InDepth: Layer-Wise Relevance Propagation

A deep dive into LRP for explaining neural network predictions.

Layer-wise Relevance Propagation (LRP) is one of the most prominent methods in explainable machine learning (XML). This article will give you a good idea about the details of LRP and some tricks for implementing it. The content is largely based on this book chapter.

To make your mouth water about the potential of LRP, check out this interactive demo first.