title
string
paper_url
string
authors
list
type
string
primary_area
string
abstract
large_string
keywords
list
TL;DR
large_string
submission_number
int64
arxiv_id
string
arxiv_id_source
string
embedding
list
Real Time Image Saliency for Black Box Classifiers
https://proceedings.neurips.cc/paper_files/paper/2017/hash/0060ef47b12160b9198302ebdb144dcf-Abstract.html
[ "Piotr Dabkowski", "Yarin Gal" ]
null
null
In this work we develop a fast saliency detection method that can be applied to any differentiable image classifier. We train a masking model to manipulate the scores of the classifier by masking salient parts of the input image. Our model generalises well to unseen images and requires a single forward pass to perform ...
[]
null
1
1705.07857
title_snapshot
[ 0.02058214507997036, -0.006060718093067408, 0.032200686633586884, 0.010825474746525288, 0.007025424856692553, 0.02776307240128517, 0.0046479105949401855, 0.022008640691637993, -0.042987555265426636, -0.042293939739465714, -0.04075806587934494, 0.01748029887676239, -0.047465428709983826, 0....
Joint distribution optimal transportation for domain adaptation
https://proceedings.neurips.cc/paper_files/paper/2017/hash/0070d23b06b1486a538c0eaa45dd167a-Abstract.html
[ "Nicolas Courty", "Rémi Flamary", "Amaury Habrard", "Alain Rakotomamonjy" ]
null
null
This paper deals with the unsupervised domain adaptation problem, where one wants to estimate a prediction function $f$ in a given target domain without any labeled sample by exploiting the knowledge available from a source domain where labels are known. Our work makes the following assumption: there exists a non-linea...
[]
null
2
1705.08848
title_snapshot
[ -0.028856877237558365, -0.011326405219733715, 0.013674633577466011, 0.042164403945207596, 0.052022647112607956, 0.026074083521962166, 0.017294714227318764, -0.03199140354990959, 0.010032943449914455, -0.03892436996102333, -0.029272504150867462, 0.02291705273091793, -0.08774011582136154, 0....
Learning A Structured Optimal Bipartite Graph for Co-Clustering
https://proceedings.neurips.cc/paper_files/paper/2017/hash/00a03ec6533ca7f5c644d198d815329c-Abstract.html
[ "Feiping Nie", "Xiaoqian Wang", "Cheng Deng", "Heng Huang" ]
null
null
Co-clustering methods have been widely applied to document clustering and gene expression analysis. These methods make use of the duality between features and samples such that the co-occurring structure of sample and feature clusters can be extracted. In graph based co-clustering methods, a bipartite graph is construc...
[]
null
3
null
null
[ -0.001826958847232163, -0.011842590756714344, -0.004485223907977343, 0.030436299741268158, 0.03538673371076584, 0.02737569436430931, 0.009838543832302094, -0.010291877202689648, 0.026403171941637993, -0.039323385804891586, 0.0021791444160044193, -0.02387409843504429, -0.06122225895524025, ...
Learning to Inpaint for Image Compression
https://proceedings.neurips.cc/paper_files/paper/2017/hash/013a006f03dbc5392effeb8f18fda755-Abstract.html
[ "Mohammad Haris Baig", "Vladlen Koltun", "Lorenzo Torresani" ]
null
null
We study the design of deep architectures for lossy image compression. We present two architectural recipes in the context of multi-stage progressive encoders and empirically demonstrate their importance on compression performance. Specifically, we show that: 1) predicting the original image data from residuals in a mu...
[]
null
4
1709.08855
title_snapshot
[ 0.010916806757450104, -0.03598533570766449, -0.02625655196607113, 0.06249323859810829, 0.059007178992033005, 0.03687189519405365, 0.015314684249460697, -0.019576173275709152, -0.0205245241522789, -0.06759639084339142, -0.01577686332166195, -0.04125043749809265, -0.030228665098547935, -0.01...
Inverse Filtering for Hidden Markov Models
https://proceedings.neurips.cc/paper_files/paper/2017/hash/01894d6f048493d2cacde3c579c315a3-Abstract.html
[ "Robert Mattila", "Cristian Rojas", "Vikram Krishnamurthy", "Bo Wahlberg" ]
null
null
This paper considers a number of related inverse filtering problems for hidden Markov models (HMMs). In particular, given a sequence of state posteriors and the system dynamics; i) estimate the corresponding sequence of observations, ii) estimate the observation likelihoods, and iii) jointly estimate the observation li...
[]
null
5
null
null
[ -0.02760612778365612, 0.026692280545830727, -0.009921400807797909, -0.003448270261287689, 0.060710757970809937, 0.032956983894109726, 0.06235614791512489, 0.018245941027998924, -0.031476058065891266, -0.03640519082546234, -0.007354554254561663, 0.018407469615340233, -0.0627746656537056, -0...
On clustering network-valued data
https://proceedings.neurips.cc/paper_files/paper/2017/hash/018dd1e07a2de4a08e6612341bf2323e-Abstract.html
[ "Soumendu Sundar Mukherjee", "Purnamrita Sarkar", "Lizhen Lin" ]
null
null
Community detection, which focuses on clustering nodes or detecting communities in (mostly) a single network, is a problem of considerable practical interest and has received a great deal of attention in the research community. While being able to cluster within a network is important, there are emerging needs to be ab...
[]
null
6
1606.02401
title_snapshot
[ 0.007732246071100235, -0.03492339327931404, 0.013300403952598572, 0.027704842388629913, 0.06283548474311829, 0.050803545862436295, 0.01922459900379181, 0.016140548512339592, -0.031283143907785416, -0.06073293089866638, 0.01325277704745531, -0.009466341696679592, -0.06439420580863953, 0.014...
Langevin Dynamics with Continuous Tempering for Training Deep Neural Networks
https://proceedings.neurips.cc/paper_files/paper/2017/hash/019d385eb67632a7e958e23f24bd07d7-Abstract.html
[ "Nanyang Ye", "Zhanxing Zhu", "Rafal Mantiuk" ]
null
null
Minimizing non-convex and high-dimensional objective functions is challenging, especially when training modern deep neural networks. In this paper, a novel approach is proposed which divides the training process into two consecutive phases to obtain better generalization performance: Bayesian sampling and stochastic op...
[]
null
7
1703.04379
title_snapshot
[ 0.003632055828347802, 0.0021275815088301897, -0.006203006953001022, 0.0540350079536438, 0.05289163440465927, 0.03700743988156319, 0.02541225031018257, -0.012493223883211613, -0.020564444363117218, -0.05335669592022896, 0.009113227017223835, 0.01670939102768898, -0.0515315867960453, 0.00004...
Beyond Worst-case: A Probabilistic Analysis of Affine Policies in Dynamic Optimization
https://proceedings.neurips.cc/paper_files/paper/2017/hash/01a0683665f38d8e5e567b3b15ca98bf-Abstract.html
[ "Omar El Housni", "Vineet Goyal" ]
null
null
Affine policies (or control) are widely used as a solution approach in dynamic optimization where computing an optimal adjustable solution is usually intractable. While the worst case performance of affine policies can be significantly bad, the empirical performance is observed to be near-optimal for a large class of p...
[]
null
8
1706.05737
title_snapshot
[ -0.06705871969461441, 0.0028085007797926664, 0.008429826237261295, 0.018591389060020447, 0.049573514610528946, 0.04172327741980553, 0.00013721513096243143, -0.023769665509462357, -0.016287103295326233, -0.03212722763419151, -0.01505959127098322, -0.026420289650559425, -0.08147238194942474, ...
Few-Shot Learning Through an Information Retrieval Lens
https://proceedings.neurips.cc/paper_files/paper/2017/hash/01e9565cecc4e989123f9620c1d09c09-Abstract.html
[ "Eleni Triantafillou", "Richard Zemel", "Raquel Urtasun" ]
null
null
Few-shot learning refers to understanding new concepts from only a few examples. We propose an information retrieval-inspired approach for this problem that is motivated by the increased importance of maximally leveraging all the available information in this low-data regime. We define a training objective that aims to...
[]
null
9
1707.02610
title_snapshot
[ -0.019581733271479607, -0.027954932302236557, -0.01675308123230934, 0.05171036347746849, 0.026097644120454788, 0.0002597525017336011, 0.025832733139395714, 0.005939655005931854, -0.030651366338133812, 0.005648090969771147, -0.03137096017599106, 0.0399160273373127, -0.05389147996902466, -0....
Accelerated consensus via Min-Sum Splitting
https://proceedings.neurips.cc/paper_files/paper/2017/hash/024d7f84fff11dd7e8d9c510137a2381-Abstract.html
[ "Patrick Rebeschini", "Sekhar C Tatikonda" ]
null
null
We apply the Min-Sum message-passing protocol to solve the consensus problem in distributed optimization. We show that while the ordinary Min-Sum algorithm does not converge, a modified version of it known as Splitting yields convergence to the problem solution. We prove that a proper choice of the tuning parameters al...
[]
null
10
1706.03807
title_snapshot
[ -0.029821030795574188, -0.011500178836286068, 0.007936990819871426, 0.029366087168455124, 0.03135768324136734, 0.02866283990442753, 0.02677057683467865, 0.0225411057472229, -0.007838202640414238, -0.0278396625071764, 0.009167024865746498, -0.0012740076053887606, -0.07252741605043411, -0.00...
Saliency-based Sequential Image Attention with Multiset Prediction
https://proceedings.neurips.cc/paper_files/paper/2017/hash/028ee724157b05d04e7bdcf237d12e60-Abstract.html
[ "Sean Welleck", "Jialin Mao", "Kyunghyun Cho", "Zheng Zhang" ]
null
null
Humans process visual scenes selectively and sequentially using attention. Central to models of human visual attention is the saliency map. We propose a hierarchical visual architecture that operates on a saliency map and uses a novel attention mechanism to sequentially focus on salient regions and take additional glim...
[]
null
11
1711.05165
title_snapshot
[ 0.010282007046043873, 0.012724803760647774, 0.0171970184892416, 0.02984268218278885, 0.017643941566348076, 0.013615220785140991, 0.01804278790950775, 0.03517058119177818, -0.04765848070383072, -0.044321704655885696, -0.04889651760458946, 0.01188328955322504, -0.06448562443256378, -0.012743...
Adaptive Bayesian Sampling with Monte Carlo EM
https://proceedings.neurips.cc/paper_files/paper/2017/hash/02a32ad2669e6fe298e607fe7cc0e1a0-Abstract.html
[ "Anirban Roychowdhury", "Srinivasan Parthasarathy" ]
null
null
We present a novel technique for learning the mass matrices in samplers obtained from discretized dynamics that preserve some energy function. Existing adaptive samplers use Riemannian preconditioning techniques, where the mass matrices are functions of the parameters being sampled. This leads to significant complexiti...
[]
null
12
1711.02159
title_snapshot
[ -0.03659958392381668, 0.008427872322499752, 0.016103355213999748, 0.06055782362818718, 0.02969324216246605, 0.008724397979676723, 0.0227255430072546, -0.014509263448417187, -0.042113058269023895, -0.07861453294754028, 0.017637616023421288, 0.006015910767018795, -0.05401403084397316, -0.008...
Scalable Levy Process Priors for Spectral Kernel Learning
https://proceedings.neurips.cc/paper_files/paper/2017/hash/02b1be0d48924c327124732726097157-Abstract.html
[ "Phillip A Jang", "Andrew Loeb", "Matthew Davidow", "Andrew G Wilson" ]
null
null
Gaussian processes are rich distributions over functions, with generalization properties determined by a kernel function. When used for long-range extrapolation, predictions are particularly sensitive to the choice of kernel parameters. It is therefore critical to account for kernel uncertainty in our predictive distri...
[]
null
13
1802.00530
title_snapshot
[ -0.023627394810318947, -0.011442461051046848, 0.02282111905515194, 0.02681805193424225, 0.030231138691306114, 0.026237644255161285, -0.0031468423549085855, -0.02153521217405796, -0.023348838090896606, -0.05112398415803909, -0.016831425949931145, 0.01843268983066082, -0.05189564451575279, 0...
Model-Powered Conditional Independence Test
https://proceedings.neurips.cc/paper_files/paper/2017/hash/02f039058bd48307e6f653a2005c9dd2-Abstract.html
[ "Rajat Sen", "Ananda Theertha Suresh", "Karthikeyan Shanmugam", "Alexandros G Dimakis", "Sanjay Shakkottai" ]
null
null
We consider the problem of non-parametric Conditional Independence testing (CI testing) for continuous random variables. Given i.i.d samples from the joint distribution $f(x,y,z)$ of continuous random vectors $X,Y$ and $Z,$ we determine whether $X \independent Y \vert Z$. We approach this by converting the conditional ...
[]
null
14
1709.06138
title_snapshot
[ -0.024819303303956985, -0.010341440327465534, -0.006282097194343805, 0.057314805686473846, 0.0488288477063179, 0.03569745272397995, 0.011116729117929935, -0.02071407437324524, 0.005269054789096117, -0.02843773551285267, -0.004642288200557232, 0.0191794466227293, -0.07751284539699554, -0.00...
Learning Multiple Tasks with Multilinear Relationship Networks
https://proceedings.neurips.cc/paper_files/paper/2017/hash/03e0704b5690a2dee1861dc3ad3316c9-Abstract.html
[ "Mingsheng Long", "ZHANGJIE CAO", "Jianmin Wang", "Philip S Yu" ]
null
null
Deep networks trained on large-scale data can learn transferable features to promote learning multiple tasks. Since deep features eventually transition from general to specific along deep networks, a fundamental problem of multi-task learning is how to exploit the task relatedness underlying parameter tensors and impro...
[]
null
15
1506.02117
title_snapshot
[ -0.015677189454436302, 0.004847750533372164, 0.007859881035983562, 0.016149334609508514, 0.039504487067461014, 0.03425127640366554, 0.0106216324493289, -0.0009702121606096625, 0.01403006911277771, -0.03456350788474083, -0.018647989258170128, 0.03480146825313568, -0.06489849090576172, -0.00...
Query Complexity of Clustering with Side Information
https://proceedings.neurips.cc/paper_files/paper/2017/hash/03e7ef47cee6fa4ae7567394b99912b7-Abstract.html
[ "Arya Mazumdar", "Barna Saha" ]
null
null
Suppose, we are given a set of $n$ elements to be clustered into $k$ (unknown) clusters, and an oracle/expert labeler that can interactively answer pair-wise queries of the form, ``do two elements $u$ and $v$ belong to the same cluster?''. The goal is to recover the optimum clustering by asking the minimum number of qu...
[]
null
16
1706.07719
title_snapshot
[ -0.051951512694358826, -0.0031680252868682146, 0.007040569558739662, 0.01528911106288433, 0.030069362372159958, 0.01640917919576168, 0.012329675257205963, -0.02251354232430458, -0.012799491174519062, -0.01622428372502327, -0.015022332780063152, 0.008070539683103561, -0.05020138621330261, 0...
Non-parametric Structured Output Networks
https://proceedings.neurips.cc/paper_files/paper/2017/hash/04048aeca2c0f5d84639358008ed2ae7-Abstract.html
[ "Andreas Lehrmann", "Leonid Sigal" ]
null
null
Deep neural networks (DNNs) and probabilistic graphical models (PGMs) are the two main tools for statistical modeling. While DNNs provide the ability to model rich and complex relationships between input and output variables, PGMs provide the ability to encode dependencies among the output variables themselves. End-to-...
[]
null
17
null
null
[ -0.0009389148908667266, -0.005252554081380367, -0.030547868460416794, 0.04653460904955864, 0.046739399433135986, 0.05443049222230911, 0.01377997174859047, -0.012405085377395153, -0.0050733801908791065, -0.05643386021256447, 0.003753785276785493, 0.013407791964709759, -0.07381996512413025, ...
Robust Imitation of Diverse Behaviors
https://proceedings.neurips.cc/paper_files/paper/2017/hash/044a23cadb567653eb51d4eb40acaa88-Abstract.html
[ "Ziyu Wang", "Josh S Merel", "Scott E Reed", "Nando de Freitas", "Gregory Wayne", "Nicolas Heess" ]
null
null
Deep generative models have recently shown great promise in imitation learning for motor control. Given enough data, even supervised approaches can do one-shot imitation learning; however, they are vulnerable to cascading failures when the agent trajectory diverges from the demonstrations. Compared to purely supervised...
[]
null
18
1707.02747
title_snapshot
[ -0.00013534913887269795, -0.03915500268340111, -0.02685600519180298, 0.05372762307524681, 0.01960994489490986, 0.02945011854171753, 0.028747472912073135, -0.0049684010446071625, -0.03671787679195404, -0.03461427241563797, -0.008666737005114555, -0.018892642110586166, -0.060591645538806915, ...
High-Order Attention Models for Visual Question Answering
https://proceedings.neurips.cc/paper_files/paper/2017/hash/051928341be67dcba03f0e04104d9047-Abstract.html
[ "Idan Schwartz", "Alexander Schwing", "Tamir Hazan" ]
null
null
The quest for algorithms that enable cognitive abilities is an important part of machine learning. A common trait in many recently investigated cognitive-like tasks is that they take into account different data modalities, such as visual and textual input. In this paper we propose a novel and generally applicable form ...
[]
null
19
1711.04323
title_snapshot
[ 0.013676810078322887, 0.003440143307670951, 0.008867193944752216, 0.05781349539756775, 0.007672747131437063, 0.02198200859129429, 0.011467683129012585, 0.02304641157388687, -0.01992005668580532, 0.00896216556429863, -0.05079123005270958, 0.04273240268230438, -0.08405845612287521, -0.018991...
FALKON: An Optimal Large Scale Kernel Method
https://proceedings.neurips.cc/paper_files/paper/2017/hash/05546b0e38ab9175cd905eebcc6ebb76-Abstract.html
[ "Alessandro Rudi", "Luigi Carratino", "Lorenzo Rosasco" ]
null
null
Kernel methods provide a principled way to perform non linear, nonparametric learning. They rely on solid functional analytic foundations and enjoy optimal statistical properties. However, at least in their basic form, they have limited applicability in large scale scenarios because of stringent computational requireme...
[]
null
20
1705.10958
title_snapshot
[ -0.03231506049633026, -0.027333378791809082, 0.040066562592983246, 0.03683837875723839, 0.012034825049340725, 0.06266648322343826, 0.0029822038486599922, -0.01177587267011404, -0.011742057278752327, -0.03765082731842995, -0.0005358225898817182, -0.026675205677747726, -0.07080549746751785, ...
Generalized Linear Model Regression under Distance-to-set Penalties
https://proceedings.neurips.cc/paper_files/paper/2017/hash/061412e4a03c02f9902576ec55ebbe77-Abstract.html
[ "Jason Xu", "Eric Chi", "Kenneth Lange" ]
null
null
Estimation in generalized linear models (GLM) is complicated by the presence of constraints. One can handle constraints by maximizing a penalized log-likelihood. Penalties such as the lasso are effective in high dimensions but often lead to severe shrinkage. This paper explores instead penalizing the squared distance t...
[]
null
21
1711.01341
title_snapshot
[ 0.0015109851956367493, -0.00017761833441909403, 0.0008286864031106234, -0.001492718351073563, 0.03932250663638115, 0.051007289439439774, 0.04338773712515831, -0.013822639361023903, -0.04530825465917587, -0.03746664524078369, -0.02108926698565483, 0.018251225352287292, -0.09097249060869217, ...
Fisher GAN
https://proceedings.neurips.cc/paper_files/paper/2017/hash/07042ac7d03d3b9911a00da43ce0079a-Abstract.html
[ "Youssef Mroueh", "Tom Sercu" ]
null
null
Generative Adversarial Networks (GANs) are powerful models for learning complex distributions. Stable training of GANs has been addressed in many recent works which explore different metrics between distributions. In this paper we introduce Fisher GAN that fits within the Integral Probability Metrics (IPM) framework fo...
[]
null
22
1705.09675
title_snapshot
[ -0.00046082769404165447, -0.03330918028950691, 0.004580507520586252, 0.05170362815260887, 0.01888209395110607, 0.035056475549936295, 0.0010653052013367414, -0.015085859224200249, -0.018628859892487526, -0.05985860526561737, -0.004668669309467077, 0.023909451439976692, -0.08121058344841003, ...
Minimax Estimation of Bandable Precision Matrices
https://proceedings.neurips.cc/paper_files/paper/2017/hash/070dbb6024b5ef93784428afc71f2146-Abstract.html
[ "Addison Hu", "Sahand Negahban" ]
null
null
The inverse covariance matrix provides considerable insight for understanding statistical models in the multivariate setting. In particular, when the distribution over variables is assumed to be multivariate normal, the sparsity pattern in the inverse covariance matrix, commonly referred to as the precision matrix, cor...
[]
null
23
1710.07006
title_snapshot
[ -0.024869590997695923, -0.008708796463906765, 0.018897265195846558, 0.009557320736348629, 0.023320134729146957, 0.026560036465525627, 0.044615283608436584, -0.007166862953454256, -0.02386936917901039, -0.05148415267467499, -0.000679929566103965, -0.011774915270507336, -0.06995054334402084, ...
Kernel functions based on triplet comparisons
https://proceedings.neurips.cc/paper_files/paper/2017/hash/07211688a0869d995947a8fb11b215d6-Abstract.html
[ "Matthäus Kleindessner", "Ulrike von Luxburg" ]
null
null
Given only information in the form of similarity triplets "Object A is more similar to object B than to object C" about a data set, we propose two ways of defining a kernel function on the data set. While previous approaches construct a low-dimensional Euclidean embedding of the data set that reflects the given similar...
[]
null
24
1607.08456
title_snapshot
[ -0.005209255497902632, -0.013162423856556416, 0.017615653574466705, 0.06411363184452057, 0.04850068688392639, 0.06259424239397049, 0.0021415266674011946, -0.04268767684698105, 0.009998506866395473, -0.0389133058488369, -0.007203016895800829, 0.03598983213305473, -0.07474052160978317, 0.020...
Breaking the Nonsmooth Barrier: A Scalable Parallel Method for Composite Optimization
https://proceedings.neurips.cc/paper_files/paper/2017/hash/072b030ba126b2f4b2374f342be9ed44-Abstract.html
[ "Fabian Pedregosa", "Rémi Leblond", "Simon Lacoste-Julien" ]
null
null
Due to their simplicity and excellent performance, parallel asynchronous variants of stochastic gradient descent have become popular methods to solve a wide range of large-scale optimization problems on multi-core architectures. Yet, despite their practical success, support for nonsmooth objectives is still lacking, ma...
[]
null
25
1707.06468
title_snapshot
[ -0.0388403944671154, -0.04213285818696022, 0.0005407692515291274, 0.02871089056134224, 0.016568109393119812, 0.042242467403411865, 0.03135142847895622, 0.012145855464041233, -0.043748632073402405, -0.043634798377752304, 0.0087747098878026, -0.01866559311747551, -0.05149780213832855, -0.004...
A New Theory for Matrix Completion
https://proceedings.neurips.cc/paper_files/paper/2017/hash/07563a3fe3bbe7e3ba84431ad9d055af-Abstract.html
[ "Guangcan Liu", "Qingshan Liu", "Xiaotong Yuan" ]
null
null
Prevalent matrix completion theories reply on an assumption that the locations of the missing data are distributed uniformly and randomly (i.e., uniform sampling). Nevertheless, the reason for observations being missing often depends on the unseen observations themselves, and thus the missing data in practice usually o...
[]
null
26
null
null
[ -0.01882489211857319, -0.024981820955872536, 0.0063912938348948956, 0.06505881249904633, 0.03169556334614754, 0.025026820600032806, 0.01502219308167696, 0.015132077038288116, -0.04091817885637283, -0.07242472469806671, -0.028043122962117195, -0.013515046797692776, -0.06100304424762726, -0....
A Bayesian Data Augmentation Approach for Learning Deep Models
https://proceedings.neurips.cc/paper_files/paper/2017/hash/076023edc9187cf1ac1f1163470e479a-Abstract.html
[ "Toan Tran", "Trung Pham", "Gustavo Carneiro", "Lyle Palmer", "Ian Reid" ]
null
null
Data augmentation is an essential part of the training process applied to deep learning models. The motivation is that a robust training process for deep learning models depends on large annotated datasets, which are expensive to be acquired, stored and processed. Therefore a reasonable alternative is to be able to aut...
[]
null
27
1710.10564
title_snapshot
[ 0.009233885444700718, -0.02443966083228588, -0.013558445498347282, 0.06282471865415573, 0.027461273595690727, 0.022490134462714195, 0.014270010404288769, -0.002376388991251588, -0.008919809013605118, -0.08251464366912842, -0.04114854335784912, -0.005523044150322676, -0.08241715282201767, 0...
Deep Hyperalignment
https://proceedings.neurips.cc/paper_files/paper/2017/hash/0768281a05da9f27df178b5c39a51263-Abstract.html
[ "Muhammad Yousefnezhad", "Daoqiang Zhang" ]
null
null
This paper proposes Deep Hyperalignment (DHA) as a regularized, deep extension, scalable Hyperalignment (HA) method, which is well-suited for applying functional alignment to fMRI datasets with nonlinearity, high-dimensionality (broad ROI), and a large number of subjects. Unlink previous methods, DHA is not limited by ...
[]
null
28
1710.03923
title_snapshot
[ -0.0036792252212762833, 0.03491402789950371, 0.004006682429462671, 0.013344597071409225, 0.009005269035696983, 0.054252881556749344, 0.05162636190652847, -0.020873768255114555, -0.03324455767869949, -0.05067555978894234, -0.007998638786375523, 0.021359028294682503, -0.06121121719479561, -0...
Best of Both Worlds: Transferring Knowledge from Discriminative Learning to a Generative Visual Dialog Model
https://proceedings.neurips.cc/paper_files/paper/2017/hash/077e29b11be80ab57e1a2ecabb7da330-Abstract.html
[ "Jiasen Lu", "Anitha Kannan", "Jianwei Yang", "Devi Parikh", "Dhruv Batra" ]
null
null
We present a novel training framework for neural sequence models, particularly for grounded dialog generation. The standard training paradigm for these models is maximum likelihood estimation (MLE), or minimizing the cross-entropy of the human responses. Across a variety of domains, a recurring problem with MLE trained...
[]
null
29
1706.01554
title_snapshot
[ 0.003228310961276293, -0.01103885006159544, 0.005443851463496685, 0.053946636617183685, 0.021590907126665115, 0.0266556516289711, 0.005695533473044634, 0.02173835225403309, 0.008464327082037926, -0.023532934486865997, -0.04118765890598297, 0.047317985445261, -0.06711234152317047, -0.013917...
PASS-GLM: polynomial approximate sufficient statistics for scalable Bayesian GLM inference
https://proceedings.neurips.cc/paper_files/paper/2017/hash/07811dc6c422334ce36a09ff5cd6fe71-Abstract.html
[ "Jonathan Huggins", "Ryan P. Adams", "Tamara Broderick" ]
null
null
Generalized linear models (GLMs)---such as logistic regression, Poisson regression, and robust regression---provide interpretable models for diverse data types. Probabilistic approaches, particularly Bayesian ones, allow coherent estimates of uncertainty, incorporation of prior information, and sharing of power across ...
[]
null
30
1709.09216
title_snapshot
[ 0.013191751204431057, -0.008493673987686634, 0.010847032070159912, 0.011151899583637714, 0.05446706712245941, 0.020596737042069435, 0.0567084364593029, -0.018586477264761925, -0.03858354687690735, -0.026092588901519775, -0.025672106072306633, 0.011282644234597683, -0.057231053709983826, 0....
Online multiclass boosting
https://proceedings.neurips.cc/paper_files/paper/2017/hash/08b255a5d42b89b0585260b6f2360bdd-Abstract.html
[ "Young Hun Jung", "Jack Goetz", "Ambuj Tewari" ]
null
null
Recent work has extended the theoretical analysis of boosting algorithms to multiclass problems and to online settings. However, the multiclass extension is in the batch setting and the online extensions only consider binary classification. We fill this gap in the literature by defining, and justifying, a weak learning...
[]
null
31
1702.07305
title_snapshot
[ -0.01761161908507347, -0.038789957761764526, 0.015586269088089466, 0.02082708477973938, 0.0230389554053545, 0.03050502948462963, 0.00003026906961167697, -0.0015525706112384796, -0.02090146578848362, -0.028083933517336845, -0.034850820899009705, 0.03205365687608719, -0.08040990680456161, -0...
State Aware Imitation Learning
https://proceedings.neurips.cc/paper_files/paper/2017/hash/08e6bea8e90ba87af3c9554d94db6579-Abstract.html
[ "Yannick Schroecker", "Charles L Isbell" ]
null
null
Imitation learning is the study of learning how to act given a set of demonstrations provided by a human expert. It is intuitively apparent that learning to take optimal actions is a simpler undertaking in situations that are similar to the ones shown by the teacher. However, imitation learning approaches do not tend t...
[]
null
32
null
null
[ -0.011456857435405254, -0.026663854718208313, -0.011950846761465073, 0.02601650543510914, 0.05038342624902725, 0.013735143467783928, 0.025381334125995636, 0.017123645171523094, -0.02387438714504242, -0.020251475274562836, -0.01495274156332016, 0.00829185638576746, -0.057013049721717834, -0...
Adaptive SVRG Methods under Error Bound Conditions with Unknown Growth Parameter
https://proceedings.neurips.cc/paper_files/paper/2017/hash/09fb05dd477d4ae6479985ca56c5a12d-Abstract.html
[ "Yi Xu", "Qihang Lin", "Tianbao Yang" ]
null
null
Error bound, an inherent property of an optimization problem, has recently revived in the development of algorithms with improved global convergence without strong convexity. The most studied error bound is the quadratic error bound, which generalizes strong convexity and is satisfied by a large family of machine learn...
[]
null
33
null
null
[ -0.03538825362920761, -0.014630119316279888, 0.0470995269715786, 0.016249464824795723, 0.02387346141040325, 0.04928367957472801, 0.050768349319696426, -0.024175036698579788, -0.05261983349919319, -0.04622220620512962, -0.023249346762895584, 0.019770538434386253, -0.04392159357666969, 0.002...
Unsupervised Learning of Disentangled and Interpretable Representations from Sequential Data
https://proceedings.neurips.cc/paper_files/paper/2017/hash/0a0a0c8aaa00ade50f74a3f0ca981ed7-Abstract.html
[ "Wei-Ning Hsu", "Yu Zhang", "James Glass" ]
null
null
We present a factorized hierarchical variational autoencoder, which learns disentangled and interpretable representations from sequential data without supervision. Specifically, we exploit the multi-scale nature of information in sequential data by formulating it explicitly within a factorized hierarchical graphical mo...
[]
null
34
1709.07902
title_snapshot
[ 0.003924272488802671, -0.003961178939789534, -0.03390643373131752, 0.04918542876839638, 0.04619190841913223, 0.0466824509203434, 0.054519619792699814, 0.013878319412469864, 0.004121201578527689, -0.026820536702871323, -0.005383169278502464, 0.027023689821362495, -0.06167064607143402, 0.015...
Recurrent Ladder Networks
https://proceedings.neurips.cc/paper_files/paper/2017/hash/0a5c79b1eaf15445da252ada718857e9-Abstract.html
[ "Isabeau Prémont-Schwarz", "Alexander Ilin", "Tele Hao", "Antti Rasmus", "Rinu Boney", "Harri Valpola" ]
null
null
We propose a recurrent extension of the Ladder networks whose structure is motivated by the inference required in hierarchical latent variable models. We demonstrate that the recurrent Ladder is able to handle a wide variety of complex learning tasks that benefit from iterative inference and temporal modeling. The arch...
[]
null
35
1707.09219
title_snapshot
[ 0.020389532670378685, -0.009024168364703655, 0.010171940550208092, 0.039848268032073975, 0.03918401524424553, 0.027161451056599617, 0.04109674319624901, 0.015644673258066177, -0.045712318271398544, -0.030931957066059113, 0.007883921265602112, 0.0073554376140236855, -0.05096780136227608, 0....
Distral: Robust multitask reinforcement learning
https://proceedings.neurips.cc/paper_files/paper/2017/hash/0abdc563a06105aee3c6136871c9f4d1-Abstract.html
[ "Yee Teh", "Victor Bapst", "Wojciech M. Czarnecki", "John Quan", "James Kirkpatrick", "Raia Hadsell", "Nicolas Heess", "Razvan Pascanu" ]
null
null
Most deep reinforcement learning algorithms are data inefficient in complex and rich environments, limiting their applicability to many scenarios. One direction for improving data efficiency is multitask learning with shared neural network parameters, where efficiency may be improved through transfer across related tas...
[]
null
36
1707.04175
title_snapshot
[ 0.01026999857276678, -0.019745012745261192, -0.013755910098552704, 0.043607644736766815, 0.034306660294532776, 0.007117207627743483, 0.01242948416620493, -0.012331939302384853, -0.008431300520896912, -0.05891932174563408, -0.018147306516766548, -0.004960660357028246, -0.07231307029724121, ...
Real-Time Bidding with Side Information
https://proceedings.neurips.cc/paper_files/paper/2017/hash/0bed45bd5774ffddc95ffe500024f628-Abstract.html
[ "arthur flajolet", "Patrick Jaillet" ]
null
null
We consider the problem of repeated bidding in online advertising auctions when some side information (e.g. browser cookies) is available ahead of submitting a bid in the form of a $d$-dimensional vector. The goal for the advertiser is to maximize the total utility (e.g. the total number of clicks) derived from display...
[]
null
37
null
null
[ -0.046044036746025085, -0.0035859569907188416, -0.002026777947321534, 0.03746937960386276, 0.013831124641001225, 0.03091992810368538, 0.01786866784095764, 0.052168287336826324, -0.006162043660879135, -0.03566452115774155, -0.004682631231844425, 0.012306053191423416, -0.05745670944452286, -...
Learning Spherical Convolution for Fast Features from 360° Imagery
https://proceedings.neurips.cc/paper_files/paper/2017/hash/0c74b7f78409a4022a2c4c5a5ca3ee19-Abstract.html
[ "Yu-Chuan Su", "Kristen Grauman" ]
null
null
While 360° cameras offer tremendous new possibilities in vision, graphics, and augmented reality, the spherical images they produce make core feature extraction non-trivial. Convolutional neural networks (CNNs) trained on images from perspective cameras yield “flat" filters, yet 360° images cannot be projected to a sin...
[]
null
38
1708.00919
title_snapshot
[ 0.040789615362882614, 0.018571989610791206, 0.015062488615512848, 0.020949063822627068, 0.019040780141949654, 0.03826109692454338, 0.008069193921983242, 0.018314078450202942, -0.024129509925842285, -0.05096793547272682, -0.02356000989675522, -0.02503477782011032, -0.09250190854072571, 0.01...
Approximate Supermodularity Bounds for Experimental Design
https://proceedings.neurips.cc/paper_files/paper/2017/hash/0d9095b0d6bbe98ea0c9c02b11b59ee3-Abstract.html
[ "Luiz Chamon", "Alejandro Ribeiro" ]
null
null
This work provides performance guarantees for the greedy solution of experimental design problems. In particular, it focuses on A- and E-optimal designs, for which typical guarantees do not apply since the mean-square error and the maximum eigenvalue of the estimation error covariance matrix are not supermodular. To do...
[]
null
39
1711.01501
title_snapshot
[ -0.023639477789402008, 0.020655108615756035, -0.01970473863184452, 0.04269790276885033, 0.06046517193317413, 0.0021911372896283865, 0.037367794662714005, -0.043497029691934586, -0.019039936363697052, -0.03449905291199684, -0.01795261539518833, -0.010336212813854218, -0.07424698024988174, -...
Differentiable Learning of Logical Rules for Knowledge Base Reasoning
https://proceedings.neurips.cc/paper_files/paper/2017/hash/0e55666a4ad822e0e34299df3591d979-Abstract.html
[ "Fan Yang", "Zhilin Yang", "William W. Cohen" ]
null
null
We study the problem of learning probabilistic first-order logical rules for knowledge base reasoning. This learning problem is difficult because it requires learning the parameters in a continuous space as well as the structure in a discrete space. We propose a framework, Neural Logic Programming, that combines the pa...
[]
null
40
1702.08367
title_snapshot
[ -0.010323097929358482, 0.003872611792758107, -0.011244279332458973, 0.024806426838040352, 0.07292705029249191, -0.013478509150445461, 0.015489483252167702, -0.015072491019964218, -0.0179173331707716, -0.005457174498587847, -0.009853227064013481, 0.051445599645376205, -0.06487210094928741, ...
When Cyclic Coordinate Descent Outperforms Randomized Coordinate Descent
https://proceedings.neurips.cc/paper_files/paper/2017/hash/0e7c7d6c41c76b9ee6445ae01cc0181d-Abstract.html
[ "Mert Gurbuzbalaban", "Asuman Ozdaglar", "Pablo A Parrilo", "Nuri Vanli" ]
null
null
The coordinate descent (CD) method is a classical optimization algorithm that has seen a revival of interest because of its competitive performance in machine learning applications. A number of recent papers provided convergence rate estimates for their deterministic (cyclic) and randomized variants that differ in the ...
[]
null
41
null
null
[ -0.02063596062362194, -0.02631657011806965, -0.0030780702363699675, 0.04551251605153084, 0.042465727776288986, 0.022133825346827507, 0.0030742890667170286, 0.00793374516069889, -0.01940419152379036, -0.042299240827560425, -0.007138072047382593, -0.0056716641411185265, -0.05638344958424568, ...
Principles of Riemannian Geometry in Neural Networks
https://proceedings.neurips.cc/paper_files/paper/2017/hash/0ebcc77dc72360d0eb8e9504c78d38bd-Abstract.html
[ "Michael Hauser", "Asok Ray" ]
null
null
This study deals with neural networks in the sense of geometric transformations acting on the coordinate representation of the underlying data manifold which the data is sampled from. It forms part of an attempt to construct a formalized general theory of neural networks in the setting of Riemannian geometry. From this...
[]
null
42
null
null
[ -0.061844147741794586, 0.017363399267196655, 0.02399279735982418, 0.035115450620651245, 0.025816388428211212, 0.048297908157110214, 0.004416181705892086, 0.029084838926792145, -0.04235813021659851, -0.05216512084007263, -0.007256083190441132, -0.0026359029579907656, -0.057080794125795364, ...
Continual Learning with Deep Generative Replay
https://proceedings.neurips.cc/paper_files/paper/2017/hash/0efbe98067c6c73dba1250d2beaa81f9-Abstract.html
[ "Hanul Shin", "Jung Kwon Lee", "Jaehong Kim", "Jiwon Kim" ]
null
null
Attempts to train a comprehensive artificial intelligence capable of solving multiple tasks have been impeded by a chronic problem called catastrophic forgetting. Although simply replaying all previous data alleviates the problem, it requires large memory and even worse, often infeasible in real world applications wher...
[]
null
43
1705.08690
title_snapshot
[ -0.0034410301595926285, -0.025728624314069748, -0.02061951905488968, 0.05353816971182823, 0.03954588249325752, 0.019961755722761154, 0.02439289540052414, 0.030941098928451538, -0.0441233329474926, -0.05011774227023125, -0.011634759604930878, -0.002852496225386858, -0.05146037042140961, 0.0...
Nonlinear random matrix theory for deep learning
https://proceedings.neurips.cc/paper_files/paper/2017/hash/0f3d014eead934bbdbacb62a01dc4831-Abstract.html
[ "Jeffrey Pennington", "Pratik Worah" ]
null
null
Neural network configurations with random weights play an important role in the analysis of deep learning. They define the initial loss landscape and are closely related to kernel and random feature methods. Despite the fact that these networks are built out of random matrices, the vast and powerful machinery of random...
[]
null
44
null
null
[ -0.0250231996178627, -0.02573213540017605, 0.014137103222310543, 0.026315871626138687, 0.004654638934880495, 0.027549123391509056, 0.027281345799565315, 0.013911345042288303, -0.020315267145633698, -0.05408221855759621, -0.013251732103526592, -0.009869513101875782, -0.03474082425236702, -0...
Identification of Gaussian Process State Space Models
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1006ff12c465532f8c574aeaa4461b16-Abstract.html
[ "Stefanos Eleftheriadis", "Tom Nicholson", "Marc Deisenroth", "James Hensman" ]
null
null
The Gaussian process state space model (GPSSM) is a non-linear dynamical system, where unknown transition and/or measurement mappings are described by GPs. Most research in GPSSMs has focussed on the state estimation problem, i.e., computing a posterior of the latent state given the model. However, the key challenge in...
[]
null
45
1705.10888
title_snapshot
[ -0.016657959669828415, -0.003234939184039831, -0.011229856871068478, 0.026535317301750183, 0.03931461274623871, 0.033780794590711594, 0.02703072875738144, 0.03745028376579285, -0.022052621468901634, -0.02167762629687786, 0.007229458540678024, -0.0017248016083613038, -0.044768575578927994, ...
Estimation of the covariance structure of heavy-tailed distributions
https://proceedings.neurips.cc/paper_files/paper/2017/hash/10c272d06794d3e5785d5e7c5356e9ff-Abstract.html
[ "Xiaohan Wei", "Stanislav Minsker" ]
null
null
We propose and analyze a new estimator of the covariance matrix that admits strong theoretical guarantees under weak assumptions on the underlying distribution, such as existence of moments of only low order. While estimation of covariance matrices corresponding to sub-Gaussian distributions is well-understood, much le...
[]
null
46
1708.00502
title_snapshot
[ -0.011536846868693829, -0.015150437131524086, 0.024912267923355103, -0.019229285418987274, 0.027280030772089958, 0.020808225497603416, 0.03274286910891533, 0.0006736012874171138, -0.0034414189867675304, -0.02444631978869438, 0.0176225733011961, -0.025411557406187057, -0.07103428989648819, ...
Robust Optimization for Non-Convex Objectives
https://proceedings.neurips.cc/paper_files/paper/2017/hash/10c66082c124f8afe3df4886f5e516e0-Abstract.html
[ "Robert S. Chen", "Brendan Lucier", "Yaron Singer", "Vasilis Syrgkanis" ]
null
null
We consider robust optimization problems, where the goal is to optimize in the worst case over a class of objective functions. We develop a reduction from robust improper optimization to stochastic optimization: given an oracle that returns $\alpha$-approximate solutions for distributions over objectives, we compute a ...
[]
null
47
1707.01047
title_snapshot
[ -0.03646554797887802, 0.017954207956790924, -0.0029879857320338488, 0.043430011719465256, 0.02509760484099388, 0.04811729118227959, 0.019281327724456787, -0.012058709748089314, -0.014959072694182396, -0.029944973066449165, -0.029098646715283394, -0.021584421396255493, -0.06268025189638138, ...
Exploring Generalization in Deep Learning
https://proceedings.neurips.cc/paper_files/paper/2017/hash/10ce03a1ed01077e3e289f3e53c72813-Abstract.html
[ "Behnam Neyshabur", "Srinadh Bhojanapalli", "David Mcallester", "Nati Srebro" ]
null
null
With a goal of understanding what drives generalization in deep networks, we consider several recently suggested explanations, including norm-based control, sharpness and robustness. We study how these measures can ensure generalization, highlighting the importance of scale normalization, and making a connection betwee...
[]
null
48
1706.08947
title_snapshot
[ -0.022067882120609283, -0.02741251327097416, -0.004316789098083973, 0.044813286513090134, 0.046924229711294174, 0.018959227949380875, 0.03238670900464058, -0.007731469813734293, -0.023858407512307167, -0.04602593183517456, -0.014101896435022354, 0.01461336761713028, -0.047499798238277435, ...
Spherical convolutions and their application in molecular modelling
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1113d7a76ffceca1bb350bfe145467c6-Abstract.html
[ "Wouter Boomsma", "Jes Frellsen" ]
null
null
Convolutional neural networks are increasingly used outside the domain of image analysis, in particular in various areas of the natural sciences concerned with spatial data. Such networks often work out-of-the box, and in some cases entire model architectures from image analysis can be carried over to other problem dom...
[]
null
49
null
null
[ -0.002222012262791395, 0.0022085311356931925, 0.01536490861326456, 0.025889527052640915, 0.024470025673508644, -0.00037821524892933667, -0.004138899967074394, -0.003676427761092782, 0.005378722678869963, -0.041411351412534714, 0.029377073049545288, -0.019399279728531837, -0.09406928718090057...
Safe Adaptive Importance Sampling
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1177967c7957072da3dc1db4ceb30e7a-Abstract.html
[ "Sebastian U Stich", "Anant Raj", "Martin Jaggi" ]
null
null
Importance sampling has become an indispensable strategy to speed up optimization algorithms for large-scale applications. Improved adaptive variants -- using importance values defined by the complete gradient information which changes during optimization -- enjoy favorable theoretical properties, but are typically com...
[]
null
50
1711.02637
title_snapshot
[ -0.05521786957979202, -0.02140357904136181, 0.02732032909989357, 0.04574398696422577, 0.04955553635954857, 0.03529023006558418, 0.006287197582423687, -0.030403632670640945, -0.016949711367487907, -0.05685839056968689, -0.02362123876810074, -0.02415776439011097, -0.06904225051403046, -0.005...
Introspective Classification with Convolutional Nets
https://proceedings.neurips.cc/paper_files/paper/2017/hash/11b921ef080f7736089c757404650e40-Abstract.html
[ "Long Jin", "Justin Lazarow", "Zhuowen Tu" ]
null
null
We propose introspective convolutional networks (ICN) that emphasize the importance of having convolutional neural networks empowered with generative capabilities. We employ a reclassification-by-synthesis algorithm to perform training using a formulation stemmed from the Bayes theory. Our ICN tries to iteratively: (1)...
[]
null
51
1704.07816
title_snapshot
[ 0.0049963402561843395, -0.02638593688607216, -0.0076028648763895035, 0.050138361752033234, 0.0312323160469532, 0.02153773605823517, 0.01683785580098629, 0.01340976357460022, -0.024027224630117416, -0.046566009521484375, -0.02147739939391613, -0.008293475955724716, -0.05291364714503288, 0.0...
Hybrid Reward Architecture for Reinforcement Learning
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1264a061d82a2edae1574b07249800d6-Abstract.html
[ "Harm Van Seijen", "Mehdi Fatemi", "Joshua Romoff", "Romain Laroche", "Tavian Barnes", "Jeffrey Tsang" ]
null
null
One of the main challenges in reinforcement learning (RL) is generalisation. In typical deep RL methods this is achieved by approximating the optimal value function with a low-dimensional representation using a deep network. While this approach works well in many domains, in domains where the optimal value function can...
[]
null
52
1706.04208
title_snapshot
[ -0.021611828356981277, 0.002475789748132229, -0.018463313579559326, 0.05651290714740753, 0.027653658762574196, 0.03116653487086296, -0.009917800314724445, -0.029517678543925285, -0.020067358389496803, -0.029268180951476097, -0.02080480009317398, 0.01235959492623806, -0.07103652507066727, -...
When Worlds Collide: Integrating Different Counterfactual Assumptions in Fairness
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1271a7029c9df08643b631b02cf9e116-Abstract.html
[ "Chris Russell", "Matt J Kusner", "Joshua Loftus", "Ricardo Silva" ]
null
null
Machine learning is now being used to make crucial decisions about people's lives. For nearly all of these decisions there is a risk that individuals of a certain race, gender, sexual orientation, or any other subpopulation are unfairly discriminated against. Our recent method has demonstrated how to use techniques fro...
[]
null
53
null
null
[ -0.03410389646887779, -0.011595458723604679, -0.027666328474879265, 0.030017271637916565, 0.02847408689558506, 0.023035846650600433, 0.019390525296330452, 0.023310862481594086, -0.038127753883600235, -0.02588922716677189, -0.012643789872527122, 0.03781870752573013, -0.09184671938419342, -0...
Dualing GANs
https://proceedings.neurips.cc/paper_files/paper/2017/hash/12a1d073d5ed3fa12169c67c4e2ce415-Abstract.html
[ "Yujia Li", "Alexander Schwing", "Kuan-Chieh Wang", "Richard Zemel" ]
null
null
Generative adversarial nets (GANs) are a promising technique for modeling a distribution from samples. It is however well known that GAN training suffers from instability due to the nature of its saddle point formulation. In this paper, we explore ways to tackle the instability problem by dualizing the discriminator. W...
[]
null
54
1706.06216
title_snapshot
[ -0.004853933583945036, -0.03331265598535538, -0.02874760329723358, 0.032445795834064484, 0.01841125078499317, 0.03811023384332657, -0.014842091128230095, -0.01256741862744093, -0.01962011680006981, -0.08198543637990952, -0.010614967904984951, -0.010645466856658459, -0.06610220670700073, -0...
A Universal Analysis of Large-Scale Regularized Least Squares Solutions
https://proceedings.neurips.cc/paper_files/paper/2017/hash/136f951362dab62e64eb8e841183c2a9-Abstract.html
[ "Ashkan Panahi", "Babak Hassibi" ]
null
null
A problem that has been of recent interest in statistical inference, machine learning and signal processing is that of understanding the asymptotic behavior of regularized least squares solutions under random measurement matrices (or dictionaries). The Least Absolute Shrinkage and Selection Operator (LASSO or least-squ...
[]
null
55
null
null
[ -0.04157949984073639, -0.011514498852193356, 0.04040856286883354, -0.012395654805004597, 0.03654104843735695, 0.050256967544555664, 0.025398068130016327, 0.012222719378769398, -0.0537337027490139, -0.03746479004621506, -0.03885161876678467, -0.0029217032715678215, -0.08115529268980026, -0....
Diffusion Approximations for Online Principal Component Estimation and Global Convergence
https://proceedings.neurips.cc/paper_files/paper/2017/hash/13f3cf8c531952d72e5847c4183e6910-Abstract.html
[ "Chris Junchi Li", "Mengdi Wang", "Han Liu", "Tong Zhang" ]
null
null
In this paper, we propose to adopt the diffusion approximation tools to study the dynamics of Oja's iteration which is an online stochastic gradient method for the principal component analysis. Oja's iteration maintains a running estimate of the true principal component from streaming data and enjoys less temporal and ...
[]
null
56
1808.09645
title_snapshot
[ -0.032263800501823425, -0.010448089800775051, 0.00743957469239831, 0.020412836223840714, 0.027314431965351105, 0.052564460784196854, 0.040494464337825775, 0.00826468225568533, -0.01240154355764389, -0.050657372921705246, 0.011371220462024212, -0.050876591354608536, -0.07685057073831558, -0...
k-Support and Ordered Weighted Sparsity for Overlapping Groups: Hardness and Algorithms
https://proceedings.neurips.cc/paper_files/paper/2017/hash/13fe9d84310e77f13a6d184dbf1232f3-Abstract.html
[ "Cong Han Lim", "Stephen Wright" ]
null
null
The k-support and OWL norms generalize the l1 norm, providing better prediction accuracy and better handling of correlated variables. We study the norms obtained from extending the k-support norm and OWL norms to the setting in which there are overlapping groups. The resulting norms are in general NP-hard to compute, b...
[]
null
57
null
null
[ -0.0025034865830093622, -0.013831016607582569, 0.023500079289078712, 0.00807881634682417, 0.03672225773334503, 0.0005462346016429365, 0.01757652312517166, 0.00011437728244345635, -0.01649560034275055, -0.007727011572569609, -0.033299122005701065, 0.004416345153003931, -0.07596374303102493, ...
Learning to Model the Tail
https://proceedings.neurips.cc/paper_files/paper/2017/hash/147ebe637038ca50a1265abac8dea181-Abstract.html
[ "Yu-Xiong Wang", "Deva Ramanan", "Martial Hebert" ]
null
null
We describe an approach to learning from long-tailed, imbalanced datasets that are prevalent in real-world settings. Here, the challenge is to learn accurate "few-shot'' models for classes in the tail of the class distribution, for which little data is available. We cast this problem as transfer learning, where knowled...
[]
null
58
null
null
[ 0.019018713384866714, -0.040064338594675064, -0.006133377552032471, 0.029949979856610298, 0.05392247810959816, 0.0009128038655035198, 0.02546992152929306, -0.010272321291267872, -0.009426428005099297, -0.01374766230583191, -0.011186361312866211, 0.018023446202278137, -0.05861654505133629, ...
Neural Variational Inference and Learning in Undirected Graphical Models
https://proceedings.neurips.cc/paper_files/paper/2017/hash/14e422f05b68cc0139988e128ee880df-Abstract.html
[ "Volodymyr Kuleshov", "Stefano Ermon" ]
null
null
Many problems in machine learning are naturally expressed in the language of undirected graphical models. Here, we propose black-box learning and inference algorithms for undirected models that optimize a variational approximation to the log-likelihood of the model. Central to our approach is an upper bound on the log-...
[]
null
59
1711.02679
title_snapshot
[ -0.01689336448907852, 0.00617222860455513, -0.021236799657344818, 0.04461849108338356, 0.045501213520765305, 0.032849591225385666, 0.03650449216365814, 0.0346982441842556, -0.021177520975470543, -0.024153152480721474, 0.0008577011176384985, 0.030587168410420418, -0.07332909107208252, 0.003...
Aggressive Sampling for Multi-class to Binary Reduction with Applications to Text Classification
https://proceedings.neurips.cc/paper_files/paper/2017/hash/14ea0d5b0cf49525d1866cb1e95ada5d-Abstract.html
[ "Bikash Joshi", "Massih R. Amini", "Ioannis Partalas", "Franck Iutzeler", "Yury Maximov" ]
null
null
We address the problem of multi-class classification in the case where the number of classes is very large. We propose a double sampling strategy on top of a multi-class to binary reduction strategy, which transforms the original multi-class problem into a binary classification problem over pairs of examples. The aim o...
[]
null
60
1701.06511
title_snapshot
[ -0.03345850855112076, -0.048522356897592545, -0.0050077675841748714, 0.055042412132024765, 0.03283108025789261, 0.033864784985780716, 0.003445324022322893, -0.012850931845605373, -0.0017182856099680066, -0.026276450604200363, -0.011552984826266766, 0.011991693638265133, -0.08454914391040802,...
Learning Linear Dynamical Systems via Spectral Filtering
https://proceedings.neurips.cc/paper_files/paper/2017/hash/165a59f7cf3b5c4396ba65953d679f17-Abstract.html
[ "Elad Hazan", "Karan Singh", "Cyril Zhang" ]
null
null
We present an efficient and practical algorithm for the online prediction of discrete-time linear dynamical systems with a symmetric transition matrix. We circumvent the non-convex optimization problem using improper learning: carefully overparameterize the class of LDSs by a polylogarithmic factor, in exchange for con...
[]
null
61
1711.00946
title_snapshot
[ -0.03772835433483124, 0.0013906682142987847, 0.00870652124285698, 0.03401032090187073, 0.03796834126114845, 0.019402233883738518, 0.03552112355828285, 0.016183944419026375, -0.030381381511688232, -0.04820685461163521, 0.024951811879873276, 0.0030563543550670147, -0.0714067816734314, 0.0082...
Efficient Modeling of Latent Information in Supervised Learning using Gaussian Processes
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1680e9fa7b4dd5d62ece800239bb53bd-Abstract.html
[ "Zhenwen Dai", "Mauricio Álvarez", "Neil Lawrence" ]
null
null
Often in machine learning, data are collected as a combination of multiple conditions, e.g., the voice recordings of multiple persons, each labeled with an ID. How could we build a model that captures the latent information related to these conditions and generalize to a new one with few data? We present a new model ca...
[]
null
62
1705.09862
title_snapshot
[ 0.0010779728181660175, -0.013335325755178928, -0.01581844873726368, 0.038966525346040726, 0.0344952754676342, 0.03144896402955055, 0.024708926677703857, 0.0031089955009520054, -0.03579763323068619, -0.028038494288921356, -0.0079413503408432, 0.03242218494415283, -0.07323440164327621, 0.007...
Semi-Supervised Learning for Optical Flow with Generative Adversarial Networks
https://proceedings.neurips.cc/paper_files/paper/2017/hash/16a5cdae362b8d27a1d8f8c7b78b4330-Abstract.html
[ "Wei-Sheng Lai", "Jia-Bin Huang", "Ming-Hsuan Yang" ]
null
null
Convolutional neural networks (CNNs) have recently been applied to the optical flow estimation problem. As training the CNNs requires sufficiently large ground truth training data, existing approaches resort to synthetic, unrealistic datasets. On the other hand, unsupervised methods are capable of leveraging real-world...
[]
null
63
null
null
[ 0.0312521830201149, -0.045193981379270554, -0.000871291384100914, 0.056744284927845, 0.015410313382744789, 0.024370364844799042, 0.014214720577001572, 0.00023090698232408613, -0.02007514424622059, -0.041647009551525116, -0.029813993722200394, -0.031168002635240555, -0.06588953733444214, 0....
Phase Transitions in the Pooled Data Problem
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1700002963a49da13542e0726b7bb758-Abstract.html
[ "Jonathan Scarlett", "Volkan Cevher" ]
null
null
In this paper, we study the {\em pooled data} problem of identifying the labels associated with a large collection of items, based on a sequence of pooled tests revealing the counts of each label within the pool. In the noiseless setting, we identify an exact asymptotic threshold on the required number of tests with op...
[]
null
64
1710.06766
title_snapshot
[ 0.006702929735183716, -0.030687378719449043, -0.004121363162994385, 0.02641831710934639, 0.031955260783433914, 0.0048008146695792675, -0.0003354961227159947, 0.02460206300020218, -0.04309852048754692, -0.01957108825445175, -0.011080017313361168, -0.0026410501450300217, -0.058171145617961884,...
Unifying PAC and Regret: Uniform PAC Bounds for Episodic Reinforcement Learning
https://proceedings.neurips.cc/paper_files/paper/2017/hash/17d8da815fa21c57af9829fb0a869602-Abstract.html
[ "Christoph Dann", "Tor Lattimore", "Emma Brunskill" ]
null
null
Statistical performance bounds for reinforcement learning (RL) algorithms can be critical for high-stakes applications like healthcare. This paper introduces a new framework for theoretically measuring the performance of such algorithms called Uniform-PAC, which is a strengthening of the classical Probably Approximatel...
[]
null
65
1703.07710
title_snapshot
[ -0.04870407283306122, -0.023564593866467476, -0.03476612642407417, 0.056119393557310104, 0.04984964057803154, 0.021828612312674522, 0.02150299958884716, -0.0037863494362682104, -0.019763968884944916, -0.04724164307117462, -0.009825537912547588, -0.010802269913256168, -0.06646496057510376, ...
Stein Variational Gradient Descent as Gradient Flow
https://proceedings.neurips.cc/paper_files/paper/2017/hash/17ed8abedc255908be746d245e50263a-Abstract.html
[ "Qiang Liu" ]
null
null
Stein variational gradient descent (SVGD) is a deterministic sampling algorithm that iteratively transports a set of particles to approximate given distributions, based on a gradient-based update constructed to optimally decrease the KL divergence within a function space. This paper develops the first theoretical analy...
[]
null
66
1704.07520
title_snapshot
[ -0.03249121457338333, -0.0029550122562795877, 0.04184907674789429, 0.021573256701231003, 0.030468042939901352, 0.009937184862792492, 0.01652507483959198, 0.006936032325029373, -0.03467969968914986, -0.04967060685157776, -0.004776857793331146, -0.018203070387244225, -0.044727638363838196, 0...
Expectation Propagation for t-Exponential Family Using q-Algebra
https://proceedings.neurips.cc/paper_files/paper/2017/hash/17fafe5f6ce2f1904eb09d2e80a4cbf6-Abstract.html
[ "Futoshi Futami", "Issei Sato", "Masashi Sugiyama" ]
null
null
Exponential family distributions are highly useful in machine learning since their calculation can be performed efficiently through natural parameters. The exponential family has recently been extended to the t-exponential family, which contains Student-t distributions as family members and thus allows us to handle noi...
[]
null
67
1705.09046
title_snapshot
[ 0.0011434988118708134, 0.022603116929531097, -0.007823185063898563, 0.010565526783466339, 0.06670691072940826, 0.027496907860040665, 0.0017909680027514696, -0.016644330695271492, -0.01422299724072218, -0.03632477670907974, -0.004815944004803896, 0.024659691378474236, -0.0646916851401329, 0...
Collaborative PAC Learning
https://proceedings.neurips.cc/paper_files/paper/2017/hash/186a157b2992e7daed3677ce8e9fe40f-Abstract.html
[ "Avrim Blum", "Nika Haghtalab", "Ariel D Procaccia", "Mingda Qiao" ]
null
null
We introduce a collaborative PAC learning model, in which k players attempt to learn the same underlying concept. We ask how much more information is required to learn an accurate classifier for all players simultaneously. We refer to the ratio between the sample complexity of collaborative PAC learning and its non-col...
[]
null
68
null
null
[ -0.0007611245382577181, -0.037073127925395966, -0.011328429915010929, 0.03881717473268509, 0.034259796142578125, 0.015454530715942383, 0.009307421743869781, -0.002601499669253826, -0.024263625964522362, -0.021885372698307037, 0.014002581126987934, 0.0025789726059883833, -0.05575234815478325,...
Polynomial time algorithms for dual volume sampling
https://proceedings.neurips.cc/paper_files/paper/2017/hash/18bb68e2b38e4a8ce7cf4f6b2625768c-Abstract.html
[ "Chengtao Li", "Stefanie Jegelka", "Suvrit Sra" ]
null
null
We study dual volume sampling, a method for selecting k columns from an n*m short and wide matrix (n <= k <= m) such that the probability of selection is proportional to the volume spanned by the rows of the induced submatrix. This method was proposed by Avron and Boutsidis (2013), who showed it to be a promising metho...
[]
null
69
1703.02674
title_snapshot
[ -0.01815069653093815, -0.020911358296871185, -0.014687689021229744, 0.02146376669406891, 0.035495586693286896, 0.02767164260149002, -0.0005921213887631893, -0.02047663740813732, -0.014826717786490917, -0.06330994516611099, 0.005539989564567804, -0.029512641951441765, -0.04582495614886284, ...
Premise Selection for Theorem Proving by Deep Graph Embedding
https://proceedings.neurips.cc/paper_files/paper/2017/hash/18d10dc6e666eab6de9215ae5b3d54df-Abstract.html
[ "Mingzhe Wang", "Yihe Tang", "Jian Wang", "Jia Deng" ]
null
null
We propose a deep learning-based approach to the problem of premise selection: selecting mathematical statements relevant for proving a given conjecture. We represent a higher-order logic formula as a graph that is invariant to variable renaming but still fully preserves syntactic and semantic information. We then embe...
[]
null
70
1709.09994
title_snapshot
[ -0.03001530095934868, 0.016207778826355934, 0.0005587003543041646, 0.06056658551096916, 0.04331168904900551, 0.03929680213332176, 0.02111980877816677, -0.020206086337566376, 0.013233548030257225, -0.012053954415023327, 0.020688418298959732, 0.028466369956731796, -0.0508292019367218, -0.002...
Differentiable Learning of Submodular Models
https://proceedings.neurips.cc/paper_files/paper/2017/hash/192fc044e74dffea144f9ac5dc9f3395-Abstract.html
[ "Josip Djolonga", "Andreas Krause" ]
null
null
Can we incorporate discrete optimization algorithms within modern machine learning models? For example, is it possible to use in deep architectures a layer whose output is the minimal cut of a parametrized graph? Given that these models are trained end-to-end by leveraging gradient information, the introduction of such...
[]
null
71
null
null
[ -0.028394244611263275, -0.013953447341918945, 0.005381554365158081, 0.03271247819066048, 0.056414391845464706, 0.036064326763153076, 0.04280296713113785, -0.007691207807511091, -0.004546359647065401, -0.03062022291123867, 0.0012814634246751666, 0.013911343179643154, -0.059254374355077744, ...
YASS: Yet Another Spike Sorter
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1943102704f8f8f3302c2b730728e023-Abstract.html
[ "Jin Hyung Lee", "David E Carlson", "Hooshmand Shokri Razaghi", "Weichi Yao", "Georges A Goetz", "Espen Hagen", "Eleanor Batty", "E. J. Chichilnisky", "Gaute T. Einevoll", "Liam Paninski" ]
null
null
Spike sorting is a critical first step in extracting neural signals from large-scale electrophysiological data. This manuscript describes an efficient, reliable pipeline for spike sorting on dense multi-electrode arrays (MEAs), where neural signals appear across many electrodes and spike sorting currently represents a ...
[]
null
72
null
null
[ -0.025860294699668884, 0.002641682978719473, -0.010876500979065895, 0.017505086958408356, 0.02576538734138012, 0.0422273650765419, 0.01948794163763523, -0.006811683066189289, -0.07078187167644501, -0.048877306282520294, 0.009415476582944393, 0.0043701729737222195, -0.04890286922454834, 0.0...
Variational Laws of Visual Attention for Dynamic Scenes
https://proceedings.neurips.cc/paper_files/paper/2017/hash/194cf6c2de8e00c05fcf16c498adc7bf-Abstract.html
[ "Dario Zanca", "Marco Gori" ]
null
null
Computational models of visual attention are at the crossroad of disciplines like cognitive science, computational neuroscience, and computer vision. This paper proposes a model of attentional scanpath that is based on the principle that there are foundational laws that drive the emergence of visual attention. We devis...
[]
null
73
null
null
[ -0.008755236864089966, 0.028667189180850983, 0.020845212042331696, -0.012765209190547466, 0.029576800763607025, 0.02045426331460476, 0.027051163837313652, 0.051998320966959, -0.05531151220202446, -0.02005769871175289, -0.03219545632600784, 0.02497963048517704, -0.06890838593244553, 0.00534...
How regularization affects the critical points in linear networks
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1abb1e1ea5f481b589da52303b091cbb-Abstract.html
[ "Amirhossein Taghvaei", "Jin W Kim", "Prashant Mehta" ]
null
null
This paper is concerned with the problem of representing and learning a linear transformation using a linear neural network. In recent years, there is a growing interest in the study of such networks, in part due to the successes of deep learning. The main question of this body of research (and also of our paper) is re...
[]
null
74
1709.09625
title_snapshot
[ -0.059709902852773666, -0.01713608019053936, 0.0011349115520715714, -0.006401808001101017, 0.049042169004678726, 0.027077220380306244, 0.009336696937680244, -0.000981782446615398, -0.046429283916950226, -0.0509912334382534, -0.0010766461491584778, -0.014777980744838715, -0.056685615330934525...
On Tensor Train Rank Minimization : Statistical Efficiency and Scalable Algorithm
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1b5230e3ea6d7123847ad55a1e06fffd-Abstract.html
[ "Masaaki Imaizumi", "Takanori Maehara", "Kohei Hayashi" ]
null
null
Tensor train (TT) decomposition provides a space-efficient representation for higher-order tensors. Despite its advantage, we face two crucial limitations when we apply the TT decomposition to machine learning problems: the lack of statistical theory and of scalable algorithms. In this paper, we address the limitations...
[]
null
75
1708.00132
title_snapshot
[ -0.030211949720978737, -0.023763667792081833, -0.005847540218383074, 0.03744039684534073, 0.017332080751657486, 0.032652828842401505, 0.022244470193982124, -0.02094338834285736, -0.027331797406077385, -0.04794842377305031, -0.055995140224695206, 0.0330086313188076, -0.07017869502305984, 0....
EX2: Exploration with Exemplar Models for Deep Reinforcement Learning
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1baff70e2669e8376347efd3a874a341-Abstract.html
[ "Justin Fu", "John Co-Reyes", "Sergey Levine" ]
null
null
Deep reinforcement learning algorithms have been shown to learn complex tasks using highly general policy classes. However, sparse reward problems remain a significant challenge. Exploration methods based on novelty detection have been particularly successful in such settings but typically require generative or predict...
[]
null
76
1703.01260
title_snapshot
[ 0.006838809233158827, -0.01701510138809681, 0.0041979094967246056, 0.056533247232437134, 0.047976695001125336, 0.009803660213947296, 0.010750805027782917, -0.018912630155682564, -0.0425044409930706, -0.05885500833392143, -0.03487712889909744, 0.02255038172006607, -0.04943331703543663, -0.0...
Training Quantized Nets: A Deeper Understanding
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1c303b0eed3133200cf715285011b4e4-Abstract.html
[ "Hao Li", "Soham De", "Zheng Xu", "Christoph Studer", "Hanan Samet", "Tom Goldstein" ]
null
null
Currently, deep neural networks are deployed on low-power portable devices by first training a full-precision model using powerful hardware, and then deriving a corresponding low-precision model for efficient inference on such systems. However, training models directly with coarsely quantized weights is a key step towa...
[]
null
77
1706.02379
title_snapshot
[ -0.0360121950507164, -0.03448500111699104, -0.04632936790585518, 0.05640043318271637, 0.04878440126776695, 0.04212815687060356, 0.003253005212172866, 0.0007871054112911224, -0.02178865484893322, -0.02372635528445244, -0.003692458150908351, -0.008983331732451916, -0.06257640570402145, -0.00...
Convolutional Gaussian Processes
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1c54985e4f95b7819ca0357c0cb9a09f-Abstract.html
[ "Mark van der Wilk", "Carl Edward Rasmussen", "James Hensman" ]
null
null
We present a practical way of introducing convolutional structure into Gaussian processes, making them more suited to high-dimensional inputs like images. The main contribution of our work is the construction of an inter-domain inducing point approximation that is well-tailored to the convolutional kernel. This allows ...
[]
null
78
1709.01894
title_snapshot
[ 0.01749282144010067, -0.023416882380843163, 0.027287377044558525, 0.03249159827828407, 0.00044783216435462236, 0.04949511960148811, 0.0013495675520971417, 0.008288679644465446, 0.005875588860362768, -0.0403696671128273, -0.013302392326295376, 0.005546897184103727, -0.06651301681995392, 0.0...
Best Response Regression
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1ce927f875864094e3906a4a0b5ece68-Abstract.html
[ "Omer Ben-Porat", "Moshe Tennenholtz" ]
null
null
In a regression task, a predictor is given a set of instances, along with a real value for each point. Subsequently, she has to identify the value of a new instance as accurately as possible. In this work, we initiate the study of strategic predictions in machine learning. We consider a regression task tackled by two p...
[]
null
79
null
null
[ -0.037942659109830856, -0.02057323418557644, 0.017805829644203186, 0.016793491318821907, 0.05372070148587227, 0.030504120513796806, 0.012076680548489094, -0.01823459006845951, -0.042401272803545, -0.021367313340306282, -0.023168623447418213, 0.04691773280501366, -0.06377089023590088, -0.00...
Elementary Symmetric Polynomials for Optimal Experimental Design
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1cecc7a77928ca8133fa24680a88d2f9-Abstract.html
[ "Zelda E. Mariet", "Suvrit Sra" ]
null
null
We revisit the classical problem of optimal experimental design (OED) under a new mathematical model grounded in a geometric motivation. Specifically, we introduce models based on elementary symmetric polynomials; these polynomials capture "partial volumes" and offer a graded interpolation between the widely used A-opt...
[]
null
80
1705.09677
title_snapshot
[ -0.034972380846738815, 0.03143057972192764, -0.012059918604791164, 0.05387358367443085, 0.028295855969190598, 0.01591571606695652, 0.01283391285687685, -0.028350962325930595, -0.007915833964943886, -0.054652296006679535, -0.013974091038107872, -0.019139979034662247, -0.057059530168771744, ...
Learning from Complementary Labels
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1dba5eed8838571e1c80af145184e515-Abstract.html
[ "Takashi Ishida", "Gang Niu", "Weihua Hu", "Masashi Sugiyama" ]
null
null
Collecting labeled data is costly and thus a critical bottleneck in real-world classification tasks. To mitigate this problem, we propose a novel setting, namely learning from complementary labels for multi-class classification. A complementary label specifies a class that a pattern does not belong to. Collecting compl...
[]
null
81
1705.07541
title_snapshot
[ 0.009242605417966843, -0.025819502770900726, -0.03994560241699219, 0.0431586317718029, 0.008054882287979126, 0.012873093597590923, 0.011462853290140629, -0.037706926465034485, -0.01980637013912201, -0.03200732544064522, -0.010115444660186768, 0.02476775273680687, -0.08088039606809616, 0.00...
Dynamic Importance Sampling for Anytime Bounds of the Partition Function
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1f1baa5b8edac74eb4eaa329f14a0361-Abstract.html
[ "Qi Lou", "Rina Dechter", "Alex Ihler" ]
null
null
Computing the partition function is a key inference task in many graphical models. In this paper, we propose a dynamic importance sampling scheme that provides anytime finite-sample bounds for the partition function. Our algorithm balances the advantages of the three major inference strategies, heuristic search, variat...
[]
null
82
null
null
[ -0.047580536454916, 0.0010182303376495838, -0.028182577341794968, 0.06752292066812515, 0.0339050367474556, 0.028214402496814728, 0.018486378714442253, 0.00032085200655274093, -0.00486037228256464, -0.05172158405184746, 0.018306121230125427, -0.029397429898381233, -0.07144849747419357, -0.0...
Process-constrained batch Bayesian optimisation
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1f71e393b3809197ed66df836fe833e5-Abstract.html
[ "Pratibha Vellanki", "Santu Rana", "Sunil Gupta", "David Rubin", "Alessandra Sutti", "Thomas Dorin", "Murray Height", "Paul Sanders", "Svetha Venkatesh" ]
null
null
Abstract Prevailing batch Bayesian optimisation methods allow all control variables to be freely altered at each iteration. Real-world experiments, however, often have physical limitations making it time-consuming to alter all settings for each recommendation in a batch. This gives rise to a unique problem in BO: in a ...
[]
null
83
null
null
[ -0.008644724264740944, 0.004568067844957113, 0.009908825159072876, 0.01902291551232338, 0.03032897785305977, 0.06697873026132584, 0.01124562043696642, 0.0006577976164408028, -0.01978854089975357, -0.033615242689847946, -0.023967474699020386, -0.020577743649482727, -0.04320525750517845, 0.0...
Uprooting and Rerooting Higher-Order Graphical Models
https://proceedings.neurips.cc/paper_files/paper/2017/hash/1ff8a7b5dc7a7d1f0ed65aaa29c04b1e-Abstract.html
[ "Mark Rowland", "Adrian Weller" ]
null
null
The idea of uprooting and rerooting graphical models was introduced specifically for binary pairwise models by Weller (2016) as a way to transform a model to any of a whole equivalence class of related models, such that inference on any one model yields inference results for all others. This is very helpful since infer...
[]
null
84
null
null
[ -0.024481182917952538, 0.005092575680464506, -0.016484197229146957, 0.04125206172466278, 0.04158193618059158, 0.02133607678115368, 0.033890362828969955, -0.0023325507063418627, -0.009861543774604797, -0.03824777901172638, 0.011836985126137733, -0.0006008947384543717, -0.10481873899698257, ...
Learned in Translation: Contextualized Word Vectors
https://proceedings.neurips.cc/paper_files/paper/2017/hash/20c86a628232a67e7bd46f76fba7ce12-Abstract.html
[ "Bryan McCann", "James Bradbury", "Caiming Xiong", "Richard Socher" ]
null
null
Computer vision has benefited from initializing multiple deep layers with weights pretrained on large supervised training sets like ImageNet. Natural language processing (NLP) typically sees initialization of only the lowest layer of deep models with pretrained word vectors. In this paper, we use a deep LSTM encoder fr...
[]
null
85
1708.00107
title_snapshot
[ 0.036277275532484055, -0.01188304740935564, -0.0016229236498475075, 0.06820763647556305, 0.01961432583630085, 0.03780033811926842, 0.011804832145571709, 0.028544416651129723, -0.01745699532330036, -0.024687496945261955, -0.03905460983514786, 0.039520613849163055, -0.03843460604548454, -0.0...
Semisupervised Clustering, AND-Queries and Locally Encodable Source Coding
https://proceedings.neurips.cc/paper_files/paper/2017/hash/2131f8ecf18db66a758f718dc729e00e-Abstract.html
[ "Arya Mazumdar", "Soumyabrata Pal" ]
null
null
Source coding is the canonical problem of data compression in information theory. In a locally encodable source coding, each compressed bit depends on only few bits of the input. In this paper, we show that a recently popular model of semisupervised clustering is equivalent to locally encodable source coding. In this m...
[]
null
86
1904.00507
title_judge
[ -0.012994556687772274, -0.04930146411061287, -0.038109228014945984, 0.05444488301873207, 0.04339428246021271, 0.018436815589666367, 0.007926441729068756, -0.00200319429859519, -0.01103934459388256, -0.03620724752545357, -0.045150063931941986, -0.0006777584203518927, -0.04406466707587242, 0...
Regularizing Deep Neural Networks by Noise: Its Interpretation and Optimization
https://proceedings.neurips.cc/paper_files/paper/2017/hash/217e342fc01668b10cb1188d40d3370e-Abstract.html
[ "Hyeonwoo Noh", "Tackgeun You", "Jonghwan Mun", "Bohyung Han" ]
null
null
Overfitting is one of the most critical challenges in deep neural networks, and there are various types of regularization methods to improve generalization performance. Injecting noises to hidden units during training, e.g., dropout, is known as a successful regularizer, but it is still not clear enough why such traini...
[]
null
87
1710.05179
title_snapshot
[ -0.007397559005767107, -0.011179260909557343, -0.025721540674567223, 0.04388224333524704, 0.05611584708094597, 0.050381701439619064, 0.022356374189257622, 0.0004040254279971123, -0.03533478453755379, -0.05321813002228737, -0.014475819654762745, 0.0038897425401955843, -0.0523437075316906, -...
Few-Shot Adversarial Domain Adaptation
https://proceedings.neurips.cc/paper_files/paper/2017/hash/21c5bba1dd6aed9ab48c2b34c1a0adde-Abstract.html
[ "Saeid Motiian", "Quinn Jones", "Seyed Iranmanesh", "Gianfranco Doretto" ]
null
null
This work provides a framework for addressing the problem of supervised domain adaptation with deep models. The main idea is to exploit adversarial learning to learn an embedded subspace that simultaneously maximizes the confusion between two domains while semantically aligning their embedding. The supervised setting b...
[]
null
88
1711.02536
title_snapshot
[ 0.0031018634326756, -0.03421732783317566, -0.017350995913147926, 0.05559510365128517, 0.03787876293063164, 0.021135548129677773, 0.04274217411875725, -0.011605565436184406, -0.0185258649289608, -0.014889138750731945, -0.039464935660362244, 0.021281031891703606, -0.06563020497560501, 0.0075...
Robust and Efficient Transfer Learning with Hidden Parameter Markov Decision Processes
https://proceedings.neurips.cc/paper_files/paper/2017/hash/2227d753dc18505031869d44673728e2-Abstract.html
[ "Taylor W Killian", "Samuel Daulton", "George Konidaris", "Finale Doshi-Velez" ]
null
null
We introduce a new formulation of the Hidden Parameter Markov Decision Process (HiP-MDP), a framework for modeling families of related tasks using low-dimensional latent embeddings. Our new framework correctly models the joint uncertainty in the latent parameters and the state space. We also replace the original Gaussi...
[]
null
89
1706.06544
title_snapshot
[ -0.045195598155260086, -0.0004147151776123792, -0.028536615893244743, 0.035562772303819656, 0.04894251376390457, 0.025878991931676865, 0.030541641637682915, -0.011514192447066307, -0.002863965230062604, -0.05472877621650696, -0.00905850064009428, 0.015373640693724155, -0.061056844890117645, ...
Multi-View Decision Processes: The Helper-AI Problem
https://proceedings.neurips.cc/paper_files/paper/2017/hash/227f6afd3b7f89b96c4bb91f95d50f6d-Abstract.html
[ "Christos Dimitrakakis", "David C. Parkes", "Goran Radanovic", "Paul Tylkin" ]
null
null
We consider a two-player sequential game in which agents have the same reward function but may disagree on the transition probabilities of an underlying Markovian model of the world. By committing to play a specific policy, the agent with the correct model can steer the behavior of the other agent, and seek to improve ...
[]
null
90
null
null
[ -0.047858208417892456, 0.010050926357507706, -0.014261951670050621, 0.03957386687397957, 0.030539393424987793, 0.016750270500779152, 0.006791544146835804, 0.014495576731860638, -0.009088587015867233, -0.044657036662101746, -0.033421557396650314, 0.032748933881521225, -0.08268715441226959, ...
Maximum Margin Interval Trees
https://proceedings.neurips.cc/paper_files/paper/2017/hash/2288f691b58edecadcc9a8691762b4fd-Abstract.html
[ "Alexandre Drouin", "Toby Hocking", "Francois Laviolette" ]
null
null
Learning a regression function using censored or interval-valued output data is an important problem in fields such as genomics and medicine. The goal is to learn a real-valued prediction function, and the training output labels indicate an interval of possible values. Whereas most existing algorithms for this task are...
[]
null
91
1710.04234
title_snapshot
[ 0.0048472024500370026, 0.01777978241443634, -0.014166947454214096, 0.04360548406839371, 0.034495096653699875, 0.03128202632069588, 0.05397069454193115, -0.0048649790696799755, -0.025453416630625725, -0.01594495214521885, 0.01370332296937704, 0.03066832572221756, -0.07396421581506729, -0.00...
Online Learning with a Hint
https://proceedings.neurips.cc/paper_files/paper/2017/hash/22b1f2e0983160db6f7bb9f62f4dbb39-Abstract.html
[ "Ofer Dekel", "arthur flajolet", "Nika Haghtalab", "Patrick Jaillet" ]
null
null
We study a variant of online linear optimization where the player receives a hint about the loss function at the beginning of each round. The hint is given in the form of a vector that is weakly correlated with the loss vector on that round. We show that the player can benefit from such a hint if the set of feasible ac...
[]
null
92
null
null
[ -0.03254365921020508, -0.0016764046158641577, -0.00851589348167181, 0.042352475225925446, 0.014275718480348587, 0.04264902323484421, 0.0016883151838555932, 0.0154265146702528, -0.02056235261261463, -0.017091473564505577, -0.03080696240067482, -0.005718918517231941, -0.06579617410898209, -0...
DPSCREEN: Dynamic Personalized Screening
https://proceedings.neurips.cc/paper_files/paper/2017/hash/22fb0cee7e1f3bde58293de743871417-Abstract.html
[ "Kartik Ahuja", "William Zame", "Mihaela van der Schaar" ]
null
null
Screening is important for the diagnosis and treatment of a wide variety of diseases. A good screening policy should be personalized to the disease, to the features of the patient and to the dynamic history of the patient (including the history of screening). The growth of electronic health records data has led to the ...
[]
null
93
null
null
[ -0.016683459281921387, -0.004408886656165123, -0.026383938267827034, 0.007524615153670311, 0.08531346172094345, 0.04011088237166405, 0.008967908099293709, -0.025377729907631874, -0.0474691241979599, -0.06503503769636154, 0.022599374875426292, -0.022989995777606964, -0.03951467573642731, 0....
Online Learning of Optimal Bidding Strategy in Repeated Multi-Commodity Auctions
https://proceedings.neurips.cc/paper_files/paper/2017/hash/23af4b45f1e166141a790d1a3126e77a-Abstract.html
[ "M. Sevi Baltaoglu", "Lang Tong", "Qing Zhao" ]
null
null
We study the online learning problem of a bidder who participates in repeated auctions. With the goal of maximizing his T-period payoff, the bidder determines the optimal allocation of his budget among his bids for $K$ goods at each period. As a bidding strategy, we propose a polynomial-time algorithm, inspired by the ...
[]
null
94
1703.02567
title_snapshot
[ -0.03566478192806244, -0.0207256730645895, -0.016796156764030457, 0.04694749414920807, 0.03313222527503967, 0.04316317290067673, -0.00632233452051878, 0.01868043839931488, -0.027416588738560677, -0.03106931410729885, 0.006763164419680834, 0.01837865822017193, -0.06389769166707993, 0.005851...
A-NICE-MC: Adversarial Training for MCMC
https://proceedings.neurips.cc/paper_files/paper/2017/hash/2417dc8af8570f274e6775d4d60496da-Abstract.html
[ "Jiaming Song", "Shengjia Zhao", "Stefano Ermon" ]
null
null
Existing Markov Chain Monte Carlo (MCMC) methods are either based on general-purpose and domain-agnostic schemes, which can lead to slow convergence, or require hand-crafting of problem-specific proposals by an expert. We propose A-NICE-MC, a novel method to train flexible parametric Markov chain kernels to produce sam...
[]
null
95
1706.07561
title_snapshot
[ -0.02445698156952858, -0.008757481351494789, 0.005524169187992811, 0.06677985191345215, 0.031293030828237534, 0.03492327779531479, 0.03511559218168259, -0.03169979900121689, -0.019207585602998734, -0.0495663620531559, -0.008624566718935966, 0.0021754945628345013, -0.06752786785364151, -0.0...
Question Asking as Program Generation
https://proceedings.neurips.cc/paper_files/paper/2017/hash/24681928425f5a9133504de568f5f6df-Abstract.html
[ "Anselm Rothe", "Brenden M Lake", "Todd Gureckis" ]
null
null
A hallmark of human intelligence is the ability to ask rich, creative, and revealing questions. Here we introduce a cognitive model capable of constructing human-like questions. Our approach treats questions as formal programs that, when executed on the state of the world, output an answer. The model specifies a probab...
[]
null
96
1711.06351
title_snapshot
[ -0.01449695136398077, -0.014827471226453781, -0.027172572910785675, 0.05854450911283493, 0.04010801762342453, 0.011997256428003311, 0.0132862264290452, 0.031078243628144264, -0.018416322767734528, -0.02770811878144741, -0.053898487240076065, 0.04052427411079407, -0.06203267350792885, -0.03...
Gradient Methods for Submodular Maximization
https://proceedings.neurips.cc/paper_files/paper/2017/hash/24b43fb034a10d78bec71274033b4096-Abstract.html
[ "Hamed Hassani", "Mahdi Soltanolkotabi", "Amin Karbasi" ]
null
null
In this paper, we study the problem of maximizing continuous submodular functions that naturally arise in many learning applications such as those involving utility functions in active learning and sensing, matrix approximations and network inference. Despite the apparent lack of convexity in such functions, we prove t...
[]
null
97
1708.03949
title_snapshot
[ -0.026558123528957367, -0.014150209724903107, 0.008221388794481754, 0.04597725346684456, 0.03826487809419632, 0.04871904477477074, 0.02287350967526436, -0.02315850742161274, -0.0017845716793090105, -0.032520391047000885, -0.0026806865353137255, 0.025845013558864594, -0.06799297779798508, -...
Recycling Privileged Learning and Distribution Matching for Fairness
https://proceedings.neurips.cc/paper_files/paper/2017/hash/250cf8b51c773f3f8dc8b4be867a9a02-Abstract.html
[ "Novi Quadrianto", "Viktoriia Sharmanska" ]
null
null
Equipping machine learning models with ethical and legal constraints is a serious issue; without this, the future of machine learning is at risk. This paper takes a step forward in this direction and focuses on ensuring machine learning models deliver fair decisions. In legal scholarships, the notion of fairness itself...
[]
null
98
null
null
[ -0.0315631628036499, -0.04858928173780441, -0.04945996403694153, 0.033574581146240234, 0.0414903499186039, 0.04470972716808319, -0.003712597768753767, -0.014926019124686718, -0.04260919615626335, -0.03577081486582756, -0.008266647346317768, 0.027173805981874466, -0.0806632712483406, -0.010...
Collecting Telemetry Data Privately
https://proceedings.neurips.cc/paper_files/paper/2017/hash/253614bbac999b38b5b60cae531c4969-Abstract.html
[ "Bolin Ding", "Janardhan Kulkarni", "Sergey Yekhanin" ]
null
null
The collection and analysis of telemetry data from user's devices is routinely performed by many software companies. Telemetry collection leads to improved user experience but poses significant risks to users' privacy. Locally differentially private (LDP) algorithms have recently emerged as the main tool that allows da...
[]
null
99
1712.01524
title_snapshot
[ 0.0350450836122036, 0.0008403286337852478, -0.001877287169918418, 0.03710182011127472, 0.06176071614027023, 0.013097656890749931, 0.03763965517282486, -0.03781639039516449, -0.011551842093467712, -0.04011458903551102, -0.004694915842264891, -0.03051057644188404, -0.0549398697912693, -0.020...
Parallel Streaming Wasserstein Barycenters
https://proceedings.neurips.cc/paper_files/paper/2017/hash/253f7b5d921338af34da817c00f42753-Abstract.html
[ "Matthew Staib", "Sebastian Claici", "Justin M Solomon", "Stefanie Jegelka" ]
null
null
Efficiently aggregating data from different sources is a challenging problem, particularly when samples from each source are distributed differently. These differences can be inherent to the inference task or present for other reasons: sensors in a sensor network may be placed far apart, affecting their individual meas...
[]
null
100
1705.07443
title_snapshot
[ -0.017642874270677567, -0.014572402462363243, 0.018400507047772408, 0.022910602390766144, 0.033431526273489, 0.030454905703663826, 0.015985438600182533, 0.010907294228672981, -0.036806490272283554, -0.06772121042013168, 0.008511831983923912, -0.03725549951195717, -0.056386083364486694, -0....