Sun Yat-sen University
- Chinese name
- 中山大学
- Pinyin
- Unified Social Credit Code
121000004558631445- Website
Sun Yat-sen University (中山大学), the major Guangzhou research university founded by Sun Yat-sen in 1924, builds digital humans as working systems as well as research subjects. Since 2022 its School of Software Engineering’s “digital village” practice team has run agriculture-assistance livestreams in Fengyang, Qingyuan (Guangdong) fronted by a self-built digital-human anchor, Huang Xiaowei (黄小伟) — modeled on a real team member off the Fay open-source digital-human stack and driven by a large language model, text-to-speech, retrieval-augmented generation and voice-driven lip-sync, with a 2.0 version in July 2024 adding automatic live-comment answering and voice trained on the person’s own audio. A sibling student project built a campus-introducer digital human for the university’s 2024 centennial.
The research side feeds the industry directly. Digital humans are a standing faculty research direction — the School of Intelligent Systems Engineering’s 黎汉汇 holds a National Natural Science Foundation project on visual-media feature learning for high-fidelity digital-human generation — and the university’s recent joint output spans the TalkVid audio-driven digital-human synthesis dataset, open-sourced with CUHK-Shenzhen and HKUST in August 2025; the MMTryon virtual try-on framework with ByteDance’s digital-human team; and the Moral RolePlay evaluation benchmark with the Tencent Hunyuan Digital Human Team (November 2025), the first systematic test of large models playing roles across moral registers. On the patent side it holds an authorized face-video generation patent aimed at improving lip-sync accuracy between virtual-human mouth movements and speech, and co-filed CN121921423A with Zhongyou Shuzhi (Xi’an) Technology (中邮数智(西安)科技) for real-time, streaming, unlimited-duration interactive avatar video.