The aim of their collaboration is to achieve machine-based translation between the languages of the European Union so that comprehensible texts are achieved for as many language combinations as possible. Two of the EU-funded research projects are being led by the Saarbrücken computer linguist Josef van Genabith.
[…] “Teaching a computer to understand all these grammatical nuances and to translate them correctly into another language is exceptionally difficult,” says Josef van Genabith, Professor of Translation-Oriented Language Technologies at Saarland University and a Scientific Director at the German Research Center for Artificial Intelligence (DFKI). His team is therefore following a different path. The computers are not fed with grammar rules and linguistic details, but are taught to recognize patterns in huge text repositories and to learn from them. In the computer linguistics community, this approach is referred to as “deep learning”. The method recently made headline news when Google used the technique to beat one of the world’s top Go players. More.
See: Science Codex
Subscribe to the translation news daily digest here. See more translation news.
Comments about this article
Canada
Local time: 18:36
Member (2008)
French to English
+ ...
So the result is grammatically incorrect text that doesn't follow the linguistics of the target language.
Isn't attempting "to recognize patterns in huge text repositories and to learn from them" exactly what Google Translate has attempted for years, with dubious su... See more
So the result is grammatically incorrect text that doesn't follow the linguistics of the target language.
Isn't attempting "to recognize patterns in huge text repositories and to learn from them" exactly what Google Translate has attempted for years, with dubious success? What is the "different path"? ▲ Collapse
Germany
Local time: 00:36
English to German
+ ...
I love this statement "Teaching a computer to understand all these grammatical nuances and to translate them correctly into another language is exceptionally difficult".
And i agr... See more
I love this statement "Teaching a computer to understand all these grammatical nuances and to translate them correctly into another language is exceptionally difficult".
And i agree, it is difficult but this is not what they are trying to do. They are not teaching the computer anything, all they apply is very advanced math, not less but also not more. And this will just result in a statistical MT sytem on steroids. Translation requires more than "pattern recognition" it requires understanding and domain/world knowledge.
[Edited at 2016-07-13 21:07 GMT] ▲ Collapse
Germany
Local time: 00:36
Member (2009)
English to German
+ ...
Grammatically incorrect "translations" are bound to back-fire. Due to unclear translations, conveying perhaps the wrong context had been the cause of wars before.
[Edited at 2016-07-14 06:10 GMT]
United States
Local time: 18:36
Russian to English
+ ...
Local time: 01:36
Romanian to English
+ ...
The sad thing is that while they probably won't succeed in what they propose, this could popularize incorrect translation (if it's more or less intelligible, albeit linguistically deplorable) - the same thing they are already doing anyway, you just have to take a look at EU texts and see how unnatural they sound with artificial, lifeless expressions I generally just call EU-speak.
[Edited at 2016-07-14 09:47 GMT]
Spain
Local time: 00:36
Spanish to English
+ ...
It looks like these wily researchers want to take advantage of the collective angst of EU accountants, worried about the vast sums apparently spent on translation and interpretation by the behemoth and desperately seeking any kind of money-saving solution. Even if it involves doling out grants for pointless tail chasing "research" projects like this.
Denmark
Local time: 00:36
Member (2003)
Danish to English
+ ...
With all due respect for the practically infinite numbers of possible moves in Go, language is far more complex.
I don't know the subtleties and am a complete beginner as far as Go is concerned, but there are two dimensions, 19 x 19 squares on the board and pieces in only two colours = two players.
•
EU legislation may be standard and predictable to a certain extent, but there are 24 languages, and infinite numbers of players with far more rules and options than a game of Go.
If EU legislation becomes so predictable that machines can make a good job of translating it, then there will be so little new content that the legislation will probably be superfluous.
Anything as complex as that calls for understanding, not just qualified guessing.
To report site rules violations or get help, contact a site moderator:
You can also contact site staff by submitting a support request »
This discussion can also be accessed via the ProZ.com forum pages.