Speaker
Elia Cellini
(Università di Torino / INFN Torino)
Description
The explosive growth of machine learning provides a novel approach to study quantum field theo- ries on the lattice. One intriguing class of algorithms recently proposed are Normalizing Flows, deep generative models able to learn highy expressive transformations between distributions. In lattice field theory, Normalizing Flows can sample uncorrelated configurations from Boltzmann distributions; moreover, they can infer the exact partition function of the target. In this talk, we present preliminary numerical computations of the thickness of the Nambu-Goto string. We sample the configurations using Continuous Normalizing Flows and validate our results using numerical and analytical studies of the Nambu-Goto action as benchmarks.
Primary author
Elia Cellini
(Università di Torino / INFN Torino)
Co-authors
Michele Caselle
(Università di Torino / INFN)
Alessandro Nada
Marco Panero
(Università di Torino / INFN)
