Kolmogorov-Arnold Network Autoencoders

TL;DR

This paper investigates Kolmogorov-Arnold Networks (KANs) as a novel alternative to traditional neural networks, focusing on their application in autoencoders for data representation and comparing their performance with CNNs on standard datasets.

Contribution

It introduces KANs with edge-based activation functions into autoencoder architectures and evaluates their effectiveness against CNNs on multiple datasets.

Findings

KAN autoencoders achieve competitive reconstruction accuracy

KANs show potential for improved interpretability

Performance comparable to CNNs on MNIST, SVHN, CIFAR-10

Abstract

Deep learning models have revolutionized various domains, with Multi-Layer Perceptrons (MLPs) being a cornerstone for tasks like data regression and image classification. However, a recent study has introduced Kolmogorov-Arnold Networks (KANs) as promising alternatives to MLPs, leveraging activation functions placed on edges rather than nodes. This structural shift aligns KANs closely with the Kolmogorov-Arnold representation theorem, potentially enhancing both model accuracy and interpretability. In this study, we explore the efficacy of KANs in the context of data representation via autoencoders, comparing their performance with traditional Convolutional Neural Networks (CNNs) on the MNIST, SVHN, and CIFAR-10 datasets. Our results demonstrate that KAN-based autoencoders achieve competitive performance in terms of reconstruction accuracy, thereby suggesting their viability as effective tools in data analysis tasks.