IBM taps into group language translation

Global company IBM seems to have found a way for its employees to get past language barriers and communicate.

IBM employees are currently using text translation software that can instantly convert documents, Web pages, and even instant messages between English and 11 other languages. The software, christened "n.Fluent," is being "crowdsourced" or tested among IBM's 400,000 employees across 170 countries.

As IBMers use n.Fluent, the software learns from its mistakes and improves itself. As the entire company potentially taps into n.Fluent, volunteers within IBM refine each translated word for greater accuracy. In just two weeks this past summer, volunteers tackled around 1.3 million words, averaging around 100,000 per day. Overall, n.Fluent has translated more than 400 million words for Big Blue staffers.

The software works as a plug-in or add-on to other applications, making it fairly seamless to use. Plug-ins translate instant messages on the fly. Text from a word processing document or other presentation is copied into one field of the software, with the immediate translation popping up in another field. IBMers can use n.Fluent on their desktops, laptops, and even smartphones.

This "universal translator" can currently tackle English, Chinese (Simplified and Traditional), Korean, Japanese, French, Italian, Russian, German, Spanish, Portuguese, Italian, and Arabic.

"To become a smarter planet, the world needs a shared vocabulary for collaboration -- particularly the business community," said David Lubensky, an IBM researcher managing the n.Fluent project, in a statement. "We see n.Fluent as just such a tool, helping to expand commerce, cement relationships and make the world that much smaller, one word at a time."

Of course, free language translators, such as Google Translate, are already available. But IBM sees n.Fluent as a better alternative. The software is more secure as it runs behind a firewall. It's also adept at handling business jargon. Right now, n.Fluent is only being used internally. But like many of IBM's research projects, it's likely to find a home outside of Big Blue's walls.

IBM spokesman Ari Fishkind said there's no fixed date as to when it might be available externally. "It would be a reasonable assumption that there's a demand in the market for a translation tool that has very good security," he said. "And also this kind of tool is uniquely tuned for a business environment that has almost a language in itself."

Other language translation tools can convert individual words. Key to n.Fluent's success will be how it handles entire sentences and paragraphs as well as colloquialisms. But the company's field tests are geared toward those goals.

"The whole point is to continually refine the idioms and the syntax and the context by people who use the language every day," said Fishkind. "And that's part of this crowdsourcing idea where hopefully at the end of the day we're going to have a system that is not only intelligible but also fluent and fluid."

[rdupuy11]

Sounds like they have no idea what they are doing.

Translations programs do not simply tackle mere words, they, and for many years now, tackle colloquialisms, phrases, and sentence structures.

To think they are so bad because someone missed such an obvious advance, is ridiculous.

It's not that easy. Good luck, there is room for improvement here, but you can't just hope throwing a crowd at it, is enough. It's not enough, not to make it 'smart.'

You'll have a lot of unintentionally miscommunication with this. It would actually make more sense, to teach everyone some man-made language...at least there would be less mistakes, even if I admit, you'd still be subject to the usual problems -- i.e. IBM management doesn't even know what they are saying in their own language...pretty much no way for a machine to correct that.

[aeyec]

I agree. All the emphasis in this article is about translating WORDS, when word-for-word translations simply do not work. A proper translator must intimately understand both languages and convert the intent of what is said in one language to phrases in another language which express the same intent.

[Global_Lingo]

While I applaud IBM's attemps I have to agree that with best will in the world it'll never be better than professional human translation

[elephanty]

I've just checked with a number of Arabic sentences from BBC News, and I was pleasantly surprised to find IBM's translation performing better than Google's. If it were merely a word translation, it would have failed miserably for Arabic, which has a language structure far different from English.