3 stories
today

Microsoft debuts ML.NET cross-platform machine learning framework

As Microsoft executive Joe Belfiore repeatedly brought up during his day two keynote, there is a new buzzword floating around the tech industry that’s sure to win some reporters a game of keyword associated bingo as their coverage bounces from conference to conference.

That word is Machine Learning.

It’s getting paired up with everything from smartphones to light sensors to cloud storage and on that note, Microsoft is putting a little bit of Machine Learning in its developer tools.

Earlier this week, Microsoft introduced ML.NET; it’s a cross-platform open sourced solution that enables developers to create their own models and inject a bit of ML into their applications for better data collection or front-end experiences.

Originally, developed down in the Microsoft Research labs as a tooling hobby, the company has refined it over the past decade, so much so, that it’s being dogfood through many of its public offerings such as Windows, Azure, and Bing.

ML.NET enables ML tasks like classification (e.g. text categorization and sentiment analysis) and regression (e.g. forecasting and price prediction). Along with these ML capabilities, this first release of ML.NET also brings the first draft of .NET APIs for training models, using models for predictions, as well as the core components of this framework, such as learning algorithms, transforms, and core ML data structures.

As this is a first pass at offering these tools to developers outside of Microsoft, the company plans to obviously refine the product in the near future to possibly include:

  • Additional ML Tasks and Scenarios
  • Deep Learning with TensorFlow & CNTK
  • ONNX support
  • Scale-out on Azure
  • Better GUI to simplify ML tasks
  • Integration with VS Tools for AI
  • Language Innovation for .NET

ML.NET follows in the same lineage of other crowdsourced feedback development participation as Office 365 or Windows 10, so if there are areas of improvement or suggestions that can be raised, Microsoft appears to be all ears. Visit the GitHub repository to not only get started but drop them some feedback.

Further reading: , , , ,

What practical uses can you think of for Machine learning in apps?