title
stringlengths
19
143
paper_url
stringlengths
41
43
authors
listlengths
1
23
type
stringclasses
3 values
primary_area
stringclasses
0 values
abstract
large_stringlengths
457
2.38k
keywords
listlengths
0
21
TL;DR
large_stringclasses
0 values
submission_number
int64
1
3.83k
arxiv_id
stringlengths
10
10
arxiv_id_source
stringclasses
2 values
embedding
listlengths
768
768
Net-DNF: Effective Deep Modeling of Tabular Data
https://openreview.net/forum?id=73WTGs96kho
[ "Liran Katzir", "Gal Elidan", "Ran El-Yaniv" ]
Poster
null
A challenging open question in deep learning is how to handle tabular data. Unlike domains such as image and natural language processing, where deep architectures prevail, there is still no widely accepted neural architecture that dominates tabular data. As a step toward bridging this gap, we present Net-DNF a novel ge...
[ "Neural Networks", "Architectures", "Tabular Data", "Predictive Modeling" ]
null
1,574
null
null
[ -0.016373762860894203, -0.027831831946969032, -0.006840958725661039, 0.051904767751693726, 0.052276212722063065, -0.001103995949961245, -0.01597399078309536, -0.011720962822437286, -0.02807473950088024, -0.040762584656476974, 0.011461503803730011, -0.00118657061830163, -0.07686426490545273, ...
Predicting Inductive Biases of Pre-Trained Models
https://openreview.net/forum?id=mNtmhaDkAr
[ "Charles Lovering", "Rohan Jha", "Tal Linzen", "Ellie Pavlick" ]
Poster
null
Most current NLP systems are based on a pre-train-then-fine-tune paradigm, in which a large neural network is first trained in a self-supervised way designed to encourage the network to extract broadly-useful linguistic features, and then fine-tuned for a specific task of interest. Recent work attempts to understand wh...
[ "information-theoretical probing", "probing", "challenge sets", "natural language processing" ]
null
3,826
null
null
[ -0.041504692286252975, -0.024559151381254196, -0.03274990990757942, 0.042177699506282806, 0.048905614763498306, 0.028540531173348427, 0.03351079300045967, 0.015673844143748283, -0.028104335069656372, -0.018578987568616867, -0.013722354546189308, 0.048581067472696304, -0.037727419286966324, ...
Optimism in Reinforcement Learning with Generalized Linear Function Approximation
https://openreview.net/forum?id=CBmJwzneppz
[ "Yining Wang", "Ruosong Wang", "Simon Shaolei Du", "Akshay Krishnamurthy" ]
Poster
null
We design a new provably efficient algorithm for episodic reinforcement learning with generalized linear function approximation. We analyze the algorithm under a new expressivity assumption that we call ``optimistic closure,'' which is strictly weaker than assumptions from prior analyses for the linear setting. With op...
[ "reinforcement learning", "optimism", "exploration", "function approximation", "theory", "regret analysis", "provable sample efficiency" ]
null
3,820
1912.04136
title_snapshot
[ -0.03719856217503548, -0.0020507273729890585, 0.024639572948217392, 0.03743778169155121, 0.02448270469903946, 0.03423384577035904, 0.01802157238125801, 0.017447255551815033, -0.030609330162405968, -0.022574499249458313, -0.022017396986484528, 0.011142074130475521, -0.08707273006439209, -0....
SCoRe: Pre-Training for Context Representation in Conversational Semantic Parsing
https://openreview.net/forum?id=oyZxhRI2RiE
[ "Tao Yu", "Rui Zhang", "Alex Polozov", "Christopher Meek", "Ahmed Hassan Awadallah" ]
Poster
null
Conversational Semantic Parsing (CSP) is the task of converting a sequence of natural language queries to formal language (e.g., SQL, SPARQL) that can be executed against a structured ontology (e.g. databases, knowledge bases). To accomplish this task, a CSP system needs to model the relation between the ...
[]
null
3,773
null
null
[ -0.005969882011413574, -0.012070818804204464, -0.005474647041410208, 0.037013940513134, 0.05538944527506828, 0.014551641419529915, 0.007138388231396675, 0.014812460169196129, 0.007725202944129705, -0.011620452627539635, -0.031187763437628746, 0.03154043108224869, -0.04080461710691452, 0.00...
A teacher-student framework to distill future trajectories
https://openreview.net/forum?id=ECuvULjFQia
[ "Alexander Neitz", "Giambattista Parascandolo", "Bernhard Schölkopf" ]
Poster
null
By learning to predict trajectories of dynamical systems, model-based methods can make extensive use of all observations from past experience. However, due to partial observability, stochasticity, compounding errors, and irrelevant dynamics, training to predict observations explicitly often results in poor models. Mode...
[ "meta-learning", "privileged information" ]
null
3,758
null
null
[ 0.0004875066224485636, -0.020398393273353577, 0.016171976923942566, 0.05223361775279045, 0.05110407620668411, -0.015469873324036598, 0.034996192902326584, 0.022610880434513092, -0.028710925951600075, -0.021938025951385498, 0.009902219288051128, 0.006432092748582363, -0.06087557226419449, -...
Certify or Predict: Boosting Certified Robustness with Compositional Architectures
https://openreview.net/forum?id=USCNapootw
[ "Mark Niklas Mueller", "Mislav Balunovic", "Martin Vechev" ]
Poster
null
A core challenge with existing certified defense mechanisms is that while they improve certified robustness, they also tend to drastically decrease natural accuracy, making it difficult to use these methods in practice. In this work, we propose a new architecture which addresses this challenge and enables one to boost ...
[ "Provable Robustness", "Network Architecture", "Robustness", "Adversarial Accuracy", "Certified Robustness" ]
null
3,751
null
null
[ 0.019359620288014412, -0.025747399777173996, -0.009083477780222893, 0.0565095953643322, 0.047875963151454926, 0.014403064735233784, 0.040992382913827896, -0.023026518523693085, -0.008816513232886791, -0.018654154613614082, 0.004606868140399456, 0.0010471675777807832, -0.054870445281267166, ...
On the Transfer of Disentangled Representations in Realistic Settings
https://openreview.net/forum?id=8VXvj1QNRl1
[ "Andrea Dittadi", "Frederik Träuble", "Francesco Locatello", "Manuel Wuthrich", "Vaibhav Agrawal", "Ole Winther", "Stefan Bauer", "Bernhard Schölkopf" ]
Poster
null
Learning meaningful representations that disentangle the underlying structure of the data generating process is considered to be of key importance in machine learning. While disentangled representations were found to be useful for diverse tasks such as abstract reasoning and fair classification, their scalability and r...
[ "representation learning", "disentanglement", "real-world" ]
null
3,746
2010.14407
title_snapshot
[ -0.011436525732278824, -0.017934395000338554, -0.0009040053118951619, 0.05002399533987045, 0.03817461058497429, -0.014136756770312786, 0.026608368381857872, 0.01588096283376217, -0.027613531798124313, -0.016441311687231064, -0.036367494612932205, -0.025966273620724678, -0.09115331619977951, ...
Robust Reinforcement Learning on State Observations with Learned Optimal Adversary
https://openreview.net/forum?id=sCZbhBvqQaU
[ "Huan Zhang", "Hongge Chen", "Duane S Boning", "Cho-Jui Hsieh" ]
Poster
null
We study the robustness of reinforcement learning (RL) with adversarially perturbed state observations, which aligns with the setting of many adversarial attacks to deep reinforcement learning (DRL) and is also important for rolling out real-world RL agent under unpredictable sensing noise. With a fixed agent policy, w...
[ "reinforcement learning", "robustness", "adversarial attacks", "adversarial defense" ]
null
3,741
2101.08452
title_snapshot
[ -0.01799892820417881, -0.02330268733203411, -0.008876688778400421, 0.05365290120244026, 0.0225211251527071, 0.003905362682417035, 0.04567841812968254, -0.00926560815423727, -0.019155820831656456, -0.03805020824074745, -0.016440559178590775, -0.010811691172420979, -0.08598136901855469, -0.0...
Practical Real Time Recurrent Learning with a Sparse Approximation
https://openreview.net/forum?id=q3KSThy2GwB
[ "Jacob Menick", "Erich Elsen", "Utku Evci", "Simon Osindero", "Karen Simonyan", "Alex Graves" ]
Spotlight
null
Recurrent neural networks are usually trained with backpropagation through time, which requires storing a complete history of network states, and prohibits updating the weights "online" (after every timestep). Real Time Recurrent Learning (RTRL) eliminates the need for history storage and allows for online weight updat...
[ "recurrent neural networks", "backpropagation", "biologically plausible", "forward mode", "real time recurrent learning", "rtrl", "bptt" ]
null
3,731
2006.07232
title_judge
[ -0.015014496631920338, -0.04429089277982712, 0.011273454874753952, 0.03204382210969925, 0.03842123970389366, 0.034894201904535294, 0.014314806088805199, 0.039932526648044586, -0.05757788196206093, -0.02317187562584877, 0.005213298834860325, 0.0017265059286728501, -0.07246686518192291, -0.0...
Retrieval-Augmented Generation for Code Summarization via Hybrid GNN
https://openreview.net/forum?id=zv-typ1gPxA
[ "Shangqing Liu", "Yu Chen", "Xiaofei Xie", "Jing Kai Siow", "Yang Liu" ]
Spotlight
null
Source code summarization aims to generate natural language summaries from structured code snippets for better understanding code functionalities. However, automatic code summarization is challenging due to the complexity of the source code and the language gap between the source code and natural language summaries. Mo...
[ "Code Summarization", "Graph Neural Network", "Retrieval", "Generation" ]
null
3,719
2006.05405
title_snapshot
[ -0.008924927562475204, -0.059218570590019226, -0.03565578907728195, 0.04624049365520477, 0.014577331952750683, 0.018286261707544327, 0.02446441538631916, 0.03342185914516449, -0.03082849085330963, -0.04051067680120468, -0.04862041771411896, 0.0055044242180883884, -0.07115976512432098, -0.0...
Learning from others' mistakes: Avoiding dataset biases without modeling them
https://openreview.net/forum?id=Hf3qXoiNkR
[ "Victor Sanh", "Thomas Wolf", "Yonatan Belinkov", "Alexander M Rush" ]
Poster
null
State-of-the-art natural language processing (NLP) models often learn to model dataset biases and surface form correlations instead of features that target the intended underlying task. Previous work has demonstrated effective methods to circumvent these issues when knowledge of the bias is available. We consider cases...
[ "dataset bias", "product of experts", "natural language processing" ]
null
3,718
2012.01300
title_snapshot
[ -0.011456144973635674, -0.017333073541522026, -0.05010826140642166, 0.049757469445466995, 0.035521458834409714, 0.009430734440684319, 0.03898055851459503, 0.013163039460778236, -0.005222110077738762, -0.016154401004314423, -0.02800501510500908, 0.044609520584344864, -0.07758931070566177, -...
Fast and Complete: Enabling Complete Neural Network Verification with Rapid and Massively Parallel Incomplete Verifiers
https://openreview.net/forum?id=nVZtXBI6LNn
[ "Kaidi Xu", "Huan Zhang", "Shiqi Wang", "Yihan Wang", "Suman Jana", "Xue Lin", "Cho-Jui Hsieh" ]
Poster
null
Formal verification of neural networks (NNs) is a challenging and important problem. Existing efficient complete solvers typically require the branch-and-bound (BaB) process, which splits the problem domain into sub-domains and solves each sub-domain using faster but weaker incomplete verifiers, such as Linear Programm...
[ "neural network verification", "branch and bound" ]
null
3,717
2011.13824
title_snapshot
[ -0.015460497699677944, -0.002680424600839615, -0.023685770109295845, 0.03439273685216904, 0.033345822244882584, 0.028357846662402153, -0.006793871987611055, -0.004273895174264908, -0.02929755486547947, -0.01435816939920187, 0.003597547532990575, -0.006517044734209776, -0.04873882606625557, ...
Self-supervised Adversarial Robustness for the Low-label, High-data Regime
https://openreview.net/forum?id=bgQek2O63w
[ "Sven Gowal", "Po-Sen Huang", "Aaron van den Oord", "Timothy Mann", "Pushmeet Kohli" ]
Poster
null
Recent work discovered that training models to be invariant to adversarial perturbations requires substantially larger datasets than those required for standard classification. Perhaps more surprisingly, these larger datasets can be "mostly" unlabeled. Pseudo-labeling, a technique simultaneously pioneered by four separ...
[ "self-supervised", "adversarial training", "robustness" ]
null
3,713
null
null
[ -0.0004883683868683875, -0.03155262768268585, -0.0022540928330272436, 0.033476557582616806, 0.020937949419021606, 0.0016851827967911959, 0.04342913255095482, -0.02876659668982029, -0.016733327880501747, -0.030736586079001427, -0.022899117320775986, -0.015369204804301262, -0.06243455410003662...
Modeling the Second Player in Distributionally Robust Optimization
https://openreview.net/forum?id=ZDnzZrTqU9N
[ "Paul Michel", "Tatsunori Hashimoto", "Graham Neubig" ]
Poster
null
Distributionally robust optimization (DRO) provides a framework for training machine learning models that are able to perform well on a collection of related data distributions (the "uncertainty set"). This is done by solving a min-max game: the model is trained to minimize its maximum expected loss among all distribut...
[ "distributionally robust optimization", "deep learning", "robustness", "adversarial learning" ]
null
3,696
2103.10282
title_snapshot
[ -0.02808738499879837, -0.0014968998730182648, 0.0003711586759891361, 0.05306160822510719, 0.03107905387878418, 0.0412275530397892, 0.006291511934250593, -0.007838503457605839, -0.002345639979466796, -0.041211873292922974, -0.00997956097126007, 0.00998740829527378, -0.054649803787469864, -0...
Neural Jump Ordinary Differential Equations: Consistent Continuous-Time Prediction and Filtering
https://openreview.net/forum?id=JFKR3WqwyXR
[ "Calypso Herrera", "Florian Krach", "Josef Teichmann" ]
Poster
null
Combinations of neural ODEs with recurrent neural networks (RNN), like GRU-ODE-Bayes or ODE-RNN are well suited to model irregularly observed time series. While those models outperform existing discrete-time approaches, no theoretical guarantees for their predictive capabilities are available. Assuming that the irregul...
[ "Neural ODE", "conditional expectation", "irregular-observed data modelling" ]
null
3,695
2006.04727
title_snapshot
[ -0.05893096700310707, -0.026742566376924515, -0.02531055361032486, 0.05361597239971161, 0.03638941049575806, 0.05244385451078415, 0.02753625065088272, 0.014122464694082737, -0.02154015749692917, -0.02856205590069294, 0.026397323235869408, 0.01196011807769537, -0.07851559668779373, -0.01767...
Gradient Origin Networks
https://openreview.net/forum?id=0O_cQfw6uEh
[ "Sam Bond-Taylor", "Chris G. Willcocks" ]
Poster
null
This paper proposes a new type of generative model that is able to quickly learn a latent representation without an encoder. This is achieved using empirical Bayes to calculate the expectation of the posterior, which is implemented by initialising a latent vector with zeros, then using the gradient of the log-likelihoo...
[ "Deep Learning", "Generative Models", "Implicit Representation" ]
null
3,694
2007.02798
title_snapshot
[ -0.006673240102827549, -0.015843885019421577, -0.0024752686731517315, 0.04067982733249664, 0.02747623436152935, 0.04825037717819214, 0.006445575505495071, 0.014715083874762058, -0.028681736439466476, -0.04003037512302399, -0.002639962127432227, -0.010782672092318535, -0.05107893794775009, ...
On the mapping between Hopfield networks and Restricted Boltzmann Machines
https://openreview.net/forum?id=RGJbergVIoO
[ "Matthew Smart", "Anton Zilman" ]
Oral
null
Hopfield networks (HNs) and Restricted Boltzmann Machines (RBMs) are two important models at the interface of statistical physics, machine learning, and neuroscience. Recently, there has been interest in the relationship between HNs and RBMs, due to their similarity under the statistical mechanics formalism. An exact m...
[ "Hopfield Networks", "Restricted Boltzmann Machines", "Statistical Physics" ]
null
3,693
2101.11744
title_snapshot
[ -0.033685389906167984, 0.012620963156223297, -0.009593351744115353, 0.02642947807908058, 0.046085864305496216, 0.024377787485718727, 0.033731020987033844, 0.0023605150636285543, -0.049443311989307404, -0.035663653165102005, -0.016570670530200005, -0.006042651366442442, -0.044283099472522736,...
Efficient Generalized Spherical CNNs
https://openreview.net/forum?id=rWZz3sJfCkm
[ "Oliver Cobb", "Christopher G. R. Wallis", "Augustine N. Mavor-Parker", "Augustin Marignier", "Matthew A. Price", "Mayeul d'Avezac", "Jason McEwen" ]
Poster
null
Many problems across computer vision and the natural sciences require the analysis of spherical data, for which representations may be learned efficiently by encoding equivariance to rotational symmetries. We present a generalized spherical CNN framework that encompasses various existing approaches and allows them to ...
[]
null
3,692
2010.11661
title_snapshot
[ 0.011570247821509838, 0.0036090391222387552, 0.05030001699924469, -0.001118840416893363, -0.004016679711639881, 0.03142698481678963, -0.0053975447081029415, 0.01603471115231514, -0.04714926332235336, -0.05367167666554451, -0.01647070236504078, -0.01875322312116623, -0.06523070484399796, 0....
DEBERTA: DECODING-ENHANCED BERT WITH DISENTANGLED ATTENTION
https://openreview.net/forum?id=XPZIaotutsD
[ "Pengcheng He", "Xiaodong Liu", "Jianfeng Gao", "Weizhu Chen" ]
Poster
null
Recent progress in pre-trained neural language models has significantly improved the performance of many natural language processing (NLP) tasks. In this paper we propose a new model architecture DeBERTa (Decoding-enhanced BERT with disentangled attention) that improves the BERT and RoBERTa models using two novel techn...
[ "Transformer", "Attention", "Natural Language Processing", "Language Model Pre-training", "Position Encoding" ]
null
3,690
2006.03654
title_snapshot
[ -0.006981974001973867, -0.029566919431090355, 0.018127979710698128, 0.053034428507089615, 0.006230969913303852, 0.017798367887735367, 0.03509972617030144, 0.014938984997570515, -0.014596540480852127, -0.011755627579987049, -0.063912034034729, 0.0554974228143692, -0.05224985256791115, 0.024...
Optimizing Memory Placement using Evolutionary Graph Reinforcement Learning
https://openreview.net/forum?id=-6vS_4Kfz0
[ "Shauharda Khadka", "Estelle Aflalo", "Mattias Marder", "Avrech Ben-David", "Santiago Miret", "Shie Mannor", "Tamir Hazan", "Hanlin Tang", "Somdeb Majumdar" ]
Poster
null
For deep neural network accelerators, memory movement is both energetically expensive and can bound computation. Therefore, optimal mapping of tensors to memory hierarchies is critical to performance. The growing complexity of neural networks calls for automated memory mapping instead of manual heuristic approaches; ye...
[ "Reinforcement Learning", "Memory Mapping", "Device Placement", "Evolutionary Algorithms" ]
null
3,688
2007.07298
title_snapshot
[ -0.04176400601863861, -0.01760202646255493, -0.011495987884700298, 0.061179690062999725, 0.03038988821208477, 0.04981020838022232, 0.015339864417910576, 0.022557780146598816, -0.030348649248480797, -0.04161548241972923, 0.00806912500411272, 0.006280864588916302, -0.0752929151058197, -0.019...
On the geometry of generalization and memorization in deep neural networks
https://openreview.net/forum?id=V8jrrnwGbuc
[ "Cory Stephenson", "suchismita padhy", "Abhinav Ganesh", "Yue Hui", "Hanlin Tang", "SueYeon Chung" ]
Poster
null
Understanding how large neural networks avoid memorizing training data is key to explaining their high generalization performance. To examine the structure of when and where memorization occurs in a deep network, we use a recently developed replica-based mean field theoretic geometric analysis method. We find that all ...
[ "deep learning theory", "representation learning", "statistical physics methods", "double descent" ]
null
3,668
2105.14602
title_snapshot
[ -0.03715723380446434, 0.0023429873399436474, -0.003200485371053219, 0.04923921078443527, 0.03327656164765358, -0.0004432083514984697, 0.03159158304333687, 0.030215341597795486, -0.0365314707159996, -0.032171741127967834, 0.00021727893908973783, -0.006216964218765497, -0.05595923215150833, ...
Continual learning in recurrent neural networks
https://openreview.net/forum?id=8xeBUgD8u9
[ "Benjamin Ehret", "Christian Henning", "Maria Cervera", "Alexander Meulemans", "Johannes von Oswald", "Benjamin F Grewe" ]
Poster
null
While a diverse collection of continual learning (CL) methods has been proposed to prevent catastrophic forgetting, a thorough investigation of their effectiveness for processing sequential data with recurrent neural networks (RNNs) is lacking. Here, we provide the first comprehensive evaluation of established CL metho...
[ "Recurrent Neural Networks", "Continual Learning" ]
null
3,663
2006.12109
title_snapshot
[ -0.013743549585342407, -0.028301876038312912, 0.0016339923022314906, 0.04311349242925644, 0.04878789559006691, 0.05509394407272339, 0.02557564154267311, 0.01714976876974106, -0.0537552572786808, -0.03955494612455368, -0.013317233882844448, 0.00611115200445056, -0.06147534027695656, 0.00625...
Witches' Brew: Industrial Scale Data Poisoning via Gradient Matching
https://openreview.net/forum?id=01olnfLIbD
[ "Jonas Geiping", "Liam H Fowl", "W. Ronny Huang", "Wojciech Czaja", "Gavin Taylor", "Michael Moeller", "Tom Goldstein" ]
Poster
null
Data Poisoning attacks modify training data to maliciously control a model trained on such data. In this work, we focus on targeted poisoning attacks which cause a reclassification of an unmodified test image and as such breach model integrity. We consider a particularly malicious poisoning attack that is both ``from s...
[ "Data Poisoning", "ImageNet", "Large-scale", "Gradient Alignment", "Security", "Backdoor Attacks", "from-scratch", "clean-label" ]
null
3,659
2009.02276
title_snapshot
[ -0.01481311023235321, -0.011391687206923962, -0.03824008256196976, 0.06390050053596497, 0.054983820766210556, 0.003850470529869199, 0.04439666122198105, -0.01828579232096672, -0.024238193407654762, -0.029070554301142693, -0.010156150907278061, -0.008955534547567368, -0.05949599668383598, -...
Overfitting for Fun and Profit: Instance-Adaptive Data Compression
https://openreview.net/forum?id=oFp8Mx_V5FL
[ "Ties van Rozendaal", "Iris AM Huijben", "Taco Cohen" ]
Poster
null
Neural data compression has been shown to outperform classical methods in terms of $RD$ performance, with results still improving rapidly. At a high level, neural compression is based on an autoencoder that tries to reconstruct the input instance from a (quantized) latent representation, coupled with a prior that is us...
[ "Neural data compression", "Learned compression", "Generative modeling", "Overfitting", "Finetuning", "Instance learning", "Instance adaptation", "Variational autoencoders", "Rate-distortion optimization", "Model compression", "Weight quantization" ]
null
3,658
2101.08687
title_snapshot
[ 0.00029691291274502873, -0.03033795952796936, -0.014065759256482124, 0.07250984758138657, 0.07357413321733475, 0.06586889922618866, 0.0012332226615399122, -0.0007365595083683729, -0.0317048616707325, -0.033605221658945084, -0.009353266097605228, 0.0108450036495924, -0.04230578988790512, -0...
A Block Minifloat Representation for Training Deep Neural Networks
https://openreview.net/forum?id=6zaTwpNSsQ2
[ "Sean Fox", "Seyedramin Rasoulinezhad", "Julian Faraone", "david boland", "Philip Leong" ]
Poster
null
Training Deep Neural Networks (DNN) with high efficiency can be difficult to achieve with native floating-point representations and commercially available hardware. Specialized arithmetic with custom acceleration offers perhaps the most promising alternative. Ongoing research is trending towards narrow floating-point r...
[]
null
3,654
null
null
[ -0.01855696737766266, -0.013917063362896442, -0.015064699575304985, 0.05430873855948448, 0.04336576163768768, 0.06128052622079849, -0.01883421465754509, -0.012842445634305477, -0.051881104707717896, -0.015550711192190647, 0.013491062447428703, -0.03471284732222557, -0.06242506951093674, 0....
Representation Learning via Invariant Causal Mechanisms
https://openreview.net/forum?id=9p2ekP904Rs
[ "Jovana Mitrovic", "Brian McWilliams", "Jacob C Walker", "Lars Holger Buesing", "Charles Blundell" ]
Poster
null
Self-supervised learning has emerged as a strategy to reduce the reliance on costly supervised signal by pretraining representations only using unlabeled data. These methods combine heuristic proxy classification tasks with data augmentations and have achieved significant success, but our theoretical understanding of t...
[ "Representation Learning", "Self-supervised Learning", "Contrastive Methods", "Causality" ]
null
3,652
2010.07922
title_snapshot
[ 0.008725563064217567, -0.024544615298509598, -0.021546246483922005, 0.0430171936750412, 0.026281991973519325, 0.018459755927324295, 0.006036241073161364, -0.0033167931251227856, -0.05471057444810867, -0.03268510475754738, -0.021969862282276154, 0.006897741463035345, -0.05484984070062637, 0...
Sparse encoding for more-interpretable feature-selecting representations in probabilistic matrix factorization
https://openreview.net/forum?id=D_KeYoqCYC
[ "Joshua C Chang", "Patrick Fletcher", "Jungmin Han", "Ted L Chang", "Shashaank Vattikuti", "Bart Desmet", "Ayah Zirikly", "Carson C Chow" ]
Poster
null
Dimensionality reduction methods for count data are critical to a wide range of applications in medical informatics and other fields where model interpretability is paramount. For such data, hierarchical Poisson matrix factorization (HPF) and other sparse probabilistic non-negative matrix factorization (NMF) methods ar...
[ "poisson matrix factorization", "generalized additive model", "probabilistic matrix factorization", "bayesian", "sparse coding", "interpretability", "factor analysis" ]
null
3,647
2012.04171
title_snapshot
[ -0.026093317195773125, -0.025397229939699173, 0.007165919523686171, 0.01080235280096531, 0.05720924586057663, 0.04139430820941925, 0.017972057685256004, -0.029741507023572922, -0.04079262167215347, -0.05520585924386978, 0.014201256446540356, 0.00034191692247986794, -0.05079945921897888, 0....
Mapping the Timescale Organization of Neural Language Models
https://openreview.net/forum?id=J3OUycKwz-
[ "Hsiang-Yun Sherry Chien", "Jinhan Zhang", "Christopher Honey" ]
Poster
null
In the human brain, sequences of language input are processed within a distributed and hierarchical architecture, in which higher stages of processing encode contextual information over longer timescales. In contrast, in recurrent neural networks which perform natural language processing, we know little about how the m...
[ "natural language processing", "LSTM", "timescale", "hierarchy", "temporal context" ]
null
3,640
2012.06717
title_snapshot
[ -0.045397572219371796, 0.010748357512056828, -0.004214481916278601, 0.011307022534310818, 0.03124270960688591, 0.050937410444021225, 0.05066429078578949, 0.027869081124663353, -0.03950035199522972, -0.017247570678591728, 0.012931493110954762, -0.010313604958355427, -0.034510936588048935, 0...
Neural networks with late-phase weights
https://openreview.net/forum?id=C0qJUx5dxFb
[ "Johannes von Oswald", "Seijin Kobayashi", "Joao Sacramento", "Alexander Meulemans", "Christian Henning", "Benjamin F Grewe" ]
Poster
null
The largely successful method of training neural networks is to learn their weights using some variant of stochastic gradient descent (SGD). Here, we show that the solutions found by SGD can be further improved by ensembling a subset of the weights in late stages of learning. At the end of learning, we obtain back a si...
[]
null
3,621
2007.12927
title_snapshot
[ -0.024499520659446716, -0.0337231419980526, 0.005250841844826937, 0.0382394902408123, 0.03225661441683769, 0.023177990689873695, 0.0028366525657474995, 0.012996891513466835, -0.035022903233766556, -0.032997507601976395, -0.004147852770984173, -0.02895207330584526, -0.04158400371670723, 0.0...
Uncertainty-aware Active Learning for Optimal Bayesian Classifier
https://openreview.net/forum?id=Mu2ZxFctAI
[ "Guang Zhao", "Edward Dougherty", "Byung-Jun Yoon", "Francis Alexander", "Xiaoning Qian" ]
Poster
null
For pool-based active learning, in each iteration a candidate training sample is chosen for labeling by optimizing an acquisition function. In Bayesian classification, expected Loss Reduction~(ELR) methods maximize the expected reduction in the classification error given a new labeled candidate based on a one-step-look...
[ "Active learning", "Bayesian classification" ]
null
3,616
null
null
[ 0.0013389770174399018, -0.005905706435441971, -0.02219744399189949, 0.028649162501096725, 0.021027421578764915, 0.018464604392647743, 0.0037891394458711147, -0.015617066994309425, -0.019927415996789932, -0.040532950311899185, -0.041502341628074646, 0.02684880793094635, -0.06952087581157684, ...
ResNet After All: Neural ODEs and Their Numerical Solution
https://openreview.net/forum?id=HxzSxSxLOJZ
[ "Katharina Ott", "Prateek Katiyar", "Philipp Hennig", "Michael Tiemann" ]
Poster
null
A key appeal of the recently proposed Neural Ordinary Differential Equation (ODE) framework is that it seems to provide a continuous-time extension of discrete residual neural networks. As we show herein, though, trained Neural ODE models actually depend on the specific numerical method used during training. If the tr...
[]
null
3,615
2007.15386
title_snapshot
[ -0.06374005228281021, -0.04320104792714119, -0.005997662898153067, 0.06447884440422058, 0.04329998791217804, 0.05476117134094238, 0.010055555030703545, -0.011501342989504337, -0.03479805216193199, -0.04406668245792389, 0.033920180052518845, 0.012336727231740952, -0.052513159811496735, -0.0...
Benefit of deep learning with non-convex noisy gradient descent: Provable excess risk bound and superiority to kernel methods
https://openreview.net/forum?id=2m0g1wEafh
[ "Taiji Suzuki", "Shunta Akiyama" ]
Spotlight
null
Establishing a theoretical analysis that explains why deep learning can outperform shallow learning such as kernel methods is one of the biggest issues in the deep learning literature. Towards answering this question, we evaluate excess risk of a deep learning estimator trained by a noisy gradient descent with ridge re...
[ "Excess risk", "minimax optimal rate", "local Rademacher complexity", "fast learning rate", "kernel method", "linear estimator" ]
null
3,614
2012.03224
title_snapshot
[ -0.04363362863659859, -0.04207625612616539, 0.009286114014685154, 0.053659047931432724, 0.03102877549827099, 0.045529600232839584, 0.03346095234155655, 0.008815043605864048, -0.015414787456393242, -0.05510873720049858, -0.02014978788793087, 0.006851447280496359, -0.05057802051305771, 0.021...
Generalized Variational Continual Learning
https://openreview.net/forum?id=_IM-AfFhna9
[ "Noel Loo", "Siddharth Swaroop", "Richard E Turner" ]
Poster
null
Continual learning deals with training models on new tasks and datasets in an online fashion. One strand of research has used probabilistic regularization for continual learning, with two of the main approaches in this vein being Online Elastic Weight Consolidation (Online EWC) and Variational Continual Learning (VCL)....
[]
null
3,595
2011.12328
title_snapshot
[ 0.021971987560391426, -0.02646126225590706, 0.018835552036762238, 0.021482378244400024, 0.03838865086436272, 0.039583709090948105, 0.028190257027745247, 0.024323375895619392, -0.06960713118314743, -0.03085312806069851, -0.03680266812443733, 0.021370960399508476, -0.07206349074840546, 0.001...
Mind the Gap when Conditioning Amortised Inference in Sequential Latent-Variable Models
https://openreview.net/forum?id=a2gqxKDvYys
[ "Justin Bayer", "Maximilian Soelch", "Atanas Mirchev", "Baris Kayalibay", "Patrick van der Smagt" ]
Poster
null
Amortised inference enables scalable learning of sequential latent-variable models (LVMs) with the evidence lower bound (ELBO). In this setting, variational posteriors are often only partially conditioned. While the true posteriors depend, e.g., on the entire sequence of observations, approximate posteriors are only in...
[ "variational inference", "state-space models", "amortized inference", "recurrent networks" ]
null
3,587
2101.07046
title_snapshot
[ 0.016350625082850456, -0.00127906643319875, 0.013649757951498032, 0.049905624240636826, 0.030832335352897644, 0.0398462750017643, 0.04060453176498413, -0.006668107584118843, -0.03404863551259041, -0.017520206049084663, -0.008011126890778542, 0.02211761474609375, -0.07276329398155212, 0.010...
CausalWorld: A Robotic Manipulation Benchmark for Causal Structure and Transfer Learning
https://openreview.net/forum?id=SK7A5pdrgov
[ "Ossama Ahmed", "Frederik Träuble", "Anirudh Goyal", "Alexander Neitz", "Manuel Wuthrich", "Yoshua Bengio", "Bernhard Schölkopf", "Stefan Bauer" ]
Poster
null
Despite recent successes of reinforcement learning (RL), it remains a challenge for agents to transfer learned skills to related environments. To facilitate research addressing this problem, we proposeCausalWorld, a benchmark for causal structure and transfer learning in a robotic manipulation environment. The environm...
[ "reinforcement learning", "transfer learning", "sim2real transfer", "domain adaptation", "causality", "generalization", "robotics" ]
null
3,586
2010.04296
title_snapshot
[ -0.016322949901223183, -0.0038645188324153423, -0.02503608539700508, 0.024995481595396996, 0.032300349324941635, 0.03172868862748146, 0.014965576119720936, 0.020977085456252098, -0.025194162502884865, -0.040101535618305206, -0.01567002944648266, 0.02242702804505825, -0.06864667683839798, -...
Transformer protein language models are unsupervised structure learners
https://openreview.net/forum?id=fylclEqgvgd
[ "Roshan Rao", "Joshua Meier", "Tom Sercu", "Sergey Ovchinnikov", "Alexander Rives" ]
Poster
null
Unsupervised contact prediction is central to uncovering physical, structural, and functional constraints for protein structure determination and design. For decades, the predominant approach has been to infer evolutionary constraints from a set of related sequences. In the past year, protein language models have emerg...
[ "proteins", "language modeling", "structure prediction", "unsupervised learning", "explainable" ]
null
3,581
null
null
[ -0.015524402260780334, -0.01392272301018238, -0.023728083819150925, -0.000388002663385123, 0.01299765333533287, 0.006692465860396624, 0.03852809965610504, 0.03128402680158615, 0.030975593253970146, -0.01238820981234312, 0.003470764495432377, 0.029575234279036522, -0.0758398175239563, 0.021...
Neural ODE Processes
https://openreview.net/forum?id=27acGyyI1BY
[ "Alexander Norcliffe", "Cristian Bodnar", "Ben Day", "Jacob Moss", "Pietro Liò" ]
Poster
null
Neural Ordinary Differential Equations (NODEs) use a neural network to model the instantaneous rate of change in the state of a system. However, despite their apparent suitability for dynamics-governed time-series, NODEs present a few disadvantages. First, they are unable to adapt to incoming data-points, a fundamental...
[ "differential equations", "neural processes", "dynamics", "deep learning", "neural ode" ]
null
3,579
2103.12413
title_snapshot
[ -0.05766534060239792, -0.021446984261274338, -0.025749139487743378, 0.050557494163513184, 0.029967542737722397, 0.04494670405983925, 0.007830863818526268, 0.006290291901677847, -0.021920569241046906, -0.0458713062107563, 0.02194027788937092, -0.016629861667752266, -0.05155118182301521, -0....
The role of Disentanglement in Generalisation
https://openreview.net/forum?id=qbH974jKUVy
[ "Milton Llera Montero", "Casimir JH Ludwig", "Rui Ponte Costa", "Gaurav Malhotra", "Jeffrey Bowers" ]
Poster
null
Combinatorial generalisation — the ability to understand and produce novel combinations of familiar elements — is a core capacity of human intelligence that current AI systems struggle with. Recently, it has been suggested that learning disentangled representations may help address this problem. It is claimed that such...
[ "disentanglement", "compositionality", "compositional generalization", "generalisation", "generative models", "variational autoencoders" ]
null
3,567
null
null
[ 0.015437041409313679, -0.007364911958575249, -0.006171267945319414, 0.041562292724847794, 0.03506309539079666, -0.0017700301250442863, 0.06068402901291847, 0.014387832954525948, -0.03804455325007439, -0.019946321845054626, -0.06678587198257446, 0.010249927639961243, -0.09030590206384659, 0...
Learning to live with Dale's principle: ANNs with separate excitatory and inhibitory units
https://openreview.net/forum?id=eU776ZYxEpz
[ "Jonathan Cornford", "Damjan Kalajdzievski", "Marco Leite", "Amélie Lamarquette", "Dimitri Michael Kullmann", "Blake Aaron Richards" ]
Poster
null
The units in artificial neural networks (ANNs) can be thought of as abstractions of biological neurons, and ANNs are increasingly used in neuroscience research. However, there are many important differences between ANN units and real neurons. One of the most notable is the absence of Dale's principle, which ensures th...
[]
null
3,565
null
null
[ -0.048680756241083145, 0.017022589221596718, -0.0012591403210535645, 0.02576247788965702, 0.010762578807771206, 0.030251668766140938, 0.030980003997683525, -0.010036863386631012, -0.0717918798327446, -0.03626548871397972, 0.008509802632033825, 0.005076349712908268, -0.0519632026553154, 0.0...
SALD: Sign Agnostic Learning with Derivatives
https://openreview.net/forum?id=7EDgLu9reQD
[ "Matan Atzmon", "Yaron Lipman" ]
Poster
null
Learning 3D geometry directly from raw data, such as point clouds, triangle soups, or unoriented meshes is still a challenging task that feeds many downstream computer vision and graphics applications. In this paper, we introduce SALD: a method for learning implicit neural representations of shapes directly from raw ...
[ "implicit neural representations", "3D shapes learning", "sign agnostic learning" ]
null
3,563
2006.05400
title_snapshot
[ -0.005329521838575602, -0.002156014321371913, 0.027602609246969223, 0.02977471984922886, 0.007983075454831123, 0.050118569284677505, -0.0009733735350891948, -0.0077360463328659534, -0.03652719408273697, -0.03661302104592323, -0.00878866482526064, 0.02092362754046917, -0.0641806423664093, 0...
Ringing ReLUs: Harmonic Distortion Analysis of Nonlinear Feedforward Networks
https://openreview.net/forum?id=TaYhv-q1Xit
[ "Christian H.X. Ali Mehmeti-Göpel", "David Hartmann", "Michael Wand" ]
Poster
null
In this paper, we apply harmonic distortion analysis to understand the effect of nonlinearities in the spectral domain. Each nonlinear layer creates higher-frequency harmonics, which we call "blueshift", whose magnitude increases with network depth, thereby increasing the “roughness” of the output landscape. Unlike dif...
[ "deep learning theory", "loss landscape", "harmonic distortion analysis", "network trainability" ]
null
3,561
null
null
[ -0.018369348719716072, -0.03533903881907463, 0.007536008488386869, 0.016916649416089058, 0.046338777989149094, 0.017913153395056725, 0.026584753766655922, 0.018715428188443184, -0.04006603732705116, -0.04741315543651581, 0.015243880450725555, 0.012475538067519665, -0.08115216344594955, -0....
CoCon: A Self-Supervised Approach for Controlled Text Generation
https://openreview.net/forum?id=VD_ozqvBy4W
[ "Alvin Chan", "Yew-Soon Ong", "Bill Pung", "Aston Zhang", "Jie Fu" ]
Poster
null
Pretrained Transformer-based language models (LMs) display remarkable natural language generation capabilities. With their immense potential, controlling text generation of such LMs is getting attention. While there are studies that seek to control high-level attributes (such as sentiment and topic) of generated text, ...
[ "Language modeling", "text generation", "controlled generation", "self-supervised learning" ]
null
3,544
2006.03535
title_snapshot
[ -0.023972008377313614, -0.01901339739561081, -0.03281904757022858, 0.05087188631296158, 0.04232991486787796, 0.012719053775072098, 0.010848627425730228, 0.032740186899900436, 0.006101741921156645, 0.010655476711690426, -0.01957652159035206, 0.048982322216033936, -0.0331699475646019, 0.0125...
Bidirectional Variational Inference for Non-Autoregressive Text-to-Speech
https://openreview.net/forum?id=o3iritJHLfO
[ "Yoonhyung Lee", "Joongbo Shin", "Kyomin Jung" ]
Poster
null
Although early text-to-speech (TTS) models such as Tacotron 2 have succeeded in generating human-like speech, their autoregressive architectures have several limitations: (1) They require a lot of time to generate a mel-spectrogram consisting of hundreds of steps. (2) The autoregressive speech generation shows a lack o...
[ "text-to-speech", "speech synthesis", "non-autoregressive", "VAE" ]
null
3,537
null
null
[ -0.007124045863747597, -0.01261735800653696, -0.025824224576354027, 0.011542028747498989, -0.007309034466743469, 0.0529303103685379, 0.0717078223824501, 0.039577651768922806, -0.032283663749694824, -0.04054950177669525, -0.009661166928708553, 0.027202704921364784, -0.057524699717760086, 0....
Learning continuous-time PDEs from sparse data with graph neural networks
https://openreview.net/forum?id=aUX5Plaq7Oy
[ "Valerii Iakovlev", "Markus Heinonen", "Harri Lähdesmäki" ]
Poster
null
The behavior of many dynamical systems follow complex, yet still unknown partial differential equations (PDEs). While several machine learning methods have been proposed to learn PDEs directly from data, previous methods are limited to discrete-time approximations or make the limiting assumption of the observations arr...
[ "dynamical systems", "partial differential equations", "PDEs", "graph neural networks", "continuous time" ]
null
3,532
2006.08956
title_snapshot
[ -0.0500887893140316, -0.007827533408999443, 0.0010896384483203292, 0.07415071129798889, 0.019349902868270874, 0.02615324780344963, 0.03094073198735714, 0.02339729480445385, -0.013865792192518711, -0.05319815129041672, 0.05024575814604759, -0.010892459191381931, -0.0717620700597763, 0.02395...
NAS-Bench-ASR: Reproducible Neural Architecture Search for Speech Recognition
https://openreview.net/forum?id=CU0APx9LMaL
[ "Abhinav Mehrotra", "Alberto Gil C. P. Ramos", "Sourav Bhattacharya", "Łukasz Dudziak", "Ravichander Vipperla", "Thomas Chau", "Mohamed S Abdelfattah", "Samin Ishtiaq", "Nicholas Donald Lane" ]
Poster
null
Powered by innovations in novel architecture design, noise tolerance techniques and increasing model capacity, Automatic Speech Recognition (ASR) has made giant strides in reducing word-error-rate over the past decade. ASR models are often trained with tens of thousand hours of high quality speech data to produce state...
[ "NAS", "ASR", "Benchmark" ]
null
3,528
null
null
[ -0.018165679648518562, -0.01843804121017456, -0.030256735160946846, 0.019411364570260048, 0.018660956993699074, 0.05976222828030586, 0.049317050725221634, -0.006334360223263502, -0.02318042702972889, -0.03836724907159805, 0.006069574970752001, 0.02224073000252247, -0.061564501374959946, -0...
Collective Robustness Certificates: Exploiting Interdependence in Graph Neural Networks
https://openreview.net/forum?id=ULQdiUTHe3y
[ "Jan Schuchardt", "Aleksandar Bojchevski", "Johannes Gasteiger", "Stephan Günnemann" ]
Poster
null
In tasks like node classification, image segmentation, and named-entity recognition we have a classifier that simultaneously outputs multiple predictions (a vector of labels) based on a single input, i.e. a single graph, image, or document respectively. Existing adversarial robustness certificates consider each predict...
[ "Robustness certificates", "Adversarial robustness", "Graph neural networks" ]
null
3,522
2302.02829
title_snapshot
[ -0.016498809680342674, -0.02913777530193329, -0.004808191210031509, 0.06329748779535294, -0.010481433942914009, 0.011623150669038296, 0.030620522797107697, 0.000007285805168066872, -0.005884724669158459, -0.02362709306180477, -0.0058623626828193665, -0.03125288709998131, -0.06961613148450851...
Adversarially Guided Actor-Critic
https://openreview.net/forum?id=_mQp5cr_iNy
[ "Yannis Flet-Berliac", "Johan Ferret", "Olivier Pietquin", "Philippe Preux", "Matthieu Geist" ]
Poster
null
Despite definite success in deep reinforcement learning problems, actor-critic algorithms are still confronted with sample inefficiency in complex environments, particularly in tasks where efficient exploration is a bottleneck. These methods consider a policy (the actor) and a value function (the critic) whose respecti...
[]
null
3,504
2102.04376
title_snapshot
[ -0.017045218497514725, -0.05467970669269562, -0.015884703025221825, 0.043758150190114975, 0.02690066769719124, 0.0217960886657238, 0.022532127797603607, -0.02123061567544937, -0.04668326675891876, -0.0303563941270113, -0.01724666729569435, 0.02168767713010311, -0.06823449581861496, -0.0253...
Training independent subnetworks for robust prediction
https://openreview.net/forum?id=OGg9XnKxFAH
[ "Marton Havasi", "Rodolphe Jenatton", "Stanislav Fort", "Jeremiah Zhe Liu", "Jasper Snoek", "Balaji Lakshminarayanan", "Andrew Mingbo Dai", "Dustin Tran" ]
Poster
null
Recent approaches to efficiently ensemble neural networks have shown that strong robustness and uncertainty performance can be achieved with a negligible gain in parameters over the original network. However, these methods still require multiple forward passes for prediction, leading to a significant runtime cost. In ...
[ "Efficient ensembles", "robustness" ]
null
3,498
2010.06610
title_snapshot
[ 0.0029844031669199467, -0.024935612455010414, -0.002504778327420354, 0.026434900239109993, 0.08438663929700851, 0.015849584713578224, 0.009495108388364315, -0.021110758185386658, -0.018128640949726105, -0.07848770171403885, 0.03268435597419739, 0.0016526380786672235, -0.08035188168287277, ...
Complex Query Answering with Neural Link Predictors
https://openreview.net/forum?id=Mos9F9kDwkz
[ "Erik Arakelyan", "Daniel Daza", "Pasquale Minervini", "Michael Cochez" ]
Oral
null
Neural link predictors are immensely useful for identifying missing edges in large scale Knowledge Graphs. However, it is still not clear how to use these models for answering more complex queries that arise in a number of domains, such as queries using logical conjunctions ($\land$), disjunctions ($\lor$) and existent...
[ "neural link prediction", "complex query answering" ]
null
3,496
2011.03459
title_snapshot
[ -0.022371187806129456, -0.032177552580833435, 0.010288788937032223, 0.05891856551170349, 0.03488396853208542, 0.010687303729355335, 0.014891059137880802, -0.01909365877509117, -0.009365435689687729, -0.026116391643881798, -0.033919062465429306, 0.0325656421482563, -0.053360432386398315, 0....
Grounding Language to Autonomously-Acquired Skills via Goal Generation
https://openreview.net/forum?id=chPj_I5KMHG
[ "Ahmed Akakzia", "Cédric Colas", "Pierre-Yves Oudeyer", "Mohamed CHETOUANI", "Olivier Sigaud" ]
Poster
null
We are interested in the autonomous acquisition of repertoires of skills. Language-conditioned reinforcement learning (LC-RL) approaches are great tools in this quest, as they allow to express abstract goals as sets of constraints on the states. However, most LC-RL agents are not autonomous and cannot learn without ext...
[ "Deep reinforcement learning", "intrinsic motivations", "symbolic representations", "autonomous learning" ]
null
3,493
2006.07185
title_snapshot
[ -0.04147777706384659, 0.0010350116062909365, -0.01833423413336277, 0.008343566209077835, 0.04353570565581322, 0.021085774526000023, 0.045949388295412064, 0.025991737842559814, -0.04137894883751869, -0.007040851283818483, -0.05493750050663948, 0.032127127051353455, -0.0642593577504158, -0.0...
Hopfield Networks is All You Need
https://openreview.net/forum?id=tL89RnzIiCd
[ "Hubert Ramsauer", "Bernhard Schäfl", "Johannes Lehner", "Philipp Seidl", "Michael Widrich", "Lukas Gruber", "Markus Holzleitner", "Thomas Adler", "David Kreil", "Michael K Kopp", "Günter Klambauer", "Johannes Brandstetter", "Sepp Hochreiter" ]
Poster
null
We introduce a modern Hopfield network with continuous states and a corresponding update rule. The new Hopfield network can store exponentially (with the dimension of the associative space) many patterns, retrieves the pattern with one update, and has exponentially small retrieval errors. It has three types of energy m...
[ "Modern Hopfield Network", "Energy", "Attention", "Convergence", "Storage Capacity", "Hopfield layer", "Associative Memory" ]
null
3,489
2008.02217
title_snapshot
[ -0.014532532542943954, -0.006951846648007631, 0.0002803085371851921, 0.027208779007196426, 0.03979121893644333, -0.0008518453687429428, 0.020021244883537292, -0.0018738317303359509, -0.02768009528517723, -0.04184650629758835, 0.013989854604005814, -0.03149009495973587, -0.08954507112503052, ...
Differentiable Trust Region Layers for Deep Reinforcement Learning
https://openreview.net/forum?id=qYZD-AO1Vn
[ "Fabian Otto", "Philipp Becker", "Vien Anh Ngo", "Hanna Carolin Maria Ziesche", "Gerhard Neumann" ]
Poster
null
Trust region methods are a popular tool in reinforcement learning as they yield robust policy updates in continuous and discrete action spaces. However, enforcing such trust regions in deep reinforcement learning is difficult. Hence, many approaches, such as Trust Region Policy Optimization (TRPO) and Proximal Policy O...
[ "reinforcement learning", "trust region", "policy gradient", "projection", "Wasserstein distance", "Kullback-Leibler divergence", "Frobenius norm" ]
null
3,480
2101.09207
title_snapshot
[ -0.032281748950481415, -0.03126596286892891, 0.0034801759757101536, 0.0686168372631073, 0.04222990572452545, 0.0269856508821249, 0.03176372870802879, -0.02553655207157135, -0.00437259441241622, -0.04626092314720154, -0.01141776517033577, -0.0017766463570296764, -0.07206014543771744, -0.000...
Self-supervised Visual Reinforcement Learning with Object-centric Representations
https://openreview.net/forum?id=xppLmXCbOw1
[ "Andrii Zadaianchuk", "Maximilian Seitzer", "Georg Martius" ]
Spotlight
null
Autonomous agents need large repertoires of skills to act reasonably on new tasks that they have not seen before. However, acquiring these skills using only a stream of high-dimensional, unstructured, and unlabeled observations is a tricky challenge for any autonomous agent. Previous methods have used variational autoe...
[ "self-supervision", "autonomous learning", "object-centric representations", "visual reinforcement learning" ]
null
3,479
2011.14381
title_snapshot
[ -0.008180283941328526, 0.00010645019938237965, -0.018509391695261, 0.036271654069423676, 0.029221197590231895, 0.023692136630415916, 0.021911388263106346, 0.007254786789417267, -0.037903301417827606, -0.03633558005094528, -0.0406411811709404, 0.0023720706813037395, -0.07689904421567917, -0...
Temporally-Extended ε-Greedy Exploration
https://openreview.net/forum?id=ONBPHFZ7zG4
[ "Will Dabney", "Georg Ostrovski", "Andre Barreto" ]
Poster
null
Recent work on exploration in reinforcement learning (RL) has led to a series of increasingly complex solutions to the problem. This increase in complexity often comes at the expense of generality. Recent empirical studies suggest that, when applied to a broader set of domains, some sophisticated exploration methods ar...
[ "reinforcement learning", "exploration" ]
null
3,475
2006.01782
title_snapshot
[ -0.011410429142415524, -0.01832977868616581, -0.01748831570148468, 0.04399494081735611, 0.05537569150328636, 0.020831700414419174, 0.0269649438560009, 0.00329385488294065, -0.041970349848270416, -0.03562932461500168, -0.013388698920607567, -0.013656246475875378, -0.05032319203019142, -0.02...
Learning Associative Inference Using Fast Weight Memory
https://openreview.net/forum?id=TuK6agbdt27
[ "Imanol Schlag", "Tsendsuren Munkhdalai", "Jürgen Schmidhuber" ]
Poster
null
Humans can quickly associate stimuli to solve problems in novel contexts. Our novel neural network model learns state representations of facts that can be composed to perform such associative inference. To this end, we augment the LSTM model with an associative memory, dubbed \textit{Fast Weight Memory} (FWM). Through ...
[ "memory-augmented neural networks", "tensor product", "fast weights" ]
null
3,467
2011.07831
title_snapshot
[ -0.03460824489593506, 0.002605004468932748, 0.007783577777445316, 0.041224975138902664, 0.007154098246246576, 0.03625628352165222, 0.005508933681994677, 0.020262889564037323, -0.04525596648454666, -0.005706895142793655, -0.013890034519135952, 0.009577506221830845, -0.08021170645952225, -0....
Multiscale Score Matching for Out-of-Distribution Detection
https://openreview.net/forum?id=xoHdgbQJohv
[ "Ahsan Mahmood", "Junier Oliva", "Martin Andreas Styner" ]
Poster
null
We present a new methodology for detecting out-of-distribution (OOD) images by utilizing norms of the score estimates at multiple noise scales. A score is defined to be the gradient of the log density with respect to the input data. Our methodology is completely unsupervised and follows a straight forward training sche...
[ "out-of-distribution detection", "score matching", "deep learning", "outlier detection" ]
null
3,465
2010.13132
title_snapshot
[ 0.012107078917324543, -0.0010239322436973453, 0.027758168056607246, 0.03335217759013176, 0.03987913578748703, 0.00594719173386693, 0.0040853749960660934, -0.008378286845982075, -0.027361106127500534, -0.05121028423309326, 0.000979319796897471, 0.007481140084564686, -0.06992431730031967, -0...
Learning to Sample with Local and Global Contexts in Experience Replay Buffer
https://openreview.net/forum?id=gJYlaqL8i8
[ "Youngmin Oh", "Kimin Lee", "Jinwoo Shin", "Eunho Yang", "Sung Ju Hwang" ]
Poster
null
Experience replay, which enables the agents to remember and reuse experience from the past, has played a significant role in the success of off-policy reinforcement learning (RL). To utilize the experience replay efficiently, the existing sampling methods allow selecting out more meaningful experiences by imposing prio...
[ "reinforcement learning", "experience replay buffer", "off-policy RL" ]
null
3,454
2007.07358
title_snapshot
[ -0.012066839262843132, -0.010763215832412243, 0.002150164684280753, 0.04326143488287926, 0.03283634036779404, 0.017409976571798325, -0.00026233517564833164, 0.010103283450007439, -0.04614844173192978, -0.04180149734020233, -0.010565572418272495, 0.01241993810981512, -0.056055307388305664, ...
Discovering a set of policies for the worst case reward
https://openreview.net/forum?id=PUkhWz65dy5
[ "Tom Zahavy", "Andre Barreto", "Daniel J Mankowitz", "Shaobo Hou", "Brendan O'Donoghue", "Iurii Kemaev", "Satinder Singh" ]
Spotlight
null
We study the problem of how to construct a set of policies that can be composed together to solve a collection of reinforcement learning tasks. Each task is a different reward function defined as a linear combination of known features. We consider a specific class of policy compositions which we call set improving pol...
[]
null
3,432
2102.04323
title_snapshot
[ -0.04469660669565201, -0.04080851376056671, -0.0063253543339669704, 0.05847424268722534, 0.051504965871572495, 0.007413219660520554, 0.0046673365868628025, -0.025228166952729225, -0.03183455392718315, -0.02431374043226242, -0.029076825827360153, 0.012271866202354431, -0.07946717739105225, ...
Parameter-Based Value Functions
https://openreview.net/forum?id=tV6oBfuyLTQ
[ "Francesco Faccio", "Louis Kirsch", "Jürgen Schmidhuber" ]
Poster
null
Traditional off-policy actor-critic Reinforcement Learning (RL) algorithms learn value functions of a single target policy. However, when value functions are updated to track the learned policy, they forget potentially useful information about old policies. We introduce a class of value functions called Parameter-Based...
[ "Reinforcement Learning", "Off-Policy Reinforcement Learning" ]
null
3,426
2006.09226
title_snapshot
[ -0.024356717243790627, -0.024231180548667908, -0.0048239813186228275, 0.04804975539445877, 0.033841293305158615, 0.0449410155415535, 0.005059358198195696, -0.018538791686296463, -0.029416244477033615, -0.03464093059301376, -0.01012678537517786, 0.03693260997533798, -0.09980384260416031, -0...
New Bounds For Distributed Mean Estimation and Variance Reduction
https://openreview.net/forum?id=t86MwoUCCNe
[ "Peter Davies", "Vijaykrishna Gurunanthan", "Niusha Moshrefi", "Saleh Ashkboos", "Dan Alistarh" ]
Poster
null
We consider the problem of distributed mean estimation (DME), in which $n$ machines are each given a local $d$-dimensional vector $\mathbf x_v \in \mathbb R^d$, and must cooperate to estimate the mean of their inputs $\mathbf \mu = \frac 1n\sum_{v = 1}^n \mathbf x_v$, while minimizing total communication cost. DME is ...
[ "distributed machine learning", "mean estimation", "variance reduction", "lattices" ]
null
3,418
2002.09268
title_snapshot
[ -0.03707395866513252, 0.0170500036329031, -0.02075030468404293, 0.026006469503045082, 0.046937134116888046, 0.0293715950101614, 0.08289553225040436, -0.020413782447576523, -0.007759482599794865, -0.05285661667585373, 0.006447851657867432, -0.03767342492938042, -0.06740065664052963, 0.00022...
Learning to Set Waypoints for Audio-Visual Navigation
https://openreview.net/forum?id=cR91FAodFMe
[ "Changan Chen", "Sagnik Majumder", "Ziad Al-Halah", "Ruohan Gao", "Santhosh Kumar Ramakrishnan", "Kristen Grauman" ]
Poster
null
In audio-visual navigation, an agent intelligently travels through a complex, unmapped 3D environment using both sights and sounds to find a sound source (e.g., a phone ringing in another room). Existing models learn to act at a fixed granularity of agent motion and rely on simple recurrent aggregations of the audio ob...
[ "visual navigation", "audio visual learning", "embodied vision" ]
null
3,405
2008.09622
title_snapshot
[ -0.011476482264697552, 0.012901809997856617, 0.013381410390138626, 0.02606838382780552, 0.0300833061337471, 0.012999145314097404, 0.03867584094405174, 0.022192208096385002, -0.05457640066742897, -0.0627681240439415, -0.039842359721660614, 0.029382726177573204, -0.05955301225185394, -0.0144...
Disambiguating Symbolic Expressions in Informal Documents
https://openreview.net/forum?id=K5j7D81ABvt
[ "Dennis Müller", "Cezary Kaliszyk" ]
Poster
null
We propose the task of \emph{disambiguating} symbolic expressions in informal STEM documents in the form of \LaTeX files -- that is, determining their precise semantics and abstract syntax tree -- as a neural machine translation task. We discuss the distinct challenges involved and present a dataset with roughly 33,000...
[]
null
3,399
2101.11716
title_snapshot
[ 0.024218695238232613, -0.022658543661236763, -0.038791462779045105, 0.041731152683496475, 0.050146471709012985, 0.020136099308729172, 0.00917800236493349, 0.005627655889838934, 0.01648850180208683, -0.0006491586682386696, -0.051550209522247314, 0.04710520803928375, -0.05062035098671913, 0....
Colorization Transformer
https://openreview.net/forum?id=5NA1PinlGFu
[ "Manoj Kumar", "Dirk Weissenborn", "Nal Kalchbrenner" ]
Poster
null
We present the Colorization Transformer, a novel approach for diverse high fidelity image colorization based on self-attention. Given a grayscale image, the colorization proceeds in three steps. We first use a conditional autoregressive transformer to produce a low resolution coarse coloring of the grayscale image. Our...
[]
null
3,388
2102.04432
title_snapshot
[ 0.01736070215702057, -0.04402635619044304, -0.03239146247506142, 0.05438314750790596, 0.028758898377418518, 0.07561412453651428, -0.005106141325086355, -0.0016325151082128286, -0.06357784569263458, -0.050412122160196304, -0.031855612993240356, 0.002188214100897312, -0.06760664284229279, 0....
Theoretical bounds on estimation error for meta-learning
https://openreview.net/forum?id=SZ3wtsXfzQR
[ "James Lucas", "Mengye Ren", "Irene Raissa KAMENI KAMENI", "Toniann Pitassi", "Richard Zemel" ]
Poster
null
Machine learning models have traditionally been developed under the assumption that the training and test distributions match exactly. However, recent success in few-shot learning and related problems are encouraging signs that these models can be adapted to more realistic settings where train and test distributions di...
[ "meta learning", "few-shot", "minimax risk", "lower bounds", "learning theory" ]
null
3,384
2010.07140
title_snapshot
[ -0.009634786285459995, 0.02287176623940468, -0.02055038884282112, 0.045276012271642685, 0.03482489660382271, 0.014800029806792736, 0.06203264743089676, -0.019806798547506332, -0.032844703644514084, -0.007430830970406532, 0.0035500130616128445, 0.030422905460000038, -0.07430269569158554, -0...
Implicit Normalizing Flows
https://openreview.net/forum?id=8PS8m9oYtNy
[ "Cheng Lu", "Jianfei Chen", "Chongxuan Li", "Qiuhao Wang", "Jun Zhu" ]
Spotlight
null
Normalizing flows define a probability distribution by an explicit invertible transformation $\boldsymbol{\mathbf{z}}=f(\boldsymbol{\mathbf{x}})$. In this work, we present implicit normalizing flows (ImpFlows), which generalize normalizing flows by allowing the mapping to be implicitly defined by the roots of an equati...
[ "Normalizing flows", "deep generative models", "probabilistic inference", "implicit functions" ]
null
3,381
2103.09527
title_snapshot
[ -0.020324831828475, -0.039659496396780014, 0.008610550314188004, 0.04668508470058441, 0.030860276892781258, 0.05457251891493797, 0.025531869381666183, -0.005126855801790953, -0.0050069899298250675, -0.04886361584067345, -0.014872484840452671, -0.037368711084127426, -0.0633147805929184, 0.0...
Variational Information Bottleneck for Effective Low-Resource Fine-Tuning
https://openreview.net/forum?id=kvhzKz-_DMF
[ "Rabeeh Karimi mahabadi", "Yonatan Belinkov", "James Henderson" ]
Poster
null
While large-scale pretrained language models have obtained impressive results when fine-tuned on a wide variety of tasks, they still often suffer from overfitting in low-resource scenarios. Since such models are general-purpose feature extractors, many of these features are inevitably irrelevant for a given target task...
[ "Transfer learning", "NLP", "large-scale pre-trained language models", "over-fitting", "robust", "biases", "variational information bottleneck" ]
null
3,366
2106.05469
title_snapshot
[ -0.012936538085341454, -0.02559026889503002, -0.001891666790470481, 0.052599962800741196, 0.03814368322491646, 0.03854846581816673, 0.04956578463315964, 0.01856072060763836, -0.018170075491070747, -0.013494385406374931, -0.024003714323043823, 0.05244878679513931, -0.08171916753053665, 0.03...
TropEx: An Algorithm for Extracting Linear Terms in Deep Neural Networks
https://openreview.net/forum?id=IqtonxWI0V3
[ "Martin Trimmel", "Henning Petzka", "Cristian Sminchisescu" ]
Poster
null
Deep neural networks with rectified linear (ReLU) activations are piecewise linear functions, where hyperplanes partition the input space into an astronomically high number of linear regions. Previous work focused on counting linear regions to measure the network's expressive power and on analyzing geometric properties...
[ "linear regions", "linear terms", "deep learning theory", "deep neural networks", "rectified linear unit", "relu network", "piecewise linear function", "tropical function" ]
null
3,365
null
null
[ 0.008740529417991638, -0.00985116045922041, 0.005262348335236311, 0.028450746089220047, 0.040883343666791916, 0.03835544362664223, 0.011571953073143959, 0.022369664162397385, -0.031364556401968, -0.007226845249533653, -0.009982435032725334, -0.010654956102371216, -0.06917225569486618, 0.00...
Seq2Tens: An Efficient Representation of Sequences by Low-Rank Tensor Projections
https://openreview.net/forum?id=dx4b7lm8jMM
[ "Csaba Toth", "Patric Bonnier", "Harald Oberhauser" ]
Poster
null
Sequential data such as time series, video, or text can be challenging to analyse as the ordered structure gives rise to complex dependencies. At the heart of this is non-commutativity, in the sense that reordering the elements of a sequence can completely change its meaning. We use a classical mathematical object -- t...
[ "time series", "sequential data", "representation learning", "low-rank tensors", "classification", "generative modelling" ]
null
3,363
2006.07027
title_snapshot
[ -0.0046758949756622314, -0.0427774079144001, -0.0084681436419487, 0.033061303198337555, 0.030210567638278008, 0.03757099062204361, 0.01000163983553648, -0.021597428247332573, 0.0018490514485165477, -0.035833753645420074, 0.003495065728202462, 0.01956692337989807, -0.06261742115020752, 0.02...
Representation learning for improved interpretability and classification accuracy of clinical factors from EEG
https://openreview.net/forum?id=TVjLza1t4hI
[ "Garrett Honke", "Irina Higgins", "Nina Thigpen", "Vladimir Miskovic", "Katie Link", "Sunny Duan", "Pramod Gupta", "Julia Klawohn", "Greg Hajcak" ]
Poster
null
Despite extensive standardization, diagnostic interviews for mental health disorders encompass substantial subjective judgment. Previous studies have demonstrated that EEG-based neural measures can function as reliable objective correlates of depression, or even predictors of depression and its course. However, their c...
[ "EEG", "ERP", "electroencephalography", "depression", "representation learning", "disentanglement", "beta-VAE" ]
null
3,359
2010.15274
title_snapshot
[ -0.016948100179433823, 0.014247508719563484, -0.004413349088281393, 0.00007468278636224568, 0.0251392163336277, 0.0547039620578289, 0.04197772219777107, -0.02866405062377453, -0.029491741210222244, -0.04128974303603172, -0.026773370802402496, 0.00014486961299553514, -0.06554020941257477, 0...
Language-Agnostic Representation Learning of Source Code from Structure and Context
https://openreview.net/forum?id=Xh5eMZVONGF
[ "Daniel Zügner", "Tobias Kirschstein", "Michele Catasta", "Jure Leskovec", "Stephan Günnemann" ]
Poster
null
Source code (Context) and its parsed abstract syntax tree (AST; Structure) are two complementary representations of the same computer program. Traditionally, designers of machine learning models have relied predominantly either on Structure or Context. We propose a new model, which jointly learns on Context and Structu...
[ "machine learning for code", "code summarization" ]
null
3,338
2103.11318
title_snapshot
[ -0.004891443066298962, -0.012570670805871487, -0.04430088773369789, 0.03161097317934036, 0.009902702644467354, 0.013425027020275593, 0.030292177572846413, 0.004673711955547333, -0.032277606427669525, -0.0203948225826025, -0.037101730704307556, 0.020782724022865295, -0.0853007584810257, -0....
Generalized Multimodal ELBO
https://openreview.net/forum?id=5Y21V0RDBV
[ "Thomas M. Sutter", "Imant Daunhawer", "Julia E Vogt" ]
Poster
null
Multiple data types naturally co-occur when describing real-world phenomena and learning from them is a long-standing goal in machine learning research. However, existing self-supervised generative models approximating an ELBO are not able to fulfill all desired requirements of multimodal models: their posterior approx...
[ "Multimodal", "VAE", "ELBO", "self-supervised", "generative learning" ]
null
3,332
2105.02470
title_snapshot
[ -0.00028652820037677884, -0.02437732182443142, 0.002438908675685525, 0.06059183180332184, 0.0242892075330019, 0.019940922036767006, 0.003000239608809352, 0.01413020770996809, -0.02959766983985901, -0.0291493758559227, -0.04613965377211571, 0.042972445487976074, -0.10310107469558716, 0.0089...
Model-based micro-data reinforcement learning: what are the crucial model properties and which model to choose?
https://openreview.net/forum?id=p5uylG94S68
[ "Balázs Kégl", "Gabriel Hurtado", "Albert Thomas" ]
Poster
null
We contribute to micro-data model-based reinforcement learning (MBRL) by rigorously comparing popular generative models using a fixed (random shooting) control agent. We find that on an environment that requires multimodal posterior predictives, mixture density nets outperform all other models by a large margin. When m...
[ "model-based reinforcement learning", "generative models", "mixture density nets", "dynamic systems", "heteroscedasticity" ]
null
3,299
2107.11587
title_snapshot
[ -0.036502011120319366, -0.02678162418305874, -0.009630262851715088, 0.07096373289823532, 0.03302517905831337, 0.031402815133333206, 0.011371494270861149, -0.008095895871520042, -0.03790650516748428, -0.0306216012686491, -0.009727086871862411, 0.028463391587138176, -0.074264757335186, -0.00...
Set Prediction without Imposing Structure as Conditional Density Estimation
https://openreview.net/forum?id=04ArenGOz3
[ "David W Zhang", "Gertjan J. Burghouts", "Cees G. M. Snoek" ]
Poster
null
Set prediction is about learning to predict a collection of unordered variables with unknown interrelations. Training such models with set losses imposes the structure of a metric space over sets. We focus on stochastic and underdefined cases, where an incorrectly chosen loss function leads to implausible predictions. ...
[ "set prediction", "energy based models" ]
null
3,297
2010.04109
title_snapshot
[ 0.0019785775803029537, -0.0031742060091346502, -0.005780951119959354, 0.05901576206088066, 0.03575442358851433, 0.02828194946050644, -0.0226745568215847, -0.010558998212218285, -0.028495902195572853, -0.04920867085456848, -0.009783808141946793, 0.033838070929050446, -0.08202408999204636, -...
Learning Value Functions in Deep Policy Gradients using Residual Variance
https://openreview.net/forum?id=NX1He-aFO_F
[ "Yannis Flet-Berliac", "reda ouhamma", "odalric-ambrym maillard", "Philippe Preux" ]
Poster
null
Policy gradient algorithms have proven to be successful in diverse decision making and control tasks. However, these methods suffer from high sample complexity and instability issues. In this paper, we address these challenges by providing a different approach for training the critic in the actor-critic framework. Our ...
[]
null
3,292
2010.04440
title_snapshot
[ -0.012032740749418736, -0.024211520329117775, -0.0074161640368402, 0.062480196356773376, 0.05876791477203369, 0.04016605764627457, 0.027777930721640587, -0.022237492725253105, -0.006168338004499674, -0.03992585092782974, 0.00994480587542057, 0.035658642649650574, -0.08916817605495453, -0.0...
IDF++: Analyzing and Improving Integer Discrete Flows for Lossless Compression
https://openreview.net/forum?id=MBOyiNnYthd
[ "Rianne van den Berg", "Alexey A. Gritsenko", "Mostafa Dehghani", "Casper Kaae Sønderby", "Tim Salimans" ]
Poster
null
In this paper we analyse and improve integer discrete flows for lossless compression. Integer discrete flows are a recently proposed class of models that learn invertible transformations for integer-valued random variables. Their discrete nature makes them particularly suitable for lossless compression with entropy cod...
[ "normalizing flows", "lossless source compression", "generative modeling" ]
null
3,291
2006.12459
title_snapshot
[ -0.01402429211884737, -0.03422108292579651, -0.00005556367977987975, 0.03797163814306259, 0.03531109169125557, 0.046293389052152634, 0.01885366626083851, -0.013104820623993874, -0.02663276717066765, -0.03666913881897926, -0.006162854377180338, -0.02768639847636223, -0.03896602988243103, -0...
Fully Unsupervised Diversity Denoising with Convolutional Variational Autoencoders
https://openreview.net/forum?id=agHLCOBM5jP
[ "Mangal Prakash", "Alexander Krull", "Florian Jug" ]
Poster
null
Deep Learning based methods have emerged as the indisputable leaders for virtually all image restoration tasks. Especially in the domain of microscopy images, various content-aware image restoration (CARE) approaches are now used to improve the interpretability of acquired data. Naturally, there are limitations to what...
[ "Diversity denoising", "Unsupervised denoising", "Variational Autoencoders", "Noise model" ]
null
3,284
2006.06072
title_snapshot
[ 0.00990725215524435, -0.026543457061052322, -0.019628964364528656, 0.04988137260079384, 0.033351968973875046, 0.03984251245856285, 0.0460568405687809, -0.004880099557340145, -0.02586132287979126, -0.07371481508016586, -0.004916362464427948, 0.022140614688396454, -0.057813964784145355, 0.03...
Is Attention Better Than Matrix Decomposition?
https://openreview.net/forum?id=1FvkSpWosOl
[ "Zhengyang Geng", "Meng-Hao Guo", "Hongxu Chen", "Xia Li", "Ke Wei", "Zhouchen Lin" ]
Poster
null
As an essential ingredient of modern deep learning, attention mechanism, especially self-attention, plays a vital role in the global correlation discovery. However, is hand-crafted attention irreplaceable when modeling the global context? Our intriguing finding is that self-attention is not better than the matrix decom...
[ "attention models", "matrix decomposition", "computer vision" ]
null
3,278
2109.04553
title_snapshot
[ 0.001876556547358632, 0.007022878620773554, 0.011993915773928165, 0.04261890426278114, 0.021034104749560356, 0.014647169969975948, 0.02273528091609478, 0.010823925957083702, -0.033087316900491714, -0.04605123773217201, -0.032685406506061554, -0.005817879922688007, -0.05744054913520813, 0.0...
Improving Transformation Invariance in Contrastive Representation Learning
https://openreview.net/forum?id=NomEDgIEBwE
[ "Adam Foster", "Rattana Pukdee", "Tom Rainforth" ]
Poster
null
We propose methods to strengthen the invariance properties of representations obtained by contrastive learning. While existing approaches implicitly induce a degree of invariance as representations are learned, we look to more directly enforce invariance in the encoding process. To this end, we first introduce a traini...
[ "contrastive learning", "representation learning", "transformation invariance" ]
null
3,273
2010.09515
title_snapshot
[ -0.006336311809718609, -0.016585715115070343, -0.014960573986172676, 0.04981964826583862, 0.030106347054243088, 0.023090293630957603, 0.04551086574792862, -0.004099319688975811, -0.034086257219314575, -0.01030658558011055, -0.010015066713094711, -0.020214427262544632, -0.05679825693368912, ...
On the Origin of Implicit Regularization in Stochastic Gradient Descent
https://openreview.net/forum?id=rq_Qr0c1Hyo
[ "Samuel L Smith", "Benoit Dherin", "David Barrett", "Soham De" ]
Poster
null
For infinitesimal learning rates, stochastic gradient descent (SGD) follows the path of gradient flow on the full batch loss function. However moderately large learning rates can achieve higher test accuracies, and this generalization benefit is not explained by convergence bounds, since the learning rate which maximiz...
[ "SGD", "learning rate", "batch size", "optimization", "generalization", "implicit regularization", "backward error analysis", "SDE", "stochastic differential equation", "ODE", "ordinary differential equation" ]
null
3,269
2101.12176
title_snapshot
[ -0.04499092325568199, -0.030397070571780205, -0.00654737651348114, 0.0397522933781147, 0.02800425887107849, 0.03585655242204666, 0.031783897429704666, 0.0333750918507576, -0.019510919228196144, -0.04132963344454765, -0.005875015631318092, -0.0049197375774383545, -0.052186962217092514, 0.00...
Share or Not? Learning to Schedule Language-Specific Capacity for Multilingual Translation
https://openreview.net/forum?id=Wj4ODo0uyCF
[ "Biao Zhang", "Ankur Bapna", "Rico Sennrich", "Orhan Firat" ]
Oral
null
Using a mix of shared and language-specific (LS) parameters has shown promise in multilingual neural machine translation (MNMT), but the question of when and where LS capacity matters most is still under-studied. We offer such a study by proposing conditional language-specific routing (CLSR). CLSR employs hard binary ...
[ "language-specific modeling", "conditional computation", "multilingual translation", "multilingual transformer" ]
null
3,265
null
null
[ -0.01730526052415371, -0.04084562510251999, -0.03551962226629257, 0.04292791709303856, 0.04594741016626358, 0.04010036215186119, 0.024766646325588226, 0.024725371971726418, -0.012091662734746933, -0.014828433282673359, -0.022282211109995842, 0.006578645668923855, -0.0656641498208046, 0.001...
Transient Non-stationarity and Generalisation in Deep Reinforcement Learning
https://openreview.net/forum?id=Qun8fv4qSby
[ "Maximilian Igl", "Gregory Farquhar", "Jelena Luketina", "Wendelin Boehmer", "Shimon Whiteson" ]
Poster
null
Non-stationarity can arise in Reinforcement Learning (RL) even in stationary environments. For example, most RL algorithms collect new data throughout training, using a non-stationary behaviour policy. Due to the transience of this non-stationarity, it is often not explicitly addressed in deep RL and a single neural ne...
[ "Reinforcement Learning", "Generalization" ]
null
3,261
2006.05826
title_snapshot
[ -0.021154241636395454, -0.05416647344827652, 0.007030697073787451, 0.04650946706533432, 0.02179909124970436, 0.01716652326285839, 0.03496883809566498, 0.023764656856656075, -0.03579032048583031, -0.033993031829595566, -0.023352021351456642, -0.00813325960189104, -0.05539965257048607, -0.00...
Lossless Compression of Structured Convolutional Models via Lifting
https://openreview.net/forum?id=oxnp2q-PGL4
[ "Gustav Sourek", "Filip Zelezny", "Ondrej Kuzelka" ]
Poster
null
Lifting is an efficient technique to scale up graphical models generalized to relational domains by exploiting the underlying symmetries. Concurrently, neural models are continuously expanding from grid-like tensor data into structured representations, such as various attributed graphs and relational databases. To addr...
[ "weight sharing", "graph neural networks", "lifted inference", "relational learning", "dynamic computation graphs", "convolutional models" ]
null
3,256
2007.06567
title_snapshot
[ -0.01895288936793804, -0.03075093775987625, -0.009602305479347706, 0.03818648308515549, 0.033674921840429306, 0.029121724888682365, -0.013856207951903343, 0.005816953256726265, 0.002164265373721719, -0.0587700754404068, -0.008885381743311882, 0.007740419823676348, -0.08202068507671356, 0.0...
Analyzing the Expressive Power of Graph Neural Networks in a Spectral Perspective
https://openreview.net/forum?id=-qh0M9XWxnv
[ "Muhammet Balcilar", "Guillaume Renton", "Pierre Héroux", "Benoit Gaüzère", "Sébastien Adam", "Paul Honeine" ]
Poster
null
In the recent literature of Graph Neural Networks (GNN), the expressive power of models has been studied through their capability to distinguish if two given graphs are isomorphic or not. Since the graph isomorphism problem is NP-intermediate, and Weisfeiler-Lehman (WL) test can give sufficient but not enough evidence ...
[ "Graph Neural Networks", "Spectral Graph Filter", "Spectral Analysis" ]
null
3,251
null
null
[ -0.034289274364709854, -0.02957279235124588, 0.014876357279717922, 0.04294486716389656, 0.028438856825232506, 0.015008865855634212, 0.030795369297266006, 0.011410115286707878, -0.00018936485867016017, -0.05353261157870293, 0.02569904737174511, -0.00168059510178864, -0.07442163676023483, -0...
End-to-end Adversarial Text-to-Speech
https://openreview.net/forum?id=rsf1z-JSj87
[ "Jeff Donahue", "Sander Dieleman", "Mikolaj Binkowski", "Erich Elsen", "Karen Simonyan" ]
Oral
null
Modern text-to-speech synthesis pipelines typically involve multiple processing stages, each of which is designed or learnt independently from the rest. In this work, we take on the challenging task of learning to synthesise speech from normalised text or phonemes in an end-to-end manner, resulting in models which oper...
[ "text-to-speech", "speech synthesis", "adversarial", "GAN", "end-to-end", "feed-forward", "generative model" ]
null
3,246
2006.03575
title_snapshot
[ 0.0016512476140633225, -0.02692040242254734, -0.015249975025653839, 0.035582225769758224, 0.0032980446703732014, 0.03357759118080139, 0.0495271272957325, 0.017914660274982452, -0.025572458282113075, -0.037030644714832306, -0.024032454937696457, 0.021144898608326912, -0.044241148978471756, ...
A unifying view on implicit bias in training linear neural networks
https://openreview.net/forum?id=ZsZM-4iMQkH
[ "Chulhee Yun", "Shankar Krishnan", "Hossein Mobahi" ]
Poster
null
We study the implicit bias of gradient flow (i.e., gradient descent with infinitesimal step size) on linear neural network training. We propose a tensor formulation of neural networks that includes fully-connected, diagonal, and convolutional networks as special cases, and investigate the linear version of the formulat...
[ "implicit bias", "implicit regularization", "convergence", "gradient flow", "gradient descent" ]
null
3,245
2010.02501
title_snapshot
[ -0.035936709493398666, -0.007350574247539043, 0.011446862481534481, 0.021700982004404068, 0.017261844128370285, 0.039575204253196716, 0.013180586509406567, 0.008984837681055069, -0.014886787161231041, -0.04974478855729103, -0.021580062806606293, -0.0023811988066881895, -0.07352209091186523, ...
Balancing Constraints and Rewards with Meta-Gradient D4PG
https://openreview.net/forum?id=TQt98Ya7UMP
[ "Dan A. Calian", "Daniel J Mankowitz", "Tom Zahavy", "Zhongwen Xu", "Junhyuk Oh", "Nir Levine", "Timothy Mann" ]
Poster
null
Deploying Reinforcement Learning (RL) agents to solve real-world applications often requires satisfying complex system constraints. Often the constraint thresholds are incorrectly set due to the complex nature of a system or the inability to verify the thresholds offline (e.g, no simulator or reasonable offline evaluat...
[ "reinforcement learning", "meta-gradients", "constraints" ]
null
3,232
2010.06324
title_snapshot
[ -0.042742013931274414, -0.027631698176264763, 0.0028252110350877047, 0.04506010189652443, 0.044406287372112274, 0.0187063068151474, 0.015887457877397537, 0.0059167612344026566, -0.05476114898920059, -0.01491981279104948, -0.038474924862384796, 0.043533116579055786, -0.0861423909664154, -0....
Robust Curriculum Learning: from clean label detection to noisy label self-correction
https://openreview.net/forum?id=lmTWnm3coJJ
[ "Tianyi Zhou", "Shengjie Wang", "Jeff Bilmes" ]
Poster
null
Neural network training can easily overfit noisy labels resulting in poor generalization performance. Existing methods address this problem by (1) filtering out the noisy data and only using the clean data for training or (2) relabeling the noisy data by the model during training or by another model trained only on a c...
[ "curriculum learning", "noisy label", "robust learning", "training dynamics", "neural networks" ]
null
3,222
null
null
[ 0.01122095063328743, -0.026385128498077393, -0.019320957362651825, 0.07374835759401321, 0.04993610084056854, 0.01467250008136034, 0.016205063089728355, -0.012444373220205307, -0.0279766246676445, -0.030691923573613167, -0.023321572691202164, 0.03892939165234566, -0.04959637299180031, -0.01...
Clairvoyance: A Pipeline Toolkit for Medical Time Series
https://openreview.net/forum?id=xnC8YwKUE3k
[ "Daniel Jarrett", "Jinsung Yoon", "Ioana Bica", "Zhaozhi Qian", "Ari Ercole", "Mihaela van der Schaar" ]
Poster
null
Time-series learning is the bread and butter of data-driven *clinical decision support*, and the recent explosion in ML research has demonstrated great potential in various healthcare settings. At the same time, medical time-series problems in the wild are challenging due to their highly *composite* nature: They entail...
[ "reproducibility", "healthcare", "medical time series", "pipeline toolkit", "software" ]
null
3,220
2310.18688
title_snapshot
[ 0.017140716314315796, -0.04185570776462555, -0.011708071455359459, 0.0019904552027583122, 0.059229690581560135, 0.04032246395945549, 0.057977814227342606, -0.00549759715795517, 0.0008261936018243432, -0.07459553331136703, 0.03085169568657875, 0.014349089935421944, -0.033342231065034866, 0....
Plan-Based Relaxed Reward Shaping for Goal-Directed Tasks
https://openreview.net/forum?id=w2Z2OwVNeK
[ "Ingmar Schubert", "Ozgur S Oguz", "Marc Toussaint" ]
Poster
null
In high-dimensional state spaces, the usefulness of Reinforcement Learning (RL) is limited by the problem of exploration. This issue has been addressed using potential-based reward shaping (PB-RS) previously. In the present work, we introduce Final-Volume-Preserving Reward Shaping (FV-RS). FV-RS relaxes the strict opti...
[ "reinforcement learning", "reward shaping", "plan-based reward shaping", "robotics", "robotic manipulation" ]
null
3,201
2107.06661
title_snapshot
[ -0.028032729402184486, -0.035959936678409576, 0.003190786112099886, 0.029037121683359146, 0.04356550797820091, 0.04295644164085388, -0.010822581127285957, -0.02363167516887188, -0.04324765503406525, -0.046405013650655746, -0.018069718033075333, 0.008043705485761166, -0.04899812862277031, -...
Improving VAEs' Robustness to Adversarial Attack
https://openreview.net/forum?id=-Hs_otp2RB
[ "Matthew JF Willetts", "Alexander Camuto", "Tom Rainforth", "S Roberts", "Christopher C Holmes" ]
Poster
null
Variational autoencoders (VAEs) have recently been shown to be vulnerable to adversarial attacks, wherein they are fooled into reconstructing a chosen target image. However, how to defend against such attacks remains an open problem. We make significant advances in addressing this issue by introducing methods for produ...
[ "deep generative models", "variational autoencoders", "robustness", "adversarial attack" ]
null
3,195
1906.00230
title_snapshot
[ 0.015616210177540779, 0.013128108344972134, -0.012845480814576149, 0.04995031654834747, 0.025701111182570457, 0.02907502092421055, 0.07300730794668198, -0.01733364537358284, -0.02733427844941616, -0.04372376203536987, -0.03424162045121193, -0.01887255348265171, -0.05967046692967415, 0.0077...
Gauge Equivariant Mesh CNNs: Anisotropic convolutions on geometric graphs
https://openreview.net/forum?id=Jnspzp-oIZE
[ "Pim De Haan", "Maurice Weiler", "Taco Cohen", "Max Welling" ]
Spotlight
null
A common approach to define convolutions on meshes is to interpret them as a graph and apply graph convolutional networks (GCNs). Such GCNs utilize isotropic kernels and are therefore insensitive to the relative orientation of vertices and thus to the geometry of the mesh as a whole. We propose Gauge Equivariant Mesh ...
[ "symmetry", "equivariance", "mesh", "geometric", "convolution" ]
null
3,192
2003.05425
title_snapshot
[ 0.010973791591823101, -0.015987757593393326, 0.021233508363366127, 0.04322570562362671, 0.009053573943674564, 0.033483851701021194, 0.015685612335801125, 0.037835028022527695, -0.029128527268767357, -0.08599479496479034, 0.006332125980407, -0.0422835499048233, -0.046331774443387985, 0.0476...
Differentiable Segmentation of Sequences
https://openreview.net/forum?id=4T489T4yav
[ "Erik Scharwächter", "Jonathan Lennartz", "Emmanuel Müller" ]
Poster
null
Segmented models are widely used to describe non-stationary sequential data with discrete change points. Their estimation usually requires solving a mixed discrete-continuous optimization problem, where the segmentation is the discrete part and all other model parameters are continuous. A number of estimation algorithm...
[ "segmented models", "segmentation", "change point detection", "concept drift", "warping functions", "gradient descent" ]
null
3,185
2006.13105
title_snapshot
[ -0.001888734521344304, -0.02850557304918766, -0.01585855521261692, 0.01707584410905838, 0.027311861515045166, 0.03303058445453644, 0.038887135684490204, 0.03694555163383484, -0.006686765234917402, -0.04277081415057182, 0.01577424630522728, -0.02333230711519718, -0.05780511349439621, 0.0173...
Generalization bounds via distillation
https://openreview.net/forum?id=EGdFhBzmAwB
[ "Daniel Hsu", "Ziwei Ji", "Matus Telgarsky", "Lan Wang" ]
Spotlight
null
This paper theoretically investigates the following empirical phenomenon: given a high-complexity network with poor generalization bounds, one can distill it into a network with nearly identical predictions but low complexity and vastly smaller generalization bounds. The main contribution is an analysis showing that t...
[ "Generalization", "statistical learning theory", "theory", "distillation" ]
null
3,178
2104.05641
title_snapshot
[ 0.0040093413554131985, -0.015920644626021385, -0.004315576981753111, 0.06867031008005142, 0.04059090465307236, 0.013593166135251522, 0.03631080314517021, -0.003435001941397786, -0.01955757662653923, -0.024738868698477745, -0.006358801387250423, -0.03630586713552475, -0.08809198439121246, -...
Learning Mesh-Based Simulation with Graph Networks
https://openreview.net/forum?id=roNqYL0_XP
[ "Tobias Pfaff", "Meire Fortunato", "Alvaro Sanchez-Gonzalez", "Peter Battaglia" ]
Spotlight
null
Mesh-based simulations are central to modeling complex physical systems in many disciplines across science and engineering. Mesh representations support powerful numerical integration methods and their resolution can be adapted to strike favorable trade-offs between accuracy and efficiency. However, high-dimensional sc...
[ "graph networks", "simulation", "mesh", "physics" ]
null
3,177
2010.03409
title_snapshot
[ -0.030339447781443596, -0.009005286730825901, 0.005405770149081945, 0.05836399644613266, 0.0399467833340168, 0.03466423228383064, -0.0011786242248490453, -0.007876611314713955, -0.04512565955519676, -0.07208554446697235, 0.03280389681458473, -0.018191158771514893, -0.04380468651652336, 0.0...
GraPPa: Grammar-Augmented Pre-Training for Table Semantic Parsing
https://openreview.net/forum?id=kyaIeYj4zZ
[ "Tao Yu", "Chien-Sheng Wu", "Xi Victoria Lin", "bailin wang", "Yi Chern Tan", "Xinyi Yang", "Dragomir Radev", "richard socher", "Caiming Xiong" ]
Poster
null
We present GraPPa, an effective pre-training approach for table semantic parsing that learns a compositional inductive bias in the joint representations of textual and tabular data. We construct synthetic question-SQL pairs over high-quality tables via a synchronous context-free grammar (SCFG). We pre-train our model o...
[ "text-to-sql", "semantic parsing", "pre-training", "nlp" ]
null
3,175
2009.13845
title_snapshot
[ -0.014219379052519798, 0.010891013778746128, -0.013582129031419754, 0.0649389773607254, 0.04057050496339798, 0.016571147367358208, 0.021685220301151276, 0.011996500194072723, -0.027163857594132423, -0.009456703439354897, -0.030620209872722626, 0.016526808962225914, -0.08076193183660507, -0...
Sliced Kernelized Stein Discrepancy
https://openreview.net/forum?id=t0TaKv0Gx6Z
[ "Wenbo Gong", "Yingzhen Li", "José Miguel Hernández-Lobato" ]
Poster
null
Kernelized Stein discrepancy (KSD), though being extensively used in goodness-of-fit tests and model learning, suffers from the curse-of-dimensionality. We address this issue by proposing the sliced Stein discrepancy and its scalable and kernelized variants, which employs kernel-based test functions defined on the opti...
[ "kernel methods", "variational inference", "particle inference" ]
null
3,174
2006.16531
title_snapshot
[ -0.02193010225892067, -0.016310498118400574, 0.01493358425796032, 0.055371470749378204, 0.036417387425899506, 0.03771425038576126, 0.03260277956724167, -0.024694260209798813, -0.015779217705130577, -0.03965235874056816, -0.0016416928265243769, 0.007374824024736881, -0.05529491603374481, 0....
Variational Intrinsic Control Revisited
https://openreview.net/forum?id=P0p33rgyoE
[ "Taehwan Kwon" ]
Poster
null
In this paper, we revisit variational intrinsic control (VIC), an unsupervised reinforcement learning method for finding the largest set of intrinsic options available to an agent. In the original work by Gregor et al. (2016), two VIC algorithms were proposed: one that represents the options explicitly, and the other t...
[ "Unsupervised reinforcement learning", "Information theory" ]
null
3,173
2010.03281
title_snapshot
[ -0.034006744623184204, -0.017942382022738457, -0.010923066176474094, 0.04268168285489082, 0.03599005565047264, 0.020514192059636116, 0.01676887646317482, 0.0039030462503433228, -0.029036641120910645, -0.04777965322136879, -0.039015159010887146, 0.03710036724805832, -0.06562073528766632, 0....
On Statistical Bias In Active Learning: How and When to Fix It
https://openreview.net/forum?id=JiYq3eqTKY
[ "Sebastian Farquhar", "Yarin Gal", "Tom Rainforth" ]
Spotlight
null
Active learning is a powerful tool when labelling data is expensive, but it introduces a bias because the training data no longer follows the population distribution. We formalize this bias and investigate the situations in which it can be harmful and sometimes even helpful. We further introduce novel corrective weight...
[ "Active Learning", "Monte Carlo", "Risk Estimation" ]
null
3,171
2101.11665
title_snapshot
[ -0.026572169736027718, -0.03464566543698311, -0.0187164805829525, 0.028944773599505424, 0.022616423666477203, 0.009618024341762066, -0.014045672491192818, -0.008858933113515377, -0.017285166308283806, -0.030669376254081726, -0.03748444840312004, 0.014825025573372841, -0.0719435065984726, -...
Winning the L2RPN Challenge: Power Grid Management via Semi-Markov Afterstate Actor-Critic
https://openreview.net/forum?id=LmUJqB1Cz8
[ "Deunsol Yoon", "Sunghoon Hong", "Byung-Jun Lee", "Kee-Eung Kim" ]
Spotlight
null
Safe and reliable electricity transmission in power grids is crucial for modern society. It is thus quite natural that there has been a growing interest in the automatic management of power grids, exemplified by the Learning to Run a Power Network Challenge (L2RPN), modeling the problem as a reinforcement learning (RL) ...
[ "power grid management", "deep reinforcement learning", "graph neural network" ]
null
3,169
null
null
[ -0.02436227537691593, -0.07886283844709396, 0.011980392970144749, 0.025554167106747627, 0.02795829437673092, -0.022237177938222885, 0.014202889055013657, -0.009782298468053341, -0.023367542773485184, -0.021147364750504494, -0.0029025464318692684, 0.01670186035335064, -0.07091744989156723, ...
HyperDynamics: Meta-Learning Object and Agent Dynamics with Hypernetworks
https://openreview.net/forum?id=pHXfe1cOmA
[ "Zhou Xian", "Shamit Lal", "Hsiao-Yu Tung", "Emmanouil Antonios Platanios", "Katerina Fragkiadaki" ]
Poster
null
We propose HyperDynamics, a dynamics meta-learning framework that conditions on an agent’s interactions with the environment and optionally its visual observations, and generates the parameters of neural dynamics models based on inferred properties of the dynamical system. Physical and visual properties of the environm...
[]
null
3,167
2103.09439
title_snapshot
[ -0.010713684372603893, -0.005772076081484556, 0.011955467984080315, 0.038090016692876816, 0.0394054539501667, 0.008315211161971092, 0.019773127511143684, -0.004952172748744488, -0.0778743177652359, -0.02733793295919895, -0.008442762307822704, 0.010131536051630974, -0.05053221806883812, -0....