Submission Type: Long Presentation

Enough with the fudges. We need to get inside the text to be searched, not write algorithms from outside it. If we transform text into an internal representation which destroys none of the meaning, and we transform a query into the same form, we have semantic search. The internal representation can’t be just a pretty diagram, but should be involved in the transformation process – an active self-extensible structure. Such a structure requires a synthesis of many logics, while a human should be comfortable with the visual display of facets of the internal structure. How large would this structure become? What resources would be required for large scale semantic search, given there are billions of pages on the internet, and Google may have a million machines dedicated to key word search, requiring less than a thousandth of the resources required for semantic search?

Abstract: Large Scale Semantic Internet Search slides

Authors: Jim Brander

Event: Fourth HCSNet Next-Generation Search Technology Workshop (NGS09)

← View all submissions for this event.