AASIST: Audio Anti-Spoofing Using Integrated Spectro-Temporal Graph Attention Networks

By Jee-weon Jung, Hee-Soo Heo, Hemlata Tak, Hye-jin Shim, Joon Son Chung, Bong-Jin Lee, Ha-jin Yu, N. Evans. IEEE International Conference on Acoustics, Speech, and Signal Processing, 2021

580 citations recorded by the source corpus.

Artefacts that differentiate spoofed from bona-fide utterances can reside in specific temporal or spectral intervals. Their reliable detection usually depends upon computationally demanding ensemble systems where each subsystem is tuned to some specific artefacts. We seek to develop an efficient, single system that can detect a broad range of different spoofing attacks without score-level ensembles. We propose a novel heterogeneous stacking graph attention layer that models artefacts spanning heterogeneous temporal and spectral intervals with a heterogeneous attention mechanism and a stack node. With a new max graph operation that involves a competitive mechanism and a new readout scheme, our approach, named AASIST, outperforms the current state-of-the-art by 20% relative. Even a lightweight variant, AASIST-L, with only 85k parameters, outperforms all competing systems.

Subfields: Artificial Intelligence, Signal Processing

View Paper: https://www.semanticscholar.org/paper/83ed100306b892da76a7468e6eca6f99074fbd02

Citations in the corpus

• Automatic speaker verification spoofing and deepfake detection using wav2vec 2.0 and data augmentation (2022, unknown): “The last set of experiments were performed in order to gauge the relative importance of the AASIST and whether the improvements in generalisation are obtained for a simpler CM solu-tion.”