bardenet at stats.ox.ac.uk
+44 (0) 1865 272860 (Department)
+44 (0) 1865 285363 (Direct)
+44 (0 )1865 272595 (Fax)
I graduated in 2009 from ENS Cachan (France) with an MSc in mathematics for vision and learning. I obtained my PhD in 2012 at University Paris-Sud XI, working under the supervision of Balazs Kégl. Since January 2013, I'm a postdoctoral researcher in Chris Holmes' group.
I am interested in numerical Bayesian methods. Particular topics include large-scale approximate inference, adaptive Markov Chain Monte Carlo (MCMC), and Bayesian optimization.
About my research
During my PhD, I was first interested in solving methodological problems motivated by MCMC inference in the Pierre Auger experiment, a large-scale cosmic ray observatory located in the Argentinian pampa. Second, I worked on automatic hyperparameter tuning methods, with the idea in mind to deliver turn-key machine learning software. Besides continuing research on these topics, I am currently interested in approximate Bayesian decision-making for large datasets.
R. Bardenet, O. Cappé, G. Fort, and B. Kégl. Adaptive MCMC with online relabeling. Submitted, preprint available as arXiv:1210.2601
R. Bardenet, O. Cappé, G. Fort, and B. Kégl. An adaptive Metropolis algorithm with online relabeling. In Proceedings of the 15th International Conference on Artificial Intelligence and Statistics (AISTATS), 2012
J. Bergstra, R. Bardenet, B. Kégl, and Y. Bengio. Algorithms for hyperparameter optimization. In Advances in Neural Information Processing Systems (NIPS) 24: Proceedings of the 2011 Conference, volume 24. The MIT Press, 2011
B. Kégl, R. Busa-Fekete, K. Louedec, R. Bardenet, X. Garrido, I.C. Mariș, D. Monnier-Ragaigne, S. Dagoret-Campagne, and M. Urban. Reconstructing Nμ19(1000). Technical Report 2011-054, Auger Project Technical Note, 2011
R. Bardenet, B. Kégl, and D. Veberic. Single muon response: The signal model. Technical Report 2010-110, Auger Project Technical Note, 2010