Each accepted paper will be presented in exactly one of the poster sessions listed below. Please find your poster session number in this list: Poster Session Link.
Poster format: 36 in (H) × 24 in (W) / 91 cm (H) × 61 cm (W). Note this is different from the main conference poster size.
Talk lengths: Oral talks are 10 minutes each and spotlight talks are 3 minutes each.
8:00 - 8:15
Opening Remarks
8:15 - 8:45
Invited Talk: Sergey Levine
8:45 - 9:15
Invited Talk: Aarti Singh
9:15 - 9:30
Spotlights (5 × 3 min)
- Soft Forward-Backward Representations for Zero-shot Reinforcement Learning with General Utilities
- From Offline Trajectories to Online Adaptation: A Multimodal JEPA Pretraining Study on Pokemon Red
- Belief-Aware Decision Transformers for Offline-to-Online Decision-Making under Partial Observability: A Geosteering Case Study
- XQCfD: Accelerating Fast Actor-Critic Algorithms with Prior Data and Prior Policies
- Position: Offline-Dataset Evaluation for Online Decision-Making Needs an Identification Standard
9:30 - 10:30
Poster Session 1
10:30 - 11:00
Invited Talk: Jacob Gardner
11:00 - 12:00
Oral Talks (6 × 10 min)
- Pitfalls and Remedies for Multi-Task Bayesian Optimization
- Can K Heads Explore Better Than One in Online Reinforcement Learning?
- Practical Bayesian Optimization for Scientific Discovery
- Rethinking Bayesian Optimization for Co-Optimizing LLM Training Configurations
- Leveraging Instruction Tuning and Merging for Reasoning Model Adaptation
- Cost-Aware Learning
12:00 - 1:00
Lunch (on your own)
1:00 - 1:30
Invited Talk: Eytan Bakshy
1:30 - 2:00
Oral Talks (3 × 10 min)
- The Three Regimes of Offline-to-Online Reinforcement Learning
- Molten Pot: Evaluations & Datasets for Social Offline Reinforcement Learning
- Hidden Failure Modes in Latent World-Model Planning from Offline Data
2:00 - 2:30
Spotlights (10 × 3 min)
- Efficient Cost-Aware LLM Evaluation via Bayesian Bandit Gittins Indices
- Can We Really Learn One Representation to Optimize All Rewards?
- Q-Flow: Stable and Expressive Reinforcement Learning with Flow-Based Policy
- Uncertainty-Guided Reward Labeling for Reinforcement Learning under Limited Feedback
- Boosting for Reinforcement Learning in Structured MDPs
- Less Tuning, Better Planning: Simplifying Offline Model-Based Planning
- Abstraction for Offline Goal-Conditioned Reinforcement Learning
- Meta-GC-TTT: Training Offline Goal-Conditioned Policies for Test-Time Adaptation
- Fast Rates for Offline Contextual Bandits with Forward-KL Regularization under Single-Policy Concentrability
- Efficient Off-Policy RL for Video Generation via Forward-Consistent Reward Matching
2:30 - 3:30
Poster Session 2
3:30 - 4:00
Invited Talk: Clara Wong-Fannjiang
4:00 - 4:30
Invited Talk: Wen Sun
4:30 - 5:00
Short Panel and Community Discussion