KAN-SAs: Efficient Acceleration of Kolmogorov-Arnold Networks on Systolic Arrays

TL;DR

This paper introduces KAN-SAs, a specialized systolic array accelerator that efficiently accelerates Kolmogorov-Arnold Networks by exploiting B-spline properties, achieving high utilization and reduced inference cycles.

Contribution

The work presents a novel SA-based accelerator, KAN-SAs, tailored for KANs, with a nonrecursive B-spline implementation and sparsity exploitation for improved efficiency.

Findings

Achieves up to 100% SA utilization.

Reduces clock cycles by up to 50%.

Demonstrates efficiency across various KAN applications.

Abstract

Kolmogorov-Arnold Networks (KANs) have garnered significant attention for their promise of improved parameter efficiency and explainability compared to traditional Deep Neural Networks (DNNs). KANs' key innovation lies in the use of learnable non-linear activation functions, which are parametrized as splines. Splines are expressed as a linear combination of basis functions (B-splines). B-splines prove particularly challenging to accelerate due to their recursive definition. Systolic Array (SA)based architectures have shown great promise as DNN accelerators thanks to their energy efficiency and low latency. However, their suitability and efficiency in accelerating KANs have never been assessed. Thus, in this work, we explore the use of SA architecture to accelerate the KAN inference. We show that, while SAs can be used to accelerate part of the KAN inference, their utilization can be reduced to 30%. Hence, we propose KAN-SAs, a novel SA-based accelerator that leverages intrinsic properties of B-splines to enable efficient KAN inference. By including a nonrecursive B-spline implementation and leveraging the intrinsic KAN sparsity, KAN-SAs enhances conventional SAs, enabling efficient KAN inference, in addition to conventional DNNs. KAN-SAs achieves up to 100% SA utilization and up to 50% clock cycles reduction compared to conventional SAs of equivalent area, as shown by hardware synthesis results on a 28nm FD-SOI technology. We also evaluate different configurations of the accelerator on various KAN applications, confirming the improved efficiency of KAN inference provided by KAN-SAs.