Existing deepfake speech detection systems lack generalizability to unseen
attacks (i.e., samples generated by generative algorithms not seen during
training). Recent studies have explored the use of universal speech
representations to tackle this issue and have obtained inspiring results. These
works, however, have focused on innovating downstream classifiers while leaving
the representation itself untouched. In this study, we argue that
characterizing the long-term temporal dynamics of these representations is
crucial for generalizability and propose a new method to assess representation
dynamics. Indeed, we show that different generative models generate similar
representation dynamics patterns with our proposed method. Experiments on the
ASVspoof 2019 and 2021 datasets validate the benefits of the proposed method to
detect deepfakes from methods unseen during training, significantly improving
on several benchmark methods.

现有的深度伪造语音检测系统对未知攻击（即在训练期间未见过的生成算法生成的样本）缺乏普适性。最近的研究探索了使用通用语音表示来解决这个问题，并取得了令人鼓舞的结果。然而，这些工作主要关注创新下游分类器，而对表示本身却没有触及。本研究中，我们认为表征这些表示的长期时间动态对于普适性至关重要，并提出了一种新的评估表示动态的方法。实验证明，不同的生成模型使用我们提出的方法生成类似的表示动态模式。在 ASVspoof 2019 和 2021 数据集上的实验验证了该方法在检测未在训练中出现的深度伪造方法方面的优势，并在几个基准方法上取得了显著改进。