Back propagation separates where perceptrons do

Neural Networks

Volumn 4, Issue 2, 1991, Pages 243-249

  Sontag, Eduardo D ^a   Sussmann, Héctor J ^a  

^a NONE

Author keywords

Back propagation; Nonlinear least squares; Pattern classification

Indexed keywords

COMPUTER PROGRAMMING - ALGORITHMS;

BACKPROPAGATION ALGORITHM; FEEDFORWARD NEURAL NETWORKS; PATTERN CLASSIFICATION; SIGMOIDAL ACTIVATION SYSTEMS;

NEURAL NETWORKS;

EID: 0025839203     PISSN: 08936080     EISSN: None     Source Type: Journal    
DOI: 10.1016/0893-6080(91)90008-S     Document Type: Article

References (12)

1. Approximation of Boolean functions by sigmoidal networks Part I XOR and other two-variable functions
2. Back propagation fails to separate where perceptrons succeed
3. Connectionist learning procedures
4. Convergence issues in perceptron based adaptive neural network models
5. Some remarks on the back propagation algorithm for neural net learning
6. Back propagation can give rise to spurious local minima even for networks without hidden layers
7. Strategies for teaching layered networks classification tasks