DEEP-IN Seminar by Elia Cellini on December 18, 2024

The speaker Elia Cellini, PhD student at the University of Turin, introduced Markov Chain Monte Carlo methods and their non-equilibrium modification based on Jarzynski's equations.
The non-equilibrium transformations can be also implemented efficiently using normalizing flows, powerful neural networks that represent transformations between probability distributions.
Applying normalizing flows and non-equilibrium MCMC, the speaker showed how to approach difficult problems in lattice gauge theories, such as computing entanglement entropies.

Reported by Enrico Rinaldi