A causal graph is essentially a map that shows how different variables impact each other. It’s a great tool for understanding cause-and-effect relationships.
What’s neat about causal graphs is that you can simulate directly changing a variable then observe what happens to the others. This act of directly manipulating a variable is called and “intervention” in causal inference, and it can be denoted as .
(Rao et al., 2021, p. 4)
Reference
Rao, Y., Chen, G., Lu, J., & Zhou, J. (2021). Counterfactual Attention Learning for Fine-Grained Visual Categorization and Re-identification. https://doi.org/10.48550/arXiv.2108.08728