Google Translate now interprets Chinese-to-English with near human-level accuracy

The very software that thrice defeated world grandmaster Lee Sedol for the ancient game of Go is now beginning to power Google Translate. Google will now begin to translate Chinese to English using a system called Google Neural Machine Translation.

Google Translate had been using the phrase-based production system to translate Chinese to English, among other language pairs, but with little success. That was particularly so because Mandarin Chinese is notably hard to convert to English due in part to the different meanings a word could take when paired with certain characters. But that was mainly because of the differences between the Chinese and English cultures, which affect language.

The rollout of Google’s neural machine translation system in both the web and mobile versions of Google Translate could significantly change that. Native Mandarin Chinese speakers may no longer cringe at Google Translate results as the AI-based tool will now look at the whole sentence structure before decoding it.

Previously, Google Translate used to parse sentences into their component words as it interpreted them, resulting in sometimes nonsense translations.

Based on Google’s tests, the Google Neural Machine Translation system reduces translation errors by between 55 percent and 85 percent. Quoc V. Le and Mike Schuster, research scientists for Google Brain Team, lauded the GNMT approach as more advantageous than Phrase-Based translation systems.

"There is still a lot of work we can do to serve our users better. However, GNMT represents a significant milestone."

However, the researchers acknowledge that machine translation isn’t solved just yet. In a blog post, the researchers wrote:

"GNMT can still make significant errors that a human translator would never make, like dropping words and mistranslating proper names or rare terms, and translating sentences in isolation rather than considering the context of the paragraph or page. There is still a lot of work we can do to serve our users better. However, GNMT represents a significant milestone."

On top of the Chinese-to-English language pair, Google also plans to implement GNMT in 10,000 other language pairs supported by Google Translate. You might want to check out the updated version of the app now and tell us how well it works.

Pilot earpiece targets language barriers with live conversation translation

From the Hitchhiker's Guide to the Galaxy's Babel fish to Star Trek's universal translator, science fiction has found ways to break down the intergalactic language barriers, but it's something those of us in the real world are still struggling with. New York startup Waverly Labs is now claiming it's ready to make fiction a reality with the Pilot earpiece, which sits in your ear to provide near real-time translations of multilingual conversations.

   

The time and technology seem close to bring another idea from science fiction into reality. The Babel fish from Douglas Adams' Hitchhiker's Guide to the Galaxy was a living, squirming animal you stuck in your ear to translate any language in the universe for you in real time – a neat plot device that let every alien in the novels understand each other.

Now, a Manhattan-based company called Waverly Labs is working on commercializing an electronic device that does a similar job. The Pilot is an earpiece that listens in to a conversation and communicates with your smartphone to give you a close-to-real time translation.

To do so, it's going to need to rely on several potential weak-link systems; it'll need a clear signal from its in-built microphone, which will need to do a decent job of converting that signal from speech into text in both speakers' languages.

Then it'll need a good, effective translation, presumably from an online translation engine like those run by Google or Microsoft. In particular, it'll need to operate super-quickly and do a good job translating each language in a spoken, chatty form.

Then it'll need to convert the translation from text to speech and send it back to the earbud. And it'll need to be able to do all these tasks concurrently if the other person keeps talking while it's thinking.

The fact is, all these systems are already out there, up and running. None are perfect, in fact most are still glitchy and inaccurate, but each is steadily improving. Waverly Labs has wisely chosen to launch with European Latin and Germanic languages only at first; these are handled far better by online translators than Euro-to-Asian language translations at the moment.

Both Google Translate and Microsoft's Skype Translator are also already attempting real-time conversation translation in mobile and desktop applications. Pilot's key innovation is to put this stuff into a wearable device such that it effectively "whispers" the translation into the listener's ear.

It'll be interesting to see how that works out in practice, with the inevitable delay the translation system is going to add to the conversation, there won't be any way for you to know when the other party has actually received the translation of the last thing you said. So until it's super quick, it might actually be better to do this stuff through a phone that's sitting on a table that both parties can hear.

In fact, that's how Waverly Labs is going to launch the Pilot system. While pre-orders via Indiegogo will start very soon, deliveries aren't expected for another 12 months. But this US summer, the team will launch a mobile app that gives you the translation experience on your smartphone.

Full retail for the Pilot earpiece system will be US$299. That'll get you a pair of earpieces, so you can use both to listen to music, or presumably give one to your foreign friend when it's time to try to communicate through these things.

Is real-time translation technology ready to make the leap into the big time yet? The proof will be in the pudding. But enough of the pieces are moving into place to suggest that the language barrier may finally be broken.