Causal Discovery via Conditional Independence Testing with Proxy Variables

arXiv:2305.05281v3 Announce Type: replace-cross
Abstract: Distinguishing causal connections from correlations is important in many scenarios. However, the presence of unobserved variables, such as the latent confounder, can introduce bias in conditional independence testing commonly employed in constraint-based causal discovery for identifying causal relations. To address this issue, existing methods introduced proxy variables to adjust for the bias caused by unobserveness. However, these methods were either limited to categorical variables or relied on strong parametric assumptions for identification. In this paper, …

abstract arxiv bias causal correlations cs.lg discovery however issue relations stat.me testing type variables via