Machine Translation – A New Convert :-)

Machine Translation, in particular in the form of free online translations, has been one of the favorite translators’ jokes for quite some time.

But things are changing now, and after many years of ignorance, then scepticism of MT, I must confess I am now slowly but surely edging towards becoming an MT zealot.

Jeff Allen has provided abundant proof that after adequate post-editing (postediting), a “translator using MT produces approximately three times more in a day than the average translator without MT”. Of course to be a good post-editor, you must first be an excellent translator, which means perfect linguistic competency, supreme cultural fluency and sound subject matter expertise.

Here is Jeff Allen’s conclusion after an experiment with MT PE (Machine Translation Postediting): “An average human translator would normally produce 2400 words per day and thus take 3-4 days to translate this translation product job (nearly 8400 words) from scratch. This implementation shows that a fully bilingual subject matter expert, with professional translation experience, combined with excellent mastery of a commercial MT system, can complete the entire translation workflow of the translation job in slightly more than 1 day of work. This clearly shows, for this specific language direction, and on a given MT commercial tool, that the results obtained on production documents by such an individual can attain production rates that are 25%-30% of the time necessary and expected via traditional translation methods.”

Lorena Guerra is another pioneer of the MT PE model. Her master’s thesis at Dublin City University in 2003 was on the subject of comparing human translation with Machine Translation and postediting. Using the ME PE method, she recorded productivity increases of 60% to 70%, even on marketing texts. Lorena Guerra has now started her own ME PE company: Euromix.

A German company, Delta, located in Bonn has been specialized in this field for years.

All this has put the final nail in the coffin of my scepticism 🙂

You can find more info on this topic by visiting the following links (If a link doesn’t work, just copy and paste it in your browser address):

http://www.geocities.com/mtpostediting
http://www.geocities.com/mtpostediting/tausfeb2005refs
http://www.geocities.com/jeffallenpubs/thematic.htm
http://www.multilingual.com/machineTranslation62.htm
http://en.wikipedia.org/wiki/Machine_translation
http://www.languageweaver.com
http://gilbane.com/globalization/machine_translation

You can also take a look at John Yunker’s “Global by Design” blog, specially the article: “The End of Translation As We Know It” ( http://www.globalbydesign.com/blog/2007/11/30/the-end-of-translation-as-we-know-it)

Says Jeff Allen: “The translation and localization industry at large does not like MT technology. MT is regarded as a threat rather than an opportunity.”

Now, I urge you to consider this as a great opportunity: you can still embrace a brilliant postediting career. All it takes is “perfect linguistic competency, supreme cultural fluency and sound subject matter expertise.”

What do you say?

A.M.Sall

P.S. Click here to download your copy of “The Insider Guide to The Strategic Marketing of Translation Services“, the book no translator/interpreter in his/her right mind should even think of going without :-). Or maybe you’d rather take a look at the Free Preview first? This is the book you’ve been waiting for so impatiently: it’s here NOW. So why wait any longer…!?

*******

Subscribe to Translator Power and successfully market your translation/interpreting services either by Email or if you prefer the feed, just look on your left and make your choice!

*********

 

Advertisements

About translatorpower

The Cross-cultural Connector, The Global Mindset Advocate
This entry was posted in Amsall, machine translation, post-editing, post-editor, postediting, posteditor, Translator Power. Bookmark the permalink.

9 Responses to Machine Translation – A New Convert :-)

  1. akuroiwa says:

    I think that it’s still too soon to start doing this. For those who have “perfect linguistic competency, supreme cultural fluency and sound subject matter expertise.” this may be true but there are too many translators who don’t have these skills and will start producing even worse translations. Although translation memory would be something I think would be more helpful I’m not sure that most translators are ready for pure machine translation post editing because they are not sufficiently proficient or are not detail oriented. Machine translations still produce a lot of mistakes and some of them are minute and hard to catch. Regardless, good topic and it is something to watch.

    Arizona Translator

    Like

  2. Sure Akuroiwa, I agree with you 100%. There are indeed prerequisites. All I can say is: those who cannot do it should either concentrate on what they CAN do, or still better, decide to take up the challenge of improving their linguistic skills, and cultural knowledge as well as mastering their subject matter.

    Thanks a lot for your truly remarkable contribution!

    A.M.Sall

    Like

  3. Tamer Elzein says:

    Akuroiwa, Sall, there’s nothing really that I could add to your comments. Very well thought and said 🙂

    But I have to admit that I’m quite astonished by Allen’s and Guerra’s findings. Yes, I do MT-PE frequently, but I never realized that they could save that much time. I’m gonna start monitoring my own MT-PE performance now 🙂

    Great info, Sall. Thanks a lot!

    Like

  4. Pingback: Chaos ex machina

  5. postediting says:

    Tamer,
    The results of projects conducted by Lorena Guerra and myself are at the links below. Please note that the productivity metrics depend on the language pair and on the MT system/software that is used.

    Getting started with Machine Translation. by Jeff ALLEN. In “Guide to Translation” of MultiLingual Computing & Technology, Number 69, Volume 16, Issue 1, January/February 2005. pp 8-12.
    https://216.18.156.115/multilingual/downloads/screenSupp69.pdf

    Postediting: an integrated part of a translation software program. by Jeff ALLEN. In Language International magazine, April 2001, Vol. 13, No. 2, 2001. pp. 26-29.
    http://www.geocities.com/jeffallenpubs/Allen-LI-article-Reverso.pdf

    Case Study: Implementing MT for the Translation of Pre-sales Marketing and Post-sales Software Deployment Documentation. by Jeff ALLEN. In Machine Translation: From Real Users to Research, 6th Conference of
    the Association for Machine Translation in the Americas, AMTA 2004.
    http://www.geocities.com/mtpostediting/Jeff-Allen-AMTA2004-paper_v1.01.pdf

    What is Post-editing? By Jeff Allen. Translation Automation Newsletter, Issue 4. February 2005.
    http://www.geocities.com/mtpostediting/TA_IssueFour.pdf

    Improved Translation Quality with Machine Translation Dictionary Building. by Jeff ALLEN. June 2006. Published by Translatorscafe.
    http://www.translatorscafe.com/cafe/Articles.asp?ArtID=59

    Human Translation versus Machine Translation and Full Post-Editing of Raw Machine Translation Output. by Lorena GUERRA. 2003. Master’s Thesis. Dublin City University.
    http://www.geocities.com/mtpostediting/lorena-guerra-masters.pdf

    Like

  6. postediting says:

    Akuroiwa,

    You stated that translators with “perfect linguistic competency, supreme cultural fluency and sound subject matter expertise” might be able to use this. I would like to come back to my original statement (a citation from one of my articles) “fully bilingual subject matter expert, with professional translation experience, combined with excellent mastery of a commercial MT system”
    What this means is that for a user to be proficient for the task of using postediting to publish translatored materials, 1) they need to fully understand both languages. This is to contrast the use of MT systems with the approach of content gisting to only “understand” texts in other languages (not spoken/written by the user) in the user’s language. 2) The translation post-editor needs to be a subject matter expert. The reason for this statement is that lack of knowledge of the vocabulary and terminology usage and coverage of these throughout the materials that are being translated will end up in wasted time. Someone who masters the content already will be much more productive that someone who is starting out with the set of content and reference materials for the first time. And 3), the mastery of the commercial of the MT system. Someone who has not taken the time to learn the tool, and the features in it to become productive, will simply spend all their time trying to figure out the tool, rather than being productive in the translation postediting task with the tool to do it.

    In other words, if a monolingual speaker expects to conduct translation publication-level postediting on texts that they are not familiar with, and very little knowledge of the MT system and features, then they need to find someone else to do it for them.

    Like

  7. Pingback: Interesting comments! « Translator Power

  8. sociolingo says:

    Hi, I was interested to read this discussion. The following freeware may be of interest to you:
    http://www.ethnologue.com/tools_docs/freeware2_intro.asp
    Particularly:
    Carla Studio 2.9 A suite of tools that are integrated to allow adaptation of texts from one language to another related language using rules based on the parsing of the original data.

    There is another collection at:
    http://www.ethnologue.com/tools_docs/freeware1bte_intro.asp

    You can download the complete disks or specific programmes from the site.

    Like

  9. Thanks a lot, sociolingo. I’ll check it out!

    Amadou M. Sall

    Like

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s