Google’s internal deep learning infrastructure DistBelief, developed in 2011, has helped the technology company advance its capabilities. It helped improve speech recognition in the Google app by 25%, assisted the image search option in Google Photos and helped in a myriad of the company’s experiments.
“Machine learning is the secret sauce for the products of tomorrow,” said Greg Corrado, a senior research scientist with Google. “It no longer makes sense to have separate tools for researchers of machine learning and people who are developing real products. There should really be one set of tools that researchers can use to try out their crazy ideas, and if those ideas work, they can move them directly into products without having to rewrite code.”
This week Google announced the open source release of its software TensorFlow, the technology company’s second-generation machine learning system.
“Part of the point of TensorFlow is to allow collaboration and communication between researchers,” Corrado said.
Machine learning is an attempt to get a computer to learn like a human brain. As the SAS Institute puts it, it “allows computers to find hidden insights without being explicitly programmed where to look.”
How do Amazon and Netflix know what to recommend to you? Machine learning.
With TensorFlow, Google said it’s improved on certain aspects of DistBelief, including speed, scalability and production readiness. “But the most important thing about TensorFlow is that it’s yours,” write Jeff Dean and Rajat Monga, both of Google. “We’ve open-sourced TensorFlow as a standalone library and associated tools, tutorials and examples with the Apache 2.0 license so you’re free to use TensorFlow at your institution (no matter where you work).”
Writing in The Conversation, David Tuffley, a lecturer in Applied Ethics and Socio-Technical Studies at Griffith Univ., said TensorFlow may lead to the development of a multi-lingual virtual assistant capable of anticipating a user’s needs by monitoring their daily patterns.
“We are making a lot easier for humans to be able to use the devices around them. We think having this as an open source tool really helps that and speeds that effort up,” said Rajat Monga, a Google technical lead.