SZU-AFS Antispoofing System for the ASVspoof 5 Challenge
Yuxiong Xu, Jiafeng Zhong, Sengui Zheng, Zefeng Liu, and Bin Li

TL;DR
This paper introduces the SZU-AFS anti-spoofing system for the ASVspoof 5 Challenge, utilizing advanced data augmentation and a co-enhancement strategy to improve spoof detection accuracy.
Contribution
The paper proposes a novel anti-spoofing system combining multiple data augmentation policies and a gradient norm aware minimization strategy for enhanced generalization.
Findings
Achieved a minDCF of 0.115 on evaluation set
Attained an EER of 4.04% on evaluation set
Demonstrated effectiveness of co-enhancement strategy
Abstract
This paper presents the SZU-AFS anti-spoofing system, designed for Track 1 of the ASVspoof 5 Challenge under open conditions. The system is built with four stages: selecting a baseline model, exploring effective data augmentation (DA) methods for fine-tuning, applying a co-enhancement strategy based on gradient norm aware minimization (GAM) for secondary fine-tuning, and fusing logits scores from the two best-performing fine-tuned models. The system utilizes the Wav2Vec2 front-end feature extractor and the AASIST back-end classifier as the baseline model. During model fine-tuning, three distinct DA policies have been investigated: single-DA, random-DA, and cascade-DA. Moreover, the employed GAM-based co-enhancement strategy, designed to fine-tune the augmented model at both data and optimizer levels, helps the Adam optimizer find flatter minima, thereby boosting model generalization.…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsParticle accelerators and beam dynamics · Ultra-Wideband Communications Technology · Gyrotron and Vacuum Electronics Research
MethodsAttentive Walk-Aggregating Graph Neural Network · Adam
