Supervisors

Suitable for

Abstract

The first part of this project aims to use advanced graphical models (including enhancements of Hidden Markov Models etc.) to discover personalized trajectories for HIV disease progression, using available electronic health record data. The second part of the project aims to learn how personalized treatment and screening plans can affect disease trajectories in the short run and in the long run, with the overall goal of identifying effective treatment plans for various types of patients. The dataset contains various types of patients and their responses to different medications over time. The project will involve also interacting with a renowned clinician specializing in HIV. In the short run, this work will lead to a publication in an important conference. In the longer run, this work – and more generally, the development of these methods – will change and advance the way medicine is practiced.

To read more about the role of machine learning in medicine – see medianetlab.ee.ucla.edu/MedAdvance

Prerequisites: This project is suitable for someone with at least basic knowledge of neural networks and/or machine learning.