UVA RL Meetup

This weekly meetup is organized by Shangtong Zhang and Chen-Yu Wei for UVA RL folks to share interesting RL papers.

Fall 2025

Date	Presenter	Paper or Topic
Sep 5	Jiuqi Wang	Convergence of Regularized Agent-State-Based Q-learning in POMDPs
Sep 12 (ICLR)
Sep 19 (ICLR)
Sep 26 (ICLR)
Oct 3	Zixuan Xie	Transformers Learn to Implement Multi-step Gradient Descent with Chain of Thought
Oct 10	Xinyu Liu	A Dynamic View of Some Anomalous Phenomena in SGD
Oct 17	Amir Moeini	On the Design of KL-Regularized Policy Gradient Algorithms for LLM Reasoning
Oct 24	Haolin Liu	Token Hidden Reward: Steering Exploration-Exploitation in Group Relative Deep Reinforcement Learning
Oct 31	Braham Snyder	The Importance of Pessimism in Fixed-Dataset Policy Optimization
Nov 7	Minjae Kwon	Constraint Policy Optimization
Nov 14	Jiuqi Wang	PokeeResearch: Effective Deep Research via Reinforcement Learning from AI Feedback and Robust Reasoning Scaffold
Nov 21	Andrew Wagenmaker	Towards Practical Online Improvement of Pretrained Policies for Robotic Manipulation
Nov 28 (Thanksgiving)

Spring 2025

Date	Presenter	Paper or Topic
Feb 7	Zixuan Xie	Analytic-DPM: An Analytic Estimate of the Optimal Reverse Variance in Diffusion Probabilistic Models Tutorial on Diffusion Models for Imaging and Vision
Feb 14	Xinyu Liu	Decoupled Functional Central Limit Theorems for Two-Time-Scale Stochastic Approximation
Feb 21	Amir Moeini	Transformers Implement Functional Gradient Descent to Learn Non-Linear Functions In Context
Feb 28 (AAAI)
Mar 7	Braham Snyder	Target Networks and Over-parameterization Stabilize Off-policy Bootstrapping with Function Approximation
Mar 14 (Spring Break)
Mar 21	Haolin Liu	Exploratory Preference Optimization: Harnessing Implicit Q*-Approximation for Sample-Efficient RLHF
Mar 28	Amin Davoodabadi (cancelled)	An Information-Theoretic Perspective on Intrinsic Motivation in Reinforcement Learning
Apr 4	Jiuqi Wang	Can Looped Transformers Learn to Implement Multi-step Gradient Descent for In-context Learning?
Apr 11	Dylan Foster (remote)	Revisiting the Foundations of Imitation Learning
Apr 18	Xiangyu Liu (remote)	Provable Partially Observable Reinforcement Learning with Privileged Information
Apr 25 (ICLR)

Fall 2024

Date	Presenter	Paper
Sep 27	Shangtong Zhang	The O.D.E. Method for Convergence of Stochastic Approximation and Reinforcement Learning
Oct 11	Chen-Yu Wei	Equivalence Between Policy Gradients and Soft Q-Learning
Oct 25	Shuze Liu	Direct Preference Optimization: Your Language Model is Secretly a Reward Model
Nov 8	Ethan Blaser	Transformers as Decision Makers: Provable In-Context Reinforcement Learning via Supervised Pretraining
Nov 22	Haolin Liu	Correcting the Mythos of KL-Regularization: Direct Alignment without Overoptimization via Χ²-Preference Optimization
Dec 6	Jiuqi Wang	Diffusion Policies as an Expressive Policy Class for Offline Reinforcement Learning Denoising Diffusion Probabilistic Models Understanding Diffusion Models: A Unified Perspective