Error correcting memorization learning for noisy training examples

Neural Networks

Volumn 14, Issue 1, 2001, Pages 79-92

  Nakashima, A ^a   Hirabayashi, A ^b   Ogawa, H ^b  

^a TOSHIBA CORPORATION   (Japan)

^b TOKYO INSTITUTE OF TECHNOLOGY   (Japan)

Author keywords

Function approximation; Mahalanobis norm; Noise suppression; Overfitting; Reproducing kernel Hilbert space; Weighted least squares estimation

Indexed keywords

APPROXIMATION THEORY; ERROR CORRECTION; LEARNING SYSTEMS; LEAST SQUARES APPROXIMATIONS; REDUNDANCY; SPURIOUS SIGNAL NOISE;

HILBERT SPACES; MAHALANOBIS NORMS; OVERFITTING; WEIGHTED LEAST SQUARES ESTIMATION;

NEURAL NETWORKS;

ARTICLE; ARTIFICIAL NEURAL NETWORK; COVARIANCE; DEFENSE MECHANISM; ERROR; LEARNING; MATHEMATICS; NOISE; OPERATOR; PRIORITY JOURNAL; REGRESSION ANALYSIS; SAMPLING; TRAINING;

MODELS, THEORETICAL; NEURAL NETWORKS (COMPUTER); NOISE;

EID: 0035191103     PISSN: 08936080     EISSN: None     Source Type: Journal    
DOI: 10.1016/S0893-6080(00)00075-7     Document Type: Article

References (30)

1. A new look at statistical model identification
2. Asymptotic statistical theory of overtraining and cross-validation
3. The effects of adding noise during backpropagation training on generalization performance
4. Training with noise is equivalent to Tikhonov regularization
5. An equivalence between sparse approximation and support vector machines
6. The family of parametric projection filters and its properties for perturbation
7. Noise injection into inputs in back propagation learning
8. Network information criterion - determining the number of hidden units for an artificial neural network model
9. Projection filter regularization of ill-conditioned problem
10. Neural network learning, generalization and over-learning
11. Projection filter, Wiener filter and Karhunen-Loéve subspaces in digital image restoration
12. Networks for approximation and learning
13. A sparse representation for function approximation
14. Computational vision and regularization theory
15. A theory of over-learning in the presence of noise
16. Properties of averaged projection filter for image restoration