Resarch Group in Mathematical Statistics
Center for Mathematical Sciences, University of Cambridge
`Uncertainty Quantification and Modern Statistical Inference (UQMSI)'
Our research group investigates the mathematical underpinnings of inference problems arising in contemporary statistics and data science, with a particular focus on uncertainty quantification methodology (such as confidence statements and tests) in high- and infinite dimensional statistical models. The mathematics involved lies at the intersection of statistics, probability theory and analysis. Currently (2015-2021) we are supported by a Consolidator Grant of the European Research Council (ERC), and also by CCIMI.
NEW: Online talk at ETH Zurich, 30/09/2020
NEW: Online talk at MIT, 02/10/2020
Research topics include:
statistics for PDEs, inverse problems & stochastic processes
Bayesian nonparametrics
confidence regions for high-dimensional statistical models
concentration of measure and empirical process theory
Current group members:
Some former group members:
Kweku Abraham, Alexandra Carpentier, Alberto Coca, Adam Kashlak, Hanne Kekkonen, Matthias Löffler, Karim Lounici, Jakob Söhl, Kolyan Ray, Bharath Sriperumpudur, Elodie Vernet
Past activities:
Zoom Conference on Mathematical and Statistical Challenges in Uncertainty Quantification, July 2020
BNP12 meeting 2019 in Oxford
Birthday conference for Aad van der Vaart, June 2019
Young Researchers meeting in Mathematical Statistics, Paris, September 24-26, 2018
2018 Uncertainty quantification in complex, nonparametric statistical models (Lorentz Centre)
2017 Meeting in Mathematical Statistics (CIRM)
June 2017 Workshop on Statistical foundations of uncertainty quantification for inverse problems
2017 YES VIII Workshop on Uncertainty Quantification
Course on Concentration Inequalities by Stephane Boucheron
2015 Oberwolfach Workshop Probabilistic Techniques in Modern Statistics