In plain words: A guide to figuring out why a deep learning model makes its predictions, showing which parts of the input it leans on. It explains the theory behind these techniques and gives tips for using them on real data, with practical examples.
Abstract
This paper provides an entry point to the problem of interpreting a deep neural network model and explaining its predictions. It is based on a tutorial given at ICASSP 2017. It introduces some recently proposed techniques of interpretation, along with theory, tricks and recommendations, to make most efficient use of these techniques on real data. It also discusses a number of practical applications.
Grégoire Montavon, Wojciech Samek, Klaus-Robert Müller
arXiv:1706.07979 · cs.LG, stat.ML · submitted Jun 24, 2017
abstract · pdf · html · 14 pages, 10 figures