E. Mayoraz and F. Aviolet. Constructive training methods for feedforward neural networks with binary weights. International Journal of Neural Networks, 7:149--166, 1996. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Constructive Learning Techniques for Designing Neural Network.. - Campbell (1997) (8 citations) (Correct)
.... substantial (for example a generalisation performance of 91:4 Sigma 3:4 on a test set was achieved for a Boolean problem for which Back Propagation gave 67:3 Sigma 5:7 ) Other constructive approaches to generating feed forward networks with binary valued weights have been considered recently [92]. 2.6 Neural networks with a single hidden layer. Instead of generating tree or cascade architectures we can also use the dichotomy procedure to generate a shallower network with a single hidden layer and weights with a value of 1 between the hidden nodes and output. As for the Target Switch ....
E. Mayoraz and F. Aviolet. Constructive training methods for feedforward neural networks with binary weights. International Journal of Neural Networks, 7:149--166, 1996.
Two Constructive Methods for Designing Compact Feedforward.. - Amaldi, Guenin (1997) (2 citations) (Correct)
....note that although the focus in this paper is on constructive algorithms using simple perceptron like training procedures for single units, it is possible to allocate additional units to break down unfaithful classes in a more natural way at the expense of using more complex local procedures. See [35] for an to feedforward networks with binary weights. 5 Experimental results Three test problems have been selected in order to compare the performances of the Shift and PTI algorithms to three alternative methods, namely the Upstart [20] the Tiling [36] and the Offset [24, 34] which derive from ....
....In spite of the poor generalization performances of PTI networks, algorithms that generate an arbitrary number of hidden layers are worth pursuing because some functions may have a much more compact representation using a larger number of hidden layers, see Section 4. It is worth noting that in [35] the basic idea of the PTI is extended to the deal with feedforward networks with binary weights and the K similarity problem is considered as a benchmark. Finally, the two algorithms we have presented can be extended to construct networks with multiple outputs. Although the current versions can ....
E. Mayoraz and F. Aviolat. Constructive training methods for feedforward neural networks with binary units. International Journal of Neural Systems, 7:149--166, 1996.
Maximizing the Robustness of a Linear Threshold Classifier.. - Mayoraz, Robert (1994) Self-citation (Mayoraz) (Correct)
....develop algorithms able to bring the stability as far as possible in order to corroborate the theoretical previsions. In the present work, we tend to produce algorithms that run fast enough to be used at each iteration of a general training algorithm for multilayer feedforward neural networks (see [26]) Thus, in our experiment with binary weights, for each problem TS executed around 1500 steps and for problems of size n = 255, the computational time was below 3 minutes per problem on a Silicon Graphics MIPS R4400 at 50Mhz. In order to improve the lower bound of the optimum stability, we also ....
E. Mayoraz and F. Aviolat, Constructive training methods for feedforward neural networks with binary weights, orwp, Swiss Federal Institute of Technology, Department of Mathematics, December 1993.
Decision Fusion Using a Multi-Linear Classifier - Verlinde, Maître, Mayoraz (1998) Self-citation (Mayoraz) (Correct)
....result of this data reduction, only the data points situated along the separation surface of the two classes are maintained. The aim of the training phase is to determine a set of jSj half spaces maximizing the global discrimination between positive examples and negative examples. As discussed in [11], when more than one separator is involved, it is more natural to consider the discrimination between the set X Theta Y of pairs of positive negative points instead of reasoning on the discrimination between the two sets of positive and negative points X and Y . Thus, our goal is to find ....
E. Mayoraz and F. Aviolat. Constructive training methods for feedforward neural networks with binary weights. International Journal of Neural Systems, 7(2):149--166, May 1996.
Two Constructive Methods for Designing Compact Feedforward.. - Amaldi, al. (1998) (2 citations) (Correct)
No context found.
E. Mayoraz and F. Aviolat 1996, "Constructive training methods for feedforward neural networks with binary units," Int. J. Neural Systems 7, 149--166.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC