Saved in:
Bibliographic Details
Main Authors: Gao, Wenhan, Luo, Jian, Wan, Fang, Xu, Ruichen, Liu, Xiang, Xing, Haipeng, Liu, Yi
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2510.02683
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914072504565760
author Gao, Wenhan
Luo, Jian
Wan, Fang
Xu, Ruichen
Liu, Xiang
Xing, Haipeng
Liu, Yi
author_facet Gao, Wenhan
Luo, Jian
Wan, Fang
Xu, Ruichen
Liu, Xiang
Xing, Haipeng
Liu, Yi
contents Recently, neural operators have emerged as powerful tools for learning mappings between function spaces, enabling data-driven simulations of complex dynamics. Despite their successes, a deeper understanding of their learning mechanisms remains underexplored. In this work, we classify neural operators into two types: (1) Spatial domain models that learn on grids and (2) Functional domain models that learn with function bases. We present several viewpoints based on this classification and focus on learning data-driven dynamics adhering to physical principles. Specifically, we provide a way to explain the prediction-making process of neural operators and show that neural operator can learn hidden physical patterns from data. However, this explanation method is limited to specific situations, highlighting the urgent need for generalizable explanation methods. Next, we show that a simple dual-space multi-scale model can achieve SOTA performance and we believe that dual-space multi-spatio-scale models hold significant potential to learn complex physics and require further investigation. Lastly, we discuss the critical need for principled frameworks to incorporate known physics into neural operators, enabling better generalization and uncovering more hidden physical phenomena.
format Preprint
id arxiv_https___arxiv_org_abs_2510_02683
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Can Data-Driven Dynamics Reveal Hidden Physics? There Is A Need for Interpretable Neural Operators
Gao, Wenhan
Luo, Jian
Wan, Fang
Xu, Ruichen
Liu, Xiang
Xing, Haipeng
Liu, Yi
Machine Learning
Artificial Intelligence
Recently, neural operators have emerged as powerful tools for learning mappings between function spaces, enabling data-driven simulations of complex dynamics. Despite their successes, a deeper understanding of their learning mechanisms remains underexplored. In this work, we classify neural operators into two types: (1) Spatial domain models that learn on grids and (2) Functional domain models that learn with function bases. We present several viewpoints based on this classification and focus on learning data-driven dynamics adhering to physical principles. Specifically, we provide a way to explain the prediction-making process of neural operators and show that neural operator can learn hidden physical patterns from data. However, this explanation method is limited to specific situations, highlighting the urgent need for generalizable explanation methods. Next, we show that a simple dual-space multi-scale model can achieve SOTA performance and we believe that dual-space multi-spatio-scale models hold significant potential to learn complex physics and require further investigation. Lastly, we discuss the critical need for principled frameworks to incorporate known physics into neural operators, enabling better generalization and uncovering more hidden physical phenomena.
title Can Data-Driven Dynamics Reveal Hidden Physics? There Is A Need for Interpretable Neural Operators
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2510.02683