Este trabalho avalia o desempenho do uso de árvores de decisão como função de ordenação para documentos e pessoas em ambientes empresariais. Para tanto, identificouse atributos relevantes das entidades a serem recuperadas a partir da análise de: (i) dinâmica de produção e consumo de informações em um ambiente empresarial; (ii) algoritmos existentes na literatura para a recuperação de documentos e pessoas; e (iii) conceitos utilizados em funções de ordenação para domínios genéricos. Montou-se um ambiente de avaliação, utilizando a coleção de referência CERC, para avaliar a aplicabilidade do algoritmo C4.5 na obtenção de funções de ordenação para o domínio empresarial. O uso do algoritmo C4.5 para a construção de funções de ordenação mostrou-se parcialmente efetivo. Para a tarefa de recuperação de documentos não trouxe resultados bons. Porém, constatou-se que é possível controlar a forma de construção da função de ordenação a fim de otimizar a precisão nas primeiras posições do ranking ou otimizar a média das precisões (MAP). Para a tarefa de recuperação de pessoas o algoritmo C4.5 obteve uma árvore de decisão que consegue resultados melhores que todas as outras funções de ordenação avaliadas. OMAP obtido pela árvore de decisão foi 0, 83, enquanto que a média do MAP das outras funções de ordenação foi de 0, 74. Percebeu-se que a árvore de decisão utilizada para representar a função de ordenação contribui para a compreensão da composição dos diversos atributos utilizados na caracterização dos documentos e pessoas. A partir da análise da árvore de decisão utilizada como função de ordenação para pessoas foi possível entender que uma pessoa é considerada especialista em algum tópico se ela aparecer em muitos documentos, aparecer muitas vezes nos documentos e os documentos onde aparece têm uma relevância alta para a consulta. / This work evaluates the performance of using decision trees as ranking functions for documents and people in enterprises. It was identified relevant attributes of the entities to be retrieved from the analysis of: (i) the production and consumption of information behavior in an enterprise, (ii) algorithms for documents and people retrieval at literature, and (iii) the concepts used in ranking functions for generic domains. It was set up an evaluation environment, using the CERC collection, to evaluate the applicability of the C4.5 algorithm to obtain a ranking function for the enterprise domain. The use of C4.5 algorithm for the construction of ranking function was proved to be partially effective. In the case of documents retrieval the C4.5 has not found good results. However, it was found that is possible to control the way of building the ranking function in order to optimize the precision in the first positions of the ranking or optimize the mean average precision (MAP). For the task of people retrieval the C4.5 algorithm developed a ranking function that obtain better results than all other ranking functions assessed. The value of MAP obtained by decision tree was 0, 83, while the average MAP of other ranking functions was 0, 74. The decision tree used to represent the ranking function contributes to understanding the attributes composition used in the characterization of documents and people. Through the analysis of the decision tree used as ranking function for people, we could realise that a person is considered expert in any topic if he/she appear in many documents, appear many times in same documents and documents where he/she appears have a high relevance to the query.

We study the problem of embedding high-dimensional visual data into low-dimensional vector representations. This is an important component in many computer vision applications involving nearest neighbor retrieval, as embedding techniques not only perform dimensionality reduction, but can also capture task-specific semantic similarities. In this thesis, we use deep neural networks to learn vector embeddings, and develop a gradient-based optimization framework that is capable of optimizing ranking-based retrieval performance metrics, such as the widely used Average Precision (AP) and Normalized Discounted Cumulative Gain (NDCG). Our framework is applied in three applications. First, we study Supervised Hashing, which is concerned with learning compact binary vector embeddings for fast retrieval, and propose two novel solutions. The first solution optimizes Mutual Information as a surrogate ranking objective, while the other directly optimizes AP and NDCG, based on the discovery of their closed-form expressions for discrete Hamming distances. These optimization problems are NP-hard, therefore we derive their continuous relaxations to enable gradient-based optimization with neural networks. Our solutions establish the state-of-the-art on several image retrieval benchmarks. Next, we learn deep neural networks to extract Local Feature Descriptors from image patches. Local features are used universally in low-level computer vision tasks that involve sparse feature matching, such as image registration and 3D reconstruction, and their matching is a nearest neighbor retrieval problem. We leverage our AP optimization technique to learn both binary and real-valued descriptors for local image patches. Compared to competing approaches, our solution eliminates complex heuristics, and performs more accurately in the tasks of patch verification, patch retrieval, and image matching. Lastly, we tackle Deep Metric Learning, the general problem of learning real-valued vector embeddings using deep neural networks. We propose a learning to rank solution through optimizing a novel quantization-based approximation of AP. For downstream tasks such as retrieval and clustering, we demonstrate promising results on standard benchmarks, especially in the few-shot learning scenario, where the number of labeled examples per class is limited.

"The goal of this thesis is to extend the theory and practice of matrix completion algorithms, and how they can be utilized, improved, and scaled up to handle large data sets. Matrix completion involves predicting missing entries in real-world data matrices using the modeling assumption that the fully observed matrix is low-rank. Low-rank matrices appear across a broad selection of domains, and such a modeling assumption is similar in spirit to Principal Component Analysis. Our focus is on large scale problems, where the matrices have millions of rows and columns. In this thesis we provide new analysis for the convergence rates of matrix completion techniques using convex nuclear norm relaxation. In addition, we validate these results on both synthetic data and data from two real-world domains (recommender systems and Internet tomography). The results we obtain show that with an empirical, data-inspired understanding of various parameters in the algorithm, this matrix completion problem can be solved more efficiently than some previous theory suggests, and therefore can be extended to much larger problems with greater ease. "

Covariance matrices are ubiquitous in computational sciences, typically describing the correlation of elements of large multivariate spatial data sets. For example, covari- ance matrices are employed in climate/weather modeling for the maximum likelihood estimation to improve prediction, as well as in computational ground-based astronomy to enhance the observed image quality by filtering out noise produced by the adap- tive optics instruments and atmospheric turbulence. The structure of these covariance matrices is dense, symmetric, positive-definite, and often data-sparse, therefore, hier- archically of low-rank. This thesis investigates the performance limit of dense matrix computations (e.g., Cholesky factorization) on covariance matrix problems as the number of unknowns grows, and in the context of the aforementioned applications. We employ recursive formulations of some of the basic linear algebra subroutines (BLAS) to accelerate the covariance matrix computation further, while reducing data traffic across the memory subsystems layers. However, dealing with large data sets (i.e., covariance matrices of billions in size) can rapidly become prohibitive in memory footprint and algorithmic complexity. Most importantly, this thesis investigates the tile low-rank data format (TLR), a new compressed data structure and layout, which is valuable in exploiting data sparsity by approximating the operator. The TLR com- pressed data structure allows approximating the original problem up to user-defined numerical accuracy. This comes at the expense of dealing with tasks with much lower arithmetic intensities than traditional dense computations. In fact, this thesis con- solidates the two trends of dense and data-sparse linear algebra for HPC. Not only does the thesis leverage recursive formulations for dense Cholesky-based matrix al- gorithms, but it also implements a novel TLR-Cholesky factorization using batched linear algebra operations to increase hardware occupancy and reduce the overhead of the API. Performance reported of the dense and TLR-Cholesky shows many-fold speedups against state-of-the-art implementations on various systems equipped with GPUs. Additionally, the TLR implementation gives the user flexibility to select the desired accuracy. This trade-off between performance and accuracy is, currently, a well-established leading trend in the convergence of the third and fourth paradigm, i.e., HPC and Big Data, when moving forward with exascale software roadmap.

As estatinas são medicamentos comumente prescritos para a prevenção da hiper-lipidemia. Além da redução do colesterol, tais medicamentos parecem estimular a osteogênese e suprimir a reabsorção óssea, o que poderia afetar a movimentação dentária induzida (MDI) e a recidiva ortodôntica. Assim, o objetivo deste estudo foi determinar se a atorvastatina (ATV) pode afetar a MDI, a recidiva e a osteoclastogênese, por meio da modulação da expressão das moléculas: - ligante do receptor ativador de NFκB (RANKL) e osteoprotegerina (OPG). Ainda foram analisados os potenciais efeitos adversos da ATV sobre a ossificação endocondral e sobre o turnover de ossos longos. No primeiro experimento, 36 ratos foram sujeitos a MDI durante 21 dias, quando o aparelho foi removido. Aos animais, foram administrados, diariamente, ATV ou solução salina (SAL), via gavagem. Após 7, 14 e 21 dias de administração de ATV / SAL, a recidiva dentária foi mensurada, e foram obtidos os cortes histológicos da maxila e fêmur, os quais foram submetidos às seguintes colorações: - H&E – para análise histomorfométrica; - fosfatase acida tartrato resistente (TRAP) – para contagem de osteoclastos e; - imunohistoquímica para RANKL e OPG. A atorvastatina resultou numa inibição da recidiva ortodôntica (p < 0,05), e numa transiente redução do número de osteoclastos (p < 0,05); havendo uma correlação positiva e significativa (p < 0,01) entre o estes dois fatores (número de osteoclastos e a taxa de recidiva). A administração de estatinas também aumentou significativamente a expressão de OPG (p < 0,01), mas não a de RANKL. Além disso, após 21 dias de administração de ATV, a espessura da cartilagem da placa de crescimento e da zona hipertrófica condrocítica foi significativamente aumentada. Já no segundo experimento, 24 ratos começaram a receber diariamente ATV ou solução salina (SAL), via gavagem. Duas semanas mais tarde, a MDI foi iniciada. O deslocamento do dente foi medido após 7, 14 e 21 dias, enquanto que os cortes histológicos da maxila e do fêmur foram obtidos após 14 e 21 dias de MDI; sendo então submetidos às colorações de H&E e TRAP, para avaliação histomorfométrica e contagem de osteoclastos. A administração de atorvastatina gerou um menor movimento dentário (p <0,05) e uma redução transitória do número de osteoclastos (p <0,05). No grupo SAL, após 14 dias de MDI, ocorreu um aumento no número de osteoclastos, assim reduzindo a taxa de volume ósseo, quando comparado com as maxilas controle (sem movimento dentário), deste mesmo grupo. Contudo, tal comportamento não foi observado no grupo ATV. Interessantemente, depois de 35 dias, a atorvastatina não afetou a remodelação óssea nas maxilas controle, nem a ossificação endocondral em fêmures. Logo, guardando as devidas limitações deste estudo pré-clínico, nossos resultados sugerem que a administração sistêmica de atorvastatina é capaz de minimizar a MDI e recidiva ortodôntica. No entanto, os seus efeitos celulares sobre a ossificação endocondral e remodelação óssea durante a MDI e recidiva, parecem ser limitados a um curto período de tempo, o que aparentemente necessita de investigações futuras. Finalmente, nossos resultados lançam luz sobre a superexpressão OPG induzida por estatinas, o que representa um alvo molecular para modular o metabolismo do ósseo e, assim minimizar a recidiva ortodôntica. / Statins are drugs commonly prescribed for prevention of hyper-lipidemia. In addition to the cholesterol-lowering, these medicines seem to enhance osteogenesis and suppress bone resorption, which could affect orthodontic tooth movement (OTM) and relapse. Therefore, the aim of this study was to determine whether atorvastatin (ATV) might affect the orthodontic relapse or tooth movement and osteoclastogenesis, through the modulation of the following molecules: receptor activator of nuclear κ B ligand (RANKL) and; - osteoprotegerin (OPG). Furthermore, we analyzed potential adverse effects of ATV on long bone turnover and endochondral ossification. In the first experiment, 36 rats were subjected to OTM for 21 days, when the appliance was removed. After, the animals were administered daily with ATV (15mg/Kg) or saline (SAL), via gavage (n=18, per group). Up to 7, 14 and 21 days of ATV/SAL administration the tooth relapse was measured while maxillary and femur histologic sections were obtained and prepared to: - H&E staining – used in histomorphometric analysis, tartrate resistant acid phosphatase histochemical staining (TRAP) – used to osteoclasts counting and, immunohistochemistry to RANKL and OPG. Atorvastatin resulted in a decreased tooth relapse (p<0.05), and a transient reduction of osteoclasts number (p<0.05). There was a positive and significant correlation (p<0.01) between these two parameters (osteoclasts number and relapse rate). The statin administration increased significantly the OPG (p<0.01), but not the RANKL expression. Furthermore, after 21 days of ATV administration, the thickness of growth plate cartilage and chondrocytic hypertrophic zone was enhanced. In the second experiment, 24 rats started to be administered daily with ATV or SAL, via gavage. Two weeks later, the OTM started. The tooth displacement was measured after 7, 14, and 21 days, while the maxillary and femur histologic sections were obtained only after 14 and 21 days of OTM. At these times, the sections were prepared to H&E and TRAP, intending to perform the histomorphometric analysis and osteoclasts count. Atorvastatin administration promoted a decreased tooth movement (p<0.05), and a transient reduction of osteoclasts number (p<0.05). In the SAL group, after 14 days of OTM, the increased number of osteoclasts was associated to a reduced bone volume rate, when compared to its control maxillae. However, this trend was not obvious in ATV group. Interestingly, after 35 days, statins did not affect the bone turnover and endochondral ossification. The big picture of our study suggests that systemic administration of statins is able to minimize OTM and relapse. However, its cellular effects on endochondral ossification and bone turnover during OTM or relapse seem to be limited to a short period, apparently requiring further investigations. Finally, our results shed light on OPG overexpression induced by statins, which represents a molecular target modulating maxillary bone metabolism thus inhibiting orthodontic relapse.

Neste trabalho estudaremos alguns resultados propostos por Benoit Kloe kner [Kl2] em sua tese de doutorado. Apresentamos prin ipalmente a prova da não-existência de compactificações diferenciáveis de Hadamard em espaços simétri os de tipo não- ompa to de posto k ≥ 2. / In this dissertation we will study some results proposed by Benoit Kloe kner [Kl2] in his do toral thesis. We mainly present the proof of non-existen e of diferentiable Hadamard compactific ations in symmetric spaces of non compact type of rank ≥ 2.

Notre travail ici concerne certaines pistes pour des constructions nouveaux groupes, et en particulier de contre-exemples à la conjecture de Cherlin-Zilber. On parvient à trouver une réponse pour la stabilité de groupes CSA existentiellement clos. On exhibe un mot de groupe en deux variables qui a la propriété d’indépendance par rapport à la classe de groupes hyperboliques sans torsion. On en déduit que l’équation correspondante donne la propriété d’indépendance des groupes CSA existentiellement clos, ce qui en particulier implique leur instabilité. En outre, on prouve que les équations, et en particulier les ensembles définissables sans quantificateurs, définissent des ensembles stables dans les boules bornées des produits libres de groupes, en utilisant la version finie du théorème de Ramsey. Enfin, on introduit certains groupes construits comme tours particulières de produits libres et d’extensions HNN. / Our work here relates to certain routes for the construction of new groups, and in particular, of counter-examples to the Cherlin-Zilber conjecture. We managed to find an answer for the stability of existentially closed CSA- groups. We build a group word in two variables that has the independence property relatively to the class of torsion-free hyperbolic groups. We deduced that the corresponding equation gives the independence property of existentially closed CSA groups which in turn implies their instability. Moreover, we demonstrate that group words, and in particular quantifierfree definable sets, define stable sets in bounded balls of free products of groups using a finite version of Ramsey’s theorem. Finally, we introduce certain groups constructed as special towers of free products and HNN-extensions.

This report deals with the managing of the waste of a company, Ericsson- site of Borås, in an analytical context. Based on sustainability (concept and aspects), they are interested to have a method to check their waste management capability whether they are in right direction. Among all studied methods, Analytic Hierarchy Process (AHP) is utilized. This method works based on a mathematical algorithm starting by making a hierarchy, continuing with pairwise comparisons between correspondent items, then doing calculations and finally checking and reviewing to be certain of the correctness of the whole process by an eligible team of decision makers. In spite of some critiques that scientifically are accepted, it remains reliable for the purpose.The method is applied to some instances of waste items, wood boxes and pallets and hard plastics, in Ericsson. Then, two controversial issues of the selected method, consistency and rank reversal, are investigated and discussed on the mentioned waste items. Application of the method for their future use is foreseen thereafter.

Data from rating scale assessments have rank-invariant properties only, which means that the data represent an ordering, but lack of standardized magnitude, inter-categorical distances, and linearity. Even though the judgments often are coded by natural numbers they are not really metric. The aim of this thesis is to further develop the nonparametric rank-based Svensson methods for paired ordinal data that are based on the rank-invariant properties only. The thesis consists of five papers. In Paper I the asymptotic properties of the measure of systematic disagreement in paired ordinal data, the Relative Position (RP), and the difference in RP between groups were studied. Based on the findings of asymptotic normality, two tests for analyses of change within group and between groups were proposed. In Paper II the asymptotic properties of rank-based measures, e.g. the Svensson’s measures of systematic disagreement and of additional individual variability were discussed, and a numerical method for approximation was suggested. In Paper III the asymptotic properties of the measures for paired ordinal data, discussed in Paper II, were verified by simulations. Furthermore, the Spearman rank-order correlation coefficient (rs) and the Svensson’s augmented rank-order agreement coefficient (ra) were compared. By demonstrating how they differ and why they differ, it is emphasized that they measure different things. In Paper IV the proposed test in Paper I for comparing two groups of systematic changes in paired ordinal data was compared with other nonparametric tests for group changes, both regarding different approaches of categorising changes. The simulation reveals that the proposed test works better for small and unbalanced samples. Paper V demonstrates that rank invariant approaches can also be used in analysis of ordinal data from multi-item scales, which is an appealing and appropriate alternative to calculating sum scores.

Ab-initio methods for calculating electronic structure represent an important field of material physics. The aim of this theses - within the project focused on developing the new method for calculating electronic states in non-periodic structures based on density functional theory, pseudopotentials, and finite elements methods - is to convert Kohn-Sham equations into the form suitable for discretisation, to suggest apropriate method for solving generalized eigenproblem resulting from this discretisation and to implement an eigenvalue solver (or to modify existing one). The thesis describes a procedure for converting the many-particle Schrödinger equation into generalized rank-k-update eigenvalue problem and discusses various methods for its solution. Eigensolver Blzpack, which makes use of the block Lanczos method, has been modified, integrated into the Sfepy framework (a tool for the finite element method calculation) and resulting code has been successfully tested.