Calibrated Explanations for Multi-class

Tuwe Löfström, Helena Löfström, Ulf Johansson
Proceedings of the Thirteenth Symposium on Conformal and Probabilistic Prediction with Applications, PMLR 230:175-194, 2024.

Abstract

Calibrated Explanations is a recently proposed feature importance explanation method providing uncertainty quantification. It utilises Venn-Abers to generate well-calibrated factual and counterfactual explanations for binary classification. In this paper, we extend the method to support multi-class classification. The paper includes an evaluation illustrating the calibration quality of the selected multi-class calibration approach, as well as a demonstration of how the explanations can help determine which explanations to trust.

Cite this Paper

BibTeX
@InProceedings{pmlr-v230-lofstrom24a, title = {Calibrated Explanations for Multi-class}, author = {L\"{o}fstr\"{o}m, Tuwe and L\"{o}fstr\"{o}m, Helena and Johansson, Ulf}, booktitle = {Proceedings of the Thirteenth Symposium on Conformal and Probabilistic Prediction with Applications}, pages = {175--194}, year = {2024}, editor = {Vantini, Simone and Fontana, Matteo and Solari, Aldo and Boström, Henrik and Carlsson, Lars}, volume = {230}, series = {Proceedings of Machine Learning Research}, month = {09--11 Sep}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v230/main/assets/lofstrom24a/lofstrom24a.pdf}, url = {https://proceedings.mlr.press/v230/lofstrom24a.html}, abstract = {Calibrated Explanations is a recently proposed feature importance explanation method providing uncertainty quantification. It utilises Venn-Abers to generate well-calibrated factual and counterfactual explanations for binary classification. In this paper, we extend the method to support multi-class classification. The paper includes an evaluation illustrating the calibration quality of the selected multi-class calibration approach, as well as a demonstration of how the explanations can help determine which explanations to trust.} }
Endnote
%0 Conference Paper %T Calibrated Explanations for Multi-class %A Tuwe Löfström %A Helena Löfström %A Ulf Johansson %B Proceedings of the Thirteenth Symposium on Conformal and Probabilistic Prediction with Applications %C Proceedings of Machine Learning Research %D 2024 %E Simone Vantini %E Matteo Fontana %E Aldo Solari %E Henrik Boström %E Lars Carlsson %F pmlr-v230-lofstrom24a %I PMLR %P 175--194 %U https://proceedings.mlr.press/v230/lofstrom24a.html %V 230 %X Calibrated Explanations is a recently proposed feature importance explanation method providing uncertainty quantification. It utilises Venn-Abers to generate well-calibrated factual and counterfactual explanations for binary classification. In this paper, we extend the method to support multi-class classification. The paper includes an evaluation illustrating the calibration quality of the selected multi-class calibration approach, as well as a demonstration of how the explanations can help determine which explanations to trust.
APA
Löfström, T., Löfström, H. & Johansson, U.. (2024). Calibrated Explanations for Multi-class. Proceedings of the Thirteenth Symposium on Conformal and Probabilistic Prediction with Applications, in Proceedings of Machine Learning Research 230:175-194 Available from https://proceedings.mlr.press/v230/lofstrom24a.html.

Related Material