Seminar : Solving Statistical Mechanics: From Mean Field to Neural Networks, then to Tensor Networks

Speaker

Prof. Pan Zhang - Institute of Theoretical Physics, Chinese Academy of Sciences

Title

Solving Statistical Mechanics: From Mean Field to Neural Networks, then to Tensor Networks

Abstract

Computing free energy, estimating physical quantities, and generating uncorrelated samples are fundamental problems in statistical mechanics. In this talk I will first overview the classic mean-field methods and message passing algorithms from the variational point of view, then introduce the variational autoregressive networks (VAN), which extends the mean-field methods using neural networks. Finally I will introduce a tensor network based method for stat. mech. on arbitrary graphs, and extend the method to quantum circuit simulations.

Join Zoom Meeting