At SQL PASS Summit 2016, Microsoft's been making a number of announcements that go beyond SQL and into the future of how data is analyzed. One of those announcements: Joseph Sirosh's unveiling of the beta of Cognitive Toolkit 2.0, one piece of Satya Nadella's vision for Microsoft to be the Gutenberg press of the AI age.
The toolkit, which is open source and available on Github, comes with over 30 code samples that cover a range of learning scenarios including working with images, numbers, speech and text.
The release also offers new support for Python and C++.
Microsoft also touted that the Cognitive Toolkit now offers "reinforcement learning," which offers particular promise when training against particularly large data sets with relatively fuzzy challenges.
Microsoft was also highlighting that Cognitive Toolkit is the only public machine learning toolkit that can split loads between different machines, helping it better scale up for particularly large data sets.
That kind of increased turnaround can dramatically open up new possibilities for how machine learning is deployed and fine-tuned within an organization. Microsoft also released a video overview of the new toolkit.
