Peter Grünwald

Peter Grünwald is a Full Professor of Statistical Learning at Leiden University's Mathematical Institute and senior researcher in the machine learning group at Centrum Wiskunde & Informatica (CWI) in Amsterdam. His pioneering work focuses on developing flexible statistical inference frameworks using e-values and e-processes, creating a novel middle ground between Bayesian and frequentist paradigms that allows optional continuation in data analysis.

Grünwald's research revolutionizes statistical methodology through:

• Development of anytime-valid tests and confidence intervals
• Novel approaches to hypothesis testing with data-driven significance levels
• E-value based methods for safe sequential testing
• Applications in clinical trials, causal inference, and high-dimensional data

His publications demonstrate consistent focus on developing robust statistical frameworks applicable to flexible experimental designs, with recent work appearing in PNAS, JRSSB, and Annals of Statistics. Research consistently addresses fundamental challenges in statistical evidence evaluation across medical, social, and computational domains.

Major Scientific Recognition:

• ERC Advanced Grant (2024) for revolutionary statistical theory development
• NWO Top Award (2016) for outstanding contributions to statistics

As primary advisor, Grünwald has supervised PhD candidates including Tyron Lardy, Yunda Hao, and Rosanne Turner on topics spanning e-statistics, conditional independence testing, and stratified data analysis. He leads methodological development for the R package 'safestats' and collaborates internationally on statistical foundations.