The ability of to explain neural network decisions goes hand in hand with their safe deployment. Several methods have been proposed to highlight features important for a given network decision. However, there is no consensus on how to measure effectiveness of these methods. We propose a new procedure for evaluating explanations. We use it to investigate visual explanations extracted from a range of possible sources in a neural network. We quantify the benefit of combining these sources and challenge a recent appeal for taking bias parameters into account. We support our conclusions with a general assessment of the impact of bias parameters in ImageNet classifiers
@article{arxiv.2003.08774,
title = {Measuring and improving the quality of visual explanations},
author = {Agnieszka Grabska-Barwińska},
journal= {arXiv preprint arXiv:2003.08774},
year = {2020}
}