提交给2026年环境感知语音和声音深度伪造检测挑战赛（ESDD2）的两篇研究论文提出了新颖的深度学习框架，用于检测经过篡操纵的音频。第一篇论文介绍了一个双分支系统，使用预训练模型XLS-R和BEATs分别分析语音和环境声音，达到了70.20%的F1分数。第二篇论文探讨了各种深度学习架构和预训练模型，发现使用三阶段策略对WavLM进行微调可获得更优异的结果，在一个基准数据集上取得了0.95的F1分数。 AI
影响 深度伪造音频检测的进步可能带来更强大的内容审核和安全系统。
排序理由 两篇arXiv论文提出了新的深度伪造音频检测方法，包括具体的模型架构和性能指标。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2605.06035v1 Announce Type: cross Abstract: Quantum machine learning has emerged as a promising tool for pattern recognition, yet many audio-focused approaches still treat spectrograms as generic images and do not explicitly exploit their time-frequency structure. We propos…
arXiv:2605.03420v1 Announce Type: cross Abstract: This paper describes a submission to the Environment-Aware Speech and Sound Deepfake Detection Challenge (ESDD2) 2026, which addresses component-level deepfake detection using the CompSpoofV2 dataset, where speech and environmenta…
arXiv:2604.19652v2 Announce Type: replace-cross Abstract: In this paper, we propose a deep-learning framework for environmental sound deepfake detection (ESDD) -- the task of identifying whether the sound scene and sound event in an input audio recording is fake or not. To this e…