Made available in DSpace on 2014-06-12T15:59:13Z (GMT). No. of bitstreams: 2 arquivo4980_1.pdf: 1807272 bytes, checksum: abdafe66edec7c505073236b251526d0 (MD5) license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 1999 / Motivados pela demanda do mercado de software musical, bem como pelo interesse científico envolvido no problema de previsão de séries temporais [Weigend, 1993], desenvolvemos um ambiente capaz de realizar previsões de acordes de canções de Jazz em tempo real. Nós propusemos uma arquitetura híbrida original que tem como base uma rede neural MLP-backpropagation atuando de forma concorrente com um rastreador de seqüências repetidas de acordes. A rede neural faz um aprendizado prévio a partir de diversos exemplos de canções, extraindo os padrões curtos de seqüências de acordes típicas. O sistema rastreador funciona capturando em tempo real as repetições (refrões, estrofes, etc.) dentro de uma dada canção, as quais escapariam à rede neural. Trata-se da problemática geral de aprendizado a priori versus aprendizado situado, em tempo real. Com a arquitetura híbrida proposta e uma representação rica do acorde musical, obtivemos resultados muito acima dos registrados na literatura dedicada ao problema

In this project a method to analyse images is presented. The images document the development of grassroots in a tilled field in order to study the movement of nitrate in the field. The final aim of the image analysis is to estimate the volume of dead and living roots in the soil. Since the roots and the soil have a broad and overlapping range of colours the fundamental problem is to find the roots in the images. Earlier methods for analysis of root images have used methods based on thresholds to extract the roots. To use a threshold the pixels of the object must have a unique range of colours separating them from the colour of the background, this is not the case for the images in this project. Instead the method uses a neural network to classify the individual pixels. In this paper a complete method to analyse images is presented and although the results are far from perfect, the method gives interesting results

Detta arbete går ut på att testa hur två olika träningsmetoder påverkar hur ett artificiellt neuralt nätverk (ANN) presterar i en 2d spelmiljö. Ett belöningsbaserat nätverk som använder genetiska algoritmer har jämförts mot ett exempelbaserat nätverk som använder backpropagation. För att göra detta möjligt att testa så behövde fyra delsteg genomföras. Dessa är utveckling av belöningsbaserad ANN, utveckling av exempelbaserad ANN, utveckling av testmiljö och evaluering av resultat. Resultaten visar att agenten belöningsbaserat nätverk har presterat bättre i det flesta testen men även att den varit mer slumpmässig. Det finns dock undantag där den agenten med exempelbaserat nätverk har varit bättre. Slutsatsen är att efter detta experiment rekommenderas en agent med belöningsbaserat nätverk över en med exempelbaserat men att detta inte är någon garanti för att få optimala resultat. Ett framtida arbete som hade varit intressant är att fokusera på endast en algoritm och se hur träning och skillnader på olika nätverksarkitekturer hade påverkat den.

Abstract Bleaching is an essential process in chemical pulp production for better pulp brightness and longer life expectancy. However, it causes costs such as chemicals, energy, equipment, and loss of yield. Non-linear reactions and several process variables, with interactions, make large plants complicated to model and optimize. As an expensive process bleaching has been a natural target of optimization, but there is still the need to either improve these methods or consider the optimization problem from a new point of view. The aim of this thesis was to develop production optimization methods for pulp bleaching, so that they are practical, usable on-line, easy to tune, and transferable. According to our assumption, neural networks could provide a practical optimization method by combining analytical knowledge with real data. In this kind of problem, the load sharing concept, recognizing interactions in chemical usage and the serial multi-stage nature of the process can simplify the task. The related work in bleaching optimization was studied as well as multi-stage serial process solving in principle, related optimization methods and especially neural networks in optimization. The data were collected during normal mill operation and modeled using neural networks. Optimization was performed based on visualizing the neural network models. The results showed that backpropagation neural networks are capable of modeling parts of the bleach plant and also the entire bleaching operation to such an extent that they are useful in the optimization. The modeling and the tuning can be performed without a profound knowledge of the system, but the process is slower and less reliable. Moving a trained neural network to another mill is inadvisable. It is more reasonable just to transfer the knowledge of variables and network structure. The important factor in on-line production optimization is the stabilization of the disturbances and a well-controlled operation towards a more economical state. Generally, more than half of the total chemicals should be used in the first bleaching stage D0 and the remaining load should be divided so that the dosage at the D1 is about 30% higher than in the D2 stage.

Artificial neural networks have shown great potential and have attracted much research interest. One problem faced when simulating such networks is speed. As the number of neurons increases, the time to simulate and train a network increases dramatically. This makes it difficult to simulate and train a large-scale network system without the support of a high-performance computer system. The solution we present is a "real" parallel system - using a parallel machine to simulate neural networks which are intrinsically parallel applications. SpiNNaker is a scalable massively-parallel computing system under development with the aim of building a general-purpose platform for the parallel simulation of large-scale neural systems. This research investigates how to model large-scale neural networks efficiently on such a parallel machine. While providing increased overall computational power, a parallel architecture introduces a new problem - the increased communication reduces the speedup gains. Modeling schemes, which take into account communication, processing, and storage requirements, are investigated to solve this problem. Since modeling schemes are application-dependent, two different types of neural network are examined - spiking neural networks with spike-time dependent plasticity, and the parallel distributed processing model with the backpropagation learning rule. Different modeling schemes are developed and evaluated for the two types of neural network. The research shows the feasibility of the approach as well as the performance of SpiNNaker as a general-purpose platform for the simulation of neural networks. The linear scalability shown in this architecture provides a path to the further development of parallel solutions for the simulation of extremely large-scale neural networks.

Made available in DSpace on 2016-01-26T18:56:03Z (GMT). No. of bitstreams: 1 Ronaldo Oikawa.pdf: 3711757 bytes, checksum: 5d13d9ef8b2d092d0bb1a6fe62a73248 (MD5) Previous issue date: 2015-03-16 / The Artificial Neural Networks are nonlinear mathematical models that resemble the human brain, and this ability to learn was applied to recognize the rain patterns in the city of Presidente Prudente, located in the region of Pontal do Paranapanema. Through these calculations, it was possible to indicate another way to rain forecast. This study used two algorithms with supervised learning, the first one the Multiple Layer Network Propagation, with 23 neurons and with one, two and three hidden layers, and the second one the Support Vector Machine (SVM) with polynomial, radial basis function and hyper tangent cores. The set analyzed covers the period from January 1996 to May 2012, collected from Weather Forecast Center and Climate Studies (CPTEC). The results showed that the atmospheric pressure, wind direction, minimum temperature and air relative humidity were the parameters more related with the rain precipitation. The SVM model with base radial function core, using sigma=0.1, showed the best results with Kappa coefficient, equal to 0.675 for first test group, equal to 0.746 to the second test group 0.746 and equal to 0.826 for the third test group. These results demonstrate the data set robustness and allowed achieve high accuracy rate in recognition of rain precipitation. / As Redes Neurais Artificiais são modelos matemáticos não lineares que se assemelham ao cérebro humano, e esta capacidade de aprender foi aplicada no reconhecimento de padrões da chuva na cidade de Presidente Prudente, localizada na região do Pontal do Paranapanema. Através desses cálculos foi possível indicar uma forma alternativa de se reconhecer o padrão da precipitação da chuva. O presente trabalho utilizou dois algoritmos com aprendizagem supervisionada, sendo o primeiro a Rede de Múltipla Camada de Retro Propagação, com 23 neurônios e com uma, duas e três camadas ocultas, já o segundo, a Máquina de Vetor de Suporte (SVM) utilizou o núcleo polinomial, função de base radial e hiper tangente. O conjunto de dados analisados compreende o período de Janeiro de 1996 até Maio de 2012, sendo obtidos do Centro de Previsão de Tempo e Estudos Climáticos (CPTEC). Os resultados demonstraram que a pressão atmosférica, direção do vento, temperatura mínima e umidade relativa do ar são os parâmetros que estão mais relacionados à precipitação da chuva. O modelo SVM, com núcleo função de base radial, utilizando o parâmetro sigma=0,1 obteve os melhores resultados, apresentando o coeficiente Kappa (resposta), igual a 0,675 para o grupo de teste um, 0,746 para o grupo de teste dois e 0,826 para o grupo de teste três. Estes resultados demonstram a robustez do conjunto de dados e permitiram atingir altos índices de acerto no reconhecimento da precipitação da chuva.

Made available in DSpace on 2016-07-18T17:46:20Z (GMT). No. of bitstreams: 1 Ronaldo Oikawa.pdf: 3711757 bytes, checksum: 5d13d9ef8b2d092d0bb1a6fe62a73248 (MD5) Previous issue date: 2015-03-16 / The Artificial Neural Networks are nonlinear mathematical models that resemble the human brain, and this ability to learn was applied to recognize the rain patterns in the city of Presidente Prudente, located in the region of Pontal do Paranapanema. Through these calculations, it was possible to indicate another way to rain forecast. This study used two algorithms with supervised learning, the first one the Multiple Layer Network Propagation, with 23 neurons and with one, two and three hidden layers, and the second one the Support Vector Machine (SVM) with polynomial, radial basis function and hyper tangent cores. The set analyzed covers the period from January 1996 to May 2012, collected from Weather Forecast Center and Climate Studies (CPTEC). The results showed that the atmospheric pressure, wind direction, minimum temperature and air relative humidity were the parameters more related with the rain precipitation. The SVM model with base radial function core, using sigma=0.1, showed the best results with Kappa coefficient, equal to 0.675 for first test group, equal to 0.746 to the second test group 0.746 and equal to 0.826 for the third test group. These results demonstrate the data set robustness and allowed achieve high accuracy rate in recognition of rain precipitation. / As Redes Neurais Artificiais são modelos matemáticos não lineares que se assemelham ao cérebro humano, e esta capacidade de aprender foi aplicada no reconhecimento de padrões da chuva na cidade de Presidente Prudente, localizada na região do Pontal do Paranapanema. Através desses cálculos foi possível indicar uma forma alternativa de se reconhecer o padrão da precipitação da chuva. O presente trabalho utilizou dois algoritmos com aprendizagem supervisionada, sendo o primeiro a Rede de Múltipla Camada de Retro Propagação, com 23 neurônios e com uma, duas e três camadas ocultas, já o segundo, a Máquina de Vetor de Suporte (SVM) utilizou o núcleo polinomial, função de base radial e hiper tangente. O conjunto de dados analisados compreende o período de Janeiro de 1996 até Maio de 2012, sendo obtidos do Centro de Previsão de Tempo e Estudos Climáticos (CPTEC). Os resultados demonstraram que a pressão atmosférica, direção do vento, temperatura mínima e umidade relativa do ar são os parâmetros que estão mais relacionados à precipitação da chuva. O modelo SVM, com núcleo função de base radial, utilizando o parâmetro sigma=0,1 obteve os melhores resultados, apresentando o coeficiente Kappa (resposta), igual a 0,675 para o grupo de teste um, 0,746 para o grupo de teste dois e 0,826 para o grupo de teste três. Estes resultados demonstram a robustez do conjunto de dados e permitiram atingir altos índices de acerto no reconhecimento da precipitação da chuva.

Das Training künstlicher neuronaler Netze mittels Gradientenverfahren wird detailliert mit allen Her- und Ableitungen beschrieben, implementiert und demonstriert.

This thesis is about using neural networks in recognition of letters A to Z and numbers 0 to 9. In the first part is theoretically described substance of neural networks and concretically described principle the method of learning multiple-layer network with backward spreaded error(a.ka Backpropagation). Basic problematic of processing the picture and resilence of network against degradation picture by a noise and compression JPEG is also described here. Second part is directed to practical realization of feed foward multiple-layer network with recognition the binary patterns of alphabetical letters and numbers 0 to 9, which was created in Matlab and Simulink environment. Next and final part is about practical realization of feed foward network with recognition the grayscale patterns of alphabetical letters and numbers 0 to 9, which was also created in Matlab and Simulink environment.

This master thesis is focused on designing and implementing a software system, that is able to trade autonomously at stock market. Neural networks are used to predict future price. Genetic algorithm was used to find good combination of input parameters.