Action selection for composable modular deep reinforcement learning

• Main Authors: GUPTA, Vaibhav, ANAND, Daksh, PARACHURI, Praveen, KUMAR, Akshat
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2021
• Subjects: Reinforcement Learning; Coordination and Control; Deep Learning; Theory and Algorithms
• Online Access: https://ink.library.smu.edu.sg/sis_research/6179; https://ink.library.smu.edu.sg/context/sis_research/article/7182/viewcontent/AAMAS_2021.pdf
• Institution: Singapore Management University

In modular reinforcement learning (MRL), a complex decision making problem is decomposed into multiple simpler subproblems each solved by a separate module. Often, these subproblems have conflicting goals, and incomparable reward scales. A composable decision making architecture requires that even t...