Inverse Reinforcement Learning for Healthcare Applications: A Survey

Mohamed-Amine Chadi, Hajar Mousannif

2021

Abstract

Reinforcement learning (RL) is a category of algorithms in machine learning that deals mainly with learning optimal sequential decision-making. And because the medical treatment process can be represented as a series of interactions between doctors and patients, RL offers promising techniques for solving complex problems in healthcare domains. However, to ensure a good performance of such applications, a reward function should be explicitly provided beforehand, which can be either too expensive to obtain, unavailable, or nonrepresentative enough of the real-world situation. Inverse reinforcement learning (IRL) is the problem of deriving the reward function of an agent, given its history of behaviour or policy. In this survey, we will discuss the theoretical foundations of IRL techniques and the problem it solves. Then, we will provide the state-of-the-art of current applications of IRL in healthcare specifically. Following that, we will summarize the challenges and what makes IRL in healthcare domains so limited despite its progress in other research areas. Finally, we shall suggest some prospective study directions for the future.

Paper Citation

in Harvard Style

Chadi M. and Mousannif H. (2021). Inverse Reinforcement Learning for Healthcare Applications: A Survey. In Proceedings of the 2nd International Conference on Big Data, Modelling and Machine Learning - Volume 1: BML, ISBN 978-989-758-559-3, pages 97-102. DOI: 10.5220/0010729200003101

in Bibtex Style

@conference{bml21,
author={Mohamed-Amine Chadi and Hajar Mousannif},
title={Inverse Reinforcement Learning for Healthcare Applications: A Survey},
booktitle={Proceedings of the 2nd International Conference on Big Data, Modelling and Machine Learning - Volume 1: BML,},
year={2021},
pages={97-102},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010729200003101},
isbn={978-989-758-559-3},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Conference on Big Data, Modelling and Machine Learning - Volume 1: BML,
TI - Inverse Reinforcement Learning for Healthcare Applications: A Survey
SN - 978-989-758-559-3
AU - Chadi M.
AU - Mousannif H.
PY - 2021
SP - 97
EP - 102
DO - 10.5220/0010729200003101