Theory-guided hard constraint projection (HCP): a knowledge-based data-driven scientific machine learning method

TL;DR

This paper introduces a theory-guided hard constraint projection (HCP) method that integrates physical laws into neural network models, ensuring predictions adhere to physical constraints while requiring less data and improving accuracy.

Contribution

The study proposes a novel HCP approach that enforces physical constraints through projection, combining domain knowledge with neural networks for more reliable scientific modeling.

Findings

HCP outperforms traditional neural networks in accuracy.

HCP requires less data than soft constraint models.

HCP demonstrates robustness to noisy data.

Abstract

Machine learning models have been successfully used in many scientific and engineering fields. However, it remains difficult for a model to simultaneously utilize domain knowledge and experimental observation data. The application of knowledge-based symbolic AI represented by an expert system is limited by the expressive ability of the model, and data-driven connectionism AI represented by neural networks is prone to produce predictions that violate physical mechanisms. In order to fully integrate domain knowledge with observations, and make full use of the prior information and the strong fitting ability of neural networks, this study proposes theory-guided hard constraint projection (HCP). This model converts physical constraints, such as governing equations, into a form that is easy to handle through discretization, and then implements hard constraint optimization through projection. Based on rigorous mathematical proofs, theory-guided HCP can ensure that model predictions strictly conform to physical mechanisms in the constraint patch. The performance of the theory-guided HCP is verified by experiments based on the heterogeneous subsurface flow problem. Due to the application of hard constraints, compared with fully connected neural networks and soft constraint models, such as theory-guided neural networks and physics-informed neural networks, theory-guided HCP requires fewer data, and achieves higher prediction accuracy and stronger robustness to noisy observations.