NIPS 2013 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.
(It looks like 0 = reinforcement learning, 1 = deep learning, 2 = structured learning?, 3 = optimization?, 4 = graphical models, 5 = theory, 6 = neuroscience)
Toggle LDA topics to sort by: TOPIC0 TOPIC1 TOPIC2 TOPIC3 TOPIC4 TOPIC5 TOPIC6
Learning Stochastic Feedforward Neural Networks
Yichuan Tang, Ruslan Salakhutdinov


[deterministic, algorithm, belief, total, state] [hidden, learning, output, test, trained, training, learn, face, dataset, sigmoid, table, weight, boltzmann, average, feedforward, input, large, image, learned, better, generative, second, performance, mlp] [binary, inference, multimodal, synthetic, network, generate] [stochastic, number, gradient, random, small, true, standard, proposed] [model, sfnns, distribution, conditional, sfnn, mixture, gaussian, log, sampling, modeling, conditioned, prior, facial, generated, posterior, figure, data, expression, exact, restricted, variational, selected, monte, carlo, residual, foreground, estimate, toronto, proposal, gibbs, hybrid] [density, note, structured, database, function, approximation, linear, set] [neural, noise]
Auditing: Active Learning with Outcome-Dependent Query Costs
Sivan Sabato, Anand D. Sarwate, Nati Srebro


[algorithm, bound, upper, return, best, setting, optimal, greedy, assume, conference, strategy, logarithmic, golovin, total] [learning, negative, class, unlabeled, large, labeled, learn, hidden, approach, international, work, achieve] [size, binary, factor] [error, number, true, random, reduction] [sample, dependence, distribution, positive, probability, respect, draw, journal, data] [active, auditing, complexity, label, hypothesis, cost, pool, query, set, realizable, agnostic, consider, point, passive, lower, general, theorem, version, machine, labeling, case, denote, lemma, result, dimension, function, approximation, minimizing, procedure, provide, easy, minimal, question, hold, depend, rectangle, consistent, access, select] [simple, threshold, corresponding]
A message-passing algorithm for multi-agent trajectory planning
Jose Bento, Nate Derbinsky, Javier Alonso-Mora, Jonathan S. Yedidia


[planning, algorithm, initial, agent, absolute, total, unconstrained, space, goal] [performance, multiple, international, based, approach, work] [local, global, associated, update, outgoing, variable, allows, graph, search, incoming, move] [optimization, method, convergence, admm, solve, number, supplementary, minimum, objective, solving, magnitude, standard, order, computing] [time, figure, wall, joint, describe, journal] [set, cost, problem, energy, solution, radius, function, bipartite, paper, min, consensus, maximum, formulation] [minimizer, velocity, trajectory, twa, minimizers, collision, coll, corresponding, static, robotics, logic, position, avoid, material, motion, kinetic, avoidance, receives, certainty, gcoll, account]
Inferring neural population dynamics from multiple partial recordings of the same neural circuit
Srini Turaga, Lars Buesing, Adam M. Packer, Henry Dalgleish, Noah Pettit, Michael Hausser, Jakob Macke


[partial, bound, upper] [predict, approach, well, prediction, large, second, accuracy, predicted, predicting, dataset, better, performance, multiple, demonstrate, applied] [real, size, underlying] [method, true, matrix, number, random] [model, observed, data, naive, latent, estimate, computational, figure, joint, gaussian, time] [linear, subset, statistical, lower] [population, activity, imaging, neural, calcium, noise, stitching, coupling, neuron, imaged, recorded, connectivity, simultaneously, dynamical, system, stimulus, estimated, session, correlation, fully, inferring, simulated, mouse, awake, anesthetized, experimental, somatosensory, correlated, circuit, recording, infer, neuronal, whisker, cortex, unobserved, overlap, functional, correspond, presented, spiking]
Multi-Prediction Deep Boltzmann Machines
Ian Goodfellow, Mehdi Mirza, Aaron Courville, Yoshua Bengio


[best] [training, dbm, learning, trained, well, hinton, good, deep, mlp, net, hidden, salakhutdinov, layerwise, train, layer, dbms, boltzmann, negative, input, approach, predict, feature, better, generative, mnist, test, prediction, unit, weight, rbm, norb, centering, based, work, outperforms, performance, average, trick, ability, maximize] [inference, missing, probabilistic, perform, graphical] [standard, order, objective, number, solving, optimization, error, solve, stochastic, rate, performs, phase] [model, variational, log, sampling, markov, figure, centered, joint, observed, three, distribution, likelihood] [procedure, machine, set, point, approximation, simply, case, example, note, lower] [recurrent, single]
Adaptive Submodular Maximization in Bandit Setting
Victor Gabillon, Branislav Kveton, Zheng Wen, Brian Eriksson, S. Muthukrishnan


[expected, policy, item, adaptive, step, regret, gain, state, bound, greedy, algorithm, upper, choosing, choose, chosen, return, assumption, cumulative, episode, setting, assume, genre, bandit, optimistic, oasm, rated, arm, movie, highest, interacting, maximizing, elicitation, restrictive] [learning, based, learn, approach, learned, work, vector, second, context, applied, propose] [computed, sensor, associated] [number, analysis, method] [model, probability, distribution, selected, figure, major, observation, three, observed] [submodular, problem, maximization, function, max, machine, set, prove, studied, loss, main, dom, lemma, result, theorem, denote] [preference, user, study, earlier, viewed]
Approximate Inference in Continuous Determinantal Processes
Raja Hafiz Affandi, Emily Fox, Ben Taskar


[algorithm, space, sampled, provided] [based, similarity, quality, propose, approach, large, directly, accuracy, learning] [compute, synthetic, computed, real] [method, phase, matrix, dual, fourier, random, distance, eigendecomposition, number, error, wide, analysis, rate, spanned] [kernel, dpp, sampling, continuous, mixture, sample, dpps, discrete, repulsive, gibbs, iid, data, rff, probability, model, sampler, gaussian, entropy, posterior, modeling, conditional, diverse, full, approximate, three, distribution, inverse, selected, iris, estimate, considered, cdf, estimation, repulsion] [density, function, point, approximation, set, consider, case, complexity, note, machine, lower, approximating] [determinantal, range, activity, location, human, spatial]
Multiclass Total Variation Clustering
Xavier Bresson, Thomas Laurent, David Uminsky, James von Brecht


[total, algorithm, previous, simplex, conference, sum, initial, fact, step] [mtv, well, international, class, unsupervised, report, learning, image, vector] [clustering, vertex, graph, splitting, recursive, weighted, partition] [variation, convex, spectral, nmf, denotes, matrix, projection, cheeger, minimize, random, number, cut, requires, operator, exactly, optimization, grant] [data, discrete, continuous, model, figure, processing, form, university, equal] [set, function, energy, proximal, solution, indicator, relaxation, multiclass, transductive, problem, theorem, proxt, smooth, version, framework, rely, lsd, machine, note, denote, minimization, nmfr, tight, easily, subset, formation, manner, case, procedure] [sharp, simple, corresponding, constraint, exponent]
Reservoir Boosting : Between Online and Offline Ensemble Learning
Leonidas Lefakis, François Fleuret


[learner, online, algorithm, greedy, computationally, strategy, choosing, optimal, inria, conference, chosen, assumption, solely, best, maximizing, step] [weak, reservoir, boosting, learning, vector, geem, training, dataset, approach, propose, table, weight, based, work, practice, entire, common, computer, boosted, large, outperforms, performance, well, mnist] [weighted, size, ensemble] [number, iteration, method, matrix, proposed, covariance, row, order, optimization, comparison, column] [sample, sampling, process, computational, three, data, inverse, full, gaussian, expectation, address, model] [set, cost, subset, consider, budget, access, complexity, note, framework, case, main, dimension, loss, batch, discard] [presented, calculation, simple, refers]
Matrix factorization with binary components
Martin Slawski, Matthias Hein, Pavlo Lutsik


[algorithm, step, linearly] [approach, achieved, second, box, performance] [binary, dna, compute, factor, candidate, contained, underlying, ground] [matrix, factorization, number, methylation, findvertices, uniqueness, reduction, solving, block, additional, small, order, alpha, random, analysis, nmf, hottopixx, scheme, quad, solve, coordinate, separable, row, singular, rank, checked] [data, figure, exact, independent, probability, mixture, model, drawn, approximate, parameter, discrete, computational, positive, indicates, bernoulli, referred, continuous, yielding] [problem, linear, case, set, theorem, solution, consider, result, dimension, condition, imposed, complexity, note, lemma, proposition, theory] [noise, corresponding, cell, constraint, single]
Learning to Pass Expectation Propagation Messages
Nicolas Heess, Daniel Tarlow, John Winn


[choice, goal] [learned, training, learning, approach, work, learn, challenging, sigmoid, target, evaluate, output, large, second, performance, context, experiment, forward, trained] [factor, message, inference, product, network, compound, probabilistic, incoming, compute, variable, implemented, propagation, passing, outgoing, xout, generate, person, graphical, directed, computation, allows, implement] [number, random, standard, true, rate, fast, computing] [model, distribution, gamma, sampling, data, approximate, gaussian, mixture, bayesian, posterior, figure, gaussians, expectation, drawn, sampler, drawing, accurate, generated, prior, family, marginal] [function, set, approximation, analytic, general, regression, framework, note, ball, formulation, program, cost] [neural, shape]
Latent Structured Active Learning
Wenjie Luo, Alex Schwing, Raquel Urtasun


[algorithm, space, setting, step, exp, uncertainty, expected, employ, initial, advantage] [learning, prediction, labeled, training, unlabeled, test, context, supervised, image, performance, approach, large, task, output, amount, good, based, learn, composed, well, input, segmentation, feature] [variable, local, graph, inference, graphical, separate, computation, reduce, perform, extra] [random, number, error, proposed, convex, required, requires, order, convergence, iteration] [latent, joint, data, entropy, model, distribution, marginal, probability, time, selected, full, figure, likelihood] [active, structured, label, labeling, set, function, note, layout, problem, queriespixelwise, weakly, query, example, batch, interested, general, cost, paper, case, consider, select] [fully, single]
Two-Target Algorithms for Infinite-Armed Bandits with Bernoulli Rewards
Thomas Bonald, Alexandre Proutiere


[arm, algorithm, regret, horizon, optimal, explore, lim, bandit, failure, policy, achieves, reward, exploration, exploit, bound, explored, anytime, best, exploited, expected, assume, fact, berry, played, current, total, previous, setting, sequence] [target, large, average, second, propose, based, consists, performance] [larger, beta, distributed] [number, unknown, random, rate, analysis, proposed, uniformly, stochastic, phase] [time, distribution, parameter, bernoulli, selected, probability, prior, length, considered, sampling] [lower, support, proposition, theorem, case, version, proof, consider, note, set, problem, equivalent, conjecture, prove, provide, statistical, inequality, close, asymptotically, general, function, denote] [uniform, depends, corresponding, respective, false, arbitrarily]
RNADE: The real-valued neural autoregressive density-estimator
Benigno Uria, Iain Murray, Hugo Larochelle


[natural, previous, conference, chosen] [rnade, mog, training, hidden, autoregressive, learning, test, pixel, better, performance, image, zoran, work, output, scale, dataset, trained, table, nade, large, measured, boltzmann, timit, datasets, international, validation, speech, input, logl, predicted] [network, factor, size, graphical, represent, probabilistic, product] [number, standard, gradient, small, order] [mixture, model, data, conditional, distribution, gaussians, gaussian, likelihood, conditionals, modeling, parameter, figure, probability, processing, journal, discrete, restricted, minibatch, drawn, measuring, hyperparameters, component, sample] [density, machine, set, laplace, statistical, linear] [neural, range, sharing]
Multiscale Dictionary Learning for Estimating Conditional Distributions
Francesca Petralia, Joshua T. Vogelstein, David Dunson


[algorithm, assume, strategy, choose, space] [approach, scale, performance, learning, dictionary, applied, compare, accuracy, bottom, large, feature] [tree, variable, partition, graph, scaling] [cpu, subspace, decomposition, faster, number, random, dimensional, fast, error, standard] [conditional, data, multiscale, bayesian, msb, distribution, time, predictive, figure, estimation, computational, posterior, sample, journal, partitioning, nonlinear, estimate, ambient, model, cart, latent, mixture, including, probability, swissroll, manifold, simulation, prior, family, indicates, gaussian, three, nonparametric, measure] [density, linear, statistical, set, regression, function, example, union, dimension, lasso, estimating, note, selection, creativity, squared] [neuroscience, described, automated, fully]
Regularized M-estimators with nonconvexity: Statistical and algorithmic theory for local optima
Po-Ling Loh, Martin J. Wainwright


[assumption, bound, observe, assume] [vector, applied, level, work, applies, predicted, class] [local, global, rsc, size] [error, gradient, optimization, descent, convex, regularized, write, require, objective, convergence, minimum] [log, plot, parameter, figure, form, sample, restricted, estimation, additive, probability, model] [statistical, linear, nonconvex, function, loss, regression, composite, theorem, theory, logistic, arg, corrected, min, optimum, program, scad, suppose, regularizers, penalty, case, corollary, opt, mcp, empirical, guaranteed, errstat, annals, consider, condition, establish, feasible, provide, convexity, main, loh, depict, lasso, constant, penalized, theoretical, hold, solution, problem, point, general] [population, strong, corresponding]
Rapid Distance-Based Outlier Detection via Sampling
Mahito Sugiyama, Karsten Borgwardt


[conference, best, assume, bound, algorithm, chosen, space, maximized] [detection, object, large, datasets, average, knowledge, score, qkthsp, international, table, performance, approach, better, discovery, qkthnn, introduced, outperforms, based, effectiveness, learning, inliers, randomly] [size, acm, implemented, local, cluster, sigkdd, computation, mining, scalability, neighbor] [outlier, number, method, qsp, small, outlierness, proposed, distance, analysis, popular, comparison, iterative, random, performs] [sample, sampling, data, time, probability, gaussian, parameter, computational, process, log, including, detect, respect] [set, complexity, theoretical, note, theorem, function, maximum, statistical, subset] [experimental, simple]
Better Approximation and Faster Algorithm Using the Proximal Average
Yao-Liang Yu


[algorithm, assumption, sum, choice] [average, better, based, work, complicated] [group, improvement, graph] [convex, gradient, method, convergence, fast, analysis, proposed, norm, accelerated, objective, rate, siam, key, optimization, faster, alternating, solving, iterative, scheme] [component, figure, continuous, smoothing, approximate, journal] [proximal, map, approximation, nonsmooth, smooth, function, proposition, subgradient, fused, case, apg, lipschitz, example, point, simply, lasso, strictly, composite, moreau, easily, general, note, constant, consider, envelop, set, minimization, theoretical, iff, structured, clearly, clear, problem, denote, technical, machine, complexity, ultimate, statistical, fij, main, smoothness, linear] [simple, overlapping, recall, property, uniform, desired]
Dynamic Clustering via Asymptotics of the Dependent Dirichlet Process Mixture
Trevor Campbell, Miao Liu, Brian Kulis, Jonathan P. How, Lawrence Carin


[algorithm, step, current, setting, ddp, conference, exp, converge] [accuracy, learning, based, international, work, weight, tracking, dataset] [cluster, clustering, inference, nkt, update, zit, assignment, evolving, revived, yit, synthetic, computation, takeoff, science, plane, represents, transitioned] [number, random, comparison, distance, analysis, stochastic] [time, data, sampling, mixture, parameter, dirichlet, gibbs, process, particle, model, prior, gaussian, bayesian, asymptotic, variational, figure, aircraft, observation, distribution, sample, equation, three, computational, posterior, limit, nonparametric] [cost, set, label, point, function, machine, statistical, hard, labeling, penalty, construction, monotonically] [dynamic, dependent, timestep, trajectory, temporal, single]
Stochastic Convex Optimization with Multiple Objectives
Mehrdad Mahdavi, Tianbao Yang, Rong Jin


[optimal, algorithm, assumption, online, bound, expected, regret, goal, sampled, assume, upper, randomness, attains, pareto, programming, ensure, long, satisfy] [multiple, learning, approach, domain, weight, average] [variable, update] [stochastic, objective, optimization, convex, convergence, dual, rate, high, violation, gradient, standard, bwt, analysis, method, order, constrained, robust, descent, proposed, exactly] [probability, respect, journal, approximate, replaced, university, estimate, distribution, address] [solution, problem, loss, function, theorem, min, machine, set, denote, max, lipschitz, inequality, consider, general, main, note, primal, proposition, linear, operational, constant, risk, decide, approximation, access] [constraint, uniform, estimated, simple, strong, term, subject]
Error-Minimizing Estimates and Universal Entry-Wise Error Bounds for Low-Rank Matrix Completion
Franz Kiraly, Louis Theran


[algorithm, multiplicative, bound, space, assumption, optimal, logarithmic] [mask, predicted, prediction, reconstruction, denoising, output, level] [path, missing, graph, edge, structure, reconstructible, cycle, partially, compute] [matrix, variance, error, rank, entry, aij, completion, oriented, basis, sign, minimum, algebraic, baij, noisy, optspace, true, norm, nuclear, method, exactly, reconstructed, order] [kernel, estimate, log, three, data, figure, observed, estimation, positive, independent, model, additive, exact] [estimator, set, theorem, unbiased, denote, combinatorial, linear, minimizing, case, consider, squared, estimating, lemma, polynomial, lower, function, proposition, theoretical, framework] [noise, blue, red, single, calculate]
Message Passing Inference with Chemical Reaction Networks
Nils E. Napp, Ryan P. Adams


[belief, sum, algorithm, called, state, denoted, action, equilibrium, deterministic] [vector, representation, represented] [reaction, factor, chemical, message, network, variable, graph, ggggb, node, product, inference, concentration, gggggggb, mass, dna, represent, implement, passing, associated, graphical, computation, steady, unassigned, loopy, david, compiled, implemented, probabilistic, erik, strand, damped, compilation, incoming, computed, propagation, catalyze, compute, continuously, physical, ical] [random, order, implementation, small, rate, entry] [figure, marginal, probability, component, simulation, distribution, exact, discrete, equation, parameter, form, time] [set, procedure, linear, design] [molecular, correct, neural, change, system, simple]
A Kernel Test for Three-Variable Interactions
Dino Sejdinovic, Arthur Gretton, Wicher Bergsma


[total, appendix, space, sum] [test, based, dataset, embeddings, embedding, approach, table, learning, well, gram, second] [product, third, associated, partition, construct, graphical, mutually] [pairwise, norm, order, matrix, denotes, random, cumulants, canonical, dimensionality, standard, distance, factorization] [independence, interaction, kernel, lancaster, pxy, measure, joint, statistic, dependence, probability, testing, three, characteristic, nonparametric, independent, reproducing, hilbert, figure, distribution, inner, marginal, positive, null, conditional, kpxy, multivariate, detecting, mutual, detect, estimation, central] [signed, case, rkhs, hypothesis, empirical, consistent, dimension, proposition, provide, notion, straightforward, map, statistical, set] [corresponding, dependent, presence, correlation, correct]
Contrastive Learning Using Spectral Methods
James Y. Zou, Daniel Hsu, David C. Parkes, Ryan P. Adams


[algorithm, state, difference, goal, sequence, appendix, trade, conference, assume, natural] [learning, learn, vector, generative, approach, word, hidden, feature, large, table, computer] [variable, associated, auc, disjoint, corresponds] [tensor, spectral, method, analysis, entry, recover, number, projection, matrix, accurately, variance, iteration] [background, foreground, contrastive, model, data, topic, mixture, latent, generated, power, chromatin, document, china, moment, markov, modeling, component, usa, emission, figure, probability, corpus, estimation, hmm, economics, independent, sample, process, university, illustrate, estimate, dirichlet, three] [proposition, set, linear, general, consider, statistical, main] [contrast, biological, exploratory, relative, neural, learns]
Lexical and Hierarchical Topic Regression
Viet-An Nguyen, Jordan Boyd-Graber, Philip Resnik


[setting, movie] [table, level, hierarchical, hierarchy, word, sentiment, amazon, sentence, supervised, negative, based, discover, vocabulary, second, jointly, type, work, better, token] [path, political, restaurant, structure, associated, shlda, lexical, node, tree, variable, agenda, congressional, assigned, chinese, nested, review, assignment, global, product, existing, sitting, group, pcc, combo, framing, slda, assigning, cnew, perspective, inference, factor, baby, science] [number, discussion] [topic, model, document, dirichlet, parameter, probability, draw, distribution, figure, latent, lda, sampling, positive, process, university, corpus, three, prior] [regression, set, mse] [response, single]
How to Hedge an Option Against an Adversary: Black-Scholes Pricing is Minimax Optimal
Jacob Abernethy, Peter Bartlett, Rafael Frongillo, Andre Wibisono


[price, option, investor, hedging, bound, minimax, pricing, asset, trading, game, sequence, strategy, upper, current, assumption, lim, hedge, assume, european, gbm, expiration, brownian, payoff, goal, totvarconstraint, previous, setting, jump, appendix, scholes, optimal, eventually, black, volatility, abernethy, payout, taylor, total, chosen, future, online, notice, supremum] [large, well, amount, learning] [underlying, path] [stochastic, call, variance, write, number, key, convergence, error, standard, largest, robust] [time, parameter, university, allow, model] [theorem, lemma, result, inf, function, proof, consider, prove, approximation, note, risk, weaker, paper, main, depend, budget, framework, converges, geometric, case, lower] [nature, constraint, described, term, desired, derivation, motion]
Discovering Hidden Variables in Noisy-Or Networks using Quartet Tests
Yacine Jernite, Yonatan Halpern, David Sontag


[algorithm, failure, share, identify, step, bound] [learning, learn, learned, hidden, test, negative, unsupervised, based, class, approach, triplet, shared] [quartet, structure, variable, network, singly, coupled, parent, depth, sontag, determine, third, halpern, binary, discovering, size, medical, subtracting, diagnosis, allows, subtraction, hauskrecht, tree, factor, candidate] [number, rank, error, method, spectral, order, analysis, tensor, supplementary] [latent, observed, bayesian, sample, parameter, probability, variational, model, figure, joint, data, mixture, three, distribution, discrete, full, mutual, university] [set, bipartite, lemma, polynomial, linear, case, example, complexity, result] [threshold]
Least Informative Dimensions
Fabian Sinz, Anna Stockl, January Grewe, January Benda


[space, choice, chosen, difference, maximizing] [learning, feature, approach, input, lnp, average, second, direct, extract] [computed, product] [subspace, optimization, method, matrix, gradient, objective, covariance, dimensionality, order, random, minimize, comparing, number, reduction, analysis] [informative, kernel, mutual, distribution, null, hsic, independent, probability, equation, data, estimation, mmd, entropy, figure, lid, triggered, electric, hilbert, estimate, form, measure, journal, reproducing, independence, respect, dependency, residual, poisson, detect, drawn] [problem, function, dimension, machine, case, linear, correctly, integral, empirical, weakly] [spike, neural, maximally, relative, neuron, single, stimulus, population, corresponding, complex]
A Scalable Approach to Probabilistic Latent Space Inference of Large-Scale Networks
Junming Yin, Qirong Ho, Eric Xing


[algorithm, space, natural, sampled, bound, total] [large, representation, accuracy, table, based, approach, prediction, bag, better, compared, work] [triangular, network, mmsb, inference, role, triangle, local, eijk, open, scalable, link, global, synthetic, real, ijk, mmtm, runtime, vertex, ptm, social, size, parsimonious, heldout, stanford, update, distinct, edge, degree, motif, methodmmtm, gibbsptm, gibbsmmsb, represents] [stochastic, gradient, method, number, convergence, recovery, high, true, magnitude] [latent, variational, data, model, probability, time, gibbs, three, journal, figure, posterior, parameter, modeling] [batch, set, closed, lower, approximation, machine, complexity, statistical, converges, equivalence, hard] [term]
Adaptive dropout for training deep neural networks
Jimmy Ba, Brendan Frey


[algorithm, belief, conference, free, achieves] [hidden, dropout, standout, learning, training, deep, trained, unit, mnist, feature, performance, norb, unsupervised, relu, discriminative, layer, input, compared, test, technique, shallow, momentum, mask, better, weight, achieved, learnt, activation, randomly, object, class, well, good, boltzmann, computer, improve, large, second, denoising, learn, dae, dataset, extracted, international] [network, binary, variable] [method, rate, error, sparse, standard, number, regularization, stochastic] [probability, data, distribution, posterior, figure, model, parameter, including, bayesian, approximate, process, likelihood] [set, function, machine, logistic, linear, result] [neural, described, viewed, produce, activity]
Efficient Online Inference for Bayesian Nonparametric Relational Models
Dae Il Kim, Prem Gopalan, David Blei, Erik Sudderth


[online, previous, natural, best, focus, bound] [learning, multiple, hierarchical, based, approach, well, large, learned, work, compare] [community, inference, pruning, relational, node, global, network, perplexity, edge, membership, ahdpr, synthetic, elbo, assortative, edgesperplexity, local, assignment, ammsb, littlesis, yij, truncation, auc, undirected, allows, relativity, pair, scalable, initialization, mixed, unbounded, link, associated, update, mass, graph, associate, degree, hdpr, pruned, nested, toy] [stochastic, number, sij, true, small, gradient] [variational, model, latent, observed, dirichlet, process, nonparametric, distribution, figure, probability, posterior, naive, allow, estimate, sampling, data, restricted] [set, structured] []
Adaptivity to Local Smoothness and Dimension in Kernel Regression
Samory Kpotufe, Vikas Garg


[space, assume, optimal, bound, adaptive, bounded, assumption, notice, algorithm, apply, changing, selecting] [metric, bandwidth, work, training, adapt, approach, input, output] [local, global, size, neighborhood, requirement, balance, diameter] [unknown, variance, high, error, method, analysis, statement] [probability, kernel, estimate, figure, selected, locally, sample, nonparametric, marginal, parameter, assumed] [lemma, regression, dimension, result, procedure, smoothness, main, problem, general, point, suppose, function, denote, proof, adaptivity, selection, eyjx, consider, vary, inequality, argument, note, notion, exists, close, theorem, stability, lepski, theory, hypothetical, depending, cover, adapts, set, argue, union, bias, proper] [intrinsic, simultaneously, change, term, viewed, corresponding]
Optimization, Learning, and Games with Predictable Sequences
Sasha Rakhlin, Karthik Sridharan


[player, regret, algorithm, mirror, step, sequence, predictable, optimistic, minimax, online, game, round, adaptive, programming, observe, flow, play, bound, partial, upper, apply, playing, fact, converge, equilibrium, previous, arbitrary, setting, bounded, assume] [learning, second, vector] [mixed, update, size, extra, factor] [convex, optimization, descent, rate, number, argmin, method, matrix, gradient, small, application, order] [time, exponential, log, approximate, draw, entropy] [problem, lemma, inf, smooth, function, consider, max, set, result, complexity, suppose, case, solution, structured, procedure, simply, corollary, theory, version, provide, idea, linear, notion, point, prox, question, constant] [simple, uniform]
Approximate inference in latent Gaussian-Markov models from continuous time observations
Botond Cseke, Manfred Opper, Guido Sanguinetti


[algorithm, assume, free, fact, notice] [approach, well, box, type] [inference, diffusion, computed, update, parallel, propagation] [canonical, stochastic, matrix, standard, supplementary, order, iteration, method] [time, continuous, process, gaussian, discrete, approximate, marginal, posterior, model, xtk, equation, likelihood, limit, prior, variational, exact, panel, latent, cseke, log, figure, expectation, marginals, form, multivariate, data, discretisation, processing, observed, journal, opper, sampling, modelling, distribution, family, heskes, university, discretised, zammit, moment] [point, function, loss, differential, general, consider, approximation, example, machine, linear, density, formulation, corrected, statistical, problem] [neural, corresponding, spike, left, term]
Linear Convergence with Condition Number Independent Access of Full Gradients
Lijun Zhang, Mehrdad Mahdavi, Rong Jin


[algorithm, optimal, step, assumption, bound, assume, deterministic, sum, sequence, named, making, unconstrained, conference] [work, domain, learning, table, applied, based, average, achieve] [mixed, reduce, update, size] [stochastic, gradient, optimization, convex, number, objective, descent, convergence, iteration, emgd, oracle, rate, epoch, method, kwt, sdca, sag, proposed, order, call, random, siam, running, computing, rithm, kbw, matrix, analysis] [full, log, computational, probability, journal, data, university, respect, independent] [function, complexity, condition, smooth, linear, solution, problem, machine, access, lower, theorem, case, lemma, smoothness, consider, approximation, cost, composite, lipschitz, minimizing, set, point, general, loss] [needed, novel, cambridge]
Multi-Task Bayesian Optimization
Kevin Swersky, Jasper Snoek, Ryan P. Adams


[strategy, expensive, algorithm, best, uncertainty, gain, assume, space, online, previous] [task, learning, average, fold, hyperparameter, svhn, multiple, approach, evaluating, dataset, primary, acquisition, applied, large, knowledge, stbo, validation, scratch, evaluation, mtbo, cheaper, evaluate, training, feature, dynamically, usps, unit, better, based, baseline, mnist] [search, global, reduce, candidate] [optimization, error, objective, order, number, minimum, optimizing, standard, matrix, additional, small] [bayesian, model, gaussian, figure, entropy, hyperparameters, time, auxiliary, estimate, distribution, latent, university, observation, prior, process, grid, full] [function, machine, set, point, cost, framework, query, problem, consider, min] [neural, location, single]
Multilinear Dynamical Systems for Tensor Time Series
Mark Rogers, Lei Li, Stuart Russell


[stock, sequence, called, algorithm, state, complete, setting] [prediction, vector, video, applied, dataset, consists, trained, initialized] [probabilistic, product, real, synthetic, factor, distributed, member, size, structure] [tensor, matrix, projection, decomposition, number, dimensionality, factorization, error, analysis, tucker, mode, covariance, convergence, standard, random, factorizable, true, alternating] [model, mlds, time, lds, multilinear, latent, series, data, distribution, figure, likelihood, transition, normal, marginal, mat, modeling, bayesian, tesla, element, observation, expectation, rij, sliceerror, conditional, imjm, estimate, tion, drawn, respect, kronecker, ocean, volume] [set, case, complexity, general, function, consider, lemma] [dynamical, system, noise, ieee]
Restricting exchangeable nonparametric distributions
Sinead A. Williamson, Steve N. MacEachern, Eric Xing


[sequence, total, arbitrary, sum, focus] [directly, feature, appropriate, better, text, negative, approach, table] [beta, represent, synthetic, binary, structure, law, distributed, atom] [number, row, random, matrix, column, error] [distribution, data, ibp, exchangeable, latent, model, restricted, process, predictive, conditional, zik, bernoulli, probability, restricting, poisson, family, indian, buffet, posterior, sample, exhibited, unrestricted, finetti, nonparametric, binomial, interpretable, equation, parameter, processing, allow, restrict, sampling, figure, restriction, university, measure, exchangeability, mixture, conditioned, gibbs, truncated, gaussian] [point, example, support, ball, set, construction, consider, subset, note, theorem, general, proposition] [described, neural, completely, corresponding]
Forgetful Bayes and myopic planning: Human learning and decision-making in a bandit setting
Shunan Zhang, Angela J. Yu


[optimal, reward, bandit, belief, decision, arm, policy, expected, exploration, choice, best, previous, highest, choosing, future, sequential, algorithm, chooses, current, wsls, exploitative, explore, making, exploitation, total, option, gain, state, horizon, choose, special] [learning, dbm, knowledge, based, performance, people, average, task, better, second] [beta, science] [number, gradient, rate, standard, pairwise] [model, prior, data, probability, bayesian, figure, likelihood, posterior, assumes, distribution, computational, time, predictive, journal, bernoulli] [solution, complexity, problem, statistical] [human, trial, agreement, dynamic, behavior, behavioral, cognitive, control, simulated, assuming, term, estimated, experimental, early, myopic, stay, change, correct]
Probabilistic Movement Primitives
Alexandros Paraschos, Christian Daniel, January Peters, Gerhard Neumann


[controller, optimal, state, conference, policy, execution, step, reinforcement, introduce] [learning, vector, approach, international, target, multiple, learned, representing, learn, representation, activation, weight, illustrated, second] [probabilistic, conditioning, represents, product, allows] [variance, basis, phase, covariance, matrix, stochastic, order, denotes, exactly] [distribution, time, figure, model, gaussian, probability, joint, generated] [framework, set, machine, linear] [movement, trajectory, robot, control, desired, promp, promps, feedback, noise, system, modulation, rhythmic, combination, primitive, single, reach, blending, speed, blue, motor, varying, red, robotics, behavior, position, match, shaded, encode, temporal, crucial, ten, angle]
Policy Shaping: Integrating Human Feedback with Reinforcement Learning
Shane Griffith, Kaushik Subramanian, Jonathan Scholz, Charles Isbell, Andrea L. Thomaz


[reward, action, policy, optimal, reinforcement, agent, state, mdp, goal, best, provided, algorithm, food, knox, exploration, assume, environment, worse] [learning, domain, approach, amount, work, performance, table, direct, performed, baseline, better, multiple, average, well, negative] [underlying] [number, true, robust, standard, oracle, updated, method] [estimate, probability, parameter, time, figure, grid, distribution, bayesian] [case, consistency, set, function, paper, water, machine, selection, problem] [feedback, human, bql, advise, control, shaping, frogger, reduced, biasing, moderate, sharing, inconsistent, infrequent, estimated, manually, integrating, ideal, signal, modify, produced, frequency]
Linear decision rule as aspiration for simple decision heuristics
Ozgur Simsek


[decision, cumulative, choose, natural, choosing, difference, making] [datasets, accuracy, object, weighting, performance, type, training, randomly, learned, unit, large, work] [binary, median] [number, error, order, regularization, comparison, rate, analysis, high] [base, figure, approximate, model, parameter, three, university, selected, estimate, positive] [linear, dom, approximation, criterion, approx, statistical, note, consider, depend, theory, max] [rule, dominance, simple, cue, lexicographic, noncompensatory, blue, numeric, environmental, relative, prevalence, corresponding, noncompensation, red, proportion, ranged, attribute, oxford, noncompensatoriness, higher, identical, psychological, behavior, cum, cognitive, refer, paired, examine, gigerenzer, dominates]
Online PCA for Contaminated Data
Jiashi Feng, Huan Xu, Shie Mannor, Shuicheng Yan


[online, algorithm, initial, previous, revealed, step, current, space, singapore, sampled, robustness, observe, sequence, provided, satisfy, total] [performance, large, based, learning, work, scale, propose, affect] [update, size, accepted, randomized, probabilistic, concentration, supported] [pca, robust, rpca, principal, proposed, authentic, outlier, small, fraction, method, incremental, rpcaonline, matrix, standard, variance, analysis, subspace, covariance, high, newly, number, updating, groundtruth, error, true, contaminated, arxiv, sequentially, preprint, sparse, supplementary, direction, mild, hrpca] [estimation, data, sample, component, process, time, national, figure, sampling] [batch, lemma, theorem, solution, case, set, proof, result, provide, ratio, converges, breakdown, theoretical, subset, point] [signal, noise, department]
Deep content-based music recommendation
Aaron van den Oord, Sander Dieleman, Benjamin Schrauwen


[conference, msd, play] [music, audio, convolutional, usage, song, learning, deep, dataset, representation, approach, international, similarity, feature, predict, recommendation, large, trained, predicted, wmf, metric, jonas, traditional, recommending, predicting, vector, daft, content, punk, prediction, table, coldplay, semantic, train, training, listened, learned, well, evaluate, affect, enjoy, brian, implicit, mcfee, based] [factor, size, recommender, computed, variable, network, auc, weighted] [collaborative, matrix, objective, factorization, popular, gap, error] [latent, data, model, topic, processing] [regression, linear, subset, problem, machine, rely, map] [neural, user, retrieval, corresponding, preference]
PAC-Bayes-Empirical-Bernstein Inequality
Ilya O. Tolstikhin, Yevgeny Seldin


[bound, expected, tighter, upper, seldin, bounded, interesting, setting, yevgeny, apply, space, errortest, provided, seeger, advantage] [learning, technique, bounding, prediction, average, based, datasets, vector, work, generalization, test, second] [synthetic, concentration, binary, supported, smaller] [variance, small, analysis, random, comparison, high, order, divergence, supplementary] [probability, distribution, sample, posterior, bayesian, draw, figure, gibbs, equation, data, prior] [inequality, loss, theorem, empirical, note, hypothesis, proof, regression, result, machine, function, john, uci, denote, ratio, provide, ball, case, derive, example, radius, sizeexpected, risk, hold] [greater, term, simultaneously, rule, presented]
Point Based Value Iteration with Optimal Belief Compression for Dec-POMDPs
Liam C. MacDermed, Charles Isbell


[belief, decpomdp, algorithm, agent, policy, action, state, pomdp, best, utility, optimal, integer, conference, decentralized, strategy, current, decpomdps, decision, bounded, game, converted, previous, choose, multiagent, perseus, reward, sequence, discount, mediator, benchmark, autonomous, cbgs, anfce, horizon, backup, cooperative, history, pure, factored, space, free] [compression, international, improve, large, original, mapping, based, view, box] [existing, factor, underlying, search, merge] [number, solving, solve, iteration, phase, order, method, standard] [bayesian, observation, distribution, probability, joint, independent] [set, function, problem, program, linear, equivalent, approximation, solution] [continuation, dynamic]
Matrix Completion From any Given Set of Observations
Troy Lee, Adi Shraibman


[algorithm, bound, initial, optimal, assumption, proved, denoted, chosen, bounded, partial, deterministic, apply, revealed, choose] [output, target, generalization, second, performance, approach, vector, good] [real, product, motivates] [matrix, norm, trace, rank, completion, distance, random, dual, error, qij, sign, singular, agrees, minimum, pij, order, exactly, submatrix, minimize, incoherence, candes, noisy, retrieve] [probability, distribution, measure, sample, observed, exact, equal, respect, figure, lowest] [complexity, theorem, set, subset, problem, general, satisfying, max, min, denote, technical, prove, answer, version, result, minimal, consider, program, condition, linear, select, example, proxy, impossible, note, generalized] [property, recall]
Online Learning with Costly Features and Labels
Navid Zolghadr, Gabor Bartok, Russell Greiner, András György, Csaba Szepesvari


[algorithm, regret, learner, online, round, bound, observing, probing, assume, space, lnnt, pay, best, bounded, special, option, achieves, expert, called, apply, goal, optimal, mannor, elp, choosing, conference, delayed, alberta] [learning, prediction, vector, feature, weight, based, predict, work, well] [purchase, priori, graph] [number, quadratic, denotes, standard, order, rate, requires] [exponential, distribution, probability, dependence, time, estimate, sampling] [loss, cost, case, problem, linear, function, label, predictor, consider, set, theorem, constant, note, lower, subset, exists, shamir, lipschitz, purchased, complexity, paper, indicator, covering, denote, discretization, construction, general, unbiased] [study]
Sparse nonnegative deconvolution for compressive calcium imaging: algorithms and phase transitions
Eftychios A. Pnevmatikakis, Liam Paninski


[assume, total, expected] [vector, approach, image, large, reconstruction, multiple, traditional] [concentration, binary, induced, relation] [number, matrix, sparse, true, compressive, convex, compressed, phase, random, nonnegative, sparsity, method, basis, high, standard, error, analysis, required, recovery, singular, proposed] [estimate, probability, time, transition, observed, prior, log, length] [case, problem, statistical, solution, dimension, general, framework, linear, note, result, constant, empirical, point, cone, map, protocol] [imaging, calcium, sensing, spatial, timestep, spike, neural, estimated, neuron, snr, signal, noiseless, voxels, temporal, fully, spiking, undersampling, deconvolution, nature, examine, location, curve, study, cgt, normalized, microscopy, depends, described]
A Stability-based Validation Procedure for Differentially Private Machine Learning
Kamalika Chaudhuri, Staal A. Vinterbo


[algorithm, best, apply, observe, provided, chosen, goal] [validation, training, learning, score, performance, output, train, good, based, generic, well, second, work, large, performed, input] [existing, auc, size, list] [random, convex, regularization, number, regularized, small, standard, optimization, objective, supplementary, practical] [parameter, data, respect, histogram, differ, figure, tion, exponential, model, alternative] [private, differentially, privacy, procedure, set, stability, differential, machine, linear, condition, density, theorem, loss, budget, logistic, adult, function, design, provide, statistical, case, regression, regularizer, denote, magic, perturbation, notion, select, hinge, result, composition] [single, change, control, experimental, ramp]
A Novel Two-Step Method for Cross Language Representation Learning
Min Xiao, Yuhong Guo


[algorithm, projected] [language, learning, unlabeled, cross, tsl, text, target, feature, semantic, bilingual, sentiment, representation, word, labeled, multilingual, vocabulary, based, training, monolingual, domain, indexing, english, performance, vector, translation, task, exploiting, represented, outperforms, average, test, concatenated, learn, baseline, efb, german, amazon, approach] [parallel, missing, source, product, perform] [matrix, method, proposed, gradient, descent, completion, analysis, convex, standard, number, row, projection, mij, optimization, operator, small] [data, document, latent, observed, three, figure, model] [set, problem, machine, function] [induce, chose, produce, adaptation]
Capacity of strong attractor patterns to model behavioural and cognitive prototypes
Abbas Edalat


[assume, sum] [large, average, temperature, learning, technique, second, learned, boltzmann, compared] [network, degree, heuristic, hand, symmetry, distributed, real, law] [random, standard, number, method, square, convergence, stochastic, updating, exceeds, solve, solving, basic, error, computing, order] [equation, model, independent, distribution, probability, equal, positive, parameter, figure] [case, solution, theorem, theory, function, spin, stability, lemma, result, consider, general, constant, condition] [strong, pattern, capacity, replica, neural, storage, mathematical, simple, tanh, single, stored, retrieving, presence, attachment, behavioural, critical, memory, rule, attractor, erf, side, multiply, load, mathematically, correlated, recall, identically]
B-test: A Non-parametric, Low Variance Kernel Two-sample Test
Wojciech Zaremba, Arthur Gretton, Matthew Blaschko


[setting, choice, sum, apply, space] [type, test, average, gram, table, large, based, approach, multiple, experiment, feature, better] [size, median, computation, synthetic, computed] [error, number, variance, block, matrix, convergence, distance, method, order, averaging, high, minimum, rate, pairwise] [kernel, distribution, null, sample, gaussian, figure, data, mmdu, asymptotic, central, limit, bootstrap, estimate, testing, time, power, probability, mmdl, statistic, family, mmd, inner, computational, normal, pearson, blocktype, alternative, equation, independent, remains, gamma, measure] [empirical, theorem, hypothesis, set, approximation, selection, consistent, asymptotically, maximum, implies, complexity, converges, lower, statistical, estimating, function] [single, presented, mathematical, increase]
q-OCSVM: A q-Quantile Estimator for High-Dimensional Distributions
Assaf Glazer, Michael Lindenbaoum, Shaul Markovitch


[algorithm, decision, introduce, expected, sequence, assume, space] [hierarchical, test, coverage, level, average, evaluated, trained, training, performance, learning, vector, introduced, hierarchy, multiple] [global, parallel, decrease] [method, number, averaged, dual, largest, outlier, objective] [data, distribution, equation, estimate, kernel, figure, estimation, probability, gaussian, hilbert] [ocsvm, quantile, program, solution, set, function, ratio, approximated, generalized, hmve, density, estimator, converges, uci, statistical, empirical, quantiles, ocsvms, suggested, support, provide, estimating, hyperplanes, note, lying, optimality, repository, generalize, solved, theoretical, paper, primal, john] [single, estimated, contrast, proportion]
Mid-level Visual Element Discovery as Discriminative Mode Seeking
Carl Doersch, Abhinav Gupta, Alexei A. Efros


[algorithm, space, previous, goal, adaptive] [visual, patch, bandwidth, dataset, discriminative, purity, feature, image, coverage, large, negative, work, detection, top, retrained, based, approach, discovery, learning, object, level, performance, multiple, well, svm, paris, street, divide, hog, training, denominator, unsupervised, maximize, visually, sidewalk, discover] [local, cluster, clustering, compute, mining, factor, nearest, perform, smaller] [number, gradient, ascent, method, mode, distance, proposed, standard, optimization, dimensionality, norm] [figure, data, element, lda, positive, kernel, distribution, include, analogous] [density, set, ratio, select, note, constant, lower, formulation] [scene, single, left, produce, higher]
Non-Linear Domain Adaptation with Boosting
Carlos J. Becker, Christos M. Christoudias, Pascal Fua


[decision, space] [domain, learning, training, approach, feature, target, labeled, task, shared, learn, performance, common, boosting, large, acquisition, weak, segmentation, applied, boosted, image, test, challenging, well, supervised, aerial, trained, mapping, representation, jmlr, transformation, datasets, illustrated, amount, undergone, negative, outperforms, multiple, object] [source, path, tree, split] [method, leverage, gradient, number, unknown, minimize, projection, error, seek] [data, kernel, boundary, latent, model, road, full, histogram, positive, figure] [problem, linear, set, regression, case, loss, function, consider, maximum, solution] [adaptation, single, highly, olfactory]
Exact and Stable Recovery of Pairwise Interaction Tensors
Shouyuan Chen, Michael R. Lyu, Irwin King, Zenglin Xu


[algorithm, optimal, step, movie, sum, special, sampled, partial, assume, apply] [performance, level, original, table, training, dataset, recommendation, large, work, tag] [degree, size, truncation, scalable, missing] [tensor, recovery, pairwise, matrix, noisy, rank, completion, convex, recover, number, error, singular, recovered, nuclear, optimization, norm, method, solving, recovering, constrained, svd, order, factorization, tijk, svt, bjk, operator, collaborative, uniformly, aij, denotes, standard, exactly, minimize] [interaction, exact, figure, time, observation, observed, model] [problem, program, solution, shrinkage, set, theorem, guarantee, result, minimization, minimizes, note, exists, suppose, easy, proposition, case, prove, theoretical, proof, general] [noise, temporal, stable, absence, relative, subject]
On Decomposing the Proximal Map
Yao-Liang Yu


[sum, assume, goal, interesting, proved, apply] [learning, class, work, feature, domain, multiple] [increasing, compute, group, scaling, combine, collection] [norm, decomposition, convex, remark, symmetric, permutation, recover, sparse, key, order, denotes, siam] [journal, positive, restricted, continuous, additive] [proximal, map, corollary, theorem, function, dom, condition, case, subdifferential, proposition, note, invariance, implies, consider, example, clearly, set, proof, easily, iff, gauge, closed, main, composition, easy, regularizer, prove, exists, result, verify, paper, min, appeared, cone, hilbertian, constant, depend, convexity, hold, weaker, machine, nonsmooth, nontrivial, orthonormal, homogeneous, formula, linear, combettes, yaoliang, clear, indicator] [equality, implication, individual, simple, completely, trivial]
Polar Operators for Structured Sparse Estimation
Xinhua Zhang, Yao-Liang Yu, Dale Schuurmans


[algorithm, appendix, optimal, current, space] [class, domain, approach, dictionary, demonstrate, unit, accuracy] [path, group, compute, variable, computed, network, local, computation, allows, edge, totally, atomic, corresponds, develop] [sparse, convex, optimization, operator, computing, norm, method, matrix, gradient, reduction, solving, sparsity, objective, number, faster, nonnegative, extend, proposed] [time, form, latent, data, computational, smoothing, figure, model, piecewise, three, log, accurate, conditional] [polar, gcg, linear, function, structured, regularizers, set, cost, max, fused, lasso, apg, consider, solution, note, problem, gauge, arg, case, proximal, easy, regularizer, secant, proposition, main, combinatorial, dom, tum, general, expressed, submodular, min, denote, subset] [simple, coding, corresponding, reduced, constraint]
A Gang of Bandits
Nicolò Cesa-Bianchi, Claudio Gentile, Giovanni Zappella


[bandit, algorithm, payoff, contextual, reward, conference, regret, delicious, best, provided, assume, item, assumption] [context, vector, learning, content, recommendation, based, approach, datasets, tested, baseline, work, instance, improve, table, dataset, international, original, created, performance] [node, graph, social, network, cit, recommender, laplacian, smaller, clustering, contained, cluster, associated, update, distributed] [matrix, number, random, running, analysis, order, stochastic, standard, popular, block, high, unknown] [figure, time, independent, model, three, parameter, kernel] [linear, set, machine, multitask, clearly] [user, noise, described, corresponding, control]
Inverse Density as an Inverse Problem: the Fredholm Equation Approach
Qichao Que, Mikhail Belkin


[space, setting, sampled, algorithm] [type, learning, shift, performance, unsupervised, test, based, labeled, training, well, table] [] [norm, regularization, optimization, analysis, error, operator, euclidean, method, random, spectral] [kernel, sample, inverse, data, parameter, estimation, sampling, hilbert, gaussian, measure, model, family, estimate, probability, independent, equation] [density, function, problem, ratio, integral, linear, rkhs, covariate, note, set, theoretical, estimating, arg, lsif, tikde, kmm, min, selection, loss, resampling, formulation, empirical, theorem, fredholm, solution, framework, fireq, firep, classical, approximation, version, representer, provide, machine, paper, fire, case, support] [simple, corresponding, assuming]
Robust Image Denoising with Multi-Column Deep Neural Networks
Forest Agostinelli, Michael R. Anderson, Honglak Lee


[optimal, adaptive, best] [denoising, ssda, image, training, type, trained, corrupted, deep, input, hidden, performance, test, mnist, stacked, speckle, ssdas, better, denoised, weight, tested, output, salt, learning, table, tang, vector, psnr, autoencoder, average, pepper, original, prediction, performed, layer, well, digit, activation, weighting, feature, pixel, good, corrupting, xie, achieved, radiation, dose, unseen] [network, determine, perform, medical, local, mixed] [column, noisy, error, sparse, method, robust, proposed, number, block] [gaussian, figure, data, poisson] [set, linear, function, statistical, border, loss] [noise, neural, signal, single, decoding, uniform, produced, ieee, encoding, performing]
EDML for Learning Parameters in Directed and Undirected Graphical Models
Khaled Refaat, Arthur Choi, Adnan Darwiche


[algorithm, complete, start, initial, optimal, interesting, incomplete, previous] [learning, work, instance, negative, dataset, vector, average, approach, evaluate] [edml, inference, network, unique, graphical, feasibility, undirected, adnan, evidence, variable, originally, cxa, perspective, optimize, computed, probabilistic] [number, iterative, optimization, method, convex, soft, proposed, constrained, gradient, objective, faster] [parameter, bayesian, markov, equation, stationary, log, estimation, data, component, likelihood, conjugate, independent, form, derived, estimate, continuous, respect, approximate] [function, set, point, feasible, case, theorem, procedure, consider, denote, instantiation, example, optimum, problem] [corresponding, subject, constraint]
A Comparative Framework for Preconditioned Lasso Algorithms
Fabian L. Wauthier, Nebojsa Jojic, Michael Jordan


[algorithm, focus, assume, upper, choosing, choice] [generative, propose, original, compare, instance, traditional] [variable, factor, induced, associated] [recovery, column, basis, preconditioning, comparing, diagonal, standard, supplementary, matrix, proposed, comparison, svd, singular, recover, lead] [parameter, model, gaussian, figure, data, journal, independent, three] [lasso, support, signed, penalty, preconditioned, lemma, selection, pbht, problem, set, theorem, general, hold, framework, paper, note, ratio, empirical, suppose, jia, continue, construction, statistical, case, comparative, rohe, jojic, theoretical, proof, provide, orthonormal, huang, denote] [estimated, noise, relative, depends, thought, left, paul, recall]
A memory frontier for complex synapses
Subhaneil Lahiri, Surya Ganguli


[upper, state, bound, space, initial, equilibrium, subsequent, previous, assume, optimal, quantity] [entire, learning, average, achieve, well, maximize, work, performance] [structure, candidate, develop, understanding, network] [order, number, stochastic, supplementary, decay] [model, time, figure, transition, form, distribution, probability, chain, discrete, markov, process] [theoretical, theory, complexity, general, linear, set, result, function] [memory, synaptic, curve, area, snr, dynamical, potentiation, complex, envelope, molecular, internal, depression, single, described, storage, increase, stored, mpot, scalar, passage, achievable, late, functional, plasticity, potf, population, neuron, neural, early, synapse, simple, change, mathematical, strong, experience, mdep, reach, capacity]
First-order Decomposition Trees
Nima Taghipour, Jesse Davis, Hendrik Blockeel


[conference, sequence] [based, domain, international, exploiting, performed] [lifted, inference, randvars, isomorphic, node, dpg, probabilistic, propositional, logvars, factor, representative, structure, size, intelligence, dtree, lifting, tree, ground, lve, parfactor, group, grounding, liftable, elimination, counted, represents, width, allows, recursive, compute, interchangeable, prvs, variable, product, randvar, graphical, represent, logical, plm, den, nima, prv, read, leaf, van, guy, jesse, logvar, symmetry, operation, bijection, recursively] [decomposition, number, largest, order, analysis, standard] [model, counting, figure, joint, exact, exponential, form, replacing] [complexity, set, solution, example, result, problem, simply, theorem] [corresponding, single]
Marginals-to-Models Reducibility
Tim Roughgarden, Michael Kearns


[programming, algorithm, optimal, appendix, provided, decision] [class, description, instance, input] [inference, graph, assignment, size, vertex, induced, graphical, compute, network, probabilistic, progress, edge, variable, pair] [number, method, small, oracle, pairwise, dual, random, objective, solving, minimize, solve, reduction, optimization] [marginals, marginal, distribution, data, markov, joint, entropy, time, exponential, full, figure, length, independent] [problem, consistency, linear, consistent, map, polynomial, ellipsoid, separation, support, closest, program, consider, set, theorem, lemma, solved, maximum, function, max, reduces, generalized, algorithmic, primal, general, formulation, main, point, denote, straightforward, prove, provide, simply, suppose] [corresponding, single, reduced, constraint, subject, interest]
Accelerating Stochastic Gradient Descent using Predictive Variance Reduction
Rie Johnson, Tong Zhang


[option, apply, choose] [learning, training, large, prediction, average, explicit, randomly, view, simpler, test, vector, net, representation, second] [update, computation, larger, stage] [variance, gradient, rate, sgd, svrg, method, convergence, stochastic, convex, sdca, reduction, descent, zhang, standard, fast, optimization, number, roux, error, dual, faster, requires, order, iteration, random, nmnist, coordinate, comparison, require, sag, analysis, regularized, ascent, small, minimum, scheduling, decay] [expectation, figure, residual, parameter, data] [loss, inequality, consider, linear, nonconvex, case, logistic, smooth, set, problem, machine, min, regression, theorem, solution, cost, prove, example, easily, structured] [neural, rule, complex, reduced, memory]
Online Variational Approximations to non-Exponential Family Change Point Models: With Application to Radar Tracking
Ryan D. Turner, Steven Bottone, Clay J. Stanek


[online, expected, algorithm, bound, complete, black] [tracking, detection, international, performance, accuracy, multiple, improved, compare] [assignment, inference, compute, associated, allows] [measurement, updating, method, standard, number, true] [posterior, data, rice, rcs, bocpd, variational, model, upm, exponential, family, log, distribution, radar, time, association, predictive, length, bayesian, prior, likelihood, exact, track, upms, latent, probability, aircraft, figure, rmse, estimate, conjugate, naive, imm, gaussian, joint, estimation, three, censored, component, marginal, process, sample, series, siap] [point, problem, lower, approximation, example, map, case, technical, machine, paper, derive] [change, clutter, snr]
Optimal Neural Population Codes for High-dimensional Stimulus Variables
Zhuo Wang, Alan Stocker, Daniel Lee


[optimal, assume, bound, appendix, arbitrary, natural, space] [fisher, input, reconstruction, maximize, work, image, compared, activation] [degree, focused, product] [matrix, covariance, basis, variance, dimensional, optimization, minimize, rate, discussed, det, diagonal, number, scheme, fourier, plotted, high] [gaussian, distribution, prior, mutual, log, independent, model, figure, nonlinear, multivariate, component, university, analytically, marginal, process] [solution, loss, lower, result, problem, constant, simply, case, estimator, general] [neural, stimulus, population, encoding, tuning, diffeomorphic, coding, neuron, infomax, noise, calculate, code, decoding, sin, term, sigmoidal, curve, correlation, sound, unitary, encode, calculation, response, perception, sensory, pennsylvania]
Fast Algorithms for Gaussian Noise Invariant Independent Component Analysis
James R. Voss, Luis Rademacher, Mikhail Belkin


[algorithm, step, arbitrary, choice, chosen] [vector, based, second, performance, invariant, work, technique, unit] [update, variable] [ica, order, matrix, random, cumulant, convergence, cumulants, fourth, fast, recover, analysis, gradient, number, robust, fastica, hessian, comparison, method, practical, iteration, whitening, noisy, requires, diagonal, indexica, samplesamari, pca, amari, orthogonalization, unknown, cubic, proposed, column, required, matlab, standard] [independent, gaussian, data, component, latent, observed, distribution, additive, estimate, white, univariate, drawn, equation, process, sample] [case, function, lemma, denote, note, theorem, separation, provide, complexity, paper] [noise, signal, blind, noiseless, neural, identity]
Minimax Theory for High-dimensional Gaussian Mixtures with Sparse Mean Separation
Martin Azizyan, Aarti Singh, Larry Wasserman


[bound, minimax, setting, assume, upper, optimal, computationally, expected, algorithm] [learning, feature, precise, computer, work, second] [clustering, relevant, variable, stage, mellon, vempala, symposium] [method, sparse, tan, small, high, spectral, standard, number, dimensional, error, principal, basis, rate, gap, require] [log, mixture, sample, probability, gaussian, gaussians, data, journal, component, distribution, estimation, spherical, university] [proposition, separation, selection, theorem, loss, proof, provide, max, case, complexity, lower, statistical, min, consider, theoretical, estimating, set, function, paper, isotropic, lemma, general, dimension, machine, subset, theoretic, chaudhuri, denote, witten, penalized, santosh, introduction] [sin, depends, simple, relative]
Predicting Parameters in Deep Learning
Misha Denil, Babak Shakibi, Laurent Dinh, Marc'Aurelio Ranzato, Nando de Freitas


[factored, space, choice, conference, natural, choose] [layer, feature, learning, deep, dictionary, convolutional, hidden, weight, technique, prediction, predict, performance, rica, image, second, approach, trained, work, knowledge, training, predicted, input, scale, pixel, train, reconstruction, learned, visible, representation, architecture, directly, extremely, unsupervised, constructing, well, output, appropriate] [network, structure, size, reduce, connected, distributed, construct, local] [number, column, matrix, random, ridge, covariance, ica] [kernel, figure, parameter, model, prior, processing, data, exponential, full, form] [linear, squared, smoothness, set, paper, intuition, regression, function, case] [neural, dynamic, single, static, interpretation, fully, corresponding, described]
Estimating the Unseen: Improved Estimators for Entropy and other Properties
Paul Valiant, Gregory Valiant


[algorithm, total, expected, appendix, bound, optimal, upper] [domain, unseen, approach, vector, large, work, performance, assigns, computer, second] [size, distinct, portion, occur, mass, science, nif] [number, distance, error, true, exactly, small, objective, random, analysis] [sample, distribution, probability, histogram, estimate, entropy, observed, naive, estimation, figure, data, poisson, drawn, independent, zipf, valiant, cae, parameter, mesh, three, returned, element, derived, sampling, likelihood, vopt, accurate] [estimator, linear, support, estimating, program, example, empirical, bub, consider, theoretical, statistical, function, note, set, constant, theorem, proof, hamlet, simplest] [uniform, property, consisting, plausible, contrast, population]
What do row and column marginals reveal about your dataset?
Behzad Golshan, John Byers, Evimaria Terzi


[algorithm, sequential, assume, step, observe, exploration, sum, introduce, computes] [hidden, input, dataset, large, datasets, generative, approach, better, experiment, accuracy, propose, vector, raw, average, jth, table] [binary, compute, size, computed, existing] [column, matrix, pdfs, row, simpleis, dataspace, number, knot, running, entry, computing, error, rasch, pdf, high, nola, implicitly, adjust, method, standard, realization, enumeration] [marginals, sampling, time, probability, estimate, data, model, rest, equation, distribution, prior, figure, sample, mcmc, university, observed, journal] [problem, set, dblp, polynomial, provide, function, statistical, estimating, framework, linear] [refer, cell, described]
One-shot learning by inverting a compositional causal process
Brenden M. Lake, Ruslan Salakhutdinov, Josh Tenenbaum


[sampled, abstract, natural, previous, conference, start, algorithm, space] [image, learning, people, hbpl, character, deep, stroke, learn, hierarchical, visual, raw, handwritten, concept, better, boltzmann, computer, learned, type, turing, performance, scale, compared, test, tested, training, class, mechanical, argmax, evaluated] [generate, structure, inference, causal, mit] [analysis, error, number, rate, standard, variance] [model, figure, bayesian, generating, log, spline, distribution, three, drawn, posterior, approximate, grid, background, include, selected, causality, data, including] [machine, set, program, approximation, example] [motor, simple, control, ieee, pattern, neural, single, produce, human, rule]
Statistical analysis of coupled time series with Kernel Cross-Spectral Density operators
Michel Besserve, Nikos K. Logothetis, Bernhard Schölkopf


[arbitrary, space, bounded, assume, lim] [test, type, approach, based, visual, well, vector, bottom, compare, performance, concept, similarity, window] [product, lfp, causal] [random, operator, norm, order, analysis, fourier, transform, matrix, pairwise] [time, kernel, series, independence, kcsd, estimate, hsic, figure, markov, biased, hilbert, measure, rbf, processing, dependency, gamma, band, independent, distribution, characteristic, centered, positive, parameter, reproducing, probability, null, bivariate, transition, data, enables] [theorem, statistical, linear, unbiased, density, general, framework, structured, lemma, case, consequence] [neural, frequency, complex, coupling, scalar, dynamical, activity, monkey]
Optimizing Instructional Policies
Robert Lindsey, Michael Mozer, William J. Huggins, Harold Pashler


[policy, space, optimal, duration, strategy, focus, involves, explored, pomdp, goal, exploration, glopnor, grey, outcome] [training, category, learning, experiment, performance, test, traditional, fading, approach, accuracy, based, concept, large, evaluating, propose, evaluation, exemplar, technique, challenge] [search, represents, determine, pair] [optimization, number, noisy, optimizing, distance, method, additional, standard] [figure, gaussian, posterior, process, data, model, time, probability, observed, boundary, alternative, family, estimate] [function, set, design, optimum, label, budget, selection, statistically, teaching, consistent, parameterized, manner] [subject, experimental, instructional, ppf, individual, blocking, gpr, population, trial, student, cognitive, correct, reliable, human, presentation, psychological, indicate]
Integrated Non-Factorized Variational Inference
Shaobo Han, Xuejun Liao, Lawrence Carin


[bound, upper, optimal, deterministic, space, assumption, conference] [based, accuracy, compare, evaluated, dataset, sense, learning, quality, hyperparameter] [inference, evidence, global, nested, elbo, parallel] [method, divergence, gradient, number, square, matrix, numerical, proposed, norm, fast] [variational, posterior, bayesian, gaussian, grid, latent, inla, approximate, distribution, marginal, figure, accurate, mcmc, bayes, hybrid, mixture, marginals, model, computational, time, full, parameter, university, hyperparameters, normal, analytically, independent, journal, monte, exact, continuous, expectation, processing, joint] [approximation, lasso, laplace, lower, consider, solution, function, provide, statistical, minimizing, problem, regression, machine, theorem, linear, lemma] [integrated, neural, integration]
More data speeds up training time in learning halfspaces over sparse vectors
Amit Daniely, Nati Linial, Shai Shalev-Shwartz


[algorithm, return, upper, bound, natural, item, assumption, pac, learner, choosing, assume] [learning, class, training, learn, large, supervised, task, output, technique, vector, instance, based, learnt, representation] [boolean, size, construct, assignment, supported, generate] [random, sparse, error, standard, small, number] [sample, data, probability, computational, positive, independent, distribution, form] [hypothesis, set, question, hardness, formula, theorem, lower, halfspaces, problem, refuting, proof, conjecture, complexity, exists, denote, answer, clause, improper, example, ratio, note, main, prove, theoretic, hard, impossible, result, case, agnostic, studied, consider, establish, israel, validity, statistical] [learns, consisting, produce]
(Nearly) Optimal Algorithms for Private Online Learning in Full-information and Bandit Settings
Abhradeep Guha Thakurta, Adam Smith


[regret, algorithm, bandit, online, bound, sum, follow, nonprivate, previous, setting, sequence, leader, appendix, space, adaptive, partial, pftal, adam, satisfy, aggregation, step, cynthia, lipshitz, complete, replace, continual, optimal, variant, oblivious, fact, best, subsequent] [learning, input, work, output, vector, based, class, technique, entire, table, unit] [tree, smaller] [convex, gradient, analysis, quadratic, minimize, random] [time, approximate, full, parameter, data, log, dependence] [private, privacy, function, cost, differential, differentially, set, lower, theorem, loss, version, min, lipschitz, arg, general, consider, provide, approximation, linear, convexity, design, access, machine, ball, note, notion] [change, strong, stream]
Curvature and Optimal Algorithms for Learning and Minimizing Submodular Functions
Rishabh K. Iyer, Stefanie Jegelka, Jeff A. Bilmes


[bound, algorithm, upper, optimal, follow, previous, tighter, cooperative, sum, logarithmic, apply] [learning, better, class, work, improved, training, concept, learn] [factor, spanning, graph, path] [constrained, cut, number, sparse, matching, minimum, nonnegative, random, optimization, imply] [log, approximate, time, form, figure, family] [submodular, function, approximation, curvature, lower, modular, corollary, lemma, set, monotone, minimization, polynomial, theoretical, combinatorial, problem, empirical, mub, version, concave, minimizing, theorem, approximating, general, tight, maximization, machine, cardinality, polymatroid, submodularity, implies, practically, prove, balcan, provide, main, solution, jegelka, factorn, pmac, surrogate, constant, case, matroid] [varying, normalized, depends, constraint]
Σ-Optimality for Active Learning on Gaussian Random Fields
Yifei Ma, Roman Garnett, Jeff Schneider


[greedy, satisfy, sum] [learning, unlabeled, class, labeled, outperforms, prediction, directly, accuracy, belonging, work] [graph, node, global, binary, undirected, loo, variable, network, toy, connected, search, larger, mellon] [random, covariance, matrix, objective, variance, reduction, nonnegative, optimization, application, high, number, largest, harmonic, spectral, diag] [predictive, model, gaussian, figure, data, markov, conditional, inverse, university, tion, alternative] [active, risk, grf, grfs, lemma, set, minimization, label, loss, theoretical, query, condition, criterion, optimality, example, machine, minimizes, subset, arg, theorem, function, citation, solution, survey, dblp, suppressor, lij, minimizing, zhu, suppose, regression, surrogate, submodularity, consider, approximation] [correlation, proportion, described, highly]
Learning Kernels Using Local Rademacher Complexity
Corinna Cortes, Marius Kloft, Mehryar Mohri


[algorithm, bound, sum, optimal, bounded, choice, unif, observe, highest] [learning, based, class, tail, performance, multiple, report, svm, training, experiment, compare, generalization, average, test] [local, compute, existing, binary, global] [convex, optimization, standard, mkl, small, order, solving, matrix, trace, analysis, eigenvalue, fast, convergence] [kernel, data, family, probability, figure, journal, measure, base, gaussian] [rademacher, complexity, conv, notion, problem, machine, risk, hypothesis, tss, theorem, loss, function, design, result, theoretical, linear, set, point, case, consider, note, general, empirical, energ, proposition, exon, concave, min, denote] [strong, angle, supplemental, combination, material, determined, constraint]
Causal Inference on Time Series using Restricted Structural Equation Models
Jonas Peters, Dominik Janzing, Bernhard Schölkopf


[algorithm, assume] [test, experiment, based, class, jointly, work, temperature, discovery, large] [causal, graph, structure, structural, inference, node, allows, dag, delay] [order, method, lead, true, require, asymmetry, call, number] [time, model, data, series, timino, instantaneous, nonlinear, additive, causality, wrong, independence, full, independent, iid, gaussian, undecided, infers, equation, granger, sample, multivariate, chu, restricted, shifted, pai, hsic, var, include, remains, figure, including, arrow, confounders, length, differ] [linear, theorem, summary, function, case, set, regression, condition, provide, simulate, lemma] [noise, correct, infer, corresponding, code]
Symbolic Opportunistic Policy Iteration for Factored-Action MDPs
Aswin Raghavan, Roni Khardon, Alan Fern, Prasad Tadepalli


[policy, action, opi, factored, state, symbolic, backup, bellman, algorithm, mpi, add, space, diagram, optimal, elevator, impact, decision, bounded, mdps, ring, actionssolution, denoted, computes, spudd, planning, reward, opportunistic, mdp, partial, violate, bottleneck, current, greedy, conference] [evaluation, better, domain, large, work, representation, computer, representing, dbn, compression, represented] [size, variable, pruning, parallel, root, scalability, leaf, structure, relational, person, larger] [iteration, number, running, true, lead, denotes, exceeds] [figure, time, limit, exact, probability, approximate, joint, full, equation] [function, compact, max, denote, procedure, example, sdp, version] [memory, constraint, dynamic, control]
Real-Time Inference for a Gamma Process Model of Neural Spiking
David Carlson, Vinayak Rao, Joshua T. Vogelstein, Lawrence Carin


[algorithm, online, space, previous, partial] [multiple, well, vector, dictionary, performance, output, compare, truth, based, work, weight, detection] [cluster, inference, associated, ground, distributed, assigned, update, event] [number, rate, random, standard, supplementary, true] [time, process, model, data, poisson, bayesian, posterior, gamma, mixture, distribution, measure, prior, nonparametric, gaussian, drawn, parameter, university, dirichlet, probability, detect, journal, positive, three, residual, panel, figure, allow] [set, approximation, batch, note] [spike, neuron, waveform, overlapping, single, channel, intracellular, shape, spiking, recording, neural, noise, false, activity, electrode, voltage, sorting, marked, threshold, extracellular, earlier, multichannel]
Phase Retrieval using Alternating Minimization
Praneeth Netrapalli, Prateek Jain, Sujay Sanghavi


[algorithm, step, sampled, natural, setting, variant] [vector, work, based, approach, applied, second, output, image, large, amount, better, good, performance, multiple] [initialization, underlying, correctness] [phase, matrix, alternating, random, phaselift, altminphase, measurement, convex, dist, recovery, standard, number, magnitude, solving, arxiv, key, preprint, sparse, method, convergence, error, solve, phasecut, diagonal, recover, analysis, success, fourier, denotes, recovering, rank, empirically, trace, whp] [log, gaussian, sample, figure, independent, computational, probability, normal] [problem, complexity, minimization, lemma, theorem, theoretical, case, sdp, note, solution, linear, proof, empirical, support, constant, paper, main, consider] [retrieval, complex, signal, greater, imaging]
Translating Embeddings for Modeling Multi-relational Data
Antoine Bordes, Nicolas Usunier, Alberto Garcia-Duran, Jason Weston, Oksana Yakhnenko


[best, conference, natural, movie] [transe, training, learning, embeddings, relationship, embedding, head, test, entity, performance, tail, unstructured, corrupted, table, validation, raw, work, large, trained, award, lfm, knowledge, freebase, prediction, evaluation, vector, appear, learn, good, translation, predicted, well, rescal, scale, mtv, international, word, better, achieve, predict, approach, wordnet, dissimilarity] [link, relational, represent, corresponds, split] [number, rank, matrix, stochastic, method, gradient, factorization, tensor, optimization] [data, model, modeling, latent, selected, processing] [set, dimension, procedure, machine, ranking, lower] [neural]
Generalizing Analytic Shrinkage for Arbitrary Covariance Structures
Daniel Bartz, Klaus-Robert Müller


[assumption, lim, behaviour, optimal, sequence, stock] [large, better, automatic, usps, average, table, performance, target, outperforms, tested, propose] [real, growth, structure, factor] [covariance, standard, matrix, orthogonal, error, largest, eigenvalue, high, number, analysis, dimensionality, regularization, random, variance, additional, small, letter, isolet, convex, low, rate, fourth, order] [data, model, sample, figure, lda, limit, journal, estimation, dependency] [shrinkage, analytic, complement, consistency, dispersion, portfolio, theorem, estimator, statistical, spoken, squared, exists, qda, derive, linear, procedure, prove, theoretical, min, selection, result, arg, eigendirections, lower, constant, tij, cost, machine, eighth, paper] [normalized, combination, left]
Flexible sampling of discrete data correlations without the marginal distributions
Alfredo Kalaitzis, Ricardo Silva


[algorithm, step, space, deterministic] [learning, approach, level, applied] [factor, compute, variable, structure, pair, smallest, graphical] [matrix, rank, number, standard, constrained, application, method, true, small, dimensionality, mode] [gaussian, sampling, copula, hmc, sample, time, data, discrete, travel, likelihood, distribution, model, latent, marginal, extended, multivariate, figure, truncated, hamiltonian, hough, mcmc, joint, bayesian, monte, chain, mixing, gibbs, univariate, carlo, cdf, conditioned, sampler, posterior, marginals, describe, markov, sinusoid, component, parameter, exact, observed, boundary] [linear, machine, framework, general, case, provide, function, hard, idea, structured, point] [correlation, blue, velocity, position, collision, red, constraint, envelope]
Adaptive Anonymity via $b$-Matching
Krzysztof M. Choromanski, Tony Jebara, Kui Tang


[adversary, assume, utility, algorithm, adaptive, succeeds, released, compatible, usernames, ypublic, goal, explored] [level, achieve, handle, output, people, vector, weight, type, better] [graph, compatibility, degree, associated, person, edge, vertex, protection, color, node, social, forest] [anonymity, matching, symmetric, attack, number, random, sustained, entry, matrix, key, minimum, obfuscated, resilience, additional, gij, existence, permutation, asymmetric, method, article] [data, probability, advance, respect, family, perfect, figure, university, allow] [bipartite, privacy, set, case, database, problem, lemma, theorem, consider, relaxation, denote, approximation, provide] [user, single, desired, uniform]
Convex Tensor Decomposition via Structured Schatten Norm Regularization
Ryota Tomioka, Taiji Suzuki


[assume, sum, optimal, bound, appendix, deterministic, choice, linearly] [approach, performance, better, predicted, predicts] [group, factor, scaling, singleton, perform] [tensor, overlapped, schatten, rank, decomposition, norm, regularization, convex, mode, unknown, error, true, duality, minimum, matrix, performs, noisy, analysis, theoretically, empirically, optimization, dual, statement, proposed, sparsity, unfolding, siam, mink, completion, basic, tucker, spectral, call] [latent, figure, panel, data, multilinear, series, observation, mixture, gaussian, component, inverse] [complexity, structured, constant, note, inequality, technical, problem, mse, minimization, squared, lemma, theorem, consistency, minimal, solution, theoretical, lasso, loss] [noise, presented, simultaneously, study, depends]
Unsupervised Structure Learning of Stochastic And-Or Grammars
Kewei Tu, Maria Pavlovskaia, Song-Chun Zhu


[algorithm, gain, natural, previous, initial, action, greedy] [learning, training, approach, image, learned, dataset, language, second, unsupervised, learn, context, video, multiple, based, table, type, better, well, applied, large, represented, representing, target, performance, computer] [grammar, fragment, nonterminal, search, event, atomic, node, terminal, structure, size, existing, third, relation, represents, product, viterbi, represent, perplexity, construct] [stochastic, method, number, order, proposed, iteration] [data, figure, probability, posterior, likelihood, sample, prior, model, testing, restrict, beam] [set, example, agnostic] [pattern, temporal]
Modeling Overlapping Communities with Node Popularities
Prem Gopalan, Chong Wang, David Blei


[algorithm, bound, natural, step, notice] [learning, similarity, better, prediction, accuracy, test, large] [node, community, network, global, mmsb, inference, popularity, link, local, social, pair, strength, elbo, degree, develop, real, update, rab, collaboration, political, blockmodel, netscience, structure, optimize, american, size, probabilistic] [stochastic, amp, gradient, number, random, noisy, precision, small, coordinate, ascent, subsampling, iteration, objective] [variational, model, data, posterior, latent, predictive, respect, figure, estimate, draw, probability, observed, include, sample, journal, conditional, distribution, parameter, likelihood] [set, lower, indicator, unbiased, statistical] [single, capture, overlapping, explain, recall]
Learning from Limited Demonstrations
Beomjoon Kim, Amir massoud Farahmand, Joelle Pineau, Doina Precup


[policy, expert, apid, optimal, bellman, lfd, space, algorithm, lspi, demonstration, reinforcement, bound, upper, drl, assumption, state, api, goal, reward, initial, choice, agent, brake, dagger, mdp, current, outperformed, assume, focus, abundant, school, behaviour, canada, acceleration, mcgill, availability, iterationsaverage] [learning, supervised, well, performance, evaluate, large, compare, good, evaluation, approach, average, task, based, computer, margin] [real, science] [error, iteration, optimization, method, number, regularization, random, argmin, standard, distance] [data, approximate, figure, probability, university, estimate, three] [function, set, problem, complexity, turn, max, note, empirical, provide, machine, framework, case, regularizer, consider, linear] [robot, avoid, reach, control, velocity]
On the Complexity and Approximation of Binary Evidence in Lifted Inference
Guy van den Broeck, Adnan Darwiche


[conference, bounded, aaai] [learning, international, vector, represented, based, knowledge, technique, large, work] [evidence, boolean, lifted, inference, binary, unary, size, conditioning, bmf, relational, probabilistic, ground, mln, den, van, relation, mcmclifted, lmcmc, perform, guy, kld, web, logical, broeck, intelligence, predicate, webkb, pauli, davis, mining, adding, graphical, compute, represent, divergenceiterationground, lifting] [rank, matrix, factorization, computing, number, reduction, error] [approximate, exact, data, model, figure, mcmc, distribution, probability, marginal, investigate, posterior, full] [complexity, approximation, example, formula, set, condition, statistical, polynomial, theorem, result, problem, machine, consider, outer, main] [corresponding, circuit]
Non-strongly-convex smooth stochastic approximation with convergence rate O(1/n)
Francis Bach, Eric Moulines


[algorithm, achieves, best, step, optimal, assumption, bound, assume] [test, datasets, performance, learning, training, average, large, achieve, better] [proportional, global, synthetic, size, smaller] [stochastic, averaged, rate, sag, convergence, gradient, newton, sgd, analysis, adagrad, square, quadratic, order, typically, sido, method, alpha, descent, convex, decaying, covertype, denotes, covariance, standard] [log, figure, data, markov, news, chain, expectation, independent, stationary, distribution, journal] [logistic, consider, approximation, constant, loss, point, regression, theoretical, support, note, function, linear, machine, close, convexity, converges, theorem, provide, complexity, effective, problem, denote, condition, excess, technical, proof, smoothness] [strong, control, novel, left, normalized]
Robust Multimodal Graph Matching: Sparse Coding Meets Graph Matching
Marcelo Fiori, Pablo Sprechmann, Joshua T. Vogelstein, Pablo Muse, Guillermo Sapiro


[algorithm, black, sum, linearized] [original, consists, performance, approach, compared, second, common, measured] [graph, multimodal, inference, group, adjacency, graphical, weighted, real, permuted, synthetic, faq, path, network, bter, elegans, vogelstein, chemical, publicly, underlying] [matching, matrix, permutation, proposed, collaborative, method, covariance, optimization, convex, det, error, solving, random, sparse, naturally, application, glag, number, doubly, analysis, solve] [data, log, inverse, figure, estimation, model, distribution, processing, university, joint] [problem, min, formulation, lasso, version, set, result, arg, case, solved, active, minimizing] [noise, fmri, presented, pattern, connection, experimental, described, brain, term, neural]
Transportability from Multiple Environments with Limited Experiments
Elias Bareinboim, Sanghack Lee, Vasant Honavar, Judea Pearl


[transportability, trmz, limited, conference, calculus, algorithm, diagram, observational, assume, estimand, menlo, interventional, aaai, bareinboim, goal, return, called, lee, previous, formal, special, age] [target, domain, knowledge, multiple, learning, work, test, based] [causal, source, structural, represents, collection, computable, induced, relation, intelligence, heterogeneous, edge, powerful, randomized, rewrite, subgraph, local, graphical, mit, variable, graph] [conducted, transport, call, computing, analysis, order, variance] [data, expression, estimate, national, university, model] [selection, set, theorem, consider, statistical, problem, condition, formula, example, note, procedure, general, provide, machine, corollary, paper, implies] [experimental, relative, rule, desired, infer]
Supervised Sparse Analysis and Synthesis Operators
Pablo Sprechmann, Roee Litman, Tal Ben Yakar, Alexander M. Bronstein, Guillermo Sapiro


[setting, algorithm, total] [learning, supervised, image, vector, pursuit, dictionary, reconstruction, music, training, transcription, learned, representation, convolutional, piano, well, denoising, performance, approach, achieved, better, metric, input, resolution, discriminative, feature, unsupervised] [network] [analysis, sparse, synthesis, fast, matrix, proposed, optimization, operator, admm, number, gradient, standard, iterative, polyphonic, regularization, solving, projection] [model, exact, data, form, inverse, respect, figure, university, computational, approximate, gaussian, extended] [problem, solution, loss, approximation, set, case, function, linear, consider, note, proximal, proposition, solved, version, map, framework, min] [bilevel, ieee, signal, neural, sensing, minimizer, depends, described]
Generalized Denoising Auto-Encoders as Generative Models
Yoshua Bengio, Li Yao, Guillaume Alain, Pascal Vincent


[algorithm, bound, apply, walk] [training, denoising, learning, reconstruction, dae, trained, input, score, generative, learned, amount, work, based, deep, rbm, feature] [local, associated] [true, regularized, implicitly, error, operator, small, matrix, random] [distribution, corruption, chain, markov, data, sampling, walkback, asymptotic, process, model, conditional, gaussian, manifold, generated, figure, sample, spurious, joint, contractive, limit, continuous, estimation, generating, marginal, estimate, likelihood, log, transition, probability, observed] [estimator, function, energy, consistent, density, converges, procedure, example, note, generalized, result, consider, criterion, minimizing, case, theorem, maximum, machine, idea, set] [noise, neural, estimated, interpretation, simple, capture]
Documents as multiple overlapping windows into grids of counts
Alessandro Perina, Nebojsa Jojic, Manuele Bicego, Andrzej Truski


[conference, algorithm, share, choose, space, best, modeled] [window, word, generative, text, based, bag, learning, learned, represented, embedding, dataset, multiple, evaluated, semantic, large, computer, usage, similarity, international] [size, characterized, generate, represents] [number, small, pick, standard, basic, analysis] [grid, counting, model, componential, distribution, lda, topic, latent, kernel, dirichlet, generated, ccgs, document, ccg, log, data, probability, mixture, modeling, component, pizza, equal, considered, multinomial, prior, figure, sample, sam, admixture, three, journal, allocation, spherical] [effective, linear, machine, set] [location, ieee, position, capacity, pattern, moving, area, highly, simple, described, thought]
Solving the multi-way matching problem by permutation synchronization
Deepti Pachauri, Risi Kondor, Vikas Singh


[algorithm, satisfy, natural, long, space] [image, large, computer, multiple, baseline, object, representation, work, supervision, feature, accuracy, based, approach, recognition, dataset, building, correspondence, common] [matchings, graph, assignment, structure, degree, ground, hand, mutation, house] [permutation, matching, matrix, synchronization, eigenvectors, pairwise, method, error, landmark, random, symmetric, orthogonal, leading, recover, tji, solving, eigenvalue, number, stereo, eigenvector, rank, array, distance, fig, column, small] [figure, form, probability, distribution] [problem, set, linear, note, proposition, solved, consistent, relaxation, arg, consistency, consider, condition, cost, paper, combinatorial] [correct, noise, normalized, shape, stable, rotation]
Robust Bloom Filters for Large MultiLabel Classification Tasks
Moustapha M. Cisse, Nicolas Usunier, Thierry Artières, Patrick Gallinari


[algorithm, best] [predicted, approach, large, training, prediction, datasets, test, learning, compared, frequent, based, improve, scale, appear, predict, performed] [bloom, size, representative, cluster, bit, mlc, binary, exclusive, mutually, incorrectly, multilabel, clustering, smaller, inference, plst, disjunctive, uhl] [number, hash, method, standard, robust, hamming, random, scheme, reduction, small, compressed, rate, error] [figure, positive, computational, probability, data, distribution, time] [label, set, loss, complexity, design, ratio, example, subset, cardinality, main, denote, theoretical] [encoding, decoding, code, false, single, strong, encode]
Visual Concept Learning: Combining Machine Vision and Bayesian Generalization on Concept Hierarchies
Yangqing Jia, Joshua T. Abbott, Joseph Austerweil, Thomas Griffiths, Trevor Darrell


[sampled, previous, total, space] [concept, learning, visual, generalization, vision, image, word, level, performance, perceptual, object, imagenet, dataset, task, work, learn, training, test, large, people, conventional, challenge, truth, score, baseline, hierarchy, instance, confusion, ilsvrc, validation, negative, compare, randomly, average, approach, better, tenenbaum, constructing] [node, leaf, existing, size, ground, subtree, science] [number, proposed, method, basic, precision, matrix, gradient] [bayesian, figure, probability, three, model, positive, prior, data, describe] [set, query, machine, hypothesis, example, note, problem, provide] [human, system, novel, cognitive, psychological]
Dirty Statistical Models
Eunho Yang, Pradeep Ravikumar


[sum, bound, assumption, previous] [multiple, work, class, well, instance, applied] [structural, structure, size, pair] [regularization, matrix, subspace, error, norm, sparse, analysis, principal, covariance, high, typically, key, robust, number, convex, incoherence, convergence, row, frobenius, exactly, elementwise, regularized, random, constrained, solve, decomposition] [parameter, model, log, component, respect, observation, probability, hybrid, restricted, sample, estimation, gaussian] [theorem, function, statistical, linear, suppose, consider, regression, loss, set, problem, condition, corollary, provide, superposition, decomposable, structured, general, proposition, framework, note, varied, max, derive, min, solution] [corresponding, strong, term, noise]
Locally Adaptive Bayesian Multivariate Time Series
Daniele Durante, Bruno Scarpa, David Dunson


[online, algorithm, state, adaptive, space, step, stock] [approach, dictionary, performance, second, dataset, based, well, large, vector, prediction] [factor, local, structure, update, computation] [covariance, updating, true, stochastic, matrix, proposed, order, analysis] [posterior, time, multivariate, model, lbcr, series, locally, bayesian, conditional, latent, figure, simulation, crisis, process, gaussian, continuous, gibbs, smoothing, distribution, prior, nonparametric, log, bcr, usa, data, nasdaq, dunson, journal, dependence, predictive, national, rapid, nsi, wishart, observed] [smoothness, regression, set, formulation, consider, shrinkage, selection] [varying, dynamic, simple, sharp, simulated, ahead]
Optimistic Concurrency Control for Distributed Unsupervised Learning
Xinghao Pan, Joseph E. Gonzalez, Stefanie Jegelka, Tamara Broderick, Michael Jordan


[algorithm, optimistic, conference, online, apply, state, inventory, execution] [learning, feature, international, work, large, preserving, unsupervised, approach, validation, second, based] [occ, distributed, cluster, serial, concurrency, ofl, accepted, scalability, parallel, facility, serially, clustering, global, runtime, scaling, size, serialization, existing, michael, product, operation, scalable, acm, communication, parallelism, supported, validated, serializable] [number, proposed, epoch, analysis, computing, additional] [data, processing, figure, process, generated, model, zik, probability, parameter, nonparametric, dirichlet, bayesian, perfect] [machine, set, approximation, point, cost, theoretical, easily, theorem, proof, database] [control, location, neural, pattern, strong, correct]
Auxiliary-variable Exact Hamiltonian Monte Carlo Samplers for Binary Distributions
Ari Pakman, Liam Paninski


[algorithm, initial, space, state, jump] [work, approach, momentum] [binary, variable, move, allows] [coordinate, method, faster, order, proposed, quadratic, rate, random, arxiv, preprint, sparse] [hmc, model, gaussian, sampler, distribution, sample, time, metropolis, chain, markov, continuous, gibbs, monte, hamiltonian, particle, wall, sampling, truncated, ising, exact, carlo, figure, bayesian, prior, illustrate, boundary, piecewise, travel, computational, posterior, log, inverse, hill, successive, exponential, slab, journal, continues, orthant, discrete, multivariate] [regression, linear, consider, energy, case, statistical, interested, cost, clearly, problem, lower, function, solution, probit] [motion, hit, trajectory, leave, periodic]
Adaptive Step-Size for Policy Gradient Methods
Matteo Pirotta, Marcello Restelli, Luca Bascetta


[policy, bound, step, state, itmax, reinforce, bounded, difference, upper, action, pgt, expected, optimal, reward, advantage, natural, assumption, future, discounted, maximized, converge, supremum, episodic, scenario, deterministic, choice] [performance, learning, large, vector, table, approach] [size, starting, allows, pair, improvement, search, larger] [gradient, variance, number, convergence, error, small, basis, direction, uniformly, order, divergence, proposed, lead, quadratic, stochastic, norm, updated] [gaussian, estimate, distribution, stationary, positive, exact, parameter, probability, model, processing, generated] [lower, approximation, theorem, corollary, lemma, problem, function, consider, derive, provide, approximated, polynomial, case, guaranteed, estimator] [estimated, neural, control, speed]
Probabilistic Principal Geodesic Analysis
Miaomiao Zhang, P.T. Fletcher


[space, exp, initial, algorithm] [vector, unit, truth, work, computer] [probabilistic, variable, ground, factor, parallel] [principal, analysis, gradient, sphere, euclidean, random, symmetric, pca, standard, matrix, denotes, proposed, variation] [riemannian, data, geodesic, manifold, model, latent, exponential, estimation, component, log, pga, tangent, derivative, monte, carlo, jacobi, corpus, ppga, distribution, callosum, normal, figure, parameter, expectation, gaussian, sample, normalizing, bayesian, processing, include, hamiltonian, nonlinear, covariant, base, likelihood, geometry, kernel, respect, hmc] [map, point, linear, set, maximization, note, procedure, function, maximum, formulation, integral, constant, density, denote, kendall, geometric] [shape, complex, estimated, rotation, term, velocity, neural, simulated]
Mixed Optimization for Smooth Functions
Mehrdad Mahdavi, Lijun Zhang, Rong Jin


[algorithm, optimal, goal, assume, assumption, bound, making, setting, online, sequence, fact, sampled, step, deterministic, regret] [domain, improve, based, achieve, training, learning, average, approach, improved, amount, introduced, better] [mixed, size, compute, distributed, allows] [stochastic, gradient, optimization, convergence, rate, oracle, convex, number, objective, method, descent, small, kth, epoch, proposed, mixedgrad, optimizing, regularization, order, key, norm, sgd, variance, stochastically, accelerated, rgit, computing, call] [full, probability, parameter, journal, sample] [function, smooth, solution, loss, theorem, proof, smoothness, access, lemma, machine, main, problem, min, note, result, constant, cost, minimizes, set, consider, example, lipschitz] [strong]
Projecting Ising Model Parameters for Fast Mixing
Justin Domke, Xianghang Liu


[algorithm, total, bound, projected, converge] [original, based, accuracy, work, well] [strength, inference, edge, graph, tree, projecting, loopy, mixed, size] [divergence, euclidean, fast, norm, spectral, random, projection, matrix, gradient, number, singular, distance, computing, stochastic, method, descent] [gibbs, mixing, sampling, piecewise, ising, interaction, distribution, marginals, approximate, exact, time, markov, tractable, model, chain, exponential, parameter, lbp, respect, dependency, family, sample, estimate, rapid, stationary, probability, variational, intractable, accurate, univariate] [set, trw, consider, case, theorem, result, simply, general, density, guaranteed, approximation, guarantee, arg] [attractive, estimated]
Which Space Partitioning Tree to Use for Search?
Parikshit Ram, Alexander Gray


[algorithm, bounded, best, conference, space, bound, chosen, worst, assumption, expected] [performance, margin, large, better, vector, based, dataset, improved, task, learning] [search, tree, quantization, depth, partition, neighbor, candidate, node, split, size, alternate, improvement, hyperplane, binary, nearest, smaller, traversal, clustering, explicitly, disjoint, corresponds, defeatist, balance] [error, distance, expansion, covariance, rank, hash, number, convex, hashing, dimensionality, imply, rate, random, principal, high, matrix] [data, partitioning, axis] [implies, set, theoretical, query, condition, result, theorem, constant, guarantee, consider, case, corollary] [region, balanced, ieee, depends]
Conditional Random Fields via Univariate Exponential Families
Eunho Yang, Pradeep Ravikumar, Genevera I. Allen, Zhandong Liu


[] [learning, class, vector, learn, well, feature, based, common, computer, compared] [graph, graphical, structure, neighborhood, compatibility, size, node, undirected, corresponds, allows] [random, method, recovery, analysis, rate, arxiv, key, preprint, sparsity, sparse, regularization] [conditional, crf, distribution, exponential, gaussian, model, crfs, covariates, family, univariate, joint, data, form, poisson, ising, conditioned, multivariate, positive, observation, gene, parameter, three, figure, estimation, glioblastoma, rest, including] [set, theorem, covariate, consider, function, statistical, selection, suppose, provide, linear, distributional, general, framework, note, mrf, question, depend, consistent] [response, false, novel, corresponding]
Provable Subspace Clustering: When LRR meets SSC
Yu-Xiang Wang, Huan Xu, Chenlei Leng


[conference, deterministic, space, algorithm, assume, difference, absolute, assumption] [computer, international, representation, segmentation, unit, better, vision, learning, vector, large] [clustering, graph, disjoint, represents, connected] [subspace, matrix, spectral, ssc, sep, sparse, lrssc, singular, dual, norm, lrr, random, gap, analysis, incoherence, noisy, skewed, numerical, convex, remark, nuclear, inradius, block, robust, relviolation, sparsity, imply, sphere, diagonal, exactly] [data, log, figure, model, sample, probability, generated, distribution, parameter, illustrate] [solution, problem, condition, lemma, min, machine, theorem, example, set, constant, ratio, theoretical, selection, guarantee, studied, proof, main, linear, version] [connectivity, pattern, ieee, range, normalized, greater, correct, motion]
Stochastic blockmodel approximation of a graphon: Theory and consistent estimation
Edoardo M. Airoldi, Thiago B. Costa, Stanley H. Chan


[algorithm, sequence, bound, step, expected, goal, optimal] [experiment, better, evaluate, large, second, performance, based, learning, similarity] [graph, missing, network, blockmodel, assign, size, vertex, mae, relational] [graphon, proposed, stochastic, number, usvt, bbi, random, block, dij, graphons, bdij, error, growing, matrix, bbk, bbj, row, method, distance, analysis, small] [sba, estimate, estimation, observed, figure, journal, exchangeable, probability, generated, independent, latent, data, parameter, nonparametric, ujv, continuous, piecewise, bayesian, model] [theorem, lemma, function, estimating, consider, consistency, estimator, procedure, idea, approximation, machine, theory, constant, empirical, consistent, statistical] [estimated]
Generalized Random Utility Models with Multiple Types
Hossein Azari Soufiani, Hansheng Diao, Zhenyu Lai, David C. Parkes


[agent, utility, algorithm, chosen, partial, choice, jump] [type, multiple, work, performance, based, propose, approach] [clustering, synthetic, real, inference, james, move, david] [number, random, matrix, method, order, true, proposed] [data, model, alternative, prior, posterior, journal, likelihood, log, interaction, grums, economics, figure, university, steven, estimate, mixture, observed, equal, estimation, economic, daniel, reversible, sample, consumer, parameter, probability, harvard, logit] [set, function, case, ranking, sushi, generalized, concave, theorem, grum, consider, proof, establish, provide, rjmcmc, suppose, theory, azari, select, theoretical, statistical, result] [involving, preference, unobserved, study, consisting, intrinsic]
Bayesian Hierarchical Community Discovery
Charles Blundell, Yee Whye Teh


[greedy, algorithm, worst, current] [hierarchical, work, hierarchy, better, large, top, propose, score, approach, learning, scale] [tree, bhcd, community, structure, network, social, adjacency, relational, inference, whilst, irm, binary, edge, forest, partition, merge, cluster, discovering, blockmodel, monastery, pair, merges, computed, formed, root, factor, local, corresponds, rose] [number, stochastic, sparse, block, iteration, matrix, faster, method] [model, bayesian, figure, likelihood, data, time, probability, prior, predictive, marginal, nonparametric, log, latent, computational, full, observed, describe] [complexity, yield, case, procedure, consistent] [potential, inferring, single, connectivity, neural]
Optimistic policy iteration and natural actor-critic: A unifying view and a non-optimality result
Paul Wagner


[policy, optimistic, natural, greedy, step, optimal, oscillation, reinforcement, methodology, vanishingly, established, special, interpolation, algorithm, action, origin, limiting, converge, markovian, conference, deterministic, current, programming, state, bounded, decision, focus, difference, permit] [learning, evaluation, temperature, class, vector, approach, well, international] [size, improvement, factor, understanding] [convergence, gradient, stochastic, iteration, uniformly, distance, small, practical, direction, lead, sustained] [parameter, approximate, gibbs, form, figure, process, extended, journal, monte, markov, carlo] [function, case, optimum, point, theorem, result, problem, effective, close, optimality, consider, machine, subset, general, version, map] [dynamic, term, neural, assuming, behavior]
Predictive PAC Learning and Process Decompositions
Cosma Shalizi, Aryeh Kontorovitch


[algorithm, bound] [image, based, large, work, better] [local, degree, structure, existing, associated] [sparse, additional, column, exactly, unknown, analysis, proposed, variance, comparison, operator, true, norm, homotopy, optimization, sparsity, projection, typically] [log, model, kernel, estimation, figure, form, process, element, bayesian, likelihood] [function, penalty, cost, concave, min, note, selection, simply, general, theoretical, result, regression, paper] [blur, deblurring, blind, camera, blurry, sharp, noise, estimated, uniform, shake, motion, relative, greater, deconvolution, concavity, normalization, single, simple, continuation, meaning, gupta, reparameterization, hirsch, blurring, undesirable, viewed, harmeling, joshi, corresponding, whyte]
From Bandits to Experts: A Tale of Domination and Independence
Noga Alon, Nicolò Cesa-Bianchi, Claudio Gentile, Yishay Mansour


[regret, bound, action, observability, setting, algorithm, dominating, bandit, upper, uninformed, player, expert, reveals, adversarial, called, partial, optimal, variant, logarithmic, playing, observe, learner, greedy, online, quantity, elp, sequence, step, current, observes] [prediction, learning, knowledge] [graph, directed, undirected, informed, acyclic, size, edge, compute, social] [number, symmetric, analysis, standard, order, random, key, minimum, grant] [observation, independence, time, independent, model, prior, distribution, generated, probability] [set, corollary, loss, lower, case, max, consider, arc, result, problem, cover, subset, linear, note, theorem, general, program, studied, israel, example] [system, simple, corresponding]
Action is in the Eye of the Beholder: Eye-gaze Driven Model for Spatio-Temporal Action Localization
Nataliya Shapovalova, Michalis Raptis, Leonid Sigal, Greg Mori


[action, algorithm, space, actor, focus] [video, localization, context, bounding, learning, bow, recognition, discriminative, box, training, feature, table, prediction, visual, representation, detection, average, output, multiple, based, score, approach, scoring, truth, object, weak, localize, performance, test, svm, amount, propose, evaluation, incorporating, ability, well, evaluate, excluding, dataset] [search, local, global, allows, person, path, ground, inference] [proposed] [latent, model, data, background, form] [function, structured, density, loss, map, linear, smooth, formulation, set, case] [gaze, saliency, region, human, eye, salient, inferred, temporal, frame, response, capture, interest, spatial, tran, perceptually, motion, potential, overlap]
Efficient Optimization for Sparse Gaussian Process Regression
Yanshuai Cao, Marcus A. Brubaker, David Fleet, Aaron Hertzmann


[algorithm, free, greedy, space, total] [training, input, test, performance, learning, vector, better, good, dataset, hyperparameter, evaluating, negative, large, based, compare, hog, domain, evaluation] [update, candidate, optimizes] [objective, optimization, sparse, factorization, random, matrix, number, covariance, swap, method, variance, small, titsias, rank] [inducing, discrete, continuous, time, gaussian, kernel, variational, process, log, approximate, spgp, cholesky, exact, likelihood, smse, data, full, hyperparameters, ivm, residual, cholqr, figure, csi, predictive, snlp, form, swapping, marginal, element] [set, point, select, approximation, selection, function, regression, cost, energy, lower, machine, complexity, subset] [change, single]
Aggregating Optimistic Planning Trees for Solving Markov Decision Processes
Gunnar Kedenburg, Raphael Fonteneau, Remi Munos


[optimal, algorithm, space, simplex, fact, sum] [metric, well, vector, mnist, large, negative, original, better, propose, image] [scaling, computed, network, computation, perform, compute] [transport, distance, matrix, pij, sinkhorn, qjk, number, divergence, classic, emd, computing, regularization, gpu, earth, euclidean, entropic, iteration, mij, regularized, rct, convergence, random, small, cpu, argmin] [entropy, joint, figure, kernel, probability, computational, log, independence, positive, equation, form, mutual, time, journal, family] [problem, set, point, cost, proof, lemma, theory, consider, linear, theorem, dimension, prove, provide, inequality, subset, extreme] [single, varying, property, speed]
Learning the Local Statistics of Optical Flow
Dan Rosenbaum, Daniel Zoran, Yair Weiss


[natural, algorithm, availability, black] [computer, learned, image, better, truth, vision, learn, patch, large, compare, training, performance, test, hidden, international, amount, work, outperforms, learning, dataset, vector, second] [local, ground, inference, conditioning, michael, improvement] [matrix, robust, eigenvectors, standard, subspace, covariance, comparing, success, row, small, expect, leading, optimization] [optical, intensity, model, gaussian, prior, likelihood, log, data, figure, estimation, kitti, component, held, mixture, boundary, sintel, generated, allow, multidimensional, transition, modeling, conditioned, conditional, restoration, markov] [gmm, smoothness, set, suggested, linear, function, constant] [motion, horn, term, noise, pattern]
Estimation Bias in Multi-Armed Bandit Algorithms for Search Advertising
Min Xu, Tao Qin, Tie-Yan Liu


[revenue, adaptive, mab, algorithm, advertiser, utility, incentive, ucb, expected, ctr, auction, payment, history, best, engine, debiasing, bound, assume, elm, bid, unif, impression, adpt, regret, satisfy, fact, gain, charge, advertising, price, slot, bandit, game, display, setting, maxk, sponsored] [learning, second, quality, vector, negative, work, propose, large] [search, larger, stage, smaller, compatibility, concentration] [number, largest, low, analysis, order, high, random] [log, sample, probability, estimation, figure, estimate, measure] [selection, bias, min, theorem, loss, suppose, estimator, analyze, select, arg, case, guarantee, corollary, lower, condition, example, unbiased, theoretic] [click, uniform, simple, increase, described]
Robust learning of low-dimensional dynamics from large neural ensembles
David Pfau, Eftychios A. Pnevmatikakis, Liam Paninski


[natural, space, history, state, optimal] [directly, learning, large, top, common, vector, prediction, direct, approach] [real, local, global] [norm, nuclear, matrix, subspace, true, principal, explained, recovered, method, divergence, number, bregman, variance, convex, recover, analysis, rate, dimensionality, spectral, singular, dimensional, rank, spanned, sparse, low, ssid] [data, model, log, latent, likelihood, component, journal, time, held, gaussian, transition, process, poisson, observed, processing, include, figure, parameter] [linear, case, minimization, penalty, statistical, generalized, point, dimension, function, paper] [neural, spike, dynamical, synaptic, system, connectivity, population, term, angle, motor, stable, presented, strong, account]
On model selection consistency of penalized M-estimators: a geometric theory
Jason Lee, Yuekai Sun, Jonathan E. Taylor


[assumption, sum, state, optimal, assume] [work, learning, large, vector] [group, graphical, neighborhood, size, van, supported, science] [norm, convex, subspace, minimize, unknown, sparsity, covariance, sparse, supplementary, error, dual, analysis, high] [model, exponential, likelihood, estimation, log, sample, parameter, probability, family, inducing, estimate] [consistency, lasso, selection, decomposable, geometrically, generalized, penalty, support, penalized, function, notion, result, maximum, main, framework, irrepresentable, hdt, condition, solution, linear, expressed, negahban, statistical, set, theorem, estimator, general, inf, select, derive, suppose, problem, establishing, consistent, constant, regression, loss, engineering, consider, generalize, corollary, commonly, convexity, establish] [strong, signal, enforce]
New Subsampling Algorithms for Fast Least Squares Regression
Paramveer Dhillon, Yichao Lu, Dean P. Foster, Lyle Ungar


[algorithm, setting, bound, assume, failure, best, long] [vector, datasets, work, amount, well, second, approach, large, compare, better, performance, accuracy] [randomized, synthetic, real, stage, subsample] [error, random, matrix, subsampling, hadamard, uluru, covariance, nsubs, transform, covs, preconditioning, subsampled, analysis, ols, number, fast, yrem, leverage, high, supplementary, corr, faster, uniformly, small, srht, fraction, remark, fourier, noisy, performs, standard, wcorrect] [data, three, probability, estimate, time, independent, full, estimation, parameter, figure] [design, theoretical, theorem, paper, linear, case, provide, estimator, set, problem, loss, estimating, squared, regression, consider, result] [described, noise, depends, varying, material, term]
Dropout Training as Adaptive Regularization
Stefan Wager, Sida Wang, Percy Liang


[adaptive, natural, online, conference, apply, linearized] [dropout, training, feature, noising, unlabeled, learning, rare, accuracy, vector, labeled, table, def, test, dataset, fisher, discriminative, level, large, imdb, based, international, noised, better, performance, well, applied, train, sida, generative, text, christopher] [size, applying] [quadratic, regularization, adagrad, matrix, sgd, descent, stochastic, random, small, method, solving, ridge] [data, form, additive, estimate, model, figure, computational, parameter, processing, simulation, gaussian, document, independent, journal] [penalty, logistic, regression, linear, approximation, regularizer, equivalent, loss, machine, design, problem, case, depend, active, empirical, function, consider] [noise, neural, connection]
Using multiple samples to learn mixture models
Jason Lee, Ran Gilad-Bachrach, Rich Caruana


[algorithm, space, assume, maximal, return, assumption, goal, maintain, projected] [learning, multiple, work, applied, large, second, appear, experiment, hidden, computer, compare] [msp, dsc, tree, clustering, underlying, disjoint, cluster, symposium, annual, leaf, associated, vempala] [projection, variance, dimensional, low, random, project, spanned, distance, require, method, high, analysis, error, spectral] [mixture, sample, data, gaussians, mixing, figure, model, probability, generated, three, log, assumes, measure, address, usa, topic] [problem, dimension, separation, case, result, set, exists, polynomial, theorem, santosh, complexity, main] [presented, single, blue, noise, ieee]
Annealing between distributions by averaging moments
Roger B. Grosse, Chris J. Maddison, Ruslan Salakhutdinov


[optimal, initial, linearly, natural] [target, hidden, rbms, learning, deep, visible, based, performance, boltzmann, rbm, mnist, table, training, average] [path, partition] [averaging, analysis, rate, high, order, number, proposed, variance] [log, intermediate, moment, sampling, exponential, perfect, intractable, schedule, family, figure, distribution, probability, model, tractable, sample, base, binned, gibbs, accurate, annealed, variational, annealing, restricted, spaced, alternative, monte, exact, markov, derived, estimate, transition, interpolate, piecewise] [geometric, linear, theorem, cost, function, set, result, estimating, machine, asymptotically, consider, statistical] [uniform, functional, viewed, indicate]
Learning Hidden Markov Models from Non-sequence Data via Tensor Decomposition
Tzu-Kuo Huang, Jeff Schneider


[state, initial, algorithm, formal, setting, appendix, sequence, expected, assume, sequential, easier] [learning, hidden, generative, multiple, based, input, large, consistently, learn] [size, structure, network] [tensor, matrix, decomposition, method, spectral, true, unknown, random, recover, order, column, small, projection, number, array, usual, arxiv] [data, markov, distribution, sample, transition, model, estimation, process, probability, estimate, latent, figure, issue, dirichlet, modeling, independent, draw, chain, topic, three, generated, observation, time, returned, discrete, observed, hmm] [set, empirical, theorem, theoretical, consistent, complexity, note, problem, proof, guarantee, result, general, procedure] [dynamic, biological, situation]
Accelerated Mini-Batch Stochastic Dual Coordinate Ascent
Shai Shalev-Shwartz, Tong Zhang


[algorithm, bound, deterministic, peter, assume, choose] [learning, better, training, table, vector, opening, nesterov, dataset, compare] [parallel, node, distributed, communication, size, martin, joseph, processed, message, runtime] [sdca, asdca, dual, iteration, gradient, agd, descent, stochastic, coordinate, accelerated, number, arxiv, preprint, computing, zhang, method, interpolates, convex, examplestest, rate, sgd, ascent, regularization, euclidean, analysis, mentioned, convergence, optimization, norm, dividing, solving, alekh, carlos] [time, data, processing] [cost, loss, machine, studied, consider, primal, shai, case, main, function, smooth, problem, version, complexity, paper, set, tong, solution, hold, example, max, subset] [single]
Faster Ridge Regression via the Subsampled Randomized Hadamard Transform
Yichao Lu, Paramveer Dhillon, Dean P. Foster, Lyle Ungar


[algorithm, space, setting, step] [large, vector, better, compare, well, feature, training, randomly, approach, dataset, setup, weight, top] [randomized, compute, median, computed, synthetic, real, size, parallel, structure, michael] [matrix, pca, psubs, true, ridge, random, dual, error, hadamard, transform, fast, computing, standard, srht, subsampled, rate, number, rank, analysis, high, principal, faster, svd, subsampling, huge, low, solutionpcarandomized, remark, ordinary, multiplication, solving] [computational, data, kernel, generated, figure, probability, power, time, equation, parameter, approximate] [risk, regression, cost, design, solution, pcr, approximation, linear, lemma, case, dimension, problem, provide, paper, statistical] [slow, mark]
Recurrent linear models of simultaneously-recorded neural populations
Marius Pachitariu, Biljana Petreska, Maneesh Sahani


[state, best, upper, deterministic] [output, better, performance, large, prediction, approach, well, hidden, directly, generative, common] [structure, factor, link, inference] [matrix, covariance, number, random, subspace, gradient, standard, pairwise, principal] [model, rlm, data, latent, lds, gaussian, distribution, poisson, ising, cglm, time, likelihood, equation, observation, describe, figure, process, replacing, estimate, three, computational, psth, alternative, conditional, instantaneous, cascaded, form, generated, plds, nonlinear] [linear, case, function, provide, note, consider] [neural, spike, dynamical, kalman, population, recurrent, simulated, recorded, temporal, estimated, noise, activity, generalisation, simple, neuronal, glm, motor]
Learning Adaptive Value of Information for Structured Prediction
David J. Weiss, Ben Taskar


[policy, state, reward, action, optimal, expected, current, space, adaptive, setting, total, reinforcement, choosing, future, add, sequence, goal, ocr] [feature, learning, learn, extraction, approach, prediction, vector, pose, work, training, extracted, accuracy, input, output, large, learned, based, test, level, video, achieve] [inference, compute, computed, computation, structure, runtime, graph] [error, method, additional, standard, rate, proposed, computing, typically] [model, time, accurate, approximate, figure, predictive, computational, markov, component, allow] [budget, structured, set, function, cost, note, simply, example, problem, effective, batch, predictor, consider, linear, selection, indicator] [control, frame, simple, trajectory, human, imitation, increase]
When are Overcomplete Topic Models Identifiable? Uniqueness of Tensor Tucker Decompositions with Structured Sparsity
Anima Anandkumar, Daniel Hsu, Majid Janzamin, Sham M. Kakade


[denoted, long, deterministic, bound, space] [hidden, class, award, vector, word, learning, generic, representation] [structure, graph, size, degree, variable, product, node] [overcomplete, random, order, matrix, matching, number, tensor, rank, persistence, sparsity, remark, expansion, uniqueness, tucker, krank, generically, recovery, incorporate, analysis, column, persistent, canonical, discussion, existence, imply, characterize] [topic, model, observed, moment, latent, admixture, perfect, log, full, figure, distribution, referred, probability, form, equation, university, aro, drawn] [condition, bipartite, theorem, establish, set, result, version, structured, general, note, consider, provide, case] [population, higher, single, novel, constraint, presence, pattern]
Estimating LASSO Risk and Noise Level
Mohsen Bayati, Murat A. Erdogdu, Andrea Montanari


[sequence, state, algorithm, lim] [level, vector, well] [evolution, applying, scaled, versus, represents, message, passing, color] [method, amp, error, standard, variance, matrix, regularization, true, random, order, iteration, compressed, remark, accurately, dimensional, basis, covariance, numerical, analysis] [gaussian, distribution, asymptotic, estimation, model, iid, estimate, inverse, figure, parameter, probability, generated, multivariate, independent, data, normal] [lasso, estimator, design, theorem, risk, general, unbiased, converging, denote, provide, squared, conjecture, corollary, consistent, statistical, problem, function, mse, linear, converges, result, empirical, weakly, estimating, main, note, consistency, annals, case, solution, suitable, formula, support, selection, msetrue, point] [noise, signal, estimated, ieee]
Relevance Topic Model for Unstructured Social Group Activity Recognition
Fang Zhao, Yongzhen Huang, Liang Wang, Tieniu Tan


[bound, action] [video, rtm, learning, relevance, replicated, training, vector, softmax, supervised, unstructured, dataset, semantic, accuracy, learn, bow, labeled, discriminative, work, better, hidden, boltzmann, wedding, recognition, svm, medlda, learned, jointly, visible, unit, class, automatic, table, gclass, discover, med, word] [social, undirected, multimodal, inference, group, graphical, binary, directed, web] [sparse, number, small, proposed, stochastic] [model, topic, variational, log, data, latent, distribution, bayesian, figure, likelihood, marginal, parameter, restricted, estimation, lda, gaussian, posterior, estimate, equation, three, conditional, wijvmi, contrastive] [function, linear, set, case, lower, maximum] [activity, complex, attribute, human, calculate]
Context-sensitive active sensing in humans
Sheeraz Ahmad, He Huang, Angela J. Yu


[belief, optimal, policy, action, switching, state, bound, strategy, decision, making, space, future, expected, assume, choice, choose, current, total, sequential, step] [target, visual, patch, better, task, performance, based, well, unit, compare, accuracy, context, work] [search, optimize] [number, proposed, block] [time, data, prior, model, probability, boundary, figure, bayesian, distribution, likelihood, sampling, processing] [cost, active, min, set, bias, function, case, approximation, theorem] [location, sensing, myopic, infomax, human, stopping, behavioral, control, behavior, dynamic, threshold, observer, presented, eye, brain, sensory, incurred, account, declare, experimental, simple, novel, trial, neural, spatial]
Confidence Intervals and Hypothesis Testing for High-Dimensional Statistical Models
Adel Javanmard, Andrea Montanari


[optimal, assume, bound, minimax, assumption, notice] [vector, test, approach, level, constructing, work] [size, construct, compatibility, binary, variable, van] [matrix, random, standard, covariance, sparse, regularized, error, order, method, dimensional, sparsity, convex, norm, high, denotes, number, variance] [testing, log, parameter, probability, gaussian, power, sample, model, estimate, asymptotic, distribution, estimation, normal, plot, exact, observed] [statistical, design, linear, hypothesis, estimator, regression, theorem, set, lasso, consider, procedure, condition, support, general, javanmard, selection, logistic, interval, consistent, provide, min, note, asymptotically, diabetes, letting, quantiles, classical, construction] [response, simple, normalized, earlier, ieee, term, corresponding, noise]
Factorized Asymptotic Bayesian Inference for Latent Feature Models
Kohei Hayashi, Ryohei Fujimaki


[algorithm, bound, satisfy, natural, acceleration, fact, elapsed, setting] [learning, feature, performance, large, achieved, representation, better, hidden, automatic] [inference, update, smaller, irrelevant, variable, real, unique] [method, analysis, matrix, regularization, number, supplementary, hessian, block, gradient, standard, empirically, convex, faster] [fab, log, latent, model, computational, fic, meibp, ibp, respect, bayesian, lfms, asymptotic, data, parameter, marginal, time, figure, znk, variational, factorized, indian, buffet, cccp, observed, form, distribution, sample, nonparametric, ficmm, yaleb, sampling, prior, shrinking, fics] [lower, approximation, shrinkage, selection, condition, set, note, lemma, case, machine, complexity, equivalent, theorem, asymptotically, deviation] [term, neural, increase, noise]
Tracking Time-varying Graphical Structure
Erich Kummerfeld, David Danks


[algorithm, online, current, assume, sequence, expected, previous] [learning, performance, learn, large, based, tracking, weight, multiple, shift, work, predict] [structure, losst, graphical, graph, relearning, size, datapoint, mahalanobis, datapoints, underlying, edge, weighted, pooled, changed, update, provably, respond, convergent, sem, perform, heuristic, undirected, focused, mellon, occurred, occur, recursive] [number, standard, distance, covariance, convergence, matrix, method] [model, data, sample, generating, bayesian, probability, locally, time, equation, figure, parameter, track, university, detect, stationary, generated, full] [effective, set, cost, provide] [change, stable, estimated, indicate]
Compressive Feature Learning
Hristo S. Paskov, Robert West, John C. Mitchell, Trevor Hastie