Google Translate just got a lot smarter

The search giant says it’s made a “leap” in giving you more natural translations. It’s just one step in a big push in machine learning that Google detailed Tuesday.

 by Richard Nieva November 15, 2016

World travelers, language nerds and everyone in between, you’re in luck.

Google said Tuesday it has vastly improved its Google Translate app, available on phones and the web.

The search giant said it’s now incorporating “neural machine translation” into the software, which means it can translate whole sentences at a time, instead of breaking the text down to smaller chunks and translating those pieces.

The result is translations coming out more natural, with better syntax and grammar, Google said.

“It has improved more in one single leap than in 10 years combined,” Barak Turovsky, the product lead for Google Translate, said during a press event at Google’s San Francisco office.

Besides English, the new translation system is coming to eight of the 103 languages supported by the app. They are French, German, Spanish, Portuguese, Chinese, Japanese, Korean and Turkish. Google says those languages represent 35 percent of all translations done on the service. Turovsky said the new method cuts down on errors by 55 to 85 percent.

Google has bet the future of the company on artificial intelligence and machine learning — a type of technology in which computers can learn without being explicitly programmed. Last month, the company unveiled several new consumer devices that feature a digital helper Google calls “the assistant,” similar to Apple’s Siri or Amazon’s Alexa. But Google bills it as the natural evolution of its iconic search box, drawing on the reams of data Google knows about you. Instead of typing a search query into a text box at your desktop computer, you’ll speak commands to your phone, watch or smart home hub.

When Google CEO Sundar Pichai first introduced the concept of the assistant in May, he said it was the cornerstone of Google’s “journey” to becoming an “AI-first” company.

The search giant isn’t the only Silicon Valley giant that has made big investments in machine learning recently. Earlier this year, Facebook started an Applied Machine Learning division, where it works on technology like photo and audio recognition.

Google on Tuesday also said it’s forming a Cloud Machine Learning group, which will focus on bringing its AI technology to other businesses. The group will be headed by Fei-Fei Li, an AI professor at Stanford, and Jia Li, former head of research at Snap, the parent company of Snapchat.

One of the group’s new projects aims to improve how job listings are categorized. One customer is FedEx, and the new service lets a job seeker more easily search for a warehouse job, even if the job title or or description doesn’t have the word “warehouse” in it. (For example, “Fork lift operator 3rd shift.)

Diane Greene, who leads the company’s cloud division, summed up the goal: “Google takes machine learning in every form and takes it to the world.”