Search papers, labs, and topics across Lattice.
This paper tackles the challenge of generalizing audio deepfake detection (ADD) to unseen attacks by focusing on hard sample classification. They generate hard samples using a diffusion-based reconstruction method, finding it superior to other reconstruction paradigms. The approach further incorporates multi-layer feature aggregation and a Regularization-Assisted Contrastive Learning (RACL) objective to improve generalizability, achieving state-of-the-art results.
Audio deepfake detectors trained on diffusion-reconstructed "hard" examples generalize far better to unseen attacks, slashing error rates compared to standard training.
Achieving robust generalization against unseen attacks remains a challenge in Audio Deepfake Detection (ADD), driven by the rapid evolution of generative models. To address this, we propose a framework centered on hard sample classification. The core idea is that a model capable of distinguishing challenging hard samples is inherently equipped to handle simpler cases effectively. We investigate multiple reconstruction paradigms, identifying the diffusion-based method as optimal for generating hard samples. Furthermore, we leverage multi-layer feature aggregation and introduce a Regularization-Assisted Contrastive Learning (RACL) objective to enhance generalizability. Experiments demonstrate the superior generalization of our approach, with our best model achieving a significant reduction in the average Equal Error Rate (EER) compared to the baseline.
