Description: Statistical inference is the de facto tool in data science to carry out hypothesis testing and draw conclusions under uncertainty. With increasingly diverse stakeholders relying on inference as data-driven solutions, to study, decipher, and articulate its strength and limitation become more important than ever. In this course, we will discuss fundamental issues in statistical inference, partly in response to a range of daunting challenges posed by modern data science such as reproducibility and interpretability at large scales. A sample of topics includes the use of p-values vs. Bayes factors, frequentist properties of Bayesian procedures for both parametric and nonparametric models, Bernstein von-Mises phenomena, variable/feature selection, post-selection inference, false discovery control. Recommended Prerequisite(s): STAT 532/533 or equivalent courses on classical statistical inference, and STAT525 or equivalent courses on Bayesian inference.