Shift-Invariant Attribute Scoring for Kolmogorov-Arnold Networks via Shapley Value
This work addresses a domain-specific challenge for improving interpretability and deployment of KANs in resource-constrained environments, representing an incremental advancement.
The paper tackles the problem of unreliable network pruning in Kolmogorov-Arnold Networks (KANs) due to sensitivity to input shifts, proposing ShapKAN, a pruning framework using Shapley value attribution that ensures consistent importance rankings and enables effective network compression, as demonstrated in experiments on synthetic and real-world datasets.
For many real-world applications, understanding feature-outcome relationships is as crucial as achieving high predictive accuracy. While traditional neural networks excel at prediction, their black-box nature obscures underlying functional relationships. Kolmogorov--Arnold Networks (KANs) address this by employing learnable spline-based activation functions on edges, enabling recovery of symbolic representations while maintaining competitive performance. However, KAN's architecture presents unique challenges for network pruning. Conventional magnitude-based methods become unreliable due to sensitivity to input coordinate shifts. We propose \textbf{ShapKAN}, a pruning framework using Shapley value attribution to assess node importance in a shift-invariant manner. Unlike magnitude-based approaches, ShapKAN quantifies each node's actual contribution, ensuring consistent importance rankings regardless of input parameterization. Extensive experiments on synthetic and real-world datasets demonstrate that ShapKAN preserves true node importance while enabling effective network compression. Our approach improves KAN's interpretability advantages, facilitating deployment in resource-constrained environments.