发表机构
The Hong Kong Polytechnic University; Nanyang Technological University（香港理工大学; 南洋理工大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对语音深度伪造检测中域偏移问题，提出域梯度手术引导的元学习方法，通过非对称投影解决梯度冲突，显著降低等错误率。
AI 中文摘要
语音深度伪造检测因域偏移而面临重大挑战。域泛化（DG），特别是用于域泛化的元学习（MLDG），通过模拟和缓解域偏移提供了一种有前景的解决方案。然而，MLDG常常受到其元训练和元测试目标之间梯度冲突的阻碍，导致性能次优。为解决此问题，我们提出了域梯度手术（DGS），一种通过非对称投影策略解决冲突的元学习方法。DGS从元测试梯度中移除有害分量，确保与元训练梯度相比无冲突的优化轨迹。此外，我们引入了分层域梯度手术（LW-DGS），它是DGS的一种高效变体，能够动态识别并仅干预易冲突的层。在具有挑战性的基准上的大量实验表明，DGS-MLDG和LW-DGS-MLDG分别实现了平均相对等错误率（EER）降低5.29%和4.04%。
英文摘要
Speech deepfake detection faces significant challenges due to domain shifts. Domain generalization (DG), particularly meta-learning for domain generalization (MLDG), offers a promising solution by simulating and mitigating domain shifts. However, MLDG is often hindered by conflicting gradients between its meta-train and meta-test objectives, leading to suboptimal performance. To address this problem, we propose domain gradient surgery (DGS), a meta-learning method that resolves conflicts through an asymmetric projection strategy. DGS removes the destructive component from the meta-test gradient, ensuring a conflict-free optimization trajectory versus the meta-train gradient. Furthermore, we introduce layer-wise DGS (LW-DGS), an efficient variant of DGS that dynamically identifies and intervenes only conflict-prone layers. Extensive experiments on challenging benchmarks demonstrate that DGS-MLDG and LW-DGS-MLDG achieve an average relative EER reduction of 5.29% and 4.04%, respectively.
CommentsAccepted by INTERSPEECH 2026
