Google Debuts New Tools To Simplify Custom AI Builds


Google is looking to further push machine learning as a differentiator for its cloud with a set of intelligent services the provider showcased Tuesday designed to simplify development of intelligent apps.

Fei-Fei Li, Google's chief scientist for AI and machine learning, presented to NEXT conference attendees a set of tools that automate building language and image capabilities without the advanced skills typically required for custom AI solutions.

"We know that many of you need more flexibility than our APIs were designed for, but aren't yet ready to make use of the advanced tools like Tensor Flow and Cloud ML engine," Li said.

[Related: Google Injects AI, Machine Learning Into G Suite Collaboration]

Sponsored post

With the set of new services, "anyone can make machine learning work for them," Li said.

Cloud AutoML lets users extend powerful machine learning models to solve challenges around their unique data.

That starts with AutoML Vision, released in beta. The service extends Google Cloud Vision API to identify entirely new categories of images.

Two more AutoML products add language-processing capabilities.

AutoML Natural Language includes capabilities for understanding subject matter specific to any industry.

And AutoML Translation enables machine learning-powered translation to recognize jargon and figures of speech unique to any customer's business. It provides "translations that more faithfully capture the nuance that your audience expects," Li said.

Google also showcased a new AI solution for contact centers that reduce requirements for skills only humans could once provide, helping businesses handle high call volumes.

Contact Center AI can navigate requests made with informal language, understand context and recognize the needs of callers that have provided limited information.

The contact center service is now in an alpha release, and open to sign-ups, Li said.

To complement those efforts, Google is preparing its third-generation Tensor Processing Units for deployment in its cloud, "demonstrating our ongoing commitment to put the best hardware in the hands of AI developers"