NIPS 2015 papers

(in nicer format than this)
maintained by @karpathy
Below every paper are TOP 100 most-occuring words in that paper and their color is based on LDA topic model with k = 7.
(This is very hard but it looks like 0 = graphical models?, 1 = reinforcement learning?, 2 = deep learning, 3 = kernels?, 4 = theory?, 5 = optimization, 6 = matrix factorization?)
Toggle LDA topics to sort by: TOPIC0 TOPIC1 TOPIC2 TOPIC3 TOPIC4 TOPIC5 TOPIC6
Also note that you can filter by the day each poster appears: click on the day next to a paper to filter by that day.
On Variance Reduction in Stochastic Gradient Descent and its Asynchronous Variants
Sashank J. Reddi, Ahmed Hefny, Suvrit Sra, Barnabas Poczos, Alex J. Smola


Poster: Tue Dec 8th 07:00 - 11:59 PM @ 210 C #77
[key, work] [framework, reduction, machine, optimal, return, set, cost, provide, based, approach] [read, performance, dataset, future] [stochastic, variance, figure, computational, compare, change, data, university, frequency] [algorithm, size, case, number, time] [asynchronous, gradient, convergence, vrg, speedup, descent, update, step, epoch, svrg, aga, parallel, schedule, general, chosen, chedule, optimization, locked, incremental, iterate, linear, sag, coordinate, convex, variant, iteration, objective, storage, additional, iterates, assume, method, note, synchronous, proximal, reduced, constant, presented, rate, sgd, mellon, denotes, unifying, dsgd, siam, csgd] [analysis, theorem, distributed, journal, consider, suppose, high, discussion, denote]
Alternating Minimization for Regression Problems with Vector-valued Outputs
Prateek Jain, Ambuj Tewari


Poster: Tue Dec 8th 07:00 - 11:59 PM @ 210 C #76
[] [] [] [] [] [] []
Interpolating Convex and Non-Convex Tensor Decompositions via the Subspace Norm
Qinqing Zheng, Ryota Tomioka


Poster: Tue Dec 8th 07:00 - 11:59 PM @ 210 C #79
[ratio, term, corresponding] [optimal, approach, based, threshold, behavior] [denoising, performance, recursive, table, experiment] [proposed, latent, gaussian, figure, kth, true, synthetic, parameter, standard, left, estimator, model, ideal, data, computed, approximation, estimated] [bound, corollary, size, prove, minimum, probability, lemma, algorithm, choice] [regularization, note, constant, theoretically, problem, general] [tensor, norm, subspace, error, noise, unfolding, rank, order, matrix, singular, relative, signal, theorem, decomposition, appendix, high, nuclear, vector, foldk, ordinary, square, recovering, random, result, kronecker, consider, odd, spectral, guarantee, orthonormal, leading, randomly, phase, version, analysis, overlapped, inner, deviation, required, condition]
Subset Selection by Pareto Optimization
Chao Qian, Yang Yu, Zhi-Hua Zhou


Poster: Tue Dec 8th 07:00 - 11:59 PM @ 210 C #78
[variable, constraint, corresponding, observation, strictly] [subset, set, optimal, criterion, select, learning, example, machine, evolutionary, integer, selecting, predictor] [performance, training, better, generated, previous, table] [data, approximation, figure, test, exponential, parameter, log, compared] [algorithm, time, greedy, number, size, lemma, implies, study, pareto, probability, exists, best, expected, prove, proof, easy, bound] [regression, optimization, problem, function, worse, squared, arg, method, pmax, note, objective, mcp, larger, convex, constant] [sparse, solution, selection, theoretical, jmax, isolation, decay, running, theorem, denote, norm, phase, guarantee, omp, journal, enpmax, recovery, newly, foba, vector, archived, iteratively, statistical, ieee, good, smaller, lasso, small, relaxation]
Semi-Proximal Mirror-Prox for Nonsmooth Composite Minimization
Niao He, Zaid Harchaoui


Poster: Tue Dec 8th 07:00 - 11:59 PM @ 210 C #87
[domain, exact, accuracy, associated, form, called] [set, optimal, allows, require, class] [sequence, representation, consists, arxiv] [variational, point, conditional, proposed, computation, assuming, full, applied] [algorithm, complexity, number, inequality, case, monotone, bound, structure, total] [convex, linear, minimization, composite, proximal, gradient, saddle, problem, objective, optimization, lmo, operator, smoothing, smooth, function, oracle, setup, convergence, inexact, rate, mirror, prox, assume, regularization, enjoys, arkadi, closed, euclidean, dual, solve, admm, max, mapping, nonsmooth, compute, technique, cmp, established, pxt, interest, anatoli] [consider, theoretical, min, norm, opt, establish, appendix, solving, solution, special]
Fast, Provable Algorithms for Isotonic Regression in all L_p-norms
Rasmus Kyng, Anup Rao, Sushant Sachdeva


Poster: Tue Dec 8th 07:00 - 11:59 PM @ 210 C #86
[domain, symmetry, work] [set, learning, supplementary, material, class, reduction, lexicographic, minimizes] [program, previous] [point, log, parameter, approximate, applied, computed] [algorithm, time, dag, lemma, best, directed, graph, probability, vertex, bound, partial, complete, starting, path, spielman, bounded, upper, implemented] [isotonic, regression, linear, ipm, strict, function, sotonic, convex, barrier, pprox, max, essian, hessian, problem, olve, method, fast, gradient, interior, compute, improved, lex, solve, computing, len, call, ood, lipschitz, stout, generalized, optimum, proved, tart] [theorem, error, running, vector, solving, weighted, denote, min, result, order, random, computes]
Sample Complexity of Episodic Fixed-Horizon Reinforcement Learning
Christoph Dann, Emma Brunskill


Poster: Tue Dec 8th 07:00 - 11:59 PM @ 210 C #90
[state, agent, transition, directly, work, decision, michael, sum, key, markov, making, accuracy] [policy, learning, mdps, reinforcement, mdp, reward, set, machine, conference, discounted, existing, optimal, international, bellman, successor, space, empirical, allows, require, class] [] [sample, model, true, equation, dependency, variance, applied, observed, introduce] [bound, number, time, complexity, pac, probability, horizon, episodic, algorithm, ucfh, lower, upper, proof, optimistic, setting, action, episode, lemma, lattimore, expected, hutter, case, total, tight, prove, pair, bounded, regret, unknown, concentration] [function, additional, local, assume, linear, satisfy] [condition, high, analysis, theorem, consider, journal, order, wmin]
A Universal Primal-Dual Convex Optimization Framework
Alp Yurtsever, Quoc Tran Dinh, Volkan Cevher


Poster: Tue Dec 8th 07:00 - 11:59 PM @ 210 C #89
[pxq, formulation, accuracy, constraint, template, state] [set, framework, optimal, algorithmic, require, based, supplementary, dist, weighting, material, learning] [performance, sequence] [computational, point, figure, standard, corresponds, numerical, parameter, correspond, density] [algorithm, complexity, number, inf] [gradient, convex, dual, universal, function, objective, iteration, proximal, method, lder, optimization, compute, scheme, primal, problem, requires, max, accelerated, convergence, operator, lipschitz, assumption, step, xpx, fista, quantum, linear, pgq, continuity, computing, minimization, oracle, choose, kmax, rate, nonsmooth, qmk, note, pkq, arg, apxq, general, solve, additional] [solution, matrix, min, norm, constrained, handle, analysis, consider, condition, completion, indicator, relative, nuclear]
HONOR: Hybrid Optimization for NOn-convex Regularized problems
Pinghua Gong, Jieping Ye


Poster: Tue Dec 8th 07:00 - 11:59 PM @ 210 C #92
[search, supplement, programming, three, subsequence, applicable] [learning, criterion, set, integer, propose] [challenging, include] [point, data, hybrid, considering, develop] [algorithm, lim, time, proposition, exists, critical, implies, number, proof, size, limit, bounded] [honor, function, optimization, convergence, convex, clarke, objective, problem, gist, regularized, cpu, solve, subdifferential, iteration, descent, scad, converges, mcp, step, lsp, scheme, constant, direction, concave, assumption, iterate, subproblem, larger, sparsenet, adopted, penalty, remark, optimality, regularizers, faster, inverse, presented, large] [sparse, small, analysis, positive, solving, iterative, annals, hold, ieee, selection, theorem, provided, journal, penalized]
Private Graphon Estimation for Sparse Graphs
Christian Borgs, Jennifer Chayes, Adam Smith


Poster: Tue Dec 8th 07:00 - 11:59 PM @ 210 C #91
[node, term, theory, work] [score, provide, consistent, set, average, needed, metric, paper, distance] [output, network, input, generated] [model, stochastic, log, density, approximation, estimator, target, data, nonparametric, full, estimating, exponential, distribution, estimate, procedure] [graph, algorithm, number, edge, bound, degree, probability, case, maximum, bounded, size, social, adjacent, dominated, lemma, best] [general, function, convergence, additional, constant, large, lipschitz, max, main, remark] [block, private, graphon, error, matrix, privacy, sparse, estimation, random, differential, graphons, nonprivate, sensitivity, analysis, consider, norm, version, statistical, square, well, consistency, adjacency, symmetric, guarantee, order]
Parallel Recursive Best-First AND/OR Search for Exact MAP Inference in Graphical Models
Akihiro Kishimoto, Radu Marinescu, Adi Botea


Poster: Thu Dec 10th 11:00 AM - 03:00 PM @ 210 C #36
[search, sprbfaoo, node, rbfaoo, graphical, cache, tree, thread, solved, protein, grid, pedigree, map, vqci, cbest, exact, corresponding, work, context, inference, working, aobb, current, sci, root, hard, bsci, promising, considerably, compilation, unsolved, load, child, associated, qci, intelligence] [heuristic, set, cost, space, virtual, optimal, instance, effective, return] [table, recursive, shared, memory] [sequential, parameter, figure, model] [time, total, algorithm, graph, number, pseudo, proof, best, limit] [parallel, problem, cpu, method, large, overhead, function, primal] [solution, distributed, solving, solves]
Bounding the Cost of Search-Based Lifted Inference
David B. Smith, Vibhav G. Gogate


Poster: Thu Dec 10th 11:00 AM - 03:00 PM @ 210 C #39
[lifted, search, inference, counting, node, tree, probabilistic, propositional, partition, decomposer, schematic, assignment, domain, exact, store, associated, variable, logic, conditioning, counted, ground, count, pxq, graphical, corresponding, markov, type, root, vibhav, list, child, mln, called, ptp, relational, mlns, induced, evidence, bounding, rooted] [set, space, cost, return, label, example, weak, conference] [representation, truth, labeled, single, generated, represent] [sampling, model, variance, sample, procedure, figure, log] [algorithm, number, size, complexity, path, structure, ordering, pair, pij, edge, tight] [function, computing, call, requires] [block, decomposition, order, consider, cluster, weighted, random, product, statistical]
Parallelizing MCMC with Random Partition Trees
Xiangyu Wang, Fangjian Guo, Katherine A. Heller, David B. Dunson


Poster: Thu Dec 10th 11:00 AM - 03:00 PM @ 210 C #44
[partition, accuracy, tree, toy] [subset, average, set, multiple, machine, based, conference, empirical, propose, distance, science, space, robert] [consists, length] [posterior, density, aggregated, data, aggregation, log, full, figure, mcmc, true, approximation, median, weierstrass, monte, dkl, nonparametric, gaussian, mixture, bayesian, procedure, carlo, drawn, fkd, parametric, uild, assuming, sample, university, likelihood, david, estimated] [algorithm, chain, total, number, time, partitioning, adaptive, probability, pairwise, bounded, versus] [assume, max, parallel, averaging, smoothing, large, combining] [random, block, theorem, provided, weighted, independent, product, randomly, error, estimation, denote, consider, statistical, covertype]
Fast Lifted MAP Inference via Partitioning
Somdeb Sarkhel, Parag Singla, Vibhav G. Gogate


Poster: Thu Dec 10th 11:00 AM - 03:00 PM @ 210 C #45
[map, ground, mln, lifted, inference, partition, domain, grounding, equivalence, markov, search, mlns, predicate, formula, propositional, ipp, argument, atom, logical, probabilistic, yield, logic, tuffy, partially, assignment, corresponding, quality, unable, three, alchemy, exchangeable, work, yielding, group] [set, approach, cost, class, conference, respect, example, lattice, return, integer, improve] [weight, network, described, processing, truth] [true, figure, approximate, cell] [algorithm, number, time, size, partial, complexity, lower, bound] [step, second] [solution, theorem, denote, consider, smaller, weighted, idea]
Sample Efficient Path Integral Control under Uncertainty
Yunpeng Pan, Evangelos Theodorou, Michail Kontitsis


Poster: Thu Dec 10th 11:00 AM - 03:00 PM @ 210 C #42
[state, theory, search, probabilistic, constraint, work, transition, belief, propagation] [control, optimal, based, policy, learned, learning, cost, pilco, dxt, approach, pddp, desirability, system, trajectory, uold, reinforcement, law, terminal, fgp, conference, existing, compositionality, machine, pde, task, international, framework, uncertain] [forward, neural, representation] [integral, sample, stochastic, gaussian, uncertainty, proposed, computational, sampling, approximation, model, computation, distribution, equation, computed, nonlinear, analytic, kernel, dynamic, data, standard, covariance, applied] [path, time, exp, algorithm, comparison, unknown] [method, composite, function, linear, optimization, compute, update, problem, generalized] [iterative, noise, order, ieee, independent, expression, sampled]
Efficient Thompson Sampling for Online Matrix-Factorization Recommendation
Jaya Kawale, Hung H. Bui, Branislav Kveton, Long Tran-Thanh, Sanjay Chawla


Poster: Thu Dec 10th 11:00 AM - 03:00 PM @ 210 C #43
[particle, probabilistic, work, current, key, agent, move] [set, datasets, system, approach, real, conference, optimal] [dataset, movie, baseline, learn, training, better, feature, performance, evaluate] [user, posterior, latent, sampling, data, bayesian, synthetic, model, recommend, mse, recommendation, icf, rating, sample, parameter, estimate, point, full, typically, observed, rit, proposed, log, prior, figure] [item, regret, algorithm, bandit, time, setting, thompson, pmf, number, movielens, cumulative, cummulative, rated, probability, action, chooses, expected, bound] [online, problem, update, method, note, linear, assume] [matrix, factorization, random, popular, row, collaborative, design, high]
Spatial Transformer Networks
Max Jaderberg, Karen Simonyan, Andrew Zisserman, koray kavukcuoglu


Spotlight: Wed Dec 9th 11:35 AM - 12:00 PM @ Room 210 A
Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #3
[grid, map, work, module, mechanism, accuracy] [transformed, multiple, allows, learning, class, task, location, showing] [spatial, transformation, transformer, feature, network, input, output, neural, localisation, cnn, convolutional, image, attention, training, mnist, single, trained, learn, table, transform, regular, deep, predicted, distorted, object, architecture, include, bilinear, baseline, perform, parameterised, previous, recognition, forward, cnns, visual, producing, thin, plate, unm, pixel, layer, warping, dataset, resolution, projective, street, vic, train] [sampling, model, data, applied, digit, kernel, invariant, test, generative, sample] [number] [differentiable, parallel] [error, subsequent, result, achieves]
A Reduced-Dimension fMRI Shared Response Model
Po-Hsuan (Cameron) Chen, Janice Chen, Yaara Yeshurun, Uri Hasson, James Haxby, Peter J. Ramadge


Oral: Wed Dec 9th 02:50 - 03:30 PM @ Room 210 A
Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #23
[group, accuracy, xit, form, factor, probabilistic] [response, space, segment, average, learned, collected, learning, datasets, basis, individual, approach, respect, informative, uninformative, allows] [shared, subject, srm, fmri, training, brain, experiment, learn, dataset, feature, voxel, spatial, movie, half, human, seg, cognitive, image, neural, raider, wit, anatomical, performance, alignment, removing, voxels, temporal, original, topography, hyperalignment, ventral] [data, model, functional, test, estimated, stimulus, held, distinct, figure, series, latent, cortical, standard, aggregation, multivariate] [time, number] [problem, objective, method, projected] [correlation, error, analysis, matrix, residual, orthogonal, component, overlapping, statistical, random, independent, dimension, application, solution]
Attractor Network Dynamics Enable Preplay and Rapid Path Planning in Maze–like Environments
Dane S. Corneil, Wulfram Gerstner


Oral: Wed Dec 9th 02:50 - 03:30 PM @ Room 210 A
Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #10
[state, position, rapid, transition, intermediate, rapidly, initial, corresponding, decoded, current, animal, move, represents] [goal, successor, attractor, environment, representing, bump, hippocampus, space, learning, location, hippocampal, reward, planning, stimulated, based, active, wik, sin, rec, set, metric, provide, preplay] [network, activity, place, representation, input, spatial, neural, encoding, recurrent, slow, represent, feature, neuron, decoding, memory, generate, generated, wij, weight, dorsal, plasticity, tend] [computational, model, approximation, cell, observed, point, figure] [path, time, shortest] [coordinate, function, mapping, large, gradient] [arbitrary, consider, eigenvectors, role, analysis, matrix, vector, stable, recovered, result]
Inferring Algorithmic Patterns with Stack-Augmented Recurrent Nets
Armand Joulin, Tomas Mikolov


Spotlight: Wed Dec 9th 03:30 - 04:00 PM @ Room 210 A
Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #9
[list, discrete, work, current, state, count, counting, term] [learning, algorithmic, set, example, machine, learned, binary, multiple, based, allows, structured, capacity, approach, goal] [stack, rnn, recurrent, pop, memory, push, learn, hidden, neural, network, sequence, generated, lstm, table, long, deep, training, language, correctly, layer, complex, predict, longer, generalize, rnns, previous, input, represent, short, predicted, stored, validation, turing, representation] [model, continuous, standard, figure, test, computational, data] [simple, time, number, study, element, focus, action, probability] [top, second, memorization, gradient, linear, note] [matrix, consider, random, studied]
Variance Reduced Stochastic Gradient Descent with Neighbors
Thomas Hofmann, Aurelien Lucchi, Simon Lacoste-Julien, Brian McWilliams


Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #78
[updated, reference, current] [learning, machine, evaluation, based, experimental, allows, science, provide] [memory, sharing, relevant, previous] [stochastic, data, variance, approximation, point, expectation, approximate, family] [number, size, algorithm, lemma, case, corollary, neighborhood, uniform, choice, regime, setting, choosing] [saga, step, gradient, sgd, note, suboptimality, rate, convergence, update, function, memorization, svrg, descent, regression, constant, requires, lyapunov, freshness, reduced, iterates, apply, chosen, cst, linear, correction, optimization, guaranteed, regularized, convex, additional, large, method, call, variant, main] [small, geometric, analysis, consider, condition, special, theorem]
Deep Convolutional Inverse Graphics Network
Tejas D. Kulkarni, William F. Whitney, Pushmeet Kohli, Josh Tenenbaum


Spotlight: Wed Dec 9th 11:35 AM - 12:00 PM @ Room 210 A
Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #6
[work, represented] [learning, set, machine, static, approach, respect, conference, intrinsic, propose] [network, encoder, representation, training, image, decoder, face, disentangled, deep, azimuth, learn, latents, trained, output, single, code, neural, rendering, represent, convolutional, chair, reconstruction, convolution, train, object, lighting, input, pose, dataset, interpretable, consists, generate, scene, elevation, force, ability, angle, arxiv, preprint, transformation, produce, generating, ztrain, architecture, render] [latent, model, figure, data, variational, procedure, light, proposed, test, stochastic, correspond, approximate, generative, distribution, approximation, varying] [algorithm] [batch, inverse, gradient, function] [random, error, ieee]
Learning to Segment Object Candidates
Pedro O. Pinheiro, Ronan Collobert, Piotr Dollar


Spotlight: Wed Dec 9th 11:35 AM - 12:00 PM @ Room 210 A
Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #8
[box, inference, scoring, bounding, edgeboxes] [recall, set, geodesic, approach, score, average, learning, multiple, prediction, label, predicts, location] [object, segmentation, image, deepmask, patch, proposal, network, mask, convolutional, branch, detection, input, output, selectivesearch, fully, trained, mcg, layer, pascal, rigor, connected, coco, generate, training, iou, pixel, performance, single, feature, deepmaskzoom, scale, architecture, generating, neural, deep, generates, auc, densely, previous, generated, convnet, train, imagenet, table, voc, vgg, evaluate] [model, figure, full, applied, test, proposed] [number] [large, fast, method, rely, apply, note] [high, positive, centered, achieves, well, dimension, small]
The Return of the Gating Network: Combining Generative Models and Discriminative Training in Natural Image Priors
Dan Rosenbaum, Yair Weiss


Spotlight: Wed Dec 9th 11:35 AM - 12:00 PM @ Room 210 A
Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #12
[inference, performing, assignment, michael, factor, visually] [learning, learned, based, approach, cost, conference, set, machine, naive, paper] [image, gating, network, training, restoration, discriminative, denoising, trained, performance, patch, natural, epll, clean, computer, discriminatively, pattern, neural, successful, vision, deblurring, perform, psnr, magnitude, architecture, stefan, learn, better, consth, layer, combine] [generative, full, gmm, gaussian, model, posterior, figure, log, computational, computation, prior, mixture, latent, comparable, difference, expensive, covariance, proposed] [probability, number, best, side, seek] [method, advantage, step, speedup, fast, top, local] [ieee, noise, component, modularity, noisy, smaller, eigenvectors]
Associative Memory via a Sparse Recovery Model
Arya Mazumdar, Ankit Singh Rawat


Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #74
[message, coherence, store, correct, associated, soft, mutual] [set, basis, recall, learning, based, capacity, approach, binary, space] [dataset, neural, memory, performance, described, stored, network] [model, gaussian, exponential] [algorithm, number, probability, feasible, property, graph, proposition, size, edge, failure] [linear, note, problem, large, constant, satisfy, iteration, arg, main, max, storage] [matrix, recovery, associative, sparse, vector, random, iterative, neurally, phase, noise, theorem, ist, bregman, consider, ieee, version, thresholding, signal, noisy, orthogonal, solution, bipartite, sparsity, subspace, spanned, orthonormal, min, error, small, pursuit, analysis]
High Dimensional EM Algorithm: Statistical Optimization and Asymptotic Normality
Zhaoran Wang, Quanquan Gu, Yang Ning, Han Liu


Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #73
[variable, inference, theory, term, exact] [score, maximization, propose, framework, optimal, respect, existing, based, approach, hypothesis, learning] [arxiv, preprint, level] [latent, log, parameter, mixture, estimator, asymptotic, sample, proposed, likelihood, gaussian, population, model, university, statistic, uncertainty, true, detailed] [algorithm, maximum, prove, size, bound, testing, upper] [rate, assumption, function, gradient, regression, large, step, converges, optimization, assume, max, implementation, local, ascent, convergence, global, linear, general] [dimensional, high, statistical, condition, error, decorrelated, sparsity, establish, estimation, low, version, theorem, order, analysis, journal, annals, init, attained, truncation, solution, appendix, geometric, theoretical, inferential, attains, iterative, nuisance, tensor, nsf]
Sparse Linear Programming via Primal and Dual Augmented Coordinate Descent
Ian En-Hsu Yen, Kai Zhong, Cho-Jui Hsieh, Pradeep K. Ravikumar, Inderjit S. Dhillon


Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #76
[constraint, programming, corresponding] [system, cost, set, approach, active, machine, optimal, learning, precision, existing] [program, relation] [point, covariance, data, log] [algorithm, complexity, number, size, bound] [linear, dual, method, coordinate, solve, primal, iteration, descent, subproblem, convergence, lagrangian, problem, augmented, alm, inverse, simplex, quadratic, optimization, binding, newton, interior, constant, gradient, note, direction, proximal, rcd, function, iterates, ipm, projected, support, update, requires, satisfying, minimization, global, objective] [matrix, sparse, solution, solving, product, estimation, vector, iterative, nonnegative, min, analysis, high, random, denote, theorem, sparsity, journal, smaller]
Matrix Completion Under Monotonic Single Index Models
Ravi Sastry Ganti, Laura Balzano, Rebecca Willett


Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #75
[work, discrete, performing, called] [set, learning, real, calibrated, datasets, effective, paper, approach, based, task, provide, distance, entire] [training, perform, single] [equation, estimate, model, observed, proposed, synthetic, kernel, gaussian, mse, sample, data] [algorithm, probability, best, bounded] [function, problem, optimization, loss, convex, squared, lipschitz, assume, gradient, solve, applying, def, linear, assumption, objective, large, call, descent, minimization, update, projected, euclidean] [matrix, completion, mmc, rank, monotonic, rmse, lpav, consider, min, high, estimation, error, low, solving, random, norm, order, lrmc, recommender, incomplete, studied, uniformly, singular, independent, well, analysis, sampled, clustering, journal]
Nonparametric von Mises Estimators for Entropies, Divergences and Mutual Informations
Kirthevasan Kandasamy, Akshay Krishnamurthy, Barnabas Poczos, Larry Wasserman, james m. robins


Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #69
[work, mutual, form, construct] [based, entropy, approach, learning, existing, framework, class, machine, paper, propose, empirical, nyi, require, computationally, multiple] [table, single, feature] [functionals, estimator, loo, estimating, divergence, density, nonparametric, distribution, functional, asymptotic, data, estimate, conditional, parametric, teaux, kde, tds, von, demonstrate, vme, semiparametric, derivative, hellinger, continuous, tloo, bandwidth, expansion, integral, compared, compare, differentiability, university, limiting, normal] [lower, bound, minimax, setting, bounded, focus, proof, prove, achieve] [rate, function, convergence, technique, assumption, method, splitting, satisfy, second, smoothness, correction] [estimation, statistical, appendix, analysis, order, ieee, knn, theorem, clustering, establish]
Action-Conditional Video Prediction using Deep Networks in Atari Games
Junhyuk Oh, Xiaoxiao Guo, Honglak Lee, Richard L. Lewis, Satinder Singh


Spotlight: Wed Dec 9th 03:30 - 04:00 PM @ Room 210 A
Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #18
[work, directly] [learning, prediction, control, paper, based, space, predicts] [encoding, recurrent, deep, feedforward, controlled, predicted, predict, neural, atari, network, frame, temporal, object, dqn, video, future, layer, input, convolutional, learn, feature, movement, previous, evaluate, image, convolution, trained, hidden, informed, cnn, better, generate, lstm, naff, decoding, involve, training, performance, modeling, henc, transformation, multiplicative, baseline, pixel, architecture] [model, figure, predictive, proposed, data, true] [action, exploration, game, time, number, similarity, play] [squared, top] [random, small, error, vector, restricted]
Local Smoothness in Variance Reduced Optimization
Daniel Vainsencher, Han Liu, Tong Zhang


Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #72
[theory, current, form, factor, proportion, reference] [empirical, locally, effective, set, risk, allows, provide, approach] [single, arxiv, relevant, preprint, half] [sampling, stochastic, figure, data, speed, variance, fewer] [algorithm, uniform, smoothed, dependent, number, expected, bound, size, lemma] [loss, sdca, local, gradient, smoothness, svrg, global, constant, duality, smooth, dual, hinge, ball, compute, strong, coordinate, convergence, svm, large, tong, gap, optimization, descent, problem, lipschitz, ascent, remark, apply, general, avg, proximal, regularized, step, optimum, certify, ill, larger, reduced] [small, solving, theorem, consider, naturally]
Bounding errors of Expectation-Propagation
Guillaume P. Dehaene, Simon Barthelmé


Poster: Wed Dec 9th 07:00 - 11:59 PM @ 210 C #70
[quality, term, propagation, bounding, inference, interesting, work, corresponding, extension] [close, distance, changing, machine, learning, cost] [natural, better, higher, third, performance] [approximation, gaussian, cga, distribution, vep, variance, url, true, target, log, approximate, expectation, slowly, point, asymptotic, divergence, introduce, computed, approximating, posterior, applied, qkl, investigate, exponential, site, hybrid, computational, dehaene] [bound, exp, limit, prove, proof, isbn, probability, number] [note, compute, method, computing, technical, large, excess, call, sense, step] [order, mode, moment, error, theoretical, good, result, well, consider, fourth, offer, odd, independent]
Collaboratively Learning Preferences from Ordinal Data
Sewoong Oh, Kiran K. Thekumparampil, Jiaming Xu


Poster: Tue Dec 8th 07:00 - 11:59 PM @ 210 C #74
[inference, logarithmic, discrete, multinomial] [learning, provide, approach, set, range, history, utility, subset] [predict, modeling, learn, hidden, category, ranked] [model, user, log, sampling, sample, proposed, dynamic, figure, parameter, data, estimate, drawn, distribution] [choice, bound, ranking, corollary, upper, number, preference, probability, lower, algorithm, case, proof, size, underlying, item, minimax, structure] [convex, presented, optimization, assume, ordinal, general, second, larger, solve, function, regularization] [error, mnl, norm, matrix, theorem, rank, collaborative, bundled, random, nuclear, purchase, denote, relaxation, appendix, solving, achieves, logit, suppose, order, fundamental, independent, rmse, analysis, noisy, row, provable]
SGD Algorithms based on Incomplete U-statistics: Large-Scale Minimization of Empirical Risk
Guillaume Papa, Stéphan Clémençon, Aurélien Bellet


Poster: Tue Dec 8th 07:00 - 11:59 PM @ 210 C #75
[tuples, form, term] [empirical, learning, risk, set, based, metric, generalization, machine, provide, distance, average, approach, paper, improve, respect, supplementary] [performance, auc, ability] [stochastic, variance, sampling, data, approximation, sample, conditional, computational, estimate, standard, distribution, estimator, proposed, asymptotic, test, figure, unbiased, kernel, xdk] [proposition, number, size, case, bound, complete, proof, probability, ranking, focus, degree, pairwise] [gradient, sgd, optimization, minimization, function, replacement, convex, batch, loss, convergence, rate, method, derive, implementation, xik, step, technical, large, generalized, descent] [incomplete, statistical, analysis, estimation, error, denote, theorem, order, independent, clustering, result, sketch]
Active Learning from Weak and Strong Labelers
Chicheng Zhang, Kamalika Chaudhuri


Poster: Thu Dec 10th 11:00 AM - 03:00 PM @ 210 C #46
[work, negative, current, key, decision, implicitly, observation] [learning, active, label, weak, disagreement, set, labeler, hypothesis, class, multiple, provide, query, false, statistically, agnostic, empirical, errd, example, dbal, querying, incorrect, consistent, labelled, goal, based, close, determine, risk, differs, labelers, predicts, generalization, medical] [region, labeled, training, learn, labeling, unlabeled, dataset, predict, annotation, train, realistic] [difference, target, data, distribution, standard, model, sample, localized] [algorithm, number, complexity, probability, observe, notation, case, setting, learner, achieve] [oracle, strong, excess, assumption, large, second, general, linear, problem, main, note, overhead, erm] [error, low, order, theorem, random, high, good]
Fast Second Order Stochastic Backpropagation for Variational Inference
Kai Fan, Ziteng Wang, Jeff Beck, James Kwok, Katherine A. Heller


Poster: Mon Dec 7th 07:00 - 11:59 PM @ 210 C #38
[inference, form, scalable, three, work] [learning, machine, cost, datasets, dimensionality] [deep, neural, train, vae, reconstruction, encoder, reduce, multiplication, hidden, james, natural] [variational, gaussian, stochastic, model, latent, backpropagation, generative, variance, parameter, expectation, approximate, computation, ada, bayesian, distribution, test, standard, continuous, computational, data, derivative, estimator, monte, approximation, carlo, prior, covariance, reparameterization, unbiased, figure, university, estimate] [lower, bound, algorithm, number, complexity, property, size] [gradient, hessian, optimization, function, method, hfsgvi, lipschitz, convergence, logistic, regression, linear, rate, fast, constant, apply, compute, problem, dsvi, note, technique, sgd] [order, matrix, theorem, small, consider]
Bidirectional Recurrent Convolutional Networks for Multi-Frame Super-Resolution
Yan Huang, Wei Wang, Liang Wang


Poster: Tue Dec 8th 07:00 - 11:59 PM @ 210 C #6
[motion, current, work, three, denoted] [learning, conference, machine, propose, international, based, existing, learned] [recurrent, video, temporal, brcn, hidden, convolutional, layer, network, bidirectional, image, feedforward, convolution, psnr, input, frame, better, previous, super, forward, computer, neural, training, original, anr, pattern, connection, table, modelling, complex, bicubic, connected, backward, visual, feature, trbm, optical, timestep, output, spatial, represent, learn, turbine] [model, conditional, dependency, figure, proposed, computational, full, speed] [time, size, achieve, comparison, focus] [method, note, second, faster] [ieee, including, running, vector, high]
Model-Based Relative Entropy Stochastic Search
Abbas Abdolmaleki, Rudolf Lioutikov, Jan R. Peters, Nuno Lau, Luis Pualo Reis, Gerhard Neumann


Poster: Mon Dec 7th 07:00 - 11:59 PM @ 210 C #40
[search, outperforms, considerably, form, state] [policy, entropy, surrogate, learning, task, robot, reaching, average, dimensionality, reward, control, machine, return, planar, premature, hole, algorithmic, xnes, beer, based, conference, reduction, evolutionary, evaluation, avoid, learned] [natural, learn, performance, higher, supervised] [distribution, stochastic, model, figure, parameter, standard, typically, gaussian, point, bayesian, equation, university, approximation, sample, pong, estimate, sampling, covariance, integral, approximate, posterior, test] [algorithm, bound, comparison] [objective, quadratic, function, optimization, gradient, linear, update, method, ball, problem, note, closed, dual, regression] [matrix, relative, order, power, noisy, vector, projection, dimensional, error, introduced]
Learnability of Influence in Networks
Harikrishna Narasimhan, David C. Parkes, Yaron Singer


Poster: Thu Dec 10th 11:00 AM - 03:00 PM @ 210 C #49
[node, observation, opinion, work, corresponding, initial, interesting] [learning, set, threshold, based, class, learned, subset, approach, binary, provide, individual<