Our Technology

Highlight is designed specifically to

• help users get more documents from multiple sources. To enlarge the search scope Highlight dispatches a query to a set of popular search engines. Individual results are merged and re-ranked.
• help users gain an overview of the entire results by developing a dynamic concept hierarchy for each search session. The hierarchy contains important conceptual terms in the documents as well as their relationships. The concept hierarchy can be viewed as a summary of the returned documents, so it can reduce the information overload.
• help users browse the returned document faster and more easily by linking each term to a set of documents which contain the term.
• help users search for specific information by enabling "search within results" when a lower level term is selected.
• help users locate specific information in a long document by highlighting the keywords in the currently selected branch, or, if the user likes, by displaying the relevant paragraphs only.

System Arfchitechture

Working as a meta-search engine, Highlight retrieves documents from several popular search engines including Google, MSN, AltaVista, AllTheWeb, Teoma, Lycos etc. When the user inputs a query, the system dispatches the query to the selected search engines through the corresponding wrappers. Each wrapper translates the query into an acceptable request to a search engine, submits the request, retrieves documents, and parses them into an internal structured format. Documents from multiple sources are merged and ranked by the averages of their orders in individual sources. Noun phrases are extracted from the title and the short text coming with each document hit. A dynamic concept hierarchy is then constructed and presented to the user.

Publications

Some of the publications are listed as follows:

1. Bot, Razvan Stefan and Wu, Yi-fang Brook and Chen, Xin and Li, Quanzhi (2005) Generating better concept hierarchies using automatic document classification. CIKM '05: Proceedings of the 14th ACM international conference on Information and knowledge management. Bremen, Germany. pdf

2. Wu, Yi-fang Brook and Shankar, Latha and Chen, Xin (2003) Finding more useful information faster from web search results. CIKM '03: Proceedings of the twelfth international conference on Information and knowledge management. New Orleans, LA, USA. pdf