MIcrosoft is joining a growing movement to make machine learning an open source option for developers. The software giant announced today that its machine learning toolkit, DMTK, is now available on GitHub for use by developers around the world.
DMTK, which stands for Distributed Machine Learning Toolkit, comes out the Microsoft Asia research lab and includes server-based frameworks that let developers easily scale machine learning programs. The toolkit also contains two algorithms that Microsoft developed to quickly train computers for various tasks.
The news comes just days after Google open-sourced its machine learning system in a bid to create a common set of tools for interfacing with computer-based neural networks. Facebook also opened up its deep-learning system earlier this year.
Each company is trying to get a foothold with researchers and creators to make its system the dominate player in the growing machine-learning market. Google may have the more consumer-focused appeal right now, with its Smart Reply system helping users respond to emails quicker. Facebook may also be using machine learning to improve things like detecting restricted content or improving its digital assistant program. But Microsoft has more appeal to developers at the moment, opening up Azure as a prime spot to host machine learning programs in the cloud.
