Learning non-linearly separable Boolean functions with linear threshold unit trees and Madaline-style networks

Authors:
Mehran Sahami
Affiliations:
Department of Computer Science, Stanford University, Stanford, CA
Venue:
AAAI'93 Proceedings of the eleventh national conference on Artificial intelligence
Year:
1993

Citing 8
Cited 2

Neural Nets for Adaptive Filtering and Adaptive Pattern Recognition

Computer
Parallel distributed processing: explorations in the microstructure of cognition, vol. 1: foundations

Parallel distributed processing: explorations in the microstructure of cognition, vol. 1: foundations
Learning internal representations by error propagation

Parallel distributed processing: explorations in the microstructure of cognition, vol. 1
Redundant noisy attributes, attribute errors, and linear-threshold learning using winnow

COLT '91 Proceedings of the fourth annual workshop on Computational learning theory
Artificial intelligence (3rd ed.)

Artificial intelligence (3rd ed.)
Induction of Decision Trees

Machine Learning
Learning Quickly When Irrelevant Attributes Abound: A New Linear-Threshold Algorithm

Machine Learning
Multivariate Versus Univariate Decision Trees

Multivariate Versus Univariate Decision Trees

Automatic Construction of Decision Trees from Data: A Multi-Disciplinary Survey

Data Mining and Knowledge Discovery
Generating neural networks through the induction of threshold logic unit trees

INBS '95 Proceedings of the First International Symposium on Intelligence in Neural and Biological Systems (INBS'95)

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper investigates an algorithm for the construction of decisions trees comprised of linear threshold umts and also presents a novel algorithm for the learning of nonlinearly separable boolean functions using Madaline style networks which are isomorphic .to decision trees. The construction of such networks is discussed, and theIr performance in learning is compared with standard Back-Propagation on a sample problem in which many irrelevant attributes are introduced. Littlestone's Winnow algorithm is also explored within this architecture as a means of learning in the presence of many Irrelevant attributes. The learning ability of this Madaline-style architecture on non-optimal (larger than necessary) networks is also explored.