Constructive algorithms for structure learning in feedforward neural networks for regression problems

By James T. Kwok, Dit-Yan Yeung. 1997

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

AdaNet: Adaptive Structural Learning of Artificial Neural Networks (2016, background): “2. can produce much sparser and/or shallower networks compared to existing methods. a r X i v : 1607 .”
A New Adaptive Merging and Growing Algorithm for Designing Artificial Neural Networks (2009, methodology): “However, the evolutionary approach is quite demanding in both time and user-defined parameters [2].”
A constructive algorithm for training cooperative neural network ensembles (2003, background): “Third, the strong convergence of a constructive algorithm follows directly from its universal approximation ability [28].”
Artificial immune systems as a novel soft computing paradigm (2003, background): “The problem of model selection in ANN can be dealt with several approaches, among which constructive (Kwok and Yeung, 1997; Fritzke, 1994) and pruning strategies (Reed, 1993; de Castro and Von Zuben, 1999) are the most common.”
Trends in extreme learning machines: A review (2015, unknown): “Compared to other incremental learning algorithms proposed in the literature (Chen et al., 1991; Huang, Song, & Wu, 2012; Kwok & Yeung, 1997; Platt, 1991), I-ELM has the notable feature that it can work with a wide type of activation functions as long as they are nonconstant piecewise…”
A new strategy for adaptively constructing multilayer feedforward neural networks (2003, background): “However, the relationship between the generalization performance and the Bayesian evidence has not been established, and those priors sometimes do not produce good results.”
Neural networks in geophysical applications (2000, background): “A recent review of constructive algorithms can be found in Kwok and Yeung (1997).”
Artificial neural networks for RF and microwave design - from theory to practice (2003, background): “The appropriate number of neurons can also be determined through adaptive processes, which add/delete neurons during training [4][22].”
Fundamentals of natural computing: an overview (2007, background): “, networks whose architecture may increase and/or decrease in size [144,174,180]; ensembles of networks [228,289]; the investigation into formal aspects of ANNs, such as convergence and universal approximation capabilities [119,131,281]; hybrids between ANN and other approaches (e.”
Neural networks for the prediction and forecasting of water resource variables: a review of modelling issues and applications (2000, methodology): “However, the number of hidden nodes is not altered, which have to be speciﬁed in advance (Kwok and Yeung, 1997a).”