LTBs-KAN: Linear-Time B-splines Kolmogorov-Arnold Networks

TL;DR

This paper introduces LTBs-KAN, a linear-time B-splines Kolmogorov-Arnold Network that enhances explainability and efficiency over traditional KANs, enabling faster computation and reduced parameters.

Contribution

The paper proposes a novel linear-time B-splines KAN architecture with reduced computational complexity and parameter count, improving practicality and scalability.

Findings

Achieves linear time complexity in B-spline computations.

Reduces model parameters via product-of-sums matrix factorization.

Demonstrates competitive performance on MNIST, Fashion-MNIST, and CIFAR-10.

Abstract

Kolmogorov-Arnold Networks (KANs) are a recent neural network architecture offering an alternative to Multilayer Perceptrons (MLPs) with improved explainability and expressibility. However, KANs are significantly slower than MLPs due to the recursive nature of B-spline function computations, limiting their application. This work addresses these issues by proposing a novel base-spline Linear-Time B-splines Kolmogorov-Arnold Network (LTBs-KAN) with linear complexity. Unlike previous methods that rely on the Boor-Mansfield-Cox spline algorithm or other computationally intensive mathematical functions, our approach significantly reduces the computational burden. Additionally, we further reduce model's parameter through product-of-sums matrix factorization in the forward pass without sacrificing performance. Experiments on MNIST, Fashion-MNIST and CIFAR-10 demonstrate that LTBs-KAN achieves good time complexity and parameter reduction, when used as building architectural blocks, compared to other KAN implementations.