Stories
Slash Boxes
Comments
typodupeerror delete not in

Hot Comments

Comments: 178 + -   Google Shooting For Smartphone Universal Translator on Monday February 08 2010, @06:12PM

Posted by ScuttleMonkey on Monday February 08 2010, @06:12PM
from the ford-why-is-this-fish-in-my-ear dept.
communications
hardware
technology
nikki4 writes to tell us that in giving some major improvement tweaks to its existing voice recognition tool for the Smartphone, Google is aiming for new translator software that will provide instant translation of foreign languages. "The company has already created an automatic system for translating text on computers, which is being honed by scanning millions of multi-lingual websites and documents. So far it covers 52 languages, adding Haitian Creole last week. Google also has a voice recognition system that enables phone users to conduct web searches by speaking commands into their phones rather than typing them in. Now it is working on combining the two technologies to produce software capable of understanding a caller’s voice and translating it into a synthetic equivalent in a foreign language."
story

Related Stories

This discussion has been archived. No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
 Full
 Abbreviated
 Hidden
More
Loading... please wait.
  • by zero_out (1705074) on Monday February 08 2010, @06:16PM (#31066088)
    Maybe my experience is atypical, but Google doesn't seem to translate pages very well. I can only imagine how bad it will be having a phone do this. "Did that guy's phone just call me what I think it did?"
    • by amRadioHed (463061) on Monday February 08 2010, @06:44PM (#31066460)

      Considering that human languages follow rules that are as convoluted and transient as Calvin-ball I'd say they do an admirable job. Machine translation is really an amazing challenge.

      • Re: (Score:2, Informative)

        by zero_out (1705074)
        As someone with a background in AI and HCI, I completely agree. Unfortunately, we still have a long way to go, and I think that Google is jumping the gun on this. It should prove to be quite humorous, even as first steps go.
        • As someone with a background in AI and HCI, I completely agree. Unfortunately, we still have a long way to go, and I think that Google is jumping the gun on this. It should prove to be quite humorous, even as first steps go.

          Well, to be fair, they did say this would be something they are shooting to accomplish "In a few years". Still a tough task, but they're giving themselves some time. Considering how far they've gotten so far, I'm really excited to see how this works out.
          -Taylor

            • Re: (Score:3, Insightful)

              by icebike (68054)

              I've been using the google App on my iPhone to speak my searches and it is amazingly accurate. For some people its approaching 98% accurate in the voice recognition portion.

              Way better than a fish in the ear.

        • As a professional translator and interpreter, I also agree that Google has done better than anyone else, and that that's still not particularly good by objective standards. I've been using their Translation Center (NOT Google Translate) for a while now, and I've seen their translation memory evolve before my eyes.

          The basic problem, however, is that the computer doesn't actually understand what it's spitting back to you. It only spits back the translations others have provided for similar phrases. It doesn't know if they're any good. Sometimes they're surprisingly good, and sometimes they're bizarrely bad.

          There's a lot of ambiguity in human writing, and even more so in speech. Even assuming you hear the words correctly, it's tricky to tease out the precise meaning they wanted to convey, and trickier still to re-express that in another language, with appropriate cultural and regional context.

          Google will get better and better at parroting good translating and interpreting decisions, but software will never be able to make those decisions, because, in the final analysis, they are subjective decisions.

            • by wisty (1335733) on Monday February 08 2010, @09:41PM (#31067888)

              Google will get better and better at parroting good translating and interpreting decisions, but software will never be able to make those decisions, because, in the final analysis, they are subjective decisions.

              Think about how successful google has been with search. Prior to the web, we would have idealized search as speaking with an expert who has all the knowledge that exists on the web. Various efforts still strive for that vision today (askjeeves, wolphramalpha, etc). But clearly it is unreachable for the forseeable future. Yet, search is very useful.

              Similarly, this universal translator may well reach a point that it is possible to visit a place, buy things, have a meal, ask where the toilets are, and get back home, particularly when both parties in the conversation are familiar with the limitations of translation. That would be extremely useful, even if it's only 1/100 of all a native bilingual speaker understands, or what you would need for nuanced treaty negotiations or to author a respectable translation of War and Peace.

              In Chinese, then back to English:

              Think about how the success has been with Google search. Prior to site, we will work with specialists who have all the knowledge and presence on the network to speak idealized search. However, efforts to fight the idea, (it is by virtue of, wolphramalpha, etc.). But obviously can not access the foreseeable future. However, the search is very useful.

              Similarly, the universal translator is likely to reach a point of view, is that we can visit places, buy things, eat a meal, and asked where the toilets and get back home, especially when the parties are familiar with the limitations of dialogue and translation. This will be very useful, even if only 1 / 100 of the machine for all those who understand the bilingual, or you need to nuanced negotiation of a treaty, or the author's respect for the translation of war and peace.

              • by SnowZero (92219) on Tuesday February 09 2010, @04:51AM (#31069768)

                (Disclaimer: I happen to work for Google, but not on anything related to machine translation.)

                You are demonstrating his very point. Translation will not get nuanced stuff, but it could greatly help everyday interactions for travelers or recent immigrants.

                Let's do English->Chinese->English on his actual examples:

                • buy things, "What is the cost of this umbrella?" -> "What is the cost of this umbrella?" (Note I didn't say "how much for" because I am familiar with the limitations of translation and know that phrase is a colloquialism.)
                • have a meal, "I would like to order the beef soup." -> "I would like to order beef soup."
                • ask where the toilets are, "Where is your bathroom?" -> "Where is the bathroom?"
                • get back home "How do I get to the Hilton hotel?" -> "How do I get in the Hilton Hotel?"

                I'd say that is pretty passable. Now, it would be better if folks could learn the local language, but for anyone who travels a lot you realize that it is not practical to learn a new language every single trip. Something like this might also help more folks travel with a little less fear, and experience places they otherwise wouldn't. Tools such as this could also allow older immigrants more access to the country they now live in.

                Machine translation has a long long way to go to even be considered "good", but having something close to the state or the art, working on improving it, and making it free for all to use seems like a good thing to me.

              • Re: (Score:3, Insightful)

                The translation party [translationparty.com] equilibrium:

                I, Google Maps and do a link and, please, I do not think it is successful. Aruuebu, all Web search experts, to discuss the background of existing knowledge. In addition, various efforts and vision, dedication, today (U.S. time askjeeves, wolphramalpha etc.) lead. But this has obviously reached the foreseeable future. However, the search is very useful.

                Similarly, from the perspective of our toilet, I can eat, and to translate the conversation reaches the end of a universal

    • by Kijori (897770)

      Maybe my experience is atypical, but Google doesn't seem to translate pages very well. I can only imagine how bad it will be having a phone do this. "Did that guy's phone just call me what I think it did?"

      If you haven't used it recently, try it now. Speaking as a linguist I am incredibly impressed by the speed of their progress.

  • Or... (Score:3, Funny)

    by fatherjoecode (1725040) on Monday February 08 2010, @06:16PM (#31066092) Homepage
    Or you could just stick this Babel fish in your ear.
  • by slyborg (524607) <fbrunner@macDEGAS.com minus painter> on Monday February 08 2010, @06:17PM (#31066096)

    ...this is a recipe for universal worldwide hilarity.

    • by sakdoctor (1087155) on Monday February 08 2010, @06:43PM (#31066444) Homepage

      I tried to translate your sentence multiple times, then back to English so I could post the ridiculous result.

      Except google's translation was actually pretty good.

      • by Tetsujin (103070) on Monday February 08 2010, @08:00PM (#31067248) Homepage Journal

        I tried to translate your sentence multiple times, then back to English so I could post the ridiculous result.

        Except google's translation was actually pretty good.

        Try a more complicated example. For instance, starting here:

        "It's probably pretty good at translating translations it produces back into the same source text. If you figure that a phrase structure in one language corresponds to a certain data structure in Google Translate, then it makes sense that this data structure would survive multiple passes through the same restructuring algorithm..."

        translating to Japanese and back to English yields this:

        "It is translated to produce translated text back to the very same source is probably a good thing. Cases, one single phrase structure of language specific data structures in the Google translation, it is this data structure makes sense and survival of multiple paths through the same algorithm structure corresponding figures ..."

        Here you've got badly handled idiomatic phrases all around... Like the Google translation to Japanese used "seiseisuru honyaku no honyaku dewa ii koto da" at the end of the first sentence ("created-translation's translation is good" or something like that). On the translation back the connection between "good" and "translation" was lost - Google slapped on a fairly generic "is probably a good thing" - picking the bit of uncertainty out of the start of the Japanese sentence and combining that with the "dewa ii koto da" - but dropping the whole idea of what it is that's good... Which is something that can be kind of vague in the structure of Japanese... Meanwhile, the phrase "source text" was transliterated into katakana, but it got broken up in the translation back to English and wound up in two different locations in the sentence...

        The whole conditional clause in the second sentence got kind of mangled. In the Japanese translation it starts with "baai wa": baai means "case" or "situation" - the structure of the sentence establishes this "case" being described as a possibility... Google lost all that, and just said "cases," Then, at the end of the sentence, after the ellipsis, "figure", from "if you figure" in the English original, was tacked on as "taiousuru zu" - "interacting drawing" or "interacting figures". In the return-to-English version this somehow wound up back before the ellipsis again.

        The rest of the second sentence in Japanese is something like "if this data structure uses the same intermediary algorithm, several passes of the algorithm should be survived and it should make sense." The apparent problem there is something analogous to operator precedence in arithmetic. The "and" is meant to mean that the surviving translation should still make sense - but this clause apparently got broken up... like the reverse translation assumed that "uses the same intermediary algorithm... should be survived" was all one stand-alone clause - and so it assumed that clause had nothing to do with "this data structure", switched the order of the "and" around, etc...

        My hobby is building Gundam models - one of the most comprehensive review sites for new Gundam kits is in Korean. Believe me, we all try using Google translate or Babelfish on Dalong's site from time to time, but the result is rarely worth the effort.

    • by MSBob (307239)
      ...or war.
    • Re: (Score:3, Interesting)

      by jellomizer (103300)

      The article... Converted from English to Hindi back to english...

      Google is developing software, the first foreign language translation of a phone almost immediately - Hitchhiker Guide's may sound like a fish galaxy.

      Building on existing technology, speech recognition and automatic translation by Google is expected to have a basic system ready in a few years time. If successful, it's finally over 6000 languages in the world can be translated into the interaction between.

      The company has set up an automated sys

  • by Ukab the Great (87152) on Monday February 08 2010, @06:17PM (#31066100)

    Caution: not for use with Hungarian Tobacconists.

  • by qoncept (599709) on Monday February 08 2010, @06:22PM (#31066168) Homepage
    Does anyone use voice recognition software? Here are a couple of my voicemails transcribed by Google Voice:

    Hey man, Hello, this is gonna ask you about Stockton uncle in a missed your call, so, so give well. Okay bye.

    Hey it's me and I for me. Long, My of the day. So Hey Jared, Here doing. If you come for another anti, gimme a call before you go to sleep and stuff, so give me a favor you familiar with it. I love you bye.
    • by rsborg (111459)
      I find Dragon to be much better, simply because they require you to upload your contact list (privacy issue flared up a while back but their new privacy agreement is pretty in-depth and satisfactory for me), so any contacts are not garbled, like they were in GVoice voicemail->text (a feature I love, but is definitely not for the easily confused).

      I'm sure if Google used their contacts list (from your google voice or gmail contacts) privacy freaks would rip them a new one... so not sure what they can do ot

      • Google voice has definitely gotten better at recognizing my name after accepting/rejecting voicemails and filing a feedback request (right near the beginning).

        The thing I have noticed is that it is a little too trigger happy to label the first bit of audio as a standard greeting (hello or something). I have a friend who seems to manage to *never* start a voicemail with an immediate greeting. There is always like a giggle or some background sounds or a last word or two of conversation with a real person.

    • I suspect your friends might have thick regional accents. The voicemail transcriptions are the main reason I use it. Perfect, no. But I'd say it usually only messes up about one or two words per paragraph of text, on average. Except for one friend with a really noticeable Texas accent. It has serious trouble with his.
      • I live in the midwest and my friends (at least in the case of these voicemails) are intelligent and well spoken. They weren't talking as though they were going to have a computer interpret what they were saying, but aside from that they were speaking about as clearly as you could hope for.
    • It doesn't work that bad for people with a clear accent. The problem is getting the software to work as well with Southerners from the US or Scottish people. But I am glad Google is continuing to work on it despite it it not being perfect.Someone has to keep pushing it if it's going to improve.
    • It is common for voice recognition software to confuse homophones, since it doesn't recognize context. But that's ok, I know you _really_ meant to say, "I love you bi!"
    • The problem with voice recognition is inherently a user related problem. All this fluid/casual conversation, regional dialects, muffled voices, uneven, laxidasical cadences not to mention you kids and your fads and lexicon of so-called 'lingo'. If everyone just spoke like robots there'd be no problems. Humph!
    • Speech to text can be so perfectly on the mark sometimes (when you expect it to be way off) and it can be way off on something so simple.

      My girlfriend is a history major and she always handwrites her papers - and because I can get 70 WPM (bursts, not constant) I usually end up typing them up for her. I decided we'd try the Speech to text service on my laptop, with the USB microphone that came with Rockband.

      The paper was on Women in Ancient Rome, so you can imagine that there would be a ton of errors when it

    • Via Stephen Fry... (Score:5, Interesting)

      by Sockatume (732728) on Monday February 08 2010, @07:07PM (#31066748) Homepage

      Stephen Fry offers... [stephenfry.com]

      "Hi, Stephen, it’s Natasha from BBC Newsnight in London. Just to say I’ve sent you two texts. One is to say that we could do it at eleven am your time after the launch, or any time sooner after the launch, or we could do it at midday as we suggested earlier. I, er, if you could text me back about that, and I’ve sent you the details of Skype that you need to do too. If you could give me a call back. Enjoy the launch and I’ll speak to you after that. Thank you Bye."

      I’ve transcribed it from the voicemail sound file that resides online on my inbox on the Google Voice site. All fine. I have also ticked the option for Google Voice to send me a text transcript of any voicemail. Below is their interpretation of Natasha’s message it’s rather endearing how hopelessly wrong the largest company on earth gets it.

      "Hi Stephen. It’s Jeff from BBC needs in nuns. And just to say I sent 80 tax, one, if to say we could do it. I left in i a m your time off to go into any time soon, or the court and full we could grab me today as we suggested at. A. F. I. If you could text me back byebye. I’ve sent you the details of skylights that you need to 3 T if you could give me a call. Bye. Enjoy the loans. I’ll speak to you after that. Thank you. Bye"

      On a more serious note, such transcripts at least allow you to get an idea of the rough content and tone of a message without having to stop and listen to it, a much more concentration-intensive task.

  • with everyone else... Google isn't great at translating and sadly it's pretty much the best. I speak a myriad of languages and Google only does well with Latin based langs and only if they are grammatically perfect.
    You could always figure it out by context but when you get to German or Russian, then you're in trouble. Hell, imagine Mandarin/Cantonese? Pretty soon though, everyone will be able to understand everyone else and I won't be as cool anymore :)

  • I hope they succeed.

  • Tablespoons, by an Apple Newton

    or [allegedly] what happens when you run Jabberwocky through a handwriting recognition program.... :-)

    -----------

    Teas Willis, and the sticky tours
    Did gym and Gibbs in the wake.
    All mimes were the borrowers,
    And the moderate Belgrade.

    'Beware the tablespoon my son,
    The jaws that bite, the Claus that catch.
    Beware the Subjects bird, and shred
    The serious Bandwidth!'

    He took his Verbal sword in hand:
    Long time the monitors fog he sought,
    So rested he by the Tumbled tree,
    Long time the monitors fog he sought,

    And as in selfish thought he stood,
    The tablespoon, with eyes of Flame,
    Came stifling through the trigger wood,
    And troubled as it came!

    One, two! One, two! And through and through,
    The Verbal blade went thicker shade.
    He left it dead, and with its head,
    He went gambling back.

    'And host Thai slash the tablespoon?
    Come to my arms my bearish boy.
    Oh various day! Cartoon! Cathay!'
    He charted in his joy.

    Teas Willis, and the sticky tours
    Did gym and Gibbs in the wake.
    All mimes were the borrowers,
    And the moderate Belgrade.

  • by Anonymous Coward on Monday February 08 2010, @06:33PM (#31066314)

    "Shaka, when the walls fell"

  • by eln (21727)
    Google's voice recognition software combined with Google's translation software. I predict this will cause World War III within hours of going live.
  • Like a f#@%ing deflector dish, then we can solve all the world's problems by reversing the polarity once it's constructed.

  • Can you imagine the possibilities? Unleash this thing at the UN. World War III started on a google phone with the mistranslation: "It was lovely eating your daughter".

  • Make it so!

  • As long as people use it to improve their understanding and not to officially communicate with others, I have no problem with that. It can be somewhat offensive to receive papers that are badly translated. If you want to communicate or sell me something, at least try to learn my language instead of faking it with computer translators. You should see the ridiculous English to French translations sometimes...

    The lack of interest in learning other languages can and will lead to embarrassing situations...

    http: [slashdot.org]

  • Of Google Planets

  • Fantasy (Score:3, Interesting)

    by cdrguru (88047) on Monday February 08 2010, @07:02PM (#31066694) Homepage

    The problem is primarily things like diction. You can "train" someone sitting in front of a computer to speak slowly and clearly with good diction. Fine.

    The problem is the most useful use model for a cell phone translator would be getting a cab or walking into a store. You talk into your phone and it says something to the other person in their language - wonderful, because you have "trained" yourself to speak clearly and slowly with good diction.

    Then the other person mumbles something back at you in their language that neither you or the cell phone can make heads or tails out of. You can't "train" them so it will never work for that.

    From my limited experience, English has its share of strange accents and such but in large measure people can speak with good diction and pronounciation. Lots of non-English languages seem to promote far less clarity and human-to-human it doesn't really impair communication that much. Human-to-machine is a whole different story and we are very far away from being able to do speech recognition with poor pronounciation and poor diction.

  • Got NSA (Score:5, Insightful)

    by dave562 (969951) on Monday February 08 2010, @07:34PM (#31067056) Journal

    This seems like something that the NSA is probably salivating over. Imagine being able to translate intercepts in near real time with accurate voice recognition. I'm sure they already have imagined it. That technology is nothing short of a Manhattan Project for the SIGINT community.

      • Re: (Score:3, Funny)

        And they probably started working on it a decade or two ago, and have a working version now :P

        I can't confirm or deny its existance. But I can offer you one reason for the denial: The FBI. If we admitted we had it, they'd want to use it. The last time they asked for forensic analysis, it was for their cross-dressing commander who wanted to know who had shit on his front doorstep. After intensive analysis, it was determined to be a squirrel. Then they called us because they couldn't identify which squirrel. We've been hesitant to offer our services ever since. We did find the squirrel though. In anot

  • Some anti-snark (Score:3, Interesting)

    by kroyd (29866) on Monday February 08 2010, @07:58PM (#31067232)
    It is really easy to make fun of translate.google.com based on how it translates Chinese to English. This is quite silly IMHO, as Chinese is possibly the hardest language in the world. (Travel around China and you'll find semi-literate taxi drivers, even in the major cities.[*]) This is a good article on why Chinese is hard: http://www.pinyin.info/readings/texts/moser.html [pinyin.info].

    A better example would be say Dutch. Translate the OP from English to Dutch and back to English (i.e. a worst case scenario), and you end up with this:

    "The company has an automatic system for translating texts on computers, sweetened by scanning millions of multilingual websites and documents. Until now includes 52 languages, adding Haitian Creole last week. Google has a system telephone speech recognition that allows users to query websites by speaking commands into their phones instead of typing them in. Now it is working on combining the two technologies to software to understand voice of a caller and translating it into a synthetic equivalent in a foreign language to produce. "

    This is perfectly legible to me, and vastly better than what you got when babelfish was introduced 11 years ago. There is a good TechTalk about the topic at http://www.youtube.com/watch?v=y_PzPDRPwlA [youtube.com] which should be required viewing before making fun of google's machine translation efforts.

    Voice recognition is harder, but for continuous untrained speech recognition google voice is pretty cool - I've gotten some barely intelligible voice messages on my google voice number, and where google voice is sure (i.e. black text) it is 95%+ correct, where it is not sure it is maybe 30% correct, but for another 30% it is not possible to figure out what was said, except when taking context into consideration. Google Voice transcribing a call from a mobile phone is better than what you got with Dragon Dictate 5 years ago even with a good microphone, so it is not unlikely that in a few years it will be better than naive human transcription. Humans will be better at guessing based on context thought.

    Basically, in 5 years the kind of system google is talking about will work good enough to successfully flirt with a french girl (see http://www.youtube.com/user/searchstories [youtube.com]) :P

    [*] This is why you should always bring a mobile phone, and have the number for the place you're going.

      • Re: (Score:3, Interesting)

        A less 19th century European perspective might be that the Chinese mandated the continuity of their literary tradition, and thus words used 2000 years before still needed to be mastered. Of course this was difficult, but this was also in a culture where scholars memorized the Confucian classics as children. The scholar class had the job of studying and passing on literature, just as the Brahmans in India had the difficult task of memorizing the Vedas precisely. Or how Buddhist monks memorized massive sutras
    • by PPalmgren (1009823) on Monday February 08 2010, @06:28PM (#31066240)

      Google is in the business of collecting data and applying it to practical problems. I imagine the voice-to-text will be vastly improved over its generations by users accepting/rejecting the vtt result and them pooling the results data. The same thing could be done for translation from one language to another.

      I see it as crowdsourcing the algorithm accuracy checks among millions of people, allowing them to improve the algo at a much faster rate than they (or their competitors) would otherwise be able to do in a closed testing environment.

      This is all speculating on the fact that google pools results of translations or VTT and whether the user accepts/declines them. I wouldn't be surprised in the slightest if they did.

    • by Idbar (1034346)

      Translation software is difficult to absorb even a simple sentence. What is a mobile phone?

      I enjoy going back and forth between languages. Go to Arabic, Korean, Chinese or sometimes Spanish and back, to realize it still needs some work.

    • For anybody elkse wondering what DRD stands for, I looked it up and it means Discrete Rectal Dongle.
I like work; it fascinates me; I can sit and look at it for hours.