Google Buys Machine Learning Startup Moodstocks

...

The underground marketplace is booming and only getting bigger, more sophisticated, and competitiv

Google has acquired Paris-based machine learning startup Moodstocks as part of its ongoing effort to improve visual recognition technology.

Google is adding to its already substantial set of machine learning technologies with the acquisition of a French company called Moodstocks, a visual recognition machine learning technology company.

Object recognition is one of the more difficult problems for machine learning, and it's a problem that Google has been working on for a while.

In a blog post announcing the deal, Google noted that many of its services including Google Translate and Smart Reply Inbox already rely on machine learning technologies. The addition of Moodstock will help with visual recognition. 

Vincent Simonet, head of the R&D Center of Google France wrote in the blog post that Google has made great strides in terms of visual recognition technology -- for instance, if you search the word "party" or "beach" you'll get a good image match.

"But there is still much to do in this area," he wrote (as translated from French to English by Google Translate), and that's where Moodstocks comes in.

[Google has expanded its R&D capabilities in Europe. Read Google Launches AI, Machine Learning Research Center.]

Google said that Paris-based Moodstocks' engineers and researchers have been developing new algorithms for visual recognition of objects, particularly those viewed by cameras in mobile devices. Now the Moodstocks team will join Google's R&D center in Paris which contributes to the development of YouTube and Chrome.

The deal is expected to close in a few weeks, according to a post at the Moodstocks website. No other terms of the deal were disclosed.

"Our dream has been to give eyes to machines by turning cameras into smart sensors able to make sense of their surroundings," Moodstocks said in a post on its website. "After introducing on-device image recognition in 2012, we've been working on extending our reach to object recognition for the past 2.5 years, using deep learning-based approaches."

The Google acquisition will now allow the company to take its technology to a much wider audience.

"Our focus will be to build great image recognition tools within Google," the company said.

Current paying customers of Moodstocks will be able to use the technology until the end of their subscription. The company said it plans to discontinue its image recognition services soon, but said "we look forward to bringing you even better tools as part of Google."

Categories
APPLICATIONS
0 Comment

Leave a Reply

Captcha image


RELATED BY

  • 5300c769af79e

    10 Factors Determining Data Pro Salaries

    These companies count data engineers, data architects, data scientists, and other data pros among the key players enabling digital transformations.That's part of the idea behind O'Reilly Media's 4th annual survey of data professionals and their compensation.
  • 5300c769af79e

    Siri Can Instruct PayPal to Empty Your Bank Account

    As Engadget explains, PayPal requires you to link Siri to your account and share your Contacts list.After that, saying "Hey Siri, send Bob $30 using PayPal" will be recognized as a valid command and Siri will bring up a confirmation screen showing the payment details.
  • 5300c769af79e

    When you live in a cellular desert, WeBoost's Eqo makes it rain

    My summers as a teenager spent on my grandparents’ farm would have been a lot more bearable if I had WeBoost’s Eqo cell phone signal booster in my arsenal.In this case, you might opt for something like the WeBoost Eqo, which claims to be discrete, compatible, and powerful enough to be worthy of your attention.
  • 5300c769af79e

    BMC's NGT Equals DB2 Utilities for the 21st Century

    Download In this analyst report, you'll learn how the operational demands of the digital age will require next generation DB2 mainframe management technology that can optimize infrastructure performance and deliver excellent response to to customers.Read the report to:- Learn about the operational challenges of dynamic databases- Explore how next generation technology can enhance mainframe data management- See how BMC can cut DB2 mainframe management costs while improving application availability