Popis:
In this tutorial-like talk we introduce a causal semantics of Bayesian networks suggested by Judea Pearl (1993) and Spirtes et al. (1993). We will illustrate the difference between an observation and an intervention in the causal model using an example of Simpson's paradox. We will also discuss the problem of identifying the effects of interventions in causal models with latent variables.
The talk is based on: Steffen Lauritzen. Causal Inference from Graphical Models. Research Report R-99-2021, Department of Mathematics, Aalborg University. Available at: http://www.utia.cas.cz/vomlel/R-99-2021.ps and Chapter 21 of Daphne Koller and Nir Friedman. Probabilistic Graphical Models: Principles and Techniques, MIT Press, 2009.