title
stringlengths
14
148
paper_url
stringlengths
105
105
authors
listlengths
0
21
type
stringclasses
0 values
primary_area
stringclasses
0 values
abstract
large_stringlengths
388
2k
keywords
listlengths
0
0
TL;DR
large_stringclasses
0 values
submission_number
int64
1
1.43k
arxiv_id
stringlengths
10
10
arxiv_id_source
stringclasses
2 values
embedding
listlengths
768
768
Compositional Plan Vectors
https://proceedings.neurips.cc/paper_files/paper/2019/hash/00989c20ff1386dc386d8124ebcba1a5-Abstract.html
[ "Coline Devin", "Daniel Geng", "Pieter Abbeel", "Trevor Darrell", "Sergey Levine" ]
null
null
Autonomous agents situated in real-world environments must be able to master large repertoires of skills. While a single short skill can be learned quickly, it would be impractical to learn every task independently. Instead, the agent should share knowledge across behaviors such that each task can be learned efficientl...
[]
null
1
1910.14033
title_judge
[ 0.02428811974823475, -0.024609901010990143, -0.0225897915661335, 0.05163717269897461, 0.022637078538537025, 0.02280416525900364, 0.022675445303320885, 0.012194259092211723, -0.017312711104750633, -0.03403792530298233, -0.013634984381496906, 0.004828173667192459, -0.08668186515569687, -0.01...
Learning to Propagate for Graph Meta-Learning
https://proceedings.neurips.cc/paper_files/paper/2019/hash/00ac8ed3b4327bdd4ebbebcb2ba10a00-Abstract.html
[ "LU LIU", "Tianyi Zhou", "Guodong Long", "Jing Jiang", "Chengqi Zhang" ]
null
null
Meta-learning extracts the common knowledge from learning different tasks and uses it for unseen tasks. It can significantly improve tasks that suffer from insufficient training data, e.g., few-shot learning. In most meta-learning methods, tasks are implicitly related by sharing parameters or optimizer. In this paper, we...
[]
null
2
1909.05024
title_snapshot
[ 0.016496365889906883, -0.03303815796971321, 0.0016099223867058754, 0.04314010217785835, 0.02586507797241211, 0.022711817175149918, 0.0468393974006176, -0.0018465738976374269, -0.017652586102485657, -0.02154218964278698, 0.013000618666410446, 0.021531635895371437, -0.08870935440063477, 0.01...
XNAS: Neural Architecture Search with Expert Advice
https://proceedings.neurips.cc/paper_files/paper/2019/hash/00e26af6ac3b1c1c49d7c3d79c60d000-Abstract.html
[ "Niv Nayman", "Asaf Noy", "Tal Ridnik", "Itamar Friedman", "Rong Jin", "Lihi Zelnik" ]
null
null
This paper introduces a novel optimization method for differential neural architecture search, based on the theory of prediction with expert advice. Its optimization criterion is well fitted for an architecture-selection, i.e., it minimizes the regret incurred by a sub-optimal selection of operations. Unlike previous s...
[]
null
3
1906.08031
title_snapshot
[ -0.005136671010404825, -0.008787848986685276, -0.00782498624175787, 0.04224070534110069, 0.052238624542951584, 0.05366754159331322, 0.00827226135879755, -0.014974184334278107, -0.009393970482051373, -0.039259713143110275, -0.015297701582312584, -0.010796187445521355, -0.041531480848789215, ...
Multi-resolution Multi-task Gaussian Processes
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0118a063b4aae95277f0bc1752c75abf-Abstract.html
[ "Oliver Hamelijnck", "Theodoros Damoulas", "Kangrui Wang", "Mark Girolami" ]
null
null
We consider evidence integration from potentially dependent observation processes under varying spatio-temporal sampling resolutions and noise levels. We offer a multi-resolution multi-task (MRGP) framework that allows for both inter-task and intra-task multi-resolution and multi-fidelity. We develop shallow Gaussian P...
[]
null
4
1906.08344
title_snapshot
[ 0.02764105424284935, -0.011207228526473045, 0.016401488333940506, 0.01277037151157856, 0.02518375776708126, 0.054857540875673294, 0.028240300714969635, 0.0011025173589587212, -0.021290116012096405, -0.06475187093019485, 0.008764483965933323, 0.010497724637389183, -0.08132120966911316, -0.0...
Deep Equilibrium Models
https://proceedings.neurips.cc/paper_files/paper/2019/hash/01386bd6d8e091c2ab4c7c7de644d37b-Abstract.html
[ "Shaojie Bai", "J. Zico Kolter", "Vladlen Koltun" ]
null
null
We present a new approach to modeling sequential data: the deep equilibrium model (DEQ). Motivated by an observation that the hidden layers of many existing deep sequence models converge towards some fixed point, we propose the DEQ approach that directly finds these equilibrium points via root-finding. Such a method is...
[]
null
5
1909.01377
title_snapshot
[ -0.019374342635273933, -0.032806091010570526, -0.015342333354055882, 0.04606028273701668, 0.02985038235783577, 0.04372020810842514, 0.010601296089589596, 0.0444536916911602, -0.0024436120875179768, -0.03562130406498909, 0.029957294464111328, 0.025483904406428337, -0.0717688724398613, -0.00...
Cross Attention Network for Few-shot Classification
https://proceedings.neurips.cc/paper_files/paper/2019/hash/01894d6f048493d2cacde3c579c315a3-Abstract.html
[ "Ruibing Hou", "Hong Chang", "Bingpeng MA", "Shiguang Shan", "Xilin Chen" ]
null
null
Few-shot classification aims to recognize unlabeled samples from unseen classes given only few labeled samples. The unseen classes and low-data problem make few-shot classification very challenging. Many existing approaches extracted features from labeled and unlabeled samples independently, as a result, the features a...
[]
null
6
1910.07677
title_snapshot
[ 0.022815566509962082, -0.009041774086654186, 0.004961302038282156, 0.022105954587459564, 0.009417076595127583, 0.013225163333117962, 0.030813585966825485, 0.012240346521139145, -0.047044895589351654, -0.0010897669708356261, -0.01308391336351633, 0.001843486912548542, -0.06353796273469925, ...
Order Optimal One-Shot Distributed Learning
https://proceedings.neurips.cc/paper_files/paper/2019/hash/018b59ce1fd616d874afad0f44ba338d-Abstract.html
[ "Arsalan Sharifnassab", "Saber Salehkaleybar", "S. Jamaloddin Golestani" ]
null
null
We consider distributed statistical optimization in one-shot setting, where there are $m$ machines each observing $n$ i.i.d samples. Based on its observed samples, each machine then sends an $O(\log(mn))$-length message to a server, at which a parameter minimizing an expected loss is to be estimated. We propose an algo...
[]
null
7
1911.00731
title_snapshot
[ -0.03461972996592522, -0.0004417299060150981, -0.014529356732964516, 0.03264106810092926, 0.04580479860305786, 0.03864313289523125, 0.05061475932598114, -0.00951378419995308, -0.02685978263616562, -0.03402736783027649, 0.014635290950536728, 0.00445416709408164, -0.07068806886672974, -0.003...
Exact Gaussian Processes on a Million Data Points
https://proceedings.neurips.cc/paper_files/paper/2019/hash/01ce84968c6969bdd5d51c5eeaa3946a-Abstract.html
[ "Ke Wang", "Geoff Pleiss", "Jacob Gardner", "Stephen Tyree", "Kilian Q. Weinberger", "Andrew Gordon Wilson" ]
null
null
Gaussian processes (GPs) are flexible non-parametric models, with a capacity that grows with the available data. However, computational constraints with standard inference procedures have limited exact GPs to problems with fewer than about ten thousand training points, necessitating approximations for larger datasets. ...
[]
null
8
1903.08114
title_snapshot
[ -0.013843861408531666, -0.018625782802700996, 0.020757092162966728, 0.038357432931661606, 0.0029913156758993864, 0.03600069135427475, 0.013703026808798313, 0.036873042583465576, -0.011210891418159008, -0.039643049240112305, -0.024273741990327835, -0.009027876891195774, -0.0846366360783577, ...
Asymmetric Valleys: Beyond Sharp and Flat Local Minima
https://proceedings.neurips.cc/paper_files/paper/2019/hash/01d8bae291b1e4724443375634ccfa0e-Abstract.html
[ "Haowei He", "Gao Huang", "Yang Yuan" ]
null
null
Despite the non-convex nature of their loss functions, deep neural networks are known to generalize well when optimized with stochastic gradient descent (SGD). Recent work conjectures that SGD with proper configuration is able to find wide and flat local minima, which are correlated with good generalization performance. I...
[]
null
9
1902.00744
title_snapshot
[ -0.044796135276556015, -0.017519034445285797, 0.0031994013115763664, 0.026875706389546394, 0.014026260003447533, 0.04623580351471901, 0.029463235288858414, 0.008015851490199566, -0.020947109907865524, -0.05671065300703049, 0.018315382301807404, -0.0013361386954784393, -0.06198834627866745, ...
Calculating Optimistic Likelihoods Using (Geodesically) Convex Optimization
https://proceedings.neurips.cc/paper_files/paper/2019/hash/021e1ea77bd91aaa0fc4d01a943a654e-Abstract.html
[ "Viet Anh Nguyen", "Soroosh Shafieezadeh-Abadeh", "Man-Chung Yue", "Daniel Huhn", "Wolfram Wiesemann" ]
null
null
A fundamental problem arising in many areas of machine learning is the evaluation of the likelihood of a given observation under different nominal distributions. Frequently, these nominal distributions are themselves estimated from data, which makes them susceptible to estimation errors. We thus propose to replace each...
[]
null
10
1910.07817
title_snapshot
[ -0.03130388259887695, 0.005684362258762121, 0.017070356756448746, 0.043052129447460175, 0.03762074187397957, 0.04903228580951691, 0.018681595101952553, 0.019353242591023445, -0.03476376831531525, -0.03743794187903404, -0.03158864006400108, 0.02438010275363922, -0.05963973328471184, -0.0085...
Think out of the "Box": Generically-Constrained Asynchronous Composite Optimization and Hedging
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0224cd598e48c5041c7947fd5cb20d53-Abstract.html
[ "Pooria Joulani", "András György", "Csaba Szepesvari" ]
null
null
We present two new algorithms, ASYNCADA and HEDGEHOG, for asynchronous sparse online and stochastic optimization. ASYNCADA is, to our knowledge, the first asynchronous stochastic optimization algorithm with finite-time data-dependent convergence guarantees for generic convex constraints. In addition, ASYNCADA: (a) allo...
[]
null
11
null
null
[ -0.025918617844581604, -0.008419360034167767, -0.005141011439263821, 0.04171709716320038, 0.027172494679689407, 0.04214901477098465, 0.014186877757310867, 0.027487097308039665, -0.01841939240694046, -0.046725817024707794, 0.019220037385821342, -0.011645388789474964, -0.06182730197906494, -...
Improved Precision and Recall Metric for Assessing Generative Models
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0234c510bc6d908b28c70ff313743079-Abstract.html
[ "Tuomas Kynkäänniemi", "Tero Karras", "Samuli Laine", "Jaakko Lehtinen", "Timo Aila" ]
null
null
The ability to automatically estimate the quality and coverage of the samples produced by a generative model is a vital requirement for driving algorithm research. We present an evaluation metric that can separately and reliably measure both of these aspects in image generation tasks by forming explicit, non-parametric...
[]
null
12
1904.06991
title_snapshot
[ 0.013229217380285263, -0.024435758590698242, -0.0009788041934370995, 0.048060525208711624, 0.02653023600578308, 0.025336800143122673, 0.011009630747139454, 0.003878481686115265, -0.003659345442429185, -0.06749463081359863, -0.01575758494436741, 0.00018352580082137138, -0.050814177840948105, ...
A Direct tilde{O}(1/epsilon) Iteration Parallel Algorithm for Optimal Transport
https://proceedings.neurips.cc/paper_files/paper/2019/hash/024d2d699e6c1a82c9ba986386f4d824-Abstract.html
[ "Arun Jambulapati", "Aaron Sidford", "Kevin Tian" ]
null
null
Optimal transportation, or computing the Wasserstein or ``earth mover's'' distance between two $n$-dimensional distributions, is a fundamental primitive which arises in many learning and statistical settings. We give an algorithm which solves the problem to additive $\epsilon$ accuracy with $\tilde{O}(1/\epsilon)$ para...
[]
null
13
null
null
[ -0.058838386088609695, -0.01000935211777687, 0.0033181458711624146, 0.042770031839609146, 0.03909362107515335, 0.035510897636413574, 0.017612894997000694, 0.020165184512734413, -0.03533675894141197, -0.061688926070928574, 0.012188605032861233, -0.050956107676029205, -0.064461350440979, 0.0...
Zero-Shot Semantic Segmentation
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0266e33d3f546cb5436a10798e657d97-Abstract.html
[ "Maxime Bucher", "Tuan-Hung VU", "Matthieu Cord", "Patrick Pérez" ]
null
null
Semantic segmentation models are limited in their ability to scale to large numbers of object classes. In this paper, we introduce the new task of zero-shot semantic segmentation: learning pixel-wise classifiers for never-seen object categories with zero training examples. To this end, we present a novel architecture, ...
[]
null
14
1906.00817
title_snapshot
[ -0.004394658375531435, -0.003921388182789087, -0.01054689846932888, 0.05368393287062645, 0.02387162856757641, 0.0286122877150774, 0.022290272638201714, 0.04639050364494324, -0.03097868338227272, -0.028898611664772034, -0.05819937586784363, -0.0010118153877556324, -0.04330068454146385, 0.01...
Hyperspherical Prototype Networks
https://proceedings.neurips.cc/paper_files/paper/2019/hash/02a32ad2669e6fe298e607fe7cc0e1a0-Abstract.html
[ "Pascal Mettes", "Elise van der Pol", "Cees Snoek" ]
null
null
This paper introduces hyperspherical prototype networks, which unify classification and regression with prototypes on hyperspherical output spaces. For classification, a common approach is to define prototypes as the mean output vector over training examples per class. Here, we propose to use hyperspheres as output spa...
[]
null
15
1901.10514
title_snapshot
[ -0.013475384563207626, 5.2048271470539476e-8, 0.01593334600329399, 0.035944242030382156, 0.026265298947691917, 0.04167505353689194, 0.016097398474812508, -0.034664906561374664, -0.020615683868527412, -0.04468787461519241, -0.024417851120233536, -0.006063083186745644, -0.07198324799537659, ...
Lower Bounds on Adversarial Robustness from Optimal Transport
https://proceedings.neurips.cc/paper_files/paper/2019/hash/02bf86214e264535e3412283e817deaa-Abstract.html
[ "Arjun Nitin Bhagoji", "Daniel Cullina", "Prateek Mittal" ]
null
null
While progress has been made in understanding the robustness of machine learning classifiers to test-time adversaries (evasion attacks), fundamental questions remain unresolved. In this paper, we use optimal transport to characterize the maximum achievable accuracy in an adversarial classification scenario. In this set...
[]
null
16
1909.12272
title_snapshot
[ -0.03908190131187439, -0.02520378865301609, -0.0002632462710607797, 0.0637378916144371, 0.03704388439655304, -0.0013863110216334462, 0.033949967473745346, -0.035700295120477676, -0.024034015834331512, -0.0643911138176918, -0.0030788076110184193, -0.008401298895478249, -0.06777440756559372, ...
A Nonconvex Approach for Exact and Efficient Multichannel Sparse Blind Deconvolution
https://proceedings.neurips.cc/paper_files/paper/2019/hash/02e656adee09f8394b402d9958389b7d-Abstract.html
[ "Qing Qu", "Xiao Li", "Zhihui Zhu" ]
null
null
We study the multi-channel sparse blind deconvolution (MCS-BD) problem, whose task is to simultaneously recover a kernel $\mathbf a$ and multiple sparse inputs $\{\mathbf x_i\}_{i=1}^p$ from their circulant convolution $\mathbf y_i = \mb a \circledast \mb x_i $ ($i=1,\cdots,p$). We formulate the task as a nonconvex opt...
[]
null
17
1908.10776
title_snapshot
[ -0.021735846996307373, -0.022190488874912262, 0.04127880185842514, 0.051509831100702286, 0.032358985394239426, 0.040761154145002365, 0.025817153975367546, 0.016343237832188606, -0.03532916307449341, -0.06189161539077759, -0.003314032917842269, 0.0146474065259099, -0.04755299538373947, 0.00...
Generalization of Reinforcement Learners with Working and Episodic Memory
https://proceedings.neurips.cc/paper_files/paper/2019/hash/02ed812220b0705fabb868ddbf17ea20-Abstract.html
[ "Meire Fortunato", "Melissa Tan", "Ryan Faulkner", "Steven Hansen", "Adrià Puigdomènech Badia", "Gavin Buttimore", "Charles Deck", "Joel Z. Leibo", "Charles Blundell" ]
null
null
Memory is an important aspect of intelligence and plays a role in many deep reinforcement learning models. However, little progress has been made in understanding when specific memory systems help more than others and how well they generalize. The field also has yet to see a prevalent consistent and rigorous approach f...
[]
null
18
1910.13406
title_snapshot
[ -0.043609004467725754, -0.005970858968794346, -0.01717083342373371, 0.058859385550022125, 0.050863802433013916, 0.008969860151410103, 0.025058520957827568, 0.01898341067135334, -0.03605141118168831, -0.021546745672822, -0.01739451102912426, 0.029910866171121597, -0.07217583805322647, -0.03...
DTWNet: a Dynamic Time Warping Network
https://proceedings.neurips.cc/paper_files/paper/2019/hash/02f063c236c7eef66324b432b748d15d-Abstract.html
[ "Xingyu Cai", "Tingyang Xu", "Jinfeng Yi", "Junzhou Huang", "Sanguthevar Rajasekaran" ]
null
null
Dynamic Time Warping (DTW) is widely used as a similarity measure in various domains. Due to its invariance against warping in the time axis, DTW provides more meaningful discrepancy measurements between two signals than other dis- tance measures. In this paper, we propose a novel component in an artificial neural netw...
[]
null
19
null
null
[ 0.010826703161001205, -0.034602176398038864, -0.014517256990075111, 0.04168996959924698, 0.027086690068244934, 0.05438145622611046, 0.031144002452492714, 0.008008775301277637, -0.01529857236891985, -0.06655927002429962, 0.025116873905062675, 0.00581733975559473, -0.043175771832466125, 0.00...
Learning Mean-Field Games
https://proceedings.neurips.cc/paper_files/paper/2019/hash/030e65da2b1c944090548d36b244b28d-Abstract.html
[ "Xin Guo", "Anran Hu", "Renyuan Xu", "Junzi Zhang" ]
null
null
This paper presents a general mean-field game (GMFG) framework for simultaneous learning and decision-making in stochastic games with a large population. It first establishes the existence of a unique Nash Equilibrium to this GMFG, and explains that naively combining Q-learning with the fixed-point approach in classica...
[]
null
20
1901.09585
title_snapshot
[ -0.045976266264915466, -0.014808247797191143, 0.013422717340290546, 0.006093365140259266, 0.030749879777431488, -0.00003723922782228328, -0.001183086889795959, 0.029152359813451767, -0.03254525363445282, -0.05473456531763077, 0.011902300640940666, 0.022928575053811073, -0.06401758641004562, ...
Learning Erdos-Renyi Random Graphs via Edge Detecting Queries
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0336dcbab05b9d5ad24f4333c7658a0e-Abstract.html
[ "Zihan Li", "Matthias Fresacher", "Jonathan Scarlett" ]
null
null
In this paper, we consider the problem of learning an unknown graph via queries on groups of nodes, with the result indicating whether or not at least one edge is present among those nodes. While learning arbitrary graphs with $n$ nodes and $k$ edges is known to be hard in the sense of requiring $\Omega( \min\{ k^2 \lo...
[]
null
21
1905.03410
title_snapshot
[ 0.004567466676235199, 0.00025545936659909785, 0.012222847901284695, 0.07770302146673203, 0.03781614825129509, 0.027857299894094467, 0.030698973685503006, 0.0168745256960392, -0.005990979261696339, -0.0625213086605072, 0.0035990141332149506, -0.01368701457977295, -0.06969046592712402, -0.00...
Cormorant: Covariant Molecular Neural Networks
https://proceedings.neurips.cc/paper_files/paper/2019/hash/03573b32b2746e6e8ca98b9123f2249b-Abstract.html
[ "Brandon Anderson", "Truong Son Hy", "Risi Kondor" ]
null
null
We propose Cormorant, a rotationally covariant neural network architecture for learning the behavior and properties of complex many-body physical systems. We apply these networks to molecular systems with two goals: learning atomic potential energy surfaces for use in Molecular Dynamics simulations, and learning ground...
[]
null
22
1906.04015
title_snapshot
[ -0.015869716182351112, 0.03295217826962471, -0.007254538591951132, 0.03787759318947792, 0.03280911222100258, -0.018742039799690247, 0.016948705539107323, 0.00316568068228662, -0.011280661448836327, -0.04253722354769707, 0.03321173042058945, 0.007463159970939159, -0.08474336564540863, -0.00...
Flattening a Hierarchical Clustering through Active Learning
https://proceedings.neurips.cc/paper_files/paper/2019/hash/03793ef7d06ffd63d34ade9d091f1ced-Abstract.html
[ "Fabio Vitale", "Anand Rajagopalan", "Claudio Gentile" ]
null
null
We investigate active learning by pairwise similarity over the leaves of trees originating from hierarchical clustering procedures. In the realizable setting, we provide a full characterization of the number of queries needed to achieve perfect reconstruction of the tree cut. In the non-realizable setting, we rely on k...
[]
null
23
1906.09458
title_snapshot
[ -0.018054630607366562, -0.02599100023508072, 0.014107612892985344, 0.025710154324769974, 0.029907388612627983, 0.028132135048508644, 0.005850031040608883, -0.012206675484776497, -0.015013913623988628, -0.03551512211561203, -0.018661338835954666, -0.02410898730158806, -0.06174958497285843, ...
Random Projections and Sampling Algorithms for Clustering of High-Dimensional Polygonal Curves
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0394ea68951e3299bcdfa75a097d7c11-Abstract.html
[ "Stefan Meintrup", "Alexander Munteanu", "Dennis Rohde" ]
null
null
We study the $k$-median clustering problem for high-dimensional polygonal curves with finite but unbounded number of vertices. We tackle the computational issue that arises from the high number of dimensions by defining a Johnson-Lindenstrauss projection for polygonal curves. We analyze the resulting error in terms of ...
[]
null
24
1907.06969
title_snapshot
[ -0.032637663185596466, -0.031279806047677994, 0.003972956910729408, 0.05984698608517647, 0.029108503833413124, 0.05774248391389847, 0.009632855653762817, 0.009513722732663155, -0.01624152436852455, -0.07016399502754211, -0.0497240275144577, -0.022639097645878792, -0.040220048278570175, 0.0...
Explicit Explore-Exploit Algorithms in Continuous State Spaces
https://proceedings.neurips.cc/paper_files/paper/2019/hash/03b264c595403666634ac75d828439bc-Abstract.html
[ "Mikael Henaff" ]
null
null
We present a new model-based algorithm for reinforcement learning (RL) which consists of explicit exploration and exploitation phases, and is applicable in large or infinite state spaces. The algorithm maintains a set of dynamics models consistent with current experience and explores by finding policies which induce hi...
[]
null
25
1911.00617
title_snapshot
[ -0.04368671029806137, -0.0018050044309347868, -0.013175100088119507, 0.04698525369167328, 0.055602140724658966, 0.003765425644814968, 0.025889595970511436, -0.006672713905572891, -0.02239377237856388, -0.021643230691552162, -0.01163812167942524, 0.005912326276302338, -0.06838839501142502, ...
How degenerate is the parametrization of neural networks with the ReLU activation function?
https://proceedings.neurips.cc/paper_files/paper/2019/hash/04115ec378e476c56d19d827bcf8db56-Abstract.html
[ "Dennis Maximilian Elbrächter", "Julius Berner", "Philipp Grohs" ]
null
null
Neural network training is usually accomplished by solving a non-convex optimization problem using stochastic gradient descent. Although one optimizes over the networks parameters, the main loss function generally only depends on the realization of the neural network, i.e. the function it computes. Studying the optimiz...
[]
null
26
1905.09803
title_snapshot
[ -0.03939472883939743, -0.00785763282328844, 0.00024363338889088482, 0.01886446587741375, 0.04210665822029114, 0.05777876451611519, 0.02376578003168106, -0.018386581912636757, -0.016694368794560432, -0.04319826886057854, 0.0017627902561798692, 0.003937114961445332, -0.04906175658106804, 0.0...
Hyperbolic Graph Convolutional Neural Networks
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0415740eaa4d9decbc8da001d3fd805f-Abstract.html
[ "Ines Chami", "Zhitao Ying", "Christopher Ré", "Jure Leskovec" ]
null
null
Graph convolutional neural networks (GCNs) embed nodes in a graph into Euclidean space, which has been shown to incur a large distortion when embedding real-world graphs with scale-free or hierarchical structure. Hyperbolic geometry offers an exciting alternative, as it enables embeddings with much smaller distortion. ...
[]
null
27
1910.12933
title_snapshot
[ -0.014506790786981583, -0.0183761864900589, 0.018293580040335655, 0.04943456873297691, 0.0434931144118309, 0.006443310528993607, 0.01348272617906332, 0.024651508778333664, -0.007186121307313442, -0.07222982496023178, 0.015157104469835758, -0.017874417826533318, -0.05355425179004669, 0.0306...
Spherical Text Embedding
https://proceedings.neurips.cc/paper_files/paper/2019/hash/043ab21fc5a1607b381ac3896176dac6-Abstract.html
[ "Yu Meng", "Jiaxin Huang", "Guangyuan Wang", "Chao Zhang", "Honglei Zhuang", "Lance Kaplan", "Jiawei Han" ]
null
null
Unsupervised text embedding has shown great power in a wide range of NLP tasks. While text embeddings are typically learned in the Euclidean space, directional similarity is often more effective in tasks such as word similarity and document clustering, which creates a gap between the training stage and usage stage of t...
[]
null
28
1911.01196
title_snapshot
[ 0.009934491477906704, -0.004312761127948761, 0.027492066845297813, 0.053161535412073135, 0.03401621803641319, 0.020774461328983307, 0.00870694499462843, 0.0052205962128937244, -0.0033797896467149258, -0.030740084126591682, -0.00808872189372778, 0.00499734515324235, -0.06680586189031601, 0....
Random Tessellation Forests
https://proceedings.neurips.cc/paper_files/paper/2019/hash/043c2ec6c6390dd0ac5519190a57c88c-Abstract.html
[ "Shufei Ge", "Shijia Wang", "Yee Whye Teh", "Liangliang Wang", "Lloyd Elliott" ]
null
null
Space partitioning methods such as random forests and the Mondrian process are powerful machine learning methods for multi-dimensional and relational data, and are based on recursively cutting a domain. The flexibility of these methods is often limited by the requirement that the cuts be axis aligned. The Ostomachion p...
[]
null
29
1906.05440
title_snapshot
[ -0.00020623128511942923, 0.017363576218485832, -0.01904291659593582, 0.025797104462981224, 0.01122227218002081, 0.06688302755355835, 0.03580750152468681, 0.013691534288227558, -0.01844039186835289, -0.06628607958555222, 0.007794501259922981, -0.014902807772159576, -0.05578972399234772, 0.0...
SpArSe: Sparse Architecture Search for CNNs on Resource-Constrained Microcontrollers
https://proceedings.neurips.cc/paper_files/paper/2019/hash/044a23cadb567653eb51d4eb40acaa88-Abstract.html
[ "Igor Fedorov", "Ryan P. Adams", "Matthew Mattina", "Paul Whatmough" ]
null
null
The vast majority of processors in the world are actually microcontroller units (MCUs), which find widespread use performing simple control tasks in applications ranging from automobiles to medical devices and office equipment. The Internet of Things (IoT) promises to inject machine learning into many of these every-da...
[]
null
30
1905.12107
title_snapshot
[ -0.009510827250778675, -0.032992299646139145, -0.01588248275220394, 0.04570954665541649, 0.046391792595386505, 0.048742711544036865, 0.010249205864965916, 0.02334393560886383, -0.027699679136276245, -0.03279220312833786, 0.02511439099907875, -0.02351786568760872, -0.060564521700143814, 0.0...
Capacity Bounded Differential Privacy
https://proceedings.neurips.cc/paper_files/paper/2019/hash/04df4d434d481c5bb723be1b6df1ee65-Abstract.html
[ "Kamalika Chaudhuri", "Jacob Imola", "Ashwin Machanavajjhala" ]
null
null
Differential privacy, a notion of algorithmic stability, is a gold standard for measuring the additional risk an algorithm's output poses to the privacy of a single record in the dataset. Differential privacy is defined as the distance between the output distribution of an algorithm on neighboring datasets that differ ...
[]
null
31
1907.02159
title_snapshot
[ -0.0053112744353711605, 0.00032366992672905326, -0.03423668071627617, 0.04154916852712631, 0.07532244920730591, 0.02423281967639923, 0.047232821583747864, -0.06563030183315277, -0.014590146951377392, -0.008830848149955273, 0.007974034175276756, -0.022943392395973206, -0.06240672990679741, ...
Information-Theoretic Generalization Bounds for SGLD via Data-Dependent Estimates
https://proceedings.neurips.cc/paper_files/paper/2019/hash/05ae14d7ae387b93370d142d82220f1b-Abstract.html
[ "Jeffrey Negrea", "Mahdi Haghifam", "Gintare Karolina Dziugaite", "Ashish Khisti", "Daniel M. Roy" ]
null
null
In this work, we improve upon the stepwise analysis of noisy iterative learning algorithms initiated by Pensia, Jog, and Loh (2018) and recently extended by Bu, Zou, and Veeravalli (2019). Our main contributions are significantly improved mutual information bounds for Stochastic Gradient Langevin Dynamics via data-depe...
[]
null
32
1911.02151
title_snapshot
[ -0.016341960057616234, 0.012474039569497108, 0.007940134033560753, 0.028726737946271896, 0.04164815694093704, 0.04597673937678337, 0.06360276788473129, -0.010933912359178066, -0.01271809171885252, -0.028797848150134087, -0.006976723205298185, 0.01065280195325613, -0.07781673222780228, -0.0...
Efficient Algorithms for Smooth Minimax Optimization
https://proceedings.neurips.cc/paper_files/paper/2019/hash/05d0abb9a864ae4981e933685b8b915c-Abstract.html
[ "Kiran K Thekumparampil", "Prateek Jain", "Praneeth Netrapalli", "Sewoong Oh" ]
null
null
This paper studies first order methods for solving smooth minimax optimization problems $\min_x \max_y g(x,y)$ where $g(\cdot,\cdot)$ is smooth and $g(x,\cdot)$ is concave for each $x$. In terms of $g(\cdot,y)$, we consider two settings -- strongly convex and nonconvex -- and improve upon the best known rates in both. ...
[]
null
33
1907.01543
title_snapshot
[ -0.04260624945163727, -0.005877708084881306, 0.03621922805905342, 0.03674769401550293, 0.025521976873278618, 0.03410254418849945, 0.019978215917944908, 0.017274631187319756, -0.008771635591983795, -0.05798456817865372, -0.00774840172380209, -0.02748255245387554, -0.06385061144828796, 0.011...
Uniform convergence may be unable to explain generalization in deep learning
https://proceedings.neurips.cc/paper_files/paper/2019/hash/05e97c207235d63ceb1db43c60db7bbb-Abstract.html
[ "Vaishnavh Nagarajan", "J. Zico Kolter" ]
null
null
Aimed at explaining the surprisingly good generalization behavior of overparameterized deep networks, recent works have developed a variety of generalization bounds for deep learning, all based on the fundamental learning-theoretic technique of uniform convergence. While it is well-known that many of these existing bou...
[]
null
34
1902.04742
title_snapshot
[ -0.014560380950570107, -0.02543838880956173, -0.0029798722825944424, 0.045025818049907684, 0.0443820022046566, 0.01343497447669506, 0.043873488903045654, 0.005232988856732845, -0.018547644838690758, -0.041937943547964096, -0.006511871237307787, -0.027020743116736412, -0.08565858751535416, ...
First order expansion of convex regularized estimators
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0609154fa35b3194026346c9cac2a248-Abstract.html
[ "Pierre Bellec", "Arun Kuchibhotla" ]
null
null
We consider first order expansions of convex penalized estimators in high-dimensional regression problems with random designs. Our setting includes linear regression and logistic regression as special cases. For a given penalty function $h$ and the corresponding penalized estimator $\hbeta$, we construct a quantity $\e...
[]
null
35
1910.05480
title_snapshot
[ -0.00910566933453083, -0.007773092482239008, -0.0008956080419011414, -0.014329451136291027, 0.04366537928581238, 0.04806055501103401, 0.02449919655919075, -0.005140721332281828, -0.03320362791419029, -0.05064813047647476, -0.003433734178543091, 0.017634207382798195, -0.06837491691112518, -...
Robust exploration in linear quadratic reinforcement learning
https://proceedings.neurips.cc/paper_files/paper/2019/hash/060fd70a06ead2e1079d27612b84aff4-Abstract.html
[ "Jack Umenberger", "Mina Ferizbegovic", "Thomas B Schön", "Håkan Hjalmarsson" ]
null
null
Learning to make decisions in an uncertain and dynamic environment is a task of fundamental performance in a number of domains. This paper concerns the problem of learning control policies for an unknown linear dynamical system so as to minimize a quadratic cost function. We present a method, based on convex optimizati...
[]
null
36
1906.01584
title_snapshot
[ -0.029700325801968575, 0.01425598282366991, -0.0020418723579496145, 0.03756002336740494, 0.04710164666175842, 0.03174171596765518, 0.0034727994352579117, -0.0064124614000320435, -0.0348893478512764, -0.035621900111436844, -0.02187236398458481, 0.0011592756491154432, -0.05409284681081772, -...
Modeling Uncertainty by Learning a Hierarchy of Deep Neural Connections
https://proceedings.neurips.cc/paper_files/paper/2019/hash/063e26c670d07bb7c4d30e6fc69fe056-Abstract.html
[ "Raanan Yehezkel Rohekar", "Yaniv Gurwicz", "Shami Nisimov", "Gal Novik" ]
null
null
Modeling uncertainty in deep neural networks, despite recent important advances, is still an open problem. Bayesian neural networks are a powerful solution, where the prior over network weights is a design choice, often a normal distribution or other distribution encouraging sparsity. However, this prior is agnostic to...
[]
null
37
1905.13195
title_snapshot
[ 0.003859465243294835, 0.020191725343465805, -0.030843568965792656, 0.056347474455833435, 0.04370575398206711, 0.04290083423256874, 0.020388685166835785, -0.011412706226110458, -0.02268311008810997, -0.06465446949005127, -0.0019078978803008795, 0.012156400829553604, -0.06112777069211006, 0....
Meta-Surrogate Benchmarking for Hyperparameter Optimization
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0668e20b3c9e9185b04b3d2a9dc8fa2d-Abstract.html
[ "Aaron Klein", "Zhenwen Dai", "Frank Hutter", "Neil Lawrence", "Javier Gonzalez" ]
null
null
Despite the recent progress in hyperparameter optimization (HPO), available benchmarks that resemble real-world scenarios consist of a few and very large problem instances that are expensive to solve. This blocks researchers and practitioners no only from systematically running large-scale comparisons that are needed t...
[]
null
38
1905.12982
title_snapshot
[ -0.016080161556601524, 0.0065997447818517685, -0.008447255939245224, 0.05290370061993599, 0.024005472660064697, 0.04826890677213669, 0.03481215611100197, -0.011629801243543625, -0.004201658070087433, -0.028580056503415108, 0.00597003661096096, -0.016508515924215317, -0.053419459611177444, ...
Time/Accuracy Tradeoffs for Learning a ReLU with respect to Gaussian Marginals
https://proceedings.neurips.cc/paper_files/paper/2019/hash/067a26d87265ea39030f5bd82408ce7c-Abstract.html
[ "Surbhi Goel", "Sushrut Karmalkar", "Adam Klivans" ]
null
null
We consider the problem of computing the best-fitting ReLU with respect to square-loss on a training set when the examples have been drawn according to a spherical Gaussian distribution (the labels can be arbitrary). Let $\opt < 1$ be the population loss of the best-fitting ReLU. We prove: \begin{itemize} \item Finding...
[]
null
39
1911.01462
title_snapshot
[ -0.0021874229423701763, 0.010450146161019802, 0.014893892221152782, 0.028971748426556587, 0.023469801992177963, 0.06156871095299721, 0.03013506717979908, -0.007856189273297787, -0.03651149943470955, -0.02728893794119358, -0.02461286447942257, 0.023750750347971916, -0.08232919126749039, -0....
Bayesian Optimization under Heavy-tailed Payoffs
https://proceedings.neurips.cc/paper_files/paper/2019/hash/06a50e3f66db4a334202d3adfd31c589-Abstract.html
[ "Sayak Ray Chowdhury", "Aditya Gopalan" ]
null
null
We consider black box optimization of an unknown function in the nonparametric Gaussian process setting when the noise in the observed function values can be heavy tailed. This is in contrast to existing literature that typically assumes sub-Gaussian noise distributions for queries. Under the assumption that the unknow...
[]
null
40
1909.07040
title_snapshot
[ -0.032810430973768234, -0.0048263706266880035, 0.02191442996263504, 0.04797448590397835, 0.033495426177978516, 0.04622660577297211, 0.01024648454040289, -0.008433463983237743, 0.01375497318804264, -0.04081416130065918, -0.029248416423797607, 0.022733930498361588, -0.04932506009936333, -0.0...
Distribution Learning of a Random Spatial Field with a Location-Unaware Mobile Sensor
https://proceedings.neurips.cc/paper_files/paper/2019/hash/06bf16f1f0372a63d520eac6cf7c5af7-Abstract.html
[ "Meera Pai", "Animesh Kumar" ]
null
null
Measurement of spatial fields is of interest in environment monitoring. Recently mobile sensing has been proposed for spatial field reconstruction, which requires a smaller number of sensors when compared to the traditional paradigm of sensing with static sensors. A challenge in mobile sensing is to overcome the locati...
[]
null
41
null
null
[ 0.016282888129353523, 0.002138959476724267, 0.0022677856031805277, 0.03327244892716408, 0.06001446768641472, 0.017679965123534203, 0.03183359652757645, 0.014102411456406116, -0.027593573555350304, -0.052911948412656784, 0.017557719722390175, -0.02207501046359539, -0.03259653598070145, -0.0...
State Aggregation Learning from Markov Transition Data
https://proceedings.neurips.cc/paper_files/paper/2019/hash/070dbb6024b5ef93784428afc71f2146-Abstract.html
[ "Yaqi Duan", "Tracy Ke", "Mengdi Wang" ]
null
null
State aggregation is a popular model reduction method rooted in optimal control. It reduces the complexity of engineering systems by mapping the system’s states into a small number of meta-states. The choice of aggregation map often depends on the data analysts’ knowledge and is largely ad hoc. In this paper, we propos...
[]
null
42
1811.02619
title_snapshot
[ -0.061203792691230774, 0.0021288732532411814, 0.017444564029574394, 0.013110723346471786, 0.042550139129161835, 0.002323330845683813, 0.03125958889722824, 0.011904262006282806, -0.009767048992216587, -0.028047552332282066, 0.008382268249988556, -0.03115980327129364, -0.09887789189815521, -...
Reliable training and estimation of variance networks
https://proceedings.neurips.cc/paper_files/paper/2019/hash/07211688a0869d995947a8fb11b215d6-Abstract.html
[ "Nicki Skafte", "Martin Jørgensen", "Søren Hauberg" ]
null
null
We propose and investigate new complementary methodologies for estimating predictive variance networks in regression neural networks. We derive a locally aware mini-batching scheme that results in sparse robust gradients, and we show how to make unbiased weight updates to a variance network. Further, we formulate a heu...
[]
null
43
1906.03260
title_snapshot
[ 0.000793601619079709, 0.013141981326043606, -0.01227279007434845, 0.0210246779024601, 0.04586091637611389, 0.06572940945625305, 0.04051954299211502, -0.021129779517650604, -0.03658395633101463, -0.04896777495741844, 0.0051015825010836124, 0.01555679365992546, -0.05353013426065445, 0.000016...
Meta-Learning with Implicit Gradients
https://proceedings.neurips.cc/paper_files/paper/2019/hash/072b030ba126b2f4b2374f342be9ed44-Abstract.html
[ "Aravind Rajeswaran", "Chelsea Finn", "Sham M. Kakade", "Sergey Levine" ]
null
null
A core capability of intelligent systems is the ability to quickly learn new tasks by drawing on prior experience. Gradient (or optimization) based meta-learning has recently emerged as an effective approach for few-shot learning. In this formulation, meta-parameters are learned in the outer loop, while task-specific m...
[]
null
44
1909.04630
title_snapshot
[ -0.015183787792921066, 0.012502637691795826, -0.011506227776408195, 0.02920861355960369, 0.03530409559607506, 0.029631853103637695, 0.03519459441304207, -0.0066458685323596, -0.03818827494978905, 0.004434701055288315, -0.018610630184412003, 0.025199342519044876, -0.07065832614898682, -0.01...
Differentially Private Markov Chain Monte Carlo
https://proceedings.neurips.cc/paper_files/paper/2019/hash/074177d3eb6371e32c16c55a3b8f706b-Abstract.html
[ "Mikko Heikkilä", "Joonas Jälkö", "Onur Dikmen", "Antti Honkela" ]
null
null
Recent developments in differentially private (DP) machine learning and DP Bayesian learning have enabled learning under strong privacy guarantees for the training data subjects. In this paper, we further extend the applicability of DP Bayesian learning by presenting the first general DP Markov chain Monte Carlo (MCMC)...
[]
null
45
1901.10275
title_snapshot
[ -0.009255885146558285, 0.04031451791524887, -0.01557136420160532, 0.07304201275110245, 0.0649862140417099, 0.02779436856508255, 0.052576422691345215, -0.04919033125042915, -0.014580518007278442, -0.030444111675024033, 0.011276323348283768, 0.01871546171605587, -0.055367615073919296, -0.004...
Universal Boosting Variational Inference
https://proceedings.neurips.cc/paper_files/paper/2019/hash/07a4e20a7bbeeb7a736682b26b16ebe8-Abstract.html
[ "Trevor Campbell", "Xinglong Li" ]
null
null
Boosting variational inference (BVI) approximates an intractable probability density by iteratively building up a mixture of simple component distributions one at a time, using techniques from sparse convex optimization to provide both computational scalability and approximation error guarantees. But the guarantees hav...
[]
null
46
1906.01235
title_snapshot
[ -0.010021080262959003, -0.004096681252121925, 0.01980522647500038, 0.04104113578796387, 0.026835180819034576, 0.03607979789376259, 0.03246794641017914, 0.003394262632355094, -0.02824944071471691, -0.046713292598724365, -0.0034620880614966154, 0.017432091757655144, -0.0752643495798111, -0.0...
LIIR: Learning Individual Intrinsic Reward in Multi-Agent Reinforcement Learning
https://proceedings.neurips.cc/paper_files/paper/2019/hash/07a9d3fed4c5ea6b17e80258dee231fa-Abstract.html
[ "Yali Du", "Lei Han", "Meng Fang", "Ji Liu", "Tianhong Dai", "Dacheng Tao" ]
null
null
A great challenge in cooperative decentralized multi-agent reinforcement learning (MARL) is generating diversified behaviors for each individual agent when receiving only a team reward. Prior studies have paid much effort on reward shaping or designing a centralized critic that can discriminatively credit the agents. I...
[]
null
47
null
null
[ -0.03770797699689865, -0.026640886440873146, -0.006400920916348696, 0.01437375321984291, 0.027289440855383873, 0.01098782941699028, 0.010577342472970486, -0.0092245414853096, -0.020755188539624214, -0.05001712962985039, -0.02550913393497467, 0.02013288624584675, -0.05234556645154953, -0.01...
A Normative Theory for Causal Inference and Bayes Factor Computation in Neural Circuits
https://proceedings.neurips.cc/paper_files/paper/2019/hash/07cb5f86508f146774a2fac4373a8e50-Abstract.html
[ "Wenhao Zhang", "Si Wu", "Brent Doiron", "Tai Sing Lee" ]
null
null
This study provides a normative theory for how Bayesian causal inference can be implemented in neural circuits. In both cognitive processes such as causal reasoning and perceptual inference such as cue integration, the nervous systems need to choose different models representing the underlying causal structures when ma...
[]
null
48
null
null
[ -0.005019127391278744, 0.041722673922777176, -0.025358013808727264, 0.003501100931316614, 0.026234108954668045, 0.02197207696735859, 0.05342729762196541, 0.023946644738316536, -0.05893813446164131, -0.04460322856903076, -0.0008099196129478514, 0.030587250366806984, -0.04055623337626457, -0...
The Geometry of Deep Networks: Power Diagram Subdivision
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0801b20e08c3242125d512808cd74302-Abstract.html
[ "Randall Balestriero", "Romain Cosentino", "Behnaam Aazhang", "Richard Baraniuk" ]
null
null
We study the geometry of deep (neural) networks (DNs) with piecewise affine and convex nonlinearities. The layers of such DNs have been shown to be max-affine spline operators (MASOs) that partition their input space and apply a region-dependent affine mapping to their input to produce their output. We demonstrate that...
[]
null
49
1905.08443
title_snapshot
[ -0.056736212223768234, 0.015206512063741684, -0.0015615113079547882, 0.04617832601070404, 0.023965934291481972, 0.04772196337580681, 0.019258663058280945, 0.00333883473649621, -0.03492053225636482, -0.044325925409793854, -0.005735922139137983, -0.025490107014775276, -0.04924832656979561, 0...
Visual Sequence Learning in Hierarchical Prediction Networks and Primate Visual Cortex
https://proceedings.neurips.cc/paper_files/paper/2019/hash/08040837089cdf46631a10aca5258e16-Abstract.html
[]
null
null
In this paper we developed a computational hierarchical network model to understand the spatiotemporal sequence learning effects observed in the primate visual cortex. The model is a hierarchical recurrent neural model that learns to predict video sequences using the incoming video signals as teaching signals. The mode...
[]
null
50
null
null
[ -0.014335324987769127, 0.013760147616267204, 0.0027523268945515156, 0.027124591171741486, 0.0281014796346426, 0.0197908952832222, 0.045137930661439896, 0.037268541753292084, -0.06828036904335022, -0.03582064434885979, -0.0014561741845682263, -0.013059251010417938, -0.04948725923895836, -0....
Equal Opportunity in Online Classification with Partial Feedback
https://proceedings.neurips.cc/paper_files/paper/2019/hash/084afd913ab1e6ea58b8ca73f6cb41a6-Abstract.html
[ "Yahav Bechavod", "Katrina Ligett", "Aaron Roth", "Bo Waggoner", "Steven Z. Wu" ]
null
null
We study an online classification problem with partial feedback in which individuals arrive one at a time from a fixed but unknown distribution, and must be classified as positive or negative. Our algorithm only observes the true label of an individual if they are given a positive classification. This setting captures ...
[]
null
51
1902.02242
title_snapshot
[ -0.012792947702109814, -0.029894182458519936, -0.0322694294154644, 0.04575083404779434, 0.027028627693653107, 0.053667232394218445, 0.011793122626841068, 0.010881932452321053, -0.04135395586490631, -0.027887869626283646, -0.021127821877598763, -0.021439695730805397, -0.07539047300815582, -...
Semi-Parametric Efficient Policy Learning with Continuous Actions
https://proceedings.neurips.cc/paper_files/paper/2019/hash/08b7dc6e8b36bcaac15847827b7951a9-Abstract.html
[ "Victor Chernozhukov", "Mert Demirer", "Greg Lewis", "Vasilis Syrgkanis" ]
null
null
We consider off-policy evaluation and optimization with continuous action spaces. We focus on observational data where the data collection policy is unknown and needs to be estimated from data. We take a semi-parametric approach where the value function takes a known parametric form in the treatment, but we are agnosti...
[]
null
52
1905.10116
title_snapshot
[ -0.005186316091567278, -0.03180881217122078, -0.010444672778248787, 0.040656521916389465, 0.023307979106903076, 0.033610038459300995, 0.02230623923242092, -0.022381428629159927, 0.004011102952063084, -0.03352378308773041, -0.013223168440163136, 0.011666701175272465, -0.07024583965539932, -...
Concentration of risk measures: A Wasserstein distance approach
https://proceedings.neurips.cc/paper_files/paper/2019/hash/091bc5440296cc0e41dd60ce22fbaf88-Abstract.html
[ "Sanjay P. Bhat", "Prashanth L.A." ]
null
null
Known finite-sample concentration bounds for the Wasserstein distance between the empirical and true distribution of a random variable are used to derive a two-sided concentration bound for the error between the true conditional value-at-risk (CVaR) of a (possibly unbounded) random variable and a standard estimate of i...
[]
null
53
1902.10709
title_judge
[ -0.010355981066823006, 0.007301768288016319, -0.001506934640929103, 0.023860175162553787, 0.0502629280090332, 0.008989955298602581, 0.0177830308675766, 0.01590149663388729, -0.009538982063531876, -0.0297083742916584, -0.011595656163990498, 0.01609036698937416, -0.04395433887839317, -0.0268...
Interior-Point Methods Strike Back: Solving the Wasserstein Barycenter Problem
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0937fb5864ed06ffb59ae5f9b5ed67a9-Abstract.html
[ "DongDong Ge", "Haoyue Wang", "Zikai Xiong", "Yinyu Ye" ]
null
null
Computing the Wasserstein barycenter of a set of probability measures under the optimal transport metric can quickly become prohibitive for traditional second-order algorithms, such as interior-point methods, as the support size of the measures increases. In this paper, we overcome the difficulty by developing a new ad...
[]
null
54
1905.12895
title_snapshot
[ -0.03829224780201912, -0.026745909824967384, 0.025171425193548203, 0.024582458660006523, 0.03584742173552513, 0.04821385070681572, 0.011287281289696693, -0.004984102677553892, -0.036547694355249405, -0.08067602664232254, -0.015146251767873764, -0.031034301966428757, -0.01585269160568714, -...
Coda: An End-to-End Neural Program Decompiler
https://proceedings.neurips.cc/paper_files/paper/2019/hash/093b60fd0557804c8ba0cbf1453da22f-Abstract.html
[ "Cheng Fu", "Huili Chen", "Haolan Liu", "Xinyun Chen", "Yuandong Tian", "Farinaz Koushanfar", "Jishen Zhao" ]
null
null
Reverse engineering of binary executables is a critical problem in the computer security domain. On the one hand, malicious parties may recover interpretable source codes from the software products to gain commercial advantages. On the other hand, binary decompilation can be leveraged for code vulnerability analysis an...
[]
null
55
null
null
[ -0.02090132050216198, -0.0059084706008434296, -0.04582208767533302, 0.03802467882633209, 0.04227346554398537, 0.03406493738293648, 0.029288141056895256, -0.0018262299709022045, -0.014874597080051899, -0.03189855441451073, -0.015261181630194187, 0.015326421707868576, -0.05547725036740303, 0...
GPipe: Efficient Training of Giant Neural Networks using Pipeline Parallelism
https://proceedings.neurips.cc/paper_files/paper/2019/hash/093f65e080a295f8076b1c5722a46aa2-Abstract.html
[ "Yanping Huang", "Youlong Cheng", "Ankur Bapna", "Orhan Firat", "Dehao Chen", "Mia Chen", "HyoukJoong Lee", "Jiquan Ngiam", "Quoc V Le", "Yonghui Wu", "zhifeng Chen" ]
null
null
Scaling up deep neural network capacity has been known as an effective approach to improving model quality for several different machine learning tasks. In many cases, increasing model capacity beyond the memory limit of a single accelerator has required developing special algorithms or infrastructure. These solutions ...
[]
null
56
1811.06965
title_snapshot
[ -0.005876851733773947, -0.05683736875653267, -0.02796659991145134, 0.01593782566487789, 0.026713699102401733, 0.03752260282635689, 0.022830398753285408, 0.013573363423347473, -0.019173694774508476, -0.03571857511997223, 0.02484203688800335, -0.009584533981978893, -0.06863449513912201, 0.03...
DiskANN: Fast Accurate Billion-point Nearest Neighbor Search on a Single Node
https://proceedings.neurips.cc/paper_files/paper/2019/hash/09853c7fb1d3f8ee67a61b6bf4a7f8e6-Abstract.html
[ "Suhas Jayaram Subramanya", "Fnu Devvrit", "Harsha Vardhan Simhadri", "Ravishankar Krishnawamy", "Rohan Kadekodi" ]
null
null
Current state-of-the-art approximate nearest neighbor search (ANNS) algorithms generate indices that must be stored in main memory for fast high-recall search. This makes them expensive and limits the size of the dataset. We present a new graph-based indexing and search system called DiskANN that can index, store, and ...
[]
null
57
null
null
[ -0.06399684399366379, -0.05034329742193222, 0.010310434736311436, 0.043723464012145996, 0.02420477382838726, 0.026016976684331894, -0.011635455302894115, 0.012013869360089302, -0.05131307244300842, -0.042872779071331024, 0.004814965184777975, -0.02876175008714199, -0.054379578679800034, -0...
Linear Stochastic Bandits Under Safety Constraints
https://proceedings.neurips.cc/paper_files/paper/2019/hash/09a8a8976abcdfdee15128b4cc02f33a-Abstract.html
[ "Sanae Amani", "Mahnoosh Alizadeh", "Christos Thrampoulidis" ]
null
null
Bandit algorithms have various application in safety-critical systems, where it is important to respect the system constraints that rely on the bandit's unknown parameters at every round. In this paper, we formulate a linear stochastic multi-armed bandit problem with safety constraints that depend (linearly) on an unkn...
[]
null
58
1908.05814
title_snapshot
[ -0.014982081949710846, -0.002558478619903326, -0.009467064402997494, 0.03694021329283714, 0.037931833416223526, 0.02470516227185726, 0.032990362495183945, 0.017891738563776016, -0.020636914297938347, -0.04766235500574112, -0.03240212798118591, -0.009516187943518162, -0.07564318180084229, -...
Power analysis of knockoff filters for correlated designs
https://proceedings.neurips.cc/paper_files/paper/2019/hash/09ab23b6b607496f095feed7aaa1259b-Abstract.html
[ "Jingbo Liu", "Philippe Rigollet" ]
null
null
The knockoff filter introduced by Barber and Cand\`es 2016 is an elegant framework for controlling the false discovery rate in variable selection. While empirical results indicate that this methodology is not too conservative, there is no conclusive theoretical result on its power. When the predictors are i.i.d.\ Gauss...
[]
null
59
1910.12428
title_snapshot
[ -0.004964002408087254, 0.017331786453723907, -0.01154231745749712, 0.03388511762022972, 0.04887649416923523, 0.02795725129544735, 0.03329458087682724, -0.025241052731871605, -0.003856905736029148, -0.048089127987623215, 0.01650010049343109, -0.01323182974010706, -0.07660020887851715, 0.000...
Implicitly learning to reason in first-order logic
https://proceedings.neurips.cc/paper_files/paper/2019/hash/09fb05dd477d4ae6479985ca56c5a12d-Abstract.html
[ "Vaishak Belle", "Brendan Juba" ]
null
null
We consider the problem of answering queries about formulas of first-order logic based on background knowledge partially represented explicitly as other formulas, and partially represented as examples independently drawn from a fixed probability distribution. PAC semantics, introduced by Valiant, is one rigorous, gener...
[]
null
60
1906.10106
title_snapshot
[ -0.037273507565259933, 0.0216495543718338, -0.015535509213805199, 0.025160392746329308, 0.037739627063274384, -0.0020525821018964052, 0.027435392141342163, -0.004898719489574432, -0.017984312027692795, 0.007628481835126877, -0.027305303141474724, 0.0578189492225647, -0.07546891272068024, 0...
Low-Rank Bandit Methods for High-Dimensional Dynamic Pricing
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0a3df70393993583a13c0dd6686f3f32-Abstract.html
[ "Jonas W Mueller", "Vasilis Syrgkanis", "Matt Taddy" ]
null
null
We consider dynamic pricing with many products under an evolving but low-dimensional demand model. Assuming the temporal variation in cross-elasticities exhibits low-rank structure based on fixed (latent) features of the products, we show that the revenue maximization problem reduces to an online bandit convex optimiza...
[]
null
61
1801.10242
title_snapshot
[ -0.039654072374105453, -0.022052427753806114, 0.016879312694072723, 0.05683770403265953, 0.03521932289004326, 0.058129556477069855, 0.019064540043473244, 0.017402537167072296, -0.01886681094765663, -0.02055084891617298, -0.009593057446181774, 0.011349976062774658, -0.04090403765439987, 0.0...
Learning Stable Deep Dynamics Models
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0a4bbceda17a6253386bc9eb45240e25-Abstract.html
[ "J. Zico Kolter", "Gaurav Manek" ]
null
null
Deep networks are commonly used to model dynamical systems, predicting how the state of a system will evolve over time (either autonomously or in response to control inputs). Despite the predictive power of these systems, it has been difficult to make formal claims about the basic properties of the learned systems. In ...
[]
null
62
2001.06116
title_snapshot
[ -0.012643085792660713, -0.013623561710119247, -0.009927960112690926, 0.0555151104927063, 0.03949080780148506, 0.02982630766928196, 0.01902717724442482, 0.013381090015172958, -0.038792967796325684, -0.04276476055383682, 0.020438149571418762, -0.007214289624243975, -0.06332390755414963, 0.01...
Beyond the Single Neuron Convex Barrier for Neural Network Certification
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0a9fdbb17feb6ccb7ec405cfb85222c4-Abstract.html
[ "Gagandeep Singh", "Rupanshu Ganvir", "Markus Püschel", "Martin Vechev" ]
null
null
We propose a new parametric framework, called k-ReLU, for computing precise and scalable convex relaxations used to certify neural networks. The key idea is to approximate the output of multiple ReLUs in a layer jointly instead of separately. This joint relaxation captures dependencies between the inputs to different R...
[]
null
63
null
null
[ -0.028679650276899338, -0.006386277265846729, -0.0031059803441166878, 0.04293786361813545, 0.04584953933954239, 0.05835674703121185, 0.031079724431037903, -0.011745152063667774, -0.04150982201099396, -0.010320638306438923, -0.0021108139771968126, 0.00023196842812467366, -0.03408016264438629,...
Variational Mixture-of-Experts Autoencoders for Multi-Modal Deep Generative Models
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0ae775a8cb3b499ad1fca944e6f5c836-Abstract.html
[ "Yuge Shi", "Siddharth N", "Brooks Paige", "Philip Torr" ]
null
null
Learning generative models that span multiple data modalities, such as vision and language, is often motivated by the desire to learn more useful, generalisable representations that faithfully capture common underlying factors between the modalities. In this work, we characterise successful learning of such models as t...
[]
null
64
1911.03393
title_snapshot
[ 0.016511356458067894, -0.00894079077988863, 0.007754335645586252, 0.05631750077009201, 0.022729380056262016, 0.04227398335933685, 0.046129610389471054, 0.011392039246857166, -0.035121720284223557, -0.03265312314033508, -0.028278052806854248, 0.02674487605690956, -0.05736171454191208, 0.012...
Language as an Abstraction for Hierarchical Deep Reinforcement Learning
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0af787945872196b42c9f73ead2565c8-Abstract.html
[ "YiDing Jiang", "Shixiang (Shane) Gu", "Kevin P. Murphy", "Chelsea Finn" ]
null
null
Solving complex, temporally-extended tasks is a long-standing problem in reinforcement learning (RL). We hypothesize that one critical element of solving such problems is the notion of compositionality. With the ability to learn sub-skills that can be composed to solve longer tasks, i.e. hierarchical RL, we can acquire...
[]
null
65
1906.07343
title_snapshot
[ -0.033197082579135895, -0.013441435061395168, -0.004815021064132452, 0.04401996359229088, 0.025734005495905876, 0.009570488706231117, 0.008838161826133728, 0.005990091245621443, -0.03808963671326637, -0.02003554441034794, -0.02099241130053997, 0.025861434638500214, -0.06401429325342178, 0....
High-dimensional multivariate forecasting with low-rank Gaussian Copula Processes
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0b105cf1504c4e241fcc6d519ea962fb-Abstract.html
[ "David Salinas", "Michael Bohlke-Schneider", "Laurent Callot", "Roberto Medico", "Jan Gasthaus" ]
null
null
Predicting the dependencies between observations from multiple time series is critical for applications such as anomaly detection, financial risk management, causal analysis, or demand forecasting. However, the computational and numerical difficulties of estimating time-varying and high-dimensional covariance matrices ...
[]
null
66
1910.03002
title_snapshot
[ -0.023427018895745277, -0.027964234352111816, 0.002241003094241023, 0.01691598631441593, 0.03916231915354729, 0.052746083587408066, 0.0263360608369112, -0.003518453799188137, -0.02006339095532894, -0.031466156244277954, 0.006317381281405687, 0.028088223189115524, -0.0702783614397049, 0.038...
Learning Macroscopic Brain Connectomes via Group-Sparse Factorization
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0bfce127947574733b19da0f30739fcd-Abstract.html
[ "Farzane Aminmansour", "Andrew Patterson", "Lei Le", "Yisu Peng", "Daniel Mitchell", "Franco Pestilli", "Cesar F. Caiafa", "Russell Greiner", "Martha White" ]
null
null
Mapping structural brain connectomes for living human brains typically requires expert analysis and rule-based models on diffusion-weighted magnetic resonance imaging. A data-driven approach, however, could overcome limitations in such rule-based approaches and improve precision mappings for individuals. In this work, ...
[]
null
67
null
null
[ -0.03810485079884529, 0.004650741349905729, 0.025297854095697403, 0.03330615162849426, 0.04281225427985191, 0.028727808967232704, 0.030377104878425598, 0.014721531420946121, -0.03224140778183937, -0.054679784923791885, 0.015377027913928032, -0.01620998978614807, -0.0615801140666008, 0.0035...
Optimal Sketching for Kronecker Product Regression and Low Rank Approximation
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0c215f194276000be6a6df6528067151-Abstract.html
[ "Huaian Diao", "Rajesh Jayaram", "Zhao Song", "Wen Sun", "David Woodruff" ]
null
null
We study the Kronecker product regression problem, in which the design matrix is a Kronecker product of two or more matrices. Formally, given $A_i \in \R^{n_i \times d_i}$ for $i=1,2,\dots,q$ where $n_i \gg d_i$ for each $i$, and $b \in \R^{n_1 n_2 \cdots n_q}$, let $\mathcal{A} = A_i \otimes A_2 \otimes \cdots \otimes...
[]
null
68
1909.13384
title_snapshot
[ -0.006024644244462252, -0.006279825232923031, 0.011115113273262978, 0.047797318547964096, 0.0720914676785469, 0.04100885987281799, 0.013587872497737408, -0.006031806580722332, -0.013656629249453545, -0.051235347986221313, -0.00849092472344637, -0.009361698292195797, -0.04555501043796539, 0...
Deep Gamblers: Learning to Abstain with Portfolio Theory
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0c4b1eeb45c90b52bfb9d07943d855ab-Abstract.html
[ "Ziyin Liu", "Zhikang Wang", "Paul Pu Liang", "Ruslan Salakhutdinov", "Louis-Philippe Morency", "Masahito Ueda" ]
null
null
We deal with the selective classification problem (supervised-learning problem with a rejection option), where we want to achieve the best performance at a certain level of coverage of the data. We transform the original $m$-class classification problem to (m+1)-class where the (m+1)-th class represents the model absta...
[]
null
69
1907.00208
title_snapshot
[ -0.04038172960281372, -0.012322955764830112, -0.0072513725608587265, 0.03434743732213974, 0.05638841167092323, 0.011600173078477383, 0.025823716074228287, 0.006631223019212484, -0.015116224996745586, -0.03258030116558075, 0.005828654393553734, 0.07333660870790482, -0.0715433731675148, -0.0...
DRUM: End-To-End Differentiable Rule Mining On Knowledge Graphs
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0c72cb7ee1512f800abe27823a792d03-Abstract.html
[ "Ali Sadeghian", "Mohammadreza Armandpour", "Patrick Ding", "Daisy Zhe Wang" ]
null
null
In this paper, we study the problem of learning probabilistic logical rules for inductive and interpretable link prediction. Despite the importance of inductive link prediction, most previous works focused on transductive link prediction and cannot manage previously unseen entities. Moreover, they are black-box models ...
[]
null
70
1911.00055
title_snapshot
[ 0.025181347504258156, -0.01399488840252161, -0.004223277326673269, 0.03132762387394905, 0.0468856580555439, 0.001548868604004383, 0.026228750124573708, -0.019317707046866417, -0.012145377695560455, -0.018318956717848778, 0.0071433852426707745, 0.011520552448928356, -0.07723475247621536, 0....
Combinatorial Inference against Label Noise
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0cb929eae7a499e50248a3a78f7acfc7-Abstract.html
[ "Paul Hongsuck Seo", "Geeho Kim", "Bohyung Han" ]
null
null
Label noise is one of the critical sources that degrade generalization performance of deep neural networks significantly. To handle the label noise issue in a principled way, we propose a unique classification framework of constructing multiple models in heterogeneous coarse-grained meta-class spaces and making joint i...
[]
null
71
null
null
[ 0.01779779978096485, -0.011899949982762337, -0.025637410581111908, 0.055642761290073395, 0.02765406295657158, 0.043203216046094894, 0.007904174737632275, -0.026500124484300613, -0.04354255273938179, -0.038412369787693024, 0.0021449539344757795, 0.02722313441336155, -0.08539381623268127, 0....
Localized Structured Prediction
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0cbed40c0d920b94126eaf5e707be1f5-Abstract.html
[ "Carlo Ciliberto", "Francis Bach", "Alessandro Rudi" ]
null
null
Key to structured prediction is exploiting the problem's structure to simplify the learning process. A major challenge arises when data exhibit a local structure (i.e., are made ``by parts'') that can be leveraged to better approximate the relation between (parts of) the input and (parts of) the output. Recent literatu...
[]
null
72
1806.02402
title_snapshot
[ 0.026946451514959335, -0.02390648052096367, 0.0027318638749420643, 0.040617723017930984, 0.046901725232601166, 0.032101716846227646, -0.006254249718040228, -0.006235487759113312, -0.024270514026284218, -0.0203451756387949, -0.004490665625780821, 0.011028851382434368, -0.06967734545469284, ...
Fast Low-rank Metric Learning for Large-scale and High-dimensional Data
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0d0fd7c6e093f7b804fa0150b875b868-Abstract.html
[ "Han Liu", "Zhizhong Han", "Yu-Shen Liu", "Ming Gu" ]
null
null
Low-rank metric learning aims to learn better discrimination of data subject to low-rank constraints. It keeps the intrinsic low-rank structure of datasets and reduces the time cost and memory usage in metric learning. However, it is still a challenge for current methods to handle datasets with both high dimensions and...
[]
null
73
1909.06297
title_snapshot
[ -0.01145261526107788, -0.015393277630209923, 0.008415552787482738, 0.016650095582008362, 0.011585489846765995, 0.016689199954271317, 0.015345560386776924, -0.000011362111763446592, -0.01455334946513176, -0.02763104997575283, -0.0282383281737566, -0.015643615275621414, -0.07785511016845703, ...
Wide Neural Networks of Any Depth Evolve as Linear Models Under Gradient Descent
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0d1a9651497a38d8b1c3871c84528bd4-Abstract.html
[ "Jaehoon Lee", "Lechao Xiao", "Samuel Schoenholz", "Yasaman Bahri", "Roman Novak", "Jascha Sohl-Dickstein", "Jeffrey Pennington" ]
null
null
A longstanding goal in deep learning research has been to precisely characterize training and generalization. However, the often complex loss landscapes of neural networks have made a theory of learning dynamics elusive. In this work, we show that for wide neural networks the learning dynamics simplify considerably and...
[]
null
74
1902.06720
title_snapshot
[ -0.016781199723482132, -0.009951756335794926, -0.0016256204107776284, 0.010457617230713367, 0.03404783830046654, 0.016132555902004242, 0.012564064934849739, 0.027589624747633934, -0.022737421095371246, -0.031204938888549805, 0.0071618324145674706, 0.018895309418439865, -0.061690133064985275,...
Retrosynthesis Prediction with Conditional Graph Logic Network
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0d2b2061826a5df3221116a5085a6052-Abstract.html
[ "Hanjun Dai", "Chengtao Li", "Connor Coley", "Bo Dai", "Le Song" ]
null
null
Retrosynthesis is one of the fundamental problems in organic chemistry. The task is to identify reactants that can be used to synthesize a specified product molecule. Recently, computer-aided retrosynthesis is finding renewed interest from both chemistry and computer science communities. Most existing approaches rely o...
[]
null
75
2001.01408
title_snapshot
[ -0.020722752436995506, 0.013871144503355026, -0.019482042640447617, 0.0374164842069149, 0.059442684054374695, -0.00420762924477458, -0.0002238064625998959, 0.014077729545533657, 0.006881227716803551, -0.035089313983917236, 0.026380328461527824, 0.012037025764584541, -0.05332643911242485, 0...
Efficient Pure Exploration in Adaptive Round model
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0d441de75945e5acbc865406fc9a2559-Abstract.html
[ "Tianyuan Jin", "Jieming SHI", "Xiaokui Xiao", "Enhong Chen" ]
null
null
In the adaptive setting, many multi-armed bandit applications allow the learner to adaptively draw samples and adjust sampling strategy in rounds. In many real applications, not only the query complexity but also the round complexity need to be optimized. In this paper, we study both PAC and exact top-$k$ arm identific...
[]
null
76
null
null
[ -0.045082103461027145, -0.009193316102027893, 0.01154420431703329, 0.040647536516189575, 0.044926248490810394, 0.03772241249680519, 0.034900564700365067, -0.0116215068846941, -0.04063326120376587, -0.04211262986063957, -0.011465009301900864, -0.01665765978395939, -0.04761962220072746, -0.0...
Unsupervised Emergence of Egocentric Spatial Structure from Sensorimotor Prediction
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0dd1bc593a91620daecf7723d2235624-Abstract.html
[ "Alban Laflaquière", "Michael Garcia Ortiz" ]
null
null
Despite its omnipresence in robotics application, the nature of spatial knowledge and the mechanisms that underlie its emergence in autonomous agents are still poorly understood. Recent theoretical works suggest that the Euclidean structure of space induces invariants in an agent’s raw sensorimotor experience. We hypot...
[]
null
77
1906.01401
title_snapshot
[ 0.00171696359757334, 0.005784991662949324, -0.02031700313091278, -0.0035077030770480633, 0.03430937975645065, 0.009583103470504284, 0.04569586366415024, 0.013557546772062778, -0.04454890266060829, -0.03321265056729317, -0.021484611555933952, -0.00430826423689723, -0.0652516633272171, -0.02...
Adversarial Robustness through Local Linearization
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0defd533d51ed0a10c5c9dbf93ee78a5-Abstract.html
[ "Chongli Qin", "James Martens", "Sven Gowal", "Dilip Krishnan", "Krishnamurthy Dvijotham", "Alhussein Fawzi", "Soham De", "Robert Stanforth", "Pushmeet Kohli" ]
null
null
Adversarial training is an effective methodology for training deep neural networks that are robust against adversarial, norm-bounded perturbations. However, the computational cost of adversarial training grows prohibitively as the size of the model and number of input dimensions increase. Further, training against less...
[]
null
78
1907.02610
title_snapshot
[ 0.01314968429505825, -0.023515138775110245, -0.00007914521847851574, 0.052952274680137634, 0.026084240525960922, 0.007800693158060312, 0.03501950204372406, -0.027412336319684982, -0.018146300688385963, -0.05536268651485443, -0.0009404701995663345, -0.02519463747739792, -0.05898139998316765, ...
Generalized Off-Policy Actor-Critic
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0e095e054ee94774d6a496099eb1cf6a-Abstract.html
[ "Shangtong Zhang", "Wendelin Boehmer", "Shimon Whiteson" ]
null
null
We propose a new objective, the counterfactual objective, unifying existing objectives for off-policy policy gradient algorithms in the continuing reinforcement learning (RL) setting. Compared to the commonly used excursion objective, which can be misleading about the performance of the target policy when deployed, our...
[]
null
79
1903.11329
title_snapshot
[ -0.018187114968895912, -0.04104721546173096, 0.023759830743074417, 0.02439456805586815, 0.016433078795671463, 0.03301393240690231, 0.01264228392392397, 0.014412498101592064, -0.061819493770599365, -0.012384491972625256, -0.004829144570976496, 0.052086468786001205, -0.09648799896240234, -0....
Average Individual Fairness: Algorithms, Generalization and Experiments
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0e1feae55e360ff05fef58199b3fa521-Abstract.html
[ "Saeed Sharifi-Malvajerdi", "Michael Kearns", "Aaron Roth" ]
null
null
We propose a new family of fairness definitions for classification problems that combine some of the best properties of both statistical and individual notions of fairness. We posit not only a distribution over individuals, but also a distribution over (or collection of) classification tasks. We then ask that standard ...
[]
null
80
1905.10607
title_snapshot
[ 0.0028678697999566793, 0.0007705476600676775, -0.00788789615035057, 0.05085277929902077, 0.016627248376607895, 0.03737635165452957, 0.023310577496886253, -0.0015886644832789898, -0.04178021103143692, -0.03581812605261803, -0.016948334872722626, -0.00018116964201908559, -0.09067996591329575, ...
Comparing distributions: $\ell_1$ geometry improves kernel two-sample testing
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0e2db0cb2c4645904a054261104b7a14-Abstract.html
[ "meyer scetbon", "Gael Varoquaux" ]
null
null
Are two sets of observations drawn from the same distribution? This problem is a two-sample test. Kernel methods lead to many appealing properties. Indeed state-of-the-art approaches use the $L^2$ distance between kernel-based distribution representatives to derive their test statistics. Here, we show that $L^p$ distan...
[]
null
81
1909.09264
title_snapshot
[ -0.017192671075463295, -0.012980467639863491, 0.006181843113154173, 0.05209251120686531, 0.04425325617194176, 0.035011861473321915, 0.010855866596102715, -0.00415899557992816, -0.00008731616981094703, -0.05695043504238129, 0.004321960732340813, -0.004295794293284416, -0.06527415663003922, ...
Nonstochastic Multiarmed Bandits with Unrestricted Delays
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0e4f5cc9f4f3f7f1651a6b9f9214e5b1-Abstract.html
[ "Tobias Sommer Thune", "Nicolò Cesa-Bianchi", "Yevgeny Seldin" ]
null
null
We investigate multiarmed bandits with delayed feedback, where the delays need neither be identical nor bounded. We first prove that "delayed" Exp3 achieves the $O(\sqrt{(KT + D)\ln K})$ regret bound conjectured by Cesa-Bianchi et al. [2016] in the case of variable, but bounded delays. Here, $K$ is the number of action...
[]
null
82
1906.00670
title_snapshot
[ -0.018638798967003822, -0.02384742721915245, -0.02840910293161869, 0.040855202823877335, 0.03599915653467178, 0.03591621667146683, 0.035182323306798935, -0.006653240416198969, -0.03436645120382309, -0.060230620205402374, -0.010729679837822914, 0.010042410343885422, -0.04192746803164482, -0...
Approximate Bayesian Inference for a Mechanistic Model of Vesicle Release at a Ribbon Synapse
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0e57098d0318a954d1443e2974a38fac-Abstract.html
[ "Cornelius Schröder", "Ben James", "Leon Lagnado", "Philipp Berens" ]
null
null
The inherent noise of neural systems makes it difficult to construct models which accurately capture experimental measurements of their activity. While much research has been done on how to efficiently model neural activity with descriptive models such as linear-nonlinear-models (LN), Bayesian inference for mechanistic...
[]
null
83
null
null
[ -0.02765500359237194, 0.034190356731414795, -0.009533152915537357, 0.011560700833797455, 0.02936789020895958, 0.01859736256301403, 0.031075557693839073, -0.01134788803756237, -0.06085507571697235, -0.02965073473751545, 0.010707338340580463, -0.007607103791087866, -0.051714491099119186, 0.0...
Data-dependent Sample Complexity of Deep Neural Networks via Lipschitz Augmentation
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0e79548081b4bd0df3c77c5ba2c23289-Abstract.html
[ "Colin Wei", "Tengyu Ma" ]
null
null
Existing Rademacher complexity bounds for neural networks rely only on norm control of the weight matrices and depend exponentially on depth via a product of the matrix norms. Lower bounds show that this exponential dependence on depth is unavoidable when no additional properties of the training data are considered. We...
[]
null
84
1905.03684
title_snapshot
[ -0.05863592028617859, -0.006992656737565994, -0.00034143784432671964, 0.047891080379486084, 0.043521441519260406, 0.0406024344265461, 0.053027279675006866, -0.026194103062152863, -0.015772895887494087, -0.035446636378765106, -0.01029293518513441, 0.00810167659074068, -0.06097903102636337, ...
Semi-supervisedly Co-embedding Attributed Networks
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0e7c7d6c41c76b9ee6445ae01cc0181d-Abstract.html
[ "Zaiqiao Meng", "Shangsong Liang", "Jinyuan Fang", "Teng Xiao" ]
null
null
Deep generative models (DGMs) have achieved remarkable advances. Semi-supervised variational auto-encoders (SVAE) as a classical DGM offers a principled framework to effective generalize from small labelled data to large unlabelled ones, but it is difficult to incorporate rich unstructured relationships within the mult...
[]
null
85
1910.14491
title_snapshot
[ 0.026656651869416237, -0.0572112612426281, 0.007963715121150017, 0.04440879821777344, 0.015324450097978115, 0.02267775498330593, 0.028230611234903336, -0.018144354224205017, 0.01551283523440361, -0.030490033328533173, -0.024069668725132942, -0.008474620059132576, -0.05122369900345802, 0.02...
Adaptive Auxiliary Task Weighting for Reinforcement Learning
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0e900ad84f63618452210ab8baae0218-Abstract.html
[ "Xingyu Lin", "Harjatin Baweja", "George Kantor", "David Held" ]
null
null
Reinforcement learning is known to be sample inefficient, preventing its application to many real-world problems, especially with high dimensional observations like images. Transferring knowledge from other auxiliary tasks is a powerful tool for improving the learning efficiency. However, the usage of auxiliary tasks h...
[]
null
86
null
null
[ -0.03168996050953865, -0.035749129951000214, 0.0028176039922982454, 0.03563053905963898, 0.03179185837507248, 0.04386332258582115, 0.0007161011453717947, -0.02076735720038414, -0.04743858426809311, -0.050388313829898834, -0.05319596081972122, 0.03427743539214134, -0.07675162702798843, -0.0...
Continuous Hierarchical Representations with Poincaré Variational Auto-Encoders
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0ec04cb3912c4f08874dd03716f80df1-Abstract.html
[ "Emile Mathieu", "Charline Le Lan", "Chris J. Maddison", "Ryota Tomioka", "Yee Whye Teh" ]
null
null
The Variational Auto-Encoder (VAE) is a popular method for learning a generative model and embeddings of the data. Many real datasets are hierarchically structured. However, traditional VAEs map data in a Euclidean latent space which cannot efficiently embed tree-like structures. Hyperbolic spaces with negative curvatu...
[]
null
87
1901.06033
title_snapshot
[ 0.008082509972155094, 0.00987868569791317, 0.0216765608638525, 0.03557266294956207, 0.0174850020557642, 0.05044207349419594, 0.04111915081739426, -0.005208456423133612, -0.03787067160010338, -0.06023028492927551, -0.04605875536799431, -0.01687793992459774, -0.049271926283836365, 0.02190932...
Training Image Estimators without Image Ground Truth
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0ed9422357395a0d4879191c66f4faa2-Abstract.html
[ "Zhihao Xia", "Ayan Chakrabarti" ]
null
null
Deep neural networks have been very successful in compressive-sensing and image restoration applications, as a means to estimate images from partial, blurry, or otherwise degraded measurements. These networks are trained on a large number of corresponding pairs of measurements and ground-truth images, and thus implicit...
[]
null
88
1906.05775
title_snapshot
[ 0.02159079909324646, -0.011433027684688568, -0.0027258950285613537, 0.05555132403969765, 0.051455944776535034, 0.024500321596860886, 0.017793988808989525, 0.024207284674048424, -0.032137151807546616, -0.05952118709683418, -0.01888054423034191, 0.03065529465675354, -0.06332123279571533, -0....
On the Convergence Rate of Training Recurrent Neural Networks
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0ee8b85a85a49346fdff9665312a5cc4-Abstract.html
[ "Zeyuan Allen-Zhu", "Yuanzhi Li", "Zhao Song" ]
null
null
How can local-search methods such as stochastic gradient descent (SGD) avoid bad local minima in training multi-layer neural networks? Why can they fit random labels even given non-convex and non-smooth architectures? Most existing theory only covers networks with one hidden layer, so can we go deeper? In this paper, w...
[]
null
89
1810.12065
title_snapshot
[ -0.034775443375110626, -0.028974752873182297, -0.0012215310707688332, 0.0400470569729805, 0.03690498322248459, 0.05856745317578316, 0.031276948750019073, 0.019377730786800385, -0.03214896470308304, -0.025575829669833183, -0.0007717043627053499, -0.00027106344350613654, -0.055791258811950684,...
Minimizers of the Empirical Risk and Risk Monotonicity
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0f9cafd014db7a619ddb4276af0d692c-Abstract.html
[ "Marco Loog", "Tom Viering", "Alexander Mey" ]
null
null
Plotting a learner's average performance against the number of training samples results in a learning curve. Studying such curves on one or more data sets is a way to get to a better understanding of the generalization properties of this learner. The behavior of learning curves is, however, not very well understood and...
[]
null
90
1907.05476
title_snapshot
[ -0.041374001652002335, -0.011442397721111774, -0.026208503171801567, 0.05843345448374748, 0.02718382515013218, 0.03893003240227699, 0.016809523105621338, -0.008916943334043026, -0.049768559634685516, -0.028238264843821526, -0.010002742521464825, 0.030548496171832085, -0.052499495446681976, ...
Factor Group-Sparse Regularization for Efficient Low-Rank Matrix Recovery
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0fc170ecbb8ff1afb2c6de48ea5343e7-Abstract.html
[ "Jicong Fan", "Lijun Ding", "Yudong Chen", "Madeleine Udell" ]
null
null
This paper develops a new class of nonconvex regularizers for low-rank matrix recovery. Many regularizers are motivated as convex relaxations of the \emph{matrix rank} function. Our new factor group-sparse regularizers are motivated as a relaxation of the \emph{number of nonzero columns} in a factorization of the matri...
[]
null
91
1911.05774
title_snapshot
[ -0.0230963546782732, -0.030297469347715378, 0.05952649563550949, 0.010522059164941311, 0.01679294928908348, 0.015958299860358238, 0.025975501164793968, -0.0020062795374542475, -0.034455254673957825, -0.05195702984929085, -0.009483400732278824, -0.00992339476943016, -0.0405803881585598, 0.0...
Möbius Transformation for Fast Inner Product Search on Graph
https://proceedings.neurips.cc/paper_files/paper/2019/hash/0fd7e4f42a8b4b4ef33394d35212b13e-Abstract.html
[ "Zhixin Zhou", "Shulong Tan", "Zhaozhuo Xu", "Ping Li" ]
null
null
We present a fast search on graph algorithm for Maximum Inner Product Search (MIPS). This optimization problem is challenging since traditional Approximate Nearest Neighbor (ANN) search methods may not perform efficiently in the non-metric similarity measure. Our proposed method is based on the property that Möbius tra...
[]
null
92
null
null
[ -0.06819097697734833, 0.01619577407836914, 0.01342061161994934, 0.020714661106467247, 0.03714966028928757, 0.046775467693805695, 0.021727479994297028, 0.005694805644452572, 0.012571594677865505, -0.0715896487236023, 0.0010752540547400713, -0.04718886688351631, -0.053887125104665756, 0.0161...
The Label Complexity of Active Learning from Observational Data
https://proceedings.neurips.cc/paper_files/paper/2019/hash/1019c8091693ef5c5f55970346633f92-Abstract.html
[ "Songbai Yan", "Kamalika Chaudhuri", "Tara Javidi" ]
null
null
Counterfactual learning from observational data involves learning a classifier on an entire population based on data that is observed conditioned on a selection policy. This work considers this problem in an active setting, where the learner additionally has access to unlabeled examples and can choose to get a subset o...
[]
null
93
1905.12791
title_snapshot
[ -0.027236947789788246, -0.022814592346549034, -0.011621735990047455, 0.022725841030478477, 0.03631896898150444, 0.013927098363637924, 0.006072629243135452, -0.017810162156820297, -0.030457470566034317, -0.017803965136408806, -0.01768265850841999, 0.036463309079408646, -0.0826011523604393, ...
Hyperbolic Graph Neural Networks
https://proceedings.neurips.cc/paper_files/paper/2019/hash/103303dd56a731e377d01f6a37badae3-Abstract.html
[ "Qi Liu", "Maximilian Nickel", "Douwe Kiela" ]
null
null
Learning from graph-structured data is an important task in machine learning and artificial intelligence, for which Graph Neural Networks (GNNs) have shown great promise. Motivated by recent advances in geometric representation learning, we propose a novel GNN architecture for learning representations on Riemannian man...
[]
null
94
1910.12892
title_snapshot
[ -0.02042151242494583, -0.0021346628200262785, 0.02324095368385315, 0.05524514615535736, 0.03338766470551491, 0.029424773529171944, 0.007394867949187756, 0.02603890746831894, -0.032299403101205826, -0.06382623314857483, 0.016656560823321342, -0.011146306991577148, -0.06957639008760452, 0.03...
Learning Fairness in Multi-Agent Systems
https://proceedings.neurips.cc/paper_files/paper/2019/hash/10493aa88605cad5ab4752b04a63d172-Abstract.html
[ "Jiechuan Jiang", "Zongqing Lu" ]
null
null
Fairness is essential for human society, contributing to stability and productivity. Similarly, fairness is also the key for many multi-agent systems. Taking fairness into multi-agent learning could help multi-agent systems become both efficient and stable. However, learning efficiency and fairness simultaneously is a ...
[]
null
95
1910.14472
title_snapshot
[ -0.03736176714301109, -0.02078641764819622, 0.0003068175574298948, 0.03903621807694435, 0.012308867648243904, 0.010545237921178341, 0.0008867760188877583, 0.0008330729906447232, -0.03537723794579506, -0.06229163333773613, 0.006428669672459364, 0.03473278135061264, -0.07534723728895187, -0....
On Robustness to Adversarial Examples and Polynomial Optimization
https://proceedings.neurips.cc/paper_files/paper/2019/hash/107878346e1d8f8fe6af7a7a588aa807-Abstract.html
[ "Pranjal Awasthi", "Abhratanu Dutta", "Aravindan Vijayaraghavan" ]
null
null
We study the design of computationally efficient algorithms with provable guarantees, that are robust to adversarial (test time) perturbations. While there has been an explosion of recent work on this topic due to its connections to test time robustness of deep networks, there is limited theoretical understanding of se...
[]
null
96
1911.04681
title_snapshot
[ -0.022822635248303413, -0.016266612336039543, -0.0079579409211874, 0.062301408499479294, 0.029963677749037743, 0.007494826801121235, 0.02995145320892334, -0.016225481405854225, 0.0006373855285346508, -0.030329497531056404, -0.008220884948968887, -0.011166623793542385, -0.0442865751683712, ...
In-Place Zero-Space Memory Protection for CNN
https://proceedings.neurips.cc/paper_files/paper/2019/hash/1091660f3dff84fd648efe31391c5524-Abstract.html
[ "Hui Guan", "Lin Ning", "Zhen Lin", "Xipeng Shen", "Huiyang Zhou", "Seung-Hwan Lim" ]
null
null
Convolutional Neural Networks (CNN) are being actively explored for safety-critical applications such as autonomous vehicles and aerospace, where it is essential to ensure the reliability of inference results in the presence of possible memory faults. Traditional methods such as error correction codes (ECC) and Triple ...
[]
null
97
1910.14479
title_snapshot
[ 0.0036730256397277117, -0.001212941249832511, -0.023101694881916046, 0.06582151353359222, 0.04588073864579201, 0.02455499768257141, 0.010586287826299667, 0.027013244107365608, -0.045010536909103394, -0.039791304618120193, 0.008034329861402512, -0.03008958138525486, -0.06285811215639114, -0...
Non-Asymptotic Gap-Dependent Regret Bounds for Tabular MDPs
https://proceedings.neurips.cc/paper_files/paper/2019/hash/10a5ab2db37feedfdeaab192ead4ac0e-Abstract.html
[ "Max Simchowitz", "Kevin G. Jamieson" ]
null
null
This paper establishes that optimistic algorithms attain gap-dependent and non-asymptotic logarithmic regret for episodic MDPs. In contrast to prior work, our bounds do not suffer a dependence on diameter-like quantities or ergodicity, and smoothly interpolate between the gap dependent logarithmic-regret, and the $\wid...
[]
null
98
1905.03814
title_snapshot
[ -0.056343622505664825, 0.0051739998161792755, -0.014905979856848717, 0.06608361005783081, 0.0546722486615181, 0.03385021165013313, 0.02764977142214775, 0.020213115960359573, -0.03718045353889465, -0.04406922310590744, -0.005022353958338499, -0.004291309509426355, -0.08953290432691574, -0.0...
Discovery of Useful Questions as Auxiliary Tasks
https://proceedings.neurips.cc/paper_files/paper/2019/hash/10ff0b5e85e5b85cc3095d431d8c08b4-Abstract.html
[ "Vivek Veeriah", "Matteo Hessel", "Zhongwen Xu", "Janarthanan Rajendran", "Richard L. Lewis", "Junhyuk Oh", "Hado P van Hasselt", "David Silver", "Satinder Singh" ]
null
null
Arguably, intelligent agents ought to be able to discover their own questions so that in learning answers for them they learn unanticipated useful knowledge and skills; this departs from the focus in much of machine learning on agents learning answers to externally defined questions. We present a novel method for a rei...
[]
null
99
1909.04607
title_snapshot
[ -0.0037875534035265446, -0.041214246302843094, 0.007925551384687424, 0.0604570209980011, 0.027878792956471443, 0.03479836508631706, 0.021085644140839577, -0.013137760572135448, -0.03251263126730919, -0.018750961869955063, -0.0277018491178751, 0.059658098965883255, -0.06488971412181854, -0....
Sequential Neural Processes
https://proceedings.neurips.cc/paper_files/paper/2019/hash/110209d8fae7417509ba71ad97c17639-Abstract.html
[ "Gautam Singh", "Jaesik Yoon", "Youngsung Son", "Sungjin Ahn" ]
null
null
Neural Processes combine the strengths of neural networks and Gaussian processes to achieve both flexible learning and fast prediction in stochastic processes. However, a large class of problems comprise underlying temporal dependency structures in a sequence of stochastic processes that Neural Processes (NP) do not ex...
[]
null
100
1906.10264
title_snapshot
[ 0.0033743625972419977, 0.024421870708465576, -0.0018553182017058134, 0.017088647931814194, 0.027481261640787125, 0.03418789058923721, 0.007181502878665924, 0.03778151422739029, -0.024400049820542336, -0.030245114117860794, -0.022566143423318863, -0.007231403607875109, -0.0467311292886734, ...