Abstract:Monte Carlo methods are essential across diverse scientific fields, yet their efficiency is frequently hampered by critical slowing down-a sharp increase in autocorrelation times near phase transitions. Although deep learning approaches, such as neural-network-based samplers, have been proposed to alleviate this issue, they face another serious problem: the difficulty of training the models. This difficulty partially stems from the overly general nature of original machine-learning architectures, which often ignore underlying physical symmetries and force networks to relearn them from scratch. In this paper, we demonstrate that incorporating physical priors into the model significantly enhances performance. Building upon existing strategies that integrate spin-spin interactions, we propose a framework that utilizes a prior probability distribution as a starting point for training. Our results for the Ising model, as well as for the Edwards-Anderson spin glass model, suggest that moving away from `blank slate' models in favor of physics-informed priors reduces the training burden and facilitates the simulation of larger system sizes in discrete spin models.
| Comments: | 28 pages, 11 figures |
| Subjects: | Machine Learning (cs.LG); Disordered Systems and Neural Networks (cond-mat.dis-nn); High Energy Physics - Lattice (hep-lat) |
| ACM classes: | J.2; I.2.6 |
| Cite as: | arXiv:2605.16020 [cs.LG] |
| (or arXiv:2605.16020v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2605.16020 arXiv-issued DOI via DataCite (pending registration) |
From: Piotr Bialas [view email]
[v1]
Fri, 15 May 2026 14:54:30 UTC (2,548 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。