In this paper we investigate the relationships between a multipreferential semantics for defeasible reasoning in knowledge representation and deep neural network model. Weighted bases description logics are considered under “concept-wise” multipreference semantics. The is further extended to fuzzy interpretations exploited provide preferential interpretation of Multilayer Perceptrons, some cond...

Nonlinear transformation is one of the major obstacles to analyzing the properties of multilayer perceptrons. In this letter, we prove that the correlation coefficient between two jointly Gaussian random variables decreases when each of them is transformed under continuous nonlinear transformations, which can be approximated by piecewise linear functions. When the inputs or the weights of a mul...

Application of the temporal processing neural networks (TPNNs) to the speech recognition is justified by the nature of the task. Indeed ASR is a sequence recognition problem and assumes incorporation of time into decision process. Static models treat elements of sequence as independent patterns, which is clearly unrealistic. On the other hand temporal processing nets, built on the basis of mult...

Generally, we allocate one output node per class in pattern recognition applications of MLPs(multilayer perceptrons). In this paper, we propose a method to improve generalization capability of MLPs through increasing the number of output nodes per class. We verify that the proposed method decreases misclassification ratios of MLPs through a short mathematical aspect. And then, simulations of is...

Classification of imbalanced data is pervasive but it is a difficult problem to solve. In order to improve the classification of imbalanced data, this letter proposes a new error function for the error backpropagation algorithm of multilayer perceptrons. The error function intensifies weight-updating for the minority class and weakens weight-updating for the majority class. We verify the effect...

In this work we use connectionist techniques to solve some problems of pattern classification involving time series. We build our classifiers using two different artificial neural networks paradigms: the self organizing maps and the multilayer perceptrons. We compare the results achieved using the single classifier and an ensemble of classifiers. The conclusions are quite hopeful, achieving res...