Tensor network approach is a powerful theoretical and computational tool for studying condensed matter systems. Simulating a two-dimensional quantum system using Projected Entangled-Pair States (PEPS) usually contains two major steps: (1) finding out the PEPS wave function using variational algorithms, and (2) evaluating the expectation values of physical observables using coarse-graining methods. In this talk, I will present a loop optimization scheme which can significantly improve the accuracy in both steps. The key idea is to deform a tensor network into small loops and then optimize tensors on each loop. Our scheme is computationally affordable even for a large bond degree of freedom. We benchmark our method on various models and compare to the results based on simple update.
Newton's Chair Postdoctoral fellow, Perimeter Institute for Theoretical Physics.
Main Research Directions:
1. Numerical simulation of quantum many-body systems
2. Topological states of matter
3. Interplay between condensed matter and quantum information
Time: 11:00 am, 23th March (Thur.), 2017
Location: Room 830, M Building of the Institue of Physics, CAS