Direct Preference Optimization for Speech Autoregressive Diffusion Models
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866916964467736576 |
|---|---|
| author | Liu, Zhijun Jia, Dongya Wang, Xiaoqiang Du, Chenpeng Wang, Shuai Chen, Zhuo Li, Haizhou |
| author_facet | Liu, Zhijun Jia, Dongya Wang, Xiaoqiang Du, Chenpeng Wang, Shuai Chen, Zhuo Li, Haizhou |
| contents | Autoregressive diffusion models (ARDMs) have recently been applied to speech generation, achieving state-of-the-art (SOTA) performance in zero-shot text-to-speech. By autoregressively generating continuous speech tokens with next-token diffusion, these models offer a promising alternative to next-token prediction, avoiding the technical complexities associated with discrete speech tokenization. As a relatively new paradigm, research on reinforcement learning (RL)-based fine-tuning of speech ARDMs remains limited. In this paper, we propose Autoregressive Diffusion-Direct Preference Optimization (ARDM-DPO) to advance this research. By fine-tuning the recently proposed zero-shot text-to-speech model DiTAR with DPO, we achieve significant improvements in terms of speech expressiveness and robustness for long texts. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_18928 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Direct Preference Optimization for Speech Autoregressive Diffusion Models Liu, Zhijun Jia, Dongya Wang, Xiaoqiang Du, Chenpeng Wang, Shuai Chen, Zhuo Li, Haizhou Audio and Speech Processing Autoregressive diffusion models (ARDMs) have recently been applied to speech generation, achieving state-of-the-art (SOTA) performance in zero-shot text-to-speech. By autoregressively generating continuous speech tokens with next-token diffusion, these models offer a promising alternative to next-token prediction, avoiding the technical complexities associated with discrete speech tokenization. As a relatively new paradigm, research on reinforcement learning (RL)-based fine-tuning of speech ARDMs remains limited. In this paper, we propose Autoregressive Diffusion-Direct Preference Optimization (ARDM-DPO) to advance this research. By fine-tuning the recently proposed zero-shot text-to-speech model DiTAR with DPO, we achieve significant improvements in terms of speech expressiveness and robustness for long texts. |
| title | Direct Preference Optimization for Speech Autoregressive Diffusion Models |
| topic | Audio and Speech Processing |
| url | https://arxiv.org/abs/2509.18928 |