Stories
Slash Boxes
Comments

SoylentNews is people

posted by Fnord666 on Sunday April 15 2018, @05:55PM   Printer-friendly
from the arguing-semantics dept.

Google's latest AI experiments let you talk to books and test word association skills

Google today announced a pair of new artificial intelligence experiments from its research division that let web users dabble in semantics and natural language processing. For Google, a company that's primary product is a search engine that traffics mostly in text, these advances in AI are integral to its business and to its goals of making software that can understand and parse elements of human language.

The website will now house any interactive AI language tools, and Google is calling the collection Semantic Experiences. The primary sub-field of AI it's showcasing is known as word vectors, a type of natural language understanding that maps "semantically similar phrases to nearby points based on equivalence, similarity or relatedness of ideas and language." It's a way to "enable algorithms to learn about the relationships between words, based on examples of actual language usage," says Ray Kurzweil, notable futurist and director of engineering at Google Research, and product manager Rachel Bernstein in a blog post. Google has published its work on the topic in a paper here, and it's also made a pre-trained module available on its TensorFlow platform for other researchers to experiment with.

The first of the two publicly available experiments released today is called Talk to Books, and it quite literally lets you converse with a machine learning-trained algorithm that surfaces answers to questions with relevant passages from human-written text. As described by Kurzweil and Bernstein, Talk to Books lets you "make a statement or ask a question, and the tool finds sentences in books that respond, with no dependence on keyword matching." The duo add that, "In a sense you are talking to the books, getting responses which can help you determine if you're interested in reading them or not."

The second experiment is Semantris, a game that tests word association skills (while collecting data from users).

The very first thing I thought of to ask "Talk to Books" is "What is a cactus?" The first result is "peyote. A species of small cactus, or the powerful drug decocted therefrom by the Indians of Mexico and the western United States and widely used for medicinal, cere­monial, and religious purposes. G.P.M." Thanks, Google.

Also at TechCrunch.


Original Submission

This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
(1)
  • (Score: 2) by maxwell demon on Sunday April 15 2018, @06:34PM (2 children)

    by maxwell demon (1608) Subscriber Badge on Sunday April 15 2018, @06:34PM (#667353) Journal

    Just from the description, it's easy to find a question that it can't answer: Just ask something which in this combination has probably nobody ever written.

    Such as: Did Rincewind visit Hogwarts? [google.com]

    Not surprisingly, no useful result comes from this. Although the first highlighted item is surprisingly irrelevant:

    And Diagon Alley was there. Did you ever see Diagon Alley?

    (I just note that following the link above gives another set of results, since the question mark got removed from the question. Adding that back manually currently reproduces my quoted result. The problem seems to be SN messing with the link.; here's the original one: https://books.google.com/talktobooks/query?q=Did%20Rincewind%20visit%20Hogwarts%3F )

    Any human who has read the relevant books would be able to answer this question: No, he didn't, as both are from different fictional worlds.

    As a side note, it's interesting that the Firefox dictionary knows Hogwarts, but not Rincewind. ;-)

    --
    The Tao of math: The numbers you can count are not the real numbers.
    • (Score: 0) by Anonymous Coward on Sunday April 15 2018, @07:04PM (1 child)

      by Anonymous Coward on Sunday April 15 2018, @07:04PM (#667363)

      It's no fun. I tried questions such as "Is gandalf gay?", "is soylent tasty?" and "who likes dick?" none of which were particularly amusing.

      Some results are strange [google.com]

      The research may be valuable but I don't see any practical use case for this product in its current form.

      • (Score: 1) by Ethanol-fueled on Sunday April 15 2018, @10:12PM

        by Ethanol-fueled (2792) Subscriber Badge on Sunday April 15 2018, @10:12PM (#667398) Homepage

        It looks like a way to sell books.

        "Why does Google engage in censorship?"

        "Harmful" content included material concerning democracy (e.g., freedom), religious cults (e.g., Falun Gong), or antigovernment protests (e.g., Tiananmen Square). Google received much criticism from human rights advocates because it censored information such as human rights.

        -- from Ethics, Technology, and Engineering: An Introduction
        by Ibo van de Poel, Lambèr Royakkers

  • (Score: 3, Interesting) by KritonK on Sunday April 15 2018, @06:49PM (3 children)

    by KritonK (465) on Sunday April 15 2018, @06:49PM (#667356)

    I asked it "Is there a way to decrease entropy?", expecting to get references to Asimov's The Last Question [wikipedia.org]. All I got were references about entropy and changes (increases, not decreases) in entropy from various textbooks.

    I then asked the question verbatim from the story: "How can the net amount of entropy of the universe be massively decreased?". I still got answers from textbooks.

    If it had understood the question, it would simply had answered "no" / "can't be done", if answering scientifically, or produced a reference to Asimov's story, at least in the second case, if it took fiction into account. However, all it seems to have done, was to identify keywords from a natural language question (no mean feat, I'll admit), apply Google's usual "I know better than you" modifications to them (a decrease is a change, so I'll look for changes in general, even though you specifically asked for decreases) then search for the modified keywords in Google Books.

    • (Score: 0) by Anonymous Coward on Sunday April 15 2018, @07:16PM (1 child)

      by Anonymous Coward on Sunday April 15 2018, @07:16PM (#667367)

      It's not just a keyword search. I asked if Obama was a massive wanker and got no results mentioning Obama but there were texts on other politicians (UK Labour leader Neil Kinnock). [google.com] So it's working via some weird form of taxonomy, classifying Obama as a politician and returning results based on politicians. And note how the result changes if you s/was/is/

      • (Score: 2) by KritonK on Monday April 16 2018, @10:17AM

        by KritonK (465) on Monday April 16 2018, @10:17AM (#667565)

        This is just Google's usual "I know better than you" modification to your query. Obama is a politician, so you got answers about various politicians, even though you specifically asked about Obama.

    • (Score: 0) by Anonymous Coward on Sunday April 15 2018, @11:37PM

      by Anonymous Coward on Sunday April 15 2018, @11:37PM (#667409)

      I tried a question I had earlier today (an older relative on digoxin and a beta blocker, currently has a bad cold):
          Is it OK to mix antihistamines with heart medications?
      Looks like it figured out that there are a whole family of decongestants but didn't do so well with the heart meds. Still, some of the results looked relevant.

      To get a good answer earlier today, I called the pharmacy that provides the heart meds (they knew exactly what had been prescribed). They recommended staying away from any of the antihistamines, there are interactions. Stick with something basic like normal Mucinex (expectorant).

  • (Score: 2, Funny) by Anonymous Coward on Sunday April 15 2018, @06:55PM

    by Anonymous Coward on Sunday April 15 2018, @06:55PM (#667359)

    "You're talking to it right now!"

  • (Score: 0) by Anonymous Coward on Sunday April 15 2018, @10:10PM (3 children)

    by Anonymous Coward on Sunday April 15 2018, @10:10PM (#667397)

    Google today announced a pair of new artificial intelligence experiments from its research division that let web users dabble in semantics and natural language processing.

    OK... well... people make mistakes.

    For Google, a company that's primary product

    Fuck it! I'm out. closes window

    • (Score: 2) by kazzie on Monday April 16 2018, @05:38AM (2 children)

      by kazzie (5309) Subscriber Badge on Monday April 16 2018, @05:38AM (#667520)

      ... lets you converse with a machine learning-trained algorithm that surfaces answers to questions with relevant passages ...

      "Serves", or perhaps "services", perhaps?

      • (Score: 2) by realDonaldTrump on Monday April 16 2018, @06:24AM (1 child)

        by realDonaldTrump (6614) on Monday April 16 2018, @06:24AM (#667529) Homepage Journal

        People keep changing the meanings to words. Great for the guys that print dictionaries. Not so great for the rest of us!!!

        --
        Sent from my iPhone
        • (Score: 2) by kazzie on Monday April 16 2018, @09:17AM

          by kazzie (5309) Subscriber Badge on Monday April 16 2018, @09:17AM (#667553)

          Remind me, what's the meaning of "covfefe"?

  • (Score: 2) by captain normal on Sunday April 15 2018, @10:20PM (3 children)

    by captain normal (2205) on Sunday April 15 2018, @10:20PM (#667402)

    If Google wants me to teach their AI they should damn well pay tutorial fees. My time is not cheap.

    • (Score: 1, Interesting) by Anonymous Coward on Monday April 16 2018, @12:44AM

      by Anonymous Coward on Monday April 16 2018, @12:44AM (#667429)

      Passages from books

              ...very little of their code is released to outsiders. One can use many of Google's services for free, as Google makes most of its money on advertising, but you cannot download any of their code to learn from it or improve it or re-use it in ways not envisioned by them. Probably 99% of the code on a typical server at...

      (view in book) [google.com]
      from After the Software Wars [google.com]
      by Keith Curtis
      “After the Software Wars” by Keith Curtis [google.com]

              Google, for example, employs a great deal of open source software in systems development. Its own software is not open source, and there are license restrictions on access to most Google services to prevent others from getting a free ride—for instance, by republishing a Google search as their own.

      (view in book) [google.com]
      from Open Source Software: Implementation and Management [google.com]
      by Paul Kavanagh
      “Open Source Software: Implementation and Management” by Paul Kavanagh [google.com]

              Why are we discussing the program in a book about Google? Because by signing up you can get free Google advertising coupons. Here are the three options for Merchant Solutions programs.

      (view in book) [google.com]
      from Google Advertising Guerrilla Tactics [google.com]
      by Bottle Tree Books LLC., Bottletree Books
      “Google Advertising Guerrilla Tactics” by Bottle Tree Books LLC., Bottletree Books [google.com]

              Anthropological Index (AI) began as a quarterly journal, published by the Royal Anthropological Institute in London, and as an index of the periodicals in Museum of Mankind Library of the British Museum. AI became an online product in 1997 and is a free service. You’ll find AI—

      (view in book) [google.com]
      from Research Methods in Anthropology: Qualitative and Quantitative Approaches [google.com]
      by H. Russell Bernard
      “Research Methods in Anthropology: Qualitative and Quantitative Approaches” by H. Russell Bernard [google.com]

              ...they had a window to be the only game in town,” says Liebman, who came over from Applied Semantics. Just to show that the system could help advertisers and publishers, Google assumed all the costs while it proved its point. Normally, the process began when a publisher signed up for the program and assigned space on a...

      (view in book) [google.com]
      from In The Plex: How Google Thinks, Works, and Shapes Our Lives [google.com]
      by Steven Levy

      “In The Plex: How Google Thinks, Works, and Shapes Our Lives” by Steven Levy [google.com]

    • (Score: -1, Flamebait) by Anonymous Coward on Monday April 16 2018, @12:56AM (1 child)

      by Anonymous Coward on Monday April 16 2018, @12:56AM (#667437)

      Google wants you to do what they failed to do: natural language processing. They're hoping that some naïve individual or individuals will have ideas to fix their broken system. As you know, jews are good at nothing except breaking the planet and making money through lending and other illegal and shady practices. AI is none of their business, so they get humans of the planet to do it for them.

      The devil is not creative. What google can't do is create. All they can do is destroy human creation and copy it. With their AI projects and all other projects, including robotics (military executioner robots), all work is done by humans, who in order to pay their (artificially high) rent have to do these slave jobs.

      • (Score: 2) by realDonaldTrump on Monday April 16 2018, @03:33AM

        by realDonaldTrump (6614) on Monday April 16 2018, @03:33AM (#667486) Homepage Journal

        Boston Dynamics, lots of smart guys there. Ethanol-fueled. And they make some terrific robots. For our Navy, for our Marine Core, for our Army. So many parts of our great military. But it's not Google anymore, they sold it to the Japs last year!

        --
        Sent from my iPhone
  • (Score: 2) by realDonaldTrump on Monday April 16 2018, @02:14AM

    by realDonaldTrump (6614) on Monday April 16 2018, @02:14AM (#667459) Homepage Journal

    I'm no libertarian. But this new Google site makes me feel like one!

    --
    Sent from my iPhone
  • (Score: 0) by Anonymous Coward on Monday April 16 2018, @03:34PM

    by Anonymous Coward on Monday April 16 2018, @03:34PM (#667661)

    How about Google pull it's head from it's ass and learn to translate numbers from Chinese websites into English. Doesn't sound too hard, but any numbers in the thousands to millions range are always translated incorrectly, factor of 10 plus or minus.

(1)