Abstract:
A virtual agent may obviously benefit from having an up-to-date model of her environment of activity. The model may include actual users' profiles, a dynamic environment characteristic or some assumptions being accepted by default. However, the agent doesn't have to stick to one model only, she can possess a set of complementary beliefs, both learned and assumed. A hierarchy of beliefs for an agent is proposed in this paper, together with a decision making scheme and some experimental results to support the idea. The experiments consisted of the agent's interactions with some simulated 0-level agents, acting as customers of an imaginary Internet banking service.

Keywords: autonomous agents, user modeling, machine learning, decision making, beliefs.

Computational Intelligence Group @ Technical University of Clausthal
Human Media Interaction Group @ University of Twente
Computer Science Group @ University of Gdansk	Last modified 2002-04-03