Physics‐informed Deep Learning

View at publisher Summarise with VIRAT
Authors: Nithin Chalapathi, Yiheng Du, Sanjeev Raja, Aditi S. Krishnapriyan
Cited by: 0
References: 109

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The rapid success of machine learning in various fields has sparked significant interest in applying and developing techniques to scientific spatial and temporal modeling. This has led to the development of the field known as physics-informed machine learning (PIML). PIML integrates traditional scientific modeling with modern data-driven techniques to enhance the understanding and prediction of complex systems across scientific and engineering domains. While traditional machine learning excels at pattern recognition and prediction from large datasets, it often operates independently of the physical principles governing the systems. Conversely, physical models offer high interpretability and reliability but struggle with real-world data complexity. PIML aims to address these limitations by combining the strengths of both approaches, promising improved accuracy, generalization, and scalability. This chapter discusses different facets of PIML including data generation, current architectures, training objectives, open problems, and the role of scaling in handling larger datasets and complex models.

Fields of Study & Subjects

Model Reduction and Neural Networks Neural Networks and Applications Machine Learning in Materials Science

Publication details

Journal / source
Wiley eBooks
Pages
269–306
Publication date
20 March 2026
Language
EN
Document type
Book chapter