Open source Microsoft Cognitive Toolkit democratizes AI and deep learning

machine-learning

Like many technology companies, Microsoft is pinning a lot on AI -- including the areas of speech and image recognition. To help speed up development, and to enable others to start working on their own projects, the company has released an updated, open source version of the Microsoft Cognitive Toolkit.

This is a deep learning toolkit, previously known as the Computational Network Toolkit (CNTK), and it's available for anyone to use completely free of charge. The toolkit has applications far beyond speech recognition, and it has already been used in Bing, and the latest version includes support for Python and C++.

Microsoft Artificial Intelligence and Research has already used the toolkit to develop technology capable of recognizing words in a sentence just as a well as human beings. Microsoft Cognitive Toolkit was considered key to reaching this milestone as it dramatically sped up development by using techniques such as working across multiple servers.

NVIDIA is one of the companies already taking advantage of the toolkit. Ian Buck, general manager of the Accelerated Computing Group at NVIDIA, says:

Microsoft Cognitive Toolkit represents tight collaboration between Microsoft and NVIDIA to bring advances to the deep learning community. Compared to the previous version, it delivers almost two times performance boost in scaling to eight Pascal GPUs in an NVIDIA DGX-1.

Writing on the Microsoft blog, Allison Linn says:

With the new version, researchers [...] can do a type of artificial intelligence work called reinforcement learning [...] The toolkit is able to deliver better performance than previous versions. It's also faster than other toolkits, especially when working on big datasets across multiple machines. That kind of large-scale deployment is necessary to do the type of deep learning across multiple GPUs that is needed to develop consumer products and professional offerings.

The Microsoft Cognitive Toolkit is available on GitHub.

Photo credit: Lightspring / Shutterstock

© 1998-2017 BetaNews, Inc. All Rights Reserved. Privacy Policy.