TV Schedule

New search engine searches by meaning, not words

  2. jpoRS
  3. related topics
Powerset, a new search engine, is setting itself up to change the way people search the web. Instead of looking for exact key words, like all major search engines do now, Powerset looks for the meaning of the word or phrase to find more detailed information.

Currently it only works on Wikipedia, but once the technology is applied to the open web, analysts predict it may be acquired by Microsoft to turn MSN into a lean, mean, Google beating machine.
jpoRS

20 responses // New search engine searches by meaning, not words

  • i already distrust it because it'll be searching wikipedia which is well known to contain false information all over the place

    its one step closer to AI, a program thinking for you
    diode
  • I don't trust it because of the dirtiest word there...Microshaft.
    echoz
  • This sounds similar to "ask.com" (formerly "ask jeeves.com").
    huntre
  • @diode: I suppose every encyclopedia published is completely unbiased right?
    _Hayko
  • I think Google is gonna take a lot of beating...

    I don't think just because it uses wikipedia as it's testing ground then this should be written off. That's like saying you wouldn't drive a car that's been tested on a traffic-less road. It doesn't make it any better/worse, it just means it's been tested somewhere a little more private in comparison to road-testing on a major motorway intersection.
    mattbrawn
  • http://www.news.com/2100-1038_3-5997332.html
    Image...
    @diode & mattbrawn

    A scientific study published in the journal Nature found no real difference in the quality of articles on Wikipedia vs. the quality of analogous articles on Britannica.

    Also, wiki is far faster to catch and repair errors.
    petervan
  • Thanks for that Peter, interesting reading.

    I wasn't discrediting wiki, merely stating how I preferred the idea of testing a new search engine on something less chaotic than the whole WWW, hence the road test analogy.


    mattbrawn
  • http://www.roughtype.com/archives/2006/02/community_and...
    Image...
    re: that Nature study...how much does it depend on the kind of information you're researching? this link provides a few specifics on how that study was conducted.

    "The article itself doesn't actually go into much detail about the survey's findings. It says that the "expert-led investigation" revealed that "the difference in accuracy [between the encyclopedias] was not particularly great: the average science entry in Wikipedia contained around four inaccuracies; Britannica, about three." But Nature subsequently released "supplementary information" about the survey, including more details on the methodology and a full list of the errors cited by the experts. (In total, Wikipedia had 162 errors while Britannica had 123.) Read together, the article and the supplementary information indicate that the survey probably exaggerated Wikipedia's overall quality considerably...

    "The Nature survey, in other words, played to Wikipedia's strength. That's fine. Nature is, after all, a scientific journal. But, unfortunately, the narrowness of the survey has tended to get lost in media coverage of it..."
    echoz
  • It doesn't have much in common with the old Ask Jeeves. The technology (some from Xerox Parc) is way better, and the focus is much more on analyzing document text rather than query text.

    Give it a try and form your own opinion. So far, I've had mixed results, but find the interface fun to experiment with.

    [Disclosure: I have friends there and was in the Beta program, but I'm no fanboy ;-]

    marcozarco
  • Give Google a run for its money? I bet Google will get its hands on this technology as soon as they can. I don't think MSN should get its hopes up....
    Oxyuscan
  • watch this comment being used here, here, here, here and here
    Things are gonna get so much more complex than this...
    nwillens
  • Actually, huntre and marco, I used to work at Ask Jeeves and the task here is almost precisely what that company once tried to do. Marco, you're right that analyzing queries versus document sets is quite different, but the notion of a "natural language" or semantic approach incorporating conceptual matching rather than mere keyword matching, is exactly what Ask did and to some extent still does with its "Smart Answers" product.

    It works great with a limited document set of a certain type - Wikipedia is pretty ideal - but pretty poorly on the Web at large. This technology is not a Google killer, IMO. It's not scalable in terms of the conceptual aspect because that inevitably requires more human-tuning than is really possible on a large scale.

    Every couple years one of these natural language companies comes along and then disappears without a trace. Ask only stuck around because it built a great brand and made some canny acquisitions, in spite of weak search performance.
    Blazesboy
  • watch this comment being used here, here, here, here and here
    LOL. so how long till Google sneaks in and snatches them up?!?!

    I doubt Microsoft will even hold a candle to the amount of cash Google is going to throw at them in order to keep its reign over search.
    clarity_kat
  • It's interesting idea and concept. Maybe it'll work and maybe it'll bust, but it's good to try and see.
    J_Jammer
  • Wikipedia stinks. Our choices are dwindling. Can you say newspeak? Orwell, are you listening?

    It all sounds pretty double plus bad to me.
    Shannon5
  • I don't know how useful this will be; I think people are so used to already using search engines to find specific words, it might be hard to transition to try to use phrase meanings to find things. Maybe in instances of asking a question or something, but I think other than that, this type of search engine isn't as useful as it may seem.
    Peewong
  • http://en.wikipedia.org/wiki/Truthiness
    Image...
    Uses "truthiness" technology
    Talixx
  • As pointed out by nwillens this tech aint new.

    To be a ground breaker it would be extensible to the web. As it is it can only work on confined document sets it's use will be confined to sites such as Wonkipedia.

    BTW, Wikipedia is a great example of our corrupt society.
    Purdey
  • Wikipedia is funny when it's brought up. People act like written word in book form is infallible.

    Everything has a chance of being corrupt. People, places, things....it's not the form that matters as much as the truth of is it or isn't it.
    J_Jammer
  • here's what I got for Suck....



    Suck A slang term that refers to something unpleasant or deficient, as in the expression "That test sucks."
    tribe10

Add your response

Login/Registration is required to add a response.