tishby.py - builds a model (architecture defined in tishby_net.py) based on synthetic dataset from Shwartz-Ziv & Tishby, 2017
MIestimation.py - computes mutual information (MI) using sample propagation estimator
plotFig.py - creates a figure with the results of training and MI estimates
*.pj - scripts to run the code in the cluster (this is for reference only, simply to see how to run the code)
saved - directory containing data after a sample run of the model and the resulting MI estimates
questions? please email to: [email protected] or [email protected]