NeurIPS 2020

1,908 papers

Top Research Directions

Research directions most represented by NeurIPS 2020 papers.

Clean Accuracy Degradation in Machine Learning Defenses 21
Bypassing the Computational Bottleneck of Adversarial Training 20
Scaling Gaussian Processes Beyond Cubic Complexity 19
Safe Reinforcement Learning via Constrained Markov Decision Processes 17
The Non-Linear Function Approximation Gap in Reinforcement Learning 17
The Mechanics of Benign Overfitting in High Dimensions 15
Optimization and Generalization with Non-Smooth Activations 14
Multi-Agent Learning Under Bandit Feedback 13
Mitigating Gradient Variance in Policy Optimization 13
The Infinite Width Approximation Gap in Deep Learning 12
The Statistical Computational Gap in High Dimensions 12
Scaling Bandit Learning to Large Action Spaces 12
The Function Approximation Gap in Reinforcement Learning Theory 12
Linearizing Deep Learning via the Neural Tangent Kernel 11
Sample Efficient Reinforcement Learning 11

Top Authors

AuthorPapers
Sergey Levine11
Mihaela van der Schaar10
Csaba Szepesvari9
Jinwoo Shin9
Michael W. Mahoney9
Stefano Ermon9
Sung Ju Hwang9
Zhaoran Wang9
Pieter Abbeel8
Zhangyang Wang8
Zhuoran Yang8
Andreas Krause7
Bo Dai7
Dacheng Tao7
Jason D. Lee7
Jun Zhu7
Masashi Sugiyama7
Michael Jordan7
Akshay Krishnamurthy6
Andrea Vedaldi6

Impactful Papers

Papers from NeurIPS 2020 that shaped research clusters.

Foundational

Influential

Papers

Title
The Statistical Cost of Robust Kernel Hyperparameter Turning
The Strong Screening Rule for SLOPE
The Surprising Simplicity of the Early-Time Learning Dynamics of Neural Networks
The Value Equivalence Principle for Model-Based Reinforcement Learning
The Wasserstein Proximal Gradient Algorithm
ThreeDWorld: A Platform for Interactive Multi-Modal Physical Simulation
Throughput-Optimal Topology Design for Cross-Silo Federated Learning
Thunder: a Fast Coordinate Selection Solver for Sparse Learning
Tight First- and Second-Order Regret Bounds for Adversarial Linear Bandits
Tight last-iterate convergence rates for no-regret learning in multi-player games
Tight Nonparametric Convergence Rates for Stochastic Gradient Descent under the Noiseless Linear Model
Time-Reversal Symmetric ODE Network
Timeseries Anomaly Detection using Temporal Hierarchical One-Class Network
TinyTL: Reduce Memory, Not Parameters for Efficient On-Device Learning
Top-KAST: Top-K Always Sparse Training
Top-k Training of GANs: Improving GAN Performance by Throwing Away Bad Samples
TorsionNet: A Reinforcement Learning Approach to Sequential Conformer Search
Towards a Better Global Loss Landscape of GANs
Towards a Combinatorial Characterization of Bounded-Memory Learning
Towards Better Generalization of Adaptive Gradient Methods
Towards Convergence Rate Analysis of Random Forests for Classification
Towards Crowdsourced Training of Large Neural Networks using Decentralized Mixture-of-Experts
Towards Deeper Graph Neural Networks with Differentiable Group Normalization
Towards Interpretable Natural Language Understanding with Explanations as Latent Variables
Towards Learning Convolutions from Scratch
Towards Maximizing the Representation Gap between In-Domain & Out-of-Distribution Examples
Towards Minimax Optimal Reinforcement Learning in Factored Markov Decision Processes
Towards More Practical Adversarial Attacks on Graph Neural Networks
Towards Neural Programming Interfaces
Towards Playing Full MOBA Games with Deep Reinforcement Learning
Towards practical differentially private causal graph discovery
Towards Problem-dependent Optimal Learning Rates
Towards robust and domain agnostic reinforcement learning competitions: MineRL 2020
Towards Safe Policy Improvement for Non-Stationary MDPs
Towards Scalable Bayesian Learning of Causal DAGs
Towards Scale-Invariant Graph-related Problem Solving by Iterative Homogeneous GNNs
Towards Theoretically Understanding Why Sgd Generalizes Better Than Adam in Deep Learning
Towards Understanding Hierarchical Learning: Benefits of Neural Representations
Toward the Fundamental Limits of Imitation Learning
Trade-offs and Guarantees of Adversarial Representation Learning for Information Obfuscation
Trading Personalization for Accuracy: Data Debugging in Collaborative Filtering
Traffic4cast at NeurIPS 2020 ? yet more on theunreasonable effectiveness of gridded geo-spatial processes
Train-by-Reconnect: Decoupling Locations of Weights from Their Values
Training Generative Adversarial Networks by Solving Ordinary Differential Equations
Training Generative Adversarial Networks with Limited Data
Training Linear Finite-State Machines
Training Normalizing Flows with the Information Bottleneck for Competitive Generative Classification
Training Stronger Baselines for Learning to Optimize
Trajectory-wise Multiple Choice Learning for Dynamics Generalization in Reinforcement Learning
Transferable Calibration with Lower Bias and Variance in Domain Adaptation

« prev Page 36/39 next »