Personal Infromation Agent can provide interesting URL to the user. Based on Multi-Agentscollaboration,we can improve the recommendation quality. The Web's interesting change may be valu-able to the user. If a user visits the unchanged URL,it can be burdensome and wasteful. The Notifica-tion Agent can notify that some modification occurred. But the user may not want to know when the au-thor changed the font. We like an Agent that notices when interesting changes occurred. The informa-tioin Agent can use machine learning to detect interesting changes of URLs. According to the user'sfeedback, the Agent will take supervised learning. By propagating interesting changes, the Agent cantake collaboration filtering and improve the detection quality.
