The AGI Institute at the School of Computer Science, Shanghai Jiao Tong University (SJTU) , focuses on the core field of Artificial General Intelligence (AGI). It is dedicated to exploring the fundamental theories and key technologies for advancing toward AGI, driving the leap of machine intelligence from specialized domains to human-level generalization capabilities. The Institute aims to build an intelligent system that transcends specific tasks, boasting robust multimodal perception and understanding, autonomous reasoning, and continuous learning abilities. Its goal is to enable machines to deeply interpret human language, complex visual scenes and image semantics, molecular language, and electroencephalographic (EEG) signals (affective cognitive computing), thereby facilitating the realization of human-like general intelligence.

The Institute has established a three-dimensional research system integrating "basic theories + core technologies + interdisciplinary applications," with systematic layout across four core research directions: fundamental AGI large models, computer vision and embodied intelligence, brain-computer interfaces (BCIs) and affective computing, as well as AI interdisciplinary applications in social sciences, natural sciences, and engineering. Leveraging high-performance computing platforms and advanced BCI equipment, it conducts end-to-end innovation spanning intelligent computing theories, algorithmic models, and system implementation, striving to become a world-class research hub.

In terms of industry-academia-research collaboration, the Institute has built a comprehensive innovation and transformation system. It deepens AI+medicine cooperation by relying on clinical resources from Ruijin Hospital Affiliated to SJTU School of Medicine and Shanghai Mental Health Center. Additionally, it takes the lead in establishing university-enterprise platforms such as the "Ruijin Hospital Encephalopathy Center - miHoYo Joint Laboratory" and the "SJTU School of Computer Science - Jiushi Autonomous Driving Technology Joint Laboratory," promoting the transformation and application of AI technologies in smart healthcare, industrial intelligence, and other fields.

For talent cultivation, the Institute adheres to a "theoretical research + systematic practice" model. Through interdisciplinary research training and project-based practice, it nurtures high-caliber interdisciplinary talents with an international perspective. It maintains close cooperation with world-class universities including the University of Oxford, Harvard Medical School, the University of North Carolina at Chapel Hill, and the Hong Kong University of Science and Technology. Through academic exchanges, joint research, and other initiatives, the Institute continuously enhances its influence and voice in the international academic community.

Looking ahead, the Institute will continue to focus on original innovations in AGI’s fundamental theories, closely align with national strategic needs, and strive to become a leading domestic and internationally renowned research institution, providing solid support for China to gain strategic initiative in the development of next-generation artificial intelligence.

1、Fundamental AI Large Models: Hai Zhao, Rui Wang

Natural Language Processing

Artificial Intelligence and Large Models

2、Computer Vision, Embodied Intelligence: Hongtao Lu, Yue Ding

Multimodal large models

Autonomous driving

Diffusion models

3D reconstruction

Multi-task dense prediction

High-, mid-, and low-level computer vision techniques

Efficient and robust Vision-Language-Action (VLA) models

Goal-oriented robotic navigation

3、Brain-Computer Interface, Affective Computing: Baoliang Lu, Weilong Zheng

Brain-Computer Interface

Affective Computing

Brain-Inspired Computing

Computational Neuroscience

Large-Scale EEG Models

4、Interdisciplinary AI Applications (Social Sciences, Natural Sciences, Engineering)

Intelligent Judiciary: Theory and Applications (Jianfeng Xu (Adjunct))

Interdisciplinary Research in AI with Biology, Chemistry, and Medicine, including Bioinformatics, Chemical Large Models, Gene and Protein Large Models, Drug Discovery, and Materials Design (Yang Yang, Shikui Tu, Ruidan Su)

Biomedical Image Processing, Multimodal Medical Foundation Models, and Intelligent Diagnosis (Yi Hong, Jing Ke)