Application of Kolmogorov-Arnold Networks in high energy physics

TL;DR

This paper investigates the use of Kolmogorov-Arnold Networks in high-energy physics, demonstrating their potential to outperform traditional neural networks and enhance interpretability in complex particle collision tasks.

Contribution

It is the first to apply KANs to high-energy physics, showing their effectiveness in classifying multijet processes and reconstructing missing transverse momentum.

Findings

KANs outperform traditional neural networks in key physics tasks

KANs provide more interpretable symbolic formulas

Successful application to dark matter event reconstruction

Abstract

Kolmogorov-Arnold Networks represent a recent advancement in machine learning, with the potential to outperform traditional perceptron-based neural networks across various domains as well as provide more interpretability with the use of symbolic formulas and pruning. This study explores the application of KANs to specific tasks in high-energy physics. We evaluate the performance of KANs in distinguishing multijet processes in proton-proton collisions and in reconstructing missing transverse momentum in events involving dark matter.