This repository is a proof-of-concept implementation of decision trees trained with the loss function proposed in The Information Bottleneck, and a presentation about it. The idea andthe usage of the code are also explained in the notebook.
This is my project for the the lecture Information Theory of Prof. Caire at TU Berlin, so thank you for the patience!
Clone the repository and run: pip install .