Google Unveils Video Cloud Intelligence, Acquires Data Community Kaggle
Until not too long ago, an online quest for videos about tigers, for instance, depended on whether the search engine could find that word in the tags, keywords and other metadata associated with those videos. So if some joker decided to upload a video about cupcakes with lots of "tiger" descriptions, you might end up with the odd baking movie in your search results.
Machine learning, though, could eliminate such problems, and Google is now offering one such solution: the Cloud Video Intelligence API, currently in private beta. Unveiled yesterday during the kickoff of the three-day Google Cloud Next '17 event in San Francisco, the application program interface (API) is the first of its kind to enable subject- and context-based searches of video content, according to Google.
Yesterday, Google also announced its acquisition of Kaggle, described as "the world's largest community of data scientists and machine learning enthusiasts." By joining forces with Kaggle, Google said it will be able to speed up efforts to advance and democratize the adoption of artificial intelligence.
'Next-Gen Apps for Unstructured Data'
The Cloud Video Intelligence API is designed to enable developers who don't have expertise in machine learning or computer vision to help users with large video catalogs better understand and identify the actual image content in their videos, Google said.
"This API is for large media organizations and consumer technology companies, who want to build their media catalogs or find easy ways to manage crowd-sourced content, and for partners like Cantemo to build it into their own video management software," Google's Fei-Fei Li wrote yesterday on the company's Cloud Platform blog. Cantemo is a Sweden-based technology company that specializes in media asset management.
The video intelligence API joins Google's growing set of interfaces for numerous tasks, including vision, speech, natural language, translation and jobs, said Li, who is the chief scientist for Google Cloud AI and machine learning.
"These APIs let customers build the next generation of applications that can see, hear and understand unstructured data -- greatly expanding the use cases for machine learning for everything from next-product recommendations, to medical-image analysis, to fraud detection and beyond," she said.
Kaggle Acquisition Aimed at 'Democratizing AI'
Based in San Francisco and founded in 2010, the Kaggle data science community provides a crowdsourced and competitive platform for solving problems involving analytics and predictive modeling. Kaggle has received just under $13 million in venture capital funding since 2011.
In its first competition in 2010, Kaggle community member Jure Zbontar beat 21 other teams for a $1,000 prize to predict voting patterns for the Eurovision Song Contest, Kaggle CEO Anthony Goldbloom said yesterday in a blog post.
The new Google-Kaggle alliance "combines the world's largest data science community with the world's most powerful machine learning cloud," Goldbloom said. "Making Google Cloud technology available to our community will allow us to offer access to powerful infrastructure, scalable training and deployment services and the ability to store and query large data sets."
Neither Google nor Kaggle disclosed financial terms of the acquisition. However, Goldbloom noted that Kaggle will continue to operate as a distinct brand within Google Cloud.
"During my keynote talk at Next '17, I emphasized the importance of democratizing AI," Google's Li said in a separate blog post. "We must lower the barriers of entry to AI and make it available to the largest community of developers, users and enterprises, so they can apply it to their own unique needs. With Kaggle joining the Google Cloud team, we can accelerate this mission."