Peter Grünwald was recognized for his foundational work in statistics and statistical learning theory, including the development of e-values, advances in Bayesian inference under model misspecification, and research on minimum description length (MDL) learning. The MDL principle, which Grünwald helped formalize in his widely used textbook The Minimum Description Length Principle, is a method for model selection based on data compression. It aims to find models that provide the shortest possible description of the data, striking a balance between simplicity and accuracy.
In recent years, Grünwald has shifted his focus to core statistical theory. Since 2020, he has been developing the concept of e-values, a new mathematical foundation for statistical testing. This work was recently awarded an ERC Advanced Grant and can, for example, lead to more reliable methods for determining statistical significance in scientific research.
Grünwald is a member of the management team at CWI and holds a professorship in Statistical Learning at the Mathematical Institute of Leiden University.
About the IMS Fellowship
For 90 years, the title of IMS Fellow has represented a mark of distinction in the statistics and probability community. Fellows are selected for excellence in research or for leadership that has had a lasting impact on the field. The IMS, founded in 1935, has more than 4,900 members worldwide. Approximately 10% of current members hold the title of Fellow.
The full list of newly elected 2025 IMS Fellows is available here.
