|Stochastic Systems Group|
Erik B. Sudderth and Alexander T. Ihler
Graphical models provide a powerful general framework for formulating and solving problems of statistical inference and machine learning. In many applications of graphical models, the hidden variables of interest are most naturally specified by continuous, non-Gaussian distributions. However, due to the limitations of existing inference algorithms, it is often necessary to form coarse, discrete approximations to such models.In this talk, we describe a nonparametric belief propagation (NBP) algorithm, which uses stochastic methods to propagate kernel-based approximations to the true continuous messages. Each NBP message update requires approximating the product of several Gaussian mixtures; we present efficient procedures for sampling from this product using multiscale representations. We apply NBP to a pair of problems motivated by distributed sensor networks. The first problem, self-calibration, entails estimation of the unknown locations of sensors using local measurements of inter-sensor distances. The second is preliminary work on an efficient method for tracking multiple, indistinguishable targets.
Problems with this site should be emailed to firstname.lastname@example.org