epann

Evolutionary Plastic Artificial Neural Networks

A Python package for implementing Compositional Pattern Producing Network (CPPN) variants for indirectly encoding artificial neural networks for neuroevolution and artificial life experiments.

• Purpose:

1. Greater understanding of the HyperNEAT algorithm and indirect encoding variants for evolving a populations of neural networks.
2. Practice in package management.
3. General purpose research pipeline for appyling EPANN neuroevolution techniques to a variety of environments.
4. Dissertation code and documentations.

Tutorials

You can find a collection of tutorials meant to introduce the major algorithms and features of the package in the docs/ subdirectory:

Start with the Tutorial Introduction