The d-separation criterion in Categorical Probability — Tobias Fritz & Andreas Klingler (2022). arXiv:2207.05740 (v3, PDF); J. Mach. Learn. Res. 24(46), 1-49 (2023).
Introduces categorical causal models and a topological notion of d-separation in Markov categories and proves an abstract d-separation criterion, covering measure-theoretic, deterministic and possibilistic networks.
Sources: the paper, arXiv:2207.05740v3, checked against the arXiv listing. Index: Papers.
Key definitions and results
- Categorical d-separation; equivalence of local and global Markov properties with causal compatibility