Kai Yu

2 articles on SOTA Papers

UltraVoice Improves Fine-Grained Control in Spoken Dialogue Models

A curated 830-hour dialogue dataset couples style-conditioned instructions with synthesized speech, raising multi-dimensional control scores without reported losses in conversational capability.

Sep 6, 20264 min2510.22588 Code available

GigaSpeechBench Exposes ASR Gaps Beyond Standard Benchmarks

A 680-hour human-annotated benchmark combines multilingual, dialect, accent, domain, and age stress tests across ASR and speech translation.

Jul 28, 20265 min2606.28884