Kolmogorov-Arnold Networks: A Critical Assessment of Claims, Performance, and Practical Viability
Yuntian Hou, Tianrui Ji, Di Zhang, Angelos Stefanidis

TL;DR
This paper critically evaluates Kolmogorov-Arnold Networks, revealing they excel only in symbolic regression and face significant limitations in other domains due to computational costs and unsubstantiated theoretical claims.
Contribution
It provides a systematic, fair comparison of KANs against traditional models, clarifies their actual performance, and identifies conditions for their effective use and research gaps.
Findings
KANs outperform MLPs only in symbolic regression
KANs are 1.36-100x slower than traditional models
Theoretical claims about breaking the curse of dimensionality are unsubstantiated
Abstract
Kolmogorov-Arnold Networks (KANs) have gained significant attention as an alternative to traditional multilayer perceptrons, with proponents claiming superior interpretability and performance through learnable univariate activation functions. However, recent systematic evaluations reveal substantial discrepancies between theoretical claims and empirical evidence. This critical assessment examines KANs' actual performance across diverse domains using fair comparison methodologies that control for parameters and computational costs. Our analysis demonstrates that KANs outperform MLPs only in symbolic regression tasks, while consistently underperforming in machine learning, computer vision, and natural language processing benchmarks. The claimed advantages largely stem from B-spline activation functions rather than architectural innovations, and computational overhead (1.36-100x slower)…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsCognitive Computing and Networks
