Este trabalho apresenta uma solução geral para a detecção de alinhamentos de dados em conjuntos multidimensionais não ordenados e ruidosos. Nesta abordagem, o tipo requerido de alinhamento de dados pode ser uma forma geométrica (e.g., linha reta, plano, círculo, esfera, seção cônica, entre outras) ou qualquer estrutura, com dimensionalidade arbitrária, que possa ser caracterizada por um subespaço linear. A detecção é realizada por meio de um procedimento composto por três etapas. Na etapa de inicialização, um espaço de parâmetros com p (n − p) dimensões é definido de modo que cada ponto neste espaço represente uma instância do alinhamento requerido, descrito por um subespaço p-dimensional em um domínio n-dimensional. Em seguida, uma grade de acumuladores é criada como sendo a representação discreta do espaço de parâmetros. Na segunda etapa do procedimento, cada elemento no conjunto de dados de entrada (também um subespaço no domínio n-dimensional) é mapeado para o espaço de parâmetros como os pontos (no espaço de parâmetros) representando os subespaços requeridos que contém ou que estão contidos no elemento de entrada. À medida que os elementos de entrada são mapeados, as células do acumulador relacionadas com o mapeamento são incrementadas pelo valor de importância do elemento mapeado. A etapa final do procedimento recupera os subespaços p-dimensionais que melhor se ajustam aos dados de entrada como sendo os máximos locais na grade de acumuladores. A parametrização proposta é independente das propriedades geométricas dos alinhamentos a serem detectados. Além disso, o procedimento de mapeamento é independente do tipo de dado de entrada e é capaz de se adaptar a elementos com dimensionalidades arbitrárias. Essas características permitem a utilização da técnica (sem a necessidade de modificações) como uma ferramenta para a detecção de padrões em uma grande quantidade de aplicações. Por conta de sua natureza geral, otimizações desenvolvidas para a abordagem proposta beneficiam, de forma imediata, todos os casos de detecção. Neste trabalho eu demonstro uma implementação em software da técnica proposta e mostro que ela pode ser aplicada tanto em casos simples de detecção, quanto na detecção concorrente de tipos diferentes de alinhamentos, com diferentes interpretações geométricas e em conjuntos de dados compostos por vários tipos de elementos. Esta dissertação também apresenta uma extensão do esquema de detecção para dados de entrada com distribuição Gaussiana de incerteza. A extensão proposta produz distribuições de valores mais suaves na grade de acumuladores e faz com que a técnica fique mais robusta à detecção de subespaços espúrios. / This dissertation presents a generalized closed-form framework for detecting data alignments in large unordered noisy multidimensional datasets. In this approach, the intended type of data alignment may be a geometric shape (e.g., straight line, plane, circle, sphere, conic section, among others) or any other structure, with arbitrary dimensionality that can be characterized by a linear subspace. The detection is performed using a three-step process. In the initialization, a p (n − p)-dimensional parameter space is defined in such a way that each point in this space represents an instance of the intended alignment described by a p-dimensional subspace in some n-dimensional domain. In turn, an accumulator array is created as the discrete representation of the parameter space. In the second step each input entry (also a subspace in the n-dimensional domain) is mapped to the parameter space as the set of points representing the intended p-dimensional subspaces that contain or are contained by the entry. As the input entries are mapped, the bins of the accumulator related to such a mapping are incremented by the importance value of the entry. The subsequent and final step retrieves the p-dimensional subspaces that best fit input data as the local maxima in the accumulator array. The proposed parameterization is independent of the geometric properties of the alignments to be detected. Also, the mapping procedure is independent of the type of input data and automatically adapts to entries of arbitrary dimensionality. This allows application of the proposed approach (without changes) in a broad range of applications as a pattern detection tool. Given its general nature, optimizations developed for the proposed framework immediately benefit all the detection cases. I demonstrate a software implementation of the proposed technique and show that it can be applied in simple detection cases as well as in concurrent detection of multiple kinds of alignments with different geometric interpretations, in datasets containing multiple types of data. This dissertation also presents an extension of the general detection scheme to data with Gaussian-distributed uncertainty. The proposed extension produces smoother distributions of values in the accumulator array and makes the framework more robust to the detection of spurious subspaces.

Submitted by MARCIA ROVADOSCHI (marciar@unifra.br) on 2018-08-20T12:55:10Z No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Tese_MarianaZancanTonel.pdf: 6177598 bytes, checksum: e830fe652c79ad380053f80cae3e2382 (MD5) / Made available in DSpace on 2018-08-20T12:55:10Z (GMT). No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Tese_MarianaZancanTonel.pdf: 6177598 bytes, checksum: e830fe652c79ad380053f80cae3e2382 (MD5) Previous issue date: 2017-08-25 / Graphene is a nanomaterial that has several applications including drug delivery and pollutant remover. However, graphene is hydrophobic, which makes it difficult to apply in biological media, and one of the alternatives is through functionalization. At the same time, aromatic molecules are precursors of several essential compounds for life, and are also important in industry. However, some of them are pollutants that can severely affect people and the environment, so developing effective removal methods is extremely interesting. Some molecules of biological interest like dopamine and serotonin have an aromatic part, recent studies show the use of these molecules as mediators for specific and selective drug delivery. Therefore, the study of the interaction of these molecules as pure and modified graphene is of great importance in the biomedical area as well as for diagnostic and treatment studies. In this work, we perform a theoretical study through the density functional theory; initially we analyze the changes caused by the effect of the concentration of the functional groups -COOH, -COH, -OH, -O- or -NH2 on graphene. The results show that in all cases it is possible to modulate the electronic properties depending on the number and location of the groups. Subsequently, we analyzed the graphene pure and functionalized with a group -COOH, -COH, -OH, -O- or -NH2, with the molecules of biological interest as dopamine, serotonin and the pollutants: benzene, aniline, benzoic acid and phenol. The results show that all interactions occur under a physical adsorption regime, there are no changes in the original geometric structures of the molecules after adsorption, it may be of interest to create possible routes as mediators for the delivery of drugs and to assist in the treatment of various diseases or in a system of removal of pollutants based on pristine and modified graphene. Finally, we developed a method to parameterize the benzene-benzene and benzene-graphene force field through the data obtained from the ab initio calculations, the results obtained agree with studies described in the literature. Thus, this work presents the understanding through the simulation of the biomedical part for drug delivery systems, and removal of pollutants, in addition to the parameterization to be used in simulations of the biological environment which may aid in the development of future experimental studies. / O grafeno é um nanomaterial que possui diversas aplicações tecnológicas entre elas como carreador de fármacos e removedor de poluentes. No entanto, o grafeno é hidrofóbico, o que dificulta sua aplicação em meios biológicos, sendo que uma das alternativas para controlar este problema é através de funcionalizações. Paralelamente, moléculas aromáticas são precursoras de diversos compostos imprescindíveis para a vida, e também são importantes na indústria. No entanto, algumas delas são poluentes que podem afetar severamente as pessoas e o meio ambiente. Assim, o desenvolvimento de métodos efetivos para remoção desses é extremante importante. Algumas moléculas de interesse biológico como a dopamina e serotonina apresentam parte aromática. Estudos recentes têm mostrado a utilização dessas moléculas como mediadoras para entrega de fármacos específica e seletiva. Portanto, o estudo dessas moléculas visando este tipo de aplicação é de grande importância na área biomédica. Realizamos um estudo teórico através da Teoria do Funcional da Densidade. Inicialmente, analisamos as alterações causadas pelo efeito da concentração de grupos funcionais -COOH, -COH, -OH, -O- ou -NH2 no grafeno. Os resultados mostram que em todos os casos é possível modular as propriedades eletrônicas dependendo do número e do local dos grupos. Posteriormente, analisamos o grafeno puro e funcionalizado com um grupo –COOH, -COH, -OH, –O- ou -NH2, com as moléculas de interesse biológico dopamina, serotonina e os poluentes benzeno, anilina, ácido benzóico e fenol. Os resultados mostram que todas as interações ocorrem sob um regime de adsorção física, sem alterações nas estruturas geométricas originais das moléculas após adsorção. Esses resultados são de interesse para criar possíveis rotas como mediadores para a entrega de fármacos e auxiliar no tratamento de várias doenças ou ainda em um sistema de remoção de poluentes baseados no grafeno puro e modificado. E, por fim, desenvolvemos um novo método para parametrizar campo de força do benzeno-benzeno e benzeno-grafeno através dos dados obtidos dos cálculos ab initio. Os resultados obtidos concordam com estudos descritos na literatura. Dessa forma, este trabalho apresenta a compreensão através da simulação da parte biomédica para sistemas de drug delivery, e remoção de poluentes, além da parametrização para ser usado em simulações do meio biológico o que poderá auxiliar no desenvolvimento de estudos experimentais futuros.

[pt] Fluidos são extremamente comuns em nosso mundo e têm papel central em muitos fenômenos naturais. A compreensão de seu comportamento tem importância fundamental em uma vasta gama de aplicações e diversas áreas de pesquisa, da análise de fluxo sanguíneo até o transporte de petróleo, da exploração do fluxo de um rio até a previsão de maremotos, tempestades e furacões. Na simulação de fluidos, a abordagem conhecida como Euleriana é capaz de gerar resultados bastante corretos e precisos, mas as computações envolvidas podem se tornar excessivamente custosas quando há a necessidade de tratar fronteiras curvas e obstáculos com formas complexas. Este trabalho aborda esse problema e apresenta uma técnica Euleriana rápida e direta para simular o escoamento de fluidos em grades estruturadas parametrizadas tridimensionais. O principal objetivo do método é tratar de forma correta e eficiente as interações de fluidos com fronteiras curvas, incluindo paredes externas e obstáculos internos. Para isso, são utilizadas matrizes Jacobianas por célula para relacionar as derivadas de campos escalares e vetoriais nos espaços do mundo e paramétrico, o que permite a resolução das equações de Navier-Stokes diretamente no segundo, onde a discretização do domínio torna-se simplesmente uma grade uniforme. O trabalho parte de um simulador baseado em grades regulares e descreve como adaptá-lo com a aplicação das matrizes Jacobianas em cada passo, incluindo a resolução de equações de Poisson e dos sistemas lineares esparsos associados, utilizando tanto iterações de Jacobi quanto o método do Gradiente Biconjugado Estabilizado. A técnica é implementada na linguagem de programação CUDA e procura explorar ao máximo a arquitetura massivamente paralela das placas gráficas atuais. / [en] Fluids are extremely common in our world and play a central role in many natural phenomena. Understanding their behavior is of great importance to a broad range of applications and several areas of research, from blood flow analysis to oil transportation, from the exploitation of river flows to the prediction of tidal waves, storms and hurricanes. When simulating fluids, the so-called Eulerian approach can generate quite correct and precise results, but the computations involved can become excessively expensive when curved boundaries and obstacles with complex shapes need to be taken into account. This work addresses this problem and presents a fast and straightforward Eulerian technique to simulate fluid flows in three-dimensional parameterized structured grids. The method s primary design goal is the correct and efficient handling of fluid interactions with curved boundary walls and internal obstacles. This is accomplished by the use of per-cell Jacobian matrices to relate field derivatives in the world and parameter spaces, which allows the Navier-Stokes equations to be solved directly in the latter, where the domain discretization becomes a simple uniform grid. The work builds on a regular-grid-based simulator and describes how to apply Jacobian matrices to each step, including the solution of Poisson equations and the related sparse linear systems using both Jacobi iterations and a Biconjugate Gradient Stabilized solver. The technique is implemented efficiently in the CUDA programming language and strives to take full advantage of the massively parallel architecture of today s graphics cards.

En régions méditerranéennes, la gestion de l'eau à partir d'outils d'aide à la décision requiert la connaissance des échanges d'énergie et de masse entre la surface et l'atmosphère, dont l'évapotranspiration, qui représente la composante majeure du cycle hydrologique. Les avancées récentes, en termes de modélisation des Transferts Sol-Végétation-Atmosphère (TSVA) pour des couverts homogènes et d'assimilation des données de télédétection, principalement à l'échelle subrégionale, permettent d'envisager le passage à l'échelle régionale pour des couverts complexes de type cultures en rang. L'objectif de ce travail est de développer une modélisation versatile et de la caler par télédétection à l'échelle régionale sur un bassin versant viticole. Les investigations sont menées sur le bassin versant de la Peyne, dans le cadre de l'ORE OMERE. Une analyse bibliographique a permis de sélectionner un modèle TSVA à vocation régionale avec un nombre réduit de paramètres. L'implémentation de ce modèle est motivée par des objectifs de versatilité mais aussi d'inclusion dans une plateforme de simulation. Parallèlement, l'évapotranspiration a été spatialisée à l'échelle régionale par synergie des données télédétectées infrarouge thermique ASTER et Landsat. Les chroniques d'évapotranspiration obtenues sont ensuite utilisées pour caler le modèle TSVA. / In Mediterranean regions, decision making tools for water management require knowledge of water and mass exchanges between land surface and atmosphere, where evapotranspiration is the main component of the hydrological cycle. Recent advances, in terms of modeling and remote sensing, mainly at the subregional scale for homogeneous canopies, allow foreseeing the regional extent for complex landscapes such as row crops. This work aims to propose and calibrate a versatile modeling at the regional scale over a vineyard watershed, the calibration relying on remote sensing. A literature review allows selecting a SVAT model with a regional scope and a limited number of parameters. Model implementation is motivated by versatility and further inclusion into a simulation platform. Then, evapotranspiration is spatialized synergistically by using thermal infrared data from ASTER and Landsat remote sensors. Next, the time series obtained for evapotranspiration are used for calibrating the selected SVAT model. These investigations are conducted over the Peyne watershed, within the framework of the OMERE Observatory for environmental research.

The main aim of this thesis is a development of ILF curves that can be used in the insurance industry when pricing general third party liability on the Czech market. Based on available data there are first estimated size of loss distribution functions used for following generating process. From generated data the increased limit factors are estimated and with a usage of Riebesell's parameterization ILF curves are derived. A substantial part of the thesis is a compilation of literature and the expansion of the statistical approach for estimating fair ILFs based on these data. Besides, the basis for the curves derivation are chapters describing basic theoretical knowledge in the field of reinsurance - in particular, the description of the basic types of reinsurance contracts, as well as the most common methods of a pricing. There is the whole mechanism of curves derivation described; their own use is then demonstrated with the example based on pseudoreal data.

Les avantages des véhicules autonomes sont formidables, mais le chemin vers une vraie autonomie sera long et semé d’incertitudes. La recherche de ces dernières années s’est basée sur des systèmes multi-capteurs capables de percevoir l’environnement dans lequel le véhicule est conduit. Ces systèmes deviennent plus complexes quand on contrôle le véhicule autonome, différents systèmes de contrôle sont activés dépendant de la décision du système multi-capteurs. Chacun de ces systèmes suit des critères de performance et de stabilité lors de leur conception. Cependant, ils doivent fonctionner ensemble, garantissant une stabilité et étant capable de se charger des changements dynamiques, structuraux et environnementaux. Cette thèse explore la paramétrisation Youla-Kucera (YK) dans des systèmes dynamiques comme les voitures, en insistant sur la stabilité quand la dynamique change, ou que le trafic impose une reconfiguration du contrôleur. Concentrons-nous sur l’obtention de résultats de simulation et expérimentaux en relation avec le "Cooperative Adaptive Cruise Control" (CACC), dans le but, non pas d’utiliser, ici, pour la première fois la paramétrisation YK dans le domaine des systèmes de transport intelligents (STI), mais d’améliorer l’état de l’art en CACC aussi. Des résultats de reconfiguration stable de contrôleurs sont donnés quand la communication avec le véhicule précédent n’est plus disponible, en cas de manœuvre d’entrées/sorties ou lorsqu’ils sont entourés de véhicules aux dynamiques différentes. Ceci démontrant l’adaptabilité, la stabilité et l’implémentation réelle de la paramétrisation YK comme structure générale de contrôle pour les véhicules autonomes. / Benefits of autonomous vehicles are genuinely exciting, but the route to true autonomy in transportation will likely be long and full of uncertainty. Research on the last years is on the development of multi-sensor systems able to perceive the environment in which the vehicle is driving in. These systems increase complexity when controlling an autonomous vehicle, as different control systems are activated depending on the multi-sensor decision system. Each of these systems follows performance and stability criteria for its design, but they all must work together, providing stability guarantees and being able to handle dynamics, structural and environmental changes. This thesis explores the Youla-Kucera (YK) parameterization in dynamics systems such as vehicles, with special emphasis on stability when some dynamics change or the traffic situation demands controller reconfiguration. Focus is in obtaining simulation and experimental results related to Cooperative Adaptive Cruise Control (CACC), with the aim not only of using for the very first time YK parameterization in the Intelligent Transportation Systems (ITS) domain, but improving CACC state-of-the art. Stable controller reconfiguration results are given when non-available communication link with the preceding vehicle, cut-in/out maneuvers or surrounding vehicles with different dynamics, proving adapability, stability and possible real implementation of the YK parameterization as general control framework for autonomous vehicles.

Tato dizertační práce se zaměřuje na použití paralelních souřadnic pro parametrizaci čar a bodů. Paralelní souřadný systém má souřadnicové osy vzájemně rovnoběžné. Bod ve dvourozměrném prostoru je v paralelních souřadnicích zobrazen jako přímka a přímka jako bod. Toho je možné využít pro Houghovu transformaci - metodu, při které body zájmu hlasují v prostoru parametrů pro danou hypotézu. Parametrizace pomocí paralelních souřadnic vyžaduje pouze rasterizaci úseček, a proto je velmi rychlá a přesná. V práci je tato parameterizace demonstrována na detekci maticových kódů a úběžníků.

Les eaux souterraines constituent une réserve d’eau potable non négligeable, leur alimentation se fait en majeure partie par les précipitations, appelée recharge des nappes. Du fait de leur grande importance, la compréhension du fonctionnement de ces ressources en eau est plus que jamais indispensable. Celle-ci passe par l’élaboration de modèles mathématiques. Ces outils nous offrent une meilleure appréhension et une bonne prévision des phénomènes physiques. Les systèmes hydrogéologiques sont généralement très complexes et caractérisés par des dynamiques hydriques très variables dans le temps et dans l’espace. Cette complexité a attiré l’attention de nombreux hydrogéologues et un grand nombre de modèles très sophistiqués a été développé afin de décrire ces systèmes avec précision. Cependant, la prise en compte de la recharge de ces réservoirs reste toujours un défi dans la modélisation hydrogéologique. En effet, le plus souvent, les modèles hydrogéologiques simulent l’écoulement dans la nappe tout en considérant la recharge comme une constante sur le domaine et indépendante du système. De plus, elle est souvent calculée de façon simplifiée. Or, la recharge traduisant la quantité des précipitations atteignant les nappes est une composante hydrologique complexe et variable car elle interagit avec les nappes et dépend des conditions climatiques, du couvert végétal et du transfert de l’eau dans le sol. Ce présent travail vise à intégrer cette recharge variable et complexe aux modèles hydrogéologiques. À cet effet, un modèle couplé a été développé. Une première partie de ce modèle permet de calculer la recharge des nappes en modélisant les interactions précipitations-sol et l’hydrodynamique dans le sol. Cette modélisation a été effectuée en utilisant des modèles conceptuels simples basés sur des lois empiriques (Gardénia, Nash) et des modèles physiques résolvant l’équation de Richards. La recharge ainsi calculée est intégrée à la deuxième partie du modèle simulant l’hydrodynamique dans les nappes décrite par l’équation de diffusivité. Des méthodes numériques précises et robustes ont été utilisées pour résoudre les équations du modèle mathématique : les éléments finis non conformes ont été utilisés pour résoudre l’équation de diffusivité et l’équation de Richards est résolue sous sa forme mixte par une méthode itérative en temps. En somme, ce modèle couplé permet de décrire les variations de niveaux de nappe à partir des données météorologiques connaissant les paramètres caractéristiques de cet aquifère. [...] / Groundwater is the main available water resource for many countries; they are mainly replenished by water from precipitation, called groundwater recharge. Due to its great importance, management of groundwater resources is more essential than ever, and is achieved through mathematical models which offer us a better understanding of physical phenomena as well as their prediction. Hydrogeological systems are generally complex thus characterized by a highly variable dynamic over time and space. These complexities have attracted the attention of many hydrogeologists and many sophisticated models that can handle these issues and describe these systems accurately were developed. Unfortunately, modeling groundwater recharge is still a challenge in groundwater resource management. Generally, groundwater models are used to simulate aquifers flow without a good estimation of recharge and its spatial-temporal distribution. As groundwater recharge rates show spatial-temporal variability due to climatic conditions, land use, and hydrogeological heterogeneity, these methods have limitations in dealing with these characteristics. To overcome these limitations, a coupled model which simulates flow in the unsaturated zone and recharge as well as groundwater flow was developed. The flow in the unsaturated zone is solved either with resolution of Richards equation or with empirical models while the diffusivity equation governs flow in the saturated zone. Robust numerical methods were used to solve these equations: we apply non-conforming finite element to solve the diffusivity equation and we used an accurate and efficient method for solving the Richards equation. [...]

Built-up areas represent an artifiial impait to natural environment with large spatial variability and speiifi meihaniit radiationt thermal and ihemiial properties. Despite of inireasing horizontal resolution of numeriial weather prediition modelst the impait of loial built-up area on mesosynoptiv weather phenomena is still not well resolved. Therefore it is neiessary to use some of urban environment modelst whiih were designed to parameterize speiifi urban prosiessest not expliiitly resolved inside the grid box. In the thesis main urban iharaiteristiis are explained (impait on the struiture of boundary layert radiation and heat balanie of urban environment or urban heat island)t basii priniiples of urbanization appliiation in the numeriial weather model are desiribedt as well as different urban parameterizations available in numeriial model WRFe (Weather Reseaih and Feoreiasting). Number of validation experiments were performed for summer and winter episode in non-hydrostatii mode at 3t3 km resolutiont where different urban parametrizationst antropogenii heat adjustment and impait of mosaii land-use were tested. April 2018 Prague weather foreiast was verifiated in ionsideration of urban heat island.

L’objectif principal de cette thèse est de souligner que le compromis biais-variance n’est pas toujours vrai (p. ex. dans les réseaux neuronaux). Nous plaidons pour que ce manque d’universalité soit reconnu dans les manuels scolaires et enseigné dans les cours d’introduction qui couvrent le compromis. Nous passons d’abord en revue l’historique du compromis entre les biais et les variances, sa prévalence dans les manuels scolaires et certaines des principales affirmations faites au sujet du compromis entre les biais et les variances. Au moyen d’expériences et d’analyses approfondies, nous montrons qu’il n’y a pas de compromis entre la variance et le biais dans les réseaux de neurones lorsque la largeur du réseau augmente. Nos conclusions semblent contredire les affirmations de l’oeuvre historique de Geman et al. (1992). Motivés par cette contradiction, nous revisitons les mesures expérimentales dans Geman et al. (1992). Nous discutons du fait qu’il n’y a jamais eu de preuves solides d’un compromis dans les réseaux neuronaux lorsque le nombre de paramètres variait. Nous observons un phénomène similaire au-delà de l’apprentissage supervisé, avec un ensemble d’expériences d’apprentissage de renforcement profond. Nous soutenons que les révisions des manuels et des cours magistraux ont pour but de transmettre cette compréhension moderne nuancée de l’arbitrage entre les biais et les variances. / The main goal of this thesis is to point out that the bias-variance tradeoff is not always true (e.g. in neural networks). We advocate for this lack of universality to be acknowledged in textbooks and taught in introductory courses that cover the tradeoff. We first review the history of the bias-variance tradeoff, its prevalence in textbooks, and some of the main claims made about the bias-variance tradeoff. Through extensive experiments and analysis, we show a lack of a bias-variance tradeoff in neural networks when increasing network width. Our findings seem to contradict the claims of the landmark work by Geman et al. (1992). Motivated by this contradiction, we revisit the experimental measurements in Geman et al. (1992). We discuss that there was never strong evidence for a tradeoff in neural networks when varying the number of parameters. We observe a similar phenomenon beyond supervised learning, with a set of deep reinforcement learning experiments. We argue that textbook and lecture revisions are in order to convey this nuanced modern understanding of the bias-variance tradeoff.