On Counterfactual Explanations under Predictive Multiplicity

Martin Pawelczyk, Klaus Broelemann, Gjergji. Kasneci
Proceedings of the 36th Conference on Uncertainty in Artificial Intelligence (UAI), PMLR 124:809-818, 2020.

Abstract

Counterfactual explanations are usually obtainedby identifying the smallest change made to an input to change a prediction made by a fixed model (hereafter called sparse methods). Recent work, however, has revitalized an old insight: there often does not exist one superior solution to a prediction problem with respect to commonly used measures of interest (e.g. error rate). In fact, often multiple different classifiers give almost equal solutions. This phenomenon is known as predictive multiplicity (Breiman, 2001; Marx et al., 2019). In this work, we derive a general upper bound for the costs of counterfactual explanations under predictive multiplicity. Most notably, it depends on a discrepancy notion between two classifiers, which describes how differently they treat negatively predicted individuals. We then compare sparse and data support approaches empirically on real-world data. The results show that data support methods are more robust to multiplicity of different models. At the same time, we show that those methods have provably higher cost of generating counterfactual explanations under one fixed model. In summary, our theoretical and empirical results challenge the commonly held view that counterfactual recommendations should be sparse in general.

Cite this Paper

BibTeX
@InProceedings{pmlr-v124-pawelczyk20a, title = {On Counterfactual Explanations under Predictive Multiplicity}, author = {Pawelczyk, Martin and Broelemann, Klaus and Kasneci, Gjergji.}, booktitle = {Proceedings of the 36th Conference on Uncertainty in Artificial Intelligence (UAI)}, pages = {809--818}, year = {2020}, editor = {Peters, Jonas and Sontag, David}, volume = {124}, series = {Proceedings of Machine Learning Research}, month = {03--06 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v124/pawelczyk20a/pawelczyk20a.pdf}, url = {https://proceedings.mlr.press/v124/pawelczyk20a.html}, abstract = {Counterfactual explanations are usually obtainedby identifying the smallest change made to an input to change a prediction made by a fixed model (hereafter called sparse methods). Recent work, however, has revitalized an old insight: there often does not exist one superior solution to a prediction problem with respect to commonly used measures of interest (e.g. error rate). In fact, often multiple different classifiers give almost equal solutions. This phenomenon is known as predictive multiplicity (Breiman, 2001; Marx et al., 2019). In this work, we derive a general upper bound for the costs of counterfactual explanations under predictive multiplicity. Most notably, it depends on a discrepancy notion between two classifiers, which describes how differently they treat negatively predicted individuals. We then compare sparse and data support approaches empirically on real-world data. The results show that data support methods are more robust to multiplicity of different models. At the same time, we show that those methods have provably higher cost of generating counterfactual explanations under one fixed model. In summary, our theoretical and empirical results challenge the commonly held view that counterfactual recommendations should be sparse in general.} }
Endnote
%0 Conference Paper %T On Counterfactual Explanations under Predictive Multiplicity %A Martin Pawelczyk %A Klaus Broelemann %A Gjergji. Kasneci %B Proceedings of the 36th Conference on Uncertainty in Artificial Intelligence (UAI) %C Proceedings of Machine Learning Research %D 2020 %E Jonas Peters %E David Sontag %F pmlr-v124-pawelczyk20a %I PMLR %P 809--818 %U https://proceedings.mlr.press/v124/pawelczyk20a.html %V 124 %X Counterfactual explanations are usually obtainedby identifying the smallest change made to an input to change a prediction made by a fixed model (hereafter called sparse methods). Recent work, however, has revitalized an old insight: there often does not exist one superior solution to a prediction problem with respect to commonly used measures of interest (e.g. error rate). In fact, often multiple different classifiers give almost equal solutions. This phenomenon is known as predictive multiplicity (Breiman, 2001; Marx et al., 2019). In this work, we derive a general upper bound for the costs of counterfactual explanations under predictive multiplicity. Most notably, it depends on a discrepancy notion between two classifiers, which describes how differently they treat negatively predicted individuals. We then compare sparse and data support approaches empirically on real-world data. The results show that data support methods are more robust to multiplicity of different models. At the same time, we show that those methods have provably higher cost of generating counterfactual explanations under one fixed model. In summary, our theoretical and empirical results challenge the commonly held view that counterfactual recommendations should be sparse in general.
APA
Pawelczyk, M., Broelemann, K. & Kasneci, G.. (2020). On Counterfactual Explanations under Predictive Multiplicity. Proceedings of the 36th Conference on Uncertainty in Artificial Intelligence (UAI), in Proceedings of Machine Learning Research 124:809-818 Available from https://proceedings.mlr.press/v124/pawelczyk20a.html.

Related Material