Google has just made TensorFlow, the very AI powering the search engine, an open source platform. This is a big decision by the tech titan which will not only boost the AI Technology overall but yield intelligent apps developed by the open source community.
What is TensorFlow? It is software belonging to deep learning category, which serves as AI Engine to make intelligent calculations, including differentiating and recognizing voices, pictures, symbols and almost everything which can form a pattern.
TensorFlow is currently implemented in many Google services, including its core service, Google Search, which cherishes a 60 billion dollar cap generated as revenue from Google ads. Why then the company is handing over such a powerful tool for free to just anybody?
There are two reasons: First off, Google acknowledges and recognizes that speedy evolution of artificial intelligence is not possible just by hiring a handful of experts, everybody must contribute their ideas. Secondly, with open source AI Engines like Caffe, Torch, and Theano on the horizon, Google had to show care.
Of course, there certain strings attached to this breaking and that’s because Google cannot just give away all the secrets that are keeping the heartbeat of the company in harmony. Fully-fledged TensorFlow contains many components that Google cannot risk leaking, so it’s releasing a downgraded version of the AI Engine.
These limitations allow TensorFlow to only run only on a single machine while Google’s original software is driving the company’s entire highly-advanced neural network, comprised of millions of servers. However, for most of the deep learning AI developers, the available resources are more than enough to harness the power of Google’s AI Engine to develop PC and even mobile apps.
Until now, TensorFlow platform was accessible by Google’s own software engineers, but now open source community can also join the party. More developers translate into the faster development of the platform, it also means that these developments will return back to Google, who can then implement the better parts of the code into its main software.
Jeff Dean is known as the father of deep learning, who is also a major asset to the company when it comes to the development of its artificial intelligence programs. Showing a positive response on the open sourcing of TensorFlow, he said that they are hoping the AI community will welcome and soon adapt to the newly unlocked AI software.
As discussed in a previous article, Google is not alone in this venture. Tech giants like Facebook, Microsoft, and Apple equally want to be the part of the advanced AI dream. They are all working on their own version of AI systems, each with a different approach, however, Google has an advantage.
According to Chris Nicholson, who runs a deep learning startup called Skymind. “Google is five to seven years ahead of the rest of the world.” He further said that Google’s gesture of goodwill towards the open source AI community by offering TensorFlow will change the current scenario.
Open sourcing TensorFlow is a big shift for Google and everybody is talking about it. Keep in mind, sharing part of the AI Engine is not affect company’s lead because true power of Google’s success also lies in its cutting-edge hardware infrastructure that drives TensorFlow. Nonetheless, Google’s AI engine is regarded by some as the world’s most advanced. To find out more, visit TensorFlow website to know more about the software.