The Daily Distraction

How weird is this? And how unsurprising that they haven’t gotten it to work properly yet?

It’s the next best thing to a Babel fish

26 October 2006

Celeste Biever

Imagine mouthing a phrase in English, only for the words to come out in Spanish. That is the promise of a device that will make anyone appear bilingual, by translating unvoiced words into synthetic speech in another language.

The device uses electrodes attached to the face and neck to detect and interpret the unique patterns of electrical signals sent to facial muscles and the tongue as the person mouths words. The effect is like the real-life equivalent of watching a television show that has been dubbed into a foreign language, says speech researcher Tanja Schultz of Carnegie Mellon University in Pittsburgh, Pennsylvania.

Existing translation systems based on automatic speech-recognition software require the user to speak the phrase out loud. This makes conversation difficult, as the speaker must speak and then push a button to play the translation. The new system allows for a more natural exchange. “The ultimate goal is to be in a position where you can just have a conversation,” says CMU speech researcher Alan Black.

In October 2005 Schultz and her colleague Alex Waibel demonstrated the first automatic translator that could pick up electrical signals from face and throat muscles and convert them into text or synthesised speech – a technique called sub-vocal speech recognition. This ran on a laptop and translated Mandarin Chinese to English or Spanish, but it could only translate around 100 words, each of which had first to be spoken into the system by the user, to “train” it on their voice.

Now the team has developed a system that can recognise a potentially limitless lexicon. Their secret is to detect not just words but also the phonemes that form the building blocks of words. The system then uses these to reconstruct the word. To translate from English to another language, the user only has to train the system on the 45 phonemes used in spoken English.

The researchers use software that has been taught to recognise which phonemes are most likely to appear next to each other and in what order. When it encounters a string of phonemes it is unfamiliar with or has only partially heard, it uses this knowledge to come up with a range of sequences that make sense given the surrounding phonemes and words, assigns a probability to each one, and then picks the one with the highest probability.

The system still has some way to go. Faced with a sequence of words it has never heard before, it picks the right phoneme sequence only 62 per cent of the time. This nevertheless ranks as “a very significant achievement” according to Chuck Jorgensen, who is working on using sub-vocal speech recognition to control robots at NASA’s Ames Research Center in Moffett Field, California. “This is showing that the technology is really within reach.”

Schultz’s team plan to attach the phoneme recognition software to their prototype Spanish or German translators, once they have improved its accuracy.

From issue 2575 of New Scientist magazine, 26 October 2006, page 32

- http://www.newscientisttech.com/article/mg19225755.800-its-the-next-best-thing-to-a-babel-fish.html

(Yes, Max, I know I have no life and spend far too much time reading random things on the Internet. But it’s a good procrastinatory tool and means that I can avoid studying)

…but that’s OK.

Here’s what I actually wanted to  put up:

Machine Translation: I’m Sick of Waiting
ARTICLE DATE:  09.18.06

By  John C. Dvorak
The way I see it, if computers can now play a credible world-class game of chess, then they should be able to translate complex sentences written in the world’s major languages. They should be able to translate to and from English, to and from French, and to and from Russian. I eventually expect a translation to and from Chinese and Japanese, too. Exactly what’s the hangup?

ADVERTISEMENT We have the computing power to make this work, so why don’t governments all demand it? Throw $10 billion at the problem, and I bet it is resolved sooner rather than later. $10 billion is less than the cost of one month of the Iraq war, just for comparison.
My French has been in decline since 1973, but I sure know enough to find machine French-to-English translations to be an abomination. For example, with rare exceptions, if you go to a wine site to find out about the latest conditions in Bordeaux, these systems will invariably translate the word chateau as “castle,” despite that Americans (and most English-speaking nations) use the word “chateau” as such. And, in fact, it is always used when referring to a Bordeaux winery such as Chateau Margaux. That’s the name of the place.

For some unknown reason, no translation system can understand this simple fact. Is this rocket science? It’s not the Castle of Margaux or Margaux Castle. How hard is this? It gets even worse if the name of the place itself can be translated. Chateau Latour comes back as the Tower Castle or Castle Tower. What good is it to read that “In big news this week Castle Tower was sold to French investors?” Wha? Huh?

This is just one example, but to me, since it is so obvious and brain-dead, it epitomizes the problem. Apparently nobody is really trying very hard. THAT’S the problem.—Continue reading…

Let’s consider an example from Sudouest , an excellent regional newspaper in France. It has a blog at sudouest.com, and I decided to cut a chunk of text and run it through various systems. Here’s the text.

Vous avez essayé d’obtenir une bourse d’étude mais celle-ci, fonction des revenus de vos parents, vous a été refusée. Vous trouvez le cas injuste? Comment cela risque-t-il d’influer sur votre scolarité? Partagez votre expérience sur ce blog en utilisant le lien “Commentaire” ci-dessous..com

A reasonable translation is as follows:

You tried to obtain a scholarship, but it was refused because of your parents’ income. Did you think this was unfair? Tell us how you think it would affect your schooling and tell us your experiences by adding a comment to the blog. Click on the comment button below.

Here are the various results from the online translation systems:

Babelfish:
You tried to obtain a grant but this one, function of the incomes of your parents, was refused to you. You find the case unjust? How that is it likely to influence your schooling? Share your experiment on this blog by using the bond “Comment” below.

Grade: D. That should be experience, NOT experiment. And bond should be link or button. Grammar screwed up.

Freetranslation.com:
You tried to obtain a study stock market but this one, function of the income of your parents, you was refused. You find the unfair case? How that does it risk influencing on your schooling? Divide your experience on this blog while using the link “Commentary” here-under. com

Grade: F. Just horrible—and wrong! (The site offered a human translation of these few sentences for $50.)

PROMT-Online:
You tried to acquire a grant but this one, function of the incomes of your parents, was refused you. You find unfair case? How does it risk having an influence on your schooling? Share your experience on this blog by using link “Comment” below com

Grade: B. You can actually make sense of this. It’s the clear winner here.

WorldLingo:
You tried to obtain a grant but this one, function of the incomes of your parents, you was refused. You find the case unjust? How that is it likely to influence your schooling? Share your experiment on this blog by using the bond “Comment” below.COM

Grade D: Similar errors to the Babelfish translation with additional errors. What does it take to use the proper noun “were” instead of “was?” You was? Please.

The computer revolution began a half-century ago. We should have been able to solve this problem by now. What we need is government resolve, because private industry can’t seem to manage it.
–http://www.pcmag.com/article2/0,1759,2017280,00.asp

 I think part of the problem might be that the people who work in the computer industry CAN’T EVEN TELL THE DIFFERENCE BETWEEN A NOUN AND A VERB. How can they be expected to know the difference between ‘was’ and ‘were’? Never mind that it’s practically impossible to take into account all the subtleties of a language. And don’t forget the importance of context. This is one of my favourite rant topics. I really have to shut up now or else I’ll be here all night.