Liang Hu
Professor and Doctoral Director, School of Computer Science and Technology, Tongji University
Liang Hu is a professor and doctoral director of School of Computer Science and Technology, Tongji University. He was selected as one of the National Overseas High-Level Young Talents and Shanghai Overseas High-Level Talents. He received his PhD degree in Computer Application Technology from Shanghai Jiaotong University and his PhD degree in Analytics from the University of Technology Sydney (UTS), Australia, respectively. He has presided over and participated in a number of national research projects. His research interests include artificial intelligence, recommender systems, data science, privacy computing, big models, and cross-disciplinary integration of next-generation intelligence technologies. He has published more than 100 papers, many of which are groundbreaking in the field. He serves as a program committee member of many high-level international conferences on AI (IJCAI, AAAI, ICDM, CIKM, KDD, CVPR, ICML, NeurIPS, etc.), and a reviewer of more than ten famous international journals, including ACM CSUR, IEEE TKDE, ACM TOIS, IEEE TPAMI, etc. He has also served as a member of the program committee of the International Conference on Artificial Intelligence (IJCAI), and a reviewer of more than ten famous international journals. He hosted two IEEE ICAACE 2023/ICAACE 2024 international conferences in 2023 and 2024, and served as the conference chair. In addition, Prof. Liang Hu serves as an executive member of the Special Committee on Granular Computing and Knowledge Discovery of the Chinese Society for Artificial Intelligence, and an executive member of the Special Committee on Collaborative Computing of the Chinese Computer Society. He is also the visiting professor of China Ship Nine Academy, the chief expert of the expert workstation of Shenlan Artificial Intelligence (Shanghai) Co., Ltd. and the chief expert of the expert workstation of Shanghai 263 Communication Co., Ltd. He actively promotes the process of combining industry, academia and research and the proposed methodology has been transformed and landed to be applied in the actual projects in different fields such as e-commerce, healthcare, taxation, finance and transportation and solved some of the core problems.
Topic
A multimodal grand model of brain-like cognition: a foundation for connecting human brains, AI brains, and robot brains
In the current era of big models of artificial intelligence, the development of big language models represented by ChatGPT is booming, and the cross-modal generative technology represented by MidJourney and SORA has also become a hot spot. This year, the emergence of domestic big models represented by DeepSeek, and the spring debut of embodied intelligent robots represented by Yushu, all mark the rapid rise of the level of new China's artificial intelligence technology to the forefront of the world. Inspired by neuroscience, this report explores the establishment of a multimodal big model with brain-like cognitive ability, which can enable the seamless connection of human brain, AI brain and robot brain, thus bringing a brand-new mode of human-computer interaction and human-computer collaboration. This paper will introduce neuroscience-inspired continuous machine learning and forgetting mechanisms, as well as mutual decoding between human brain information encoding (e.g., fMRI) and generative AI models, which provide new paradigms for the construction of multimodal macromodels of brain-like cognition and their applications.
