On the generalizability of artificial neural networks in spin models

Abstract

The applicability of artificial neural networks (ANNs) is typically limited to the models they are trained with and little is known about their generalizability, which is a pressing issue in the practical application of trained ANNs to unseen problems. Here, by using the task of identifying phase transitions in spin models, we establish a systematic generalizability such that simple ANNs trained with the two-dimensional ferromagnetic Ising model can be applied to the ferromagnetic $q$-state Potts model in different dimensions for $q\ge 2$. The same scheme can be applied to the highly nontrivial antiferromagnetic $q$-state Potts model. We demonstrate that similar results can be obtained by reducing the exponentially large state space spanned by the training data to one that comprises only three representative configurations artificially constructed through symmetry considerations. We expect our findings to simplify and accelerate the development of machine learning-assisted tasks in spin-model related disciplines in physics and materials science.

31 pages, 25 figures, 3 tables; v2 - published version

References in corpus (7)

• Learning phase transitions by confusion
• Mapping distinct phase transitions to a neural network
• Phase Transition Properties of 3D Potts Models
• Exploring neural network training strategies to determine phase transitions in frustrated magnetic models
• Machine-learning physics from unphysics: Finding deconfinement temperature in lattice Yang-Mills theories from outside the scaling window
• autoScale.py - A program for automatic finite-size scaling analyses: A user's guide
• Normalized entropy density of the 3D 3-state Potts model

Cited by in corpus (5)

• Mapping distinct phase transitions to a neural network
• Machine-learning physics from unphysics: Finding deconfinement temperature in lattice Yang-Mills theories from outside the scaling window
• Study of phase transition of Potts model with Domain Adversarial Neural Network
• Neural Network Statistical Mechanics
• Discovering quasiorder parameters in the Potts model: A bridge between machine learning and critical phenomena