Abstract

The focus of this session is on advanced Bayesian computations for complicated dependent systems. Bayesian inference provides an efficient framework for quantifying uncertainty. However, quantification requires the computation of integrals with respect to the posterior distribution. Markov chain Monte Carlo methods have been the gold standard for this purpose for the past thirty years. However, the classical methodologies are no longer suitable for modern high-dimensional and complicated dependent systems. The researchers in this session address this problem by proposing new efficient methodologies.

Prof. Jasra is well known in the field of sequential Monte Carlo methods and has recently developed a combination of filtering methods and multilevel computing. Dr. Kamatani is working on scalable Bayesian computational methods such as piecewise deterministic Markov processes. Dr. Duncan is a young talented researcher working on machine learning and scalable computation. Dr. Choi is also a young talented researcher actively working on non-reversible Markov chains and annealing methods for Bayesian analysis.

Two speakers are from East Asia (Singapore and Japan), one from the Middle East and one from Western Europe.