Comparative Evaluation of Kolmogorov-Arnold Autoencoders and Orthogonal Autoencoders for Fault Detection with Varying Training Set Sizes

TL;DR

This study compares various Kolmogorov-Arnold autoencoders with orthogonal autoencoders for fault detection in chemical processes, highlighting their data efficiency and performance across different training set sizes.

Contribution

It introduces a comprehensive evaluation of four KAN-AE variants for unsupervised fault detection, demonstrating their effectiveness and robustness in low-data scenarios.

Findings

WavKAN-AE achieves ≥92% FDR with 4,000 samples.

EfficientKAN-AE reaches ≥90% FDR with only 500 samples.

KAN-AEs outperform the orthogonal autoencoder, especially with limited data.

Abstract

Kolmogorov-Arnold Networks (KANs) have recently emerged as a flexible and parameter-efficient alternative to conventional neural networks. Unlike standard architectures that use fixed node-based activations, KANs place learnable functions on edges, parameterized by different function families. While they have shown promise in supervised settings, their utility in unsupervised fault detection remains largely unexplored. This study presents a comparative evaluation of KAN-based autoencoders (KAN-AEs) for unsupervised fault detection in chemical processes. We investigate four KAN-AE variants, each based on a different KAN implementation (EfficientKAN, FastKAN, FourierKAN, and WavKAN), and benchmark them against an Orthogonal Autoencoder (OAE) on the Tennessee Eastman Process. Models are trained on normal operating data across 13 training set sizes and evaluated on 21 fault types, using Fault Detection Rate (FDR) as the performance metric. WavKAN-AE achieves the highest overall FDR ($\geq$92\%) using just 4,000 training samples and remains the top performer, even as other variants are trained on larger datasets. EfficientKAN-AE reaches $\geq$90\% FDR with only 500 samples, demonstrating robustness in low-data settings. FastKAN-AE becomes competitive at larger scales ($\geq$50,000 samples), while FourierKAN-AE consistently underperforms. The OAE baseline improves gradually but requires substantially more data to match top KAN-AE performance. These results highlight the ability of KAN-AEs to combine data efficiency with strong fault detection performance. Their use of structured basis functions suggests potential for improved model transparency, making them promising candidates for deployment in data-constrained industrial settings.