A generalization of the optimal diagonal approximate inverse preconditioner

[EN ]The classical optimal (in the Frobenius sense) diagonal preconditioner for large sparse linear systems Ax = b is generalized and improved. The new proposed approximate inverse preconditioner N is based on the minimization of the Frobenius norm of the residual matrix AM − I, where M runs over a certain linear subspace of n × n real matrices, defined by a prescribed sparsity pattern. The number of nonzero entries of the n×n preconditioning matrix N is less than or equal to 2n, and n of them are selected as the optimal positions in each of the n columns of matrix N. All theoretical results are justified in detail…

Heterogeneous multicore systems for signal processing

This thesis explores the capabilities of heterogeneous multi-core systems, based on multiple Graphics Processing Units (GPUs) in a standard desktop framework. Multi-GPU accelerated desk side computers are an appealing alternative to other high performance computing (HPC) systems: being composed of commodity hardware components fabricated in large quantities, their price-performance ratio is unparalleled in the world of high performance computing. Essentially bringing “supercomputing to the masses”, this opens up new possibilities for application fields where investing in HPC resources had been considered unfeasible before. One of these is the field of bioelectrical imaging, a class of medical imaging technologies that occupy a low-cost niche next to million-dollar systems like functional Magnetic Resonance Imaging (fMRI). In the scope of this work, several computational challenges encountered in bioelectrical imaging are tackled with this new kind of computing resource, striving to help these methods approach their true potential. Specifically, the following main contributions were made: Firstly, a novel dual-GPU implementation of parallel triangular matrix inversion (TMI) is presented, addressing an crucial kernel in computation of multi-mesh head models of encephalographic (EEG) source localization. This includes not only a highly efficient implementation of the routine itself achieving excellent speedups versus an optimized CPU implementation, but also a novel GPU-friendly compressed storage scheme for triangular matrices. Secondly, a scalable multi-GPU solver for non-hermitian linear systems was implemented. It is integrated into a simulation environment for electrical impedance tomography (EIT) that requires frequent solution of complex systems with millions of unknowns, a task that this solution can perform within seconds. In terms of computational throughput, it outperforms not only an highly optimized multi-CPU reference, but related GPU-based work as well. Finally, a GPU-accelerated graphical EEG real-time source localization software was implemented. Thanks to acceleration, it can meet real-time requirements in unpreceeded anatomical detail running more complex localization algorithms. Additionally, a novel implementation to extract anatomical priors from static Magnetic Resonance (MR) scansions has been included.

Sea Level Variations from Decades to Millennia: A few Case Studies

The main goals of this Ph.D. study are to investigate the regional and global geophysical components related to present polar ice melting and to provide independent cross validation checks of GIA models using both geophysical data detected by satellite mission, and geological observations from far field sites, in order to determine a lower and upper bound of uncertainty of GIA effect. The subject of this Thesis is the sea level change from decades to millennia scale. Within ice2sea collaboration, we developed a Fortran numerical code to analyze the local short-term sea level change and vertical deformation resulting from the loss of ice mass. This method is used to investigate polar regions: Greenland and Antarctica. We have used mass balance based on ICESat data for Greenland ice sheet and a plausible mass balance for Antarctic ice sheet. We have determined the regional and global fingerprint of sea level variations, vertical deformations of the solid surface of the Earth and variations of shape of the geoid for each ice source mentioned above. The coastal areas are affected by the long wavelength component of GIA process. Hence understanding the response of the Earth to loading is crucial in various contexts. Based on the hypothesis that Earth mantle materials obey to a linear rheology, and that the physical parameters of this rheology can be only characterized by their depth dependence, we investigate the Glacial Isostatic Effect upon the far field sites of Mediterranean area using an improved SELEN program. We presented new and revised observations for archaeological fish tanks located along the Tyrrhenian and Adriatic coast of Italy and new RSL for the SE Tunisia. Spatial and temporal variations of the Holocene sea levels studied in central Italy and Tunisia, provided important constraints on the melting history of the major ice sheets.

Spectral estimates and preconditioning for saddle point systems arising from optimization problems

In this thesis, we consider the problem of solving large and sparse linear systems of saddle point type stemming from optimization problems. The focus of the thesis is on iterative methods, and new preconditioning srategies are proposed, along with novel spectral estimtates for the matrices involved.

Visualization and validation of (Q)SAR models

Analyzing and modeling relationships between the structure of chemical compounds, their physico-chemical properties, and biological or toxic effects in chemical datasets is a challenging task for scientific researchers in the field of cheminformatics. Therefore, (Q)SAR model validation is essential to ensure future model predictivity on unseen compounds. Proper validation is also one of the requirements of regulatory authorities in order to approve its use in real-world scenarios as an alternative testing method. However, at the same time, the question of how to validate a (Q)SAR model is still under discussion. In this work, we empirically compare a k-fold cross-validation with external test set validation. The introduced workflow allows to apply the built and validated models to large amounts of unseen data, and to compare the performance of the different validation approaches. Our experimental results indicate that cross-validation produces (Q)SAR models with higher predictivity than external test set validation and reduces the variance of the results. Statistical validation is important to evaluate the performance of (Q)SAR models, but does not support the user in better understanding the properties of the model or the underlying correlations. We present the 3D molecular viewer CheS-Mapper (Chemical Space Mapper) that arranges compounds in 3D space, such that their spatial proximity reflects their similarity. The user can indirectly determine similarity, by selecting which features to employ in the process. The tool can use and calculate different kinds of features, like structural fragments as well as quantitative chemical descriptors. Comprehensive functionalities including clustering, alignment of compounds according to their 3D structure, and feature highlighting aid the chemist to better understand patterns and regularities and relate the observations to established scientific knowledge. Even though visualization tools for analyzing (Q)SAR information in small molecule datasets exist, integrated visualization methods that allows for the investigation of model validation results are still lacking. We propose visual validation, as an approach for the graphical inspection of (Q)SAR model validation results. New functionalities in CheS-Mapper 2.0 facilitate the analysis of (Q)SAR information and allow the visual validation of (Q)SAR models. The tool enables the comparison of model predictions to the actual activity in feature space. Our approach reveals if the endpoint is modeled too specific or too generic and highlights common properties of misclassified compounds. Moreover, the researcher can use CheS-Mapper to inspect how the (Q)SAR model predicts activity cliffs. The CheS-Mapper software is freely available at http://ches-mapper.org.

Informatics for cross-sample analysis with comprehensive two-dimensional gas chromatography and high-resolution mass spectrometry (GCxGC-HRMS)

This paper describes informatics for cross-sample analysis with comprehensive two-dimensional gas chromatography (GCxGC) and high-resolution mass spectrometry (HRMS). GCxGC-HRMS analysis produces large data sets that are rich with information, but highly complex. The size of the data and volume of information requires automated processing for comprehensive cross-sample analysis, but the complexity poses a challenge for developing robust methods. The approach developed here analyzes GCxGC-HRMS data from multiple samples to extract a feature template that comprehensively captures the pattern of peaks detected in the retention-times plane. Then, for each sample chromatogram, the template is geometrically transformed to align with the detected peak pattern and generate a set of feature measurements for cross-sample analyses such as sample classification and biomarker discovery. The approach avoids the intractable problem of comprehensive peak matching by using a few reliable peaks for alignment and peak-based retention-plane windows to define comprehensive features that can be reliably matched for cross-sample analysis. The informatics are demonstrated with a set of 18 samples from breast-cancer tumors, each from different individuals, six each for Grades 1-3. The features allow classification that matches grading by a cancer pathologist with 78% success in leave-one-out cross-validation experiments. The HRMS signatures of the features of interest can be examined for determining elemental compositions and identifying compounds.

Late Holocene air temperature variability reconstructed from the sediments of Laguna Escondida, Patagonia, Chile (45 degrees 30 ' S)

Climate and environmental reconstructions from natural archives are important for the interpretation of current climatic change. Few quantitative high-resolution reconstructions exist for South America which is the only land mass extending from the tropics to the southern high latitudes at 56°S. We analyzed sediment cores from two adjacent lakes in Northern Chilean Patagonia, Lago Castor (45°36′S, 71°47′W) and Laguna Escondida (45°31′S, 71°49′W). Radiometric dating (210Pb, 137Cs, 14C-AMS) suggests that the cores reach back to c. 900 BC (Laguna Escondida) and c. 1900 BC (Lago Castor). Both lakes show similarities and reproducibility in sedimentation rate changes and tephra layer deposition. We found eight macroscopic tephras (0.2–5.5 cm thick) dated at 1950 BC, 1700 BC, at 300 BC, 50 BC, 90 AD, 160 AD, 400 AD and at 900 AD. These can be used as regional time-synchronous stratigraphic markers. The two thickest tephras represent known well-dated explosive eruptions of Hudson volcano around 1950 and 300 BC. Biogenic silica flux revealed in both lakes a climate signal and correlation with annual temperature reanalysis data (calibration 1900–2006 AD; Lago Castor r = 0.37; Laguna Escondida r = 0.42, seven years filtered data). We used a linear inverse regression plus scaling model for calibration and leave-one-out cross-validation (RMSEv = 0.56 °C) to reconstruct sub decadal-scale temperature variability for Laguna Escondida back to AD 400. The lower part of the core from Laguna Escondida prior to AD 400 and the core of Lago Castor are strongly influenced by primary and secondary tephras and, therefore, not used for the temperature reconstruction. The temperature reconstruction from Laguna Escondida shows cold conditions in the 5th century (relative to the 20th century mean), warmer temperatures from AD 600 to AD 1150 and colder temperatures from AD 1200 to AD 1450. From AD 1450 to AD 1700 our reconstruction shows a period with stronger variability and on average higher values than the 20th century mean. Until AD 1900 the temperature values decrease but stay slightly above the 20th century mean. Most of the centennial-scale features are reproduced in the few other natural climate archives in the region. The early onset of cool conditions from c. AD 1200 onward seems to be confirmed for this region.

Application and comparison of classification algorithms for recognition of Alzheimer's disease in electrical brain activity (EEG)

The early detection of subjects with probable Alzheimer's disease (AD) is crucial for effective appliance of treatment strategies. Here we explored the ability of a multitude of linear and non-linear classification algorithms to discriminate between the electroencephalograms (EEGs) of patients with varying degree of AD and their age-matched control subjects. Absolute and relative spectral power, distribution of spectral power, and measures of spatial synchronization were calculated from recordings of resting eyes-closed continuous EEGs of 45 healthy controls, 116 patients with mild AD and 81 patients with moderate AD, recruited in two different centers (Stockholm, New York). The applied classification algorithms were: principal component linear discriminant analysis (PC LDA), partial least squares LDA (PLS LDA), principal component logistic regression (PC LR), partial least squares logistic regression (PLS LR), bagging, random forest, support vector machines (SVM) and feed-forward neural network. Based on 10-fold cross-validation runs it could be demonstrated that even tough modern computer-intensive classification algorithms such as random forests, SVM and neural networks show a slight superiority, more classical classification algorithms performed nearly equally well. Using random forests classification a considerable sensitivity of up to 85% and a specificity of 78%, respectively for the test of even only mild AD patients has been reached, whereas for the comparison of moderate AD vs. controls, using SVM and neural networks, values of 89% and 88% for sensitivity and specificity were achieved. Such a remarkable performance proves the value of these classification algorithms for clinical diagnostics.

A coronary heart disease risk model for predicting the effect of potent antiretroviral therapy in HIV-1 infected men

BACKGROUND: Many HIV-infected patients on highly active antiretroviral therapy (HAART) experience metabolic complications including dyslipidaemia and insulin resistance, which may increase their coronary heart disease (CHD) risk. We developed a prognostic model for CHD tailored to the changes in risk factors observed in patients starting HAART. METHODS: Data from five cohort studies (British Regional Heart Study, Caerphilly and Speedwell Studies, Framingham Offspring Study, Whitehall II) on 13,100 men aged 40-70 and 114,443 years of follow up were used. CHD was defined as myocardial infarction or death from CHD. Model fit was assessed using the Akaike Information Criterion; generalizability across cohorts was examined using internal-external cross-validation. RESULTS: A parametric model based on the Gompertz distribution generalized best. Variables included in the model were systolic blood pressure, total cholesterol, high-density lipoprotein cholesterol, triglyceride, glucose, diabetes mellitus, body mass index and smoking status. Compared with patients not on HAART, the estimated CHD hazard ratio (HR) for patients on HAART was 1.46 (95% CI 1.15-1.86) for moderate and 2.48 (95% CI 1.76-3.51) for severe metabolic complications. CONCLUSIONS: The change in the risk of CHD in HIV-infected men starting HAART can be estimated based on typical changes in risk factors, assuming that HRs estimated using data from non-infected men are applicable to HIV-infected men. Based on this model the risk of CHD is likely to increase, but increases may often be modest, and could be offset by lifestyle changes.

Use of hydroclimatic forecasts for improved water management in central Texas

Accurate seasonal to interannual streamflow forecasts based on climate information are critical for optimal management and operation of water resources systems. Considering most water supply systems are multipurpose, operating these systems to meet increasing demand under the growing stresses of climate variability and climate change, population and economic growth, and environmental concerns could be very challenging. This study was to investigate improvement in water resources systems management through the use of seasonal climate forecasts. Hydrological persistence (streamflow and precipitation) and large-scale recurrent oceanic-atmospheric patterns such as the El Niño/Southern Oscillation (ENSO), Pacific Decadal Oscillation (PDO), North Atlantic Oscillation (NAO), the Atlantic Multidecadal Oscillation (AMO), the Pacific North American (PNA), and customized sea surface temperature (SST) indices were investigated for their potential to improve streamflow forecast accuracy and increase forecast lead-time in a river basin in central Texas. First, an ordinal polytomous logistic regression approach is proposed as a means of incorporating multiple predictor variables into a probabilistic forecast model. Forecast performance is assessed through a cross-validation procedure, using distributions-oriented metrics, and implications for decision making are discussed. Results indicate that, of the predictors evaluated, only hydrologic persistence and Pacific Ocean sea surface temperature patterns associated with ENSO and PDO provide forecasts which are statistically better than climatology. Secondly, a class of data mining techniques, known as tree-structured models, is investigated to address the nonlinear dynamics of climate teleconnections and screen promising probabilistic streamflow forecast models for river-reservoir systems. Results show that the tree-structured models can effectively capture the nonlinear features hidden in the data. Skill scores of probabilistic forecasts generated by both classification trees and logistic regression trees indicate that seasonal inflows throughout the system can be predicted with sufficient accuracy to improve water management, especially in the winter and spring seasons in central Texas. Lastly, a simplified two-stage stochastic economic-optimization model was proposed to investigate improvement in water use efficiency and the potential value of using seasonal forecasts, under the assumption of optimal decision making under uncertainty. Model results demonstrate that incorporating the probabilistic inflow forecasts into the optimization model can provide a significant improvement in seasonal water contract benefits over climatology, with lower average deficits (increased reliability) for a given average contract amount, or improved mean contract benefits for a given level of reliability compared to climatology. The results also illustrate the trade-off between the expected contract amount and reliability, i.e., larger contracts can be signed at greater risk.

Population transcriptomics of life-history variation in the genus Salmo

In this study, we demonstrate the power of applying complementary DNA (cDNA) microarray technology to identifying candidate loci that exhibit subtle differences in expression levels associated with a complex trait in natural populations of a nonmodel organism. Using a highly replicated experimental design involving 180 cDNA microarray experiments, we measured gene-expression levels from 1098 transcript probes in 90 individuals originating from six brown trout (Salmo trutta) and one Atlantic salmon (Salmo salar) population, which follow either a migratory or a sedentary life history. We identified several candidate genes associated with preparatory adaptations to different life histories in salmonids, including genes encoding for transaldolase 1, constitutive heat-shock protein HSC70-1 and endozepine. Some of these genes clustered into functional groups, providing insight into the physiological pathways potentially involved in the expression of life-history related phenotypic differences. Such differences included the down-regulation of genes involved in the respiratory system of future migratory individuals. In addition, we used linear discriminant analysis to identify a set of 12 genes that correctly classified immature individuals as migratory or sedentary with high accuracy. Using the expression levels of these 12 genes, 17 out of 18 individuals used for cross-validation were correctly assigned to their respective life-history phenotype. Finally, we found various candidate genes associated with physiological changes that are likely to be involved in preadaptations to seawater in anadromous populations of the genus Salmo, one of which was identified to encode for nucleophosmin 1. Our findings thus provide new molecular insights into salmonid life-history variation, opening new perspectives in the study of this complex trait.

FACTS: Fully Automatic CT Segmentation of a Hip Joint

Extraction of surface models of a hip joint from CT data is a pre-requisite step for computer assisted diagnosis and planning (CADP) of periacetabular osteotomy (PAO). Most of existing CADP systems are based on manual segmentation, which is time-consuming and hard to achieve reproducible results. In this paper, we present a Fully Automatic CT Segmentation (FACTS) approach to simultaneously extract both pelvic and femoral models. Our approach works by combining fast random forest (RF) regression based landmark detection, multi-atlas based segmentation, with articulated statistical shape model (aSSM) based fitting. The two fundamental contributions of our approach are: (1) an improved fast Gaussian transform (IFGT) is used within the RF regression framework for a fast and accurate landmark detection, which then allows for a fully automatic initialization of the multi-atlas based segmentation; and (2) aSSM based fitting is used to preserve hip joint structure and to avoid penetration between the pelvic and femoral models. Taking manual segmentation as the ground truth, we evaluated the present approach on 30 hip CT images (60 hips) with a 6-fold cross validation. When the present approach was compared to manual segmentation, a mean segmentation accuracy of 0.40, 0.36, and 0.36 mm was found for the pelvis, the left proximal femur, and the right proximal femur, respectively. When the models derived from both segmentations were used to compute the PAO diagnosis parameters, a difference of 2.0 ± 1.5°, 2.1 ± 1.6°, and 3.5 ± 2.3% were found for anteversion, inclination, and acetabular coverage, respectively. The achieved accuracy is regarded as clinically accurate enough for our target applications.

Bedform measurements from the San Francisco Bay Coastal System

The morphology of ~45,000 bedforms from 13 multibeam bathymetry surveys was used as a proxy for identifying net bedload sediment transport directions and pathways throughout the San Francisco Bay estuary and adjacent outer coast. The spatially-averaged shape asymmetry of the bedforms reveals distinct pathways of ebb and flood transport. Additionally, the region-wide, ebb-oriented asymmetry of 5% suggests net seaward-directed transport within the estuarine-coastal system, with significant seaward asymmetry at the mouth of San Francisco Bay (11%), through the northern reaches of the Bay (7-8%), and among the largest bedforms (21% for lambda > 50 m). This general indication for the net transport of sand to the open coast strongly suggests that anthropogenic removal of sediment from the estuary, particularly along clearly defined seaward transport pathways, will limit the supply of sand to chronically eroding, open-coast beaches. The bedform asymmetry measurements significantly agree (up to ~ 76%) with modeled annual residual transport directions derived from a hydrodynamically-calibrated numerical model, and the orientation of adjacent, flow-sculpted seafloor features such as mega-flute structures, providing a comprehensive validation of the technique. The methods described in this paper to determine well-defined, cross-validated sediment transport pathways can be applied to estuarine-coastal systems globally where bedforms are present. The results can inform and improve regional sediment management practices to more efficiently utilize often limited sediment resources and mitigate current and future sediment supply-related impacts.

X-parameters based analytical design of non-linear microwave circuits Application to oscillator design

Resumen El diseño clásico de circuitos de microondas se basa fundamentalmente en el uso de los parámetros s, debido a su capacidad para caracterizar de forma exitosa el comportamiento de cualquier circuito lineal. La relación existente entre los parámetros s con los sistemas de medida actuales y con las herramientas de simulación lineal han facilitado su éxito y su uso extensivo tanto en el diseño como en la caracterización de circuitos y subsistemas de microondas. Sin embargo, a pesar de la gran aceptación de los parámetros s en la comunidad de microondas, el principal inconveniente de esta formulación reside en su limitación para predecir el comportamiento de sistemas no lineales reales. En la actualidad, uno de los principales retos de los diseñadores de microondas es el desarrollo de un contexto análogo que permita integrar tanto el modelado no lineal, como los sistemas de medidas de gran señal y los entornos de simulación no lineal, con el objetivo de extender las capacidades de los parámetros s a regímenes de operación en gran señal y por tanto, obtener una infraestructura que permita tanto la caracterización como el diseño de circuitos no lineales de forma fiable y eficiente. De acuerdo a esta filosofía, en los últimos años se han desarrollado diferentes propuestas como los parámetros X, de Agilent Technologies, o el modelo de Cardiff que tratan de proporcionar esta plataforma común en el ámbito de gran señal. Dentro de este contexto, uno de los objetivos de la presente Tesis es el análisis de la viabilidad del uso de los parámetros X en el diseño y simulación de osciladores para transceptores de microondas. Otro aspecto relevante en el análisis y diseño de circuitos lineales de microondas es la disposición de métodos analíticos sencillos, basados en los parámetros s del transistor, que permitan la obtención directa y rápida de las impedancias de carga y fuente necesarias para cumplir las especificaciones de diseño requeridas en cuanto a ganancia, potencia de salida, eficiencia o adaptación de entrada y salida, así como la determinación analítica de parámetros de diseño clave como el factor de estabilidad o los contornos de ganancia de potencia. Por lo tanto, el desarrollo de una formulación de diseño analítico, basada en los parámetros X y similar a la existente en pequeña señal, permitiría su uso en aplicaciones no lineales y supone un nuevo reto que se va a afrontar en este trabajo. Por tanto, el principal objetivo de la presente Tesis consistiría en la elaboración de una metodología analítica basada en el uso de los parámetros X para el diseño de circuitos no lineales que jugaría un papel similar al que juegan los parámetros s en el diseño de circuitos lineales de microondas. Dichos métodos de diseño analíticos permitirían una mejora significativa en los actuales procedimientos de diseño disponibles en gran señal, así como una reducción considerable en el tiempo de diseño, lo que permitiría la obtención de técnicas mucho más eficientes. Abstract In linear world, classical microwave circuit design relies on the s-parameters due to its capability to successfully characterize the behavior of any linear circuit. Thus the direct use of s-parameters in measurement systems and in linear simulation analysis tools, has facilitated its extensive use and success in the design and characterization of microwave circuits and subsystems. Nevertheless, despite the great success of s-parameters in the microwave community, the main drawback of this formulation is its limitation in the behavior prediction of real non-linear systems. Nowadays, the challenge of microwave designers is the development of an analogue framework that allows to integrate non-linear modeling, large-signal measurement hardware and non-linear simulation environment in order to extend s-parameters capabilities to non-linear regimen and thus, provide the infrastructure for non-linear design and test in a reliable and efficient way. Recently, different attempts with the aim to provide this common platform have been introduced, as the Cardiff approach and the Agilent X-parameters. Hence, this Thesis aims to demonstrate the X-parameter capability to provide this non-linear design and test framework in CAD-based oscillator context. Furthermore, the classical analysis and design of linear microwave transistorbased circuits is based on the development of simple analytical approaches, involving the transistor s-parameters, that are able to quickly provide an analytical solution for the input/output transistor loading conditions as well as analytically determine fundamental parameters as the stability factor, the power gain contours or the input/ output match. Hence, the development of similar analytical design tools that are able to extend s-parameters capabilities in small-signal design to non-linear ap- v plications means a new challenge that is going to be faced in the present work. Therefore, the development of an analytical design framework, based on loadindependent X-parameters, constitutes the core of this Thesis. These analytical nonlinear design approaches would enable to significantly improve current large-signal design processes as well as dramatically decrease the required design time and thus, obtain more efficient approaches.