Machine Learning

Google Cloud Acquires Kaggle, Releases New Machine Learning Products

• By Richard Chang
• 03/08/17

Google Cloud has acquired Kaggle, home to the world’s largest community of data scientists and machine learning enthusiasts. More than 800,000 data experts use Kaggle to explore, analyze and understand the latest updates in machine learning and data analytics.

Google Cloud is also releasing new machine learning products, including a Video Intelligence application program interface (API) that automatically recognizes objects in videos and makes them searchable and discoverable. Users can type in words such as “dog,” “flower” or human,” or verbs such as “run,” “swim” or fly,” and search within every shot or frame of every video file in their catalogs.

The API quickly annotates videos stored in Google Cloud Storage with video, shot and frame-level context. Here are some main features:

• Label detection, identifying objects such as everyday items, places and things (i.e. elephants);
• Temporal annotations: One can search for relevant entities across one’s entire video catalog and automatically aggregate one’s search by video and each frame location; and
• Shot change detection: Detect scene changes within the video.

Some applications of this new technology include media archiving, allowing publishing platforms to recommend digital content to consumers, and video content discovery.

Google also revealed new capabilities for Vision API, as well as an Advanced Solutions Lab (ASL) training facility.

Vision API has added new capabilities that support new enterprise use cases for image search:

• Expansion of video API’s metadata to recognize millions of entities from Google’s Knowledge Graph. Google is now using the same metadata that powers Google image search. This also enables the API to detect and group similar images.
• Enhanced optical character recognition that can extract text from images of text-heavy documents such as books or legal contracts.

The ASL is a new Google facility in Mountain View, CA, that is designed for customers to directly collaborate with Google’s machine learning experts. Customers will be able to explore how machine learning can solve their specific use cases, while getting trained on the principles of machine learning and how to use Cloud Machine Learning Engine, which is Google’s fully managed service for building and training machine learning models at scale. The training is based on Google’s own internal education and training for machine learning.

Cloud Machine Learning Engine (CMLE, formerly known as Cloud Machine Learning Platform) is now generally available for all businesses. CMLE can create a rich environment across TensorFlow and cloud computing tools such as Google Cloud Dataflow, BigQuery, Cloud Storage and Cloud Datalab.

CMLE has been deployed by companies such as Airbus and Ocado.

Google announced these new products and developments at its Google Cloud Next conference, taking place today through March 10 in San Francisco.

For more information on Google’s new products and developments, visit Google Cloud’s Big Data and Machine Learning Blog.