HyperPotter: Spell the Charm of High-Order Interactions in Audio Deepfake Detection

TL;DR

HyperPotter introduces a hypergraph-based approach to model high-order interactions in audio deepfake detection, significantly improving accuracy and generalization across multiple datasets and attack types.

Contribution

It is the first to explicitly model high-order interactions in audio deepfake detection using hypergraphs with clustering-based hyperedges.

Findings

Achieves an average of 22.15% relative gain over baseline.

Outperforms state-of-the-art methods by 13.96% on cross-domain datasets.

Demonstrates strong generalization to diverse attacks and speakers.

Abstract

Advances in AIGC technologies have enabled the synthesis of highly realistic audio deepfakes capable of deceiving human auditory perception. Although numerous audio deepfake detection (ADD) methods have been developed, most rely on local temporal/spectral features or pairwise relations, overlooking high-order interactions (HOIs). HOIs capture discriminative patterns that emerge from multiple feature components beyond their individual contributions. We propose HyperPotter, a hypergraph-based framework that explicitly models these synergistic HOIs through clustering-based hyperedges with class-aware prototype initialization. Extensive experiments demonstrate that HyperPotter surpasses its baseline by an average relative gain of 22.15% across 11 datasets and outperforms state-of-the-art methods by 13.96% on 4 challenging cross-domain datasets, demonstrating superior generalization to diverse attacks and speakers.