CODE: A global approach to ODE dynamics learning

TL;DR

This paper introduces ChaosODE (CODE), a polynomial chaos expansion method for data-driven learning of ODE dynamics that demonstrates superior extrapolation and robustness over neural and kernel methods, especially with sparse and noisy data.

Contribution

The paper presents a novel global polynomial chaos approach for learning ODE dynamics directly from data, improving extrapolation and robustness compared to existing neural and kernel methods.

Findings

CODE outperforms neural and kernel methods in extrapolation tasks.

It maintains robustness under noisy and sparse data conditions.

The approach provides practical guidelines for optimization in dynamics learning.

Abstract

Ordinary differential equations (ODEs) are a conventional way to describe the observed dynamics of physical systems. Scientists typically hypothesize about dynamical behavior, propose a mathematical model, and compare its predictions to data. However, modern computing and algorithmic advances now enable purely data-driven learning of governing dynamics directly from observations. In data-driven settings, one learns the ODE's right-hand side (RHS). Dense measurements are often assumed, yet high temporal resolution is typically both cumbersome and expensive. Consequently, one usually has only sparsely sampled data. In this work we introduce ChaosODE (CODE), a Polynomial Chaos ODE Expansion in which we use an arbitrary Polynomial Chaos Expansion (aPCE) for the ODE's right-hand side, resulting in a global orthonormal polynomial representation of dynamics. We evaluate the performance of CODE in several experiments on the Lotka-Volterra system, across varying noise levels, initial conditions, and predictions far into the future, even on previously unseen initial conditions. CODE exhibits remarkable extrapolation capabilities even when evaluated under novel initial conditions and shows advantages compared to well-examined methods using neural networks (NeuralODE) or kernel approximators (KernelODE) as the RHS representer. We observe that the high flexibility of NeuralODE and KernelODE degrades extrapolation capabilities under scarce data and measurement noise. Finally, we provide practical guidelines for robust optimization of dynamics-learning problems and illustrate them in the accompanying code.