Sequential Monte Carlo Methods to Train Neural Network Models

Authors:
J. F. G. De Freitas;M. A. Niranjan;A. H. Gee;A. Doucet
Affiliations:
Cambridge University Engineering Department, Cambridge CB2 1PZ, England, U.K.;Cambridge University Engineering Department, Cambridge CB2 1PZ, England, U.K.;Cambridge University Engineering Department, Cambridge CB2 1PZ, England, U.K.;Cambridge University Engineering Department, Cambridge CB2 1PZ, England, U.K.
Venue:
Neural Computation
Year:
2000

Citing 11
Cited 16

Recursive Bayesian estimation using piece-wise constant approximations

Automatica (Journal of IFAC)
Training multilayer perceptrons with the extended Kalman algorithm

Advances in neural information processing systems 1
Comparative Analysis of Backpropagation and the Extended Kalman Filter for Training Multilayer Perceptrons

IEEE Transactions on Pattern Analysis and Machine Intelligence
Bayesian interpolation

Neural Computation
Original Contribution: Optimal filtering algorithms for fast learning in feedforward neural networks

Neural Networks
Exact adaptive filters for Markov chains observed in Gaussian noise

Automatica (Journal of IFAC)
Issues in Bayesian analysis of neural network models

Neural Computation
Regularisation in sequential learning algorithms

NIPS '97 Proceedings of the 1997 conference on Advances in neural information processing systems 10
Bayesian Learning for Neural Networks

Bayesian Learning for Neural Networks
Neural Networks for Pattern Recognition

Neural Networks for Pattern Recognition
Contour Tracking by Stochastic Propagation of Conditional Density

ECCV '96 Proceedings of the 4th European Conference on Computer Vision-Volume I - Volume I

Robust Full Bayesian Learning for Radial Basis Networks

Neural Computation
Annealing stochastic approximation Monte Carlo algorithm for neural network training

Machine Learning
A graphical model for evolutionary optimization

Evolutionary Computation
Visual Tracking Using Particle Filters with Gaussian Process Regression

PSIVT '09 Proceedings of the 3rd Pacific Rim Symposium on Advances in Image and Video Technology
Gauss-Newton Particle Filter

IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
FastSLAM 2.0: an improved particle filtering algorithm for simultaneous localization and mapping that provably converges

IJCAI'03 Proceedings of the 18th international joint conference on Artificial intelligence
Gaussian sum approach with optimal experiment design for neural network

SIP '07 Proceedings of the Ninth IASTED International Conference on Signal and Image Processing
A one-step unscented particle filter for nonlinear dynamical systems

ICANN'07 Proceedings of the 17th international conference on Artificial neural networks
Nonlinear identification based on diagonal recurrent neural network and particle filter

ICNC'09 Proceedings of the 5th international conference on Natural computation
Online variational inference for state-space models with point-process observations

Neural Computation
Sequential support vector machine control of nonlinear systems via lyapunov function derivative estimation

ICNC'05 Proceedings of the First international conference on Advances in Natural Computation - Volume Part II
A new learning algorithm for diagonal recurrent neural network

ICNC'05 Proceedings of the First international conference on Advances in Natural Computation - Volume Part I
Support vector machine adaptive control of nonlinear systems

ICIC'05 Proceedings of the 2005 international conference on Advances in Intelligent Computing - Volume Part II
Sequential support vector machine control of nonlinear systems by state feedback

ISNN'05 Proceedings of the Second international conference on Advances in Neural Networks - Volume Part III
A smarter particle filter

ACCV'09 Proceedings of the 9th Asian conference on Computer Vision - Volume Part II
Stochastic volatility modeling with computational intelligence particle filters

Proceedings of the 15th annual conference on Genetic and evolutionary computation

Quantified Score

Hi-index	0.00

Visualization

Abstract

We discuss a novel strategy for training neural networks using sequential Monte Carlo algorithms and propose a new hybrid gradient descent / sampling importance resampling algorithm (HySIR). In terms of computational time and accuracy, the hybrid SIR is a clear improvement over conventional sequential Monte Carlo techniques. The new algorithm may be viewed as a global optimization strategy that allows us to learn the probability distributions of the network weights and outputs in a sequential framework. It is well suited to applications involving on-line, nonlinear, and nongaussian signal processing. We show how the new algorithm outperforms extended Kalman filter training on several problems. In particular, we address the problem of pricing option contracts, traded in financial markets. In this context, we are able to estimate the one-step-ahead probability density functions of the options prices.