![[2602.17750] Inelastic Constitutive Kolmogorov-Arnold Networks: A generalized framework for automated discovery of interpretable inelastic material models](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.17750] Inelastic Constitutive Kolmogorov-Arnold Networks: A generalized framework for automated discovery of interpretable inelastic material models
Summary
This article presents Inelastic Constitutive Kolmogorov-Arnold Networks (iCKANs), a novel machine learning framework for automating the discovery of interpretable inelastic material models, enhancing the understanding of material behavior under various conditions.
Why It Matters
The development of iCKANs represents a significant advancement in materials science and machine learning, enabling researchers to derive symbolic constitutive laws from experimental data. This has implications for improving material design and understanding complex material behaviors, particularly in engineering applications.
Key Takeaways
- iCKANs automate the discovery of constitutive laws for both elastic and inelastic materials.
- The framework preserves physical interpretability while capturing complex material behaviors.
- iCKANs can incorporate additional information, such as temperature effects, enhancing their utility.
Condensed Matter > Materials Science arXiv:2602.17750 (cond-mat) [Submitted on 19 Feb 2026] Title:Inelastic Constitutive Kolmogorov-Arnold Networks: A generalized framework for automated discovery of interpretable inelastic material models Authors:Chenyi Ji, Kian P. Abdolazizi, Hagen Holthusen, Christian J. Cyron, Kevin Linka View a PDF of the paper titled Inelastic Constitutive Kolmogorov-Arnold Networks: A generalized framework for automated discovery of interpretable inelastic material models, by Chenyi Ji and 4 other authors View PDF HTML (experimental) Abstract:A key problem of solid mechanics is the identification of the constitutive law of a material, that is, the relation between strain and stress. Machine learning has lead to considerable advances in this field lately. Here we introduce inelastic Constitutive Kolmogorov-Arnold Networks (iCKANs). This novel artificial neural network architecture can discover in an automated manner symbolic constitutive laws describing both the elastic and inelastic behavior of materials. That is, it can translate data from material testing into corresponding elastic and inelastic potential functions in closed mathematical form. We demonstrate the advantages of iCKANs using both synthetic data and experimental data of the viscoelastic polymer materials VHB 4910 and VHB 4905. The results demonstrate that iCKANs accurately capture complex viscoelastic behavior while preserving physical interpretability. It is a particular strength of ...
