Unnamed: 0
int64
0
7.24k
id
int64
1
7.28k
raw_text
stringlengths
9
124k
vw_text
stringlengths
12
15k
6,000
6,428
The Sound of APALM Clapping: Faster Nonsmooth Nonconvex Optimization with Stochastic Asynchronous PALM Damek Davis and Madeleine Udell Cornell University {dsd95,mru8}@cornell.edu Brent Edmunds University of California, Los Angeles brent.edmunds@math.ucla.edu Abstract We introduce the Stochastic Asynchronous Proximal...
6428 |@word mild:1 version:2 norm:1 vldb:1 semicontinuous:1 linearized:4 decomposition:1 hsieh:1 carry:1 reduction:1 liu:3 existing:1 kwjk:1 current:1 written:1 devin:1 numerical:3 enables:1 remove:1 concert:1 update:13 v:4 stationary:2 slowing:1 accordingly:1 xk:22 short:1 core:3 iterates:5 math:1 bittorf:1 zhang:1 un...
6,001
6,429
Optimistic Bandit Convex Optimization Mehryar Mohri Courant Institute and Google 251 Mercer Street New York, NY 10012 Scott Yang Courant Institute 251 Mercer Street New York, NY 10012 mohri@cims.nyu.edu yangs@cims.nyu.edu Abstract 1 We introduce the general and powerful scheme of predicting information re-use in ...
6429 |@word mild:2 exploitation:1 version:2 achievable:1 polynomial:7 norm:5 dekel:11 c0:13 open:1 d2:8 linearized:1 decomposition:2 incurs:1 reduction:3 ftrl:13 past:1 existing:1 ka:3 current:1 dikin:1 designed:2 update:5 v:1 preemptively:1 greedy:1 selected:2 guess:1 grfp:1 zhang:2 along:1 prove:1 shorthand:1 introdu...
6,002
643
Extended Regularization Methods for N onconvergent Model Selection W. Finnoff, F. Hergert and H.G. Zimmermann Siemens AG, Corporate Research and Development Otto-Hahn-Ring 6 8000 Munich 83, Fed. Rep. Germany Abstract Many techniques for model selection in the field of neural networks correspond to well established st...
643 |@word version:4 seems:1 instrumental:1 nd:1 simulation:3 reduction:1 initial:1 contains:2 comparing:1 activation:2 additive:1 remove:2 designed:1 update:2 precaution:1 alone:1 selected:1 n_o:2 parametrization:2 short:1 detecting:1 constructed:1 consists:2 fitting:2 baldi:1 manner:1 deteriorate:1 frequently:1 brain...
6,003
6,430
Linear dynamical neural population models through nonlinear embeddings Yuanjun Gao? 1 , Evan Archer?12 , Liam Paninski12 , John P. Cunningham12 Department of Statistics1 and Grossman Center2 Columbia University New York, NY, United States yg2312@columbia.edu, evan@stat.columbia.edu, liam@stat.columbia.edu, jpc2181@colu...
6430 |@word neurophysiology:1 trial:30 private:1 briefly:1 proportion:2 norm:1 busing:1 seek:1 simulation:8 covariance:3 decomposition:1 q1:3 datagenerating:1 thereby:1 carry:1 reduction:13 initial:1 series:1 contains:1 united:1 uncovered:1 ours:1 interestingly:1 outperforms:1 existing:1 recovered:2 comparing:1 nt:2 co...
6,004
6,431
Improved Error Bounds for Tree Representations of Metric Spaces Samir Chowdhury Department of Mathematics The Ohio State University Columbus, OH 43210 chowdhury.57@osu.edu Facundo M?moli Department of Mathematics Department of Computer Science and Engineering The Ohio State University Columbus, OH 43210 memoli@math.o...
6431 |@word polynomial:1 norm:1 open:1 bn:4 kent:1 invoking:3 contains:2 united:1 interestingly:1 past:1 existing:3 bitmap:1 current:1 ddim:1 dx:99 written:2 john:3 realize:1 fn:1 numerical:4 additive:12 partition:6 thrust:1 subsequent:1 noche:1 lemy:1 inspection:1 xk:1 smith:2 realizing:1 characterization:2 math:1 gx:...
6,005
6,432
Exact Recovery of Hard Thresholding Pursuit Xiao-Tong Yuan B-DAT Lab Nanjing University of Info. Sci.&Tech. Nanjing, Jiangsu, 210044, China xtyuan@nuist.edu.cn Ping Li?? Tong Zhang? ?Depart. of Statistics and ?Depart. of CS Rutgers University Piscataway, NJ, 08854, USA {pingli,tzhang}@stat.rutgers.edu Abstract The H...
6432 |@word mild:1 trial:1 determinant:1 norm:2 replicate:1 nd:1 open:3 gaussion:1 confirms:1 simulation:3 r:4 covariance:2 decomposition:1 bahmani:3 configuration:1 liu:1 ours:2 outperforms:1 existing:2 current:2 comparing:2 recovered:6 numerical:6 subsequent:1 designed:1 update:1 greedy:6 selected:1 intelligence:2 xk...
6,006
6,433
Spatiotemporal Residual Networks for Video Action Recognition Christoph Feichtenhofer Graz University of Technology Axel Pinz Graz University of Technology Richard P. Wildes York University, Toronto feichtenhofer@tugraz.at axel.pinz@tugraz.at wildes@cse.yorku.ca Abstract Two-stream Convolutional Networks (ConvNe...
6433 |@word multitask:1 middle:1 advantageous:1 underline:2 wla:1 rgb:5 sgd:3 thereby:2 tr:1 carry:1 moment:2 reduction:2 born:1 series:1 score:4 denoting:1 ours:2 interestingly:2 guadarrama:1 com:1 comparing:3 skipping:2 anne:1 activation:3 yet:3 gpu:1 readily:1 gavves:1 additive:3 shape:1 enables:1 christian:3 design...
6,007
6,434
Adaptive Smoothed Online Multi-Task Learning Keerthiram Murugesan? Carnegie Mellon University kmuruges@cs.cmu.edu Hanxiao Liu? Carnegie Mellon University hanxiaol@cs.cmu.edu Jaime Carbonell Carnegie Mellon University jgc@cs.cmu.edu Yiming Yang Carnegie Mellon University yiming@cs.cmu.edu Abstract This paper addres...
6434 |@word multitask:8 version:3 manageable:1 middle:2 advantageous:2 norm:1 stronger:1 dekel:4 briefly:1 nemirovsky:1 covariance:2 jacob:1 blender:1 keerthiram:1 moment:1 venkatasubramanian:1 liu:1 contains:1 score:1 ours:2 outperforms:2 existing:5 current:1 transferability:1 attracted:1 john:2 interpretable:1 update...
6,008
6,435
A Pseudo-Bayesian Algorithm for Robust PCA Tae-Hyun Oh1 Yasuyuki Matsushita2 In So Kweon1 David Wipf3? 1 Electrical Engineering, KAIST, Daejeon, South Korea 2 Multimedia Engineering, Osaka University, Osaka, Japan 3 Microsoft Research, Beijing, China thoh.kaist.ac.kr@gmail.com yasumat@ist.osaka-u.ac.jp iskweon@kaist.a...
6435 |@word mild:1 trial:1 determinant:2 version:2 norm:9 stronger:3 replicate:1 seek:1 accounting:1 covariance:1 decomposition:4 tr:8 accommodate:1 liu:2 selecting:1 zij:1 mag:1 egt:3 kweon:3 ours:1 amp:2 outperforms:1 existing:6 recovered:1 optim:1 com:2 yet:1 gmail:1 pcp:16 chu:1 must:4 exposing:1 distant:1 subseque...
6,009
6,436
SPALS: Fast Alternating Least Squares via Implicit Leverage Scores Sampling Dehua Cheng University of Southern California dehua.cheng@usc.edu Ioakeim Perros Georgia Institute of Technology perros@gatech.edu Richard Peng Georgia Institute of Technology rpeng@cc.gatech.edu Yan Liu University of Southern California yanli...
6436 |@word trial:1 version:1 polynomial:2 norm:3 decomposition:41 sgd:3 mcauley:1 reduction:2 initial:1 liu:4 contains:1 score:37 series:1 woodruff:4 tuned:1 ours:1 existing:2 comparing:4 com:1 must:1 numerical:7 subsequent:3 informative:1 kdd:2 enables:1 analytic:1 remove:1 interpretable:1 update:1 v:1 alone:1 fewer:...
6,010
6,437
Selective inference for group-sparse linear models Rina Foygel Barber Department of Statistics University of Chicago rina@uchicago.edu Fan Yang Department of Statistics University of Chicago fyang1@uchicago.edu Prateek Jain Microsoft Research India prajain@microsoft.com John Lafferty Depts. of Statistics and Compute...
6437 |@word trial:5 version:1 polynomial:1 stronger:1 norm:2 open:4 jacob:1 carry:1 initial:2 contains:2 series:3 selecting:1 interestingly:1 past:1 existing:1 com:1 nt:4 dx:1 must:2 john:1 chicago:3 numerical:4 partition:1 enables:2 designed:2 plot:2 update:4 greedy:1 selected:23 core:1 record:1 characterization:1 org...
6,011
6,438
Accelerating Stochastic Composition Optimization Mengdi Wang? , Ji Liu? , and Ethan X. Fang Princeton University, University of Rochester, Pennsylvania State University mengdiw@princeton.edu, ji.liu.uwisc@gmail.com, xxf13@psu.edu Abstract Consider the stochastic composition optimization problem where the objective is ...
6438 |@word version:1 norm:4 stronger:1 twelfth:1 open:1 r:4 simulation:5 pg:32 liu:12 contains:2 series:1 current:3 com:1 deteriorating:1 surprising:1 gmail:1 must:1 numerical:2 plot:1 update:5 juditsky:1 v:1 intelligence:1 kyk:2 xk:25 short:2 bwt:1 provides:3 iterates:1 zhang:2 mathematical:4 constructed:1 x0:1 expec...
6,012
6,439
Bayesian optimization under mixed constraints with a slack-variable augmented Lagrangian Victor Picheny MIAT, Universit? de Toulouse, INRA Castanet-Tolosan, France victor.picheny@toulouse.inra.fr Stefan Wild Argonne National Laboratory Argonne, IL, USA wildmcs.anl.gov Robert B. Gramacy Virginia Tech Blacksburg, VA, U...
6439 |@word mild:1 exploitation:2 version:6 stronger:1 proportion:5 nd:1 mockus:1 open:1 termination:1 simplifying:1 accounting:1 thereby:1 solid:4 accommodate:1 reduction:1 moment:1 ndez:1 contains:2 series:1 initial:6 denoting:1 outperforms:1 freitas:1 current:3 com:1 optim:8 written:2 readily:1 must:1 mesh:1 numeric...
6,013
644
? ? ?                "!    # $  "% & '   (  ! ? ? ? ? ? < ) B @ D A F C H E G ) < : ? = ; > 8 * 4 7 9 )  , * . + 0 2 / 4 1 6 3 5 IKJLJNM0OMQPSR? TV`,?{UX? WYW J[gZ]b \_? J l?^a`,ZmbNcN? d?r?cNe ??J c?W ?.v?f Zhw gKikjml bonqp tr s cLuYvxw P R ? w ` gon ...
644 |@word agf:1 km:1 hu:1 tr:1 v2o:1 n8:1 bc:4 ala:1 o2:1 ka:1 wd:2 chu:1 bd:3 fn:1 gv:1 e65:1 wlm:1 d5i:1 rts:1 lr:2 vxw:2 ooj:1 ry:1 phj:1 anj:1 q2:1 acbed:1 y3:1 nf:1 rm:1 hkm:1 ap:1 kml:2 au:1 lop:1 bi:1 qdq:1 vu:4 eut:1 z7:1 oqp:3 py:1 go:1 l:1 onqp:1 qc:2 s6:1 mq:1 tht:1 qq:3 pt:1 gm:2 enx:1 jk:1 gon:1 cy:2 wj:1...
6,014
6,440
Avoiding Imposters and Delinquents: Adversarial Crowdsourcing and Peer Prediction Jacob Steinhardt Stanford University Gregory Valiant Stanford University Moses Charikar Stanford University Abstract We consider a crowdsourcing model in which n workers are asked to rate the quality of n items previously generated by...
6440 |@word mild:1 version:2 briefly:1 polynomial:1 norm:17 seems:1 judgement:1 open:5 km:6 condon:2 jacob:1 dishonest:3 harder:1 reduction:1 liu:1 score:1 karger:2 neeman:4 bc:5 interestingly:1 ours:2 semirandom:7 imposter:1 tuned:1 recovered:2 ka:2 collude:3 must:3 john:1 partition:1 informative:1 enables:1 wanted:1 ...
6,015
6,441
Direct Feedback Alignment Provides Learning in Deep Neural Networks Arild N?kland Trondheim, Norway arild.nokland@gmail.com Abstract Artificial neural networks are most commonly trained with the back-propagation algorithm, where the gradient for learning is provided by back-propagating the error, layer by layer, from...
6441 |@word version:1 seems:3 twelfth:1 grey:1 propagate:1 linearized:1 contrastive:2 initial:7 configuration:2 daniel:1 com:1 surprising:1 activation:8 gmail:1 visible:1 distant:1 happen:1 plasticity:1 enables:1 x240:3 christian:1 update:26 intelligence:1 reciprocal:3 steepest:5 short:1 provides:3 relayed:2 org:1 zhan...
6,016
6,442
Computational and Statistical Tradeoffs in Learning to Rank Ashish Khetan and Sewoong Oh Department of ISE, University of Illinois at Urbana-Champaign Email: {khetan2,swoh}@illinois.edu Abstract For massive and heterogeneous modern data sets, it is of fundamental interest to provide guarantees on the accuracy of estim...
6442 |@word illustrating:1 middle:5 achievable:2 logit:1 c0:2 d2:3 willing:2 simulation:1 tr:3 reduction:1 moment:1 offering:2 e2b:1 khetan:2 bradley:1 comparing:1 yet:1 assigning:1 written:2 stemming:1 numerical:3 partition:11 confirming:2 predetermined:1 remove:1 treating:3 item:28 parkes:3 provides:10 node:1 prefere...
6,017
6,443
Gaussian Processes for Survival Analysis Tamara Fern?ndez Department of Statistics, University of Oxford. Oxford, UK. fernandez@stats.ox.ac.uk Nicol?s Rivera Department of Informatics, King?s College London. London, UK. nicolas.rivera@kcl.ac.uk Yee Whye Teh Department of Statistics, University of Oxford. Oxford, UK....
6443 |@word trial:3 cox:12 middle:2 version:1 inversion:3 seems:3 proportionality:1 covariance:2 p0:3 rivera:3 harder:1 initial:4 ndez:2 contains:2 score:59 series:1 denoting:1 current:2 elliptical:2 riihim:1 si:1 additive:2 numerical:2 drop:1 treating:1 update:4 stationary:5 generative:2 beginning:3 chile:1 accepting:...
6,018
6,444
Variational Information Maximization for Feature Selection Shuyang Gao Greg Ver Steeg Aram Galstyan University of Southern California, Information Sciences Institute gaos@usc.edu, gregv@isi.edu, galstyan@isi.edu Abstract Feature selection is one of the most fundamental problems in machine learning. An extensive body o...
6444 |@word madelon:2 repository:1 version:1 advantageous:1 stronger:2 d2:2 motoda:1 decomposition:8 citeseer:1 elisseeff:1 pick:1 recursively:2 wrapper:3 liu:2 lichman:1 selecting:1 jimenez:1 rightmost:1 outperforms:3 existing:7 spambase:1 current:1 past:2 com:1 written:2 john:3 cruz:1 plot:2 progressively:1 joy:1 gre...
6,019
6,445
Fast Algorithms for Robust PCA via Gradient Descent Xinyang Yi? Dohyung Park? Yudong Chen? Constantine Caramanis? ? ? The University of Texas at Austin Cornell University ? ? {yixy,dhpark,constantine}@utexas.edu yudong.chen@cornell.edu Abstract We consider the problem of Robust PCA in the fully and partially observed ...
6445 |@word polynomial:1 seems:1 norm:9 c0:2 d2:31 simulation:1 decomposition:8 contraction:2 minming:1 thereby:1 initial:1 liu:3 contains:2 series:2 daniel:2 denoting:1 woodruff:1 xinyang:2 existing:3 ksk1:1 ka:3 recovered:1 luo:1 yet:1 must:1 john:3 numerical:2 plot:3 designed:1 update:2 rd2:6 intelligence:1 prohibit...
6,020
6,446
Multimodal Residual Learning for Visual QA Jin-Hwa Kim Sang-Woo Lee Donghyun Kwak Min-Oh Heo Seoul National University {jhkim,slee,dhkwak,moheo}@bi.snu.ac.kr Jeonghee Kim Jung-Woo Ha Naver Labs, Naver Corp. {jeonghee.kim,jungwoo.ha}@navercorp.com Byoung-Tak Zhang Seoul National University & Surromind Robotics btzhan...
6446 |@word cnn:4 nd:1 open:10 shuicheng:1 jacob:1 slee:1 mengye:1 denoting:1 outperforms:1 existing:1 com:2 contextual:1 activation:1 yet:2 readily:1 ronald:1 latt:3 christian:1 treating:1 update:2 bart:1 sukhbaatar:1 intelligence:1 selected:2 half:1 num:5 provides:1 contribute:1 successive:1 firstly:1 zhang:3 qualita...
6,021
6,447
The Power of Optimization from Samples Eric Balkanski Harvard University ericbalkanski@g.harvard.edu Aviad Rubinstein University of California, Berkeley aviad@eecs.berkeley.edu Yaron Singer Harvard University yaron@seas.harvard.edu Abstract We consider the problem of optimization from samples of monotone submodular ...
6447 |@word worsens:1 private:1 polynomial:1 simulation:1 pick:1 contains:2 score:2 document:6 outperforms:1 si:15 must:3 v:2 generative:3 greedy:8 fewer:1 short:1 junta:1 math:1 node:1 unbounded:2 c2:34 focs:1 consists:5 manner:1 tagging:4 expected:7 hardness:3 behavior:3 multi:1 decreasing:1 cardinality:6 increasing:...
6,022
6,448
Combining Fully Convolutional and Recurrent Neural Networks for 3D Biomedical Image Segmentation Jianxu Chen University of Notre Dame jchen16@nd.edu Yizhe Zhang University of Notre Dame yzhang29@nd.edu Lin Yang University of Notre Dame lyang5@nd.edu Mark Alber University of Notre Dame malber@nd.edu Danny Z. Chen Un...
6448 |@word trial:1 cnn:10 briefly:1 illustrating:1 fcns:4 nd:5 bf:2 seek:1 propagate:1 mention:1 minus:1 shot:1 moment:2 initial:1 series:2 score:5 bc:2 ours:3 rightmost:1 outperforms:1 comparing:3 contextual:6 activation:1 yet:1 danny:1 finest:1 gpu:6 subsequent:1 shape:3 alone:2 half:1 fewer:1 selected:5 intelligenc...
6,023
6,449
Clustering with Same-Cluster Queries Hassan Ashtiani , Shrinu Kushagra and Shai Ben-David David R. Cheriton School of Computer Science University of Waterloo, Waterloo, Ontario, Canada {mhzokaei,skushagr,shai}@uwaterloo.ca Abstract We propose a framework for Semi-Supervised Active Clustering framework (SSAC), where t...
6449 |@word trial:1 kulis:1 version:2 polynomial:11 stronger:1 simulation:2 sheffet:1 asks:4 mention:1 reduction:2 contains:1 ours:1 interestingly:1 sugato:3 current:1 si:20 must:1 seeding:1 aside:1 selected:2 plane:5 location:4 simpler:1 constructed:2 prove:7 consists:3 combine:2 inside:1 yingyu:1 manner:1 introduce:2...
6,024
645
Using hippocampal 'place cells' for navigation, exploiting phase coding Neil Burgess, John O'Keefe and Michael Recce Department of Anatomy, University College London, London WC1E 6BT, England. (e-mail: n.burgess<Ducl.ac . uk) Abstract A model of the hippocampus as a central element in rat navigation is presented. Sim...
645 |@word middle:1 hippocampus:5 seems:1 open:1 cm2:1 simulation:3 crucially:1 excited:1 minus:1 harder:1 moment:1 rearing:1 ranck:1 current:3 activation:1 must:2 john:1 subsequent:2 motor:2 update:1 cue:1 fewer:1 beginning:1 short:1 nearness:1 location:11 successive:1 constructed:1 direct:1 become:1 examine:1 brain:2...
6,025
6,450
Hardness of Online Sleeping Combinatorial Optimization Problems Satyen Kale? ? Yahoo Research satyen@satyenkale.com Chansoo Lee? Univ. of Michigan, Ann Arbor chansool@umich.edu D?avid P?al Yahoo Research dpal@yahoo-inc.com Abstract We show that several online combinatorial optimization problems that admit efficient...
6450 |@word version:6 polynomial:5 stronger:1 open:9 km:3 carry:1 reduction:7 contains:2 current:2 com:2 michal:1 must:1 benign:1 update:1 implying:1 intelligence:1 warmuth:8 beginning:1 manfred:5 boosting:1 bijection:1 node:6 become:1 symposium:2 prove:6 consists:1 specialize:1 indeed:1 hardness:14 yasin:1 decreasing:...
6,026
6,451
Learned Region Sparsity and Diversity Also Predict Visual Attention Zijun Wei1? , Hossein Adeli2? , Gregory Zelinsky1,2 , Minh Hoai1 , Dimitris Samaras1 1. Department of Computer Science 2. Department of Psychology ? Stony Brook University 1.{zijwei, minhhoai, samaras}@cs.stonybrook.edu 2.{hossein.adelijelodar, gr...
6451 |@word stronger:2 kokkinos:1 everingham:1 attended:2 thereby:1 contains:2 score:27 selecting:1 hereafter:1 trainval:2 offering:1 tuned:2 interestingly:1 outperforms:3 current:2 activation:5 si:1 stony:1 must:1 intriguing:2 indistinguishably:1 cottrell:2 informative:1 blur:3 shape:1 enables:1 plot:4 intelligence:1 ...
6,027
6,452
Batched Gaussian Process Bandit Optimization via Determinantal Point Processes Tarun Kathuria, Amit Deshpande, Pushmeet Kohli Microsoft Research t-takat@microsoft.com, amitdesh@microsoft.com, pkohli@microsoft.com Abstract Gaussian Process bandit optimization has emerged as a powerful tool for optimizing noisy black bo...
6452 |@word kohli:1 determinant:5 version:2 exploitation:4 repository:1 seems:1 simulation:5 crucially:1 covariance:2 pick:1 thereby:1 nystr:1 tr:1 initial:1 contains:1 score:2 selecting:4 series:1 bibtex:4 outperforms:1 existing:3 ka:2 com:4 contextual:1 must:1 determinantal:10 pe1:1 kdd:2 burdick:1 greedy:8 selected:...
6,028
6,453
Using Social Dynamics to Make Individual Predictions: Variational Inference with a Stochastic Kinetic Model Zhen Xu, Wen Dong, and Sargur Srihari Department of Computer Science and Engineering University at Buffalo {zxu8,wendong,srihari}@buffalo.edu Abstract Social dynamics is concerned primarily with interactions amo...
6453 |@word briefly:1 willing:1 simulation:1 fifteen:1 minus:1 moment:1 contains:2 daniel:1 past:2 reaction:7 outperforms:1 current:6 comparing:1 si:1 peyton:1 must:2 written:2 john:1 realistic:1 subsequent:1 christian:1 update:3 discrimination:1 v:1 stationary:1 selected:2 intelligence:1 record:2 manfred:1 santo:1 nod...
6,029
6,454
A Non-parametric Learning Method for Confidently Estimating Patient?s Clinical State and Dynamics William Hoiles Department of Electrical Engineering University of California Los Angeles Los Angeles, CA 90024 whoiles@ucla.edu Mihaela van der Schaar Department of Electrical Engineering University of California Los Ang...
6454 |@word eliminating:1 polynomial:1 norm:4 km:2 covariance:16 accounting:1 citeseer:1 pressure:2 dramatic:1 tr:1 solid:1 initial:1 contains:7 score:3 selecting:1 series:2 bhattacharyya:2 outperforms:2 reaction:1 current:2 mihaela:2 must:5 written:3 john:1 remove:1 interpretable:1 update:2 prohibitive:1 selected:3 re...
6,030
6,455
Following the Leader and Fast Rates in Linear Prediction: Curved Constraint Sets and Other Regularities Ruitong Huang Department of Computing Science University of Alberta, AB, Canada ruitong@ualberta.ca Tor Lattimore School of Informatics and Computing Indiana University, IN, USA tor.lattimore@gmail.com Andr?s Gy?r...
6455 |@word innovates:1 version:4 norm:6 seems:1 replicate:1 nd:2 open:1 mehta:1 simulation:1 attainable:2 pick:4 thereby:1 tr:1 ftrl:2 selecting:2 chervonenkis:1 ours:1 erven:2 existing:1 com:1 nt:1 od:1 gmail:1 yet:1 bd:23 fn:1 belmont:1 shape:2 remove:1 update:1 intelligence:1 selected:2 parameterization:1 plane:5 r...
6,031
6,456
Multi-view Anomaly Detection via Robust Probabilistic Latent Variable Models Tomoharu Iwata NTT Communication Science Laboratories iwata.tomoharu@lab.ntt.co.jp Makoto Yamada Kyoto University makoto.m.yamada@ieee.org Abstract We propose probabilistic latent variable models for multi-view anomaly detection, which is th...
6456 |@word private:7 nd:1 covariance:1 liu:2 series:2 score:13 disparity:3 tist:1 document:4 existing:4 current:2 wd:14 comparing:1 contextual:1 written:2 romance:2 bd:1 ranka:1 kdd:1 shape:1 enables:1 generative:3 selected:2 pursued:1 item:2 intelligence:2 discovering:1 yamada:3 eskin:1 detecting:4 node:1 org:1 zhang...
6,032
6,457
CMA-ES with Optimal Covariance Update and Storage Complexity Oswin Krause Dept. of Computer Science University of Copenhagen Copenhagen, Denmark oswin.krause@di.ku.dk D?dac R. Arbon?s Dept. of Computer Science University of Copenhagen Copenhagen, Denmark didac@di.ku.dk Christian Igel Dept. of Computer Science Univer...
6457 |@word trial:7 version:1 briefly:2 norm:2 open:1 d2:6 ajj:1 covariance:30 decomposition:14 initial:1 omidvar:2 att:1 selecting:1 genetic:5 existing:1 comparing:1 numerical:1 shape:2 christian:1 designed:1 plot:1 update:32 drop:1 fund:1 intelligence:2 accordingly:1 beginning:1 smith:2 rosenbrock:6 provides:1 succes...
6,033
6,458
Large Margin Discriminant Dimensionality Reduction in Prediction Space Mohammad Saberian Netflix esaberian@netflix.com Can Xu Google canxu@google.com Jose Costa Pereira INESCTEC jose.c.pereira@inesctec.pt Jian Yang Yahoo Research jianyang@yahoo-inc.com Nuno Vasconcelos UC San Diego nvasconcelos@ucsd.edu Abstract I...
6458 |@word kulis:2 middle:3 dekel:1 seek:1 rgb:1 lpp:2 reduction:18 initial:3 liu:1 contains:2 score:1 document:1 current:8 com:3 guadarrama:1 goldberger:1 activation:1 must:1 john:1 distant:1 numerical:2 shape:2 enables:1 gist:1 update:8 v:1 hash:6 half:2 prohibitive:1 selected:2 plane:1 sys:1 steepest:1 short:1 iter...
6,034
6,459
Ef?cient Globally Convergent Stochastic Optimization for Canonical Correlation Analysis Weiran Wang1? Jialei Wang2? Dan Garber1 Nathan Srebro1 2 1 Toyota Technological Institute at Chicago University of Chicago {weiranwang,dgarber,nati}@ttic.edu jialei@uchicago.edu Abstract We study the stochastic optimization of can...
6459 |@word version:2 nd:3 reused:1 decomposition:4 covariance:5 concise:1 sgd:8 reduction:3 initial:2 document:1 si:35 dx:4 written:1 readily:1 numerical:1 chicago:2 enables:1 remove:2 plot:1 update:3 v:2 instantiate:2 warmuth:1 beginning:1 provides:3 iterates:7 allerton:2 zhang:2 dn:11 along:1 become:1 consists:2 dou...
6,035
646
On the Use of Projection Pursuit Constraints for Training Neural Networks Nathan Illtl'ator'" Comput.er Science Department Tel-Aviv Universit.y Ramat.-A viv, 69978 ISRAEL and Inst.itute for Brain and Neural Systems, Brown University nin~math,tau.ac.il Abstract \Ve present a novel classifica t.ioll and regression met....
646 |@word mild:1 neurophysiology:1 version:2 underst:1 polynomial:2 compression:3 norm:1 nd:1 ivit:1 seitz:1 seek:1 simplifying:1 awij:1 initial:1 erms:1 nowlan:2 lang:1 numerical:1 plasticity:2 nemal:1 ial:3 erat:1 math:1 location:1 ional:4 banff:1 sigmoidal:2 along:1 c2:1 direct:3 become:2 ect:1 edelman:1 combine:1 ...
6,036
6,460
Dynamic matrix recovery from incomplete observations under an exact low-rank constraint Liangbei Xu Mark A. Davenport Department of Electrical and Computer Engineering Georgia Institute of Technology Atlanta, GA 30318 lxu66@gatech.edu mdav@gatech.edu Abstract Low-rank matrix factorizations arise in a wide variety of a...
6460 |@word trial:2 version:1 briefly:1 norm:13 stronger:1 c0:2 d2:2 simulation:5 decomposition:1 tr:1 reduction:1 liu:1 contains:4 bc:1 outperforms:2 existing:1 ka:3 wd:3 luo:1 visible:1 numerical:1 timestamps:1 kdd:2 treating:1 v:2 item:3 ith:2 prize:1 fa9550:1 math:1 preference:4 c2:2 focs:2 symp:3 n22:2 expected:2 ...
6,037
6,461
Learning to learn by gradient descent by gradient descent Marcin Andrychowicz1 , Misha Denil1 , Sergio G?mez Colmenarejo1 , Matthew W. Hoffman1 , David Pfau1 , Tom Schaul1 , Brendan Shillingford1,2 , Nando de Freitas1,2,3 1 Google DeepMind 2 University of Oxford 3 Canadian Institute for Advanced Research marcin.a...
6461 |@word version:1 norm:1 bptt:2 gradual:1 decomposition:3 pick:1 sgd:1 solid:3 initial:2 series:1 contains:1 selecting:1 daniel:2 tuned:1 prefix:1 outperforms:3 com:5 optim:1 activation:7 gmail:1 john:1 informative:1 enables:1 designed:5 plot:10 update:19 v:1 intelligence:3 advancement:2 short:2 recherche:1 iterate...
6,038
6,462
Solving Marginal MAP Problems with NP Oracles and Parity Constraints Yexiang Xue Department of Computer Science Cornell University yexiang@cs.cornell.edu Stefano Ermon Department of Computer Science Stanford University ermon@cs.stanford.edu Zhiyuan Li? Institute of Interdisciplinary Information Sciences Tsinghua Univ...
6462 |@word middle:1 polynomial:3 chakraborty:1 replicate:3 nd:1 adnan:3 propagate:1 decomposition:1 pick:1 harder:1 reduction:1 moment:1 configuration:1 series:5 contains:2 liu:2 daniel:3 outperforms:2 comparing:2 conjunctive:1 written:1 dechter:3 plot:1 kuldeep:1 bart:5 aside:1 hash:1 selected:6 intelligence:6 greedy...
6,039
6,463
PerforatedCNNs: Acceleration through Elimination of Redundant Convolutions Michael Figurnov1,2 , Aijan Ibraimova4 , Dmitry Vetrov1,3 , and Pushmeet Kohli5 1 National Research University Higher School of Economics 2 Lomonosov Moscow State University 3 Yandex 4 Skolkovo Institute of Science and Technology 5 Microsoft Re...
6463 |@word kohli:1 cnn:15 retraining:2 d2:2 rgb:1 decomposition:3 perfo:1 reduction:7 configuration:5 contains:1 tuned:4 ours:1 outperforms:2 existing:1 freitas:1 guadarrama:1 com:6 comparing:1 skipping:2 activation:7 gmail:1 gpu:13 numerical:1 remove:1 moczulski:1 v:1 greedy:3 device:3 fried:1 podoprikhin:1 vanishing...
6,040
6,464
Learning Deep Embeddings with Histogram Loss Evgeniya Ustinova and Victor Lempitsky Skolkovo Institute of Science and Technology (Skoltech) Moscow, Russia Abstract We suggest a loss for learning deep embeddings. The new loss does not introduce parameters that need to be tuned and results in very good embeddings acros...
6464 |@word cnn:3 version:2 middle:1 polynomial:1 kokkinos:1 wexler:1 contrastive:6 tr:12 shot:2 moment:1 initial:1 liu:1 series:1 score:2 swansea:1 tuned:4 interestingly:1 outperforms:3 guadarrama:1 comparing:1 com:2 si:3 assigning:1 dx:2 moreno:1 drop:1 intelligence:2 ith:1 node:4 location:1 firstly:1 bowman:1 become...
6,041
6,465
R-FCN: Object Detection via Region-based Fully Convolutional Networks Jifeng Dai Microsoft Research Yi Li? Tsinghua University Kaiming He Microsoft Research Jian Sun Microsoft Research Abstract We present region-based, fully convolutional networks for accurate and efficient object detection. In contrast to previou...
6465 |@word cnn:50 fcns:3 kokkinos:1 everingham:1 c0:1 decomposition:1 minus:1 incarnation:1 shot:1 liu:2 series:1 score:34 trainval:9 ours:2 com:2 yet:2 gpu:4 enables:1 lcls:2 hypothesize:1 designed:1 remove:2 drop:1 rpn:19 v:2 aside:1 selected:2 parameterization:1 detecting:1 simpler:1 zhang:4 rc:4 constructed:1 cons...
6,042
6,466
Bayesian optimization for automated model selection Gustavo Malkomes,? Chip Schaff,? Roman Garnett Department of Computer Science and Engineering Washington University in St. Louis St. Louis, MO 63130 {luizgustavo, cbschaff, garnett}@wustl.edu Abstract Despite the success of kernel-based nonparametric methods, kernel...
6466 |@word exploitation:3 briefly:2 stronger:1 termination:1 closure:1 seek:2 covariance:14 accounting:1 automl:1 series:1 selecting:3 sobol:1 past:1 freitas:1 must:1 concatenate:1 numerical:1 distant:1 additive:2 treating:1 plot:1 update:2 stationary:1 generative:1 fewer:1 greedy:4 selected:1 intelligence:2 beginning...
6,043
6,467
Generalization of ERM in Stochastic Convex Optimization: The Dimension Strikes Back? Vitaly Feldman IBM Research ? Almaden Abstract In stochastic convex optimization the goal is to minimize a convex function . F (x) = Ef ?D [f (x)] over a convex set K ? Rd where D is some unknown distribution and each f (?) in the sup...
6467 |@word briefly:1 version:6 polynomial:1 norm:5 stronger:1 open:2 hu:1 elisseeff:1 thereby:2 boundedness:1 existing:1 current:1 optim:1 yet:1 gv:11 maxv:2 juditsky:1 leaf:1 smith:1 lr:4 district:1 simpler:1 zhang:1 dn:1 differential:1 consists:2 prove:5 interscience:1 privacy:1 huber:1 expected:3 examine:3 moulines...
6,044
6,468
One-vs-Each Approximation to Softmax for Scalable Estimation of Probabilities Michalis K. Titsias Department of Informatics Athens University of Economics and Business mtitsias@aueb.gr Abstract The softmax representation of probabilities for categorical variables plays a prominent role in modern machine learning with ...
6468 |@word repository:1 version:1 norm:6 nd:1 palma:1 crucially:1 jacob:1 citeseer:1 sgd:12 solid:1 carry:1 initial:3 qatar:1 score:13 selecting:1 bibtex:6 blackout:1 interestingly:2 bradley:3 current:1 com:3 surprising:1 must:1 john:1 fn:1 analytic:2 update:3 v:21 stationary:4 half:2 prohibitive:1 alone:1 item:1 para...
6,045
6,469
Dual Learning for Machine Translation Di He1,?, Yingce Xia2,? , Tao Qin3 , Liwei Wang1 , Nenghai Yu2 , Tie-Yan Liu3 , Wei-Ying Ma3 1 Key Laboratory of Machine Perception (MOE), School of EECS, Peking University 2 University of Science and Technology of China 3 Microsoft Research 1 {dih,wanglw}@cis.pku.edu.cn; 2 xiayin...
6469 |@word middle:6 briefly:1 open:2 seek:1 citeseer:1 initial:3 liu:1 contains:4 score:6 qatar:1 document:2 past:1 outperforms:9 recovered:1 com:3 contextual:1 comparing:1 jeopardy:1 gpu:1 subsequent:1 happen:1 informative:1 plot:2 update:6 half:2 accordingly:2 beginning:5 short:3 provides:1 barrault:1 five:1 constru...
6,046
647
Second order derivatives for network pruning: Optimal Brain Surgeon Babak Hassibi* and David G. Stork Ricoh California Research Center 2882 Sand Hill Road, Suite 115 Menlo Park, CA 94025-7022 stork@crc.ricoh.com and * Department of Electrical Engineering Stanford University Stanford, CA 94305 Abstract We investigate ...
647 |@word eliminating:1 inversion:2 seems:1 retraining:12 hu:2 simulation:2 gradual:1 covariance:3 fonn:1 dramatic:1 thereby:4 solid:1 reduction:3 initial:1 series:1 mag:3 qth:1 recovered:1 com:1 must:1 written:1 subsequent:1 remove:4 unintelligible:1 update:2 tenn:2 monk:8 isotropic:1 plane:1 ith:1 short:1 pointer:1 ...
6,047
6,470
Efficient Neural Codes under Metabolic Constraints Zhuo Wang ?? Department of Mathematics University of Pennsylvania wangzhuo@nyu.edu Xue-Xin Wei ?? Department of Psychology University of Pennsylvania weixxpku@gmail.com Alan A. Stocker Department of Psychology University of Pennsylvania astocker@sas.upenn.edu Daniel...
6470 |@word mild:1 trial:2 achievable:1 seems:1 grey:1 seek:1 solid:4 reduction:3 configuration:1 series:1 daniel:3 interestingly:3 current:4 com:1 nt:3 gmail:1 yet:1 must:2 readily:1 john:1 additive:3 numerical:2 informative:2 shape:3 bart:1 stationary:2 half:1 cue:1 metabolism:1 parameterization:1 short:4 characteriz...
6,048
6,471
Stochastic Variance Reduction Methods for Saddle-Point Problems P. Balamurugan INRIA - Ecole Normale Sup?rieure, Paris balamurugan.palaniappan@inria.fr Francis Bach INRIA - Ecole Normale Sup?rieure, Paris francis.bach@ens.fr Abstract We consider convex-concave saddle-point problems where the objective functions may b...
6471 |@word middle:1 version:1 norm:8 stronger:1 nd:7 unif:3 subcase:1 reduction:10 woodruff:2 ecole:2 existing:7 written:1 readily:1 refresh:2 numerical:1 plot:2 update:11 resampling:3 v:2 selected:1 fewer:1 xk:2 isotropic:1 provides:1 math:2 herbrich:1 simpler:1 zhang:3 mathematical:1 dn:1 ik:3 prove:1 introductory:1...
6,049
6,472
Simple and Efficient Weighted Minwise Hashing Anshumali Shrivastava Department of Computer Science Rice University Houston, TX, 77005 anshumali@rice.edu Abstract Weighted minwise hashing (WMH) is one of the fundamental subroutine, required by many celebrated approximation algorithms, commonly adopted in industrial pr...
6472 |@word msr:1 briefly:1 manageable:2 eliminating:1 proportion:2 seems:2 advantageous:1 loading:1 faculty:1 compression:1 dalal:1 triggs:1 rajaraman:1 scg:1 rgb:1 tr:1 reduction:2 necessity:1 celebrated:2 series:1 document:3 existing:9 nally:1 comparing:2 surprising:3 clara:1 realistic:1 kdd:1 cheap:1 designed:1 plo...
6,050
6,473
Incremental Variational Sparse Gaussian Process Regression Ching-An Cheng Institute for Robotics and Intelligent Machines Georgia Institute of Technology Atlanta, GA 30332 cacheng@gatech.edu Byron Boots Institute for Robotics and Intelligent Machines Georgia Institute of Technology Atlanta, GA 30332 bboots@cc.gatech....
6473 |@word determinant:1 version:1 illustrating:1 nd:1 heuristically:1 linearized:2 covariance:12 tr:2 solid:1 moment:1 precluding:1 rkhs:11 recovered:1 written:3 john:1 multioutput:1 numerical:2 subsequent:1 partition:1 j1:3 christian:1 designed:1 update:12 juditsky:1 intelligence:5 selected:3 isotropic:1 parametriza...
6,051
6,474
Combining Adversarial Guarantees and Stochastic Fast Rates in Online Learning Wouter M. Koolen Centrum Wiskunde & Informatica Science Park 123, 1098 XG Amsterdam, the Netherlands wmkoolen@cwi.nl Peter Gr?nwald CWI and Leiden University pdg@cwi.nl Tim van Erven Leiden University Niels Bohrweg 1, 2333 CA Leiden, the N...
6474 |@word illustrating:2 version:1 norm:2 mehta:2 d2:1 gradual:1 crucially:1 linearized:1 incurs:1 harder:1 interestingly:1 erven:16 z2:1 luo:2 worsening:1 surprising:1 yet:4 must:1 readily:1 gerchinovitz:1 succeeding:1 v:1 selected:1 plane:1 chiang:2 provides:3 characterization:1 boosting:1 along:1 c2:7 prove:3 doub...
6,052
6,475
A scalable end-to-end Gaussian process adapter for irregularly sampled time series classification Steven Cheng-Xian Li Benjamin Marlin College of Information and Computer Sciences University of Massachusetts Amherst Amherst, MA 01003 {cxl,marlin}@cs.umass.edu Abstract We present a general framework for classification...
6475 |@word middle:3 norm:1 vi1:1 nd:5 scalably:1 covariance:13 decomposition:1 recursively:1 carry:1 reduction:1 liu:1 series:55 uma:1 contains:2 daniel:1 ours:1 interestingly:1 rightmost:1 outperforms:2 existing:1 ka:4 com:1 si:5 diederik:1 written:1 must:1 numerical:3 partition:1 wx:5 kdd:2 enables:2 drop:2 plot:6 u...
6,053
6,476
Inference by Reparameterization in Neural Population Codes Rajkumar V. Raju Department of ECE Rice University Houston, TX 77005 rv12@rice.edu Xaq Pitkow Dept. of Neuroscience, Dept. of ECE Baylor College of Medicine, Rice University Houston, TX 77005 xaq@rice.edu Abstract Behavioral experiments on humans and animals ...
6476 |@word trial:4 cingulate:1 version:2 polynomial:1 nd:1 simulation:4 covariance:3 ttn:1 att:1 past:2 ka:1 anterior:1 dx:2 must:4 readily:1 written:1 mst:2 distant:1 subsequent:1 pseudomarginals:11 motor:1 gv:3 treating:1 update:17 intelligence:2 leaf:1 parameterization:1 ith:1 filtered:1 provides:2 node:20 five:2 m...
6,054
6,477
Understanding Probabilistic Sparse Gaussian Process Approximations Matthias Bauer?? Mark van der Wilk? Carl Edward Rasmussen? ? Department of Engineering, University of Cambridge, Cambridge, UK ? Max Planck Institute for Intelligent Systems, T?ubingen, Germany {msb55, mv310, cer54}@cam.ac.uk Abstract Good sparse appr...
6477 |@word worsens:2 briefly:1 inversion:2 twelfth:1 calculus:1 grey:1 covariance:10 tr:1 harder:1 initial:6 configuration:7 contains:1 series:3 initialisation:1 kuf:1 existing:1 recovered:1 must:1 fn:1 numerical:1 happen:1 remove:3 plot:1 update:1 clumping:6 intelligence:7 prohibitive:1 fewer:2 selected:1 greedy:1 is...
6,055
6,478
Fast and Provably Good Seedings for k-Means Olivier Bachem Department of Computer Science ETH Zurich olivier.bachem@inf.ethz.ch Mario Lucic Department of Computer Science ETH Zurich lucic@inf.ethz.ch S. Hamed Hassani Department of Computer Science ETH Zurich hamed@inf.ethz.ch Andreas Krause Department of Computer Sc...
6478 |@word stronger:1 nd:3 unif:1 vldb:1 d2:15 bahmani:2 initial:6 kingravi:1 selecting:1 ktv:1 daniel:1 outperforms:6 csn:7 yet:2 dx:2 sergei:2 additive:1 subsequent:1 kdd:6 seeding:29 stationary:1 half:2 selected:1 obsolete:2 intelligence:1 record:1 hypersphere:1 quantizer:2 completeness:1 provides:2 firstly:1 sympo...
6,056
6,479
Optimal spectral transportation with application to music transcription R?mi Flamary Universit? C?te d?Azur, CNRS, OCA remi.flamary@unice.fr Nicolas Courty Universit? de Bretagne Sud, CNRS, IRISA courty@univ-ubs.fr C?dric F?votte CNRS, IRIT, Toulouse cedric.fevotte@irit.fr Valentin Emiya Aix-Marseille Universit?, CNR...
6479 |@word middle:1 eliminating:1 villani:2 plsa:1 iki:1 km:2 decomposition:6 simplifying:1 contains:2 score:1 daniel:2 denoting:1 tuned:1 outperforms:1 thre:1 com:1 activation:2 negentropy:1 must:2 realistic:1 subsequent:1 additive:1 shape:1 hofmann:2 remove:1 designed:1 plot:3 sampl:1 polyphonic:2 discrimination:2 h...
6,057
648
On-Line Estimation of the Optimal Value Function: HJB-Estimators James K. Peterson Department of Mathematical Sciences Martin Hall Box 341907 Clemson University Clemson, SC 29634-1907 email: petersonOmath. clemson. edu Abstract In this paper, we discuss on-line estimation strategies that model the optimal value functi...
648 |@word automat:1 reduction:1 initial:10 series:1 zij:1 expositional:1 must:8 mesh:2 drop:1 update:3 alone:1 argm:1 dissertation:1 coarse:5 characterization:1 math:1 successive:1 ofo:1 mathematical:2 differential:2 corridor:7 supply:2 ooj:1 hjb:17 manner:1 introduce:1 ra:1 indeed:1 roughly:1 dist:3 planning:2 discre...
6,058
6,480
Coevolutionary Latent Feature Processes for Continuous-Time User-Item Interactions Yichen Wang? , Nan Du? , Rakshit Trivedi? , Le Song? ? Google Research ? College of Computing, Georgia Institute of Technology {yichen.wang, rstrivedi}@gatech.edu, dunan@google.com lsong@cc.gatech.edu Abstract Matching users to the righ...
6480 |@word cox:2 norm:5 proportion:1 flach:2 simulation:1 jacob:1 moment:1 liu:1 contains:7 series:1 past:5 outperforms:2 err:4 current:1 com:2 comparing:1 nell:1 attracted:1 romance:2 happen:1 informative:1 kdd:4 shape:1 designed:5 drop:1 update:2 v:3 selected:2 website:2 item:107 desktop:1 xk:8 parametrization:1 sho...
6,059
6,481
Nested Mini-Batch K-Means Franc?ois Fleuret Idiap Research Institue & EPFL francois.fleuret@idiap.ch James Newling Idiap Research Institue & EPFL james.newling@idiap.ch Abstract A new algorithm is proposed which accelerates the mini-batch k-means algorithm of Sculley (2010) by using the distance bounding approach of ...
6481 |@word eliminating:1 compression:1 proportion:1 nd:1 reused:4 closure:2 curtail:1 thereby:1 initial:3 contains:2 kingravi:1 initialisation:5 ours:3 dubourg:1 existing:1 current:1 comparing:3 com:1 must:3 written:2 subsequent:1 partition:2 kdd:1 enables:1 remove:2 drop:2 plot:1 update:14 pursued:1 selected:2 intell...
6,060
6,482
Blind Attacks on Machine Learners Alex Beatson Department of Computer Science Princeton University abeatson@princeton.edu Zhaoran Wang Department of Operations Research and Financial Engineering Princeton University zhaoran@princeton.edu Han Liu Department of Operations Research and Financial Engineering Princeton Un...
6482 |@word private:5 briefly:2 polynomial:1 proportion:2 nd:1 pick:6 accommodate:1 shot:1 carry:4 reduction:5 liu:2 omniscient:1 counterterrorism:2 current:1 comparing:2 si:1 yet:2 tackling:1 must:3 s2max:2 intelligence:4 advancement:2 ith:1 provides:6 completeness:2 attack:59 firstly:2 direct:3 differential:7 symposi...
6,061
6,483
Minimax Estimation of Maximum Mean Discrepancy with Radial Kernels Ilya Tolstikhin Department of Empirical Inference MPI for Intelligent Systems T?bingen 72076, Germany ilya@tuebingen.mpg.de Bharath K. Sriperumbudur Department of Statistics Pennsylvania State University University Park, PA 16802, USA bks18@psu.edu Be...
6483 |@word version:1 norm:2 seems:1 nd:1 open:2 cm2:1 covariance:2 p0:11 q1:5 thereby:3 mention:1 ipm:2 moment:2 rkhs:6 interestingly:1 past:1 existing:1 universality:1 nt1:3 fn:3 tailoring:1 intelligence:1 kyk:2 beginning:1 provides:5 math:1 c22:1 zhang:1 dn:2 constructed:2 c2:24 lopez:1 consists:1 prove:1 introduce:...
6,062
6,484
Fast recovery from a union of subspaces Chinmay Hegde Iowa State University Piotr Indyk MIT Ludwig Schmidt MIT Abstract We address the problem of recovering a high-dimensional but structured vector from linear observations in a general setting where the vector can come from an arbitrary union of subspaces. This setu...
6484 |@word torsten:1 version:2 polynomial:3 compression:2 norm:6 stronger:1 d2:11 vldb:2 decomposition:4 accounting:1 incurs:1 initial:1 liu:1 contains:3 ours:1 pprox:4 past:2 ka:1 written:1 must:1 john:1 additive:2 numerical:1 enables:2 update:1 v:1 greedy:1 instantiate:5 item:1 propack:7 cormode:1 volkan:2 iterates:...
6,063
6,485
Structured Prediction Theory Based on Factor Graph Complexity Corinna Cortes Google Research New York, NY 10011 Vitaly Kuznetsov Google Research New York, NY 10011 corinna@google.com vitaly@cims.nyu.edu Mehryar Mohrii Courant Institute and Google New York, NY 10012 Scott Yang Courant Institute New York, NY 10012 ...
6485 |@word mild:1 version:2 eliminating:1 norm:5 tadepalli:1 nd:1 r:4 crucially:1 seek:1 decomposition:11 contraction:5 reduction:1 substitution:1 series:3 score:1 past:1 existing:5 com:1 yet:2 must:1 parsing:4 additive:5 hofmann:2 enables:1 designed:1 fewer:1 mccallum:1 grfp:1 provides:3 boosting:3 m3n:4 node:7 simpl...
6,064
6,486
Coresets for Scalable Bayesian Logistic Regression Jonathan H. Huggins Trevor Campbell Tamara Broderick Computer Science and Artificial Intelligence Laboratory, MIT {jhuggins@, tdjc@, tbroderick@csail.}mit.edu Abstract The use of Bayesian methods in large-scale data settings is attractive because of the rich hierarch...
6486 |@word version:1 polynomial:2 norm:1 d2:1 seek:1 crucially:2 q1:1 thereby:1 tr:3 boundedness:1 reduction:1 moment:1 efficacy:1 score:1 genetic:2 document:2 past:1 existing:3 must:3 kqj:1 informative:1 wanted:1 seeding:1 plot:2 drop:1 n0:3 intelligence:5 generative:3 haario:1 hamiltonian:1 blei:1 location:1 org:5 z...
6,065
6,487
Universal Correspondence Network Christopher B. Choy Stanford University chrischoy@ai.stanford.edu JunYoung Gwak Stanford University jgwak@ai.stanford.edu Silvio Savarese Stanford University ssilvio@stanford.edu Manmohan Chandraker NEC Laboratories America, Inc. manu@nec-labs.com Abstract We present a deep learnin...
6487 |@word cnn:12 version:1 dalal:1 stronger:1 advantageous:1 everingham:1 triggs:1 open:1 choy:2 decomposition:2 jacob:2 contrastive:17 pick:1 tr:2 lepetit:2 reduction:1 configuration:1 series:2 disparity:1 liu:2 ours:19 outperforms:4 guadarrama:1 com:1 activation:10 yet:1 si:2 gpu:2 shape:10 enables:1 designed:4 plo...
6,066
6,488
Protein contact prediction from amino acid co-evolution using convolutional networks for graph-valued images Vladimir Golkov1 , Marcin J. Skwark2 , Antonij Golkov3 , Alexey Dosovitskiy4 , Thomas Brox4 , Jens Meiler2 , and Daniel Cremers1 1 Technical University of Munich, Germany 2 Vanderbilt University, Nashville, TN,...
6488 |@word mri:1 version:1 compression:1 norm:2 stronger:1 open:1 simulation:3 covariance:2 pressure:6 arous:1 reduction:1 configuration:2 contains:1 series:2 score:1 exclusively:1 daniel:2 uncovered:1 united:2 envision:1 outperforms:2 reaction:1 existing:1 recovered:1 com:1 current:1 videolearn:1 scatter:1 diederik:1...
6,067
6,489
Single-Image Depth Perception in the Wild Weifeng Chen Zhao Fu Dawei Yang Jia Deng University of Michigan, Ann Arbor {wfchen,zhaofu,ydawei,jiadeng}@umich.edu Abstract This paper studies single-image depth perception in the wild, i.e., recovering depth from a single image taken in unconstrained settings. We introduce ...
6489 |@word kohli:1 version:2 middle:1 judgement:1 seems:1 rgb:17 jacob:1 decomposition:1 pick:1 incurs:1 shading:2 harder:1 liu:7 series:2 exclusively:1 hoiem:2 salzmann:1 tuned:1 ours:4 outperforms:5 existing:6 lichtenberg:1 current:4 comparing:2 recovered:1 yet:5 must:1 shape:4 remove:2 hourglass:3 zik:2 cue:2 websi...
6,068
649
Learning Fuzzy Rule-Based Neural Networks for Control Charles M. Higgins and Rodney M. Goodman Department of Electrical Engineering, 116-81 California Institute of Technology Pasadena, CA 91125 Abstract A three-step method for function approximation with a fuzzy system is proposed. First, the membership functions and ...
649 |@word compression:10 loading:2 simulation:1 gradual:1 simplifying:1 decomposition:1 initial:5 contains:4 xiy:1 hereafter:1 existing:1 current:1 must:2 numerical:1 partition:1 remove:1 designed:1 v:1 fewer:1 timo:2 node:8 successive:1 mathematical:1 constructed:5 direct:1 yuhas:4 combine:1 manner:1 automatically:1 ...
6,069
6,490
On statistical learning via the lens of compression Ofir David Department of Mathematics Technion - Israel Institute of Technology ofirdav@tx.technion.ac.il Shay Moran Department of Computer Science Technion - Israel Institute of Technology shaymrn@cs.technion.ac.il Amir Yehudayoff Department of Mathematics Technion -...
6490 |@word version:9 polynomial:2 compression:113 open:3 ld:12 zij:1 chervonenkis:6 ramsey:4 com:1 z2:1 gmail:1 must:1 john:1 ligett:1 cue:1 selected:1 amir:3 warmuth:8 core:1 manfred:5 provides:1 boosting:3 characterization:1 compressible:1 herbrich:1 c2:2 aryeh:1 prove:2 consists:1 theoretically:1 indeed:1 hardness:...
6,070
6,491
Robust Spectral Detection of Global Structures in the Data by Learning a Regularization Pan Zhang Institute of Theoretical Physics, Chinese Academy of Sciences, Beijing 100190, China panzhang@itp.ac.cn Abstract Spectral methods are popular in detecting global structures in the given data that can be represented as a ...
6491 |@word illustrating:1 norm:1 tried:1 decomposition:2 asks:1 carry:1 initial:1 contains:3 selecting:1 itp:1 denoting:3 neeman:2 suppressing:1 outperforms:3 existing:3 si:5 written:2 numerical:3 partition:11 informative:8 remove:1 plot:5 update:1 v:1 generative:1 selected:3 fewer:1 item:5 short:1 detecting:3 complet...
6,071
6,492
Quantized Random Projections and Non-Linear Estimation of Cosine Similarity Ping Li Rutgers University Michael Mitzenmacher Harvard University Martin Slawski Rutgers University pingli@stat.rutgers.edu michaelm@eecs.harvard.edu martin.slawski@rutgers.edu Abstract Random projections constitute a simple, yet effecti...
6492 |@word repository:2 version:1 briefly:1 compression:1 norm:8 inversion:4 manageable:1 faculty:1 decomposition:1 jacob:1 attainable:2 thereby:1 tr:15 moment:1 reduction:5 celebrated:1 series:1 zij:1 tabulate:2 interestingly:1 kx0:1 comparing:1 yet:1 subsequent:1 numerical:1 kdd:2 treating:1 v:7 accordingly:4 vanish...
6,072
6,493
Adaptive Concentration Inequalities for Sequential Decision Problems Shengjia Zhao Tsinghua University zhaosj12@stanford.edu Enze Zhou Tsinghua University zhouez_thu_12@126.com Ashish Sabharwal Allen Institute for AI AshishS@allenai.org Stefano Ermon Stanford University ermon@cs.stanford.edu Abstract A key challen...
6493 |@word version:1 simulation:2 bn:5 kalyanakrishnan:1 necessity:1 configuration:1 ours:1 interestingly:1 existing:2 com:1 comparing:1 must:2 readily:1 fn:1 subsequent:2 analytic:1 remove:2 plot:6 drop:2 v:1 intelligence:1 half:1 fewer:1 provides:1 math:2 mannor:2 org:5 simpler:1 constructed:1 symposium:1 prove:1 wa...
6,073
6,494
Threshold Learning for Optimal Decision Making Nathan F. Lepora Department of Engineering Mathematics, University of Bristol, UK n.lepora@bristol.ac.uk Abstract Decision making under uncertainty is commonly modelled as a process of competitive stochastic evidence accumulation to threshold (the drift-diffusion model)....
6494 |@word neurophysiology:1 trial:58 exploitation:1 version:1 hu:2 simulation:1 gradual:1 covariance:2 recursively:1 initial:1 substitution:1 selecting:1 past:2 freitas:1 com:1 written:1 must:1 john:1 plasticity:1 shape:1 motor:2 moreno:1 plot:3 designed:1 update:2 v:4 discrimination:4 fewer:3 beginning:1 proficient:...
6,074
6,495
Sorting out typicality with the inverse moment matrix SOS polynomial Jean-Bernard Lasserre LAAS-CNRS & IMT Universit? de Toulouse 31400 Toulouse, France lasserre@laas.fr Edouard Pauwels IRIT & IMT Universit? Toulouse 3 Paul Sabatier 31400 Toulouse, France edouard.pauwels@irit.fr Abstract We study a surprising phenom...
6495 |@word repository:1 inversion:4 polynomial:80 proportion:3 nd:2 r:6 simulation:1 accounting:1 covariance:1 decomposition:1 independant:1 dishonest:1 ld:2 moment:31 contains:1 score:9 series:1 lichman:1 ours:2 existing:2 current:1 surprising:2 smtp:2 written:1 determinantal:1 numerical:4 additive:1 kdd:2 shape:17 p...
6,075
6,496
Sublinear Time Orthogonal Tensor Decomposition? Zhao Song? David P. Woodruff? Huan Zhang? Dept. of Computer Science, University of Texas, Austin, USA ? IBM Almaden Research Center, San Jose, USA ? Dept. of Electrical and Computer Engineering, University of California, Davis, USA zhaos@utexas.edu, dpwoodru@us.ibm.com, ...
6496 |@word h:1 mild:1 repository:1 version:6 mri:1 polynomial:1 norm:33 proportion:1 private:1 c0:5 hu:3 seek:1 decomposition:23 contraction:10 weekday:1 moment:1 initial:3 liu:2 selecting:1 woodruff:3 ours:1 fa8750:1 pprox:12 existing:3 current:1 com:3 si:7 kdd:2 cheap:1 remove:2 update:1 generative:1 core:1 short:1 ...
6,076
6,497
Neural Universal Discrete Denoiser Taesup Moon DGIST Daegu, Korea 42988 tsmoon@dgist.ac.kr Seonwoo Min, Byunghan Lee, Sungroh Yoon Seoul National University Seoul, Korea 08826 {mswzeus, styxkr, sryoon}@snu.ac.kr Abstract We present a new framework of applying deep neural networks (DNN) to devise a universal discrete ...
6497 |@word faculty:1 version:1 evaluating:1 sgd:4 inpainting:1 substitution:3 selecting:1 tuned:1 outperforms:2 activation:2 enables:1 plot:1 drop:1 update:1 fund:1 v:1 half:1 device:1 core:1 short:1 pascanu:1 node:3 location:7 lx:1 toronto:1 firstly:1 simpler:1 bioinform:1 along:1 become:2 consists:1 inside:1 deterio...
6,077
6,498
Online and Differentially-Private Tensor Decomposition Animashree Anandkumar Department of EECS University of California, Irvine a.anandkumar@uci.edu Yining Wang Machine Learning Department Carnegie Mellon University yiningwa@cs.cmu.edu Abstract Tensor decomposition is an important tool for big data analysis. In this...
6498 |@word private:22 faculty:1 version:1 polynomial:2 norm:8 sharpens:1 stronger:1 c0:2 open:2 simulation:4 decomposition:47 sgd:6 moment:8 liu:1 series:1 document:1 existing:3 recovered:1 current:1 protection:1 written:2 must:2 numerical:1 additive:1 j1:2 subsequent:1 designed:1 update:2 implying:1 fa9550:1 boosting...
6,078
6,499
Crowdsourced Clustering: Querying Edges vs Triangles Ramya Korlakai Vinayak Department of Electrical Engineering Caltech, Pasadena ramya@caltech.edu Babak Hassibi Department of Electrical Engineering Caltech, Pasadena hassibi@systems.caltech.edu Abstract We consider the task of clustering items using answers from no...
6499 |@word norm:5 km:1 condon:1 simulation:3 bn:1 pick:1 tr:1 klk:1 configuration:19 contains:1 liu:2 karger:2 series:1 daniel:1 hermosillo:1 outperforms:2 err:5 recovered:2 ksk1:1 comparing:1 manuel:1 anne:1 si:4 luis:1 john:1 planet:3 partition:3 shape:1 cheap:1 designed:1 drop:2 v:2 generative:4 intelligence:1 item...
6,079
65
348 Minkowski-r Back-Propaaation: Learnine in Connectionist Models with Non-Euclidian Error Silllais Stephen Jose Hanson and David J. Burr Bell Communications Research Morristown, New Jersey 07960 Abstract Many connectionist learning models are implemented using a gradient descent in a least squares error function of ...
65 |@word illustrating:1 compression:1 seems:2 simulation:2 euclidian:6 moment:1 reduction:4 emn:1 recovered:5 activation:5 dx:1 mesh:3 partition:2 shape:12 update:3 discrimination:1 tenn:1 plane:4 lr:1 simpler:2 five:1 differential:1 replication:1 burr:2 expected:3 roughly:1 examine:1 decreasing:2 increasing:3 underly...
6,080
650
Diffusion Approximations for the Constant Learning Rate Backpropagation Algorithm and Resistence to Local Minima William Finnoff Siemens AG, Corporate Research and Development Otto-Hahn-Ring 6 8000 Munich 83, Fed. Rep. Germany Abstract In this paper we discuss the asymptotic properties of the most commonly used varian...
650 |@word version:3 tedious:1 covariance:3 boundedness:1 exclusively:1 denoting:1 document:1 activation:2 additive:1 update:4 xk:2 ifx:1 provides:1 mathematical:1 constructed:1 differential:3 combine:2 indeed:1 expected:4 ra:1 themselves:1 decreasing:1 actual:1 considering:1 notation:2 bounded:1 developed:1 ag:1 every...
6,081
6,500
Linear Relaxations for Finding Diverse Elements in Metric Spaces Aditya Bhaskara University of Utah bhaskara@cs.utah.edu Mehrdad Ghadiri Sharif University of Technology ghadiri@ce.sharif.edu Vahab Mirrokni Google Research mirrokni@google.com Ola Svensson EPFL ola.svensson@epfl.ch Abstract Choosing a diverse subset ...
6500 |@word madelon:1 cu:7 repository:1 polynomial:2 nd:5 open:2 vldb:1 pick:3 concise:1 mention:1 reduction:3 liu:1 contains:4 lichman:1 selecting:2 score:1 freitas:1 nonmonotone:1 com:1 comparing:3 yet:1 written:2 partition:2 happen:1 cant:1 kdd:1 remove:3 drop:1 treating:1 update:1 v:1 greedy:9 selected:2 half:1 ite...
6,082
6,501
Deep Exploration via Bootstrapped DQN Ian Osband1,2 , Charles Blundell2 , Alexander Pritzel2 , Benjamin Van Roy1 1 Stanford University, 2 Google DeepMind {iosband, cblundell, apritzel}@google.com, bvr@stanford.edu Abstract Efficient exploration remains a major challenge for reinforcement learning (RL). Common ditherin...
6501 |@word exploitation:3 version:1 pieter:1 crucially:1 propagate:2 q1:3 pick:1 carry:1 initial:5 series:1 efficacy:1 selecting:1 score:2 daniel:1 tuned:1 bootstrapped:64 ours:1 rightmost:1 outperforms:2 existing:1 bradley:2 hasselt:1 com:1 freitas:1 guez:2 must:5 john:1 ronald:1 realistic:1 informative:5 enables:1 d...
6,083
6,502
SURGE: Surface Regularized Geometry Estimation from a Single Image Peng Wang1 Xiaohui Shen2 Bryan Russell2 Scott Cohen2 Brian Price2 Alan Yuille3 1 University of California, Los Angeles 2 Adobe Research 3 Johns Hopkins University Abstract This paper introduces an approach to regularize 2.5D surface normal and de...
6502 |@word kohli:1 cnn:16 kokkinos:2 paredes:1 nd:1 seek:1 propagate:6 rgb:8 ndez:1 contains:1 liu:1 hoiem:3 ours:3 romera:1 outperforms:1 existing:1 lichtenberg:1 current:1 guadarrama:1 written:1 readily:1 john:1 designed:1 drop:3 v:1 alone:1 cue:2 inspection:1 plane:36 vanishing:1 coarse:1 revisited:1 location:2 fir...
6,084
6,503
A Locally Adaptive Normal Distribution Georgios Arvanitidis, Lars Kai Hansen and S?ren Hauberg Technical University of Denmark, Lyngby, Denmark DTU Compute, Section for Cognitive Systems {gear,lkai,sohau}@dtu.dk Abstract The multivariate normal density is a monotonic function of the distance to the mean, and its elli...
6503 |@word trial:1 e215:1 nd:1 open:2 hu:1 covariance:20 thereby:1 reduction:3 initial:2 selecting:1 ours:1 outperforms:1 contextual:1 goldberger:1 yet:1 intriguing:1 written:1 numerical:1 shape:1 enables:1 moreno:1 v:7 generative:5 half:1 intelligence:4 tone:1 gear:1 isotropic:1 merger:1 steepest:1 short:1 provides:1...
6,085
6,504
Learning Structured Sparsity in Deep Neural Networks Wei Wen University of Pittsburgh wew57@pitt.edu Chunpeng Wu University of Pittsburgh chw127@pitt.edu Yiran Chen University of Pittsburgh yic52@pitt.edu Yandan Wang University of Pittsburgh yaw46@pitt.edu Hai Li University of Pittsburgh hal66@pitt.edu Abstract Hi...
6504 |@word cnn:1 middle:2 version:1 compression:4 norm:10 averagely:1 pg:1 solid:1 reduction:6 necessity:1 liu:4 configuration:1 series:1 offering:1 tuned:3 document:1 freitas:1 err:1 recovered:1 com:1 comparing:1 guadarrama:1 gemm:7 activation:1 written:1 gpu:16 john:2 grain:2 shape:33 enables:1 christian:2 remove:4 ...
6,086
6,505
Fast Active Set Methods for Online Spike Inference from Calcium Imaging 1 Johannes Friedrich1,2 , Liam Paninski1 Grossman Center and Department of Statistics, Columbia University, New York, NY 2 Janelia Research Campus, Ashburn, VA j.friedrich@columbia.edu, liam@stat.columbia.edu Abstract Fluorescent calcium indicat...
6505 |@word neurophysiology:1 faculty:1 version:3 polynomial:3 c0:15 disk:1 proportionality:1 r:8 holy:1 solid:1 deisseroth:2 initial:2 series:6 contains:1 optically:2 daniel:1 denoting:1 outperforms:1 ksk1:3 current:4 optim:1 skipping:1 chu:2 john:1 numerical:2 realistic:2 confirming:2 enables:2 remove:1 succeeding:1 ...
6,087
6,506
NESTT: A Nonconvex Primal-Dual Splitting Method for Distributed and Stochastic Optimization Davood Hajinezhad, Mingyi Hong ? Tuo Zhao? Zhaoran Wang? Abstract We study a stochastic and distributed algorithm for nonconvex problems whose objective consists of a sum of N nonconvex Li /N -smooth functions, plus a nonsmo...
6506 |@word mild:1 version:4 norm:1 seems:1 logit:1 confirms:1 covariance:1 q1:2 pick:4 sgd:7 reduction:2 initial:2 liu:2 past:2 existing:2 surprising:1 luo:3 activation:1 chu:1 hajinezhad:3 written:1 additive:1 numerical:1 designed:4 update:8 stationary:11 lky:1 selected:3 half:1 antoniadis:1 ith:1 characterization:1 ...
6,088
6,507
LazySVD: Even Faster SVD Decomposition Yet Without Agonizing Pain? Zeyuan Allen-Zhu zeyuan@csail.mit.edu Institute for Advanced Study & Princeton University Yuanzhi Li yuanzhil@cs.princeton.edu Princeton University Abstract We study k-SVD that is to obtain the first k singular vectors of a matrix A. Recently, a few ...
6507 |@word version:6 inversion:5 knd:8 compression:1 norm:18 nd:3 polynomial:4 stronger:2 open:5 cleanly:1 km:1 tried:2 decomposition:4 incurs:1 reduction:6 liu:1 woodruff:1 denoting:1 ours:1 outperforms:4 kmk:4 ka:20 comparing:1 yet:1 numerical:3 plot:4 update:2 v:25 fewer:1 website:2 selected:1 short:1 core:3 provid...
6,089
6,508
Statistical Inference for Cluster Trees Jisu Kim Department of Statistics Carnegie Mellon University Pittsburgh, USA jisuk1@andrew.cmu.edu Yen-Chi Chen Department of Statistics University of Washington Seattle, USA yenchic@uw.edu Alessandro Rinaldo Department of Statistics Carnegie Mellon University Pittsburgh, USA a...
6508 |@word mild:2 eliminating:1 norm:1 open:2 simulation:3 crucially:1 p0:14 pick:2 concise:1 solid:6 contains:4 efficacy:1 pbh:2 existing:1 comparing:1 john:2 dtq:1 shape:3 remove:1 interpretable:2 generative:1 leaf:15 fewer:1 smith:1 provides:1 node:3 complication:1 firstly:1 simpler:4 phylogenetic:2 height:7 along:...
6,090
6,509
Deep Learning for Predicting Human Strategic Behavior Jason Hartford, James R. Wright, Kevin Leyton-Brown Department of Computer Science University of British Columbia {jasonhar, jrwright, kevinlb}@cs.ubc.ca Abstract Predicting the behavior of human participants in strategic settings is an important problem in many do...
6509 |@word middle:3 version:5 proportion:3 nd:1 seek:1 simplifying:1 paid:1 sgd:1 thereby:1 recursively:1 initial:1 configuration:4 selecting:1 tuned:2 ours:2 offering:1 denoting:3 interestingly:1 outperforms:1 existing:5 current:2 si:2 activation:2 guez:1 must:4 john:1 subsequent:3 drop:1 update:1 aside:1 implying:1 ...
6,091
651
Memory-based Reinforcement Learning: Efficient Computation with Prioritized Sweeping Andrew W. Moore awm@ai.mit.edu NE43-759 MIT AI Lab. 545 Technology Square Cambridge MA 02139 Christopher G. At:iteson cga@ai.mit.edu NE43-771 MIT AI Lab. 545 Technology Square Cambridge MA 02139 Abstract We present a new algorithm, ...
651 |@word trial:1 version:1 tr:1 initial:1 tuned:2 existing:1 current:1 subsequent:1 remove:1 designed:1 alone:1 intelligence:1 prohibitive:1 fewer:1 record:1 draft:1 quantized:1 nom:1 five:1 rc:1 along:2 predecessor:5 qij:2 combine:1 peng:3 forgetting:1 themselves:1 examine:1 planning:1 terminal:3 globally:1 td:8 lit...
6,092
6,510
Depth from a Single Image by Harmonizing Overcomplete Local Network Predictions Ayan Chakrabarti TTI-Chicago Chicago, IL ayanc@ttic.edu Jingyu Shao Dept. of Statistics, UCLA? Los Angeles, CA shaojy15@ucla.edu Gregory Shakhnarovich TTI-Chicago Chicago, IL gregory@ttic.edu Abstract A single color image can contain ma...
6510 |@word kohli:1 version:1 norm:1 replicate:1 rgb:4 decomposition:1 sgd:1 shading:3 carry:2 initial:1 liu:3 contains:2 efficacy:1 disparity:1 hoiem:1 interestingly:1 current:2 z2:1 activation:3 gpu:1 chicago:5 informative:2 shape:1 cheap:1 cue:11 fewer:1 intelligence:1 parameterization:1 accordingly:1 plane:2 recipr...
6,093
6,511
Combinatorial Multi-Armed Bandit with General Reward Functions Wei Chen? Wei Hu? Fu Li? Jian Li? Yu Liu? Pinyan Luk Abstract In this paper, we study the stochastic combinatorial multi-armed bandit (CMAB) framework that allows a general nonlinear reward function, whose expected value may not depend only on the mea...
6511 |@word luk:1 exploitation:1 version:2 private:1 polynomial:3 laurence:1 hu:1 d2:2 r:3 jacob:1 profit:1 boundedness:1 liu:2 contains:1 celebrated:1 selecting:1 daniel:1 past:1 existing:4 yajun:1 com:4 discretization:5 si:4 gmail:3 conjunctive:1 attracted:1 john:1 underly:1 numerical:1 partition:1 enables:1 remove:1...
6,094
6,512
LightRNN: Memory and Computation-Efficient Recurrent Neural Networks 1 Xiang Li1 Tao Qin2 Jian Yang1 Tie-Yan Liu2 Nanjing University of Science and Technology 2 Microsoft Research Asia 1 implusdream@gmail.com 1 csjyang@njust.edu.cn 2 {taoqin, tie-yan.liu}@microsoft.com Abstract Recurrent neural networks (RNNs) have ...
6512 |@word luk:1 msr:1 norm:1 hu:1 heuristically:1 tried:1 pavel:1 citeseer:1 solid:1 tnlist:1 reduction:2 initial:1 liu:2 contains:2 score:1 pub:1 blackout:3 document:2 outperforms:2 existing:2 com:5 comparing:1 activation:1 gmail:1 njust:1 gpu:8 partition:1 xcj:2 drop:1 fund:1 v:1 half:1 leaf:4 device:7 yr:1 data2:1...
6,095
6,513
Contextual semibandits via supervised learning oracles Akshay Krishnamurthy? akshay@cs.umass.edu ? Alekh Agarwal? alekha@microsoft.com College of Information and Computer Sciences University of Massachusetts, Amherst, MA Miroslav Dud?k? mdudik@microsoft.com ? Microsoft Research New York, NY Abstract We study an on...
6513 |@word trial:1 exploitation:7 version:1 seems:1 norm:1 nd:3 suitably:1 c0:3 unif:2 open:1 termination:1 crucially:1 harder:1 moment:3 reduction:2 contains:1 uma:1 exclusively:1 selecting:2 tuned:3 document:7 longitudinal:1 outperforms:3 existing:6 contextual:33 com:4 discretization:1 chu:3 must:3 benign:1 enables:...
6,096
6,514
Stochastic Gradient Richardson-Romberg Markov Chain Monte Carlo ? Alain Durmus1 , Umut S?ims?ekli1 , Eric Moulines2 , Roland Badeau1 , Ga?el Richard1 1: LTCI, CNRS, T?el?ecom ParisTech, Universit?e Paris-Saclay, 75013, Paris, France ? 2: Centre de Math?ematiques Appliqu?ees, UMR 7641, Ecole Polytechnique, France Abstr...
6514 |@word mild:3 polynomial:1 norm:1 confirms:2 covariance:4 sgd:3 moment:1 initial:2 liu:1 contains:3 ecole:1 document:8 current:1 discretization:4 z2:1 yet:1 dx:1 additive:1 numerical:6 kdd:1 sdes:4 drop:1 update:2 stationary:1 generative:1 selected:1 intelligence:1 accordingly:4 xk:1 hamiltonian:4 core:2 provides:...
6,097
6,515
Riemannian SVRG: Fast Stochastic Optimization on Riemannian Manifolds Hongyi Zhang Sashank J. Reddi Suvrit Sra MIT Carnegie Mellon University MIT Abstract We study optimization of finite sums of geodesically smooth functions on Riemannian manifolds. Although variance reduction techniques for optimizing finite-sum...
6515 |@word briefly:1 version:1 middle:1 norm:4 trigonometry:1 stronger:1 seems:2 advantageous:1 nd:1 wiesel:2 hu:2 d2:2 simulation:2 bn:1 covariance:4 pick:1 sgd:2 sepulchre:1 reduction:11 liu:1 cherian:1 offering:1 bc:1 existing:2 mishra:1 elliptical:1 comparing:1 yet:1 john:1 numerical:1 analytic:1 plot:1 update:4 s...
6,098
6,516
Tensor Switching Networks Chuan-Yung Tsai?, Andrew Saxe?, David Cox Center for Brain Science, Harvard University, Cambridge, MA 02138 {chuanyungtsai,asaxe,davidcox}@fas.harvard.edu Abstract We present a novel neural network algorithm, the Tensor Switching (TS) network, which generalizes the Rectified Linear Unit (ReL...
6516 |@word cox:1 briefly:1 cnn:10 compression:6 advantageous:1 middle:1 hu:1 propagate:1 lobe:1 contraction:11 pick:1 sgd:1 reduction:4 bai:1 liu:2 rippel:1 tuned:2 interestingly:2 rightmost:1 current:1 comparing:1 com:1 activation:24 tackling:1 written:1 readily:1 must:3 gpu:4 wx:4 confirming:1 shape:2 remove:1 plot:...
6,099
6,517
The non-convex Burer?Monteiro approach works on smooth semide?nite programs Vladislav Voroninski? Department of Mathematics Massachusetts Institute of Technology vvlad@math.mit.edu Nicolas Boumal? Department of Mathematics Princeton University nboumal@math.princeton.edu Afonso S. Bandeira Department of Mathematics an...
6517 |@word version:1 polynomial:2 norm:7 stronger:1 nd:1 open:1 linearized:1 jacob:1 tr:5 sepulchre:3 initial:1 celebrated:1 interestingly:1 mishra:1 recovered:1 z2:2 com:1 toh:1 must:4 numerical:2 informative:1 benign:1 guess:1 caveat:3 provides:1 math:2 node:2 firstly:1 org:1 zhang:1 mathematical:7 consists:1 naor:2...