White House innovation white paper advocates MT
The last section of the White House’s report, A Strategy for American Innovation: Driving towards Sustainable Growth and Quality Jobs issued September 20, 2009, presents “the President’s renewed commitment to science, technology and innovation [that] will allow the nation to set and meet ambitious goals that will improve our quality of life and establish the foundation for the industries and jobs of the future.”
The last of these 8 proposals supports:
Automatic, highly accurate and real-time translation between the major languages of the world – greatly lowering the barriers to international commerce and collaboration.
Let’s see how popular MT services translate the sentence into French.
- Google translate:
Automatique, très précises et en temps réel de traduction entre les principales langues du monde – qui réduira considérablement les obstacles au commerce international et la collaboration.
- Reverso:
Traduction automatique, fortement précise et en temps réel entre les langues majeures du monde – énormément baisse des barrières à commerce international et collaboration.
- Systran:
Automatique, traduction fortement précise et en temps réel entre les langues principales du monde – considérablement abaissement des barrières au commerce et à la collaboration internationaux.
- Babelfish:
Automatique, traduction fortement précise et en temps réel entre les langues principales du monde – considérablement abaissement des barrières au commerce et à la collaboration internationaux.
None of these translations come close to being grammatically correct and well-written comprehensible.
Some believe (I don’t) that back translations are an effective way to judge a translation’s quality accuracy. In this sample, Google’s back translation produces the original text even though the translation into French was abysmal. It is because I did not empty my browser’s cache or because the sample phrase is already part of the TAUS Data Association (TDA) language data exchange portal? I do not have the answer, and I am not an MT expert.
The other services’ back translation faithfully transpose their French gibberish into English gobbledygook:
Reverso:: Machine, strongly precise and real-time translation between the major languages(tongues) of the world – enormously lowers(dims) barriers to international trade and collaboration.
Systran: Automatic, translation strongly precise and in real time enters the principal languages of the world – considerably lowering of the barriers to the trade and collaboration international.
Babelfish: Automatic, translation strongly precise and in real-time enters the principal languages of the world – considerably lowering of the barriers to the trade and collaboration international.
If, in the near future, the US government relies on MT to “lower the barriers to international commerce and collaboration”, and then on back translations to gauge the accuracy of the tools used, the risk may well be an increase, not a decrease, in miscommunication.
I guess we human translators won’t be going out of business any time soon!
At least not those who focus on the foreign language writing of translations! When you take a look at Renato’s presentation at the ATA Annual Congress, you start to wonder about the various directions the industry is taking and some of their legal ramifications.
For the record, I just used GT on your Fr output and got back the original phrase too, which suggests to me that they are stored somewhere centrally as a pair of segments, as it were.