arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.12403cs.CLcs.AI
收藏

偏好对齐改进基于语言模型的文本到语音合成

Preference Alignment Improves Language Model-Based TTS

Language Technologies Institute, Carnegie Mellon University(卡内基梅隆大学语言技术研究所);Tencent AI LAB(腾讯AI实验室)

Jinchuan Tian, Chunlei Zhang, Jiatong Shi, Hao Zhang, Jianwei Yu, Shinji Watanabe, Dong Yu

AI总结:

本研究通过实证评估证明,偏好对齐算法(特别是DPO)能显著提升基于语言模型的TTS在可懂度、说话人相似度及主观评分上的表现,并适用于低资源与域外场景。

摘要:

文本到语音(TTS)领域的最新进展表明,基于语言模型(LM)的系统在性能上可与同类系统相媲美。通过偏好对齐算法可以进一步优化,这些算法调整语言模型以符合奖励模型的偏好,从而提升生成内容的合意性。本研究对偏好对齐算法(尤其是直接偏好优化(DPO))如何增强基于LM的TTS进行了全面的实证评估。使用一个1.15B参数规模的基于LM的TTS模型,我们证明偏好对齐持续改善了可懂度、说话人相似度和代理主观评价分数,其中后两项指标在某些评估中甚至超过了人类语音。我们还表明,偏好对齐适用于低资源场景,并能有效泛化到域外应用。

英文摘要:

Recent advancements in text-to-speech (TTS) have shown that language model (LM)-based systems offer competitive performance to their counterparts. Further optimization can be achieved through preference alignment algorithms, which adjust LMs to align with the preferences of reward models, enhancing the desirability of the generated content. This study presents a thorough empirical evaluation of how preference alignment algorithms, particularly Direct Preference Optimization (DPO), enhance LM-based TTS. With a 1.15B parameter LM-based TTS model, we demonstrate that preference alignment consistently improves intelligibility, speaker similarity, and proxy subjective evaluation scores, with the latter two metrics surpassing even human speech in certain evaluations. We also show preference alignment is applicable to low-resource scenarios and effectively generalized to out-of-domain applications.

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。

↑