Section: Software and Platforms

BiiPS software

BiiPS is a general software, developped by Adrien Todeschini, for Bayesian inference with interacting particle systems, a.k.a. sequential Monte Carlo (SMC) methods. It aims at popularizing the use of these methods to non-statistician researchers and students, thanks to its automated “black box" inference engine.

It borrows from the BUGS/JAGS software, widely used in Bayesian statistics, the statistical modeling with graphical models and the language associated with their descriptions.

Unlike MCMC methods used by BUGS/JAGS, SMC methods are more adapted to dynamic problems (tracking, signal filtering, etc).

A beta version of the software can be downloaded from the website of the BiiPS project . This software has been presented at the international workshop BayesComp in Kyoto, the international conference ISBA in Tokyo, the conference on Premières Rencontres R in Bordeaux, and the international workshop on efficient simulation in finance in Paris.

Adrien Todeschini participated to a session focused on Recent Developments in Software for MCMC (and SMC) (MCMSki IV, Fifth IMS-ISBA joint meeting MCMSki IV, Chamonix Mont-Blanc, France).

This invited panel features four leading researchers working on software development for Bayesian computation. Each panelist will highlight their particular software, including its history, development, and relative strengths and weaknesses. Looking forward, panelists will discuss and debate the future of Bayesian computation and software development, including challenges, opportunities and bottlenecks. Emphasis throughout will be on simplifying and automating the implementation of Monte Carlo methods, with an eye towards scalability to larger and more complex models and data.