Game-theoretic statistical inference

(betting, e-values and martingales --- history, philosophy, statistics and probability)

Location/time: Zoom from 10am to 1pm ET (with a break from 11:45am to noon), every Friday.

This class will be recorded due to differing university semester start and end dates.

Glenn Shafer (Rutgers), Ruodu Wang (Waterloo), Aaditya Ramdas (Stat/ML)



Course description We will study the game-theoretic foundation for statistics and its applications to prediction, testing and estimation. It will distinctively employ modern concepts around betting, e-values, and martingales, effectively bridging Bayesian, frequentist and model-free or adversarial perspectives on these topics.

Target audience: PhD students who are interested in some subset of these topics.

Prerequisites: Advanced undergraduate, or basic graduate probability/statistics.



Additional Links

SAVI (safe, anytime-valid inference) workshop --- Jun 28 to Jul 2, 2021 in Eindhoven.