Results 1 to 8 of 8
  1. #1
    Thailand Expat CaptainNemo's Avatar
    Join Date
    Sep 2009
    Last Online
    18-07-2020 @ 11:25 PM
    Location
    in t' naughty lass
    Posts
    5,525

    Computer-aided language learning thread

    Google uses neural networks to translate without transcribing


    Daily news
    4 April 2017
    Google uses neural networks to translate without transcribing

    I hear youFloresco Images/Getty

    By Matt Reynolds
    Google’s latest take on machine translation could make it easier for people to communicate with those speaking a different language, by translating speech directly into text in a language they understand.
    Machine translation of speech normally works by first converting it into text, then translating that into text in another language. But any error in speech recognition will lead to an error in transcription and a mistake in the translation.
    Researchers at Google Brain, the tech giant’s deep learning research arm, have turned to neural networks to cut out the middle step. By skipping transcription, the approach could potentially allow for more accurate and quicker translations.
    The team trained its system on hundreds of hours of Spanish audio with corresponding English text. In each case, it used several layers of neural networks – computer systems loosely modelled on the human brain – to match sections of the spoken Spanish with the written translation. To do this, it analysed the waveform of the Spanish audio to learn which parts seemed to correspond with which chunks of written English. When it was then asked to translate, each neural layer used this knowledge to manipulate the audio waveform until it was turned into the corresponding section of written English.
    Corresponding patterns

    “It learns to find patterns of correspondence between the waveforms in the source language and the written text,” says Dzmitry Bahdanau at the University of Montreal in Canada, who wasn’t involved with the work.
    After a learning period, Google’s system produced a better-quality English translation of Spanish speech than one that transcribed the speech into written Spanish first. It was evaluated using the BLEU score, which is designed to judge machine translations based on how close they are to that by a professional human.
    The system could be particularly useful for translating speech in languages that are spoken by very few people, says Sharon Goldwater at the University of Edinburgh in the UK.
    International disaster relief teams, for instance, could use it to quickly put together a translation system to communicate with people they are trying to assist. When an earthquake hit Haiti in 2010, says Goldwater, there was no translation software available for Haitian Creole.
    Goldwater’s team is using a similar method to translate speech from Arapaho, a language spoken by only 1000 or so people in the Native American tribe of the same name, and Ainu, a language spoken by a handful of people in Japan.
    Rare languages

    The system could also be used to translate languages that are rarely written down, since it doesn’t require a written version of the source language to produce successful translations.
    Until it is tested on a much larger dataset, it’s hard to tell how the new approach really compares with more conventional translation systems, says Goldwater. But she thinks it could set the standard for future machine translation.
    Some services already use machine translation to let people who speak different languages have conversations in real time. Skype introduced a live speech-to-text translation feature in 2014 and now supports nine languages, including Mandarin and Arabic as well as the most common European languages. But like other existing translation methods, Skype’s transcribes speech into text before translating that text into a different language.
    And text translation service Google Translate already uses neural networks on its most popular language pairs, which lets it analyse entire sentences at once to figure out the best written translation. Intriguingly, this system appears to use an “interlingua” – a common representation of sentences that have the same meaning in different languages – to translate from one language to another, meaning it could translate between a language pair it hasn’t explicitly been trained on. The Google Brain researchers suggest the new speech-to-text approach may also be able to produce a system that can translate multiple languages.
    But while machine translation keeps improving, it’s difficult to tell how neural networks are coming to their solutions, says Bahdanau. “It’s very hard to understand what’s happening inside.”
    Journal reference: arXiv, DOI: arxiv.org/abs/1703.08581
    https://www.newscientist.com/article...-transcribing/

    This is interesting, because technology is approaching that point where language-learning without computers may become a novely like paper books and vinyl. That's a long way off for Thai, Google Translate is still useless at that.

  2. #2
    Harbinger of Doom

    Join Date
    Apr 2015
    Last Online
    @
    Posts
    1,622
    That's a long way off for Thai, Google Translate is still useless at that.
    It's actually not too bad now. Here's a sample from an article on Thaipublica:

    อีไอซีมองว่าคณะกรรมการนโยบายการเงินจะยังคงอัตราดอก เบี้ยนโยบายไว้ที่ 1.5% ถึงแม้อัตราเงินเฟ้อทั่วไปปี 2017 จะเพิ่มขึ้น เนื่องจากอัตราเงินเฟ้อเพิ่มขึ้นจากระดับต่ำและยังอย ู่ในกรอบเป้าหมายของธนาคารแห่งประเทศไทยที่ 2.5% ± 1.5% และยังต้องคงอัตราดอกเบี้ยอยู่ในระดับต่ำเพื่อสนับสน ุนการฟื้นตัวของเศรษฐกิจภายในประเทศที่เต็มไปด้วยควา มเสี่ยงต่อไป
    EIC views the MPC will maintain its policy rate at 1.5%. Although headline inflation in 2017 is expected to increase as inflation rises from a low level and remains within the Bank of Thailand's target of 2.5% ± 1.5% and still need to maintain low interest rates to support the recovery of the domestic economy at risk.
    The end is a bit off (hasn't quite captured the widespread sense of risk) but on the whole, that's pretty impressive.

    ---

    The forum software still doesn't render Thai properly.

  3. #3
    Thailand Expat CaptainNemo's Avatar
    Join Date
    Sep 2009
    Last Online
    18-07-2020 @ 11:25 PM
    Location
    in t' naughty lass
    Posts
    5,525
    It's definitely got better, but it's still quite shite. With a lot of languages you have paste back the translation to check it, and tweak it with what you know. I notice that usually when you type "I" you get "chun", but for some phrases you get "pom", so when you have no Thai keyboard, and you are copying and pasting it (a tip for ShrewedPunter and his messaging hobby) you have to write wonky bits of English to assemble what you know makes sense in Thai.
    It's a nuisance that they don't allow options for words like if you were typing pinyin or something... I think the user interface needs rethinking a bit to handle non-latin script languages that have different ways of building words.

  4. #4
    Newbie
    Join Date
    Aug 2017
    Last Online
    07-08-2017 @ 01:22 PM
    Posts
    6
    yeah it still doesn't work good for less popular languages

  5. #5
    Member

    Join Date
    Aug 2017
    Last Online
    @
    Location
    Shanghai
    Posts
    386
    Quite an interesting article. Hopefully they can figure it out for rare languages to work.

    I have learned quite recently of many apps on my iphone where I can take a picture of something in Mandarin and it automatically translates it to English or other languages. You can choose what language you want it to translate it to. I see Thai is not one of the languages on the list. This pic app doesn't work for everything, but it is quite good. I took a pic of my washing machine labelling in Mandarin and now I know what it all means. If I see a menu in Chinese I can take a pic and have it translated into English.

  6. #6
    Member

    Join Date
    Jan 2017
    Last Online
    @
    Posts
    520
    I don't reckon it'll be more than five years until you have a "translate to" option before making a phone call.

  7. #7
    Thailand Expat CaptainNemo's Avatar
    Join Date
    Sep 2009
    Last Online
    18-07-2020 @ 11:25 PM
    Location
    in t' naughty lass
    Posts
    5,525
    Takes the fun out of using languages to chat up women... unless you get the translations wrong on purpose

  8. #8
    Member

    Join Date
    Jan 2017
    Last Online
    @
    Posts
    520
    "I meant do you like to flirt, not do you like to squirt"

    "Honestly"


Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •