This is small library to turn binary classifiers into multi class classifiers using a Reed-Muller error correction code.
It is compatible with the scikit-learn API for multiclass and it is conceptually similar to OutputCodeClassifier.
I've tried it on a few datasets and it is useful in the same scenarios as hierarchical softmax. I haven't done extensive testing but it could promising for language modelling.
As far as I know, there is no content on Reed Muller for Machine Learning (papers or code), but if you know of any please let me know! (email is in the commits).