This page contains my publications and related documents. Bibliographic information is available in this BibTeX file.
Note that, like many Dutch names, my family name 'Van Erven' consists of multiple words. In the Netherlands, the prefix 'van' is capitalised, except when directly preceded by a given name (e.g. Tim) or initials.
Preprints and Pending Submissions
-
Mixability is Bayes Risk Curvature Relative to Log Loss.
T. van Erven, M. D. Reid and R. C. Williamson. Submitted to the Journal of Machine Learning Research, 2011. This is an extended version of our COLT paper, with an improved presentation and several new results. -
Catching up faster by switching sooner: A predictive approach to adaptive estimation with an application to the AIC-BIC dilemma.
T. van Erven, P. Grünwald and S. de Rooij. Accepted for publication by the Journal of the Royal Statistical Society, Series B, 2011. Read at the ordinary meeting on October 19, 2011. An earlier version of this paper was runner-up in the student paper competition of the Risk Analysis Section of the ASA. [Matlab code]
Refereed Publications
-
Adaptive Hedge.
T. van Erven, P. D. Grünwald, W. Koolen and S. de Rooij. Advances in Neural Information Processing Systems 24 (NIPS 2011), 2012. -
Mixability is Bayes Risk Curvature Relative to Log Loss.
T. van Erven, M. D. Reid and R. C. Williamson. Proceedings of the 24th Annual Conference on Learning Theory (COLT), 2011. -
Rényi Divergence and Majorization.
T. van Erven and P. Harremoës. IEEE International Symposium on Information Theory (ISIT), 2010. -
Learning the Switching Rate by Discretising Bernoulli Sources Online.
S. de Rooij and T. van Erven. Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics (AISTATS), 2009. [details] -
Catching Up Faster in Bayesian Model Selection and Model Averaging.
T. van Erven, P. D. Grünwald and S. de Rooij. Advances in Neural Information Processing Systems 20 (NIPS 2007), 2008. [details, NIPS poster]
PhD Thesis
-
When Data Compression and Statistics Disagree: Two Frequentist Challenges for the Minimum Description Length Principle.
T. van Erven. PhD thesis, Leiden University, 2010. Promotor: Peter Grünwald. [details]
Other
Unrefereed publications, publications at local conferences, and unpublished work
-
Freezing and Sleeping: Tracking Experts that Learn by Evolving Past Posteriors.
W. M. Koolen and T. van Erven. Unpublished, 2010. -
Switching between Hidden Markov Models using Fixed Share.
W. M. Koolen and T. van Erven. Unpublished, 2010. -
Catching Up Faster by Switching Sooner: A Prequential Solution to the AIC-BIC Dilemma.
T. van Erven, P. D. Grünwald and S. de Rooij. Preprint posted on the math arXiv, arXiv:0807.1005 [math.ST], July 2008. Runner-up in the student paper competition of the Risk Analysis Section of the ASA. [details] -
Switching between Predictors with an Application in Density Estimation.
T. van Erven, S. de Rooij and P. Grünwald. Proceedings of the 28th Symposium on Information Theory in the Benelux, Enschede, The Netherlands, 2007. -
The Momentum Problem in MDL and Bayesian Prediction.
T. van Erven. Master's thesis, University of Amsterdam, The Netherlands, May 2006. Supervisors: Peter Grünwald, Steven de Rooij. [details, tex]