Tensor Network Methods and Large Deviations

Speaker: Dr. Juan Garrahan, University of Nottingham

Host: Dr. Pratyush Tiwary, UMCP

Abstract: I will describe the application of tensor networks for the study of dynamical fluctuations in systems with stochastic dynamics. I will discuss several related questions: (i) how to obtain the long-time statistics, i.e. the large deviations, of trajectory observables from tilted Markov generators (in analogy with finding quantum ground states); (ii) how to exploit these methods to efficiently sample rare events; (iii) how to extend this approach to finite time trajectories (in analogy with calculating quantum partition sums); (iv) the emergence of symmetry protected topological phases in atypical dynamics.

Statistical Physics Seminar