written by:

On Monday, Powerset launched the beta version of their own search engine. While this is not interesting in and of itself, they have attracted a great deal of attention by using a unique algorithm to generate more relevant results. Using ‘natural language search,’ their system actively interprets the meaning of words rather than simply matching them or analyzing how frequently they are mentioned in a document. In the blogosphere world, this form of artificial intelligence has propelled them into direct competition with our favorite search engines. In addition, it has sparked rumors Microsoft may be interested in buying them out. I have to give them some credit, they were considered competition to and acquisition material for the most advanced search engines on the planet, all pre-beta release.

What makes this system so interesting is the engine’s ability to interpret the meaning of words, and actually deliver on that promise. Rather than simply indexing all the material contained in a document, the engine literally reads and assigns meaning to every single word in every single document. Talk about heavy processing capacity. By analyzing the literal meaning of a word, and then analyzing the linguistic meaning of a word and finding synonyms, the engine can find what they claim to be more relevant results than the ‘exact match’ system used on our favorite engines. I would normally smash any engine trying to compete with Google, but how can you blame anyone for trying to deliver more relevant results to the end user. The only problem with their system now is they have only indexed pages from Wikipedia and Freebase, another open database. So, as with all online content, the question becomes whether they can transform their system into a source of revenue. Given Wikipedia has failed to do this, I am curios to see what happens down the line.

And since you are all curious, click the Powerset link above and try asking the engine a question yourself. You may be surprised by the results it delivers.

Comments

5 thoughts on “Powerset Search Engine – Natural Language Search
  1. Mike Block says:

    Wow, this is really impressive! Searching within Wikipedia has always been difficult. It seemed like you had to get EVERYTHING right, even capitalization, in order to find what you were looking for. Searching through Google meant that Wiki results were likely, but not guaranteed to be numerous. This changes everything. I’m not sure how well the actual format will compare to Google’s, but it’s off to a good start in Wiki.

  2. Ryan says:

    I’m very interested to see how this new technology will hold true outside of wikipedia. It seems like a great idea though!

  3. I was checking out this site the other day and it was cool to input full sentences and see the results. Sometimes it was right on, other times….eh.. not so much, but all in all I welcome the Semantic Web!

  4. edwin says:

    They finally released the beta. Jeez! When i spoke with the CEO of Powerset in November he said they were going to release the open beta in December of 07. I guess it took a little longer than they expected. I suspect it will take several more years for their technology to really mature and begin to impress us- but – like the poster above me said, ” welcome the Semantic Web!! “

  5. Theodore Cohen says:

    Good stuff. It’s always great to watch as we try and get a web that is closer and closer to AI. As for going against Google…well…Google has become so much more that I think it’s ok too. Always good to have a little competition, or the possibilities of it, in the marketplace.

Leave a Reply

Your email address will not be published.

Business

EarthLink Wi-Fi Drops It’s Signal, Forever.
Google

Web Spam and You
Internet News

A Smarter Web
Become An Insider!

Thanks for signing up to be a Wpromote Insider.
You’ll be the first to get the scoop on our latest services, promotions and industry news.

CONNECT
  • Los Angeles HQ: 866.977.6668
  • Chicago: 310.529.4578
  • San Francisco: 310.683.0435