对比学习中用余弦相似度+延迟队列，能更好检测深度伪造语音。
Similarity Choice and Negative Scaling in Supervised Contrastive Learning for Deepfake Audio Detection
- 用余弦与球面角相似度对比，测试对伪造语音识别的影响。
- 余弦相似度配延迟队列，使ITW错误率降至8.29%，综合错误率4.44%。
- 球面角相似度无需大量负样本也能表现好，适合资源有限场景。
监督对比学习（SupCon）广泛用于表征学习，但在语音深度伪造检测中研究较少。现有工作通常将对比损失嵌入整体框架，缺乏对SupCon本身机制的深入分析。本文以wav2vec2 XLS-R（300M）为基础，在控制条件下考察了两个关键因素：(i) SupCon中的相似度度量方式（余弦相似度与基于超球面角度的角相似度），以及 (ii) 使用冷启动全局跨批队列进行负样本缩放。实验分为两阶段：第一阶段使用SupCon微调编码器与投影头；第二阶段冻结模型，仅训练线性分类器（采用BCE损失）。在ASVspoof 2019 LA上训练，评估在ASV19 eval、ITW和ASVspoof 2021 DF/LA数据集上。结果表明，余弦相似度配合延迟队列的设置在ITW上的错误率最低（8.29%），综合错误率为4.44%；而角相似度在无队列负样本时表现良好（ITW EER 8.70%），说明其对大规模负样本依赖较低。
原文摘要 · Abstract (English)
Supervised contrastive learning (SupCon) is widely used to shape representations, but has seen limited targeted study for audio deepfake detection. Existing work typically combines contrastive terms with broader pipelines; however, the focus on SupCon itself is missing. In this work, we run a controlled study on wav2vec2 XLS-R (300M) that varies (i) similarity in SupCon (cosine vs angular similarity derived from the hyperspherical angle) and (ii) negative scaling using a warm-started global cross-batch queue. Stage 1 fine-tunes the encoder and projection head with SupCon; Stage 2 freezes them and trains a linear classifier with BCE. Trained on ASVspoof 2019 LA and evaluated on ASV19 eval plus ITW and ASVspoof 2021 DF/LA, Cosine SupCon with a delayed queue achieves the best ITW EER (8.29%) and pooled EER (4.44), while angular similarity performs strongly without queued negatives (ITW 8.70), indicating reduced reliance on large negative sets.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。