Machine Learning of Quantum and Topological Physics
Student No.:80
Time:Fri 16:30-17:30, Jun.8
Instructor:翟荟 Zhai Hui  
Place:Lecture Hall, 3rd floor of Jin Chun Yuan West Bldg.
Starting Date:2018-6-8
Ending Date:2018-6-8


The motivation of this work is to apply the machine learning method to well known physics problems, and try to understand how it works and what exact the neural network learns when dealing with these problems. In this talk I will discuss two examples. In the first example, I will discuss using neural network to classify topological phases. I will discuss that the neural network indeed finds out the right formula for calculating topological invariant after training. In the second example, I will discuss using neural network to solve the quantum mechanical scattering problems. After training the neural network, we find that the neural network automatically develops systematical perturbation theory.


Zhai Hui obtains his B.S. in physics department of Tsinghua University in 2002, and PhD in Institute for Advanced Study in 2005. After postdoc in Ohio-State University and UC Berkeley, he joins Institute for Advanced Study in 2009, and becomes tenured member in 2012 and full Professor in 2015. His research focuses on cold atom and condensed matter theory, application of machine learning method in physics and holographic principle. He has been awarded National Science Fund for Distinguished Young Scholar and Changjiang Professorship.