Research Focus
Associate Professor • CUHK-Shenzhen
Zhizheng Wu is Associate Professor at The Chinese University of Hong Kong, Shenzhen, Jointly Appointed Professor at Shenzhen Loop Area Institute, Director of the Artificial Intelligence and Robotics Taught Post-Graduate Program, and Deputy Director of the Shenzhen Key Laboratory of Cross-Modal Cognitive Computing.
Open Science
Projects including Merlin, Amphion, and Emilia adopted by more than 1,000 organizations.
Recognition
National-level Young Talent, Stanford Top 2% Scientist, multiple best paper awards.
Biography
Academic profile
Professor Wu has been selected as a National-level Young Talent and has been consecutively listed in Stanford University's "World's Top 2% Scientists." He has received multiple Best Paper Awards and has held research and technical leadership roles at Meta, Apple, JD.com, the University of Edinburgh, and Microsoft Research Asia.
He initiated several influential open-source efforts, including Merlin, Amphion, and Emilia. Amphion has repeatedly appeared at the top of GitHub Trending, and Emilia became one of the most liked audio datasets on HuggingFace. His work spans speech synthesis, voice conversion, speech restoration, speaker security, and speech generation at scale.
Latest Updates
Selected news
Best Poster Award at 2025年声纹处理研究与应用学术研讨会
Yuancheng Wang received the Best Poster Award.
GenSR-Pref accepted to AAAI 2026
GenSR-Pref is accepted to AAAI 2026. Congrats to Junan Zhang.
Invited talk at Huawei Media Tech Summit
Research progress in Speech Tokenize/Codec at 华为2025媒体技术峰会.
Invited talk at RTE Conference
Research progress in speech processing technologies at RTE大会.
Publications
Selected research output
MaskGCT: Zero-Shot Text-to-Speech with Masked Generative Codec Transformer
A novel zero-shot TTS model utilizing masked generative codec transformers.
Metis: A Foundation Speech Generation Model with Masked Generative Pre-training
Foundational model for speech generation leveraging masked generative pre-training techniques.
AnyEnhance: A Unified Generative Model with Prompt-Guidance and Self-Critic for Voice Enhancement
A unified generative approach for voice enhancement using prompt guidance.
Emilia: A Large-Scale, Extensive, Multilingual, and Diverse Dataset for Speech Generation
High-quality training data from under-explored in-the-wild sources that capture spontaneous human speech in real-world contexts.
ASVspoof: the Automatic Speaker Verification Spoofing and Countermeasures Challenge
A benchmark challenge paper defining the standards for spoofing detection in speaker verification.
Professional Path
Work history
Associate Professor
The Chinese University of Hong Kong, Shenzhen
Founding member / Advisor
Sanas
Tech Lead / Research Scientist
Meta Platforms Inc, USA
Engineering Director / Research Scientist
JD.COM Silicon Valley Research Center, USA
Research Scientist
Apple Inc, USA
Research Fellow
University of Edinburgh, UK
Research Intern
Microsoft Research Asia
Team Culture
Open science, ambitious execution, and useful AI.
Values: be a leader not a follower, be bold and fight for excellence, seek expertise and experience. The group actively collaborates with academic and industry partners to push AI systems into practical impact.
Research Group
Current members
Ruixing Jin
PhD Student
Aug 2025 - PresentYingda Shen
MPhil Student
Aug 2025 - PresentQinke Ni
MPhil Student
Aug 2025 - PresentChong Jing
MPhil Student
Aug 2025 - PresentMinghao Xu
PhD Student
Aug 2025 - PresentHuan Liao
PhD Student
Aug 2025 - PresentYuxiang Wang
PhD Student
Aug 2025 - PresentYudong Li
PhD Student
Aug 2025 - PresentJiaqi Li
PhD Student
Aug 2024 - PresentJun'an Zhang
PhD Student
Aug 2024 - PresentDekun Chen
PhD Student
Aug 2024 - PresentZihao Fang
MPhil Student
Aug 2024 - PresentYuancheng Wang
PhD Student
Aug 2023 - PresentLi Wang
PhD Student
Aug 2023 - PresentXueyao Zhang
PhD Student
Aug 2022 - PresentAlumni
Past members
XXX
Alumni (XXX)
Jun 2020 - Sep 2020