ML.NET lets Windows devs infuse machine learning into apps

Radu Tyrsina
by Radu Tyrsina
CEO & Founder
Download PDF

At Build 2018, Microsoft announced the preview of ML.NET, a cross-platform, open source machine learning framework. The company’s target are .NET developers who will get the chance to develop their very own models and infuse custom ML into their apps without needing to have expertise in developing or tuning machine learning models.

ML.NET enables ML tasks

NET was initially developed by Microsoft Research and evolved into a massive framework over the past ten years. Now, it’s being used across lots of product groups in Microsoft including Azure, Bing, Windows and more.

As shown in the preview release, ML.NET enables ML tasks such as classification (sentiment analysis and text categorization) and regression (price prediction and forecasting).

Microsoft ML.NET sentiment classification algorithm

Microsoft ML.NET sentiment classification algorithm

iv>Besides these ML capabilities, the first release of the ML.NET also packs the first draft of .NET APIs for training models, using models for prediction and the core components of the framework including transforms, algorithms, and core ML data structures.

<p>ML.NET </pcan also be extended to add popular ML Libraries such as TensorFlow, Accord.NET, and CNTK. Microsoft stated in its official announcement that the company is “committed to bringing the full experience of ML.NET’s internal capabilities to ML.NET in open source. To sum it all up, ML.NET is our commitment to making ML great in .NET.” 

ML.NET will enable more scenarios over time

ML.NET will allow other situations in the future such as anomaly detection, recommendation systems, and approaches like deep learning by leveraging popular deep learning libraries such as TensorFlow, Caffe2, and CNTK, and also general machine learning libraries such as Accord.NET.

ML.NET will also support and enhance the experience that Azure Machine Learning and Cognitive Services provides by allowing a code-first approach, supporting app-local deployment and the possibility to build personal models.</p>

Join Microsoft on GitHub to support shaping the future of ML in .NET.