Optimal Abstractions for Verifying Properties of Kolmogorov-Arnold Networks (KANs)
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866908817984323584 |
|---|---|
| author | Schwartz, Noah Nagesh, Chandra Kanth Sankaranarayanan, Sriram Kaur, Ramneet Sahai, Tuhin Jha, Susmit |
| author_facet | Schwartz, Noah Nagesh, Chandra Kanth Sankaranarayanan, Sriram Kaur, Ramneet Sahai, Tuhin Jha, Susmit |
| contents | We present a novel approach for verifying properties of Kolmogorov-Arnold Networks (KANs), a class of neural networks characterized by nonlinear, univariate activation functions typically implemented as piecewise polynomial splines or Gaussian processes. Our method creates mathematical ``abstractions'' by replacing each KAN unit with a piecewise affine (PWA) function, providing both local and global error estimates between the original network and its approximation. These abstractions enable property verification by encoding the problem as a Mixed Integer Linear Program (MILP), determining whether outputs satisfy specified properties when inputs belong to a given set. A critical challenge lies in balancing the number of pieces in the PWA approximation: too many pieces add binary variables that make verification computationally intractable, while too few pieces create excessive error margins that yield uninformative bounds. Our key contribution is a systematic framework that exploits KAN structure to find optimal abstractions. By combining dynamic programming at the unit level with a knapsack optimization across the network, we minimize the total number of pieces while guaranteeing specified error bounds. This approach determines the optimal approximation strategy for each unit while maintaining overall accuracy requirements. Empirical evaluation across multiple KAN benchmarks demonstrates that the upfront analysis costs of our method are justified by superior verification results. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_06737 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Optimal Abstractions for Verifying Properties of Kolmogorov-Arnold Networks (KANs) Schwartz, Noah Nagesh, Chandra Kanth Sankaranarayanan, Sriram Kaur, Ramneet Sahai, Tuhin Jha, Susmit Machine Learning Artificial Intelligence Logic in Computer Science 68T07 We present a novel approach for verifying properties of Kolmogorov-Arnold Networks (KANs), a class of neural networks characterized by nonlinear, univariate activation functions typically implemented as piecewise polynomial splines or Gaussian processes. Our method creates mathematical ``abstractions'' by replacing each KAN unit with a piecewise affine (PWA) function, providing both local and global error estimates between the original network and its approximation. These abstractions enable property verification by encoding the problem as a Mixed Integer Linear Program (MILP), determining whether outputs satisfy specified properties when inputs belong to a given set. A critical challenge lies in balancing the number of pieces in the PWA approximation: too many pieces add binary variables that make verification computationally intractable, while too few pieces create excessive error margins that yield uninformative bounds. Our key contribution is a systematic framework that exploits KAN structure to find optimal abstractions. By combining dynamic programming at the unit level with a knapsack optimization across the network, we minimize the total number of pieces while guaranteeing specified error bounds. This approach determines the optimal approximation strategy for each unit while maintaining overall accuracy requirements. Empirical evaluation across multiple KAN benchmarks demonstrates that the upfront analysis costs of our method are justified by superior verification results. |
| title | Optimal Abstractions for Verifying Properties of Kolmogorov-Arnold Networks (KANs) |
| topic | Machine Learning Artificial Intelligence Logic in Computer Science 68T07 |
| url | https://arxiv.org/abs/2602.06737 |