Representations in the hidden layers of Deep Neural Networks (DNN) are often hard to interpret since it is difficult to project them into an interpretable domain. Graph Convolutional Networks (GCN) allow this projection, but existing explainability methods do not exploit this fact, i.e. do not focus their explanations on intermediate states. In this work, we present a novel method that traces and visualizes features that contribute to a classification decision in the visible and hidden layers of a GCN. Our method exposes hidden cross-layer dynamics in the input graph structure. We experimentally demonstrate that it yields meaningful layerwise explanations for a GCN sentence classifier.
@article{arxiv.1909.10911,
title = {Layerwise Relevance Visualization in Convolutional Text Graph Classifiers},
author = {Robert Schwarzenberg and Marc Hübner and David Harbecke and Christoph Alt and Leonhard Hennig},
journal= {arXiv preprint arXiv:1909.10911},
year = {2019}
}
Comments
Accepted at EMNLP 2019 Workshop on Graph-Based Natural Language Processing