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.

Symmetries of systems of first order ODEs: Symbolic symmetry computations, mechanistic model construction and applications in biology

Symmetry-Informed Governing Equation Discovery

Energy translation symmetries and dynamics of separable autonomous two-dimensional ODEs

Finite symmetric systems and their analysis

Dn Symmetric Hamiltonian System: A Network of Coupled Gyroscopes as a Case Study

From Physics to Biology by Extending Criticality and Symmetry Breakings

Structural and Symmetry Analysis of Discrete Dynamical Systems

Symmetry operators and generation of symmetry transformations of partial differential equations

On discrete symmetries of robotics systems: A group-theoretic and data-driven analysis

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

Symmetry in Signals: A New Insight

Morphological Symmetries in Robotics

Modeling and Analyzing Biological Oscillations in Molecular Networks

Symmetry and Numerical Solutions for Systems of Non-linear Reaction Diffusion Equations

Symmetry in models of natural selection

Symmetry within and between solutions

Symmetries of nonlinear ordinary differential equations: the modified Emden equation as a case study

On Exact and Approximate Symmetries of Algebraic and Ordinary Differential Equations with a Small Parameter

Symmetry-preserving numerical schemes

Involution symmetry quantification using recurrences

Symmetry properties and explicit solutions of the generalized Weierstrass system