Learning and cooperating in information gathering

With the explosive increase of available resources, the Internet has become the main platform for information promulgating and communicating. There are many complicated characteristics of the Internet, so that the effectiveness and efficiency of current information retrieval systems is not satisfying. Internet information retrieval is one of the most active areas of computer science. Software agents have been investigated widely in recent years and agent technology has been applied in some intelligent systems. We introduce a distributed cooperative information retrieval multi-agent system, ICMAS. The system learns to understand user's information demand accurately through interactions with the user. The learning agent can improve the system's ability by acquiring new knowledge and sharing knowledge with other individuals.