Reproduce “Causal models in string diagrams” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Robin Lorenz, Sean Tull
Robin Lorenz and Sean Tull from Quantinuum present a reformulation of the causal model framework using string diagrams and category theory, providing a unified and intuitive graphical language for causal reasoning. This approach offers enhanced conceptual clarity and rigor for representing causal structures, interventions, and counterfactuals, facilitating fully diagrammatic inference calculations and establishing foundations for advanced causal theory.
Key insights
- String diagrams provide an intuitive and rigorous mathematical language that elegantly unifies causal structure (DAGs) and stochastic/deterministic mechanisms.
- The introduction of "open causal models" allows for compositional reasoning about causal systems, representing expressions like P(Y; do(X)) as channels.
- Latent variables encoded in ADMGs can be formally incorporated and analyzed within the string diagram framework through a process of "rootification."
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.