Layer-Wise Relevance Propagation: An Overview

Authors: Grégoire Montavon, Wojciech Samek, Klaus-Robert Müller
Year: 2019
Summary: This paper provides an overview of Layer-Wise Relevance Propagation (LRP), a technique for explaining the predictions of deep neural networks. LRP propagates the prediction backwards through the network to assign a relevance score to each input feature (e.g., pixel). This allows for the creation of heatmaps that visualize which parts of the input were most important for the decision.
Link: https://link.springer.com/chapter/10.1007/978-3-030-28954-6_10