Thesis (M.S.)--State University of New York at Binghamton, Thomas J. Watson School of Engineering and Applied Science, Department of Electrical and Computer Engineering, 2008. / Includes bibliographical references.

Thesis (Ph.D.)--University of North Carolina at Charlotte, 2005. / Includes bibliographical references (leaves 71-79).

Philosophiae Doctor - PhD / Promoter detection, especially in prokaryotes, has always been an uphill task and may remain so, because of the many varieties of sigma factors employed by various organisms in transcription. The situation is made more complex by the fact, that any seemingly unimportant sequence segment may be turned into a promoter sequence by an activator or repressor (if the actual promoter sequence is made unavailable). Nevertheless, a computational approach to promoter detection has to be performed due to number of reasons. The obvious that comes to mind is the long and tedious process involved in elucidating promoters in the ‘wet’ laboratories not to mention the financial aspect of such endeavors. Promoter detection/prediction of an organism with few characterized promoters (M.tuberculosis) as envisaged at the beginning of this work was never going to be easy. Even for the few known Mycobacterial promoters, most of the respective sigma factors associated with their transcription were not known. If the information (promoter-sigma) were available, the research would have been focused on categorizing the promoters according to sigma factors and training the methods on the respective categories. That is assuming that, there would be enough training data for the respective categories. Most promoter detection/prediction studies have been carried out on E.coli because of the availability of a number of experimentally characterized promoters (+- 310). Even then, no researcher to date has extended the research to the entire E.coli genome. / South Africa

Associative Memories (AMs) are essential building blocks for brain-like intelligent computing with applications in artificial vision, speech recognition, artificial intelligence, and robotics. Computations for such applications typically rely on spatial and temporal associations in the input patterns and need to be robust against noise and incomplete patterns. The conventional method for implementing AMs is through Artificial Neural Networks (ANNs). Improving the density of ANN based on conventional circuit elements poses a challenge as devices reach their physical scalability limits. Furthermore, stored information in AMs is vulnerable to destructive input signals. Novel nano-scale components, such as memristors, represent one solution to the density problem. Memristors are non-linear time-dependent circuit elements with an inherently small form factor. However, novel neuromorphic circuits typically use memristors to replace synapses in conventional ANN circuits. This sub-optimal use is primarily because there is no established design methodology to exploit the memristor's non-linear properties in a more encompassing way. The objective of this thesis is to explore denser and more robust AM designs using memristor networks. We hypothesize that such network AMs will be more area-efficient than the traditional ANN designs if we can use the memristor's non-linear property for spatial and time-dependent temporal association. We have built a comprehensive simulation framework that employs Genetic Programming (GP) to evolve AM circuits with memristors. The framework is based on the ParadisEO metaheuristics API and uses ngspice for the circuit evaluation. Our results show that we can evolve efficient memristor-based networks that have the potential to replace conventional ANNs used for AMs. We obtained AMs that a) can learn spatial and temporal correlation in the input patterns; b) optimize the trade-off between the size and the accuracy of the circuits; and c) are robust against destructive noise in the inputs. This robustness was achieved at the expense of additional components in the network. We have shown that automated circuit discovery is a promising tool for memristor-based circuits. Future work will focus on evolving circuits that can be used as a building block for more complicated intelligent computing architectures.

Chemical reaction networks are an unconventional computing medium that could benefit from the ability to form basic control systems. In this work, we demonstrate the functionality of a chemical control system by evaluating classic genetic algorithm problems: Koza's Santa Fe trail, Jefferson's John Muir trail, and three Santa Fe trail segments. Both Jefferson and Koza found that memory, such as a recurrent neural network or memories in a genetic program, are required to solve the task. Our approach presents the first instance of a chemical system acting as a control system. We propose a delay line connected with an artificial neural network in a chemical reaction network to determine the artificial ant's moves. We first search for the minimal required delay line size connected to a feed forward neural network in a chemical system. Our experiments show a delay line of length four is sufficient. Next, we used these findings to implement a chemical reaction network with a length four delay line and an artificial neural network. We use genetic algorithms to find an optimal set of weights for the artificial neural network. This chemical system is capable of consuming 100% of the food on a subset and greater than 44% of the food on Koza's Santa Fe trail. We also show the first implementation of a simulated chemical memory in two different models that can reliably capture and store information over time. The ability to store data over time gives rise to basic control systems that can perform more complex tasks. The integration of a memory storage unit and a control system in a chemistry has applications in biomedicine, like smart drug delivery. We show that we can successfully store the information over time and use it to act as a memory for a control system navigating an agent through a maze.

Thesis (Ph. D.)--Brandeis University, 2009. / "UMI:3369218." MICROFILM COPY ALSO AVAILABLE IN THE UNIVERSITY ARCHIVES. Includes bibliographical references.

Bibliography: leaves 179-190. / xii, 189 leaves : ill. ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / This thesis proposes a genetic-programming-based classifier system for the diagnosis of coronary artery disease. Based on genetic programming, a software system called Evolutionary Pre-Processor has been developed as a new method for the automatic extraction of non-linear features for supervised classification. Two different hybrid intelligent system techniques are presented; fuzzy systems integrated with genetic algorithms and genetic algorithms combined with back-propagation algorithms. All approaches were tested on a real-world problem of coronary artery disease data. / Thesis (Ph.D.)--University of Adelaide, Dept. of Applied Mathematics, 1998

Thesis (M.Sc)--Athabasca University, 2004. / Includes bibliographical references (leaves 31-32). Also available online.

Orientador: Ricardo da Silva Torres / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Computação / Made available in DSpace on 2018-08-08T22:04:09Z (GMT). No. of bitstreams: 1 Ferreira_CristianoDalmaschio_M.pdf: 3661487 bytes, checksum: 589a6834c502d67559dbb716e1dd4645 (MD5) Previous issue date: 2007 / Resumo: A técnica de realimentação de relevância tem sido utilizada com o intuito de incorporar a subjetividade da percepção visual de usuários à recuperação de imagens por conteúdo. Basicamente, o processo de realimentação de relevância consiste na: (i) exibição de um pequeno conjunto de imagens; (ii) rotulação dessas imagens pelo usuário, indicando quais são relevantes ou não; (iii) e finalmente, aprendizado das preferências do usuário a partir das imagens rotuladas e seleção de um novo conjunto de imagens para exibição. O processo se repete até que o usuário esteja satisfeito. Esta dissertação apresenta dois arcabouços para recuperação de imagens por conteúdo com realimentação de relevância. Esses arcabouços utilizam programação genética para assimilar a percepção visual do usuário por meio de uma combinação de descritores. A utilização de programação genética é motivada pela sua capacidade exploratória do espaço de busca uma vez que esse espaço se adequa ao objetivo principal dos arcabouços propostos: encontrar, dentre todas as possíveis funções de combinação de descritores, aquela que melhor representa as características visuais que um usuário deseja ressaltar na realização de uma consulta. Os arcabouços desenvolvidos foram validados por meio de uma série de experimentos, envolvendo três diferentes bases de imagens e descritores de cor, forma e textura para a caracterização do conteúdo dessas imagens. Os arcabouços propostos foram comparados com três outros métodos de recuperação de imagens por conteúdo com realimentação de relevância, considerando-se a eficiência e a efetividade no processo de recuperação. Os resultados experimentais mostraram a superioridade dos arcabouços propostos. As contribuições dessa dissertação são: (i) estudo sobre diferentes técnicas de realimentação de relevância; (ii) proposta de dois arcabouços para recuperação de imagens por conteúdo com realimentação de relevância baseado em programação genética; (iii) implementação dos métodos propostos, validando-os por meio de uma série de experimentos e comparações com outros métodos / Abstract: Relevance Feedback has been used to incorporate the subjectivity of user visual perception in content-based image retrieval tasks. The relevance feedback process consists in the following steps: (i) showing a small set of images; (ii) indication of relevant or irrelevant images by the user; (iii) and finally, learning the user needs from her feedback, and selecting a new set of images to be showed. This procedure is repeated until the user is satisfied. This dissertation presents two content-based image retrieval frameworks with relevance feedback. These frameworks employ Genetic Programming to discover a combination of descriptors that characterize the user perception of image similarity. The use of genetic programming is motivated by its capability of exploring the search space, which deals with the major goal of the proposed frameworks: find, among all combination functions of descriptors, the one that best represents the user needs. Several experiments were conducted to validate the proposed frameworks. These experiments employed three different images databases and color, shape and texture descriptors to represent the content of database images. The proposed frameworks were compared with three other content-based image retrieval methods regarding their efficiency and effectiveness in the retrieval process. Experiment results demonstrate the superiority of the proposed methods. The contributions of this work are: (i) study of different relevance feedback techniques; (ii) proposal of two content-based image retrieval frameworks with relevance feedback, based on genetic programming; (ii) implementation of the proposed methods and their validation with several experiments, and comparison with other methods / Mestrado / Banco de Dados / Mestre em Ciência da Computação

Orientadores: Ariadne Maria Brito Rizzoni Carvalho, Ricardo da Silva Torres / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Computação / Made available in DSpace on 2018-08-24T12:42:46Z (GMT). No. of bitstreams: 1 Godoi_ThiagoAnzolinde_M.pdf: 1782345 bytes, checksum: d9ede100469835a7820e3cc67caae355 (MD5) Previous issue date: 2013 / Resumo: Este trabalho apresenta um novo método semiautomático para desambiguação de nomes que explora a utilização de iterações com realimentação de relevância. Uma etapa não supervisionada é utilizada para definir exemplos puros para o treinamento, e uma etapa híbrida supervisionada é empregada para aprender a função de classificação que irá atribuir autores a referências. O modelo combina um classificador por floresta de caminhos ótimos (OPF - Optimum-Path Forest) com uma função de similaridade complexa gerada por um algoritmo de Programação Genética (PG). As principais contribuições deste trabalho são: (i) proposta de um novo método para desambiguação de nomes de autores; (ii) avaliação em uma nova aplicação, da combinação entre os algoritmos OPF e PG, também conhecida como GOPF (Genetic Programming e Optimum-Path Forest), incrementada por uma etapa de realimentação de relevância; (iii) avaliação do algoritmo do GOPF em um problema de classificação multiclasse; e (iv) adaptação do algoritmo do GOPF para lidar com problemas de classificação de conjunto aberto, isto é, que não possuem todas as classes definidas previamente. O método proposto foi validado em duas coleções tradicionais muito utilizadas para avaliação de métodos de desambiguação de nomes de autores. A primeira é a coleção extraída da DBLP e que possui 4.287 referências associadas a 220 autores distintos; a segunda é chamada de KISTI, gerada pelo Korea Institute of Science Technology Information, e que contém os primeiros 1000 autores mais frequentes na versão do banco de dados da DBLP no final de 2007. Após 5 iterações de realimentação do usuário, nossa abordagem atingiu os melhores resultados para a desambiguação de nomes de autores quando comparado com os outros métodos existentes que utilizam somente as informações básicas da referência / Abstract: This work presents a new name disambiguation method that exploits user feedback on ambiguous references across iterations. An unsupervised step is used to define pure training samples, and a hybrid supervised step is employed to learn a classification model for assigning references to authors. Our disambiguation method combines the Optimum-Path Forest (OPF) classifier with complex reference similarity functions generated by a Genetic Programming (GP) framework. The main contributions of this work are: (i) proposal of a novel author name desambiguation method; (ii) evaluation in a new application of the combination between GP and OPF algorithms, also known as GOPF, in interaction learning systems; (iii) evaluation of the GOPF algorithm in a multi-class classification problem; and (iv) extension of the GOPF algorithm to handle open-set classification problems, i.e., classification problems in which class samples are not known in advance. The proposed method was validated with two traditional databases largely used for the evaluation of author name disambiguation methods: one is a collection extracted from DBLP which sums up 4,287 references associated with 220 distinct authors; the other is called KISTI and was built by the Korea Institute of Science and Technology Information; it contains the top 1000 most frequent author names from the late-2007 DBLP database. After 5 iterations of relevance feedback, our approach yielded the best results for author name disambiguation when compared with the state-of-the-art methods that just consider basic reference information, such as author names, publication title, and venue title / Mestrado / Ciência da Computação / Mestre em Ciência da Computação