Finding Symmetry Breaking Order Parameters with Euclidean Neural Networks

Abstract

Curie's principle states that "when effects show certain asymmetry, this asymmetry must be found in the causes that gave rise to them". We demonstrate that symmetry equivariant neural networks uphold Curie's principle and can be used to articulate many symmetry-relevant scientific questions into simple optimization problems. We prove these properties mathematically and demonstrate them numerically by training a Euclidean symmetry equivariant neural network to learn symmetry-breaking input to deform a square into a rectangle and to generate octahedra tilting patterns in perovskites.

arXiv comment6 pages, 3 figures

References in corpus (4)

• On representing chemical environments
• Machine learning and the physical sciences
• SchNet - a deep learning architecture for molecules and materials
• Symmetry-Adapted Machine-Learning for Tensorial Properties of Atomistic Systems

Cited by in corpus (9)

• Self-Consistent Determination of Long-Range Electrostatics in Neural Network Potentials
• Machine Learning on Neutron and X-Ray Scattering
• Applications and Techniques for Fast Machine Learning in Science
• Direct prediction of phonon density of states with Euclidean neural networks
• MLGWSC-1: The first Machine Learning Gravitational-Wave Search Mock Data Challenge
• Reinforcement-learning-based control of convectively-unstable flows
• Towards out of distribution generalization for problems in mechanics
• Data-driven discovery and extrapolation of parameterized pattern-forming dynamics
• Artificial Intelligence for Science in Quantum, Atomistic, and Continuum Systems