Contributing

If you are interested in contributing to Syntropy, there are several fronts for development. The primary area is the code itself (features, bugfixes, efficiency), but we are also interested in tutorials, educational material, and example notebooks that can serve as a reference to new users.

To-do

The space of information theory is vast, and while Syntropy aims to be comprehensive, there’s a lot that has been untouched. If you are interested in implementing any of the following, feel free:

• GPU acceleration for the neural estimators.

• Mixed estimators (e.g. mutual information between discrete and continuous random variables).

• The i_sx and i_min redundancy functions for KNN-based PID.

• Bug fixes in any existing implementations.