AAAI 2025 accepted papers on Reinforcement learning

Visual and representation RL · 50 papers

• Deep Reinforcement Learning for Robotics: A Survey of Real-World Successes (172 citations)
• FlowPolicy: Enabling Fast and Robust 3D Flow-Based Policy via Consistency Flow Matching for Robot Manipulation (20 citations)
• Pareto Set Learning for Multi-Objective Reinforcement Learning (14 citations)
• GenAL: Generative Agent for Adaptive Learning (9 citations)
• Bootstrapped Reward Shaping (5 citations)
• APIRL: Deep Reinforcement Learning for REST API Fuzzing (4 citations)
• ConfigX: Modular Configuration for Evolutionary Algorithms via Multitask Reinforcement Learning (4 citations)
• Learning Personalized Decision Support Policies (4 citations)
• SkillTree: Explainable Skill-Based Deep Reinforcement Learning for Long-Horizon Control Tasks (4 citations)
• A Deployed Online Reinforcement Learning Algorithm in an Oral Health Clinical Trial (3 citations)
• Autonomous Option Invention for Continual Hierarchical Reinforcement Learning and Planning (3 citations)
• Enhancing Question Generation through Diversity-Seeking Reinforcement Learning with Bilevel Policy Decomposition (3 citations)

and 38 more in the interactive map.

Policy gradient and actor-critic · 16 papers

• CTD4 – a Deep Continuous Distributional Actor-Critic Agent with a Kalman Fusion of Multiple Critics (3 citations)
• Neural Control and Certificate Repair via Runtime Monitoring (3 citations)
• Cross-Validated Off-Policy Evaluation (2 citations)
• Decoupled Policy Actor-Critic: Bridging Pessimism and Risk Awareness in Reinforcement Learning (1 citation)
• Optimizing Vital Sign Monitoring in Resource-Constrained Maternal Care: An RL-Based Restless Bandit Approach (1 citation)
• Single-Loop Federated Actor-Critic across Heterogeneous Environments (1 citation)
• The Distributional Reward Critic Framework for Reinforcement Learning Under Perturbed Rewards (1 citation)
• Truncated Gaussian Policy for Debiased Continuous Control (1 citation)
• Augmented Lagrangian Risk-constrained Reinforcement Learning for Portfolio Optimization (Student Abstract)
• Distributionally Robust Policy Evaluation and Learning for Continuous Treatment with Observational Data
• Efficient Multi-Policy Evaluation for Reinforcement Learning
• Leveraging Constraint Violation Signals for Action Constrained Reinforcement Learning

and 4 more in the interactive map.

Offline RL · 13 papers

• Constraint-Adaptive Policy Switching for Offline Safe Reinforcement Learning (7 citations)
• Are Expressive Models Truly Necessary for Offline RL? (4 citations)
• SUMO: Search-Based Uncertainty Estimation for Model-Based Offline Reinforcement Learning (4 citations)
• Active Reinforcement Learning Strategies for Offline Policy Improvement (1 citation)
• Distribution-Free Uncertainty Quantification in Mechanical Ventilation Treatment: A Conformal Deep Q-Learning Framework (1 citation)
• Entropy Regularized Task Representation Learning for Offline Meta-Reinforcement Learning (1 citation)
• In-Dataset Trajectory Return Regularization for Offline Preference-based Reinforcement Learning (1 citation)
• Offline Safe Reinforcement Learning Using Trajectory Classification (1 citation)
• Behaviour Preference Regression for Offline Reinforcement Learning
• Dynamic Uncertainty Estimation for Offline Reinforcement Learning
• Improving Generalization in Offline Reinforcement Learning via Latent Distribution Representation Learning
• MGDA: Model-based Goal Data Augmentation for Offline Goal-conditioned Weighted Supervised Learning

and 1 more in the interactive map.

MDPs and planning · 13 papers

• Learning Logic Specifications for Policy Guidance in POMDPs: an Inductive Logic Programming Approach (5 citations)
• Probabilistic Shielding for Safe Reinforcement Learning (5 citations)
• Real-Time Recurrent Reinforcement Learning (4 citations)
• Deterministic Policy Gradient Primal-Dual Methods for Continuous-Space Constrained MDPs (2 citations)
• Formally Verified Approximate Policy Iteration (1 citation)
• GRSN: Gated Recurrent Spiking Neurons for POMDPs and MARL (1 citation)
• Semi-Markovian Planning to Coordinate Aerial and Maritime Medical Evacuation Platforms (1 citation)
• AoI-MDP: An AoI Optimized Markov Decision Process Dedicated in the Underwater Task (Student Abstract)
• Approximate Bilevel Difference Convex Programming for Bayesian Risk Markov Decision Processes
• On Shallow Planning Under Partial Observability
• Risk-averse Total-reward MDPs with ERM and EVaR
• Scalable Solutions for Decision-Making Systems Using Explainable Policy Representations

and 1 more in the interactive map.

RL from feedback · 7 papers

• Radiology Report Generation via Multi-objective Preference Optimization (9 citations)
• ERFSL: An Efficient Reward Function Searcher via Large Language Models for Custom-Environment Multi-Objective Reinforcement Learning (Student Abstract) (2 citations)
• Forward KL Regularized Preference Optimization for Aligning Diffusion Policies (2 citations)
• Universal Post-Processing Networks for Joint Optimization of Modules in Task-Oriented Dialogue Systems (2 citations)
• Alleviating Shifted Distribution in Human Preference Alignment through Meta-Learning (1 citation)
• DUO: Diverse, Uncertain, On-Policy Query Generation and Selection for Reinforcement Learning from Human Feedback (1 citation)
• Efficient and Robust Reinforcement Learning from Human Feedback (1 citation)

Imitation learning · 7 papers

• Inverse Reinforcement Learning by Estimating Expertise of Demonstrators (3 citations)
• SORREL: Suboptimal-Demonstration-Guided Reinforcement Learning for Learning to Branch (3 citations)
• POI Recommendation via Multi-Objective Adversarial Imitation Learning (2 citations)
• Partial Identifiability in Inverse Reinforcement Learning for Agents with Non-Exponential Discounting (1 citation)
• Enhancing Online Reinforcement Learning with Meta-Learned Objective from Offline Data
• Imitation Learning Backoff: Reinforcement Learning-based Channel Access for Guaranteeing Fairness (Student Abstract)
• Markov Balance Satisfaction Improves Performance in Strictly Batch Offline Imitation Learning