Descrizione

Titolo del progetto di ricerca: "Privacy-preserving Explainable AI for Health - HealthExplAI"

Descrizione dell'attività da svolgere dal/la collaboratore/trice:
The successful completion of the project HealthExplAI of Dr. Panagiotis Symeonidis (IN 200H) require additional skills, which are expected to be covered by an occasional collaborator, who is needed only for a specific time period of 4 months, and will be assigned with the following tasks:
- The occasional collaborator will write code in Python for the implementation of a notebook in Python Jupyter with the task to build an intelligent recommendation agent for personalized health along with explanations:
- The occasional collaborator will build a drug recommendation agent that provides explanations to the medical doctors based on meta paths over heterogeneous graphs.
- The occasional collaborator will develop on- and off-policy reinforcement learning methods, which are needed to capture the short-term changes of the recommendation agent behavior.
- The occasional collaborator will extract similarities of diseases, drugs and their side effects for patients based on their daily drug treatment of a specific time period only. Based on these similarities/relevancies, the occasional collaborator will code a recommendation agent to recommend the suitable drug therapeutic for each patient.

Responsabile: dr. Panagiotis Symeonidis