Sequential recurrent connectionist algorithms for time series modeling of nonlinear dynamical systems

Mirikitani, Derrick Takeshi. 2010. Sequential recurrent connectionist algorithms for time series modeling of nonlinear dynamical systems. Doctoral thesis, Goldsmiths, University of London [Thesis]

[img]
Preview
Text
COMP_thesis_Mirikitani_2010.pdf - Submitted Version

Download (45MB) | Preview

Abstract or Description

This thesis deals with the methodology of building data driven models of nonlinear
systems through the framework of dynamic modeling. More specifically
this thesis focuses on sequential optimization of nonlinear dynamic
models called recurrent neural networks (RNNs). In particular, the thesis
considers fully connected recurrent neural networks with one hidden layer of
neurons for modeling of nonlinear dynamical systems. The general objective
is to improve sequential training of the RNN through sequential second-order
methods and to improve generalization of the RNN by regularization. The
total contributions of the proposed thesis can be summarized as follows:

1. First, a sequential Bayesian training and regularization strategy for recurrent
neural networks based on an extension of the Evidence Framework
is developed.

2. Second, an efficient ensemble method for Sequential Monte Carlo filtering
is proposed. The methodology allows for efficient O(H 2 ) sequential
training of the RNN.

3. Last, the Expectation Maximization (EM) framework is proposed for
training RNNs sequentially.

Item Type:

Thesis (Doctoral)

Departments, Centres and Research Units:

Computing

Date:

9 June 2010

Item ID:

3239

Date Deposited:

01 Jul 2010 14:47

Last Modified:

08 Sep 2022 13:43

URI:

https://research.gold.ac.uk/id/eprint/3239

View statistics for this item...

Edit Record Edit Record (login required)