Multimodal Representation Learning for Healthcare Data

The abundance of electronic health record (EHR) data has accelerated the adoption of data-driven methods to automate various tasks ranging from patient care to resource management in hospitals around the world. The use of specific types of data such as X-ray images, doctor’s note and others have been used individually to implement machine learning models suited for a specific task. However, such single mode of data is not able to provide an overall picture of the patient health. In this project, we aim to develop an end-to-end machine learning model that can learn a combined representation of all these different types of data which can be fine-tuned for multiple downstream tasks such as treatment outcome prediction, hospitalization and length of stay prediction and many more.

Faculty Supervisor:

Rahul G. Krishnan

Student:

Partner:

Layer 6 AI

Discipline:

Computer science

Sector:

Artificial Intelligence; Health and Related Sciences & Technology

University:

University of Toronto

Program: