🤖 Offline RL: Approaching Reinforcement Learning in a data-driven manner

Student: Artur Magalhaes Rodrigues dos Santos

Supervisor: Prof. Dr. Denis Deratani Maua

Abstract

This work consists in an overview on Offline Reinforcement Learning, a research area part of Reinforcement Learning. We go through modern RL approaches, which only use online data, up to full offline RL algorithms. Recently, the field has gained attention due to its usage of static data. We are going to present, discuss and compare the Offline RL algorithms, providing a literature review for future discussions.

📕 Monography: click here.

📄 (outdated - Reduced scope to just cover Offline RL methods) Work proposal: click here

🔤 Keywords

Offline Reinforcement Learning, Reinforcement Learning, Machine Learning, Artificial Intelligence