CAUSE: Learning Granger Causality from Event Sequences using Attribution Methods

Wei Zhang, Thomas Panum, Somesh Jha, Prasad Chalasani, David Page
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:11235-11245, 2020.

Abstract

We study the problem of learning Granger causality between event types from asynchronous, interdependent, multi-type event sequences. Existing work suffers from either limited model flexibility or poor model explainability and thus fails to uncover Granger causality across a wide variety of event sequences with diverse event interdependency. To address these weaknesses, we propose CAUSE (Causality from AttribUtions on Sequence of Events), a novel framework for the studied task. The key idea of CAUSE is to first implicitly capture the underlying event interdependency by fitting a neural point process, and then extract from the process a Granger causality statistic using an axiomatic attribution method. Across multiple datasets riddled with diverse event interdependency, we demonstrate that CAUSE achieves superior performance on correctly inferring the inter-type Granger causality over a range of state-of-the-art methods.

Cite this Paper

BibTeX
@InProceedings{pmlr-v119-zhang20v, title = {{CAUSE}: Learning {G}ranger Causality from Event Sequences using Attribution Methods}, author = {Zhang, Wei and Panum, Thomas and Jha, Somesh and Chalasani, Prasad and Page, David}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {11235--11245}, 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/zhang20v/zhang20v.pdf}, url = {https://proceedings.mlr.press/v119/zhang20v.html}, abstract = {We study the problem of learning Granger causality between event types from asynchronous, interdependent, multi-type event sequences. Existing work suffers from either limited model flexibility or poor model explainability and thus fails to uncover Granger causality across a wide variety of event sequences with diverse event interdependency. To address these weaknesses, we propose CAUSE (Causality from AttribUtions on Sequence of Events), a novel framework for the studied task. The key idea of CAUSE is to first implicitly capture the underlying event interdependency by fitting a neural point process, and then extract from the process a Granger causality statistic using an axiomatic attribution method. Across multiple datasets riddled with diverse event interdependency, we demonstrate that CAUSE achieves superior performance on correctly inferring the inter-type Granger causality over a range of state-of-the-art methods.} }
Endnote
%0 Conference Paper %T CAUSE: Learning Granger Causality from Event Sequences using Attribution Methods %A Wei Zhang %A Thomas Panum %A Somesh Jha %A Prasad Chalasani %A David Page %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-zhang20v %I PMLR %P 11235--11245 %U https://proceedings.mlr.press/v119/zhang20v.html %V 119 %X We study the problem of learning Granger causality between event types from asynchronous, interdependent, multi-type event sequences. Existing work suffers from either limited model flexibility or poor model explainability and thus fails to uncover Granger causality across a wide variety of event sequences with diverse event interdependency. To address these weaknesses, we propose CAUSE (Causality from AttribUtions on Sequence of Events), a novel framework for the studied task. The key idea of CAUSE is to first implicitly capture the underlying event interdependency by fitting a neural point process, and then extract from the process a Granger causality statistic using an axiomatic attribution method. Across multiple datasets riddled with diverse event interdependency, we demonstrate that CAUSE achieves superior performance on correctly inferring the inter-type Granger causality over a range of state-of-the-art methods.
APA
Zhang, W., Panum, T., Jha, S., Chalasani, P. & Page, D.. (2020). CAUSE: Learning Granger Causality from Event Sequences using Attribution Methods. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:11235-11245 Available from https://proceedings.mlr.press/v119/zhang20v.html.

Related Material