期刊名称：Revista Brasileira de Informática na Educação
印刷版ISSN：1414-5685
出版年度：2004
卷号：12
期号：2
页码：69-77
语种：Portuguese
出版社：Revista Brasileira de Informática na Educação
摘要：The goal of this paper is to present a diagnostic module included in an Intelligent Tutoring System (ITS) architeture. In this module, a Reinforcement Learning technique (Q-Learning algorithm) is applied, making it possible to autonomous modelling of the learner. An utility value is calculated based on a state-action table upon which the algorithm estimates future rewards which represent the cognitive states of the learner. The best action policy to be used by the tutor at any cognitive state is then learned and made available by the Reinforcement Learning algorithm.
其他摘要：The goal of this paper is to present a diagnostic module included in an Intelligent Tutoring System (ITS) architeture. In this module, a Reinforcement Learning technique (Q-Learning algorithm) is applied, making it possible to autonomous modelling of the learner. An utility value is calculated based on a state-action table upon which the algorithm estimates future rewards which represent the cognitive states of the learner. The best action policy to be used by the tutor at any cognitive state is then learned and made available by the Reinforcement Learning algorithm.
关键词：Intelligent Tutoring Systems;Reinforcement learning;Q-Learning algorithm;Sistemas Tutores Inteligentes;Aprendizado por Reforço;Algoritmo Q-Learning
