PhD and Postdoc Positions in Theoretical Machine Learning/Statistics at CWI Amsterdam

One postdoc and two PhD positions are offered within the Machine Learning group at the CWI in Amsterdam, the Netherlands, in the recently awarded NWO TOP-project "Safe Bayesian Inference".
Application Deadline for each position: January 20, 2017
Starting date: before October, 2017
The topic is making Bayesian and related inference methods more robust in a variety of settings, most notably:
- under misspecification, i.e. when the model is 'wrong but useful'
- in hypothesis testing, when optional stopping and other violations of the sampling plan are allowed
Robustified Bayesian methods often work by equipping the likelihood with a 'temperature' or 'learning rate' parameter, which also plays a crucial role in sequential prediction methods that continue to work well in worst-case environments. Part of the project is about determining optimal theoretical values for the learning rates that lead to fast learning; another part of the project is about implementing computationally fast algorithms for inferring this learning rate from the data in sequential prediction contexts; yet another part of the project is about developing general (not necessarily Bayesian) theory for inference under misspecification - yet another part of the project is about designing and implementing robust hypothesis tests and investigating whether this can mitigate the so-called 'replicability crisis'. Despite the implementation aspect, the research has a theoretical focus and candidates are expected to have demonstrable ability for mathematical analysis.
Background papers/technical reports:
? P.D. Grünwald and T. van Ommen. Inconsistency of Bayesian Inference for Misspecified Linear Models, and a Proposal for Repairing It. Preprint, December 2014, available on arXiv as http://arxiv.org/abs/1412.3730.
? P.D. Grünwald. Toetsen als gokken: een redelijk alternatief voor de p-waarde. Nieuw Archief voor de Wiskunde 5/17, Nr. 4, December 2016 (in Dutch)
? P.D. Grünwald. Safe Probability. To appear in Journal of Statistical Planning and Inference. Preprint, December 2016 available on arXiv as http://arxiv.org/abs/1604.01785.
Postdoc Details
The successful postdoc applicant will have a strong mathematical background with a Ph.D. in mathematics, computer science or statistics.
To apply for the postdoc position, please contact Peter Grunwald (pdg dot cwi dot nl), attaching a CV, a cover letter specifying why you are interested in the topic, your two most important papers and the name of a scientist whom we can contact for a letter of recommendation. Directly supplying such a recommendation letter is possible as well.
We especially invite qualified women to apply.
*About CWI*
CWI is the national research institute for mathematics and computer science in the Netherlands, located in Amsterdam. It conducts pioneering research in these fields and transfers its results to society. With 55 permanent research staff, 40 postdocs and 70 PhD students, CWI is a compact institute that lies at the heart of European research in mathematics and computer science. It was the birthplace of the European internet and was home to the invention of the popular programming language Python. CWI is located at Science Park Amsterdam that is presently developing into a major location of research in the natural sciences in The Netherlands, housing the sciences of the University of Amsterdam and of the Vrije Universiteit as well as several other national research institutes next to CWI. It is within easy biking distance from the centre of one of Europe's most beautiful, lively and international cities.