AI for Medical Group

Overview

Medical data intelligence and computational pathology

The AI for Medical Group studies medical data intelligence and computational pathology, bringing together multimodal clinical data, machine learning and domain knowledge. Working with hospital and industry partners, the group develops methods and systems for disease-specific research and clinical decision support.

Its research connects data integration and annotation with medical foundation models, pathology retrieval and clinical agents, supporting diagnosis, treatment-response prediction and continuous patient care.

Research areas

• Medical data governance and quality. Integrating clinical data, improving annotation and dataset quality, and studying fair, quality-based data sharing. [FQora · WWW Demo 2025]
• Computational pathology. Learning representations of whole-slide images and localizing diagnostically relevant regions through multiple instance learning. [PaM-MIL · CVPR Findings 2026]
• Multimodal medical representation learning. Aligning medical images with pathology, genomics and clinical knowledge to learn robust representations from limited paired data. [KAMP · CVPR 2026]
• Medical image and knowledge retrieval. Retrieving arbitrary pathology regions and relevant biomedical evidence through image indexing, semantic search and medical knowledge graphs. [URICA · CVPR 2026]
• Clinical agents and continuous care. Developing agents for triage, pre-consultation, evidence-based question answering, report interpretation, patient management and follow-up.