This project allows one to generate distributions in network and find the best 
local model approaching these distributions with a neural network based method.

Project made by Massi Rashidi and Antoine Girardin.

The file main_generate_distributions.py computes distributions for given 
networks, measurements and states.

The file main_find_local_model.py run the neural network to find the best local model.

The file main_read_results.py reads the model found in main_find_local_models.py.

The folder utils contains the definition of some useful measurement, networks and states.

The folder main_codes_for_exact_strategies contains an example to use another approach to
find local model in networks. It build explicit strategies with a finite cardinalities of 
the symbols distributed by the sources. This method works best for small network with few 
outputs, like the triangle with two outputs.