The most popular and comprehensive Open Source ECM platform
In November 2015, a group within Google called ‘Google Brain’ introduced an open-source API called TensorFlow for numerical computing and managing deep learning and artificial intelligence tasks like speech and image recognition. TensorFlow can be deployed on anywhere from one to a distributed set of CPUs or GPUs.
Matthew Zeiler, founder of AI service Clarifai, said that “Neural nets are getting used everywhere these days… Deep learning is a huge opportunity right now because it enables developers to create applications in a way that was never possible before. Neural networks are a new way of programming computers … It’s a new way of handling data.”
With TensorFlow, tasks are broken down into a dataflow-graph representation. The nodes of the graph represent CPUs or GPUs that perform the computations. The edges of the graph represent the flow of the data between the nodes. The data is organized as large multidimensional arrays or tensors that flow into and out of the processing nodes, which is the reason for the name ‘TensorFlow’. TensorFlow itself is written in C++, but processing at the nodes is performed in Python.
Greg Corrado, a senior search scientist at Google, said that “machine learning is the secret sauce for the products of tomorrow. It no longer makes sense to have separate tools for researchers to use machine learning and people who are developing real product. There should 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 them.”
Michael A. Cusumano, a professor at MIT, told the New York Times that “the software itself is open source, but if this is successful, it will feed Google’s money-making machine. There are so many applications of machine learning to the bread and butter of what Google does.”