Skip to main navigation Skip to search Skip to main content

Perturbation Effect: A Metric to Counter Misleading Validation of Feature Attribution

Research output: Chapter in Book/Report/Conference proceeding › Conference paper › peer-review

Abstract

This paper provides evidence indicating that the most commonly used metric for validating feature attribution methods in eXplainable AI (XAI) is misleading when applied to time series data. To evaluate whether an XAI method attributes importance to relevant features, these are systematically perturbed while measuring the impact on the performance of the classifier. The assumption is that a drastic performance reduction with increasing perturbation of relevant features indicates that these are indeed relevant. We demonstrate empirically that this assumption is incomplete without considering low relevance features in the used metrics. We introduce a novel metric, the Perturbation Effect Size, and demonstrate how it complements existing metrics to offer a more faithful assessment of importance attribution. Finally, we contribute a comprehensive evaluation of attribution methods on time series data, considering the influence of perturbation methods and region size selection.
Original languageEnglish
Title of host publicationCIKM '22: Proceedings of the 31st ACM International Conference on Information & Knowledge Management
Place of PublicationNew York, NY
PublisherAssociation for Computing Machinery (ACM)
Pages1798-1807
ISBN (Electronic)978-1-4503-9236-5
DOIs
Publication statusPublished - 17 Oct 2022
Event31st ACM International Conference on Information and Knowledge Management: CIKM 2022 - Atlanta, United States
Duration: 17 Oct 2022 → 21 Oct 2022

Conference

Conference31st ACM International Conference on Information and Knowledge Management
Abbreviated titleCIKM '22
Country/TerritoryUnited States
CityAtlanta
Period17/10/22 → 21/10/22

Keywords

  • deep learning
  • explainable AI
  • trustworthy AI
  • feature attribution

Cite this