Personal Assistants for the Web: An MIT Perspective
- 127 Downloads
The growing complexity of sources of information and users’ needs for information is fast outstripping the manual browsing and searching interfaces of the past. A long-term approach to the complexity problem is to design intelligent information agents that provide active assistance in the process of finding and organizing information. These agents differ from conventional information retrieval in how they interact with the user and the dynamic nature of the information they deal with. This article surveys some work at the MIT Media Lab in developing intelligent information agents, especially as assistants to users browsing the Web. These agents don’t replace conventional browsing or direct manipulation interfaces, but work with such interfaces to learn from interaction with the user, anticipate the user’s needs, and connect the user with other people who may share his or her interests.
KeywordsRecommendation System Inverse Document Frequency Personal Assistant Information Agent Semantic Neighborhood
Unable to display preview. Download preview PDF.