Evaluating Input Perturbation Methods for Interpreting CNNs and Saliency Map Comparison
This is an incremental study for researchers in interpretable AI, highlighting reliability issues in existing saliency map techniques.
The paper tackles the problem of input perturbation methods for interpreting CNNs and evaluating saliency maps, showing that neutral baseline images and hyperparameter choices impact results and cause inconsistencies, revealing a lack of robustness in these methods.
Input perturbation methods occlude parts of an input to a function and measure the change in the function's output. Recently, input perturbation methods have been applied to generate and evaluate saliency maps from convolutional neural networks. In practice, neutral baseline images are used for the occlusion, such that the baseline image's impact on the classification probability is minimal. However, in this paper we show that arguably neutral baseline images still impact the generated saliency maps and their evaluation with input perturbations. We also demonstrate that many choices of hyperparameters lead to the divergence of saliency maps generated by input perturbations. We experimentally reveal inconsistencies among a selection of input perturbation methods and find that they lack robustness for generating saliency maps and for evaluating saliency maps as saliency metrics.