Abstract: Reconstructing the structure of causal graphical models from observational data is crucial for identifying causal mechanisms in scientific research. However, real-world noise and hidden ...