Powerset presents semantic search engine for Wikipedia

The Californian company Powerset, which wants to enable internet searches based on ‘natural language’, presented its first product on Sunday: a search engine for the English version of Wikipedia.

With the new search tool users can ask questions such as “how many people were evacuated from New Orleans during hurricane Katrina” or “what was banned by the FDA”. They will then be presented with a list of articles in which the answer is most likely to be found. The traditional search engines can only search for words or phrases, something that the Powerset search engine can also do.

Powerset chose Wikipedia to start with, because this is a highly structured collection of data, in which it is relatively easy to search. The company plans to extend its search to other well-organized data collections such as patents, the CIA factbook or Wiki clones. Ultimately, it is the intention that searches can be made all over the internet, but the realization of this can take a while, so message Reuters.

According to Greg Sterling, an analyst with the research firm Sterling Market Intelligence, it will be many years before Powerset is able to search the World Wide Web on the same scale as Google. However, when this is the case, Powerset could become a formidable competitor to Google. Rumors have it that Microsoft, which recently withdrew its bid for Yahoo, is interested in Powerset. Spokespersons for both Microsoft and Powerset declined to comment on a potential merger between the two companies.