计算机学院系列讲座名家讲坛第33期——Empowering Personalized Healthcare with Multi-Modal AI and Sensing Systems

报告题目(Title)：Empowering Personalized Healthcare with Multi-Modal AI and Sensing Systems

时间(Date & Time)：2026.7.3; 10:00 -11:30

腾讯会议：378 875 708

主讲人(Speaker)：Guoliang Xing（邢国良）

邀请人(Host)：Daqing Zhang（张大庆）

报告摘要(Abstract)：

Aging populations and chronic conditions have emerged as a major global health challenge. Yet current diagnostic and monitoring solutions remain largely confined to clinical settings, limiting both their accessibility and their capacity for continuous care. AI and smart sensors offer a transformative paradigm—one that supports aging in place and enables the proactive management of chronic conditions within natural living environments.

In this talk, we present a comprehensive suite of AI-driven systems designed to bridge this gap. We will first introduce ADMarker, which detects digital biomarkers of Alzheimer's disease through multimodal sensors and large language models (LLMs) and is currently being validated in a five-year clinical trial involving 1,500 participants. We will then turn to two systems that have been commercialized and deployed to thousands of users: Acure Care, which integrates radar and infrared sensing for privacy-preserving fall detection in the elderly, and Nuna, the first smart pendant that leverages on-device LLMs and multimodal sensors to generate life journals capturing daily emotions, events, and social activities.

The key enabling technology behind these personalized healthcare systems is human action recognition, understanding, and reasoning across multiple modalities—a challenge that has remained open for decades, due largely to the lack of high-quality, scalable datasets. To address this, we introduce CUHK-X, a large-scale multimodal dataset and benchmark suite comprising 64K samples in 7 different sensor modalities that span 40 actions performed by 30 participants across two indoor environments. Building on CUHK-X, we have launched an international competition to encourage the community to apply and develop data-intensive learning methods for robust, multimodal human activity analysis.

主讲人简介(Bio)：

Guoliang Xing received his B.S. and M.S. degrees from Xi'an Jiao Tong University, China, in 1998 and 2001, respectively, and his D.Sc. degree from Washington University in St. Louis in 2006. He is currently a Professor in the Department of Information Engineering at The Chinese University of Hong Kong (CUHK). Prior to joining CUHK, he served on the faculty of Michigan State University, USA, from 2008 to 2017.

Prof. Xing's research lies at the intersection of systems and Embedded AI, with transformative applications in healthcare and autonomous driving. He is the recipient of numerous prestigious honors, including the US NSF CAREER Award (2010), the Withrow Rising Scholar Award from Michigan State University (2014), the Research Excellence Award from CUHK (2024), and the Hong Kong Research Grants Council (RGC) Senior Research Fellowship, which is granted to only 10 recipients annually across all disciplines. His work has received 6 Best Paper Awards, 5 Best Demo/Poster/Artifact Awards, and 7 Best Paper Finalist distinctions at top-tier international conferences, including MobiCom, MobiSys, SenSys, ICNP, and IPSN. He is a Fellow of both the ACM and the IEEE.

欢迎关注计算机学院微信公众号，了解更多讲座信息！

北京大学计算机学院