Technical forums » Machine Translation (MT) »
Pages in topic:   < [1 2 3]
Machine Translation - anyone using it successfully?
Thread poster: gianfranco
Jeff Allen
Jeff Allen  Identity Verified
France
Local time: 21:57
Multiplelanguages
+ ...
MT corporate user not praising their vendor, but their own in-house expertise Nov 25, 2009




b) MT only good for related languages. There have been successful implementations of non-related languages. Take for example Symantec which has implemented it not just between EN and other FIGES languages, but also with Japanese and Chinese:

first page of PDF under the topic "Solutions" of:
http://www.systran.co.uk/systran/corporate-profile/translation-case-studies/symantec

section 4.3 of:
http://www.mt-archive.info/MTS-2009-Roturier.pdf

slide 3 of
http://www.mt-archive.info/Translingual-Europe-2009-Roturier-ppt.pdf

[...]

[/quote]

Tim Drayton wrote:

The links you have quoted are all to pieces written in praise of a certain company's software on that company's own website. That is hardly the most objective source of information. I have recently glanced at quite a few academic artiles about developing statistical MT translation software to translate from Turkish into English, and they all seem to be suggesting that a significant number of transformations need to be applied to a sentence before it can be submitted to statistical processing. These are all tentative suggestions at the moment and I do not feel that anybody is close to cracking this particular 'nut' soon.


Tim,

It's not that Symantec is trying to praise the MT vendor, but that Symantec is showing how well they themselves have created an in-house team to bring together a variety of commercially sold 3rd party tools to be one of the most complex translation workflows, because no Language Service Provider at the time was capable and willing to do what they wanted to achieve. They are partnering with university programs to fund research of students focused in translation automation.

Here are a few more links to translation and L10n industry events where the localization team of that corporation continues to present.

http://www.translationautomation.com/user-cases/symantecs-localization-revolution.html

http://www.lisa.org/Final-Program.851.0.html#c536

http://www.localizationworld.com/lwber2007/programSynopses.htm
B1: Translation Automation 2.0 – A Shopping List

http://www.localizationworld.com/lwber2008/program2.php
look up Fred Hollowood
http://www.localizationworld.com/lwber2008/speakers.php#fHollowood

http://www.dcu.ie/salis/newsletter1107.pdf
bottom of page 6

http://www.iti.org.uk/ice/uploadedFiles/108_Johann-Roturier-Guest-Lecture-Imperial.doc


The proof is in the pudding. It has been successful. The people are still on the project, and still making presentations about the project after several few years. Good sign compared with the history of MT implementations where projects start and then there is no more external communication about them.

If you take note at the vendor who provided the MT solution and then check my Proz Profile, it will make sense that I know the project mentioned above quite well.
But I only cite presentations that are publicly available.

Tim Drayton wrote:
I think it is obvious to anyone with a good grounding in language and linguistics and at least a passing knowledge of the huge diversity that exists in natural languages around the world that it is going to be much easier to develop MT software between cognate languages than between those with radically different structures.


Check out my Proz profile again, not just the education section, but also the advisory board section.

MT can be implemented from a 1) computational linguistics perspective, from a 2) linguistics perspective, and also from that of a 3) professional translator. And it is good to see that several of the MT vendors are hiring more experts who are from backgrounds 2 and 3, and not just 1.

Tim Drayton wrote:
developing statistical MT translation software to translate from Turkish into English,


A good article I read on Turkish language computing a few years ago, and mentioned to others at an MT conference in 2003 is:

Computing in Turkish
http://www.multilingual.com/articleDetail.php?id=613
October/November, 2002


and for Turkish, I partially replied to this in:
about MT Turkish>English
http://www.proz.com/post/1239887#1239887

and would supplement with the info that several MT vendors were present at the AMTA98 panel on "The Forgotten Majority: Neglected Languages"
http://www.isi.edu/natural-language/conferences/amta98/program.html
If you are ready to put down 300K, 400K, 500K USD then, "Have money, will travel", and any MT vendor would be ready to kick-off the project. Without dedicated funding, then the other regular market analysis factors come into play on where they choose to make the investment on enhancements and development.

The past (5 years back) and forecasted (5 years forward) availability of languages for MT systems were mentioned in a series of formal surveys conducted in 1999 and 2000. See:

Survey of Language Engineering needs: a Language Resources perspective. In proceedings of the Second International Conference on Language Resources and Evaluation (LREC2000), 31 May - 2 June 2000, Athens, Greece.
http://linguistlist.org/pubs/papers/browse-papers-action.cfm?PaperID=7066

Survey on multilingual issues, and the evolution of languages in speech and machine translation products. Report conducted and written within the framework of the Language Resources Production and Packaging project (LRsPP - LE4-8335). European Language Resources Distribution Agency & European Commission.
http://linguistlist.org/pubs/papers/browse-papers-action.cfm?PaperID=7085

These surveys were based on the general outlook at that time. Google and Microsoft started making their free online systems available afterward and somewhat changed the landscape of the MT free offer with free statistical-based MT. But if you choose to use the free carrot, which is the result of industry internal research by such companies, then it comes along with all the conditions and constraints of such freebie systems.

Jeff


[Edited at 2009-11-26 06:30 GMT]
 
Kirti Vashee
Kirti Vashee  Identity Verified
United States
Local time: 12:57
Successful use of MT at Lexis Nexis Jan 16, 2010

In the world of patents, the sheer volume of material that needs to be translated require that some kind of automation be used. Clearly the expectation for MT is not to produce what a reasonably competent human would, but rather enough info. Better than the gist quality that one would get from the free MT translation services. As thousands of patents are filed across the globe on a monthly basis - MT is necessary.

This MT output is often used to decide and identify the most interes... See more
In the world of patents, the sheer volume of material that needs to be translated require that some kind of automation be used. Clearly the expectation for MT is not to produce what a reasonably competent human would, but rather enough info. Better than the gist quality that one would get from the free MT translation services. As thousands of patents are filed across the globe on a monthly basis - MT is necessary.

This MT output is often used to decide and identify the most interesting material and what should be sent to a human for a real translation.

Lexis Nexis currently uses highly customized Rule based systems for European languages and will increasingly shift to SMT because translation quality continues to improve.

Here are some video presentations from the Localization & Translation Conference in Thailand December 2009 ( a joint LISA, ProZ, RTI and Asia Online event.)

Implementing large scale Machine Translation in Patent Information by Andrew Ruffner CTO Lexis Nexis on comparison of Japanese Patent MT systems http://dotsub.com/view/159ce97c-dbd4-4d6a-90c2-427a3a3e755f

The On-going Evolution of the Localization Business by Renato Beninatto CEO Milengo where he talks about the increasing use of MT by translators
http://dotsub.com/view/7da1f3a0-4df2-45a4-b62d-99434c2cf75f

Beyond Machine Translation: Collaboration, Integration, Quality, Change and Jobs – Dion Wiggins on Asia Online Vision which clearly states and warns that the best MT systems need intensive human translator feedback and that MT without human feedback can produce garbage
http://dotsub.com/view/727cdacf-9653-40ef-b6f3-6145ca107db0


[Edited at 2010-01-16 17:22 GMT] ▲ Collapse
 
Pages in topic:   < [1 2 3]


To report site rules violations or get help, contact a site moderator:

Moderator(s) of this forum
Mahmoud Akbari[Call to this topic]
Prachya Mruetusatorn[Call to this topic]

You can also contact site staff by submitting a support request »

Machine Translation - anyone using it successfully?






Wordfast Pro
Translation Memory Software for Any Platform

Exclusive discount for ProZ.com users! Save over 13% when purchasing Wordfast Pro through ProZ.com. Wordfast is the world's #1 provider of platform-independent Translation Memory software. Consistently ranked the most user-friendly and highest value

Buy now! »
Trados Business Manager Lite
Create customer quotes and invoices from within Trados Studio

Trados Business Manager Lite helps to simplify and speed up some of the daily tasks, such as invoicing and reporting, associated with running your freelance translation business.

More info »