Japan’s DoCoMo Invents Realtime Translation Device
In The Hitchhiker’s Guide to the Galaxy, the English sci-fi humorist Douglas Adams famously conjured the Babel Fish, a “small, yellow, leech-like creature” able to tap directly into the speech center of the brain, thereby providing instant and universal language translation. The creature was so useful, wrote Adams, that some chose to see it as prove of God’s non-existence. “The argument goes something like this: ‘I refuse to prove that I exist,’ says God, ‘for proof denies faith, and without faith I am nothing’. ‘But,’ says man, ‘the Babel fish is a dead giveaway, isn’t it? It proves you exist and so therefore you don’t. QED.’ ‘Oh dear,’ says God, ‘I hadn’t thought of that,’ and promptly vanishes in a puff of logic.”
Convoluted proofs notwithstanding, NTT Docomo unveiled a feat of technology at the recent Wireless Japan 2011 conference that would, at the very least, have God whistling with admiration. The company demonstrated a translation system that is something like a prototypical Babel Fish. An NTT Docomo hire read a newspaper article in Japanese, and the machine very quickly turned the voice to text, the text to translation—and finally that translation into synthesized, spoken English.
English of a sort, at any rate. The resulting sentence in this DigInfo video isn’t exactly the Queen’s finest: “It was under one month until the end of current Diet session of June 22,” it began, continuing, “and, in the whereabouts of the deficit-covering bond publication bill necessary to carry out…” Are you still with me? The “sentence” goes on from there. Still, the fact that it was able to communicate any meaning at all so quickly, even if there are kinks to be ironed out, is impressive. Voice recognition alone—setting aside translation—has been shown to be difficult enough.
As NTT Docomo’s Atsushi Sato explained, though voice recognition wasn’t “at 100% yet,” admittedly, there might still possibly be a demand for a speed-over-accuracy translation device. “We don’t really know how much time it will take to bring accuracy to 100%,” he told DigInfo; rather than wait, he said, “we are considering scenarios in which customers who could accept a certain level of inaccuracy could use this.” He added that the company planned to offer trials of the service in the coming year.
NTT Docomo is joining a race with some major players (though it’s first to admit that it’s not exclusively using its own technology, but rather combining some of the best out there into its own product). In January, Google showed off its latest effort at its own proto-Babel Fish, in a somewhat less ambitious demonstration that extended simple phrases between English and Spanish (the Google translation, while more accurate, was slower than what NTT Docomo put on display). And DARPA, joining forces with the National Institute of Standards and Technology, has long taken an interest in insta-translation, to help with the paucity of American speakers of Pashto and other languages spoken in Afghanistan and elsewhere.
Who in this race will create the first functional Babel Fish? God only knows.
Keep Reading
Most Popular
Large language models can do jaw-dropping things. But nobody knows exactly why.
And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.
How scientists traced a mysterious covid case back to six toilets
When wastewater surveillance turns into a hunt for a single infected individual, the ethics get tricky.
The problem with plug-in hybrids? Their drivers.
Plug-in hybrids are often sold as a transition to EVs, but new data from Europe shows we’re still underestimating the emissions they produce.
Stay connected
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.