BSRBF-KAN: A combination of B-splines and Radial Basis Functions in Kolmogorov-Arnold Networks

TL;DR

BSRBF-KAN is introduced, a Kolmogorov Arnold Network that combines B-splines and radial basis functions (RBFs) to fit input vectors during data training and expects BSRBF-KAN to open many combinations of mathematical functions to design KANs.

Abstract

In this paper, we introduce BSRBF-KAN, a Kolmogorov Arnold Network (KAN) that combines B-splines and radial basis functions (RBFs) to fit input vectors during data training. We perform experiments with BSRBF-KAN, multi-layer perception (MLP), and other popular KANs, including EfficientKAN, FastKAN, FasterKAN, and GottliebKAN over the MNIST and Fashion-MNIST datasets. BSRBF-KAN shows stability in 5 training runs with a competitive average accuracy of 97.55% on MNIST and 89.33% on Fashion-MNIST and obtains convergence better than other networks. We expect BSRBF-KAN to open many combinations of mathematical functions to design KANs. Our repo is publicly available at: https://github.com/hoangthangta/BSRBF_KAN.

Figures (2)

• Figure 1: The logarithmic values of training losses during a training run over 15 epochs on MNIST and 25 epochs on Fashion-MNIST.
• Figure 2: Heatmap of the misclassified images in the test set by models over MNIST and Fashion-MNIST.