Learning Bayesian Statistics BITESIZE | How Probability Becomes Causality?
8 snips
Sep 24, 2025 In this engaging discussion, Sam Witty, a researcher from the Cairo project, dives into the fascinating world of causal inference. He explains the differences between do-calculus and Cairo’s parametric Bayesian methods, and how regression discontinuity designs enable causal estimation. Sam also explores how Cairo automates the construction of interventions, providing users easy access to complex statistical tools. The talk highlights the significance of efficient estimators, making causal queries more accessible without needing extensive expertise.
AI Snips
Chapters
Transcript
Episode notes
Do-Calculus Is Nonparametric By Design
- Do-calculus translates causal queries from a graph into probabilistic expressions for nonparametric structural causal models.
- Parametric assumptions can often make causal inference easier than purely nonparametric approaches.
Regression Discontinuity Example
- Sam uses regression discontinuity as an example where parametric smoothness assumptions enable causal identification near a threshold.
- He explains comparing units just left and right of the cutoff gives an estimate of the conditional average treatment effect.
Treat Causal Models As Bayesian Objects
- Bayesian structural causal inference treats a distribution over causal models and updates it with observed data using ordinary Bayesian conditioning.
- Interventions map models to intervened models and let you derive posterior distributions over counterfactuals and causal queries.