Microsoft is making available its homegrown deep learning toolkit used by its researchers to advance artificial intelligence, on GitHub and MIT open source license.
CNTK (Computational Network Toolkit) was launched by Microsoft in April of 2015 and at that time it was only restricted to strict academic license.
Earlier the toolkit was available in its own CodePlex site, but now the company is moving it to the popular open source repository hosting service, GitHub.
Microsoft’s chief speech scientist, Xuedong Huang said he and his team were anxious to make faster improvements to how well computers can understand speech, and the tools they had to work with were slowing them down.
So, this task was assigned to group of volunteers to solve this problem using homegrown solutions and the results were promising.
Internal tests showed that CNTK proved its efficiency when compared to similar computational toolkits that were developed to create deep learning models for speech and image recognition.
Huang said: "The CNTK toolkit is just insanely more efficient than anything we have ever seen."
The results are particularly important in the advancement of deep learning capabilities in machines because other toolkits took weeks to finish some of the biggest deep learning tasks when compared to CNTK.
Content from our partners
In the recent years, more and more researchers are developing machine learning algorithms with the help of deep neural networks which are based on the biological processes of human brain.
These improvements have helped in developing systems that can recognise and translate conversations, recognise images and even answer questions about them.
In the internal tests and research, Microsoft scientists had also found that Graphics Processing Units (GPUs) which are designed to handle complex computer graphics were ideal for processing algorithms that can advance technology to recognise and understand speech, recognise images and movements.
Microsoft, principal development manager, Chris Basoglu has said that one of the advantages of CNTK is that it can be used by anyone from a researcher on a limited budget, with a single computer, to someone who has the ability to create their own large cluster of GPU-based computers.
Huang said it was important for his team to be able to address Microsoft’s internal needs with a tool like CNTK, but they also want to provide the same resources to other researchers who are making similar advances in deep learning.
Huang said: "With CNTK, they can actually join us to drive artificial intelligence breakthroughs."
Internally, the company is using CNTK on a set of powerful computers that use graphics processing units, or GPUs.
Microsoft claimed that CNTK is the only toolkit which can scale beyond single machine and the toll on the Azure Lab can be scaled beyond 8 GPUs across multiple machines.
Apart from Microsoft, Baidu, Facebook, Google are also working on deep learning technology and released open-source code in the past. Microsoft uses CNTK for speech recognition in its Cortana personal digital assistant and Skype Translator.
