Google Translate just got way better

29 Jul 2015 | Author: | No comments yet »

Google Translate adds 20 languages to visual translate, speeds up voice translations.

The search engine giant added 20 languages to their instant virtual translation app, a big leap from the original seven languages (English, French, German, Italian, Portuguese, Russian and Spanish) people could use to instantly identify words in a foreign language.

Google today announced that within the next few days its Google Translate app for iOS and Android will be able to give users immediate visual translations of text in 27 languages. Google also says the offline capabilities work much faster, which should enable you to quickly find out what that sign means whether or not you have a Internet connection. Instant visual translation is now available for Bulgarian, Catalan, Croatian, Czech, Danish, Dutch, English, Filipino, Finnish, French, German, Hindi, Hungarian, Indonesian, Italian, Lithuanian, Norwegian, Polish, Portuguese, Romanian, Russian, Slovak, Spanish, Swedish, Thai, Turkish, and Ukrainian, according to a blog post today from Google Translate product lead Barak Turovsky.

In an update released today, Google is adding dozens of new languages to some of the Translate app’s most powerful features, and smoothing out the app to make it friendlier to slow connections. Additionally, if a user takes a photo of the text through camera mode, Google can translate the text into an additional 10 languages, for a total of 37. Its most handy, and impressive, tool is the six-month-old instant translation feature, using the goodies from the acquired Word Lens, that lets you point your camera at something written in another language, say a sign, and it’ll translate into your language with ridiculous accuracy in almost real-time. In particular, the update makes Translate’s visual translation features significantly more powerful, letting mobile users translate 37 languages via photo, 32 via voice, and 27 through real-time video. Just open the app, click on the camera, and point it at the text you need to translate-a street sign, ingredient list, instruction manual, dials on a washing machine.

Google’s secret to recognizing this vast array of languages is hidden in deep neural networks, the same networks that create hallucinative images and eerie art. Today’s changes are updates to two features added to translate back in January: a real-time video translation feature called Word Lens and a conversation feature that translates bilingual speech in real-time.

TechCrunch spoke with Julie Cattiau, the product manager for Google GOOG 0.14% Translate, who said: Our mission is to help overcome language barriers. So no matter where you are or which country you’re visiting, you’ll always have a quick, mostly-reliable language translator, which now works in 27 different languages. The images undergo a step-by-step process, which begins when the app finds letters in the image through identifying blocks of pixels with similar colors next to each other. Google’s secret for the speed, and upgrade in speed for voice translations (to better serve slower 3G networks in ), is its use of convolutional neural networks.

It’s designed to work entirely offline, without making any queries to Google’s servers — convenient for those backpacking trips around Europe or Asia. The story behind the story: Google is making the most of its acquisition of Word Lens, which powers the augmented features of the new version of Google Translate. Google has trained its artificial neural network — a key technology for deep learning — on images showing letters as well as on fake images marred by imperfections, to simulate real-life scenes. The best part about this functionality too is that Google makes it so translations happen without having to be connected to the internet, so you can still see things happen in front of your eyes even when there’s no network connection around.

But part of that is also the overall user experience, which is why we also invest in things like instant camera translation and multi-language conversation. The app can also translate text into Hindi and Thai but cannot translate back, since the script makes it difficult for the program to resolve individual letters. Along with adding more languages to the app, Google Translate has also made improvements to its voice conversation mode so it will work faster and more naturally – even on slow mobile networks. The new languages aren’t built directly into the app, but they’re available as 2-megabyte modules, which can be downloaded by accessing the languages within the app.

In order to have the app still recognize imperfect letters, the Google team created ”all kinds of fake ‘dirt’ to convincingly mimic the noisiness of the real world,” according to Google’s blog post. I spoke to Julie Cattiau, Product Manager, Google Translate, yesterday about the Translate project as a whole, and asked specifically about Translate as an aid, rather than a replacement: JC: Translate is not a replacement for a language learning course; it’s not going to teach you a language from scratch or all of the intricacies involved in learning a language. The update will also make speech-to-speech translation easier to use on a slow connection, after Googlers noticed more and more users relying on the feature in areas with bad connections. After the program has found the letters, it then looks up approximate definitions to find the correct word, but also to account for mistakes that the program might have made earlier in the process. For instance, a handwritten “s” could have been interpreted as a “5” in the translation process, but the program would still ultimately translate the mistaken “5uper” as “super.” But not everyone has a data processing center with the capabilities of Google.

Every day, Translate processes 100 billion words across the globe, and 95 percent of those translations take place outside the US, so the recalibration toward slower connections could have real implications in countries with less developed cell infrastructure. Thanks to the millions of language lovers who have already pitched in-more than 100 million words so far!-we’ve been updating our translations for over 90 language pairs, and plan to update many more as our community grows. To counter that issue, Google had to develop a small neural net to limit the information density that the computer processor handles. “Sometimes new technology can seem very abstract, and it’s not always obvious what the applications for things like convolutional neural nets could be,” notes the Google research blog. “We think breaking down language barriers is one great use.” And Google promises that there more barriers to be broken: “More than half of the content on the Internet is in English, but only around 20% of the world’s population speaks English. JC: Personally, I really enjoy reading and hearing user stories around how we were able to improve their lives by enabling them to connect with other people — whether it’s someone successfully ordering off of a foreign menu, a couple getting married who didn’t speak each other’s languages, or a life being saved because an emergency responder could communicate with someone.

And the Translate app has also helped friends of mine – a Czech friend whose father was seriously ill in hospital was able to have a conversation in Czech / English with her boyfriend thanks to Google Translate.

Here you can write a commentary on the recording "Google Translate just got way better".

* Required fields
Our partners
Follow us
Contact us
Our contacts

ICQ: 423360519

About this site