Cost-effectively Identifying Causal Effects When Only Response Variable is Observable

Tian-Zuo Wang, Xi-Zhu Wu, Sheng-Jun Huang, Zhi-Hua Zhou
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:10060-10069, 2020.

Abstract

In many real tasks, we care about how to make decisions rather than mere predictions on an event, e.g. how to increase the revenue next month instead of merely knowing it will drop. The key is to identify the causal effects on the desired event. It is achievable with do-calculus if the causal structure is known; however, in many real tasks it is not easy to infer the whole causal structure with the observational data. Introducing external interventions is needed to achieve it. In this paper, we study the situation where only the response variable is observable under intervention. We propose a novel approach which is able to cost-effectively identify the causal effects, by an active strategy introducing limited interventions, and thus guide decision-making. Theoretical analysis and empirical studies validate the effectiveness of the proposed approach.

Cite this Paper

BibTeX
@InProceedings{pmlr-v119-wang20w, title = {Cost-effectively Identifying Causal Effects When Only Response Variable is Observable}, author = {Wang, Tian-Zuo and Wu, Xi-Zhu and Huang, Sheng-Jun and Zhou, Zhi-Hua}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {10060--10069}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/wang20w/wang20w.pdf}, url = {https://proceedings.mlr.press/v119/wang20w.html}, abstract = {In many real tasks, we care about how to make decisions rather than mere predictions on an event, e.g. how to increase the revenue next month instead of merely knowing it will drop. The key is to identify the causal effects on the desired event. It is achievable with do-calculus if the causal structure is known; however, in many real tasks it is not easy to infer the whole causal structure with the observational data. Introducing external interventions is needed to achieve it. In this paper, we study the situation where only the response variable is observable under intervention. We propose a novel approach which is able to cost-effectively identify the causal effects, by an active strategy introducing limited interventions, and thus guide decision-making. Theoretical analysis and empirical studies validate the effectiveness of the proposed approach.} }
Endnote
%0 Conference Paper %T Cost-effectively Identifying Causal Effects When Only Response Variable is Observable %A Tian-Zuo Wang %A Xi-Zhu Wu %A Sheng-Jun Huang %A Zhi-Hua Zhou %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-wang20w %I PMLR %P 10060--10069 %U https://proceedings.mlr.press/v119/wang20w.html %V 119 %X In many real tasks, we care about how to make decisions rather than mere predictions on an event, e.g. how to increase the revenue next month instead of merely knowing it will drop. The key is to identify the causal effects on the desired event. It is achievable with do-calculus if the causal structure is known; however, in many real tasks it is not easy to infer the whole causal structure with the observational data. Introducing external interventions is needed to achieve it. In this paper, we study the situation where only the response variable is observable under intervention. We propose a novel approach which is able to cost-effectively identify the causal effects, by an active strategy introducing limited interventions, and thus guide decision-making. Theoretical analysis and empirical studies validate the effectiveness of the proposed approach.
APA
Wang, T., Wu, X., Huang, S. & Zhou, Z.. (2020). Cost-effectively Identifying Causal Effects When Only Response Variable is Observable. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:10060-10069 Available from https://proceedings.mlr.press/v119/wang20w.html.

Related Material