Projected Neural Differential Equations for Learning Constrained Dynamics

TL;DR

This paper introduces projected neural differential equations (PNDEs), a novel method that enforces known constraints on learned dynamics, improving accuracy and stability in complex systems like chaotic models and power grids.

Contribution

The paper proposes PNDEs, a new approach that incorporates constraints into neural differential equations via projection, enhancing their generalizability and numerical stability.

Findings

PNDEs outperform existing methods on challenging dynamical systems.

PNDEs require fewer hyperparameters than previous approaches.

PNDEs demonstrate improved modeling of constrained systems in complex domains.

Abstract

Neural differential equations offer a powerful approach for learning dynamics from data. However, they do not impose known constraints that should be obeyed by the learned model. It is well-known that enforcing constraints in surrogate models can enhance their generalizability and numerical stability. In this paper, we introduce projected neural differential equations (PNDEs), a new method for constraining neural differential equations based on projection of the learned vector field to the tangent space of the constraint manifold. In tests on several challenging examples, including chaotic dynamical systems and state-of-the-art power grid models, PNDEs outperform existing methods while requiring fewer hyperparameters. The proposed approach demonstrates significant potential for enhancing the modeling of constrained dynamical systems, particularly in complex domains where accuracy and reliability are essential.