Download:
|
by G. C. Vasconcelos, M. C. Fairhurst, D. L. Bisset
Neural Computing & Applic., Spring-Verlag
http://cimbres.di.ufpe.br/~compint/./papers/nca.ps.gz
Add To MetaCart
Abstract:
The problem of the rejection of patterns not belonging to identified training classes is investigated with respect to multilayer perceptron networks (MLPs). The reason for the inherent unreliability of the standard MLP in this respect is explained and some mechanisms for the enhancement of its rejection performance are considered. Two network configurations are presented as candidates for a more reliable structure and are compared to the so-called "negative training " approach. The first configuration is an MLP which uses a Gaussian as its activation function and the second is an MLP with direct connections from the input to the output layer of the network. The networks are examined and evaluated both through the technique of network inversion and through practical experiments in a pattern classification application. Finally, the model of radial basis function networks (RBFs) is also considered in this respect and its performance is compared to that attained with the other networks described.
Citations
|
2138
|
Learning Internal Representations by Error Propagation
– Rumelhart, Hinton, et al.
- 1986
|
|
396
|
Fast learning in networks of locally-tuned processing units
– Darken
- 1989
|
|
308
|
Multivariable functional interpolation and adaptive networks
– Broomhead, Lowe
- 1988
|
|
70
|
Review of Neural Networks for Speech Recognition
– Lippmann
- 1989
|
|
60
|
Novelty Detection and Neural Network Validation
– Bishop
- 1994
|
|
45
|
Pattern classification using neural networks
– Lippmann
- 1989
|
|
27
|
Recognizing handprinted letters and digits using backpropagation learning
– Martin, Pittman
- 1991
|
|
27
|
Inversion of multilayer nets
– Linden, Kindermann
- 1989
|
|
22
|
Handwritten digit recognition using k nearest-neighbor, radial-basis function, and backpropagation neural networks
– Lee
- 1991
|
|
15
|
Modifying the generalized delta rule to train networks of non-monotonic processors for pattern classification
– Dawson, Schopflocher
- 1992
|
|
12
|
Reflective modular neural network systems
– Smieja, Muhlenbein
- 1992
|
|
10
|
Improving Rejection Performance on Handwritten Digits by Training with Rubbish
– Bromley, Denker
- 1993
|
|
4
|
Enhanced reliability of multilayer perceptron networks through controlled pattern rejection
– Vasconcelos, Fairhurst, et al.
- 1993
|
|
4
|
On the recognition capabilities of feedforward nets
– Sontag
- 1990
|
