Abstract:Symmetry is present throughout nature and continues to play an increasingly central role in physics and machine learning. Fundamental symmetries, such as Poincaré invariance, allow physical laws discovered in laboratories on Earth to be extrapolated to the farthest reaches of the universe. Symmetry is essential to achieving this extrapolatory power in machine learning applications. For example, translation invariance in image classification allows models with fewer parameters, such as convolutional neural networks, to be trained on smaller data sets and achieve state-of-the-art performance. In this paper, we provide a unifying theoretical and methodological framework for incorporating symmetry into machine learning models in three ways: 1. enforcing known symmetry when training a model; 2. discovering unknown symmetries of a given model or data set; and 3. promoting symmetry during training by learning a model that breaks symmetries within a user-specified group of candidates when there is sufficient evidence in the data. We show that these tasks can be cast within a common mathematical framework whose central object is the Lie derivative associated with fiber-linear Lie group actions on vector bundles. We extend and unify several existing results by showing that enforcing and discovering symmetry are linear-algebraic tasks that are dual with respect to the bilinear structure of the Lie derivative. We also propose a novel way to promote symmetry by introducing a class of convex regularization functions based on the Lie derivative and nuclear norm relaxation to penalize symmetry breaking during training of machine learning models. We explain how these ideas can be applied to a wide range of machine learning models including basis function regression, dynamical systems discovery, neural networks, and neural operators acting on fields.

Machine Learning Symmetry Discovery for Classical Mechanics

Identifying the Group-Theoretic Structure of Machine-Learned Symmetries

Deep Learning Symmetries and Their Lie Groups, Algebras, and Subalgebras from First Principles

Finding discrete symmetry groups via Machine Learning

Machine learning Lie structures & applications to physics

A Unified Framework to Enforce, Discover, and Promote Symmetry in Machine Learning

Hamiltonian Neural Networks with Automatic Symmetry Detection

Accelerated Discovery of Machine-Learned Symmetries: Deriving the Exceptional Lie Groups G2, F4 and E6

Robust Symmetry Detection via Riemannian Langevin Dynamics

Probing the effects of broken symmetries in machine learning

Learning Lie Group Symmetry Transformations with Neural Networks

Machine Learning Phase Space Quantum Dynamics Approaches

Symmetry Preservation in Hamiltonian Systems: Simulation and Learning

Learning Infinitesimal Generators of Continuous Symmetries from Data

Discovering Nonlinear Dynamics Through Scientific Machine Learning

Lie symmetry analysis, exact solutions, and conservation laws to multi-component nonlinear Schrödinger equations

Unsupervised machine learning of quenched gauge symmetries: A proof-of-concept demonstration

Adding machine learning within Hamiltonians: Renormalization group transformations, symmetry breaking and restoration

Symmetry Discovery Beyond Affine Transformations

Symmetry and Orbit Detection Via Lie-Algebra Voting.