LASSP Special Seminar - Vadim Oganesyan - CUNY

Hosted by: Eun-Ah Kim

Title: Statistical mechanics with tensor renormalization group

Abstract: Tensor renormalization group (TRG) is an algorithm introduced by Levin and Nave to implement a Migdal-Kadanoff-like iterative coarse-graining scheme for lattice partition sums. It proved remarkably accurate for the 2D Ising models and stimulated considerable further work to understand and improve the technical aspects of the algorithm leading to qualitative improvements in the method. This talk will cover three types of new results -- (i) discovery of novel phases with quasi-long range modulated space or time (not both) order in 2D Ising models at complex coupling (mappable to non-unitary quantum circuits);  (ii) detailed study of universal content of three critical points, including large-separation correlation functions (up to 2^20 sites!) that seems to indicate an extreme paucity of finite bond dimension effects; (iii) some ideas and preliminary results on combining monte carlo and TRG