{"id":"arxiv/2605.15899","url":"https://arxiv.org/abs/2605.15899","tags":["cond-mat.dis-nn","quant-ph"],"score":0,"title":"Solving Classical and Quantum Spin Glasses with Deep Boltzmann Quantum States","domain":"cond-mat.dis-nn","source":"arxiv","authors":["Luca Leone","Arka Dutta","Markus Heyl","Enrico Prati","Pietro Torta"],"reviews":0,"subject":"cond-mat.dis-nn","summary":"Variational neural network models have achieved remarkable success in solving ground-state problems of quantum many-body systems. However, addressing classical and quantum spin glasses remains challenging, as disorder and energy frustration give rise to an exponentially large number of local energy minima separated by high-energy barriers, hindering the efficiency of conventional Metropolis-based Monte Carlo methods. To bridge this gap, we introduce Deep Boltzmann Quantum States, a class of neural quantum states inspired by deep Boltzmann machines that inherit efficient block Gibbs sampling.…","category":"research","comments":0,"disputed":false,"division":"physics","mentions":0,"discussed":0,"published":"2026-05-15","reactions":{}}