Dissertations / Theses on the topic 'Artificials neurons'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 50 dissertations / theses for your research on the topic 'Artificials neurons.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
Henniquau, Dimitri. "Conception d’une interface fonctionnelle permettant la communication de neurones artificiels et biologiques pour des applications dans le domaine des neurosciences." Thesis, Université de Lille (2018-2021), 2021. http://www.theses.fr/2021LILUN032.
Neuromorphic engineering is an exciting emerging new field, which combines skills in electronics, mathematics, computer sciences and biomorphic engineering with the aim of developing artificial neuronal networks capable of reproducing the brain’s data processing. Thus, neuromorphic systems not only offer more effective and energy efficient solutions than current data processing technologies, but also set the bases for developing novel original therapeutic strategies in the context of pathological brain dysfunctions. The research group Circuits Systèmes Applications des Micro-ondes (CSAM) of the Institute for Electronics, Microelectronics and Nanotechnologies (IEMN) in Lille, in which this thesis work was carried out, has contributed to the generation of such neuromorphic systems by developing a toolbox constituted of artificial neurons and synapses. In order to implement neuromorphic engineering in the therapeutic arsenal for treating neurologic disorders, we need to interface living and artificial neurons to ensure real communication between these different components. In this context and using the original tools developed by the CSAM group, the main goal of this thesis work was to design and produce a functional interface allowing a bidirectional communication loop to be established between living and artificial neurons. These artificial neurons have been developed by the CSAM group using CMOS technology and are able to emit biomimetic electrical signals. Living neurons were obtained from differentiated PC-12 cells. A first step in this work consisted in modeling and simulating this interface between artificial and living neurons; a second part of the thesis was dedicated to the fabrication and characterization of neurobiohybrid interfaces, and to the growth and characterization of living neurons before studying their capacities to communicate with artificial neurons. First, a model of neuronal membrane representing a living neuron interfaced with a metallic planar electrode has been developed. We thus showed that it is possible to excite neurons using biomimetic signals produced by artificial neurons while maintaining a low excitation voltage. Low voltage excitation would improve energy efficiency of neurobiohybrid systems integrating artificial neurons and reduce the impact of harmful electrical signals on living neurons. Then, the neurobiohybrid interfacing living and artificial neurons has been designed and produced. The results obtained by experimental characterization of this interface validate the approach consisting in exciting living neurons through a metallic planar electrode. Finally, living neurons from PC-12 cells were grown and differentiated directly onto neurobiohybrids. Then, an experimental proof of the ability of biomimetic electrical signals to excite living neurons was obtained using calcium imaging. To conclude, the work presented in this manuscript clearly establishes a proof of concept for the excitation of living neurons using a biomimetic signal in our experimental conditions and thus substantiates the first part of the bidirectional communication loop between artificial neurons and living neurons
Cottens, Pablo Eduardo Pereira de Araujo. "Development of an artificial neural network architecture using programmable logic." Universidade do Vale do Rio dos Sinos, 2016. http://www.repositorio.jesuita.org.br/handle/UNISINOS/5411.
Made available in DSpace on 2016-06-29T14:42:16Z (GMT). No. of bitstreams: 1 Pablo Eduardo Pereira de Araujo Cottens_.pdf: 1315690 bytes, checksum: 78ac4ce471c2b51e826c7523a01711bd (MD5) Previous issue date: 2016-03-07
Nenhuma
Normalmente Redes Neurais Artificiais (RNAs) necessitam estações de trabalho para o seu processamento, por causa da complexidade do sistema. Este tipo de arquitetura de processamento requer que instrumentos de campo estejam localizados na vizinhança da estação de trabalho, caso exista a necessidade de processamento em tempo real, ou que o dispositivo de campo possua como única tarefa a de coleta de dados para processamento futuro. Este projeto visa criar uma arquitetura em lógica programável para um neurônio genérico, no qual as RNAs podem fazer uso da natureza paralela de FPGAs para executar a aplicação de forma rápida. Este trabalho mostra que a utilização de lógica programável para a implementação de RNAs de baixa resolução de bits é viável e as redes neurais, devido à natureza paralelizável, se beneficiam pela implementação em hardware, podendo obter resultados de forma muito rápida.
Currently, modern Artificial Neural Networks (ANN), according to their complexity, require a workstation for processing all their input data. This type of processing architecture requires that the field device is located somewhere in the vicintity of a workstation, in case real-time processing is required, or that the field device at hand will have the sole task of collecting data for future processing, when field data is required. This project creates a generic neuron architecture in programmabl logic, where Artifical Neural Networks can use the parallel nature of FPGAs to execute applications in a fast manner, albeit not using the same resolution for its otputs. This work shows that the utilization of programmable logic for the implementation of low bit resolution ANNs is not only viable, but the neural network, due to its parallel nature, benefits greatly from the hardware implementation, giving fast and accurate results.
Ogden, James M. "Construction of fully equivalent neuronal cables : an analysis of neuron morphology." Thesis, University of Glasgow, 1999. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.301502.
Guahyba, Adriano da Silva. "Utilização de inteligência artificial (redes neurais artificiais) no gerenciamento de reprodutoras pesadas." reponame:Biblioteca Digital de Teses e Dissertações da UFRGS, 2001. http://hdl.handle.net/10183/3322.
Bénédic, Yohann. "Approche analytique pour l'optimisation de réseaux de neurones artificiels." Phd thesis, Université de Haute Alsace - Mulhouse, 2007. http://tel.archives-ouvertes.fr/tel-00605216.
Martinez, Regis. "Dynamique des systèmes cognitifs et des systèmes complexes : étude du rôle des délais de transmission de l’information." Thesis, Lyon 2, 2011. http://www.theses.fr/2011LYO20054/document.
How memory information is represented is still an open question in neurobiology, but also, from the computer science point of view, in machine learning. Some artificial neuron networks models have to face the problem of retrieving information, knowing that, in regard to the model performance, this information is actually stored but in an unknown form or too complex to be easily accessible. This is one of the problems met in large neuron networks and which « reservoir computing » intends to answer.« Reservoir computing » is a category of models that has emerged at the same period as, and has propoerties similar to the model we present here. It is composed of three parts that are (1) an input layer that allows to inject learning examples, (2) a « reservoir » composed of neurons connected with or without a particular predefined, and where there can be adaptation mecanisms, (3) an output layer, called « readout », on which a supervised learning if performed. We bring a particularity that consists in using axonal delays, the propagation time of information from one neuron to another through an axonal connexion. Using delays is a computational improvement in the light of machin learning but also a biological argument for information representation.We show that our model is capable of a improvable but efficient and promising artificial learning. Based on this observation and in the aim of improving performance we seek to understand the internal dynamics of the model. More precisely we study how the topology of the reservoir can influence the dynamics. To do so, we make use of the theory of polychronous groups. We have developped complexe algorithms allowing us to detect those topologicodynamic structures in a network, and in a network activity having a given topology.If we succeed in understanding the links between topology and dynamics, we may take advantage of it to be able to create reservoir with specific properties, suited for learning. Finally, we have conducted an exhaustive study of network expressivness in terms of polychronous groups, based on various types of topologies (random, regular, small-world) and different parameters (number of neurones, conectivity, etc.). We are able to formulate some recommandations to create a network whose topology can be rich in terms of possible representations. We propose to link with the cognitive theory of multiple trace memory that can, in principle, be implemented and studied in the light of polychronous groups
Reali, Egidio Henrique. "Utilização de inteligência artificial - (Redes neurais artificiais) no gerenciamento da produção de frangos de corte." reponame:Biblioteca Digital de Teses e Dissertações da UFRGS, 2004. http://hdl.handle.net/10183/6339.
Monaldi, Jessica. "neuroni artificiali e loro applicazioni." Bachelor's thesis, Alma Mater Studiorum - Università di Bologna, 2020. http://amslaurea.unibo.it/21371/.
Wang, Shengrui Robert François Verjus Jean-Pierre Cosnard Michel Mazaré Guy. "Réseaux multicouches de neurones artificiels." S.l. : Université Grenoble 1, 2008. http://tel.archives-ouvertes.fr/tel-00335818.
Tápia, Milena. "Redes neurais artificiais." Florianópolis, SC, 2000. http://repositorio.ufsc.br/xmlui/handle/123456789/78807.
Made available in DSpace on 2012-10-17T19:37:03Z (GMT). No. of bitstreams: 0Bitstream added on 2014-09-25T17:12:29Z : No. of bitstreams: 1 178322.pdf: 8164173 bytes, checksum: 58dff9972980056ae164ad29c6b70fd0 (MD5)
Pesquisa que aborda o uso de Redes Neurais Artificiais (RNAs) - modelos biologicamente inspirados - no problema de processamento temporal, onde o principal objetivo é a previsão. Com base na Taxinomia de MOZER (1994) para processamento temporal, o foco do estudo recaiu em duas questões: 1) Definir a forma da memória de curto tempo, o conteúdo que deveria ser armazenado nesta, e como seus parametros serião atualizados; 2) e definir a topologia da rede (tamanho, estrutura e conexões), assim como os parâmetros do algoritmo de treinamento (taxa de aprendizado, termo de momento e outros). O modelo resultante foi comparado com a Metodologia de Box & Jenkins para modelos univariados, avaliado e criticado em termos de: capacidade representativa, processo de identificação e capacidade preditiva. Os resultados mostram que uma RNA, quando bem modelada, têm potencial para representar qualquer mapeamento complexo, não-linear, que pode governar mudanças em uma série de tempo. No estudo de caso foi possível prever o preço do ovo para um período de quatorze meses à frente
Dartora, Gery Antonio. "Redes neurais artificiais." Florianópolis, SC, 2003. http://repositorio.ufsc.br/xmlui/handle/123456789/84537.
Made available in DSpace on 2012-10-20T10:14:25Z (GMT). No. of bitstreams: 1 195587.pdf: 955703 bytes, checksum: b8985042df5f6b7ddeaa25bb04b73172 (MD5)
Ariza, Zambrano William Camilo 1989. "Controle ativo de vibrações usando redes neurais artificiais : Active vibration control using artificial neural networks." [s.n.], 2013. http://repositorio.unicamp.br/jspui/handle/REPOSIP/264927.
Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecânica
Made available in DSpace on 2018-08-23T23:04:02Z (GMT). No. of bitstreams: 1 ArizaZambrano_WilliamCamilo_M.pdf: 5789145 bytes, checksum: 151f5e3ef1780d5448a7073b85b4715f (MD5) Previous issue date: 2013
Resumo: Este trabalho tem como objetivo principal o estudo de um método de controle baseado no uso de redes neurais artificiais aplicado ao problema de controle de vibrações em estruturas flexíveis. Este trabalho centra-se no estudo do esquema de controle inverso-direto, que consiste em identificar a dinâmica inversa da planta através de uma rede neural artificial para ser usada como controlador. Três exemplos de aplicação foram resolvidos utilizando-se controladores projetados com o método inverso-direto. A primeira aplicação é o controle de vibrações em uma estrutura mecânica de parâmetros concentrados. O segundo exemplo de aplicação é o controle de vibrações de uma placa engastada em uma de suas extremidades. Neste caso, a placa engastada foi modelada utilizando-se o método de elementos finitos. No seguinte exemplo, o modelo da placa usado no exemplo anterior foi reduzido, deixando apenas os primeiros modos de vibração. No último exemplo tratou-se o problema de controle não colocado das vibrações em uma placa engastada. Os resultados foram analisados a partir da resposta temporal e da resposta em frequência do sistema em malha fechada. Para comparar os resultados obtidos utilizando-se o método de controle baseado em redes neurais artificiais, os exemplos citados anteriormente foram também resolvidos utilizando-se o método de controle ??. Os resultados obtidos demonstram que o método de controle baseado em modelo inverso usando redes neurais foi eficaz na resolução deste tipo de problema
Abstract: The goal of this work is to study a control method based on artificial neural networks applied to the vibration control of flexible structures problem. This work focuses in the direct-inverse control scheme which consists of identifing the inverse dynamics of the plant through an artificial neural network to be used as the controller. Three application examples using the direct-inverse method were solved. The first application is the vibration control in a mechanical structure of concentrated parameters. The second application example is the vibration control of a cantilever plate. The cantilever plate was modeled using the finite elements method. In the third example, a reduction of the cantilever plate model was made. In the last example a non-collocated control problem of vibration in a cantilever plate was treated. The results of the scheme were evaluated according to the temporal response and the frequency response of the closed-loop system. In order to compare the results obtained using the control method based on artificial neural networks, the previous examples were also solved using the ?? control method. The obtained results show that the control method based on inverse model using neural networks was effective in solving this kind of problem
Mestrado
Mecanica dos Sólidos e Projeto Mecanico
Mestre em Engenharia Mecânica
Neto, Camilo Rodrigues. "Propriedades de recuperação de memória em redes neurais atratoras." Universidade de São Paulo, 1997. http://www.teses.usp.br/teses/disponiveis/76/76131/tde-31102008-173551/.
Attractor neural networks are feedback neural networks with no pre-defined connection structure. These types of neural networks present a rich dissipative dynamics and, in general, are used as associative memory devices. Such devices have the capacity to retrieve a previously stored memory, even when exposed to partial or degraded information. To store a memory means to create an attractor for it in the network dynamics, and this is done by specifying the set of synaptic weighs. In this thesis, we concentrate on two classical ways of specifying the synaptics weighs: the pseudo-inverse and the optimal weighs models. For extremely diluted neural networks, for which the connectivity C and the number of neurons N satisfy the condition C « In N, we obtain the phase diagrams in the complete space of the model parameters through the analytical study of the retrieval overlap dynamics. We also investigate the retrieval properties of fully connected neural networks using two approaches: the analytical study of the neighborhood of the stored patterns, and the exhaustive enumeration of the attractors via numerical simulations. Finally, we study analytically the problem of categorization in the pseudo-inverse model. Categorization in attractor neural networks is the capacity to create an attractor for a concept to which the network has had access only through a finite number of examples.
Sales, Daniel Oliva. "NeuroFSM: aprendizado de Autômatos Finitos através do uso de Redes Neurais Artificiais aplicadas à robôs móveis e veículos autônomos." Universidade de São Paulo, 2012. http://www.teses.usp.br/teses/disponiveis/55/55134/tde-06122012-143729/.
Autonomous navigation is a fundamental task in mobile robotics. In order to accurately perform this task it is necessary an intelligent navigation and control system associated to the sensorial system. This project presents the development of a control system for autonomous mobile robots and vehicles navigation. The adopted approach uses Artificial Neural Networks for Finite State Machine learning, allowing the robots to deal with sensorial data even when this data is not precise and correct. Simultaneously, it allows the robots to consider the different situations and states they are inserted in (context detection). This way, it is possible to decide how to proceed with motion control and then execute navigation and control tasks from the most simple ones until the most complex and high level tasks. So, this work uses Artificial Neural Networks to recognize the robots current state (context) at the environment where it is inserted. Once the state is detected, including identification of robots position according to environment elements, the robot will be able to determine the action/- behavior to be executed. The navigation and control system implements a Finite State Machine deciding the current action from current state, being able to identify state changes, alternating between different previously defined behaviors. In order to validade this approach, many experiments were performed with the use of a robotic simulator (Player-Stage), and carrying out tests with real robots (Pioneer P3-AT, SRV-1 and autonomous vehicles)
Miguel, Cesar Gomes. "Evolução estrutural e paramétrica de redes neurais dinâmicas em vida artificial." Universidade de São Paulo, 2009. http://www.teses.usp.br/teses/disponiveis/3/3142/tde-22042009-151501/.
The evolution of artificial neural networks has a wide range of applicability in diverse areas in the field of machine learning, particularly, in artificial life simulations where a population of individuals, controlled by neural networks, adapts in a virtual environment in order to solve a given task. Resembling the natural process in which an organism\'s behavior is subjected to phylogenetic modifications through the complexification of the nervous system, such simulations offer a new synthetic approach in the investigation of intelligence, counter posing traditional symbolic methods. A recent method known as NEAT (NeuroEvolution of Augmenting Topologies), is able to obtain the synaptic weights and the topology with the aid of genetic algorithms. The encoding used by NEAT is flexible enough to allow for open-ended evolution and arbitrary neural architectures. This work presents a NEAT implementation especially suitable to be used with a general purpose simulator known as Breve, constituting a framework for artificial life experiments. The proposed implementation extends NEAT to include dynamical neuron models, where their inner state continuously varies over time. The new model is then compared to the traditional method in a classic unsupervised control benchmark task, showing an efficiency increase while solving the problem. The obtained results motivate the proposed framework for general experiments in artificial life, in which a population of individuals continuously interact with a dynamical environment, adapting through generations.
ALMEIDA, Marcelo Barbosa de. "Um estudo comparativo de técnicas conexionistas na implementação de um sistema de reconhecimento de padrões para um nariz artificial." Universidade Federal de Pernambuco, 2003. https://repositorio.ufpe.br/handle/123456789/2502.
O principal objetivo desta dissertação é fazer um estudo sistemático sobre os diversos tipos de redes neurais artificiais (e seus respectivos algoritmos de aprendizagem) que vêm sendo utilizados na implementação do sistema de reconhecimento de padrões do nariz artificial proposto em [Santos, 2000], apontando suas vantagens e desvantagens. Os modelos analisados são as Multi-layer Perceptrons (MLPs) com o backpropagation, Levenberg-Marquardt e tabu search, e as redes de funções de base radiais (Redes RBF). Também serão investigadas as MLPs com o Resilient backpropagation (Rprop). O algoritmo Rprop foi escolhido por duas razões principais: em geral ele possui um tempo de convergência inferior ao tradicional backpropagation, e até o momento não existe na literatura nenhum trabalho que aplique este algoritmo (junto com as MLPs) como parte do sistema de reconhecimento de padrões do nariz artificial estudado. Para cada modelo de arquitetura (por exemplo, MLP) e algoritmo de treinamento (por exemplo, backpropagation) três topologias diferentes serão investigadas. Para cada uma destas topologias serão feitas trinta inicializações de pesos diferentes (aleatórias), em que cada uma destas inicializações será executada com cada uma das três diferentes partições do conjunto de dados. A partir disto, os resultados obtidos serão analisados através de testes estatísticos (teste de hipóteses). Isto tudo contrasta com os trabalhos anteriores, os quais usavam apenas uma partição dos dados, somente dez execuções para cada topologia, e nenhum teste estatístico era feito
Barnett, William Halbert. "Duty Cycle Maintenance in an Artificial Neuron." Digital Archive @ GSU, 2009. http://digitalarchive.gsu.edu/phy_astr_theses/7.
Murray, Andrew Gerard William, and n/a. "Micro-net the parallel path artificial neuron." Swinburne University of Technology, 2006. http://adt.lib.swin.edu.au./public/adt-VSWT20070423.121528.
Murray, Andrew Gerard William. "Micro-net the parallel path artificial neuron /." Australasian Digital Thesis Program, 2007. http://adt.lib.swin.edu.au/public/adt-VSWT20070423.121528/index.html.
A dissertation presented for the fulfilment of the requirements for the award of Doctor of Philosophy, Faculty of Information and Communication Technology, Swinburne University of Technology, 2007. Typescript. Includes bibliography.
Srinivasan, Vikram. "HDL Descriptions of Artificial Neuron Activation Functions." University of Cincinnati / OhioLINK, 2005. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1121113992.
Capanni, Niccolo Francesco. "The functionality of spatial and time domain artificial neural models." Thesis, Robert Gordon University, 2006. http://hdl.handle.net/10059/241.
Ciliegi, Federico. "Topologie non convenzionali per reti di neuroni artificiali." Bachelor's thesis, Alma Mater Studiorum - Università di Bologna, 2019. http://amslaurea.unibo.it/19497/.
Koiran, Pascal. "Puissance de calcul des réseaux de neurones artificiels." Lyon 1, 1993. http://www.theses.fr/1993LYO19003.
Rocha, Ana Cristina Gonçalves Pinto da. "Utilização de inteligência artificial (redes neurais artificiais) para a classificação de patogenicidade de amostras de Escherichia coli isoladas de frangos de corte." reponame:Biblioteca Digital de Teses e Dissertações da UFRGS, 2006. http://hdl.handle.net/10183/17363.
Escherichia coli (E. coli), which has been forgotten as a potential pathogen for a long time, is now being seriously considered due to the economical losses it generates. In Brazil from 2001 to 2005 the condemnation of carcasses represented a loss of about 58 million dollars to the poultry industry (Brasil/ Ministerio da Agricultura, Pecuária e Abastecimento, 2006). From this total 19 million dollars were due to skin lesions and 39 million dollars to systemic lesions. There have been improvements in the comprehension of the mechanisms of pathogenicity. However, the differentiation of virulent from non-virulent samples is still a problem for veterinarians to come to a diagnosis and, as a consequence, to make decisions. Article 1 is about factors of virulence in E. coli isolates from chicken flocks presenting respiratory problems. The samples from litter were probed for the presence of factors of virulence and antimicrobial resistance. The following properties were analyzed: hemolisin production, motility, operon pap presence, colicin production and serum resistance. The capacity of hemagglutination was verified in 84.1% of the samples. In 76.4% of the samples operon pap presence was detected. Colicin production was observed in 87.3% and in 88.9% serum resistance was verified. In article 2, 238 E. coli samples were probed by Polimerase Chain Reaction (PCR) for the presence of seven virulence genes responsible for adhesion capacity, P fimbriae (papC) and F11 fimbriae (felA), colicin production (cavC), aerobactin presence (iutA), serun resistence (iss) temperature-sensitive hemmaglutinin (tsh) and presence of K1 and K5 capsular antigens (kpsll). The CvaC gene was detected in 31.3% of the cellulitis samples and in 11.5% of feces samples. In the present study 80.6% of the cellulitis samples and 53.8% of feces samples presented the iss gen. The kps was positive in 27% of the cellulitis and in 7% of the feces samples. The papC gene occurred in 46.9% of cellulitis and in 30.8% in feces samples. In cellulitis samples 3.8% were positive for the felA gene whereas in feces samples were 1.3%. The tsh gene was positive in 83% of the isolates from lesions and in 14% of feces samples.For cmvaC, iss, iutA, kpsll ,papc and tsh genes significant statistical differences were detected for isolates from lesions and litter. In article 3 an amount of 61 Escherichia coli isolates from chicken flocks with respiratory problems were probed by the Polimerase chain Reaction PCR) for the presence of the genes which are responsible for the adhesion capacity, P fimbria (papC) and F11 fimbria (felA), colicin production (cvaC), aerobactin presence (iutA), serum resistance (iss), temperature-sensitive hemmaglutin (tsh) and presence of K1 and K5 capsular antigens (kpsII). The iss gene was detected in 73.8%, the tsh in 55.7%, the iutA in 45.9%, the felA in 39.3%, the papC in 24.3%, the cvaC in 23% and kpsII in 18%. Article 4 presents three neural nets of artificial intelligence which were developed through the analysis of papC, felA, cvaC, iutA, iss, tsh and kpsII genes, motility and pathogenicity index (PI) in order to establish predictions and classifications of the pathogenicty of E. coli litter without using animals. Net 1 obtained 54.27% of correctness using 11 categories of IP. In order to improve the performance of the model, a second net was created using 3 categories of IP the correct classification of 80.55%. Trying to get an even better performance, we worked with only two categories, building this way the third net. With this new configuration the correct classification was 83.96%.
Zanchettin, Cleber. "Otimização Global em Redes Neurais Artificiais." Universidade Federal de Pernambuco, 2008. https://repositorio.ufpe.br/handle/123456789/1288.
Conselho Nacional de Desenvolvimento Científico e Tecnológico
Esta tese apresenta um método de otimização global e local, baseado na integração das heurísticas das técnicas Simulated Annealing, Tabu Search, Algoritmos Genéticos e Backpropagation. O desempenho deste método é investigado na otimização simultânea da topologia e dos valores dos pesos das conexões entre as unidades de processamento de redes neurais artificiais Multi-layer Perceptron, a fim de gerar topologias com poucas conexões e alto desempenho para qualquer conjunto de dados. A heurística proposta realiza a busca de forma construtiva e baseada na poda das conexões entre as unidades de processamento da rede. Assim, são geradas redes com arquitetura variável e que podem ser ajustadas para cada problema de forma automática. Experimentos demonstram que o método pode também ser utilizado para a seleção de atributos relevantes. Durante a otimização da arquitetura da rede, unidades de processamento de entrada podem ser eliminadas de acordo com sua relevância para o desempenho do modelo. Desta forma, é obtida uma seleção de atributos inerente ao processo de otimização das redes neurais artificiais. Os principais parâmetros de configuração do método tiveram sua influência estimada através da técnica de planejamento fatorial de experimentos. Com base no planejamento fatorial de experimentos, foi possível verificar a influência, interação e a inter-relação entre os parâmetros de configuração do modelo. Estas análises são importantes para identificar a influência de cada parâmetro e possivelmente diminuir a quantidade de parâmetros ajustáveis no projeto deste método. Nesta tese são realizados experimentos com cinco diferentes bases de dados de classificação e duas bases de dados de previsão. A técnica proposta apresentou resultados estatisticamente relevantes em comparação com outras técnicas de otimização global e local
Zhao, Le. "Adaptive neurocomputation with spiking semiconductor neurons." Thesis, University of Bath, 2015. https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.675688.
Miranda, João Felipe Nunes de. "Modelos de regressão e de redes neurais artificiais na quantificação de carbono e biomassa lenhosa em floresta estacional decidual no Brasil Central." reponame:Repositório Institucional da UnB, 2015. http://dx.doi.org/10.26512/2015.06.D.20198.
Submitted by Fernanda Percia França (fernandafranca@bce.unb.br) on 2015-11-20T13:57:49Z No. of bitstreams: 1 2015_JoãoFelipeNunesDeMiranda.pdf: 1961695 bytes, checksum: 32c59daa230ca8115731bef2bb3357b2 (MD5)
Approved for entry into archive by Raquel Viana(raquelviana@bce.unb.br) on 2016-05-12T21:05:28Z (GMT) No. of bitstreams: 1 2015_JoãoFelipeNunesDeMiranda.pdf: 1961695 bytes, checksum: 32c59daa230ca8115731bef2bb3357b2 (MD5)
Made available in DSpace on 2016-05-12T21:05:28Z (GMT). No. of bitstreams: 1 2015_JoãoFelipeNunesDeMiranda.pdf: 1961695 bytes, checksum: 32c59daa230ca8115731bef2bb3357b2 (MD5)
O presente estudo avaliou e comparou a qualidade de ajuste de modelos alométricos (Spurr, Schumacher & Hall e Husch) e de redes neurais artificiais na estimativa de estoque de biomassa seca e de carbono de árvores com diâmetro a altura do peito (DAP) igual ou superior a 5 cm, em área de floresta estacional decidual, localizada no Município de Campos Belos – GO, no Brasil central. Um total de 74 árvores pertencente a diferentes espécies registradas na área e distribuído em diferentes classes de diâmetro foram cubadas rigorosamente. Tanto para a biomassa seca (total e do fuste) quanto para o carbono (total e do fuste), o modelo de Schumacher & Hall foi o que apresentou melhores medidas de precisão. Foram treinadas 300 RNAs, do tipo MLP (multilayer perceptron) para cada variável dependente e as 10 redes com melhores resultados foram retidas para a análise das medidas de precisão. Todas as melhores redes encontradas apresentaram medidas de precisão sensivelmente melhores do que as alcançadas pelo modelo de Schumacher & Hall. Os estoques de biomassa seca total e do fuste foram iguais a respectivamente 65,61 ±15,52 t.ha-1 e 34,17 ± 7,85 t.ha-1. Para carbono, os estoques total e do fuste foram 29,47 ± 6,93 t.ha-1 e 15,16 ± 3,48 t.ha-1. Guazuma ulmifolia, Callisthene fasciculata, Myracrodruon urundeuva e Dilodendron bipinnatum, além de se destacarem na área por apresentarem maiores valores de IVI (índice de Valor de Importânicia), também se destacaram por apresentarem maiores estoques de biomassa e carbono.
The aim of this study was to evaluate and compare the quality of the adjustments from the use of allometric models (Spurr, Schumacher & Hall and Husch) and artificial neural networks, and generate stock estimates of dry biomass and carbon from the best method. 15 permanent plots located in a fragment of dry seasonal forest, located in the municipality of Campos Belos - GO were inventoried of 737 individuals sampled 74 were strictly cubed. The model that presented best precision measurements, R² adjusted (0.88 to 0.96) and Syx% (9.2% to 28.0%) for both dry biomass (total and bole) and for Carbon (total and bole), was the template of Schumacher & Hall. 300 ANNs of the MLP type (multilayer perceptron) for each dependent variable were trained and the 10 networks with best results were retained for the analysis of precision measurements. All the best networks found had significantly better accuracy measures than those achieved by the Schumacher and Hall model. The stocks of dry biomass, total and bole, and carbon, total and bole, were respectively 65.61 ± 15.52 t ha-1, 34.17 ± 7.85 t ha-1, 29,47 ± 6,93 t.ha-1 e 15,16 ± 3,48 t.ha-1. The species with highest values of IVI and stock of dry biomass and carbon were Guazuma ulmifolia, Callisthene fasciculata, Myracrodruon urundeuva, Dilodendron bipinnatum.
Paiva, Medeiros de Farias Gilles. "Detecção de intrusão em redes de computadores: uma abordagem usando extreme learning machines." Universidade Federal de Pernambuco, 2011. https://repositorio.ufpe.br/handle/123456789/2820.
O mundo dos negócios é definido por qualidade, competitividade e luta por fatias de mercado. A informação é uma ferramenta indispensável nesse meio, onde organizações a usam como diferencial competitivo, uma forma de obter vantagem frente aos competidores. Segundo a Techoje (revista de opinião do IETEC - Instituto de Educação e Tecnologia), a quantidade de informação criada no ano de 2006 seria bastante para escrever 12 pilhas de livros, cada uma medindo 150 milhões de quilômetros, o que corresponde à distância da Terra ao Sol. De acordo com a Techoje, estudos estipulam que essa quantidade teria aumentado até 6 vezes até o ano de 2010. As redes de computadores são os meios utilizados para o compartilhamento dessas tão valiosas informações e sofrem com constantes tentativas de intrusão e com surgimentos cada vez mais acelerados de softwares maliciosos, que se disseminam pelos sistemas computacionais. Frente a essa realidade, IDS (Intrusion Detection Systems - Sistemas de Detecção de Intrusão) são ferramentas que auxiliam desde usuários comuns até grandes organizações a se manter seguros, contra invasores e ataques das mais diversas naturezas. Apesar de serem ferramentas úteis a seu propósito, IDS´s necessitam de implantação planejada e estruturada, ou efeitos, tais como lentidão no ambiente, alarmes falsos ou intrusões não detectadas podem vir a acontecer. O presente trabalho foca no estudo da construção de IDS´s, levando em conta as técnicas ELM (Extreme Learning Machine) e OS-ELM (Online Sequential Extreme Learning Machine) aplicadas ao problema. As técnicas citadas são usadas para o treinamento de redes neurais artificias do tipo feedforward e vêm sendo usadas em vários estudos em outras áreas de aplicação. Tais técnicas conseguem resolver problemas de forma mais rápida que técnicas tradicionais de treinamento de redes neurais, como o algoritmo backpropagation. Os resultados obtidos no estudo mostraram-se relevantes, pois alcançaram boas taxas de generalização e tempo computacional, que são fatores críticos para a área de segurança. Dessa forma, o presente estudo utiliza de forma pioneira as duas técnicas citadas, que pelas suas características, conseguem dar respostas rápidas frente ao surgimento de novos ataques
ATHREYA, JOOTHIRAM JAYAMANI. "A COMPONENT LIBRARY OF ARTIFICIAL NEURON ACTIVATION FUNCTIONS." University of Cincinnati / OhioLINK, 2005. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1122071899.
Nagaoka, Maria Eiko [UNESP]. "Uma abordagem baseada em redes neurais artificiais para a estimação de densidade de solo." Universidade Estadual Paulista (UNESP), 2003. http://hdl.handle.net/11449/101853.
Este trabalho apresenta a aplicação de um sistema inteligente utilizando redes neurais artificiais para estimar valores de densidade do solo, a partir de parâmetros referentes à resistência do solo à penetração. Foram considerados solos preparados e não preparados, os não preparados foram os seguintes : teor de argila menor que 30 % (solo tipo 1), de 30 a 50 % (solo tipo 2) e maior que 50 % (solo tipo 3). Os preparados foram os seguintes: um com teor de argila menor que 30 % (solo tipo 1) e o outro com teor de argila maior que 50 % (solo tipo 3). O objetivo principal deste trabalho foi implementar diversas redes neurais do tipo perceptron multicamadas, alimentando-as com resistência do solo à penetração, teor de água e teor de argila, tendo como variável de saída a densidade do solo. Cada rede foi treinada variando o número de camadas escondidas e também variando o número de neurônios, de 10 a 40, em cada camada. Para cada arquitetura, a rede foi treinada 10 vezes, escolhendo-se no final do treinamento a arquitetura com menor erro relativo médio e menor variância em relação aos dados de validação. As análises realizadas mostraram que as arquiteturas de rede com apenas uma camada escondida forneceram melhores resultados. Todas as redes tiveram melhor desempenho em solo não preparado do que em solo preparado. A rede de arquitetura de 3 entradas, uma camada escondida com 30 neurônios e 1 saída forneceu excelente resultado para solo não preparado (com teor de argila entre 30 e 50 %). Constatou-se que a rede quando treinada com dados do solo preparado, juntamente com dados do solo não preparado, melhorou os resultados de estimação para o solo preparado, mas piorou para os solos não preparados. Constatou também que a rede quando treinada junto com dados que contém solo solto fornece resultados imprecisos. O mesmo ocorreu para dados com teor de água elevado.
This work presents the development of an intelligent system using artificial neural networks to estimate values of soil density. Prepared and non-prepared soils were considered in this work. The non-prepared soils were the following ones: clay content lesser than 30 % (soil type 1), 30 to 50 % (soil type 2) and larger than 50 % (soil type 3). The prepared soils were the following ones: soil with clay content lesser than 30 % (soil type 1) and soil with clay content larger than 50 % (soil type 3). The main objective of this work was to implement several neural networks of type multilayer perceptron, feeding them with data concerning to the soil compaction characteristics. The output computed by the neural network was the respective density of these soils. Each neural network was trained varying both number of hidden layers and number of neurons, which was changed from 10 to 40 neurons in each layer. In each architecture the network was trained 10 times and selected architecture was always that having either the least mean relative error or the least variance in relation to validation data. The carried out analyses showed that the neural architectures having only a hidden layer were those that provided the best results. All neural networks have presented more efficient results for non-prepared soils than prepared soils. The neural network constituted by three inputs and one output, having 30 neurons at hidden layer, has provided excellent results for non-prepared soils (clay content between 30 and 50 %). It was also verified that the neural network when trained with data referent to non-prepared and soils, which were put in the same data set, it became the results referent to prepared soils more efficient, but the results for non-prepared soils become worse. Another observed point was when the network had been trained with data constituted by soft soil... (Complete abstract, click electronic address below).
Varotto, Sebastiao Eduardo Corsatto. "Controle de atitude de satélites artificiais terrestres utilizando redes neurais artificiais." Instituto Nacional de Pesquisas Espaciais (INPE), 1997. http://urlib.net/sid.inpe.br/iris@1905/2005/07.29.07.24.
This work evaluates the use of artificial neural networks for satellite attitude dynamics identification and control. In order to exemplify this application, a satellite with a rigid main body, three reaction wheel and three flexible solar panels is chosen (lay-out similar to Brazilian Remote Sensing Satellite). The differential equations therefore show the nonlinear dynamic effects to be identified by neural nets. The equations of motion used to generate training and testing data are derived by the Lagrangian approach for quasi-coordinates (rotational motion) and for generalized coordinates (elastic motion). To validate the attitude control the situation of fine pointing, with open solar panels is considered in the presence of perturbing torques and errors in the satellite moments of inertia. The identification of neural nets parameters is performed by a Kahnan filtering algorithm local parallel processing and control structure used is that of the Internal Model Control (IMC).
Pegorini, Vinicius. "Classificação de padrões de mastigação de ruminantes utilizando aprendizagem de máquina." Universidade Tecnológica Federal do Paraná, 2015. http://repositorio.utfpr.edu.br/jspui/handle/1/1508.
In this work, a system to automate the classification of chewing patterns of ruminants is developed. Pattern classification of ingestive behavior in grazing animals has extreme importance in studies related to animal nutrition, growing and health. The proposal is based on data collected by optical fiber Bragg grating sensors (FBG). The collected data are processed by machine learning techniques. The FBG sensors measure the biomechanical strain during jaw movements. In vitro tests were performed based on a cadaveric mandible of a goat and in vivo tests were performed by monitoring the food intake process of a steer. For the pattern classification a classic model of decision tree and artificial neural network were used. Experimental results show that the proposed approaches for pattern classification have been capable to differentiate the materials and events involved in the chewing process. Experimental results show that it is possible to classify different forage and events involved in the ingestive behaviour of ruminants, that contributes to improving the current methodology for monitoring the animal consumption efficiency.
Alvado, Ludovic. "Neurones artificiels sur silicium : une évolution vers les réseaux." Bordeaux 1, 2003. http://www.theses.fr/2003BOR12674.
This thesis describes a new approach for modelling biological neuron networks. This approach uses analogue specific integrated circuit (ASIC) in which Hodgkin-Huxley formalism as been implemented to integrate medium density artificial neural network, modelled at a biological realistic level. This thesis also deals with the component mismatches problem and the pertinent choice of optimized structure dedicated to network applications
Schuler, Joao Paulo Schwarz. "Inteligência artificial popperiana." reponame:Biblioteca Digital de Teses e Dissertações da UFRGS, 2002. http://hdl.handle.net/10183/5901.
Nagaoka, Maria Eiko. "Uma abordagem baseada em redes neurais artificiais para a estimação de densidade de solo /." Botucatu : [s.n.], 2003. http://hdl.handle.net/11449/101853.
Abstract: This work presents the development of an intelligent system using artificial neural networks to estimate values of soil density. Prepared and non-prepared soils were considered in this work. The non-prepared soils were the following ones: clay content lesser than 30 % (soil type 1), 30 to 50 % (soil type 2) and larger than 50 % (soil type 3). The prepared soils were the following ones: soil with clay content lesser than 30 % (soil type 1) and soil with clay content larger than 50 % (soil type 3). The main objective of this work was to implement several neural networks of type multilayer perceptron, feeding them with data concerning to the soil compaction characteristics. The output computed by the neural network was the respective density of these soils. Each neural network was trained varying both number of hidden layers and number of neurons, which was changed from 10 to 40 neurons in each layer. In each architecture the network was trained 10 times and selected architecture was always that having either the least mean relative error or the least variance in relation to validation data. The carried out analyses showed that the neural architectures having only a hidden layer were those that provided the best results. All neural networks have presented more efficient results for non-prepared soils than prepared soils. The neural network constituted by three inputs and one output, having 30 neurons at hidden layer, has provided excellent results for non-prepared soils (clay content between 30 and 50 %). It was also verified that the neural network when trained with data referent to non-prepared and soils, which were put in the same data set, it became the results referent to prepared soils more efficient, but the results for non-prepared soils become worse. Another observed point was when the network had been trained with data constituted by soft soil... (Complete abstract, click electronic address below).
Orientador: Ivan Nunes da Silva
Coorientador: Kléber Pereira Lanças
Doutor
Herng, Eduardo Wu Jyh. "Detecção de bordas em imagens de ecocardiografia utilizando redes neurais artificiais." Universidade de São Paulo, 2012. http://www.teses.usp.br/teses/disponiveis/98/98131/tde-04062012-083028/.
Being non-invasive and having low cost, the echocardiography has been largely applied as diagnostic technique for left ventricle systolic and diastolic volumes determination that indirectly are used to calculate the left ventricle ejection volume, the cardiac cavities muscular contraction, the regional and global ejection fraction, the myocardial thickness, the ventricular mass, etc. For this reason, the detection of the left ventricle endocardial borders become necessary, but hampered by the noise that impairs the echocardiography images definition. In spite of having many image segmentation techniques, this work intend to detect the borders of left ventricle on echocardiography images by using a artificial neural network to recognize border patterns. To accelerate the process and facilitate the procedure, the operator uses the mouse to define a rectangular region inside the acoustic window of the pacient where all analyses and border recognitions will be accomplished. After labeling the recognized points as \'border\', gradient techniques and mobile boundary are used to connect the points of greater probability and delineate the left ventricle border. This technique has proved to be efficient when compared to the borders traced by the specialist. The ability of the operator is important in choosing of the region to be analyzed. After training with 50 samples of \"border\" pattern and 10 samples of \"no-border\" pattern, this technique was tested on 108 images, achieving good results on precision and velocitiy when we compared the calculated left ventricle area with the results of other techniques published on national and international literature.
Brady, Patrick. "Internal representation and biological plausibility in an artificial neural network." Thesis, Brunel University, 1995. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.311273.
Barbieri, Heitor. "Sirena : um simulador de redes neurais artificiais." [s.n.], 1994. http://repositorio.unicamp.br/jspui/handle/REPOSIP/276032.
Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Matematica, Estatistica e Ciencia da Computação
Made available in DSpace on 2018-07-19T14:51:10Z (GMT). No. of bitstreams: 1 Barbieri_Heitor_M.pdf: 3338070 bytes, checksum: 7e01fc370d8436edb4d799783be57ff8 (MD5) Previous issue date: 1994
Resumo: Rede Neural Artificial (RNA) é um modelo que tenta emular uma Rede Neural Biológica. A área de RNA tem se mostrado bastante promissora, o que pode ser comprovado pela quantidade de trabalhos publicados e de eventos científicos. Mas para que as RNAs atinjam o escopo de aplicações desejado, muitas de suas limitações atuais terão que ser superadas. Ainda não é claro e bem estabelecido o funcionamento das RNAs, não existem metodologias boas e completas para a utilização das mesmas em aplicações, isto é, metodologias que diante de um problema específico a ser resolvido, indiquem qual a topologia de rede, o algoritmo de aprendizagem e a amostragem de informações adequadas ao funcionamento desejado. Em não se tendo uma metodologia que indique a combinação ótima dos elementos de uma RNA para uma determinada aplicação, resta aos usuários a opção de partir de uma base teórica e, utilizando-se de métodos empíricos, ir formando regras individuais de como conseguir as melhores combinações dos elementos formadores da rede. Esta técnica, porém, apresenta muitas dificuldades em sua realização devido à grande quantidade de variáveis que precisam ser avaliadas durante todo o processo de desenvolvimento da rede. O presente trabalho busca facilitar o entendimento do funcionamento das RNAs através da familiarização do usuário com os seus elementos formadores. Foi desenvolvido um simulador de RNAs, denominado Sirena, que através de sua interface gráfica procura minimizar a dificuldade de entendimento dos processos de baixo nível realizados pelas RNAs. Durante o processo de simulação pode-se ter acesso a diversas representações, tanto qualitativas quanto quantitativas, que visam refletir as alterações que ocorrem na rede rias fases de aprendizagem e inferência.
Abstract: Artificial Neural Net (ANN) is a model that emulates a Biological Neural Net. The ANN field has showed very promising which can be verified by the number of published papers and scientific events. In spite, to reach the desired ANN applications scope, many of ANN current limitations have to be overcome since it is not yet and well established the ANN functioning . There is no good and complete methodologies for construct ANN applications, i.e., for a specific problem to be solved, no methodology indicates what the net topology is, the learning algorithm and the sample of information suitable to the desired performance. If there is no methodology that indicates the better combination of the ANN elements to a specific application, the users have the option to start from a theoretical base and, by using empirical methods, begin constructing personal rules that indicates. the better combination of neural elements. The execution of this technique is difficulty because the number of variables that need to be evaluated during the net development process The focus of this work is facilitate the understanding of the ANN functioning through the user familiarization with its elements. A ANN simulator named Sirena was developed and its graphical interface aim to minimize the understanding difficulties of the low level processes executed by ANNs. During the simulation process one can access to several qualitative and quantitative representations that reflect the net alterations in the learning and inference phases.
Mestrado
Mestre em Ciência da Computação
Martineli, Edmar. "Extração de conhecimento de redes neurais artificiais." Universidade de São Paulo, 1999. http://www.teses.usp.br/teses/disponiveis/55/55134/tde-19102001-100256/.
This work describes experiments carried out witch Artificial Neural Networks and symbolic learning algorithms. Two algorithms for knowledge extraction from Artificial Neural Networks are also investigates. This experiments are performed whit three data set with the objective of compare the performance obtained. The data set used in this work are: Brazilians banks bankruptcy data set, tic-tac-toe data set and credit analysis data set. Three techniques for data set performance improvements are investigates. These techniques are: partition for the smallest class, noise increment in the examples of the smallest class and selection of more important attributes. Besides the analysis of the performance obtained, an analysis of the understanding difficulty of the knowledge extracted by each method in each data bases is made.
Oliveira, Patrícia Rufino. "Redes Neurais Artificiais para Extração de Características." Universidade de São Paulo, 1997. http://www.teses.usp.br/teses/disponiveis/55/55134/tde-20032018-141939/.
Methods for feature extraction are used to select from an initial data set, some features that represent the most important information of this set or that are essential to differentiate one class of objects from other. In this work, two methodologies that can be used for feature extraction are presented. The first uses classical statistical methods such as Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA) and Cluster Analysis. The other approach is based on Artificial Neural Networks architectures that implement the same statistical methods. The performance of the presented neural network models is appraised considering the use of these in the feature extraction of a small data set. Also, to investigate the usability of these models in applications of image processing, one of the neural networks that implements PCA is used for compressing some medical images. The results obtained by the PCA network are compared with others obtained by applying classical PCA and JPEG compression standard to the same group of images.
Ballini, Rosangela. "REDES NEURAIS ARTIFICIAIS PARA PREVISAO CHUVA/VAZAO." Universidade de São Paulo, 1996. http://www.teses.usp.br/teses/disponiveis/55/55134/tde-24112017-103742/.
Artificial Neural Networks have been widely used in a variety of arcas. One of these arcas is time forecasting. In this work, neural network models known as Kohonen and Multi-layer perceptron with algorithm back-propagation are utilized in inflow forecasting. Moreover, these methods are compared with the nearest-neighbor method which have been utilized in inflow forecasting. A comparative analyze is made using the data of the Atibaia River basin and the results show the advantages and disadvantages of the techniques used.
Guimarães, Lourenço da Rocha. "Previsão de inadimplência e redes neurais artificiais." Universidade do Estado do Rio de Janeiro, 2006. http://www.bdtd.uerj.br/tde_busca/arquivo.php?codArquivo=383.
The objective of this work is the insolvency forecast. They were implemented two insolvency forecast models, so that the first model used a feed forward neural network utilizing the retro propagation algorithm, and the second utilized a non-supervised neural network (Kohonen networks). The prominent characteristics of credit users were presented for the neural networks, for their training and test. The results obtained showed that the supervised network as well the non-supervised neural network showed themselves efficient instruments for the insolvency forecast trial.
Ribeiro, Eduardo Ferreira. "Caracterização de imagens utilizando redes neurais artificiais." Universidade Federal de Uberlândia, 2009. https://repositorio.ufu.br/handle/123456789/12481.
Image representation in Content Based Image Retrieval systems is a fundamental task. The results obtained by these systems strongly depend on the choice of features selected to represent an image. Works in the literature show that intelligent techniques are used to minimize the semantic gap between the limited power of machine interpretation and human subjectivity. In this work the use of artificial neural networks to characterize images in a high-level space from an initial characterization based on low-level features (color, shape and texture) is proposed. Experiments on 3 databases of various kinds, one with general images (BD-12750 ), one with texture images (Vistex-167 ) and other with buildings (ZuBuD) are performed to exemplify the application of the method and to show the effectiveness of the model. Furthermore, the application of the proposed method in the high-level characterization of complex motions patterns is presented.
Em sistemas de Recuperação de Imagens Baseada em Conteúdo a representação das imagens desempenham um papel fundamental. Os resultados obtidos por esses sistemas dependem fortemente da escolha das características selecionadas para representar uma imagem. Trabalhos existentes na literatura evidenciam que técnicas inteligentes conseguem minimizar o gap- semântico existente entre o poder de interpretação limitado das máquinas e a subjetividade humana. Neste trabalho é proposto o uso das redes neurais artificiais para caracterizar imagens neurosemânticamente à partir de uma caracterização inicial baseada em características de baixo nível (cor, forma e textura). Testes em 3 bases de dados de naturezas diferentes, um de imagens mais gerais (BD-12750 ), um de texturas (Vistex-167 ) e outro de prédios (ZuBuD) exemplificam a aplicação do método como também mostram a eficácia do modelo. Ainda é apresentada a aplicação do método proposto na caracterização neurosemântica de movimentos complexos em vídeos.
Mestre em Ciência da Computação
Muller, Daniel Nehme. "Reconhecimento semântico através de redes neurais artificiais." reponame:Biblioteca Digital de Teses e Dissertações da UFRGS, 1996. http://hdl.handle.net/10183/23955.
One of the great challenges of computation nowadays is to cross the abyss between man and machine. Thus, the challenge becomes the formalization of mental states and its computational modelling. This is necessary since man will only get to communicate with a machine when this machine is able to give and receive information without man needs to learn a special way to communicate. Therefore, it is necessary that the machine learns to communicate with man. In this sense, the study of the language becomes an open door in order to create a computation that may be adapted to man. and, at the same time, may help researches which aim at a better comprehension of the brain functioning of the language and of man's learning. This work shows that the computer has a potential for communication that has not been explored yet. For this reason, in prior studies we tried to verify the present stage of man-machine communication modelling in comparison with the human language evolution. We verified, then, that the machine can reach an effective communication with man, but never spontaneous, as we see in scientific fiction (Sci-Fi). What can be possible is the self-organization by computer of signals deriving from its own environment, aiming at realization of specifics tasks. Those signals of the computer environment are exactly what justifies its actions. what gives meaning to what is transmitted to it in the same way that happens with man. In order to mould the Semantic Recognition of phrases it is necessary to find out a way of codifying the signals of the environment so that these signals. accompanying a phrase, may permit recognition of its meaning. However, as the purpose of this work is the implementation of the Semantic Recognition, and not the reception of signals, we have opted for a representative codification of external signals. This codification allows that, through the Artificial Neural Nets technology, the implementation of semantic relations among words and phrases may be possible, permitting the classification for posterior recognition. The computational implementation realized permits the recognition of phrases, even with alteration of words and number of words. The prototype presented here shows that, even with one structure extremely simpler than other systems of Natural Language Recognition, an adequate identification of phrases is possible.
Schwingel, Dinamerico. "Um simulador distribuido para redes neurais artificiais." reponame:Biblioteca Digital de Teses e Dissertações da UFRGS, 1995. http://hdl.handle.net/10183/25775.
The use of workstation networks as distributed multicomputers to solve resource demanding problems that cannot be feasibly solved in one node is the main concern of this work. Two different artificial neural network algorithms, Combinatorial Neural Model and Back Propagation, are faced and a scheme for distributing this class of algorithms is presented. The several advantages of the environment are focused in the proposal along with its disadvantages. This work also presents the implementation of the proposed scheme allowing an in loco performance evaluation. At the end results are shown and a more in depth evaluation of the Back Propagation parallelization is presented.
Scheidt, Felippe Alex. "Modelagem chuva-vazão utilizando redes neurais artificiais." Universidade Estadual de Londrina. Centro de Ciências Exatas. Programa de Pós-Graduação em Ciência da Computação, 2010. http://www.bibliotecadigital.uel.br/document/?code=vtls000160789.
In this work it is proposed a methodology for modeling the rainfall-runo_ relantionshipof a speci_c watershed, throught articial neural networks (ANN) and geneticalgorithm. This model was developed on events based on daily and monthly observations.The study case is a watershed of the state of Parana, Brasil, named Piquiri river Basin. The results are compared with an autorregressive and movingaverage model, and showed that neural networks has superior capacity to representthe rainfall-runo relationship. Besides, the ANN methodology was compared witha hybrid model, which coupling ANN with genetic algorithm, and the results showedthat the hybrid model tted better than ANN, when representing the rainfall-runo transformation.
Alves, Daniel Pedrosa. "Predição da área abaixo da curva de progresso da requeima em tomateiro utilizando inteligência artificial." Universidade Federal de Viçosa, 2014. http://www.locus.ufv.br/handle/123456789/6862.
Made available in DSpace on 2015-12-02T14:33:09Z (GMT). No. of bitstreams: 1 texto completo.pdf: 869288 bytes, checksum: 1ec9e1ebef3bae322c6fce4583cbe976 (MD5) Previous issue date: 2014-03-27
Conselho Nacional de Desenvolvimento Científico e Tecnológico
Redes neurais artificiais (RNA) são modelos computacionais inspirados no sistema nervoso de seres vivos, capazes de aprender a partir de exemplos e empregá-lo na solução de problemas tais como predição não linear, reconhecimento de padrões e diversas outras aplicações. Neste trabalho utilizamos uma RNA para predizer o valor da área abaixo da curva de progresso da doença (AACPD) para o patossistema tomate x requeima. A AACPD é uma medida de ampla utilização na epidemiologia de doenças policíclicas, especialmente em estudos que inferem a respeito da resistência quantitativa dos genótipos. Contudo, para a obtenção do valor final desta área são necessárias, neste patossistema, uma série de seis avaliações ao longo do tempo. O objetivo deste trabalho é propor a utilização das RNAs para a obtenção da AACPD no patossistema tomate x requeima, utilizando um número reduzido de avaliações de severidade. Para tanto, foram considerados quatro experimentos independentes, totalizando 1836 plantas infectadas com o patógeno Phytophthora infestans e avaliadas a cada três dias em um total de seis oportunidades, sendo procedido o cálculo da AACPD por método convencional. A RNA criada permitiu predizer AACPD com correlação de 0,97 e 0,84 quando comparado com os métodos convencionais, utilizando-se de um número 50% e 67% menor de avaliações por genótipo respectivamente. Ao se utilizar a RNA gerada por um experimento para predizer a AACPD para os demais experimentos ocorreu correlação média de 0,94, com duas avaliações, e 0,96, com três avaliações, entre os valores preditos pela RNA e os observados com seis avaliações. Apresentamos neste trabalho um novo paradigma para a utilização da informação da AACPD em experimentos de tomateiro confrontado com P. infestans. Este novo paradigma proposto pode ser adaptado para diferentes patossistemas.
Artificial neural networks (ANN) are computational models, inspired in the nervous system of living organisms, that is able to learn from examples and uses it to solve problems such as non-linear prediction, pattern recognition, and many other applications. In this work we use an ANN to predict the value of the area under the disease progress curve (AUDPC) for pathosystem tomato x late blight. The AUDPC is a widely used measure in the epidemiology of polycyclic diseases, especially in studies about quantitative resistance of genotypes. However, to obtain the final value of this area is required, in this pathossystem, a series of six evaluations along time. The objective of this paper is to propose a new use of ANN, based on the principles of learning, for to obtain the AUDPC in pathosystem tomato x late blight, using a reduced number of disease severity evaluations. We considered four independent experiments, a total of 1836 infected plants with the pathogen Phytophthora infestans and assessed every three days for six times, and proceeded to calculate the AUDPC by conventional methods. The ANN created possible to predict the AUDPC with a correlation coefficient of 0.97 and 0.84 compared with conventional methods, using a number 50% and 67% less ratings for genotypes respectively. Using ANN generated by an experiment to predict the AUDPC for the other experiments there was an average correlation of 0.94, with two ratings, and 0.96, with three evaluations, between the value predicted from ANN and value observed with six evaluations. We present in this work a new paradigm for obtaining AUDPC in tomato experiments inoculated with P. infestans. This proposed new paradigm can be adapted to different pathosystems.
Voysey, Matthew David. "Inexact analogue CMOS neurons for VLSI neural network design." Thesis, University of Southampton, 1998. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.264387.
Henseler, Johan. "Connections, neurons and activation the organization of representation in artificial neural networks /." Maastricht : Maastricht : Rijksuniversiteit Limburg ; University Library, Maastricht University [Host], 1993. http://arno.unimaas.nl/show.cgi?fid=5754.
Sporl, Christiane. "Metodologia para elaboração de modelos de fragilidade ambiental utilizando redes neurais." Universidade de São Paulo, 2007. http://www.teses.usp.br/teses/disponiveis/8/8135/tde-10102007-145640/.
This paper deals with the challenge in modeling environmental fragility, which implies not only the understanding of the intrinsic and dynamic relationship that exists between the physical, biotic and socio-economic components of environmental systems, but also in translating this knowledge in a mathematical model. In order to shed light on this difficulty, the results generated by two empirical models of environmental fragility were presented and compared, models that are widely used in Brazilian physical-territorial planning. (CREPANI et al. 2001 and ROSS, 1994). These two models were applied in two thesis-areas with very diverging results. Within this context of uncertainties, this paper tested the feasibility and reliability of a new tool to be applied in the elaboration of environmental fragility models, the artificial neural networks (ANN). Tapping on the knowledge and experience of specialists in this area, extracted from the answers given by them during the comparison of variables and scenarios applied in programs adapted for this objective: Gauging Research, Scheduling of Variables Research and Scenario Evaluation Research. These programs generated a databank related to the evaluation format of each specialist regarding environmental fragility applied in the training of ANNs, so that the network would assimilate the evaluation standard of that specialist. The results proved that it is possible to emulate, with reasonable reliability, the evaluation standard of specialists in the definition of environmental systems fragility, eliminating in this way, arbitrariness and subjectivity in the elaboration process of environmental fragility models. This work does not presuppose a new model, rather a methodology for the construction of models, using artificial neural networks, taking the first step in the search of new techniques, albeit feared by the geographers, however, necessary for the evolution of geographic science.