Global Optimisation of Neural Network Models via Sequential Sampling

João F. G. de Freitas, Mahesan Niranjan, Arnaud Doucet, Andrew H. Gee

Abstract: We propose a novel strategy for training neural networks using se(cid:173) quential sampling-importance resampling algorithms. This global optimisation strategy allows us to learn the probability distribu(cid:173) tion of the network weights in a sequential framework. It is well suited to applications involving on-line, nonlinear, non-Gaussian or non-stationary signal processing.