提出双分支门控融合模型，提升语音伪造溯源的准确性与泛化能力。
Dual-Branch Gated Fusion for Open-Set Audio Deepfake Source Tracing
- 用XLSR-53和66维多维度特征CORES并行提取语音合成痕迹。
- 在MLAAD数据集上实现97.6%识别准确率，误报率降低83.5%。
- 门控机制自适应融合特征，适合应对未知合成器的开放集场景。
将合成语音追溯到其生成系统仍是未解难题：封闭集模型无法拒绝未知合成器，且预测过于自信。为此，我们提出双分支门控融合框架，结合XLSR-53与CORES——一种66维描述符，涵盖倒谱、振荡、节奏、能量与频谱等维度，可捕捉互补的合成痕迹。分析表明，XLSR-53在域内（ID）保持判别力，而CORES在分布偏移（OOD）下稳定泛化，但简单拼接因自监督表示不平衡而失效。为此，设计输入感知门控，在联合训练中结合交叉熵、能量边界损失（用于区分ID/OOD）与门控多样性项。在MLAAD基准上，系统达到97.6%的ID准确率，4.9%的EERc，相比Interspeech 2025基线，相对FPR95降低83.5%。
原文摘要 · Abstract (English)
Attributing a synthetic utterance to its originating system remains an open challenge: closed-set models fail to reject unseen synthesizers and produce overconfident predictions. To address this, we propose a dual-branch gated fusion framework that pairs XLSR-53 with CORES, a 66-dimensional descriptor that, unlike prior Linear Filter Bank (LFB)-only work, spans cepstral, oscillatory, rhythmic, energy, and spectral dimensions to capture complementary synthesis artifacts. Our analysis shows XLSR-53 remains discriminative in-domain (ID) while CORES generalizes stably under distribution shift (OOD), yet their naive concatenation fails due to SSL representational imbalance. To resolve this, an input-conditioned gate adaptively weights each branch under joint training with cross-entropy, an energy margin loss for ID/OOD separation, and a gate diversity term. On the MLAAD benchmark, our system achieves 97.6\% ID accuracy, 4.9\% EERc, and an 83.5\% relative FPR95 reduction over the Interspeech 2025 baseline.
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。