The Web as an Autobiographical Agent

The reward-based autobiographical memory approach has been applied to the Web search agent. The approach is based on the analogy between the Web and the environmental exploration by a robot and has branched off from a currently developed method for autonomous agent learning of novel environments and consolidating the learned information for efficient further use. The paper describes a model of an agent with “autobiographical memories”, inspired by studies on neurobiology of human memory, the experiments of search path categorisation by the model and its application to Web agent design.

• Type: Conference
• Year: 2004
• Where: AIMSA
• Authors: Maya Dimitrova, Emilia I. Barakova, Tino Lourens, Petia Radeva
• Keywords: AIMSA 2004, Autonomous Agent Learning, Reward-based Autobiographical Memory, Web Search Agent