In this paper, we use the neural property known as intrinsic plasticity to develop neural network models that resemble the koniocortex, the fourth layer of sensory cortices. These models evolved from a very basic two-layered neural network to a complex associative koniocortex network. In the initial network, intrinsic and synaptic plasticity govern the shifting of the activation function, and the modification of synaptic weights, respectively. In this first version, competition is forced, so that the most activated neuron is arbitrarily set to one and the others to zero, while in the second, competition occurs naturally due to inhibition between second layer neurons. In the third version of the network, whose architecture is similar to the koniocortex, competition also occurs naturally owing to the interplay between inhibitory interneurons and synaptic and intrinsic plasticity. A more complex associative neural network was developed based on this basic koniocortex-like neural network, capable of dealing with incomplete patterns and ideally suited to operating similarly to a learning vector quantization network. We also discuss the biological plausibility of the networks and their role in a more complex thalamocortical model.

Several research studies point to the fact that sensory and cognitive reductions like cataracts, deafness, macular degeneration, or even lack of activity after job retirement, precede the onset of Alzheimer’s disease. To simulate Alzheimer’s disease earlier stages, which manifest in sensory cortices, we used a computational model of the koniocortex that is the first cortical stage processing sensory information. The architecture and physiology of the modeled koniocortex resemble those of its cerebral counterpart being capable of continuous learning. This model allows one to analyze the initial phases of Alzheimer’s disease by “aging” the artificial koniocortex through synaptic pruning, by the modification of acetylcholine and GABA-A signaling, and by reducing sensory stimuli, among other processes. The computational model shows that during aging, a GABA-A deficit followed by a reduction in sensory stimuli leads to a dysregulation of neural excitability, which in the biological brain is associated with hypermetabolism, one of the earliest symptoms of Alzheimer’s disease.

No abstract

Resumo-No planejamento e operação de redes de distribuição de energia elétrica, integradas com fontes renováveis como a solar, a estimativa de valores de irradiância torna-se um grande diferencial estratégico. Para auxiliar na resolução deste problema, pesquisadores da área propõem inúmeras soluções como, por exemplo, aplicação de séries temporais regressivas, utilização de dados previstos por agências meteorológicas e emprego de técnicas de inteligência artificial, entre outros. Considerando a abordagem conexionista, as redes neurais artificiais (RNA) vêm sendo cada vez mais utilizadas em problemas de classificação e regressão não-linear devido à sua aplicabilidade na análise de dados e previsão, incluindo em aplicações na área de energias renováveis. A disponibilidade de dados históricos em bancos de dados meteorológicos torna a RNA ainda mais atrativa nos problemas de estimação de irradiância solar, uma vez que esta é capaz de executar um mapeamento não-linear entre conjuntos de variáveis de entrada e de saída. Assim, este trabalho objetiva desenvolver duas diferentes arquiteturas de RNA, a Multilayer Perceptron (MLP), com algoritmo de treinamento Backpropagation, e a Generalized Regression Neural Network (GRNN), para analisar a contribuição de cada uma para a determinação de uma curva típica de irradiância solar para o mês de janeiro de 2019 na cidade de Curitiba-PR. Foram implementadas duas arquiteturas distintas para cada tipo de RNA, variando o número de neurônios na camada oculta da MLP e o valor de espalhamento (spread) para a GRNN, visando avaliar o impacto dos parâmetros estruturais no desempenho de cada algoritmo. Embora ambas as arquiteturas demonstrarem, assim como menciona a literatura, resultados

A computational model for self-recovery of electricity distribution network was developed to simulate it, emulated by the IEEE 123 node model. The electrical system considered has automatic switches capable of identifying a momentary failure in the line and finding the best reconfiguration for its reclosing. An artificial neural network (ANN), backpropagation, was used to classify the type of failure and determine the best reconfiguration of the distribution network. Initially, five power failure scenarios were simulated in certain different parts of the power grid, and power flow analysis via OpenDSS was performed. Next, the most suitable switching was observed within the shortest time interval to restore the power supply. With the purpose of better visualization to identify the reclosing, an implementation was carried out via ELIPSE SCADA. In this way, it is possible to identify the faulted segment in order to isolate it, leaving the smallest number of consumers without power supply in shortest possible time. With the results of the simulations, tests and analyzes were performed to verify their robustness and speed, in the expectation that the model developed be faster than an experienced Operating Distribution Center.

No abstract

No abstract