Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
The Military The Internet

DARPA Seeks the Holy Grail of Search Engines 78

coondoggie writes "The scientists at DARPA say the current methods of searching the Internet for all manner of information just won't cut it in the future. Today the agency announced a program that would aim to totally revamp Internet search and 'revolutionize the discovery, organization and presentation of search results.' Specifically, the goal of DARPA's Memex program is to develop software that will enable domain-specific indexing of public web content and domain-specific search capabilities. According to the agency the technologies developed in the program will also provide the mechanisms for content discovery, information extraction, information retrieval, user collaboration, and other areas needed to address distributed aggregation, analysis, and presentation of web content."
This discussion has been archived. No new comments can be posted.

DARPA Seeks the Holy Grail of Search Engines

Comments Filter:
  • by Anonymous Coward

    When they can just google it?

    • by aliquis ( 678370 )

      They aren't spying and censoring the information.

      They are just making a search engine and present the information which fit your profile.

    • by Anonymous Coward

      Have you tried Google lately for a complex search? Or a search for a specific topic which uses words that are common to other, more popular topics?

      It's a disgrace, and getting worse every year.

      • Have you tried Google lately for a complex search?

        I believe that the presence of the words "information extraction" and "information retrieval" in the summary means that the search engine should be able to answer questions of the type "which kings were assassinated less then two years after their coronation?" by going through the facts available on the web, extracting them, and figuring stuff out. For any given random question, it's unlikely that someone already has it answered on the web, but the facts are all there anyway.

        • Sounds like they need to hire Watson, then.
          • Sort of. But Watson is probably too small (or primitive) a system to be efficient at a large scale (as in, serving a Google-sized audience). I have no idea how DARPA wants to tackle that. But hey, that's why they call it research! ;-)
      • yes, many times most days and i usually find what i want
      • Google is not the answer.

        Want proof? Try these two searches:

        "Thor"
        "atm"

        How is Google supposed to know what to do with that? Do you want norse mythology, a comic book, or something else? Do you want packet switching information, bank information, or "Acrylic Tank Manufacturing" -- that's a new one.

        About a decade ago, "Cow9" -- that was the name of the alta vista search engine -- had a wonderful solution to this, that required loading a java applet into your browser as part of the search. I loved it, and was

    • FBI M-O-U-S-E. next best thing to surfing along looking over your shoulder, taking notes, texting the US Marshals Service what they need to subpoena before driving over and clapping on the leg irons.

    • by Anonymous Coward

      Google really is an exceptionally good search engine if it wasn't for a career known as "Internet Marketing". Scam and ad sites are the #1 reason why it's so damn hard to find relevant pages in a search. I should not have to construct a search query that spans the entire width of a 1680 monitor just to filter out all the criminally insane "Internet Marketing" guys grab for cash.

    • by Anonymous Coward

      Actually there is more truth in this answer than most people would think. With wolfram alpha and knowledge graphs Google did recently included features that actually do waht DARPA has in mind: make sense of crawled data, draw conclusions and display accordingly.

      The difference with Google is: They do introduce the changes step by step. But it is defininately out there: I just recently noticed that the first search results are more and more spread around different possible meanings of my search terms, when in

  • ....needed to address distributed aggregation, analysis, and presentation of citizens' content."
    FTFY
  • by hAckz0r ( 989977 ) on Monday February 10, 2014 @02:34PM (#46211245)
    ... that Snowden didn't have good enough tools available.
  • by mmell ( 832646 ) on Monday February 10, 2014 @02:37PM (#46211263)
    First, remember that this is essentially where the world wide web of internetworked computers began. The words 'unintended consequences' come screaming to mind . . . (although I have to admit - I love those wicked pipes!)

    Second - we (common netizens) may welcome the sort of information availability DARPA is seeking - sort of like the scifi future where you just ask the nearest terminal whatever you want to know and magically get the answer you need - but there are lots of bad people still running around on this planet (scamsters, governments, jilted ex-lovers, religious extremists, etc.). The problem isn't the technology, the problem is our ability to handle it.

    I very much suspect DARPA may be onto something. I wonder if it will be as beneficial as the WWW has been.

  • Just how many buzzwords can we fit in here, anyhow?
  • Google already does domain-specific indexing, certain sites get indexed faster or deeper than others based on a number of secret rules.

    For site specific search prefix your query with "site:foo.com"

    • Addendum: It just hit me that they might've meant *topic* specific.

      Google has something like this as well but it can be hard to notice. I remember one time I was searching for something automotive-related and my query had a word in it that meant something in the fashion world. It took hell to get out of "fashion mode" and get automotive results, IIRC I had to completely rephrase the question.

  • So do I... One that is robust and can't be censored by anybody, ever.

  • Just search in Google for "holy grail of search engines". Job done.

  • by Anonymous Coward

    Sometimes, as you read the RFP document you can tell it's been written with a specific proposer in mind, and everybody else gets a short window to come up with an idea + jump through endless list of government hoops.

  • by koan ( 80826 )

    Google, Deepmind, etc.

  • Today unstructured information search findability is limited by the Shannon Limit, this is a fundamental physical limit since all pattern search engines are statistical decoders. Google does a little better than the Shannon limit by looking at which search results are selected, this a communal intelligence technique based on how we "vote" for the right result. Unfortunately this only works well for high volume searches, that's why Google work's best if you know exactly what you're looking for or you're look

  • If they can't find the Holy Grail of search engines, I can direct them to this alternative that is just a Grail Shaped Beacon of a search tool instead...

  • This sounds like progmatically generated RDF(http://www.w3.org/RDF/), and could be handled via OWL(http://www.w3.org/TR/owl-features/)... Current search engines and metadata focus on content (which makes sense), but leave out context. W3C have been supporting research on Ontological networks, where semantics are collected and linked. The W3C concept is self-assigning meaning to the content. DARPA's Memex sounds like they are assigning meaning via a 3rd party. Less privacy, more meaning.
  • Have one without "sponsored ads".

    Google has produce less useful hits over the last five years, as its advertising income skyrocketed. Search for something, and add -"photo", and you'll still get photographer. Try mens riding boots -women -womens -women's, and you'll probably get a "sponsored ad" for women's boots, with the letters "men" bolded.

    This isn't even counting Target, which will claim it has anything you're looking for, but if you follow the link, oh, no, sorry... but we have !!!!!

    Add the way librar

    • ... not just with AND and OR, but also "within so many words"...

      Try Google's firstterm AROUND(3) secondterm.

  • My screen just came up in beta, slow to load & impossible to read comments. Tried clicking on "classic" at the bottom, it's been trashed also. I logged in today for the first time in ten years just to say Beta Sucks and I will be finding another site to read every morning with my coffee. I will be canceling my subscription...so sad but this "new" look with its intention run by greed and ego is NOT why I signed up for slashdot many years ago.

One man's constant is another man's variable. -- A.J. Perlis

Working...