Evaluation of Induced Expert Knowledge in Causal Structure Learning by NOTEARS

Jawad Chowdhury, Rezaur Rashid, Gabriel Terejanu

2023

Abstract

Causal modeling provides us with powerful counterfactual reasoning and interventional mechanism to generate predictions and reason under various what-if scenarios. However, causal discovery using observation data remains a nontrivial task due to unobserved confounding factors, finite sampling, and changes in the data distribution. These can lead to spurious cause-effect relationships. To mitigate these challenges in practice, researchers augment causal learning with known causal relations. The goal of the paper is to study the impact of expert knowledge on causal relations in the form of additional constraints used in the formulation of the nonparametric NOTEARS. We provide a comprehensive set of comparative analyses of biasing the model using different types of knowledge. We found that (i) knowledge that correct the mistakes of the NOTEARS model can lead to statistically significant improvements, (ii) constraints on active edges have a larger positive impact on causal discovery than inactive edges, and surprisingly, (iii) the induced knowledge does not correct on average more incorrect active and/or inactive edges than expected. We also demonstrate the behavior of the model and the effectiveness of domain knowledge on a real-world dataset.

Paper Citation

in Harvard Style

Chowdhury J., Rashid R. and Terejanu G. (2023). Evaluation of Induced Expert Knowledge in Causal Structure Learning by NOTEARS. In Proceedings of the 12th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM, ISBN 978-989-758-626-2, pages 136-146. DOI: 10.5220/0011716000003411

in Bibtex Style

@conference{icpram23,
author={Jawad Chowdhury and Rezaur Rashid and Gabriel Terejanu},
title={Evaluation of Induced Expert Knowledge in Causal Structure Learning by NOTEARS},
booktitle={Proceedings of the 12th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,},
year={2023},
pages={136-146},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011716000003411},
isbn={978-989-758-626-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 12th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,
TI - Evaluation of Induced Expert Knowledge in Causal Structure Learning by NOTEARS
SN - 978-989-758-626-2
AU - Chowdhury J.
AU - Rashid R.
AU - Terejanu G.
PY - 2023
SP - 136
EP - 146
DO - 10.5220/0011716000003411