Artificial intelligence and machine learning for phases and transitions in the Ising model: an overview

VK Babu and R Pandit, EUROPEAN PHYSICAL JOURNAL B, 98, 258 (2025).

DOI: 10.1140/epjb/s10051-025-01105-y

The Ising spin model has long been the foundational model for studying phase transitions in theoretical statistical physics. In the wake of the explosion of machine-learning (ML) techniques in recent years, the Ising model has been used for ML applications to phase transitions. In this overview, we discuss the applications of ML, via both supervised and unsupervised learning, to the study of phases and transitions in the Ising model. We also discuss transfer learning and provide some examples of its use for neural networks trained on Ising model spin configurations. We conclude with a summary and some future directions.

Return to Publications page