Aprendizado não-supervisionado em redes neurais pulsadas de base radial: um estudo da capacidade de agrupamento para a classificação de pixels

Redes neurais pulsadas - redes que utilizam uma codificação temporal da informação - têm despontado como uma promissora abordagem dentro do paradigma conexionista, emergente da ciência cognitiva. Um desses novos modelos é a rede neural pulsada com função de base radial, que é capaz de armazenar informação nos tempos de atraso axonais dos neurônios. Um algoritmo de aprendizado foi aplicado com sucesso nesta rede pulsada, que se mostrou capaz de mapear uma seqüência de pulsos de entrada em uma seqüência de pulsos de saída. Mais recentemente, um método baseado no uso de campos receptivos gaussianos foi proposto para codificar dados constantes em uma seqüência de pulsos temporais. Este método tornou possível a essa rede lidar com dados computacionais. O processo de aprendizado desta nova rede não se encontra plenamente compreendido e investigações mais profundas são necessárias para situar este modelo dentro do contexto do aprendizado de máquinas e também para estabelecer as habilidades e limitações desta rede. Este trabalho apresenta uma investigação desse novo classificador e um estudo de sua capacidade de agrupar dados em três dimensões, particularmente procurando estabelecer seus domínios de aplicação e horizontes no campo da visão computacional.

Radiographic study of distal radial physeal closure in thoroughbred horses

Monthly radiography was performed to study distal radial physeal closure in ten male and ten female Throughbred horses. The height, thoracic circumference and metacarpus circumference were also measured, Distal radial physeal closure time was sooner in females than males, and took 701 +/- 37 and 748 +/- 55 days respectively.

THEORY OF ELASTICITY OF THE ABRIKOSOV FLUX-LINE LATTICE FOR UNIAXIAL SUPERCONDUCTORS - PARALLEL FLUX LINES

In this paper, we consider the extension of the Brandt theory of elasticity of the Abrikosov flux-line lattice for a uniaxial superconductor for the case of parallel flux lines. The results show that the effect of the anisotropy is to rescale the components of the wave vector k and the magnetic field and order-parameter wave vector cut off by a geometrical parameter previously introduced by Kogan.

Effect of the heat flux direction on electrical properties of SrBi2Nb2O9 thin films crystallized using a microwave oven

Ferroelectric SrBi2Nb2O9 (SBN) thin films were prepared by the polymeric precursors method and deposited by spin coating onto Pt/Ti/SiO2/Si substrate and crystallized using a domestic microwave oven. It was studied the influence of the heat flux direction and the duration of the thermal treatment on the films crystallization. An element with high dielectric loss, a SiC susceptor, was used to absorb the microwave energy and transfers the heat to the film. Influence of the susceptor position to the sample crystallization was verified, the susceptor was, placed or below the substrate or above the film. The SBN perovskite phase was observed after a thermal treatment at 700 degreesC for 10 min when the susceptor was placed below the substrate and for 30 min when the susceptor was placed above the film. Electrical measurements revealed that the film crystallized at 700 degreesC for 10 min, with the susceptor placed below the film, presented dielectric constant, dielectric loss, remanent polarization and coercive field of, 67, 0.011, 4.2 muC/cm(2) and 27.5 kV/cm, respectively. When the films were crystallized at 700 degreesC for 30 min, with the susceptor placed above the film, the dielectric constant was 115 and the dissipation factor was around of 0.033, remanent polarization and coercive field were 10.8 muC/cm(2) and 170 kV/cm, respectively. (C) 2003 Elsevier B.V. All rights reserved.

The implications of laboratory Rn-222 flux measurements to the radioactivity in groundwaters: the case of a karstic limestone aquifer

Laboratory time-scale experiments were conducted on Carboniferous Limestone gravels from the Mendip Hills area, England, with the purpose of evaluating the release of Rn-222 to the water phase. The specific surface areas of the samples were 4.14 and 1.69 cm g(-1), which provided, respectively, values of 50.6 and 12.7 pCi for the released Rn. These results allowed the calculation of the emanation coefficient of this rock matrix with respect to the release of Rn, where completely different values corresponding to 23% and 6% were found, suggesting that the extent to which grain boundaries or imperfections in aggregates of micro-crystals of calcite intersect the particle surface certainly affects the Rn release. They also permitted the evaluation of models for the generation of Rn in rocks and transfer to water, in order to interpret the radioactivity due to this gas in groundwaters from the karstic aquifer of the Mendip Hills area, where the calculated activities in groundwater based on the values of 23% and 6% for the emanation coefficient were about 51 and 15 times higher than actually measured in groundwater. Therefore, the emanation coefficient in nature is considerably smaller than in the lab experiment, and another factor k (0 < k < 1) may be introduced into the equations related to the modelling, with the aim of adjusting the theoretical-practical results. (C) 1997 Elsevier B.V. Ltd. All rights reserved.

Maximum number of flux lines inside columnar defects

We set new limits for the maximum number of fluxons trapped inside columnar defects, for both cases of metallic and insulated cores. We show that the saturation limit predicted by the Mkrtchyan and Shmidt theory is just a first approximation to our results, because it does not consider the presence of the remaining flux lines exterior to the defects.

Optimal capacitor placement in radial distribution networks

The capacitor placement (replacement) problem for radial distribution networks determines capacitor types, sizes, locations and control schemes. Optimal capacitor placement is a hard combinatorial problem that can be formulated as a mixed integer nonlinear program. Since this is a NP complete problem (Non Polynomial time) the solution approach uses a combinatorial search algorithm. The paper proposes a hybrid method drawn upon the Tabu Search approach, extended with features taken from other combinatorial approaches such as genetic algorithms and simulated annealing, and from practical heuristic approaches. The proposed method has been tested in a range of networks available in the literature with superior results regarding both quality and cost of solutions.

Anthropogenic influences on annual flux of cations and anions at meio stream basin, São Paulo State, Brazil

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Optimization of Radial Basis Function neural network employed for prediction of surface roughness in hard turning process using Taguchi's orthogonal arrays

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Performance of low/correct pressure radial ply tires

Extensive field testes were conducted using the UCD single wheel tester employing three large radial ply tractor tires in two different soils, four different soil conditions, two axle load levels, and three levels of tire inflation pressures in order to quantify the benefits of using low/correct inflation pressures. During these tests slip, net traction, gross traction, and dynamic axle load were recorded. Furthermore, soil moisture content, cone index, and dry bulk density data were obtained at test locations. The results of the analysis showed a significant increase in net traction and traction efficiency when low/correct inflation was used. Benefits of using low/correct pressure was higher in tilled soil conditions.

Conformai field theory of AdS background with Ramond-Ramond flux

We review a formalism of superstring quantization with manifest six-dimensional spacetime supersymmetry, and apply it to AdS3 × S3 backgrounds with Ramond-Ramond flux. The resulting description is a conformal field theory based on a sigma model whose target space is a certain supergroup SU′(2|2).

Transient flow of the oil-refrigerant mixture through the radial clearance in rolling piston compressors

Unsteady flow of oil and refrigerant gas through radial clearance in rolling piston compressors has been modeled as a heterogeneous mixture, where the properties are determined from the species conservation transport equation coupled with momentum and energy equations. Time variations of pressure, tangential velocity of the rolling piston and radial clearance due to pump setting have been included in the mixture flow model. Those variables have been obtained by modeling the compression process, rolling piston dynamics and by using geometric characteristics of the pump, respectively. An important conclusion concerning this work is the large variation of refrigerant concentration in the oil-filled radial clearance during the compression cycle. That is particularly true for large values of mass flow rates, and for those cases the flow mixture cannot be considered as having uniform concentration. In presence of low mass flow rates homogeneous flow prevail and the mixture tend to have a uniform concentration. In general, it was observed that for calculating the refrigerant mass flow rate using the difference in refrigerant concentration between compression and suction chambers, a time average value for the gas concentration should be used at the clearance inlet.

An analog implementation of radial basis neural networks (RBNN) using BiCMOS technology

This paper describes a analog implementation of radial basis neural networks (RBNN) in BiCMOS technology. The RBNN uses a gaussian function obtained through the characteristic of the bipolar differential pair. The gaussian parameters (gain, center and width) is changed with programmable current source. Results obtained with PSPICE software is showed.