Yuxuan Wang

End-to-end Listen, Look, Speak and Act

ParaS2S: Benchmarking and Aligning Spoken Language Models for Paralinguistic-aware Speech-to-Speech Interaction

DuPO: Enabling Reliable Self-Verification via Dual Preference Optimization

VGPO: Fine-Tuning Speech Autoregressive Diffusion Models with Value Guided Policy Optimization

SplitMeanFlow: Interval Splitting Consistency in Few-Step Generative Modeling

DiSTAR: Diffusion over a Scalable Token Autoregressive Representation for Speech Generation

SALMONN-omni: A Standalone Speech LLM without Codec Injection for Full-duplex Conversation

Sounding that Object: Interactive Object-Aware Image to Audio Generation

DiTAR: Diffusion Transformer Autoregressive Modeling for Speech Generation

Object-Aware Audio-Visual Sound Generation

SALMONN-omni: A Speech Understanding and Generation LLM in a Codec-free Full-duplex Framework

AudioMorphix: Training-free audio editing with diffusion probabilistic models