Google has released more details about three new tools it announced last month to help developers more easily integrate machine learning and artificial intelligence capabilities into their applications.
Google’s AutoML Vision, AutoML Natural Language, and AutoML Translation are part of a broader Cloud AutoML suite of Google machine learning offerings aimed at helping developers train high-quality machine learning models, even if they have only limited prior experience with the technology.
Google introduced the three tools at its Next ’18 partner conference in San Francisco in July. This week the company announced beta availability of the products along with more details on the technology.
“This new suite of products aligns with our mission to democratize AI, and make it easy, fast and useful for all developers and enterprises,” said Levent Besik, Google Cloud product manager in a blog Aug. 13. The AutoML tools enable both ease of use and high quality machine learning models, he said.
Besik described AutoML Vision as a more specialized version of Google’s existing Cloud Vision Application Programming Interface for injecting image analysis capabilities in applications.
AutoML Vision is a more specialized version of the Cloud Vision API and is designed for developers looking for a way to integrate image classification and analysis capabilities for specific business needs, he said. AutoML Vision allows developers to upload their own image datasets and create custom machine learning models for image analysis.
AutoML Natural Language is aimed at making it easier for developers to implement natural language processing capabilities in their products. Like AutoML Vision, AutoML Natural Language is a specialized version of Cloud Natural Language, an existing capability for helping embed powerful text analysis into applications.
The tool helps developers build applications capable of extracting information about people, places and event mentioned in text documents, blog posts and other sources. Google has described the tool as helping developers enable capabilities for doing sentiment analysis in their applications.
AutoML Translation allows developers to create apps that take advantage of Google’s translation capabilities to create language translation models for their specific domains. For example, an organization could use the technology to build capabilities for translating specific taxonomies for financial news, Besik said.
Besides the specialized tools, Google is also continuing to develop its Cloud AI APIs. For example, Google has added a handwriting recognition enhancement in the Cloud Vision API that makes it possible for developers to add capabilities in their apps for identifying handwritten text. Enhancements to the Vision API include one feature that enables apps to identify items that are visually similar to a company’s own products.
Google has also made improvements to its Cloud Text-to-Speech API, including the ability for an app to automatically identify the language being spoken, and the ability to identify and record individual participants in multiple participant conversations.