Physics-Augmented Learning: A New Paradigm Beyond Physics-Informed Learning

TL;DR

This paper introduces physics-augmented learning (PAL), a new framework that extends physics-informed learning (PIL) by enhancing generative capabilities and improving performance in cases where PIL is limited.

Contribution

The paper generalizes PIL into PAL, integrating physical inductive biases to better handle both discriminative and generative tasks in machine learning.

Findings

PAL performs well where PIL is ineffective

PAL enhances model generalizability with physical biases

PAL complements PIL by addressing generative properties

Abstract

Integrating physical inductive biases into machine learning can improve model generalizability. We generalize the successful paradigm of physics-informed learning (PIL) into a more general framework that also includes what we term physics-augmented learning (PAL). PIL and PAL complement each other by handling discriminative and generative properties, respectively. In numerical experiments, we show that PAL performs well on examples where PIL is inapplicable or inefficient.