Explanations can be manipulated and geometry is to blame
Ann-Kathrin Dombrowski, Maximillian Alber, Christopher Anders, Marcel Ackermann, Klaus-Robert Müller, Pan Kessel
Abstract
Explanation methods aim to make neural networks more trustworthy and interpretable. In this paper, we demonstrate a property of explanation methods which is disconcerting for both of these purposes. Namely, we show that explanations can be manipulated arbitrarily by applying visually hardly perceptible perturbations to the input that keep the network's output approximately constant. We establish theoretically that this phenomenon can be related to certain geometrical properties of neural networks. This allows us to derive an upper bound on the susceptibility of explanations to manipulations. Based on this result, we propose effective mechanisms to enhance the robustness of explanations.
BibTeX
@inproceedings{NEURIPS2019_bb836c01,
author = {Dombrowski, Ann-Kathrin and Alber, Maximillian and Anders, Christopher and Ackermann, Marcel and M\"{u}ller, Klaus-Robert and Kessel, Pan},
booktitle = {Advances in Neural Information Processing Systems},
editor = {H. Wallach and H. Larochelle and A. Beygelzimer and F. d\textquotesingle Alch\'{e}-Buc and E. Fox and R. Garnett},
pages = {},
publisher = {Curran Associates, Inc.},
title = {Explanations can be manipulated and geometry is to blame},
url = {https://proceedings.neurips.cc/paper_files/paper/2019/file/bb836c01cdc9120a9c984c525e4b1a4a-Paper.pdf},
volume = {32},
year = {2019}
}