Gaskett, Chris (2003) Reinforcement learning under circumstances beyond its control. In: Proceedings of the international conference on computational intelligence, robotics and autonomous systems. From: Proceedings of the international conference on computational intelligence for modelling control and automation (CIMCA2003), Vienna, Austria.

Preview

PDF Download (967kB)

Abstract

Decision theory addresses the task of choosing an action; it provides robust decision-making criteria that support decision-making under conditions of uncertainty or risk. Decision theory has been applied to produce reinforcement learning algorithms that manage uncertainty in state-transitions. However, performance when there is uncertainty regarding the selection of future actions must also be considered, since reinforcement learning tasks are multiple-step decision problems. This work proposes beta-pessimistic Q-learning—a reinforcement learning algorithm that does not assume complete control.

Item ID:	632
Item Type:	Conference Item (Research - E1)
Keywords:	Reinforcement learning, Beta-pessimistic, Risk, Uncertainty
Date Deposited:	04 Oct 2006
FoR Codes:	01 MATHEMATICAL SCIENCES > 0102 Applied Mathematics > 010205 Financial Mathematics @ 0% 17 PSYCHOLOGY AND COGNITIVE SCIENCES > 1702 Cognitive Science > 170203 Knowledge Representation and Machine Learning @ 0%
Downloads:	Total: 1165 Last 12 Months: 21
	More Statistics