A Benchmark for Interpretability Methods in Deep Neural Networks

Part of Advances in Neural Information Processing Systems 32 (NeurIPS 2019)

AuthorFeedback »Bibtex »Bibtex »MetaReview »Metadata »Paper »Reviews »Supplemental »


Sara Hooker, Dumitru Erhan, Pieter-Jan Kindermans, Been Kim


<p>We propose an empirical measure of the approximate accuracy of feature importance estimates in deep neural networks. Our results across several large-scale image classification datasets show that many popular interpretability methods produce estimates of feature importance that are not better than a random designation of feature importance. Only certain ensemble based approaches---VarGrad and SmoothGrad-Squared---outperform such a random assignment of importance. The manner of ensembling remains critical, we show that some approaches do no better then the underlying method but carry a far higher computational burden.</p>