演讲人： Alexey Naumov 俄罗斯高等经济大学
时间： 2025-04-28 10:30-2025-04-28 12:00
地点：FIT 1-222
内容：

The recently proposed generative flow networks (GFlowNets) are a method of training a policy to sample compositional discrete objects with probabilities proportional to a given reward via a sequence of actions. GFlowNets exploit the sequential nature of the problem, drawing parallels with reinforcement learning (RL). Our work extends the connection between RL and GFlowNets to a general case. We demonstrate how the task of learning a generative flow network can be efficiently redefined as an entropy-regularized RL problem with a specific reward and regularizer structure. Furthermore, we illustrate the practical efficiency of this reformulation by applying standard soft RL algorithms to GFlowNet training across several probabilistic modeling tasks. Contrary to previously reported results, we show that entropic RL approaches can be competitive against established GFlowNet training methods. This perspective opens a direct path for integrating RL principles into the realm of generative flow networks.

The talk is based on the joint work with N. Morozov, D. Tiapkin and D. Vetrov. 

个人简介:

Alexey Naumov is a professor and scientific advisor at the International Laboratory of Stochastic Algorithms and High-Dimensional Inference at HSE University in Moscow. He graduated from Lomonosov Moscow State University (MSU) in 2010 and received a PhD in mathematics from Bielefeld University in 2013, along with a Candidate of Sciences degree in mathematics and physics from MSU the same year. His research focuses on high-dimensional probability and the mathematics of data science. He has authored over 40 papers in leading journals and conferences.