Arthur Gretton is a distinguished computational statistician and leading authority in machine learning theory and applications. He currently serves as Professor with the Gatsby Computational Neuroscience Unit and Director of the Centre for Computational Statistics and Machine Learning at University College London, while also holding a position as Research Scientist at Google DeepMind. Gretton received his degrees in Physics and Systems Engineering from the Australian National University and received his PhD from the University of Cambridge, conducted in collaboration with Microsoft Research and the Signal Processing and Communications Laboratory. His distinguished career includes significant appointments at the Max Planck Institute for Biological Cybernetics and the Machine Learning Department at Carnegie Mellon University, establishing him as a prominent figure bridging theoretical foundations with practical applications in computational statistics.
Professor Gretton's groundbreaking research has fundamentally transformed the field of nonparametric statistical hypothesis testing through his pioneering development of kernel methods, particularly the widely adopted Maximum Mean Discrepancy framework. His influential work on designing and training both implicit generative models including Wasserstein gradient flows and GANs, and explicit models such as energy-based approaches, has provided essential tools for modern machine learning practitioners. The kernel-based methods he developed for causal inference and representation learning have become standard techniques in machine learning research, enabling researchers to address complex problems of hidden confounding and statistical dependence. His contributions have significantly advanced the theoretical understanding of machine learning models while providing practical solutions to challenging problems in data analysis.
Beyond his research achievements, Professor Gretton has been instrumental in shaping the machine learning community through his extensive service as program chair for AISTATS, tutorials chair for ICML, and senior area chair for multiple NeurIPS and ICML conferences. He serves as an active Action Editor for the Journal of Machine Learning Research and contributes significantly to the Royal Statistical Society Research Section Committee, demonstrating his commitment to advancing statistical methodology. As an ELLIS Fellow, Gretton continues to lead cutting-edge research at the intersection of computational statistics and machine learning, with his current work focusing on spectral methods for causal estimation with hidden confounders and novel approaches to proxy causal learning. His ongoing contributions position him at the forefront of developing theoretically rigorous yet practically applicable methods for addressing complex challenges in modern data science.
