Microsoft is currently extremely active in this era of Artificial Intelligence and has very recently open sourced its award-winning Infer.NET AI framework.
Microsoft’s ‘love’ for AI research is growing each day and at times it’s in the good direction. They have pledged $115 million for humanitarian related AI programs under its AI for Good initiative. Some of the AI for Good programs by Microsoft are AI for Humanitarian Action, AI for Accessibility and AI for Earth. Infer.NET is part of the program.
What is Infer.NET all about?
Infer.NET is a Machine Learning framework that can use a developer’s customized model to create a Machine Learning algorithm built around that model only. It uses Bayesian inference in graphical models and can also implement Probabilistic Programming.
This is in contrast to many learning models that require a separate learning algorithm that has already been developed previously.
Since the algorithm is completely based upon the model fed to the Infer.NET framework, interpretation and debugging becomes much easier as the developers can focus on their own individual models instead of a separate pre-existing learning algorithm.
The Model-based Machine Learning algorithm will evolve and work in only the way the developer has specifically designed the model.
You can learn more about Model-based Machine Learning from this free eBook from the developers.
Infer.NET has been in development since 2004 via Microsoft’s research centre in Cambridge, UK. It was released for academic use in 2008 and finally went Open Source on October 5, 2018.
Applications include but are not limited to:
- Information retrieval
Till now there have been hundreds of research applications of Infer.NET. Some of the following research applications can be used as examples to understand it better and its significance now as FOSS:
Three notable research initiatives through Infer.NET
- Trajectories of Lung Function During Childhood
- Bridging Medical Data Inference to Achilles Tendon Rupture Rehabilitation
- How do disturbances and environmental heterogeneity affect the pace of forest distribution shifts under climate change?
The above three papers make it clear as to why the framework, now being Open Source, is a Positive Step towards Humanity.
Infer.NET is available on GitHub and also as NuGet packages.
Do you like Microsoft’s new initiative? Are you interested in Machine Learning? Share your thoughts with us on the comments below.