黑料网

In-person听触听Mary Ellen Jones 3116

Virtual听触听聽

TITLE: On exact energy guided diffusion model and diffusion-based offline reinforcement learning

Abstract: Guided generative models are pivotal in advancing the applications of generative modeling. In this talk, I will explore energy guidance in diffusion and flow matching models–a generalized formulation that extends beyond conventional diffusion models. By leveraging energy guidance, generative models are encouraged to produce samples with higher energy from the target data distribution. I will introduce energy-weighted diffusion model and flow matching model, with efficient implementation and offering new theoretical insights. In the second half of the presentation, I will discuss the extension of this approach to offline reinforcement learning through Q-weighted iterative policy optimization, which shows notable performance improvements across various offline RL tasks.

Short Bio:聽Weitong聽Zhang聽joined the School of Data Science and Society at the University of North Carolina at Chapel Hill as an assistant professor after completing his Ph.D. degree in computer science at the University of California, Los Angeles. His research focuses on developing robust and efficient reinforcement learning algorithms, emphasizing generative models and their applications in scientific discovery.