Pushing the Performance of Synthetic Speech Detection with Kolmogorov-Arnold Networks and Self-Supervised Learning Models

TL;DR

This paper introduces a novel architecture combining Kolmogorov-Arnold Networks with self-supervised learning models to significantly improve synthetic speech detection performance, addressing challenges from advanced speech synthesis spoofing attacks.

Contribution

It proposes replacing the MLP in SSL models with KAN, based on the Kolmogorov-Arnold theorem, achieving substantial performance gains in synthetic speech detection.

Findings

60.55% relative improvement on ASVspoof2021 LA and DF sets

Achieved 0.70% EER on the 21LA set

Demonstrated effectiveness of KAN in SSL-based speech detection

Abstract

Recent advancements in speech synthesis technologies have led to increasingly advanced spoofing attacks, posing significant challenges for automatic speaker verification systems. While systems based on self-supervised learning (SSL) models, particularly the XLSR-Conformer model, have demonstrated remarkable performance in synthetic speech detection, there remains room for architectural improvements. In this paper, we propose a novel approach that replaces the traditional Multi-Layer Perceptron in the XLSR-Conformer model with a Kolmogorov-Arnold Network (KAN), a novel architecture based on the Kolmogorov-Arnold representation theorem. Our results on ASVspoof2021 demonstrate that integrating KAN into the SSL-based models can improve the performance by 60.55% relatively on LA and DF sets, further achieving 0.70% EER on the 21LA set. These findings suggest that incorporating KAN into SSL-based models is a promising direction for advances in synthetic speech detection.