Quantum Matter SG seminar by Prof. Daw-Wei Wang on August 23, 2021
On August 23rd, 2021, Quantum Matter Study Group invited Prof. Daw-Wei Wang to give a talk on the application of machine learning in condensed matter physics. First, he briefly introduced learning and emphasized using supervised machine learning in condensed matter physics. The practical condensed matter problem he focused on is the physics of many-body systems since the many-body problem is challenging to solve. By randomly choosing small portions of a many-body Hamiltonian, machine learning can almost accurately predict the energy of the many-body system. Prof. Wang used the 1D Fermi-Hubbard model and the 1D Ising model to show the consistency between the learning prediction and the known solutions. The limitation of this approach is that the system size has to be fixed. To resolve this problem, he used the transfer learning approach to extend the prediction to a larger system size by learning from small systems. In the end, he talked about identifying the topological phase transition points by improving the machine learning approach in the literature. The talk is very comprehensive and informative. We thank Prof. Wang for giving a wonderful talk.
Reported by Ching-Kai Chiu