We consider likelihood score-based methods for causal discovery in structural causal models. In particular, we focus on Gaussian scoring and analyze the effect of model misspecification in terms of non-Gaussian error distribution. We present a surprising negative result for Gaussian likelihood scoring in combination with nonparametric regression methods.
@article{arxiv.2210.11104,
title = {On the pitfalls of Gaussian likelihood scoring for causal discovery},
author = {Christoph Schultheiss and Peter Bühlmann},
journal= {arXiv preprint arXiv:2210.11104},
year = {2023}
}