Machine Translation and Everyday Life

Machine Translation and Everyday Life

How do you compete with Google? Well, step one might be to partner with a company that holds relatively similar prowess.

 

I have been using Facebook for over a decade, and being an expat American in a foreign country has led to my share of (facebook) friends that speak a plethora of languages that I simply can’t keep up with. Just over three years ago Facebook introduced the “translate” feature, which allows you to translate what someone posted, replied to, captioned in a photo, etc. in other languages with just a single click.

 

And who landed this massive project, which was destined to make the site a global giant in the machine translation world? Bing Translator, of course. Wait, who? After three years as the go-to machine translation for the world’s second most-trafficked website, according to the Alexa study, one would expect big things. Given the productivity of online machine translators, which allows users to individually correct problems as they go, Bing’s exposure to so many clients in recent years should mean sound translation by now, shouldn’t it?

 

I’ll show you a few examples, and you can be the judge:

1)
Italian to English translation agency

2)

Italian to English translation agency

And the grand finale:

3)
Italian to English translation agency

 

Not too shabby. While you might sit there scratching your head before you just give up and go to another page, this issue certainly says something about the difficulty of automatic translation with smaller pieces that lack context. But how much context does a machine truly utilize? The latest trend in machine translation seems to be the real-time translator, used for live socialization.

 

Take Skype’s new real-time chat translation. Skype, a sort of online telephone, is ready to take a global product to an even more global level. If you need to speak with someone who does not understand your language, and vice versa, simply turn on this feature and let Skype’s real-time translation do the rest. Talk, wait for translation, listen, wait for translation, rinse and repeat.

 

A ground-breaking idea, no question about it. It’s almost as if there is a little Bart Simpson in all of us, calling halfway across the world just to see if the toilet flushes a different way (only now minus the hefty phone bill, and in a new language):

 

 

But how far can a system like this get when it is driven by smaller, context-deprived conversations, like Bing’s unsuccessful Facebook campaign?

 

As an English teacher here in Italy, I have one recommendation for you Facebook users lost in (Bing) translation: enroll in a language course.