Tuesday 4th November, 14:30-15:30
Room MR3 in the Department of Pure Mathematics and Mathematical Statistics, University of Cambridge (Please note this is not the usual location for the BSU Seminars)
For directions please click here. Map here.
Dr Alex Beskos
University College London
“Sequential Monte Carlo methods: some Developments and Applications”
Abstract: Sequential Monte Carlo (SMC) methods are nowadays routinely applied in a variety of complex applications: hidden Markov models, dynamical systems, target tracking, control problems, just to name a few. SMC algorithms have been greatly refined in the last decades and their theoretical properties are now much better understood. The talk will review recent advancements in SMC algorithms, and present some attempts (where the speaker is involved) for applying such methods in complex models in Data Assimilation and Biostatistics, illustrating the relevant computational challenges. Important properties of SMC methods (e.g. parallelisation, adaptation, tempering) will be highlighted and discussed.