NeurIPS 2023 accepted papers on Reinforcement learning

Visual and representation RL · 102 papers

• A Toolkit for Reliable Benchmarking and Research in Multi-Objective Reinforcement Learning (20 citations)
• Learning to Search Feasible and Infeasible Regions of Routing Problems with Flexible Neural k-Opt (18 citations)
• Minigrid & Miniworld: Modular & Customizable Reinforcement Learning Environments for Goal-Oriented Tasks (16 citations)
• Safety Gymnasium: A Unified Safe Reinforcement Learning Benchmark (15 citations)
• Diffusion Model is an Effective Planner and Data Synthesizer for Multi-Task Reinforcement Learning (12 citations)
• Synthetic Experience Replay (12 citations)
• Structured State Space Models for In-Context Reinforcement Learning (11 citations)
• Learning Better with Less: Effective Augmentation for Sample-Efficient Visual Reinforcement Learning (10 citations)
• STORM: Efficient Stochastic Transformer based World Models for Reinforcement Learning (10 citations)
• Swarm Reinforcement Learning for Adaptive Mesh Refinement (9 citations)
• KuaiSim: A Comprehensive Simulator for Recommender Systems (8 citations)
• Latent exploration for Reinforcement Learning (7 citations)

and 90 more in the interactive map.

Policy gradient and actor-critic · 58 papers

• Policy Optimization for Continuous Reinforcement Learning (6 citations)
• Distributional Pareto-Optimal Multi-Objective Reinforcement Learning (4 citations)
• Gradient Informed Proximal Policy Optimization (4 citations)
• Prioritizing Samples in Reinforcement Learning with Reducible Loss (4 citations)
• Safe Exploration in Reinforcement Learning: A Generalized Formulation and Algorithms (4 citations)
• Adjustable Robust Reinforcement Learning for Online 3D Bin Packing (3 citations)
• An Alternative to Variance: Gini Deviation for Risk-averse Policy Gradient (3 citations)
• Constraint-Conditioned Policy Optimization for Versatile Safe Reinforcement Learning (3 citations)
• Natural Actor-Critic for Robust Reinforcement Learning with Function Approximation (3 citations)
• Sample-Efficient and Safe Deep Reinforcement Learning via Reset Deep Ensemble Agents (3 citations)
• Small batch deep reinforcement learning (3 citations)
• Generative Modelling of Stochastic Actions with Arbitrary Constraints in Reinforcement Learning (2 citations)

and 46 more in the interactive map.

Offline RL · 49 papers

• Cal-QL: Calibrated Offline RL Pre-Training for Efficient Online Fine-Tuning (20 citations)
• Efficient Diffusion Policies For Offline Reinforcement Learning (12 citations)
• CORL: Research-oriented Deep Offline Reinforcement Learning Library (9 citations)
• Context Shift Reduction for Offline Meta-Reinforcement Learning (7 citations)
• Adversarial Counterfactual Environment Model Learning (6 citations)
• Learning to Influence Human Behavior with Offline Reinforcement Learning (6 citations)
• Adversarial Model for Offline Reinforcement Learning (4 citations)
• Beyond Uniform Sampling: Offline Reinforcement Learning with Imbalanced Datasets (4 citations)
• Revisiting the Minimalist Approach to Offline Reinforcement Learning (4 citations)
• Train Once, Get a Family: State-Adaptive Balances for Offline-to-Online Reinforcement Learning (4 citations)
• Conformal Prediction for Uncertainty-Aware Planning with Diffusion Dynamics Model (3 citations)
• Offline Reinforcement Learning with Differential Privacy (3 citations)

and 37 more in the interactive map.

MDPs and planning · 37 papers

• BQ-NCO: Bisimulation Quotienting for Efficient Neural Combinatorial Optimization (10 citations)
• A Theoretical Analysis of Optimistic Proximal Policy Optimization in Linear Markov Decision Processes (7 citations)
• On the Convergence and Sample Complexity Analysis of Deep Q-Networks with $\epsilon$-Greedy Exploration (5 citations)
• Policy Gradient for Rectangular Robust Markov Decision Processes (5 citations)
• Inverse Reinforcement Learning with the Average Reward Criterion (4 citations)
• Loss Dynamics of Temporal Difference Reinforcement Learning (4 citations)
• Weakly Coupled Deep Q-Networks (3 citations)
• Learning non-Markovian Decision-Making from State-only Sequences (2 citations)
• Model-Free Active Exploration in Reinforcement Learning (2 citations)
• ODE-based Recurrent Model-free Reinforcement Learning for POMDPs (2 citations)
• Optimal Treatment Allocation for Efficient Policy Evaluation in Sequential Decision Making (2 citations)
• Replicability in Reinforcement Learning (2 citations)

and 25 more in the interactive map.

Imitation learning · 19 papers

• Data Quality in Imitation Learning (7 citations)
• RoboCLIP: One Demonstration is Enough to Learn Robot Policies (7 citations)
• Large Language Models can Implement Policy Iteration (6 citations)
• Imitation Learning from Imperfection: Theoretical Justifications and Algorithms (3 citations)
• Learning Shared Safety Constraints from Multi-task Demonstrations (3 citations)
• Offline Imitation Learning with Variational Counterfactual Reasoning (3 citations)
• Residual Q-Learning: Offline and Online Policy Customization without Value (3 citations)
• CEIL: Generalized Contextual Imitation Learning (2 citations)
• Multi-Modal Inverse Constrained Reinforcement Learning from a Mixture of Demonstrations (1 citation)
• When Demonstrations meet Generative World Models: A Maximum Likelihood Framework for Offline Inverse Reinforcement Learning (1 citation)
• Action Inference by Maximising Evidence: Zero-Shot Imitation from Observation with World Models
• Coherent Soft Imitation Learning

and 7 more in the interactive map.

RL from feedback · 8 papers

• DPOK: Reinforcement Learning for Fine-tuning Text-to-Image Diffusion Models (39 citations)
• Inverse Preference Learning: Preference-based RL without a Reward Function (6 citations)
• Direct Preference-based Policy Optimization without Reward Modeling (5 citations)
• Sequential Preference Ranking for Efficient Reinforcement Learning from Human Feedback (5 citations)
• How to Fine-tune the Model: Unified Model Shift and Model Bias Policy Optimization (3 citations)
• Is RLHF More Difficult than Standard RL? A Theoretical Perspective (2 citations)
• Off-Policy Evaluation for Human Feedback (2 citations)
• Language Model Alignment with Elastic Reset (1 citation)