This Small Business Innovation Research Phase I project shall prove the technical feasibility of semantic search using meta or search documents. Today's Internet search services use various levels of semantic understanding to enhance the results; however, most limit their conception of a search to a single query. Nevertheless, queries are usually one step in a larger process; users often visit multiple websites, collect results of interest, and evolve queries over time as focus shifts from concept to concept and website to website. The goal of this effort is to incorporate the entire context of a user's search into a meta document, semantically disambiguate all terms in that document, and provide automated results based on discoveries of other users' prior and contemporaneous searches.
Search is an integral part of how people use the Internet. Meta documents make it easy for people to save, share, and re-use complex searches. Availability of a meta document will transform search from a tool for discovering individual web pages into a means to create and share knowledge. Transforming complex searches from an abstract process of repeated queries across multiple websites into a meta document is a simple, yet profound, shift in how search is conceived. Having an explicit representation of users' multi-site, multi-query searches enables any number of strategies for improving search results in multiple ways. If successful, the meta document will create a tabula rasa for future third party innovations and thus have significant economic impact.
