A New Approach to Backtracking Counterfactual Explanations: A Unified Causal Framework for Efficient Model Interpretability

Pouria Fatemi, Ehsan Sharifian, Mohammad Hossein Yassaee
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:16273-16285, 2025.

Abstract

Counterfactual explanations enhance interpretability by identifying alternative inputs that produce different outputs, offering localized insights into model decisions. However, traditional methods often neglect causal relationships, leading to unrealistic examples. While newer approaches integrate causality, they are computationally expensive. To address these challenges, we propose an efficient method called BRACE based on backtracking counterfactuals that incorporates causal reasoning to generate actionable explanations. We first examine the limitations of existing methods and then introduce our novel approach and its features. We also explore the relationship between our method and previous techniques, demonstrating that it generalizes them in specific scenarios. Finally, experiments show that our method provides deeper insights into model outputs.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-fatemi25a, title = {A New Approach to Backtracking Counterfactual Explanations: A Unified Causal Framework for Efficient Model Interpretability}, author = {Fatemi, Pouria and Sharifian, Ehsan and Yassaee, Mohammad Hossein}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {16273--16285}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/fatemi25a/fatemi25a.pdf}, url = {https://proceedings.mlr.press/v267/fatemi25a.html}, abstract = {Counterfactual explanations enhance interpretability by identifying alternative inputs that produce different outputs, offering localized insights into model decisions. However, traditional methods often neglect causal relationships, leading to unrealistic examples. While newer approaches integrate causality, they are computationally expensive. To address these challenges, we propose an efficient method called BRACE based on backtracking counterfactuals that incorporates causal reasoning to generate actionable explanations. We first examine the limitations of existing methods and then introduce our novel approach and its features. We also explore the relationship between our method and previous techniques, demonstrating that it generalizes them in specific scenarios. Finally, experiments show that our method provides deeper insights into model outputs.} }
Endnote
%0 Conference Paper %T A New Approach to Backtracking Counterfactual Explanations: A Unified Causal Framework for Efficient Model Interpretability %A Pouria Fatemi %A Ehsan Sharifian %A Mohammad Hossein Yassaee %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-fatemi25a %I PMLR %P 16273--16285 %U https://proceedings.mlr.press/v267/fatemi25a.html %V 267 %X Counterfactual explanations enhance interpretability by identifying alternative inputs that produce different outputs, offering localized insights into model decisions. However, traditional methods often neglect causal relationships, leading to unrealistic examples. While newer approaches integrate causality, they are computationally expensive. To address these challenges, we propose an efficient method called BRACE based on backtracking counterfactuals that incorporates causal reasoning to generate actionable explanations. We first examine the limitations of existing methods and then introduce our novel approach and its features. We also explore the relationship between our method and previous techniques, demonstrating that it generalizes them in specific scenarios. Finally, experiments show that our method provides deeper insights into model outputs.
APA
Fatemi, P., Sharifian, E. & Yassaee, M.H.. (2025). A New Approach to Backtracking Counterfactual Explanations: A Unified Causal Framework for Efficient Model Interpretability. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:16273-16285 Available from https://proceedings.mlr.press/v267/fatemi25a.html.

Related Material