Clicking on a poster will zoom in on it; double clicking will zoom in further.
 |  |
Presenter: Evan Welty Assmus (University of Michigan) Collaborator(s): Qining Zhang, Lei Ying Title: SP3O: Reinforcement Learning from Segment Preferences without Reward Modeling | Presenter: Nirmit Joshi (Toyota Technological Institute at Chicago) Collaborator(s): Gene Li, Siddharth Bhandari, Shiva Kasiviswanathan, Cong Ma, Nathan Srebro Title: Learning to Answer Correct Demonstrations |
 |  |
Presenter: Kellen Kanarios (University of Michigan) Collaborator(s): Lei Ying Title: Motion Planning via Contrastive Reinforcement Learning and Monte-Carlo Tree Search | Presenter: Kihyun Kim (Massachusetts Institute of Technology (MIT)) Collaborator(s): Jiawei Zhang (University of Wisconsin-Madison), Asuman Ozdaglar (MIT), Pablo Parrilo (MIT) Title: A Linear Programming Framework for Offline Inverse Reinforcement Learning |
 |  |
Presenter: Chunyin Lei (University of California, Santa Barbara (UCSB)) Collaborator(s): Annie Qu Title: Risk-Aware Quantile Dynamic Treatment Regimes | Presenter: Pangpang Liu (Yale University) Collaborator(s): Junwei Lu (Harvard University), Will Wei Sun (Purdue University) Title: Uncertainty Quantification for Large Language Model Reward Learning under Heterogeneous Human Feedback |
 |  |
Presenter: Sunil Madhow (University of California, San Diego (UCSD)) Collaborator(s): Yuchen Liang, Yingbin Liang, Ness Shroff, Yu-Xiang Wang Title: Learnable Chernoff Baselines for Inference-Time Alignment | Presenter: Jongha Jon Ryu (Massachusetts Institute of Technology (MIT)) Collaborator(s): Jeongyeol Kwon (Meta), Benjamin Koppe (Cornell), Kwang-Sung Jun (POSTECH) Title: Improved Offline Contextual Bandits with Second-Order Bounds: Betting and Freezing |
 |  |
Presenter: Braham Snyder (University of Virginia) Collaborator(s): Haolin Liu, Chen-Yu Wei Title: On the Complexity of Offline Reinforcement Learning with Q^star-Approximation and Partial Coverage | Presenter: Deyi Wang (University of Michigan) Collaborator(s): Qining Zhang, Lei Ying Title: Efficient Federated RLHF via Zeroth-Order Policy Optimization |
 | |
Presenter: Qining Zhang (University of Michigan) Collaborator(s): Lei Ying Title: Provable Preference-based Reinforcement Learning with an Unknown Link Function: A Stochastic Zeroth-Order Framework |