PHYSICS267

(Formerly numbered PHYSICS 366) Foundations of principled inference from data, primarily in the Bayesian framework, with applications in astrophysics and cosmology. Topics include probabilistic modeling of data, parameter constraints and model comparison, numerical methods including Markov Chain Monte Carlo, and connections to frequentist and machine learning frameworks. The course is organized around tutorial notebooks using Python and Numpy, providing hands-on experience with real data. Prerequisites: programming in Python at the level of CS 106A or PHYSICS 113; probability at the level of STATS 116, CS 109, or PHYSICS 166/266; or permission of instructor. Normally offered every 2 years.