NeurIPS 2021 accepted papers on Reinforcement learning
216 of the 2,508 papers accepted at NeurIPS 2021 (8.6%) are on Reinforcement learning, compared with 8.6% at NeurIPS 2020. They fall into 6 topics.
See NeurIPS 2021 in the interactive map
Submitting to NeurIPS? Checklist →
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Visual and representation RL · 87 papers
- Isaac Gym: High Performance GPU Based Physics Simulation For Robot Learning (59 citations)
- A Hierarchical Reinforcement Learning Based Optimization Framework for Large-scale Dynamic Pickup and Delivery Problems (49 citations)
- Learning Collaborative Policies to Solve NP-hard Routing Problems (43 citations)
- Mastering Atari Games with Limited Data (39 citations)
- Stabilizing Deep Q-Learning with ConvNets and Vision Transformers under Data Augmentation (37 citations)
- On Joint Learning for Solving Placement and Routing in Chip Design (34 citations)
- Pretraining Representations for Data-Efficient Reinforcement Learning (33 citations)
- Automatic Data Augmentation for Generalization in Reinforcement Learning (30 citations)
- Habitat 2.0: Training Home Assistants to Rearrange their Habitat (24 citations)
- Brax - A Differentiable Physics Engine for Large Scale Rigid Body Simulation (21 citations)
- Hit and Lead Discovery with Explorative RL and Fragment-based Molecule Generation (20 citations)
- Landmark-Guided Subgoal Generation in Hierarchical Reinforcement Learning (19 citations)
and 75 more in the interactive map.
Policy gradient and actor-critic · 43 papers
- Empirical Study of Off-Policy Policy Evaluation for Reinforcement Learning (61 citations)
- Deep Reinforcement Learning at the Edge of the Statistical Precipice (41 citations)
- Generalized Proximal Policy Optimization with Sample Reuse (21 citations)
- Accelerating Quadratic Optimization with Reinforcement Learning (20 citations)
- On the Convergence and Sample Efficiency of Variance-Reduced Policy Gradient Method (18 citations)
- Local policy search with Bayesian optimization (17 citations)
- Tactical Optimism and Pessimism for Deep Reinforcement Learning (13 citations)
- Learning Large Neighborhood Search Policy for Integer Programming (11 citations)
- Stabilizing Dynamical Systems via Policy Gradient Methods (11 citations)
- Derivative-Free Policy Optimization for Linear Risk-Sensitive and Robust Control Design: Implicit Regularization and Sample Complexity (8 citations)
- Time Discretization-Invariant Safe Action Repetition for Policy Gradient Methods (7 citations)
- Counterexample Guided RL Policy Refinement Using Bayesian Optimization (6 citations)
and 31 more in the interactive map.
MDPs and planning · 39 papers
- Bellman Eluder Dimension: New Rich Classes of RL Problems, and Sample-Efficient Algorithms (28 citations)
- Near-Optimal Offline Reinforcement Learning via Double Variance Reduction (18 citations)
- Policy Finetuning: Bridging Sample-Efficient Offline and Online Reinforcement Learning (16 citations)
- MICo: Improved representations via sampling-based state similarity for Markov decision processes (10 citations)
- Why Generalization in RL is Difficult: Epistemic POMDPs and Implicit Partial Observability (10 citations)
- Learning MDPs from Features: Predict-Then-Optimize for Sequential Decision Making by Reinforcement Learning (8 citations)
- Reinforcement Learning with State Observation Costs in Action-Contingent Noiselessly Observable Markov Decision Processes (7 citations)
- Finite Sample Analysis of Average-Reward TD Learning and $Q$-Learning (6 citations)
- Learning Markov State Abstractions for Deep Reinforcement Learning (6 citations)
- Nearly Horizon-Free Offline Reinforcement Learning (6 citations)
- Online Robust Reinforcement Learning with Model Uncertainty (6 citations)
- Reward-Free Model-Based Reinforcement Learning with Linear Function Approximation (6 citations)
and 27 more in the interactive map.
Offline RL · 24 papers
- Decision Transformer: Reinforcement Learning via Sequence Modeling (464 citations)
- A Minimalist Approach to Offline Reinforcement Learning (162 citations)
- Bridging Offline Reinforcement Learning and Imitation Learning: A Tale of Pessimism (55 citations)
- COMBO: Conservative Offline Model-Based Policy Optimization (51 citations)
- Uncertainty-Based Offline Reinforcement Learning with Diversified Q-Ensemble (48 citations)
- Offline Reinforcement Learning as One Big Sequence Modeling Problem (41 citations)
- Online and Offline Reinforcement Learning by Planning with a Learned Model (23 citations)
- Offline Reinforcement Learning with Reverse Model-based Imagination (21 citations)
- Offline RL Without Off-Policy Evaluation (16 citations)
- Bellman-consistent Pessimism for Offline Reinforcement Learning (15 citations)
- Conservative Offline Distributional Reinforcement Learning (15 citations)
- Provable Benefits of Actor-Critic Methods for Offline Reinforcement Learning (14 citations)
and 12 more in the interactive map.
Imitation learning · 21 papers
- Confidence-Aware Imitation Learning from Demonstrations with Varying Optimality (19 citations)
- What Matters for Adversarial Imitation Learning? (18 citations)
- Provable Representation Learning for Imitation with Contrastive Fourier Features (7 citations)
- An Empirical Investigation of Representation Learning for Imitation (6 citations)
- Policy Learning Using Weak Supervision (6 citations)
- Curriculum Offline Imitating Learning (5 citations)
- Identifiability in inverse reinforcement learning (5 citations)
- Mitigating Covariate Shift in Imitation Learning via Offline Data With Partial Coverage (5 citations)
- Visual Adversarial Imitation Learning using Variational Models (5 citations)
- Imitation with Neural Density Models (4 citations)
- Curriculum Design for Teaching via Demonstrations: Theory and Applications (3 citations)
- Generalizable Imitation Learning from Observation via Inferring Goal Proximity (3 citations)
and 9 more in the interactive map.
