A flexible navigation system for autonomous robots operating in heterogeneous environments

Dissertação para obtenção do Grau de Mestre em Engenharia Electrotécnica e de Computadores

Data queries over heterogeneous sources

Dissertação para obtenção do Grau de Mestre em Engenharia Informática

The Edges of Areal Units: a case study in the heterogeneous effects of assessment district edges

Dissertation submitted in partial fulfillment of the requirements for the Degree of Master of Science in Geospatial Technologies

Heterogeneous computing with an algorithmic skeleton framework

The Graphics Processing Unit (GPU) is present in almost every modern day personal computer. Despite its specific purpose design, they have been increasingly used for general computations with very good results. Hence, there is a growing effort from the community to seamlessly integrate this kind of devices in everyday computing. However, to fully exploit the potential of a system comprising GPUs and CPUs, these devices should be presented to the programmer as a single platform. The efficient combination of the power of CPU and GPU devices is highly dependent on each device’s characteristics, resulting in platform specific applications that cannot be ported to different systems. Also, the most efficient work balance among devices is highly dependable on the computations to be performed and respective data sizes. In this work, we propose a solution for heterogeneous environments based on the abstraction level provided by algorithmic skeletons. Our goal is to take full advantage of the power of all CPU and GPU devices present in a system, without the need for different kernel implementations nor explicit work-distribution.To that end, we extended Marrow, an algorithmic skeleton framework for multi-GPUs, to support CPU computations and efficiently balance the work-load between devices. Our approach is based on an offline training execution that identifies the ideal work balance and platform configurations for a given application and input data size. The evaluation of this work shows that the combination of CPU and GPU devices can significantly boost the performance of our benchmarks in the tested environments, when compared to GPU-only executions.

Integrating data from heterogeneous DNA microarray platforms

DNA microarrays are one of the most used technologies for gene expression measurement. However, there are several distinct microarray platforms, from different manufacturers, each with its own measurement protocol, resulting in data that can hardly be compared or directly integrated. Data integration from multiple sources aims to improve the assertiveness of statistical tests, reducing the data dimensionality problem. The integration of heterogeneous DNA microarray platforms comprehends a set of tasks that range from the re-annotation of the features used on gene expression, to data normalization and batch effect elimination. In this work, a complete methodology for gene expression data integration and application is proposed, which comprehends a transcript-based re-annotation process and several methods for batch effect attenuation. The integrated data will be used to select the best feature set and learning algorithm for a brain tumor classification case study. The integration will consider data from heterogeneous Agilent and Affymetrix platforms, collected from public gene expression databases, such as The Cancer Genome Atlas and Gene Expression Omnibus.

Patrones dinámicos de abundancia y dispersión de una plaga en un paisaje heterogéneo. Dynamic patterns of abundance and dispersion of a pest in a heterogeneous landscape.

El desarrollo de conocimiento empírico sobre cómo la heterogeneidad espacial de un paisaje afecta los patrones de movimiento de una especie animal es considerado una prioridad para el manejo y la conservación de las especies y sus hábitats. En el caso de los insectos plaga, estos estudios resultan importantes ya que aportan las bases teóricas y empíricas fundamentales para su manejo. La persistencia de éstas especies en un paisaje modificado depende de la interrelación entre procesos ecológicos y la estructura del paisaje, tales como la interacción entre especies, la disponibilidad de parches hábitat y la influencia de las prácticas de manejo. El análisis de éstos procesos en un agroecosistema permite simplificar los modelos de heterogeneidad espacial, debido a que los lotes de cultivo son internamente homogéneos y los disturbios antropogénicos generalmente ocurren a la escala de parche, permitiendo determinar las respuestas de los insectos a dicha escala. La alfalfa (Medicago sativa) es un recurso fundamental para la producción agropecuaria y en Argentina, es el recurso forrajero más importante, constituyendo la base de la producción ganadera del país. Actualmente se cultivan alrededor de 5 millones de hectáreas, de las cuales un millón se siembran en la provincia de Córdoba. Además, cumple un rol importante en la sustentabilidad de los sistemas de producción por su función de recuperación de la fertilidad y estabilidad edáfica. La isoca de la alfalfa (Colias lesbia) es la plaga principal del cultivo, produciendo en promedio la pérdida de un corte por año. La hipótesis principal de nuestro trabajo es que los patrones de abundancia y movilidad de la isoca de la alfalfa son afectados por la estructura del paisaje y las prácticas de manejo. Los objetivos específicos del proyecto son: (a) Establecer el efecto de la estructura del paisaje y y el manejo del cultivo en la abundancia de los distintos estadios de Colias lesbia. (b) Determinar los patrones de dispersión de Colias lesbia en relación a la heterogeneidad espacial del paisaje (c) Generar un modelo predictivo de la abundancia de Colias lesbia según la estructura espacial del paisaje, el clima y el manejo del cultivo. (d) Desarrollar un conjunto de recomendaciones de manejo a escala regional para el control de la isoca de la alfalfa. Para ello se elegirán lotes de alfalfa en la región este de la provincia de Córdoba, en el departamento de San Justo, donde se realizará un relevamiento inicial del área de estudio y se dialogará con los productores. Paralelamente, se realizará una clasificación supervisada del área de estudio a partir de escenas de imágenes Landsat TM. En los parches seleccionados, durante 3 años y durante los meses de verano, se muestrearán quincenalmente los distintos estadios de Colias lesbia. Se realizarán análisis de correlación y regresión entre las variables independientes (métricas de la configuración y dinámica del paisaje) y las variables dependientes, (abundancia media de los diferentes estadios de las poblaciones). Asimismo, se realizarán experimentos de marcado-liberación-recaptura para determinar cómo el movimiento de la especie depende de la estructura del paisaje. Para modelar el movimiento inherente de la especie se combinará la información obtenida en el campo con un modelo de difusión utilizando métodos bayesianos. Se espera obtener modelos que permitan comprender los mecanismos que generan los patrones observados. Con esta información se propondrán lineamientos generales y específicos para un manejo de la isoca de la alfalfa a escala regional. En tal sentido, se espera aportar información para restringir la dispersión de la plaga, y reducir los costos y perjuicios del control químico que podrían evitarse con la aplicación de prácticas de manejo integrado y de "manejo de área" que minimicen el impacto de la plaga como también contribuir al conocimiento general de la ecología de insectos.

Degradación de efluentes de las industrias del cuero y textil mediante procesos fotocatalíticos heterogéneos empleando la luz solar. Degradation of effluents from the leather and textile industries through heterogeneous photocatalytic processes using sunlight.

El creciente desarrollo de la industria del cuero y textil en nuestro país, y específicamente en la provincia de Córdoba, ha hecho resurgir en los ultimos años una problemática aún no resuelta que es la elevada contaminación de los recursos hídricos. En ambas industrias, la operación de teñido involucra principalmente colorantes de tipo azoico los cuales son "no biodegradables" y se fragmentan liberando aminas aromáticas cancerígenas. Para abordar esta problemática, la fotocatálisis heterogénea aparece como una nueva tecnología que permitiría la completa mineralización de estos colorantes. A través de radiación y un fotocatalizador sólido adecuado se pueden generan radicales libres eficientes para la oxidación de materia orgánica (colorantes) en medio acuoso. En este sentido, se proponen tamices moleculares mesoporosos modificados con metales de transición (MT) como fotocatalizadores potencialmente aptos para la degradación de estos contaminantes. El propósito principal de este proyecto es el diseño, síntesis, caracterización y evaluación de materiales mesoporosos que presenten actividad fotocatalítica ya sea mediante la modificación de su estructura con diversos metales fotosensibles y/o empleándolos como soporte de óxido de titanio. Se pretende evaluar estos materiales en la degradación de colorantes intentando desplazar su fotosensibilidad hacia la radiación visible para desarrollar nuevas tecnologías con menor impacto ambiental y mayor aprovechamiento de la energía solar. Para ello se sintetizarán materiales del tipo MCM-41 modificados con distintos MT tales como Fe, Cr, Co, Ni y Zn mediante incorporación directa del ión metálico o impregnación. Al mismo tiempo, tanto estos últimos materiales como el MCM-41 silíceo serán empleados como soporte de TiO2. Sus propiedades fisicoquímicas se caracterizarán mediante distintas técnicas instrumentales y su actividad fotocatalítica se evaluará en la degradación de colorantes azoicos bajo radiación visible. Se seleccionará el catalizador más eficiente y se estudiarán los diversos factores que afectan el proceso de fotodegradación. Así mismo, el análisis de la concentración del colorante y los productos presentes en el medio en función del tiempo de reacción permitirá inferir sobre la cinética de la decoloración y postular posibles mecanismos de fotodegradación. Con esta propuesta se espera contribuír al desarrollo de un sector industrial importante en nuestra provincia como es el de las industrias del cuero y textil, mediante la generación de nuevas tecnologías que empleen la energía solar para la degradación de sus efluentes (colorantes). En este sentido, se espera desarrollar nuevos materiales optimizados para lograr la mayor eficiencia fotocatalítica. Esto conduciría entonces hacia la remediación de un problema ambiental de alto impacto tanto para nuestra provincia y nuestro país como para la población mundial, como es la contaminación de los recursos hídricos. Finalmente, con este proyecto se contribuirá a la formación de dos doctorandos y un maestrando, cuyos temas de tesis están vinculados con nuestro objeto de estudio. The increasing development of the textile and leather industries in our country, and specifically in Córdoba, has revived an unresolved problem that is the high contamination of water resources. In both industries, the dyeing involves mainly type azoic dyes which are not biodegradable and break releasing carcinogenic aromatic amines. Heterogeneous photocatalysis appears as a new technology that would allow the complete mineralization of these pollutants. Through radiation and a suitable solid it is possible to generate free radicals for efficient oxidation of organic matter (dyes) in aqueous medium. In this respect, mesoporous molecular sieves modified with transition metals are proposed as potential photocatalysts. The main purpose of this project is the synthesis of mesoporous materials having photocatalytic activity for the degradation of dyes. We will try to move their photosensitivity to visible radiation to develop new technologies with lower environmental impact and greater use of solar energy. Materials MCM-41 modified with metals (Fe, Cr, Co, Ni and Zn) will be synthesized by direct incorporation or impregnation. These materials and the siliceous MCM-41 will be then employed as support of TiO2. The materials will be evaluated in the photocatalytic degradation of azoic dyes under visible radiation. The influence of different factors on the photodegradation proccess will be studied. Kinetic studies will be carried out and a possible reaction way will be proposed. Thus, this work will contribute to the advancement of an important industrial sector and the remediation of an environmental problem with high impact for our province and our country. Moreover, this proyect will contribute to the development of two doctoral tesis and one magister tesis which are vinculated with our study subject.

Coalition formation in a contest game with three heterogeneous players

We analyze the incentives for cooperation of three players differing in their efficiency of effort in a contest game. We concentrate on the non-cooperative bargaining foundation of coalition formation, and therefore, we adopt a two-stage model. In the first stage, individuals form coalitions following a bargaining protocol similar to the one proposed by Gul (1989). Afterwards, coalitions play the contest game of Esteban and Ray (1999) within the resulting coalition structure of the first stage. We find that the grand coalition forms whenever the distribution of the bargaining power in the coalition formation game is equal to the distribution of the relative efficiency of effort. Finally, we use the case of equal bargaining power for all individuals to show that other types of coalition structures may be observed as well.

Effects of selection and drift on G matrix evolution in a heterogeneous environment: a multivariate Qst-Fst Test with the freshwater snail Galba truncatula.

Unraveling the effect of selection vs. drift on the evolution of quantitative traits is commonly achieved by one of two methods. Either one contrasts population differentiation estimates for genetic markers and quantitative traits (the Q(st)-F(st) contrast) or multivariate methods are used to study the covariance between sets of traits. In particular, many studies have focused on the genetic variance-covariance matrix (the G matrix). However, both drift and selection can cause changes in G. To understand their joint effects, we recently combined the two methods into a single test (accompanying article by Martin et al.), which we apply here to a network of 16 natural populations of the freshwater snail Galba truncatula. Using this new neutrality test, extended to hierarchical population structures, we studied the multivariate equivalent of the Q(st)-F(st) contrast for several life-history traits of G. truncatula. We found strong evidence of selection acting on multivariate phenotypes. Selection was homogeneous among populations within each habitat and heterogeneous between habitats. We found that the G matrices were relatively stable within each habitat, with proportionality between the among-populations (D) and the within-populations (G) covariance matrices. The effect of habitat heterogeneity is to break this proportionality because of selection for habitat-dependent optima. Individual-based simulations mimicking our empirical system confirmed that these patterns are expected under the selective regime inferred. We show that homogenizing selection can mimic some effect of drift on the G matrix (G and D almost proportional), but that incorporating information from molecular markers (multivariate Q(st)-F(st)) allows disentangling the two effects.

Hydrogeophysical data integration at larger scales: Application of Bayesian sequential simulation for the characterization of heterogeneous alluvial aquifers

Knowledge of the spatial distribution of hydraulic conductivity (K) within an aquifer is critical for reliable predictions of solute transport and the development of effective groundwater management and/or remediation strategies. While core analyses and hydraulic logging can provide highly detailed information, such information is inherently localized around boreholes that tend to be sparsely distributed throughout the aquifer volume. Conversely, larger-scale hydraulic experiments like pumping and tracer tests provide relatively low-resolution estimates of K in the investigated subsurface region. As a result, traditional hydrogeological measurement techniques contain a gap in terms of spatial resolution and coverage, and they are often alone inadequate for characterizing heterogeneous aquifers. Geophysical methods have the potential to bridge this gap. The recent increased interest in the application of geophysical methods to hydrogeological problems is clearly evidenced by the formation and rapid growth of the domain of hydrogeophysics over the past decade (e.g., Rubin and Hubbard, 2005).

A Bootstrap Neural Network Based Heterogeneous Panel Unit Root Test: Application to Exchange Rates

This paper proposes a bootstrap artificial neural network based panel unit root test in a dynamic heterogeneous panel context. An application to a panel of bilateral real exchange rate series with the US Dollar from the 20 major OECD countries is provided to investigate the Purchase Power Parity (PPP). The combination of neural network and bootstrapping significantly changes the findings of the economic study in favour of PPP.

Monetary Policy and Heterogeneous Expectations

This paper studies the implications for monetary policy of heterogeneous expectations in a New Keynesian model. The assumption of rational expectations is replaced with parsimonious forecasting models where agents select between predictors that are underparameterized. In a Misspecification Equilibrium agents only select the best-performing statistical models. We demonstrate that, even when monetary policy rules satisfy the Taylor principle by adjusting nominal interest rates more than one for one with inflation, there may exist equilibria with Intrinsic Heterogeneity. Under certain conditions, there may exist multiple misspecification equilibria. We show that these findings have important implications for business cycle dynamics and for the design of monetary policy.

Heterogeneous T-cell response to MAGE-A10(254-262): high avidity-specific cytolytic T lymphocytes show superior antitumor activity.

MAGE-encoded antigens, which are expressed by tumors of many histological types but not in normal tissues, are suitable candidates for vaccine-based immunotherapy of cancers. Thus far, however, T-cell responses to MAGE antigens have been detected only occasionally in cancer patients. In contrast, by using HLA/peptide fluorescent tetramers, we have observed recently that CD8(+) T cells specific for peptide MAGE-A10(254-262) can be detected frequently in peptide-stimulated peripheral blood mononuclear cells from HLA-A2-expressing melanoma patients and healthy donors. On the basis of these results, antitumoral vaccination trials using peptide MAGE-A10(254-262) have been implemented recently. In the present study, we have characterized MAGE-A10(254-262)-specific CD8(+) T cells in polyclonal cultures and at the clonal level. The results indicate that the repertoire of MAGE-A10(254-262)-specific CD8(+) T cells is diverse both in terms of clonal composition, efficiency of peptide recognition, and tumor-specific lytic activity. Importantly, only CD8(+) T cells able to recognize the antigenic peptide with high efficiency are able to lyse MAGE-A10-expressing tumor cells. Under defined experimental conditions, the tetramer staining intensity exhibited by MAGE-A10(254-262)-specific CD8(+) T cells correlates with efficiency of peptide recognition so that "high" and "low" avidity cells can be separated by FACS. Altogether, the data reported here provide evidence for functional diversity of MAGE-A10(254-262)-specific T cells and will be instrumental for the monitoring of peptide MAGE-A10(254-262)-based clinical trials.

Sharp Bounds on Heterogeneous Individual Treatment Responses

This paper discusses how to identify individual-specific causal effects of an ordered discrete endogenous variable. The counterfactual heterogeneous causal information is recovered by identifying the partial differences of a structural relation. The proposed refutable nonparametric local restrictions exploit the fact that the pattern of endogeneity may vary across the level of the unobserved variable. The restrictions adopted in this paper impose a sense of order to an unordered binary endogeneous variable. This allows for a uni.ed structural approach to studying various treatment effects when self-selection on unobservables is present. The usefulness of the identi.cation results is illustrated using the data on the Vietnam-era veterans. The empirical findings reveal that when other observable characteristics are identical, military service had positive impacts for individuals with low (unobservable) earnings potential, while it had negative impacts for those with high earnings potential. This heterogeneity would not be detected by average effects which would underestimate the actual effects because different signs would be cancelled out. This partial identification result can be used to test homogeneity in response. When homogeneity is rejected, many parameters based on averages may deliver misleading information.

Heterogeneous dynamics, aggregation and the persistence of economic shocks

It has been recently emphasized that, if individuals have heterogeneous dynamics, estimates of shock persistence based on aggregate data are significatively higher than those derived from its disaggregate counterpart. However, a careful examination of the implications of this statement on the various tools routinely employed to measure persistence is missing in the literature. This paper formally examines this issue. We consider a disaggregate linear model with heterogeneous dynamics and compare the values of several measures of persistence across aggregation levels. Interestingly, we show that the average persistence of aggregate shocks, as measured by the impulse response function (IRF) of the aggregate model or by the average of the individual IRFs, is identical on all horizons. This result remains true even in situations where the units are (short-memory) stationary but the aggregate process is long-memory or even nonstationary. In contrast, other popular persistence measures, such as the sum of the autoregressive coefficients or the largest autoregressive root, tend to be higher the higher the aggregation level. We argue, however, that this should be seen more as an undesirable property of these measures than as evidence of different average persistence across aggregation levels. The results are illustrated in an application using U.S. inflation data.