Improving Generalizability of Kolmogorov-Arnold Networks via Error-Correcting Output Codes

2025-05-09

AI Summary

Kolmogorov–Arnold Networks (KANs) suffer from weak generalization and high sensitivity to hyperparameters in multi-class medical image classification. Method: This work introduces the Error-Correcting Output Codes (ECOC) framework to KANs for the first time, decomposing multi-class tasks into robust binary subproblems and employing Hamming-distance-based decoding to enhance discriminative stability. The approach is compatible with FastKAN, FasterKAN, and other variants, and integrates spline basis functions to enable interpretable modeling. Contribution/Results: Evaluated on a blood cell classification dataset, the proposed method achieves significant accuracy improvements and maintains consistent superiority across diverse hyperparameter configurations. Ablation studies confirm that ECOC delivers uniform performance gains across various KAN architectures. This work substantially advances the practicality and reliability of activation-free neural networks in high-stakes medical AI applications.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Computer Vision: Multi-modal Vision

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Kolmogorov-Arnold Networks (KAN) offer universal function approximation using univariate spline compositions without nonlinear activations. In this work, we integrate Error-Correcting Output Codes (ECOC) into the KAN framework to transform multi-class classification into multiple binary tasks, improving robustness via Hamming-distance decoding. Our proposed KAN with ECOC method outperforms vanilla KAN on a challenging blood cell classification dataset, achieving higher accuracy under diverse hyperparameter settings. Ablation studies further confirm that ECOC consistently enhances performance across FastKAN and FasterKAN variants. These results demonstrate that ECOC integration significantly boosts KAN generalizability in critical healthcare AI applications. To the best of our knowledge, this is the first integration of ECOC with KAN for enhancing multi-class medical image classification performance.

Problem

• Enhancing Kolmogorov-Arnold Networks generalizability via ECOC
• Improving multi-class classification robustness with Hamming decoding
• Boosting medical image classification accuracy in healthcare AI

Innovation

• Integrates ECOC into KAN for robustness
• Transforms multi-class to binary tasks
• Enhances medical image classification accuracy