Keynote: The computer science behind the Microsoft Cognitive Toolkit: An open source large-scale deep learning toolkit for Windows and Linux

Deep Learning is redefining computing. Deep Neural Networks, or DNNs, have led to breakthrough accuracy improvements for tasks formerly considered AI, like speech recognition, image classification, and translation. Recurrent DNNs are differentiable universal computers. DNNs are layered structures of relatively simple functions with millions to billions of learnable model parameters. The challenge is that these model parameters are obtained by machine learning of sometimes billions of data samples, which often requires harnessing a farm of powerful multi-GPU servers. Microsoft's open source Cognitive Toolkit (CNTK) is used to create the DNNs that power many Microsoft services and products. It enables researchers and data scientists to easily code such neural networks at the right abstraction level, and to efficiently train and test them on production-scale data. This talk will discuss the Cognitive Toolkit and how it takes a functional-style, differentiable user program, compiles it into a computation graph for GPU execution, and distributes execution of its training across a GPU-server farm. This talk will explain how the toolkit's design intersects with several topics of interest to the three co-located conferences.