Parameterized complexity is a part of computer science dealing with the computational complexity of problems measured not only by the length of their input but also some parameter of the input. Nei- ghborhood diversity is a recently introduced parameter describing a certain structure of a graph. is parameter is aractive for resear especially because some problems whi are hard with respect to other parameters that are incomparable with neighborhood diversity become fixed-parameter tractable with respect to neighborhood diversity. In this thesis we show fixed-parameter tractability for three problems that are hard with respect to treewidth. is constitutes the main part of this thesis and it is our original work. Next it contains an overview of other interesting problems and also a survey of the state of the art in the area of parameters for sparse and dense graphs. 1

We present Chandra/ACIS-S subarray observations of the quiescent neutron star (NS) low-mass X-ray binaries X7 and X5 in the globular cluster 47 Tuc. The large reduction in photon pile-up compared to previous deep exposures enables a substantial improvement in the spectroscopic determination of the NS radius and mass of these NSs. Modeling the thermal emission from the NS surface with a non-magnetized hydrogen atmosphere and accounting for numerous sources of uncertainties, we obtain for the NS in X7 a radius of R = 11.1(-0.7)(+0.8) km for an assumed stellar mass of M = 1.4 M-circle dot (68% confidence level). We argue, based on astrophysical grounds, that the presence of a He atmosphere is unlikely for this source. Due to the excision of data affected by eclipses and variable absorption, the quiescent low-mass X-ray binary X5 provides less stringent constraints, leading to a radius of R = 9.6(-1.1)(+0.9) km, assuming a hydrogen atmosphere and a mass of M. =. 1.4 Me. When combined with all existing spectroscopic radius measurements from other quiescent low-mass X-ray binaries and Type I X-ray bursts, these measurements strongly favor radii in the 9.9-11.2 km range for a similar to 1.5 M-circle dot NS and point to a dense matter equation of state that is somewhat softer than the nucleonic ones that are consistent with laboratory experiments at low densities.

This study examined how attentional sub-processes contribute to binge-eating. Dense-array EEG and a version of the canonical attentional blink task were used to ascertain the neural correlates underlying the attentional sub-processes that comprise the Posner model of attention (alerting, orienting, and executive control) and how attentional activation differs for binge-eaters vs. non-binge eaters. Furthermore, we examined a number of the event-related potentials (ERP), including P2 activation, which has been linked with orientating of attention, and N2 activation which has been linked with attentional conflict. We found decreased P2 activation for binge-eaters, in the negative condition, for incorrect target 2 (T2) detection trials. We also found more N2 activation for binge-eaters than non-binge eaters, in negative trials when T2 was not detected. This pattern of results suggest that binge-eaters showed deficiencies in allocating attention to stimuli that followed negative images; this attention deficiency may be a key factor for binge-eating behavior.

<div>For decades nuclear fusion space propulsion has been studied but due to technological set backs for self-sustaining fusion, it has been repeatedly abandoned in favor of more near-term or present day solutions. While these present day solutions of chemical and electric propulsion have been able to accomplish their missions, as the human race looks to explore Mars, a near term solution utilizing nuclear fusion propulsion must be sought as the fusion powered thruster case currently does not meet the minimum 0.2 thrust-to-weight ratio requirement. The current work seeks to investigate the use of a ssion powered magnetic fusion thruster for a manned Mars mission with an emphasis on creating a very near-term propulsion system. This will be accomplished by utilizing present day readily available technology and adapting methods of nuclear electric and nuclear fusion propulsion to build this ssion assisted propulsion system. Near term solutions have been demonstrated utilizing both DT and D-He3 fuels for a ssion powered and ssion assisted Dense Plasma Focus fusion device capable of achieving thrust-to-weight ratios greater than 0.2 for V's of 20 km/s. The Dense Plasma Focus can achieve thrust-to-weight ratios of 0.34 and 0.4 for ssion assisted and ssion powered cases, respectively, however, the Gasdynamic Mirror device proved to be an infeasible design as a ssion powered thruster due to the increased weight of a ssion reactor.</div>

Esquemas de diagnóstico auxiliado por computador (CAD, do inglês \"computer aided diagnosis\") têm sido desenvolvidos com o objetivo de fornecer subsídios para a detecção precoce do câncer de mama. Nesses esquemas, técnicas de processamento de imagens são utilizadas para indicar a existência de estruturas suspeitas em imagens mamográficas. Entre essas estruturas estão os agrupamentos de microcalcificações (clusters), cuja existência é indicativo da necessidade de uma inspeção mais minuciosa no local. As imagens de mamas densas, características principalmente de mulheres jovens, constituem, no entanto, um desafio para esses esquemas devido ao baixo contraste entre as estruturas de interesse e os demais tecidos representados na imagem mamográfica. Nesta pesquisa, portanto, foram feitas investigações sobre as características das imagens radiográficas de mamas densas, a partir das quais foram desenvolvidas técnicas de realce de contraste que, somadas a outras técnicas de processamento digital, proporcionaram um desempenho mais apropriado de um esquema de processamento para detecção de c1usters. Foram ainda elaboradas técnicas que permitam o trabalho com uma resolução de contraste mais adequada, além de outras que consideram alguns dos parâmetros físicos envolvidos na obtenção das imagens e procedimentos para reduzir diagnósticos falsos-positivos. Os resultados registrados nos testes com diferentes conjuntos de imagens de uma base de dados montada para esse projeto indicam que a combinação das técnicas desenvolvidas permite incrementar o desempenho de um esquema de processamento para detectar agrupamentos de microcalcificações, possibilitando a identificação de estruturas em imagens de baixo contraste, não detectadas em processamento convencional antes do realce de contrate. Como efeito, essa investigação mostra a possibilidade de esquemas CAD em mamografia atingir agora desempenho satisfatório na detecção de microcalcificações em imagens de mamas densas. / Computer-aided diagnosis (CAD) schemes have been developed intended to provide information for early detection of breast cancer. Image processing techniques are used in these schemes in order to indicate suspicious structures in mammographic images. Among these structures there are clustered microcalcifications, which usually drive to a more detailed examination in the location where they are. Images corresponding to dense breasts, which are characteristic mainly of young women, are however a challenge to CAD schemes due to the low contrast between the structures of c1inical interest and the other tissues registered on the film. Therefore, investigations were performed in this work on the characteristics of dense breasts radiographic images, from which contrast enhancement techniques were developed. These procedures were joined to other digital processing techniques to provide a better performance for a processing scheme intended to clusters detection. In addition, techniques which allow to work with a more suitable contrast resolution and others which take into account some of the physical parameters involved in the image acquisition process were developed together with a procedure designed to reduce false positive diagnoses. The results obtained during tests with different images sets from a data base developed for this research indicate that combining all the techniques developed here allow to improve the performance of a processing scheme designed to detect microcalcifications clusters, and it also allows to distinguish some of these structures in low contrast images, which were not detected in conventional processing before the contrast enhancement. As consequence, this investigation shows the possibility now for CAD schemes in mammography reaches a better performance in microcalcifications detection in dense breasts images.

A reparação de danos provocados pelo homem aos ecossistemas não é recente no Brasil. Plantações florestais têm sido estabelecidas desde o século XIX sem, no entanto, até recentemente, terem vínculos estreitos com concepções teóricas, sendo executadas normalmente como uma prática de plantio de mudas, com objetivos muito específicos, como controle de erosão, estabilização de taludes, melhoria visual, entre outros. Com o desenvolvimento dos conceitos e teorias ecológicos, sobretudo nos últimos cinqüenta anos, os programas de manejo e restauração florestal têm deixado de ser meras aplicações de práticas agronômicas e silviculturais, para buscar a reconstrução de interações ecológicas da comunidade, ou seja, os projetos de restauração estão deixando de ser o plantio de um amontoado de indivíduos e passando a considerar o potencial de auto-regeneração do ambiente. Este trabalho foi desenvolvido em um sítio degradado da Reserva Legal do aeroporto municipal de Registro (SP), cidade situada no Vale do Ribeira, com o objetivo de avaliar o potencial da transferência de banco de propágulos alóctone em recuperar áreas degradadas. O uso de banco de propágulos como método de recuperação de áreas degradadas é uma prática já bastante utilizada para a recuperação de áreas mineradas em todo o mundo. Nesses casos, o banco é comumente espalhado na área total, normalmente com grande espessura (20 - 50 cm). Neste trabalho foi testada a deposição do banco de propágulos em sulcos, partindo-se da premissa que esta técnica pode permitir que um mesmo volume de banco seja capaz de recuperar uma área degradada maior, aumentando seu rendimento. Assim, na área experimental foi testado o potencial regenerativo do banco de propágulos através de quatro tratamentos: testemunha (Controle), Calcário (Ca), Calcário + Nitrogênio inicial (Ca+Ni) e Calcário + Nitrogênio em cobertura (Ca+Nc). Cada tratamento foi alocado em um sulco de dimensões 0,3m X 0,3m X 20,0m (1,8 m³), e repetido quatro vezes, totalizando 16 sulcos (28,8 m³). A coleta da serapilheira e do solo até uma profundidade de 20 cm, que corresponde ao banco de propágulos aqui estudado, foi feita em 12 parcelas amostrais de dimensão 3,0mX3,0mX0,2m (1,8m³), distribuídas a esmo nas manchas da floresta classificadas em estágios médio e avançado de regeneração, na faixa de floresta condenada à supressão. O solo e a serapilheira da parcela foram revolvidos com enxada. Este material foi acondicionado em sacos de ráfia de 100L. Entretanto, cada saco continha apenas 60L (~70Kg). Para obedecer este padrão, utilizou-se uma lata de tinta de 18L, estabelecendo-se 3,3 latas por saco. Este volume estabeleceu uma média de 21,33 sacos por parcela. Foi coletado um total de 256 sacos. Estes sacos foram distribuídos em 16 sulcos, com 16 sacos por sulco. Em um ano, iniciado em Setembro de 2006, foram coletados dados de emergência, crescimento, recrutamento e mortalidade dos indivíduos arbustivo-arbóreos. Durante os 12 meses de monitoramento, surgiram 538 indivíduos de 40 espécies, apenas desses hábitos, em uma área de 96m². Aos 12 meses, 409 indivíduos estavam vivos. O tratamento Controle foi aquele que apresentou as maiores quantidades de indivíduos e espécies, apresentou menor número de 7 indivíduos mortos e menores taxas de crescimento. Os demais tratamentos indicaram que os insumos favoreceram apenas o crescimento, tendo havido efeito neutro ou negativo sobre os outros parâmetros avaliados. / The relief of the human damage caused on natural ecosystems is not recent in Brazil. Forest plantations have been established since XIX century, although only recently it have been based on a conceptual framework; before that plantations were planned to attend specific purposes, such as eroding control, embankment slopes, visual improvement, among others. The development of ecological concepts and theories, mostly in the last 50 years, promoted relevant changes on restoration projects, which evolved to consider the environment\'s self-regeneration potential. This study was developed at a degraded site within the Legal Reserve of the airport of Registro, SP, located at Ribeira Valley. The main purpose was to evaluate the alochtonous seed bank transference potential as a restoration technique. Transference of alochtonous propagules is an usual technique for mining areas restoration around the world. In these cases the propagules bank (topsoil) is usually spread as a thick layer (20 - 50cm) over the target area. Here the topsoil was spread over sulcos (arroyos, wales), accepting the idea that with this technique a certain amount of topsoil may recover a larger area, enhancing its income/output. Thus, the regenerative potential of the topsoil was tested over four experimental treatments: attester (Control), Calcareous/Limestone (Ca), initial Calcareous + Nitrogen (Ca+Ni) and on surface Calcareous + Nitrogen (Ca+Nc). Each treatment was allocated into 0.3m x 0.3m x 20m (1.8m3) furrows and repeated four times, in a total of 16 furrows (28.8m3). Topsoil was collected at a 20cm depth, sampled by twelve 3.0m x 3.0m x 0.2m (1.8m2) plots, distributed without direction within secondary forest remnants in late sucessional stages, in the logged area. Topsoil was removed with hoe and bagged into 100L raffia bags. Therefore, each bag was fulfilled with 60L (~70Kg) only. To follow the pattern, 18L paint tin box were used to keep an average number of 3,3 paint tin box per bag. This volume correspond to an average of 21.33 sacos per plot. From September 2006 to September 2007, emergence, growth, recruitment and mortality were monitored. A total of 538 individuals (shrubs and trees) of 40 species were recorded, in a total area of 96m2. After 12 twelve months 409 individuals survived. Control treatment presented the highest individuals and species quantities, the lowest growth rates and mortality. The other treatments revealed that fertilizers indulge growth, with neutral or negative effects over demais evaluated parameters.

Enabling machines to solve computer vision tasks with natural language components can greatly improve human interaction with computers. In this thesis, we address vision and language tasks with deep learning methods that explicitly localize relevant visual evidence. Spatial evidence localization in images enhances the interpretability of the model, while temporal localization in video is necessary to remove irrelevant content. We apply our methods to various vision and language tasks, including visual question answering, temporal activity detection, dense video captioning and cross-modal retrieval. First, we tackle the problem of image question answering, which requires the model to predict answers to questions posed about images. We design a memory network with a question-guided spatial attention mechanism which assigns higher weights to regions that are more relevant to the question. The visual evidence used to derive the answer can be shown by visualizing the attention weights in images. We then address the problem of localizing temporal evidence in videos. For most language/vision tasks, only part of the video is relevant to the linguistic component, so we need to detect these relevant events in videos. We propose an end-to-end model for temporal activity detection, which can detect arbitrary length activities by coordinate regression with respect to anchors and contains a proposal stage to filter out background segments, saving computation time. We further extend activity category detection to event captioning, which can express richer semantic meaning compared to a class label. This derives the problem of dense video captioning, which involves two sub-problems: localizing distinct events in long video and generating captions for the localized events. We propose an end-to-end hierarchical captioning model with vision and language context modeling in which the captioning training affects the activity localization. Lastly, the task of text-to-clip video retrieval requires one to localize the specified query instead of detecting and captioning all events. We propose a model based on the early fusion of words and visual features, outperforming standard approaches which embed the whole sentence before performing late feature fusion. Furthermore, we use queries to regulate the proposal network to generate query related proposals. In conclusion, our proposed visual localization mechanism applies across a variety of vision and language tasks and achieves state-of-the-art results. Together with the inference module, our work can contribute to solving other tasks such as video question answering in future research.

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.

Árvores emergentes e de dossel da Floresta Ombrófila Densa Sub-montana de uma Parcela Permanente (10,24ha) no Parque Estadual &quotCarlos Botelho&quot, SP, foram estudadas e descritas morfologicamente considerando padrões de tronco, raízes e casca externa com a finalidade de auxiliar na identificação de campo. O critério de inclusão (DAP igual ou superior a 48cm) abrangeu 324 indivíduos de 64 espécies pertencentes a 36 famílias. São fornecidas descrições, ilustrações e chave de identificação para as espécies estudadas baseadas em características do tronco à altura do peito e da base; das raízes aparentes e rentes ao tronco; das raízes superficiais; e das texturas da casca externa nas regiões da árvore mencionadas. Considerando que não há uma terminologia padrão para descrição dos caracteres estudados, é proposta aqui uma nomenclatura, além de serem utilizados termos propostos por Torres et al. (1994) e Ribeiro et al. (1999), com adaptações para as variações observadas no local de estudo. O tronco foi diferenciado em cilíndrico ou acanalado, com base reta ou dilatada. As raízes variaram entre digitadas e sapopemas. Foram categorizados 13 padrões básicos de textura da casca externa: liso, áspero, sujo, rugoso, fissurado, fendido, lenticelado, reticulado, com placas, escamoso, laminado, variegado e com depressões. O uso de terminologia padronizada na descrição de características de tronco, raízes e casca permite sua aplicação e comparação com dados de estudos desenvolvidos em diferentes regiões, sendo importante ressaltar sua utilização aliada a outros caracteres da planta para uma identificação mais eficiente. / Emergent and canopy trees of lowmontane dense ombrophyllous forest in a permanent plot (10,24ha) in &quotCarlos Botelho&quot State Park, SP, were studied and morphologically described. Trunk, roots and outer bark patterns were considered in order to make field identification easier. A total of 324 trees, 64 species and 36 families were included (DAP &#8805 48cm) and analyzed. We provide descriptions, figures and an identification key based on the following features: trunk at breast high; trunk base; apparent roots; superficial roots; and texture of outer bark. Since there is not a standard morphological nomenclature for these patterns, we propose a nomenclature here, besides using those terms proposed by Torres et al. (1994) and Ribeiro et al. (1999) with some adaptations. Trunk are named terete or angular, straight or with an expanded base. Roots are named digitated or buttressed. Thirteen basic patterns of outer bark texture were named: entire, rough, messy, rugose, fissured, cracked, lenticellate, reticulate, flaky, scaly, laminate, variegated and dippled. Using this terminology allows its application and comparison with data from different areas. They should be used with other features, reproductive and vegetative ones, in order to allow a correct identification.

O presente trabalho tem como tema eficiência de corte dos ciclones de meio denso da planta de beneficiamento da mina de carvão da Vale- Moatize, Moçambique. O foco deste trabalho centra-se na avaliação de eficiência de ciclone de meio denso. As amostras utilizadas nos estudos foram provenientes de ciclones de meio denso da usina da mineradora da Vale- Moatize, de onde foram coletadas amostras em dois ciclones (Dmc1 e Dmc2) sendo que o concentrado do primeiro Ciclone alimenta o segundo ciclone, de onde se obtém dois produtos finais (Carvão metalúrgico e térmico). Este roteiro compreendeu a caracterização, comparação das características de lavabilidade e curvas de Tromp. Estas amostras foram processadas de forma padronizada de modo que os resultados obtidos pudessem ser comparados. As amostras foram secadas a 60°C de modo a retirar o excesso de humidade. Em seguida foram quarteadas e emcaminhadas para o processo de densimétricos. Após o processo densimétrico nas densidades 1,3 á 2,2, seguiu se a análise imediata de cinzas. Com os resultados densimétricos e químicas foi plotado as curvas de separabilidade densimétrica. Quanto aos aspectos de lavabilidade do carvão, apesar da diferença na densidade de corte nos dois ciclones de meio denso, verifica-se uma alta liberação do material orgânico em relação ao material inorgânico nas densidades 1,3; 1,4; e 1,5g/cm3 devido ao processo de formação do carvão e o tempo de carbonificação com que esta matéria orgânica ficou submetida apresentando maiores dificuldades de separação na densidade 1,3 nos dois ciclones. Os parâmetros avaliados foram: densidade de corte (D50), Erro provável médio ( EPM) e Imperfeição (I). Os ciclones mostraram desempenhos satisfatórios. O Dmc1 e Dmc2 apresentaram uma boa eficiência de corte com índice de precisão de 0,02 e 0,05 respectivamente. Portanto mesmo com alimentação bastante variaveis os ciclones apresentaram um densepenho muito similar isto se deve a independéncia da distribuição densimentrica na alimentação. / The present work focuses on cutting efficiency of dense medium cyclones from the Vale- Moatize coal mine beneficiation plant, Mozambique. The focus of this work focuses on the evaluation of dense medium cyclone efficiency. The samples used in the studies were obtained from dense medium cyclones of the Vale-Moatize mining plant, from which samples were collected in two cyclones (Dmc1 and Dmc2) and the concentrate of the first Cyclone feeds the second cyclone, from which one obtains Two final products (metallurgical and thermal coal). This script comprised the characterization, comparison of the washability characteristics and Tromp curves. These samples were processed in a standardized way so that the results obtained could be compared. The samples were dried at 60 ° C to remove excess moisture. Then they were quarteted and headed for the densimetric process. After the densimetric process at densities 1.3 to 2.2, immediate ash analysis was followed. With the densimetric and chemical results the densimetric separability curves were plotted. Regarding the washability aspects of the coal, despite the difference in cut density in the two cyclones of dense medium, there is a high release of the organic material in relation to the inorganic material in the densities 1,3; 1.4; And 1.5g / cm3 due to the process of formation of the coal and the carbonization time with which this organic matter was submitted presenting greater difficulties of separation in the density 1,3 in the two cyclones. The evaluated parameters were: cut density (D50), Mean Probable Error (EPM) and Imperfection (I). Cyclones showed satisfactory performances. The Dmc1 and Dmc2 presented a good cutting efficiency with an accuracy index of 0.02 and 0.05 respectively. Therefore, even with very variable feeding, the cyclones showed a very similar performance, due to the independence of the density distribution in the feed.