Representações reduzidas por redes neurais com aprendizado local não-supervisionado sobre campos receptivos restritos

Algoritmos ótimos na extração de componentes principais com aprendizado não-supervisionado em redes neurais de múltiplos neurônios de saída são não-locais, ou seja, as modificações em uma dada sinapse entre dois neurônios dependem também da atividade de outros neurônios. Esta rede ótima extrairá as principais componentes dos dados e submetidos à sua primeira camada. As principais componentes são as projeções destes vetores nos autovalores máximos da matriz de correlação Gij = (eiej), onde a média (-) é sobre a distribuição de e. Existem fortes evidências indicando que sinapses biológicas só se modificam via regras locais, como por exemplo a regra de Hebb. Mas se aplicarmos regras locais numa rede com múltiplas saídas, todos os neurônios da saída serão equivalentes e darão respostas redundantes. A rede será bastante ineficiente. Um modo de contornar este problema é através da restrição dos campos receptivos dos neurônios de saída. Se cada neurônio acessar diferentes partes dos estímulos de entrada, a redundância diminui significativamente. Em contrapartida, ao mesmo tempo que a redundância diminui, também diminui a informação contida em cada neurônio; assim, devemos balancear os dois efeitos otimizando o campo receptivo. O valor ótimo, em geral, depende da natureza dos estímulos, sua estatística, e também do ruído intrínseco à rede. Objetivamos com este trabalho determinar a estrutura ótima de campos receptivos com aprendizado não-supervisionado para uma rede neural de uma camada em diversas condições medindo seu desempenho a partir de técnicas de reconstrução.

Aplicação de redes neurais para previsão de contrato de dólar futuro no mercado brasileiro

Este trabalho tem por objetivo a construção de uma rede neural para previsão do movimento dos contratos de dólar futuro e a construção de estratégias de negociação, para prover uma ferramenta para estimar o movimento do câmbio e para a negociação desses ativos. Essa ferramenta pode auxiliar empresas que necessitam fazer hedge de ativos e passivos e players do mercado que necessitam rentabilizar carteiras. Neste trabalho utilizamos como input dados de ativos do mercado financeiro, de janeiro de 2001 até setembro de 2013, disponíveis via terminal Bloomberg. Para o cálculo dos resultados financeiros das estratégias utilizamos dados de preços referenciais disponibilizados pela BM&F.

Implementação de classificador de tarefas mentais baseado em EEG

BARBOSA, André F. ; SOUZA, Bryan C. ; PEREIRA JUNIOR, Antônio ; MEDEIROS, Adelardo A. D.de, . Implementação de Classificador de Tarefas Mentais Baseado em EEG. In: CONGRESSO BRASILEIRO DE REDES NEURAIS, 9., 2009, Ouro Preto, MG. Anais... Ouro Preto, MG, 2009

Um Modelo híbrido para previsão de curvas de produção de petróleo

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior

Um sistema para a inferência de variáveis usando redes neurais artificiais

This work aims to obtain a low-cost virtual sensor to estimate the quality of LPG. For the acquisition of data from a distillation tower, software HYSYS ® was used to simulate chemical processes. These data will be used for training and validation of an Artificial Neural Network (ANN). This network will aim to estimate from available simulated variables such as temperature, pressure and discharge flow of a distillation tower, the mole fraction of pentane present in LPG. Thus, allowing a better control of product quality

Sistema inteligente para estimar a porosidade em sedimentos a partir da análise de sinais GPR

This Thesis presents the elaboration of a methodological propose for the development of an intelligent system, able to automatically achieve the effective porosity, in sedimentary layers, from a data bank built with information from the Ground Penetrating Radar GPR. The intelligent system was built to model the relation between the porosity (response variable) and the electromagnetic attribute from the GPR (explicative variables). Using it, the porosity was estimated using the artificial neural network (Multilayer Perceptron MLP) and the multiple linear regression. The data from the response variable and from the explicative variables were achieved in laboratory and in GPR surveys outlined in controlled sites, on site and in laboratory. The proposed intelligent system has the capacity of estimating the porosity from any available data bank, which has the same variables used in this Thesis. The architecture of the neural network used can be modified according to the existing necessity, adapting to the available data bank. The use of the multiple linear regression model allowed the identification and quantification of the influence (level of effect) from each explicative variable in the estimation of the porosity. The proposed methodology can revolutionize the use of the GPR, not only for the imaging of the sedimentary geometry and faces, but mainly for the automatically achievement of the porosity one of the most important parameters for the characterization of reservoir rocks (from petroleum or water)

Emprego de redes neurais artificiais supervisionadas e não supervisionadas no estudo de parâmetros reológicos de excipientes farmacêuticos sólidos

In this paper artificial neural network (ANN) based on supervised and unsupervised algorithms were investigated for use in the study of rheological parameters of solid pharmaceutical excipients, in order to develop computational tools for manufacturing solid dosage forms. Among four supervised neural networks investigated, the best learning performance was achieved by a feedfoward multilayer perceptron whose architectures was composed by eight neurons in the input layer, sixteen neurons in the hidden layer and one neuron in the output layer. Learning and predictive performance relative to repose angle was poor while to Carr index and Hausner ratio (CI and HR, respectively) showed very good fitting capacity and learning, therefore HR and CI were considered suitable descriptors for the next stage of development of supervised ANNs. Clustering capacity was evaluated for five unsupervised strategies. Network based on purely unsupervised competitive strategies, classic "Winner-Take-All", "Frequency-Sensitive Competitive Learning" and "Rival-Penalize Competitive Learning" (WTA, FSCL and RPCL, respectively) were able to perform clustering from database, however this classification was very poor, showing severe classification errors by grouping data with conflicting properties into the same cluster or even the same neuron. On the other hand it could not be established what was the criteria adopted by the neural network for those clustering. Self-Organizing Maps (SOM) and Neural Gas (NG) networks showed better clustering capacity. Both have recognized the two major groupings of data corresponding to lactose (LAC) and cellulose (CEL). However, SOM showed some errors in classify data from minority excipients, magnesium stearate (EMG) , talc (TLC) and attapulgite (ATP). NG network in turn performed a very consistent classification of data and solve the misclassification of SOM, being the most appropriate network for classifying data of the study. The use of NG network in pharmaceutical technology was still unpublished. NG therefore has great potential for use in the development of software for use in automated classification systems of pharmaceutical powders and as a new tool for mining and clustering data in drug development

Técnicas de inteligência artificial para a geração dinâmica de set points para uma coluna de destilação

Artificial Intelligence techniques are applied to improve performance of a simulated oil distillation system. The chosen system was a debutanizer column. At this process, the feed, which comes to the column, is segmented by heating. The lightest components become steams, by forming the LPG (Liquefied Petroleum Gas). The others components, C5+, continue liquid. In the composition of the LPG, ideally, we have only propane and butanes, but, in practice, there are contaminants, for example, pentanes. The objective of this work is to control pentane amount in LPG, by means of intelligent set points (SP s) determination for PID controllers that are present in original instrumentation (regulatory control) of the column. A fuzzy system will be responsible for adjusting the SP's, driven by the comparison between the molar fraction of the pentane present in the output of the plant (LPG) and the desired amount. However, the molar fraction of pentane is difficult to measure on-line, due to constraints such as: long intervals of measurement, high reliability and low cost. Therefore, an inference system was used, based on a multilayer neural network, to infer the pentane molar fraction through secondary variables of the column. Finally, the results shown that the proposed control system were able to control the value of pentane molar fraction under different operational situations