Neural Network Studies
Abstract
Research at Logicon RDA in neural networks under the ARPA contract includes three main areas: theoretical research in the fundamentals of neural networks, applications of neural networks to practical problems of interest, and the development of new and more powerful neural networks and techniques for solving these problems. The papers which follow are compendium of our research in these areas. The theoretical studies include an overview of the basic useful theorems and general rules which apply to neural networks (in 'Overview of Neural Network Theory'), studies of training time as the network is scaled to larger dimensions (in 'Scaling of Back-Propagation Training Time to Large Dimensions'), an analysis of the classification and function fitting capability of neural networks (in 'Classification and Function Fitting as a Function of Scale and Complexity'), a Comparison of Classifiers: The Neural Network, Bayes- Gaussian, and k-Nearest Neighbor Classifiers'), an analysis of fuzzy logic and its relationship to neural network (in 'Fuzzy Logic and the Relation to Neural Networks'), an analysis of the Reduced Coulomb Energy (RCE) network (in 'The Reduced Coulomb Energy Classifier'), radial basis functions (in 'Radial Basis Approximations'), and the Lynch-Granger models (in 'Lynch-Granger Model of Olfactory Cortex)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 1993
- Accession Number
- ADA271593
Entities
People
- Gregg Wilensky
- Joseph Neuhaus
- Narbik Manukian
- Natalie Rivetti