Context-Aware Personal Information Agents

Abstract

We present a society of personal information agents that work for a community of users and that are aware of the physical and social context of their users. We show how context-awareness is a feature that allows the agents to improve their performance when they work with limited resources in information spaces with a large amount of information. The use of context information allows the agents to focus their information search and, as a result of this, increase the quantity and quality of information delivered to the user. Moreover, we propose an implemented agent architecture with context-aware capabilities. We discuss this architecture in detail, focusing on the capability of exploiting the windows of opportunity provided by the awareness of the users' activity. The agents use these windows of opportunity to furnish the user with information and advice in the situation where it can be most useful. In particular, we show how context-aware information agents can assist a community of attendees to big conferences and fairs. In this application, an information agent gathers relevant information based on a model of specific interests of a user. Given the multiplicity of interesting events, and their distribution in time and space, an information agent has to deliver the gathered information in a few hours and comply to the schedule constraints. Finally, we report some experimentation results to illustrate how context-awareness improve the service a society of information agents provides to a community of users in the conference application.