DISCRET: Synthesizing Faithful Explanations For Treatment Effect Estimation

Yinjun Wu, Mayank Keoliya, Kan Chen, Neelay Velingker, Ziyang Li, Emily J Getzen, Qi Long, Mayur Naik, Ravi B Parikh, Eric Wong
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:53597-53618, 2024.

Abstract

Designing faithful yet accurate AI models is challenging, particularly in the field of individual treatment effect estimation (ITE). ITE prediction models deployed in critical settings such as healthcare should ideally be (i) accurate, and (ii) provide faithful explanations. However, current solutions are inadequate: state-of-the-art black-box models do not supply explanations, post-hoc explainers for black-box models lack faithfulness guarantees, and self-interpretable models greatly compromise accuracy. To address these issues, we propose DISCRET, a self-interpretable ITE framework that synthesizes faithful, rule-based explanations for each sample. A key insight behind DISCRET is that explanations can serve dually as database queries to identify similar subgroups of samples. We provide a novel RL algorithm to efficiently synthesize these explanations from a large search space. We evaluate DISCRET on diverse tasks involving tabular, image, and text data. DISCRET outperforms the best self-interpretable models and has accuracy comparable to the best black-box models while providing faithful explanations. DISCRET is available at https://github.com/wuyinjun-1993/DISCRET-ICML2024.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-wu24n, title = {{DISCRET}: Synthesizing Faithful Explanations For Treatment Effect Estimation}, author = {Wu, Yinjun and Keoliya, Mayank and Chen, Kan and Velingker, Neelay and Li, Ziyang and Getzen, Emily J and Long, Qi and Naik, Mayur and Parikh, Ravi B and Wong, Eric}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {53597--53618}, 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/wu24n/wu24n.pdf}, url = {https://proceedings.mlr.press/v235/wu24n.html}, abstract = {Designing faithful yet accurate AI models is challenging, particularly in the field of individual treatment effect estimation (ITE). ITE prediction models deployed in critical settings such as healthcare should ideally be (i) accurate, and (ii) provide faithful explanations. However, current solutions are inadequate: state-of-the-art black-box models do not supply explanations, post-hoc explainers for black-box models lack faithfulness guarantees, and self-interpretable models greatly compromise accuracy. To address these issues, we propose DISCRET, a self-interpretable ITE framework that synthesizes faithful, rule-based explanations for each sample. A key insight behind DISCRET is that explanations can serve dually as database queries to identify similar subgroups of samples. We provide a novel RL algorithm to efficiently synthesize these explanations from a large search space. We evaluate DISCRET on diverse tasks involving tabular, image, and text data. DISCRET outperforms the best self-interpretable models and has accuracy comparable to the best black-box models while providing faithful explanations. DISCRET is available at https://github.com/wuyinjun-1993/DISCRET-ICML2024.} }
Endnote
%0 Conference Paper %T DISCRET: Synthesizing Faithful Explanations For Treatment Effect Estimation %A Yinjun Wu %A Mayank Keoliya %A Kan Chen %A Neelay Velingker %A Ziyang Li %A Emily J Getzen %A Qi Long %A Mayur Naik %A Ravi B Parikh %A Eric Wong %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-wu24n %I PMLR %P 53597--53618 %U https://proceedings.mlr.press/v235/wu24n.html %V 235 %X Designing faithful yet accurate AI models is challenging, particularly in the field of individual treatment effect estimation (ITE). ITE prediction models deployed in critical settings such as healthcare should ideally be (i) accurate, and (ii) provide faithful explanations. However, current solutions are inadequate: state-of-the-art black-box models do not supply explanations, post-hoc explainers for black-box models lack faithfulness guarantees, and self-interpretable models greatly compromise accuracy. To address these issues, we propose DISCRET, a self-interpretable ITE framework that synthesizes faithful, rule-based explanations for each sample. A key insight behind DISCRET is that explanations can serve dually as database queries to identify similar subgroups of samples. We provide a novel RL algorithm to efficiently synthesize these explanations from a large search space. We evaluate DISCRET on diverse tasks involving tabular, image, and text data. DISCRET outperforms the best self-interpretable models and has accuracy comparable to the best black-box models while providing faithful explanations. DISCRET is available at https://github.com/wuyinjun-1993/DISCRET-ICML2024.
APA
Wu, Y., Keoliya, M., Chen, K., Velingker, N., Li, Z., Getzen, E.J., Long, Q., Naik, M., Parikh, R.B. & Wong, E.. (2024). DISCRET: Synthesizing Faithful Explanations For Treatment Effect Estimation. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:53597-53618 Available from https://proceedings.mlr.press/v235/wu24n.html.

Related Material