Wang Guangrun

Faculty Profile

Guangrun Wang is a Young Researcher, Associate Professor, and Ph.D. Supervisor. He is a recipient of a national-level young talent fund, the National Excellent Young Scientist Fund (Overseas), and the Huawei Strategic Research Institute Talent Fund.

Before returning to China, he served as a Research Fellow at the University of Oxford. His collaborating advisor was Professor Philip H. S. Torr, a Fellow of the Royal Society (the Royal Academy of Sciences, of which Isaac Newton once served as president), a Fellow of the Royal Academy of Engineering, and a Fellow of the Alan Turing Institute. During his time in the UK, he also concurrently served as a project lead for a research team at Aistetic (UK) and maintained close collaborations with leading international technology companies. He earned dual bachelor’s degrees in Engineering and Management from the former School of Information Science and Technology and the School of Management at Sun Yat-sen University, respectively, and obtained his Ph.D. degree from the School of Computer Science at Sun Yat-sen University (including a research visit to the Department of Information Engineering at The Chinese University of Hong Kong), under the supervision of Professor Liang Lin. He has also conducted research at Dark Matter Intelligent Technology Co., Ltd. He has published more than 50 papers in top-tier (A-class) conferences and CAS Zone-1 journals, including 12 oral presentations at major international conferences, which typically have very low acceptance rates.

He was selected for the highest tier of Huawei’s “Genius Youth” Program; received the Wu Wenjun Award for Outstanding Doctoral Dissertation in Artificial Intelligence (only nine recipients nationwide that year); won the Pattern Recognition Best Paper Award (only one paper globally that year); was recognized as an IJCAI Distinguished Senior Program Committee Member (only 18 worldwide that year); and was listed in the Global AI Chinese Rising Stars (only 25 worldwide in machine learning that year). He has earned top honors in multiple international competitions, winning gold awards among thousands of participating teams. He has served as an Area Chair for ICLR and IJCAI. He has been recognized six times as an outstanding/top reviewer for CCF A-class conferences (IJCAI 2021, NeurIPS 2022, ICLR 2022, ICLR 2021, ICCV 2021, NeurIPS 2021), and his papers have been cited and used as supporting evidence by researchers including Yann LeCun. He co-organized workshops at ICML 2024 and CVPR 2023, co-supervised multiple students from the University of Oxford and Sun Yat-sen University, and his research outcomes have been deployed in multiple companies.

Research Areas

His primary research focuses on next-generation AI architectures (see: https://thegreatailab.github.io/).

The motto of Sun Yat-sen University—“Extensive Learning, Careful Inquiry, Thorough Reflection, Clear Discrimination, and Earnest Practice”—is, in a certain sense, highly aligned with the core vision of artificial general intelligence. In particular, “earnest practice” emphasizes the transition from cognition to action, which corresponds to the action capability in Physical Intelligence (π, Pi), namely the ability to carry out effective physical actions in the real world.

For example, when you leave home in the morning the house may be messy, yet when you return at night it is clean and orderly, and you cannot tell whether the cleaning was done by a human or by an AI (adapted from The Physical Turing Test: Jim Fan on NVIDIA’s Roadmap for Embodied AI). This provides an intuitive manifestation of artificial general intelligence in the physical world.

Keywords / Topics

• Next-generation AI architectures
• Large Physical Models (encompassing multimodal large models and large language models)
• Multimodal generative AI: 2D/3D/4D visual modeling; language modeling; sequence modeling
• Natural sciences