A Novel Model for Performance Optimization of Reinforcement Learning (RL) Agent in Stochastic Clinical Scenarios

Principal Investigator

Name
Madeha Arif

Degrees
Ph.D

Institution
National University of Sciences and Technology, Islamabad Pakistan

Position Title
Student

Email
madeha_746@yahoo.com

About this CDAS Project

Study
PLCO (Learn more about this study)

Project ID
PLCO-837

Initial CDAS Request Approval
Sep 27, 2021

Title
A Novel Model for Performance Optimization of Reinforcement Learning (RL) Agent in Stochastic Clinical Scenarios

Summary
To propose an optimized and generic Deep Reinforcement Learning Model (DRL) that will have dynamic adaptive reward function for stochastic clinical environment and will intelligently cover maximum dynamic behaviors of patients for treating cancer disease.

Aims

In this research we will address:
• Stochastic nature of chronic diseases (handling noisy or incomplete states).
• Credit assignment problem.
• Cost of exploitation and exploration.
• Misleading reward function.

Collaborators

Usman Qamar (National University of sciences and Technology )