Machine learning the Ising transition: A comparison between discriminative and generative approaches

Difei Zhang,Frank Schäfer,Julian Arnold
2024-11-29
Abstract:The detection of phase transitions is a central task in many-body physics. To automate this process, the task can be phrased as a classification problem. Classification problems can be approached in two fundamentally distinct ways: through either a discriminative or a generative method. In general, it is unclear which of these two approaches is most suitable for a given problem. The choice is expected to depend on factors such as the availability of system knowledge, dataset size, desired accuracy, computational resources, and other considerations. In this work, we answer the question of how one should approach the solution of phase-classification problems by performing a numerical case study on the thermal phase transition in the classical two-dimensional square-lattice ferromagnetic Ising model.
Disordered Systems and Neural Networks,Statistical Mechanics,Machine Learning
What problem does this paper attempt to address?