Variational encoder-decoders (VEDs) have shown promising results in dialogue generation. However, the latent variable distributions are usually approximated by a much simpler model than the powerful RNN structure used for encoding and decoding, yielding the KL-vanishing problem and inconsistent training objective. In this paper, we separate the training step into two phases: The first phase lea...

Fuzzy class label adjustment is a classification preprocessing strategy that compensates for the possible imprecision of class labels. Using training vectors, robust measures of location and dispersion are computed for each class center. Based on distances from these centers, fuzzy sets are constructed that determine the degree to which each input vector belongs to each class. These membership ...

In this contribution we describe the modelling of the differential pressure behavior of isolation materials at a sieve by artificial neural networks (ANN). The subject arranges itself in the area of the reactor safety research. Compared with [3] the number of the inputs for the connection which can be modelled was increased. Thereby the number of necessary connections is reduced and the model q...

In automatic speech understanding, the division of continuously running speech into syntactic chunks is a great problem. Syntactic boundaries are often marked by prosodic means. For the training of statistic models for prosodic boundaries large data-bases are necessary. For the GermanVerbmobil project (automatic speech{to{speech translation), we developed a syntactic-prosodic labeling scheme wh...

Automatic language identification (LID) systems generally exploit acoustic knowledge, possibly enriched by explicit language specific phonotactic or lexical constraints. This paper investigates a new LID approach based on hierarchical multilayer perceptron (MLP) classifiers, where the first layer is a “universal phoneme set MLP classifier”. The resulting (multilingual) phoneme posterior sequenc...

This paper gives the definition of Transparent Neural Network “TNN” for the simulation of the globallocal vision and its application to the segmentation of administrative document image. We have developed and have adapted a recognition method which models the contextual effects reported from studies in experimental psychology. Then, we evaluated and tested the TNN and the multi-layer perceptron...

A concept of fuzzy discretization of feature space for a rough set theoretic classifier is explained. Fuzzy discretization is characterised by membership value, group number and affinity corresponding to an attribute value, unlike crisp discretization which is characterised only by the group number. The merit of this approach over both crisp discretization in terms of classification accuracy, i...

Pointing refers to orienting a hand, arm, head or body towards an object and is possible without calculating the object’s depth and 3D position. We show that pointing can be learned as holistic direct mapping from an object’s pixel coordinates in the visual field to joint angles, which define pose and orientation of a human or robot. To this aim, we record real world and noisy training images t...

In this paper, both the architecture and learning procedure underlying the parallel layer perceptron is presented. This topology, di1erent to the previous ones, uses parallel layers of perceptrons to map nonlinear input–output relationships. Comparisons between the parallel layer perceptron, multi-layer perceptron and ANFIS are included and show the e1ectiveness of the proposed topology. c © 20...

While depth tends to improve network performances, it also makes gradient-based training more difficult since deeper networks tend to be more non-linear. The recently proposed knowledge distillation approach is aimed at obtaining small and fast-to-execute models, and it has shown that a student network could imitate the soft output of a larger teacher network or ensemble of networks. In this pa...