Margin-distancing for safe model explanation

Tom Yan, Chicheng Zhang
Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, PMLR 151:5104-5134, 2022.

Abstract

The growing use of machine learning models in consequential settings has highlighted an important and seemingly irreconcilable tension between transparency and vulnerability to gaming. While this has sparked sizable debate in legal literature, there has been comparatively less technical study of this contention. In this work, we propose a clean-cut formulation of this tension and a way to make the tradeoff between transparency and gaming. We identify the source of gaming as being points close to the decision boundary of the model. And we initiate an investigation on how to provide example-based explanations that are expansive and yet consistent with a version space that is sufficiently uncertain with respect to the boundary points’ labels. Finally, we furnish our theoretical results with empirical investigations of this tradeoff on real-world datasets.

Cite this Paper

BibTeX
@InProceedings{pmlr-v151-yan22a, title = { Margin-distancing for safe model explanation }, author = {Yan, Tom and Zhang, Chicheng}, booktitle = {Proceedings of The 25th International Conference on Artificial Intelligence and Statistics}, pages = {5104--5134}, year = {2022}, editor = {Camps-Valls, Gustau and Ruiz, Francisco J. R. and Valera, Isabel}, volume = {151}, series = {Proceedings of Machine Learning Research}, month = {28--30 Mar}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v151/yan22a/yan22a.pdf}, url = {https://proceedings.mlr.press/v151/yan22a.html}, abstract = { The growing use of machine learning models in consequential settings has highlighted an important and seemingly irreconcilable tension between transparency and vulnerability to gaming. While this has sparked sizable debate in legal literature, there has been comparatively less technical study of this contention. In this work, we propose a clean-cut formulation of this tension and a way to make the tradeoff between transparency and gaming. We identify the source of gaming as being points close to the decision boundary of the model. And we initiate an investigation on how to provide example-based explanations that are expansive and yet consistent with a version space that is sufficiently uncertain with respect to the boundary points’ labels. Finally, we furnish our theoretical results with empirical investigations of this tradeoff on real-world datasets. } }
Endnote
%0 Conference Paper %T Margin-distancing for safe model explanation %A Tom Yan %A Chicheng Zhang %B Proceedings of The 25th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2022 %E Gustau Camps-Valls %E Francisco J. R. Ruiz %E Isabel Valera %F pmlr-v151-yan22a %I PMLR %P 5104--5134 %U https://proceedings.mlr.press/v151/yan22a.html %V 151 %X The growing use of machine learning models in consequential settings has highlighted an important and seemingly irreconcilable tension between transparency and vulnerability to gaming. While this has sparked sizable debate in legal literature, there has been comparatively less technical study of this contention. In this work, we propose a clean-cut formulation of this tension and a way to make the tradeoff between transparency and gaming. We identify the source of gaming as being points close to the decision boundary of the model. And we initiate an investigation on how to provide example-based explanations that are expansive and yet consistent with a version space that is sufficiently uncertain with respect to the boundary points’ labels. Finally, we furnish our theoretical results with empirical investigations of this tradeoff on real-world datasets.
APA
Yan, T. & Zhang, C.. (2022). Margin-distancing for safe model explanation . Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 151:5104-5134 Available from https://proceedings.mlr.press/v151/yan22a.html.

Related Material