Professor Yang Hongxia, Associate Dean (Global Engagement) of the Faculty of Computer and Mathematical Sciences and Professor in the Department of Computing at The Hong Kong Polytechnic University (PolyU), has announced a major milestone in AI research and healthcare innovation.
Professor Yang's team has developed the project “Enhancing Edge-based Foundation Models for Advanced Reasoning,” which leverages the computing power of Cyberport’s Artificial Intelligence Supercomputing Centre (AISC). The project adopts an innovative approach by integrating high-quality small language models to efficiently train large-scale models. This strategy significantly reduces dependency on centralized computing infrastructure while improving the accuracy of generated information by 28%.
The research has also made substantial progress in medical application scenarios, particularly in the field of cancer treatment. The team is currently working in close collaboration with leading hospitals in Hong Kong and Mainland China, applying vertical large models and domain-specific models in tandem with supercomputing resources to enhance data analysis for oncology.
This AI-driven approach enables more accurate and localized treatment planning, effectively reducing the need for complex diagnostic tests, and thus alleviating the physical and psychological burden on patients. Furthermore, it brings substantial efficiency gains by saving human resources and reducing time costs associated with clinical testing.
This development underscores PolyU’s commitment to interdisciplinary AI innovation with tangible societal impact, particularly in healthcare. The initiative exemplifies how edge AI and federated modeling can help democratize access to powerful AI tools while addressing real-world challenges.
