Here are some of my most recent (2017 and 2018) publications and preprints. See here for an exhaustive publication list.
- P.D. Grünwald and T. van Ommen. Inconsistency of Bayesian Inference for Misspecified Linear Models, and a Proposal for Repairing It . Bayesian Analysis, 2017, pp. 1069-1103. Here is a video recording of a talk I gave about the subject at the NIPS 2014 workshop From Bad Models to Good Policies.
- P.D. Grünwald. Safe Probability. Journal of Statistical Planning and Inference 195, 2018, pp. 47-63. Earlier and more complete version available on arXiv as http://arxiv.org/abs/1604.01785.
- S. van der Pas and P.D. Grünwald. Almost the Best of Three Worlds: Risk, Consistency and Optional Stopping for the Switch Criterion in Single Parameter Model Selection. Statistica Sinica 28(I), January 2018, page 229-255. Available on arXiv as arXiv:1408.5724.
- Wouter Koolen, Peter D. Grünwald and T. van Erven. Combining adversarial guarantees and stochastic fast rates in online learning In Advances in Neural Information Processing Systems (NIPS) 29, pages 4457-4465, December 2016. Here is a longer version containing proofs.
- N. Lewis and P.D. Grünwald. Objectively combining AR5 instrumental period and paleoclimate climate sensitivity evidence. Climate Dynamics, 2017. (author’s accepted manuscript; the final publication is available as read-only pdf at Springer via this link). R code that implements the method is avalaible at Nic Lewis’ website.
- P.D. Grünwald and N. Mehta, 2017a. A Tight Excess Risk Bound via a Unified PAC-Bayesian-Rademacher-Shtarkov-MDL Complexity. Available on arXiv as https://arxiv.org/abs/1710.07732.
- P.D. Grünwald and N. Mehta, 2017b. Fast Rates for General Unbounded Loss Functions: from ERM to Generalized Bayes. Available on arXiv as https://arxiv.org/abs/1605.00252, 2017b.
- P.D. Grünwald and R. de Heide, 2018. Why Optional Stopping is a problem for Bayesians, 2017. Available on arXiv as https://arxiv.org/abs/1708.08278.