ICML 2022spotlight71 citations
Causal structure-based root cause analysis of outliers
Kailash Budhathoki, Lenon Minorics, Patrick Bloebaum, Dominik Janzing
Abstract
Current techniques for explaining outliers cannot tell what caused the outliers. We present a formal method to identify "root causes" of outliers, amongst variables. The method requires a causal graph of the variables along with the functional causal model. It quantifies the contribution of each variable to the target outlier score, which explains to what extent each variable is a "root cause" of the target outlier. We study the empirical performance of the method through simulations and present a real-world case study identifying "root causes" of extreme river flows.
BibTeX
@InProceedings{pmlr-v162-budhathoki22a,
title = {Causal structure-based root cause analysis of outliers},
author = {Budhathoki, Kailash and Minorics, Lenon and Bloebaum, Patrick and Janzing, Dominik},
booktitle = {Proceedings of the 39th International Conference on Machine Learning},
pages = {2357--2369},
year = {2022},
editor = {Chaudhuri, Kamalika and Jegelka, Stefanie and Song, Le and Szepesvari, Csaba and Niu, Gang and Sabato, Sivan},
volume = {162},
series = {Proceedings of Machine Learning Research},
month = {17--23 Jul},
publisher = {PMLR},
pdf = {https://proceedings.mlr.press/v162/budhathoki22a/budhathoki22a.pdf},
url = {https://proceedings.mlr.press/v162/budhathoki22a.html},
abstract = {Current techniques for explaining outliers cannot tell what caused the outliers. We present a formal method to identify "root causes" of outliers, amongst variables. The method requires a causal graph of the variables along with the functional causal model. It quantifies the contribution of each variable to the target outlier score, which explains to what extent each variable is a "root cause" of the target outlier. We study the empirical performance of the method through simulations and present a real-world case study identifying "root causes" of extreme river flows.}
}