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.
@article{arxiv.2604.26465,
title = {Diffusion Reconstruction towards Generalizable Audio Deepfake Detection},
author = {Bo Cheng and Songjun Cao and Xiaoming Zhang and Jie Chen and Long Ma and Fei Chen},
journal= {arXiv preprint arXiv:2604.26465},
year = {2026}
}
Comments
5 pages, this paper was submitted to Interspeech2026 for review