AttributionLab: Faithfulness of Feature Attribution Under Controllable Environments

arXiv:2310.06514v2 Announce Type: replace
Abstract: Feature attribution explains neural network outputs by identifying relevant input features. The attribution has to be faithful, meaning that the attributed features must mirror the input features that influence the output. One recent trend to test faithfulness is to fit a model on designed data with known relevant features and then compare attributions with ground truth input features.This idea assumes that the model learns to use all and only these designed features, for which there …

abstract arxiv attribution cs.lg data environments feature features influence meaning network neural network test trend type