Microsoft’s Machine Learning Framework Goes Open Source

The global technology giant Microsoft has opened its multiplatform infrastructure Infer.NET – a machine learning engine used in Office, Xbox and Azure – for everyone and the world.

“We are extremely excited today about using Open Source Infer.NET on GitHub under the MIT license for free use in commercial applications,” wrote Yordan Zaykov, Senior Software Engineering Manager. at Microsoft, in an article published on Friday.

Developed in 2014 by Microsoft’s research laboratory in Cambridge, Infer.NET was originally designed as a research tool before being released in 2008 for academic purposes.

Infer.NET provides a model-based approach for machine learning. It allows users to integrate domain knowledge into their model.

The framework can then build a custom machine learning algorithm directly from this model. “This means that instead of having to map your problem to a pre-existing learning algorithm that was provided to you, Infer.NET actually builds a learning algorithm, based on the model you provided,” Zaykov said.

He noted that the Infer.NET team is eager to collaborate with the open-source community in development. Infer.NET will become part of ML.NET – the machine learning framework for .NET developers.

“We have already taken several steps to integrate ML.NET, such as configuring the repository under .NET Foundation and moving the package and namespaces to Microsoft.ML.Probabilistic.Infer.NET will extend ML.NET to statistical modeling and online learning,” said Zaykov.

Infer.NET has been used to publish hundreds of research papers using a variety of fields, ranging from information retrieval to healthcare.