ORCID IDBiography
Chief Supervisor
Project Title
Towards a Generalizable Multimodal Tool-Use Agent for Ophthalmology
Synopsis
I propose to investigate the development of a multimodal tool-use agent for ophthalmology that actively integrates diverse imaging modalities such as fundus photography, optical coherence tomography, and angiography. This agent will be designed not only to process and analyze images with high precision but also to interact with external knowledge bases, retrieve pertinent medical information, and dynamically determine the optimal use of various diagnostic tools to support clinical decision-making.
The primary objective is to enhance the accuracy, adaptability, and reliability of ophthalmic diagnosis and decision-making. Leveraging few-shot generalization, cutting-edge model architectures, and retrieval-augmented mechanisms, the agent will be capable of performing complex tasks including disease classification, risk prediction, automated medical report generation, and assisting ophthalmologists in diagnosis and personalized treatment planning. Integrating agentic AI with advanced deep learning techniques in optometry, I aspire to create innovative solutions that autonomously tackle complex challenges and drive transformative impact in both academic research and clinical practice.
