On Network Science and Mutual Information for Explaining Deep Neural Networks
Machine Learning
2020-05-05 v2 Machine Learning
Abstract
In this paper, we present a new approach to interpret deep learning models. By coupling mutual information with network science, we explore how information flows through feedforward networks. We show that efficiently approximating mutual information allows us to create an information measure that quantifies how much information flows between any two neurons of a deep learning model. To that end, we propose NIF, Neural Information Flow, a technique for codifying information flow that exposes deep learning model internals and provides feature attributions.
Cite
@article{arxiv.1901.08557,
title = {On Network Science and Mutual Information for Explaining Deep Neural Networks},
author = {Brian Davis and Umang Bhatt and Kartikeya Bhardwaj and Radu Marculescu and José M. F. Moura},
journal= {arXiv preprint arXiv:1901.08557},
year = {2020}
}
Comments
ICASSP 2020 (shorter version appeared at AAAI-19 Workshop on Network Interpretability for Deep Learning)