2020/01/02 by Gavin S. Hartnett, Masoud Mohseni, Hartnett, Gavin S. +1 · 6 citations
Computer Science · Mathematics · Physics and Astronomy · #Artificial intelligence #Complex Network Analysis Techniques #Computer science #Disordered Systems and Neural Networks (cond-mat.dis-nn) #FOS: Computer and information sciences #FOS: Physical sciences #Generative grammar #Generative model #Granularity #Inference #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Metastability #Neural Networks and Applications #Physics #Quantum Physics (quant-ph) #Quantum mechanics #Renormalization group #Spin (aerodynamics) #Spin glass #Statistical physics #Theoretical and Computational Physics #Theoretical computer science #cond-mat.dis-nn #cs.LG #quant-ph #stat.ML
paper · pdf · doi:10.48550/arxiv.2001.00585
published in arXiv (Cornell University) (Cornell University) · 16 pages, 7 figures
openalex publication_date 2020/01/02 · arxiv created 2020/01/10 · arxiv updated 2020/01/14 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Spin-glasses are universal models that can capture complex behavior of many-body systems at the interface of statistical physics and computer science including discrete optimization, inference in graphical models, and automated reasoning. Computing the underlying structure and dynamics of such complex systems is extremely difficult due to the combinatorial explosion of their state space. Here, we develop deep generative continuous spin-glass distributions with normalizing flows to model correlations in generic discrete problems. We use a self-supervised learning paradigm by automatically generating the data from the spin-glass itself. We demonstrate that key physical and computational properties of the spin-glass phase can be successfully learned, including multi-modal steady-state distributions and topological structures among metastable states. Remarkably, we observe that the learning itself corresponds to a spin-glass phase transition within the layers of the trained normalizing flows. The inverse normalizing flows learns to perform reversible multi-scale coarse-graining operations which are very different from the typical irreversible renormalization group techniques.