© Andrew Lipson. Escher's Relativity in Lego

All models are wrong...

Model uncertainty & selection in complex models

On Tuesday morning March 16 1971 Hirotake Akaike sat in the metro in a Tokyo suburb with an impending deadline for a paper. On that journey the general idea behind the AIC suddenly dawned on him. Exactly fourty years later, the model selection landscape has become much more complex, involving themes such as sparsity, high-dimensionality, explanatory vs predictive ability, etc.

This workshop brings together statisticians, philosophers and quantitative scientists to discuss the complexity of model uncertainty, cutting-edge strategies to deal with this uncertainty and common (or not so common) pitfalls in model selection. We look at modern, complex problems where we need to go beyond Akaike to find a satisfactory solution.

The workshop starts on Monday 14 March with a short course by Gerda Claeskens (KU Leuven) and has other great, confirmed speakers such as Kenneth P. Burnham (Colorado State), John B. Copas (Warwick), Nial Friel (University College Dublin), Peter J. Green (Bristol), Peter Grunwald (CWI, Leiden), Herbert Hoijtink (Utrecht), Angelika van der Linde (Bremen), Arthur Petersen (PBL, Bilthoven), Erik-Jan Wagenmakers (Amsterdam).

NEW (30 March 2011) Some of the presentations are now available for download.