Sebastiano Pilati (University of Camerino)

Accelerating Spin-Glass Simulations Using Quantum Annealers Through Deep Learning

Simulating the low-temperature equilibrium properties of spin glasses is a notoriously hard computational task. It plays a central role in condensed matter physics, and it is also related to relevant NP-hard optimization problems one faces in various field of science and engineering. 

In this talk, I will first discuss how generative neural networks can be used to accurately mimic Boltzmann distributions and to accelerate Monte Carlo simulations of classical statistical models. Next, I will discuss how D-WAVE quantum annealers can be used to produce adequate training datasets to optimize generative neural networks. Hybrid neural-Metropolis algorithms will be described, as well as the use of hybrid quantum-classical training datasets. We obtain a remarkable suppression of the long correlation times that plague spin-glass simulations in the low temperature regime. These results demonstrate that quantum devices combined with deep learning algorithms allow tackling otherwise intractable computational problems.