Unnamed: 0
int64
0
7.24k
id
int64
1
7.28k
raw_text
stringlengths
9
124k
vw_text
stringlengths
12
15k
6,200
6,609
Attentional Pooling for Action Recognition Rohit Girdhar Deva Ramanan The Robotics Institute, Carnegie Mellon University http://rohitgirdhar.github.io/AttentionalPoolingAction Abstract We introduce a simple yet surprisingly powerful model to incorporate attention in action recognition and human object interaction tas...
6609 |@word kong:1 cnn:14 version:1 nd:1 everingham:1 chopping:4 tried:1 rgb:16 bn:3 forestry:2 mention:1 tr:1 harder:2 moment:1 initial:4 liu:2 contains:4 score:8 quo:1 ours:11 interestingly:4 batista:1 past:1 existing:5 current:3 contextual:2 com:1 skipping:2 guadarrama:1 yet:2 written:2 reminiscent:1 exposing:1 parm...
6,201
661
Improving Convergence in Hierarchical Matching Networks for Object Recognition Joachim Utans* Gene Gindi t Department of Electrical Engineering Yale University P. O. Box 2157 Yale Station New Haven, CT 06520 Abstract We are interested in the use of analog neural networks for recognizing visual objects. Objects are de...
661 |@word simulation:6 tr:1 solid:1 recursively:1 yaleu:2 initial:7 denoting:1 current:1 comparing:1 stony:2 must:8 j1:1 shape:3 designed:2 v:1 alone:1 xk:1 supplying:1 coarse:8 node:1 successive:1 simpler:1 incorrect:1 advocate:1 combine:1 expected:1 behavior:2 freeman:1 actual:2 becomes:1 estimating:1 notation:1 mat...
6,202
6,610
On the Consistency of Quick Shift Heinrich Jiang Google Inc. 1600 Amphitheatre Parkway, Mountain View, CA 94043 heinrich.jiang@gmail.com Abstract Quick Shift is a popular mode-seeking and clustering algorithm. We present finite sample statistical consistency guarantees for Quick Shift on mode and cluster recovery und...
6610 |@word mild:7 version:2 stronger:2 rsl:2 bf:4 r:21 pick:1 contains:3 interestingly:2 xinyang:1 current:3 com:1 gmail:1 must:1 john:1 partition:1 kdd:1 designed:1 update:1 v:1 intelligence:4 discovering:1 epanechnikov:1 provides:1 characterization:1 simpler:1 constructed:1 become:3 prove:1 manner:1 x0:72 amphitheat...
6,203
6,611
Breaking the Nonsmooth Barrier: A Scalable Parallel Method for Composite Optimization Fabian Pedregosa INRIA/ENS? Paris, France R?emi Leblond INRIA/ENS? Paris, France Simon Lacoste-Julien MILA and DIRO Universit?e de Montr?eal, Canada Abstract Due to their simplicity and excellent performance, parallel asynchronous ...
6611 |@word version:6 norm:3 stronger:1 c0:6 open:2 semicontinuous:1 crucially:1 hsieh:4 arti:1 eld:1 reduction:5 liu:6 contains:1 ecole:1 interestingly:1 outperforms:2 existing:5 past:1 current:2 com:1 leblond:11 si:10 yet:2 dx:1 written:4 grain:1 realistic:1 partition:2 kdd:5 cant:2 designed:1 update:29 greedy:2 sele...
6,204
6,612
Dual-Agent GANs for Photorealistic and Identity Preserving Profile Face Synthesis Jian Zhao1,2?? Lin Xiong3 Karlekar Jayashree3 Jianshu Li1 Fang Zhao1 Zhecan Wang4? Sugiri Pranata3 Shengmei Shen3 Shuicheng Yan1,5 Jiashi Feng1 1 3 National University of Singapore 2 National University of Defense Technology Panason...
6612 |@word cnn:2 version:1 middle:1 advantageous:1 proportion:1 tedious:1 annoying:1 shuicheng:1 bn:1 lpp:8 ld:2 liu:3 contains:1 tuned:1 ours:2 franklin:1 outperforms:4 existing:1 com:2 contextual:2 yet:4 parsing:1 refines:2 realistic:3 csc:1 shape:2 designed:3 interpretable:1 alone:1 generative:11 advancement:1 real...
6,205
6,613
Dilated Recurrent Neural Networks Shiyu Chang1?, Yang Zhang1?, Wei Han2?, Mo Yu1 , Xiaoxiao Guo1 , Wei Tan1 , Xiaodong Cui1 , Michael Witbrock1 , Mark Hasegawa-Johnson2 , Thomas S. Huang2 1 IBM Thomas J. Watson Research Center, Yorktown, NY 10598, USA 2 University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA ...
6613 |@word cnn:26 middle:3 version:3 norm:3 bn:1 harder:1 cyclic:4 contains:2 liu:1 daniel:1 document:1 past:1 existing:2 outperforms:2 com:4 comparing:1 si:2 yet:3 activation:1 must:2 gpu:1 john:1 devin:1 timestamps:4 happen:1 informative:1 subsequent:1 remove:1 designed:1 drop:1 atlas:1 half:1 fewer:6 guess:4 genera...
6,206
6,614
Hunt For The Unique, Stable, Sparse And Fast Feature Learning On Graphs Saurabh Verma Department of Computer Science University of Minnesota, Twin Cities verma@cs.umn.edu Zhi-Li Zhang Department of Computer Science University of Minnesota, Twin Cities zhang@cs.umn.edu Abstract For the purpose of learning on graphs, ...
6614 |@word kondor:5 inversion:1 polynomial:8 norm:2 loading:1 flach:1 nd:1 open:1 sg2:1 accounting:1 decomposition:3 elisseeff:1 kutzkov:1 sgd:1 mlk:1 accommodate:1 ld:1 reduction:5 substitution:1 series:2 contains:3 tist:1 tuned:1 interestingly:1 outperforms:3 past:1 current:4 comparing:1 com:1 si:1 yet:3 must:5 mesh...
6,207
6,615
Scalable Generalized Linear Bandits: Online Computation and Hashing Kwang-Sung Jun UW-Madison kjun@discovery.wisc.edu Aniruddha Bhargava UW-Madison aniruddha@wisc.edu Robert Nowak UW-Madison rdnowak@wisc.edu Rebecca Willett UW-Madison willett@discovery.wisc.edu Abstract Generalized Linear Bandits (GLBs), a natural...
6615 |@word mild:2 trial:3 exploitation:6 version:4 ruiqi:1 polynomial:2 norm:1 katja:1 logit:6 c0:7 dekel:2 d2:2 confirms:2 jingdong:1 q1:1 ld:1 initial:1 series:1 efficacy:1 document:1 ours:2 interestingly:1 past:1 existing:14 comparing:1 contextual:9 chu:3 must:3 readily:1 written:4 bd:4 john:3 sanjiv:1 confirming:1...
6,208
6,616
Probabilistic Models for Integration Error in the Assessment of Functional Cardiac Models Chris. J. Oates1,5 , Steven Niederer2 , Angela Lee2 , Fran?ois-Xavier Briol3 , Mark Girolami4,5 1 Newcastle University, 2 King?s College London, 3 University of Warwick, 4 Imperial College London, 5 Alan Turing Institute Abstract...
6616 |@word mri:2 version:1 polynomial:3 simulation:6 crucially:1 seek:1 covariance:4 contraction:4 p0:13 pressure:2 incurs:1 tr:1 initial:5 configuration:1 series:3 precluding:1 interestingly:1 activation:3 dx:40 reminiscent:1 written:1 mesh:1 numerical:31 subsequent:1 partition:1 interpretable:1 generative:1 selected...
6,209
6,617
Machine Learning with Adversaries: Byzantine Tolerant Gradient Descent El Mahdi El Mhamdi? EPFL, Switzerland elmahdi.elmhamdi@epfl.ch Peva Blanchard EPFL, Switzerland peva.blanchard@epfl.ch Rachid Guerraoui EPFL, Switzerland rachid.guerraoui@epfl.ch Julien Stainer EPFL, Switzerland julien.stainer@epfl.ch Abstract W...
6617 |@word mild:1 kgk:7 repository:3 version:4 hampson:1 polynomial:1 stronger:2 norm:6 nd:1 bf:6 seems:1 open:2 d2:1 confirms:1 covariance:1 sgd:15 thereby:1 solid:1 reduction:2 moment:8 liu:1 score:8 lichman:1 ours:1 interestingly:2 omniscient:5 outperforms:1 spambase:6 current:2 com:1 collude:1 yet:1 devin:2 realis...
6,210
6,618
Dynamic Safe Interruptibility for Decentralized Multi-Agent Reinforcement Learning El Mahdi El Mhamdi EPFL, Switzerland elmahdi.elmhamdi@epfl.ch Rachid Guerraoui EPFL, Switzerland rachid.guerraoui@epfl.ch Hadrien Hendrikx? ? Ecole Polytechnique, France hadrien.hendrikx@gmail.com Alexandre Maurer EPFL, Switzerland a...
6618 |@word achievable:1 norm:1 nd:1 c0:2 open:1 r:2 tried:1 pick:2 moment:1 initial:1 selecting:1 afraid:2 ecole:1 past:1 freitas:1 current:4 com:2 nt:3 gmail:1 yet:3 interrupted:18 realistic:2 remove:2 designed:2 update:11 alone:1 intelligence:6 greedy:8 stationary:1 item:1 short:2 provides:1 org:2 kristjan:1 wierstr...
6,211
6,619
Interactive Submodular Bandit 1 Lin Chen1,2 , Andreas Krause3 , Amin Karbasi1,2 Department of Electrical Engineering, 2 Yale Institute for Network Science, Yale University 3 Department of Computer Science, ETH Z?rich {lin.chen, amin.karbasi}@yale.edu, krausea@ethz.ch Abstract In many machine learning applications, s...
6619 |@word private:1 exploitation:1 faculty:1 polynomial:2 norm:17 laurence:1 crucially:1 decomposition:1 citeseer:1 mention:1 mcauley:1 bai:1 contains:1 selecting:5 daniel:4 rkhs:13 ours:2 document:2 outperforms:5 yajun:1 current:2 contextual:17 wd:1 com:1 beygelzimer:1 si:8 yet:1 chu:2 written:1 manuel:1 john:2 mani...
6,212
662
Mapping Between Neural and Physical Activities of the Lobster Gastric Mill Kenji Doya Mary E. T. Boyle Allen I. Selverston Department of Biology University of California, San Diego La Jolla, CA 92093-0322 Abstract A computer model of the musculoskeletal system of the lobster gastric mill was constructed in order t...
662 |@word neurophysiology:4 simulation:2 contraction:6 configuration:2 series:1 activation:5 must:1 entrance:1 motor:5 medial:8 update:1 pacemaker:1 nervous:5 plane:1 record:2 beauchamp:1 digestive:1 mathematical:1 constructed:2 differential:1 consists:1 combine:1 behavioral:7 inside:1 pairwise:1 behavior:1 mechanic:1...
6,213
6,620
Learning to See Physics via Visual De-animation Jiajun Wu MIT CSAIL Erika Lu University of Oxford William T. Freeman MIT CSAIL, Google Research Pushmeet Kohli DeepMind Joshua B. Tenenbaum MIT CSAIL Abstract We introduce a paradigm for understanding physical scenes without human annotations. At the core of our syst...
6620 |@word kohli:4 inversion:1 open:1 pieter:1 simulation:24 rgb:4 decomposition:1 blender:2 jacob:1 sgd:1 bai:2 liu:1 contains:1 initial:5 jimenez:1 salzmann:2 daniel:1 ours:4 past:1 existing:2 freitas:1 recovered:2 comparing:1 current:1 attracted:1 ronald:1 realistic:2 happen:3 blur:1 ronan:1 shape:3 christian:1 rem...
6,214
6,621
Label Efficient Learning of Transferable Representations across Domains and Tasks Zelun Luo Stanford University zelunluo@stanford.edu Yuliang Zou Virginia Tech ylzou@vt.edu Judy Hoffman University of California, Berkeley jhoffman@eecs.berkeley.edu Li Fei-Fei Stanford University feifeili@cs.stanford.edu Abstract We...
6621 |@word cnn:6 compression:1 fcns:1 pieter:1 d2:9 seek:3 propagate:1 rgb:1 vicky:1 shot:11 ld:1 initial:4 liu:4 contains:1 score:7 salzmann:1 tuned:7 ours:5 document:1 outperforms:1 existing:2 current:1 guadarrama:1 nt:3 luo:2 activation:6 diederik:2 written:1 hou:1 subsequent:1 blur:1 christian:1 update:1 depict:1 ...
6,215
6,622
Decoding with Value Networks for Neural Machine Translation Di He1 di_he@pku.edu.cn Tao Qin4 taoqin@microsoft.com Hanqing Lu2 hanqinglu@cmu.edu Liwei Wang1,5 wanglw@cis.pku.edu.cn Yingce Xia3 xiayingc@mail.ustc.edu.cn Tie-Yan Liu4 tie-yan.liu@microsoft.com 1 Key Laboratory of Machine Perception, MOE, School of EECS...
6622 |@word middle:1 briefly:1 open:2 pick:1 sgd:1 liu:10 contains:2 score:21 qatar:1 ours:3 past:1 outperforms:6 current:1 com:3 contextual:1 activation:1 guez:1 attracted:1 gpu:1 concatenate:2 designed:2 plot:1 drop:3 krikun:1 short:6 provides:3 contribute:2 node:1 c2:3 constructed:1 become:1 htx:3 incorrect:1 combin...
6,216
6,623
Parametric Simplex Method for Sparse Learning ? Haotian Pang? Robert Vanderbei? Han Liu?? Tuo Zhao? ? ? Princeton University Tencent AI Lab Northwestern University ? Georgia Tech? Abstract High dimensional sparse learning has imposed a great computational challenge to large scale data analysis. In this paper, we are...
6623 |@word briefly:1 polynomial:1 norm:8 nd:2 d2:15 hu:2 covariance:6 ipm:2 liu:1 series:1 zij:1 tuned:1 denoting:1 suppressing:1 genetic:1 amp:1 existing:7 blank:1 comparing:1 current:1 recovered:1 si:2 written:2 must:3 bd:2 hou:1 numerical:3 partition:8 remove:1 designed:1 update:3 rd2:1 selected:2 flare:9 according...
6,217
6,624
Group Sparse Additive Machine 1 Hong Chen1 , Xiaoqian Wang1 , Cheng Deng2 , Heng Huang1? Department of Electrical and Computer Engineering, University of Pittsburgh, USA 2 School of Electronic Engineering, Xidian University, China chenh@mail.hzau.edu.cn,xqwang1991@gmail.com chdeng@mail.xidian.edu.cn,heng.huang@pitt.e...
6624 |@word mild:3 repository:2 version:1 polynomial:2 norm:12 c0:1 decomposition:2 covariance:1 pick:1 liu:2 contains:1 lichman:1 rkhs:6 outperforms:1 existing:1 current:1 com:1 gmail:1 attracted:1 written:1 john:1 additive:59 informative:2 enables:1 statis:1 half:2 selected:3 kandasamy:1 math:1 zhang:2 unbounded:1 co...
6,218
6,625
Uprooting and Rerooting Higher-Order Graphical Models Mark Rowland? University of Cambridge mr504@cam.ac.uk Adrian Weller? University of Cambridge and Alan Turing Institute aw665@cam.ac.uk Abstract The idea of uprooting and rerooting graphical models was introduced specifically for binary pairwise models by Weller [1...
6625 |@word kohli:1 determinant:1 polynomial:1 stronger:4 advantageous:1 adrian:1 open:2 unif:1 barahona:1 pick:2 harder:1 kappen:1 initial:1 configuration:15 series:2 score:27 selecting:2 sherali:15 recovered:1 comparing:1 surprising:4 yet:2 forbidding:1 must:4 subsequent:2 partition:10 visible:1 plot:3 intelligence:4...
6,219
6,626
The Unreasonable Effectiveness of Structured Random Orthogonal Embeddings Krzysztof Choromanski ? Google Brain Robotics kchoro@google.com Mark Rowland ? University of Cambridge mr504@cam.ac.uk Adrian Weller University of Cambridge and Alan Turing Institute aw665@cam.ac.uk Abstract We examine a class of embeddings b...
6626 |@word msr:4 middle:2 version:3 illustrating:1 norm:2 ruiqi:1 adrian:1 open:1 unif:6 km:3 cos2:1 vldb:1 recursively:1 reduction:12 initial:1 contains:1 interestingly:4 kx0:1 com:1 chazelle:3 si:2 yet:2 dx:2 sanjiv:1 razenshteyn:1 shape:1 analytic:1 kdd:1 moreno:1 plot:2 sundaram:2 stationary:1 half:3 fewer:1 selec...
6,220
6,627
From Parity to Preference-based Notions of Fairness in Classification Muhammad Bilal Zafar MPI-SWS mzafar@mpi-sws.org Krishna P. Gummadi MPI-SWS gummadi@mpi-sws.org Isabel Valera MPI-IS isabel.valera@tue.mpg.de Manuel Gomez Rodriguez MPI-SWS manuelgr@mpi-sws.org Adrian Weller University of Cambridge & Alan Turing In...
6627 |@word repository:1 version:1 middle:1 norm:3 sex:1 adrian:1 open:2 incurs:2 profit:3 offending:1 reduction:2 venkatasubramanian:2 contains:2 disparity:5 selecting:1 score:1 precluding:1 bilal:1 existing:4 contextual:1 com:1 manuel:1 yet:1 kdd:4 hypothesize:1 drop:2 propublica:4 discrimination:10 v:2 smith:1 core:...
6,221
6,628
Inferring Generative Model Structure with Static Analysis Paroma Varma1 , Bryan He2 , Payal Bajaj2 , Nishith Khandwala2 , Imon Banerjee3 , Daniel Rubin3,4 , Christopher R?2 1 Electrical Engineering, 2 Computer Science, 3 Biomedical Data Science, 4 Radiology Stanford University {paroma,bryanhe,pabajaj,nishith,imonb,rub...
6628 |@word cnn:1 seems:1 nd:1 vldb:1 simulation:4 seek:1 programmatically:5 contrastive:2 pick:3 mention:1 configuration:1 score:5 daniel:1 tuned:3 fa8750:3 outperforms:2 existing:1 recovered:1 comparing:2 written:2 must:3 realistic:1 happen:1 distant:4 benign:1 shape:2 treating:1 interpretable:3 sponsored:1 takamatsu...
6,222
6,629
Structured Embedding Models for Grouped Data Maja Rudolph Columbia Univ. maja@cs.columbia.edu Francisco Ruiz Univ. of Cambridge Columbia Univ. Susan Athey Stanford Univ. David Blei Columbia Univ. Abstract Word embeddings are a powerful approach for analyzing language, and exponential family embeddings (EFE) extend...
6629 |@word proportion:2 stronger:1 open:1 uncovers:1 sgd:1 yih:1 harder:1 carry:1 reduction:1 contains:7 document:3 fa8750:1 outperforms:5 existing:1 past:1 com:1 si:4 must:1 john:1 devin:1 enables:1 remove:2 hypothesize:2 intelligence:9 discovering:1 fewer:3 item:11 generative:1 smith:1 blei:4 barkan:2 provides:3 mat...
6,223
663
A Note on Learning Vector Quantization Virginia R. de Sa Dana H. Ballard Department of Computer Science University of Rochester Rochester, NY 14627 Department of Computer Science University of Rochester Rochester, NY 14627 Abstract Vector Quantization is useful for data compression. Competitive Learning which mini...
663 |@word middle:1 version:2 compression:2 pulse:1 seek:1 simulation:1 barney:1 initial:3 nowlan:2 assigning:1 dx:3 class1:6 xlclass:1 update:2 alit:1 draft:1 quantizer:3 revisited:1 math:1 codebook:25 zii:1 c2:1 consists:1 ra:1 expected:1 formants:1 decreasing:8 little:1 window:26 minimizes:2 classifier:1 grant:2 pos...
6,224
6,630
A Linear-Time Kernel Goodness-of-Fit Test Wittawat Jitkrittum Gatsby Unit, UCL Wenkai Xu Gatsby Unit, UCL Zolt?n Szab?? CMAP, ?cole Polytechnique wittawatj@gmail.com wenkaix@gatsby.ucl.ac.uk zoltan.szabo@polytechnique.edu Kenji Fukumizu The Institute of Statistical Mathematics fukumizu@ism.ac.jp Arthur Gretton? ...
6630 |@word trial:2 version:1 briefly:1 eliminating:1 norm:4 stronger:1 smirnov:1 c0:4 open:2 d2:5 seek:1 simulation:3 covariance:2 zolt:2 tr:1 harder:1 initial:1 liu:1 series:1 score:1 tuned:1 rkhs:6 com:3 gmail:2 universality:1 written:1 readily:1 john:1 numerical:1 chicago:3 informative:1 wx:1 analytic:10 plot:4 int...
6,225
6,631
Cortical microcircuits as gated-recurrent neural networks Rui Ponte Costa? Centre for Neural Circuits and Behaviour Dept. of Physiology, Anatomy and Genetics University of Oxford, Oxford, UK rui.costa@cncb.ox.ac.uk Yannis M. Assael? Dept. of Computer Science University of Oxford, Oxford, UK and DeepMind, London, UK y...
6631 |@word h:1 neurophysiology:1 cox:2 version:1 middle:1 hippocampus:3 open:1 pulse:1 kappen:1 initial:1 contains:1 series:1 initialisation:1 bc:1 interestingly:1 past:1 freitas:2 existing:1 current:4 com:1 contextual:4 analysed:2 activation:4 yet:2 must:1 exposing:1 hyperpolarizing:1 subsequent:1 plasticity:13 enabl...
6,226
6,632
k-Support and Ordered Weighted Sparsity for Overlapping Groups: Hardness and Algorithms Cong Han Lim University of Wisconsin-Madison clim9@wisc.edu Stephen J. Wright University of Wisconsin-Madison swright@cs.wisc.edu Abstract The k-support and OWL norms generalize the `1 norm, providing better prediction accuracy a...
6632 |@word multitask:1 trial:1 cox:2 middle:1 version:3 polynomial:1 norm:85 replicate:1 open:1 palma:1 decomposition:24 jacob:1 pick:2 harder:1 configuration:2 contains:1 efficacy:1 series:2 selecting:1 tuned:1 current:1 com:1 incidence:1 si:2 written:1 numerical:2 partition:1 plot:2 update:5 v:1 greedy:1 leaf:6 weig...
6,227
6,633
A simple model of recognition and recall memory Nisheeth Srivastava Computer Science, IIT Kanpur Kanpur, UP 208016 nsrivast@cse.iitk.ac.in Edward Vul Dept of Psychology, UCSD 9500 Gilman Drive La Jolla CA 92093 evul@ucsd.edu Abstract We show that several striking differences in memory performance between recognition...
6633 |@word trial:7 illustrating:1 inversion:1 stronger:3 approved:1 hippocampus:1 extinction:1 termination:2 additively:1 simulation:9 irb:1 attended:1 asks:1 harder:4 contains:1 exclusively:1 uncovered:1 ours:2 reaction:1 current:1 comparing:1 activation:13 assigning:1 atop:1 john:3 realistic:1 engendered:1 wanted:1 ...
6,228
6,634
On Structured Prediction Theory with Calibrated Convex Surrogate Losses Anton Osokin INRIA/ENS?, Paris, France HSE?, Moscow, Russia Francis Bach INRIA/ENS?, Paris, France Simon Lacoste-Julien MILA and DIRO Universit? de Montr?al, Canada Abstract We provide novel theoretical insights on structured prediction in the ...
6634 |@word msr:1 illustrating:1 version:1 norm:10 c0:2 simplifying:2 pick:1 sgd:1 harder:1 contains:4 score:43 selecting:1 rkhs:8 ours:1 existing:2 dx:3 attracted:1 john:2 refines:1 kpf:1 informative:2 shape:1 hofmann:2 christian:1 update:2 juditsky:1 v:1 implying:4 mackey:1 intelligence:1 xk:2 mccallum:1 smith:2 coar...
6,229
6,635
Best of Both Worlds: Transferring Knowledge from Discriminative Learning to a Generative Visual Dialog Model Jiasen Lu1?, Anitha Kannan2?, Jianwei Yang1 , Devi Parikh3,1 , Dhruv Batra3,1 1 Georgia Institute of Technology, 2 Curai, 3 Facebook AI Research {jiasenlu, jw2yang, parikh, dbatra}@gatech.edu Abstract We prese...
6635 |@word cnn:3 stronger:1 norm:2 q1:1 attended:4 mengye:1 ld:3 lantao:1 liu:2 ndez:1 score:14 att:14 ours:13 bilal:1 outperforms:4 existing:1 guadarrama:2 current:5 comparing:5 anne:2 haoyuan:2 com:1 yet:1 z2:4 must:1 john:1 concatenate:1 realistic:1 informative:3 confirming:2 christian:1 enables:2 drop:1 update:4 v...
6,230
6,636
MaskRNN: Instance Level Video Object Segmentation Yuan-Ting Hu UIUC ythu2@illinois.edu Jia-Bin Huang Virginia Tech jbhuang@vt.edu Alexander G. Schwing UIUC aschwing@illinois.edu Abstract Instance level video object segmentation is an important technique for video editing and compression. To capture the temporal coh...
6636 |@word cnn:2 compression:2 nd:2 hu:1 shot:1 configuration:2 contains:5 denoting:1 ours:4 past:1 current:10 assigning:2 gpu:1 refines:1 subsequent:1 shape:5 hvs:1 occlude:1 stationary:1 cue:5 beginning:1 short:3 chua:1 provides:1 detecting:1 location:7 successive:1 org:1 zhang:1 height:1 along:1 brostow:1 yuan:2 co...
6,231
6,637
Gated Recurrent Convolution Neural Network for OCR Jianfeng Wang? Beijing University of Posts and Telecommunications Beijing 100876, China jianfengwang1991@gmail.com Xiaolin Hu Tsinghua National Laboratory for Information Science and Technology (TNList) Department of Computer Science and Technology Center for Brain-Ins...
6637 |@word neurophysiology:2 cnn:15 pw:10 wco:1 wiesel:1 bf:1 hu:4 bn:9 concise:1 tnlist:1 bai:2 configuration:4 contains:6 score:1 hereafter:1 liu:3 ours:1 document:2 outperforms:3 existing:2 past:2 blank:2 com:2 comparing:1 mishra:4 sosa:1 babenko:1 gmail:1 written:2 parsing:1 explorative:1 realistic:1 alphanumeric:...
6,232
6,638
Towards Accurate Binary Convolutional Neural Network Xiaofan Lin Cong Zhao Wei Pan* DJI Innovations Inc, Shenzhen, China {xiaofan.lin, cong.zhao, wei.pan}@dji.com Abstract We introduce a novel scheme to train binary convolutional neural networks (CNNs) ? CNNs with weights and activations constrained to {-1,+1} at run-...
6638 |@word cnn:6 version:1 seems:2 instruction:1 propagate:1 bn:2 teich:1 citeseer:1 reduction:1 electronics:1 configuration:3 contains:3 liu:2 tuned:1 bitwise:12 com:1 luo:1 activation:73 must:2 numerical:1 drop:3 update:1 half:1 leaf:1 device:2 short:1 quantized:6 node:1 zhang:1 five:1 height:1 along:1 symposium:2 i...
6,233
6,639
Semi-Supervised Learning for Optical Flow with Generative Adversarial Networks Wei-Sheng Lai1 Jia-Bin Huang2 Ming-Hsuan Yang1,3 2 3 University of California, Merced Virginia Tech Nvidia Research 1 2 {wlai24|mhyang}@ucmerced.edu jbhuang@vt.edu 1 Abstract Convolutional neural networks (CNNs) have recently been applied ...
6639 |@word cnn:11 fcns:1 open:1 brightness:26 inpainting:2 tr:1 ld:2 initial:1 minmax:1 series:1 score:2 liu:1 disparity:1 tuned:3 ours:2 outperforms:3 existing:4 current:1 activation:1 refines:1 realistic:3 blur:1 enables:1 update:5 generative:16 intelligence:2 advancement:1 provides:1 zhang:2 five:1 qualitative:1 ij...
6,234
664
Forecasting Demand for Electric Power Jen-Lun Yuan and Terrence L. Fine School of Electrical Engineering Cornell University Ithaca, NY 14853 Abstract We are developing a forecaster for daily extremes of demand for electric power encountered in the service area of a large midwestern utility and using this application ...
664 |@word economically:1 version:2 achievable:1 termination:1 forecaster:2 decomposition:1 covariance:3 weekday:1 thereby:2 reduction:6 initial:3 series:1 denoting:1 past:1 current:4 scatter:3 additive:1 numerical:1 plot:3 atlas:1 v:3 fewer:1 devising:1 plane:1 short:4 node:6 monday:8 mathematical:1 along:1 constructe...
6,235
6,640
Learning a Multi-View Stereo Machine Abhishek Kar UC Berkeley akar@berkeley.edu Christian H?ne UC Berkeley chaene@berkeley.edu Jitendra Malik UC Berkeley malik@berkeley.edu Abstract We present a learnt system for multi-view stereopsis. In contrast to recent learning based methods for 3D reconstruction, we leverage ...
6640 |@word cnn:6 version:2 repository:1 replicate:1 choy:2 seitz:2 shading:2 reduction:2 bai:1 liu:1 disparity:9 score:1 ours:1 current:1 comparing:1 yet:2 conforming:2 gpu:2 mesh:2 subsequent:1 realistic:1 visible:1 shape:23 christian:2 enables:1 visibility:2 designed:2 drop:1 depict:1 progressively:2 cue:13 fewer:4 ...
6,236
6,641
Phase Transitions in the Pooled Data Problem Jonathan Scarlett and Volkan Cevher Laboratory for Information and Inference Systems (LIONS) ?cole Polytechnique F?d?rale de Lausanne (EPFL) {jonathan.scarlett,volkan.cevher}@epfl.ch Abstract In this paper, we study the pooled data problem of identifying the labels associat...
6641 |@word mild:1 version:1 polynomial:1 proportion:4 norm:1 nd:10 open:2 seek:1 pg:6 initial:1 contains:2 series:1 selecting:1 denoting:1 existing:5 nt:9 jaynes:1 must:1 readily:2 written:1 john:1 item:19 accordingly:2 vanishing:1 volkan:2 characterization:3 provides:1 math:1 allerton:1 org:3 simpler:1 unbounded:3 ma...
6,237
6,642
Universal Style Transfer via Feature Transforms Yijun Li UC Merced yli62@ucmerced.edu Zhaowen Wang Adobe Research zhawang@adobe.com Chen Fang Adobe Research cfang@adobe.com Xin Lu Adobe Research xinl@adobe.com Jimei Yang Adobe Research jimyang@adobe.com Ming-Hsuan Yang UC Merced, NVIDIA Research mhyang@ucmerced.edu ...
6642 |@word h:2 middle:2 inversion:1 advantageous:1 kokkinos:1 nd:1 open:1 rgb:2 covariance:12 decomposition:2 jacob:1 thereby:1 shot:1 accommodate:1 carry:3 shechtman:2 inefficiency:1 tuned:1 ours:4 subjective:1 existing:7 com:5 yet:2 gpu:1 shape:1 enables:2 remove:1 designed:2 generative:1 accordingly:1 coarse:6 pref...
6,238
6,643
On the Model Shrinkage Effect of Gamma Process Edge Partition Models Iku Ohama?? Issei Sato? Takuya Kida? Hiroki Arimura? ? ? ? Panasonic Corp., Japan The Univ. of Tokyo, Japan Hokkaido Univ., Japan ohama.iku@jp.panasonic.com sato@k.u-tokyo.ac.jp {kida,arim}@ist.hokudai.ac.jp Abstract The edge partition model (EPM) is...
6643 |@word version:2 stronger:1 c0:20 hu:1 takuya:1 liu:1 score:1 hereafter:1 com:1 analysed:2 written:1 additive:2 partition:9 j1:1 shape:1 designed:2 update:2 generative:13 discovering:2 selected:1 parameterization:1 geyer:1 yamada:1 colored:1 blei:2 org:3 five:2 c2:4 constructed:1 beta:2 issei:1 introduce:1 manner:...
6,239
6,644
Pose Guided Person Image Generation Liqian Ma1 Xu Jia2? Qianru Sun3? Bernt Schiele3 Tinne Tuytelaars2 Luc Van Gool1,4 KU-Leuven/PSI, TRACE (Toyota Res in Europe) 2 KU-Leuven/PSI, IMEC 3 Max Planck Institute for Informatics, Saarland Informatics Campus 4 ETH Zurich {liqian.ma, xu.jia, tinne.tuytelaars, luc.vangool}@es...
6644 |@word trial:1 pieter:1 propagate:1 tenka:1 jingdong:1 ld:1 initial:12 liu:2 contains:1 score:6 jimenez:1 daniel:1 denoting:1 ours:3 interestingly:1 outperforms:1 freitas:1 cvae:1 comparing:1 luo:1 diederik:2 connectomics:1 john:2 refines:1 realistic:6 concatenate:3 distant:1 numerical:1 shape:1 uria:1 motor:1 vis...
6,240
6,645
Inference in Graphical Models via Semidefinite Programming Hierarchies Murat A. Erdogdu Microsoft Research erdogdu@cs.toronto.edu Yash Deshpande MIT and Microsoft Research yash@mit.edu Andrea Montanari Stanford University montanari@stanford.edu Abstract Maximum A posteriori Probability (MAP) inference in graphical m...
6645 |@word briefly:1 version:2 polynomial:3 norm:1 nd:1 open:2 barahona:3 adrian:1 accounting:2 contraction:2 recursively:1 carry:2 moment:1 reduction:4 contains:1 sherali:4 shum:1 denoting:3 outperforms:3 ka:4 surprising:1 si:1 assigning:1 danny:1 written:5 numerical:4 partition:1 plot:5 update:7 intelligence:2 fewer...
6,241
6,646
Variable Importance using Decision Trees Jalil Kazemitabar UCLA sjalilk@ucla.edu Arash A. Amini UCLA aaamini@ucla.edu Adam Bloniarz UC Berkeley? adam@stat.berkeley.edu Ameet Talwalkar CMU talwalkar@cmu.edu Abstract Decision trees and random forests are well established models that not only offer good predictive pe...
6646 |@word mild:1 trial:1 version:8 norm:4 unif:8 simulation:7 r:1 simplifying:1 covariance:1 boundedness:1 reduction:13 liu:1 contains:1 score:7 series:3 ours:1 existing:2 xnj:1 current:1 comparing:1 recovered:1 si:12 yet:1 additive:17 realistic:2 thrust:1 numerical:1 plot:3 greedy:4 leaf:1 selected:2 generative:3 ha...
6,242
6,647
Preventing Gradient Explosions in Gated Recurrent Units Sekitoshi Kanai, Yasuhiro Fujiwara, Sotetsu Iwamura NTT Software Innovation Center 3-9-11, Midori-cho, Musashino-shi, Tokyo {kanai.sekitoshi, fujiwara.yasuhiro, iwamura.sotetsu}@lab.ntt.co.jp Abstract A gated recurrent unit (GRU) is a successful recurrent neural ...
6647 |@word trial:1 briefly:1 norm:20 bptt:3 heuristically:1 linearized:2 decomposition:2 jingdong:1 prokhorov:2 pg:1 sgd:10 initial:4 series:4 bppt:1 tuned:1 kurt:2 past:8 outperforms:2 comparing:1 manuel:1 activation:3 diederik:1 must:1 designed:1 update:7 midori:1 polyphonic:5 bart:2 selected:1 vanishing:4 short:4 p...
6,243
6,648
On the Power of Truncated SVD for General High-rank Matrix Estimation Problems Simon S. Du Carnegie Mellon University ssdu@cs.cmu.edu Yining Wang Carnegie Mellon University yiningwa@cs.cmu.edu Aarti Singh Carnegie Mellon University aartisingh@cmu.edu Abstract ? that is close to a general high-rank positive semiWe sh...
6648 |@word mild:1 private:1 version:1 rising:1 polynomial:5 norm:41 nd:2 r:3 decomposition:8 covariance:21 arti:1 eld:1 asks:1 mention:1 tr:3 nystr:1 liu:2 xinyang:1 existing:7 comparing:1 luo:3 yet:2 attracted:1 must:1 additive:4 cant:1 remove:1 mackey:2 intelligence:1 cult:1 isotropic:1 contribute:1 simpler:1 zhang:...
6,244
6,649
f -GANs in an Information Geometric Nutshell Richard Nock?,?,? Zac Cranko?,? Aditya Krishna Menon?,? ?,? Lizhen Qu Robert C. Williamson?,? ? ? Data61, the Australian National University and ? the University of Sydney {firstname.lastname, aditya.menon, bob.williamson}@data61.csiro.au Abstract Nowozin et al showed last...
6649 |@word deformed:10 mild:3 version:2 briefly:2 nd:1 suitably:1 open:2 d2:2 simplifying:1 pick:2 naudts:1 carry:1 moment:1 initial:1 liu:1 series:2 exclusively:1 denoting:2 document:1 current:2 com:1 activation:28 yet:1 tackling:1 distant:1 happen:1 shape:1 drop:1 plot:1 v:3 generative:8 guess:1 warmuth:1 ith:1 shor...
6,245
665
Generalization Abilities of Cascade Network Architectures E. Littmann* H. Ritter Department of Information Science Bielefeld University D-4800 Bielefeld, FRG littmann@techfak.uni-bielefeld.de Department of Information Science Bielefeld University D-4800 Bielefeld, FRG helge@techfak.uni- bielefeld.de Abstract In [5]...
665 |@word inversion:1 covariance:2 paid:1 tr:1 series:7 lapedes:2 meyering:1 comparing:1 must:2 mackey:4 iso:4 node:28 ron:1 differential:2 symposium:1 edelman:1 expected:1 roughly:1 little:1 increasing:1 becomes:2 provided:1 matched:2 kaufman:1 rm:1 control:1 unit:12 grant:1 local:3 limit:1 severely:1 consequence:1 i...
6,246
6,650
Toward Multimodal Image-to-Image Translation Jun-Yan Zhu UC Berkeley Trevor Darrell UC Berkeley Richard Zhang UC Berkeley Alexei A. Efros UC Berkeley Oliver Wang Adobe Research Deepak Pathak UC Berkeley Eli Shechtman Adobe Research Abstract Many image-to-image translation problems are ambiguous, as a single input ...
6650 |@word cnn:1 middle:1 version:3 inversion:1 unpopulated:1 tried:1 propagate:2 asks:1 inpainting:2 shechtman:3 contains:2 score:10 document:1 existing:2 cvae:23 written:1 subsequent:1 concatenate:1 realistic:14 confirming:1 shape:2 interpretable:1 v:3 generative:22 instantiate:1 website:2 intelligence:1 parameteriz...
6,247
6,651
Mixture-Rank Matrix Approximation for Collaborative Filtering Dongsheng Li1 Chao Chen1 Wei Liu2? Tun Lu3,4 Ning Gu3,4 Stephen M. Chu1 1 IBM Research - China 2 Tencent AI Lab, China 3 School of Computer Science, Fudan University, China 4 Shanghai Key Laboratory of Data Science, Fudan University, China {ldsli, cs...
6651 |@word private:1 norm:3 km:6 confirms:4 initial:5 contains:1 score:1 series:1 outperforms:4 existing:2 com:1 contextual:1 toh:1 yet:1 chu:2 shakespeare:1 kdd:2 remove:1 update:1 v:1 alone:2 intelligence:2 mackey:1 weighing:1 item:63 accordingly:1 isotropic:1 prize:1 dear:1 boosting:1 location:2 club:1 toronto:1 si...
6,248
6,652
Non-monotone Continuous DR-submodular Maximization: Structure and Algorithms An Bian ETH Zurich ybian@inf.ethz.ch Kfir Y. Levy ETH Zurich yehuda.levy@inf.ethz.ch Andreas Krause ETH Zurich krausea@ethz.ch Joachim M. Buhmann ETH Zurich jbuhmann@inf.ethz.ch Abstract DR-submodular continuous functions are important obj...
6652 |@word mild:1 shayan:1 briefly:1 polynomial:5 norm:1 stronger:2 nd:1 laurence:1 open:1 bachman:2 x2p:1 tr:1 ld:2 reduction:4 initial:1 zij:2 daniel:1 document:3 past:1 existing:2 current:1 com:2 yet:3 attracted:1 luis:1 determinantal:4 john:1 partition:1 enables:6 plot:1 update:2 stationary:15 greedy:2 selected:1 ...
6,249
6,653
Learning with Average Top-k Loss Yanbo Fan3,4,1 , Siwei Lyu1?, Yiming Ying2 , Bao-Gang Hu3,4 1 Department of Computer Science, University at Albany, SUNY 2 Department of Mathematics and Statistics, University at Albany, SUNY 3 National Laboratory of Pattern Recognition, CASIA 4 University of Chinese Academy of Science...
6653 |@word madelon:1 repository:1 norm:2 proportion:1 calculus:1 hu:3 wexler:1 accommodate:2 liu:1 tist:1 rkhs:6 spambase:1 existing:2 outperforms:1 yet:2 must:1 subsequent:1 plot:5 fund:1 v:4 intelligence:2 selected:1 rudin:1 monk:5 provides:3 uca:2 completeness:1 dn:1 direct:1 shorthand:1 combine:3 introduce:3 pairw...
6,250
6,654
Learning multiple visual domains with residual adapters Sylvestre-Alvise Rebuffi1 Hakan Bilen1,2 1 Visual Geometry Group University of Oxford {srebuffi,hbilen,vedaldi}@robots.ox.ac.uk Andrea Vedaldi1 2 School of Informatics University of Edinburgh Abstract There is a growing interest in learning data representa...
6654 |@word aircraft:6 multitask:2 cnn:1 kokkinos:2 rgb:1 bn:11 decomposition:3 shot:2 reduction:2 initial:1 configuration:2 contains:9 liu:3 selecting:1 score:10 hoiem:1 tuned:3 document:1 outperforms:3 existing:1 current:1 wd:1 luo:1 dx:3 reminiscent:1 must:1 written:1 john:1 gavves:1 realistic:1 designed:1 drop:3 v:...
6,251
6,655
Dykstra?s Algorithm, ADMM, and Coordinate Descent: Connections, Insights, and Extensions Ryan J. Tibshirani Department of Statistics and Machine Learning Department Carnegie Mellon University Pittsburgh, PA 15213 ryantibs@stat.cmu.edu Abstract We study connections between Dykstra?s algorithm for projecting onto an in...
6655 |@word mild:1 version:18 seems:4 replicate:1 stronger:1 c0:3 adrian:1 confirms:1 seek:2 decomposition:4 jacob:1 cyclic:5 liu:1 series:2 daniel:2 denoting:1 interestingly:1 reinvented:1 existing:3 comparing:1 luo:6 surprising:1 rpi:14 chu:1 must:1 john:2 realize:1 stemming:1 numerical:6 subsequent:1 additive:1 wenj...
6,252
6,656
Learning Spherical Convolution for Fast Features from 360? Imagery Yu-Chuan Su Kristen Grauman The University of Texas at Austin Abstract While 360? cameras offer tremendous new possibilities in vision, graphics, and augmented reality, the spherical images they produce make core feature extraction non-trivial. Convol...
6656 |@word multitask:1 cnn:16 version:2 rising:1 compression:2 replicate:4 everingham:1 hu:2 zelnik:1 azimuthal:1 rgb:2 photographer:1 liu:1 offering:1 ours:1 interestingly:1 document:1 outperforms:5 existing:13 current:2 com:3 comparing:1 guadarrama:1 yet:1 must:1 subsequent:1 romero:1 shape:8 analytic:2 remove:2 plo...
6,253
6,657
MarrNet: 3D Shape Reconstruction via 2.5D Sketches Jiajun Wu* MIT CSAIL Yifan Wang* ShanghaiTech University Tianfan Xue MIT CSAIL William T. Freeman MIT CSAIL, Google Research Xingyuan Sun Shanghai Jiao Tong University Joshua B. Tenenbaum MIT CSAIL Abstract 3D object reconstruction from a single image is a highl...
6657 |@word mild:1 kohli:2 repository:2 middle:1 choy:5 rgb:11 decomposition:1 jacob:1 sgd:1 harder:1 shading:4 contains:3 hoiem:4 jimenez:1 tuned:7 past:2 recovered:2 cad:1 yet:1 dx:10 gpu:1 parsing:1 john:1 realistic:4 happen:1 informative:1 ronan:1 shape:97 ashutosh:1 v:1 generative:3 amir:1 plane:1 davison:2 provid...
6,254
6,658
Multimodal Learning and Reasoning for Visual Question Answering Ilija Ilievski Integrative Sciences and Engineering National University of Singapore ilija.ilievski@u.nus.edu Jiashi Feng Electrical and Computer Engineering National University of Singapore elefjia@nus.edu.sg Abstract Reasoning about entities and their...
6658 |@word multitask:2 cnn:5 version:2 hu:3 integrative:1 shuicheng:1 jacob:4 contains:4 score:5 hoiem:1 outperforms:1 existing:1 current:4 com:4 haoyuan:1 activation:1 diederik:1 must:1 parsing:1 john:1 ronan:2 enables:3 christian:1 remove:1 designed:1 interpretable:1 intelligence:4 item:1 kyoung:1 short:4 num:4 ment...
6,255
6,659
Adversarial Surrogate Losses for Ordinal Regression Rizal Fathony Mohammad Bashiri Brian D. Ziebart Department of Computer Science University of Illinois at Chicago Chicago, IL 60607 {rfatho2, mbashi4, bziebart}@uic.edu Abstract Ordinal regression seeks class label predictions when the penalty incurred for mistakes ...
6659 |@word repository:2 version:2 nd:1 c0:2 triazine:4 seek:4 moment:1 reduction:8 liu:5 contains:3 series:1 lichman:1 undiscovered:1 existing:4 jaynes:1 anqi:2 chu:4 must:1 written:2 joaquim:1 john:1 indistinguishably:2 chicago:2 distant:1 partition:1 shape:1 enables:2 hofmann:1 remove:1 plot:1 intelligence:4 selecte...
6,256
666
Transient Signal Detection with Neural Networks: The Search for the Desired Signal Jose C. Principe and Abir Zahalka Computational NeuroEngineering Laboratory Department of Electrical Engineering University of Florida, CSE 447 Gainesville, FL 32611 principe@synapse.ee.ufl.edu Abstract Matched filtering has been one of...
666 |@word seems:2 gainesville:1 configuration:4 outperforms:1 must:1 readily:2 happen:1 shape:8 analytic:1 extrapolating:1 discrimination:1 stationary:1 fewer:2 selected:1 short:1 lr:1 compo:1 detecting:1 node:5 cse:1 five:1 ladendorf:1 constructed:1 supply:1 consists:3 symp:1 olfactory:2 theoretically:1 nor:1 brain:2...
6,257
6,660
Hypothesis Transfer Learning via Transformation Functions Simon S. Du Carnegie Mellon University ssdu@cs.cmu.edu Jayanth Koushik Carnegie Mellon University jayanthkoushik@cmu.edu Barnab?s P?czos Carnegie Mellon University bapoczos@cs.cmu.edu Aarti Singh Carnegie Mellon University aartisingh@cmu.edu Abstract We consi...
6660 |@word neurophysiology:1 trial:3 middle:1 norm:3 proportion:1 nd:1 elisseeff:6 arti:1 shot:1 liu:2 contains:1 rkhs:3 ours:1 fa8750:1 outperforms:1 existing:3 reaction:3 scovel:1 written:1 john:2 sanjiv:1 happen:1 enables:1 krikamol:1 plot:5 stroop:3 medial:1 intelligence:3 greedy:1 cult:1 awg:9 toronto:1 simpler:3...
6,258
6,661
Controllable Invariance through Adversarial Feature Learning Qizhe Xie, Zihang Dai, Yulun Du, Eduard Hovy, Graham Neubig Language Technologies Institute Carnegie Mellon University {qizhex, dzihang, yulund, hovy, gneubig}@cs.cmu.edu Abstract Learning meaningful representations that maintain the content necessary for a...
6661 |@word cnn:1 repository:1 eliminating:3 open:2 pieter:1 seek:1 dramatic:1 shot:1 moment:3 contains:1 score:2 jimenez:1 denoting:1 ours:10 document:1 subword:2 fa8750:1 outperforms:2 existing:1 com:1 comparing:1 diederik:2 dx:3 attracted:1 john:1 enables:1 christian:2 remove:3 drop:1 interpretable:1 stationary:1 ge...
6,259
6,662
Convergence Analysis of Two-layer Neural Networks with ReLU Activation Yuanzhi Li Computer Science Department Princeton University yuanzhil@cs.princeton.edu Yang Yuan Computer Science Department Cornell University yangyuan@cs.cornell.edu Abstract In recent years, stochastic gradient descent (SGD) based techniques has...
6662 |@word trial:2 version:1 eliminating:1 polynomial:6 seems:1 norm:11 open:2 d2:1 simulation:2 tried:1 decomposition:2 sgd:26 solid:1 arous:2 initial:6 contains:2 kurt:1 err:1 comparing:1 activation:18 diederik:1 john:1 hanie:2 plot:1 update:2 v:1 alone:1 pascanu:3 node:1 hyperplanes:1 sigmoidal:3 org:1 zhang:4 beco...
6,260
6,663
Doubly Accelerated Stochastic Variance Reduced Dual Averaging Method for Regularized Empirical Risk Minimization Tomoya Murata NTT DATA Mathematical Systems Inc. , Tokyo, Japan murata@msi.co.jp Taiji Suzuki Department of Mathematical Informatics Graduate School of Information Science and Technology, The University of ...
6663 |@word version:1 briefly:1 middle:2 norm:1 seems:1 nd:2 pg:16 unstably:1 pick:2 sgd:5 reduction:10 initial:3 tuned:3 past:1 existing:1 current:2 numerical:4 enables:2 update:3 intelligence:2 xk:13 steepest:1 core:1 zhang:4 mathematical:4 direct:6 become:1 symposium:1 ik:3 yuan:1 prove:1 doubly:3 nlog2:1 introducto...
6,261
6,664
Langevin Dynamics with Continuous Tempering for Training Deep Neural Networks Nanyang Ye University of Cambridge Cambridge, United Kingdom yn272@cam.ac.uk Zhanxing Zhu Center for Data Science, Peking University Beijing Institute of Big Data Research (BIBDR) Beijing, China zhanxing.zhu@pku.edu.cn Rafal K.Mantiuk Unive...
6664 |@word polynomial:3 simulation:3 crucially:1 covariance:1 sgd:15 arous:1 configuration:3 united:2 existing:3 current:1 comparing:1 activation:2 numerical:3 sdes:5 designed:2 plot:1 update:4 stationary:6 greedy:1 selected:1 accordingly:1 trapping:4 hamiltonian:5 prize:1 provides:1 completeness:1 pascanu:1 firstly:1...
6,262
6,665
Efficient Online Linear Optimization with Approximation Algorithms Dan Garber Technion - Israel Institute of Technology dangar@technion.ac.il Abstract We revisit the problem of online linear optimization in case the set of feasible actions is accessible through an approximated linear optimization oracle with a factor ...
6665 |@word briefly:1 version:4 achievable:1 polynomial:1 seems:3 norm:1 nd:1 d2:6 decomposition:4 jacob:1 q1:2 incurs:3 reduction:2 celebrated:3 etric:1 document:1 current:3 luo:1 yet:1 kft:2 readily:1 must:1 ligett:2 update:1 greedy:1 selected:1 prohibitive:2 instantiate:1 kyk:1 item:1 plane:1 xk:1 vanishing:1 comple...
6,263
6,666
Geometric Descent Method for Convex Composite Minimization Shixiang Chen1 , Shiqian Ma2 , and Wei Liu3 1 Department of SEEM, The Chinese University of Hong Kong, Hong Kong 2 Department of Mathematics, UC Davis, USA 3 Tencent AI Lab, China Abstract In this paper, we extend the geometric descent method recently propose...
6666 |@word kong:2 repository:1 briefly:2 version:2 middle:2 norm:3 stronger:1 c0:3 dekker:4 open:1 tr:4 reduction:1 bai:1 initial:7 contains:2 series:1 past:3 comparing:2 tackling:1 written:1 numerical:9 implying:1 selected:1 xk:77 vanishing:1 ck2:1 successive:1 mathematical:2 differential:1 prove:3 introductory:1 int...
6,264
6,667
Diffusion Approximations for Online Principal Component Estimation and Global Convergence Chris Junchi Li Mengdi Wang Princeton University Department of Operations Research and Financial Engineering, Princeton, NJ 08544 {junchil,mengdiw}@princeton.edu Tong Zhang Tencent AI Lab Shennan Ave, Nanshan District, Shenzhen, ...
6667 |@word version:1 briefly:1 polynomial:1 norm:1 stronger:1 nd:3 d2:1 simulation:3 covariance:7 decomposition:6 sgd:4 thereby:1 tr:3 initial:8 liu:2 denoting:1 existing:2 written:1 john:1 analytic:1 plot:2 drop:1 update:1 clumping:1 stationary:8 xk:1 beginning:1 short:1 blei:1 provides:3 characterization:6 iterates:...
6,265
6,668
Avoiding Discrimination through Causal Reasoning Niki Kilbertus?? nkilbertus@tue.mpg.de Moritz Hardt? hardt@berkeley.edu Mateo Rojas-Carulla?? mrojas@tue.mpg.de Dominik Janzing? janzing@tue.mpg.de Giambattista Parascandolo?? gparascandolo@tue.mpg.de Bernhard Sch?olkopf? bs@tue.mpg.de ? Max Planck Institute for Int...
6668 |@word trial:1 middle:1 version:2 manageable:1 judgement:1 instrumental:2 justice:2 tedious:1 sex:1 calculus:1 willing:1 seek:1 p0:10 asks:1 recursively:1 venkatasubramanian:2 substitution:1 score:4 exclusively:1 sendhil:1 envision:1 bilal:3 past:1 existing:4 mishra:1 comparing:2 qureshi:1 manuel:2 yet:2 assigning...
6,266
6,669
Nonparametric Online Regression while Learning the Metric Ilja Kuzborskij EPFL Switzerland ilja.kuzborskij@gmail.com Nicol`o Cesa-Bianchi Dipartimento di Informatica Universit`a degli Studi di Milano Milano 20135, Italy nicolo.cesa-bianchi@unimi.it Abstract We study algorithms for online nonparametric regression tha...
6669 |@word mild:2 determinant:3 version:1 norm:5 km:7 bn:1 covariance:1 pick:1 incurs:2 versatile:1 initial:1 contains:1 rkhs:1 past:2 current:3 com:1 gmail:1 bd:1 must:1 subsequent:1 gerchinovitz:2 update:2 discrimination:1 intelligence:1 beginning:1 core:1 provides:1 simpler:1 outerproduct:2 mathematical:1 along:5 c...
6,267
667
Deriving Receptive Fields Using An Optimal Encoding Criterion Ralph Linsker IBM T. J. Watson Research Center P. O. Box 218, Yorktown Heights, NY 10598 Abstract An information-theoretic optimization principle ('infomax') has previously been used for unsupervised learning of statistical regularities in an input ensembl...
667 |@word version:1 nd:1 heuristically:2 seek:1 covariance:12 tr:1 carry:1 substitution:1 contains:2 si:3 yet:1 perturbative:1 fn:1 numerical:1 remove:1 plot:3 progressively:1 aside:1 v:5 fewer:2 accordingly:1 provides:1 node:16 location:1 height:1 mathematical:1 along:1 pairwise:3 expected:1 indeed:2 rapid:1 roughly:...
6,268
6,670
Recycling Privileged Learning and Distribution Matching for Fairness Novi Quadrianto? Predictive Analytics Lab (PAL) University of Sussex Brighton, United Kingdom n.quadrianto@sussex.ac.uk Viktoriia Sharmanska Department of Computing Imperial College London London, United Kingdom sharmanska.v@gmail.com Abstract Equi...
6670 |@word compression:2 polynomial:1 fairer:2 p0:2 harder:1 reduction:4 initial:1 venkatasubramanian:2 ndez:3 score:1 united:2 moment:1 daniel:2 contains:1 rkhs:2 sendhil:1 genetic:2 ours:2 bilal:3 existing:6 current:2 comparing:1 contextual:1 manuel:2 com:1 yet:1 gmail:1 must:4 gpu:1 universality:1 john:3 additive:3...
6,269
6,671
Safe and Nested Subgame Solving for Imperfect-Information Games Noam Brown Computer Science Department Carnegie Mellon University Pittsburgh, PA 15217 noamb@cs.cmu.edu Tuomas Sandholm Computer Science Department Carnegie Mellon University Pittsburgh, PA 15217 sandholm@cs.cmu.edu Abstract In imperfect-information gam...
6671 |@word version:4 proportion:2 szafron:2 vi1:2 stronger:1 tried:1 decomposition:3 pick:1 thereby:2 reduction:2 initial:4 contains:1 past:6 outperforms:1 current:1 guez:1 must:10 john:1 subsequent:2 christian:3 sponsored:1 v:1 alone:1 intelligence:15 guess:6 warmuth:1 libratus:2 prize:1 aja:1 consulting:1 node:41 ma...
6,270
6,672
Unsupervised Image-to-Image Translation Networks Ming-Yu Liu, Thomas Breuel, Jan Kautz NVIDIA {mingyul,tbreuel,jkautz}@nvidia.com Abstract Unsupervised image-to-image translation aims at learning a joint distribution of images in different domains by using images from the marginal distributions in individual domains....
6672 |@word trial:1 version:2 nd:1 cha:1 d2:15 eng:1 q1:14 shot:1 harder:3 liu:3 configuration:2 existing:1 current:1 com:3 z2:18 luo:1 realize:1 realistic:4 designed:1 update:3 alone:1 generative:13 gan2:3 bissacco:1 zhang:1 wierstra:1 qualitative:1 consists:3 lopez:1 mingyuliutw:2 wild:1 ex2:1 aitken:1 cheetah:2 mult...
6,271
6,673
Coded Distributed Computing for Inverse Problems Yaoqing Yang, Pulkit Grover and Soummya Kar Carnegie Mellon University {yyaoqing, pgrover, soummyak}@andrew.cmu.edu Abstract Computationally intensive distributed and parallel computing is often bottlenecked by a small set of slow workers known as stragglers. In this p...
6673 |@word version:3 inversion:1 polynomial:1 norm:2 replicate:4 nd:1 vi1:1 vldb:2 covariance:4 pick:2 carry:1 mcauley:1 ld:5 liu:2 initial:7 outperforms:1 existing:4 current:1 comparing:1 si:2 assigning:1 yet:1 written:2 fn:2 additive:1 j1:3 designed:2 selected:1 xk:1 short:2 iterates:1 provides:2 node:4 preference:1...
6,272
6,674
A Screening Rule for `1-Regularized Ising Model Estimation Zhaobin Kuang1 , Sinong Geng2 , David Page3 University of Wisconsin zkuang@wisc.edu1 , sgeng2@wisc.edu2 , page@biostat.wisc.edu3 Abstract We discover a screening rule for `1 -regularized Ising model estimation. The simple closed-form screening rule is a neces...
6674 |@word trial:4 rising:1 polynomial:1 stronger:1 eliminating:1 nd:5 twelfth:1 checkable:1 prominence:1 covariance:8 hsieh:2 attainable:1 pick:1 contrastive:1 moment:10 necessity:1 liu:17 contains:1 safeguarded:1 karger:2 initial:1 configuration:1 series:3 current:1 comparing:1 luo:2 yet:1 readily:2 stemming:1 parti...
6,273
6,675
Improved Dynamic Regret for Non-degenerate Functions Lijun Zhang? , Tianbao Yang? , Jinfeng Yi? , Rong Jin? , Zhi-Hua Zhou? National Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China ? Department of Computer Science, The University of Iowa, Iowa City, USA ? AI Foundations Lab, IBM Thomas ...
6675 |@word polynomial:1 achievable:1 norm:1 nd:1 open:2 gradual:1 bellevue:1 incurs:1 past:1 current:3 com:2 z2:1 dikin:1 must:1 realize:1 update:4 stationary:2 intelligence:1 warmuth:2 chiang:2 successive:2 zhang:2 height:1 along:1 differential:1 clairvoyant:1 prove:2 naor:1 introductory:1 redefine:2 polyhedral:1 int...
6,274
6,676
Learning Efficient Object Detection Models with Knowledge Distillation Guobin Chen1,2 Wongun Choi1 Xiang Yu1 Tony Han2 Manmohan Chandraker1,3 1 2 3 NEC Labs America University of Missouri University of California, San Diego Abstract Despite significant accuracy improvement in convolutional neural networks (CNN) based...
6676 |@word cnn:16 middle:1 compression:14 stronger:1 norm:2 loading:1 seems:1 everingham:1 r:4 rgb:1 decomposition:5 thereby:3 harder:3 shot:1 configuration:1 lightweight:2 score:2 trainval:2 liu:2 tuned:2 interestingly:1 outperforms:1 freitas:1 blank:1 activation:1 written:1 gpu:3 subsequent:1 romero:3 informative:1 ...
6,275
6,677
One-Sided Unsupervised Domain Mapping Sagie Benaim1 and Lior Wolf1,2 1 The Blavatnik School of Computer Science , Tel Aviv University, Israel 2 Facebook AI Research Abstract In unsupervised domain mapping, the learner is given two unmatched datasets A and B. The goal is to learn a mapping GAB that translates a sample...
6677 |@word cnn:1 version:2 fcns:1 seems:3 norm:1 cha:1 rgb:3 d0k:5 initial:1 configuration:1 liu:3 score:6 ours:1 hyunsoo:1 deconvolutional:2 outperforms:3 existing:2 com:1 comparing:2 luo:2 activation:5 numerical:3 realistic:2 distant:1 happen:1 shape:1 christian:1 remove:1 v:4 generative:9 selected:1 half:6 alec:1 b...
6,276
6,678
Deep Mean-Shift Priors for Image Restoration Siavash A. Bigdeli University of Bern bigdeli@inf.unibe.ch Meiguang Jin University of Bern jin@inf.unibe.ch Paolo Favaro University of Bern favaro@inf.unibe.ch Matthias Zwicker University of Bern, and University of Maryland, College Park zwicker@cs.umd.edu Abstract In t...
6678 |@word cnn:2 version:1 seems:1 underline:1 rgb:1 p0:5 set5:2 initial:1 substitution:1 daniel:3 ours:11 interestingly:1 reaction:1 com:1 blur:2 remove:1 designed:2 update:4 half:3 selected:1 intelligence:5 rudin:1 kyoung:1 core:1 provides:1 location:1 simpler:1 zhang:9 favaro:5 hazirbas:1 constructed:1 qualitative:...
6,277
6,679
Greedy Algorithms for Cone Constrained Optimization with Convergence Guarantees Francesco Locatello MPI for Intelligent Systems - ETH Zurich Michael Tschannen ETH Zurich locatelf@ethz.ch michaelt@nari.ee.ethz.ch Gunnar R?tsch ETH Zurich Martin Jaggi EPFL raetsch@inf.ethz.ch martin.jaggi@epfl.ch Abstract Greedy...
6679 |@word repository:1 version:1 mri:2 briefly:2 norm:13 nd:1 open:1 crucially:2 hsieh:1 decomposition:2 kz1:1 thereby:2 boundedness:1 reduction:1 initial:1 cyclic:1 cristina:1 electronics:1 selecting:1 contains:1 ap1:1 kahles:1 daniel:5 document:1 amp:10 hyunsoo:1 frankwolfe:1 existing:5 kx0:2 ka:7 current:2 z2:2 lu...
6,278
668
Single-iteration Threshold Hamming Networks Eytan Ruppin Isaac Meilijson Moshe Sipper School of Mathematical Sciences Raymond and Beverly Sackler Faculty of Exact Sciences Tel Aviv University, 69978 Tel Aviv, Israel Abstract We analyze in detail the performance of a Hamming network classifying inputs that are distor...
668 |@word effect:1 predicted:2 version:2 faculty:1 rk2:1 proportion:2 hence:1 society:1 moshe:1 correct:5 laboratory:1 eng:1 tr:1 bin:3 subnet:9 distance:2 berlin:1 rat:1 initial:1 capacity:4 selecting:1 opt:1 tuned:1 renewed:1 biological:1 evident:1 tn:1 performs:2 fj:2 practically:1 activation:2 normal:4 dx:1 exp:11...
6,279
6,680
A New Theory for Matrix Completion Guangcan Liu? Qingshan Liu? Xiao-Tong Yuan? School of Information & Control, Nanjing University of Information Science & Technology NO 219 Ningliu Road, Nanjing, Jiangsu, China, 210044 {gcliu,qsliu,xtyuan}@nuist.edu.cn Abstract Prevalent matrix completion theories reply on an ass...
6680 |@word mild:3 trial:2 middle:1 polynomial:1 norm:17 km:3 shuicheng:3 seek:2 simulation:2 theran:1 decomposition:1 thereby:1 klk:2 necessity:1 liu:13 series:1 configuration:1 selecting:1 hereafter:1 contains:1 seriously:1 initial:2 interestingly:3 bc:1 daniel:1 existing:4 recovered:1 comparing:4 luo:1 must:1 john:1...
6,280
6,681
Robust Hypothesis Test for Nonlinear Effect with Gaussian Processes Jeremiah Zhe Liu, Brent Coull Department of Biostatistics Harvard University Cambridge, MA 02138 {zhl112@mail, bcoull@hsph}.harvard.edu Abstract This work constructs a hypothesis test for detecting whether an data-generating function h : Rp ? R belong...
6681 |@word trial:1 version:1 polynomial:4 norm:2 open:1 km:1 hu:1 simulation:7 covariance:1 decomposition:2 elisseeff:2 tr:4 solid:4 carry:1 moment:1 liu:2 contains:2 score:3 selecting:1 genetic:1 rkhs:7 intake:2 yet:2 subsequent:2 additive:2 partition:1 selected:3 rp1:3 ith:1 record:1 detecting:4 zhang:1 unbiasedly:1...
6,281
6,682
Lower bounds on the robustness to adversarial perturbations Jonathan Peck1,2 , Joris Roels2,3 , Bart Goossens3 , and Yvan Saeys1,2 1 Department of Applied Mathematics, Computer Science and Statistics, Ghent University, Ghent, 9000, Belgium 2 Data Mining and Modeling for Biomedicine, VIB Inflammation Research Center, G...
6682 |@word moosavi:1 cnn:1 eliminating:1 norm:18 nd:2 c0:2 rgb:1 reduction:1 liu:1 series:1 score:1 bc:3 document:1 guadarrama:1 comparing:1 com:1 protection:1 activation:1 intriguing:1 must:2 written:2 drop:1 designed:1 bart:1 intelligence:1 oldest:1 krkf:3 characterization:5 provides:2 location:1 toronto:1 tahoe:1 b...
6,282
6,683
Minimizing a Submodular Function from Samples Eric Balkanski Harvard University ericbalkanski@g.harvard.edu Yaron Singer Harvard University yaron@seas.harvard.edu Abstract In this paper we consider the problem of minimizing a submodular function from training data. Submodular functions can be efficiently minimized an...
6683 |@word private:1 polynomial:9 seems:2 norm:1 nd:1 seek:3 tat:1 cla:1 contains:3 united:1 surprising:1 si:17 must:2 additive:7 partition:4 drop:2 selected:1 tahoe:1 mathematical:1 constructed:6 direct:1 symposium:2 prove:1 combine:2 introduce:1 pairwise:1 indeed:1 hardness:3 behavior:1 multi:3 pitfall:1 paclearnabl...
6,283
6,684
Introspective Classification with Convolutional Nets Long Jin UC San Diego longjin@ucsd.edu Justin Lazarow UC San Diego jlazarow@ucsd.edu Zhuowen Tu UC San Diego ztu@ucsd.edu Abstract We propose introspective convolutional networks (ICN) that emphasize the importance of having convolutional neural networks empowere...
6684 |@word cnn:52 seems:1 logit:1 simulation:1 contrastive:2 sgd:20 carry:5 reduction:1 initial:4 liu:2 series:1 ours:8 past:2 existing:3 outperforms:2 current:2 icn:69 comparing:1 contextual:1 com:2 stemmed:2 activation:1 dx:1 partition:1 informative:1 grumman:1 designed:2 plot:1 progressively:4 update:5 v:21 generat...
6,284
6,685
Label Distribution Learning Forests Wei Shen1,2 , Kai Zhao1 , Yilu Guo1 , Alan Yuille2 Key Laboratory of Specialty Fiber Optics and Optical Access Networks, Shanghai Institute for Advanced Communication and Data Science, School of Communication and Information Engineering, Shanghai University 2 Department of Computer ...
6685 |@word version:1 eng:1 contains:1 ours:6 longitudinal:1 past:1 existing:3 current:4 com:1 guadarrama:1 comparing:1 gmail:1 assigning:1 john:1 hou:2 numerical:1 additive:1 partition:1 shape:1 enables:2 update:18 v:2 alone:6 greedy:1 leaf:31 selected:1 cook:1 smith:1 short:1 boosting:3 node:57 cse:1 zhang:3 rc:1 con...
6,285
6,686
Unsupervised learning of object frames by dense equivariant image labelling James Thewlis1 Hakan Bilen2 1 Andrea Vedaldi1 2 Visual Geometry Group University of Oxford {jdt,vedaldi}@robots.ox.ac.uk School of Informatics University of Edinburgh hbilen@ed.ac.uk Abstract One of the key challenges of visual perceptio...
6686 |@word cnn:8 middle:4 version:1 dalal:1 norm:1 triggs:1 seitz:1 shuicheng:1 rgb:1 decomposition:2 jacob:1 brightness:1 inpainting:2 harder:1 liu:3 score:2 daniel:1 interestingly:1 animated:2 existing:1 com:1 luo:3 must:7 gavves:1 subsequent:1 realistic:1 shape:2 enables:1 remove:1 plot:1 localise:1 generative:2 se...
6,286
6,687
Compression-aware Training of Deep Networks Mathieu Salzmann EPFL - CVLab Lausanne, Switzerland mathieu.salzmann@epfl.ch Jose M. Alvarez Toyota Research Institute Los Altos, CA 94022 jose.alvarez@tri.global Abstract In recent years, great progress has been made in a variety of application domains thanks to the devel...
6687 |@word kohli:1 compression:34 stronger:1 seems:1 norm:2 hu:1 seek:4 accounting:3 decomposition:3 simplifying:1 sgd:1 incurs:1 reduction:6 initial:5 configuration:3 liu:2 salzmann:3 tuned:4 ours:4 interestingly:2 existing:4 freitas:1 current:3 activation:5 yet:1 guez:1 written:2 gpu:2 numerical:1 alphanumeric:1 rem...
6,287
6,688
Multiscale Semi-Markov Dynamics for Intracortical Brain-Computer Interfaces Daniel J. Milstein ? daniel_milstein@alumni.brown.edu John D. Simeral ? ? john_simeral@brown.edu Jason L. Pacheco ? Leigh R. Hochberg ? ? ? pachecoj@mit.edu leigh_hochberg@brown.edu Beata Jarosiewicz k ? ?? beataj@stanford.edu Erik B. S...
6688 |@word trial:10 open:1 tried:1 covariance:6 eng:1 thereby:1 accommodate:1 initial:1 configuration:7 series:1 liu:1 selecting:1 united:1 daniel:1 tuned:5 bc:2 past:1 existing:1 reaction:2 current:3 ka:14 discretization:1 yet:1 written:1 must:2 john:1 enables:1 motor:10 designed:4 plot:2 update:1 intelligence:2 cue:...
6,288
6,689
PredRNN: Recurrent Neural Networks for Predictive Learning using Spatiotemporal LSTMs Yunbo Wang School of Software Tsinghua University wangyb15@mails.tsinghua.edu.cn Jianmin Wang School of Software Tsinghua University jimwang@tsinghua.edu.cn Mingsheng Long? School of Software Tsinghua University mingsheng@tsinghua.e...
6689 |@word trial:1 cnn:3 version:1 wmf:1 wco:3 cox:1 bf:3 open:1 tried:1 rgb:4 tnlist:1 recursively:3 reduction:1 contains:2 score:2 past:1 existing:1 outperforms:3 current:7 guadarrama:1 anne:1 activation:1 devin:1 concatenate:1 happen:1 blur:3 subsequent:1 shape:4 visibility:1 designed:1 update:1 fund:1 occlude:1 ge...
6,289
669
Hidden Markov Model Induction by Bayesian Model Merging Andreas Stolcke*'** *Computer Science Division University of California Berkeley, CA 94720 stolcke@icsi.berkeley.edu Stephen Omohundro" **International Computer Science Institute 1947 Center Street, Suite 600 Berkeley, CA 94704 om@icsi.berkeley.edu Abstract Thi...
669 |@word trial:4 version:1 briefly:1 seems:1 replicate:1 accounting:2 tr:2 accommodate:1 carry:2 initial:12 contains:1 series:1 bc:7 interestingly:1 soules:2 current:2 must:1 cruz:1 shape:2 drop:1 cfo:1 update:3 alone:1 greedy:2 fewer:2 intelligence:1 smith:2 provides:1 complication:1 simpler:2 mathematical:1 along:6...
6,290
6,690
Detrended Partial Cross Correlation for Brain Connectivity Analysis Jaime S Ide? Yale University New Haven, CT 06519 jaime.ide@yale.edu Fabio A Cappabianco Federal University of Sao Paulo S.J. dos Campos, 12231, Brazil cappabianco@unifesp.br Fabio A Faria Federal University of Sao Paulo S.J. dos Campos, 12231, Brazi...
6690 |@word middle:1 version:1 mri:2 polynomial:1 approved:1 open:1 hu:2 simulation:5 seek:2 covariance:5 fifteen:1 tr:2 shot:1 reduction:4 initial:1 series:18 interestingly:1 dubourg:1 past:1 reaction:1 current:3 ka:1 activation:7 connectomics:2 realistic:2 additive:1 oxygenation:1 analytic:1 motor:5 hypothesize:3 atl...
6,291
6,691
Contrastive Learning for Image Captioning Bo Dai Dahua Lin Department of Information Engineering, The Chinese University of Hong Kong db014@ie.cuhk.edu.hk dhlin@ie.cuhk.edu.hk Abstract Image captioning, a popular topic in computer vision, has achieved substantial progress in recent years. However, the distinctivenes...
6691 |@word kong:3 cnn:1 compression:1 stronger:7 hyv:1 jacob:1 contrastive:12 pg:2 mention:1 ytn:1 initial:1 liu:1 contains:4 score:1 att:2 ours:3 suppressing:1 existing:1 guadarrama:1 comparing:2 nt:5 luo:1 written:1 readily:2 john:1 neuraltalk2:9 numerical:1 partition:1 periodically:2 designed:1 fund:1 alone:1 gener...
6,292
6,692
Safe Model-based Reinforcement Learning with Stability Guarantees Felix Berkenkamp Department of Computer Science ETH Zurich befelix@inf.ethz.ch Matteo Turchetta Department of Computer Science, ETH Zurich matteotu@inf.ethz.ch Angela P. Schoellig Institute for Aerospace Studies University of Toronto schoellig@utias.ut...
6692 |@word middle:1 norm:3 mockus:1 c0:3 open:1 pieter:2 simulation:1 linearized:1 covariance:1 jacob:1 schoellig:5 evaluating:2 pick:1 sgd:1 thereby:1 initial:16 ndez:1 series:1 contains:2 selecting:2 daniel:2 rkhs:1 steiner:1 current:8 discretization:7 com:1 anne:1 activation:1 intriguing:1 must:1 john:2 devin:1 aca...
6,293
6,693
Online Multiclass Boosting Young Hun Jung Jack Goetz Department of Statistics University of Michigan Ann Arbor, MI 48109 {yhjung, jrgoetz, tewaria}@umich.edu Ambuj Tewari Abstract Recent work has extended the theoretical analysis of boosting algorithms to multiclass problems and to online settings. However, the mul...
6693 |@word repository:1 version:3 polynomial:1 stronger:1 norm:1 cochleagram:1 open:1 hu:2 pick:2 incurs:1 mention:1 tr:1 solid:1 series:1 contains:1 past:1 subjective:1 existing:1 current:1 outperforms:1 com:1 beygelzimer:12 luo:1 si:14 yet:1 mushroom:1 enables:1 update:2 v:1 intelligence:3 guess:4 warmuth:2 ith:2 sh...
6,294
6,694
Matching on Balanced Nonlinear Representations for Treatment Effects Estimation Yun Fu Northeastern University Boston, MA yunfu@ece.neu.edu Sheng Li Adobe Research San Jose, CA sheli@adobe.com Abstract Estimating treatment effects from observational data is challenging due to the missing counterfactuals. Matching is ...
6694 |@word exploitation:1 version:1 prognostic:1 nd:1 johansson:2 essay:1 seek:1 simulation:1 lpp:7 tr:6 klk:1 carry:1 reduction:2 hunting:1 liu:1 contains:6 score:17 att:16 selecting:1 series:2 genetic:2 rkhs:4 document:4 ours:1 daniel:1 past:1 existing:5 outperforms:1 current:1 com:1 comparing:1 nt:7 protection:1 sc...
6,295
6,695
Learning Overcomplete HMMs Vatsal Sharan Stanford University vsharan@stanford.edu Sham Kakade University of Washington sham@cs.washington.edu Percy Liang Stanford University pliang@cs.stanford.edu Gregory Valiant Stanford University valiant@stanford.edu Abstract We study the problem of learning overcomplete HMMs?t...
6695 |@word mild:2 trial:1 koopmans:1 polynomial:22 seems:4 norm:3 stronger:1 open:5 d2:2 simulation:2 crucially:1 decomposition:15 p0:2 pick:1 recursively:1 carry:2 reduction:1 moment:27 cyclic:2 contains:1 initial:6 necessity:1 denoting:1 document:2 ours:3 past:3 recovered:1 surprising:1 must:4 written:1 additive:1 d...
6,296
6,696
GP CaKe: Effective brain connectivity with causal kernels Luca Ambrogioni Radboud University l.ambrogioni@donders.ru.nl Max Hinne Radboud University m.hinne@donders.ru.nl Marcel A. J. van Gerven Radboud University m.vangerven@donders.ru.nl Eric Maris Radboud University e.maris@donders.ru.nl Abstract A fundamental g...
6696 |@word trial:2 mri:1 version:1 polynomial:2 coombes:1 d2:2 simulation:5 propagate:1 covariance:24 accounting:1 decomposition:5 thereby:1 moment:1 series:12 efficacy:1 selecting:1 outperforms:1 imaginary:1 recovered:2 mari:4 discretization:1 current:2 scaffolding:1 must:1 written:1 connectomics:1 kiebel:1 realistic...
6,297
6,697
Decoupling ?when to update? from ?how to update? Eran Malach School of Computer Science The Hebrew University, Israel eran.malach@mail.huji.ac.il Shai Shalev-Shwartz School of Computer Science The Hebrew University, Israel shais@cs.huji.ac.il Abstract Deep learning requires data. A useful approach to obtain data is t...
6697 |@word version:6 seems:2 open:1 jacob:2 palso:1 sgd:1 mention:1 tr:1 moment:1 initial:4 configuration:1 contains:1 liu:1 daniel:1 past:1 existing:2 outperforms:1 current:2 comparing:2 com:2 bootkrajang:2 goldberger:2 tackling:1 must:3 realistic:1 happen:2 klaas:1 christian:1 atlas:2 mislabelled:1 update:57 sukhbaa...
6,298
6,698
Self-Normalizing Neural Networks G?nter Klambauer Thomas Unterthiner Andreas Mayr Sepp Hochreiter LIT AI Lab & Institute of Bioinformatics, Johannes Kepler University Linz A-4040 Linz, Austria {klambauer,unterthiner,mayr,hochreit}@bioinf.jku.at Abstract Deep Learning has revolutionized vision via convolutional neu...
6698 |@word mild:1 arabic:1 repository:3 version:1 cnn:1 norm:6 stronger:3 nd:1 contraction:5 mammal:1 sgd:2 thereby:1 delgado:1 moment:4 initial:1 configuration:1 ndez:1 series:1 jku:1 bradley:1 activation:51 readily:1 subsequent:1 shape:1 enables:1 hochreit:1 update:1 selected:3 device:1 proficient:1 marine:1 vanishi...
6,299
6,699
Learning to Pivot with Adversarial Networks Gilles Louppe New York University g.louppe@nyu.edu Michael Kagan SLAC National Accelerator Laboratory makagan@slac.stanford.edu Kyle Cranmer New York University kyle.cranmer@nyu.edu Abstract Several techniques for domain adaptation have been proposed to account for differ...
6699 |@word illustrating:1 middle:2 stronger:1 open:2 simulation:1 propagate:1 eng:1 thereby:1 carry:1 venkatasubramanian:1 phy:2 initial:1 score:6 efficacy:1 pub:2 salzmann:1 interestingly:1 com:1 activation:8 written:1 evans:2 shape:1 plot:3 atlas:5 update:2 generative:4 selected:1 accordingly:1 lr:24 characterizatio...