Approved Research

Develop multi-modality medical QA (Question Answering) systems

Lay summary

Our study aims to enhance the precision, adaptability, and usefulness of artificial intelligence models in the medical field. We plan to develop a comprehensive multimodal foundation model infused with medical-specific knowledge to create a general Medical Question Answering system. This system will be capable of quickly answering medical questions based on user's input in any format, such as text or images, facilitating convenient access to top-notch medical resources for healthcare centers and online services. By utilizing advanced deep learning techniques and leveraging data from the biobank, we aim to improve public health by providing faster and more accessible medical services, including online diagnosis, treatment advice, and medication recommendations. This ambitious project is expected to span 36 months.