5 papers
Repairing Speech Access After Text Updates with Frozen Affine Fields
Ming-Hao Hsu, Yuxiang Wang, Yingda Shen +1
Speech-capable language models inherit most of their knowledge from text-rich language backbones, and continued model maintenance is likewise dominated by text updates. Knowledge a…
Long-Prompt Zero-Shot TTS: Benchmarking Prosody Reproduction Beyond Voice Cloning
Dekun Chen, Renjie Xu, Yingda Shen +2
Zero-shot text-to-speech (TTS) typically conditions on a short speech prompt, which conveys timbre but provides only limited evidence of a speaker's prosodic patterns. We formulate…
What Matters in Preference Post-Training of Continuous Autoregressive Speech Language Models
Yingda Shen, Yao Qian, Dekun Chen +6
Continuous autoregressive speech language models combine an autoregressive (AR) backbone with a diffusion or flow-matching head. Preference post-training requires choosing how to c…
AuraSE: Low-Hallucination Generative Speech Enhancement via Multimodal Flow Matching and Inference Policy Optimization
Yingda Shen, Yao Qian, Yuxuan Hu +7
Generative speech enhancement models can produce cleaner and more natural-sounding speech than conventional discriminative approaches, but may hallucinate by changing speech conten…
Interpreting and Evaluating Dynamic-Rate Speech Codec Boundaries
Han Wang, Jiaqi Li, Yingda Shen +2
Dynamic-frame-rate neural speech codecs replace a uniform frame grid with variable-duration tokens, making boundary placement part of the representation itself. Yet it is unclear w…