1.43132913e+00 -7.59744644e-02 3.15990746e-01 -1.84354797e-01
2.70494342e-01 -4.05867070e-01 -1.25319672e+00 3.61005247e-01
6.76857889e-01 -1.57462463e-01 2.90077478e-01 9.65290189e-01
3.08401048e-01 3.64604853e-02 -7.90353119e-01 1.44196141e+00
2.95653790e-01 -1.40797198e+00 2.09170748e-02 -1.37623459e-01
3.09938610e-01 3.22798818e-01 6.74018189e-02 3.63866627e-01
1.30685687e-01 -7.86493659e-01 8.65223050e-01 1.80043370e-01
1.08440340e+00 -6.91999853e-01 5.36272287e-01 3.08563143e-01
-1.53098810e+00 1.01628555e-02 -7.38079309e-01 3.36924225e-01
-1.74637154e-01 6.39185309e-01 -1.04644454e+00 2.78469801e-01
4.89802986e-01 1.11213326e+00 -7.69461811e-01 1.11316872e+00
-5.13047120e-03 7.29458690e-01 -6.91595852e-01 -7.23194659e-01
3.85726690e-01 1.04693517e-01 2.42210746e-01 1.44489884e+00
2.68538296e-01 -2.80211091e-01 3.38958025e-01 7.34152019e-01
-3.58487368e-01 3.20193172e-01 -1.53929755e-01 -1.61407471e-01
4.45301980e-01 1.27236259e+00 -1.03480804e+00 -4.53969717e-01
-3.60661477e-01 1.39247549e+00 1.62591100e-01 2.38326356e-01
-9.10447896e-01 -1.05313253e+00 2.09220394e-01 6.55162185e-02
5.56337118e-01 -3.10936838e-01 -2.76548207e-01 -1.48323405e+00
2.24549189e-01 -5.86376727e-01 1.24875724e-01 -8.24312687e-01
-1.06851840e+00 4.02039647e-01 -5.72342098e-01 -1.26363063e+00
1.39705300e-01 -8.11035872e-01 -7.00134695e-01 7.43911207e-01
-1.54373538e+00 -1.28011441e+00 -6.80965662e-01 8.25532675e-01
8.10041964e-01 -3.27169955e-01 2.97247738e-01 4.17890310e-01
-1.09449661e+00 1.16896307e+00 5.89693546e-01 5.80556214e-01
7.62389183e-01 -9.17876065e-01 7.24528253e-01 9.04559195e-01
5.40021136e-02 4.29743290e-01 4.56701398e-01 -7.54102707e-01
-1.69682276e+00 -1.29331303e+00 4.58042592e-01 7.25739868e-03
5.36855876e-01 -8.25724244e-01 -8.74153674e-01 5.64874291e-01
-4.25792299e-02 -1.70551483e-02 1.08228996e-01 -1.83704514e-02
-2.28124112e-01 -1.42883196e-01 -8.30392540e-01 6.33208990e-01
9.67764318e-01 -5.13583899e-01 -2.58283854e-01 7.24338293e-01
7.19709337e-01 -7.13596463e-01 -4.34409022e-01 -7.15807155e-02
4.73134488e-01 -6.07021987e-01 5.92932999e-01 1.31011650e-01
1.59689471e-01 -7.42662549e-02 5.60401864e-02 -4.65251237e-01
-1.32190898e-01 -7.67517745e-01 -5.43001369e-02 1.09908342e+00
7.74347484e-02 -5.57330132e-01 1.20562923e+00 5.62728271e-02
-2.57803887e-01 -7.03301609e-01 -7.11212635e-01 -7.87858248e-01
-5.99201620e-02 -5.38685203e-01 4.64008957e-01 6.49343669e-01
-6.82503507e-02 5.33757925e-01 -4.89999324e-01 3.20063345e-02
5.72408080e-01 -2.16848757e-02 8.98797870e-01 -1.05998456e+00
-1.96418725e-02 -4.72041696e-01 -5.20292819e-01 -1.94939840e+00
1.10183313e-01 -7.50911891e-01 2.68407345e-01 -1.41563892e+00
3.65518570e-01 -4.20600146e-01 1.01257786e-01 3.46639752e-01
-1.93288475e-01 -1.53563261e-01 8.85712951e-02 4.52553064e-01
-1.06286573e+00 1.07538366e+00 1.14100873e+00 -1.54453382e-01
-1.73764676e-01 -2.77065001e-02 -2.94018239e-01 5.87251902e-01
5.65523267e-01 -5.20406485e-01 -4.30044293e-01 -7.41457760e-01
1.11481979e-01 -1.30080774e-01 1.56950310e-01 -1.09494841e+00
7.47419417e-01 3.94034147e-01 6.52787805e-01 -1.30908501e+00
2.84869820e-01 -7.74476945e-01 -6.98653996e-01 4.49049890e-01
-3.33121657e-01 -1.11358978e-01 2.42628172e-01 6.43418729e-01
-3.68057005e-02 -3.05857211e-01 7.41744518e-01 5.14519215e-01
-5.48373640e-01 5.60509980e-01 -2.56615281e-01 -1.42216086e-02
6.38271689e-01 -2.73118258e-01 -3.88176322e-01 -3.74813467e-01
5.15315086e-02 5.01908302e-01 4.40404624e-01 3.31525743e-01
8.73567104e-01 -1.27553320e+00 -5.57954788e-01 2.40459159e-01
5.08424733e-03 6.62783906e-02 7.35997334e-02 7.71040499e-01
-8.93802762e-01 9.02146995e-01 2.76691616e-01 -7.96877027e-01
-1.15832567e+00 2.94127405e-01 4.61137354e-01 -4.41829234e-01
-1.12714660e+00 7.31947660e-01 1.70152619e-01 -4.30193484e-01
7.04291761e-01 -4.47483033e-01 2.61989385e-01 -4.50948209e-01
6.72204137e-01 2.42288068e-01 1.47577599e-01 -2.95329630e-01
-2.98966020e-01 7.82397985e-01 -6.11036301e-01 -3.85952145e-02
9.43518937e-01 -2.51455069e-01 4.04836953e-01 6.48245681e-03
1.30831921e+00 -1.20184980e-01 -1.13084531e+00 -6.88969135e-01
-1.67780817e-01 -4.42982167e-01 6.75318122e-01 -5.58468461e-01
-1.13970160e+00 1.11909616e+00 8.13633204e-01 -2.46241197e-01
1.19093919e+00 -4.85223532e-01 1.27234530e+00 9.59915876e-01
5.12234271e-02 -1.36043477e+00 5.28320849e-01 7.70613253e-01
5.18626630e-01 -1.23667979e+00 2.97782421e-01 -2.53984451e-01
-4.62674201e-01 1.38578713e+00 7.26621151e-01 -1.38337925e-01
5.01551807e-01 3.29932928e-01 -7.21674114e-02 -2.93743104e-01
-5.98126590e-01 -1.26046389e-01 2.44125590e-01 7.38449097e-02
2.35448331e-01 -3.87820631e-01 -5.60089797e-02 4.20699924e-01
-1.52580999e-02 -1.54137230e-02 2.13873386e-01 1.00117588e+00
-8.26174378e-01 -7.14119852e-01 -3.11650664e-01 7.70012498e-01
-3.70702624e-01 -3.98856014e-01 -4.09996659e-01 7.88476646e-01
-5.50637245e-01 6.77370131e-01 2.61337787e-01 -6.10868394e-01
1.81421235e-01 -2.98028409e-01 1.54116258e-01 -3.47651541e-01
-5.55655599e-01 5.13146520e-01 -2.68319309e-01 -2.79398233e-01
6.65798113e-02 -3.21501136e-01 -1.71110547e+00 -8.39475095e-01
-1.08024359e+00 -1.36304051e-01 8.60210776e-01 9.72456038e-01
5.08997798e-01 3.32995087e-01 8.48603427e-01 -6.40385747e-01
-6.25680149e-01 -1.17037058e+00 -6.88593090e-01 -7.76781291e-02
7.19755962e-02 -3.96362692e-01 -3.21203351e-01 -5.56704514e-02] | [12.007072448730469, 2.2172110080718994] |
d797d544-d032-407e-ba19-a1ebb8767ae3 | malunet-a-multi-attention-and-light-weight | 2211.01784 | null | https://arxiv.org/abs/2211.01784v1 | https://arxiv.org/pdf/2211.01784v1.pdf | MALUNet: A Multi-Attention and Light-weight UNet for Skin Lesion Segmentation | Recently, some pioneering works have preferred applying more complex modules to improve segmentation performances. However, it is not friendly for actual clinical environments due to limited computing resources. To address this challenge, we propose a light-weight model to achieve competitive performances for skin lesion segmentation at the lowest cost of parameters and computational complexity so far. Briefly, we propose four modules: (1) DGA consists of dilated convolution and gated attention mechanisms to extract global and local feature information; (2) IEA, which is based on external attention to characterize the overall datasets and enhance the connection between samples; (3) CAB is composed of 1D convolution and fully connected layers to perform a global and local fusion of multi-stage features to generate attention maps at channel axis; (4) SAB, which operates on multi-stage features by a shared 2D convolution to generate attention maps at spatial axis. We combine four modules with our U-shape architecture and obtain a light-weight medical image segmentation model dubbed as MALUNet. Compared with UNet, our model improves the mIoU and DSC metrics by 2.39% and 1.49%, respectively, with a 44x and 166x reduction in the number of parameters and computational complexity. In addition, we conduct comparison experiments on two skin lesion segmentation datasets (ISIC2017 and ISIC2018). Experimental results show that our model achieves state-of-the-art in balancing the number of parameters, computational complexity and segmentation performances. Code is available at https://github.com/JCruan519/MALUNet. | ['Yuzhuo Fu', 'Ting Liu', 'Mingye Xie', 'Suncheng Xiang', 'Jiacheng Ruan'] | 2022-11-03 | null | null | null | null | ['skin-lesion-segmentation'] | ['medical'] | [ 4.09237921e-01 1.02997579e-01 -6.74412027e-02 -2.59761035e-01
-8.68735731e-01 -1.31897658e-01 2.09322169e-01 1.92553326e-01
-6.41119242e-01 4.01831865e-01 -6.60646409e-02 -4.16149497e-01
-6.94946796e-02 -8.50037575e-01 -5.05626976e-01 -8.24372649e-01
1.33040309e-01 3.76285501e-02 5.13726056e-01 2.65960936e-02
1.05917290e-01 3.69275719e-01 -9.47808027e-01 3.11491787e-01
1.30476725e+00 1.20383012e+00 4.17922705e-01 6.33091390e-01
-1.51426002e-01 5.04412174e-01 -8.09668228e-02 -3.40970039e-01
3.49187776e-02 -4.76029098e-01 -1.02314413e+00 3.69812883e-02
-9.07188430e-02 -2.69806743e-01 -2.65112162e-01 1.06598711e+00
8.83934438e-01 -1.23034365e-01 4.73965704e-01 -6.66070879e-01
-4.91951048e-01 5.72979867e-01 -9.39926803e-01 3.70393634e-01
-1.78893402e-01 3.01244915e-01 5.15823007e-01 -6.10652864e-01
4.10475194e-01 8.08684766e-01 7.48194993e-01 5.76898873e-01
-9.09876883e-01 -5.01551688e-01 8.82704854e-02 1.57431215e-01
-1.39620626e+00 -3.20713259e-02 5.06548524e-01 -1.68241426e-01
7.76221931e-01 5.16854286e-01 4.80130911e-01 8.17728221e-01
2.51251101e-01 8.80658269e-01 1.03942645e+00 -3.04998428e-01
9.88660157e-02 -1.50723025e-01 1.43174872e-01 8.08721423e-01
3.52617577e-02 -3.55080903e-01 1.07712038e-01 1.19039714e-01
1.01518011e+00 1.91334561e-02 -4.19660091e-01 1.27544805e-01
-9.97715592e-01 8.64618957e-01 9.35810208e-01 4.93576795e-01
-5.15404701e-01 6.79656863e-02 3.67928296e-01 -1.99088857e-01
4.02719140e-01 2.75236875e-01 -1.70686632e-01 3.46640795e-02
-6.51565731e-01 -1.63805738e-01 3.19753587e-01 7.38998830e-01
4.13620412e-01 -4.11905587e-01 -4.78196025e-01 9.88748252e-01
1.72576323e-01 2.25061566e-01 7.55444586e-01 -5.98433673e-01
2.73912251e-01 8.90002131e-01 -2.88735807e-01 -6.71422601e-01
-8.08071375e-01 -7.24877477e-01 -1.07473814e+00 -1.72219992e-01
2.92363226e-01 -2.81705320e-01 -1.33083963e+00 1.44288588e+00
2.90273219e-01 2.86220253e-01 -3.08416158e-01 1.15007424e+00
1.03257930e+00 3.57936502e-01 3.39877814e-01 5.64251430e-02
1.50474620e+00 -1.30265319e+00 -5.54884911e-01 -1.44020543e-01
8.28207016e-01 -8.87615383e-01 9.56048667e-01 -1.34640490e-03
-1.32184374e+00 -5.20552099e-01 -8.17068636e-01 -1.73864752e-01
-2.71877497e-01 3.28931063e-01 7.35807002e-01 5.49320877e-01
-1.07085729e+00 3.85837287e-01 -1.06301725e+00 -4.10031378e-01
8.61327767e-01 5.05096555e-01 -7.16638565e-02 2.88436282e-02
-9.98055756e-01 6.68996334e-01 2.92299271e-01 3.11855763e-01
-6.34881675e-01 -8.89221966e-01 -8.48352611e-01 6.55899197e-02
2.49148503e-01 -7.40521967e-01 1.09325182e+00 -7.12778687e-01
-1.46972120e+00 8.10184956e-01 2.17834581e-02 -2.78236002e-01
6.21200383e-01 -3.07633225e-02 -1.76337793e-01 3.38241428e-01
5.28165027e-02 8.05493295e-01 2.26791337e-01 -8.85020256e-01
-6.63881481e-01 -3.71937722e-01 3.81974168e-02 3.03222060e-01
-2.98311800e-01 -1.66020368e-03 -1.11968672e+00 -6.22840464e-01
1.93127692e-01 -8.22008967e-01 -6.43627405e-01 1.14694640e-01
-6.01766765e-01 1.64405912e-01 5.79562306e-01 -8.87509584e-01
1.44780922e+00 -2.00636005e+00 -4.44602557e-02 2.91437536e-01
3.08360279e-01 6.35100961e-01 -1.63907111e-01 -4.25646678e-02
-1.17223620e-01 3.58700842e-01 -5.60297668e-01 -2.41808414e-01
-3.36778998e-01 -1.44226044e-01 3.59182507e-01 4.41158086e-01
3.24877322e-01 1.30333102e+00 -8.08706820e-01 -7.30827034e-01
4.07284081e-01 7.33934820e-01 -5.86029172e-01 3.71361077e-02
1.85293540e-01 5.68271041e-01 -6.74189031e-01 7.43661761e-01
8.51009727e-01 -5.47604620e-01 7.20577687e-02 -3.15374881e-01
-7.85045978e-03 -3.51453349e-02 -8.81429732e-01 1.88999760e+00
-5.24972618e-01 2.23733261e-01 2.26388946e-01 -8.97470772e-01
6.32764995e-01 2.69418061e-01 7.25747764e-01 -7.75867760e-01
7.76883662e-01 2.41392031e-01 1.60097703e-01 -6.31517231e-01
7.19134435e-02 1.22609563e-01 6.42799437e-02 1.01280928e-01
6.99855685e-02 2.03797624e-01 1.92573428e-01 4.36601564e-02
1.13970578e+00 -4.55195941e-02 1.03519343e-01 -4.17184979e-01
7.37500966e-01 -1.79834798e-01 5.02673686e-01 4.83352005e-01
-3.15858334e-01 6.91192329e-01 4.60123718e-01 -2.46685326e-01
-7.63616800e-01 -8.77272725e-01 -3.73932600e-01 6.54766321e-01
4.21951950e-01 -1.37477681e-01 -1.08606637e+00 -6.12094223e-01
-2.80453861e-01 3.52697194e-01 -9.05503511e-01 -3.67912650e-02
-5.37343621e-01 -1.19569838e+00 3.68048638e-01 8.40524495e-01
8.03584576e-01 -1.12189746e+00 -7.47449696e-01 2.28451133e-01
-2.07797468e-01 -1.01856065e+00 -8.01853836e-01 1.10719083e-02
-8.06707561e-01 -1.00941300e+00 -1.10475433e+00 -9.85727608e-01
9.66779351e-01 8.93383399e-02 7.22304285e-01 2.64279366e-01
-7.87289083e-01 -1.13554142e-01 -4.12047595e-01 -4.29905534e-01
2.64109839e-02 3.69029641e-01 -6.94525898e-01 2.31379360e-01
1.69138059e-01 -4.15963113e-01 -1.04751563e+00 2.81952947e-01
-1.06484878e+00 3.42107892e-01 9.30482447e-01 9.25554514e-01
7.05776632e-01 -2.69752175e-01 4.26160932e-01 -9.98888075e-01
5.82074940e-01 -3.73279691e-01 -3.37546229e-01 2.30201855e-01
-3.03036332e-01 -2.35635653e-01 3.96250904e-01 -2.88537860e-01
-1.00432932e+00 8.10916051e-02 -3.60658884e-01 -2.07286432e-01
-3.92097272e-02 3.91195893e-01 6.65254369e-02 -2.92556822e-01
4.76477981e-01 2.77422398e-01 1.37409613e-01 -4.26285058e-01
1.64378420e-01 7.80879676e-01 4.38726038e-01 -2.87084103e-01
3.71929884e-01 5.31394005e-01 -1.24246888e-01 -5.68463206e-01
-6.61439836e-01 -4.71667975e-01 -6.19678438e-01 -1.47039175e-01
1.18487179e+00 -7.25938320e-01 -6.89819217e-01 8.25391531e-01
-8.51541340e-01 -4.87592369e-01 -1.66261733e-01 5.01280546e-01
-3.78151745e-01 3.14332128e-01 -9.85411584e-01 -3.76823336e-01
-9.64242756e-01 -1.49337792e+00 9.85550940e-01 8.17760348e-01
8.72777700e-02 -1.00187969e+00 -3.61226410e-01 3.27685565e-01
7.02833235e-01 4.96498048e-01 7.89430320e-01 -3.82774621e-01
-3.70453179e-01 -2.00426817e-01 -6.00112379e-01 2.20343068e-01
1.18515834e-01 -2.11064994e-01 -8.36616457e-01 -4.01399523e-01
-3.47015202e-01 -1.28471181e-01 9.86976743e-01 8.88271987e-01
1.73779333e+00 1.45672798e-01 -5.70520580e-01 1.03575647e+00
1.52908349e+00 4.09816355e-01 7.41738319e-01 4.91130762e-02
6.77287817e-01 3.26911032e-01 2.81208545e-01 2.58191496e-01
3.86298388e-01 3.81352812e-01 4.74234819e-01 -8.44568551e-01
-3.86062145e-01 8.55603963e-02 -2.52026469e-01 9.23545301e-01
-1.75714895e-01 -6.41494617e-02 -9.55658972e-01 8.19177747e-01
-1.79008543e+00 -4.45213556e-01 -9.37470421e-02 1.78942275e+00
7.80380368e-01 1.24558687e-01 1.54151013e-02 -6.59986660e-02
8.21636379e-01 -2.72815265e-02 -6.66605592e-01 -4.03486758e-01
1.36951566e-01 5.77974737e-01 5.70383906e-01 3.92880678e-01
-1.28383541e+00 7.46642053e-01 4.92950964e+00 1.11623526e+00
-1.32180560e+00 2.18928114e-01 1.10904527e+00 -3.41039710e-02
-1.73328016e-02 -3.66165578e-01 -5.14098823e-01 6.65248632e-01
5.05981743e-01 1.09442167e-01 1.47348017e-01 4.88000512e-01
-2.10128445e-02 -5.16459998e-03 -4.87017959e-01 7.95904219e-01
-3.66769806e-02 -1.32425499e+00 -5.56714758e-02 -1.69684691e-03
7.05253661e-01 6.09562993e-02 3.56373712e-02 1.64945573e-01
2.91063972e-02 -1.11463034e+00 1.89447865e-01 5.12867153e-01
9.39589620e-01 -7.51409471e-01 1.05655241e+00 2.71986593e-02
-1.14163673e+00 1.86041519e-02 -1.78961039e-01 4.60057646e-01
1.34833589e-01 4.87534791e-01 -6.52042747e-01 8.39691818e-01
7.19852269e-01 4.08925056e-01 -6.02415740e-01 1.39286089e+00
-1.41495898e-01 5.66851020e-01 -2.55789608e-01 -1.49787053e-01
5.96639514e-01 -1.71446964e-01 2.25789845e-01 1.22497535e+00
2.28762552e-01 2.77051151e-01 3.20317447e-02 7.69259930e-01
-6.24922998e-02 2.00145513e-01 2.32446820e-01 4.06484783e-01
2.10411638e-01 1.67219925e+00 -9.93612468e-01 -2.27270424e-01
-3.59973848e-01 9.51092839e-01 8.39544684e-02 3.02488536e-01
-1.13995266e+00 -8.33444536e-01 3.59591872e-01 2.19048068e-01
3.06468368e-01 3.60068798e-01 -4.95111793e-01 -8.65732491e-01
-2.96545848e-02 -5.40681005e-01 4.73549217e-01 -4.77337599e-01
-1.03491163e+00 9.40264940e-01 -4.01267856e-01 -8.92641127e-01
2.74725497e-01 -5.22178411e-01 -8.76760006e-01 1.19994068e+00
-1.69200242e+00 -1.30228293e+00 -7.51685143e-01 5.34990072e-01
4.32180673e-01 2.50842124e-01 7.70013452e-01 4.73297864e-01
-9.36910987e-01 8.84941638e-01 -8.56265351e-02 3.61221850e-01
5.05722404e-01 -1.13998520e+00 3.89784187e-01 7.64777839e-01
-6.02744937e-01 3.90847951e-01 -1.42620459e-01 -4.31996971e-01
-9.04607475e-01 -1.32750392e+00 6.17617190e-01 5.51536232e-02
5.10769546e-01 -1.41618341e-01 -7.52731800e-01 3.18856329e-01
1.34106308e-01 3.60441923e-01 6.74778402e-01 -2.36678258e-01
1.66083336e-01 -3.33860889e-02 -1.32955563e+00 5.99436402e-01
8.94369304e-01 -1.62041217e-01 5.35307340e-02 2.22661212e-01
6.79848433e-01 -7.81670153e-01 -1.12533140e+00 5.82596362e-01
4.34997946e-01 -6.39802516e-01 8.10811996e-01 -2.23790407e-01
5.78954339e-01 -2.37445384e-01 1.95468754e-01 -9.73846734e-01
-5.81185400e-01 -4.88261491e-01 2.14219078e-01 8.56595099e-01
5.30080140e-01 -7.49920130e-01 6.68141007e-01 3.97661090e-01
-5.15665293e-01 -1.72650695e+00 -7.81899869e-01 -2.83496588e-01
2.20307007e-01 -1.32280335e-01 5.97644746e-01 8.34737003e-01
-1.83590293e-01 1.14916466e-01 1.10162176e-01 4.67808247e-02
4.33729112e-01 8.01078975e-02 1.33663699e-01 -6.61769748e-01
-1.50072366e-01 -8.17953110e-01 -2.44465128e-01 -8.64485681e-01
-4.34864670e-01 -1.13083470e+00 -1.84881210e-01 -1.80627155e+00
4.57978100e-01 -3.96606296e-01 -5.80912828e-01 7.10723579e-01
-5.41482925e-01 5.47034025e-01 1.58753052e-01 -1.08323105e-01
-4.98448998e-01 2.38706127e-01 1.77958703e+00 -9.07777026e-02
-2.24389955e-01 -8.36749673e-02 -1.05292475e+00 6.76013768e-01
9.85229135e-01 2.48005381e-03 -2.72327363e-01 -6.05771840e-01
-4.89338696e-01 2.45530698e-02 3.22927386e-01 -1.12271702e+00
2.80572206e-01 1.57959893e-01 6.50091946e-01 -5.05410492e-01
8.23697522e-02 -5.18885016e-01 -2.05247417e-01 8.00070524e-01
-1.95029616e-01 -3.06850225e-01 4.51545984e-01 1.15094237e-01
-1.80084974e-01 -8.58313441e-02 1.17903876e+00 -2.07865741e-02
-5.41626811e-01 6.98463738e-01 -5.11093736e-02 -1.99120685e-01
1.33542991e+00 -1.29967347e-01 -3.89371365e-01 6.36959001e-02
-8.15379620e-01 5.34731209e-01 1.87986940e-01 2.91659534e-01
4.18584526e-01 -1.19605458e+00 -8.45960915e-01 3.13151628e-01
-1.09172285e-01 2.23718092e-01 9.09173608e-01 1.23735619e+00
-8.71517599e-01 5.41026831e-01 -2.33622819e-01 -5.90815723e-01
-1.13725126e+00 2.73100227e-01 5.25995851e-01 -6.29038990e-01
-5.36792338e-01 1.18976092e+00 4.01840061e-01 -3.75881672e-01
3.41992266e-02 -4.68582898e-01 -2.13932991e-01 -1.49595082e-01
5.04712701e-01 2.89145410e-01 1.69077501e-01 -4.93227065e-01
-5.65070331e-01 6.84450090e-01 -3.47390234e-01 3.60671610e-01
1.26224792e+00 7.08976313e-02 -1.75009854e-02 -3.69502991e-01
1.27343988e+00 -4.83717978e-01 -1.29102385e+00 -2.91289061e-01
-3.83094937e-01 -2.21626744e-01 3.62335265e-01 -1.23529863e+00
-1.48446929e+00 9.44601238e-01 9.58814204e-01 -8.56223777e-02
1.72808087e+00 -9.84190926e-02 1.17270792e+00 -4.88540649e-01
-1.55164460e-02 -8.76114011e-01 -2.53878444e-01 1.89740062e-01
8.12240779e-01 -1.02864063e+00 -1.52408242e-01 -6.62689209e-01
-6.00006938e-01 8.13766360e-01 7.41110325e-01 -1.29762098e-01
7.70704329e-01 4.47655797e-01 8.76390338e-02 -1.48839384e-01
-3.80334020e-01 -3.25014651e-01 3.89076084e-01 3.37844670e-01
4.55962390e-01 2.09240973e-01 -7.13958085e-01 9.59354162e-01
1.46475270e-01 2.16039233e-02 9.55199301e-02 8.16939473e-01
-1.32405803e-01 -9.91598487e-01 6.50483668e-02 7.33904719e-01
-6.73279464e-01 -1.63824767e-01 -5.35617769e-02 7.99117029e-01
3.37289929e-01 7.30558276e-01 1.21187933e-01 -4.20737147e-01
3.93599868e-01 -3.84345412e-01 3.01872730e-01 -4.00411159e-01
-8.82953405e-01 4.21813041e-01 -2.06326589e-01 -7.08325326e-01
-2.48364508e-01 -3.12283605e-01 -1.45003545e+00 -1.63760066e-01
-3.50139737e-01 -2.74039917e-02 6.31407797e-01 5.53253531e-01
5.18311858e-01 1.05713344e+00 5.71728170e-01 -5.54815471e-01
-3.81161600e-01 -1.03564155e+00 -3.95583063e-01 4.00458872e-01
-8.60416889e-02 -3.33423465e-01 1.03999212e-01 -1.31808132e-01] | [14.699980735778809, -2.620511054992676] |
e0edb632-9675-4c26-b81b-49a77b0deae0 | textray-contour-based-geometric-modeling-for | 2008.04851 | null | https://arxiv.org/abs/2008.04851v2 | https://arxiv.org/pdf/2008.04851v2.pdf | TextRay: Contour-based Geometric Modeling for Arbitrary-shaped Scene Text Detection | Arbitrary-shaped text detection is a challenging task due to the complex geometric layouts of texts such as large aspect ratios, various scales, random rotations and curve shapes. Most state-of-the-art methods solve this problem from bottom-up perspectives, seeking to model a text instance of complex geometric layouts with simple local units (e.g., local boxes or pixels) and generate detections with heuristic post-processings. In this work, we propose an arbitrary-shaped text detection method, namely TextRay, which conducts top-down contour-based geometric modeling and geometric parameter learning within a single-shot anchor-free framework. The geometric modeling is carried out under polar system with a bidirectional mapping scheme between shape space and parameter space, encoding complex geometric layouts into unified representations. For effective learning of the representations, we design a central-weighted training strategy and a content loss which builds propagation paths between geometric encodings and visual content. TextRay outputs simple polygon detections at one pass with only one NMS post-processing. Experiments on several benchmark datasets demonstrate the effectiveness of the proposed approach. The code is available at https://github.com/LianaWang/TextRay. | ['Yifeng Chen', 'Xi Li', 'Fangfang Wang', 'Fei Wu'] | 2020-08-11 | null | null | null | null | ['scene-text-detection'] | ['computer-vision'] | [ 3.16915751e-01 -1.69020575e-02 1.24605224e-01 -2.32682511e-01
-7.28893816e-01 -7.04596519e-01 7.03373790e-01 4.62946206e-01
-7.17878565e-02 1.06471470e-02 -1.41935095e-01 -2.61849225e-01
2.09172964e-01 -9.60785627e-01 -7.77169645e-01 -5.71922839e-01
1.67305961e-01 5.95968723e-01 4.80379730e-01 -2.26644084e-01
6.40236735e-01 6.33830369e-01 -1.16428626e+00 2.09768608e-01
9.14494634e-01 7.46880233e-01 4.35179025e-01 8.60439003e-01
-6.19499922e-01 1.65956095e-01 -4.40430611e-01 -6.12528503e-01
1.90464616e-01 1.62930861e-01 -1.96300998e-01 4.13438827e-01
5.16804576e-01 -1.77243084e-01 -3.07407856e-01 1.03840697e+00
5.66498697e-01 1.97364554e-01 1.02062428e+00 -9.05677617e-01
-9.31225896e-01 5.39865971e-01 -1.10100484e+00 -2.99973190e-01
1.23728871e-01 9.88800526e-02 1.16340148e+00 -1.38986826e+00
5.96290767e-01 1.54571903e+00 8.15212190e-01 1.38328984e-01
-9.73437369e-01 -3.49049836e-01 6.71215534e-01 -6.84785619e-02
-1.54004395e+00 -1.31781818e-02 1.04187930e+00 -4.26535189e-01
5.71568787e-01 6.52118325e-02 3.83163631e-01 7.66254663e-01
9.60068963e-03 1.23354888e+00 2.90842444e-01 -4.99060243e-01
1.30230501e-01 5.08249663e-02 -3.96444136e-03 1.09088409e+00
3.91954005e-01 -4.63943839e-01 -1.49026275e-01 1.97339822e-02
9.36406970e-01 8.08124691e-02 -7.81053454e-02 -6.63011611e-01
-9.40877259e-01 7.87449658e-01 6.19568825e-01 1.79053575e-01
2.75291912e-02 1.57751173e-01 3.50955367e-01 -6.63923845e-02
6.45852506e-01 1.46947354e-01 -1.39174730e-01 3.93577129e-01
-1.04606164e+00 2.58151829e-01 4.41633880e-01 1.36691082e+00
5.66899180e-01 6.86836913e-02 -2.66804516e-01 1.02546895e+00
4.12998378e-01 7.36096561e-01 1.74140707e-01 -1.07189482e-02
9.09964323e-01 8.08765709e-01 -2.70156693e-02 -1.37018287e+00
-5.52798212e-01 -4.14592296e-01 -7.10234880e-01 1.51399270e-01
3.98834527e-01 -2.62900032e-02 -9.51872289e-01 1.16502690e+00
5.49513638e-01 -7.09699318e-02 -1.88359454e-01 7.31326282e-01
7.00707138e-01 1.05211222e+00 -4.04525697e-02 3.74919683e-01
1.52224171e+00 -1.12237751e+00 -5.60726106e-01 -2.97835499e-01
7.16127098e-01 -1.05866075e+00 1.09904432e+00 3.20021957e-01
-1.11725700e+00 -5.57049990e-01 -1.24101174e+00 -4.47842419e-01
-7.63488293e-01 8.71902108e-01 2.27699831e-01 5.69855928e-01
-6.67443633e-01 3.19347531e-01 -7.06003368e-01 -2.95407921e-01
5.48486829e-01 9.27931368e-02 1.97830945e-01 -2.36251578e-01
-6.80652380e-01 4.14198786e-01 3.88156176e-01 2.39709929e-01
-3.78589034e-01 -5.06370008e-01 -9.96596038e-01 7.97150806e-02
6.00809574e-01 -4.43974018e-01 9.14705336e-01 -5.96242130e-01
-1.16002953e+00 5.97477973e-01 -6.38731420e-02 8.20854586e-03
9.42855239e-01 -2.93535262e-01 -1.64810613e-01 2.05552265e-01
-7.82768503e-02 4.88630027e-01 1.36891699e+00 -1.57683480e+00
-4.93815690e-01 -4.16013867e-01 -2.56286085e-01 5.21504343e-01
-4.29746687e-01 -9.54769403e-02 -1.00092947e+00 -1.08156502e+00
5.39185405e-01 -6.37089193e-01 -2.04062492e-01 5.91496646e-01
-7.18618691e-01 -2.47201219e-01 1.05435991e+00 -6.77497923e-01
1.30040586e+00 -2.00351405e+00 5.17602228e-02 3.18649352e-01
3.33623663e-02 1.18250057e-01 -1.61182836e-01 4.78584111e-01
2.08854690e-01 2.24611297e-01 -2.85387099e-01 -7.27025211e-01
1.57759607e-01 -3.98745477e-01 -4.10397261e-01 6.37162268e-01
4.33125764e-01 9.73416507e-01 -8.23869944e-01 -7.87970424e-01
6.05151892e-01 5.47136784e-01 -2.43212819e-01 8.08686465e-02
-6.31160080e-01 -5.12529314e-02 -6.55942619e-01 9.50155139e-01
1.07320428e+00 -2.78272539e-01 3.11906431e-02 -3.74356776e-01
-3.70026261e-01 -3.68008077e-01 -1.40751565e+00 1.66676283e+00
-3.81468058e-01 5.91328859e-01 -4.50241975e-02 -8.28470767e-01
1.24483716e+00 -6.72557503e-02 1.81198586e-02 -5.84008694e-01
3.56242597e-01 1.19639128e-01 -5.59959233e-01 -3.62017721e-01
8.51684093e-01 4.74090070e-01 -1.34629562e-01 3.35894704e-01
-2.43107602e-01 -4.78311002e-01 1.27966091e-01 1.43803477e-01
6.35546863e-01 4.59241301e-01 2.04171464e-01 1.14310412e-02
6.49733245e-01 -7.89842382e-02 8.91572982e-02 7.53835559e-01
2.15640709e-01 1.01629663e+00 3.86514574e-01 -4.82402712e-01
-1.35326922e+00 -9.88701820e-01 -1.62323549e-01 1.27543128e+00
3.38267326e-01 -4.66535479e-01 -7.54499137e-01 -3.48898351e-01
-2.41402015e-02 5.18838882e-01 -5.37760913e-01 1.02493055e-01
-8.84815753e-01 -6.74795270e-01 4.12837744e-01 5.16384840e-01
4.58296835e-01 -8.76343787e-01 -4.93479133e-01 2.06092089e-01
2.04096511e-01 -1.01145351e+00 -7.82849133e-01 -3.84956524e-02
-9.65697050e-01 -7.73125410e-01 -1.14979887e+00 -1.13413072e+00
1.15260100e+00 5.33153355e-01 8.14816594e-01 2.15255991e-01
-8.98507535e-01 1.83508784e-01 -4.79820698e-01 -3.53329003e-01
-3.84934880e-02 1.31765634e-01 -6.33100510e-01 2.35174105e-01
-1.69296712e-01 -2.04956625e-03 -6.99004412e-01 2.64559627e-01
-8.42975259e-01 3.53598982e-01 8.05095375e-01 6.61183596e-01
7.86528885e-01 9.90911946e-02 2.57484645e-01 -8.62726390e-01
5.64850450e-01 -3.83779794e-01 -8.28181565e-01 4.73386943e-01
-2.23677069e-01 -8.13042223e-02 8.13105762e-01 -3.77372503e-01
-1.08636355e+00 2.33954683e-01 2.52167314e-01 -5.54096997e-01
1.97407473e-02 1.95109800e-01 -2.71837503e-01 6.56529218e-02
4.30639535e-01 5.07903874e-01 -4.43742275e-01 -4.16332394e-01
8.19252193e-01 6.33755684e-01 3.79471242e-01 -8.16295981e-01
1.17586899e+00 6.16255820e-01 -1.10124290e-01 -1.28816974e+00
-4.90521759e-01 -6.08420253e-01 -8.86280358e-01 -2.74528176e-01
7.63041258e-01 -7.57593036e-01 -3.30917299e-01 6.45264566e-01
-1.21894157e+00 -2.88739532e-01 1.46512598e-01 -1.17433921e-01
-4.22345132e-01 6.04213297e-01 -2.77328908e-01 -1.15426254e+00
-5.19180119e-01 -8.55039299e-01 1.60304725e+00 2.24930182e-01
1.61019906e-01 -9.80782092e-01 -2.76036382e-01 1.64869323e-01
1.04135454e-01 2.52930552e-01 1.11221421e+00 -3.92467737e-01
-8.78523767e-01 -3.93372297e-01 -7.29914188e-01 -9.01724175e-02
-9.42991450e-02 4.13836837e-01 -9.24313307e-01 -3.12343568e-01
-5.47511697e-01 -1.69880971e-01 8.58638108e-01 3.73992682e-01
1.47046745e+00 -1.79728419e-01 -4.35437888e-01 7.13088095e-01
1.66414750e+00 1.61525697e-01 3.28542501e-01 5.52980527e-02
1.07763922e+00 6.00546241e-01 7.54036307e-01 7.11000144e-01
2.65031338e-01 5.69806576e-01 4.10674423e-01 -2.96791553e-01
-1.42175511e-01 -4.88433391e-01 1.75833721e-02 6.03005767e-01
2.39385903e-01 -6.86910748e-01 -9.19835985e-01 4.05311286e-01
-1.98136270e+00 -6.40521049e-01 -3.27050567e-01 2.13399076e+00
4.60854292e-01 4.25036550e-01 -3.64456624e-02 3.71312872e-02
1.10990226e+00 3.51099640e-01 -6.51818275e-01 -2.11070850e-01
-1.98930800e-01 -1.22160405e-01 5.33933640e-01 3.98577243e-01
-1.36010003e+00 1.33556843e+00 4.68035555e+00 1.11336315e+00
-9.78909731e-01 -2.72284657e-01 6.86362863e-01 2.46799797e-01
-2.14869380e-01 -3.10399741e-01 -8.69749725e-01 2.83982575e-01
1.80430219e-01 -4.18469124e-02 1.62976325e-01 8.74140918e-01
3.05414468e-01 1.44695625e-01 -7.39158690e-01 1.02403247e+00
3.55434328e-01 -1.35889149e+00 2.91235030e-01 -3.60922724e-01
7.59398818e-01 -1.66520491e-01 2.04900146e-01 8.44276547e-02
1.31880447e-01 -9.01861608e-01 1.10955119e+00 5.48660755e-01
1.03395784e+00 -7.06899703e-01 2.68129289e-01 1.63819656e-01
-1.88290083e+00 -1.12213738e-01 -5.34097731e-01 5.63060582e-01
-2.87656356e-02 5.14690697e-01 -9.45670068e-01 5.21246433e-01
3.29287499e-01 7.20575988e-01 -6.69020474e-01 1.25420845e+00
-1.72584534e-01 2.45477587e-01 -3.90349239e-01 -4.96594578e-01
3.42594773e-01 -3.45307142e-01 4.71549362e-01 1.61372340e+00
3.96153837e-01 -1.73985332e-01 3.73478413e-01 1.15532112e+00
-1.18096516e-01 5.23171782e-01 -6.31077945e-01 -1.18043264e-02
6.39811933e-01 1.40340292e+00 -1.32243276e+00 -2.31955677e-01
-2.98874915e-01 9.10970449e-01 2.67642915e-01 3.67260575e-01
-8.60441566e-01 -7.08226740e-01 -1.21410288e-01 1.01200551e-01
5.23859143e-01 -4.16602343e-01 -7.84027457e-01 -1.01465321e+00
2.25269392e-01 -3.59469354e-01 1.46759391e-01 -8.50283623e-01
-1.06351626e+00 2.93087929e-01 -1.72998115e-01 -1.48946500e+00
4.38341290e-01 -8.12307656e-01 -1.08602834e+00 7.25275934e-01
-1.51455510e+00 -1.55235779e+00 -4.33919221e-01 3.63757372e-01
1.21583164e+00 -2.34742146e-02 4.45387036e-01 -2.94853412e-02
-7.31833994e-01 6.99318886e-01 4.42184269e-01 3.09612572e-01
5.36663413e-01 -1.42795765e+00 7.82852232e-01 7.17387140e-01
3.73143628e-02 4.23633486e-01 5.59715331e-01 -8.23875964e-01
-1.46764219e+00 -1.41351950e+00 4.54038501e-01 -7.18240365e-02
5.88181078e-01 -9.43807960e-01 -9.57054436e-01 2.70250946e-01
-2.11813897e-01 -3.27334329e-02 1.73202991e-01 -1.94879964e-01
-3.93046170e-01 2.43385360e-01 -1.01554680e+00 9.67284262e-01
9.87637341e-01 -2.36818343e-01 -3.15757394e-01 5.45190692e-01
5.68156540e-01 -4.64396715e-01 -4.31819588e-01 -1.40307490e-02
5.40541351e-01 -5.63083172e-01 1.09014690e+00 -1.15487263e-01
3.14852595e-01 -4.46451068e-01 1.26333535e-01 -9.94159102e-01
-1.71077341e-01 -6.14204466e-01 -6.28228933e-02 1.14564180e+00
3.62387061e-01 -2.31686220e-01 8.98999333e-01 -2.04391107e-02
-2.99231857e-01 -9.24789131e-01 -6.42603517e-01 -5.29874861e-01
1.85074463e-01 -4.19636965e-01 5.10460615e-01 6.49244249e-01
-2.26084426e-01 1.87740445e-01 -1.20039426e-01 3.09852362e-01
7.38922000e-01 1.99473158e-01 7.31572211e-01 -1.26095355e+00
4.31613363e-02 -6.92929089e-01 -2.61399150e-01 -1.27993429e+00
-1.91108480e-01 -7.40824521e-01 2.02434137e-01 -1.65265834e+00
-1.19252883e-01 -6.33201420e-01 2.64754683e-01 1.70961171e-01
-2.68085510e-01 -9.85154435e-02 2.18889102e-01 8.02444294e-02
-6.23912752e-01 6.68890595e-01 1.40052831e+00 -4.98663336e-01
-2.20981345e-01 -3.99275385e-02 -4.38301206e-01 8.66180599e-01
9.35446501e-01 -2.01289251e-01 -5.73733568e-01 -5.92251897e-01
3.27941149e-01 -8.70391130e-02 1.52882338e-01 -9.43235338e-01
4.27529365e-01 -2.93812752e-02 6.45999551e-01 -1.17767406e+00
4.21826184e-01 -6.70905530e-01 -5.88878751e-01 1.78090587e-01
-3.53531986e-01 1.51054449e-02 2.58748323e-01 7.97570348e-01
3.04208308e-01 -5.25254190e-01 8.98655117e-01 -5.98212481e-02
-5.91287494e-01 4.07414973e-01 -1.02701206e-02 8.27706456e-02
9.10854578e-01 -4.10913914e-01 -4.32828069e-01 -5.56499064e-02
-5.21286070e-01 3.78218532e-01 4.91746992e-01 5.13808250e-01
9.09072042e-01 -1.04615450e+00 -7.26784110e-01 2.74987131e-01
2.66166985e-01 4.36069816e-01 3.97385329e-01 3.20767492e-01
-7.90234506e-01 3.01241994e-01 2.57069290e-01 -6.12075329e-01
-1.24608648e+00 4.46151137e-01 1.57006845e-01 -1.74387336e-01
-1.05222094e+00 8.19286466e-01 4.55590665e-01 -4.46088850e-01
4.77453411e-01 -3.97275895e-01 -3.31548929e-01 1.26147583e-01
4.95668828e-01 3.27463478e-01 1.10646158e-01 -5.11941910e-01
3.18507031e-02 1.17075193e+00 -3.23820919e-01 3.92651483e-02
1.02513587e+00 -1.85518801e-01 3.06033760e-01 4.08483863e-01
1.08163726e+00 7.15020904e-03 -1.43664122e+00 -2.52386689e-01
-4.35027778e-02 -5.94327629e-01 3.19084004e-02 -4.86872733e-01
-9.39982593e-01 1.12589490e+00 5.40915549e-01 1.41224831e-01
6.54365838e-01 -1.51157588e-01 5.67506373e-01 5.74800670e-01
2.94074900e-02 -1.42065930e+00 4.24743652e-01 4.99253213e-01
1.08670592e+00 -1.12683272e+00 1.44736066e-01 -6.57612681e-01
-4.28270906e-01 1.63509536e+00 5.82736731e-01 -2.87498921e-01
4.86538619e-01 2.15499729e-01 -2.13802457e-01 -2.61618495e-01
-5.18100381e-01 -1.13681965e-01 3.00935835e-01 4.31764305e-01
3.47569764e-01 7.20981434e-02 1.58117455e-03 4.17231292e-01
4.17908058e-02 -5.01691997e-01 5.35264969e-01 9.46378529e-01
-7.28693008e-01 -6.91848874e-01 -6.38099372e-01 4.55331236e-01
-5.30674979e-02 -9.27281082e-02 -3.74029011e-01 7.18350053e-01
-4.64037023e-02 5.47241211e-01 2.74871826e-01 -2.85358820e-02
3.64474475e-01 -2.34109219e-02 2.83329040e-01 -5.99855900e-01
-1.91027701e-01 4.35854703e-01 -1.91466093e-01 -1.18343160e-01
2.89748460e-01 -8.75238836e-01 -1.64419138e+00 2.48288363e-02
-4.71986175e-01 -4.05780911e-01 9.25337911e-01 5.26291013e-01
9.17076990e-02 6.51894331e-01 7.97909260e-01 -1.20580256e+00
-5.48147440e-01 -7.50680387e-01 -4.18906301e-01 2.89348513e-01
1.04125477e-01 -6.15153074e-01 -2.83132672e-01 9.27474201e-02] | [12.070260047912598, 2.271317958831787] |
e806ff1c-29c7-4dc2-8438-09691df7b1a4 | c-pack-of-ipas-a-c90-program-benchmark-of | 2206.08768 | null | https://arxiv.org/abs/2206.08768v1 | https://arxiv.org/pdf/2206.08768v1.pdf | C-Pack of IPAs: A C90 Program Benchmark of Introductory Programming Assignments | Due to the vast number of students enrolled in Massive Open Online Courses (MOOCs), there has been an increasing number of automated program repair techniques focused on introductory programming assignments (IPAs). Such techniques take advantage of previous correct student implementations in order to provide automated, comprehensive, and personalized feedback to students. This paper presents C-Pack-IPAs, a publicly available benchmark of students' programs submitted for 25 different IPAs. C-Pack-IPAs contains semantically correct, semantically incorrect, and syntactically incorrect programs plus a test suite for each IPA. Hence, C-Pack-IPAs can be used to help evaluate the development of novel semantic, as well as syntactic, automated program repair frameworks, focused on providing feedback to novice programmers. | ['Vasco Manquinho', 'Mikoláš Janota', 'Pedro Orvalho'] | 2022-06-17 | null | null | null | null | ['program-repair', 'program-repair'] | ['computer-code', 'reasoning'] | [-2.29648337e-01 -1.41975850e-01 3.43647599e-01 -4.04037207e-01
-4.31806087e-01 -1.09226036e+00 -1.74550638e-01 9.92578208e-01
6.25079870e-02 1.50313243e-01 -3.20691705e-01 -7.88926780e-01
-1.03155375e-01 -1.11119974e+00 -7.83595324e-01 9.14902538e-02
4.17460352e-02 -3.58866714e-02 7.31976032e-01 -5.28884292e-01
6.37233317e-01 4.67837840e-01 -2.04699230e+00 4.62261349e-01
1.46226859e+00 3.21283221e-01 5.22415340e-01 7.70159662e-01
-6.09292746e-01 6.77510560e-01 -6.73994541e-01 -7.99315035e-01
-2.15949818e-01 -1.50748640e-01 -9.01379883e-01 -6.32863164e-01
7.35625327e-01 8.83714780e-02 2.31856778e-01 1.36756659e+00
1.44634098e-01 5.64819336e-01 -4.96357903e-02 -1.22674394e+00
-7.28307903e-01 7.77413309e-01 -4.08309661e-02 2.27370083e-01
9.87989366e-01 4.68306132e-02 9.06675637e-01 -7.54390419e-01
7.16348529e-01 8.81711006e-01 7.56621182e-01 4.67306793e-01
-1.09273267e+00 -4.31094736e-01 -2.80398279e-01 1.27993554e-01
-9.83739674e-01 4.34435368e-01 5.47572970e-01 -7.86807060e-01
9.54565942e-01 4.96432811e-01 8.36775124e-01 5.03415108e-01
2.06031248e-01 6.48040414e-01 8.35294068e-01 -8.42434764e-01
1.19094066e-01 6.82454705e-01 8.56175005e-01 1.15862131e+00
3.81452948e-01 -8.83015171e-02 -2.66950339e-01 -1.22010320e-01
1.25932738e-01 -3.27856839e-02 -3.61022145e-01 -2.90801883e-01
-6.61471784e-01 4.90757525e-01 1.31996930e-01 4.24192786e-01
2.27878943e-01 -9.82191712e-02 4.31331187e-01 5.16469061e-01
-3.23376417e-01 9.12488461e-01 -7.01896667e-01 -6.95506036e-01
-3.87753993e-01 6.88586235e-01 1.01035702e+00 1.29320991e+00
7.73050964e-01 -7.39525929e-02 -7.33571947e-02 6.65206313e-01
2.21943915e-01 5.87493032e-02 5.55715978e-01 -7.95775354e-01
5.54437101e-01 1.40679574e+00 -4.64732766e-01 -1.18131423e+00
4.15092967e-02 -3.02586764e-01 6.20920882e-02 1.83986992e-01
3.14256251e-01 1.60913646e-01 -3.18211317e-01 1.43963838e+00
1.63708314e-01 2.74404734e-02 -1.22433454e-01 3.17803562e-01
1.21425998e+00 5.64968824e-01 6.18165508e-02 4.29222494e-01
1.25240040e+00 -1.22518301e+00 -2.61800349e-01 3.84630412e-02
1.06450212e+00 -1.00606978e+00 1.66063952e+00 7.87870944e-01
-1.16170585e+00 -8.41468930e-01 -8.45482290e-01 -1.04950920e-01
-7.11178899e-01 -2.57386893e-01 3.97990197e-01 1.26810586e+00
-1.01731277e+00 9.92047369e-01 -3.88072670e-01 -8.56839642e-02
2.36642912e-01 1.89415783e-01 -2.94609666e-01 -3.91591966e-01
-6.75064027e-01 6.18927479e-01 3.50311190e-01 -6.01229072e-01
-4.01451766e-01 -1.48753512e+00 -9.80143547e-01 7.24560201e-01
-5.44158816e-02 -4.26978767e-01 1.48198414e+00 -5.93226850e-01
-1.52777112e+00 9.22319710e-01 2.42198691e-01 2.80782044e-01
1.98797628e-01 -7.92811587e-02 7.26715401e-02 -8.90845060e-02
-1.92830324e-01 3.66945803e-01 -1.63189337e-01 -8.21319222e-01
-6.38607085e-01 -5.99357374e-02 4.09131318e-01 -6.86203986e-02
-4.74411905e-01 3.19353700e-01 3.39733027e-02 -5.57980597e-01
2.33201206e-01 -7.30725467e-01 6.29166141e-02 -3.53888929e-01
1.38500065e-01 -4.55028772e-01 5.54022133e-01 -7.61447489e-01
1.34938920e+00 -2.15044999e+00 1.30092248e-01 1.93870217e-01
4.55899201e-02 3.31778318e-01 -1.52162895e-01 2.65749156e-01
-4.15173233e-01 4.21032101e-01 1.35582596e-01 -3.75945643e-02
3.26720744e-01 -4.10866626e-02 -2.82043546e-01 -3.73709917e-01
-2.98475325e-01 4.98508215e-01 -1.12975895e+00 -2.50952482e-01
2.74074107e-01 -2.61627734e-01 -1.18639326e+00 4.89214420e-01
-3.50614607e-01 3.37470502e-01 -2.27694079e-01 7.46614754e-01
6.40946209e-01 3.26154500e-01 -1.11550458e-01 7.28193104e-01
-3.30903351e-01 4.97021765e-01 -1.08257937e+00 1.67197275e+00
-6.34446204e-01 6.75549090e-01 -2.36654475e-01 -7.70221293e-01
1.12215734e+00 2.82888651e-01 5.44157960e-02 -4.67856377e-01
-1.04615286e-01 3.17863613e-01 -9.15359240e-03 -6.54502809e-01
8.60866904e-01 4.65335459e-01 -2.24789068e-01 3.56991738e-01
3.02300334e-01 -3.75399202e-01 6.75203979e-01 4.34697539e-01
1.35950661e+00 3.40498298e-01 -1.42733539e-02 -3.71643066e-01
8.40744317e-01 9.94089395e-02 4.96317476e-01 7.74155259e-01
-3.17288965e-01 4.07213718e-01 5.62290907e-01 -5.21456063e-01
-8.71430516e-01 -8.58095646e-01 1.30487740e-01 1.43238568e+00
-2.58584827e-01 -1.02212620e+00 -1.04638517e+00 -5.73049188e-01
-1.84229016e-01 7.88057685e-01 7.84956664e-02 -2.76575416e-01
-5.76772809e-01 -6.64756671e-02 4.59186316e-01 4.07374501e-01
3.44382405e-01 -1.04572546e+00 -6.30157292e-01 2.66953558e-01
-4.25685644e-02 -1.00686574e+00 -1.21065967e-01 4.00047079e-02
-9.41807747e-01 -1.13065767e+00 -5.65668344e-02 -1.37118423e+00
9.69452322e-01 3.65747035e-01 1.33881724e+00 7.31721461e-01
-4.71763402e-01 7.33251393e-01 -5.66629291e-01 -3.77007276e-01
-7.21543550e-01 -3.05868927e-02 -2.73514986e-01 -8.43209803e-01
3.75727206e-01 -7.33887017e-01 6.56670928e-02 9.48326755e-03
-7.51944542e-01 8.24163333e-02 6.47693127e-02 4.21552479e-01
2.55515218e-01 2.95931965e-01 1.97672322e-01 -1.25832224e+00
7.01288521e-01 -3.36968213e-01 -1.25686097e+00 7.14586914e-01
-4.21698719e-01 -8.12520236e-02 9.01701272e-01 -1.29850119e-01
-1.02574193e+00 -2.11092710e-01 -6.12730622e-01 -3.45778838e-02
-4.51986820e-01 8.91205430e-01 -1.82080284e-01 -7.81977594e-01
9.09073949e-01 1.31158113e-01 -6.97070599e-01 -4.79479194e-01
-2.21276313e-01 1.99718952e-01 4.90738213e-01 -1.33082187e+00
6.98769927e-01 -7.84834564e-01 -2.31720239e-01 -6.40993357e-01
-5.67946970e-01 -5.32174587e-01 -4.53993052e-01 -2.25452721e-01
2.68560857e-01 -6.26839042e-01 -9.88281071e-01 5.29121995e-01
-1.27074015e+00 -4.87891018e-01 -1.93028048e-01 1.88987598e-01
-3.05319950e-02 4.75316286e-01 -7.24610806e-01 -3.06954980e-01
-9.55545157e-02 -1.61323142e+00 2.26581842e-01 6.71182394e-01
-4.27634209e-01 -9.94787216e-01 3.34574401e-01 7.17207491e-01
2.76616037e-01 -2.78568324e-02 1.43292737e+00 -6.23793483e-01
-7.02375829e-01 -3.62605691e-01 2.78126419e-01 4.48033214e-01
-4.07928914e-01 4.93257582e-01 -8.22076499e-01 8.08764920e-02
-2.59575397e-01 -5.56586012e-02 1.40463576e-01 -2.82827318e-01
1.48647809e+00 -2.73308545e-01 1.97488070e-02 4.94850069e-01
1.55134976e+00 1.29262239e-01 7.48910785e-01 6.63565874e-01
5.11798084e-01 5.94012201e-01 4.59724039e-01 1.02286443e-01
4.01782900e-01 3.20857495e-01 3.00937653e-01 9.24881756e-01
2.11292073e-01 -3.49739373e-01 6.30298674e-01 1.18765748e+00
1.94193333e-01 9.27419364e-02 -1.26515436e+00 6.11151874e-01
-1.33389342e+00 -6.71453238e-01 -8.27286541e-01 2.09519339e+00
8.65337133e-01 8.91316161e-02 -8.59886110e-02 3.80448580e-01
5.98558128e-01 -4.65819061e-01 4.70583528e-01 -1.17253733e+00
5.13185143e-01 9.46070671e-01 -9.13484097e-02 5.03746390e-01
-5.56795120e-01 7.55325019e-01 5.89448595e+00 6.03425384e-01
-8.55838060e-01 -1.79900154e-02 -6.76916614e-02 7.33126044e-01
-6.14652991e-01 2.51782715e-01 -1.02088785e+00 4.47836995e-01
1.21186793e+00 -2.46538430e-01 4.41577673e-01 1.42964983e+00
-3.13157469e-01 -3.18781137e-01 -1.09383833e+00 4.78257060e-01
-3.49077344e-01 -1.52504849e+00 -1.92201689e-01 -3.35610092e-01
1.22066605e+00 -4.92130369e-01 -6.74213246e-02 8.52423906e-01
3.54112267e-01 -5.87685704e-01 5.89308977e-01 1.30714342e-01
-9.30738263e-03 -8.77266824e-01 8.19572806e-01 4.74410027e-01
-1.05528975e+00 -1.46027625e-01 -4.23459649e-01 -4.40801501e-01
-8.15900862e-01 4.09788430e-01 -8.68578017e-01 5.29057026e-01
9.59313750e-01 1.13598429e-01 -1.09349716e+00 1.60821831e+00
-6.34306669e-01 6.19334042e-01 8.20177980e-03 -2.90638566e-01
-2.71224707e-01 -1.38827071e-01 4.18354869e-01 1.16651261e+00
7.08667994e-01 3.65403205e-01 4.53560263e-01 1.26547539e+00
9.04779583e-02 2.89101422e-01 -3.15649241e-01 -1.58723339e-01
6.76320672e-01 1.36934340e+00 -6.50932729e-01 -1.85700029e-01
-4.33822542e-01 5.79365909e-01 5.00244319e-01 -4.83645573e-02
-6.53377712e-01 -9.76229608e-01 5.59634566e-01 8.49512219e-02
-8.20706487e-02 -3.77601713e-01 -5.18895030e-01 -9.03628767e-01
1.03761278e-01 -9.07483697e-01 3.22255582e-01 -8.00953329e-01
-7.30064034e-01 3.40828896e-01 -3.10639977e-01 -7.80806184e-01
3.24557304e-01 -7.82999337e-01 -1.27952862e+00 9.56761420e-01
-1.22944736e+00 -5.23756146e-01 -8.70152831e-01 1.68738320e-01
2.41656855e-01 -2.68126547e-01 1.04428673e+00 5.77417850e-01
-7.73744762e-01 8.75715554e-01 -1.67373389e-01 -9.70326364e-02
5.08376956e-01 -1.54528141e+00 1.79601684e-01 1.09065628e+00
-3.34538758e-01 9.19185281e-01 7.63912320e-01 -4.38481182e-01
-1.46114814e+00 -1.04312849e+00 1.13385415e+00 -3.97256732e-01
5.26150703e-01 2.76948903e-02 -1.24668252e+00 7.57990062e-01
-9.16286409e-02 7.96029642e-02 9.98833418e-01 1.19255416e-01
-4.08235192e-01 -7.85352364e-02 -1.13996792e+00 4.04928476e-01
6.69906437e-01 -3.56169701e-01 -9.59981740e-01 3.12848687e-01
7.88198948e-01 -9.78771567e-01 -1.06677127e+00 -1.51426371e-04
1.41617224e-01 -1.01333463e+00 7.91922390e-01 -4.84036326e-01
1.07232010e+00 -4.28825850e-03 5.13210241e-03 -1.24029219e+00
-1.44703453e-02 -4.09988046e-01 2.18777254e-01 1.53064525e+00
8.69760364e-02 -3.68292361e-01 8.74162138e-01 1.11215770e+00
-1.01039982e+00 -4.79403347e-01 -5.50928302e-02 -6.50970042e-01
1.85957938e-01 -4.00559634e-01 6.19840384e-01 1.27314317e+00
5.43126464e-01 -3.32823336e-01 5.57614386e-01 2.10174754e-01
2.79096484e-01 3.31102699e-01 9.99519646e-01 -1.48610628e+00
-6.06756210e-01 -7.95906723e-01 -4.94119227e-01 -6.48450494e-01
1.17946491e-01 -1.23105836e+00 6.34488165e-02 -9.65315163e-01
-5.62018231e-02 -8.34125400e-01 1.58903569e-01 5.78254104e-01
-3.30960810e-01 -2.04167906e-02 2.73723565e-02 -3.90291274e-01
-5.62292457e-01 -8.99836868e-02 9.73906159e-01 2.81504959e-01
-3.35273802e-01 1.86378825e-02 -6.18430674e-01 9.20364797e-01
6.34512544e-01 -5.59319615e-01 -5.84174953e-02 -2.88340360e-01
6.86636031e-01 5.48003167e-02 2.78985798e-01 -1.38573813e+00
4.13475603e-01 -4.64454442e-01 -6.03614897e-02 -3.52488577e-01
-6.00950480e-01 -6.02629125e-01 -4.50496636e-02 6.49965465e-01
-2.50576854e-01 2.22581044e-01 6.47642076e-01 -7.34418035e-02
-4.52590942e-01 -1.33119917e+00 6.95877016e-01 -4.83912468e-01
-9.60320294e-01 -2.72536546e-01 -5.21856844e-01 -6.52771294e-02
1.14602673e+00 -2.74866164e-01 -6.07698500e-01 3.70292276e-01
-6.17842138e-01 1.30367562e-01 2.98888713e-01 2.65328616e-01
6.93135917e-01 -1.00617647e+00 -1.25493467e-01 3.10555249e-01
1.44482583e-01 2.40908384e-01 5.02565444e-01 4.60857630e-01
-1.40016520e+00 5.59850693e-01 -6.92311227e-01 -5.24920762e-01
-1.71037626e+00 5.06404757e-01 -3.77660170e-02 -4.06807899e-01
-4.71708417e-01 1.18542159e+00 -9.38859656e-02 -1.33864033e+00
3.08244914e-01 -5.77375710e-01 -4.43632677e-02 -3.95680040e-01
7.13680446e-01 4.35403198e-01 5.27556181e-01 2.75505066e-01
1.07476607e-01 1.47359878e-01 1.36234611e-01 4.55668867e-01
1.36606097e+00 2.54397601e-01 -2.78779536e-01 1.36575043e-01
8.89628828e-01 5.24674833e-01 -2.41835386e-01 1.65076390e-01
4.65385348e-01 -7.39824295e-01 -3.06578904e-01 -9.43753898e-01
-8.33039343e-01 9.92177606e-01 3.39583039e-01 1.29645532e-02
7.34609962e-01 -6.36760294e-01 5.18941104e-01 6.59475803e-01
6.80246770e-01 -7.09782898e-01 2.24283621e-01 8.76858175e-01
5.48577666e-01 -9.83811140e-01 -2.15373665e-01 -8.60042214e-01
1.41236037e-01 1.59404051e+00 1.32704616e+00 -1.05924509e-01
1.60021871e-01 -4.74039353e-02 -5.55278361e-01 1.29480392e-01
-4.86584663e-01 5.46086133e-01 1.86395541e-01 3.53316605e-01
8.89719963e-01 -8.15378278e-02 -1.13723435e-01 1.26319706e+00
-5.08045673e-01 8.85379687e-02 1.18186641e+00 1.36616039e+00
-7.91572571e-01 -1.51346576e+00 -6.93449736e-01 2.94979848e-02
-2.58600295e-01 -1.86191142e-01 -2.07886562e-01 4.74704236e-01
2.59828061e-01 4.83205616e-01 -3.11279237e-01 -3.70146006e-01
5.06866574e-01 5.80773294e-01 9.39721823e-01 -1.09595537e+00
-1.40181446e+00 -9.46075618e-01 -1.49688154e-01 -2.05913648e-01
3.37697923e-01 -5.84846377e-01 -1.45511794e+00 -9.69904542e-01
-3.55373174e-01 3.55405867e-01 7.61611223e-01 7.42896140e-01
4.15654182e-01 9.33719099e-01 1.70884773e-01 -2.97866464e-01
-5.39129794e-01 -5.23997605e-01 -1.73035979e-01 2.77697951e-01
-2.84332037e-01 -3.40032607e-01 -1.24165557e-01 5.71106300e-02] | [9.449725151062012, 7.4172749519348145] |
3dc8f50e-e593-4ad1-8a65-45641bfa0aa3 | unsupervised-neural-machine-translation-with | 1901.04112 | null | http://arxiv.org/abs/1901.04112v1 | http://arxiv.org/pdf/1901.04112v1.pdf | Unsupervised Neural Machine Translation with SMT as Posterior Regularization | Without real bilingual corpus available, unsupervised Neural Machine
Translation (NMT) typically requires pseudo parallel data generated with the
back-translation method for the model training. However, due to weak
supervision, the pseudo data inevitably contain noises and errors that will be
accumulated and reinforced in the subsequent training process, leading to bad
translation performance. To address this issue, we introduce phrase based
Statistic Machine Translation (SMT) models which are robust to noisy data, as
posterior regularizations to guide the training of unsupervised NMT models in
the iterative back-translation process. Our method starts from SMT models built
with pre-trained language models and word-level translation tables inferred
from cross-lingual embeddings. Then SMT and NMT models are optimized jointly
and boost each other incrementally in a unified EM framework. In this way, (1)
the negative effect caused by errors in the iterative back-translation process
can be alleviated timely by SMT filtering noises from its phrase tables;
meanwhile, (2) NMT can compensate for the deficiency of fluency inherent in
SMT. Experiments conducted on en-fr and en-de translation tasks show that our
method outperforms the strong baseline and achieves new state-of-the-art
unsupervised machine translation performance. | ['Zhirui Zhang', 'Shuai Ma', 'Ming Zhou', 'Shuo Ren', 'Shujie Liu'] | 2019-01-14 | null | null | null | null | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [ 2.18776792e-01 -9.59554538e-02 -4.19401854e-01 -4.00812209e-01
-1.09824765e+00 -3.55474412e-01 5.81073523e-01 -2.33073562e-01
-3.84612173e-01 9.12512422e-01 4.53592569e-01 -6.08257651e-01
5.21410048e-01 -6.19033098e-01 -1.11794341e+00 -5.97307742e-01
7.38024175e-01 9.45142686e-01 -4.73307997e-01 -3.04845572e-01
-7.52758831e-02 -2.31508866e-01 -6.74058139e-01 2.64257520e-01
1.48600948e+00 3.29220384e-01 5.29689729e-01 1.01582646e-01
-5.97585320e-01 5.27085185e-01 -3.70128244e-01 -6.92400873e-01
2.78031170e-01 -7.31134176e-01 -4.60415065e-01 1.26465961e-01
6.90861642e-02 -2.11433932e-01 -2.77184486e-01 1.43179297e+00
5.83489299e-01 -1.01749569e-01 4.89374518e-01 -7.61673570e-01
-1.22127235e+00 1.07441795e+00 -5.71895480e-01 -4.73240949e-02
-8.34664777e-02 2.41017431e-01 9.76008415e-01 -1.56370032e+00
6.28685534e-01 1.34593797e+00 5.63424468e-01 7.65483141e-01
-1.55114174e+00 -6.44658923e-01 -2.52041817e-02 -4.45623398e-02
-1.04040611e+00 -5.24533391e-01 5.91598153e-01 -8.50550979e-02
1.11570621e+00 5.39940083e-03 4.45197582e-01 1.47877228e+00
3.81879330e-01 9.66649771e-01 1.18496418e+00 -6.55343890e-01
-4.09394056e-02 4.06747669e-01 -2.75068074e-01 4.40120518e-01
7.61358067e-02 1.28328398e-01 -7.25969255e-01 1.77496567e-01
6.68190122e-01 -2.17446819e-01 -1.07843871e-03 9.20519382e-02
-1.52464914e+00 7.11252570e-01 -1.27531597e-02 3.90308321e-01
-4.90026921e-01 -1.28547534e-01 5.21772504e-01 7.41344750e-01
9.30778742e-01 3.63513887e-01 -7.96586812e-01 -3.37685049e-01
-1.00397491e+00 -3.29392940e-01 3.77944142e-01 1.17857635e+00
9.05766904e-01 1.77589208e-01 -2.90942013e-01 1.27527010e+00
2.51331061e-01 9.66403782e-01 8.11269701e-01 -3.13451201e-01
1.06127942e+00 4.69657332e-01 -1.34519562e-01 -6.79814696e-01
1.82036370e-01 -8.22412252e-01 -9.08978403e-01 -6.03582263e-01
1.25000075e-01 -3.00781369e-01 -8.92814815e-01 1.87631679e+00
1.10360771e-01 2.17264332e-02 1.49849340e-01 8.17782581e-01
3.74641359e-01 8.78466308e-01 -9.34020206e-02 -5.74568510e-01
9.64736640e-01 -1.31283474e+00 -1.19688284e+00 -5.26029527e-01
9.71596777e-01 -1.23935032e+00 1.33320773e+00 1.32178918e-01
-1.10708201e+00 -7.17660427e-01 -9.20012891e-01 -1.29880831e-01
-8.96507502e-02 5.08735478e-01 3.89993846e-01 2.57558972e-01
-8.85874152e-01 5.50519288e-01 -1.01762271e+00 -2.64658689e-01
3.08473140e-01 3.33537012e-01 -2.11760581e-01 -2.37853870e-01
-1.35452199e+00 1.28040898e+00 1.31174311e-01 5.32992184e-01
-7.70977497e-01 -3.45901668e-01 -6.83257580e-01 -2.59547710e-01
1.79985389e-01 -7.77979136e-01 1.26926339e+00 -1.27285373e+00
-1.91434062e+00 7.11042702e-01 -6.44817233e-01 -2.63604015e-01
5.37218869e-01 -5.36083996e-01 -3.50129724e-01 -4.42144275e-01
4.91423905e-01 4.48061407e-01 7.10394919e-01 -9.81381118e-01
-3.95530164e-01 -2.85874575e-01 -7.41247177e-01 3.47407907e-01
-4.69133973e-01 2.92277098e-01 -3.79881173e-01 -8.44135344e-01
3.04006457e-01 -9.66824830e-01 -3.19324523e-01 -6.73613966e-01
-4.13476795e-01 -2.46661872e-01 4.70980704e-01 -9.12990093e-01
1.20620787e+00 -2.05438519e+00 5.79400241e-01 -1.31696641e-01
-3.17314744e-01 2.80997247e-01 -5.72551489e-01 4.67096508e-01
9.51853767e-02 7.64821544e-02 -2.53873646e-01 -6.80499852e-01
-8.07103142e-03 5.24190307e-01 -5.49026966e-01 2.24146247e-01
4.73900527e-01 1.23996675e+00 -1.06834733e+00 -5.17025173e-01
-5.39948530e-02 1.69054776e-01 -2.18642309e-01 2.22517267e-01
-3.32106143e-01 8.55920851e-01 -3.87895912e-01 5.90513527e-01
6.18443489e-01 2.98983473e-02 3.75277132e-01 8.65599811e-02
-5.38632236e-02 9.43379164e-01 -4.74287301e-01 2.00686216e+00
-6.73338950e-01 5.59857726e-01 -2.75610685e-01 -1.06721759e+00
1.07856905e+00 4.92370844e-01 1.67786226e-01 -7.92227805e-01
2.83049852e-01 8.05407345e-01 1.10724784e-01 -4.85883534e-01
4.02300566e-01 -3.86131704e-01 1.78468063e-01 6.39599800e-01
2.88713694e-01 -1.30003646e-01 8.74395445e-02 -5.74814528e-02
6.57736719e-01 5.82314670e-01 -1.83609158e-01 -2.14434892e-01
3.04933816e-01 2.83370376e-01 1.04232550e+00 2.35107556e-01
3.48196849e-02 4.01304066e-01 4.17055674e-02 -2.70624846e-01
-1.34386301e+00 -9.84264791e-01 -7.26749003e-02 9.06562150e-01
-1.74861029e-01 -5.07522166e-01 -7.42418408e-01 -7.25892425e-01
-4.20659065e-01 8.32804739e-01 -2.67269641e-01 -2.05203995e-01
-1.00988293e+00 -1.09482491e+00 4.12789553e-01 4.46400762e-01
4.41791326e-01 -8.99978697e-01 5.02218723e-01 6.21309817e-01
-7.95570552e-01 -1.14403510e+00 -9.23247516e-01 2.66061038e-01
-1.44218826e+00 -3.17032963e-01 -5.62117040e-01 -1.12752163e+00
9.63224947e-01 2.22522363e-01 1.21649182e+00 -1.37113392e-01
4.68187720e-01 -4.79229510e-01 -2.97868907e-01 -2.41273969e-01
-8.26899648e-01 1.44415826e-01 6.06770873e-01 -7.97862038e-02
9.44708467e-01 -7.81903267e-01 -1.03751056e-01 3.39797676e-01
-4.98396456e-01 4.14002746e-01 8.93186629e-01 1.35328221e+00
8.01187158e-01 -1.88428879e-01 5.22529304e-01 -7.69814968e-01
6.54145241e-01 -3.06194514e-01 -4.62052345e-01 3.23226005e-01
-8.39094877e-01 1.73870519e-01 8.25786471e-01 -8.49733889e-01
-1.04337358e+00 -3.31599534e-01 -7.85265956e-03 -4.22390640e-01
4.06709582e-01 6.22403920e-01 -3.75944167e-01 4.44845468e-01
5.94203472e-01 8.54735732e-01 -5.92792481e-02 -5.62451184e-01
3.85222584e-01 9.02306497e-01 2.58238226e-01 -7.34954357e-01
1.10326982e+00 -1.07218504e-01 -4.01519179e-01 -2.77688533e-01
-8.36382508e-01 1.15690313e-01 -8.13313961e-01 9.59882364e-02
6.17058754e-01 -1.14815402e+00 1.79065138e-01 2.98823535e-01
-1.36201096e+00 -3.13312590e-01 6.58906111e-03 9.40749168e-01
-4.23404336e-01 2.56976873e-01 -1.11509275e+00 -6.79441035e-01
-5.51424801e-01 -1.37168777e+00 9.46718395e-01 -8.34268406e-02
-2.44370356e-01 -9.53552186e-01 2.29361251e-01 4.48372155e-01
5.14582217e-01 -4.86837834e-01 1.04952931e+00 -4.48948711e-01
-6.56283498e-01 1.03767030e-01 -3.72607331e-03 8.66338909e-01
3.60977352e-01 -1.40229657e-01 -5.89259148e-01 -1.71374351e-01
3.64592344e-01 -2.07090601e-01 6.09878719e-01 2.17105955e-01
5.11280775e-01 -3.91543657e-01 -9.35671553e-02 6.12550318e-01
1.22184086e+00 -2.18690950e-02 5.16119540e-01 3.48368287e-01
7.26675332e-01 3.51092994e-01 7.16583431e-01 -2.41118044e-01
3.24672669e-01 5.88556230e-01 -2.03252569e-01 -1.18771702e-01
-1.72992468e-01 -8.02587330e-01 1.02179289e+00 1.95809507e+00
1.04246877e-01 1.75269872e-01 -5.97672939e-01 5.15038311e-01
-1.99573374e+00 -5.96562147e-01 -3.63635898e-01 2.17342615e+00
1.44524539e+00 1.38151973e-01 -4.05338496e-01 -2.01810896e-01
8.22471440e-01 -1.59170017e-01 -3.21029902e-01 -5.44157624e-01
-4.69954044e-01 2.34956965e-01 4.18075800e-01 4.14754480e-01
-5.35266817e-01 1.54144645e+00 5.94663477e+00 1.02639151e+00
-1.15601909e+00 5.75307310e-01 5.92379153e-01 -8.65323842e-03
-5.28389871e-01 2.85459876e-01 -6.51320219e-01 5.99245369e-01
1.10321951e+00 -4.39260639e-02 7.93608487e-01 5.75420618e-01
5.60304463e-01 2.96976030e-01 -1.12990856e+00 8.29465508e-01
-1.19902670e-01 -1.03347158e+00 1.55440867e-01 1.36774704e-01
1.08121216e+00 5.06628215e-01 6.91901371e-02 6.61740243e-01
4.41051602e-01 -7.82665431e-01 6.33381367e-01 1.21874951e-01
8.77337754e-01 -6.47892773e-01 7.99884379e-01 6.65397644e-01
-6.80886328e-01 3.00990045e-01 -8.09645176e-01 -7.09480196e-02
1.17295466e-01 9.68437850e-01 -7.68098533e-01 6.78172827e-01
1.49034217e-01 9.10770237e-01 -1.37303218e-01 3.44232589e-01
-7.90308237e-01 1.03266823e+00 -1.85076505e-01 1.03935469e-02
2.77552217e-01 -7.69430518e-01 6.02624059e-01 1.11351240e+00
5.38689256e-01 -4.58311707e-01 -4.51674871e-02 9.72778082e-01
-4.23900515e-01 4.62212592e-01 -4.30329829e-01 -4.24229980e-01
4.25087214e-01 8.86940300e-01 -3.34872872e-01 -5.33197641e-01
-5.19656360e-01 1.30476487e+00 5.67125440e-01 5.71119785e-01
-7.36187041e-01 6.80651069e-02 6.08574927e-01 -2.49683231e-01
-9.08292737e-03 -4.60977226e-01 -7.43017137e-01 -1.77173567e+00
4.33487415e-01 -1.16774464e+00 -3.75816524e-01 -5.00760019e-01
-1.47196507e+00 7.81325221e-01 -6.79666877e-01 -1.45933175e+00
-1.64126605e-01 -3.59371841e-01 -4.19209480e-01 1.27511811e+00
-1.42288423e+00 -1.08269477e+00 6.31973088e-01 2.05432713e-01
9.36470449e-01 -3.07381809e-01 9.85553980e-01 5.23284912e-01
-8.44906271e-01 7.38466442e-01 4.90347475e-01 3.11218262e-01
1.10413623e+00 -1.05243778e+00 7.49976218e-01 1.18792796e+00
4.09024239e-01 9.27271187e-01 6.04536533e-01 -9.00343657e-01
-1.51069760e+00 -1.31520748e+00 1.68337059e+00 -8.27960014e-01
7.38983572e-01 -6.05903566e-01 -7.91222095e-01 7.55995512e-01
2.85383403e-01 -3.92513633e-01 6.81058705e-01 3.35486382e-01
-3.76812309e-01 -3.40229161e-02 -6.47113621e-01 7.79240489e-01
1.00720668e+00 -8.33670795e-01 -8.95191908e-01 5.76574743e-01
9.29629862e-01 -2.97794014e-01 -6.55189693e-01 3.15555006e-01
3.34095657e-01 -2.65074223e-01 3.91473502e-01 -7.53809750e-01
8.31918001e-01 -3.20824265e-01 -2.22354218e-01 -1.82911205e+00
-2.71812737e-01 -8.31664026e-01 -9.70036816e-03 1.24715376e+00
8.83474350e-01 -5.89207530e-01 3.03964823e-01 1.13382630e-01
-3.29746008e-01 -7.47669935e-01 -9.94141400e-01 -9.48729634e-01
4.19054896e-01 -4.74272132e-01 6.10908329e-01 1.08321607e+00
1.99185073e-01 8.34603846e-01 -7.67047644e-01 -5.96963651e-02
5.08399069e-01 -1.89591721e-02 7.87764966e-01 -6.26686633e-01
-3.56140167e-01 -3.83784503e-01 2.67053485e-01 -1.54257870e+00
1.60978466e-01 -1.17655396e+00 3.44471484e-01 -1.49087417e+00
4.14076477e-01 -2.49606594e-01 -2.40109816e-01 3.43132883e-01
-5.79691768e-01 1.56391963e-01 -8.80865380e-02 6.10276580e-01
-1.69463322e-01 9.69414055e-01 1.67887592e+00 -9.57302451e-02
-2.06004366e-01 -9.29112881e-02 -5.57211697e-01 2.80684501e-01
7.20142484e-01 -8.56335938e-01 -1.95530295e-01 -1.14561260e+00
3.23751152e-01 3.97538505e-02 -3.47574383e-01 -3.99210572e-01
6.51427731e-02 -2.29468524e-01 1.98939085e-01 -4.57779467e-01
8.54489580e-03 -5.87234020e-01 -2.27561057e-01 1.97386444e-01
-2.94476867e-01 3.36196005e-01 4.59986068e-02 3.36000800e-01
-3.61189395e-01 1.14431098e-01 4.62461919e-01 -1.66818108e-02
1.71272710e-01 3.23284119e-01 -4.44226563e-01 -8.55877921e-02
2.11058527e-01 1.90623831e-02 -6.81572333e-02 -1.16710037e-01
-4.82029855e-01 3.77048612e-01 4.02462840e-01 6.75605595e-01
2.90032953e-01 -1.72895551e+00 -1.11746001e+00 3.47218841e-01
-9.70448088e-03 1.91526376e-02 -1.04890414e-01 1.25319338e+00
-6.65814104e-03 3.53797495e-01 3.74181680e-02 -5.72958112e-01
-6.97532833e-01 4.78124440e-01 1.40599743e-01 -4.49918330e-01
-5.02763987e-01 7.74356961e-01 -4.65497635e-02 -8.86829853e-01
3.26357447e-02 -3.01050037e-01 5.01752853e-01 -3.69813830e-01
3.56138557e-01 -1.01047106e-01 1.49909735e-01 -6.99638367e-01
-9.22780633e-02 3.97092402e-01 -2.68963575e-01 -3.84202212e-01
1.33712816e+00 -4.54986513e-01 -5.18765807e-01 6.18753314e-01
1.06670821e+00 1.99443907e-01 -7.83051312e-01 -7.35471010e-01
1.05348334e-01 -3.46246034e-01 -1.03596151e-01 -9.61382210e-01
-7.75308490e-01 1.04386628e+00 1.34048700e-01 -5.54288089e-01
1.02063406e+00 -4.19069469e-01 1.31194365e+00 4.01564091e-01
4.73215580e-01 -1.50939596e+00 -1.77765012e-01 7.94821024e-01
4.55629021e-01 -1.30365610e+00 -5.12647212e-01 -2.58907884e-01
-4.77928579e-01 1.05303121e+00 5.26835084e-01 2.36555971e-02
2.12450564e-01 1.66168556e-01 5.00657797e-01 3.19516480e-01
-9.71974850e-01 1.12635419e-01 1.13635175e-01 2.63878763e-01
6.43022656e-01 1.95375025e-01 -7.20987082e-01 6.02496624e-01
-2.93121070e-01 7.14768991e-02 1.07874148e-01 5.82715392e-01
-2.60925144e-01 -1.66454053e+00 -2.47907400e-01 2.08315834e-01
-3.99381846e-01 -6.64930761e-01 -3.61592919e-01 2.69031167e-01
1.27904549e-01 1.06913042e+00 -2.31874049e-01 -6.65513992e-01
7.41953626e-02 3.81310612e-01 5.12181401e-01 -7.88132429e-01
-4.84748363e-01 3.92133862e-01 1.26762062e-01 -2.89010674e-01
-1.93989471e-01 -4.97857511e-01 -1.02739811e+00 -8.76053423e-02
-5.76968014e-01 4.95166659e-01 8.47890556e-01 1.24006987e+00
3.25274527e-01 4.33898509e-01 8.44751120e-01 -3.14984143e-01
-9.88992691e-01 -1.61556864e+00 -1.08111367e-01 3.21956724e-01
5.23132943e-02 -2.46466026e-01 -3.65019053e-01 -4.28166389e-02] | [11.637060165405273, 10.217557907104492] |
3a776ca3-71f3-4ae7-826f-8f4774b639cb | identity-construction-in-a-misogynist-incels | 2306.15745 | null | https://arxiv.org/abs/2306.15745v2 | https://arxiv.org/pdf/2306.15745v2.pdf | Identity Construction in a Misogynist Incels Forum | Online communities of involuntary celibates (incels) are a prominent source of misogynist hate speech. In this paper, we use quantitative text and network analysis approaches to examine how identity groups are discussed on <incels.is>, the largest black-pilled incels forum. We find that this community produces a wide range of novel identity terms and, while terms for women are most common, mentions of other minoritized identities are increasing. An analysis of the associations made with identity groups suggests an essentialist ideology where physical appearance, as well as gender and racial hierarchies, determine human value. We discuss implications for research into automated misogynist hate speech detection. | ['Meredith Pruden', 'Kathleen M. Carley', 'David West Brown', 'Chloe Perry', 'Michael Miller Yoder'] | 2023-06-27 | null | null | null | null | ['hate-speech-detection'] | ['natural-language-processing'] | [-1.13707691e-01 2.42041916e-01 -4.49552089e-01 2.02105224e-01
3.08692664e-01 -8.80213022e-01 1.04638541e+00 5.99387884e-01
-1.72944412e-01 4.75334018e-01 1.00361598e+00 -5.03313184e-01
1.09394722e-01 -6.90004945e-01 -3.03203445e-02 -3.88558567e-01
3.42667639e-01 2.20972046e-01 -1.64084271e-01 -7.20249176e-01
9.05430257e-01 1.35075301e-01 -7.83634663e-01 8.48147199e-02
7.86658883e-01 9.19954572e-03 -6.61620319e-01 3.95516068e-01
-3.32453668e-01 1.44126844e+00 -9.52837110e-01 -9.53299046e-01
-1.66967481e-01 -6.48821414e-01 -6.30926728e-01 -1.02266006e-01
9.35312808e-01 -4.33521390e-01 -4.73540038e-01 1.29406953e+00
4.56222951e-01 -2.21121565e-01 5.42899847e-01 -9.74156439e-01
-1.14257634e+00 1.15257752e+00 -1.03195035e+00 5.70597887e-01
4.00840133e-01 5.82361892e-02 8.20555329e-01 -6.74884558e-01
1.34569824e+00 1.76025975e+00 1.05624080e+00 4.59667444e-01
-1.50672758e+00 -1.06667852e+00 -3.04800302e-01 -2.37454534e-01
-1.49624443e+00 -6.45684063e-01 8.43010008e-01 -1.25327384e+00
3.24801624e-01 4.09484595e-01 9.38077748e-01 1.00522339e+00
1.10126123e-01 -1.14733391e-01 1.10031712e+00 -4.42902654e-01
-2.04294011e-01 4.17281032e-01 1.48416787e-01 9.21961606e-01
6.55768931e-01 -4.61126596e-01 -7.13088989e-01 -8.87718260e-01
5.13116896e-01 -3.02400198e-02 2.00959355e-01 4.14746106e-01
-9.53116238e-01 1.41306973e+00 3.39476794e-01 8.03016424e-01
7.16227069e-02 2.39906490e-01 7.74997711e-01 1.69759467e-01
1.03918624e+00 6.72853649e-01 6.46149099e-01 -4.92242485e-01
-9.02001023e-01 1.35091469e-01 7.48663127e-01 6.85678273e-02
4.13653493e-01 -8.69795978e-02 1.85664579e-01 1.14747548e+00
1.30245522e-01 4.02468979e-01 -7.96018243e-02 -1.00537801e+00
1.55603215e-01 5.47775447e-01 -1.59419432e-01 -1.96802771e+00
-4.81961556e-02 -1.87992930e-01 -4.57842767e-01 3.24779898e-01
5.22935808e-01 -3.07702422e-01 -2.51845449e-01 1.45457554e+00
2.93781877e-01 -5.35320461e-01 -7.69580603e-01 5.70369124e-01
5.56115210e-01 4.42629486e-01 4.41992700e-01 -6.72881082e-02
1.28660858e+00 -4.12326396e-01 -9.31523085e-01 1.01621546e-01
4.29729491e-01 -1.13922298e+00 5.01537442e-01 -2.17298716e-02
-9.90627944e-01 1.33737460e-01 -7.33919919e-01 -2.07637116e-01
-4.41511691e-01 -5.19142389e-01 3.64312112e-01 1.21313787e+00
-9.51847970e-01 8.15568268e-01 -3.12778503e-01 -9.66425002e-01
8.07733595e-01 -2.70678461e-01 -1.13919921e-01 3.49596381e-01
-7.25068390e-01 1.16356170e+00 6.62351307e-03 -2.99274862e-01
-4.08106953e-01 -5.25570393e-01 -6.42340004e-01 -6.98879063e-01
2.87522733e-01 -9.37476754e-02 6.50462925e-01 -1.25530899e+00
-6.91617846e-01 1.50063229e+00 1.70346960e-01 -2.12358177e-01
4.67594266e-01 -1.44027546e-01 -7.44322062e-01 3.98586392e-01
3.70924890e-01 3.13096672e-01 1.02117860e+00 -1.32034624e+00
-2.70027101e-01 -3.29780310e-01 2.58172542e-01 -2.64456332e-01
-8.13746154e-01 1.11552691e+00 9.44242477e-01 -8.42778265e-01
7.47354478e-02 -7.50947297e-01 3.73145431e-01 1.71011072e-02
-8.99153411e-01 -1.34162471e-01 1.01070786e+00 -1.31995404e+00
1.88406706e+00 -2.41738200e+00 -2.60006070e-01 3.84352207e-01
1.17417252e+00 -2.10771546e-01 6.91967785e-01 1.12351871e+00
1.55613109e-01 1.00409234e+00 2.85509437e-01 1.52263582e-01
4.93625551e-02 -1.54012561e-01 -1.98150709e-01 9.75927949e-01
-3.03368181e-01 4.36279982e-01 -1.10187304e+00 -6.78390265e-01
-1.45874560e-01 5.18994510e-01 -4.16985214e-01 -6.20030046e-01
3.16251874e-01 4.32920098e-01 -6.82864711e-02 7.11281240e-01
5.47379553e-01 -1.99252173e-01 3.23631972e-01 2.33150274e-01
-8.83606374e-01 3.89571965e-01 -1.52508438e-01 7.85947502e-01
2.36331597e-01 1.31727481e+00 4.55212206e-01 3.95532362e-02
8.31088841e-01 -1.63790658e-02 3.33639055e-01 -5.69287911e-02
5.90924382e-01 1.76068515e-01 5.51099479e-01 -2.19307184e-01
8.20795476e-01 -4.21448410e-01 1.02629922e-02 8.23443651e-01
-3.84271830e-01 4.67835143e-02 -1.02945574e-01 6.93504930e-01
1.05851161e+00 -1.50496036e-01 5.88432074e-01 -8.06362629e-01
2.49759154e-03 2.90994287e-01 4.56934273e-01 4.35955912e-01
-6.85407937e-01 1.87120110e-01 8.39347184e-01 -5.29525757e-01
-1.54699397e+00 -1.06134200e+00 -1.07680604e-01 1.37543023e+00
1.87015623e-01 -7.05609739e-01 -8.22433233e-01 -2.43949413e-01
2.55538791e-01 7.44925022e-01 -5.43040335e-01 1.35560602e-01
-3.09447825e-01 -3.89936417e-01 8.29640448e-01 -3.52207154e-01
3.64374608e-01 -8.30649436e-01 -6.84891403e-01 -9.32583362e-02
-6.63629696e-02 -6.17114902e-01 -5.06046176e-01 -6.22986495e-01
-2.92533398e-01 -1.17031634e+00 -3.80409390e-01 -5.42295694e-01
6.71861649e-01 2.04499573e-01 7.92577207e-01 7.34504223e-01
-5.71091235e-01 7.42112994e-02 -3.94045800e-01 -3.34485710e-01
-9.49134886e-01 -1.54169813e-01 -9.30864066e-02 1.59829352e-02
4.01637971e-01 -8.92369449e-01 -3.47987562e-01 -1.63685009e-01
-5.26054621e-01 2.55101651e-01 -3.91736090e-01 4.52464461e-01
-6.64220929e-01 -1.13914572e-01 4.02748674e-01 -1.51791334e+00
6.98702693e-01 -9.12189960e-01 1.85406968e-01 -1.56194896e-01
-3.79865885e-01 -8.17083120e-01 2.96051770e-01 -3.51075232e-01
-1.00134683e+00 -9.63813007e-01 2.42593542e-01 -3.15803010e-03
6.63716123e-02 9.97959152e-02 6.13513947e-01 -2.75304765e-01
1.10426283e+00 -4.79290992e-01 3.84632200e-01 -3.76858890e-01
1.93308711e-01 7.18097627e-01 1.36310205e-01 -5.65808237e-01
1.20165658e+00 7.60670424e-01 -4.55271542e-01 -1.31242907e+00
-4.72493410e-01 -3.72000307e-01 -4.00333703e-01 -7.60035038e-01
1.23036492e+00 -8.76850605e-01 -1.02863896e+00 4.36099976e-01
-1.19682360e+00 -2.44156748e-01 2.33005136e-02 2.29774669e-04
2.26151913e-01 4.11932468e-01 -1.27987719e+00 -1.27672708e+00
-2.71061718e-01 -7.60469660e-02 2.25503460e-01 2.13974312e-01
-1.19135320e+00 -1.18884611e+00 2.96715826e-01 5.60765982e-01
5.72379708e-01 1.14612591e+00 9.94476080e-01 -5.54829061e-01
2.22985655e-01 4.83678371e-01 -1.77747056e-01 -1.57072350e-01
3.87503028e-01 5.29932976e-01 -5.23730993e-01 -8.01849067e-02
-8.52464810e-02 -3.76641929e-01 2.94347405e-01 -2.04518035e-01
3.20031524e-01 -1.00023973e+00 -2.71333516e-01 -8.86138901e-02
1.50998330e+00 2.64267653e-01 3.38525146e-01 4.63117391e-01
9.35312629e-01 7.81989396e-01 4.68446910e-02 1.05745697e+00
-7.08733276e-02 1.03455678e-01 5.65426610e-02 1.83697462e-01
-1.84655502e-01 -5.12204289e-01 4.49829221e-01 9.51630473e-01
-4.03042942e-01 2.40564093e-01 -1.26617742e+00 6.85843825e-01
-1.42566121e+00 -1.71620846e+00 -4.66544002e-01 1.55168879e+00
8.14213276e-01 6.52246103e-02 5.09888411e-01 -2.42234439e-01
1.39742875e+00 7.60333359e-01 -1.15417004e-01 -8.27354193e-01
-4.69091572e-02 3.97724301e-01 5.73394060e-01 7.39422619e-01
-8.92727554e-01 1.01976693e+00 6.99422407e+00 7.80187607e-01
-7.46670127e-01 6.26342356e-01 7.04823017e-01 -3.00766855e-01
-5.19915283e-01 2.61781439e-02 -3.75731081e-01 5.31982005e-01
4.82507437e-01 -5.19001007e-01 6.50604010e-01 4.94575143e-01
2.08753556e-01 -2.42149189e-01 -5.13086736e-01 8.37027311e-01
4.73873228e-01 -1.43480539e+00 -5.46273947e-01 3.49484205e-01
9.62753534e-01 -2.96914369e-01 2.18434930e-01 -1.83905721e-01
7.09971249e-01 -1.01065612e+00 1.02518225e+00 -4.73935753e-02
9.49054241e-01 -7.46602893e-01 -5.15480675e-02 5.17861210e-02
-4.97219652e-01 -2.33732626e-01 -2.18696073e-01 -6.95968091e-01
2.22867563e-01 5.19124150e-01 -4.84143227e-01 -1.81776613e-01
3.41048241e-01 7.58732855e-01 -7.17700183e-01 3.26394469e-01
-1.26272812e-01 8.69224250e-01 1.83676388e-02 -4.19958681e-02
2.57053733e-01 -5.56244552e-01 9.42898154e-01 1.30481172e+00
-2.01050658e-02 1.04639299e-01 -1.70811266e-01 1.15269506e+00
2.08614208e-02 6.25944883e-02 -1.01790261e+00 -8.78576636e-01
7.95697153e-01 1.36219859e+00 -1.02266610e+00 1.84339983e-03
-1.60137400e-01 7.12665081e-01 2.93317795e-01 3.10095120e-02
-7.97530651e-01 -2.47249216e-01 8.64241421e-01 9.01257515e-01
-5.20193458e-01 -4.99520719e-01 -3.54462475e-01 -7.73960173e-01
-7.74906278e-01 -1.07438362e+00 1.95245013e-01 -3.44963402e-01
-1.57496524e+00 7.17608863e-03 -2.11745262e-01 -3.07700545e-01
1.78757772e-01 -5.78539027e-03 -7.81450272e-01 4.68098760e-01
-3.47659081e-01 -1.25922430e+00 -9.32634622e-02 1.20696798e-01
3.09591629e-02 -1.16514534e-01 5.79142928e-01 2.70463943e-01
-6.28743649e-01 3.91196460e-01 1.21385679e-01 6.48757696e-01
6.83615148e-01 -7.55340338e-01 3.80180478e-01 4.73678648e-01
-1.33056194e-01 8.50379467e-01 9.86065447e-01 -1.30171955e+00
-9.06773329e-01 -4.47195441e-01 9.54778969e-01 -5.11177957e-01
1.51079714e+00 -4.03914839e-01 -3.52764219e-01 7.76571572e-01
8.59292924e-01 -9.17467415e-01 1.09238684e+00 5.40075362e-01
-7.35456347e-01 7.29604244e-01 -1.41594470e+00 1.05813217e+00
1.69964373e+00 -9.84032989e-01 -5.37942529e-01 5.33433914e-01
5.37886679e-01 2.45333970e-01 -8.23289335e-01 -6.37474954e-01
7.12619424e-01 -9.05797064e-01 4.89371002e-01 -4.16617304e-01
1.12365103e+00 2.47764647e-01 2.63737291e-01 -1.03029346e+00
-7.81718314e-01 -1.13046563e+00 3.29597861e-01 1.76135445e+00
5.70579059e-02 -8.47998798e-01 6.92852437e-01 7.26665378e-01
2.07466677e-01 4.11291905e-02 -8.28295588e-01 -4.88844156e-01
2.63170689e-01 5.30575037e-01 -7.22778291e-02 2.14178729e+00
5.64841270e-01 4.98953313e-01 -5.03061473e-01 -5.85369468e-01
8.26495469e-01 -5.15625298e-01 5.71140468e-01 -1.34312534e+00
5.06808050e-02 -8.11026096e-01 -6.53273702e-01 2.51622200e-01
3.43790501e-02 -8.76452148e-01 -5.75931072e-01 -1.30700636e+00
6.54142499e-01 -1.90802336e-01 5.39923966e-01 2.16141135e-01
4.49313611e-01 5.68995833e-01 7.59112418e-01 3.33188981e-01
-2.77384102e-01 -6.94745332e-02 1.09840345e+00 -4.84559238e-02
-1.82355270e-01 -8.77403855e-01 -1.03293526e+00 9.69064236e-01
8.80125046e-01 -5.53261042e-01 -1.98068134e-02 4.37712297e-02
7.48883128e-01 -5.29284894e-01 4.06122357e-01 -8.06292355e-01
-1.58163905e-01 -5.07956147e-01 1.06827676e-01 -9.22779068e-02
2.51157343e-01 -4.08266246e-01 6.71965420e-01 9.55398679e-01
-2.10023329e-01 -1.09725848e-01 -4.47896905e-02 1.46121711e-01
1.27310574e-01 -2.99604744e-01 9.70425129e-01 -5.83719313e-02
-3.01377978e-02 -2.67614841e-01 -8.07750881e-01 4.26608652e-01
1.01132441e+00 -4.02424634e-01 -1.22748291e+00 -6.65823340e-01
-6.24833286e-01 -3.18167269e-01 1.19314432e+00 6.78268075e-02
1.17717952e-01 -1.46857572e+00 -1.02536809e+00 -5.87318778e-01
-4.30833995e-02 -1.10007036e+00 -2.53809523e-02 8.55188906e-01
-1.07632136e+00 -4.63063627e-01 -5.81028163e-01 2.35513464e-01
-1.34060562e+00 3.24359864e-01 -1.32867485e-01 3.37017626e-01
-7.63700962e-01 6.35629356e-01 2.38216639e-01 -1.11149922e-02
-4.82210994e-01 8.23530853e-01 -8.00776929e-02 5.52243352e-01
6.00201130e-01 1.00962651e+00 -1.12461209e+00 -1.30815315e+00
-3.55303705e-01 1.15375601e-01 -6.93308115e-02 -2.71496147e-01
1.17842555e+00 -2.09209606e-01 -1.06560731e+00 8.20842743e-01
1.15755594e+00 7.23361254e-01 -5.15509903e-01 3.37414235e-01
2.70253450e-01 -1.24866772e+00 -5.92185795e-01 -6.66566074e-01
-4.21130627e-01 4.38480765e-01 -1.80278309e-02 8.27525914e-01
1.21788286e-01 1.25816628e-01 6.90855384e-01 -3.01352531e-01
2.97569692e-01 -1.34907329e+00 2.46879265e-01 4.29258734e-01
9.33625102e-01 -6.79075718e-01 3.42827380e-01 -8.06803524e-01
-4.29362386e-01 9.44338560e-01 6.29189670e-01 -3.38576227e-01
6.61773562e-01 1.30665135e-02 -1.03194110e-01 -6.84957445e-01
-1.95874453e-01 2.84937590e-01 -4.39000726e-01 4.54089999e-01
7.91262507e-01 3.55681151e-01 -1.21930552e+00 1.18808776e-01
-6.00882351e-01 -6.72425270e-01 8.82007480e-01 6.43485725e-01
-7.69972801e-01 -5.99552333e-01 -9.36360776e-01 3.46099496e-01
-1.13634920e+00 -6.95004612e-02 -1.49202478e+00 9.81036603e-01
5.76280832e-01 1.08303416e+00 3.67592186e-01 -5.70612609e-01
-4.93020207e-01 4.89768274e-02 3.92432362e-01 -6.48703933e-01
-1.48631275e+00 -2.99281217e-02 9.18123543e-01 -2.13222858e-02
-2.47004300e-01 -6.56703115e-01 -9.22958076e-01 -1.76676714e+00
-2.87001699e-01 -1.08935498e-01 5.89849651e-01 4.98727828e-01
1.05763465e-01 -8.78752321e-02 3.05073112e-01 -2.94775873e-01
7.74252415e-02 -9.26578522e-01 -9.95481491e-01 6.87812150e-01
-3.38193960e-02 -4.03422087e-01 -4.47508097e-01 2.87318267e-02] | [8.778169631958008, 10.441375732421875] |
58e3b342-0351-492f-93b6-c5f9cd47fb5a | an-evolutionary-computing-enriched-rs-attack | 1709.08362 | null | http://arxiv.org/abs/1709.08362v2 | http://arxiv.org/pdf/1709.08362v2.pdf | An Evolutionary Computing Enriched RS Attack Resilient Medical Image Steganography Model for Telemedicine Applications | The recent advancement in computing technologies and resulting vision based
applications have gives rise to a novel practice called telemedicine that
requires patient diagnosis images or allied information to recommend or even
perform diagnosis practices being located remotely. However, to ensure accurate
and optimal telemedicine there is the requirement of seamless or flawless
biomedical information about patient. On the contrary, medical data transmitted
over insecure channel often remains prone to get manipulated or corrupted by
attackers. The existing cryptosystems alone are not sufficient to deal with
these issues and hence in this paper a highly robust reversible image
steganography model has been developed for secret information hiding. Unlike
traditional wavelet transform techniques, we incorporated Discrete Ripplet
Transformation (DRT) technique for message embedding in the medical cover
images. In addition, to assure seamless communication over insecure channel, a
dual cryptosystem model containing proposed steganography scheme and RSA
cryptosystem has been developed. One of the key novelties of the proposed
research work is the use of adaptive genetic algorithm (AGA) for optimal pixel
adjustment process (OPAP) that enriches data hiding capacity as well as
imperceptibility features. The performance assessment reveals that the proposed
steganography model outperforms other wavelet transformation based approaches
in terms of high PSNR, embedding capacity, imperceptibility etc. | ['Elsaid MD. Abdelrahim', 'Romany F. Mansour'] | 2017-09-25 | null | null | null | null | ['image-steganography'] | ['computer-vision'] | [ 1.03109729e+00 1.09171286e-01 3.48110467e-01 -2.91677788e-02
-7.51118883e-02 -6.06346190e-01 4.42087471e-01 2.48970419e-01
-2.60083288e-01 6.99557900e-01 1.14366829e-01 -5.16191185e-01
-3.63742828e-01 -9.77490187e-01 -3.22483748e-01 -8.25922310e-01
-2.27845553e-02 -2.24359576e-02 1.76170737e-01 -4.46899891e-01
4.67430234e-01 2.90245444e-01 -1.02006245e+00 1.74444899e-01
7.31541753e-01 5.57166100e-01 1.24171093e-01 9.88312840e-01
1.44421443e-01 4.83667612e-01 -3.30988765e-01 -8.35401788e-02
5.29724002e-01 -6.65905714e-01 -7.52304912e-01 2.07166523e-01
-3.33254129e-01 -3.58049840e-01 -2.90253848e-01 1.15470791e+00
5.28416336e-01 -3.16711217e-01 5.38441658e-01 -1.00551248e+00
-6.74913824e-01 3.72303665e-01 -7.51334786e-01 3.14651787e-01
3.31298530e-01 1.00337379e-01 1.62174612e-01 -2.70595014e-01
6.91720486e-01 1.02659845e+00 5.79325855e-01 2.59046406e-01
-8.18501592e-01 -5.24719119e-01 -6.43605053e-01 2.81515062e-01
-1.14169335e+00 -1.17044300e-01 8.21725011e-01 9.90292653e-02
7.38000929e-01 6.52441323e-01 7.06336796e-01 2.84751505e-01
9.64536071e-01 -1.63437165e-02 1.73706770e+00 -7.05819070e-01
-2.72673726e-01 3.97676498e-01 -1.85118005e-01 8.51668835e-01
5.99547148e-01 1.00107215e-01 1.32137433e-01 -3.38745415e-02
5.96230745e-01 3.75401229e-01 -4.77951825e-01 2.62064278e-01
-1.40318787e+00 6.28457308e-01 2.42854565e-01 8.40467095e-01
-4.24239427e-01 8.41730386e-02 4.17556673e-01 6.94782853e-01
-2.63681352e-01 -1.26635611e-01 -1.20252627e-03 1.80709958e-01
-5.45143545e-01 -2.28861928e-01 5.99165261e-01 4.83679563e-01
2.16621548e-01 2.14093834e-01 4.73267794e-01 -8.20765570e-02
6.28703117e-01 5.71282446e-01 6.24691069e-01 -5.90608418e-01
3.00874054e-01 5.84829450e-01 -2.15060830e-01 -1.56570935e+00
-1.50967926e-01 -3.86699885e-01 -8.20064187e-01 4.27091122e-01
8.44312087e-03 1.22488532e-02 -1.08813381e+00 1.19361877e+00
4.68659818e-01 3.51387829e-01 7.24672794e-01 5.33837378e-01
4.94768232e-01 9.33217049e-01 -5.79772145e-03 -1.78046450e-01
1.56360877e+00 -4.31213796e-01 -1.13175893e+00 3.12948227e-01
3.52592736e-01 -1.45510924e+00 2.23755330e-01 4.17914242e-01
-8.90349627e-01 -2.36910388e-01 -1.43119454e+00 4.16309237e-01
-4.00504917e-01 -6.23252392e-01 5.41141331e-01 1.17691767e+00
-9.84673023e-01 3.07552367e-01 -7.60439813e-01 -4.11377162e-01
3.16153169e-01 8.30480099e-01 -6.01868272e-01 -2.60809332e-01
-1.13051808e+00 8.36415648e-01 5.29529035e-01 1.72193259e-01
-4.03567344e-01 -9.08860043e-02 -5.84713399e-01 -1.99718609e-01
-1.52129486e-01 -5.65110028e-01 3.71150285e-01 -1.21511996e+00
-1.17154574e+00 6.95517957e-01 3.90319765e-01 -7.28129983e-01
5.21199942e-01 6.21108115e-01 -6.10814929e-01 5.54373085e-01
-6.10542715e-01 3.64710838e-01 8.83820951e-01 -9.72673118e-01
-5.47466040e-01 -4.76004303e-01 -2.58502156e-01 2.42233664e-01
-2.21495301e-01 -2.34509353e-02 4.15073186e-01 -7.31951535e-01
6.30225122e-01 -9.60053205e-01 4.35236208e-02 -8.19229260e-02
-1.99471354e-01 6.12020135e-01 1.54775023e+00 -1.03920841e+00
9.21702743e-01 -1.96851373e+00 -1.86273277e-01 6.55625880e-01
-3.24859992e-02 5.55694938e-01 3.90886664e-01 7.22898126e-01
-5.89800850e-02 1.80253968e-01 -4.09803301e-01 5.26022613e-01
-4.06515092e-01 2.09962830e-01 1.28976271e-01 9.10845220e-01
-3.60682905e-01 5.48205137e-01 -4.11526352e-01 -6.96702600e-01
4.52973455e-01 1.20391655e+00 -2.69198209e-01 -4.83988643e-01
4.42690879e-01 7.63611794e-01 -5.90063751e-01 7.83569098e-01
1.02965569e+00 8.68678913e-02 2.21332744e-01 -2.32862413e-01
4.61848639e-02 -4.28586930e-01 -1.27690327e+00 1.05251193e+00
-5.69213368e-02 4.76578414e-01 -2.98730284e-02 -1.08578193e+00
9.12737370e-01 1.09933615e+00 6.44717395e-01 -3.55965525e-01
5.34585118e-01 4.91926342e-01 1.80777907e-01 -9.79535758e-01
1.01096667e-01 -1.86430112e-01 5.59156001e-01 5.05903780e-01
-4.60335881e-01 4.43826646e-01 -4.93237615e-01 -1.03380062e-01
1.07468438e+00 1.56775638e-01 2.93324649e-01 -8.88790712e-02
1.01268995e+00 7.27475360e-02 3.35092723e-01 1.35676533e-01
-4.16868657e-01 1.55848563e-01 -2.75011927e-01 -3.74820381e-01
-1.22539377e+00 -6.30923688e-01 -1.80056423e-01 -1.88133672e-01
2.57354856e-01 4.27573055e-01 -7.90366292e-01 -3.58583778e-01
-3.32030296e-01 2.69886907e-02 -1.61390647e-01 -1.58691257e-01
-6.77242994e-01 -6.64137304e-01 6.99065089e-01 -2.30053201e-01
1.23056936e+00 -1.07043362e+00 -9.82617855e-01 4.74201500e-01
3.87384072e-02 -9.51869190e-01 -1.79727733e-01 -4.76480633e-01
-1.24197233e+00 -1.16101420e+00 -6.41186237e-01 -1.12238443e+00
1.08954370e+00 4.40047681e-01 4.95302640e-02 6.54314220e-01
-7.44397104e-01 2.17270687e-01 -6.70833051e-01 -4.67088670e-01
-8.86405587e-01 -5.03090262e-01 -3.23939800e-01 1.09000742e-01
2.92954743e-01 -6.40550971e-01 -1.13860905e+00 -1.33457601e-01
-1.20771420e+00 -4.46030907e-02 7.47466326e-01 7.24824786e-01
2.61351019e-01 9.41604197e-01 5.26188135e-01 -9.34367180e-01
3.17037940e-01 -3.19595635e-01 -4.16550040e-01 2.82940775e-01
-9.43905652e-01 -4.57092049e-03 2.97488570e-01 -1.13913894e-01
-9.30364966e-01 -5.75771108e-02 -1.48175254e-01 3.77120167e-01
-1.23863645e-01 3.41519266e-01 1.85793489e-01 -7.89931595e-01
1.45707011e-01 6.60071313e-01 3.94903749e-01 -1.37817234e-01
-2.51839638e-01 9.19457316e-01 4.50564653e-01 2.83082098e-01
1.06394577e+00 8.14130783e-01 4.40363079e-01 -7.26635456e-01
4.34723735e-01 -1.91431656e-01 -6.91821277e-02 -2.47606933e-01
1.13943326e+00 -5.23703873e-01 -9.17306662e-01 5.81317067e-01
-9.93321955e-01 6.19367778e-01 5.23138225e-01 6.49381161e-01
-1.49183899e-01 7.48959541e-01 -4.75597322e-01 -1.07341206e+00
-8.05525184e-01 -1.22865701e+00 2.28235841e-01 8.34962279e-02
8.21676627e-02 -1.12118125e+00 -3.63285661e-01 7.01530993e-01
7.73447871e-01 1.17747581e+00 9.55792725e-01 -3.14597696e-01
-9.38913584e-01 -6.22753859e-01 -2.44374033e-02 2.17152670e-01
5.87012172e-01 -2.11884752e-01 -5.17308056e-01 -4.60151941e-01
5.92513442e-01 2.90367275e-01 4.61615443e-01 3.11009586e-01
4.19701666e-01 -7.04659879e-01 -1.57747731e-01 5.40270150e-01
2.09682918e+00 7.86599994e-01 1.07447565e+00 7.24966347e-01
3.98432493e-01 5.67359924e-01 4.20593321e-01 3.72598380e-01
1.37129530e-01 1.16668299e-01 4.87856776e-01 -9.68598500e-02
-4.14243713e-02 5.36457403e-03 2.26330489e-01 8.70823264e-01
-3.01452994e-01 -3.73282582e-01 -6.20365798e-01 2.58973449e-01
-1.27410769e+00 -9.25702333e-01 -4.45858985e-01 2.02319241e+00
6.01875544e-01 -2.08844207e-02 -3.59193772e-01 1.01076841e+00
9.16624546e-01 -1.94954485e-01 7.77284876e-02 -6.93037629e-01
1.17510244e-01 4.09920484e-01 1.19071758e+00 6.99410558e-01
-7.49301255e-01 4.08377647e-01 4.63885832e+00 4.26834166e-01
-1.36618686e+00 3.65591675e-01 2.83852637e-01 5.91853917e-01
-5.72880924e-01 2.51228660e-01 -1.41598418e-01 4.79018480e-01
7.09446907e-01 1.07169308e-01 3.08883995e-01 9.72067490e-02
4.89789307e-01 -3.00283909e-01 -8.75404403e-02 6.22512639e-01
2.71385699e-03 -1.10457540e+00 9.48702171e-02 3.96281421e-01
6.77235723e-01 -8.07151318e-01 3.44947606e-01 -8.54986370e-01
-3.22171032e-01 -9.39736784e-01 4.10677195e-02 2.94563740e-01
4.95293707e-01 -9.81678247e-01 9.85241592e-01 3.93145569e-02
-9.33293879e-01 -5.80913275e-02 -2.35606655e-01 2.45227650e-01
1.87967181e-01 -6.79936931e-02 -9.70099747e-01 6.48137808e-01
3.77918750e-01 2.02626407e-01 -1.19069912e-01 8.81147265e-01
2.93382853e-01 5.14640570e-01 -2.43950441e-01 2.61756271e-01
5.31423986e-01 -1.21359639e-01 7.34194100e-01 8.21483731e-01
7.78758645e-01 3.78348678e-01 -4.12231565e-01 1.53552398e-01
3.31445873e-01 3.65957737e-01 -9.40749347e-01 -7.06531107e-02
3.87731433e-01 8.28375995e-01 -1.21735847e+00 -5.14479540e-02
-2.64362425e-01 1.08567870e+00 -1.06951451e+00 8.45878571e-02
-5.67210674e-01 -6.33471489e-01 -3.53814512e-02 5.01394331e-01
2.67645806e-01 -2.73891985e-01 -1.93531305e-01 -5.77023566e-01
-3.85938466e-01 -1.17497981e+00 3.44321877e-01 -2.57256776e-01
-2.72807658e-01 4.72189456e-01 -3.11722636e-01 -1.41226983e+00
3.41910720e-01 -2.18012333e-01 -3.92954648e-01 5.66628754e-01
-1.70551133e+00 -1.59789729e+00 -1.65866196e-01 9.24592614e-01
4.65065181e-01 -3.96638751e-01 9.09388125e-01 3.91525924e-01
-2.71735564e-02 3.47967327e-01 2.11397454e-01 -8.32081959e-02
4.96626526e-01 -7.05778480e-01 -1.71628267e-01 1.11834931e+00
-2.12293610e-01 6.39918268e-01 1.15436792e+00 -8.29901278e-01
-1.73376453e+00 -6.34202778e-01 8.65492165e-01 1.59894884e-01
2.25363955e-01 3.37063342e-01 -6.85850620e-01 7.29380727e-01
7.04196274e-01 -2.59147882e-01 4.93354440e-01 -1.22909713e+00
1.79212362e-01 -8.75966102e-02 -1.85341549e+00 2.05764800e-01
1.74989849e-01 -1.44691885e-01 -4.12963986e-01 2.65255243e-01
4.10580605e-01 -3.37438941e-01 -8.88459980e-01 1.89063191e-01
6.98702753e-01 -9.11945939e-01 9.57668364e-01 -5.61854094e-02
5.77948429e-02 -5.13327003e-01 -1.03748351e-01 -3.86602581e-01
2.02828661e-01 -9.23294604e-01 4.96879905e-01 9.42923427e-01
1.84591874e-01 -1.06107783e+00 6.97895765e-01 3.58961821e-01
2.19620511e-01 -4.58490968e-01 -9.01496470e-01 -3.30151826e-01
-4.18775290e-01 1.00384898e-01 5.08821726e-01 1.16959953e+00
-1.36905685e-01 -5.24465084e-01 -5.68939686e-01 3.31820250e-01
9.81108904e-01 -4.53981042e-01 2.02887103e-01 -8.67795527e-01
-2.68613935e-01 4.12613958e-01 -9.97833490e-01 1.30967006e-01
-6.80756390e-01 -7.53037035e-01 -4.69059765e-01 -1.51771832e+00
-7.94930011e-02 -6.60887122e-01 -3.64951998e-01 1.45468637e-01
2.49743551e-01 8.27258468e-01 9.63028744e-02 3.71785641e-01
2.72130281e-01 -3.51199776e-01 1.41549444e+00 -1.18924258e-02
-1.34287357e-01 -5.03187999e-02 -4.77254450e-01 4.65877414e-01
9.96244669e-01 -8.48367333e-01 -6.93072081e-01 -1.72191769e-01
3.27150583e-01 3.91613394e-01 3.73188049e-01 -1.03571677e+00
2.50265509e-01 4.80358899e-02 1.23893179e-01 -2.50828594e-01
2.17291014e-03 -1.41446185e+00 7.86697507e-01 1.34774911e+00
-6.12122752e-02 3.01850200e-01 -2.10057333e-01 7.69979060e-01
-1.31470904e-01 -2.89573491e-01 7.44449615e-01 -3.62852186e-01
-5.69418311e-01 -1.11250699e-01 -6.18640065e-01 -8.81236017e-01
1.48124790e+00 -1.11486006e+00 -3.03031385e-01 -2.53250122e-01
-6.80158019e-01 -2.01576561e-01 4.54694480e-01 -2.20908169e-02
9.73953307e-01 -8.79716814e-01 -7.68963635e-01 3.48575771e-01
-3.19239140e-01 -5.50593674e-01 4.20209110e-01 1.11093044e+00
-1.53242028e+00 3.72949749e-01 -7.77422249e-01 -2.56933779e-01
-1.94442618e+00 3.47570866e-01 7.51223192e-02 -1.02055661e-01
-1.00825882e+00 3.34453285e-01 -6.99473977e-01 3.22999030e-01
-1.10996723e-01 -1.32934609e-02 -4.96417910e-01 -5.48786044e-01
4.52221483e-01 4.58696306e-01 -1.87377855e-01 -8.43430698e-01
-1.02434181e-01 8.80863070e-01 -6.25280431e-03 -2.56824672e-01
1.24206924e+00 -6.14867389e-01 -5.59009433e-01 -2.97013849e-01
1.21395266e+00 -5.85789122e-02 -6.45702958e-01 1.75050914e-01
-1.19601555e-01 -6.66350842e-01 3.93893212e-01 -8.02524507e-01
-9.80382323e-01 5.59447050e-01 1.26740944e+00 2.29126468e-01
1.31033397e+00 -8.19170833e-01 1.33177686e+00 2.52821296e-01
3.32733274e-01 -8.69381845e-01 -2.91412652e-01 -4.25350070e-01
3.23594719e-01 -1.24606073e+00 9.39002112e-02 -5.87393641e-01
-4.44745332e-01 1.32466722e+00 -2.18737960e-01 -2.39048690e-01
8.05874407e-01 3.70171189e-01 2.46686578e-01 -3.34740937e-01
-2.27479532e-01 3.57823074e-01 -2.56123841e-01 8.74005139e-01
2.88440019e-01 -2.01337859e-01 -1.01245630e+00 -3.94823492e-01
-8.47546943e-03 3.54021877e-01 8.43369901e-01 1.60690153e+00
-8.37264597e-01 -1.36838114e+00 -1.06182861e+00 4.21920940e-02
-1.25795341e+00 1.85512409e-01 5.54354675e-02 7.85180271e-01
3.26999813e-01 1.10555267e+00 -5.22180557e-01 -2.74442106e-01
-2.52021313e-01 -1.17530480e-01 5.44494927e-01 -2.41129361e-02
-6.73658729e-01 1.87143713e-01 -1.93316549e-01 1.70306787e-01
-8.94237280e-01 -3.66772443e-01 -1.48354006e+00 -7.17328310e-01
-3.44381809e-01 1.56702623e-01 1.26061273e+00 1.02842855e+00
-9.01951045e-02 7.53933370e-01 6.35875344e-01 1.16752297e-01
-1.70779794e-01 -4.67619270e-01 -4.65926737e-01 2.34856158e-01
4.58340675e-01 3.47223207e-02 -1.00821495e-01 4.80833441e-01] | [4.298358917236328, 8.051968574523926] |
3e43c0ed-21a1-42d2-9dfe-ccca710a7df0 | rethinking-zero-shot-video-classification-end | 2003.01455 | null | https://arxiv.org/abs/2003.01455v4 | https://arxiv.org/pdf/2003.01455v4.pdf | Rethinking Zero-shot Video Classification: End-to-end Training for Realistic Applications | Trained on large datasets, deep learning (DL) can accurately classify videos into hundreds of diverse classes. However, video data is expensive to annotate. Zero-shot learning (ZSL) proposes one solution to this problem. ZSL trains a model once, and generalizes to new tasks whose classes are not present in the training dataset. We propose the first end-to-end algorithm for ZSL in video classification. Our training procedure builds on insights from recent video classification literature and uses a trainable 3D CNN to learn the visual features. This is in contrast to previous video ZSL methods, which use pretrained feature extractors. We also extend the current benchmarking paradigm: Previous techniques aim to make the test task unknown at training time but fall short of this goal. We encourage domain shift across training and test data and disallow tailoring a ZSL model to a specific test dataset. We outperform the state-of-the-art by a wide margin. Our code, evaluation procedure and model weights are available at github.com/bbrattoli/ZeroShotVideoClassification. | ['Biagio Brattoli', 'Krzysztof Chalupka', 'Joseph Tighe', 'Fedor Zhdanov', 'Pietro Perona'] | 2020-03-03 | rethinking-zero-shot-video-classification-end-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Brattoli_Rethinking_Zero-Shot_Video_Classification_End-to-End_Training_for_Realistic_Applications_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Brattoli_Rethinking_Zero-Shot_Video_Classification_End-to-End_Training_for_Realistic_Applications_CVPR_2020_paper.pdf | cvpr-2020-6 | ['zero-shot-action-recognition'] | ['computer-vision'] | [ 1.82553485e-01 -1.67620361e-01 -5.79211891e-01 -4.31231916e-01
-8.73025298e-01 -9.46947932e-01 6.53008401e-01 -2.19719097e-01
-2.80019611e-01 6.18386745e-01 1.23660505e-01 -2.19070867e-01
1.01585269e-01 -5.84464610e-01 -1.02680302e+00 -3.10711443e-01
-5.25726862e-02 4.32531655e-01 4.62341577e-01 1.21066116e-01
6.09354265e-02 1.01371579e-01 -1.63838112e+00 7.28843331e-01
4.66255933e-01 1.30639124e+00 -2.04976043e-03 7.99969375e-01
2.44420059e-02 1.02939987e+00 -5.00540674e-01 -3.92732263e-01
5.50161183e-01 -4.43237334e-01 -9.34289873e-01 3.70601922e-01
9.62309957e-01 -4.90613133e-01 -6.17203951e-01 7.95539141e-01
4.45933461e-01 1.65383384e-01 7.00788379e-01 -1.67621434e+00
-7.95890331e-01 1.64619312e-01 -1.33953199e-01 3.33508790e-01
4.24824625e-01 2.92968154e-01 9.52779472e-01 -1.13776946e+00
9.81031299e-01 8.46152067e-01 8.42901587e-01 8.98280740e-01
-1.01450384e+00 -5.97105324e-01 2.42387444e-01 4.98041511e-01
-1.30331993e+00 -6.95692360e-01 5.92728078e-01 -8.52466345e-01
1.09821999e+00 1.73264667e-01 6.12766981e-01 1.63928938e+00
-1.28121838e-01 1.13033307e+00 8.35162818e-01 -3.03624183e-01
2.87608266e-01 3.45397852e-02 1.47756323e-01 6.88783109e-01
1.36615828e-01 -9.09222215e-02 -6.14447832e-01 1.41818732e-01
6.85899973e-01 5.98796755e-02 -4.85073805e-01 -9.38396215e-01
-1.19402182e+00 6.93329453e-01 2.14912787e-01 2.98911557e-02
7.18798861e-02 8.92586857e-02 8.29669893e-01 7.03353941e-01
6.05909824e-01 4.70506549e-01 -8.73883128e-01 -4.91836846e-01
-1.02973711e+00 2.22282171e-01 7.92922676e-01 1.14020908e+00
5.89866996e-01 -2.11717766e-02 -2.91734248e-01 7.53304720e-01
-1.80988476e-01 5.41710444e-02 6.48832321e-01 -1.15802419e+00
5.20573020e-01 3.87334704e-01 -1.86600775e-01 -5.16275108e-01
-9.29623097e-02 -3.49560767e-01 -5.08930862e-01 3.31617236e-01
4.10100967e-01 -1.38846830e-01 -1.16048086e+00 1.47955608e+00
1.18276358e-01 6.61343753e-01 1.18738003e-02 9.13581192e-01
1.19753814e+00 5.54587603e-01 -1.10646747e-01 1.10579602e-01
8.20796311e-01 -1.31544054e+00 -2.41306141e-01 -1.87264502e-01
7.64236450e-01 -3.64264369e-01 1.22500622e+00 5.57625175e-01
-7.85273612e-01 -6.55528665e-01 -1.03936863e+00 -2.32540946e-02
-4.16541964e-01 -9.54749528e-03 5.72366416e-01 5.42137146e-01
-1.13935685e+00 6.87666953e-01 -7.57821798e-01 -5.29924214e-01
8.74800324e-01 3.05545717e-01 -6.70012474e-01 -3.23664129e-01
-9.70398366e-01 8.15059304e-01 3.43839765e-01 -2.07267910e-01
-1.27698946e+00 -7.46981740e-01 -1.05298758e+00 -2.27568910e-01
7.30842113e-01 -6.49158418e-01 1.49433267e+00 -1.50310194e+00
-1.31603277e+00 1.11108375e+00 1.33678183e-01 -3.61207068e-01
6.50982082e-01 -2.04514667e-01 -2.36590475e-01 2.92218447e-01
1.25923127e-01 7.99292147e-01 1.02017915e+00 -1.06136918e+00
-5.67655027e-01 8.87954012e-02 5.53552926e-01 1.90442204e-02
-3.91961813e-01 -2.30865508e-01 -8.04659843e-01 -6.48521781e-01
-4.93102759e-01 -8.39033961e-01 8.73026811e-03 3.13837975e-01
-1.22066028e-01 -1.99847057e-01 8.90703321e-01 -3.55133563e-01
1.01800692e+00 -2.25254512e+00 3.26045960e-01 -2.45568633e-01
3.74431252e-01 3.89506280e-01 -3.63282412e-01 2.51651675e-01
-1.62035450e-01 3.18753645e-02 -1.57343000e-01 -2.76674479e-01
8.93486440e-02 2.46881410e-01 -7.67314732e-02 5.48895061e-01
4.33163524e-01 9.13457274e-01 -1.05424440e+00 -4.27199453e-01
2.92172074e-01 2.83663899e-01 -9.96270239e-01 4.61419314e-01
-2.70814061e-01 1.53956652e-01 -1.14644364e-01 9.11953747e-01
3.95583630e-01 -4.32722121e-01 -1.03434436e-02 -2.35658795e-01
1.27410308e-01 1.17771029e-01 -1.01422858e+00 1.92382193e+00
-2.97705889e-01 8.54539335e-01 -2.94441819e-01 -1.41733515e+00
5.45486093e-01 4.83618498e-01 5.94942391e-01 -4.18019325e-01
2.47999094e-02 6.80071339e-02 -2.74324596e-01 -8.75538647e-01
2.51244996e-02 2.94318330e-02 -7.27302134e-02 5.10067195e-02
9.95302975e-01 1.12614110e-01 4.29082155e-01 1.07958935e-01
1.33767092e+00 7.21764743e-01 3.27340603e-01 -1.08611576e-01
2.12100103e-01 7.87583068e-02 7.31539845e-01 8.57929349e-01
-4.68672484e-01 8.59802186e-01 6.28121078e-01 -7.29414940e-01
-1.09554565e+00 -1.04665363e+00 -5.01494855e-02 1.33179545e+00
7.20126033e-02 -5.62517703e-01 -7.15140581e-01 -1.28878295e+00
1.52690724e-01 3.05582464e-01 -8.72995079e-01 -7.65028447e-02
-1.84418529e-01 -2.46959180e-01 3.47872436e-01 6.59910023e-01
3.19119811e-01 -9.04447436e-01 -6.02266073e-01 6.31131008e-02
4.41882797e-02 -1.23458493e+00 -2.99408108e-01 1.38307407e-01
-6.05606139e-01 -1.46174359e+00 -8.35696578e-01 -8.72968495e-01
5.16999185e-01 2.41505638e-01 1.34366703e+00 4.81790155e-02
-3.32636803e-01 7.67285228e-01 -7.08861232e-01 -2.50756234e-01
-8.07122737e-02 2.28141293e-01 4.24468368e-02 -2.64828242e-02
6.89400673e-01 -2.68550098e-01 -7.46567667e-01 2.69514233e-01
-8.59642625e-01 1.15355656e-01 3.09267640e-01 7.75557160e-01
3.34449291e-01 -3.50751013e-01 4.92365122e-01 -9.61038768e-01
2.13861182e-01 -8.18841457e-01 -3.03192973e-01 2.20498458e-01
-2.31432140e-01 -2.16080651e-01 5.81937373e-01 -7.22935498e-01
-4.58303630e-01 1.02029763e-01 -1.03595160e-01 -1.12176740e+00
-4.37528282e-01 4.16899770e-01 -9.03120637e-02 -1.43831044e-01
6.56697929e-01 -3.57194394e-02 -5.76162748e-02 -3.95788670e-01
2.12858796e-01 5.58259845e-01 3.25176239e-01 -4.38356072e-01
6.77611709e-01 2.54696697e-01 -3.97913694e-01 -6.49246871e-01
-1.07651806e+00 -4.70052272e-01 -9.43949580e-01 -6.09484375e-01
7.86535740e-01 -1.16167212e+00 -3.69137019e-01 2.90904075e-01
-8.30180049e-01 -7.04844236e-01 -4.10322636e-01 4.16448683e-01
-7.57908642e-01 1.21818341e-01 -5.44645607e-01 -2.88047105e-01
-4.09684293e-02 -1.12668622e+00 1.04870415e+00 -1.46467581e-01
-3.90272856e-01 -9.72949803e-01 4.46724780e-02 2.33983189e-01
3.65538597e-01 4.07838881e-01 5.31728804e-01 -7.73225069e-01
-4.52886879e-01 -2.25651577e-01 -1.52889282e-01 5.95596850e-01
-4.23500910e-02 3.15608606e-02 -1.12125432e+00 -5.78072727e-01
-2.34344393e-01 -9.94215429e-01 1.07612669e+00 2.75413334e-01
1.49535596e+00 -2.88530231e-01 -1.99879438e-01 8.22010279e-01
1.40411162e+00 -5.59405945e-02 4.00452256e-01 6.35358810e-01
8.10980856e-01 3.35678697e-01 5.96949220e-01 4.17811692e-01
3.67179632e-01 6.87350035e-01 4.91006702e-01 -4.48626131e-02
-1.98849663e-01 -1.63756222e-01 5.79898000e-01 5.97962737e-01
-2.18237072e-01 -4.03076977e-01 -8.74089241e-01 5.92542529e-01
-1.85787177e+00 -1.21435523e+00 3.35783541e-01 2.06238890e+00
7.03961194e-01 3.59280944e-01 2.94327319e-01 8.00997987e-02
4.06050950e-01 2.95236707e-01 -5.89794457e-01 -1.85556024e-01
2.05378696e-01 2.93894947e-01 4.90833998e-01 2.81732738e-01
-1.52223682e+00 1.05216384e+00 6.20281839e+00 8.40646207e-01
-1.32667553e+00 2.18053460e-01 5.94606698e-01 -4.72867012e-01
-2.63115037e-02 -9.64895114e-02 -7.09116042e-01 4.42740172e-01
8.77813816e-01 -1.47590563e-01 2.77842015e-01 1.02770984e+00
-1.57647491e-01 3.08380932e-01 -1.52069497e+00 1.27328634e+00
5.04723668e-01 -1.54654336e+00 1.50025129e-01 -2.02356830e-01
7.81542063e-01 3.61538261e-01 -2.57369000e-02 8.06505620e-01
1.01279624e-01 -1.01321483e+00 8.54424119e-01 4.22418296e-01
1.11519861e+00 -4.73076820e-01 4.89434153e-01 1.54379621e-01
-1.11905050e+00 -2.89432019e-01 -3.80504757e-01 -1.39382377e-01
-5.09375073e-02 8.72999877e-02 -6.88737452e-01 3.52244556e-01
8.31622005e-01 1.36627233e+00 -8.24099898e-01 1.28589296e+00
-5.40938154e-02 7.52043843e-01 1.14043191e-01 2.60958165e-01
4.42567617e-01 5.55621088e-02 3.56292903e-01 1.35435832e+00
3.06461155e-01 -6.19539693e-02 5.25707126e-01 3.43014956e-01
-3.12128693e-01 -1.19036414e-01 -8.98641586e-01 -1.00685984e-01
1.31775022e-01 9.31542397e-01 -5.06486416e-01 -6.01309538e-01
-9.26125288e-01 1.18579400e+00 4.51762646e-01 5.18171608e-01
-8.56530905e-01 -2.57769704e-01 6.88267469e-01 2.83899844e-01
5.68324685e-01 -1.93019226e-01 1.78672761e-01 -1.66261768e+00
3.77010256e-02 -1.10691559e+00 4.75032628e-01 -6.51408374e-01
-1.41493011e+00 6.41299486e-01 2.83700433e-02 -1.81718874e+00
-3.64645422e-01 -1.02125108e+00 -3.20891589e-01 1.78182364e-01
-1.46163535e+00 -1.20804000e+00 -4.33938086e-01 7.14518189e-01
1.13780177e+00 -4.23719317e-01 8.93050730e-01 4.72069949e-01
-4.45337504e-01 6.37139380e-01 1.52891323e-01 5.10731816e-01
8.73994410e-01 -1.14196074e+00 4.10915792e-01 6.61626160e-01
2.67553747e-01 1.55682236e-01 5.79423964e-01 -3.72150421e-01
-1.48255229e+00 -1.24234140e+00 5.89580595e-01 -8.72646272e-01
8.52727294e-01 -7.23813593e-01 -8.29937637e-01 9.59804595e-01
2.19664812e-01 6.49390280e-01 9.53406155e-01 1.49124920e-01
-6.33474052e-01 -4.72280905e-02 -8.14422786e-01 3.56688082e-01
1.44077587e+00 -6.83526933e-01 -4.75177735e-01 3.82470787e-01
6.09528899e-01 -5.48078358e-01 -8.59134495e-01 5.00124335e-01
7.84014702e-01 -8.52956295e-01 8.48722219e-01 -1.08174133e+00
7.99773276e-01 -1.18075773e-01 -2.97764331e-01 -1.34755123e+00
-2.82236964e-01 -2.85128087e-01 -4.58644837e-01 1.00918055e+00
4.74109888e-01 -3.12590688e-01 8.96762192e-01 4.71172273e-01
-3.70330870e-01 -8.52212310e-01 -8.61267984e-01 -1.03637147e+00
-2.39247289e-02 -5.70669413e-01 2.25269824e-01 1.34039092e+00
-2.24750526e-02 2.51953244e-01 -5.03163278e-01 -7.50853121e-02
6.90216720e-01 8.02426040e-02 1.03883195e+00 -1.24931431e+00
-5.70011735e-01 -3.83970618e-01 -8.40112925e-01 -1.00927567e+00
3.90157312e-01 -9.91869748e-01 -1.13467075e-01 -1.45836103e+00
3.62265319e-01 -2.52064317e-01 -4.61993873e-01 6.75089478e-01
3.92429344e-02 6.64091051e-01 3.93597633e-01 7.53289014e-02
-1.17206228e+00 4.19692487e-01 1.00734270e+00 -3.09180915e-01
7.04013780e-02 -4.08811532e-02 -4.21556443e-01 6.05431318e-01
6.71906352e-01 -2.65775114e-01 -6.73126161e-01 -6.40681565e-01
-1.29135221e-01 -1.22432537e-01 4.55360204e-01 -1.21969688e+00
-2.35493518e-02 -1.62339494e-01 5.94790578e-01 -1.87268704e-01
3.43469292e-01 -7.75305331e-01 -5.67488745e-02 8.70906711e-02
-4.80201572e-01 -2.01222301e-01 1.60419285e-01 4.60395068e-01
-2.80134678e-01 -2.52676606e-01 6.25827909e-01 -2.82396346e-01
-1.32740474e+00 8.03757071e-01 -3.32358629e-01 2.78109819e-01
1.27610743e+00 -4.09726679e-01 -4.43534821e-01 -3.00159723e-01
-9.14328933e-01 1.85901225e-01 7.27295518e-01 8.75754714e-01
6.63349748e-01 -1.41608131e+00 -6.00436985e-01 2.37950608e-01
4.86754209e-01 -1.12881735e-01 8.81321132e-02 5.44606745e-01
-5.59191763e-01 2.76571423e-01 -4.03994650e-01 -6.17217898e-01
-1.31422353e+00 7.52162695e-01 4.63993162e-01 1.89864650e-01
-6.71633005e-01 1.05468261e+00 1.37888670e-01 -3.98829579e-01
4.28404003e-01 -1.69209957e-01 -1.63283303e-01 1.06648013e-01
5.15961170e-01 1.05075352e-01 6.40276074e-02 -5.49642622e-01
-3.84097397e-01 5.11423528e-01 -6.26141131e-02 2.05382451e-01
1.56163228e+00 1.84397817e-01 4.41794544e-01 7.10352719e-01
1.58991027e+00 -4.21210021e-01 -1.77042294e+00 -1.63485318e-01
-2.75346916e-02 -7.25140989e-01 -6.10960461e-02 -5.72373509e-01
-1.03396320e+00 9.28525925e-01 5.91260791e-01 1.32420242e-01
9.91385698e-01 1.58469215e-01 5.16717315e-01 2.96773314e-01
3.47927600e-01 -1.13335967e+00 3.42065722e-01 6.04246318e-01
7.87939370e-01 -1.64109528e+00 -1.08701751e-01 -1.72342435e-01
-8.19206953e-01 1.17885208e+00 9.70157802e-01 -3.58114272e-01
6.51975334e-01 1.36482507e-01 2.84226350e-02 -1.49934784e-01
-1.09038973e+00 -2.54926234e-01 3.76319408e-01 7.08741963e-01
5.01241744e-01 -2.50366449e-01 1.41583934e-01 4.77350116e-01
1.28451988e-01 4.78157014e-01 4.71158326e-01 1.21337974e+00
-3.70238900e-01 -1.00782633e+00 4.78166044e-02 6.71200097e-01
-3.93772155e-01 -6.52740477e-03 -2.62836814e-01 8.12571287e-01
3.00652921e-01 6.43875778e-01 2.14263111e-01 -5.99551976e-01
4.06318665e-01 1.35991275e-01 5.09729087e-01 -8.29218030e-01
-1.78231135e-01 -2.09465072e-01 1.69422761e-01 -7.60073304e-01
-6.03533745e-01 -5.32215238e-01 -6.86510205e-01 -1.24335706e-01
-7.63416290e-02 4.45960797e-02 1.91047907e-01 9.60307956e-01
3.50914538e-01 3.94289404e-01 6.24209046e-01 -1.22881699e+00
-3.17699641e-01 -7.92328715e-01 -3.03878546e-01 6.79271460e-01
6.35468602e-01 -9.93395269e-01 -2.94752866e-01 4.79386240e-01] | [8.998577117919922, 1.1414693593978882] |
07c29d15-c33c-4b9b-981a-7915cca21e72 | improving-multi-hop-knowledge-base-question | 2101.03737 | null | https://arxiv.org/abs/2101.03737v2 | https://arxiv.org/pdf/2101.03737v2.pdf | Improving Multi-hop Knowledge Base Question Answering by Learning Intermediate Supervision Signals | Multi-hop Knowledge Base Question Answering (KBQA) aims to find the answer entities that are multiple hops away in the Knowledge Base (KB) from the entities in the question. A major challenge is the lack of supervision signals at intermediate steps. Therefore, multi-hop KBQA algorithms can only receive the feedback from the final answer, which makes the learning unstable or ineffective. To address this challenge, we propose a novel teacher-student approach for the multi-hop KBQA task. In our approach, the student network aims to find the correct answer to the query, while the teacher network tries to learn intermediate supervision signals for improving the reasoning capacity of the student network. The major novelty lies in the design of the teacher network, where we utilize both forward and backward reasoning to enhance the learning of intermediate entity distributions. By considering bidirectional reasoning, the teacher network can produce more reliable intermediate supervision signals, which can alleviate the issue of spurious reasoning. Extensive experiments on three benchmark datasets have demonstrated the effectiveness of our approach on the KBQA task. The code to reproduce our analysis is available at https://github.com/RichardHGL/WSDM2021_NSM. | ['Ji-Rong Wen', 'Wayne Xin Zhao', 'Jing Jiang', 'Yunshi Lan', 'Gaole He'] | 2021-01-11 | null | null | null | null | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-2.32995614e-01 4.73638445e-01 -2.82330066e-01 -3.41388673e-01
-9.60613489e-01 -5.21147370e-01 8.90959576e-02 3.16738427e-01
-2.35904828e-01 1.07497013e+00 5.25453649e-02 -4.60007548e-01
-4.24446225e-01 -1.12509704e+00 -9.21446383e-01 -5.75350821e-01
4.57927436e-01 5.36267340e-01 8.35964322e-01 -5.27118683e-01
5.93859935e-03 3.38301845e-02 -1.16687977e+00 5.45637429e-01
1.38507533e+00 1.00097334e+00 5.34196682e-02 4.37243998e-01
-3.59629840e-01 1.34434235e+00 -5.07613420e-01 -6.84200585e-01
-8.32714513e-02 -5.66771328e-01 -1.26143074e+00 -6.88923120e-01
3.80211473e-01 -3.74728292e-01 -3.55562150e-01 1.10888135e+00
5.13241053e-01 2.14733988e-01 2.51112044e-01 -1.38311052e+00
-7.69083917e-01 8.86169136e-01 -2.95745313e-01 3.95489573e-01
4.08019751e-01 -1.59000903e-01 1.48057258e+00 -9.12136078e-01
3.10706049e-01 1.15079534e+00 4.53808278e-01 6.78367734e-01
-8.42376947e-01 -6.66019380e-01 2.57199407e-01 7.78839469e-01
-1.43622792e+00 -3.43856722e-01 7.24777102e-01 -4.52484228e-02
4.98113632e-01 1.24610901e-01 3.42554659e-01 6.06506705e-01
-4.27265227e-01 1.12019587e+00 6.79068387e-01 -3.72354954e-01
1.17556460e-01 4.19779181e-01 5.44655979e-01 1.02580583e+00
3.50017622e-02 -3.33273530e-01 -7.51969159e-01 -3.54120255e-01
3.24970126e-01 -2.76716709e-01 -5.15263736e-01 -3.71699065e-01
-8.98973107e-01 8.25825632e-01 8.77610922e-01 3.00998807e-01
-2.85952181e-01 7.40268901e-02 9.32834446e-02 5.30524194e-01
1.81942388e-01 5.00745773e-01 -7.42606223e-01 9.88775678e-03
-5.08204579e-01 2.02564582e-01 9.13007319e-01 7.90281951e-01
1.03083122e+00 -4.90346014e-01 -2.85575598e-01 6.72497630e-01
4.63056475e-01 5.08877993e-01 1.58911511e-01 -9.45516229e-01
8.42654288e-01 8.91902566e-01 1.51016012e-01 -9.34433103e-01
-1.55254593e-02 -3.18202794e-01 -4.25202250e-01 -3.45722884e-01
6.86845541e-01 -3.65577698e-01 -6.72858417e-01 1.69788027e+00
8.50903273e-01 4.19225305e-01 2.97543705e-01 1.02272892e+00
1.13253629e+00 7.33728170e-01 3.55538120e-03 1.40055820e-01
1.32563293e+00 -1.19166684e+00 -5.62572837e-01 -4.42494787e-02
8.51298451e-01 -3.91155511e-01 8.20441663e-01 1.30376086e-01
-1.08221543e+00 -3.81830573e-01 -7.72140384e-01 -3.12714905e-01
-2.34366730e-01 -8.66713282e-03 2.71945179e-01 2.31137112e-01
-8.95240545e-01 9.32573453e-02 -5.03651202e-01 1.13131158e-01
3.86472046e-01 3.03080738e-01 -6.38308600e-02 -3.90098631e-01
-1.87282741e+00 8.25948954e-01 3.94882083e-01 2.59439528e-01
-6.84796929e-01 -8.12786877e-01 -6.15512967e-01 3.59565854e-01
7.50668764e-01 -7.14074552e-01 1.37793100e+00 -8.73248637e-01
-1.25222862e+00 2.11756021e-01 -3.78415316e-01 -3.77842337e-01
4.46405530e-01 -6.74071461e-02 -2.43146256e-01 4.38858479e-01
1.81461036e-01 6.56547844e-01 4.86440539e-01 -1.28756261e+00
-9.41060483e-01 -3.09111118e-01 4.10164535e-01 3.29708040e-01
-2.94675946e-01 -4.99697864e-01 -4.69598889e-01 -8.11429992e-02
9.75776464e-02 -6.43823028e-01 5.43395206e-02 -2.45655820e-01
-2.27457076e-01 -7.42644608e-01 7.19021976e-01 -6.61468148e-01
1.24109828e+00 -2.10695195e+00 1.29386514e-01 3.91159415e-01
3.72674435e-01 3.66200447e-01 -1.99972942e-01 3.35955679e-01
1.89199954e-01 -1.63132828e-02 6.47062762e-03 1.47961870e-01
-1.66956007e-01 3.11488211e-01 -5.52131832e-01 8.57881531e-02
3.25623631e-01 1.03158224e+00 -1.16306484e+00 -6.33951902e-01
-3.96142066e-01 1.77458271e-01 -5.30060053e-01 3.34647566e-01
-6.43865883e-01 3.88511151e-01 -9.19227958e-01 5.58716416e-01
5.46421945e-01 -6.50362790e-01 1.20391257e-01 -1.72657266e-01
2.94191688e-01 6.62580729e-01 -1.25064719e+00 1.33874547e+00
-4.34312880e-01 2.66159177e-01 1.56913504e-01 -1.01437545e+00
8.18243265e-01 4.44797337e-01 1.42649308e-01 -9.51343894e-01
-2.18862310e-01 4.32563931e-01 1.55718878e-01 -5.76035202e-01
3.05222899e-01 -9.34498087e-02 1.63890988e-01 3.52165401e-01
3.85644920e-02 1.25870183e-01 1.26130715e-01 5.14470458e-01
1.08839262e+00 -2.60309666e-01 -5.19714803e-02 1.01399370e-01
8.74685585e-01 2.39621326e-01 7.56993830e-01 7.00378180e-01
-1.90807745e-01 6.32397607e-02 6.57147706e-01 -1.00418203e-01
-3.68452102e-01 -1.11277306e+00 2.50743568e-01 1.27729094e+00
5.01259625e-01 -2.65371829e-01 -4.16863263e-01 -1.15471327e+00
2.53939122e-01 7.58436501e-01 -4.72626060e-01 -3.85420471e-01
-6.20371103e-01 -1.34378180e-01 5.26420355e-01 5.61737120e-01
5.74147522e-01 -9.64391351e-01 -6.19835109e-02 2.07836982e-02
-6.23734891e-01 -8.68600070e-01 -2.40876839e-01 -6.23869477e-03
-8.24000537e-01 -1.31721878e+00 -4.84608233e-01 -8.59072387e-01
8.45387280e-01 2.54517913e-01 1.02489722e+00 5.60951889e-01
2.49350414e-01 3.63112271e-01 -4.12255347e-01 -1.55261993e-01
-3.02945197e-01 3.93272400e-01 -2.78303504e-01 -1.06794931e-01
6.68037951e-01 -3.30262184e-01 -6.76766396e-01 3.96728635e-01
-7.20892251e-01 -1.32221580e-01 6.13150775e-01 9.71310318e-01
3.74944627e-01 1.63854137e-01 1.16307294e+00 -1.05825412e+00
7.23703086e-01 -8.37427795e-01 -6.12551093e-01 5.95030785e-01
-6.96645975e-01 3.42016995e-01 8.27730358e-01 -2.95223057e-01
-1.23744094e+00 -2.58325458e-01 -2.97546566e-01 -2.80175835e-01
2.03389481e-01 6.81119621e-01 -8.80866349e-02 5.19335829e-02
5.61343133e-01 1.85549185e-01 -2.58359522e-01 -2.98213840e-01
4.97496098e-01 5.33559740e-01 2.35056594e-01 -7.96656549e-01
8.35193694e-01 2.62573451e-01 -3.16241264e-01 -3.82430851e-01
-1.28004801e+00 -6.55766070e-01 -3.82513851e-01 -6.48330972e-02
5.23978770e-01 -9.76777673e-01 -9.62441742e-01 1.28943413e-01
-1.15737116e+00 -3.33987296e-01 -3.18559319e-01 2.87795126e-01
-8.99581015e-02 1.12928949e-01 -5.79777360e-01 -7.15426385e-01
-2.74994373e-01 -9.58625793e-01 5.69386423e-01 5.85241556e-01
4.50610779e-02 -1.06563795e+00 5.15623912e-02 8.42297912e-01
2.29971081e-01 -4.92738754e-01 1.21998346e+00 -9.76788461e-01
-8.74181390e-01 1.11493632e-01 -3.41793627e-01 3.06672215e-01
7.67194107e-02 -2.93452859e-01 -8.56119931e-01 -3.79033722e-02
-1.91514075e-01 -7.62111008e-01 1.05228865e+00 -1.98842391e-01
8.73563826e-01 -3.03987026e-01 -2.66042262e-01 -3.00193788e-03
1.16378856e+00 -5.44423349e-02 3.12726200e-01 2.12893620e-01
7.18049228e-01 7.22007632e-01 8.55735064e-01 7.68263042e-02
9.52307105e-01 3.81341308e-01 3.39355916e-01 1.15420863e-01
-5.44645414e-02 -5.91197073e-01 3.08276236e-01 9.45466340e-01
3.24888408e-01 -1.98916569e-01 -1.12346900e+00 8.87661636e-01
-2.00463057e+00 -7.53077447e-01 -2.04639807e-01 1.87983143e+00
1.22708845e+00 -1.25394300e-01 -1.02413006e-01 -2.32665669e-02
4.97069269e-01 -1.02064610e-01 -7.26260006e-01 -1.09550700e-01
2.15705931e-01 7.53044933e-02 9.73614082e-02 8.52667630e-01
-5.44694185e-01 9.27204013e-01 4.86990786e+00 7.71158218e-01
-8.99416983e-01 1.58576205e-01 3.87764513e-01 1.77450225e-01
-7.17164874e-01 6.74802586e-02 -1.10077155e+00 3.94605815e-01
8.57967377e-01 -1.24045238e-01 2.42227778e-01 6.34959280e-01
-1.05256349e-01 -2.39050299e-01 -1.04167438e+00 3.41300607e-01
-7.27856457e-02 -1.24037611e+00 1.52905568e-01 -4.03633058e-01
7.28053510e-01 4.99321558e-02 -1.00516707e-01 5.94664693e-01
5.39267421e-01 -7.36600578e-01 3.88402075e-01 5.50157368e-01
2.30565622e-01 -8.11400950e-01 8.48229527e-01 6.66534305e-01
-1.10008836e+00 -3.06507856e-01 -3.86583984e-01 2.58214980e-01
-3.65341641e-02 6.85043991e-01 -1.22389627e+00 7.85369098e-01
7.38994241e-01 3.30200881e-01 -6.00982845e-01 8.34290326e-01
-7.64751554e-01 7.80809104e-01 -2.75793582e-01 -3.16621274e-01
2.36410692e-01 -9.47535560e-02 2.39427224e-01 7.81981528e-01
1.52438551e-01 4.16271985e-01 9.48322788e-02 8.16664517e-01
-5.01212180e-01 1.34570211e-01 -2.97062933e-01 -3.26916091e-02
8.19999695e-01 1.01967812e+00 -1.28634602e-01 -3.08800697e-01
-5.06409287e-01 6.99481308e-01 9.34189200e-01 5.24961412e-01
-7.04801738e-01 -5.07114828e-01 4.66835231e-01 -8.72341022e-02
4.25761342e-01 1.97793186e-01 6.31257668e-02 -1.03941238e+00
2.89213955e-01 -9.15442467e-01 8.78798723e-01 -7.86963046e-01
-1.39630711e+00 2.23913014e-01 -2.37374604e-01 -6.62753761e-01
-2.16896847e-01 -3.11627179e-01 -5.65134823e-01 9.58474457e-01
-1.99788642e+00 -8.85359466e-01 -3.47080797e-01 7.73217201e-01
1.01514697e-01 1.28827587e-01 4.80978459e-01 5.03116548e-01
-4.28624183e-01 6.48565292e-01 -4.33116406e-02 3.34193170e-01
7.91154981e-01 -1.34422898e+00 -1.59762368e-01 4.94014770e-01
1.51866302e-01 6.41945422e-01 4.39599246e-01 -5.79156280e-01
-1.50765419e+00 -1.00699818e+00 9.89048362e-01 -5.32931328e-01
8.19716036e-01 -6.93339109e-02 -1.35858667e+00 6.88476443e-01
4.90440764e-02 6.33314531e-03 7.56322980e-01 3.97277385e-01
-5.37930250e-01 -4.23132867e-01 -1.08283865e+00 5.30185342e-01
5.05151570e-01 -6.07175112e-01 -9.15318191e-01 1.78655788e-01
9.90334392e-01 -4.03087199e-01 -8.88020515e-01 3.91141623e-01
2.44676664e-01 -6.52353525e-01 8.11501861e-01 -6.67115629e-01
3.65174681e-01 -6.55121207e-01 1.81209505e-01 -1.33344698e+00
-2.58428156e-02 -4.14845310e-02 -5.83599865e-01 1.25944352e+00
9.37100947e-01 -7.34129310e-01 8.66425276e-01 5.22569418e-01
1.56337559e-01 -1.06435847e+00 -9.92649615e-01 -3.69106829e-01
1.90078214e-01 -2.84752287e-02 5.80464542e-01 9.77855861e-01
3.45152989e-02 6.00165427e-01 1.88052189e-02 6.55298650e-01
2.85650134e-01 5.30891657e-01 6.33275211e-01 -1.08196771e+00
-3.28216612e-01 -8.14047009e-02 9.84226689e-02 -1.56859732e+00
2.51209050e-01 -8.89046550e-01 1.88641459e-01 -1.75002730e+00
1.09874204e-01 -8.05606306e-01 -3.77246886e-01 6.54533386e-01
-6.28076494e-01 -2.18870193e-01 2.95089465e-02 -1.23599879e-02
-1.06258738e+00 6.74715161e-01 1.49329984e+00 -1.86592087e-01
-6.64180666e-02 2.44087458e-01 -9.91497695e-01 4.35713768e-01
8.28891516e-01 -7.36546338e-01 -8.32999825e-01 -6.31657779e-01
6.34965658e-01 1.63574457e-01 3.56253594e-01 -4.84401643e-01
8.88935924e-01 -7.01497868e-02 2.03592151e-01 -5.80563128e-01
3.23865116e-01 -8.54485452e-01 -4.28629875e-01 4.33983237e-01
-4.60674614e-01 -1.62373513e-01 3.08093335e-02 6.55691266e-01
-4.84334022e-01 -4.70455438e-01 4.25960034e-01 -3.96305881e-02
-5.79322278e-01 2.39272192e-01 5.64737320e-02 5.03536582e-01
8.43747199e-01 4.00010914e-01 -8.33404660e-01 -4.96285617e-01
-5.83423257e-01 1.12057877e+00 4.68972512e-02 2.61802465e-01
6.59426510e-01 -1.18815804e+00 -7.56868124e-01 -1.41516432e-01
1.00656196e-01 5.70118129e-01 2.80161023e-01 1.02206743e+00
-1.85623616e-01 5.91794074e-01 2.87025660e-01 -3.92707765e-01
-1.12656558e+00 3.61772716e-01 4.90862221e-01 -3.83668900e-01
-2.59010136e-01 1.03056788e+00 -6.92648217e-02 -7.26937473e-01
2.66361684e-01 -1.58943415e-01 -2.96418965e-01 1.72348231e-01
5.33566058e-01 4.25226301e-01 1.10383965e-01 -1.77704006e-01
-3.13343734e-01 9.99965817e-02 -3.63583893e-01 6.06622361e-02
1.13295770e+00 -1.63126156e-01 -2.72755742e-01 3.95754427e-01
9.14429903e-01 2.51120925e-01 -8.10517907e-01 -5.50339520e-01
2.78879941e-01 -2.37668276e-01 -1.46068469e-01 -9.21595991e-01
-1.02290618e+00 8.29722464e-01 1.29219323e-01 3.04260194e-01
1.02611470e+00 3.34664404e-01 9.80571985e-01 9.36114311e-01
2.28605106e-01 -8.49326193e-01 1.65796697e-01 6.56459570e-01
6.04520261e-01 -1.22399676e+00 -2.51154542e-01 -4.83136952e-01
-6.02614403e-01 8.61675382e-01 1.10728538e+00 3.26001793e-01
5.25950611e-01 -1.28401607e-01 2.98465997e-01 -3.38782817e-01
-1.11122239e+00 -1.47178143e-01 2.59360403e-01 7.81298652e-02
2.98356563e-01 -2.15560198e-01 -1.49551302e-01 6.76826358e-01
-8.53289757e-03 -1.62401125e-01 2.90837437e-01 8.80999744e-01
-6.45380855e-01 -1.04629683e+00 -1.94334328e-01 4.32778418e-01
-3.15760732e-01 -1.86303392e-01 -5.59026778e-01 4.78173465e-01
2.41194926e-02 1.12147629e+00 -3.00208539e-01 -1.13148384e-01
4.86015230e-01 4.04518127e-01 2.34895572e-01 -6.12842560e-01
-5.62693000e-01 -5.97278833e-01 1.87546864e-01 -3.92925173e-01
-3.49673837e-01 -1.60615340e-01 -1.64916587e+00 -2.81309068e-01
-6.77408397e-01 9.99740243e-01 1.87403530e-01 1.00092447e+00
5.38738132e-01 5.51203966e-01 5.95194757e-01 4.39587146e-01
-7.87812412e-01 -7.21883178e-01 -2.89590150e-01 1.47795260e-01
6.71507776e-01 -5.32453597e-01 -4.09286439e-01 -2.89555371e-01] | [10.787293434143066, 7.922609806060791] |
055f4f49-dd4b-4c04-924b-5a27f64f0cc9 | the-offense-defense-balance-of-scientific | 2001.00463 | null | https://arxiv.org/abs/2001.00463v2 | https://arxiv.org/pdf/2001.00463v2.pdf | The Offense-Defense Balance of Scientific Knowledge: Does Publishing AI Research Reduce Misuse? | There is growing concern over the potential misuse of artificial intelligence (AI) research. Publishing scientific research can facilitate misuse of the technology, but the research can also contribute to protections against misuse. This paper addresses the balance between these two effects. Our theoretical framework elucidates the factors governing whether the published research will be more useful for attackers or defenders, such as the possibility for adequate defensive measures, or the independent discovery of the knowledge outside of the scientific community. The balance will vary across scientific fields. However, we show that the existing conversation within AI has imported concepts and conclusions from prior debates within computer security over the disclosure of software vulnerabilities. While disclosure of software vulnerabilities often favours defence, this cannot be assumed for AI research. The AI research community should consider concepts and policies from a broad set of adjacent fields, and ultimately needs to craft policy well-suited to its particular challenges. | ['Toby Shevlane', 'Allan Dafoe'] | 2019-12-27 | null | null | null | null | ['computer-security'] | ['miscellaneous'] | [ 3.77511799e-01 2.58121818e-01 -3.20118219e-01 2.93877963e-02
-2.17638507e-01 -1.04734302e+00 7.40087688e-01 3.21515769e-01
-6.17701054e-01 4.63075221e-01 2.24029377e-01 -1.42539418e+00
-1.64210543e-01 -7.06356049e-01 -5.81787884e-01 -6.77803278e-01
4.59666461e-01 -1.42344594e-01 1.02114722e-01 1.80294681e-02
1.08841527e+00 7.84708083e-01 -1.25967908e+00 -3.47625017e-01
5.81179380e-01 4.26574469e-01 -5.10086954e-01 5.04333138e-01
-1.99825048e-01 8.78326416e-01 -9.76978779e-01 -5.78951180e-01
4.14529860e-01 -2.59267658e-01 -8.13999236e-01 -2.74167150e-01
4.81008366e-02 -2.13205829e-01 7.14418367e-02 1.33682692e+00
1.46609023e-01 -4.10726219e-01 3.06386232e-01 -1.43106580e+00
-6.16147339e-01 7.56495178e-01 -6.58484936e-01 5.38588822e-01
-5.47007881e-02 4.38336045e-01 5.64183831e-01 -3.88348661e-02
6.49915159e-01 1.03641450e+00 2.95683593e-01 4.39240843e-01
-9.66490328e-01 -1.12751341e+00 1.40019342e-01 -1.78355649e-01
-1.14956057e+00 -5.19848883e-01 4.90095645e-01 -6.72428370e-01
4.89190042e-01 5.43977618e-01 4.91621077e-01 1.09959424e+00
4.20173049e-01 -4.95010540e-02 1.33653879e+00 -5.27544498e-01
3.85492563e-01 4.02426064e-01 4.28860813e-01 8.47332552e-02
1.53175545e+00 8.49666670e-02 -3.16324204e-01 -7.95854628e-01
3.92590970e-01 -6.08877242e-01 -2.98790962e-01 -1.29999027e-01
-9.42334592e-01 7.02043295e-01 -2.21909985e-01 8.13583851e-01
-2.68418044e-01 -4.95843738e-02 4.36951399e-01 2.77008235e-01
3.08738202e-02 1.03187919e+00 -2.72765517e-01 -4.43032295e-01
-5.18589318e-01 2.67331749e-01 1.07388473e+00 1.27900928e-01
1.87351212e-01 1.06026925e-01 7.65020192e-01 -1.01963729e-01
5.65350950e-01 1.32332265e-01 -1.13277826e-02 -1.02524281e+00
1.31348506e-01 4.13643867e-01 2.25267082e-01 -1.31537282e+00
1.61877230e-01 -3.77007872e-01 -1.69228658e-01 6.90767765e-01
6.85862362e-01 -3.51550013e-01 -4.06939179e-01 1.57879317e+00
1.91909134e-01 -1.16277218e-01 1.93017066e-01 7.11534023e-01
2.22937971e-01 5.37340701e-01 4.50262338e-01 -9.47929099e-02
1.25485766e+00 8.55575968e-03 -7.09068179e-01 -4.09558386e-01
6.07898116e-01 -7.25547314e-01 4.10616368e-01 4.83851224e-01
-9.53316450e-01 2.14120269e-01 -1.39243972e+00 2.71950185e-01
-6.25843525e-01 -7.51165748e-01 6.53180718e-01 1.44244051e+00
-7.60144293e-01 3.10102850e-01 -7.70815372e-01 -4.21758592e-01
3.79097879e-01 1.37841597e-01 4.99372184e-03 2.76514113e-01
-1.16501248e+00 1.10759771e+00 3.95435870e-01 2.93365806e-01
-3.41597527e-01 -5.35876632e-01 -4.20752436e-01 -1.64263129e-01
4.54773337e-01 -4.62656319e-01 8.90551150e-01 -1.17544580e+00
-9.17321980e-01 8.13659489e-01 5.20511210e-01 -5.74252129e-01
3.56455684e-01 -1.64972097e-02 -5.45782030e-01 6.22865409e-02
1.76913649e-01 6.16611354e-02 5.92638671e-01 -1.32265842e+00
-7.31900871e-01 -5.87690771e-01 2.95252562e-01 -1.68917656e-01
-4.81044441e-01 7.34244645e-01 3.83246183e-01 -5.64166009e-01
-2.92840749e-01 -9.81752932e-01 -1.09816231e-01 -1.65144295e-01
-1.85298771e-01 -4.98385355e-02 1.09344077e+00 -3.91620696e-01
1.17343235e+00 -2.12137747e+00 -5.34786224e-01 2.33959302e-01
1.05245344e-01 3.79648149e-01 3.94012570e-01 4.24886793e-01
-1.23442579e-02 1.12270439e+00 7.19675869e-02 6.53239489e-01
-1.43076867e-01 4.01895940e-02 -7.69141376e-01 8.30676436e-01
8.45694542e-02 3.70350271e-01 -5.54451942e-01 -8.26100931e-02
-2.50911385e-01 3.67438227e-01 -2.53276937e-02 -1.80960298e-01
1.66538358e-01 3.65065992e-01 -9.56310809e-01 7.36284971e-01
6.62763000e-01 -9.48349983e-02 4.65269268e-01 4.13224250e-01
-8.11469674e-01 5.47463775e-01 -7.78735280e-01 8.12545776e-01
7.93486089e-02 5.91544211e-01 7.24066436e-01 -9.53580081e-01
7.77065396e-01 4.32030767e-01 6.42880350e-02 -3.05946767e-01
4.33843881e-01 3.22573960e-01 7.60694325e-01 -3.26268315e-01
2.25713313e-01 -9.26852077e-02 3.38513590e-02 1.22680044e+00
-7.55738318e-01 -1.15824662e-01 -3.19628865e-01 2.46082872e-01
1.03630435e+00 1.66728124e-01 1.09560065e-01 -6.39351189e-01
2.85230339e-01 2.84953535e-01 5.41903019e-01 8.24561834e-01
-4.02217835e-01 -3.20057303e-01 5.81189632e-01 -3.81473690e-01
-8.37580562e-01 -5.52299261e-01 -4.33672816e-01 9.36129570e-01
2.30755299e-01 -3.88420433e-01 -7.70352602e-01 -8.37414563e-01
1.16748296e-01 1.28323197e+00 -5.02974689e-01 -3.52357775e-01
-3.86136383e-01 -6.51076496e-01 9.16999161e-01 1.75573125e-01
1.71447054e-01 -8.16777587e-01 -1.57393837e+00 9.61116329e-03
2.68719316e-01 -8.78482044e-01 1.35444656e-01 6.78188726e-02
-6.44040585e-01 -1.14586866e+00 -2.04375595e-01 -1.23772718e-01
6.23093426e-01 3.03770721e-01 6.41867101e-01 5.74498355e-01
-2.18009353e-01 6.65187180e-01 -2.28309274e-01 -1.22657335e+00
-1.15423369e+00 -1.08011477e-01 -1.70595407e-01 -3.79643381e-01
7.18753517e-01 -5.76582551e-01 -2.28046943e-02 8.57711285e-02
-1.19173062e+00 -3.41161221e-01 4.84130055e-01 2.99543828e-01
-3.42729062e-01 3.68788421e-01 8.16151381e-01 -9.23306286e-01
7.84477115e-01 -7.50620782e-01 -6.40223205e-01 3.03777754e-01
-9.99047697e-01 -1.70846865e-01 2.54443079e-01 -4.14213389e-01
-1.07700574e+00 -6.87261522e-01 2.56613106e-01 7.34639317e-02
-4.88534838e-01 5.85842669e-01 -4.89063621e-01 -5.05842924e-01
6.48768365e-01 -2.86568940e-01 2.27297112e-01 -3.88634533e-01
2.44707167e-01 7.84727633e-01 2.02932745e-01 -9.94529366e-01
9.82379079e-01 5.03193974e-01 -2.38182232e-01 -1.00733733e+00
-3.39616865e-01 9.34835374e-02 2.78945849e-03 -7.20968172e-02
5.06447732e-01 -3.37463498e-01 -4.95640457e-01 1.67865261e-01
-1.12673461e+00 2.70514041e-01 1.97657302e-01 4.31928217e-01
1.09169520e-02 5.97127259e-01 -2.43434921e-01 -1.06571889e+00
-2.42845923e-01 -1.33217478e+00 1.12989672e-01 3.19470793e-01
-7.42495358e-01 -7.18678832e-01 -2.03943536e-01 8.07015181e-01
5.76233625e-01 4.71133798e-01 9.76779759e-01 -1.06277514e+00
-5.56147695e-01 -3.48832816e-01 8.74815285e-02 1.43049270e-01
1.95179954e-01 3.75291854e-01 -1.05732250e+00 -9.78946909e-02
6.24022007e-01 -1.75302744e-01 1.96680695e-01 -3.37600410e-02
8.11785817e-01 -8.53017807e-01 -3.70855927e-01 1.49498656e-01
1.24767232e+00 1.03218627e+00 6.73422456e-01 1.16544569e+00
4.01881844e-01 1.14002192e+00 3.61123353e-01 1.62650183e-01
-5.51631786e-02 1.05369903e-01 9.59620625e-02 3.20813477e-01
7.56685734e-01 1.31155819e-01 2.93896616e-01 1.73267201e-01
-1.31482437e-01 1.75436621e-03 -1.33523202e+00 4.53812838e-01
-1.49339283e+00 -9.58453596e-01 -7.30647566e-03 2.23811173e+00
7.12179780e-01 8.05144012e-01 2.20765978e-01 2.25451306e-01
7.35292196e-01 1.83733940e-01 -5.47841549e-01 -1.00148916e+00
6.68713972e-02 3.37290429e-02 7.32806861e-01 2.95997620e-01
-8.65632296e-01 5.58376729e-01 6.89696312e+00 5.02390802e-01
-1.13518143e+00 -1.38404638e-01 8.14467430e-01 2.55904615e-01
-6.98955059e-01 5.87241650e-01 -4.76610005e-01 3.84225130e-01
1.27623332e+00 -1.00178337e+00 4.36906330e-02 6.59724116e-01
-4.38265204e-02 -1.99923307e-01 -5.45098126e-01 6.83258381e-03
-4.39935587e-02 -1.23721433e+00 -1.92137703e-01 5.10214686e-01
1.71296403e-01 -2.96613574e-01 1.33221045e-01 -2.51765728e-01
6.02928817e-01 -1.12734222e+00 8.07713211e-01 -3.28590497e-02
1.34143367e-01 -1.06870341e+00 5.56570470e-01 4.39253062e-01
-3.20086539e-01 -9.06519145e-02 -1.51569229e-02 -6.25331461e-01
-1.38177380e-01 1.44800723e-01 -5.14531612e-01 2.51316786e-01
6.60075665e-01 -1.04308620e-01 -4.80399489e-01 5.84977567e-01
-7.43606612e-02 1.07839715e+00 -2.11198613e-01 -3.56415696e-02
5.62604308e-01 -3.00130934e-01 8.76781404e-01 9.15826857e-01
-7.25954399e-02 3.89067262e-01 -3.31845820e-01 8.80366206e-01
5.41197777e-01 -2.27867723e-01 -9.68568742e-01 -1.08610260e+00
7.54023433e-01 9.89435494e-01 -1.10164988e+00 1.86572000e-01
-6.07517183e-01 1.52805194e-01 -3.61684382e-01 1.80346489e-01
-3.23610038e-01 -3.74090612e-01 8.42147052e-01 4.15396810e-01
-9.11392458e-03 -2.78337806e-01 -9.02490616e-01 -6.76236331e-01
-1.20408051e-01 -1.59271848e+00 2.67359018e-01 -1.93652764e-01
-1.00867105e+00 2.20985919e-01 2.31386367e-02 -4.01644170e-01
4.45561223e-02 -5.61010897e-01 -5.78361213e-01 9.19718206e-01
-9.96394455e-01 -1.13281667e+00 6.65314794e-01 -2.47148678e-01
-1.82772726e-01 -7.52459988e-02 6.77114904e-01 -5.29743254e-01
-4.78122085e-01 4.05985296e-01 -2.03898892e-01 6.16044588e-02
4.94073361e-01 -6.74693346e-01 6.80822432e-01 1.10061717e+00
7.03070452e-03 1.40136278e+00 9.80787933e-01 -9.69662130e-01
-1.81371093e+00 -3.70770544e-01 7.44576991e-01 -9.53730047e-01
1.14707506e+00 -7.63349012e-02 -1.19089937e+00 7.00394928e-01
4.63239312e-01 -8.78243268e-01 7.77044833e-01 1.82422265e-01
-6.95660949e-01 4.21282053e-01 -1.18586445e+00 9.23109710e-01
5.26913404e-01 -5.23084164e-01 -7.79292762e-01 -2.11382836e-01
6.19273603e-01 2.15162367e-01 -6.50553524e-01 2.64723301e-01
7.37266898e-01 -7.82146633e-01 8.37739468e-01 -7.64175832e-01
-2.35532667e-03 -2.46118784e-01 2.96247918e-02 -5.85347712e-01
-1.17847420e-01 -8.39520454e-01 4.58037287e-01 1.38113129e+00
4.68436062e-01 -1.19228625e+00 5.69306433e-01 1.47327971e+00
2.75941670e-01 -5.60375929e-01 -8.19157541e-01 -4.82044041e-01
7.00975835e-01 -4.80091333e-01 5.75355589e-01 1.54587412e+00
3.34004134e-01 1.14488877e-01 1.16507262e-01 5.49654543e-01
7.59516060e-01 -2.46035099e-01 6.86169446e-01 -1.34084761e+00
-4.88394918e-03 -7.69979894e-01 -6.33580923e-01 1.07132010e-01
1.10668413e-01 -5.05537808e-01 -4.32609677e-01 -1.07772911e+00
-4.03081365e-02 -2.39755899e-01 -1.32035837e-01 3.67067367e-01
-1.74291134e-02 -4.86679167e-01 5.18627346e-01 3.42050225e-01
2.96380937e-01 -2.74384439e-01 7.75563359e-01 1.55454800e-01
3.88352051e-02 -9.78527069e-02 -1.83183026e+00 1.04742551e+00
1.03677094e+00 -5.87494731e-01 -4.75075245e-01 -5.51290391e-03
4.37217593e-01 -3.21379125e-01 3.54983598e-01 -7.70918369e-01
4.07728851e-01 -7.78150499e-01 9.54667330e-02 4.37247567e-02
-3.06158632e-01 -1.06758738e+00 5.11655629e-01 5.95827997e-01
-3.57201010e-01 1.41678363e-01 5.38122058e-01 2.94580609e-01
3.12419415e-01 -4.80425328e-01 5.93266547e-01 -6.55517876e-02
-8.78654718e-02 -2.90171027e-01 -8.94536555e-01 6.26884401e-02
1.24044597e+00 -4.42603081e-01 -6.76376343e-01 -2.23854974e-01
-1.94201857e-01 5.86605556e-02 1.04276943e+00 5.41248977e-01
9.87853408e-02 -5.69356740e-01 -4.71720815e-01 -4.87276167e-02
-1.70720533e-01 -5.42649090e-01 -1.99741974e-01 5.51642299e-01
-4.65479970e-01 3.94485384e-01 -1.58318594e-01 2.86242366e-01
-1.19616163e+00 7.92645514e-01 2.12400243e-01 3.67297679e-01
-7.82652140e-01 6.11308873e-01 2.92787194e-01 2.42438555e-01
2.13929608e-01 3.06690961e-01 -1.86416790e-01 6.62353402e-03
9.04179871e-01 5.15079081e-01 -3.42784196e-01 -6.91890121e-01
-5.74461401e-01 5.73606528e-02 -5.81328869e-01 -5.03895223e-01
1.13478017e+00 1.81669723e-02 -4.34708983e-01 4.85109895e-01
5.70601940e-01 4.04684156e-01 -7.11708844e-01 3.19553941e-01
3.81663442e-01 -7.18262374e-01 1.86919615e-01 -1.01038241e+00
-7.13348806e-01 5.28654158e-01 5.25597855e-02 7.33428776e-01
7.21194923e-01 -1.68958396e-01 2.63794124e-01 1.67212695e-01
5.10696888e-01 -9.77184176e-01 -4.47434872e-01 -2.09599063e-01
8.39210629e-01 -5.19692004e-01 2.64451414e-01 -2.88585484e-01
-6.84198856e-01 1.02561116e+00 5.16241848e-01 2.21426100e-01
7.45281219e-01 7.67545819e-01 5.26525080e-01 -3.24876279e-01
-5.38593233e-01 4.03342724e-01 -3.07187080e-01 6.44186914e-01
5.59171915e-01 -4.71933652e-03 -1.09555793e+00 6.28740966e-01
-1.06974542e-01 8.96377116e-02 9.71348524e-01 1.45700967e+00
-4.95657742e-01 -1.31164682e+00 -1.04125082e+00 1.01446860e-01
-1.25303662e+00 4.89225164e-02 -1.23058343e+00 1.01266158e+00
2.76036292e-01 1.12482297e+00 -3.95057127e-02 -2.43500993e-01
-1.19563423e-01 1.88161731e-01 -2.24244684e-01 -4.39204425e-01
-8.32416654e-01 2.17366233e-01 4.77093399e-01 -8.14280212e-02
-3.62016588e-01 -9.18448627e-01 -9.20988321e-01 -5.22963047e-01
-4.43258464e-01 6.44232571e-01 9.18817401e-01 7.98664570e-01
4.83303934e-01 3.09839826e-02 2.63521761e-01 9.92590562e-03
-6.71872139e-01 -3.71185541e-01 -6.00919425e-01 -2.30714574e-01
8.39606524e-02 -4.60616827e-01 -7.51814961e-01 -2.17348516e-01] | [8.945463180541992, 6.536774158477783] |
6eacc117-d362-4a0e-99d5-11f83397f2ae | diffusion-model-augmented-behavioral-cloning | 2302.13335 | null | https://arxiv.org/abs/2302.13335v2 | https://arxiv.org/pdf/2302.13335v2.pdf | Diffusion Model-Augmented Behavioral Cloning | Imitation learning addresses the challenge of learning by observing an expert's demonstrations without access to reward signals from environments. Most existing imitation learning methods that do not require interacting with environments either model the expert distribution as the conditional probability p(a|s) (e.g., behavioral cloning, BC) or the joint probability p(s, a) (e.g., implicit behavioral cloning). Despite its simplicity, modeling the conditional probability with BC usually struggles with generalization. While modeling the joint probability can lead to improved generalization performance, the inference procedure can be time-consuming and it often suffers from manifold overfitting. This work proposes an imitation learning framework that benefits from modeling both the conditional and joint probability of the expert distribution. Our proposed diffusion model-augmented behavioral cloning (DBC) employs a diffusion model trained to model expert behaviors and learns a policy to optimize both the BC loss (conditional) and our proposed diffusion model loss (joint). DBC outperforms baselines in various continuous control tasks in navigation, robot arm manipulation, dexterous manipulation, and locomotion. We design additional experiments to verify the limitations of modeling either the conditional probability or the joint probability of the expert distribution as well as compare different generative models. | ['Chun-Mao Lai', 'Ming-Hao Hsu', 'Shao-Hua Sun', 'Shang-Fu Chen', 'Hsiang-Chun Wang'] | 2023-02-26 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [ 4.19136584e-02 1.07976101e-01 -2.23889694e-01 -1.24077894e-01
-3.74938488e-01 -4.70090389e-01 7.13277459e-01 -2.59708971e-01
-6.36808455e-01 7.91496813e-01 -1.13522455e-01 -3.21244299e-01
-6.51299953e-03 -5.64039826e-01 -1.23903346e+00 -8.27258408e-01
-2.17712224e-01 4.00384814e-01 1.81244299e-01 4.17991653e-02
2.17741221e-01 4.11445439e-01 -1.45356476e+00 -2.38101840e-01
1.30519283e+00 6.74442768e-01 6.49151623e-01 7.16974378e-01
3.51075143e-01 9.91031468e-01 -8.18692565e-01 -8.92345756e-02
3.68693501e-01 -6.28249049e-01 -4.67293382e-01 -2.66875237e-01
1.01013213e-01 -7.90851772e-01 -6.55242920e-01 1.17858231e+00
2.42162749e-01 4.28703427e-01 1.07350290e+00 -1.68458629e+00
-7.50567853e-01 5.55943370e-01 -3.65725696e-01 -1.38425499e-01
1.48393705e-01 6.29855216e-01 9.17715669e-01 -3.20901483e-01
5.04007041e-01 1.39155746e+00 5.62723100e-01 5.72059453e-01
-1.33307207e+00 -6.14091516e-01 4.19743270e-01 3.26953620e-01
-1.12420225e+00 1.18514255e-01 4.93273258e-01 -4.46993858e-01
1.04739475e+00 -4.01262075e-01 5.68536401e-01 1.69895387e+00
6.37048483e-01 1.23577905e+00 1.15929031e+00 -2.32864376e-02
5.01376212e-01 -1.04114927e-01 -2.09179014e-01 6.68823183e-01
-9.34581757e-02 6.14433825e-01 -3.06509048e-01 -1.56977415e-01
1.01794267e+00 1.83956757e-01 -2.67279804e-01 -7.52781391e-01
-1.12454367e+00 6.41110957e-01 5.92899919e-01 -1.40065178e-01
-6.32073164e-01 6.48443937e-01 1.39320642e-01 5.62869966e-01
-2.89562374e-01 4.76950437e-01 -4.84328091e-01 -5.47358155e-01
-3.77951235e-01 5.65342486e-01 9.86742556e-01 1.26961112e+00
6.28318071e-01 1.99948236e-01 -2.26339057e-01 6.59553468e-01
3.00489306e-01 6.97138846e-01 5.53127944e-01 -1.30071509e+00
4.04402286e-01 3.50283325e-01 4.63784456e-01 -5.02345324e-01
-3.03886354e-01 -4.09563750e-01 -3.60101193e-01 7.29491830e-01
5.31120777e-01 -3.92969102e-01 -1.08753002e+00 2.33437991e+00
1.38492927e-01 2.69623160e-01 1.04983412e-01 8.75393867e-01
1.81453496e-01 4.52961683e-01 1.89230889e-01 4.07017648e-01
7.36759365e-01 -1.25198281e+00 -4.89377916e-01 -1.16834760e-01
5.44063985e-01 -3.25367391e-01 1.29259777e+00 3.84850264e-01
-7.90612161e-01 -3.85515332e-01 -9.59925890e-01 1.45178095e-01
-1.66994274e-01 3.71778756e-01 3.60268056e-01 1.73891008e-01
-8.28184605e-01 1.02693713e+00 -1.34787405e+00 -4.22705591e-01
2.34893531e-01 4.55594480e-01 -2.39605650e-01 8.54416657e-03
-8.87835383e-01 1.23424613e+00 1.18273839e-01 -2.83414215e-01
-1.55524290e+00 -4.75231498e-01 -8.31818759e-01 2.89536357e-01
4.57133532e-01 -7.78777063e-01 1.48057842e+00 -7.99725711e-01
-2.01419377e+00 7.63456523e-02 1.35550186e-01 -7.17573524e-01
7.72425890e-01 -7.39179432e-01 3.51251900e-01 5.60222082e-02
1.13853067e-01 7.90515184e-01 1.12292993e+00 -1.10175443e+00
-5.74581802e-01 -9.71035734e-02 2.76639253e-01 4.14631039e-01
-1.28871519e-02 -6.77476645e-01 -1.61860093e-01 -6.55883551e-01
-2.16834113e-01 -1.23014247e+00 -1.55145392e-01 2.98101634e-01
-2.68178970e-01 -3.08468074e-01 9.19510841e-01 -6.79569066e-01
7.36981809e-01 -2.16635227e+00 7.03967035e-01 -7.04988390e-02
-6.49116635e-02 3.22755203e-02 -2.77477175e-01 4.98613536e-01
4.23893481e-01 -3.55762899e-01 -3.98668259e-01 -3.64667058e-01
4.20173705e-01 5.61057627e-01 -3.55211794e-01 4.06264842e-01
2.44541436e-01 8.77546728e-01 -1.15067065e+00 -1.11691229e-01
1.84658542e-01 4.73110944e-01 -9.29269433e-01 3.61386120e-01
-4.33438480e-01 7.74919212e-01 -4.34604734e-01 4.90555853e-01
2.00058877e-01 -2.35963792e-01 1.65836468e-01 2.24736392e-01
1.53364077e-01 1.96063429e-01 -8.94048452e-01 1.58993661e+00
-6.13848746e-01 4.31700200e-01 2.39689901e-01 -1.07332754e+00
6.16729617e-01 1.25739321e-01 3.29847038e-01 -4.99328315e-01
1.87534057e-02 3.30779910e-01 4.08656657e-01 -5.41848540e-01
3.09940308e-01 -1.38766974e-01 -2.60110777e-02 5.34249902e-01
3.44955862e-01 -2.04484075e-01 -3.22812557e-01 2.64463514e-01
1.44439423e+00 9.84738648e-01 7.54932910e-02 -7.51584442e-03
-1.96111985e-02 -7.51017034e-02 6.10021472e-01 1.12840676e+00
-4.90811348e-01 2.72284657e-01 6.19430959e-01 3.04054111e-01
-1.15635324e+00 -1.40491986e+00 2.05734491e-01 9.95048523e-01
4.42286938e-01 -1.17205396e-01 -5.62638164e-01 -8.02018166e-01
4.52002317e-01 1.03627360e+00 -7.33074486e-01 -7.23272800e-01
-5.47889650e-01 -2.39584059e-01 5.61741710e-01 7.93413818e-01
4.61953431e-01 -1.03795326e+00 -8.53886724e-01 2.25137353e-01
-2.79222559e-02 -7.95073092e-01 -4.66594130e-01 4.27694142e-01
-7.69161165e-01 -9.28204477e-01 -8.25726569e-01 -5.93473256e-01
5.92399061e-01 -1.67162716e-02 6.54807210e-01 -1.58204958e-01
-2.52465993e-01 9.64003861e-01 -4.12144661e-01 -3.62958372e-01
-4.07641411e-01 -8.92998427e-02 2.29859129e-01 -4.36407357e-01
4.26185690e-02 -7.77488887e-01 -7.14039326e-01 3.63342404e-01
-6.96224749e-01 -2.72186637e-01 9.82245028e-01 1.33247650e+00
3.15687209e-01 -1.11906216e-01 5.37982047e-01 -3.14897358e-01
6.06509745e-01 -6.69551790e-01 -5.56950152e-01 1.23923413e-01
-5.28275132e-01 3.37066412e-01 7.59124637e-01 -1.00336242e+00
-1.07884622e+00 -1.27865970e-01 6.43581823e-02 -8.46559644e-01
-2.07057685e-01 3.65648329e-01 1.62682161e-01 2.40805119e-01
3.92694414e-01 3.91330808e-01 4.44558918e-01 -4.94980097e-01
3.03453922e-01 3.21856201e-01 4.51142609e-01 -1.00442564e+00
5.63748181e-01 2.29685917e-01 8.77686143e-02 -5.80572248e-01
-2.85134882e-01 -1.52522832e-01 -4.36563849e-01 -2.37020656e-01
7.23326623e-01 -8.82782102e-01 -8.86686385e-01 6.06640041e-01
-9.13230360e-01 -1.05921292e+00 -3.20080698e-01 9.18892801e-01
-1.12172782e+00 4.29432213e-01 -7.14395404e-01 -9.89487886e-01
2.47412458e-01 -1.25304878e+00 1.05596375e+00 1.93029746e-01
-3.52240682e-01 -9.01566923e-01 -1.58436984e-01 -1.30483955e-01
5.82558155e-01 1.98171943e-01 1.05654204e+00 -5.39333165e-01
-7.34683454e-01 -3.16338420e-01 6.21287599e-02 4.45729792e-01
8.34257305e-02 -2.74418026e-01 -4.84179586e-01 -4.42372203e-01
-6.23748451e-02 -6.87068701e-01 7.88715422e-01 3.23030382e-01
9.42779779e-01 -3.49554509e-01 -3.89036834e-01 2.38839492e-01
1.02675247e+00 3.83473694e-01 4.68911976e-01 2.79784054e-01
5.63504934e-01 4.51498181e-01 7.81132340e-01 6.13145769e-01
4.08028960e-01 7.22911477e-01 6.40528142e-01 5.16159892e-01
-4.00013961e-02 -5.68115592e-01 8.73550415e-01 3.72661024e-01
1.56485271e-02 -1.44367903e-01 -6.18645847e-01 4.19687748e-01
-1.98514676e+00 -8.50090325e-01 4.98516470e-01 2.09309697e+00
7.69081414e-01 6.31097034e-02 2.20082849e-01 -3.95698518e-01
3.46537411e-01 -3.27812582e-01 -1.25425637e+00 -2.00027898e-01
9.22884718e-02 -1.29280472e-03 3.76492739e-01 6.64825857e-01
-7.84897327e-01 7.54470587e-01 5.83491373e+00 6.33653104e-01
-1.06621730e+00 7.51276020e-05 -6.94154277e-02 -2.40400270e-01
7.02918395e-02 1.34228513e-01 -5.83717644e-01 6.19286001e-01
5.01829863e-01 3.79519649e-02 8.15393984e-01 1.17376781e+00
-1.33286193e-01 -3.90429586e-01 -1.46801305e+00 7.31040716e-01
-1.41815946e-01 -4.69194770e-01 -5.83509505e-02 2.08366275e-01
5.38740098e-01 2.55831420e-01 3.26044500e-01 9.56549168e-01
8.52852285e-01 -8.62128139e-01 9.76274967e-01 6.17859960e-01
2.33098626e-01 -3.88787955e-01 5.22230089e-01 7.94152617e-01
-7.28089750e-01 -4.18723643e-01 -2.87930489e-01 -1.36416569e-01
2.80351918e-02 -3.18567693e-01 -6.25756323e-01 3.23808193e-01
6.72119677e-01 8.73923600e-01 -1.02409318e-01 9.08366919e-01
-5.09638369e-01 5.92244387e-01 -4.54494178e-01 -3.85169238e-01
4.48362917e-01 -4.59495634e-01 9.89773571e-01 6.53673708e-01
3.93593132e-01 -1.93483263e-01 2.46926203e-01 1.37039793e+00
9.51485112e-02 -2.88414150e-01 -6.27461672e-01 -2.93143362e-01
5.52399993e-01 6.02617681e-01 -3.34433913e-01 -3.10754240e-01
-3.35143626e-01 1.35482907e+00 6.07123911e-01 7.96997368e-01
-1.03346658e+00 -5.03389180e-01 6.68012559e-01 -3.26075345e-01
6.16854429e-01 -4.54588979e-01 -1.03286642e-03 -1.00989187e+00
1.04184665e-01 -8.36271346e-01 -7.05103129e-02 -5.87231457e-01
-1.38028765e+00 1.48089662e-01 1.65378615e-01 -1.31744218e+00
-3.00512344e-01 -7.65550673e-01 -7.25845277e-01 8.19406509e-01
-1.33102965e+00 -6.71208620e-01 -8.48613959e-03 3.70436639e-01
4.67162222e-01 -1.13005280e-01 5.11188924e-01 7.82538950e-02
-5.58451951e-01 6.10592186e-01 3.16433609e-01 -1.71260566e-01
7.21210837e-01 -1.52539277e+00 2.64869239e-02 2.96759129e-01
-3.86921495e-01 8.96375835e-01 9.89875853e-01 -6.99111342e-01
-1.40338802e+00 -7.69463301e-01 7.62378573e-02 -5.02183318e-01
7.00832129e-01 -1.12696409e-01 -9.11397099e-01 8.59398305e-01
1.33058196e-02 -4.54671502e-01 2.03275099e-01 -1.52318910e-01
-2.10538730e-01 4.05651629e-01 -1.14817607e+00 9.95597303e-01
9.83357668e-01 -4.47916299e-01 -6.22616291e-01 8.60181823e-02
6.10732198e-01 -3.79111230e-01 -8.58947814e-01 3.62025231e-01
8.67065370e-01 -7.28011608e-01 7.95751333e-01 -8.02218974e-01
5.34913659e-01 -1.24514520e-01 -1.18585542e-01 -1.61810982e+00
-9.34540406e-02 -4.08545166e-01 -3.39592785e-01 7.54298210e-01
3.74806702e-01 -8.86420786e-01 5.83000839e-01 7.23195672e-01
-2.05006287e-01 -8.67942274e-01 -1.07696378e+00 -1.41805947e+00
5.24056137e-01 -1.14781680e-02 3.87904912e-01 6.53227150e-01
1.74575374e-01 1.38964877e-01 -5.73684096e-01 1.40867218e-01
6.02869272e-01 1.04013510e-01 1.10272825e+00 -9.20327306e-01
-7.69865632e-01 -6.85477972e-01 -2.80750185e-01 -1.65210891e+00
4.22793001e-01 -7.39406466e-01 5.32596886e-01 -1.48494327e+00
6.17799610e-02 -3.83426309e-01 -9.64570716e-02 3.40603560e-01
-2.75406390e-01 -6.51715577e-01 1.82713926e-01 2.97718763e-01
-5.36010921e-01 1.28324294e+00 1.47514272e+00 -5.89728095e-02
-1.88569307e-01 -6.85672686e-02 -1.22987321e-02 6.55400038e-01
8.48789811e-01 -4.74450052e-01 -5.37758589e-01 -3.25235307e-01
-1.53852493e-01 7.59396926e-02 5.58417499e-01 -9.66882467e-01
2.86228895e-01 -2.06833482e-01 1.00579366e-01 -2.55982488e-01
6.17374659e-01 -8.18679392e-01 -2.28896067e-01 8.75728667e-01
-4.31873918e-01 -1.02764539e-01 8.94707814e-02 1.31360030e+00
3.69894169e-02 -9.46163088e-02 7.98081934e-01 -6.54417723e-02
-5.00266016e-01 1.47231882e-02 -7.89232552e-01 -1.02059402e-01
1.01621532e+00 -9.24994424e-02 -2.39213362e-01 -7.04849601e-01
-7.76901305e-01 5.96795917e-01 8.11356068e-01 5.25517762e-01
5.81584454e-01 -1.09558833e+00 -2.04675481e-01 1.85874440e-02
-1.53389554e-02 -9.93004441e-02 1.22484632e-01 1.11050022e+00
-2.52346158e-01 1.04838565e-01 -3.38330030e-01 -5.95878243e-01
-7.90644407e-01 5.47943771e-01 4.15518343e-01 -1.35832101e-01
-8.03750515e-01 8.57469141e-01 3.67324591e-01 -8.10965180e-01
6.58307850e-01 -4.42303896e-01 1.49667442e-01 -6.24815583e-01
-1.40272483e-01 5.21131456e-01 -8.93440187e-01 -3.19137216e-01
-6.39820695e-02 2.11812764e-01 -1.69668242e-01 -2.97848046e-01
1.11468339e+00 -2.07158625e-02 1.96555644e-01 6.35090351e-01
9.35674846e-01 -5.37922621e-01 -2.01826835e+00 -3.28185350e-01
-1.51132539e-01 -3.34118932e-01 -2.88631946e-01 -9.97102559e-01
-6.34571016e-01 9.66600835e-01 4.29274052e-01 -3.74563575e-01
5.25356233e-01 -1.30546600e-01 7.49591291e-01 8.13418329e-01
7.50570774e-01 -1.30850780e+00 5.27083814e-01 7.36863852e-01
9.98175263e-01 -1.30717647e+00 -2.31149077e-01 2.99308505e-02
-9.01715398e-01 7.50385046e-01 1.01458395e+00 -5.11888981e-01
7.33292639e-01 1.38836458e-01 -9.28509608e-02 8.67298841e-02
-8.53131592e-01 -5.62866665e-02 -1.15933586e-02 6.75810695e-01
-3.52114402e-02 1.73206612e-01 -7.72809014e-02 5.60136259e-01
-4.31250222e-02 -8.94663110e-02 3.10618073e-01 1.47762251e+00
-3.55124712e-01 -1.01404655e+00 -1.03238232e-01 5.37640393e-01
-1.21421918e-01 1.80079386e-01 -4.26348060e-01 7.69859850e-01
-2.73627192e-01 6.86799943e-01 -1.11505069e-01 -4.15440321e-01
3.46847862e-01 1.13771692e-01 7.84319997e-01 -5.60891628e-01
-1.99162081e-01 -4.42790650e-02 -4.19168353e-01 -6.24413192e-01
-1.30635887e-01 -8.23931396e-01 -1.33196807e+00 -3.67186368e-02
-3.35695684e-01 -4.24076617e-02 6.49791420e-01 9.13199961e-01
3.44767094e-01 5.92460513e-01 2.72810280e-01 -9.57055151e-01
-1.40134418e+00 -1.19570446e+00 -7.81270504e-01 4.99541134e-01
5.46800494e-01 -1.34334981e+00 -5.42246521e-01 -1.96929023e-01] | [4.359593868255615, 1.2818074226379395] |
4e8dd7ec-6f62-4cac-8331-d61358b64895 | identifying-cover-songs-using-information | 1407.2433 | null | http://arxiv.org/abs/1407.2433v3 | http://arxiv.org/pdf/1407.2433v3.pdf | Identifying Cover Songs Using Information-Theoretic Measures of Similarity | This paper investigates methods for quantifying similarity between audio
signals, specifically for the task of of cover song detection. We consider an
information-theoretic approach, where we compute pairwise measures of
predictability between time series. We compare discrete-valued approaches
operating on quantised audio features, to continuous-valued approaches. In the
discrete case, we propose a method for computing the normalised compression
distance, where we account for correlation between time series. In the
continuous case, we propose to compute information-based measures of similarity
as statistics of the prediction error between time series. We evaluate our
methods on two cover song identification tasks using a data set comprised of
300 Jazz standards and using the Million Song Dataset. For both datasets, we
observe that continuous-valued approaches outperform discrete-valued
approaches. We consider approaches to estimating the normalised compression
distance (NCD) based on string compression and prediction, where we observe
that our proposed normalised compression distance with alignment (NCDA)
improves average performance over NCD, for sequential compression algorithms.
Finally, we demonstrate that continuous-valued distances may be combined to
improve performance with respect to baseline approaches. Using a large-scale
filter-and-refine approach, we demonstrate state-of-the-art performance for
cover song identification using the Million Song Dataset. | ['Simon Dixon', 'Peter Foster', 'Anssi Klapuri'] | 2014-07-09 | null | null | null | null | ['cover-song-identification'] | ['music'] | [ 7.70446360e-01 -4.35659170e-01 1.23501025e-01 -2.34992504e-01
-1.52560842e+00 -9.02545333e-01 5.95334172e-01 6.22610986e-01
-4.79622036e-01 2.31966510e-01 3.55006248e-01 1.11404369e-02
-6.97695315e-01 -6.15090191e-01 -4.38332379e-01 -5.37200332e-01
-8.87180686e-01 1.10525765e-01 1.65058270e-01 -5.20264320e-02
5.74214697e-01 2.04585269e-01 -1.98702645e+00 3.20195556e-01
2.99246818e-01 1.45800936e+00 -1.87444896e-01 1.04522848e+00
3.11153233e-01 3.67730647e-01 -8.56304526e-01 -2.79336542e-01
4.80925888e-01 -5.75375736e-01 -6.94024444e-01 -1.22286759e-01
6.00999415e-01 -1.72458012e-02 5.68007194e-02 9.37132239e-01
5.16851783e-01 3.19170319e-02 6.97154403e-01 -1.15611148e+00
5.63957430e-02 8.63016665e-01 -2.23056674e-01 4.58274156e-01
7.86497533e-01 -2.40694553e-01 1.56542230e+00 -5.83438694e-01
3.21011901e-01 9.50257301e-01 1.07985389e+00 -2.30248272e-01
-1.51938283e+00 -4.33367252e-01 -4.56316918e-01 2.00948417e-01
-1.66619337e+00 -5.25412738e-01 5.88708103e-01 -5.10744393e-01
8.41599882e-01 7.38995314e-01 5.75224698e-01 4.01737809e-01
-1.45617530e-01 5.09156048e-01 9.20377493e-01 -7.32109487e-01
2.78070241e-01 -3.28404874e-01 -2.25058477e-02 1.46861643e-01
-4.04996611e-02 4.69998389e-01 -6.69015348e-01 -6.14684701e-01
1.85793176e-01 -2.17105433e-01 -3.11047107e-01 -4.85749468e-02
-1.43833852e+00 7.61928558e-01 -1.86421290e-01 5.29061019e-01
-2.52665848e-01 1.23567604e-01 6.36861861e-01 8.47828090e-01
5.76917827e-01 7.90542185e-01 -2.23227441e-01 -6.28182888e-01
-1.52441406e+00 5.73747158e-01 1.08443582e+00 7.99691021e-01
4.93321627e-01 -1.28707722e-01 -3.66491467e-01 8.55811119e-01
-1.02407455e-01 2.30465531e-01 5.38270533e-01 -1.07360244e+00
3.75606775e-01 3.34861763e-02 -1.65289789e-01 -1.06736708e+00
-2.07138509e-01 -5.82316458e-01 -5.23392618e-01 -1.63075402e-01
3.65723342e-01 3.34368676e-01 -2.31498361e-01 1.65498841e+00
-1.28210867e-02 4.66896683e-01 4.05888483e-02 5.31587899e-01
1.98323414e-01 4.39860255e-01 -5.59363961e-01 -6.54743493e-01
1.12662137e+00 -3.69017810e-01 -4.40213025e-01 6.00820243e-01
6.24595761e-01 -1.05426633e+00 9.01460767e-01 9.32154775e-01
-1.15487647e+00 -5.45565486e-01 -1.23066473e+00 5.59551060e-01
-3.28795820e-01 -1.31828815e-01 -2.95346323e-02 9.19980526e-01
-9.46157157e-01 1.28823793e+00 -4.71971333e-01 -9.52035040e-02
1.01276475e-03 2.53715724e-01 -9.44594890e-02 4.30429369e-01
-1.19155180e+00 2.39956066e-01 4.76740032e-01 -5.31100035e-01
-7.03963101e-01 -7.88660169e-01 -5.41480303e-01 1.25507161e-01
9.15649161e-02 1.11986533e-01 1.40647960e+00 -4.73258346e-01
-1.03975165e+00 7.45483160e-01 3.04843932e-01 -1.06049919e+00
1.91451460e-01 -9.80440155e-02 -7.23214209e-01 4.48480308e-01
-1.22391388e-01 2.98166186e-01 9.55293894e-01 -7.33347595e-01
-6.99262559e-01 1.24399327e-01 -1.37592465e-01 -1.08006299e-01
-5.01821995e-01 2.10710153e-01 3.36391553e-02 -1.19450116e+00
7.86496326e-02 -9.17288244e-01 9.88991335e-02 -1.75513372e-01
-3.03753614e-01 -5.07631861e-02 7.24710882e-01 -6.12809122e-01
1.86447918e+00 -2.44142318e+00 7.64711872e-02 5.64172626e-01
-4.23121545e-03 1.22288465e-01 -2.02426791e-01 6.92901015e-01
-7.98632428e-02 2.15314731e-01 -6.56081498e-01 -2.74563670e-01
1.99055851e-01 -2.56342981e-02 -6.19712174e-01 5.17273009e-01
8.98722485e-02 2.48175383e-01 -7.38758385e-01 -4.30496812e-01
-9.55289528e-02 3.61097544e-01 -7.15935647e-01 3.27579468e-01
-9.10223946e-02 2.61742380e-02 2.01496139e-01 5.08880675e-01
4.11313534e-01 2.94242114e-01 3.09290346e-02 -1.31592840e-01
-1.71777338e-01 6.32489145e-01 -1.43064606e+00 1.69412363e+00
-3.46215695e-01 8.97555232e-01 -1.74402952e-01 -1.07532763e+00
1.12616909e+00 5.11758208e-01 7.10355699e-01 -3.39870512e-01
-3.46809849e-02 4.27332252e-01 6.72762468e-02 1.82437599e-02
5.67153335e-01 -3.34793590e-02 -2.83799112e-01 7.13774443e-01
9.30242687e-02 -5.22070885e-01 3.65717113e-01 6.85540810e-02
1.36682117e+00 -3.74041557e-01 5.34303963e-01 -2.72872329e-01
5.88736951e-01 -4.12881106e-01 4.67418227e-04 7.04228759e-01
6.93084523e-02 8.17738950e-01 4.88513410e-01 3.65920551e-02
-1.07104063e+00 -1.06525111e+00 -4.41187590e-01 1.10362351e+00
-2.29587033e-01 -1.10544801e+00 -7.02417195e-01 -1.97425216e-01
2.15390727e-01 4.26991731e-01 -2.97136784e-01 -8.18700492e-02
-4.90795285e-01 -4.30241078e-01 1.05571318e+00 2.21654594e-01
6.65854514e-02 -7.45485306e-01 -6.69550180e-01 3.69888067e-01
-1.00166038e-01 -9.13010657e-01 -6.97779536e-01 4.06141341e-01
-7.94366837e-01 -7.39177644e-01 -4.70011652e-01 -4.61264282e-01
-2.21105784e-01 8.12885687e-02 1.19972646e+00 -4.79559861e-02
-4.25347477e-01 7.48583853e-01 -8.07780147e-01 -3.99163336e-01
-5.00684679e-01 -8.53851289e-02 2.88703889e-01 1.54773081e-02
2.64820546e-01 -1.02216887e+00 -5.14474988e-01 4.28094864e-01
-1.12911093e+00 -6.92021847e-01 2.58706123e-01 7.75762677e-01
8.49961221e-01 1.37451634e-01 5.11806190e-01 -4.21816945e-01
9.75418985e-01 -2.96964288e-01 -3.26223850e-01 1.92120168e-02
-7.60090888e-01 9.37000215e-02 6.76896036e-01 -6.08358681e-01
7.99067318e-02 -1.37225345e-01 -1.31631307e-02 -5.35216212e-01
-6.29202276e-02 5.80871940e-01 2.44528741e-01 -2.55870938e-01
7.35410094e-01 4.35400933e-01 -1.54854208e-01 -6.82048857e-01
3.49049360e-01 1.05273247e+00 8.66985738e-01 -6.29988909e-01
7.01454341e-01 2.24141330e-01 -9.41248331e-03 -8.15945625e-01
-4.49536711e-01 -8.79558623e-01 -7.16623902e-01 -1.21108569e-01
4.70138192e-01 -7.37081945e-01 -6.37358785e-01 7.31033459e-02
-8.26280534e-01 1.48227647e-01 -7.06783295e-01 7.39894092e-01
-1.09293354e+00 5.05540431e-01 -3.66975069e-01 -1.04757011e+00
-3.23900908e-01 -6.13435149e-01 1.33516514e+00 -4.87119496e-01
-5.67626834e-01 -8.43121707e-01 5.94729185e-01 -2.39860177e-01
3.54153454e-01 5.00617445e-01 8.15074921e-01 -1.15683281e+00
-2.23703936e-01 -4.02136713e-01 1.91343337e-01 5.28738678e-01
1.30608389e-02 -2.14605018e-01 -9.84711707e-01 -5.29519379e-01
1.09569103e-01 -2.83401310e-01 9.02342796e-01 1.21168293e-01
1.38289475e+00 -5.39949894e-01 1.68705091e-01 6.32225096e-01
1.31392515e+00 1.85173497e-01 4.47872341e-01 2.12259144e-01
1.28461704e-01 5.39711773e-01 1.05136192e+00 9.91581321e-01
-1.70979381e-01 9.11273837e-01 1.96171984e-01 3.78908992e-01
7.15850070e-02 -1.68874338e-01 4.26551014e-01 1.32776523e+00
-1.87522277e-01 -3.22655439e-02 -7.01219380e-01 6.10751331e-01
-1.48516262e+00 -1.27120292e+00 2.01011628e-01 2.65492630e+00
8.72785151e-01 3.18936050e-01 7.22624481e-01 1.11884713e+00
6.95439339e-01 2.33960509e-01 -1.79052576e-01 -5.63512206e-01
-4.32525091e-02 5.05914330e-01 4.71426636e-01 2.41696462e-01
-1.22954679e+00 5.33333421e-02 6.78405619e+00 1.21346557e+00
-8.12261522e-01 5.64078465e-02 1.56278059e-01 -3.10551345e-01
-2.33784560e-02 -4.21862826e-02 -4.34727579e-01 5.14875710e-01
1.57069826e+00 -4.54854876e-01 6.66085601e-01 6.16335392e-01
-5.06420583e-02 1.67072520e-01 -1.46513009e+00 1.13212657e+00
1.09708853e-01 -1.07647491e+00 -2.30942518e-01 2.65676588e-01
4.56478179e-01 -2.24109709e-01 2.06577912e-01 3.91871296e-02
-3.76749903e-01 -8.70462239e-01 9.99111891e-01 4.23871845e-01
1.02449882e+00 -9.63428915e-01 4.32832956e-01 1.77535266e-01
-1.77381802e+00 3.29824388e-02 -2.75822252e-01 -6.05433583e-02
4.09427471e-02 6.05103910e-01 -8.14416766e-01 6.52895212e-01
5.87950230e-01 9.98811126e-01 -5.06829262e-01 1.42578876e+00
4.51885343e-01 9.53831851e-01 -5.26427090e-01 4.58067805e-02
-3.94754969e-02 -8.60345885e-02 9.10853386e-01 1.57681894e+00
7.28594065e-01 -1.08303159e-01 1.16357177e-01 5.39244592e-01
2.00764850e-01 2.21586823e-01 -4.90730286e-01 -3.26497495e-01
7.56109595e-01 8.65749359e-01 -5.08107781e-01 -2.11414084e-01
-1.24350980e-01 6.61748052e-01 -2.27242872e-01 -2.22269312e-01
-6.05244815e-01 -9.21520412e-01 6.54874384e-01 2.56176293e-02
7.01837957e-01 -2.32108891e-01 9.21937674e-02 -8.03050816e-01
7.38724470e-02 -1.20230103e+00 6.32782340e-01 -1.99000210e-01
-1.24466574e+00 6.09148204e-01 2.48218074e-01 -2.13739872e+00
-7.78954983e-01 -3.31481695e-01 -5.03961504e-01 5.97976446e-01
-9.73864198e-01 -4.09705698e-01 3.31668258e-01 5.18870354e-01
2.77469784e-01 -2.75762111e-01 1.13328910e+00 4.26451802e-01
2.50523657e-01 8.73360157e-01 4.56784964e-01 -7.70186484e-02
6.30148649e-01 -1.27826977e+00 5.28484821e-01 5.19208372e-01
9.73424971e-01 4.22681302e-01 9.35858607e-01 -2.62500644e-01
-1.26922894e+00 -8.94983411e-01 7.26631999e-01 -2.25135401e-01
8.40028644e-01 -3.94864082e-01 -9.13312554e-01 1.30855963e-01
-2.19980806e-01 -1.74876243e-01 1.21175015e+00 7.30844140e-02
-6.98441207e-01 -1.98702917e-01 -1.20359719e+00 -6.38469234e-02
1.07572412e+00 -1.13319528e+00 -7.35212743e-01 3.85997057e-01
8.47500324e-01 1.71057656e-02 -1.50381696e+00 4.54804957e-01
8.67772341e-01 -1.09423137e+00 1.09179115e+00 -2.25094035e-01
9.49832126e-02 -3.82342070e-01 -7.53490746e-01 -1.08058929e+00
-1.60190880e-01 -1.00707996e+00 -2.80576915e-01 1.29027367e+00
2.83991188e-01 -3.57189417e-01 3.32865655e-01 -2.96062738e-01
1.43166482e-01 -7.41708457e-01 -1.21468115e+00 -1.38587308e+00
-1.31229237e-01 -8.40740383e-01 6.64664328e-01 7.61326313e-01
2.75160521e-01 -9.13475305e-02 -4.65034544e-01 -5.59592098e-02
6.43664718e-01 3.80298585e-01 5.08291662e-01 -1.32011199e+00
-9.72441792e-01 -4.94426787e-01 -1.12697875e+00 -7.86584198e-01
-2.56911039e-01 -8.98730278e-01 -2.69171372e-02 -5.41313767e-01
-1.08289249e-01 -1.47280172e-01 -7.21264362e-01 6.18960448e-02
3.75398755e-01 6.94888890e-01 4.14604425e-01 4.19598460e-01
-6.40612960e-01 2.54313618e-01 3.98581654e-01 -1.63734347e-01
-1.75129756e-01 2.30225071e-01 -3.40179026e-01 4.23913538e-01
6.15973353e-01 -6.63970828e-01 -2.00136408e-01 1.35109037e-01
1.51515871e-01 1.75002128e-01 3.51869434e-01 -1.59178472e+00
2.00241357e-01 3.06695372e-01 -1.13970637e-01 -8.11237335e-01
3.54322493e-01 -6.98464930e-01 2.46877253e-01 5.12350082e-01
-7.82903314e-01 1.60252109e-01 2.52728071e-02 8.33746910e-01
-7.10671008e-01 -2.90223151e-01 5.18476248e-01 3.50119680e-01
-2.47848600e-01 -1.97755173e-02 -1.96743846e-01 4.51230928e-02
6.04878962e-01 -2.75011837e-01 8.36021602e-02 -6.12326086e-01
-7.69069672e-01 -2.91486084e-01 2.52674699e-01 2.17030004e-01
6.58360600e-01 -1.47578669e+00 -8.87637138e-01 2.18080893e-01
2.72592843e-01 -7.96006501e-01 -2.74307877e-01 9.44635928e-01
-2.29112789e-01 4.99879867e-01 5.05925417e-02 -7.71432459e-01
-1.79029298e+00 4.85045314e-01 -4.64386679e-02 -2.09343195e-01
-1.73179626e-01 7.12087154e-01 -4.28394496e-01 -8.83785635e-02
6.16429329e-01 -5.61541498e-01 -6.89383969e-02 4.14765775e-01
7.05085039e-01 5.13323486e-01 3.88688862e-01 -7.19860435e-01
-4.08633411e-01 8.70147347e-01 2.53421724e-01 -5.63713193e-01
1.27853191e+00 -2.88516693e-02 -1.18802516e-02 7.82000303e-01
1.72168493e+00 2.15700954e-01 -6.38132632e-01 -3.66575330e-01
3.20934355e-01 -7.42449462e-01 -1.03729025e-01 -4.15066004e-01
-4.28373635e-01 8.24286163e-01 8.68884623e-01 9.23154473e-01
1.62365627e+00 -7.42330849e-02 6.50768638e-01 4.23334032e-01
4.97485638e-01 -8.18355203e-01 -1.76454559e-01 3.15239161e-01
7.41235971e-01 -8.30576837e-01 2.63302565e-01 -2.36272350e-01
-2.21162185e-01 1.26766932e+00 -4.96959478e-01 -2.94288784e-01
7.11683929e-01 2.37153098e-01 -3.68665189e-01 5.45347333e-02
-8.05321336e-01 -3.98343772e-01 7.43700564e-01 4.58519697e-01
5.34220278e-01 1.24661587e-01 -5.83800673e-01 5.27007818e-01
-9.03801441e-01 -4.84261513e-01 1.97947696e-01 8.79034162e-01
-6.43980920e-01 -1.36230683e+00 -5.13822019e-01 7.25566089e-01
-6.77120566e-01 -2.89437860e-01 -5.77283084e-01 5.12091160e-01
7.44899064e-02 1.09575486e+00 2.24021494e-01 -9.75679517e-01
2.44891718e-01 1.55798963e-03 3.96622837e-01 -4.09837484e-01
-9.22806084e-01 2.51873523e-01 2.15614900e-01 -4.33684736e-01
-6.93330884e-01 -9.41542566e-01 -6.60230994e-01 -1.50845960e-01
-4.29704726e-01 4.71088171e-01 5.71071148e-01 6.75686121e-01
1.92878038e-01 2.12423861e-01 1.16661406e+00 -8.22029471e-01
-9.34853375e-01 -8.60847533e-01 -9.01503265e-01 5.03324866e-01
5.31735480e-01 -2.84790158e-01 -6.17739677e-01 1.44229800e-01] | [15.69459342956543, 5.323807239532471] |
bba1f87b-72e1-4857-86cb-34c15fb137d6 | evaluation-metrics-for-headline-generation | null | null | https://aclanthology.org/2020.lrec-1.222 | https://aclanthology.org/2020.lrec-1.222.pdf | Evaluation Metrics for Headline Generation Using Deep Pre-Trained Embeddings | With the explosive growth in textual data, it is becoming increasingly important to summarize text automatically. Recently, generative language models have shown promise in abstractive text summarization tasks. Since these models rephrase text and thus use similar but different words as found in the summarized text, existing metrics such as ROUGE that use n-gram overlap may not be optimal. Therefore we evaluate two embedding-based evaluation metrics that are applicable to abstractive summarization: Fr �echet embedding distance, which has been introduced recently, and angular embedding similarity, which is our proposed metric. To demonstrate the utility of both metrics, we analyze the headline generation capacity of two state-of-the-art language models: GPT-2 and ULMFiT. In particular, our proposed metric shows close relation with human judgments in our experiments and has overall better correlations with them. To provide reproducibility, the source code plus human assessments of our experiments is available on GitHub. | ['Georg Groh', 'Gerhard Hagerer', 'Yang An', 'Abdul Moeed'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['headline-generation'] | ['natural-language-processing'] | [-1.49477075e-03 1.77772045e-01 -1.92057490e-01 -6.33173212e-02
-8.86213720e-01 -5.66057205e-01 8.77407551e-01 8.06149662e-01
-4.82398510e-01 9.47131455e-01 9.63034749e-01 -1.54474348e-01
-1.54715791e-01 -5.52002370e-01 -1.79488435e-01 -2.28958875e-01
-2.26552468e-02 3.94515216e-01 2.60750085e-01 -3.27451319e-01
6.81330264e-01 2.31947869e-01 -1.28529370e+00 1.97383463e-01
1.40476108e+00 5.23389280e-01 1.17655024e-01 8.75239611e-01
-2.86303967e-01 6.17785752e-01 -9.38335538e-01 -5.44160128e-01
-2.75795460e-01 -7.14637339e-01 -9.29281175e-01 -3.26174229e-01
3.38002086e-01 -1.45879477e-01 -4.88388389e-01 8.74051511e-01
6.16907954e-01 2.47564659e-01 9.58980381e-01 -8.74635279e-01
-8.61634970e-01 8.84141982e-01 -4.14882362e-01 4.45796013e-01
7.19975889e-01 -2.46530816e-01 1.51482153e+00 -8.61417472e-01
6.41589701e-01 1.17080796e+00 5.10160446e-01 2.28651196e-01
-1.15914285e+00 -1.69174984e-01 -9.54051986e-02 1.78001076e-01
-1.14455009e+00 -4.04036343e-01 8.20081949e-01 -1.62251204e-01
1.18490136e+00 6.13353133e-01 6.19343936e-01 9.68288898e-01
3.81935358e-01 1.02748644e+00 7.82034218e-01 -5.98740101e-01
1.72257051e-01 2.90286839e-01 2.28380099e-01 4.28921223e-01
7.90117323e-01 -4.96303290e-01 -5.89638054e-01 -2.92332977e-01
1.47564664e-01 -3.06884736e-01 -4.30448651e-01 -1.18536785e-01
-1.21813214e+00 1.00435698e+00 3.14370096e-01 6.08079314e-01
-4.41305697e-01 1.06497698e-01 5.81904113e-01 2.37418879e-02
8.72397840e-01 9.62218404e-01 8.16211924e-02 -5.39231181e-01
-1.32251978e+00 4.67576653e-01 1.03514183e+00 6.11103654e-01
4.17112142e-01 -1.27518594e-01 -4.94193524e-01 1.08498359e+00
1.37348339e-01 2.62857288e-01 8.08469474e-01 -7.66219676e-01
5.43139338e-01 6.92569494e-01 7.96786770e-02 -1.26098466e+00
-1.99247941e-01 -4.99112248e-01 -7.92701602e-01 -4.29535002e-01
-1.55960724e-01 -1.45320818e-01 -4.94849026e-01 1.33776319e+00
-1.96462646e-01 -3.09862435e-01 3.04543704e-01 5.33319533e-01
1.08712912e+00 8.43751431e-01 -1.67925909e-01 -4.51108813e-01
1.14755416e+00 -1.01846206e+00 -1.02410126e+00 -2.70203084e-01
6.37419581e-01 -9.15331542e-01 1.02287078e+00 6.65814504e-02
-1.13207459e+00 -2.59867996e-01 -1.19701886e+00 -2.20017835e-01
-3.79597515e-01 1.51699752e-01 4.20278102e-01 5.17823935e-01
-1.20296228e+00 7.16929615e-01 -7.20083714e-01 -8.60849261e-01
1.36283860e-01 -1.30233407e-01 -2.46957302e-01 1.68304771e-01
-1.13540542e+00 1.00140381e+00 5.36555588e-01 -4.00503546e-01
-2.57556736e-01 -3.23640317e-01 -6.59004748e-01 3.67994547e-01
1.78326845e-01 -7.96048820e-01 1.20210481e+00 -2.21129268e-01
-1.40093434e+00 5.27594447e-01 -2.47498855e-01 -4.73387659e-01
4.09887612e-01 -3.93756181e-01 -2.97359675e-01 3.63707900e-01
5.25920689e-02 4.92154717e-01 3.68928581e-01 -1.27097404e+00
-5.05371809e-01 -1.45568982e-01 -2.67920271e-02 4.20310080e-01
-8.21812272e-01 -1.58851489e-01 -2.76058495e-01 -7.69293070e-01
-1.22740760e-01 -6.57805622e-01 3.58975008e-02 -4.76050466e-01
-6.86054885e-01 -5.44167042e-01 5.96865475e-01 -8.07381511e-01
1.98052061e+00 -1.84024143e+00 1.27375796e-01 1.74701083e-02
1.47820249e-01 3.59887540e-01 -6.30198047e-02 1.11684358e+00
2.51181513e-01 4.30842638e-01 -2.90056467e-01 -4.03223544e-01
1.25102162e-01 -4.65114489e-02 -3.51281852e-01 -3.46912965e-02
-1.25258472e-02 9.53272879e-01 -1.11653590e+00 -7.76257217e-01
-4.02907953e-02 2.33560577e-01 -4.18597907e-01 6.78404644e-02
-7.46023729e-02 -2.33015716e-01 -4.94117945e-01 1.41003326e-01
1.56470075e-01 -2.73922384e-01 -9.73444991e-03 -1.22820206e-01
-9.80240330e-02 4.56706285e-01 -6.91873848e-01 1.53607464e+00
-3.72993976e-01 8.88285816e-01 -6.32954121e-01 -7.65845835e-01
8.50183249e-01 2.74633735e-01 3.22159916e-01 -4.61986274e-01
-1.23605179e-03 1.68215781e-01 -5.37217855e-02 -3.27293307e-01
1.38352394e+00 2.11283058e-01 -4.31289263e-02 6.68876290e-01
-3.25286426e-02 -4.34029371e-01 7.43881643e-01 8.12402368e-01
1.24355543e+00 -1.89620644e-01 7.33878553e-01 -3.45051080e-01
4.81509864e-01 1.66413218e-01 2.44650105e-03 8.59197974e-01
1.42186478e-01 6.79782867e-01 6.52477205e-01 1.13016993e-01
-1.06413078e+00 -1.00188947e+00 2.53786389e-02 7.43358433e-01
2.35314164e-02 -1.04026997e+00 -7.64626384e-01 -6.12158418e-01
-1.17485947e-03 1.28616297e+00 -5.52416623e-01 -2.82588810e-01
-3.47642392e-01 -6.78902030e-01 5.76236725e-01 5.14480650e-01
2.24632502e-01 -9.24811482e-01 -5.89006484e-01 2.32057974e-01
-4.17664319e-01 -7.31857717e-01 -5.90288401e-01 -1.13553189e-01
-8.42988849e-01 -6.30230963e-01 -9.16393518e-01 -4.85728174e-01
3.76005322e-01 3.94262701e-01 1.27261353e+00 5.10482006e-02
2.92175496e-03 5.68973184e-01 -7.78466463e-01 -4.98074412e-01
-6.96399629e-01 6.60044432e-01 -1.37915574e-02 -4.74291295e-01
1.59433931e-01 -5.47114730e-01 -6.67715669e-01 -1.37936532e-01
-1.25534654e+00 4.72090468e-02 5.83827853e-01 7.37884700e-01
1.53880745e-01 -2.14084044e-01 7.25095987e-01 -7.35814154e-01
1.50796056e+00 -3.41440439e-01 4.46981266e-02 4.34374869e-01
-8.74986053e-01 3.42417091e-01 6.58238351e-01 -4.44233902e-02
-8.47600698e-01 -7.04918027e-01 -2.04126596e-01 5.93701862e-02
1.89985037e-01 9.68692183e-01 3.63653004e-01 4.78879571e-01
7.38481700e-01 3.79443854e-01 -7.48063028e-02 -4.51391220e-01
4.94410247e-01 9.70058799e-01 3.44127625e-01 -2.48300582e-01
6.27326250e-01 1.91999540e-01 -3.20706785e-01 -1.11899412e+00
-8.59193146e-01 -6.28655910e-01 -2.24501312e-01 -8.88697803e-02
7.27639198e-01 -6.57207847e-01 -3.18254411e-01 2.72149015e-02
-1.35331345e+00 2.26304665e-01 -4.55158800e-01 4.66007560e-01
-5.32696605e-01 8.69550645e-01 -5.71959794e-01 -7.43079007e-01
-8.06628883e-01 -7.69975066e-01 1.07199574e+00 3.12524021e-01
-8.33124042e-01 -1.14122677e+00 4.83588278e-01 9.72153768e-02
4.85216230e-01 1.46411479e-01 1.09502923e+00 -1.07508481e+00
-1.89336762e-02 -4.99046832e-01 -1.64369166e-01 3.29560369e-01
3.18255484e-01 7.84937963e-02 -6.45065606e-01 -2.86527514e-01
-2.92363375e-01 -1.69214293e-01 1.17604482e+00 3.68197858e-01
7.93787062e-01 -4.67381805e-01 -3.33122015e-01 1.92506574e-02
1.18302453e+00 4.04150411e-02 7.55164206e-01 3.70498270e-01
4.44774598e-01 5.86100757e-01 5.32695949e-01 5.55549741e-01
4.54941154e-01 6.39240861e-01 1.27632096e-02 1.62278950e-01
-1.66097715e-01 -4.81520206e-01 3.78549039e-01 1.58801925e+00
-1.60195213e-02 -8.52188051e-01 -7.39701509e-01 5.21353006e-01
-1.96446538e+00 -1.18629360e+00 -1.69767141e-02 2.10554981e+00
8.53894532e-01 2.24109381e-01 1.39516979e-01 1.56111971e-01
4.36970174e-01 5.72713614e-01 -1.46619484e-01 -7.16744125e-01
-1.25800088e-01 7.93782324e-02 1.24744356e-01 5.66942096e-01
-6.33249462e-01 7.48815358e-01 6.71381330e+00 9.73614991e-01
-8.73109102e-01 -8.45944136e-02 5.30422866e-01 -1.33901043e-02
-6.27704799e-01 -5.51747754e-02 -5.37294447e-01 4.57064241e-01
9.38090026e-01 -1.02052128e+00 3.27124819e-02 6.82872117e-01
2.72555411e-01 -3.40725482e-01 -1.11991751e+00 7.75359035e-01
5.00458717e-01 -1.36592364e+00 4.05948341e-01 -2.44705975e-02
7.70580888e-01 -1.99121863e-01 -6.43141419e-02 2.84846842e-01
1.90362602e-01 -7.62689650e-01 4.69524950e-01 6.05555475e-01
4.74816859e-01 -7.15588987e-01 8.99096191e-01 2.76677161e-01
-9.16900337e-01 2.20745087e-01 -3.58164161e-01 1.68582812e-01
3.65483224e-01 6.76380575e-01 -8.98422658e-01 9.87493277e-01
2.23382339e-01 7.22272217e-01 -8.74118149e-01 1.11617684e+00
-2.52189279e-01 5.34316361e-01 -9.85526145e-02 -5.78985691e-01
2.40988389e-01 -2.42490053e-01 9.42346513e-01 1.33387983e+00
5.39297581e-01 -1.93970278e-01 -1.17759921e-01 6.58243477e-01
-9.76677611e-02 5.53713739e-01 -6.26729250e-01 -5.14610410e-01
3.99758130e-01 1.24737668e+00 -7.21111417e-01 -5.52004397e-01
-1.81339905e-01 1.11188889e+00 1.81432858e-01 2.15020835e-01
-6.54677808e-01 -7.74998963e-01 4.35043931e-01 1.38865039e-02
9.00292546e-02 -2.86342472e-01 -2.35660985e-01 -1.17076540e+00
7.08019733e-02 -7.34374821e-01 1.97038665e-01 -7.99558818e-01
-1.20982873e+00 8.95836949e-01 3.02254438e-01 -1.19789064e+00
-6.43211186e-01 -1.61194578e-01 -7.67133176e-01 6.60256863e-01
-1.11402082e+00 -6.92944944e-01 -1.68456599e-01 -1.59906879e-01
7.15819359e-01 -1.74089134e-01 8.43762696e-01 -1.51345842e-02
-4.46226925e-01 6.26015604e-01 4.45591003e-01 -1.61634654e-01
7.60460019e-01 -1.32904553e+00 5.55392981e-01 8.54163229e-01
3.58042985e-01 7.24734902e-01 1.24084413e+00 -7.15668023e-01
-1.10269749e+00 -8.93041670e-01 1.16966379e+00 -4.59688753e-01
7.12253809e-01 -4.29141708e-02 -9.69614565e-01 3.88817877e-01
6.49296105e-01 -8.50499153e-01 7.20735610e-01 1.04790300e-01
-8.64200071e-02 1.07324809e-01 -6.12686694e-01 8.03800583e-01
8.78961504e-01 -2.56722629e-01 -9.84921098e-01 4.24992323e-01
7.68266797e-01 -9.08578113e-02 -7.58134604e-01 2.34590352e-01
3.76001596e-01 -9.74996090e-01 6.55431211e-01 -3.92100006e-01
7.80065775e-01 -9.57087576e-02 -7.41050171e-04 -1.85979819e+00
-3.64967972e-01 -5.30992270e-01 -2.21044257e-01 1.34630549e+00
6.41849875e-01 -8.00174892e-01 5.74865580e-01 2.15478376e-01
-1.67319939e-01 -9.05788064e-01 -6.98516548e-01 -9.33572292e-01
1.86850682e-01 6.18808642e-02 3.98740649e-01 6.34697258e-01
5.01957178e-01 7.85906196e-01 -2.10278288e-01 -3.79040807e-01
2.48334333e-01 -8.99148509e-02 7.14533091e-01 -1.31696808e+00
-2.28283610e-02 -7.28496313e-01 -3.40114564e-01 -1.09053195e+00
1.40176699e-01 -1.00790536e+00 -3.59912179e-02 -2.13880348e+00
5.35404980e-01 6.53809458e-02 -9.63148177e-02 1.06502064e-01
-5.33781767e-01 8.07075202e-02 2.30487198e-01 3.50886762e-01
-7.42969513e-01 1.08358932e+00 9.18551862e-01 -1.05451740e-01
-2.80520976e-01 -3.01006615e-01 -1.00476921e+00 4.42331016e-01
7.51648247e-01 -3.88876826e-01 -3.96213084e-01 -3.48999172e-01
2.85357594e-01 -6.93843812e-02 -1.17200658e-01 -1.12797213e+00
2.50470638e-01 1.69658780e-01 1.45132720e-01 -6.53271198e-01
3.07138771e-01 -3.67927432e-01 -2.93357596e-02 3.91998023e-01
-6.09912038e-01 4.95623976e-01 7.75542557e-02 4.38357472e-01
-4.58036870e-01 -5.47221661e-01 3.15867960e-01 7.58370534e-02
-3.40549499e-01 -7.70049915e-02 -4.33572710e-01 1.50645345e-01
6.25655413e-01 -3.57789755e-01 -4.61437166e-01 -8.61911237e-01
-9.29544792e-02 1.74617082e-01 6.04476452e-01 5.08819222e-01
7.01837480e-01 -1.19108665e+00 -1.04083395e+00 -4.62636948e-01
3.14849079e-01 -4.68478590e-01 -1.03174798e-01 8.21263969e-01
-6.65009379e-01 7.69625783e-01 4.31149118e-02 -3.12294841e-01
-1.21740556e+00 2.72955298e-01 -2.17525527e-01 -5.69473803e-01
-5.07419765e-01 3.65358710e-01 2.66078929e-03 -2.98051499e-02
-1.54507384e-01 -3.21841747e-01 -4.56924796e-01 3.13603163e-01
6.57498419e-01 6.93776786e-01 1.75308421e-01 -5.11451185e-01
-2.41504490e-01 2.10070685e-01 -3.13893020e-01 -3.85015547e-01
1.24990296e+00 -2.08072290e-01 -2.27813646e-01 6.89951241e-01
1.16543639e+00 3.08762014e-01 -5.16314685e-01 -5.14443070e-02
3.18378359e-01 -2.71657616e-01 -2.15842593e-02 -5.64926445e-01
-3.30348432e-01 6.62730157e-01 1.58950672e-01 7.85331845e-01
9.67098713e-01 3.85530144e-02 8.02147985e-01 4.26570624e-01
-6.69409856e-02 -1.21212614e+00 2.64439553e-01 7.30833292e-01
1.10631263e+00 -9.64402914e-01 4.76634085e-01 -2.45123684e-01
-7.23423243e-01 1.05599809e+00 2.70612061e-01 1.10579036e-01
1.28862068e-01 -1.44395560e-01 -1.12293124e-01 -7.46197626e-02
-8.80837440e-01 -1.53584838e-01 5.03808856e-01 2.60132432e-01
8.53111207e-01 -3.59459929e-02 -8.93315852e-01 3.32610875e-01
-6.29443944e-01 -2.74510324e-01 7.70907760e-01 7.29204774e-01
-6.30786657e-01 -1.02991331e+00 -7.56913275e-02 8.98428917e-01
-4.88586396e-01 -2.10619524e-01 -6.53626025e-01 7.21995711e-01
-7.33725965e-01 1.14071167e+00 1.47282034e-02 -2.56453127e-01
3.59658122e-01 1.87182024e-01 2.92772233e-01 -8.15329850e-01
-4.38994110e-01 -1.88455537e-01 3.61791760e-01 -1.45099401e-01
-3.72731775e-01 -6.23796999e-01 -9.47772503e-01 -4.15957183e-01
-4.68817294e-01 5.95984280e-01 6.38996124e-01 7.96469331e-01
4.81453538e-01 5.07894874e-01 4.69721407e-01 -7.62559652e-01
-6.59852087e-01 -1.48617756e+00 -4.47350085e-01 3.97689730e-01
-2.00728644e-02 -2.94291794e-01 -4.95257348e-01 -7.68825039e-02] | [12.297891616821289, 9.475226402282715] |
0a4ba403-51b0-45cb-84de-a2b8bd2363be | on-batching-variable-size-inputs-for-training | 2301.10587 | null | https://arxiv.org/abs/2301.10587v2 | https://arxiv.org/pdf/2301.10587v2.pdf | On Batching Variable Size Inputs for Training End-to-End Speech Enhancement Systems | The performance of neural network-based speech enhancement systems is primarily influenced by the model architecture, whereas training times and computational resource utilization are primarily affected by training parameters such as the batch size. Since noisy and reverberant speech mixtures can have different duration, a batching strategy is required to handle variable size inputs during training, in particular for state-of-the-art end-to-end systems. Such strategies usually strive for a compromise between zero-padding and data randomization, and can be combined with a dynamic batch size for a more consistent amount of data in each batch. However, the effect of these strategies on resource utilization and more importantly network performance is not well documented. This paper systematically investigates the effect of different batching strategies and batch sizes on the training statistics and speech enhancement performance of a Conv-TasNet, evaluated in both matched and mismatched conditions. We find that using a small batch size during training improves performance in both conditions for all batching strategies. Moreover, using sorted or bucket batching with a dynamic batch size allows for reduced training time and GPU memory usage while achieving similar performance compared to random batching with a fixed batch size. | ['Tobias May', 'Tommy Sonne Alstrøm', 'Philippe Gonzalez'] | 2023-01-25 | null | null | null | null | ['speech-enhancement'] | ['speech'] | [-1.22052833e-01 -5.63851476e-01 2.17666060e-01 -4.46715653e-01
-2.96098948e-01 -2.45582610e-01 4.00642455e-01 7.28250667e-02
-9.38943923e-01 2.73614973e-01 9.06765908e-02 -5.84795713e-01
3.50791123e-03 -4.52857196e-01 -3.34760576e-01 -9.90761459e-01
-5.18684983e-02 3.32294963e-02 4.81102854e-01 -1.48527473e-01
-8.72823372e-02 5.48957169e-01 -1.74273574e+00 2.22559243e-01
5.56091070e-01 9.70075071e-01 6.84313893e-01 8.41161907e-01
-6.06181324e-02 5.38268328e-01 -1.00556850e+00 -6.60708472e-02
2.74756014e-01 -5.03918350e-01 -2.68391281e-01 3.72276396e-01
5.19133985e-01 -3.14640850e-01 -4.65090603e-01 9.26364303e-01
1.24797690e+00 5.53827226e-01 3.19436997e-01 -8.87077332e-01
3.81439105e-02 5.78273058e-01 -1.74009353e-01 5.33607602e-01
-3.45630378e-01 3.34483892e-01 4.23426867e-01 -7.30709612e-01
2.16690898e-01 8.57097208e-01 7.56295562e-01 4.91892308e-01
-1.26228011e+00 -8.53371680e-01 1.55326352e-01 2.36497983e-01
-1.04360867e+00 -1.18103921e+00 4.33784485e-01 -2.07956374e-01
1.16348815e+00 1.62378252e-01 6.22007728e-01 8.17028224e-01
8.02030340e-02 2.42588282e-01 9.08731520e-01 -5.65114439e-01
4.67236996e-01 3.61660004e-01 8.71164277e-02 1.23333283e-01
2.33096853e-02 7.86706731e-02 -5.80635309e-01 1.90994665e-02
7.39358842e-01 -3.63888443e-01 -3.84255677e-01 1.85584888e-01
-8.00089538e-01 5.26408374e-01 -7.66831115e-02 4.94730741e-01
-5.05067885e-01 -5.43699926e-03 8.51264834e-01 5.48119187e-01
5.13274848e-01 4.06411022e-01 -5.83883166e-01 -3.88312370e-01
-1.38947332e+00 1.05125122e-01 7.35094309e-01 6.23513699e-01
3.43663275e-01 6.26313865e-01 -5.32105923e-01 1.50832403e+00
5.21720387e-02 2.37342730e-01 7.48602033e-01 -6.46981835e-01
4.69557852e-01 -1.17651813e-01 -1.57974809e-01 -5.75880945e-01
-5.46688557e-01 -7.98407614e-01 -8.50620031e-01 3.94949257e-01
5.69302559e-01 -5.94647706e-01 -1.14762592e+00 1.74690187e+00
3.32795709e-01 8.37208237e-03 4.46870513e-02 9.24057126e-01
9.83317733e-01 6.02183044e-01 1.55916318e-01 -3.03097039e-01
1.30807912e+00 -9.50845122e-01 -9.31020677e-01 -3.66552621e-01
5.08993506e-01 -1.00790429e+00 1.09655631e+00 4.08625185e-01
-1.29708552e+00 -6.67463541e-01 -1.05831003e+00 1.53569117e-01
-1.69258490e-01 4.42431360e-01 2.32539833e-01 1.26348174e+00
-1.29602206e+00 8.01400006e-01 -9.79184926e-01 -2.37700030e-01
2.09220290e-01 5.00616670e-01 -1.39368579e-01 9.81712416e-02
-8.88493836e-01 7.87007272e-01 3.55253577e-01 3.28205228e-01
-4.88626122e-01 -8.10012698e-01 -7.23618329e-01 2.24638149e-01
1.58332326e-02 -3.77059400e-01 1.31600714e+00 -8.88354957e-01
-1.92127252e+00 4.38638389e-01 -1.93957537e-01 -6.62056208e-01
4.84970450e-01 -6.52436242e-02 -5.60605288e-01 -2.23046318e-01
-5.14868498e-01 5.59910774e-01 1.11706793e+00 -7.80558348e-01
-2.99262911e-01 -1.56666741e-01 -3.36635113e-01 3.25495303e-01
-6.55874193e-01 3.41674536e-01 -4.14352626e-01 -5.97609460e-01
1.46753177e-01 -7.85915673e-01 -2.82590449e-01 -3.35908026e-01
6.71482272e-03 2.96618581e-01 7.51165986e-01 -7.05454051e-01
1.23870385e+00 -2.29975843e+00 -3.53386223e-01 1.99503731e-02
-5.58270104e-02 6.68511450e-01 -1.60036996e-01 1.82164997e-01
-1.52601972e-01 -1.73144042e-01 1.66567340e-01 -7.87467182e-01
-2.26391792e-01 8.44825357e-02 -2.73077693e-02 5.48223615e-01
1.10717081e-01 2.49719605e-01 -5.21589816e-01 -2.48862490e-01
5.12231946e-01 7.99511611e-01 -8.30890596e-01 3.66094589e-01
1.92655832e-01 5.27192593e-01 2.28993997e-01 -2.82875057e-02
8.30301464e-01 1.55690700e-01 1.34731665e-01 -2.44987264e-01
-3.87056798e-01 5.69542408e-01 -1.41987336e+00 1.13834941e+00
-9.15361941e-01 1.02687383e+00 4.47772741e-01 -7.69041538e-01
9.99366343e-01 6.69456720e-01 2.04990163e-01 -8.42408895e-01
3.17173123e-01 1.82723582e-01 4.49860722e-01 -3.47214431e-01
8.61784279e-01 -1.72941014e-01 6.98128521e-01 2.07027629e-01
8.97364914e-02 -7.33892545e-02 2.15588033e-01 -1.57196701e-01
1.03060472e+00 -3.54128689e-01 -7.35745057e-02 -1.78688407e-01
1.27510056e-01 -3.46767932e-01 5.68896413e-01 9.37251985e-01
-1.93435624e-01 6.87423050e-01 2.62337655e-01 -6.28948659e-02
-1.20298791e+00 -6.52061462e-01 -4.83422905e-01 1.33752573e+00
-2.38500789e-01 -3.62364411e-01 -7.86718190e-01 -2.83850767e-02
-5.14461637e-01 5.89626133e-01 -1.72426030e-01 -2.12840810e-01
-6.32318735e-01 -9.44962978e-01 6.07770383e-01 4.05831218e-01
6.59667075e-01 -1.07939136e+00 -8.11457098e-01 4.42283124e-01
7.77530968e-02 -1.34123719e+00 -5.88975191e-01 7.00217187e-01
-1.01098979e+00 -4.70989555e-01 -8.22015166e-01 -5.75737417e-01
5.47975719e-01 3.36033940e-01 1.10267949e+00 3.62006202e-02
1.19771518e-01 1.13957405e-01 -3.48562479e-01 -4.31234717e-01
-5.07098973e-01 4.38676365e-02 2.05116197e-01 -3.48696783e-02
-1.50361329e-01 -6.00100279e-01 -7.92890489e-01 4.32069331e-01
-9.43901479e-01 5.88350743e-02 4.03237045e-01 9.79064941e-01
3.41487616e-01 3.69655818e-01 5.47543824e-01 -5.91941416e-01
6.82129145e-01 -1.87137157e-01 -5.77728570e-01 -1.15283020e-01
-5.13641000e-01 -8.53153467e-02 7.76509047e-01 -9.62524116e-01
-1.20876980e+00 -2.50255167e-01 -5.61960101e-01 -5.02693713e-01
-2.29274139e-01 2.70958751e-01 -1.18973702e-01 -3.10604703e-02
6.43172681e-01 -1.23502156e-02 -6.85203224e-02 -5.20741463e-01
4.86307405e-02 8.55520844e-01 3.56796384e-01 -2.40004689e-01
1.55651391e-01 1.01119742e-01 -3.96127939e-01 -1.28606761e+00
-1.95911184e-01 -5.43357968e-01 -2.75521249e-01 -4.20147002e-01
6.15088105e-01 -8.98652196e-01 -3.11446249e-01 9.87817168e-01
-9.67148900e-01 -8.48098099e-01 -3.04490507e-01 8.61624897e-01
-1.07912853e-01 6.02743477e-02 -6.12137735e-01 -9.01866436e-01
-4.84835088e-01 -1.39988160e+00 6.75100863e-01 2.98045605e-01
-9.91106853e-02 -9.36297596e-01 -9.94903296e-02 7.03337863e-02
1.03717792e+00 -3.69439781e-01 5.94483614e-01 -7.04492331e-01
1.47181660e-01 1.41641861e-02 -2.00922713e-02 6.51379406e-01
1.25318676e-01 -4.91563044e-02 -1.37337577e+00 -5.46150982e-01
3.53891283e-01 7.35758245e-02 7.82130480e-01 9.35409069e-01
9.38910902e-01 -1.71358794e-01 9.22657028e-02 5.90679765e-01
1.02830207e+00 3.12978327e-01 7.59293258e-01 3.40247452e-01
3.70781839e-01 4.49068487e-01 3.09907198e-01 5.86779475e-01
-1.75639361e-01 7.83366859e-01 1.70061797e-01 -2.24682301e-01
-4.12126839e-01 4.13391232e-01 3.03652972e-01 8.08107913e-01
1.28594711e-01 -2.40672261e-01 -6.90209866e-01 4.11677480e-01
-1.28626823e+00 -9.65259373e-01 1.72316693e-02 2.66740489e+00
1.03988373e+00 2.00940952e-01 2.88827568e-01 4.83008951e-01
8.04603040e-01 2.80864835e-01 -2.73609489e-01 -5.36517203e-01
-1.06177032e-01 3.68337095e-01 6.20377362e-01 3.16913903e-01
-9.19060171e-01 6.78048074e-01 6.40485239e+00 9.97062922e-01
-1.78742051e+00 2.59404123e-01 7.71735489e-01 -6.89209044e-01
2.39073381e-01 -3.64097625e-01 -8.50286841e-01 4.69537109e-01
1.39303124e+00 2.37719283e-01 3.48061323e-01 7.18343437e-01
6.59702241e-01 -2.06738129e-01 -7.32362151e-01 1.02963459e+00
-3.58467609e-01 -1.04303384e+00 -6.77474082e-01 5.85180782e-02
5.65436304e-01 4.41803753e-01 8.21917728e-02 3.26720148e-01
-1.53957009e-01 -6.55920684e-01 9.80598271e-01 -9.13473368e-02
5.78419983e-01 -6.03990316e-01 7.51035273e-01 1.16812155e-01
-1.03794777e+00 -2.26727817e-02 -4.55746204e-01 1.77007448e-02
8.23561251e-02 9.98757362e-01 -7.85571218e-01 -6.92684948e-02
7.84337699e-01 -2.62815543e-02 -3.14549595e-01 1.45076168e+00
1.58078119e-01 1.02114439e+00 -4.45073664e-01 -7.64242858e-02
7.07744062e-02 -1.42101303e-01 5.36857903e-01 1.54080951e+00
2.71629721e-01 -7.32946098e-02 -2.28050858e-01 3.15857589e-01
1.41806602e-01 -1.94867793e-02 -1.84841193e-02 7.58461282e-02
7.89246202e-01 1.18664968e+00 -7.86551237e-01 -2.45201796e-01
-2.19323665e-01 6.33100450e-01 9.50336754e-02 5.26886225e-01
-8.25765729e-01 -4.71057236e-01 8.42410564e-01 2.91992635e-01
6.77576780e-01 -5.63201249e-01 -3.59423369e-01 -6.72489405e-01
5.03978096e-02 -9.19727921e-01 4.51948456e-02 -4.54774767e-01
-7.72529721e-01 1.11593747e+00 -1.41662911e-01 -1.02959204e+00
-1.29439726e-01 -4.97728884e-01 -8.39263201e-01 9.77681518e-01
-1.24955869e+00 -5.03849328e-01 -3.60715508e-01 4.41441357e-01
6.32252812e-01 -6.94599673e-02 6.23861074e-01 7.08078623e-01
-9.01472032e-01 8.84877205e-01 2.22334787e-01 -6.44508600e-02
7.78037012e-01 -9.66839552e-01 4.70091224e-01 9.69658375e-01
1.70676962e-01 5.48712969e-01 1.08121407e+00 -2.46815860e-01
-1.12046731e+00 -9.48227942e-01 4.67812568e-01 3.82519484e-01
1.76539376e-01 -3.55496109e-01 -1.15820074e+00 7.20207095e-02
1.06206998e-01 9.70561355e-02 6.13988757e-01 2.74526298e-01
-1.39283702e-01 -2.10069165e-01 -1.02962089e+00 8.32942545e-01
6.13369823e-01 -4.38751280e-01 1.41562432e-01 1.92664877e-01
5.64508915e-01 -8.63111258e-01 -7.83693433e-01 2.26570740e-01
5.39192677e-01 -1.12211478e+00 7.45731056e-01 -3.07796270e-01
-3.23135555e-02 -6.73844144e-02 -2.99397088e-03 -1.52646601e+00
-2.26609841e-01 -5.91647804e-01 1.60676718e-01 1.29264534e+00
4.68730420e-01 -7.03593493e-01 7.16416419e-01 9.14909303e-01
-3.33507210e-01 -8.87532294e-01 -9.86698985e-01 -9.22090232e-01
-2.21518472e-01 -8.20464611e-01 4.76846844e-01 5.05569339e-01
-3.69195729e-01 1.17666563e-02 -2.37455472e-01 1.77609876e-01
3.78080048e-02 -6.09279692e-01 6.01352751e-01 -6.03184402e-01
-5.66942871e-01 -5.23782074e-01 -3.08159083e-01 -8.92051935e-01
-1.99116334e-01 -3.20970684e-01 2.57615894e-01 -1.22625065e+00
-3.74862939e-01 -8.15779746e-01 -1.13768272e-01 3.03794384e-01
-2.14175537e-01 2.85904258e-01 2.05835670e-01 -5.18002845e-02
-9.48657468e-02 5.45494199e-01 8.81931484e-01 1.93234682e-01
-8.64170372e-01 2.08513588e-01 -2.69840509e-01 5.02603590e-01
9.20738876e-01 -4.58795696e-01 -4.87530261e-01 -7.03285277e-01
-2.35833690e-01 -3.56730111e-02 2.27679908e-01 -1.31085646e+00
4.70653027e-01 8.80705416e-02 2.66070038e-01 -3.51414710e-01
6.49153113e-01 -6.42914236e-01 1.90219671e-01 1.61122859e-01
-2.88055629e-01 9.25460085e-02 7.51017690e-01 6.67244270e-02
-1.21111237e-01 -5.27827442e-01 1.05702531e+00 3.38145107e-01
-4.34240490e-01 1.33378610e-01 -6.60285890e-01 -9.18633416e-02
4.85700190e-01 -5.13216734e-01 -7.81449154e-02 -3.37711692e-01
-6.25598669e-01 -1.85552001e-01 1.08178884e-01 2.91495502e-01
2.94611901e-01 -1.03448427e+00 -7.06305921e-01 4.38316494e-01
-4.02910084e-01 -1.30503923e-01 5.68851590e-01 1.11077619e+00
-2.77047187e-01 1.80832431e-01 -5.69398254e-02 -6.15598023e-01
-1.67482591e+00 -4.34434302e-02 7.39859641e-01 -1.15321174e-01
-5.33077359e-01 1.06135190e+00 -2.40932718e-01 -1.56384066e-01
6.32759511e-01 -2.86132216e-01 -1.75095826e-01 5.33922799e-02
6.93975210e-01 4.57601935e-01 7.49405921e-01 -3.46660346e-01
-1.49852201e-01 2.38490030e-02 -2.85855383e-01 -4.71391678e-01
1.23722267e+00 -5.30622117e-02 1.38043702e-01 3.51487577e-01
9.76669073e-01 -6.29363731e-02 -1.27674890e+00 -2.94912130e-01
-4.68448281e-01 -4.69123065e-01 7.18527079e-01 -4.36332732e-01
-1.40998340e+00 6.78099573e-01 9.64150786e-01 1.75968334e-01
1.52879441e+00 -3.36683631e-01 5.99321008e-01 1.89465448e-01
5.59245273e-02 -1.07851195e+00 -8.09418932e-02 6.69278860e-01
6.95850194e-01 -1.02000594e+00 -1.44343749e-01 -6.71302602e-02
-6.39045537e-01 9.72003341e-01 5.15542150e-01 2.25971922e-01
6.92685783e-01 7.28622854e-01 2.63972729e-01 1.74211189e-01
-7.63415575e-01 -1.23970486e-01 1.29667014e-01 5.34122169e-01
7.48793602e-01 -9.70642790e-02 -1.36254460e-01 4.16968018e-01
-4.26230580e-01 -2.83317447e-01 2.86190420e-01 8.33425820e-01
-3.51126343e-01 -1.13628447e+00 -5.97188890e-01 6.29863262e-01
-6.62445307e-01 -3.84706914e-01 2.94288576e-01 2.90257215e-01
1.18040383e-01 1.41176212e+00 4.46473151e-01 -4.57462847e-01
3.97892833e-01 -2.23632138e-02 4.78008777e-01 -5.32711685e-01
-1.23763490e+00 2.09538788e-01 2.22251654e-01 -1.81076854e-01
-1.11814372e-01 -6.51651084e-01 -9.58692014e-01 -4.21539932e-01
-8.34086716e-01 1.39664710e-01 1.14847684e+00 9.17016029e-01
3.85390550e-01 9.57965791e-01 5.37060618e-01 -1.05353892e+00
-6.07682168e-01 -1.33514655e+00 -5.88000417e-01 -4.29851264e-02
3.38420123e-01 -3.88908058e-01 -3.90019447e-01 7.18727475e-04] | [14.879294395446777, 5.9369707107543945] |
030ff9cc-958c-4e59-97de-e2a19eeddeba | au-aware-graph-convolutional-network-for | 2303.09114 | null | https://arxiv.org/abs/2303.09114v1 | https://arxiv.org/pdf/2303.09114v1.pdf | AU-aware graph convolutional network for Macro- and Micro-expression spotting | Automatic Micro-Expression (ME) spotting in long videos is a crucial step in ME analysis but also a challenging task due to the short duration and low intensity of MEs. When solving this problem, previous works generally lack in considering the structures of human faces and the correspondence between expressions and relevant facial muscles. To address this issue for better performance of ME spotting, this paper seeks to extract finer spatial features by modeling the relationships between facial Regions of Interest (ROIs). Specifically, we propose a graph convolutional-based network, called Action-Unit-aWare Graph Convolutional Network (AUW-GCN). Furthermore, to inject prior information and to cope with the problem of small datasets, AU-related statistics are encoded into the network. Comprehensive experiments show that our results outperform baseline methods consistently and achieve new SOTA performance in two benchmark datasets,CAS(ME)^2 and SAMM-LV. Our code is available at https://github.com/xjtupanda/AUW-GCN. | ['Enhong Chen', 'Sirui Zhao', 'Shifeng Liu', 'Tong Xu', 'Shiwei Wu', 'Shukang Yin'] | 2023-03-16 | null | null | null | null | ['micro-expression-spotting'] | ['computer-vision'] | [ 2.16403097e-01 8.65687504e-02 -3.19526464e-01 -5.13108909e-01
-3.32923889e-01 -1.85307384e-01 3.25640291e-01 -4.41254020e-01
-9.94447246e-02 4.32679236e-01 2.81681061e-01 2.36724257e-01
-5.02503589e-02 -6.23821437e-01 -4.98585105e-01 -5.88983834e-01
-2.23931447e-01 -9.55917761e-02 -3.97700705e-02 -2.93024868e-01
4.60059606e-02 6.84292197e-01 -1.30764985e+00 2.43630201e-01
4.09634650e-01 1.28663743e+00 -8.53363872e-02 1.95043847e-01
-1.52348038e-02 8.70917499e-01 -6.64301738e-02 -6.31833613e-01
2.90896744e-01 -7.28757858e-01 -8.16607773e-01 1.89681366e-01
6.30540788e-01 -1.78968385e-01 -3.70812595e-01 1.24826336e+00
3.77761662e-01 1.45160452e-01 6.09163642e-01 -1.49802327e+00
-2.44224370e-01 3.29144269e-01 -1.11947739e+00 2.29191318e-01
1.88867450e-01 -3.77509417e-03 9.91240859e-01 -9.12453413e-01
9.64583576e-01 1.31108749e+00 5.15936553e-01 6.34848714e-01
-9.24445689e-01 -9.66735840e-01 8.99745151e-02 3.72324318e-01
-1.62512529e+00 -7.32791305e-01 1.12773228e+00 -2.84350723e-01
5.64093709e-01 1.97320059e-01 5.75860262e-01 1.20245993e+00
1.06561467e-01 7.80900598e-01 9.73197103e-01 -1.30772263e-01
-2.11477250e-01 -2.73954451e-01 -2.08017543e-01 1.15629947e+00
-3.12442750e-01 -2.48991549e-01 -5.91096401e-01 1.06837228e-01
9.42317545e-01 -1.08071305e-01 -2.91619778e-01 -1.46224439e-01
-8.12175989e-01 6.10171437e-01 4.99229312e-01 5.48600197e-01
-4.43954766e-01 3.27376008e-01 6.10623062e-01 2.59352535e-01
7.10444331e-01 1.27410218e-01 -1.93336517e-01 -8.76023844e-02
-1.06505311e+00 3.80279683e-02 4.48740900e-01 6.89125776e-01
8.40049624e-01 9.06529948e-02 -3.59795451e-01 8.78856003e-01
1.77206948e-01 3.92228663e-02 1.27581730e-01 -1.03322434e+00
3.39697272e-01 8.35077763e-01 -4.22082603e-01 -1.56910682e+00
-4.48129684e-01 -2.69232869e-01 -1.12315154e+00 1.05723530e-01
3.63798648e-01 -2.58510888e-01 -7.15873003e-01 1.88535368e+00
4.64184523e-01 3.91886234e-01 -4.88464415e-01 9.82700467e-01
1.02434075e+00 5.30664980e-01 2.34721273e-01 -2.83656329e-01
1.26918995e+00 -9.85993326e-01 -7.86607504e-01 -1.92738488e-01
6.66572750e-01 -6.03591144e-01 8.29216540e-01 2.47050598e-01
-8.74453545e-01 -4.74555939e-01 -5.94143033e-01 -1.10791348e-01
-1.87123209e-01 5.86423576e-01 7.38359869e-01 1.79716483e-01
-1.12369406e+00 5.84863961e-01 -6.58594728e-01 -3.22674513e-01
9.09810364e-01 5.24272621e-01 -7.83200860e-01 1.38485953e-01
-1.01070678e+00 6.14151776e-01 3.66607271e-02 5.85387409e-01
-7.12925971e-01 -3.98688078e-01 -9.91412282e-01 -1.39631312e-02
4.93437439e-01 -3.63708347e-01 8.27129185e-01 -1.58229792e+00
-1.48518133e+00 1.07165098e+00 -2.26401374e-01 -2.38133911e-02
4.76319373e-01 -3.81091833e-02 -3.16790283e-01 3.02983969e-01
-8.92269388e-02 8.08920085e-01 1.00260341e+00 -9.75689232e-01
-2.19215065e-01 -4.88137633e-01 7.45321289e-02 -7.09296167e-02
-3.42991889e-01 3.89643282e-01 -6.10114574e-01 -8.11226606e-01
-8.13187957e-02 -8.89620543e-01 -1.81755975e-01 3.40711325e-01
-5.46263695e-01 -4.77340341e-01 9.16105151e-01 -8.25150430e-01
1.35628378e+00 -2.23944879e+00 3.81945163e-01 3.57020169e-01
4.71567661e-01 3.38449150e-01 -4.72710907e-01 4.81315926e-02
-2.88897902e-01 -1.56487972e-02 -1.81721643e-01 -4.31051761e-01
-1.18857682e-01 -2.44383216e-02 2.63449848e-01 6.34350777e-01
3.66741776e-01 1.18864393e+00 -6.35470450e-01 -8.17852139e-01
1.16656654e-01 6.46809876e-01 -4.13008183e-01 1.41417801e-01
-1.31911725e-01 6.55884087e-01 -5.81791580e-01 9.35704589e-01
7.22616076e-01 -2.82850474e-01 2.12790459e-01 -6.14481270e-01
2.84819990e-01 -3.14668566e-01 -1.05467153e+00 1.74479711e+00
-3.88898998e-01 8.08510303e-01 5.02913892e-01 -1.15607202e+00
9.36662734e-01 1.94908395e-01 7.98150599e-01 -7.16287434e-01
6.44384146e-01 7.22145895e-04 -1.15257740e-01 -7.39188910e-01
3.92154455e-02 9.25850719e-02 2.71959871e-01 2.49670804e-01
2.64771521e-01 3.68521214e-01 2.08829671e-01 3.56928371e-02
8.19934487e-01 3.68647456e-01 4.41675872e-01 -2.16095805e-01
6.56316876e-01 -4.82539535e-01 8.71457458e-01 -6.34977520e-02
-3.49684864e-01 5.04401445e-01 9.49609935e-01 -4.23606575e-01
-5.37484467e-01 -4.20314670e-01 -3.79047655e-02 1.06998920e+00
-1.41611397e-02 -7.12121069e-01 -9.49388742e-01 -8.38734150e-01
-1.84768438e-01 1.25129268e-01 -1.09118044e+00 1.44639695e-02
-7.17042208e-01 -5.90986550e-01 4.29407448e-01 5.09699821e-01
6.50683105e-01 -1.16375971e+00 -3.48016828e-01 4.43553232e-04
-2.89149046e-01 -1.20698571e+00 -6.03934526e-01 -3.84293854e-01
-5.33967733e-01 -1.18652797e+00 -7.20074475e-01 -7.89431512e-01
8.45134616e-01 -8.12840983e-02 1.05230045e+00 2.47580051e-01
-6.04870498e-01 1.57428637e-01 -2.80291796e-01 -8.98942426e-02
2.45915558e-02 -1.52227012e-02 -2.69932896e-01 5.95607042e-01
4.50719982e-01 -7.38077104e-01 -7.05403686e-01 3.86346251e-01
-7.78129041e-01 7.82419667e-02 5.91670454e-01 6.20533943e-01
6.01791143e-01 -2.48623058e-01 3.27245831e-01 -8.27523053e-01
4.60317373e-01 -5.29237509e-01 -4.54438418e-01 3.12490433e-01
-1.32682547e-01 -1.93573549e-01 2.88985252e-01 -2.53398299e-01
-9.71802473e-01 3.39144915e-01 -2.77732193e-01 -7.15221643e-01
-1.98965818e-01 3.18537593e-01 -3.35536689e-01 -4.62678432e-01
3.36234659e-01 -1.13822361e-02 3.45249236e-01 -2.13799179e-01
1.91712454e-01 4.54108834e-01 1.96298957e-01 -3.58862758e-01
4.12843913e-01 4.76635396e-01 4.15387094e-01 -7.77902901e-01
-9.35950816e-01 -2.63716191e-01 -8.91945064e-01 -5.47318518e-01
9.80779886e-01 -8.19782138e-01 -6.80514812e-01 2.99087614e-01
-1.00237787e+00 -3.58083576e-01 2.37706274e-01 1.42556682e-01
-4.79919642e-01 4.95631099e-01 -6.53467655e-01 -5.61292827e-01
-4.75360334e-01 -1.14690316e+00 1.19931948e+00 1.94656432e-01
-3.30768377e-01 -9.52907562e-01 -3.07881441e-02 4.03779924e-01
3.03840280e-01 8.53496671e-01 4.99659747e-01 -1.04451321e-01
-4.02149826e-01 7.48918951e-03 -4.27407295e-01 3.89203966e-01
2.89124906e-01 2.85235167e-01 -9.40701723e-01 -9.61324796e-02
-3.89546871e-01 -4.15796459e-01 8.95271659e-01 4.66439486e-01
1.57124281e+00 -3.24222833e-01 -3.72830480e-01 9.95838344e-01
1.14552343e+00 -1.67471722e-01 6.44920170e-01 -1.86724234e-02
8.91613185e-01 9.25921440e-01 7.77818441e-01 4.78883356e-01
7.55745322e-02 8.37310135e-01 6.30856156e-01 -3.44319612e-01
-2.91142792e-01 2.09167190e-02 3.76734108e-01 5.87994456e-01
-4.92571831e-01 -1.66218445e-01 -6.44128799e-01 5.05264759e-01
-1.88410532e+00 -8.34110081e-01 -4.51880842e-02 1.52945340e+00
6.48119450e-01 -2.57568091e-01 2.40273610e-01 -1.60470486e-01
5.58723032e-01 5.14250994e-01 -3.15894395e-01 -3.73652905e-01
-9.36013833e-02 3.63516182e-01 1.71943635e-01 3.51479560e-01
-1.22117698e+00 1.04057813e+00 5.07641554e+00 1.07444751e+00
-1.38378239e+00 1.87466413e-01 9.73274052e-01 -1.45291999e-01
1.21192783e-01 -3.36600184e-01 -6.27585769e-01 4.32276160e-01
6.10439122e-01 1.72322392e-01 2.92845577e-01 7.57800400e-01
2.94835478e-01 9.61162150e-02 -9.06554401e-01 1.24045336e+00
1.94820270e-01 -1.25017726e+00 -1.27253503e-01 -1.75601859e-02
6.10158622e-01 -1.30634457e-01 -2.51264889e-02 7.35869780e-02
-2.06701502e-01 -1.31823516e+00 3.43724877e-01 4.90030468e-01
1.08308578e+00 -8.47764134e-01 6.99188173e-01 -1.35193586e-01
-1.47472632e+00 9.59251970e-02 -3.53764623e-01 1.06913172e-01
-2.10689195e-02 4.76743698e-01 -5.21700680e-01 4.80622530e-01
7.74028122e-01 1.13694060e+00 -5.78212619e-01 6.50834322e-01
-3.42920244e-01 6.11869037e-01 -2.47216806e-01 4.82622236e-02
2.47849345e-01 -3.66435498e-01 3.78391832e-01 1.37966645e+00
2.47048005e-01 1.47776499e-01 -1.37390122e-01 1.06771123e+00
-3.07332009e-01 4.44702089e-01 -6.23547614e-01 -1.78314880e-01
-1.52836397e-01 1.82010484e+00 -6.42932057e-01 9.02120024e-02
-4.07489330e-01 1.16283083e+00 4.89067912e-01 3.35455656e-01
-8.92771244e-01 -1.78565398e-01 8.56660128e-01 2.31029600e-01
1.02493435e-01 -2.17022255e-01 3.11529011e-01 -1.14504361e+00
8.40688273e-02 -7.83862948e-01 3.68921638e-01 -6.28649831e-01
-1.12709749e+00 6.42889857e-01 -1.41062260e-01 -1.18373442e+00
-1.14318080e-01 -7.32944787e-01 -7.32409537e-01 5.15147328e-01
-1.52922976e+00 -1.49791658e+00 -6.52362525e-01 8.68267179e-01
5.09831488e-01 1.65020786e-02 6.74171388e-01 6.00214779e-01
-8.40325415e-01 8.37884843e-01 -4.37730372e-01 5.79239249e-01
7.33911991e-01 -7.27091789e-01 -2.78675631e-02 5.99393845e-01
1.80581823e-01 3.81342262e-01 3.43326777e-01 -4.84943897e-01
-1.20793223e+00 -1.23675537e+00 7.44691670e-01 -2.22178698e-01
5.38698494e-01 -4.94066745e-01 -8.15558493e-01 7.81370163e-01
1.94299549e-01 4.38598573e-01 5.51032424e-01 -9.23984647e-02
-2.02301085e-01 -2.76490718e-01 -1.01350951e+00 5.57647347e-01
1.36945176e+00 -3.72833848e-01 7.39508644e-02 3.10804844e-01
1.77063659e-01 -4.06582296e-01 -9.81159687e-01 5.30467808e-01
5.79083443e-01 -1.02302122e+00 6.93992794e-01 -5.51901400e-01
6.24278963e-01 -1.48422003e-01 1.46254346e-01 -1.11722898e+00
-2.31978461e-01 -6.89385831e-01 1.40585536e-02 1.23003089e+00
2.03006148e-01 -1.81466863e-01 9.52930331e-01 2.08262205e-01
2.58503318e-01 -1.02557158e+00 -9.40612793e-01 -4.08099920e-01
-2.40615427e-01 -2.48585820e-01 4.66233701e-01 1.24622536e+00
-9.69570205e-02 2.10267872e-01 -5.53699017e-01 -9.83397439e-02
5.19176722e-01 1.89892888e-01 7.40945637e-01 -9.42129314e-01
-1.88342724e-02 -6.06568754e-01 -7.09099710e-01 -6.56829596e-01
5.23205817e-01 -9.26319897e-01 -2.95841783e-01 -1.27087116e+00
2.07787618e-01 -1.76806018e-01 -3.09917510e-01 7.91947484e-01
-2.26070229e-02 5.07407606e-01 9.67533290e-02 -1.18038781e-01
-6.13802671e-01 6.51524901e-01 1.52863514e+00 -2.45551467e-02
9.25344005e-02 -1.99608102e-01 -3.32003325e-01 7.47734427e-01
9.04208243e-01 -2.52737463e-01 -4.16028261e-01 -2.27093562e-01
1.55840963e-01 6.96356222e-02 4.91978705e-01 -8.13208163e-01
1.04347281e-01 -2.76951879e-01 3.95923793e-01 -2.90876120e-01
3.85507137e-01 -7.37403274e-01 1.77704662e-01 7.43802339e-02
-1.72099426e-01 -2.06941869e-02 1.53742447e-01 2.12495431e-01
-5.46531558e-01 7.50963241e-02 8.78837049e-01 -2.40635946e-01
-1.07576668e+00 8.05228174e-01 8.90395325e-03 -1.19485311e-01
1.14999449e+00 -1.27445087e-01 8.93638935e-03 -5.69908202e-01
-8.25071633e-01 2.22627193e-01 1.67575657e-01 4.70293760e-01
7.19783485e-01 -1.38940513e+00 -6.74748421e-01 1.32810563e-01
1.36364326e-01 -1.06876493e-01 3.55139762e-01 1.24381411e+00
-5.03922343e-01 1.50280863e-01 -6.11987889e-01 -4.64397460e-01
-1.63358831e+00 1.40307352e-01 6.03942454e-01 -7.39106734e-04
-4.09561813e-01 1.19366074e+00 3.89795542e-01 -2.38955379e-01
2.94026017e-01 -1.79491222e-01 -5.17627835e-01 2.43748605e-01
4.03011829e-01 1.64885342e-01 -2.33525157e-01 -1.07343018e+00
-4.16689456e-01 8.76564085e-01 1.88968316e-01 4.59029287e-01
1.48279703e+00 -1.09567232e-01 -5.80337405e-01 -9.06116143e-02
1.54308736e+00 7.56115541e-02 -1.26731765e+00 -1.95576385e-01
5.15144765e-02 -5.22973120e-01 1.03236042e-01 -3.08250606e-01
-1.82286859e+00 9.28862989e-01 5.39610088e-01 -4.27857161e-01
1.37131429e+00 5.82334921e-02 6.67707503e-01 -8.04034807e-03
9.83534828e-02 -9.22240019e-01 2.95309395e-01 4.55029085e-02
1.08918726e+00 -1.30937958e+00 -1.15952846e-02 -7.93921232e-01
-6.56726301e-01 1.32129383e+00 1.03590322e+00 -3.44838314e-02
9.27112699e-01 1.47910014e-01 1.35490164e-01 -5.81597090e-01
-4.14106488e-01 -1.90373629e-01 5.42254508e-01 2.93447018e-01
6.08762622e-01 -7.60681853e-02 -2.75385857e-01 4.05692428e-01
-5.15224226e-02 2.50455022e-01 6.93353564e-02 7.31897473e-01
-1.10168695e-01 -1.07088375e+00 2.07856983e-01 3.84224236e-01
-7.67038882e-01 9.26250443e-02 -5.90545952e-01 7.86695004e-01
2.96387434e-01 5.83728909e-01 2.36576851e-02 -4.47268039e-01
1.69142157e-01 -1.23670861e-01 5.32769144e-01 -3.50775093e-01
-3.94529730e-01 2.38971949e-01 2.45860994e-01 -1.17839754e+00
-8.37136388e-01 -5.04817903e-01 -1.16009617e+00 -2.71919489e-01
-1.51395410e-01 -2.03004405e-01 4.11451638e-01 6.39852405e-01
5.50451994e-01 3.48102152e-01 5.83182216e-01 -7.77369440e-01
-9.91678890e-03 -9.14852321e-01 -7.58933187e-01 6.44640446e-01
2.22616285e-01 -8.16639423e-01 -2.22730100e-01 -2.30501480e-02] | [13.642552375793457, 1.6052922010421753] |
405fb63a-043b-4d4d-bef3-f986d30bbdc4 | for-the-sake-of-privacy-skeleton-based | null | null | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9897358 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9897358 | FOR THE SAKE OF PRIVACY: SKELETON-BASED SALIENT BEHAVIOR RECOGNITION | Authorities as well as emergency and rescue services have an increasing interest in smart support systems to ensure public safety which includes in particular behavioral analysis of pedestrians by using video surveillance systems. In order to accommodate concerns of citizens regarding their personal rights, the demand for data privacy friendly approaches, using as few information as possible, arises. In this paper, we examine existing approaches tackling the recognition of anomalous or salient behavior based solely on person pose information within the context of real-world surveillance applications. Particularly, we chose two existing state-of-the-art approaches and evaluate them on two public and an internal dataset in order to examine the overall performance of these methods for the desired task. Furthermore, we present our
own approach achieving comparable results to these methods. Finally, we extend the aforementioned methods with a memory extension for modeling normal behavior, which yields on average a 4.3% higher recognition performance. | ['Jurgen Beyerer', 'Mickael Cormier', 'Johanna Thiemich', 'Thomas Golda'] | 2022-10-18 | null | null | null | 2022-ieee-international-conference-on-image | ['anomaly-detection-in-surveillance-videos', 'anomaly-detection-in-surveillance-videos'] | ['computer-vision', 'methodology'] | [ 1.63643584e-01 4.24397886e-02 -1.18899317e-02 -6.63231611e-01
-5.83828092e-01 -4.43774849e-01 7.16289401e-01 2.34332472e-01
-8.51712048e-01 7.22017348e-01 3.11439842e-01 -3.00562203e-01
2.79777646e-01 -5.54828942e-01 -2.41023153e-01 -5.97001493e-01
3.09658479e-02 1.00825354e-01 4.91816580e-01 -2.17222482e-01
9.25532356e-02 7.86543548e-01 -1.78348017e+00 3.85647207e-01
5.75603783e-01 9.95430112e-01 -6.34087265e-01 5.86297870e-01
5.07501483e-01 5.52170157e-01 -5.69711268e-01 -8.55429947e-01
5.67260802e-01 3.55861425e-01 -5.57234585e-01 2.72566587e-01
8.36876631e-01 -7.81722069e-01 -6.77639842e-01 8.77122760e-01
4.50299352e-01 4.65078741e-01 2.97220945e-01 -1.46576834e+00
-1.43223554e-01 -3.56207818e-01 -2.86760420e-01 4.78267610e-01
8.50139797e-01 4.15443450e-01 5.75674415e-01 -4.01726812e-01
3.48894328e-01 9.91736710e-01 5.79099774e-01 7.28926122e-01
-1.09650493e+00 -5.08547008e-01 4.79458809e-01 2.73034364e-01
-1.43123662e+00 -1.02454150e+00 5.08433282e-01 -3.64841282e-01
9.85005260e-01 6.95357561e-01 5.91624558e-01 1.38939834e+00
-1.74262241e-01 8.40704620e-01 8.74618769e-01 -7.42093474e-02
2.01989785e-01 5.94010592e-01 5.70011616e-01 4.14982229e-01
6.48605943e-01 4.08419296e-02 -2.38004193e-01 -6.80621028e-01
1.04915030e-01 1.28339320e-01 -3.48888129e-01 -5.21192491e-01
-8.29076588e-01 5.78963637e-01 -9.40321013e-03 8.50507170e-02
-3.46436977e-01 -3.72406244e-01 6.50882602e-01 -1.62402168e-02
5.48354626e-01 -6.47492856e-02 -2.94416994e-01 -1.91308662e-01
-9.13951159e-01 4.89842415e-01 8.69824290e-01 8.38788927e-01
3.40645850e-01 -1.91357329e-01 -2.81883538e-01 4.14791644e-01
2.97371894e-01 4.64243144e-01 4.88733947e-02 -7.82636285e-01
5.33190131e-01 7.02106714e-01 4.15405303e-01 -1.19241345e+00
-4.42620814e-01 -3.86397280e-02 -7.52149701e-01 7.15428889e-02
5.13500750e-01 -5.70410751e-02 -5.37141919e-01 1.60280144e+00
4.54675674e-01 7.30720982e-02 1.13431558e-01 7.87535250e-01
5.63696265e-01 4.65185583e-01 2.69699693e-01 -4.32498574e-01
1.55876899e+00 -4.42927986e-01 -6.48168385e-01 -2.40172893e-01
5.25437474e-01 -3.29627961e-01 6.63387775e-01 4.39378440e-01
-8.02798271e-01 -1.94658801e-01 -6.82261825e-01 1.50869697e-01
-6.71565413e-01 7.51081705e-02 2.54276454e-01 1.19471443e+00
-1.03603375e+00 1.41251206e-01 -8.41112316e-01 -9.87871706e-01
4.28243697e-01 4.97646600e-01 -5.61637640e-01 8.17846507e-03
-1.03357029e+00 1.00547278e+00 7.63369352e-02 -7.72026777e-02
-5.15742064e-01 -2.16350049e-01 -8.85699391e-01 -1.95238903e-01
4.86584336e-01 -4.96791869e-01 1.08294964e+00 -5.00993907e-01
-9.76055145e-01 1.11464894e+00 -3.86789650e-01 -7.13100612e-01
7.66034186e-01 -3.80669951e-01 -9.72219408e-01 1.04585513e-01
5.30243032e-02 4.65706587e-01 5.44122577e-01 -9.78281260e-01
-8.31496596e-01 -6.26695752e-01 9.12717506e-02 -3.02936118e-02
-5.29303730e-01 3.99962783e-01 -3.78734708e-01 -4.74960417e-01
-4.05969679e-01 -1.04152858e+00 -1.88038245e-01 1.11159518e-01
-2.79394746e-01 -2.33129673e-02 1.13777876e+00 -7.83537745e-01
1.30953538e+00 -2.26898193e+00 -4.14085180e-01 3.24641705e-01
1.18163750e-01 7.61778831e-01 2.88761884e-01 2.35450387e-01
1.63854077e-01 -1.10249616e-01 -1.12759314e-01 -4.96146351e-01
-1.07729562e-01 2.09293306e-01 -2.58242637e-01 8.68479848e-01
1.75101280e-01 4.22456086e-01 -5.18184602e-01 -4.12459075e-01
3.63648176e-01 4.29556370e-01 -4.92076606e-01 2.07200065e-01
2.94226497e-01 4.28791672e-01 -5.50702214e-01 8.34640443e-01
6.04832411e-01 2.61482984e-01 8.31868798e-02 -9.23265368e-02
-3.32326084e-01 -4.10499377e-03 -1.00480711e+00 9.13667977e-01
2.51521677e-01 5.98316252e-01 1.79443374e-01 -9.55886841e-01
7.70040274e-01 5.70507348e-01 6.64201498e-01 -5.56490004e-01
2.90938854e-01 -1.43309802e-01 -3.59208167e-01 -6.26212895e-01
7.22280622e-01 3.50984603e-01 -1.54529810e-01 1.35607257e-01
-5.78238249e-01 8.30581665e-01 1.73661664e-01 1.02968588e-01
1.16516483e+00 -2.30615944e-01 6.09056413e-01 -2.49520913e-01
9.13932264e-01 -5.98706119e-03 4.89891022e-01 7.05270946e-01
-1.01531279e+00 4.33320105e-01 1.29393116e-01 -8.82057309e-01
-7.33001888e-01 -8.18502009e-01 -4.74004224e-02 8.49322915e-01
3.48849669e-02 -4.38230485e-01 -9.32430029e-01 -9.20056581e-01
-7.46987835e-02 8.75891507e-01 -4.46543813e-01 -1.84383184e-01
-6.58027530e-01 -8.03179860e-01 8.64525557e-01 3.60299140e-01
7.93965876e-01 -7.46837378e-01 -1.01247203e+00 -7.49389455e-02
-2.83369333e-01 -1.47047901e+00 -4.31043893e-01 -5.65244555e-01
-4.88007933e-01 -1.31084502e+00 -5.14084339e-01 -1.71480566e-01
6.77015543e-01 5.66528797e-01 7.77769804e-01 9.43453908e-02
-2.68062353e-01 9.20849204e-01 -1.81810185e-01 -2.65901297e-01
-2.94154137e-01 1.00591250e-01 5.04479408e-01 5.34525454e-01
1.02156103e+00 -2.68288434e-01 -5.21782994e-01 4.95137662e-01
-7.85763562e-01 -4.57942814e-01 2.25102186e-01 5.23897521e-02
1.66277662e-02 -2.41545752e-01 2.11544439e-01 -4.87446487e-01
5.95670044e-01 -4.97533441e-01 -6.89197063e-01 2.50078291e-01
-4.85797763e-01 -3.15583587e-01 4.06813145e-01 -1.73718870e-01
-1.09870636e+00 3.28873724e-01 -1.06415629e-01 -8.44857022e-02
-9.34657454e-01 -3.42731297e-01 -3.68529260e-01 -1.25641719e-01
4.45170701e-01 3.58082205e-01 1.22090295e-01 -3.74501348e-01
-8.54384229e-02 8.38402867e-01 5.72396815e-01 -2.56663650e-01
5.85441053e-01 7.50782132e-01 -1.20394357e-01 -1.14297795e+00
-6.37923777e-01 -8.24609995e-01 -5.71681917e-01 -3.95677686e-01
9.94384706e-01 -9.08818781e-01 -1.08889139e+00 5.60922444e-01
-1.23344707e+00 4.94126350e-01 3.67684197e-03 3.77047628e-01
-4.79760706e-01 7.09334850e-01 -2.68739343e-01 -1.30570221e+00
-1.30607590e-01 -1.18481684e+00 1.16484809e+00 8.86781979e-03
-4.23192143e-01 -6.44378781e-01 4.97165434e-02 7.03075409e-01
4.73660856e-01 4.37237024e-01 1.10093862e-01 -1.05350947e+00
-4.93857026e-01 -7.03818440e-01 -2.00199649e-01 1.28738105e-01
1.53917119e-01 -2.61239797e-01 -1.17621958e+00 -5.14965773e-01
-6.75572529e-02 1.11693978e-01 6.76543891e-01 2.02775151e-01
8.12299848e-01 -6.45969033e-01 -5.70239186e-01 2.61668652e-01
8.81267250e-01 2.49879241e-01 8.21907759e-01 5.30644298e-01
4.49637741e-01 8.43708754e-01 6.45185292e-01 7.47981012e-01
6.11528873e-01 9.52963829e-01 3.56002897e-01 1.45837203e-01
3.99919719e-01 6.95564300e-02 6.03391588e-01 -1.29250556e-01
-3.25734675e-01 -3.36556852e-01 -8.22522342e-01 4.37078476e-01
-2.15719914e+00 -1.44966280e+00 -1.60800815e-01 2.52395225e+00
1.76975951e-01 3.77847329e-02 6.84803128e-01 1.80472225e-01
7.57629752e-01 2.66293555e-01 -3.62232029e-01 -2.66736269e-01
8.77920315e-02 -6.12874508e-01 7.99901187e-01 2.26322576e-01
-1.53164554e+00 7.10128009e-01 6.79754114e+00 2.85303116e-01
-8.60284626e-01 -9.89936516e-02 8.41741741e-01 -3.25514346e-01
3.01781952e-01 -3.08779866e-01 -1.19797969e+00 5.01158595e-01
1.13770962e+00 6.83174431e-02 2.38082781e-01 1.01733279e+00
5.80520272e-01 -2.14159504e-01 -9.42118287e-01 9.17485714e-01
2.47876123e-01 -8.70778203e-01 -6.89138379e-03 4.36626822e-01
1.58595532e-01 -1.96041167e-01 -9.59736705e-02 -7.00345635e-03
-8.23164582e-02 -5.77926219e-01 5.99441290e-01 6.03920341e-01
2.89647251e-01 -7.49743700e-01 7.16300011e-01 5.24759591e-01
-1.20784438e+00 -1.68073177e-01 -7.19417334e-02 -2.12282255e-01
4.41815495e-01 2.58393139e-01 -5.64106286e-01 3.18019122e-01
9.99285281e-01 5.22440553e-01 -7.68982589e-01 1.01341438e+00
3.95377636e-01 4.55759048e-01 -4.56624150e-01 9.41952243e-02
1.12743758e-01 -6.98147807e-03 8.87091875e-01 1.38672638e+00
1.69617906e-01 3.61150086e-01 3.44545931e-01 2.89266855e-01
3.88037950e-01 -8.46971851e-03 -1.12577045e+00 3.62583637e-01
1.38832316e-01 1.01470435e+00 -4.39837039e-01 -3.89282435e-01
-7.59085834e-01 8.41695189e-01 1.51882231e-01 2.05699429e-01
-8.89803529e-01 1.23891734e-01 1.25284183e+00 5.44744492e-01
1.60934672e-01 -3.45448732e-01 6.49795160e-02 -1.23895061e+00
3.43484610e-01 -1.02586377e+00 6.84643209e-01 -2.75640160e-01
-1.08165622e+00 7.52556384e-01 4.42411721e-01 -1.39114416e+00
-3.51391286e-01 -6.40631557e-01 -4.18287992e-01 4.34249282e-01
-1.27468538e+00 -1.14334404e+00 -3.81307602e-01 9.00118351e-01
3.43064249e-01 -4.12493795e-01 7.76382923e-01 5.19572437e-01
-8.06562066e-01 6.39086127e-01 -3.13032657e-01 1.93852454e-01
7.04148948e-01 -5.65540552e-01 5.80338001e-01 1.27477348e+00
-2.16706395e-01 6.43805981e-01 8.56778979e-01 -6.98869050e-01
-1.23415804e+00 -1.17700636e+00 9.14003313e-01 -7.66483903e-01
2.39088655e-01 -3.83722305e-01 -8.33422005e-01 8.19579065e-01
1.24954015e-01 1.20269343e-01 7.76951075e-01 -1.36449650e-01
-2.44263873e-01 -1.25877231e-01 -1.66570771e+00 6.59403741e-01
1.02836299e+00 -1.57477409e-01 -5.33306658e-01 2.09961191e-01
1.73476249e-01 -1.13340780e-01 -4.48199481e-01 2.57043898e-01
5.41188717e-01 -1.40976715e+00 1.03157544e+00 -8.54043543e-01
-5.23858428e-01 -4.21384484e-01 -3.68891597e-01 -7.25217044e-01
-3.81240815e-01 -5.82092881e-01 -2.35391855e-01 1.19169521e+00
5.25222793e-02 -9.94098783e-01 9.66764569e-01 1.54443157e+00
3.00241560e-01 -2.97644019e-01 -1.06206453e+00 -7.37472594e-01
-6.65626407e-01 -5.73849142e-01 5.59434950e-01 6.72205806e-01
-8.73232037e-02 -9.17483270e-02 -7.84515381e-01 5.67036688e-01
7.75799513e-01 -4.11880493e-01 9.95983958e-01 -1.19957089e+00
2.58692533e-01 -9.66566205e-02 -9.89468217e-01 -8.36599946e-01
9.59762260e-02 -2.16960341e-01 -2.56198019e-01 -9.59509134e-01
2.41857409e-01 1.60666749e-01 -1.49413913e-01 3.40603858e-01
2.53793355e-02 2.96982974e-01 1.21835195e-01 -3.06347013e-02
-7.71853447e-01 3.15284103e-01 3.67760390e-01 -1.04143620e-01
-3.67784016e-02 5.31612217e-01 -8.39843929e-01 8.31862330e-01
9.72041070e-01 -2.34770879e-01 -1.89168349e-01 -1.58915922e-01
-2.58532077e-01 -2.79995203e-01 8.19659889e-01 -1.19056284e+00
3.98092777e-01 -1.39347434e-01 2.86550879e-01 -5.27743995e-01
6.11791074e-01 -1.36416459e+00 8.25193971e-02 3.45279574e-01
-1.12377062e-01 2.53501534e-01 2.35339805e-01 7.11682081e-01
-1.43942058e-01 1.71519205e-01 7.16005206e-01 -6.75487891e-02
-8.00418019e-01 3.59535336e-01 -8.43040943e-01 -1.54954687e-01
1.51731539e+00 -5.81435680e-01 -4.61003959e-01 -7.01432884e-01
-6.63460612e-01 2.83591807e-01 6.18732035e-01 5.39643168e-01
5.79513311e-01 -1.16455770e+00 -5.10735452e-01 3.86894464e-01
3.71901631e-01 -5.77512443e-01 2.02797309e-01 8.75129461e-01
-1.13811232e-01 7.05934405e-01 -2.42671013e-01 -3.66897613e-01
-1.80933142e+00 7.89633036e-01 2.06354931e-01 -8.18477720e-02
-5.30647337e-01 1.14400186e-01 -1.05569132e-01 -1.03731103e-01
4.30842072e-01 -1.66115969e-01 -3.07577729e-01 -1.74999237e-02
1.02490783e+00 8.13736141e-01 1.22815289e-01 -1.28374040e+00
-8.55469763e-01 1.71499863e-01 -2.40902811e-01 1.32167071e-01
1.07864213e+00 -2.74358064e-01 2.64362782e-01 -1.62929073e-01
1.16305888e+00 3.33032124e-02 -1.23294723e+00 -4.45340686e-02
1.59282684e-01 -7.87141919e-01 -1.39051631e-01 -4.38870549e-01
-9.92731571e-01 3.40016842e-01 7.70363092e-01 6.39721870e-01
1.15156567e+00 -1.35269880e-01 8.13906908e-01 5.51474035e-01
5.42533219e-01 -9.26054955e-01 -6.18566215e-01 1.35559797e-01
5.76102853e-01 -1.56930351e+00 1.75431877e-01 -5.21725476e-01
-7.79571891e-01 7.82930315e-01 5.66527367e-01 1.26010671e-01
4.80022281e-01 4.07170840e-02 -2.42756270e-02 4.30258140e-02
-6.04325056e-01 -2.39216462e-01 3.60631317e-01 8.21165025e-01
1.38206437e-01 -4.74730432e-02 -6.90214485e-02 4.43306059e-01
1.55267701e-01 -5.11918683e-03 4.19397175e-01 1.14083612e+00
-4.50187564e-01 -1.01201832e+00 -6.08248234e-01 4.10860866e-01
-5.67352295e-01 2.05042422e-01 -6.38830423e-01 8.10145676e-01
-9.28277820e-02 1.19752610e+00 3.24433558e-02 -3.25115889e-01
8.04005325e-01 1.03240229e-01 -4.38382439e-02 -1.72285676e-01
-6.14212692e-01 -4.10303414e-01 5.66401124e-01 -9.25579906e-01
-6.68446004e-01 -1.25181198e+00 -5.73743641e-01 -5.31818807e-01
1.75454333e-01 -1.40962526e-01 2.66061693e-01 9.20828998e-01
5.75166404e-01 -3.76670286e-02 4.00495827e-01 -7.30596483e-01
-4.95532364e-01 -5.02959251e-01 -2.66181469e-01 5.37192643e-01
4.56983775e-01 -5.88940799e-01 -1.78248689e-01 1.00054957e-01] | [7.85410737991333, 1.1266200542449951] |
b6a51459-c12e-4d3d-a9f0-636ba006e6a2 | bring-your-own-data-self-supervised | 2306.13651 | null | https://arxiv.org/abs/2306.13651v2 | https://arxiv.org/pdf/2306.13651v2.pdf | Bring Your Own Data! Self-Supervised Evaluation for Large Language Models | With the rise of Large Language Models (LLMs) and their ubiquitous deployment in diverse domains, measuring language model behavior on realistic data is imperative. For example, a company deploying a client-facing chatbot must ensure that the model will not respond to client requests with profanity. Current evaluations approach this problem using small, domain-specific datasets with human-curated labels. These evaluation sets are often sampled from a narrow and simplified distribution, and data sources can unknowingly be leaked into the training set which can lead to misleading evaluations. To bypass these drawbacks, we propose a framework for self-supervised evaluation of LLMs by analyzing their sensitivity or invariance to transformations on the input text. Self-supervised evaluation can directly monitor LLM behavior on datasets collected in the wild or streamed during live model deployment. We demonstrate self-supervised evaluation strategies for measuring closed-book knowledge, toxicity, and long-range context dependence, in addition to sensitivity to grammatical structure and tokenization errors. When comparisons to similar human-labeled benchmarks are available, we find strong correlations between self-supervised and human-supervised evaluations. The self-supervised paradigm complements current evaluation strategies that rely on labeled data. | ['Tom Goldstein', 'Jonas Geiping', 'Micah Goldblum', 'Aniruddha Saha', 'Manli Shu', 'John Kirchenbauer', 'Yuxin Wen', 'Khalid Saifullah', 'Neel Jain'] | 2023-06-23 | null | null | null | null | ['chatbot', 'chatbot'] | ['methodology', 'natural-language-processing'] | [-1.51274085e-01 -1.38967363e-02 -3.09904367e-01 -7.29678333e-01
-1.12000787e+00 -1.01056552e+00 6.12079144e-01 2.94570237e-01
-6.09632194e-01 6.88079715e-01 8.26501325e-02 -4.29130733e-01
1.50830030e-01 -5.41312456e-01 -3.84213835e-01 -1.21418945e-01
1.92123890e-01 9.58783567e-01 1.11507669e-01 -2.76065379e-01
3.09711993e-01 2.45397359e-01 -1.16459608e+00 5.00293016e-01
9.82519865e-01 5.84863067e-01 7.79329706e-03 8.66171181e-01
-3.44146669e-01 1.42770195e+00 -1.22803617e+00 -6.34781897e-01
-1.96964249e-01 -2.35528499e-01 -1.25614583e+00 -1.77854955e-01
3.30991745e-01 -4.48128909e-01 4.45911922e-02 9.98007476e-01
5.04217386e-01 -9.24543291e-02 4.73423809e-01 -1.51692569e+00
-5.33996582e-01 9.61330295e-01 8.16609710e-03 1.44525364e-01
3.50401610e-01 4.11163211e-01 1.03339100e+00 -5.77634573e-01
8.68765950e-01 1.33096862e+00 7.55831301e-01 6.33464932e-01
-1.46455860e+00 -4.82199252e-01 -1.06362272e-02 -1.06973633e-01
-1.17777336e+00 -6.91163480e-01 4.31285471e-01 -5.29755950e-01
1.02408063e+00 1.61773339e-01 -1.76791176e-01 1.51544583e+00
-1.88168705e-01 7.37537324e-01 1.09788406e+00 -4.28092510e-01
3.45439345e-01 7.58003712e-01 4.78970975e-01 4.11919296e-01
1.08553357e-01 -4.61196117e-02 -7.27686822e-01 -5.95500588e-01
-7.45740766e-03 -3.69974256e-01 1.82454929e-01 -1.30652830e-01
-9.82224405e-01 7.68702447e-01 -2.65450478e-01 4.52688664e-01
-5.43794110e-02 -7.93501884e-02 7.61661053e-01 6.14055037e-01
5.56287587e-01 7.76403069e-01 -6.95961952e-01 -8.81146073e-01
-8.71708632e-01 3.24469805e-01 1.22714579e+00 9.96232152e-01
6.83219135e-01 -1.49724290e-01 -2.83111125e-01 9.52008665e-01
1.52252883e-01 6.25160754e-01 6.86056256e-01 -1.10364020e+00
7.45283186e-01 7.56666303e-01 3.17896485e-01 -6.87616885e-01
-3.38513881e-01 -3.28028888e-01 -1.74709722e-01 2.80249193e-02
7.73710132e-01 -6.41058087e-02 -2.41290331e-01 1.90796888e+00
7.48982653e-02 -4.67400193e-01 1.38285803e-02 4.14972901e-01
5.32965302e-01 1.10091671e-01 2.81479210e-01 -4.07153964e-02
8.32067490e-01 -6.47263288e-01 -5.61262727e-01 -3.28850001e-01
1.41950071e+00 -5.28538287e-01 1.86375821e+00 3.66942316e-01
-8.88421178e-01 -6.14211187e-02 -5.63659668e-01 3.84919755e-02
-2.67887533e-01 -3.28413576e-01 3.49548846e-01 8.12080503e-01
-8.38042855e-01 6.22318447e-01 -7.48262167e-01 -4.87357795e-01
3.22823673e-01 6.07852638e-02 -3.37851673e-01 -9.20827538e-02
-1.06335235e+00 1.15379763e+00 -2.69691851e-02 -1.79883748e-01
-9.54820633e-01 -4.77433383e-01 -7.02432692e-01 -4.14930172e-02
3.99506122e-01 1.56394288e-01 1.80364633e+00 -8.46259534e-01
-1.11588776e+00 9.06017780e-01 -1.57055050e-01 -3.06324214e-01
7.14726329e-01 -1.21373124e-02 -3.84695560e-01 -2.18327507e-01
2.75792450e-01 -2.96058203e-03 4.44384247e-01 -1.19297612e+00
-3.59149724e-01 -4.17900562e-01 7.71595240e-02 -4.16052295e-03
-5.79946280e-01 5.49822628e-01 1.87874347e-01 -9.54496488e-02
-4.46202427e-01 -6.77822113e-01 6.40087947e-02 -5.72263002e-01
-6.10264480e-01 -1.77430421e-01 8.08818996e-01 -5.59350431e-01
1.50109160e+00 -2.00661063e+00 -3.54466051e-01 2.57077366e-01
2.61687517e-01 5.86694002e-01 -3.93322825e-01 5.56305230e-01
3.68578941e-01 7.06690252e-01 -1.01724833e-01 -6.34232759e-01
3.37420076e-01 2.86683202e-01 -3.16380680e-01 1.63353413e-01
1.01634199e-02 9.70528960e-01 -1.06393313e+00 -8.15118790e-01
-2.92218477e-02 -1.92299369e-03 -5.70624650e-01 5.10518432e-01
-4.37405676e-01 2.20063865e-01 -1.63135722e-01 7.82679021e-01
2.87775576e-01 -2.80413091e-01 2.70963401e-01 4.48580354e-01
1.11426257e-01 7.50979066e-01 -7.58183539e-01 1.28226483e+00
-7.11574554e-01 6.73673391e-01 2.33741984e-01 -6.83704972e-01
8.40481341e-01 2.05950245e-01 -1.00981228e-01 -7.63994634e-01
1.00311570e-01 2.43942007e-01 1.55327218e-02 -6.69723392e-01
5.36934376e-01 -5.26223369e-02 -2.46164292e-01 1.10384166e+00
1.71572402e-01 -8.29656944e-02 3.51812899e-01 5.83311558e-01
1.43134117e+00 -2.61015862e-01 -1.36443928e-01 -9.85423550e-02
1.95647210e-01 1.77689984e-01 4.53760952e-01 1.08180153e+00
-3.64102453e-01 2.57705599e-01 8.86835992e-01 -2.06435397e-01
-9.34838235e-01 -8.84344578e-01 -1.06998935e-01 1.60070550e+00
-3.32128257e-01 -6.52740657e-01 -1.11135173e+00 -9.53058124e-01
-2.95742992e-02 1.07870042e+00 -2.36583099e-01 -4.18657929e-01
-3.19319278e-01 -3.37178856e-01 1.27202666e+00 4.87207383e-01
1.46569178e-01 -1.31291449e+00 -4.77315754e-01 3.05565238e-01
-4.48656827e-01 -1.11639273e+00 -2.67530501e-01 2.58463085e-01
-6.29868388e-01 -1.08136320e+00 1.04578547e-01 -4.41431075e-01
4.90061253e-01 -2.85184801e-01 1.57409847e+00 2.98269182e-01
-3.21760252e-02 5.06891549e-01 -3.70111406e-01 -4.04270977e-01
-1.17311227e+00 4.66390878e-01 2.58969128e-01 -2.40147412e-01
6.98323905e-01 -5.67367613e-01 3.29765342e-02 5.48118532e-01
-6.87406778e-01 -5.03574669e-01 1.68716803e-01 8.47088516e-01
-2.11311746e-02 -3.24477762e-01 6.61523759e-01 -1.37344766e+00
1.20484793e+00 -5.45461595e-01 -5.08249283e-01 5.37160158e-01
-7.43622184e-01 1.81174487e-01 7.79364645e-01 -5.77217519e-01
-9.86740291e-01 -4.50360060e-01 1.20927811e-01 -1.04751170e-01
-4.32814300e-01 6.45546794e-01 -1.92356855e-01 1.67329952e-01
1.12833035e+00 8.51959586e-02 1.34594455e-01 -6.12897635e-01
2.55622298e-01 1.19647515e+00 5.40294647e-01 -1.00543153e+00
6.24030828e-01 1.51082482e-02 -7.92360306e-01 -7.92324185e-01
-8.01576734e-01 -2.08280772e-01 -5.19465029e-01 -1.38905481e-01
2.22005203e-01 -5.00162661e-01 -8.44985843e-01 5.61811864e-01
-1.21871519e+00 -9.65579689e-01 -3.02823454e-01 1.72041774e-01
-4.29610401e-01 3.60138267e-01 -8.11855316e-01 -1.05748844e+00
-2.45683074e-01 -1.05780923e+00 8.30826223e-01 -2.09202468e-01
-9.25357044e-01 -1.17516553e+00 2.46852934e-01 5.64193904e-01
7.27131724e-01 -1.98854893e-01 1.15381539e+00 -1.37917495e+00
-1.12451583e-01 -5.97336650e-01 7.05051720e-02 7.76840806e-01
-9.18330103e-02 2.23189577e-01 -1.08455002e+00 -1.08309418e-01
-8.45531300e-02 -1.13115823e+00 4.05154750e-02 -3.14926714e-01
9.18137372e-01 -5.68741798e-01 6.52058795e-02 2.41277128e-01
7.69316137e-01 -1.10484540e-01 1.72081158e-01 2.18973100e-01
5.02152860e-01 6.94836736e-01 4.75432187e-01 4.72867489e-01
5.08720577e-01 4.98047620e-01 1.75274655e-01 3.45568150e-01
3.10885578e-01 -6.30383015e-01 5.22334933e-01 6.83496773e-01
4.85606462e-01 -4.48452145e-01 -1.34352207e+00 4.27428514e-01
-1.65783417e+00 -9.50326920e-01 -1.72236643e-03 2.32154775e+00
1.11251605e+00 3.56521785e-01 1.45033151e-01 1.61221530e-02
5.40395975e-01 4.63117519e-03 -6.59591079e-01 -7.39080787e-01
-1.82928607e-01 -4.58070226e-02 3.75880361e-01 7.68711448e-01
-5.84828854e-01 9.17149186e-01 7.11425924e+00 6.36649609e-01
-1.15738714e+00 4.06315327e-01 7.94897676e-01 -2.21503377e-01
-5.59089363e-01 1.11743428e-01 -7.30553508e-01 6.56301439e-01
1.39916623e+00 -3.17667633e-01 6.65031195e-01 1.03005838e+00
2.24024311e-01 -1.94629878e-01 -1.42874682e+00 8.30711842e-01
2.83012912e-02 -9.21773851e-01 -4.04597372e-01 9.79641974e-02
3.52498323e-01 3.99043113e-01 -7.01736510e-02 6.80164993e-01
8.36321592e-01 -1.18829632e+00 8.36254179e-01 3.20414662e-01
8.78834784e-01 -4.79852825e-01 7.52470016e-01 1.11323893e+00
-4.05155718e-01 -1.26929551e-01 -6.34218752e-02 -2.86529154e-01
1.11219369e-01 3.47350001e-01 -1.04255927e+00 -3.17385256e-01
3.93761069e-01 2.93823719e-01 -8.55152845e-01 2.86399215e-01
-1.89616680e-01 1.14995944e+00 -4.28329229e-01 -2.49648616e-01
1.48312682e-02 1.18223175e-01 3.13714802e-01 1.25226974e+00
-2.92091429e-01 -3.60794395e-01 -2.26064101e-02 1.15651965e+00
-1.33064047e-01 1.15735047e-01 -7.26478100e-01 -5.11063457e-01
9.07180369e-01 1.14858687e+00 -2.34312654e-01 -3.50571007e-01
-2.20012903e-01 6.74566984e-01 7.79284954e-01 4.50656682e-01
-5.48260152e-01 -3.20171744e-01 6.44118309e-01 2.34205991e-01
-4.12554860e-01 -1.80858210e-01 -4.36972350e-01 -1.21274137e+00
2.42896259e-01 -1.32097018e+00 2.06559151e-01 -5.69823444e-01
-1.46420813e+00 7.24514484e-01 2.44814754e-02 -6.56688631e-01
-7.04407513e-01 -4.46577460e-01 -6.01046085e-01 7.81434953e-01
-1.03892803e+00 -8.81728411e-01 -9.92466807e-02 4.89777684e-01
1.72925323e-01 -3.49243999e-01 1.07516360e+00 2.84594148e-01
-5.96766412e-01 1.13512754e+00 1.68702275e-01 3.83078814e-01
8.93680751e-01 -1.12608314e+00 3.87424678e-01 5.16054034e-01
2.73516059e-01 8.33755374e-01 5.93706846e-01 -5.68570554e-01
-1.02030802e+00 -8.69794905e-01 1.40775669e+00 -1.26398814e+00
7.90504932e-01 -7.99796581e-01 -1.16211104e+00 8.21199179e-01
-2.64992714e-01 -2.63221785e-02 6.78781986e-01 5.33814371e-01
-7.64259577e-01 3.19918431e-02 -1.25265622e+00 4.14936602e-01
9.73257899e-01 -1.17415094e+00 -4.40632164e-01 5.65932155e-01
4.71611977e-01 -1.80871397e-01 -6.84119523e-01 -1.39198042e-02
3.92991632e-01 -9.77018595e-01 2.81912327e-01 -1.25165570e+00
3.22798997e-01 1.07333951e-01 -2.61953712e-01 -1.28925014e+00
1.42158717e-01 -8.27238202e-01 -7.32469261e-02 1.61481559e+00
7.17221320e-01 -5.68040907e-01 6.81964815e-01 1.30228436e+00
2.44698808e-01 -3.85638237e-01 -7.21268773e-01 -9.55819964e-01
2.96388298e-01 -7.42358983e-01 7.11588740e-01 1.05496609e+00
5.53991497e-01 4.20755476e-01 -1.48472950e-01 -1.54581025e-01
5.05738556e-01 -3.70125622e-01 9.80952084e-01 -1.06158173e+00
-1.63256794e-01 -3.13660860e-01 -2.56370455e-01 -5.68508565e-01
4.91397440e-01 -9.37196493e-01 1.54854298e-01 -1.14308417e+00
2.33460203e-01 -7.88779378e-01 1.28894066e-02 5.87337017e-01
7.97298774e-02 -8.17821100e-02 1.01995967e-01 1.67985305e-01
-1.06761336e+00 3.10758978e-01 4.72573727e-01 1.13615327e-01
-1.33163229e-01 -5.59322275e-02 -6.09416842e-01 7.97617018e-01
8.77584517e-01 -7.39897549e-01 -3.92229766e-01 -6.03365541e-01
4.50522065e-01 -1.48520514e-01 2.41280809e-01 -7.61635065e-01
3.04272652e-01 -2.27196157e-01 -3.58641773e-01 -1.06904805e-01
-6.92692176e-02 -5.59247613e-01 -3.28092486e-01 -2.66069602e-02
-9.18023050e-01 1.90923601e-01 -8.90763178e-02 1.55627087e-01
-2.56565481e-01 -4.63983715e-01 6.16720259e-01 -2.14011833e-01
-3.06687385e-01 -2.15637852e-02 -5.83377421e-01 7.76049733e-01
5.28242588e-01 -6.83182990e-03 -5.98531783e-01 -6.95733786e-01
-4.42545593e-01 3.39967906e-01 6.35656774e-01 2.83534855e-01
1.45950511e-01 -1.06826639e+00 -5.91280699e-01 1.62156761e-01
4.68198806e-01 -6.18495159e-02 -4.04818431e-02 6.41254663e-01
-4.27926451e-01 2.42014095e-01 1.87635019e-01 -5.59752405e-01
-1.14291239e+00 2.43287370e-01 4.84756023e-01 -4.04430985e-01
6.45642728e-02 6.93521440e-01 -3.46673489e-01 -1.17928123e+00
5.58423579e-01 -1.98269352e-01 3.36178482e-01 -9.42484289e-02
5.39709806e-01 3.90371323e-01 2.96094656e-01 -5.45494139e-01
-3.00122529e-01 -3.06166649e-01 -1.84857994e-01 -3.50977600e-01
1.10691416e+00 -5.11901118e-02 -1.62954748e-01 7.89809227e-01
1.24045372e+00 9.24594775e-02 -9.42243218e-01 -5.86594820e-01
4.28729892e-01 -2.60120630e-01 -3.21843952e-01 -1.08476269e+00
-4.95215625e-01 9.33049858e-01 7.88214281e-02 3.39878649e-01
3.77108902e-01 1.08608097e-01 5.73916376e-01 9.40427184e-01
5.52407384e-01 -1.41371679e+00 1.84842944e-01 7.37620652e-01
4.69595313e-01 -1.47448349e+00 -5.09503365e-01 1.02658987e-01
-9.69439387e-01 6.09447300e-01 8.78876984e-01 4.12061125e-01
3.89882386e-01 5.96074104e-01 5.40391445e-01 -1.11823387e-01
-1.15495121e+00 4.17438954e-01 -3.74938488e-01 6.94168210e-01
5.70127606e-01 2.69761235e-01 1.24543421e-01 8.82033050e-01
-4.04454827e-01 -1.40287742e-01 5.41905761e-01 9.98908222e-01
-2.25418553e-01 -1.26414120e+00 -2.05632478e-01 6.17066681e-01
-4.20250893e-01 -3.20308208e-01 -7.58205891e-01 6.32590771e-01
-3.33668172e-01 1.20786822e+00 -1.84507258e-02 -4.70653236e-01
4.64255244e-01 5.73144078e-01 7.57067502e-02 -9.11830902e-01
-8.61574233e-01 -5.46517015e-01 6.13222837e-01 -6.42529011e-01
4.44504805e-02 -7.23934054e-01 -9.65227306e-01 -5.78764677e-01
-3.65685493e-01 3.60772729e-01 6.02064848e-01 1.07633448e+00
3.83239776e-01 -2.46865049e-01 6.12524271e-01 -3.02750528e-01
-1.37992263e+00 -1.31255090e+00 -5.69240451e-01 8.48013580e-01
1.57930255e-01 -2.39421070e-01 -5.81564128e-01 -1.08944729e-01] | [12.588116645812988, 7.917078495025635] |
18a8dc4b-4708-482e-b017-07ab0f777ae7 | triplenet-triple-attention-network-for-multi | 1909.10666 | null | https://arxiv.org/abs/1909.10666v2 | https://arxiv.org/pdf/1909.10666v2.pdf | TripleNet: Triple Attention Network for Multi-Turn Response Selection in Retrieval-based Chatbots | We consider the importance of different utterances in the context for selecting the response usually depends on the current query. In this paper, we propose the model TripleNet to fully model the task with the triple <context, query, response> instead of <context, response> in previous works. The heart of TripleNet is a novel attention mechanism named triple attention to model the relationships within the triple at four levels. The new mechanism updates the representation for each element based on the attention with the other two concurrently and symmetrically. We match the triple <C, Q, R> centered on the response from char to context level for prediction. Experimental results on two large-scale multi-turn response selection datasets show that the proposed model can significantly outperform the state-of-the-art methods. TripleNet source code is available at https://github.com/wtma/TripleNet | ['Wei-Nan Zhang', 'Shijin Wang', 'Yiming Cui', 'Ting Liu', 'Su He', 'Nan Shao', 'Guoping Hu', 'Wentao Ma'] | 2019-09-24 | triplenet-triple-attention-network-for-multi-1 | https://aclanthology.org/K19-1069 | https://aclanthology.org/K19-1069.pdf | conll-2019-11 | ['conversational-response-selection'] | ['natural-language-processing'] | [ 6.91642538e-02 -2.29807898e-01 -3.12925935e-01 -7.38672733e-01
-1.04841626e+00 -2.41420791e-01 2.88050503e-01 1.62632734e-01
-4.53435898e-01 4.37811464e-01 6.26758039e-01 -2.01409563e-01
6.72533214e-02 -4.16938603e-01 -5.31117558e-01 -3.62972379e-01
5.33017278e-01 5.85028291e-01 5.26112497e-01 -5.35468757e-01
4.56845105e-01 -5.53978011e-02 -1.44969356e+00 9.02383268e-01
6.57673240e-01 1.16110396e+00 5.36113620e-01 6.35467350e-01
-3.47330749e-01 1.01134706e+00 -2.83767194e-01 -2.83520788e-01
-9.39630195e-02 -3.72513831e-01 -1.23170829e+00 -4.23724055e-01
3.03611845e-01 -3.17454606e-01 -3.93043309e-01 6.50219977e-01
6.95604622e-01 4.27775055e-01 5.50025292e-02 -1.09506416e+00
-8.97237122e-01 7.20597088e-01 -3.85830969e-01 5.36090076e-01
8.00114512e-01 8.80849585e-02 1.55651379e+00 -1.39590740e+00
4.75102961e-01 1.46885991e+00 2.28271782e-01 3.59730631e-01
-1.04232550e+00 -6.85732007e-01 7.41354764e-01 7.78356254e-01
-1.34223235e+00 -3.16086352e-01 7.12825894e-01 -1.84761360e-01
1.37935078e+00 6.26780570e-01 3.58492047e-01 9.59318697e-01
-9.69960988e-02 1.05967247e+00 8.79468620e-01 -4.02684957e-01
-3.61469239e-01 3.16898078e-02 6.61322892e-01 3.36190581e-01
-6.74012482e-01 -1.71272188e-01 -6.24384820e-01 -3.57160538e-01
2.32131809e-01 -2.35333052e-02 -3.16594154e-01 7.02745989e-02
-1.13518798e+00 8.12385082e-01 6.02598250e-01 2.76013315e-01
-6.64035618e-01 2.71984249e-01 3.65941763e-01 4.20830905e-01
4.06175822e-01 1.95477352e-01 -7.64630497e-01 -7.57060423e-02
-9.62681547e-02 3.39585781e-01 6.73968732e-01 1.16042531e+00
8.84071648e-01 -2.58579433e-01 -7.84520209e-01 1.04908657e+00
4.25087452e-01 1.97325438e-01 4.25730884e-01 -6.66967273e-01
6.99584305e-01 7.95041740e-01 3.31986368e-01 -8.75614464e-01
-4.70962942e-01 -2.73317099e-01 -4.30279762e-01 -6.55916274e-01
-1.48228496e-01 2.43914463e-02 -8.08354259e-01 1.70411623e+00
7.18320310e-01 7.57308528e-02 -3.82483751e-01 1.16192126e+00
1.30815959e+00 7.31871963e-01 3.71609002e-01 -3.75150293e-02
1.31810987e+00 -1.30537891e+00 -6.85811520e-01 -2.04059660e-01
5.06895781e-01 -9.72411454e-01 1.42920685e+00 -3.82721100e-06
-8.69552016e-01 -7.64876068e-01 -4.93156821e-01 -4.59893674e-01
-1.98814332e-01 1.13603741e-01 3.78253907e-01 1.68343261e-02
-1.04643095e+00 1.37784996e-03 -1.60157993e-01 -3.93482864e-01
-2.82600373e-01 4.85186905e-01 -1.14951342e-01 -1.17097676e-01
-1.87842739e+00 8.40068400e-01 2.37747893e-01 1.01495497e-01
-7.00612307e-01 -4.41401780e-01 -6.18662894e-01 2.37663060e-01
6.76665187e-01 -7.35313475e-01 1.73146212e+00 -1.01134348e+00
-1.47358453e+00 7.62677848e-01 -5.45356274e-01 -4.33614440e-02
1.34922147e-01 -4.74186182e-01 -3.41995776e-01 -8.69776607e-02
2.45208163e-02 7.62788534e-01 4.94305521e-01 -1.19574356e+00
-8.47196877e-01 -2.32989311e-01 4.27946508e-01 5.01094222e-01
1.06684729e-01 6.45843267e-01 -9.71569657e-01 -2.89785981e-01
2.73594987e-02 -9.90646780e-01 -1.61549404e-01 -6.07706785e-01
-2.79524773e-01 -7.96054542e-01 4.68008131e-01 -5.48972607e-01
1.86634982e+00 -1.79361522e+00 7.86054805e-02 -4.62710708e-02
-1.40545303e-02 2.44236544e-01 -3.08326393e-01 9.02583420e-01
-2.60811836e-01 2.47494057e-01 1.26572028e-01 -2.60110199e-01
4.70138900e-02 2.93067340e-02 -3.72386038e-01 1.66410208e-01
1.09771043e-01 8.59731257e-01 -9.03484285e-01 -4.79325682e-01
-5.04982397e-02 3.84062290e-01 -7.40798712e-01 5.77390015e-01
-4.36434239e-01 4.56356913e-01 -8.26734662e-01 6.52426958e-01
4.06297117e-01 -5.08156955e-01 3.88742447e-01 -3.03132236e-01
-1.40776932e-01 8.21067154e-01 -9.41376746e-01 1.35659957e+00
-4.50784266e-01 2.25862920e-01 2.64746130e-01 -7.58110940e-01
7.85909653e-01 4.54939663e-01 4.44830537e-01 -1.03250134e+00
1.47402599e-01 -2.15011401e-04 2.11581990e-01 -7.25477338e-01
8.37782741e-01 3.25442016e-01 -3.87553126e-01 5.26761413e-01
4.23378907e-02 2.67299771e-01 1.05329715e-01 3.71845603e-01
7.19923675e-01 2.40837224e-02 4.58424449e-01 -1.29995614e-01
8.10136974e-01 -3.30192685e-01 8.25195491e-01 9.49721038e-01
-2.88162827e-01 3.59956980e-01 4.77795392e-01 -5.88181794e-01
-6.00256324e-01 -3.77382815e-01 2.73408681e-01 2.00117731e+00
2.69700319e-01 -5.36217213e-01 -3.48279148e-01 -6.74991906e-01
-3.25883366e-02 6.22928917e-01 -7.90725768e-01 -4.07378841e-03
-8.23863685e-01 -3.01399916e-01 1.72260702e-01 5.78290045e-01
2.75963217e-01 -1.49795401e+00 -3.31438482e-01 3.03578496e-01
-8.03930819e-01 -9.53796983e-01 -1.06363702e+00 -1.52277416e-02
-2.02761844e-01 -9.47515607e-01 -1.79452151e-01 -6.93406343e-01
2.33581394e-01 4.06648666e-01 1.41574025e+00 6.22889638e-01
2.37476394e-01 4.22797829e-01 -7.79211044e-01 -3.26599836e-01
4.45587300e-02 2.47482851e-01 -1.73308626e-01 4.24483865e-01
4.54850674e-01 -2.03777686e-01 -7.50384092e-01 6.42832816e-01
-8.27332020e-01 9.41565707e-02 3.52451801e-01 9.48638558e-01
8.27679515e-01 -7.50628412e-01 7.77509809e-01 -1.08530343e+00
7.76046932e-01 -8.39859009e-01 -1.86274499e-01 5.10867059e-01
-4.80137199e-01 -2.11070642e-01 4.32200491e-01 -3.81171048e-01
-9.96691465e-01 -6.27835914e-02 -4.25308257e-01 -1.90156549e-01
-1.26053333e-01 8.10594738e-01 -1.11766487e-01 3.02811623e-01
2.18579754e-01 1.31891847e-01 -6.68987572e-01 -6.10768497e-01
3.68052661e-01 7.07673073e-01 5.32496125e-02 -6.78701818e-01
-6.79824501e-02 -3.02557787e-03 -6.63906872e-01 -2.44324937e-01
-8.99650455e-01 -7.49894381e-01 -3.56917739e-01 -2.88509846e-01
7.52736628e-01 -7.50162601e-01 -8.84844124e-01 2.99890131e-01
-1.40792310e+00 -3.46381038e-01 1.99980810e-01 3.88266474e-01
-3.65721405e-01 7.43275136e-02 -6.60864532e-01 -7.86817253e-01
-5.63839495e-01 -1.35486376e+00 8.92573237e-01 2.65815556e-01
-1.25373483e-01 -6.77269101e-01 -3.57039794e-02 5.00750363e-01
5.24586737e-01 -4.24313277e-01 1.07927346e+00 -1.01999664e+00
-7.37378597e-01 -8.76848237e-04 -3.09063405e-01 -1.81121990e-01
-1.06088892e-01 4.33274219e-03 -8.65998745e-01 -1.48373097e-01
-1.43213645e-01 -5.31235814e-01 1.08043575e+00 1.25720277e-01
1.38817549e+00 -4.27002490e-01 -4.05294716e-01 2.58562982e-01
1.31793582e+00 5.12118101e-01 5.34009874e-01 5.56849577e-02
6.96598828e-01 5.95993578e-01 9.31501031e-01 4.82350796e-01
8.12651098e-01 1.13481438e+00 5.44804871e-01 8.10425505e-02
1.74359024e-01 -2.26135761e-01 1.51746213e-01 9.45451796e-01
3.09432805e-01 -7.33677387e-01 -9.93489981e-01 6.60170317e-01
-2.10688066e+00 -1.11483014e+00 -1.65387228e-01 2.09912658e+00
8.60453546e-01 -8.08637664e-02 7.95024186e-02 -4.12538350e-01
8.18765402e-01 3.86968046e-01 -4.88239527e-01 -6.92059338e-01
1.54511169e-01 -2.04170849e-02 -6.29370585e-02 9.22128737e-01
-9.82573807e-01 1.08865738e+00 5.96361208e+00 7.62009740e-01
-1.11212695e+00 2.44565874e-01 5.53135931e-01 -7.64682814e-02
-6.34402692e-01 1.60801277e-01 -1.10994053e+00 4.89474565e-01
8.16944599e-01 -2.20419213e-01 3.53093773e-01 5.41624606e-01
1.71965629e-01 -8.31610709e-02 -1.00363624e+00 5.70633292e-01
1.68646529e-01 -9.68771160e-01 2.08977088e-01 -5.21646976e-01
3.29372287e-01 1.17554471e-01 1.28481969e-01 7.43714154e-01
4.26159889e-01 -7.34648645e-01 6.25542700e-01 5.22748470e-01
6.96117163e-01 -5.34952760e-01 6.77574277e-01 3.29761773e-01
-1.53299224e+00 -3.61406267e-01 -9.77868140e-02 5.85569292e-02
2.67151982e-01 1.19976126e-01 -8.41722012e-01 7.13184834e-01
8.27559471e-01 4.05364484e-01 -3.31907064e-01 8.51412416e-01
-3.26586843e-01 6.31656229e-01 -1.60683528e-01 -2.37174794e-01
3.31609100e-01 5.92322834e-02 5.59065282e-01 1.38679290e+00
1.51591286e-01 5.86759984e-01 6.72357321e-01 5.24010599e-01
-2.24954128e-01 4.51747209e-01 -1.41055286e-01 3.21297795e-01
8.32198501e-01 1.08875930e+00 -2.25014195e-01 -4.27928597e-01
-6.50686979e-01 5.62439442e-01 6.85788333e-01 5.06537318e-01
-1.02791834e+00 -3.00578505e-01 6.22925758e-01 2.98469290e-02
5.07967532e-01 3.54210109e-01 1.75787061e-01 -8.05627823e-01
1.51697174e-01 -1.05915010e+00 7.81029403e-01 -9.62609589e-01
-1.12815988e+00 9.64650214e-01 2.47092173e-02 -1.09430671e+00
-1.72983527e-01 -1.52900979e-01 -5.88186860e-01 1.20485258e+00
-1.34357297e+00 -1.33427787e+00 -1.25466943e-01 6.74033165e-01
8.32075536e-01 2.42808491e-01 8.65512908e-01 4.94809300e-01
-6.56677008e-01 7.78056860e-01 -3.28509331e-01 -2.39175614e-02
8.48528624e-01 -9.00755823e-01 3.63817364e-01 5.16036570e-01
-5.85690215e-02 8.14008176e-01 4.00004685e-01 -5.75912654e-01
-1.05283022e+00 -7.81050205e-01 1.53705812e+00 -3.61027300e-01
4.79809761e-01 -2.43781164e-01 -1.07789481e+00 8.88269722e-01
6.48086905e-01 -1.34399787e-01 8.98120880e-01 3.46675158e-01
-5.61602414e-01 -1.72480494e-01 -7.46718287e-01 3.66443813e-01
8.64144087e-01 -7.17172801e-01 -4.07326579e-01 2.30712637e-01
1.26749289e+00 -6.98758483e-01 -6.60555065e-01 4.45567310e-01
6.71934783e-01 -8.41916323e-01 9.13952172e-01 -9.47650075e-01
3.45483631e-01 -1.80194795e-01 -4.46103811e-01 -1.28837144e+00
-7.41882145e-01 -5.82804024e-01 -2.50437856e-01 1.18750894e+00
6.11260951e-01 -5.49291372e-01 3.63978684e-01 6.62815273e-01
-4.38024014e-01 -1.35012722e+00 -9.38258886e-01 -9.88847017e-02
-2.48009071e-01 -3.01003516e-01 9.33236539e-01 8.38690579e-01
7.80029735e-03 7.27605820e-01 -6.01149142e-01 6.89164922e-03
-1.27716422e-01 5.92378974e-01 5.29550731e-01 -9.32060301e-01
-3.35555136e-01 -4.13046390e-01 1.68073207e-01 -1.50388885e+00
2.12379038e-01 -8.63441408e-01 2.26426143e-02 -1.58827853e+00
2.88918287e-01 -5.07904291e-01 -7.75460660e-01 5.73448300e-01
-7.69351542e-01 -3.21912646e-01 4.28115457e-01 1.56130835e-01
-1.05098760e+00 6.35899425e-01 1.09577870e+00 -2.12920547e-01
-2.39603385e-01 1.28184766e-01 -9.53748822e-01 3.70709240e-01
8.21915627e-01 -6.32732809e-01 -4.12931919e-01 -9.41846311e-01
2.86336899e-01 4.14063215e-01 -6.41580522e-02 -5.13025761e-01
4.93289083e-01 -4.76011842e-01 -7.04979748e-02 -8.45071137e-01
4.51606452e-01 -7.73980081e-01 1.81482527e-02 8.67016092e-02
-9.49478924e-01 5.75405478e-01 4.80877683e-02 5.77768385e-01
-3.64835024e-01 7.83824325e-02 5.47499537e-01 -1.02593414e-01
-8.51721942e-01 5.71218610e-01 -1.70320392e-01 1.59542158e-01
6.08652949e-01 3.05152774e-01 -3.72208446e-01 -6.93897188e-01
-6.71585321e-01 8.15813780e-01 -7.26679564e-02 7.77079701e-01
8.51278424e-01 -1.43443811e+00 -8.50653887e-01 -3.82777341e-02
4.48806167e-01 -2.40158439e-01 6.53987408e-01 9.76217389e-01
6.80203661e-02 8.11827540e-01 4.25463030e-03 -3.00923914e-01
-1.50170970e+00 6.19980931e-01 3.78248096e-01 -7.35310912e-01
-3.68704461e-03 1.33702230e+00 4.57168669e-01 -6.06368124e-01
2.42293894e-01 -1.96594536e-01 -5.80120146e-01 -4.89984676e-02
5.70492625e-01 -2.42952798e-02 3.16638090e-02 -1.00581944e+00
-6.13204896e-01 4.01355982e-01 -5.67020535e-01 -3.74512449e-02
9.95695949e-01 -3.19082558e-01 -2.14827821e-01 4.30935800e-01
1.12964535e+00 -1.21887922e-01 -8.79752398e-01 -6.76623464e-01
-4.50222529e-02 -7.08560824e-01 -2.70603627e-01 -1.00991213e+00
-1.04605532e+00 6.86331332e-01 3.07119578e-01 2.44884491e-01
1.20248950e+00 6.76224530e-02 7.31075644e-01 3.46839577e-01
3.20012033e-01 -1.11796582e+00 1.06506579e-01 9.18554485e-01
1.41310024e+00 -1.17140293e+00 -1.62726671e-01 -3.32767040e-01
-1.04275930e+00 4.92029607e-01 1.21076941e+00 1.27338722e-01
6.46563530e-01 -2.23207906e-01 1.98706821e-01 -1.95393592e-01
-1.59890056e+00 -3.59030157e-01 3.18638504e-01 2.54163384e-01
1.01906562e+00 8.77456516e-02 -6.27588153e-01 7.41460502e-01
3.64366561e-01 -9.86075401e-02 1.55758888e-01 9.00590956e-01
-4.56083685e-01 -1.45949018e+00 -1.07809439e-01 4.47785467e-01
-5.34865677e-01 -3.13963443e-01 -6.99402332e-01 5.03658831e-01
9.18412115e-03 1.28012359e+00 -4.21442650e-02 -8.84211659e-01
7.64796793e-01 2.26727441e-01 -1.42075801e-02 -6.47259951e-01
-1.28151524e+00 1.19618684e-01 3.01110238e-01 -7.42279887e-01
-2.95863062e-01 -5.34232378e-01 -1.20531666e+00 -1.72667295e-01
-3.22424352e-01 2.45523900e-01 1.12867236e-01 6.53946519e-01
7.43740857e-01 4.49378163e-01 8.83968771e-01 -5.74130714e-01
-5.86235344e-01 -1.19752228e+00 -1.84818171e-02 5.31202018e-01
3.87678862e-01 -8.02482247e-01 1.55383497e-01 -3.95650864e-01] | [12.362241744995117, 7.837774753570557] |
9f0beeb1-2ff1-4520-93bc-b13829d00d7a | ssl4eo-s12-a-large-scale-multi-modal-multi | 2211.07044 | null | https://arxiv.org/abs/2211.07044v2 | https://arxiv.org/pdf/2211.07044v2.pdf | SSL4EO-S12: A Large-Scale Multi-Modal, Multi-Temporal Dataset for Self-Supervised Learning in Earth Observation | Self-supervised pre-training bears potential to generate expressive representations without human annotation. Most pre-training in Earth observation (EO) are based on ImageNet or medium-size, labeled remote sensing (RS) datasets. We share an unlabeled RS dataset SSL4EO-S12 (Self-Supervised Learning for Earth Observation - Sentinel-1/2) to assemble a large-scale, global, multimodal, and multi-seasonal corpus of satellite imagery from the ESA Sentinel-1 \& -2 satellite missions. For EO applications we demonstrate SSL4EO-S12 to succeed in self-supervised pre-training for a set of methods: MoCo-v2, DINO, MAE, and data2vec. Resulting models yield downstream performance close to, or surpassing accuracy measures of supervised learning. In addition, pre-training on SSL4EO-S12 excels compared to existing datasets. We make openly available the dataset, related source code, and pre-trained models at https://github.com/zhu-xlab/SSL4EO-S12. | ['Xiao Xiang Zhu', 'Conrad M Albrecht', 'Chenying Liu', 'Zhitong Xiong', 'Nassim Ait Ali Braham', 'Yi Wang'] | 2022-11-13 | null | null | null | null | ['multi-label-image-classification'] | ['computer-vision'] | [ 2.66418725e-01 1.77144215e-01 -1.16898753e-01 -7.77256846e-01
-9.49807286e-01 -6.19562745e-01 6.76648200e-01 1.96575105e-01
-4.39093918e-01 7.16591418e-01 4.02387708e-01 -7.26831555e-01
-5.78224175e-02 -9.55272079e-01 -6.05706871e-01 -4.85036612e-01
-7.61730134e-01 2.77684689e-01 -4.63328987e-01 -6.42569304e-01
-3.30515027e-01 4.26624954e-01 -1.47230899e+00 3.66879016e-01
7.60137022e-01 1.12948227e+00 1.58874005e-01 8.96257102e-01
1.83651134e-01 7.35286474e-01 1.60244912e-01 2.14003503e-01
5.73722243e-01 -2.64108926e-01 -1.01375520e+00 -8.87713134e-02
5.01743197e-01 -1.80536762e-01 -2.25185543e-01 9.59643185e-01
5.71685553e-01 2.31972545e-01 6.81826353e-01 -1.09303629e+00
-6.59888029e-01 5.67725301e-01 -3.14080536e-01 4.63287950e-01
-1.86065346e-01 1.80851385e-01 1.14074445e+00 -1.02121377e+00
6.04978800e-01 9.34036553e-01 9.05135930e-01 4.52105522e-01
-1.20228815e+00 -7.59269118e-01 -2.58089006e-01 -3.93364757e-01
-1.54583037e+00 -6.38403177e-01 1.22630425e-01 -6.46463096e-01
9.67649102e-01 2.84635752e-01 2.92023480e-01 8.90175879e-01
-1.90635383e-01 7.03021646e-01 1.35181391e+00 -3.13357621e-01
1.03842147e-01 -2.00479664e-02 1.30341589e-01 5.12430668e-01
-6.70947284e-02 5.08664668e-01 -7.01462552e-02 3.49539705e-02
6.69209838e-01 5.94602451e-02 -3.44018936e-01 3.19243908e-01
-1.28928876e+00 1.00750208e+00 9.92687345e-01 2.78662771e-01
-5.02097845e-01 1.40717281e-02 2.79555917e-01 6.50355458e-01
9.53958333e-01 5.69977820e-01 -8.76924038e-01 2.59065181e-01
-1.07629049e+00 1.74185969e-02 4.11303788e-01 8.80039036e-01
1.14175427e+00 4.20895249e-01 3.01485777e-01 7.02691257e-01
3.42433363e-01 9.99939978e-01 3.95841718e-01 -7.71676958e-01
4.43230659e-01 4.76366669e-01 2.07182929e-01 -7.45014668e-01
-8.22103858e-01 -6.60442829e-01 -1.11688995e+00 5.59004955e-02
-2.84932882e-01 -7.60977089e-01 -1.09771693e+00 1.57265806e+00
1.87318936e-01 2.33114082e-02 5.84913552e-01 8.93596292e-01
1.40552294e+00 1.15503132e+00 3.48065257e-01 2.68022537e-01
1.01698768e+00 -9.86811638e-01 -2.92486310e-01 -7.59181142e-01
9.78897929e-01 -2.05524161e-01 8.12089503e-01 -1.46679044e-01
-2.98012435e-01 -7.68461466e-01 -6.82583928e-01 3.11626107e-01
-7.69868970e-01 4.59116578e-01 9.36531365e-01 1.61179140e-01
-1.23310006e+00 5.77537477e-01 -8.09474409e-01 -7.26681948e-01
5.38172007e-01 3.06395465e-03 -6.96574926e-01 4.76264209e-02
-1.48441684e+00 6.96131170e-01 8.93859863e-01 2.01389685e-01
-1.36312950e+00 -8.26774180e-01 -1.20949435e+00 -1.17181070e-01
4.16538641e-02 -3.57480764e-01 1.00262868e+00 -1.29656863e+00
-8.53315830e-01 1.19963908e+00 3.20326835e-01 -5.84838033e-01
1.37863150e-02 -3.35531473e-01 -1.00229001e+00 1.50812477e-01
3.89202088e-01 1.30216646e+00 5.03504574e-01 -1.34465301e+00
-5.52303374e-01 -2.86253005e-01 -8.72782543e-02 2.72088051e-01
-2.44989857e-01 8.83282162e-03 1.28854543e-01 -7.08986282e-01
1.46946602e-03 -9.89043772e-01 -6.73415542e-01 -9.82732028e-02
-1.96977422e-01 3.89971770e-02 4.44488615e-01 -7.24184334e-01
8.97735059e-01 -2.46382141e+00 -4.28459018e-01 6.67627752e-02
8.16953182e-02 5.34421265e-01 -7.84826279e-01 7.05256701e-01
-6.29610837e-01 4.89785790e-01 -5.42646706e-01 -2.04889700e-01
-1.93156108e-01 3.40787470e-01 -3.81116122e-01 5.15909433e-01
3.85470331e-01 9.62113857e-01 -1.13957059e+00 -2.24804401e-01
1.28757983e-01 2.28877902e-01 -1.86954513e-01 2.54811853e-01
-3.26616287e-01 7.78002262e-01 -3.20609301e-01 7.72999585e-01
6.40582025e-01 -3.59657139e-01 -6.60765544e-02 -1.50260357e-02
-3.18965703e-01 9.97817591e-02 -8.32961977e-01 1.70974016e+00
-4.11144316e-01 7.60139585e-01 1.15343168e-01 -8.59843552e-01
1.03492403e+00 2.01893434e-01 4.80576396e-01 -7.68256485e-01
-3.33790272e-03 2.91261256e-01 -4.29498494e-01 -6.54777169e-01
6.39483273e-01 -3.39726806e-01 -5.07104583e-02 3.73855591e-01
3.34241897e-01 -1.08068913e-01 2.70663977e-01 1.35055497e-01
6.04949892e-01 3.14611852e-01 4.80677187e-01 -4.02110904e-01
1.87657848e-01 3.39919060e-01 4.72201288e-01 7.72315800e-01
-8.51406157e-02 6.29619300e-01 -1.74276426e-01 -6.75582945e-01
-1.07340229e+00 -5.37187636e-01 -6.80535495e-01 1.35699701e+00
-4.59537983e-01 -3.10018510e-01 -2.07654715e-01 -5.24668813e-01
1.50382787e-01 8.03955376e-01 -9.48333919e-01 3.24095577e-01
1.85860232e-01 -1.01677418e+00 1.07367909e+00 4.55748856e-01
6.15205705e-01 -1.00697923e+00 -4.25250173e-01 1.09487474e-01
-2.28985786e-01 -1.01083314e+00 2.77027518e-01 3.79736930e-01
-1.11452651e+00 -8.62066925e-01 -6.37765825e-01 -3.92242044e-01
6.14480853e-01 2.50315398e-01 1.23163986e+00 -1.52160019e-01
-7.33463764e-02 3.92846495e-01 -8.26618195e-01 -4.13406968e-01
-3.00289392e-01 1.34857386e-01 1.44568998e-02 2.80089229e-02
3.81737590e-01 -5.00967741e-01 -4.63201135e-01 1.86931729e-01
-1.34646070e+00 7.72216022e-02 6.09380722e-01 8.02357793e-01
5.27635992e-01 -1.72500193e-01 3.55343401e-01 -6.29615724e-01
-2.00004980e-01 -1.13452864e+00 -4.14865792e-01 2.34454088e-02
-6.06423974e-01 -1.54743478e-01 3.30162704e-01 2.24695668e-01
-9.10749972e-01 2.02645198e-01 -2.94800967e-01 -2.92222679e-01
-6.94485843e-01 1.44653237e+00 2.69265741e-01 1.01185732e-01
1.31085777e+00 2.57444739e-01 -1.80123225e-01 -6.86007798e-01
5.19759357e-01 1.17677605e+00 6.20424986e-01 -8.89789313e-02
8.42684329e-01 5.96646488e-01 -3.88988167e-01 -1.18256652e+00
-1.48368144e+00 -6.90177977e-01 -5.89608788e-01 -7.43361786e-02
7.31888831e-01 -1.78015268e+00 1.68714657e-01 3.52880329e-01
-4.99342144e-01 -8.98008823e-01 -3.66014630e-01 7.14378059e-01
-6.59531727e-02 1.12036459e-01 -2.99513221e-01 -9.02179837e-01
-7.20341444e-01 -3.85566831e-01 1.13238585e+00 1.71441913e-01
1.46142215e-01 -1.12674868e+00 4.04652119e-01 2.36000404e-01
6.92335486e-01 5.52389324e-01 1.51984960e-01 -8.60681355e-01
-5.34313358e-02 -1.89442962e-01 -4.50794160e-01 5.96119225e-01
8.78948346e-02 -1.01722844e-01 -9.97840226e-01 -5.79476058e-01
-5.18119097e-01 -9.42465782e-01 1.40637445e+00 2.49326214e-01
7.32148826e-01 -4.47092742e-01 1.16501845e-01 1.04979491e+00
1.67762983e+00 -3.18446517e-01 7.15207756e-01 7.03413546e-01
5.54935038e-01 5.17272532e-01 5.94228864e-01 7.39718199e-01
5.88169575e-01 -5.14789624e-03 6.90068007e-01 -5.96832395e-01
2.61396348e-01 -1.34342209e-01 3.42213422e-01 5.21980405e-01
-1.84953794e-01 -1.99436888e-01 -1.48383915e+00 9.66640711e-01
-1.56221116e+00 -9.97453332e-01 -4.38850492e-01 1.92605174e+00
8.35037708e-01 -4.87726569e-01 -2.14672714e-01 -1.83765084e-01
3.01173598e-01 5.75302064e-01 -3.19340169e-01 1.98168918e-01
-5.54297984e-01 1.12495258e-01 1.05717099e+00 5.04062951e-01
-1.88106394e+00 1.22101164e+00 5.42197609e+00 6.18425190e-01
-1.39808083e+00 3.16777855e-01 5.83178818e-01 1.94112867e-01
-4.83306110e-01 1.36682346e-01 -8.63632739e-01 5.59861138e-02
1.40248096e+00 2.23965973e-01 1.10681482e-01 8.86921346e-01
2.20482841e-01 -5.44431694e-02 -6.43432081e-01 7.36554980e-01
-7.81771317e-02 -1.41536605e+00 -3.91347036e-02 -7.00137839e-02
1.09657741e+00 1.33836961e+00 -1.06521234e-01 5.54414928e-01
5.75269878e-01 -1.17294848e+00 4.73017424e-01 4.40056115e-01
1.27231514e+00 -3.48012596e-01 1.00238764e+00 3.03349465e-01
-1.13498485e+00 -1.06128365e-01 -4.03692752e-01 1.88946985e-02
-1.00701958e-01 6.79026484e-01 -4.73051757e-01 9.63152170e-01
1.11842525e+00 1.33202243e+00 -8.26582670e-01 7.96675086e-01
-3.24112862e-01 1.15393758e+00 -5.57670176e-01 4.36016589e-01
7.77938306e-01 -2.03178320e-02 4.51585263e-01 1.49196160e+00
3.23969096e-01 5.03448784e-01 2.84337729e-01 4.90943432e-01
-7.89992213e-02 1.92965925e-01 -8.13814998e-01 -5.57988882e-01
2.01242223e-01 1.37052882e+00 -2.59173900e-01 -5.00665903e-01
-3.64967108e-01 5.92512310e-01 1.14356883e-01 4.81054246e-01
-4.04225051e-01 -2.88693488e-01 4.92759675e-01 -1.47263944e-01
2.21286431e-01 -2.57963002e-01 -8.01864604e-04 -1.41438174e+00
-6.45387292e-01 -9.69913363e-01 5.15021265e-01 -1.11454618e+00
-1.39840615e+00 9.47638690e-01 -2.56078802e-02 -1.65734828e+00
-8.39766674e-03 -4.56822485e-01 -3.86054605e-01 6.72715485e-01
-2.00577521e+00 -1.61240363e+00 -6.76702023e-01 4.38475966e-01
2.19364703e-01 -3.07029456e-01 1.18072116e+00 2.96284020e-01
-3.79214853e-01 1.23106726e-02 4.65005517e-01 5.12265265e-01
5.18099308e-01 -1.05762196e+00 6.03173971e-01 8.06498230e-01
1.45069987e-01 2.83381224e-01 5.11304379e-01 -5.46734333e-01
-1.12776315e+00 -1.77927327e+00 9.12301898e-01 -1.53071269e-01
7.71323085e-01 3.58213373e-02 -8.07626307e-01 1.23856258e+00
1.77971825e-01 4.70240504e-01 1.08578861e+00 -1.06171280e-01
-3.14850509e-01 -2.97039658e-01 -8.94177377e-01 9.67773795e-02
7.47496545e-01 -6.37403488e-01 -5.59574842e-01 6.24034584e-01
5.57552278e-01 -1.65684283e-01 -1.19710982e+00 7.17293799e-01
1.91010550e-01 -7.42857516e-01 7.93934822e-01 -9.35470045e-01
8.43780100e-01 -3.23284358e-01 -6.72965944e-01 -1.47493196e+00
-4.32007164e-01 -2.18726560e-01 5.17168999e-01 1.04107678e+00
8.10288608e-01 -5.84123969e-01 2.24558800e-01 1.19533971e-01
-3.80859703e-01 -1.83854058e-01 -6.29927337e-01 -8.31564665e-01
2.00739220e-01 -6.10283673e-01 3.78420234e-01 1.67555022e+00
-4.09959316e-01 2.56805927e-01 -4.57892567e-01 6.06609285e-01
7.48221159e-01 2.70414203e-01 9.47516799e-01 -1.22962618e+00
-1.63784459e-01 -3.94971296e-02 -9.09257457e-02 -9.94989038e-01
-4.61592562e-02 -1.23694837e+00 1.12902813e-01 -1.51944733e+00
-9.36107710e-02 -6.66359127e-01 -4.77689981e-01 1.32641339e+00
-4.94621843e-02 5.45716882e-01 -1.40509978e-01 5.09238720e-01
-5.36523879e-01 8.94797981e-01 7.79953241e-01 -1.36395380e-01
-2.10754961e-01 -3.84726018e-01 -4.21423852e-01 4.56013530e-01
1.07904601e+00 -6.90093756e-01 1.21363394e-01 -6.73593342e-01
4.49527532e-01 1.08982660e-01 4.75403756e-01 -1.03347683e+00
-1.87238246e-01 -2.39536256e-01 3.53835672e-01 -6.73525929e-01
-9.31135863e-02 -5.79162598e-01 3.65489870e-01 2.63604373e-01
-3.85181963e-01 -3.24060261e-01 4.33691591e-01 2.94778436e-01
-5.92054129e-01 -2.91335613e-01 6.59914970e-01 -4.91596431e-01
-1.26189160e+00 5.63215792e-01 -5.26249349e-01 2.03277111e-01
5.96623480e-01 2.47310355e-01 -3.31110328e-01 -4.50065255e-01
-8.96126688e-01 7.27858126e-01 2.68962532e-01 2.23097414e-01
3.25889826e-01 -1.10967243e+00 -1.44319212e+00 3.22040856e-01
6.59217715e-01 2.29313701e-01 5.39574146e-01 7.89103210e-01
-7.53397167e-01 4.35341924e-01 -3.42636615e-01 -6.44212723e-01
-7.36258447e-01 1.00495689e-01 3.61830890e-01 -1.64718434e-01
-4.36094254e-01 9.41173673e-01 -1.40326321e-01 -1.31115568e+00
-3.02907795e-01 -1.24432862e-01 -4.19954151e-01 2.80462086e-01
4.96059865e-01 1.01414517e-01 -1.32469192e-01 -9.01612759e-01
-3.36378425e-01 1.33884639e-01 5.58323920e-01 -1.99825883e-01
2.08652091e+00 -2.37655237e-01 -3.95019241e-02 5.50450146e-01
1.22817695e+00 -3.23598236e-01 -1.03225589e+00 -5.25168836e-01
-9.17024687e-02 -1.54773772e-01 5.55009484e-01 -7.86821723e-01
-1.18950999e+00 9.27051485e-01 7.70317733e-01 2.40600272e-03
1.16227853e+00 4.34299419e-03 4.82755810e-01 8.59433532e-01
1.28269702e-01 -9.05610621e-01 -3.57009113e-01 8.07917118e-01
1.15712893e+00 -1.76876462e+00 2.02282816e-01 2.16227233e-01
-9.04565275e-01 9.17397380e-01 4.66956764e-01 7.26629943e-02
9.66655076e-01 -1.22983553e-01 4.73124385e-01 -5.37901938e-01
-4.80134457e-01 -7.70949006e-01 3.75763416e-01 4.53121305e-01
4.76640105e-01 3.01460326e-01 1.99672744e-01 3.48230839e-01
-6.63668811e-02 3.00123721e-01 3.07762295e-01 1.14009976e+00
-5.05181432e-01 -6.64170563e-01 -2.37975925e-01 4.83486950e-01
-3.94879788e-01 -7.17523992e-01 -7.15713799e-02 5.68430901e-01
-3.22339125e-02 9.48497951e-01 1.69592187e-01 -2.85882056e-01
-4.15710323e-02 -3.69296446e-02 -3.43710244e-01 -6.53050065e-01
-4.89119262e-01 4.57402617e-02 4.94054407e-01 -4.55752194e-01
-1.01572371e+00 -5.54066837e-01 -1.14160359e+00 -8.39107558e-02
-8.88803601e-02 2.71074533e-01 6.97422028e-01 7.57035017e-01
5.44638991e-01 1.12936413e-02 9.39644694e-01 -1.21343243e+00
-3.79611373e-01 -1.56091774e+00 -8.64201188e-01 4.16805059e-01
4.82350767e-01 -1.84022695e-01 -5.68672478e-01 1.44617647e-01] | [9.576154708862305, -1.4506242275238037] |
6107ed92-b2a1-42e3-aef4-fa25ed25c176 | fine-robotic-manipulation-without-force | 2301.13413 | null | https://arxiv.org/abs/2301.13413v1 | https://arxiv.org/pdf/2301.13413v1.pdf | Fine Robotic Manipulation without Force/Torque Sensor | Force Sensing and Force Control are essential to many industrial applications. Typically, a 6-axis Force/Torque (F/T) sensor is mounted between the robot's wrist and the end-effector in order to measure the forces and torques exerted by the environment onto the robot (the external wrench). Although a typical 6-axis F/T sensor can provide highly accurate measurements, it is expensive and vulnerable to drift and external impacts. Existing methods aiming at estimating the external wrench using only the robot's internal signals are limited in scope: for example, wrench estimation accuracy was mostly validated in free-space motions and simple contacts as opposed to tasks like assembly that require high-precision force control. Here we present a Neural Network based method and argue that by devoting particular attention to the training data structure, it is possible to accurately estimate the external wrench in a wide range of scenarios based solely on internal signals. As an illustration, we demonstrate a pin insertion experiment with 100-micron clearance and a hand-guiding experiment, both performed without external F/T sensors or joint torque sensors. Our result opens the possibility of equipping the existing 2.7 million industrial robots with Force Sensing and Force Control capabilities without any additional hardware. | ['Quang-Cuong Pham', 'Shilin Shan'] | 2023-01-31 | null | null | null | null | ['industrial-robots'] | ['robots'] | [ 3.08161825e-01 1.06791377e-01 -1.86093360e-01 1.35689810e-01
-1.88212365e-01 -5.78068793e-01 1.90374210e-01 -1.04923882e-01
-4.52224642e-01 5.73704779e-01 -4.51215595e-01 -1.59756511e-01
-3.82273853e-01 -4.95100200e-01 -8.56212199e-01 -4.55686331e-01
8.17867462e-03 4.55772728e-01 2.54722327e-01 -3.58790070e-01
4.04763430e-01 6.66663468e-01 -1.32255888e+00 -5.84881544e-01
5.60876727e-01 1.12736583e+00 6.28026664e-01 5.17310202e-01
6.83358133e-01 3.21435362e-01 -8.14235568e-01 1.20400272e-01
4.23808813e-01 1.83811024e-01 -4.99489516e-01 6.00968078e-02
-1.04875781e-01 -6.88930929e-01 -1.56063901e-03 7.86038935e-01
3.19949985e-01 1.87078103e-01 5.46196759e-01 -1.00563991e+00
-3.85032088e-01 4.60682899e-01 -3.99311870e-01 -5.19157171e-01
5.97449720e-01 1.19068787e-01 5.34302950e-01 -5.19235611e-01
8.37711155e-01 7.22451448e-01 8.38945448e-01 3.91375750e-01
-1.30043101e+00 -3.49853575e-01 -3.18559736e-01 -2.45826021e-01
-9.77316618e-01 -7.26789087e-02 1.08911574e+00 -6.24564767e-01
8.90604973e-01 1.05219431e-01 4.29257929e-01 1.16880226e+00
8.78977656e-01 2.34615460e-01 8.31977367e-01 -3.84029955e-01
1.30102798e-01 -1.36444658e-01 -2.01935634e-01 2.20580459e-01
3.16792607e-01 4.52076979e-02 2.19368219e-01 -3.66361905e-03
1.52774417e+00 1.58258826e-01 -3.38981062e-01 -5.57121933e-01
-1.41046035e+00 3.72709066e-01 5.64419329e-01 2.45945349e-01
-5.87656379e-01 2.12019101e-01 4.16872114e-01 4.45562005e-01
-1.75657779e-01 9.54408109e-01 -7.56231725e-01 -6.10729873e-01
-3.01435322e-01 2.93912590e-01 9.08710480e-01 7.29459286e-01
2.91511983e-01 1.75132617e-01 4.54730034e-01 7.70836890e-01
4.77103293e-02 4.58999246e-01 4.26429182e-01 -1.14355922e+00
5.82748175e-01 5.07645726e-01 4.74655807e-01 -1.05233181e+00
-7.19327569e-01 -1.05425589e-01 -5.12688994e-01 6.35828316e-01
6.45184517e-01 -5.61402202e-01 -5.25699317e-01 1.33712733e+00
3.06061089e-01 -6.46672666e-01 -1.69663966e-01 1.36256194e+00
-1.00649223e-01 -5.97311892e-02 -4.98661578e-01 -8.18577409e-02
1.08612001e+00 -4.45180595e-01 -6.79800212e-01 -3.11706722e-01
1.44715890e-01 -8.17719162e-01 1.28877735e+00 6.46052420e-01
-8.42367887e-01 -7.35374391e-01 -1.39399385e+00 1.26676010e-02
-9.04138312e-02 3.47325742e-01 3.06824386e-01 1.78097084e-01
-2.42301971e-01 1.24237692e+00 -1.16348231e+00 -7.69389048e-02
-5.83741665e-01 6.63138509e-01 -5.29834032e-01 5.15194774e-01
-1.15671146e+00 1.38991475e+00 6.81470558e-02 3.89075190e-01
-3.20360154e-01 -2.81730652e-01 -3.95947784e-01 -4.14452702e-01
6.15765393e-01 -2.69126028e-01 1.25767899e+00 -2.17741087e-01
-2.22836590e+00 1.10951766e-01 5.14904141e-01 -2.09195167e-01
7.38147974e-01 -7.57770658e-01 -3.67614031e-02 9.02879313e-02
-4.06863801e-02 7.60699958e-02 1.04179287e+00 -8.80976677e-01
2.01234445e-01 -3.70932490e-01 3.45355831e-02 -1.23208135e-01
-2.61010736e-01 -3.45390886e-01 2.05854103e-01 -8.05066228e-01
2.90106177e-01 -1.25867176e+00 -5.86988516e-02 -1.43175258e-03
-5.37186205e-01 7.17054009e-02 9.14447963e-01 -5.11780560e-01
6.54964626e-01 -1.89012682e+00 2.34234467e-01 2.30541736e-01
-1.13883473e-01 2.11690784e-01 2.76569217e-01 8.66735876e-01
1.60030842e-01 -3.87075275e-01 1.46910787e-01 4.30139184e-01
-1.59041584e-01 1.23392545e-01 -1.29781947e-01 4.24183995e-01
2.59482205e-01 6.10412121e-01 -6.21586025e-01 1.05414279e-01
3.57955426e-01 3.92026991e-01 -1.89008951e-01 1.91372365e-01
-2.22803820e-02 6.45779967e-01 -6.31676435e-01 6.36536360e-01
1.55525148e-01 9.78383049e-02 1.71506986e-01 -4.23708588e-01
-2.30032876e-01 3.77647549e-01 -1.21500659e+00 1.37623298e+00
-6.36184216e-01 3.09134841e-01 5.04309237e-01 -8.71326149e-01
1.67425978e+00 6.03335917e-01 4.89799798e-01 -4.13237870e-01
6.52492702e-01 6.08696222e-01 1.94925547e-01 -5.87139606e-01
3.93976092e-01 -1.20850556e-01 -7.53966644e-02 4.96339262e-01
-8.96464735e-02 -6.25065207e-01 -6.17027283e-02 -4.44170922e-01
1.06141543e+00 7.31628299e-01 1.74392909e-01 -2.05634423e-02
2.41939232e-01 -1.25339508e-01 3.80185813e-01 2.26895183e-01
2.57964171e-02 4.46500987e-01 2.96616763e-01 -2.78370142e-01
-1.46168447e+00 -1.04322660e+00 -1.19171496e-02 8.24594855e-01
1.76074699e-01 -4.17697430e-02 -6.17046237e-01 -1.09103583e-02
6.28854871e-01 1.91387162e-01 -2.09357932e-01 -3.19357574e-01
-7.97876120e-01 -6.05725981e-02 2.55290687e-01 9.69740868e-01
8.59037507e-03 -8.17790985e-01 -1.13827550e+00 6.15245163e-01
2.51215100e-01 -1.19967163e+00 -4.78256419e-02 3.61135304e-01
-1.14596999e+00 -1.23247516e+00 -5.33447623e-01 -4.09050763e-01
5.70991039e-01 -1.31361037e-01 5.11952162e-01 -3.61876428e-01
-3.42394233e-01 1.60095751e-01 -3.31421852e-01 -5.12247682e-01
-2.84752756e-01 7.34986737e-02 4.51052427e-01 -5.68772376e-01
-1.59738764e-01 -8.36269319e-01 -5.37164390e-01 8.35394323e-01
-2.86913663e-01 -2.27881849e-01 8.01600039e-01 5.85413754e-01
2.01098830e-01 -1.87572196e-01 7.05928504e-01 -2.58795053e-01
8.67136240e-01 7.70560130e-02 -7.25248098e-01 -3.80409896e-01
-3.29159170e-01 -1.48301378e-01 1.03322434e+00 -1.01914644e+00
-9.27570522e-01 2.04405472e-01 -2.85182476e-01 -4.54560518e-01
-5.22735752e-02 4.53310043e-01 2.51560628e-01 -5.49541637e-02
6.79447353e-01 -8.65163952e-02 6.07334435e-01 -7.53602862e-01
4.15112861e-02 7.98898160e-01 8.86410892e-01 -9.31460202e-01
6.46906793e-01 5.45461103e-03 4.61314648e-01 -7.42433786e-01
3.37350368e-02 -7.68808424e-02 -9.41991389e-01 -4.51134294e-01
3.47690016e-01 -4.43836778e-01 -1.39628124e+00 3.74419093e-01
-1.04219544e+00 -3.34571242e-01 -2.85455346e-01 9.53438520e-01
-8.28242898e-01 3.03611934e-01 -1.19800651e+00 -9.12070453e-01
-5.10949135e-01 -1.43853533e+00 1.09341443e+00 -9.63657647e-02
-7.21451879e-01 -5.02941549e-01 -1.88320592e-01 1.18918680e-01
5.64738274e-01 8.24039757e-01 5.33336043e-01 -6.96838275e-02
1.99563783e-02 -8.84845376e-01 3.30157071e-01 4.87439543e-01
3.71437699e-01 1.47292055e-02 -5.03084958e-01 -5.14394045e-01
5.62099516e-01 -5.34519851e-01 4.35436666e-02 2.36408219e-01
5.86743116e-01 -1.10186026e-01 -2.74640441e-01 -1.17398657e-01
1.33970404e+00 2.80727535e-01 2.92438358e-01 4.63829070e-01
7.32786953e-01 6.80314839e-01 8.43305647e-01 2.33488888e-01
-2.86579430e-01 9.67094302e-01 7.19443917e-01 3.56931716e-01
3.58759463e-01 -1.04653232e-01 4.19348419e-01 8.06968987e-01
-7.62247562e-01 3.90789539e-01 -5.61369061e-01 1.62808061e-01
-1.39389050e+00 -3.81361276e-01 -2.94850469e-01 2.45382905e+00
9.10733938e-01 4.92507964e-01 2.20015272e-02 8.33214104e-01
7.11639762e-01 -3.91582519e-01 -9.02854800e-01 -4.75745171e-01
6.47676110e-01 5.01128323e-02 6.27292097e-01 3.74680221e-01
-6.47331536e-01 2.38750026e-01 6.27330256e+00 2.39729896e-01
-1.64493001e+00 -3.12699765e-01 -3.13746989e-01 -5.59169166e-02
4.09966022e-01 -3.19660395e-01 -5.32881200e-01 5.49263954e-01
6.49145901e-01 2.42367744e-01 3.90876830e-01 1.22669542e+00
5.68434931e-02 -1.83038503e-01 -1.22855818e+00 2.80971557e-01
-4.00680661e-01 -7.76230395e-01 -6.73307657e-01 -8.75514373e-03
2.97006547e-01 -1.98610440e-01 -1.91504449e-01 -4.36016321e-02
-2.15332404e-01 -4.50991422e-01 7.73515880e-01 4.68979329e-01
8.98735344e-01 -3.35526109e-01 6.84728205e-01 6.82817221e-01
-1.03376675e+00 -3.04571271e-01 -4.54012096e-01 -8.63785207e-01
4.46569979e-01 8.66909742e-01 -1.15277600e+00 5.40421069e-01
2.92872220e-01 1.40422866e-01 8.68610367e-02 4.34622735e-01
-3.76520157e-01 2.48040780e-01 -5.81672907e-01 -3.95737916e-01
-1.90962285e-01 -2.77872622e-01 7.32841015e-01 4.42640156e-01
3.35163087e-01 -4.78370786e-01 9.06990990e-02 8.98233771e-01
3.23636323e-01 -2.96557099e-01 -5.58050156e-01 3.32692498e-03
4.45881426e-01 1.33819354e+00 -5.60665429e-01 1.40095189e-01
6.14945553e-02 7.59908795e-01 1.01255151e-02 7.89560750e-02
-6.51365995e-01 -1.07070041e+00 5.60947180e-01 3.19877595e-01
8.86211842e-02 -8.68945956e-01 -1.81224450e-01 -8.01421881e-01
5.94832480e-01 -4.22983646e-01 -5.12698233e-01 -9.58558917e-01
-1.06877387e+00 2.06851274e-01 -9.20640901e-02 -1.46121931e+00
-9.26307142e-01 -1.18723178e+00 -4.24328476e-01 8.17080736e-01
-5.85745037e-01 -6.60979629e-01 -1.14056282e-01 2.77245671e-01
3.11546117e-01 2.43239820e-01 7.91072845e-01 7.25981966e-02
-1.65805921e-01 5.29747680e-02 -1.97054744e-02 1.71561137e-01
7.40250230e-01 -1.13277805e+00 3.34967524e-01 3.08019489e-01
-6.01203322e-01 1.05621266e+00 1.18270767e+00 -7.30291367e-01
-2.08125162e+00 -4.37936544e-01 4.39878494e-01 -2.41545513e-01
9.84436631e-01 -4.67437863e-01 -8.10752630e-01 7.93084919e-01
-3.39874208e-01 1.23421952e-01 -8.59848931e-02 -1.38355568e-01
1.38747901e-01 -1.62618503e-01 -1.12620497e+00 5.62690794e-01
6.67534471e-01 -5.15861273e-01 -9.38841045e-01 1.70501903e-01
5.25032341e-01 -8.12124491e-01 -1.49754643e+00 7.42024779e-01
1.21843946e+00 -5.37449062e-01 9.77040648e-01 1.33451261e-03
3.77377391e-01 -3.36946636e-01 2.15380281e-01 -1.18466675e+00
-3.41168433e-01 -5.63230336e-01 -2.93886691e-01 9.65069056e-01
1.38121927e-02 -8.87821496e-01 7.51667142e-01 3.69776309e-01
-2.56942473e-02 -1.06693459e+00 -9.34134364e-01 -1.01980186e+00
-1.25461161e-01 -3.29607397e-01 4.58302826e-01 6.38046861e-01
5.66693783e-01 3.89080435e-01 -4.59674507e-01 9.04866830e-02
1.06740378e-01 -3.18349600e-02 9.93793070e-01 -1.42792952e+00
-4.09073085e-01 -1.36755079e-01 -4.69722360e-01 -8.20790648e-01
-2.69949317e-01 -1.40018150e-01 2.44332984e-01 -1.30458295e+00
-5.22003710e-01 -1.82321802e-01 5.45528829e-02 3.31106782e-01
1.74730048e-01 2.32864380e-01 1.07363969e-01 2.56363869e-01
3.18698555e-01 1.38203412e-01 1.42309296e+00 1.59275308e-01
-3.67864847e-01 3.22555810e-01 1.11830972e-01 7.27392077e-01
8.44616592e-01 -3.98717821e-02 -1.81182399e-01 -3.16808134e-01
2.12765634e-01 3.43906611e-01 3.49915564e-01 -1.21964562e+00
2.72088736e-01 -2.57928353e-02 6.71407342e-01 -1.48663998e-01
3.11647207e-01 -1.06150317e+00 5.16100883e-01 8.32294941e-01
-9.25175697e-02 2.78087318e-01 -1.60440251e-01 1.86744899e-01
-2.11428389e-01 -2.28283092e-01 3.80926311e-01 -3.51148918e-02
-6.23888671e-02 -3.99815589e-01 -3.78253669e-01 -7.27916598e-01
9.37416434e-01 -4.97162044e-01 -2.74669737e-01 -1.34156004e-01
-7.30435789e-01 -1.95707545e-01 6.13851786e-01 5.48399329e-01
3.92133236e-01 -1.09271240e+00 -7.52883926e-02 5.02565026e-01
-2.33804569e-01 -8.36198255e-02 -7.56829605e-02 9.81617689e-01
-5.09904146e-01 5.01818597e-01 -6.06397569e-01 -7.20169723e-01
-8.43955755e-01 5.46761632e-01 1.48851424e-01 -4.35449108e-02
-6.90683126e-01 2.81593919e-01 -5.56139171e-01 -6.17036104e-01
-1.63242176e-01 -5.80019891e-01 1.09909557e-01 -3.30098748e-01
2.32009757e-02 6.55780673e-01 3.35189819e-01 -2.57127225e-01
-4.66639623e-02 1.06386757e+00 4.07960653e-01 -1.22756049e-01
1.27841318e+00 8.15672353e-02 8.62795860e-02 6.76009893e-01
9.13019836e-01 6.61746562e-02 -1.54709482e+00 2.85265028e-01
-2.92494968e-02 -4.73276854e-01 -4.54411983e-01 -6.62779868e-01
-7.46618211e-01 7.19510555e-01 2.09015340e-01 4.02016670e-01
8.33916724e-01 -2.28423223e-01 7.54009306e-01 6.50353074e-01
7.95370698e-01 -1.41811585e+00 2.09820285e-01 3.30493271e-01
1.25165522e+00 -7.25726962e-01 1.45626560e-01 -6.08537853e-01
-2.39950821e-01 1.42273891e+00 4.98556226e-01 -6.35309815e-01
3.07172328e-01 8.13252985e-01 1.64437279e-01 3.06778371e-01
-3.41030210e-01 4.91066635e-01 2.89179504e-01 4.88160133e-01
5.86455822e-01 1.48146495e-01 -5.37984967e-01 3.94464254e-01
-3.00992340e-01 4.13504124e-01 4.02409613e-01 1.56828535e+00
-5.73485613e-01 -1.12247515e+00 -7.54409850e-01 4.53870624e-01
-6.29193842e-01 7.49283612e-01 -1.48387760e-01 1.13198233e+00
-9.97243300e-02 7.31956661e-01 8.09927210e-02 -7.28294671e-01
8.37822616e-01 1.03545943e-02 6.72689676e-01 -2.27146775e-01
-4.42652732e-01 1.86994225e-01 4.03650366e-02 -8.32047939e-01
-1.39652997e-01 -5.05180657e-01 -1.39111960e+00 3.43906954e-02
-5.90671897e-01 -2.83430338e-01 1.12862229e+00 7.35530555e-01
2.57989734e-01 6.11594498e-01 3.96604508e-01 -1.60486591e+00
-1.15924799e+00 -1.45542037e+00 -7.60718882e-01 3.11643243e-01
3.42358917e-01 -1.24237561e+00 -2.70072222e-01 -1.54725075e-01] | [4.880010604858398, 1.24251127243042] |
4e658c96-cf09-402f-99c1-471031b66f7a | towards-high-order-complementary | 2212.04966 | null | https://arxiv.org/abs/2212.04966v1 | https://arxiv.org/pdf/2212.04966v1.pdf | Towards High-Order Complementary Recommendation via Logical Reasoning Network | Complementary recommendation gains increasing attention in e-commerce since it expedites the process of finding frequently-bought-with products for users in their shopping journey. Therefore, learning the product representation that can reflect this complementary relationship plays a central role in modern recommender systems. In this work, we propose a logical reasoning network, LOGIREC, to effectively learn embeddings of products as well as various transformations (projection, intersection, negation) between them. LOGIREC is capable of capturing the asymmetric complementary relationship between products and seamlessly extending to high-order recommendations where more comprehensive and meaningful complementary relationship is learned for a query set of products. Finally, we further propose a hybrid network that is jointly optimized for learning a more generic product representation. We demonstrate the effectiveness of our LOGIREC on multiple public real-world datasets in terms of various ranking-based metrics under both low-order and high-order recommendation scenarios. | ['Dawei Zhou', 'Yao Zhou', 'Longfeng Wu'] | 2022-12-09 | null | null | null | null | ['logical-reasoning'] | ['reasoning'] | [-1.05883954e-02 -2.84064919e-01 -5.83868921e-01 -9.02229965e-01
-1.75525129e-01 -9.33298349e-01 6.10400558e-01 4.20686930e-01
-6.22117408e-02 -4.73072156e-02 4.68202353e-01 -4.11952406e-01
-9.05273914e-01 -1.15111279e+00 -5.97009838e-01 -2.48995319e-01
-1.99986950e-01 4.27161783e-01 -1.99886575e-01 -6.49785638e-01
1.11783773e-01 3.85325313e-01 -1.65292954e+00 4.23150361e-01
8.09295297e-01 1.29743350e+00 3.32777239e-02 1.27229571e-01
-5.42704351e-02 5.50234318e-01 2.55243480e-01 -1.21412396e+00
6.85760856e-01 2.93664578e-02 -5.77066839e-01 -2.26586480e-02
4.00263250e-01 -2.31391072e-01 -1.45524219e-01 9.42739904e-01
6.81500882e-02 5.36331236e-01 8.31804216e-01 -1.36366892e+00
-1.23879468e+00 9.23061013e-01 -1.96867034e-01 -2.08492562e-01
5.11545658e-01 -2.70378768e-01 2.07369733e+00 -1.06358874e+00
4.17673945e-01 1.05726826e+00 5.82456589e-01 -9.10876617e-02
-1.14778721e+00 -3.99812609e-01 3.53599787e-01 2.63048470e-01
-1.23678243e+00 2.55119447e-02 7.08168864e-01 -1.82261631e-01
9.12609339e-01 4.25307572e-01 6.28243148e-01 6.96195185e-01
-1.51806166e-02 8.30771565e-01 6.17930174e-01 -1.65335268e-01
-5.48899062e-02 5.85944593e-01 5.61076880e-01 3.80608469e-01
5.79325080e-01 1.16978414e-01 -4.64436650e-01 9.38935671e-03
4.30972964e-01 7.48944223e-01 1.05182296e-02 -4.24105972e-01
-7.49072194e-01 8.94661546e-01 5.17293155e-01 -3.53764854e-02
-7.15111136e-01 -2.45920271e-02 2.18621373e-01 4.78219897e-01
1.81906819e-01 7.56760001e-01 -7.07186997e-01 2.66044855e-01
-2.76778042e-01 5.13429463e-01 8.71345997e-01 1.23511171e+00
6.81955576e-01 -5.39258182e-01 -1.06358133e-01 8.96597683e-01
7.60855734e-01 2.56885648e-01 1.84431821e-01 -7.27017581e-01
3.76831830e-01 8.91982138e-01 2.77810067e-01 -1.22260070e+00
-3.42588514e-01 -9.03390646e-01 -7.61043727e-01 -3.70127499e-01
7.28499666e-02 3.21068168e-01 -1.21571071e-01 1.45407557e+00
2.09895492e-01 9.01665464e-02 1.37455717e-01 8.78186166e-01
8.36501479e-01 5.45431912e-01 -1.60260513e-01 2.30261013e-01
1.48859441e+00 -1.11489737e+00 -5.02052903e-01 -3.36514153e-02
6.62774086e-01 -7.36896336e-01 1.23019421e+00 5.79953372e-01
-6.82739735e-01 -6.08442783e-01 -1.19670641e+00 -4.54958260e-01
-7.77591467e-01 4.70199808e-02 1.34039521e+00 6.36663795e-01
-5.69198072e-01 7.55939305e-01 -1.24898292e-01 -2.45456412e-01
2.31330663e-01 4.21012223e-01 -2.71196127e-01 -6.09918654e-01
-1.48226857e+00 7.48268783e-01 2.55488366e-01 2.51923293e-01
-2.02223942e-01 -1.00075948e+00 -9.49992955e-01 3.94448668e-01
4.87378657e-01 -6.76877558e-01 9.67833340e-01 -3.92940521e-01
-1.29836965e+00 3.92772794e-01 2.73106813e-01 -3.62190813e-01
-1.78976003e-02 -3.46335173e-01 -1.20045483e+00 -3.28129321e-01
-1.17106654e-01 3.27154726e-01 4.45542872e-01 -1.00029647e+00
-1.02960038e+00 -3.63747984e-01 7.55222142e-01 2.74392307e-01
-6.07005060e-01 -1.02052771e-01 -6.34268582e-01 -5.61020434e-01
6.04563020e-02 -8.55536580e-01 -1.58771262e-01 1.32299304e-01
-1.48835942e-01 -5.57395041e-01 2.07726315e-01 -2.45549366e-01
1.42075634e+00 -2.08719373e+00 -8.13759789e-02 5.96641600e-01
1.10171840e-01 1.06727846e-01 -4.26114053e-01 7.44736731e-01
2.76904721e-02 -6.34665713e-02 3.99521917e-01 -2.44183913e-01
6.95054650e-01 3.33293825e-01 -4.71812129e-01 4.81816269e-02
1.51985720e-01 1.23506260e+00 -9.40488398e-01 -8.02993327e-02
4.99562919e-02 4.91463274e-01 -9.51302886e-01 1.15504496e-01
-1.59839645e-01 -3.49224925e-01 -4.79518920e-01 7.09591746e-01
7.50264943e-01 -4.17105705e-01 3.76845568e-01 -6.66671157e-01
2.45005235e-01 5.26484072e-01 -1.36446285e+00 1.61048043e+00
-8.01981032e-01 -3.60862985e-02 -3.40954214e-01 -6.38661206e-01
1.08338642e+00 -1.92942992e-01 4.11866933e-01 -1.01870394e+00
-3.89861390e-02 1.50355905e-01 -9.66050923e-02 -3.91925573e-01
1.14266324e+00 3.72599140e-02 -1.17835470e-01 6.53228581e-01
-6.71155155e-02 3.52403134e-01 3.34966093e-01 3.05357099e-01
7.35832691e-01 1.84773728e-01 2.79497921e-01 -2.42933854e-01
6.52819932e-01 -3.32070529e-01 4.97872978e-01 5.42674959e-01
2.57676154e-01 2.46578425e-01 3.91668081e-01 -5.63448668e-01
-5.69522440e-01 -1.01633763e+00 -1.46667629e-01 1.46255124e+00
5.08382380e-01 -8.01065087e-01 1.38075307e-01 -9.38988864e-01
6.25187755e-01 1.02993977e+00 -5.13851941e-01 -4.29594606e-01
-1.80642590e-01 -4.67055351e-01 3.74770164e-02 5.90237796e-01
2.04915881e-01 -7.63388097e-01 7.56696016e-02 1.05755590e-01
-7.46476576e-02 -9.25361991e-01 -8.19689393e-01 2.23280281e-01
-7.85852194e-01 -1.12767231e+00 -7.23015517e-02 -8.24043870e-01
5.93766749e-01 6.72999561e-01 1.34203041e+00 -7.87099078e-02
2.76958168e-01 2.97336966e-01 -6.50673270e-01 -6.36952370e-02
7.52541646e-02 1.65737458e-02 9.23489109e-02 4.12238151e-01
7.51441479e-01 -6.23072922e-01 -7.19216228e-01 5.62495053e-01
-1.06708121e+00 -3.34508866e-01 6.08863592e-01 7.27757335e-01
5.87442756e-01 6.48515761e-01 6.44327581e-01 -1.28411841e+00
1.00190055e+00 -7.13970542e-01 -4.21705008e-01 5.64970195e-01
-1.10278511e+00 1.85135752e-01 9.01584804e-01 -3.99059474e-01
-1.06419301e+00 -3.12595606e-01 -2.89591461e-01 -1.30482152e-01
3.05267543e-01 8.81252646e-01 -3.26894373e-01 3.14943761e-01
4.38014001e-01 -1.35773988e-02 -4.04942930e-01 -8.53047729e-01
9.36214268e-01 7.22957969e-01 5.09619772e-01 -5.10476887e-01
6.44512713e-01 2.72226036e-01 8.90473276e-02 -1.28740326e-01
-1.00937223e+00 -8.53015125e-01 -4.35411125e-01 3.03978831e-01
2.78126113e-02 -7.49404907e-01 -1.03478765e+00 -1.29967541e-01
-6.68441355e-01 2.62559503e-01 -4.53113794e-01 6.39058292e-01
-1.23396687e-01 8.77379626e-02 -4.99337494e-01 -4.04825985e-01
-4.37171429e-01 -8.97928476e-01 6.54442906e-01 -8.17910477e-04
-8.12009275e-02 -1.14861405e+00 4.05812375e-02 2.81573176e-01
2.99667060e-01 -3.32798183e-01 1.27015495e+00 -9.80399728e-01
-4.89452034e-01 -4.79111969e-01 -3.03009599e-01 4.69585866e-01
3.36167902e-01 -2.70067185e-01 -4.06370997e-01 -2.36838460e-01
-4.68364388e-01 -2.50498597e-02 8.80714059e-01 -8.57797414e-02
6.83824778e-01 -5.17549574e-01 -2.80103292e-02 6.14602864e-01
1.54562986e+00 9.36063677e-02 6.39538229e-01 1.91217750e-01
5.21634281e-01 8.79124582e-01 1.07292473e+00 4.78199571e-01
7.13010430e-01 7.81441212e-01 4.72578228e-01 2.29845047e-01
1.38778239e-01 -7.93440282e-01 2.83315629e-01 6.23283088e-01
1.95785649e-02 -3.80232036e-01 -7.25305155e-02 3.52978617e-01
-1.91543961e+00 -8.24166596e-01 1.62630212e-02 2.23245621e+00
4.82098460e-01 1.49422571e-01 9.69812945e-02 -7.89819136e-02
3.48684937e-01 -6.13980256e-02 -6.78229809e-01 -4.91534859e-01
-4.59137261e-02 2.12554038e-01 4.39075589e-01 1.91301182e-01
-1.04407287e+00 6.45886600e-01 5.82306576e+00 6.51785672e-01
-7.86992013e-01 -9.09224078e-02 2.50936210e-01 1.16475984e-01
-9.83112335e-01 -1.15550473e-01 -8.53624344e-01 1.76687121e-01
6.15392864e-01 -1.83316141e-01 5.75198948e-01 9.46897447e-01
-1.08665280e-01 3.62247050e-01 -1.39549482e+00 1.02814114e+00
5.38867414e-02 -1.35207188e+00 2.38880992e-01 3.09168071e-01
7.60600030e-01 -2.90821254e-01 5.86407244e-01 5.63168705e-01
6.40706658e-01 -7.63414562e-01 6.29971981e-01 5.37000716e-01
5.64786017e-01 -9.09319341e-01 8.73785079e-01 9.24319178e-02
-1.38255572e+00 -2.80210584e-01 -4.47464228e-01 1.59077328e-02
5.50571010e-02 5.42927623e-01 -3.51797819e-01 8.82148743e-01
5.48697054e-01 1.36969960e+00 -6.28868520e-01 8.34649861e-01
-2.04079747e-01 1.96232468e-01 -1.99672341e-01 -8.71201530e-02
1.47276327e-01 -7.01372564e-01 -9.24705938e-02 1.05147660e+00
5.51407874e-01 8.51307660e-02 -1.20034382e-01 8.96396875e-01
-3.76735032e-01 3.33483458e-01 -3.25539947e-01 -3.20034087e-01
4.84275281e-01 1.64098692e+00 -3.45680654e-01 -3.37653421e-02
-7.07273245e-01 8.36079240e-01 2.23343477e-01 2.48561397e-01
-7.06134439e-01 -4.29195970e-01 1.26281857e+00 -1.10581284e-02
6.29243970e-01 7.92942569e-02 4.77639735e-02 -1.14415824e+00
2.33083814e-01 -6.92801416e-01 4.65145677e-01 -4.30862188e-01
-1.81410968e+00 3.92351717e-01 -1.92777246e-01 -1.47440207e+00
-1.16879985e-01 -7.68515706e-01 -4.28110838e-01 6.99663341e-01
-1.80877602e+00 -1.35673702e+00 5.28966561e-02 5.72900951e-01
9.39592347e-02 -1.37172699e-01 8.21680129e-01 6.73268914e-01
-5.24710476e-01 9.14686263e-01 3.32320720e-01 -2.72007018e-01
6.24780834e-01 -1.10873890e+00 3.06699097e-01 4.43701446e-01
4.87907022e-01 1.43916833e+00 4.74559009e-01 -2.89059877e-01
-1.86946833e+00 -1.23351932e+00 1.22261691e+00 -2.80514538e-01
7.46691167e-01 -2.24884227e-01 -6.12765312e-01 8.75548840e-01
-1.22204490e-01 6.73880354e-02 1.22521269e+00 6.80554032e-01
-1.07831550e+00 -7.27461457e-01 -1.21227801e+00 4.94113356e-01
1.17401659e+00 -5.00857830e-01 -5.35011768e-01 2.41698831e-01
8.02061141e-01 1.15796730e-01 -1.24143338e+00 1.82466134e-01
1.07625389e+00 -8.87464285e-01 1.18037450e+00 -8.19861293e-01
4.26329404e-01 -3.08079213e-01 -5.46442330e-01 -1.35694170e+00
-6.83261395e-01 -2.86514461e-01 -3.49981427e-01 1.26184118e+00
6.15527332e-01 -7.62546301e-01 6.12599730e-01 6.15804195e-01
-7.47082904e-02 -1.05924106e+00 -3.33469436e-02 -7.12404490e-01
-3.33994091e-01 -6.24144197e-01 1.35267556e+00 7.27836847e-01
1.24439746e-01 4.62272644e-01 -4.67848867e-01 2.71123409e-01
4.24899280e-01 6.36411607e-01 5.50009906e-01 -1.36981869e+00
-6.83105469e-01 -3.95885020e-01 -4.17904526e-01 -1.70754838e+00
-1.93002209e-01 -1.40102184e+00 -4.82050925e-01 -1.61869740e+00
1.19163118e-01 -7.37421632e-01 -1.03647113e+00 2.53681690e-01
2.33429447e-01 1.81463197e-01 1.21668786e-01 1.40483648e-01
-9.19407725e-01 3.75619024e-01 1.19053721e+00 -2.64542282e-01
-1.87693805e-01 3.55864763e-01 -1.56084704e+00 3.07868510e-01
4.62533772e-01 -1.74363226e-01 -7.44357049e-01 -2.65074968e-01
1.18047237e+00 -3.76701534e-01 9.31804813e-03 -3.35445970e-01
1.68503091e-01 -1.65652707e-01 5.16384542e-02 -6.79734766e-01
1.83566689e-01 -1.27669561e+00 3.62538546e-01 -1.57006249e-01
-7.96990573e-01 2.07059644e-02 -4.20688540e-01 7.01274514e-01
-1.92283675e-01 -4.14622605e-01 -1.23776495e-03 1.67114168e-01
-9.46117997e-01 4.04426008e-01 3.17749470e-01 -3.35412592e-01
6.54945970e-01 9.97345150e-02 -2.34101206e-01 -3.04914027e-01
-4.92836654e-01 3.74069542e-01 1.56363383e-01 9.23258185e-01
7.00254202e-01 -1.61928821e+00 -3.75214338e-01 3.27717543e-01
7.76603341e-01 -4.67946261e-01 1.53223976e-01 6.87249362e-01
-2.76043981e-01 7.57221699e-01 -5.70466779e-02 1.06176622e-01
-9.16872263e-01 8.82920086e-01 -7.94924051e-03 -6.41615570e-01
-4.34849352e-01 7.47589290e-01 5.52305095e-02 -8.38074327e-01
2.51550823e-01 -5.27744949e-01 -5.13080120e-01 1.34122580e-01
6.58684969e-01 4.87161398e-01 2.47807294e-01 -6.89146399e-01
-1.30432248e-01 4.03236210e-01 -4.71019238e-01 3.70254070e-01
1.41464531e+00 -3.43866110e-01 -2.19380185e-01 -1.31859416e-02
1.52822208e+00 8.46676975e-02 -9.59924638e-01 -8.22948098e-01
7.37599209e-02 -5.81714571e-01 9.34008583e-02 -8.52227569e-01
-1.30350208e+00 4.43882912e-01 2.82536000e-01 2.16111600e-01
1.19028771e+00 1.97905116e-02 1.03260088e+00 8.59615505e-01
5.38270354e-01 -8.63559902e-01 -5.21211475e-02 1.52773395e-01
6.58649206e-01 -1.12630582e+00 2.62894392e-01 -6.44937098e-01
-8.13582540e-01 1.01334167e+00 2.93209463e-01 -3.76641862e-02
1.16664255e+00 -2.66125858e-01 -2.78298676e-01 -3.77687693e-01
-8.31982672e-01 -3.47326458e-01 6.23212993e-01 5.34108758e-01
3.73805165e-01 3.62116486e-01 -3.86968285e-01 1.16137350e+00
-1.84220701e-01 3.43998447e-02 4.76229712e-02 6.38594985e-01
-1.27436419e-03 -1.39274597e+00 3.43945831e-01 7.04754233e-01
-8.31024349e-02 -2.13237599e-01 1.07166441e-02 4.82024729e-01
4.37230945e-01 1.18380499e+00 1.70669854e-01 -8.10890615e-01
7.17449188e-01 -3.73362005e-01 3.40164453e-01 -5.39973557e-01
-7.62837946e-01 -1.51126698e-01 2.34936833e-01 -6.17228925e-01
-1.87582478e-01 -4.83551770e-01 -1.13637233e+00 -4.16048616e-01
-4.24015135e-01 1.94083586e-01 6.42594635e-01 6.63736701e-01
6.66092753e-01 2.80352205e-01 8.89464498e-01 -3.42353612e-01
-7.82787502e-01 -5.78573227e-01 -1.26764679e+00 9.18529332e-01
-1.61509868e-02 -8.00931156e-01 -1.05710365e-01 -3.20164502e-01] | [10.108487129211426, 5.747125625610352] |
8d3af08d-ee04-44a1-939e-c7d82ada250a | double-and-single-descent-in-causal-inference | 2305.00700 | null | https://arxiv.org/abs/2305.00700v1 | https://arxiv.org/pdf/2305.00700v1.pdf | Double and Single Descent in Causal Inference with an Application to High-Dimensional Synthetic Control | Motivated by a recent literature on the double-descent phenomenon in machine learning, we consider highly over-parametrized models in causal inference, including synthetic control with many control units. In such models, there may be so many free parameters that the model fits the training data perfectly. As a motivating example, we first investigate high-dimensional linear regression for imputing wage data, where we find that models with many more covariates than sample size can outperform simple ones. As our main contribution, we document the performance of high-dimensional synthetic control estimators with many control units. We find that adding control units can help improve imputation performance even beyond the point where the pre-treatment fit is perfect. We then provide a unified theoretical perspective on the performance of these high-dimensional models. Specifically, we show that more complex models can be interpreted as model-averaging estimators over simpler ones, which we link to an improvement in average performance. This perspective yields concrete insights into the use of synthetic control when control units are many relative to the number of pre-treatment periods. | ['Amar Venugopal', 'Guido Imbens', 'Jann Spiess'] | 2023-05-01 | null | null | null | null | ['causal-inference', 'causal-inference'] | ['knowledge-base', 'miscellaneous'] | [ 2.47415692e-01 4.28787947e-01 -8.22806418e-01 -4.40705717e-01
-1.04244065e+00 -2.38734707e-01 7.64272809e-01 8.78003240e-03
-4.53051746e-01 1.30615389e+00 7.38170743e-01 -2.71708876e-01
-5.17076552e-01 -8.62274289e-01 -1.02958751e+00 -7.61100173e-01
-4.79193516e-02 5.26027203e-01 -5.67962945e-01 1.89359169e-02
7.58788586e-02 5.80948368e-02 -1.08500278e+00 -6.40849620e-02
1.28949785e+00 3.16464156e-03 -1.91886827e-01 3.39259863e-01
3.70210558e-01 4.04633433e-01 -3.51557016e-01 -5.76720357e-01
3.14792037e-01 -3.24907273e-01 -4.36215013e-01 -6.47449568e-02
4.45893317e-01 -2.29980409e-01 -2.29342237e-01 8.06678951e-01
5.20535767e-01 1.14410520e-01 1.07168972e+00 -1.17906654e+00
-8.26264143e-01 1.07320940e+00 -7.48607159e-01 -1.40869305e-01
1.07190169e-01 -8.07379037e-02 9.46154177e-01 -4.97864813e-01
6.71770215e-01 1.74229193e+00 8.82875979e-01 4.03328568e-01
-2.07851267e+00 -4.70149398e-01 6.57294765e-02 -3.07781339e-01
-8.84766340e-01 -5.78583002e-01 4.66966182e-01 -7.49263823e-01
3.33349794e-01 2.70821273e-01 2.52496570e-01 1.40231192e+00
2.09477291e-01 7.13679910e-01 1.15122414e+00 -7.09323227e-01
3.17739427e-01 -8.18613991e-02 6.19903684e-01 3.00473183e-01
9.27690566e-01 4.63981450e-01 -3.24350357e-01 -6.00007057e-01
9.05255914e-01 5.38471043e-02 -1.95524395e-01 -6.08485699e-01
-1.18046701e+00 1.38880861e+00 7.91055188e-02 1.69981122e-02
-4.82033670e-01 5.28434336e-01 4.27718610e-01 3.82051796e-01
7.90838718e-01 5.49720883e-01 -6.34372652e-01 6.80859163e-02
-9.84856248e-01 6.49239480e-01 7.38915741e-01 7.89322555e-01
6.01964712e-01 1.54208811e-02 -6.51478708e-01 9.53379333e-01
-2.34492123e-01 5.34452379e-01 3.61681618e-02 -1.36149824e+00
6.16250575e-01 1.71533301e-01 5.89815855e-01 -5.59638858e-01
-3.96880925e-01 -5.12332916e-01 -1.37419784e+00 -1.00938700e-01
9.93629754e-01 -8.34659278e-01 -5.08021414e-01 2.36099052e+00
1.25981405e-01 1.95071384e-01 -1.13251165e-01 5.40620029e-01
-2.68080205e-01 3.72105122e-01 3.09202880e-01 -8.61009598e-01
1.04053390e+00 -4.90562469e-01 -8.78003299e-01 -2.22160414e-01
9.35734749e-01 -3.19397271e-01 9.80080843e-01 1.66752607e-01
-1.31099296e+00 -3.62110853e-01 -4.77158099e-01 -5.96200339e-02
4.46986174e-03 1.38946861e-01 1.15541255e+00 6.96130395e-01
-7.39108264e-01 8.78607571e-01 -7.04126716e-01 -2.63382137e-01
3.62237364e-01 2.94483930e-01 -1.14039905e-01 -1.32001027e-01
-1.16104543e+00 7.12414563e-01 -2.68715564e-02 -1.47560567e-01
-4.81201977e-01 -1.46956694e+00 -7.23334312e-01 3.38637829e-01
5.37208140e-01 -1.11574721e+00 1.10989416e+00 -6.75864041e-01
-7.00697899e-01 4.91050899e-01 -4.70352978e-01 -5.09810090e-01
7.62996137e-01 -3.38989377e-01 6.79612011e-02 -6.07726932e-01
4.29187804e-01 8.43052641e-02 4.56015468e-01 -1.13367021e+00
-2.81374902e-01 -7.52270043e-01 -7.65416026e-02 -1.52985379e-01
4.25492600e-02 2.33765990e-02 2.30190024e-01 -8.30088437e-01
-1.63567275e-01 -7.78747141e-01 -4.68076944e-01 -3.83307248e-01
-5.42171657e-01 -2.62992054e-01 -2.54936427e-01 -2.83986002e-01
1.48335326e+00 -1.93929839e+00 3.21452856e-01 1.17525958e-01
2.54036248e-01 -3.36400688e-01 -5.49600124e-02 4.06550467e-01
-4.25762147e-01 3.58504176e-01 -2.71340281e-01 -4.07390803e-01
4.60795552e-01 2.07367048e-01 -2.71343470e-01 5.44586420e-01
8.00264552e-02 9.00058746e-01 -6.69169843e-01 -1.82722166e-01
-1.19319156e-01 4.01080064e-02 -8.83692741e-01 -6.98428154e-02
6.09682947e-02 2.78963864e-01 -5.32947361e-01 1.96587548e-01
8.28976750e-01 -1.52020752e-01 2.05064163e-01 2.35807747e-01
-2.39145756e-01 2.75223076e-01 -1.25104988e+00 1.17364347e+00
-3.89210701e-01 2.84550428e-01 1.75013989e-01 -1.26692009e+00
3.62074554e-01 2.85202205e-01 5.02628565e-01 -1.87992617e-01
-2.63754100e-01 8.23133811e-02 2.00698618e-02 -4.01607394e-01
1.84648544e-01 -3.78957748e-01 -2.96737134e-01 3.40834498e-01
-1.93657324e-01 2.54889488e-01 3.85433674e-01 1.38507396e-01
1.04448974e+00 -4.16126475e-02 4.71785009e-01 -5.18703163e-01
-7.78896287e-02 -8.98694843e-02 9.58972394e-01 1.36167645e+00
3.59455019e-01 4.57655549e-01 1.15521109e+00 -2.47899041e-01
-1.38050079e+00 -1.10392499e+00 -6.69870079e-01 1.16606486e+00
-5.81403136e-01 -1.08306617e-01 -8.11411798e-01 -2.75783181e-01
6.85818493e-01 9.04335141e-01 -1.03543544e+00 -5.50253177e-03
-5.48909724e-01 -1.57716942e+00 4.13386673e-01 4.90045339e-01
1.14640981e-01 -4.59726155e-01 -3.39155644e-02 1.94286153e-01
9.95802656e-02 -4.32364196e-01 -4.18743581e-01 2.64798462e-01
-1.12997568e+00 -8.72620881e-01 -8.83182824e-01 -7.70228431e-02
4.43039447e-01 -1.57077044e-01 1.28948069e+00 -1.90664843e-01
1.13123000e-01 4.75066528e-03 -3.01229358e-02 -5.25612473e-01
-4.27044779e-01 1.97692692e-01 2.55117118e-01 -1.96325436e-01
5.18611789e-01 -6.87631249e-01 -3.81935328e-01 -2.28692055e-01
-5.23778975e-01 -1.71040855e-02 4.09574538e-01 1.20564389e+00
7.69750103e-02 -5.81240058e-01 1.11936009e+00 -1.36066484e+00
8.29884589e-01 -7.44064450e-01 -6.53914332e-01 4.17509407e-01
-8.18590224e-01 4.25242633e-01 2.46643826e-01 -6.92704380e-01
-1.16409445e+00 -3.52044165e-01 2.83046365e-01 6.49804994e-02
-5.57600753e-03 5.24491072e-01 -1.66101251e-02 3.97428870e-01
7.84134507e-01 -2.86459863e-01 1.20805642e-02 -7.43036509e-01
5.03670692e-01 4.19656068e-01 6.95481524e-02 -1.04188657e+00
5.42899191e-01 1.85304791e-01 2.90001571e-01 -5.31200707e-01
-1.01234388e+00 9.18483287e-02 -6.30840421e-01 3.01979512e-01
7.44076312e-01 -9.05993700e-01 -8.93980026e-01 2.34395042e-02
-9.91117239e-01 -5.20934463e-01 -4.94490117e-01 9.48239267e-01
-8.45724225e-01 -1.03269117e-02 -5.44988394e-01 -1.16264236e+00
1.67514056e-01 -8.33800375e-01 9.80331779e-01 -6.98113963e-02
-2.56867945e-01 -1.33446884e+00 2.59195566e-01 7.02946410e-02
1.39250621e-01 4.41178083e-01 1.34435034e+00 -4.93070483e-01
-2.93772429e-01 -1.01757273e-02 -2.09426865e-01 -1.14973255e-01
-6.15970464e-03 -1.59868762e-01 -7.90541232e-01 -2.76016384e-01
-2.14818660e-02 -2.05215901e-01 1.11007285e+00 1.24166477e+00
1.22861946e+00 -3.64090532e-01 -5.15943229e-01 5.40414035e-01
1.45743418e+00 -1.15112826e-01 5.19324183e-01 2.87357662e-02
6.19249403e-01 6.75285757e-01 3.73084784e-01 7.26515234e-01
2.92162359e-01 7.29619026e-01 -9.19068679e-02 -3.80463004e-01
3.80604416e-01 -3.10848802e-01 4.79633473e-02 2.77600199e-01
-2.66750485e-01 2.73581594e-02 -5.02617717e-01 5.67585528e-01
-2.00710678e+00 -1.18524694e+00 -6.30722821e-01 2.66955805e+00
1.16072810e+00 -9.13395286e-02 3.42569113e-01 -1.11413881e-01
6.32590413e-01 -5.62330410e-02 -5.54567158e-01 -4.94811803e-01
-1.48543820e-01 3.36610198e-01 9.64371979e-01 6.81137562e-01
-9.86052334e-01 4.59545791e-01 7.70456791e+00 9.01968658e-01
-2.97649741e-01 3.57284695e-01 8.31185818e-01 -2.79414147e-01
-4.80432302e-01 1.19635865e-01 -6.94947422e-01 5.35705268e-01
1.07484651e+00 -6.39802992e-01 4.20068890e-01 6.02514267e-01
8.42689991e-01 -2.36758739e-01 -1.56300843e+00 3.62221211e-01
-5.99251390e-01 -1.46300197e+00 -2.61889428e-01 6.59757137e-01
1.13426316e+00 -3.51780325e-01 -1.19637847e-01 4.27832514e-01
8.79997015e-01 -1.08497727e+00 2.14276016e-01 7.43215680e-01
8.32999706e-01 -8.29216123e-01 5.55632472e-01 5.84235847e-01
-3.97364318e-01 -3.53320152e-01 -8.02609801e-01 -5.30928314e-01
2.33032584e-01 1.05508351e+00 -2.52419949e-01 6.80609941e-01
6.52168319e-02 5.46449542e-01 -2.12092936e-01 9.09903705e-01
2.73657858e-01 8.31048608e-01 -2.48732910e-01 3.41909260e-01
-1.20560654e-01 -6.15278363e-01 1.13505609e-01 9.69316065e-01
4.09862787e-01 3.51756305e-01 -1.71349585e-01 1.07471919e+00
-2.46753618e-01 -1.07450984e-01 -9.17779624e-01 3.02071840e-01
5.27733445e-01 5.54413795e-01 4.98418026e-02 -3.89178962e-01
-5.47706783e-01 3.20051223e-01 4.82128859e-01 5.28269827e-01
-5.92140317e-01 2.93213911e-02 9.23447132e-01 1.51280344e-01
-1.79727197e-01 -7.00091645e-02 -7.65438080e-01 -1.53249216e+00
-1.66265056e-01 -6.33741915e-01 3.52839321e-01 -5.17181396e-01
-1.59199345e+00 -4.96025622e-01 4.77025837e-01 -5.33780992e-01
-5.23225605e-01 -3.56596142e-01 -5.66036522e-01 1.01547050e+00
-8.59456658e-01 -6.53859794e-01 4.78141397e-01 1.72230884e-01
3.75150263e-01 1.62268147e-01 8.74143600e-01 3.01917732e-01
-7.63406277e-01 6.36674345e-01 8.07775438e-01 -1.27391428e-01
8.17512691e-01 -1.52141845e+00 3.20164949e-01 6.56713605e-01
-2.12521955e-01 1.00777137e+00 8.24554801e-01 -7.86169529e-01
-1.35325646e+00 -8.31171751e-01 1.12976635e+00 -6.35931611e-01
6.03138804e-01 -5.39932191e-01 -8.22267771e-01 1.07607627e+00
1.41467825e-01 -4.81456876e-01 5.77099204e-01 9.83749032e-01
3.42762301e-04 -9.46910083e-02 -1.23048043e+00 7.24209905e-01
1.06188405e+00 -3.22363786e-02 -6.83882415e-01 6.03219986e-01
6.98853016e-01 4.86890338e-02 -1.07694912e+00 4.27301913e-01
5.45445681e-01 -9.34428275e-01 1.07017303e+00 -1.12870359e+00
6.03454173e-01 3.78866702e-01 -5.60647808e-02 -1.44755864e+00
-4.98725027e-01 -6.11616373e-01 2.35606551e-01 1.25941479e+00
4.35020298e-01 -5.58756113e-01 6.40195370e-01 1.12266672e+00
2.95290738e-01 -5.74173927e-01 -9.98343647e-01 -6.21386349e-01
9.36240673e-01 -2.94912428e-01 5.30556619e-01 1.30467618e+00
-1.21049784e-01 3.42523515e-01 -8.02287459e-01 -6.49556890e-02
1.18531919e+00 -3.50358151e-02 7.98347771e-01 -1.74728262e+00
-3.60485792e-01 -8.69756341e-02 2.50803083e-01 -9.60803926e-01
4.52355683e-01 -7.62708664e-01 -5.02973437e-01 -1.20112789e+00
7.56492376e-01 -5.45232773e-01 1.41977444e-01 3.11872661e-01
-3.08283299e-01 -1.41051888e-01 7.32846260e-02 -6.96368217e-02
1.86750337e-01 3.17049384e-01 1.28782904e+00 8.35576430e-02
-3.09638083e-01 3.63933921e-01 -1.03904092e+00 5.91229320e-01
6.73335791e-01 -7.14715481e-01 -3.07192773e-01 -3.91722113e-01
2.43169203e-01 5.90050161e-01 4.50464994e-01 -2.64478058e-01
-1.32495433e-01 -7.69636989e-01 5.39958715e-01 -1.85301781e-01
4.75355126e-02 -4.34334040e-01 3.89106423e-01 2.84375638e-01
-9.46163237e-01 -1.82695910e-02 -1.96017697e-01 4.75671321e-01
3.41949105e-01 -4.39660370e-01 4.99831498e-01 -1.02963641e-01
3.66602957e-01 -4.44695204e-02 -3.01920682e-01 2.76894748e-01
6.09625518e-01 1.33143634e-01 -3.71585786e-01 -3.82005095e-01
-9.88772154e-01 3.41893047e-01 3.21034372e-01 3.86026986e-02
-5.96791543e-02 -1.47898281e+00 -1.12486422e+00 1.08238928e-01
-2.28009403e-01 -3.30997378e-01 1.16997629e-01 9.97239888e-01
2.55115837e-01 3.62635612e-01 9.56273079e-02 -2.42252544e-01
-9.20535266e-01 7.15965152e-01 1.44558758e-01 -2.35964775e-01
-3.16862166e-01 2.30470628e-01 5.07726371e-01 -3.62089545e-01
1.00728765e-01 -4.88435656e-01 2.19425395e-01 2.27944940e-01
3.70531082e-01 7.68417835e-01 -4.38363254e-01 1.66136604e-02
1.57689169e-01 4.13775265e-01 7.27891326e-02 -1.21769257e-01
1.63493979e+00 -2.75399715e-01 -3.36949497e-01 7.41443455e-01
8.21042240e-01 2.25960359e-01 -1.31908000e+00 -1.06271140e-01
2.16669276e-01 -5.11875510e-01 -2.08605617e-01 -6.29040420e-01
-6.87310696e-01 6.65905654e-01 2.43540734e-01 3.83734018e-01
8.13340604e-01 -1.02899492e-01 -2.40497217e-01 4.21558648e-01
3.04316133e-01 -9.20996070e-01 -5.58850586e-01 3.16152841e-01
8.11906993e-01 -1.22212732e+00 2.33866289e-01 -1.98202088e-01
-2.51983702e-01 5.78835309e-01 1.34827361e-01 -4.06319052e-01
4.64804560e-01 1.78487837e-01 -5.34611762e-01 3.29313837e-02
-9.68801200e-01 -8.01866725e-02 1.00240707e-01 6.33025706e-01
6.56911552e-01 3.59108299e-01 -9.54220176e-01 6.74550295e-01
-3.07147592e-01 1.28560022e-01 8.19532871e-01 2.69158393e-01
-2.51726955e-01 -1.39123750e+00 -6.15750730e-01 9.85172212e-01
-8.98475528e-01 -2.26501808e-01 -9.30356681e-02 1.18246758e+00
-1.54526100e-01 1.00543892e+00 2.48796746e-01 3.55596811e-01
4.65661615e-01 4.13679838e-01 5.57601035e-01 -6.95239484e-01
-3.62874836e-01 3.49834591e-01 2.29934469e-01 -2.76386708e-01
-5.43538809e-01 -9.57393706e-01 -4.88050818e-01 -9.24841046e-01
-2.15988770e-01 2.67157972e-01 2.67413467e-01 8.56337130e-01
1.90043375e-01 4.79015768e-01 7.01409936e-01 -8.33236456e-01
-1.13327980e+00 -1.28638053e+00 -8.83057117e-01 3.26162666e-01
6.28077626e-01 -8.38662684e-01 -5.91880739e-01 4.07068105e-03] | [7.965087890625, 5.245583534240723] |
7daf0f3e-5772-4d01-88df-95f954767be2 | tracking-public-attitudes-toward-chatgpt-on | 2306.12951 | null | https://arxiv.org/abs/2306.12951v1 | https://arxiv.org/pdf/2306.12951v1.pdf | Tracking public attitudes toward ChatGPT on Twitter using sentiment analysis and topic modeling | ChatGPT sets a new record with the fastest-growing user base, as a chatbot powered by a large language model (LLM). While it demonstrates state-of-the-art capabilities in a variety of language-generating tasks, it also raises widespread public concerns regarding its societal impact. In this paper, we utilize natural language processing approaches to investigate the public attitudes towards ChatGPT by applying sentiment analysis and topic modeling techniques to Twitter data. Our result shows that the overall sentiment is largely neutral to positive, which also holds true across different occupation groups. Among a wide range of topics mentioned in tweets, the most popular topics are Artificial Intelligence, Search Engines, Education, Writing, and Question Answering. | ['Hyeju Jang', 'Yanling Pan', 'Ratanond Koonchanok'] | 2023-06-22 | null | null | null | null | ['chatbot', 'question-answering', 'sentiment-analysis', 'chatbot'] | ['methodology', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-3.46801192e-01 4.44283277e-01 -5.52756727e-01 -1.79551467e-01
-9.70618606e-01 -2.75893450e-01 9.81209159e-01 4.85025406e-01
-3.21799219e-01 8.54981184e-01 7.24161506e-01 -5.21665692e-01
1.68958575e-01 -8.24685693e-01 5.53221069e-02 -2.87677884e-01
4.52783465e-01 4.19648230e-01 1.96851000e-01 -8.54830325e-01
6.12249553e-01 -7.36137629e-02 -1.10713136e+00 4.20181125e-01
9.53361273e-01 5.47032416e-01 -9.45022032e-02 3.07492018e-01
-9.02339339e-01 1.65910351e+00 -8.27867389e-01 -7.82599449e-01
-4.51626271e-01 -2.48959541e-01 -1.14152193e+00 -1.25358582e-01
1.60481587e-01 1.07164718e-01 -2.18797818e-01 6.94319725e-01
3.83535683e-01 8.08812305e-02 3.99408966e-01 -1.37028801e+00
-7.04284430e-01 7.38686800e-01 -4.26758766e-01 7.31977373e-02
6.00692511e-01 4.04921882e-02 1.10785735e+00 -6.30866110e-01
9.68525648e-01 1.58348286e+00 6.56337082e-01 1.73056260e-01
-8.79205525e-01 -7.35356450e-01 3.00420504e-02 -4.84554589e-01
-1.24322987e+00 -1.09530307e-01 5.58594704e-01 -7.79865324e-01
1.05289614e+00 2.11414203e-01 4.63507921e-01 1.19021344e+00
5.75615168e-01 5.16322196e-01 1.34656668e+00 -3.45153928e-01
2.40861833e-01 8.88417125e-01 3.48969340e-01 4.07102644e-01
-1.15803607e-01 -9.30354059e-01 -9.49202359e-01 -8.01613808e-01
1.12257846e-01 -8.05520173e-03 4.80243027e-01 5.58157563e-01
-9.46279049e-01 1.40341687e+00 2.81865578e-02 5.01537859e-01
-5.84537089e-01 -2.89585609e-02 4.21758801e-01 3.39531422e-01
1.36513925e+00 4.37320828e-01 -4.63051528e-01 -8.75127256e-01
-5.05174398e-01 4.56148565e-01 1.43917632e+00 6.61201358e-01
6.46231294e-01 -3.22106212e-01 -1.97498143e-01 1.14917779e+00
4.97138023e-01 4.00300294e-01 4.57782835e-01 -9.08886850e-01
8.64982009e-01 9.67138052e-01 -8.38627815e-02 -1.28027570e+00
-2.76786238e-01 -1.05165057e-01 -3.34327966e-01 -3.97700876e-01
4.66091484e-01 -6.57706320e-01 -1.00643009e-01 1.45724714e+00
1.09605081e-01 -3.75197113e-01 -1.16227850e-01 3.42351317e-01
9.45456684e-01 8.40131998e-01 7.30138421e-01 4.62569222e-02
1.60239291e+00 -5.20084620e-01 -9.15181875e-01 -6.67512119e-01
7.97168136e-01 -1.03667045e+00 1.19069886e+00 1.28680140e-01
-8.72128367e-01 -6.96142390e-02 -3.30759883e-01 -1.23992443e-01
-8.07158887e-01 -1.59005761e-01 8.32566202e-01 1.00423777e+00
-8.66476536e-01 4.47946861e-02 -2.99878806e-01 -1.05319512e+00
3.43812883e-01 -2.20265433e-01 -1.93264470e-01 2.62367465e-02
-1.27114820e+00 1.09659374e+00 -1.88189000e-01 -6.42820418e-01
7.83167109e-02 -7.13778317e-01 -6.22541547e-01 -1.09538347e-01
3.41945618e-01 -2.42527768e-01 1.47791731e+00 -4.61326122e-01
-1.50895333e+00 9.76862729e-01 -4.18088913e-01 -3.25987190e-01
2.73918033e-01 -1.47178382e-01 -4.82603192e-01 -3.26705873e-02
3.95861924e-01 3.62657607e-01 3.85137349e-01 -6.41732574e-01
-4.90814745e-01 -1.18195608e-01 -3.95006984e-02 -8.31896812e-02
-7.76087284e-01 8.36750627e-01 -4.09624875e-02 -2.55178213e-01
-1.97978094e-01 -8.44559073e-01 -3.24219644e-01 -5.20035565e-01
-3.70895654e-01 -7.34394968e-01 9.72374141e-01 -6.57706439e-01
1.50515938e+00 -1.78868151e+00 -6.98423326e-01 1.56780630e-01
2.80460477e-01 -2.75617301e-01 3.51327837e-01 1.31653070e+00
4.64134216e-01 6.58632576e-01 5.45349836e-01 -2.33363271e-01
1.68805644e-01 -1.59691889e-02 -3.96875620e-01 2.35586941e-01
3.27374846e-01 7.63642192e-01 -7.94624925e-01 -6.92844212e-01
-7.37484079e-03 3.30288023e-01 -5.71610868e-01 -2.78803557e-01
-2.83848822e-01 5.74430004e-02 -9.66180623e-01 5.33533752e-01
1.50096819e-01 -6.13004148e-01 1.56350628e-01 7.69702911e-01
-7.12054431e-01 9.04998362e-01 -3.43196303e-01 1.15921736e+00
-6.14698887e-01 1.09940851e+00 1.26524866e-01 -5.13486147e-01
1.06713724e+00 4.85035032e-01 5.19226313e-01 -4.01883632e-01
4.18973476e-01 7.37664402e-02 -3.72812450e-02 -6.50778949e-01
8.82935405e-01 -2.42257029e-01 -5.84352672e-01 8.82495821e-01
-6.83923960e-02 -3.97048891e-01 1.72269493e-02 6.54025018e-01
9.27875936e-01 -5.40548563e-01 3.35062534e-01 -5.53612888e-01
1.21666044e-01 3.22059602e-01 -4.97090295e-02 6.37982190e-01
-1.96952850e-01 1.90379784e-01 8.37988257e-01 -8.03081095e-02
-5.19655347e-01 -5.64018190e-01 -6.12770207e-02 1.58287454e+00
-3.48030895e-01 -7.55368352e-01 -7.43184507e-01 -2.45114818e-01
-2.53517106e-02 9.22184169e-01 -3.40550691e-01 2.58827955e-01
-1.18936211e-01 -7.86375880e-01 3.27351540e-01 -1.68845162e-01
5.81899166e-01 -1.23703742e+00 -1.33061275e-01 2.72908777e-01
-5.92403531e-01 -1.33165872e+00 -1.63267508e-01 -2.36148417e-01
-5.65870166e-01 -6.08858168e-01 -5.98848403e-01 -7.59788752e-01
2.19856277e-01 1.29157856e-01 1.35422850e+00 -1.36408150e-01
-2.62945313e-02 4.88197595e-01 -3.31074387e-01 -1.00652230e+00
-6.97393060e-01 6.34507358e-01 -2.61374891e-01 -2.23518267e-01
5.37256896e-01 -4.30801272e-01 -8.15537795e-02 7.17747360e-02
-4.04877484e-01 -2.51417398e-01 3.05761069e-01 1.92090735e-01
-3.49417388e-01 2.65785586e-02 1.12832189e+00 -1.27425802e+00
1.37696826e+00 -1.10731673e+00 1.70353219e-01 -1.06751040e-01
-4.13652360e-01 -5.93417466e-01 8.28652903e-02 -3.59270811e-01
-1.27778387e+00 -7.27557063e-01 -3.24439585e-01 6.05265260e-01
-2.37040401e-01 8.26224625e-01 2.42301568e-01 1.43145978e-01
7.78869092e-01 -1.03945419e-01 1.54191300e-01 -3.20386052e-01
2.14802966e-01 1.15838397e+00 -2.05325797e-01 -3.78606290e-01
6.41931474e-01 4.72526431e-01 -7.38356411e-01 -1.56294525e+00
-9.27707732e-01 -8.00247848e-01 -1.55859357e-02 -6.96066499e-01
8.74829292e-01 -8.25246155e-01 -1.17149949e+00 4.52509910e-01
-1.10294092e+00 -4.16649640e-01 -4.15539630e-02 3.24456662e-01
8.04836154e-02 2.94423569e-02 -9.86129463e-01 -1.16684902e+00
-3.36861968e-01 -6.81344807e-01 8.22398067e-01 1.94738001e-01
-9.33166742e-01 -1.36474144e+00 8.79874453e-02 9.07135308e-01
8.10082614e-01 5.61158638e-03 1.05638611e+00 -9.85389054e-01
-1.79617435e-01 -4.83405501e-01 -8.71872753e-02 -5.55336326e-02
7.21234828e-02 -3.83382253e-02 -8.77680421e-01 1.87362865e-01
-6.70544729e-02 -6.59099638e-01 2.06642762e-01 1.73957407e-01
7.16746867e-01 -6.04244649e-01 -3.40084791e-01 -7.32062936e-01
1.02595556e+00 -4.40697325e-03 4.64906096e-01 5.55193245e-01
2.54491091e-01 9.37195063e-01 4.97506917e-01 8.12952578e-01
8.02245140e-01 2.76220024e-01 5.92446513e-03 2.09439903e-01
3.97188872e-01 -5.12885988e-01 6.63309216e-01 9.76202369e-01
1.77084029e-01 -4.28797215e-01 -1.27343857e+00 5.55242181e-01
-1.66964030e+00 -1.05529404e+00 -4.40439016e-01 1.54822564e+00
7.35660374e-01 3.59637767e-01 2.59124339e-01 -2.74965107e-01
3.66274387e-01 3.73851538e-01 4.41793241e-02 -7.46374130e-01
3.11824922e-02 2.78178573e-01 3.33758056e-01 3.15879494e-01
-5.74403763e-01 1.01346159e+00 7.27250814e+00 8.46789956e-01
-1.02258718e+00 3.39822859e-01 9.91358161e-01 2.54557371e-01
-2.83122480e-01 6.13511167e-02 -7.77625978e-01 3.67107481e-01
1.21384716e+00 -5.63661873e-01 1.60192810e-02 1.02867806e+00
6.57004893e-01 -4.78323430e-01 -2.31868610e-01 5.94968259e-01
1.44589260e-01 -1.52746451e+00 -2.03873605e-01 3.16277355e-01
8.14823151e-01 4.27560687e-01 1.85813114e-01 7.03708231e-01
5.61125636e-01 -8.73601496e-01 6.04800284e-01 1.45022839e-01
2.68538296e-01 -5.79587936e-01 8.06701303e-01 8.66396964e-01
-7.02472508e-01 7.05362707e-02 -1.04852960e-01 -8.41937661e-01
3.48591864e-01 6.43660367e-01 -1.08633292e+00 3.84951308e-02
6.03680134e-01 4.95276183e-01 -5.68894804e-01 4.54640925e-01
5.61914630e-02 1.21323729e+00 -1.77264661e-01 -7.10814059e-01
4.70226854e-01 -3.93241495e-01 3.60220075e-01 1.51341724e+00
1.61934420e-01 1.58286020e-01 1.89234525e-01 6.15306675e-01
-2.68280357e-01 5.49872875e-01 -8.83012891e-01 -5.22276938e-01
6.64122581e-01 1.37141132e+00 -8.66833627e-01 -3.52435380e-01
-7.52806485e-01 3.59134912e-01 1.42010599e-01 2.03484222e-01
-4.18111920e-01 -2.62211293e-01 6.52162671e-01 3.97935748e-01
-3.74930441e-01 -1.48839340e-01 -5.71767747e-01 -8.18151295e-01
-3.47218424e-01 -7.53031015e-01 1.36330828e-01 -7.18547106e-01
-1.57281160e+00 3.34701270e-01 -9.43606645e-02 -7.43904471e-01
-2.12763220e-01 -3.26541394e-01 -1.04263794e+00 7.50708938e-01
-1.11149013e+00 -8.79881322e-01 -8.67638588e-02 4.30106640e-01
6.11617327e-01 -1.86852470e-01 6.94657743e-01 2.90553719e-01
-3.52226317e-01 2.36597791e-01 -2.21728832e-01 1.38323784e-01
7.50656545e-01 -7.46962965e-01 3.75655025e-01 1.21797256e-01
-3.90308827e-01 7.52791703e-01 7.03493416e-01 -6.94020987e-01
-1.13618016e+00 -9.18772519e-01 1.85040069e+00 -6.75831079e-01
1.34890306e+00 -4.58008349e-01 -5.83143234e-01 6.87989593e-01
8.10669363e-01 -9.97887790e-01 9.85353112e-01 6.32550359e-01
-9.19726193e-02 4.93206143e-01 -1.04652178e+00 6.81317329e-01
4.04501677e-01 -8.83247554e-01 -5.66466153e-01 7.90512383e-01
6.15618050e-01 -9.48404893e-02 -1.03372681e+00 -5.34071505e-01
4.10532355e-01 -6.57526314e-01 6.09610975e-01 -6.52204216e-01
7.98891187e-01 6.24762177e-01 -3.26200947e-02 -1.06152761e+00
-8.99240077e-02 -1.04468274e+00 4.82518703e-01 1.82508612e+00
7.20911801e-01 -1.13482249e+00 1.03001451e+00 1.16039312e+00
4.98856837e-03 -6.37385190e-01 -7.83865988e-01 -3.52877378e-01
3.74806345e-01 -6.65688813e-01 1.62339315e-01 1.30720448e+00
7.49158621e-01 9.96040404e-01 -3.00797671e-01 -5.60649037e-01
1.18707076e-01 -1.44318938e-01 1.02374876e+00 -1.49693716e+00
1.82096213e-01 -4.88597095e-01 -9.03708339e-02 -7.42384732e-01
1.64264202e-01 -7.24227667e-01 -1.81230664e-01 -1.80454779e+00
4.35633600e-01 -2.48753965e-01 4.82302397e-01 3.95896494e-01
1.14388250e-01 1.26806870e-01 1.25557944e-01 2.42303293e-02
-6.98865354e-01 2.63048619e-01 1.10526860e+00 1.03596501e-01
-1.64861143e-01 2.94972271e-01 -1.36563420e+00 1.09574926e+00
1.13585544e+00 -4.57508057e-01 -1.66558206e-01 5.03767096e-02
7.56463230e-01 -1.71720907e-02 2.56419480e-02 -5.04244447e-01
-2.57905889e-02 -3.16006780e-01 -4.26589280e-01 -5.57324111e-01
4.04967904e-01 -4.51160520e-01 -3.50793302e-01 2.65064210e-01
-7.66736567e-01 2.07316384e-01 1.01098120e-01 2.11401626e-01
-4.15657341e-01 9.60515216e-02 4.20800418e-01 -9.24286693e-02
-2.25558415e-01 9.74709615e-02 -1.43749011e+00 5.70212960e-01
5.47973156e-01 -1.74352210e-02 -7.10927904e-01 -1.14932609e+00
-2.84006804e-01 2.81439096e-01 -1.30148577e-02 6.54757559e-01
1.38455659e-01 -1.02832830e+00 -7.88070679e-01 -3.65949422e-01
1.88780501e-01 -5.55759728e-01 1.66783288e-01 9.56166029e-01
-1.81871191e-01 5.84479511e-01 2.60597169e-01 -1.37125686e-01
-1.10212147e+00 -2.34315887e-01 -2.78018653e-01 -3.94104242e-01
-2.87437499e-01 7.56229579e-01 -1.72233477e-01 -6.86172724e-01
-1.60988867e-02 -3.37989807e-01 -3.29219639e-01 4.17682976e-01
4.75366592e-01 6.14521503e-01 -1.73187867e-01 -6.90944910e-01
-1.51827246e-01 -1.95209533e-01 2.78226268e-02 -2.76273906e-01
1.26412570e+00 -1.19905747e-01 -4.58141625e-01 9.40343082e-01
1.33402753e+00 3.66157740e-01 -2.50809044e-01 -1.43217936e-01
4.43790346e-01 -3.96605909e-01 -1.89344555e-01 -7.22080410e-01
-5.80483496e-01 6.45331144e-01 -1.36730388e-01 9.06320930e-01
2.52005428e-01 4.46642220e-01 7.74271786e-01 5.83626091e-01
2.94700354e-01 -1.33014786e+00 2.15765104e-01 9.34938192e-01
7.42975712e-01 -1.32128727e+00 -1.43182114e-01 -6.18660688e-01
-1.05359709e+00 7.87977040e-01 4.62204874e-01 2.33605951e-01
1.11720324e+00 1.97268039e-01 4.03522134e-01 -5.93143761e-01
-1.02046394e+00 1.85582712e-01 -2.99966633e-01 2.69702315e-01
1.00955117e+00 1.15444630e-01 -5.89539051e-01 7.09049582e-01
-8.27360511e-01 2.52854191e-02 7.50653028e-01 7.48896539e-01
-5.80864131e-01 -9.83892381e-01 -3.05279016e-01 7.28344917e-01
-1.06896150e+00 -1.38123900e-01 -8.66148412e-01 1.09056747e+00
-4.33229208e-01 1.49362206e+00 -3.91855128e-02 -3.07878762e-01
1.40122965e-01 4.36196625e-01 -4.41674232e-01 -9.35605288e-01
-1.16616058e+00 -2.21880704e-01 7.55690396e-01 -1.88359916e-01
-4.83506918e-01 -9.11577821e-01 -9.75699902e-01 -7.86594450e-01
-4.15378064e-01 6.88765645e-01 7.27851033e-01 1.05572319e+00
4.36193228e-01 1.61095291e-01 4.12532359e-01 -3.15438092e-01
-1.59501091e-01 -1.45708025e+00 -5.82397223e-01 1.77570954e-01
-2.00643331e-01 -2.08303615e-01 -2.24551886e-01 -2.34296054e-01] | [10.758415222167969, 7.309314250946045] |
a24faf5e-9330-448a-8d0b-ec5d7243d7aa | a-survey-of-some-density-based-clustering | 2306.09256 | null | https://arxiv.org/abs/2306.09256v1 | https://arxiv.org/pdf/2306.09256v1.pdf | A Survey of Some Density Based Clustering Techniques | Density Based Clustering are a type of Clustering methods using in data mining for extracting previously unknown patterns from data sets. There are a number of density based clustering methods such as DBSCAN, OPTICS, DENCLUE, VDBSCAN, DVBSCAN, DBCLASD and ST-DBSCAN. In this paper, a study of these methods is done along with their characteristics, advantages and disadvantages and most importantly, their applicability to different types of data sets to mine useful and appropriate patterns. | ['Samarjeet Borah', 'Rupanka Bhuyan'] | 2023-06-15 | null | null | null | null | ['clustering'] | ['methodology'] | [-7.44591117e-01 -5.13089597e-01 -7.96090364e-02 -5.71107745e-01
1.35572508e-01 -5.10049462e-01 4.98859674e-01 3.98231745e-01
-1.95222512e-01 1.05272460e+00 2.76352078e-01 -2.73427069e-01
-1.07689941e+00 -9.89642203e-01 3.32379565e-02 -8.47866416e-01
-5.08010745e-01 1.16996503e+00 8.16627145e-01 2.44217053e-01
5.99476159e-01 9.46351349e-01 -1.92884827e+00 1.32316083e-01
7.59193361e-01 8.43736112e-01 5.53284287e-02 4.80546057e-01
-5.91999888e-01 5.89369774e-01 -8.49479079e-01 2.54685104e-01
1.25709474e-01 -3.09512645e-01 -4.10585850e-01 -1.45852759e-01
-5.50636530e-01 -8.02123845e-02 -2.00116977e-01 6.95113063e-01
3.59448642e-01 4.29706931e-01 1.23454082e+00 -1.51220453e+00
-4.43505913e-01 5.59223890e-01 -1.05890954e+00 9.54922318e-01
3.85688543e-01 -2.94672370e-01 5.56924306e-02 -8.18919301e-01
4.45546925e-01 1.29741657e+00 6.34425282e-01 1.02983266e-01
-8.34041178e-01 -6.57865405e-01 -4.86197650e-01 6.38969243e-01
-1.90068901e+00 -1.08966105e-01 5.24698734e-01 -3.88499856e-01
9.69682157e-01 2.95498729e-01 8.23118627e-01 3.14759851e-01
3.66415493e-02 5.37191749e-01 1.18286192e+00 -2.93889642e-01
9.39792037e-01 2.82932460e-01 3.82445753e-01 -1.52560756e-01
7.57576168e-01 1.74897760e-02 -3.45976889e-01 -3.75946730e-01
4.43440974e-01 1.97222695e-01 -1.05327450e-01 -4.45014298e-01
-5.55400014e-01 9.45953727e-01 1.67233106e-02 6.02076054e-01
-2.65363425e-01 -4.68156397e-01 2.69904077e-01 -8.32784250e-02
2.04124570e-01 -5.63060790e-02 -1.71193764e-01 -4.04673576e-01
-1.39947283e+00 2.71053046e-01 8.35697472e-01 1.25468981e+00
9.05003250e-01 3.04002594e-02 3.17943543e-01 7.27833152e-01
5.74990690e-01 4.71682578e-01 5.97851634e-01 -4.29105669e-01
-4.31663431e-02 8.39323282e-01 1.90982431e-01 -1.11176622e+00
-5.28276145e-01 6.32571518e-01 -9.62542772e-01 8.01523551e-02
7.71772042e-02 -2.67034382e-01 -1.07463408e+00 5.72071433e-01
5.83849072e-01 5.18577278e-01 8.24690312e-02 5.45873225e-01
8.33545089e-01 8.90760541e-01 -2.24587217e-01 -5.31805634e-01
6.80758417e-01 7.78033882e-02 -7.02096820e-01 3.44824135e-01
1.32412314e-01 -5.65550327e-01 3.72926146e-01 7.78440893e-01
-7.01490760e-01 -2.61406988e-01 -7.48868644e-01 4.78934079e-01
-7.34600961e-01 -5.21370888e-01 6.77830219e-01 9.50788736e-01
-1.02900314e+00 4.67652440e-01 -9.41698551e-01 -8.47621739e-01
5.15643716e-01 4.97668713e-01 1.02456860e-01 -1.37286171e-01
-5.99799931e-01 7.56687582e-01 9.44950283e-01 -3.79379034e-01
-4.86660749e-01 -4.96953785e-01 -5.26943386e-01 -2.01630026e-01
-6.02080859e-03 -4.52024769e-03 5.96339464e-01 1.29405633e-02
-9.14029419e-01 5.36808312e-01 -3.77585292e-01 -5.74953735e-01
-1.67191122e-02 3.25605333e-01 -9.54486549e-01 -5.61222062e-02
5.98558858e-02 2.16091543e-01 1.23757452e-01 -1.21246350e+00
-9.91573811e-01 -8.31655025e-01 -9.48153555e-01 2.63736602e-02
8.64651427e-02 4.03355621e-02 -2.08485916e-01 -2.45114416e-01
2.15654969e-01 -3.75265032e-01 -2.16667593e-01 -7.26068497e-01
-4.77503002e-01 -7.62828112e-01 1.61576509e+00 -1.99816197e-01
1.63240409e+00 -2.07681942e+00 -6.28173798e-02 1.02220917e+00
1.07871159e-03 1.51764870e-01 7.85447121e-01 7.75550306e-01
2.78722793e-01 6.27132729e-02 -3.75716776e-01 3.09120417e-01
-3.90206307e-01 8.87142301e-01 2.84631550e-01 6.71808541e-01
-3.00924540e-01 -4.84956615e-03 -7.28762388e-01 -6.58958435e-01
8.03832650e-01 3.16144526e-01 -1.91257209e-01 -3.61346006e-02
1.28788248e-01 2.07779050e-01 -3.75409722e-01 1.07577980e+00
1.09437776e+00 3.93341720e-01 1.53126031e-01 2.04870641e-01
-5.76417208e-01 -1.48632765e-01 -1.77275920e+00 1.01735795e+00
2.41444007e-01 6.35821998e-01 3.69330831e-02 -1.40692914e+00
1.38538539e+00 3.32111828e-02 8.49976301e-01 -2.91544437e-01
2.52224356e-01 2.20062509e-01 1.32997766e-01 -5.65223217e-01
4.26493257e-01 -1.13874607e-01 8.94644186e-02 5.08400559e-01
1.73664942e-01 5.05636111e-02 5.95240176e-01 2.41012692e-01
1.14463973e+00 -6.84719801e-01 5.85657835e-01 -8.62642825e-01
3.64808828e-01 5.92740774e-01 6.53854907e-01 3.91435653e-01
-1.62398636e-01 5.53068399e-01 1.64179176e-01 -3.00387442e-01
-1.29418683e+00 -1.27624559e+00 -4.67237055e-01 2.32543558e-01
3.08455259e-01 -3.74159724e-01 -3.85577738e-01 -2.43462488e-01
3.87014896e-01 8.07223320e-01 -3.91857833e-01 1.02174222e-01
-2.73799539e-01 -1.07010996e+00 2.70554245e-01 4.52743024e-01
4.86316711e-01 -9.05449450e-01 -4.79723394e-01 1.60994291e-01
4.83738899e-01 -4.85606849e-01 1.82169288e-01 1.80256873e-01
-1.19932246e+00 -1.41141808e+00 -2.04636171e-01 -6.24006271e-01
5.12106001e-01 3.64992112e-01 9.58915174e-01 -1.83421388e-01
-5.33485234e-01 1.15094014e-01 -6.61033273e-01 -8.00427079e-01
-1.75413638e-01 -1.76798075e-01 5.79563498e-01 -3.49816769e-01
1.54317963e+00 -1.15353954e+00 -5.15591681e-01 5.24132550e-01
-6.31482661e-01 -9.00566161e-01 3.36284757e-01 5.38257062e-02
7.82258749e-01 9.00479078e-01 6.47063136e-01 -8.95442843e-01
1.07760990e+00 -1.28230345e+00 -6.45819902e-01 -1.29109591e-01
-9.58637893e-01 -3.31199259e-01 3.66829842e-01 -3.48868556e-02
-7.16666937e-01 -1.81600973e-02 1.82971910e-01 -6.95054173e-01
-6.76747262e-01 4.59412098e-01 -1.90829501e-01 4.40322310e-01
9.44550037e-01 3.78580809e-01 2.97665685e-01 -8.27890933e-01
4.49366029e-03 1.17806280e+00 3.93663555e-01 -1.66042969e-01
5.57933927e-01 5.85330963e-01 -2.09291831e-01 -1.15263176e+00
-2.51412969e-02 -1.33381069e+00 -1.07373190e+00 -3.37081909e-01
8.35177243e-01 -4.27776694e-01 -5.13431609e-01 3.09844762e-01
-4.77035999e-01 1.78726643e-01 -2.97844082e-01 5.92826247e-01
-1.70484528e-01 2.68219292e-01 -1.09458655e-01 -1.29659510e+00
-2.17038736e-01 -5.93402147e-01 3.83751839e-01 6.46797180e-01
-4.32499111e-01 -1.08484364e+00 3.01094204e-01 -2.68999368e-01
4.10260737e-01 5.80583096e-01 6.87108278e-01 -1.03649819e+00
-4.70640212e-02 -9.13172960e-02 -3.80450557e-03 6.21941127e-03
5.96376836e-01 4.76668358e-01 -2.06215322e-01 -9.38352291e-03
-8.12722817e-02 1.73663810e-01 1.34330928e-01 6.55469894e-01
1.06455159e+00 -4.51087952e-01 -1.06366038e+00 4.00278658e-01
1.70028329e+00 1.18905962e+00 1.01619947e+00 3.68023276e-01
2.52648324e-01 3.01550984e-01 5.79322219e-01 7.90439963e-01
3.82085770e-01 2.27361009e-01 6.05590641e-02 2.89329767e-01
1.76486626e-01 1.07679062e-01 -3.65940511e-01 5.92301250e-01
-1.38982818e-01 -9.36030149e-02 -1.37640011e+00 8.33971143e-01
-1.89337063e+00 -1.54436350e+00 -7.41768479e-01 2.22416615e+00
4.51778382e-01 -1.85553618e-02 9.07864869e-01 7.78586030e-01
1.05343187e+00 -4.88321006e-01 -5.77526629e-01 -5.61042488e-01
5.81113435e-02 3.31658900e-01 7.18631864e-01 4.03392762e-02
-9.39901173e-01 5.89624286e-01 7.55367041e+00 6.52189851e-01
-7.14869559e-01 -1.87595949e-01 3.21352750e-01 -2.97651976e-01
1.30632490e-01 -5.56079559e-02 -8.85089278e-01 1.14065874e+00
1.09356260e+00 -6.94743156e-01 2.83132792e-01 1.07294750e+00
4.80478764e-01 -9.02756035e-01 -6.78149700e-01 1.14143133e+00
4.12870906e-02 -1.31798959e+00 -1.68813288e-01 3.05975586e-01
6.60161018e-01 1.52989134e-01 -5.98980546e-01 4.16397024e-03
9.37432349e-01 -1.07399046e+00 7.32780620e-02 6.23934209e-01
4.13608283e-01 -1.46506059e+00 8.90735209e-01 4.38370287e-01
-9.05824006e-01 -1.93789780e-01 -6.81193590e-01 -1.76221326e-01
4.74301040e-01 1.20735466e+00 -1.13433182e+00 6.65942132e-01
1.42567778e+00 8.00300241e-01 -2.31865391e-01 1.71241832e+00
2.88633704e-01 7.11981356e-01 -1.01060081e+00 -5.00978529e-01
1.43917665e-01 -6.68702841e-01 2.51820534e-01 1.15095198e+00
7.97180772e-01 3.34913880e-01 -7.45073408e-02 8.14324677e-01
5.84733009e-01 -1.57949463e-01 -8.34699512e-01 2.74673343e-01
1.45920837e+00 9.61605489e-01 -9.03648138e-01 -3.46850932e-01
-9.61502567e-02 2.44181931e-01 -2.07320824e-01 -8.63157138e-02
-5.80690801e-01 -6.42576993e-01 6.36642098e-01 9.29933965e-01
2.98934758e-01 -4.18099225e-01 -5.24684966e-01 -5.44710040e-01
-5.48281789e-01 -4.23107803e-01 9.26291287e-01 -5.42489886e-01
-1.56385577e+00 1.01631925e-01 8.96276534e-01 -1.28094101e+00
-2.34447718e-01 -3.53018105e-01 -1.06635153e+00 6.12998664e-01
-5.43047667e-01 -2.38580644e-01 -4.65162963e-01 1.10910225e+00
3.60993087e-01 -7.11763382e-01 4.78635877e-01 4.26723450e-01
-3.78328651e-01 -1.31832138e-01 8.71124864e-01 6.72128871e-02
2.45058253e-01 -1.40791738e+00 -1.11834131e-01 5.65204799e-01
-1.74210101e-01 4.25225377e-01 8.44210386e-01 -7.86840141e-01
-1.20785081e+00 -9.20425653e-01 2.52952844e-01 -3.34049851e-01
3.98130000e-01 -3.91723514e-01 -9.28998232e-01 2.96053767e-01
1.08136073e-01 -3.05694669e-01 9.68507111e-01 1.38537809e-01
4.16907489e-01 -2.53004104e-01 -1.86198592e+00 4.71579805e-02
7.75740027e-01 3.19052547e-01 -7.82946944e-01 1.95072308e-01
-3.02858651e-01 7.17014521e-02 -1.06400526e+00 1.11375242e-01
8.24783891e-02 -1.55053139e+00 7.36098647e-01 -7.87222683e-02
-1.49392277e-01 -6.69660687e-01 -7.98845887e-02 -1.41807806e+00
-4.17843491e-01 -9.88211185e-02 -2.20919907e-01 1.53769064e+00
3.15463722e-01 -3.55974257e-01 1.19943964e+00 3.71948391e-01
-2.06748154e-02 -2.42704704e-01 -1.25197244e+00 -8.20878923e-01
-8.87814313e-02 -1.81413725e-01 7.33881176e-01 1.17628300e+00
5.58921576e-01 1.87195942e-01 -3.17400247e-02 -8.02393854e-02
8.13198745e-01 -2.81038225e-01 8.88151050e-01 -1.77344108e+00
2.85843164e-01 -1.95570827e-01 -1.00360227e+00 -3.41000646e-01
-4.80743408e-01 -3.12775016e-01 -2.35895559e-01 -2.05715108e+00
-2.05221578e-01 -9.83341336e-01 1.26237854e-01 3.49383168e-02
3.12414527e-01 -2.59604454e-01 -4.60238099e-01 8.10512483e-01
-3.87541384e-01 8.56714323e-03 4.37273979e-01 1.99526340e-01
-5.97140312e-01 1.54815063e-01 -5.30823588e-01 4.53043610e-01
1.05129671e+00 -5.55349290e-01 -8.70739877e-01 1.65789247e-01
-3.23442280e-01 -2.64757901e-01 -4.41808671e-01 -1.29266798e+00
6.42491758e-01 -5.59893966e-01 7.13287055e-01 -1.31863940e+00
4.12211381e-02 -9.22825038e-01 5.43218076e-01 2.60286719e-01
7.02018201e-01 4.14375544e-01 1.62788808e-01 8.04302812e-01
-4.88837361e-01 -2.69792318e-01 8.97909760e-01 -2.01917246e-01
-1.15668213e+00 1.77764818e-01 -9.07412767e-01 -1.65321261e-01
1.87754202e+00 -9.16126907e-01 -1.52406633e-01 -2.05920562e-01
-7.77119637e-01 4.02606666e-01 4.82107639e-01 2.08519176e-01
6.43453240e-01 -1.41728997e+00 -2.23857760e-01 4.46380526e-02
-1.58769786e-01 5.29944003e-01 1.53772786e-01 8.04062247e-01
-8.02906930e-01 3.49852592e-01 -4.68435645e-01 -8.53355885e-01
-1.26718616e+00 8.77903700e-01 1.25281125e-01 3.77097994e-01
-6.68351352e-01 4.40681845e-01 -6.77829683e-01 -1.35375708e-01
9.66037586e-02 -1.54904023e-01 -5.14251053e-01 1.51600599e-01
6.90922916e-01 1.01275706e+00 9.22189206e-02 -5.71266472e-01
-9.37496006e-01 4.76931483e-01 1.57162234e-01 -2.03708876e-02
1.44749379e+00 -9.06022936e-02 -2.24529594e-01 6.59254313e-01
7.95529127e-01 -3.88925403e-01 -7.95455456e-01 2.69512534e-01
7.10082531e-01 -6.84338033e-01 -1.91288874e-01 -6.83073163e-01
-8.24219704e-01 4.07170385e-01 7.01120555e-01 7.59129703e-01
1.09004223e+00 2.90559024e-01 2.54930854e-01 1.04585486e-02
5.56672931e-01 -1.49827671e+00 -4.26670730e-01 3.35647494e-01
3.93623739e-01 -9.76110578e-01 1.44124374e-01 -1.88647673e-01
-6.08953238e-01 9.99264419e-01 6.03074431e-01 -4.61531490e-01
1.48932588e+00 7.88546860e-01 -3.85888547e-01 -5.04826427e-01
-5.61772406e-01 -1.04600020e-01 -2.18676209e-01 1.50179410e+00
3.35833609e-01 6.33629858e-02 -4.42779899e-01 3.95176470e-01
-4.48859692e-01 1.21016696e-01 7.12644935e-01 1.33169341e+00
-1.04379392e+00 -8.36210012e-01 -7.73238122e-01 9.77643728e-01
-2.24368513e-01 3.77334237e-01 -7.43097782e-01 1.28819013e+00
2.35344812e-01 1.21128631e+00 7.57386208e-01 -6.23364329e-01
5.25241911e-01 -7.23952651e-02 8.97469074e-02 -6.15284979e-01
-1.53888255e-01 -1.73770845e-01 -5.98602518e-02 -7.45305419e-02
-7.67485201e-01 -9.14019644e-01 -1.53651214e+00 -1.09246993e+00
-2.42783204e-01 7.80475676e-01 6.17018938e-01 6.86223567e-01
5.71342826e-01 3.41666117e-02 5.39216876e-01 -4.91026938e-01
3.57191682e-01 -1.28732133e+00 -1.39310849e+00 8.39981139e-02
-2.39688203e-01 -9.99626994e-01 -4.48035955e-01 -8.56113061e-02] | [7.478668689727783, 4.542597770690918] |
59218a8e-d616-45be-a15a-f0db0a49d3db | multimodal-pathophysiological-dataset-of | 2002.09154 | null | https://arxiv.org/abs/2002.09154v2 | https://arxiv.org/pdf/2002.09154v2.pdf | Multimodal pathophysiological dataset of gradual cerebral ischemia in a cohort of juvenile pigs | Ischemic brain injuries are frequent and difficult to detect reliably or early. We present the multi-modal data set containing cardiovascular (blood pressure, blood flow, electrocardiogram) and brain electrical activities to derive electroencephalogram (EEG) biomarkers of corticothalamic communication under normal, sedation and hypoxic/ischemic conditions with ensuing recovery. We provide technical validation using EEGLAB. We also delineate the corresponding changes in the electrocardiogram (ECG)-derived heart rate variability (HRV) with the potential for future in-depth analyses of joint EEG-ECG dynamics. We review an open-source methodology to derive signatures of coupling between the ECoG and electrothalamogram (EThG) signals contained in the presented data set to better characterize the dynamics of thalamocortical communication during these clinically relevant states. The data set is presented in full band sampled at 2000 Hz, so the additional potential exists for insights from the full-band EEG and high-frequency oscillations under the bespoke experimental conditions. Future studies on the dataset may contribute to the development of new brain monitoring technologies, which will facilitate the prevention of neurological injuries. | ['Reinhard Bauer', 'Christophe L. Herry', 'Bernd Walter', 'Martin G. Frasch'] | 2020-02-21 | null | null | null | null | ['heart-rate-variability'] | ['medical'] | [ 6.81339353e-02 -4.81847554e-01 5.84474444e-01 -1.81745395e-01
-2.88569123e-01 -3.84915441e-01 1.55549601e-01 2.91340351e-01
-4.66242880e-01 9.29117322e-01 3.22094440e-01 3.45375948e-02
-5.67887604e-01 -4.27044719e-01 -2.88097233e-01 -9.36054230e-01
-9.45887566e-01 -1.73308089e-01 -3.55043143e-01 4.44369838e-02
2.74014741e-01 4.77653682e-01 -1.37138617e+00 2.07107380e-01
7.05698133e-01 1.13055789e+00 9.04587507e-02 6.39088750e-01
5.08049309e-01 2.56113797e-01 -8.92708004e-01 2.01545283e-01
-4.60415147e-02 -6.41775966e-01 -4.28267241e-01 -3.03069443e-01
-3.50917727e-01 -1.81398287e-01 -3.62094402e-01 7.43694842e-01
1.04493022e+00 -2.21894339e-01 3.46819401e-01 -1.02157378e+00
9.40367803e-02 4.02862132e-01 -2.70977706e-01 1.07406926e+00
2.74445385e-01 2.27905512e-01 -2.14944053e-02 -5.95878661e-01
3.32269639e-01 3.10712487e-01 5.58986723e-01 1.21360589e-02
-1.24253821e+00 -7.94299424e-01 -3.01353067e-01 6.69616818e-01
-1.49456084e+00 -3.27823013e-01 7.37834632e-01 -4.80428517e-01
9.98503387e-01 1.84042588e-01 1.34102607e+00 1.41353881e+00
8.95634472e-01 -3.74503106e-01 1.55454791e+00 1.93344860e-03
1.77250654e-01 5.36300465e-02 7.49087870e-01 1.19448461e-01
3.71188581e-01 2.84725666e-01 -1.00249028e+00 -5.28473496e-01
5.57441354e-01 -2.58147210e-01 -1.18097091e+00 7.56656975e-02
-1.47959828e+00 4.26615506e-01 -1.47757486e-01 4.31958288e-01
-8.91703129e-01 1.13591179e-01 6.31723464e-01 6.09360516e-01
2.79679716e-01 4.66695160e-01 -5.85894108e-01 -6.52970612e-01
-8.59564364e-01 -2.51067221e-01 7.27509975e-01 5.04343092e-01
4.36419696e-01 1.22479573e-01 -8.16533938e-02 6.47710443e-01
1.51493549e-01 4.23028439e-01 7.09569633e-01 -1.03314781e+00
1.73886910e-01 -7.86059946e-02 -7.18923211e-02 -7.29496419e-01
-9.12536919e-01 -5.01269937e-01 -7.40798533e-01 -1.04538269e-01
1.60468444e-01 -5.78833938e-01 1.20286115e-01 1.36532462e+00
-1.06808662e-01 2.46610746e-01 -2.36456767e-01 9.26861882e-01
5.62186778e-01 1.09050646e-01 1.05359256e-01 -7.60104299e-01
1.63321221e+00 2.90993869e-01 -8.68325412e-01 -4.61593196e-02
2.71019906e-01 -3.87133546e-02 5.29564202e-01 5.73052108e-01
-9.02068496e-01 -2.98740834e-01 -1.02064109e+00 7.20624447e-01
-1.07556939e-01 -3.25298786e-01 2.98348099e-01 8.74099255e-01
-1.01631570e+00 6.72941148e-01 -1.23940659e+00 -3.82608175e-01
4.28484142e-01 3.64849478e-01 -4.75828975e-01 5.08580506e-01
-1.40462697e+00 1.31381166e+00 1.18069373e-01 1.85764074e-01
-8.26477706e-01 -9.13601995e-01 -3.17035496e-01 -5.21902256e-02
-4.35359925e-01 -6.72599494e-01 2.82097042e-01 -1.45714417e-01
-1.46177113e+00 6.35016143e-01 -2.58404408e-02 -4.01967049e-01
1.95524618e-01 8.18425883e-03 -3.74336660e-01 6.06831551e-01
-3.34030360e-01 2.66011693e-02 3.22416961e-01 -5.59953868e-01
2.66960174e-01 -9.66875196e-01 -6.52224600e-01 -2.07987009e-03
-2.67018616e-01 3.32413137e-01 7.26461828e-01 -4.40298617e-01
1.36984512e-01 -5.26737094e-01 2.38794923e-01 -6.57928407e-01
1.02989934e-01 5.07767141e-01 5.30671813e-02 -6.81977689e-01
9.65577900e-01 -2.08657694e+00 7.72036836e-02 2.64309585e-01
2.24421144e-01 -3.63401204e-01 2.15681456e-02 5.74040473e-01
-4.19161797e-01 -1.15293346e-01 -1.57473326e-01 2.27891326e-01
-3.95508260e-01 -3.75710636e-01 -2.39338040e-01 1.32202983e+00
-7.70918727e-02 8.95322144e-01 -4.90566730e-01 1.86206415e-01
2.18457326e-01 6.85324967e-01 5.28258942e-02 7.93088302e-02
8.29923630e-01 8.51390719e-01 -6.78324373e-03 2.02617869e-01
5.39635539e-01 4.29038674e-01 -1.75185546e-01 -2.97895104e-01
-2.69706190e-01 3.19227159e-01 -6.58219397e-01 1.63234329e+00
-8.99801776e-02 1.02135527e+00 1.53625712e-01 -1.30831087e+00
1.13060880e+00 7.93873787e-01 6.80157065e-01 -9.56628025e-01
7.47672915e-01 -2.32598633e-02 5.90207100e-01 -7.63283849e-01
-4.86084193e-01 -2.79260904e-01 3.26647043e-01 5.79131663e-01
4.12005723e-01 -3.27919483e-01 2.78144684e-02 -1.16358094e-01
1.07879639e+00 -8.84235576e-02 3.44827175e-01 -9.99337316e-01
4.13032860e-01 -4.77898449e-01 1.55390739e-01 5.25600970e-01
-5.13272643e-01 4.87373829e-01 6.66971028e-01 -1.39083013e-01
-3.03431928e-01 -8.96367729e-01 -8.70768011e-01 3.56636107e-01
4.26465161e-02 -2.80803651e-01 -7.31998563e-01 4.35151398e-01
-1.73184678e-01 7.14824378e-01 -4.93503064e-01 -7.13371158e-01
-2.17859581e-01 -1.51677072e+00 5.77619970e-01 4.24449980e-01
1.53529555e-01 -1.13621736e+00 -1.37974226e+00 5.69684148e-01
-3.23148787e-01 -1.11902153e+00 2.76487052e-01 6.54913068e-01
-1.12231040e+00 -1.07302964e+00 -5.63835561e-01 -2.05328912e-01
7.31231347e-02 -4.32738096e-01 6.50349319e-01 -3.10285270e-01
-6.88439190e-01 6.34657979e-01 -1.99946791e-01 -6.61119044e-01
5.49132638e-02 -5.09313881e-01 3.50676090e-01 3.45315300e-02
6.05073154e-01 -1.38930607e+00 -7.79054940e-01 1.87073737e-01
-2.14221045e-01 -2.78074175e-01 1.78308204e-01 3.78671020e-01
3.00853968e-01 -2.55567968e-01 1.33995378e+00 -7.38385543e-02
1.03230107e+00 -7.30472684e-01 -3.56681883e-01 -1.60886034e-01
-6.72303498e-01 -4.30869877e-01 6.07236445e-01 -3.21682274e-01
-7.00088382e-01 -6.24672890e-01 2.89829940e-01 6.55033812e-02
-5.60711801e-01 4.09093291e-01 5.86535446e-02 5.89971542e-02
9.32719767e-01 6.09738946e-01 2.40226034e-02 -4.00828309e-02
-3.91204059e-01 6.20270789e-01 1.01398969e+00 -4.99588311e-01
-9.43445638e-02 3.17742139e-01 5.30770347e-02 -1.04395735e+00
1.67792328e-02 -2.94965893e-01 -6.86654866e-01 -3.50907892e-01
8.84140790e-01 -1.03442609e+00 -7.78801501e-01 6.85035408e-01
-1.05746055e+00 -4.66514736e-01 8.34235996e-02 1.18636048e+00
-5.77549517e-01 3.96156043e-01 -7.64383674e-01 -1.07719755e+00
-8.62332523e-01 -8.13927174e-01 3.01601231e-01 -2.43668240e-02
-6.10305429e-01 -6.70479119e-01 4.26271290e-01 1.46188527e-01
5.26879966e-01 6.02081001e-01 8.50433350e-01 -4.31970268e-01
5.08682951e-02 -2.59986937e-01 3.97134751e-01 1.26694694e-01
1.24691622e-02 -2.38908157e-01 -1.01772296e+00 -6.43308163e-02
8.86116505e-01 3.98827270e-02 1.37818426e-01 6.97042048e-01
8.87503147e-01 4.19053525e-01 -1.34790018e-02 7.29543388e-01
1.17789447e+00 5.43073237e-01 8.22889864e-01 1.61983252e-01
-2.17168462e-02 7.18691885e-01 -2.64948994e-01 6.21673286e-01
1.84235066e-01 1.14541717e-01 2.06568211e-01 3.29341263e-01
1.76379681e-01 5.85395694e-01 4.19280231e-01 8.76141250e-01
-5.78169405e-01 1.38496101e-01 -9.42740142e-01 4.42423433e-01
-1.21404922e+00 -8.74410152e-01 -7.67747343e-01 2.40102983e+00
6.48296475e-01 -1.97217673e-01 2.30012089e-02 4.31811243e-01
6.60218775e-01 -4.90804046e-01 -4.36226696e-01 -4.67201918e-01
-2.08861217e-01 7.48709977e-01 2.73609042e-01 5.28114364e-02
-3.06454152e-01 -2.83388477e-02 6.90619278e+00 -4.05650586e-01
-1.24000192e+00 3.87763083e-01 2.83599436e-01 -4.99466956e-01
2.72474438e-01 -3.12471658e-01 -1.24829069e-01 8.03785920e-01
2.01207018e+00 -6.15863800e-01 8.15042436e-01 1.62895769e-01
7.22059131e-01 -5.83385587e-01 -1.06468129e+00 1.29742348e+00
-8.81222934e-02 -9.77939308e-01 -8.05805564e-01 1.25854865e-01
-5.69811370e-03 5.92060626e-01 -4.79552299e-01 -1.87176093e-01
-8.80147815e-01 -8.23810756e-01 7.60533988e-01 8.16369653e-01
1.04085982e+00 -7.47456849e-01 7.36763597e-01 4.81368601e-01
-8.54417324e-01 -1.14895768e-01 -2.10459784e-01 -2.43814871e-01
1.11548118e-01 7.29326546e-01 -6.15198128e-02 4.06262904e-01
6.62957489e-01 8.43625903e-01 -4.29319799e-01 1.29191411e+00
-6.51868433e-02 7.77532279e-01 -4.34453905e-01 1.99619159e-01
-6.28420353e-01 -4.28952903e-01 6.58871353e-01 9.58257675e-01
4.87709194e-01 4.57148969e-01 -7.31544614e-01 1.20855045e+00
5.21741092e-01 -1.47631019e-01 -5.67761004e-01 1.38797149e-01
3.39095652e-01 1.18895710e+00 -9.05571401e-01 -3.46131027e-02
-4.03109580e-01 5.34587681e-01 -1.31544262e-01 5.91276288e-01
-6.11158669e-01 -6.01722717e-01 5.75584352e-01 5.16677983e-02
-6.28635764e-01 -3.26512277e-01 -8.32492530e-01 -1.29507911e+00
5.86894751e-02 -5.00365436e-01 8.47888961e-02 -9.44276273e-01
-7.35645294e-01 6.66624367e-01 4.93301809e-01 -7.22234070e-01
-2.46972702e-02 -2.71586508e-01 -1.01683986e+00 1.47251892e+00
-1.25721228e+00 3.50279391e-01 -6.58298373e-01 8.99747968e-01
-8.94570276e-02 2.73953259e-01 1.25867307e+00 2.72221535e-01
-7.18858600e-01 -5.84754487e-03 -3.46331805e-01 -1.14464514e-01
4.94961947e-01 -7.68642426e-01 -9.92169380e-02 7.81535983e-01
-2.84618229e-01 6.62597358e-01 5.90872228e-01 -3.53700221e-01
-1.26667297e+00 -6.06828690e-01 4.07603502e-01 -3.31467688e-01
6.69034123e-01 -6.83065057e-01 -9.00931954e-01 1.11880459e-01
3.29675257e-01 -3.74615379e-02 8.95833433e-01 -3.69967312e-01
3.46753865e-01 -2.51915485e-01 -1.11898553e+00 1.49975136e-01
7.39797473e-01 -8.49562764e-01 -1.07271564e+00 6.15464859e-02
-8.92671719e-02 1.16172351e-01 -1.40192056e+00 3.19121093e-01
7.75705874e-01 -1.17494392e+00 5.50412118e-01 -1.49674609e-01
-1.67816445e-01 8.09988976e-02 2.72746533e-01 -1.78123188e+00
-1.37045696e-01 -1.01418233e+00 -3.26974271e-03 8.21004748e-01
-1.60754658e-02 -1.43838120e+00 3.23494263e-02 7.50949323e-01
-2.39877105e-01 -6.29535794e-01 -9.46422458e-01 -6.42330587e-01
1.11140288e-01 -6.32407606e-01 3.86200666e-01 3.69124025e-01
1.23759818e+00 9.89729315e-02 1.47861108e-01 -7.05578253e-02
4.52408403e-01 -1.89935207e-01 -1.35657534e-01 -1.20369828e+00
8.41586739e-02 -2.94627756e-01 -6.51635110e-01 1.94003567e-01
8.02445877e-03 -1.14652562e+00 -1.86732799e-01 -1.33014381e+00
2.31055617e-01 2.07646146e-01 -6.01214409e-01 2.68057548e-02
-2.61931978e-02 2.22922921e-01 -2.43834838e-01 -2.13338062e-01
3.80885452e-01 5.14541209e-01 7.14830875e-01 3.51805121e-01
-4.40485418e-01 -4.72183287e-01 -5.63316166e-01 3.40411037e-01
9.12690163e-01 -7.91003227e-01 -6.68955505e-01 2.03790311e-02
-4.72439788e-02 7.08263814e-01 3.88249576e-01 -1.37057304e+00
1.64710209e-01 4.15956587e-01 6.67865753e-01 -2.67846972e-01
1.55292839e-01 -6.41703784e-01 5.13982356e-01 6.03449702e-01
-9.38610062e-02 2.11595044e-01 2.37804145e-01 3.45105320e-01
-1.35757521e-01 2.74383873e-01 8.35056186e-01 -8.40802193e-02
1.35427803e-01 3.55872691e-01 -1.19739282e+00 1.88252494e-01
1.27632499e+00 -5.47422707e-01 -6.86787903e-01 -1.33663073e-01
-8.22755218e-01 -8.22398812e-02 -5.83159830e-03 1.26087770e-01
4.28374231e-01 -9.84851241e-01 -8.82205844e-01 6.93975031e-01
-8.64551309e-03 -8.01625252e-01 4.57835674e-01 2.04986429e+00
-3.45385224e-01 4.48407263e-01 -8.65655363e-01 -6.13801420e-01
-7.40577877e-01 4.96215560e-02 7.16898263e-01 6.13947451e-01
-1.01154649e+00 6.79543555e-01 -7.81629458e-02 8.80833924e-01
2.54593164e-01 -3.12789112e-01 -5.63698828e-01 5.51441193e-01
7.07935750e-01 5.95156848e-01 5.37983954e-01 -2.42113814e-01
-5.85573316e-01 2.46584520e-01 6.40477121e-01 -2.64744163e-01
1.47690475e+00 -4.00440007e-01 -5.87894261e-01 8.47706139e-01
1.08122051e+00 -5.54271281e-01 -8.81075025e-01 6.76197171e-01
-2.19864145e-01 -4.34769765e-02 1.82073846e-01 -9.29359972e-01
-1.21467459e+00 1.20597327e+00 9.36731637e-01 3.03346694e-01
1.29509866e+00 -3.92601520e-01 3.94981802e-01 1.83261648e-01
9.60431635e-01 -9.59786415e-01 -5.06129444e-01 -1.70261040e-01
1.02743268e+00 -2.38281816e-01 -4.37738709e-02 1.87499195e-01
-3.58811378e-01 1.50306189e+00 1.83042973e-01 -4.12074983e-01
1.13368118e+00 7.14325488e-01 4.51218970e-02 -4.23919678e-01
-8.64943922e-01 2.37893775e-01 -1.07921727e-01 8.62752616e-01
5.38798869e-01 4.39038575e-02 -6.84955895e-01 1.06613410e+00
-4.16394085e-01 1.66401967e-01 1.02627099e+00 6.64420784e-01
-1.31504506e-01 -3.79598379e-01 -4.90896463e-01 9.66598511e-01
-5.89982629e-01 -1.55838713e-01 -1.92613289e-01 4.67208922e-01
-5.32004498e-02 1.23436487e+00 1.28321322e-02 -1.63746208e-01
4.24934953e-01 9.05229390e-01 7.82874882e-01 -4.70079839e-01
-7.71859646e-01 7.76578635e-02 2.86260676e-02 -7.53323972e-01
-3.92605484e-01 -8.17039311e-01 -1.31328392e+00 2.05850467e-01
-2.78580070e-01 4.39377539e-02 9.38266873e-01 8.43612075e-01
6.06467366e-01 8.10394287e-01 2.67814636e-01 -9.09703612e-01
-3.00689545e-02 -1.41087377e+00 -1.09398723e+00 -6.12445585e-02
3.89549494e-01 -6.68477893e-01 -6.95974410e-01 3.23654823e-02] | [13.216248512268066, 3.3816192150115967] |
4a6997dc-bc46-49da-a2b0-706c8aad5b02 | deuteros-2-0-peptide-level-significance | 2005.08380 | null | http://arxiv.org/abs/2005.08380v1 | http://arxiv.org/pdf/2005.08380v1.pdf | Deuteros 2.0: Peptide-level significance testing of data from hydrogen deuterium exchange mass spectrometry | Summary: Hydrogen deuterium exchange mass spectrometry (HDX-MS) is becoming
increasing routine for monitoring changes in the structural dynamics of
proteins. Differential HDX-MS allows comparison of individual protein states,
such as in the absence or presence of a ligand. This can be used to attribute
changes in conformation to binding events, allowing the mapping of entire
con-formational networks. As such, the number of necessary cross-state
comparisons quickly increas-es as additional states are introduced to the
system of study. There are currently very few software packages available that
offer quick and informative comparison of HDX-MS datasets and even few-er which
offer statistical analysis and advanced visualization. Following the feedback
from our origi-nal software Deuteros, we present Deuteros 2.0 which has been
redesigned from the ground up to fulfil a greater role in the HDX-MS analysis
pipeline. Deuteros 2.0 features a repertoire of facilities for back exchange
correction, data summarization, peptide-level statistical analysis and advanced
data plotting features. Availability: Deuteros 2.0 can be downloaded from
https://github.com/andymlau/Deuteros_2.0 under the Apache 2.0 license.
Installation of Deuteros 2.0 requires the MATLAB Runtime Library available free
of charge from MathWorks
(https://www.mathworks.com/products/compiler/matlab-runtime.html) and is
available for both Windows and Mac operating systems. | [] | 2020-05-17 | null | null | null | null | ['data-summarization'] | ['miscellaneous'] | [ 9.58487913e-02 -3.85117978e-01 -1.84242904e-01 -5.11796236e-01
-4.37887698e-01 -7.11583614e-01 3.59078467e-01 7.74907172e-01
-3.80636722e-01 8.32538307e-01 -1.81487277e-01 -6.55672789e-01
2.57211309e-02 -3.98036420e-01 -4.61313516e-01 -6.24677837e-01
-2.55955070e-01 7.88002372e-01 3.12965035e-01 -1.59173116e-01
1.84053972e-01 6.92856550e-01 -1.24442744e+00 2.24037930e-01
4.84145254e-01 4.65715617e-01 6.51973188e-01 7.95239806e-01
-1.68840826e-01 1.10498898e-01 -1.95306271e-01 -5.28395735e-02
1.12596475e-01 -5.82745969e-01 -6.29614115e-01 -5.12879312e-01
6.51333630e-02 1.79855116e-02 -6.87086061e-02 9.19651985e-01
9.49868202e-01 -8.10586512e-02 8.07174891e-02 -1.06741202e+00
-1.73709050e-01 2.87799448e-01 -3.43213290e-01 7.05383420e-01
5.35608530e-01 4.60834622e-01 6.53472602e-01 -9.24592078e-01
1.00601518e+00 1.07687318e+00 4.02867913e-01 3.42676342e-01
-1.56941450e+00 -8.07741106e-01 -3.97782296e-01 1.18523045e-02
-1.11528683e+00 -6.65506303e-01 5.14705300e-01 -4.56464857e-01
1.27387202e+00 4.97202277e-01 5.94869673e-01 7.34255075e-01
2.65815735e-01 2.99605876e-01 1.37355006e+00 -3.26165766e-01
1.07289307e-01 -1.52101055e-01 2.76675075e-01 4.76890981e-01
2.03934580e-01 -2.37125568e-02 -6.18951976e-01 -4.90271598e-01
3.61652344e-01 1.40668780e-01 -1.70636639e-01 -5.38081944e-01
-1.34675992e+00 5.00683129e-01 -7.82090724e-02 2.64288127e-01
-4.71125811e-01 -2.20435798e-01 6.29061580e-01 4.17441726e-01
2.49177068e-01 3.42106789e-01 -7.14790165e-01 -3.99834484e-01
-7.84966350e-01 4.44633335e-01 7.00117111e-01 6.23609066e-01
5.33764899e-01 -5.17298162e-01 6.14925802e-01 7.70598769e-01
3.28120142e-01 9.03885141e-02 6.28816307e-01 -7.92830765e-01
-2.22073913e-01 5.29527128e-01 2.65287071e-01 -5.88370442e-01
-7.32035995e-01 -9.78685841e-02 -4.70777094e-01 3.46820801e-01
6.91305757e-01 -1.86783791e-01 -5.43903172e-01 1.42315328e+00
8.19887877e-01 -4.93501663e-01 -1.05394863e-01 9.72296536e-01
6.31608069e-01 4.76668656e-01 3.62411946e-01 -5.85007310e-01
1.38351524e+00 -2.53182411e-01 -6.72132671e-01 1.89841345e-01
4.92480457e-01 -1.06274617e+00 6.14473581e-01 4.16442454e-01
-1.27904868e+00 -3.12759072e-01 -1.09172559e+00 -1.29317893e-02
-7.90124714e-01 -4.25674736e-01 6.76587045e-01 1.75876305e-01
-8.85736525e-01 1.03893399e+00 -1.57139337e+00 -9.35044825e-01
1.74174890e-01 5.47692478e-01 -4.98196214e-01 2.47136399e-01
-1.23621571e+00 9.60697830e-01 6.34358168e-01 -1.87188730e-01
-3.18370789e-01 -9.26464379e-01 -5.67794383e-01 -3.20595235e-01
5.67177273e-02 -2.88253039e-01 1.27327538e+00 -4.34094548e-01
-1.05710506e+00 1.01917958e+00 -1.00659929e-01 -1.36381462e-01
3.51555139e-01 2.41383478e-01 -4.66982514e-01 4.19098824e-01
1.10216632e-01 4.80428994e-01 3.21114659e-02 -6.36148572e-01
-4.01157737e-01 -6.52336955e-01 -2.79194295e-01 1.18751854e-01
3.66940320e-01 6.08549178e-01 -3.49081941e-02 -3.35632414e-01
2.19449252e-01 -8.44177246e-01 -1.74362212e-01 1.94777161e-01
-2.27651879e-01 1.72963887e-01 7.44167328e-01 -6.79051578e-01
1.02296984e+00 -1.97311890e+00 -1.14889340e-02 -3.07289939e-02
2.85380065e-01 3.92851472e-01 3.33222598e-01 1.23920345e+00
-8.47246706e-01 1.46470480e-02 -1.63137674e-01 2.19663247e-01
1.65160626e-01 -2.72571534e-01 2.85695910e-01 7.39576101e-01
-2.26284768e-02 6.62839115e-01 -9.94491279e-01 -1.78764462e-01
4.10466343e-01 4.84893084e-01 2.65407506e-02 -7.21813040e-03
-2.75815930e-02 6.78303421e-01 -2.53269523e-01 5.83841801e-01
7.81126618e-01 -3.61345857e-01 1.00508714e+00 -2.26339355e-01
-7.92321801e-01 3.98294836e-01 -1.28830123e+00 1.49907243e+00
4.39324498e-01 1.95587113e-01 2.37010583e-01 -8.68680418e-01
1.02272856e+00 2.42047146e-01 6.79442644e-01 -4.55398768e-01
3.86091545e-02 1.73350886e-01 2.67897457e-01 -2.30262905e-01
9.35820937e-02 -3.66047114e-01 3.69537532e-01 5.23437679e-01
-1.04658060e-01 1.95143208e-01 4.94148612e-01 6.11862898e-01
9.96065676e-01 2.47195765e-01 8.00634205e-01 -5.64257801e-01
2.77849346e-01 3.48242134e-01 3.95414859e-01 1.08692437e-01
-3.84266973e-01 2.65807263e-03 6.60067737e-01 -3.65866542e-01
-1.37807810e+00 -8.95845234e-01 -6.96934760e-01 1.05552268e+00
-1.29634038e-01 -7.22418904e-01 -4.24105316e-01 1.56885535e-01
2.74354249e-01 3.79730225e-01 -2.97252834e-01 -2.09438149e-02
-1.10465147e-01 -9.36706960e-01 3.36557865e-01 1.48104504e-01
3.19579337e-03 -9.95060503e-01 -7.83414066e-01 2.81890094e-01
3.34247351e-01 -2.92878598e-01 -2.13764802e-01 7.45126069e-01
-7.64886260e-01 -1.23673511e+00 -3.65597099e-01 -3.29942316e-01
4.79912400e-01 1.39145657e-01 5.34794569e-01 5.73180132e-02
-7.39600480e-01 3.81213635e-01 -3.85730445e-01 -5.18055677e-01
-4.20724303e-01 -1.97770491e-01 3.81153882e-01 -6.14250898e-01
7.48808265e-01 -8.90827000e-01 -8.98521960e-01 1.68269202e-01
-9.60189342e-01 9.42005739e-02 3.05147499e-01 5.70453525e-01
9.44120884e-01 -2.02890098e-01 6.09665751e-01 -8.57802331e-01
6.11710489e-01 -6.30043209e-01 -7.65768588e-01 -1.29850075e-01
-9.11239743e-01 -1.67964458e-01 4.35948253e-01 -8.34073424e-02
-6.64893448e-01 3.13784927e-01 -1.97517082e-01 -2.27838635e-01
-3.60503674e-01 8.09664965e-01 -2.41919860e-01 2.32483923e-01
4.89613116e-01 1.97066680e-01 7.62678981e-01 -7.70028472e-01
1.54346311e-02 6.56716466e-01 4.04865444e-01 -4.21604365e-01
1.86955035e-01 2.96580285e-01 -1.25709940e-02 -7.63478816e-01
-1.77316919e-01 -6.73390508e-01 -8.05313408e-01 -7.76630789e-02
6.73188150e-01 -5.96854687e-01 -9.13565099e-01 6.14270568e-01
-3.52405876e-01 -4.51597899e-01 1.26660585e-01 3.88962269e-01
-3.36060107e-01 7.42410421e-01 -6.45311832e-01 -3.48738134e-01
-4.40097421e-01 -1.02527869e+00 4.79657143e-01 1.19096085e-01
-6.16116405e-01 -8.05352509e-01 4.99657035e-01 2.29024976e-01
2.02577502e-01 5.49603820e-01 6.60672724e-01 -8.79420221e-01
-7.12033957e-02 -2.73514956e-01 4.34777051e-01 -1.48899807e-02
1.84213072e-01 4.36650723e-01 -4.94917452e-01 -5.07375956e-01
-3.64439994e-01 -7.51947165e-02 3.92448574e-01 4.10464674e-01
6.24972343e-01 6.37953952e-02 -5.80095828e-01 3.22129995e-01
1.20111382e+00 5.61549902e-01 5.60721874e-01 5.24883151e-01
2.70991087e-01 5.91461778e-01 9.14298534e-01 6.84505224e-01
-6.60902038e-02 6.54757023e-01 2.63110995e-01 -4.61261757e-02
4.07454342e-01 3.64707746e-02 3.00288528e-01 4.17667389e-01
2.20713466e-02 1.89475507e-01 -9.80793953e-01 1.20500892e-01
-1.71208954e+00 -1.21380889e+00 -6.55055344e-01 2.41091895e+00
1.11113143e+00 1.02465730e-02 6.12314999e-01 2.32252274e-02
8.27465832e-01 -7.73931220e-02 -9.14822519e-01 -4.07946199e-01
6.60742000e-02 3.81532311e-01 5.07424533e-01 5.42065322e-01
-7.15527892e-01 6.57882273e-01 5.59016562e+00 2.96508878e-01
-1.37257588e+00 -7.06710592e-02 1.14799879e-01 -3.64701986e-01
2.25285664e-01 2.80486912e-01 -6.54647827e-01 6.80934608e-01
1.33086264e+00 -8.18248332e-01 2.48382166e-01 6.05457664e-01
7.42849350e-01 -2.81354785e-01 -7.85682797e-01 6.76532447e-01
-7.14714885e-01 -1.40485048e+00 -5.91776669e-01 3.36574644e-01
-1.99973643e-01 4.97489572e-01 -3.50148261e-01 -2.65824437e-01
2.31829043e-02 -3.46702188e-01 3.77952814e-01 5.75446546e-01
9.75147665e-01 -8.35447729e-01 3.49970132e-01 1.71297371e-01
-8.84896457e-01 4.77124959e-01 -3.83250952e-01 -1.48593038e-01
3.65561515e-01 8.28738689e-01 -8.91789138e-01 5.05066454e-01
5.24957359e-01 6.74739122e-01 -4.09732580e-01 7.33710170e-01
3.52396071e-01 4.00793582e-01 -2.28878319e-01 2.27702990e-01
-3.45733702e-01 -4.56181109e-01 4.42632616e-01 1.09498763e+00
-6.08125515e-02 5.44495940e-01 2.62968361e-01 5.03775597e-01
3.09816569e-01 1.81878924e-01 -2.28704795e-01 -6.17088735e-01
5.98768830e-01 1.55138838e+00 -9.62580383e-01 -4.80540186e-01
-3.99733156e-01 8.10284972e-01 7.04757497e-02 -6.19031452e-02
-5.99555314e-01 -5.96533835e-01 1.09064794e+00 4.80814189e-01
1.63772732e-01 -3.00389022e-01 5.42149007e-01 -8.74801338e-01
-3.73960167e-01 -1.01540482e+00 6.18007004e-01 -9.39452112e-01
-9.57636535e-01 1.50132120e-01 1.56576872e-01 -6.70011401e-01
2.85440730e-03 -7.46074557e-01 -2.46004671e-01 9.25716281e-01
-9.76576447e-01 -6.33396864e-01 -2.33175489e-03 2.62419909e-01
1.28044561e-01 3.37829381e-01 9.56209719e-01 4.06450361e-01
-7.94228613e-01 8.29615369e-02 4.43003118e-01 -3.36788863e-01
1.13723564e+00 -1.43376124e+00 2.70978659e-01 4.86785382e-01
-6.13148749e-01 1.37473094e+00 1.17013407e+00 -9.64892447e-01
-1.50137663e+00 -7.70394146e-01 9.22165394e-01 -5.07398009e-01
9.72995758e-01 -3.17557961e-01 -1.04765487e+00 8.41337621e-01
1.12888135e-01 -2.32132509e-01 1.32881796e+00 -1.45032540e-01
4.49270681e-02 2.66045988e-01 -1.13338137e+00 1.93578601e-01
5.98372877e-01 -3.37042779e-01 -4.55118030e-01 5.12346804e-01
2.75388390e-01 -5.74037850e-01 -1.64238000e+00 1.22844530e-02
8.45775247e-01 -1.05409741e+00 7.10138083e-01 -8.26699853e-01
-4.20038253e-02 -7.08629072e-01 -8.28783140e-02 -7.71350324e-01
-5.14588237e-01 -9.00113761e-01 4.70298380e-02 1.06864071e+00
3.61990213e-01 -1.10104465e+00 3.18959683e-01 7.17453480e-01
-4.10955608e-01 -6.90082967e-01 -5.22292614e-01 -4.37744379e-01
-2.86075056e-01 -1.14657901e-01 4.51662987e-01 1.14509833e+00
8.93176556e-01 2.37584203e-01 7.77335614e-02 4.75630164e-02
4.49799299e-01 1.14928618e-01 5.69733977e-01 -1.22006321e+00
-3.28524977e-01 -1.35467127e-01 -5.13409197e-01 -3.34703714e-01
-3.18864405e-01 -1.02893400e+00 -4.66745555e-01 -1.31548584e+00
3.29809666e-01 2.03498434e-02 -3.37801307e-01 6.87082946e-01
2.37982888e-02 5.96397333e-02 5.32563142e-02 5.25205374e-01
-4.29143697e-01 -6.77615553e-02 9.25743639e-01 3.41564089e-01
-2.51322508e-01 -2.59469032e-01 -7.23243773e-01 1.88507006e-01
1.19282830e+00 -4.75645065e-01 -2.65410066e-01 4.91314024e-01
3.04242760e-01 -5.85874841e-02 2.43822262e-01 -6.97391748e-01
-7.93383177e-03 -6.15797490e-02 5.06401181e-01 -7.40811050e-01
1.52109608e-01 -5.60671151e-01 1.00735068e+00 9.02954400e-01
-2.20594451e-01 3.96680474e-01 3.93729359e-01 2.45043516e-01
1.98554516e-01 -3.68455984e-02 1.12895072e+00 -4.79305536e-01
-6.31968439e-01 2.08911106e-01 -4.87724304e-01 -3.69991839e-01
1.08622789e+00 -4.24854398e-01 -5.14190495e-01 -1.02666587e-01
-1.26693034e+00 3.95643592e-01 1.13023901e+00 -3.93109694e-02
9.54044387e-02 -8.84200871e-01 -2.26897001e-01 1.07703164e-01
2.66310692e-01 -3.71447593e-01 5.17668188e-01 1.37312782e+00
-9.21964347e-01 5.78994751e-01 -6.23847008e-01 -2.45081797e-01
-1.62694144e+00 8.28187764e-01 3.12489480e-01 2.20039576e-01
-6.13436222e-01 2.32642010e-01 -8.57600719e-02 -3.74555528e-01
-3.52306962e-01 -1.80364624e-02 2.05663815e-02 1.22581333e-01
8.68508458e-01 5.67798913e-01 1.05056413e-01 -6.52295530e-01
-5.97905457e-01 -1.52384818e-01 -3.21753889e-01 1.97838306e-01
1.58650398e+00 -3.01498979e-01 -5.38791955e-01 6.12795174e-01
9.54901695e-01 -2.07473412e-01 -9.91982341e-01 2.89319158e-01
-1.72616839e-02 -2.13443398e-01 -2.85651952e-01 -1.01022089e+00
-5.25282323e-01 5.18451929e-01 7.17796266e-01 1.40981987e-01
9.23036098e-01 1.72193512e-01 4.18276697e-01 5.94945997e-02
1.43327519e-01 -8.40791523e-01 -5.90944529e-01 1.19896278e-01
5.52401185e-01 -8.96562278e-01 3.99539053e-01 -4.90965955e-02
-5.36130846e-01 1.15432978e+00 2.70877063e-01 1.93215847e-01
4.14208561e-01 3.15094858e-01 2.24021360e-01 -8.21406901e-01
-9.34793830e-01 7.78321400e-02 -3.97854745e-01 4.65516746e-01
1.01177001e+00 8.63286108e-02 -9.31516707e-01 3.91218483e-01
-4.54251990e-02 -1.05596695e-03 6.04083002e-01 1.40210485e+00
-4.56473053e-01 -1.54741931e+00 -2.97127336e-01 4.73223060e-01
-7.03229725e-01 -6.52057678e-02 -7.27562964e-01 9.09979403e-01
-3.46851468e-01 5.56537986e-01 -1.26206487e-01 9.78674069e-02
3.29505563e-01 6.69936299e-01 3.61348927e-01 -5.75117648e-01
-2.64294505e-01 3.21245462e-01 2.19145060e-01 -7.40279555e-01
-6.37220442e-01 -1.24956870e+00 -1.68477881e+00 -5.71018219e-01
-3.47721875e-01 5.38835287e-01 9.48995829e-01 3.49302620e-01
9.83718753e-01 2.79216141e-01 4.56104167e-02 -8.14843714e-01
-2.62133360e-01 -9.89170909e-01 -1.03669965e+00 9.44494754e-02
1.06425986e-01 -5.91316760e-01 -2.85760909e-01 2.56938785e-01] | [4.781728267669678, 5.385626316070557] |
bc1adf68-b64b-4fb9-a42b-c577b66c3729 | animc-a-soft-framework-for-auto-weighted | 2011.10331 | null | https://arxiv.org/abs/2011.10331v3 | https://arxiv.org/pdf/2011.10331v3.pdf | ANIMC: A Soft Framework for Auto-weighted Noisy and Incomplete Multi-view Clustering | Multi-view clustering has wide applications in many image processing scenarios. In these scenarios, original image data often contain missing instances and noises, which is ignored by most multi-view clustering methods. However, missing instances may make these methods difficult to use directly and noises will lead to unreliable clustering results. In this paper, we propose a novel Auto-weighted Noisy and Incomplete Multi-view Clustering framework (ANIMC) via a soft auto-weighted strategy and a doubly soft regular regression model. Firstly, by designing adaptive semi-regularized nonnegative matrix factorization (adaptive semi-RNMF), the soft auto-weighted strategy assigns a proper weight to each view and adds a soft boundary to balance the influence of noises and incompleteness. Secondly, by proposing{\theta}-norm, the doubly soft regularized regression model adjusts the sparsity of our model by choosing different{\theta}. Compared with existing methods, ANIMC has three unique advantages: 1) it is a soft algorithm to adjust our framework in different scenarios, thereby improving its generalization ability; 2) it automatically learns a proper weight for each view, thereby reducing the influence of noises; 3) it performs doubly soft regularized regression that aligns the same instances in different views, thereby decreasing the impact of missing instances. Extensive experimental results demonstrate its superior advantages over other state-of-the-art methods. | ['Dapeng Oliver Wu', 'Pan Zhou', 'Yuchong Hu', 'Xiang Fang'] | 2020-11-20 | null | null | null | null | ['incomplete-multi-view-clustering'] | ['computer-vision'] | [ 9.64818057e-03 -5.87616086e-01 6.85917512e-02 -3.87223572e-01
-5.19061923e-01 -2.66143769e-01 9.83311459e-02 -3.28796417e-01
-2.48846307e-01 1.93607897e-01 4.05867100e-01 2.29663193e-01
-3.83724719e-01 -5.26480913e-01 -2.86361635e-01 -1.18553865e+00
4.69255030e-01 2.80960143e-01 1.76351666e-01 -7.38133267e-02
-4.85793874e-03 -3.33230346e-02 -1.53338349e+00 3.16902667e-01
1.05239308e+00 7.79548526e-01 4.25418794e-01 2.13972926e-02
-1.35742068e-01 6.22324884e-01 -3.94761950e-01 -2.72969335e-01
3.42185974e-01 -3.25004488e-01 -3.51224482e-01 7.84422517e-01
7.69766122e-02 1.68440551e-01 -1.98795125e-02 1.21828723e+00
5.77896416e-01 2.26481825e-01 3.70630801e-01 -1.29011095e+00
-7.99901128e-01 5.51809728e-01 -1.26022661e+00 4.34742644e-02
3.36263552e-02 -1.51317297e-02 7.43056297e-01 -1.08374858e+00
3.42131525e-01 1.21275795e+00 5.04990160e-01 3.77065390e-01
-1.31495595e+00 -7.35608697e-01 6.55209661e-01 2.10157588e-01
-1.45855987e+00 -2.31106833e-01 1.08109879e+00 -4.27520931e-01
1.17666841e-01 1.93945482e-01 5.07309258e-01 9.03923929e-01
-2.44441062e-01 8.56101930e-01 1.32979107e+00 -1.50286406e-01
1.73538834e-01 1.13808177e-01 -9.44775194e-02 3.81045550e-01
1.92175567e-01 -3.80015820e-01 -2.69120306e-01 -1.86118677e-01
4.84217793e-01 5.30960083e-01 -5.09794712e-01 -6.73094571e-01
-1.41686618e+00 6.20758295e-01 1.59302205e-01 3.71974200e-01
-4.14655089e-01 -3.06924552e-01 5.36219418e-01 4.35821861e-02
4.04578388e-01 -2.05523893e-01 -4.20257688e-01 2.36285761e-01
-7.32973278e-01 -2.13942707e-01 1.06463738e-01 8.05901825e-01
7.71716535e-01 2.10366622e-01 1.38939321e-01 1.41349709e+00
3.41736823e-01 5.87214470e-01 7.53219604e-01 -9.17964458e-01
6.86154425e-01 9.53052282e-01 -1.90916017e-01 -1.32664418e+00
-3.72476906e-01 -5.51970005e-01 -1.52183270e+00 -8.65531415e-02
5.58390729e-02 -4.57442440e-02 -6.08125985e-01 1.59628570e+00
5.13901830e-01 2.73960531e-01 -6.96299747e-02 9.73750234e-01
8.90990913e-01 5.90673268e-01 -2.72055060e-01 -7.91959822e-01
1.13471222e+00 -8.33368421e-01 -8.47589016e-01 -1.56662479e-01
3.15021239e-02 -9.09705102e-01 9.97249663e-01 6.97145224e-01
-8.59228075e-01 -6.58471286e-01 -9.29014802e-01 4.17207837e-01
-5.60476631e-02 2.12417409e-01 3.53109747e-01 3.98597807e-01
-7.29048073e-01 9.24360901e-02 -6.66722596e-01 1.75912324e-02
1.31269485e-01 2.19593704e-01 -5.25026619e-01 -3.16670418e-01
-8.06070924e-01 2.51179874e-01 3.79188448e-01 3.22105497e-01
-2.29182348e-01 -4.26162839e-01 -8.77386093e-01 -1.69038624e-01
8.51744711e-01 -5.69032967e-01 5.91971159e-01 -1.06411147e+00
-1.22141576e+00 5.47267020e-01 -5.10057688e-01 1.56503469e-01
3.24605614e-01 -4.79871891e-02 -8.04340422e-01 1.46552324e-01
4.41314816e-01 2.18137443e-01 1.20757198e+00 -1.83637440e+00
-5.18018305e-01 -7.32369840e-01 -3.90686572e-01 3.36467415e-01
-5.05392790e-01 -2.04522535e-01 -9.69295144e-01 -1.01482809e+00
7.89526463e-01 -8.83628190e-01 -5.88619947e-01 -3.47163767e-01
-1.94654137e-01 2.04417715e-03 1.02658272e+00 -3.20027292e-01
1.45718133e+00 -2.31779218e+00 4.37881529e-01 4.45243925e-01
3.84737462e-01 2.53037393e-01 -4.45500575e-02 1.70743152e-01
-3.62740457e-01 -3.14345360e-02 -4.14986938e-01 -2.88500071e-01
-4.24744666e-01 3.99001747e-01 3.88441384e-02 5.04446983e-01
-1.81857422e-01 2.66474932e-01 -8.45271587e-01 -5.72512209e-01
4.22308981e-01 5.73128164e-01 -4.41285908e-01 5.35437539e-02
2.52799422e-01 4.93688822e-01 -3.61672550e-01 5.23980975e-01
1.01042974e+00 -3.74577790e-01 2.53352255e-01 -5.06008685e-01
-7.82791823e-02 -5.65468371e-01 -2.03430581e+00 1.38585734e+00
-3.80021840e-01 -1.40107796e-01 3.83438766e-01 -1.13796544e+00
9.41568315e-01 2.90683568e-01 9.19118941e-01 -4.43298936e-01
1.47907898e-01 -3.84584256e-02 -2.82183826e-01 -5.41453481e-01
2.73905993e-01 -1.80045784e-01 6.43179193e-02 4.15971756e-01
-1.71584338e-01 1.79814786e-01 2.77926892e-01 3.46370697e-01
5.96557200e-01 -1.03842176e-01 2.41738409e-01 -1.21623896e-01
1.08809757e+00 -5.06248593e-01 1.26294434e+00 2.21957326e-01
-1.47930458e-01 9.82581615e-01 1.43945336e-01 -4.75129396e-01
-6.37961864e-01 -8.82136226e-01 -2.96607111e-02 8.44906688e-01
4.68278021e-01 -5.14914274e-01 -5.81508696e-01 -7.46515751e-01
-2.03341112e-01 1.70582876e-01 -5.65885246e-01 -1.13373302e-01
-4.07043844e-01 -1.19157612e+00 -7.08850548e-02 5.25898159e-01
6.44424796e-01 -6.16529167e-01 9.27059799e-02 1.16016857e-01
-6.83950901e-01 -1.08126068e+00 -8.34957063e-01 8.14576820e-02
-8.37160826e-01 -1.20902312e+00 -6.44380331e-01 -7.12064445e-01
1.02773440e+00 1.18093264e+00 8.09714735e-01 1.47562325e-01
1.05925784e-01 4.60560948e-01 -6.50151968e-01 4.86042863e-03
-6.97180331e-02 -3.37657511e-01 3.74054849e-01 6.13145709e-01
3.84285659e-01 -6.51377320e-01 -6.44255817e-01 8.56062472e-01
-1.23506045e+00 -1.30483806e-01 4.18749988e-01 9.14044917e-01
1.06014645e+00 5.07985115e-01 4.70969409e-01 -1.08149052e+00
4.57049370e-01 -4.77865070e-01 -4.40342158e-01 2.31983334e-01
-8.30824077e-01 -2.69482106e-01 9.60364044e-01 -4.83251840e-01
-1.19103932e+00 2.41425410e-01 1.38297901e-01 -7.51116693e-01
3.56881954e-02 3.52054596e-01 -5.48659980e-01 7.81309232e-02
2.34284624e-01 4.01313186e-01 1.86498180e-01 -4.14831221e-01
4.16919917e-01 6.08861744e-01 3.42116743e-01 -3.65041167e-01
9.89710093e-01 7.30021119e-01 -2.73230046e-01 -6.52186871e-01
-8.76957774e-01 -8.82479012e-01 -7.84609675e-01 -4.05832738e-01
7.98823297e-01 -1.23405218e+00 -7.22486734e-01 5.80841422e-01
-4.81975406e-01 3.31462711e-01 -6.56263083e-02 4.96044725e-01
-2.50209272e-01 9.73856807e-01 -3.21210653e-01 -6.79586411e-01
-1.77271411e-01 -1.23974037e+00 8.36938739e-01 2.92279005e-01
1.36557892e-01 -7.70325601e-01 -2.14254901e-01 8.94634306e-01
6.23771846e-02 1.11617535e-01 6.22021556e-01 -2.56092429e-01
-1.05530731e-01 1.33809447e-02 -2.53244072e-01 6.80779755e-01
3.99628073e-01 1.48245469e-01 -6.83214068e-01 -3.97470027e-01
3.25710207e-01 -4.00256850e-02 8.80978763e-01 4.24812734e-01
1.48931456e+00 -1.97963253e-01 -1.79205686e-01 6.55996799e-01
1.47434282e+00 1.57119557e-01 4.19974834e-01 2.66147703e-01
9.37758267e-01 4.29208338e-01 6.17443085e-01 6.63195312e-01
6.27763808e-01 5.45097351e-01 5.26068270e-01 -2.46344343e-01
2.33978763e-01 -4.19696560e-03 2.51887351e-01 1.48341322e+00
-1.90952748e-01 6.27450347e-02 -4.81796533e-01 6.00121439e-01
-2.18825102e+00 -1.10607243e+00 -4.06569779e-01 2.25358582e+00
6.13345385e-01 -5.75643964e-02 2.38943517e-01 4.78975594e-01
1.04535484e+00 3.77247065e-01 -4.87190813e-01 1.64036140e-01
-4.02325988e-01 -2.79070169e-01 2.28036672e-01 9.08095464e-02
-1.14514732e+00 4.53854382e-01 4.78136206e+00 9.71180141e-01
-8.47926497e-01 1.54444203e-01 4.91718829e-01 -1.12051584e-01
-2.99353540e-01 -9.64204520e-02 -5.29116094e-01 8.06477904e-01
6.85719177e-02 2.15696111e-01 5.44475317e-01 7.48759210e-01
4.56309378e-01 -9.81531814e-02 -4.99240637e-01 1.27727175e+00
4.71384525e-01 -8.30179513e-01 1.14003971e-01 2.62698475e-02
9.61520553e-01 -2.30520159e-01 1.47062503e-02 2.43560430e-02
3.36490273e-01 -4.69770521e-01 4.67406124e-01 4.12508845e-01
4.84931052e-01 -9.80723679e-01 7.66781092e-01 4.80515629e-01
-1.36103666e+00 -2.89274484e-01 -4.12807554e-01 2.90768534e-01
1.55291677e-01 1.11885297e+00 -7.64038488e-02 8.20207298e-01
9.66162086e-01 9.82571900e-01 -5.32155395e-01 7.63123989e-01
-1.17878109e-01 4.05202270e-01 -1.71944082e-01 4.89822328e-01
1.32020921e-01 -7.38749087e-01 5.94781876e-01 8.88123810e-01
1.36204556e-01 3.36298525e-01 6.29430532e-01 3.36688429e-01
1.50920928e-01 2.11976334e-01 -3.93414766e-01 6.19297266e-01
4.18582618e-01 1.48677301e+00 -8.19388926e-01 -9.65202525e-02
-6.19729817e-01 9.51570809e-01 1.34884313e-01 3.82559925e-01
-7.07232773e-01 -1.10767558e-01 4.03374523e-01 -3.52255292e-02
5.25993288e-01 -1.01703875e-01 -3.07845324e-01 -1.45748079e+00
3.21872890e-01 -1.05176270e+00 6.97457790e-01 -6.47447884e-01
-1.57726240e+00 5.22378981e-01 -2.45507747e-01 -1.73919964e+00
2.17194930e-01 -1.90122426e-01 -4.93319392e-01 4.02965844e-01
-1.19200313e+00 -9.63140249e-01 -4.28695291e-01 9.51957703e-01
7.27802038e-01 -3.21679413e-01 5.24414837e-01 4.99875844e-01
-9.80675876e-01 3.35514009e-01 2.49220476e-01 6.77077249e-02
9.48764384e-01 -1.09154284e+00 -3.81232202e-01 9.06714022e-01
3.97274978e-02 7.64823377e-01 3.75317633e-01 -5.41433036e-01
-1.33292007e+00 -1.17484879e+00 2.06347585e-01 -1.42902523e-01
4.47390348e-01 6.60875207e-03 -1.03871882e+00 4.28146988e-01
8.78945168e-04 2.67786235e-01 1.07468283e+00 1.28285781e-01
-4.08293128e-01 -6.22804701e-01 -1.10233510e+00 4.43349987e-01
7.49186933e-01 -1.76038340e-01 -2.87064314e-01 3.40971410e-01
5.57742536e-01 -1.32824928e-01 -1.00641572e+00 5.85660160e-01
2.90950418e-01 -1.29642832e+00 1.05943215e+00 -1.44346878e-01
3.20446521e-01 -1.00390363e+00 -2.64750570e-01 -1.45243979e+00
-7.93091953e-01 -2.08410874e-01 -8.99185166e-02 1.65905058e+00
1.54664442e-01 -5.52205920e-01 5.49175322e-01 3.64278316e-01
1.25022888e-01 -8.12001944e-01 -8.32371652e-01 -6.30445361e-01
-4.16002750e-01 -4.66948360e-01 5.85397482e-01 1.22582257e+00
-2.48458311e-01 3.08876038e-01 -6.09836280e-01 3.22539181e-01
8.05383921e-01 2.09788442e-01 7.70182490e-01 -1.18763483e+00
-2.45038584e-01 -2.58615166e-01 -2.35516429e-01 -7.40746140e-01
-8.16929936e-02 -5.89144111e-01 -1.02738596e-01 -1.36559868e+00
4.64247286e-01 -3.99745941e-01 -3.39774549e-01 3.65141779e-01
-5.54379761e-01 3.36571604e-01 2.30824903e-01 4.39000875e-01
-8.04345191e-01 6.56617522e-01 1.15686393e+00 -8.18373337e-02
-1.79612711e-01 3.39922935e-01 -1.02863693e+00 1.10745907e+00
5.04876316e-01 -3.56232852e-01 -4.97084737e-01 -4.40683365e-01
2.87312269e-01 -3.53499688e-03 3.70731987e-02 -7.93179333e-01
2.88456231e-01 -2.83142090e-01 6.03649378e-01 -7.72793651e-01
6.25263304e-02 -1.33496535e+00 1.81954876e-01 6.41080439e-02
2.19500214e-01 1.19263686e-01 -2.23376989e-01 1.01294208e+00
-4.59959418e-01 -6.87024593e-02 8.87476265e-01 -4.48418111e-01
-6.15649164e-01 2.72065520e-01 -3.01649928e-01 1.67568326e-01
1.07293904e+00 -4.75819051e-01 1.68308884e-01 -4.02648389e-01
-7.57031143e-01 7.93025792e-01 5.42881310e-01 5.60931504e-01
7.10494936e-01 -1.48547256e+00 -5.35081744e-01 2.94506639e-01
1.85489908e-01 3.44274133e-01 7.25113332e-01 1.00394213e+00
8.63560140e-02 -3.03559005e-01 2.43642598e-01 -8.17322612e-01
-1.43689716e+00 8.23091328e-01 -1.06281638e-01 -1.81859612e-01
-4.99749243e-01 5.26588857e-01 1.76018462e-01 -6.64673924e-01
1.57444820e-01 2.29638591e-01 -4.55183119e-01 4.06888694e-01
6.14781320e-01 5.93579292e-01 7.76393758e-03 -9.60148036e-01
-3.24253887e-01 9.78046536e-01 -8.42499584e-02 1.90386057e-01
1.57348490e+00 -5.51120043e-01 -4.52144027e-01 5.97294807e-01
1.01455438e+00 1.91594422e-01 -1.07422101e+00 -5.77101588e-01
-4.35771167e-01 -6.12830758e-01 1.66486427e-02 -4.82119352e-01
-1.70779979e+00 5.57489276e-01 3.30555052e-01 2.31123701e-01
1.56625926e+00 -2.69034415e-01 8.49922895e-01 2.30434444e-02
2.61316240e-01 -1.35403001e+00 1.60139814e-01 2.62675911e-01
5.48806310e-01 -1.32727635e+00 2.73181021e-01 -6.87742949e-01
-1.07697427e+00 9.38105702e-01 8.18592072e-01 2.17879172e-02
7.65805781e-01 2.68314183e-01 2.39673331e-01 -2.01220050e-01
-4.94760722e-01 -1.36446685e-01 2.51074851e-01 5.32431066e-01
1.67270973e-01 -1.72864031e-02 -2.12067142e-01 8.32814455e-01
2.61791587e-01 -1.16977103e-01 4.89913434e-01 5.80567360e-01
-2.68199205e-01 -1.03620887e+00 -7.89702356e-01 4.66147959e-01
-4.07632828e-01 1.74096242e-01 -1.86995238e-01 4.12827909e-01
4.58518296e-01 1.23729944e+00 -2.39598960e-01 -5.88299990e-01
4.16522294e-01 -2.16702595e-01 -2.44808174e-03 -4.36546892e-01
-5.26252329e-01 5.46036243e-01 -5.15552998e-01 -4.42906886e-01
-9.12158012e-01 -9.34192836e-01 -1.12767148e+00 -9.53898802e-02
-6.33346140e-01 1.85786128e-01 2.25393817e-01 8.85775089e-01
4.68170613e-01 5.20532370e-01 1.19067729e+00 -6.47979975e-01
-3.07774693e-01 -6.21237278e-01 -6.62795305e-01 6.37191355e-01
5.68620451e-02 -6.63381219e-01 -4.89969790e-01 2.14980468e-01] | [8.328559875488281, 4.613122463226318] |
e8f7ebb7-62a1-4ddb-9996-6a9be75f13b0 | data-driven-mitigation-of-adversarial-text | 2202.09483 | null | https://arxiv.org/abs/2202.09483v1 | https://arxiv.org/pdf/2202.09483v1.pdf | Data-Driven Mitigation of Adversarial Text Perturbation | Social networks have become an indispensable part of our lives, with billions of people producing ever-increasing amounts of text. At such scales, content policies and their enforcement become paramount. To automate moderation, questionable content is detected by Natural Language Processing (NLP) classifiers. However, high-performance classifiers are hampered by misspellings and adversarial text perturbations. In this paper, we classify intentional and unintentional adversarial text perturbation into ten types and propose a deobfuscation pipeline to make NLP models robust to such perturbations. We propose Continuous Word2Vec (CW2V), our data-driven method to learn word embeddings that ensures that perturbations of words have embeddings similar to those of the original words. We show that CW2V embeddings are generally more robust to text perturbations than embeddings based on character ngrams. Our robust classification pipeline combines deobfuscation and classification, using proposed defense methods and word embeddings to classify whether Facebook posts are requesting engagement such as likes. Our pipeline results in engagement bait classification that goes from 0.70 to 0.67 AUC with adversarial text perturbation, while character ngram-based word embedding methods result in downstream classification that goes from 0.76 to 0.64. | ['Igor Markov', 'Anand Bhaskar', 'Mohammad Al-Rubaie', 'Rasika Bhalerao'] | 2022-02-19 | null | null | null | null | ['adversarial-text'] | ['adversarial'] | [ 2.12680608e-01 2.31373250e-01 -4.19997811e-01 2.11973675e-02
-5.60156763e-01 -1.08642650e+00 1.16939628e+00 7.55165100e-01
-5.44034481e-01 5.92979491e-01 8.03019762e-01 -5.91990829e-01
2.98509032e-01 -1.01149869e+00 -4.60328996e-01 -1.94522396e-01
1.51451007e-01 1.34478882e-01 1.49466112e-01 -4.85500515e-01
4.42516476e-01 5.87616563e-01 -1.02489352e+00 3.51333767e-01
6.03548706e-01 7.58093834e-01 -9.14459348e-01 9.42714095e-01
-2.86732435e-01 4.60282713e-01 -8.73163521e-01 -7.37156153e-01
3.62084687e-01 1.42290862e-02 -5.85811198e-01 -4.90773708e-01
5.48478067e-01 -5.62992156e-01 -1.05973339e+00 1.17957044e+00
5.93613505e-01 7.11016506e-02 7.50825882e-01 -1.23086619e+00
-1.14537799e+00 6.66415930e-01 -3.27169329e-01 3.78093362e-01
6.19326115e-01 5.36411762e-01 9.38827753e-01 -4.35376406e-01
4.95373219e-01 1.41006351e+00 6.31823540e-01 6.20754302e-01
-1.31821382e+00 -8.13693285e-01 -2.32437447e-01 1.65223535e-02
-1.05667019e+00 -3.97376478e-01 4.94460642e-01 -8.02258611e-01
8.56901407e-01 4.87154305e-01 3.09348464e-01 1.86483431e+00
4.33108330e-01 3.79965037e-01 7.54345894e-01 -1.31914809e-01
6.67131692e-02 2.81767756e-01 2.43173152e-01 5.53553581e-01
5.98601937e-01 -2.23052740e-01 -3.70903850e-01 -7.34715939e-01
2.43458953e-02 2.71986604e-01 -8.05211067e-02 5.35754859e-01
-9.09881830e-01 1.23047352e+00 2.47113764e-01 3.62338126e-01
-2.58358687e-01 3.03615510e-01 8.65241885e-01 2.37551689e-01
6.15754604e-01 1.05168152e+00 -1.07076518e-01 -4.82029021e-01
-6.46925926e-01 2.59545386e-01 7.32694626e-01 5.47889471e-01
3.27659428e-01 -6.68482780e-02 -7.46408939e-01 7.31916428e-01
-7.67163783e-02 6.38997018e-01 7.06776738e-01 -4.76816088e-01
5.28790951e-01 6.65283322e-01 7.39260316e-02 -1.73081088e+00
-1.80655628e-01 -1.40356764e-01 -6.81088924e-01 -1.91219017e-01
3.47770721e-01 -3.28207374e-01 -7.15629935e-01 1.39777398e+00
2.13723212e-01 -6.39794171e-02 -2.21636385e-01 3.25098157e-01
5.47841907e-01 6.41163766e-01 1.64525062e-01 1.45102859e-01
1.42414463e+00 -5.04573107e-01 -8.21959794e-01 -2.87506878e-01
7.59665251e-01 -5.49798906e-01 1.18056333e+00 2.02639818e-01
-7.21980989e-01 -1.15782291e-01 -9.57855821e-01 -1.79714233e-01
-8.61105084e-01 -9.16773677e-01 2.55973160e-01 1.24492693e+00
-6.78357482e-01 8.00941825e-01 -3.37152451e-01 -4.32531297e-01
7.70853758e-01 9.41524878e-02 -5.41580141e-01 5.58956563e-02
-1.56207800e+00 9.00725961e-01 2.75848862e-02 -4.56079185e-01
-4.88721311e-01 -8.79799008e-01 -7.78585911e-01 -1.03616565e-01
-1.44531846e-01 1.10164240e-01 9.15732145e-01 -5.67330718e-01
-9.38707471e-01 7.76883960e-01 -5.05943522e-02 -5.81969082e-01
4.99903858e-01 -2.30735660e-01 -8.06121886e-01 2.54074693e-01
6.02406301e-02 1.40023619e-01 1.21254361e+00 -6.70062184e-01
-1.20901182e-01 -3.50766301e-01 -5.50017580e-02 -4.54769045e-01
-1.52480972e+00 2.18051299e-01 1.59906685e-01 -7.18567371e-01
-5.58331966e-01 -7.20911503e-01 -4.87451851e-02 4.38860208e-02
-6.37981355e-01 -1.00805901e-01 1.05289328e+00 -1.08895361e+00
1.78417802e+00 -2.06926012e+00 -2.36999184e-01 2.73603201e-01
5.64145565e-01 7.46617913e-01 -2.75287807e-01 7.95755506e-01
9.53960195e-02 9.59804654e-01 3.89314666e-02 -2.09860727e-01
2.39238888e-01 -6.50296360e-02 -6.86452627e-01 6.46759748e-01
1.69893175e-01 1.14853048e+00 -1.07603145e+00 1.07282456e-02
-8.73671100e-02 2.40996063e-01 -6.45036936e-01 9.87060890e-02
-2.48552978e-01 6.95577636e-03 -3.56702954e-01 6.98894978e-01
4.36487913e-01 2.29316980e-01 -9.26476866e-02 1.27542406e-01
1.03891045e-01 3.27504367e-01 -3.80166769e-01 6.66637421e-01
-3.92274857e-01 1.10349274e+00 -9.22729671e-02 -5.49370944e-01
8.91876578e-01 -1.04233623e-02 3.18964630e-01 -2.81609833e-01
4.53054845e-01 -1.78336233e-01 4.91852835e-02 -6.81531131e-01
6.04578614e-01 1.99332722e-02 -4.15321171e-01 5.59162915e-01
-2.22031325e-01 5.55166081e-02 -1.89463735e-01 5.98745465e-01
1.84881341e+00 -8.48584116e-01 1.77718773e-01 7.37838596e-02
2.46097237e-01 -2.07835734e-01 2.06620395e-02 7.40543723e-01
-5.48637509e-01 3.55668724e-01 9.51919138e-01 -2.25320518e-01
-1.39203143e+00 -7.83189178e-01 -1.46710157e-01 1.03712535e+00
-3.27965975e-01 -6.62712574e-01 -7.91038394e-01 -1.01237226e+00
6.95999503e-01 1.02575541e+00 -8.43584657e-01 -6.84761643e-01
-1.94168538e-01 -4.04638231e-01 1.30125558e+00 2.04967484e-01
2.26910353e-01 -8.44069839e-01 2.93046951e-01 2.28887960e-01
-3.63342389e-02 -1.16826844e+00 -9.72794533e-01 -4.35354233e-01
-2.77051955e-01 -1.05381608e+00 -3.43175262e-01 -1.77792475e-01
4.50197697e-01 1.01777293e-01 4.70429778e-01 -8.33457336e-04
-4.69667286e-01 4.52801883e-01 -5.96037447e-01 -3.94820392e-01
-6.59107387e-01 1.12032279e-01 5.53753614e-01 2.94932842e-01
8.20733666e-01 -6.59488082e-01 -4.68088329e-01 3.25787328e-02
-1.34811044e+00 -8.69222641e-01 1.01564024e-02 4.54301238e-01
-1.84390739e-01 -1.16269790e-01 5.79376519e-01 -1.06773424e+00
1.19112062e+00 -9.90969300e-01 6.07098080e-02 -3.05298328e-01
-6.24489784e-01 -2.25852370e-01 8.95058513e-01 -1.08120489e+00
-1.75472483e-01 -5.63072681e-01 -3.83482158e-01 -3.98344338e-01
-7.02431574e-02 1.70932963e-01 -1.59116723e-02 -9.12001580e-02
1.23733377e+00 1.12672165e-01 1.97404146e-01 -2.64526844e-01
5.54557264e-01 1.35603809e+00 1.06740423e-01 -1.45426780e-01
1.39711106e+00 4.56600428e-01 -3.00454199e-01 -1.34296632e+00
-8.31976652e-01 -4.40495372e-01 -2.86803126e-01 -1.70434371e-01
9.34115827e-01 -4.52348500e-01 -7.87543952e-01 3.40338469e-01
-1.14106417e+00 -2.51227051e-01 -1.12118885e-01 -1.18014216e-02
1.63098127e-01 8.12992573e-01 -6.30156875e-01 -6.95653796e-01
-5.22219777e-01 -6.07578635e-01 5.75105667e-01 -1.99408278e-01
-8.33438873e-01 -9.39872742e-01 1.87991351e-01 5.61656833e-01
6.21733427e-01 6.57325566e-01 9.14678812e-01 -1.31702495e+00
2.36145273e-01 -1.08009100e+00 -1.79926619e-01 5.91660142e-01
2.48004094e-01 1.27767250e-01 -8.97850215e-01 -3.10298413e-01
-2.84252942e-01 -3.39522064e-01 6.16659522e-01 -3.02808642e-01
1.42712176e+00 -1.09386039e+00 -1.93009600e-01 3.69441926e-01
9.84914124e-01 -2.91464716e-01 9.97499585e-01 3.61385405e-01
9.56951976e-01 6.12884700e-01 1.57125577e-01 7.03529775e-01
-1.01832166e-01 3.80342215e-01 6.21152639e-01 5.81187069e-01
2.13288531e-01 -8.52114081e-01 7.13285327e-01 2.89715707e-01
5.75984299e-01 -4.60089415e-01 -1.01437783e+00 6.09529018e-01
-1.30400109e+00 -1.23829722e+00 -4.40024257e-01 1.95506990e+00
9.60292220e-01 4.90631193e-01 2.91127384e-01 2.30492562e-01
7.73233414e-01 6.65745258e-01 -5.60341775e-01 -1.00234795e+00
1.58569828e-01 2.80431777e-01 9.12469804e-01 6.41569614e-01
-8.89436722e-01 1.09632969e+00 5.48843193e+00 8.33656251e-01
-9.44752276e-01 3.96913290e-01 3.42735082e-01 -2.64773041e-01
-5.76105654e-01 -5.21900892e-01 -9.12996709e-01 9.60944057e-01
1.31886482e+00 -2.79717028e-01 5.45821786e-01 7.15785980e-01
3.98050934e-01 5.27628064e-01 -8.41241002e-01 9.20231223e-01
3.90826255e-01 -1.41644776e+00 2.39062190e-01 2.90639758e-01
7.37706065e-01 -7.58109940e-03 4.12041247e-01 4.48085219e-01
4.04605836e-01 -1.31746507e+00 2.52537459e-01 2.25464284e-01
1.08868539e+00 -6.63063765e-01 4.28028345e-01 2.47162536e-01
-5.68106413e-01 -4.05950725e-01 -3.89897972e-01 -8.85989890e-02
5.26622385e-02 8.77789199e-01 -9.09979582e-01 -2.52016097e-01
2.23737419e-01 6.47532046e-01 -5.84129155e-01 3.63532364e-01
-4.53870922e-01 9.35337722e-01 -1.53713554e-01 -5.00471950e-01
2.70322382e-01 7.05742463e-02 7.59700477e-01 1.38257432e+00
8.53744000e-02 1.20691791e-01 -2.67717928e-01 5.74631929e-01
-6.67565107e-01 6.80930242e-02 -1.12886035e+00 -9.45027709e-01
7.15050101e-01 1.24171317e+00 -2.47390289e-02 -2.66510874e-01
-1.41578615e-01 1.19461274e+00 4.55916792e-01 2.31614217e-01
-8.63248765e-01 -7.94577360e-01 1.33863044e+00 6.43050194e-01
-1.10010475e-01 -3.46320122e-01 -3.02986681e-01 -1.17848778e+00
-1.55565649e-01 -9.92019534e-01 1.22517183e-01 -3.16144198e-01
-1.67634654e+00 2.72414684e-01 -3.71659607e-01 -7.84018517e-01
6.24789856e-02 -6.58667088e-01 -8.20663691e-01 8.30127597e-01
-1.00114536e+00 -8.06314588e-01 -1.25979066e-01 3.14730674e-01
3.78516138e-01 -1.93851635e-01 8.12564254e-01 2.31901497e-01
-6.64875090e-01 1.19693470e+00 2.62088180e-01 5.98222017e-01
8.70460987e-01 -1.05055559e+00 7.27337480e-01 6.76942766e-01
-3.17361027e-01 8.10078025e-01 6.89692855e-01 -1.03375459e+00
-1.36102784e+00 -1.68421841e+00 1.11098194e+00 -9.06164885e-01
1.46303916e+00 -7.74692774e-01 -9.28440928e-01 7.42528439e-01
-3.69466618e-02 -1.38179332e-01 9.94620085e-01 -7.50555936e-03
-9.67909992e-01 2.47298121e-01 -1.47978044e+00 1.06264424e+00
1.05402613e+00 -9.71417904e-01 -5.07046521e-01 9.39287066e-01
1.00267088e+00 1.92628115e-01 -9.22365487e-01 -1.70986891e-01
5.39184451e-01 -3.63649964e-01 8.34338963e-01 -1.29103565e+00
8.87304187e-01 2.98114657e-01 -1.74838394e-01 -1.13115573e+00
-4.72644687e-01 -1.02369094e+00 -5.94550729e-01 1.36416411e+00
4.80849266e-01 -1.09263074e+00 6.48423135e-01 7.65379846e-01
2.43019328e-01 -5.15025258e-01 -7.13620722e-01 -8.90384436e-01
4.29958612e-01 -5.11309922e-01 3.69914025e-01 1.42887151e+00
4.11024034e-01 2.05701232e-01 -3.86854678e-01 2.31925696e-01
4.92920041e-01 -9.28914011e-01 8.88142884e-01 -9.46752071e-01
-1.48514479e-01 -5.53128481e-01 -7.83216655e-01 -6.52275980e-01
4.04847682e-01 -1.07811141e+00 -5.04160583e-01 -1.29431295e+00
2.40637437e-02 -1.15058929e-01 7.48000741e-02 5.32033980e-01
-2.42173970e-01 6.49726808e-01 -1.65264495e-02 2.28816308e-02
-1.10199079e-01 4.42722321e-01 5.96625149e-01 -4.96859699e-01
-8.80569369e-02 -1.51831791e-01 -1.06380749e+00 5.46361744e-01
1.17308331e+00 -5.77808976e-01 1.00726768e-01 -1.42530233e-01
5.81933618e-01 -6.74230874e-01 2.21905395e-01 -5.69882333e-01
-2.53701091e-01 -4.53656837e-02 1.27559423e-01 -9.94555801e-02
1.77280232e-01 -5.67824364e-01 -6.41711771e-01 5.99875569e-01
-6.50141120e-01 -8.41427222e-02 2.32950151e-01 8.97208393e-01
3.38271320e-01 -2.94680655e-01 9.11509752e-01 3.39522243e-01
1.35509819e-01 5.18288612e-01 -8.65810990e-01 3.35326612e-01
1.07272744e+00 -8.38818848e-02 -6.62774026e-01 -7.39597917e-01
-6.27186775e-01 7.16257989e-02 4.82548505e-01 6.45985305e-01
5.16173959e-01 -1.27116072e+00 -7.58549333e-01 4.05317321e-02
1.45025462e-01 -8.04810643e-01 -7.80970510e-03 4.66189384e-01
-4.44269031e-01 1.52478233e-01 9.46784765e-02 9.91954207e-02
-1.37644506e+00 7.38030255e-01 2.82392576e-02 1.35772973e-02
-3.00903916e-01 7.76272476e-01 -5.66104233e-01 -4.20929134e-01
1.94099367e-01 -1.59708634e-02 -1.34444207e-01 2.65845001e-01
1.04212832e+00 6.47848904e-01 -1.80334166e-01 -5.50398707e-01
-1.21286705e-01 1.79145485e-04 -2.30778232e-01 -9.46047716e-03
1.08712721e+00 2.99494535e-01 -2.55343556e-01 2.69248128e-01
1.88031912e+00 5.81500292e-01 -5.30168712e-01 -7.36963004e-02
-6.75157979e-02 -7.57419467e-01 1.48559269e-03 -5.45495749e-01
-4.59036052e-01 8.30506206e-01 2.40203053e-01 7.85057187e-01
3.85039747e-01 -1.47974834e-01 1.29303837e+00 2.76379824e-01
-2.13673294e-01 -7.97314763e-01 4.02894944e-01 6.49247706e-01
9.07288194e-01 -1.02285159e+00 -4.17956244e-03 -1.82207927e-01
-4.68139738e-01 8.78886878e-01 4.45071250e-01 -1.38820201e-01
6.19391084e-01 -5.33897206e-02 -3.39857817e-01 -5.04091941e-02
-3.46324623e-01 2.17137069e-01 1.14318237e-01 7.92198181e-01
-8.63906294e-02 4.56613898e-02 -4.17957008e-01 5.26361823e-01
-3.07465494e-01 -7.33144879e-01 7.42876172e-01 6.42111182e-01
-6.51753426e-01 -1.12069440e+00 -4.44161594e-01 9.95170355e-01
-9.38155890e-01 -3.48843008e-01 -9.05722380e-01 3.29092145e-01
-1.90474898e-01 1.21980572e+00 3.24956030e-02 -8.90881479e-01
3.71507585e-01 2.45780215e-01 -1.26129553e-01 -9.19526935e-01
-8.74930978e-01 -8.22844565e-01 4.15458530e-01 -5.19895196e-01
5.40389180e-01 -6.02389991e-01 -8.53542209e-01 -1.03811646e+00
-4.13613096e-02 2.42378823e-02 7.06567049e-01 6.97237253e-01
6.23885036e-01 1.92410722e-01 9.36073422e-01 -6.13140821e-01
-9.58493352e-01 -1.20561659e+00 -4.44822013e-01 8.03187191e-01
4.67368841e-01 -2.44505867e-01 -8.90574872e-01 -3.87605429e-01] | [8.3579683303833, 10.25439167022705] |
ddfd3d56-c22d-4efa-8c01-2e42199edcf4 | real-time-patient-specific-ecg-classification | 2110.02215 | null | https://arxiv.org/abs/2110.02215v1 | https://arxiv.org/pdf/2110.02215v1.pdf | Real-Time Patient-Specific ECG Classification by 1D Self-Operational Neural Networks | Despite the proliferation of numerous deep learning methods proposed for generic ECG classification and arrhythmia detection, compact systems with the real-time ability and high accuracy for classifying patient-specific ECG are still few. Particularly, the scarcity of patient-specific data poses an ultimate challenge to any classifier. Recently, compact 1D Convolutional Neural Networks (CNNs) have achieved the state-of-the-art performance level for the accurate classification of ventricular and supraventricular ectopic beats. However, several studies have demonstrated the fact that the learning performance of the conventional CNNs is limited because they are homogenous networks with a basic (linear) neuron model. In order to address this deficiency and further boost the patient-specific ECG classification performance, in this study, we propose 1D Self-organized Operational Neural Networks (1D Self-ONNs). Due to its self-organization capability, Self-ONNs have the utmost advantage and superiority over conventional ONNs where the prior operator search within the operator set library to find the best possible set of operators is entirely avoided. As the first study where 1D Self-ONNs are ever proposed for a classification task, our results over the MIT-BIH arrhythmia benchmark database demonstrate that 1D Self-ONNs can surpass 1D CNNs with a significant margin while having a similar computational complexity. Under AAMI recommendations and with minimal common training data used, over the entire MIT-BIH dataset 1D Self-ONNs have achieved 98% and 99.04% average accuracies, 76.6% and 93.7% average F1 scores on supra-ventricular and ventricular ectopic beat (VEB) classifications, respectively, which is the highest performance level ever reported. | ['Moncef Gabbouj', 'Turker Ince', 'Serkan Kiranyaz', 'Ozer Can Devecioglu', 'Junaid Malik'] | 2021-09-30 | null | null | null | null | ['arrhythmia-detection', 'ecg-classification'] | ['medical', 'medical'] | [ 2.34792084e-02 3.25957723e-02 -1.34634480e-01 -2.32672691e-01
-4.30082351e-01 -3.11145961e-01 -1.92567304e-01 2.48973817e-01
-3.81543845e-01 7.61442304e-01 -3.94321084e-01 -3.68502468e-01
-5.06447196e-01 -4.26638186e-01 -3.44376087e-01 -5.43087959e-01
-4.45950925e-01 4.08335954e-01 -1.44993901e-01 4.63158078e-02
-2.21872360e-01 6.75026298e-01 -1.27241325e+00 2.26722658e-01
9.78949964e-01 1.42158926e+00 -6.16741031e-02 5.64615250e-01
4.15138662e-01 4.80150312e-01 -7.45073557e-01 -1.40369520e-01
3.13929826e-01 -5.76108456e-01 -4.06594515e-01 -3.16918701e-01
1.59999728e-01 -2.69724965e-01 -2.10200325e-01 5.98255277e-01
1.28878725e+00 -4.13604915e-01 4.93467540e-01 -9.98801529e-01
-2.59930313e-01 6.80202901e-01 -1.75248921e-01 5.17903686e-01
-1.58779532e-01 2.09452480e-01 7.46509969e-01 -8.39440644e-01
4.65731531e-01 3.46514374e-01 1.18724954e+00 5.25371909e-01
-1.18007088e+00 -7.50066280e-01 -3.73169005e-01 -7.79412091e-02
-1.71063387e+00 -1.41894966e-01 7.54379570e-01 -3.62742245e-01
1.09039676e+00 3.19540739e-01 9.81091738e-01 8.66596341e-01
2.67208099e-01 4.58864480e-01 7.90898442e-01 -3.47739518e-01
1.75164431e-01 7.43075367e-03 1.72508508e-01 6.74050987e-01
4.76471394e-01 1.31510615e-01 -5.04683912e-01 -1.69468671e-01
9.55010533e-01 7.10761547e-02 -5.65569401e-01 -2.44431213e-01
-1.25502789e+00 5.61409116e-01 4.84933913e-01 6.46750867e-01
-5.52970171e-01 -8.86498839e-02 8.01554263e-01 3.65629256e-01
1.66591182e-01 9.25806940e-01 -6.70341134e-01 -4.14177984e-01
-1.07161927e+00 1.55870482e-01 6.35993421e-01 5.99477053e-01
1.76347733e-01 4.93858844e-01 -3.17200154e-01 8.69569480e-01
-3.17058206e-01 1.72701821e-01 5.25304615e-01 -5.37772238e-01
3.34454477e-01 9.03281987e-01 -2.93869764e-01 -1.05542576e+00
-8.87950122e-01 -1.51545131e+00 -1.53663194e+00 1.87440664e-02
1.49385631e-01 -2.87716478e-01 -9.60282385e-01 1.64845598e+00
-1.72706515e-01 -2.15117577e-02 2.64013231e-01 8.24223816e-01
1.09704959e+00 1.99732482e-01 -6.69497326e-02 -4.49337780e-01
1.26868713e+00 -4.45455849e-01 -6.61377907e-01 -1.19682786e-03
9.13605630e-01 -2.22144842e-01 7.13836074e-01 4.60120916e-01
-9.76055026e-01 -7.48272359e-01 -1.23190069e+00 4.45919126e-01
5.22821061e-02 4.22017723e-01 6.31454051e-01 8.82748902e-01
-9.42184567e-01 7.59982169e-01 -6.91164255e-01 -3.10575485e-01
9.55723882e-01 6.38515651e-01 -3.23954672e-01 3.80009681e-01
-1.25696576e+00 7.49332547e-01 5.32263815e-01 4.12809402e-01
-8.37561071e-01 -9.70321894e-01 -5.10377228e-01 2.45161071e-01
1.66958541e-01 -7.09339499e-01 8.78355324e-01 -7.13663816e-01
-1.19184589e+00 8.78458202e-01 3.69923234e-01 -1.06485391e+00
4.51919764e-01 4.59467508e-02 -5.48089027e-01 1.96003854e-01
8.31554364e-03 4.63361323e-01 4.85793859e-01 -7.98688769e-01
-4.77421075e-01 -2.33002201e-01 -1.87673178e-02 5.76643720e-02
-6.68797255e-01 -2.71444529e-01 -1.84198812e-01 -8.36362720e-01
2.55948603e-01 -9.38561976e-01 -1.09566867e-01 -9.73237455e-02
-4.26728994e-01 -3.99185307e-02 7.10437834e-01 -4.12829131e-01
1.66617680e+00 -2.36257720e+00 4.53321449e-02 1.95813149e-01
7.36377954e-01 8.30796242e-01 2.37290785e-01 1.18886694e-01
-3.16668689e-01 2.43054003e-01 -4.44681942e-01 4.29737493e-02
-3.07132095e-01 1.20521024e-01 1.70350805e-01 3.48524302e-01
2.38864049e-01 1.08257639e+00 -5.31884670e-01 -4.01817888e-01
1.24258235e-01 3.77711892e-01 -4.85349745e-01 3.76042537e-02
3.02107513e-01 3.92884493e-01 -2.97077000e-01 6.35525167e-01
4.32183981e-01 -3.76074433e-01 3.83206636e-01 -4.15889829e-01
8.49562287e-02 6.62333611e-03 -1.08212280e+00 1.70989406e+00
-2.04141691e-01 5.64228594e-01 -2.14291632e-01 -1.26642573e+00
1.11805618e+00 7.21786261e-01 8.33086669e-01 -7.52778232e-01
2.14393169e-01 6.21514499e-01 7.94959068e-01 -3.13545048e-01
-1.81521162e-01 -2.20148131e-01 8.04331228e-02 2.40992889e-01
1.67703539e-01 2.18468785e-01 4.38325293e-02 -1.13251589e-01
1.17854905e+00 -1.66597798e-01 4.42370623e-01 -6.86907470e-01
3.23624551e-01 -2.58173734e-01 8.64613891e-01 1.14339697e+00
-3.51557910e-01 8.78488958e-01 5.53112388e-01 -1.03693140e+00
-7.62313426e-01 -8.16132307e-01 -6.79491878e-01 2.38486066e-01
-1.89543635e-01 -3.77759665e-01 -7.55747437e-01 -4.08889472e-01
-1.21030465e-01 1.68131098e-01 -4.98213619e-01 -2.04211503e-01
-7.15860248e-01 -1.08734643e+00 1.27802825e+00 8.02580476e-01
7.58035541e-01 -1.25604784e+00 -1.35880446e+00 6.04662716e-01
-1.37268100e-02 -1.03675508e+00 2.31371280e-02 6.26912355e-01
-1.20709491e+00 -9.59901571e-01 -9.51621175e-01 -8.93757164e-01
3.43272209e-01 -4.96193260e-01 1.16122890e+00 2.64486581e-01
-6.31077290e-01 -2.26324499e-01 -1.73772022e-01 -7.70565033e-01
-1.91465378e-01 5.40105343e-01 1.93467662e-01 1.26238674e-01
-5.92950499e-03 -7.25534678e-01 -8.34451139e-01 1.35789767e-01
-4.99458969e-01 9.09126773e-02 5.99519670e-01 1.02838850e+00
4.58489925e-01 5.34398220e-02 1.14583921e+00 -8.49375010e-01
6.31167591e-01 -1.02164708e-01 -2.06337824e-01 1.25740571e-02
-1.02610505e+00 -4.50646281e-01 6.65968657e-01 -2.10528433e-01
-3.83267373e-01 9.46475193e-02 -1.59632996e-01 -4.80075449e-01
-4.30082530e-02 6.94021344e-01 6.95641488e-02 6.47250041e-02
9.56238329e-01 2.99035698e-01 6.09015636e-02 -1.93662867e-01
-5.19090414e-01 6.20802164e-01 4.62705910e-01 -3.58281821e-01
4.04480368e-01 1.58673614e-01 2.21776798e-01 -6.89209163e-01
-5.96879184e-01 -2.41223440e-01 -6.34636819e-01 -4.17626649e-02
9.66680884e-01 -8.79350483e-01 -7.59877563e-01 7.35666037e-01
-9.62238431e-01 -1.85489073e-01 -4.97186244e-01 4.00238961e-01
-2.92105794e-01 3.02236408e-01 -5.43668687e-01 -8.21715534e-01
-9.49100435e-01 -1.14047718e+00 5.86263776e-01 8.57450441e-02
-5.21843612e-01 -8.52775216e-01 -4.56344873e-01 -5.27379215e-02
7.14174151e-01 7.09092736e-01 1.10787582e+00 -8.41608047e-01
-3.37466896e-01 -4.08055961e-01 -5.84909357e-02 6.33618236e-01
2.92587489e-01 -4.05244827e-01 -9.47989345e-01 -3.48998219e-01
8.62908289e-02 -4.30954695e-02 5.01102149e-01 6.91102207e-01
1.22366929e+00 1.75751194e-01 -3.16324174e-01 8.91828060e-01
1.34590244e+00 6.54707134e-01 7.14295030e-01 2.78002471e-01
5.24675488e-01 -1.05018251e-01 1.92417532e-01 4.59701419e-01
-9.43729430e-02 5.90533137e-01 4.11740929e-01 -5.84207773e-01
1.05969077e-02 2.89144367e-01 -3.55354816e-01 8.00064147e-01
-3.98511708e-01 -1.62617192e-01 -1.28122187e+00 6.38947546e-01
-1.71517599e+00 -6.85079336e-01 -1.51736915e-01 2.22294140e+00
5.71869612e-01 5.28656721e-01 9.95857120e-02 6.88423991e-01
5.82647324e-01 -4.05210406e-02 -7.06231833e-01 -4.33091104e-01
-4.33316678e-01 5.17378211e-01 1.63918361e-01 -2.32680246e-01
-1.15701270e+00 2.14379892e-01 5.77782774e+00 6.41033649e-01
-1.38310826e+00 5.77462874e-02 8.81547153e-01 -6.50743023e-02
4.42688793e-01 -3.47462624e-01 -5.60999870e-01 2.46403158e-01
9.53999698e-01 1.58124454e-02 -1.75630804e-02 6.53608620e-01
1.46085754e-01 3.99976730e-01 -1.08338189e+00 1.45594180e+00
-1.33297846e-01 -1.67749453e+00 -1.23433016e-01 1.55354701e-02
5.73206782e-01 -4.03100811e-02 -1.26844585e-01 2.68400043e-01
-8.97395790e-01 -1.33014214e+00 5.35888433e-01 3.32484215e-01
1.34165990e+00 -8.80021095e-01 1.29011929e+00 3.74932855e-01
-1.10493147e+00 -4.33030874e-01 -7.69275501e-02 -2.06027463e-01
3.00613083e-02 6.65705502e-01 -7.04305589e-01 7.22696841e-01
9.34599936e-01 9.59334373e-01 -4.03744191e-01 1.09628212e+00
2.90636122e-01 8.55578184e-01 -2.60759145e-01 3.94609757e-02
1.41229779e-01 1.55052066e-01 5.17405093e-01 1.17725182e+00
3.02258074e-01 1.13845259e-01 7.49411210e-02 6.84897482e-01
-1.93548486e-01 1.76134080e-01 -4.55175161e-01 3.56664136e-02
2.26100177e-01 1.01857030e+00 -7.99355567e-01 -3.75333101e-01
-6.34707212e-02 6.21271133e-01 5.18446155e-02 1.69032335e-01
-9.30242360e-01 -7.51162171e-01 4.32356447e-01 3.46250355e-01
1.07376792e-01 2.48592138e-01 -8.79836798e-01 -8.03073168e-01
3.37020159e-01 -1.15277040e+00 3.37459832e-01 -3.70631337e-01
-9.18309510e-01 1.13505900e+00 -3.12736064e-01 -1.53755569e+00
4.29801457e-02 -6.06144488e-01 -3.85187298e-01 9.43398535e-01
-1.06613958e+00 -8.16935837e-01 -3.96685004e-01 3.29521745e-01
3.92843008e-01 -5.85274458e-01 1.32943285e+00 6.40990734e-01
-5.37814915e-01 8.88159215e-01 -1.68492839e-01 4.90100414e-01
3.22247326e-01 -1.05244970e+00 1.66474432e-01 6.51238799e-01
-6.03825711e-02 7.37713933e-01 2.37963930e-01 -3.45009595e-01
-1.07115388e+00 -9.91770208e-01 7.93701589e-01 -2.02871874e-01
4.06868085e-02 -3.26921016e-01 -9.03920710e-01 2.35282093e-01
-2.19732197e-03 3.09576958e-01 5.70163369e-01 1.72864243e-01
2.60760449e-02 -5.38522780e-01 -8.92578125e-01 4.01630729e-01
1.24106491e+00 -3.92994821e-01 -2.77577162e-01 2.42792249e-01
2.07963377e-01 -6.28336847e-01 -1.30529594e+00 1.10357690e+00
7.15266049e-01 -1.19145548e+00 8.75277042e-01 -4.53575253e-01
4.08941478e-01 -2.41192609e-01 1.70206636e-01 -9.92040336e-01
-3.32996637e-01 -7.57743835e-01 -1.24468565e-01 9.29232836e-01
5.90468585e-01 -9.03877795e-01 8.20498109e-01 1.64208487e-01
-6.25454843e-01 -1.54484141e+00 -1.21723199e+00 -8.69742870e-01
-1.85019612e-01 -4.11085516e-01 4.98581052e-01 8.46048176e-01
-2.55067617e-01 3.32603246e-01 -4.37295496e-01 3.70097090e-03
4.68736470e-01 9.29607004e-02 3.13179702e-01 -1.54234684e+00
-2.89252043e-01 -5.18925428e-01 -8.64073277e-01 -5.16053736e-01
-2.41074547e-01 -1.16309738e+00 -3.70887667e-01 -1.41310394e+00
2.56521285e-01 -7.86724031e-01 -8.69198978e-01 6.08151674e-01
-2.34137364e-02 5.63092172e-01 9.68276039e-02 2.11685181e-01
-3.02826196e-01 8.74313563e-02 1.06178355e+00 -5.26229590e-02
-4.54842448e-01 1.08922288e-01 -6.87057614e-01 5.09233594e-01
8.96042347e-01 -5.57404995e-01 -5.05515575e-01 -3.28823298e-01
-3.25206146e-02 2.50266850e-01 4.14827585e-01 -1.58155346e+00
9.23352391e-02 6.56948328e-01 4.37411994e-01 -5.06273091e-01
2.03499824e-01 -5.77347934e-01 4.22747523e-01 1.01530147e+00
-1.76203340e-01 1.82097793e-01 3.68101388e-01 2.95234591e-01
-4.59243387e-01 3.97283658e-02 7.66419530e-01 3.62141989e-02
-3.36318851e-01 5.60280085e-01 -3.82236391e-01 1.57052934e-01
9.30124879e-01 -7.06929982e-01 2.00773086e-02 -7.60329887e-02
-8.14740419e-01 -1.10901214e-01 -2.55845524e-02 2.90225029e-01
6.11699164e-01 -1.09615827e+00 -8.08031619e-01 3.24140400e-01
2.50545889e-01 1.78419948e-01 5.32075286e-01 1.07702219e+00
-8.73063505e-01 6.65935934e-01 -5.18633187e-01 -1.01983941e+00
-1.06973565e+00 1.87525555e-01 8.76488149e-01 -3.73118907e-01
-1.15724945e+00 6.60625756e-01 1.41494855e-01 -2.15449780e-01
2.83363551e-01 -3.66091102e-01 -8.43743607e-02 -5.92331961e-02
1.07462466e-01 2.67073691e-01 6.92418694e-01 -1.95893809e-01
-5.64472258e-01 5.27666450e-01 1.37902707e-01 6.07880294e-01
1.46237302e+00 4.89544302e-01 1.19009800e-02 4.26887721e-01
9.73628104e-01 -4.87442493e-01 -7.99794912e-01 9.75382924e-02
-1.54767126e-01 8.94333702e-03 2.11441354e-03 -1.07016277e+00
-1.33002400e+00 1.09855914e+00 1.09875774e+00 2.84428209e-01
1.47808254e+00 -3.64605635e-01 7.84018099e-01 2.08476141e-01
2.96978712e-01 -8.80049706e-01 -1.55692011e-01 2.48558044e-01
8.08348656e-01 -8.74427915e-01 -4.83864099e-02 -2.94101775e-01
-5.57540536e-01 1.19413662e+00 5.13296306e-01 1.82382215e-03
7.78992414e-01 3.31962973e-01 4.03094500e-01 -3.70894670e-01
-4.11405772e-01 2.13128567e-01 1.52013794e-01 6.29611909e-01
5.12092650e-01 7.94403106e-02 -4.95047241e-01 9.18167233e-01
-1.42502580e-02 2.62105346e-01 2.57788509e-01 8.50795746e-01
8.89903829e-02 -7.49973774e-01 1.51060224e-01 9.35001493e-01
-7.53284156e-01 -2.03630492e-01 1.28690243e-01 8.53075564e-01
2.82508910e-01 5.70993781e-01 -1.69520095e-01 -3.00933391e-01
5.84591985e-01 2.74882764e-01 2.00921923e-01 -5.61808705e-01
-1.08374953e+00 -1.61792994e-01 3.36888582e-02 -2.07018748e-01
-2.07390472e-01 -3.27266127e-01 -1.17489767e+00 2.15747148e-01
-3.11316490e-01 -3.80107462e-02 5.35773098e-01 7.13044286e-01
8.56561184e-01 9.87410903e-01 2.74176806e-01 -5.33447683e-01
-7.51705706e-01 -9.78273273e-01 -8.30098152e-01 3.64153206e-01
2.53971517e-01 -5.11261046e-01 -1.17625289e-01 -1.54783681e-01] | [14.263278007507324, 3.2765533924102783] |
f41be975-06c4-448b-8682-3068c6d48a81 | gligen-open-set-grounded-text-to-image | 2301.07093 | null | https://arxiv.org/abs/2301.07093v2 | https://arxiv.org/pdf/2301.07093v2.pdf | GLIGEN: Open-Set Grounded Text-to-Image Generation | Large-scale text-to-image diffusion models have made amazing advances. However, the status quo is to use text input alone, which can impede controllability. In this work, we propose GLIGEN, Grounded-Language-to-Image Generation, a novel approach that builds upon and extends the functionality of existing pre-trained text-to-image diffusion models by enabling them to also be conditioned on grounding inputs. To preserve the vast concept knowledge of the pre-trained model, we freeze all of its weights and inject the grounding information into new trainable layers via a gated mechanism. Our model achieves open-world grounded text2img generation with caption and bounding box condition inputs, and the grounding ability generalizes well to novel spatial configurations and concepts. GLIGEN's zero-shot performance on COCO and LVIS outperforms that of existing supervised layout-to-image baselines by a large margin. | ['Yong Jae Lee', 'Chunyuan Li', 'Jianfeng Gao', 'Jianwei Yang', 'Fangzhou Mu', 'Qingyang Wu', 'Haotian Liu', 'Yuheng Li'] | 2023-01-17 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Li_GLIGEN_Open-Set_Grounded_Text-to-Image_Generation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_GLIGEN_Open-Set_Grounded_Text-to-Image_Generation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['image-inpainting'] | ['computer-vision'] | [ 4.06593800e-01 3.15668672e-01 -2.37913623e-01 -1.67943642e-01
-7.67160118e-01 -9.18789089e-01 1.08658206e+00 -5.76251335e-02
-1.54212549e-01 7.68222749e-01 5.25944173e-01 -4.09074068e-01
2.83467084e-01 -9.17453766e-01 -8.63683760e-01 -5.00849426e-01
9.60658193e-02 5.72363973e-01 4.17115569e-01 -5.40144145e-01
3.90995383e-01 2.81213611e-01 -1.08599603e+00 6.67564154e-01
8.41080010e-01 5.98913610e-01 4.56271768e-01 9.27717686e-01
-3.87087584e-01 1.15186894e+00 -3.91071051e-01 -3.22067708e-01
1.90612480e-01 -6.53169811e-01 -7.80263722e-01 4.46133725e-02
6.06955409e-01 -4.77363318e-01 -5.79187155e-01 5.54603517e-01
3.53904635e-01 -6.37944564e-02 7.46729195e-01 -1.14972031e+00
-1.55536783e+00 7.59897172e-01 -4.55003619e-01 6.85417093e-03
9.11647007e-02 4.65772718e-01 1.02855742e+00 -9.52188551e-01
1.31348217e+00 1.19211555e+00 3.50878805e-01 8.34714174e-01
-1.48915255e+00 -5.21145046e-01 4.55290198e-01 -4.02136236e-01
-1.24523008e+00 -4.56899971e-01 5.37233353e-01 -6.73443794e-01
1.21908259e+00 -5.28836250e-02 7.77632713e-01 1.45085096e+00
1.57239437e-01 8.77594471e-01 9.66798842e-01 -4.47716266e-01
3.66221040e-01 1.61692396e-01 -4.52756971e-01 1.10255527e+00
1.43643796e-01 -2.37668660e-02 -7.87465870e-01 2.35377178e-01
1.00793779e+00 -9.63355973e-02 -1.82938471e-01 -5.77976465e-01
-1.31037486e+00 9.02786911e-01 8.53738725e-01 2.28730798e-01
3.83999534e-02 6.47110760e-01 5.83680980e-02 1.42187551e-01
6.60779178e-01 6.66645706e-01 -4.55867760e-02 1.02012493e-01
-1.24183786e+00 2.67083615e-01 5.59046268e-01 1.29102385e+00
7.90480375e-01 1.57070801e-01 -6.12586081e-01 2.87337184e-01
-2.79078055e-02 7.29737937e-01 4.15115952e-01 -9.55935001e-01
7.23464787e-01 6.07534468e-01 7.88810104e-02 -6.63631082e-01
-6.02598600e-02 -4.48363304e-01 -6.92697585e-01 2.51836121e-01
2.63788939e-01 -2.49721602e-01 -1.45194244e+00 1.54416215e+00
-9.36896428e-02 -2.37167954e-01 -8.14599320e-02 7.13595212e-01
4.30007905e-01 8.58118236e-01 4.20099497e-03 3.27586025e-01
5.88513553e-01 -1.35898912e+00 -3.40749472e-01 -4.02530909e-01
7.83995569e-01 -5.58192670e-01 1.40269566e+00 1.16308868e-01
-1.09676898e+00 -2.28951335e-01 -1.08302081e+00 -3.95691872e-01
-6.06215239e-01 -8.36627707e-02 6.17974222e-01 4.49204117e-01
-1.48454225e+00 5.83936512e-01 -5.98700583e-01 -5.01202047e-01
7.32851803e-01 1.69912696e-01 -2.94250280e-01 -3.53373140e-01
-8.54837716e-01 5.51297963e-01 3.20289016e-01 -2.12476745e-01
-1.21531641e+00 -6.10213161e-01 -6.74502671e-01 -2.62663309e-02
3.39328289e-01 -1.09744751e+00 1.05688679e+00 -1.04101944e+00
-1.55308378e+00 6.70412242e-01 -2.03882251e-03 -4.61343080e-01
6.66922867e-01 -6.44330755e-02 1.01286449e-01 1.88338086e-01
1.95967451e-01 1.50492799e+00 1.01649773e+00 -1.41567457e+00
-3.36689800e-01 3.56817096e-02 1.82885587e-01 2.07883671e-01
-6.62180364e-01 -4.80421424e-01 -6.54055357e-01 -7.43014574e-01
-2.46505097e-01 -8.63799751e-01 -3.43678713e-01 4.13282871e-01
-5.98168194e-01 2.40999982e-01 7.40638077e-01 -3.21545988e-01
1.10352492e+00 -1.99140131e+00 3.94016594e-01 1.25048980e-01
3.39383513e-01 1.57688651e-02 -6.66771412e-01 8.47437859e-01
2.46872783e-01 5.90364397e-01 -2.62575626e-01 -5.21599650e-01
5.74664176e-02 2.22432330e-01 -4.97320652e-01 1.55741721e-02
4.08274174e-01 1.57269061e+00 -1.03724253e+00 -6.59348369e-01
1.86006308e-01 4.37912345e-01 -7.99450219e-01 -1.11755943e-02
-7.70273983e-01 3.91576082e-01 -4.44997400e-01 3.60031754e-01
3.00889403e-01 -4.65259552e-01 -3.80545929e-02 2.16106579e-01
-1.29815921e-01 -5.18476358e-03 -5.73773682e-01 2.24800181e+00
-7.13307559e-01 8.39329898e-01 -1.14413798e-01 -4.45915490e-01
6.67147815e-01 5.30894585e-02 1.91037461e-01 -8.55497539e-01
-1.82532892e-01 1.27738312e-01 -2.09065631e-01 -2.15467349e-01
6.94047272e-01 -4.09493111e-02 -7.60391727e-02 5.36060274e-01
1.05364762e-01 -6.10548437e-01 1.98844001e-01 9.36858654e-01
1.14160538e+00 5.31187475e-01 -2.82583952e-01 -3.74006301e-01
-1.10289715e-01 2.86052138e-01 -3.69957499e-02 9.97699916e-01
2.94565946e-01 1.02986372e+00 4.56669778e-01 -9.80489030e-02
-1.43817806e+00 -1.33383405e+00 4.37341362e-01 1.00653791e+00
8.65197927e-03 -5.12519836e-01 -1.00452900e+00 -7.46718109e-01
-1.73975408e-01 7.78387725e-01 -8.48017693e-01 -6.80491328e-04
-3.44266206e-01 -2.75677085e-01 4.63095993e-01 6.03434861e-01
5.23077071e-01 -8.73517513e-01 -3.43523413e-01 2.68492609e-01
-4.69106883e-02 -1.01852536e+00 -8.34366560e-01 8.07030872e-02
-8.69476318e-01 -5.22005498e-01 -1.13498580e+00 -7.32209504e-01
9.38182533e-01 3.03113610e-01 1.27588224e+00 1.31197691e-01
-4.49970871e-01 3.30077320e-01 -3.37955087e-01 -2.21608236e-01
-3.51275772e-01 5.68878889e-01 -5.59091985e-01 -2.22836450e-01
-3.34721595e-01 -3.80114675e-01 -9.02964115e-01 5.88188739e-03
-1.18159008e+00 5.02753973e-01 6.76244080e-01 8.84579480e-01
2.94330239e-01 -2.65527993e-01 4.31640983e-01 -1.00220537e+00
7.15885341e-01 -3.40765595e-01 -5.25734365e-01 2.42124915e-01
-6.23181820e-01 3.55555415e-01 6.13014817e-01 -4.95090634e-01
-1.04800451e+00 5.84468208e-02 8.60443339e-02 -3.98959130e-01
1.50104254e-01 3.17720681e-01 -7.95309804e-03 8.96066055e-02
9.46808755e-01 1.64929762e-01 -3.78854990e-01 -1.87012732e-01
1.19276810e+00 3.94215375e-01 3.94265085e-01 -7.69323349e-01
1.01393998e+00 6.91580117e-01 -2.24256694e-01 -7.96452522e-01
-6.99696124e-01 -2.79515442e-02 -6.45477176e-01 1.92571357e-02
1.08790731e+00 -9.07509625e-01 -8.77678618e-02 3.92253876e-01
-1.16871166e+00 -1.19594514e+00 -5.33055663e-01 -1.89018488e-01
-4.52997237e-01 2.34450549e-02 -7.90939987e-01 -6.56957567e-01
-3.31134468e-01 -9.67581987e-01 1.26065505e+00 -6.46417961e-02
-1.72006562e-01 -1.16032207e+00 -9.92158428e-02 1.05056517e-01
7.96516597e-01 2.12099239e-01 9.12155688e-01 -8.25690180e-02
-1.20875514e+00 9.08794254e-03 -3.34037781e-01 1.96825966e-01
-1.20687909e-01 7.72956572e-03 -7.91444957e-01 -2.67410129e-01
-6.04492486e-01 -7.12012768e-01 1.19823992e+00 1.50575578e-01
9.66155231e-01 -5.32751620e-01 -5.10622025e-01 5.55109918e-01
1.48519504e+00 -1.66902721e-01 7.37008095e-01 1.91117600e-01
9.11169946e-01 4.47993279e-01 -6.79180324e-02 1.30192205e-01
4.37293917e-01 4.04479384e-01 3.06167930e-01 -5.51846325e-01
-5.53093910e-01 -9.99444783e-01 2.98385710e-01 5.85355759e-01
2.00959519e-01 -8.85824442e-01 -8.20980728e-01 7.11017609e-01
-1.77400506e+00 -9.35487092e-01 1.58405557e-01 1.94794619e+00
7.56456196e-01 3.48888397e-01 -2.16153368e-01 -4.08518761e-01
3.64251763e-01 3.61128241e-01 -5.78328609e-01 -2.95180202e-01
-3.35752130e-01 2.44746014e-01 6.05333626e-01 6.83683097e-01
-6.46673262e-01 1.63011134e+00 6.72370672e+00 9.43734944e-01
-1.28013599e+00 2.10979179e-01 9.16431129e-01 -4.36839193e-01
-9.51716721e-01 2.37915635e-01 -5.49061000e-01 2.75517162e-02
5.47485352e-01 7.93026760e-02 6.84026241e-01 4.84846354e-01
8.26938227e-02 -1.18236750e-01 -9.81793404e-01 6.50957167e-01
2.04743698e-01 -1.80368030e+00 4.61835742e-01 2.18073249e-01
1.30749011e+00 2.25464866e-01 4.80961889e-01 1.19623341e-01
8.77434552e-01 -1.21971977e+00 1.08733737e+00 5.16341507e-01
1.15832388e+00 -3.82115602e-01 3.49302553e-02 3.12888503e-01
-9.21905220e-01 -6.20260648e-02 -2.03942537e-01 -6.59008399e-02
3.22871000e-01 4.22880441e-01 -7.35870540e-01 2.82430798e-01
2.82857209e-01 5.47066450e-01 -6.82108402e-01 5.28695643e-01
-3.91317487e-01 6.73295677e-01 -3.16219479e-01 -2.18156055e-01
7.26517498e-01 2.56048050e-02 2.27668002e-01 1.18204987e+00
4.23220992e-01 -1.29884705e-01 1.90425500e-01 1.38920200e+00
-3.08426231e-01 1.11374177e-01 -1.04870546e+00 -4.51444387e-01
2.48251483e-01 1.17693794e+00 -9.09565628e-01 -4.90198880e-01
-2.92571276e-01 1.73176897e+00 5.32898188e-01 8.30544770e-01
-6.90154850e-01 -3.60559672e-01 2.69852310e-01 4.99052674e-01
5.37251413e-01 -6.26452625e-01 -4.25451547e-01 -1.24194586e+00
-1.42801926e-01 -5.22379339e-01 -1.06984138e-01 -1.19823396e+00
-1.06093061e+00 6.56682074e-01 -6.17890544e-02 -5.51236808e-01
-2.00199798e-01 -5.57285249e-01 -6.68064892e-01 9.31720972e-01
-1.40610826e+00 -1.75241661e+00 -1.21924736e-01 5.45842648e-01
5.74402750e-01 -5.86070456e-02 5.74193716e-01 -2.49292105e-02
-3.75404030e-01 5.63771963e-01 1.30360201e-01 2.19070256e-01
6.47499621e-01 -1.23847115e+00 1.04713929e+00 9.12590265e-01
4.74158973e-01 7.21564293e-01 5.15646398e-01 -8.46933544e-01
-1.32138979e+00 -1.28642070e+00 7.61689961e-01 -7.82452404e-01
8.15141678e-01 -9.85777199e-01 -4.53401834e-01 7.36908495e-01
7.58760273e-01 -1.44666791e-01 1.42907158e-01 -5.57471365e-02
-6.02748334e-01 8.68959129e-02 -6.92656279e-01 9.98177707e-01
1.60054779e+00 -6.64954066e-01 -1.21503711e-01 3.46141666e-01
9.63433325e-01 -3.07317644e-01 -3.97083253e-01 -1.97964832e-01
3.02156568e-01 -8.47912788e-01 9.11149323e-01 -5.18413067e-01
8.45553637e-01 -1.48053676e-01 -5.70973158e-02 -1.40407550e+00
-4.38594222e-01 -9.68746722e-01 1.76353857e-01 1.23208904e+00
8.64089489e-01 -4.49660629e-01 9.48727190e-01 4.09948975e-01
-2.04684258e-01 -6.21143639e-01 -5.63118935e-01 -7.22652912e-01
6.19652331e-01 -1.65553451e-01 4.95669156e-01 6.80187404e-01
5.11881895e-02 6.70767784e-01 -3.29435915e-01 -3.84343028e-01
4.26077217e-01 -6.97526783e-02 7.45979786e-01 -6.39197648e-01
-4.36225206e-01 -5.33193648e-01 9.96440649e-05 -1.47851646e+00
-6.30602539e-02 -1.08597386e+00 1.47070348e-01 -2.01970887e+00
1.71040162e-01 -4.71184343e-01 8.34187865e-02 6.14592612e-01
-6.37517720e-02 5.89594603e-01 5.10890305e-01 2.20588729e-01
-6.38767779e-01 6.69120729e-01 1.62807906e+00 -4.99141693e-01
-1.95553899e-01 -7.14446485e-01 -7.46005237e-01 2.38054767e-01
7.34358490e-01 -3.02717119e-01 -9.20029461e-01 -9.50372696e-01
5.24334252e-01 -2.73857147e-01 4.22858596e-01 -1.06230342e+00
3.63804847e-01 -1.67558253e-01 4.53187078e-01 -2.65379310e-01
3.00235003e-01 -3.32318574e-01 -1.43865600e-01 2.69059986e-01
-6.25095546e-01 3.65477167e-02 1.76552966e-01 6.42554820e-01
9.63539854e-02 3.89303155e-02 5.01021326e-01 -4.09140974e-01
-6.04914665e-01 4.59565252e-01 -5.15853763e-01 3.98267597e-01
8.72649848e-01 -2.04597786e-01 -7.26250172e-01 -6.19423211e-01
-3.71934652e-01 9.59262848e-02 9.24285233e-01 5.79715312e-01
6.03440344e-01 -1.14584768e+00 -6.28237724e-01 1.03076391e-01
2.67345011e-01 -2.45257244e-02 7.25361472e-03 3.72121096e-01
-6.39356434e-01 5.63609064e-01 3.76938432e-02 -4.90095854e-01
-5.86040974e-01 8.90837669e-01 1.63554981e-01 -1.47741094e-01
-6.72713220e-01 9.54208076e-01 5.12224495e-01 -2.12342501e-01
-8.24736506e-02 -3.97482425e-01 5.24117053e-01 -2.96724290e-01
2.43359491e-01 -1.57769233e-01 -2.38199934e-01 -2.54382432e-01
1.26963720e-01 6.59247398e-01 -1.63066596e-01 -8.14996481e-01
1.09595811e+00 -1.74281240e-01 1.66346446e-01 2.78685540e-01
1.12039483e+00 5.19331805e-02 -1.59953630e+00 -3.41576375e-02
-1.95004880e-01 -2.87188917e-01 1.03064053e-01 -1.05895102e+00
-9.01854336e-01 1.33501959e+00 2.50263780e-01 -1.88481495e-01
7.02202082e-01 7.78032839e-02 8.87253225e-01 4.91304636e-01
3.17393512e-01 -1.19246566e+00 7.52677500e-01 4.91374582e-01
9.96007621e-01 -1.04232824e+00 -2.05804393e-01 -3.00072879e-01
-7.45136678e-01 9.10800993e-01 5.35255849e-01 -5.03905267e-02
3.47878069e-01 4.92192775e-01 1.68029085e-01 -1.68614790e-01
-8.82071018e-01 -2.92261571e-01 8.01291540e-02 8.87070954e-01
4.14096028e-01 -1.94523469e-01 1.30042031e-01 -1.64265394e-01
-8.25417135e-03 1.58960775e-01 3.55462193e-01 9.01541770e-01
-4.97733355e-01 -1.14917076e+00 -7.48539269e-02 2.49718890e-01
6.44979402e-02 -6.40961766e-01 -7.14842975e-01 7.65292645e-01
2.18086094e-02 8.34608018e-01 2.99633685e-02 -2.70534784e-01
3.18444036e-02 -3.56424716e-03 4.82218266e-01 -8.22259009e-01
-4.70054537e-01 5.82126416e-02 6.44041523e-02 -6.01263285e-01
1.81995891e-02 -2.24066734e-01 -1.33642387e+00 -3.59706849e-01
-1.50284767e-01 -1.12275898e-01 5.82768202e-01 6.97228372e-01
6.17597103e-01 5.32945871e-01 2.27265850e-01 -9.76699889e-01
-2.55291253e-01 -7.89173901e-01 -3.14486057e-01 4.93794650e-01
3.18985254e-01 -2.28616759e-01 2.52289027e-02 2.67509431e-01] | [11.270124435424805, -0.07982894778251648] |
b1bda8d5-af23-42d1-a6f2-1a4213a5e5a6 | a-review-on-generative-adversarial-networks-1 | 2302.09119 | null | https://arxiv.org/abs/2302.09119v3 | https://arxiv.org/pdf/2302.09119v3.pdf | A Review on Generative Adversarial Networks for Data Augmentation in Person Re-Identification Systems | Interest in automatic people re-identification systems has significantly grown in recent years, mainly for developing surveillance and smart shops software. Due to the variability in person posture, different lighting conditions, and occluded scenarios, together with the poor quality of the images obtained by different cameras, it is currently an unsolved problem. In machine learning-based computer vision applications with reduced data sets, one possibility to improve the performance of re-identification system is through the augmentation of the set of images or videos available for training the neural models. Currently, one of the most robust ways to generate synthetic information for data augmentation, whether it is video, images or text, are the generative adversarial networks. This article reviews the most relevant recent approaches to improve the performance of person re-identification models through data augmentation, using generative adversarial networks. We focus on three categories of data augmentation approaches: style transfer, pose transfer, and random generation. | ['Laura Alvarez-Gonzalez', 'Anabel Martin-Gonzalez', 'Victor Uc-Cetina'] | 2023-02-17 | null | null | null | null | ['person-re-identification', 'pose-transfer'] | ['computer-vision', 'computer-vision'] | [ 4.98896599e-01 -4.49839979e-03 2.03035265e-01 -2.46621490e-01
-6.73952922e-02 -6.74026906e-01 8.61599684e-01 -1.05171010e-01
-5.51811576e-01 8.88544321e-01 1.41692653e-01 3.54205340e-01
2.21092343e-01 -7.36191154e-01 -6.20865464e-01 -5.99576175e-01
4.34146762e-01 9.38244402e-01 -1.55935779e-01 -2.69981027e-01
7.61293545e-02 7.53191113e-01 -1.77511263e+00 7.50603080e-02
6.92918658e-01 5.07671893e-01 -2.17757851e-01 8.11083913e-01
-1.45318002e-01 9.30897221e-02 -9.39398170e-01 -6.70598686e-01
3.03078800e-01 -5.26975214e-01 -5.09229898e-01 3.44563276e-01
5.85574269e-01 -3.63356322e-01 -3.48455489e-01 1.14302444e+00
5.83812356e-01 1.57078281e-01 6.77358389e-01 -1.38582098e+00
-6.52873516e-01 1.99118271e-01 -2.58111447e-01 -2.90333088e-02
6.99001431e-01 1.17723837e-01 -9.58126858e-02 -5.47512233e-01
6.43876612e-01 1.54320061e+00 5.21707475e-01 1.21921861e+00
-1.40292275e+00 -6.52911663e-01 -1.65240154e-01 2.42553338e-01
-1.27529597e+00 -4.22628164e-01 7.50978053e-01 -4.89814997e-01
3.06122899e-01 2.90827423e-01 7.26209819e-01 1.68717921e+00
-4.18119043e-01 3.83191705e-01 1.32871234e+00 -7.12081850e-01
2.27124467e-02 6.55146837e-01 -1.95035562e-02 3.52089703e-01
4.02029008e-01 2.06183732e-01 -2.59826779e-01 -1.49159953e-01
9.04313266e-01 8.26722383e-02 -1.83385313e-01 -2.89149970e-01
-1.12338328e+00 6.29863441e-01 7.13093802e-02 2.97880620e-01
-1.40394390e-01 -1.93974316e-01 4.37823474e-01 1.75817341e-01
3.27842832e-01 4.81196910e-01 -2.17064649e-01 1.56895667e-01
-7.85320401e-01 5.24548829e-01 6.26814067e-01 7.79500723e-01
5.19383609e-01 1.38386682e-01 -2.49019533e-01 8.21027875e-01
2.39078268e-01 8.72712493e-01 7.10524499e-01 -5.03684521e-01
3.57618839e-01 6.17189527e-01 2.32287616e-01 -8.13208878e-01
-2.10299775e-01 -1.46081047e-02 -9.88710940e-01 3.83281261e-01
8.73162150e-01 -3.44753385e-01 -9.65279758e-01 1.66583133e+00
3.60326409e-01 1.70234829e-01 1.75171182e-01 6.58753932e-01
7.70611644e-01 3.81234705e-01 1.75314113e-01 -3.31712477e-02
1.25120592e+00 -7.98468471e-01 -6.81063473e-01 -1.70019478e-01
1.27962366e-01 -9.22018409e-01 5.02841890e-01 2.52228141e-01
-8.07669520e-01 -1.08687568e+00 -8.02569985e-01 2.77904361e-01
-6.94169402e-01 1.73454165e-01 8.20647851e-02 1.16597390e+00
-8.43667328e-01 6.11002266e-01 -3.69680226e-01 -7.27217793e-01
2.80253708e-01 6.03762746e-01 -7.50799775e-01 -4.64558862e-02
-1.07399392e+00 8.01614165e-01 5.76396346e-01 -7.07289651e-02
-6.29654765e-01 -3.48180145e-01 -7.81113923e-01 -1.76698089e-01
3.24413143e-02 -6.79805934e-01 6.52282119e-01 -1.39939892e+00
-1.46843958e+00 1.16180503e+00 4.43376228e-02 -5.24604499e-01
7.78258860e-01 -2.68063545e-01 -5.25854945e-01 -1.94134414e-01
-2.67734438e-01 8.34932864e-01 1.28530240e+00 -1.38804352e+00
-2.46526062e-01 -6.43774033e-01 -1.19361751e-01 1.07020929e-01
-4.24509019e-01 1.96504712e-01 -3.77306134e-01 -7.53148258e-01
-5.73909402e-01 -1.37179053e+00 -1.97468072e-01 -3.38129073e-01
-4.31633651e-01 -1.27397175e-03 1.05159378e+00 -1.03680313e+00
5.23478210e-01 -2.03174400e+00 4.07143831e-01 3.20662618e-01
-1.47895202e-01 8.49257588e-01 -1.08326294e-01 9.04363394e-02
-3.39824915e-01 1.51180118e-01 -6.43697102e-03 -5.12892187e-01
-3.05086255e-01 1.23708189e-01 -7.15721212e-03 2.16229409e-01
-1.79691613e-02 7.26047814e-01 -5.80378830e-01 -5.10218740e-01
6.55714869e-01 7.74713159e-01 -2.22566992e-01 4.74762082e-01
2.80591752e-02 1.25669789e+00 -2.48152778e-01 2.97919601e-01
6.95730388e-01 3.57506156e-01 -3.36514562e-02 -3.68215829e-01
1.64961278e-01 -5.05526841e-01 -1.40959513e+00 1.37360585e+00
-2.41688803e-01 5.37761092e-01 -1.73525766e-01 -8.67955506e-01
1.01670039e+00 4.84723747e-01 4.92326975e-01 -4.61659789e-01
1.88672587e-01 -5.78862876e-02 -1.99463144e-01 -4.15196985e-01
4.67010736e-01 3.41804400e-02 3.37578909e-04 2.82618374e-01
1.58566266e-01 9.39579234e-02 2.74007112e-01 -3.08506578e-01
4.12622362e-01 2.88014293e-01 8.15773830e-02 1.39206484e-01
1.01462436e+00 -1.78556293e-01 2.85354078e-01 3.99979174e-01
-2.31974140e-01 9.32567894e-01 -3.49773318e-02 -5.71351111e-01
-1.56704009e+00 -7.87693620e-01 9.50067956e-03 6.90540612e-01
-4.50668037e-02 8.04116651e-02 -1.35296750e+00 -7.28841662e-01
-1.08108647e-01 3.27595711e-01 -7.14249313e-01 -2.39549696e-01
-7.35203505e-01 -8.27005684e-01 7.73458481e-01 4.28209275e-01
7.15486765e-01 -1.21511185e+00 -6.04046956e-02 6.50501326e-02
-2.80563027e-01 -1.27811623e+00 -3.93920273e-01 -4.31658417e-01
-1.01107275e+00 -1.02997041e+00 -1.23790860e+00 -6.90772772e-01
9.74984229e-01 3.97043936e-02 9.77394938e-01 4.49318439e-02
-4.08192962e-01 6.83850706e-01 -4.36002105e-01 -5.90264797e-01
-8.48682761e-01 -6.50193393e-02 4.34212685e-01 3.91966581e-01
3.84328604e-01 -2.85324544e-01 -4.11480635e-01 4.51982796e-01
-9.24952030e-01 1.21050319e-02 3.48298073e-01 8.24532211e-01
4.89613801e-01 -7.53462613e-02 3.07921499e-01 -9.18530464e-01
5.79782128e-01 2.00202595e-02 -4.22425479e-01 3.14470470e-01
-3.27055514e-01 4.60079573e-02 6.50959194e-01 -6.38218701e-01
-1.19573045e+00 4.09554183e-01 -3.17485511e-01 -4.29824412e-01
-9.52844262e-01 -2.72895455e-01 -3.33069086e-01 -3.18498403e-01
7.91160643e-01 8.87988806e-02 2.23341480e-01 -5.06655157e-01
1.78862214e-01 6.42644644e-01 5.39336145e-01 -2.32475534e-01
1.15194869e+00 3.79159659e-01 -1.03500508e-01 -1.17996609e+00
-3.46089363e-01 -2.68813342e-01 -9.66367424e-01 -4.14407134e-01
1.18607664e+00 -6.08737409e-01 -5.36203682e-01 9.38264251e-01
-1.25371456e+00 3.12121958e-02 -4.29790854e-01 3.21122199e-01
-3.25953811e-01 4.87243891e-01 -3.12019795e-01 -7.62573957e-01
-4.69362855e-01 -1.10741711e+00 9.26714480e-01 7.39554286e-01
-2.68238157e-01 -1.02188289e+00 2.33609155e-01 6.91735983e-01
3.16252828e-01 4.89459306e-01 5.43230176e-01 -6.92292273e-01
-2.66114444e-01 -6.78185582e-01 -4.09919433e-02 5.15752912e-01
3.87947768e-01 -1.76268548e-01 -1.10387337e+00 -3.92678857e-01
-1.91233382e-01 -1.65296510e-01 4.68172312e-01 2.92012662e-01
1.05664659e+00 -1.94685698e-01 -4.23960090e-01 5.10482728e-01
1.10025990e+00 2.68898845e-01 9.80049074e-01 2.20356315e-01
8.78392041e-01 8.79761755e-01 3.95801097e-01 2.64632314e-01
-7.28850588e-02 1.04999697e+00 2.53825545e-01 -3.55163157e-01
-2.67558217e-01 -2.11025849e-01 2.12744594e-01 1.97243109e-01
-8.49355698e-01 -2.56268889e-01 -4.72795337e-01 1.74702302e-01
-1.67594504e+00 -1.28410518e+00 -1.82961822e-02 2.68898726e+00
3.30171049e-01 -1.61204919e-01 5.07923007e-01 4.28099930e-01
1.30904770e+00 -2.74892986e-01 -4.36066896e-01 -1.91530392e-01
-2.32857898e-01 1.33134589e-01 4.69088167e-01 2.79766083e-01
-1.26266515e+00 7.08820820e-01 6.17652225e+00 3.86968225e-01
-8.89749348e-01 1.35966977e-02 6.68125749e-01 2.64464498e-01
2.14356661e-01 -5.30167639e-01 -1.01557624e+00 7.96977520e-01
8.79864931e-01 7.88339749e-02 5.14977157e-01 7.43283153e-01
8.10036063e-02 1.53553501e-01 -1.15298688e+00 1.41318154e+00
5.84487379e-01 -1.02353024e+00 2.42610618e-01 1.79980248e-01
8.40033591e-01 -5.08229375e-01 1.68720484e-01 6.82374164e-02
-2.90017854e-02 -1.11488569e+00 3.83184463e-01 7.18344390e-01
7.31943429e-01 -7.93355823e-01 1.02653575e+00 1.38016254e-01
-8.95267665e-01 1.53531581e-01 -1.34104371e-01 3.21214765e-01
2.02759534e-01 7.04490691e-02 -8.20859253e-01 3.31400901e-01
8.02216232e-01 2.46452048e-01 -8.76323581e-01 1.04969704e+00
3.26466165e-03 2.59493142e-01 -1.68227196e-01 1.38404101e-01
-4.67375398e-01 -2.49775678e-01 6.49994791e-01 1.20803618e+00
1.76432103e-01 -1.85463116e-01 7.97517151e-02 6.63804948e-01
-3.71510629e-03 1.74575411e-02 -7.50981152e-01 7.01270029e-02
1.01544596e-02 1.16443956e+00 -5.62019229e-01 -3.64668459e-01
-2.26236165e-01 1.44658303e+00 -2.14531600e-01 2.17009902e-01
-7.44568050e-01 -3.01365871e-02 4.79605079e-01 3.05679888e-01
-1.95080847e-01 -1.90720215e-01 6.65031970e-02 -1.09861088e+00
-7.74988160e-02 -1.23863518e+00 2.86918283e-01 -5.26475310e-01
-1.22788095e+00 7.69185543e-01 1.99134052e-01 -1.26020777e+00
-5.75679898e-01 -7.40273893e-01 -3.58194917e-01 1.04644310e+00
-9.48862731e-01 -1.34927166e+00 -6.66865587e-01 9.05036211e-01
6.95050657e-01 -7.64868975e-01 1.10993648e+00 4.63773549e-01
-5.11371315e-01 7.32068598e-01 8.30583870e-02 3.46184939e-01
9.61373270e-01 -1.17404819e+00 4.86347377e-01 7.55945027e-01
9.23878253e-02 2.91456431e-01 8.97705495e-01 -5.41046798e-01
-1.06295288e+00 -1.04555523e+00 5.66704392e-01 -4.66214240e-01
-4.34972942e-02 -3.44041914e-01 -8.36817443e-01 5.35105526e-01
2.19629869e-01 -1.56460494e-01 7.46512711e-01 -2.23944753e-01
-8.09625834e-02 -1.56666175e-01 -1.43841636e+00 5.36466837e-01
6.72111869e-01 -2.93159395e-01 -4.68335390e-01 2.99995989e-01
1.63343862e-01 -4.12654519e-01 -7.72018194e-01 3.12924713e-01
6.20679557e-01 -8.63091230e-01 1.43762565e+00 -6.07435763e-01
7.37070814e-02 -1.54821813e-01 2.83294648e-01 -1.41268957e+00
-1.23383366e-01 -3.68789673e-01 7.71708935e-02 1.60440230e+00
5.73121458e-02 -5.24122357e-01 1.19768751e+00 8.10788870e-01
6.26215398e-01 1.16953023e-01 -5.79877973e-01 -5.00944674e-01
-1.84267804e-01 1.55135065e-01 6.97298765e-01 7.42936671e-01
-6.06617570e-01 2.21115068e-01 -8.89153779e-01 8.39508101e-02
8.75520051e-01 -2.99330443e-01 1.19763398e+00 -1.52290916e+00
-3.65846157e-01 -1.45730525e-01 -7.05782652e-01 -4.79763061e-01
1.32787853e-01 -4.98771101e-01 -2.89504707e-01 -1.24846363e+00
1.49016455e-01 -3.84942085e-01 1.81646898e-01 2.39392921e-01
-1.45439073e-01 7.64561832e-01 3.97094548e-01 1.91533983e-01
-1.38412654e-01 3.71523976e-01 1.11719108e+00 -2.90887564e-01
-3.23836714e-01 3.46174061e-01 -2.03218937e-01 9.06652331e-01
8.64519835e-01 -1.60511598e-01 -1.10558592e-01 -2.45940343e-01
-1.59596741e-01 -1.29159719e-01 5.43428421e-01 -1.32949364e+00
-1.20598683e-02 6.66126162e-02 1.11702609e+00 -3.16340178e-01
5.34772635e-01 -9.78896022e-01 5.05338728e-01 6.52547479e-01
-1.86523557e-01 1.84672326e-01 2.94479162e-01 3.99820149e-01
-1.66421846e-01 -4.50002581e-01 1.04369390e+00 -4.19869989e-01
-6.04307294e-01 2.54867375e-01 -3.29704821e-01 -2.42992073e-01
1.06855977e+00 -4.73647326e-01 -9.46731046e-02 -4.17660117e-01
-1.05036235e+00 -3.15009207e-01 5.42065680e-01 7.52497911e-01
5.00241578e-01 -1.43976879e+00 -7.84334898e-01 3.94612908e-01
4.85948324e-02 -2.83326864e-01 4.46856439e-01 2.43868709e-01
-5.20369828e-01 3.14370632e-01 -9.18712020e-01 -3.82468164e-01
-1.88215470e+00 7.93497086e-01 4.64471638e-01 -3.61553222e-01
-2.38885283e-01 3.42415929e-01 5.48529886e-02 -3.87390912e-01
1.29340917e-01 3.66294116e-01 -8.21815729e-01 -1.47692636e-01
7.46005952e-01 7.10374236e-01 6.38708174e-02 -1.13473594e+00
-9.62069705e-02 7.44995952e-01 -8.16459656e-02 -1.48872867e-01
9.22840416e-01 7.23699108e-03 1.13856725e-01 6.47832081e-02
7.32638836e-01 -5.94970882e-02 -9.51388597e-01 -3.31362411e-02
-4.46201831e-01 -4.24350113e-01 -4.97062594e-01 -6.07266188e-01
-9.94316161e-01 6.95742726e-01 1.24927223e+00 3.75281453e-01
1.01418066e+00 -2.70063579e-01 5.85394442e-01 2.32766002e-01
4.31380361e-01 -1.21760690e+00 1.62042499e-01 1.77478075e-01
9.31671321e-01 -1.40157616e+00 -2.22076252e-01 -4.45539713e-01
-5.59961200e-01 1.13861585e+00 6.77823544e-01 -8.33703950e-02
4.06889915e-01 7.98024237e-02 1.35310799e-01 3.83651495e-01
2.22686917e-01 -1.82162091e-01 4.70944643e-01 1.14385021e+00
3.59959871e-01 7.02735558e-02 -1.87091872e-01 1.21230155e-01
-1.90800637e-01 -1.15790807e-01 2.85898030e-01 4.18650955e-01
1.03257243e-02 -1.77514219e+00 -1.14081740e+00 3.01583886e-01
-6.12363398e-01 4.11686271e-01 -6.52736485e-01 6.08939350e-01
4.92861956e-01 9.32629347e-01 5.81174269e-02 -3.09362650e-01
4.54878330e-01 7.66844898e-02 8.13237727e-01 -4.56668526e-01
-6.21999979e-01 -3.32502872e-01 -1.46595836e-01 -6.73422962e-02
-7.66881526e-01 -8.27412665e-01 -7.00536907e-01 -2.89734304e-01
-2.58266211e-01 -4.36337292e-02 8.92740548e-01 1.05234420e+00
1.16730042e-01 3.53584379e-01 4.87334102e-01 -1.25047684e+00
-2.72309184e-01 -1.06413686e+00 -1.82757556e-01 1.07959592e+00
6.02653204e-03 -6.71652675e-01 -2.40383260e-02 5.29365897e-01] | [14.665138244628906, 1.004502296447754] |
f2f23280-9030-4a3e-8c5a-dc2af6105bc5 | doc2edag-an-end-to-end-document-level | 1904.07535 | null | https://arxiv.org/abs/1904.07535v2 | https://arxiv.org/pdf/1904.07535v2.pdf | Doc2EDAG: An End-to-End Document-level Framework for Chinese Financial Event Extraction | Most existing event extraction (EE) methods merely extract event arguments within the sentence scope. However, such sentence-level EE methods struggle to handle soaring amounts of documents from emerging applications, such as finance, legislation, health, etc., where event arguments always scatter across different sentences, and even multiple such event mentions frequently co-exist in the same document. To address these challenges, we propose a novel end-to-end model, Doc2EDAG, which can generate an entity-based directed acyclic graph to fulfill the document-level EE (DEE) effectively. Moreover, we reformalize a DEE task with the no-trigger-words design to ease the document-level event labeling. To demonstrate the effectiveness of Doc2EDAG, we build a large-scale real-world dataset consisting of Chinese financial announcements with the challenges mentioned above. Extensive experiments with comprehensive analyses illustrate the superiority of Doc2EDAG over state-of-the-art methods. Data and codes can be found at https://github.com/dolphin-zs/Doc2EDAG. | ['Wei Xu', 'Shun Zheng', 'Wei Cao', 'Jiang Bian'] | 2019-04-16 | doc2edag-an-end-to-end-document-level-1 | https://aclanthology.org/D19-1032 | https://aclanthology.org/D19-1032.pdf | ijcnlp-2019-11 | ['document-level-event-extraction'] | ['natural-language-processing'] | [-1.97070524e-01 8.75634514e-03 -1.49450243e-01 -4.70192045e-01
-1.00879121e+00 -7.62860000e-01 6.90556884e-01 5.72615862e-01
-3.34887654e-01 7.17870891e-01 7.85350025e-01 -5.75132191e-01
-1.27657354e-01 -1.00147092e+00 -5.05596161e-01 -1.75607473e-01
-1.53612047e-02 1.40748858e-01 4.66795057e-01 -2.10464731e-01
1.91964790e-01 1.50079280e-01 -9.53776777e-01 6.39810622e-01
6.74505353e-01 7.42504716e-01 -9.42953676e-02 2.13379323e-01
-3.51055056e-01 9.76937950e-01 -6.98928773e-01 -8.11527848e-01
-8.13279599e-02 -3.46700937e-01 -7.36404359e-01 -9.44423899e-02
-3.12017590e-01 -1.51315257e-01 -5.97926915e-01 1.00468481e+00
6.18395030e-01 8.26085825e-03 4.39268380e-01 -1.09715974e+00
-5.52642584e-01 1.07971478e+00 -5.54075778e-01 6.38562918e-01
4.70502406e-01 -2.00624391e-01 1.31199121e+00 -1.09160435e+00
9.98031735e-01 1.15202582e+00 4.21256244e-01 1.06511652e-01
-5.51114678e-01 -7.11289644e-01 4.60045904e-01 2.30722085e-01
-1.09398293e+00 -2.13788301e-01 9.03257132e-01 -1.73540279e-01
9.22879159e-01 3.66436154e-01 5.07132709e-01 1.15961814e+00
1.94217160e-01 1.12225640e+00 8.22858393e-01 -1.29208878e-01
-7.46368838e-04 -4.00358319e-01 3.80575597e-01 4.05788332e-01
4.39502776e-01 -2.64529347e-01 -5.99872172e-01 -2.72420317e-01
2.29502022e-01 1.46327361e-01 -1.65151238e-01 5.02611339e-01
-1.24599731e+00 8.10559630e-01 1.83370914e-02 4.09886897e-01
-5.02719343e-01 -1.94679365e-01 8.29157710e-01 -3.41187418e-02
6.28452599e-01 7.69829527e-02 -6.36510372e-01 -2.08873838e-01
-6.30392909e-01 4.74502116e-01 8.60172749e-01 1.09054899e+00
1.50720477e-01 -3.28504503e-01 -5.41848958e-01 6.07442379e-01
2.98951775e-01 1.32494718e-01 3.83883804e-01 -3.07272822e-01
1.17268634e+00 8.55671227e-01 8.33418667e-02 -1.33240008e+00
-5.55219889e-01 -5.13409853e-01 -6.49971247e-01 -6.74183428e-01
1.16447650e-01 -5.72818875e-01 -4.81410682e-01 1.58845901e+00
7.78575540e-01 1.58983603e-01 2.25239977e-01 7.32415736e-01
1.36254549e+00 9.74719346e-01 2.43436202e-01 -3.43150377e-01
1.87257707e+00 -7.41237164e-01 -1.08330405e+00 -4.23167944e-01
5.35153508e-01 -8.01610827e-01 8.40550721e-01 1.04488380e-01
-8.30320537e-01 5.21069877e-02 -9.21558559e-01 -1.22926950e-01
-4.24937457e-01 1.26177207e-01 6.93058550e-01 2.26122156e-01
-2.48560254e-02 1.22851238e-01 -8.10186267e-01 4.60492298e-02
3.43528867e-01 -2.19616890e-01 -1.39160350e-01 1.86376020e-01
-1.88199186e+00 5.84902048e-01 8.58303964e-01 1.56848177e-01
-5.73797643e-01 -5.55955887e-01 -8.52627814e-01 2.33820558e-01
8.91183496e-01 -3.17331940e-01 1.46242368e+00 -9.78619084e-02
-9.02363956e-01 7.68055320e-01 -1.37504891e-01 -3.58080536e-01
3.68306071e-01 -4.18315917e-01 -9.83962119e-01 7.48770609e-02
4.05942947e-01 -1.84802964e-01 1.79912940e-01 -6.97433293e-01
-8.64984751e-01 -1.56106070e-01 2.12613568e-01 7.49536380e-02
-3.62255722e-01 5.60651839e-01 -6.59554243e-01 -1.03357792e+00
1.12739496e-01 -5.27964294e-01 -2.40446165e-01 -9.00453031e-01
-9.78079200e-01 -6.77559972e-01 6.95092499e-01 -8.32652271e-01
2.03194427e+00 -2.28122640e+00 -3.35847288e-01 -5.39771654e-02
2.55240470e-01 -4.29013558e-02 8.75691921e-02 1.00275457e+00
-1.75199524e-01 2.75097072e-01 -1.30129531e-01 -2.03828439e-01
2.75501877e-01 3.21035162e-02 -5.58563292e-01 1.61784083e-01
2.37883106e-01 8.71413291e-01 -1.14712930e+00 -7.45117724e-01
-2.62073278e-01 -6.29316550e-03 -3.27245444e-01 -3.71813402e-02
-2.03234494e-01 1.32149458e-01 -1.13157582e+00 6.43403411e-01
4.88262266e-01 -4.61539000e-01 4.55798626e-01 -2.42682114e-01
-1.97686225e-01 9.65951204e-01 -1.24064505e+00 1.45602643e+00
-1.45508498e-01 3.64255935e-01 -1.84273198e-01 -8.67283821e-01
7.15190351e-01 4.84473288e-01 5.25196493e-01 -6.15034282e-01
2.50762343e-01 3.38588119e-01 -1.61509678e-01 -4.21926528e-01
7.05036640e-01 -1.68103212e-03 -7.38997698e-01 2.97706544e-01
-1.81163102e-01 2.58261740e-01 7.58492351e-01 7.27818131e-01
1.22079134e+00 -2.60244817e-01 6.21262729e-01 -3.22180763e-02
4.52721536e-01 5.51346280e-02 1.11155045e+00 6.07663870e-01
-4.95257974e-02 6.43962741e-01 9.77643430e-01 -2.35543266e-01
-7.09090173e-01 -6.62840724e-01 -8.90329406e-02 7.05079317e-01
8.40294585e-02 -1.05078995e+00 -6.81958914e-01 -1.14554143e+00
-2.04389513e-01 8.67452979e-01 -3.99986565e-01 2.43191794e-01
-6.96920574e-01 -1.16837037e+00 7.65507460e-01 3.22224915e-01
6.13415062e-01 -1.09673190e+00 -3.07279795e-01 8.10510635e-01
-7.76210606e-01 -1.46349943e+00 -5.69570720e-01 -8.31589028e-02
-2.86511630e-01 -1.28401566e+00 -3.00855607e-01 -6.48193598e-01
3.89912248e-01 -1.23115964e-01 1.35787416e+00 -1.88861117e-01
-2.86126323e-02 -2.06158116e-01 -7.12958038e-01 -6.77420676e-01
-1.98624641e-01 1.29119322e-01 -2.81921148e-01 -7.88342506e-02
7.27672637e-01 -4.09163117e-01 -6.05323911e-01 1.93165377e-01
-1.25071692e+00 7.95221627e-02 2.98887998e-01 6.01403594e-01
6.56927526e-01 5.12811005e-01 1.04802871e+00 -1.24899852e+00
9.95275795e-01 -8.80228877e-01 -5.64424455e-01 2.28227735e-01
-5.33554733e-01 -2.60213166e-01 5.65988004e-01 -3.25239629e-01
-1.17348087e+00 -2.95094430e-01 -5.11058271e-01 2.68341869e-01
1.64994538e-01 1.21030807e+00 -3.15379053e-01 1.01310039e+00
3.20125937e-01 1.15898363e-01 -7.84844398e-01 -5.23036301e-01
3.30524117e-01 8.45100343e-01 5.75511158e-01 -5.41212738e-01
6.01879478e-01 3.24665993e-01 -3.55777830e-01 -1.95284009e-01
-1.54510903e+00 -4.36079562e-01 -1.55359162e-02 -1.14596881e-01
8.06259871e-01 -9.91750538e-01 -5.62222958e-01 3.97873849e-01
-1.40764654e+00 6.62062466e-02 -1.63342819e-01 6.07225597e-01
7.25836605e-02 3.12233508e-01 -9.23174620e-01 -6.33058310e-01
-4.85831648e-01 -6.86710835e-01 8.10703754e-01 3.03017825e-01
-8.01540688e-02 -7.96836913e-01 1.15352653e-01 1.14384264e-01
-2.15759337e-01 6.22353911e-01 8.74951899e-01 -1.09370255e+00
-3.50083739e-01 -3.67722392e-01 -1.75374731e-01 -1.74306557e-02
1.79174468e-01 -4.29230891e-02 -3.98664713e-01 1.00928962e-01
1.16933912e-01 7.86443502e-02 7.76299834e-01 7.42027834e-02
1.15968990e+00 -6.14557385e-01 -3.08370560e-01 2.14712210e-02
1.19410491e+00 2.87575334e-01 5.58190763e-01 6.25676870e-01
4.48288977e-01 5.25965452e-01 8.97711635e-01 8.08783054e-01
7.27886617e-01 4.95194077e-01 2.65182823e-01 7.15200379e-02
-5.99876717e-02 -5.72947502e-01 3.97143573e-01 1.10710812e+00
2.67286211e-01 -7.54154801e-01 -8.72060895e-01 7.52433419e-01
-2.01218224e+00 -1.10977542e+00 -5.75101912e-01 1.61643136e+00
1.07125282e+00 6.38512850e-01 -1.20260447e-01 1.94884449e-01
8.60744536e-01 4.73391801e-01 -2.86353469e-01 1.06369387e-02
-3.91426384e-01 -1.79977521e-01 2.85113484e-01 1.51347160e-01
-1.31162059e+00 7.74786115e-01 4.96314049e+00 1.14133120e+00
-7.31002033e-01 2.59889394e-01 5.95920265e-01 1.34371324e-02
-5.84826767e-01 2.44926929e-01 -1.28597808e+00 8.83839846e-01
9.01047111e-01 -5.09512722e-01 -1.91868678e-01 6.75194085e-01
3.34408283e-01 1.96282715e-01 -7.58829713e-01 7.37978578e-01
-1.87951863e-01 -1.56747508e+00 -4.59971353e-02 -1.99502874e-02
5.86169362e-01 3.36606838e-02 -3.27178121e-01 4.98223960e-01
2.77486056e-01 -3.10876518e-01 9.29858148e-01 9.94182155e-02
4.53158110e-01 -6.80043876e-01 8.31556082e-01 3.85099798e-01
-1.57078207e+00 1.09193679e-02 -7.83895776e-02 2.02691153e-01
8.95438254e-01 1.28379679e+00 -6.43716872e-01 1.23564863e+00
5.81425905e-01 8.44110072e-01 -3.51024657e-01 7.57089853e-01
-6.61243260e-01 9.72964227e-01 -4.08253580e-01 -2.73350388e-01
4.97390747e-01 7.00971432e-05 8.35120201e-01 1.62991512e+00
3.76185179e-01 6.70579195e-01 1.98438674e-01 5.89050233e-01
-4.36388463e-01 2.85836875e-01 -3.30909371e-01 -3.92779291e-01
6.89269781e-01 1.32712841e+00 -9.19093490e-01 -5.31171501e-01
-6.78360164e-01 6.13570929e-01 1.97668910e-01 2.15697274e-01
-1.12285280e+00 -7.68223524e-01 2.92985052e-01 -2.97918394e-02
3.17626268e-01 -1.76611826e-01 2.46335622e-02 -1.45040524e+00
3.11809510e-01 -9.22084987e-01 9.32518661e-01 -5.21870792e-01
-1.45891523e+00 6.95490956e-01 1.43984452e-01 -1.47515440e+00
-1.98402703e-01 -3.81967068e-01 -7.09188581e-01 4.53069866e-01
-1.27187383e+00 -7.94366658e-01 -2.11084392e-02 4.23658133e-01
6.08950377e-01 5.91727681e-02 4.05323148e-01 8.83672535e-01
-8.94624412e-01 3.39812845e-01 -6.68325126e-02 7.35814810e-01
5.53118050e-01 -1.04021168e+00 7.54650593e-01 1.26684785e+00
2.49501035e-01 6.67047918e-01 6.32951498e-01 -1.11116230e+00
-1.17156637e+00 -1.19260395e+00 1.41013575e+00 -3.33936542e-01
1.01867545e+00 -3.99208605e-01 -8.46915245e-01 7.82067001e-01
2.50209093e-01 -9.51826572e-02 6.16386592e-01 2.19964772e-01
-2.56691724e-01 7.95527622e-02 -8.34840715e-01 6.94161654e-01
1.13742483e+00 -4.58451748e-01 -9.60389674e-01 6.76114798e-01
9.42760229e-01 -8.05396318e-01 -8.15157294e-01 5.25517821e-01
7.18632760e-03 -4.62940842e-01 8.79274249e-01 -9.52750683e-01
7.54800081e-01 -5.31290770e-01 -5.72922193e-02 -1.00545859e+00
-6.53390288e-02 -7.49581933e-01 -5.61470926e-01 1.80198884e+00
8.12253475e-01 -6.59444094e-01 3.22623849e-01 3.66303295e-01
-3.25019807e-01 -8.41591895e-01 -8.50339651e-01 -7.99977660e-01
-3.53674471e-01 -6.30655289e-01 8.95780265e-01 1.09777987e+00
2.14296043e-01 6.13007367e-01 -2.67670900e-01 3.62174988e-01
3.04097205e-01 4.69638586e-01 2.33037964e-01 -9.32637990e-01
-1.65075421e-01 -1.82164952e-01 4.74529602e-02 -1.03435516e+00
-1.51071576e-02 -9.75894868e-01 -1.07578620e-01 -1.91716969e+00
2.99438685e-01 -2.73570508e-01 -3.93853128e-01 5.55146754e-01
-5.38867056e-01 -2.82386869e-01 3.79311331e-02 6.86662644e-02
-8.42837930e-01 5.52560151e-01 1.19242740e+00 2.46686861e-02
-7.10579455e-02 -1.59420237e-01 -8.80871117e-01 8.03242087e-01
7.60805011e-01 -9.24012959e-01 -2.36743182e-01 -3.58922958e-01
6.91907108e-01 1.93300307e-01 8.20049196e-02 -4.05046016e-01
3.26063335e-01 -2.44428307e-01 8.36972371e-02 -9.79780555e-01
-1.90701216e-01 -4.76744026e-01 2.17270091e-01 2.01504827e-01
-2.74534822e-01 3.78489166e-01 1.37287036e-01 6.75371885e-01
-7.16555238e-01 -3.66906255e-01 -4.28444147e-02 -3.26287337e-02
-6.72284245e-01 4.13685739e-01 -2.85591632e-01 6.40618742e-01
8.70112121e-01 4.37675983e-01 -6.90674961e-01 -2.57817805e-01
-5.72941005e-01 4.32668000e-01 -2.47812122e-01 5.33829093e-01
5.01244664e-01 -1.39514720e+00 -1.16340768e+00 -4.25365239e-01
1.50306612e-01 2.57132649e-01 3.56477052e-01 8.68559659e-01
-2.72884965e-01 3.29424024e-01 4.98944074e-01 7.98369274e-02
-1.05943644e+00 4.78723645e-01 -1.57954931e-01 -8.22240174e-01
-8.69498312e-01 6.08297169e-01 1.76622614e-01 -3.50294292e-01
-1.59987152e-01 -3.89000565e-01 -3.46001238e-01 3.44844222e-01
6.23852313e-01 1.15123846e-01 2.21684635e-01 -1.80708274e-01
-6.36742532e-01 -1.63273767e-01 -2.83374488e-01 -2.39764690e-01
1.35174596e+00 -6.27411827e-02 -1.02432065e-01 9.85430852e-02
9.10541832e-01 5.74372113e-01 -9.61152911e-01 -2.24848345e-01
3.76649141e-01 -1.88977838e-01 -1.19354218e-01 -7.05752015e-01
-1.01279688e+00 3.89115393e-01 -3.23495269e-01 6.32911921e-01
1.12469149e+00 1.53812081e-01 1.22826719e+00 2.62059569e-01
2.20584854e-01 -1.12176406e+00 -2.31355324e-01 6.49337769e-01
8.94811690e-01 -8.30335081e-01 1.50681570e-01 -8.14227521e-01
-7.48496234e-01 8.58269453e-01 3.18150461e-01 3.36190052e-02
7.29130507e-01 3.98985237e-01 -2.26533443e-01 -4.97632086e-01
-7.45802522e-01 -9.26940963e-02 2.70199984e-01 -3.01823646e-01
4.31047797e-01 6.25099912e-02 -1.12038255e+00 1.30526578e+00
-1.87893391e-01 -1.90792233e-01 5.44755399e-01 1.19327378e+00
-1.08189605e-01 -1.19962263e+00 3.81839983e-02 4.39880759e-01
-1.18872070e+00 -4.33068961e-01 -2.44860455e-01 7.83014596e-01
-7.83527642e-02 1.11388743e+00 -2.44260117e-01 -1.90134376e-01
6.83469594e-01 3.04714590e-02 -2.33730450e-01 -5.53581536e-01
-7.71804631e-01 1.40623271e-01 7.58181632e-01 -3.44189048e-01
-3.98283392e-01 -8.94440234e-01 -1.69940317e+00 -1.77409947e-01
-2.65033394e-01 4.87315476e-01 4.13054198e-01 9.56065595e-01
5.69730520e-01 8.88072431e-01 6.73177660e-01 8.12504366e-02
-3.00000429e-01 -9.32374895e-01 -5.19619823e-01 5.23579538e-01
2.40264810e-03 -5.74413896e-01 -2.43548259e-01 9.04060900e-02] | [9.066614151000977, 9.174013137817383] |
167a5efb-f515-4fec-aa46-ac89430fc56e | audio-source-separation-using-a-deep | 1412.7193 | null | http://arxiv.org/abs/1412.7193v1 | http://arxiv.org/pdf/1412.7193v1.pdf | Audio Source Separation Using a Deep Autoencoder | This paper proposes a novel framework for unsupervised audio source
separation using a deep autoencoder. The characteristics of unknown source
signals mixed in the mixed input is automatically by properly configured
autoencoders implemented by a network with many layers, and separated by
clustering the coefficient vectors in the code layer. By investigating the
weight vectors to the final target, representation layer, the primitive
components of the audio signals in the frequency domain are observed. By
clustering the activation coefficients in the code layer, the previously
unknown source signals are segregated. The original source sounds are then
separated and reconstructed by using code vectors which belong to different
clusters. The restored sounds are not perfect but yield promising results for
the possibility in the success of many practical applications. | ['Yung-Hwan Oh', 'Han-Gyu Kim', 'Giljin Jang'] | 2014-12-22 | null | null | null | null | ['audio-source-separation'] | ['audio'] | [ 6.66038021e-02 -1.00549959e-01 2.80607671e-01 -5.41860685e-02
-3.90298009e-01 -4.13406223e-01 1.21845797e-01 -3.70892650e-03
1.41584137e-02 4.23123896e-01 3.69591922e-01 4.26958352e-01
-5.36171198e-01 -6.52978957e-01 -2.69546568e-01 -1.22707808e+00
-5.19391179e-01 1.68099165e-01 1.80475637e-02 -1.23144269e-01
-3.19753885e-02 2.32982576e-01 -2.40813279e+00 8.13066661e-01
7.47072279e-01 9.48379278e-01 1.20000236e-01 1.06892514e+00
-3.39000106e-01 8.93194556e-01 -1.07647395e+00 3.42473298e-01
1.74389273e-01 -8.75499725e-01 -4.38608557e-01 1.78594902e-01
1.13005619e-02 5.51680587e-02 -1.71361174e-02 1.31109941e+00
4.57372427e-01 3.61684352e-01 8.42856765e-01 -1.04462266e+00
-3.10138643e-01 1.41936421e+00 1.03651129e-01 8.64866599e-02
2.42075324e-01 -4.49872553e-01 9.30062771e-01 -1.08083606e+00
1.72706485e-01 1.03985405e+00 7.42076576e-01 4.20881599e-01
-1.42495859e+00 -5.01720369e-01 -3.89602333e-01 2.74310738e-01
-1.39607799e+00 -6.37153149e-01 1.25502455e+00 -7.94157088e-01
7.21542120e-01 3.35793704e-01 5.49157798e-01 9.47221518e-01
1.19733484e-03 4.55257922e-01 7.19714165e-01 -5.64446509e-01
4.64130968e-01 3.48207265e-01 2.56432772e-01 2.10378647e-01
-2.17185467e-01 2.79263467e-01 -3.34051788e-01 -1.09851561e-01
4.48446810e-01 4.80523147e-02 -5.30007660e-01 -1.05305165e-01
-8.54758918e-01 7.60874391e-01 4.55549896e-01 1.09453034e+00
-6.93982780e-01 -2.10613653e-01 2.56956905e-01 4.25275385e-01
1.60094425e-01 4.20889705e-01 -3.75722736e-01 2.06251815e-01
-1.20631492e+00 -1.36758640e-01 8.49650621e-01 5.25945961e-01
6.00865245e-01 8.55819941e-01 3.17244917e-01 9.95450556e-01
2.67996728e-01 3.01216930e-01 9.59641755e-01 -9.09688056e-01
4.70659174e-02 4.59246308e-01 -2.43525803e-01 -1.08212602e+00
-3.14681947e-01 -8.40713382e-01 -9.84493732e-01 5.65319061e-01
2.83096582e-01 -4.03756291e-01 -6.59604013e-01 1.37282276e+00
6.69059455e-02 5.11589825e-01 5.46241939e-01 7.72237360e-01
1.00279415e+00 1.15007496e+00 -2.69633740e-01 -3.52257639e-01
1.03995335e+00 -7.53360689e-01 -9.54684138e-01 8.12641829e-02
-1.27234891e-01 -8.49457204e-01 6.33815348e-01 9.69088733e-01
-1.23595917e+00 -1.36867261e+00 -1.17754686e+00 5.58305502e-01
-4.04004693e-01 5.73024750e-01 3.25352699e-01 4.14039075e-01
-1.19364846e+00 7.65607893e-01 -4.14903522e-01 1.14434369e-01
-9.27227437e-02 4.63617265e-01 -2.46519864e-01 6.24443710e-01
-1.14892459e+00 2.79807389e-01 7.18537092e-01 2.35985070e-01
-1.16396606e+00 -4.89192307e-01 -6.37726367e-01 4.66454387e-01
-3.45511168e-01 -1.32251352e-01 8.82729590e-01 -1.80476618e+00
-1.59774470e+00 1.93590000e-01 1.53354749e-01 -3.74250382e-01
-1.45382822e-01 3.62383612e-02 -1.05999708e+00 4.73499805e-01
-1.69773370e-01 2.47270778e-01 1.42811871e+00 -1.60549188e+00
-7.91856766e-01 9.02605951e-02 -6.14537239e-01 -2.59231310e-02
-5.07590771e-01 -5.73627725e-02 2.11279482e-01 -9.48095679e-01
3.10494632e-01 -4.32698011e-01 -1.09008603e-01 -7.27341056e-01
-1.17579982e-01 -5.46717346e-02 7.00352550e-01 -6.39867663e-01
1.22629821e+00 -2.84698296e+00 4.94253814e-01 4.90161180e-01
1.83183476e-01 -1.27055541e-01 -1.76140919e-01 1.27919510e-01
-6.01109505e-01 -3.95592391e-01 -2.44835958e-01 3.68448235e-02
-2.85684735e-01 -8.19159895e-02 -6.02681100e-01 7.74030834e-02
2.74399482e-02 -5.33430614e-02 -8.34222913e-01 -3.55178058e-01
1.55199543e-01 7.39454985e-01 -5.37914753e-01 5.09681523e-01
8.73266309e-02 2.61016130e-01 1.41419508e-02 1.54335395e-01
5.13192296e-01 3.98885369e-01 4.50091362e-02 -1.49938360e-01
-1.79294407e-01 1.81836128e-01 -1.59602737e+00 1.25541103e+00
-2.43395209e-01 9.16816533e-01 6.45306408e-01 -1.22325325e+00
1.06296897e+00 8.36530566e-01 7.17692554e-01 1.56937197e-01
6.14372611e-01 1.79290935e-01 4.35832173e-01 -5.66235602e-01
4.28357780e-01 -3.22261631e-01 5.63065223e-02 2.06160873e-01
9.16748762e-01 -5.63207678e-02 7.90272802e-02 -1.74009159e-01
4.92918551e-01 -3.21564108e-01 -1.73955578e-02 -2.94457525e-01
6.03813112e-01 -2.14676276e-01 1.62429303e-01 5.39262116e-01
6.23168536e-02 6.90749884e-01 9.11917612e-02 -2.34744892e-01
-7.73615420e-01 -1.28156447e+00 -3.11685860e-01 1.12886441e+00
-3.63950819e-01 -1.38457835e-01 -9.72345471e-01 -9.63394716e-02
-9.15935636e-02 4.78514791e-01 -5.07984877e-01 -3.85064542e-01
-3.53025198e-01 -5.82571507e-01 6.35890245e-01 3.39039832e-01
1.47336587e-01 -1.29895508e+00 -2.97389001e-01 8.08006763e-01
-1.69769868e-01 -6.12091720e-01 1.65580541e-01 9.24050987e-01
-9.18311596e-01 -7.66323984e-01 -6.22632205e-01 -1.12588501e+00
4.60037977e-01 9.08605009e-03 7.97045052e-01 -6.73656911e-02
-4.71060462e-02 2.07130700e-01 -4.54783797e-01 -4.89016712e-01
-7.78657138e-01 -2.68393219e-01 4.12294149e-01 6.56144619e-01
1.98440015e-01 -1.03503001e+00 -1.43003628e-01 -3.19478661e-02
-7.78311312e-01 -5.61622441e-01 2.88424641e-01 5.02060056e-01
4.28886235e-01 1.08368039e+00 5.59772253e-01 -1.31779686e-01
8.00133228e-01 -7.03008473e-01 -2.54037857e-01 -4.50733900e-01
-3.59556377e-02 -4.41533998e-02 1.10987544e+00 -8.78915250e-01
-1.15403926e+00 2.11701274e-01 -1.63178951e-01 -3.85862261e-01
-6.34148955e-01 3.83696973e-01 -2.87694931e-01 3.15235287e-01
1.00786555e+00 3.94232541e-01 -1.16727144e-01 -6.87251985e-01
2.56358385e-01 8.90011072e-01 8.74197721e-01 -1.83965117e-01
6.63290799e-01 1.84585780e-01 -7.51573682e-01 -1.12329602e+00
-3.82094592e-01 -3.15689296e-01 -7.00046360e-01 -4.84174222e-01
9.61952984e-01 -9.40940082e-01 -3.33684355e-01 5.75779676e-01
-1.10296774e+00 1.00025989e-01 -8.25742841e-01 9.41717446e-01
-2.61068225e-01 8.83427858e-02 -6.87279403e-01 -1.01358974e+00
1.06750518e-01 -9.60022211e-01 4.44102019e-01 4.72724974e-01
-3.22035044e-01 -8.73997748e-01 2.81456620e-01 -3.89684260e-01
2.31279358e-01 -1.36148572e-01 9.35411751e-01 -9.80121791e-01
8.53037811e-04 -9.51551273e-02 6.57947063e-01 9.71170545e-01
2.40823790e-01 3.29651445e-01 -1.49685860e+00 4.15849835e-02
7.22905278e-01 2.95097474e-02 9.58233654e-01 6.28029764e-01
1.00552356e+00 -3.76895458e-01 -3.50773111e-02 4.93233681e-01
1.18981707e+00 7.94121802e-01 5.23970485e-01 -2.06240222e-01
4.63048995e-01 7.08498418e-01 -2.49062002e-01 2.65540957e-01
-3.26896667e-01 9.80604663e-02 1.43896833e-01 -2.22875550e-01
-2.88277239e-01 -5.92065081e-02 7.19912648e-01 1.67177808e+00
-3.04001272e-01 -4.87349071e-02 -4.72263426e-01 7.58076966e-01
-1.35062718e+00 -1.42985022e+00 -2.41524324e-01 1.87925053e+00
8.78192425e-01 3.01822364e-01 1.64961606e-01 1.22918487e+00
1.03967285e+00 1.00819366e-02 -7.91402757e-02 -5.13990819e-01
-3.97492290e-01 5.48905492e-01 -2.79091835e-01 6.08280361e-01
-1.18738246e+00 4.88924950e-01 7.36392593e+00 8.05620968e-01
-9.72281516e-01 -7.82537758e-02 -5.74230440e-02 -2.79926099e-02
-1.49767309e-01 -5.07662743e-02 -2.27042258e-01 5.30996978e-01
1.26123369e+00 2.25942105e-01 7.46672332e-01 9.82039452e-01
-6.20595354e-04 6.42498136e-02 -9.12458241e-01 9.47498381e-01
1.05067216e-01 -9.28448915e-01 3.51919271e-02 -3.08132350e-01
6.98099971e-01 -3.13859344e-01 1.90788954e-01 2.99164563e-01
4.64950770e-01 -8.69338453e-01 9.46304023e-01 6.05690241e-01
3.58736575e-01 -1.02839375e+00 6.26791060e-01 1.93776205e-01
-1.22648752e+00 -5.61256945e-01 -6.16162121e-01 -2.95372903e-01
-3.47534239e-01 8.36001277e-01 -6.99800253e-01 4.41485286e-01
7.67433107e-01 5.75337291e-01 -2.83276200e-01 9.79728222e-01
-1.98956504e-01 1.12554955e+00 -1.55649558e-01 1.82192862e-01
4.25022244e-02 -2.73058444e-01 7.62387216e-01 1.34497738e+00
4.29158539e-01 1.10213809e-01 -1.43268466e-01 8.84966075e-01
1.97753936e-01 4.18739729e-02 -4.38329548e-01 -7.48265088e-02
2.70419091e-01 1.27689862e+00 -7.58491695e-01 -5.12786806e-01
1.24535672e-01 6.57428265e-01 -2.54489809e-01 6.62831485e-01
-5.35353065e-01 -9.22230601e-01 6.02980852e-01 -2.77811736e-01
4.89699990e-01 1.42098349e-02 -8.04022886e-03 -1.06331110e+00
-2.47420743e-01 -8.90264988e-01 3.70648623e-01 -8.48107398e-01
-1.30310798e+00 1.26023734e+00 -2.09669590e-01 -1.69129574e+00
-5.56112051e-01 -4.70257461e-01 -7.65898705e-01 7.42570996e-01
-9.09100771e-01 -1.99512243e-01 -7.85232335e-02 9.98955667e-01
4.78500694e-01 -8.67527843e-01 1.16531479e+00 2.74403751e-01
-2.56565571e-01 1.41872823e-01 5.41784883e-01 2.55258083e-01
3.57664973e-01 -1.22684944e+00 -4.04255569e-01 7.07286417e-01
6.90595090e-01 4.37886626e-01 7.12823749e-01 -3.30069929e-01
-6.08908176e-01 -7.07068086e-01 7.41807103e-01 -5.70007861e-02
4.57644761e-01 -4.31060106e-01 -1.01029134e+00 9.85895172e-02
5.98487258e-01 -3.10745388e-01 1.04066503e+00 -1.48297787e-01
-2.62347162e-01 -4.23454732e-01 -9.25688207e-01 1.86841875e-01
2.76251793e-01 -7.68206358e-01 -1.32674706e+00 -5.49127236e-02
6.19543552e-01 1.96557432e-01 -6.14799500e-01 -1.56551629e-01
3.88050526e-01 -1.24040735e+00 9.99704480e-01 -6.60301328e-01
5.45861900e-01 -6.07359469e-01 -1.71517581e-01 -1.65595639e+00
-6.92099988e-01 -3.87448728e-01 -1.31980225e-03 1.45160031e+00
4.00579691e-01 -2.78750628e-01 3.44061464e-01 -3.22838724e-01
-5.05227521e-02 1.10408820e-01 -8.48069310e-01 -5.58277905e-01
-1.92186415e-01 -6.84284389e-01 5.95652640e-01 9.60519314e-01
2.85412431e-01 3.85675967e-01 -1.84105337e-01 4.16903466e-01
5.87365568e-01 2.01778531e-01 1.35896161e-01 -1.64406645e+00
-5.94338596e-01 -6.04763091e-01 -4.22555894e-01 -5.82534432e-01
2.22607166e-01 -9.87694800e-01 3.31811130e-01 -1.12709558e+00
-3.40962052e-01 -2.17956491e-02 -7.16014147e-01 2.06378624e-01
1.21548213e-01 1.43315807e-01 1.11871287e-01 3.04558158e-01
4.86138426e-02 5.15951216e-01 4.30757999e-01 -3.40879589e-01
-6.01632595e-01 2.82225311e-01 -4.15906906e-01 1.17168057e+00
8.63084435e-01 -6.62732124e-01 -4.12670910e-01 -3.17106128e-01
-1.94890469e-01 2.20037371e-01 1.88629195e-01 -1.74056983e+00
1.68710425e-01 2.63546467e-01 7.15723336e-01 -6.21840715e-01
2.26313025e-01 -1.28124452e+00 3.51558596e-01 2.87031829e-01
-4.86802608e-01 -4.44119692e-01 9.74109024e-02 3.32509100e-01
-7.71736085e-01 -5.15544176e-01 7.79492617e-01 -9.87715274e-02
-3.02424461e-01 -4.34357524e-01 -1.07249463e+00 -2.78393924e-01
5.87475181e-01 -4.13347214e-01 4.62733716e-01 -6.54730916e-01
-1.31599140e+00 -4.58244532e-01 -3.78836870e-01 2.33476758e-01
7.13176250e-01 -1.49227285e+00 -6.59570217e-01 7.36623287e-01
-4.23844010e-01 -2.85077065e-01 4.62281376e-01 2.87809432e-01
-9.76131931e-02 -2.73961555e-02 -5.97180843e-01 -4.91470963e-01
-1.02122259e+00 7.19701886e-01 5.59919059e-01 3.08928758e-01
-3.05680752e-01 9.47350204e-01 1.75619319e-01 -3.61073948e-02
5.21803975e-01 -4.33887035e-01 -9.21738625e-01 6.70966506e-01
8.69431376e-01 3.44909340e-01 -2.51613501e-02 -8.83937001e-01
-1.17187634e-01 5.62474728e-01 6.08183742e-01 -3.25728595e-01
1.64064181e+00 1.91445649e-01 -3.80935073e-01 7.79116452e-01
1.30724514e+00 5.20716727e-01 -9.83519018e-01 9.99277383e-02
-5.19409515e-02 6.62582219e-02 2.98702102e-02 -6.09497547e-01
-1.36985910e+00 9.78794158e-01 7.96529174e-01 7.70567298e-01
1.56232047e+00 -1.60375088e-01 2.38599837e-01 1.93690747e-01
-7.34598860e-02 -1.06883252e+00 8.34972039e-02 4.29956555e-01
8.23560178e-01 -4.34940428e-01 -4.64775980e-01 4.99931313e-02
-4.44008142e-01 1.54475379e+00 6.86285868e-02 -6.60674691e-01
1.07921314e+00 8.26584399e-01 4.02154744e-01 2.74737459e-02
-4.95403945e-01 -2.37873465e-01 4.41698313e-01 7.97520578e-01
2.62289912e-01 -5.61099015e-02 1.48260340e-01 1.51763117e+00
-5.71620464e-01 -4.58875030e-01 6.33329749e-01 3.77979517e-01
-8.29363942e-01 -8.64762604e-01 -1.18318188e+00 1.17817381e-03
-4.23150033e-01 -8.31437111e-02 -5.41032910e-01 2.84046471e-01
6.10517323e-01 1.27709651e+00 2.46300131e-01 -9.01042223e-01
3.92653644e-01 5.29094398e-01 9.62636918e-02 -4.04685944e-01
-9.81636703e-01 8.76968503e-01 -3.52560580e-01 -5.43327257e-02
-5.72031558e-01 -3.21945518e-01 -1.40551734e+00 3.95388573e-01
-3.64400774e-01 8.25769663e-01 4.58895028e-01 4.38158661e-01
1.68407142e-01 9.45416451e-01 1.07727289e+00 -1.10540009e+00
-2.92464703e-01 -9.95406985e-01 -8.84663343e-01 2.47200325e-01
7.36187875e-01 -3.44354510e-01 -9.19147670e-01 8.24492455e-01] | [15.388568878173828, 5.678995609283447] |
62869af5-ac49-417a-a8df-b8c2f6eb1bf3 | spartans-lt-edi-eacl2021-inclusive-speech | null | null | https://aclanthology.org/2021.ltedi-1.28 | https://aclanthology.org/2021.ltedi-1.28.pdf | Spartans@LT-EDI-EACL2021: Inclusive Speech Detection using Pretrained Language Models | We describe our system that ranked first in Hope Speech Detection (HSD) shared task and fourth in Offensive Language Identification (OLI) shared task, both in Tamil language. The goal of HSD and OLI is to identify if a code-mixed comment or post contains hope speech or offensive content respectively. We pre-train a transformer-based model RoBERTa using synthetically generated code-mixed data and use it in an ensemble along with their pre-trained ULMFiT model available from iNLTK. | ['Gaurav Arora', 'Megha Sharma'] | null | null | null | null | eacl-ltedi-2021-4 | ['hope-speech-detection'] | ['natural-language-processing'] | [-1.2813293e-02 1.3988262e-01 -1.9950454e-01 -3.4774326e-02
-1.5886772e+00 -8.3222026e-01 1.0682065e+00 1.9955160e-01
-1.3524994e-01 2.3953508e-01 8.3203930e-01 -1.0409194e+00
4.5472875e-01 -1.2889838e-01 -3.5260081e-01 -1.5267566e-01
-9.1940634e-02 5.6297588e-01 -1.5571526e-01 -6.0603738e-01
4.4517061e-01 6.0138382e-02 -5.2413243e-01 1.0226108e+00
9.0820634e-01 6.4555854e-01 -2.3707379e-01 1.3821377e+00
-3.3293251e-02 1.9695385e+00 -5.7567579e-01 -8.2616293e-01
4.7237936e-02 -3.0794805e-01 -1.2333610e+00 -3.6251402e-01
4.0607396e-01 -1.1051736e-01 -4.2782557e-01 1.0781491e+00
5.8436298e-01 -8.1168734e-02 8.7401509e-01 -1.0273349e+00
-8.6498213e-01 1.5147067e+00 -2.4796255e-01 4.1904792e-01
5.9096748e-01 1.8895280e-01 1.0705664e+00 -1.3005359e+00
4.9552375e-01 1.3891776e+00 7.7053803e-01 4.4520399e-01
-9.7852451e-01 -8.8559622e-01 -7.5803208e-01 -9.9902945e-03
-1.0912045e+00 -9.6975476e-01 1.0176538e+00 -8.2740498e-01
1.3030934e+00 2.8795397e-01 2.2893423e-01 1.7380255e+00
2.2202255e-01 1.3201306e+00 9.9803150e-01 -2.0941915e-01
-1.6772735e-01 5.4660684e-01 8.3634302e-02 9.3441957e-01
-7.8414375e-01 4.9864337e-02 -6.4056057e-01 -4.7343338e-01
-2.1053429e-01 -4.5493311e-01 -6.4471215e-02 3.3384776e-01
-1.2315431e+00 1.5233642e+00 1.9387461e-01 4.0615901e-01
9.4350919e-02 -1.6679995e-01 1.1875678e+00 7.2272009e-01
8.3825785e-01 7.2920990e-01 -1.2480815e-01 -5.6107926e-01
-1.6055729e+00 7.0340149e-02 8.4768128e-01 7.6516664e-01
2.3780599e-01 3.4017679e-01 -3.4733805e-01 1.3141105e+00
2.0816202e-01 5.1605046e-01 9.0994608e-01 -2.7293950e-01
1.0063134e+00 3.3302683e-01 -3.8949808e-01 -7.6359499e-01
-9.2822328e-02 -5.4397243e-01 -4.5093361e-01 2.2792645e-02
-1.9615816e-02 -3.6568898e-01 -7.8765440e-01 1.0721160e+00
-6.6234922e-01 -1.1486331e-01 5.1287758e-01 2.1681945e-01
9.9551451e-01 8.4632295e-01 -1.5118383e-01 1.5900092e-01
9.9149626e-01 -1.5130711e+00 -3.4846497e-01 -3.4053114e-01
8.4234583e-01 -1.2287163e+00 1.0370249e+00 3.9715153e-01
-8.2555836e-01 -3.3782282e-01 -8.5238892e-01 -2.0579977e-01
-3.6858612e-01 4.5845947e-01 -6.7390157e-03 7.7652091e-01
-1.1246486e+00 3.3064970e-01 -1.1735419e-01 -3.0105057e-01
3.6072367e-01 -2.4620779e-01 -3.6215284e-01 5.0391710e-01
-1.3172356e+00 9.5806849e-01 2.0544708e-01 -5.8149582e-01
-1.7095904e+00 -5.9800208e-01 -1.0036681e+00 -2.2871535e-01
-4.0073477e-02 2.7244747e-01 1.5780573e+00 -1.0003979e+00
-1.4027468e+00 1.3353280e+00 1.8872070e-01 -1.0082405e+00
7.1595836e-01 -1.6508242e-01 -8.9536661e-01 -9.3016498e-02
3.0455959e-01 2.8125072e-01 1.1938775e+00 -8.9791870e-01
-3.0282503e-01 1.5395804e-01 -4.8582155e-02 1.8490290e-02
-3.9496902e-01 8.2208836e-01 4.0010747e-01 -7.8093755e-01
-5.8345282e-01 -7.8336102e-01 3.9079159e-01 -8.2356060e-01
-1.2366319e+00 -2.3115598e-01 1.2258044e+00 -1.1933901e+00
1.5211660e+00 -2.1640086e+00 5.0794769e-02 -8.7819602e-03
1.3058269e-01 5.8096772e-01 -4.9049112e-01 9.5815361e-01
-4.5485434e-01 4.6805313e-01 6.4101154e-03 -6.8115371e-01
-1.4050774e-01 -9.0727948e-02 -1.0230774e+00 3.7223965e-01
3.4810415e-01 7.0102078e-01 -1.0837650e+00 -4.9107894e-01
-8.5107507e-03 2.4664821e-01 -6.7511046e-01 4.9807537e-01
9.1447048e-02 1.0813483e-01 -1.6485330e-01 7.8308064e-01
3.5435244e-01 2.5911367e-01 -4.1679239e-01 1.9167632e-01
-5.6715572e-01 9.3265229e-01 -1.4134070e-01 1.1950036e+00
-8.9256084e-01 1.1424440e+00 1.8670995e-02 -5.5888987e-01
1.1653805e+00 5.8360463e-01 -5.1171202e-02 -2.6875594e-01
3.5800460e-01 3.6063695e-01 4.6460602e-02 -4.1055334e-01
6.3085520e-01 -2.4822339e-01 -3.1396383e-01 3.1975105e-01
2.7640769e-01 -2.1821436e-01 -2.1661428e-01 6.9570577e-01
1.4991972e+00 -6.2878877e-01 4.0496016e-01 -3.4468332e-01
1.0741634e+00 -7.8154579e-02 -2.0695640e-01 8.4904623e-01
-4.3812963e-01 5.1442230e-01 7.1592128e-01 -1.1395706e-01
-1.0941145e+00 -8.4902954e-01 -1.9894631e-01 1.3982146e+00
-7.6930279e-01 -6.7693734e-01 -5.0305867e-01 -1.1209471e+00
-2.0822264e-01 1.1902747e+00 -8.1913888e-01 -1.6299143e-01
-2.4339303e-01 1.3655105e-01 1.4097284e+00 3.1512233e-03
3.6144876e-01 -1.0556980e+00 -7.2074465e-02 2.1254028e-01
-3.2927996e-01 -8.7050271e-01 -8.3262908e-01 6.2526572e-01
2.1106713e-02 -8.7167662e-01 -6.3560766e-01 -6.9254071e-01
-4.2697292e-02 -4.5261834e-02 1.0217732e+00 -7.0898719e-02
-2.3398777e-02 -1.4635518e-01 -7.8692794e-01 -3.6499456e-01
-1.4356308e+00 3.6151829e-01 -1.6196834e-01 -1.1459449e-02
6.4291492e-02 -1.4974155e-01 3.3658963e-01 -2.2148637e-01
-4.2111918e-01 1.5567069e-01 2.5350279e-01 9.9346197e-01
-7.0673400e-01 -4.1197202e-01 3.0338442e-01 -1.0045056e+00
1.1139275e+00 -8.0119598e-01 -7.3139399e-02 1.3530149e-01
-6.1566565e-02 1.3555168e-01 8.5719901e-01 -3.7971061e-01
-7.6289254e-01 -1.6403347e-02 -6.0106444e-01 -7.4106419e-01
1.7343813e-01 7.9157937e-01 2.7341065e-01 -2.3338083e-02
9.7644389e-01 3.7126800e-01 -2.3671277e-01 -6.2550586e-01
2.7420753e-01 1.3992393e+00 6.3659424e-01 -2.0433681e-01
1.0631824e+00 -1.4678060e-01 -8.8960242e-01 -9.3404156e-01
-1.1532869e+00 -8.8220906e-01 -5.4956478e-01 -2.2925594e-01
9.3683136e-01 -1.0352013e+00 -2.7960923e-01 5.0230801e-01
-1.4595370e+00 -2.5541902e-01 2.0615059e-01 -2.3651227e-02
-3.5959691e-01 2.3048757e-01 -7.9511219e-01 -1.1418424e+00
-7.0761967e-01 -1.2989473e+00 1.1397885e+00 -4.4423088e-01
-3.0519074e-01 -1.0704591e+00 6.6512901e-01 6.6418689e-01
5.1474750e-01 8.9082038e-03 9.4298333e-01 -1.4223721e+00
4.7672707e-01 -3.3387849e-01 -2.2459447e-01 6.0545653e-01
-3.6055990e-02 3.5325733e-01 -1.2123629e+00 -1.5520357e-01
-1.3357924e-01 -1.1277354e+00 9.8288769e-01 -2.0039791e-01
6.4422470e-01 -8.5490072e-01 4.3396551e-02 4.1291681e-01
9.9947876e-01 -8.1611432e-02 3.5902750e-01 1.5372117e-01
7.5274891e-01 3.3419314e-01 8.2696147e-02 5.8844119e-01
7.6672547e-02 6.9159657e-01 3.6393306e-01 6.4526781e-02
-1.4350238e-01 -7.3540467e-01 1.3622646e+00 1.2801811e+00
4.8463181e-01 -1.9887872e-01 -1.4471778e+00 6.9214517e-01
-1.4531569e+00 -1.4285411e+00 -3.3584011e-01 1.5979363e+00
1.0851210e+00 2.6882246e-01 5.2842939e-01 -2.7742544e-02
4.4884896e-01 5.1237774e-01 7.5135581e-02 -1.2626961e+00
-5.2577883e-02 6.3938396e-03 4.6391168e-01 9.1213977e-01
-1.6554440e+00 1.2512687e+00 6.4893374e+00 1.1721363e+00
-1.2994068e+00 5.8684969e-01 5.1187849e-01 4.1287772e-02
-2.5025535e-01 6.2160682e-02 -6.5055573e-01 4.4533989e-01
1.7064085e+00 -6.4651489e-01 4.7742477e-01 1.2259631e+00
1.2402243e-01 4.8818466e-01 -9.5737922e-01 8.8854349e-01
5.7523137e-01 -1.2491642e+00 7.5512074e-02 3.8725112e-02
6.0384095e-01 8.9685345e-01 3.4423557e-01 9.6512365e-01
9.4701451e-01 -1.2142369e+00 1.1312102e+00 -1.0930289e-01
7.5231200e-01 -7.9921824e-01 7.6566273e-01 7.7164531e-01
-8.9711440e-01 -2.0737408e-01 -1.7823496e-01 1.4079472e-01
-1.4225556e-01 2.9097676e-01 -1.4153594e+00 1.2610179e-01
1.2156474e-01 1.1494294e+00 -9.3163097e-01 5.0890648e-01
-2.8395894e-01 1.1580687e+00 1.8939435e-01 -8.7685063e-02
9.5766109e-01 3.4635141e-01 9.9115747e-01 1.8400607e+00
3.4265146e-02 -6.2621474e-01 5.1441479e-01 9.6923912e-01
-3.3454841e-01 2.8321254e-01 -8.8627410e-01 -6.1441535e-01
2.8182453e-01 1.4751744e+00 4.1628659e-02 -5.6406927e-01
-2.4016592e-01 1.1438763e+00 4.5200956e-01 -1.5324621e-01
-9.5280272e-01 -3.9375204e-01 4.1945702e-01 -4.3959897e-02
-8.1088163e-02 4.7105808e-02 1.6133124e-01 -1.4634949e+00
-7.8617430e-01 -1.2848527e+00 3.6795598e-01 -7.3063517e-01
-1.2551951e+00 1.1824671e+00 -4.4606572e-01 -1.2519498e+00
-6.8570465e-01 -5.1224411e-01 -1.0748093e+00 9.5513326e-01
-1.2262541e+00 -1.5318236e+00 3.1998390e-01 6.7776185e-01
1.3015169e+00 -9.3488425e-01 8.0699813e-01 1.4293614e-01
-6.2064695e-01 7.8218836e-01 5.3340502e-02 8.5822755e-01
7.8168184e-01 -1.3661522e+00 6.6732508e-01 8.7605280e-01
2.1434100e-01 4.6954733e-01 8.6795205e-01 -6.4059663e-01
-1.0553956e+00 -1.3875852e+00 1.3207945e+00 -8.3575892e-01
1.6250337e+00 -7.5692374e-01 -6.2849915e-01 6.0097337e-01
5.7742965e-01 -3.6203232e-01 7.1249431e-01 4.7122305e-03
-1.0313691e+00 6.9720113e-01 -9.3090868e-01 2.9422012e-01
4.2202765e-01 -1.3963299e+00 -7.9448205e-01 7.5467378e-01
6.6282076e-01 -1.5521891e-01 -6.4620167e-01 -2.2261123e-01
7.9491220e-02 -9.4463050e-01 4.5859748e-01 -8.2278454e-01
1.1970594e+00 1.8050608e-01 -3.3711919e-01 -1.6075273e+00
-2.1160309e-01 -8.6545879e-01 2.9613498e-01 1.3916709e+00
8.1057292e-01 -1.3014756e-01 2.9395032e-01 -5.7894465e-02
-3.9751688e-01 -2.9120323e-01 -9.9754089e-01 -7.9632324e-01
3.6682698e-01 -5.6441170e-01 -2.1521351e-01 1.4119416e+00
7.8964996e-01 9.1769987e-01 -8.0990267e-01 -2.7729177e-01
3.6574122e-01 -3.7417990e-01 6.9022095e-01 -9.2557532e-01
-3.8505176e-01 -5.2848065e-01 -2.1473376e-01 -5.0851899e-01
7.1906012e-01 -1.6223189e+00 -7.6411523e-02 -9.7300005e-01
5.4515523e-01 -1.5368126e-02 4.8934121e-02 9.3398917e-01
2.8611553e-01 2.5312638e-01 3.0067751e-01 3.7055585e-01
-3.5156438e-01 5.1021057e-01 2.7123848e-01 -6.1084080e-01
7.3863298e-02 7.4214442e-03 -4.6221414e-01 6.4956850e-01
6.5346956e-01 -6.0418481e-01 -2.5306958e-01 2.0572942e-02
2.0571677e-01 3.5880753e-01 2.3518677e-01 -1.0484029e+00
-4.9600162e-02 1.2909225e-01 -3.3046764e-01 -7.1376902e-01
1.0114056e-01 -1.5453799e-01 -4.9882731e-01 7.8288430e-01
-1.0381353e+00 1.1848036e-01 1.1821270e-01 2.5850296e-02
-4.1341203e-01 -5.8234358e-01 1.0176734e+00 -5.4050013e-02
-2.9077393e-01 -6.2835231e-02 -9.7799635e-01 3.6605945e-01
5.7822561e-01 2.3996480e-01 -6.7090738e-01 -7.9678243e-01
-4.6554831e-01 -8.9194298e-02 2.8540826e-01 9.2493194e-01
6.7589986e-01 -1.0905447e+00 -1.3479322e+00 1.3967425e-01
5.0347739e-01 -1.2757359e+00 -2.7137458e-01 7.9647350e-01
-4.2868927e-01 5.5497551e-01 8.4810443e-02 -9.7709335e-02
-1.5724695e+00 5.1572496e-01 3.9468813e-01 -4.4353896e-01
-5.0225532e-01 1.1411471e+00 -3.4073484e-01 -7.5389594e-01
3.1470448e-02 5.8590092e-02 -5.0698638e-01 2.9820699e-02
8.5128897e-01 1.4132725e-01 5.8232386e-02 -1.2524644e+00
-4.6030477e-01 -2.1264069e-01 -3.6112761e-01 -2.8488711e-01
1.0440222e+00 4.6041790e-01 -1.3836750e-01 6.2332660e-01
1.8406419e+00 6.4531559e-01 -3.8571751e-01 -1.1449945e-02
1.6665322e-01 -1.4953458e-01 5.1115167e-01 -9.1761774e-01
-7.1371871e-01 1.0841790e+00 3.2691273e-01 3.7357527e-01
4.2610079e-01 7.2172165e-02 8.0153495e-01 2.5855961e-01
-7.3523529e-02 -9.5506310e-01 3.7328836e-01 1.3057286e+00
1.2761316e+00 -1.2037536e+00 -5.7671732e-01 -7.1400531e-02
-1.1138077e+00 1.1275469e+00 1.3997027e-01 -1.6442162e-01
6.4996958e-01 4.7044089e-01 4.3787888e-01 -2.2717251e-01
-1.1310430e+00 -6.9866076e-02 3.3782756e-01 5.2796143e-01
9.4520092e-01 1.6060202e-01 -2.6161049e-02 2.0019549e-01
-8.1337959e-01 -4.7423157e-01 8.1480503e-01 3.0917656e-01
-5.6224960e-01 -8.2622766e-01 -1.5599731e-01 4.5809522e-01
-7.1164888e-01 -8.9759070e-01 -1.0697461e+00 3.9378825e-01
-1.5789255e-01 1.3047296e+00 -2.8663537e-01 -1.2139724e+00
-1.1764305e-01 2.2208923e-01 -5.0537443e-01 -9.4762164e-01
-1.4398986e+00 9.9829815e-02 6.6355062e-01 -3.2723850e-01
1.7651176e-01 -6.7365724e-01 -8.5525799e-01 -2.8721261e-01
-2.0487796e-01 5.7852781e-01 6.9599229e-01 8.3417684e-01
-9.1650426e-02 2.9784435e-01 8.6428010e-01 -6.1587095e-01
-9.1949427e-01 -1.2716502e+00 -3.5477605e-01 4.0061182e-01
7.4557263e-01 -1.2451849e-01 -6.8758440e-01 -1.0424011e-01] | [9.137288093566895, 10.616043090820312] |
3db80d62-beef-4bd0-b3e6-05eb6c8bef1e | plasticity-and-evolvability-under | 2202.08834 | null | https://arxiv.org/abs/2202.08834v3 | https://arxiv.org/pdf/2202.08834v3.pdf | Plasticity and evolvability under environmental variability: the joint role of fitness-based selection and niche-limited competition | The diversity and quality of natural systems have been a puzzle and inspiration for communities studying artificial life. It is now widely admitted that the adaptation mechanisms enabling these properties are largely influenced by the environments they inhabit. Organisms facing environmental variability have two alternative adaptation mechanisms operating at different timescales: \textit{plasticity}, the ability of a phenotype to survive in diverse environments and \textit{evolvability}, the ability to adapt through mutations. Although vital under environmental variability, both mechanisms are associated with fitness costs hypothesized to render them unnecessary in stable environments. In this work, we study the interplay between environmental dynamics and adaptation in a minimal model of the evolution of plasticity and evolvability. We experiment with different types of environments characterized by the presence of niches and a climate function that determines the fitness landscape. We empirically show that environmental dynamics affect plasticity and evolvability differently and that the presence of diverse ecological niches favors adaptability even in stable environments. We perform ablation studies of the selection mechanisms to separate the role of fitness-based selection and niche-limited competition. Results obtained from our minimal model allow us to propose promising research directions in the study of open-endedness in biological and artificial systems. | ['Clément Moulin-Frier', 'Eleni Nisioti'] | 2022-02-17 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 1.12093940e-01 -3.08519661e-01 1.76685467e-01 1.26819208e-01
8.55746269e-01 -6.97744250e-01 6.38900936e-01 3.28045398e-01
-6.30224109e-01 1.05091989e+00 -1.05286427e-01 -3.06465507e-01
-3.56656134e-01 -9.65450227e-01 -6.60855114e-01 -1.10113418e+00
-4.21878248e-01 2.26056799e-01 4.49358076e-01 -5.40222168e-01
1.52244180e-01 5.30436635e-01 -2.08744574e+00 -5.21852672e-01
1.04678512e+00 1.44845650e-01 5.75902998e-01 7.79532373e-01
8.11278597e-02 -1.62291288e-01 -6.00706339e-01 2.80277193e-01
4.12225723e-01 -7.64869630e-01 -3.28021497e-01 -1.73893049e-01
-3.95831436e-01 4.15118068e-01 2.48103961e-03 7.17613518e-01
6.43196762e-01 -1.11549161e-02 7.80879855e-01 -9.32372749e-01
-8.11462164e-01 4.43989307e-01 -1.83419913e-01 3.68624955e-01
-2.22722143e-02 5.74795067e-01 7.66608655e-01 -4.14938211e-01
7.20078409e-01 1.11683381e+00 4.18846846e-01 4.00589466e-01
-1.52175581e+00 -1.30707733e-02 1.67666793e-01 -4.11680490e-02
-1.32012773e+00 -4.02583927e-01 2.70845503e-01 -5.84659934e-01
8.93070281e-01 5.64625204e-01 1.15363681e+00 8.07802439e-01
8.42337430e-01 5.76575985e-03 1.09934425e+00 -4.00161177e-01
6.16569757e-01 1.31321386e-01 -3.32487971e-01 1.95276275e-01
7.91775465e-01 4.72460836e-01 -5.44912517e-01 -2.53016412e-01
8.86870801e-01 -1.77170366e-01 -6.67036474e-01 -3.58817130e-01
-1.03103578e+00 4.18864429e-01 4.01423395e-01 4.46135759e-01
-3.04628760e-01 1.55340791e-01 -6.34924769e-02 5.54866374e-01
-1.45767943e-03 1.03649795e+00 -8.36611271e-01 -2.08436564e-01
-1.84635803e-01 1.04465233e-02 9.19019520e-01 3.42234761e-01
7.44080663e-01 2.64150947e-02 3.28337759e-01 9.53890622e-01
2.99877226e-01 7.84403563e-01 7.17101097e-01 -5.97044647e-01
-5.61293602e-01 6.57046199e-01 2.13703111e-01 -9.44508135e-01
-4.15550083e-01 -7.58742332e-01 -7.20454216e-01 3.86882514e-01
7.09707379e-01 -1.54569685e-01 -4.11309391e-01 2.21733999e+00
3.58682692e-01 -1.86738417e-01 -1.39383703e-01 6.19888246e-01
-1.10640740e-02 5.39764941e-01 -3.03273369e-03 -4.96134341e-01
9.23502922e-01 -4.13652807e-01 -2.72228301e-01 -1.64295062e-01
2.67754376e-01 -4.32892710e-01 1.35083640e+00 4.75612246e-02
-8.99900734e-01 -1.57098800e-01 -1.00342011e+00 7.80047357e-01
-5.69907606e-01 -6.26966715e-01 3.70941639e-01 9.01053250e-01
-1.17467332e+00 8.44294131e-01 -9.14259195e-01 -8.94023716e-01
-1.44609988e-01 1.07952230e-01 -1.21879093e-01 6.59955919e-01
-1.02958643e+00 9.75529373e-01 6.36871066e-03 -5.69855347e-02
-6.74958408e-01 -3.45102727e-01 -3.07891965e-01 1.59675777e-01
1.11504100e-01 -9.81228113e-01 4.15391177e-01 -1.19835925e+00
-1.45172298e+00 7.26957202e-01 3.80639471e-02 -1.84899852e-01
6.19590700e-01 1.53646439e-01 1.49482995e-01 -3.38125855e-01
-1.39434114e-02 3.07376087e-01 3.76213908e-01 -1.24993122e+00
-6.30735755e-02 -4.49725628e-01 -1.28378868e-01 3.09611529e-01
-4.27221328e-01 -2.01896295e-01 2.65778810e-01 -4.68589991e-01
9.39660370e-02 -1.29098344e+00 -5.15492260e-01 1.25041783e-01
7.24298656e-02 2.17069358e-01 2.53541440e-01 3.97152975e-02
9.95844841e-01 -2.00739408e+00 6.84702456e-01 1.84521884e-01
-5.87882586e-02 1.85812134e-02 -2.78956741e-01 7.57939816e-01
3.26422632e-01 4.61543024e-01 -4.47805464e-01 4.22134936e-01
-1.07108526e-01 4.83463854e-01 1.18550919e-01 4.91645902e-01
2.24725261e-01 5.76749802e-01 -8.47435594e-01 -2.20400080e-01
-8.09359550e-02 5.99804759e-01 -5.01361370e-01 -5.71713373e-02
-1.27675727e-01 7.09148526e-01 -2.66573995e-01 5.37855446e-01
3.79663229e-01 -9.16371793e-02 4.31320369e-01 8.32846284e-01
-6.65984273e-01 1.01927808e-02 -1.09042692e+00 8.88953924e-01
-3.39535236e-01 5.48044860e-01 1.87257960e-01 -6.37919605e-01
9.09690201e-01 -1.40206426e-01 1.36411592e-01 -5.75375199e-01
2.69888610e-01 5.93032598e-01 8.77965450e-01 -2.99607843e-01
4.57429886e-01 -1.31389350e-01 2.48111308e-01 5.05113661e-01
-3.91666293e-01 -4.66316581e-01 1.58387288e-01 -2.16869920e-01
1.07398975e+00 2.01269165e-01 4.62242872e-01 -9.67662811e-01
3.99493366e-01 -2.69189894e-01 7.67431676e-01 8.12677324e-01
-5.76713264e-01 1.83530420e-01 4.47024673e-01 -3.15970451e-01
-1.27783406e+00 -1.08348012e+00 -7.27970898e-01 1.20244360e+00
5.93828499e-01 -9.89021882e-02 -4.66507107e-01 3.35116237e-01
1.16556406e-01 4.90571588e-01 -9.84056711e-01 -6.17634058e-01
-6.44627929e-01 -1.40018821e+00 3.82690370e-01 4.90484275e-02
1.45878777e-01 -1.27361429e+00 -1.31316984e+00 2.08679080e-01
3.00019324e-01 -2.55163759e-01 -1.00279681e-01 5.99386334e-01
-8.25404108e-01 -7.16521561e-01 -4.19620365e-01 -2.98732936e-01
6.02032065e-01 1.23832025e-01 1.14747345e+00 7.35678494e-01
-6.21498168e-01 2.87821084e-01 -2.91659504e-01 -4.33085322e-01
-7.00807810e-01 9.68193635e-02 5.58531821e-01 -3.01410854e-01
-2.98670799e-01 -1.27094710e+00 -7.20187068e-01 7.59018779e-01
-1.06923366e+00 -2.97873855e-01 3.85426372e-01 9.27045882e-01
5.46061575e-01 1.56116411e-02 7.80111372e-01 -3.38890791e-01
6.68469310e-01 -6.74674094e-01 -4.33099568e-01 3.95216703e-01
-7.13108420e-01 2.52469182e-01 6.85455203e-01 -7.40062714e-01
-7.12819934e-01 -9.09440592e-02 5.98845109e-02 5.86063027e-01
-4.48666774e-02 2.84386992e-01 -3.26578081e-01 -9.65566933e-02
7.68760383e-01 3.69027615e-01 -6.79719821e-02 -1.75234661e-01
1.47119343e-01 3.81297141e-01 -1.11004084e-01 -8.47957790e-01
7.23549306e-01 4.12278354e-01 2.77716190e-01 -1.32980907e+00
1.48474962e-01 3.86999369e-01 -7.47363448e-01 -4.45433021e-01
5.41409194e-01 -4.08694714e-01 -6.01003110e-01 5.35776138e-01
-6.16028130e-01 -6.06639326e-01 -6.91422641e-01 2.25033760e-01
-4.60691541e-01 2.02212363e-01 -8.74800533e-02 -1.05952287e+00
-6.04714379e-02 -9.54987526e-01 4.60379779e-01 5.61408639e-01
-2.25612715e-01 -1.02805960e+00 5.42674661e-01 -4.46231991e-01
8.57414842e-01 4.65497315e-01 8.99868071e-01 -2.68907338e-01
-5.80807924e-01 4.24747020e-01 3.93138319e-01 -1.54419065e-01
2.61763573e-01 5.32113135e-01 -5.96329927e-01 -1.50171667e-01
-4.14532162e-02 1.15002394e-01 8.82917106e-01 4.88511831e-01
3.05231631e-01 -8.60945955e-02 -3.19933295e-01 6.93913102e-01
1.43520105e+00 4.80942637e-01 5.08646548e-01 8.78586590e-01
1.74830221e-02 9.49026406e-01 1.06641605e-01 5.32832265e-01
-6.14921823e-02 4.49366242e-01 6.03654027e-01 2.07620278e-01
2.83878475e-01 1.91267520e-01 3.61623585e-01 6.42740905e-01
-1.78983331e-01 -3.79404306e-01 -1.05152249e+00 6.54267848e-01
-1.73864603e+00 -9.55573082e-01 -2.49071911e-01 2.56513977e+00
8.81311655e-01 8.42548832e-02 2.01937035e-01 -6.37314245e-02
5.73191941e-01 -2.93109834e-01 -9.24508095e-01 -5.31405091e-01
-9.11441386e-01 -6.25619739e-02 4.86959517e-01 6.93427682e-01
-4.60685641e-01 5.62244654e-01 7.30761576e+00 1.41969025e-01
-1.18427610e+00 -2.17951640e-01 5.04190743e-01 -2.56646186e-01
-6.00438297e-01 2.06738740e-01 -4.10199612e-01 4.68624800e-01
9.76247847e-01 -4.25455749e-01 4.63232100e-01 4.47141469e-01
5.74877203e-01 -4.02534574e-01 -7.70323694e-01 3.53900865e-02
-4.53280091e-01 -7.65523016e-01 -3.15230936e-01 3.36721420e-01
6.30039692e-01 6.30808249e-02 1.46583140e-01 -2.27570757e-01
3.35176557e-01 -9.81669247e-01 7.47865498e-01 4.85758185e-01
2.83378541e-01 -3.43453854e-01 3.32537651e-01 4.05054033e-01
-9.73195553e-01 -2.50313163e-01 -3.64157796e-01 -6.46710753e-01
1.01280339e-01 5.61582625e-01 -4.33845490e-01 -1.43602602e-02
6.51138663e-01 3.00281912e-01 -8.22597802e-01 1.16074932e+00
-7.43799508e-02 4.36992824e-01 -5.56116700e-01 -3.69830310e-01
-1.76508740e-01 -4.82734263e-01 9.88331258e-01 6.68434322e-01
2.77931184e-01 -1.20675936e-01 -3.15758169e-01 1.05632889e+00
5.39540589e-01 2.11653784e-01 -6.20943785e-01 -1.02832399e-01
6.33788943e-01 8.18458319e-01 -1.09700239e+00 6.19047061e-02
1.97860926e-01 6.99886918e-01 2.26178616e-02 2.49572322e-01
-6.16642356e-01 -1.18263692e-01 1.04581928e+00 2.98004687e-01
9.17780027e-02 -5.66579700e-01 -3.64758074e-01 -1.15813398e+00
-2.23925278e-01 -6.26042664e-01 1.03268605e-02 -2.34932542e-01
-1.11824214e+00 6.53587818e-01 -1.85071543e-01 -5.19579709e-01
1.92279607e-01 -4.02706295e-01 -6.22939110e-01 8.78751397e-01
-1.31051016e+00 -4.83515561e-01 -1.35821325e-03 1.50031075e-01
3.28741759e-01 3.97526398e-02 5.87712646e-01 -2.19873875e-01
-9.14195836e-01 1.94843054e-01 9.68036890e-01 -8.12682867e-01
5.82990110e-01 -1.18192959e+00 2.88321882e-01 8.70221913e-01
-3.40506405e-01 1.04264331e+00 1.15183163e+00 -9.67944980e-01
-1.35192633e+00 -7.06258774e-01 4.86995727e-01 -3.80038440e-01
5.65095246e-01 -3.94838303e-01 -9.23194170e-01 7.99543262e-02
1.52084738e-01 -4.66599524e-01 8.40353251e-01 1.04945205e-01
-1.71913236e-01 2.05067188e-01 -1.02460146e+00 9.08775926e-01
1.14729095e+00 -8.68273841e-04 -1.02840483e-01 -5.74535057e-02
7.88816154e-01 5.18750310e-01 -5.89975119e-01 2.47728452e-01
1.00079191e+00 -9.72872257e-01 8.24202180e-01 -5.32299042e-01
1.50523961e-01 -4.80451852e-01 -2.40223259e-01 -1.24303186e+00
-5.09678960e-01 -5.82970679e-01 5.09813130e-01 9.46286380e-01
6.58854902e-01 -1.23619032e+00 1.09995887e-01 3.74809712e-01
1.43720821e-01 -6.57440960e-01 -9.64610994e-01 -1.02976596e+00
6.39959157e-01 4.60884690e-01 5.85924327e-01 8.17688525e-01
-1.22050248e-01 1.35430053e-01 2.87052244e-02 -6.94544911e-02
3.25914353e-01 -6.74699768e-02 6.17668927e-01 -1.51431978e+00
-4.99452680e-01 -8.88959050e-01 -2.76700616e-01 -1.83315665e-01
-2.23909885e-01 -3.63143265e-01 2.35324368e-01 -1.07037735e+00
1.46120355e-01 -5.89350700e-01 -2.97449946e-01 6.52839020e-02
-3.41711789e-01 -8.29863846e-02 1.49434358e-01 5.62140942e-01
-4.19120900e-02 7.35753775e-01 1.00714540e+00 3.40646327e-01
-8.36968899e-01 -3.32624674e-01 -6.43969417e-01 5.20380616e-01
9.06967461e-01 -4.57407564e-01 -4.00143921e-01 -1.28241152e-01
6.75189435e-01 -3.06925654e-01 -5.46501204e-02 -8.37222993e-01
-3.24250549e-01 -7.75487959e-01 -3.91309299e-02 3.06652039e-01
1.53510794e-01 -6.54868364e-01 4.82471347e-01 1.15155363e+00
-1.94808215e-01 2.17217490e-01 1.81054637e-01 7.04699874e-01
5.05401254e-01 -5.85286207e-02 1.08117270e+00 -1.43861979e-01
-1.54700398e-01 -1.15018673e-01 -1.09227848e+00 -1.60036400e-01
1.10197616e+00 -4.64637548e-01 -3.36407602e-01 -9.96751040e-02
-5.69365442e-01 5.54950163e-02 1.30135643e+00 3.24863136e-01
1.24443352e-01 -7.99092352e-01 -7.90399849e-01 1.80818379e-01
-1.52688712e-01 -7.05452561e-01 -2.79761627e-02 8.37193370e-01
-8.59826505e-01 -7.04865009e-02 -7.81018257e-01 -5.03964722e-01
-1.08147418e+00 5.45915723e-01 7.51527727e-01 1.07139349e-01
-9.11880806e-02 9.69105542e-01 3.99389207e-01 -3.82504821e-01
-2.92793870e-01 2.07642606e-03 -1.10028237e-01 -1.83779281e-02
2.98125058e-01 3.32398266e-01 -4.15517598e-01 -7.02751517e-01
-5.50411284e-01 6.04131222e-01 5.07380962e-01 -2.95700461e-01
1.40853405e+00 -4.58795369e-01 -5.28361678e-01 8.67696524e-01
2.71699965e-01 1.03371873e-01 -9.54258144e-01 2.71363854e-01
-2.98482017e-03 -6.66681468e-01 -5.25498748e-01 -6.97832644e-01
-4.65739071e-01 5.56438446e-01 5.59243917e-01 6.89377785e-01
1.11947405e+00 -2.02842146e-01 1.59921214e-01 3.88967633e-01
3.56310785e-01 -1.04307187e+00 3.01568620e-02 5.31610668e-01
8.73971522e-01 -7.23757029e-01 -2.71166563e-01 -1.49331614e-01
-1.15028054e-01 8.11459541e-01 7.57175148e-01 -3.60119306e-02
7.16799676e-01 3.44649702e-01 -6.76188245e-02 3.42494309e-01
-1.14829934e+00 -4.30899560e-01 -3.22015673e-01 6.58036947e-01
7.24180996e-01 2.37355769e-01 -1.15524566e+00 1.53057009e-01
-3.51926625e-01 -6.10343695e-01 8.67109001e-01 1.02430248e+00
-8.83136749e-01 -1.42713475e+00 -6.09674335e-01 1.19623192e-01
-1.49153873e-01 1.20708719e-02 -1.12696743e+00 5.79909205e-01
4.44729805e-01 8.50383341e-01 -7.55963251e-02 -1.35619357e-01
7.60385254e-03 1.51588872e-01 4.02518868e-01 -4.49969769e-01
-8.06867003e-01 5.21852300e-02 -1.51145220e-01 1.91092975e-02
-2.91375339e-01 -1.19031560e+00 -1.01990449e+00 -2.83075780e-01
-7.09126890e-01 3.40951651e-01 7.30500996e-01 5.53579569e-01
4.87360686e-01 5.20115852e-01 4.42381978e-01 -4.31175172e-01
-3.12064916e-01 -7.59660065e-01 -7.71575272e-01 1.71231598e-01
9.83387381e-02 -6.64670348e-01 -6.91748083e-01 -8.25436600e-03] | [5.6186089515686035, 4.160006999969482] |
1a1af09a-b73b-49ae-9376-a7aeed9215e5 | mutux-at-semeval-2018-task-1-exploring | null | null | https://aclanthology.org/S18-1052 | https://aclanthology.org/S18-1052.pdf | Mutux at SemEval-2018 Task 1: Exploring Impacts of Context Information On Emotion Detection | This paper describes MuTuX, our system that is designed for task 1-5a, emotion classification analysis of tweets on SemEval2018. The system aims at exploring the potential of context information of terms for emotion analysis. A Recurrent Neural Network is adopted to capture the context information of terms in tweets. Only term features and the sequential relations are used in our system. The results submitted ranks 16th out of 35 systems on the task of emotion detection in English-language tweets. | ['Jian-Yun Nie', 'Pan Du'] | 2018-06-01 | null | null | null | semeval-2018-6 | ['product-recommendation'] | ['miscellaneous'] | [-5.76555468e-02 -7.72909373e-02 -1.50346771e-01 -6.70789480e-01
-2.58617282e-01 -2.19297424e-01 6.88464105e-01 4.97725427e-01
-7.50823915e-01 3.47458005e-01 4.59107518e-01 -2.94885159e-01
2.26718441e-01 -4.53001916e-01 -5.61741106e-02 -3.32344055e-01
-5.43316543e-01 -2.27136821e-01 -6.06771588e-01 -8.23773026e-01
2.59675890e-01 2.30595604e-01 -1.75968504e+00 8.43542457e-01
2.62534738e-01 1.71275342e+00 -3.10534298e-01 7.58293867e-01
-8.04785550e-01 1.38536954e+00 -5.78479409e-01 -3.08686405e-01
-3.01657170e-01 -2.83877641e-01 -1.25366211e+00 -4.13860828e-01
-4.07310128e-01 4.68429655e-01 -2.18925595e-01 7.96416461e-01
3.94503444e-01 4.08763081e-01 5.66227973e-01 -1.20434630e+00
-6.21268637e-02 1.21600676e+00 -1.75921321e-02 5.79388022e-01
6.70324624e-01 -6.93836689e-01 1.04621994e+00 -1.33646595e+00
7.55005360e-01 8.87156188e-01 6.49062872e-01 4.89760965e-01
-5.36361754e-01 -6.64924443e-01 2.80830801e-01 5.10590374e-01
-1.48809290e+00 -4.26131189e-01 9.34553683e-01 -2.45540112e-01
1.89151990e+00 4.65277344e-01 7.96740651e-01 1.60222769e+00
4.61016566e-01 8.65607977e-01 1.02131093e+00 -5.02585411e-01
1.56620234e-01 2.39672258e-01 8.75488579e-01 2.79835165e-01
-6.02084100e-01 -4.43041563e-01 -9.50280368e-01 -2.99980342e-01
-3.90857577e-01 -2.88186491e-01 1.40659213e-01 8.75028849e-01
-6.18123531e-01 1.01388431e+00 3.05771142e-01 8.29550624e-01
-8.59102428e-01 -1.51131064e-01 1.24786234e+00 6.56624198e-01
1.22666669e+00 6.00993633e-01 -7.10576057e-01 -6.99196696e-01
-6.26009464e-01 5.64888567e-02 9.70471144e-01 7.73450136e-01
3.38379204e-01 2.68577766e-02 -2.22539723e-01 8.66145730e-01
4.71317470e-02 1.15122110e-01 1.02658701e+00 -3.07514369e-01
3.17151964e-01 8.38945866e-01 -1.52817488e-01 -1.16713989e+00
-9.52199280e-01 -2.52861828e-01 -5.93143404e-01 -6.22459769e-01
-6.37687862e-01 -7.56267250e-01 -7.19849646e-01 1.40121269e+00
1.89747721e-01 1.78887650e-01 3.36702198e-01 5.45127928e-01
1.48572445e+00 1.08349335e+00 4.49096531e-01 -6.12351596e-01
1.60040474e+00 -7.46989548e-01 -1.23519957e+00 -2.89911419e-01
9.10791278e-01 -8.52154374e-01 6.35633528e-01 3.37439597e-01
-8.51202607e-01 -9.40639898e-02 -5.97911179e-01 -3.25845070e-02
-1.28985858e+00 -6.54305667e-02 8.31768095e-01 1.72843605e-01
-9.02486145e-01 1.34576604e-01 -1.65303245e-01 -4.36823905e-01
-1.34046763e-01 1.85272172e-01 -1.72068894e-01 7.27533340e-01
-1.93766737e+00 1.13264358e+00 4.04323846e-01 3.71131986e-01
-1.17640078e-01 -2.95343578e-01 -1.02203429e+00 4.69338596e-02
2.41625085e-01 8.29524100e-02 1.25586820e+00 -1.05516720e+00
-1.60346293e+00 1.00328267e+00 -6.18547857e-01 -5.95350027e-01
-2.28281170e-01 -4.31698859e-01 -1.11875296e+00 -2.13398226e-03
-1.75049543e-01 1.82103470e-01 4.94597852e-01 -6.08100235e-01
-6.25362873e-01 4.31898385e-02 -4.35908169e-01 1.43916413e-01
-6.13962591e-01 8.72629941e-01 -4.38571647e-02 -6.56713486e-01
-3.01167965e-01 -9.13135469e-01 -3.13790381e-01 -1.26474619e+00
-2.92456180e-01 -9.42956924e-01 9.72554505e-01 -4.57223147e-01
1.70225203e+00 -2.15394831e+00 -2.08883435e-01 5.31708121e-01
-2.18399554e-01 1.46033779e-01 -1.85127720e-01 7.51601398e-01
-5.48607886e-01 3.45785648e-01 3.57728064e-01 -4.64578331e-01
3.45155075e-02 3.50740850e-02 -7.50994444e-01 -1.39306352e-01
2.72238225e-01 9.75102603e-01 -7.98634171e-01 -4.27428663e-01
-1.38050944e-01 4.12783921e-01 3.24578136e-02 2.87627846e-01
-8.77839774e-02 -3.93463559e-02 -6.49167120e-01 6.10199988e-01
1.26240626e-01 4.44372334e-02 1.16438076e-01 -8.32758024e-02
-4.43490386e-01 8.35671306e-01 -5.72503686e-01 1.42739403e+00
-8.63890588e-01 9.28397954e-01 -1.56418085e-01 -7.96268821e-01
9.84623432e-01 8.95313740e-01 6.80925310e-01 -8.02813888e-01
8.80288482e-01 -4.29314598e-02 -2.67862886e-01 -9.32589293e-01
1.03111887e+00 -6.87764511e-02 -7.41595387e-01 3.22871715e-01
7.36395121e-02 -4.79798801e-02 2.62572557e-01 1.50071666e-01
9.52704787e-01 -3.59973550e-01 3.70904744e-01 -9.61940438e-02
4.31587249e-01 -1.41291276e-01 2.88454115e-01 5.22278547e-01
-2.30064854e-01 -1.72653142e-02 4.62807328e-01 -6.97403967e-01
-3.07815850e-01 -1.96643025e-01 -1.18808918e-01 1.68251514e+00
-2.73631901e-01 -1.04243147e+00 -3.76032680e-01 -6.06607437e-01
-5.07830143e-01 8.26349378e-01 -1.19298470e+00 -2.31863111e-01
-2.02464968e-01 -6.54941142e-01 7.97642708e-01 2.08249599e-01
2.47385025e-01 -1.54441321e+00 -6.90165341e-01 3.99709523e-01
-7.83005357e-01 -1.35452342e+00 6.93799108e-02 7.60824621e-01
-3.26411635e-01 -5.65923572e-01 4.03390750e-02 -6.45695269e-01
-2.35389434e-02 -2.43754178e-01 1.25279069e+00 -2.25416839e-01
-2.63385832e-01 2.41800368e-01 -9.19034719e-01 -8.81122530e-01
9.45239961e-02 3.21459562e-01 3.67834605e-02 2.03071594e-01
9.02189612e-01 -3.69243711e-01 -2.40558073e-01 -1.74548790e-01
-7.07015991e-01 -3.57886970e-01 2.45066844e-02 4.97647911e-01
1.18488252e-01 5.01135327e-02 5.71087003e-01 -6.52220130e-01
1.33734167e+00 -9.88286018e-01 2.54008591e-01 1.25805691e-01
-4.86187339e-01 -3.38597298e-01 3.49124610e-01 -4.52905238e-01
-1.05075848e+00 -1.37239501e-01 -3.99225265e-01 -1.42735228e-01
-1.85572520e-01 1.15676105e+00 5.17990589e-01 1.34623677e-01
6.21074259e-01 -4.44062836e-02 -6.86688960e-01 -2.41582453e-01
7.69388676e-02 1.02983725e+00 1.46376118e-01 -3.26578170e-01
-3.46213013e-01 3.03226970e-02 -1.97670907e-01 -9.13496375e-01
-1.03430939e+00 -8.30621183e-01 9.37372297e-02 -5.63672483e-01
7.92018831e-01 -9.11897361e-01 -8.67390692e-01 3.65462899e-01
-1.18098068e+00 -3.72390375e-02 -4.05072346e-02 5.29760659e-01
-1.34458736e-01 -4.58184242e-01 -1.04122150e+00 -1.14825320e+00
-8.89614820e-01 -6.43718123e-01 8.52442086e-01 1.36135057e-01
-9.18222904e-01 -8.90041411e-01 2.44862527e-01 -1.24911830e-01
6.61916554e-01 2.89157242e-01 5.28559923e-01 -1.20681548e+00
8.29263926e-01 -4.49063152e-01 1.97144628e-01 5.44053577e-02
-4.92792157e-03 2.26238757e-01 -1.11147332e+00 3.37118208e-01
2.48518109e-01 -8.33086133e-01 8.66927922e-01 1.89656302e-01
1.34444559e+00 -4.34436917e-01 -2.50844032e-01 1.99569073e-02
1.01866531e+00 3.03876400e-01 3.22852224e-01 3.20043415e-01
4.11196709e-01 9.35829401e-01 7.37428665e-01 8.51356328e-01
5.32784522e-01 3.72814178e-01 3.23340863e-01 7.80890509e-02
7.44620740e-01 2.20055237e-01 5.48619032e-01 1.26521099e+00
2.14382377e-03 -4.56435084e-01 -1.30594599e+00 4.97694731e-01
-1.88689780e+00 -8.47763240e-01 -3.80044639e-01 1.07516062e+00
7.24291921e-01 7.64004216e-02 -1.00376315e-01 1.38004161e-02
4.94832575e-01 2.33333305e-01 -2.71178503e-02 -1.39991033e+00
-1.27753928e-01 3.63036811e-01 -1.13801233e-01 3.53494227e-01
-1.33377171e+00 1.22406518e+00 6.91091537e+00 7.08555222e-01
-1.37137592e+00 4.66408096e-02 6.77285552e-01 -4.52001631e-01
8.41113180e-03 -4.31101054e-01 -5.24038315e-01 3.66137236e-01
1.62333834e+00 -1.83224931e-01 1.46587580e-01 9.37292397e-01
2.97242105e-01 -1.57868251e-01 -8.10547352e-01 9.45295453e-01
2.80920982e-01 -9.68048096e-01 -9.40423459e-02 -3.73187125e-01
5.64455807e-01 4.23511595e-01 2.14094758e-01 8.09244156e-01
1.12151071e-01 -1.07031202e+00 5.91441274e-01 6.29622281e-01
2.03967124e-01 -1.25741804e+00 1.24812102e+00 1.44485191e-01
-1.06214213e+00 1.96604878e-02 -1.17765536e-04 -4.67816085e-01
1.40244246e-01 6.34496927e-01 -8.54761422e-01 5.31897426e-01
1.05205631e+00 9.26432729e-01 -3.15246105e-01 2.51975209e-01
-2.62033135e-01 7.87113488e-01 -3.42172116e-01 -5.31456709e-01
6.78462803e-01 1.90313011e-01 5.70814788e-01 2.16448569e+00
6.37738332e-02 3.49979699e-01 5.62239289e-02 1.35670125e-01
-3.64440084e-01 6.80577159e-01 -7.00358927e-01 -2.22574830e-01
2.74598390e-01 1.57986331e+00 -7.57025361e-01 -5.24996161e-01
-4.95619141e-02 6.20128512e-01 4.05521214e-01 3.26000184e-01
-7.32153058e-01 -6.23107195e-01 6.04777634e-01 -8.12708199e-01
1.78182393e-01 1.53942883e-01 -1.88608408e-01 -1.11158371e+00
-1.59330115e-01 -6.71585321e-01 6.03514135e-01 -8.96530330e-01
-1.42751300e+00 1.34087551e+00 -7.04353303e-02 -7.63834417e-01
-6.62766814e-01 -5.55704892e-01 -9.30963635e-01 5.26115775e-01
-1.34811735e+00 -1.00914550e+00 9.58009511e-02 6.71516061e-01
5.57650685e-01 -6.78619370e-04 1.22912443e+00 1.36289895e-01
-6.54734910e-01 3.59718710e-01 -3.85590792e-01 1.68456376e-01
6.64462566e-01 -8.91506612e-01 1.13629602e-01 5.16352594e-01
-2.21707653e-02 6.12352550e-01 8.68695498e-01 -4.85075057e-01
-1.05827653e+00 -9.63974953e-01 1.79208386e+00 -1.48877457e-01
1.10107386e+00 -4.84879553e-01 -6.27379894e-01 5.91364145e-01
7.06052125e-01 -8.85439217e-02 1.20567071e+00 6.50210857e-01
-3.40902656e-01 1.98951483e-01 -7.80554414e-01 5.76036155e-01
6.49744987e-01 -9.35071170e-01 -7.49916792e-01 -1.80604886e-02
8.44579518e-01 -3.12197030e-01 -1.02217472e+00 5.25947213e-01
5.08102298e-01 -2.74128735e-01 5.56876838e-01 -8.92961919e-01
7.04742134e-01 3.30605745e-01 -2.70839781e-01 -1.35956609e+00
8.42297673e-02 -6.68303072e-01 -2.79994011e-01 1.35644591e+00
7.82878518e-01 -4.47229445e-01 1.60692990e-01 6.42845452e-01
3.70617695e-02 -1.06249571e+00 -9.72834408e-01 -2.60694742e-01
-8.22618157e-02 -1.07347310e+00 4.55225199e-01 1.41598213e+00
7.97833085e-01 6.85033083e-01 -3.17415208e-01 -3.27409297e-01
-3.52482945e-01 1.38267145e-01 1.27795771e-01 -8.99551094e-01
5.08075118e-01 -7.36614883e-01 -1.92548275e-01 -1.66377321e-01
7.42463827e-01 -7.72255123e-01 1.12865545e-01 -1.13401604e+00
-2.24499553e-01 1.54023811e-01 -7.60973692e-01 6.49596870e-01
5.38017191e-02 1.67973936e-01 4.54314537e-02 -2.82096028e-01
-9.52794850e-01 7.01304436e-01 3.06656510e-01 -6.48805872e-02
-3.07729393e-01 -1.15185902e-01 -4.47631717e-01 8.61814737e-01
1.17288506e+00 -5.59770882e-01 -1.09297939e-01 4.39710990e-02
9.96460140e-01 -2.37481043e-01 -7.35370889e-02 -4.18261051e-01
2.47452378e-01 -5.36855645e-02 1.24028243e-01 -7.60408521e-01
4.49041694e-01 -8.77170920e-01 -1.61134213e-01 -6.23411536e-02
-7.63217330e-01 3.63044590e-01 4.38954622e-01 5.57780266e-02
-6.82256579e-01 -2.33957887e-01 5.65640181e-02 1.39141092e-02
-7.41327822e-01 7.81600550e-02 -1.29846466e+00 7.56112561e-02
6.88932955e-01 4.15329814e-01 -1.76910255e-02 -5.70682168e-01
-1.03551674e+00 2.88599461e-01 -4.35723126e-01 8.88964951e-01
6.88428581e-01 -1.32947826e+00 -6.48430407e-01 -3.49468529e-01
4.45358008e-01 -5.43905795e-01 1.84098318e-01 6.53682888e-01
2.90929288e-01 3.75000179e-01 1.79404050e-01 -8.19958597e-02
-1.48886812e+00 3.42265278e-01 2.66434044e-01 -6.35303557e-01
-2.18537256e-01 9.84863281e-01 -4.72422600e-01 -1.83001488e-01
2.52098352e-01 -4.54504311e-01 -1.01338112e+00 7.53297448e-01
6.38978004e-01 6.81239367e-02 3.26588959e-01 -8.71246517e-01
-6.54730082e-01 2.05121741e-01 -2.22946495e-01 -2.64204264e-01
1.40635824e+00 -3.62393498e-01 -4.72617716e-01 1.13847625e+00
1.51933587e+00 -2.33130246e-01 1.89731829e-02 -3.11557084e-01
5.91649771e-01 3.72123867e-01 3.99597794e-01 -9.70183730e-01
-6.91296339e-01 5.46122491e-01 3.22553873e-01 7.23048329e-01
1.16202307e+00 1.49867745e-04 7.65596390e-01 7.62766182e-01
-9.14985035e-03 -1.61165011e+00 -2.06371561e-01 1.56157887e+00
9.76497233e-01 -1.25093639e+00 -3.30270082e-01 3.95210013e-02
-8.59339774e-01 1.24999440e+00 4.81168538e-01 1.81973092e-02
1.07221997e+00 2.96421289e-01 4.07951444e-01 -6.85245693e-01
-1.31641209e+00 -3.09198976e-01 3.86667490e-01 -2.76111066e-01
7.95124769e-01 1.19026013e-01 -5.83324611e-01 9.77967918e-01
-5.80671608e-01 -1.88369751e-01 2.25847095e-01 1.24052930e+00
-1.16323724e-01 -7.11596310e-01 -8.08143616e-02 2.94454247e-01
-1.13780367e+00 -3.74577522e-01 -9.28217590e-01 4.69437957e-01
1.41517356e-01 1.40534270e+00 1.67003959e-01 -6.76783562e-01
3.49882185e-01 7.49483585e-01 -3.81814808e-01 -2.25735739e-01
-1.41666877e+00 -1.42949149e-01 7.12297678e-01 -7.90396392e-01
-8.03114712e-01 -5.29806912e-01 -1.35621798e+00 -4.87953387e-02
-2.89837658e-01 7.83627391e-01 1.10733700e+00 1.15736115e+00
6.15415156e-01 6.87392056e-01 1.08195317e+00 -6.56790674e-01
5.12377501e-01 -1.31385386e+00 -3.36304009e-01 3.47226381e-01
3.49780291e-01 -3.69660974e-01 -4.14469779e-01 -1.74236506e-01] | [12.913999557495117, 6.203357696533203] |
66077df1-de3a-4247-bcee-2812681278fa | extract-select-and-rewrite-a-new-modular | null | null | https://openreview.net/forum?id=fg1zmL4XRu2 | https://openreview.net/pdf?id=fg1zmL4XRu2 | Extract, Select and Rewrite: A New Modular Summarization Method | Prior works on supervised summarization are mainly based on end-to-end models, leading to low modularity, unfaithfulness and low interpretability. To address this, we propose a new three-phase modular abstractive sentence summarization method.We split up the summarization problem explicitly into three stages, namely knowledge extraction, content selection and rewriting.We utilize multiple knowledge extractors to obtain relation triples from the text, learn a fine-tuned classifier to select content to be included in the summary and use a fine-tuned BART rewriter to rewrite the selected triples into a natural language summary.We find our model shows good modularity as the modules can be trained separately and on different datasets. The automatic and human evaluations demonstrate that our new method is competitive with state-of-the-art methods and more faithful than end-to-end baseline models. | ['Anonymous'] | 2022-01-16 | null | null | null | acl-arr-january-2022-1 | ['abstractive-sentence-summarization'] | ['natural-language-processing'] | [ 4.01276022e-01 7.52877355e-01 -4.86974925e-01 -2.57446259e-01
-1.04843366e+00 -7.48102307e-01 6.17623627e-01 4.35635716e-01
-3.12087119e-01 8.41389179e-01 9.50493097e-01 5.06016314e-02
2.27343515e-01 -5.28675795e-01 -7.51601636e-01 8.84319395e-02
2.27993056e-01 5.46121001e-01 2.10156471e-01 -3.45926523e-01
3.99369597e-01 -8.78614262e-02 -1.02997530e+00 6.43595815e-01
1.42925131e+00 4.62030560e-01 3.29567380e-02 8.69819701e-01
2.08542552e-02 1.02074814e+00 -7.21144855e-01 -8.97245944e-01
-1.28865063e-01 -7.49506295e-01 -1.43425119e+00 6.50040433e-02
4.59137887e-01 -2.03717366e-01 -1.96668282e-01 9.34809506e-01
4.83238429e-01 -6.01290818e-03 8.61053407e-01 -6.45839572e-01
-6.53270781e-01 1.28634632e+00 -4.63648319e-01 -1.70662794e-02
6.28898382e-01 -2.89174058e-02 1.57702565e+00 -9.10298049e-01
8.42356324e-01 1.21718335e+00 6.20021641e-01 4.55246419e-01
-1.38246477e+00 -1.79210771e-02 2.75988191e-01 1.83657810e-01
-9.09590006e-01 -8.71526659e-01 6.53569400e-01 -1.46270528e-01
1.49446678e+00 5.53490579e-01 6.67324543e-01 9.60707486e-01
3.60600770e-01 9.06595290e-01 3.62890393e-01 -5.27478993e-01
1.42735645e-01 -1.07891105e-01 2.88849056e-01 9.21785533e-01
6.05461955e-01 -7.71495461e-01 -7.03759193e-01 -8.57848823e-02
1.48185462e-01 -4.63189662e-01 -2.54146457e-01 -1.38290316e-01
-1.33368146e+00 6.32461190e-01 3.81289184e-01 1.14620060e-01
-5.01516402e-01 6.85922876e-02 6.90745950e-01 3.09961200e-01
6.14733338e-01 8.78897965e-01 -4.43064749e-01 -1.38043761e-01
-1.17815435e+00 3.21167022e-01 1.22891593e+00 1.00876379e+00
7.03679979e-01 -2.84165621e-01 -6.83438957e-01 8.79056513e-01
-1.44408988e-02 1.98956043e-01 4.54471886e-01 -1.00806558e+00
9.07253921e-01 9.55356300e-01 -1.88766181e-01 -8.67180526e-01
-2.90241629e-01 -4.70828593e-01 -9.53213573e-01 -4.67265576e-01
-3.37973386e-01 -2.10237041e-01 -5.88738441e-01 1.56900489e+00
2.79090218e-02 -3.77307743e-01 3.16391766e-01 4.55131739e-01
1.26177156e+00 4.96577233e-01 -1.08494915e-01 -3.52737963e-01
1.50686729e+00 -1.48265290e+00 -5.38738608e-01 -5.83210647e-01
7.80864000e-01 -5.24168611e-01 9.00532603e-01 1.83222696e-01
-1.58472347e+00 -2.41156012e-01 -1.33891439e+00 -5.82638085e-01
1.09944128e-01 4.37526762e-01 2.65514165e-01 2.10299283e-01
-9.32118893e-01 7.95486689e-01 -8.22754800e-01 -4.52449471e-01
5.17522216e-01 2.49079943e-01 -4.35920894e-01 1.78049311e-01
-9.86745954e-01 1.22117579e+00 6.95394039e-01 -2.59551078e-01
-6.17104590e-01 -4.90446001e-01 -1.02855277e+00 2.69829541e-01
5.56348622e-01 -1.72420824e+00 1.59207273e+00 -8.75170529e-01
-1.79659092e+00 7.96009779e-01 -4.43177104e-01 -5.84774554e-01
3.51290047e-01 -5.06899297e-01 2.48301834e-01 3.82339746e-01
3.17547083e-01 4.41140920e-01 5.57231843e-01 -1.00372529e+00
-5.20884454e-01 3.99176450e-03 1.78265810e-01 4.74558026e-01
-2.99094558e-01 5.36325574e-02 -5.05254269e-01 -5.76719701e-01
-1.10875353e-01 -6.39961660e-01 -1.58566073e-01 -8.11195672e-01
-1.05358672e+00 -2.11365581e-01 1.05516180e-01 -9.56040561e-01
1.70426548e+00 -1.50919795e+00 8.13061357e-01 -3.35998923e-01
3.32716048e-01 3.30890596e-01 -1.41566336e-01 8.33268702e-01
7.47538432e-02 3.85623693e-01 -4.17951137e-01 -7.20764816e-01
3.76049206e-02 -1.38706833e-01 -2.66346514e-01 6.60599694e-02
4.57320571e-01 1.25431287e+00 -1.21397388e+00 -8.29114556e-01
-2.55935043e-01 -1.49976760e-01 -7.10186839e-01 3.10161233e-01
-4.72740710e-01 1.94350824e-01 -3.24231088e-01 3.47794592e-01
2.35655993e-01 -2.26735383e-01 2.79392540e-01 -3.80483180e-01
1.20185614e-01 8.95808280e-01 -7.15428114e-01 2.00951338e+00
-3.98717254e-01 4.81291473e-01 -2.27569401e-01 -8.59182835e-01
5.47489703e-01 2.24670321e-01 -1.09818988e-01 -3.41016233e-01
7.88588002e-02 1.12679638e-01 -1.89335898e-01 -6.53661847e-01
1.02781904e+00 -7.87843615e-02 -3.53378743e-01 5.98790407e-01
4.59302992e-01 -3.27078164e-01 6.38668537e-01 8.27301741e-01
1.31986856e+00 2.28414655e-01 8.50597978e-01 -1.24869235e-01
5.34941137e-01 5.75787947e-02 6.31121933e-01 6.96639597e-01
3.91879737e-01 7.14367867e-01 9.44868624e-01 5.86889721e-02
-9.40403521e-01 -8.58483732e-01 5.95567882e-01 9.74141479e-01
-2.95183398e-02 -1.15456903e+00 -1.01077509e+00 -8.98613632e-01
-3.46523255e-01 1.11868560e+00 -3.87782127e-01 -4.78279680e-01
-7.10347474e-01 -3.57917964e-01 7.29655385e-01 3.98251444e-01
5.58109343e-01 -8.68543327e-01 -5.20214021e-01 1.86773121e-01
-7.72922695e-01 -1.07035160e+00 -7.27069080e-01 -1.25342444e-01
-1.01780248e+00 -8.70865226e-01 -2.16906786e-01 -8.48794639e-01
5.55133402e-01 9.16141421e-02 1.54676723e+00 -8.69177431e-02
1.54052824e-01 1.85645431e-01 -3.97188991e-01 -3.89823169e-01
-7.36438274e-01 9.36480343e-01 -2.15787739e-01 -2.24944770e-01
-1.58438355e-01 -6.88756227e-01 -5.57601988e-01 -2.80339867e-01
-7.85669029e-01 5.34159362e-01 8.66474271e-01 7.28047848e-01
2.64215827e-01 -4.12072748e-01 8.61702979e-01 -1.20894325e+00
1.04392743e+00 -2.52232105e-01 1.15448236e-01 6.49044931e-01
-5.24440587e-01 4.34865505e-01 8.14054906e-01 -7.55880699e-02
-1.16179538e+00 -6.76378384e-02 1.37095794e-01 2.01057777e-01
2.29683191e-01 8.70839059e-01 -2.55049884e-01 5.27941346e-01
9.41539049e-01 2.28338584e-01 -1.58012137e-01 -3.48944843e-01
8.51583064e-01 6.32093966e-01 6.92052722e-01 -5.33047557e-01
6.44797146e-01 1.49839193e-01 -3.47689331e-01 -8.42202961e-01
-1.26447940e+00 -1.86260164e-01 -7.59754300e-01 1.21503830e-01
5.99897206e-01 -1.07410741e+00 -2.93970257e-01 1.11268379e-01
-1.51261997e+00 -7.22842067e-02 -5.18727839e-01 7.57039636e-02
-6.33275092e-01 8.52314949e-01 -7.04267502e-01 -4.48880613e-01
-1.29116106e+00 -7.48693347e-01 1.18519914e+00 3.15522432e-01
-8.45150292e-01 -9.36989129e-01 3.26647460e-01 5.86623788e-01
1.57585099e-01 1.46733522e-01 1.03659022e+00 -8.15219700e-01
-3.58144045e-01 -1.90613061e-01 -8.45336020e-02 4.19202179e-01
7.80025572e-02 2.31310487e-01 -4.57809448e-01 -3.21774818e-02
-1.57647371e-01 -5.54108143e-01 1.34499216e+00 8.03129598e-02
6.07778430e-01 -8.44551206e-01 -2.65899360e-01 4.15863127e-01
9.77259338e-01 -7.07825601e-01 7.93094575e-01 3.35900217e-01
7.34491169e-01 5.46263218e-01 4.52541023e-01 3.33327800e-01
8.77156198e-01 4.06492561e-01 1.60991713e-01 1.75023049e-01
-2.71938622e-01 -5.92450798e-01 7.34599173e-01 1.26652622e+00
1.08682588e-02 -4.01813388e-01 -5.74308693e-01 6.72752440e-01
-2.33933210e+00 -1.10922408e+00 -1.38149383e-02 1.85229576e+00
1.38622904e+00 2.57332534e-01 2.66197324e-01 -1.65358871e-01
5.15692472e-01 3.39226812e-01 -2.19546109e-01 -6.32419765e-01
-1.68117657e-01 2.18291283e-01 1.92806914e-01 5.48764586e-01
-9.03201222e-01 1.33026659e+00 6.12001467e+00 7.58772671e-01
-8.38334799e-01 -1.77705078e-03 4.08598810e-01 -3.86476099e-01
-5.13701797e-01 4.51761574e-01 -5.92011213e-01 1.02274463e-01
8.86824608e-01 -5.22627831e-01 1.65341094e-01 4.12563026e-01
2.99393296e-01 -1.98062479e-01 -1.42345130e+00 5.98432183e-01
4.03848559e-01 -1.59471822e+00 5.23198247e-01 -4.65404302e-01
8.04490089e-01 -7.01452494e-02 -5.82972169e-01 3.50330561e-01
4.20996875e-01 -9.69262898e-01 8.40966880e-01 6.98615730e-01
5.78587532e-01 -6.22539341e-01 5.53984284e-01 5.68608642e-01
-9.16635811e-01 1.52612627e-01 -2.80393243e-01 -2.19871998e-01
2.59503990e-01 5.87889135e-01 -9.23472106e-01 1.11814308e+00
2.09235605e-02 9.04451787e-01 -9.17753577e-01 7.61019468e-01
-8.42719138e-01 6.04533315e-01 -3.84707935e-02 -2.11353719e-01
-9.64443758e-02 -1.47242695e-01 8.62393618e-01 1.59828591e+00
9.33522452e-03 -7.41515458e-02 1.76079303e-01 7.63511062e-01
-5.92539251e-01 2.01841474e-01 -4.17120963e-01 -1.15826972e-01
3.31745297e-01 1.44741476e+00 -3.48648012e-01 -7.18879640e-01
-1.61904946e-01 1.33670056e+00 8.78604412e-01 2.10530415e-01
-5.97415209e-01 -6.87163234e-01 6.49695396e-02 -8.99005905e-02
4.27038491e-01 -9.42376703e-02 -6.16846502e-01 -1.74574935e+00
3.35729450e-01 -9.82098043e-01 3.17962497e-01 -6.01947606e-01
-9.86441195e-01 5.21161139e-01 7.24642947e-02 -9.69234467e-01
-4.00230885e-01 1.61958069e-01 -1.09784210e+00 5.30211687e-01
-1.29990113e+00 -1.37347758e+00 4.51596044e-02 1.82358176e-02
6.94986403e-01 6.20785207e-02 8.27227235e-01 -1.31503955e-01
-9.07410860e-01 6.95191741e-01 -1.41868323e-01 1.52956128e-01
8.23991001e-01 -1.41882479e+00 6.83720291e-01 1.18062687e+00
1.82602219e-02 7.93880463e-01 8.51023972e-01 -7.38695562e-01
-1.34451699e+00 -1.21110070e+00 1.36797857e+00 -5.16281784e-01
5.19991219e-01 -2.64125705e-01 -6.57390475e-01 8.15493941e-01
7.65744448e-01 -7.02782869e-01 6.15401626e-01 2.70738095e-01
-4.71047550e-01 -6.22295178e-02 -7.57509947e-01 8.64340484e-01
1.19797730e+00 -3.10496300e-01 -1.33831835e+00 3.10771137e-01
9.39967930e-01 -3.31034899e-01 -6.63083971e-01 2.74190754e-01
3.14482719e-01 -7.61090934e-01 6.46312416e-01 -8.40005040e-01
1.10354042e+00 -1.05533861e-01 2.98643500e-01 -1.67101634e+00
-4.70750481e-01 -9.33487833e-01 -4.26968724e-01 1.63939178e+00
8.05673838e-01 -4.12308693e-01 4.73846793e-01 3.81533891e-01
-4.56399828e-01 -8.47014666e-01 -7.19027638e-01 -5.55379808e-01
1.08894348e-01 2.49377787e-01 1.90271080e-01 4.76576447e-01
9.03481960e-01 1.55424786e+00 -2.93353915e-01 -2.29527026e-01
4.85845417e-01 3.88660371e-01 1.00361288e+00 -1.03594851e+00
-4.03462350e-01 -6.71489060e-01 3.81585881e-02 -1.15400219e+00
2.81945229e-01 -1.31065834e+00 6.52238876e-02 -2.23570013e+00
7.88457453e-01 4.27095890e-01 2.26923004e-01 5.49221694e-01
-6.47581935e-01 -1.79028183e-01 1.87868744e-01 3.10071558e-01
-1.27799582e+00 8.74070525e-01 9.80152488e-01 -3.44721794e-01
-3.70437413e-01 -1.44886196e-01 -1.35071445e+00 6.47775114e-01
7.35202789e-01 -3.65532964e-01 -2.29114622e-01 -5.57612896e-01
5.71230412e-01 8.38721618e-02 -2.28213090e-02 -7.87680149e-01
3.89196157e-01 -2.95637194e-02 2.50423811e-02 -4.57656473e-01
3.49583551e-02 -4.07381430e-02 -1.82373703e-01 3.27170938e-01
-6.43672287e-01 -3.23344320e-02 1.45218456e-02 3.78876090e-01
-1.78526461e-01 -4.39230442e-01 5.84280312e-01 -2.80097872e-01
-5.87008670e-02 -7.19941705e-02 -2.89552838e-01 5.26795387e-01
6.60684645e-01 8.42455868e-03 -6.63795531e-01 -4.82798040e-01
-3.97504270e-01 4.54036951e-01 6.09640062e-01 2.18404561e-01
4.43288386e-01 -1.03302085e+00 -1.18376243e+00 -4.33315903e-01
1.48644581e-01 2.83672571e-01 2.63372101e-02 9.10358727e-01
-6.52783632e-01 3.89123827e-01 1.10901497e-01 -2.71365076e-01
-1.29035652e+00 2.92381912e-01 1.46563947e-01 -8.68467569e-01
-5.32421291e-01 6.25428200e-01 -1.40790597e-01 -3.29804510e-01
1.97496098e-02 -5.24654031e-01 -2.15729713e-01 2.37855420e-01
4.57131743e-01 5.01761675e-01 1.93030789e-01 -4.02321219e-01
-3.85874271e-01 4.19270098e-01 -3.64674777e-01 -1.83788285e-01
1.43871427e+00 -8.88807178e-02 -5.19531012e-01 3.52626294e-01
1.01506507e+00 2.37673998e-01 -7.17043698e-01 -3.08460116e-01
1.52647167e-01 1.40211612e-01 -3.59057456e-01 -8.63854825e-01
-5.33204019e-01 5.52924514e-01 -5.61105907e-01 2.64385670e-01
9.61293936e-01 1.04585670e-01 8.89947712e-01 8.25050414e-01
-1.35164484e-01 -1.23845780e+00 2.40019187e-01 8.66909266e-01
1.15810335e+00 -9.31671679e-01 6.06132686e-01 -5.40896356e-01
-9.26585972e-01 1.00923967e+00 4.08354938e-01 -3.11015517e-01
-3.63829508e-02 -6.32283762e-02 -3.86498421e-01 -7.80542865e-02
-1.25710750e+00 -9.49984118e-02 4.49380219e-01 2.35328063e-01
6.14726186e-01 7.66811101e-03 -6.58644557e-01 9.30391133e-01
-5.40181816e-01 -2.16284934e-02 7.64059722e-01 8.33319783e-01
-6.06250882e-01 -1.12161887e+00 1.10986710e-01 6.03257716e-01
-5.36534190e-01 -2.48823002e-01 -1.05054820e+00 2.18158007e-01
-4.27312106e-01 1.17399383e+00 -3.91224205e-01 -4.56577152e-01
6.81407571e-01 2.10598096e-01 7.44001925e-01 -1.16645432e+00
-1.18290675e+00 -2.16870293e-01 8.64924550e-01 -3.64810884e-01
-2.59711146e-01 -5.80740333e-01 -1.37233436e+00 -2.81538337e-01
-3.74300539e-01 1.69226125e-01 2.24349543e-01 1.08186841e+00
9.61632013e-01 6.78881407e-01 5.11617541e-01 -8.99116814e-01
-8.00457239e-01 -1.26359105e+00 8.04256573e-02 4.45997179e-01
2.18364477e-01 3.78797799e-02 -1.24512173e-01 2.47158632e-01] | [12.44015121459961, 9.44245433807373] |
9d43878a-cc2b-4c36-b53e-60c3336f1bf2 | chatbot-with-a-discourse-structure-driven | null | null | https://aclanthology.org/E17-3022 | https://aclanthology.org/E17-3022.pdf | Chatbot with a Discourse Structure-Driven Dialogue Management | We build a chat bot with iterative content exploration that leads a user through a personalized knowledge acquisition session. The chat bot is designed as an automated customer support or product recommendation agent assisting a user in learning product features, product usability, suitability, troubleshooting and other related tasks. To control the user navigation through content, we extend the notion of a linguistic discourse tree (DT) towards a set of documents with multiple sections covering a topic. For a given paragraph, a DT is built by DT parsers. We then combine DTs for the paragraphs of documents to form what we call extended DT, which is a basis for interactive content exploration facilitated by the chat bot. To provide cohesive answers, we use a measure of rhetoric agreement between a question and an answer by tree kernel learning of their DTs. | ['Boris Galitsky', 'Dmitry Ilvovsky'] | 2017-04-01 | null | null | null | eacl-2017-4 | ['product-recommendation'] | ['miscellaneous'] | [ 6.93946257e-02 8.00031900e-01 -3.32720220e-01 -3.45673919e-01
-5.27184486e-01 -1.07912278e+00 6.00748062e-01 5.20111382e-01
1.49705768e-01 3.58472615e-01 5.66258907e-01 -6.46454632e-01
-1.14581995e-01 -7.99375713e-01 -1.13579467e-01 -3.12429398e-01
2.64467776e-01 6.52487397e-01 4.05983955e-01 -3.21743876e-01
4.97720718e-01 5.15411841e-04 -1.19158745e+00 8.72942030e-01
9.44944441e-01 7.88609743e-01 5.86208284e-01 5.70305228e-01
-9.75331724e-01 1.36430168e+00 -7.38259792e-01 -5.84244311e-01
-3.25518489e-01 -5.67797124e-01 -1.60824978e+00 6.98477626e-02
-4.39864546e-01 1.33013465e-02 5.37792802e-01 9.11914229e-01
-1.49178177e-01 -2.07439680e-02 3.77955735e-01 -1.06529248e+00
-5.58075786e-01 1.51554954e+00 -7.54086897e-02 -1.18663833e-01
8.04154992e-01 -1.08789571e-01 1.27929044e+00 -4.69386667e-01
1.15949631e+00 1.49101853e+00 4.34978515e-01 4.07928944e-01
-1.19006121e+00 -8.59446228e-02 2.62707740e-01 4.83960599e-01
-6.39915347e-01 6.15766132e-03 1.03791225e+00 -9.66195047e-01
8.90332878e-01 3.18163544e-01 6.46693587e-01 8.00222337e-01
7.36634359e-02 8.86316359e-01 7.67163813e-01 -6.63832426e-01
4.47444558e-01 8.45686674e-01 6.13084555e-01 6.32319212e-01
-4.02218729e-01 -7.23063231e-01 -4.91495490e-01 -2.86047071e-01
3.36183012e-01 -6.21821523e-01 2.15511858e-01 -3.98301691e-01
-6.54956162e-01 1.24192643e+00 1.62949815e-01 4.44225103e-01
-3.37978959e-01 -3.87082845e-01 6.55286670e-01 5.75787485e-01
1.93309292e-01 7.91019976e-01 -4.66928840e-01 -4.78866935e-01
-1.22212403e-01 2.42742807e-01 1.60314751e+00 9.04363990e-01
7.77003646e-01 -6.12554491e-01 -3.21156889e-01 7.72989810e-01
2.79296100e-01 -1.86424837e-01 2.85898298e-01 -1.17929649e+00
2.79585719e-01 1.22670543e+00 2.27556050e-01 -7.61595368e-01
-2.55694360e-01 -1.26334295e-01 -4.88172807e-02 2.06227928e-01
2.92951673e-01 -3.37963551e-01 -2.13573158e-01 1.37623167e+00
6.61522865e-01 -8.81341100e-01 2.84698252e-02 3.94186676e-01
9.68725681e-01 4.49850082e-01 3.45080376e-01 -3.92188549e-01
1.84419441e+00 -9.79497194e-01 -1.21498466e+00 -1.55346366e-02
1.11433828e+00 -8.59882057e-01 1.14206362e+00 6.52239978e-01
-8.68563354e-01 -4.67694193e-01 -7.80915260e-01 -3.01835656e-01
-6.02731168e-01 8.66285115e-02 2.65000880e-01 3.41599375e-01
-4.54745591e-01 6.70886457e-01 -3.43235880e-01 -4.20478672e-01
2.81455427e-01 2.70081628e-02 2.64851928e-01 4.56910104e-01
-1.24389541e+00 1.19342124e+00 3.64527792e-01 -3.54429811e-01
-4.87140626e-01 -6.58393741e-01 -7.78658807e-01 3.38450037e-02
8.12799096e-01 -4.58743662e-01 1.87671828e+00 -8.27786505e-01
-1.94899464e+00 7.57570744e-01 8.57359543e-02 -2.54589796e-01
1.67170167e-01 -4.56577569e-01 -1.96031883e-01 -4.52023596e-02
1.28641620e-01 3.21588308e-01 8.08761299e-01 -1.06594145e+00
-6.75076544e-01 -1.90828934e-01 4.09763128e-01 1.50367588e-01
-2.23358572e-01 4.27206844e-01 -3.47432584e-01 -2.77506918e-01
-2.88741827e-01 -6.99625731e-01 -7.02955946e-02 -5.22380590e-01
-9.18585479e-01 -8.78851533e-01 1.19130588e+00 -8.44599009e-01
1.61578071e+00 -1.85351872e+00 3.06415617e-01 2.25268558e-01
3.53156894e-01 1.95125833e-01 -1.68222785e-02 8.89643550e-01
2.25079492e-01 3.24880898e-01 3.01140606e-01 -4.69343998e-02
1.36677772e-01 3.77931148e-02 -7.66795054e-02 -3.57013345e-01
9.02958661e-02 7.97122836e-01 -9.29140687e-01 -7.21675575e-01
1.03798330e-01 1.94930974e-02 -5.61982214e-01 4.77255583e-01
-1.13174248e+00 5.82392633e-01 -9.89120305e-01 1.30906120e-01
1.81652933e-01 -4.26722944e-01 5.34236491e-01 -1.48758277e-01
-4.71691400e-01 8.91000807e-01 -7.10903227e-01 1.48166239e+00
-5.94516158e-01 7.23125577e-01 1.03068441e-01 -6.27311766e-01
1.22041106e+00 1.75978646e-01 1.29864156e-01 -4.26391870e-01
4.75206047e-01 -2.75710881e-01 -1.04060650e-01 -1.04865134e+00
6.08913541e-01 3.05506468e-01 -2.57018059e-01 9.19949830e-01
6.66727964e-03 8.40550289e-02 2.84846276e-01 4.25243258e-01
1.11514926e+00 3.47801834e-01 5.80488086e-01 -1.84905574e-01
5.63033521e-01 3.52413237e-01 -2.68265177e-02 5.87671161e-01
2.49183103e-01 -2.72628546e-01 1.12883818e+00 -3.12143236e-01
-9.36155021e-01 -5.84835052e-01 4.95979898e-02 1.44454610e+00
-2.99680352e-01 -1.08278954e+00 -1.20972049e+00 -1.00299942e+00
-2.03600541e-01 1.04052138e+00 -6.36966050e-01 2.75169134e-01
-5.98845303e-01 3.06964666e-01 3.33167799e-02 1.45298243e-01
4.84052539e-01 -1.57839286e+00 -9.30573702e-01 4.06552136e-01
-5.67556679e-01 -1.08542502e+00 -2.70962954e-01 3.76538813e-01
-5.54338157e-01 -1.09040177e+00 -1.42561093e-01 -1.03041625e+00
2.13148311e-01 -2.52804846e-01 1.09219873e+00 7.89753050e-02
-8.10138211e-02 2.28881657e-01 -8.51561964e-01 -3.75842690e-01
-1.28961444e+00 3.52148384e-01 -5.20252645e-01 -1.79254398e-01
3.55279356e-01 -5.45390368e-01 -3.12599450e-01 4.34035748e-01
-5.17470360e-01 3.11869949e-01 8.08212906e-03 4.21439111e-01
1.59229517e-01 7.61113465e-02 4.30112809e-01 -1.24498546e+00
1.16700077e+00 -5.40581524e-01 -6.13495708e-01 5.69743693e-01
-4.18347180e-01 3.94890487e-01 5.74714422e-01 -6.10827148e-01
-1.53442550e+00 6.68144599e-02 -7.53720999e-02 3.18768412e-01
-2.16685414e-01 8.86716902e-01 -8.92381370e-02 3.39231223e-01
9.91860986e-01 -3.78205448e-01 -1.22241348e-01 -9.41703677e-01
8.78639519e-01 9.73876953e-01 3.31278980e-01 -4.06965315e-01
3.77354681e-01 -1.58128768e-01 -4.94045824e-01 -8.15727651e-01
-1.09025311e+00 -4.42140430e-01 -8.39544952e-01 -4.19019341e-01
1.01286840e+00 -9.07645375e-02 -1.41795647e+00 -2.55358636e-01
-1.48688114e+00 -4.63728696e-01 -4.84906048e-01 -1.13886200e-01
-6.24354959e-01 3.87581915e-01 -5.22391319e-01 -1.02890873e+00
-3.47720742e-01 -7.85467446e-01 5.84834635e-01 2.79726893e-01
-1.10766137e+00 -1.06041574e+00 2.18173027e-01 4.14678305e-01
2.16555029e-01 2.77583361e-01 1.57841480e+00 -1.10293436e+00
-3.66867036e-01 -2.89595798e-02 1.07211374e-01 1.27861202e-01
2.85763502e-01 2.48859427e-03 -8.41180980e-01 3.81661087e-01
7.58354217e-02 -4.16992933e-01 1.86731711e-01 2.31454447e-02
8.48383665e-01 -9.51185942e-01 -5.48932970e-01 -1.70156628e-01
7.66435146e-01 7.40976751e-01 5.11747479e-01 6.59659386e-01
4.44505364e-01 1.24822342e+00 7.32350826e-01 5.31479239e-01
2.41177201e-01 8.83790910e-01 1.51239544e-01 4.78123605e-01
6.86745485e-03 -5.47444403e-01 2.25669935e-01 5.71286798e-01
3.86917263e-01 -2.34504610e-01 -7.18140662e-01 2.72396028e-01
-2.00861001e+00 -6.93716407e-01 -1.76629573e-01 1.65150261e+00
1.10969365e+00 3.26471299e-01 4.58564907e-01 -2.88507659e-02
6.67671502e-01 -1.94580287e-01 -4.40378636e-01 -9.08644974e-01
4.46957827e-01 -8.41886550e-02 -1.86270624e-01 7.33428121e-01
-6.61205709e-01 1.17697573e+00 6.14201212e+00 7.49068379e-01
-5.12313545e-01 1.06726572e-01 2.53371954e-01 4.91018951e-01
-4.06242192e-01 1.64905354e-01 -8.61763716e-01 8.43622684e-02
8.61934066e-01 -3.25122267e-01 3.12371463e-01 1.20364690e+00
4.03463155e-01 -5.15655041e-01 -1.25256205e+00 3.98106694e-01
-3.49832296e-01 -1.73377073e+00 -1.11530304e-01 -1.45886272e-01
3.40494186e-01 -5.34010172e-01 -2.17150882e-01 2.54222661e-01
8.48586202e-01 -6.72812879e-01 6.46498680e-01 3.58208001e-01
2.96364427e-01 -4.31049228e-01 3.30766588e-01 6.06797636e-01
-1.04113388e+00 -2.74271548e-01 1.19401157e-01 1.13040190e-02
4.67912257e-01 1.76230893e-01 -1.45475984e+00 2.71090716e-01
6.62003756e-01 3.76949310e-01 -4.26193923e-01 8.09248388e-01
-5.29309571e-01 6.11682892e-01 2.74280347e-02 -6.16232872e-01
6.50309026e-02 -2.45683685e-01 6.96697831e-01 1.36442459e+00
-8.09108242e-02 2.24364132e-01 1.81166232e-01 1.21150637e+00
8.43399540e-02 3.37769419e-01 -3.82126600e-01 -2.58012772e-01
5.83768368e-01 1.52405119e+00 -8.49795997e-01 -3.43334347e-01
5.04076295e-02 8.20427716e-01 3.56483161e-01 2.14961797e-01
-2.44346246e-01 -6.10666156e-01 5.24453998e-01 7.14917257e-02
5.54521799e-01 3.29037500e-03 -5.13662815e-01 -4.13446963e-01
-1.33725017e-01 -9.47468519e-01 3.98022085e-01 -1.01249063e+00
-9.57704663e-01 8.45294535e-01 5.33685237e-02 -1.04922950e+00
-6.33737385e-01 -4.81663197e-01 -9.80652034e-01 7.06363082e-01
-8.97774339e-01 -8.53435934e-01 -1.72306269e-01 3.28112900e-01
8.20925832e-01 -2.57729590e-02 9.41299558e-01 -3.83874476e-01
-2.36112565e-01 1.92053840e-01 -3.29411119e-01 1.39624253e-01
1.40845552e-01 -1.23076606e+00 4.88567919e-01 5.96278384e-02
3.42309214e-02 6.71690404e-01 1.01052952e+00 -1.05160344e+00
-1.04503596e+00 -6.08401239e-01 1.35422754e+00 -8.82063508e-01
9.50831115e-01 -5.20798862e-01 -1.04947495e+00 5.13242245e-01
6.26509190e-01 -1.04856181e+00 7.15107203e-01 4.52531666e-01
-3.21086824e-01 3.50006253e-01 -1.00705290e+00 6.92055881e-01
7.48875439e-01 -5.82909346e-01 -9.20014262e-01 5.13107002e-01
1.21329689e+00 -4.27803367e-01 -9.41405535e-01 -2.88295448e-01
3.98606598e-01 -8.52123678e-01 3.97321403e-01 -8.39149415e-01
4.73693758e-01 -1.66127816e-01 3.06159616e-01 -1.08873212e+00
-3.10499072e-01 -1.30047297e+00 -9.54376906e-02 1.55560029e+00
6.46634638e-01 -6.23182580e-02 7.30292976e-01 5.28176785e-01
-8.60142708e-02 -4.75195229e-01 -4.89378601e-01 -5.62884033e-01
-1.23233989e-01 -5.10996580e-01 5.12196124e-01 7.11173773e-01
1.28337514e+00 1.06251371e+00 -9.32503119e-02 -3.30385715e-01
2.88167566e-01 1.44614458e-01 5.78459978e-01 -1.50507164e+00
-2.15155050e-01 -4.49977428e-01 4.55974013e-01 -1.14491796e+00
-1.08666621e-01 -8.39565575e-01 1.49832526e-02 -1.69739509e+00
-8.19123816e-03 -3.49875301e-01 5.75076818e-01 4.55631793e-01
4.31771398e-01 -4.61926371e-01 1.64190739e-01 2.69161850e-01
-8.00785899e-01 1.67956650e-01 1.29671848e+00 -1.01525132e-02
-8.40404272e-01 3.31951410e-01 -9.39067602e-01 9.71781790e-01
7.35057235e-01 -4.66348916e-01 -3.90317142e-01 2.24328920e-01
6.36789143e-01 2.67233908e-01 -1.41548470e-01 -2.94394106e-01
3.60621005e-01 -1.98330760e-01 -1.11162983e-01 -6.66050494e-01
1.32032428e-02 -6.40921772e-01 9.47963074e-02 3.15448284e-01
-1.12675738e+00 -9.36550125e-02 -3.51649299e-02 1.68896973e-01
-3.50476243e-02 -6.54828131e-01 3.67183387e-01 -2.11158618e-01
-4.17362243e-01 -3.86170626e-01 -9.75681484e-01 -1.47192433e-01
7.15441763e-01 -2.22598642e-01 -3.73203069e-01 -6.87258005e-01
-1.25687599e+00 4.17917669e-01 -2.79890243e-02 6.13309383e-01
2.77113199e-01 -1.05675876e+00 -1.05171733e-01 -7.57917911e-02
1.97432935e-01 -1.98747396e-01 -1.26186058e-01 4.83707696e-01
-8.80108550e-02 4.76863325e-01 5.29712439e-02 -3.61320436e-01
-1.65236795e+00 5.99498510e-01 -7.43018789e-03 -6.24182105e-01
-6.97801650e-01 5.51590621e-01 7.47078061e-02 -1.64780721e-01
6.59098268e-01 -4.57243681e-01 -1.01921153e+00 4.13302571e-01
7.09972203e-01 3.58770579e-01 -1.13619082e-02 -7.47488886e-02
-5.77979796e-02 4.27795351e-01 -1.34586141e-01 -2.58881867e-01
1.06406510e+00 -4.51643318e-01 -1.72992110e-01 6.04424596e-01
9.99843895e-01 -1.23855695e-01 -1.01983762e+00 -5.66006720e-01
7.81662822e-01 1.39805768e-03 -4.04451042e-01 -1.30710578e+00
-2.84625381e-01 4.64487255e-01 1.37549162e-01 1.21669149e+00
5.96047997e-01 6.32891297e-01 5.33275366e-01 8.57705712e-01
-8.56194925e-03 -1.28148806e+00 5.53479850e-01 8.57030690e-01
1.21993935e+00 -7.83367932e-01 -3.27174872e-01 -7.16641068e-01
-1.20091903e+00 1.59443951e+00 4.57993716e-01 4.19317633e-01
6.98642135e-01 4.78471696e-01 3.34131628e-01 -5.44222057e-01
-9.39767003e-01 -2.41125882e-01 2.68990129e-01 8.31802726e-01
4.23406601e-01 -1.89379677e-01 -2.00224996e-01 7.97929585e-01
-6.65484071e-01 1.88938245e-01 4.63791460e-01 8.99031937e-01
-9.27939594e-01 -1.35456061e+00 -3.54844285e-03 4.40636128e-02
-6.41563460e-02 1.22875929e-01 -9.10811543e-01 5.58351696e-01
-3.80634815e-02 1.31851566e+00 -1.08140451e-03 -7.17031062e-01
4.51523840e-01 4.46898133e-01 1.83855504e-01 -9.83007252e-01
-1.03503001e+00 7.14236684e-03 8.36287618e-01 -3.74488562e-01
-3.52724642e-01 -7.01602817e-01 -1.34119165e+00 -1.95615962e-01
-4.06443566e-01 6.37775540e-01 7.07096696e-01 1.28658760e+00
1.67002603e-01 3.99779946e-01 7.92422771e-01 -2.85311520e-01
-2.74898052e-01 -1.14812887e+00 -1.88500509e-01 1.78907573e-01
1.61501110e-01 -2.97224760e-01 6.79195970e-02 4.23393011e-01] | [12.791146278381348, 7.985222339630127] |
ee91854c-254d-4f3e-9497-10cafca501fa | a-causal-framework-to-quantify-the-robustness | 2210.12023 | null | https://arxiv.org/abs/2210.12023v3 | https://arxiv.org/pdf/2210.12023v3.pdf | A Causal Framework to Quantify the Robustness of Mathematical Reasoning with Language Models | We have recently witnessed a number of impressive results on hard mathematical reasoning problems with language models. At the same time, the robustness of these models has also been called into question; recent works have shown that models can rely on shallow patterns in the problem description when generating a solution. Building on the idea of behavioral testing, we propose a novel framework, which pins down the causal effect of various factors in the input, e.g., the surface form of the problem text, the operands, and math operators on the output solution. By grounding the behavioral analysis in a causal graph describing an intuitive reasoning process, we study the behavior of language models in terms of robustness and sensitivity to direct interventions in the input space. We apply our framework on a test bed of math word problems. Our analysis shows that robustness does not appear to continuously improve as a function of size, but the GPT-3 Davinci models (175B) achieve a dramatic improvement in both robustness and sensitivity compared to all other GPT variants. | ['Mrinmaya Sachan', 'Bernhard Schölkopf', 'Kumar Shridhar', 'Zhijing Jin', 'Alessandro Stolfo'] | 2022-10-21 | null | null | null | null | ['mathematical-reasoning'] | ['natural-language-processing'] | [ 1.21334471e-01 3.38776141e-01 1.25367597e-01 -9.24917310e-02
-2.28635043e-01 -9.41342652e-01 7.05277026e-01 5.15677392e-01
1.18456557e-01 3.45537722e-01 1.34264916e-01 -8.16779196e-01
-4.18868840e-01 -1.32074404e+00 -1.06261456e+00 -1.07185163e-01
-1.43257171e-01 3.14805686e-01 4.36929226e-01 -5.12230814e-01
3.67762029e-01 1.95844755e-01 -1.41828060e+00 3.25758308e-01
9.42075431e-01 3.71643722e-01 -2.23618105e-01 8.02182257e-01
-1.41934827e-01 9.86823857e-01 -6.11629307e-01 -3.96969259e-01
2.32648432e-01 -4.71553624e-01 -1.01849067e+00 -1.39791235e-01
2.88041592e-01 -1.24895563e-02 -3.34170103e-01 1.26109982e+00
-2.23122742e-02 -2.08429843e-01 4.32985634e-01 -1.50651515e+00
-5.55626810e-01 1.38273084e+00 -2.68503040e-01 -9.14721095e-05
7.06991136e-01 5.71366608e-01 1.12126791e+00 -2.50665903e-01
5.47260642e-01 1.54135323e+00 5.08867800e-01 4.76465136e-01
-1.63980031e+00 -4.00659740e-01 1.74161881e-01 9.89435939e-04
-1.29018998e+00 -7.46857673e-02 2.39023656e-01 -6.68654144e-01
1.08479691e+00 4.63031650e-01 4.21500385e-01 8.08215439e-01
3.53282541e-01 4.84421372e-01 1.19791317e+00 -7.11836517e-01
3.03329140e-01 2.32770890e-01 5.78424037e-01 1.08293307e+00
5.47034442e-01 9.96773914e-02 -3.64253461e-01 -3.02226484e-01
6.06172800e-01 -4.52578127e-01 -5.10021336e-02 -4.65119004e-01
-8.47450435e-01 7.53908217e-01 1.37287781e-01 3.52941245e-01
4.06176955e-01 4.82357085e-01 7.32594654e-02 6.29889369e-01
-2.34255362e-02 8.73618305e-01 -5.20240426e-01 -1.96210250e-01
-5.10045171e-01 5.22011042e-01 1.07255888e+00 7.61358380e-01
4.29084092e-01 -2.59726554e-01 -3.67390424e-01 2.81817824e-01
1.70474440e-01 4.87591058e-01 1.52286828e-01 -7.22797334e-01
6.89428627e-01 1.07146919e+00 7.88866132e-02 -9.32147324e-01
-4.17187929e-01 -2.84902781e-01 -1.71318382e-01 1.85670197e-01
8.03437769e-01 -1.19013868e-01 -5.86216509e-01 2.15612769e+00
-9.11281481e-02 -6.19092509e-02 -3.22071952e-03 4.96994853e-01
2.74001867e-01 6.32143676e-01 7.42336065e-02 8.33252221e-02
1.01366138e+00 -5.87001920e-01 -3.31620991e-01 -3.58951926e-01
1.11549926e+00 -2.62546897e-01 1.34070122e+00 4.78829861e-01
-1.49247897e+00 -2.05667511e-01 -1.15985680e+00 9.05297101e-02
-2.80183345e-01 -4.89627570e-01 7.79299259e-01 9.57036495e-01
-1.13010180e+00 7.54983485e-01 -7.49883175e-01 -1.43143460e-01
1.03448287e-01 4.06512499e-01 -7.10486844e-02 -3.70950043e-01
-1.11697757e+00 7.86417067e-01 3.69702518e-01 -2.62971252e-01
-7.73305237e-01 -8.61382902e-01 -9.27066207e-01 4.47296202e-01
7.47642159e-01 -7.83436596e-01 1.19138885e+00 -5.70599973e-01
-1.44723308e+00 5.22110581e-01 -1.39389215e-02 -4.43959594e-01
6.90744519e-01 1.31687326e-02 -9.98836979e-02 -2.15944618e-01
-2.26643994e-01 2.51625836e-01 4.31511402e-01 -8.83147001e-01
-2.49854267e-01 -4.16219890e-01 7.72063017e-01 -3.89142990e-01
-1.54794350e-01 -6.63216971e-03 -1.35624766e-01 -4.14295226e-01
-6.52450770e-02 -9.95982468e-01 -2.71231622e-01 -3.94929379e-01
-5.42304873e-01 -2.77784407e-01 1.38753369e-01 -2.49502197e-01
1.38308156e+00 -2.35348868e+00 6.23063326e-01 5.71000993e-01
3.36923927e-01 -1.94540381e-01 -2.92912632e-01 5.21249652e-01
-3.73162597e-01 8.29899907e-01 -1.62858546e-01 2.11916536e-01
5.96442580e-01 5.12150675e-02 -5.40379941e-01 3.66793126e-01
4.51370060e-01 1.24516189e+00 -7.47010827e-01 -1.60748556e-01
-2.13925913e-02 -1.62102148e-01 -1.08526695e+00 1.44053265e-01
-8.18378448e-01 -5.77574559e-02 -4.10239249e-01 2.18734369e-01
2.45720714e-01 -4.54040647e-01 3.45443517e-01 5.14243186e-01
5.86278886e-02 5.27401149e-01 -1.35369956e+00 1.19919610e+00
-3.48932415e-01 3.24999452e-01 -2.80833244e-01 -6.57263279e-01
4.57178891e-01 6.35291561e-02 -3.33928287e-01 -8.24225843e-01
-3.14194113e-02 -1.17829246e-02 5.63806772e-01 -5.82448483e-01
2.23199412e-01 -2.10665673e-01 -4.19057161e-01 6.62839651e-01
-1.87317222e-01 -4.34819758e-01 5.57631612e-01 2.89798051e-01
1.57729471e+00 -8.11709911e-02 6.65916353e-02 -5.06222904e-01
4.86238778e-01 1.23213165e-01 1.90063938e-01 1.19943118e+00
5.87827444e-01 1.00861602e-01 1.29148436e+00 -2.07306132e-01
-9.98886645e-01 -9.67812061e-01 -1.97117198e-02 9.29329574e-01
-4.15603258e-02 -1.00114846e+00 -9.65123415e-01 -3.23199630e-01
6.92989156e-02 1.08403039e+00 -7.77716875e-01 -6.50471270e-01
-5.50426483e-01 -4.63058174e-01 6.62344575e-01 5.18472612e-01
1.45858556e-01 -7.74081767e-01 -5.79703450e-01 3.97741608e-03
2.36419082e-01 -9.65166271e-01 1.84812382e-01 1.25724301e-01
-9.55895901e-01 -1.02204823e+00 9.93302539e-02 -2.53955394e-01
7.41656363e-01 -1.53157339e-01 1.13602626e+00 5.41232169e-01
-2.80160904e-01 5.56175470e-01 -7.51614422e-02 -3.17783415e-01
-7.48695254e-01 -1.32710366e-02 -2.59254664e-01 -4.45135653e-01
-6.48724586e-02 -4.61457551e-01 7.53032416e-02 2.62017190e-01
-1.07488000e+00 -2.76205316e-02 2.45663255e-01 5.94924033e-01
-5.53395189e-02 5.50961256e-01 -2.07547545e-01 -9.54192817e-01
8.30513060e-01 -5.70387602e-01 -1.13041925e+00 4.88223702e-01
-4.39197928e-01 6.34480357e-01 6.96711600e-01 -3.57130051e-01
-4.54232037e-01 -3.63731325e-01 3.47259194e-01 3.38602699e-02
-9.81428195e-03 7.82115698e-01 -1.64333582e-01 5.49175553e-02
7.21670568e-01 -1.06772251e-01 -2.52942920e-01 -1.36953756e-01
2.42827520e-01 -2.24661827e-01 1.40872106e-01 -1.35393548e+00
1.23827338e+00 -1.19063370e-01 2.84458578e-01 -4.40578699e-01
-4.85317290e-01 3.36628914e-01 -4.10688221e-02 2.00206831e-01
5.44285178e-01 -3.28621715e-01 -1.12800014e+00 2.86427051e-01
-9.31591809e-01 -9.23536241e-01 -2.02220500e-01 -5.49134091e-02
-5.63838303e-01 1.75606504e-01 -6.85309589e-01 -8.28940034e-01
3.83581728e-01 -1.50534701e+00 7.78769255e-01 6.79726079e-02
-6.51228964e-01 -9.78514910e-01 -8.57633427e-02 2.45929547e-02
3.08544099e-01 1.85421761e-02 2.09193730e+00 -7.20114291e-01
-9.55577672e-01 -1.59209788e-01 1.18463971e-02 -3.45773771e-02
-3.87402236e-01 1.32486835e-01 -6.37319684e-01 -4.98230346e-02
7.13568777e-02 -1.05844252e-01 3.54534537e-01 1.08249016e-01
1.21523595e+00 -4.50075239e-01 -8.12837109e-02 3.12025219e-01
1.50894916e+00 -9.88370255e-02 7.51966119e-01 2.74579227e-01
3.34011734e-01 5.00702739e-01 1.16914995e-01 1.02377519e-01
3.35453898e-02 5.14806271e-01 2.10734397e-01 3.16478908e-01
1.77831456e-01 -5.48864603e-01 5.09036064e-01 1.88265562e-01
9.84856263e-02 -4.12402600e-01 -1.37351453e+00 2.19488800e-01
-1.72907162e+00 -7.91829407e-01 -1.16187826e-01 2.12873101e+00
9.59821761e-01 5.76873779e-01 1.63422033e-01 3.89555663e-01
3.00715983e-01 -2.19341278e-01 -2.29372516e-01 -5.80512464e-01
8.27703401e-02 5.17361581e-01 4.06414300e-01 8.67636144e-01
-3.31315368e-01 8.29152405e-01 7.14940977e+00 5.01201630e-01
-1.01081324e+00 -2.30430990e-01 4.29903209e-01 4.63741086e-02
-7.51498878e-01 1.96618557e-01 -5.16957879e-01 2.16777429e-01
1.07193422e+00 -5.52383661e-01 9.69341040e-01 6.51293933e-01
-1.50122195e-01 -2.64723867e-01 -1.74670386e+00 3.37103039e-01
-2.10527211e-01 -1.24885607e+00 5.69339693e-02 2.00981006e-01
4.52415735e-01 -5.09722710e-01 1.54398918e-01 4.94948000e-01
7.22671270e-01 -1.49294233e+00 1.05605948e+00 3.81927699e-01
2.96218932e-01 -6.11666262e-01 5.05832016e-01 5.57643712e-01
-5.81557691e-01 -4.22669590e-01 2.51147430e-02 -6.97576046e-01
-4.65897352e-01 2.32769534e-01 -8.45336676e-01 2.36859828e-01
3.35686982e-01 1.56537488e-01 -9.99663711e-01 8.33848894e-01
-6.16348863e-01 8.94938290e-01 -2.61286497e-01 -1.76953450e-01
6.88869506e-02 -8.79019052e-02 4.27485973e-01 9.25881743e-01
8.19846243e-02 3.75730067e-01 3.76871713e-02 1.38475072e+00
1.79043278e-01 -7.63447136e-02 -6.40945673e-01 -5.84205747e-01
6.68907091e-02 7.11656809e-01 -8.70324194e-01 -2.59438664e-01
-2.19104826e-01 2.62941927e-01 2.97782540e-01 4.09324914e-01
-9.59597170e-01 6.32174462e-02 5.09827673e-01 4.10211682e-01
2.52575874e-01 -4.48207110e-01 -6.20781362e-01 -1.05637360e+00
2.12038323e-01 -1.20230460e+00 2.35704437e-01 -9.87185001e-01
-9.59364593e-01 2.31905449e-02 3.94521087e-01 -3.06084812e-01
-7.02670664e-02 -8.23769093e-01 -5.44218421e-01 9.36738372e-01
-7.64726579e-01 -3.50296676e-01 1.28312722e-01 3.88395876e-01
1.10629648e-01 1.99643090e-01 7.80502379e-01 -2.41761252e-01
-7.48730958e-01 6.65875554e-01 -5.89049041e-01 8.30194131e-02
1.29355699e-01 -1.21130323e+00 6.33500755e-01 1.12715232e+00
1.89548414e-02 1.34707713e+00 1.02520204e+00 -5.36946595e-01
-1.97093308e+00 -6.16148710e-01 7.38377810e-01 -9.10534561e-01
1.22995138e+00 -6.12115383e-01 -7.36217797e-01 9.48609710e-01
3.04771513e-02 -3.33284795e-01 3.31988424e-01 4.03942764e-01
-8.02507758e-01 1.37759566e-01 -9.12818968e-01 1.05420315e+00
1.26699638e+00 -5.83587110e-01 -8.04646611e-01 2.82794625e-01
1.05499446e+00 -5.28594732e-01 -6.55741513e-01 2.77234018e-01
2.76751101e-01 -8.36690545e-01 7.81916678e-01 -1.08182144e+00
9.45012927e-01 -3.35350960e-01 -2.72691637e-01 -1.09573197e+00
-5.15388966e-01 -7.45989680e-01 -4.98706140e-02 1.01855111e+00
7.66760290e-01 -8.23239684e-01 3.59691173e-01 1.18983197e+00
4.86198366e-02 -5.12428701e-01 -5.15436709e-01 -6.46426082e-01
3.18295687e-01 -7.07966208e-01 7.16211736e-01 6.42250776e-01
5.47679543e-01 4.62520421e-01 2.00101569e-01 4.14039642e-01
1.59387380e-01 3.25482756e-01 8.04670274e-01 -1.19682169e+00
-8.91899049e-01 -7.34036267e-01 -3.77104968e-01 -5.58194280e-01
2.30565935e-01 -1.16986835e+00 -1.57387555e-01 -1.31340742e+00
3.73128563e-01 -4.57158297e-01 -2.67381594e-02 7.79088974e-01
-1.29241422e-01 -2.99134552e-01 3.60927790e-01 -3.08139354e-01
-4.10946608e-01 -1.14225909e-01 1.02487481e+00 -2.10295767e-01
-1.38153341e-02 -2.39435256e-01 -8.37351799e-01 6.48172140e-01
6.53583765e-01 -4.82598871e-01 -5.01304328e-01 -4.53490525e-01
1.15013945e+00 3.23233187e-01 5.75559735e-01 -9.86429274e-01
2.00292543e-01 -6.04330957e-01 -4.71516699e-02 1.13085017e-01
-1.97262496e-01 -6.58127069e-01 2.05337927e-01 7.61540949e-01
-7.88697660e-01 4.12014037e-01 6.93831146e-01 2.77398765e-01
2.01249331e-01 -3.82903397e-01 3.45514208e-01 -6.54776320e-02
-4.39045906e-01 -1.79841474e-01 -4.54736173e-01 3.04188877e-01
7.96118259e-01 -5.14084008e-03 -5.62807977e-01 -6.84471503e-02
-5.47528863e-01 2.70574868e-01 7.00830579e-01 4.26041752e-01
8.00005123e-02 -7.41845369e-01 -4.15211767e-01 3.14864457e-01
1.10952646e-01 -2.14950487e-01 1.47659089e-02 9.00584817e-01
-3.74803841e-01 3.21607351e-01 -7.91030899e-02 -4.97652620e-01
-1.13039434e+00 9.14439559e-01 3.56931418e-01 -2.53654361e-01
-3.86595935e-01 7.21908033e-01 3.41531277e-01 -2.74921000e-01
1.78590059e-01 -1.04785180e+00 3.22805405e-01 -4.51011598e-01
4.32953030e-01 6.76172078e-02 6.90181702e-02 2.96062857e-01
-3.05072188e-01 4.11856025e-01 1.68820411e-01 -2.33134538e-01
1.42978513e+00 4.37147528e-01 -5.40777206e-01 6.02442265e-01
5.17700016e-01 3.90918136e-01 -5.56644559e-01 4.53390144e-02
1.00652762e-01 -2.61468291e-01 -3.54889423e-01 -1.04609895e+00
-6.15036428e-01 7.49312997e-01 -1.81676224e-02 7.24918664e-01
7.98181236e-01 4.93969657e-02 1.60333827e-01 5.93810618e-01
5.34938395e-01 -7.43835688e-01 3.75733897e-02 7.86331534e-01
7.54512250e-01 -5.27126908e-01 -1.75301746e-01 -6.87788367e-01
1.35847731e-02 1.03515685e+00 4.84978825e-01 -1.58078417e-01
1.42816871e-01 7.07369030e-01 -5.32853305e-01 -2.25945815e-01
-9.79786396e-01 2.00233832e-01 -7.58505687e-02 1.75898030e-01
3.45938802e-01 6.01185784e-02 -1.43045381e-01 7.79089868e-01
-5.51695585e-01 5.65559827e-02 8.44957054e-01 9.38322783e-01
-2.75821120e-01 -1.25915980e+00 -6.04757071e-01 6.53433055e-02
-3.17261308e-01 -2.75066465e-01 -6.37427926e-01 1.00577807e+00
-7.53533542e-02 8.26645732e-01 -1.25861272e-01 -3.90455753e-01
4.69301969e-01 4.29313600e-01 9.95189130e-01 -1.01930511e+00
-6.56506240e-01 -5.35762548e-01 1.56014100e-01 -7.08659768e-01
2.45086700e-01 -4.96819228e-01 -1.15746999e+00 -6.04201376e-01
2.41578966e-02 1.72532514e-01 3.25811118e-01 1.17116606e+00
2.26317674e-01 5.99415720e-01 1.17970772e-01 -2.07328349e-01
-8.34087670e-01 -5.61699390e-01 -2.80550241e-01 2.89275527e-01
1.54290959e-01 -4.82219905e-01 -3.47191542e-01 -1.74691126e-01] | [9.154844284057617, 7.233509540557861] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.