Validating and Expanding Approximate Bayesian Computation Methods (17w5025)


(Paris Dauphine University)

(Simon Fraser University)

(University of Helskinki)

(Monash University)

(University of Newcastle)

(University of Sheffield)


The Banff International Research Station will host the "Validating and Expanding Approximate Bayesian Computation Methods" workshop from February 19th to February 24th, 2017.

The increasing computational power at our disposal allows us to simulate many aspects of life -- ranging from the common ancestry of far-away communities of invasive insects to handling data bases such as those of a search-engine daily search patterns or of a major on-line retailer customer browsing habits. It is nowadays possible to quantitatively predict the behavior of systems without performing experiments, or to efficiently complement experiences. For example, new drugs are now pre-selected based on molecular models, while the advertising focus behind social networks like Facebook rely on simulated designs of experiment. ABC (Approximate Bayesian computation) methods are generic algorithms that can handle such complex structures with a reduced amount of calibration and monitoring, provided the connected models can be simulated on computers.

However, despite advances made in applied mathematics and in computational statistics towards processing such systems, the ever-growing flow of data and the increase in modeling needs implies to develop even more efficient computing tools. It also calls for mathematical insight about the validation of generic simulation tools, beyond mere computer simulations that are necessarily restricted in terms of the system sizes and integration times considered therein. The aim of the workshop is to bring together a mixed audience of statisticians who are using and developing computational methods and of applied probabilists and computer scientists studying approximation theory and probabilistic programming. The workshop will give these researchers the opportunity to present their most recent work on the topic of ABC convergence and to exchange ideas and methods in order to plan the steps ahead and to facilitate the scaling challenges that face the method. Since more participants are also involved in at least one area of applications for ABC methods, we expect further diffusions and cross-fertilization with the communities relying on ABC.

The Banff International Research Station for Mathematical Innovation and Discovery (BIRS) is a collaborative Canada-US-Mexico venture that provides
an environment for creative interaction as well as the exchange of ideas, knowledge, and methods within the Mathematical Sciences, with related disc
iplines and with industry. The research station is located at The Banff Centre in Alberta and is supported by Canada's Natural Science and Engineeri
ng Research Council (NSERC), the U.S. National Science Foundation (NSF), Alberta's Advanced Education and Technology, and Mexico's Consejo Nacional
de Ciencia y Tecnología (CONACYT).