Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Kolmogorov-Arnold Networks are Radial Basis Function Networks

About

This short paper is a fast proof-of-concept that the 3-order B-splines used in Kolmogorov-Arnold Networks (KANs) can be well approximated by Gaussian radial basis functions. Doing so leads to FastKAN, a much faster implementation of KAN which is also a radial basis function (RBF) network.

Ziyao Li• 2024

Related benchmarks

TaskDatasetResultRank
Human Activity RecognitionPAMAP2--
54
Human Activity RecognitionOpportunity
Macro F143.4
43
Activity RecognitionmHealth--
35
Human Activity RecognitionSKODA
Macro F188.7
29
Human Activity RecognitionMotionSense
Macro-F182.8
29
Human Activity RecognitionHAPT
Macro-F167.3
20
Activity RecognitionDSADS
Macro F148.6
20
Activity RecognitionDG
Macro F1 Score56.2
20
Physics-Informed PDE Solving2D Helmholtz (test)
Time (sec/iter)0.0228
14
Function ApproximationFunction Approximation [np, 6, 1]
Time (sec/iter)0.0305
14
Showing 10 of 14 rows

Other info

Follow for update