Isn't Google Translate just another translation memory? Thread poster: Diana Singureanu
|
I had an argument with a colleague recently who states that it is much more complex (I agree the memory corpus with segments in various language combination must be very wide) but surely this is the basic principle.... Also has anyone ever heard of a software which interprets live - automatically? Basically there is a language recognition tool involved as well but again it's based on translation memory... I would really appreciate your replies.
[Edited at 2014-02-13 07:38 GM... See more I had an argument with a colleague recently who states that it is much more complex (I agree the memory corpus with segments in various language combination must be very wide) but surely this is the basic principle.... Also has anyone ever heard of a software which interprets live - automatically? Basically there is a language recognition tool involved as well but again it's based on translation memory... I would really appreciate your replies.
[Edited at 2014-02-13 07:38 GMT] ▲ Collapse | | | Of course it is! | Feb 12, 2014 |
The difference is that YOUR translation memory, viz, from MemoQ, Trados, WordFast, etc. will have been created by YOU. You know your competence as a translator, or at least you should. Google Translate's giant multilingual TMs have been collectively created by millions of unidentified people of all shades, ranging from wannabes and self-claimed "tranzlaters" to some world-class masters of professional translation who published bilingual stuff on the web. You never know ... See more The difference is that YOUR translation memory, viz, from MemoQ, Trados, WordFast, etc. will have been created by YOU. You know your competence as a translator, or at least you should. Google Translate's giant multilingual TMs have been collectively created by millions of unidentified people of all shades, ranging from wannabes and self-claimed "tranzlaters" to some world-class masters of professional translation who published bilingual stuff on the web. You never know what you'll get. Regarding your other question, you may visit YouTube and request subtitles based on voice recognition and automatic translation. The result is usually hilarious. What's missing there - YouTube has no use for it - is to capture that translated text and shoot it through some text-to-speech program (they exist since the days of Windows 3.1). ▲ Collapse | | | Erik Freitag Germany Local time: 23:43 Member (2006) Dutch to German + ... Of course it isn't | Feb 12, 2014 |
Dear Diana, I'm afraid your colleague is absolutely right here. Although Google Translate may look like a simple TM system to the unitiated eye, it works fundamentally different. If you're interested to learn how it works, you could start reading here: http://en.wikipedia.org/wiki/Statistical_machine_translation It has its uses, but certainly not for ... See more Dear Diana, I'm afraid your colleague is absolutely right here. Although Google Translate may look like a simple TM system to the unitiated eye, it works fundamentally different. If you're interested to learn how it works, you could start reading here: http://en.wikipedia.org/wiki/Statistical_machine_translation It has its uses, but certainly not for a professional translator.
[Bearbeitet am 2014-02-12 21:49 GMT] ▲ Collapse | | | John Fossey Canada Local time: 17:43 Member (2008) French to English + ... No, it's not a translation memory | Feb 12, 2014 |
Google translate is not simply a translation memory. It has a huge "corpus" of translations but it uses complex statistical calculations to determine that when a phrase or series of words appears in the source text in a certain position and under particular conditions then there is a high probability that another certain phrase or series of words will appear in the target text in a certain place and order. Each word and each phrase, along with its position, is weighted statistically. ... See more Google translate is not simply a translation memory. It has a huge "corpus" of translations but it uses complex statistical calculations to determine that when a phrase or series of words appears in the source text in a certain position and under particular conditions then there is a high probability that another certain phrase or series of words will appear in the target text in a certain place and order. Each word and each phrase, along with its position, is weighted statistically. A translation memory is much simpler. It just compares the degree of match between the source text in the translation and the source text in the translation memory, and if the level of match exceeds a certain threshold it produces the entire translation segment. ▲ Collapse | |
|
|
Phil Hand China Local time: 05:43 Chinese to English Similar in some ways | Feb 13, 2014 |
GT works on a principle that is similar to a TM. For example, it can't invent any language that isn't in its corpus. But it is more granular than most TMs: it works at the level of the word and the phrase, not the sentence. And it applies an extra layer or two of processing, so instead of just telling you how close a phrase is, it works out a best guess for the context. I believe there is a Japanese device which claims to do live interpreting of tourist phrases. And the US military ... See more GT works on a principle that is similar to a TM. For example, it can't invent any language that isn't in its corpus. But it is more granular than most TMs: it works at the level of the word and the phrase, not the sentence. And it applies an extra layer or two of processing, so instead of just telling you how close a phrase is, it works out a best guess for the context. I believe there is a Japanese device which claims to do live interpreting of tourist phrases. And the US military I think has a device which does English-Arabic or English-Pashtu for use in its Middle Eastern wars. Not sure if that one is actually working yet. But I assume they are extremely limited in the range of language they can recognise and translate. ▲ Collapse | | | google translation is not TM | Feb 13, 2014 |
TM must be edited and checked by human, unchecked materials are not reliable. | | |
Diana Singureanu wrote: Also has anyone ever heard of a software which interprets live - automatically? Basically there is a language recognition tool involved as well but again it's based on translation memory... And, I doubt anyone will ever hear. (About a reliable tool of that kind--perhaps when robots start walking in the street, doing different chores, which I personally doubt will ever happen). Plus, who would need it?
[Edited at 2014-02-13 11:43 GMT] | | | Alex Lago Spain Local time: 23:43 English to Spanish + ...
A TM is basically a database in which one of the fields stores the original segment and another field stores the translation, plus some fields with the date and name of person who created the entry. GT is not a database. Yes there are programs that interpret (I'm not saying... See more A TM is basically a database in which one of the fields stores the original segment and another field stores the translation, plus some fields with the date and name of person who created the entry. GT is not a database. Yes there are programs that interpret (I'm not saying they're any good), for example https://itunes.apple.com/us/app/voice-translator/id640034983?mt=8 And even specific devices http://buysigmo.com/ ▲ Collapse | | | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » Isn't Google Translate just another translation memory? TM-Town | Manage your TMs and Terms ... and boost your translation business
Are you ready for something fresh in the industry? TM-Town is a unique new site for you -- the freelance translator -- to store, manage and share translation memories (TMs) and glossaries...and potentially meet new clients on the basis of your prior work.
More info » |
| CafeTran Espresso | You've never met a CAT tool this clever!
Translate faster & easier, using a sophisticated CAT tool built by a translator / developer.
Accept jobs from clients who use Trados, MemoQ, Wordfast & major CAT tools.
Download and start using CafeTran Espresso -- for free
Buy now! » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |