Rate My Professor Patrick Rebeschini

Professor Patrick Rebeschini is the Statutory Professor of Statistical Science in the Department of Statistics at the University of Oxford, a position held since 2026. He is also Professorial Fellow at St Anne’s College, Oxford, since 2026. Previously, he was Professor of Statistics and Machine Learning from 2023 to 2026 and Associate Professor from 2017 to 2023 in the Department of Statistics, as well as Tutorial Fellow at University College, Oxford, from 2017 to 2026. Before joining Oxford, Rebeschini held positions at Yale University, including Associate Research Scientist in the Electrical Engineering Department and Lecturer in the Computer Science Department in 2016–2017, and Postdoctoral Associate at the Yale Institute for Network Science from 2014 to 2016. He received his Ph.D. in Operations Research and Financial Engineering from Princeton University in 2014, with a thesis titled 'Nonlinear filtering in high dimension' advised by Ramon van Handel. His earlier qualifications include an M.A. in Operations Research and Financial Engineering from Princeton in 2011, an M.S. in Theoretical Physics from the University of Padova in 2009, an Imperial College International Diploma in Physics in 2006, and a B.S. in Physics from the University of Padova in 2006.

Rebeschini's research specializes in the intersection of probability, statistics, and computer science, focusing on high-dimensional probability, statistics, and optimization to design efficient and optimal machine learning algorithms. Core interests encompass implicit regularization in machine learning, generalization bounds, optimization techniques like mirror descent and stochastic gradient descent, statistical learning theory, reinforcement learning, bandit problems, decentralized and distributed learning, network optimization, phase retrieval, and high-dimensional filtering. Key publications include 'Can local particle filters beat the curse of dimensionality?' with Ramon van Handel (Annals of Applied Probability, 2015), 'Phase transitions in nonlinear filtering' with Ramon van Handel (Electronic Journal of Probability, 2015), 'Graph-dependent implicit regularisation for distributed stochastic subgradient descent' with Dominic Richards (Journal of Machine Learning Research, 2020), 'Nearly minimax-optimal rates for noisy sparse phase retrieval via early-stopped mirror descent' with Fan Wu (Information and Inference: A Journal of the IMA, 2023), and 'Optimal convergence rate for exact policy mirror descent in discounted Markov decision processes' with Emmeran Johnson and Ciara Pike-Burke (NeurIPS, 2023). Major awards include the ERC Consolidator Grant (€1,999,929, 2023–2028), Amazon Research Award ($120,000, 2025), UKRI AI Research Resource (2026), Google TensorFlow Award ($10,000, 2020), Turing Fellowship (2017–2021), and MPLS Teaching Award (2019). He holds leadership roles as Director of the Oxford ELLIS Unit and Deputy Head of Department, and serves as Senior Area Chair for ICML (2025–2026).