🔭 I’m currently working on speech and natural language processing, especially large-scale pre-trained models.

🎓 I obtained my Ph.D. degree at Beihang University, China. Now, I am a senior researcher at Microsoft Research Asia.

📫 How to reach me: Wu.Yu at microsoft.com

📄 Here are my selected publications:

• Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers
  • Chengyi Wang, Sanyuan Chen, Yu Wu (Corresponding author) , Ziqiang Zhang, Long Zhou, Shujie Liu, Zhuo Chen, Yanqing Liu, Huaming Wang, Jinyu Li, Lei He, Sheng Zhao, Furu Wei.
  • A language model based TTS system, which could clone your voice with a 3-second recording.
  • Demo and Paper
  • VALL-E X a cross-lingual version VALL-E that can help anyone speak a foreign language in their own voice.
• WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing
  • Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu (Corresponding author), Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Xiangzhan Yu, Furu Wei.
  • [Accepted in J-STSP in June 2022] [code] [demo]
  • Ranks 1st in the SUPERB leaderboard and SLT2022 SUPERB Challenge.
  • Ranks 1st on VoxSRC 2021 speaker verification permanent leaderboard.
  • Integrate into official torchaudio
• Response Generation by Context-aware Prototype Editing
  • Yu Wu, Furu Wei, Shaohan Huang, Yunli Wang, Zhoujun Li, Ming Zhou.
  • [Accepted in AAAI 2019] [code]
  • The first paper studies prototype based response generation.
• Sequential Matching Network: A New Architecture for Multi-turn Response Selection in Retrieval-Based Chatbots
  • Yu Wu, Wei Wu, Chen Xing, Ming Zhou, Zhoujun Li.
  • [Accepted in ACL 2017] [code]
  • The first paper studies multi-turn response selection.