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.

Teaching "Symmetry" in the Introductory Physics Curriculum

Symmetries and the fundamental forces of Nature

Symmetries and the fundamental forces of Nature II

Music of the Spheres: Teaching Quantum Field Theory at the Introductory Level

Symmetry, Broken Symmetry, and Topology in Modern Physics

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

The Symmetry Principle in Condensed Matter Physics (I)

Introduction to Mechanics and Symmetry: A Basic Exposition of Classical Mechanical Systems

Teaching Physics in Combination with Natural Phenomena in Daily Life

The role physics can play in a multi-disciplinary curriculum for non-physics scientists and engineers

A Positive Role of Symmetry Analysis in Electromagnetism Course

An equitable and effective approach to introductory mechanics

In Search of Symmetry Lost

From Physics to Biology by Extending Criticality and Symmetry Breakings

Are we teaching students to think like scientists?

Teaching Physics with Music

Introductory concept for teaching chirality - symmetry of the asymmetric

Quantum and classical symmetries

Introducing interdisciplinary science to second year undergraduates in a Current Topics in Biophysics course

Reinventing college physics for biologists: Explicating an epistemological curriculum

A new introductory quantum mechanics curriculum