Web Search Is About to Get a Whole Lot Smarter

July 26th, 2010 Posted in Information Retrieval, Other News

Google’s recently-announced acquisition of Metaweb, a company that Google describes rather vaguely as maintaining “an open database of things in the world,” promises to realize the type of free, far-reaching human-language semantic web search that has up till now has been more speculative fantasy than reality.

Simply put, Metaweb maintains a database of “entities” — specific people, places, or things. Current web search relies on words, which are almost always ambiguous (to borrow the example from the video below, “Boston” could refer to a number of different cities around the world, a sports team, or a rock band), and so word-based web searches like we use now yield imperfect results that confuse all the different possible meanings of a given word or phrase.

But using a service such as Metaweb’s, each individual meaning is identified as its own separate entity — a specific concept, separate from the words that may be used to refer to that concept — which is entered into a database. Then, individual instances of this entity are identified with metadata that references the database, thus precluding any confusion between, for example, Boston the city and Boston the band.

For a clear, jargon-free illustration of Metaweb’s services and the concept of semantic search, watch the video below.