| Subcribe via RSS 

AMTA Conference San Diego

November 13th, 2012 | by LTD

Reviewed by: Rubén de la Fuente

I attended the first day of AMTA conference, which brings together MT vendors, LSPs and end-clients. I was pleased to see attendees from ATA conference as well. ATA and AMTA have decided to collocate their conferences in order to build bridges between both communities. It was encouraging to see people taking advantage of the opportunity.

I missed ATA President Caitlin Walsh’s keynote on Literature and Poetry, but I got on time for Luis von Ahn’s on Duolingo. Duolingo seeks to “translate the web for free” by giving language learners sentences to translate. Here’s a TED video explaining Duolingo in more detail. Although it could be seen as a threat for translators, I don’t think it will be: Duolingo will not charge for translation only if content is creative commons-licensed (Wikipedia and the like) or if there’s no required deadline for translations. Most commercial projects will not meet either of these characteristics.

There was a panel with SDL, CA, Adobe and PayPal, where innovative uses beyond MT and post editing were presented. The key take-away for me was that MT can be used to validate the translatability of the source (spotting ambiguities or internationalization issues) and make sure the source content is more translation-friendly.

Omnilingua presented its findings after analyzing its MT projects in terms of productivity and quality (using SAE J4250 metric). MT will actually decrease productivity when segments have very poor quality, so translators waste time deciding if anything can be used and then deleting the whole segment. If these segments can be filtered out using MT’s confidence scores, so that translators do not need to waste time discarding them, productivity gains will be even higher and translators’ frustration with MT will be reduced. Also, it’s very important for translators to give precise detailed feedback on what can be improved in MT, to the benefit of both parties.

Last, PayPal presented on its experience with 3 types of MT technology: rule-based, statistical and hybrid. People often ask which technology is better. The key message from PayPal’s presentation was that all three technologies have strength and weaknesses, and anyone considering MT should look at all three to find their best fit, depending on their budget, their existing linguistic assets and their know-how.

The AMTA conference went on until Thursday. For people wanting to know more, I’d suggest checking out GTS’s blog entries listed below or browse the conference papers.

Rubén R. de la Fuente has a BA in translation and interpreting from the University of Granada. He has over 10 years of experience in localization in various capacities, including as a freelance and in-house translator, reviewer, project manager, and machine translation specialist. He is currently taking a graduate course on computational linguistics. He has taught several courses and workshops about translation tools for the Universidad Alfonso X and organizations such as the Institute of Localisation Professionals, ProZ, and ecpdwebinars.co.uk. He has written articles on translation tools for ATA’s Language Technology Division. You can reach Rubén at rubo@wordbonds.es.

Leave a Reply