May 13, 2008

Semantic Search Engine Web 3.0 Tools

As you know, I've had my eye on Radar Networks and recently became approved as a beta tester. I'm continuing to eyeball Radar's "Twine" technology, which renders search results in a more inter-relational way than Google.

Sunday, a company called Powerset rolled out a new tool which also offers a new twist on search: complete language search (versus the stand "keyword" search technique common through AOL, Google, Yahoo, and MSN today).

The idea of searching with complete language is intriguing to me. The initial roll-out of Powerset (http://www.powerset.com) allows a person to search the Wikipedia database using any combination of words. I actually found myself, habitually, typing in phrases, and seeing Powerset deliver results in a well-organized format, based upon those phrases. Not much difference there. However, I noticed videos were off to one side. The primary response returned in an upper-left box, followed by a Google-esque list of search URLs and corresponding results similar to other search engines. However, when I changed this phrase from simple searches, like "Tony Robbins," to more complex searches like "I like Dulce De Leche Ice Cream," the engine shined. First of all, Powerset returned a clarification question (I misspelled the phrase) which I ignored. Then, a count of matches within Wikipedia. After that, what looked like a Google search response; however, upon closer examination I noticed the results were more varied and offered a 360 degree view of the phrase, rather than a specific, narrowed search.

To clarify, it seems that Powerset offers the narrow phrase response first, then widens the results as you go deeper. Certainly, this technology is worth examining further. I'm curious if Powerset is going to include graphic search response in the future, or if they will concentrate on textual search as they are now. In addition, I could watch a video on the right. I could also use the "Add This" bookmark to share the link I found.

A quote from an article on Powerset by Reuters mentioned that "The Wikipedia is becoming a microcosm of the most useful parts of the Web," said Greg Sterling, an Internet analyst with Sterling Market Intelligence. "This offers a powerful way to find what you are looking for against this subset of the Web."

"Our system is a little more forgiving," Scott Prevost, general manager of Powerset, said in an interview on Sunday. "It is not looking for hard-word matches. We are not searching for exact words, but concepts," he said.

Intriguing, wouldn't you say?

The technology for natural language search comes out of Xerox PARC, the same lab that brought us the initial technology we now are so comfortable with through Windows and the Mac interfaces. 2008 appears to be the year that semantic search and language processes are becoming reality.

Additional articles related to this topic include:
Semantic Search Closer To Reality Than You Think
Tying It Together With Twine

Are you interested in being a Beta tester for ARRiiVE.com's new collaboration software based upon semantic concepts? Please email us or contact us to let us know. Thanks!
________________________________________

Copyright © 1999-2008 by ARRiiVE Business Solutions. All Rights Reserved. Powerset is a trademark of Powerlabs. Google, Yahoo, AOL, Xerox, Yahoo, Radar Networks, and Wikipedia are trademarks of their respective companies. SUBSCRIBE.

No comments:

Post a Comment