Statistical Limits of Causal Inference

Explore the fundamental limits of statistical estimation in causal inference through this comprehensive lecture by Sivaraman Balakrishnan from Carnegie Mellon University. Delve into the challenges of estimating causal effects from observational data across various scientific fields. Examine classical concepts and three distinct vignettes that investigate the inherent difficulties in causal effect estimation under different structural assumptions. Learn about the limitations of estimating personalized causal effects, derive rates for causal effect estimation without relying on smoothness assumptions, and understand the intrinsic challenges of estimation in discrete settings. No prior knowledge of causal inference is required for this insightful presentation, which is part of the Modern Paradigms in Generalization Boot Camp at the Simons Institute.