ActiveTutor

In this paper we present an architecture dedicated to an intelligently assisted educational tool which integrates within a unified framework software rational agents both at the microscopic level and macroscopic level, in conjunction with an Embodied Conversational Agent (ECA) which communicates at the same time with the agents but also with the user. This architecture is based on the concept of active representation, containing rational micro-agents which control the atomic elements of the tutorials and which are able to interact with the student in two different and complementary ways: initially while displaying themselves dynamically in the interface (and this in a ‘smart” manner) and then by using the Embodied Conversational Agent to express what a human pedagogue would wish to express in his task of learning support at the time of the sessions. The conversational agent is based on the LEA (LIMSI Embodied Agent) technology developed at the LIMSI-CNRS within the framework of EU...

• Type: Conference
• Year: 2005
• Where: ICALT
• Authors: Jean Pierre Fournier
• Keywords: Embodied Conversational Agents, ICALT 2005, LIMSI Embodied Agent, Machine Learning, Software Rational Agents