Unobserved discrete data are ubiquitous in many scientific disciplines, and how to learn the
causal structure of these latent variables is crucial for uncovering data patterns. Most studies …