Reasoning about Changes and Uncertainty in Browser Customization

Browsers such as Mosaic are popular tools to explore large repositories of information (e.g., the Internet). Yet users of a browser will be served even better by a tool that is capable of customizing itself to the individual needs and interests of its current user. This paper reports on the design of a browser customization tool which automatically and autonomously identifies personally relevant information, and adapts its support to the changing personal interests of its users. We adopted a recently published probabilistic approach to reasoning about dynamic change and uncertainty in order to model our envisaged reasoning of an autonomous agent for browser customization.