Feature Attribution with Necessity and Sufficiency via Dual-stage Perturbation Test for Causal Explanation

Xuexin Chen, Ruichu Cai, Zhengting Huang, Yuxuan Zhu, Julien Horwood, Zhifeng Hao, Zijian Li, José Miguel Hernández-Lobato
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:6486-6502, 2024.

Abstract

We investigate the problem of explainability for machine learning models, focusing on Feature Attribution Methods (FAMs) that evaluate feature importance through perturbation tests. Despite their utility, FAMs struggle to distinguish the contributions of different features, when their prediction changes are similar after perturbation. To enhance FAMs’ discriminative power, we introduce Feature Attribution with Necessity and Sufficiency (FANS), which find a neighborhood of the input such that perturbing samples within this neighborhood have a high Probability of being Necessity and Sufficiency (PNS) cause for the change in predictions, and use this PNS as the importance of the feature. Specifically, FANS compute this PNS via a heuristic strategy for estimating the neighborhood and a perturbation test involving two stages (factual and interventional) for counterfactual reasoning. To generate counterfactual samples, we use a resampling-based approach on the observed samples to approximate the required conditional distribution. We demonstrate that FANS outperforms existing attribution methods on six benchmarks. Please refer to the source code via https://github.com/DMIRLAB-Group/FANS.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-chen24d, title = {Feature Attribution with Necessity and Sufficiency via Dual-stage Perturbation Test for Causal Explanation}, author = {Chen, Xuexin and Cai, Ruichu and Huang, Zhengting and Zhu, Yuxuan and Horwood, Julien and Hao, Zhifeng and Li, Zijian and Hern\'{a}ndez-Lobato, Jos\'{e} Miguel}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {6486--6502}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/chen24d/chen24d.pdf}, url = {https://proceedings.mlr.press/v235/chen24d.html}, abstract = {We investigate the problem of explainability for machine learning models, focusing on Feature Attribution Methods (FAMs) that evaluate feature importance through perturbation tests. Despite their utility, FAMs struggle to distinguish the contributions of different features, when their prediction changes are similar after perturbation. To enhance FAMs’ discriminative power, we introduce Feature Attribution with Necessity and Sufficiency (FANS), which find a neighborhood of the input such that perturbing samples within this neighborhood have a high Probability of being Necessity and Sufficiency (PNS) cause for the change in predictions, and use this PNS as the importance of the feature. Specifically, FANS compute this PNS via a heuristic strategy for estimating the neighborhood and a perturbation test involving two stages (factual and interventional) for counterfactual reasoning. To generate counterfactual samples, we use a resampling-based approach on the observed samples to approximate the required conditional distribution. We demonstrate that FANS outperforms existing attribution methods on six benchmarks. Please refer to the source code via https://github.com/DMIRLAB-Group/FANS.} }
Endnote
%0 Conference Paper %T Feature Attribution with Necessity and Sufficiency via Dual-stage Perturbation Test for Causal Explanation %A Xuexin Chen %A Ruichu Cai %A Zhengting Huang %A Yuxuan Zhu %A Julien Horwood %A Zhifeng Hao %A Zijian Li %A José Miguel Hernández-Lobato %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-chen24d %I PMLR %P 6486--6502 %U https://proceedings.mlr.press/v235/chen24d.html %V 235 %X We investigate the problem of explainability for machine learning models, focusing on Feature Attribution Methods (FAMs) that evaluate feature importance through perturbation tests. Despite their utility, FAMs struggle to distinguish the contributions of different features, when their prediction changes are similar after perturbation. To enhance FAMs’ discriminative power, we introduce Feature Attribution with Necessity and Sufficiency (FANS), which find a neighborhood of the input such that perturbing samples within this neighborhood have a high Probability of being Necessity and Sufficiency (PNS) cause for the change in predictions, and use this PNS as the importance of the feature. Specifically, FANS compute this PNS via a heuristic strategy for estimating the neighborhood and a perturbation test involving two stages (factual and interventional) for counterfactual reasoning. To generate counterfactual samples, we use a resampling-based approach on the observed samples to approximate the required conditional distribution. We demonstrate that FANS outperforms existing attribution methods on six benchmarks. Please refer to the source code via https://github.com/DMIRLAB-Group/FANS.
APA
Chen, X., Cai, R., Huang, Z., Zhu, Y., Horwood, J., Hao, Z., Li, Z. & Hernández-Lobato, J.M.. (2024). Feature Attribution with Necessity and Sufficiency via Dual-stage Perturbation Test for Causal Explanation. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:6486-6502 Available from https://proceedings.mlr.press/v235/chen24d.html.

Related Material