arXiv 2603.029142026-03-03eess.AS

基于强化学习的微调是否能提高二元语音深度伪造检测的泛化能力？

Does Fine-tuning by Reinforcement Learning Improve Generalization in Binary Speech Deepfake Detection?

Xin Wang, Ge Wanying, Junichi Yamagishi

备注：: Submitted to Interspeech 2026; put on arxiv based on requirement of paper open-access rule; quote from Interspeech: "Interspeech no longer enforces an anonymity period for submissions. While uploading a version online is permitted, your official submission to Interspeech must not contain any author-identifying information"

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文研究了基于GRPO的强化学习微调对二元语音深度伪造检测泛化能力的提升作用，发现其在无领域测试集上表现优于传统SFT方法。

摘要：

构建能够泛化到未见过攻击的语音深度伪造检测模型仍然是一个具有挑战性的问题。尽管该领域已转向使用语音基础模型的预训练和微调范式，但大多数方法仅依赖于有监督微调（SFT）。受大型语言模型领域的启发，其中强化学习（RL）用于模型微调，我们研究了RL，特别是群相对策略优化（GRPO）的影响。通过使用多个检测器和测试集的实验结果表明，纯基于GRPO的微调在无领域测试集上提高了性能，同时在目标领域测试数据上保持性能。这种方法优于仅SFT和混合设置。我们的消融研究进一步表明，GRPO中的负奖励可能是这一改进的关键因素。

英文摘要：

Building speech deepfake detection models that are generalizable to unseen attacks remains a challenging problem. Although the field has shifted toward a pre-training and fine-tuning paradigm using speech foundation models, most approaches rely solely on supervised fine-tuning (SFT). Inspired by the field of large language models, wherein reinforcement learning (RL) is used for model fine-tuning, we investigate the impact of RL, specifically Group Relative Policy Optimization (GRPO). The results from experiments using multiple detectors and test sets indicate that pure GRPO-based fine-tuning improves performance on out-of-domain test sets while maintaining performance on target-domain test data. This approach outperforms both SFT-only and hybrid setups. Our ablation studies further suggest that the negative reward in GRPO may be a key factor in this improvement.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。