Accepted Papers

Poster Session 1 (December 6 - 11:15 a.m.)

• Improving Value Estimation Critically Enhances Vanilla Policy Gradient
• Data-Dependent Regret Bounds for MABs with Constraints
• Policy Optimization in CMDPs with Bandit Feedback: Learning with Stochastic and Adversarial Constraints
• Test Time Risk Adaption with Mixture of Agents
• Optimal Regret Bounds for Policy Optimization in Contextual Bandits
• Thinking vs. Doing: Agents that Reason by Scaling Test-Time Interaction
• Large Language Model-Enhanced RL for Diverse and Novel Recommendations
• Convergence and Sample Complexity of First-Order Methods for Agnostic Reinforcement Learning
• Towards shutdownable agents via stochastic choice
• When Maximum Entropy Misleads Policy Optimization
• Regret Bounds for Adversarial Contextual Bandits with General Function Approximation and Delayed Feedback
• Unifying Agent Interaction and World Information for Multi-agent Coordination
• Efficiently Robust In-Context Reinforcement Learning with Adversarial Generalization and Adaptation
• What Makes a Reward Model a Good Teacher? An Optimization Perspective
• A Regularized Actor-Critic Algorithm for Bi-Level Reinforcement Learning
• Horizon Reduction Makes RL Scalable
• Policy Testing in Markov Decision Processes
• Scaling Offline RL via Efficient and Expressive Shortcut Models
• RL’s Razor: Why On-Policy Reinforcement Learning Forgets Less
• Real-World Reinforcement Learning of Active Perception Behaviors
• Linear Dynamics meets Linear MDPs: Closed-Form Optimal Policies via Reinforcement Learning
• Safe Guaranteed Dynamics Exploration with Probabilistic Models
• Idea: Sharpe Ratio-Optimized Thompson Sampling for Risk-Aware Online Learning
• Policy Compatible Skill Incremental Learning via Lazy Learning Interface
• Oracle-RLAIF: An Improved Fine-Tuning Framework for Multi-modal Video Models through Reinforcement Learning from Ranking Feedback
• Revisiting Actor-Critic Methods in Discrete Action Off-Policy Reinforcement Learning
• Speaking the Language of Teamwork: LLM-Guided Credit Assignment in Multi-Agent Reinforcement Learning
• Robust Constrained Offline Reinforcement Learning with Linear Function Approximation
• LLM-Driven Policy Diffusion: Enhancing Generalization in Offline Reinforcement Learning
• Hybrid Training for Enhanced Multi-task Generalization in Multi-agent Reinforcement Learning
• Structure Matters: Dynamic Policy Gradient
• SUSD: Structured Unsupervised Skill Discovery through State Factorization
• State Entropy Regularization for Robust Reinforcement Learning
• Shift Before You Learn: Enabling Low-Rank Representations in Reinforcement Learning
• Stackelberg Learning from Human Feedback: Preference Optimization as a Sequential Game
• How to Provably Improve Return Conditioned Supervised Learning?
• Accelerating Diffusion Planners in Offline RL via Reward-Aware Consistency Trajectory Distillation
• Exploration Implies Data Augmentation: Reachability and Generalisation in Contextual MDPs
• Learning a Pessimistic Reward in RLHF: KL Regularization is Not Necessary
• Beyond Marginals: Capturing Correlated Returns through Joint Distributional Reinforcement Learning
• Safe Exploration via Policy Priors
• Spectral Collapse Drives Loss of Plasticity in Deep Continual Learning
• Improved Training Mechanisms for Reinforcement Learning via Online Model Selection
• Ada-Diffuser: Latent-Aware Adaptive Diffusion for Decision-Making
• Improved Regret Bounds for Linear Bandits with Heavy-Tailed Rewards
• TARC: Time-Adaptive Robotic Control
• Quantile Reward Policy Optimization: Alignment with Pointwise Regression and Exact Partition Functions
• Policy Search via Bayesian Optimization with Temporal Difference Gaussian Processes
• Bigger, Regularized, Categorical: High-Capacity Value Functions are Efficient Multi-Task Learners
• Compute-Optimal Scaling for Value-Based Deep RL
• When Data is the Algorithm: A Systematic Study and Curation of Preference Optimization Datasets

Poster Session 2 (December 6 - 3:30 p.m.)

• A Reinforcement Learning Approach for Health-Behavioural Recommendations to Reduce Cancer Risk
• DHP: Discrete Hierarchical Planning for HRL Agents
• From Contextual Combinatorial Semi-Bandits to Bandit List Classification: Improved Sample Complexity with Sparse Rewards
• Efficient Adversarial Attacks on High-dimensional Offline Bandits
• Beyond RLHF and NLHF: Population-Proportional Alignment under an Axiomatic Framework
• Policy Gradient Guidance Enables Test Time Control
• Generalized Linear Bandits: Almost Optimal Regret with One-Pass Update
• Provably Efficient and Agile Randomized Q-Learning
• Idea: Bridging Theoretical Fairness Definitions with Multi-Agent Coordination in the Real World
• Human-Inspired Multi-Level Reinforcement Learning
• Safe, Trust Region Policy Optimization for Constrained Reinforcement Learning
• Long-Horizon Model-Based Offline Reinforcement Learning Without Conservatism
• Reward Model Overoptimisation in Iterated RLHF
• Demystifying the Mechanisms Behind Emergent Exploration in Goal-conditioned RL
• Uncertainty-Aware Policy-Preserving Abstractions with Abstention for One-Shot Decisions
• A Theoretical Analysis of Information Bottlenecks for Zero-Shot Transfer in Reinforcement Learning
• Robust Policy Gradient Optimization through Parameter Perturbation in Reinforcement Learning
• All Roads Lead to Likelihood: The Value of RL in Fine-Tuning
• Enhancing Diversity in Large Language Models via Determinantal Point Processes
• Efficient Restarts in Non-Stationary Model-Free Reinforcement Learning
• Bootstrap Ensemble Uncertainty for State-Adaptive Regularization in Offline Reinforcement Learning
• Behavior-Aware Off-Policy Selection in High-Stake Human-Centric Environments
• The Role of Preference Data and Unembeddings in the Convergence Rate of DPO
• Active Learning for Stochastic Contextual Linear Bandits
• Constrained Linear Thompson Sampling
• Towards Parameter-Free Temporal Difference Learning
• Generating Auxiliary Tasks with Reinforcement Learning
• Open Problem: Order Optimal Regret Bounds for Non-Markovian Rewards
• Exploring Time-Step Size in Reinforcement Learning for Sepsis Treatment
• Intent‑Based Reward Inference for Value‑Aligned Reinforcement Learning
• MOBODY: Model-Based Off-Dynamics Offline Reinforcement Learning
• Outcome-based Exploration for LLM Reasoning
• Idea: Fairness Constraints as Reliability Guarantees for RLHF Reward Models
• Principled Learning-to-Communicate in Cooperative MARL: An Information-Structure Perspective
• Steering Diffusion Policies with Value-Guided Denoising
• The Minimax Complexity of Preference-Based Decision Making in Multi-Objective Reinforcement Learning
• Sail into the Headwind: Alignment via Robust Rewards and Dynamic Labels against Reward Hacking
• Replicable Reinforcement Learning with Linear Function Approximation
• floq: Training Critics via Flow-Matching for Scaling Compute in Value-Based RL
• Optimistic Actor-Critic with Parametric Policies: Unifying Sample Efficiency and Practicality
• SpatialThinker: Reinforcing 3D Reasoning in Multimodal LLMs via Spatial Rewards
• Revisiting Mixture Policies in Entropy-Regularized Actor-Critic
• Bandit and Delayed Feedback in Online Structured Prediction
• Sharp Gap-Dependent Variance-Aware Regret Bounds for Tabular MDPs
• Fictive Learning Augments Model-Based Reinforcement Learning in the Two-Step Task
• On the relation of bisimulation, model irrelevance, and corresponding regret bounds
• Unsupervised Contrastive Goal Reaching
• Automatic Reward Shaping from Multi-Objective Human Heuristics
• Bandit Learning on Dynamic Graphs
• The Good, The Bad, and The Hybrid: A Reward Structure Showdown in Reasoning Models Training