Microsoft open-sources Infer.NET machine learning framework
https://slashdot.org/submission/8747783/microsoft-open-sources-infernet-machine-learning-framework
Yordan Zaykov writes:Earlier this month at the annual international conference in probabilistic programming Microsoft announced the open-sourcing of Infer.NET — a cross-platform machine learning framework the company has been developing for over a decade. Infer.NET differs from other mainstream frameworks in this domain in that it employs a model-based approach to machine learning. In this paradigm, the user defines a model of how the real world works — a set of assumptions that gave rise to the data at hand. Then, a compiler automatically generates the inference algorithm for the specified model. According to the announcement, this has a number of advantages:
... you’d want to use Infer.NET when you have extensive knowledge about the domain you’re solving a problem in, or if interpreting the behaviour of the system is of importance for you, or if you have a production system that needs to learn as new data arrives.
The model-based approach is explained in more detail in a free online book by Microsoft Research. The company already started integrating Infer.NET with ML.NET and welcomes everyone to contribute on GitHub.
Comments
Post a Comment