AI estimation of partition functions with the PULSE method for the thermodynamic properties of disordered compounds
Abstract
Predicting the thermodynamic properties of chemically disordered materials is constrained by the combinatorial growth of possible atomic configurations, which makes conventional sampling approaches computationally costly. In this work, we extend the PULSE (Partition function Unsupervised Learning Sampling and Evaluation) approach that enables the estimation of observable-dependent partition functions, thereby extending its applicability from local atomic quantities to general thermodynamic observables. This is based on a generative machine-learning approach in which a single model both generates statistically representative atomic configurations and provides explicit partition function estimates. Our primary goal is to assess it on the two-dimensional Ising model, an exactly solvable benchmark system sharing the same statistical-mechanical formulation as chemically disordered binary materials, and compare its performance against exact brute-force calculations as well as Wolff and Wang-Landau Monte Carlo algorithms. PULSE reproduces the relevant partition functions accurately using a limited number of configurations, at a sampling cost that does not grow with system size. These key results demonstrate the relevance of our method to enable the study of disordered materials at scales that are difficult to achieve using conventional approaches, and with limited computational effort.
Related
- Improved Adaptive Estimation of Quantum Partition Functions with Heisenberg Scaling
- Spin–orbit–controlled flat-band transition, thermodynamic and magnetic response in a disordered Lieb-5 lattice
- Finding CFTs from spatial positivity of partition functions
- Trace Relations and Partition Functions
- From Z to all: cyclic sum rules for thermal CFT partition functions
BibTeX
@misc{bernard2026estimation,
title = {{AI estimation of partition functions with the PULSE method for the thermodynamic properties of disordered compounds}},
author = {Baptiste Bernard and Luca Messina and Eiji Kawasaki and Emeric Bourasseau},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2605.28594},
eprint = {2605.28594},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2605.28594},
}
RIS
TY - UNPB
TI - AI estimation of partition functions with the PULSE method for the thermodynamic properties of disordered compounds
AU - Baptiste Bernard
AU - Luca Messina
AU - Eiji Kawasaki
AU - Emeric Bourasseau
PY - 2026
DA - 2026/10/19
PB - arXiv
DO - 10.48550/arxiv.2605.28594
UR - https://arxiv.org/abs/2605.28594
AB - Predicting the thermodynamic properties of chemically disordered materials is constrained by the combinatorial growth of possible atomic configurations, which makes conventional sampling approaches computationally costly. In this work, we extend the PULSE (Partition function Unsupervised Learning Sampling and Evaluation) approach that enables the estimation of observable-dependent partition functions, thereby extending its applicability from local atomic quantities to general thermodynamic observables. This is based on a generative machine-learning approach in which a single model both generates statistically representative atomic configurations and provides explicit partition function estimates. Our primary goal is to assess it on the two-dimensional Ising model, an exactly solvable benchmark system sharing the same statistical-mechanical formulation as chemically disordered binary materials, and compare its performance against exact brute-force calculations as well as Wolff and Wang-Landau Monte Carlo algorithms. PULSE reproduces the relevant partition functions accurately using a limited number of configurations, at a sampling cost that does not grow with system size. These key results demonstrate the relevance of our method to enable the study of disordered materials at scales that are difficult to achieve using conventional approaches, and with limited computational effort.
ER -