Scrutinizing XAI using linear ground-truth data with suppressor variables

DOI10.1007/S10994-022-06167-YMaRDI QIDQ2163233zbMATH OpenOpenAlexWikidataFDO

Authors Rick Wilming, Céline Budding, Klaus-Robert Müller, Stefan Haufe

Publication date 10 August 2022

Published in Machine Learning (Search for Journal in Brave)

Copyright license Creative Commons Attribution 4.0 International

Full work available at URL https://arxiv.org/abs/2111.07473

zbMATH Keywords

benchmarklinear classificationground truthexplainable AIsaliency methodssuppressor variables

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05)

Recommendations

• Correction of AI systems by linear discriminants: probabilistic foundations
• Provable training set debugging for linear regression
• Testing for neglected nonlinearity using extreme learning machines
• Linear classifiers are nearly optimal when hidden variables have diverse effects
• Exact learning when irrelevant variables abound
• Towards formal XAI: formally approximate minimal explanations of neural networks
• Debiased machine learning of set-identified linear models

Cites work

• scientific article; zbMATH DE number 3078997 (Why is no real title available?)
• All models are wrong, but many are useful: learning a variable's importance by studying an entire class of prediction models simultaneously
• Definitions, methods, and applications in interpretable machine learning
• How to explain individual classification decisions
• Random forests
• Scikit-learn: machine learning in Python

Cited in

(1)

• Stable graph based decision route explanation in siamese neural networks

Describes a project that uses

Uses Software

• AlexNet
• shap
• ImageNet
• Scikit

This page was built for publication: Scrutinizing XAI using linear ground-truth data with suppressor variables

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2163233)