Sequential Monte Carlo for Graphical Models
2014 (English)In: Advances in Neural Information Processing Systems, 2014, 1862-1870 p.Conference paper (Refereed)
We propose a new framework for how to use sequential Monte Carlo (SMC) algorithms for inference in probabilistic graphical models (PGM). Via a sequential decomposition of the PGM we find a sequence of auxiliary distributions defined on a monotonically increasing sequence of probability spaces. By targeting these auxiliary distributions using SMC we are able to approximate the full joint distribution defined by the PGM. One of the key merits of the SMC sampler is that it provides an unbiased estimate of the partition function of the model. We also show how it can be used within a particle Markov chain Monte Carlo framework in order to construct high-dimensional block-sampling algorithms for general PGMs.
Place, publisher, year, edition, pages
2014. 1862-1870 p.
Computer Science Probability Theory and Statistics Control Engineering
IdentifiersURN: urn:nbn:se:liu:diva-112967OAI: oai:DiVA.org:liu-112967DiVA: diva2:775992
Neural Information Processing Systems (NIPS)