Graphical models and point set matching

Point pattern matching in Euclidean Spaces is one of the fundamental problems in Pattern Recognition, having applications ranging from Computer Vision to Computational Chemistry. Whenever two complex patterns are encoded by two sets of points identifying their key features, their comparison can be seen as a point pattern matching problem. This work proposes a single approach to both exact and inexact point set matching in Euclidean Spaces of arbitrary dimension. In the case of exact matching, it is assured to find an optimal solution. For inexact matching (when noise is involved), experimental results confirm the validity of the approach. We start by regarding point pattern matching as a weighted graph matching problem. We then formulate the weighted graph matching problem as one of Bayesian inference in a probabilistic graphical model. By exploiting the existence of fundamental constraints in patterns embedded in Euclidean Spaces, we prove that for exact point set matching a simple graphical model is equivalent to the full model. It is possible to show that exact probabilistic inference in this simple model has polynomial time complexity with respect to the number of elements in the patterns to be matched. This gives rise to a technique that for exact matching provably finds a global optimum in polynomial time for any dimensionality of the underlying Euclidean Space. Computational experiments comparing this technique with well-known probabilistic relaxation labeling show significant performance improvement for inexact matching. The proposed approach is significantly more robust under augmentation of the sizes of the involved patterns. In the absence of noise, the results are always perfect.

Como tomadores de decisão experts percebem cenários complexos?

In this thesis, the basic research of Chase and Simon (1973) is questioned, and we seek new results by analyzing the errors of experts and beginners chess players in experiments to reproduce chess positions. Chess players with different levels of expertise participated in the study. The results were analyzed by a Brazilian grandmaster, and quantitative analysis was performed with the use of statistical methods data mining. The results challenge significantly, the current theories of expertise, memory and decision making in this area, because the present theory predicts piece on square encoding, in which players can recognize the strategic situation reproducing it faithfully, but commit several errors that the theory can¿t explain. The current theory can¿t fully explain the encoding used by players to register a board. The errors of intermediary players preserved fragments of the strategic situation, although they have committed a series of errors in the reconstruction of the positions. The encoding of chunks therefore includes more information than that predicted by current theories. Currently, research on perception, trial and decision is heavily concentrated on the idea of pattern recognition". Based on the results of this research, we explore a change of perspective. The idea of "pattern recognition" presupposes that the processing of relevant information is on "patterns" (or data) that exist independently of any interpretation. We propose that the theory suggests the vision of decision-making via the recognition of experience.

Autodeterminação local e desenvolvimento: uma análise da dinâmica social no município de São Roque de Minas

In this thesis, the basic research of Chase and Simon (1973) is questioned, and we seek new results by analyzing the errors of experts and beginners chess players in experiments to reproduce chess positions. Chess players with different levels of expertise participated in the study. The results were analyzed by a Brazilian grandmaster, and quantitative analysis was performed with the use of statistical methods data mining. The results challenge significantly, the current theories of expertise, memory and decision making in this area, because the present theory predicts piece on square encoding, in which players can recognize the strategic situation reproducing it faithfully, but commit several errors that the theory can¿t explain. The current theory can¿t fully explain the encoding used by players to register a board. The errors of intermediary players preserved fragments of the strategic situation, although they have committed a series of errors in the reconstruction of the positions. The encoding of chunks therefore includes more information than that predicted by current theories. Currently, research on perception, trial and decision is heavily concentrated on the idea of 'pattern recognition'. Based on the results of this research, we explore a change of perspective. The idea of 'pattern recognition' presupposes that the processing of relevant information is on 'patterns' (or data) that exist independently of any interpretation. We propose that the theory suggests the vision of decision-making via the recognition of experience.

Aplicação de modelos gráficos probabilísticos computacionais em economia

O objetivo deste trabalho é testar a aplicação de um modelo gráfico probabilístico, denominado genericamente de Redes Bayesianas, para desenvolver modelos computacionais que possam ser utilizados para auxiliar a compreensão de problemas e/ou na previsão de variáveis de natureza econômica. Com este propósito, escolheu-se um problema amplamente abordado na literatura e comparou-se os resultados teóricos e experimentais já consolidados com os obtidos utilizando a técnica proposta. Para tanto,foi construído um modelo para a classificação da tendência do "risco país" para o Brasil a partir de uma base de dados composta por variáveis macroeconômicas e financeiras. Como medida do risco adotou-se o EMBI+ (Emerging Markets Bond Index Plus), por ser um indicador amplamente utilizado pelo mercado.

Técnicas de computação natural para segmentação de imagens médicas

Image segmentation is one of the image processing problems that deserves special attention from the scientific community. This work studies unsupervised methods to clustering and pattern recognition applicable to medical image segmentation. Natural Computing based methods have shown very attractive in such tasks and are studied here as a way to verify it's applicability in medical image segmentation. This work treats to implement the following methods: GKA (Genetic K-means Algorithm), GFCMA (Genetic FCM Algorithm), PSOKA (PSO and K-means based Clustering Algorithm) and PSOFCM (PSO and FCM based Clustering Algorithm). Besides, as a way to evaluate the results given by the algorithms, clustering validity indexes are used as quantitative measure. Visual and qualitative evaluations are realized also, mainly using data given by the BrainWeb brain simulator as ground truth

Fabrication and characterization of chemical sensors made from nanostructured films of poly(o-ethoxyaniline) prepared with different doping acids

Chemical sensors made from nanostructured films of poly(o-ethoxyaniline) POEA and poly(sodium 4-styrene sulfonate) PSS are produced and used to detect and distinguish 4 chemicals in solution at 20 mM, including sucrose, NaCl, HCl, and caffeine. These substances are used in order to mimic the 4 basic tastes recognized by humans, namely sweet, salty, sour, and bitter, respectively. The sensors are produced by the deposition of POEA/PSS films at the top of interdigitated microelectrodes via the layer-by-layer technique, using POEA solutions containing different dopant acids. Besides the different characteristics of the POEA/PSS films investigated by UV-Vis and Raman spectroscopies, and by atomic force microscopy.. it is observed that their electrical response to the different chemicals in liquid media is very fast, in the order of seconds, systematical, reproducible, and extremely dependent on the type of acid used for film fabrication. The responses of the as-prepared sensors are reproducible and repetitive after many cycles of operation. Furthermore, the use of an "electronic tongue" composed by an array of these sensors and principal component analysis as pattern recognition tool allows one to reasonably distinguish test solutions according to their chemical composition. (c) 2007 Published by Elsevier B.V.

HOW FAR do WE GET USING MACHINE LEARNING BLACK-BOXES?

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Multidimensional polynomial powers of sigmoid (PPS) Wavelet neural networks

Wavelet functions have been used as the activation function in feedforward neural networks. An abundance of R&D has been produced on wavelet neural network area. Some successful algorithms and applications in wavelet neural network have been developed and reported in the literature. However, most of the aforementioned reports impose many restrictions in the classical backpropagation algorithm, such as low dimensionality, tensor product of wavelets, parameters initialization, and, in general, the output is one dimensional, etc. In order to remove some of these restrictions, a family of polynomial wavelets generated from powers of sigmoid functions is presented. We described how a multidimensional wavelet neural networks based on these functions can be constructed, trained and applied in pattern recognition tasks. As an example of application for the method proposed, it is studied the exclusive-or (XOR) problem.

Structural FRF acquisition via electric impedance measurement applied to damage location

Continuing development of new materials makes systems lighter and stronger permitting more complex systems to provide more functionality and flexibility that demands a more effective evaluation of their structural health. Smart material technology has become an area of increasing interest in this field. The combination of smart materials and artificial neural networks can be used as an excellent tool for pattern recognition, turning their application adequate for monitoring and fault classification of equipment and structures. In order to identify the fault, the neural network must be trained using a set of solutions to its corresponding forward Variational problem. After the training process, the net can successfully solve the inverse variational problem in the context of monitoring and fault detection because of their pattern recognition and interpolation capabilities. The use of structural frequency response function is a fundamental portion of structural dynamic analysis, and it can be extracted from measured electric impedance through the electromechanical interaction of a piezoceramic and a structure. In this paper we use the FRF obtained by a mathematical model (FEM) in order to generate the training data for the neural networks, and the identification of damage can be done by measuring electric impedance, since suitable data normalization correlates FRF and electrical impedance.

Otimização em comitês de classificadores: uma abordagem baseada em filtro para seleção de subconjuntos de atributos

Traditional applications of feature selection in areas such as data mining, machine learning and pattern recognition aim to improve the accuracy and to reduce the computational cost of the model. It is done through the removal of redundant, irrelevant or noisy data, finding a representative subset of data that reduces its dimensionality without loss of performance. With the development of research in ensemble of classifiers and the verification that this type of model has better performance than the individual models, if the base classifiers are diverse, comes a new field of application to the research of feature selection. In this new field, it is desired to find diverse subsets of features for the construction of base classifiers for the ensemble systems. This work proposes an approach that maximizes the diversity of the ensembles by selecting subsets of features using a model independent of the learning algorithm and with low computational cost. This is done using bio-inspired metaheuristics with evaluation filter-based criteria

Uma nova forma de calcular os centros dos Clusters em algoritmos de agrupamento tipo fuzzy c-means

Clustering data is a very important task in data mining, image processing and pattern recognition problems. One of the most popular clustering algorithms is the Fuzzy C-Means (FCM). This thesis proposes to implement a new way of calculating the cluster centers in the procedure of FCM algorithm which are called ckMeans, and in some variants of FCM, in particular, here we apply it for those variants that use other distances. The goal of this change is to reduce the number of iterations and processing time of these algorithms without affecting the quality of the partition, or even to improve the number of correct classifications in some cases. Also, we developed an algorithm based on ckMeans to manipulate interval data considering interval membership degrees. This algorithm allows the representation of data without converting interval data into punctual ones, as it happens to other extensions of FCM that deal with interval data. In order to validate the proposed methodologies it was made a comparison between a clustering for ckMeans, K-Means and FCM algorithms (since the algorithm proposed in this paper to calculate the centers is similar to the K-Means) considering three different distances. We used several known databases. In this case, the results of Interval ckMeans were compared with the results of other clustering algorithms when applied to an interval database with minimum and maximum temperature of the month for a given year, referring to 37 cities distributed across continents

Uma análise da aplicação do modelo de Rede Neural RePART em Comitês de classificadores

RePART (Reward/Punishment ART) is a neural model that constitutes a variation of the Fuzzy Artmap model. This network was proposed in order to minimize the inherent problems in the Artmap-based model, such as the proliferation of categories and misclassification. RePART makes use of additional mechanisms, such as an instance counting parameter, a reward/punishment process and a variable vigilance parameter. The instance counting parameter, for instance, aims to minimize the misclassification problem, which is a consequence of the sensitivity to the noises, frequently presents in Artmap-based models. On the other hand, the use of the variable vigilance parameter tries to smoouth out the category proliferation problem, which is inherent of Artmap-based models, decreasing the complexity of the net. RePART was originally proposed in order to minimize the aforementioned problems and it was shown to have better performance (higer accuracy and lower complexity) than Artmap-based models. This work proposes an investigation of the performance of the RePART model in classifier ensembles. Different sizes, learning strategies and structures will be used in this investigation. As a result of this investigation, it is aimed to define the main advantages and drawbacks of this model, when used as a component in classifier ensembles. This can provide a broader foundation for the use of RePART in other pattern recognition applications