A core task abstraction approach to hierarchical reinforcement learning [Extended abstract]

• Main Authors: LI, Zhuoru, NARAYAN, Akshay, Tze-Yun LEONG
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2016
• Subjects: hierarchical reinforcement learning; core task abstraction; Computer Sciences; Numerical Analysis and Scientific Computing; Theory and Algorithms
• Online Access: https://ink.library.smu.edu.sg/sis_research/3431; https://ink.library.smu.edu.sg/context/sis_research/article/4432/viewcontent/Acoretaskabstractionapproachto.pdf
• Institution: Singapore Management University

We propose a new, core task abstraction (CTA) approach to learning the relevant transition functions in model-based hierarchical reinforcement learning. CTA exploits contextual independences of the state variables conditional on the task-specific actions; its promising performance is demonstrated th...