|
Abstract:
|
Bayesian inference of causal effects has received increasing interest in recent years, fueled by the vast development in Bayesian modeling, causal inference and machine learning. There are some unsettled issues regarding the probabilistic and statistical foundations of Bayesian causal inference, e.g. the role of propensity score in inference under ignorability, prior dogmatism, separation of design and analysis, and choice of priors. Despite these, much methodological and empirical advancement in this field has been achieved, and new applications are rapidly emerging. Examples include environmental science, spatial statistics, clinical trial design and analysis, imaging analysis. This roundtable brings experts and consumers of Bayesian causal inference to discuss the basics, challenges, opportunities in theory and practice, as well as the aforementioned foundational problems in this field.
|
