Microsoft Infer.NET machine learning framework has evolved from a mere research tool to become the machine learning engine in a number of Microsoft products, including: Office, Xbox and Azure.
Now, the company has open-sourced the Infer.NET framework, and it will eventually become part of the ML.NET machine learning framework for .Net developers, extending ML.NET for statistical modeling and online learning.
Initially, Infer.NET was envisioned as a research tool released for academic use in 2008, and there have been hundreds of papers published using the framework across a variety of fields, from information retrieval to healthcare.
In this age of abundance of machine learning libraries, Infer.NET will enable model-based approach to machine learning, which allow incorporation of domain knowledge into statistical model.
The framework can build a bespoke machine learning algorithm directly from a model, instead of having to map a problem onto a pre-existing learning algorithm, Infer.NET will actually constructs a learning algorithm based on the model provided.
The model is compiled by the framework into high-performance code to implement deterministic approximate Bayesian inference, which allows for substantial scalability. And the use of deterministic inference algorithms is complementary to the predominantly sampling-based methods of most other probabilistic programming frameworks.
The Infer.NET team is looking to engage with the open-source community in developing and growing the framework further, and Infer.NET will become part of ML.NET – the machine learning framework for .NET developers.