Saved in:
Bibliographic Details
Main Authors: Abdolazizi, Kian P., Aydin, Roland C., Cyron, Christian J., Linka, Kevin
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2502.05682
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910847293456384
author Abdolazizi, Kian P.
Aydin, Roland C.
Cyron, Christian J.
Linka, Kevin
author_facet Abdolazizi, Kian P.
Aydin, Roland C.
Cyron, Christian J.
Linka, Kevin
contents Hybrid constitutive modeling integrates two complementary approaches for describing and predicting a material's mechanical behavior: purely data-driven black-box methods and physically constrained, theory-based models. While black-box methods offer high accuracy, they often lack interpretability and extrapolability. Conversely, physics-based models provide theoretical insight and generalizability but may not capture complex behaviors with the same accuracy. Traditionally, hybrid modeling has required a trade-off between these aspects. In this paper, we show how recent advances in symbolic machine learning, specifically Kolmogorov-Arnold Networks (KANs), help to overcome this limitation. We introduce Constitutive Kolmogorov-Arnold Networks (CKANs) as a new class of hybrid constitutive models. By incorporating a post-processing symbolification step, CKANs combine the predictive accuracy of data-driven models with the interpretability and extrapolation capabilities of symbolic expressions, bridging the gap between machine learning and physical modeling.
format Preprint
id arxiv_https___arxiv_org_abs_2502_05682
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Constitutive Kolmogorov-Arnold Networks (CKANs): Combining Accuracy and Interpretability in Data-Driven Material Modeling
Abdolazizi, Kian P.
Aydin, Roland C.
Cyron, Christian J.
Linka, Kevin
Computational Physics
Materials Science
Hybrid constitutive modeling integrates two complementary approaches for describing and predicting a material's mechanical behavior: purely data-driven black-box methods and physically constrained, theory-based models. While black-box methods offer high accuracy, they often lack interpretability and extrapolability. Conversely, physics-based models provide theoretical insight and generalizability but may not capture complex behaviors with the same accuracy. Traditionally, hybrid modeling has required a trade-off between these aspects. In this paper, we show how recent advances in symbolic machine learning, specifically Kolmogorov-Arnold Networks (KANs), help to overcome this limitation. We introduce Constitutive Kolmogorov-Arnold Networks (CKANs) as a new class of hybrid constitutive models. By incorporating a post-processing symbolification step, CKANs combine the predictive accuracy of data-driven models with the interpretability and extrapolation capabilities of symbolic expressions, bridging the gap between machine learning and physical modeling.
title Constitutive Kolmogorov-Arnold Networks (CKANs): Combining Accuracy and Interpretability in Data-Driven Material Modeling
topic Computational Physics
Materials Science
url https://arxiv.org/abs/2502.05682