发表机构
University of Surrey; Guangxi University; Shenzhen University of Advanced Technology; Adelaide University（萨里大学; 广西大学; 深圳理工大学; 阿德莱德大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对音频深度伪造检测中新兴攻击的泛化问题，提出SE-ADD自进化框架，利用ALM自身错误驱动的监督和低秩适配迭代调整，显著降低等错误率。
AI 中文摘要
音频深度伪造检测（ADD）必须在部署后出现新的欺骗攻击时保持有效性。新兴的基于音频语言模型（ALM）的ADD方法建立在来自真实标签或已验证取证依据的预定义监督之上。然而，这种范式忽视了ALM自身的错误，而这些错误恰恰指示了最需要针对性监督的地方。为此，我们首先为基于ALM的ADD引入进化欺骗环境，在该环境中，新攻击成为主导，同时先前观察到的攻击持续存在。受上述从错误中学习视角的启发，我们进一步提出SE-ADD，一种自进化框架，通过低秩适配（LoRA）迭代地调整ALM，利用基于其判定和自生成取证线索构建的错误驱动监督。所有训练样本都接受直接的真实性监督，而错误分类的样本则接受额外的线索增强监督。随着更新后的ALM重新生成判定和线索，相应的监督也随之进化。在两个ALM上的实验证明了SE-ADD在泛化到未见攻击方面的有效性，将Qwen2-Audio的等错误率（EER）从36.72%降至7.52%，将MOSS-Audio的等错误率从19.93%降至3.97%。
英文摘要
Audio deepfake detection (ADD) must remain effective when new spoofing attacks emerge after deployment. Emerging audio language model (ALM)-based ADD methods are built on predefined supervision from ground-truth labels or verified forensic rationales. However, this paradigm overlooks an ALM's own mistakes, which indicate where targeted supervision is most needed. To this end, we first introduce evolving spoofing environments for ALM-based ADD, where a new attack becomes dominant while previously observed attacks persist. Motivated by the above learning-from-mistakes perspective, we further propose SE-ADD, a self-evolving framework that iteratively adapts an ALM via low-rank adaptation (LoRA) using mistake-driven supervision built from its verdicts and self-generated forensic cues. All training samples receive direct authenticity supervision, while misclassified ones receive additional cue-augmented supervision. As verdicts and cues are regenerated by the updated ALM, the resulting supervision evolves accordingly. Experiments on two ALMs demonstrate the effectiveness of SE-ADD in generalizing to unseen attacks, reducing the equal error rate (EER) from $36.72\%$ to $7.52\%$ for Qwen2-Audio and from $19.93\%$ to $3.97\%$ for MOSS-Audio.
